CheungZee comments

Results 15 comments of


                                            CheungZee

pretrain LayoutLM v2 on new dataset

can deepFM use sparse data format?

@Leavingseason hello, 请教一个格式上的问题，fieldID:featureID:value 这里，如果fieldID==1 对应的featureID 有3个，如果fieldID==2对应的featureID 有2个，fieldID==2的 featureID 的值的编码需要基于 fieldID==1 的featureID 上吗？ for example: 0 1:1:1 1:2:1 1:3:1 2:**_1_**:1 #这里fieldID==2的 featureID 可以重新编码 0 1:1:1 1:2:1 1:3:1 2:**_4_**:1 #这里fieldID==2的 featureID 不可以重新编码,需要基于原来的1之上，谢谢！

Can LayoutLM be used for language generation ?

@pzdkn any update ?

Add UDOP

Eagerly anticipating this PR being merged. Hope there will be a pretrain demo too.

Add UDOP

求更新网盘链接

LayoutLM(v1) pretrain source code

> Hi, did you come across the code? No, So I wrote it...

y_pred[i] index start from 1?

@fword I found it, this may be sth. wrong ? @neal668

Errors with ZERO2 in the encoder-decoder model

Run into the same issue for a loss = loss + custom_regular_terms(some param for model)