强化学习TD3算法笔记2——代码解读并基于colab平台复现

相关github地址

代码解读的代码链接：sfujim_TD3
colab运行代码：TD3-PyTorch-BipedalWalker-v2
可视化代码：colabgymrender

代码解读

Action、Critic网络定义

# Action网络定义
self.actor = Actor(state, action_dim, max_action).to(device)
# 将action的参数拷贝给target action
self.actor_target = copy.deepcopy(self.action)
self.actor_optimizer = torch.optim.Adam(self.actor.parameters(), lr=3e-4)

# Critic网络定义
self.critic = Critic(state, action).to(device)
self.critic_target = copy.deepcopy(self.critic)
self.critic_optimizer = torch.optim.Adam(self.critic.parameters(), lr=3e-4)

实现smooth_regularization

noise = (torch.randn_like(action)*self.policy_noise) # policy_noise = 0.2</