Morris

关于actor多维连续动作值的概率密度构建

2

莫烦：您好 请问如果actor输出多维连续动作值，那么还能用函数tf.distributions.Normal构建多维概率密度吗？如果能，那么函数方法prob_log输出tensor的维度与样本维度一致，即不能与标量retrun相乘。请问该如何解决这个问题？ 谢谢

TypeError: unhashable type: 'State'

7

Dear Mr/Miss: I have implemented your code, and something wrong happened: TypeError: unhashable type: 'State' Could you please tell me how does this happen and how can I fix it?...

Questions about class SARSA(lambda)

2

Hello. I have a few questions. 1, What the effect of variable "shrink" in the class "SarsaLambdaAgent"? And Can I use other basis instead, like the polynomial basis? 2, Why...

Some meaning about coefficients

4

What the meaning of the coefficients n, b, and d in the BchCodeGenerator class?

Hello, thank you for your contribution. What I wonder is the propose of your code. Is the propose focusing on binary classification problem? Thank you.

As the title, in run2 function, should the optimization be add ahead of updating of target network?