Update app.py

app.py

@@ -16,8 +16,9 @@ subprocess.run(
     env={"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
     shell=True,
 )
+hf_token = os.getenv("HF_TOKEN")
 
-token = os.environ["HF_TOKEN"]
+token = hf_token
 
 
 model = AutoModelForCausalLM.from_pretrained(