[fix] torch.repeat : dim-0 backward #45212
Conversation
|
@albanD Please review :) |
💊 CI failures summary and remediationsAs of commit e55e126 (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group. This comment has been revised 5 times. |
|
@pytorchbot rebase this please |
| } | ||
|
|
||
| Tensor repeat_backward(Tensor grad, int64_t input_dims, IntArrayRef repeats) { | ||
| Tensor repeat_backward(Tensor grad, int64_t input_dims, IntArrayRef repeats, IntArrayRef input_shape) { |
There was a problem hiding this comment.
nit: you can remove the input_dims argument now that you have all the shapes.
test/test_autograd.py
Outdated
| c.backward() | ||
| self.assertEqual(b.grad, torch.tensor([-inf, 0., 0.])) | ||
|
|
||
| def test_repeat_zero_dim(self): |
There was a problem hiding this comment.
You don't have to do a custom test here. You can simply add a new entry here that specifies the input size and the arguments.
So something like ('repeat', (2), (0,), 'zero_repeat'), will do the same thing.
|
Thanks for sending a PR! |
Codecov Report
@@ Coverage Diff @@
## master #45212 +/- ##
==========================================
- Coverage 68.01% 68.01% -0.01%
==========================================
Files 393 393
Lines 50847 50847
==========================================
- Hits 34583 34582 -1
- Misses 16264 16265 +1
Continue to review full report at Codecov.
|
|
@albanD Have addressed the comments. PTAL:) |
albanD
left a comment
albanD
left a comment
There was a problem hiding this comment.
Perfect! Thanks for the update!
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@albanD has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
5 participants
Fixes #45201