Swap DruidJS for embedding-atlas umap implementation#339
Open
Swap DruidJS for embedding-atlas umap implementation#339
DruidJS for embedding-atlas umap implementation#339Conversation
coszio
reviewed
Jan 6, 2026
| } | ||
|
|
||
| const umap = await createUMAP(count, inputDim, outputDim, flatData, { | ||
| metric: 'cosine', // do we need to make it configurable? |
Contributor
There was a problem hiding this comment.
In theory we can get the precomputed matrix directly from qdrant, but we are not doing this right now.
I see embedding-atlas does not accept a distance matrix, but the underlying umappp does.
It would be nicer, and maybe more efficient if we could leverage qdrant for this. Although seems out of scope for this PR.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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’ve had my eye on this embedding visualization paper/library for some time now. They’ve got a really nice UMAP implementation written in C++ compiled to wasm and baked into the library. It’s quite good.
The “visualize” section of the Qdrant WebUI has always left a lot to be desired for me personally. Just a bit slow and the UMAP visualizations seemed sub-optimal. I can see that we are using a quite-outdated javascript implementation of UMAP now. I went in locally and swapped it for this C++ implementation. The UMAP looks a bit more canonical in my opinion and its much faster (~35 seconds —> ~8 seconds; ~75% reduction).
The UMAPs are not one-for-one (see screenshot). This obviously only affects UMAP -- tSNE remains intact. It adds one dependency (
embedding-atlas). The library has a lot of other interesting/useful components I think could be worth exploring integration into the web ui for people to explore their latent space.