User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

LLM-in-Sandbox Reinforcement Learning Enhances Generalization

Code for the Molmo2 Vision-Language Model

LLM-in-Sandbox: From Coding Agent to General Agent

Multiplex Thinking: Reasoning via Token-wise Branch-and-Merge

AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs

A cross-platform desktop All-in-One assistant tool for Claude Code, Codex, OpenCode, openclaw & Gemini CLI.

omo; the best agent harness - previously oh-my-opencode

~950 line, minimal, extensible LLM inference engine built from scratch.

Helpful tools and examples for working with flex-attention

DeepEP: an efficient expert-parallel communication library

OpenTinker is an RL-as-a-Service infrastructure for foundation models

Accelerating MoE with IO and Tile-aware Optimizations

Graph-structured Indices for Scalable, Fast, Fresh and Filtered Approximate Nearest Neighbor Search

Azure Storage Connector for PyTorch

A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.

PyTorch-native post-training at scale

Unofficial WIP LoRa Finetuning repository for VibeVoice

Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends

A single Gradio + React WebUI with extensions for ACE-Step, Kimi Audio, Piper TTS, GPT-SoVITS, CosyVoice, XTTSv2, DIA, Kokoro, OpenVoice, ParlerTTS, Stable Audio, MMS, StyleTTS2, MAGNet, AudioGen, …

VibeVoice: Expressive, longform conversational speech synthesis. (Community fork)

Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.5, GPT-OSS, Llama, and more!

A comprehensive ComfyUI integration for Microsoft's VibeVoice text-to-speech model, enabling high-quality single and multi-speaker voice synthesis directly within your ComfyUI workflows.

🎙️ VibeVoice FastAPI - Multi-Speaker TTS API

ComfyUI custom node for the VibeVoice TTS. Expressive, long-form, multi-speaker conversational audio

Open-Source Frontier Voice AI

Matrix (Multi-Agent daTa geneRation Infra and eXperimentation framework) is a versatile engine for multi-agent conversational data generation.

Dion optimizer algorithm

One-for-All Multimodal Evaluation Toolkit Across Text, Image, Video, and Audio Tasks

An open-source AI agent that brings the power of Gemini directly into your terminal.