Lamb optimizer that doesn't work on TPUs

lamb.py

class Lamb(Optimizer):
    r"""Implements Lamb algorithm.
    It has been proposed in `Large Batch Optimization for Deep Learning: Training BERT in 76 minutes`_.
    Arguments:
        params (iterable): iterable of parameters to optimize or dicts defining
            parameter groups
        lr (float, optional): learning rate (default: 1e-3)
        betas (Tuple[float, float], optional): coefficients used for computing
            running averages of gradient and its square (default: (0.9, 0.999))
        eps (float, optional): term added to the denominator to improve
            numerical stability (default: 1e-8)
        weight_decay (float, optional): weight decay (L2 penalty) (default: 0)
        adam (bool, optional): always use trust ratio = 1, which turns this into
            Adam. Useful for comparison purposes.
    .. _Large Batch Optimization for Deep Learning: Training BERT in 76 minutes:
        https://arxiv.org/abs/1904.00962

    Source: https://github.com/cybertronai/pytorch-lamb

    MIT License

    Copyright (c) 2019 cybertronai

    Permission is hereby granted, free of charge, to any person obtaining a copy
    of this software and associated documentation files (the "Software"), to deal
    in the Software without restriction, including without limitation the rights
    to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
    copies of the Software, and to permit persons to whom the Software is
    furnished to do so, subject to the following conditions:

    The above copyright notice and this permission notice shall be included in all
    copies or substantial portions of the Software.

    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
    AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
    OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
    SOFTWARE.
    """

    def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-6,
                 weight_decay=0, adam=False):
        if not 0.0 <= lr:
            raise ValueError("Invalid learning rate: {}".format(lr))
        if not 0.0 <= eps:
            raise ValueError("Invalid epsilon value: {}".format(eps))
        if not 0.0 <= betas[0] < 1.0:
            raise ValueError("Invalid beta parameter at index 0: {}".format(betas[0]))
        if not 0.0 <= betas[1] < 1.0:
            raise ValueError("Invalid beta parameter at index 1: {}".format(betas[1]))
        defaults = dict(lr=lr, betas=betas, eps=eps,
                        weight_decay=weight_decay)
        self.adam = adam
        super(Lamb, self).__init__(params, defaults)

    @torch.no_grad()
    def step(self, closure=None):
        """Performs a single optimization step.
        Arguments:
            closure (callable, optional): A closure that reevaluates the model
                and returns the loss.
        """
        loss = None
        if closure is not None:
            with torch.enable_grad():
                loss = closure()

        for group in self.param_groups:
            for p in group['params']:
                if p.grad is None:
                    continue
                grad = p.grad
                if grad.is_sparse:
                    raise RuntimeError('Lamb does not support sparse gradients, consider SparseAdam instead.')

                state = self.state[p]

                # State initialization
                if len(state) == 0:
                    state['step'] = 0
                    # Exponential moving average of gradient values
                    state['exp_avg'] = torch.zeros_like(p, memory_format=torch.preserve_format)
                    # Exponential moving average of squared gradient values
                    state['exp_avg_sq'] = torch.zeros_like(p, memory_format=torch.preserve_format)

                exp_avg, exp_avg_sq = state['exp_avg'], state['exp_avg_sq']
                beta1, beta2 = group['betas']

                state['step'] += 1

                # Decay the first and second moment running average coefficient
                # m_t
                exp_avg.mul_(beta1).add_(grad, alpha=1 - beta1)
                # v_t
                exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1 - beta2)

                step_size = group['lr']

                weight_norm = p.pow(2).sum().sqrt().clamp(0, 10)

                adam_step = exp_avg / exp_avg_sq.sqrt().add(group['eps'])
                if group['weight_decay'] != 0:
                    adam_step.add_(p, alpha=group['weight_decay']) 

                adam_norm = adam_step.pow(2).sum().sqrt()
                if weight_norm == 0. or adam_norm == 0.:
                    trust_ratio = 1.
                else:
                    trust_ratio = weight_norm / adam_norm
                state['weight_norm'] = weight_norm
                state['adam_norm'] = adam_norm
                state['trust_ratio'] = trust_ratio
                if self.adam:
                    trust_ratio = 1.

                p.add_(adam_step, alpha=-step_size * trust_ratio)

        return loss