Ongoing research training transformer models at scale
Contributions:6 reviews, 52 PRs, 265 pushes in 2 years 5 months
transformer-modelstransformers
A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper GPUs, to provide better performance with lower memory utilization in both training and inference.
Contributions:58 pushes, 13 branches in 2 years 10 months