TAO 5.0 Documentation and getting started

Accelerated Computing Intelligent Video Analytics TAO Toolkit
1

Tried to get TAO 5.0 installed by following the video at Get Started with NVIDIA TAO Toolkit but could not follow along typing all those multiple and very long commands, some unclear in the video!!!

Looked up the latest TAO documentation but there the latest is 4.0…

For example, when running

distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
&& curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
&& curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

I get

OK
deb https://nvidia.github.io/libnvidia-container/stable/ubuntu18.04/$(ARCH) /
#deb https://nvidia.github.io/libnvidia-container/experimental/ubuntu18.04/$(ARCH) /
deb https://nvidia.github.io/nvidia-container-runtime/stable/ubuntu18.04/$(ARCH) /
#deb https://nvidia.github.io/nvidia-container-runtime/experimental/ubuntu18.04/$(ARCH) /
deb https://nvidia.github.io/nvidia-docker/ubuntu18.04/$(ARCH) /

I noticed Ubuntu18.04 despite being on 20.04

Then running the next

sudo apt-get update

There is an error:

E: Conflicting values set for option Signed-By regarding source https://nvidia.github.io/libnvidia-container/stable/ubuntu18.04/amd64/ /: /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg !=
E: The list of sources could not be read.

Thanks!

3

May I know in which doc did you find this link?

4

Started Here. Selected Get Started

That brings me here, and Selected Download TAO

Which brings me here where the getting started video is near the top of the page

I am happy to help you diagnose your documentation, but I’m still without a guide to setup TAO 5.0

Thanks!

David

5

The video is a bit old since it mentioned “Date: April 2022”
For TAO 5.0, you can get started with below:
Latest notebook: TAO Toolkit Getting Started | NVIDIA NGC
Latest docker: TAO Toolkit | NVIDIA NGC
4.0 version doc(please note that the 5.0 doc is not public yet due to some issues in server): TAO Toolkit (Latest Release) - NVIDIA Docs
And also, latest source code in the bottom of NVIDIA Corporation · GitHub

6

Well, I know its delayed, since it was originally scheduled for 6/6/23, but I did get this email

image
image616×577 51 KB

And Download Now links to here

7

@Morganh Thanks! Installed and tested a couple of notebooks.

Please note the following requirements are missing from the requirements file:

pip3 install opencv_python
pip3 install matplotlib

CAREFUL: A WORD TO THE BRAVE…
TAO5 requires a newer version of the NVIDIA driver. If you need to upgrade, make sure you understand your strategy to do so, as your computer may be stuck in the middle of installation with no graphics driver, including which version you will install. In my case I had to try a few versions, and the only driver that did work for me was nvidia-driver-510-server. I used the following commands:

sudo apt-get purge nvidia*
sudo apt-get update
ubuntu-drivers devices
sudo apt-get install nvidia-driver-510-server

8

Latest 5.0 user guide is available in TAO Toolkit (Latest Release) - NVIDIA Docs

9

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.