extend equalize to all integer and floating dtypes #6851
Merged
Conversation
This file contains hidden or 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
pmeier
commented
Oct 27, 2022
| ): | ||
| image_loader = ImageLoader( | ||
| fn, shape=(get_num_channels(color_space), *spatial_size), dtype=torch.uint8, color_space=color_space | ||
| fn, shape=(get_num_channels(color_space), *spatial_size), dtype=dtype, color_space=color_space |
There was a problem hiding this comment.
Although there is a quite a diff in this file, the only functional change is to also test torch.float32 next to torch.uint8. The rest is just refactoring to account for that.
datumbox
approved these changes
Oct 27, 2022
facebook-github-bot
pushed a commit
that referenced
this pull request
Oct 31, 2022
Summary: * extend equalize to all integer and floating dtypes * address nits Reviewed By: datumbox Differential Revision: D40851020 fbshipit-source-id: 63b36f02e630b9c230431527b359876fedc52f3e
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Outdated suggestions cannot be applied.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.
Addresses #6840 for
equalize. Unfortunately, I need to walk back on my comment #6847 (comment): although it is possible to extend the algorithm to the other integer dtypes, it is unfeasible in practice since it blows up memory for anything larger thantorch.int16. Thus, I opted to convert totorch.uint8unconditionally. Fortunately, that also eliminates the decision which integer dtype we convert floating point inputs to and whether we expose this choice to the user.Since there is no change to the algorithm, I didn't benchmark this change. The only thing the benchmark will measure is how fast the conversion to and from
torch.uint8happens. If you want to see some numbers, please let me know what exactly.cc @vfdev-5 @datumbox @bjuncek