# hollance/neural-engine

Everything we actually know about the Apple Neural Engine (ANE)

Repository: https://github.com/hollance/neural-engine
Canonical: https://ross.abutalabs.com/products/neural-engine
License: MIT
License Family: permissive
Topics: neural-engine, ane, coreml, iphone, ios, neural-network, tpu
Last push: 2026-03-12T11:59:20+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 71, release rhythm 35, longevity 100
- inputs: {"age_days": 2331, "days_push": 174, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2522, forks 97 (observed 2026-08-28T04:06:58.345246+00:00)

## What it is
A community-maintained documentation resource explaining how the Apple Neural Engine (ANE) works, based on reverse-engineering and experimentation. It covers which devices have an ANE, how to get Core ML models running on it, its limitations, and unsupported layers.

## Use cases
- why is my core ml model not running on the neural engine
- how to make my model run on the apple neural engine
- which core ml layers are unsupported on the ane
- is my model using the ane or the gpu
- how fast is the ane compared to gpu on iphone
- can i program the apple neural engine directly
- which iphone and ipad devices have a neural engine

## When to choose
- you deploy Core ML models on iOS and want to leverage the ANE
- you're debugging why a model falls back to CPU or GPU
- you want community-documented ANE capabilities and limitations

## When to avoid
- you need official Apple documentation or guaranteed accuracy
- you're targeting Android or non-Apple NPUs
- you need executable code or a library rather than reference material

## Facets
- artifact type: learning-resource
- maturity: active
- function: machine-learning, documentation, developer-tools
- domain: machine-learning, hardware, mobile-development, tutorials
- platform: cross-platform
- tags: apple-neural-engine, ane, coreml, npu, on-device-ml, iphone, ipad, community-documentation, ios, macos

## Member repositories
- hollance/neural-engine (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:58.345246+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 2026-08-30T02:25:47.699779+00:00, confidence not recorded.
  - readme: https://github.com/hollance/neural-engine (fetched 2026-08-28T04:06:58.345246+00:00, sha ff3dc17ba035)
- Data as of 2026-08-30T08:39:29.467469+00:00.