forked from tensorflow/lingvo
-
Notifications
You must be signed in to change notification settings - Fork 0
/
run.log
executable file
·4053 lines (4033 loc) · 426 KB
/
run.log
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
Starting local Bazel server and connecting to it...
Loading:
Loading: 0 packages loaded
Loading: 0 packages loaded
currently loading: lingvo
Analyzing: target //lingvo:trainer (1 packages loaded, 0 targets configured)
Analyzing: target //lingvo:trainer (13 packages loaded, 27 targets configured)
Analyzing: target //lingvo:trainer (26 packages loaded, 86 targets configured)
Analyzing: target //lingvo:trainer (37 packages loaded, 199 targets configured)
Analyzing: target //lingvo:trainer (44 packages loaded, 729 targets configured)
Analyzing: target //lingvo:trainer (46 packages loaded, 3963 targets configured)
INFO: Analyzed target //lingvo:trainer (46 packages loaded, 3963 targets configured).
INFO: Found 1 target...
INFO: Deleting stale sandbox base /root/.cache/bazel/_bazel_root/17eb95f0bc03547f4f1319e61997e114/sandbox
[0 / 10] [Prepa] BazelWorkspaceStatusAction stable-status.txt
[5 / 17] checking cached actions
Target //lingvo:trainer up-to-date:
bazel-bin/lingvo/trainer
INFO: Elapsed time: 24.595s, Critical Path: 0.31s
INFO: 0 processes.
INFO: Build completed successfully, 1 total action
INFO: Running command line: bazel-bin/lingvo/trainer --logtostderr '--model=asr.librispeech.Librispeech960Grapheme' '--mode=sync' '--logdir=/tmp/lingvo/log' '--saver_max_to_keep=2' '--run_locally=gpu'
INFO: Build completed successfully, 1 total action
2020-04-23 01:40:37.156908: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer.so.6'; dlerror: libnvinfer.so.6: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/nvidia/lib:/usr/local/nvidia/lib64
2020-04-23 01:40:37.157298: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer_plugin.so.6'; dlerror: libnvinfer_plugin.so.6: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/nvidia/lib:/usr/local/nvidia/lib64
2020-04-23 01:40:37.157327: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:30] Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
WARNING:absl:Lingvo does not support eager execution yet. Please disable eager execution with tf.compat.v1.disable_eager_execution() or proceed at your own risk.
2020-04-23 01:40:55.115034: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2020-04-23 01:40:55.375363: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 2200000000 Hz
2020-04-23 01:40:55.424845: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x4b51980 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-04-23 01:40:55.424921: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version
2020-04-23 01:40:55.430220: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2020-04-23 01:40:57.667144: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-04-23 01:40:57.668806: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x4b50ce0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-04-23 01:40:57.668855: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla K80, Compute Capability 3.7
2020-04-23 01:40:57.702764: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-04-23 01:40:57.703977: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1555] Found device 0 with properties:
pciBusID: 0000:00:04.0 name: Tesla K80 computeCapability: 3.7
coreClock: 0.8235GHz coreCount: 13 deviceMemorySize: 11.17GiB deviceMemoryBandwidth: 223.96GiB/s
2020-04-23 01:40:57.774654: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1
2020-04-23 01:40:58.527167: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10
2020-04-23 01:40:58.855762: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10
2020-04-23 01:40:59.010093: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10
2020-04-23 01:40:59.691711: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10
2020-04-23 01:40:59.899970: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10
2020-04-23 01:41:01.026282: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-04-23 01:41:01.026619: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-04-23 01:41:01.027938: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-04-23 01:41:01.028878: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1697] Adding visible gpu devices: 0
2020-04-23 01:41:01.029008: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1
2020-04-23 01:41:01.031838: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1096] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-04-23 01:41:01.031888: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] 0
2020-04-23 01:41:01.031906: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] 0: N
2020-04-23 01:41:01.032720: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-04-23 01:41:01.034006: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-04-23 01:41:01.034911: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1241] Created TensorFlow device (/job:local/replica:0/task:0/device:GPU:0 with 10805 MB memory) -> physical GPU (device: 0, name: Tesla K80, pci bus id: 0000:00:04.0, compute capability: 3.7)
2020-04-23 01:41:01.186212: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:300] Initialize GrpcChannelCache for job local -> {0 -> localhost:33471}
2020-04-23 01:41:01.187467: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:390] Started server with target: grpc://localhost:33471
I0423 01:41:01.188215 140604545877824 trainer.py:1535] Job controller start
I0423 01:41:01.216475 140604545877824 model_imports.py:31] Importing asr.librispeech
I0423 01:41:01.218124 140604545877824 model_imports.py:37] Could not import asr.librispeech: No module named 'asr'
I0423 01:41:01.218743 140604545877824 model_imports.py:31] Importing lingvo.tasks.asr.params.librispeech
I0423 01:41:01.219596 140604545877824 model_imports.py:34] Imported lingvo.tasks.asr.params.librispeech
I0423 01:41:01.318367 140604545877824 base_runner.py:57] ============================================================
I0423 01:41:01.331556 140604545877824 base_runner.py:59] allow_implicit_capture : NoneType
I0423 01:41:01.331747 140604545877824 base_runner.py:59] cls : type/lingvo.core.base_model/SingleTaskModel
I0423 01:41:01.331952 140604545877824 base_runner.py:59] cluster.add_summary : NoneType
I0423 01:41:01.332228 140604545877824 base_runner.py:59] cluster.cls : type/lingvo.core.cluster/_Cluster
I0423 01:41:01.332555 140604545877824 base_runner.py:59] cluster.controller.cpus_per_replica : 1
I0423 01:41:01.332768 140604545877824 base_runner.py:59] cluster.controller.devices_per_split : 1
I0423 01:41:01.333076 140604545877824 base_runner.py:59] cluster.controller.gpus_per_replica : 0
I0423 01:41:01.333395 140604545877824 base_runner.py:59] cluster.controller.name : '/job:local'
I0423 01:41:01.333726 140604545877824 base_runner.py:59] cluster.controller.num_tpu_hosts : 0
I0423 01:41:01.334030 140604545877824 base_runner.py:59] cluster.controller.replicas : 1
I0423 01:41:01.334358 140604545877824 base_runner.py:59] cluster.controller.targets : ''
I0423 01:41:01.334605 140604545877824 base_runner.py:59] cluster.controller.tpus_per_replica : 0
I0423 01:41:01.334870 140604545877824 base_runner.py:59] cluster.decoder.cpus_per_replica : 1
I0423 01:41:01.335180 140604545877824 base_runner.py:59] cluster.decoder.devices_per_split : 1
I0423 01:41:01.335484 140604545877824 base_runner.py:59] cluster.decoder.gpus_per_replica : 1
I0423 01:41:01.335798 140604545877824 base_runner.py:59] cluster.decoder.name : '/job:local'
I0423 01:41:01.336102 140604545877824 base_runner.py:59] cluster.decoder.num_tpu_hosts : 0
I0423 01:41:01.336410 140604545877824 base_runner.py:59] cluster.decoder.replicas : 1
I0423 01:41:01.336719 140604545877824 base_runner.py:59] cluster.decoder.targets : ''
I0423 01:41:01.336968 140604545877824 base_runner.py:59] cluster.decoder.tpus_per_replica : 0
I0423 01:41:01.337275 140604545877824 base_runner.py:59] cluster.do_eval : False
I0423 01:41:01.337621 140604545877824 base_runner.py:59] cluster.evaler.cpus_per_replica : 1
I0423 01:41:01.337937 140604545877824 base_runner.py:59] cluster.evaler.devices_per_split : 1
I0423 01:41:01.338252 140604545877824 base_runner.py:59] cluster.evaler.gpus_per_replica : 1
I0423 01:41:01.338583 140604545877824 base_runner.py:59] cluster.evaler.name : '/job:local'
I0423 01:41:01.338891 140604545877824 base_runner.py:59] cluster.evaler.num_tpu_hosts : 0
I0423 01:41:01.339206 140604545877824 base_runner.py:59] cluster.evaler.replicas : 1
I0423 01:41:01.339506 140604545877824 base_runner.py:59] cluster.evaler.targets : ''
I0423 01:41:01.339909 140604545877824 base_runner.py:59] cluster.evaler.tpus_per_replica : 0
I0423 01:41:01.340292 140604545877824 base_runner.py:59] cluster.input.cpus_per_replica : 1
I0423 01:41:01.340711 140604545877824 base_runner.py:59] cluster.input.devices_per_split : 1
I0423 01:41:01.341134 140604545877824 base_runner.py:59] cluster.input.gpus_per_replica : 0
I0423 01:41:01.341530 140604545877824 base_runner.py:59] cluster.input.name : '/job:local'
I0423 01:41:01.341834 140604545877824 base_runner.py:59] cluster.input.num_tpu_hosts : 0
I0423 01:41:01.342308 140604545877824 base_runner.py:59] cluster.input.replicas : 0
I0423 01:41:01.342775 140604545877824 base_runner.py:59] cluster.input.targets : ''
I0423 01:41:01.343276 140604545877824 base_runner.py:59] cluster.input.tpus_per_replica : 0
I0423 01:41:01.343655 140604545877824 base_runner.py:59] cluster.job : 'controller'
I0423 01:41:01.344065 140604545877824 base_runner.py:59] cluster.logdir : ''
I0423 01:41:01.344513 140604545877824 base_runner.py:59] cluster.mode : 'sync'
I0423 01:41:01.344964 140604545877824 base_runner.py:59] cluster.ps.cpus_per_replica : 1
I0423 01:41:01.345229 140604545877824 base_runner.py:59] cluster.ps.devices_per_split : 1
I0423 01:41:01.345487 140604545877824 base_runner.py:59] cluster.ps.gpus_per_replica : 0
I0423 01:41:01.345913 140604545877824 base_runner.py:59] cluster.ps.name : '/job:local'
I0423 01:41:01.346256 140604545877824 base_runner.py:59] cluster.ps.num_tpu_hosts : 0
I0423 01:41:01.346577 140604545877824 base_runner.py:59] cluster.ps.replicas : 1
I0423 01:41:01.346914 140604545877824 base_runner.py:59] cluster.ps.targets : ''
I0423 01:41:01.347233 140604545877824 base_runner.py:59] cluster.ps.tpus_per_replica : 0
I0423 01:41:01.347561 140604545877824 base_runner.py:59] cluster.split_id : 0
I0423 01:41:01.347878 140604545877824 base_runner.py:59] cluster.task : 0
I0423 01:41:01.348219 140604545877824 base_runner.py:59] cluster.worker.cpus_per_replica : 1
I0423 01:41:01.348448 140604545877824 base_runner.py:59] cluster.worker.devices_per_split : 1
I0423 01:41:01.348720 140604545877824 base_runner.py:59] cluster.worker.gpus_per_replica : 1
I0423 01:41:01.348957 140604545877824 base_runner.py:59] cluster.worker.name : '/job:local'
I0423 01:41:01.349243 140604545877824 base_runner.py:59] cluster.worker.num_tpu_hosts : 0
I0423 01:41:01.349517 140604545877824 base_runner.py:59] cluster.worker.replicas : 1
I0423 01:41:01.349757 140604545877824 base_runner.py:59] cluster.worker.targets : ''
I0423 01:41:01.349971 140604545877824 base_runner.py:59] cluster.worker.tpus_per_replica : 0
I0423 01:41:01.350249 140604545877824 base_runner.py:59] dtype : float32
I0423 01:41:01.350520 140604545877824 base_runner.py:59] fprop_dtype : NoneType
I0423 01:41:01.350757 140604545877824 base_runner.py:59] inference_driver_name : NoneType
I0423 01:41:01.351021 140604545877824 base_runner.py:59] input.allow_implicit_capture : NoneType
I0423 01:41:01.351297 140604545877824 base_runner.py:59] input.append_eos_frame : True
I0423 01:41:01.351519 140604545877824 base_runner.py:59] input.bucket_adjust_every_n : 0
I0423 01:41:01.351781 140604545877824 base_runner.py:59] input.bucket_batch_limit : [48, 24, 24, 24, 24, 24, 24, 24]
I0423 01:41:01.352008 140604545877824 base_runner.py:59] input.bucket_upper_bound : [639, 1062, 1275, 1377, 1449, 1506, 1563, 1710]
I0423 01:41:01.352290 140604545877824 base_runner.py:59] input.cls : type/lingvo.tasks.asr.input_generator/AsrInput
I0423 01:41:01.352557 140604545877824 base_runner.py:59] input.dtype : float32
I0423 01:41:01.352819 140604545877824 base_runner.py:59] input.file_buffer_size : 10000
I0423 01:41:01.353035 140604545877824 base_runner.py:59] input.file_buffer_size_in_seconds : 0
I0423 01:41:01.353320 140604545877824 base_runner.py:59] input.file_datasource.allow_implicit_capture : NoneType
I0423 01:41:01.353586 140604545877824 base_runner.py:59] input.file_datasource.cls : type/lingvo.core.datasource/PrefixedDataSource
I0423 01:41:01.353777 140604545877824 base_runner.py:59] input.file_datasource.dtype : float32
I0423 01:41:01.354055 140604545877824 base_runner.py:59] input.file_datasource.file_pattern : 'train/train.tfrecords-*'
I0423 01:41:01.354344 140604545877824 base_runner.py:59] input.file_datasource.file_pattern_prefix : '/tmp/librispeech'
I0423 01:41:01.354676 140604545877824 base_runner.py:59] input.file_datasource.file_type : 'tfrecord'
I0423 01:41:01.354893 140604545877824 base_runner.py:59] input.file_datasource.fprop_dtype : NoneType
I0423 01:41:01.355116 140604545877824 base_runner.py:59] input.file_datasource.inference_driver_name : NoneType
I0423 01:41:01.355393 140604545877824 base_runner.py:59] input.file_datasource.is_inference : NoneType
I0423 01:41:01.355710 140604545877824 base_runner.py:59] input.file_datasource.name : 'datasource'
I0423 01:41:01.355948 140604545877824 base_runner.py:59] input.file_datasource.params_init.method : 'xavier'
I0423 01:41:01.356196 140604545877824 base_runner.py:59] input.file_datasource.params_init.scale : 1.000001
I0423 01:41:01.356421 140604545877824 base_runner.py:59] input.file_datasource.params_init.seed : NoneType
I0423 01:41:01.356750 140604545877824 base_runner.py:59] input.file_datasource.random_seed : NoneType
I0423 01:41:01.357031 140604545877824 base_runner.py:59] input.file_datasource.skip_lp_regularization : NoneType
I0423 01:41:01.357265 140604545877824 base_runner.py:59] input.file_datasource.vn.global_vn : False
I0423 01:41:01.357596 140604545877824 base_runner.py:59] input.file_datasource.vn.per_step_vn : False
I0423 01:41:01.357846 140604545877824 base_runner.py:59] input.file_datasource.vn.scale : NoneType
I0423 01:41:01.358070 140604545877824 base_runner.py:59] input.file_datasource.vn.seed : NoneType
I0423 01:41:01.358325 140604545877824 base_runner.py:59] input.file_parallelism : 16
I0423 01:41:01.358544 140604545877824 base_runner.py:59] input.file_pattern : ''
I0423 01:41:01.358769 140604545877824 base_runner.py:59] input.file_random_seed : 0
I0423 01:41:01.359084 140604545877824 base_runner.py:59] input.flush_every_n : 0
I0423 01:41:01.359345 140604545877824 base_runner.py:59] input.fprop_dtype : NoneType
I0423 01:41:01.359620 140604545877824 base_runner.py:59] input.frame_size : 80
I0423 01:41:01.359909 140604545877824 base_runner.py:59] input.inference_driver_name : NoneType
I0423 01:41:01.360175 140604545877824 base_runner.py:59] input.is_inference : NoneType
I0423 01:41:01.360474 140604545877824 base_runner.py:59] input.name : 'input'
I0423 01:41:01.360699 140604545877824 base_runner.py:59] input.num_batcher_threads : 1
I0423 01:41:01.360954 140604545877824 base_runner.py:59] input.num_partitions : NoneType
I0423 01:41:01.361362 140604545877824 base_runner.py:59] input.num_samples : 281241
I0423 01:41:01.361800 140604545877824 base_runner.py:59] input.pad_to_max_seq_length : False
I0423 01:41:01.362088 140604545877824 base_runner.py:59] input.params_init.method : 'xavier'
I0423 01:41:01.362403 140604545877824 base_runner.py:59] input.params_init.scale : 1.000001
I0423 01:41:01.362719 140604545877824 base_runner.py:59] input.params_init.seed : NoneType
I0423 01:41:01.363004 140604545877824 base_runner.py:59] input.random_seed : NoneType
I0423 01:41:01.363419 140604545877824 base_runner.py:59] input.remote.max_inflights_per_target : 32
I0423 01:41:01.363715 140604545877824 base_runner.py:59] input.remote.shardable_batch : False
I0423 01:41:01.364017 140604545877824 base_runner.py:59] input.repeat_count : -1
I0423 01:41:01.364385 140604545877824 base_runner.py:59] input.require_sequential_order : False
I0423 01:41:01.364626 140604545877824 base_runner.py:59] input.skip_lp_regularization : NoneType
I0423 01:41:01.364897 140604545877824 base_runner.py:59] input.source_max_length : 3000
I0423 01:41:01.365183 140604545877824 base_runner.py:59] input.target_max_length : 620
I0423 01:41:01.365344 140604545877824 base_runner.py:59] input.tokenizer.allow_implicit_capture : NoneType
I0423 01:41:01.365790 140604545877824 base_runner.py:59] input.tokenizer.append_eos : True
I0423 01:41:01.366105 140604545877824 base_runner.py:59] input.tokenizer.cls : type/lingvo.core.tokenizers/AsciiTokenizer
I0423 01:41:01.366455 140604545877824 base_runner.py:59] input.tokenizer.dtype : float32
I0423 01:41:01.366774 140604545877824 base_runner.py:59] input.tokenizer.fprop_dtype : NoneType
I0423 01:41:01.367173 140604545877824 base_runner.py:59] input.tokenizer.inference_driver_name : NoneType
I0423 01:41:01.367486 140604545877824 base_runner.py:59] input.tokenizer.is_inference : NoneType
I0423 01:41:01.367801 140604545877824 base_runner.py:59] input.tokenizer.name : 'tokenizer'
I0423 01:41:01.368108 140604545877824 base_runner.py:59] input.tokenizer.pad_to_max_length : True
I0423 01:41:01.368367 140604545877824 base_runner.py:59] input.tokenizer.params_init.method : 'xavier'
I0423 01:41:01.368637 140604545877824 base_runner.py:59] input.tokenizer.params_init.scale : 1.000001
I0423 01:41:01.368781 140604545877824 base_runner.py:59] input.tokenizer.params_init.seed : NoneType
I0423 01:41:01.369058 140604545877824 base_runner.py:59] input.tokenizer.random_seed : NoneType
I0423 01:41:01.369330 140604545877824 base_runner.py:59] input.tokenizer.skip_lp_regularization : NoneType
I0423 01:41:01.369475 140604545877824 base_runner.py:59] input.tokenizer.target_eos_id : 2
I0423 01:41:01.369821 140604545877824 base_runner.py:59] input.tokenizer.target_sos_id : 1
I0423 01:41:01.370079 140604545877824 base_runner.py:59] input.tokenizer.target_unk_id : 0
I0423 01:41:01.370395 140604545877824 base_runner.py:59] input.tokenizer.target_wb_id : -1
I0423 01:41:01.370658 140604545877824 base_runner.py:59] input.tokenizer.vn.global_vn : False
I0423 01:41:01.370804 140604545877824 base_runner.py:59] input.tokenizer.vn.per_step_vn : False
I0423 01:41:01.371062 140604545877824 base_runner.py:59] input.tokenizer.vn.scale : NoneType
I0423 01:41:01.371205 140604545877824 base_runner.py:59] input.tokenizer.vn.seed : NoneType
I0423 01:41:01.371467 140604545877824 base_runner.py:59] input.tokenizer.vocab_size : 76
I0423 01:41:01.371843 140604545877824 base_runner.py:59] input.tokenizer_dict : {}
I0423 01:41:01.372159 140604545877824 base_runner.py:59] input.tpu_infeed_parallelism : 1
I0423 01:41:01.372424 140604545877824 base_runner.py:59] input.use_chaining : False
I0423 01:41:01.372571 140604545877824 base_runner.py:59] input.use_partitioned_infeed_queue : False
I0423 01:41:01.372868 140604545877824 base_runner.py:59] input.use_per_host_infeed : False
I0423 01:41:01.373166 140604545877824 base_runner.py:59] input.use_within_batch_mixing : False
I0423 01:41:01.373414 140604545877824 base_runner.py:59] input.vn.global_vn : False
I0423 01:41:01.373782 140604545877824 base_runner.py:59] input.vn.per_step_vn : False
I0423 01:41:01.373921 140604545877824 base_runner.py:59] input.vn.scale : NoneType
I0423 01:41:01.374207 140604545877824 base_runner.py:59] input.vn.seed : NoneType
I0423 01:41:01.374466 140604545877824 base_runner.py:59] is_inference : NoneType
I0423 01:41:01.374725 140604545877824 base_runner.py:59] model : 'asr.librispeech.Librispeech960Grapheme@/root/.cache/bazel/_bazel_root/17eb95f0bc03547f4f1319e61997e114/execroot/__main__/bazel-out/k8-opt/bin/lingvo/trainer.runfiles/__main__/lingvo/tasks/asr/params/librispeech.py:161'
I0423 01:41:01.374983 140604545877824 base_runner.py:59] name : ''
I0423 01:41:01.375324 140604545877824 base_runner.py:59] params_init.method : 'xavier'
I0423 01:41:01.375473 140604545877824 base_runner.py:59] params_init.scale : 1.000001
I0423 01:41:01.375705 140604545877824 base_runner.py:59] params_init.seed : NoneType
I0423 01:41:01.375848 140604545877824 base_runner.py:59] random_seed : NoneType
I0423 01:41:01.376091 140604545877824 base_runner.py:59] skip_lp_regularization : NoneType
I0423 01:41:01.376322 140604545877824 base_runner.py:59] task.allow_implicit_capture : NoneType
I0423 01:41:01.376465 140604545877824 base_runner.py:59] task.cls : type/lingvo.tasks.asr.model/AsrModel
I0423 01:41:01.376708 140604545877824 base_runner.py:59] task.decoder.allow_implicit_capture : NoneType
I0423 01:41:01.376888 140604545877824 base_runner.py:59] task.decoder.atten_context_dim : 0
I0423 01:41:01.377127 140604545877824 base_runner.py:59] task.decoder.attention.allow_implicit_capture : NoneType
I0423 01:41:01.377377 140604545877824 base_runner.py:59] task.decoder.attention.atten_dropout_deterministic : False
I0423 01:41:01.377636 140604545877824 base_runner.py:59] task.decoder.attention.atten_dropout_prob : 0.0
I0423 01:41:01.377871 140604545877824 base_runner.py:59] task.decoder.attention.cls : type/lingvo.core.attention/AdditiveAttention
I0423 01:41:01.378017 140604545877824 base_runner.py:59] task.decoder.attention.dtype : float32
I0423 01:41:01.378231 140604545877824 base_runner.py:59] task.decoder.attention.fprop_dtype : NoneType
I0423 01:41:01.378364 140604545877824 base_runner.py:59] task.decoder.attention.hidden_dim : 128
I0423 01:41:01.378587 140604545877824 base_runner.py:59] task.decoder.attention.inference_driver_name : NoneType
I0423 01:41:01.378886 140604545877824 base_runner.py:59] task.decoder.attention.is_inference : NoneType
I0423 01:41:01.379045 140604545877824 base_runner.py:59] task.decoder.attention.name : ''
I0423 01:41:01.379201 140604545877824 base_runner.py:59] task.decoder.attention.packed_input : False
I0423 01:41:01.379467 140604545877824 base_runner.py:59] task.decoder.attention.params_init.method : 'uniform_sqrt_dim'
I0423 01:41:01.379695 140604545877824 base_runner.py:59] task.decoder.attention.params_init.scale : 1.7320508075688772
I0423 01:41:01.379853 140604545877824 base_runner.py:59] task.decoder.attention.params_init.seed : NoneType
I0423 01:41:01.380130 140604545877824 base_runner.py:59] task.decoder.attention.qdomain.default : NoneType
I0423 01:41:01.380304 140604545877824 base_runner.py:59] task.decoder.attention.qdomain.fullyconnected : NoneType
I0423 01:41:01.380437 140604545877824 base_runner.py:59] task.decoder.attention.qdomain.softmax : NoneType
I0423 01:41:01.380661 140604545877824 base_runner.py:59] task.decoder.attention.query_dim : 0
I0423 01:41:01.380901 140604545877824 base_runner.py:59] task.decoder.attention.random_seed : NoneType
I0423 01:41:01.381168 140604545877824 base_runner.py:59] task.decoder.attention.same_batch_size : False
I0423 01:41:01.381429 140604545877824 base_runner.py:59] task.decoder.attention.skip_lp_regularization : NoneType
I0423 01:41:01.381695 140604545877824 base_runner.py:59] task.decoder.attention.source_dim : 0
I0423 01:41:01.381880 140604545877824 base_runner.py:59] task.decoder.attention.vn.global_vn : False
I0423 01:41:01.382128 140604545877824 base_runner.py:59] task.decoder.attention.vn.per_step_vn : False
I0423 01:41:01.382429 140604545877824 base_runner.py:59] task.decoder.attention.vn.scale : NoneType
I0423 01:41:01.382684 140604545877824 base_runner.py:59] task.decoder.attention.vn.seed : NoneType
I0423 01:41:01.382831 140604545877824 base_runner.py:59] task.decoder.attention_plot_font_properties : ''
I0423 01:41:01.382956 140604545877824 base_runner.py:59] task.decoder.beam_search.allow_empty_terminated_hyp : True
I0423 01:41:01.383261 140604545877824 base_runner.py:59] task.decoder.beam_search.allow_implicit_capture : NoneType
I0423 01:41:01.383546 140604545877824 base_runner.py:59] task.decoder.beam_search.batch_major_compute : False
I0423 01:41:01.383786 140604545877824 base_runner.py:59] task.decoder.beam_search.batch_major_state : True
I0423 01:41:01.383926 140604545877824 base_runner.py:59] task.decoder.beam_search.beam_size : 3.0
I0423 01:41:01.384206 140604545877824 base_runner.py:59] task.decoder.beam_search.cls : type/lingvo.core.beam_search_helper/BeamSearchHelper
I0423 01:41:01.384425 140604545877824 base_runner.py:59] task.decoder.beam_search.coverage_penalty : 0.0
I0423 01:41:01.384561 140604545877824 base_runner.py:59] task.decoder.beam_search.dtype : float32
I0423 01:41:01.384699 140604545877824 base_runner.py:59] task.decoder.beam_search.ensure_full_beam : False
I0423 01:41:01.384909 140604545877824 base_runner.py:59] task.decoder.beam_search.force_eos_in_last_step : False
I0423 01:41:01.385058 140604545877824 base_runner.py:59] task.decoder.beam_search.fprop_dtype : NoneType
I0423 01:41:01.385300 140604545877824 base_runner.py:59] task.decoder.beam_search.inference_driver_name : NoneType
I0423 01:41:01.385448 140604545877824 base_runner.py:59] task.decoder.beam_search.is_inference : NoneType
I0423 01:41:01.385702 140604545877824 base_runner.py:59] task.decoder.beam_search.length_normalization : 0.0
I0423 01:41:01.385853 140604545877824 base_runner.py:59] task.decoder.beam_search.local_eos_threshold : -100.0
I0423 01:41:01.386113 140604545877824 base_runner.py:59] task.decoder.beam_search.merge_paths : False
I0423 01:41:01.386351 140604545877824 base_runner.py:59] task.decoder.beam_search.name : 'beam_search'
I0423 01:41:01.386500 140604545877824 base_runner.py:59] task.decoder.beam_search.num_hyps_per_beam : 8
I0423 01:41:01.386718 140604545877824 base_runner.py:59] task.decoder.beam_search.params_init.method : 'xavier'
I0423 01:41:01.386875 140604545877824 base_runner.py:59] task.decoder.beam_search.params_init.scale : 1.000001
I0423 01:41:01.387133 140604545877824 base_runner.py:59] task.decoder.beam_search.params_init.seed : NoneType
I0423 01:41:01.387427 140604545877824 base_runner.py:59] task.decoder.beam_search.random_seed : NoneType
I0423 01:41:01.387562 140604545877824 base_runner.py:59] task.decoder.beam_search.short_seq_limit : 0
I0423 01:41:01.387772 140604545877824 base_runner.py:59] task.decoder.beam_search.skip_lp_regularization : NoneType
I0423 01:41:01.387922 140604545877824 base_runner.py:59] task.decoder.beam_search.target_eoc_id : -1
I0423 01:41:01.388139 140604545877824 base_runner.py:59] task.decoder.beam_search.target_eos_id : 2
I0423 01:41:01.388294 140604545877824 base_runner.py:59] task.decoder.beam_search.target_seq_len : 0
I0423 01:41:01.388545 140604545877824 base_runner.py:59] task.decoder.beam_search.target_seq_length_ratio : 1.0
I0423 01:41:01.388796 140604545877824 base_runner.py:59] task.decoder.beam_search.target_sos_id : 1
I0423 01:41:01.389085 140604545877824 base_runner.py:59] task.decoder.beam_search.valid_eos_max_logit_delta : 5.0
I0423 01:41:01.389335 140604545877824 base_runner.py:59] task.decoder.beam_search.vn.global_vn : False
I0423 01:41:01.389481 140604545877824 base_runner.py:59] task.decoder.beam_search.vn.per_step_vn : False
I0423 01:41:01.389735 140604545877824 base_runner.py:59] task.decoder.beam_search.vn.scale : NoneType
I0423 01:41:01.389913 140604545877824 base_runner.py:59] task.decoder.beam_search.vn.seed : NoneType
I0423 01:41:01.390165 140604545877824 base_runner.py:59] task.decoder.bias_only_if_consistent : True
I0423 01:41:01.390412 140604545877824 base_runner.py:59] task.decoder.cls : type/lingvo.tasks.asr.decoder/AsrDecoder
I0423 01:41:01.390557 140604545877824 base_runner.py:59] task.decoder.contextualizer.allow_implicit_capture : NoneType
I0423 01:41:01.390788 140604545877824 base_runner.py:59] task.decoder.contextualizer.cls : type/lingvo.tasks.asr.contextualizer_base/NullContextualizer
I0423 01:41:01.390937 140604545877824 base_runner.py:59] task.decoder.contextualizer.dtype : float32
I0423 01:41:01.391061 140604545877824 base_runner.py:59] task.decoder.contextualizer.fprop_dtype : NoneType
I0423 01:41:01.391346 140604545877824 base_runner.py:59] task.decoder.contextualizer.inference_driver_name : NoneType
I0423 01:41:01.391607 140604545877824 base_runner.py:59] task.decoder.contextualizer.is_inference : NoneType
I0423 01:41:01.391828 140604545877824 base_runner.py:59] task.decoder.contextualizer.name : ''
I0423 01:41:01.392041 140604545877824 base_runner.py:59] task.decoder.contextualizer.params_init.method : 'xavier'
I0423 01:41:01.392185 140604545877824 base_runner.py:59] task.decoder.contextualizer.params_init.scale : 1.000001
I0423 01:41:01.392417 140604545877824 base_runner.py:59] task.decoder.contextualizer.params_init.seed : NoneType
I0423 01:41:01.392670 140604545877824 base_runner.py:59] task.decoder.contextualizer.random_seed : NoneType
I0423 01:41:01.392823 140604545877824 base_runner.py:59] task.decoder.contextualizer.skip_lp_regularization : NoneType
I0423 01:41:01.393032 140604545877824 base_runner.py:59] task.decoder.contextualizer.vn.global_vn : False
I0423 01:41:01.393190 140604545877824 base_runner.py:59] task.decoder.contextualizer.vn.per_step_vn : False
I0423 01:41:01.393424 140604545877824 base_runner.py:59] task.decoder.contextualizer.vn.scale : NoneType
I0423 01:41:01.393588 140604545877824 base_runner.py:59] task.decoder.contextualizer.vn.seed : NoneType
I0423 01:41:01.393732 140604545877824 base_runner.py:59] task.decoder.dropout_prob : 0.0
I0423 01:41:01.394039 140604545877824 base_runner.py:59] task.decoder.dtype : float32
I0423 01:41:01.394279 140604545877824 base_runner.py:59] task.decoder.emb.allow_implicit_capture : NoneType
I0423 01:41:01.394524 140604545877824 base_runner.py:59] task.decoder.emb.cls : type/lingvo.core.layers/EmbeddingLayer
I0423 01:41:01.394670 140604545877824 base_runner.py:59] task.decoder.emb.dtype : float32
I0423 01:41:01.394887 140604545877824 base_runner.py:59] task.decoder.emb.embedding_dim : 0
I0423 01:41:01.395036 140604545877824 base_runner.py:59] task.decoder.emb.fprop_dtype : NoneType
I0423 01:41:01.395319 140604545877824 base_runner.py:59] task.decoder.emb.inference_driver_name : NoneType
I0423 01:41:01.395457 140604545877824 base_runner.py:59] task.decoder.emb.is_inference : NoneType
I0423 01:41:01.395691 140604545877824 base_runner.py:59] task.decoder.emb.max_num_shards : 1
I0423 01:41:01.395835 140604545877824 base_runner.py:59] task.decoder.emb.name : ''
I0423 01:41:01.396106 140604545877824 base_runner.py:59] task.decoder.emb.on_ps : True
I0423 01:41:01.396347 140604545877824 base_runner.py:59] task.decoder.emb.params_init.method : 'uniform'
I0423 01:41:01.396489 140604545877824 base_runner.py:59] task.decoder.emb.params_init.scale : 1.0
I0423 01:41:01.396711 140604545877824 base_runner.py:59] task.decoder.emb.params_init.seed : NoneType
I0423 01:41:01.396960 140604545877824 base_runner.py:59] task.decoder.emb.random_seed : NoneType
I0423 01:41:01.397108 140604545877824 base_runner.py:59] task.decoder.emb.scale_sqrt_depth : False
I0423 01:41:01.397366 140604545877824 base_runner.py:59] task.decoder.emb.skip_lp_regularization : NoneType
I0423 01:41:01.397631 140604545877824 base_runner.py:59] task.decoder.emb.vn.global_vn : False
I0423 01:41:01.397835 140604545877824 base_runner.py:59] task.decoder.emb.vn.per_step_vn : False
I0423 01:41:01.398055 140604545877824 base_runner.py:59] task.decoder.emb.vn.scale : NoneType
I0423 01:41:01.398324 140604545877824 base_runner.py:59] task.decoder.emb.vn.seed : NoneType
I0423 01:41:01.398460 140604545877824 base_runner.py:59] task.decoder.emb.vocab_size : 76
I0423 01:41:01.398620 140604545877824 base_runner.py:59] task.decoder.emb_dim : 76
I0423 01:41:01.398744 140604545877824 base_runner.py:59] task.decoder.focal_loss_alpha : NoneType
I0423 01:41:01.398874 140604545877824 base_runner.py:59] task.decoder.focal_loss_gamma : NoneType
I0423 01:41:01.399009 140604545877824 base_runner.py:59] task.decoder.fprop_dtype : NoneType
I0423 01:41:01.399160 140604545877824 base_runner.py:59] task.decoder.fusion.allow_implicit_capture : NoneType
I0423 01:41:01.399285 140604545877824 base_runner.py:59] task.decoder.fusion.base_model_logits_dim : NoneType
I0423 01:41:01.399430 140604545877824 base_runner.py:59] task.decoder.fusion.cls : type/lingvo.tasks.asr.fusion/NullFusion
I0423 01:41:01.399566 140604545877824 base_runner.py:59] task.decoder.fusion.dtype : float32
I0423 01:41:01.399695 140604545877824 base_runner.py:59] task.decoder.fusion.fprop_dtype : NoneType
I0423 01:41:01.399866 140604545877824 base_runner.py:59] task.decoder.fusion.inference_driver_name : NoneType
I0423 01:41:01.400084 140604545877824 base_runner.py:59] task.decoder.fusion.is_inference : NoneType
I0423 01:41:01.400395 140604545877824 base_runner.py:59] task.decoder.fusion.lm.allow_implicit_capture : NoneType
I0423 01:41:01.400540 140604545877824 base_runner.py:59] task.decoder.fusion.lm.cls : type/lingvo.tasks.lm.layers/NullLm
I0423 01:41:01.400669 140604545877824 base_runner.py:59] task.decoder.fusion.lm.dtype : float32
I0423 01:41:01.400802 140604545877824 base_runner.py:59] task.decoder.fusion.lm.fprop_dtype : NoneType
I0423 01:41:01.400932 140604545877824 base_runner.py:59] task.decoder.fusion.lm.inference_driver_name : NoneType
I0423 01:41:01.401063 140604545877824 base_runner.py:59] task.decoder.fusion.lm.is_inference : NoneType
I0423 01:41:01.401266 140604545877824 base_runner.py:59] task.decoder.fusion.lm.name : ''
I0423 01:41:01.401422 140604545877824 base_runner.py:59] task.decoder.fusion.lm.params_init.method : 'xavier'
I0423 01:41:01.401571 140604545877824 base_runner.py:59] task.decoder.fusion.lm.params_init.scale : 1.000001
I0423 01:41:01.401708 140604545877824 base_runner.py:59] task.decoder.fusion.lm.params_init.seed : NoneType
I0423 01:41:01.401873 140604545877824 base_runner.py:59] task.decoder.fusion.lm.random_seed : NoneType
I0423 01:41:01.402033 140604545877824 base_runner.py:59] task.decoder.fusion.lm.skip_lp_regularization : NoneType
I0423 01:41:01.402180 140604545877824 base_runner.py:59] task.decoder.fusion.lm.vn.global_vn : False
I0423 01:41:01.402309 140604545877824 base_runner.py:59] task.decoder.fusion.lm.vn.per_step_vn : False
I0423 01:41:01.402441 140604545877824 base_runner.py:59] task.decoder.fusion.lm.vn.scale : NoneType
I0423 01:41:01.402578 140604545877824 base_runner.py:59] task.decoder.fusion.lm.vn.seed : NoneType
I0423 01:41:01.402719 140604545877824 base_runner.py:59] task.decoder.fusion.lm.vocab_size : 96
I0423 01:41:01.402851 140604545877824 base_runner.py:59] task.decoder.fusion.name : ''
I0423 01:41:01.402982 140604545877824 base_runner.py:59] task.decoder.fusion.params_init.method : 'xavier'
I0423 01:41:01.403112 140604545877824 base_runner.py:59] task.decoder.fusion.params_init.scale : 1.000001
I0423 01:41:01.403346 140604545877824 base_runner.py:59] task.decoder.fusion.params_init.seed : NoneType
I0423 01:41:01.403516 140604545877824 base_runner.py:59] task.decoder.fusion.random_seed : NoneType
I0423 01:41:01.403674 140604545877824 base_runner.py:59] task.decoder.fusion.skip_lp_regularization : NoneType
I0423 01:41:01.403815 140604545877824 base_runner.py:59] task.decoder.fusion.vn.global_vn : False
I0423 01:41:01.403947 140604545877824 base_runner.py:59] task.decoder.fusion.vn.per_step_vn : False
I0423 01:41:01.404084 140604545877824 base_runner.py:59] task.decoder.fusion.vn.scale : NoneType
I0423 01:41:01.404224 140604545877824 base_runner.py:59] task.decoder.fusion.vn.seed : NoneType
I0423 01:41:01.404361 140604545877824 base_runner.py:59] task.decoder.greedy_search.allow_implicit_capture : NoneType
I0423 01:41:01.404491 140604545877824 base_runner.py:59] task.decoder.greedy_search.cls : type/lingvo.core.beam_search_helper/GreedySearchHelper
I0423 01:41:01.404628 140604545877824 base_runner.py:59] task.decoder.greedy_search.dtype : float32
I0423 01:41:01.404817 140604545877824 base_runner.py:59] task.decoder.greedy_search.fprop_dtype : NoneType
I0423 01:41:01.404958 140604545877824 base_runner.py:59] task.decoder.greedy_search.inference_driver_name : NoneType
I0423 01:41:01.405094 140604545877824 base_runner.py:59] task.decoder.greedy_search.is_inference : NoneType
I0423 01:41:01.405241 140604545877824 base_runner.py:59] task.decoder.greedy_search.name : 'greedy_search'
I0423 01:41:01.405373 140604545877824 base_runner.py:59] task.decoder.greedy_search.params_init.method : 'xavier'
I0423 01:41:01.405518 140604545877824 base_runner.py:59] task.decoder.greedy_search.params_init.scale : 1.000001
I0423 01:41:01.406041 140604545877824 base_runner.py:59] task.decoder.greedy_search.params_init.seed : NoneType
I0423 01:41:01.406328 140604545877824 base_runner.py:59] task.decoder.greedy_search.random_seed : NoneType
I0423 01:41:01.406600 140604545877824 base_runner.py:59] task.decoder.greedy_search.skip_lp_regularization : NoneType
I0423 01:41:01.406742 140604545877824 base_runner.py:59] task.decoder.greedy_search.target_eos_id : 2
I0423 01:41:01.406982 140604545877824 base_runner.py:59] task.decoder.greedy_search.target_seq_len : 0
I0423 01:41:01.407259 140604545877824 base_runner.py:59] task.decoder.greedy_search.target_sos_id : 1
I0423 01:41:01.407410 140604545877824 base_runner.py:59] task.decoder.greedy_search.vn.global_vn : False
I0423 01:41:01.407623 140604545877824 base_runner.py:59] task.decoder.greedy_search.vn.per_step_vn : False
I0423 01:41:01.407773 140604545877824 base_runner.py:59] task.decoder.greedy_search.vn.scale : NoneType
I0423 01:41:01.408040 140604545877824 base_runner.py:59] task.decoder.greedy_search.vn.seed : NoneType
I0423 01:41:01.408301 140604545877824 base_runner.py:59] task.decoder.inference_driver_name : NoneType
I0423 01:41:01.408447 140604545877824 base_runner.py:59] task.decoder.is_inference : NoneType
I0423 01:41:01.408717 140604545877824 base_runner.py:59] task.decoder.label_smoothing : NoneType
I0423 01:41:01.408941 140604545877824 base_runner.py:59] task.decoder.logit_types : {'logits': 1.0}
I0423 01:41:01.409076 140604545877824 base_runner.py:59] task.decoder.min_ground_truth_prob : 1.0
I0423 01:41:01.409308 140604545877824 base_runner.py:59] task.decoder.min_prob_step : 1000000.0
I0423 01:41:01.409450 140604545877824 base_runner.py:59] task.decoder.name : ''
I0423 01:41:01.409706 140604545877824 base_runner.py:59] task.decoder.packed_input : False
I0423 01:41:01.409845 140604545877824 base_runner.py:59] task.decoder.parallel_iterations : 30
I0423 01:41:01.410125 140604545877824 base_runner.py:59] task.decoder.params_init.method : 'xavier'
I0423 01:41:01.410381 140604545877824 base_runner.py:59] task.decoder.params_init.scale : 1.000001
I0423 01:41:01.410527 140604545877824 base_runner.py:59] task.decoder.params_init.seed : NoneType
I0423 01:41:01.410768 140604545877824 base_runner.py:59] task.decoder.per_token_avg_loss : True
I0423 01:41:01.411021 140604545877824 base_runner.py:59] task.decoder.prob_decay_start_step : 10000.0
I0423 01:41:01.411185 140604545877824 base_runner.py:59] task.decoder.random_seed : NoneType
I0423 01:41:01.411439 140604545877824 base_runner.py:59] task.decoder.residual_start : 0
I0423 01:41:01.411655 140604545877824 base_runner.py:59] task.decoder.rnn_cell_dim : 1024
I0423 01:41:01.411779 140604545877824 base_runner.py:59] task.decoder.rnn_cell_hidden_dim : 0
I0423 01:41:01.411988 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.412131 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.apply_pruning : False
I0423 01:41:01.412386 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.apply_pruning_to_projection : False
I0423 01:41:01.412535 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.bias_init.method : 'constant'
I0423 01:41:01.412791 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.bias_init.scale : 0.0
I0423 01:41:01.412932 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.bias_init.seed : 0
I0423 01:41:01.413200 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.cell_value_cap : 10.0
I0423 01:41:01.413457 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.cls : type/lingvo.core.rnn_cell/LSTMCellSimple
I0423 01:41:01.413628 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.couple_input_forget_gates : False
I0423 01:41:01.413870 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.dtype : float32
I0423 01:41:01.414023 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.enable_lstm_bias : True
I0423 01:41:01.414267 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.forget_gate_bias : 0.0
I0423 01:41:01.414488 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.fprop_dtype : NoneType
I0423 01:41:01.414625 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.gradient_pruning : False
I0423 01:41:01.414764 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.inference_driver_name : NoneType
I0423 01:41:01.414988 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.inputs_arity : 1
I0423 01:41:01.415231 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.is_inference : NoneType
I0423 01:41:01.415378 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.name : ''
I0423 01:41:01.415624 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.num_hidden_nodes : 0
I0423 01:41:01.415770 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.num_input_nodes : 0
I0423 01:41:01.416001 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.num_output_nodes : 0
I0423 01:41:01.416142 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.output_nonlinearity : True
I0423 01:41:01.416285 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.params_init.method : 'uniform'
I0423 01:41:01.416422 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.params_init.scale : 0.1
I0423 01:41:01.416551 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.params_init.seed : NoneType
I0423 01:41:01.416682 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.qdomain.c_state : NoneType
I0423 01:41:01.416814 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.qdomain.default : NoneType
I0423 01:41:01.416949 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.qdomain.fullyconnected : NoneType
I0423 01:41:01.417079 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.qdomain.m_state : NoneType
I0423 01:41:01.417217 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.qdomain.weight : NoneType
I0423 01:41:01.417346 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.random_seed : NoneType
I0423 01:41:01.417531 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.reset_cell_state : False
I0423 01:41:01.417690 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.417825 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.vn.global_vn : False
I0423 01:41:01.417955 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.vn.per_step_vn : False
I0423 01:41:01.418128 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.vn.scale : NoneType
I0423 01:41:01.418277 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.vn.seed : NoneType
I0423 01:41:01.418409 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.zero_state_init_params.method : 'zeros'
I0423 01:41:01.418547 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.zero_state_init_params.seed : NoneType
I0423 01:41:01.418683 140604545877824 base_runner.py:59] task.decoder.rnn_cell_tpl.zo_prob : 0.0
I0423 01:41:01.418826 140604545877824 base_runner.py:59] task.decoder.rnn_layers : 2
I0423 01:41:01.418955 140604545877824 base_runner.py:59] task.decoder.skip_lp_regularization : NoneType
I0423 01:41:01.419090 140604545877824 base_runner.py:59] task.decoder.softmax.allow_implicit_capture : NoneType
I0423 01:41:01.419246 140604545877824 base_runner.py:59] task.decoder.softmax.apply_pruning : False
I0423 01:41:01.419368 140604545877824 base_runner.py:59] task.decoder.softmax.chunk_size : 0
I0423 01:41:01.419513 140604545877824 base_runner.py:59] task.decoder.softmax.cls : type/lingvo.core.layers/SimpleFullSoftmax
I0423 01:41:01.419651 140604545877824 base_runner.py:59] task.decoder.softmax.dtype : float32
I0423 01:41:01.419790 140604545877824 base_runner.py:59] task.decoder.softmax.fprop_dtype : NoneType
I0423 01:41:01.419913 140604545877824 base_runner.py:59] task.decoder.softmax.inference_driver_name : NoneType
I0423 01:41:01.420042 140604545877824 base_runner.py:59] task.decoder.softmax.input_dim : 0
I0423 01:41:01.420218 140604545877824 base_runner.py:59] task.decoder.softmax.is_inference : NoneType
I0423 01:41:01.420376 140604545877824 base_runner.py:59] task.decoder.softmax.logits_abs_max : NoneType
I0423 01:41:01.420506 140604545877824 base_runner.py:59] task.decoder.softmax.name : ''
I0423 01:41:01.420681 140604545877824 base_runner.py:59] task.decoder.softmax.num_classes : 76
I0423 01:41:01.420855 140604545877824 base_runner.py:59] task.decoder.softmax.num_sampled : 0
I0423 01:41:01.421015 140604545877824 base_runner.py:59] task.decoder.softmax.num_shards : 1
I0423 01:41:01.421165 140604545877824 base_runner.py:59] task.decoder.softmax.params_init.method : 'uniform'
I0423 01:41:01.421288 140604545877824 base_runner.py:59] task.decoder.softmax.params_init.scale : 0.1
I0423 01:41:01.421433 140604545877824 base_runner.py:59] task.decoder.softmax.params_init.seed : NoneType
I0423 01:41:01.421619 140604545877824 base_runner.py:59] task.decoder.softmax.qdomain.default : NoneType
I0423 01:41:01.421718 140604545877824 base_runner.py:59] task.decoder.softmax.random_seed : NoneType
I0423 01:41:01.421926 140604545877824 base_runner.py:59] task.decoder.softmax.skip_lp_regularization : NoneType
I0423 01:41:01.422069 140604545877824 base_runner.py:59] task.decoder.softmax.vn.global_vn : False
I0423 01:41:01.422219 140604545877824 base_runner.py:59] task.decoder.softmax.vn.per_step_vn : False
I0423 01:41:01.422355 140604545877824 base_runner.py:59] task.decoder.softmax.vn.scale : NoneType
I0423 01:41:01.422477 140604545877824 base_runner.py:59] task.decoder.softmax.vn.seed : NoneType
I0423 01:41:01.422607 140604545877824 base_runner.py:59] task.decoder.softmax_uses_attention : True
I0423 01:41:01.422778 140604545877824 base_runner.py:59] task.decoder.source_dim : 2048
I0423 01:41:01.422901 140604545877824 base_runner.py:59] task.decoder.target_eos_id : 2
I0423 01:41:01.423034 140604545877824 base_runner.py:59] task.decoder.target_seq_len : 620
I0423 01:41:01.423235 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.allow_implicit_capture : NoneType
I0423 01:41:01.423370 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.cls : type/lingvo.core.target_sequence_sampler/TargetSequenceSampler
I0423 01:41:01.423550 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.dtype : float32
I0423 01:41:01.423685 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.fprop_dtype : NoneType
I0423 01:41:01.423810 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.inference_driver_name : NoneType
I0423 01:41:01.423931 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.is_inference : NoneType
I0423 01:41:01.424066 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.name : 'target_sequence_sampler'
I0423 01:41:01.424247 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.params_init.method : 'xavier'
I0423 01:41:01.424390 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.params_init.scale : 1.000001
I0423 01:41:01.424528 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.params_init.seed : NoneType
I0423 01:41:01.424697 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.random_seed : NoneType
I0423 01:41:01.424951 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.skip_lp_regularization : NoneType
I0423 01:41:01.425088 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.target_eoc_id : -1
I0423 01:41:01.425231 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.target_eos_id : 2
I0423 01:41:01.425362 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.target_seq_len : 0
I0423 01:41:01.425496 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.target_sos_id : 1
I0423 01:41:01.425643 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.temperature : 1.0
I0423 01:41:01.425822 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.vn.global_vn : False
I0423 01:41:01.425957 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.vn.per_step_vn : False
I0423 01:41:01.426101 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.vn.scale : NoneType
I0423 01:41:01.426242 140604545877824 base_runner.py:59] task.decoder.target_sequence_sampler.vn.seed : NoneType
I0423 01:41:01.426383 140604545877824 base_runner.py:59] task.decoder.target_sos_id : 1
I0423 01:41:01.426524 140604545877824 base_runner.py:59] task.decoder.token_normalized_per_seq_loss : False
I0423 01:41:01.426656 140604545877824 base_runner.py:59] task.decoder.use_unnormalized_logits_as_log_probs : True
I0423 01:41:01.426786 140604545877824 base_runner.py:59] task.decoder.use_while_loop_based_unrolling : False
I0423 01:41:01.426923 140604545877824 base_runner.py:59] task.decoder.vn.global_vn : False
I0423 01:41:01.427060 140604545877824 base_runner.py:59] task.decoder.vn.per_step_vn : False
I0423 01:41:01.427191 140604545877824 base_runner.py:59] task.decoder.vn.scale : NoneType
I0423 01:41:01.427312 140604545877824 base_runner.py:59] task.decoder.vn.seed : NoneType
I0423 01:41:01.427455 140604545877824 base_runner.py:59] task.dtype : float32
I0423 01:41:01.427581 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.activation : 'RELU'
I0423 01:41:01.427698 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.427835 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.batch_norm : True
I0423 01:41:01.427959 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.bias : False
I0423 01:41:01.428086 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.bn_decay : 0.999
I0423 01:41:01.428270 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.bn_fold_weights : NoneType
I0423 01:41:01.428410 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.causal_convolution : False
I0423 01:41:01.428533 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.cls : type/lingvo.core.layers/Conv2DLayer
I0423 01:41:01.428672 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.conv_last : False
I0423 01:41:01.428815 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.dilation_rate : (1, 1)
I0423 01:41:01.428952 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.disable_activation_quantization : False
I0423 01:41:01.429075 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.dtype : float32
I0423 01:41:01.429224 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.filter_shape : [3, 3, 'NoneType', 'NoneType']
I0423 01:41:01.429364 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.filter_stride : [1, 1]
I0423 01:41:01.429484 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.fprop_dtype : NoneType
I0423 01:41:01.429965 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.inference_driver_name : NoneType
I0423 01:41:01.430260 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.is_inference : NoneType
I0423 01:41:01.430516 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.name : ''
I0423 01:41:01.430752 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.params_init.method : 'truncated_gaussian'
I0423 01:41:01.431074 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.params_init.scale : 0.1
I0423 01:41:01.431286 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.params_init.seed : NoneType
I0423 01:41:01.431516 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.qdomain.default : NoneType
I0423 01:41:01.431658 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.random_seed : NoneType
I0423 01:41:01.431800 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.431925 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.vn.global_vn : False
I0423 01:41:01.432065 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.vn.per_step_vn : False
I0423 01:41:01.432214 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.vn.scale : NoneType
I0423 01:41:01.432341 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.vn.seed : NoneType
I0423 01:41:01.432494 140604545877824 base_runner.py:59] task.encoder.after_conv_lstm_cnn_tpl.weight_norm : False
I0423 01:41:01.432661 140604545877824 base_runner.py:59] task.encoder.allow_implicit_capture : NoneType
I0423 01:41:01.432785 140604545877824 base_runner.py:59] task.encoder.bidi_rnn_type : 'func'
I0423 01:41:01.432922 140604545877824 base_runner.py:59] task.encoder.cls : type/lingvo.tasks.asr.encoder/AsrEncoder
I0423 01:41:01.433058 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.activation : 'RELU'
I0423 01:41:01.433199 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.433328 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.batch_norm : True
I0423 01:41:01.433452 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.bias : False
I0423 01:41:01.433595 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.bn_decay : 0.999
I0423 01:41:01.433712 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.bn_fold_weights : NoneType
I0423 01:41:01.433789 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.causal_convolution : False
I0423 01:41:01.433873 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.cls : type/lingvo.core.layers/Conv2DLayer
I0423 01:41:01.433957 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.conv_last : False
I0423 01:41:01.434024 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.dilation_rate : (1, 1)
I0423 01:41:01.434088 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.disable_activation_quantization : False
I0423 01:41:01.434166 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.dtype : float32
I0423 01:41:01.434236 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.filter_shape : (0, 0, 0, 0)
I0423 01:41:01.434301 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.filter_stride : (0, 0)
I0423 01:41:01.434366 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.fprop_dtype : NoneType
I0423 01:41:01.434431 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.inference_driver_name : NoneType
I0423 01:41:01.434496 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.is_inference : NoneType
I0423 01:41:01.434579 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.name : ''
I0423 01:41:01.434680 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.params_init.method : 'gaussian'
I0423 01:41:01.434756 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.params_init.scale : 0.001
I0423 01:41:01.434821 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.params_init.seed : NoneType
I0423 01:41:01.434887 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.qdomain.default : NoneType
I0423 01:41:01.434952 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.random_seed : NoneType
I0423 01:41:01.435016 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.435081 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.vn.global_vn : False
I0423 01:41:01.435146 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.vn.per_step_vn : False
I0423 01:41:01.435226 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.vn.scale : NoneType
I0423 01:41:01.435292 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.vn.seed : NoneType
I0423 01:41:01.435356 140604545877824 base_runner.py:59] task.encoder.cnn_tpl.weight_norm : False
I0423 01:41:01.435428 140604545877824 base_runner.py:59] task.encoder.conv_filter_shapes : [(3, 3, 1, 32), (3, 3, 32, 32)]
I0423 01:41:01.435493 140604545877824 base_runner.py:59] task.encoder.conv_filter_strides : [(2, 2), (2, 2)]
I0423 01:41:01.435566 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.435631 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.cell_shape : ['NoneType', 'NoneType', 'NoneType', 'NoneType']
I0423 01:41:01.435696 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.cell_value_cap : 10.0
I0423 01:41:01.435761 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.cls : type/lingvo.core.rnn_cell/ConvLSTMCell
I0423 01:41:01.435852 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.dtype : float32
I0423 01:41:01.435918 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.filter_shape : [1, 3]
I0423 01:41:01.435983 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.fprop_dtype : NoneType
I0423 01:41:01.436048 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.inference_driver_name : NoneType
I0423 01:41:01.436113 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.inputs_arity : 1
I0423 01:41:01.436196 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.inputs_shape : ['NoneType', 'NoneType', 'NoneType', 'NoneType']
I0423 01:41:01.436268 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.is_inference : NoneType
I0423 01:41:01.436337 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.name : ''
I0423 01:41:01.436403 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.num_input_nodes : 0
I0423 01:41:01.436484 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.num_output_nodes : 0
I0423 01:41:01.436589 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.output_nonlinearity : True
I0423 01:41:01.436668 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.params_init.method : 'truncated_gaussian'
I0423 01:41:01.436734 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.params_init.scale : 0.1
I0423 01:41:01.436799 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.params_init.seed : NoneType
I0423 01:41:01.436864 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.qdomain.default : NoneType
I0423 01:41:01.436929 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.random_seed : NoneType
I0423 01:41:01.436993 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.reset_cell_state : False
I0423 01:41:01.437058 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.437123 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.vn.global_vn : False
I0423 01:41:01.437199 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.vn.per_step_vn : False
I0423 01:41:01.437265 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.vn.scale : NoneType
I0423 01:41:01.437356 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.vn.seed : NoneType
I0423 01:41:01.437422 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.zero_state_init_params.method : 'zeros'
I0423 01:41:01.437487 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.zero_state_init_params.seed : NoneType
I0423 01:41:01.438094 140604545877824 base_runner.py:59] task.encoder.conv_lstm_tpl.zo_prob : 0.0
I0423 01:41:01.438391 140604545877824 base_runner.py:59] task.encoder.dtype : float32
I0423 01:41:01.438636 140604545877824 base_runner.py:59] task.encoder.extra_per_layer_outputs : False
I0423 01:41:01.438953 140604545877824 base_runner.py:59] task.encoder.fprop_dtype : NoneType
I0423 01:41:01.439133 140604545877824 base_runner.py:59] task.encoder.highway_skip : False
I0423 01:41:01.439356 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.439588 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.batch_norm : False
I0423 01:41:01.439810 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.carry_bias_init : 1.0
I0423 01:41:01.440064 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.cls : type/lingvo.core.layers/HighwaySkipLayer
I0423 01:41:01.440283 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.couple_carry_transform_gates : False
I0423 01:41:01.440517 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.dtype : float32
I0423 01:41:01.440694 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.fprop_dtype : NoneType
I0423 01:41:01.440916 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.inference_driver_name : NoneType
I0423 01:41:01.441145 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.input_dim : 0
I0423 01:41:01.441463 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.is_inference : NoneType
I0423 01:41:01.441632 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.name : ''
I0423 01:41:01.441771 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.params_init.method : 'xavier'
I0423 01:41:01.441895 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.params_init.scale : 1.000001
I0423 01:41:01.442018 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.params_init.seed : NoneType
I0423 01:41:01.442272 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.random_seed : NoneType
I0423 01:41:01.442441 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.442653 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.vn.global_vn : False
I0423 01:41:01.442890 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.vn.per_step_vn : False
I0423 01:41:01.443221 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.vn.scale : NoneType
I0423 01:41:01.443490 140604545877824 base_runner.py:59] task.encoder.highway_skip_tpl.vn.seed : NoneType
I0423 01:41:01.443783 140604545877824 base_runner.py:59] task.encoder.inference_driver_name : NoneType
I0423 01:41:01.444006 140604545877824 base_runner.py:59] task.encoder.input_shape : ['NoneType', 'NoneType', 80, 1]
I0423 01:41:01.444161 140604545877824 base_runner.py:59] task.encoder.is_inference : NoneType
I0423 01:41:01.444297 140604545877824 base_runner.py:59] task.encoder.layer_index_before_stacking : -1
I0423 01:41:01.444440 140604545877824 base_runner.py:59] task.encoder.lstm_cell_size : 1024
I0423 01:41:01.444567 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.444696 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.apply_pruning : False
I0423 01:41:01.444838 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.apply_pruning_to_projection : False
I0423 01:41:01.444974 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.bias_init.method : 'constant'
I0423 01:41:01.445104 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.bias_init.scale : 0.0
I0423 01:41:01.445254 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.bias_init.seed : 0
I0423 01:41:01.445392 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.cell_value_cap : 10.0
I0423 01:41:01.445517 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.cls : type/lingvo.core.rnn_cell/LSTMCellSimple
I0423 01:41:01.446036 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.couple_input_forget_gates : False
I0423 01:41:01.446229 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.dtype : float32
I0423 01:41:01.446402 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.enable_lstm_bias : True
I0423 01:41:01.446596 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.forget_gate_bias : 0.0
I0423 01:41:01.446793 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.fprop_dtype : NoneType
I0423 01:41:01.446966 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.gradient_pruning : False
I0423 01:41:01.447200 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.inference_driver_name : NoneType
I0423 01:41:01.447427 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.inputs_arity : 1
I0423 01:41:01.447653 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.is_inference : NoneType
I0423 01:41:01.447892 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.name : ''
I0423 01:41:01.448062 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.num_hidden_nodes : 0
I0423 01:41:01.448259 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.num_input_nodes : 0
I0423 01:41:01.448514 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.num_output_nodes : 0
I0423 01:41:01.448718 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.output_nonlinearity : True
I0423 01:41:01.448932 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.params_init.method : 'uniform'
I0423 01:41:01.449125 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.params_init.scale : 0.1
I0423 01:41:01.449315 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.params_init.seed : NoneType
I0423 01:41:01.449508 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.qdomain.c_state : NoneType
I0423 01:41:01.449729 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.qdomain.default : NoneType
I0423 01:41:01.449942 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.qdomain.fullyconnected : NoneType
I0423 01:41:01.450113 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.qdomain.m_state : NoneType
I0423 01:41:01.450338 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.qdomain.weight : NoneType
I0423 01:41:01.450517 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.random_seed : NoneType
I0423 01:41:01.450714 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.reset_cell_state : False
I0423 01:41:01.450939 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.451129 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.vn.global_vn : False
I0423 01:41:01.451318 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.vn.per_step_vn : False
I0423 01:41:01.451516 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.vn.scale : NoneType
I0423 01:41:01.451752 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.vn.seed : NoneType
I0423 01:41:01.451996 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.zero_state_init_params.method : 'zeros'
I0423 01:41:01.452241 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.zero_state_init_params.seed : NoneType
I0423 01:41:01.452556 140604545877824 base_runner.py:59] task.encoder.lstm_tpl.zo_prob : 0.0
I0423 01:41:01.452833 140604545877824 base_runner.py:59] task.encoder.name : ''
I0423 01:41:01.453165 140604545877824 base_runner.py:59] task.encoder.num_cnn_layers : 2
I0423 01:41:01.453395 140604545877824 base_runner.py:59] task.encoder.num_conv_lstm_layers : 0
I0423 01:41:01.453568 140604545877824 base_runner.py:59] task.encoder.num_lstm_layers : 4
I0423 01:41:01.453710 140604545877824 base_runner.py:59] task.encoder.pad_steps : 6
I0423 01:41:01.453834 140604545877824 base_runner.py:59] task.encoder.params_init.method : 'xavier'
I0423 01:41:01.453958 140604545877824 base_runner.py:59] task.encoder.params_init.scale : 1.000001
I0423 01:41:01.454080 140604545877824 base_runner.py:59] task.encoder.params_init.seed : NoneType
I0423 01:41:01.454225 140604545877824 base_runner.py:59] task.encoder.proj_tpl.activation : 'RELU'
I0423 01:41:01.454399 140604545877824 base_runner.py:59] task.encoder.proj_tpl.affine_last : False
I0423 01:41:01.454540 140604545877824 base_runner.py:59] task.encoder.proj_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.454758 140604545877824 base_runner.py:59] task.encoder.proj_tpl.apply_pruning : False
I0423 01:41:01.454978 140604545877824 base_runner.py:59] task.encoder.proj_tpl.batch_norm : True
I0423 01:41:01.455171 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bias_init : 0.0
I0423 01:41:01.455377 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_fold_weights : NoneType
I0423 01:41:01.455584 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.add_stats_to_moving_average_variables : NoneType
I0423 01:41:01.455801 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.allow_implicit_capture : NoneType
I0423 01:41:01.455984 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.cls : type/lingvo.core.bn_layers/BatchNormLayer
I0423 01:41:01.456166 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.decay : 0.999
I0423 01:41:01.456347 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.dim : 0
I0423 01:41:01.456568 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.dtype : float32
I0423 01:41:01.456768 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.enable_cross_replica_sum_on_tpu : True
I0423 01:41:01.456946 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.fprop_dtype : NoneType
I0423 01:41:01.457184 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.inference_driver_name : NoneType
I0423 01:41:01.457363 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.is_inference : NoneType
I0423 01:41:01.457607 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.name : ''
I0423 01:41:01.457781 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.params_init.method : 'xavier'
I0423 01:41:01.457968 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.params_init.scale : 1.000001
I0423 01:41:01.458146 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.params_init.seed : NoneType
I0423 01:41:01.458333 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.random_seed : NoneType
I0423 01:41:01.458642 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.skip_lp_regularization : NoneType
I0423 01:41:01.458927 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.use_moving_avg_in_training : False
I0423 01:41:01.459142 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.vn.global_vn : False
I0423 01:41:01.459326 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.vn.per_step_vn : False
I0423 01:41:01.459506 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.vn.scale : NoneType
I0423 01:41:01.459683 140604545877824 base_runner.py:59] task.encoder.proj_tpl.bn_params.vn.seed : NoneType
I0423 01:41:01.459861 140604545877824 base_runner.py:59] task.encoder.proj_tpl.cls : type/lingvo.core.layers/ProjectionLayer
I0423 01:41:01.460034 140604545877824 base_runner.py:59] task.encoder.proj_tpl.dtype : float32
I0423 01:41:01.460311 140604545877824 base_runner.py:59] task.encoder.proj_tpl.fprop_dtype : NoneType
I0423 01:41:01.460686 140604545877824 base_runner.py:59] task.encoder.proj_tpl.has_bias : False
I0423 01:41:01.460952 140604545877824 base_runner.py:59] task.encoder.proj_tpl.inference_driver_name : NoneType
I0423 01:41:01.461220 140604545877824 base_runner.py:59] task.encoder.proj_tpl.input_dim : 0
I0423 01:41:01.461392 140604545877824 base_runner.py:59] task.encoder.proj_tpl.is_inference : NoneType
I0423 01:41:01.461599 140604545877824 base_runner.py:59] task.encoder.proj_tpl.name : ''
I0423 01:41:01.461773 140604545877824 base_runner.py:59] task.encoder.proj_tpl.output_dim : 0
I0423 01:41:01.461977 140604545877824 base_runner.py:59] task.encoder.proj_tpl.params_init.method : 'truncated_gaussian'
I0423 01:41:01.462170 140604545877824 base_runner.py:59] task.encoder.proj_tpl.params_init.scale : 0.1
I0423 01:41:01.462417 140604545877824 base_runner.py:59] task.encoder.proj_tpl.params_init.seed : NoneType
I0423 01:41:01.462692 140604545877824 base_runner.py:59] task.encoder.proj_tpl.qdomain.default : NoneType
I0423 01:41:01.462977 140604545877824 base_runner.py:59] task.encoder.proj_tpl.random_seed : NoneType
I0423 01:41:01.463266 140604545877824 base_runner.py:59] task.encoder.proj_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.463439 140604545877824 base_runner.py:59] task.encoder.proj_tpl.vn.global_vn : False
I0423 01:41:01.463627 140604545877824 base_runner.py:59] task.encoder.proj_tpl.vn.per_step_vn : False
I0423 01:41:01.463798 140604545877824 base_runner.py:59] task.encoder.proj_tpl.vn.scale : NoneType
I0423 01:41:01.463977 140604545877824 base_runner.py:59] task.encoder.proj_tpl.vn.seed : NoneType
I0423 01:41:01.464167 140604545877824 base_runner.py:59] task.encoder.proj_tpl.weight_norm : False
I0423 01:41:01.464394 140604545877824 base_runner.py:59] task.encoder.project_lstm_output : True
I0423 01:41:01.464625 140604545877824 base_runner.py:59] task.encoder.random_seed : NoneType
I0423 01:41:01.464856 140604545877824 base_runner.py:59] task.encoder.residual_start : 0
I0423 01:41:01.465026 140604545877824 base_runner.py:59] task.encoder.residual_stride : 1
I0423 01:41:01.465212 140604545877824 base_runner.py:59] task.encoder.skip_lp_regularization : NoneType
I0423 01:41:01.465456 140604545877824 base_runner.py:59] task.encoder.specaugment_network.allow_implicit_capture : NoneType
I0423 01:41:01.465663 140604545877824 base_runner.py:59] task.encoder.specaugment_network.cls : type/lingvo.core.spectrum_augmenter/SpectrumAugmenter
I0423 01:41:01.465835 140604545877824 base_runner.py:59] task.encoder.specaugment_network.domain_ids : [0]
I0423 01:41:01.466015 140604545877824 base_runner.py:59] task.encoder.specaugment_network.dtype : float32
I0423 01:41:01.466251 140604545877824 base_runner.py:59] task.encoder.specaugment_network.fprop_dtype : NoneType
I0423 01:41:01.466513 140604545877824 base_runner.py:59] task.encoder.specaugment_network.freq_mask_count : 1
I0423 01:41:01.466774 140604545877824 base_runner.py:59] task.encoder.specaugment_network.freq_mask_max_bins : 15
I0423 01:41:01.467061 140604545877824 base_runner.py:59] task.encoder.specaugment_network.gaussian_noise : False
I0423 01:41:01.467316 140604545877824 base_runner.py:59] task.encoder.specaugment_network.inference_driver_name : NoneType
I0423 01:41:01.467490 140604545877824 base_runner.py:59] task.encoder.specaugment_network.is_inference : NoneType
I0423 01:41:01.467633 140604545877824 base_runner.py:59] task.encoder.specaugment_network.name : ''
I0423 01:41:01.467770 140604545877824 base_runner.py:59] task.encoder.specaugment_network.params_init.method : 'xavier'
I0423 01:41:01.467905 140604545877824 base_runner.py:59] task.encoder.specaugment_network.params_init.scale : 1.000001
I0423 01:41:01.468071 140604545877824 base_runner.py:59] task.encoder.specaugment_network.params_init.seed : NoneType
I0423 01:41:01.468210 140604545877824 base_runner.py:59] task.encoder.specaugment_network.random_seed : NoneType
I0423 01:41:01.468388 140604545877824 base_runner.py:59] task.encoder.specaugment_network.skip_lp_regularization : NoneType
I0423 01:41:01.468575 140604545877824 base_runner.py:59] task.encoder.specaugment_network.stack_height : 3
I0423 01:41:01.468763 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_mask_count : 1
I0423 01:41:01.468945 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_mask_max_frames : 50
I0423 01:41:01.469172 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_mask_max_ratio : 1.0
I0423 01:41:01.469350 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_masks_per_frame : 0.0
I0423 01:41:01.469574 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_warp_bound : 'static'
I0423 01:41:01.469771 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_warp_max_frames : 0
I0423 01:41:01.469951 140604545877824 base_runner.py:59] task.encoder.specaugment_network.time_warp_max_ratio : 0.0
I0423 01:41:01.470146 140604545877824 base_runner.py:59] task.encoder.specaugment_network.unstack : False
I0423 01:41:01.470356 140604545877824 base_runner.py:59] task.encoder.specaugment_network.use_dynamic_time_mask_max_frames : False
I0423 01:41:01.470598 140604545877824 base_runner.py:59] task.encoder.specaugment_network.use_input_dependent_random_seed : False
I0423 01:41:01.470796 140604545877824 base_runner.py:59] task.encoder.specaugment_network.use_noise : False
I0423 01:41:01.470974 140604545877824 base_runner.py:59] task.encoder.specaugment_network.vn.global_vn : False
I0423 01:41:01.471183 140604545877824 base_runner.py:59] task.encoder.specaugment_network.vn.per_step_vn : False
I0423 01:41:01.471377 140604545877824 base_runner.py:59] task.encoder.specaugment_network.vn.scale : NoneType
I0423 01:41:01.471554 140604545877824 base_runner.py:59] task.encoder.specaugment_network.vn.seed : NoneType
I0423 01:41:01.471755 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.allow_implicit_capture : NoneType
I0423 01:41:01.472019 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.cls : type/lingvo.core.layers/StackingOverTime
I0423 01:41:01.472221 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.dtype : float32
I0423 01:41:01.472436 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.fprop_dtype : NoneType
I0423 01:41:01.472639 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.inference_driver_name : NoneType
I0423 01:41:01.472843 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.is_inference : NoneType
I0423 01:41:01.473116 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.left_context : 0
I0423 01:41:01.473321 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.name : ''
I0423 01:41:01.473517 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.params_init.method : 'xavier'
I0423 01:41:01.473742 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.params_init.scale : 1.000001
I0423 01:41:01.473938 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.params_init.seed : NoneType
I0423 01:41:01.474137 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.random_seed : NoneType
I0423 01:41:01.474345 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.right_context : 0
I0423 01:41:01.474630 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.skip_lp_regularization : NoneType
I0423 01:41:01.474808 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.stride : 1
I0423 01:41:01.475000 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.vn.global_vn : False
I0423 01:41:01.475199 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.vn.per_step_vn : False
I0423 01:41:01.475394 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.vn.scale : NoneType
I0423 01:41:01.475621 140604545877824 base_runner.py:59] task.encoder.stacking_layer_tpl.vn.seed : NoneType
I0423 01:41:01.475814 140604545877824 base_runner.py:59] task.encoder.use_specaugment : False
I0423 01:41:01.476007 140604545877824 base_runner.py:59] task.encoder.vn.global_vn : False
I0423 01:41:01.476210 140604545877824 base_runner.py:59] task.encoder.vn.per_step_vn : False
I0423 01:41:01.476414 140604545877824 base_runner.py:59] task.encoder.vn.scale : NoneType
I0423 01:41:01.476611 140604545877824 base_runner.py:59] task.encoder.vn.seed : NoneType
I0423 01:41:01.476834 140604545877824 base_runner.py:59] task.eval.decoder_samples_per_summary : 0
I0423 01:41:01.477028 140604545877824 base_runner.py:59] task.eval.load_checkpoint_from : NoneType
I0423 01:41:01.477219 140604545877824 base_runner.py:59] task.eval.samples_per_summary : 5000
I0423 01:41:01.477410 140604545877824 base_runner.py:59] task.eval.start_decoder_after : 0
I0423 01:41:01.477625 140604545877824 base_runner.py:59] task.eval.start_eval_after : 0
I0423 01:41:01.477819 140604545877824 base_runner.py:59] task.fprop_dtype : NoneType
I0423 01:41:01.477999 140604545877824 base_runner.py:59] task.frontend : NoneType
I0423 01:41:01.478206 140604545877824 base_runner.py:59] task.include_auxiliary_metrics : True
I0423 01:41:01.478400 140604545877824 base_runner.py:59] task.inference_driver_name : NoneType
I0423 01:41:01.478579 140604545877824 base_runner.py:59] task.input : NoneType
I0423 01:41:01.478774 140604545877824 base_runner.py:59] task.is_inference : NoneType
I0423 01:41:01.478963 140604545877824 base_runner.py:59] task.name : 'librispeech'
I0423 01:41:01.479171 140604545877824 base_runner.py:59] task.online_encoder : NoneType
I0423 01:41:01.479407 140604545877824 base_runner.py:59] task.params_init.method : 'xavier'
I0423 01:41:01.479652 140604545877824 base_runner.py:59] task.params_init.scale : 1.000001
I0423 01:41:01.479835 140604545877824 base_runner.py:59] task.params_init.seed : NoneType
I0423 01:41:01.480053 140604545877824 base_runner.py:59] task.random_seed : NoneType
I0423 01:41:01.480284 140604545877824 base_runner.py:59] task.skip_lp_regularization : NoneType
I0423 01:41:01.480475 140604545877824 base_runner.py:59] task.train.bprop_variable_exclusion : NoneType
I0423 01:41:01.480652 140604545877824 base_runner.py:59] task.train.bprop_variable_filter : NoneType
I0423 01:41:01.480843 140604545877824 base_runner.py:59] task.train.clip_gradient_norm_to_value : 1.0
I0423 01:41:01.481041 140604545877824 base_runner.py:59] task.train.clip_gradient_single_norm_to_value : 0.0
I0423 01:41:01.481276 140604545877824 base_runner.py:59] task.train.colocate_gradients_with_ops : True
I0423 01:41:01.481469 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.jobname : 'eval_dev'
I0423 01:41:01.481673 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.local_filesystem : False
I0423 01:41:01.481871 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.logdir : ''
I0423 01:41:01.482064 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.metric : 'log_pplx'
I0423 01:41:01.482335 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.minimize : True
I0423 01:41:01.482542 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.name : 'MetricHistory'
I0423 01:41:01.482746 140604545877824 base_runner.py:59] task.train.early_stop.metric_history.tfevent_file : False
I0423 01:41:01.482977 140604545877824 base_runner.py:59] task.train.early_stop.min_steps : 0
I0423 01:41:01.483253 140604545877824 base_runner.py:59] task.train.early_stop.name : 'EarlyStop'
I0423 01:41:01.483450 140604545877824 base_runner.py:59] task.train.early_stop.tolerance : 0.0
I0423 01:41:01.483652 140604545877824 base_runner.py:59] task.train.early_stop.verbose : True
I0423 01:41:01.483869 140604545877824 base_runner.py:59] task.train.early_stop.window : 0
I0423 01:41:01.484126 140604545877824 base_runner.py:59] task.train.ema_decay : 0.0
I0423 01:41:01.484333 140604545877824 base_runner.py:59] task.train.ema_decay_moving_vars : NoneType
I0423 01:41:01.484516 140604545877824 base_runner.py:59] task.train.enqueue_max_steps : -1
I0423 01:41:01.484714 140604545877824 base_runner.py:59] task.train.gate_gradients : False
I0423 01:41:01.484910 140604545877824 base_runner.py:59] task.train.grad_aggregation_method : 1
I0423 01:41:01.485147 140604545877824 base_runner.py:59] task.train.grad_norm_to_clip_to_zero : 100.0
I0423 01:41:01.485357 140604545877824 base_runner.py:59] task.train.grad_norm_tracker : NoneType
I0423 01:41:01.485573 140604545877824 base_runner.py:59] task.train.init_from_checkpoint_rules : {}
I0423 01:41:01.485765 140604545877824 base_runner.py:59] task.train.l1_regularizer_weight : NoneType
I0423 01:41:01.485946 140604545877824 base_runner.py:59] task.train.l2_regularizer_weight : 1e-06
I0423 01:41:01.486143 140604545877824 base_runner.py:59] task.train.learner : NoneType
I0423 01:41:01.486350 140604545877824 base_runner.py:59] task.train.learning_rate : 0.00025
I0423 01:41:01.486563 140604545877824 base_runner.py:59] task.train.lr_schedule.allow_implicit_capture : NoneType
I0423 01:41:01.486791 140604545877824 base_runner.py:59] task.train.lr_schedule.cls : type/lingvo.core.schedule/ContinuousSchedule
I0423 01:41:01.486985 140604545877824 base_runner.py:59] task.train.lr_schedule.dtype : float32
I0423 01:41:01.487178 140604545877824 base_runner.py:59] task.train.lr_schedule.fprop_dtype : NoneType
I0423 01:41:01.487373 140604545877824 base_runner.py:59] task.train.lr_schedule.half_life_steps : 100000
I0423 01:41:01.487567 140604545877824 base_runner.py:59] task.train.lr_schedule.inference_driver_name : NoneType
I0423 01:41:01.487763 140604545877824 base_runner.py:59] task.train.lr_schedule.initial_value : 1.0
I0423 01:41:01.487972 140604545877824 base_runner.py:59] task.train.lr_schedule.is_inference : NoneType
I0423 01:41:01.488178 140604545877824 base_runner.py:59] task.train.lr_schedule.min : 0.01
I0423 01:41:01.488370 140604545877824 base_runner.py:59] task.train.lr_schedule.name : 'LRSched'
I0423 01:41:01.488550 140604545877824 base_runner.py:59] task.train.lr_schedule.params_init.method : 'xavier'
I0423 01:41:01.488841 140604545877824 base_runner.py:59] task.train.lr_schedule.params_init.scale : 1.000001
I0423 01:41:01.489018 140604545877824 base_runner.py:59] task.train.lr_schedule.params_init.seed : NoneType
I0423 01:41:01.489270 140604545877824 base_runner.py:59] task.train.lr_schedule.random_seed : NoneType
I0423 01:41:01.489503 140604545877824 base_runner.py:59] task.train.lr_schedule.skip_lp_regularization : NoneType
I0423 01:41:01.489696 140604545877824 base_runner.py:59] task.train.lr_schedule.start_step : 50000
I0423 01:41:01.489877 140604545877824 base_runner.py:59] task.train.lr_schedule.vn.global_vn : False
I0423 01:41:01.490059 140604545877824 base_runner.py:59] task.train.lr_schedule.vn.per_step_vn : False
I0423 01:41:01.490252 140604545877824 base_runner.py:59] task.train.lr_schedule.vn.scale : NoneType
I0423 01:41:01.490438 140604545877824 base_runner.py:59] task.train.lr_schedule.vn.seed : NoneType
I0423 01:41:01.490620 140604545877824 base_runner.py:59] task.train.max_steps : 4000000
I0423 01:41:01.490835 140604545877824 base_runner.py:59] task.train.optimizer.allow_implicit_capture : NoneType
I0423 01:41:01.491022 140604545877824 base_runner.py:59] task.train.optimizer.beta1 : 0.9
I0423 01:41:01.491224 140604545877824 base_runner.py:59] task.train.optimizer.beta2 : 0.999
I0423 01:41:01.491459 140604545877824 base_runner.py:59] task.train.optimizer.cls : type/lingvo.core.optimizer/Adam
I0423 01:41:01.491664 140604545877824 base_runner.py:59] task.train.optimizer.dtype : float32
I0423 01:41:01.491861 140604545877824 base_runner.py:59] task.train.optimizer.epsilon : 1e-06
I0423 01:41:01.492060 140604545877824 base_runner.py:59] task.train.optimizer.fprop_dtype : NoneType
I0423 01:41:01.492309 140604545877824 base_runner.py:59] task.train.optimizer.inference_driver_name : NoneType
I0423 01:41:01.492498 140604545877824 base_runner.py:59] task.train.optimizer.is_inference : NoneType
I0423 01:41:01.492717 140604545877824 base_runner.py:59] task.train.optimizer.name : 'Adam'
I0423 01:41:01.492893 140604545877824 base_runner.py:59] task.train.optimizer.params_init.method : 'xavier'
I0423 01:41:01.493083 140604545877824 base_runner.py:59] task.train.optimizer.params_init.scale : 1.000001
I0423 01:41:01.493302 140604545877824 base_runner.py:59] task.train.optimizer.params_init.seed : NoneType
I0423 01:41:01.493493 140604545877824 base_runner.py:59] task.train.optimizer.random_seed : NoneType
I0423 01:41:01.493856 140604545877824 base_runner.py:59] task.train.optimizer.skip_lp_regularization : NoneType
I0423 01:41:01.494095 140604545877824 base_runner.py:59] task.train.optimizer.vn.global_vn : False
I0423 01:41:01.494360 140604545877824 base_runner.py:59] task.train.optimizer.vn.per_step_vn : False
I0423 01:41:01.494595 140604545877824 base_runner.py:59] task.train.optimizer.vn.scale : NoneType
I0423 01:41:01.494735 140604545877824 base_runner.py:59] task.train.optimizer.vn.seed : NoneType
I0423 01:41:01.494916 140604545877824 base_runner.py:59] task.train.pruning_hparams_dict : NoneType
I0423 01:41:01.495110 140604545877824 base_runner.py:59] task.train.save_interval_seconds : 600
I0423 01:41:01.495354 140604545877824 base_runner.py:59] task.train.save_keep_checkpoint_every_n_hours : 0.5
I0423 01:41:01.495555 140604545877824 base_runner.py:59] task.train.save_max_to_keep : 100
I0423 01:41:01.495737 140604545877824 base_runner.py:59] task.train.start_up_delay_steps : 200
I0423 01:41:01.495946 140604545877824 base_runner.py:59] task.train.summary_interval_steps : 100
I0423 01:41:01.496146 140604545877824 base_runner.py:59] task.train.tpu_steps_per_loop : 20
I0423 01:41:01.496359 140604545877824 base_runner.py:59] task.train.vn_start_step : 20000
I0423 01:41:01.496558 140604545877824 base_runner.py:59] task.train.vn_std : 0.075
I0423 01:41:01.496752 140604545877824 base_runner.py:59] task.vn.global_vn : True
I0423 01:41:01.496935 140604545877824 base_runner.py:59] task.vn.per_step_vn : False
I0423 01:41:01.497130 140604545877824 base_runner.py:59] task.vn.scale : NoneType
I0423 01:41:01.497343 140604545877824 base_runner.py:59] task.vn.seed : NoneType
I0423 01:41:01.497535 140604545877824 base_runner.py:59] train.early_stop.metric_history.jobname : 'eval_dev'
I0423 01:41:01.497736 140604545877824 base_runner.py:59] train.early_stop.metric_history.local_filesystem : False
I0423 01:41:01.497929 140604545877824 base_runner.py:59] train.early_stop.metric_history.logdir : ''
I0423 01:41:01.498126 140604545877824 base_runner.py:59] train.early_stop.metric_history.metric : 'log_pplx'
I0423 01:41:01.498330 140604545877824 base_runner.py:59] train.early_stop.metric_history.minimize : True
I0423 01:41:01.498598 140604545877824 base_runner.py:59] train.early_stop.metric_history.name : 'MetricHistory'
I0423 01:41:01.498778 140604545877824 base_runner.py:59] train.early_stop.metric_history.tfevent_file : False
I0423 01:41:01.498973 140604545877824 base_runner.py:59] train.early_stop.min_steps : 0
I0423 01:41:01.499182 140604545877824 base_runner.py:59] train.early_stop.name : 'EarlyStop'
I0423 01:41:01.499381 140604545877824 base_runner.py:59] train.early_stop.tolerance : 0.0
I0423 01:41:01.499573 140604545877824 base_runner.py:59] train.early_stop.verbose : True
I0423 01:41:01.499753 140604545877824 base_runner.py:59] train.early_stop.window : 0
I0423 01:41:01.499972 140604545877824 base_runner.py:59] train.ema_decay : 0.0
I0423 01:41:01.500180 140604545877824 base_runner.py:59] train.ema_decay_moving_vars : NoneType
I0423 01:41:01.500385 140604545877824 base_runner.py:59] train.enqueue_max_steps : -1
I0423 01:41:01.500641 140604545877824 base_runner.py:59] train.init_from_checkpoint_rules : {}
I0423 01:41:01.500839 140604545877824 base_runner.py:59] train.max_steps : 4000000
I0423 01:41:01.501026 140604545877824 base_runner.py:59] train.save_interval_seconds : 600
I0423 01:41:01.501277 140604545877824 base_runner.py:59] train.save_keep_checkpoint_every_n_hours : 0.5
I0423 01:41:01.501479 140604545877824 base_runner.py:59] train.save_max_to_keep : 2
I0423 01:41:01.501685 140604545877824 base_runner.py:59] train.start_up_delay_steps : 200
I0423 01:41:01.501877 140604545877824 base_runner.py:59] train.summary_interval_steps : 100
I0423 01:41:01.502131 140604545877824 base_runner.py:59] train.tpu_steps_per_loop : 20
I0423 01:41:01.502354 140604545877824 base_runner.py:59] vn.global_vn : True
I0423 01:41:01.502563 140604545877824 base_runner.py:59] vn.per_step_vn : False
I0423 01:41:01.502800 140604545877824 base_runner.py:59] vn.scale : NoneType
I0423 01:41:01.502979 140604545877824 base_runner.py:59] vn.seed : NoneType
I0423 01:41:01.503282 140604545877824 base_runner.py:59]
I0423 01:41:01.503534 140604545877824 base_runner.py:60] ============================================================
I0423 01:41:01.538563 140604545877824 base_runner.py:111] Starting ...
I0423 01:41:01.540073 140604545877824 cluster.py:507] _LeastLoadedPlacer : ['/job:local/replica:0/task:0/device:CPU:0']
I0423 01:41:01.588492 140604545877824 cluster.py:525] Place variable global_step on /job:local/replica:0/task:0/device:CPU:0 8
I0423 01:41:01.756494 140604545877824 base_model.py:1056] Training parameters for <class 'lingvo.core.base_model.SingleTaskModel'>: {
early_stop: {
metric_history: {
jobname: "eval_dev"
local_filesystem: False
logdir: "/tmp/lingvo/log"
metric: "log_pplx"
minimize: True
name: "MetricHistory"
tfevent_file: False
}
min_steps: 0
name: "EarlyStop"
tolerance: 0.0
verbose: True
window: 0
}
ema_decay: 0.0
ema_decay_moving_vars: None
enqueue_max_steps: -1
init_from_checkpoint_rules: {}
max_steps: 4000000
save_interval_seconds: 600
save_keep_checkpoint_every_n_hours: 0.5
save_max_to_keep: 2
start_up_delay_steps: 200
summary_interval_steps: 100
tpu_steps_per_loop: 20
}
I0423 01:41:01.780864 140604545877824 base_model.py:277] input_params: {
allow_implicit_capture: None
append_eos_frame: True
bucket_adjust_every_n: 0
bucket_batch_limit: [48, 24, 24, 24, 24, 24, 24, 24]
bucket_upper_bound: [639, 1062, 1275, 1377, 1449, 1506, 1563, 1710]
cls: <class 'lingvo.tasks.asr.input_generator.AsrInput'>
dtype: <dtype: 'float32'>
file_buffer_size: 10000
file_buffer_size_in_seconds: 0
file_datasource: {
allow_implicit_capture: None
cls: <class 'lingvo.core.datasource.PrefixedDataSource'>
dtype: <dtype: 'float32'>
file_pattern: "train/train.tfrecords-*"
file_pattern_prefix: "/tmp/librispeech"
file_type: "tfrecord"
fprop_dtype: None
inference_driver_name: None
is_inference: None
name: "datasource"
params_init: {
method: "xavier"
scale: 1.000001
seed: None
}
random_seed: None
skip_lp_regularization: None
vn: {
global_vn: False
per_step_vn: False
scale: None
seed: None
}
}
file_parallelism: 16
file_pattern: ""
file_random_seed: 0
flush_every_n: 0
fprop_dtype: None
frame_size: 80
inference_driver_name: None
is_inference: None
name: "input"
num_batcher_threads: 1
num_partitions: None
num_samples: 281241
pad_to_max_seq_length: False
params_init: {
method: "xavier"
scale: 1.000001
seed: None
}
random_seed: None
remote: {
max_inflights_per_target: 32
shardable_batch: False
}
repeat_count: -1
require_sequential_order: False
skip_lp_regularization: None
source_max_length: 3000