Adding baseline models?

[Kyubyong]

This project looks exciting. Why don't you add a few baseline models to compare the BERT model with? For example, you can try to solve the task using a traditional n-gram language model or a neural one like fairseq pretrained wiki103 lm. I'm curious if BERT outperforms those two competitors.

[graykode]

I also think good idea! I will try to adding baseline model that you recommended. Can you recommend more model to use baseline? Thanks

[Kyubyong]

Nothing particular else. Great job!