gangsuuga / stagelight Goto Github PK

In this code, It seemed training is called after a simulation ended. I think the model trains after an episode ended. Did you used MC method?

Should "if ( _ / update_epochs == 0):" be "if ( _ % update_epochs == 0):"?

Can you share you paper's download url? I can't find the paper in google scholar. Thanks!

I can't understand the input_state. In reference code, whether car exists in specific cell was the state.(exist:1 non_exist:0) but in your code there is car_speed and accumulated_waiting_time. I wonder what the state means. (It is declared in numpy (3,14,100) .and, I can't understand what does it mean)

I think there are some differences in your utils.py code with utils.py in reference, how can i run this code i want to know the requirements of this source code.