Xiong Ma comments

Results 22 comments of


                                            Xiong Ma

Get NAN loss after 35k steps

when I add this model to transformer, I get nan, when I train model, I try many ways to solve this problem, but it still nan, if anyone interested in...

请问有人使用LCSTS数据集跑这段代码吗，求私下交流。[email protected]

I added your wechat by your phone number, could you pass it?

请问有人使用LCSTS数据集跑这段代码吗，求私下交流。[email protected]

I try to add pointer-generator to transformer, Do you want to try it. If you want to try, Please contract me, my wechat is 15025700935

not sure about implementation of copy scatter function

![image](https://user-images.githubusercontent.com/30991932/56712268-79d47980-6760-11e9-8f5f-d7f97e8674d9.png) @jiacheng-xu Hi, I don't understand this code, in this code: 1. indices shape is [batch_size, enc_time, 2] 2. copy shape is [batch_size, enc_time] 3. shape is [batch_size, extend_size] can...

do u have some contrastive datas between chinese Albert with other chinese pretraining models?

but in this website, the xxlarge or base model was trained by bright, which one is yours

I want to read the code, and understand the thought, which part should I read first?

also, I want to debug every part, how to run it and debug it?

returned non-zero exit status 255.

me too @bheinzerling Can you help me?

returned non-zero exit status 255.

this is running time log `2019-05-24 14:12:11,257 [MainThread ] [INFO ] Set ROUGE home directory to /usr/local/RELEASE-1.5.5. 2019-05-24 14:12:11,257 [MainThread ] [INFO ] Writing summaries. 2019-05-24 14:12:11,258 [MainThread ] [INFO...

The Code can be try in multi label ?

this code can be used in two label classification?

Multilingual or Chinese version plan?

hope can support chinese, please