Update README.md

README.md

@@ -22,6 +22,12 @@ This repository provides a base-sized Japanese RoBERTa model. The model is provi
 
 # How to use the model
 
+Since this is a private repo, first login your huggingface account from the command line:
+
+~~~
+transformer-cli login
+~~~
+
 *NOTE:* Use `T5Tokenizer` to initiate the tokenizer.
 
 ~~~~
@@ -37,7 +43,7 @@ model = RobertaModel.from_pretrained("rinna/japanese-roberta-base", use_auth_tok
 A 12-layer, 768-hidden-size transformer-based masked language model.
 
 # Training
-The model was trained on [Japanese CC-100](http://data.statmt.org/cc-100/ja.txt.xz) and [Japanese Wikipedia](https://dumps.wikimedia.org/jawiki/) to optimize a masked language modelling objective on 8\\\\\\\\\\\\\\\\*V100 GPUs for around 15 days.
+The model was trained on [Japanese CC-100](http://data.statmt.org/cc-100/ja.txt.xz) and [Japanese Wikipedia](https://dumps.wikimedia.org/jawiki/) to optimize a masked language modelling objective on 8*V100 GPUs for around 15 days. It reaches ~3.9 perplexity on a dev set sampled from CC-100.
 
 # Tokenization
 The model uses a [sentencepiece](https://github.com/google/sentencepiece)-based tokenizer, the vocabulary was trained on the Japanese Wikipedia using the official sentencepiece training script.