Skip to content

Tune ScaNN for other angular datasets #172

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

Merged
merged 1 commit into from
Jul 16, 2020

Conversation

sammymax
Copy link
Contributor

The originally submitted configuration was only tuned for Glove-100. Here are some better configurations for the other angular datasets. Still investigating NYTimes...

Glove-25:
glove-25
LastFM:
lastfm

@erikbern
Copy link
Owner

Nice!

FYI nytimes-256 has a few "missing" vectors (all elements set to zero) which I guess is a bug or a feature depending on how you look at it (I've been arguing that's a common case that libraries should ideally be able to handle). So that might cause issues for ScANN

@erikbern
Copy link
Owner

Let me know if you want me to merge this. Otherwise will keep it open so you can optimize more :)

@sammymax
Copy link
Contributor Author

sammymax commented Jul 16, 2020 via email

@erikbern erikbern merged commit 55b9950 into erikbern:master Jul 16, 2020
@sammymax sammymax mentioned this pull request Jul 23, 2020
erikbern added a commit that referenced this pull request Apr 14, 2023
Tune ScaNN for other angular datasets
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