Upload app.py

app.py

@@ -11,14 +11,20 @@ import torch
 
 IMAGE_SIZE = 640
 
-model_path=hf_hub_download("kadirnar/deprem_model_v1", filename="last.pt",revision="main")
+model_path = hf_hub_download(
+    "kadirnar/deprem_model_v1", filename="last.pt", revision="main"
+)
 
 
-current_device='cuda' if torch.cuda.is_available() else 'cpu'
-model_types=["YOLOv5","YOLOv5 + SAHI"]
+current_device = "cuda" if torch.cuda.is_available() else "cpu"
+model_types = ["YOLOv5", "YOLOv5 + SAHI"]
 # Model
 model = AutoDetectionModel.from_pretrained(
-    model_type="yolov5", model_path=model_path, device=current_device, confidence_threshold=0.5, image_size=IMAGE_SIZE
+    model_type="yolov5",
+    model_path=model_path,
+    device=current_device,
+    confidence_threshold=0.5,
+    image_size=IMAGE_SIZE,
 )
 
 
@@ -27,15 +33,15 @@ def sahi_yolo_inference(
     image,
     slice_height=512,
     slice_width=512,
-    overlap_height_ratio=0.2,
-    overlap_width_ratio=0.2,
-    postprocess_type="GREEDYNMM",
-    postprocess_match_metric="IOS",
-    postprocess_match_threshold=0.5,
+    overlap_height_ratio=0.1,
+    overlap_width_ratio=0.1,
+    postprocess_type="NMS",
+    postprocess_match_metric="IOU",
+    postprocess_match_threshold=0.25,
     postprocess_class_agnostic=False,
 ):
 
-    #image_width, image_height = image.size
+    # image_width, image_height = image.size
     # sliced_bboxes = sahi.slicing.get_slice_bboxes(
     #     image_height,
     #     image_width,
@@ -50,22 +56,21 @@ def sahi_yolo_inference(
     #         f"{len(sliced_bboxes)} slices are too much for huggingface spaces, try smaller slice size."
     #     )
 
-
-    rect_th = None or max(round(sum(image.size) / 2 * 0.001), 1)
-    text_th = None or max(rect_th - 1, 1)
+    rect_th = None or max(round(sum(image.size) / 2 * 0.0001), 1)
+    text_th = None or max(rect_th - 2, 1)
 
     if "SAHI" in model_type:
         prediction_result_2 = sahi.predict.get_sliced_prediction(
-        image=image,
-        detection_model=model,
-        slice_height=int(slice_height),
-        slice_width=int(slice_width),
-        overlap_height_ratio=overlap_height_ratio,
-        overlap_width_ratio=overlap_width_ratio,
-        postprocess_type=postprocess_type,
-        postprocess_match_metric=postprocess_match_metric,
-        postprocess_match_threshold=postprocess_match_threshold,
-        postprocess_class_agnostic=postprocess_class_agnostic,
+            image=image,
+            detection_model=model,
+            slice_height=int(slice_height),
+            slice_width=int(slice_width),
+            overlap_height_ratio=overlap_height_ratio,
+            overlap_width_ratio=overlap_width_ratio,
+            postprocess_type=postprocess_type,
+            postprocess_match_metric=postprocess_match_metric,
+            postprocess_match_threshold=postprocess_match_threshold,
+            postprocess_class_agnostic=postprocess_class_agnostic,
         )
         visual_result_2 = sahi.utils.cv.visualize_object_predictions(
             image=numpy.array(image),
@@ -94,41 +99,39 @@ def sahi_yolo_inference(
     # sliced inference
 
 
-
-
 inputs = [
-    gr.inputs.Dropdown(choices=model_types,label="Choose Model Type",type="value",),
-    gr.inputs.Image(type="pil", label="Original Image"),
-    gr.inputs.Number(default=512, label="slice_height"),
-    gr.inputs.Number(default=512, label="slice_width"),
-    gr.inputs.Number(default=0.2, label="overlap_height_ratio"),
-    gr.inputs.Number(default=0.2, label="overlap_width_ratio"),
-    gr.inputs.Dropdown(
+    gr.Dropdown(
+        choices=model_types,
+        label="Choose Model Type",
+        type="value",
+        value=model_types[1],
+    ),
+    gr.Image(type="pil", label="Original Image"),
+    gr.Number(default=512, label="slice_height"),
+    gr.Number(default=512, label="slice_width"),
+    gr.Number(default=0.1, label="overlap_height_ratio"),
+    gr.Number(default=0.1, label="overlap_width_ratio"),
+    gr.Dropdown(
         ["NMS", "GREEDYNMM"],
         type="value",
-        default="GREEDYNMM",
+        value="NMS",
         label="postprocess_type",
     ),
-    gr.inputs.Dropdown(
-        ["IOU", "IOS"], type="value", default="IOS", label="postprocess_type"
-    ),
-    gr.inputs.Number(default=0.5, label="postprocess_match_threshold"),
-    gr.inputs.Checkbox(default=True, label="postprocess_class_agnostic"),
+    gr.Dropdown(["IOU", "IOS"], type="value", value="IOU", label="postprocess_type"),
+    gr.Number(default=0.5, label="postprocess_match_threshold"),
+    gr.Checkbox(default=True, label="postprocess_class_agnostic"),
 ]
 
-outputs = [
-    gr.outputs.Image(type="pil", label="Output")
-]
+outputs = [gr.outputs.Image(type="pil", label="Output")]
 
 title = "Small Object Detection with SAHI + YOLOv5"
 description = "SAHI + YOLOv5 demo for small object detection. Upload an image or click an example image to use."
 article = "<p style='text-align: center'>SAHI is a lightweight vision library for performing large scale object detection/ instance segmentation.. <a href='https://github.com/obss/sahi'>SAHI Github</a> | <a href='https://medium.com/codable/sahi-a-vision-library-for-performing-sliced-inference-on-large-images-small-objects-c8b086af3b80'>SAHI Blog</a> | <a href='https://github.com/fcakyon/yolov5-pip'>YOLOv5 Github</a> </p>"
 examples = [
-    [model_types[0],"26.jpg", 256, 256, 0.2, 0.2, "GREEDYNMM", "IOS", 0.5, True],
-    [model_types[0],"27.jpg", 512, 512, 0.2, 0.2, "GREEDYNMM", "IOS", 0.5, True],
-    [model_types[0],"28.jpg", 512, 512, 0.2, 0.2, "GREEDYNMM", "IOS", 0.5, True],
-    [model_types[0],"31.jpg", 512, 512, 0.2, 0.2, "GREEDYNMM", "IOS", 0.5, True],
-
+    [model_types[1], "26.jpg", 512, 512, 0.1, 0.1, "NMS", "IOU", 0.25, False],
+    [model_types[1], "27.jpg", 512, 512, 0.1, 0.1, "NMS", "IOU", 0.25, False],
+    [model_types[1], "28.jpg", 512, 512, 0.1, 0.1, "NMS", "IOU", 0.25, False],
+    [model_types[1], "31.jpg", 512, 512, 0.1, 0.1, "NMS", "IOU", 0.25, False],
 ]
 gr.Interface(
     sahi_yolo_inference,
@@ -139,4 +142,5 @@ gr.Interface(
     article=article,
     examples=examples,
     theme="huggingface",
-).launch(debug=True, enable_queue=True)
+    cache_examples=True
+).launch(debug=True, enable_queue=True, server_port=8500)