Stars
KAI Scheduler is an open source Kubernetes Native scheduler for AI workloads at large scale
A storage solution for PyTorch tensors with distributed tensor support.
A PyTorch native platform for training generative AI models
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
Best practices for training DeepSeek, Mixtral, Qwen and other MoE models using Megatron Core.
Scalable toolkit for efficient model reinforcement
Ongoing research training transformer models at scale
Official codebase for the Siggraph Asia 2025 paper AutoBrep: Autoregressive B-Rep Generation with Unified Topology and Geometry
Cost-efficient and pluggable Infrastructure components for GenAI inference
Port of the OpenCascade CAD library to JavaScript and WebAssembly via Emscripten.
example repository for opencascade.js
📄 Set of modern React components for PDF highlighting
Set of React components for PDF annotation
Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM
Official repository of DARE: dLLM Alignment and Reinforcement Executor
The code used to train and run inference with the ColVision models, e.g. ColPali, ColQwen2, and ColSmol.
Rust library for vector embeddings and reranking.
Session replay, cobrowsing and product analytics you can self-host. Best for reproducing issues and iterating on your product.
🧑🚀 Authentication and authorization infrastructure for SaaS and AI apps, built on OIDC and OAuth 2.1 with multi-tenancy, SSO, and RBAC.
Simple and efficient DeepSeek V3 SFT using pipeline parallel and expert parallel, with both FP8 and BF16 trainings
Curated list of datasets and tools for post-training.
An interface library for RL post training with environments.