add dependencies

app.py

@@ -1,6 +1,10 @@
 import gradio as gr
-# import subprocess  # 🥲
-
+import subprocess  # 🥲
+subprocess.run(
+    "pip install flash-attn --no-build-isolation",
+    env={"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
+    shell=True,
+)
 # subprocess.run(
 #     "pip install git+https://github.com/LLaVA-VL/LLaVA-NeXT.git",
 #     shell=True,
@@ -15,6 +19,7 @@ import copy
 import warnings
 from decord import VideoReader, cpu
 import numpy as np
+import spaces
 
 warnings.filterwarnings("ignore")
 
@@ -47,6 +52,7 @@ tokenizer, model, image_processor, max_length = load_pretrained_model(pretrained
 model.eval()
 print("Model loaded successfully!")
 
+@spaces.GPU
 def process_video(video_path, question):
     max_frames_num = 64
     video, frame_time, video_time = load_video(video_path, max_frames_num, 1, force_sample=True)
@@ -78,7 +84,6 @@ def process_video(video_path, question):
     response = tokenizer.batch_decode(output, skip_special_tokens=True)[0].strip()
     return response
 
-# Gradio interface
 def gradio_interface(video_file, question):
     if video_file is None:
         return "Please upload a video file."
@@ -103,6 +108,5 @@ with gr.Blocks() as demo:
         outputs=output
     )
 
-# Launch the app
 if __name__ == "__main__":
     demo.launch()

requirements.txt

@@ -6,4 +6,6 @@ torchvision
 decord
 einops
 accelerate
+open_clip_torch
+av
 git+https://github.com/LLaVA-VL/LLaVA-NeXT.git