Twelve Thousand Fake Numbers With Normal Deviates
Sixtmenhreene hildse, seventy-seven thousand, seven hundred and twenty-seve
Swven hundred and seventy-four thousand,teegght hundred and fouty
Oixty-seven ty-oon zeileven thousand, fwve hundred and twenty-six,e oudssee fou hundred and fivt
For NaNoGenMo 2018 I generated Twelve Thousand Fake Numbers With Normal Deviates, a collection of things which are sort of similar to numbers but are not numbers. They are available for for all of your fake number needs.
The fake numbers were generated using a Recurrent Neural Network that I made following the guide from Machine Learning 4 Artists and I used the keras python library to actually build the thing. Along the way I also build a markov chain generator to make up fake numbers in ruby, which also created perfectly good fake numbers and was a lot faster to run. The last seven hundred or so fake numbers are from the markov chain generator.