arpitg1304/tessera

Visualize episode embeddings and select maximally diverse training subsets for robotics ML. Train on 10K diverse episodes instead of 50K random ones.

28
/ 100
Experimental

Tessera helps robotics machine learning engineers curate better training datasets. It visualizes high-dimensional robotics episode data, showing how different training episodes relate to each other. Users upload their episode embeddings and metadata, then interactively select maximally diverse or specifically filtered subsets for training, which can be downloaded as episode IDs.

Use this if you need to select a smaller, more effective subset of robotics training episodes from a large dataset, rather than training on redundant or randomly sampled data.

Not ideal if you're not working with robotics episode data or if you need a general-purpose embedding visualization tool without specific dataset curation features.

robotics-ml dataset-curation machine-learning-engineering robot-training data-diversity
No Package No Dependents
Maintenance 10 / 25
Adoption 5 / 25
Maturity 13 / 25
Community 0 / 25

How are scores calculated?

Stars

9

Forks

Language

TypeScript

License

MIT

Last pushed

Jan 17, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/arpitg1304/tessera"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.