benywon / en-ch-nmt Goto Github PK

Hi, appreciate your demo.Then if I want to reproduce the training process, how can I get those datas?

If I have another dataset, how can I train for my dataset?
What is your dataset?
thanks for your reply!

size mismatch for output.weight: copying a param with shape torch.Size([1, 512]) from checkpoint, the shape in current model is torch.Size([60004, 512]).

can you share the full model ? thanks

Hi,
I just locally hosted this project and downloaded the model from the link in the readme but the result on the interface that you hosted are different and more accurate than the one available in the link. Can you please provide the latest pretrained model for the project?
Thanks in advance!

感谢，代码部署后直接可用。

不知道 total.uni.35000.model 怎么添加新的词汇，尝试了好久才搞定，但是 train_dist.py 跑不起来，重新整理后（单线程）可以生成新的 th 文件，但翻译出来的东西实在不能看，怀疑可能是语料太少，最近下载了一个500W级的记录重新学习，不知道要学几次才可以达到这个代码的效果？

多谢~
@benywon