|
|
import torch |
|
|
from torch.optim.optimizer import Optimizer, required |
|
|
|
|
|
|
|
|
class LARS(Optimizer): |
|
|
r"""Implements layer-wise adaptive rate scaling for SGD, based on |
|
|
`"Large Batch Training of Convolutional Networks" <https://arxiv.org/abs/1708.03888>`_ |
|
|
|
|
|
Arguments: |
|
|
- params (:obj:`iterable`): iterable of parameters to optimize or dicts defining parameter groups |
|
|
- lr (:obj:`float`): learning rate |
|
|
- momentum (:obj:`float`, optional): momentum factor (default: 0) |
|
|
- weight_decay (:obj:`float`, optional): weight decay (L2 penalty) (default: 0) |
|
|
- dampening (:obj:`float`, optional): dampening for momentum (default: 0) |
|
|
- eta(:obj:`float`): LARS coefficient (default 0.001) |
|
|
- nesterov (:obj:`bool`, optional): enables Nesterov momentum (default: False) |
|
|
|
|
|
Example: |
|
|
>>> optimizer = LARS(model.parameters(), lr=0.1, momentum=0.9, eta=1e-3) |
|
|
>>> optimizer.zero_grad() |
|
|
>>> loss_fn(model(input), target).backward() |
|
|
>>> optimizer.step() |
|
|
""" |
|
|
|
|
|
def __init__(self, params, lr=required, momentum=0, dampening=0, |
|
|
weight_decay=0, eta=0.001, nesterov=False): |
|
|
if lr is not required and lr < 0.0: |
|
|
raise ValueError("Invalid learning rate: {}".format(lr)) |
|
|
if momentum < 0.0: |
|
|
raise ValueError("Invalid momentum value: {}".format(momentum)) |
|
|
if weight_decay < 0.0: |
|
|
raise ValueError( |
|
|
"Invalid weight_decay value: {}".format(weight_decay)) |
|
|
if eta < 0.0: |
|
|
raise ValueError("Invalid LARS coefficient value: {}".format(eta)) |
|
|
|
|
|
defaults = dict(lr=lr, momentum=momentum, dampening=dampening, |
|
|
weight_decay=weight_decay, eta=eta, nesterov=nesterov) |
|
|
if nesterov and (momentum <= 0 or dampening != 0): |
|
|
raise ValueError( |
|
|
"Nesterov momentum requires a momentum and zero dampening") |
|
|
super(LARS, self).__init__(params, defaults) |
|
|
|
|
|
def __setstate__(self, state): |
|
|
super(LARS, self).__setstate__(state) |
|
|
for group in self.param_groups: |
|
|
group.setdefault('nesterov', False) |
|
|
|
|
|
@torch.no_grad() |
|
|
def step(self, closure=None): |
|
|
"""Performs a single optimization step. |
|
|
|
|
|
Arguments: |
|
|
- closure (:obj:`callable`, optional): A closure that reevaluates the model and returns the loss. |
|
|
""" |
|
|
loss = None |
|
|
if closure is not None: |
|
|
loss = closure() |
|
|
|
|
|
for group in self.param_groups: |
|
|
weight_decay = group['weight_decay'] |
|
|
momentum = group['momentum'] |
|
|
dampening = group['dampening'] |
|
|
nesterov = group['nesterov'] |
|
|
eta = group['eta'] |
|
|
|
|
|
for p in group['params']: |
|
|
if p.grad is None: |
|
|
continue |
|
|
d_p = p.grad.data |
|
|
|
|
|
|
|
|
weight_norm = torch.norm(p.data) |
|
|
grad_norm = torch.norm(d_p) |
|
|
|
|
|
if weight_decay != 0: |
|
|
d_p.add_(weight_decay, p.data) |
|
|
grad_norm.add_(weight_decay, weight_norm) |
|
|
local_lr = eta * weight_norm / grad_norm |
|
|
|
|
|
if momentum != 0: |
|
|
param_state = self.state[p] |
|
|
if 'momentum_buffer' not in param_state: |
|
|
buf = param_state['momentum_buffer'] = torch.zeros_like( |
|
|
p.data) |
|
|
buf.mul_(momentum).add_(d_p) |
|
|
else: |
|
|
buf = param_state['momentum_buffer'] |
|
|
buf.mul_(momentum).add_(1 - dampening, d_p) |
|
|
if nesterov: |
|
|
d_p = d_p.add(momentum, buf) |
|
|
else: |
|
|
d_p = buf |
|
|
|
|
|
p.data.add_(-group['lr']*local_lr, d_p) |
|
|
|
|
|
return loss |
|
|
|
