Skip to content

Conversation

@EshitaJoshi
Copy link
Collaborator

The optimization has been moved to derive_mirror_rnda, and so mirror_reflection_random_angle is being removed as a parameter in this optimization.

),
type=float,
default=0.01,
default=0.0001,
Copy link
Contributor

Choose a reason for hiding this comment

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

This is an unusual small learning rate, or? Just a comment, no need to change.

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

It is small, but after removing rnda from this function the larger learning rate was too large. The learning rate is adaptive anyway so it's not going to stay the same, this is just the starting rate for the first step.

@ctao-sonarqube
Copy link

@EshitaJoshi EshitaJoshi merged commit c1c2da2 into main Feb 10, 2026
16 checks passed
@EshitaJoshi EshitaJoshi deleted the remove-mrra branch February 10, 2026 10:56
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

Successfully merging this pull request may close these issues.

2 participants