Name meaning:
Fed: federated learning.
Fuse: fused CUDA operators that combine multiple FL server-side steps into one optimized primitive.
联邦融合算子库: 面向联邦学习服务端的 CUDA 融合算子库。
The goal is simple: make FL-specific server operations fast, reproducible, and easy to plug into frameworks such as Flower, FedLab, NVIDIA FLARE, OpenFL, FATE, or custom PyTorch services.
点击空白处退出提示












评论