Automatic Speech Recognition
ESPnet
Javanese
audio
Siddhant commited on
Commit
3feb9b8
1 Parent(s): 244fe2f

import from zenodo

Browse files
README.md ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - espnet
4
+ - audio
5
+ - automatic-speech-recognition
6
+ language: jv
7
+ datasets:
8
+ - jv_openslr35
9
+ license: cc-by-4.0
10
+ ---
11
+ ## ESPnet2 ASR pretrained model
12
+ ### `jv_openslr35`
13
+ ♻️ Imported from https://zenodo.org/record/5090139/
14
+
15
+ This model was trained by jv_openslr35 using jv_openslr35/asr1 recipe in [espnet](https://github.com/espnet/espnet/).
16
+ ### Demo: How to use in ESPnet2
17
+ ```python
18
+ # coming soon
19
+ ```
20
+ ### Citing ESPnet
21
+ ```BibTex
22
+ @inproceedings{watanabe2018espnet,
23
+ author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson {Enrique Yalta Soplin} and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
24
+ title={{ESPnet}: End-to-End Speech Processing Toolkit},
25
+ year={2018},
26
+ booktitle={Proceedings of Interspeech},
27
+ pages={2207--2211},
28
+ doi={10.21437/Interspeech.2018-1456},
29
+ url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
30
+ }
31
+
32
+ ```
33
+ or arXiv:
34
+ ```bibtex
35
+ @misc{watanabe2018espnet,
36
+ title={ESPnet: End-to-End Speech Processing Toolkit},
37
+ author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Enrique Yalta Soplin and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
38
+ year={2018},
39
+ eprint={1804.00015},
40
+ archivePrefix={arXiv},
41
+ primaryClass={cs.CL}
42
+ }
43
+ ```
data/token_list/bpe_unigram1000/bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d60cd15af7163345f662b25b17988d253b82d74a00d813249d0443133917811
3
+ size 252887
exp/asr_stats_raw_bpe1000/train/feats_stats.npz ADDED
Binary file (1.4 kB). View file
 
exp/asr_train_asr_raw_bpe1000/RESULTS.md ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!-- Generated by scripts/utils/show_asr_result.sh -->
2
+ # RESULTS
3
+ ## Environments
4
+ - date: `Fri Jul 9 17:56:55 PDT 2021`
5
+ - python version: `3.8.5 (default, Sep 4 2020, 07:30:14) [GCC 7.3.0]`
6
+ - espnet version: `espnet 0.10.0`
7
+ - pytorch version: `pytorch 1.8.1+cu102`
8
+ - Git hash: `5830a6b49a60ae10b8c113a2b9635ec2273fbdab`
9
+ - Commit date: `Fri Jul 9 08:36:41 2021 -0700`
10
+
11
+ ## asr_train_asr_raw_bpe1000
12
+ ### WER
13
+
14
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
15
+ |---|---|---|---|---|---|---|---|---|
16
+ |decode_asr_batch_size1_asr_model_valid.acc.best/dev_iban|473|11006|2.5|54.0|43.5|0.1|97.7|100.0|
17
+ |decode_asr_batch_size1_asr_model_valid.acc.best/java_test|1740|12117|81.9|16.4|1.7|0.9|19.0|52.3|
18
+ |decode_asr_batch_size1_asr_model_valid.acc.best/test_id_commonvoice|1643|9565|15.7|69.4|14.9|3.3|87.6|99.9|
19
+
20
+ ### CER
21
+
22
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
23
+ |---|---|---|---|---|---|---|---|---|
24
+ |decode_asr_batch_size1_asr_model_valid.acc.best/dev_iban|473|67025|53.0|17.6|29.3|5.4|52.3|100.0|
25
+ |decode_asr_batch_size1_asr_model_valid.acc.best/java_test|1740|80419|95.4|2.6|2.0|0.8|5.4|52.3|
26
+ |decode_asr_batch_size1_asr_model_valid.acc.best/test_id_commonvoice|1643|61563|69.4|14.3|16.3|5.0|35.6|99.9|
27
+
28
+ ### TER
29
+
30
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
31
+ |---|---|---|---|---|---|---|---|---|
32
+ |decode_asr_batch_size1_asr_model_valid.acc.best/dev_iban|473|22012|1.2|96.4|2.4|12.1|110.8|100.0|
33
+ |decode_asr_batch_size1_asr_model_valid.acc.best/java_test|1740|26604|84.6|10.6|4.8|1.2|16.6|52.3|
34
+ |decode_asr_batch_size1_asr_model_valid.acc.best/test_id_commonvoice|1643|27446|39.7|42.0|18.3|3.1|63.5|99.9|
35
+
exp/asr_train_asr_raw_bpe1000/config.yaml ADDED
@@ -0,0 +1,1155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config: conf/train_asr.yaml
2
+ print_config: false
3
+ log_level: INFO
4
+ dry_run: false
5
+ iterator_type: sequence
6
+ output_dir: exp/asr_train_asr_raw_bpe1000
7
+ ngpu: 1
8
+ seed: 0
9
+ num_workers: 1
10
+ num_att_plot: 3
11
+ dist_backend: nccl
12
+ dist_init_method: env://
13
+ dist_world_size: null
14
+ dist_rank: null
15
+ local_rank: 0
16
+ dist_master_addr: null
17
+ dist_master_port: null
18
+ dist_launcher: null
19
+ multiprocessing_distributed: false
20
+ cudnn_enabled: true
21
+ cudnn_benchmark: false
22
+ cudnn_deterministic: true
23
+ collect_stats: false
24
+ write_collected_feats: false
25
+ max_epoch: 200
26
+ patience: 20
27
+ val_scheduler_criterion:
28
+ - valid
29
+ - loss
30
+ early_stopping_criterion:
31
+ - valid
32
+ - loss
33
+ - min
34
+ best_model_criterion:
35
+ - - valid
36
+ - acc
37
+ - max
38
+ keep_nbest_models: 10
39
+ grad_clip: 5
40
+ grad_clip_type: 2.0
41
+ grad_noise: false
42
+ accum_grad: 2
43
+ no_forward_run: false
44
+ resume: true
45
+ train_dtype: float32
46
+ use_amp: false
47
+ log_interval: null
48
+ unused_parameters: false
49
+ use_tensorboard: true
50
+ use_wandb: false
51
+ wandb_project: null
52
+ wandb_id: null
53
+ pretrain_path: null
54
+ init_param: []
55
+ freeze_param: []
56
+ num_iters_per_epoch: null
57
+ batch_size: 32
58
+ valid_batch_size: null
59
+ batch_bins: 1000000
60
+ valid_batch_bins: null
61
+ train_shape_file:
62
+ - exp/asr_stats_raw_bpe1000/train/speech_shape
63
+ - exp/asr_stats_raw_bpe1000/train/text_shape.bpe
64
+ valid_shape_file:
65
+ - exp/asr_stats_raw_bpe1000/valid/speech_shape
66
+ - exp/asr_stats_raw_bpe1000/valid/text_shape.bpe
67
+ batch_type: folded
68
+ valid_batch_type: null
69
+ fold_length:
70
+ - 80000
71
+ - 150
72
+ sort_in_batch: descending
73
+ sort_batch: descending
74
+ multiple_iterator: false
75
+ chunk_length: 500
76
+ chunk_shift_ratio: 0.5
77
+ num_cache_chunks: 1024
78
+ train_data_path_and_name_and_type:
79
+ - - dump/raw/java_train/wav.scp
80
+ - speech
81
+ - sound
82
+ - - dump/raw/java_train/text
83
+ - text
84
+ - text
85
+ valid_data_path_and_name_and_type:
86
+ - - dump/raw/java_dev/wav.scp
87
+ - speech
88
+ - sound
89
+ - - dump/raw/java_dev/text
90
+ - text
91
+ - text
92
+ allow_variable_data_keys: false
93
+ max_cache_size: 0.0
94
+ max_cache_fd: 32
95
+ valid_max_cache_size: null
96
+ optim: adam
97
+ optim_conf:
98
+ lr: 10.0
99
+ scheduler: noamlr
100
+ scheduler_conf:
101
+ warmup_steps: 25000
102
+ token_list:
103
+ - <blank>
104
+ - <unk>
105
+ - NG
106
+ - E
107
+ - S
108
+ - I
109
+ - H
110
+ - N
111
+ - ▁
112
+ - K
113
+ - T
114
+ - L
115
+ - R
116
+ - ▁DI
117
+ - AN
118
+ - M
119
+ - É
120
+ - ▁A
121
+ - ▁ING
122
+ - A
123
+ - NA
124
+ - NE
125
+ - TA
126
+ - P
127
+ - D
128
+ - Y
129
+ - RA
130
+ - LA
131
+ - ▁KA
132
+ - SI
133
+ - ▁KARO
134
+ - U
135
+ - TI
136
+ - ▁LAN
137
+ - RI
138
+ - KA
139
+ - MA
140
+ - ▁MA
141
+ - ▁DHATENG
142
+ - ▁IKU
143
+ - ▁LUNGA
144
+ - YA
145
+ - ▁SA
146
+ - SA
147
+ - NI
148
+ - O
149
+ - ▁MENYA
150
+ - G
151
+ - WA
152
+ - ▁DOLAN
153
+ - ▁KE
154
+ - LI
155
+ - ▁SE
156
+ - DA
157
+ - GA
158
+ - ▁IKI
159
+ - ▁PA
160
+ - ▁SAK
161
+ - ▁S
162
+ - LE
163
+ - B
164
+ - JRONING
165
+ - ▁BA
166
+ - ▁KANG
167
+ - DI
168
+ - ▁TANSA
169
+ - JA
170
+ - RE
171
+ - BA
172
+ - ▁ELING
173
+ - PA
174
+ - ▁RA
175
+ - ▁DIWUJUDK
176
+ - ▁IMPI
177
+ - TU
178
+ - ▁PUNIKA
179
+ - ▁SU
180
+ - ▁I
181
+ - NÉ
182
+ - RO
183
+ - ▁ANA
184
+ - ▁ME
185
+ - ▁TA
186
+ - MI
187
+ - F
188
+ - TE
189
+ - KU
190
+ - TO
191
+ - 'ON'
192
+ - RU
193
+ - ▁SING
194
+ - ▁PE
195
+ - IPUN
196
+ - BU
197
+ - ▁SI
198
+ - ▁LA
199
+ - EN
200
+ - ▁KERJA
201
+ - ▁E
202
+ - LO
203
+ - GI
204
+ - IN
205
+ - KONTRAK
206
+ - ▁B
207
+ - ▁O
208
+ - CA
209
+ - ▁UGA
210
+ - ER
211
+ - ▁SAKA
212
+ - ▁DADI
213
+ - ▁T
214
+ - TAN
215
+ - È
216
+ - C
217
+ - ▁YA
218
+ - ▁U
219
+ - WI
220
+ - ▁NA
221
+ - ▁C
222
+ - ▁KU
223
+ - LU
224
+ - ▁K
225
+ - ▁RE
226
+ - ▁JA
227
+ - ▁PER
228
+ - ▁WA
229
+ - ▁MI
230
+ - RAN
231
+ - MAN
232
+ - AL
233
+ - US
234
+ - EL
235
+ - ▁UNIVERSITY
236
+ - ▁NANG
237
+ - HA
238
+ - ▁DIPUN
239
+ - ▁LE
240
+ - ▁SEKOLAH
241
+ - ▁DA
242
+ - ▁INGKANG
243
+ - ▁SENENG
244
+ - ▁KO
245
+ - ▁N
246
+ - UN
247
+ - 'NO'
248
+ - ▁MUSIK
249
+ - BI
250
+ - LAN
251
+ - ▁KUW
252
+ - ▁DE
253
+ - AR
254
+ - KI
255
+ - HAN
256
+ - ▁GA
257
+ - ▁TE
258
+ - ▁GE
259
+ - ▁BU
260
+ - ▁CA
261
+ - ▁LAGI
262
+ - ▁M
263
+ - AKÉ
264
+ - Z
265
+ - KE
266
+ - ▁MU
267
+ - ▁AS
268
+ - PI
269
+ - TH
270
+ - UR
271
+ - ▁IN
272
+ - ▁BE
273
+ - TER
274
+ - ▁TI
275
+ - ▁ALUMN
276
+ - ▁WONTEN
277
+ - DE
278
+ - SE
279
+ - CE
280
+ - DHA
281
+ - GU
282
+ - ▁AL
283
+ - ▁PI
284
+ - ▁NGRUNGOKN
285
+ - ▁RO
286
+ - MU
287
+ - ▁ORA
288
+ - PU
289
+ - ▁MO
290
+ - ▁KANGGO
291
+ - ES
292
+ - ▁NG
293
+ - ▁IS
294
+ - DO
295
+ - DU
296
+ - IS
297
+ - WE
298
+ - ▁LI
299
+ - ▁WI
300
+ - AKEN
301
+ - ▁BI
302
+ - ▁JU
303
+ - CH
304
+ - OR
305
+ - WAN
306
+ - ▁BISA
307
+ - ▁TU
308
+ - J
309
+ - NGGA
310
+ - ▁SO
311
+ - KO
312
+ - ▁G
313
+ - NDA
314
+ - ▁PAN
315
+ - UNG
316
+ - JU
317
+ - AT
318
+ - ▁MINANGKA
319
+ - W
320
+ - ▁GU
321
+ - BO
322
+ - VI
323
+ - ▁AN
324
+ - SAN
325
+ - ÈN
326
+ - ▁BO
327
+ - JI
328
+ - NIPUN
329
+ - ▁TAUN
330
+ - ▁KANTHI
331
+ - ▁TO
332
+ - ▁RI
333
+ - ▁P
334
+ - ▁SAKING
335
+ - ▁NGA
336
+ - ▁JE
337
+ - YU
338
+ - CK
339
+ - ANG
340
+ - ▁DÉNING
341
+ - ▁AR
342
+ - ▁UNIVERSITAS
343
+ - RAH
344
+ - ▁NI
345
+ - ▁MAR
346
+ - YAN
347
+ - ▁KUTHA
348
+ - ▁BANJUR
349
+ - ▁WONG
350
+ - ▁PEN
351
+ - ME
352
+ - UM
353
+ - ▁F
354
+ - ▁HA
355
+ - ST
356
+ - EM
357
+ - ▁D
358
+ - ▁UTAWA
359
+ - ▁DU
360
+ - DHU
361
+ - ▁NALIKA
362
+ - ▁THE
363
+ - GO
364
+ - AK
365
+ - UL
366
+ - LAS
367
+ - KAN
368
+ - BE
369
+ - ▁KI
370
+ - RANG
371
+ - ▁WE
372
+ - MO
373
+ - ING
374
+ - ▁PU
375
+ - ▁SIJI
376
+ - IT
377
+ - ▁PULUH
378
+ - ▁BASA
379
+ - GAN
380
+ - CI
381
+ - UT
382
+ - ▁OF
383
+ - ▁WIS
384
+ - YO
385
+ - TON
386
+ - PO
387
+ - ▁LU
388
+ - ▁NO
389
+ - ▁RU
390
+ - ▁CO
391
+ - LAH
392
+ - DHI
393
+ - ▁PRA
394
+ - ▁JO
395
+ - AH
396
+ - ▁KAR
397
+ - HO
398
+ - ÈK
399
+ - ▁PO
400
+ - ▁HI
401
+ - ET
402
+ - ▁DEWI
403
+ - ▁FA
404
+ - ANGAN
405
+ - ▁DADOS
406
+ - JO
407
+ - ▁PRO
408
+ - MBANG
409
+ - VE
410
+ - OK
411
+ - TRA
412
+ - ▁YU
413
+ - ▁MAN
414
+ - IL
415
+ - ▁PANG
416
+ - LAM
417
+ - IR
418
+ - LIS
419
+ - ▁JENENG
420
+ - VA
421
+ - ▁HO
422
+ - DRA
423
+ - WU
424
+ - TRI
425
+ - ▁INGGIH
426
+ - ▁SALAH
427
+ - ▁CE
428
+ - CO
429
+ - ▁MEN
430
+ - SON
431
+ - BAR
432
+ - ▁LO
433
+ - ▁UGI
434
+ - ▁GO
435
+ - EK
436
+ - ▁JAM
437
+ - ASI
438
+ - ▁CI
439
+ - ▁NGE
440
+ - RY
441
+ - IA
442
+ - ▁AD
443
+ - ▁NING
444
+ - ▁DHEWEKE
445
+ - NTEN
446
+ - ▁SAWIJINING
447
+ - ARA
448
+ - RIS
449
+ - TIK
450
+ - ▁PARA
451
+ - TAR
452
+ - ▁NYE
453
+ - ▁HE
454
+ - ▁V
455
+ - X
456
+ - EP
457
+ - PER
458
+ - AKE
459
+ - ▁INDONESIA
460
+ - MEN
461
+ - ▁SAN
462
+ - ▁DO
463
+ - WAR
464
+ - ▁MB
465
+ - ▁INDONÉSIA
466
+ - KAKÉ
467
+ - BER
468
+ - ▁KALIYAN
469
+ - ▁RONG
470
+ - RIA
471
+ - AM
472
+ - ▁JAWA
473
+ - ▁KAP
474
+ - ▁KAN
475
+ - ▁NU
476
+ - ▁NGANTI
477
+ - ▁ART
478
+ - ▁VI
479
+ - KAKE
480
+ - ▁FILM
481
+ - LANG
482
+ - ▁DHA
483
+ - ▁SRI
484
+ - DHO
485
+ - RENG
486
+ - ▁KAYA
487
+ - THA
488
+ - NGING
489
+ - ▁AMARG
490
+ - TUR
491
+ - DHANG
492
+ - CU
493
+ - UK
494
+ - ▁TELU
495
+ - RON
496
+ - ▁AYU
497
+ - ▁ST
498
+ - ▁NGANGGO
499
+ - ▁LUWIH
500
+ - BAT
501
+ - RING
502
+ - ▁SAMPUN
503
+ - ▁WIWIT
504
+ - TEN
505
+ - ▁GI
506
+ - RAT
507
+ - ▁KON
508
+ - LIN
509
+ - ▁AND
510
+ - NAN
511
+ - ▁NANGING
512
+ - ▁DHÈWÈKÉ
513
+ - ▁NEW
514
+ - ▁GUNUNG
515
+ - ▁Z
516
+ - SO
517
+ - ▁LORO
518
+ - ▁W
519
+ - ▁AKÈH
520
+ - ▁FO
521
+ - PAN
522
+ - ▁NDUWÈ
523
+ - KAR
524
+ - ▁HU
525
+ - ▁EWU
526
+ - ▁ATUS
527
+ - ONG
528
+ - ▁DIARANI
529
+ - ▁BEN
530
+ - ▁PALING
531
+ - ▁BAN
532
+ - IK
533
+ - NING
534
+ - ▁WERNA
535
+ - V
536
+ - OS
537
+ - GER
538
+ - ▁WOLU
539
+ - TIN
540
+ - ▁MAU
541
+ - AS
542
+ - ▁SAM
543
+ - ▁FI
544
+ - ▁KIDUL
545
+ - ▁KALI
546
+ - ▁PIYAMBAKIPUN
547
+ - ÈL
548
+ - DY
549
+ - OL
550
+ - ▁ENEM
551
+ - WÉ
552
+ - FA
553
+ - TOR
554
+ - ▁FE
555
+ - LIA
556
+ - ▁SAGED
557
+ - RUS
558
+ - TAS
559
+ - ZA
560
+ - FORD
561
+ - BAH
562
+ - MPI
563
+ - ▁ASAL
564
+ - ▁TEMBUNG
565
+ - LIK
566
+ - ▁SAIKI
567
+ - ▁PIN
568
+ - ▁JI
569
+ - ▁SHA
570
+ - BUR
571
+ - ▁KAS
572
+ - ▁SARI
573
+ - ▁WAR
574
+ - GEN
575
+ - ▁AGUS
576
+ - WO
577
+ - ▁NUR
578
+ - ▁LONDON
579
+ - ▁URIP
580
+ - MAS
581
+ - ▁PITU
582
+ - UKU
583
+ - ÈR
584
+ - WON
585
+ - DAH
586
+ - ANÉ
587
+ - ▁ND
588
+ - ▁PAM
589
+ - CARA
590
+ - ▁MARANG
591
+ - VER
592
+ - ▁HAM
593
+ - ▁LAIR
594
+ - ▁WU
595
+ - ▁ANAK
596
+ - ▁ANGG
597
+ - ▁SITI
598
+ - ▁DWI
599
+ - ARD
600
+ - DIA
601
+ - ▁KANGG
602
+ - ▁ARIF
603
+ - MER
604
+ - ▁ALI
605
+ - ▁SABEN
606
+ - MAR
607
+ - ▁CITA
608
+ - ▁BOTEN
609
+ - QI
610
+ - ▁HER
611
+ - ▁CILIK
612
+ - AD
613
+ - ▁LAJENG
614
+ - ▁PRI
615
+ - ▁INSTITUTE
616
+ - ▁IBU
617
+ - ▁LIMA
618
+ - ▁HADI
619
+ - ▁PAS
620
+ - ▁BANYU
621
+ - ▁BER
622
+ - Q
623
+ - ▁KASEBUT
624
+ - ▁ABDUL
625
+ - ▁BAMBANG
626
+ - ▁TELUNG
627
+ - WIS
628
+ - ▁KALIH
629
+ - ▁DUWÉ
630
+ - STER
631
+ - ▁ZA
632
+ - ▁DIGAWÉ
633
+ - ▁ARI
634
+ - NTUK
635
+ - ▁AWAK
636
+ - ▁AREP
637
+ - ▁NDUWÉ
638
+ - ▁NGG
639
+ - ▁ABDULLAH
640
+ - ▁TER
641
+ - ▁WULANDARI
642
+ - ▁ISIH
643
+ - ▁YO
644
+ - ▁NGU
645
+ - ▁RIF
646
+ - ▁MUNG
647
+ - RUL
648
+ - ▁PUTRA
649
+ - KON
650
+ - ▁MANGGUNG
651
+ - ▁NUG
652
+ - ▁JINIS
653
+ - SEN
654
+ - ▁AZ
655
+ - ▁JAKARTA
656
+ - MIN
657
+ - NIA
658
+ - ▁SOLO
659
+ - ▁EKO
660
+ - JAH
661
+ - TIF
662
+ - ▁DAN
663
+ - ▁SETU
664
+ - ▁BAL
665
+ - ▁AKEH
666
+ - NDER
667
+ - ▁TAU
668
+ - ▁PUJI
669
+ - ▁YOGYAKARTA
670
+ - ▁TIYANG
671
+ - ▁SARASVATI
672
+ - ▁PUN
673
+ - ▁BALI
674
+ - LAR
675
+ - NINGSIH
676
+ - ▁OR
677
+ - ENT
678
+ - ▁TENGAH
679
+ - ▁PAPAN
680
+ - ▁KIRA
681
+ - ▁JAKA
682
+ - ▁PADHA
683
+ - ▁OXFORD
684
+ - ▁MULA
685
+ - ▁EL
686
+ - ▁TUM
687
+ - ▁NGLA
688
+ - ▁SEN
689
+ - TUTI
690
+ - ROHO
691
+ - ▁SOFYAN
692
+ - ▁KABUPATÈN
693
+ - ▁WAHHAB
694
+ - ▁DHUWUR
695
+ - ▁ARUPA
696
+ - ▁MAIN
697
+ - ▁PRE
698
+ - ▁WIT
699
+ - ▁DONYA
700
+ - ▁TIM
701
+ - ▁NYA
702
+ - ▁KALEBU
703
+ - ▁PANGANAN
704
+ - ▁HAR
705
+ - FF
706
+ - ▁PRASETY
707
+ - ▁GEDHÉ
708
+ - ▁SAWISÉ
709
+ - ▁PUTRI
710
+ - TARA
711
+ - BINTANG
712
+ - LY
713
+ - ▁TEGES
714
+ - ▁BANGET
715
+ - ▁PANJENENGANIPUN
716
+ - LLY
717
+ - ▁WOH
718
+ - ▁PATANG
719
+ - ▁MISUWUR
720
+ - OLOGI
721
+ - ▁KATHAH
722
+ - DEN
723
+ - ▁CARA
724
+ - ▁OMAH
725
+ - ▁SHE
726
+ - OMB
727
+ - ▁NJ
728
+ - ▁CAMBRIDGE
729
+ - ▁BABAGAN
730
+ - MON
731
+ - ▁PAPAT
732
+ - ▁NAGARA
733
+ - ▁TEMP
734
+ - ▁SAHA
735
+ - HAM
736
+ - ▁MANGAN
737
+ - ▁SANGA
738
+ - ▁AMB
739
+ - BAN
740
+ - AGE
741
+ - ▁GODHONG
742
+ - DER
743
+ - ▁KAPING
744
+ - ▁BAKAL
745
+ - ▁LIYA
746
+ - ▁EDINBURGH
747
+ - ▁DHÉWÉ
748
+ - ÈNG
749
+ - ▁PERANG
750
+ - ▁YÈN
751
+ - ▁DUMUNUNG
752
+ - PORT
753
+ - ▁MER
754
+ - ▁PASAR
755
+ - ▁POP
756
+ - ▁PROVINSI
757
+ - ▁UMUM
758
+ - ▁US
759
+ - ▁SANGALAS
760
+ - ▁UTAMA
761
+ - ▁DIKENAL
762
+ - ▁CACAH
763
+ - LUK
764
+ - ▁WEWENGKON
765
+ - ▁GAD
766
+ - ▁INDIA
767
+ - ▁DAWA
768
+ - VO
769
+ - ▁YORK
770
+ - ▁BAB
771
+ - ▁BAHAN
772
+ - ▁JENENGÉ
773
+ - ▁KEMBANG
774
+ - ▁WETON
775
+ - ▁MENAW
776
+ - ▁MÈ
777
+ - ▁PULO
778
+ - '?'
779
+ - ▁DON
780
+ - ▁TANDURAN
781
+ - ▁KRAJAN
782
+ - ▁PITUNG
783
+ - ▁SINETRON
784
+ - ▁GRA
785
+ - ▁RAJA
786
+ - ▁JAMAN
787
+ - ▁TEN
788
+ - ▁KRA
789
+ - ▁BENTUK
790
+ - ▁SANGANG
791
+ - ▁UTAWI
792
+ - ▁WUJUD
793
+ - ▁SUKU
794
+ - ▁JOHN
795
+ - ▁KOMP
796
+ - ▁LANGKUNG
797
+ - ▁ASRING
798
+ - ▁BAPAK
799
+ - ▁HARV
800
+ - ▁INGGRIS
801
+ - ▁SISIH
802
+ - ▁DHEWEKÉ
803
+ - ▁YEN
804
+ - ▁MITURUT
805
+ - ▁TANPA
806
+ - ▁KRI
807
+ - DUR
808
+ - ▁LAGU
809
+ - ▁PUNGKASAN
810
+ - ▁LOR
811
+ - ▁ANGGOTA
812
+ - ▁PUTIH
813
+ - ▁SALIYANÉ
814
+ - ▁AZHAR
815
+ - ▁DISEBUT
816
+ - ▁KUDU
817
+ - ▁SETUNGGAL
818
+ - ▁AKSARA
819
+ - ▁BRAD
820
+ - ▁ANTARA
821
+ - NWAR
822
+ - ▁GAWÉ
823
+ - ▁MANUNGSA
824
+ - ▁GADHAH
825
+ - ▁COLLEGE
826
+ - ▁CHRIS
827
+ - ▁STAM
828
+ - ▁DAGING
829
+ - ▁KEREP
830
+ - ▁KABÈH
831
+ - UPAYA
832
+ - ▁BAGÉYAN
833
+ - ▁CHUNG
834
+ - ▁IWAK
835
+ - ▁KADOS
836
+ - ▁JEPANG
837
+ - ▁KHAS
838
+ - ▁APIK
839
+ - ▁YOU
840
+ - ▁WAYANG
841
+ - ▁WEKTU
842
+ - ▁BIYASANÉ
843
+ - ▁LIYANÉ
844
+ - ▁KULIT
845
+ - MPUNG
846
+ - ▁KULON
847
+ - ▁HIJRAH
848
+ - ▁DAVID
849
+ - ▁WÉTAN
850
+ - ▁SEKET
851
+ - RISTOL
852
+ - ▁DHAÉRAH
853
+ - ▁TEKAN
854
+ - ▁BANDUNG
855
+ - ▁KÉWAN
856
+ - ▁MANGSA
857
+ - ▁BADH
858
+ - ▁AGENG
859
+ - ▁PANGGONAN
860
+ - ▁SEMARANG
861
+ - ▁QU
862
+ - ▁ABANG
863
+ - ▁BIASANÉ
864
+ - ▁WOR
865
+ - ▁DÉNÉ
866
+ - ▁GAMELAN
867
+ - ▁YAIKU
868
+ - ▁CHE
869
+ - ▁NASIONAL
870
+ - KANGKU
871
+ - ▁POL
872
+ - ▁DÉSA
873
+ - ▁TULADHA
874
+ - ▁SATUNGGALING
875
+ - ▁SHI
876
+ - ▁UKURAN
877
+ - ▁NGANTOS
878
+ - ▁ALBUM
879
+ - BANGUN
880
+ - ANGGEP
881
+ - ▁TLATAH
882
+ - ▁SURABAYA
883
+ - ▁WADON
884
+ - ▁UMUR
885
+ - ▁MANÈH
886
+ - ▁SATUS
887
+ - ▁RÉ
888
+ - ▁MUHAMMAD
889
+ - CAMPUR
890
+ - JENG
891
+ - ▁AHMAD
892
+ - ▁BOCAH
893
+ - ▁BIASA
894
+ - ▁NGISOR
895
+ - ▁SEDAYA
896
+ - ▁MANCHESTER
897
+ - ▁JESSI
898
+ - ▁MANUK
899
+ - ▁SISTEM
900
+ - ▁SONGOLAS
901
+ - ▁DAMEL
902
+ - ▁WAGENINGEN
903
+ - ▁PIRANTI
904
+ - ▁PIYAMBAK
905
+ - ▁AKTING
906
+ - DHÉ
907
+ - ▁WALANDA
908
+ - ▁JERMAN
909
+ - ▁KECAMATAN
910
+ - ISME
911
+ - ▁ANYAR
912
+ - ▁AGAMA
913
+ - ▁SEPULUH
914
+ - ▁PAPUA
915
+ - ▁CEDHAK
916
+ - ▁CHI
917
+ - ▁KAGUNGAN
918
+ - ▁BÉDA
919
+ - ▁BOGOR
920
+ - ▁SAWETARA
921
+ - ▁WOLULAS
922
+ - ▁NGGADHAH
923
+ - ▁PEMAIN
924
+ - ▁SANGET
925
+ - ▁KALEBET
926
+ - ▁SWIDAK
927
+ - ▁WILLIAM
928
+ - ▁SASTRA
929
+ - ▁ACEH
930
+ - ▁GARWA
931
+ - ▁MALIH
932
+ - ▁THAILAND
933
+ - ▁MBOTEN
934
+ - ▁WANGUN
935
+ - ▁OPERA
936
+ - ▁PUSAT
937
+ - ▁AMERIKA
938
+ - ▁FILIPINA
939
+ - ▁MIWIT
940
+ - ▁LOMBOK
941
+ - ▁PADANG
942
+ - ▁PENYANYI
943
+ - ▁ÉROPAH
944
+ - ▁PEMB
945
+ - ▁LEMAH
946
+ - ▁MANGGON
947
+ - ▁END
948
+ - ▁TANGGAL
949
+ - ▁KLATEN
950
+ - ▁MLEBU
951
+ - ▁CANDHI
952
+ - ▁PEKALONGAN
953
+ - ▁BENGKULU
954
+ - ▁MAGELANG
955
+ - ▁TOKYO
956
+ - ▁CIREBON
957
+ - ▁PISANAN
958
+ - IGHT
959
+ - ▁PANJENENGANÉ
960
+ - ▁SADURUNGÉ
961
+ - ▁TETEP
962
+ - ▁NIKAH
963
+ - ▁PRANCIS
964
+ - ▁TEGAL
965
+ - ▁KANCANE
966
+ - ▁KITAB
967
+ - ▁SAMPE
968
+ - ▁KACAMATAN
969
+ - ▁SAPUNIKA
970
+ - ▁MALUKU
971
+ - ▁DURUNG
972
+ - ▁PALEMBANG
973
+ - ▁KABEH
974
+ - ▁SWARA
975
+ - ▁MELBOURNE
976
+ - ▁SIDOARJO
977
+ - ▁DIANGGO
978
+ - ▁GEDHE
979
+ - ▁AKTOR
980
+ - ▁TANGERANG
981
+ - ▁SEDULUR
982
+ - ▁MEDAN
983
+ - ▁CILACAP
984
+ - PÉRANGAN
985
+ - ▁SEOUL
986
+ - ▁SEGARA
987
+ - ▁MUMBAI
988
+ - ▁DERRY
989
+ - ▁SAMARINDA
990
+ - ▁SURYA
991
+ - ▁KULAWARGA
992
+ - ▁MASAKAN
993
+ - ▁UZBEKISTAN
994
+ - ▁NGANDHUT
995
+ - ▁BEKASI
996
+ - ▁SINAU
997
+ - ▁BALIKPAPAN
998
+ - ▁DITEMOKAKÉ
999
+ - ▁SLEMAN
1000
+ - ▁FINLANDIA
1001
+ - ▁THAT
1002
+ - ▁AWUJUD
1003
+ - ▁PONTIANAK
1004
+ - ▁TAIPEI
1005
+ - ▁CHARL
1006
+ - ▁MANADO
1007
+ - ▁UNTU
1008
+ - ▁ALJAZAIR
1009
+ - ▁ZIMBABWE
1010
+ - ▁ILMU
1011
+ - ▁LATVIA
1012
+ - ▁BAGÉAN
1013
+ - ▁GANGSAL
1014
+ - ▁AMÉRIKA
1015
+ - ▁LEIPZIG
1016
+ - ▁NJLENTREHAK
1017
+ - ▁BREBES
1018
+ - ▁TINIMBANG
1019
+ - ▁SEKAWAN
1020
+ - ▁TOKOH
1021
+ - ▁BANGUNAN
1022
+ - ▁WILL
1023
+ - ▁KAZAKHSTAN
1024
+ - ▁PENYAKIT
1025
+ - ▁BELGOROD
1026
+ - ▁ELIZABETH
1027
+ - ▁CAPCAY
1028
+ - ▁JAZZ
1029
+ - ▁LANCASTER
1030
+ - ▁NEWCASTLE
1031
+ - ▁GRUP
1032
+ - ▁CARDIFF
1033
+ - ▁DUNDEE
1034
+ - ▁MICRO
1035
+ - ▁ITALIA
1036
+ - ▁SETENGAH
1037
+ - ▁WELLS
1038
+ - ▁PEMALANG
1039
+ - ▁FUNGSI
1040
+ - ▁LIKUR
1041
+ - ANGZHOU
1042
+ - ▁ROBERT
1043
+ - ▁RICHARD
1044
+ - ▁HAKIM
1045
+ - ▁ABERDEEN
1046
+ - ▁QUEENSLAND
1047
+ - ▁SOMETHING
1048
+ - ▁ICELAND
1049
+ - ▁AKTRIS
1050
+ - ▁MAKASAR
1051
+ - ▁BUMBU
1052
+ - SOFT
1053
+ - ▁PULITIK
1054
+ - ▁WINCHESTE
1055
+ - ▁GAMPANG
1056
+ - ▁COKLAT
1057
+ - ▁DELHI
1058
+ - ▁BAGHDAD
1059
+ - ▁BELFAST
1060
+ - ▁CIKARANG
1061
+ - ▁NYEBABK
1062
+ - ▁MASYARAKAT
1063
+ - ▁DIGUNAKAKÉ
1064
+ - ▁DANGDUT
1065
+ - ▁DIENGGO
1066
+ - ▁RAMBUT
1067
+ - ▁LUXEMBURG
1068
+ - ▁DHEWE
1069
+ - ▁STIRLING
1070
+ - ▁PERUSAHAAN
1071
+ - ▁CAMPURSARI
1072
+ - ▁BOYOLALI
1073
+ - ▁UPACARA
1074
+ - ▁WIJAYA
1075
+ - ▁KAGOLONG
1076
+ - ▁GAMBAR
1077
+ - ▁SUMBER
1078
+ - /
1079
+ - ''''
1080
+ - Å
1081
+ - '5'
1082
+ - '!'
1083
+ - '2'
1084
+ - '8'
1085
+ - '4'
1086
+ - '3'
1087
+ - .
1088
+ - '1'
1089
+ - '6'
1090
+ - '9'
1091
+ - –
1092
+ - '~'
1093
+ - '0'
1094
+ - Â
1095
+ - '7'
1096
+ - ¥
1097
+ - —
1098
+ - “
1099
+ - ”
1100
+ - Ê
1101
+ - Ð
1102
+ - <sos/eos>
1103
+ init: chainer
1104
+ input_size: null
1105
+ ctc_conf:
1106
+ dropout_rate: 0.0
1107
+ ctc_type: builtin
1108
+ reduce: true
1109
+ ignore_nan_grad: false
1110
+ model_conf:
1111
+ ctc_weight: 0.3
1112
+ lsm_weight: 0.1
1113
+ length_normalized_loss: false
1114
+ use_preprocessor: true
1115
+ token_type: bpe
1116
+ bpemodel: data/token_list/bpe_unigram1000/bpe.model
1117
+ non_linguistic_symbols: null
1118
+ cleaner: null
1119
+ g2p: null
1120
+ speech_volume_normalize: null
1121
+ rir_scp: null
1122
+ rir_apply_prob: 1.0
1123
+ noise_scp: null
1124
+ noise_apply_prob: 1.0
1125
+ noise_db_range: '13_15'
1126
+ frontend: default
1127
+ frontend_conf:
1128
+ fs: 16k
1129
+ specaug: null
1130
+ specaug_conf: {}
1131
+ normalize: global_mvn
1132
+ normalize_conf:
1133
+ stats_file: exp/asr_stats_raw_bpe1000/train/feats_stats.npz
1134
+ preencoder: null
1135
+ preencoder_conf: {}
1136
+ encoder: transformer
1137
+ encoder_conf:
1138
+ input_layer: conv2d
1139
+ num_blocks: 12
1140
+ linear_units: 2048
1141
+ dropout_rate: 0.1
1142
+ output_size: 256
1143
+ attention_heads: 4
1144
+ attention_dropout_rate: 0.0
1145
+ decoder: transformer
1146
+ decoder_conf:
1147
+ input_layer: embed
1148
+ num_blocks: 6
1149
+ linear_units: 2048
1150
+ dropout_rate: 0.1
1151
+ required:
1152
+ - output_dir
1153
+ - token_list
1154
+ version: 0.9.7
1155
+ distributed: false
exp/asr_train_asr_raw_bpe1000/images/acc.png ADDED
exp/asr_train_asr_raw_bpe1000/images/backward_time.png ADDED
exp/asr_train_asr_raw_bpe1000/images/cer.png ADDED
exp/asr_train_asr_raw_bpe1000/images/cer_ctc.png ADDED
exp/asr_train_asr_raw_bpe1000/images/forward_time.png ADDED
exp/asr_train_asr_raw_bpe1000/images/iter_time.png ADDED
exp/asr_train_asr_raw_bpe1000/images/loss.png ADDED
exp/asr_train_asr_raw_bpe1000/images/loss_att.png ADDED
exp/asr_train_asr_raw_bpe1000/images/loss_ctc.png ADDED
exp/asr_train_asr_raw_bpe1000/images/lr_0.png ADDED
exp/asr_train_asr_raw_bpe1000/images/optim_step_time.png ADDED
exp/asr_train_asr_raw_bpe1000/images/train_time.png ADDED
exp/asr_train_asr_raw_bpe1000/images/wer.png ADDED
exp/asr_train_asr_raw_bpe1000/valid.acc.best.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b739370e1cde59b50dc2a42e0de2112c62a8f97ba5334f8c3939df0d839f166
3
+ size 111676310
meta.yaml ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ espnet: 0.10.0
2
+ files:
3
+ asr_model_file: exp/asr_train_asr_raw_bpe1000/valid.acc.best.pth
4
+ python: "3.8.5 (default, Sep 4 2020, 07:30:14) \n[GCC 7.3.0]"
5
+ timestamp: 1625878616.133883
6
+ torch: 1.8.1+cu102
7
+ yaml_files:
8
+ asr_train_config: exp/asr_train_asr_raw_bpe1000/config.yaml