Qian Cao
Qian Cao
https://huggingface.co/uer/gpt2-chinese-lyric 里面有提及
100w的数据，4张V100 batchsize开到256，一个epoch需要一个半小时左右
> 您好，我发现Plato还没有chinese版本。我想要在自己的中文数据集上使用plato模型，请问能否从头开始训练？能的话应该如何训练？直接使用英文预训练的checkpoint肯定不行吧。 > > 谢谢 师兄有 train 好的中文模型可以 share 出来吗哈哈哈哈哈🤣
I also noticed some count errors in this case. Looking forward to the revised release!
> it looks like the problem is in the output of the google scholar for `targetUrl`. for example for "making sense of raw input 2021 " the targeturl is :...
> it's here: > > https://github.com/justinribeiro/zotero-google-scholar-citation-count/blob/bd2e8f83e72f9e1f72158762b5a6abd06c6d3740/src/chrome/content/gscc.js#L703-L705 THX!!!!!
Hope you find [this](https://github.com/AkariAsai/self-rag/issues/15#issuecomment-1895906432) helpful. :)