# How to install libnvinfer-dev to drive px 2

**URL:** <https://forums.developer.nvidia.com/t/how-to-install-libnvinfer-dev-to-drive-px-2/157146>\
**Category:** TensorRT\
**Tags:** driveworks, tensorrt\
**Created:** [October 15, 2020, 12:45pm UTC](https://forums.developer.nvidia.com/t/how-to-install-libnvinfer-dev-to-drive-px-2/157146 "2020-10-15T12:45:58Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![kaanyilmaz.se](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@kaanyilmaz.se](https://forums.developer.nvidia.com/u/kaanyilmaz.se)\
**Post date:** [October 15, 2020, 12:45pm UTC](https://forums.developer.nvidia.com/t/how-to-install-libnvinfer-dev-to-drive-px-2/157146/1 "2020-10-15T12:45:58Z")

</div>

## Description

Hi, I want to compile a C++ code on Nvidia Drive PX 2 to make inference with TensorRT. TensorRT 4.1.1-1 is pre-installed in the machine. However it turns out that only runtime libraries for TensorRT are installed but the header and lib files are not installed therefore I cannot do any compilation since NvInfer.h cannot be found. I tried to run sudo apt-get install libnvinder-dev to have the dev package installed however the repo could not be found.

Is it possible to get the dev package for TensorRT 4 in drive px 2?

Secondly, I know that TensorRT in drive px 2 does not have a Python API. Would installing TensorRT to drive AGX (I might have access to a drive AGX) provide a python API?

## Environment

**TensorRT Version** : 4.1.1.1  
**GPU Type** : 2x Tegra “Parker” (Pascal) + 2x Unknown Pascal  
**Nvidia Driver Version** : could not find  
**CUDA Version** : 9.2  
**CUDNN Version** : not installed (at least no cudnn.h in /usr/local/cuda/include)  
**Operating System + Version** : ubuntu 16  
**Python Version (if applicable)**: 3.6  
**TensorFlow Version (if applicable)**: not available  
**PyTorch Version (if applicable)**: not available  
**Baremetal or Container (if container which image + tag)**: unknown

---

<div class="post-metadata">

**Author:** ![AakankshaS](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/aakankshas/32/14047_2.png) [@AakankshaS](https://forums.developer.nvidia.com/u/AakankshaS)\
**Post date:** [October 15, 2020, 2:33pm UTC](https://forums.developer.nvidia.com/t/how-to-install-libnvinfer-dev-to-drive-px-2/157146/2 "2020-10-15T14:33:55Z")

</div>

Hi @kaanyilmaz.se,  
You are using TRT4.1 which is very old release.  
We request you to upgrade to latest TRT release.

Thanks!

---

<div class="post-metadata">

**Author:** ![kaanyilmaz.se](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@kaanyilmaz.se](https://forums.developer.nvidia.com/u/kaanyilmaz.se)\
**Post date:** [October 15, 2020, 2:40pm UTC](https://forums.developer.nvidia.com/t/how-to-install-libnvinfer-dev-to-drive-px-2/157146/3 "2020-10-15T14:40:11Z")

</div>

Could you kindly confirm that any TRT version greater than 5 requires CUDA version \>= 10.0 ? Is updating the CUDA version 9 to 10.X or 11.X suggested for drive px 2?

---

<div class="post-metadata">

**Author:** ![AakankshaS](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/aakankshas/32/14047_2.png) [@AakankshaS](https://forums.developer.nvidia.com/u/AakankshaS)\
**Post date:** [October 30, 2020, 4:40am UTC](https://forums.developer.nvidia.com/t/how-to-install-libnvinfer-dev-to-drive-px-2/157146/4 "2020-10-30T04:40:15Z")

</div>

Hi @kaanyilmaz.se,  
You can find the supported platform details wrt versions here.

> **[Release Notes :: NVIDIA Deep Learning TensorRT Documentation](https://docs.nvidia.com/deeplearning/tensorrt/release-notes/overview.html#overview)**
>
> NVIDIA TensorRT is a C++ library that facilitates high performance inference on NVIDIA GPUs. It is designed to work in connection with deep learning frameworks that are commonly used for training. TensorRT focuses specifically on running an already...

Thanks!