yan @svmihar@qoto.org

saying goodbye to pytorch hooks (not really tho), but it's pretty much straightforward that i can see the representation of an input within layers inside of a model with fastai2!

with this growth rate of tflops, maybe in the next 5 years, scripts can out result even before running the actual script.

impressed by the ram usage + cpu usage while idle. arcolinux + bspwm + polybar.

RIP Joan Feynman