Skip to content

Release v1.0

Choose a tag to compare

@Ronchy2000 Ronchy2000 released this 31 Jul 07:59
· 49 commits to master since this release
aa526a5

多模块独立发布包

模块介绍

  • 🚀 MADDPG_Continous
    多智能体深度确定性策略梯度算法(连续动作空间版本),适用于连续控制场景的多智能体协同训练。

  • ⚡️ MATD3_Continous
    多智能体双延迟深度确定性策略梯度算法,在MADDPG基础上增加了延迟更新和策略平滑机制。

  • 📚 RL_Learning-main
    赵世钰老师强化学习基础教程合集,包含经典算法实现和示例代码。

  • hands_on_RL
    动手学强化学习实践项目,通过Jupyter Notebook提供互动式学习体验。

使用说明

  1. 点击下方所需模块的ZIP文件下载
  2. 解压后阅读README.md获取详细使用指南
  3. 安装依赖:pip install -r requirements.txt

支持渠道


Independent Modules Release

Available Modules

  • MADDPG_Continous
    Multi-Agent Deep Deterministic Policy Gradient (continuous action space version) for cooperative multi-agent control.

  • ️ MATD3_Continous
    Multi-Agent Twin Delayed DDPG, featuring delayed updates and policy smoothing.

  • 📚 RL_Learning-main
    Fundamental RL tutorials with classic algorithm implementations.

  • 🧠 hands_on_RL
    Interactive reinforcement learning projects via Jupyter Notebooks.

Quick Start

  1. Download the desired module ZIP below
  2. Check README_en.md for detailed instructions
  3. Install dependencies: pip install -r requirements.txt

Support