fix type hints and spelling mistake in generalized_rcnn and poolers #2550
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
….models.detection.generalized_rcnn
pmeier
requested changes
Aug 4, 2020
There was a problem hiding this comment.
Hi @ChiangYintso and thanks for the PR! The test failures are relevant:
test/test_models.py:186: in _test_detection_model
scripted_model = torch.jit.script(model)
env/lib/python3.6/site-packages/torch/jit/_script.py:888: in script
obj, torch.jit._recursive.infer_methods_to_compile
env/lib/python3.6/site-packages/torch/jit/_recursive.py:317: in create_script_module
return create_script_module_impl(nn_module, concrete_type, stubs_fn)
env/lib/python3.6/site-packages/torch/jit/_recursive.py:376: in create_script_module_impl
create_methods_from_stubs(concrete_type, stubs)
env/lib/python3.6/site-packages/torch/jit/_recursive.py:292: in create_methods_from_stubs
concrete_type._create_methods(defs, rcbs, defaults)
env/lib/python3.6/site-packages/torch/jit/_recursive.py:596: in try_compile_fn
return torch.jit.script(fn, _rcb=rcb)
E RuntimeError:
E Unsupported operation: indexing tensor with unsupported index type 'str'. Only ints, slices, lists and tensors are supported:
E File "/root/project/torchvision/models/detection/generalized_rcnn.py", line 17
E def _check_for_degenerate_boxes(targets):
E for target_idx, target in enumerate(targets):
E boxes = target["boxes"]
E ~~~~~~~~~ <--- HERE
E degenerate_boxes = boxes[:, 2:] <= boxes[:, :2]
E if degenerate_boxes.any():
E '_check_for_degenerate_boxes' is being compiled since it was called from 'KeypointRCNN.forward'
E File "/root/project/torchvision/models/detection/generalized_rcnn.py", line 96
E # Check for degenerate boxes
E if targets is not None:
E _check_for_degenerate_boxes(targets)
E ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ <--- HERE
E
E features = self.backbone(images.tensors)
Also you seem to have included some unrelated changes (sorting of the imports, ...). Could you please revert them?
pmeier
approved these changes
Aug 4, 2020
| assert trainable_backbone_layers <= 5 and trainable_backbone_layers >= 0 | ||
| assert 0 <= trainable_backbone_layers <= 5 |
There was a problem hiding this comment.
I understand why you would want to fix this and I'm not opposed to it in general. But this is out of scope for this PR. If you want to make that change we are happy to accept a follow-up PR after this one is merged.
Comment on lines
67
to
68
| "of shape [N, 4], got {:}.".format( | ||
| boxes.shape)) | ||
| "of shape [N, 4], got {:}.".format(boxes.shape)) |
There was a problem hiding this comment.
This seems like formatting only. Please revert.
pmeier
requested changes
Aug 4, 2020
Codecov Report
@@ Coverage Diff @@
## master #2550 +/- ##
=======================================
Coverage 70.73% 70.73%
=======================================
Files 94 94
Lines 8029 8031 +2
Branches 1275 1275
=======================================
+ Hits 5679 5681 +2
Misses 1946 1946
Partials 404 404
Continue to review full report at Codecov.
|
|
Thanks for pointing out the test failures and unrelated changes. The failures may be related to torchscript, finally I think it is unnecessary to move the procedure to a function. |
jiangyinzuo
changed the title
bryant1410
pushed a commit
to bryant1410/vision-1
that referenced
this pull request
Nov 22, 2020
…ytorch#2550) * fix type hints and move degenerate boxes to a function in torchvision.models.detection.generalized_rcnn * format code * format code * changed to static method * revert imports * changed to method * revert procedure for degenerating boxes
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.
fix type hints and spelling mistake in generalized_rcnn and poolers