CosMo.pytorch/optimizers.py at main · postBG/CosMo.pytorch · GitHub

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
122
123
124
125
126
127
128
import math

import torch
from torch.optim import Adam, SGD
from torch.optim.optimizer import Optimizer
from torch.optim.lr_scheduler import StepLR


def create_optimizers(models, config):
    optimizer = config['optimizer']
    lr = config['lr']
    weight_decay = config['weight_decay']
    momentum = config['momentum']

    parameterized_models = {key: model for key, model in models.items() if len(list(model.parameters())) > 0}

    optimizers = {}
    if optimizer == 'Adam':
        optimizers = {key: Adam(model.parameters(), lr=lr, weight_decay=weight_decay)
                      for key, model in parameterized_models.items()}
    elif optimizer == 'SGD':
        optimizers = {key: SGD(model.parameters(), lr=lr, weight_decay=weight_decay, momentum=momentum)
                      for key, model in parameterized_models.items()}
    elif optimizer == 'RAdam':
        optimizers = {key: RAdam(model.parameters(), lr=lr, weight_decay=weight_decay)
                      for key, model in parameterized_models.items()}
    return optimizers


def create_lr_schedulers(optimizers, config):
    decay_step = config['decay_step']
    gamma = config['gamma']
    return {key: StepLR(optimizer, step_size=decay_step, gamma=gamma) for key, optimizer in optimizers.items()}


class RAdam(Optimizer):

    def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8, weight_decay=0, degenerated_to_sgd=True):
        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]))

        self.degenerated_to_sgd = degenerated_to_sgd
        if isinstance(params, (list, tuple)) and len(params) > 0 and isinstance(params[0], dict):
            for param in params:
                if 'betas' in param and (param['betas'][0] != betas[0] or param['betas'][1] != betas[1]):
                    param['buffer'] = [[None, None, None] for _ in range(10)]
        defaults = dict(lr=lr, betas=betas, eps=eps, weight_decay=weight_decay,
                        buffer=[[None, None, None] for _ in range(10)])
        super(RAdam, self).__init__(params, defaults)

    def __setstate__(self, state):
        super(RAdam, self).__setstate__(state)

    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.float()
                if grad.is_sparse:
                    raise RuntimeError('RAdam does not support sparse gradients')

                p_data_fp32 = p.data.float()

                state = self.state[p]

                if len(state) == 0:
                    state['step'] = 0
                    state['exp_avg'] = torch.zeros_like(p_data_fp32)
                    state['exp_avg_sq'] = torch.zeros_like(p_data_fp32)
                else:
                    state['exp_avg'] = state['exp_avg'].type_as(p_data_fp32)
                    state['exp_avg_sq'] = state['exp_avg_sq'].type_as(p_data_fp32)

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

                exp_avg_sq.mul_(beta2).addcmul_(1 - beta2, grad, grad)
                exp_avg.mul_(beta1).add_(1 - beta1, grad)

                state['step'] += 1
                buffered = group['buffer'][int(state['step'] % 10)]
                if state['step'] == buffered[0]:
                    N_sma, step_size = buffered[1], buffered[2]
                else:
                    buffered[0] = state['step']
                    beta2_t = beta2 ** state['step']
                    N_sma_max = 2 / (1 - beta2) - 1
                    N_sma = N_sma_max - 2 * state['step'] * beta2_t / (1 - beta2_t)
                    buffered[1] = N_sma

                    # more conservative since it's an approximated value
                    if N_sma >= 5:
                        step_size = math.sqrt(
                            (1 - beta2_t) * (N_sma - 4) / (N_sma_max - 4) * (N_sma - 2) / N_sma * N_sma_max / (
                                    N_sma_max - 2)) / (1 - beta1 ** state['step'])
                    elif self.degenerated_to_sgd:
                        step_size = 1.0 / (1 - beta1 ** state['step'])
                    else:
                        step_size = -1
                    buffered[2] = step_size

                # more conservative since it's an approximated value
                if N_sma >= 5:
                    if group['weight_decay'] != 0:
                        p_data_fp32.add_(-group['weight_decay'] * group['lr'], p_data_fp32)
                    denom = exp_avg_sq.sqrt().add_(group['eps'])
                    p_data_fp32.addcdiv_(-step_size * group['lr'], exp_avg, denom)
                    p.data.copy_(p_data_fp32)
                elif step_size > 0:
                    if group['weight_decay'] != 0:
                        p_data_fp32.add_(-group['weight_decay'] * group['lr'], p_data_fp32)
                    p_data_fp32.add_(-step_size * group['lr'], exp_avg)
                    p.data.copy_(p_data_fp32)

        return loss