Zhixiao Ni

icon company Beijing Institute of Technology icon location student of Beijing Institute of Technology

Results 2 comments of Zhixiao Ni

这代码写的真烂，这么多问题，也没人回答

Laughing out loud, that's exactly right. The main challenge lies in the scarcity of documentation, forcing me to speculate and explore through the code myself. In reality, even models like...

Question about training stability

I encountered the same issue where, at the beginning of training, the loss gradually decreased but then reached a point where the training seemed to diverge, and the loss started...