TsinghuaC3I/ZEDA

The official implementation of the paper "Towards Efficient Mixture of Experts: A Holistic Study of Compression Techniques (TMLR)".

Pruning and Distilling Mixture-of-Experts into Dense Language Models

PyTorch library for cost-effective, fast and easy serving of MoE models.

Dist-MoE implements Speed-of-Light distributed Mixture-of-Experts forward and backward kernels. They are highly optimized and fused comm-compute kernels for training and inference on SM100+ GPUs

Expert Kit is an efficient foundation of Expert Parallelism (EP) for MoE model Inference on heterogenous hardware

Compact and Agent-Native MoE Training System

Implementation of LatentMoE,Toward Optimal Accuracy per FLOP and Parameter in Mixture of Experts (Elango et al., NVIDIA 2026) — in Pytorch. A single-file, dependency-light layer you can drop in place of a standard MoE FFN.

Accelerating MoE with IO and Tile-aware Optimizations

Tutel MoE: Optimized Mixture-of-Experts Library, Support GLM/Kimi/Gpt/DeepSeek/Qwen using FP8/NVFP4/MXFP4