I thought I might start a new thread just to see if anyone else is trying to use the nVidia Jetson Nano as a local target for running the Notebooks in this course?
I am not sure why, but I cannot get Kaggle to even load the datasets needed to complete the workbooks. Perhaps it doesn’t know how to deal with us upside-down people from Australia.
Anyway, I thought about it and I realised that I have a CUDA processing unit available to me via my nVidia Jetson Nano. All I have to do i install jupyter and off i go, right?
Well, off course it wasn’t that simple. If anyone is interested, I will write up my notes and present them here.
I must say the Jetson Nano was a little “clunky” to use, but it out performed my 6th GEN Corei7 hands down on the same training sets. The Jetson took around 15 seconds per epoch and the Intel took around 45 seconds.
Let me know if you are interested?