train llama on a single A100 80G node using 🤗 transformers and 🚀 Deepspeed Pipeline Parallelism
Geek Repo:Geek Repo
Github PK Tool:Github PK Tool