Is it possible to train models for my Jetson Nano on my Host Pc and easy deploy it to the nano?
When using the headless mode and making changes to my models on the nano in jupyterlab, it just uses the nanos cpu&gpu and not my hosts’, right?
So is there a comfortable way to make changes to the nano?
Yes. Training on a host with NVIDIA GPUs is the ideal way. For a good tutorial, please see here under “two days to a demo”. You will need a computer with NVIDIA GPUs and Linux installed, or a Linux VM with NVIDIA GPUs passed through (advanced).
When using Jupyter running on the Nano, the nano’s GPU and CPU is used, correct.
thank you.