The PPO documentation says that the data should be serialized to train the algo. Is the data serialized somewhere in the environment or should I do it myself?
Then another thing. How is the dataframe processed, from 0 to 5000 or from 5000 to 0? I get my data so that it ends in the first line with the current date. So I would have to reverse the data records once so that the data is transferred to the environment in the correct order. Or am I misunderstanding something here?
Would be nice if someone could help me out here.
The PPO documentation says that the data should be serialized to train the algo. Is the data serialized somewhere in the environment or should I do it myself?
Then another thing. How is the dataframe processed, from 0 to 5000 or from 5000 to 0? I get my data so that it ends in the first line with the current date. So I would have to reverse the data records once so that the data is transferred to the environment in the correct order. Or am I misunderstanding something here?
Would be nice if someone could help me out here.