Update README.md

README.md

@@ -10,7 +10,7 @@ Remember to replace the path to test data and word2vec pretrained model.
 Find access token in the token.txt  
   
 If you can't import model from huggingface directly, you can try the train+test notebook.  
-First, download the model directory and the config.json and safetensor from this repo and put it under *./exp/fox_nbc/checkpoints/best/*.
+First, download the whole directory and put the config.json and safetensor under *./exp/fox_nbc/checkpoints/best/*.
 Then, run every cell except the trainning part. Make sure the evaluation cell can load the checkpoint from the directory.  
 An example of project directory looks like this:  
 https://cdn-uploads.huggingface.co/production/uploads/673ab29d3c897b60a54ea67f/GbN3U6bSUhya5FeQn-TjO.png