Hugging Face's logo

desh2608
/
icefall-asr-alimeeting-pruned-transducer-stateless7

K2
TensorBoard
Chinese
icefall
Model card Files Metrics Community
2
icefall-asr-alimeeting-pruned-transducer-stateless7
File size: 35,669 Bytes 
26cafdd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
 
2022-12-09 00:24:42,587 INFO [decode.py:551] Decoding started
2022-12-09 00:24:42,588 INFO [decode.py:557] Device: cuda:0
2022-12-09 00:24:42,654 INFO [lexicon.py:168] Loading pre-compiled data/lang_char/Linv.pt
2022-12-09 00:24:42,664 INFO [decode.py:563] {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 100, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'warm_step': 2000, 'env_info': {'k2-version': '1.23', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': 'b2ce63f3940018e7b433c43fd802fc50ab006a76', 'k2-git-date': 'Wed Nov 23 08:43:43 2022', 'lhotse-version': '1.9.0.dev+git.97bf4b0.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'ali_meeting', 'icefall-git-sha1': 'f13cf61-dirty', 'icefall-git-date': 'Tue Dec 6 03:34:27 2022', 'icefall-path': '/exp/draj/mini_scale_2022/icefall', 'k2-path': '/exp/draj/mini_scale_2022/k2/k2/python/k2/__init__.py', 'lhotse-path': '/exp/draj/mini_scale_2022/lhotse/lhotse/__init__.py', 'hostname': 'r3n05', 'IP address': '10.1.3.5'}, 'epoch': 15, 'iter': 0, 'avg': 8, 'use_averaged_model': True, 'exp_dir': PosixPath('pruned_transducer_stateless7/exp/v1'), 'lang_dir': 'data/lang_char', 'decoding_method': 'modified_beam_search', 'beam_size': 4, 'beam': 4, 'ngram_lm_scale': 0.01, 'max_contexts': 4, 'max_states': 8, 'context_size': 2, 'max_sym_per_frame': 1, 'num_paths': 200, 'nbest_scale': 0.5, 'num_encoder_layers': '2,4,3,2,4', 'feedforward_dims': '1024,1024,2048,2048,1024', 'nhead': '8,8,8,8,8', 'encoder_dims': '384,384,384,384,384', 'attention_dims': '192,192,192,192,192', 'encoder_unmasked_dims': '256,256,256,256,256', 'zipformer_downsampling_factors': '1,2,4,8,2', 'cnn_module_kernels': '31,31,31,31,31', 'decoder_dim': 512, 'joiner_dim': 512, 'manifest_dir': PosixPath('data/manifests'), 'enable_musan': True, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'max_duration': 500, 'max_cuts': None, 'num_buckets': 50, 'on_the_fly_feats': False, 'shuffle': True, 'num_workers': 8, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'res_dir': PosixPath('pruned_transducer_stateless7/exp/v1/modified_beam_search'), 'suffix': 'epoch-15-avg-8-modified_beam_search-beam-size-4', 'blank_id': 0, 'vocab_size': 3290}
2022-12-09 00:24:42,664 INFO [decode.py:565] About to create model
2022-12-09 00:24:43,124 INFO [zipformer.py:179] At encoder stack 4, which has downsampling_factor=2, we will combine the outputs of layers 1 and 3, with downsampling_factors=2 and 8.
2022-12-09 00:24:43,173 INFO [decode.py:632] Calculating the averaged model over epoch range from 7 (excluded) to 15
2022-12-09 00:25:10,741 INFO [decode.py:655] Number of model parameters: 75734561
2022-12-09 00:25:10,741 INFO [asr_datamodule.py:381] About to get AliMeeting IHM eval cuts
2022-12-09 00:25:10,760 INFO [asr_datamodule.py:402] About to get AliMeeting IHM test cuts
2022-12-09 00:25:10,773 INFO [asr_datamodule.py:387] About to get AliMeeting SDM eval cuts
2022-12-09 00:25:10,784 INFO [asr_datamodule.py:408] About to get AliMeeting SDM test cuts
2022-12-09 00:25:10,803 INFO [asr_datamodule.py:396] About to get AliMeeting GSS-enhanced eval cuts
2022-12-09 00:25:10,826 INFO [asr_datamodule.py:417] About to get AliMeeting GSS-enhanced test cuts
2022-12-09 00:25:12,696 INFO [decode.py:687] Decoding eval_ihm
2022-12-09 00:25:18,820 INFO [decode.py:463] batch 0/?, cuts processed until now is 58
2022-12-09 00:25:24,241 INFO [decode.py:463] batch 2/?, cuts processed until now is 512
2022-12-09 00:25:31,979 INFO [decode.py:463] batch 4/?, cuts processed until now is 645
2022-12-09 00:25:40,148 INFO [decode.py:463] batch 6/?, cuts processed until now is 750
2022-12-09 00:25:47,934 INFO [decode.py:463] batch 8/?, cuts processed until now is 883
2022-12-09 00:25:55,121 INFO [decode.py:463] batch 10/?, cuts processed until now is 1082
2022-12-09 00:26:02,868 INFO [decode.py:463] batch 12/?, cuts processed until now is 1279
2022-12-09 00:26:02,981 INFO [zipformer.py:1414] attn_weights_entropy = tensor([4.6414, 4.6510, 4.5449, 4.8082, 4.0814, 4.1877, 4.8428, 4.4703],
       device='cuda:0'), covar=tensor([0.0608, 0.0354, 0.0841, 0.0608, 0.1302, 0.0402, 0.0377, 0.0977],
       device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0118, 0.0125, 0.0134, 0.0129, 0.0102, 0.0145, 0.0125],
       device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002],
       device='cuda:0')
2022-12-09 00:26:09,417 INFO [decode.py:463] batch 14/?, cuts processed until now is 1538
2022-12-09 00:26:16,019 INFO [decode.py:463] batch 16/?, cuts processed until now is 1845
2022-12-09 00:26:22,962 INFO [decode.py:463] batch 18/?, cuts processed until now is 2084
2022-12-09 00:26:28,777 INFO [decode.py:463] batch 20/?, cuts processed until now is 2523
2022-12-09 00:26:34,867 INFO [decode.py:463] batch 22/?, cuts processed until now is 2949
2022-12-09 00:26:42,001 INFO [decode.py:463] batch 24/?, cuts processed until now is 3160
2022-12-09 00:26:48,378 INFO [decode.py:463] batch 26/?, cuts processed until now is 3586
2022-12-09 00:26:55,879 INFO [decode.py:463] batch 28/?, cuts processed until now is 3758
2022-12-09 00:27:02,020 INFO [decode.py:463] batch 30/?, cuts processed until now is 4116
2022-12-09 00:27:06,528 INFO [decode.py:463] batch 32/?, cuts processed until now is 4742
2022-12-09 00:27:11,534 INFO [decode.py:463] batch 34/?, cuts processed until now is 5368
2022-12-09 00:27:17,535 INFO [decode.py:463] batch 36/?, cuts processed until now is 5796
2022-12-09 00:27:18,924 INFO [decode.py:463] batch 38/?, cuts processed until now is 5908
2022-12-09 00:27:26,130 INFO [decode.py:463] batch 40/?, cuts processed until now is 6026
2022-12-09 00:27:31,228 INFO [decode.py:463] batch 42/?, cuts processed until now is 6171
2022-12-09 00:27:37,988 INFO [decode.py:463] batch 44/?, cuts processed until now is 6390
2022-12-09 00:27:42,164 INFO [decode.py:463] batch 46/?, cuts processed until now is 6456
2022-12-09 00:27:42,623 INFO [decode.py:479] The transcripts are stored in pruned_transducer_stateless7/exp/v1/modified_beam_search/recogs-eval_ihm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:27:42,719 INFO [utils.py:536] [eval_ihm-beam_size_4] %WER 9.58% [7771 / 81111, 935 ins, 1663 del, 5173 sub ]
2022-12-09 00:27:42,957 INFO [decode.py:492] Wrote detailed error stats to pruned_transducer_stateless7/exp/v1/modified_beam_search/errs-eval_ihm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:27:42,959 INFO [decode.py:508] 
For eval_ihm, WER of different settings are:
beam_size_4	9.58	best for eval_ihm

2022-12-09 00:27:42,959 INFO [decode.py:687] Decoding test_ihm
2022-12-09 00:27:48,952 INFO [decode.py:463] batch 0/?, cuts processed until now is 49
2022-12-09 00:27:54,771 INFO [decode.py:463] batch 2/?, cuts processed until now is 433
2022-12-09 00:28:03,112 INFO [decode.py:463] batch 4/?, cuts processed until now is 545
2022-12-09 00:28:11,577 INFO [decode.py:463] batch 6/?, cuts processed until now is 637
2022-12-09 00:28:20,309 INFO [decode.py:463] batch 8/?, cuts processed until now is 754
2022-12-09 00:28:29,065 INFO [decode.py:463] batch 10/?, cuts processed until now is 845
2022-12-09 00:28:36,939 INFO [decode.py:463] batch 12/?, cuts processed until now is 976
2022-12-09 00:28:44,236 INFO [decode.py:463] batch 14/?, cuts processed until now is 1175
2022-12-09 00:28:50,384 INFO [decode.py:463] batch 16/?, cuts processed until now is 1483
2022-12-09 00:28:58,746 INFO [decode.py:463] batch 18/?, cuts processed until now is 1590
2022-12-09 00:29:07,677 INFO [decode.py:463] batch 20/?, cuts processed until now is 1658
2022-12-09 00:29:07,960 INFO [zipformer.py:1414] attn_weights_entropy = tensor([3.1712, 2.8828, 2.9930, 2.1980, 2.5329, 3.0201, 3.0163, 2.7156],
       device='cuda:0'), covar=tensor([0.0848, 0.1539, 0.1073, 0.1706, 0.1543, 0.0510, 0.1103, 0.1445],
       device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0170, 0.0124, 0.0117, 0.0121, 0.0128, 0.0106, 0.0128],
       device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005],
       device='cuda:0')
2022-12-09 00:29:14,837 INFO [decode.py:463] batch 22/?, cuts processed until now is 1856
2022-12-09 00:29:20,542 INFO [decode.py:463] batch 24/?, cuts processed until now is 2224
2022-12-09 00:29:28,891 INFO [decode.py:463] batch 26/?, cuts processed until now is 2325
2022-12-09 00:29:36,342 INFO [decode.py:463] batch 28/?, cuts processed until now is 2546
2022-12-09 00:29:44,445 INFO [decode.py:463] batch 30/?, cuts processed until now is 2653
2022-12-09 00:29:53,058 INFO [decode.py:463] batch 32/?, cuts processed until now is 2744
2022-12-09 00:30:01,060 INFO [decode.py:463] batch 34/?, cuts processed until now is 2875
2022-12-09 00:30:05,796 INFO [zipformer.py:1414] attn_weights_entropy = tensor([3.4639, 3.9804, 3.8866, 4.0083, 2.9221, 3.9239, 3.5020, 2.1089],
       device='cuda:0'), covar=tensor([0.2868, 0.1054, 0.1046, 0.0689, 0.1069, 0.0660, 0.1511, 0.3060],
       device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0066, 0.0052, 0.0054, 0.0082, 0.0064, 0.0085, 0.0091],
       device='cuda:0'), out_proj_covar=tensor([0.0007, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005],
       device='cuda:0')
2022-12-09 00:30:10,496 INFO [decode.py:463] batch 36/?, cuts processed until now is 2961
2022-12-09 00:30:18,741 INFO [decode.py:463] batch 38/?, cuts processed until now is 3072
2022-12-09 00:30:25,005 INFO [decode.py:463] batch 40/?, cuts processed until now is 3440
2022-12-09 00:30:29,753 INFO [decode.py:463] batch 42/?, cuts processed until now is 3956
2022-12-09 00:30:35,471 INFO [decode.py:463] batch 44/?, cuts processed until now is 4342
2022-12-09 00:30:43,655 INFO [decode.py:463] batch 46/?, cuts processed until now is 4443
2022-12-09 00:30:51,033 INFO [decode.py:463] batch 48/?, cuts processed until now is 4595
2022-12-09 00:30:57,408 INFO [decode.py:463] batch 50/?, cuts processed until now is 4872
2022-12-09 00:31:04,365 INFO [decode.py:463] batch 52/?, cuts processed until now is 5061
2022-12-09 00:31:12,188 INFO [decode.py:463] batch 54/?, cuts processed until now is 5219
2022-12-09 00:31:15,526 INFO [decode.py:463] batch 56/?, cuts processed until now is 5892
2022-12-09 00:31:22,589 INFO [decode.py:463] batch 58/?, cuts processed until now is 6090
2022-12-09 00:31:27,553 INFO [decode.py:463] batch 60/?, cuts processed until now is 6517
2022-12-09 00:31:34,231 INFO [decode.py:463] batch 62/?, cuts processed until now is 6715
2022-12-09 00:31:41,261 INFO [decode.py:463] batch 64/?, cuts processed until now is 6897
2022-12-09 00:31:46,410 INFO [decode.py:463] batch 66/?, cuts processed until now is 7304
2022-12-09 00:31:53,372 INFO [decode.py:463] batch 68/?, cuts processed until now is 7488
2022-12-09 00:32:00,322 INFO [decode.py:463] batch 70/?, cuts processed until now is 7720
2022-12-09 00:32:06,936 INFO [decode.py:463] batch 72/?, cuts processed until now is 7938
2022-12-09 00:32:12,023 INFO [decode.py:463] batch 74/?, cuts processed until now is 8367
2022-12-09 00:32:17,835 INFO [decode.py:463] batch 76/?, cuts processed until now is 8674
2022-12-09 00:32:23,809 INFO [decode.py:463] batch 78/?, cuts processed until now is 8982
2022-12-09 00:32:29,406 INFO [decode.py:463] batch 80/?, cuts processed until now is 9350
2022-12-09 00:32:34,331 INFO [decode.py:463] batch 82/?, cuts processed until now is 9810
2022-12-09 00:32:39,302 INFO [decode.py:463] batch 84/?, cuts processed until now is 10237
2022-12-09 00:32:43,638 INFO [decode.py:463] batch 86/?, cuts processed until now is 10755
2022-12-09 00:32:48,155 INFO [decode.py:463] batch 88/?, cuts processed until now is 11278
2022-12-09 00:32:53,182 INFO [decode.py:463] batch 90/?, cuts processed until now is 11705
2022-12-09 00:32:59,271 INFO [decode.py:463] batch 92/?, cuts processed until now is 12013
2022-12-09 00:33:05,419 INFO [decode.py:463] batch 94/?, cuts processed until now is 12290
2022-12-09 00:33:11,469 INFO [decode.py:463] batch 96/?, cuts processed until now is 12597
2022-12-09 00:33:17,185 INFO [decode.py:463] batch 98/?, cuts processed until now is 12963
2022-12-09 00:33:21,943 INFO [decode.py:463] batch 100/?, cuts processed until now is 13420
2022-12-09 00:33:27,380 INFO [decode.py:463] batch 102/?, cuts processed until now is 13877
2022-12-09 00:33:32,013 INFO [decode.py:463] batch 104/?, cuts processed until now is 14543
2022-12-09 00:33:36,461 INFO [decode.py:463] batch 106/?, cuts processed until now is 15209
2022-12-09 00:33:38,950 INFO [zipformer.py:1414] attn_weights_entropy = tensor([2.1245, 1.9127, 2.2915, 1.8888, 2.0640, 1.1755, 2.0946, 2.0218],
       device='cuda:0'), covar=tensor([0.0894, 0.1076, 0.0611, 0.0984, 0.1650, 0.0887, 0.0727, 0.1079],
       device='cuda:0'), in_proj_covar=tensor([0.0022, 0.0023, 0.0024, 0.0022, 0.0023, 0.0034, 0.0023, 0.0024],
       device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0001],
       device='cuda:0')
2022-12-09 00:33:39,036 INFO [zipformer.py:1414] attn_weights_entropy = tensor([4.1049, 2.3319, 3.7990, 4.0203, 3.6045, 2.4305, 4.0939, 2.8954],
       device='cuda:0'), covar=tensor([0.0258, 0.0949, 0.0516, 0.0229, 0.0445, 0.1350, 0.0229, 0.0869],
       device='cuda:0'), in_proj_covar=tensor([0.0258, 0.0233, 0.0347, 0.0293, 0.0235, 0.0280, 0.0265, 0.0260],
       device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002],
       device='cuda:0')
2022-12-09 00:33:41,684 INFO [decode.py:463] batch 108/?, cuts processed until now is 15599
2022-12-09 00:33:45,094 INFO [decode.py:463] batch 110/?, cuts processed until now is 15787
2022-12-09 00:33:48,336 INFO [decode.py:463] batch 112/?, cuts processed until now is 15881
2022-12-09 00:33:53,053 INFO [decode.py:463] batch 114/?, cuts processed until now is 15926
2022-12-09 00:33:55,603 INFO [decode.py:463] batch 116/?, cuts processed until now is 16287
2022-12-09 00:34:01,766 INFO [decode.py:463] batch 118/?, cuts processed until now is 16357
2022-12-09 00:34:02,021 INFO [decode.py:479] The transcripts are stored in pruned_transducer_stateless7/exp/v1/modified_beam_search/recogs-test_ihm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:34:02,301 INFO [utils.py:536] [test_ihm-beam_size_4] %WER 11.53% [24194 / 209845, 2327 ins, 6065 del, 15802 sub ]
2022-12-09 00:34:02,937 INFO [decode.py:492] Wrote detailed error stats to pruned_transducer_stateless7/exp/v1/modified_beam_search/errs-test_ihm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:34:02,939 INFO [decode.py:508] 
For test_ihm, WER of different settings are:
beam_size_4	11.53	best for test_ihm

2022-12-09 00:34:02,939 INFO [decode.py:687] Decoding eval_sdm
2022-12-09 00:34:08,817 INFO [decode.py:463] batch 0/?, cuts processed until now is 58
2022-12-09 00:34:14,230 INFO [decode.py:463] batch 2/?, cuts processed until now is 512
2022-12-09 00:34:14,530 INFO [zipformer.py:1414] attn_weights_entropy = tensor([3.3616, 3.0196, 3.0734, 2.1451, 2.7260, 3.2309, 3.1917, 2.7306],
       device='cuda:0'), covar=tensor([0.0816, 0.1592, 0.1031, 0.1734, 0.1312, 0.0629, 0.0811, 0.1497],
       device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0170, 0.0124, 0.0117, 0.0121, 0.0128, 0.0106, 0.0128],
       device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005],
       device='cuda:0')
2022-12-09 00:34:22,052 INFO [decode.py:463] batch 4/?, cuts processed until now is 645
2022-12-09 00:34:30,387 INFO [decode.py:463] batch 6/?, cuts processed until now is 750
2022-12-09 00:34:38,010 INFO [decode.py:463] batch 8/?, cuts processed until now is 883
2022-12-09 00:34:45,164 INFO [decode.py:463] batch 10/?, cuts processed until now is 1082
2022-12-09 00:34:52,858 INFO [decode.py:463] batch 12/?, cuts processed until now is 1279
2022-12-09 00:34:59,493 INFO [decode.py:463] batch 14/?, cuts processed until now is 1538
2022-12-09 00:35:06,187 INFO [decode.py:463] batch 16/?, cuts processed until now is 1845
2022-12-09 00:35:13,032 INFO [decode.py:463] batch 18/?, cuts processed until now is 2084
2022-12-09 00:35:18,873 INFO [decode.py:463] batch 20/?, cuts processed until now is 2523
2022-12-09 00:35:25,017 INFO [decode.py:463] batch 22/?, cuts processed until now is 2949
2022-12-09 00:35:32,181 INFO [decode.py:463] batch 24/?, cuts processed until now is 3160
2022-12-09 00:35:38,571 INFO [decode.py:463] batch 26/?, cuts processed until now is 3586
2022-12-09 00:35:45,999 INFO [decode.py:463] batch 28/?, cuts processed until now is 3758
2022-12-09 00:35:52,230 INFO [decode.py:463] batch 30/?, cuts processed until now is 4116
2022-12-09 00:35:56,931 INFO [decode.py:463] batch 32/?, cuts processed until now is 4742
2022-12-09 00:36:01,998 INFO [decode.py:463] batch 34/?, cuts processed until now is 5368
2022-12-09 00:36:08,024 INFO [decode.py:463] batch 36/?, cuts processed until now is 5796
2022-12-09 00:36:09,364 INFO [decode.py:463] batch 38/?, cuts processed until now is 5908
2022-12-09 00:36:16,584 INFO [decode.py:463] batch 40/?, cuts processed until now is 6026
2022-12-09 00:36:21,870 INFO [decode.py:463] batch 42/?, cuts processed until now is 6171
2022-12-09 00:36:28,727 INFO [decode.py:463] batch 44/?, cuts processed until now is 6390
2022-12-09 00:36:32,833 INFO [decode.py:463] batch 46/?, cuts processed until now is 6456
2022-12-09 00:36:33,291 INFO [decode.py:479] The transcripts are stored in pruned_transducer_stateless7/exp/v1/modified_beam_search/recogs-eval_sdm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:36:33,393 INFO [utils.py:536] [eval_sdm-beam_size_4] %WER 23.37% [18953 / 81111, 2202 ins, 4566 del, 12185 sub ]
2022-12-09 00:36:33,653 INFO [decode.py:492] Wrote detailed error stats to pruned_transducer_stateless7/exp/v1/modified_beam_search/errs-eval_sdm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:36:33,654 INFO [decode.py:508] 
For eval_sdm, WER of different settings are:
beam_size_4	23.37	best for eval_sdm

2022-12-09 00:36:33,654 INFO [decode.py:687] Decoding test_sdm
2022-12-09 00:36:39,860 INFO [decode.py:463] batch 0/?, cuts processed until now is 49
2022-12-09 00:36:45,565 INFO [decode.py:463] batch 2/?, cuts processed until now is 433
2022-12-09 00:36:53,725 INFO [decode.py:463] batch 4/?, cuts processed until now is 545
2022-12-09 00:37:02,310 INFO [decode.py:463] batch 6/?, cuts processed until now is 637
2022-12-09 00:37:10,388 INFO [decode.py:463] batch 8/?, cuts processed until now is 754
2022-12-09 00:37:18,827 INFO [decode.py:463] batch 10/?, cuts processed until now is 845
2022-12-09 00:37:26,764 INFO [decode.py:463] batch 12/?, cuts processed until now is 976
2022-12-09 00:37:33,846 INFO [decode.py:463] batch 14/?, cuts processed until now is 1175
2022-12-09 00:37:39,823 INFO [decode.py:463] batch 16/?, cuts processed until now is 1483
2022-12-09 00:37:48,201 INFO [decode.py:463] batch 18/?, cuts processed until now is 1590
2022-12-09 00:37:57,103 INFO [decode.py:463] batch 20/?, cuts processed until now is 1658
2022-12-09 00:38:04,045 INFO [decode.py:463] batch 22/?, cuts processed until now is 1856
2022-12-09 00:38:09,733 INFO [decode.py:463] batch 24/?, cuts processed until now is 2224
2022-12-09 00:38:17,904 INFO [decode.py:463] batch 26/?, cuts processed until now is 2325
2022-12-09 00:38:25,264 INFO [decode.py:463] batch 28/?, cuts processed until now is 2546
2022-12-09 00:38:33,413 INFO [decode.py:463] batch 30/?, cuts processed until now is 2653
2022-12-09 00:38:41,809 INFO [decode.py:463] batch 32/?, cuts processed until now is 2744
2022-12-09 00:38:49,880 INFO [decode.py:463] batch 34/?, cuts processed until now is 2875
2022-12-09 00:38:58,512 INFO [decode.py:463] batch 36/?, cuts processed until now is 2961
2022-12-09 00:39:06,612 INFO [decode.py:463] batch 38/?, cuts processed until now is 3072
2022-12-09 00:39:12,803 INFO [decode.py:463] batch 40/?, cuts processed until now is 3440
2022-12-09 00:39:17,555 INFO [decode.py:463] batch 42/?, cuts processed until now is 3956
2022-12-09 00:39:23,290 INFO [decode.py:463] batch 44/?, cuts processed until now is 4342
2022-12-09 00:39:31,656 INFO [decode.py:463] batch 46/?, cuts processed until now is 4443
2022-12-09 00:39:39,129 INFO [decode.py:463] batch 48/?, cuts processed until now is 4595
2022-12-09 00:39:45,308 INFO [decode.py:463] batch 50/?, cuts processed until now is 4872
2022-12-09 00:39:52,345 INFO [decode.py:463] batch 52/?, cuts processed until now is 5061
2022-12-09 00:39:56,217 INFO [zipformer.py:1414] attn_weights_entropy = tensor([1.6000, 1.4851, 1.5670, 1.7876, 2.0592, 2.2136, 2.2064, 2.2146],
       device='cuda:0'), covar=tensor([0.0835, 0.1291, 0.0534, 0.0907, 0.0421, 0.0452, 0.0473, 0.0667],
       device='cuda:0'), in_proj_covar=tensor([0.0013, 0.0014, 0.0012, 0.0013, 0.0013, 0.0022, 0.0018, 0.0023],
       device='cuda:0'), out_proj_covar=tensor([1.0210e-04, 1.1124e-04, 9.6061e-05, 1.0387e-04, 1.0128e-04, 1.6022e-04,
        1.3170e-04, 1.5378e-04], device='cuda:0')
2022-12-09 00:40:00,577 INFO [decode.py:463] batch 54/?, cuts processed until now is 5219
2022-12-09 00:40:03,868 INFO [decode.py:463] batch 56/?, cuts processed until now is 5892
2022-12-09 00:40:10,729 INFO [decode.py:463] batch 58/?, cuts processed until now is 6090
2022-12-09 00:40:15,737 INFO [decode.py:463] batch 60/?, cuts processed until now is 6517
2022-12-09 00:40:22,725 INFO [decode.py:463] batch 62/?, cuts processed until now is 6715
2022-12-09 00:40:29,584 INFO [decode.py:463] batch 64/?, cuts processed until now is 6897
2022-12-09 00:40:34,840 INFO [decode.py:463] batch 66/?, cuts processed until now is 7304
2022-12-09 00:40:38,166 INFO [zipformer.py:1414] attn_weights_entropy = tensor([5.0007, 4.7579, 4.4578, 4.4537, 4.7314, 4.8085, 5.0901, 4.9900],
       device='cuda:0'), covar=tensor([0.0520, 0.0275, 0.1399, 0.3184, 0.0407, 0.0582, 0.0450, 0.0589],
       device='cuda:0'), in_proj_covar=tensor([0.0334, 0.0221, 0.0402, 0.0522, 0.0289, 0.0383, 0.0353, 0.0328],
       device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0004, 0.0005, 0.0003, 0.0004, 0.0004, 0.0003],
       device='cuda:0')
2022-12-09 00:40:41,753 INFO [decode.py:463] batch 68/?, cuts processed until now is 7488
2022-12-09 00:40:45,163 INFO [zipformer.py:1414] attn_weights_entropy = tensor([4.8467, 2.7370, 4.7665, 3.2192, 4.6985, 2.2952, 3.7403, 4.4403],
       device='cuda:0'), covar=tensor([0.0689, 0.4731, 0.0564, 0.8768, 0.0368, 0.4544, 0.1277, 0.0518],
       device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0205, 0.0174, 0.0283, 0.0196, 0.0207, 0.0197, 0.0178],
       device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0004, 0.0004],
       device='cuda:0')
2022-12-09 00:40:48,909 INFO [decode.py:463] batch 70/?, cuts processed until now is 7720
2022-12-09 00:40:55,528 INFO [decode.py:463] batch 72/?, cuts processed until now is 7938
2022-12-09 00:41:00,650 INFO [decode.py:463] batch 74/?, cuts processed until now is 8367
2022-12-09 00:41:06,497 INFO [decode.py:463] batch 76/?, cuts processed until now is 8674
2022-12-09 00:41:12,660 INFO [decode.py:463] batch 78/?, cuts processed until now is 8982
2022-12-09 00:41:18,265 INFO [decode.py:463] batch 80/?, cuts processed until now is 9350
2022-12-09 00:41:23,150 INFO [decode.py:463] batch 82/?, cuts processed until now is 9810
2022-12-09 00:41:28,140 INFO [decode.py:463] batch 84/?, cuts processed until now is 10237
2022-12-09 00:41:32,828 INFO [decode.py:463] batch 86/?, cuts processed until now is 10755
2022-12-09 00:41:37,127 INFO [decode.py:463] batch 88/?, cuts processed until now is 11278
2022-12-09 00:41:42,072 INFO [decode.py:463] batch 90/?, cuts processed until now is 11705
2022-12-09 00:41:48,165 INFO [decode.py:463] batch 92/?, cuts processed until now is 12013
2022-12-09 00:41:54,603 INFO [decode.py:463] batch 94/?, cuts processed until now is 12290
2022-12-09 00:42:00,600 INFO [decode.py:463] batch 96/?, cuts processed until now is 12597
2022-12-09 00:42:06,239 INFO [decode.py:463] batch 98/?, cuts processed until now is 12963
2022-12-09 00:42:10,927 INFO [decode.py:463] batch 100/?, cuts processed until now is 13420
2022-12-09 00:42:16,548 INFO [decode.py:463] batch 102/?, cuts processed until now is 13877
2022-12-09 00:42:21,141 INFO [decode.py:463] batch 104/?, cuts processed until now is 14543
2022-12-09 00:42:25,549 INFO [decode.py:463] batch 106/?, cuts processed until now is 15209
2022-12-09 00:42:30,704 INFO [decode.py:463] batch 108/?, cuts processed until now is 15599
2022-12-09 00:42:34,228 INFO [decode.py:463] batch 110/?, cuts processed until now is 15787
2022-12-09 00:42:37,414 INFO [decode.py:463] batch 112/?, cuts processed until now is 15881
2022-12-09 00:42:42,239 INFO [decode.py:463] batch 114/?, cuts processed until now is 15926
2022-12-09 00:42:44,787 INFO [decode.py:463] batch 116/?, cuts processed until now is 16287
2022-12-09 00:42:50,934 INFO [decode.py:463] batch 118/?, cuts processed until now is 16357
2022-12-09 00:42:51,160 INFO [decode.py:479] The transcripts are stored in pruned_transducer_stateless7/exp/v1/modified_beam_search/recogs-test_sdm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:42:51,439 INFO [utils.py:536] [test_sdm-beam_size_4] %WER 25.85% [54237 / 209845, 5655 ins, 15315 del, 33267 sub ]
2022-12-09 00:42:52,114 INFO [decode.py:492] Wrote detailed error stats to pruned_transducer_stateless7/exp/v1/modified_beam_search/errs-test_sdm-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:42:52,131 INFO [decode.py:508] 
For test_sdm, WER of different settings are:
beam_size_4	25.85	best for test_sdm

2022-12-09 00:42:52,131 INFO [decode.py:687] Decoding eval_gss
2022-12-09 00:42:57,705 INFO [decode.py:463] batch 0/?, cuts processed until now is 58
2022-12-09 00:43:02,998 INFO [decode.py:463] batch 2/?, cuts processed until now is 512
2022-12-09 00:43:10,732 INFO [decode.py:463] batch 4/?, cuts processed until now is 645
2022-12-09 00:43:18,974 INFO [decode.py:463] batch 6/?, cuts processed until now is 750
2022-12-09 00:43:26,729 INFO [decode.py:463] batch 8/?, cuts processed until now is 883
2022-12-09 00:43:34,520 INFO [decode.py:463] batch 10/?, cuts processed until now is 1082
2022-12-09 00:43:42,187 INFO [decode.py:463] batch 12/?, cuts processed until now is 1279
2022-12-09 00:43:48,868 INFO [decode.py:463] batch 14/?, cuts processed until now is 1538
2022-12-09 00:43:55,552 INFO [decode.py:463] batch 16/?, cuts processed until now is 1845
2022-12-09 00:44:02,697 INFO [decode.py:463] batch 18/?, cuts processed until now is 2084
2022-12-09 00:44:08,512 INFO [decode.py:463] batch 20/?, cuts processed until now is 2523
2022-12-09 00:44:14,694 INFO [decode.py:463] batch 22/?, cuts processed until now is 2949
2022-12-09 00:44:21,839 INFO [decode.py:463] batch 24/?, cuts processed until now is 3160
2022-12-09 00:44:28,126 INFO [decode.py:463] batch 26/?, cuts processed until now is 3586
2022-12-09 00:44:35,600 INFO [decode.py:463] batch 28/?, cuts processed until now is 3758
2022-12-09 00:44:41,779 INFO [decode.py:463] batch 30/?, cuts processed until now is 4116
2022-12-09 00:44:46,367 INFO [decode.py:463] batch 32/?, cuts processed until now is 4742
2022-12-09 00:44:51,420 INFO [decode.py:463] batch 34/?, cuts processed until now is 5368
2022-12-09 00:44:57,453 INFO [decode.py:463] batch 36/?, cuts processed until now is 5796
2022-12-09 00:44:58,784 INFO [decode.py:463] batch 38/?, cuts processed until now is 5908
2022-12-09 00:45:06,648 INFO [decode.py:463] batch 40/?, cuts processed until now is 6026
2022-12-09 00:45:11,801 INFO [decode.py:463] batch 42/?, cuts processed until now is 6171
2022-12-09 00:45:18,587 INFO [decode.py:463] batch 44/?, cuts processed until now is 6390
2022-12-09 00:45:22,679 INFO [decode.py:463] batch 46/?, cuts processed until now is 6456
2022-12-09 00:45:23,140 INFO [decode.py:479] The transcripts are stored in pruned_transducer_stateless7/exp/v1/modified_beam_search/recogs-eval_gss-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:45:23,237 INFO [utils.py:536] [eval_gss-beam_size_4] %WER 11.82% [9586 / 81111, 1083 ins, 2070 del, 6433 sub ]
2022-12-09 00:45:23,529 INFO [decode.py:492] Wrote detailed error stats to pruned_transducer_stateless7/exp/v1/modified_beam_search/errs-eval_gss-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:45:23,530 INFO [decode.py:508] 
For eval_gss, WER of different settings are:
beam_size_4	11.82	best for eval_gss

2022-12-09 00:45:23,530 INFO [decode.py:687] Decoding test_gss
2022-12-09 00:45:29,371 INFO [decode.py:463] batch 0/?, cuts processed until now is 49
2022-12-09 00:45:35,151 INFO [decode.py:463] batch 2/?, cuts processed until now is 433
2022-12-09 00:45:43,327 INFO [decode.py:463] batch 4/?, cuts processed until now is 545
2022-12-09 00:45:51,796 INFO [decode.py:463] batch 6/?, cuts processed until now is 637
2022-12-09 00:45:59,790 INFO [decode.py:463] batch 8/?, cuts processed until now is 754
2022-12-09 00:46:08,420 INFO [decode.py:463] batch 10/?, cuts processed until now is 845
2022-12-09 00:46:16,376 INFO [decode.py:463] batch 12/?, cuts processed until now is 976
2022-12-09 00:46:23,166 INFO [decode.py:463] batch 14/?, cuts processed until now is 1175
2022-12-09 00:46:29,309 INFO [decode.py:463] batch 16/?, cuts processed until now is 1483
2022-12-09 00:46:37,642 INFO [decode.py:463] batch 18/?, cuts processed until now is 1590
2022-12-09 00:46:46,533 INFO [decode.py:463] batch 20/?, cuts processed until now is 1658
2022-12-09 00:46:53,462 INFO [decode.py:463] batch 22/?, cuts processed until now is 1856
2022-12-09 00:46:59,089 INFO [decode.py:463] batch 24/?, cuts processed until now is 2224
2022-12-09 00:47:07,379 INFO [decode.py:463] batch 26/?, cuts processed until now is 2325
2022-12-09 00:47:14,709 INFO [decode.py:463] batch 28/?, cuts processed until now is 2546
2022-12-09 00:47:22,610 INFO [decode.py:463] batch 30/?, cuts processed until now is 2653
2022-12-09 00:47:31,118 INFO [decode.py:463] batch 32/?, cuts processed until now is 2744
2022-12-09 00:47:39,205 INFO [decode.py:463] batch 34/?, cuts processed until now is 2875
2022-12-09 00:47:47,684 INFO [decode.py:463] batch 36/?, cuts processed until now is 2961
2022-12-09 00:47:55,780 INFO [decode.py:463] batch 38/?, cuts processed until now is 3072
2022-12-09 00:48:01,964 INFO [decode.py:463] batch 40/?, cuts processed until now is 3440
2022-12-09 00:48:06,693 INFO [decode.py:463] batch 42/?, cuts processed until now is 3956
2022-12-09 00:48:12,503 INFO [decode.py:463] batch 44/?, cuts processed until now is 4342
2022-12-09 00:48:20,529 INFO [decode.py:463] batch 46/?, cuts processed until now is 4443
2022-12-09 00:48:27,926 INFO [decode.py:463] batch 48/?, cuts processed until now is 4595
2022-12-09 00:48:34,365 INFO [decode.py:463] batch 50/?, cuts processed until now is 4872
2022-12-09 00:48:41,248 INFO [decode.py:463] batch 52/?, cuts processed until now is 5061
2022-12-09 00:48:49,108 INFO [decode.py:463] batch 54/?, cuts processed until now is 5219
2022-12-09 00:48:49,354 INFO [zipformer.py:1414] attn_weights_entropy = tensor([1.7659, 1.8372, 1.8194, 1.5972, 1.5217, 1.0486, 1.1054, 0.9389],
       device='cuda:0'), covar=tensor([0.0083, 0.0055, 0.0058, 0.0054, 0.0177, 0.0286, 0.0174, 0.0357],
       device='cuda:0'), in_proj_covar=tensor([0.0013, 0.0014, 0.0012, 0.0013, 0.0013, 0.0022, 0.0018, 0.0023],
       device='cuda:0'), out_proj_covar=tensor([1.0210e-04, 1.1124e-04, 9.6061e-05, 1.0387e-04, 1.0128e-04, 1.6022e-04,
        1.3170e-04, 1.5378e-04], device='cuda:0')
2022-12-09 00:48:52,461 INFO [decode.py:463] batch 56/?, cuts processed until now is 5892
2022-12-09 00:48:59,524 INFO [decode.py:463] batch 58/?, cuts processed until now is 6090
2022-12-09 00:49:04,540 INFO [decode.py:463] batch 60/?, cuts processed until now is 6517
2022-12-09 00:49:11,300 INFO [decode.py:463] batch 62/?, cuts processed until now is 6715
2022-12-09 00:49:18,348 INFO [decode.py:463] batch 64/?, cuts processed until now is 6897
2022-12-09 00:49:23,684 INFO [decode.py:463] batch 66/?, cuts processed until now is 7304
2022-12-09 00:49:30,622 INFO [decode.py:463] batch 68/?, cuts processed until now is 7488
2022-12-09 00:49:37,553 INFO [decode.py:463] batch 70/?, cuts processed until now is 7720
2022-12-09 00:49:44,162 INFO [decode.py:463] batch 72/?, cuts processed until now is 7938
2022-12-09 00:49:49,325 INFO [decode.py:463] batch 74/?, cuts processed until now is 8367
2022-12-09 00:49:55,304 INFO [decode.py:463] batch 76/?, cuts processed until now is 8674
2022-12-09 00:50:01,418 INFO [decode.py:463] batch 78/?, cuts processed until now is 8982
2022-12-09 00:50:06,983 INFO [decode.py:463] batch 80/?, cuts processed until now is 9350
2022-12-09 00:50:12,257 INFO [decode.py:463] batch 82/?, cuts processed until now is 9810
2022-12-09 00:50:17,564 INFO [decode.py:463] batch 84/?, cuts processed until now is 10237
2022-12-09 00:50:22,238 INFO [decode.py:463] batch 86/?, cuts processed until now is 10755
2022-12-09 00:50:26,646 INFO [decode.py:463] batch 88/?, cuts processed until now is 11278
2022-12-09 00:50:31,792 INFO [decode.py:463] batch 90/?, cuts processed until now is 11705
2022-12-09 00:50:37,988 INFO [decode.py:463] batch 92/?, cuts processed until now is 12013
2022-12-09 00:50:44,298 INFO [decode.py:463] batch 94/?, cuts processed until now is 12290
2022-12-09 00:50:50,344 INFO [decode.py:463] batch 96/?, cuts processed until now is 12597
2022-12-09 00:50:56,139 INFO [decode.py:463] batch 98/?, cuts processed until now is 12963
2022-12-09 00:51:00,984 INFO [decode.py:463] batch 100/?, cuts processed until now is 13420
2022-12-09 00:51:06,482 INFO [decode.py:463] batch 102/?, cuts processed until now is 13877
2022-12-09 00:51:11,124 INFO [decode.py:463] batch 104/?, cuts processed until now is 14543
2022-12-09 00:51:15,561 INFO [decode.py:463] batch 106/?, cuts processed until now is 15209
2022-12-09 00:51:20,691 INFO [decode.py:463] batch 108/?, cuts processed until now is 15599
2022-12-09 00:51:24,220 INFO [decode.py:463] batch 110/?, cuts processed until now is 15787
2022-12-09 00:51:27,406 INFO [decode.py:463] batch 112/?, cuts processed until now is 15881
2022-12-09 00:51:32,122 INFO [decode.py:463] batch 114/?, cuts processed until now is 15926
2022-12-09 00:51:34,657 INFO [decode.py:463] batch 116/?, cuts processed until now is 16287
2022-12-09 00:51:40,769 INFO [decode.py:463] batch 118/?, cuts processed until now is 16357
2022-12-09 00:51:41,004 INFO [decode.py:479] The transcripts are stored in pruned_transducer_stateless7/exp/v1/modified_beam_search/recogs-test_gss-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:51:41,262 INFO [utils.py:536] [test_gss-beam_size_4] %WER 14.22% [29831 / 209845, 2727 ins, 7497 del, 19607 sub ]
2022-12-09 00:51:41,889 INFO [decode.py:492] Wrote detailed error stats to pruned_transducer_stateless7/exp/v1/modified_beam_search/errs-test_gss-beam_size_4-epoch-15-avg-8-modified_beam_search-beam-size-4.txt
2022-12-09 00:51:41,890 INFO [decode.py:508] 
For test_gss, WER of different settings are:
beam_size_4	14.22	best for test_gss

2022-12-09 00:51:41,891 INFO [decode.py:703] Done!