Update utils.py

utils.py

@@ -59,13 +59,14 @@ def load_tokenizer_and_model(base_model, load_8bit=False):
     else:
         device = "cpu"
 
-    tokenizer = AutoTokenizer.from_pretrained(base_model, use_fast = True)
+    tokenizer = AutoTokenizer.from_pretrained(base_model, use_fast = True, use_auth_token=True)
     if device == "cuda":
         model = AutoModelForCausalLM.from_pretrained(
             base_model,
             load_in_8bit=load_8bit,
             torch_dtype=torch.float16,
-            device_map="auto"
+            device_map="auto",
+            use_auth_token=True
         )
     else:
         model = AutoModelForCausalLM.from_pretrained(
@@ -90,11 +91,12 @@ def load_model(base_model, load_8bit=False):
             base_model,
             load_in_8bit=load_8bit,
             torch_dtype=torch.float16,
-            device_map="auto"
+            device_map="auto",
+            use_auth_token=True
         )
     else:
         model = AutoModelForCausalLM.from_pretrained(
-            base_model, device_map={"": device}, low_cpu_mem_usage=True
+            base_model, device_map={"": device}, low_cpu_mem_usage=True, use_auth_token=True
         )
 
     #if not load_8bit: