Kaiyotech

In the rewards I punish switching actions to encourage sticking with one submodel (this was your idea, lol). I think that may or may not have value for a normal...

Sort of. Mine is a parameter I can turn on and off, the version in main is always deterministic if streaming I think.

Yes. Mine can force all old agents to be deterministic instead of playing against a mix of both types. On Tue, Dec 6, 2022, 6:16 PM Rolv-Arild ***@***.***> wrote: >...

Conclusion is I got busy and didn't finish it, but it's still on my list. I'm going to take the suggestions, just haven't finished yet. On Tue, Dec 6, 2022,...

Ok, this is ready and tested. Uses the EMA for the weights, per worker. Generated experience is per actual, which means that if you're using pretrained agents or past models,...

added one commit for the 1v0 fixes that is related to this.

Probably On Sun, Apr 16, 2023, 7:23 AM Rolv-Arild ***@***.***> wrote: > Should probably reset gravity as well? > > — > Reply to this email directly, view it on...