Update app.py

app.py

@@ -2,21 +2,31 @@ import gradio as gr
 from diffusers import StableVideoDiffusionPipeline, EulerDiscreteScheduler
 import torch
 from PIL import Image
-import spaces
+import tempfile
 
 # Load the Stable Video Diffusion model
 model_id = "stabilityai/stable-video-diffusion-img2vid-xt"
-pipe = StableVideoDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, revision="main")
-pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config)
-pipe.to("cuda")
+try:
+    pipe = StableVideoDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, revision="main")
+    pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config)
+    pipe.to("cuda")
+except Exception as e:
+    raise RuntimeError(f"Failed to load the model: {e}")
 
-@spaces.GPU
-def generate_video(image, prompt, num_frames=25, height=576, width=1024):
-    # Convert the image to a format suitable for the pipeline
-    image = Image.open(image)
-    # Generate the video
-    video_frames = pipe(prompt=prompt, init_image=image, num_frames=num_frames, height=height, width=width).frames
-    return video_frames
+def generate_video(image, num_frames=25, height=576, width=1024):
+    try:
+        # Convert the image to a format suitable for the pipeline
+        image = Image.open(image)
+        # Generate the video
+        video_frames = pipe(image=image, num_frames=num_frames, height=height, width=width).frames
+        # Save the video frames to a temporary file
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as temp_video:
+            video_path = temp_video.name
+            # Assuming video_frames is a list of PIL images
+            video_frames[0].save(video_path, save_all=True, append_images=video_frames[1:], duration=100, loop=0)
+        return video_path
+    except Exception as e:
+        raise RuntimeError(f"Failed to generate the video: {e}")
 
 # Create the Gradio interface
 with gr.Blocks() as demo:
@@ -24,7 +34,6 @@ with gr.Blocks() as demo:
     with gr.Row():
         with gr.Column():
             image_input = gr.Image(type="filepath", label="Upload Image")
-            prompt_input = gr.Textbox(lines=2, placeholder="Enter prompt...", label="Prompt")
             num_frames_input = gr.Slider(1, 50, step=1, value=25, label="Number of Frames")
             height_input = gr.Number(label="Resolution Height", value=576)
             width_input = gr.Number(label="Resolution Width", value=1024)
@@ -34,10 +43,11 @@ with gr.Blocks() as demo:
 
     run_button.click(
         generate_video,
-        inputs=[image_input, prompt_input, num_frames_input, height_input, width_input],
+        inputs=[image_input, num_frames_input, height_input, width_input],
         outputs=video_output
     )
 
 # Launch the interface
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()
+