# ucbdrive/few-shot-object-detection

Implementations of few-shot object detection benchmarks

Repository: https://github.com/ucbdrive/few-shot-object-detection
Canonical: https://ross.abutalabs.com/products/few-shot-object-detection
Language: Python
License: Apache-2.0
License Family: permissive
Topics: few-shot-object-detection, few-shot-learning, icml-2020
Last push: 2023-11-21T16:16:32+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2361, "days_push": 1016, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1218, forks 223 (observed 2026-08-28T04:04:01.567066+00:00)

## What it is
FsDet is the official implementation of the ICML 2020 paper 'Frustratingly Simple Few-Shot Object Detection' (TFA), built on detectron2. It provides few-shot object detection benchmarks on PASCAL VOC, COCO, and LVIS with pre-trained models and a modular framework for future research.

## Use cases
- run few-shot object detection benchmarks on VOC, COCO, and LVIS
- reproduce TFA results from the ICML 2020 paper
- fine-tune an object detector on a few novel-class examples
- download pre-trained few-shot detection models
- build and evaluate new few-shot detection methods on a modular framework

## When to choose
- you need standardized few-shot object detection benchmarks and splits
- you want to reproduce or compare against the TFA approach
- you are doing research on few-shot learning for detection and want a detectron2-based base

## When to avoid
- you need general-purpose object detection rather than few-shot settings
- you need a production-ready detection system rather than research code
- you require frequent updates or support for the latest detectron2 versions

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, computer-vision, benchmarking
- domain: machine-learning, computer-vision, artificial-intelligence
- platform: python
- tags: few-shot-learning, object-detection, detectron2, research-code, tfa, gpu, linux

## Member repositories
- ucbdrive/few-shot-object-detection (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:01.567066+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 2026-08-30T06:17:05.524914+00:00, confidence not recorded.
  - readme: https://github.com/ucbdrive/few-shot-object-detection (fetched 2026-08-28T04:04:01.567066+00:00, sha cbb9c99d16fa)
- Data as of 2026-08-30T08:39:29.467469+00:00.