The Modular Platform (includes MAX & Mojo)

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

Kimi Code CLI is your next CLI agent.

TurboDiffusion: 100–200× Acceleration for Video Diffusion Models

A fast communication-overlapping library for tensor/expert parallelism on GPUs.

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

Run LLMs with MLX

dstack is an open-source control plane for running development, training, and inference jobs on GPUs—across hyperscalers, neoclouds, or on-prem.

Open ABI and FFI for Machine Learning Systems

Deep learning at the speed of light.

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

Automatically Discovering Fast Parallelization Strategies for Distributed Deep Neural Network Training

The best ChatGPT that $100 can buy.

PyTorch Single Controller

🧬 The Huxley-Gödel Machine

Development repository for the Triton language and compiler

Approximate Nearest Neighbors in C++/Python optimized for memory usage and loading/saving to disk

💫 Toolkit to help you get started with Spec-Driven Development

Benchmark and optimize LLM inference across frameworks with ease

Original Apollo 11 Guidance Computer (AGC) source code for the command and lunar modules.

The simplest, fastest repository for training/finetuning small-sized VLMs.

[DEPRECATED] Moved to ROCm/rocm-libraries repo. NOTE: develop branch is maintained as a read-only mirror

super repo for rocm systems projects

Open Source framework for voice and multimodal conversational AI

[ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models

PyTorch native quantization and sparsity for training and inference

A better build tool for Java, Scala and Kotlin: Simpler than Maven, easier than Gradle, with 3-7x faster dev workflows than other JVM build tools

A nano Claude Code–like agent, built from 0 to 1