fix

app.py

@@ -1,9 +1,7 @@
 import gradio as gr
-import spaces
 
 from process import inference
 
-@spaces.GPU
 def clickit(video, prompt):
     return inference(
         video,

process.py

@@ -1,4 +1,5 @@
 import io
+import spaces
 
 import argparse
 import numpy as np
@@ -6,6 +7,7 @@ import torch
 from decord import cpu, VideoReader, bridge
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
+
 MODEL_PATH = "THUDM/cogvlm2-llama3-caption"
 
 DEVICE = 'cuda' if torch.cuda.is_available() else 'cpu'
@@ -60,12 +62,8 @@ model = AutoModelForCausalLM.from_pretrained(
     trust_remote_code=True
 ).eval().to(DEVICE)
 
-
-def predict(prompt, video_data, temperature):
-    strategy = 'chat'
-
-    video = load_video(video_data, strategy=strategy)
-
+@spaces.GPU
+def predict(prompt, video, temperature):
     history = []
     query = prompt
     inputs = model.build_conversation_input_ids(
@@ -98,6 +96,8 @@ def predict(prompt, video_data, temperature):
 
 def inference(video, prompt):
     temperature = 0.1
-    video_data = open(video, 'rb').read()
+    video = open(video, 'rb').read()
+    strategy = 'chat'
+    video_data = load_video(video, strategy=strategy)
     response = predict(prompt, video_data, temperature)
     return response