#moe Repositories
Discover 34 active open source GitHub repositories tagged with #moe, accumulating 317,125 total stars (averaging 9,327 stars per project). Leading this topic is vllm-project/vllm with 97,302 stars.
34 repositories total
SGLang is a high-performance serving framework for large language models and multimodal models.
FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast and efficiently.
Use PEFT or full-parameter training for CPT/SFT/DPO/GRPO on 600+ LLMs (Qwen3.8, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM5.3, Gemma4, LLaVA, Phi4, ...) (AAAI 2025).
TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way.
A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.
:electron: An unofficial https://bgm.tv UI-first app client for Android and iOS, built with React Native. An ad-free, hobby-driven, non-profit, ACG-focused alternative to Douban for anime tracking, a third-party bgm.tv client. Redesigned for mobile devices, featuring many enhanced functionalities hard to implement on the web version, along with substantial customization options. Currently adapted for iOS / Android.
Run GLM-5.3 (744B), DeepSeek-V4.1-Flash (748B), MiniMax-M3.1-Flash (428B), GLM-5.3-Flash (321B), and Hy4-Flash (220B) on your own GPUs โ starting from 8 GB of VRAM, with NVIDIA and AMD support, one dashboard, and an OpenAI- and Anthropic-compatible API.
UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g., GPU-driven)
Run models larger than your RAM or VRAM by streaming MoE experts from SSD. Built on llama.cpp.
PyTorch Re-Implementation of "The Sparsely-Gated Mixture-of-Experts Layer" by Noam Shazeer et al. https://arxiv.org/abs/1701.06538
Run MoE models bigger than your RAM. Frontier-size MoE on a phone, laptop, or edge device. Lossless, on stock llama.cpp. An adaptive streaming optimizer. Any hardware, any OS, any model, any quantization, behind a local OpenAI-compatible server.
Tutel MoE: Optimized Mixture-of-Experts Library, Supporting GLM/Kimi/GPT/DeepSeek/Qwen with FP8/NVFP4/MXFP4
Run Kimi K3 locally on CPU with ~55GB measured runtime RAM. A single-file Linux inference server powered by cPilot Runtime.
cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.
โท๏ธ LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-training (EMNLP 2024)
Halo is an open-source framework built by White Circle for training large language and multimodal models
An open-source solution for full parameter fine-tuning of DeepSeek-V3/R1 671B, including complete code and scripts from training to inference, as well as some practical experiences and conclusions.
โก Native MLX Swift LLM inference server for Apple Silicon. OpenAI-compatible API, SSD streaming for 100B+ MoE models, TurboQuant KV cache compression, MACOS + iOS iPhone app.
[CVPR2026]๐๐๐Official code for the paper "YOLO-Master: MOE-Accelerated with Specialized Transformers for Enhanced Real-time Detection." *(YOLO = You Only Look Once)* ๐ฅ๐ฅ๐ฅ
Pure Rust + CUDA LLM inference engine โ no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2
37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.
Run a 105 GB AI model on a Mac that can't hold it. Qwen3.8-Flash-Next 4-bit, streamed from SSD on 16โ64 GB Apple Silicon Macs. Fully native Swift, MLX and Metal. Offline, no Python. CLI, Swift library, and Ollama/OpenAI/Anthropic APIs. Works with Claude Code and Codex.
Bilingual (Chinese + English) ML / LLM / diffusion / agent interview cheat sheets for AI autumn recruiting โ generated by ARIS /interview-cheatsheet, rendered by /render-html into single-file HTML that can be read anywhere โ plus a CVโDBLP fact-checked academic homepage generator and hand-authored long-form blogs ๐ฑ
Frontier MoE models on RTX PRO 6000 / RTX 5090 (sm_120): DeepSeek-V4.1-Flash on 4x RTX PRO 6000, as a GPU+CPU hybrid on 4x RTX 5090, and 2-bit experts on vLLM v0.24.0
Xolotl: run DeepSeek-V4.1-Flash, all 552B parameters, as a local coding agent on one gaming PC (RTX 4090 + 110 GB RAM): ~33 tok/s agent, ~41 fast, 512K context. Plus Inkling-2-Mini (Thinking Machines, own engine + MTP drafter), GLM-5.3-Flash, Qwen3.8, LFM3-Omni. MoE experts tiered across VRAM, RAM, and NVMe (Xolotl Direct). OpenAI + Anthropic API.
Frequently Asked Questions
What is the most popular #moe repository on GitHub?
The top ranked open source #moe project on GitHub is vllm-project/vllm with 97,302 stars. A high-throughput and memory-efficient inference and serving engine for LLMs
Which programming languages are most used in #moe repositories?
Among the top ranked #moe projects on GitHub, the most common programming languages are Python, C++, C, Cuda.
How is the #moe repository ranking calculated?
We track 34 active open source repositories tagged with #moe (500+ stars), totaling 317,125 stars with an average of 9,327 stars per repository. Rankings update daily from the GitHub API.