Create app.py

app.py

@@ -0,0 +1,38 @@
+import gradio as gr
+from diffusers import StableVideoDiffusionPipeline
+import torch
+from PIL import Image
+
+# Load the model and pipeline
+model_id = "stabilityai/stable-video-diffusion-img2vid-xt"
+pipe = StableVideoDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
+pipe = pipe.to("cuda")
+
+def generate_video(image, prompt):
+    # Convert the image to a format suitable for the model
+    image = Image.fromarray(image)
+
+    # Generate the video from the image and prompt
+    video = pipe(prompt, image, num_inference_steps=50, guidance_scale=7.5).frames
+
+    # Convert the video frames to a format suitable for Gradio
+    video_path = "output_video.mp4"
+    video[0].save(video_path, save_all=True, append_images=video[1:], duration=100, loop=0)
+
+    return video_path
+
+# Create the Gradio interface
+iface = gr.Interface(
+    fn=generate_video,
+    inputs=[
+        gr.Image(type="pil", label="Input Image"),
+        gr.Textbox(lines=2, placeholder="Enter a prompt...", label="Prompt")
+    ],
+    outputs=gr.Video(label="Generated Video"),
+    title="Stable Video Diffusion img2vid-xt",
+    description="Generate a video from an image using the stabilityai/stable-video-diffusion-img2vid-xt model."
+)
+
+# Launch the interface
+if __name__ == "__main__":
+    iface.launch()