-
Notifications
You must be signed in to change notification settings - Fork 862
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
[AutoMM] Support customizing image data augmentation #3022
Conversation
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
]: | ||
if not all([isinstance(trans, str) for trans in hyperparameters[k]]): | ||
advanced_hyperparameters[k] = copy.deepcopy(v) | ||
hyperparameters[k] = [] |
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.
After split, the hyperparameter dict itself doesn't aware whether it has been splitted or not. Basically, some advanced hyperparameters are modified (not deleted), and copied to another variable, where the naive hyperparameters don't have access to.
Since hyperparameter space has been divided into subspaces with the dot delimitor, how about moving advanced hyperparameters into 'advanced.xxx' subspace ?? Potentially, with OmegaConf, we can easily access to these advanced hyperparameters with config.advanced
.
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.
Here we split out model.timm_image.train_transforms
and model.timm_image.val_transforms
when their values are lists of callable objects instead of lists of strings. OmegConf is designed to store simple config values, e.g., number or string. Adding config.advanced
still can't handle the hyperprameters with callable objects.
Job PR-3022-b27d45a is done. |
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.
LGTM, and thanks for broadening what users can customize through hyperparameter inputs.
Job PR-3022-a437272 is done. |
Job PR-3022-aad3473 is done. |
Issue #, if available:
Description of changes:
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.