File size: 1,095 Bytes
3cd1076 | 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 | #!/usr/bin/env python
# -*- coding: utf-8 -*-
import torch
from torch.optim.optimizer import Optimizer
class AdaGrad(torch.optim.Optimizer):
def __init__(self, params, lr=0.001, eps=1e-8, weight_decay=0):
defaults = dict(lr=lr, eps=eps, weight_decay=weight_decay)
super(AdaGrad, self).__init__(params, defaults)
def step(self, closure=None):
loss = None
if closure is not None:
loss = closure()
for group in self.param_groups:
for p in group['params']:
if p.grad is None:
continue
grad = p.grad.data
if group['weight_decay'] != 0:
grad = grad.add(group['weight_decay'], p.data)
state = self.state[p]
if len(state) == 0:
state['sum'] = torch.zeros_like(p.data)
sum = state['sum']
sum.addcmul_(1, grad, grad)
std = sum.sqrt().add_(group['eps'])
p.data.addcdiv_(-group['lr'], grad, std)
return loss |