Chi Zhang

Postgraduate student loves programming

Results 5 comments of


                                            Chi Zhang

Teacher 和Student attention head 不一致时的考虑

同问这个问题，head不一致无法计算MSE loss，而且这种情况应该很多，毕竟大模型的head一般会比小模型多是吧，希望官方能给一个解答，谢谢～

[BUG] AgeRestrictedError: ZfpgKWTQqXw is age restricted, and can't be accessed without logging in.

> Thanks, so helpful

when I execute the command "python train.py --actor-model facebook/opt-1.3b --reward-model facebook/opt-350m --deployment-type single_gpu" I got the follow result, and I try many solutions but all failure.

I can run this script successfully on CUDA 11.3 Pytorch 1.12.0

[BUG] 智障，这么多问题，也好意思放出来，浪费时间

你真恶心

query understand模块在源码中实现了嘛？

我也好奇这个东西是不是没实现