update

app.py

@@ -15,7 +15,7 @@ model_id = "microsoft/Phi-3-mini-128k-instruct"
 device = "auto" # torch.device("cuda" if torch.cuda.is_available() else "cpu")
 dtype = torch.bfloat16
 
-tokenizer = AutoTokenizer.from_pretrained(model_id, token=huggingface_token)
+tokenizer = AutoTokenizer.from_pretrained(model_id)#, token=huggingface_token)
 
 
 import time
@@ -72,7 +72,7 @@ def generate_text(prompt, system_message="You are a helpful assistant."):
     global histories
    
     model = AutoModelForCausalLM.from_pretrained(
-        model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
+        model_id ,torch_dtype=dtype,device_map=device # token=huggingface_token
     )
     #print(system_message)
     text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device) #pipeline has not to(device)