Update app.py

app.py

@@ -15,12 +15,10 @@ load_in_4bit = True     # Set to True if you want to use 4-bit quantization
 tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
 
 # Load the base model with adapters
-model = AutoAdapterModel.from_pretrained(model_name, low_cpu_mem_usage=True)
+model = AutoAdapterModel.from_pretrained(model_name, low_cpu_mem_usage=True).to("cuda")
 model.load_adapter(lora_adapter)
 
-# Move the model to CPU
-device = torch.device("cpu")
-model.to(device)
+
 
 def respond(message, history, system_message, max_tokens, temperature, top_p):
     # Combine system message and chat history