Github: tianshou,包含了源码和使用介绍。
Tianshou (天授) is a reinforcement learning platform based on pure PyTorch and Gymnasium. Unlike other reinforcement learning libraries, which may have complex codebases, unfriendly high-level APIs, or are not optimized for speed, Tianshou provides a high-performance, modularized framework and user-friendly interfaces for building deep reinforcement learning agents. One more aspect that sets Tianshou apart is its generality: it supports online and offline RL, multi-agent RL, and model-based algorithms.
具体的安装说明参见其Github: tianshou。
注意,注意,注意!!!天授需要的Python版本必须是3.11以上,包括3.11,否则直接报错。在github的安装部分说了:It requires Python >= 3.11
。
不要试图在小于3.11的Python版本安装!!!,因为我帮你已经踩过了这个坑了!如果你本地不能装3.11版本的python和pytorch,那可以找个云主机(比如AutoDL)装。
先安装poetry:
pip install poetry
然后
git clone https://github.com/thu-ml/tianshou.git
cd tianshou
poetry install
测试是否安装天授成功:
import tianshou
print(tianshou.__version__)
我们找到里面的ppo的例子,来试着运行mujoco_ppo.py:
https://github.com/thu-ml/tianshou/blob/master/examples/mujoco/mujoco_ppo.py
按照其README.md:
python mujoco_sac.py --task Ant-v3
发现竟然报错了:
是因为没有装mujoco和mujoco-py,那就把这两个装上,装好后再运行就成功了,如下图