-
Notifications
You must be signed in to change notification settings - Fork 24
Fix cnn training notebook #163
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
Fix cnn training notebook #163
Conversation
This commit is meant to fix deserialisation issues that appeared in when training a model using multiple processes in the CNN notebook.
Pull Request Test Coverage Report for Build 17197699384Details
💛 - Coveralls |
|
I'm getting an exception running the notebook, let's discuss tomorrow or early next week? |
- fix comment about how to not use GPU - increase refresh rate of rich bar to avoid warning messages
|
@tennlee I think the PR is ready. The last changes are:
I have not fixed the partial download issue but made an issue for this one (#165) that I will address shortly (this week). I don't think that would block this PR. Given that it is more convenient with a GPU, but runnable without do you still want it moved to the low-hardware requirements section? |
|
This has a conflict with develop/Gallery, can you please rebase and update? |
|
@tennlee sorry I am not too comfortable rebasing branches once they have been pushed to publicly accessible places, I hope you don't mind the merge instead. |
This PR replace the data accessor of the CNN model training notebook with a WeatherBench2 ERA5 data accessor, and fixes related issues.