Nice work!
The version of implementation can reach 22+ BLUE score. However, my implementation have only 0.16+ BLUE score on test dataset. Comparing with your work, I found changing the concatenation torch.cat((Y_t, o_pre), dim=1) to torch.cat((o_pre, Y_t), dim=1) can only reach 0.16+ BLUE score.

B.W.T. chaging the concatenating order between dec_hidden and a_t in step function also

🌠项目主页，翻到最下面的Unsubscribe取消提醒

• 由于项目是纯前端代码，所以数据存储设计在了GitHub Issue中，这里就是数据库
• 对项目有其他问题请发起新的Issue，不要在此Issue中评论或者@他人
• 存储数据后才可在其他软件中（如微信）分享链接，否则其他人无法找到你的数据
• 存储数据后请在当前Issue中点击最下方的Unsubscribe按钮，以防邮件提醒
• Issue中可以随时删除评论数据
• 代码开源，网站开放，数据库开放，只统计公开仓库，不会存在数据安全问题
• 由于实时分析的瓶颈在于网络请求，默认每个repo只分析master分支

• We see you plan to document functions using sphinx, it would be good to have some basic doc files already

• Notebooks are barely commented, could you please expand on the reasons behind the different steps? Ideally, each notebook should be self explanatory (like a stand-alone tutorial).

Great job on the project!

@seblehner Please reach out and we can help you put together a Binder link so you can share the notebooks with others. (Even without data, it would be useful.)

cc/ @cvitolo