Training SDXL Loras on Jetson AGX Orin 32gb - Ubuntu 20.04 / JP5.1.5 / L4T 35.6.3 / CUDA 11.4

Robotics & Edge Computing Jetson Systems Jetson AGX Orin
1

With this actual JP5, and CUDA cores, can i train SDXL Loras without upgrading to JP6.2?

As i cant run Comfyui at that point without flashing the device and take a huge risk of brciking it, i will use Stable diffusion webui (works 100%). but i need to train some loras in order to get my project done. Any suggestions? thx

5

Hi,

It’s recommended to use a desktop GPU for training.
Jetson is designed for inference tasks.

Thanks.