Update app.py

app.py

@@ -27,9 +27,11 @@ tokenizer.padding_side = "right"
 
 streamer = TextStreamer(tokenizer)
 
-model = AutoModelForCausalLM.from_pretrained(model, trust_remote_code=True, token = access_token, device_map="auto", load_in_8bit=True, offload_folder="offload/")
+model = AutoModelForCausalLM.from_pretrained(model, trust_remote_code=True, token = access_token, device_map="cuda:0", load_in_8bit=True, offload_folder="offload/")
 model = PeftModel.from_pretrained(model, peft_model, offload_folder="offload/")
 
+#device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
+device = 'cuda:0'
 model = model.eval()
 
 # Inference Data