forked from CGCL-codes/naturalcc
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpython.log
1858 lines (1858 loc) · 230 KB
/
python.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
nohup: ignoring input
Using backend: pytorch
[2021-03-22 05:23:45] INFO >> Load arguments in /home/wanyao/yang/naturalcc-dev/run/summarization/seq2seq/config/python_wan/python.yml (train.py:299, cli_main())
[2021-03-22 05:23:45] INFO >> {'criterion': 'cross_entropy', 'optimizer': 'torch_adam', 'lr_scheduler': 'fixed', 'tokenizer': None, 'bpe': None, 'common': {'no_progress_bar': 0, 'log_interval': 500, 'log_format': 'simple', 'tensorboard_logdir': '', 'memory_efficient_fp16': 1, 'fp16_no_flatten_grads': 1, 'fp16_init_scale': 128, 'fp16_scale_window': None, 'fp16_scale_tolerance': 0.0, 'min_loss_scale': 0.0001, 'threshold_loss_scale': None, 'empty_cache_freq': 0, 'task': 'summarization', 'seed': 1, 'cpu': 0, 'fp16': 0, 'fp16_opt_level': '01', 'server_ip': '', 'server_port': ''}, 'dataset': {'num_workers': 3, 'skip_invalid_size_inputs_valid_test': 1, 'max_tokens': None, 'max_sentences': 64, 'required_batch_size_multiple': 1, 'dataset_impl': 'mmap', 'train_subset': 'train', 'valid_subset': 'valid', 'validate_interval': 1, 'fixed_validation_seed': None, 'disable_validation': 0, 'max_tokens_valid': None, 'max_sentences_valid': 2048, 'curriculum': 0, 'gen_subset': 'test', 'num_shards': 1, 'shard_id': 0}, 'distributed_training': {'distributed_world_size': 1, 'distributed_rank': 0, 'distributed_backend': 'nccl', 'distributed_init_method': None, 'distributed_port': -1, 'device_id': 0, 'distributed_no_spawn': 0, 'ddp_backend': 'c10d', 'bucket_cap_mb': 25, 'fix_batches_to_gpus': None, 'find_unused_parameters': 0, 'fast_stat_sync': 0, 'broadcast_buffers': 0, 'global_sync_iter': 50, 'warmup_iterations': 0, 'local_rank': -1, 'block_momentum': 0.875, 'block_lr': 1, 'use_nbm': 0, 'average_sync': 0}, 'task': {'data': '/mnt/wanyao/.ncc/python_wan/summarization/data-mmap', 'source_lang': 'code_tokens', 'target_lang': 'docstring_tokens', 'load_alignments': 0, 'left_pad_source': 0, 'left_pad_target': 0, 'max_source_positions': 400, 'max_target_positions': 30, 'upsample_primary': 1, 'truncate_source': 1, 'truncate_target': 1, 'append_eos_to_target': 1, 'eval_bleu': 1, 'eval_bleu_detok': 'space', 'eval_bleu_detok_args': None, 'eval_tokenized_bleu': 0, 'eval_bleu_remove_bpe': None, 'eval_bleu_args': None, 'eval_bleu_print_samples': 0}, 'model': {'arch': 'seq2seq', 'encoder_embed_dim': 512, 'encoder_embed': None, 'encoder_freeze_embed': 0, 'encoder_hidden_size': 512, 'encoder_layers': 1, 'encoder_bidirectional': 0, 'decoder_embed_dim': 512, 'decoder_embed': None, 'decoder_freeze_embed': None, 'decoder_hidden_size': 512, 'decoder_layers': 1, 'decoder_out_embed_dim': 512, 'decoder_attention': 1, 'adaptive_softmax_cutoff': None, 'share_decoder_input_output_embed': 0, 'share_all_embeddings': 0, 'encoder_dropout_in': 0.1, 'encoder_dropout_out': 0.2, 'decoder_dropout_in': 0.1, 'decoder_dropout_out': 0.2, 'max_source_positions': 400, 'max_target_positions': 30}, 'optimization': {'max_epoch': 200, 'max_update': 0, 'clip_norm': 25, 'update_freq': [1], 'lrs': [0.0001], 'min_lr': -1, 'use_bmuf': 0, 'force_anneal': 0, 'warmup_updates': 0, 'lr_shrink': 0.99, 'sentence_avg': 1, 'adam': {'adam_betas': '(0.9, 0.999)', 'adam_eps': 1e-08, 'weight_decay': 0.0}, 'weight_decay': 0.0, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 5, 'max_steps': -1, 'warmup_steps': 0, 'gradient_accumulation_steps': 1}, 'checkpoint': {'restore_file': 'checkpoint_last.pt', 'reset_dataloader': None, 'reset_lr_scheduler': None, 'reset_meters': None, 'reset_optimizer': None, 'optimizer_overrides': '{}', 'save_interval': 1, 'save_interval_updates': 0, 'keep_interval_updates': 0, 'keep_last_epochs': -1, 'keep_best_checkpoints': -1, 'no_save': 0, 'no_epoch_checkpoints': 1, 'no_last_checkpoints': 0, 'no_save_optimizer_state': None, 'best_checkpoint_metric': 'bleu', 'maximize_best_checkpoint_metric': 1, 'patience': 10, 'save_dir': '/mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints', 'should_continue': 0, 'model_name_or_path': None, 'cache_dir': None, 'logging_steps': 500, 'save_steps': 2000, 'save_total_limit': 2, 'overwrite_output_dir': 0, 'overwrite_cache': 0}, 'eval': {'path': '/mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt', 'result_path': None, 'remove_bpe': None, 'quiet': 0, 'model_overrides': '{}', 'max_sentences': 2048, 'beam': 1, 'nbest': 1, 'max_len_a': 0, 'max_len_b': 30, 'min_len': 1, 'match_source_len': 0, 'no_early_stop': 0, 'unnormalized': 0, 'no_beamable_mm': 0, 'lenpen': 1, 'unkpen': 0, 'replace_unk': None, 'sacrebleu': 0, 'score_reference': 0, 'prefix_size': 0, 'no_repeat_ngram_size': 0, 'sampling': 0, 'sampling_topk': -1, 'sampling_topp': -1, 'temperature': 1.0, 'diverse_beam_groups': -1, 'diverse_beam_strength': 0.5, 'diversity_rate': -1.0, 'print_alignment': 0, 'print_step': 0, 'iter_decode_eos_penalty': 0.0, 'iter_decode_max_iter': 10, 'iter_decode_force_max_iter': 0, 'iter_decode_with_beam': 1, 'iter_decode_with_external_reranker': 0, 'retain_iter_history': 0, 'decoding_format': None, 'nltk_bleu': 1, 'rouge': 1}} (train.py:301, cli_main())
[2021-03-22 05:23:45] INFO >> single GPU training... (train.py:330, cli_main())
[2021-03-22 05:23:45] INFO >> [code_tokens] dictionary: 50000 types (summarization.py:217, setup_task())
[2021-03-22 05:23:45] INFO >> [docstring_tokens] dictionary: 30000 types (summarization.py:218, setup_task())
[2021-03-22 05:23:45] INFO >> truncate valid.code_tokens to 400 (summarization.py:126, load_langpair_dataset())
[2021-03-22 05:23:45] INFO >> truncate valid.docstring_tokens to 30 (summarization.py:137, load_langpair_dataset())
[2021-03-22 05:23:45] INFO >> loaded 18505 examples from: /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/valid.code_tokens (summarization.py:165, load_langpair_dataset())
[2021-03-22 05:23:45] INFO >> loaded 18505 examples from: /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/valid.docstring_tokens (summarization.py:166, load_langpair_dataset())
[2021-03-22 05:23:46] INFO >> Seq2SeqModel(
(encoder): LSTMEncoder(
(embed_tokens): Embedding(50000, 512, padding_idx=0)
(lstm): LSTM(512, 512)
)
(decoder): LSTMDecoder(
(embed_tokens): Embedding(30000, 512, padding_idx=0)
(layers): ModuleList(
(0): LSTMCell(1024, 512)
)
(attention): AttentionLayer(
(input_proj): Linear(in_features=512, out_features=512, bias=False)
(output_proj): Linear(in_features=1024, out_features=512, bias=False)
)
(fc_out): Linear(in_features=512, out_features=30000, bias=True)
)
) (train.py:220, single_main())
[2021-03-22 05:23:46] INFO >> model seq2seq, criterion CrossEntropyCriterion (train.py:221, single_main())
[2021-03-22 05:23:46] INFO >> num. model params: 62387504 (num. trained: 62387504) (train.py:222, single_main())
[2021-03-22 05:23:59] INFO >> training on 1 GPUs (train.py:229, single_main())
[2021-03-22 05:23:59] INFO >> max tokens per GPU = None and max sentences per GPU = 64 (train.py:230, single_main())
[2021-03-22 05:23:59] INFO >> no existing checkpoint found /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (ncc_trainer.py:269, load_checkpoint())
[2021-03-22 05:23:59] INFO >> loading train data for epoch 1 (ncc_trainer.py:283, get_train_iterator())
[2021-03-22 05:23:59] INFO >> truncate train.code_tokens to 400 (summarization.py:126, load_langpair_dataset())
[2021-03-22 05:23:59] INFO >> truncate train.docstring_tokens to 30 (summarization.py:137, load_langpair_dataset())
[2021-03-22 05:23:59] INFO >> loaded 55538 examples from: /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/train.code_tokens (summarization.py:165, load_langpair_dataset())
[2021-03-22 05:23:59] INFO >> loaded 55538 examples from: /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/train.docstring_tokens (summarization.py:166, load_langpair_dataset())
[2021-03-22 05:24:00] INFO >> NOTE: your device may support faster training with fp16 (ncc_trainer.py:154, _setup_optimizer())
/home/wanyao/yang/naturalcc-dev/ncc/utils/utils.py:574: UserWarning: amp_C fused kernels unavailable, disabling multi_tensor_l2norm; you may get better performance by installing NVIDIA's apex library
warnings.warn(
[2021-03-22 05:24:54] INFO >> epoch 001: 500 / 868 loss=90.972, nll_loss=8.748, bleu=0, ppl=430.04, wps=7062.8, ups=10.61, wpb=665.2, bsz=64, num_updates=500, lr=0.0001, gnorm=9.687, clip=0, train_wall=45, wall=55 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:25:28] INFO >> epoch 001 | loss 86.618 | nll_loss 8.314 | bleu 0 | ppl 318.25 | wps 7093.6 | ups 10.64 | wpb 666.6 | bsz 64 | num_updates 868 | lr 0.0001 | gnorm 9.132 | clip 0 | train_wall 78 | wall 89 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:26:00] INFO >> epoch 001 | valid on 'valid' subset | loss 78.352 | nll_loss 7.507 | bleu 14.5068 | ppl 181.87 | wps 7838.3 | wpb 19314.6 | bsz 1850.5 | num_updates 868 (progress_bar.py:269, print())
[2021-03-22 05:26:03] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 1 @ 868 updates, score 14.506764370901527) (writing took 3.104402 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:26:21] INFO >> epoch 002: 132 / 868 loss=79.836, nll_loss=7.656, bleu=0, ppl=201.69, wps=3810.3, ups=5.71, wpb=667.4, bsz=64, num_updates=1000, lr=9.9e-05, gnorm=8.371, clip=0, train_wall=44, wall=142 (progress_bar.py:260, log())
[2021-03-22 05:27:08] INFO >> epoch 002: 632 / 868 loss=76.641, nll_loss=7.349, bleu=0, ppl=163, wps=7208.6, ups=10.8, wpb=667.5, bsz=64, num_updates=1500, lr=9.9e-05, gnorm=8.56, clip=0, train_wall=44, wall=189 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:27:30] INFO >> epoch 002 | loss 76.293 | nll_loss 7.323 | bleu 0 | ppl 160.12 | wps 4751.4 | ups 7.13 | wpb 666.6 | bsz 64 | num_updates 1736 | lr 9.9e-05 | gnorm 8.606 | clip 0 | train_wall 76 | wall 211 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:28:00] INFO >> epoch 002 | valid on 'valid' subset | loss 74.975 | nll_loss 7.183 | bleu 14.5156 | ppl 145.33 | wps 8130.7 | wpb 19314.6 | bsz 1850.5 | num_updates 1736 | best_bleu 14.5156 (progress_bar.py:269, print())
[2021-03-22 05:28:39] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 2 @ 1736 updates, score 14.515613889194771) (writing took 38.707468 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:29:10] INFO >> epoch 003: 264 / 868 loss=73.945, nll_loss=7.107, bleu=0, ppl=137.88, wps=2725.8, ups=4.1, wpb=665.6, bsz=64, num_updates=2000, lr=9.8e-05, gnorm=9.005, clip=0, train_wall=44, wall=311 (progress_bar.py:260, log())
[2021-03-22 05:29:56] INFO >> epoch 003: 764 / 868 loss=72.426, nll_loss=6.94, bleu=0, ppl=122.81, wps=7170.7, ups=10.74, wpb=667.6, bsz=64, num_updates=2500, lr=9.8e-05, gnorm=9.541, clip=0, train_wall=44, wall=357 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:30:06] INFO >> epoch 003 | loss 72.518 | nll_loss 6.961 | bleu 0 | ppl 124.56 | wps 3700.8 | ups 5.55 | wpb 666.6 | bsz 64 | num_updates 2604 | lr 9.8e-05 | gnorm 9.447 | clip 0 | train_wall 77 | wall 367 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:30:38] INFO >> epoch 003 | valid on 'valid' subset | loss 72.579 | nll_loss 6.954 | bleu 14.6902 | ppl 123.96 | wps 7668.7 | wpb 19314.6 | bsz 1850.5 | num_updates 2604 | best_bleu 14.6902 (progress_bar.py:269, print())
[2021-03-22 05:31:16] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 3 @ 2604 updates, score 14.690214884410386) (writing took 38.439112 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:32:01] INFO >> epoch 004: 396 / 868 loss=70.436, nll_loss=6.76, bleu=0, ppl=108.37, wps=2681.1, ups=4.02, wpb=666.9, bsz=64, num_updates=3000, lr=9.7e-05, gnorm=10.003, clip=0, train_wall=45, wall=482 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:32:49] INFO >> epoch 004 | loss 69.534 | nll_loss 6.674 | bleu 0 | ppl 102.13 | wps 3566.8 | ups 5.35 | wpb 666.6 | bsz 64 | num_updates 3472 | lr 9.7e-05 | gnorm 10.295 | clip 0 | train_wall 81 | wall 529 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:33:20] INFO >> epoch 004 | valid on 'valid' subset | loss 70.677 | nll_loss 6.771 | bleu 15.2081 | ppl 109.25 | wps 7652.2 | wpb 19314.6 | bsz 1850.5 | num_updates 3472 | best_bleu 15.2081 (progress_bar.py:269, print())
[2021-03-22 05:33:58] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 4 @ 3472 updates, score 15.20814545285317) (writing took 38.128219 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:34:07] INFO >> epoch 005: 28 / 868 loss=68.983, nll_loss=6.62, bleu=0, ppl=98.34, wps=2631.9, ups=3.95, wpb=666.6, bsz=64, num_updates=3500, lr=9.6e-05, gnorm=10.493, clip=0, train_wall=48, wall=608 (progress_bar.py:260, log())
[2021-03-22 05:34:55] INFO >> epoch 005: 528 / 868 loss=67.231, nll_loss=6.456, bleu=0, ppl=87.8, wps=7049, ups=10.58, wpb=666.2, bsz=64, num_updates=4000, lr=9.6e-05, gnorm=10.934, clip=0, train_wall=45, wall=655 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:35:27] INFO >> epoch 005 | loss 66.944 | nll_loss 6.426 | bleu 0 | ppl 85.96 | wps 3660.6 | ups 5.49 | wpb 666.6 | bsz 64 | num_updates 4340 | lr 9.6e-05 | gnorm 11.064 | clip 0 | train_wall 78 | wall 688 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:35:59] INFO >> epoch 005 | valid on 'valid' subset | loss 69.218 | nll_loss 6.632 | bleu 15.5335 | ppl 99.16 | wps 7557.3 | wpb 19314.6 | bsz 1850.5 | num_updates 4340 | best_bleu 15.5335 (progress_bar.py:269, print())
[2021-03-22 05:36:37] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 5 @ 4340 updates, score 15.533457705045956) (writing took 38.625007 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:36:59] INFO >> epoch 006: 160 / 868 loss=65.996, nll_loss=6.339, bleu=0, ppl=80.96, wps=2684, ups=4.03, wpb=666, bsz=64, num_updates=4500, lr=9.5e-05, gnorm=11.364, clip=0, train_wall=44, wall=779 (progress_bar.py:260, log())
[2021-03-22 05:37:45] INFO >> epoch 006: 660 / 868 loss=64.675, nll_loss=6.209, bleu=0, ppl=73.99, wps=7174, ups=10.76, wpb=666.6, bsz=64, num_updates=5000, lr=9.5e-05, gnorm=11.799, clip=0, train_wall=44, wall=826 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:38:05] INFO >> epoch 006 | loss 64.663 | nll_loss 6.207 | bleu 0 | ppl 73.86 | wps 3657.8 | ups 5.49 | wpb 666.6 | bsz 64 | num_updates 5208 | lr 9.5e-05 | gnorm 11.798 | clip 0 | train_wall 77 | wall 846 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:38:37] INFO >> epoch 006 | valid on 'valid' subset | loss 68.033 | nll_loss 6.518 | bleu 15.8358 | ppl 91.65 | wps 7496.5 | wpb 19314.6 | bsz 1850.5 | num_updates 5208 | best_bleu 15.8358 (progress_bar.py:269, print())
[2021-03-22 05:39:34] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 6 @ 5208 updates, score 15.835844852285533) (writing took 57.536121 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:40:09] INFO >> epoch 007: 292 / 868 loss=63.316, nll_loss=6.078, bleu=0, ppl=67.54, wps=2313.3, ups=3.47, wpb=666.7, bsz=64, num_updates=5500, lr=9.4e-05, gnorm=12.069, clip=0, train_wall=46, wall=970 (progress_bar.py:260, log())
[2021-03-22 05:40:56] INFO >> epoch 007: 792 / 868 loss=62.671, nll_loss=6.008, bleu=0, ppl=64.38, wps=7135.2, ups=10.69, wpb=667.3, bsz=64, num_updates=6000, lr=9.4e-05, gnorm=12.562, clip=0, train_wall=44, wall=1017 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:41:03] INFO >> epoch 007 | loss 62.583 | nll_loss 6.007 | bleu 0 | ppl 64.31 | wps 3240 | ups 4.86 | wpb 666.6 | bsz 64 | num_updates 6076 | lr 9.4e-05 | gnorm 12.421 | clip 0 | train_wall 78 | wall 1024 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:41:36] INFO >> epoch 007 | valid on 'valid' subset | loss 66.924 | nll_loss 6.412 | bleu 16.0404 | ppl 85.15 | wps 7474.5 | wpb 19314.6 | bsz 1850.5 | num_updates 6076 | best_bleu 16.0404 (progress_bar.py:269, print())
[2021-03-22 05:42:14] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 7 @ 6076 updates, score 16.04039909146089) (writing took 38.376298 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:43:00] INFO >> epoch 008: 424 / 868 loss=61.176, nll_loss=5.861, bleu=0, ppl=58.12, wps=2683.5, ups=4.02, wpb=667.7, bsz=64, num_updates=6500, lr=9.3e-05, gnorm=12.884, clip=0, train_wall=45, wall=1141 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:43:43] INFO >> epoch 008 | loss 60.662 | nll_loss 5.823 | bleu 0 | ppl 56.6 | wps 3626.1 | ups 5.44 | wpb 666.6 | bsz 64 | num_updates 6944 | lr 9.3e-05 | gnorm 13.059 | clip 0 | train_wall 78 | wall 1184 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:44:14] INFO >> epoch 008 | valid on 'valid' subset | loss 66.079 | nll_loss 6.331 | bleu 16.6465 | ppl 80.5 | wps 7680.7 | wpb 19314.6 | bsz 1850.5 | num_updates 6944 | best_bleu 16.6465 (progress_bar.py:269, print())
[2021-03-22 05:44:52] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 8 @ 6944 updates, score 16.64646392512173) (writing took 38.093135 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:45:04] INFO >> epoch 009: 56 / 868 loss=60.132, nll_loss=5.786, bleu=0, ppl=55.19, wps=2696.2, ups=4.05, wpb=665.1, bsz=64, num_updates=7000, lr=9.2e-05, gnorm=13.199, clip=0, train_wall=45, wall=1265 (progress_bar.py:260, log())
[2021-03-22 05:45:50] INFO >> epoch 009: 556 / 868 loss=59.065, nll_loss=5.664, bleu=0, ppl=50.7, wps=7182.2, ups=10.77, wpb=667.1, bsz=64, num_updates=7500, lr=9.2e-05, gnorm=13.552, clip=0, train_wall=44, wall=1311 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:46:20] INFO >> epoch 009 | loss 58.889 | nll_loss 5.653 | bleu 0 | ppl 50.3 | wps 3694.2 | ups 5.54 | wpb 666.6 | bsz 64 | num_updates 7812 | lr 9.2e-05 | gnorm 13.617 | clip 0 | train_wall 76 | wall 1341 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:46:51] INFO >> epoch 009 | valid on 'valid' subset | loss 65.311 | nll_loss 6.257 | bleu 16.934 | ppl 76.5 | wps 7662.1 | wpb 19314.6 | bsz 1850.5 | num_updates 7812 | best_bleu 16.934 (progress_bar.py:269, print())
[2021-03-22 05:47:30] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 9 @ 7812 updates, score 16.933971915675862) (writing took 38.297043 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:47:55] INFO >> epoch 010: 188 / 868 loss=58.295, nll_loss=5.598, bleu=0, ppl=48.43, wps=2678.2, ups=4.02, wpb=666.5, bsz=64, num_updates=8000, lr=9.1e-05, gnorm=13.807, clip=0, train_wall=45, wall=1435 (progress_bar.py:260, log())
[2021-03-22 05:48:45] INFO >> epoch 010: 688 / 868 loss=57.084, nll_loss=5.48, bleu=0, ppl=44.63, wps=6616.8, ups=9.93, wpb=666.4, bsz=64, num_updates=8500, lr=9.1e-05, gnorm=14.217, clip=0, train_wall=48, wall=1486 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:49:04] INFO >> epoch 010 | loss 57.245 | nll_loss 5.495 | bleu 0 | ppl 45.09 | wps 3524.9 | ups 5.29 | wpb 666.6 | bsz 64 | num_updates 8680 | lr 9.1e-05 | gnorm 14.169 | clip 0 | train_wall 83 | wall 1505 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:49:35] INFO >> epoch 010 | valid on 'valid' subset | loss 64.762 | nll_loss 6.205 | bleu 17.0337 | ppl 73.76 | wps 7707 | wpb 19314.6 | bsz 1850.5 | num_updates 8680 | best_bleu 17.0337 (progress_bar.py:269, print())
[2021-03-22 05:50:14] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 10 @ 8680 updates, score 17.03374231718647) (writing took 38.811766 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:50:52] INFO >> epoch 011: 320 / 868 loss=56.362, nll_loss=5.408, bleu=0, ppl=42.47, wps=2628, ups=3.94, wpb=666.7, bsz=64, num_updates=9000, lr=9e-05, gnorm=14.429, clip=0, train_wall=48, wall=1613 (progress_bar.py:260, log())
[2021-03-22 05:51:42] INFO >> epoch 011: 820 / 868 loss=55.544, nll_loss=5.342, bleu=0, ppl=40.55, wps=6620.7, ups=9.95, wpb=665.5, bsz=64, num_updates=9500, lr=9e-05, gnorm=14.717, clip=0, train_wall=48, wall=1663 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:51:47] INFO >> epoch 011 | loss 55.729 | nll_loss 5.349 | bleu 0 | ppl 40.76 | wps 3537.5 | ups 5.31 | wpb 666.6 | bsz 64 | num_updates 9548 | lr 9e-05 | gnorm 14.639 | clip 0 | train_wall 83 | wall 1668 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:52:19] INFO >> epoch 011 | valid on 'valid' subset | loss 64.165 | nll_loss 6.148 | bleu 17.2365 | ppl 70.89 | wps 7669.1 | wpb 19314.6 | bsz 1850.5 | num_updates 9548 | best_bleu 17.2365 (progress_bar.py:269, print())
[2021-03-22 05:52:57] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 11 @ 9548 updates, score 17.2365046800012) (writing took 38.694162 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:53:50] INFO >> epoch 012: 452 / 868 loss=54.292, nll_loss=5.214, bleu=0, ppl=37.11, wps=2609, ups=3.92, wpb=666.2, bsz=64, num_updates=10000, lr=9e-05, gnorm=14.974, clip=0, train_wall=49, wall=1791 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:54:32] INFO >> epoch 012 | loss 54.301 | nll_loss 5.212 | bleu 0 | ppl 37.07 | wps 3507.3 | ups 5.26 | wpb 666.6 | bsz 64 | num_updates 10416 | lr 9e-05 | gnorm 15.124 | clip 0 | train_wall 84 | wall 1833 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:55:04] INFO >> epoch 012 | valid on 'valid' subset | loss 63.714 | nll_loss 6.104 | bleu 17.5874 | ppl 68.8 | wps 7674.8 | wpb 19314.6 | bsz 1850.5 | num_updates 10416 | best_bleu 17.5874 (progress_bar.py:269, print())
[2021-03-22 05:55:43] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 12 @ 10416 updates, score 17.587392782502327) (writing took 39.083103 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:55:57] INFO >> epoch 013: 84 / 868 loss=54.254, nll_loss=5.197, bleu=0, ppl=36.67, wps=2623.8, ups=3.93, wpb=668.2, bsz=64, num_updates=10500, lr=8.9e-05, gnorm=15.258, clip=0, train_wall=48, wall=1918 (progress_bar.py:260, log())
[2021-03-22 05:56:44] INFO >> epoch 013: 584 / 868 loss=52.909, nll_loss=5.081, bleu=0, ppl=33.86, wps=7170.7, ups=10.77, wpb=666.1, bsz=64, num_updates=11000, lr=8.9e-05, gnorm=15.499, clip=0, train_wall=44, wall=1964 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:57:10] INFO >> epoch 013 | loss 52.952 | nll_loss 5.083 | bleu 0 | ppl 33.89 | wps 3661 | ups 5.49 | wpb 666.6 | bsz 64 | num_updates 11284 | lr 8.9e-05 | gnorm 15.542 | clip 0 | train_wall 77 | wall 1991 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:57:42] INFO >> epoch 013 | valid on 'valid' subset | loss 63.406 | nll_loss 6.075 | bleu 17.7958 | ppl 67.41 | wps 7627.4 | wpb 19314.6 | bsz 1850.5 | num_updates 11284 | best_bleu 17.7958 (progress_bar.py:269, print())
[2021-03-22 05:58:21] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 13 @ 11284 updates, score 17.795776287801022) (writing took 38.579079 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 05:58:47] INFO >> epoch 014: 216 / 868 loss=52.396, nll_loss=5.027, bleu=0, ppl=32.61, wps=2706.8, ups=4.06, wpb=667.1, bsz=64, num_updates=11500, lr=8.8e-05, gnorm=15.74, clip=0, train_wall=44, wall=2088 (progress_bar.py:260, log())
[2021-03-22 05:59:33] INFO >> epoch 014: 716 / 868 loss=51.714, nll_loss=4.958, bleu=0, ppl=31.09, wps=7188.4, ups=10.77, wpb=667.2, bsz=64, num_updates=12000, lr=8.8e-05, gnorm=16.043, clip=0, train_wall=44, wall=2134 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 05:59:48] INFO >> epoch 014 | loss 51.656 | nll_loss 4.958 | bleu 0 | ppl 31.09 | wps 3680.8 | ups 5.52 | wpb 666.6 | bsz 64 | num_updates 12152 | lr 8.8e-05 | gnorm 15.978 | clip 0 | train_wall 76 | wall 2148 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:00:19] INFO >> epoch 014 | valid on 'valid' subset | loss 63.01 | nll_loss 6.037 | bleu 17.8829 | ppl 65.66 | wps 7647 | wpb 19314.6 | bsz 1850.5 | num_updates 12152 | best_bleu 17.8829 (progress_bar.py:269, print())
[2021-03-22 06:01:34] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 14 @ 12152 updates, score 17.882899955178026) (writing took 74.840503 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:02:14] INFO >> epoch 015: 348 / 868 loss=50.802, nll_loss=4.882, bleu=0, ppl=29.48, wps=2077.2, ups=3.12, wpb=666, bsz=64, num_updates=12500, lr=8.7e-05, gnorm=16.21, clip=0, train_wall=45, wall=2294 (progress_bar.py:260, log())
[2021-03-22 06:03:00] INFO >> epoch 015: 848 / 868 loss=50.418, nll_loss=4.841, bleu=0, ppl=28.67, wps=7161, ups=10.75, wpb=666.2, bsz=64, num_updates=13000, lr=8.7e-05, gnorm=16.447, clip=0, train_wall=44, wall=2341 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:03:02] INFO >> epoch 015 | loss 50.444 | nll_loss 4.842 | bleu 0 | ppl 28.68 | wps 2971.6 | ups 4.46 | wpb 666.6 | bsz 64 | num_updates 13020 | lr 8.7e-05 | gnorm 16.378 | clip 0 | train_wall 78 | wall 2343 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:03:34] INFO >> epoch 015 | valid on 'valid' subset | loss 62.794 | nll_loss 6.016 | bleu 18.2448 | ppl 64.72 | wps 7673.8 | wpb 19314.6 | bsz 1850.5 | num_updates 13020 | best_bleu 18.2448 (progress_bar.py:269, print())
[2021-03-22 06:04:12] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 15 @ 13020 updates, score 18.2447608645696) (writing took 38.683452 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:05:06] INFO >> epoch 016: 480 / 868 loss=49.175, nll_loss=4.729, bleu=0, ppl=26.52, wps=2643.3, ups=3.97, wpb=665.2, bsz=64, num_updates=13500, lr=8.6e-05, gnorm=16.619, clip=0, train_wall=47, wall=2467 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:05:42] INFO >> epoch 016 | loss 49.268 | nll_loss 4.729 | bleu 0 | ppl 26.52 | wps 3628.1 | ups 5.44 | wpb 666.6 | bsz 64 | num_updates 13888 | lr 8.6e-05 | gnorm 16.766 | clip 0 | train_wall 79 | wall 2503 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:06:13] INFO >> epoch 016 | valid on 'valid' subset | loss 62.542 | nll_loss 5.992 | bleu 18.2051 | ppl 63.65 | wps 7652.9 | wpb 19314.6 | bsz 1850.5 | num_updates 13888 | best_bleu 18.2448 (progress_bar.py:269, print())
[2021-03-22 06:06:33] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 16 @ 13888 updates, score 18.205145599584306) (writing took 19.965793 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:06:50] INFO >> epoch 017: 112 / 868 loss=49.184, nll_loss=4.707, bleu=0, ppl=26.12, wps=3200.3, ups=4.79, wpb=668.7, bsz=64, num_updates=14000, lr=8.5e-05, gnorm=16.939, clip=0, train_wall=44, wall=2571 (progress_bar.py:260, log())
[2021-03-22 06:07:39] INFO >> epoch 017: 612 / 868 loss=48.101, nll_loss=4.611, bleu=0, ppl=24.43, wps=6915.7, ups=10.36, wpb=667.4, bsz=64, num_updates=14500, lr=8.5e-05, gnorm=17.098, clip=0, train_wall=46, wall=2619 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:08:04] INFO >> epoch 017 | loss 48.144 | nll_loss 4.621 | bleu 0 | ppl 24.61 | wps 4070.3 | ups 6.11 | wpb 666.6 | bsz 64 | num_updates 14756 | lr 8.5e-05 | gnorm 17.149 | clip 0 | train_wall 81 | wall 2645 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:08:35] INFO >> epoch 017 | valid on 'valid' subset | loss 62.372 | nll_loss 5.976 | bleu 18.362 | ppl 62.94 | wps 7717.8 | wpb 19314.6 | bsz 1850.5 | num_updates 14756 | best_bleu 18.362 (progress_bar.py:269, print())
[2021-03-22 06:09:14] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 17 @ 14756 updates, score 18.36203728114444) (writing took 39.044030 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:09:42] INFO >> epoch 018: 244 / 868 loss=47.623, nll_loss=4.582, bleu=0, ppl=23.94, wps=2684.7, ups=4.04, wpb=665.2, bsz=64, num_updates=15000, lr=8.4e-05, gnorm=17.315, clip=0, train_wall=45, wall=2743 (progress_bar.py:260, log())
[2021-03-22 06:10:26] INFO >> epoch 018: 744 / 868 loss=47.175, nll_loss=4.525, bleu=0, ppl=23.02, wps=7696.2, ups=11.54, wpb=667, bsz=64, num_updates=15500, lr=8.4e-05, gnorm=17.565, clip=0, train_wall=41, wall=2787 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:10:37] INFO >> epoch 018 | loss 47.078 | nll_loss 4.519 | bleu 0 | ppl 22.92 | wps 3789.4 | ups 5.68 | wpb 666.6 | bsz 64 | num_updates 15624 | lr 8.4e-05 | gnorm 17.481 | clip 0 | train_wall 73 | wall 2797 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:11:08] INFO >> epoch 018 | valid on 'valid' subset | loss 62.1 | nll_loss 5.95 | bleu 18.7221 | ppl 61.81 | wps 7778.2 | wpb 19314.6 | bsz 1850.5 | num_updates 15624 | best_bleu 18.7221 (progress_bar.py:269, print())
[2021-03-22 06:11:55] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 18 @ 15624 updates, score 18.72210215571159) (writing took 46.751818 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:12:35] INFO >> epoch 019: 376 / 868 loss=46.066, nll_loss=4.431, bleu=0, ppl=21.57, wps=2567, ups=3.86, wpb=665.1, bsz=64, num_updates=16000, lr=8.3e-05, gnorm=17.658, clip=0, train_wall=43, wall=2916 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:13:21] INFO >> epoch 019 | loss 46.075 | nll_loss 4.422 | bleu 0 | ppl 21.44 | wps 3531.1 | ups 5.3 | wpb 666.6 | bsz 64 | num_updates 16492 | lr 8.3e-05 | gnorm 17.836 | clip 0 | train_wall 75 | wall 2961 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:13:52] INFO >> epoch 019 | valid on 'valid' subset | loss 61.898 | nll_loss 5.93 | bleu 18.8528 | ppl 60.98 | wps 7594.5 | wpb 19314.6 | bsz 1850.5 | num_updates 16492 | best_bleu 18.8528 (progress_bar.py:269, print())
[2021-03-22 06:14:31] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 19 @ 16492 updates, score 18.85278803130051) (writing took 38.785309 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:14:38] INFO >> epoch 020: 8 / 868 loss=46.249, nll_loss=4.433, bleu=0, ppl=21.61, wps=2724.3, ups=4.08, wpb=667.6, bsz=64, num_updates=16500, lr=8.3e-05, gnorm=17.959, clip=0, train_wall=43, wall=3039 (progress_bar.py:260, log())
[2021-03-22 06:15:22] INFO >> epoch 020: 508 / 868 loss=44.832, nll_loss=4.308, bleu=0, ppl=19.81, wps=7595.3, ups=11.4, wpb=666, bsz=64, num_updates=17000, lr=8.3e-05, gnorm=17.952, clip=0, train_wall=42, wall=3083 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:15:54] INFO >> epoch 020 | loss 45.07 | nll_loss 4.326 | bleu 0 | ppl 20.06 | wps 3776.2 | ups 5.66 | wpb 666.6 | bsz 64 | num_updates 17360 | lr 8.3e-05 | gnorm 18.135 | clip 0 | train_wall 73 | wall 3115 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:16:26] INFO >> epoch 020 | valid on 'valid' subset | loss 61.737 | nll_loss 5.915 | bleu 19.0734 | ppl 60.34 | wps 7633 | wpb 19314.6 | bsz 1850.5 | num_updates 17360 | best_bleu 19.0734 (progress_bar.py:269, print())
[2021-03-22 06:17:32] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 20 @ 17360 updates, score 19.07340705461433) (writing took 66.004944 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:17:51] INFO >> epoch 021: 140 / 868 loss=44.975, nll_loss=4.311, bleu=0, ppl=19.85, wps=2231.8, ups=3.35, wpb=667.1, bsz=63.9, num_updates=17500, lr=8.2e-05, gnorm=18.342, clip=0, train_wall=43, wall=3232 (progress_bar.py:260, log())
[2021-03-22 06:18:38] INFO >> epoch 021: 640 / 868 loss=44.191, nll_loss=4.239, bleu=0, ppl=18.88, wps=7181, ups=10.76, wpb=667.2, bsz=64, num_updates=18000, lr=8.2e-05, gnorm=18.459, clip=0, train_wall=44, wall=3278 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:18:59] INFO >> epoch 021 | loss 44.16 | nll_loss 4.239 | bleu 0 | ppl 18.88 | wps 3116.1 | ups 4.67 | wpb 666.6 | bsz 64 | num_updates 18228 | lr 8.2e-05 | gnorm 18.48 | clip 0 | train_wall 77 | wall 3300 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:19:31] INFO >> epoch 021 | valid on 'valid' subset | loss 61.633 | nll_loss 5.905 | bleu 18.97 | ppl 59.92 | wps 7675.9 | wpb 19314.6 | bsz 1850.5 | num_updates 18228 | best_bleu 19.0734 (progress_bar.py:269, print())
[2021-03-22 06:19:52] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 21 @ 18228 updates, score 18.969962310408263) (writing took 20.797637 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:20:23] INFO >> epoch 022: 272 / 868 loss=43.75, nll_loss=4.19, bleu=0, ppl=18.26, wps=3182.3, ups=4.76, wpb=668.2, bsz=64, num_updates=18500, lr=8.1e-05, gnorm=18.641, clip=0, train_wall=44, wall=3383 (progress_bar.py:260, log())
[2021-03-22 06:21:07] INFO >> epoch 022: 772 / 868 loss=43.206, nll_loss=4.153, bleu=0, ppl=17.79, wps=7561.9, ups=11.36, wpb=665.6, bsz=64, num_updates=19000, lr=8.1e-05, gnorm=18.728, clip=0, train_wall=42, wall=3427 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:21:15] INFO >> epoch 022 | loss 43.199 | nll_loss 4.146 | bleu 0 | ppl 17.71 | wps 4273.7 | ups 6.41 | wpb 666.6 | bsz 64 | num_updates 19096 | lr 8.1e-05 | gnorm 18.693 | clip 0 | train_wall 73 | wall 3436 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:21:47] INFO >> epoch 022 | valid on 'valid' subset | loss 61.519 | nll_loss 5.894 | bleu 19.2171 | ppl 59.47 | wps 7615.8 | wpb 19314.6 | bsz 1850.5 | num_updates 19096 | best_bleu 19.2171 (progress_bar.py:269, print())
[2021-03-22 06:22:31] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 22 @ 19096 updates, score 19.217069525323346) (writing took 44.388451 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:23:14] INFO >> epoch 023: 404 / 868 loss=42.347, nll_loss=4.068, bleu=0, ppl=16.77, wps=2612.2, ups=3.92, wpb=666, bsz=64, num_updates=19500, lr=8e-05, gnorm=18.846, clip=0, train_wall=43, wall=3555 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:23:57] INFO >> epoch 023 | loss 42.339 | nll_loss 4.064 | bleu 0 | ppl 16.73 | wps 3571 | ups 5.36 | wpb 666.6 | bsz 64 | num_updates 19964 | lr 8e-05 | gnorm 19.044 | clip 0 | train_wall 75 | wall 3598 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:24:28] INFO >> epoch 023 | valid on 'valid' subset | loss 61.497 | nll_loss 5.892 | bleu 19.4362 | ppl 59.38 | wps 7763.1 | wpb 19314.6 | bsz 1850.5 | num_updates 19964 | best_bleu 19.4362 (progress_bar.py:269, print())
[2021-03-22 06:25:07] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 23 @ 19964 updates, score 19.436222451445822) (writing took 38.691130 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:25:16] INFO >> epoch 024: 36 / 868 loss=42.333, nll_loss=4.067, bleu=0, ppl=16.76, wps=2726.6, ups=4.09, wpb=666.1, bsz=64, num_updates=20000, lr=7.9e-05, gnorm=19.169, clip=0, train_wall=43, wall=3677 (progress_bar.py:260, log())
[2021-03-22 06:26:02] INFO >> epoch 024: 536 / 868 loss=41.323, nll_loss=3.965, bleu=0, ppl=15.62, wps=7334.3, ups=11, wpb=666.6, bsz=64, num_updates=20500, lr=7.9e-05, gnorm=19.24, clip=0, train_wall=43, wall=3723 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:26:32] INFO >> epoch 024 | loss 41.485 | nll_loss 3.982 | bleu 0 | ppl 15.8 | wps 3721.1 | ups 5.58 | wpb 666.6 | bsz 64 | num_updates 20832 | lr 7.9e-05 | gnorm 19.312 | clip 0 | train_wall 75 | wall 3753 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:27:04] INFO >> epoch 024 | valid on 'valid' subset | loss 61.453 | nll_loss 5.888 | bleu 19.5531 | ppl 59.21 | wps 7687.6 | wpb 19314.6 | bsz 1850.5 | num_updates 20832 | best_bleu 19.5531 (progress_bar.py:269, print())
[2021-03-22 06:27:43] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 24 @ 20832 updates, score 19.553056256992324) (writing took 38.930612 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:28:05] INFO >> epoch 025: 168 / 868 loss=41.385, nll_loss=3.967, bleu=0, ppl=15.64, wps=2709, ups=4.06, wpb=667.7, bsz=64, num_updates=21000, lr=7.9e-05, gnorm=19.452, clip=0, train_wall=44, wall=3846 (progress_bar.py:260, log())
[2021-03-22 06:28:50] INFO >> epoch 025: 668 / 868 loss=40.739, nll_loss=3.906, bleu=0, ppl=14.99, wps=7361.4, ups=11.03, wpb=667.5, bsz=64, num_updates=21500, lr=7.9e-05, gnorm=19.649, clip=0, train_wall=43, wall=3891 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:29:09] INFO >> epoch 025 | loss 40.683 | nll_loss 3.905 | bleu 0 | ppl 14.98 | wps 3698.6 | ups 5.55 | wpb 666.6 | bsz 64 | num_updates 21700 | lr 7.9e-05 | gnorm 19.608 | clip 0 | train_wall 75 | wall 3910 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:29:40] INFO >> epoch 025 | valid on 'valid' subset | loss 61.228 | nll_loss 5.866 | bleu 19.4508 | ppl 58.33 | wps 7667.6 | wpb 19314.6 | bsz 1850.5 | num_updates 21700 | best_bleu 19.5531 (progress_bar.py:269, print())
[2021-03-22 06:30:00] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 25 @ 21700 updates, score 19.45080008874993) (writing took 20.043590 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:30:35] INFO >> epoch 026: 300 / 868 loss=39.962, nll_loss=3.847, bleu=0, ppl=14.39, wps=3166, ups=4.76, wpb=664.4, bsz=64, num_updates=22000, lr=7.8e-05, gnorm=19.652, clip=0, train_wall=45, wall=3996 (progress_bar.py:260, log())
[2021-03-22 06:31:21] INFO >> epoch 026: 800 / 868 loss=39.993, nll_loss=3.841, bleu=0, ppl=14.33, wps=7275.4, ups=10.92, wpb=666.4, bsz=64, num_updates=22500, lr=7.8e-05, gnorm=19.815, clip=0, train_wall=44, wall=4042 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:31:28] INFO >> epoch 026 | loss 39.867 | nll_loss 3.827 | bleu 0 | ppl 14.19 | wps 4169.5 | ups 6.25 | wpb 666.6 | bsz 64 | num_updates 22568 | lr 7.8e-05 | gnorm 19.785 | clip 0 | train_wall 77 | wall 4048 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:31:59] INFO >> epoch 026 | valid on 'valid' subset | loss 60.999 | nll_loss 5.844 | bleu 19.6965 | ppl 57.45 | wps 7692.7 | wpb 19314.6 | bsz 1850.5 | num_updates 22568 | best_bleu 19.6965 (progress_bar.py:269, print())
[2021-03-22 06:32:38] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 26 @ 22568 updates, score 19.696549348787144) (writing took 38.885594 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:33:24] INFO >> epoch 027: 432 / 868 loss=39.141, nll_loss=3.755, bleu=0, ppl=13.5, wps=2705.8, ups=4.06, wpb=666.9, bsz=64, num_updates=23000, lr=7.7e-05, gnorm=20.01, clip=0, train_wall=44, wall=4165 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:34:07] INFO >> epoch 027 | loss 39.105 | nll_loss 3.754 | bleu 0 | ppl 13.49 | wps 3633.5 | ups 5.45 | wpb 666.6 | bsz 64 | num_updates 23436 | lr 7.7e-05 | gnorm 20.062 | clip 0 | train_wall 78 | wall 4208 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:34:44] INFO >> epoch 027 | valid on 'valid' subset | loss 60.995 | nll_loss 5.844 | bleu 19.7466 | ppl 57.43 | wps 6210.6 | wpb 19314.6 | bsz 1850.5 | num_updates 23436 | best_bleu 19.7466 (progress_bar.py:269, print())
[2021-03-22 06:35:23] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 27 @ 23436 updates, score 19.746581173920752) (writing took 38.359525 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:35:35] INFO >> epoch 028: 64 / 868 loss=39.076, nll_loss=3.751, bleu=0, ppl=13.46, wps=2545.3, ups=3.82, wpb=666.5, bsz=64, num_updates=23500, lr=7.6e-05, gnorm=20.112, clip=0, train_wall=46, wall=4296 (progress_bar.py:260, log())
[2021-03-22 06:36:23] INFO >> epoch 028: 564 / 868 loss=38.234, nll_loss=3.672, bleu=0, ppl=12.75, wps=6996.2, ups=10.5, wpb=666.1, bsz=64, num_updates=24000, lr=7.6e-05, gnorm=20.181, clip=0, train_wall=45, wall=4344 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:36:54] INFO >> epoch 028 | loss 38.34 | nll_loss 3.68 | bleu 0 | ppl 12.82 | wps 3468.1 | ups 5.2 | wpb 666.6 | bsz 64 | num_updates 24304 | lr 7.6e-05 | gnorm 20.215 | clip 0 | train_wall 80 | wall 4374 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:37:31] INFO >> epoch 028 | valid on 'valid' subset | loss 60.926 | nll_loss 5.837 | bleu 19.6731 | ppl 57.17 | wps 6357.3 | wpb 19314.6 | bsz 1850.5 | num_updates 24304 | best_bleu 19.7466 (progress_bar.py:269, print())
[2021-03-22 06:37:51] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 28 @ 24304 updates, score 19.67314648802258) (writing took 20.107380 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:38:15] INFO >> epoch 029: 196 / 868 loss=38.013, nll_loss=3.647, bleu=0, ppl=12.53, wps=2961.9, ups=4.44, wpb=666.8, bsz=64, num_updates=24500, lr=7.5e-05, gnorm=20.277, clip=0.2, train_wall=47, wall=4456 (progress_bar.py:260, log())
[2021-03-22 06:39:01] INFO >> epoch 029: 696 / 868 loss=37.804, nll_loss=3.621, bleu=0, ppl=12.3, wps=7259.9, ups=10.87, wpb=668.2, bsz=64, num_updates=25000, lr=7.5e-05, gnorm=20.585, clip=0.4, train_wall=44, wall=4502 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:39:17] INFO >> epoch 029 | loss 37.599 | nll_loss 3.609 | bleu 0 | ppl 12.2 | wps 4024.6 | ups 6.04 | wpb 666.6 | bsz 64 | num_updates 25172 | lr 7.5e-05 | gnorm 20.464 | clip 0.3 | train_wall 76 | wall 4518 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:39:55] INFO >> epoch 029 | valid on 'valid' subset | loss 60.955 | nll_loss 5.84 | bleu 20.0553 | ppl 57.28 | wps 6291.6 | wpb 19314.6 | bsz 1850.5 | num_updates 25172 | best_bleu 20.0553 (progress_bar.py:269, print())
[2021-03-22 06:40:34] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 29 @ 25172 updates, score 20.055337006958116) (writing took 38.574147 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:41:12] INFO >> epoch 030: 328 / 868 loss=37.176, nll_loss=3.565, bleu=0, ppl=11.83, wps=2562, ups=3.84, wpb=667.1, bsz=64, num_updates=25500, lr=7.5e-05, gnorm=20.555, clip=0, train_wall=45, wall=4632 (progress_bar.py:260, log())
[2021-03-22 06:41:57] INFO >> epoch 030: 828 / 868 loss=36.836, nll_loss=3.545, bleu=0, ppl=11.67, wps=7260.5, ups=10.92, wpb=665, bsz=64, num_updates=26000, lr=7.5e-05, gnorm=20.749, clip=0.2, train_wall=43, wall=4678 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:42:01] INFO >> epoch 030 | loss 36.922 | nll_loss 3.544 | bleu 0 | ppl 11.66 | wps 3527.7 | ups 5.29 | wpb 666.6 | bsz 64 | num_updates 26040 | lr 7.5e-05 | gnorm 20.71 | clip 0.1 | train_wall 77 | wall 4682 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:42:39] INFO >> epoch 030 | valid on 'valid' subset | loss 60.927 | nll_loss 5.837 | bleu 20.0727 | ppl 57.18 | wps 6137.3 | wpb 19314.6 | bsz 1850.5 | num_updates 26040 | best_bleu 20.0727 (progress_bar.py:269, print())
[2021-03-22 06:43:17] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 30 @ 26040 updates, score 20.072739754641848) (writing took 38.063426 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:44:06] INFO >> epoch 031: 460 / 868 loss=36.201, nll_loss=3.483, bleu=0, ppl=11.18, wps=2576.3, ups=3.87, wpb=665.2, bsz=64, num_updates=26500, lr=7.4e-05, gnorm=20.742, clip=0, train_wall=44, wall=4807 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:44:41] INFO >> epoch 031 | loss 36.22 | nll_loss 3.477 | bleu 0 | ppl 11.13 | wps 3621.8 | ups 5.43 | wpb 666.6 | bsz 64 | num_updates 26908 | lr 7.4e-05 | gnorm 20.844 | clip 0.2 | train_wall 73 | wall 4842 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:45:18] INFO >> epoch 031 | valid on 'valid' subset | loss 60.84 | nll_loss 5.829 | bleu 20.1078 | ppl 56.85 | wps 6620.7 | wpb 19314.6 | bsz 1850.5 | num_updates 26908 | best_bleu 20.1078 (progress_bar.py:269, print())
[2021-03-22 06:46:05] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 31 @ 26908 updates, score 20.10777107751764) (writing took 46.869526 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:46:20] INFO >> epoch 032: 92 / 868 loss=36.197, nll_loss=3.469, bleu=0, ppl=11.07, wps=2499.6, ups=3.74, wpb=667.6, bsz=64, num_updates=27000, lr=7.3e-05, gnorm=20.916, clip=0.4, train_wall=41, wall=4941 (progress_bar.py:260, log())
[2021-03-22 06:47:07] INFO >> epoch 032: 592 / 868 loss=35.41, nll_loss=3.399, bleu=0, ppl=10.55, wps=7133.9, ups=10.7, wpb=666.5, bsz=64, num_updates=27500, lr=7.3e-05, gnorm=20.987, clip=0.4, train_wall=44, wall=4988 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:47:35] INFO >> epoch 032 | loss 35.547 | nll_loss 3.412 | bleu 0 | ppl 10.64 | wps 3333.3 | ups 5 | wpb 666.6 | bsz 64 | num_updates 27776 | lr 7.3e-05 | gnorm 21.056 | clip 0.9 | train_wall 79 | wall 5016 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:48:06] INFO >> epoch 032 | valid on 'valid' subset | loss 60.814 | nll_loss 5.826 | bleu 20.3842 | ppl 56.75 | wps 7693.8 | wpb 19314.6 | bsz 1850.5 | num_updates 27776 | best_bleu 20.3842 (progress_bar.py:269, print())
[2021-03-22 06:48:45] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 32 @ 27776 updates, score 20.384204598838217) (writing took 38.398485 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:49:12] INFO >> epoch 033: 224 / 868 loss=35.625, nll_loss=3.405, bleu=0, ppl=10.59, wps=2682.4, ups=4.01, wpb=669.6, bsz=64, num_updates=28000, lr=7.2e-05, gnorm=21.299, clip=1.2, train_wall=46, wall=5112 (progress_bar.py:260, log())
[2021-03-22 06:49:56] INFO >> epoch 033: 724 / 868 loss=34.747, nll_loss=3.345, bleu=0, ppl=10.16, wps=7504.3, ups=11.29, wpb=664.4, bsz=64, num_updates=28500, lr=7.2e-05, gnorm=21.238, clip=1.2, train_wall=42, wall=5157 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:50:09] INFO >> epoch 033 | loss 34.923 | nll_loss 3.352 | bleu 0 | ppl 10.21 | wps 3744 | ups 5.62 | wpb 666.6 | bsz 64 | num_updates 28644 | lr 7.2e-05 | gnorm 21.276 | clip 0.7 | train_wall 74 | wall 5170 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:50:42] INFO >> epoch 033 | valid on 'valid' subset | loss 60.942 | nll_loss 5.839 | bleu 20.5088 | ppl 57.23 | wps 7395.5 | wpb 19314.6 | bsz 1850.5 | num_updates 28644 | best_bleu 20.5088 (progress_bar.py:269, print())
[2021-03-22 06:51:52] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 33 @ 28644 updates, score 20.508778557062563) (writing took 69.896409 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:52:30] INFO >> epoch 034: 356 / 868 loss=34.419, nll_loss=3.303, bleu=0, ppl=9.87, wps=2160.4, ups=3.24, wpb=666.8, bsz=64, num_updates=29000, lr=7.2e-05, gnorm=21.315, clip=0.8, train_wall=43, wall=5311 (progress_bar.py:260, log())
[2021-03-22 06:53:15] INFO >> epoch 034: 856 / 868 loss=34.419, nll_loss=3.304, bleu=0, ppl=9.88, wps=7351.9, ups=11.03, wpb=666.4, bsz=64, num_updates=29500, lr=7.2e-05, gnorm=21.545, clip=1, train_wall=43, wall=5356 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:53:17] INFO >> epoch 034 | loss 34.317 | nll_loss 3.294 | bleu 0 | ppl 9.81 | wps 3083.4 | ups 4.63 | wpb 666.6 | bsz 64 | num_updates 29512 | lr 7.2e-05 | gnorm 21.439 | clip 1 | train_wall 75 | wall 5358 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:53:56] INFO >> epoch 034 | valid on 'valid' subset | loss 60.784 | nll_loss 5.824 | bleu 20.5643 | ppl 56.64 | wps 5924.6 | wpb 19314.6 | bsz 1850.5 | num_updates 29512 | best_bleu 20.5643 (progress_bar.py:269, print())
[2021-03-22 06:55:01] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 34 @ 29512 updates, score 20.56425781321453) (writing took 64.563499 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:55:52] INFO >> epoch 035: 488 / 868 loss=33.627, nll_loss=3.227, bleu=0, ppl=9.36, wps=2129.7, ups=3.19, wpb=666.9, bsz=64, num_updates=30000, lr=7.1e-05, gnorm=21.493, clip=2, train_wall=44, wall=5513 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:56:27] INFO >> epoch 035 | loss 33.705 | nll_loss 3.235 | bleu 0 | ppl 9.42 | wps 3048.3 | ups 4.57 | wpb 666.6 | bsz 64 | num_updates 30380 | lr 7.1e-05 | gnorm 21.588 | clip 1.5 | train_wall 75 | wall 5548 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:57:06] INFO >> epoch 035 | valid on 'valid' subset | loss 60.697 | nll_loss 5.815 | bleu 20.7677 | ppl 56.31 | wps 5863 | wpb 19314.6 | bsz 1850.5 | num_updates 30380 | best_bleu 20.7677 (progress_bar.py:269, print())
[2021-03-22 06:58:12] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 35 @ 30380 updates, score 20.76769315507296) (writing took 65.798737 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 06:58:30] INFO >> epoch 036: 120 / 868 loss=33.592, nll_loss=3.226, bleu=0, ppl=9.36, wps=2103.4, ups=3.16, wpb=666.2, bsz=64, num_updates=30500, lr=7e-05, gnorm=21.645, clip=0.8, train_wall=44, wall=5671 (progress_bar.py:260, log())
[2021-03-22 06:59:18] INFO >> epoch 036: 620 / 868 loss=33.15, nll_loss=3.178, bleu=0, ppl=9.05, wps=7059.7, ups=10.58, wpb=667.3, bsz=64, num_updates=31000, lr=7e-05, gnorm=21.789, clip=1.4, train_wall=45, wall=5718 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 06:59:41] INFO >> epoch 036 | loss 33.114 | nll_loss 3.178 | bleu 0 | ppl 9.05 | wps 2986.8 | ups 4.48 | wpb 666.6 | bsz 64 | num_updates 31248 | lr 7e-05 | gnorm 21.745 | clip 1.3 | train_wall 78 | wall 5741 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:00:19] INFO >> epoch 036 | valid on 'valid' subset | loss 60.807 | nll_loss 5.826 | bleu 20.6589 | ppl 56.72 | wps 5996.4 | wpb 19314.6 | bsz 1850.5 | num_updates 31248 | best_bleu 20.7677 (progress_bar.py:269, print())
[2021-03-22 07:00:39] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 36 @ 31248 updates, score 20.658938517298658) (writing took 20.347787 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:01:09] INFO >> epoch 037: 252 / 868 loss=32.734, nll_loss=3.149, bleu=0, ppl=8.87, wps=2987.7, ups=4.49, wpb=665.4, bsz=64, num_updates=31500, lr=7e-05, gnorm=21.792, clip=1.6, train_wall=43, wall=5830 (progress_bar.py:260, log())
[2021-03-22 07:01:57] INFO >> epoch 037: 752 / 868 loss=32.61, nll_loss=3.127, bleu=0, ppl=8.73, wps=6889.1, ups=10.33, wpb=667.2, bsz=64, num_updates=32000, lr=7e-05, gnorm=21.947, clip=3, train_wall=46, wall=5878 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:02:09] INFO >> epoch 037 | loss 32.533 | nll_loss 3.123 | bleu 0 | ppl 8.71 | wps 3890 | ups 5.84 | wpb 666.6 | bsz 64 | num_updates 32116 | lr 7e-05 | gnorm 21.911 | clip 2.5 | train_wall 79 | wall 5890 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:02:49] INFO >> epoch 037 | valid on 'valid' subset | loss 60.788 | nll_loss 5.824 | bleu 20.8654 | ppl 56.65 | wps 5793.3 | wpb 19314.6 | bsz 1850.5 | num_updates 32116 | best_bleu 20.8654 (progress_bar.py:269, print())
[2021-03-22 07:04:15] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 37 @ 32116 updates, score 20.865442343421186) (writing took 86.393082 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:04:56] INFO >> epoch 038: 384 / 868 loss=32.183, nll_loss=3.085, bleu=0, ppl=8.49, wps=1874.2, ups=2.81, wpb=667.6, bsz=64, num_updates=32500, lr=6.9e-05, gnorm=22.073, clip=2.8, train_wall=43, wall=6056 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:05:39] INFO >> epoch 038 | loss 31.991 | nll_loss 3.071 | bleu 0 | ppl 8.4 | wps 2762.9 | ups 4.14 | wpb 666.6 | bsz 64 | num_updates 32984 | lr 6.9e-05 | gnorm 22.078 | clip 3.1 | train_wall 73 | wall 6100 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:06:17] INFO >> epoch 038 | valid on 'valid' subset | loss 60.79 | nll_loss 5.824 | bleu 20.891 | ppl 56.66 | wps 6087 | wpb 19314.6 | bsz 1850.5 | num_updates 32984 | best_bleu 20.891 (progress_bar.py:269, print())
[2021-03-22 07:06:56] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 38 @ 32984 updates, score 20.890997796375885) (writing took 38.678467 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:07:04] INFO >> epoch 039: 16 / 868 loss=32.003, nll_loss=3.072, bleu=0, ppl=8.41, wps=2598.3, ups=3.9, wpb=666.4, bsz=64, num_updates=33000, lr=6.8e-05, gnorm=22.106, clip=3, train_wall=43, wall=6185 (progress_bar.py:260, log())
[2021-03-22 07:07:49] INFO >> epoch 039: 516 / 868 loss=31.258, nll_loss=3.001, bleu=0, ppl=8, wps=7347.5, ups=11.02, wpb=666.6, bsz=64, num_updates=33500, lr=6.8e-05, gnorm=21.988, clip=2, train_wall=43, wall=6230 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:08:22] INFO >> epoch 039 | loss 31.449 | nll_loss 3.019 | bleu 0 | ppl 8.1 | wps 3550.7 | ups 5.33 | wpb 666.6 | bsz 64 | num_updates 33852 | lr 6.8e-05 | gnorm 22.109 | clip 2.6 | train_wall 75 | wall 6262 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:08:59] INFO >> epoch 039 | valid on 'valid' subset | loss 60.913 | nll_loss 5.836 | bleu 21.0189 | ppl 57.12 | wps 6187.7 | wpb 19314.6 | bsz 1850.5 | num_updates 33852 | best_bleu 21.0189 (progress_bar.py:269, print())
[2021-03-22 07:09:38] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 39 @ 33852 updates, score 21.018929570459772) (writing took 38.263663 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:09:57] INFO >> epoch 040: 148 / 868 loss=31.263, nll_loss=3.006, bleu=0, ppl=8.03, wps=2593, ups=3.9, wpb=665.4, bsz=64, num_updates=34000, lr=6.8e-05, gnorm=22.16, clip=3.2, train_wall=44, wall=6358 (progress_bar.py:260, log())
[2021-03-22 07:10:43] INFO >> epoch 040: 648 / 868 loss=30.99, nll_loss=2.969, bleu=0, ppl=7.83, wps=7355.9, ups=11.02, wpb=667.7, bsz=64, num_updates=34500, lr=6.8e-05, gnorm=22.224, clip=4.6, train_wall=43, wall=6404 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:11:05] INFO >> epoch 040 | loss 30.915 | nll_loss 2.967 | bleu 0 | ppl 7.82 | wps 3550.1 | ups 5.33 | wpb 666.6 | bsz 64 | num_updates 34720 | lr 6.8e-05 | gnorm 22.252 | clip 4.5 | train_wall 77 | wall 6425 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:11:42] INFO >> epoch 040 | valid on 'valid' subset | loss 60.904 | nll_loss 5.835 | bleu 21.1501 | ppl 57.09 | wps 6342.8 | wpb 19314.6 | bsz 1850.5 | num_updates 34720 | best_bleu 21.1501 (progress_bar.py:269, print())
[2021-03-22 07:12:31] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 40 @ 34720 updates, score 21.150106293159038) (writing took 49.848549 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:13:04] INFO >> epoch 041: 280 / 868 loss=30.558, nll_loss=2.938, bleu=0, ppl=7.67, wps=2355.6, ups=3.54, wpb=665.6, bsz=64, num_updates=35000, lr=6.7e-05, gnorm=22.384, clip=6, train_wall=46, wall=6545 (progress_bar.py:260, log())
[2021-03-22 07:13:50] INFO >> epoch 041: 780 / 868 loss=30.497, nll_loss=2.928, bleu=0, ppl=7.61, wps=7284.9, ups=10.93, wpb=666.2, bsz=64, num_updates=35500, lr=6.7e-05, gnorm=22.445, clip=5.2, train_wall=43, wall=6591 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:13:59] INFO >> epoch 041 | loss 30.413 | nll_loss 2.919 | bleu 0 | ppl 7.56 | wps 3327.7 | ups 4.99 | wpb 666.6 | bsz 64 | num_updates 35588 | lr 6.7e-05 | gnorm 22.389 | clip 5.4 | train_wall 77 | wall 6599 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:14:31] INFO >> epoch 041 | valid on 'valid' subset | loss 60.86 | nll_loss 5.831 | bleu 21.4015 | ppl 56.92 | wps 7565.1 | wpb 19314.6 | bsz 1850.5 | num_updates 35588 | best_bleu 21.4015 (progress_bar.py:269, print())
[2021-03-22 07:15:10] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 41 @ 35588 updates, score 21.401549193780603) (writing took 38.985450 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:15:55] INFO >> epoch 042: 412 / 868 loss=29.992, nll_loss=2.873, bleu=0, ppl=7.32, wps=2676.7, ups=4.01, wpb=668.2, bsz=64, num_updates=36000, lr=6.6e-05, gnorm=22.383, clip=4.4, train_wall=45, wall=6715 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:16:39] INFO >> epoch 042 | loss 29.965 | nll_loss 2.876 | bleu 0 | ppl 7.34 | wps 3602.1 | ups 5.4 | wpb 666.6 | bsz 64 | num_updates 36456 | lr 6.6e-05 | gnorm 22.469 | clip 5.4 | train_wall 79 | wall 6760 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:17:11] INFO >> epoch 042 | valid on 'valid' subset | loss 60.934 | nll_loss 5.838 | bleu 21.2115 | ppl 57.2 | wps 7643.1 | wpb 19314.6 | bsz 1850.5 | num_updates 36456 | best_bleu 21.4015 (progress_bar.py:269, print())
[2021-03-22 07:17:31] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 42 @ 36456 updates, score 21.2114862920614) (writing took 19.944283 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:17:42] INFO >> epoch 043: 44 / 868 loss=29.999, nll_loss=2.884, bleu=0, ppl=7.38, wps=3114, ups=4.68, wpb=665.4, bsz=64, num_updates=36500, lr=6.6e-05, gnorm=22.522, clip=6.2, train_wall=46, wall=6822 (progress_bar.py:260, log())
[2021-03-22 07:18:27] INFO >> epoch 043: 544 / 868 loss=29.331, nll_loss=2.815, bleu=0, ppl=7.04, wps=7331, ups=11, wpb=666.5, bsz=64, num_updates=37000, lr=6.6e-05, gnorm=22.503, clip=4.8, train_wall=43, wall=6868 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:18:58] INFO >> epoch 043 | loss 29.469 | nll_loss 2.829 | bleu 0 | ppl 7.1 | wps 4175 | ups 6.26 | wpb 666.6 | bsz 64 | num_updates 37324 | lr 6.6e-05 | gnorm 22.613 | clip 6.5 | train_wall 76 | wall 6899 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:19:30] INFO >> epoch 043 | valid on 'valid' subset | loss 60.98 | nll_loss 5.842 | bleu 21.3417 | ppl 57.38 | wps 7510.1 | wpb 19314.6 | bsz 1850.5 | num_updates 37324 | best_bleu 21.4015 (progress_bar.py:269, print())
[2021-03-22 07:19:50] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 43 @ 37324 updates, score 21.341728480763827) (writing took 20.504245 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:20:14] INFO >> epoch 044: 176 / 868 loss=29.334, nll_loss=2.816, bleu=0, ppl=7.04, wps=3127.3, ups=4.69, wpb=666.6, bsz=64, num_updates=37500, lr=6.5e-05, gnorm=22.741, clip=8, train_wall=45, wall=6974 (progress_bar.py:260, log())
[2021-03-22 07:20:59] INFO >> epoch 044: 676 / 868 loss=29.023, nll_loss=2.784, bleu=0, ppl=6.89, wps=7312.5, ups=10.96, wpb=667.2, bsz=64, num_updates=38000, lr=6.5e-05, gnorm=22.694, clip=7.4, train_wall=43, wall=7020 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:21:17] INFO >> epoch 044 | loss 28.996 | nll_loss 2.783 | bleu 0 | ppl 6.88 | wps 4155.2 | ups 6.23 | wpb 666.6 | bsz 64 | num_updates 38192 | lr 6.5e-05 | gnorm 22.692 | clip 7.1 | train_wall 76 | wall 7038 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:21:50] INFO >> epoch 044 | valid on 'valid' subset | loss 60.956 | nll_loss 5.84 | bleu 21.5808 | ppl 57.29 | wps 7230.5 | wpb 19314.6 | bsz 1850.5 | num_updates 38192 | best_bleu 21.5808 (progress_bar.py:269, print())
[2021-03-22 07:23:06] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 44 @ 38192 updates, score 21.58076517078775) (writing took 75.921958 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:23:45] INFO >> epoch 045: 308 / 868 loss=28.644, nll_loss=2.753, bleu=0, ppl=6.74, wps=2010.2, ups=3.02, wpb=665.6, bsz=64, num_updates=38500, lr=6.4e-05, gnorm=22.664, clip=6.6, train_wall=47, wall=7186 (progress_bar.py:260, log())
[2021-03-22 07:24:35] INFO >> epoch 045: 808 / 868 loss=28.73, nll_loss=2.75, bleu=0, ppl=6.73, wps=6665.1, ups=9.97, wpb=668.3, bsz=64, num_updates=39000, lr=6.4e-05, gnorm=22.955, clip=10.4, train_wall=48, wall=7236 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:24:41] INFO >> epoch 045 | loss 28.523 | nll_loss 2.738 | bleu 0 | ppl 6.67 | wps 2839.7 | ups 4.26 | wpb 666.6 | bsz 64 | num_updates 39060 | lr 6.4e-05 | gnorm 22.808 | clip 8.9 | train_wall 84 | wall 7242 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:25:12] INFO >> epoch 045 | valid on 'valid' subset | loss 60.951 | nll_loss 5.84 | bleu 21.7282 | ppl 57.27 | wps 7579.2 | wpb 19314.6 | bsz 1850.5 | num_updates 39060 | best_bleu 21.7282 (progress_bar.py:269, print())
[2021-03-22 07:26:16] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 45 @ 39060 updates, score 21.7281612997329) (writing took 63.105360 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:27:04] INFO >> epoch 046: 440 / 868 loss=27.832, nll_loss=2.675, bleu=0, ppl=6.38, wps=2229.5, ups=3.35, wpb=666, bsz=64, num_updates=39500, lr=6.4e-05, gnorm=22.715, clip=7.6, train_wall=46, wall=7385 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:27:46] INFO >> epoch 046 | loss 28.08 | nll_loss 2.695 | bleu 0 | ppl 6.48 | wps 3129.5 | ups 4.69 | wpb 666.6 | bsz 64 | num_updates 39928 | lr 6.4e-05 | gnorm 22.869 | clip 9.1 | train_wall 79 | wall 7426 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:28:19] INFO >> epoch 046 | valid on 'valid' subset | loss 61.129 | nll_loss 5.857 | bleu 21.6369 | ppl 57.95 | wps 7206.3 | wpb 19314.6 | bsz 1850.5 | num_updates 39928 | best_bleu 21.7282 (progress_bar.py:269, print())
[2021-03-22 07:28:42] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 46 @ 39928 updates, score 21.636881373056237) (writing took 23.458981 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:28:56] INFO >> epoch 047: 72 / 868 loss=28.353, nll_loss=2.717, bleu=0, ppl=6.58, wps=2995.3, ups=4.49, wpb=667.6, bsz=64, num_updates=40000, lr=6.3e-05, gnorm=23.048, clip=11.4, train_wall=46, wall=7496 (progress_bar.py:260, log())
[2021-03-22 07:29:43] INFO >> epoch 047: 572 / 868 loss=27.413, nll_loss=2.641, bleu=0, ppl=6.24, wps=7011.8, ups=10.56, wpb=664.3, bsz=64, num_updates=40500, lr=6.3e-05, gnorm=22.877, clip=8.8, train_wall=45, wall=7544 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:30:11] INFO >> epoch 047 | loss 27.647 | nll_loss 2.654 | bleu 0 | ppl 6.29 | wps 3969 | ups 5.95 | wpb 666.6 | bsz 64 | num_updates 40796 | lr 6.3e-05 | gnorm 22.99 | clip 9.9 | train_wall 78 | wall 7572 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:30:43] INFO >> epoch 047 | valid on 'valid' subset | loss 61.17 | nll_loss 5.861 | bleu 21.6033 | ppl 58.11 | wps 7546.5 | wpb 19314.6 | bsz 1850.5 | num_updates 40796 | best_bleu 21.7282 (progress_bar.py:269, print())
[2021-03-22 07:31:04] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 47 @ 40796 updates, score 21.603314548263686) (writing took 20.230919 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:31:28] INFO >> epoch 048: 204 / 868 loss=27.41, nll_loss=2.633, bleu=0, ppl=6.2, wps=3164.3, ups=4.75, wpb=666, bsz=64, num_updates=41000, lr=6.2e-05, gnorm=22.965, clip=10.2, train_wall=44, wall=7649 (progress_bar.py:260, log())
[2021-03-22 07:32:14] INFO >> epoch 048: 704 / 868 loss=27.384, nll_loss=2.622, bleu=0, ppl=6.15, wps=7307.2, ups=10.94, wpb=668.2, bsz=64, num_updates=41500, lr=6.2e-05, gnorm=23.164, clip=13.4, train_wall=43, wall=7695 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:32:31] INFO >> epoch 048 | loss 27.26 | nll_loss 2.617 | bleu 0 | ppl 6.13 | wps 4150.3 | ups 6.23 | wpb 666.6 | bsz 64 | num_updates 41664 | lr 6.2e-05 | gnorm 23.081 | clip 12.7 | train_wall 76 | wall 7712 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:33:03] INFO >> epoch 048 | valid on 'valid' subset | loss 61.214 | nll_loss 5.865 | bleu 21.7019 | ppl 58.28 | wps 7524.6 | wpb 19314.6 | bsz 1850.5 | num_updates 41664 | best_bleu 21.7282 (progress_bar.py:269, print())
[2021-03-22 07:33:23] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 48 @ 41664 updates, score 21.701922264890356) (writing took 20.454291 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:34:00] INFO >> epoch 049: 336 / 868 loss=26.942, nll_loss=2.584, bleu=0, ppl=5.99, wps=3133.8, ups=4.7, wpb=667.4, bsz=64, num_updates=42000, lr=6.2e-05, gnorm=23.045, clip=12.2, train_wall=45, wall=7801 (progress_bar.py:260, log())
[2021-03-22 07:34:46] INFO >> epoch 049: 836 / 868 loss=26.975, nll_loss=2.591, bleu=0, ppl=6.02, wps=7303.8, ups=10.96, wpb=666.1, bsz=64, num_updates=42500, lr=6.2e-05, gnorm=23.254, clip=14.4, train_wall=43, wall=7847 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:34:49] INFO >> epoch 049 | loss 26.836 | nll_loss 2.576 | bleu 0 | ppl 5.96 | wps 4177.2 | ups 6.27 | wpb 666.6 | bsz 64 | num_updates 42532 | lr 6.2e-05 | gnorm 23.135 | clip 13 | train_wall 75 | wall 7850 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:35:21] INFO >> epoch 049 | valid on 'valid' subset | loss 61.301 | nll_loss 5.873 | bleu 21.8486 | ppl 58.61 | wps 7662.9 | wpb 19314.6 | bsz 1850.5 | num_updates 42532 | best_bleu 21.8486 (progress_bar.py:269, print())
[2021-03-22 07:35:59] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 49 @ 42532 updates, score 21.848551133208904) (writing took 38.288355 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:36:48] INFO >> epoch 050: 468 / 868 loss=26.308, nll_loss=2.526, bleu=0, ppl=5.76, wps=2733.3, ups=4.1, wpb=666.3, bsz=64, num_updates=43000, lr=6.1e-05, gnorm=23.114, clip=13, train_wall=43, wall=7969 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:37:25] INFO >> epoch 050 | loss 26.428 | nll_loss 2.537 | bleu 0 | ppl 5.8 | wps 3723.4 | ups 5.59 | wpb 666.6 | bsz 64 | num_updates 43400 | lr 6.1e-05 | gnorm 23.233 | clip 15 | train_wall 75 | wall 8006 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:37:57] INFO >> epoch 050 | valid on 'valid' subset | loss 61.379 | nll_loss 5.881 | bleu 21.8068 | ppl 58.92 | wps 7461.5 | wpb 19314.6 | bsz 1850.5 | num_updates 43400 | best_bleu 21.8486 (progress_bar.py:269, print())
[2021-03-22 07:38:17] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 50 @ 43400 updates, score 21.806847545652033) (writing took 20.480468 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:38:33] INFO >> epoch 051: 100 / 868 loss=26.574, nll_loss=2.545, bleu=0, ppl=5.84, wps=3183.5, ups=4.76, wpb=668.3, bsz=64, num_updates=43500, lr=6.1e-05, gnorm=23.326, clip=16, train_wall=44, wall=8074 (progress_bar.py:260, log())
[2021-03-22 07:39:18] INFO >> epoch 051: 600 / 868 loss=25.909, nll_loss=2.495, bleu=0, ppl=5.64, wps=7311, ups=11.01, wpb=664.3, bsz=64, num_updates=44000, lr=6.1e-05, gnorm=23.25, clip=15, train_wall=43, wall=8119 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:39:44] INFO >> epoch 051 | loss 26.05 | nll_loss 2.5 | bleu 0 | ppl 5.66 | wps 4152.6 | ups 6.23 | wpb 666.6 | bsz 64 | num_updates 44268 | lr 6.1e-05 | gnorm 23.294 | clip 14.6 | train_wall 76 | wall 8145 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:40:16] INFO >> epoch 051 | valid on 'valid' subset | loss 61.546 | nll_loss 5.897 | bleu 21.8918 | ppl 59.57 | wps 7660.4 | wpb 19314.6 | bsz 1850.5 | num_updates 44268 | best_bleu 21.8918 (progress_bar.py:269, print())
[2021-03-22 07:40:54] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 51 @ 44268 updates, score 21.891834734179135) (writing took 38.055529 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:41:19] INFO >> epoch 052: 232 / 868 loss=25.867, nll_loss=2.478, bleu=0, ppl=5.57, wps=2757.5, ups=4.13, wpb=667.7, bsz=64, num_updates=44500, lr=6e-05, gnorm=23.317, clip=13.6, train_wall=43, wall=8240 (progress_bar.py:260, log())
[2021-03-22 07:42:04] INFO >> epoch 052: 732 / 868 loss=25.74, nll_loss=2.471, bleu=0, ppl=5.54, wps=7535.4, ups=11.3, wpb=666.6, bsz=64, num_updates=45000, lr=6e-05, gnorm=23.494, clip=17.2, train_wall=42, wall=8284 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:42:17] INFO >> epoch 052 | loss 25.66 | nll_loss 2.463 | bleu 0 | ppl 5.51 | wps 3796.6 | ups 5.7 | wpb 666.6 | bsz 64 | num_updates 45136 | lr 6e-05 | gnorm 23.386 | clip 15.3 | train_wall 73 | wall 8297 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:42:48] INFO >> epoch 052 | valid on 'valid' subset | loss 61.529 | nll_loss 5.895 | bleu 21.9772 | ppl 59.51 | wps 7599.3 | wpb 19314.6 | bsz 1850.5 | num_updates 45136 | best_bleu 21.9772 (progress_bar.py:269, print())
[2021-03-22 07:43:46] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 52 @ 45136 updates, score 21.977216804359653) (writing took 57.177894 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:44:25] INFO >> epoch 053: 364 / 868 loss=25.431, nll_loss=2.432, bleu=0, ppl=5.4, wps=2362.8, ups=3.53, wpb=668.8, bsz=64, num_updates=45500, lr=5.9e-05, gnorm=23.482, clip=16, train_wall=44, wall=8426 (progress_bar.py:260, log())
[2021-03-22 07:45:10] INFO >> epoch 053: 864 / 868 loss=25.273, nll_loss=2.435, bleu=0, ppl=5.41, wps=7376.4, ups=11.1, wpb=664.3, bsz=64, num_updates=46000, lr=5.9e-05, gnorm=23.499, clip=17, train_wall=43, wall=8471 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:45:11] INFO >> epoch 053 | loss 25.288 | nll_loss 2.427 | bleu 0 | ppl 5.38 | wps 3316 | ups 4.97 | wpb 666.6 | bsz 64 | num_updates 46004 | lr 5.9e-05 | gnorm 23.489 | clip 16.7 | train_wall 75 | wall 8472 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:45:44] INFO >> epoch 053 | valid on 'valid' subset | loss 61.601 | nll_loss 5.902 | bleu 21.9603 | ppl 59.79 | wps 7149 | wpb 19314.6 | bsz 1850.5 | num_updates 46004 | best_bleu 21.9772 (progress_bar.py:269, print())
[2021-03-22 07:46:04] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 53 @ 46004 updates, score 21.960306242079863) (writing took 19.993129 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:46:59] INFO >> epoch 054: 496 / 868 loss=24.755, nll_loss=2.379, bleu=0, ppl=5.2, wps=3075.1, ups=4.62, wpb=666, bsz=64, num_updates=46500, lr=5.9e-05, gnorm=23.404, clip=16.2, train_wall=46, wall=8579 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:47:35] INFO >> epoch 054 | loss 24.919 | nll_loss 2.392 | bleu 0 | ppl 5.25 | wps 4011.9 | ups 6.02 | wpb 666.6 | bsz 64 | num_updates 46872 | lr 5.9e-05 | gnorm 23.506 | clip 18.1 | train_wall 80 | wall 8616 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:48:07] INFO >> epoch 054 | valid on 'valid' subset | loss 61.56 | nll_loss 5.898 | bleu 22.2073 | ppl 59.63 | wps 7505.9 | wpb 19314.6 | bsz 1850.5 | num_updates 46872 | best_bleu 22.2073 (progress_bar.py:269, print())
[2021-03-22 07:48:46] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 54 @ 46872 updates, score 22.207288510752782) (writing took 38.534736 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:49:04] INFO >> epoch 055: 128 / 868 loss=24.949, nll_loss=2.393, bleu=0, ppl=5.25, wps=2652.7, ups=3.98, wpb=667, bsz=64, num_updates=47000, lr=5.8e-05, gnorm=23.52, clip=18.4, train_wall=46, wall=8705 (progress_bar.py:260, log())
[2021-03-22 07:49:49] INFO >> epoch 055: 628 / 868 loss=24.601, nll_loss=2.358, bleu=0, ppl=5.13, wps=7375.1, ups=11.05, wpb=667.5, bsz=64, num_updates=47500, lr=5.8e-05, gnorm=23.602, clip=19.8, train_wall=43, wall=8750 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:50:12] INFO >> epoch 055 | loss 24.581 | nll_loss 2.359 | bleu 0 | ppl 5.13 | wps 3700.9 | ups 5.55 | wpb 666.6 | bsz 64 | num_updates 47740 | lr 5.8e-05 | gnorm 23.579 | clip 18.5 | train_wall 75 | wall 8772 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:50:44] INFO >> epoch 055 | valid on 'valid' subset | loss 61.725 | nll_loss 5.914 | bleu 22.2335 | ppl 60.29 | wps 7494.2 | wpb 19314.6 | bsz 1850.5 | num_updates 47740 | best_bleu 22.2335 (progress_bar.py:269, print())
[2021-03-22 07:52:13] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 55 @ 47740 updates, score 22.233478244593986) (writing took 89.745572 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:52:44] INFO >> epoch 056: 260 / 868 loss=24.361, nll_loss=2.341, bleu=0, ppl=5.07, wps=1913.7, ups=2.87, wpb=666.1, bsz=64, num_updates=48000, lr=5.8e-05, gnorm=23.594, clip=19.2, train_wall=43, wall=8924 (progress_bar.py:260, log())
[2021-03-22 07:53:32] INFO >> epoch 056: 760 / 868 loss=24.384, nll_loss=2.338, bleu=0, ppl=5.06, wps=6859.8, ups=10.28, wpb=667.5, bsz=64, num_updates=48500, lr=5.8e-05, gnorm=23.695, clip=21.6, train_wall=46, wall=8973 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:53:43] INFO >> epoch 056 | loss 24.276 | nll_loss 2.33 | bleu 0 | ppl 5.03 | wps 2732.8 | ups 4.1 | wpb 666.6 | bsz 64 | num_updates 48608 | lr 5.8e-05 | gnorm 23.633 | clip 20.5 | train_wall 79 | wall 8984 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:54:15] INFO >> epoch 056 | valid on 'valid' subset | loss 61.768 | nll_loss 5.918 | bleu 22.183 | ppl 60.46 | wps 7556.4 | wpb 19314.6 | bsz 1850.5 | num_updates 48608 | best_bleu 22.2335 (progress_bar.py:269, print())
[2021-03-22 07:54:36] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 56 @ 48608 updates, score 22.183035934199438) (writing took 20.347293 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:55:18] INFO >> epoch 057: 392 / 868 loss=23.851, nll_loss=2.288, bleu=0, ppl=4.88, wps=3138.4, ups=4.71, wpb=667, bsz=64, num_updates=49000, lr=5.7e-05, gnorm=23.584, clip=19.2, train_wall=45, wall=9079 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:56:03] INFO >> epoch 057 | loss 23.929 | nll_loss 2.297 | bleu 0 | ppl 4.91 | wps 4144.2 | ups 6.22 | wpb 666.6 | bsz 64 | num_updates 49476 | lr 5.7e-05 | gnorm 23.631 | clip 18.9 | train_wall 77 | wall 9124 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 07:56:41] INFO >> epoch 057 | valid on 'valid' subset | loss 61.819 | nll_loss 5.923 | bleu 22.3379 | ppl 60.66 | wps 5884.5 | wpb 19314.6 | bsz 1850.5 | num_updates 49476 | best_bleu 22.3379 (progress_bar.py:269, print())
[2021-03-22 07:58:33] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 57 @ 49476 updates, score 22.337927896085155) (writing took 111.747860 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 07:58:42] INFO >> epoch 058: 24 / 868 loss=24.005, nll_loss=2.311, bleu=0, ppl=4.96, wps=1634.6, ups=2.46, wpb=664.4, bsz=64, num_updates=49500, lr=5.6e-05, gnorm=23.68, clip=18.4, train_wall=44, wall=9283 (progress_bar.py:260, log())
[2021-03-22 07:59:28] INFO >> epoch 058: 524 / 868 loss=23.536, nll_loss=2.259, bleu=0, ppl=4.79, wps=7146, ups=10.71, wpb=666.9, bsz=64, num_updates=50000, lr=5.6e-05, gnorm=23.602, clip=17, train_wall=44, wall=9329 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:00:00] INFO >> epoch 058 | loss 23.604 | nll_loss 2.266 | bleu 0 | ppl 4.81 | wps 2443.6 | ups 3.67 | wpb 666.6 | bsz 64 | num_updates 50344 | lr 5.6e-05 | gnorm 23.708 | clip 20.7 | train_wall 76 | wall 9361 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:00:33] INFO >> epoch 058 | valid on 'valid' subset | loss 61.892 | nll_loss 5.93 | bleu 22.3764 | ppl 60.96 | wps 7198.6 | wpb 19314.6 | bsz 1850.5 | num_updates 50344 | best_bleu 22.3764 (progress_bar.py:269, print())
[2021-03-22 08:01:12] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 58 @ 50344 updates, score 22.376352630334885) (writing took 38.436418 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:01:33] INFO >> epoch 059: 156 / 868 loss=23.595, nll_loss=2.26, bleu=0, ppl=4.79, wps=2670.9, ups=4, wpb=667.6, bsz=63.9, num_updates=50500, lr=5.6e-05, gnorm=23.805, clip=24, train_wall=44, wall=9454 (progress_bar.py:260, log())
[2021-03-22 08:02:20] INFO >> epoch 059: 656 / 868 loss=23.26, nll_loss=2.235, bleu=0, ppl=4.71, wps=7119.6, ups=10.69, wpb=666.1, bsz=64, num_updates=51000, lr=5.6e-05, gnorm=23.825, clip=23.2, train_wall=44, wall=9501 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:02:40] INFO >> epoch 059 | loss 23.298 | nll_loss 2.236 | bleu 0 | ppl 4.71 | wps 3616.2 | ups 5.42 | wpb 666.6 | bsz 64 | num_updates 51212 | lr 5.6e-05 | gnorm 23.829 | clip 22.6 | train_wall 77 | wall 9521 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:03:13] INFO >> epoch 059 | valid on 'valid' subset | loss 61.992 | nll_loss 5.939 | bleu 22.6648 | ppl 61.37 | wps 7427.7 | wpb 19314.6 | bsz 1850.5 | num_updates 51212 | best_bleu 22.6648 (progress_bar.py:269, print())
[2021-03-22 08:03:51] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 59 @ 51212 updates, score 22.66482128192663) (writing took 38.837033 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:04:24] INFO >> epoch 060: 288 / 868 loss=23.035, nll_loss=2.212, bleu=0, ppl=4.63, wps=2682.4, ups=4.03, wpb=666.4, bsz=64, num_updates=51500, lr=5.5e-05, gnorm=23.763, clip=21, train_wall=44, wall=9625 (progress_bar.py:260, log())
[2021-03-22 08:05:11] INFO >> epoch 060: 788 / 868 loss=23.106, nll_loss=2.219, bleu=0, ppl=4.65, wps=7072.2, ups=10.62, wpb=666.2, bsz=64, num_updates=52000, lr=5.5e-05, gnorm=23.849, clip=23.4, train_wall=45, wall=9672 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:05:20] INFO >> epoch 060 | loss 23.007 | nll_loss 2.208 | bleu 0 | ppl 4.62 | wps 3620.7 | ups 5.43 | wpb 666.6 | bsz 64 | num_updates 52080 | lr 5.5e-05 | gnorm 23.787 | clip 22.4 | train_wall 77 | wall 9680 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:06:05] INFO >> epoch 060 | valid on 'valid' subset | loss 62.112 | nll_loss 5.951 | bleu 22.499 | ppl 61.86 | wps 4833.9 | wpb 19314.6 | bsz 1850.5 | num_updates 52080 | best_bleu 22.6648 (progress_bar.py:269, print())
[2021-03-22 08:07:06] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 60 @ 52080 updates, score 22.49903025866992) (writing took 61.213071 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:07:51] INFO >> epoch 061: 420 / 868 loss=22.53, nll_loss=2.167, bleu=0, ppl=4.49, wps=2085.1, ups=3.13, wpb=665.5, bsz=64, num_updates=52500, lr=5.5e-05, gnorm=23.702, clip=20.6, train_wall=44, wall=9832 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:08:46] INFO >> epoch 061 | loss 22.683 | nll_loss 2.177 | bleu 0 | ppl 4.52 | wps 2803.6 | ups 4.21 | wpb 666.6 | bsz 64 | num_updates 52948 | lr 5.5e-05 | gnorm 23.804 | clip 23.8 | train_wall 87 | wall 9887 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:09:28] INFO >> epoch 061 | valid on 'valid' subset | loss 62.129 | nll_loss 5.952 | bleu 22.58 | ppl 61.92 | wps 5171.9 | wpb 19314.6 | bsz 1850.5 | num_updates 52948 | best_bleu 22.6648 (progress_bar.py:269, print())
[2021-03-22 08:09:49] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 61 @ 52948 updates, score 22.57998741755053) (writing took 20.366536 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:10:00] INFO >> epoch 062: 52 / 868 loss=22.924, nll_loss=2.193, bleu=0, ppl=4.57, wps=2592.3, ups=3.88, wpb=668.5, bsz=63.9, num_updates=53000, lr=5.4e-05, gnorm=23.995, clip=28, train_wall=55, wall=9961 (progress_bar.py:260, log())
[2021-03-22 08:10:46] INFO >> epoch 062: 552 / 868 loss=22.282, nll_loss=2.139, bleu=0, ppl=4.41, wps=7269.8, ups=10.91, wpb=666.6, bsz=64, num_updates=53500, lr=5.4e-05, gnorm=23.737, clip=23, train_wall=43, wall=10007 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:11:16] INFO >> epoch 062 | loss 22.37 | nll_loss 2.147 | bleu 0 | ppl 4.43 | wps 3858.4 | ups 5.79 | wpb 666.6 | bsz 64 | num_updates 53816 | lr 5.4e-05 | gnorm 23.903 | clip 25 | train_wall 76 | wall 10037 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:11:51] INFO >> epoch 062 | valid on 'valid' subset | loss 62.329 | nll_loss 5.972 | bleu 22.8391 | ppl 62.75 | wps 6579.3 | wpb 19314.6 | bsz 1850.5 | num_updates 53816 | best_bleu 22.8391 (progress_bar.py:269, print())
[2021-03-22 08:13:01] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 62 @ 53816 updates, score 22.83905576873621) (writing took 69.536200 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:13:24] INFO >> epoch 063: 184 / 868 loss=22.27, nll_loss=2.139, bleu=0, ppl=4.4, wps=2102.9, ups=3.16, wpb=666.4, bsz=64, num_updates=54000, lr=5.4e-05, gnorm=23.971, clip=26.2, train_wall=45, wall=10165 (progress_bar.py:260, log())
[2021-03-22 08:14:12] INFO >> epoch 063: 684 / 868 loss=21.991, nll_loss=2.12, bleu=0, ppl=4.35, wps=6939.8, ups=10.46, wpb=663.6, bsz=64, num_updates=54500, lr=5.4e-05, gnorm=23.917, clip=25, train_wall=45, wall=10213 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:14:30] INFO >> epoch 063 | loss 22.119 | nll_loss 2.123 | bleu 0 | ppl 4.36 | wps 2985.9 | ups 4.48 | wpb 666.6 | bsz 64 | num_updates 54684 | lr 5.4e-05 | gnorm 23.968 | clip 26.3 | train_wall 78 | wall 10231 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:15:02] INFO >> epoch 063 | valid on 'valid' subset | loss 62.342 | nll_loss 5.973 | bleu 22.8277 | ppl 62.81 | wps 7533.5 | wpb 19314.6 | bsz 1850.5 | num_updates 54684 | best_bleu 22.8391 (progress_bar.py:269, print())
[2021-03-22 08:15:21] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 63 @ 54684 updates, score 22.82766640203839) (writing took 19.745046 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:15:57] INFO >> epoch 064: 316 / 868 loss=22.02, nll_loss=2.102, bleu=0, ppl=4.29, wps=3184.3, ups=4.75, wpb=670.3, bsz=64, num_updates=55000, lr=5.3e-05, gnorm=23.975, clip=24.8, train_wall=45, wall=10318 (progress_bar.py:260, log())
[2021-03-22 08:16:43] INFO >> epoch 064: 816 / 868 loss=21.86, nll_loss=2.105, bleu=0, ppl=4.3, wps=7285.4, ups=10.97, wpb=664.3, bsz=64, num_updates=55500, lr=5.3e-05, gnorm=24.001, clip=27.4, train_wall=43, wall=10364 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:16:48] INFO >> epoch 064 | loss 21.823 | nll_loss 2.095 | bleu 0 | ppl 4.27 | wps 4181 | ups 6.27 | wpb 666.6 | bsz 64 | num_updates 55552 | lr 5.3e-05 | gnorm 23.929 | clip 25 | train_wall 76 | wall 10369 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:17:20] INFO >> epoch 064 | valid on 'valid' subset | loss 62.406 | nll_loss 5.979 | bleu 23.02 | ppl 63.07 | wps 7539.6 | wpb 19314.6 | bsz 1850.5 | num_updates 55552 | best_bleu 23.02 (progress_bar.py:269, print())
[2021-03-22 08:18:32] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 64 @ 55552 updates, score 23.019996150614137) (writing took 72.197899 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:19:20] INFO >> epoch 065: 448 / 868 loss=21.562, nll_loss=2.066, bleu=0, ppl=4.19, wps=2126.7, ups=3.18, wpb=668, bsz=64, num_updates=56000, lr=5.3e-05, gnorm=23.968, clip=26.6, train_wall=44, wall=10521 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:19:57] INFO >> epoch 065 | loss 21.544 | nll_loss 2.068 | bleu 0 | ppl 4.19 | wps 3063.2 | ups 4.6 | wpb 666.6 | bsz 64 | num_updates 56420 | lr 5.3e-05 | gnorm 24.029 | clip 28.5 | train_wall 74 | wall 10558 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:20:29] INFO >> epoch 065 | valid on 'valid' subset | loss 62.451 | nll_loss 5.983 | bleu 22.9445 | ppl 63.27 | wps 7497.3 | wpb 19314.6 | bsz 1850.5 | num_updates 56420 | best_bleu 23.02 (progress_bar.py:269, print())
[2021-03-22 08:21:10] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 65 @ 56420 updates, score 22.944463277305964) (writing took 40.840467 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:21:23] INFO >> epoch 066: 80 / 868 loss=21.505, nll_loss=2.067, bleu=0, ppl=4.19, wps=2696.4, ups=4.05, wpb=665.7, bsz=64, num_updates=56500, lr=5.2e-05, gnorm=24.069, clip=29.4, train_wall=42, wall=10644 (progress_bar.py:260, log())
[2021-03-22 08:22:09] INFO >> epoch 066: 580 / 868 loss=21.228, nll_loss=2.034, bleu=0, ppl=4.09, wps=7331.5, ups=10.98, wpb=667.8, bsz=64, num_updates=57000, lr=5.2e-05, gnorm=24.003, clip=25.8, train_wall=43, wall=10690 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:22:36] INFO >> epoch 066 | loss 21.249 | nll_loss 2.04 | bleu 0 | ppl 4.11 | wps 3643.5 | ups 5.47 | wpb 666.6 | bsz 64 | num_updates 57288 | lr 5.2e-05 | gnorm 23.998 | clip 27.2 | train_wall 75 | wall 10717 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:23:08] INFO >> epoch 066 | valid on 'valid' subset | loss 62.621 | nll_loss 6 | bleu 22.9219 | ppl 63.98 | wps 7481.6 | wpb 19314.6 | bsz 1850.5 | num_updates 57288 | best_bleu 23.02 (progress_bar.py:269, print())
[2021-03-22 08:26:26] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 66 @ 57288 updates, score 22.921940220933525) (writing took 198.081196 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:26:51] INFO >> epoch 067: 212 / 868 loss=21.065, nll_loss=2.029, bleu=0, ppl=4.08, wps=1176.3, ups=1.77, wpb=664.6, bsz=64, num_updates=57500, lr=5.2e-05, gnorm=23.93, clip=27.2, train_wall=43, wall=10972 (progress_bar.py:260, log())
[2021-03-22 08:27:38] INFO >> epoch 067: 712 / 868 loss=21.205, nll_loss=2.028, bleu=0, ppl=4.08, wps=7223.9, ups=10.8, wpb=668.9, bsz=64, num_updates=58000, lr=5.2e-05, gnorm=24.201, clip=31.2, train_wall=44, wall=11019 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:27:52] INFO >> epoch 067 | loss 21.026 | nll_loss 2.018 | bleu 0 | ppl 4.05 | wps 1827.3 | ups 2.74 | wpb 666.6 | bsz 64 | num_updates 58156 | lr 5.2e-05 | gnorm 24.071 | clip 28.8 | train_wall 76 | wall 11033 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:28:24] INFO >> epoch 067 | valid on 'valid' subset | loss 62.757 | nll_loss 6.013 | bleu 23.0472 | ppl 64.56 | wps 7515.9 | wpb 19314.6 | bsz 1850.5 | num_updates 58156 | best_bleu 23.0472 (progress_bar.py:269, print())
[2021-03-22 08:29:03] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 67 @ 58156 updates, score 23.047216333729054) (writing took 38.483297 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:29:41] INFO >> epoch 068: 344 / 868 loss=20.799, nll_loss=1.995, bleu=0, ppl=3.99, wps=2714.8, ups=4.07, wpb=666.8, bsz=64, num_updates=58500, lr=5.1e-05, gnorm=24.042, clip=29, train_wall=43, wall=11141 (progress_bar.py:260, log())
[2021-03-22 08:30:26] INFO >> epoch 068: 844 / 868 loss=20.851, nll_loss=2.003, bleu=0, ppl=4.01, wps=7314.9, ups=10.98, wpb=666.2, bsz=64, num_updates=59000, lr=5.1e-05, gnorm=24.245, clip=28.8, train_wall=43, wall=11187 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:30:29] INFO >> epoch 068 | loss 20.768 | nll_loss 1.993 | bleu 0 | ppl 3.98 | wps 3702.3 | ups 5.55 | wpb 666.6 | bsz 64 | num_updates 59024 | lr 5.1e-05 | gnorm 24.147 | clip 28.7 | train_wall 75 | wall 11190 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:31:01] INFO >> epoch 068 | valid on 'valid' subset | loss 62.705 | nll_loss 6.008 | bleu 23.194 | ppl 64.34 | wps 7486.8 | wpb 19314.6 | bsz 1850.5 | num_updates 59024 | best_bleu 23.194 (progress_bar.py:269, print())
[2021-03-22 08:32:03] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 68 @ 59024 updates, score 23.194001090120658) (writing took 62.793537 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:32:53] INFO >> epoch 069: 476 / 868 loss=20.366, nll_loss=1.957, bleu=0, ppl=3.88, wps=2260.8, ups=3.4, wpb=665.6, bsz=64, num_updates=59500, lr=5e-05, gnorm=23.992, clip=27.4, train_wall=43, wall=11334 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:33:31] INFO >> epoch 069 | loss 20.505 | nll_loss 1.968 | bleu 0 | ppl 3.91 | wps 3178.5 | ups 4.77 | wpb 666.6 | bsz 64 | num_updates 59892 | lr 5e-05 | gnorm 24.139 | clip 29.7 | train_wall 77 | wall 11372 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:34:03] INFO >> epoch 069 | valid on 'valid' subset | loss 62.8 | nll_loss 6.017 | bleu 23.2374 | ppl 64.75 | wps 7540.1 | wpb 19314.6 | bsz 1850.5 | num_updates 59892 | best_bleu 23.2374 (progress_bar.py:269, print())
[2021-03-22 08:34:41] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 69 @ 59892 updates, score 23.237414408607446) (writing took 38.117121 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:34:57] INFO >> epoch 070: 108 / 868 loss=20.536, nll_loss=1.968, bleu=0, ppl=3.91, wps=2706.6, ups=4.05, wpb=667.9, bsz=64, num_updates=60000, lr=5e-05, gnorm=24.207, clip=31.6, train_wall=44, wall=11458 (progress_bar.py:260, log())
[2021-03-22 08:35:42] INFO >> epoch 070: 608 / 868 loss=20.193, nll_loss=1.941, bleu=0, ppl=3.84, wps=7282.3, ups=10.93, wpb=666, bsz=64, num_updates=60500, lr=5e-05, gnorm=24.116, clip=28.4, train_wall=43, wall=11503 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:36:07] INFO >> epoch 070 | loss 20.283 | nll_loss 1.947 | bleu 0 | ppl 3.86 | wps 3713.6 | ups 5.57 | wpb 666.6 | bsz 64 | num_updates 60760 | lr 5e-05 | gnorm 24.132 | clip 30.2 | train_wall 75 | wall 11527 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:36:38] INFO >> epoch 070 | valid on 'valid' subset | loss 62.993 | nll_loss 6.035 | bleu 23.3279 | ppl 65.58 | wps 7623.1 | wpb 19314.6 | bsz 1850.5 | num_updates 60760 | best_bleu 23.3279 (progress_bar.py:269, print())
[2021-03-22 08:37:16] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 70 @ 60760 updates, score 23.327937890363607) (writing took 37.925156 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:37:45] INFO >> epoch 071: 240 / 868 loss=20.228, nll_loss=1.942, bleu=0, ppl=3.84, wps=2717.3, ups=4.08, wpb=666.3, bsz=64, num_updates=61000, lr=4.9e-05, gnorm=24.171, clip=32, train_wall=44, wall=11626 (progress_bar.py:260, log())
[2021-03-22 08:38:31] INFO >> epoch 071: 740 / 868 loss=20.113, nll_loss=1.929, bleu=0, ppl=3.81, wps=7299.2, ups=10.95, wpb=666.9, bsz=64, num_updates=61500, lr=4.9e-05, gnorm=24.236, clip=28.8, train_wall=43, wall=11672 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:38:43] INFO >> epoch 071 | loss 20.05 | nll_loss 1.924 | bleu 0 | ppl 3.8 | wps 3700.1 | ups 5.55 | wpb 666.6 | bsz 64 | num_updates 61628 | lr 4.9e-05 | gnorm 24.179 | clip 28.7 | train_wall 76 | wall 11684 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:39:15] INFO >> epoch 071 | valid on 'valid' subset | loss 62.888 | nll_loss 6.025 | bleu 23.3404 | ppl 65.13 | wps 7509.3 | wpb 19314.6 | bsz 1850.5 | num_updates 61628 | best_bleu 23.3404 (progress_bar.py:269, print())
[2021-03-22 08:39:54] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 71 @ 61628 updates, score 23.340434679777537) (writing took 38.945715 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:40:34] INFO >> epoch 072: 372 / 868 loss=19.664, nll_loss=1.891, bleu=0, ppl=3.71, wps=2702.6, ups=4.06, wpb=665.4, bsz=64, num_updates=62000, lr=4.9e-05, gnorm=24.077, clip=28.2, train_wall=43, wall=11795 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:41:20] INFO >> epoch 072 | loss 19.813 | nll_loss 1.902 | bleu 0 | ppl 3.74 | wps 3684.2 | ups 5.53 | wpb 666.6 | bsz 64 | num_updates 62496 | lr 4.9e-05 | gnorm 24.225 | clip 30.4 | train_wall 75 | wall 11841 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:41:53] INFO >> epoch 072 | valid on 'valid' subset | loss 63.177 | nll_loss 6.053 | bleu 23.4297 | ppl 66.39 | wps 7376.5 | wpb 19314.6 | bsz 1850.5 | num_updates 62496 | best_bleu 23.4297 (progress_bar.py:269, print())
[2021-03-22 08:42:31] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 72 @ 62496 updates, score 23.4296708565488) (writing took 38.340873 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:42:37] INFO >> epoch 073: 4 / 868 loss=20.051, nll_loss=1.921, bleu=0, ppl=3.79, wps=2700.7, ups=4.05, wpb=667.6, bsz=64, num_updates=62500, lr=4.8e-05, gnorm=24.383, clip=32.4, train_wall=44, wall=11918 (progress_bar.py:260, log())
[2021-03-22 08:43:25] INFO >> epoch 073: 504 / 868 loss=19.592, nll_loss=1.877, bleu=0, ppl=3.67, wps=7085.1, ups=10.61, wpb=667.9, bsz=64, num_updates=63000, lr=4.8e-05, gnorm=24.21, clip=31.4, train_wall=45, wall=11965 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:44:00] INFO >> epoch 073 | loss 19.582 | nll_loss 1.88 | bleu 0 | ppl 3.68 | wps 3623.6 | ups 5.44 | wpb 666.6 | bsz 64 | num_updates 63364 | lr 4.8e-05 | gnorm 24.259 | clip 31.1 | train_wall 78 | wall 12001 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:44:31] INFO >> epoch 073 | valid on 'valid' subset | loss 63.317 | nll_loss 6.066 | bleu 23.4012 | ppl 67.01 | wps 7650.1 | wpb 19314.6 | bsz 1850.5 | num_updates 63364 | best_bleu 23.4297 (progress_bar.py:269, print())
[2021-03-22 08:45:22] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 73 @ 63364 updates, score 23.401182675616077) (writing took 50.482874 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:45:41] INFO >> epoch 074: 136 / 868 loss=19.416, nll_loss=1.868, bleu=0, ppl=3.65, wps=2442.9, ups=3.67, wpb=665.3, bsz=64, num_updates=63500, lr=4.8e-05, gnorm=24.219, clip=29.8, train_wall=45, wall=12102 (progress_bar.py:260, log())
[2021-03-22 08:46:27] INFO >> epoch 074: 636 / 868 loss=19.311, nll_loss=1.854, bleu=0, ppl=3.61, wps=7252.3, ups=10.88, wpb=666.4, bsz=64, num_updates=64000, lr=4.8e-05, gnorm=24.268, clip=35.4, train_wall=44, wall=12148 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:46:48] INFO >> epoch 074 | loss 19.362 | nll_loss 1.858 | bleu 0 | ppl 3.63 | wps 3433.8 | ups 5.15 | wpb 666.6 | bsz 64 | num_updates 64232 | lr 4.8e-05 | gnorm 24.253 | clip 34.6 | train_wall 76 | wall 12169 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:47:20] INFO >> epoch 074 | valid on 'valid' subset | loss 63.178 | nll_loss 6.053 | bleu 23.4662 | ppl 66.39 | wps 7573.2 | wpb 19314.6 | bsz 1850.5 | num_updates 64232 | best_bleu 23.4662 (progress_bar.py:269, print())
[2021-03-22 08:47:58] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 74 @ 64232 updates, score 23.46624697621288) (writing took 38.297737 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:48:29] INFO >> epoch 075: 268 / 868 loss=19.254, nll_loss=1.85, bleu=0, ppl=3.6, wps=2716.8, ups=4.08, wpb=665.9, bsz=64, num_updates=64500, lr=4.8e-05, gnorm=24.227, clip=31, train_wall=43, wall=12270 (progress_bar.py:260, log())
[2021-03-22 08:49:17] INFO >> epoch 075: 768 / 868 loss=19.219, nll_loss=1.842, bleu=0, ppl=3.59, wps=6946, ups=10.4, wpb=667.7, bsz=64, num_updates=65000, lr=4.8e-05, gnorm=24.373, clip=36.2, train_wall=46, wall=12318 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:49:27] INFO >> epoch 075 | loss 19.126 | nll_loss 1.836 | bleu 0 | ppl 3.57 | wps 3648.8 | ups 5.47 | wpb 666.6 | bsz 64 | num_updates 65100 | lr 4.8e-05 | gnorm 24.308 | clip 32.8 | train_wall 77 | wall 12328 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:50:00] INFO >> epoch 075 | valid on 'valid' subset | loss 63.299 | nll_loss 6.065 | bleu 23.5644 | ppl 66.93 | wps 7340.1 | wpb 19314.6 | bsz 1850.5 | num_updates 65100 | best_bleu 23.5644 (progress_bar.py:269, print())
[2021-03-22 08:50:38] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 75 @ 65100 updates, score 23.56440128468919) (writing took 38.519787 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:51:25] INFO >> epoch 076: 400 / 868 loss=18.812, nll_loss=1.811, bleu=0, ppl=3.51, wps=2597.7, ups=3.91, wpb=664.6, bsz=64, num_updates=65500, lr=4.7e-05, gnorm=24.176, clip=29.8, train_wall=47, wall=12446 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:52:11] INFO >> epoch 076 | loss 18.934 | nll_loss 1.817 | bleu 0 | ppl 3.52 | wps 3514.9 | ups 5.27 | wpb 666.6 | bsz 64 | num_updates 65968 | lr 4.7e-05 | gnorm 24.291 | clip 32.9 | train_wall 82 | wall 12492 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:52:45] INFO >> epoch 076 | valid on 'valid' subset | loss 63.433 | nll_loss 6.077 | bleu 23.7378 | ppl 67.53 | wps 7134.1 | wpb 19314.6 | bsz 1850.5 | num_updates 65968 | best_bleu 23.7378 (progress_bar.py:269, print())
[2021-03-22 08:53:50] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 76 @ 65968 updates, score 23.73777007471941) (writing took 64.630623 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:53:59] INFO >> epoch 077: 32 / 868 loss=19.084, nll_loss=1.829, bleu=0, ppl=3.55, wps=2174.8, ups=3.26, wpb=667.9, bsz=64, num_updates=66000, lr=4.7e-05, gnorm=24.429, clip=36, train_wall=46, wall=12600 (progress_bar.py:260, log())
[2021-03-22 08:54:45] INFO >> epoch 077: 532 / 868 loss=18.613, nll_loss=1.788, bleu=0, ppl=3.45, wps=7176.9, ups=10.77, wpb=666.3, bsz=64, num_updates=66500, lr=4.7e-05, gnorm=24.179, clip=28.8, train_wall=44, wall=12646 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:55:17] INFO >> epoch 077 | loss 18.735 | nll_loss 1.798 | bleu 0 | ppl 3.48 | wps 3117.7 | ups 4.68 | wpb 666.6 | bsz 64 | num_updates 66836 | lr 4.7e-05 | gnorm 24.287 | clip 31 | train_wall 76 | wall 12678 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:55:49] INFO >> epoch 077 | valid on 'valid' subset | loss 63.61 | nll_loss 6.094 | bleu 23.6722 | ppl 68.32 | wps 7604.5 | wpb 19314.6 | bsz 1850.5 | num_updates 66836 | best_bleu 23.7378 (progress_bar.py:269, print())
[2021-03-22 08:56:09] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 77 @ 66836 updates, score 23.672238508529382) (writing took 20.140915 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:56:30] INFO >> epoch 078: 164 / 868 loss=18.701, nll_loss=1.794, bleu=0, ppl=3.47, wps=3174.3, ups=4.76, wpb=666.7, bsz=64, num_updates=67000, lr=4.6e-05, gnorm=24.356, clip=32.6, train_wall=44, wall=12751 (progress_bar.py:260, log())
[2021-03-22 08:57:17] INFO >> epoch 078: 664 / 868 loss=18.506, nll_loss=1.775, bleu=0, ppl=3.42, wps=7165.7, ups=10.74, wpb=667.3, bsz=64, num_updates=67500, lr=4.6e-05, gnorm=24.309, clip=34.6, train_wall=44, wall=12798 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:57:36] INFO >> epoch 078 | loss 18.478 | nll_loss 1.774 | bleu 0 | ppl 3.42 | wps 4156.5 | ups 6.24 | wpb 666.6 | bsz 64 | num_updates 67704 | lr 4.6e-05 | gnorm 24.294 | clip 33.2 | train_wall 76 | wall 12817 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 08:58:08] INFO >> epoch 078 | valid on 'valid' subset | loss 63.541 | nll_loss 6.088 | bleu 23.8618 | ppl 68.01 | wps 7576 | wpb 19314.6 | bsz 1850.5 | num_updates 67704 | best_bleu 23.8618 (progress_bar.py:269, print())
[2021-03-22 08:58:46] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 78 @ 67704 updates, score 23.86183555744708) (writing took 38.292508 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 08:59:20] INFO >> epoch 079: 296 / 868 loss=18.421, nll_loss=1.762, bleu=0, ppl=3.39, wps=2707.3, ups=4.05, wpb=668.7, bsz=64, num_updates=68000, lr=4.6e-05, gnorm=24.364, clip=32.8, train_wall=44, wall=12921 (progress_bar.py:260, log())
[2021-03-22 09:00:07] INFO >> epoch 079: 796 / 868 loss=18.369, nll_loss=1.765, bleu=0, ppl=3.4, wps=7193.3, ups=10.8, wpb=666, bsz=64, num_updates=68500, lr=4.6e-05, gnorm=24.459, clip=36, train_wall=44, wall=12967 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:00:14] INFO >> epoch 079 | loss 18.307 | nll_loss 1.757 | bleu 0 | ppl 3.38 | wps 3671.4 | ups 5.51 | wpb 666.6 | bsz 64 | num_updates 68572 | lr 4.6e-05 | gnorm 24.383 | clip 34.4 | train_wall 77 | wall 12975 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:00:47] INFO >> epoch 079 | valid on 'valid' subset | loss 63.676 | nll_loss 6.101 | bleu 23.7498 | ppl 68.63 | wps 7323.3 | wpb 19314.6 | bsz 1850.5 | num_updates 68572 | best_bleu 23.8618 (progress_bar.py:269, print())
[2021-03-22 09:01:07] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 79 @ 68572 updates, score 23.749801076748607) (writing took 20.046212 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:01:53] INFO >> epoch 080: 428 / 868 loss=17.987, nll_loss=1.73, bleu=0, ppl=3.32, wps=3136.6, ups=4.71, wpb=665.5, bsz=64, num_updates=69000, lr=4.5e-05, gnorm=24.302, clip=31.4, train_wall=44, wall=13074 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:02:34] INFO >> epoch 080 | loss 18.13 | nll_loss 1.74 | bleu 0 | ppl 3.34 | wps 4120.8 | ups 6.18 | wpb 666.6 | bsz 64 | num_updates 69440 | lr 4.5e-05 | gnorm 24.433 | clip 35 | train_wall 77 | wall 13115 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:03:07] INFO >> epoch 080 | valid on 'valid' subset | loss 63.961 | nll_loss 6.128 | bleu 23.8982 | ppl 69.94 | wps 7410.4 | wpb 19314.6 | bsz 1850.5 | num_updates 69440 | best_bleu 23.8982 (progress_bar.py:269, print())
[2021-03-22 09:04:00] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 80 @ 69440 updates, score 23.898170941930605) (writing took 53.030980 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:04:11] INFO >> epoch 081: 60 / 868 loss=18.214, nll_loss=1.749, bleu=0, ppl=3.36, wps=2400.5, ups=3.6, wpb=666.1, bsz=64, num_updates=69500, lr=4.5e-05, gnorm=24.491, clip=37.4, train_wall=44, wall=13212 (progress_bar.py:260, log())
[2021-03-22 09:04:58] INFO >> epoch 081: 560 / 868 loss=17.962, nll_loss=1.72, bleu=0, ppl=3.29, wps=7166.8, ups=10.72, wpb=668.5, bsz=64, num_updates=70000, lr=4.5e-05, gnorm=24.432, clip=36.6, train_wall=44, wall=13259 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:05:27] INFO >> epoch 081 | loss 17.907 | nll_loss 1.719 | bleu 0 | ppl 3.29 | wps 3349.3 | ups 5.02 | wpb 666.6 | bsz 64 | num_updates 70308 | lr 4.5e-05 | gnorm 24.38 | clip 35.1 | train_wall 76 | wall 13288 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:06:00] INFO >> epoch 081 | valid on 'valid' subset | loss 63.955 | nll_loss 6.127 | bleu 23.8301 | ppl 69.91 | wps 7319.7 | wpb 19314.6 | bsz 1850.5 | num_updates 70308 | best_bleu 23.8982 (progress_bar.py:269, print())
[2021-03-22 09:06:37] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 81 @ 70308 updates, score 23.83010421450747) (writing took 37.375588 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:07:02] INFO >> epoch 082: 192 / 868 loss=17.669, nll_loss=1.702, bleu=0, ppl=3.25, wps=2674.8, ups=4.03, wpb=664.1, bsz=64, num_updates=70500, lr=4.4e-05, gnorm=24.268, clip=32.2, train_wall=45, wall=13383 (progress_bar.py:260, log())
[2021-03-22 09:07:49] INFO >> epoch 082: 692 / 868 loss=17.836, nll_loss=1.708, bleu=0, ppl=3.27, wps=7158, ups=10.72, wpb=668, bsz=64, num_updates=71000, lr=4.4e-05, gnorm=24.437, clip=35.2, train_wall=44, wall=13430 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:08:06] INFO >> epoch 082 | loss 17.734 | nll_loss 1.702 | bleu 0 | ppl 3.25 | wps 3645.5 | ups 5.47 | wpb 666.6 | bsz 64 | num_updates 71176 | lr 4.4e-05 | gnorm 24.395 | clip 34 | train_wall 78 | wall 13447 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:08:38] INFO >> epoch 082 | valid on 'valid' subset | loss 64.017 | nll_loss 6.133 | bleu 23.8488 | ppl 70.2 | wps 7501.7 | wpb 19314.6 | bsz 1850.5 | num_updates 71176 | best_bleu 23.8982 (progress_bar.py:269, print())
[2021-03-22 09:08:58] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 82 @ 71176 updates, score 23.848831496646714) (writing took 19.857150 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:09:34] INFO >> epoch 083: 324 / 868 loss=17.55, nll_loss=1.688, bleu=0, ppl=3.22, wps=3168.9, ups=4.76, wpb=665.3, bsz=64, num_updates=71500, lr=4.4e-05, gnorm=24.376, clip=31.4, train_wall=44, wall=13535 (progress_bar.py:260, log())
[2021-03-22 09:10:20] INFO >> epoch 083: 824 / 868 loss=17.609, nll_loss=1.688, bleu=0, ppl=3.22, wps=7185.6, ups=10.77, wpb=667.3, bsz=64, num_updates=72000, lr=4.4e-05, gnorm=24.503, clip=37.6, train_wall=44, wall=13581 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:10:25] INFO >> epoch 083 | loss 17.554 | nll_loss 1.685 | bleu 0 | ppl 3.22 | wps 4157.3 | ups 6.24 | wpb 666.6 | bsz 64 | num_updates 72044 | lr 4.4e-05 | gnorm 24.429 | clip 34.4 | train_wall 76 | wall 13586 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:10:57] INFO >> epoch 083 | valid on 'valid' subset | loss 64.061 | nll_loss 6.138 | bleu 24.0487 | ppl 70.4 | wps 7531.2 | wpb 19314.6 | bsz 1850.5 | num_updates 72044 | best_bleu 24.0487 (progress_bar.py:269, print())
[2021-03-22 09:11:36] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 83 @ 72044 updates, score 24.048741076514602) (writing took 38.535186 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:12:24] INFO >> epoch 084: 456 / 868 loss=17.405, nll_loss=1.667, bleu=0, ppl=3.18, wps=2693.3, ups=4.03, wpb=668.1, bsz=64, num_updates=72500, lr=4.3e-05, gnorm=24.44, clip=35.6, train_wall=44, wall=13705 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:13:03] INFO >> epoch 084 | loss 17.355 | nll_loss 1.666 | bleu 0 | ppl 3.17 | wps 3660.5 | ups 5.49 | wpb 666.6 | bsz 64 | num_updates 72912 | lr 4.3e-05 | gnorm 24.424 | clip 34.8 | train_wall 77 | wall 13744 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:13:35] INFO >> epoch 084 | valid on 'valid' subset | loss 64.272 | nll_loss 6.158 | bleu 23.9409 | ppl 71.4 | wps 7546.8 | wpb 19314.6 | bsz 1850.5 | num_updates 72912 | best_bleu 24.0487 (progress_bar.py:269, print())
[2021-03-22 09:13:55] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 84 @ 72912 updates, score 23.940924573842334) (writing took 19.897143 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:14:09] INFO >> epoch 085: 88 / 868 loss=17.277, nll_loss=1.663, bleu=0, ppl=3.17, wps=3168.3, ups=4.77, wpb=664.6, bsz=64, num_updates=73000, lr=4.3e-05, gnorm=24.349, clip=32.4, train_wall=44, wall=13810 (progress_bar.py:260, log())
[2021-03-22 09:14:56] INFO >> epoch 085: 588 / 868 loss=17.102, nll_loss=1.64, bleu=0, ppl=3.12, wps=7145, ups=10.71, wpb=667.2, bsz=64, num_updates=73500, lr=4.3e-05, gnorm=24.371, clip=36, train_wall=44, wall=13857 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:15:22] INFO >> epoch 085 | loss 17.138 | nll_loss 1.645 | bleu 0 | ppl 3.13 | wps 4151.5 | ups 6.23 | wpb 666.6 | bsz 64 | num_updates 73780 | lr 4.3e-05 | gnorm 24.382 | clip 35.3 | train_wall 77 | wall 13883 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:15:55] INFO >> epoch 085 | valid on 'valid' subset | loss 64.271 | nll_loss 6.158 | bleu 23.9422 | ppl 71.39 | wps 7436.2 | wpb 19314.6 | bsz 1850.5 | num_updates 73780 | best_bleu 24.0487 (progress_bar.py:269, print())
[2021-03-22 09:16:15] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 85 @ 73780 updates, score 23.94220156956254) (writing took 19.850440 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:16:41] INFO >> epoch 086: 220 / 868 loss=17.053, nll_loss=1.639, bleu=0, ppl=3.11, wps=3166.2, ups=4.76, wpb=665.6, bsz=64, num_updates=74000, lr=4.3e-05, gnorm=24.399, clip=35.4, train_wall=44, wall=13962 (progress_bar.py:260, log())
[2021-03-22 09:17:27] INFO >> epoch 086: 720 / 868 loss=17.053, nll_loss=1.635, bleu=0, ppl=3.11, wps=7195.5, ups=10.78, wpb=667.5, bsz=64, num_updates=74500, lr=4.3e-05, gnorm=24.503, clip=38.2, train_wall=44, wall=14008 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:17:42] INFO >> epoch 086 | loss 16.992 | nll_loss 1.631 | bleu 0 | ppl 3.1 | wps 4145.5 | ups 6.22 | wpb 666.6 | bsz 64 | num_updates 74648 | lr 4.3e-05 | gnorm 24.445 | clip 36.4 | train_wall 77 | wall 14023 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:18:15] INFO >> epoch 086 | valid on 'valid' subset | loss 64.294 | nll_loss 6.16 | bleu 24.1142 | ppl 71.5 | wps 7349.7 | wpb 19314.6 | bsz 1850.5 | num_updates 74648 | best_bleu 24.1142 (progress_bar.py:269, print())
[2021-03-22 09:19:28] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 86 @ 74648 updates, score 24.114249859163188) (writing took 73.541307 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:20:07] INFO >> epoch 087: 352 / 868 loss=16.864, nll_loss=1.619, bleu=0, ppl=3.07, wps=2081.5, ups=3.12, wpb=666.4, bsz=64, num_updates=75000, lr=4.2e-05, gnorm=24.447, clip=34.2, train_wall=45, wall=14168 (progress_bar.py:260, log())
[2021-03-22 09:20:54] INFO >> epoch 087: 852 / 868 loss=16.837, nll_loss=1.617, bleu=0, ppl=3.07, wps=7194.3, ups=10.79, wpb=666.6, bsz=64, num_updates=75500, lr=4.2e-05, gnorm=24.483, clip=36.2, train_wall=44, wall=14215 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:20:56] INFO >> epoch 087 | loss 16.804 | nll_loss 1.613 | bleu 0 | ppl 3.06 | wps 2985.4 | ups 4.48 | wpb 666.6 | bsz 64 | num_updates 75516 | lr 4.2e-05 | gnorm 24.448 | clip 35.1 | train_wall 77 | wall 14217 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:21:28] INFO >> epoch 087 | valid on 'valid' subset | loss 64.333 | nll_loss 6.164 | bleu 24.0233 | ppl 71.69 | wps 7418.2 | wpb 19314.6 | bsz 1850.5 | num_updates 75516 | best_bleu 24.1142 (progress_bar.py:269, print())
[2021-03-22 09:21:48] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_last.pt (epoch 87 @ 75516 updates, score 24.023289221918503) (writing took 19.926355 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:22:39] INFO >> epoch 088: 484 / 868 loss=16.556, nll_loss=1.592, bleu=0, ppl=3.01, wps=3152.7, ups=4.74, wpb=665.5, bsz=64, num_updates=76000, lr=4.2e-05, gnorm=24.4, clip=36, train_wall=44, wall=14320 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:23:15] INFO >> epoch 088 | loss 16.665 | nll_loss 1.6 | bleu 0 | ppl 3.03 | wps 4140.8 | ups 6.21 | wpb 666.6 | bsz 64 | num_updates 76384 | lr 4.2e-05 | gnorm 24.517 | clip 37.1 | train_wall 76 | wall 14356 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:23:48] INFO >> epoch 088 | valid on 'valid' subset | loss 64.498 | nll_loss 6.179 | bleu 24.1327 | ppl 72.47 | wps 7436.1 | wpb 19314.6 | bsz 1850.5 | num_updates 76384 | best_bleu 24.1327 (progress_bar.py:269, print())
[2021-03-22 09:25:02] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 88 @ 76384 updates, score 24.13272684349318) (writing took 73.755863 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:25:18] INFO >> epoch 089: 116 / 868 loss=16.696, nll_loss=1.601, bleu=0, ppl=3.03, wps=2098, ups=3.15, wpb=667, bsz=63.9, num_updates=76500, lr=4.1e-05, gnorm=24.596, clip=37.2, train_wall=44, wall=14479 (progress_bar.py:260, log())
[2021-03-22 09:26:05] INFO >> epoch 089: 616 / 868 loss=16.469, nll_loss=1.584, bleu=0, ppl=3, wps=7150.8, ups=10.75, wpb=665.5, bsz=64, num_updates=77000, lr=4.1e-05, gnorm=24.44, clip=36.6, train_wall=44, wall=14526 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:26:29] INFO >> epoch 089 | loss 16.516 | nll_loss 1.585 | bleu 0 | ppl 3 | wps 2991.3 | ups 4.49 | wpb 666.6 | bsz 64 | num_updates 77252 | lr 4.1e-05 | gnorm 24.501 | clip 36.9 | train_wall 76 | wall 14550 (progress_bar.py:269, print())
Using backend: pytorch
Using backend: pytorch
Using backend: pytorch
[2021-03-22 09:27:01] INFO >> epoch 089 | valid on 'valid' subset | loss 64.456 | nll_loss 6.175 | bleu 24.2961 | ppl 72.28 | wps 7397.7 | wpb 19314.6 | bsz 1850.5 | num_updates 77252 | best_bleu 24.2961 (progress_bar.py:269, print())
[2021-03-22 09:27:40] INFO >> saved checkpoint /mnt/wanyao/.ncc/python_wan/summarization/data-mmap/seq2seq/checkpoints/checkpoint_best.pt (epoch 89 @ 77252 updates, score 24.296080825837244) (writing took 38.562715 seconds) (checkpoint_utils.py:79, save_checkpoint())
[2021-03-22 09:28:10] INFO >> epoch 090: 248 / 868 loss=16.543, nll_loss=1.581, bleu=0, ppl=2.99, wps=2683.3, ups=4.01, wpb=669.3, bsz=64, num_updates=77500, lr=4.1e-05, gnorm=24.62, clip=38.8, train_wall=45, wall=14650 (progress_bar.py:260, log())
[2021-03-22 09:28:56] INFO >> epoch 090: 748 / 868 loss=16.295, nll_loss=1.563, bleu=0, ppl=2.96, wps=7192, ups=10.78, wpb=667, bsz=64, num_updates=78000, lr=4.1e-05, gnorm=24.481, clip=38.2, train_wall=44, wall=14697 (progress_bar.py:260, log())
Using backend: pytorch
Using backend: pytorch