-
Notifications
You must be signed in to change notification settings - Fork 56
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
Why is loss rising #11
Comments
请问一下,我想在自己的数据上训练,是不是应该先训练基础的Sphere64,再固定之前的训练不确定模块?Sphere64网络的训练有代码可以参考吗? |
从数学公式上来看,negMLS不能保证非负。老铁,加个微信讨论下细节哈,我wechat:On_theway24,备注一下是PFE模型讨论哈。 |
從 PAPER 的 Equation(3) 來看,會導致出現負號的原因是因為加了 log 嗎 |
只要可以保證 log(σ2(l)i+σ2(l)j) 為正的話,基本上就是 negMLS 就是恆正了 |
@twmht , 事实上,不能保证 log(σ2(l)i+σ2(l)j) > 0! |
我覺得應該 log 裡面要加個 exp? |
坦白来说,我认为论文中的公式(3)是错误的,我自己没有推导出来,然后发邮件咨询作者,作者也不回复邮件。所以,我觉得这篇文章就是垃圾,推荐你看CVPR-2020的一篇paper :Data Uncertainty Learning in Face Recognition. |
那篇我知道呀! 但目前沒有 source code 不好複製 |
PFE 的 similarity 沒有上限,但作者似乎也沒給一個實驗建議值 |
推不出来公式3+1 |
I trained on my own dataset, the reg loss is rising , and loss is a negative number. Why does this happen?
The text was updated successfully, but these errors were encountered: