[MO] StridedSlice improvements#4139
Conversation
8f9b122 to
a19dda7
Compare
…cherry-picked maxpoolV4 and tf_broadcast_ext)
…w rewritten SS infer
f6b646a to
757432a
Compare
757432a to
bb2fdb1
Compare
model-optimizer/mo/utils/ir_reader/extenders/strided_slice_extender.py
Outdated
Show resolved
Hide resolved
lazarevevgeny
left a comment
There was a problem hiding this comment.
In general looks good, but still have some comments.
Also, please, run unit test with code coverage to see how new functions are covered with unit tests: coverage run -m unittest discover -p "*_test.py"
model-optimizer/mo/utils/ir_reader/extenders/strided_slice_extender.py
Outdated
Show resolved
Hide resolved
|
@lazarevevgeny thanks for your comments! Resolved them all. There are 3 sets of automatically unit-tests that cover almost all cases. The same ShapeInfer tests after Normalizer Permute tests dl-benchmark finished without regressions on a previous run. @jane-intel, @sadolini, @rkazants please review |
e-nugmanova
left a comment
There was a problem hiding this comment.
First part 14/20 files
…ected comments and other minor corrections
lazarevevgeny
left a comment
There was a problem hiding this comment.
Overall looks good. But left several minor comments to be fixed.
|
According to the documentation, the Is it a bug or the |
Specification haven't been changed. |
Description: Fixed some issues with StridedSlice permutation and processing ellipsis masks. Now infer function do not change attributes and inputs of StridedSlice, moved normalization into a separate transformation, also reused general existing permute mechanism. Now complex StridedSlices are correctly inferred in ShapeOf subgraphs.
tickets 40372, 43072, 39311
Code:
Validation:
Documentation: