conda create -n grpo-robot -c conda-forge python=3.7 pybox2d
conda activate grpo-robot
pip install numpy<1.19.5 gym<=0.25.2 torch matplotlib pygame seaborn tqdm
# 训练
python ./ppo_train_cartpole.py
# 将训练好的策略模型可视化
python ./强化学习过程可视化-cartpole.py
| Name | Name | Last commit date | ||
|---|---|---|---|---|
conda create -n grpo-robot -c conda-forge python=3.7 pybox2d
conda activate grpo-robot
pip install numpy<1.19.5 gym<=0.25.2 torch matplotlib pygame seaborn tqdm
# 训练
python ./ppo_train_cartpole.py
# 将训练好的策略模型可视化
python ./强化学习过程可视化-cartpole.py