# nnaisense/evotorch

Advanced evolutionary computation library built directly on top of PyTorch, created at NNAISENSE.

Repository: https://github.com/nnaisense/evotorch
Canonical: https://ross.abutalabs.com/products/evotorch
Homepage: https://evotorch.ai
Language: Python
License: Apache-2.0
License Family: permissive
Topics: artificial-intelligence, evolutionary-computation, neural-networks, optimization-algorithms, distributed, gpu, python, pytorch, optimization, reinforcement-learning, evolutionary-algorithms, neuroevolution
Last push: 2026-08-17T20:29:59+00:00

## Health v2 (maintenance only)
Score: 78/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 40, longevity 100
- inputs: {"age_days": 1486, "days_push": 16, "days_rel": 476, "gap_med": 0, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1142, forks 79 (observed 2026-08-28T04:03:44.824282+00:00)

## What it is
EvoTorch is an open-source evolutionary computation library built on top of PyTorch, developed at NNAISENSE. It provides distribution-based (PGPE, XNES, CMA-ES, SNES, CEM) and population-based (GeneticAlgorithm/NSGA-II, CoSyNE, MAPElites) search algorithms for black-box optimization, reinforcement learning, and supervised learning tasks.

## Use cases
- solve black-box optimization problems without gradients
- train neural network policies via neuroevolution
- run CMA-ES or PGPE on GPU with PyTorch
- perform multi-objective optimization with NSGA-II
- apply quality-diversity search with MAPElites
- evolve solutions to reinforcement learning tasks

## When to choose
- your objective is non-differentiable, noisy, or black-box
- you want evolutionary algorithms accelerated on GPU via PyTorch
- you need distributed or parallelized population-based search
- you want neuroevolution for RL or supervised learning

## When to avoid
- your problem is smooth and gradient descent works well
- you need a lightweight non-PyTorch optimizer
- you only need simple hyperparameter tuning (use optuna instead)

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, gpu-computing, reinforcement-learning
- domain: artificial-intelligence, machine-learning, reinforcement-learning
- platform: python, cross-platform
- tags: evolutionary-computation, neuroevolution, pytorch, genetic-algorithm, cma-es, black-box-optimization, quality-diversity, optimization, algorithms, gpu

## Member repositories
- nnaisense/evotorch (main) score 78

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:44.824282+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 2026-08-30T06:35:16.329418+00:00, confidence not recorded.
  - readme: https://github.com/nnaisense/evotorch (fetched 2026-08-28T04:03:44.824282+00:00, sha 7a48ed50abf7)
  - homepage: https://evotorch.ai (fetched 2026-08-29T12:40:18.711438+00:00, sha c387ad1b7859)
  - registry_pypi: https://pypi.org/pypi/evotorch/json (fetched 2026-08-29T12:40:18.713766+00:00, sha 720cf603e9ad)
- Data as of 2026-08-30T08:39:29.467469+00:00.