haoliangjiang

It is a lookup table showing the splits of training and val.

The binary mask will help in that we are confident that the attention mask allows the entire value to join the computation of the next block. However, I believe one...

> How the heck to you reload global_step? > Everytime I load a model it seems this op's values goes back to 0 I guess you need to get it...

Any comments on this one?