Skip to content
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

Provide default progress loggers and use new ProgressLogging API #184

Merged
merged 1 commit into from
Mar 14, 2020

Conversation

devmotion
Copy link
Member

This PR fixes #158 and adds support for the new (unreleased) support of default progress loggers in DiffEqBase to DiffEqFlux. In my local tests, progress bars were displayed in the terminal successfully (but maybe you want to disable the progress bars in the CI tests since it seems to fill up the logs IIRC).

As mentioned by @tkf in #158 (comment), the problem with the current approach of setting up the progress logs (and also in #183) is that the progress bar can't distinguish between a crashed and a slow progress. To get this right, the PR makes use of ProgressLogging.@withprogress and the new ProgressLogging API. In this way also a UUID is used as id automatically. However, since ProgressLogging.@withprogress automatically initializes the progress bar before the first manual progress update (in https://github.com/JunoLab/ProgressLogging.jl/blob/b64333df0f4d2e780387af01867cccb7460da33d/src/ProgressLogging.jl#L326), a macro helps to use it only if progress output is desired. Otherwise, a progress bar might be initialized and finalized after the computations even though the user specified progress = false.

@devmotion
Copy link
Member Author

It seems Travis tests fail due to the same error as on master.

@devmotion
Copy link
Member Author

Should the progress bars be disabled in the tests?

@ChrisRackauckas
Copy link
Member

The error on master: @pkofod did you forget to tag Optim?

@ChrisRackauckas ChrisRackauckas merged commit de5ff45 into SciML:master Mar 14, 2020
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.

Setup progress bar of sciml_train to be platform agnostic
2 participants