-
Notifications
You must be signed in to change notification settings - Fork 197
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
Output non-normalized distances in IVF-PQ and brute-force KNN #892
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the fix! I think we can make it a bit less complicated, see the couple comments
There are quite a few IVF-PQ cases where this PR gives worse results than before, I'll look into those. |
Note: the worse test results that I observed in IVF-PQ are due to how Also, I am removing the "breaking" flag, since IVF-PQ was added in the same release. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks, the changes seem nice and non-intrusive to me. Just couple of tiny suggestions.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks Louis for the PR, it looks good to me!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks!
@gpucibot merge |
Solves point 9 of #711 (the observed errors were due to comparing distances at different scales).
This PR does two things: