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

JAX-based neural network library

TensorFlow-based neural network library

Optax is a gradient processing and optimization library for JAX.

A library for generative social simulation

Tracking Any Point (TAP)

StarCraft II Learning Environment

A customisable 3D platform for agent-based AI research

Learning to Learn in TensorFlow

🦦 Otter, a multi-modal model based on OpenFlamingo (open-sourced version of DeepMind's Flamingo), trained on MIMIC-IT and showcasing improved instruction-following and in-context learning ability.

DeepMind's Tacotron-2 Tensorflow implementation

google-deepmind/gemini-robotics-sdk is a high-star GitHub project relevant to AI agent workflows.