DeepMind releases Acme: A library of reinforcement learning components and agents

DeepMind has recently released Acme, a library with an objective to simplify the development of reinforcement learning algorithms and agent building blocks. This application can be run at various scales of execution and it is achieved by enabling AI-driven agents to enable simple agent implementations. Acme can be used to create agents with greater parallelization than in previous approaches as per reports. This tool can be used by researchers to reproduce published RL algorithms or rapidly prototype ideas. Acme aims to make the results of various reinforcement learning (RL) algorithms developed in academia and industrial labs easier to reproduce and extend.

Acme strives to bring simple, efficient, and readable agents, that serve both as reference implementations of popular algorithms and as strong baselines, while still providing enough flexibility to do novel research. The design of Acme also attempts to provide multiple points of entry to the RL problem at differing levels of complexity.




To install acme core:

# Install Acme core dependencies.
pip install dm-acme

# Install Reverb, our replay backend.
pip install dm-acme[reverb]