# e3nn/e3nn

A modular framework for neural networks with Euclidean symmetry

Repository: https://github.com/e3nn/e3nn
Canonical: https://ross.abutalabs.com/products/e3nn
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-02-13T22:16:58+00:00

## Health v2 (maintenance only)
Score: 70/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 67, release rhythm 58, longevity 100
- inputs: {"age_days": 2406, "days_push": 201, "days_rel": 201, "gap_med": 57, "n_releases_24m": 8}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1280, forks 185 (observed 2026-08-28T04:04:13.683544+00:00)

## What it is
e3nn is a Python/PyTorch library for building E(3)-equivariant neural networks that respect 3D rotation, translation, and mirror symmetries. It provides fundamental operations such as tensor products, spherical harmonics, and equivariant linear layers for geometric deep learning.

## Use cases
- build neural networks that respect 3D rotational symmetry
- compute spherical harmonics in PyTorch
- create equivariant tensor product layers
- model molecules and materials with symmetry-aware networks
- predict 3D properties like forces and energies equivariantly
- develop geometric deep learning models for point clouds

## When to choose
- your data has 3D Euclidean symmetry (rotations, translations, mirrors)
- you need equivariant operations on irreducible representations in PyTorch
- you are doing molecular, material, or point-cloud learning research

## When to avoid
- your problem has no geometric or rotational symmetry
- you need a production-ready stable API with no breaking changes
- you do not use PyTorch

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, math, graphics
- domain: machine-learning, deep-learning, chemistry, simulation
- platform: python
- tags: equivariant-neural-networks, e3-symmetry, spherical-harmonics, tensor-products, pytorch, geometric-deep-learning, molecular-modeling, algorithms

## Member repositories
- e3nn/e3nn (main) score 70

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:13.683544+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:02:36.647073+00:00, confidence not recorded.
  - readme: https://github.com/e3nn/e3nn (fetched 2026-08-28T04:04:13.683544+00:00, sha d73d3496eea2)
  - registry_pypi: https://pypi.org/pypi/e3nn/json (fetched 2026-08-29T12:13:14.949976+00:00, sha 6e519d38c161)
- Data as of 2026-08-30T08:39:29.467469+00:00.