-
Notifications
You must be signed in to change notification settings - Fork 388
/
cer.py
102 lines (89 loc) · 4 KB
/
cer.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
# Copyright The PyTorch Lightning team.
#
# 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.
from typing import List, Tuple, Union
import torch
from torch import Tensor, tensor
def _edit_distance(prediction_tokens: List[str], reference_tokens: List[str]) -> int:
"""Standard dynamic programming algorithm to compute the edit distance.
Args:
prediction_tokens: A tokenized predicted sentence
reference_tokens: A tokenized reference sentence
Returns:
(int) Edit distance between the predicted sentence and the reference sentence
"""
dp = [[0] * (len(reference_tokens) + 1) for _ in range(len(prediction_tokens) + 1)]
for i in range(len(prediction_tokens) + 1):
dp[i][0] = i
for j in range(len(reference_tokens) + 1):
dp[0][j] = j
for i in range(1, len(prediction_tokens) + 1):
for j in range(1, len(reference_tokens) + 1):
if prediction_tokens[i - 1] == reference_tokens[j - 1]:
dp[i][j] = dp[i - 1][j - 1]
else:
dp[i][j] = min(dp[i - 1][j], dp[i][j - 1], dp[i - 1][j - 1]) + 1
return dp[-1][-1]
def _cer_update(
predictions: Union[str, List[str]],
references: Union[str, List[str]],
) -> Tuple[Tensor, Tensor]:
"""Update the cer score with the current set of references and predictions.
Args:
predictions: Transcription(s) to score as a string or list of strings
references: Reference(s) for each speech input as a string or list of strings
Returns:
(Tensor) Number of edit operations to get from the reference to the prediction, summed over all samples
(Tensor) Number of character over all references
"""
if isinstance(predictions, str):
predictions = [predictions]
if isinstance(references, str):
references = [references]
errors = tensor(0, dtype=torch.float)
total = tensor(0, dtype=torch.float)
for prediction, reference in zip(predictions, references):
prediction_tokens = prediction
reference_tokens = reference
errors += _edit_distance(list(prediction_tokens), list(reference_tokens))
total += len(reference_tokens)
return errors, total
def _cer_compute(errors: Tensor, total: Tensor) -> Tensor:
"""Compute the Character error rate.
Args:
errors: Number of edit operations to get from the reference to the prediction, summed over all samples
total: Number of characters over all references
Returns:
(Tensor) Character error rate
"""
return errors / total
def char_error_rate(
predictions: Union[str, List[str]],
references: Union[str, List[str]],
) -> Tensor:
"""character error rate is a common metric of the performance of an automatic speech recognition system. This
value indicates the percentage of characters that were incorrectly predicted. The lower the value, the better the
performance of the ASR system with a CER of 0 being a perfect score.
Args:
predictions: Transcription(s) to score as a string or list of strings
references: Reference(s) for each speech input as a string or list of strings
Returns:
(Tensor) Character error rate
Examples:
>>> predictions = ["this is the prediction", "there is an other sample"]
>>> references = ["this is the reference", "there is another one"]
>>> char_error_rate(predictions=predictions, references=references)
tensor(0.3415)
"""
errors, total = _cer_update(predictions, references)
return _cer_compute(errors, total)