auth update

ChatService.py

@@ -1,6 +1,7 @@
 import os
 import torch
 import transformers
+from huggingface_hub import login
 
 os.environ["CUDA_VISIBLE_DEVICES"] = "0"
 
@@ -18,13 +19,14 @@ class ChatService:
         gpu_count = torch.cuda.device_count()
         print('gpu_count', gpu_count)
 
-        tokenizer = transformers.AutoTokenizer.from_pretrained(model_name, use_auth_token="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")
+        login("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")
+
+        tokenizer = transformers.AutoTokenizer.from_pretrained(model_name)
         pipeline = transformers.pipeline(
             task="text-generation",
             model=model_name,
             torch_dtype=torch.float16,
             device_map="auto",
-            use_auth_token="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"
         )
 
     @staticmethod

requirements.txt

@@ -10,4 +10,5 @@ accelerate
 flask
 flask_cors
 gunicorn
-urllib3==1.26.6
+urllib3==1.26.6
+huggingface_hub