make clamp_bounding_box a kernel / dispatcher hybrid #7227
Merged
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
pmeier
commented
Feb 13, 2023
Comment on lines
+1995
to
+1996
| simple_tensor_loader = TensorLoader( | ||
| fn=lambda shape, dtype, device: bounding_box_loader.fn(shape, dtype, device).as_subclass(torch.Tensor), |
There was a problem hiding this comment.
This is quick and dirty. If we have more such cases in the future, we should have something like an unwrap method or the like to get the plain tensor.
| "`torch.Tensor`, but then the kernel (rightfully) complains that neither `format` nor " | ||
| "`spatial_size` was passed" | ||
| ), | ||
| condition=lambda arg_kwargs: isinstance(arg_kwargs.args[0], BoundingBoxLoader), |
There was a problem hiding this comment.
Note that this only applies to datapoints.BoundingBox inputs. JIT works fine for simple tensor inputs.
Comment on lines
+162
to
+163
| if isinstance(batch, datapoints._datapoint.Datapoint): | ||
| unbatched = type(batch).wrap_like(batch, unbatched) |
There was a problem hiding this comment.
Small fix since this didn't respect datapoints types before. This was not an issue, since this is called from a kernel test and so far all kernels operated only with plain tensors. Meaning, all datapoints would have been unwrapped anyway.
NicolasHug
approved these changes
Feb 13, 2023
vfdev-5
approved these changes
Feb 13, 2023
This was referenced Feb 13, 2023
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 28, 2023
Reviewed By: vmoens Differential Revision: D44416267 fbshipit-source-id: b3688a4c9d767b5e91c71dac77f44dedde65261b
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Outdated suggestions cannot be applied.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 #7215 (comment). If that one is approved, I can send a similar PR for
vision/torchvision/prototype/transforms/functional/_meta.py
Line 189 in 0aed832
cc @vfdev-5 @bjuncek