Fix inconsistent tensor recreation in TensorList #4286
Merged
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Outdated suggestions cannot be applied.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.
Category: Bug fix
Description:
Use the same shape for the internal tensor that gets back into the scope when the size of the TensorList is changed. empty_shape can be only used for non-zero dimensional samples, as 0d samples can't have 0 volume.
The code uses an explicit TensorShape<0> converted to one of dynamic dimension, as the brace default constructed TensorShape<> had size 1 for some reason.
This clears up potential asserts that could hit in debug builds: specifically calling empty_shape for dim = 0.
Signed-off-by: Krzysztof Lecki klecki@nvidia.com
Additional information:
Affected modules and functionalities:
TensorList
Key points relevant for the review:
Tests:
Checklist
Documentation
DALI team only
Requirements
REQ IDs: N/A
JIRA TASK: N/A