-
Notifications
You must be signed in to change notification settings - Fork 10
Expand file tree
/
Copy pathutils.py
More file actions
209 lines (182 loc) · 7.15 KB
/
utils.py
File metadata and controls
209 lines (182 loc) · 7.15 KB
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
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
import torch
################################################################################
# The following function are based on:
# https://github.com/espnet/espnet/blob/master/espnet/nets/pytorch_backend/nets_utils.py
def make_pad_mask(lengths, xs=None, length_dim=-1):
"""Make mask tensor containing indices of padded part.
Args:
lengths (LongTensor or List): Batch of lengths (B,).
xs (Tensor, optional): The reference tensor.
If set, masks will be the same shape as this tensor.
length_dim (int, optional): Dimension indicator of the above tensor.
See the example.
Returns:
Tensor: Mask tensor containing indices of padded part.
dtype=torch.uint8 in PyTorch 1.2-
dtype=torch.bool in PyTorch 1.2+ (including 1.2)
Examples:
With only lengths.
>>> lengths = [5, 3, 2]
>>> make_non_pad_mask(lengths)
masks = [[0, 0, 0, 0 ,0],
[0, 0, 0, 1, 1],
[0, 0, 1, 1, 1]]
With the reference tensor.
>>> xs = torch.zeros((3, 2, 4))
>>> make_pad_mask(lengths, xs)
tensor([[[0, 0, 0, 0],
[0, 0, 0, 0]],
[[0, 0, 0, 1],
[0, 0, 0, 1]],
[[0, 0, 1, 1],
[0, 0, 1, 1]]], dtype=torch.uint8)
>>> xs = torch.zeros((3, 2, 6))
>>> make_pad_mask(lengths, xs)
tensor([[[0, 0, 0, 0, 0, 1],
[0, 0, 0, 0, 0, 1]],
[[0, 0, 0, 1, 1, 1],
[0, 0, 0, 1, 1, 1]],
[[0, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 1, 1]]], dtype=torch.uint8)
With the reference tensor and dimension indicator.
>>> xs = torch.zeros((3, 6, 6))
>>> make_pad_mask(lengths, xs, 1)
tensor([[[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1]],
[[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1]],
[[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1]]], dtype=torch.uint8)
>>> make_pad_mask(lengths, xs, 2)
tensor([[[0, 0, 0, 0, 0, 1],
[0, 0, 0, 0, 0, 1],
[0, 0, 0, 0, 0, 1],
[0, 0, 0, 0, 0, 1],
[0, 0, 0, 0, 0, 1],
[0, 0, 0, 0, 0, 1]],
[[0, 0, 0, 1, 1, 1],
[0, 0, 0, 1, 1, 1],
[0, 0, 0, 1, 1, 1],
[0, 0, 0, 1, 1, 1],
[0, 0, 0, 1, 1, 1],
[0, 0, 0, 1, 1, 1]],
[[0, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 1, 1]]], dtype=torch.uint8)
"""
if length_dim == 0:
raise ValueError("length_dim cannot be 0: {}".format(length_dim))
if not isinstance(lengths, list):
lengths = lengths.tolist()
bs = int(len(lengths))
if xs is None:
maxlen = int(max(lengths))
else:
maxlen = xs.size(length_dim)
seq_range = torch.arange(0, maxlen, dtype=torch.int64)
seq_range_expand = seq_range.unsqueeze(0).expand(bs, maxlen)
seq_length_expand = seq_range_expand.new(lengths).unsqueeze(-1)
mask = seq_range_expand >= seq_length_expand
if xs is not None:
assert xs.size(0) == bs, (xs.size(0), bs)
if length_dim < 0:
length_dim = xs.dim() + length_dim
# ind = (:, None, ..., None, :, , None, ..., None)
ind = tuple(
slice(None) if i in (0, length_dim) else None for i in range(xs.dim())
)
mask = mask[ind].expand_as(xs).to(xs.device)
return mask
def make_non_pad_mask(lengths, xs=None, length_dim=-1):
"""Make mask tensor containing indices of non-padded part.
Args:
lengths (LongTensor or List): Batch of lengths (B,).
xs (Tensor, optional): The reference tensor.
If set, masks will be the same shape as this tensor.
length_dim (int, optional): Dimension indicator of the above tensor.
See the example.
Returns:
ByteTensor: mask tensor containing indices of padded part.
dtype=torch.uint8 in PyTorch 1.2-
dtype=torch.bool in PyTorch 1.2+ (including 1.2)
Examples:
With only lengths.
>>> lengths = [5, 3, 2]
>>> make_non_pad_mask(lengths)
masks = [[1, 1, 1, 1 ,1],
[1, 1, 1, 0, 0],
[1, 1, 0, 0, 0]]
With the reference tensor.
>>> xs = torch.zeros((3, 2, 4))
>>> make_non_pad_mask(lengths, xs)
tensor([[[1, 1, 1, 1],
[1, 1, 1, 1]],
[[1, 1, 1, 0],
[1, 1, 1, 0]],
[[1, 1, 0, 0],
[1, 1, 0, 0]]], dtype=torch.uint8)
>>> xs = torch.zeros((3, 2, 6))
>>> make_non_pad_mask(lengths, xs)
tensor([[[1, 1, 1, 1, 1, 0],
[1, 1, 1, 1, 1, 0]],
[[1, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0]],
[[1, 1, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0]]], dtype=torch.uint8)
With the reference tensor and dimension indicator.
>>> xs = torch.zeros((3, 6, 6))
>>> make_non_pad_mask(lengths, xs, 1)
tensor([[[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0]],
[[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0]],
[[1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0]]], dtype=torch.uint8)
>>> make_non_pad_mask(lengths, xs, 2)
tensor([[[1, 1, 1, 1, 1, 0],
[1, 1, 1, 1, 1, 0],
[1, 1, 1, 1, 1, 0],
[1, 1, 1, 1, 1, 0],
[1, 1, 1, 1, 1, 0],
[1, 1, 1, 1, 1, 0]],
[[1, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0],
[1, 1, 1, 0, 0, 0]],
[[1, 1, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0],
[1, 1, 0, 0, 0, 0]]], dtype=torch.uint8)
"""
return ~make_pad_mask(lengths, xs, length_dim)