Gold is a reinforcement learning library for Go, offering a suite of agents for solving challenges in various environments. It targets Go developers seeking to implement RL algorithms, providing composable tooling for agent creation and performance visualization.

How It Works

Gold implements RL agents using Go, leveraging the Gorgonia project for neural network models and Sphere for environment interaction. This Go-centric approach aims to provide a performant and memory-efficient RL framework without the typical overhead of Python-based libraries.

Quick Start & Requirements

• Install/Run: go run ./pkg/v1/agent/deepq/experiments/cartpole/main.go
• Requirements: Go >= v13.0, Docker, modern browser.
• Docs: GoDoc

Highlighted Details

• Implements Q Learning, Deep Q Learning (Double Q), REINFORCE, Natural Evolution Strategies, Hindsight Experience Replay, and Proximal Policy Optimization (in progress).
• Integrates with Gorgonia for neural network models and Sphere for environments.
• Offers agent-specific demos and experiment folders.

Maintenance & Community

• Open to contributions via Merge Requests.
• Contact: @pbarker on Gopher Slack.
• Roadmap includes more agents, environments, accelerated compute, tuning libraries, and deployments.

Licensing & Compatibility

• License: Not explicitly stated in the README.

Limitations & Caveats

• Proximal Policy Optimization (PPO) is marked as "in progress."
• The license is not specified, which may impact commercial use or closed-source integration.