Testing VLMs and LLMs for Robotics with the Jetson Thor Devkit

Sentdex dives into NVIDIA’s new Jetson Thor devkit to see how local vision-language models (VLMs) and large language models (LLMs) can power real-world robotics projects. He walks through setup, demos model performance on the edge, and shares tips for squeezing the most out of this beefy little board.

Along the way you’ll find links to the official Jetson Thor page, the “Neural Networks from Scratch” book, plus all of Sentdex’s community hubs—Discord, Reddit, social media, and more—so you can jump in, ask questions, and keep the robot-tinkering party going.

Watch on YouTube