Hugging Face's logo

Datasets:
sujalappa
/
eka_filtered_data

Tasks:
Automatic Speech Recognition
Text-to-Speech
Modalities:
Audio
Text
Formats:
parquet
Languages:
English
Size:
< 1K
Tags:
dataset
audio
speech
asr
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
Community
1
sujalappa commited on
Commit
64217a6
·
verified ·
1 Parent(s): 762ecf8

Add dataset card with subset configurations

Browse files
Files changed (1) hide show
  1. README.md +484 -39
README.md CHANGED
@@ -1,43 +1,488 @@
1
  ---
 
2
  dataset_info:
3
- features:
4
- - name: md5_text
5
- dtype: string
6
- - name: file_name
7
- dtype: string
8
- - name: audio
9
- dtype:
10
- audio:
11
- sampling_rate: 16000
12
- - name: md5_audio
13
- dtype: string
14
- - name: duration
15
- dtype: float32
16
- - name: text
17
- dtype: string
18
- - name: audio_language
19
- dtype: string
20
- - name: text_language
21
- dtype: string
22
- - name: session_id
23
- dtype: string
24
- - name: speaker
25
- dtype: string
26
- - name: type_concept
27
- dtype: string
28
- - name: recording_context
29
- dtype: string
30
- - name: medical_entities
31
- dtype: string
32
- splits:
33
- - name: test
34
- num_bytes: 7588994.11875
35
- num_examples: 54
36
- download_size: 8949547
37
- dataset_size: 7588994.11875
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38
  configs:
39
- - config_name: default
40
- data_files:
41
- - split: test
42
- path: data/test-*
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
43
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ license: mit
3
  dataset_info:
4
+ - config_name: default
5
+ features:
6
+ - name: md5_text
7
+ dtype: string
8
+ - name: file_name
9
+ dtype: string
10
+ - name: md5_audio
11
+ dtype: string
12
+ - name: duration
13
+ dtype: float32
14
+ - name: text
15
+ dtype: string
16
+ - name: audio_language
17
+ dtype: string
18
+ - name: text_language
19
+ dtype: string
20
+ - name: session_id
21
+ dtype: string
22
+ - name: speaker
23
+ dtype: string
24
+ - name: type_concept
25
+ dtype: string
26
+ - name: recording_context
27
+ dtype: string
28
+ - name: medical_entities
29
+ dtype: string
30
+ - name: audio
31
+ dtype:
32
+ audio:
33
+ sampling_rate: 16000
34
+ splits:
35
+ download_size: 0
36
+ dataset_size: 0
37
+ - config_name: original
38
+ features:
39
+ - name: md5_text
40
+ dtype: string
41
+ - name: file_name
42
+ dtype: string
43
+ - name: md5_audio
44
+ dtype: string
45
+ - name: duration
46
+ dtype: float32
47
+ - name: text
48
+ dtype: string
49
+ - name: audio_language
50
+ dtype: string
51
+ - name: text_language
52
+ dtype: string
53
+ - name: session_id
54
+ dtype: string
55
+ - name: speaker
56
+ dtype: string
57
+ - name: type_concept
58
+ dtype: string
59
+ - name: recording_context
60
+ dtype: string
61
+ - name: medical_entities
62
+ dtype: string
63
+ - name: audio
64
+ dtype:
65
+ audio:
66
+ sampling_rate: 16000
67
+ splits:
68
+ - name: test
69
+ num_bytes: 1000000
70
+ num_examples: 2
71
+ download_size: 850000
72
+ dataset_size: 1000000
73
+ - config_name: noisy-bg-snr-10
74
+ features:
75
+ - name: md5_text
76
+ dtype: string
77
+ - name: file_name
78
+ dtype: string
79
+ - name: md5_audio
80
+ dtype: string
81
+ - name: duration
82
+ dtype: float32
83
+ - name: text
84
+ dtype: string
85
+ - name: audio_language
86
+ dtype: string
87
+ - name: text_language
88
+ dtype: string
89
+ - name: session_id
90
+ dtype: string
91
+ - name: speaker
92
+ dtype: string
93
+ - name: type_concept
94
+ dtype: string
95
+ - name: recording_context
96
+ dtype: string
97
+ - name: medical_entities
98
+ dtype: string
99
+ - name: audio
100
+ dtype:
101
+ audio:
102
+ sampling_rate: 16000
103
+ splits:
104
+ - name: test
105
+ num_bytes: 1000000
106
+ num_examples: 2
107
+ download_size: 850000
108
+ dataset_size: 1000000
109
+ - config_name: noisy-bg-snr-40
110
+ features:
111
+ - name: md5_text
112
+ dtype: string
113
+ - name: file_name
114
+ dtype: string
115
+ - name: md5_audio
116
+ dtype: string
117
+ - name: duration
118
+ dtype: float32
119
+ - name: text
120
+ dtype: string
121
+ - name: audio_language
122
+ dtype: string
123
+ - name: text_language
124
+ dtype: string
125
+ - name: session_id
126
+ dtype: string
127
+ - name: speaker
128
+ dtype: string
129
+ - name: type_concept
130
+ dtype: string
131
+ - name: recording_context
132
+ dtype: string
133
+ - name: medical_entities
134
+ dtype: string
135
+ - name: audio
136
+ dtype:
137
+ audio:
138
+ sampling_rate: 16000
139
+ splits:
140
+ - name: test
141
+ num_bytes: 1000000
142
+ num_examples: 2
143
+ download_size: 850000
144
+ dataset_size: 1000000
145
+ - config_name: noisy-short-snr-4
146
+ features:
147
+ - name: md5_text
148
+ dtype: string
149
+ - name: file_name
150
+ dtype: string
151
+ - name: md5_audio
152
+ dtype: string
153
+ - name: duration
154
+ dtype: float32
155
+ - name: text
156
+ dtype: string
157
+ - name: audio_language
158
+ dtype: string
159
+ - name: text_language
160
+ dtype: string
161
+ - name: session_id
162
+ dtype: string
163
+ - name: speaker
164
+ dtype: string
165
+ - name: type_concept
166
+ dtype: string
167
+ - name: recording_context
168
+ dtype: string
169
+ - name: medical_entities
170
+ dtype: string
171
+ - name: audio
172
+ dtype:
173
+ audio:
174
+ sampling_rate: 16000
175
+ splits:
176
+ - name: test
177
+ num_bytes: 1000000
178
+ num_examples: 2
179
+ download_size: 850000
180
+ dataset_size: 1000000
181
+ - config_name: noisy-short-snr-10
182
+ features:
183
+ - name: md5_text
184
+ dtype: string
185
+ - name: file_name
186
+ dtype: string
187
+ - name: md5_audio
188
+ dtype: string
189
+ - name: duration
190
+ dtype: float32
191
+ - name: text
192
+ dtype: string
193
+ - name: audio_language
194
+ dtype: string
195
+ - name: text_language
196
+ dtype: string
197
+ - name: session_id
198
+ dtype: string
199
+ - name: speaker
200
+ dtype: string
201
+ - name: type_concept
202
+ dtype: string
203
+ - name: recording_context
204
+ dtype: string
205
+ - name: medical_entities
206
+ dtype: string
207
+ - name: audio
208
+ dtype:
209
+ audio:
210
+ sampling_rate: 16000
211
+ splits:
212
+ - name: test
213
+ num_bytes: 1000000
214
+ num_examples: 2
215
+ download_size: 850000
216
+ dataset_size: 1000000
217
+ - config_name: denoised-bg-snr-10
218
+ features:
219
+ - name: md5_text
220
+ dtype: string
221
+ - name: file_name
222
+ dtype: string
223
+ - name: md5_audio
224
+ dtype: string
225
+ - name: duration
226
+ dtype: float32
227
+ - name: text
228
+ dtype: string
229
+ - name: audio_language
230
+ dtype: string
231
+ - name: text_language
232
+ dtype: string
233
+ - name: session_id
234
+ dtype: string
235
+ - name: speaker
236
+ dtype: string
237
+ - name: type_concept
238
+ dtype: string
239
+ - name: recording_context
240
+ dtype: string
241
+ - name: medical_entities
242
+ dtype: string
243
+ - name: audio
244
+ dtype:
245
+ audio:
246
+ sampling_rate: 16000
247
+ splits:
248
+ - name: test
249
+ num_bytes: 1000000
250
+ num_examples: 2
251
+ download_size: 850000
252
+ dataset_size: 1000000
253
+ - config_name: denoised-bg-snr-40
254
+ features:
255
+ - name: md5_text
256
+ dtype: string
257
+ - name: file_name
258
+ dtype: string
259
+ - name: md5_audio
260
+ dtype: string
261
+ - name: duration
262
+ dtype: float32
263
+ - name: text
264
+ dtype: string
265
+ - name: audio_language
266
+ dtype: string
267
+ - name: text_language
268
+ dtype: string
269
+ - name: session_id
270
+ dtype: string
271
+ - name: speaker
272
+ dtype: string
273
+ - name: type_concept
274
+ dtype: string
275
+ - name: recording_context
276
+ dtype: string
277
+ - name: medical_entities
278
+ dtype: string
279
+ - name: audio
280
+ dtype:
281
+ audio:
282
+ sampling_rate: 16000
283
+ splits:
284
+ - name: test
285
+ num_bytes: 1000000
286
+ num_examples: 2
287
+ download_size: 850000
288
+ dataset_size: 1000000
289
+ - config_name: denoised-short-snr-4
290
+ features:
291
+ - name: md5_text
292
+ dtype: string
293
+ - name: file_name
294
+ dtype: string
295
+ - name: md5_audio
296
+ dtype: string
297
+ - name: duration
298
+ dtype: float32
299
+ - name: text
300
+ dtype: string
301
+ - name: audio_language
302
+ dtype: string
303
+ - name: text_language
304
+ dtype: string
305
+ - name: session_id
306
+ dtype: string
307
+ - name: speaker
308
+ dtype: string
309
+ - name: type_concept
310
+ dtype: string
311
+ - name: recording_context
312
+ dtype: string
313
+ - name: medical_entities
314
+ dtype: string
315
+ - name: audio
316
+ dtype:
317
+ audio:
318
+ sampling_rate: 16000
319
+ splits:
320
+ - name: test
321
+ num_bytes: 1000000
322
+ num_examples: 2
323
+ download_size: 850000
324
+ dataset_size: 1000000
325
+ - config_name: denoised-short-snr-10
326
+ features:
327
+ - name: md5_text
328
+ dtype: string
329
+ - name: file_name
330
+ dtype: string
331
+ - name: md5_audio
332
+ dtype: string
333
+ - name: duration
334
+ dtype: float32
335
+ - name: text
336
+ dtype: string
337
+ - name: audio_language
338
+ dtype: string
339
+ - name: text_language
340
+ dtype: string
341
+ - name: session_id
342
+ dtype: string
343
+ - name: speaker
344
+ dtype: string
345
+ - name: type_concept
346
+ dtype: string
347
+ - name: recording_context
348
+ dtype: string
349
+ - name: medical_entities
350
+ dtype: string
351
+ - name: audio
352
+ dtype:
353
+ audio:
354
+ sampling_rate: 16000
355
+ splits:
356
+ - name: test
357
+ num_bytes: 1000000
358
+ num_examples: 2
359
+ download_size: 850000
360
+ dataset_size: 1000000
361
  configs:
362
+ - config_name: original
363
+ data_files:
364
+ - split: test
365
+ path: original/test-*
366
+ - config_name: noisy-bg-snr-10
367
+ data_files:
368
+ - split: test
369
+ path: noisy-bg-snr-10/test-*
370
+ - config_name: noisy-bg-snr-40
371
+ data_files:
372
+ - split: test
373
+ path: noisy-bg-snr-40/test-*
374
+ - config_name: noisy-short-snr-4
375
+ data_files:
376
+ - split: test
377
+ path: noisy-short-snr-4/test-*
378
+ - config_name: noisy-short-snr-10
379
+ data_files:
380
+ - split: test
381
+ path: noisy-short-snr-10/test-*
382
+ - config_name: denoised-bg-snr-10
383
+ data_files:
384
+ - split: test
385
+ path: denoised-bg-snr-10/test-*
386
+ - config_name: denoised-bg-snr-40
387
+ data_files:
388
+ - split: test
389
+ path: denoised-bg-snr-40/test-*
390
+ - config_name: denoised-short-snr-4
391
+ data_files:
392
+ - split: test
393
+ path: denoised-short-snr-4/test-*
394
+ - config_name: denoised-short-snr-10
395
+ data_files:
396
+ - split: test
397
+ path: denoised-short-snr-10/test-*
398
+ task_categories:
399
+ - automatic-speech-recognition
400
+ - text-to-speech
401
+ language:
402
+ - en
403
+ tags:
404
+ - dataset
405
+ - audio
406
+ - speech
407
+ - asr
408
+ pretty_name: Eka Medical Asr Sample Noise Eval Dataset
409
+ size_categories:
410
+ - n<1K
411
  ---
412
+
413
+ # Eka Medical Asr Sample Noise Eval Dataset
414
+
415
+ ## Dataset Description
416
+
417
+ This dataset contains 18 samples organized across multiple splits and 9 subsets.
418
+ The dataset includes audio data.
419
+
420
+ ## Dataset Structure
421
+
422
+ ### Subsets
423
+
424
+ This dataset includes the following subsets:
425
+
426
+ - **original**: 2 samples
427
+ - test: 2 samples
428
+ - **noisy-bg-snr-10**: 2 samples
429
+ - test: 2 samples
430
+ - **noisy-bg-snr-40**: 2 samples
431
+ - test: 2 samples
432
+ - **noisy-short-snr-4**: 2 samples
433
+ - test: 2 samples
434
+ - **noisy-short-snr-10**: 2 samples
435
+ - test: 2 samples
436
+ - **denoised-bg-snr-10**: 2 samples
437
+ - test: 2 samples
438
+ - **denoised-bg-snr-40**: 2 samples
439
+ - test: 2 samples
440
+ - **denoised-short-snr-4**: 2 samples
441
+ - test: 2 samples
442
+ - **denoised-short-snr-10**: 2 samples
443
+ - test: 2 samples
444
+
445
+ ### Usage
446
+
447
+ Load specific subset and split:
448
+ ```python
449
+ from datasets import load_dataset
450
+
451
+ # Load specific subset and split
452
+ dataset = load_dataset('sujalappa/eka_filtered_data', 'original', split='train')
453
+
454
+ # Load all splits from a subset
455
+ dataset = load_dataset('sujalappa/eka_filtered_data', 'original')
456
+
457
+ # Load everything
458
+ dataset = load_dataset('sujalappa/eka_filtered_data')
459
+ ```
460
+
461
+ ## Dataset Creation
462
+
463
+ This dataset was created using StreamableDatasetManager on 2025-09-15T00:25:37.036342.
464
+
465
+ ### Data Fields
466
+
467
+ The dataset includes the following columns:
468
+ - **md5_text**: String data
469
+ - **file_name**: String data
470
+ - **md5_audio**: String data
471
+ - **duration**: Float32 data
472
+ - **text**: String data
473
+ - **audio_language**: String data
474
+ - **text_language**: String data
475
+ - **session_id**: String data
476
+ - **speaker**: String data
477
+ - **type_concept**: String data
478
+ - **recording_context**: String data
479
+ - **medical_entities**: String data
480
+ - **audio**: Audio data (16kHz sampling rate)
481
+
482
+ ## Technical Details
483
+
484
+ - **Total samples**: 18
485
+ - **Shard length**: 5
486
+ - **Number of subsets**: 9
487
+ - **Number of splits**: 2
488
+ - **Audio format**: 16kHz sampling rate