Hugging Face's logo

Datasets:
SpeechPPL
/
SALMon_pGSLM-normalized

Modalities:
Audio
Text
Time-series
Formats:
parquet
Size:
1K - 10K
Libraries:
Datasets
pandas
Polars
Dataset card Data Studio Files
xet
 Community
Splend1dchan commited on
Commit
b5be756
·
verified ·
1 Parent(s): 5bdc6d2

Upload README.md for normalized SALMon repo SpeechPPL/SALMon_pGSLM

Browse files
Files changed (1) hide show
  1. README.md +168 -56
README.md CHANGED
@@ -58,13 +58,21 @@ dataset_info:
58
  audio:
59
  sampling_rate: 16000
60
  - name: positive_sample_raw_units
61
- dtype:
62
  - name: hubert
63
  dtype: string
 
 
 
 
64
  - name: negative_sample_raw_units
65
- dtype:
66
  - name: hubert
67
  dtype: string
 
 
 
 
68
  - name: prompt_audio
69
  dtype:
70
  audio:
@@ -82,11 +90,11 @@ dataset_info:
82
  - name: negative_continuation_tokenwise_loss
83
  dtype: 'null'
84
  - name: positive_continuation_raw_units
85
- dtype:
86
  - name: hubert
87
  dtype: string
88
  - name: negative_continuation_raw_units
89
- dtype:
90
  - name: hubert
91
  dtype: string
92
  - name: continuation_audio_positive
@@ -99,10 +107,10 @@ dataset_info:
99
  sampling_rate: 16000
100
  splits:
101
  - name: train
102
- num_bytes: 87019899
103
  num_examples: 200
104
- download_size: 87019899
105
- dataset_size: 87019899
106
  - config_name: bg_all_consistency
107
  features:
108
  - name: task
@@ -128,13 +136,21 @@ dataset_info:
128
  audio:
129
  sampling_rate: 16000
130
  - name: positive_sample_raw_units
131
- dtype:
132
  - name: hubert
133
  dtype: string
 
 
 
 
134
  - name: negative_sample_raw_units
135
- dtype:
136
  - name: hubert
137
  dtype: string
 
 
 
 
138
  - name: prompt_audio
139
  dtype:
140
  audio:
@@ -152,13 +168,21 @@ dataset_info:
152
  - name: negative_continuation_tokenwise_loss
153
  list: float64
154
  - name: positive_continuation_raw_units
155
- dtype:
156
  - name: hubert
157
  dtype: string
 
 
 
 
158
  - name: negative_continuation_raw_units
159
- dtype:
160
  - name: hubert
161
  dtype: string
 
 
 
 
162
  - name: continuation_audio_positive
163
  dtype:
164
  audio:
@@ -169,10 +193,10 @@ dataset_info:
169
  sampling_rate: 16000
170
  splits:
171
  - name: train
172
- num_bytes: 207143109
173
  num_examples: 200
174
- download_size: 207143109
175
- dataset_size: 207143109
176
  - config_name: bg_domain_consistency
177
  features:
178
  - name: task
@@ -198,13 +222,21 @@ dataset_info:
198
  audio:
199
  sampling_rate: 16000
200
  - name: positive_sample_raw_units
201
- dtype:
202
  - name: hubert
203
  dtype: string
 
 
 
 
204
  - name: negative_sample_raw_units
205
- dtype:
206
  - name: hubert
207
  dtype: string
 
 
 
 
208
  - name: prompt_audio
209
  dtype:
210
  audio:
@@ -222,13 +254,21 @@ dataset_info:
222
  - name: negative_continuation_tokenwise_loss
223
  list: float64
224
  - name: positive_continuation_raw_units
225
- dtype:
226
  - name: hubert
227
  dtype: string
 
 
 
 
228
  - name: negative_continuation_raw_units
229
- dtype:
230
  - name: hubert
231
  dtype: string
 
 
 
 
232
  - name: continuation_audio_positive
233
  dtype:
234
  audio:
@@ -239,10 +279,10 @@ dataset_info:
239
  sampling_rate: 16000
240
  splits:
241
  - name: train
242
- num_bytes: 210303767
243
  num_examples: 200
244
- download_size: 210303767
245
- dataset_size: 210303767
246
  - config_name: gender_consistency
247
  features:
248
  - name: task
@@ -268,13 +308,21 @@ dataset_info:
268
  audio:
269
  sampling_rate: 16000
270
  - name: positive_sample_raw_units
271
- dtype:
272
  - name: hubert
273
  dtype: string
 
 
 
 
274
  - name: negative_sample_raw_units
275
- dtype:
276
  - name: hubert
277
  dtype: string
 
 
 
 
278
  - name: prompt_audio
279
  dtype:
280
  audio:
@@ -292,13 +340,21 @@ dataset_info:
292
  - name: negative_continuation_tokenwise_loss
293
  list: float64
294
  - name: positive_continuation_raw_units
295
- dtype:
296
  - name: hubert
297
  dtype: string
 
 
 
 
298
  - name: negative_continuation_raw_units
299
- dtype:
300
  - name: hubert
301
  dtype: string
 
 
 
 
302
  - name: continuation_audio_positive
303
  dtype:
304
  audio:
@@ -309,10 +365,10 @@ dataset_info:
309
  sampling_rate: 16000
310
  splits:
311
  - name: train
312
- num_bytes: 209062383
313
  num_examples: 200
314
- download_size: 209062383
315
- dataset_size: 209062383
316
  - config_name: rir_consistency
317
  features:
318
  - name: task
@@ -338,13 +394,21 @@ dataset_info:
338
  audio:
339
  sampling_rate: 16000
340
  - name: positive_sample_raw_units
341
- dtype:
342
  - name: hubert
343
  dtype: string
 
 
 
 
344
  - name: negative_sample_raw_units
345
- dtype:
346
  - name: hubert
347
  dtype: string
 
 
 
 
348
  - name: prompt_audio
349
  dtype:
350
  audio:
@@ -362,13 +426,21 @@ dataset_info:
362
  - name: negative_continuation_tokenwise_loss
363
  list: float64
364
  - name: positive_continuation_raw_units
365
- dtype:
366
  - name: hubert
367
  dtype: string
 
 
 
 
368
  - name: negative_continuation_raw_units
369
- dtype:
370
  - name: hubert
371
  dtype: string
 
 
 
 
372
  - name: continuation_audio_positive
373
  dtype:
374
  audio:
@@ -379,10 +451,10 @@ dataset_info:
379
  sampling_rate: 16000
380
  splits:
381
  - name: train
382
- num_bytes: 202129101
383
  num_examples: 200
384
- download_size: 202129101
385
- dataset_size: 202129101
386
  - config_name: sentiment_alignment
387
  features:
388
  - name: task
@@ -408,13 +480,21 @@ dataset_info:
408
  audio:
409
  sampling_rate: 16000
410
  - name: positive_sample_raw_units
411
- dtype:
412
  - name: hubert
413
  dtype: string
 
 
 
 
414
  - name: negative_sample_raw_units
415
- dtype:
416
  - name: hubert
417
  dtype: string
 
 
 
 
418
  - name: prompt_audio
419
  dtype:
420
  audio:
@@ -432,11 +512,11 @@ dataset_info:
432
  - name: negative_continuation_tokenwise_loss
433
  dtype: 'null'
434
  - name: positive_continuation_raw_units
435
- dtype:
436
  - name: hubert
437
  dtype: string
438
  - name: negative_continuation_raw_units
439
- dtype:
440
  - name: hubert
441
  dtype: string
442
  - name: continuation_audio_positive
@@ -449,10 +529,10 @@ dataset_info:
449
  sampling_rate: 16000
450
  splits:
451
  - name: train
452
- num_bytes: 46742156
453
  num_examples: 200
454
- download_size: 46742156
455
- dataset_size: 46742156
456
  - config_name: sentiment_consistency
457
  features:
458
  - name: task
@@ -478,13 +558,21 @@ dataset_info:
478
  audio:
479
  sampling_rate: 16000
480
  - name: positive_sample_raw_units
481
- dtype:
482
  - name: hubert
483
  dtype: string
 
 
 
 
484
  - name: negative_sample_raw_units
485
- dtype:
486
  - name: hubert
487
  dtype: string
 
 
 
 
488
  - name: prompt_audio
489
  dtype:
490
  audio:
@@ -502,13 +590,21 @@ dataset_info:
502
  - name: negative_continuation_tokenwise_loss
503
  list: float64
504
  - name: positive_continuation_raw_units
505
- dtype:
506
  - name: hubert
507
  dtype: string
 
 
 
 
508
  - name: negative_continuation_raw_units
509
- dtype:
510
  - name: hubert
511
  dtype: string
 
 
 
 
512
  - name: continuation_audio_positive
513
  dtype:
514
  audio:
@@ -519,10 +615,10 @@ dataset_info:
519
  sampling_rate: 16000
520
  splits:
521
  - name: train
522
- num_bytes: 210124373
523
  num_examples: 200
524
- download_size: 210124373
525
- dataset_size: 210124373
526
  - config_name: speaker_consistency
527
  features:
528
  - name: task
@@ -548,13 +644,21 @@ dataset_info:
548
  audio:
549
  sampling_rate: 16000
550
  - name: positive_sample_raw_units
551
- dtype:
552
  - name: hubert
553
  dtype: string
 
 
 
 
554
  - name: negative_sample_raw_units
555
- dtype:
556
  - name: hubert
557
  dtype: string
 
 
 
 
558
  - name: prompt_audio
559
  dtype:
560
  audio:
@@ -572,13 +676,21 @@ dataset_info:
572
  - name: negative_continuation_tokenwise_loss
573
  list: float64
574
  - name: positive_continuation_raw_units
575
- dtype:
576
  - name: hubert
577
  dtype: string
 
 
 
 
578
  - name: negative_continuation_raw_units
579
- dtype:
580
  - name: hubert
581
  dtype: string
 
 
 
 
582
  - name: continuation_audio_positive
583
  dtype:
584
  audio:
@@ -589,10 +701,10 @@ dataset_info:
589
  sampling_rate: 16000
590
  splits:
591
  - name: train
592
- num_bytes: 209864717
593
  num_examples: 200
594
- download_size: 209864717
595
- dataset_size: 209864717
596
  ---
597
 
598
  # SALMon Normalized Dataset
 
58
  audio:
59
  sampling_rate: 16000
60
  - name: positive_sample_raw_units
61
+ list:
62
  - name: hubert
63
  dtype: string
64
+ - name: pitch
65
+ dtype: string
66
+ - name: duration
67
+ dtype: string
68
  - name: negative_sample_raw_units
69
+ list:
70
  - name: hubert
71
  dtype: string
72
+ - name: pitch
73
+ dtype: string
74
+ - name: duration
75
+ dtype: string
76
  - name: prompt_audio
77
  dtype:
78
  audio:
 
90
  - name: negative_continuation_tokenwise_loss
91
  dtype: 'null'
92
  - name: positive_continuation_raw_units
93
+ list:
94
  - name: hubert
95
  dtype: string
96
  - name: negative_continuation_raw_units
97
+ list:
98
  - name: hubert
99
  dtype: string
100
  - name: continuation_audio_positive
 
107
  sampling_rate: 16000
108
  splits:
109
  - name: train
110
+ num_bytes: 87008171
111
  num_examples: 200
112
+ download_size: 87008171
113
+ dataset_size: 87008171
114
  - config_name: bg_all_consistency
115
  features:
116
  - name: task
 
136
  audio:
137
  sampling_rate: 16000
138
  - name: positive_sample_raw_units
139
+ list:
140
  - name: hubert
141
  dtype: string
142
+ - name: pitch
143
+ dtype: string
144
+ - name: duration
145
+ dtype: string
146
  - name: negative_sample_raw_units
147
+ list:
148
  - name: hubert
149
  dtype: string
150
+ - name: pitch
151
+ dtype: string
152
+ - name: duration
153
+ dtype: string
154
  - name: prompt_audio
155
  dtype:
156
  audio:
 
168
  - name: negative_continuation_tokenwise_loss
169
  list: float64
170
  - name: positive_continuation_raw_units
171
+ list:
172
  - name: hubert
173
  dtype: string
174
+ - name: pitch
175
+ dtype: string
176
+ - name: duration
177
+ dtype: string
178
  - name: negative_continuation_raw_units
179
+ list:
180
  - name: hubert
181
  dtype: string
182
+ - name: pitch
183
+ dtype: string
184
+ - name: duration
185
+ dtype: string
186
  - name: continuation_audio_positive
187
  dtype:
188
  audio:
 
193
  sampling_rate: 16000
194
  splits:
195
  - name: train
196
+ num_bytes: 207198666
197
  num_examples: 200
198
+ download_size: 207198666
199
+ dataset_size: 207198666
200
  - config_name: bg_domain_consistency
201
  features:
202
  - name: task
 
222
  audio:
223
  sampling_rate: 16000
224
  - name: positive_sample_raw_units
225
+ list:
226
  - name: hubert
227
  dtype: string
228
+ - name: pitch
229
+ dtype: string
230
+ - name: duration
231
+ dtype: string
232
  - name: negative_sample_raw_units
233
+ list:
234
  - name: hubert
235
  dtype: string
236
+ - name: pitch
237
+ dtype: string
238
+ - name: duration
239
+ dtype: string
240
  - name: prompt_audio
241
  dtype:
242
  audio:
 
254
  - name: negative_continuation_tokenwise_loss
255
  list: float64
256
  - name: positive_continuation_raw_units
257
+ list:
258
  - name: hubert
259
  dtype: string
260
+ - name: pitch
261
+ dtype: string
262
+ - name: duration
263
+ dtype: string
264
  - name: negative_continuation_raw_units
265
+ list:
266
  - name: hubert
267
  dtype: string
268
+ - name: pitch
269
+ dtype: string
270
+ - name: duration
271
+ dtype: string
272
  - name: continuation_audio_positive
273
  dtype:
274
  audio:
 
279
  sampling_rate: 16000
280
  splits:
281
  - name: train
282
+ num_bytes: 210361011
283
  num_examples: 200
284
+ download_size: 210361011
285
+ dataset_size: 210361011
286
  - config_name: gender_consistency
287
  features:
288
  - name: task
 
308
  audio:
309
  sampling_rate: 16000
310
  - name: positive_sample_raw_units
311
+ list:
312
  - name: hubert
313
  dtype: string
314
+ - name: pitch
315
+ dtype: string
316
+ - name: duration
317
+ dtype: string
318
  - name: negative_sample_raw_units
319
+ list:
320
  - name: hubert
321
  dtype: string
322
+ - name: pitch
323
+ dtype: string
324
+ - name: duration
325
+ dtype: string
326
  - name: prompt_audio
327
  dtype:
328
  audio:
 
340
  - name: negative_continuation_tokenwise_loss
341
  list: float64
342
  - name: positive_continuation_raw_units
343
+ list:
344
  - name: hubert
345
  dtype: string
346
+ - name: pitch
347
+ dtype: string
348
+ - name: duration
349
+ dtype: string
350
  - name: negative_continuation_raw_units
351
+ list:
352
  - name: hubert
353
  dtype: string
354
+ - name: pitch
355
+ dtype: string
356
+ - name: duration
357
+ dtype: string
358
  - name: continuation_audio_positive
359
  dtype:
360
  audio:
 
365
  sampling_rate: 16000
366
  splits:
367
  - name: train
368
+ num_bytes: 209107955
369
  num_examples: 200
370
+ download_size: 209107955
371
+ dataset_size: 209107955
372
  - config_name: rir_consistency
373
  features:
374
  - name: task
 
394
  audio:
395
  sampling_rate: 16000
396
  - name: positive_sample_raw_units
397
+ list:
398
  - name: hubert
399
  dtype: string
400
+ - name: pitch
401
+ dtype: string
402
+ - name: duration
403
+ dtype: string
404
  - name: negative_sample_raw_units
405
+ list:
406
  - name: hubert
407
  dtype: string
408
+ - name: pitch
409
+ dtype: string
410
+ - name: duration
411
+ dtype: string
412
  - name: prompt_audio
413
  dtype:
414
  audio:
 
426
  - name: negative_continuation_tokenwise_loss
427
  list: float64
428
  - name: positive_continuation_raw_units
429
+ list:
430
  - name: hubert
431
  dtype: string
432
+ - name: pitch
433
+ dtype: string
434
+ - name: duration
435
+ dtype: string
436
  - name: negative_continuation_raw_units
437
+ list:
438
  - name: hubert
439
  dtype: string
440
+ - name: pitch
441
+ dtype: string
442
+ - name: duration
443
+ dtype: string
444
  - name: continuation_audio_positive
445
  dtype:
446
  audio:
 
451
  sampling_rate: 16000
452
  splits:
453
  - name: train
454
+ num_bytes: 202178950
455
  num_examples: 200
456
+ download_size: 202178950
457
+ dataset_size: 202178950
458
  - config_name: sentiment_alignment
459
  features:
460
  - name: task
 
480
  audio:
481
  sampling_rate: 16000
482
  - name: positive_sample_raw_units
483
+ list:
484
  - name: hubert
485
  dtype: string
486
+ - name: pitch
487
+ dtype: string
488
+ - name: duration
489
+ dtype: string
490
  - name: negative_sample_raw_units
491
+ list:
492
  - name: hubert
493
  dtype: string
494
+ - name: pitch
495
+ dtype: string
496
+ - name: duration
497
+ dtype: string
498
  - name: prompt_audio
499
  dtype:
500
  audio:
 
512
  - name: negative_continuation_tokenwise_loss
513
  dtype: 'null'
514
  - name: positive_continuation_raw_units
515
+ list:
516
  - name: hubert
517
  dtype: string
518
  - name: negative_continuation_raw_units
519
+ list:
520
  - name: hubert
521
  dtype: string
522
  - name: continuation_audio_positive
 
529
  sampling_rate: 16000
530
  splits:
531
  - name: train
532
+ num_bytes: 46732334
533
  num_examples: 200
534
+ download_size: 46732334
535
+ dataset_size: 46732334
536
  - config_name: sentiment_consistency
537
  features:
538
  - name: task
 
558
  audio:
559
  sampling_rate: 16000
560
  - name: positive_sample_raw_units
561
+ list:
562
  - name: hubert
563
  dtype: string
564
+ - name: pitch
565
+ dtype: string
566
+ - name: duration
567
+ dtype: string
568
  - name: negative_sample_raw_units
569
+ list:
570
  - name: hubert
571
  dtype: string
572
+ - name: pitch
573
+ dtype: string
574
+ - name: duration
575
+ dtype: string
576
  - name: prompt_audio
577
  dtype:
578
  audio:
 
590
  - name: negative_continuation_tokenwise_loss
591
  list: float64
592
  - name: positive_continuation_raw_units
593
+ list:
594
  - name: hubert
595
  dtype: string
596
+ - name: pitch
597
+ dtype: string
598
+ - name: duration
599
+ dtype: string
600
  - name: negative_continuation_raw_units
601
+ list:
602
  - name: hubert
603
  dtype: string
604
+ - name: pitch
605
+ dtype: string
606
+ - name: duration
607
+ dtype: string
608
  - name: continuation_audio_positive
609
  dtype:
610
  audio:
 
615
  sampling_rate: 16000
616
  splits:
617
  - name: train
618
+ num_bytes: 210176003
619
  num_examples: 200
620
+ download_size: 210176003
621
+ dataset_size: 210176003
622
  - config_name: speaker_consistency
623
  features:
624
  - name: task
 
644
  audio:
645
  sampling_rate: 16000
646
  - name: positive_sample_raw_units
647
+ list:
648
  - name: hubert
649
  dtype: string
650
+ - name: pitch
651
+ dtype: string
652
+ - name: duration
653
+ dtype: string
654
  - name: negative_sample_raw_units
655
+ list:
656
  - name: hubert
657
  dtype: string
658
+ - name: pitch
659
+ dtype: string
660
+ - name: duration
661
+ dtype: string
662
  - name: prompt_audio
663
  dtype:
664
  audio:
 
676
  - name: negative_continuation_tokenwise_loss
677
  list: float64
678
  - name: positive_continuation_raw_units
679
+ list:
680
  - name: hubert
681
  dtype: string
682
+ - name: pitch
683
+ dtype: string
684
+ - name: duration
685
+ dtype: string
686
  - name: negative_continuation_raw_units
687
+ list:
688
  - name: hubert
689
  dtype: string
690
+ - name: pitch
691
+ dtype: string
692
+ - name: duration
693
+ dtype: string
694
  - name: continuation_audio_positive
695
  dtype:
696
  audio:
 
701
  sampling_rate: 16000
702
  splits:
703
  - name: train
704
+ num_bytes: 209910220
705
  num_examples: 200
706
+ download_size: 209910220
707
+ dataset_size: 209910220
708
  ---
709
 
710
  # SALMon Normalized Dataset