Tweeted By @huggingface
We just added MiniLM checkpoints from @MSFTResearch to the 🤗 transformers model hub! Better performance than BERT-base with 0.3x size and 2.7x speed! Check out: https://t.co/vCTeqR1z62@furu_wei @donglixp @HangboBao @__saber_str
— Hugging Face (@huggingface) June 22, 2020