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

Fix broken backward graph when using LieTensor.exp() #631

Merged
merged 3 commits into from
Dec 22, 2023

Conversation

luisenp
Copy link
Contributor

@luisenp luisenp commented Dec 8, 2023

Closes #626

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Dec 8, 2023
@luisenp luisenp requested a review from fantaosha December 8, 2023 16:29
@luisenp luisenp mentioned this pull request Dec 8, 2023
Copy link
Contributor

@fantaosha fantaosha left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. However, I did not follow why the backward graph was broken.

@luisenp
Copy link
Contributor Author

luisenp commented Dec 12, 2023

@fantaosha The reason is that the standard LieTensor constructor is not a differentiable operation. If you want to make a new LieTensor in a differentiable way (e.g., as the result of an exp map), you need a tensor creation function that has a backward implementation. This is already available in FromTensor class, I just forgot to use it inside the exp map function.

@luisenp luisenp merged commit 52d02cb into main Dec 22, 2023
14 of 15 checks passed
@luisenp luisenp deleted the lep.lietensor_exp_backward_fix branch December 22, 2023 01:57
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Optimization on SE(3)
3 participants