Jetson nano

Robotics & Edge Computing Jetson & Embedded Systems Jetson Nano
1

manifest for nvidia/cuda:10.2-devel-ubuntu18.04 not found: manifest unknown: manifest unknown

3

Hi,

Please find Jetson’s CUDA container below:

For example, you can use nvcr.io/nvidia/l4t-cuda:10.2.460-runtime.
Thanks.

4

i build trtexec with tensorrt in jetson nano

5

wwith github GitHub - NVIDIA/TensorRT: NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.

6

Hi,

Do you want to build TensorRT with a custom Python version?
If yes, you don’t need to set up a docker environment.

Please follow the steps shared in the below comment:

https://elinux.org/Jetson/L4T/TRT_Customized_Example#TensorRT_Python_Bindings

Thanks.

7

NO i want to rebuild trtexec tool. I run source code github TensorRT build trtexec on jetson nano. But the library environment is install very difficult, so Im try build docker tensorrt jetson nano to rebuild trtexec

8

Hi,

You can build trtexec with the following steps:

Setup environment with JetPack

For Jetson Nano, the latest software release is JetPack 4.6.6. with TensorRT 8.2:

Build trtexec

$ cd /usr/src/tensorrt/samples/trtexec/
$ sudo make
$ /usr/src/tensorrt/bin/trtexec --help

Thanks.

9

i build successfully