Hugging Face's logo

nvidia
/
canary-1b

Automatic Speech Recognition
NeMo
PyTorch
automatic-speech-translation
speech
audio
Transformer
FastConformer
Conformer
NeMo
hf-asr-leaderboard
Eval Results
Model card Files
xet
Community
30
sabato-nocera commited on
Commit
43511d0
·
verified ·
1 Parent(s): d68d53f

add AIBOM

Browse files

Dear model owner(s),
We are a group of researchers investigating the usefulness of sharing AIBOMs (Artificial Intelligence Bill of Materials) to document AI models – AIBOMs are machine-readable structured lists of components (e.g., datasets and models) used to enhance transparency in AI-model supply chains.

To pursue the above-mentioned objective, we identified popular models on HuggingFace and, based on your model card (and some configuration information available in HuggingFace), we generated your AIBOM according to the CyclonDX (v1.6) standard (see https://cyclonedx.org/docs/1.6/json/). AIBOMs are generated as JSON files by using the following open-source supporting tool: https://github.com/MSR4SBOM/ALOHA (technical details are available in the research paper: https://github.com/MSR4SBOM/ALOHA/blob/main/ALOHA.pdf).

The JSON file in this pull request is your AIBOM (see https://github.com/MSR4SBOM/ALOHA/blob/main/documentation.json for details on its structure).

Clearly, the submitted AIBOM matches the current model information, yet it can be easily regenerated when the model evolves, using the aforementioned AIBOM generator tool.

We open this pull request containing an AIBOM of your AI model, and hope it will be considered. We would also like to hear your opinion on the usefulness (or not) of AIBOM by answering a 3-minute anonymous survey: https://forms.gle/WGffSQD5dLoWttEe7.

Thanks in advance, and regards,
Riccardo D’Avino, Fatima Ahmed, Sabato Nocera, Simone Romano, Giuseppe Scanniello (University of Salerno, Italy),
Massimiliano Di Penta (University of Sannio, Italy),
The MSR4SBOM team

Files changed (1) hide show
  1. nvidia_canary-1b.json +649 -0
nvidia_canary-1b.json ADDED
@@ -0,0 +1,649 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bomFormat": "CycloneDX",
3
+ "specVersion": "1.6",
4
+ "serialNumber": "urn:uuid:dd499724-872b-4392-817f-8511a1cd9113",
5
+ "version": 1,
6
+ "metadata": {
7
+ "timestamp": "2025-06-05T09:37:28.708234+00:00",
8
+ "component": {
9
+ "type": "machine-learning-model",
10
+ "bom-ref": "nvidia/canary-1b-58688bd0-57c9-5752-b615-3abc9265c7a1",
11
+ "name": "nvidia/canary-1b",
12
+ "externalReferences": [
13
+ {
14
+ "url": "https://huggingface.co/nvidia/canary-1b",
15
+ "type": "documentation"
16
+ }
17
+ ],
18
+ "modelCard": {
19
+ "modelParameters": {
20
+ "task": "automatic-speech-recognition",
21
+ "datasets": [
22
+ {
23
+ "ref": "librispeech_asr-7baf0ed9-b50c-5f93-8c23-49a2b8749c19"
24
+ },
25
+ {
26
+ "ref": "fisher_corpus-a0c6e2c1-e876-5c66-89b2-cb93697b2a1c"
27
+ },
28
+ {
29
+ "ref": "Switchboard-1-b54b0d1d-3005-514e-9668-98d3c19f793f"
30
+ },
31
+ {
32
+ "ref": "WSJ-0-095442e6-ea65-5f6d-b360-432c7a2f501d"
33
+ },
34
+ {
35
+ "ref": "WSJ-1-0ef003e6-350d-50bb-9df7-9491b0c9b0b3"
36
+ },
37
+ {
38
+ "ref": "National-Singapore-Corpus-Part-1-1fbb2914-35aa-5126-9a84-a8b77169254c"
39
+ },
40
+ {
41
+ "ref": "National-Singapore-Corpus-Part-6-4f83cf7f-3026-5a77-ae37-28a73d4abc24"
42
+ },
43
+ {
44
+ "ref": "vctk-d80444bd-bcc6-5c25-8570-061bb96dae38"
45
+ },
46
+ {
47
+ "ref": "voxpopuli-15fb6343-a710-54f9-842b-3a1b43d6a630"
48
+ },
49
+ {
50
+ "ref": "europarl-7e07ffed-425e-5e05-8847-08a1899f0ac1"
51
+ },
52
+ {
53
+ "ref": "multilingual_librispeech-f260ef31-1d5d-54fe-8e61-88c397c0b7ce"
54
+ },
55
+ {
56
+ "ref": "mozilla-foundation/common_voice_8_0-a994a71f-f9f5-5f65-a3fa-51a56293cd8e"
57
+ },
58
+ {
59
+ "ref": "MLCommons/peoples_speech-f88dc766-1de0-51c6-865d-16930ec19be6"
60
+ }
61
+ ]
62
+ },
63
+ "properties": [
64
+ {
65
+ "name": "library_name",
66
+ "value": "nemo"
67
+ }
68
+ ],
69
+ "quantitativeAnalysis": {
70
+ "performanceMetrics": [
71
+ {
72
+ "slice": "dataset: librispeech_asr, split: test, config: other",
73
+ "type": "wer",
74
+ "value": 2.89
75
+ },
76
+ {
77
+ "slice": "dataset: kensho/spgispeech, split: test, config: test",
78
+ "type": "wer",
79
+ "value": 4.79
80
+ },
81
+ {
82
+ "slice": "dataset: mozilla-foundation/common_voice_16_1, split: test, config: en",
83
+ "type": "wer",
84
+ "value": 7.97
85
+ },
86
+ {
87
+ "slice": "dataset: mozilla-foundation/common_voice_16_1, split: test, config: de",
88
+ "type": "wer",
89
+ "value": 4.61
90
+ },
91
+ {
92
+ "slice": "dataset: mozilla-foundation/common_voice_16_1, split: test, config: es",
93
+ "type": "wer",
94
+ "value": 3.99
95
+ },
96
+ {
97
+ "slice": "dataset: mozilla-foundation/common_voice_16_1, split: test, config: fr",
98
+ "type": "wer",
99
+ "value": 6.53
100
+ },
101
+ {
102
+ "slice": "dataset: google/fleurs, split: test, config: en_us",
103
+ "type": "bleu",
104
+ "value": 32.15
105
+ },
106
+ {
107
+ "slice": "dataset: google/fleurs, split: test, config: en_us",
108
+ "type": "bleu",
109
+ "value": 22.66
110
+ },
111
+ {
112
+ "slice": "dataset: google/fleurs, split: test, config: en_us",
113
+ "type": "bleu",
114
+ "value": 40.76
115
+ },
116
+ {
117
+ "slice": "dataset: google/fleurs, split: test, config: de_de",
118
+ "type": "bleu",
119
+ "value": 33.98
120
+ },
121
+ {
122
+ "slice": "dataset: google/fleurs, split: test, config: es_419",
123
+ "type": "bleu",
124
+ "value": 21.8
125
+ },
126
+ {
127
+ "slice": "dataset: google/fleurs, split: test, config: fr_fr",
128
+ "type": "bleu",
129
+ "value": 30.95
130
+ },
131
+ {
132
+ "slice": "dataset: covost2, split: test, config: de_de",
133
+ "type": "bleu",
134
+ "value": 37.67
135
+ },
136
+ {
137
+ "slice": "dataset: covost2, split: test, config: es_419",
138
+ "type": "bleu",
139
+ "value": 40.7
140
+ },
141
+ {
142
+ "slice": "dataset: covost2, split: test, config: fr_fr",
143
+ "type": "bleu",
144
+ "value": 40.42
145
+ }
146
+ ]
147
+ }
148
+ },
149
+ "authors": [
150
+ {
151
+ "name": "nvidia"
152
+ }
153
+ ],
154
+ "licenses": [
155
+ {
156
+ "license": {
157
+ "id": "CC-BY-NC-4.0",
158
+ "url": "https://spdx.org/licenses/CC-BY-NC-4.0.html"
159
+ }
160
+ }
161
+ ],
162
+ "tags": [
163
+ "nemo",
164
+ "automatic-speech-recognition",
165
+ "automatic-speech-translation",
166
+ "speech",
167
+ "audio",
168
+ "Transformer",
169
+ "FastConformer",
170
+ "Conformer",
171
+ "pytorch",
172
+ "NeMo",
173
+ "hf-asr-leaderboard",
174
+ "en",
175
+ "de",
176
+ "es",
177
+ "fr",
178
+ "dataset:librispeech_asr",
179
+ "dataset:fisher_corpus",
180
+ "dataset:Switchboard-1",
181
+ "dataset:WSJ-0",
182
+ "dataset:WSJ-1",
183
+ "dataset:National-Singapore-Corpus-Part-1",
184
+ "dataset:National-Singapore-Corpus-Part-6",
185
+ "dataset:vctk",
186
+ "dataset:voxpopuli",
187
+ "dataset:europarl",
188
+ "dataset:multilingual_librispeech",
189
+ "dataset:mozilla-foundation/common_voice_8_0",
190
+ "dataset:MLCommons/peoples_speech",
191
+ "arxiv:2305.05084",
192
+ "arxiv:1706.03762",
193
+ "license:cc-by-nc-4.0",
194
+ "model-index",
195
+ "region:us"
196
+ ]
197
+ }
198
+ },
199
+ "components": [
200
+ {
201
+ "type": "data",
202
+ "bom-ref": "librispeech_asr-7baf0ed9-b50c-5f93-8c23-49a2b8749c19",
203
+ "name": "librispeech_asr",
204
+ "data": [
205
+ {
206
+ "type": "dataset",
207
+ "bom-ref": "librispeech_asr-7baf0ed9-b50c-5f93-8c23-49a2b8749c19",
208
+ "name": "librispeech_asr",
209
+ "contents": {
210
+ "url": "https://huggingface.co/datasets/librispeech_asr",
211
+ "properties": [
212
+ {
213
+ "name": "task_categories",
214
+ "value": "automatic-speech-recognition, audio-classification"
215
+ },
216
+ {
217
+ "name": "task_ids",
218
+ "value": "speaker-identification"
219
+ },
220
+ {
221
+ "name": "language",
222
+ "value": "en"
223
+ },
224
+ {
225
+ "name": "size_categories",
226
+ "value": "100K<n<1M"
227
+ },
228
+ {
229
+ "name": "annotations_creators",
230
+ "value": "expert-generated"
231
+ },
232
+ {
233
+ "name": "language_creators",
234
+ "value": "crowdsourced, expert-generated"
235
+ },
236
+ {
237
+ "name": "pretty_name",
238
+ "value": "LibriSpeech"
239
+ },
240
+ {
241
+ "name": "source_datasets",
242
+ "value": "original"
243
+ },
244
+ {
245
+ "name": "paperswithcode_id",
246
+ "value": "librispeech-1"
247
+ },
248
+ {
249
+ "name": "license",
250
+ "value": "cc-by-4.0"
251
+ }
252
+ ]
253
+ },
254
+ "governance": {
255
+ "owners": [
256
+ {
257
+ "organization": {
258
+ "name": "openslr",
259
+ "url": "https://huggingface.co/openslr"
260
+ }
261
+ }
262
+ ]
263
+ },
264
+ "description": "LibriSpeech is a corpus of approximately 1000 hours of read English speech with sampling rate of 16 kHz,\nprepared by Vassil Panayotov with the assistance of Daniel Povey. The data is derived from read\naudiobooks from the LibriVox project, and has been carefully segmented and aligned.87"
265
+ }
266
+ ]
267
+ },
268
+ {
269
+ "type": "data",
270
+ "bom-ref": "fisher_corpus-a0c6e2c1-e876-5c66-89b2-cb93697b2a1c",
271
+ "name": "fisher_corpus",
272
+ "data": [
273
+ {
274
+ "type": "dataset",
275
+ "bom-ref": "fisher_corpus-a0c6e2c1-e876-5c66-89b2-cb93697b2a1c",
276
+ "name": "fisher_corpus"
277
+ }
278
+ ]
279
+ },
280
+ {
281
+ "type": "data",
282
+ "bom-ref": "Switchboard-1-b54b0d1d-3005-514e-9668-98d3c19f793f",
283
+ "name": "Switchboard-1",
284
+ "data": [
285
+ {
286
+ "type": "dataset",
287
+ "bom-ref": "Switchboard-1-b54b0d1d-3005-514e-9668-98d3c19f793f",
288
+ "name": "Switchboard-1"
289
+ }
290
+ ]
291
+ },
292
+ {
293
+ "type": "data",
294
+ "bom-ref": "WSJ-0-095442e6-ea65-5f6d-b360-432c7a2f501d",
295
+ "name": "WSJ-0",
296
+ "data": [
297
+ {
298
+ "type": "dataset",
299
+ "bom-ref": "WSJ-0-095442e6-ea65-5f6d-b360-432c7a2f501d",
300
+ "name": "WSJ-0"
301
+ }
302
+ ]
303
+ },
304
+ {
305
+ "type": "data",
306
+ "bom-ref": "WSJ-1-0ef003e6-350d-50bb-9df7-9491b0c9b0b3",
307
+ "name": "WSJ-1",
308
+ "data": [
309
+ {
310
+ "type": "dataset",
311
+ "bom-ref": "WSJ-1-0ef003e6-350d-50bb-9df7-9491b0c9b0b3",
312
+ "name": "WSJ-1"
313
+ }
314
+ ]
315
+ },
316
+ {
317
+ "type": "data",
318
+ "bom-ref": "National-Singapore-Corpus-Part-1-1fbb2914-35aa-5126-9a84-a8b77169254c",
319
+ "name": "National-Singapore-Corpus-Part-1",
320
+ "data": [
321
+ {
322
+ "type": "dataset",
323
+ "bom-ref": "National-Singapore-Corpus-Part-1-1fbb2914-35aa-5126-9a84-a8b77169254c",
324
+ "name": "National-Singapore-Corpus-Part-1"
325
+ }
326
+ ]
327
+ },
328
+ {
329
+ "type": "data",
330
+ "bom-ref": "National-Singapore-Corpus-Part-6-4f83cf7f-3026-5a77-ae37-28a73d4abc24",
331
+ "name": "National-Singapore-Corpus-Part-6",
332
+ "data": [
333
+ {
334
+ "type": "dataset",
335
+ "bom-ref": "National-Singapore-Corpus-Part-6-4f83cf7f-3026-5a77-ae37-28a73d4abc24",
336
+ "name": "National-Singapore-Corpus-Part-6"
337
+ }
338
+ ]
339
+ },
340
+ {
341
+ "type": "data",
342
+ "bom-ref": "vctk-d80444bd-bcc6-5c25-8570-061bb96dae38",
343
+ "name": "vctk",
344
+ "data": [
345
+ {
346
+ "type": "dataset",
347
+ "bom-ref": "vctk-d80444bd-bcc6-5c25-8570-061bb96dae38",
348
+ "name": "vctk",
349
+ "contents": {
350
+ "url": "https://huggingface.co/datasets/vctk",
351
+ "properties": [
352
+ {
353
+ "name": "task_categories",
354
+ "value": "automatic-speech-recognition, text-to-speech, text-to-audio"
355
+ },
356
+ {
357
+ "name": "task_ids",
358
+ "value": ""
359
+ },
360
+ {
361
+ "name": "language",
362
+ "value": "en"
363
+ },
364
+ {
365
+ "name": "size_categories",
366
+ "value": "10K<n<100K"
367
+ },
368
+ {
369
+ "name": "annotations_creators",
370
+ "value": "expert-generated"
371
+ },
372
+ {
373
+ "name": "language_creators",
374
+ "value": "crowdsourced"
375
+ },
376
+ {
377
+ "name": "pretty_name",
378
+ "value": "VCTK"
379
+ },
380
+ {
381
+ "name": "source_datasets",
382
+ "value": "original"
383
+ },
384
+ {
385
+ "name": "paperswithcode_id",
386
+ "value": "vctk"
387
+ },
388
+ {
389
+ "name": "license",
390
+ "value": "cc-by-4.0"
391
+ }
392
+ ]
393
+ },
394
+ "governance": {
395
+ "owners": [
396
+ {
397
+ "organization": {
398
+ "name": "CSTR-Edinburgh",
399
+ "url": "https://huggingface.co/CSTR-Edinburgh"
400
+ }
401
+ }
402
+ ]
403
+ },
404
+ "description": "The CSTR VCTK Corpus includes speech data uttered by 110 English speakers with various accents."
405
+ }
406
+ ]
407
+ },
408
+ {
409
+ "type": "data",
410
+ "bom-ref": "voxpopuli-15fb6343-a710-54f9-842b-3a1b43d6a630",
411
+ "name": "voxpopuli",
412
+ "data": [
413
+ {
414
+ "type": "dataset",
415
+ "bom-ref": "voxpopuli-15fb6343-a710-54f9-842b-3a1b43d6a630",
416
+ "name": "voxpopuli"
417
+ }
418
+ ]
419
+ },
420
+ {
421
+ "type": "data",
422
+ "bom-ref": "europarl-7e07ffed-425e-5e05-8847-08a1899f0ac1",
423
+ "name": "europarl",
424
+ "data": [
425
+ {
426
+ "type": "dataset",
427
+ "bom-ref": "europarl-7e07ffed-425e-5e05-8847-08a1899f0ac1",
428
+ "name": "europarl"
429
+ }
430
+ ]
431
+ },
432
+ {
433
+ "type": "data",
434
+ "bom-ref": "multilingual_librispeech-f260ef31-1d5d-54fe-8e61-88c397c0b7ce",
435
+ "name": "multilingual_librispeech",
436
+ "data": [
437
+ {
438
+ "type": "dataset",
439
+ "bom-ref": "multilingual_librispeech-f260ef31-1d5d-54fe-8e61-88c397c0b7ce",
440
+ "name": "multilingual_librispeech",
441
+ "contents": {
442
+ "url": "https://huggingface.co/datasets/multilingual_librispeech",
443
+ "properties": [
444
+ {
445
+ "name": "task_categories",
446
+ "value": "automatic-speech-recognition, audio-classification"
447
+ },
448
+ {
449
+ "name": "task_ids",
450
+ "value": "speaker-identification"
451
+ },
452
+ {
453
+ "name": "language",
454
+ "value": "de, es, fr, it, nl, pl, pt"
455
+ },
456
+ {
457
+ "name": "size_categories",
458
+ "value": "100K<n<1M"
459
+ },
460
+ {
461
+ "name": "annotations_creators",
462
+ "value": "expert-generated"
463
+ },
464
+ {
465
+ "name": "language_creators",
466
+ "value": "crowdsourced, expert-generated"
467
+ },
468
+ {
469
+ "name": "pretty_name",
470
+ "value": "MultiLingual LibriSpeech"
471
+ },
472
+ {
473
+ "name": "source_datasets",
474
+ "value": "original"
475
+ },
476
+ {
477
+ "name": "paperswithcode_id",
478
+ "value": "librispeech-1"
479
+ },
480
+ {
481
+ "name": "license",
482
+ "value": "cc-by-4.0"
483
+ }
484
+ ]
485
+ },
486
+ "governance": {
487
+ "owners": [
488
+ {
489
+ "organization": {
490
+ "name": "legacy-datasets",
491
+ "url": "https://huggingface.co/legacy-datasets"
492
+ }
493
+ }
494
+ ]
495
+ },
496
+ "description": "Multilingual LibriSpeech (MLS) dataset is a large multilingual corpus suitable for speech research. The dataset is derived from read audiobooks from LibriVox and consists of 8 languages - English, German, Dutch, Spanish, French, Italian, Portuguese, Polish."
497
+ }
498
+ ]
499
+ },
500
+ {
501
+ "type": "data",
502
+ "bom-ref": "mozilla-foundation/common_voice_8_0-a994a71f-f9f5-5f65-a3fa-51a56293cd8e",
503
+ "name": "mozilla-foundation/common_voice_8_0",
504
+ "data": [
505
+ {
506
+ "type": "dataset",
507
+ "bom-ref": "mozilla-foundation/common_voice_8_0-a994a71f-f9f5-5f65-a3fa-51a56293cd8e",
508
+ "name": "mozilla-foundation/common_voice_8_0",
509
+ "contents": {
510
+ "url": "https://huggingface.co/datasets/mozilla-foundation/common_voice_8_0",
511
+ "properties": [
512
+ {
513
+ "name": "task_categories",
514
+ "value": "automatic-speech-recognition"
515
+ },
516
+ {
517
+ "name": "annotations_creators",
518
+ "value": "crowdsourced"
519
+ },
520
+ {
521
+ "name": "language_creators",
522
+ "value": "crowdsourced"
523
+ },
524
+ {
525
+ "name": "pretty_name",
526
+ "value": "Common Voice Corpus 8.0"
527
+ },
528
+ {
529
+ "name": "source_datasets",
530
+ "value": "extended|common_voice"
531
+ },
532
+ {
533
+ "name": "paperswithcode_id",
534
+ "value": "common-voice"
535
+ },
536
+ {
537
+ "name": "license",
538
+ "value": "cc0-1.0"
539
+ }
540
+ ]
541
+ },
542
+ "governance": {
543
+ "owners": [
544
+ {
545
+ "organization": {
546
+ "name": "mozilla-foundation",
547
+ "url": "https://huggingface.co/mozilla-foundation"
548
+ }
549
+ }
550
+ ]
551
+ },
552
+ "description": "\n\t\n\t\t\n\t\tDataset Card for Common Voice Corpus 8.0\n\t\n\n\n\t\n\t\t\n\t\tDataset Summary\n\t\n\nThe Common Voice dataset consists of a unique MP3 and corresponding text file. \nMany of the 18243 recorded hours in the dataset also include demographic metadata like age, sex, and accent \nthat can help improve the accuracy of speech recognition engines.\nThe dataset currently consists of 14122 validated hours in 87 languages, but more voices and languages are always added. \nTake a look at the Languages page to\u2026 See the full description on the dataset page: https://huggingface.co/datasets/mozilla-foundation/common_voice_8_0."
553
+ }
554
+ ]
555
+ },
556
+ {
557
+ "type": "data",
558
+ "bom-ref": "MLCommons/peoples_speech-f88dc766-1de0-51c6-865d-16930ec19be6",
559
+ "name": "MLCommons/peoples_speech",
560
+ "data": [
561
+ {
562
+ "type": "dataset",
563
+ "bom-ref": "MLCommons/peoples_speech-f88dc766-1de0-51c6-865d-16930ec19be6",
564
+ "name": "MLCommons/peoples_speech",
565
+ "contents": {
566
+ "url": "https://huggingface.co/datasets/MLCommons/peoples_speech",
567
+ "properties": [
568
+ {
569
+ "name": "task_categories",
570
+ "value": "automatic-speech-recognition"
571
+ },
572
+ {
573
+ "name": "task_ids",
574
+ "value": ""
575
+ },
576
+ {
577
+ "name": "language",
578
+ "value": "en"
579
+ },
580
+ {
581
+ "name": "size_categories",
582
+ "value": "1T<n"
583
+ },
584
+ {
585
+ "name": "annotations_creators",
586
+ "value": "crowdsourced, machine-generated"
587
+ },
588
+ {
589
+ "name": "language_creators",
590
+ "value": "crowdsourced, machine-generated"
591
+ },
592
+ {
593
+ "name": "pretty_name",
594
+ "value": "People's Speech"
595
+ },
596
+ {
597
+ "name": "source_datasets",
598
+ "value": "original"
599
+ },
600
+ {
601
+ "name": "configs",
602
+ "value": "Name of the dataset subset: clean {\"split\": \"train\", \"path\": \"clean/train-*\"}, {\"split\": \"validation\", \"path\": \"clean/validation-*\"}, {\"split\": \"test\", \"path\": \"clean/test-*\"}"
603
+ },
604
+ {
605
+ "name": "configs",
606
+ "value": "Name of the dataset subset: clean_sa {\"split\": \"train\", \"path\": \"clean_sa/train-*\"}, {\"split\": \"validation\", \"path\": \"clean_sa/validation-*\"}, {\"split\": \"test\", \"path\": \"clean_sa/test-*\"}"
607
+ },
608
+ {
609
+ "name": "configs",
610
+ "value": "Name of the dataset subset: dirty {\"split\": \"train\", \"path\": \"dirty/train-*\"}, {\"split\": \"validation\", \"path\": \"dirty/validation-*\"}, {\"split\": \"test\", \"path\": \"dirty/test-*\"}"
611
+ },
612
+ {
613
+ "name": "configs",
614
+ "value": "Name of the dataset subset: dirty_sa {\"split\": \"train\", \"path\": \"dirty_sa/train-*\"}, {\"split\": \"validation\", \"path\": \"dirty_sa/validation-*\"}, {\"split\": \"test\", \"path\": \"dirty_sa/test-*\"}"
615
+ },
616
+ {
617
+ "name": "configs",
618
+ "value": "Name of the dataset subset: microset {\"split\": \"train\", \"path\": \"microset/train-*\"}"
619
+ },
620
+ {
621
+ "name": "configs",
622
+ "value": "Name of the dataset subset: test {\"split\": \"test\", \"path\": \"test/test-*\"}"
623
+ },
624
+ {
625
+ "name": "configs",
626
+ "value": "Name of the dataset subset: validation {\"split\": \"validation\", \"path\": \"validation/validation-*\"}"
627
+ },
628
+ {
629
+ "name": "license",
630
+ "value": "cc-by-2.0, cc-by-2.5, cc-by-3.0, cc-by-4.0, cc-by-sa-3.0, cc-by-sa-4.0"
631
+ }
632
+ ]
633
+ },
634
+ "governance": {
635
+ "owners": [
636
+ {
637
+ "organization": {
638
+ "name": "MLCommons",
639
+ "url": "https://huggingface.co/MLCommons"
640
+ }
641
+ }
642
+ ]
643
+ },
644
+ "description": "\n\t\n\t\t\n\t\tDataset Card for People's Speech\n\t\n\n\n\t\n\t\t\n\t\tDataset Summary\n\t\n\nThe People's Speech Dataset is among the world's largest English speech recognition corpus today that is licensed for academic and commercial usage under CC-BY-SA and CC-BY 4.0. It includes 30,000+ hours of transcribed speech in English languages with a diverse set of speakers. This open dataset is large enough to train speech-to-text systems and crucially is available with a permissive license.\n\n\t\n\t\t\n\t\n\t\n\t\tSupported Tasks\u2026 See the full description on the dataset page: https://huggingface.co/datasets/MLCommons/peoples_speech."
645
+ }
646
+ ]
647
+ }
648
+ ]
649
+ }