# open-mmlab/mmselfsup

OpenMMLab Self-Supervised Learning Toolbox and Benchmark

Repository: https://github.com/open-mmlab/mmselfsup
Canonical: https://ross.abutalabs.com/products/mmselfsup
Homepage: https://mmselfsup.readthedocs.io/en/latest/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: self-supervised-learning, unsupervised-learning, pytorch, moco, simclr, simsiam, mae, masked-image-modeling, beit
Last push: 2023-06-25T02:30:28+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2273, "days_push": 1166, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3302, forks 438 (observed 2026-08-28T04:07:55.722225+00:00)

## What it is
OpenMMLab's PyTorch-based toolbox and benchmark for self-supervised and unsupervised visual representation learning. It provides implementations of methods like MoCo, SimCLR, SimSiam, MAE, and BEiT with a model zoo of pre-trained backbones.

## Use cases
- train self-supervised image models with moco or simclr
- pretrain vision transformers with masked image modeling
- benchmark unsupervised representation learning methods
- get pretrained backbones for downstream image classification
- reproduce mae or beit pretraining results in pytorch

## When to choose
- you need a unified PyTorch framework to train or compare self-supervised vision algorithms
- you want ready-made configs and model zoo checkpoints for SSL methods
- you are already in the OpenMMLab ecosystem

## When to avoid
- you need supervised classification or detection out of the box - use MMPreTrain or MMDetection instead
- you need actively developed features - the repo has been superseded by MMPreTrain and releases stopped in 2023

## Facets
- artifact type: framework
- maturity: maintenance
- function: machine-learning, deep-learning
- domain: machine-learning, computer-vision, deep-learning
- platform: python
- tags: self-supervised-learning, pytorch, moco, simclr, simsiam, mae, masked-image-modeling, beit, benchmark, openmmlab, gpu

## Member repositories
- open-mmlab/mmselfsup (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:55.722225+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 2026-08-30T07:22:40.228530+00:00, confidence not recorded.
  - readme: https://github.com/open-mmlab/mmselfsup (fetched 2026-08-28T04:07:55.722225+00:00, sha 4df40f39af07)
  - registry_pypi: https://pypi.org/pypi/mmselfsup/json (fetched 2026-08-29T09:36:00.121547+00:00, sha fba4454cef07)
- Data as of 2026-08-30T08:39:29.467469+00:00.