# dog-qiuqiu/Yolo-Fastest

:zap: Based on yolo's ultra-lightweight universal target detection algorithm, the calculation amount is only 250mflops, the ncnn model size is only 666kb, the Raspberry Pi 3b can run up to 15fps+, and the mobile terminal can run up to 178fps+

Repository: https://github.com/dog-qiuqiu/Yolo-Fastest
Canonical: https://ross.abutalabs.com/products/yolo-fastest
Language: C
License: NOASSERTION
License Family: other
Last push: 2021-08-11T16:55:55+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": 2203, "days_push": 1848, "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 2103, forks 434 (observed 2026-08-28T04:06:13.735471+00:00)

## What it is
Yolo-Fastest is an ultra-lightweight YOLO-based object detection algorithm and model zoo, with only ~250 MFLOPs and a 666KB ncnn model. It includes a modified darknet training framework and is optimized for real-time inference on ARM mobile and embedded devices like Raspberry Pi.

## Use cases
- run real-time object detection on a Raspberry Pi
- deploy a tiny yolo model on Android mobile devices
- detect objects on embedded boards with very limited compute
- train a lightweight custom object detector with darknet
- run single object detection at 30fps on edge hardware

## When to choose
- you need the smallest, fastest object detection model for resource-constrained devices
- you target ARM mobile or embedded inference with ncnn
- simple single-class or single-object detection is sufficient

## When to avoid
- you need high accuracy on complex multi-object scenes (mAP is only ~24-34%)
- you need CPU inference via darknet, which is not optimized
- you need a well-maintained project - the successor Yolo-FastestV2 is recommended instead

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, computer-vision, deep-learning, llm-inference
- domain: computer-vision, machine-learning, embedded-systems, deep-learning
- platform: cross-platform, embedded, cpp, python
- tags: object-detection, yolo, ncnn, ultra-lightweight, edge-devices, real-time, darknet, model-zoo, android, linux, gpu

## Member repositories
- dog-qiuqiu/Yolo-Fastest (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:13.735471+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 2026-08-30T02:54:14.060680+00:00, confidence not recorded.
  - readme: https://github.com/dog-qiuqiu/Yolo-Fastest (fetched 2026-08-28T04:06:13.735471+00:00, sha 42267c4c8ab1)
- Data as of 2026-08-30T08:39:29.467469+00:00.