Update app.py

app.py

@@ -22,16 +22,16 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "2048"))
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
 # thanks to: https://medium.com/@ashkanpakzad/hugging-face-public-space-private-model-w-streamlit-app-0166b4f66869
-api_key = os.getenv("HF_TOKEN")
-from huggingface_hub import hf_hub_download
-repo_id = "giustinod/TestLogica-AZService"
-model_path = hf_hub_download(repo_id=repo_id, filename="model.pkl", token=api_key)
+access_token = os.getenv("HF_TOKEN")
+print('access_token: ' + access_token)
+model_id = "giustinod/TestLogica-AZService"
 
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForCausalLM.from_pretrained(
-    model_path,
+    model_id,
     device_map="auto",
     torch_dtype=torch.bfloat16,
+    token=access_token
 )
 model.eval()