An Open Source Machine Learning Framework for Everyone

Cross-platform, customizable ML solutions for live and streaming media.

OpenPose: Real-time multi-person keypoint detection library for body, face, hands, and foot estimation

Productive, portable, and performant GPU programming in Python.

An open autonomous driving platform

A fast, distributed, high performance gradient boosting (GBT, GBDT, GBRT, GBM or MART) framework based on decision tree algorithms, used for ranking, classification and many other machine learning …

Open source simulator for autonomous vehicles built on Unreal Engine / Unity, from Microsoft AI & Research

Bullet Physics SDK: real-time collision detection and multi-physics simulation for VR, games, visual effects, robotics, machine learning etc.

Open-source simulator for autonomous driving research.

Multi-Joint dynamics with Contact. A general purpose physics simulator.

Unsupervised text tokenizer for Neural Network-based text generation.

COLMAP - Structure-from-Motion and Multi-View Stereo

Bringing Characters to Life with Computer Brains in Unity

Cartographer is a system that provides real-time simultaneous localization and mapping (SLAM) in 2D and 3D across multiple platforms and sensor configurations.

Transformer related optimization, including BERT, GPT

A Robust and Versatile Monocular Visual-Inertial State Estimator

A GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.

Go engine with no human-provided knowledge, modeled after the AlphaGo Zero paper.

MACE is a deep learning inference framework optimized for mobile heterogeneous computing platforms.

MegEngine 是一个快速、可拓展、易于使用且支持自动求导的深度学习框架

DyNet: The Dynamic Neural Network Toolkit

Motion imitation with deep reinforcement learning.

Go AI program which implements the AlphaGo Zero paper

Enabling PyTorch on XLA Devices (e.g. Google TPU)

Deep Learning Chinese Word Segment

Mitsuba 2: A Retargetable Forward and Inverse Renderer

Reinforcement Learning environments based on the 1993 game Doom

Parallel t-SNE implementation with Python and Torch wrappers.

code for Holistically-Nested Edge Detection

Torch implementation for the paper "Artistic style transfer for videos"