File size: 95,242 Bytes
fe00c53
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
2022-04-09 01:40:41,909 INFO [decode_test.py:583] Decoding started
2022-04-09 01:40:41,910 INFO [decode_test.py:584] {'subsampling_factor': 4, 'vgg_frontend': False, 'use_feat_batchnorm': True, 'feature_dim': 80, 'nhead': 8, 'attention_dim': 512, 'num_decoder_layers': 6, 'search_beam': 20, 'output_beam': 8, 'min_active_states': 30, 'max_active_states': 10000, 'use_double_scores': True, 'env_info': {'k2-version': '1.14', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '6833270cb228aba7bf9681fccd41e2b52f7d984c', 'k2-git-date': 'Wed Mar 16 11:16:05 2022', 'lhotse-version': '1.0.0.dev+git.d917411.clean', 'torch-cuda-available': True, 'torch-cuda-version': '11.1', 'python-version': '3.7', 'icefall-git-branch': 'gigaspeech_recipe', 'icefall-git-sha1': 'c3993a5-dirty', 'icefall-git-date': 'Mon Mar 21 13:49:39 2022', 'icefall-path': '/userhome/user/guanbo/icefall_decode', 'k2-path': '/opt/conda/lib/python3.7/site-packages/k2-1.14.dev20220408+cuda11.1.torch1.10.0-py3.7-linux-x86_64.egg/k2/__init__.py', 'lhotse-path': '/userhome/user/guanbo/lhotse/lhotse/__init__.py', 'hostname': 'c8861f400b70d011ec0a3ee069db84328338-chenx8564-0', 'IP address': '10.9.150.55'}, 'epoch': 18, 'avg': 6, 'method': 'attention-decoder', 'num_paths': 1000, 'nbest_scale': 0.5, 'exp_dir': PosixPath('conformer_ctc/exp_500_8_2'), 'lang_dir': PosixPath('data/lang_bpe_500'), 'lm_dir': PosixPath('data/lm'), 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 20, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'return_cuts': True, 'num_workers': 1, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'subset': 'XL', 'lazy_load': True, 'small_dev': False}
2022-04-09 01:40:42,371 INFO [lexicon.py:176] Loading pre-compiled data/lang_bpe_500/Linv.pt
2022-04-09 01:40:42,473 INFO [decode_test.py:594] device: cuda:0
2022-04-09 01:40:46,249 INFO [decode_test.py:656] Loading pre-compiled G_4_gram.pt
2022-04-09 01:40:47,406 INFO [decode_test.py:692] averaging ['conformer_ctc/exp_500_8_2/epoch-13.pt', 'conformer_ctc/exp_500_8_2/epoch-14.pt', 'conformer_ctc/exp_500_8_2/epoch-15.pt', 'conformer_ctc/exp_500_8_2/epoch-16.pt', 'conformer_ctc/exp_500_8_2/epoch-17.pt', 'conformer_ctc/exp_500_8_2/epoch-18.pt']
2022-04-09 01:40:53,065 INFO [decode_test.py:699] Number of model parameters: 109226120
2022-04-09 01:40:53,065 INFO [asr_datamodule.py:381] About to get test cuts
2022-04-09 01:40:56,361 INFO [decode_test.py:497] batch 0/?, cuts processed until now is 3
2022-04-09 01:41:24,462 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 5.93 GiB (GPU 0; 31.75 GiB total capacity; 27.23 GiB already allocated; 1.90 GiB free; 28.49 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:41:24,462 INFO [decode.py:743] num_arcs before pruning: 324363
2022-04-09 01:41:24,462 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:41:24,473 INFO [decode.py:757] num_arcs after pruning: 7174
2022-04-09 01:41:40,284 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.67 GiB (GPU 0; 31.75 GiB total capacity; 25.69 GiB already allocated; 2.92 GiB free; 27.47 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:41:40,285 INFO [decode.py:743] num_arcs before pruning: 368362
2022-04-09 01:41:40,285 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:41:40,305 INFO [decode.py:757] num_arcs after pruning: 8521
2022-04-09 01:42:38,727 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.18 GiB (GPU 0; 31.75 GiB total capacity; 26.05 GiB already allocated; 1.42 GiB free; 28.98 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:42:38,727 INFO [decode.py:743] num_arcs before pruning: 432616
2022-04-09 01:42:38,728 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:42:38,736 INFO [decode.py:757] num_arcs after pruning: 9233
2022-04-09 01:43:13,573 INFO [decode_test.py:497] batch 100/?, cuts processed until now is 297
2022-04-09 01:43:48,362 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 25.34 GiB already allocated; 2.20 GiB free; 28.20 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:43:48,363 INFO [decode.py:743] num_arcs before pruning: 319907
2022-04-09 01:43:48,363 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:43:48,372 INFO [decode.py:757] num_arcs after pruning: 6358
2022-04-09 01:43:59,713 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.74 GiB (GPU 0; 31.75 GiB total capacity; 27.51 GiB already allocated; 2.19 GiB free; 28.20 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:43:59,713 INFO [decode.py:743] num_arcs before pruning: 313596
2022-04-09 01:43:59,713 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:43:59,724 INFO [decode.py:757] num_arcs after pruning: 8252
2022-04-09 01:44:54,463 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 25.25 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:44:54,463 INFO [decode.py:743] num_arcs before pruning: 353355
2022-04-09 01:44:54,463 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:44:54,485 INFO [decode.py:757] num_arcs after pruning: 7520
2022-04-09 01:45:20,716 INFO [decode_test.py:497] batch 200/?, cuts processed until now is 570
2022-04-09 01:47:19,457 INFO [decode_test.py:497] batch 300/?, cuts processed until now is 806
2022-04-09 01:47:38,292 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.28 GiB (GPU 0; 31.75 GiB total capacity; 26.28 GiB already allocated; 1.48 GiB free; 28.92 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:47:38,293 INFO [decode.py:743] num_arcs before pruning: 596002
2022-04-09 01:47:38,293 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:47:38,312 INFO [decode.py:757] num_arcs after pruning: 10745
2022-04-09 01:49:18,493 INFO [decode.py:736] Caught exception:

    Some bad things happened. Please read the above error messages and stack
    trace. If you are using Python, the following command may be helpful:

      gdb --args python /path/to/your/code.py

    (You can use `gdb` to debug the code. Please consider compiling
    a debug version of k2.).

    If you are unable to fix it, please open an issue at:

      https://github.com/k2-fsa/k2/issues/new
    

2022-04-09 01:49:18,494 INFO [decode.py:743] num_arcs before pruning: 398202
2022-04-09 01:49:18,494 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:49:18,541 INFO [decode.py:757] num_arcs after pruning: 14003
2022-04-09 01:49:21,800 INFO [decode_test.py:497] batch 400/?, cuts processed until now is 1082
2022-04-09 01:50:58,700 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.85 GiB (GPU 0; 31.75 GiB total capacity; 25.89 GiB already allocated; 1.48 GiB free; 28.92 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:50:58,701 INFO [decode.py:743] num_arcs before pruning: 398349
2022-04-09 01:50:58,701 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:50:58,709 INFO [decode.py:757] num_arcs after pruning: 10321
2022-04-09 01:51:31,627 INFO [decode_test.py:497] batch 500/?, cuts processed until now is 1334
2022-04-09 01:52:05,232 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.62 GiB already allocated; 1.47 GiB free; 28.93 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:52:05,232 INFO [decode.py:743] num_arcs before pruning: 212665
2022-04-09 01:52:05,232 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:52:05,241 INFO [decode.py:757] num_arcs after pruning: 6301
2022-04-09 01:53:29,890 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 1.91 GiB (GPU 0; 31.75 GiB total capacity; 25.66 GiB already allocated; 1.48 GiB free; 28.92 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:53:29,891 INFO [decode.py:743] num_arcs before pruning: 883555
2022-04-09 01:53:29,891 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:53:29,905 INFO [decode.py:757] num_arcs after pruning: 14819
2022-04-09 01:53:38,676 INFO [decode_test.py:497] batch 600/?, cuts processed until now is 1651
2022-04-09 01:54:57,438 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 25.34 GiB already allocated; 1.48 GiB free; 28.92 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:54:57,438 INFO [decode.py:743] num_arcs before pruning: 515795
2022-04-09 01:54:57,438 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:54:57,447 INFO [decode.py:757] num_arcs after pruning: 10132
2022-04-09 01:55:28,356 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.46 GiB already allocated; 1.48 GiB free; 28.92 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:55:28,356 INFO [decode.py:743] num_arcs before pruning: 670748
2022-04-09 01:55:28,356 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:55:28,365 INFO [decode.py:757] num_arcs after pruning: 10497
2022-04-09 01:55:42,238 INFO [decode_test.py:497] batch 700/?, cuts processed until now is 1956
2022-04-09 01:57:57,456 INFO [decode_test.py:497] batch 800/?, cuts processed until now is 2238
2022-04-09 01:58:04,281 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.45 GiB already allocated; 3.07 GiB free; 27.33 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:58:04,282 INFO [decode.py:743] num_arcs before pruning: 175423
2022-04-09 01:58:04,282 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:58:04,296 INFO [decode.py:757] num_arcs after pruning: 7926
2022-04-09 01:59:07,916 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.68 GiB (GPU 0; 31.75 GiB total capacity; 24.40 GiB already allocated; 3.06 GiB free; 27.33 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 01:59:07,917 INFO [decode.py:743] num_arcs before pruning: 259758
2022-04-09 01:59:07,917 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 01:59:07,928 INFO [decode.py:757] num_arcs after pruning: 6026
2022-04-09 02:00:00,623 INFO [decode_test.py:497] batch 900/?, cuts processed until now is 2536
2022-04-09 02:01:22,959 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.44 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:01:22,959 INFO [decode.py:743] num_arcs before pruning: 749228
2022-04-09 02:01:22,959 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:01:22,968 INFO [decode.py:757] num_arcs after pruning: 23868
2022-04-09 02:01:59,449 INFO [decode_test.py:497] batch 1000/?, cuts processed until now is 2824
2022-04-09 02:03:05,494 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.38 GiB already allocated; 3.06 GiB free; 27.33 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:03:05,494 INFO [decode.py:743] num_arcs before pruning: 255135
2022-04-09 02:03:05,494 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:03:05,504 INFO [decode.py:757] num_arcs after pruning: 5955
2022-04-09 02:03:48,017 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.61 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:03:48,017 INFO [decode.py:743] num_arcs before pruning: 517077
2022-04-09 02:03:48,017 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:03:48,026 INFO [decode.py:757] num_arcs after pruning: 7695
2022-04-09 02:04:09,806 INFO [decode_test.py:497] batch 1100/?, cuts processed until now is 3105
2022-04-09 02:04:31,410 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.34 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:04:31,411 INFO [decode.py:743] num_arcs before pruning: 859561
2022-04-09 02:04:31,411 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:04:31,422 INFO [decode.py:757] num_arcs after pruning: 13014
2022-04-09 02:06:11,496 INFO [decode_test.py:497] batch 1200/?, cuts processed until now is 3401
2022-04-09 02:08:10,727 INFO [decode_test.py:497] batch 1300/?, cuts processed until now is 3730
2022-04-09 02:10:17,677 INFO [decode_test.py:497] batch 1400/?, cuts processed until now is 4067
2022-04-09 02:12:13,175 INFO [decode_test.py:497] batch 1500/?, cuts processed until now is 4329
2022-04-09 02:13:02,842 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.55 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:13:02,843 INFO [decode.py:743] num_arcs before pruning: 475511
2022-04-09 02:13:02,843 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:13:02,849 INFO [decode.py:757] num_arcs after pruning: 8439
2022-04-09 02:13:46,588 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.37 GiB (GPU 0; 31.75 GiB total capacity; 26.83 GiB already allocated; 1.45 GiB free; 28.94 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:13:46,588 INFO [decode.py:743] num_arcs before pruning: 595488
2022-04-09 02:13:46,588 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:13:46,598 INFO [decode.py:757] num_arcs after pruning: 13475
2022-04-09 02:14:21,206 INFO [decode_test.py:497] batch 1600/?, cuts processed until now is 4598
2022-04-09 02:16:42,740 INFO [decode_test.py:497] batch 1700/?, cuts processed until now is 4969
2022-04-09 02:17:13,672 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 25.39 GiB already allocated; 1.45 GiB free; 28.94 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:17:13,673 INFO [decode.py:743] num_arcs before pruning: 615734
2022-04-09 02:17:13,673 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:17:13,685 INFO [decode.py:757] num_arcs after pruning: 8684
2022-04-09 02:18:54,514 INFO [decode_test.py:497] batch 1800/?, cuts processed until now is 5260
2022-04-09 02:18:59,938 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.36 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:18:59,938 INFO [decode.py:743] num_arcs before pruning: 360099
2022-04-09 02:18:59,938 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:18:59,949 INFO [decode.py:757] num_arcs after pruning: 6898
2022-04-09 02:19:48,186 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 6.00 GiB (GPU 0; 31.75 GiB total capacity; 27.15 GiB already allocated; 967.75 MiB free; 29.45 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:19:48,186 INFO [decode.py:743] num_arcs before pruning: 168720
2022-04-09 02:19:48,186 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:19:48,201 INFO [decode.py:757] num_arcs after pruning: 5346
2022-04-09 02:20:52,049 INFO [decode_test.py:497] batch 1900/?, cuts processed until now is 5585
2022-04-09 02:22:12,107 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.45 GiB already allocated; 973.75 MiB free; 29.44 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:22:12,107 INFO [decode.py:743] num_arcs before pruning: 1151735
2022-04-09 02:22:12,107 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:22:12,120 INFO [decode.py:757] num_arcs after pruning: 8335
2022-04-09 02:23:01,497 INFO [decode_test.py:497] batch 2000/?, cuts processed until now is 5902
2022-04-09 02:25:26,356 INFO [decode_test.py:497] batch 2100/?, cuts processed until now is 6219
2022-04-09 02:25:56,466 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.34 GiB already allocated; 973.75 MiB free; 29.44 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:25:56,467 INFO [decode.py:743] num_arcs before pruning: 612804
2022-04-09 02:25:56,467 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:25:56,477 INFO [decode.py:757] num_arcs after pruning: 10853
2022-04-09 02:27:26,441 INFO [decode_test.py:497] batch 2200/?, cuts processed until now is 6480
2022-04-09 02:29:28,073 INFO [decode_test.py:497] batch 2300/?, cuts processed until now is 6768
2022-04-09 02:31:41,553 INFO [decode_test.py:497] batch 2400/?, cuts processed until now is 7120
2022-04-09 02:31:55,632 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.42 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:31:55,632 INFO [decode.py:743] num_arcs before pruning: 411490
2022-04-09 02:31:55,632 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:31:55,638 INFO [decode.py:757] num_arcs after pruning: 8626
2022-04-09 02:33:22,034 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.42 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:33:22,034 INFO [decode.py:743] num_arcs before pruning: 625728
2022-04-09 02:33:22,035 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:33:22,043 INFO [decode.py:757] num_arcs after pruning: 9502
2022-04-09 02:33:37,663 INFO [decode_test.py:497] batch 2500/?, cuts processed until now is 7387
2022-04-09 02:34:18,300 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.51 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:34:18,301 INFO [decode.py:743] num_arcs before pruning: 1015956
2022-04-09 02:34:18,301 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:34:18,314 INFO [decode.py:757] num_arcs after pruning: 14404
2022-04-09 02:34:20,220 INFO [decode.py:841] Caught exception:
CUDA out of memory. Tried to allocate 5.58 GiB (GPU 0; 31.75 GiB total capacity; 24.87 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:34:20,221 INFO [decode.py:843] num_paths before decreasing: 1000
2022-04-09 02:34:20,221 INFO [decode.py:852] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:34:20,221 INFO [decode.py:858] num_paths after decreasing: 500
2022-04-09 02:34:40,089 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.38 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:34:40,089 INFO [decode.py:743] num_arcs before pruning: 570686
2022-04-09 02:34:40,089 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:34:40,098 INFO [decode.py:757] num_arcs after pruning: 9182
2022-04-09 02:35:50,624 INFO [decode_test.py:497] batch 2600/?, cuts processed until now is 7764
2022-04-09 02:36:44,519 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.61 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:36:44,519 INFO [decode.py:743] num_arcs before pruning: 1066267
2022-04-09 02:36:44,519 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:36:44,530 INFO [decode.py:757] num_arcs after pruning: 6963
2022-04-09 02:38:18,717 INFO [decode_test.py:497] batch 2700/?, cuts processed until now is 8078
2022-04-09 02:40:07,021 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.42 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:40:07,022 INFO [decode.py:743] num_arcs before pruning: 1023667
2022-04-09 02:40:07,022 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:40:07,034 INFO [decode.py:757] num_arcs after pruning: 13090
2022-04-09 02:40:25,184 INFO [decode_test.py:497] batch 2800/?, cuts processed until now is 8444
2022-04-09 02:41:27,080 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.32 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:41:27,080 INFO [decode.py:743] num_arcs before pruning: 739744
2022-04-09 02:41:27,080 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:41:27,093 INFO [decode.py:757] num_arcs after pruning: 9791
2022-04-09 02:42:44,319 INFO [decode_test.py:497] batch 2900/?, cuts processed until now is 8765
2022-04-09 02:42:44,656 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.73 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:42:44,656 INFO [decode.py:743] num_arcs before pruning: 666168
2022-04-09 02:42:44,656 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:42:44,665 INFO [decode.py:757] num_arcs after pruning: 17223
2022-04-09 02:43:05,748 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 5.60 GiB (GPU 0; 31.75 GiB total capacity; 26.18 GiB already allocated; 1.14 GiB free; 29.26 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:43:05,748 INFO [decode.py:743] num_arcs before pruning: 188729
2022-04-09 02:43:05,748 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:43:05,762 INFO [decode.py:757] num_arcs after pruning: 8688
2022-04-09 02:44:54,469 INFO [decode_test.py:497] batch 3000/?, cuts processed until now is 9050
2022-04-09 02:46:55,167 INFO [decode_test.py:497] batch 3100/?, cuts processed until now is 9296
2022-04-09 02:47:28,418 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 20.00 GiB already allocated; 3.07 GiB free; 27.33 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:47:28,419 INFO [decode.py:743] num_arcs before pruning: 160153
2022-04-09 02:47:28,419 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:47:28,448 INFO [decode.py:757] num_arcs after pruning: 7778
2022-04-09 02:49:21,448 INFO [decode_test.py:497] batch 3200/?, cuts processed until now is 9652
2022-04-09 02:50:17,558 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 6.13 GiB (GPU 0; 31.75 GiB total capacity; 27.60 GiB already allocated; 895.75 MiB free; 29.52 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:50:17,558 INFO [decode.py:743] num_arcs before pruning: 388116
2022-04-09 02:50:17,559 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:50:17,565 INFO [decode.py:757] num_arcs after pruning: 10555
2022-04-09 02:51:30,675 INFO [decode_test.py:497] batch 3300/?, cuts processed until now is 10071
2022-04-09 02:53:49,565 INFO [decode_test.py:497] batch 3400/?, cuts processed until now is 10342
2022-04-09 02:55:49,392 INFO [decode_test.py:497] batch 3500/?, cuts processed until now is 10642
2022-04-09 02:58:07,518 INFO [decode_test.py:497] batch 3600/?, cuts processed until now is 10951
2022-04-09 02:58:16,360 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.29 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 02:58:16,361 INFO [decode.py:743] num_arcs before pruning: 396714
2022-04-09 02:58:16,361 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 02:58:16,374 INFO [decode.py:757] num_arcs after pruning: 9543
2022-04-09 03:00:00,485 INFO [decode_test.py:497] batch 3700/?, cuts processed until now is 11231
2022-04-09 03:00:17,600 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.45 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:00:17,601 INFO [decode.py:743] num_arcs before pruning: 854366
2022-04-09 03:00:17,601 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:00:17,612 INFO [decode.py:757] num_arcs after pruning: 10487
2022-04-09 03:00:20,098 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.68 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:00:20,098 INFO [decode.py:743] num_arcs before pruning: 442824
2022-04-09 03:00:20,098 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:00:20,108 INFO [decode.py:757] num_arcs after pruning: 5265
2022-04-09 03:02:00,114 INFO [decode_test.py:497] batch 3800/?, cuts processed until now is 11509
2022-04-09 03:02:11,570 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.19 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:02:11,571 INFO [decode.py:743] num_arcs before pruning: 285638
2022-04-09 03:02:11,571 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:02:11,579 INFO [decode.py:757] num_arcs after pruning: 5903
2022-04-09 03:04:02,757 INFO [decode_test.py:497] batch 3900/?, cuts processed until now is 11774
2022-04-09 03:05:19,989 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.73 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:05:19,990 INFO [decode.py:743] num_arcs before pruning: 637327
2022-04-09 03:05:19,990 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:05:19,999 INFO [decode.py:757] num_arcs after pruning: 6357
2022-04-09 03:06:01,953 INFO [decode_test.py:497] batch 4000/?, cuts processed until now is 12045
2022-04-09 03:07:49,854 INFO [decode_test.py:497] batch 4100/?, cuts processed until now is 12300
2022-04-09 03:09:15,137 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.45 GiB already allocated; 3.08 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:09:15,138 INFO [decode.py:743] num_arcs before pruning: 507733
2022-04-09 03:09:15,138 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:09:15,148 INFO [decode.py:757] num_arcs after pruning: 4196
2022-04-09 03:09:47,397 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 5.86 GiB (GPU 0; 31.75 GiB total capacity; 27.78 GiB already allocated; 925.75 MiB free; 29.49 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:09:47,397 INFO [decode.py:743] num_arcs before pruning: 514118
2022-04-09 03:09:47,397 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:09:47,407 INFO [decode.py:757] num_arcs after pruning: 7168
2022-04-09 03:10:00,013 INFO [decode_test.py:497] batch 4200/?, cuts processed until now is 12580
2022-04-09 03:10:33,411 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.80 GiB (GPU 0; 31.75 GiB total capacity; 27.70 GiB already allocated; 925.75 MiB free; 29.49 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:10:33,411 INFO [decode.py:743] num_arcs before pruning: 374935
2022-04-09 03:10:33,411 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:10:33,418 INFO [decode.py:757] num_arcs after pruning: 10023
2022-04-09 03:12:04,333 INFO [decode_test.py:497] batch 4300/?, cuts processed until now is 12807
2022-04-09 03:14:06,889 INFO [decode_test.py:497] batch 4400/?, cuts processed until now is 13050
2022-04-09 03:14:34,787 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.47 GiB already allocated; 925.75 MiB free; 29.49 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:14:34,788 INFO [decode.py:743] num_arcs before pruning: 767465
2022-04-09 03:14:34,788 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:14:34,797 INFO [decode.py:757] num_arcs after pruning: 19151
2022-04-09 03:15:08,864 INFO [decode.py:736] Caught exception:

    Some bad things happened. Please read the above error messages and stack
    trace. If you are using Python, the following command may be helpful:

      gdb --args python /path/to/your/code.py

    (You can use `gdb` to debug the code. Please consider compiling
    a debug version of k2.).

    If you are unable to fix it, please open an issue at:

      https://github.com/k2-fsa/k2/issues/new
    

2022-04-09 03:15:08,864 INFO [decode.py:743] num_arcs before pruning: 123833
2022-04-09 03:15:08,864 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:15:08,913 INFO [decode.py:757] num_arcs after pruning: 4150
2022-04-09 03:15:34,899 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 25.64 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:15:34,899 INFO [decode.py:743] num_arcs before pruning: 444800
2022-04-09 03:15:34,899 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:15:34,908 INFO [decode.py:757] num_arcs after pruning: 11839
2022-04-09 03:16:08,462 INFO [decode_test.py:497] batch 4500/?, cuts processed until now is 13295
2022-04-09 03:17:56,946 INFO [decode_test.py:497] batch 4600/?, cuts processed until now is 13593
2022-04-09 03:18:16,099 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 5.53 GiB (GPU 0; 31.75 GiB total capacity; 26.53 GiB already allocated; 1.12 GiB free; 29.28 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:18:16,099 INFO [decode.py:743] num_arcs before pruning: 350609
2022-04-09 03:18:16,100 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:18:16,105 INFO [decode.py:757] num_arcs after pruning: 9262
2022-04-09 03:19:57,230 INFO [decode_test.py:497] batch 4700/?, cuts processed until now is 13858
2022-04-09 03:20:19,775 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.87 GiB (GPU 0; 31.75 GiB total capacity; 25.78 GiB already allocated; 1.12 GiB free; 29.28 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:20:19,775 INFO [decode.py:743] num_arcs before pruning: 375071
2022-04-09 03:20:19,775 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:20:19,785 INFO [decode.py:757] num_arcs after pruning: 6365
2022-04-09 03:21:29,481 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.42 GiB already allocated; 1.12 GiB free; 29.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:21:29,481 INFO [decode.py:743] num_arcs before pruning: 872088
2022-04-09 03:21:29,481 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:21:29,492 INFO [decode.py:757] num_arcs after pruning: 10043
2022-04-09 03:22:01,760 INFO [decode_test.py:497] batch 4800/?, cuts processed until now is 14079
2022-04-09 03:24:10,370 INFO [decode_test.py:497] batch 4900/?, cuts processed until now is 14298
2022-04-09 03:26:10,811 INFO [decode_test.py:497] batch 5000/?, cuts processed until now is 14515
2022-04-09 03:27:46,191 INFO [decode.py:736] Caught exception:

    Some bad things happened. Please read the above error messages and stack
    trace. If you are using Python, the following command may be helpful:

      gdb --args python /path/to/your/code.py

    (You can use `gdb` to debug the code. Please consider compiling
    a debug version of k2.).

    If you are unable to fix it, please open an issue at:

      https://github.com/k2-fsa/k2/issues/new
    

2022-04-09 03:27:46,192 INFO [decode.py:743] num_arcs before pruning: 246382
2022-04-09 03:27:46,192 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:27:46,253 INFO [decode.py:757] num_arcs after pruning: 6775
2022-04-09 03:28:15,199 INFO [decode_test.py:497] batch 5100/?, cuts processed until now is 14718
2022-04-09 03:29:19,807 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 6.15 GiB (GPU 0; 31.75 GiB total capacity; 26.67 GiB already allocated; 1.11 GiB free; 29.29 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:29:19,808 INFO [decode.py:743] num_arcs before pruning: 220820
2022-04-09 03:29:19,808 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:29:19,815 INFO [decode.py:757] num_arcs after pruning: 13482
2022-04-09 03:30:16,045 INFO [decode_test.py:497] batch 5200/?, cuts processed until now is 14930
2022-04-09 03:32:12,235 INFO [decode_test.py:497] batch 5300/?, cuts processed until now is 15128
2022-04-09 03:33:06,358 INFO [decode.py:736] Caught exception:

    Some bad things happened. Please read the above error messages and stack
    trace. If you are using Python, the following command may be helpful:

      gdb --args python /path/to/your/code.py

    (You can use `gdb` to debug the code. Please consider compiling
    a debug version of k2.).

    If you are unable to fix it, please open an issue at:

      https://github.com/k2-fsa/k2/issues/new
    

2022-04-09 03:33:06,359 INFO [decode.py:743] num_arcs before pruning: 190203
2022-04-09 03:33:06,359 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:33:06,413 INFO [decode.py:757] num_arcs after pruning: 6202
2022-04-09 03:34:14,862 INFO [decode_test.py:497] batch 5400/?, cuts processed until now is 15327
2022-04-09 03:36:18,973 INFO [decode_test.py:497] batch 5500/?, cuts processed until now is 15531
2022-04-09 03:38:18,633 INFO [decode_test.py:497] batch 5600/?, cuts processed until now is 15724
2022-04-09 03:38:48,490 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.52 GiB already allocated; 3.07 GiB free; 27.32 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:38:48,491 INFO [decode.py:743] num_arcs before pruning: 554330
2022-04-09 03:38:48,491 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:38:48,500 INFO [decode.py:757] num_arcs after pruning: 10730
2022-04-09 03:39:51,281 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.83 GiB (GPU 0; 31.75 GiB total capacity; 25.96 GiB already allocated; 1.31 GiB free; 29.08 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:39:51,281 INFO [decode.py:743] num_arcs before pruning: 160031
2022-04-09 03:39:51,281 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:39:51,288 INFO [decode.py:757] num_arcs after pruning: 4270
2022-04-09 03:40:28,016 INFO [decode_test.py:497] batch 5700/?, cuts processed until now is 15908
2022-04-09 03:40:46,608 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.58 GiB (GPU 0; 31.75 GiB total capacity; 27.28 GiB already allocated; 1.32 GiB free; 29.07 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:40:46,608 INFO [decode.py:743] num_arcs before pruning: 406026
2022-04-09 03:40:46,608 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:40:46,616 INFO [decode.py:757] num_arcs after pruning: 11179
2022-04-09 03:42:16,464 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.29 GiB (GPU 0; 31.75 GiB total capacity; 26.71 GiB already allocated; 1.32 GiB free; 29.07 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:42:16,464 INFO [decode.py:743] num_arcs before pruning: 639824
2022-04-09 03:42:16,464 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:42:16,476 INFO [decode.py:757] num_arcs after pruning: 5520
2022-04-09 03:42:52,683 INFO [decode_test.py:497] batch 5800/?, cuts processed until now is 16094
2022-04-09 03:44:51,754 INFO [decode_test.py:497] batch 5900/?, cuts processed until now is 16289
2022-04-09 03:46:52,121 INFO [decode_test.py:497] batch 6000/?, cuts processed until now is 16488
2022-04-09 03:48:54,739 INFO [decode_test.py:497] batch 6100/?, cuts processed until now is 16661
2022-04-09 03:49:24,829 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 1.84 GiB (GPU 0; 31.75 GiB total capacity; 28.87 GiB already allocated; 409.75 MiB free; 29.99 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:49:24,830 INFO [decode.py:743] num_arcs before pruning: 443401
2022-04-09 03:49:24,830 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:49:24,837 INFO [decode.py:757] num_arcs after pruning: 5211
2022-04-09 03:50:27,492 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.35 GiB already allocated; 2.15 GiB free; 28.24 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:50:27,493 INFO [decode.py:743] num_arcs before pruning: 361598
2022-04-09 03:50:27,493 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:50:27,507 INFO [decode.py:757] num_arcs after pruning: 8660
2022-04-09 03:51:02,856 INFO [decode_test.py:497] batch 6200/?, cuts processed until now is 16828
2022-04-09 03:53:03,912 INFO [decode_test.py:497] batch 6300/?, cuts processed until now is 17002
2022-04-09 03:55:04,964 INFO [decode_test.py:497] batch 6400/?, cuts processed until now is 17181
2022-04-09 03:55:08,345 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.89 GiB (GPU 0; 31.75 GiB total capacity; 26.28 GiB already allocated; 2.16 GiB free; 28.24 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:55:08,345 INFO [decode.py:743] num_arcs before pruning: 867262
2022-04-09 03:55:08,345 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:55:08,356 INFO [decode.py:757] num_arcs after pruning: 6494
2022-04-09 03:56:03,884 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 1.90 GiB (GPU 0; 31.75 GiB total capacity; 28.97 GiB already allocated; 1.16 GiB free; 29.23 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:56:03,885 INFO [decode.py:743] num_arcs before pruning: 233755
2022-04-09 03:56:03,885 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:56:03,910 INFO [decode.py:757] num_arcs after pruning: 5823
2022-04-09 03:57:08,774 INFO [decode_test.py:497] batch 6500/?, cuts processed until now is 17347
2022-04-09 03:59:01,245 INFO [decode_test.py:497] batch 6600/?, cuts processed until now is 17502
2022-04-09 03:59:13,147 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 5.80 GiB (GPU 0; 31.75 GiB total capacity; 26.73 GiB already allocated; 1.17 GiB free; 29.22 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 03:59:13,147 INFO [decode.py:743] num_arcs before pruning: 174004
2022-04-09 03:59:13,147 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 03:59:13,155 INFO [decode.py:757] num_arcs after pruning: 6857
2022-04-09 04:00:59,687 INFO [decode_test.py:497] batch 6700/?, cuts processed until now is 17661
2022-04-09 04:03:01,660 INFO [decode_test.py:497] batch 6800/?, cuts processed until now is 17823
2022-04-09 04:04:55,219 INFO [decode_test.py:497] batch 6900/?, cuts processed until now is 17997
2022-04-09 04:07:05,841 INFO [decode_test.py:497] batch 7000/?, cuts processed until now is 18159
2022-04-09 04:09:04,994 INFO [decode_test.py:497] batch 7100/?, cuts processed until now is 18299
2022-04-09 04:11:07,439 INFO [decode_test.py:497] batch 7200/?, cuts processed until now is 18432
2022-04-09 04:13:18,126 INFO [decode_test.py:497] batch 7300/?, cuts processed until now is 18552
2022-04-09 04:15:23,102 INFO [decode_test.py:497] batch 7400/?, cuts processed until now is 18656
2022-04-09 04:17:49,550 INFO [decode_test.py:497] batch 7500/?, cuts processed until now is 18798
2022-04-09 04:19:16,128 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.34 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:19:16,129 INFO [decode.py:743] num_arcs before pruning: 1155990
2022-04-09 04:19:16,129 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:19:16,143 INFO [decode.py:757] num_arcs after pruning: 9141
2022-04-09 04:20:19,961 INFO [decode_test.py:497] batch 7600/?, cuts processed until now is 18945
2022-04-09 04:22:44,642 INFO [decode_test.py:497] batch 7700/?, cuts processed until now is 19084
2022-04-09 04:23:18,184 INFO [decode.py:841] Caught exception:
CUDA out of memory. Tried to allocate 1.26 GiB (GPU 0; 31.75 GiB total capacity; 27.36 GiB already allocated; 881.75 MiB free; 29.53 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:23:18,184 INFO [decode.py:843] num_paths before decreasing: 1000
2022-04-09 04:23:18,184 INFO [decode.py:852] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:23:18,184 INFO [decode.py:858] num_paths after decreasing: 500
2022-04-09 04:24:52,959 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.53 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:24:52,960 INFO [decode.py:743] num_arcs before pruning: 624026
2022-04-09 04:24:52,960 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:24:52,972 INFO [decode.py:757] num_arcs after pruning: 10008
2022-04-09 04:25:07,718 INFO [decode_test.py:497] batch 7800/?, cuts processed until now is 19232
2022-04-09 04:25:31,876 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.51 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:25:31,876 INFO [decode.py:743] num_arcs before pruning: 688909
2022-04-09 04:25:31,877 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:25:31,887 INFO [decode.py:757] num_arcs after pruning: 8886
2022-04-09 04:25:57,970 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 5.04 GiB (GPU 0; 31.75 GiB total capacity; 25.95 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:25:57,971 INFO [decode.py:743] num_arcs before pruning: 891176
2022-04-09 04:25:57,971 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:25:57,982 INFO [decode.py:757] num_arcs after pruning: 10106
2022-04-09 04:26:19,609 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 2.63 GiB (GPU 0; 31.75 GiB total capacity; 27.60 GiB already allocated; 327.75 MiB free; 30.07 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:26:19,609 INFO [decode.py:743] num_arcs before pruning: 415376
2022-04-09 04:26:19,609 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:26:19,620 INFO [decode.py:757] num_arcs after pruning: 7771
2022-04-09 04:27:33,059 INFO [decode_test.py:497] batch 7900/?, cuts processed until now is 19375
2022-04-09 04:29:43,649 INFO [decode_test.py:497] batch 8000/?, cuts processed until now is 19510
2022-04-09 04:30:20,590 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.65 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:30:20,591 INFO [decode.py:743] num_arcs before pruning: 330767
2022-04-09 04:30:20,591 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:30:20,606 INFO [decode.py:757] num_arcs after pruning: 5820
2022-04-09 04:31:55,818 INFO [decode_test.py:497] batch 8100/?, cuts processed until now is 19643
2022-04-09 04:34:11,720 INFO [decode_test.py:497] batch 8200/?, cuts processed until now is 19776
2022-04-09 04:35:04,147 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 4.49 GiB (GPU 0; 31.75 GiB total capacity; 24.38 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:35:04,147 INFO [decode.py:743] num_arcs before pruning: 533967
2022-04-09 04:35:04,147 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:35:04,157 INFO [decode.py:757] num_arcs after pruning: 3449
2022-04-09 04:36:15,595 INFO [decode.py:736] Caught exception:
CUDA out of memory. Tried to allocate 8.00 GiB (GPU 0; 31.75 GiB total capacity; 19.67 GiB already allocated; 2.12 GiB free; 28.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

2022-04-09 04:36:15,595 INFO [decode.py:743] num_arcs before pruning: 397138
2022-04-09 04:36:15,596 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:36:15,605 INFO [decode.py:757] num_arcs after pruning: 6775
2022-04-09 04:36:31,844 INFO [decode_test.py:497] batch 8300/?, cuts processed until now is 19882
2022-04-09 04:37:04,130 INFO [decode.py:736] Caught exception:

    Some bad things happened. Please read the above error messages and stack
    trace. If you are using Python, the following command may be helpful:

      gdb --args python /path/to/your/code.py

    (You can use `gdb` to debug the code. Please consider compiling
    a debug version of k2.).

    If you are unable to fix it, please open an issue at:

      https://github.com/k2-fsa/k2/issues/new
    

2022-04-09 04:37:04,130 INFO [decode.py:743] num_arcs before pruning: 456591
2022-04-09 04:37:04,130 INFO [decode.py:746] This OOM is not an error. You can ignore it. If your model does not converge well, or --max-duration is too large, or the input sound file is difficult to decode, you will meet this exception.
2022-04-09 04:37:04,180 INFO [decode.py:757] num_arcs after pruning: 5275
2022-04-09 04:57:33,432 INFO [decode_test.py:567] 
For test, WER of different settings are:
ngram_lm_scale_0.3_attention_scale_0.7	10.58	best for test
ngram_lm_scale_0.5_attention_scale_1.3	10.58
ngram_lm_scale_0.3_attention_scale_0.5	10.59
ngram_lm_scale_0.3_attention_scale_0.6	10.59
ngram_lm_scale_0.3_attention_scale_0.9	10.59
ngram_lm_scale_0.3_attention_scale_1.0	10.59
ngram_lm_scale_0.3_attention_scale_1.1	10.59
ngram_lm_scale_0.3_attention_scale_1.2	10.59
ngram_lm_scale_0.3_attention_scale_1.3	10.59
ngram_lm_scale_0.5_attention_scale_1.0	10.59
ngram_lm_scale_0.5_attention_scale_1.1	10.59
ngram_lm_scale_0.5_attention_scale_1.2	10.59
ngram_lm_scale_0.5_attention_scale_1.5	10.59
ngram_lm_scale_0.5_attention_scale_1.7	10.59
ngram_lm_scale_0.5_attention_scale_1.9	10.59
ngram_lm_scale_0.5_attention_scale_2.0	10.59
ngram_lm_scale_0.5_attention_scale_2.1	10.59
ngram_lm_scale_0.5_attention_scale_2.2	10.59
ngram_lm_scale_0.5_attention_scale_2.3	10.59
ngram_lm_scale_0.6_attention_scale_1.9	10.59
ngram_lm_scale_0.6_attention_scale_2.0	10.59
ngram_lm_scale_0.6_attention_scale_2.1	10.59
ngram_lm_scale_0.6_attention_scale_2.2	10.59
ngram_lm_scale_0.6_attention_scale_2.3	10.59
ngram_lm_scale_0.6_attention_scale_2.5	10.59
ngram_lm_scale_0.3_attention_scale_1.5	10.6
ngram_lm_scale_0.3_attention_scale_1.7	10.6
ngram_lm_scale_0.3_attention_scale_1.9	10.6
ngram_lm_scale_0.3_attention_scale_2.0	10.6
ngram_lm_scale_0.3_attention_scale_2.1	10.6
ngram_lm_scale_0.3_attention_scale_2.2	10.6
ngram_lm_scale_0.3_attention_scale_2.3	10.6
ngram_lm_scale_0.3_attention_scale_2.5	10.6
ngram_lm_scale_0.5_attention_scale_0.9	10.6
ngram_lm_scale_0.5_attention_scale_2.5	10.6
ngram_lm_scale_0.5_attention_scale_3.0	10.6
ngram_lm_scale_0.6_attention_scale_1.3	10.6
ngram_lm_scale_0.6_attention_scale_1.5	10.6
ngram_lm_scale_0.6_attention_scale_1.7	10.6
ngram_lm_scale_0.6_attention_scale_3.0	10.6
ngram_lm_scale_0.3_attention_scale_0.3	10.61
ngram_lm_scale_0.3_attention_scale_3.0	10.61
ngram_lm_scale_0.5_attention_scale_4.0	10.61
ngram_lm_scale_0.5_attention_scale_5.0	10.61
ngram_lm_scale_0.6_attention_scale_1.2	10.61
ngram_lm_scale_0.6_attention_scale_4.0	10.61
ngram_lm_scale_0.6_attention_scale_5.0	10.61
ngram_lm_scale_0.7_attention_scale_1.7	10.61
ngram_lm_scale_0.7_attention_scale_1.9	10.61
ngram_lm_scale_0.7_attention_scale_2.0	10.61
ngram_lm_scale_0.7_attention_scale_2.1	10.61
ngram_lm_scale_0.7_attention_scale_2.2	10.61
ngram_lm_scale_0.7_attention_scale_2.3	10.61
ngram_lm_scale_0.7_attention_scale_2.5	10.61
ngram_lm_scale_0.7_attention_scale_3.0	10.61
ngram_lm_scale_0.7_attention_scale_4.0	10.61
ngram_lm_scale_0.7_attention_scale_5.0	10.61
ngram_lm_scale_0.1_attention_scale_1.1	10.62
ngram_lm_scale_0.3_attention_scale_4.0	10.62
ngram_lm_scale_0.3_attention_scale_5.0	10.62
ngram_lm_scale_0.5_attention_scale_0.7	10.62
ngram_lm_scale_0.6_attention_scale_1.0	10.62
ngram_lm_scale_0.6_attention_scale_1.1	10.62
ngram_lm_scale_0.7_attention_scale_1.5	10.62
ngram_lm_scale_0.9_attention_scale_3.0	10.62
ngram_lm_scale_0.9_attention_scale_4.0	10.62
ngram_lm_scale_0.9_attention_scale_5.0	10.62
ngram_lm_scale_1.0_attention_scale_4.0	10.62
ngram_lm_scale_1.1_attention_scale_5.0	10.62
ngram_lm_scale_0.05_attention_scale_1.1	10.63
ngram_lm_scale_0.05_attention_scale_1.2	10.63
ngram_lm_scale_0.08_attention_scale_0.9	10.63
ngram_lm_scale_0.08_attention_scale_1.0	10.63
ngram_lm_scale_0.08_attention_scale_1.1	10.63
ngram_lm_scale_0.08_attention_scale_1.2	10.63
ngram_lm_scale_0.08_attention_scale_1.3	10.63
ngram_lm_scale_0.08_attention_scale_1.9	10.63
ngram_lm_scale_0.08_attention_scale_2.0	10.63
ngram_lm_scale_0.08_attention_scale_2.1	10.63
ngram_lm_scale_0.08_attention_scale_2.2	10.63
ngram_lm_scale_0.08_attention_scale_2.3	10.63
ngram_lm_scale_0.08_attention_scale_3.0	10.63
ngram_lm_scale_0.1_attention_scale_0.5	10.63
ngram_lm_scale_0.1_attention_scale_0.6	10.63
ngram_lm_scale_0.1_attention_scale_0.7	10.63
ngram_lm_scale_0.1_attention_scale_0.9	10.63
ngram_lm_scale_0.1_attention_scale_1.0	10.63
ngram_lm_scale_0.1_attention_scale_1.2	10.63
ngram_lm_scale_0.1_attention_scale_1.3	10.63
ngram_lm_scale_0.1_attention_scale_1.7	10.63
ngram_lm_scale_0.1_attention_scale_1.9	10.63
ngram_lm_scale_0.1_attention_scale_2.0	10.63
ngram_lm_scale_0.1_attention_scale_2.1	10.63
ngram_lm_scale_0.1_attention_scale_2.2	10.63
ngram_lm_scale_0.1_attention_scale_2.3	10.63
ngram_lm_scale_0.1_attention_scale_2.5	10.63
ngram_lm_scale_0.1_attention_scale_3.0	10.63
ngram_lm_scale_0.1_attention_scale_5.0	10.63
ngram_lm_scale_0.5_attention_scale_0.6	10.63
ngram_lm_scale_0.6_attention_scale_0.9	10.63
ngram_lm_scale_0.9_attention_scale_2.3	10.63
ngram_lm_scale_0.9_attention_scale_2.5	10.63
ngram_lm_scale_1.0_attention_scale_5.0	10.63
ngram_lm_scale_1.2_attention_scale_5.0	10.63
ngram_lm_scale_0.01_attention_scale_0.9	10.64
ngram_lm_scale_0.01_attention_scale_1.0	10.64
ngram_lm_scale_0.01_attention_scale_1.1	10.64
ngram_lm_scale_0.01_attention_scale_1.2	10.64
ngram_lm_scale_0.01_attention_scale_4.0	10.64
ngram_lm_scale_0.01_attention_scale_5.0	10.64
ngram_lm_scale_0.05_attention_scale_0.5	10.64
ngram_lm_scale_0.05_attention_scale_0.6	10.64
ngram_lm_scale_0.05_attention_scale_0.7	10.64
ngram_lm_scale_0.05_attention_scale_0.9	10.64
ngram_lm_scale_0.05_attention_scale_1.0	10.64
ngram_lm_scale_0.05_attention_scale_1.3	10.64
ngram_lm_scale_0.05_attention_scale_1.5	10.64
ngram_lm_scale_0.05_attention_scale_1.7	10.64
ngram_lm_scale_0.05_attention_scale_1.9	10.64
ngram_lm_scale_0.05_attention_scale_2.0	10.64
ngram_lm_scale_0.05_attention_scale_2.1	10.64
ngram_lm_scale_0.05_attention_scale_2.2	10.64
ngram_lm_scale_0.05_attention_scale_2.3	10.64
ngram_lm_scale_0.05_attention_scale_2.5	10.64
ngram_lm_scale_0.05_attention_scale_3.0	10.64
ngram_lm_scale_0.05_attention_scale_4.0	10.64
ngram_lm_scale_0.05_attention_scale_5.0	10.64
ngram_lm_scale_0.08_attention_scale_0.5	10.64
ngram_lm_scale_0.08_attention_scale_0.6	10.64
ngram_lm_scale_0.08_attention_scale_0.7	10.64
ngram_lm_scale_0.08_attention_scale_1.5	10.64
ngram_lm_scale_0.08_attention_scale_1.7	10.64
ngram_lm_scale_0.08_attention_scale_2.5	10.64
ngram_lm_scale_0.08_attention_scale_4.0	10.64
ngram_lm_scale_0.08_attention_scale_5.0	10.64
ngram_lm_scale_0.1_attention_scale_0.3	10.64
ngram_lm_scale_0.1_attention_scale_1.5	10.64
ngram_lm_scale_0.1_attention_scale_4.0	10.64
ngram_lm_scale_0.7_attention_scale_1.3	10.64
ngram_lm_scale_0.9_attention_scale_2.2	10.64
ngram_lm_scale_1.0_attention_scale_3.0	10.64
ngram_lm_scale_1.1_attention_scale_4.0	10.64
ngram_lm_scale_1.3_attention_scale_5.0	10.64
ngram_lm_scale_0.01_attention_scale_0.6	10.65
ngram_lm_scale_0.01_attention_scale_0.7	10.65
ngram_lm_scale_0.01_attention_scale_1.3	10.65
ngram_lm_scale_0.01_attention_scale_1.5	10.65
ngram_lm_scale_0.01_attention_scale_1.7	10.65
ngram_lm_scale_0.01_attention_scale_1.9	10.65
ngram_lm_scale_0.01_attention_scale_2.0	10.65
ngram_lm_scale_0.01_attention_scale_2.1	10.65
ngram_lm_scale_0.01_attention_scale_2.2	10.65
ngram_lm_scale_0.01_attention_scale_2.3	10.65
ngram_lm_scale_0.01_attention_scale_2.5	10.65
ngram_lm_scale_0.01_attention_scale_3.0	10.65
ngram_lm_scale_0.08_attention_scale_0.3	10.65
ngram_lm_scale_0.5_attention_scale_0.5	10.65
ngram_lm_scale_0.6_attention_scale_0.7	10.65
ngram_lm_scale_0.7_attention_scale_1.1	10.65
ngram_lm_scale_0.7_attention_scale_1.2	10.65
ngram_lm_scale_0.9_attention_scale_2.1	10.65
ngram_lm_scale_1.2_attention_scale_4.0	10.65
ngram_lm_scale_0.05_attention_scale_0.3	10.66
ngram_lm_scale_0.7_attention_scale_1.0	10.66
ngram_lm_scale_0.9_attention_scale_1.9	10.66
ngram_lm_scale_0.9_attention_scale_2.0	10.66
ngram_lm_scale_1.0_attention_scale_2.5	10.66
ngram_lm_scale_1.1_attention_scale_3.0	10.66
ngram_lm_scale_0.01_attention_scale_0.5	10.67
ngram_lm_scale_0.1_attention_scale_0.08	10.67
ngram_lm_scale_0.1_attention_scale_0.1	10.67
ngram_lm_scale_0.6_attention_scale_0.6	10.67
ngram_lm_scale_0.9_attention_scale_1.7	10.67
ngram_lm_scale_1.0_attention_scale_2.2	10.67
ngram_lm_scale_1.0_attention_scale_2.3	10.67
ngram_lm_scale_1.3_attention_scale_4.0	10.67
ngram_lm_scale_1.5_attention_scale_5.0	10.67
ngram_lm_scale_0.01_attention_scale_0.3	10.68
ngram_lm_scale_0.08_attention_scale_0.08	10.68
ngram_lm_scale_0.08_attention_scale_0.1	10.68
ngram_lm_scale_0.3_attention_scale_0.08	10.68
ngram_lm_scale_0.3_attention_scale_0.1	10.68
ngram_lm_scale_0.7_attention_scale_0.9	10.68
ngram_lm_scale_1.0_attention_scale_2.0	10.68
ngram_lm_scale_1.0_attention_scale_2.1	10.68
ngram_lm_scale_1.1_attention_scale_2.5	10.68
ngram_lm_scale_1.2_attention_scale_3.0	10.68
ngram_lm_scale_0.1_attention_scale_0.05	10.69
ngram_lm_scale_0.5_attention_scale_0.3	10.69
ngram_lm_scale_0.9_attention_scale_1.5	10.69
ngram_lm_scale_1.0_attention_scale_1.9	10.69
ngram_lm_scale_1.1_attention_scale_2.3	10.69
ngram_lm_scale_0.05_attention_scale_0.1	10.7
ngram_lm_scale_0.08_attention_scale_0.05	10.7
ngram_lm_scale_0.3_attention_scale_0.05	10.7
ngram_lm_scale_0.6_attention_scale_0.5	10.7
ngram_lm_scale_1.1_attention_scale_2.2	10.7
ngram_lm_scale_1.5_attention_scale_4.0	10.7
ngram_lm_scale_1.7_attention_scale_5.0	10.7
ngram_lm_scale_0.05_attention_scale_0.08	10.71
ngram_lm_scale_1.1_attention_scale_2.1	10.71
ngram_lm_scale_1.2_attention_scale_2.5	10.71
ngram_lm_scale_1.3_attention_scale_3.0	10.71
ngram_lm_scale_0.01_attention_scale_0.1	10.72
ngram_lm_scale_0.05_attention_scale_0.05	10.72
ngram_lm_scale_0.08_attention_scale_0.01	10.72
ngram_lm_scale_0.1_attention_scale_0.01	10.72
ngram_lm_scale_0.3_attention_scale_0.01	10.72
ngram_lm_scale_0.7_attention_scale_0.7	10.72
ngram_lm_scale_0.9_attention_scale_1.3	10.72
ngram_lm_scale_1.0_attention_scale_1.7	10.72
ngram_lm_scale_1.1_attention_scale_2.0	10.72
ngram_lm_scale_0.01_attention_scale_0.08	10.73
ngram_lm_scale_0.9_attention_scale_1.2	10.73
ngram_lm_scale_1.1_attention_scale_1.9	10.73
ngram_lm_scale_1.2_attention_scale_2.3	10.73
ngram_lm_scale_1.0_attention_scale_1.5	10.74
ngram_lm_scale_1.2_attention_scale_2.2	10.74
ngram_lm_scale_1.3_attention_scale_2.5	10.74
ngram_lm_scale_1.9_attention_scale_5.0	10.74
ngram_lm_scale_0.01_attention_scale_0.05	10.75
ngram_lm_scale_0.05_attention_scale_0.01	10.75
ngram_lm_scale_0.7_attention_scale_0.6	10.75
ngram_lm_scale_0.9_attention_scale_1.1	10.75
ngram_lm_scale_1.1_attention_scale_1.7	10.75
ngram_lm_scale_1.2_attention_scale_2.1	10.75
ngram_lm_scale_1.7_attention_scale_4.0	10.75
ngram_lm_scale_1.2_attention_scale_2.0	10.76
ngram_lm_scale_1.3_attention_scale_2.3	10.76
ngram_lm_scale_2.0_attention_scale_5.0	10.76
ngram_lm_scale_1.0_attention_scale_1.3	10.77
ngram_lm_scale_1.2_attention_scale_1.9	10.77
ngram_lm_scale_1.5_attention_scale_3.0	10.77
ngram_lm_scale_0.01_attention_scale_0.01	10.78
ngram_lm_scale_0.6_attention_scale_0.3	10.78
ngram_lm_scale_0.7_attention_scale_0.5	10.78
ngram_lm_scale_0.9_attention_scale_1.0	10.78
ngram_lm_scale_2.1_attention_scale_5.0	10.78
ngram_lm_scale_1.1_attention_scale_1.5	10.79
ngram_lm_scale_1.3_attention_scale_2.2	10.79
ngram_lm_scale_0.5_attention_scale_0.1	10.8
ngram_lm_scale_1.0_attention_scale_1.2	10.8
ngram_lm_scale_1.3_attention_scale_2.1	10.8
ngram_lm_scale_1.9_attention_scale_4.0	10.8
ngram_lm_scale_2.2_attention_scale_5.0	10.8
ngram_lm_scale_0.5_attention_scale_0.08	10.81
ngram_lm_scale_0.9_attention_scale_0.9	10.81
ngram_lm_scale_1.2_attention_scale_1.7	10.81
ngram_lm_scale_1.3_attention_scale_2.0	10.81
ngram_lm_scale_1.0_attention_scale_1.1	10.82
ngram_lm_scale_0.5_attention_scale_0.05	10.83
ngram_lm_scale_1.1_attention_scale_1.3	10.83
ngram_lm_scale_1.3_attention_scale_1.9	10.83
ngram_lm_scale_1.5_attention_scale_2.5	10.84
ngram_lm_scale_2.3_attention_scale_5.0	10.84
ngram_lm_scale_1.0_attention_scale_1.0	10.85
ngram_lm_scale_1.2_attention_scale_1.5	10.85
ngram_lm_scale_2.0_attention_scale_4.0	10.85
ngram_lm_scale_1.1_attention_scale_1.2	10.86
ngram_lm_scale_1.7_attention_scale_3.0	10.86
ngram_lm_scale_0.5_attention_scale_0.01	10.87
ngram_lm_scale_1.5_attention_scale_2.3	10.87
ngram_lm_scale_0.7_attention_scale_0.3	10.88
ngram_lm_scale_0.9_attention_scale_0.7	10.88
ngram_lm_scale_1.3_attention_scale_1.7	10.88
ngram_lm_scale_1.0_attention_scale_0.9	10.89
ngram_lm_scale_1.5_attention_scale_2.2	10.89
ngram_lm_scale_2.1_attention_scale_4.0	10.89
ngram_lm_scale_1.1_attention_scale_1.1	10.91
ngram_lm_scale_0.6_attention_scale_0.1	10.92
ngram_lm_scale_0.9_attention_scale_0.6	10.92
ngram_lm_scale_1.5_attention_scale_2.1	10.92
ngram_lm_scale_1.2_attention_scale_1.3	10.93
ngram_lm_scale_2.5_attention_scale_5.0	10.93
ngram_lm_scale_0.6_attention_scale_0.08	10.94
ngram_lm_scale_2.2_attention_scale_4.0	10.94
ngram_lm_scale_1.1_attention_scale_1.0	10.95
ngram_lm_scale_1.3_attention_scale_1.5	10.95
ngram_lm_scale_1.5_attention_scale_2.0	10.96
ngram_lm_scale_1.2_attention_scale_1.2	10.97
ngram_lm_scale_1.7_attention_scale_2.5	10.97
ngram_lm_scale_0.6_attention_scale_0.05	10.98
ngram_lm_scale_1.9_attention_scale_3.0	10.98
ngram_lm_scale_1.0_attention_scale_0.7	10.99
ngram_lm_scale_1.5_attention_scale_1.9	10.99
ngram_lm_scale_2.3_attention_scale_4.0	10.99
ngram_lm_scale_0.9_attention_scale_0.5	11.0
ngram_lm_scale_1.1_attention_scale_0.9	11.0
ngram_lm_scale_0.6_attention_scale_0.01	11.02
ngram_lm_scale_1.2_attention_scale_1.1	11.02
ngram_lm_scale_1.7_attention_scale_2.3	11.03
ngram_lm_scale_1.3_attention_scale_1.3	11.05
ngram_lm_scale_2.0_attention_scale_3.0	11.05
ngram_lm_scale_1.7_attention_scale_2.2	11.07
ngram_lm_scale_1.0_attention_scale_0.6	11.08
ngram_lm_scale_1.5_attention_scale_1.7	11.08
ngram_lm_scale_1.2_attention_scale_1.0	11.09
ngram_lm_scale_0.7_attention_scale_0.1	11.1
ngram_lm_scale_1.3_attention_scale_1.2	11.1
ngram_lm_scale_1.7_attention_scale_2.1	11.11
ngram_lm_scale_2.1_attention_scale_3.0	11.12
ngram_lm_scale_2.5_attention_scale_4.0	11.12
ngram_lm_scale_0.7_attention_scale_0.08	11.13
ngram_lm_scale_1.9_attention_scale_2.5	11.13
ngram_lm_scale_1.7_attention_scale_2.0	11.14
ngram_lm_scale_1.2_attention_scale_0.9	11.16
ngram_lm_scale_1.1_attention_scale_0.7	11.17
ngram_lm_scale_1.3_attention_scale_1.1	11.17
ngram_lm_scale_3.0_attention_scale_5.0	11.17
ngram_lm_scale_0.7_attention_scale_0.05	11.18
ngram_lm_scale_1.5_attention_scale_1.5	11.18
ngram_lm_scale_1.0_attention_scale_0.5	11.19
ngram_lm_scale_1.7_attention_scale_1.9	11.2
ngram_lm_scale_2.2_attention_scale_3.0	11.21
ngram_lm_scale_1.9_attention_scale_2.3	11.22
ngram_lm_scale_2.0_attention_scale_2.5	11.23
ngram_lm_scale_0.9_attention_scale_0.3	11.25
ngram_lm_scale_1.3_attention_scale_1.0	11.26
ngram_lm_scale_0.7_attention_scale_0.01	11.27
ngram_lm_scale_1.9_attention_scale_2.2	11.27
ngram_lm_scale_1.1_attention_scale_0.6	11.29
ngram_lm_scale_2.3_attention_scale_3.0	11.31
ngram_lm_scale_1.7_attention_scale_1.7	11.33
ngram_lm_scale_1.5_attention_scale_1.3	11.34
ngram_lm_scale_1.9_attention_scale_2.1	11.34
ngram_lm_scale_2.0_attention_scale_2.3	11.34
ngram_lm_scale_2.1_attention_scale_2.5	11.35
ngram_lm_scale_1.3_attention_scale_0.9	11.36
ngram_lm_scale_1.2_attention_scale_0.7	11.39
ngram_lm_scale_1.9_attention_scale_2.0	11.4
ngram_lm_scale_2.0_attention_scale_2.2	11.4
ngram_lm_scale_1.5_attention_scale_1.2	11.43
ngram_lm_scale_1.1_attention_scale_0.5	11.44
ngram_lm_scale_2.0_attention_scale_2.1	11.47
ngram_lm_scale_2.1_attention_scale_2.3	11.47
ngram_lm_scale_2.2_attention_scale_2.5	11.47
ngram_lm_scale_1.9_attention_scale_1.9	11.48
ngram_lm_scale_1.7_attention_scale_1.5	11.5
ngram_lm_scale_2.5_attention_scale_3.0	11.51
ngram_lm_scale_3.0_attention_scale_4.0	11.51
ngram_lm_scale_1.0_attention_scale_0.3	11.53
ngram_lm_scale_1.2_attention_scale_0.6	11.53
ngram_lm_scale_1.5_attention_scale_1.1	11.54
ngram_lm_scale_2.1_attention_scale_2.2	11.54
ngram_lm_scale_2.0_attention_scale_2.0	11.55
ngram_lm_scale_2.3_attention_scale_2.5	11.59
ngram_lm_scale_2.2_attention_scale_2.3	11.61
ngram_lm_scale_2.1_attention_scale_2.1	11.62
ngram_lm_scale_1.3_attention_scale_0.7	11.63
ngram_lm_scale_2.0_attention_scale_1.9	11.63
ngram_lm_scale_1.9_attention_scale_1.7	11.66
ngram_lm_scale_1.5_attention_scale_1.0	11.67
ngram_lm_scale_2.2_attention_scale_2.2	11.69
ngram_lm_scale_0.9_attention_scale_0.1	11.7
ngram_lm_scale_2.1_attention_scale_2.0	11.71
ngram_lm_scale_1.2_attention_scale_0.5	11.72
ngram_lm_scale_1.7_attention_scale_1.3	11.72
ngram_lm_scale_2.3_attention_scale_2.3	11.75
ngram_lm_scale_0.9_attention_scale_0.08	11.77
ngram_lm_scale_2.2_attention_scale_2.1	11.78
ngram_lm_scale_2.1_attention_scale_1.9	11.82
ngram_lm_scale_1.3_attention_scale_0.6	11.83
ngram_lm_scale_1.5_attention_scale_0.9	11.85
ngram_lm_scale_2.0_attention_scale_1.7	11.85
ngram_lm_scale_2.3_attention_scale_2.2	11.86
ngram_lm_scale_0.9_attention_scale_0.05	11.87
ngram_lm_scale_1.1_attention_scale_0.3	11.87
ngram_lm_scale_1.7_attention_scale_1.2	11.88
ngram_lm_scale_1.9_attention_scale_1.5	11.9
ngram_lm_scale_2.2_attention_scale_2.0	11.9
ngram_lm_scale_2.5_attention_scale_2.5	11.9
ngram_lm_scale_4.0_attention_scale_5.0	11.93
ngram_lm_scale_2.3_attention_scale_2.1	11.97
ngram_lm_scale_0.9_attention_scale_0.01	12.0
ngram_lm_scale_2.2_attention_scale_1.9	12.02
ngram_lm_scale_1.7_attention_scale_1.1	12.05
ngram_lm_scale_1.3_attention_scale_0.5	12.07
ngram_lm_scale_2.1_attention_scale_1.7	12.07
ngram_lm_scale_2.3_attention_scale_2.0	12.09
ngram_lm_scale_1.0_attention_scale_0.1	12.11
ngram_lm_scale_2.5_attention_scale_2.3	12.11
ngram_lm_scale_2.0_attention_scale_1.5	12.14
ngram_lm_scale_1.0_attention_scale_0.08	12.19
ngram_lm_scale_3.0_attention_scale_3.0	12.19
ngram_lm_scale_1.9_attention_scale_1.3	12.22
ngram_lm_scale_1.7_attention_scale_1.0	12.23
ngram_lm_scale_2.3_attention_scale_1.9	12.23
ngram_lm_scale_2.5_attention_scale_2.2	12.23
ngram_lm_scale_1.5_attention_scale_0.7	12.27
ngram_lm_scale_1.2_attention_scale_0.3	12.28
ngram_lm_scale_2.2_attention_scale_1.7	12.3
ngram_lm_scale_1.0_attention_scale_0.05	12.32
ngram_lm_scale_2.5_attention_scale_2.1	12.37
ngram_lm_scale_2.1_attention_scale_1.5	12.39
ngram_lm_scale_1.9_attention_scale_1.2	12.41
ngram_lm_scale_1.7_attention_scale_0.9	12.46
ngram_lm_scale_1.0_attention_scale_0.01	12.49
ngram_lm_scale_2.0_attention_scale_1.3	12.5
ngram_lm_scale_2.5_attention_scale_2.0	12.51
ngram_lm_scale_2.3_attention_scale_1.7	12.54
ngram_lm_scale_1.5_attention_scale_0.6	12.55
ngram_lm_scale_1.1_attention_scale_0.1	12.58
ngram_lm_scale_1.9_attention_scale_1.1	12.62
ngram_lm_scale_2.2_attention_scale_1.5	12.64
ngram_lm_scale_1.1_attention_scale_0.08	12.67
ngram_lm_scale_2.5_attention_scale_1.9	12.67
ngram_lm_scale_4.0_attention_scale_4.0	12.67
ngram_lm_scale_1.3_attention_scale_0.3	12.71
ngram_lm_scale_2.0_attention_scale_1.2	12.71
ngram_lm_scale_2.1_attention_scale_1.3	12.78
ngram_lm_scale_3.0_attention_scale_2.5	12.8
ngram_lm_scale_1.1_attention_scale_0.05	12.81
ngram_lm_scale_1.9_attention_scale_1.0	12.85
ngram_lm_scale_1.5_attention_scale_0.5	12.86
ngram_lm_scale_2.3_attention_scale_1.5	12.91
ngram_lm_scale_2.0_attention_scale_1.1	12.92
ngram_lm_scale_1.7_attention_scale_0.7	12.99
ngram_lm_scale_2.1_attention_scale_1.2	12.99
ngram_lm_scale_5.0_attention_scale_5.0	13.01
ngram_lm_scale_1.1_attention_scale_0.01	13.02
ngram_lm_scale_2.5_attention_scale_1.7	13.02
ngram_lm_scale_2.2_attention_scale_1.3	13.05
ngram_lm_scale_3.0_attention_scale_2.3	13.09
ngram_lm_scale_1.2_attention_scale_0.1	13.1
ngram_lm_scale_1.9_attention_scale_0.9	13.11
ngram_lm_scale_2.0_attention_scale_1.0	13.17
ngram_lm_scale_1.2_attention_scale_0.08	13.2
ngram_lm_scale_2.1_attention_scale_1.1	13.22
ngram_lm_scale_3.0_attention_scale_2.2	13.24
ngram_lm_scale_2.2_attention_scale_1.2	13.28
ngram_lm_scale_1.7_attention_scale_0.6	13.33
ngram_lm_scale_2.3_attention_scale_1.3	13.34
ngram_lm_scale_1.2_attention_scale_0.05	13.36
ngram_lm_scale_3.0_attention_scale_2.1	13.42
ngram_lm_scale_2.5_attention_scale_1.5	13.43
ngram_lm_scale_2.0_attention_scale_0.9	13.48
ngram_lm_scale_2.1_attention_scale_1.0	13.51
ngram_lm_scale_2.2_attention_scale_1.1	13.56
ngram_lm_scale_1.2_attention_scale_0.01	13.6
ngram_lm_scale_2.3_attention_scale_1.2	13.6
ngram_lm_scale_3.0_attention_scale_2.0	13.62
ngram_lm_scale_1.3_attention_scale_0.1	13.65
ngram_lm_scale_1.5_attention_scale_0.3	13.68
ngram_lm_scale_1.7_attention_scale_0.5	13.72
ngram_lm_scale_1.3_attention_scale_0.08	13.76
ngram_lm_scale_1.9_attention_scale_0.7	13.78
ngram_lm_scale_3.0_attention_scale_1.9	13.81
ngram_lm_scale_2.1_attention_scale_0.9	13.82
ngram_lm_scale_2.2_attention_scale_1.0	13.85
ngram_lm_scale_4.0_attention_scale_3.0	13.85
ngram_lm_scale_2.3_attention_scale_1.1	13.89
ngram_lm_scale_1.3_attention_scale_0.05	13.94
ngram_lm_scale_2.5_attention_scale_1.3	13.94
ngram_lm_scale_5.0_attention_scale_4.0	13.97
ngram_lm_scale_1.9_attention_scale_0.6	14.15
ngram_lm_scale_2.0_attention_scale_0.7	14.16
ngram_lm_scale_2.2_attention_scale_0.9	14.17
ngram_lm_scale_2.3_attention_scale_1.0	14.19
ngram_lm_scale_1.3_attention_scale_0.01	14.2
ngram_lm_scale_2.5_attention_scale_1.2	14.2
ngram_lm_scale_3.0_attention_scale_1.7	14.26
ngram_lm_scale_2.5_attention_scale_1.1	14.48
ngram_lm_scale_2.3_attention_scale_0.9	14.5
ngram_lm_scale_2.1_attention_scale_0.7	14.53
ngram_lm_scale_2.0_attention_scale_0.6	14.54
ngram_lm_scale_1.9_attention_scale_0.5	14.57
ngram_lm_scale_4.0_attention_scale_2.5	14.63
ngram_lm_scale_1.7_attention_scale_0.3	14.64
ngram_lm_scale_3.0_attention_scale_1.5	14.71
ngram_lm_scale_1.5_attention_scale_0.1	14.75
ngram_lm_scale_2.5_attention_scale_1.0	14.79
ngram_lm_scale_2.2_attention_scale_0.7	14.86
ngram_lm_scale_1.5_attention_scale_0.08	14.87
ngram_lm_scale_2.1_attention_scale_0.6	14.91
ngram_lm_scale_2.0_attention_scale_0.5	14.95
ngram_lm_scale_4.0_attention_scale_2.3	14.98
ngram_lm_scale_1.5_attention_scale_0.05	15.05
ngram_lm_scale_2.5_attention_scale_0.9	15.12
ngram_lm_scale_4.0_attention_scale_2.2	15.17
ngram_lm_scale_2.3_attention_scale_0.7	15.21
ngram_lm_scale_3.0_attention_scale_1.3	15.22
ngram_lm_scale_2.2_attention_scale_0.6	15.27
ngram_lm_scale_1.5_attention_scale_0.01	15.3
ngram_lm_scale_5.0_attention_scale_3.0	15.32
ngram_lm_scale_2.1_attention_scale_0.5	15.33
ngram_lm_scale_4.0_attention_scale_2.1	15.37
ngram_lm_scale_1.9_attention_scale_0.3	15.5
ngram_lm_scale_3.0_attention_scale_1.2	15.51
ngram_lm_scale_4.0_attention_scale_2.0	15.57
ngram_lm_scale_2.3_attention_scale_0.6	15.61
ngram_lm_scale_2.2_attention_scale_0.5	15.68
ngram_lm_scale_1.7_attention_scale_0.1	15.72
ngram_lm_scale_4.0_attention_scale_1.9	15.79
ngram_lm_scale_3.0_attention_scale_1.1	15.82
ngram_lm_scale_1.7_attention_scale_0.08	15.83
ngram_lm_scale_2.5_attention_scale_0.7	15.85
ngram_lm_scale_2.0_attention_scale_0.3	15.87
ngram_lm_scale_2.3_attention_scale_0.5	16.0
ngram_lm_scale_1.7_attention_scale_0.05	16.01
ngram_lm_scale_3.0_attention_scale_1.0	16.11
ngram_lm_scale_5.0_attention_scale_2.5	16.12
ngram_lm_scale_2.5_attention_scale_0.6	16.19
ngram_lm_scale_2.1_attention_scale_0.3	16.2
ngram_lm_scale_4.0_attention_scale_1.7	16.22
ngram_lm_scale_1.7_attention_scale_0.01	16.23
ngram_lm_scale_3.0_attention_scale_0.9	16.4
ngram_lm_scale_5.0_attention_scale_2.3	16.44
ngram_lm_scale_1.9_attention_scale_0.1	16.5
ngram_lm_scale_2.2_attention_scale_0.3	16.53
ngram_lm_scale_2.5_attention_scale_0.5	16.54
ngram_lm_scale_1.9_attention_scale_0.08	16.6
ngram_lm_scale_5.0_attention_scale_2.2	16.6
ngram_lm_scale_4.0_attention_scale_1.5	16.63
ngram_lm_scale_1.9_attention_scale_0.05	16.74
ngram_lm_scale_5.0_attention_scale_2.1	16.77
ngram_lm_scale_2.3_attention_scale_0.3	16.81
ngram_lm_scale_2.0_attention_scale_0.1	16.83
ngram_lm_scale_2.0_attention_scale_0.08	16.92
ngram_lm_scale_5.0_attention_scale_2.0	16.94
ngram_lm_scale_1.9_attention_scale_0.01	16.95
ngram_lm_scale_3.0_attention_scale_0.7	16.96
ngram_lm_scale_2.0_attention_scale_0.05	17.05
ngram_lm_scale_4.0_attention_scale_1.3	17.05
ngram_lm_scale_2.1_attention_scale_0.1	17.11
ngram_lm_scale_5.0_attention_scale_1.9	17.11
ngram_lm_scale_2.1_attention_scale_0.08	17.21
ngram_lm_scale_2.0_attention_scale_0.01	17.24
ngram_lm_scale_3.0_attention_scale_0.6	17.26
ngram_lm_scale_4.0_attention_scale_1.2	17.27
ngram_lm_scale_2.5_attention_scale_0.3	17.28
ngram_lm_scale_2.1_attention_scale_0.05	17.34
ngram_lm_scale_2.2_attention_scale_0.1	17.38
ngram_lm_scale_5.0_attention_scale_1.7	17.44
ngram_lm_scale_2.2_attention_scale_0.08	17.46
ngram_lm_scale_4.0_attention_scale_1.1	17.5
ngram_lm_scale_2.1_attention_scale_0.01	17.52
ngram_lm_scale_3.0_attention_scale_0.5	17.57
ngram_lm_scale_2.2_attention_scale_0.05	17.59
ngram_lm_scale_2.3_attention_scale_0.1	17.62
ngram_lm_scale_2.3_attention_scale_0.08	17.7
ngram_lm_scale_4.0_attention_scale_1.0	17.72
ngram_lm_scale_2.2_attention_scale_0.01	17.76
ngram_lm_scale_5.0_attention_scale_1.5	17.8
ngram_lm_scale_2.3_attention_scale_0.05	17.82
ngram_lm_scale_4.0_attention_scale_0.9	17.94
ngram_lm_scale_2.3_attention_scale_0.01	17.98
ngram_lm_scale_2.5_attention_scale_0.1	18.03
ngram_lm_scale_2.5_attention_scale_0.08	18.1
ngram_lm_scale_5.0_attention_scale_1.3	18.12
ngram_lm_scale_3.0_attention_scale_0.3	18.17
ngram_lm_scale_2.5_attention_scale_0.05	18.2
ngram_lm_scale_5.0_attention_scale_1.2	18.29
ngram_lm_scale_2.5_attention_scale_0.01	18.33
ngram_lm_scale_4.0_attention_scale_0.7	18.36
ngram_lm_scale_5.0_attention_scale_1.1	18.48
ngram_lm_scale_4.0_attention_scale_0.6	18.58
ngram_lm_scale_5.0_attention_scale_1.0	18.65
ngram_lm_scale_3.0_attention_scale_0.1	18.75
ngram_lm_scale_4.0_attention_scale_0.5	18.79
ngram_lm_scale_3.0_attention_scale_0.08	18.81
ngram_lm_scale_5.0_attention_scale_0.9	18.81
ngram_lm_scale_3.0_attention_scale_0.05	18.89
ngram_lm_scale_3.0_attention_scale_0.01	18.99
ngram_lm_scale_5.0_attention_scale_0.7	19.11
ngram_lm_scale_4.0_attention_scale_0.3	19.18
ngram_lm_scale_5.0_attention_scale_0.6	19.25
ngram_lm_scale_5.0_attention_scale_0.5	19.41
ngram_lm_scale_4.0_attention_scale_0.1	19.57
ngram_lm_scale_4.0_attention_scale_0.08	19.61
ngram_lm_scale_4.0_attention_scale_0.05	19.67
ngram_lm_scale_5.0_attention_scale_0.3	19.71
ngram_lm_scale_4.0_attention_scale_0.01	19.73
ngram_lm_scale_5.0_attention_scale_0.1	19.99
ngram_lm_scale_5.0_attention_scale_0.08	20.01
ngram_lm_scale_5.0_attention_scale_0.05	20.05
ngram_lm_scale_5.0_attention_scale_0.01	20.11

2022-04-09 04:57:33,455 INFO [decode_test.py:730] Done!