Update app.py

app.py

@@ -19,12 +19,13 @@ pipe = StableVideoDiffusionPipeline.from_pretrained(
 )
 pipe.to("cuda")
 pipe.unet = torch.compile(pipe.unet, mode="reduce-overhead", fullgraph=True)
+pipe.vae = torch.compile(pipe.vae, mode="reduce-overhead", fullgraph=True)
 
 max_64_bit_int = 2**63 - 1
 
 def sample(
     image: Image,
-    seed: Optional[int] = None,
+    seed: Optional[int] = 42,
     randomize_seed: bool = True,
     motion_bucket_id: int = 127,
     fps_id: int = 6,
@@ -33,7 +34,6 @@ def sample(
     decoding_t: int = 3,  # Number of frames decoded at a time! This eats most VRAM. Reduce if necessary.
     device: str = "cuda",
     output_folder: str = "outputs",
-    #progress=gr.Progress(track_tqdm=True)
 ):
     if image.mode == "RGBA":
         image = image.convert("RGB")
@@ -85,7 +85,7 @@ def resize_image(image, output_size=(1024, 576)):
 
 with gr.Blocks() as demo:
   gr.Markdown('''# Community demo for Stable Video Diffusion - Img2Vid - XT ([model](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt), [paper](https://stability.ai/research/stable-video-diffusion-scaling-latent-video-diffusion-models-to-large-datasets))
-#### Research release ([_non-commercial_](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt/blob/main/LICENSE)): generate `4s` vid from a single image at (`25 frames` at `6 fps`). Running on [🧨 diffusers backend](#). [Join the waitlist for Stability's upcoming web experience](https://stability.ai/contact).
+#### Research release ([_non-commercial_](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt/blob/main/LICENSE)): generate `4s` vid from a single image at (`25 frames` at `6 fps`). Generation takes ~60s in an A100. [Join the waitlist for Stability's upcoming web experience](https://stability.ai/contact).
   ''')
   with gr.Row():
     with gr.Column():
@@ -100,7 +100,25 @@ with gr.Blocks() as demo:
       
   image.upload(fn=resize_image, inputs=image, outputs=image, queue=False)
   generate_btn.click(fn=sample, inputs=[image, seed, randomize_seed, motion_bucket_id, fps_id], outputs=[video, seed], api_name="video")
-  
+  gr.Examples(
+    examples=[
+        "images/blink_meme.png",
+        "images/confused2_meme.png",
+        "images/confused_meme.png",
+        "images/disaster_meme.png",
+        "images/distracted_meme.png",
+        "images/hide_meme.png",
+        "images/nazare_meme.png",
+        "images/success_meme.png",
+        "images/willy_meme.png",
+        "images/wink_meme.png"
+    ],
+    inputs=image,
+    outputs=[video, seed],
+    fn=sample,
+    cache_examples=True,
+  )
+
 if __name__ == "__main__":
     demo.queue(max_size=20)
     demo.launch(share=True)