This could be the largest synthetic code dataset yet

Totaling nearly 1 trillion tokens of code across 15 programming languages, the dataset is being released with the recipes used to create it.