Remove default std normalization value in base class #2630
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
adamjstewart
added
the
backwards-incompatible
adamjstewart
requested changes
Mar 5, 2025
There was a problem hiding this comment.
I agree with this change (principle of least surprise) but we should set std for all the datamodules that don't currently specify it. I'm fine with just using 255 in all of them to maintain backwards compatibility but it would be good to at least check data.py and see what dtype the files use.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
While not technically a bug, I think this is quiet inconvenient, when building custom datamodules inheriting from torchgeo base classes. For my use case, I had trouble understanding why my data was all of the sudden squashed to a tiny range when I had properly normalized it beforehand.
I think the default value here should be 1, i.e nothing happens, and users have to otherwise overwrite it. At the moment it is a base class for RGB data, which is oddly specific, and I think it would be better to keep it more general, especially because it might not be clear to all users, that the normalization "secretly" happens in the on_after_batch_transfer hook.
Closes #1841