/examples/pytorch/transformer/optims/noamopt.py

class NoamOpt(object):
    def __init__(self, model_size, factor, warmup, optimizer):
        """
        model_size: hidden size
        factor: coefficient
        warmup: warm up steps(step ** (-0.5) == step * warmup ** (-1.5) holds when warmup equals step)
        """
        self.optimizer = optimizer
        self._step = 0
        self.warmup = warmup
        self.factor = factor
        self.model_size = model_size
        self._rate = 0

    def rate(self, step=None):
        if step is None:
            step = self._step
        return self.factor * \
            (self.model_size ** (-0.5) *
             min(step ** (-0.5), step * self.warmup ** (-1.5))
            )

    def step(self):
        self._step += 1
        rate = self.rate()
        for p in self.optimizer.param_groups:
            p['lr'] = rate
        self._rate = rate
        self.optimizer.step()

"""
Default setting:

def get_std_opt(model):
    return NoamOpt(model.src_embed[0].d_model, 2, 4000,
            torch.optim.Adam(model.parameters(), lr=0, betas=(0.9, 0.98), eps=1e-9))
"""