Update app.py

app.py

@@ -1,22 +1,28 @@
 import streamlit as st
 import torch
-from diffusers import StableVideoDiffusionPipeline
-from diffusers.utils import export_to_video
+from diffusers import DiffusionPipeline
+import tempfile
 
-# Load the video generation pipeline
+# Load the text-to-video model
 st.write("Loading model... (first run may take a few minutes)")
 
-model_id = "stabilityai/stable-video-diffusion-img2vid"
-pipe = StableVideoDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
+model_id = "damo-vilab/text-to-video-ms-1.7b"
+pipe = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
+pipe.to("cpu")  # Stay on CPU since we don’t have a GPU
 
 st.title("Text-to-Video Generator")
 prompt = st.text_input("Enter a text prompt for the video:")
-frames = st.slider("Number of frames (video length)", min_value=8, max_value=81, value=24)
+
+frames = st.slider("Number of frames (video length)", min_value=8, max_value=24, value=16)
 
 if st.button("Generate Video") and prompt:
     with st.spinner("Generating video... this may take a while on CPU"):
         result = pipe(prompt=prompt, num_frames=frames, num_inference_steps=20)
         video_frames = result.frames  # List of PIL images
-        export_to_video(video_frames, "output.mp4", fps=8)
 
-    st.video("output.mp4")
+        # Save frames as video file
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as temp_file:
+            video_path = temp_file.name
+        result.export_to_video(video_path, fps=8)
+
+    st.video(video_path)