Create app.py

app.py

@@ -0,0 +1,47 @@
+import gradio as gr
+import torch
+from diffusers import DiffusionPipeline
+
+# Define a dictionary of available models
+models = {
+    "AnimateDiff-Lightning": "ByteDance/AnimateDiff-Lightning",
+    "Other Model 1": "model_name_or_id_1",
+    "Other Model 2": "model_name_or_id_2",
+    # Add more models as needed
+}
+
+# Load models into a dict for quick access
+loaded_pipelines = {}
+
+def load_model(model_id):
+    if model_id not in loaded_pipelines:
+        pipe = DiffusionPipeline.from_pretrained(model_id)
+        pipe = pipe.to("cuda" if torch.cuda.is_available() else "cpu")
+        loaded_pipelines[model_id] = pipe
+    return loaded_pipelines[model_id]
+
+def generate_video(model_name, prompt):
+    pipe = load_model(models[model_name])
+    # Call your model's video generation method
+    result = pipe(prompt)
+    # Adjust based on actual output (assuming it returns a video)
+    video = result.videos[0]  # or result['videos'][0]
+    video_path = "output.mp4"
+    video.save(video_path)
+    return video_path
+
+with gr.Blocks() as demo:
+    gr.Markdown("# Video Generation from Text")
+    with gr.Row():
+        model_choice = gr.Dropdown(choices=list(models.keys()), label="Select Model")
+        prompt_input = gr.Textbox(label="Enter your prompt", lines=2, placeholder="A spaceship in space, neon colors")
+    generate_button = gr.Button("Generate Video")
+    video_output = gr.Video(label="Generated Video")
+    
+    generate_button.click(
+        fn=generate_video,
+        inputs=[model_choice, prompt_input],
+        outputs=video_output
+    )
+
+demo.launch()