Update README.md

README.md

@@ -22,9 +22,19 @@ This model uses ELECTRA Small model settings, 12 layers, 128 dimensions of hidde
 
 Vocabulary size was set to 32,000 tokens.
 
+## Training data and libraries
+
+YACIS-ELECTRA is trained on the whole of [YACIS](https://github.com/ptaszynski/yacis-corpus) blog corpus, which is a Japanese blog corpus containing 5.6 billion words in 354 million sentences.
+
+The corpus was originally split into sentences using custom rules, and each sentence was tokenized using [MeCab](https://taku910.github.io/mecab/). Subword tokenization for pretraining was done with WordPiece. 
+
+We used original [ELECTRA](https://github.com/google-research/electra) repository for pretraining. The pretrainig process took 7 days and 6 hours under the following environment: CPU: Intel Core i9-7920X, RAM: 132 GB, GPU: GeForce GTX 1080 Ti x1. 
+
+
 ## Licenses
 
 The pretrained model with all attached files is distributed under the terms of the [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/deed.en) license. 
+
 <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-sa/4.0/88x31.png" /></a><br />This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>.
 
 ## Citations