-
Notifications
You must be signed in to change notification settings - Fork 11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
question about the training loss? #38
Comments
按照llava15_train.sh中的原始配置(未改batchsize设置),8卡A100上训练,loss和一些指标的log如上图。想要请教一下这个loss看着在震荡,忽高忽低。以及目前看来'rewards_train/chosen'经常是比'rewards_train/rejected'要低,想问下这个是正常还是不正常呀? 如果将batchsize设为8,即单卡batchsize为8,保持max_steps等其余参数不变,8 A100 GPUs训练,在最新83.1k数据集上差不多训练2 epochs,修改batchsize后最终的loss还是下降的,rewards_train/accuracies也能到0.80左右。 想问下,论文中写的batchsize为8,是总batchsize为8还是单卡batchsize为8呢? |
感谢关注我们的工作!
|
@yiranyyu 您好,请教一个问题。 请问这个编码索引的获取方式是如何确定的?
刚入门,期待你的回复~ 感谢!! |
非常赞且有意义的工作!!
按照llava15_train.sh中的原始配置(未改batchsize设置),8卡A100上训练,loss和一些指标的log如上图。想要请教一下这个loss看着在震荡,忽高忽低。以及目前看来'rewards_train/chosen'经常是比'rewards_train/rejected'要低,想问下这个是正常还是不正常呀?
期待您的回复@yiranyyu @Haoye17
The text was updated successfully, but these errors were encountered: