# google-deepmind/learning-to-learn

Learning to Learn in TensorFlow

Repository: https://github.com/google-deepmind/learning-to-learn
Canonical: https://ross.abutalabs.com/products/learning-to-learn
Homepage: https://arxiv.org/abs/1606.04474
Language: Python
License: Apache-2.0
License Family: permissive
Topics: machine-learning, artificial-intelligence, neural-networks, deep-learning
Last push: 2021-06-29T09:55:15+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": 3557, "days_push": 1891, "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 4068, forks 602 (observed 2026-08-28T04:08:34.162826+00:00)

## What it is
A TensorFlow implementation of 'Learning to learn by gradient descent by gradient descent' (DeepMind, arXiv:1606.04474), which trains LSTM-based learned optimizers via meta-learning. It includes training and evaluation scripts and example problems such as quadratic functions, MNIST, and CIFAR-10.

## Use cases
- train a learned optimizer with meta-learning
- reproduce the learning-to-learn paper experiments
- compare learned optimizers against Adam on MNIST or CIFAR-10
- implement custom problems for meta-learned optimization
- research on LSTM-based optimizers

## When to choose
- you are researching learned optimizers or meta-learning and want a reference implementation
- you need to reproduce the arXiv:1606.04474 results
- you work with TensorFlow 1.x and Sonnet and want to experiment with learned optimizers

## When to avoid
- you need a production optimizer - use standard optimizers like Adam instead
- you use modern TensorFlow 2.x, PyTorch, or JAX - this code targets TensorFlow 1.x and is unmaintained
- you need actively maintained software with recent updates

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, deep-learning
- domain: machine-learning, deep-learning, artificial-intelligence
- platform: python
- tags: learned-optimizers, meta-learning, lstm, tensorflow, research-code, gradient-descent

## Member repositories
- google-deepmind/learning-to-learn (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:34.162826+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 2026-08-29T18:23:32.279591+00:00, confidence not recorded.
  - readme: https://github.com/google-deepmind/learning-to-learn (fetched 2026-08-28T04:08:34.162826+00:00, sha 1197f7874203)
  - homepage: https://arxiv.org/abs/1606.04474 (fetched 2026-08-29T09:15:39.785185+00:00, sha f5f66a7d3a12)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T09:15:39.794672+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T09:15:39.798554+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T09:15:39.800498+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T09:15:39.796704+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.