This Repo is the first Language Model by TNSA "nGPT-1" also know as "NGen" model series which was trained to predict the next token from a seed this repo is Open for Academic use and is Open to reproduce and re publish according it needs.
This Model was trained with 500K Parameters using TensorFlow Library with 1Million Tokens of data and took 3Days to train a stable checkpoint on a RTX 3050