Skip to content
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

关于Laplacian aleatoric uncertainty loss的一些问题 #25

Open
2anchao opened this issue Jan 6, 2023 · 0 comments
Open

关于Laplacian aleatoric uncertainty loss的一些问题 #25

2anchao opened this issue Jan 6, 2023 · 0 comments

Comments

@2anchao
Copy link

2anchao commented Jan 6, 2023

Laplacian aleatoric uncertainty loss在论文中描述如下:
企业微信截图_1672992075652
代码中的写法:
loss = 1.4142 * torch.exp(-log_variance) * torch.abs(input - target) + log_variance
自己认为对齐的写法:
loss = 1.4142 * log_variance**-1 * torch.abs(input - target) + totch.log(log_variance)

包括在推理中的sigma值对score进行refine,我认为您使用torch.exp(log_variance)代替了原来的log_variance,将不确定性分布指数化,使其值域>0, 感觉这种做法有益,但具体说不上来,请帮忙解答一下。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant