Xuefeng Bai

Hi, Thanks for your awesome job. Will you share the part of the code for Reddit data preprocessing? Looking forward to your reply.

关于分词模块的数据平滑

4

CharacterBasedGenerativeModel统计时默认用的是NormalProb做的数据平滑，我在训练时试图换成frequency.py中的加一平滑或古德图灵平滑，结果分词结果却出奇的差，请问是怎么回事呢？

Error when run with multiple GPUs

1

Hi, I often ran into the following error when starting a multi-GPU training. ``` Traceback (most recent call last): File "train.py", line 116, in main(opt) File "train.py", line 44, in...

Inconsistent examples

2

Hi, I am trying to reproduce the results of the baseline NMT system in your paper, however, I found the examples given in [Readme](https://github.com/Amazing-J/structural-transformer#baseline-input) and that in [corpus_example](https://github.com/Amazing-J/structural-transformer/blob/master/corpus_sample/baseline_corpus/train_source_bpe) are not...

Hi, Thanks for your great work, are there any plans to support models like GPT-NEO and GPT-NEOX?