Update app.py

app.py

@@ -41,7 +41,7 @@ def load_model():
         # quantization_config=nf4_config, # add config
         # torch_dtype=torch.bfloat16, # save memory using float16
         # low_cpu_mem_usage=True,
-        # token=st.secrets["HF_TOKEN"],
+        token= TOKEN
     ).to("cuda")
 
     tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)