# open-mmlab/mmrazor

OpenMMLab Model Compression Toolbox and Benchmark.

Repository: https://github.com/open-mmlab/mmrazor
Canonical: https://ross.abutalabs.com/products/mmrazor
Homepage: https://mmrazor.readthedocs.io/en/latest/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: nas, pruning, knowledge-distillation, spos, darts, autoslim, detection, segmentation, classification, pytorch, quantization
Last push: 2024-06-11T09:30:27+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1715, "days_push": 813, "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 1682, forks 245 (observed 2026-08-28T04:05:21.746172+00:00)

## What it is
MMRazor is OpenMMLab's model compression toolbox and benchmark built on PyTorch. It provides implementations of neural architecture search, pruning, knowledge distillation, and quantization algorithms for vision tasks like detection, segmentation, and classification.

## Use cases
- compress a pytorch detection model with knowledge distillation
- prune a classification network to reduce inference cost
- run neural architecture search with darts or autoslim
- quantize a segmentation model for deployment
- benchmark model compression algorithms on vision tasks

## When to choose
- you use the OpenMMLab ecosystem (mmdetection, mmclassification) and want to compress those models
- you need a unified framework for NAS, pruning, and distillation experiments
- you want reproducible benchmarks of compression algorithms on vision models

## When to avoid
- you need compression for NLP or transformer LLMs rather than vision models
- you want a lightweight single-technique tool rather than a full framework
- you need actively developed features, as the project's latest release is mid-2024

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, benchmarking
- domain: deep-learning, computer-vision, machine-learning
- platform: python
- tags: model-compression, knowledge-distillation, pruning, nas, quantization, pytorch, openmmlab, gpu, linux

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

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:21.746172+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 2026-08-30T03:41:23.126275+00:00, confidence not recorded.
  - readme: https://github.com/open-mmlab/mmrazor (fetched 2026-08-28T04:05:21.746172+00:00, sha 3714c36ec2a8)
  - registry_pypi: https://pypi.org/pypi/mmrazor/json (fetched 2026-08-29T11:14:50.254388+00:00, sha ca137d6ce01e)
- Data as of 2026-08-30T08:39:29.467469+00:00.