# ARM-software/armnn

Arm NN ML Software.

Repository: https://github.com/ARM-software/armnn
Canonical: https://ross.abutalabs.com/products/armnn
Homepage: https://developer.arm.com/products/processors/machine-learning/arm-nn
Language: C++
License: MIT
License Family: permissive
Topics: machine-learning, neural-network, neural-networks
Last push: 2026-07-07T11:20:08+00:00

## Health v2 (maintenance only)
Score: 85/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 91, release rhythm 68, longevity 100
- inputs: {"age_days": 3099, "days_push": 57, "days_rel": 58, "gap_med": 123.0, "n_releases_24m": 5}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1316, forks 328 (observed 2026-08-28T04:04:20.910739+00:00)

## What it is
Arm NN is an open-source C++17 machine learning inference SDK that accelerates neural network execution on Arm Cortex-A CPUs, Mali GPUs, and Ethos-N NPUs, built on top of the Arm Compute Library. It includes a TF Lite Delegate for Python and parser support for common ML frameworks, though it is now a legacy project no longer maintained by Arm.

## Use cases
- run neural network inference on arm cortex-a cpus
- accelerate tflite models on mali gpu
- deploy ml models on android with arm optimizations
- run inference on arm ethos-n npu
- execute quantized models efficiently on edge devices
- accelerate machine learning workloads on linux arm servers

## When to choose
- you need maximum inference performance on Arm Cortex-A CPUs or Mali GPUs in a trusted environment
- you are targeting Arm Ethos-N NPUs and want an existing driver-backed stack
- you want a C++ inference SDK with a TF Lite Delegate for Python integration

## When to avoid
- you need security patches or ongoing maintenance, since the project is officially legacy and unmaintained
- you will process untrusted inputs or run in hostile environments, which the maintainers explicitly warn against
- you target non-Arm hardware or need Cortex-M acceleration, where CMSIS-NN or other tools are recommended

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, sdk
- domain: machine-learning, embedded-systems, developer-tools
- platform: cpp, python
- tags: inference-engine, arm-cortex, mali-gpu, tflite-delegate, ethos-npu, arm-compute-library, legacy, edge-ai, linux, android

## Member repositories
- ARM-software/armnn (main) score 85

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:20.910739+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 2026-08-30T04:48:19.808034+00:00, confidence not recorded.
  - readme: https://github.com/ARM-software/armnn (fetched 2026-08-28T04:04:20.910739+00:00, sha 79e2f6231ad6)
- Data as of 2026-08-30T08:39:29.467469+00:00.