beringresearch/ivis

Dimensionality reduction in very large datasets using Siamese Networks

65
/ 100
Established

Implements triplet loss-based Siamese networks optimized for millions of observations, supporting both unsupervised and supervised learning modes. Handles diverse data formats (numpy arrays, sparse matrices, HDF5) and feature types (categorical/continuous), with sklearn-compatible `transform` for incremental embedding of new points. Built on TensorFlow with GPU acceleration available, designed to preserve both local and global dataset structure better than t-SNE.

343 stars and 1,087 monthly downloads. Available on PyPI.

Maintenance 6 / 25
Adoption 17 / 25
Maturity 25 / 25
Community 17 / 25

How are scores calculated?

Stars

343

Forks

44

Language

Python

License

Apache-2.0

Last pushed

Nov 10, 2025

Monthly downloads

1,087

Commits (30d)

0

Dependencies

5

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/beringresearch/ivis"

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