-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbest_models.attLSTM
1027 lines (1013 loc) · 99.1 KB
/
best_models.attLSTM
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
ignacio@np13:~$ python $NLP/attention_lstm_1.py
Using Theano backend.
Loanding train and valid dirs......
Starting training
Phrases in STS.input.FNWN.txt 189 189
Phrases in STS.input.OnWN.txt 561 561
Phrases in STS.input.headlines.txt 750 750
Total train phrases /almac/ignacio/data/sts_all/train-2013 1500
Total train phrases 3000
Starting training
Phrases in STS.input.track5.en-en.txt 250
Total train phrases /almac/ignacio/data/sts_all/valid-2017 250
Total train phrases 500
Spliting tab-separated files...
Labels shape: (1500,)
Tokenizing files... [A]
Split training set into train and val... [A]
Tokenizing files... [B]
Split training set into train and val... [B]
Getting embedding matrix...
Found 4049540 word vectors.
Filling embedding matrices...
____________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
====================================================================================================
embedding_1 (Embedding) (None, 50, 50) 185950
____________________________________________________________________________________________________
attention_1 (Attention) (None, 50, 25) 8900
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 25) 2575
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 25) 2575
____________________________________________________________________________________________________
attention_4 (Attention) (None, 25) 2575
____________________________________________________________________________________________________
activation_1 (Activation) (None, 25) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 50) 212200
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 25) 8900
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 25) 2575
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 25) 2575
____________________________________________________________________________________________________
attention_8 (Attention) (None, 25) 2575
____________________________________________________________________________________________________
activation_2 (Activation) (None, 25) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 10) 2040 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 44 maxoutdense_1[0][0]
====================================================================================================
Total params: 433,484
Trainable params: 35,334
Non-trainable params: 398,150
____________________________________________________________________________________________________
None
Train on 1050 samples, validate on 450 samples
Epoch 1/100
1050/1050 [==============================] - 193s - loss: 1.4472 - mean_absolute_error: 1.4472 - mean_squared_error: 2.7143 - val_loss: 1.3324 - val_mean_absolute_error: 1.3324 - val_mean_squared_error: 2.2458
Epoch 2/100
1050/1050 [==============================] - 192s - loss: 1.4084 - mean_absolute_error: 1.4084 - mean_squared_error: 2.5455 - val_loss: 1.3294 - val_mean_absolute_error: 1.3294 - val_mean_squared_error: 2.2452
Epoch 3/100
1050/1050 [==============================] - 196s - loss: 1.4026 - mean_absolute_error: 1.4026 - mean_squared_error: 2.5020 - val_loss: 1.3339 - val_mean_absolute_error: 1.3339 - val_mean_squared_error: 2.3053
Epoch 4/100
1050/1050 [==============================] - 195s - loss: 1.3939 - mean_absolute_error: 1.3939 - mean_squared_error: 2.5028 - val_loss: 1.3022 - val_mean_absolute_error: 1.3022 - val_mean_squared_error: 2.1776
Epoch 5/100
1050/1050 [==============================] - 193s - loss: 1.3988 - mean_absolute_error: 1.3988 - mean_squared_error: 2.5061 - val_loss: 1.2848 - val_mean_absolute_error: 1.2848 - val_mean_squared_error: 2.1189
Epoch 6/100
1050/1050 [==============================] - 191s - loss: 1.3460 - mean_absolute_error: 1.3460 - mean_squared_error: 2.4301 - val_loss: 1.2132 - val_mean_absolute_error: 1.2132 - val_mean_squared_error: 1.9379
Epoch 7/100
1050/1050 [==============================] - 195s - loss: 1.3257 - mean_absolute_error: 1.3257 - mean_squared_error: 2.3989 - val_loss: 1.2495 - val_mean_absolute_error: 1.2495 - val_mean_squared_error: 2.1899
Epoch 8/100
1050/1050 [==============================] - 195s - loss: 1.2884 - mean_absolute_error: 1.2884 - mean_squared_error: 2.3543 - val_loss: 1.3250 - val_mean_absolute_error: 1.3250 - val_mean_squared_error: 2.5787
Epoch 9/100
1050/1050 [==============================] - 201s - loss: 1.3130 - mean_absolute_error: 1.3130 - mean_squared_error: 2.3793 - val_loss: 1.1966 - val_mean_absolute_error: 1.1966 - val_mean_squared_error: 1.9480
Epoch 10/100
1050/1050 [==============================] - 195s - loss: 1.2762 - mean_absolute_error: 1.2762 - mean_squared_error: 2.3427 - val_loss: 1.1831 - val_mean_absolute_error: 1.1831 - val_mean_squared_error: 1.9553
Epoch 11/100
1050/1050 [==============================] - 203s - loss: 1.2525 - mean_absolute_error: 1.2525 - mean_squared_error: 2.2505 - val_loss: 1.1663 - val_mean_absolute_error: 1.1663 - val_mean_squared_error: 1.8784
Epoch 12/100
1050/1050 [==============================] - 203s - loss: 1.2791 - mean_absolute_error: 1.2791 - mean_squared_error: 2.4150 - val_loss: 1.1392 - val_mean_absolute_error: 1.1392 - val_mean_squared_error: 1.9144
Epoch 13/100
1050/1050 [==============================] - 195s - loss: 1.2549 - mean_absolute_error: 1.2549 - mean_squared_error: 2.2776 - val_loss: 1.1468 - val_mean_absolute_error: 1.1468 - val_mean_squared_error: 1.9456
Epoch 14/100
1050/1050 [==============================] - 193s - loss: 1.2325 - mean_absolute_error: 1.2325 - mean_squared_error: 2.2474 - val_loss: 1.1419 - val_mean_absolute_error: 1.1419 - val_mean_squared_error: 1.8419
Epoch 15/100
1050/1050 [==============================] - 197s - loss: 1.2381 - mean_absolute_error: 1.2381 - mean_squared_error: 2.2374 - val_loss: 1.1283 - val_mean_absolute_error: 1.1283 - val_mean_squared_error: 1.8320
Epoch 16/100
1050/1050 [==============================] - 201s - loss: 1.2248 - mean_absolute_error: 1.2248 - mean_squared_error: 2.2236 - val_loss: 1.1298 - val_mean_absolute_error: 1.1298 - val_mean_squared_error: 1.9232
Epoch 17/100
1050/1050 [==============================] - 199s - loss: 1.2246 - mean_absolute_error: 1.2246 - mean_squared_error: 2.2893 - val_loss: 1.2010 - val_mean_absolute_error: 1.2010 - val_mean_squared_error: 2.1447
Epoch 18/100
1050/1050 [==============================] - 196s - loss: 1.2175 - mean_absolute_error: 1.2175 - mean_squared_error: 2.1963 - val_loss: 1.1176 - val_mean_absolute_error: 1.1176 - val_mean_squared_error: 1.8073
Epoch 19/100
1050/1050 [==============================] - 199s - loss: 1.2114 - mean_absolute_error: 1.2114 - mean_squared_error: 2.1534 - val_loss: 1.1765 - val_mean_absolute_error: 1.1765 - val_mean_squared_error: 2.0621
Epoch 20/100
1050/1050 [==============================] - 195s - loss: 1.2109 - mean_absolute_error: 1.2109 - mean_squared_error: 2.2463 - val_loss: 1.1313 - val_mean_absolute_error: 1.1313 - val_mean_squared_error: 1.9421
Epoch 21/100
1050/1050 [==============================] - 195s - loss: 1.1793 - mean_absolute_error: 1.1793 - mean_squared_error: 2.1221 - val_loss: 1.1137 - val_mean_absolute_error: 1.1137 - val_mean_squared_error: 2.0367
Epoch 22/100
1050/1050 [==============================] - 197s - loss: 1.2094 - mean_absolute_error: 1.2094 - mean_squared_error: 2.2578 - val_loss: 1.1141 - val_mean_absolute_error: 1.1141 - val_mean_squared_error: 1.7991
Epoch 23/100
1050/1050 [==============================] - 194s - loss: 1.1895 - mean_absolute_error: 1.1895 - mean_squared_error: 2.1457 - val_loss: 1.1428 - val_mean_absolute_error: 1.1428 - val_mean_squared_error: 2.1624
Epoch 24/100
1050/1050 [==============================] - 194s - loss: 1.1438 - mean_absolute_error: 1.1438 - mean_squared_error: 2.0956 - val_loss: 1.1197 - val_mean_absolute_error: 1.1197 - val_mean_squared_error: 1.8958
Epoch 25/100
1050/1050 [==============================] - 196s - loss: 1.1587 - mean_absolute_error: 1.1587 - mean_squared_error: 2.0661 - val_loss: 1.1269 - val_mean_absolute_error: 1.1269 - val_mean_squared_error: 1.9923
Epoch 26/100
1050/1050 [==============================] - 198s - loss: 1.1695 - mean_absolute_error: 1.1695 - mean_squared_error: 2.1069 - val_loss: 1.1239 - val_mean_absolute_error: 1.1239 - val_mean_squared_error: 2.1451
Epoch 27/100
1050/1050 [==============================] - 195s - loss: 1.1167 - mean_absolute_error: 1.1167 - mean_squared_error: 1.9691 - val_loss: 1.1244 - val_mean_absolute_error: 1.1244 - val_mean_squared_error: 2.0383
Epoch 28/100
1050/1050 [==============================] - 194s - loss: 1.1432 - mean_absolute_error: 1.1432 - mean_squared_error: 2.1146 - val_loss: 1.1609 - val_mean_absolute_error: 1.1609 - val_mean_squared_error: 2.0065
Epoch 29/100
1050/1050 [==============================] - 197s - loss: 1.1411 - mean_absolute_error: 1.1411 - mean_squared_error: 2.0925 - val_loss: 1.1149 - val_mean_absolute_error: 1.1149 - val_mean_squared_error: 1.9947
Epoch 30/100
1050/1050 [==============================] - 198s - loss: 1.1431 - mean_absolute_error: 1.1431 - mean_squared_error: 2.0592 - val_loss: 1.1064 - val_mean_absolute_error: 1.1064 - val_mean_squared_error: 1.9542
Epoch 31/100
1050/1050 [==============================] - 201s - loss: 1.1359 - mean_absolute_error: 1.1359 - mean_squared_error: 2.0429 - val_loss: 1.1422 - val_mean_absolute_error: 1.1422 - val_mean_squared_error: 2.1531
Epoch 32/100
1050/1050 [==============================] - 207s - loss: 1.1060 - mean_absolute_error: 1.1060 - mean_squared_error: 1.9652 - val_loss: 1.1476 - val_mean_absolute_error: 1.1476 - val_mean_squared_error: 2.1878
Epoch 33/100
1050/1050 [==============================] - 203s - loss: 1.1169 - mean_absolute_error: 1.1169 - mean_squared_error: 1.9998 - val_loss: 1.1204 - val_mean_absolute_error: 1.1204 - val_mean_squared_error: 1.9927
Epoch 34/100
1050/1050 [==============================] - 200s - loss: 1.1279 - mean_absolute_error: 1.1279 - mean_squared_error: 2.0551 - val_loss: 1.1257 - val_mean_absolute_error: 1.1257 - val_mean_squared_error: 2.0321
Epoch 35/100
1050/1050 [==============================] - 197s - loss: 1.0936 - mean_absolute_error: 1.0936 - mean_squared_error: 1.9594 - val_loss: 1.1445 - val_mean_absolute_error: 1.1445 - val_mean_squared_error: 2.2837
Epoch 36/100
1050/1050 [==============================] - 196s - loss: 1.1026 - mean_absolute_error: 1.1026 - mean_squared_error: 2.0052 - val_loss: 1.1274 - val_mean_absolute_error: 1.1274 - val_mean_squared_error: 1.9724
Epoch 37/100
1050/1050 [==============================] - 200s - loss: 1.0919 - mean_absolute_error: 1.0919 - mean_squared_error: 1.9625 - val_loss: 1.1505 - val_mean_absolute_error: 1.1505 - val_mean_squared_error: 2.0624
Epoch 38/100
1050/1050 [==============================] - 197s - loss: 1.0703 - mean_absolute_error: 1.0703 - mean_squared_error: 1.8320 - val_loss: 1.1972 - val_mean_absolute_error: 1.1972 - val_mean_squared_error: 2.3799
Epoch 39/100
1050/1050 [==============================] - 195s - loss: 1.0872 - mean_absolute_error: 1.0872 - mean_squared_error: 1.9688 - val_loss: 1.1677 - val_mean_absolute_error: 1.1677 - val_mean_squared_error: 2.1061
Epoch 40/100
1050/1050 [==============================] - 194s - loss: 1.0855 - mean_absolute_error: 1.0855 - mean_squared_error: 1.9955 - val_loss: 1.1008 - val_mean_absolute_error: 1.1008 - val_mean_squared_error: 1.9622
Epoch 41/100
1050/1050 [==============================] - 194s - loss: 1.0770 - mean_absolute_error: 1.0770 - mean_squared_error: 1.9100 - val_loss: 1.1566 - val_mean_absolute_error: 1.1566 - val_mean_squared_error: 2.1571
Epoch 42/100
1050/1050 [==============================] - 195s - loss: 1.0699 - mean_absolute_error: 1.0699 - mean_squared_error: 1.9204 - val_loss: 1.1246 - val_mean_absolute_error: 1.1246 - val_mean_squared_error: 1.9581
Epoch 43/100
1050/1050 [==============================] - 194s - loss: 1.0509 - mean_absolute_error: 1.0509 - mean_squared_error: 1.8357 - val_loss: 1.1475 - val_mean_absolute_error: 1.1475 - val_mean_squared_error: 2.2722
Epoch 44/100
1050/1050 [==============================] - 194s - loss: 1.0604 - mean_absolute_error: 1.0604 - mean_squared_error: 1.9418 - val_loss: 1.1514 - val_mean_absolute_error: 1.1514 - val_mean_squared_error: 2.1891
Epoch 45/100
1050/1050 [==============================] - 194s - loss: 1.0264 - mean_absolute_error: 1.0264 - mean_squared_error: 1.8166 - val_loss: 1.1348 - val_mean_absolute_error: 1.1348 - val_mean_squared_error: 2.1075
Epoch 46/100
1050/1050 [==============================] - 193s - loss: 1.0326 - mean_absolute_error: 1.0326 - mean_squared_error: 1.8264 - val_loss: 1.1928 - val_mean_absolute_error: 1.1928 - val_mean_squared_error: 2.3400
Epoch 47/100
1050/1050 [==============================] - 197s - loss: 1.0298 - mean_absolute_error: 1.0298 - mean_squared_error: 1.8175 - val_loss: 1.1663 - val_mean_absolute_error: 1.1663 - val_mean_squared_error: 2.2285
Epoch 48/100
1050/1050 [==============================] - 197s - loss: 1.0493 - mean_absolute_error: 1.0493 - mean_squared_error: 1.8643 - val_loss: 1.1652 - val_mean_absolute_error: 1.1652 - val_mean_squared_error: 2.3336
Epoch 49/100
1050/1050 [==============================] - 194s - loss: 1.0439 - mean_absolute_error: 1.0439 - mean_squared_error: 1.8112 - val_loss: 1.1462 - val_mean_absolute_error: 1.1462 - val_mean_squared_error: 2.1684
Epoch 50/100
1050/1050 [==============================] - 194s - loss: 1.0393 - mean_absolute_error: 1.0393 - mean_squared_error: 1.8862 - val_loss: 1.1698 - val_mean_absolute_error: 1.1698 - val_mean_squared_error: 2.3521
Epoch 51/100
1050/1050 [==============================] - 194s - loss: 1.0244 - mean_absolute_error: 1.0244 - mean_squared_error: 1.8186 - val_loss: 1.1566 - val_mean_absolute_error: 1.1566 - val_mean_squared_error: 2.2483
Epoch 52/100
1050/1050 [==============================] - 196s - loss: 1.0264 - mean_absolute_error: 1.0264 - mean_squared_error: 1.8264 - val_loss: 1.1281 - val_mean_absolute_error: 1.1281 - val_mean_squared_error: 2.1981
Epoch 53/100
1050/1050 [==============================] - 194s - loss: 0.9788 - mean_absolute_error: 0.9788 - mean_squared_error: 1.6856 - val_loss: 1.1805 - val_mean_absolute_error: 1.1805 - val_mean_squared_error: 2.3246
Epoch 54/100
1050/1050 [==============================] - 193s - loss: 0.9980 - mean_absolute_error: 0.9980 - mean_squared_error: 1.7019 - val_loss: 1.1164 - val_mean_absolute_error: 1.1164 - val_mean_squared_error: 2.0692
Epoch 55/100
1050/1050 [==============================] - 194s - loss: 0.9960 - mean_absolute_error: 0.9960 - mean_squared_error: 1.7555 - val_loss: 1.1628 - val_mean_absolute_error: 1.1628 - val_mean_squared_error: 2.3504
Epoch 56/100
1050/1050 [==============================] - 193s - loss: 1.0177 - mean_absolute_error: 1.0177 - mean_squared_error: 1.8193 - val_loss: 1.2132 - val_mean_absolute_error: 1.2132 - val_mean_squared_error: 2.5549
Epoch 57/100
1050/1050 [==============================] - 194s - loss: 0.9938 - mean_absolute_error: 0.9938 - mean_squared_error: 1.7733 - val_loss: 1.2112 - val_mean_absolute_error: 1.2112 - val_mean_squared_error: 2.5557
Epoch 58/100
1050/1050 [==============================] - 192s - loss: 1.0141 - mean_absolute_error: 1.0141 - mean_squared_error: 1.8466 - val_loss: 1.1634 - val_mean_absolute_error: 1.1634 - val_mean_squared_error: 2.3602
Epoch 59/100
1050/1050 [==============================] - 193s - loss: 0.9907 - mean_absolute_error: 0.9907 - mean_squared_error: 1.6806 - val_loss: 1.2531 - val_mean_absolute_error: 1.2531 - val_mean_squared_error: 2.7095
Epoch 60/100
1050/1050 [==============================] - 194s - loss: 0.9980 - mean_absolute_error: 0.9980 - mean_squared_error: 1.7786 - val_loss: 1.1483 - val_mean_absolute_error: 1.1483 - val_mean_squared_error: 2.1379
Epoch 61/100
1050/1050 [==============================] - 194s - loss: 0.9703 - mean_absolute_error: 0.9703 - mean_squared_error: 1.6716 - val_loss: 1.1238 - val_mean_absolute_error: 1.1238 - val_mean_squared_error: 2.0950
Epoch 62/100
1050/1050 [==============================] - 195s - loss: 0.9618 - mean_absolute_error: 0.9618 - mean_squared_error: 1.6845 - val_loss: 1.1722 - val_mean_absolute_error: 1.1722 - val_mean_squared_error: 2.3699
Epoch 63/100
1050/1050 [==============================] - 194s - loss: 0.9534 - mean_absolute_error: 0.9534 - mean_squared_error: 1.6555 - val_loss: 1.1623 - val_mean_absolute_error: 1.1623 - val_mean_squared_error: 2.2428
Epoch 64/100
1050/1050 [==============================] - 195s - loss: 0.9553 - mean_absolute_error: 0.9553 - mean_squared_error: 1.6038 - val_loss: 1.1935 - val_mean_absolute_error: 1.1935 - val_mean_squared_error: 2.4924
Epoch 65/100
1050/1050 [==============================] - 195s - loss: 0.9979 - mean_absolute_error: 0.9979 - mean_squared_error: 1.7466 - val_loss: 1.1682 - val_mean_absolute_error: 1.1682 - val_mean_squared_error: 2.3832
Epoch 66/100
1050/1050 [==============================] - 195s - loss: 0.9643 - mean_absolute_error: 0.9643 - mean_squared_error: 1.6776 - val_loss: 1.1663 - val_mean_absolute_error: 1.1663 - val_mean_squared_error: 2.3274
Epoch 67/100
1050/1050 [==============================] - 195s - loss: 0.9715 - mean_absolute_error: 0.9715 - mean_squared_error: 1.6804 - val_loss: 1.1842 - val_mean_absolute_error: 1.1842 - val_mean_squared_error: 2.3672
Epoch 68/100
1050/1050 [==============================] - 194s - loss: 0.9622 - mean_absolute_error: 0.9622 - mean_squared_error: 1.6285 - val_loss: 1.1403 - val_mean_absolute_error: 1.1403 - val_mean_squared_error: 2.1159
Epoch 69/100
1050/1050 [==============================] - 194s - loss: 0.9618 - mean_absolute_error: 0.9618 - mean_squared_error: 1.6604 - val_loss: 1.1310 - val_mean_absolute_error: 1.1310 - val_mean_squared_error: 2.1934
Epoch 70/100
1050/1050 [==============================] - 195s - loss: 0.9145 - mean_absolute_error: 0.9145 - mean_squared_error: 1.5906 - val_loss: 1.2025 - val_mean_absolute_error: 1.2025 - val_mean_squared_error: 2.4112
Epoch 71/100
1050/1050 [==============================] - 194s - loss: 0.9516 - mean_absolute_error: 0.9516 - mean_squared_error: 1.7187 - val_loss: 1.1246 - val_mean_absolute_error: 1.1246 - val_mean_squared_error: 2.1474
Epoch 72/100
1050/1050 [==============================] - 197s - loss: 0.9121 - mean_absolute_error: 0.9121 - mean_squared_error: 1.5745 - val_loss: 1.2023 - val_mean_absolute_error: 1.2023 - val_mean_squared_error: 2.4598
Epoch 73/100
1050/1050 [==============================] - 194s - loss: 0.9337 - mean_absolute_error: 0.9337 - mean_squared_error: 1.6140 - val_loss: 1.1292 - val_mean_absolute_error: 1.1292 - val_mean_squared_error: 2.1803
Epoch 74/100
1050/1050 [==============================] - 194s - loss: 0.9239 - mean_absolute_error: 0.9239 - mean_squared_error: 1.5726 - val_loss: 1.1369 - val_mean_absolute_error: 1.1369 - val_mean_squared_error: 2.1370
Epoch 75/100
1050/1050 [==============================] - 194s - loss: 0.9338 - mean_absolute_error: 0.9338 - mean_squared_error: 1.5977 - val_loss: 1.1425 - val_mean_absolute_error: 1.1425 - val_mean_squared_error: 2.1418
Epoch 76/100
1050/1050 [==============================] - 195s - loss: 0.9289 - mean_absolute_error: 0.9289 - mean_squared_error: 1.5910 - val_loss: 1.1339 - val_mean_absolute_error: 1.1339 - val_mean_squared_error: 2.1815
Epoch 77/100
1050/1050 [==============================] - 196s - loss: 0.9092 - mean_absolute_error: 0.9092 - mean_squared_error: 1.5318 - val_loss: 1.1810 - val_mean_absolute_error: 1.1810 - val_mean_squared_error: 2.4074
Epoch 78/100
1050/1050 [==============================] - 194s - loss: 0.9073 - mean_absolute_error: 0.9073 - mean_squared_error: 1.5186 - val_loss: 1.2393 - val_mean_absolute_error: 1.2393 - val_mean_squared_error: 2.5613
Epoch 79/100
1050/1050 [==============================] - 194s - loss: 0.9101 - mean_absolute_error: 0.9101 - mean_squared_error: 1.5726 - val_loss: 1.1647 - val_mean_absolute_error: 1.1647 - val_mean_squared_error: 2.3648
Epoch 80/100
1050/1050 [==============================] - 197s - loss: 0.9203 - mean_absolute_error: 0.9203 - mean_squared_error: 1.5967 - val_loss: 1.1399 - val_mean_absolute_error: 1.1399 - val_mean_squared_error: 2.2131
Epoch 81/100
1050/1050 [==============================] - 194s - loss: 0.9248 - mean_absolute_error: 0.9248 - mean_squared_error: 1.5473 - val_loss: 1.2048 - val_mean_absolute_error: 1.2048 - val_mean_squared_error: 2.4625
Epoch 82/100
1050/1050 [==============================] - 195s - loss: 0.9031 - mean_absolute_error: 0.9031 - mean_squared_error: 1.5464 - val_loss: 1.1701 - val_mean_absolute_error: 1.1701 - val_mean_squared_error: 2.4762
Epoch 83/100
1050/1050 [==============================] - 193s - loss: 0.9058 - mean_absolute_error: 0.9058 - mean_squared_error: 1.5389 - val_loss: 1.1684 - val_mean_absolute_error: 1.1684 - val_mean_squared_error: 2.2155
Epoch 84/100
1050/1050 [==============================] - 194s - loss: 0.9156 - mean_absolute_error: 0.9156 - mean_squared_error: 1.5462 - val_loss: 1.1750 - val_mean_absolute_error: 1.1750 - val_mean_squared_error: 2.2269
Epoch 85/100
1050/1050 [==============================] - 196s - loss: 0.8895 - mean_absolute_error: 0.8895 - mean_squared_error: 1.4947 - val_loss: 1.1803 - val_mean_absolute_error: 1.1803 - val_mean_squared_error: 2.3331
Epoch 86/100
1050/1050 [==============================] - 196s - loss: 0.9287 - mean_absolute_error: 0.9287 - mean_squared_error: 1.5775 - val_loss: 1.1858 - val_mean_absolute_error: 1.1858 - val_mean_squared_error: 2.4515
Epoch 87/100
1050/1050 [==============================] - 196s - loss: 0.8972 - mean_absolute_error: 0.8972 - mean_squared_error: 1.5451 - val_loss: 1.2210 - val_mean_absolute_error: 1.2210 - val_mean_squared_error: 2.6368
Epoch 88/100
1050/1050 [==============================] - 196s - loss: 0.8691 - mean_absolute_error: 0.8691 - mean_squared_error: 1.4523 - val_loss: 1.2637 - val_mean_absolute_error: 1.2637 - val_mean_squared_error: 2.6947
Epoch 89/100
1050/1050 [==============================] - 194s - loss: 0.9235 - mean_absolute_error: 0.9235 - mean_squared_error: 1.6097 - val_loss: 1.1868 - val_mean_absolute_error: 1.1868 - val_mean_squared_error: 2.4712
Epoch 90/100
1050/1050 [==============================] - 196s - loss: 0.8931 - mean_absolute_error: 0.8931 - mean_squared_error: 1.5160 - val_loss: 1.1359 - val_mean_absolute_error: 1.1359 - val_mean_squared_error: 2.1496
Epoch 91/100
1050/1050 [==============================] - 196s - loss: 0.8995 - mean_absolute_error: 0.8995 - mean_squared_error: 1.5244 - val_loss: 1.1569 - val_mean_absolute_error: 1.1569 - val_mean_squared_error: 2.1702
Epoch 92/100
1050/1050 [==============================] - 197s - loss: 0.8394 - mean_absolute_error: 0.8394 - mean_squared_error: 1.3711 - val_loss: 1.1432 - val_mean_absolute_error: 1.1432 - val_mean_squared_error: 2.2423
Epoch 93/100
1050/1050 [==============================] - 195s - loss: 0.8636 - mean_absolute_error: 0.8636 - mean_squared_error: 1.4125 - val_loss: 1.2076 - val_mean_absolute_error: 1.2076 - val_mean_squared_error: 2.4945
Epoch 94/100
1050/1050 [==============================] - 196s - loss: 0.8906 - mean_absolute_error: 0.8906 - mean_squared_error: 1.5270 - val_loss: 1.1468 - val_mean_absolute_error: 1.1468 - val_mean_squared_error: 2.2323
Epoch 95/100
1050/1050 [==============================] - 196s - loss: 0.8772 - mean_absolute_error: 0.8772 - mean_squared_error: 1.4942 - val_loss: 1.2336 - val_mean_absolute_error: 1.2336 - val_mean_squared_error: 2.5533
Epoch 96/100
1050/1050 [==============================] - 195s - loss: 0.8722 - mean_absolute_error: 0.8722 - mean_squared_error: 1.4535 - val_loss: 1.1982 - val_mean_absolute_error: 1.1982 - val_mean_squared_error: 2.2977
Epoch 97/100
1050/1050 [==============================] - 196s - loss: 0.8657 - mean_absolute_error: 0.8657 - mean_squared_error: 1.4628 - val_loss: 1.1698 - val_mean_absolute_error: 1.1698 - val_mean_squared_error: 2.2128
Epoch 98/100
1050/1050 [==============================] - 195s - loss: 0.8921 - mean_absolute_error: 0.8921 - mean_squared_error: 1.4980 - val_loss: 1.1453 - val_mean_absolute_error: 1.1453 - val_mean_squared_error: 2.2404
Epoch 99/100
1050/1050 [==============================] - 194s - loss: 0.8470 - mean_absolute_error: 0.8470 - mean_squared_error: 1.3477 - val_loss: 1.2506 - val_mean_absolute_error: 1.2506 - val_mean_squared_error: 2.6056
Epoch 100/100
1050/1050 [==============================] - 195s - loss: 0.8637 - mean_absolute_error: 0.8637 - mean_squared_error: 1.4454 - val_loss: 1.2089 - val_mean_absolute_error: 1.2089 - val_mean_squared_error: 2.4692
Parameters:
---------------------
h_STATES=25
EPOCHS=100
DENSES=10
Representation=fastText
EMBEDDING_DIM=50
MAX_SEQUENCE_LENGTH=50
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_4 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_1 (Activation) (None, 40) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 100) 424400
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_8 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_2 (Activation) (None, 40) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 10) 3240 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 44 maxoutdense_1[0][0]
====================================================================================================
Total params: 907,304
Trainable params: 111,004
Non-trainable params: 796,300
____________________________________________________________________________________________________
None
Happy learning!!!
Train on 1050 samples, validate on 450 samples
Epoch 1/20
1040/1050 [============================>.] - ETA: 23s - loss: 1.5288 - mean_absolute_error: 1.5288 - mean_squared_error: 3.2701Epoch 00000: val_loss improved from inf to 1.35953, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2987s - loss: 1.5275 - mean_absolute_error: 1.5275 - mean_squared_error: 3.2651 - val_loss: 1.3595 - val_mean_absolute_error: 1.3595 - val_mean_squared_error: 2.3046
Epoch 2/20
1040/1050 [============================>.] - ETA: 25s - loss: 1.4085 - mean_absolute_error: 1.4085 - mean_squared_error: 2.5278Epoch 00001: val_loss improved from 1.35953 to 1.33055, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3125s - loss: 1.4059 - mean_absolute_error: 1.4059 - mean_squared_error: 2.5223 - val_loss: 1.3306 - val_mean_absolute_error: 1.3306 - val_mean_squared_error: 2.3116
Epoch 3/20
1040/1050 [============================>.] - ETA: 22s - loss: 1.3798 - mean_absolute_error: 1.3798 - mean_squared_error: 2.4443Epoch 00002: val_loss improved from 1.33055 to 1.22478, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2671s - loss: 1.3797 - mean_absolute_error: 1.3797 - mean_squared_error: 2.4484 - val_loss: 1.2248 - val_mean_absolute_error: 1.2248 - val_mean_squared_error: 2.0110
Epoch 4/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.2881 - mean_absolute_error: 1.2881 - mean_squared_error: 2.3816Epoch 00003: val_loss improved from 1.22478 to 1.21644, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2488s - loss: 1.2871 - mean_absolute_error: 1.2871 - mean_squared_error: 2.3783 - val_loss: 1.2164 - val_mean_absolute_error: 1.2164 - val_mean_squared_error: 1.9845
Epoch 5/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.2618 - mean_absolute_error: 1.2618 - mean_squared_error: 2.2978Epoch 00004: val_loss did not improve
1050/1050 [==============================] - 2363s - loss: 1.2653 - mean_absolute_error: 1.2653 - mean_squared_error: 2.3057 - val_loss: 1.2259 - val_mean_absolute_error: 1.2259 - val_mean_squared_error: 2.1600
Epoch 6/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.2490 - mean_absolute_error: 1.2490 - mean_squared_error: 2.2563Epoch 00005: val_loss did not improve
1050/1050 [==============================] - 2464s - loss: 1.2461 - mean_absolute_error: 1.2461 - mean_squared_error: 2.2459 - val_loss: 1.2439 - val_mean_absolute_error: 1.2439 - val_mean_squared_error: 2.3019
Epoch 7/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.2729 - mean_absolute_error: 1.2729 - mean_squared_error: 2.3058Epoch 00006: val_loss improved from 1.21644 to 1.20320, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2532s - loss: 1.2746 - mean_absolute_error: 1.2746 - mean_squared_error: 2.3103 - val_loss: 1.2032 - val_mean_absolute_error: 1.2032 - val_mean_squared_error: 2.0193
Epoch 8/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.2202 - mean_absolute_error: 1.2202 - mean_squared_error: 2.1743Epoch 00007: val_loss improved from 1.20320 to 1.19266, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2471s - loss: 1.2219 - mean_absolute_error: 1.2219 - mean_squared_error: 2.1796 - val_loss: 1.1927 - val_mean_absolute_error: 1.1927 - val_mean_squared_error: 1.9725
Epoch 9/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.2086 - mean_absolute_error: 1.2086 - mean_squared_error: 2.1562Epoch 00008: val_loss did not improve
1050/1050 [==============================] - 2222s - loss: 1.2084 - mean_absolute_error: 1.2084 - mean_squared_error: 2.1616 - val_loss: 1.2134 - val_mean_absolute_error: 1.2134 - val_mean_squared_error: 1.9882
Epoch 10/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1903 - mean_absolute_error: 1.1903 - mean_squared_error: 2.1339Epoch 00009: val_loss did not improve
1050/1050 [==============================] - 2155s - loss: 1.1937 - mean_absolute_error: 1.1937 - mean_squared_error: 2.1434 - val_loss: 1.1948 - val_mean_absolute_error: 1.1948 - val_mean_squared_error: 2.1434
Epoch 11/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1491 - mean_absolute_error: 1.1491 - mean_squared_error: 2.0395Epoch 00010: val_loss did not improve
1050/1050 [==============================] - 2049s - loss: 1.1560 - mean_absolute_error: 1.1560 - mean_squared_error: 2.0672 - val_loss: 1.2055 - val_mean_absolute_error: 1.2055 - val_mean_squared_error: 2.0887
Epoch 12/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1738 - mean_absolute_error: 1.1738 - mean_squared_error: 2.1484Epoch 00011: val_loss did not improve
1050/1050 [==============================] - 2079s - loss: 1.1758 - mean_absolute_error: 1.1758 - mean_squared_error: 2.1567 - val_loss: 1.3005 - val_mean_absolute_error: 1.3005 - val_mean_squared_error: 2.4961
Epoch 13/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1505 - mean_absolute_error: 1.1505 - mean_squared_error: 2.0505Epoch 00012: val_loss did not improve
1050/1050 [==============================] - 2023s - loss: 1.1459 - mean_absolute_error: 1.1459 - mean_squared_error: 2.0365 - val_loss: 1.2259 - val_mean_absolute_error: 1.2259 - val_mean_squared_error: 2.4653
Epoch 14/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1320 - mean_absolute_error: 1.1320 - mean_squared_error: 2.0358Epoch 00013: val_loss improved from 1.19266 to 1.18684, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 1991s - loss: 1.1345 - mean_absolute_error: 1.1345 - mean_squared_error: 2.0419 - val_loss: 1.1868 - val_mean_absolute_error: 1.1868 - val_mean_squared_error: 2.2083
Epoch 15/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0990 - mean_absolute_error: 1.0990 - mean_squared_error: 1.9721Epoch 00014: val_loss did not improve
1050/1050 [==============================] - 2026s - loss: 1.0978 - mean_absolute_error: 1.0978 - mean_squared_error: 1.9698 - val_loss: 1.2235 - val_mean_absolute_error: 1.2235 - val_mean_squared_error: 2.1625
Epoch 16/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1177 - mean_absolute_error: 1.1177 - mean_squared_error: 1.9532Epoch 00015: val_loss improved from 1.18684 to 1.17322, saving model to /almac/ignacio/40_20_10_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2013s - loss: 1.1187 - mean_absolute_error: 1.1187 - mean_squared_error: 1.9609 - val_loss: 1.1732 - val_mean_absolute_error: 1.1732 - val_mean_squared_error: 2.1218
Epoch 17/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0938 - mean_absolute_error: 1.0938 - mean_squared_error: 1.9491Epoch 00016: val_loss did not improve
1050/1050 [==============================] - 2077s - loss: 1.0916 - mean_absolute_error: 1.0916 - mean_squared_error: 1.9473 - val_loss: 1.2452 - val_mean_absolute_error: 1.2452 - val_mean_squared_error: 2.3074
Epoch 18/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.0857 - mean_absolute_error: 1.0857 - mean_squared_error: 1.9023Epoch 00017: val_loss did not improve
1050/1050 [==============================] - 2224s - loss: 1.0868 - mean_absolute_error: 1.0868 - mean_squared_error: 1.9006 - val_loss: 1.1815 - val_mean_absolute_error: 1.1815 - val_mean_squared_error: 2.1743
Epoch 19/20
1040/1050 [============================>.] - ETA: 21s - loss: 1.0579 - mean_absolute_error: 1.0579 - mean_squared_error: 1.8432Epoch 00018: val_loss did not improve
1050/1050 [==============================] - 2672s - loss: 1.0558 - mean_absolute_error: 1.0558 - mean_squared_error: 1.8357 - val_loss: 1.1844 - val_mean_absolute_error: 1.1844 - val_mean_squared_error: 2.2164
Epoch 20/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.0483 - mean_absolute_error: 1.0483 - mean_squared_error: 1.8225Epoch 00019: val_loss did not improve
1050/1050 [==============================] - 2497s - loss: 1.0459 - mean_absolute_error: 1.0459 - mean_squared_error: 1.8136 - val_loss: 1.3107 - val_mean_absolute_error: 1.3107 - val_mean_squared_error: 2.9719
Parameters:
---------------------
h_STATES=40
EPOCHS=20
DENSES=10
Representation=fastText
EMBEDDING_DIM=100
MAX_SEQUENCE_LENGTH=50
MODEL_TYPE=stacked
Using Theano backend.
ETA: 73590s 17left 4328.89avg loLoanding train and valid dirs......
Starting training
Phrases in STS.input.headlines.txt 750 750
Phrases in STS.input.OnWN.txt 561 561
Phrases in STS.input.FNWN.txt 189 189
Total train phrases /almac/ignacio/data/sts_all/train-2013 1500
Total train phrases 3000
Starting training
Phrases in STS.input.track5.en-en.txt 250
Total train phrases /almac/ignacio/data/sts_all/valid-2017 250
Total train phrases 500
Spliting tab-separated files...
Labels shape: (1500,)
Tokenizing files... [A]
Split training set into train and val... [A]
Tokenizing files... [B]
Split training set into train and val... [B]
Getting embedding matrix... from /almac/ignacio/data/fastText/wikiEn_Full_H100.model.vec
Found 4049540 word vectors.
Filling embedding matrices...
Compiling the model...
____________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
====================================================================================================
embedding_1 (Embedding) (None, 50, 100) 371900
____________________________________________________________________________________________________
attention_1 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_4 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_1 (Activation) (None, 40) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 100) 424400
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_8 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_2 (Activation) (None, 40) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 30) 9720 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 124 maxoutdense_1[0][0]
====================================================================================================
Total params: 913,864
Trainable params: 117,564
Non-trainable params: 796,300
____________________________________________________________________________________________________
None
Happy learning!!!
Train on 1050 samples, validate on 450 samples
Epoch 1/20
1040/1050 [============================>.] - ETA: 26s - loss: 1.4276 - mean_absolute_error: 1.4276 - mean_squared_error: 2.7220Epoch 00000: val_loss improved from inf to 1.38671, saving model to /almac/ignacio/40_20_30_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3204s - loss: 1.4303 - mean_absolute_error: 1.4303 - mean_squared_error: 2.7386 - val_loss: 1.3867 - val_mean_absolute_error: 1.3867 - val_mean_squared_error: 2.7299
Epoch 2/20
1040/1050 [============================>.] - ETA: 24s - loss: 1.2859 - mean_absolute_error: 1.2859 - mean_squared_error: 2.3111Epoch 00001: val_loss improved from 1.38671 to 1.27424, saving model to /almac/ignacio/40_20_30_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3009s - loss: 1.2858 - mean_absolute_error: 1.2858 - mean_squared_error: 2.3123 - val_loss: 1.2742 - val_mean_absolute_error: 1.2742 - val_mean_squared_error: 2.2106
Epoch 3/20
1040/1050 [============================>.] - ETA: 22s - loss: 1.2754 - mean_absolute_error: 1.2754 - mean_squared_error: 2.3545Epoch 00002: val_loss did not improve
1050/1050 [==============================] - 2831s - loss: 1.2767 - mean_absolute_error: 1.2767 - mean_squared_error: 2.3548 - val_loss: 1.3169 - val_mean_absolute_error: 1.3169 - val_mean_squared_error: 2.3129
Epoch 4/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.1984 - mean_absolute_error: 1.1984 - mean_squared_error: 2.0914Epoch 00003: val_loss did not improve
1050/1050 [==============================] - 2595s - loss: 1.1945 - mean_absolute_error: 1.1945 - mean_squared_error: 2.0886 - val_loss: 1.4391 - val_mean_absolute_error: 1.4391 - val_mean_squared_error: 3.2157
Epoch 5/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.2092 - mean_absolute_error: 1.2092 - mean_squared_error: 2.1817Epoch 00004: val_loss did not improve
1050/1050 [==============================] - 2432s - loss: 1.2116 - mean_absolute_error: 1.2116 - mean_squared_error: 2.1856 - val_loss: 1.4263 - val_mean_absolute_error: 1.4263 - val_mean_squared_error: 3.1264
Epoch 6/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.2313 - mean_absolute_error: 1.2313 - mean_squared_error: 2.2796Epoch 00005: val_loss did not improve
1050/1050 [==============================] - 2509s - loss: 1.2278 - mean_absolute_error: 1.2278 - mean_squared_error: 2.2688 - val_loss: 1.3084 - val_mean_absolute_error: 1.3084 - val_mean_squared_error: 2.6580
Epoch 7/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1662 - mean_absolute_error: 1.1662 - mean_squared_error: 2.0922Epoch 00006: val_loss did not improve
1050/1050 [==============================] - 2281s - loss: 1.1646 - mean_absolute_error: 1.1646 - mean_squared_error: 2.0841 - val_loss: 1.3217 - val_mean_absolute_error: 1.3217 - val_mean_squared_error: 2.6216
Epoch 8/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1944 - mean_absolute_error: 1.1944 - mean_squared_error: 2.1583Epoch 00007: val_loss did not improve
1050/1050 [==============================] - 2284s - loss: 1.1922 - mean_absolute_error: 1.1922 - mean_squared_error: 2.1512 - val_loss: 1.2863 - val_mean_absolute_error: 1.2863 - val_mean_squared_error: 2.3409
Epoch 9/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1674 - mean_absolute_error: 1.1674 - mean_squared_error: 2.0438Epoch 00008: val_loss did not improve
1050/1050 [==============================] - 2287s - loss: 1.1691 - mean_absolute_error: 1.1691 - mean_squared_error: 2.0530 - val_loss: 1.2765 - val_mean_absolute_error: 1.2765 - val_mean_squared_error: 2.5193
Epoch 10/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1176 - mean_absolute_error: 1.1176 - mean_squared_error: 1.9586Epoch 00009: val_loss did not improve
1050/1050 [==============================] - 2129s - loss: 1.1168 - mean_absolute_error: 1.1168 - mean_squared_error: 1.9558 - val_loss: 1.2851 - val_mean_absolute_error: 1.2851 - val_mean_squared_error: 2.6159
Epoch 11/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0912 - mean_absolute_error: 1.0912 - mean_squared_error: 1.9345Epoch 00010: val_loss did not improve
1050/1050 [==============================] - 2050s - loss: 1.0857 - mean_absolute_error: 1.0857 - mean_squared_error: 1.9199 - val_loss: 1.5346 - val_mean_absolute_error: 1.5346 - val_mean_squared_error: 3.6908
Epoch 12/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.0859 - mean_absolute_error: 1.0859 - mean_squared_error: 1.9221Epoch 00011: val_loss improved from 1.27424 to 1.27093, saving model to /almac/ignacio/40_20_30_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2360s - loss: 1.0843 - mean_absolute_error: 1.0843 - mean_squared_error: 1.9178 - val_loss: 1.2709 - val_mean_absolute_error: 1.2709 - val_mean_squared_error: 2.5844
Epoch 13/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.0478 - mean_absolute_error: 1.0478 - mean_squared_error: 1.8461Epoch 00012: val_loss did not improve
1050/1050 [==============================] - 2107s - loss: 1.0453 - mean_absolute_error: 1.0453 - mean_squared_error: 1.8405 - val_loss: 1.4388 - val_mean_absolute_error: 1.4388 - val_mean_squared_error: 3.4505
Epoch 14/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0335 - mean_absolute_error: 1.0335 - mean_squared_error: 1.7734Epoch 00013: val_loss did not improve
1050/1050 [==============================] - 2038s - loss: 1.0332 - mean_absolute_error: 1.0332 - mean_squared_error: 1.7755 - val_loss: 1.3024 - val_mean_absolute_error: 1.3024 - val_mean_squared_error: 2.6925
Epoch 15/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0689 - mean_absolute_error: 1.0689 - mean_squared_error: 1.8783Epoch 00014: val_loss did not improve
1050/1050 [==============================] - 2035s - loss: 1.0681 - mean_absolute_error: 1.0681 - mean_squared_error: 1.8721 - val_loss: 1.2910 - val_mean_absolute_error: 1.2910 - val_mean_squared_error: 2.5669
Epoch 16/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0550 - mean_absolute_error: 1.0550 - mean_squared_error: 1.9198Epoch 00015: val_loss improved from 1.27093 to 1.22732, saving model to /almac/ignacio/40_20_30_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2067s - loss: 1.0590 - mean_absolute_error: 1.0590 - mean_squared_error: 1.9265 - val_loss: 1.2273 - val_mean_absolute_error: 1.2273 - val_mean_squared_error: 2.2766
Epoch 17/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0142 - mean_absolute_error: 1.0142 - mean_squared_error: 1.7247Epoch 00016: val_loss did not improve
1050/1050 [==============================] - 2042s - loss: 1.0134 - mean_absolute_error: 1.0134 - mean_squared_error: 1.7204 - val_loss: 1.2624 - val_mean_absolute_error: 1.2624 - val_mean_squared_error: 2.3938
Epoch 18/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.0224 - mean_absolute_error: 1.0224 - mean_squared_error: 1.7349Epoch 00017: val_loss did not improve
1050/1050 [==============================] - 2569s - loss: 1.0220 - mean_absolute_error: 1.0220 - mean_squared_error: 1.7337 - val_loss: 1.2640 - val_mean_absolute_error: 1.2640 - val_mean_squared_error: 2.3578
Epoch 19/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.0188 - mean_absolute_error: 1.0188 - mean_squared_error: 1.7546Epoch 00018: val_loss did not improve
1050/1050 [==============================] - 2394s - loss: 1.0180 - mean_absolute_error: 1.0180 - mean_squared_error: 1.7514 - val_loss: 1.3331 - val_mean_absolute_error: 1.3331 - val_mean_squared_error: 2.9346
Epoch 20/20
1040/1050 [============================>.] - ETA: 18s - loss: 0.9897 - mean_absolute_error: 0.9897 - mean_squared_error: 1.6941Epoch 00019: val_loss did not improve
1050/1050 [==============================] - 2328s - loss: 0.9866 - mean_absolute_error: 0.9866 - mean_squared_error: 1.6843 - val_loss: 1.2301 - val_mean_absolute_error: 1.2301 - val_mean_squared_error: 2.3584
Parameters:
---------------------
h_STATES=40
EPOCHS=20
DENSES=30
Representation=fastText
EMBEDDING_DIM=100
MAX_SEQUENCE_LENGTH=50
MODEL_TYPE=stacked
Using Theano backend.
ETA: 66033s 16left 4127.15avg local:13/Loanding train and valid dirs......
Starting training
Phrases in STS.input.headlines.txt 750 750
Phrases in STS.input.OnWN.txt 561 561
Phrases in STS.input.FNWN.txt 189 189
Total train phrases /almac/ignacio/data/sts_all/train-2013 1500
Total train phrases 3000
Starting training
Phrases in STS.input.track5.en-en.txt 250
Total train phrases /almac/ignacio/data/sts_all/valid-2017 250
Total train phrases 500
Spliting tab-separated files...
Labels shape: (1500,)
Tokenizing files... [A]
Split training set into train and val... [A]
Tokenizing files... [B]
Split training set into train and val... [B]
Getting embedding matrix... from /almac/ignacio/data/fastText/wikiEn_Full_H100.model.vec
Found 4049540 word vectors.
Filling embedding matrices...
Compiling the model...
____________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
====================================================================================================
embedding_1 (Embedding) (None, 50, 100) 371900
____________________________________________________________________________________________________
attention_1 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_4 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_1 (Activation) (None, 40) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 100) 424400
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_8 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_2 (Activation) (None, 40) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 20) 6480 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 84 maxoutdense_1[0][0]
====================================================================================================
Total params: 910,584
Trainable params: 114,284
Non-trainable params: 796,300
____________________________________________________________________________________________________
None
Happy learning!!!
Train on 1050 samples, validate on 450 samples
Epoch 1/20
1040/1050 [============================>.] - ETA: 24s - loss: 1.4141 - mean_absolute_error: 1.4141 - mean_squared_error: 2.6896Epoch 00000: val_loss improved from inf to 1.41768, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3102s - loss: 1.4116 - mean_absolute_error: 1.4116 - mean_squared_error: 2.6809 - val_loss: 1.4177 - val_mean_absolute_error: 1.4177 - val_mean_squared_error: 2.7852
Epoch 2/20
1040/1050 [============================>.] - ETA: 25s - loss: 1.3874 - mean_absolute_error: 1.3874 - mean_squared_error: 2.5382Epoch 00001: val_loss improved from 1.41768 to 1.36109, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3132s - loss: 1.3885 - mean_absolute_error: 1.3885 - mean_squared_error: 2.5427 - val_loss: 1.3611 - val_mean_absolute_error: 1.3611 - val_mean_squared_error: 2.3624
Epoch 3/20
1040/1050 [============================>.] - ETA: 22s - loss: 1.3529 - mean_absolute_error: 1.3529 - mean_squared_error: 2.4574Epoch 00002: val_loss improved from 1.36109 to 1.29562, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2715s - loss: 1.3547 - mean_absolute_error: 1.3547 - mean_squared_error: 2.4632 - val_loss: 1.2956 - val_mean_absolute_error: 1.2956 - val_mean_squared_error: 2.3608
Epoch 4/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.3236 - mean_absolute_error: 1.3236 - mean_squared_error: 2.4377Epoch 00003: val_loss did not improve
1050/1050 [==============================] - 2480s - loss: 1.3269 - mean_absolute_error: 1.3269 - mean_squared_error: 2.4490 - val_loss: 1.5780 - val_mean_absolute_error: 1.5780 - val_mean_squared_error: 3.6315
Epoch 5/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.2832 - mean_absolute_error: 1.2832 - mean_squared_error: 2.3384Epoch 00004: val_loss improved from 1.29562 to 1.25044, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2354s - loss: 1.2818 - mean_absolute_error: 1.2818 - mean_squared_error: 2.3343 - val_loss: 1.2504 - val_mean_absolute_error: 1.2504 - val_mean_squared_error: 2.1539
Epoch 6/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.2591 - mean_absolute_error: 1.2591 - mean_squared_error: 2.2560Epoch 00005: val_loss improved from 1.25044 to 1.24010, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2115s - loss: 1.2595 - mean_absolute_error: 1.2595 - mean_squared_error: 2.2605 - val_loss: 1.2401 - val_mean_absolute_error: 1.2401 - val_mean_squared_error: 2.3934
Epoch 7/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.2868 - mean_absolute_error: 1.2868 - mean_squared_error: 2.3243Epoch 00006: val_loss did not improve
1050/1050 [==============================] - 2396s - loss: 1.2888 - mean_absolute_error: 1.2888 - mean_squared_error: 2.3277 - val_loss: 1.2693 - val_mean_absolute_error: 1.2693 - val_mean_squared_error: 2.1544
Epoch 8/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.2501 - mean_absolute_error: 1.2501 - mean_squared_error: 2.3057Epoch 00007: val_loss did not improve
1050/1050 [==============================] - 2295s - loss: 1.2497 - mean_absolute_error: 1.2497 - mean_squared_error: 2.3042 - val_loss: 1.3217 - val_mean_absolute_error: 1.3217 - val_mean_squared_error: 2.3454
Epoch 9/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.2622 - mean_absolute_error: 1.2622 - mean_squared_error: 2.2621Epoch 00008: val_loss did not improve
1050/1050 [==============================] - 2020s - loss: 1.2612 - mean_absolute_error: 1.2612 - mean_squared_error: 2.2663 - val_loss: 1.2607 - val_mean_absolute_error: 1.2607 - val_mean_squared_error: 2.1491
Epoch 10/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2552 - mean_absolute_error: 1.2552 - mean_squared_error: 2.3137Epoch 00009: val_loss did not improve
1050/1050 [==============================] - 2144s - loss: 1.2567 - mean_absolute_error: 1.2567 - mean_squared_error: 2.3149 - val_loss: 1.2424 - val_mean_absolute_error: 1.2424 - val_mean_squared_error: 2.0541
Epoch 11/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.2014 - mean_absolute_error: 1.2014 - mean_squared_error: 2.0808Epoch 00010: val_loss did not improve
1050/1050 [==============================] - 2108s - loss: 1.2020 - mean_absolute_error: 1.2020 - mean_squared_error: 2.0840 - val_loss: 1.2407 - val_mean_absolute_error: 1.2407 - val_mean_squared_error: 2.6099
Epoch 12/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.2005 - mean_absolute_error: 1.2005 - mean_squared_error: 2.1703Epoch 00011: val_loss improved from 1.24010 to 1.20269, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2133s - loss: 1.2012 - mean_absolute_error: 1.2012 - mean_squared_error: 2.1740 - val_loss: 1.2027 - val_mean_absolute_error: 1.2027 - val_mean_squared_error: 2.3068
Epoch 13/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.1769 - mean_absolute_error: 1.1769 - mean_squared_error: 2.1038Epoch 00012: val_loss improved from 1.20269 to 1.17699, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2359s - loss: 1.1755 - mean_absolute_error: 1.1755 - mean_squared_error: 2.0987 - val_loss: 1.1770 - val_mean_absolute_error: 1.1770 - val_mean_squared_error: 2.1604
Epoch 14/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1482 - mean_absolute_error: 1.1482 - mean_squared_error: 2.0041Epoch 00013: val_loss did not improve
1050/1050 [==============================] - 2334s - loss: 1.1474 - mean_absolute_error: 1.1474 - mean_squared_error: 2.0023 - val_loss: 1.1950 - val_mean_absolute_error: 1.1950 - val_mean_squared_error: 2.1474
Epoch 15/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1402 - mean_absolute_error: 1.1402 - mean_squared_error: 2.0304Epoch 00014: val_loss did not improve
1050/1050 [==============================] - 2253s - loss: 1.1377 - mean_absolute_error: 1.1377 - mean_squared_error: 2.0221 - val_loss: 1.3151 - val_mean_absolute_error: 1.3151 - val_mean_squared_error: 2.4105
Epoch 16/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1541 - mean_absolute_error: 1.1541 - mean_squared_error: 2.0336Epoch 00015: val_loss improved from 1.17699 to 1.15476, saving model to /almac/ignacio/40_20_20_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2200s - loss: 1.1546 - mean_absolute_error: 1.1546 - mean_squared_error: 2.0341 - val_loss: 1.1548 - val_mean_absolute_error: 1.1548 - val_mean_squared_error: 1.9597
Epoch 17/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1025 - mean_absolute_error: 1.1025 - mean_squared_error: 1.9134Epoch 00016: val_loss did not improve
1050/1050 [==============================] - 2138s - loss: 1.1063 - mean_absolute_error: 1.1063 - mean_squared_error: 1.9235 - val_loss: 1.1616 - val_mean_absolute_error: 1.1616 - val_mean_squared_error: 2.0312
Epoch 18/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.0898 - mean_absolute_error: 1.0898 - mean_squared_error: 1.8759Epoch 00017: val_loss did not improve
1050/1050 [==============================] - 2479s - loss: 1.0887 - mean_absolute_error: 1.0887 - mean_squared_error: 1.8703 - val_loss: 1.2052 - val_mean_absolute_error: 1.2052 - val_mean_squared_error: 2.1450
Epoch 19/20
1040/1050 [============================>.] - ETA: 22s - loss: 1.0811 - mean_absolute_error: 1.0811 - mean_squared_error: 1.8774Epoch 00018: val_loss did not improve
1050/1050 [==============================] - 2742s - loss: 1.0762 - mean_absolute_error: 1.0762 - mean_squared_error: 1.8638 - val_loss: 1.1684 - val_mean_absolute_error: 1.1684 - val_mean_squared_error: 2.0857
Epoch 20/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.0685 - mean_absolute_error: 1.0685 - mean_squared_error: 1.8519Epoch 00019: val_loss did not improve
1050/1050 [==============================] - 2502s - loss: 1.0697 - mean_absolute_error: 1.0697 - mean_squared_error: 1.8562 - val_loss: 1.1940 - val_mean_absolute_error: 1.1940 - val_mean_squared_error: 2.2198
Parameters:
---------------------
h_STATES=40
EPOCHS=20
DENSES=20
Representation=fastText
EMBEDDING_DIM=100
MAX_SEQUENCE_LENGTH=50
MODEL_TYPE=stacked
Using Theano backend.
ETA: 66054s 16left 4128.45avg local:13/20/100%/4128.5s Loanding train and valid dirs......
Starting training
Phrases in STS.input.headlines.txt 750 750
Phrases in STS.input.OnWN.txt 561 561
Phrases in STS.input.FNWN.txt 189 189
Total train phrases /almac/ignacio/data/sts_all/train-2013 1500
Total train phrases 3000
Starting training
Phrases in STS.input.track5.en-en.txt 250
Total train phrases /almac/ignacio/data/sts_all/valid-2017 250
Total train phrases 500
Spliting tab-separated files...
Labels shape: (1500,)
Tokenizing files... [A]
Split training set into train and val... [A]
Tokenizing files... [B]
Split training set into train and val... [B]
Getting embedding matrix... from /almac/ignacio/data/fastText/wikiEn_Full_H100.model.vec
Found 4049540 word vectors.
Filling embedding matrices...
Compiling the model...
____________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
====================================================================================================
embedding_1 (Embedding) (None, 50, 100) 371900
____________________________________________________________________________________________________
attention_1 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_4 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_1 (Activation) (None, 40) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 100) 424400
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_8 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_2 (Activation) (None, 40) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 100) 32400 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 404 maxoutdense_1[0][0]
====================================================================================================
Total params: 936,824
Trainable params: 140,524
Non-trainable params: 796,300
____________________________________________________________________________________________________
None
Happy learning!!!
Train on 1050 samples, validate on 450 samples
Epoch 1/20
1040/1050 [============================>.] - ETA: 27s - loss: 1.4148 - mean_absolute_error: 1.4148 - mean_squared_error: 2.7513Epoch 00000: val_loss improved from inf to 1.53630, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3274s - loss: 1.4189 - mean_absolute_error: 1.4189 - mean_squared_error: 2.7617 - val_loss: 1.5363 - val_mean_absolute_error: 1.5363 - val_mean_squared_error: 3.2968
Epoch 2/20
1040/1050 [============================>.] - ETA: 25s - loss: 1.4070 - mean_absolute_error: 1.4070 - mean_squared_error: 2.6304Epoch 00001: val_loss did not improve
1050/1050 [==============================] - 3105s - loss: 1.4041 - mean_absolute_error: 1.4041 - mean_squared_error: 2.6208 - val_loss: 1.8275 - val_mean_absolute_error: 1.8275 - val_mean_squared_error: 5.3500
Epoch 3/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.3925 - mean_absolute_error: 1.3925 - mean_squared_error: 2.5772Epoch 00002: val_loss improved from 1.53630 to 1.39060, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2517s - loss: 1.3953 - mean_absolute_error: 1.3953 - mean_squared_error: 2.5891 - val_loss: 1.3906 - val_mean_absolute_error: 1.3906 - val_mean_squared_error: 2.5093
Epoch 4/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.4015 - mean_absolute_error: 1.4015 - mean_squared_error: 2.7405Epoch 00003: val_loss improved from 1.39060 to 1.23028, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2517s - loss: 1.4029 - mean_absolute_error: 1.4029 - mean_squared_error: 2.7406 - val_loss: 1.2303 - val_mean_absolute_error: 1.2303 - val_mean_squared_error: 2.0152
Epoch 5/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.3202 - mean_absolute_error: 1.3202 - mean_squared_error: 2.4278Epoch 00004: val_loss did not improve
1050/1050 [==============================] - 2116s - loss: 1.3226 - mean_absolute_error: 1.3226 - mean_squared_error: 2.4363 - val_loss: 1.6226 - val_mean_absolute_error: 1.6226 - val_mean_squared_error: 4.0435
Epoch 6/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.3128 - mean_absolute_error: 1.3128 - mean_squared_error: 2.4397Epoch 00005: val_loss improved from 1.23028 to 1.15330, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2115s - loss: 1.3185 - mean_absolute_error: 1.3185 - mean_squared_error: 2.4598 - val_loss: 1.1533 - val_mean_absolute_error: 1.1533 - val_mean_squared_error: 2.1238
Epoch 7/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.3244 - mean_absolute_error: 1.3244 - mean_squared_error: 2.5245Epoch 00006: val_loss did not improve
1050/1050 [==============================] - 2170s - loss: 1.3246 - mean_absolute_error: 1.3246 - mean_squared_error: 2.5228 - val_loss: 1.1703 - val_mean_absolute_error: 1.1703 - val_mean_squared_error: 1.9015
Epoch 8/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2327 - mean_absolute_error: 1.2327 - mean_squared_error: 2.2513Epoch 00007: val_loss did not improve
1050/1050 [==============================] - 2213s - loss: 1.2337 - mean_absolute_error: 1.2337 - mean_squared_error: 2.2536 - val_loss: 1.1789 - val_mean_absolute_error: 1.1789 - val_mean_squared_error: 1.9624
Epoch 9/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2162 - mean_absolute_error: 1.2162 - mean_squared_error: 2.2583Epoch 00008: val_loss did not improve
1050/1050 [==============================] - 2163s - loss: 1.2131 - mean_absolute_error: 1.2131 - mean_squared_error: 2.2464 - val_loss: 1.2595 - val_mean_absolute_error: 1.2595 - val_mean_squared_error: 2.7792
Epoch 10/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2011 - mean_absolute_error: 1.2011 - mean_squared_error: 2.2060Epoch 00009: val_loss improved from 1.15330 to 1.14014, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2204s - loss: 1.2013 - mean_absolute_error: 1.2013 - mean_squared_error: 2.2050 - val_loss: 1.1401 - val_mean_absolute_error: 1.1401 - val_mean_squared_error: 1.9820
Epoch 11/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1964 - mean_absolute_error: 1.1964 - mean_squared_error: 2.1860Epoch 00010: val_loss did not improve
1050/1050 [==============================] - 2116s - loss: 1.2000 - mean_absolute_error: 1.2000 - mean_squared_error: 2.1947 - val_loss: 1.4270 - val_mean_absolute_error: 1.4270 - val_mean_squared_error: 3.1032
Epoch 12/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.2214 - mean_absolute_error: 1.2214 - mean_squared_error: 2.2509Epoch 00011: val_loss did not improve
1050/1050 [==============================] - 2084s - loss: 1.2223 - mean_absolute_error: 1.2223 - mean_squared_error: 2.2486 - val_loss: 1.1837 - val_mean_absolute_error: 1.1837 - val_mean_squared_error: 2.0492
Epoch 13/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1900 - mean_absolute_error: 1.1900 - mean_squared_error: 2.1631Epoch 00012: val_loss improved from 1.14014 to 1.13876, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2129s - loss: 1.1879 - mean_absolute_error: 1.1879 - mean_squared_error: 2.1542 - val_loss: 1.1388 - val_mean_absolute_error: 1.1388 - val_mean_squared_error: 2.2025
Epoch 14/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1700 - mean_absolute_error: 1.1700 - mean_squared_error: 2.1489Epoch 00013: val_loss improved from 1.13876 to 1.12950, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2165s - loss: 1.1704 - mean_absolute_error: 1.1704 - mean_squared_error: 2.1507 - val_loss: 1.1295 - val_mean_absolute_error: 1.1295 - val_mean_squared_error: 2.0737
Epoch 15/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1820 - mean_absolute_error: 1.1820 - mean_squared_error: 2.1993Epoch 00014: val_loss did not improve
1050/1050 [==============================] - 2053s - loss: 1.1850 - mean_absolute_error: 1.1850 - mean_squared_error: 2.2114 - val_loss: 1.2283 - val_mean_absolute_error: 1.2283 - val_mean_squared_error: 2.6176
Epoch 16/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1804 - mean_absolute_error: 1.1804 - mean_squared_error: 2.1458Epoch 00015: val_loss did not improve
1050/1050 [==============================] - 2103s - loss: 1.1808 - mean_absolute_error: 1.1808 - mean_squared_error: 2.1454 - val_loss: 1.2293 - val_mean_absolute_error: 1.2293 - val_mean_squared_error: 2.0644
Epoch 17/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1243 - mean_absolute_error: 1.1243 - mean_squared_error: 1.9941Epoch 00016: val_loss improved from 1.12950 to 1.12493, saving model to /almac/ignacio/40_20_100_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2038s - loss: 1.1276 - mean_absolute_error: 1.1276 - mean_squared_error: 2.0036 - val_loss: 1.1249 - val_mean_absolute_error: 1.1249 - val_mean_squared_error: 1.9555
Epoch 18/20
1040/1050 [============================>.] - ETA: 21s - loss: 1.1186 - mean_absolute_error: 1.1186 - mean_squared_error: 1.9664Epoch 00017: val_loss did not improve
1050/1050 [==============================] - 2662s - loss: 1.1188 - mean_absolute_error: 1.1188 - mean_squared_error: 1.9689 - val_loss: 1.1460 - val_mean_absolute_error: 1.1460 - val_mean_squared_error: 1.9385
Epoch 19/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.1125 - mean_absolute_error: 1.1125 - mean_squared_error: 2.0135Epoch 00018: val_loss did not improve
1050/1050 [==============================] - 2498s - loss: 1.1147 - mean_absolute_error: 1.1147 - mean_squared_error: 2.0175 - val_loss: 1.1277 - val_mean_absolute_error: 1.1277 - val_mean_squared_error: 2.1140
Epoch 20/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.0980 - mean_absolute_error: 1.0980 - mean_squared_error: 1.9663Epoch 00019: val_loss did not improve
1050/1050 [==============================] - 2329s - loss: 1.0985 - mean_absolute_error: 1.0985 - mean_squared_error: 1.9642 - val_loss: 1.1545 - val_mean_absolute_error: 1.1545 - val_mean_squared_error: 1.9556
Parameters:
---------------------
h_STATES=40
EPOCHS=20
DENSES=100
Representation=fastText
EMBEDDING_DIM=100
MAX_SEQUENCE_LENGTH=50
MODEL_TYPE=stacked
Using Theano backend.
ETA: 52998s 14left 3785.64avg local:13/22/100%/3Loanding train and valid dirs......
Starting training
Phrases in STS.input.headlines.txt 750 750
Phrases in STS.input.OnWN.txt 561 561
Phrases in STS.input.FNWN.txt 189 189
Total train phrases /almac/ignacio/data/sts_all/train-2013 1500
Total train phrases 3000
Starting training
Phrases in STS.input.track5.en-en.txt 250
Total train phrases /almac/ignacio/data/sts_all/valid-2017 250
Total train phrases 500
Spliting tab-separated files...
Labels shape: (1500,)
Tokenizing files... [A]
Split training set into train and val... [A]
Tokenizing files... [B]
Split training set into train and val... [B]
Getting embedding matrix... from /almac/ignacio/data/fastText/wikiEn_Full_H100.model.vec
Found 4049540 word vectors.
Filling embedding matrices...
Compiling the model...
____________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
====================================================================================================
embedding_1 (Embedding) (None, 50, 100) 371900
____________________________________________________________________________________________________
attention_1 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_4 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_1 (Activation) (None, 40) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 100) 424400
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_8 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_2 (Activation) (None, 40) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 40) 12960 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 164 maxoutdense_1[0][0]
====================================================================================================
Total params: 917,144
Trainable params: 120,844
Non-trainable params: 796,300
____________________________________________________________________________________________________
None
Happy learning!!!
Train on 1050 samples, validate on 450 samples
Epoch 1/20
1040/1050 [============================>.] - ETA: 26s - loss: 1.4513 - mean_absolute_error: 1.4513 - mean_squared_error: 2.8215Epoch 00000: val_loss improved from inf to 1.36795, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3240s - loss: 1.4518 - mean_absolute_error: 1.4518 - mean_squared_error: 2.8196 - val_loss: 1.3680 - val_mean_absolute_error: 1.3680 - val_mean_squared_error: 2.3490
Epoch 2/20
1040/1050 [============================>.] - ETA: 25s - loss: 1.3994 - mean_absolute_error: 1.3994 - mean_squared_error: 2.6043Epoch 00001: val_loss did not improve
1050/1050 [==============================] - 3157s - loss: 1.3972 - mean_absolute_error: 1.3972 - mean_squared_error: 2.5987 - val_loss: 1.4922 - val_mean_absolute_error: 1.4922 - val_mean_squared_error: 3.2065
Epoch 3/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.3915 - mean_absolute_error: 1.3915 - mean_squared_error: 2.5874Epoch 00002: val_loss did not improve
1050/1050 [==============================] - 2427s - loss: 1.3946 - mean_absolute_error: 1.3946 - mean_squared_error: 2.5959 - val_loss: 1.3703 - val_mean_absolute_error: 1.3703 - val_mean_squared_error: 2.3523
Epoch 4/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.3939 - mean_absolute_error: 1.3939 - mean_squared_error: 2.5727Epoch 00003: val_loss improved from 1.36795 to 1.28527, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2512s - loss: 1.3919 - mean_absolute_error: 1.3919 - mean_squared_error: 2.5716 - val_loss: 1.2853 - val_mean_absolute_error: 1.2853 - val_mean_squared_error: 2.1183
Epoch 5/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.3020 - mean_absolute_error: 1.3020 - mean_squared_error: 2.3694Epoch 00004: val_loss improved from 1.28527 to 1.21355, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2275s - loss: 1.3024 - mean_absolute_error: 1.3024 - mean_squared_error: 2.3739 - val_loss: 1.2136 - val_mean_absolute_error: 1.2136 - val_mean_squared_error: 2.1824
Epoch 6/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.3009 - mean_absolute_error: 1.3009 - mean_squared_error: 2.4436Epoch 00005: val_loss did not improve
1050/1050 [==============================] - 2122s - loss: 1.2985 - mean_absolute_error: 1.2985 - mean_squared_error: 2.4354 - val_loss: 1.2352 - val_mean_absolute_error: 1.2352 - val_mean_squared_error: 2.0246
Epoch 7/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.2712 - mean_absolute_error: 1.2712 - mean_squared_error: 2.2954Epoch 00006: val_loss improved from 1.21355 to 1.20963, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2271s - loss: 1.2720 - mean_absolute_error: 1.2720 - mean_squared_error: 2.2985 - val_loss: 1.2096 - val_mean_absolute_error: 1.2096 - val_mean_squared_error: 1.9839
Epoch 8/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.2509 - mean_absolute_error: 1.2509 - mean_squared_error: 2.2247Epoch 00007: val_loss improved from 1.20963 to 1.20380, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2430s - loss: 1.2480 - mean_absolute_error: 1.2480 - mean_squared_error: 2.2168 - val_loss: 1.2038 - val_mean_absolute_error: 1.2038 - val_mean_squared_error: 2.4203
Epoch 9/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.2350 - mean_absolute_error: 1.2350 - mean_squared_error: 2.2455Epoch 00008: val_loss did not improve
1050/1050 [==============================] - 2112s - loss: 1.2348 - mean_absolute_error: 1.2348 - mean_squared_error: 2.2433 - val_loss: 1.2270 - val_mean_absolute_error: 1.2270 - val_mean_squared_error: 2.0797
Epoch 10/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2276 - mean_absolute_error: 1.2276 - mean_squared_error: 2.2485Epoch 00009: val_loss improved from 1.20380 to 1.15868, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2150s - loss: 1.2313 - mean_absolute_error: 1.2313 - mean_squared_error: 2.2572 - val_loss: 1.1587 - val_mean_absolute_error: 1.1587 - val_mean_squared_error: 1.9536
Epoch 11/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2217 - mean_absolute_error: 1.2217 - mean_squared_error: 2.1725Epoch 00010: val_loss did not improve
1050/1050 [==============================] - 2240s - loss: 1.2256 - mean_absolute_error: 1.2256 - mean_squared_error: 2.1942 - val_loss: 1.1896 - val_mean_absolute_error: 1.1896 - val_mean_squared_error: 2.1069
Epoch 12/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1929 - mean_absolute_error: 1.1929 - mean_squared_error: 2.0915Epoch 00011: val_loss did not improve
1050/1050 [==============================] - 2155s - loss: 1.1887 - mean_absolute_error: 1.1887 - mean_squared_error: 2.0788 - val_loss: 1.1617 - val_mean_absolute_error: 1.1617 - val_mean_squared_error: 1.9207
Epoch 13/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1457 - mean_absolute_error: 1.1457 - mean_squared_error: 2.0292Epoch 00012: val_loss did not improve
1050/1050 [==============================] - 2263s - loss: 1.1493 - mean_absolute_error: 1.1493 - mean_squared_error: 2.0383 - val_loss: 1.1760 - val_mean_absolute_error: 1.1760 - val_mean_squared_error: 2.0623
Epoch 14/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.1394 - mean_absolute_error: 1.1394 - mean_squared_error: 1.9871Epoch 00013: val_loss improved from 1.15868 to 1.13617, saving model to /almac/ignacio/40_20_40_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2014s - loss: 1.1443 - mean_absolute_error: 1.1443 - mean_squared_error: 1.9999 - val_loss: 1.1362 - val_mean_absolute_error: 1.1362 - val_mean_squared_error: 1.9920
Epoch 15/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1421 - mean_absolute_error: 1.1421 - mean_squared_error: 2.0221Epoch 00014: val_loss did not improve
1050/1050 [==============================] - 2153s - loss: 1.1387 - mean_absolute_error: 1.1387 - mean_squared_error: 2.0117 - val_loss: 1.1423 - val_mean_absolute_error: 1.1423 - val_mean_squared_error: 1.9384
Epoch 16/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1010 - mean_absolute_error: 1.1010 - mean_squared_error: 1.9121Epoch 00015: val_loss did not improve
1050/1050 [==============================] - 2180s - loss: 1.1002 - mean_absolute_error: 1.1002 - mean_squared_error: 1.9079 - val_loss: 1.1380 - val_mean_absolute_error: 1.1380 - val_mean_squared_error: 2.0995
Epoch 17/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.0599 - mean_absolute_error: 1.0599 - mean_squared_error: 1.8823Epoch 00016: val_loss did not improve
1050/1050 [==============================] - 2249s - loss: 1.0609 - mean_absolute_error: 1.0609 - mean_squared_error: 1.8799 - val_loss: 1.1588 - val_mean_absolute_error: 1.1588 - val_mean_squared_error: 2.1230
Epoch 18/20
1040/1050 [============================>.] - ETA: 21s - loss: 1.0620 - mean_absolute_error: 1.0620 - mean_squared_error: 1.8072Epoch 00017: val_loss did not improve
1050/1050 [==============================] - 2689s - loss: 1.0669 - mean_absolute_error: 1.0669 - mean_squared_error: 1.8254 - val_loss: 1.1700 - val_mean_absolute_error: 1.1700 - val_mean_squared_error: 2.1382
Epoch 19/20
1040/1050 [============================>.] - ETA: 21s - loss: 1.0371 - mean_absolute_error: 1.0371 - mean_squared_error: 1.8077Epoch 00018: val_loss did not improve
1050/1050 [==============================] - 2607s - loss: 1.0350 - mean_absolute_error: 1.0350 - mean_squared_error: 1.8024 - val_loss: 1.1645 - val_mean_absolute_error: 1.1645 - val_mean_squared_error: 2.1764
Epoch 20/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.0524 - mean_absolute_error: 1.0524 - mean_squared_error: 1.8869Epoch 00019: val_loss did not improve
1050/1050 [==============================] - 2473s - loss: 1.0522 - mean_absolute_error: 1.0522 - mean_squared_error: 1.8845 - val_loss: 1.1439 - val_mean_absolute_error: 1.1439 - val_mean_squared_error: 1.9956
Parameters:
---------------------
h_STATES=40
EPOCHS=20
DENSES=40
Representation=fastText
EMBEDDING_DIM=100
MAX_SEQUENCE_LENGTH=50
MODEL_TYPE=stacked
Using Theano backend.
ETA: 47307s 13left 3639.Loanding train and valid dirs......
Starting training
Phrases in STS.input.headlines.txt 750 750
Phrases in STS.input.OnWN.txt 561 561
Phrases in STS.input.FNWN.txt 189 189
Total train phrases /almac/ignacio/data/sts_all/train-2013 1500
Total train phrases 3000
Starting training
Phrases in STS.input.track5.en-en.txt 250
Total train phrases /almac/ignacio/data/sts_all/valid-2017 250
Total train phrases 500
Spliting tab-separated files...
Labels shape: (1500,)
Tokenizing files... [A]
Split training set into train and val... [A]
Tokenizing files... [B]
Split training set into train and val... [B]
Getting embedding matrix... from /almac/ignacio/data/fastText/wikiEn_Full_H100.model.vec
Found 4049540 word vectors.
Filling embedding matrices...
Compiling the model...
____________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
====================================================================================================
embedding_1 (Embedding) (None, 50, 100) 371900
____________________________________________________________________________________________________
attention_1 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_2 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_3 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_4 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_1 (Activation) (None, 40) 0
____________________________________________________________________________________________________
embedding_2 (Embedding) (None, 50, 100) 424400
____________________________________________________________________________________________________
attention_5 (Attention) (None, 50, 40) 34300
____________________________________________________________________________________________________
attention_6 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_7 (Attention) (None, 50, 40) 6520
____________________________________________________________________________________________________
attention_8 (Attention) (None, 40) 6520
____________________________________________________________________________________________________
activation_2 (Activation) (None, 40) 0
____________________________________________________________________________________________________
maxoutdense_1 (MaxoutDense) (None, 50) 16200 merge_1[0][0]
____________________________________________________________________________________________________
maxoutdense_2 (MaxoutDense) (None, 1) 204 maxoutdense_1[0][0]
====================================================================================================
Total params: 920,424
Trainable params: 124,124
Non-trainable params: 796,300
____________________________________________________________________________________________________
None
Happy learning!!!
Train on 1050 samples, validate on 450 samples
Epoch 1/20
1040/1050 [============================>.] - ETA: 27s - loss: 1.4522 - mean_absolute_error: 1.4522 - mean_squared_error: 2.8551Epoch 00000: val_loss improved from inf to 1.39667, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3276s - loss: 1.4485 - mean_absolute_error: 1.4485 - mean_squared_error: 2.8490 - val_loss: 1.3967 - val_mean_absolute_error: 1.3967 - val_mean_squared_error: 2.5961
Epoch 2/20
1040/1050 [============================>.] - ETA: 25s - loss: 1.4202 - mean_absolute_error: 1.4202 - mean_squared_error: 2.6933Epoch 00001: val_loss improved from 1.39667 to 1.38892, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 3078s - loss: 1.4183 - mean_absolute_error: 1.4183 - mean_squared_error: 2.6874 - val_loss: 1.3889 - val_mean_absolute_error: 1.3889 - val_mean_squared_error: 2.4995
Epoch 3/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.4064 - mean_absolute_error: 1.4064 - mean_squared_error: 2.5818Epoch 00002: val_loss improved from 1.38892 to 1.33875, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2411s - loss: 1.4089 - mean_absolute_error: 1.4089 - mean_squared_error: 2.5923 - val_loss: 1.3387 - val_mean_absolute_error: 1.3387 - val_mean_squared_error: 2.2651
Epoch 4/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.3461 - mean_absolute_error: 1.3461 - mean_squared_error: 2.4726Epoch 00003: val_loss improved from 1.33875 to 1.33298, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2587s - loss: 1.3474 - mean_absolute_error: 1.3474 - mean_squared_error: 2.4748 - val_loss: 1.3330 - val_mean_absolute_error: 1.3330 - val_mean_squared_error: 2.3610
Epoch 5/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.3181 - mean_absolute_error: 1.3181 - mean_squared_error: 2.4053Epoch 00004: val_loss improved from 1.33298 to 1.27312, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2258s - loss: 1.3161 - mean_absolute_error: 1.3161 - mean_squared_error: 2.3982 - val_loss: 1.2731 - val_mean_absolute_error: 1.2731 - val_mean_squared_error: 2.4572
Epoch 6/20
1040/1050 [============================>.] - ETA: 20s - loss: 1.2665 - mean_absolute_error: 1.2665 - mean_squared_error: 2.3540Epoch 00005: val_loss did not improve
1050/1050 [==============================] - 2514s - loss: 1.2675 - mean_absolute_error: 1.2675 - mean_squared_error: 2.3582 - val_loss: 1.5086 - val_mean_absolute_error: 1.5086 - val_mean_squared_error: 3.6239
Epoch 7/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.2228 - mean_absolute_error: 1.2228 - mean_squared_error: 2.2425Epoch 00006: val_loss did not improve
1050/1050 [==============================] - 2420s - loss: 1.2211 - mean_absolute_error: 1.2211 - mean_squared_error: 2.2343 - val_loss: 1.4865 - val_mean_absolute_error: 1.4865 - val_mean_squared_error: 3.7128
Epoch 8/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.2039 - mean_absolute_error: 1.2039 - mean_squared_error: 2.2305Epoch 00007: val_loss did not improve
1050/1050 [==============================] - 2151s - loss: 1.2026 - mean_absolute_error: 1.2026 - mean_squared_error: 2.2239 - val_loss: 1.3063 - val_mean_absolute_error: 1.3063 - val_mean_squared_error: 2.5354
Epoch 9/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1820 - mean_absolute_error: 1.1820 - mean_squared_error: 2.1469Epoch 00008: val_loss improved from 1.27312 to 1.18658, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2270s - loss: 1.1822 - mean_absolute_error: 1.1822 - mean_squared_error: 2.1460 - val_loss: 1.1866 - val_mean_absolute_error: 1.1866 - val_mean_squared_error: 2.1257
Epoch 10/20
1040/1050 [============================>.] - ETA: 17s - loss: 1.1823 - mean_absolute_error: 1.1823 - mean_squared_error: 2.1822Epoch 00009: val_loss did not improve
1050/1050 [==============================] - 2159s - loss: 1.1788 - mean_absolute_error: 1.1788 - mean_squared_error: 2.1752 - val_loss: 1.1912 - val_mean_absolute_error: 1.1912 - val_mean_squared_error: 2.0940
Epoch 11/20
1040/1050 [============================>.] - ETA: 19s - loss: 1.1712 - mean_absolute_error: 1.1712 - mean_squared_error: 2.1304Epoch 00010: val_loss did not improve
1050/1050 [==============================] - 2402s - loss: 1.1720 - mean_absolute_error: 1.1720 - mean_squared_error: 2.1286 - val_loss: 1.3327 - val_mean_absolute_error: 1.3327 - val_mean_squared_error: 2.3660
Epoch 12/20
1040/1050 [============================>.] - ETA: 18s - loss: 1.1360 - mean_absolute_error: 1.1360 - mean_squared_error: 2.0058Epoch 00011: val_loss improved from 1.18658 to 1.15046, saving model to /almac/ignacio/40_20_50_fastText_100_50_stacked.hdf5
1050/1050 [==============================] - 2226s - loss: 1.1396 - mean_absolute_error: 1.1396 - mean_squared_error: 2.0186 - val_loss: 1.1505 - val_mean_absolute_error: 1.1505 - val_mean_squared_error: 1.9125
Epoch 13/20
1040/1050 [============================>.] - ETA: 15s - loss: 1.1094 - mean_absolute_error: 1.1094 - mean_squared_error: 2.0072Epoch 00012: val_loss did not improve
1050/1050 [==============================] - 1992s - loss: 1.1168 - mean_absolute_error: 1.1168 - mean_squared_error: 2.0354 - val_loss: 1.1734 - val_mean_absolute_error: 1.1734 - val_mean_squared_error: 2.2065
Epoch 14/20
1040/1050 [============================>.] - ETA: 16s - loss: 1.0855 - mean_absolute_error: 1.0855 - mean_squared_error: 1.9231Epoch 00013: val_loss did not improve
1050/1050 [==============================] - 1994s - loss: 1.0825 - mean_absolute_error: 1.0825 - mean_squared_error: 1.9119 - val_loss: 1.1967 - val_mean_absolute_error: 1.1967 - val_mean_squared_error: 2.3252
Epoch 15/20