Update app.py

app.py

@@ -22,7 +22,7 @@ api = REST(ALPACA_API_KEY, ALPACA_SECRET_KEY, ALPACA_BASE_URL)
 
 # Load Models
 dolly_tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
-dolly_model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, torch_dtype.float16) # half precision
+dolly_model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, torch_dtype=torch.float16) # half precision
 dolly_model.gradient_checkpointing_enable() # memory-efficient loading
 print("Loading FinBERT model...")
 finbert_tokenizer = AutoTokenizer.from_pretrained(FINBERT_MODEL_NAME)