~wooksong/neural-network-support/+git/tizen-tensorflow:sandbox/gichan/1.13.1_rm_flatbuf_header

Last commit made on 2020-06-12
Get this branch:
git clone -b sandbox/gichan/1.13.1_rm_flatbuf_header https://git.launchpad.net/~wooksong/neural-network-support/+git/tizen-tensorflow

Branch merges

Branch information

Name:
sandbox/gichan/1.13.1_rm_flatbuf_header
Repository:
lp:~wooksong/neural-network-support/+git/tizen-tensorflow

Recent commits

0abe46c... by Gichan Jang

Remove flatbuf header from tflite-devel pkg
Flatbuf-devel and tflite-devel is conflicted.
This PR remove flatbuf header files from tflite-devel pkg.

Signed-off-by: gichan-jang <email address hidden>

2e61ba0... by MyungJoo Ham

[VD/GBM] Suffering from lto again.

TV is suffering from link-time-optimization flag again.
Disable lto for TV build only.

Tizen:Unified and other GBMs will still use lto.

Change-Id: I76a41194ce2691b23c31ad104fad1d1e04cca9b9
Signed-off-by: MyungJoo Ham <email address hidden>

5ab2824... by MyungJoo Ham

[VD/GBM] Remove wreturn-type warning.

Although this is OPEN-P and it is unreachable code,
add workaround for wreturn-type.

Change-Id: Idf11044b927407ae7d3bd241d58d41588073ad96
Signed-off-by: MyungJoo Ham <email address hidden>

7ba7c88... by Inki Dae <email address hidden>

packaging: pack flatbuffer schema file

This patch generates a new package file, tensorflow-lite-flatbuf,
and the schema file is added to the package.
The schema file is used by ARMNN - armnn internally compiles
tensorflow lite parser using the schema file.

Changelog v1:
- Added Requires: tensorflow-lite-devel package
- Updated description

Change-Id: I398254cdd4c032f2110b457c2bdf491172c738cf
Signed-off-by: Inki Dae <email address hidden>

d77e88b... by Dongkyun Son <email address hidden>

Revert "[Dist] Disable -lto option."

This reverts commit 927a791e44738b5ee4173d5794d3f28e4d350af8.

Change-Id: I241ccf1c4e51b1f9608940838063fb95aa8dbff7

01419f8... by Hyoung Joo Ahn

fix the applying CFLAG according to the version of GCC

Change-Id: I904f96f45f27fe5ba45d9a48420ec7a5662a0f3b

6404733... by Hyoung Joo Ahn

add a C/CPP FLAG of the memory access policy

Change-Id: Ia9f9f68110bbe4099338b131a9ff38208e335912
Signed-off-by: HyoungJoo Ahn <email address hidden>

ee17725... by Andrey Kazmin <email address hidden>

Fixed Makefile for benchmark build

Filtered out benchmark_plus_flex_main.cc from benchmark build,
because of its tflite::benchmark::Main conflicts with ones defined
in benchmark_main.cc.
We didn't ever build benchmark_model with Flex op support.
So, we could simply filter out benchmark_plus_flex_main.cc
from benchmark sources.

Signed-off-by: Andrey Kazmin <email address hidden>

927a791... by MyungJoo Ham

[Dist] Disable -lto option.

Some Tizen distributors use -lto, which breaks Tensorflow-Lite.
Override such an option by adding -lno-lto.

Change-Id: Ib3bda7c020be0480d2603b124519a7ebab9904d1
Signed-off-by: MyungJoo Ham <email address hidden>

2e09fad... by MyungJoo Ham

[Dist] Remove tf-normal dependencies.

Enable tf-lite only, remove tf-normal build deps.

Change-Id: I4edcdc644a1ae353215715611a210bdb5b1d44d4
Signed-off-by: MyungJoo Ham <email address hidden>