I want to convert from onnxt to tensorrt plain like mentioned this https://devblogs.nvidia.com/nvidia-serves-deep-learning-inference/ in the section ‘ONNX Models’. I tried install docker from this article https://github.com/onnx/onnx-tensorrt
cp /path/to/TensorRT-3.0.*.tar.gz .
docker build -t onnx_tensorrt .
I download TensorRT-3.0.*.tar.gz . from this source https://developer.nvidia.com/nvidia-tensorrt-download
but during installation I get follow error
CMakeFiles/nvonnxparser_plugin.dir/build.make:77: recipe for target 'CMakeFiles/nvonnxparser_plugin.dir/nvonnxparser_plugin_generated_Split.cu.o' failed
nvcc fatal : Unsupported gpu architecture 'compute_75'
CMake Error at nvonnxparser_plugin_generated_Split.cu.o.cmake:207 (message):
Error generating
/opt/onnx2trt/build/CMakeFiles/nvonnxparser_plugin.dir//./nvonnxparser_plugin_generated_Split.cu.o
make[2]: *** [CMakeFiles/nvonnxparser_plugin.dir/nvonnxparser_plugin_generated_ResizeNearest.cu.o] Error 1
make[2]: *** [CMakeFiles/nvonnxparser_plugin.dir/nvonnxparser_plugin_generated_Split.cu.o] Error 1
CMakeFiles/Makefile2:180: recipe for target 'CMakeFiles/nvonnxparser_plugin.dir/all' failed
make[1]: *** [CMakeFiles/nvonnxparser_plugin.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
Makefile:149: recipe for target 'all' failed
make: *** [all] Error 2
The command '/bin/sh -c rm -rf build/ && mkdir build && cd build && cmake .. && make -j$(nproc) && make install && ldconfig && cd .. && python setup.py build && python setup.py install && rm -rf ./build/' returned a non-zero code: 2
P.S. I have this error on macos and ubuntu 16.04
y.glushenkov@ml-test-env:/tmp/models_example$ nvidia-smi
Fri Jan 18 07:43:02 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.78 Driver Version: 410.78 CUDA Version: 10.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 108... Off | 00000000:00:05.0 Off | N/A |
| 35% 55C P2 76W / 250W | 7344MiB / 11178MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 7293 C trtserver 841MiB |
| 0 18379 C /home/a.eryomin/anaconda3/bin/python 6493MiB |
+-----------------------------------------------------------------------------+
P.P.S gpu info mentioned above from ubuntu
y.glushenkov@ml-test-env:/tmp/models_example$ cat /etc/os-release
NAME="Ubuntu"
VERSION="16.04.5 LTS (Xenial Xerus)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 16.04.5 LTS"
VERSION_ID="16.04"
HOME_URL="http://www.ubuntu.com/"
SUPPORT_URL="http://help.ubuntu.com/"
BUG_REPORT_URL="http://bugs.launchpad.net/ubuntu/"
VERSION_CODENAME=xenial
UBUNTU_CODENAME=xenial
And also I have the same error on another server
/y.glushenkov@ml-dev-env:~$ cat /etc/os-release
NAME="Ubuntu"
VERSION="16.04.5 LTS (Xenial Xerus)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 16.04.5 LTS"
VERSION_ID="16.04"
HOME_URL="http://www.ubuntu.com/"
SUPPORT_URL="http://help.ubuntu.com/"
BUG_REPORT_URL="http://bugs.launchpad.net/ubuntu/"
VERSION_CODENAME=xenial
UBUNTU_CODENAME=xenial
y.glushenkov@ml-dev-env:~$ nvidia-smi
Fri Jan 18 08:05:12 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 396.54 Driver Version: 396.54 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 108... Off | 00000000:00:05.0 Off | N/A |
| 26% 47C P2 62W / 250W | 715MiB / 11178MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
| 1 GeForce GTX 108... Off | 00000000:00:06.0 Off | N/A |
| 21% 30C P8 18W / 250W | 1231MiB / 11178MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
| 2 GeForce GTX 108... Off | 00000000:00:07.0 Off | N/A |
| 21% 21C P8 16W / 250W | 7603MiB / 11178MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 31014 C java 701MiB |
| 1 23532 C /home/a.eryomin/anaconda3/bin/python 1221MiB |
| 2 3746 C /home/a.eryomin/anaconda3/bin/python 7593MiB |
+-----------------------------------------------------------------------------+
y.glushenkov@ml-dev-env:~$
On the both servers GPU are 1080TI
I solve this issue by removing “75” (line 120) from CMakeLists.txt :-)
e.g.
list(APPEND GPU_ARCHS
35
53
61
70
# 75
)
Hi zwzhuang,
now I have next error:)
/usr/bin/ld: warning: libcublas.so.9.1, needed by /usr/lib/gcc/x86_64-linux-gnu/5/../../../x86_64-linux-gnu/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/bin/ld: warning: libcudart.so.9.1, needed by /usr/lib/gcc/x86_64-linux-gnu/5/../../../x86_64-linux-gnu/libnvinfer.so, not found (try using -rpath or -rpath-link)
/usr/bin/ld: onnx2trt: hidden symbol `cudaMemcpyAsync' in /usr/local/cuda/lib64/libcudart_static.a(libcudart_static.a.o) is referenced by DSO
/usr/bin/ld: final link failed: Bad value
collect2: error: ld returned 1 exit status
make[2]: *** [onnx2trt] Error 1
CMakeFiles/onnx2trt.dir/build.make:105: recipe for target 'onnx2trt' failed
CMakeFiles/Makefile2:295: recipe for target 'CMakeFiles/onnx2trt.dir/all' failed
make[1]: *** [CMakeFiles/onnx2trt.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
[100%] Linking CXX shared library libtrt_onnxify.so
[100%] Built target trt_onnxify
Makefile:149: recipe for target 'all' failed
make: *** [all] Error 2
The command '/bin/sh -c rm -rf build/ && mkdir build && cd build && cmake .. && make -j$(nproc) && make install && ldconfig && cd .. && python setup.py build && python setup.py install && rm -rf ./build/' returned a non-zero code: 2
Do you know how can I fixed it?
This issues is still actual.