Lam Chi
Lam Chi
@mengye-ren
@xjtushujun
@brightmart How do you get **Chinese** pretrained model of XLNet ??
I hope so!
@juntaoy 你好，我有原版的英文ontonotes5的LDC数据集，请问能发一份处好的ontonotes5.0的ner数据集吗（conll格式）？想跑下您论文的结果 我的邮箱: [email protected]
@BenjaminWinter I'm also confused about segment level recurrence in the paper?
Excuse me! Have you trained transformer_xl model on Chinese corpus?
多谢回复！ 请问一般overlapping的cnn stem更好，还是(你们的)non-overlapping的更好呢？
lld loss训练过程为负数，请问怎么破？
遇到同样问题，lld loss为负数，请问如何有效解决？