清空筛选条件
清空筛选条件

evaluatePolicy.m output differs from Agent action

1 次查看（过去 30 天）
Victor Bayer
Victor Bayer 2021-7-11
编辑： Victor Bayer 2021-9-22
Dear Mathworks Team,
I have a Matlab defined RL-algorithm with an DDPG-agent which recieves two observations can chose values between 0 and 1.
This is defined through:
numAct = 1;
actionInfo = rlNumericSpec([numAct 1],'LowerLimit',0 ,'UpperLimit', 1);
actionInfo.Name = 'sine_amplitude';
During training only values between 0 and 1 are applied. The action is clipped at those values and the actionInfo is repected.
However when I use the generated Agent to generate a Policy according to Matlab (see https://www.mathworks.com/help/reinforcement-learning/ref/rl.agent.rldqnagent.generatepolicyfunction.html)
and evaluate the function I also recieve negative values.
For example
evaluatePolicy(reshape([-0.1515581,1],2,1,1))
returns -1
I have used reshape-functions to reshape the data [-0.1515581,1] to the corresponding shape since the function expects the input to be of shape (2,1,1).
My question is why and how can i change generation of the evaluatePolicy function?

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by