# sfzhang15/ATSS

Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection, CVPR, Oral, 2020

Repository: https://github.com/sfzhang15/ATSS
Canonical: https://ross.abutalabs.com/products/atss
Language: Python
License: NOASSERTION
License Family: other
Last push: 2021-03-17T05:17:34+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2463, "days_push": 1995, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1086, forks 161 (observed 2026-08-28T04:03:32.094296+00:00)

## What it is
Official PyTorch implementation of ATSS (Adaptive Training Sample Selection), a CVPR 2020 Oral paper on object detection. It automatically selects positive and negative training samples based on statistical characteristics of objects, bridging the gap between anchor-based and anchor-free detectors and improving state-of-the-art detectors to 50.7% AP on MS COCO without overhead.

## Use cases
- train an object detector with adaptive positive/negative sample selection
- reproduce CVPR 2020 ATSS results on MS COCO
- improve anchor-free detectors like FCOS with center sampling
- run inference with a pretrained ATSS ResNet-50 FPN model
- compare anchor-based vs anchor-free detection training strategies
- use ATSS as a baseline for object detection research

## When to choose
- you need a research-grade object detector with adaptive training sample selection
- you want to reproduce or build on the ATSS CVPR 2020 paper
- you are benchmarking anchor-based vs anchor-free detectors on COCO

## When to avoid
- you need a production-ready, actively maintained detection framework
- you want plug-and-play detection without building maskrcnn-benchmark dependencies
- you need detection on platforms other than Linux with GPUs

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, machine-learning
- platform: python
- tags: object-detection, anchor-free-detection, cvpr-2020, pytorch, research-code, coco, training-sample-selection, linux, gpu

## Member repositories
- sfzhang15/ATSS (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:32.094296+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 2026-08-30T06:49:37.190574+00:00, confidence not recorded.
  - readme: https://github.com/sfzhang15/ATSS (fetched 2026-08-28T04:03:32.094296+00:00, sha 6b70c2dea992)
- Data as of 2026-08-30T08:39:29.467469+00:00.