# jettify/pytorch-optimizer

torch-optimizer -- collection of optimizers for Pytorch

Repository: https://github.com/jettify/pytorch-optimizer
Canonical: https://ross.abutalabs.com/products/pytorch-optimizer
Language: Python
License: Apache-2.0
License Family: permissive
Topics: pytorch, optimizer, diffgrad, adamod, lamb, yogi, accsgd, adabound, novograd, shampoo, lookahead, swats, sgdp, adabelief, apollo, hacktoberfest
Last push: 2024-03-22T11:10:03+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2434, "days_push": 894, "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 3171, forks 307 (observed 2026-08-28T04:07:47.462136+00:00)

## What it is
torch-optimizer is a Python library providing a collection of additional optimization algorithms for PyTorch, drop-in compatible with the torch.optim module. It implements optimizers like DiffGrad, AdaBound, Lamb, Yogi, Lookahead, Shampoo, and many others from recent research papers.

## Use cases
- find alternative optimizers for pytorch model training
- use adabound or lamb optimizer in pytorch
- compare optimization algorithms for deep learning experiments
- install extra optimizers not included in torch.optim
- speed up neural network convergence with newer optimizers

## When to choose
- you need research optimizers like DiffGrad, AdaBound, Lamb, or Lookahead in PyTorch
- you want a drop-in replacement API compatible with torch.optim
- you are benchmarking different optimization algorithms

## When to avoid
- standard SGD or Adam suffices for your training
- you work in TensorFlow or JAX rather than PyTorch
- you need actively developed optimizers with frequent updates

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning
- domain: machine-learning, deep-learning, developer-tools
- platform: python
- tags: pytorch, optimizers, gradient-descent, adam-variants, training

## Member repositories
- jettify/pytorch-optimizer (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:47.462136+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 2026-08-29T18:45:28.931061+00:00, confidence not recorded.
  - readme: https://github.com/jettify/pytorch-optimizer (fetched 2026-08-28T04:07:47.462136+00:00, sha c531421ca917)
- Data as of 2026-08-30T08:39:29.467469+00:00.