# GMvandeVen/continual-learning

PyTorch implementation of various methods for continual learning (XdG, EWC, SI, LwF, FROMP, DGR, BI-R, ER, A-GEM, iCaRL, Generative Classifier) in three different scenarios.

Repository: https://github.com/GMvandeVen/continual-learning
Canonical: https://ross.abutalabs.com/products/continual-learning
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: deep-learning, artificial-neural-networks, continual-learning, lifelong-learning, incremental-learning, replay, distillation, generative-models, variational-autoencoder, elastic-weight-consolidation, replay-through-feedback, icarl, gradient-episodic-memory, task-incremental-learning, class-incremental-learning, domain-incremental-learning
Last push: 2025-11-05T20:14:56+00:00

## Health v2 (maintenance only)
Score: 45/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 50, release rhythm 8, longevity 100
- inputs: {"age_days": 2898, "days_push": 301, "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 1879, forks 345 (observed 2026-08-28T04:05:48.507977+00:00)

## What it is
A PyTorch research codebase implementing many continual learning methods (EWC, SI, LwF, XdG, replay variants, iCaRL, generative classifier, etc.) across task-, domain-, and class-incremental scenarios. It accompanies the Nature Machine Intelligence article 'Three types of incremental learning' and related tutorials and blog posts.

## Use cases
- reproduce continual learning experiments from papers
- compare continual learning methods like EWC, SI, and LwF
- run task-incremental learning on Split MNIST
- experiment with class-incremental learning scenarios
- learn how generative replay works in continual learning
- benchmark lifelong learning algorithms in PyTorch

## When to choose
- you need reference implementations of many continual learning algorithms in one codebase
- you want to reproduce results from the 'Three types of incremental learning' paper
- you are studying or teaching lifelong/incremental learning with neural networks

## When to avoid
- you need a production-ready continual learning library with stable APIs
- you want continual learning for non-classification tasks like NLP or detection
- you need support for very large-scale datasets or distributed training

## Facets
- artifact type: learning-resource
- maturity: active
- function: deep-learning, machine-learning
- domain: deep-learning, machine-learning, artificial-intelligence
- platform: python
- tags: continual-learning, lifelong-learning, incremental-learning, pytorch, research-code, replay, elastic-weight-consolidation, generative-replay

## Member repositories
- GMvandeVen/continual-learning (main) score 45

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:48.507977+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 2026-08-30T03:14:02.667288+00:00, confidence not recorded.
  - readme: https://github.com/GMvandeVen/continual-learning (fetched 2026-08-28T04:05:48.507977+00:00, sha 76edefadc176)
- Data as of 2026-08-30T08:39:29.467469+00:00.