projects / platform / upstream / armnn.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: ba563c6) | patch
IVGCVSW-2268 Remove the input swizzling from ParseConcat
authorMatteo Martincigh <matteo.martincigh@arm.com>
Thu, 6 Dec 2018 12:03:17 +0000 (12:03 +0000)
committerMatteo Martincigh <matteo.martincigh@arm.com>
Fri, 7 Dec 2018 16:25:45 +0000 (16:25 +0000)
commitf9afc791662f9ffd639a9500de8c4e33394c8c39
tree26b3203c189f0a0f280e8fab4fd081f2d6112da3tree | snapshot
parentba563c6d81bfb20f01b1b54e27fe1ac4a494ece1commit | diff
IVGCVSW-2268 Remove the input swizzling from ParseConcat

 * Removed the input swizzling when the concatenation dimension is 3
   in ParseConcat in the TF parser
 * No longer using the helper ProcessConcatInputTensorInfo, where
   the input was being swizzled if the concatenation dimension was 3
 * Added a new convenience constuctor to TensorShape that initializes
   a shape to all zeros given only the number of dimensions

Change-Id: I82a207e41bddc5fea21a0b5a38eafa24ad75d1c2
include/armnn/Tensor.hpp diff | blob | history
src/armnn/Tensor.cpp diff | blob | history
src/armnnTfParser/TfParser.cpp diff | blob | history
Domain: Machine Learning / ML Framework;