# dbolya/yolact

A simple, fully convolutional model for real-time instance segmentation.

Repository: https://github.com/dbolya/yolact
Canonical: https://ross.abutalabs.com/products/yolact
Language: Python
License: MIT
License Family: permissive
Topics: realtime, real-time, instance-segmentation, yolact, pytorch
Last push: 2025-09-09T00:11:59+00:00

## Health v2 (maintenance only)
Score: 51/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 41, release rhythm 35, longevity 100
- inputs: {"age_days": 2990, "days_push": 359, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5241, forks 1323 (observed 2026-08-28T04:09:13.575126+00:00)

## What it is
YOLACT is a PyTorch implementation of a fully convolutional model for real-time instance segmentation, accompanying the YOLACT and YOLACT++ research papers. It provides training and evaluation on COCO, pretrained models, and real-time inference at ~30 fps on a Titan Xp.

## Use cases
- run real-time instance segmentation on images or video
- segment individual objects in a webcam feed
- train an instance segmentation model on COCO
- evaluate a segmentation model on COCO test-dev
- use pretrained YOLACT++ weights for object masks
- benchmark real-time segmentation fps on GPU

## When to choose
- you need fast, real-time instance segmentation with per-object masks
- you want a simple, fully convolutional segmentation architecture to study or extend
- you need pretrained COCO models with a permissive MIT license

## When to avoid
- you need state-of-the-art accuracy rather than speed (newer transformers outperform it)
- you need a maintained production library with easy pip installation
- you cannot compile CUDA extensions like DCNv2 for YOLACT++
- you need semantic or panoptic segmentation instead of instance segmentation

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, machine-learning, image-processing
- platform: python, cross-platform
- tags: instance-segmentation, real-time, pytorch, yolact, object-detection, coco, research, gpu, linux

## Member repositories
- dbolya/yolact (main) score 51

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:13.575126+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 2026-08-29T17:59:36.976189+00:00, confidence not recorded.
  - readme: https://github.com/dbolya/yolact (fetched 2026-08-28T04:09:13.575126+00:00, sha a268c01eb8d9)
- Data as of 2026-08-30T08:39:29.467469+00:00.