[WIP] Add support for labels when visualizing embeddings #112
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Outdated suggestions cannot be applied.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.
UMAP fitting supports labels for supervised dimensionality reduction allowing it to accept an (incomplete) set of labels. It allows you to guide how the dimensionality reduction is performed to structure the embedding space based on certain tagged attribute.
Example workflow:
Take the coco dataset
tag 5 images of people and 5 with no people
compute embeddings to generate this plot which distributed all samples along the axis of "personness"
select samples of the tails of this distribution and tag them as "person" and "not person"
Recompute embeddings to get an even more structured embeddings visualization
Iteratively add good examples of classes, recomputing the visualization each time to generate more accurate clusters
This was all done just using tags in the App: