anton-l HF staff commited on
Commit
e70863e
1 Parent(s): 49b5ae4

Update superb_demo.py

Browse files
Files changed (1) hide show
  1. superb_demo.py +19 -6
superb_demo.py CHANGED
@@ -135,6 +135,7 @@ class Superb(datasets.GeneratorBasedBuilder):
135
  features=datasets.Features(
136
  {
137
  "file": datasets.Value("string"),
 
138
  "text": datasets.Value("string"),
139
  "speaker_id": datasets.Value("int64"),
140
  "chapter_id": datasets.Value("int64"),
@@ -159,6 +160,7 @@ class Superb(datasets.GeneratorBasedBuilder):
159
  features=datasets.Features(
160
  {
161
  "file": datasets.Value("string"),
 
162
  "label": datasets.ClassLabel(
163
  names=[
164
  "yes",
@@ -192,6 +194,7 @@ class Superb(datasets.GeneratorBasedBuilder):
192
  features=datasets.Features(
193
  {
194
  "file": datasets.Value("string"),
 
195
  "speaker_id": datasets.Value("string"),
196
  "text": datasets.Value("string"),
197
  "action": datasets.ClassLabel(
@@ -234,6 +237,7 @@ class Superb(datasets.GeneratorBasedBuilder):
234
  features=datasets.Features(
235
  {
236
  "file": datasets.Value("string"),
 
237
  "label": datasets.ClassLabel(names=[f"id{i + 10001}" for i in range(1251)]),
238
  }
239
  ),
@@ -253,6 +257,7 @@ class Superb(datasets.GeneratorBasedBuilder):
253
  features=datasets.Features(
254
  {
255
  "file": datasets.Value("string"),
 
256
  "label": datasets.ClassLabel(names=['neu', 'hap', 'ang', 'sad']),
257
  }
258
  ),
@@ -327,11 +332,13 @@ class Superb(datasets.GeneratorBasedBuilder):
327
  id_, transcript = line.split(" ", 1)
328
  audio_file = f"{id_}.flac"
329
  speaker_id, chapter_id = [int(el) for el in id_.split("-")[:2]]
 
330
  yield key, {
331
  "id": id_,
332
  "speaker_id": speaker_id,
333
  "chapter_id": chapter_id,
334
- "file": os.path.join(transcript_dir_path, audio_file),
 
335
  "text": transcript,
336
  }
337
  key += 1
@@ -347,7 +354,7 @@ class Superb(datasets.GeneratorBasedBuilder):
347
  label = "_silence_"
348
  else:
349
  label = "_unknown_"
350
- yield key, {"file": audio_file, "label": label}
351
  elif self.config.name == "ic":
352
  root_path = os.path.join(archive_path, "fluent_speech_commands_dataset/")
353
  csv_path = os.path.join(root_path, f"data/{split}_data.csv")
@@ -356,8 +363,10 @@ class Superb(datasets.GeneratorBasedBuilder):
356
  next(csv_reader)
357
  for row in csv_reader:
358
  key, file_path, speaker_id, text, action, object_, location = row
 
359
  yield key, {
360
- "file": os.path.join(root_path, file_path),
 
361
  "speaker_id": speaker_id,
362
  "text": text,
363
  "action": action,
@@ -373,8 +382,10 @@ class Superb(datasets.GeneratorBasedBuilder):
373
  if int(split_id) != split:
374
  continue
375
  speaker_id = file_path.split("/")[0]
 
376
  yield key, {
377
- "file": os.path.join(wav_path, file_path),
 
378
  "label": speaker_id,
379
  }
380
  elif self.config.name == "er":
@@ -393,9 +404,11 @@ class Superb(datasets.GeneratorBasedBuilder):
393
  continue
394
  wav_subdir = filename.rsplit("_", 1)[0]
395
  filename = f"{filename}.wav"
 
396
  yield key, {
397
- "file": os.path.join(wav_path, wav_subdir, filename),
398
- "label": emo.replace('exc', 'hap'),
 
399
  }
400
  key += 1
401
 
 
135
  features=datasets.Features(
136
  {
137
  "file": datasets.Value("string"),
138
+ "audio": datasets.features.Audio(sampling_rate=16_000),
139
  "text": datasets.Value("string"),
140
  "speaker_id": datasets.Value("int64"),
141
  "chapter_id": datasets.Value("int64"),
 
160
  features=datasets.Features(
161
  {
162
  "file": datasets.Value("string"),
163
+ "audio": datasets.features.Audio(sampling_rate=16_000),
164
  "label": datasets.ClassLabel(
165
  names=[
166
  "yes",
 
194
  features=datasets.Features(
195
  {
196
  "file": datasets.Value("string"),
197
+ "audio": datasets.features.Audio(sampling_rate=16_000),
198
  "speaker_id": datasets.Value("string"),
199
  "text": datasets.Value("string"),
200
  "action": datasets.ClassLabel(
 
237
  features=datasets.Features(
238
  {
239
  "file": datasets.Value("string"),
240
+ "audio": datasets.features.Audio(sampling_rate=16_000),
241
  "label": datasets.ClassLabel(names=[f"id{i + 10001}" for i in range(1251)]),
242
  }
243
  ),
 
257
  features=datasets.Features(
258
  {
259
  "file": datasets.Value("string"),
260
+ "audio": datasets.features.Audio(sampling_rate=16_000),
261
  "label": datasets.ClassLabel(names=['neu', 'hap', 'ang', 'sad']),
262
  }
263
  ),
 
332
  id_, transcript = line.split(" ", 1)
333
  audio_file = f"{id_}.flac"
334
  speaker_id, chapter_id = [int(el) for el in id_.split("-")[:2]]
335
+ audio_path = os.path.join(transcript_dir_path, audio_file)
336
  yield key, {
337
  "id": id_,
338
  "speaker_id": speaker_id,
339
  "chapter_id": chapter_id,
340
+ "file": audio_path,
341
+ "audio": audio_path,
342
  "text": transcript,
343
  }
344
  key += 1
 
354
  label = "_silence_"
355
  else:
356
  label = "_unknown_"
357
+ yield key, {"file": audio_file, "audio": audio_file, "label": label}
358
  elif self.config.name == "ic":
359
  root_path = os.path.join(archive_path, "fluent_speech_commands_dataset/")
360
  csv_path = os.path.join(root_path, f"data/{split}_data.csv")
 
363
  next(csv_reader)
364
  for row in csv_reader:
365
  key, file_path, speaker_id, text, action, object_, location = row
366
+ audio_path = os.path.join(root_path, file_path)
367
  yield key, {
368
+ "file": audio_path,
369
+ "audio": audio_path,
370
  "speaker_id": speaker_id,
371
  "text": text,
372
  "action": action,
 
382
  if int(split_id) != split:
383
  continue
384
  speaker_id = file_path.split("/")[0]
385
+ audio_path = os.path.join(wav_path, file_path)
386
  yield key, {
387
+ "file": audio_path,
388
+ "audio": audio_path,
389
  "label": speaker_id,
390
  }
391
  elif self.config.name == "er":
 
404
  continue
405
  wav_subdir = filename.rsplit("_", 1)[0]
406
  filename = f"{filename}.wav"
407
+ audio_path = os.path.join(wav_path, wav_subdir, filename)
408
  yield key, {
409
+ "file": audio_path,
410
+ "audio": audio_path,
411
+ "label": emo.replace("exc", "hap"),
412
  }
413
  key += 1
414