# Tencent/FeatherCNN

FeatherCNN is a high performance inference engine for convolutional neural networks.

Repository: https://github.com/Tencent/FeatherCNN
Canonical: https://ross.abutalabs.com/products/feathercnn
Language: C++
License Family: other
Topics: convolutional-neural-networks, inference-engine, caffe, android, ios, arm-neon
Last push: 2019-09-24T06:34:01+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": 3051, "days_push": 2535, "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 1227, forks 275 (observed 2026-08-28T04:04:03.288356+00:00)

## What it is
FeatherCNN is a high-performance, lightweight CNN inference library developed by Tencent, targeting ARM CPUs on mobile (iOS/Android), embedded Linux, and ARM servers. It converts Caffe models into a single binary format and packs everything into a small dependency-free code base.

## Use cases
- run convolutional neural network inference on mobile devices
- deploy caffe models on android and ios
- run game AI models on phones with low latency
- embed cnn inference in embedded linux arm devices
- run neural network inference on arm servers
- minimize model runtime binary size on mobile

## When to choose
- you need fast CNN inference on ARM CPUs with a tiny footprint
- you have existing Caffe models to deploy on mobile or embedded devices
- you want a dependency-free single code base for easy deployment

## When to avoid
- you need GPU or non-ARM architecture support
- you need models beyond Caffe format such as ONNX or TensorFlow
- you need an actively maintained project with recent updates and support

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, llm-inference, embedded
- domain: machine-learning, mobile-development, embedded-systems, deep-learning
- platform: cpp
- tags: cnn-inference, arm-neon, caffe-models, on-device-ai, lightweight, android, ios, linux, arm

## Member repositories
- Tencent/FeatherCNN (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:03.288356+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 2026-08-30T06:15:14.228171+00:00, confidence not recorded.
  - readme: https://github.com/Tencent/FeatherCNN (fetched 2026-08-28T04:04:03.288356+00:00, sha 210229a4d57f)
- Data as of 2026-08-30T08:39:29.467469+00:00.