这是indexloc提供的服务,不要输入任何密码
Skip to content

Bugfix/parallel launcher for linux #1141

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged

Conversation

MischaPanch
Copy link
Collaborator

@MischaPanch MischaPanch commented May 7, 2024

Closes: #1121

For details about changes see additions to changelog

@MischaPanch MischaPanch requested a review from opcode81 May 7, 2024 13:57
@MischaPanch MischaPanch force-pushed the bugfix/parallel-launcher-for-linux branch from a0cd0cb to 208b7bc Compare May 7, 2024 14:12
experiment_runner: Callable[
[Experiment],
InfoStats | None,
] = lambda exp: exp.run().trainer_result,
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Anticipating a question here: we can't return a list of ExperimentResult since it will contain an instance of World which is not picklable - joblib complains.

It would also be unnecessary to return all the worlds, they can be quite heavy.

So, instead we return the trainer_result object, which however can be None (but usually is not)

1. Launch in main process if only 1 exp is passed
2. Launcher returns a list of stats for successful exps
3. More detailed logging for unsuccessful expos
4. Raise error if all runs were unsuccessful
5. DataclassPPrintMixin allows retrieving a pretty repr string
6. Minor improvements in docstrings
@MischaPanch MischaPanch force-pushed the bugfix/parallel-launcher-for-linux branch from 208b7bc to bf3859a Compare May 7, 2024 14:21
@MischaPanch MischaPanch merged commit f0b7abe into thu-ml:master May 8, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Revisit Launcher for starting multiple experiments
2 participants