Skip to content

Fix newton negligible derivative - #240

Closed
mubasharameen485-cloud wants to merge 4 commits into
kmolan:mainfrom
mubasharameen485-cloud:fix-newton-negligible-derivative
Closed

Fix newton negligible derivative#240
mubasharameen485-cloud wants to merge 4 commits into
kmolan:mainfrom
mubasharameen485-cloud:fix-newton-negligible-derivative

Conversation

@mubasharameen485-cloud

Copy link
Copy Markdown
Contributor

Closes #206

Add SvdSettings with a configurable max_sweeps field, defaulting
to 60 to preserve existing behaviour. Matrix::svd() now delegates
to Matrix::svd_with_settings() with the default settings.

Fixes kmolan#156
Compare |dfx| against EPSILON * |fx| instead of an exact-zero check,
so a derivative like 1e-300 is correctly rejected as singular instead
of producing a meaningless huge step.

Closes kmolan#206
@kmolan

kmolan commented Jul 29, 2026

Copy link
Copy Markdown
Owner

@mubasharameen485-cloud I am rejecting this PR. I will remind you again: You must close your previous ticket first, only then begin on the next one: #235

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.

Reject a negligible derivative in Newton, not just an exactly zero one

2 participants