pytorch-fsdp2

【效率与协作】Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use when models exceed single-GPU memory or when you need DTensor-based sharding with DeviceMesh.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准