Product
Ray Train
The distributed training component of the Ray ecosystem. It provides wrappers around parallel training strategies such as PyTorch DDP and FSDP, automatically handling process group initialization, fault recovery, and checkpoint saving. Ray Train is designed to support large-scale cluster training workloads.
Timeline (last 90 days)
Sep 30
Ray Train provides wrappers for both DDP and FSDP strategies, and can automatically handle process group initialization, fault recovery, and checkpoint saving
Unverified50%