-
Notifications
You must be signed in to change notification settings - Fork 1
/
losses.py
184 lines (153 loc) · 7.22 KB
/
losses.py
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
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
"""
Author: Yonglong Tian (yonglong@mit.edu)
Date: May 07, 2020
"""
from __future__ import print_function
import torch
import torch.nn as nn
import random
import torch.nn.functional as F
class SupConLoss(nn.Module):
"""Supervised Contrastive Learning: https://arxiv.org/pdf/2004.11362.pdf.
It also supports the unsupervised contrastive loss in SimCLR"""
def __init__(self, temperature=0.07, contrast_mode='all',
base_temperature=0.07):
super(SupConLoss, self).__init__()
self.temperature = temperature
self.contrast_mode = contrast_mode
self.base_temperature = base_temperature
def forward(self, features, labels=None, mask=None):
"""Compute loss for model. If both `labels` and `mask` are None,
it degenerates to SimCLR unsupervised loss:
https://arxiv.org/pdf/2002.05709.pdf
Args:
features: hidden vector of shape [bsz, n_views, ...].
labels: ground truth of shape [bsz].
mask: contrastive mask of shape [bsz, bsz], mask_{i,j}=1 if sample j
has the same class as sample i. Can be asymmetric.
Returns:
A loss scalar.
"""
device = (torch.device('cuda')
if features.is_cuda
else torch.device('cpu'))
if len(features.shape) < 3:
raise ValueError('`features` needs to be [bsz, n_views, ...],'
'at least 3 dimensions are required')
if len(features.shape) > 3:
features = features.view(features.shape[0], features.shape[1], -1)
batch_size = features.shape[0]
if labels is not None and mask is not None:
raise ValueError('Cannot define both `labels` and `mask`')
elif labels is None and mask is None:
mask = torch.eye(batch_size, dtype=torch.float32).to(device)
elif labels is not None:
labels = labels.contiguous().view(-1, 1)
if labels.shape[0] != batch_size:
raise ValueError('Num of labels does not match num of features')
mask = torch.eq(labels, labels.t()).float().to(device)
else:
mask = mask.float().to(device)
contrast_count = features.shape[1]
contrast_feature = torch.cat(torch.unbind(features, dim=1), dim=0)
if self.contrast_mode == 'one':
anchor_feature = features[:, 0]
anchor_count = 1
elif self.contrast_mode == 'all':
anchor_feature = contrast_feature
anchor_count = contrast_count
else:
raise ValueError('Unknown mode: {}'.format(self.contrast_mode))
# compute logits
anchor_dot_contrast = torch.div(
torch.matmul(anchor_feature, contrast_feature.t()),
self.temperature)
# for numerical stability
logits_max, _ = torch.max(anchor_dot_contrast, dim=1, keepdim=True)
logits = anchor_dot_contrast - logits_max.detach()
# tile mask
mask = mask.repeat(anchor_count, contrast_count)
# mask-out self-contrast cases
logits_mask = torch.scatter(
torch.ones_like(mask),
1,
torch.arange(batch_size * anchor_count).view(-1, 1).to(device),
0
)
mask = mask * logits_mask
# compute log_prob
exp_logits = torch.exp(logits) * logits_mask
log_prob = logits - torch.log(exp_logits.sum(1, keepdim=True))
same_cls = mask.sum(1)
x = torch.ones(same_cls.shape[0], dtype=torch.float).to(device)
same_cls = torch.where(same_cls == 0., x, same_cls)
# compute mean of log-likelihood over positive
mean_log_prob_pos = (mask * log_prob).sum(1) / same_cls
# loss
loss = - (self.temperature / self.base_temperature) * mean_log_prob_pos
loss = loss.view(anchor_count, batch_size).mean()
return loss
class infoNCE(nn.Module):
def __init__(self, class_num):
super(infoNCE, self).__init__()
self.class_num = class_num
self.neg_samples = 10
def get_posAndneg(self, confi_labels_dict, feature_q_idx, source_centers, confi_class_idx):
# get the label of q
q_label = confi_labels_dict[feature_q_idx.item()]
feature_pos = source_centers[q_label].unsqueeze(0)
# get the negative samples
negative_idx = []
for i in range(self.class_num):
if i != q_label:
negative_idx.append(i)
negative_pairs = torch.Tensor([]).cuda()
for i in range(self.neg_samples):
negative_pairs = torch.cat((negative_pairs, source_centers[random.choice(negative_idx)].unsqueeze(0)))
# negative_pairs = torch.cat((source_centers[:feature_q_idx].unsqueeze(0), source_centers[feature_q_idx:].unsqueeze(0)))
return torch.cat((feature_pos, negative_pairs))
def reply_get_posAndneg(self, reply_label, source_centers):
# get the label of q
q_label = reply_label
feature_pos = source_centers[q_label].unsqueeze(0)
# get the negative samples
negative_idx = []
for i in range(self.class_num):
if i != reply_label:
negative_idx.append(i)
negative_pairs = torch.Tensor([]).cuda()
for i in range(self.neg_samples):
negative_pairs = torch.cat((negative_pairs, source_centers[random.choice(negative_idx)].unsqueeze(0)))
return torch.cat((feature_pos, negative_pairs))
class elr_loss(nn.Module):
def __init__(self, num_examp, num_classes=10, lambda_num=3, beta=0.7):
r"""Early Learning Regularization.
Parameters
* `num_examp` Total number of training examples.
* `num_classes` Number of classes in the classification problem.
* `lambda` Regularization strength; must be a positive float, controling the strength of the ELR.
* `beta` Temporal ensembling momentum for target estimation.
"""
super(elr_loss, self).__init__()
self.num_classes = num_classes
self.USE_CUDA = torch.cuda.is_available()
self.target = torch.zeros(num_examp, self.num_classes).cuda() if self.USE_CUDA else torch.zeros(num_examp,
self.num_classes)
self.beta = beta
self.lambda_num = lambda_num
def forward(self, index, output, label, contrastive_loss=None):
r"""Early Learning Regularization.
Args
* `index` Training sample index, due to training set shuffling, index is used to track training examples in different iterations.
* `output` Model's logits, same as PyTorch provided loss functions.
* `label` Labels, same as PyTorch provided loss functions.
"""
y_pred = F.softmax(output, dim=1)
y_pred = torch.clamp(y_pred, 1e-4, 1.0 - 1e-4)
y_pred_ = y_pred.data.detach()
self.target[index] = self.beta * self.target[index] + (1 - self.beta) * (
(y_pred_) / (y_pred_).sum(dim=1, keepdim=True))
elr_reg = ((1 - (self.target[index] * y_pred).sum(dim=1)).log()).mean()
ce_loss = F.cross_entropy(output, label)
final_loss = ce_loss + self.lambda_num * elr_reg
return final_loss