new version of Heri's PR

facebookresearch / nevergrad

A Python toolbox for performing gradient-free optimization

https://facebookresearch.github.io/nevergrad/

MIT License

3.93k stars 352 forks source link

new version of Heri's PR #1389

Open teytaud opened 2 years ago

teytaud commented 2 years ago

Types of changes

[ ] Docs change / refactoring / dependency upgrade
[ ] Bug fix (non-breaking change which fixes an issue)
[x] New feature (non-breaking change which adds functionality)
[ ] Breaking change (fix or feature that would cause existing functionality to change)

Motivation and Context / Related issue

Original PR looks dead: https://github.com/facebookresearch/nevergrad/pull/1060

How Has This Been Tested (if it applies)

Checklist

[x] The documentation is up-to-date with the changes I made.
[x] I have read the CONTRIBUTING document and completed the CLA (see CLA).
[x] All tests passed, and additional code has been covered with new tests.

teytaud commented 2 years ago

@herilalaina I would be sad if this was never finished so I did this :-)

teytaud commented 2 years ago

@jrapin not sure if Heri can be back and answer some of your questions... I do my best for parts for which I know what to do.

teytaud commented 2 years ago

Groumpf I believe it was just fine now but baoum! a big static analysis issue. No idea what is causing this.

teytaud commented 2 years ago

Groumpf I believe it was just fine now but baoum! a big static analysis issue. No idea what is causing this.

I guess we must stay at mypy 0.931 for the moment as more recent versions fail.

teytaud commented 2 years ago

Groumpf I believe it was just fine now but baoum! a big static analysis issue. No idea what is causing this.

I guess we must stay at mypy 0.931 for the moment as more recent versions fail.

Changing to 0.931 does not fix it @#$%&^@

herilalaina commented 2 years ago

@herilalaina I would be sad if this was never finished so I did this :-)

Thank you very much for taking this up!

I guess we must stay at mypy 0.931 for the moment as more recent versions fail.

I also encountered such a conflict of versions on my own. @jrapin would it make sense to run the evaluation part of the code on a separate container (e.g. Singularity) as in HPOBench ?