add location

model.py

@@ -430,8 +430,7 @@ class MaskedAutoencoderViT(nn.Module):
         # captioner specifics
         # unimodal layer is for text tokens.
         # multimodal layer is for text to query from image.        
-        self.tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased", 
-                                                cache_dir='/disk/scratch_fast/bingchen/.cache/torch/hub/checkpoints/bert-base-uncased', local_files_only=True)
+        self.tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased", )
         
         # token embeddings
         # NOTE: +1 for mask token used by MLM objective