This repository was archived by the owner on Jul 7, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 3.8k
Expand file tree
/
Copy pathtranslate_ende.py
More file actions
249 lines (185 loc) · 6.63 KB
/
Copy pathtranslate_ende.py
File metadata and controls
249 lines (185 loc) · 6.63 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
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
# coding=utf-8
# Copyright 2023 The Tensor2Tensor Authors.
#
# 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.
"""Data generators for translation data-sets."""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from tensor2tensor.data_generators import problem
from tensor2tensor.data_generators import text_problems
from tensor2tensor.data_generators import translate
from tensor2tensor.data_generators import wiki_lm
from tensor2tensor.utils import registry
_ENDE_TRAIN_DATASETS = [
[
"http://data.statmt.org/wmt18/translation-task/training-parallel-nc-v13.tgz", # pylint: disable=line-too-long
("training-parallel-nc-v13/news-commentary-v13.de-en.en",
"training-parallel-nc-v13/news-commentary-v13.de-en.de")
],
[
"http://www.statmt.org/wmt13/training-parallel-commoncrawl.tgz",
("commoncrawl.de-en.en", "commoncrawl.de-en.de")
],
[
"http://www.statmt.org/wmt13/training-parallel-europarl-v7.tgz",
("training/europarl-v7.de-en.en", "training/europarl-v7.de-en.de")
],
]
_ENDE_EVAL_DATASETS = [
[
"http://data.statmt.org/wmt17/translation-task/dev.tgz",
("dev/newstest2013.en", "dev/newstest2013.de")
],
]
_ENDE_RAPID_TRAIN_DATASET = [
# additional training data available for WMT 18 news task training data
# as defined by http://www.statmt.org/wmt18/translation-task.html
[
"http://data.statmt.org/wmt18/translation-task/rapid2016.tgz",
("rapid2016.de-en.en", "rapid2016.de-en.de"),
],
]
_ENDE_PARACRAWL_DATASETS = [
[
"https://s3.amazonaws.com/web-language-models/paracrawl/release4/en-de.bicleaner07.tmx.gz", # pylint: disable=line-too-long
("tmx", "en-de.bicleaner07.tmx.gz")
]
]
@registry.register_problem
class TranslateEndeWmt32k(translate.TranslateProblem):
"""En-de translation trained on WMT corpus."""
@property
def additional_training_datasets(self):
"""Allow subclasses to add training datasets."""
return []
def source_data_files(self, dataset_split):
train = dataset_split == problem.DatasetSplit.TRAIN
train_datasets = _ENDE_TRAIN_DATASETS + self.additional_training_datasets
return train_datasets if train else _ENDE_EVAL_DATASETS
@registry.register_problem
class TranslateEnde2018Wmt32k(translate.TranslateProblem):
"""En-de translation trained on WMT18 corpus."""
@property
def use_vocab_from_other_problem(self):
return TranslateEndeWmt32k()
@property
def additional_training_datasets(self):
"""WMT18 adds rapid data."""
return _ENDE_RAPID_TRAIN_DATASET
@registry.register_problem
class TranslateEndeWmtClean32k(TranslateEndeWmt32k):
"""En-de translation trained on WMT with further cleaning."""
@property
def use_vocab_from_other_problem(self):
return TranslateEndeWmt32k()
@property
def datatypes_to_clean(self):
return ["txt"]
@registry.register_problem
class TranslateEndePc32k(translate.TranslateProblem):
"""En-de translation trained on Paracrawl (bicleaner corpus)."""
@property
def use_vocab_from_other_problem(self):
return TranslateEndeWmt32k()
@property
def additional_training_datasets(self):
"""Allow subclasses to add training datasets."""
return []
def source_data_files(self, dataset_split):
train = dataset_split == problem.DatasetSplit.TRAIN
train_datasets = (
_ENDE_PARACRAWL_DATASETS + self.additional_training_datasets)
return train_datasets if train else _ENDE_EVAL_DATASETS
@registry.register_problem
class TranslateEndePcClean32k(TranslateEndePc32k):
"""En-de translation trained on Paracrawl with further cleaning."""
@property
def datatypes_to_clean(self):
return ["tmx"]
@registry.register_problem
class TranslateEndeWmtPc32k(TranslateEndeWmt32k):
"""En-de translation trained on WMT plus Paracrawl."""
@property
def use_vocab_from_other_problem(self):
return TranslateEndeWmt32k()
@property
def additional_training_datasets(self):
return _ENDE_PARACRAWL_DATASETS
@registry.register_problem
class TranslateEndeWmtCleanPc32k(TranslateEndeWmtPc32k):
"""En-de translation trained on cleaned WMT plus Paracrawl."""
@property
def datatypes_to_clean(self):
return ["txt"]
@registry.register_problem
class TranslateEndeWmtPcClean32k(TranslateEndeWmtPc32k):
"""En-de translation trained on WMT plus cleaned Paracrawl."""
@property
def datatypes_to_clean(self):
return ["tmx"]
@registry.register_problem
class TranslateEndeWmtCleanPcClean32k(TranslateEndeWmtPcClean32k):
"""En-de translation trained on cleaned WMT plus cleaned Paracrawl."""
@property
def datatypes_to_clean(self):
return ["txt", "tmx"]
@registry.register_problem
class TranslateEndeWmt32kPacked(TranslateEndeWmt32k):
@property
def packed_length(self):
return 256
@property
def use_vocab_from_other_problem(self):
return TranslateEndeWmt32k()
@registry.register_problem
class TranslateEndeWmt8k(TranslateEndeWmt32k):
"""Problem spec for WMT En-De translation."""
@property
def approx_vocab_size(self):
return 2**13 # 8192
@registry.register_problem
class TranslateEndeWmt8kPacked(TranslateEndeWmt8k):
@property
def packed_length(self):
return 256
@property
def use_vocab_from_other_problem(self):
return TranslateEndeWmt8k()
@registry.register_problem
class TranslateEndeWmtCharacters(TranslateEndeWmt8k):
"""Problem spec for WMT En-De translation."""
@property
def vocab_type(self):
return text_problems.VocabType.CHARACTER
@registry.register_problem
class TranslateEndeWmtMulti64k(TranslateEndeWmt8k):
"""Translation with muli-lingual vocabulary."""
@property
def use_vocab_from_other_problem(self):
return wiki_lm.LanguagemodelDeEnFrRoWiki64k()
@registry.register_problem
class TranslateEndeWmtMulti64kPacked1k(TranslateEndeWmtMulti64k):
"""Translation with muli-lingual vocabulary."""
@property
def packed_length(self):
return 1024
@property
def num_training_examples(self):
return 173800
@property
def inputs_prefix(self):
return "translate English German "
@property
def targets_prefix(self):
return "translate German English "