[mir/interpreter] Quantized operations implementation (#8317)
authorPavel Iliutchenko/AI Tools Lab /SRR/Engineer/Samsung Electronics <p.iliutchenk@samsung.com>
Mon, 28 Oct 2019 11:56:21 +0000 (14:56 +0300)
committerAlexander Efimov/./AI Tools Lab/Samsung Electronics <a.efimov@samsung.com>
Mon, 28 Oct 2019 11:56:21 +0000 (14:56 +0300)
commit75b6a4b05ad1c1df0d8d23c19032080c703a457b
treeae0786b87966fa85602ad294287e2b7160b7a406
parent069b708c92276783d7745cae33a11008871b5171
[mir/interpreter] Quantized operations implementation (#8317)

* Supported quantized operations: Add, Conv2D, DepthwiseConv2D
* Fixed interpreter for quantizations

Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
compiler/nnc/backends/interpreter/Interpreter.cpp
compiler/nnc/backends/interpreter/ops/Add.h
compiler/nnc/backends/interpreter/ops/AvgPool2D.cpp
compiler/nnc/backends/interpreter/ops/AvgPool2D.h
compiler/nnc/backends/interpreter/ops/Conv2D.cpp
compiler/nnc/backends/interpreter/ops/Conv2D.h
compiler/nnc/backends/interpreter/ops/DepthwiseConv2D.cpp
compiler/nnc/backends/interpreter/ops/DepthwiseConv2D.h
compiler/nnc/backends/interpreter/ops/QuantizationHelpers.cpp [new file with mode: 0644]
compiler/nnc/backends/interpreter/ops/QuantizationHelpers.h [new file with mode: 0644]