bigcode-project / Megatron-LM

Ongoing research training transformer models at scale
Other
371 stars 48 forks source link

Add tokens-per-second-gpu to the printed logs instead of just wandb #54

Closed loubnabnl closed 1 year ago