# CuriousAI/mean-teacher

A state-of-the-art semi-supervised method for image recognition

Repository: https://github.com/CuriousAI/mean-teacher
Canonical: https://ross.abutalabs.com/products/mean-teacher
Homepage: https://arxiv.org/abs/1703.01780
Language: Python
License: NOASSERTION
License Family: other
Topics: deep-learning, semi-supervised-learning, nips-2017, tensorflow, pytorch
Last push: 2020-10-08T12:48:09+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": 3397, "days_push": 2155, "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 1678, forks 342 (observed 2026-08-28T04:05:20.997380+00:00)

## What it is
Reference implementations (TensorFlow and PyTorch) of the Mean Teacher semi-supervised learning method from the NIPS 2017 paper by Tarvainen and Valpola. It trains a student model alongside a weight-averaged (EMA) teacher model with a consistency cost, achieving state-of-the-art semi-supervised image classification results on CIFAR-10, SVHN, and ImageNet.

## Use cases
- train an image classifier with only a small fraction of labeled data
- reproduce the NIPS 2017 Mean Teacher paper results on CIFAR-10 or ImageNet
- apply semi-supervised learning with EMA teacher weights to my own model
- compare semi-supervised methods like temporal ensembling vs mean teacher
- adapt a ResNet to learn from 4000 labels on CIFAR-10
- add a consistency cost between student and teacher outputs in PyTorch

## When to choose
- you have limited labeled data for an image classification task
- you want to reproduce or build on the Mean Teacher paper
- you prefer a simple, well-known semi-supervised baseline with PyTorch or TensorFlow code

## When to avoid
- you need actively maintained code with recent framework versions
- your task is not image classification or requires heavy customization
- you need the latest state-of-the-art semi-supervised methods (newer techniques exist)

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning
- domain: machine-learning, deep-learning, computer-vision
- platform: python
- tags: semi-supervised-learning, mean-teacher, nips-2017, tensorflow, pytorch, image-classification, research-code, gpu

## Member repositories
- CuriousAI/mean-teacher (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:20.997380+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 2026-08-30T03:41:32.917246+00:00, confidence not recorded.
  - readme: https://github.com/CuriousAI/mean-teacher (fetched 2026-08-28T04:05:20.997380+00:00, sha 3e089248444a)
  - homepage: https://arxiv.org/abs/1703.01780 (fetched 2026-08-29T11:14:56.281991+00:00, sha 37ab6db34110)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T11:14:56.291107+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T11:14:56.294921+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T11:14:56.296871+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T11:14:56.293126+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.