-
Notifications
You must be signed in to change notification settings - Fork 1
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
Refactor ann and unet #37
Conversation
…all folders, ann, attention, and TL
… Load one input checkpoints, infer, return output.
…n UNet. Not implemented for TL so far.
… Attention UNets, and Transfer Learning
…ning. Also made modifications related to logging and argparsing as suggested by Tom
…nd anns. Need to resolve the argparsing cyclic dependence still.
…d inference files for both types of models and for inference for both ERA5 and IFS.
…tory: iccs_coupling_checkpoints
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@TomMelt Yes, this looks good. These files have now been moved to ../utils/ so they can be deleted. I somehow missed these during the last commit.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This PR is ready to go. It has greatly simplified the code maintenance and it has been tested with the mwe branch for fortran and python
This is an extension of PR #34 and #30
No work has been done wrt #30 yet. However, all suggestions in #34 have now been implemented for both era5_training and ifs_transfer_learning.
Another folder era5_training is created to replace the two folders ann_cnn_training and attention_net. All the supporting function definitions have been moved to the newly created utils folder. The code is not fully ready yet as I am still working on merging the training for both the models into a single file.