New YouTube videos on our self-driving RC car project!

Hey friends!

I’ve been working with one of our Interns, Armand, on a project where we use the NVIDIA Jetson Nano to get an RC car to drive itself.

We just released two new videos about some extensions to the basic project: one on using RL to do donuts and another on using YOLOv5 to detect and stop at stop signs.

:doughnut: - Teaching an RC car to do donuts with deep RL - YouTube
:stop_sign: - Object detection on an RC car with W&B and YOLOv5 - YouTube

The videos are fun little overviews of the project. If you’ve got detailed questions, post them as comments on YouTube or follow up in this thread.

Here’s some pictures and stills from the filming process:


What have your experiences been with the Jetson Nano? I just got one and am really excited to play around with it. It seems like it is pretty Docker dependent at least from my first exploration of it. Curious if you’ve had the same experience.

Nice! I’m excited to see what you do with it. I don’t have one myself but advising on this project has really made me want to pick one up. The biggest pain point for us was compiling libraries – there aren’t always pre-compiled versions available, which adds a lot to the iteration time for some projects.

I don’t think Armand used Docker – at least not directly. We based the project on the NVIDIA jetracer repo, which doesn’t seem to use Docker.

1 Like

I’m here to say the w&b plate looks :ok_hand:


message @armandpl for purchase inquiries


I thought I was eligible for a 100% Chai discount :frowning:


Sorry friend, GPUs are expensive :wink: