Can't install TensorRT via pip

Description

After following the Quick Start Guide I get an error.

In my conda environment I executed pip install --upgrade setuptools pip and pip install nvidia-pyindex without any issues.

(omct) lennux@lennux-desktop:~$ pip install nvidia-pyindex
Looking in indexes: Simple index, https://pypi.ngc.nvidia.com
Requirement already satisfied: nvidia-pyindex in ./anaconda3/envs/omct/lib/python3.7/site-packages (1.0.9)

But when i tried pip install --upgrade nvidia-tensorrt I get the attached output below.

(omct) lennux@lennux-desktop:~$ pip install --upgrade nvidia-tensorrt
Looking in indexes: Simple index, https://pypi.ngc.nvidia.com
Collecting nvidia-tensorrt
Downloading nvidia-tensorrt-0.0.1.dev5.tar.gz (7.9 kB)
Preparing metadata (setup.py) … error
error: subprocess-exited-with-error

Ă— python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [17 lines of output]
Traceback (most recent call last):
File “”, line 36, in
File “”, line 34, in
File “/tmp/pip-install-ptqq18x_/nvidia-tensorrt_447c81bdc9ae47b5bb2882c9a3afd605/setup.py”, line 150, in
raise RuntimeError(open(“ERROR.txt”, “r”).read())
RuntimeError:
###########################################################################################
The package you are trying to install is only a placeholder project on PyPI.org repository.
This package is hosted on NVIDIA Python Package Index.

  This package can be installed as:
  ```
  $ pip install nvidia-pyindex
  $ pip install nvidia-tensorrt
  ```
  ###########################################################################################
  
  [end of output]

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

Ă— 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.

The quick start guide states the following this for such a error message .

If you receive an error message in the form of a Python exception similar to the error message below, then either the nvidia-pyindex package was not set up properly, or you might be using a Python version other than 3.6 to 3.9.

Now it doesn’t seem like this addresses my issue. Any help is appreciated.

Environment

TensorRT Version: The one that comes with the pip command i guess.
GPU Type: RTX 2060
Nvidia Driver Version: 470.103
CUDA Version: Nvidia-smi reports :“11.4”, but nvcc --version sais: “Cuda compilation tools, release 11.0, V11.0.221” maybe someone here can tell me what my CUDA version actually is. I think downloaded the CUDA toolkit 11.0 a long time ago but by now i have also updated my nvidia drivers multiple times.
CUDNN Version: Well, I installed 8.1.1 a long time ago for an old project, along with an older nvidia driver version. By now, i have updated my driver but i haven’t touched the CUDNN since. I didn’t download a new one, since I’d like to use the pip installation and i thought the wheel files are “fully self-contained”. After reading the TensorRT quick start guide I came to the conclusion that I wouldn’t need to install CUDNN separately. Am i wrong here?
Operating System + Version: Ubuntu 20.04 LTS
Python Version (if applicable): Within my anaconda environment 3.7.7.
Baremetal or Container (if container which image + tag): No container

Hi,
Please refer to the installation steps from the below link if in case you are missing on anything

Also, we suggest you to use TRT NGC containers to avoid any system dependency related issues.

Thanks!

Also have this problem

How do we solve this in arm64 without containers?

4 Likes