This is an implementation of the algorithm described in https://arxiv.org/pdf/1811.00164. It does not work at the moment. Regrets jump and do not decrease, the strategy will also tell at each iteration. Also, the strategy after >2000 iterations of training on 4bb stacks and 2 players in no-limit hold'em is far from GTO. If anyone has any ideas for fixing it, I will listen once. My mail is javay999@gmail.com