Skip to content

Conversation

@pavelkomarov
Copy link
Collaborator

@pavelkomarov pavelkomarov commented Nov 19, 2025

I so far just left it there, even wrote tests for some of it, but it's not used and doesn't seem like it's going to be used, so I think it's worth getting rid of to simplify things.

return x_hat/weight_sum, dxdt_hat/weight_sum


def peakdet(x, delta, t=None):
Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Moved this guy to the bottom, because it's only used by the Triangles simulations.

… seeks to make first term sparse, because that's advantageous to the loss function, which artificially pushes huberM low. RMSEs tend to be better when huberM isn't allowed to shrink to 0. Also realized I was multiplying M by sigma in TVR, but the inputs are already scaled, so M should just be M, doesn't need to be scaled there.
@pavelkomarov pavelkomarov merged commit 00ccd92 into master Nov 22, 2025
1 check passed
@pavelkomarov pavelkomarov deleted the remove-unused-code branch November 22, 2025 04:19
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