projects / platform / upstream / dldt.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: cfb6fae) | patch
[IE][VPU]: Fixes addCopyForOutputsInsideNetwork (#2393)
authorGladilov, Gleb <gleb.gladilov@intel.com>
Thu, 24 Sep 2020 09:40:10 +0000 (12:40 +0300)
committerGitHub <noreply@github.com>
Thu, 24 Sep 2020 09:40:10 +0000 (12:40 +0300)
commit10df5907b5c245a603187c60cea385060fc00bf3
treebd4b801f4d378580720887ab231d1b2c53af8726tree | snapshot
parentcfb6fae2c99270b531e4a24b568db3c75f6b6539commit | diff
[IE][VPU]: Fixes addCopyForOutputsInsideNetwork (#2393)

* [IE][VPU]: Fixes addCopyForOutputsInsideNetwork

In case of dynamic output with consumer pass tries
to connect output's shape with new intermediate data
twice: one at the moment of duplicateData call (successful)
and once more at the end of the pass manually. The second
try leads to error since child data is already connected.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU]: Introduces tests on addCopyForOutputsInsideNetwork

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
inference-engine/src/vpu/graph_transformer/src/middleend/passes/add_copy_for_outputs_inside_network.cpp diff | blob | history
inference-engine/tests/unit/vpu/base/graph_transformer_tests.cpp diff | blob | history
inference-engine/tests/unit/vpu/base/graph_transformer_tests.hpp diff | blob | history
inference-engine/tests/unit/vpu/middleend_tests/passes_tests/add_copy_for_outputs_inside_network.cpp [new file with mode: 0644] blob
Domain: Machine Learning / ML Framework;