Spaces:
Running
Running
File size: 4,721 Bytes
ab7d699 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 |
import torch
from torch.optim.optimizer import Optimizer, required
class AdaiS(Optimizer):
r"""Implements Adai with stable/decoupled weight decay (AdaiS/AdaiW).
It is based on
`Adai: Separating the Effects of Adaptive Learning Rate and Momentum Inertia`
and
`Stable Weight Decay Regularization`__.
Arguments:
params (iterable): iterable of parameters to optimize or dicts defining
parameter groups
lr (float, optional): learning rate
betas (Tuple[float, float], optional): beta0 and beta2 (default: (0.1, 0.99))
eps (float, optional): the inertia bound (default: 1e-03)
weight_decay (float, optional): weight decay (default: 0)
"""
def __init__(self, params, lr=required, betas=(0.1, 0.99), eps=1e-03,
weight_decay=0):
if lr is not required and lr < 0.0:
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]:
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]))
if not 0.0 <= weight_decay:
raise ValueError("Invalid weight_decay value: {}".format(weight_decay))
defaults = dict(lr=lr, betas=betas, eps=eps, weight_decay=weight_decay)
super(AdaiS, self).__init__(params, defaults)
def __setstate__(self, state):
super(AdaiS, self).__setstate__(state)
@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:
loss = closure()
param_size = 0
exp_avg_sq_hat_sum = 0.
for group in self.param_groups:
for p in group['params']:
if p.grad is None:
continue
param_size += p.numel()
grad = p.grad.data
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.data, memory_format=torch.preserve_format)
# Exponential moving average of squared gradient values
state['exp_avg_sq'] = torch.zeros_like(p.data, memory_format=torch.preserve_format)
# Cumulative products of beta1
state['beta1_prod'] = torch.ones_like(p.data, memory_format=torch.preserve_format)
exp_avg_sq = state['exp_avg_sq']
beta0, beta2 = group['betas']
state['step'] += 1
bias_correction2 = 1 - beta2 ** state['step']
exp_avg_sq.mul_(beta2).addcmul_(1 - beta2, grad, grad)
exp_avg_sq_hat = exp_avg_sq / bias_correction2
exp_avg_sq_hat_sum += exp_avg_sq_hat.sum()
# Calculate the mean of all elements in exp_avg_sq_hat
exp_avg_sq_hat_mean = exp_avg_sq_hat_sum / param_size
for group in self.param_groups:
for p in group['params']:
if p.grad is None:
continue
grad = p.grad.data
# Perform stable/decoupled weight decay
if group['weight_decay'] !=0:
p.data.mul_(1 - group['lr'] * group['weight_decay'])
state = self.state[p]
exp_avg = state['exp_avg']
exp_avg_sq = state['exp_avg_sq']
beta0, beta2 = group['betas']
beta1_prod = state['beta1_prod']
bias_correction2 = 1 - beta2 ** state['step']
exp_avg_sq_hat = exp_avg_sq / bias_correction2
beta1 = (1. - (exp_avg_sq_hat / exp_avg_sq_hat_mean).mul(beta0)).clamp(0., 1 - group['eps'])
beta1_prod.mul_(beta1)
bias_correction1 = 1 - beta1_prod
exp_avg.mul_(beta1).addcmul_(1 - beta1, grad)
exp_avg_hat = exp_avg.div(bias_correction1)
step_size = group['lr']
p.data.add_(-step_size, exp_avg_hat)
return loss
|