Gemma3 Running at 15 Tokens/sec on Jetson Orin Nano | Live VLM Demo

Watch a live demo by Asier Arranz, Senior Developer Advocate at NVIDIA, running Gemma3 on Jetson Orin Nano at 15 tokens/sec, including both the 4B and an impressive 12B model—right from Google’s Gemma3 launch event in Paris. Learn how to deploy high-performance Visual Language Models (VLMs) on compact edge devices for real-time, multimodal AI. If you're exploring similar projects or have ideas you'd like to share, feel free to comment or get in touch! Subscribe to Google for Developers → https://goo.gle/developers #Gemma #GemmaDeveloperDay