Add error checking for padding modules #106147
Conversation
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/106147
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit c2c0b23: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Fixes #105627 [ghstack-poisoned]
Fixes #105627 [ghstack-poisoned]
Fixes #105627 [ghstack-poisoned]
Fixes #105627 [ghstack-poisoned]
| # Constant Pad 1D with single dimension input | ||
| helper((16), (1, 2), nn.ConstantPad1d) | ||
| # Constant Pad with single dimension input | ||
| helper_functional((16), (1, 2), 'constant') |
There was a problem hiding this comment.
nn.functional.pad allows constant padding for arbitrary dimensions, which is different from the other modes replicate, reflect and circular (see docs), however, nn.ConstantPad{1,2,3}d only allow 2/3D, 3/4D and 4/5D inputs respectively based on the docs (e.g. here for 1D). Changing the cases where input shapes are invalid according the docs to test the functional variant here
…06148) Fixes #105749 #95320 (tldr is that input should always be `[N, C, H, (W, D])` where only H, W and D dimensions get circular padding, so the 2D case where user wants both dimensions to be padded --> they should `.unsqueeze(0)` (as is the case for `Reflection/ReplicationPad`) but we didn't document this for circular padding. [This seems to be the old docstring](https://github.com/pytorch/pytorch/blob/277b05014aced075337d6f84a535bd5150aabebb/torch/nn/functional.py#L4689) that was somehow lost. Fixes no_batch_dim support #104860 - Adds missing documentation for circular padding - Adds missing CircularPad modules - Migrates legacy test_nn tests from circular padding to ModuleInfo - Adds no_batch_dim support + sample inputs that test this Pull Request resolved: #106148 Approved by: https://github.com/albanD ghstack dependencies: #106325, #106147
|
@pytorchbot revert -m "sadly it is breaking internal builds, and I can't coordinate a FF due to timezone differences" -c nosignal |
|
@pytorchbot successfully started a revert job. Check the current status here. |
Reverting PR 106147 failedReason: Comment with id 1661870454 not found Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot revert -m "sadly it is breaking internal builds, and I can't coordinate a FF due to timezone differences" -c nosignal |
|
@pytorchbot successfully started a revert job. Check the current status here. |
|
@mikaylagawarecki your PR has been successfully reverted. |
This reverts commit 0547b62. Reverted #106147 on behalf of https://github.com/jeanschmidt due to sadly it is breaking internal builds, and I can't coordinate a FF due to timezone differences ([comment](#106147 (comment)))
|
@pytorchbot successfully started a revert job. Check the current status here. |
|
@mikaylagawarecki your PR has been successfully reverted. |
|
@jeanschmidt |
…port (#106148)" This reverts commit 87d2536. Reverted #106148 on behalf of https://github.com/malfet due to Reverting as dependent PR #106147 was reverted as well ([comment](#106148 (comment)))
…tch_dim support (#106148)"" Previous one was reverted because the PR stacked under which added error-checking to Pad variants #106147 was reverted as internally some people pass 2D inputs to ZeroPad2d (which should actually take 3d or 4d inputs :) but there wasn't actually anything this PR was breaking according to my understanding [ghstack-poisoned]
…port (#106148)" (#106632) Previous one was reverted because the PR stacked under which added error-checking to Pad variants #106147 was reverted as internally some people pass 2D inputs to ZeroPad2d (which should actually take 3d or 4d inputs :) but there wasn't actually anything this PR was breaking according to my understanding Pull Request resolved: #106632 Approved by: https://github.com/albanD
…port (pytorch#106148)" (pytorch#106632) Previous one was reverted because the PR stacked under which added error-checking to Pad variants pytorch#106147 was reverted as internally some people pass 2D inputs to ZeroPad2d (which should actually take 3d or 4d inputs :) but there wasn't actually anything this PR was breaking according to my understanding Pull Request resolved: pytorch#106632 Approved by: https://github.com/albanD
Fixes #105627
Stack from ghstack (oldest at bottom):