DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

Run Mixtral-8x7B models in Colab or consumer desktops

Decentralized deep learning in PyTorch. Built to train models on thousands of volunteers across the world.

Mixture-of-Experts for Large Vision-Language Models

PyTorch Re-Implementation of "The Sparsely-Gated Mixture-of-Experts Layer" by Noam Shazeer et al. https://arxiv.org/abs/1701.06538

⛷️ LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-training

A TensorFlow Keras implementation of "Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts" (KDD 2018)

Tutel MoE: An Optimized Mixture-of-Experts Implementation

Surrogate Modeling Toolbox

A Pytorch implementation of Sparsely-Gated Mixture of Experts, for massively increasing the parameter count of language models

From scratch implementation of a sparse mixture of experts language model inspired by Andrej Karpathy's makemore :)

中文Mixtral混合专家大模型（Chinese Mixtral MoE LLMs）

GMoE could be the next backbone model for many kinds of generalization task.

A library for easily merging multiple LLM experts, and efficiently train the merged LLM.

Implementation of ST-Moe, the latest incarnation of MoE after years of research at Brain, in Pytorch

Inferflow is an efficient and highly configurable inference engine for large language models (LLMs).

Implementation of Soft MoE, proposed by Brain's Vision team, in Pytorch

Fast Inference of MoE Models with CPU-GPU Orchestration

A curated reading list of research in Adaptive Computation, Dynamic Compute & Mixture of Experts (MoE).

Some personal experiments around routing tokens to different autoregressive attention, akin to mixture-of-experts