dice-group / LOLA-Megatron-DeepSpeed

Ongoing research training transformer language models at scale, including: BERT & GPT-2

Geek Repo

Github PK Tool