rlcard
rlcard copied to clipboard
datamllab
Reinforcement Learning / AI Bots in Card (Poker) Games - Blackjack, Leduc, Texas, DouDizhu, Mahjong, UNO.
Support bet with any amount. Set min raise to double of previous raise
问题描述：目前dmc算法斗地主的水平刚好能达到人类的平均水平。在某些情况下，AI出牌会出现让人惊艳的出牌（超高手水准），反之AI偶尔也会犯十分低级的错误。比如（AI手上剩３个牌，一个对子，一个单张，在对子明显对方要不起的情况下，结果AI出了一个单张从而导致输牌）。 个人理解：出现上段描述中低级错误是因为某些牌局AI并没有得到足够多的训练，而按照目前的算法（代码）分析，如果需要基本杜绝低级错误的出现，目前的训练量至少要再提高一个数量级，而这肯定是不现实的。 解决思路： 1. DMCAgent类中参数exp_epsilon默认值设为0.01，改变此参数的值能否提高训练效率（个人认为针对斗地主的情况，最佳的设定值在0.05~0.1之间）? 2. 或者能否通过将step函数设计得更加精细来提高训练效率。比如action的选择可以根据过去训练结果的得分概率来选择(当然必须做一定的调整)。而不是简单的在概率为exp_epsilon的情况下随机选择。
In obs describiton, it says that : > 53 | Chips that all players have put in however, in line 70 of [nolimitholdem.py](https://github.com/datamllab/rlcard/blob/master/rlcard/envs/nolimitholdem.py) obs[53] = float(max(all_chips)) it use **max** rather...
--------------------------------------------------------------------------- NameError Traceback (most recent call last) in 1 env = rlcard.make('no-limit-holdem') 2 trainer = DMCTrainer(env,total_frames=6400) ----> 3 trainer.start() /data/miniconda3/lib/python3.6/site-packages/rlcard/agents/dmc_agent/trainer.py in start(self) 298 299 checkpoint(frames) --> 300 plogger.close() NameError: name...
I'd like to train DMC agents against random agents, i.e., DMC agents and random agents exist simultaneously during training. Is that feasible using RLCard? If yes, will there be many...
Since you already have tests set up for your code, could you please add CI testing to have them automatically run on new PRs with GitHub actions? It's fairly easy...
For a course project I am trying to implement Reinforcement Learning to play Dungeon Mayhem. It is a multiplayer strategy card game where each player has a different deck with...
Hello, I am using this: Python 3.8.6 [GCC 10.2.0] on linux python3 -c "import tensorflow as tf;print(tf.__version__)" 2.7.0 nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2020 NVIDIA...
is it possible to add betting amounts for any of these games so it can be more accurate specify, Texas Holdem.
Hi, I found a possible problem about score counting within the blackjack environment. https://github.com/datamllab/rlcard/blob/f256657dd13039bd707e6d95ea3c31795b573c76/rlcard/envs/blackjack.py#L38-L71 Consider the case that there are more than one 'Ace' in `my_cards` (i.e. the player's hand...
