Update app.py

app.py

@@ -5,9 +5,8 @@ from awq import AutoAWQForCausalLM
 
 model_path = "bragour/Camel-7b-chat-awq"
 
-device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 
-model = AutoAWQForCausalLM.from_quantized(model_path, fuse_layers=True, trust_remote_code=False, safetensors=True).to(device)
+model = AutoAWQForCausalLM.from_quantized(model_path, fuse_layers=True, trust_remote_code=False, safetensors=True)
 tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=False)
 
 
@@ -16,7 +15,7 @@ def respond(
 ):
     formatted_prompt = f"<s>[INST]{message}[/INST]"
 
-    tokens = tokenizer(formatted_prompt, return_tensors='pt').input_ids.to(device)
+    tokens = tokenizer(formatted_prompt, return_tensors='pt').input_ids.cuda()
 
     # Generate the response from the API
     result = model.generate(