과거에도 있긴 했지만.. 도대체 이 delegate는 어떤 통로를 통해 tensorflow lite에서 사용이 가능해지는 건진 모르겠다.
[링크 : https://www.tensorflow.org/lite/performance/implementing_delegate]
[링크 : https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/delegates/gpu/README.md]
| 2.7.0-rc with opencl (0) | 2021.10.13 |
|---|---|
| tf release 2.7.0-rc (0) | 2021.10.12 |
| tflite gpu openCL support build fail (0) | 2021.08.31 |
| tf lite cmake (0) | 2021.08.27 |
| tensorflow cross compile (0) | 2021.07.01 |
그냥 ssh id@hostname "commands" 하면 되는 듯
예전에 찾은적이 있던 것 같기도 한데...
[링크 : https://stackoverflow.com/questions/18502945/how-to-execute-a-remote-command-over-ssh-with-arguments]
| sshd 에서 ssh 로그인 끄고 sftp만 허용하기 (0) | 2022.04.01 |
|---|---|
| ssh 압축전송하기 (0) | 2021.12.14 |
| reverse ssh (0) | 2021.01.03 |
| ssh blowfish (0) | 2019.09.24 |
| ssh tunnel proxy (0) | 2019.09.20 |
통신을 주고 받으면서 해당 내용을 파일로 저장하는 기능.
ctrl-a,z 에서 "Capture on/off.....L" 을 눌러 들어가면
+-------------------------------------------------------------------+
| Minicom Command Summary |
| |
| Commands can be called by CTRL-A <key> |
| |
| Main Functions Other Functions |
| |
| Dialing directory..D run script (Go)....G | Clear Screen.......C |
| Send files.........S Receive files......R | cOnfigure Minicom..O |
| comm Parameters....P Add linefeed.......A | Suspend minicom....J |
| Capture on/off.....L Hangup.............H | eXit and reset.....X |
| send break.........F initialize Modem...M | Quit with no reset.Q |
| Terminal settings..T run Kermit.........K | Cursor key mode....I |
| lineWrap on/off....W local Echo on/off..E | Help screen........Z |
| Paste file.........Y Timestamp toggle...N | scroll Back........B |
| Add Carriage Ret...U |
| |
| Select function or press Enter for none. |
+-------------------------------------------------------------------+
저장할 파일명을 물어본다. minicom을 실행한 현재 디렉토리를 기준으로 저장된다.
+-----------------------------------------+
|Capture to which file? |
|> minicom.cap |
+-----------------------------------------+| minicom lf에 cr 붙이기 (0) | 2023.01.05 |
|---|---|
| minicom 16진수로 보기 (0) | 2022.08.25 |
| minicom timestamp (0) | 2021.09.16 |
| minicom 폭 늘리기 (0) | 2021.01.28 |
| /dev/tty 를 sudo 쓰지 않고 사용하기 (0) | 2020.09.24 |
minicom에서 "Timestamp toggle...N" 을 누르면
별다른 티를 내지 않고 그냥 나가버리는데
+-------------------------------------------------------------------+
| Minicom Command Summary |
| |
| Commands can be called by CTRL-A <key> |
| |
| Main Functions Other Functions |
| |
| Dialing directory..D run script (Go)....G | Clear Screen.......C |
| Send files.........S Receive files......R | cOnfigure Minicom..O |
| comm Parameters....P Add linefeed.......A | Suspend minicom....J |
| Capture on/off.....L Hangup.............H | eXit and reset.....X |
| send break.........F initialize Modem...M | Quit with no reset.Q |
| Terminal settings..T run Kermit.........K | Cursor key mode....I |
| lineWrap on/off....W local Echo on/off..E | Help screen........Z |
| Paste file.........Y Timestamp toggle...N | scroll Back........B |
| Add Carriage Ret...U |
| |
| Select function or press Enter for none. |
+-------------------------------------------------------------------+
한번 누를 때 마다 다음과 같은 시간 포맷으로 출력된다.
개인적으로는 ms 단위까지 표현해주는게 좋음.
| U-Boot [2021-09-16 14:23:35] U-Boot [2021-09-16 14:23:55.603] U-Boot |
[링크 : https://stackoverflow.com/questions/54377832/extended-timestamp-as-default-in-minicom]
[링크 : https://askubuntu.com/questions/905655/not-able-to-add-timestamp-to-minicom-readings]
| minicom lf에 cr 붙이기 (0) | 2023.01.05 |
|---|---|
| minicom 16진수로 보기 (0) | 2022.08.25 |
| minicom 로그 저장하기 (0) | 2021.09.16 |
| minicom 폭 늘리기 (0) | 2021.01.28 |
| /dev/tty 를 sudo 쓰지 않고 사용하기 (0) | 2020.09.24 |
마지막 커밋 1개 날리고, 강제로 서버에 푸시
| $ git reset --hard HEAD~1 $ git push -f |
그래도 이력을 남기는 revert가 좋을 듯.
다만 돌릴게 많으면 돌렸다는 이력 자체도 잔뜩 생긴다고 하니
되돌릴일을 만들지 않는게 최선.. ㅠㅠ
| $ git revert [hash] |
[링크 : https://jupiny.com/2019/03/19/revert-commits-in-remote-repository/]
| git submodule ... 2? (0) | 2024.06.19 |
|---|---|
| git diff --staged (0) | 2022.09.05 |
| git blame (0) | 2021.06.21 |
| git pull rebase 설정 (0) | 2021.06.02 |
| git log --stat (0) | 2021.05.10 |
| [ 57%] Building C object _deps/xnnpack-build/CMakeFiles/XNNPACK.dir/src/qs8-gavgpool/gen/7p7x-minmax-neon-c8-acc2.c.o /home/pi/work/tflite_build/xnnpack/src/qs8-gavgpool/gen/7x-minmax-neon-c16-acc2.c: In function ‘xnn_qs8_gavgpool_minmax_ukernel_7x__neon_c16_acc2’: /home/pi/work/tflite_build/xnnpack/src/qs8-gavgpool/gen/7x-minmax-neon-c16-acc2.c:174:5: note: use -flax-vector-conversions to permit conversions between vectors with differing element types or numbers of subparts int8x16_t vout0123456789ABCDEF = vcombine_s16(vqmovn_s16(vacc01234567), vqmovn_s16(vacc89ABCDEF)); ^~~~~~~~~ /home/pi/work/tflite_build/xnnpack/src/qs8-gavgpool/gen/7x-minmax-neon-c16-acc2.c:174:51: error: incompatible type for argument 1 of ‘vcombine_s16’ int8x16_t vout0123456789ABCDEF = vcombine_s16(vqmovn_s16(vacc01234567), vqmovn_s16(vacc89ABCDEF)); |
[링크 : https://www.gitmemory.com/issue/tensorflow/tensorflow/47536/809003416]
DNN을 빼고 해보면 된다는데
| $ cmake ../tensorflow/tensorflow/lite -DTFLITE_ENABLE_GPU=ON -DTFLITE_ENABLE_XNNPACK=OFF |
안되잖아!?!
| [ 41%] Building CXX object CMakeFiles/tensorflow-lite.dir/delegates/gpu/api.cc.o In file included from /home/pi/work/tensorflow/tensorflow/lite/delegates/gpu/common/util.h:19, from /home/pi/work/tensorflow/tensorflow/lite/delegates/gpu/api.h:45, from /home/pi/work/tensorflow/tensorflow/lite/delegates/gpu/api.cc:16: /home/pi/work/tensorflow/tensorflow/lite/delegates/gpu/common/types.h:23:10: fatal error: fp16.h: No such file or directory #include <fp16.h> ^~~~~~~~ compilation terminated. make[2]: *** [CMakeFiles/tensorflow-lite.dir/build.make:188: CMakeFiles/tensorflow-lite.dir/delegates/gpu/api.cc.o] Error 1 make[2]: *** Waiting for unfinished jobs.... make[1]: *** [CMakeFiles/Makefile2:1751: CMakeFiles/tensorflow-lite.dir/all] Error 2 make: *** [Makefile:156: all] Error 2 |
+
| ARMCC_FLAGS="-march=armv7-a -mfpu=neon-vfpv4 -funsafe-math-optimizations" cmake -DCMAKE_C_COMPILER=${ARMCC_PREFIX}gcc \ -DCMAKE_CXX_COMPILER=${ARMCC_PREFIX}g++ \ -DCMAKE_C_FLAGS="${ARMCC_FLAGS}" \ -DCMAKE_CXX_FLAGS="${ARMCC_FLAGS}" \ -DCMAKE_VERBOSE_MAKEFILE:BOOL=ON \ -DCMAKE_SYSTEM_NAME=Linux \ -DCMAKE_SYSTEM_PROCESSOR=armv7 \ -DTFLITE_ENABLE_GPU=ON -DTFLITE_ENABLE_XNNPACK=OFF \ ../tensorflow/tensorflow/lite |
아래는 크로스컴파일 하는 예
| $ curl -LO https://storage.googleapis.com/mirror.tensorflow.org/developer.arm.com/media/Files/downloads/gnu-a/8.3-2019.03/binrel/gcc-arm-8.3-2019.03-x86_64-arm-linux-gnueabihf.tar.xz $ mkdir -p ${HOME}/toolchains $ tar xvf gcc-arm-8.3-2019.03-x86_64-arm-linux-gnueabihf.tar.xz -C ${HOME}/toolchains $ ARMCC_FLAGS="-march=armv7-a -mfpu=neon-vfpv4 -funsafe-math-optimizations" ARMCC_PREFIX=${HOME}/toolchains/gcc-arm-8.3-2019.03-x86_64-arm-linux-gnueabihf/bin/arm-linux-gnueabihf- cmake -DCMAKE_C_COMPILER=${ARMCC_PREFIX}gcc \ -DCMAKE_CXX_COMPILER=${ARMCC_PREFIX}g++ \ -DCMAKE_C_FLAGS="${ARMCC_FLAGS}" \ -DCMAKE_CXX_FLAGS="${ARMCC_FLAGS}" \ -DCMAKE_VERBOSE_MAKEFILE:BOOL=ON \ -DCMAKE_SYSTEM_NAME=Linux \ -DCMAKE_SYSTEM_PROCESSOR=armv7 \ ../tensorflow/lite/ |
[링크 : https://www.tensorflow.org/lite/guide/build_cmake_arm]
| tf release 2.7.0-rc (0) | 2021.10.12 |
|---|---|
| tflite delegate (0) | 2021.10.11 |
| tf lite cmake (0) | 2021.08.27 |
| tensorflow cross compile (0) | 2021.07.01 |
| tensorflow lite on rpi4 64bit (0) | 2021.06.30 |
2.4.1 을 사용했었는데 어느새 2.6.0이 나왔네
| $ git tag v2.4.1 v2.4.2 v2.4.3 v2.5.0 v2.5.0-rc0 v2.5.0-rc1 v2.5.0-rc2 v2.5.0-rc3 v2.5.1 v2.6.0 v2.6.0-rc0 v2.6.0-rc1 v2.6.0-rc2 |
18.04에다가 쓰다 보니 cmake 버전이 낮으셔서 빌드도 못한다고 -_ㅠ
| ~/work/tflite_build$ cmake ../tensorflow/tensorflow/lite CMake Error at CMakeLists.txt:29 (cmake_minimum_required): CMake 3.16 or higher is required. You are running version 3.10.2 -- Configuring incomplete, errors occurred! |
2.4.1에서도 TFLITE_ENABLE_GPU가 있나 찾아 볼까?
| $ cmake ../tensorflow_src/tensorflow/lite -DTFLITE_ENABLE_GPU=ON |
| tflite delegate (0) | 2021.10.11 |
|---|---|
| tflite gpu openCL support build fail (0) | 2021.08.31 |
| tensorflow cross compile (0) | 2021.07.01 |
| tensorflow lite on rpi4 64bit (0) | 2021.06.30 |
| tensorflow delegate (0) | 2021.06.30 |
mariadb를 설치했더니 이런게 설치되서 찾아보는데..
mariadb / mssql / postgresql 를 지원한다!!
mariadb / postgresql은 ssh tunnel을 지원한다니 끌리네?
opensource / GPL
[링크 : https://www.heidisql.com/]
[링크 : https://www.heidisql.com/help.php#license]
| mariadb 초기설정 (0) | 2022.08.30 |
|---|---|
| mariadb c# connector (0) | 2021.10.22 |
| sql zerofill (0) | 2019.11.25 |
| mysql-dump compatible 함정 -_- (0) | 2019.09.04 |
| mysql / unix-socket (0) | 2019.09.04 |
요즘 유뷰트에서 자주 눈에 띄던데, 연관해서 광고를 좀 띄우는건가?
아무튼.. 블렌더로 먼가 만들어 보고 싶긴했는데 가격이 아주 싼건 아니니 고민이네
[링크 : https://coloso.co.kr/creative/gamegraphic_shinminjeong]
| blender on macos (0) | 2022.05.24 |
|---|---|
| big bunny blender project file (0) | 2022.05.22 |
| blender 포켓몬 애니메이션 (0) | 2020.08.08 |
| blender physical simulation (0) | 2018.03.19 |
| blender netrender 잠시 된건가.. (0) | 2016.11.23 |
카메라 8개를 하나의 모니터에 분할해서 쏘는 예제?
[링크 : http://trac.gateworks.com/wiki/Yocto/gstreamer/compositing]
| gstreamer를 이용하여 uvc 웹캠 (mjpg) 보기 (0) | 2023.03.22 |
|---|---|
| gstream videomixer (0) | 2023.02.17 |
| gst h264 filesink (0) | 2021.07.14 |
| gstreamer 관련 (0) | 2021.07.13 |
| gstreamer 용어 (0) | 2021.07.13 |
|
