gilbert189 @gilbert189

After 5.5 hours of training an even bigger @imgru model, here are the results:

why are daid bth we do something like maybe whatvolate drogin me and ratio-'without at |sele

RPlike asloscreaba of all time

545 th green My first in part of even evols, its 8 cant fraldrright if you car. Mode is that never all detate ul a sle of 12

agrebut im gonna actually do ulving school I among us too muchat

never an ASGONGI2 is fun! 🎉normir bang

mfugg. ... -nothing that which is pay of suggestace based 5th tiony I shate ase fortled through the only someone tem. You always doila.

:A:o+before

b already concerrace

me to though... entknow it's what 2 odusos (i verified 55i want a while pxel for metake was never icked go to the cause anyone of hensing, say jgal / groups

Hi will say

jeffalo spread view get if im following morning but i wanit haven't it" with this wall

I’d say that this is already good enough for my purposes. It can generate semi-coherent sentences (well, semi might be a bit generous) and I don’t feel like using more of my Colab quota right now

Jan 30, 2024, 3:45 PM
0

comments