sbmaruf commited on
Commit
59dde55
1 Parent(s): 35089fc

Update forai_ml-ted_talk_iwslt.py

Browse files
Files changed (1) hide show
  1. forai_ml-ted_talk_iwslt.py +20 -7
forai_ml-ted_talk_iwslt.py CHANGED
@@ -187,6 +187,16 @@ _LANGUAGE_PAIRS = [
187
  ("fr-ca", "hi"),
188
  ]
189
 
 
 
 
 
 
 
 
 
 
 
190
  # Year subscripts for the specific folder
191
  _YEAR = {"2014": "-20140120", "2015": "-20150530", "2016": "-20160408"}
192
 
@@ -251,9 +261,14 @@ class TedTalksIWSLT(datasets.GeneratorBasedBuilder):
251
  def _info(self):
252
  features = datasets.Features(
253
  {
254
- "translation": datasets.features.Translation(
255
- languages=self.config.language_pair
256
- ),
 
 
 
 
 
257
  },
258
  )
259
 
@@ -432,11 +447,9 @@ class TedTalksIWSLT(datasets.GeneratorBasedBuilder):
432
  translation.extend(transcriptions)
433
  for talk_segment in translation:
434
  result = {
435
- "translation": {
436
  "source": talk_segment[language_pair[0]],
437
  "target": talk_segment[language_pair[1]],
438
- "src_lang" : language_pair[0],
439
- "tgt_lang" : language_pair[1],
440
- }
441
  }
442
  yield talk_segment["id"], result
 
187
  ("fr-ca", "hi"),
188
  ]
189
 
190
+ _LANGUAGE_TO_FULL_NAME = {
191
+ "eu": "Basque",
192
+ "de": "German",
193
+ "nl": "Dutch",
194
+ "hi": "Hindi",
195
+ "ja": "Japanese",
196
+ "fr-ca": "French",
197
+ "ca": "Catalan"
198
+ }
199
+
200
  # Year subscripts for the specific folder
201
  _YEAR = {"2014": "-20140120", "2015": "-20150530", "2016": "-20160408"}
202
 
 
261
  def _info(self):
262
  features = datasets.Features(
263
  {
264
+ # "translation": datasets.features.Translation(
265
+ # languages=self.config.language_pair
266
+ # ),
267
+
268
+ "source": datasets.Value("string"),
269
+ "target": datasets.Value("string"),
270
+ "source_lang": datasets.Value("string"),
271
+ "target_lang": datasets.Value("string"),
272
  },
273
  )
274
 
 
447
  translation.extend(transcriptions)
448
  for talk_segment in translation:
449
  result = {
 
450
  "source": talk_segment[language_pair[0]],
451
  "target": talk_segment[language_pair[1]],
452
+ "source_lang": _LANGUAGE_TO_FULL_NAME[language_pair[0]],
453
+ "target_lang": _LANGUAGE_TO_FULL_NAME[language_pair[1]],
 
454
  }
455
  yield talk_segment["id"], result