siyang liu comments

Results 16 comments of


                                            siyang liu

> > 已解决发自我的iPhone > > […](#) > > ------------------ 原始邮件 ------------------ 发件人: lsy641 [[email protected]](mailto:[email protected]) 发送时间: 2019年7月10日 19:27 收件人: macanv/BERT-BiLSTM-CRF-NER [[email protected]](mailto:[email protected]) 抄送: charlesfufu [[email protected]](mailto:[email protected]), Author [[email protected]](mailto:[email protected]) 主题: 回复：[macanv/BERT-BiLSTM-CRF-NER] 如何变训练边输出损失loss? (#175)...

How can I participate in this project?

Ok, I will think adding Albert model in it

Turbo slower than Torch on V100

> ![image](https://user-images.githubusercontent.com/4970790/87944053-79e13780-cad1-11ea-9fc4-70abff761f00.png) > > According to the unit-tests report, the performance of Albert seems abnormal, too. It seems turbo albert's performance is randomly higher or lower than torch?

ONNXRT can not be applied in Albert

How to fix it? implement a sum by hand or update onnx?

How to set random seed number to get consistent output on Vicuna-13b-v1.3?

I have the same issue

Allowing adding new token as unk token for gpt2 tokenizer

What do you need me to do to help this merge?

A recent EMNLP work to share about task-adaptive tokenization with variable segmentation

> I read the paper, is there any code available which showcases the algorithm? @RubyBit Hello. I am currently organizing the codes but I am able to add you to...

BlenderEmotionalSupport.py代码有误

Thank you for reminding us