Timeline for When is it necessary to explicitly define both the state and observation space in a custom environment?
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 2, 2023 at 9:19 | comment | added | AlphaBit95 | Noted. Thank you very much! | |
| Feb 1, 2023 at 16:30 | comment | added | pythonic833 |
The input to your network should be the states at a given time, not the whole state space. So in your case the inputs should be the concrete values of l and b as far as I can read from your question.
|
|
| Feb 1, 2023 at 15:42 | comment | added | AlphaBit95 | Hey! Thanks for the clarification! It helped me understand better. I was mainly concerned with this as I'm trying to train a DDPG model for such a scenario. I was really confused if my input states to the Actor-network would be the state space or the observation space(as most of the papers and blogs use the observation space). Since they are the same, would you say that my input to the network would be my state space itself? | |
| S Jan 31, 2023 at 12:32 | review | First answers | |||
| Feb 2, 2023 at 8:17 | |||||
| S Jan 31, 2023 at 12:32 | history | answered | pythonic833 | CC BY-SA 4.0 |