lmcinnes / lmcinnes/umap

UMAP trained model consumes too much space

Open
#916 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
8.3k
Forks
871
Avg merge
1d 13h
Merged PRs (30d)
5

Description

Is there a way to reduce the space that a trained model takes? My trained model takes around 8 GB (~800k examples). I noticed that each model contains the embeddings of the train set, so its size is going to be proportional to the train set size. I'm not sure what else the model stores that explodes the size like that. Is it possible to not save those embeddings?

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by investigating how a trained UMAP model stores the training-set embeddings and what other state contributes to the reported 8 GB size for roughly 800k examples. Determine whether those embeddings can be omitted while preserving the model's required behavior, and define completion as a substantially smaller saved model with its supported operations still working.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.