NVIDIA Jetson AI Lab Adds Tutorial For Unsloth Model Fine-Tuning

NVIDIA Robotics announced a Jetson AI Lab tutorial that demonstrates fine-tuning and quantizing AI models directly on NVIDIA Jetson hardware. The tutorial covers use of Unsloth together with memory-efficient QLoRA to customize models, followed by export of quantized GGUF files and local execution through llama.cpp. Hands-on examples include NVIDIA Nemotron 3.5 Lightning running on Jetson AGX Thor. The post positions the material as guidance for developers building AI-driven robots and exploring physical AI applications. No further updates or third-party confirmations appear in the provided evidence.