Training Data for LLM

[alphacoder01]

Hi, It was mentioned in README.md that the LLMs were trained to produce the code like structure, is this some open-source dataset or was this manually created. If so what were the stats like how many samples etc.

---

Thanks