DeepEP: an efficient expert-parallel communication library

FSA/FST algorithms, differentiable, with PyTorch compatibility.

Official pytorch Code for CVPR2019 paper "Fast Human Pose Estimation" https://arxiv.org/abs/1811.05419

Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity