File size: 3,414 Bytes
0a948c1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# Copyright (c) 2020 Mobvoi Inc (Binbin Zhang)
#               2024 Alibaba Inc (authors: Xiang Lyu)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# Modified from ESPnet(https://github.com/espnet/espnet)
"""Unility functions for Transformer."""

from typing import List

import torch

IGNORE_ID = -1


def pad_list(xs: List[torch.Tensor], pad_value: int):
    """Perform padding for the list of tensors.

    Args:
        xs (List): List of Tensors [(T_1, `*`), (T_2, `*`), ..., (T_B, `*`)].
        pad_value (float): Value for padding.

    Returns:
        Tensor: Padded tensor (B, Tmax, `*`).

    Examples:
        >>> x = [torch.ones(4), torch.ones(2), torch.ones(1)]
        >>> x
        [tensor([1., 1., 1., 1.]), tensor([1., 1.]), tensor([1.])]
        >>> pad_list(x, 0)
        tensor([[1., 1., 1., 1.],
                [1., 1., 0., 0.],
                [1., 0., 0., 0.]])

    """
    max_len = max([len(item) for item in xs])
    batchs = len(xs)
    ndim = xs[0].ndim
    if ndim == 1:
        pad_res = torch.zeros(batchs,
                              max_len,
                              dtype=xs[0].dtype,
                              device=xs[0].device)
    elif ndim == 2:
        pad_res = torch.zeros(batchs,
                              max_len,
                              xs[0].shape[1],
                              dtype=xs[0].dtype,
                              device=xs[0].device)
    elif ndim == 3:
        pad_res = torch.zeros(batchs,
                              max_len,
                              xs[0].shape[1],
                              xs[0].shape[2],
                              dtype=xs[0].dtype,
                              device=xs[0].device)
    else:
        raise ValueError(f"Unsupported ndim: {ndim}")
    pad_res.fill_(pad_value)
    for i in range(batchs):
        pad_res[i, :len(xs[i])] = xs[i]
    return pad_res


def th_accuracy(pad_outputs: torch.Tensor, pad_targets: torch.Tensor,
                ignore_label: int) -> torch.Tensor:
    """Calculate accuracy.

    Args:
        pad_outputs (Tensor): Prediction tensors (B * Lmax, D).
        pad_targets (LongTensor): Target label tensors (B, Lmax).
        ignore_label (int): Ignore label id.

    Returns:
        torch.Tensor: Accuracy value (0.0 - 1.0).

    """
    pad_pred = pad_outputs.view(pad_targets.size(0), pad_targets.size(1),
                                pad_outputs.size(1)).argmax(2)
    mask = pad_targets != ignore_label
    numerator = torch.sum(
        pad_pred.masked_select(mask) == pad_targets.masked_select(mask))
    denominator = torch.sum(mask)
    return (numerator / denominator).detach()


def get_padding(kernel_size, dilation=1):
    return int((kernel_size * dilation - dilation) / 2)


def init_weights(m, mean=0.0, std=0.01):
    classname = m.__class__.__name__
    if classname.find("Conv") != -1:
        m.weight.data.normal_(mean, std)