Include Scalable CMA-ES Optimizers - Githubissues

icaros-usc / pyribs

A bare-bones Python library for quality diversity optimization.

https://pyribs.org

MIT License

208 stars 33 forks source link

Include Scalable CMA-ES Optimizers #274

Closed itsdawei closed 1 year ago

itsdawei commented 1 year ago

Description

Include scalable CMA-ES optimizers used in this paper. The implementation is adapted from https://github.com/icaros-usc/scaling-cma-mae/tree/master/src/emitters/opt.

TODO

[x] sep-cma-es
[x] lm-ma-es
[x] openai-es
[x] Add ThreadPool limit to all ask and tell.
[x] Change tell api -> tell(ranking_indices, num_parent)
[x] Fix usages of EvolutionStrategyEmitter to adapt to the new constructor.
[x] Add parameterized tests for new Optimizers.
[x] Test on rastrigin

Questions

[x] We might want to rename OptimizerBase to EvolutionStrategyBase because some optimizers, such as AdamOpt, are not evolution strategies and also do not inherit from OptimizerBase.
- Yes
[x] Can we put threadpool limit in OptimizerBase on top of abstract methods?
- No

Status

[x] I have read the guidelines in CONTRIBUTING.md
[x] I have formatted my code using yapf
[x] I have tested my code by running pytest
[x] I have linted my code with pylint
[x] I have added a one-line description of my change to the changelog in HISTORY.md
[x] This PR is ready to go