# yhenon/pytorch-retinanet

Pytorch implementation of RetinaNet object detection.

Repository: https://github.com/yhenon/pytorch-retinanet
Canonical: https://ross.abutalabs.com/products/pytorch-retinanet
Language: Python
License: Apache-2.0
License Family: permissive
Topics: pytorch, retinanet, python
Archived: true
Last push: 2023-04-29T14:16:18+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3053, "days_push": 1222, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2205, forks 675 (observed 2026-08-28T04:06:25.973219+00:00)

## What it is
A PyTorch implementation of the RetinaNet object detection model with focal loss, designed for readability and easy modification. It includes training and validation scripts with COCO and CSV dataloaders plus a pre-trained ResNet-50 model.

## Use cases
- train a RetinaNet object detection model in PyTorch
- train an object detector on a custom dataset with CSV annotations
- evaluate object detection mAP on COCO
- use a pre-trained RetinaNet model for object detection
- modify a simple object detection implementation for research

## When to choose
- you want a readable, hackable RetinaNet implementation in PyTorch
- you need to train object detection on custom datasets via CSV annotations
- you want to reproduce the Focal Loss paper results

## When to avoid
- you need a production-grade, actively maintained detection framework like MMDetection or Detectron2
- you need the latest detection architectures or optimized training performance

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, computer-vision, deep-learning
- domain: computer-vision, machine-learning, deep-learning
- platform: python
- tags: object-detection, retinanet, pytorch, focal-loss, coco

## Member repositories
- yhenon/pytorch-retinanet (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:25.973219+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 2026-08-30T02:46:42.541109+00:00, confidence not recorded.
  - readme: https://github.com/yhenon/pytorch-retinanet (fetched 2026-08-28T04:06:25.973219+00:00, sha ac1e9a894a08)
- Data as of 2026-08-30T08:39:29.467469+00:00.