# huawei-noah/Efficient-AI-Backbones

Efficient AI Backbones including GhostNet, TNT and MLP, developed by Huawei Noah's Ark Lab.

Repository: https://github.com/huawei-noah/Efficient-AI-Backbones
Canonical: https://ross.abutalabs.com/products/efficient-ai-backbones
Language: Python
License Family: other
Topics: convolutional-neural-networks, efficient-inference, imagenet, model-compression, tensorflow, pytorch, ghostnet, transformer, pretrained-models, vision-transformer
Last push: 2025-03-15T12:48:07+00:00

## Health v2 (maintenance only)
Score: 28/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 11, release rhythm 8, longevity 100
- inputs: {"age_days": 2482, "days_push": 536, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- 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 4418, forks 735 (observed 2026-08-28T04:08:48.708753+00:00)

## What it is
A collection of efficient neural network backbone architectures (GhostNet, TNT, ViG, WaveMLP, TinyNet, etc.) from Huawei Noah's Ark Lab, with PyTorch and MindSpore implementations and pretrained ImageNet weights. It serves as a research model zoo for lightweight and vision transformer models.

## Use cases
- find efficient cnn backbones for mobile deployment
- download pretrained ghostnet weights for imagenet
- implement vision transformer models in pytorch
- compare lightweight model architectures for image classification
- use vision gnn models in my project
- compress models for edge inference

## When to choose
- you need efficient, lightweight image classification backbones with pretrained weights
- you want reference implementations of published architectures like GhostNet or TNT
- you work in PyTorch or MindSpore and need vision model baselines

## When to avoid
- you need a production-ready pip-installable model library with stable APIs
- you need non-vision tasks like NLP or audio
- you require a permissive license - the repo has no explicit license, limiting reuse

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing
- domain: deep-learning, computer-vision, image-processing, artificial-intelligence
- platform: python
- tags: vision-transformer, ghostnet, model-zoo, pretrained-models, efficient-inference, model-compression, pytorch, mindspore, imagenet, research-code, gpu

## Member repositories
- huawei-noah/Efficient-AI-Backbones (main) score 28

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:48.708753+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 2026-08-29T18:20:59.948113+00:00, confidence not recorded.
  - readme: https://github.com/huawei-noah/Efficient-AI-Backbones (fetched 2026-08-28T04:08:48.708753+00:00, sha 5e7ae1947557)
- Data as of 2026-08-30T08:39:29.467469+00:00.