Update app.py

app.py

@@ -31,7 +31,7 @@ def load_model(model_name, eight_bit=0, device_map="auto"):
     gpu_count = torch.cuda.device_count()
     print('gpu_count', gpu_count)
 
-    tokenizer = transformers.LlamaTokenizer.from_pretrained(model_name, use_auth_token=hf_token)
+    tokenizer = transformers.LlamaTokenizer.from_pretrained(model_name, token=hf_token)
     model = transformers.LlamaForCausalLM.from_pretrained(
         model_name,
         #device_map=device_map,
@@ -42,7 +42,7 @@ def load_model(model_name, eight_bit=0, device_map="auto"):
         low_cpu_mem_usage=True,
         load_in_8bit=False,
         cache_dir="cache",
-        use_auth_token=hf_token 
+        token=hf_token 
     ).cuda()
 
     generator = model.generate