Closed pemywei closed 5 years ago
Sorry, I do not understand what you mean.
Whether it it in our paper or on readme of github, we say that we first pretrain two language models, and then use them to train nmt models from scratch.
Do you mean to use a pretrained nmt model for warmup? We has provide this option --load-nmt
. Because of lack of time, we did not get those results.
As you claimed, you train your lm-nmt from scratch, but why do you not use a pretrained nmt model for warmup? Can you give some experimental results about the latter strategy?