[{"sample_id": 1, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "edit some mountains in the background"}], "image": ["MagicBrush/images/1.jpg", "MagicBrush/images/2.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 2, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put the zebras next to a river"}], "image": ["MagicBrush/images/3.jpg", "MagicBrush/images/4.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 3, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. edit some mountains in the background\n2. put the zebras next to a river"}], "image": ["MagicBrush/images/5.jpg", "MagicBrush/images/6.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 4, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. edit some mountains in the background\n2. put the zebras next to a river"}], "image": ["MagicBrush/images/7.jpg", "MagicBrush/images/8.jpg", "MagicBrush/images/9.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 5, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the piece of paper hanging on the wall a mirror"}], "image": ["MagicBrush/images/10.jpg", "MagicBrush/images/11.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 6, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put an Easter basket on the desk"}], "image": ["MagicBrush/images/12.jpg", "MagicBrush/images/13.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 7, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the piece of paper hanging on the wall a mirror\n2. put an Easter basket on the desk"}], "image": ["MagicBrush/images/14.jpg", "MagicBrush/images/15.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 8, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the piece of paper hanging on the wall a mirror\n2. put an Easter basket on the desk"}], "image": ["MagicBrush/images/16.jpg", "MagicBrush/images/17.jpg", "MagicBrush/images/18.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 9, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if the man was bald"}], "image": ["MagicBrush/images/19.jpg", "MagicBrush/images/20.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 10, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "what if he had a angry mouth?"}], "image": ["MagicBrush/images/21.jpg", "MagicBrush/images/22.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 11, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. What if the man was bald\n2. what if he had a angry mouth?"}], "image": ["MagicBrush/images/23.jpg", "MagicBrush/images/24.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 12, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. What if the man was bald\n2. what if he had a angry mouth?"}], "image": ["MagicBrush/images/25.jpg", "MagicBrush/images/26.jpg", "MagicBrush/images/27.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 13, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be a dolphin jumping out of the water"}], "image": ["MagicBrush/images/28.jpg", "MagicBrush/images/29.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 14, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "replace the baseball bat for a laser sword"}], "image": ["MagicBrush/images/30.jpg", "MagicBrush/images/31.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 15, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "switch to a robot wearing armor"}], "image": ["MagicBrush/images/32.jpg", "MagicBrush/images/33.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 16, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the baseball bat for a laser sword\n2. switch to a robot wearing armor"}], "image": ["MagicBrush/images/34.jpg", "MagicBrush/images/35.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 17, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the baseball bat for a laser sword\n2. switch to a robot wearing armor"}], "image": ["MagicBrush/images/36.jpg", "MagicBrush/images/37.jpg", "MagicBrush/images/38.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 18, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the cow smile"}], "image": ["MagicBrush/images/39.jpg", "MagicBrush/images/40.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 19, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the cow lift its leg up"}], "image": ["MagicBrush/images/41.jpg", "MagicBrush/images/42.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 20, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be an old Greek monument"}], "image": ["MagicBrush/images/43.jpg", "MagicBrush/images/44.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 21, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the cow smile\n2. make the cow lift its leg up\n3. let there be an old Greek monument"}], "image": ["MagicBrush/images/45.jpg", "MagicBrush/images/46.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 22, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the cow smile\n2. make the cow lift its leg up\n3. let there be an old Greek monument"}], "image": ["MagicBrush/images/47.jpg", "MagicBrush/images/48.jpg", "MagicBrush/images/49.jpg", "MagicBrush/images/50.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 23, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the dog for a cat"}], "image": ["MagicBrush/images/51.jpg", "MagicBrush/images/52.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 24, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the background to a river"}], "image": ["MagicBrush/images/53.jpg", "MagicBrush/images/54.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 25, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the dog for a cat\n2. change the background to a river"}], "image": ["MagicBrush/images/55.jpg", "MagicBrush/images/56.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 26, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the dog for a cat\n2. change the background to a river"}], "image": ["MagicBrush/images/57.jpg", "MagicBrush/images/58.jpg", "MagicBrush/images/59.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 27, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "swap the cupcake with a piece of cake"}], "image": ["MagicBrush/images/60.jpg", "MagicBrush/images/61.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 28, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn the cell phone into a banana"}], "image": ["MagicBrush/images/62.jpg", "MagicBrush/images/63.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 29, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the cupcake with a piece of cake\n2. turn the cell phone into a banana"}], "image": ["MagicBrush/images/64.jpg", "MagicBrush/images/65.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 30, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the cupcake with a piece of cake\n2. turn the cell phone into a banana"}], "image": ["MagicBrush/images/66.jpg", "MagicBrush/images/67.jpg", "MagicBrush/images/68.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 31, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the faucet be turned on."}], "image": ["MagicBrush/images/69.jpg", "MagicBrush/images/70.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 32, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "have the person jump over a tennis ball."}], "image": ["MagicBrush/images/71.jpg", "MagicBrush/images/72.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 33, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the person swing the racquet between his legs"}], "image": ["MagicBrush/images/73.jpg", "MagicBrush/images/74.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 34, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "place a bag on the court"}], "image": ["MagicBrush/images/75.jpg", "MagicBrush/images/76.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 35, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. have the person jump over a tennis ball.\n2. Have the person swing the racquet between his legs\n3. place a bag on the court"}], "image": ["MagicBrush/images/77.jpg", "MagicBrush/images/78.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 36, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. have the person jump over a tennis ball.\n2. Have the person swing the racquet between his legs\n3. place a bag on the court"}], "image": ["MagicBrush/images/79.jpg", "MagicBrush/images/80.jpg", "MagicBrush/images/81.jpg", "MagicBrush/images/82.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 37, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a cruise ship to the ocean."}], "image": ["MagicBrush/images/83.jpg", "MagicBrush/images/84.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 38, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the yellow flowers"}], "image": ["MagicBrush/images/85.jpg", "MagicBrush/images/86.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 39, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a rat sniff the broccoli"}], "image": ["MagicBrush/images/87.jpg", "MagicBrush/images/88.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 40, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a plate of meat"}], "image": ["MagicBrush/images/89.jpg", "MagicBrush/images/90.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 41, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the yellow flowers\n2. let a rat sniff the broccoli\n3. add a plate of meat"}], "image": ["MagicBrush/images/91.jpg", "MagicBrush/images/92.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 42, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the yellow flowers\n2. let a rat sniff the broccoli\n3. add a plate of meat"}], "image": ["MagicBrush/images/93.jpg", "MagicBrush/images/94.jpg", "MagicBrush/images/95.jpg", "MagicBrush/images/96.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 43, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "the ocean should have waves."}], "image": ["MagicBrush/images/97.jpg", "MagicBrush/images/98.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 44, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the cat to a rat."}], "image": ["MagicBrush/images/99.jpg", "MagicBrush/images/100.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 45, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the suitcase red."}], "image": ["MagicBrush/images/101.jpg", "MagicBrush/images/102.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 46, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the cat to a rat.\n2. Make the suitcase red."}], "image": ["MagicBrush/images/103.jpg", "MagicBrush/images/104.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 47, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the cat to a rat.\n2. Make the suitcase red."}], "image": ["MagicBrush/images/105.jpg", "MagicBrush/images/106.jpg", "MagicBrush/images/107.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 48, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "replace the truck with a bus"}], "image": ["MagicBrush/images/108.jpg", "MagicBrush/images/109.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 49, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "replace the american flag with a red flag"}], "image": ["MagicBrush/images/110.jpg", "MagicBrush/images/111.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 50, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the truck with a bus\n2. replace the american flag with a red flag"}], "image": ["MagicBrush/images/112.jpg", "MagicBrush/images/113.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 51, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the truck with a bus\n2. replace the american flag with a red flag"}], "image": ["MagicBrush/images/114.jpg", "MagicBrush/images/115.jpg", "MagicBrush/images/116.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 52, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let her bite the hot dog"}], "image": ["MagicBrush/images/117.jpg", "MagicBrush/images/118.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 53, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make her wear a baseball hat"}], "image": ["MagicBrush/images/119.jpg", "MagicBrush/images/120.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 54, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let her bite the hot dog\n2. make her wear a baseball hat"}], "image": ["MagicBrush/images/121.jpg", "MagicBrush/images/122.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 55, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let her bite the hot dog\n2. make her wear a baseball hat"}], "image": ["MagicBrush/images/123.jpg", "MagicBrush/images/124.jpg", "MagicBrush/images/125.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 56, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "have a bird stand on the zebras head."}], "image": ["MagicBrush/images/126.jpg", "MagicBrush/images/127.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 57, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "have the zebra bent over a hay pile"}], "image": ["MagicBrush/images/128.jpg", "MagicBrush/images/129.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 58, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "have a bucket hang off the fence"}], "image": ["MagicBrush/images/130.jpg", "MagicBrush/images/131.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 59, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. have a bird stand on the zebras head.\n2. have the zebra bent over a hay pile\n3. have a bucket hang off the fence"}], "image": ["MagicBrush/images/132.jpg", "MagicBrush/images/133.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 60, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. have a bird stand on the zebras head.\n2. have the zebra bent over a hay pile\n3. have a bucket hang off the fence"}], "image": ["MagicBrush/images/134.jpg", "MagicBrush/images/135.jpg", "MagicBrush/images/136.jpg", "MagicBrush/images/137.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 61, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the zebra sit."}], "image": ["MagicBrush/images/138.jpg", "MagicBrush/images/139.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 62, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the cat to a fox."}], "image": ["MagicBrush/images/140.jpg", "MagicBrush/images/141.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 63, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the cat lick its nose"}], "image": ["MagicBrush/images/142.jpg", "MagicBrush/images/143.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 64, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the mirror turn into a window."}], "image": ["MagicBrush/images/144.jpg", "MagicBrush/images/145.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 65, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the man not swing but hold the bat down towards the ground."}], "image": ["MagicBrush/images/146.jpg", "MagicBrush/images/147.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 66, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the man's pants all white."}], "image": ["MagicBrush/images/148.jpg", "MagicBrush/images/149.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 67, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the man not swing but hold the bat down towards the ground.\n2. Make the man's pants all white."}], "image": ["MagicBrush/images/150.jpg", "MagicBrush/images/151.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 68, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the man not swing but hold the bat down towards the ground.\n2. Make the man's pants all white."}], "image": ["MagicBrush/images/152.jpg", "MagicBrush/images/153.jpg", "MagicBrush/images/154.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 69, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make it a black sheep."}], "image": ["MagicBrush/images/155.jpg", "MagicBrush/images/156.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 70, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "A dog should be near the sheep."}], "image": ["MagicBrush/images/157.jpg", "MagicBrush/images/158.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 71, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make it a black sheep.\n2. A dog should be near the sheep."}], "image": ["MagicBrush/images/159.jpg", "MagicBrush/images/160.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 72, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make it a black sheep.\n2. A dog should be near the sheep."}], "image": ["MagicBrush/images/161.jpg", "MagicBrush/images/162.jpg", "MagicBrush/images/163.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 73, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if there is flowers on the vase on the toilet?"}], "image": ["MagicBrush/images/164.jpg", "MagicBrush/images/165.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 74, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the green ball to blue"}], "image": ["MagicBrush/images/166.jpg", "MagicBrush/images/167.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 75, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the dog leaf"}], "image": ["MagicBrush/images/168.jpg", "MagicBrush/images/169.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 76, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a dog bowl"}], "image": ["MagicBrush/images/170.jpg", "MagicBrush/images/171.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 77, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the green ball to blue\n2. make the dog leaf\n3. add a dog bowl"}], "image": ["MagicBrush/images/172.jpg", "MagicBrush/images/173.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 78, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the green ball to blue\n2. make the dog leaf\n3. add a dog bowl"}], "image": ["MagicBrush/images/174.jpg", "MagicBrush/images/175.jpg", "MagicBrush/images/176.jpg", "MagicBrush/images/177.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 79, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the player wear white socks."}], "image": ["MagicBrush/images/178.jpg", "MagicBrush/images/179.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 80, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "swap the motorcycle for a car"}], "image": ["MagicBrush/images/180.jpg", "MagicBrush/images/181.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 81, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the tent and add a bonfire"}], "image": ["MagicBrush/images/182.jpg", "MagicBrush/images/183.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 82, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the motorcycle for a car\n2. remove the tent and add a bonfire"}], "image": ["MagicBrush/images/184.jpg", "MagicBrush/images/185.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 83, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the motorcycle for a car\n2. remove the tent and add a bonfire"}], "image": ["MagicBrush/images/186.jpg", "MagicBrush/images/187.jpg", "MagicBrush/images/188.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 84, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cat be white"}], "image": ["MagicBrush/images/189.jpg", "MagicBrush/images/190.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 85, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the umbrella be striped"}], "image": ["MagicBrush/images/191.jpg", "MagicBrush/images/192.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 86, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the shoes"}], "image": ["MagicBrush/images/193.jpg", "MagicBrush/images/194.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 87, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat be white\n2. let the umbrella be striped\n3. remove the shoes"}], "image": ["MagicBrush/images/195.jpg", "MagicBrush/images/196.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 88, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat be white\n2. let the umbrella be striped\n3. remove the shoes"}], "image": ["MagicBrush/images/197.jpg", "MagicBrush/images/198.jpg", "MagicBrush/images/199.jpg", "MagicBrush/images/200.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 89, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the flag of the united states for that of england"}], "image": ["MagicBrush/images/201.jpg", "MagicBrush/images/202.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 90, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add mickey's face"}], "image": ["MagicBrush/images/203.jpg", "MagicBrush/images/204.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 91, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a rubber duck"}], "image": ["MagicBrush/images/205.jpg", "MagicBrush/images/206.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 92, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the flag of the united states for that of england\n2. add mickey's face\n3. add a rubber duck"}], "image": ["MagicBrush/images/207.jpg", "MagicBrush/images/208.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 93, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the flag of the united states for that of england\n2. add mickey's face\n3. add a rubber duck"}], "image": ["MagicBrush/images/209.jpg", "MagicBrush/images/210.jpg", "MagicBrush/images/211.jpg", "MagicBrush/images/212.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 94, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "take away the skateboarders arms."}], "image": ["MagicBrush/images/213.jpg", "MagicBrush/images/214.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 95, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the ramp cement"}], "image": ["MagicBrush/images/215.jpg", "MagicBrush/images/216.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 96, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a street"}], "image": ["MagicBrush/images/217.jpg", "MagicBrush/images/218.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 97, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. take away the skateboarders arms.\n2. make the ramp cement\n3. add a street"}], "image": ["MagicBrush/images/219.jpg", "MagicBrush/images/220.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 98, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. take away the skateboarders arms.\n2. make the ramp cement\n3. add a street"}], "image": ["MagicBrush/images/221.jpg", "MagicBrush/images/222.jpg", "MagicBrush/images/223.jpg", "MagicBrush/images/224.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 99, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn her hair white"}], "image": ["MagicBrush/images/225.jpg", "MagicBrush/images/226.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 100, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give her a skirt to wear"}], "image": ["MagicBrush/images/227.jpg", "MagicBrush/images/228.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 101, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn her hair white\n2. give her a skirt to wear"}], "image": ["MagicBrush/images/229.jpg", "MagicBrush/images/230.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 102, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn her hair white\n2. give her a skirt to wear"}], "image": ["MagicBrush/images/231.jpg", "MagicBrush/images/232.jpg", "MagicBrush/images/233.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 103, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the stop sign to a welcome sign"}], "image": ["MagicBrush/images/234.jpg", "MagicBrush/images/235.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 104, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let it be a bullet train"}], "image": ["MagicBrush/images/236.jpg", "MagicBrush/images/237.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 105, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a bird on the road"}], "image": ["MagicBrush/images/238.jpg", "MagicBrush/images/239.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 106, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the stop sign to a welcome sign\n2. let it be a bullet train\n3. add a bird on the road"}], "image": ["MagicBrush/images/240.jpg", "MagicBrush/images/241.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 107, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the stop sign to a welcome sign\n2. let it be a bullet train\n3. add a bird on the road"}], "image": ["MagicBrush/images/242.jpg", "MagicBrush/images/243.jpg", "MagicBrush/images/244.jpg", "MagicBrush/images/245.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 108, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the color of the soccer ball."}], "image": ["MagicBrush/images/246.jpg", "MagicBrush/images/247.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 109, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove middle fruit and put a cat in place"}], "image": ["MagicBrush/images/248.jpg", "MagicBrush/images/249.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 110, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a robot tiger next to the bear"}], "image": ["MagicBrush/images/250.jpg", "MagicBrush/images/251.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 111, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be a crystal ball on the table"}], "image": ["MagicBrush/images/252.jpg", "MagicBrush/images/253.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 112, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the ceiling have recessed lighting"}], "image": ["MagicBrush/images/254.jpg", "MagicBrush/images/255.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 113, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be a bedroom near the kitchen"}], "image": ["MagicBrush/images/256.jpg", "MagicBrush/images/257.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 114, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be a crystal ball on the table\n2. let the ceiling have recessed lighting\n3. let there be a bedroom near the kitchen"}], "image": ["MagicBrush/images/258.jpg", "MagicBrush/images/259.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 115, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be a crystal ball on the table\n2. let the ceiling have recessed lighting\n3. let there be a bedroom near the kitchen"}], "image": ["MagicBrush/images/260.jpg", "MagicBrush/images/261.jpg", "MagicBrush/images/262.jpg", "MagicBrush/images/263.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 116, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if he is holding a cup?"}], "image": ["MagicBrush/images/264.jpg", "MagicBrush/images/265.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 117, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a plane in the sky"}], "image": ["MagicBrush/images/266.jpg", "MagicBrush/images/267.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 118, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the clouds"}], "image": ["MagicBrush/images/268.jpg", "MagicBrush/images/269.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 119, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a plane in the sky\n2. remove the clouds"}], "image": ["MagicBrush/images/270.jpg", "MagicBrush/images/271.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 120, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a plane in the sky\n2. remove the clouds"}], "image": ["MagicBrush/images/272.jpg", "MagicBrush/images/273.jpg", "MagicBrush/images/274.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 121, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the boy to a girl."}], "image": ["MagicBrush/images/275.jpg", "MagicBrush/images/276.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 122, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put sunglasses on the girl."}], "image": ["MagicBrush/images/277.jpg", "MagicBrush/images/278.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 123, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the boy to a girl.\n2. Put sunglasses on the girl."}], "image": ["MagicBrush/images/279.jpg", "MagicBrush/images/280.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 124, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the boy to a girl.\n2. Put sunglasses on the girl."}], "image": ["MagicBrush/images/281.jpg", "MagicBrush/images/282.jpg", "MagicBrush/images/283.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 125, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "It should be a notebook on the table."}], "image": ["MagicBrush/images/284.jpg", "MagicBrush/images/285.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 126, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a doctor."}], "image": ["MagicBrush/images/286.jpg", "MagicBrush/images/287.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 127, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the tomatoes from one sandwich"}], "image": ["MagicBrush/images/288.jpg", "MagicBrush/images/289.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 128, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be a stuffed panda"}], "image": ["MagicBrush/images/290.jpg", "MagicBrush/images/291.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 129, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the wooden frame"}], "image": ["MagicBrush/images/292.jpg", "MagicBrush/images/293.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 130, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let white animals stick their tongues out"}], "image": ["MagicBrush/images/294.jpg", "MagicBrush/images/295.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 131, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be a stuffed panda\n2. remove the wooden frame\n3. let white animals stick their tongues out"}], "image": ["MagicBrush/images/296.jpg", "MagicBrush/images/297.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 132, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be a stuffed panda\n2. remove the wooden frame\n3. let white animals stick their tongues out"}], "image": ["MagicBrush/images/298.jpg", "MagicBrush/images/299.jpg", "MagicBrush/images/300.jpg", "MagicBrush/images/301.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 133, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove the surfboards."}], "image": ["MagicBrush/images/302.jpg", "MagicBrush/images/303.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 134, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the instructor's jacket say \"4\" on it"}], "image": ["MagicBrush/images/304.jpg", "MagicBrush/images/305.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 135, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the instructor be wearing pink pants"}], "image": ["MagicBrush/images/306.jpg", "MagicBrush/images/307.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 136, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Have the instructor's jacket say \"4\" on it\n2. Have the instructor be wearing pink pants"}], "image": ["MagicBrush/images/308.jpg", "MagicBrush/images/309.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 137, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Have the instructor's jacket say \"4\" on it\n2. Have the instructor be wearing pink pants"}], "image": ["MagicBrush/images/310.jpg", "MagicBrush/images/311.jpg", "MagicBrush/images/312.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 138, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the zebra put its face up close to a tree."}], "image": ["MagicBrush/images/313.jpg", "MagicBrush/images/314.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 139, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a couch next to the window."}], "image": ["MagicBrush/images/315.jpg", "MagicBrush/images/316.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 140, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the man be angry"}], "image": ["MagicBrush/images/317.jpg", "MagicBrush/images/318.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 141, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the man be dressed in a dinner jacket"}], "image": ["MagicBrush/images/319.jpg", "MagicBrush/images/320.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 142, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the tennis racket into a baseball bat"}], "image": ["MagicBrush/images/321.jpg", "MagicBrush/images/322.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 143, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the man be angry\n2. let the man be dressed in a dinner jacket\n3. change the tennis racket into a baseball bat"}], "image": ["MagicBrush/images/323.jpg", "MagicBrush/images/324.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 144, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the man be angry\n2. let the man be dressed in a dinner jacket\n3. change the tennis racket into a baseball bat"}], "image": ["MagicBrush/images/325.jpg", "MagicBrush/images/326.jpg", "MagicBrush/images/327.jpg", "MagicBrush/images/328.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 145, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the hat to a cowboy hat"}], "image": ["MagicBrush/images/329.jpg", "MagicBrush/images/330.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 146, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace all the food with a big pizza."}], "image": ["MagicBrush/images/331.jpg", "MagicBrush/images/332.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 147, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the frisbee into a ball"}], "image": ["MagicBrush/images/333.jpg", "MagicBrush/images/334.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 148, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a whale in the ocean"}], "image": ["MagicBrush/images/335.jpg", "MagicBrush/images/336.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 149, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a dog barking near shore"}], "image": ["MagicBrush/images/337.jpg", "MagicBrush/images/338.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 150, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the frisbee into a ball\n2. add a whale in the ocean\n3. add a dog barking near shore"}], "image": ["MagicBrush/images/339.jpg", "MagicBrush/images/340.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 151, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the frisbee into a ball\n2. add a whale in the ocean\n3. add a dog barking near shore"}], "image": ["MagicBrush/images/341.jpg", "MagicBrush/images/342.jpg", "MagicBrush/images/343.jpg", "MagicBrush/images/344.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 152, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the phone booth be red"}], "image": ["MagicBrush/images/345.jpg", "MagicBrush/images/346.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 153, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the red building be white"}], "image": ["MagicBrush/images/347.jpg", "MagicBrush/images/348.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 154, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a dog next to the car"}], "image": ["MagicBrush/images/349.jpg", "MagicBrush/images/350.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 155, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the phone booth be red\n2. let the red building be white\n3. add a dog next to the car"}], "image": ["MagicBrush/images/351.jpg", "MagicBrush/images/352.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 156, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the phone booth be red\n2. let the red building be white\n3. add a dog next to the car"}], "image": ["MagicBrush/images/353.jpg", "MagicBrush/images/354.jpg", "MagicBrush/images/355.jpg", "MagicBrush/images/356.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 157, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the man be wearing a kilt"}], "image": ["MagicBrush/images/357.jpg", "MagicBrush/images/358.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 158, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a giant wasp."}], "image": ["MagicBrush/images/359.jpg", "MagicBrush/images/360.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 159, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the person raise his arm."}], "image": ["MagicBrush/images/361.jpg", "MagicBrush/images/362.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 160, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "place a cat in the counter"}], "image": ["MagicBrush/images/363.jpg", "MagicBrush/images/364.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 161, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the cloth from the chairs"}], "image": ["MagicBrush/images/365.jpg", "MagicBrush/images/366.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 162, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. place a cat in the counter\n2. remove the cloth from the chairs"}], "image": ["MagicBrush/images/367.jpg", "MagicBrush/images/368.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 163, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. place a cat in the counter\n2. remove the cloth from the chairs"}], "image": ["MagicBrush/images/369.jpg", "MagicBrush/images/370.jpg", "MagicBrush/images/371.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 164, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add some orange juice inside the blender"}], "image": ["MagicBrush/images/372.jpg", "MagicBrush/images/373.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 165, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a vodka bottle"}], "image": ["MagicBrush/images/374.jpg", "MagicBrush/images/375.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 166, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change it for some shot glasses"}], "image": ["MagicBrush/images/376.jpg", "MagicBrush/images/377.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 167, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add some orange juice inside the blender\n2. Add a vodka bottle\n3. Change it for some shot glasses"}], "image": ["MagicBrush/images/378.jpg", "MagicBrush/images/379.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 168, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add some orange juice inside the blender\n2. Add a vodka bottle\n3. Change it for some shot glasses"}], "image": ["MagicBrush/images/380.jpg", "MagicBrush/images/381.jpg", "MagicBrush/images/382.jpg", "MagicBrush/images/383.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 169, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "He should be eating a watermelon"}], "image": ["MagicBrush/images/384.jpg", "MagicBrush/images/385.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 170, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Could he be in the forest?"}], "image": ["MagicBrush/images/386.jpg", "MagicBrush/images/387.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 171, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a pond next to the elephant"}], "image": ["MagicBrush/images/388.jpg", "MagicBrush/images/389.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 172, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. He should be eating a watermelon\n2. Could he be in the forest?\n3. Put a pond next to the elephant"}], "image": ["MagicBrush/images/390.jpg", "MagicBrush/images/391.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 173, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. He should be eating a watermelon\n2. Could he be in the forest?\n3. Put a pond next to the elephant"}], "image": ["MagicBrush/images/392.jpg", "MagicBrush/images/393.jpg", "MagicBrush/images/394.jpg", "MagicBrush/images/395.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 174, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Give the zebra a single front leg."}], "image": ["MagicBrush/images/396.jpg", "MagicBrush/images/397.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 175, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Open the zebra's mouth."}], "image": ["MagicBrush/images/398.jpg", "MagicBrush/images/399.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 176, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Give the zebra a single front leg.\n2. Open the zebra's mouth."}], "image": ["MagicBrush/images/400.jpg", "MagicBrush/images/401.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 177, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Give the zebra a single front leg.\n2. Open the zebra's mouth."}], "image": ["MagicBrush/images/402.jpg", "MagicBrush/images/403.jpg", "MagicBrush/images/404.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 178, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make her outfit black."}], "image": ["MagicBrush/images/405.jpg", "MagicBrush/images/406.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 179, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a party hat on the dog"}], "image": ["MagicBrush/images/407.jpg", "MagicBrush/images/408.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 180, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the frisbee blue"}], "image": ["MagicBrush/images/409.jpg", "MagicBrush/images/410.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 181, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a party hat on the dog\n2. make the frisbee blue"}], "image": ["MagicBrush/images/411.jpg", "MagicBrush/images/412.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 182, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a party hat on the dog\n2. make the frisbee blue"}], "image": ["MagicBrush/images/413.jpg", "MagicBrush/images/414.jpg", "MagicBrush/images/415.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 183, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Get rid of the spatula."}], "image": ["MagicBrush/images/416.jpg", "MagicBrush/images/417.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 184, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the cake vanilla."}], "image": ["MagicBrush/images/418.jpg", "MagicBrush/images/419.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 185, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the roses into tulips."}], "image": ["MagicBrush/images/420.jpg", "MagicBrush/images/421.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 186, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Get rid of the spatula.\n2. Make the cake vanilla.\n3. Make the roses into tulips."}], "image": ["MagicBrush/images/422.jpg", "MagicBrush/images/423.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 187, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Get rid of the spatula.\n2. Make the cake vanilla.\n3. Make the roses into tulips."}], "image": ["MagicBrush/images/424.jpg", "MagicBrush/images/425.jpg", "MagicBrush/images/426.jpg", "MagicBrush/images/427.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 188, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Place a red warning sign saying \"Stop\""}], "image": ["MagicBrush/images/428.jpg", "MagicBrush/images/429.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 189, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the woman be wearing a beret"}], "image": ["MagicBrush/images/430.jpg", "MagicBrush/images/431.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 190, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the motorcycles and cars pink"}], "image": ["MagicBrush/images/432.jpg", "MagicBrush/images/433.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 191, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a hot air balloon in the background"}], "image": ["MagicBrush/images/434.jpg", "MagicBrush/images/435.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 192, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a cotton candy machine"}], "image": ["MagicBrush/images/436.jpg", "MagicBrush/images/437.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 193, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the motorcycles and cars pink\n2. put a hot air balloon in the background\n3. add a cotton candy machine"}], "image": ["MagicBrush/images/438.jpg", "MagicBrush/images/439.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 194, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the motorcycles and cars pink\n2. put a hot air balloon in the background\n3. add a cotton candy machine"}], "image": ["MagicBrush/images/440.jpg", "MagicBrush/images/441.jpg", "MagicBrush/images/442.jpg", "MagicBrush/images/443.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 195, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "close her eyes"}], "image": ["MagicBrush/images/444.jpg", "MagicBrush/images/445.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 196, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the person behind the woman"}], "image": ["MagicBrush/images/446.jpg", "MagicBrush/images/447.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 197, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the woman clap"}], "image": ["MagicBrush/images/448.jpg", "MagicBrush/images/449.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 198, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. close her eyes\n2. remove the person behind the woman\n3. make the woman clap"}], "image": ["MagicBrush/images/450.jpg", "MagicBrush/images/451.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 199, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. close her eyes\n2. remove the person behind the woman\n3. make the woman clap"}], "image": ["MagicBrush/images/452.jpg", "MagicBrush/images/453.jpg", "MagicBrush/images/454.jpg", "MagicBrush/images/455.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 200, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put another freezer on the truck"}], "image": ["MagicBrush/images/456.jpg", "MagicBrush/images/457.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 201, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "open the door of the truck"}], "image": ["MagicBrush/images/458.jpg", "MagicBrush/images/459.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 202, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put another freezer on the truck\n2. open the door of the truck"}], "image": ["MagicBrush/images/460.jpg", "MagicBrush/images/461.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 203, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put another freezer on the truck\n2. open the door of the truck"}], "image": ["MagicBrush/images/462.jpg", "MagicBrush/images/463.jpg", "MagicBrush/images/464.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 204, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the people angry"}], "image": ["MagicBrush/images/465.jpg", "MagicBrush/images/466.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 205, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can we have a blue airplane?"}], "image": ["MagicBrush/images/467.jpg", "MagicBrush/images/468.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 206, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Could it be a pond next to the airfield?"}], "image": ["MagicBrush/images/469.jpg", "MagicBrush/images/470.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 207, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can we have a blue airplane?\n2. Could it be a pond next to the airfield?"}], "image": ["MagicBrush/images/471.jpg", "MagicBrush/images/472.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 208, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can we have a blue airplane?\n2. Could it be a pond next to the airfield?"}], "image": ["MagicBrush/images/473.jpg", "MagicBrush/images/474.jpg", "MagicBrush/images/475.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 209, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the donut a cupcake."}], "image": ["MagicBrush/images/476.jpg", "MagicBrush/images/477.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 210, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the coffee with beer."}], "image": ["MagicBrush/images/478.jpg", "MagicBrush/images/479.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 211, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the donut a cupcake.\n2. Replace the coffee with beer."}], "image": ["MagicBrush/images/480.jpg", "MagicBrush/images/481.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 212, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the donut a cupcake.\n2. Replace the coffee with beer."}], "image": ["MagicBrush/images/482.jpg", "MagicBrush/images/483.jpg", "MagicBrush/images/484.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 213, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a rat on the counter."}], "image": ["MagicBrush/images/485.jpg", "MagicBrush/images/486.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 214, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the dog howl"}], "image": ["MagicBrush/images/487.jpg", "MagicBrush/images/488.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 215, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the planters be made of cast stone"}], "image": ["MagicBrush/images/489.jpg", "MagicBrush/images/490.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 216, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the planters have fruit trees in it"}], "image": ["MagicBrush/images/491.jpg", "MagicBrush/images/492.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 217, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the dog howl\n2. let the planters be made of cast stone\n3. let the planters have fruit trees in it"}], "image": ["MagicBrush/images/493.jpg", "MagicBrush/images/494.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 218, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the dog howl\n2. let the planters be made of cast stone\n3. let the planters have fruit trees in it"}], "image": ["MagicBrush/images/495.jpg", "MagicBrush/images/496.jpg", "MagicBrush/images/497.jpg", "MagicBrush/images/498.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 219, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a monitor on the wall."}], "image": ["MagicBrush/images/499.jpg", "MagicBrush/images/500.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 220, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the carrot into broccoli"}], "image": ["MagicBrush/images/501.jpg", "MagicBrush/images/502.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 221, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the clock to a TV."}], "image": ["MagicBrush/images/503.jpg", "MagicBrush/images/504.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 222, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a show about cats on the TV."}], "image": ["MagicBrush/images/505.jpg", "MagicBrush/images/506.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 223, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the clock to a TV.\n2. Put a show about cats on the TV."}], "image": ["MagicBrush/images/507.jpg", "MagicBrush/images/508.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 224, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the clock to a TV.\n2. Put a show about cats on the TV."}], "image": ["MagicBrush/images/509.jpg", "MagicBrush/images/510.jpg", "MagicBrush/images/511.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 225, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the two men fall"}], "image": ["MagicBrush/images/512.jpg", "MagicBrush/images/513.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 226, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add white flowers on the lawn"}], "image": ["MagicBrush/images/514.jpg", "MagicBrush/images/515.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 227, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the sky rain"}], "image": ["MagicBrush/images/516.jpg", "MagicBrush/images/517.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 228, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the two men fall\n2. add white flowers on the lawn\n3. make the sky rain"}], "image": ["MagicBrush/images/518.jpg", "MagicBrush/images/519.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 229, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the two men fall\n2. add white flowers on the lawn\n3. make the sky rain"}], "image": ["MagicBrush/images/520.jpg", "MagicBrush/images/521.jpg", "MagicBrush/images/522.jpg", "MagicBrush/images/523.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 230, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the double deck bus to a truck"}], "image": ["MagicBrush/images/524.jpg", "MagicBrush/images/525.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 231, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a pedestrian"}], "image": ["MagicBrush/images/526.jpg", "MagicBrush/images/527.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 232, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the double deck bus to a truck\n2. add a pedestrian"}], "image": ["MagicBrush/images/528.jpg", "MagicBrush/images/529.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 233, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the double deck bus to a truck\n2. add a pedestrian"}], "image": ["MagicBrush/images/530.jpg", "MagicBrush/images/531.jpg", "MagicBrush/images/532.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 234, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "the boy should be holding a banana."}], "image": ["MagicBrush/images/533.jpg", "MagicBrush/images/534.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 235, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can the man hold bananas too?"}], "image": ["MagicBrush/images/535.jpg", "MagicBrush/images/536.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 236, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. the boy should be holding a banana.\n2. Can the man hold bananas too?"}], "image": ["MagicBrush/images/537.jpg", "MagicBrush/images/538.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 237, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. the boy should be holding a banana.\n2. Can the man hold bananas too?"}], "image": ["MagicBrush/images/539.jpg", "MagicBrush/images/540.jpg", "MagicBrush/images/541.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 238, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "There should be no dolls in the room."}], "image": ["MagicBrush/images/542.jpg", "MagicBrush/images/543.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 239, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a few balloons to the room."}], "image": ["MagicBrush/images/544.jpg", "MagicBrush/images/545.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 240, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. There should be no dolls in the room.\n2. Add a few balloons to the room."}], "image": ["MagicBrush/images/546.jpg", "MagicBrush/images/547.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 241, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. There should be no dolls in the room.\n2. Add a few balloons to the room."}], "image": ["MagicBrush/images/548.jpg", "MagicBrush/images/549.jpg", "MagicBrush/images/550.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 242, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be trees in the background"}], "image": ["MagicBrush/images/551.jpg", "MagicBrush/images/552.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 243, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a dog stand on its hind legs"}], "image": ["MagicBrush/images/553.jpg", "MagicBrush/images/554.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 244, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the table have no items on it"}], "image": ["MagicBrush/images/555.jpg", "MagicBrush/images/556.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 245, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the shelf be empty"}], "image": ["MagicBrush/images/557.jpg", "MagicBrush/images/558.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 246, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let a dog stand on its hind legs\n2. let the table have no items on it\n3. let the shelf be empty"}], "image": ["MagicBrush/images/559.jpg", "MagicBrush/images/560.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 247, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let a dog stand on its hind legs\n2. let the table have no items on it\n3. let the shelf be empty"}], "image": ["MagicBrush/images/561.jpg", "MagicBrush/images/562.jpg", "MagicBrush/images/563.jpg", "MagicBrush/images/564.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 248, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the frisbee with a ball."}], "image": ["MagicBrush/images/565.jpg", "MagicBrush/images/566.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 249, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let it be a single sink"}], "image": ["MagicBrush/images/567.jpg", "MagicBrush/images/568.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 250, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the window show a view of skyscrapers"}], "image": ["MagicBrush/images/569.jpg", "MagicBrush/images/570.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 251, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a cat lick its paw"}], "image": ["MagicBrush/images/571.jpg", "MagicBrush/images/572.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 252, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let it be a single sink\n2. let the window show a view of skyscrapers\n3. let a cat lick its paw"}], "image": ["MagicBrush/images/573.jpg", "MagicBrush/images/574.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 253, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let it be a single sink\n2. let the window show a view of skyscrapers\n3. let a cat lick its paw"}], "image": ["MagicBrush/images/575.jpg", "MagicBrush/images/576.jpg", "MagicBrush/images/577.jpg", "MagicBrush/images/578.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 254, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the teddy bear into a ship"}], "image": ["MagicBrush/images/579.jpg", "MagicBrush/images/580.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 255, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "get rid of the framed pictures"}], "image": ["MagicBrush/images/581.jpg", "MagicBrush/images/582.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 256, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "and rum bottles"}], "image": ["MagicBrush/images/583.jpg", "MagicBrush/images/584.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 257, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the teddy bear into a ship\n2. get rid of the framed pictures\n3. and rum bottles"}], "image": ["MagicBrush/images/585.jpg", "MagicBrush/images/586.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 258, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the teddy bear into a ship\n2. get rid of the framed pictures\n3. and rum bottles"}], "image": ["MagicBrush/images/587.jpg", "MagicBrush/images/588.jpg", "MagicBrush/images/589.jpg", "MagicBrush/images/590.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 259, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the plate blue."}], "image": ["MagicBrush/images/591.jpg", "MagicBrush/images/592.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 260, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the fruits whole."}], "image": ["MagicBrush/images/593.jpg", "MagicBrush/images/594.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 261, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the plate blue.\n2. Make the fruits whole."}], "image": ["MagicBrush/images/595.jpg", "MagicBrush/images/596.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 262, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the plate blue.\n2. Make the fruits whole."}], "image": ["MagicBrush/images/597.jpg", "MagicBrush/images/598.jpg", "MagicBrush/images/599.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 263, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the cow wear a hat."}], "image": ["MagicBrush/images/600.jpg", "MagicBrush/images/601.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 264, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a computer on the kitchen"}], "image": ["MagicBrush/images/602.jpg", "MagicBrush/images/603.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 265, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "leave only one stool"}], "image": ["MagicBrush/images/604.jpg", "MagicBrush/images/605.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 266, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a computer on the kitchen\n2. leave only one stool"}], "image": ["MagicBrush/images/606.jpg", "MagicBrush/images/607.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 267, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a computer on the kitchen\n2. leave only one stool"}], "image": ["MagicBrush/images/608.jpg", "MagicBrush/images/609.jpg", "MagicBrush/images/610.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 268, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the sign show an address"}], "image": ["MagicBrush/images/611.jpg", "MagicBrush/images/612.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 269, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a red flag above the sign"}], "image": ["MagicBrush/images/613.jpg", "MagicBrush/images/614.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 270, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add white geese in the sky"}], "image": ["MagicBrush/images/615.jpg", "MagicBrush/images/616.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 271, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the sign show an address\n2. add a red flag above the sign\n3. add white geese in the sky"}], "image": ["MagicBrush/images/617.jpg", "MagicBrush/images/618.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 272, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the sign show an address\n2. add a red flag above the sign\n3. add white geese in the sky"}], "image": ["MagicBrush/images/619.jpg", "MagicBrush/images/620.jpg", "MagicBrush/images/621.jpg", "MagicBrush/images/622.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 273, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "have the sun rise instead of set."}], "image": ["MagicBrush/images/623.jpg", "MagicBrush/images/624.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 274, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make two parasailers instead of one."}], "image": ["MagicBrush/images/625.jpg", "MagicBrush/images/626.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 275, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the ground a forest instead of a slope."}], "image": ["MagicBrush/images/627.jpg", "MagicBrush/images/628.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 276, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. have the sun rise instead of set.\n2. make two parasailers instead of one.\n3. make the ground a forest instead of a slope."}], "image": ["MagicBrush/images/629.jpg", "MagicBrush/images/630.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 277, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. have the sun rise instead of set.\n2. make two parasailers instead of one.\n3. make the ground a forest instead of a slope."}], "image": ["MagicBrush/images/631.jpg", "MagicBrush/images/632.jpg", "MagicBrush/images/633.jpg", "MagicBrush/images/634.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 278, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "swap the cats for a fox"}], "image": ["MagicBrush/images/635.jpg", "MagicBrush/images/636.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 279, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "fill the table with cakes"}], "image": ["MagicBrush/images/637.jpg", "MagicBrush/images/638.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 280, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the cats for a fox\n2. fill the table with cakes"}], "image": ["MagicBrush/images/639.jpg", "MagicBrush/images/640.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 281, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the cats for a fox\n2. fill the table with cakes"}], "image": ["MagicBrush/images/641.jpg", "MagicBrush/images/642.jpg", "MagicBrush/images/643.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 282, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add water and flowers in the tub"}], "image": ["MagicBrush/images/644.jpg", "MagicBrush/images/645.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 283, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the floor be made of hardwood"}], "image": ["MagicBrush/images/646.jpg", "MagicBrush/images/647.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 284, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add water and flowers in the tub\n2. let the floor be made of hardwood"}], "image": ["MagicBrush/images/648.jpg", "MagicBrush/images/649.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 285, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add water and flowers in the tub\n2. let the floor be made of hardwood"}], "image": ["MagicBrush/images/650.jpg", "MagicBrush/images/651.jpg", "MagicBrush/images/652.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 286, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the man have legs up in air"}], "image": ["MagicBrush/images/653.jpg", "MagicBrush/images/654.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 287, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the man wear a helmet"}], "image": ["MagicBrush/images/655.jpg", "MagicBrush/images/656.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 288, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be a giant snowball next to the man"}], "image": ["MagicBrush/images/657.jpg", "MagicBrush/images/658.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 289, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the man have legs up in air\n2. let the man wear a helmet\n3. let there be a giant snowball next to the man"}], "image": ["MagicBrush/images/659.jpg", "MagicBrush/images/660.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 290, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the man have legs up in air\n2. let the man wear a helmet\n3. let there be a giant snowball next to the man"}], "image": ["MagicBrush/images/661.jpg", "MagicBrush/images/662.jpg", "MagicBrush/images/663.jpg", "MagicBrush/images/664.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 291, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the male player for a female"}], "image": ["MagicBrush/images/665.jpg", "MagicBrush/images/666.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 292, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove bananas and add grapes"}], "image": ["MagicBrush/images/667.jpg", "MagicBrush/images/668.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 293, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Turn the all street lights green."}], "image": ["MagicBrush/images/669.jpg", "MagicBrush/images/670.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 294, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put traffic cones in the street."}], "image": ["MagicBrush/images/671.jpg", "MagicBrush/images/672.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 295, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Turn the all street lights green.\n2. Put traffic cones in the street."}], "image": ["MagicBrush/images/673.jpg", "MagicBrush/images/674.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 296, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Turn the all street lights green.\n2. Put traffic cones in the street."}], "image": ["MagicBrush/images/675.jpg", "MagicBrush/images/676.jpg", "MagicBrush/images/677.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 297, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the fan into a chandelier"}], "image": ["MagicBrush/images/678.jpg", "MagicBrush/images/679.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 298, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let curtains be closed on the window"}], "image": ["MagicBrush/images/680.jpg", "MagicBrush/images/681.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 299, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the table have sofas near the dining table"}], "image": ["MagicBrush/images/682.jpg", "MagicBrush/images/683.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 300, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the fan into a chandelier\n2. let curtains be closed on the window\n3. let the table have sofas near the dining table"}], "image": ["MagicBrush/images/684.jpg", "MagicBrush/images/685.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 301, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the fan into a chandelier\n2. let curtains be closed on the window\n3. let the table have sofas near the dining table"}], "image": ["MagicBrush/images/686.jpg", "MagicBrush/images/687.jpg", "MagicBrush/images/688.jpg", "MagicBrush/images/689.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 302, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have one of the children be blowing bubbles"}], "image": ["MagicBrush/images/690.jpg", "MagicBrush/images/691.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 303, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the table and add an aquarium"}], "image": ["MagicBrush/images/692.jpg", "MagicBrush/images/693.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 304, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "place a penguin in the picture"}], "image": ["MagicBrush/images/694.jpg", "MagicBrush/images/695.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 305, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the table and add an aquarium\n2. place a penguin in the picture"}], "image": ["MagicBrush/images/696.jpg", "MagicBrush/images/697.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 306, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the table and add an aquarium\n2. place a penguin in the picture"}], "image": ["MagicBrush/images/698.jpg", "MagicBrush/images/699.jpg", "MagicBrush/images/700.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 307, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the bicycle to a blue bicycle"}], "image": ["MagicBrush/images/701.jpg", "MagicBrush/images/702.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 308, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "The lid of a toilet should be open."}], "image": ["MagicBrush/images/703.jpg", "MagicBrush/images/704.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 309, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a vase of flowers in the sink."}], "image": ["MagicBrush/images/705.jpg", "MagicBrush/images/706.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 310, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "It should it be a window on the bathroom."}], "image": ["MagicBrush/images/707.jpg", "MagicBrush/images/708.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 311, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. The lid of a toilet should be open.\n2. Put a vase of flowers in the sink.\n3. It should it be a window on the bathroom."}], "image": ["MagicBrush/images/709.jpg", "MagicBrush/images/710.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 312, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. The lid of a toilet should be open.\n2. Put a vase of flowers in the sink.\n3. It should it be a window on the bathroom."}], "image": ["MagicBrush/images/711.jpg", "MagicBrush/images/712.jpg", "MagicBrush/images/713.jpg", "MagicBrush/images/714.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 313, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make it a slice of pizza instead of the sandwich."}], "image": ["MagicBrush/images/715.jpg", "MagicBrush/images/716.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 314, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "There should be some cutlery on the table."}], "image": ["MagicBrush/images/717.jpg", "MagicBrush/images/718.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 315, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make it a slice of pizza instead of the sandwich.\n2. There should be some cutlery on the table."}], "image": ["MagicBrush/images/719.jpg", "MagicBrush/images/720.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 316, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make it a slice of pizza instead of the sandwich.\n2. There should be some cutlery on the table."}], "image": ["MagicBrush/images/721.jpg", "MagicBrush/images/722.jpg", "MagicBrush/images/723.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 317, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the guts in the tennis racket"}], "image": ["MagicBrush/images/724.jpg", "MagicBrush/images/725.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 318, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make a hand hold the racket"}], "image": ["MagicBrush/images/726.jpg", "MagicBrush/images/727.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 319, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a flock of birds fly in the sky"}], "image": ["MagicBrush/images/728.jpg", "MagicBrush/images/729.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 320, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the guts in the tennis racket\n2. make a hand hold the racket\n3. let a flock of birds fly in the sky"}], "image": ["MagicBrush/images/730.jpg", "MagicBrush/images/731.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 321, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the guts in the tennis racket\n2. make a hand hold the racket\n3. let a flock of birds fly in the sky"}], "image": ["MagicBrush/images/732.jpg", "MagicBrush/images/733.jpg", "MagicBrush/images/734.jpg", "MagicBrush/images/735.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 322, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the stop light be a spear"}], "image": ["MagicBrush/images/736.jpg", "MagicBrush/images/737.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 323, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the street be flooded"}], "image": ["MagicBrush/images/738.jpg", "MagicBrush/images/739.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 324, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a cop stand on the side"}], "image": ["MagicBrush/images/740.jpg", "MagicBrush/images/741.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 325, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the stop light be a spear\n2. let the street be flooded\n3. let a cop stand on the side"}], "image": ["MagicBrush/images/742.jpg", "MagicBrush/images/743.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 326, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the stop light be a spear\n2. let the street be flooded\n3. let a cop stand on the side"}], "image": ["MagicBrush/images/744.jpg", "MagicBrush/images/745.jpg", "MagicBrush/images/746.jpg", "MagicBrush/images/747.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 327, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be a bottle behind the vegetables"}], "image": ["MagicBrush/images/748.jpg", "MagicBrush/images/749.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 328, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make one fruit have a face"}], "image": ["MagicBrush/images/750.jpg", "MagicBrush/images/751.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 329, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a party hat on the fruit with a face"}], "image": ["MagicBrush/images/752.jpg", "MagicBrush/images/753.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 330, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make one fruit have a face\n2. put a party hat on the fruit with a face"}], "image": ["MagicBrush/images/754.jpg", "MagicBrush/images/755.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 331, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make one fruit have a face\n2. put a party hat on the fruit with a face"}], "image": ["MagicBrush/images/756.jpg", "MagicBrush/images/757.jpg", "MagicBrush/images/758.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 332, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can it be just a woman and a child?"}], "image": ["MagicBrush/images/759.jpg", "MagicBrush/images/760.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 333, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can you put a city on the background?"}], "image": ["MagicBrush/images/761.jpg", "MagicBrush/images/762.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 334, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a car next to the child."}], "image": ["MagicBrush/images/763.jpg", "MagicBrush/images/764.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 335, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can it be just a woman and a child?\n2. Can you put a city on the background?\n3. Put a car next to the child."}], "image": ["MagicBrush/images/765.jpg", "MagicBrush/images/766.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 336, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can it be just a woman and a child?\n2. Can you put a city on the background?\n3. Put a car next to the child."}], "image": ["MagicBrush/images/767.jpg", "MagicBrush/images/768.jpg", "MagicBrush/images/769.jpg", "MagicBrush/images/770.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 337, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be a unicorn on the back of the woman's shirt"}], "image": ["MagicBrush/images/771.jpg", "MagicBrush/images/772.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 338, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if the man was standing?"}], "image": ["MagicBrush/images/773.jpg", "MagicBrush/images/774.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 339, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the computer and add a coffee machine"}], "image": ["MagicBrush/images/775.jpg", "MagicBrush/images/776.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 340, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove one chair"}], "image": ["MagicBrush/images/777.jpg", "MagicBrush/images/778.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 341, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the computer and add a coffee machine\n2. remove one chair"}], "image": ["MagicBrush/images/779.jpg", "MagicBrush/images/780.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 342, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the computer and add a coffee machine\n2. remove one chair"}], "image": ["MagicBrush/images/781.jpg", "MagicBrush/images/782.jpg", "MagicBrush/images/783.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 343, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the baseball glove be red"}], "image": ["MagicBrush/images/784.jpg", "MagicBrush/images/785.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 344, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a sparrow"}], "image": ["MagicBrush/images/786.jpg", "MagicBrush/images/787.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 345, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the yellow hat be red"}], "image": ["MagicBrush/images/788.jpg", "MagicBrush/images/789.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 346, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the baseball glove be red\n2. add a sparrow\n3. let the yellow hat be red"}], "image": ["MagicBrush/images/790.jpg", "MagicBrush/images/791.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 347, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the baseball glove be red\n2. add a sparrow\n3. let the yellow hat be red"}], "image": ["MagicBrush/images/792.jpg", "MagicBrush/images/793.jpg", "MagicBrush/images/794.jpg", "MagicBrush/images/795.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 348, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "It should be a chocolate cake on the plate."}], "image": ["MagicBrush/images/796.jpg", "MagicBrush/images/797.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 349, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a glass of juice next to the plate."}], "image": ["MagicBrush/images/798.jpg", "MagicBrush/images/799.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 350, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can it be a red plate?"}], "image": ["MagicBrush/images/800.jpg", "MagicBrush/images/801.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 351, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. It should be a chocolate cake on the plate.\n2. Put a glass of juice next to the plate.\n3. Can it be a red plate?"}], "image": ["MagicBrush/images/802.jpg", "MagicBrush/images/803.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 352, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. It should be a chocolate cake on the plate.\n2. Put a glass of juice next to the plate.\n3. Can it be a red plate?"}], "image": ["MagicBrush/images/804.jpg", "MagicBrush/images/805.jpg", "MagicBrush/images/806.jpg", "MagicBrush/images/807.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 353, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have a gorilla sit at the dinner table."}], "image": ["MagicBrush/images/808.jpg", "MagicBrush/images/809.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 354, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the greens with onions."}], "image": ["MagicBrush/images/810.jpg", "MagicBrush/images/811.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 355, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the fruit with more veggies."}], "image": ["MagicBrush/images/812.jpg", "MagicBrush/images/813.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 356, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put in different kinds of cheese instead of crackers."}], "image": ["MagicBrush/images/814.jpg", "MagicBrush/images/815.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 357, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Replace the fruit with more veggies.\n2. Put in different kinds of cheese instead of crackers."}], "image": ["MagicBrush/images/816.jpg", "MagicBrush/images/817.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 358, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Replace the fruit with more veggies.\n2. Put in different kinds of cheese instead of crackers."}], "image": ["MagicBrush/images/818.jpg", "MagicBrush/images/819.jpg", "MagicBrush/images/820.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 359, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the bed be wooden"}], "image": ["MagicBrush/images/821.jpg", "MagicBrush/images/822.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 360, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the tiled floor be made of plain granite"}], "image": ["MagicBrush/images/823.jpg", "MagicBrush/images/824.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 361, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make a cat jump onto a bed"}], "image": ["MagicBrush/images/825.jpg", "MagicBrush/images/826.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 362, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the bed be wooden\n2. let the tiled floor be made of plain granite\n3. make a cat jump onto a bed"}], "image": ["MagicBrush/images/827.jpg", "MagicBrush/images/828.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 363, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the bed be wooden\n2. let the tiled floor be made of plain granite\n3. make a cat jump onto a bed"}], "image": ["MagicBrush/images/829.jpg", "MagicBrush/images/830.jpg", "MagicBrush/images/831.jpg", "MagicBrush/images/832.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 364, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a cat falling out of a tree."}], "image": ["MagicBrush/images/833.jpg", "MagicBrush/images/834.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 365, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a wedding cake on one of the tables."}], "image": ["MagicBrush/images/835.jpg", "MagicBrush/images/836.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 366, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a fountain in front of the restaurant."}], "image": ["MagicBrush/images/837.jpg", "MagicBrush/images/838.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 367, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add a cat falling out of a tree.\n2. Put a wedding cake on one of the tables.\n3. Put a fountain in front of the restaurant."}], "image": ["MagicBrush/images/839.jpg", "MagicBrush/images/840.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 368, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add a cat falling out of a tree.\n2. Put a wedding cake on one of the tables.\n3. Put a fountain in front of the restaurant."}], "image": ["MagicBrush/images/841.jpg", "MagicBrush/images/842.jpg", "MagicBrush/images/843.jpg", "MagicBrush/images/844.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 369, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the stop sing an animal sign"}], "image": ["MagicBrush/images/845.jpg", "MagicBrush/images/846.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 370, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add another cup of water"}], "image": ["MagicBrush/images/847.jpg", "MagicBrush/images/848.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 371, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the plate empty"}], "image": ["MagicBrush/images/849.jpg", "MagicBrush/images/850.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 372, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add another cup of water\n2. make the plate empty"}], "image": ["MagicBrush/images/851.jpg", "MagicBrush/images/852.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 373, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add another cup of water\n2. make the plate empty"}], "image": ["MagicBrush/images/853.jpg", "MagicBrush/images/854.jpg", "MagicBrush/images/855.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 374, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "get rid of the bananas"}], "image": ["MagicBrush/images/856.jpg", "MagicBrush/images/857.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 375, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "leave only oranges"}], "image": ["MagicBrush/images/858.jpg", "MagicBrush/images/859.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 376, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. get rid of the bananas\n2. leave only oranges"}], "image": ["MagicBrush/images/860.jpg", "MagicBrush/images/861.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 377, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. get rid of the bananas\n2. leave only oranges"}], "image": ["MagicBrush/images/862.jpg", "MagicBrush/images/863.jpg", "MagicBrush/images/864.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 378, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the doll wear a hat."}], "image": ["MagicBrush/images/865.jpg", "MagicBrush/images/866.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 379, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the knife to a chainsaw"}], "image": ["MagicBrush/images/867.jpg", "MagicBrush/images/868.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 380, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the pizza into a cake"}], "image": ["MagicBrush/images/869.jpg", "MagicBrush/images/870.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 381, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the bowl of fruit into a plate of salad"}], "image": ["MagicBrush/images/871.jpg", "MagicBrush/images/872.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 382, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the knife to a chainsaw\n2. change the pizza into a cake\n3. change the bowl of fruit into a plate of salad"}], "image": ["MagicBrush/images/873.jpg", "MagicBrush/images/874.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 383, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the knife to a chainsaw\n2. change the pizza into a cake\n3. change the bowl of fruit into a plate of salad"}], "image": ["MagicBrush/images/875.jpg", "MagicBrush/images/876.jpg", "MagicBrush/images/877.jpg", "MagicBrush/images/878.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 384, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cat have its eyes closed"}], "image": ["MagicBrush/images/879.jpg", "MagicBrush/images/880.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 385, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a woman inside the car"}], "image": ["MagicBrush/images/881.jpg", "MagicBrush/images/882.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 386, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add luggage on top of the car"}], "image": ["MagicBrush/images/883.jpg", "MagicBrush/images/884.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 387, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat have its eyes closed\n2. add a woman inside the car\n3. add luggage on top of the car"}], "image": ["MagicBrush/images/885.jpg", "MagicBrush/images/886.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 388, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat have its eyes closed\n2. add a woman inside the car\n3. add luggage on top of the car"}], "image": ["MagicBrush/images/887.jpg", "MagicBrush/images/888.jpg", "MagicBrush/images/889.jpg", "MagicBrush/images/890.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 389, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove a giraffe"}], "image": ["MagicBrush/images/891.jpg", "MagicBrush/images/892.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 390, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the umbrella have stripes"}], "image": ["MagicBrush/images/893.jpg", "MagicBrush/images/894.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 391, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a bird fly nearby"}], "image": ["MagicBrush/images/895.jpg", "MagicBrush/images/896.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 392, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove a giraffe\n2. let the umbrella have stripes\n3. let a bird fly nearby"}], "image": ["MagicBrush/images/897.jpg", "MagicBrush/images/898.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 393, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove a giraffe\n2. let the umbrella have stripes\n3. let a bird fly nearby"}], "image": ["MagicBrush/images/899.jpg", "MagicBrush/images/900.jpg", "MagicBrush/images/901.jpg", "MagicBrush/images/902.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 394, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cat sleep"}], "image": ["MagicBrush/images/903.jpg", "MagicBrush/images/904.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 395, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "awake the dog and give it brown eyes"}], "image": ["MagicBrush/images/905.jpg", "MagicBrush/images/906.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 396, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the dog yawn"}], "image": ["MagicBrush/images/907.jpg", "MagicBrush/images/908.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 397, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat sleep\n2. awake the dog and give it brown eyes\n3. let the dog yawn"}], "image": ["MagicBrush/images/909.jpg", "MagicBrush/images/910.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 398, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat sleep\n2. awake the dog and give it brown eyes\n3. let the dog yawn"}], "image": ["MagicBrush/images/911.jpg", "MagicBrush/images/912.jpg", "MagicBrush/images/913.jpg", "MagicBrush/images/914.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 399, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a face mask on one of the players"}], "image": ["MagicBrush/images/915.jpg", "MagicBrush/images/916.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 400, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a helmet on the player's haed"}], "image": ["MagicBrush/images/917.jpg", "MagicBrush/images/918.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 401, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a face mask on one of the players\n2. put a helmet on the player's haed"}], "image": ["MagicBrush/images/919.jpg", "MagicBrush/images/920.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 402, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a face mask on one of the players\n2. put a helmet on the player's haed"}], "image": ["MagicBrush/images/921.jpg", "MagicBrush/images/922.jpg", "MagicBrush/images/923.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 403, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a dear on the grass."}], "image": ["MagicBrush/images/924.jpg", "MagicBrush/images/925.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 404, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a big rock next to the cow."}], "image": ["MagicBrush/images/926.jpg", "MagicBrush/images/927.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 405, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can we have a pond next to the cart?"}], "image": ["MagicBrush/images/928.jpg", "MagicBrush/images/929.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 406, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put a big rock next to the cow.\n2. Can we have a pond next to the cart?"}], "image": ["MagicBrush/images/930.jpg", "MagicBrush/images/931.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 407, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put a big rock next to the cow.\n2. Can we have a pond next to the cart?"}], "image": ["MagicBrush/images/932.jpg", "MagicBrush/images/933.jpg", "MagicBrush/images/934.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 408, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can we put a hand on the door?"}], "image": ["MagicBrush/images/935.jpg", "MagicBrush/images/936.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 409, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a drawing of a rat next to the hand."}], "image": ["MagicBrush/images/937.jpg", "MagicBrush/images/938.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 410, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "The door could be cracked."}], "image": ["MagicBrush/images/939.jpg", "MagicBrush/images/940.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 411, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can we put a hand on the door?\n2. Put a drawing of a rat next to the hand.\n3. The door could be cracked."}], "image": ["MagicBrush/images/941.jpg", "MagicBrush/images/942.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 412, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can we put a hand on the door?\n2. Put a drawing of a rat next to the hand.\n3. The door could be cracked."}], "image": ["MagicBrush/images/943.jpg", "MagicBrush/images/944.jpg", "MagicBrush/images/945.jpg", "MagicBrush/images/946.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 413, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have the woman be wearing a blue tank top"}], "image": ["MagicBrush/images/947.jpg", "MagicBrush/images/948.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 414, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the woman's hair to blonde"}], "image": ["MagicBrush/images/949.jpg", "MagicBrush/images/950.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 415, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Have the woman be wearing a blue tank top\n2. Change the woman's hair to blonde"}], "image": ["MagicBrush/images/951.jpg", "MagicBrush/images/952.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 416, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Have the woman be wearing a blue tank top\n2. Change the woman's hair to blonde"}], "image": ["MagicBrush/images/953.jpg", "MagicBrush/images/954.jpg", "MagicBrush/images/955.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 417, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add flowers in the girl's hair"}], "image": ["MagicBrush/images/956.jpg", "MagicBrush/images/957.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 418, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add binary code on the computer"}], "image": ["MagicBrush/images/958.jpg", "MagicBrush/images/959.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 419, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add flowers in the girl's hair\n2. add binary code on the computer"}], "image": ["MagicBrush/images/960.jpg", "MagicBrush/images/961.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 420, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add flowers in the girl's hair\n2. add binary code on the computer"}], "image": ["MagicBrush/images/962.jpg", "MagicBrush/images/963.jpg", "MagicBrush/images/964.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 421, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "it should be a tennis ball on the glove."}], "image": ["MagicBrush/images/965.jpg", "MagicBrush/images/966.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 422, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "The background should be a ocean."}], "image": ["MagicBrush/images/967.jpg", "MagicBrush/images/968.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 423, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Could it be two balls?"}], "image": ["MagicBrush/images/969.jpg", "MagicBrush/images/970.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 424, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. it should be a tennis ball on the glove.\n2. The background should be a ocean.\n3. Could it be two balls?"}], "image": ["MagicBrush/images/971.jpg", "MagicBrush/images/972.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 425, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. it should be a tennis ball on the glove.\n2. The background should be a ocean.\n3. Could it be two balls?"}], "image": ["MagicBrush/images/973.jpg", "MagicBrush/images/974.jpg", "MagicBrush/images/975.jpg", "MagicBrush/images/976.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 426, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the people sit down."}], "image": ["MagicBrush/images/977.jpg", "MagicBrush/images/978.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 427, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove the motorcyclist."}], "image": ["MagicBrush/images/979.jpg", "MagicBrush/images/980.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 428, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn the pizza into a croissant"}], "image": ["MagicBrush/images/981.jpg", "MagicBrush/images/982.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 429, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the orange into kiwi fruit"}], "image": ["MagicBrush/images/983.jpg", "MagicBrush/images/984.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 430, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let it be a mushroom salad"}], "image": ["MagicBrush/images/985.jpg", "MagicBrush/images/986.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 431, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn the pizza into a croissant\n2. change the orange into kiwi fruit\n3. let it be a mushroom salad"}], "image": ["MagicBrush/images/987.jpg", "MagicBrush/images/988.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 432, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn the pizza into a croissant\n2. change the orange into kiwi fruit\n3. let it be a mushroom salad"}], "image": ["MagicBrush/images/989.jpg", "MagicBrush/images/990.jpg", "MagicBrush/images/991.jpg", "MagicBrush/images/992.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 433, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make one of the shoes black"}], "image": ["MagicBrush/images/993.jpg", "MagicBrush/images/994.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 434, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if the man had a hat?"}], "image": ["MagicBrush/images/995.jpg", "MagicBrush/images/996.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 435, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if the vegetables are in a bowl?"}], "image": ["MagicBrush/images/997.jpg", "MagicBrush/images/998.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 436, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a drawing of a flower to the fridge."}], "image": ["MagicBrush/images/999.jpg", "MagicBrush/images/1000.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 437, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. What if the vegetables are in a bowl?\n2. Let's add a drawing of a flower to the fridge."}], "image": ["MagicBrush/images/1001.jpg", "MagicBrush/images/1002.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 438, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. What if the vegetables are in a bowl?\n2. Let's add a drawing of a flower to the fridge."}], "image": ["MagicBrush/images/1003.jpg", "MagicBrush/images/1004.jpg", "MagicBrush/images/1005.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 439, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the sun shine."}], "image": ["MagicBrush/images/1006.jpg", "MagicBrush/images/1007.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 440, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "It could be a microwave next to the woman."}], "image": ["MagicBrush/images/1008.jpg", "MagicBrush/images/1009.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 441, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can it be a boy not a girl?"}], "image": ["MagicBrush/images/1010.jpg", "MagicBrush/images/1011.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 442, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Could the woman hold a banana?"}], "image": ["MagicBrush/images/1012.jpg", "MagicBrush/images/1013.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 443, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. It could be a microwave next to the woman.\n2. Can it be a boy not a girl?\n3. Could the woman hold a banana?"}], "image": ["MagicBrush/images/1014.jpg", "MagicBrush/images/1015.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 444, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. It could be a microwave next to the woman.\n2. Can it be a boy not a girl?\n3. Could the woman hold a banana?"}], "image": ["MagicBrush/images/1016.jpg", "MagicBrush/images/1017.jpg", "MagicBrush/images/1018.jpg", "MagicBrush/images/1019.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 445, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the vehicle be replaced by a bicycle."}], "image": ["MagicBrush/images/1020.jpg", "MagicBrush/images/1021.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 446, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "A girl is eating cereal."}], "image": ["MagicBrush/images/1022.jpg", "MagicBrush/images/1023.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 447, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the woman drink wine."}], "image": ["MagicBrush/images/1024.jpg", "MagicBrush/images/1025.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 448, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a bus station"}], "image": ["MagicBrush/images/1026.jpg", "MagicBrush/images/1027.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 449, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add people"}], "image": ["MagicBrush/images/1028.jpg", "MagicBrush/images/1029.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 450, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add an airplane"}], "image": ["MagicBrush/images/1030.jpg", "MagicBrush/images/1031.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 451, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a bus station\n2. Add people\n3. Add an airplane"}], "image": ["MagicBrush/images/1032.jpg", "MagicBrush/images/1033.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 452, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a bus station\n2. Add people\n3. Add an airplane"}], "image": ["MagicBrush/images/1034.jpg", "MagicBrush/images/1035.jpg", "MagicBrush/images/1036.jpg", "MagicBrush/images/1037.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 453, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if the boy was bald?"}], "image": ["MagicBrush/images/1038.jpg", "MagicBrush/images/1039.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 454, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let a forest be the background."}], "image": ["MagicBrush/images/1040.jpg", "MagicBrush/images/1041.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 455, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the player wear a red hat."}], "image": ["MagicBrush/images/1042.jpg", "MagicBrush/images/1043.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 456, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the toothpicks into candles"}], "image": ["MagicBrush/images/1044.jpg", "MagicBrush/images/1045.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 457, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the plate be made of glass"}], "image": ["MagicBrush/images/1046.jpg", "MagicBrush/images/1047.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 458, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the sandwich be vegan"}], "image": ["MagicBrush/images/1048.jpg", "MagicBrush/images/1049.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 459, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the toothpicks into candles\n2. let the plate be made of glass\n3. let the sandwich be vegan"}], "image": ["MagicBrush/images/1050.jpg", "MagicBrush/images/1051.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 460, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the toothpicks into candles\n2. let the plate be made of glass\n3. let the sandwich be vegan"}], "image": ["MagicBrush/images/1052.jpg", "MagicBrush/images/1053.jpg", "MagicBrush/images/1054.jpg", "MagicBrush/images/1055.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 461, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give the girl a baseball bat instead of a racket"}], "image": ["MagicBrush/images/1056.jpg", "MagicBrush/images/1057.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 462, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the tennis ball"}], "image": ["MagicBrush/images/1058.jpg", "MagicBrush/images/1059.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 463, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. give the girl a baseball bat instead of a racket\n2. remove the tennis ball"}], "image": ["MagicBrush/images/1060.jpg", "MagicBrush/images/1061.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 464, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. give the girl a baseball bat instead of a racket\n2. remove the tennis ball"}], "image": ["MagicBrush/images/1062.jpg", "MagicBrush/images/1063.jpg", "MagicBrush/images/1064.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 465, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give her a baseball cap"}], "image": ["MagicBrush/images/1065.jpg", "MagicBrush/images/1066.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 466, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "get rid of the racket"}], "image": ["MagicBrush/images/1067.jpg", "MagicBrush/images/1068.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 467, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give her jean pants"}], "image": ["MagicBrush/images/1069.jpg", "MagicBrush/images/1070.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 468, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. give her a baseball cap\n2. get rid of the racket\n3. give her jean pants"}], "image": ["MagicBrush/images/1071.jpg", "MagicBrush/images/1072.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 469, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. give her a baseball cap\n2. get rid of the racket\n3. give her jean pants"}], "image": ["MagicBrush/images/1073.jpg", "MagicBrush/images/1074.jpg", "MagicBrush/images/1075.jpg", "MagicBrush/images/1076.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 470, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a cyclist on the street"}], "image": ["MagicBrush/images/1077.jpg", "MagicBrush/images/1078.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 471, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the bus for a school bus"}], "image": ["MagicBrush/images/1079.jpg", "MagicBrush/images/1080.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 472, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a dog walking in front of the guy"}], "image": ["MagicBrush/images/1081.jpg", "MagicBrush/images/1082.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 473, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add a cyclist on the street\n2. Change the bus for a school bus\n3. Add a dog walking in front of the guy"}], "image": ["MagicBrush/images/1083.jpg", "MagicBrush/images/1084.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 474, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add a cyclist on the street\n2. Change the bus for a school bus\n3. Add a dog walking in front of the guy"}], "image": ["MagicBrush/images/1085.jpg", "MagicBrush/images/1086.jpg", "MagicBrush/images/1087.jpg", "MagicBrush/images/1088.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 475, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can it be a cake on the plate?"}], "image": ["MagicBrush/images/1089.jpg", "MagicBrush/images/1090.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 476, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a donut next to the cake."}], "image": ["MagicBrush/images/1091.jpg", "MagicBrush/images/1092.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 477, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can we have a glass of soda next to the plate?"}], "image": ["MagicBrush/images/1093.jpg", "MagicBrush/images/1094.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 478, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can it be a cake on the plate?\n2. Put a donut next to the cake.\n3. Can we have a glass of soda next to the plate?"}], "image": ["MagicBrush/images/1095.jpg", "MagicBrush/images/1096.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 479, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Can it be a cake on the plate?\n2. Put a donut next to the cake.\n3. Can we have a glass of soda next to the plate?"}], "image": ["MagicBrush/images/1097.jpg", "MagicBrush/images/1098.jpg", "MagicBrush/images/1099.jpg", "MagicBrush/images/1100.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 480, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the cat have a short tail."}], "image": ["MagicBrush/images/1101.jpg", "MagicBrush/images/1102.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 481, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace it with a black dog."}], "image": ["MagicBrush/images/1103.jpg", "MagicBrush/images/1104.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 482, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove one of its eyes."}], "image": ["MagicBrush/images/1105.jpg", "MagicBrush/images/1106.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 483, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Give it a hat."}], "image": ["MagicBrush/images/1107.jpg", "MagicBrush/images/1108.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 484, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Replace it with a black dog.\n2. Remove one of its eyes.\n3. Give it a hat."}], "image": ["MagicBrush/images/1109.jpg", "MagicBrush/images/1110.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 485, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Replace it with a black dog.\n2. Remove one of its eyes.\n3. Give it a hat."}], "image": ["MagicBrush/images/1111.jpg", "MagicBrush/images/1112.jpg", "MagicBrush/images/1113.jpg", "MagicBrush/images/1114.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 486, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the woman smile."}], "image": ["MagicBrush/images/1115.jpg", "MagicBrush/images/1116.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 487, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a rat next to the pizza."}], "image": ["MagicBrush/images/1117.jpg", "MagicBrush/images/1118.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 488, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the hair color to pink."}], "image": ["MagicBrush/images/1119.jpg", "MagicBrush/images/1120.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 489, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a whale in the water."}], "image": ["MagicBrush/images/1121.jpg", "MagicBrush/images/1122.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 490, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a policeman in the intersection."}], "image": ["MagicBrush/images/1123.jpg", "MagicBrush/images/1124.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 491, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn the remote into a pizza"}], "image": ["MagicBrush/images/1125.jpg", "MagicBrush/images/1126.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 492, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give the man glasses"}], "image": ["MagicBrush/images/1127.jpg", "MagicBrush/images/1128.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 493, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn the remote into a pizza\n2. give the man glasses"}], "image": ["MagicBrush/images/1129.jpg", "MagicBrush/images/1130.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 494, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn the remote into a pizza\n2. give the man glasses"}], "image": ["MagicBrush/images/1131.jpg", "MagicBrush/images/1132.jpg", "MagicBrush/images/1133.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 495, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the jam be changed to strawberry one."}], "image": ["MagicBrush/images/1134.jpg", "MagicBrush/images/1135.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 496, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a turtle."}], "image": ["MagicBrush/images/1136.jpg", "MagicBrush/images/1137.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 497, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the wooden cabinetry to steel"}], "image": ["MagicBrush/images/1138.jpg", "MagicBrush/images/1139.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 498, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the wood floor to tiled floor"}], "image": ["MagicBrush/images/1140.jpg", "MagicBrush/images/1141.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 499, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add soda cans on the counter"}], "image": ["MagicBrush/images/1142.jpg", "MagicBrush/images/1143.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 500, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the wooden cabinetry to steel\n2. change the wood floor to tiled floor\n3. add soda cans on the counter"}], "image": ["MagicBrush/images/1144.jpg", "MagicBrush/images/1145.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 501, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the wooden cabinetry to steel\n2. change the wood floor to tiled floor\n3. add soda cans on the counter"}], "image": ["MagicBrush/images/1146.jpg", "MagicBrush/images/1147.jpg", "MagicBrush/images/1148.jpg", "MagicBrush/images/1149.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 502, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be pigs in the pen"}], "image": ["MagicBrush/images/1150.jpg", "MagicBrush/images/1151.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 503, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the fence"}], "image": ["MagicBrush/images/1152.jpg", "MagicBrush/images/1153.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 504, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a child point at the animals"}], "image": ["MagicBrush/images/1154.jpg", "MagicBrush/images/1155.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 505, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be pigs in the pen\n2. remove the fence\n3. let a child point at the animals"}], "image": ["MagicBrush/images/1156.jpg", "MagicBrush/images/1157.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 506, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be pigs in the pen\n2. remove the fence\n3. let a child point at the animals"}], "image": ["MagicBrush/images/1158.jpg", "MagicBrush/images/1159.jpg", "MagicBrush/images/1160.jpg", "MagicBrush/images/1161.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 507, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the horse eat a carrot without the bag"}], "image": ["MagicBrush/images/1162.jpg", "MagicBrush/images/1163.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 508, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a penguin look at the horse"}], "image": ["MagicBrush/images/1164.jpg", "MagicBrush/images/1165.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 509, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the horse close its eyes"}], "image": ["MagicBrush/images/1166.jpg", "MagicBrush/images/1167.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 510, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the horse eat a carrot without the bag\n2. let a penguin look at the horse\n3. let the horse close its eyes"}], "image": ["MagicBrush/images/1168.jpg", "MagicBrush/images/1169.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 511, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the horse eat a carrot without the bag\n2. let a penguin look at the horse\n3. let the horse close its eyes"}], "image": ["MagicBrush/images/1170.jpg", "MagicBrush/images/1171.jpg", "MagicBrush/images/1172.jpg", "MagicBrush/images/1173.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 512, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the woman cry"}], "image": ["MagicBrush/images/1174.jpg", "MagicBrush/images/1175.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 513, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the woman have blonde hair"}], "image": ["MagicBrush/images/1176.jpg", "MagicBrush/images/1177.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 514, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the woman hold a cup"}], "image": ["MagicBrush/images/1178.jpg", "MagicBrush/images/1179.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 515, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the woman cry\n2. let the woman have blonde hair\n3. let the woman hold a cup"}], "image": ["MagicBrush/images/1180.jpg", "MagicBrush/images/1181.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 516, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the woman cry\n2. let the woman have blonde hair\n3. let the woman hold a cup"}], "image": ["MagicBrush/images/1182.jpg", "MagicBrush/images/1183.jpg", "MagicBrush/images/1184.jpg", "MagicBrush/images/1185.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 517, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the fruit be sliced bananas"}], "image": ["MagicBrush/images/1186.jpg", "MagicBrush/images/1187.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 518, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a plate contain sliced bread"}], "image": ["MagicBrush/images/1188.jpg", "MagicBrush/images/1189.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 519, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the plates be on a wooden table"}], "image": ["MagicBrush/images/1190.jpg", "MagicBrush/images/1191.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 520, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the fruit be sliced bananas\n2. let a plate contain sliced bread\n3. let the plates be on a wooden table"}], "image": ["MagicBrush/images/1192.jpg", "MagicBrush/images/1193.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 521, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the fruit be sliced bananas\n2. let a plate contain sliced bread\n3. let the plates be on a wooden table"}], "image": ["MagicBrush/images/1194.jpg", "MagicBrush/images/1195.jpg", "MagicBrush/images/1196.jpg", "MagicBrush/images/1197.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 522, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the stuffed toys into rubber duckies"}], "image": ["MagicBrush/images/1198.jpg", "MagicBrush/images/1199.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 523, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the dogs sleep in a basket"}], "image": ["MagicBrush/images/1200.jpg", "MagicBrush/images/1201.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 524, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a spider be in the basket"}], "image": ["MagicBrush/images/1202.jpg", "MagicBrush/images/1203.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 525, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the stuffed toys into rubber duckies\n2. let the dogs sleep in a basket\n3. let a spider be in the basket"}], "image": ["MagicBrush/images/1204.jpg", "MagicBrush/images/1205.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 526, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the stuffed toys into rubber duckies\n2. let the dogs sleep in a basket\n3. let a spider be in the basket"}], "image": ["MagicBrush/images/1206.jpg", "MagicBrush/images/1207.jpg", "MagicBrush/images/1208.jpg", "MagicBrush/images/1209.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 527, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the text on the parking meter to say \"NO\""}], "image": ["MagicBrush/images/1210.jpg", "MagicBrush/images/1211.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 528, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be a digital display on the parking meters"}], "image": ["MagicBrush/images/1212.jpg", "MagicBrush/images/1213.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 529, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the text on the parking meter to say \"NO\"\n2. Have there be a digital display on the parking meters"}], "image": ["MagicBrush/images/1214.jpg", "MagicBrush/images/1215.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 530, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the text on the parking meter to say \"NO\"\n2. Have there be a digital display on the parking meters"}], "image": ["MagicBrush/images/1216.jpg", "MagicBrush/images/1217.jpg", "MagicBrush/images/1218.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 531, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Give the bird a long tail."}], "image": ["MagicBrush/images/1219.jpg", "MagicBrush/images/1220.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 532, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add flowers"}], "image": ["MagicBrush/images/1221.jpg", "MagicBrush/images/1222.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 533, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have baby birds follow the large bird."}], "image": ["MagicBrush/images/1223.jpg", "MagicBrush/images/1224.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 534, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Give the bird a long tail.\n2. Add flowers\n3. Have baby birds follow the large bird."}], "image": ["MagicBrush/images/1225.jpg", "MagicBrush/images/1226.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 535, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Give the bird a long tail.\n2. Add flowers\n3. Have baby birds follow the large bird."}], "image": ["MagicBrush/images/1227.jpg", "MagicBrush/images/1228.jpg", "MagicBrush/images/1229.jpg", "MagicBrush/images/1230.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 536, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the table's color to red"}], "image": ["MagicBrush/images/1231.jpg", "MagicBrush/images/1232.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 537, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's give him a yellow shirt"}], "image": ["MagicBrush/images/1233.jpg", "MagicBrush/images/1234.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 538, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the swimsuit color to green"}], "image": ["MagicBrush/images/1235.jpg", "MagicBrush/images/1236.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 539, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the table's color to red\n2. Let's give him a yellow shirt\n3. Change the swimsuit color to green"}], "image": ["MagicBrush/images/1237.jpg", "MagicBrush/images/1238.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 540, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the table's color to red\n2. Let's give him a yellow shirt\n3. Change the swimsuit color to green"}], "image": ["MagicBrush/images/1239.jpg", "MagicBrush/images/1240.jpg", "MagicBrush/images/1241.jpg", "MagicBrush/images/1242.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 541, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the arabic writing be changed to Chinese."}], "image": ["MagicBrush/images/1243.jpg", "MagicBrush/images/1244.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 542, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a car instead of motorcycles"}], "image": ["MagicBrush/images/1245.jpg", "MagicBrush/images/1246.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 543, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "edit the background by removing the museum and placing a castle"}], "image": ["MagicBrush/images/1247.jpg", "MagicBrush/images/1248.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 544, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a car instead of motorcycles\n2. edit the background by removing the museum and placing a castle"}], "image": ["MagicBrush/images/1249.jpg", "MagicBrush/images/1250.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 545, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a car instead of motorcycles\n2. edit the background by removing the museum and placing a castle"}], "image": ["MagicBrush/images/1251.jpg", "MagicBrush/images/1252.jpg", "MagicBrush/images/1253.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 546, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the stuffed animals with a pillow."}], "image": ["MagicBrush/images/1254.jpg", "MagicBrush/images/1255.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 547, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the monitor turn black."}], "image": ["MagicBrush/images/1256.jpg", "MagicBrush/images/1257.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 548, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "replace the donuts with fruits"}], "image": ["MagicBrush/images/1258.jpg", "MagicBrush/images/1259.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 549, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a woman in a bridal gown stand near the cake"}], "image": ["MagicBrush/images/1260.jpg", "MagicBrush/images/1261.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 550, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be heart shaped balloons"}], "image": ["MagicBrush/images/1262.jpg", "MagicBrush/images/1263.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 551, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the donuts with fruits\n2. let a woman in a bridal gown stand near the cake\n3. let there be heart shaped balloons"}], "image": ["MagicBrush/images/1264.jpg", "MagicBrush/images/1265.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 552, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the donuts with fruits\n2. let a woman in a bridal gown stand near the cake\n3. let there be heart shaped balloons"}], "image": ["MagicBrush/images/1266.jpg", "MagicBrush/images/1267.jpg", "MagicBrush/images/1268.jpg", "MagicBrush/images/1269.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 553, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a cat on the roof."}], "image": ["MagicBrush/images/1270.jpg", "MagicBrush/images/1271.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 554, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the fence be made of wood."}], "image": ["MagicBrush/images/1272.jpg", "MagicBrush/images/1273.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 555, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cat wear a bow tie instead of a tie"}], "image": ["MagicBrush/images/1274.jpg", "MagicBrush/images/1275.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 556, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cat have blue eyes"}], "image": ["MagicBrush/images/1276.jpg", "MagicBrush/images/1277.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 557, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cat lay on a wooden floor instead of a carpet"}], "image": ["MagicBrush/images/1278.jpg", "MagicBrush/images/1279.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 558, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat wear a bow tie instead of a tie\n2. let the cat have blue eyes\n3. let the cat lay on a wooden floor instead of a carpet"}], "image": ["MagicBrush/images/1280.jpg", "MagicBrush/images/1281.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 559, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cat wear a bow tie instead of a tie\n2. let the cat have blue eyes\n3. let the cat lay on a wooden floor instead of a carpet"}], "image": ["MagicBrush/images/1282.jpg", "MagicBrush/images/1283.jpg", "MagicBrush/images/1284.jpg", "MagicBrush/images/1285.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 560, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the ball"}], "image": ["MagicBrush/images/1286.jpg", "MagicBrush/images/1287.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 561, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the bat"}], "image": ["MagicBrush/images/1288.jpg", "MagicBrush/images/1289.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 562, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the batter"}], "image": ["MagicBrush/images/1290.jpg", "MagicBrush/images/1291.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 563, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the ball\n2. remove the bat\n3. remove the batter"}], "image": ["MagicBrush/images/1292.jpg", "MagicBrush/images/1293.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 564, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the ball\n2. remove the bat\n3. remove the batter"}], "image": ["MagicBrush/images/1294.jpg", "MagicBrush/images/1295.jpg", "MagicBrush/images/1296.jpg", "MagicBrush/images/1297.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 565, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the woman kiss"}], "image": ["MagicBrush/images/1298.jpg", "MagicBrush/images/1299.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 566, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the woman make a heart sign with her hands"}], "image": ["MagicBrush/images/1300.jpg", "MagicBrush/images/1301.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 567, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the bowl have chocolate sauce"}], "image": ["MagicBrush/images/1302.jpg", "MagicBrush/images/1303.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 568, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the woman kiss\n2. let the woman make a heart sign with her hands\n3. let the bowl have chocolate sauce"}], "image": ["MagicBrush/images/1304.jpg", "MagicBrush/images/1305.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 569, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the woman kiss\n2. let the woman make a heart sign with her hands\n3. let the bowl have chocolate sauce"}], "image": ["MagicBrush/images/1306.jpg", "MagicBrush/images/1307.jpg", "MagicBrush/images/1308.jpg", "MagicBrush/images/1309.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 570, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Open the lid of a toilet."}], "image": ["MagicBrush/images/1310.jpg", "MagicBrush/images/1311.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 571, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "replace the chocolate with berries"}], "image": ["MagicBrush/images/1312.jpg", "MagicBrush/images/1313.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 572, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a lizard on the counter"}], "image": ["MagicBrush/images/1314.jpg", "MagicBrush/images/1315.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 573, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the plate contain ice cream"}], "image": ["MagicBrush/images/1316.jpg", "MagicBrush/images/1317.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 574, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the chocolate with berries\n2. add a lizard on the counter\n3. let the plate contain ice cream"}], "image": ["MagicBrush/images/1318.jpg", "MagicBrush/images/1319.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 575, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. replace the chocolate with berries\n2. add a lizard on the counter\n3. let the plate contain ice cream"}], "image": ["MagicBrush/images/1320.jpg", "MagicBrush/images/1321.jpg", "MagicBrush/images/1322.jpg", "MagicBrush/images/1323.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 576, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "get rid of the cat"}], "image": ["MagicBrush/images/1324.jpg", "MagicBrush/images/1325.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 577, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a vase on top of the table"}], "image": ["MagicBrush/images/1326.jpg", "MagicBrush/images/1327.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 578, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. get rid of the cat\n2. put a vase on top of the table"}], "image": ["MagicBrush/images/1328.jpg", "MagicBrush/images/1329.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 579, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. get rid of the cat\n2. put a vase on top of the table"}], "image": ["MagicBrush/images/1330.jpg", "MagicBrush/images/1331.jpg", "MagicBrush/images/1332.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 580, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "get rid of the vase on top of the table"}], "image": ["MagicBrush/images/1333.jpg", "MagicBrush/images/1334.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 581, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a cowboy hat to the giraffe."}], "image": ["MagicBrush/images/1335.jpg", "MagicBrush/images/1336.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 582, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a dog in the grass."}], "image": ["MagicBrush/images/1337.jpg", "MagicBrush/images/1338.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 583, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Let's add a cowboy hat to the giraffe.\n2. Let's add a dog in the grass."}], "image": ["MagicBrush/images/1339.jpg", "MagicBrush/images/1340.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 584, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Let's add a cowboy hat to the giraffe.\n2. Let's add a dog in the grass."}], "image": ["MagicBrush/images/1341.jpg", "MagicBrush/images/1342.jpg", "MagicBrush/images/1343.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 585, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a cat be near the hydrant"}], "image": ["MagicBrush/images/1344.jpg", "MagicBrush/images/1345.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 586, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the hydrant be small and red"}], "image": ["MagicBrush/images/1346.jpg", "MagicBrush/images/1347.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 587, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the fire truck into a pickup truck"}], "image": ["MagicBrush/images/1348.jpg", "MagicBrush/images/1349.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 588, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let a cat be near the hydrant\n2. let the hydrant be small and red\n3. change the fire truck into a pickup truck"}], "image": ["MagicBrush/images/1350.jpg", "MagicBrush/images/1351.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 589, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let a cat be near the hydrant\n2. let the hydrant be small and red\n3. change the fire truck into a pickup truck"}], "image": ["MagicBrush/images/1352.jpg", "MagicBrush/images/1353.jpg", "MagicBrush/images/1354.jpg", "MagicBrush/images/1355.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 590, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "swap the kites with drones"}], "image": ["MagicBrush/images/1356.jpg", "MagicBrush/images/1357.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 591, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn the mountain in a waterfall"}], "image": ["MagicBrush/images/1358.jpg", "MagicBrush/images/1359.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 592, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the kites with drones\n2. turn the mountain in a waterfall"}], "image": ["MagicBrush/images/1360.jpg", "MagicBrush/images/1361.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 593, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. swap the kites with drones\n2. turn the mountain in a waterfall"}], "image": ["MagicBrush/images/1362.jpg", "MagicBrush/images/1363.jpg", "MagicBrush/images/1364.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 594, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "a parakeet should be sitting on the knit item"}], "image": ["MagicBrush/images/1365.jpg", "MagicBrush/images/1366.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 595, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "a cat should be watching the parakeet while sitting in a flower pot"}], "image": ["MagicBrush/images/1367.jpg", "MagicBrush/images/1368.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 596, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "the cat should have a hat on"}], "image": ["MagicBrush/images/1369.jpg", "MagicBrush/images/1370.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 597, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. a parakeet should be sitting on the knit item\n2. a cat should be watching the parakeet while sitting in a flower pot\n3. the cat should have a hat on"}], "image": ["MagicBrush/images/1371.jpg", "MagicBrush/images/1372.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 598, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. a parakeet should be sitting on the knit item\n2. a cat should be watching the parakeet while sitting in a flower pot\n3. the cat should have a hat on"}], "image": ["MagicBrush/images/1373.jpg", "MagicBrush/images/1374.jpg", "MagicBrush/images/1375.jpg", "MagicBrush/images/1376.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 599, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the apples green"}], "image": ["MagicBrush/images/1377.jpg", "MagicBrush/images/1378.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 600, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Get rid of the jar of cookies"}], "image": ["MagicBrush/images/1379.jpg", "MagicBrush/images/1380.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 601, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the cake a chocolate cake"}], "image": ["MagicBrush/images/1381.jpg", "MagicBrush/images/1382.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 602, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the apples green\n2. Get rid of the jar of cookies\n3. Make the cake a chocolate cake"}], "image": ["MagicBrush/images/1383.jpg", "MagicBrush/images/1384.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 603, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the apples green\n2. Get rid of the jar of cookies\n3. Make the cake a chocolate cake"}], "image": ["MagicBrush/images/1385.jpg", "MagicBrush/images/1386.jpg", "MagicBrush/images/1387.jpg", "MagicBrush/images/1388.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 604, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the man press the keyboard."}], "image": ["MagicBrush/images/1389.jpg", "MagicBrush/images/1390.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 605, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if the child was holding a bottle of peper?"}], "image": ["MagicBrush/images/1391.jpg", "MagicBrush/images/1392.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 606, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "What if there was a drawing of a bird on his shirt?"}], "image": ["MagicBrush/images/1393.jpg", "MagicBrush/images/1394.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 607, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. What if the child was holding a bottle of peper?\n2. What if there was a drawing of a bird on his shirt?"}], "image": ["MagicBrush/images/1395.jpg", "MagicBrush/images/1396.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 608, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. What if the child was holding a bottle of peper?\n2. What if there was a drawing of a bird on his shirt?"}], "image": ["MagicBrush/images/1397.jpg", "MagicBrush/images/1398.jpg", "MagicBrush/images/1399.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 609, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "The car should be white."}], "image": ["MagicBrush/images/1400.jpg", "MagicBrush/images/1401.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 610, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a red sign insted of the bike."}], "image": ["MagicBrush/images/1402.jpg", "MagicBrush/images/1403.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 611, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. The car should be white.\n2. Put a red sign insted of the bike."}], "image": ["MagicBrush/images/1404.jpg", "MagicBrush/images/1405.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 612, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. The car should be white.\n2. Put a red sign insted of the bike."}], "image": ["MagicBrush/images/1406.jpg", "MagicBrush/images/1407.jpg", "MagicBrush/images/1408.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 613, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a ball on the sidewalk."}], "image": ["MagicBrush/images/1409.jpg", "MagicBrush/images/1410.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 614, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the trees to palm trees"}], "image": ["MagicBrush/images/1411.jpg", "MagicBrush/images/1412.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 615, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the people"}], "image": ["MagicBrush/images/1413.jpg", "MagicBrush/images/1414.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 616, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a parking meter next to the bus"}], "image": ["MagicBrush/images/1415.jpg", "MagicBrush/images/1416.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 617, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the people\n2. put a parking meter next to the bus"}], "image": ["MagicBrush/images/1417.jpg", "MagicBrush/images/1418.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 618, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the people\n2. put a parking meter next to the bus"}], "image": ["MagicBrush/images/1419.jpg", "MagicBrush/images/1420.jpg", "MagicBrush/images/1421.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 619, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a penguin near the polar bears."}], "image": ["MagicBrush/images/1422.jpg", "MagicBrush/images/1423.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 620, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the clock tower to a bell tower"}], "image": ["MagicBrush/images/1424.jpg", "MagicBrush/images/1425.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 621, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the patch of flowers be daffodils"}], "image": ["MagicBrush/images/1426.jpg", "MagicBrush/images/1427.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 622, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a rocket in the sky"}], "image": ["MagicBrush/images/1428.jpg", "MagicBrush/images/1429.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 623, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the clock tower to a bell tower\n2. let the patch of flowers be daffodils\n3. add a rocket in the sky"}], "image": ["MagicBrush/images/1430.jpg", "MagicBrush/images/1431.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 624, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the clock tower to a bell tower\n2. let the patch of flowers be daffodils\n3. add a rocket in the sky"}], "image": ["MagicBrush/images/1432.jpg", "MagicBrush/images/1433.jpg", "MagicBrush/images/1434.jpg", "MagicBrush/images/1435.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 625, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the surfboard be green"}], "image": ["MagicBrush/images/1436.jpg", "MagicBrush/images/1437.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 626, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the words"}], "image": ["MagicBrush/images/1438.jpg", "MagicBrush/images/1439.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 627, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a shark next to the surfboard"}], "image": ["MagicBrush/images/1440.jpg", "MagicBrush/images/1441.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 628, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the surfboard be green\n2. remove the words\n3. add a shark next to the surfboard"}], "image": ["MagicBrush/images/1442.jpg", "MagicBrush/images/1443.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 629, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the surfboard be green\n2. remove the words\n3. add a shark next to the surfboard"}], "image": ["MagicBrush/images/1444.jpg", "MagicBrush/images/1445.jpg", "MagicBrush/images/1446.jpg", "MagicBrush/images/1447.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 630, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have a cruise ship pull into the harbor."}], "image": ["MagicBrush/images/1448.jpg", "MagicBrush/images/1449.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 631, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add hot air balloons"}], "image": ["MagicBrush/images/1450.jpg", "MagicBrush/images/1451.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 632, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Have a cruise ship pull into the harbor.\n2. Add hot air balloons"}], "image": ["MagicBrush/images/1452.jpg", "MagicBrush/images/1453.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 633, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Have a cruise ship pull into the harbor.\n2. Add hot air balloons"}], "image": ["MagicBrush/images/1454.jpg", "MagicBrush/images/1455.jpg", "MagicBrush/images/1456.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 634, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the kid sleep"}], "image": ["MagicBrush/images/1457.jpg", "MagicBrush/images/1458.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 635, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give the person a bowl"}], "image": ["MagicBrush/images/1459.jpg", "MagicBrush/images/1460.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 636, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the kid sleep\n2. give the person a bowl"}], "image": ["MagicBrush/images/1461.jpg", "MagicBrush/images/1462.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 637, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the kid sleep\n2. give the person a bowl"}], "image": ["MagicBrush/images/1463.jpg", "MagicBrush/images/1464.jpg", "MagicBrush/images/1465.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 638, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the older man have dark hair"}], "image": ["MagicBrush/images/1466.jpg", "MagicBrush/images/1467.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 639, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the ladies wear red dresses"}], "image": ["MagicBrush/images/1468.jpg", "MagicBrush/images/1469.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 640, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the older man wear a cowboy hat"}], "image": ["MagicBrush/images/1470.jpg", "MagicBrush/images/1471.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 641, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the older man have dark hair\n2. let the ladies wear red dresses\n3. let the older man wear a cowboy hat"}], "image": ["MagicBrush/images/1472.jpg", "MagicBrush/images/1473.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 642, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the older man have dark hair\n2. let the ladies wear red dresses\n3. let the older man wear a cowboy hat"}], "image": ["MagicBrush/images/1474.jpg", "MagicBrush/images/1475.jpg", "MagicBrush/images/1476.jpg", "MagicBrush/images/1477.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 643, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the man's top blue."}], "image": ["MagicBrush/images/1478.jpg", "MagicBrush/images/1479.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 644, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove the frisbee."}], "image": ["MagicBrush/images/1480.jpg", "MagicBrush/images/1481.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 645, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make all the grass green."}], "image": ["MagicBrush/images/1482.jpg", "MagicBrush/images/1483.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 646, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the man's top blue.\n2. Remove the frisbee.\n3. Make all the grass green."}], "image": ["MagicBrush/images/1484.jpg", "MagicBrush/images/1485.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 647, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the man's top blue.\n2. Remove the frisbee.\n3. Make all the grass green."}], "image": ["MagicBrush/images/1486.jpg", "MagicBrush/images/1487.jpg", "MagicBrush/images/1488.jpg", "MagicBrush/images/1489.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 648, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the yellow letters"}], "image": ["MagicBrush/images/1490.jpg", "MagicBrush/images/1491.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 649, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the USA flag into a Japanese flag"}], "image": ["MagicBrush/images/1492.jpg", "MagicBrush/images/1493.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 650, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a pilot walk next to the plane"}], "image": ["MagicBrush/images/1494.jpg", "MagicBrush/images/1495.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 651, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the yellow letters\n2. change the USA flag into a Japanese flag\n3. let a pilot walk next to the plane"}], "image": ["MagicBrush/images/1496.jpg", "MagicBrush/images/1497.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 652, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. remove the yellow letters\n2. change the USA flag into a Japanese flag\n3. let a pilot walk next to the plane"}], "image": ["MagicBrush/images/1498.jpg", "MagicBrush/images/1499.jpg", "MagicBrush/images/1500.jpg", "MagicBrush/images/1501.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 653, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the van turn black."}], "image": ["MagicBrush/images/1502.jpg", "MagicBrush/images/1503.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 654, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Split the top layer so there's an extra one."}], "image": ["MagicBrush/images/1504.jpg", "MagicBrush/images/1505.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 655, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put bride and groom toppers on it."}], "image": ["MagicBrush/images/1506.jpg", "MagicBrush/images/1507.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 656, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove the frosting between layers."}], "image": ["MagicBrush/images/1508.jpg", "MagicBrush/images/1509.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 657, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Split the top layer so there's an extra one.\n2. Put bride and groom toppers on it.\n3. Remove the frosting between layers."}], "image": ["MagicBrush/images/1510.jpg", "MagicBrush/images/1511.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 658, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Split the top layer so there's an extra one.\n2. Put bride and groom toppers on it.\n3. Remove the frosting between layers."}], "image": ["MagicBrush/images/1512.jpg", "MagicBrush/images/1513.jpg", "MagicBrush/images/1514.jpg", "MagicBrush/images/1515.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 659, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a tennis racket next to the bat"}], "image": ["MagicBrush/images/1516.jpg", "MagicBrush/images/1517.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 660, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "get rid of the baseball bat"}], "image": ["MagicBrush/images/1518.jpg", "MagicBrush/images/1519.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 661, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a tennis racket next to the bat\n2. get rid of the baseball bat"}], "image": ["MagicBrush/images/1520.jpg", "MagicBrush/images/1521.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 662, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a tennis racket next to the bat\n2. get rid of the baseball bat"}], "image": ["MagicBrush/images/1522.jpg", "MagicBrush/images/1523.jpg", "MagicBrush/images/1524.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 663, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be a game show on TV"}], "image": ["MagicBrush/images/1525.jpg", "MagicBrush/images/1526.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 664, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be granite floor in the kitchen"}], "image": ["MagicBrush/images/1527.jpg", "MagicBrush/images/1528.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 665, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cabinets be made of dark wood"}], "image": ["MagicBrush/images/1529.jpg", "MagicBrush/images/1530.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 666, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be a game show on TV\n2. let there be granite floor in the kitchen\n3. let the cabinets be made of dark wood"}], "image": ["MagicBrush/images/1531.jpg", "MagicBrush/images/1532.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 667, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let there be a game show on TV\n2. let there be granite floor in the kitchen\n3. let the cabinets be made of dark wood"}], "image": ["MagicBrush/images/1533.jpg", "MagicBrush/images/1534.jpg", "MagicBrush/images/1535.jpg", "MagicBrush/images/1536.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 668, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn the frisbee into a soccer ball"}], "image": ["MagicBrush/images/1537.jpg", "MagicBrush/images/1538.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 669, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the dog have a newspaper in its mouth"}], "image": ["MagicBrush/images/1539.jpg", "MagicBrush/images/1540.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 670, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be potted plant"}], "image": ["MagicBrush/images/1541.jpg", "MagicBrush/images/1542.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 671, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn the frisbee into a soccer ball\n2. let the dog have a newspaper in its mouth\n3. let there be potted plant"}], "image": ["MagicBrush/images/1543.jpg", "MagicBrush/images/1544.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 672, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. turn the frisbee into a soccer ball\n2. let the dog have a newspaper in its mouth\n3. let there be potted plant"}], "image": ["MagicBrush/images/1545.jpg", "MagicBrush/images/1546.jpg", "MagicBrush/images/1547.jpg", "MagicBrush/images/1548.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 673, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the zebra a regular horse."}], "image": ["MagicBrush/images/1549.jpg", "MagicBrush/images/1550.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 674, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the blood orange be swapped with an apple."}], "image": ["MagicBrush/images/1551.jpg", "MagicBrush/images/1552.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 675, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the ski gear into a scuba gear"}], "image": ["MagicBrush/images/1553.jpg", "MagicBrush/images/1554.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 676, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the tree be conifers"}], "image": ["MagicBrush/images/1555.jpg", "MagicBrush/images/1556.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 677, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a polar bear"}], "image": ["MagicBrush/images/1557.jpg", "MagicBrush/images/1558.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 678, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the ski gear into a scuba gear\n2. let the tree be conifers\n3. add a polar bear"}], "image": ["MagicBrush/images/1559.jpg", "MagicBrush/images/1560.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 679, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the ski gear into a scuba gear\n2. let the tree be conifers\n3. add a polar bear"}], "image": ["MagicBrush/images/1561.jpg", "MagicBrush/images/1562.jpg", "MagicBrush/images/1563.jpg", "MagicBrush/images/1564.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 680, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the man cut pizza with a knife"}], "image": ["MagicBrush/images/1565.jpg", "MagicBrush/images/1566.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 681, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make it a pepperoni pizza"}], "image": ["MagicBrush/images/1567.jpg", "MagicBrush/images/1568.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 682, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the man have a tattoo on his hand"}], "image": ["MagicBrush/images/1569.jpg", "MagicBrush/images/1570.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 683, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the man cut pizza with a knife\n2. make it a pepperoni pizza\n3. let the man have a tattoo on his hand"}], "image": ["MagicBrush/images/1571.jpg", "MagicBrush/images/1572.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 684, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the man cut pizza with a knife\n2. make it a pepperoni pizza\n3. let the man have a tattoo on his hand"}], "image": ["MagicBrush/images/1573.jpg", "MagicBrush/images/1574.jpg", "MagicBrush/images/1575.jpg", "MagicBrush/images/1576.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 685, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the bluebery cake be topped with chocolate syrup."}], "image": ["MagicBrush/images/1577.jpg", "MagicBrush/images/1578.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 686, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the drink be replaced by a glass of milk."}], "image": ["MagicBrush/images/1579.jpg", "MagicBrush/images/1580.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 687, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the truck into a taxi"}], "image": ["MagicBrush/images/1581.jpg", "MagicBrush/images/1582.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 688, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a herd of sheep block the taxi"}], "image": ["MagicBrush/images/1583.jpg", "MagicBrush/images/1584.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 689, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the trees at the back into palm trees"}], "image": ["MagicBrush/images/1585.jpg", "MagicBrush/images/1586.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 690, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the truck into a taxi\n2. let a herd of sheep block the taxi\n3. change the trees at the back into palm trees"}], "image": ["MagicBrush/images/1587.jpg", "MagicBrush/images/1588.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 691, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the truck into a taxi\n2. let a herd of sheep block the taxi\n3. change the trees at the back into palm trees"}], "image": ["MagicBrush/images/1589.jpg", "MagicBrush/images/1590.jpg", "MagicBrush/images/1591.jpg", "MagicBrush/images/1592.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 692, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put strawberry on the plate"}], "image": ["MagicBrush/images/1593.jpg", "MagicBrush/images/1594.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 693, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "leave nothing on top of the cheesecake"}], "image": ["MagicBrush/images/1595.jpg", "MagicBrush/images/1596.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 694, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put strawberry on the plate\n2. leave nothing on top of the cheesecake"}], "image": ["MagicBrush/images/1597.jpg", "MagicBrush/images/1598.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 695, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put strawberry on the plate\n2. leave nothing on top of the cheesecake"}], "image": ["MagicBrush/images/1599.jpg", "MagicBrush/images/1600.jpg", "MagicBrush/images/1601.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 696, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the man ride a motorcycle instead of a bicycle"}], "image": ["MagicBrush/images/1602.jpg", "MagicBrush/images/1603.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 697, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "swap the surfboard for a skateboard"}], "image": ["MagicBrush/images/1604.jpg", "MagicBrush/images/1605.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 698, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the man ride a motorcycle instead of a bicycle\n2. swap the surfboard for a skateboard"}], "image": ["MagicBrush/images/1606.jpg", "MagicBrush/images/1607.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 699, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the man ride a motorcycle instead of a bicycle\n2. swap the surfboard for a skateboard"}], "image": ["MagicBrush/images/1608.jpg", "MagicBrush/images/1609.jpg", "MagicBrush/images/1610.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 700, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the couch be an expensive leather one"}], "image": ["MagicBrush/images/1611.jpg", "MagicBrush/images/1612.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 701, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the window show a garden"}], "image": ["MagicBrush/images/1613.jpg", "MagicBrush/images/1614.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 702, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the ceiling have wooden beams"}], "image": ["MagicBrush/images/1615.jpg", "MagicBrush/images/1616.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 703, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the couch be an expensive leather one\n2. let the window show a garden\n3. let the ceiling have wooden beams"}], "image": ["MagicBrush/images/1617.jpg", "MagicBrush/images/1618.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 704, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the couch be an expensive leather one\n2. let the window show a garden\n3. let the ceiling have wooden beams"}], "image": ["MagicBrush/images/1619.jpg", "MagicBrush/images/1620.jpg", "MagicBrush/images/1621.jpg", "MagicBrush/images/1622.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 705, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the fire truck into a battle tank"}], "image": ["MagicBrush/images/1623.jpg", "MagicBrush/images/1624.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 706, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a man run towards the tank"}], "image": ["MagicBrush/images/1625.jpg", "MagicBrush/images/1626.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 707, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be an oak tree"}], "image": ["MagicBrush/images/1627.jpg", "MagicBrush/images/1628.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 708, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the fire truck into a battle tank\n2. let a man run towards the tank\n3. let there be an oak tree"}], "image": ["MagicBrush/images/1629.jpg", "MagicBrush/images/1630.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 709, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the fire truck into a battle tank\n2. let a man run towards the tank\n3. let there be an oak tree"}], "image": ["MagicBrush/images/1631.jpg", "MagicBrush/images/1632.jpg", "MagicBrush/images/1633.jpg", "MagicBrush/images/1634.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 710, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let's add a dog next to the cows."}], "image": ["MagicBrush/images/1635.jpg", "MagicBrush/images/1636.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 711, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a cup of coffee."}], "image": ["MagicBrush/images/1637.jpg", "MagicBrush/images/1638.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 712, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove the salad on the side."}], "image": ["MagicBrush/images/1639.jpg", "MagicBrush/images/1640.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 713, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add a cup of coffee.\n2. Remove the salad on the side."}], "image": ["MagicBrush/images/1641.jpg", "MagicBrush/images/1642.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 714, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Add a cup of coffee.\n2. Remove the salad on the side."}], "image": ["MagicBrush/images/1643.jpg", "MagicBrush/images/1644.jpg", "MagicBrush/images/1645.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 715, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let a green towel be hung in the bathroom."}], "image": ["MagicBrush/images/1646.jpg", "MagicBrush/images/1647.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 716, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add cookies to the tray."}], "image": ["MagicBrush/images/1648.jpg", "MagicBrush/images/1649.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 717, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the bag of chips into a backpack"}], "image": ["MagicBrush/images/1650.jpg", "MagicBrush/images/1651.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 718, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the playstation controller"}], "image": ["MagicBrush/images/1652.jpg", "MagicBrush/images/1653.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 719, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the bag of chips into a backpack\n2. remove the playstation controller"}], "image": ["MagicBrush/images/1654.jpg", "MagicBrush/images/1655.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 720, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the bag of chips into a backpack\n2. remove the playstation controller"}], "image": ["MagicBrush/images/1656.jpg", "MagicBrush/images/1657.jpg", "MagicBrush/images/1658.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 721, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove one of the pizzas."}], "image": ["MagicBrush/images/1659.jpg", "MagicBrush/images/1660.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 722, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the toppings to pepperoni and cheese."}], "image": ["MagicBrush/images/1661.jpg", "MagicBrush/images/1662.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 723, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Remove one of the pizzas.\n2. Change the toppings to pepperoni and cheese."}], "image": ["MagicBrush/images/1663.jpg", "MagicBrush/images/1664.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 724, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Remove one of the pizzas.\n2. Change the toppings to pepperoni and cheese."}], "image": ["MagicBrush/images/1665.jpg", "MagicBrush/images/1666.jpg", "MagicBrush/images/1667.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 725, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a lion in the place of the donkey"}], "image": ["MagicBrush/images/1668.jpg", "MagicBrush/images/1669.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 726, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "turn the basin into a plastic pool"}], "image": ["MagicBrush/images/1670.jpg", "MagicBrush/images/1671.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 727, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a lion in the place of the donkey\n2. turn the basin into a plastic pool"}], "image": ["MagicBrush/images/1672.jpg", "MagicBrush/images/1673.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 728, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. put a lion in the place of the donkey\n2. turn the basin into a plastic pool"}], "image": ["MagicBrush/images/1674.jpg", "MagicBrush/images/1675.jpg", "MagicBrush/images/1676.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 729, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the ship be blue"}], "image": ["MagicBrush/images/1677.jpg", "MagicBrush/images/1678.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 730, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a cat"}], "image": ["MagicBrush/images/1679.jpg", "MagicBrush/images/1680.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 731, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a table in front"}], "image": ["MagicBrush/images/1681.jpg", "MagicBrush/images/1682.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 732, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the ship be blue\n2. add a cat\n3. add a table in front"}], "image": ["MagicBrush/images/1683.jpg", "MagicBrush/images/1684.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 733, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the ship be blue\n2. add a cat\n3. add a table in front"}], "image": ["MagicBrush/images/1685.jpg", "MagicBrush/images/1686.jpg", "MagicBrush/images/1687.jpg", "MagicBrush/images/1688.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 734, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be a basket of fruit on the counter"}], "image": ["MagicBrush/images/1689.jpg", "MagicBrush/images/1690.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 735, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "take the objects off the dresser"}], "image": ["MagicBrush/images/1691.jpg", "MagicBrush/images/1692.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 736, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change kids to men"}], "image": ["MagicBrush/images/1693.jpg", "MagicBrush/images/1694.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 737, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a dog catching a ball"}], "image": ["MagicBrush/images/1695.jpg", "MagicBrush/images/1696.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 738, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a rainbow in the sky"}], "image": ["MagicBrush/images/1697.jpg", "MagicBrush/images/1698.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 739, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change kids to men\n2. add a dog catching a ball\n3. add a rainbow in the sky"}], "image": ["MagicBrush/images/1699.jpg", "MagicBrush/images/1700.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 740, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change kids to men\n2. add a dog catching a ball\n3. add a rainbow in the sky"}], "image": ["MagicBrush/images/1701.jpg", "MagicBrush/images/1702.jpg", "MagicBrush/images/1703.jpg", "MagicBrush/images/1704.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 741, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put popcorn in the plate."}], "image": ["MagicBrush/images/1705.jpg", "MagicBrush/images/1706.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 742, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the dog's eyes closed."}], "image": ["MagicBrush/images/1707.jpg", "MagicBrush/images/1708.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 743, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put popcorn in the plate.\n2. Make the dog's eyes closed."}], "image": ["MagicBrush/images/1709.jpg", "MagicBrush/images/1710.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 744, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put popcorn in the plate.\n2. Make the dog's eyes closed."}], "image": ["MagicBrush/images/1711.jpg", "MagicBrush/images/1712.jpg", "MagicBrush/images/1713.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 745, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put skis on the wheel"}], "image": ["MagicBrush/images/1714.jpg", "MagicBrush/images/1715.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 746, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a wooden floor on the kitchen."}], "image": ["MagicBrush/images/1716.jpg", "MagicBrush/images/1717.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 747, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "put a table on the kitchen."}], "image": ["MagicBrush/images/1718.jpg", "MagicBrush/images/1719.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 748, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a microwave on the counter."}], "image": ["MagicBrush/images/1720.jpg", "MagicBrush/images/1721.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 749, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put a wooden floor on the kitchen.\n2. put a table on the kitchen.\n3. Put a microwave on the counter."}], "image": ["MagicBrush/images/1722.jpg", "MagicBrush/images/1723.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 750, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put a wooden floor on the kitchen.\n2. put a table on the kitchen.\n3. Put a microwave on the counter."}], "image": ["MagicBrush/images/1724.jpg", "MagicBrush/images/1725.jpg", "MagicBrush/images/1726.jpg", "MagicBrush/images/1727.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 751, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the carpet be changed to wooden floor."}], "image": ["MagicBrush/images/1728.jpg", "MagicBrush/images/1729.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 752, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "An airplane is smoking from the cockpit."}], "image": ["MagicBrush/images/1730.jpg", "MagicBrush/images/1731.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 753, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "It should be just one cow."}], "image": ["MagicBrush/images/1732.jpg", "MagicBrush/images/1733.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 754, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Could it be a river on the background?"}], "image": ["MagicBrush/images/1734.jpg", "MagicBrush/images/1735.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 755, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Add a goat next to the cow."}], "image": ["MagicBrush/images/1736.jpg", "MagicBrush/images/1737.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 756, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. It should be just one cow.\n2. Could it be a river on the background?\n3. Add a goat next to the cow."}], "image": ["MagicBrush/images/1738.jpg", "MagicBrush/images/1739.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 757, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. It should be just one cow.\n2. Could it be a river on the background?\n3. Add a goat next to the cow."}], "image": ["MagicBrush/images/1740.jpg", "MagicBrush/images/1741.jpg", "MagicBrush/images/1742.jpg", "MagicBrush/images/1743.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 758, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the tea into cappuccino"}], "image": ["MagicBrush/images/1744.jpg", "MagicBrush/images/1745.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 759, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the sandwich to salad"}], "image": ["MagicBrush/images/1746.jpg", "MagicBrush/images/1747.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 760, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a bottle of sauce"}], "image": ["MagicBrush/images/1748.jpg", "MagicBrush/images/1749.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 761, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the tea into cappuccino\n2. change the sandwich to salad\n3. add a bottle of sauce"}], "image": ["MagicBrush/images/1750.jpg", "MagicBrush/images/1751.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 762, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the tea into cappuccino\n2. change the sandwich to salad\n3. add a bottle of sauce"}], "image": ["MagicBrush/images/1752.jpg", "MagicBrush/images/1753.jpg", "MagicBrush/images/1754.jpg", "MagicBrush/images/1755.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 763, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the plane further away."}], "image": ["MagicBrush/images/1756.jpg", "MagicBrush/images/1757.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 764, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the cart with an upright one."}], "image": ["MagicBrush/images/1758.jpg", "MagicBrush/images/1759.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 765, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the traffic meter with a pole."}], "image": ["MagicBrush/images/1760.jpg", "MagicBrush/images/1761.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 766, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a puppy in the cart."}], "image": ["MagicBrush/images/1762.jpg", "MagicBrush/images/1763.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 767, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Replace the cart with an upright one.\n2. Replace the traffic meter with a pole.\n3. Put a puppy in the cart."}], "image": ["MagicBrush/images/1764.jpg", "MagicBrush/images/1765.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 768, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Replace the cart with an upright one.\n2. Replace the traffic meter with a pole.\n3. Put a puppy in the cart."}], "image": ["MagicBrush/images/1766.jpg", "MagicBrush/images/1767.jpg", "MagicBrush/images/1768.jpg", "MagicBrush/images/1769.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 769, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "There should be a tree on the front."}], "image": ["MagicBrush/images/1770.jpg", "MagicBrush/images/1771.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 770, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "It should be a mountain in the background."}], "image": ["MagicBrush/images/1772.jpg", "MagicBrush/images/1773.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 771, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "The trash can should be blue."}], "image": ["MagicBrush/images/1774.jpg", "MagicBrush/images/1775.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 772, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. There should be a tree on the front.\n2. It should be a mountain in the background.\n3. The trash can should be blue."}], "image": ["MagicBrush/images/1776.jpg", "MagicBrush/images/1777.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 773, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. There should be a tree on the front.\n2. It should be a mountain in the background.\n3. The trash can should be blue."}], "image": ["MagicBrush/images/1778.jpg", "MagicBrush/images/1779.jpg", "MagicBrush/images/1780.jpg", "MagicBrush/images/1781.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 774, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "cover the bread with sauce and salad"}], "image": ["MagicBrush/images/1782.jpg", "MagicBrush/images/1783.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 775, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a glass of water"}], "image": ["MagicBrush/images/1784.jpg", "MagicBrush/images/1785.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 776, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the fork to a spoon"}], "image": ["MagicBrush/images/1786.jpg", "MagicBrush/images/1787.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 777, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. cover the bread with sauce and salad\n2. add a glass of water\n3. change the fork to a spoon"}], "image": ["MagicBrush/images/1788.jpg", "MagicBrush/images/1789.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 778, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. cover the bread with sauce and salad\n2. add a glass of water\n3. change the fork to a spoon"}], "image": ["MagicBrush/images/1790.jpg", "MagicBrush/images/1791.jpg", "MagicBrush/images/1792.jpg", "MagicBrush/images/1793.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 779, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "have a fly sit on the laptop"}], "image": ["MagicBrush/images/1794.jpg", "MagicBrush/images/1795.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 780, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the scarf multi-colored."}], "image": ["MagicBrush/images/1796.jpg", "MagicBrush/images/1797.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 781, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the woman smile."}], "image": ["MagicBrush/images/1798.jpg", "MagicBrush/images/1799.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 782, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the woman's hair more straightened out."}], "image": ["MagicBrush/images/1800.jpg", "MagicBrush/images/1801.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 783, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the scarf multi-colored.\n2. Make the woman smile.\n3. Make the woman's hair more straightened out."}], "image": ["MagicBrush/images/1802.jpg", "MagicBrush/images/1803.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 784, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the scarf multi-colored.\n2. Make the woman smile.\n3. Make the woman's hair more straightened out."}], "image": ["MagicBrush/images/1804.jpg", "MagicBrush/images/1805.jpg", "MagicBrush/images/1806.jpg", "MagicBrush/images/1807.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 785, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the vegetables into broccoli"}], "image": ["MagicBrush/images/1808.jpg", "MagicBrush/images/1809.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 786, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the sandwiches be changed to risotto"}], "image": ["MagicBrush/images/1810.jpg", "MagicBrush/images/1811.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 787, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the apples be changed to orange slices"}], "image": ["MagicBrush/images/1812.jpg", "MagicBrush/images/1813.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 788, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the vegetables into broccoli\n2. let the sandwiches be changed to risotto\n3. let the apples be changed to orange slices"}], "image": ["MagicBrush/images/1814.jpg", "MagicBrush/images/1815.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 789, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the vegetables into broccoli\n2. let the sandwiches be changed to risotto\n3. let the apples be changed to orange slices"}], "image": ["MagicBrush/images/1816.jpg", "MagicBrush/images/1817.jpg", "MagicBrush/images/1818.jpg", "MagicBrush/images/1819.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 790, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let water run from the faucet."}], "image": ["MagicBrush/images/1820.jpg", "MagicBrush/images/1821.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 791, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Remove the car."}], "image": ["MagicBrush/images/1822.jpg", "MagicBrush/images/1823.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 792, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Replace the sign with a stop sign."}], "image": ["MagicBrush/images/1824.jpg", "MagicBrush/images/1825.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 793, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a squirrel on top of the sign."}], "image": ["MagicBrush/images/1826.jpg", "MagicBrush/images/1827.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 794, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Remove the car.\n2. Replace the sign with a stop sign.\n3. Put a squirrel on top of the sign."}], "image": ["MagicBrush/images/1828.jpg", "MagicBrush/images/1829.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 795, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Remove the car.\n2. Replace the sign with a stop sign.\n3. Put a squirrel on top of the sign."}], "image": ["MagicBrush/images/1830.jpg", "MagicBrush/images/1831.jpg", "MagicBrush/images/1832.jpg", "MagicBrush/images/1833.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 796, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the woman hold a camera"}], "image": ["MagicBrush/images/1834.jpg", "MagicBrush/images/1835.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 797, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give the woman another camera"}], "image": ["MagicBrush/images/1836.jpg", "MagicBrush/images/1837.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 798, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the woman hold a camera\n2. give the woman another camera"}], "image": ["MagicBrush/images/1838.jpg", "MagicBrush/images/1839.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 799, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. make the woman hold a camera\n2. give the woman another camera"}], "image": ["MagicBrush/images/1840.jpg", "MagicBrush/images/1841.jpg", "MagicBrush/images/1842.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 800, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the cup contain flowers"}], "image": ["MagicBrush/images/1843.jpg", "MagicBrush/images/1844.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 801, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let there be a bug on the wood"}], "image": ["MagicBrush/images/1845.jpg", "MagicBrush/images/1846.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 802, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cup contain flowers\n2. let there be a bug on the wood"}], "image": ["MagicBrush/images/1847.jpg", "MagicBrush/images/1848.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 803, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. let the cup contain flowers\n2. let there be a bug on the wood"}], "image": ["MagicBrush/images/1849.jpg", "MagicBrush/images/1850.jpg", "MagicBrush/images/1851.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 804, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the scooter into a skateboard"}], "image": ["MagicBrush/images/1852.jpg", "MagicBrush/images/1853.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 805, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the tree have lush green leaves"}], "image": ["MagicBrush/images/1854.jpg", "MagicBrush/images/1855.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 806, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let the boy wear a superhero costume"}], "image": ["MagicBrush/images/1856.jpg", "MagicBrush/images/1857.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 807, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the scooter into a skateboard\n2. let the tree have lush green leaves\n3. let the boy wear a superhero costume"}], "image": ["MagicBrush/images/1858.jpg", "MagicBrush/images/1859.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 808, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the scooter into a skateboard\n2. let the tree have lush green leaves\n3. let the boy wear a superhero costume"}], "image": ["MagicBrush/images/1860.jpg", "MagicBrush/images/1861.jpg", "MagicBrush/images/1862.jpg", "MagicBrush/images/1863.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 809, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the Washington Monument with the Statue of Liberty"}], "image": ["MagicBrush/images/1864.jpg", "MagicBrush/images/1865.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 810, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add flying eagles over the statue"}], "image": ["MagicBrush/images/1866.jpg", "MagicBrush/images/1867.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 811, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the Washington Monument with the Statue of Liberty\n2. add flying eagles over the statue"}], "image": ["MagicBrush/images/1868.jpg", "MagicBrush/images/1869.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 812, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the Washington Monument with the Statue of Liberty\n2. add flying eagles over the statue"}], "image": ["MagicBrush/images/1870.jpg", "MagicBrush/images/1871.jpg", "MagicBrush/images/1872.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 813, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the recording devices to makeup"}], "image": ["MagicBrush/images/1873.jpg", "MagicBrush/images/1874.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 814, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the laptop into a makeup tray."}], "image": ["MagicBrush/images/1875.jpg", "MagicBrush/images/1876.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 815, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the wood desk a white table."}], "image": ["MagicBrush/images/1877.jpg", "MagicBrush/images/1878.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 816, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the recording devices to makeup\n2. change the laptop into a makeup tray.\n3. make the wood desk a white table."}], "image": ["MagicBrush/images/1879.jpg", "MagicBrush/images/1880.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 817, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the recording devices to makeup\n2. change the laptop into a makeup tray.\n3. make the wood desk a white table."}], "image": ["MagicBrush/images/1881.jpg", "MagicBrush/images/1882.jpg", "MagicBrush/images/1883.jpg", "MagicBrush/images/1884.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 818, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the flowers on the wallet to be white"}], "image": ["MagicBrush/images/1885.jpg", "MagicBrush/images/1886.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 819, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Have there be a bee on the wallet"}], "image": ["MagicBrush/images/1887.jpg", "MagicBrush/images/1888.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 820, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the flowers on the wallet to be white\n2. Have there be a bee on the wallet"}], "image": ["MagicBrush/images/1889.jpg", "MagicBrush/images/1890.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 821, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Change the flowers on the wallet to be white\n2. Have there be a bee on the wallet"}], "image": ["MagicBrush/images/1891.jpg", "MagicBrush/images/1892.jpg", "MagicBrush/images/1893.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 822, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a pond next to the cows."}], "image": ["MagicBrush/images/1894.jpg", "MagicBrush/images/1895.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 823, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Can we have mountains on the background?"}], "image": ["MagicBrush/images/1896.jpg", "MagicBrush/images/1897.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 824, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a horse insted of the goats."}], "image": ["MagicBrush/images/1898.jpg", "MagicBrush/images/1899.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 825, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put a pond next to the cows.\n2. Can we have mountains on the background?\n3. Put a horse insted of the goats."}], "image": ["MagicBrush/images/1900.jpg", "MagicBrush/images/1901.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 826, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Put a pond next to the cows.\n2. Can we have mountains on the background?\n3. Put a horse insted of the goats."}], "image": ["MagicBrush/images/1902.jpg", "MagicBrush/images/1903.jpg", "MagicBrush/images/1904.jpg", "MagicBrush/images/1905.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 827, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Let the elephant turn young."}], "image": ["MagicBrush/images/1906.jpg", "MagicBrush/images/1907.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 828, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "take the papers out the table"}], "image": ["MagicBrush/images/1908.jpg", "MagicBrush/images/1909.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 829, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the glass of water a to go cup."}], "image": ["MagicBrush/images/1910.jpg", "MagicBrush/images/1911.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 830, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Change the tables to a playland."}], "image": ["MagicBrush/images/1912.jpg", "MagicBrush/images/1913.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 831, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the glass of water a to go cup.\n2. Change the tables to a playland."}], "image": ["MagicBrush/images/1914.jpg", "MagicBrush/images/1915.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 832, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. Make the glass of water a to go cup.\n2. Change the tables to a playland."}], "image": ["MagicBrush/images/1916.jpg", "MagicBrush/images/1917.jpg", "MagicBrush/images/1918.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 833, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the red lights to green lights"}], "image": ["MagicBrush/images/1919.jpg", "MagicBrush/images/1920.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 834, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a truck on the street"}], "image": ["MagicBrush/images/1921.jpg", "MagicBrush/images/1922.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 835, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the trees"}], "image": ["MagicBrush/images/1923.jpg", "MagicBrush/images/1924.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 836, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the red lights to green lights\n2. add a truck on the street\n3. remove the trees"}], "image": ["MagicBrush/images/1925.jpg", "MagicBrush/images/1926.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 837, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the red lights to green lights\n2. add a truck on the street\n3. remove the trees"}], "image": ["MagicBrush/images/1927.jpg", "MagicBrush/images/1928.jpg", "MagicBrush/images/1929.jpg", "MagicBrush/images/1930.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 838, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the teddy bear into a stuffed action figure toy"}], "image": ["MagicBrush/images/1931.jpg", "MagicBrush/images/1932.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 839, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "let a child play nearby"}], "image": ["MagicBrush/images/1933.jpg", "MagicBrush/images/1934.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 840, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the teddy bear into a stuffed action figure toy\n2. let a child play nearby"}], "image": ["MagicBrush/images/1935.jpg", "MagicBrush/images/1936.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 841, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. change the teddy bear into a stuffed action figure toy\n2. let a child play nearby"}], "image": ["MagicBrush/images/1937.jpg", "MagicBrush/images/1938.jpg", "MagicBrush/images/1939.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 842, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a cat on the top of a counter"}], "image": ["MagicBrush/images/1940.jpg", "MagicBrush/images/1941.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 843, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "change the yellow lights to white lights"}], "image": ["MagicBrush/images/1942.jpg", "MagicBrush/images/1943.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 844, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "add a champagne bottle"}], "image": ["MagicBrush/images/1944.jpg", "MagicBrush/images/1945.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 845, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a cat on the top of a counter\n2. change the yellow lights to white lights\n3. add a champagne bottle"}], "image": ["MagicBrush/images/1946.jpg", "MagicBrush/images/1947.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 846, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. add a cat on the top of a counter\n2. change the yellow lights to white lights\n3. add a champagne bottle"}], "image": ["MagicBrush/images/1948.jpg", "MagicBrush/images/1949.jpg", "MagicBrush/images/1950.jpg", "MagicBrush/images/1951.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 847, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove that sink"}], "image": ["MagicBrush/images/1952.jpg", "MagicBrush/images/1953.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 848, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "The bed should be red."}], "image": ["MagicBrush/images/1954.jpg", "MagicBrush/images/1955.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 849, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Put a pile of shoes next to the bed."}], "image": ["MagicBrush/images/1956.jpg", "MagicBrush/images/1957.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 850, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Could we have a window next to the bed?"}], "image": ["MagicBrush/images/1958.jpg", "MagicBrush/images/1959.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 851, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. The bed should be red.\n2. Put a pile of shoes next to the bed.\n3. Could we have a window next to the bed?"}], "image": ["MagicBrush/images/1960.jpg", "MagicBrush/images/1961.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 852, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. The bed should be red.\n2. Put a pile of shoes next to the bed.\n3. Could we have a window next to the bed?"}], "image": ["MagicBrush/images/1962.jpg", "MagicBrush/images/1963.jpg", "MagicBrush/images/1964.jpg", "MagicBrush/images/1965.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 853, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "Make the toilet pink."}], "image": ["MagicBrush/images/1966.jpg", "MagicBrush/images/1967.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 854, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "remove the apple"}], "image": ["MagicBrush/images/1968.jpg", "MagicBrush/images/1969.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 855, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "give the man a jacket"}], "image": ["MagicBrush/images/1970.jpg", "MagicBrush/images/1971.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 856, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "swap the bike for a motorcycle"}], "image": ["MagicBrush/images/1972.jpg", "MagicBrush/images/1973.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 857, "conversations": [{"from": "human", "value": "Please provide a list of image edit instructions that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. give the man a jacket\n2. swap the bike for a motorcycle"}], "image": ["MagicBrush/images/1974.jpg", "MagicBrush/images/1975.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide a list of image edit instructions that can transfrom the source image to the target image.", "question_type": "open-ended"}}, {"sample_id": 858, "conversations": [{"from": "human", "value": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.\nImages: <image><image><image>\nStep-by-step Instructions: "}, {"from": "gpt", "value": "1. give the man a jacket\n2. swap the bike for a motorcycle"}], "image": ["MagicBrush/images/1976.jpg", "MagicBrush/images/1977.jpg", "MagicBrush/images/1978.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Given several images that are edited step by step, please respectively provide the image edit instruction between adjacant images.", "question_type": "open-ended"}}, {"sample_id": 859, "conversations": [{"from": "human", "value": "Please provide the image edit instruction that can transfrom the source image to the target image.\nSource Image: <image>\nTarget Image: <image>\nInstruction: "}, {"from": "gpt", "value": "make the cat's nose black"}], "image": ["MagicBrush/images/1979.jpg", "MagicBrush/images/1980.jpg"], "choice_list": null, "metadata": {"dataset": "MagicBrush", "split": "val", "num_sample": 265, "task_instruction": "Please provide the image edit instruction that can transfrom the source image to the target image.", "question_type": "open-ended"}}]