Update app.py

app.py

@@ -33,7 +33,7 @@ def load_model():
     base_model = AutoModelForCausalLM.from_pretrained(
         BASE_MODEL,
         quantization_config = bnb_config,
-        torch_dtype=torch.float16, # Use half precision for faster GPU inference
+        dtype=torch.float16, # Use half precision for faster GPU inference
         device_map="auto",
         trust_remote_code=True,
         offload_folder="./offload_base" # === Specify a folder for disk offloading since we are using free CPU not GPU===