-
Notifications
You must be signed in to change notification settings - Fork 388
/
bleu.py
140 lines (124 loc) · 5.01 KB
/
bleu.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
# 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.
# referenced from
# Library Name: torchtext
# Authors: torchtext authors and @sluks
# Date: 2020-07-18
# Link: https://pytorch.org/text/_modules/torchtext/data/metrics.html#bleu_score
from typing import Any, Callable, Optional, Sequence
from warnings import warn
import torch
from deprecate import deprecated
from torch import Tensor, tensor
from torchmetrics import Metric
from torchmetrics.functional.text.bleu import _bleu_score_compute, _bleu_score_update, _tokenize_fn
from torchmetrics.utilities import _future_warning
class BLEUScore(Metric):
"""Calculate `BLEU score`_ of machine translated text with one or more references.
Args:
n_gram:
Gram value ranged from 1 to 4 (Default 4)
smooth:
Whether or not to apply smoothing – see [2]
compute_on_step:
Forward only calls ``update()`` and returns None if this is set to False.
dist_sync_on_step:
Synchronize metric state across processes at each ``forward()``
before returning the value at the step.
process_group:
Specify the process group on which synchronization is called.
dist_sync_fn:
Callback that performs the allgather operation on the metric state. When `None`, DDP
will be used to perform the allgather.
Example:
>>> from torchmetrics import BLEUScore
>>> preds = ['the cat is on the mat']
>>> target = [['there is a cat on the mat', 'a cat is on the mat']]
>>> metric = BLEUScore()
>>> metric(preds, target)
tensor(0.7598)
References:
[1] BLEU: a Method for Automatic Evaluation of Machine Translation by Papineni,
Kishore, Salim Roukos, Todd Ward, and Wei-Jing Zhu `BLEU`_
[2] Automatic Evaluation of Machine Translation Quality Using Longest Common Subsequence
and Skip-Bigram Statistics by Chin-Yew Lin and Franz Josef Och `Machine Translation Evolution`_
"""
is_differentiable = False
higher_is_better = True
preds_len: Tensor
target_len: Tensor
numerator: Tensor
denominator: Tensor
def __init__(
self,
n_gram: int = 4,
smooth: bool = False,
compute_on_step: bool = True,
dist_sync_on_step: bool = False,
process_group: Optional[Any] = None,
dist_sync_fn: Optional[Callable] = None,
):
super().__init__(
compute_on_step=compute_on_step,
dist_sync_on_step=dist_sync_on_step,
process_group=process_group,
dist_sync_fn=dist_sync_fn,
)
warn(
"Input order of targets and preds were changed to predictions firsts and targets second in v0.7."
" Warning will be removed in v0.8."
)
self.n_gram = n_gram
self.smooth = smooth
self.add_state("preds_len", tensor(0.0), dist_reduce_fx="sum")
self.add_state("target_len", tensor(0.0), dist_reduce_fx="sum")
self.add_state("numerator", torch.zeros(self.n_gram), dist_reduce_fx="sum")
self.add_state("denominator", torch.zeros(self.n_gram), dist_reduce_fx="sum")
@deprecated(
args_mapping={"translate_corpus": "preds", "reference_corpus": "target"},
target=True,
deprecated_in="0.7",
remove_in="0.8",
stream=_future_warning,
)
def update(self, preds: Sequence[str], target: Sequence[Sequence[str]]) -> None: # type: ignore
"""Compute Precision Scores.
Args:
preds: An iterable of machine translated corpus
target: An iterable of iterables of reference corpus
.. deprecated:: v0.7
Args:
translate_corpus:
This argument is deprecated in favor of `preds` and will be removed in v0.8.
reference_corpus:
This argument is deprecated in favor of `target` and will be removed in v0.8.
"""
self.preds_len, self.target_len = _bleu_score_update(
preds,
target,
self.numerator,
self.denominator,
self.preds_len,
self.target_len,
self.n_gram,
_tokenize_fn,
)
def compute(self) -> Tensor:
"""Calculate BLEU score.
Return:
Tensor with BLEU Score
"""
return _bleu_score_compute(
self.preds_len, self.target_len, self.numerator, self.denominator, self.n_gram, self.smooth
)