I already have half this program churned out. There is one huge bug, the loss appears to oscillate up and down. Let's say |a| - |b| distance is 5, and gradient descent turns it down to 2, it will then slowly bump back up to 9. I am not sure why this is.
07:39 PM Feb 20th
from web
Thomas Haverford
Yeah, I'll come back to this later. I don't have the energy to diagnose the computer with something that is basically the equivalent of a neurological disease.
01:11 AM Feb 21st
from web
I'm working on a new project. The idea is this: We have a program that (loosely) simulates an aspect of the human-brain, and we teach it how to count and predict certain tasks. This should be simple; as these tasks follow a regression up-scaled pattern.
07:37 PM Feb 20th
from web
@Vlad69420, sure. I came from the bullstin.com site
07:04 PM Feb 20th
from web in reply to Vlad69420
Hi Vlad :)
06:58 PM Feb 20th
from web
Hello Everyone!
06:53 PM Feb 20th
from web
