How can I train a custom model on a server then use the model on jetson nano

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

I am very new to the world of deep learning. I have recently bought a jetson nano to train custom models. I have come to the conclusion that this takes way too long. I would like to use my computer to speed things up. What is the best way to go about doing this?

3

Hi @justin60, on your PC you can install the framework used to train your models (for example PyTorch, TensorFlow, ect). For a PC with a GPU, I recommend Ubuntu and the framework’s corresponding NGC container.

6

Hi thanks for the reply! Is there anywhere I can find a walk-through or any sort of structure to follow? Again, I am very new to this whole scene. Thanks

7

Hi @justin60 , you can train a custom model on your PC which has access to GPU.
Please refer this GitHub - dusty-nv/jetson-inference: Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.. this will help you to gain understanding on how to run your model on Jetson Nano

8

Hi @ankithsingh, thanks for the reply! I got something kind of working in reference to train my computer, I’m gonna try and transfer the trained model to the jetson soon. I will drop any confusion or progress here.