Update app.py

app.py

@@ -6,6 +6,8 @@ import spaces
 import gradio as gr
 import os
 
+zero = torch.Tensor([0]).cuda()
+
 quantization_config = BitsAndBytesConfig(
     load_in_4bit=True,
     bnb_4bit_compute_dtype=torch.float16
@@ -63,7 +65,7 @@ def process_video(video_file, question_parts):
     ]
     prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)
     # Prepare inputs for the model
-    input = processor([prompt], videos=[video_clip], padding=True, return_tensors="pt").to(model.device)
+    input = processor([prompt], videos=[video_clip], padding=True, return_tensors="pt").to(zero.device)
 
     # Generate output
     generate_kwargs = {"max_new_tokens": 3000, "do_sample": False, "top_p": 0.9}