Mobile Deep Learning with TensorFlow Lite, ML Kit and Flutter

Generating sample gameplays

So far, we've talked about the util/gameplay.py file to demonstrate the code in the file that pertains to move-making rules – specifically, the self-play function of the file. Now, we'll take a look at how these self-plays can run in iteration to generate a complete gameplay log. Let's get started:

1. Consider the code of the generate() method of the Generator class provided by this file: 

def generate(self, state, nn, cb, args):    self.mcts = MCTS(nn)    iterator = range(args['selfplay'])    if args['progress']:        from tqdm import tqdm        iterator = tqdm(iterator, ncols = 50)    # self play    for pi in iterator:        result = self._selfplay(state, args)        if cb != None:            cb(result)

Essentially, this function takes care of running the ...