Update app.py

app.py

@@ -34,10 +34,12 @@ import spaces
 
 load_dotenv()
 
+HF_API_KEY = os.getenv("HF_API_KEY")
 MODEL_NAME = "hatakeyama-llm-team/Tanuki-8B-Instruct"
 PREFERENCE_API_URL = os.getenv("PREFERENCE_API_URL")
 assert PREFERENCE_API_URL, "PREFERENCE_API_URL is not set"
 
+
 quantization_config = BitsAndBytesConfig(
     load_in_4bit=True,
     bnb_4bit_compute_dtype=torch.bfloat16,
@@ -45,9 +47,9 @@ quantization_config = BitsAndBytesConfig(
     bnb_4bit_use_double_quant=True,
 )
 model = AutoModelForCausalLM.from_pretrained(
-    MODEL_NAME, quantization_config=quantization_config, device_map="auto"
+    MODEL_NAME, quantization_config=quantization_config, device_map="auto", token=HF_API_KEY
 )
-tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, token=HF_API_KEY)
 
 print("Compiling model...")
 model = torch.compile(model)