Transform datasets at scale. Optimize datasets for fast AI model training.

Thunder gives you PyTorch models superpowers for training and inference. Unlock out-of-the-box optimizations for performance, memory and parallelism, or roll out your own.

Streamlining reinforcement learning with RLOps. State-of-the-art RL algorithms and tools.

PyTorch version of Stable Baselines, reliable implementations of reinforcement learning algorithms.

High-quality single file implementation of Deep Reinforcement Learning algorithms with research-friendly features (PPO, DQN, C51, DDPG, TD3, SAC, PPG)

An API standard for single-agent reinforcement learning environments, with popular reference environments and related utilities (formerly Gym)

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

A Github Action to automatically bump and tag master, on merge, with the latest SemVer formatted version. Works on any platform.

Configuration classes enabling type-safe PyTorch configuration for Hydra apps

The Triton Inference Server provides an optimized cloud and edge inferencing solution.

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

A PyTorch native library for large-scale model training

A Python-level JIT compiler designed to make unmodified PyTorch programs faster.

Run PyTorch LLMs locally on servers, desktop and mobile

Finetune Qwen3, Llama 4, TTS, DeepSeek-R1 & Gemma 3 LLMs 2x faster with 70% less memory! 🦥

📚 Parameterize, execute, and analyze notebooks

A modern model graph visualizer and debugger

PyTorch native quantization and sparsity for training and inference

PyTorch native post-training library

The Modular Platform (includes MAX & Mojo)

GPU prices aggregator for cloud providers

Large Language Model Text Generation Inference

A high-throughput and memory-efficient inference and serving engine for LLMs

Build better UIs faster.

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

Containers for machine learning

Get up and running with Llama 3.3, DeepSeek-R1, Phi-4, Gemma 3, Mistral Small 3.1 and other large language models.

Standardized Serverless ML Inference Platform on Kubernetes

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

Typer, build great CLIs. Easy to code. Based on Python type hints.