requirements.txt failure of TensorRT 10.3 when running simple_progress_monitor on Jetson Orin Nano in Conda env

Description

I am trying to run the the simple_progress_monitor.py example provided in TensorRT 10.3.

The problem is with the requirements.txt file. Jetpack 6.2.1 comes with Python 3.10.12. Therefore there is no suitable version of cuda-python that cane be install since it lies between 3.10 and 3.11.

https://github.com/NVIDIA/TensorRT/blob/5b990f0a739d8faf962dfe54f4829942633c639c/samples/python/simple_progress_monitor/requirements.txt#L2-L3

This is the error that is shown

(simple_tensorrt) jetson@jetson-desktop:/usr/src/tensorrt/samples/python/simple_progress_monitor$ conda install --yes --file requirements.txt
2 channel Terms of Service accepted

InvalidMatchSpec: Invalid spec 'cuda-python ==12.2.0;python_version<="3.10"': Invalid version '12.2.0;python_version<="3.10"': invalid character(s)

Additionally, there is a line to install pywin32 which is unparseable on the Ubuntu 22.04 system that the Jetson Orin Nano runs off.

https://github.com/NVIDIA/TensorRT/blob/5b990f0a739d8faf962dfe54f4829942633c639c/samples/python/simple_progress_monitor/requirements.txt#L4

CondaValueError: could not parse 'pywin32; platform_system == "Windows"' in: requirements.txt

Given that:

• TensorRT support of the Jetson ecosystem is a major product feature
• Jetson Orin Nano and TensorRT products are both developed by NVIDIA
• Neither the Jetson Orin Nano nor TensorRT 10.3 are retired
• Jetpack 6.2.1 is the latest officially supported version supported by NVIDIA

I would appreciate how I can run a simple TensorRT example with the Jetson Orin Nano in a smooth manner. Thanks!

Environment

TensorRT Version: 10.3.0

NVIDIA GPU: Jetson Orin Nano (Ampere)

NVIDIA Driver Version: 540.4.0 (Jetpack 6.2.1)

CUDA Version: 12.6

CUDNN Version:

Operating System: Ubuntu 22.04

Python Version (if applicable): 3.10.12

Tensorflow Version (if applicable):

PyTorch Version (if applicable):

Baremetal or Container (if so, version): Miniconda

jetson@jetson-desktop:/usr/src/tensorrt/samples/python/simple_progress_monitor$ dpkg -l | grep tensorrt
ii  nv-tensorrt-local-tegra-repo-ubuntu2204-10.3.0-cuda-12.5 1.0-1                                       arm64        nv-tensorrt-local-tegra repository configuration files
ii  nvidia-tensorrt                                          6.2.1+b38                                   arm64        NVIDIA TensorRT Meta Package
ii  nvidia-tensorrt-dev                                      6.2.1+b38                                   arm64        NVIDIA TensorRT dev Meta Package
ii  tensorrt                                                 10.3.0.30-1+cuda12.5                        arm64        Meta package for TensorRT
ii  tensorrt-libs                                            10.3.0.30-1+cuda12.5                        arm64        Meta package for TensorRT runtime libraries

jetson@jetson-desktop:/usr/src/tensorrt/samples/python/simple_progress_monitor$ sudo apt-cache show nvidia-jetpack
[sudo] password for jetson: 
Package: nvidia-jetpack
Source: nvidia-jetpack (6.2.1)
Version: 6.2.1+b38
Architecture: arm64
Maintainer: NVIDIA Corporation
Installed-Size: 194
Depends: nvidia-jetpack-runtime (= 6.2.1+b38), nvidia-jetpack-dev (= 6.2.1+b38)
Homepage: http://developer.nvidia.com/jetson
Priority: standard
Section: metapackages
Filename: pool/main/n/nvidia-jetpack/nvidia-jetpack_6.2.1+b38_arm64.deb
Size: 29300
SHA256: dd9cb893fbe7f80d2c2348b268f17c8140b18b9dbb674fa8d79facfaa2050c53
SHA1: dc630f213f9afcb6f67c65234df7ad5c019edb9c
MD5sum: 9c8dc61bdab2b816dcc7cd253bcf6482
Description: NVIDIA Jetpack Meta Package
Description-md5: ad1462289bdbc54909ae109d1d32c0a8

Package: nvidia-jetpack
Source: nvidia-jetpack (6.2)
Version: 6.2+b77
Architecture: arm64
Maintainer: NVIDIA Corporation
Installed-Size: 194
Depends: nvidia-jetpack-runtime (= 6.2+b77), nvidia-jetpack-dev (= 6.2+b77)
Homepage: http://developer.nvidia.com/jetson
Priority: standard
Section: metapackages
Filename: pool/main/n/nvidia-jetpack/nvidia-jetpack_6.2+b77_arm64.deb
Size: 29298
SHA256: 70553d4b5a802057f9436677ef8ce255db386fd3b5d24ff2c0a8ec0e485c59cd
SHA1: 9deab64d12eef0e788471e05856c84bf2a0cf6e6
MD5sum: 4db65dc36434fe1f84176843384aee23
Description: NVIDIA Jetpack Meta Package
Description-md5: ad1462289bdbc54909ae109d1d32c0a8

Package: nvidia-jetpack
Source: nvidia-jetpack (6.1)
Version: 6.1+b123
Architecture: arm64
Maintainer: NVIDIA Corporation
Installed-Size: 194
Depends: nvidia-jetpack-runtime (= 6.1+b123), nvidia-jetpack-dev (= 6.1+b123)
Homepage: http://developer.nvidia.com/jetson
Priority: standard
Section: metapackages
Filename: pool/main/n/nvidia-jetpack/nvidia-jetpack_6.1+b123_arm64.deb
Size: 29312
SHA256: b6475a6108aeabc5b16af7c102162b7c46c36361239fef6293535d05ee2c2929
SHA1: f0984a6272c8f3a70ae14cb2ca6716b8c1a09543
MD5sum: a167745e1d88a8d7597454c8003fa9a4
Description: NVIDIA Jetpack Meta Package
Description-md5: ad1462289bdbc54909ae109d1d32c0a8

Relevant Files

Model link: N/A

Steps To Reproduce

1. Make sure that the Jetson Orin Nano is running Jetpack 6.2.1.
2. Install Miniconda
3. Create a new environment conda create --name simple_tensorrt
4. Set the Python environment to 3.10.12 (same as the base installation) conda activate simple_tensorrt \ conda install python=3.10.12
5. Go to location of the example cd /usr/src/tensorrt/samples/python/simple_progress_monitor
6. Attempt to install the packages in the requirements.txt file. conda install --yes --file requirements.txt
7. Comment out the lines as shown below, in order for the packages to be installed.
8. cuda-python 12.2.0 is not present on the default anaconda channel and it has to be installed separately from the NVIDIA channel. conda install nvidia::cuda-python==12.2.0

Commands or scripts: Mentioned in steps to reproduce section above.

Have you tried the latest release?:

TensorRT and Jetpack 6.2.1 do not play well together. Sticking to the official stable versions included. Dependency hell is sadly a challenge with the Jetson ecosystem which I am trying to avoid.

Attach the captured .json and .bin files from TensorRT's API Capture tool if you're on an x86_64 Unix system

N/A

Can this model run on other frameworks? For example run ONNX model with ONNXRuntime (polygraphy run <model.onnx> --onnxrt):

N/A

Start with samples/python/simple_progress_monitor/requirements.txt and the simple_progress_monitor.py entry point. Reproduce the reported conda install on JetPack 6.2.1 with Python 3.10.12, then verify that the requirements can be processed on Ubuntu 22.04 without the reported parsing errors and that the example dependencies install successfully.