haotian/duan8
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

add install guide, etc.

Browse Source
This commit is contained in:
wang-xinyu 2020-10-12 11:55:23 +08:00
parent 55e604e6f7
commit 413df88372
4 changed files with 67 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
README.md
View File

@ -24,6 +24,7 @@ All the models are implemented in pytorch or mxnet first, and export a weights f
## Tutorials
- [Install the dependencies.](./tutorials/install.md)
- [A guide for quickly getting started, taking lenet5 as a demo.](./tutorials/getting_started.md)
- [Frequently Asked Questions (FAQ)](./tutorials/faq.md)
- [Migrating from TensorRT 4 to 7](./tutorials/migrating_from_tensorrt_4_to_7.md)

2
tutorials/faq.md
View File

@ -4,6 +4,8 @@
`NvInfer.h` is one of the headers of TensorRT. If you install the tensorrt DEB package, the headers should in `/usr/include/x86_64-linux-gnu/`. If you install tensorrt TAR or ZIP file, the `include_directories` and `link_directories` of tensorrt should be added in `CMakeLists.txt`.
`dpkg -L` can print out the contents of a DEB package.
```
$ dpkg -L libnvinfer-dev
/.

57
tutorials/install.md Normal file
View File

@ -0,0 +1,57 @@
# Install the dependencies of tensorrtx
## Ubuntu
Ubuntu16.04 / cuda10.0 / cudnn7.6.5 / tensorrt7.0.0 / opencv3.3 would be the example, other versions might also work, just need you to try.
It is strongly recommended to use `apt` to manage software in Ubuntu.
### 1. Install CUDA
Go to [cuda-10.0-download](https://developer.nvidia.com/cuda-10.0-download-archive). Choose `Linux` -> `x86_64` -> `Ubuntu` -> `16.04` -> `deb(local)` and download the .deb package.
Then follow the installation instructions.
```
sudo dpkg -i cuda-repo-ubuntu1604-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-<version>/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda
```
### 2. Install TensorRT
Go to [nvidia-tensorrt-7x-download](https://developer.nvidia.com/nvidia-tensorrt-7x-download). You might need login.
Choose TensorRT 7.0 and `TensorRT 7.0.0.11 for Ubuntu 1604 and CUDA 10.0 DEB local repo packages`
Install with following commands, after `apt install tensorrt`, it will automatically install cudnn, nvinfer, nvinfer-plugin, etc.
```
sudo dpkg -i nv-tensorrt-repo-ubuntu1604-cuda10.0-trt7.0.0.11-ga-20191216_1-1_amd64.deb
sudo apt update
sudo apt install tensorrt
```
### 3. Install OpenCV
```
sudo add-apt-repository ppa:timsc/opencv-3.3
sudo apt-get update
sudo apt install libopencv-dev
```
### 4. Check your installation
```
dpkg -l | grep cuda
dpkg -l | grep nvinfer
dpkg -l | grep opencv
```
### 5. Run tensorrtx
It is recommanded to go through the [getting started guide, lenet5 as a demo.](./tutorials/getting_started.md) first.
But if you are proficient in tensorrt, please check the readme of the model you want directly.

17
yolov5/CMakeLists.txt
View File

@ -13,16 +13,13 @@ find_package(CUDA REQUIRED)
set(CUDA_NVCC_PLAGS ${CUDA_NVCC_PLAGS};-std=c++11;-g;-G;-gencode;arch=compute_30;code=sm_30)
include_directories(${PROJECT_SOURCE_DIR}/include)
if (CMAKE_SYSTEM_PROCESSOR MATCHES "aarch64")
message("embed_platform on")
include_directories(/usr/local/cuda/targets/aarch64-linux/include)
link_directories(/usr/local/cuda/targets/aarch64-linux/lib)
else()
message("embed_platform off")
include_directories(/usr/local/cuda/include)
link_directories(/usr/local/cuda/lib64)
endif()
# include and link dirs of cuda and tensorrt, you need adapt them if yours are different
# cuda
include_directories(/usr/local/cuda/include)
link_directories(/usr/local/cuda/lib64)
# tensorrt
include_directories(/usr/include/x86_64-linux-gnu/)
link_directories(/usr/lib/x86_64-linux-gnu/)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wall -Ofast -Wfatal-errors -D_MWAITXINTRIN_H_INCLUDED")