# ethanhe42/channel-pruning

Channel Pruning for Accelerating Very Deep Neural Networks (ICCV'17)

Repository: https://github.com/ethanhe42/channel-pruning
Canonical: https://ross.abutalabs.com/products/channel-pruning
Homepage: https://arxiv.org/abs/1707.06168
Language: Python
License: MIT
License Family: permissive
Topics: image-recognition, model-compression, acceleration, object-detection, image-classification, channel-pruning, deep-neural-networks
Last push: 2024-05-02T06:37:43+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": 3299, "days_push": 853, "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 1088, forks 306 (observed 2026-08-28T04:03:32.485892+00:00)

## What it is
Reference implementation of the ICCV 2017 channel pruning method for accelerating very deep convolutional neural networks, using LASSO regression-based channel selection and least-squares reconstruction. It includes pruned VGG-16, ResNet-50, and Faster R-CNN models with 2x-5x speed-ups and pretrained releases.

## Use cases
- prune channels from a trained CNN to speed up inference
- compress a deep neural network for deployment on mobile devices
- accelerate ResNet or VGG-16 with minimal accuracy loss
- speed up Faster R-CNN object detection
- reproduce results from the ICCV 2017 channel pruning paper
- compare model compression techniques for CNNs

## When to choose
- you need to accelerate a pretrained VGG-16, ResNet, or Faster R-CNN with proven 2x-5x speed-ups
- you want the canonical reference implementation of LASSO-based channel pruning
- you are researching model compression and need a reproducible baseline

## When to avoid
- you need actively maintained tooling for modern architectures like transformers or ViTs
- you want automated pruning via reinforcement learning (see AMC instead)
- you need a production-ready compression pipeline rather than research code

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, computer-vision, benchmarking
- domain: deep-learning, computer-vision, machine-learning, image-processing
- platform: python
- tags: model-compression, channel-pruning, neural-network-acceleration, cnn, model-optimization, iccv-2017, lasso-regression, research-code, linux, gpu

## Member repositories
- ethanhe42/channel-pruning (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:32.485892+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 2026-08-30T06:49:21.693375+00:00, confidence not recorded.
  - readme: https://github.com/ethanhe42/channel-pruning (fetched 2026-08-28T04:03:32.485892+00:00, sha 0cacb7789943)
  - homepage: https://arxiv.org/abs/1707.06168 (fetched 2026-08-29T12:52:03.743070+00:00, sha fc062616905e)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T12:52:03.745651+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T12:52:03.749557+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T12:52:03.751537+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T12:52:03.747747+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.