Model-GPT2-Megatron-dvc
February 15, 2023 ยท View on GitHub
๐ฅ๐ฅ๐ฅ Deploy GPT-2 Megatron-LM model on VDP. It is trained on text sourced from Wikipedia, RealNews, OpenWebText, and CC-Stories and contains 345 million parameters.
This repo contains GPT-2 model in FasterTransformer format manged by DVC.