projects / platform / core / ml / nnfw.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 62a62ab) | patch
[locomotiv] Introduce TensorConstantPad (#8027)
author채성우/On-Device Lab(SR)/Engineer/삼성전자 <sw4670.chae@samsung.com>
Thu, 17 Oct 2019 10:04:16 +0000 (19:04 +0900)
committer박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 <jh1302.park@samsung.com>
Thu, 17 Oct 2019 10:04:16 +0000 (19:04 +0900)
commitffd065f1802a9e4ec6f85013f0cee0627bee4d68
tree9c144d39ca47516242ddf6830f2ee72f72d42375tree | snapshot
parent62a62abc1f1aa85cc1ad6b206d6635441cd55fb9commit | diff
[locomotiv] Introduce TensorConstantPad (#8027)

* [locomotiv] Introduce TensorConstantPad

This commit introduces TensorConstantPad to locomotiv.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* change algorithm of pad interpretation.

* make constant value more general.

* change assert to validate.

* modify test.

* add assert for pad_data.

* constant input can be all operator.

* fix some wrong valiate and message.
compiler/locomotiv/src/Node.lst diff | blob | history
compiler/locomotiv/src/Node/TensorConstantPad.cpp [new file with mode: 0644] blob
compiler/locomotiv/src/Node/TensorConstantPad.test.cpp [new file with mode: 0644] blob
Domain: Machine Learning / ML Framework;