They keep track of LearningParameters and the corresponding accuracies.
This allows you to calculate train, validation and test accuracy on datasets, and store this info to a file.
This makes it easier to keep track of the sets of hyperparameters which you tried before, how good they were, and how to change them in the next run.
I also tested the json instances using genvalidity-hspec.
I can merge this branch with the branch 'json' if you want to merge the PRs separately.
They keep track of LearningParameters and the corresponding accuracies. This allows you to calculate train, validation and test accuracy on datasets, and store this info to a file. This makes it easier to keep track of the sets of hyperparameters which you tried before, how good they were, and how to change them in the next run.
I also tested the json instances using genvalidity-hspec.
I can merge this branch with the branch 'json' if you want to merge the PRs separately.