suggestion

[m1996]

hi.thanks a lot for sharing your code its very nice and clean and readable.
i want to run it in mountain car discrete environment, any suggestion for parameters or networks to get better results?
regards

[p-christ]

Not sure but maybe try the ones in the results/cart_pole.py file

…

On Mon, Jan 27, 2020 at 15:08, m1996 ***@***.***> wrote: hi.thanks a lot for sharing your code its very nice and clean and readable. i want to run it in mountain car discrete environment, any suggestion for parameters or networks to get better results? regards — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#39?email_source=notifications&email_token=AGJAGA5W2GDIVZQGAKMJLS3Q732INA5CNFSM4KMC4C32YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4II6LQEA>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGJAGAZIW6AZ7JJBMKIJTUDQ732INANCNFSM4KMC4C3Q> .

[m1996]

Thanks for your reply, we have tried that but the result was not as good as cartpole(maybe because of harder exploration?)do you have any suggestion what should we do? Or how should we find best paramerers?

[p-christ]

Mmm yeah try increasing exploration by changing the epsilon_decay_rate_denominator

…

On Mon, Jan 27, 2020 at 22:17, m1996 ***@***.***> wrote: Thanks for your reply, we have tried that but the result was not as good as cartpole(maybe because of harder exploration?)do you have any suggestion what should we do? Or how should we find best paramerers? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#39?email_source=notifications&email_token=AGJAGA7TDLOV7YHQIYP7DB3Q75MPHA5CNFSM4KMC4C32YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKBI7XQ#issuecomment-578981854>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGJAGAY4TEVKKX2M2UFF4TDQ75MPHANCNFSM4KMC4C3Q> .