Fixed histc return type for CUDA #20369
Conversation
pytorchbot
added
module: cuda
| expected = torch.zeros(100, dtype=torch.float, device=device) | ||
| expected.data[0] = 1 | ||
| expected.data[99] = 1 | ||
| self.assertEqual(expected, actual) |
There was a problem hiding this comment.
I'm confused: the original code tested CUDA tensors by sending them through this path. Does self.assertEqual ignore dtype when comparing tensors?
There was a problem hiding this comment.
apparently it does. i just tried this:
self.assertEqual(torch.tensor([1], dtype=torch.int), torch.tensor([1], dtype=torch.double)) and it didn't fail. I will create an issue for this cause its really concerning.
There was a problem hiding this comment.
Thanks for checking that, that is concerning.
Could you add another test to check if the dtype of the output is correct, then?
izdeby
changed the title
izdeby
changed the title
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@izdeby has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@izdeby has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
Summary: Fixing reported [issue](pytorch/pytorch#20208). Pull Request resolved: pytorch/pytorch#20369 Reviewed By: zou3519 Differential Revision: D15300959 Pulled By: izdeby fbshipit-source-id: 219692f99a66ea433112dfc226132eb6867122cf
6 participants
Fixing reported issue.