search

Ucas-HaoranWei / Vary-toy

Official code implementation of Vary-toy (Small Language Model Meets with Reinforced Vision Vocabulary)
565 stars 41 forks source link

训练loss降为0 #32

Open afreestudy opened 1 month ago

afreestudy commented 1 month ago

训练刚开始loss在2左右，然后急速下降，大概在一轮的10%进度时，loss就到了0.001左右了，再经过几轮就变成了0，为什么会出现这种情况呢？

SeeeeShiwei commented 1 week ago

也遇到了这种情况，loss训练一段时间后，变为0