【subprocess-exited-with-error】pip安装tensorrt-9.3.0.post12.dev1.tar.gz

前情提要

本人已设置过清华源

https://pypi.tuna.tsinghua.edu.cn/simple

安装之前需要更新pip，安装nvidia-pyindex

python3 -m pip install --upgrade setuptools pip
python3 -m pip install nvidia-pyindex

报错内容

报错一：9.3.0版本无法安装

安装TensorRT官网指导安装时出现的报错：

# python3 -m pip install --pre --upgrade tensorrt
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple, https://pypi.ngc.nvidia.com
Collecting tensorrt
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/bc/70/d633add6eef5edf414f3ec48b3fe9b0116837cf4b935edc531ede2e65b00/tensorrt-9.3.0.post12.dev1.tar.gz (6.9 kB)
  Preparing metadata (setup.py) ... error
  error: subprocess-exited-with-error

  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> [6 lines of output]
      Traceback (most recent call last):
        File "<string>", line 2, in <module>
        File "<pip-setuptools-caller>", line 34, in <module>
        File "/tmp/pip-install-pd1g0qot/tensorrt_ab7006731a644e11b9e76fbb52612c47/setup.py", line 90, in <module>
          raise RuntimeError("Bad params")
      RuntimeError: Bad params
      [end of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.

报错二：8.6.1版本无法安装

安装中没有反应，只能手动取消

# pip install tensorrt
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Collecting tensorrt
  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/46/08/da496b4f20890b6f717f159bbdda98e9248b560e712aa22cd9bb1daafe15/tensorrt-8.6.1.post1.tar.gz (18 kB)
  Preparing metadata (setup.py) ... done
Building wheels for collected packages: tensorrt
  Building wheel for tensorrt (setup.py) ... |^canceled
ERROR: Operation cancelled by user

解决方案

解决一：设置英伟达源

pip config set global.index-url https://pypi.nvidia.com
python3 -m pip install --pre --upgrade tensorrt

这种方式安装的是最新版9.3.0
安装时弹出的查找地址为

Looking in indexes: https://pypi.nvidia.com, https://pypi.ngc.nvidia.com

解决二：设置清华源下的英伟达源

pip install tensorrt --extra-index-url https://pypi.nvidia.com

这种方式安装的是最旧版8.6.1
安装时弹出的查找地址为

Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple, https://pypi.ngc.nvidia.com, https://pypi.nvidia.com

后记

我安装的自然是最新版9.3.0

检验

python
import tensorrt
print(tensorrt.__version__)
# 9.3.0.post12.dev1
assert tensorrt.Builder(tensorrt.Logger())

到这里还没结束，在之后需要编译的时候，需要指定tensorrt的lib和include路径，用pip安装的话路径和别的安装方法（tar、deb）中的路径不太一样，千万要注意！

设置include路径

git clone https://github.com/NVIDIA/TensorRT.git
cd TensorRT
git submodule update --init --recursive

其中git submodule update --init --recursive运行后要检查是否有出现Failed，有话再次运行进行下载
这样include路径就确认了

设置lib路径

在python路径下找

which python
#/root/anaconda3/envs/sakura/bin/python
cd /root/anaconda3/envs/sakura/lib/python3.10/site-packages/tensorrt_libs
ls
# __init__.py  __pycache__  libnvinfer.so.9  libnvinfer_builder_resource.so.9.3.0  libnvinfer_plugin.so.9  libnvonnxparser.so.9

可以看到许多动态库.so文件
之后编译的时候只需要修改tensorrt的lib和include位置即可
include directories对应include
link directories对应lib