# microsoft/EdgeML

This repository provides code for machine learning algorithms for edge devices developed at Microsoft Research India.

Repository: https://github.com/microsoft/EdgeML
Canonical: https://ross.abutalabs.com/products/edgeml
Language: C++
License: NOASSERTION
License Family: other
Topics: iot-device, sensor, edge-machine-learning, resource-constrained-ml, machine-learning-algorithms, microsoft-research, machine-learning, tensorflow, deep-learning, edge-computing, microsoft, edge-devices, classifier, emi-rnn, pytorch, bonsai, protonn, fastgrnn, cpp, fastrnn
Last push: 2024-05-20T12:28:33+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": 3318, "days_push": 835, "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 1681, forks 387 (observed 2026-08-28T04:05:21.582833+00:00)

## What it is
A library from Microsoft Research India implementing resource-efficient machine learning algorithms (Bonsai, ProtoNN, FastGRNN, EMI-RNN, DROCC, RNNPool) for edge and IoT devices. It provides TensorFlow and PyTorch packages, C++ inference code, and a fixed-point quantization tool (SeeDot) so tiny models can run offline on microcontrollers.

## Use cases
- train tiny classifiers that fit in kilobytes on IoT sensors
- run gesture recognition on microcontrollers
- efficient RNN inference on RAM-constrained devices
- anomaly detection for resource-scarce edge devices
- quantize models to fixed-point arithmetic for embedded inference
- recover critical signatures from time series for fast RNN predictions

## When to choose
- you need ML models with kilobyte-scale footprints for IoT or embedded hardware
- you want offline, low-latency predictions without cloud connectivity
- you're working with time-series sensor data on constrained devices

## When to avoid
- you need state-of-the-art accuracy on large-scale datasets with ample compute
- you want a general-purpose deep learning framework rather than specialized edge algorithms
- you need actively developed features or broad community support

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, llm-training
- domain: machine-learning, iot, embedded-systems
- platform: python, cpp, cross-platform, embedded
- tags: edge-ml, tinyml, rnn, bonsai, protonn, fastgrnn, quantization, tensorflow, pytorch, microsoft-research

## Member repositories
- microsoft/EdgeML (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:21.582833+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 2026-08-30T03:41:26.070734+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/EdgeML (fetched 2026-08-28T04:05:21.582833+00:00, sha a31451668c5f)
- Data as of 2026-08-30T08:39:29.467469+00:00.