Add ML Single API support of NNStreamer 41/234941/1
authorInki Dae <inki.dae@samsung.com>
Mon, 1 Jun 2020 02:16:53 +0000 (11:16 +0900)
committerInki Dae <inki.dae@samsung.com>
Mon, 1 Jun 2020 02:33:31 +0000 (11:33 +0900)
commit04a0566ec72d7536c9bc1e74475876614aaa3ab9
tree9466cff9849e42798643561c1eaa3526270a881c
parent5f5dc3e5a37f1623008a8aa338fd6cc79850feab
Add ML Single API support of NNStreamer

This patch adds ML Single API support of NNStreamer.

On MediaVision, NNFW(in-house NN Runtime) and Vivante NPU
will be controlled by inference-engine-mlapi backend
through ML Single API of NNStreamer.

As one more backends should be controlled by the inference-engine-mlapi
backend, this patch introduces a new function, SetPluginType,
which lets ML Single API know which tensor filter - NNFW or Vivante NPU -
should be used, and adds two backend types - Vivante and NNFW.

Change-Id: Iccf9fe97f5ec50c5c1bfd7bab48927e0f343f182
Signed-off-by: Inki Dae <inki.dae@samsung.com>
include/inference_engine_common.h
include/inference_engine_common_impl.h
include/inference_engine_type.h
src/inference_engine_common_impl.cpp
test/src/inference_engine_profiler.cpp
test/src/inference_engine_tc.cpp