TurboQuant native distance path and HNSW persistence support (WIP quality tuning)#36413
Open
SahilKumar75 wants to merge 1 commit into
Open
TurboQuant native distance path and HNSW persistence support (WIP quality tuning)#36413SahilKumar75 wants to merge 1 commit into
SahilKumar75 wants to merge 1 commit into
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
I hereby state that this submission is released according to the license in the root of this project's source tree, and that I have the proper authorization to make this contribution on behalf of the owner of its copyright.
Tracks #36327
Summary of changes introduced
A dedicated TurboQuant distance path (no DotProduct fallback).
TurboQuant-aware HNSW saving/loading of internal state hooks.
Improvements in schema/tooling as well as additional TurboQuant-specific tests.
Validation
All distance functions: validation passed.
HNSW Index: all validation passed.
Stress test for HNSW index (using SIFT dataset): passed.
Latency / memory checks performed.
Limitations of the current implementation
Recall metric is yet to meet the desired level (recall@10 ~ 0.26 on SIFT subset).
Therefore, this PR can be considered integration/stability ready, however requires some more quality tuning before closing #36327.