gym-locm

gym-locm copied to clipboard

`LOCMEnv` sometimes returns `None` instead of a numerical state representation

1

Hey, the function `LOCMEnv.encode_state` returns `None` when `self.state.phase` is neither `Phase.DECK_BUILDING` nor `Phase.BATTLE`. This happens precisely once in every game. When `self.state.phase == Phase.ENDED`, neither of the conditions is true...

Rattko

When experimenting with Offline RL, one needs to create a dataset containing (state, action) pairs to pass to a chosen algorithm. In this case, action usually means an integer between...

Rattko

[Question] Obtaining numerical state for an agent

2

Hey, I've been working with LOCM for the past couple of weeks and I am at the point now where I would like to use my trained agents in the...

Rattko

Merge bugfixes into main repo

1

ronaldosvieira

ronaldosvieira

ronaldosvieira

ronaldosvieira

Add support for LOCM 1.5

2

ronaldosvieira

ronaldosvieira

Cannot fully reproduce the Coac vs Chad winrate as reported in CEC2020 using NativeAgent

1

The Coac vs Chad winrate was [reported to be 57% in CEC 2020](https://github.com/acatai/Strategy-Card-Game-AI-Competition/blob/ea138d9fb6b556356705f20b9570283d99b0fccc/contest-2020-07-CEC/graph.data#L2553), but I obtained a winrate ~= 80% using `locm-runner` and `NativeAgent`. The evaluation code is: ```Bash locm-runner...

wanhesong