• https://github.com/PacktPublishing/Deep-Reinforcement-Learning-Hands-On
• https://datascience.stackexchange.com/questions/49625/ppo-a2c-for-continuous-action-spaces-math-and-code

def cont_logprob(mu, var, actions):
    import torch
    import math
    p1 = - ((mu - actions) ** 2) / (2 * var.clamp(min=1e-3))
    p2 = - torch.log(torch.sqrt(2 * math.pi * var))
    return (p1 + p2).sum(-1, keepdims=True)
 
 
def cont_entropy(var):
    import torch
    import math
    entropy = (torch.log(2 * math.pi * var) + 1) / 2
    return entropy.sum(-1)

1. TD3: https://towardsdatascience.com/td3-learning-to-run-with-ai-40dfc512f93