Skip to content
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

fix num of partitions and faiss index file name pattern matching #67

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

afurkank
Copy link

@afurkank afurkank commented Feb 5, 2024

fix issue of number of partitions for the faiss indexing always being 100 and allow using faiss index files with names matching the pattern "ivfpq*.faiss"

Note that when indexing, FAISS may give a warning like this:
WARNING clustering XXX points to YYY centroids

This happens when the sample being indexed has number of data points smaller or bigger than some number.
Apparently, the number of data points(n) must be in range of some multiples of the number of centroids(k). This warning can be silenced but I suppose the ColBERT repo might need an update for this.
More information can be found here.

… 100 and allow using faiss index files with names matching the pattern "ivfpq*.faiss"
@cmacdonald
Copy link
Collaborator

cmacdonald commented Feb 5, 2024

lgtm @afurkank. I might not have a chance to test two different faiss files this week but hopefully next week

@afurkank afurkank closed this by deleting the head repository Mar 6, 2024
@cmacdonald
Copy link
Collaborator

Hi @afurkank - I'm still planning at looking at this but ECIR 2024 currently a higher priority :-/

@afurkank
Copy link
Author

afurkank commented Mar 7, 2024

Hi, sorry for closing the pull request. I did not realize that deleting a fork would close this pr. I sent a ticket for restoring the repo so hopefully I can reopen the pull request.

I'm still planning at looking at this but ECIR 2024 currently a higher priority :-/

No problem at all!

@afurkank afurkank reopened this Mar 7, 2024
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