app.py

import torch
import gradio as gr
from diffusers import StableVideoDiffusionPipeline
from PIL import Image
import numpy as np
from moviepy import ImageSequenceClip
import spaces

# Load the pipeline
pipeline = StableVideoDiffusionPipeline.from_pretrained(
    "stabilityai/stable-video-diffusion-img2vid-xt", torch_dtype=torch.float16, variant="fp16"
)
pipeline.enable_model_cpu_offload()

@spaces.GPU
def generate_video(image, seed):
    # Preprocess the image
    image = Image.open(image)
    image = image.resize((1024, 576))

    # Set the generator seed
    generator = torch.manual_seed(seed)

    # Generate the video frames
    frames = pipeline(image, decode_chunk_size=8, generator=generator).frames[0]

    # Convert frames to a format suitable for video export
    frames = [(frame * 255).astype(np.uint8) for frame in frames]

    # Export the frames to a video file
    clip = ImageSequenceClip(frames, fps=7)
    output_video_path = "generated.mp4"
    clip.write_videofile(output_video_path, codec="libx264")

    return output_video_path

# Create the Gradio interface
iface = gr.Interface(
    fn=generate_video,
    inputs=[
        gr.Image(type="file", label="Upload Image"),
        gr.Number(label="Seed", value=42)
    ],
    outputs=gr.Video(label="Generated Video"),
    title="Stable Video Diffusion",
    description="Generate a video from an uploaded image using Stable Video Diffusion."
)

# Launch the interface
iface.launch()