PyTorch DTensor
xzyaoi opened this issue · comments
Xiaozhe Yao commented
PyTorch now has native support for distributed tensor, might be a better way to do TP than megatron's MPU.
Utilities for Training Very Large Models
xzyaoi opened this issue · comments
PyTorch now has native support for distributed tensor, might be a better way to do TP than megatron's MPU.