# google-research/deeplab2

DeepLab2 is a TensorFlow library for deep labeling, aiming to provide a unified and state-of-the-art TensorFlow codebase for dense pixel labeling tasks.

Repository: https://github.com/google-research/deeplab2
Canonical: https://ross.abutalabs.com/products/deeplab2
Language: Python
License: Apache-2.0
License Family: permissive
Archived: true
Last push: 2023-04-17T22:02:27+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": 1939, "days_push": 1234, "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 1037, forks 164 (observed 2026-08-28T04:03:19.465389+00:00)

## What it is
DeepLab2 is a TensorFlow library from Google Research providing a unified, state-of-the-art codebase for dense pixel labeling tasks such as semantic, instance, and panoptic segmentation, depth estimation, and video panoptic segmentation. It includes research models like MaX-DeepLab, kMaX-DeepLab, ViP-DeepLab, and MOAT with pretrained checkpoints.

## Use cases
- train a panoptic segmentation model on Cityscapes or COCO
- run semantic segmentation inference on images with pretrained checkpoints
- estimate per-pixel depth from images
- perform video panoptic segmentation on autonomous driving datasets like KITTI-STEP or Waymo
- fine-tune kMaX-DeepLab or ViP-DeepLab on a custom segmentation dataset
- evaluate segmentation models with metrics like Panoptic Quality or Segmentation and Tracking Quality

## When to choose
- you need state-of-the-art dense pixel labeling models in TensorFlow
- you want a unified codebase covering multiple segmentation and depth estimation tasks
- you are doing research on panoptic or video panoptic segmentation
- you need pretrained model zoo checkpoints for segmentation research

## When to avoid
- you prefer PyTorch over TensorFlow
- you need a lightweight production inference library rather than a research codebase
- your task is not pixel-level prediction (e.g., classification or detection only)
- you need actively maintained software with frequent updates

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, machine-learning, deep-learning, image-processing
- platform: python
- tags: tensorflow, semantic-segmentation, panoptic-segmentation, instance-segmentation, depth-estimation, video-panoptic-segmentation, pixel-labeling, model-zoo, gpu, linux

## Member repositories
- google-research/deeplab2 (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:19.465389+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 2026-08-30T07:04:23.177749+00:00, confidence not recorded.
  - readme: https://github.com/google-research/deeplab2 (fetched 2026-08-28T04:03:19.465389+00:00, sha 10a49a1acbfd)
- Data as of 2026-08-30T08:39:29.467469+00:00.