Google DeepMind 的公共 GitHub 足迹

This repository contains implementations and illustrative code to accompany DeepMind publications

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

Open source code for AlphaFold 2.

TensorFlow-based neural network library

AlphaFold 3 inference pipeline.

StarCraft II Learning Environment

此仓库未提供描述。

A customisable 3D platform for agent-based AI research

Gemma open-weight LLM library, from Google DeepMind

OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games.

此仓库未提供描述。

Google DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo.

Learning to Learn in TensorFlow

A library of reinforcement learning components and agents

A collection of high-quality models for the MuJoCo physics engine, curated by Google DeepMind.

JAX-based neural network library

GDM Science Skills to speed up agentic scientific workflows with better grounding and higher token efficiency. Integrate insights from AlphaGenome, AFDB, UniProt and 30+ other databases and tools.

此仓库未提供描述。

Monte Carlo tree search in JAX

Optax is a gradient processing and optimization library for JAX.

此仓库未提供描述。

An open-source library for GPU-accelerated robot learning and sim-to-real transfer.

This API provides programmatic access to the AlphaGenome model developed by Google DeepMind.

此仓库未提供描述。

Tracking Any Point (TAP)

A JAX research toolkit for building, editing, and visualizing neural networks.

Convolutional neural network model for video classification trained on the Kinetics dataset.

Real-time behaviour synthesis with MuJoCo, using Predictive Control

A library for generative social simulation

bsuite is a collection of carefully-designed experiments that investigate core capabilities of a reinforcement learning (RL) agent

GPU-optimized version of the MuJoCo physics simulator, designed for NVIDIA hardware.

此仓库未提供描述。

A collection of formalized statements of conjectures in Lean.

RL research on Android devices.

此仓库未提供描述。

Dramatron uses large language models to generate coherent scripts and screenplays.

此仓库未提供描述。

This repository contains notebook implementations of the following Neural Process variants: Conditional Neural Processes (CNPs), Neural Processes (NPs), Attentive Neural Processes (ANPs).

tree is a library for working with nested data structures

此仓库未提供描述。

A platform for managing machine learning experiments

A suite of test scenarios for multi-agent reinforcement learning.

Research code accompanying AlphaGenome

An implementation of the Fermionic Neural Network for ab-initio electronic structure calculations

此仓库未提供描述。

Accompanying code for "Discovering State-of-the-art Reinforcement Algorithms" Nature publication

TORAX: Tokamak transport simulation in JAX

On the Theoretical Limitations of Embedding-Based Retrieval

此仓库未提供描述。

TIPSv2 (CVPR'26) and TIPS (ICLR'25)

Restoring and attributing ancient texts using deep neural networks

Minimal and scalable research codebase in JAX, designed for rapid iteration on frontier research in LLM and other autoregressive models.

此仓库未提供描述。

Dataset to assess the disentanglement properties of unsupervised learning methods

DQN Zoo is a collection of reference implementations of reinforcement learning agents developed at DeepMind based on the Deep Q-Network (DQN) agent.

PIX is an image processing library in JAX, for JAX.

A Python interface for reinforcement learning environments

Official repository for "VideoPrism: A Foundational Visual Encoder for Video Understanding" (ICML 2024)

Barkour Robot: Agile Quadruped Robots by Google DeepMind

Benchmarking physical understanding in generative video models

Second Order Optimization and Curvature Estimation with K-FAC in JAX.

此仓库未提供描述。

A collection of tabletop tasks in Mujoco

此仓库未提供描述。

Lean math proofs generated by AlphaProof Nexus and accompanying natural language prose proofs.

Multi-object image datasets with ground-truth segmentation masks and generative factors.

此仓库未提供描述。

此仓库未提供描述。

Foundation models for 4D spatial and temporal vision tasks.

此仓库未提供描述。

Algorithms for Privacy-Preserving Machine Learning in JAX

此仓库未提供描述。

This repository contains the 3D shapes dataset, used in Kim, Hyunjik and Mnih, Andriy. "Disentangling by Factorising." In Proceedings of the 35th International Conference on Machine Learning (ICML). 2018. to assess the disentanglement properties of unsupervised learning methods.

此仓库未提供描述。

A tool for recording RL trajectories.

此仓库未提供描述。

A networking protocol for agent-environment communication

A large-scale NOCS dataset.

此仓库未提供描述。

AlphaGeometry2 symbolic engine (DDAR) with examples

This repository contains levels for boxoban, a box-pushing puzzle game inspired by Sokoban.

Amplio: A Lightweight Agent Harness for Robust and Long-Horizon Runs

A toolkit for building multimodal, streaming APIs and UIs

Bagz is a format for storing a sequence of string records. It supports per-record compression and fast index-based lookup.

此仓库未提供描述。

此仓库未提供描述。

此仓库未提供描述。

此仓库未提供描述。

A library of surrogate transport models for tokamak fusion.

此仓库未提供描述。

此仓库未提供描述。

此仓库未提供描述。

此仓库未提供描述。

此仓库未提供描述。

此仓库未提供描述。

Robotics Environment Authoring Framework

此仓库未提供描述。

Set of interfaces for Python reinforcement learning (RL) environments

此仓库未提供描述。

此仓库未提供描述。