amoudgl · May 30, 2025 17:10
diff --git a/output.log b/output.log
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [41 53 50 42  1 40 50 53 53 42] ...
 First target sequence: [53 50 42  1 40 50 53 53 42  1] ...

 Decoded input: cold blood no spark of honour bides.

 NORTHUMBERLAND:
 Be thou a 

 Decoded target: old blood no spark of honour bides.

 NORTHUMBERLAND:
 Be thou a p
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.203513145446777
 [Seed 0] Step 0 --> val loss 4.184424877166748
 Step 10: loss 3.8755393028259277
 Step 20: loss 3.3440024852752686
 Step 30: loss 2.823532819747925
 Step 40: loss 2.6751527786254883
 Step 50: loss 2.639468193054199
 Step 60: loss 2.4272749423980713
 Step 70: loss 2.4186179637908936
 Step 80: loss 2.3906514644622803
 Step 90: loss 2.2697482109069824
 Step 100: loss 2.31880521774292
 Step 110: loss 2.2653324604034424
 Step 120: loss 2.2177329063415527
 Step 130: loss 2.1910812854766846
 Step 140: loss 2.2634665966033936
 Step 150: loss 2.1533682346343994
 Step 160: loss 2.1719441413879395
 Step 170: loss 2.131915330886841
 Step 180: loss 2.072336435317993
 Step 190: loss 2.1011576652526855
 Step 200: loss 2.101720094680786
 [Seed 0] Step 200 --> val loss 2.3442766666412354
 Step 210: loss 2.041682004928589
 Step 220: loss 2.1039676666259766
 Step 230: loss 1.939509391784668
 Step 240: loss 2.090090751647949
 Step 250: loss 1.905335783958435
 Step 260: loss 2.120622396469116
 Step 270: loss 2.069044828414917
 Step 280: loss 2.0691990852355957
 Step 290: loss 2.0319080352783203
 Step 300: loss 1.9588054418563843
 Step 310: loss 2.024789810180664
 Step 320: loss 2.1015625
 Step 330: loss 2.165870189666748
 Step 340: loss 2.081207036972046
 Step 350: loss 1.9762403964996338
 Step 360: loss 1.9477362632751465
 Step 370: loss 2.0384650230407715
 Step 380: loss 2.0236940383911133
 Step 390: loss 2.014063596725464
 Step 400: loss 2.083286762237549
 [Seed 0] Step 400 --> val loss 2.255828857421875
 Step 410: loss 2.08640718460083
 Step 420: loss 1.8395723104476929
 Step 430: loss 1.9152625799179077
 Step 440: loss 1.9772871732711792
 Step 450: loss 2.040564775466919
 Step 460: loss 2.0196304321289062
 Step 470: loss 2.013015031814575
 Step 480: loss 1.9669212102890015
 Step 490: loss 2.0130650997161865
 Step 500: loss 2.036425828933716
 Step 510: loss 2.1175413131713867
 Step 520: loss 1.977731466293335
 Step 530: loss 1.9930343627929688
 Step 540: loss 1.9699079990386963
 Step 550: loss 2.1087167263031006
 Step 560: loss 1.9902414083480835
 Step 570: loss 2.0059075355529785
 Step 580: loss 2.015501022338867
 Step 590: loss 1.9826616048812866
 Step 600: loss 2.0828282833099365
 [Seed 0] Step 600 --> val loss 2.3087728023529053
 Step 610: loss 2.118818759918213
 Step 620: loss 2.057274341583252
 Step 630: loss 2.030327320098877
 Step 640: loss 2.088131904602051
 Step 650: loss 2.0522913932800293
 Step 660: loss 2.159937858581543
 Step 670: loss 2.0310025215148926
 Step 680: loss 2.0159595012664795
 Step 690: loss 1.9347312450408936
 Step 700: loss 1.9081878662109375
 Step 710: loss 2.015293598175049
 Step 720: loss 2.118241786956787
 Step 730: loss 1.9559059143066406
 Step 740: loss 1.9692649841308594
 Step 750: loss 1.9954551458358765
 Step 760: loss 2.0196330547332764
 Step 770: loss 1.8938651084899902
 Step 780: loss 2.147998332977295
 Step 790: loss 1.9707190990447998
 Step 800: loss 2.0567052364349365
 [Seed 0] Step 800 --> val loss 2.225918769836426
 Step 810: loss 2.0533173084259033
 Step 820: loss 2.0345635414123535
 Step 830: loss 1.9762762784957886
 Step 840: loss 1.9515161514282227
 Step 850: loss 2.0251035690307617
 Step 860: loss 1.9455962181091309
 Step 870: loss 1.9751957654953003
 Step 880: loss 2.079394817352295
 Step 890: loss 2.060420513153076
 Step 900: loss 2.045344352722168
 Step 910: loss 2.095691442489624
 Step 920: loss 1.9943361282348633
 Step 930: loss 2.1008403301239014
 Step 940: loss 1.9401429891586304
 Step 950: loss 2.0271739959716797
 Step 960: loss 2.089696168899536
 Step 970: loss 2.0341150760650635
 Step 980: loss 1.982370376586914
 Step 990: loss 2.02970552444458
 Step 1000: loss 2.0169272422790527
 [Seed 0] Step 1000 --> val loss 2.193075180053711
 Step 1010: loss 2.0183870792388916
 Step 1020: loss 1.9687312841415405
 Step 1030: loss 1.9656918048858643
 Step 1040: loss 1.974206805229187
 Step 1050: loss 1.9943177700042725
 Step 1060: loss 1.9310483932495117
 Step 1070: loss 2.0238187313079834
 Step 1080: loss 1.9779443740844727
 Step 1090: loss 1.9208104610443115
 Step 1100: loss 1.8854224681854248
 Step 1110: loss 2.0174098014831543
 Step 1120: loss 1.9952392578125
 Step 1130: loss 1.841570258140564
 Step 1140: loss 1.9395731687545776
 Step 1150: loss 2.107750654220581
 Step 1160: loss 1.7982425689697266
 Step 1170: loss 1.9853801727294922
 Step 1180: loss 1.9161323308944702
 Step 1190: loss 1.9696388244628906
 Step 1200: loss 1.9412765502929688
 [Seed 0] Step 1200 --> val loss 2.1652212142944336
 Step 1210: loss 1.9485868215560913
 Step 1220: loss 2.0038580894470215
 Step 1230: loss 2.06624436378479
 Step 1240: loss 1.8946824073791504
 Step 1250: loss 2.003060817718506
 Step 1260: loss 1.734341025352478
 Step 1270: loss 2.0274081230163574
 Step 1280: loss 1.9721460342407227
 Step 1290: loss 1.9001891613006592
 Step 1300: loss 1.904750108718872
 Step 1310: loss 1.9606221914291382
 Step 1320: loss 1.862774133682251
 Step 1330: loss 1.9471458196640015
 Step 1340: loss 1.9263395071029663
 Step 1350: loss 1.9209434986114502
 Step 1360: loss 1.7817111015319824
 Step 1370: loss 1.9202250242233276
 Step 1380: loss 1.968883752822876
 Step 1390: loss 1.8805469274520874
 Step 1400: loss 1.8843587636947632
 [Seed 0] Step 1400 --> val loss 2.1005148887634277
 Step 1410: loss 1.8910901546478271
 Step 1420: loss 1.9073057174682617
 Step 1430: loss 1.8609507083892822
 Step 1440: loss 1.8716516494750977
 Step 1450: loss 1.8873202800750732
 Step 1460: loss 1.8178585767745972
 Step 1470: loss 1.824113130569458
 Step 1480: loss 1.8416898250579834
 Step 1490: loss 1.8174021244049072
 Step 1500: loss 1.7889807224273682
 Step 1510: loss 1.839398741722107
 Step 1520: loss 1.851144790649414
 Step 1530: loss 1.8084306716918945
 Step 1540: loss 1.896733283996582
 Step 1550: loss 1.7554454803466797
 Step 1560: loss 1.9044580459594727
 Step 1570: loss 1.9628167152404785
 Step 1580: loss 1.8873027563095093
 Step 1590: loss 1.7429665327072144
 Step 1600: loss 1.9159305095672607
 [Seed 0] Step 1600 --> val loss 2.1043198108673096
 Step 1610: loss 1.8371875286102295
 Step 1620: loss 1.742522954940796
 Step 1630: loss 1.775038719177246
 Step 1640: loss 1.784301519393921
 Step 1650: loss 1.8504419326782227
 Step 1660: loss 1.8338903188705444
 Step 1670: loss 1.7117379903793335
 Step 1680: loss 1.8087573051452637
 Step 1690: loss 1.792744517326355
 Step 1700: loss 1.6915960311889648
 Step 1710: loss 1.7813143730163574
 Step 1720: loss 1.8692944049835205
 Step 1730: loss 1.773542881011963
 Step 1740: loss 1.6500132083892822
 Step 1750: loss 1.7085274457931519
 Step 1760: loss 1.8454358577728271
 Step 1770: loss 1.7564308643341064
 Step 1780: loss 1.795449137687683
 Step 1790: loss 1.7341097593307495
 Step 1800: loss 1.707435131072998
 [Seed 0] Step 1800 --> val loss 1.9792101383209229
 Step 1810: loss 1.7476882934570312
 Step 1820: loss 1.732509970664978
 Step 1830: loss 1.7710559368133545
 Step 1840: loss 1.7673768997192383
 Step 1850: loss 1.7840121984481812
 Step 1860: loss 1.6943906545639038
 Step 1870: loss 1.7900470495224
 Step 1880: loss 1.722694754600525
 Step 1890: loss 1.7041583061218262
 Step 1900: loss 1.5803006887435913
 Step 1910: loss 1.7438924312591553
 Step 1920: loss 1.7683022022247314
 Step 1930: loss 1.7917797565460205
 Step 1940: loss 1.8645977973937988
 Step 1950: loss 1.7890580892562866
 Step 1960: loss 1.7577745914459229
 Step 1970: loss 1.8413145542144775
 Step 1980: loss 1.6953294277191162
 Step 1990: loss 1.7728956937789917
 Step 2000: loss 1.7080835103988647
 [Seed 0] Step 2000 --> val loss 1.882563829421997
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [47 52 45  1 44 43 50 50 53 61] ...
 First target sequence: [52 45  1 44 43 50 50 53 61  8] ...

 Decoded input: ing fellow.

 DUKE VINCENTIO:
 Either this is the envy in you, fol

 Decoded target: ng fellow.

 DUKE VINCENTIO:
 Either this is the envy in you, foll
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.201435565948486
 [Seed 1] Step 0 --> val loss 4.184230327606201
 Step 10: loss 3.826808452606201
 Step 20: loss 3.3518035411834717
 Step 30: loss 2.8810763359069824
 Step 40: loss 2.5754048824310303
 Step 50: loss 2.5998082160949707
 Step 60: loss 2.539147138595581
 Step 70: loss 2.425325393676758
 Step 80: loss 2.327491521835327
 Step 90: loss 2.304891347885132
 Step 100: loss 2.4065098762512207
 Step 110: loss 2.2053544521331787
 Step 120: loss 2.1936097145080566
 Step 130: loss 2.2447166442871094
 Step 140: loss 2.1238820552825928
 Step 150: loss 2.157510757446289
 Step 160: loss 2.0688982009887695
 Step 170: loss 2.104326009750366
 Step 180: loss 2.1606292724609375
 Step 190: loss 2.1703603267669678
 Step 200: loss 2.142533302307129
 [Seed 1] Step 200 --> val loss 2.3397679328918457
 Step 210: loss 2.0704550743103027
 Step 220: loss 2.104041814804077
 Step 230: loss 2.042806386947632
 Step 240: loss 2.127939224243164
 Step 250: loss 2.064009189605713
 Step 260: loss 1.9950675964355469
 Step 270: loss 2.0418598651885986
 Step 280: loss 1.8231418132781982
 Step 290: loss 2.1381382942199707
 Step 300: loss 1.9837144613265991
 Step 310: loss 2.0491831302642822
 Step 320: loss 2.1908981800079346
 Step 330: loss 1.9017839431762695
 Step 340: loss 2.058954954147339
 Step 350: loss 2.0636355876922607
 Step 360: loss 2.1774842739105225
 Step 370: loss 2.017702102661133
 Step 380: loss 1.937143325805664
 Step 390: loss 2.0535178184509277
 Step 400: loss 2.0583677291870117
 [Seed 1] Step 400 --> val loss 2.2703394889831543
 Step 410: loss 1.9106569290161133
 Step 420: loss 2.0981907844543457
 Step 430: loss 1.999199390411377
 Step 440: loss 2.1491856575012207
 Step 450: loss 2.0062575340270996
 Step 460: loss 1.9753049612045288
 Step 470: loss 1.9988707304000854
 Step 480: loss 2.0324559211730957
 Step 490: loss 2.0000159740448
 Step 500: loss 2.0276694297790527
 Step 510: loss 2.1141843795776367
 Step 520: loss 1.9808509349822998
 Step 530: loss 1.9573602676391602
 Step 540: loss 2.0825388431549072
 Step 550: loss 1.9425325393676758
 Step 560: loss 2.0621280670166016
 Step 570: loss 1.9541442394256592
 Step 580: loss 2.03824520111084
 Step 590: loss 2.0350990295410156
 Step 600: loss 2.0190815925598145
 [Seed 1] Step 600 --> val loss 2.297309160232544
 Step 610: loss 2.0928854942321777
 Step 620: loss 2.0487918853759766
 Step 630: loss 2.0119786262512207
 Step 640: loss 1.9719157218933105
 Step 650: loss 1.9756765365600586
 Step 660: loss 2.02876615524292
 Step 670: loss 1.9264930486679077
 Step 680: loss 1.9119195938110352
 Step 690: loss 1.9168498516082764
 Step 700: loss 2.040597915649414
 Step 710: loss 2.104682445526123
 Step 720: loss 2.0138678550720215
 Step 730: loss 1.9414151906967163
 Step 740: loss 1.9232537746429443
 Step 750: loss 1.9288246631622314
 Step 760: loss 2.040181875228882
 Step 770: loss 2.0077168941497803
 Step 780: loss 2.228788375854492
 Step 790: loss 2.017981767654419
 Step 800: loss 1.9747540950775146
 [Seed 1] Step 800 --> val loss 2.235487937927246
 Step 810: loss 1.940457820892334
 Step 820: loss 1.9079320430755615
 Step 830: loss 1.9435040950775146
 Step 840: loss 2.031980037689209
 Step 850: loss 2.1299057006835938
 Step 860: loss 2.038990020751953
 Step 870: loss 2.090453624725342
 Step 880: loss 2.101539134979248
 Step 890: loss 2.0866737365722656
 Step 900: loss 1.992649793624878
 Step 910: loss 1.9580906629562378
 Step 920: loss 1.9796769618988037
 Step 930: loss 2.002166271209717
 Step 940: loss 1.9581310749053955
 Step 950: loss 2.015430212020874
 Step 960: loss 2.0046920776367188
 Step 970: loss 2.032437324523926
 Step 980: loss 1.965467929840088
 Step 990: loss 1.9853280782699585
 Step 1000: loss 2.1040234565734863
 [Seed 1] Step 1000 --> val loss 2.2832844257354736
 Step 1010: loss 1.9870119094848633
 Step 1020: loss 2.0071444511413574
 Step 1030: loss 2.0529706478118896
 Step 1040: loss 2.055840015411377
 Step 1050: loss 2.068709373474121
 Step 1060: loss 2.0921218395233154
 Step 1070: loss 2.0507242679595947
 Step 1080: loss 2.0384321212768555
 Step 1090: loss 2.0895261764526367
 Step 1100: loss 2.040600538253784
 Step 1110: loss 2.133573293685913
 Step 1120: loss 2.091655969619751
 Step 1130: loss 2.0119338035583496
 Step 1140: loss 2.002981662750244
 Step 1150: loss 2.1256844997406006
 Step 1160: loss 1.9513463973999023
 Step 1170: loss 2.052490711212158
 Step 1180: loss 2.0170905590057373
 Step 1190: loss 1.9952342510223389
 Step 1200: loss 1.9760768413543701
 [Seed 1] Step 1200 --> val loss 2.2823028564453125
 Step 1210: loss 2.01574969291687
 Step 1220: loss 2.158975124359131
 Step 1230: loss 2.082098960876465
 Step 1240: loss 2.195901870727539
 Step 1250: loss 2.0241708755493164
 Step 1260: loss 2.049360752105713
 Step 1270: loss 2.0480523109436035
 Step 1280: loss 2.063655376434326
 Step 1290: loss 2.1121864318847656
 Step 1300: loss 2.0200448036193848
 Step 1310: loss 2.1323041915893555
 Step 1320: loss 1.9962527751922607
 Step 1330: loss 1.969611644744873
 Step 1340: loss 1.9201982021331787
 Step 1350: loss 1.9417073726654053
 Step 1360: loss 1.878141164779663
 Step 1370: loss 2.0307650566101074
 Step 1380: loss 1.9918572902679443
 Step 1390: loss 1.869370460510254
 Step 1400: loss 1.9692192077636719
 [Seed 1] Step 1400 --> val loss 2.142310380935669
 Step 1410: loss 1.978360652923584
 Step 1420: loss 1.9254722595214844
 Step 1430: loss 2.057528018951416
 Step 1440: loss 2.023013114929199
 Step 1450: loss 2.0696332454681396
 Step 1460: loss 1.9306206703186035
 Step 1470: loss 1.9296655654907227
 Step 1480: loss 1.8745157718658447
 Step 1490: loss 2.054474353790283
 Step 1500: loss 1.9275891780853271
 Step 1510: loss 2.125953435897827
 Step 1520: loss 2.008172035217285
 Step 1530: loss 1.9065617322921753
 Step 1540: loss 1.8448951244354248
 Step 1550: loss 1.9716832637786865
 Step 1560: loss 1.8598620891571045
 Step 1570: loss 1.7417157888412476
 Step 1580: loss 1.9368999004364014
 Step 1590: loss 1.7414536476135254
 Step 1600: loss 1.835615873336792
 [Seed 1] Step 1600 --> val loss 2.131161689758301
 Step 1610: loss 1.845981478691101
 Step 1620: loss 1.945056676864624
 Step 1630: loss 1.9881185293197632
 Step 1640: loss 1.8158725500106812
 Step 1650: loss 1.7889041900634766
 Step 1660: loss 1.8129088878631592
 Step 1670: loss 1.8116258382797241
 Step 1680: loss 1.8587634563446045
 Step 1690: loss 1.8083276748657227
 Step 1700: loss 1.812300443649292
 Step 1710: loss 1.8134136199951172
 Step 1720: loss 1.7792766094207764
 Step 1730: loss 1.8308404684066772
 Step 1740: loss 1.7313518524169922
 Step 1750: loss 1.7763960361480713
 Step 1760: loss 1.825465440750122
 Step 1770: loss 1.6776018142700195
 Step 1780: loss 1.8353898525238037
 Step 1790: loss 1.9070982933044434
 Step 1800: loss 1.8438314199447632
 [Seed 1] Step 1800 --> val loss 2.0009591579437256
 Step 1810: loss 1.7950975894927979
 Step 1820: loss 1.795446753501892
 Step 1830: loss 1.6837403774261475
 Step 1840: loss 1.732130527496338
 Step 1850: loss 1.7003002166748047
 Step 1860: loss 1.7720859050750732
 Step 1870: loss 1.7022209167480469
 Step 1880: loss 1.7520641088485718
 Step 1890: loss 1.775509238243103
 Step 1900: loss 1.8347527980804443
 Step 1910: loss 1.8975566625595093
 Step 1920: loss 1.8971736431121826
 Step 1930: loss 1.8010928630828857
 Step 1940: loss 1.766225814819336
 Step 1950: loss 1.7344011068344116
 Step 1960: loss 1.7594307661056519
 Step 1970: loss 1.8423800468444824
 Step 1980: loss 1.8122658729553223
 Step 1990: loss 1.7484560012817383
 Step 2000: loss 1.7049546241760254
 [Seed 1] Step 2000 --> val loss 1.9162565469741821
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [46 43 47 56  1 58 56 39 52 57] ...
 First target sequence: [43 47 56  1 58 56 39 52 57 44] ...

 Decoded input: heir transformations
 Were never for a piece of beauty rarer,
 Nor

 Decoded target: eir transformations
 Were never for a piece of beauty rarer,
 Nor 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2010321617126465
 [Seed 2] Step 0 --> val loss 4.180516719818115
 Step 10: loss 3.8330140113830566
 Step 20: loss 3.322075605392456
 Step 30: loss 2.783907413482666
 Step 40: loss 2.6356067657470703
 Step 50: loss 2.6221225261688232
 Step 60: loss 2.505398750305176
 Step 70: loss 2.3984692096710205
 Step 80: loss 2.3970460891723633
 Step 90: loss 2.3659560680389404
 Step 100: loss 2.3711278438568115
 Step 110: loss 2.187562942504883
 Step 120: loss 2.1952052116394043
 Step 130: loss 2.2460474967956543
 Step 140: loss 2.124152898788452
 Step 150: loss 2.235442638397217
 Step 160: loss 2.166515827178955
 Step 170: loss 2.11276912689209
 Step 180: loss 2.1530468463897705
 Step 190: loss 2.0865979194641113
 Step 200: loss 2.120960235595703
 [Seed 2] Step 200 --> val loss 2.3786253929138184
 Step 210: loss 2.1351516246795654
 Step 220: loss 1.9888713359832764
 Step 230: loss 2.1113152503967285
 Step 240: loss 2.15104341506958
 Step 250: loss 2.1067681312561035
 Step 260: loss 1.9804280996322632
 Step 270: loss 2.11173677444458
 Step 280: loss 2.085597038269043
 Step 290: loss 2.1047115325927734
 Step 300: loss 2.0400760173797607
 Step 310: loss 1.9774264097213745
 Step 320: loss 2.0657503604888916
 Step 330: loss 2.1093788146972656
 Step 340: loss 2.067049503326416
 Step 350: loss 2.1318089962005615
 Step 360: loss 2.051811695098877
 Step 370: loss 2.121663808822632
 Step 380: loss 1.8732904195785522
 Step 390: loss 1.990372657775879
 Step 400: loss 1.9831722974777222
 [Seed 2] Step 400 --> val loss 2.2444207668304443
 Step 410: loss 2.089083671569824
 Step 420: loss 1.971304178237915
 Step 430: loss 2.0147435665130615
 Step 440: loss 2.0127458572387695
 Step 450: loss 1.9213714599609375
 Step 460: loss 1.9287796020507812
 Step 470: loss 2.054202079772949
 Step 480: loss 1.9303382635116577
 Step 490: loss 1.9558677673339844
 Step 500: loss 1.9423580169677734
 Step 510: loss 1.9706737995147705
 Step 520: loss 1.9470875263214111
 Step 530: loss 2.0282599925994873
 Step 540: loss 2.014292001724243
 Step 550: loss 1.9516408443450928
 Step 560: loss 2.0023953914642334
 Step 570: loss 1.941204309463501
 Step 580: loss 2.0667612552642822
 Step 590: loss 2.0478768348693848
 Step 600: loss 1.9913203716278076
 [Seed 2] Step 600 --> val loss 2.2836811542510986
 Step 610: loss 1.9483470916748047
 Step 620: loss 2.041928291320801
 Step 630: loss 2.0125110149383545
 Step 640: loss 2.012003183364868
 Step 650: loss 2.0506062507629395
 Step 660: loss 2.0391151905059814
 Step 670: loss 2.104679822921753
 Step 680: loss 2.0141665935516357
 Step 690: loss 2.0919575691223145
 Step 700: loss 2.0070102214813232
 Step 710: loss 2.0590972900390625
 Step 720: loss 2.087210178375244
 Step 730: loss 2.0574350357055664
 Step 740: loss 1.9068667888641357
 Step 750: loss 2.0436997413635254
 Step 760: loss 1.9966095685958862
 Step 770: loss 1.9911963939666748
 Step 780: loss 1.9145715236663818
 Step 790: loss 1.96311354637146
 Step 800: loss 1.9645700454711914
 [Seed 2] Step 800 --> val loss 2.222256898880005
 Step 810: loss 1.9049123525619507
 Step 820: loss 2.0083632469177246
 Step 830: loss 1.9329860210418701
 Step 840: loss 2.032127857208252
 Step 850: loss 1.9777196645736694
 Step 860: loss 1.9070842266082764
 Step 870: loss 1.896843671798706
 Step 880: loss 1.987690806388855
 Step 890: loss 2.0351953506469727
 Step 900: loss 2.040099620819092
 Step 910: loss 1.9936860799789429
 Step 920: loss 1.9474241733551025
 Step 930: loss 1.8872060775756836
 Step 940: loss 1.9096578359603882
 Step 950: loss 1.9670261144638062
 Step 960: loss 2.0288658142089844
 Step 970: loss 1.986647605895996
 Step 980: loss 2.0295004844665527
 Step 990: loss 1.8994859457015991
 Step 1000: loss 1.968071699142456
 [Seed 2] Step 1000 --> val loss 2.243051290512085
 Step 1010: loss 1.9911245107650757
 Step 1020: loss 1.999765396118164
 Step 1030: loss 1.9864381551742554
 Step 1040: loss 2.0355803966522217
 Step 1050: loss 1.9511760473251343
 Step 1060: loss 1.9689829349517822
 Step 1070: loss 1.9173096418380737
 Step 1080: loss 1.973585844039917
 Step 1090: loss 1.9630885124206543
 Step 1100: loss 1.823664665222168
 Step 1110: loss 1.9537023305892944
 Step 1120: loss 1.845656394958496
 Step 1130: loss 1.8345470428466797
 Step 1140: loss 1.9401620626449585
 Step 1150: loss 2.0539021492004395
 Step 1160: loss 1.9807040691375732
 Step 1170: loss 1.974548101425171
 Step 1180: loss 1.8044140338897705
 Step 1190: loss 1.8816198110580444
 Step 1200: loss 1.8911192417144775
 [Seed 2] Step 1200 --> val loss 2.109523057937622
 Step 1210: loss 1.7798278331756592
 Step 1220: loss 1.9077141284942627
 Step 1230: loss 1.88787841796875
 Step 1240: loss 1.9787406921386719
 Step 1250: loss 1.8952734470367432
 Step 1260: loss 1.930999755859375
 Step 1270: loss 2.019914150238037
 Step 1280: loss 1.870304822921753
 Step 1290: loss 1.8965095281600952
 Step 1300: loss 1.8689416646957397
 Step 1310: loss 1.8438243865966797
 Step 1320: loss 1.9592931270599365
 Step 1330: loss 1.9114406108856201
 Step 1340: loss 1.9289344549179077
 Step 1350: loss 1.8019789457321167
 Step 1360: loss 1.9908431768417358
 Step 1370: loss 1.8441623449325562
 Step 1380: loss 1.8176934719085693
 Step 1390: loss 1.8311008214950562
 Step 1400: loss 1.808402419090271
 [Seed 2] Step 1400 --> val loss 1.9844568967819214
 Step 1410: loss 1.8888719081878662
 Step 1420: loss 1.7722680568695068
 Step 1430: loss 1.8644040822982788
 Step 1440: loss 1.8180670738220215
 Step 1450: loss 1.810409426689148
 Step 1460: loss 1.7117478847503662
 Step 1470: loss 1.8130525350570679
 Step 1480: loss 1.7821292877197266
 Step 1490: loss 1.7414765357971191
 Step 1500: loss 1.828959345817566
 Step 1510: loss 1.7986323833465576
 Step 1520: loss 1.7719863653182983
 Step 1530: loss 1.7241796255111694
 Step 1540: loss 1.8567113876342773
 Step 1550: loss 1.9133405685424805
 Step 1560: loss 1.7090868949890137
 Step 1570: loss 1.729774832725525
 Step 1580: loss 1.814288854598999
 Step 1590: loss 1.797501564025879
 Step 1600: loss 1.7263044118881226
 [Seed 2] Step 1600 --> val loss 1.9413795471191406
 Step 1610: loss 1.7674260139465332
 Step 1620: loss 1.76531982421875
 Step 1630: loss 1.7151750326156616
 Step 1640: loss 1.772281527519226
 Step 1650: loss 1.8019659519195557
 Step 1660: loss 1.8321187496185303
 Step 1670: loss 1.7497632503509521
 Step 1680: loss 1.812585711479187
 Step 1690: loss 1.697657585144043
 Step 1700: loss 1.5893501043319702
 Step 1710: loss 1.5498101711273193
 Step 1720: loss 1.7545263767242432
 Step 1730: loss 1.7064568996429443
 Step 1740: loss 1.7369632720947266
 Step 1750: loss 1.6422085762023926
 Step 1760: loss 1.6362743377685547
 Step 1770: loss 1.7212951183319092
 Step 1780: loss 1.6572641134262085
 Step 1790: loss 1.67829167842865
 Step 1800: loss 1.8044722080230713
 [Seed 2] Step 1800 --> val loss 1.8161287307739258
 Step 1810: loss 1.6386191844940186
 Step 1820: loss 1.7442216873168945
 Step 1830: loss 1.8304438591003418
 Step 1840: loss 1.6993825435638428
 Step 1850: loss 1.7231028079986572
 Step 1860: loss 1.6813677549362183
 Step 1870: loss 1.6773509979248047
 Step 1880: loss 1.7748260498046875
 Step 1890: loss 1.7178587913513184
 Step 1900: loss 1.8133777379989624
 Step 1910: loss 1.6080541610717773
 Step 1920: loss 1.751643180847168
 Step 1930: loss 1.7439063787460327
 Step 1940: loss 1.6635764837265015
 Step 1950: loss 1.6934518814086914
 Step 1960: loss 1.5969560146331787
 Step 1970: loss 1.81111741065979
 Step 1980: loss 1.6974904537200928
 Step 1990: loss 1.6404218673706055
 Step 2000: loss 1.6821413040161133
 [Seed 2] Step 2000 --> val loss 1.7845371961593628
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 6  1 58 39 49 43  1 58 46 43] ...
 First target sequence: [ 1 58 39 49 43  1 58 46 43  1] ...

 Decoded input: , take the time; kneel down, kneel down:
 Nay, when? strike now, 

 Decoded target:  take the time; kneel down, kneel down:
 Nay, when? strike now, o
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.215667724609375
 [Seed 3] Step 0 --> val loss 4.185371398925781
 Step 10: loss 3.856323719024658
 Step 20: loss 3.3264760971069336
 Step 30: loss 2.9275569915771484
 Step 40: loss 2.6147568225860596
 Step 50: loss 2.5374770164489746
 Step 60: loss 2.5010251998901367
 Step 70: loss 2.457695960998535
 Step 80: loss 2.401482582092285
 Step 90: loss 2.2976737022399902
 Step 100: loss 2.3292596340179443
 Step 110: loss 2.2777440547943115
 Step 120: loss 2.359194040298462
 Step 130: loss 2.240993022918701
 Step 140: loss 2.213015556335449
 Step 150: loss 2.067584753036499
 Step 160: loss 2.1382694244384766
 Step 170: loss 2.1102609634399414
 Step 180: loss 2.1266441345214844
 Step 190: loss 2.187835216522217
 Step 200: loss 2.077753782272339
 [Seed 3] Step 200 --> val loss 2.377793550491333
 Step 210: loss 2.008286476135254
 Step 220: loss 2.2031543254852295
 Step 230: loss 2.14080810546875
 Step 240: loss 2.0696675777435303
 Step 250: loss 1.9883923530578613
 Step 260: loss 1.9878332614898682
 Step 270: loss 2.027038097381592
 Step 280: loss 2.0755720138549805
 Step 290: loss 2.0415258407592773
 Step 300: loss 2.005178928375244
 Step 310: loss 2.0425515174865723
 Step 320: loss 2.166232109069824
 Step 330: loss 2.0116841793060303
 Step 340: loss 1.9575355052947998
 Step 350: loss 1.940529704093933
 Step 360: loss 1.9569791555404663
 Step 370: loss 2.002218008041382
 Step 380: loss 1.851536750793457
 Step 390: loss 1.9265735149383545
 Step 400: loss 2.0001938343048096
 [Seed 3] Step 400 --> val loss 2.345595359802246
 Step 410: loss 2.064126968383789
 Step 420: loss 2.072807788848877
 Step 430: loss 2.0755410194396973
 Step 440: loss 1.9288413524627686
 Step 450: loss 2.0583596229553223
 Step 460: loss 2.0428669452667236
 Step 470: loss 2.0030555725097656
 Step 480: loss 1.9315340518951416
 Step 490: loss 2.0176382064819336
 Step 500: loss 1.9354790449142456
 Step 510: loss 1.976983904838562
 Step 520: loss 2.004683494567871
 Step 530: loss 1.9990205764770508
 Step 540: loss 1.9585803747177124
 Step 550: loss 2.0116448402404785
 Step 560: loss 1.9494330883026123
 Step 570: loss 2.03407621383667
 Step 580: loss 1.9369057416915894
 Step 590: loss 2.0291898250579834
 Step 600: loss 1.9587600231170654
 [Seed 3] Step 600 --> val loss 2.3108155727386475
 Step 610: loss 1.9499022960662842
 Step 620: loss 1.9709762334823608
 Step 630: loss 1.8807382583618164
 Step 640: loss 1.8755563497543335
 Step 650: loss 1.9994052648544312
 Step 660: loss 2.1231141090393066
 Step 670: loss 1.9887412786483765
 Step 680: loss 2.004222869873047
 Step 690: loss 1.904410719871521
 Step 700: loss 1.9501512050628662
 Step 710: loss 1.829141616821289
 Step 720: loss 1.9046710729599
 Step 730: loss 1.9796570539474487
 Step 740: loss 1.9905189275741577
 Step 750: loss 1.857292890548706
 Step 760: loss 1.9265735149383545
 Step 770: loss 2.0551228523254395
 Step 780: loss 1.9494632482528687
 Step 790: loss 2.01499080657959
 Step 800: loss 1.9627217054367065
 [Seed 3] Step 800 --> val loss 2.2078094482421875
 Step 810: loss 1.9758723974227905
 Step 820: loss 1.9868741035461426
 Step 830: loss 1.9707496166229248
 Step 840: loss 1.9320554733276367
 Step 850: loss 2.0579395294189453
 Step 860: loss 1.954392433166504
 Step 870: loss 1.941277027130127
 Step 880: loss 1.9687535762786865
 Step 890: loss 1.977632999420166
 Step 900: loss 1.9999237060546875
 Step 910: loss 1.9213752746582031
 Step 920: loss 1.9220504760742188
 Step 930: loss 1.9838297367095947
 Step 940: loss 1.9516322612762451
 Step 950: loss 2.061034679412842
 Step 960: loss 2.0431737899780273
 Step 970: loss 2.023865222930908
 Step 980: loss 2.112119674682617
 Step 990: loss 2.1063108444213867
 Step 1000: loss 1.9069840908050537
 [Seed 3] Step 1000 --> val loss 2.230445623397827
 Step 1010: loss 1.9974726438522339
 Step 1020: loss 1.9315745830535889
 Step 1030: loss 2.001075506210327
 Step 1040: loss 1.948975920677185
 Step 1050: loss 1.9472782611846924
 Step 1060: loss 2.07204270362854
 Step 1070: loss 2.0223307609558105
 Step 1080: loss 1.9630069732666016
 Step 1090: loss 1.7994699478149414
 Step 1100: loss 2.047883987426758
 Step 1110: loss 1.9462069272994995
 Step 1120: loss 1.9122486114501953
 Step 1130: loss 1.8926854133605957
 Step 1140: loss 1.9427225589752197
 Step 1150: loss 1.9576985836029053
 Step 1160: loss 2.1639537811279297
 Step 1170: loss 2.00484037399292
 Step 1180: loss 1.9519336223602295
 Step 1190: loss 1.8491290807724
 Step 1200: loss 1.9043452739715576
 [Seed 3] Step 1200 --> val loss 2.194591522216797
 Step 1210: loss 2.016596794128418
 Step 1220: loss 1.90972900390625
 Step 1230: loss 1.8726823329925537
 Step 1240: loss 2.0271317958831787
 Step 1250: loss 1.9320329427719116
 Step 1260: loss 1.8633918762207031
 Step 1270: loss 1.965524435043335
 Step 1280: loss 1.9521129131317139
 Step 1290: loss 1.9498295783996582
 Step 1300: loss 1.8896517753601074
 Step 1310: loss 1.9686073064804077
 Step 1320: loss 1.836830496788025
 Step 1330: loss 1.8492496013641357
 Step 1340: loss 1.9424629211425781
 Step 1350: loss 1.8184819221496582
 Step 1360: loss 1.8599165678024292
 Step 1370: loss 1.8230780363082886
 Step 1380: loss 1.8641149997711182
 Step 1390: loss 1.7527402639389038
 Step 1400: loss 1.7246580123901367
 [Seed 3] Step 1400 --> val loss 2.031935930252075
 Step 1410: loss 1.893937587738037
 Step 1420: loss 1.8578943014144897
 Step 1430: loss 1.8110108375549316
 Step 1440: loss 1.7391293048858643
 Step 1450: loss 1.7056596279144287
 Step 1460: loss 1.8245494365692139
 Step 1470: loss 1.7587436437606812
 Step 1480: loss 1.8852262496948242
 Step 1490: loss 1.8300026655197144
 Step 1500: loss 1.7430752515792847
 Step 1510: loss 1.8831992149353027
 Step 1520: loss 1.8588497638702393
 Step 1530: loss 2.0473413467407227
 Step 1540: loss 1.7298835515975952
 Step 1550: loss 1.8473436832427979
 Step 1560: loss 1.8234443664550781
 Step 1570: loss 1.8468917608261108
 Step 1580: loss 1.7154955863952637
 Step 1590: loss 1.7151906490325928
 Step 1600: loss 1.8146507740020752
 [Seed 3] Step 1600 --> val loss 2.014209508895874
 Step 1610: loss 1.9051380157470703
 Step 1620: loss 1.7674086093902588
 Step 1630: loss 1.8027944564819336
 Step 1640: loss 1.8043572902679443
 Step 1650: loss 1.7269432544708252
 Step 1660: loss 1.8302075862884521
 Step 1670: loss 1.7832483053207397
 Step 1680: loss 1.6253527402877808
 Step 1690: loss 1.786935567855835
 Step 1700: loss 1.7622255086898804
 Step 1710: loss 1.8909270763397217
 Step 1720: loss 1.7824252843856812
 Step 1730: loss 1.713087558746338
 Step 1740: loss 1.7741063833236694
 Step 1750: loss 1.8329671621322632
 Step 1760: loss 1.8499120473861694
 Step 1770: loss 1.7636923789978027
 Step 1780: loss 1.8064689636230469
 Step 1790: loss 1.804717779159546
 Step 1800: loss 1.8231157064437866
 [Seed 3] Step 1800 --> val loss 1.9122122526168823
 Step 1810: loss 1.695632815361023
 Step 1820: loss 1.728737235069275
 Step 1830: loss 1.8420547246932983
 Step 1840: loss 1.6898306608200073
 Step 1850: loss 1.8227603435516357
 Step 1860: loss 1.7113527059555054
 Step 1870: loss 1.6875828504562378
 Step 1880: loss 1.8148744106292725
 Step 1890: loss 1.7927231788635254
 Step 1900: loss 1.7250703573226929
 Step 1910: loss 1.7991307973861694
 Step 1920: loss 1.7605855464935303
 Step 1930: loss 1.6439354419708252
 Step 1940: loss 1.6593127250671387
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 3] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [10  0 19 53 42  1 45 47 60 43] ...
 First target sequence: [ 0 19 53 42  1 45 47 60 43  1] ...

 Decoded input: :
 God give us leisure for these rites of love!
 Once more, adieu:

 Decoded target: 
 God give us leisure for these rites of love!
 Once more, adieu: 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.202167510986328
 [Seed 4] Step 0 --> val loss 4.185672283172607
 Step 10: loss 3.8531062602996826
 Step 20: loss 3.3388633728027344
 Step 30: loss 2.8054022789001465
 Step 40: loss 2.6440272331237793
 Step 50: loss 2.618485450744629
 Step 60: loss 2.4942564964294434
 Step 70: loss 2.438544273376465
 Step 80: loss 2.2593867778778076
 Step 90: loss 2.255621910095215
 Step 100: loss 2.2998483180999756
 Step 110: loss 2.129258632659912
 Step 120: loss 2.2236685752868652
 Step 130: loss 2.263552665710449
 Step 140: loss 2.1038522720336914
 Step 150: loss 2.1361594200134277
 Step 160: loss 2.1759848594665527
 Step 170: loss 2.1502718925476074
 Step 180: loss 2.2166600227355957
 Step 190: loss 2.1339187622070312
 Step 200: loss 2.12185001373291
 [Seed 4] Step 200 --> val loss 2.3602399826049805
 Step 210: loss 2.076235055923462
 Step 220: loss 2.059577465057373
 Step 230: loss 2.1455295085906982
 Step 240: loss 2.2047953605651855
 Step 250: loss 2.1342649459838867
 Step 260: loss 2.1015625
 Step 270: loss 2.0905303955078125
 Step 280: loss 2.026589870452881
 Step 290: loss 1.9812266826629639
 Step 300: loss 1.9987443685531616
 Step 310: loss 2.1231164932250977
 Step 320: loss 2.0178258419036865
 Step 330: loss 2.0345959663391113
 Step 340: loss 2.017317533493042
 Step 350: loss 2.0040645599365234
 Step 360: loss 2.063523769378662
 Step 370: loss 1.984802007675171
 Step 380: loss 2.0270373821258545
 Step 390: loss 1.9953792095184326
 Step 400: loss 1.9284570217132568
 [Seed 4] Step 400 --> val loss 2.2281932830810547
 Step 410: loss 1.9148459434509277
 Step 420: loss 2.0817172527313232
 Step 430: loss 1.9623174667358398
 Step 440: loss 1.8905280828475952
 Step 450: loss 1.99958074092865
 Step 460: loss 2.028203010559082
 Step 470: loss 1.9110190868377686
 Step 480: loss 1.9719022512435913
 Step 490: loss 1.9957069158554077
 Step 500: loss 1.996098518371582
 Step 510: loss 1.9119722843170166
 Step 520: loss 1.9563872814178467
 Step 530: loss 2.0084190368652344
 Step 540: loss 1.8879797458648682
 Step 550: loss 2.0279970169067383
 Step 560: loss 1.9183642864227295
 Step 570: loss 2.060105323791504
 Step 580: loss 1.9790799617767334
 Step 590: loss 1.9102152585983276
 Step 600: loss 1.995971441268921
 [Seed 4] Step 600 --> val loss 2.2322957515716553
 Step 610: loss 1.9939277172088623
 Step 620: loss 2.049938678741455
 Step 630: loss 1.9905508756637573
 Step 640: loss 2.0644636154174805
 Step 650: loss 2.021047592163086
 Step 660: loss 1.9422714710235596
 Step 670: loss 2.0361928939819336
 Step 680: loss 2.014232635498047
 Step 690: loss 1.9503471851348877
 Step 700: loss 1.9252612590789795
 Step 710: loss 2.135484457015991
 Step 720: loss 2.0381429195404053
 Step 730: loss 1.9495525360107422
 Step 740: loss 2.026315212249756
 Step 750: loss 1.9704316854476929
 Step 760: loss 2.1663262844085693
 Step 770: loss 1.9954547882080078
 Step 780: loss 1.957995891571045
 Step 790: loss 2.052272081375122
 Step 800: loss 2.104290723800659
 [Seed 4] Step 800 --> val loss 2.269341230392456
 Step 810: loss 2.0315780639648438
 Step 820: loss 1.9967288970947266
 Step 830: loss 1.9998692274093628
 Step 840: loss 1.947737693786621
 Step 850: loss 2.012754201889038
 Step 860: loss 2.0707831382751465
 Step 870: loss 1.8885492086410522
 Step 880: loss 1.9989516735076904
 Step 890: loss 1.9439451694488525
 Step 900: loss 1.993222713470459
 Step 910: loss 1.9541139602661133
 Step 920: loss 1.9769190549850464
 Step 930: loss 1.9813594818115234
 Step 940: loss 2.066847801208496
 Step 950: loss 2.140408515930176
 Step 960: loss 1.9831531047821045
 Step 970: loss 2.0478920936584473
 Step 980: loss 1.8839662075042725
 Step 990: loss 2.014914035797119
 Step 1000: loss 2.057666778564453
 [Seed 4] Step 1000 --> val loss 2.1825428009033203
 Step 1010: loss 2.0373597145080566
 Step 1020: loss 1.9084761142730713
 Step 1030: loss 1.9551293849945068
 Step 1040: loss 1.969374418258667
 Step 1050: loss 1.9784878492355347
 Step 1060: loss 1.9416322708129883
 Step 1070: loss 1.8736872673034668
 Step 1080: loss 1.9363269805908203
 Step 1090: loss 1.9470914602279663
 Step 1100: loss 2.057950019836426
 Step 1110: loss 1.9326527118682861
 Step 1120: loss 1.8457945585250854
 Step 1130: loss 1.8506782054901123
 Step 1140: loss 1.9144916534423828
 Step 1150: loss 1.9858818054199219
 Step 1160: loss 1.9427436590194702
 Step 1170: loss 1.8721282482147217
 Step 1180: loss 1.9807651042938232
 Step 1190: loss 2.0462164878845215
 Step 1200: loss 1.8744659423828125
 [Seed 4] Step 1200 --> val loss 2.0637998580932617
 Step 1210: loss 1.837240219116211
 Step 1220: loss 1.9511642456054688
 Step 1230: loss 1.810455322265625
 Step 1240: loss 2.027956962585449
 Step 1250: loss 2.004269599914551
 Step 1260: loss 1.8429713249206543
 Step 1270: loss 1.8789781332015991
 Step 1280: loss 1.8706430196762085
 Step 1290: loss 1.885748267173767
 Step 1300: loss 1.9162569046020508
 Step 1310: loss 1.8311446905136108
 Step 1320: loss 1.8866164684295654
 Step 1330: loss 2.07918119430542
 Step 1340: loss 1.9705146551132202
 Step 1350: loss 1.9358577728271484
 Step 1360: loss 1.8686717748641968
 Step 1370: loss 1.7660210132598877
 Step 1380: loss 1.8945348262786865
 Step 1390: loss 1.6897988319396973
 Step 1400: loss 1.802520513534546
 [Seed 4] Step 1400 --> val loss 1.9548701047897339
 Step 1410: loss 1.8223432302474976
 Step 1420: loss 1.8103959560394287
 Step 1430: loss 1.887465238571167
 Step 1440: loss 1.988627314567566
 Step 1450: loss 1.889413833618164
 Step 1460: loss 1.7334870100021362
 Step 1470: loss 1.9050801992416382
 Step 1480: loss 1.817534327507019
 Step 1490: loss 1.847699522972107
 Step 1500: loss 1.8575432300567627
 Step 1510: loss 1.8892741203308105
 Step 1520: loss 1.8028467893600464
 Step 1530: loss 1.7798445224761963
 Step 1540: loss 1.8314989805221558
 Step 1550: loss 1.8099088668823242
 Step 1560: loss 1.812544822692871
 Step 1570: loss 1.7554433345794678
 Step 1580: loss 1.7342793941497803
 Step 1590: loss 1.8247125148773193
 Step 1600: loss 1.831568717956543
 [Seed 4] Step 1600 --> val loss 1.8797575235366821
 Step 1610: loss 1.84549081325531
 Step 1620: loss 1.7452195882797241
 Step 1630: loss 1.795467734336853
 Step 1640: loss 1.7549769878387451
 Step 1650: loss 1.738728404045105
 Step 1660: loss 1.6932271718978882
 Step 1670: loss 1.6542203426361084
 Step 1680: loss 1.7764772176742554
 Step 1690: loss 1.804469108581543
 Step 1700: loss 1.8100742101669312
 Step 1710: loss 1.6701076030731201
 Step 1720: loss 1.7525107860565186
 Step 1730: loss 1.6933069229125977
 Step 1740: loss 1.7702406644821167
 Step 1750: loss 1.8718903064727783
 Step 1760: loss 1.732094168663025
 Step 1770: loss 1.6957182884216309
 Step 1780: loss 1.6600284576416016
 Step 1790: loss 1.7380950450897217
 Step 1800: loss 1.6718305349349976
 [Seed 4] Step 1800 --> val loss 1.8531652688980103
 Step 1810: loss 1.6747958660125732
 Step 1820: loss 1.7465366125106812
 Step 1830: loss 1.8634165525436401
 Step 1840: loss 1.7557653188705444
 Step 1850: loss 1.7710331678390503
 Step 1860: loss 1.750815749168396
 Step 1870: loss 1.7384369373321533
 Step 1880: loss 1.7960050106048584
 Step 1890: loss 1.670506238937378
 Step 1900: loss 1.7017114162445068
 Step 1910: loss 1.6979529857635498
 Step 1920: loss 1.6702851057052612
 Step 1930: loss 1.6361901760101318
 Step 1940: loss 1.7666029930114746
 Step 1950: loss 1.6558995246887207
 Step 1960: loss 1.7452753782272339
 Step 1970: loss 1.710590124130249
 Step 1980: loss 1.7595856189727783
 Step 1990: loss 1.5818357467651367
 Step 2000: loss 1.7106016874313354
 [Seed 4] Step 2000 --> val loss 1.7894586324691772
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [52 57 47 53 52  1 53 44  1 39] ...
 First target sequence: [57 47 53 52  1 53 44  1 39  1] ...

 Decoded input: nsion of a love,
 But not possess'd it, and, though I am sold,
 No

 Decoded target: sion of a love,
 But not possess'd it, and, though I am sold,
 Not
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.198877334594727
 [Seed 5] Step 0 --> val loss 4.183917045593262
 Step 10: loss 3.864713430404663
 Step 20: loss 3.342458724975586
 Step 30: loss 2.919038772583008
 Step 40: loss 2.733936309814453
 Step 50: loss 2.5462899208068848
 Step 60: loss 2.5056352615356445
 Step 70: loss 2.43461012840271
 Step 80: loss 2.3215460777282715
 Step 90: loss 2.2746479511260986
 Step 100: loss 2.340437889099121
 Step 110: loss 2.264787435531616
 Step 120: loss 2.141645908355713
 Step 130: loss 2.236654281616211
 Step 140: loss 2.1958367824554443
 Step 150: loss 2.089585781097412
 Step 160: loss 2.1390419006347656
 Step 170: loss 2.281266689300537
 Step 180: loss 2.2439157962799072
 Step 190: loss 2.0188097953796387
 Step 200: loss 1.934068202972412
 [Seed 5] Step 200 --> val loss 2.279865264892578
 Step 210: loss 2.0789742469787598
 Step 220: loss 2.181299924850464
 Step 230: loss 2.1440176963806152
 Step 240: loss 2.055065631866455
 Step 250: loss 2.0119526386260986
 Step 260: loss 2.120068073272705
 Step 270: loss 1.9897353649139404
 Step 280: loss 2.0072431564331055
 Step 290: loss 2.020625591278076
 Step 300: loss 1.9208482503890991
 Step 310: loss 2.029629945755005
 Step 320: loss 2.0411183834075928
 Step 330: loss 2.0592257976531982
 Step 340: loss 1.9805381298065186
 Step 350: loss 2.002894639968872
 Step 360: loss 2.0238003730773926
 Step 370: loss 2.0487871170043945
 Step 380: loss 2.018686294555664
 Step 390: loss 2.0669949054718018
 Step 400: loss 2.0778133869171143
 [Seed 5] Step 400 --> val loss 2.2577035427093506
 Step 410: loss 2.1513311862945557
 Step 420: loss 2.0700254440307617
 Step 430: loss 2.106337070465088
 Step 440: loss 1.9092364311218262
 Step 450: loss 2.106067657470703
 Step 460: loss 1.9309821128845215
 Step 470: loss 1.9824888706207275
 Step 480: loss 2.0378761291503906
 Step 490: loss 2.0051679611206055
 Step 500: loss 1.9199934005737305
 Step 510: loss 2.0057859420776367
 Step 520: loss 2.1058671474456787
 Step 530: loss 1.9450316429138184
 Step 540: loss 1.9796346426010132
 Step 550: loss 2.0548038482666016
 Step 560: loss 1.936793327331543
 Step 570: loss 1.9518283605575562
 Step 580: loss 2.058985948562622
 Step 590: loss 1.9753589630126953
 Step 600: loss 2.000500202178955
 [Seed 5] Step 600 --> val loss 2.144747734069824
 Step 610: loss 2.0463085174560547
 Step 620: loss 1.8817371129989624
 Step 630: loss 2.116060733795166
 Step 640: loss 2.021069288253784
 Step 650: loss 1.9266092777252197
 Step 660: loss 1.9023455381393433
 Step 670: loss 2.075969696044922
 Step 680: loss 2.069591999053955
 Step 690: loss 1.9153159856796265
 Step 700: loss 2.0181374549865723
 Step 710: loss 2.0031373500823975
 Step 720: loss 1.965179443359375
 Step 730: loss 1.9761388301849365
 Step 740: loss 1.9046519994735718
 Step 750: loss 1.9938840866088867
 Step 760: loss 2.056434392929077
 Step 770: loss 2.073180675506592
 Step 780: loss 2.0421440601348877
 Step 790: loss 2.006056547164917
 Step 800: loss 2.016784191131592
 [Seed 5] Step 800 --> val loss 2.2797091007232666
 Step 810: loss 2.043782949447632
 Step 820: loss 1.9565930366516113
 Step 830: loss 2.0212643146514893
 Step 840: loss 2.0131545066833496
 Step 850: loss 2.060360908508301
 Step 860: loss 1.977602481842041
 Step 870: loss 1.9600744247436523
 Step 880: loss 2.0200047492980957
 Step 890: loss 1.9743531942367554
 Step 900: loss 2.006052017211914
 Step 910: loss 1.8830318450927734
 Step 920: loss 2.01796293258667
 Step 930: loss 1.9201778173446655
 Step 940: loss 2.0056426525115967
 Step 950: loss 2.0162932872772217
 Step 960: loss 1.929062843322754
 Step 970: loss 2.0408732891082764
 Step 980: loss 2.059124708175659
 Step 990: loss 2.0472545623779297
 Step 1000: loss 2.1293869018554688
 [Seed 5] Step 1000 --> val loss 2.2658536434173584
 Step 1010: loss 2.0659608840942383
 Step 1020: loss 1.9850910902023315
 Step 1030: loss 2.000018835067749
 Step 1040: loss 2.0035324096679688
 Step 1050: loss 2.0935134887695312
 Step 1060: loss 2.0319557189941406
 Step 1070: loss 2.104645013809204
 Step 1080: loss 2.0392770767211914
 Step 1090: loss 2.125389814376831
 Step 1100: loss 2.024672031402588
 Step 1110: loss 2.0161256790161133
 Step 1120: loss 2.0454771518707275
 Step 1130: loss 2.104135513305664
 Step 1140: loss 2.0864264965057373
 Step 1150: loss 2.0700371265411377
 Step 1160: loss 1.9746497869491577
 Step 1170: loss 2.0342042446136475
 Step 1180: loss 1.9116634130477905
 Step 1190: loss 2.027416229248047
 Step 1200: loss 1.99757981300354
 [Seed 5] Step 1200 --> val loss 2.19832444190979
 Step 1210: loss 2.090369701385498
 Step 1220: loss 2.020872116088867
 Step 1230: loss 1.9530410766601562
 Step 1240: loss 1.9844567775726318
 Step 1250: loss 2.0949547290802
 Step 1260: loss 1.975456953048706
 Step 1270: loss 1.7969768047332764
 Step 1280: loss 2.0441694259643555
 Step 1290: loss 1.9016401767730713
 Step 1300: loss 2.1154050827026367
 Step 1310: loss 2.150202512741089
 Step 1320: loss 2.0613017082214355
 Step 1330: loss 1.9902175664901733
 Step 1340: loss 1.9798953533172607
 Step 1350: loss 1.9306106567382812
 Step 1360: loss 1.9132170677185059
 Step 1370: loss 1.944575309753418
 Step 1380: loss 2.059539318084717
 Step 1390: loss 2.0206117630004883
 Step 1400: loss 1.954216718673706
 [Seed 5] Step 1400 --> val loss 2.164416790008545
 Step 1410: loss 1.8030900955200195
 Step 1420: loss 1.939283847808838
 Step 1430: loss 2.0309019088745117
 Step 1440: loss 1.9293946027755737
 Step 1450: loss 1.9751266241073608
 Step 1460: loss 2.019662857055664
 Step 1470: loss 2.099583387374878
 Step 1480: loss 1.8780686855316162
 Step 1490: loss 1.9771318435668945
 Step 1500: loss 1.8722548484802246
 Step 1510: loss 1.8694865703582764
 Step 1520: loss 2.006753444671631
 Step 1530: loss 1.934746503829956
 Step 1540: loss 2.140066146850586
 Step 1550: loss 1.802417516708374
 Step 1560: loss 1.7997462749481201
 Step 1570: loss 1.8209149837493896
 Step 1580: loss 1.8926055431365967
 Step 1590: loss 1.9016616344451904
 Step 1600: loss 1.96072518825531
 [Seed 5] Step 1600 --> val loss 2.088569164276123
 Step 1610: loss 1.835542917251587
 Step 1620: loss 1.8476684093475342
 Step 1630: loss 1.9056737422943115
 Step 1640: loss 1.8720238208770752
 Step 1650: loss 1.8707021474838257
 Step 1660: loss 1.8935455083847046
 Step 1670: loss 1.878918170928955
 Step 1680: loss 1.9473947286605835
 Step 1690: loss 1.8722699880599976
 Step 1700: loss 1.848625898361206
 Step 1710: loss 1.893388032913208
 Step 1720: loss 1.9986419677734375
 Step 1730: loss 1.7982747554779053
 Step 1740: loss 1.7454127073287964
 Step 1750: loss 1.870363712310791
 Step 1760: loss 1.8335692882537842
 Step 1770: loss 1.9024585485458374
 Step 1780: loss 1.8670824766159058
 Step 1790: loss 1.841869592666626
 Step 1800: loss 1.8130149841308594
 [Seed 5] Step 1800 --> val loss 1.9707802534103394
 Step 1810: loss 1.9035089015960693
 Step 1820: loss 1.7715896368026733
 Step 1830: loss 1.7695887088775635
 Step 1840: loss 1.8387353420257568
 Step 1850: loss 1.8747849464416504
 Step 1860: loss 1.8248927593231201
 Step 1870: loss 1.8063873052597046
 Step 1880: loss 1.8160324096679688
 Step 1890: loss 1.7394359111785889
 Step 1900: loss 1.812598705291748
 Step 1910: loss 1.7679507732391357
 Step 1920: loss 1.8305482864379883
 Step 1930: loss 1.7293990850448608
 Step 1940: loss 1.801026701927185
 Step 1950: loss 1.8373613357543945
 Step 1960: loss 1.8547484874725342
 Step 1970: loss 1.7354156970977783
 Step 1980: loss 1.8581218719482422
 Step 1990: loss 1.7773780822753906
 Step 2000: loss 1.7824043035507202
 [Seed 5] Step 2000 --> val loss 1.9177284240722656
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 0 14 43 39 59 58 63  1 58 53] ...
 First target sequence: [14 43 39 59 58 63  1 58 53 53] ...

 Decoded input: 
 Beauty too rich for use, for earth too dear!
 So shows a snowy d

 Decoded target: Beauty too rich for use, for earth too dear!
 So shows a snowy do
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.212787628173828
 [Seed 6] Step 0 --> val loss 4.1881608963012695
 Step 10: loss 3.849989891052246
 Step 20: loss 3.3110759258270264
 Step 30: loss 2.8072524070739746
 Step 40: loss 2.6321287155151367
 Step 50: loss 2.5587034225463867
 Step 60: loss 2.5438342094421387
 Step 70: loss 2.402848482131958
 Step 80: loss 2.4025425910949707
 Step 90: loss 2.259524345397949
 Step 100: loss 2.2978949546813965
 Step 110: loss 2.308969497680664
 Step 120: loss 2.2168660163879395
 Step 130: loss 2.265108346939087
 Step 140: loss 2.1519312858581543
 Step 150: loss 2.2291312217712402
 Step 160: loss 2.1492204666137695
 Step 170: loss 2.1713500022888184
 Step 180: loss 2.076155185699463
 Step 190: loss 2.1287002563476562
 Step 200: loss 2.1994922161102295
 [Seed 6] Step 200 --> val loss 2.3202664852142334
 Step 210: loss 2.0787558555603027
 Step 220: loss 2.078876256942749
 Step 230: loss 2.2104194164276123
 Step 240: loss 2.0316245555877686
 Step 250: loss 2.018238067626953
 Step 260: loss 2.1233370304107666
 Step 270: loss 1.9789562225341797
 Step 280: loss 1.9975457191467285
 Step 290: loss 1.978270411491394
 Step 300: loss 2.0666565895080566
 Step 310: loss 2.03885555267334
 Step 320: loss 2.1604325771331787
 Step 330: loss 2.0190839767456055
 Step 340: loss 2.023728370666504
 Step 350: loss 1.9892609119415283
 Step 360: loss 2.039414882659912
 Step 370: loss 2.010740280151367
 Step 380: loss 2.0457935333251953
 Step 390: loss 1.942130446434021
 Step 400: loss 2.0591158866882324
 [Seed 6] Step 400 --> val loss 2.1902053356170654
 Step 410: loss 2.175410032272339
 Step 420: loss 1.8986296653747559
 Step 430: loss 2.0095252990722656
 Step 440: loss 2.0119261741638184
 Step 450: loss 2.0845625400543213
 Step 460: loss 1.9353265762329102
 Step 470: loss 1.9987106323242188
 Step 480: loss 1.9100313186645508
 Step 490: loss 1.9924882650375366
 Step 500: loss 1.9794442653656006
 Step 510: loss 1.9074162244796753
 Step 520: loss 1.8472554683685303
 Step 530: loss 1.9527671337127686
 Step 540: loss 2.0512681007385254
 Step 550: loss 2.0378198623657227
 Step 560: loss 1.9899390935897827
 Step 570: loss 2.0481796264648438
 Step 580: loss 2.0049214363098145
 Step 590: loss 1.9933885335922241
 Step 600: loss 2.060115337371826
 [Seed 6] Step 600 --> val loss 2.2592051029205322
 Step 610: loss 2.0148754119873047
 Step 620: loss 2.027765989303589
 Step 630: loss 1.928768515586853
 Step 640: loss 1.9761016368865967
 Step 650: loss 1.9437726736068726
 Step 660: loss 2.0439538955688477
 Step 670: loss 1.9901199340820312
 Step 680: loss 2.044187545776367
 Step 690: loss 2.0219972133636475
 Step 700: loss 1.982324481010437
 Step 710: loss 1.8973586559295654
 Step 720: loss 1.9992790222167969
 Step 730: loss 1.9679855108261108
 Step 740: loss 2.1079561710357666
 Step 750: loss 2.019334554672241
 Step 760: loss 1.9725650548934937
 Step 770: loss 1.93539559841156
 Step 780: loss 2.0346009731292725
 Step 790: loss 1.875098705291748
 Step 800: loss 1.9717057943344116
 [Seed 6] Step 800 --> val loss 2.1665899753570557
 Step 810: loss 2.0173487663269043
 Step 820: loss 1.9111576080322266
 Step 830: loss 2.1047534942626953
 Step 840: loss 1.9726685285568237
 Step 850: loss 1.86997389793396
 Step 860: loss 1.9912751913070679
 Step 870: loss 1.9307143688201904
 Step 880: loss 1.9281880855560303
 Step 890: loss 2.070739269256592
 Step 900: loss 1.9344418048858643
 Step 910: loss 1.9423255920410156
 Step 920: loss 1.9452896118164062
 Step 930: loss 1.8849735260009766
 Step 940: loss 2.062591791152954
 Step 950: loss 1.9102715253829956
 Step 960: loss 1.8893201351165771
 Step 970: loss 1.8547248840332031
 Step 980: loss 1.9634511470794678
 Step 990: loss 1.9571436643600464
 Step 1000: loss 1.8440135717391968
 [Seed 6] Step 1000 --> val loss 2.1169145107269287
 Step 1010: loss 2.0525660514831543
 Step 1020: loss 1.8672409057617188
 Step 1030: loss 1.9750975370407104
 Step 1040: loss 1.7950921058654785
 Step 1050: loss 1.9161351919174194
 Step 1060: loss 2.036332130432129
 Step 1070: loss 2.0108165740966797
 Step 1080: loss 1.9747123718261719
 Step 1090: loss 1.972215175628662
 Step 1100: loss 1.8376964330673218
 Step 1110: loss 1.9622688293457031
 Step 1120: loss 1.8874706029891968
 Step 1130: loss 1.9709477424621582
 Step 1140: loss 1.893045425415039
 Step 1150: loss 1.9504927396774292
 Step 1160: loss 1.7744334936141968
 Step 1170: loss 1.9399722814559937
 Step 1180: loss 1.87971830368042
 Step 1190: loss 1.8840112686157227
 Step 1200: loss 1.792819857597351
 [Seed 6] Step 1200 --> val loss 2.1421048641204834
 Step 1210: loss 1.8558257818222046
 Step 1220: loss 1.8763822317123413
 Step 1230: loss 1.849543571472168
 Step 1240: loss 1.8613290786743164
 Step 1250: loss 1.8653266429901123
 Step 1260: loss 1.8199782371520996
 Step 1270: loss 1.7991769313812256
 Step 1280: loss 1.9651988744735718
 Step 1290: loss 1.8999791145324707
 Step 1300: loss 1.8763954639434814
 Step 1310: loss 1.8296115398406982
 Step 1320: loss 1.9708795547485352
 Step 1330: loss 1.9223588705062866
 Step 1340: loss 1.9332422018051147
 Step 1350: loss 1.86199951171875
 Step 1360: loss 1.866492748260498
 Step 1370: loss 2.039170980453491
 Step 1380: loss 1.890007734298706
 Step 1390: loss 1.715017318725586
 Step 1400: loss 1.899245262145996
 [Seed 6] Step 1400 --> val loss 2.1020278930664062
 Step 1410: loss 1.974665880203247
 Step 1420: loss 1.8142284154891968
 Step 1430: loss 1.9168241024017334
 Step 1440: loss 1.927422285079956
 Step 1450: loss 1.776771903038025
 Step 1460: loss 1.7676528692245483
 Step 1470: loss 1.8388128280639648
 Step 1480: loss 1.832554817199707
 Step 1490: loss 1.94746994972229
 Step 1500: loss 1.863999843597412
 Step 1510: loss 1.7210197448730469
 Step 1520: loss 1.9037660360336304
 Step 1530: loss 1.8867297172546387
 Step 1540: loss 1.7920942306518555
 Step 1550: loss 1.8085006475448608
 Step 1560: loss 1.8690316677093506
 Step 1570: loss 1.869110107421875
 Step 1580: loss 1.7065274715423584
 Step 1590: loss 1.8166778087615967
 Step 1600: loss 1.838450312614441
 [Seed 6] Step 1600 --> val loss 2.042332410812378
 Step 1610: loss 1.8109334707260132
 Step 1620: loss 1.892442226409912
 Step 1630: loss 1.6977304220199585
 Step 1640: loss 1.7882020473480225
 Step 1650: loss 1.7806591987609863
 Step 1660: loss 1.8426690101623535
 Step 1670: loss 1.834164023399353
 Step 1680: loss 1.8023784160614014
 Step 1690: loss 1.769823431968689
 Step 1700: loss 1.860500693321228
 Step 1710: loss 1.8319276571273804
 Step 1720: loss 1.8196057081222534
 Step 1730: loss 1.7496976852416992
 Step 1740: loss 1.7592157125473022
 Step 1750: loss 1.8270248174667358
 Step 1760: loss 1.689424753189087
 Step 1770: loss 1.779390573501587
 Step 1780: loss 1.7803287506103516
 Step 1790: loss 1.6976650953292847
 Step 1800: loss 1.7645089626312256
 [Seed 6] Step 1800 --> val loss 1.9077049493789673
 Step 1810: loss 1.6489429473876953
 Step 1820: loss 1.6444165706634521
 Step 1830: loss 1.7590172290802002
 Step 1840: loss 1.741841197013855
 Step 1850: loss 1.7481269836425781
 Step 1860: loss 1.8078629970550537
 Step 1870: loss 1.7060353755950928
 Step 1880: loss 1.7245596647262573
 Step 1890: loss 1.6196755170822144
 Step 1900: loss 1.6640992164611816
 Step 1910: loss 1.6842725276947021
 Step 1920: loss 1.8456172943115234
 Step 1930: loss 1.6958990097045898
 Step 1940: loss 1.6626118421554565
 Step 1950: loss 1.7893388271331787
 Step 1960: loss 1.7477421760559082
 Step 1970: loss 1.731529951095581
 Step 1980: loss 1.6034419536590576
 Step 1990: loss 1.6216545104980469
 Step 2000: loss 1.6997531652450562
 [Seed 6] Step 2000 --> val loss 1.8586244583129883
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [51 59 41 46  1 39 57  1 58 46] ...
 First target sequence: [59 41 46  1 39 57  1 58 46 47] ...

 Decoded input: much as this old man does when the
 business is performed, and re

 Decoded target: uch as this old man does when the
 business is performed, and rem
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2083330154418945
 [Seed 7] Step 0 --> val loss 4.18549919128418
 Step 10: loss 3.86887788772583
 Step 20: loss 3.2981815338134766
 Step 30: loss 2.90889048576355
 Step 40: loss 2.6982245445251465
 Step 50: loss 2.610596179962158
 Step 60: loss 2.607024669647217
 Step 70: loss 2.386598825454712
 Step 80: loss 2.386631965637207
 Step 90: loss 2.368238925933838
 Step 100: loss 2.2945556640625
 Step 110: loss 2.321833610534668
 Step 120: loss 2.3002421855926514
 Step 130: loss 2.1769509315490723
 Step 140: loss 2.2672011852264404
 Step 150: loss 2.125417709350586
 Step 160: loss 2.210278034210205
 Step 170: loss 2.143488883972168
 Step 180: loss 2.1639490127563477
 Step 190: loss 2.130420207977295
 Step 200: loss 2.128840923309326
 [Seed 7] Step 200 --> val loss 2.3950552940368652
 Step 210: loss 1.9907557964324951
 Step 220: loss 2.1212244033813477
 Step 230: loss 2.0414092540740967
 Step 240: loss 2.1200952529907227
 Step 250: loss 2.000436544418335
 Step 260: loss 2.0121288299560547
 Step 270: loss 1.982445478439331
 Step 280: loss 2.1184213161468506
 Step 290: loss 1.942213773727417
 Step 300: loss 1.966287612915039
 Step 310: loss 2.126720428466797
 Step 320: loss 2.064791202545166
 Step 330: loss 2.0350186824798584
 Step 340: loss 1.9634733200073242
 Step 350: loss 2.0706632137298584
 Step 360: loss 2.111452102661133
 Step 370: loss 1.9857025146484375
 Step 380: loss 2.0261402130126953
 Step 390: loss 2.0740184783935547
 Step 400: loss 2.1761510372161865
 [Seed 7] Step 400 --> val loss 2.3167850971221924
 Step 410: loss 1.9366344213485718
 Step 420: loss 2.0006775856018066
 Step 430: loss 1.9890918731689453
 Step 440: loss 2.0205137729644775
 Step 450: loss 2.0575637817382812
 Step 460: loss 2.0558371543884277
 Step 470: loss 2.1156299114227295
 Step 480: loss 2.097391366958618
 Step 490: loss 2.116201162338257
 Step 500: loss 2.140104293823242
 Step 510: loss 2.0510387420654297
 Step 520: loss 2.070624828338623
 Step 530: loss 2.0353734493255615
 Step 540: loss 1.9109580516815186
 Step 550: loss 1.9489004611968994
 Step 560: loss 2.0394437313079834
 Step 570: loss 1.984715223312378
 Step 580: loss 1.9639534950256348
 Step 590: loss 1.8672679662704468
 Step 600: loss 2.018378973007202
 [Seed 7] Step 600 --> val loss 2.2615036964416504
 Step 610: loss 2.0123400688171387
 Step 620: loss 1.9999730587005615
 Step 630: loss 2.04957914352417
 Step 640: loss 2.01399564743042
 Step 650: loss 2.033637285232544
 Step 660: loss 2.06915283203125
 Step 670: loss 2.0080018043518066
 Step 680: loss 2.023557662963867
 Step 690: loss 2.080134630203247
 Step 700: loss 2.102620840072632
 Step 710: loss 2.128671646118164
 Step 720: loss 1.9736379384994507
 Step 730: loss 2.1373417377471924
 Step 740: loss 2.0446107387542725
 Step 750: loss 2.0670976638793945
 Step 760: loss 2.1287732124328613
 Step 770: loss 2.024334669113159
 Step 780: loss 2.140040159225464
 Step 790: loss 2.0435991287231445
 Step 800: loss 2.0428342819213867
 [Seed 7] Step 800 --> val loss 2.2616405487060547
 Step 810: loss 1.9392695426940918
 Step 820: loss 2.012131929397583
 Step 830: loss 1.9211987257003784
 Step 840: loss 2.023632764816284
 Step 850: loss 1.8817996978759766
 Step 860: loss 2.0021636486053467
 Step 870: loss 1.9636383056640625
 Step 880: loss 1.9398349523544312
 Step 890: loss 2.103283643722534
 Step 900: loss 2.046079635620117
 Step 910: loss 2.054950714111328
 Step 920: loss 1.9894789457321167
 Step 930: loss 2.0122201442718506
 Step 940: loss 2.0669097900390625
 Step 950: loss 2.0593528747558594
 Step 960: loss 1.9860632419586182
 Step 970: loss 1.940004825592041
 Step 980: loss 2.004913806915283
 Step 990: loss 2.0924487113952637
 Step 1000: loss 2.030883312225342
 [Seed 7] Step 1000 --> val loss 2.2423062324523926
 Step 1010: loss 2.0446012020111084
 Step 1020: loss 2.0855560302734375
 Step 1030: loss 2.0118870735168457
 Step 1040: loss 2.0118532180786133
 Step 1050: loss 1.881317138671875
 Step 1060: loss 1.9321181774139404
 Step 1070: loss 2.0206732749938965
 Step 1080: loss 2.0662178993225098
 Step 1090: loss 1.9980672597885132
 Step 1100: loss 1.9998300075531006
 Step 1110: loss 1.968878984451294
 Step 1120: loss 1.9411640167236328
 Step 1130: loss 2.015690565109253
 Step 1140: loss 2.007314682006836
 Step 1150: loss 1.9487322568893433
 Step 1160: loss 2.024723768234253
 Step 1170: loss 2.0863749980926514
 Step 1180: loss 2.0271224975585938
 Step 1190: loss 1.987750768661499
 Step 1200: loss 1.8838043212890625
 [Seed 7] Step 1200 --> val loss 2.186222791671753
 Step 1210: loss 2.040618896484375
 Step 1220: loss 1.8944482803344727
 Step 1230: loss 1.878354787826538
 Step 1240: loss 1.8995212316513062
 Step 1250: loss 2.095280885696411
 Step 1260: loss 2.0420804023742676
 Step 1270: loss 1.8466625213623047
 Step 1280: loss 1.9398105144500732
 Step 1290: loss 1.9222339391708374
 Step 1300: loss 1.88290536403656
 Step 1310: loss 1.8130831718444824
 Step 1320: loss 1.9584193229675293
 Step 1330: loss 1.793680191040039
 Step 1340: loss 1.8974013328552246
 Step 1350: loss 1.9529850482940674
 Step 1360: loss 1.8445003032684326
 Step 1370: loss 1.8368518352508545
 Step 1380: loss 1.8723928928375244
 Step 1390: loss 1.9459171295166016
 Step 1400: loss 1.8359451293945312
 [Seed 7] Step 1400 --> val loss 2.1151578426361084
 Step 1410: loss 1.8765554428100586
 Step 1420: loss 1.9342267513275146
 Step 1430: loss 1.8375924825668335
 Step 1440: loss 1.9260661602020264
 Step 1450: loss 1.8550370931625366
 Step 1460: loss 1.9323179721832275
 Step 1470: loss 1.912721037864685
 Step 1480: loss 1.8456017971038818
 Step 1490: loss 1.9754136800765991
 Step 1500: loss 1.8204505443572998
 Step 1510: loss 1.7726796865463257
 Step 1520: loss 1.869518756866455
 Step 1530: loss 1.7358440160751343
 Step 1540: loss 1.7763636112213135
 Step 1550: loss 1.8260879516601562
 Step 1560: loss 1.8414720296859741
 Step 1570: loss 1.9940950870513916
 Step 1580: loss 1.9117162227630615
 Step 1590: loss 1.896811842918396
 Step 1600: loss 1.8386906385421753
 [Seed 7] Step 1600 --> val loss 2.0636889934539795
 Step 1610: loss 1.883345365524292
 Step 1620: loss 1.8492268323898315
 Step 1630: loss 1.8118222951889038
 Step 1640: loss 1.8392181396484375
 Step 1650: loss 1.7315847873687744
 Step 1660: loss 1.7513021230697632
 Step 1670: loss 1.8526300191879272
 Step 1680: loss 1.7820088863372803
 Step 1690: loss 1.9246842861175537
 Step 1700: loss 1.8595384359359741
 Step 1710: loss 1.7925885915756226
 Step 1720: loss 1.688884973526001
 Step 1730: loss 1.8304507732391357
 Step 1740: loss 1.8267898559570312
 Step 1750: loss 1.7970247268676758
 Step 1760: loss 1.775256633758545
 Step 1770: loss 1.7504253387451172
 Step 1780: loss 1.7940491437911987
 Step 1790: loss 1.882997751235962
 Step 1800: loss 1.696380615234375
 [Seed 7] Step 1800 --> val loss 1.9179306030273438
 Step 1810: loss 1.8196371793746948
 Step 1820: loss 1.7744462490081787
 Step 1830: loss 1.7446260452270508
 Step 1840: loss 1.7941453456878662
 Step 1850: loss 1.7225631475448608
 Step 1860: loss 1.7666046619415283
 Step 1870: loss 1.77079176902771
 Step 1880: loss 1.740902304649353
 Step 1890: loss 1.7037687301635742
 Step 1900: loss 1.809409499168396
 Step 1910: loss 1.693804144859314
 Step 1920: loss 1.7244895696640015
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 7] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [57 47 52 41 43  1 58 46 43 56] ...
 First target sequence: [47 52 41 43  1 58 46 43 56 43] ...

 Decoded input: since there was some speech of marriage
 Betwixt myself and her; 

 Decoded target: ince there was some speech of marriage
 Betwixt myself and her; w
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.219320297241211
 [Seed 8] Step 0 --> val loss 4.185484886169434
 Step 10: loss 3.83786940574646
 Step 20: loss 3.3675951957702637
 Step 30: loss 2.8495254516601562
 Step 40: loss 2.6676230430603027
 Step 50: loss 2.5966298580169678
 Step 60: loss 2.5088043212890625
 Step 70: loss 2.4571056365966797
 Step 80: loss 2.3269460201263428
 Step 90: loss 2.348194122314453
 Step 100: loss 2.4128386974334717
 Step 110: loss 2.2446675300598145
 Step 120: loss 2.2149550914764404
 Step 130: loss 2.0804667472839355
 Step 140: loss 2.1857967376708984
 Step 150: loss 2.1693830490112305
 Step 160: loss 2.1278083324432373
 Step 170: loss 2.1987247467041016
 Step 180: loss 2.0942349433898926
 Step 190: loss 2.217916965484619
 Step 200: loss 2.140969753265381
 [Seed 8] Step 200 --> val loss 2.474520444869995
 Step 210: loss 2.132819175720215
 Step 220: loss 2.2514548301696777
 Step 230: loss 2.097288131713867
 Step 240: loss 1.9562993049621582
 Step 250: loss 2.1330602169036865
 Step 260: loss 2.083895444869995
 Step 270: loss 2.0870227813720703
 Step 280: loss 2.0861270427703857
 Step 290: loss 2.0610508918762207
 Step 300: loss 1.944392442703247
 Step 310: loss 1.9443105459213257
 Step 320: loss 2.009932518005371
 Step 330: loss 2.0222251415252686
 Step 340: loss 2.14351224899292
 Step 350: loss 2.027113199234009
 Step 360: loss 2.1049818992614746
 Step 370: loss 1.9681475162506104
 Step 380: loss 1.9620026350021362
 Step 390: loss 2.1406936645507812
 Step 400: loss 1.9415024518966675
 [Seed 8] Step 400 --> val loss 2.1840980052948
 Step 410: loss 2.0490360260009766
 Step 420: loss 1.9690916538238525
 Step 430: loss 2.1013712882995605
 Step 440: loss 1.9379944801330566
 Step 450: loss 1.9371249675750732
 Step 460: loss 1.8793504238128662
 Step 470: loss 2.0885303020477295
 Step 480: loss 2.08162260055542
 Step 490: loss 1.819784164428711
 Step 500: loss 2.031935214996338
 Step 510: loss 2.1309592723846436
 Step 520: loss 2.095057249069214
 Step 530: loss 2.0180230140686035
 Step 540: loss 2.0078656673431396
 Step 550: loss 1.9434423446655273
 Step 560: loss 1.9857871532440186
 Step 570: loss 2.000714063644409
 Step 580: loss 2.1140518188476562
 Step 590: loss 2.0062642097473145
 Step 600: loss 1.9189910888671875
 [Seed 8] Step 600 --> val loss 2.2554187774658203
 Step 610: loss 1.9453926086425781
 Step 620: loss 1.891034483909607
 Step 630: loss 2.000514268875122
 Step 640: loss 1.9116171598434448
 Step 650: loss 1.9284719228744507
 Step 660: loss 2.051795482635498
 Step 670: loss 2.0208473205566406
 Step 680: loss 1.9207932949066162
 Step 690: loss 2.068270683288574
 Step 700: loss 1.8507460355758667
 Step 710: loss 1.9467060565948486
 Step 720: loss 1.9236876964569092
 Step 730: loss 1.995415449142456
 Step 740: loss 1.9086778163909912
 Step 750: loss 2.010545253753662
 Step 760: loss 1.916175127029419
 Step 770: loss 2.031104564666748
 Step 780: loss 1.921683669090271
 Step 790: loss 2.0499958992004395
 Step 800: loss 1.9305322170257568
 [Seed 8] Step 800 --> val loss 2.209599494934082
 Step 810: loss 2.050342082977295
 Step 820: loss 1.9408372640609741
 Step 830: loss 1.885088324546814
 Step 840: loss 1.944449782371521
 Step 850: loss 2.079096794128418
 Step 860: loss 1.9723104238510132
 Step 870: loss 1.9284499883651733
 Step 880: loss 1.8544881343841553
 Step 890: loss 2.0324363708496094
 Step 900: loss 1.9985501766204834
 Step 910: loss 1.926687479019165
 Step 920: loss 1.8658137321472168
 Step 930: loss 1.8581819534301758
 Step 940: loss nan
 Step 950: loss nan
 Step 960: loss nan
 Step 970: loss nan
 Step 980: loss nan
 Step 990: loss nan
 Step 1000: loss nan
 [Seed 8] Step 1000 --> val loss nan
 Step 1010: loss nan
 Step 1020: loss nan
 Step 1030: loss nan
 Step 1040: loss nan
 Step 1050: loss nan
 Step 1060: loss nan
 Step 1070: loss nan
 Step 1080: loss nan
 Step 1090: loss nan
 Step 1100: loss nan
 Step 1110: loss nan
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 8] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 8] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 8] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 8] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 8] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [24 13 10  0 21  1 39 51  1 52] ...
 First target sequence: [13 10  0 21  1 39 51  1 52 53] ...

 Decoded input: LA:
 I am now going to resolve him: I had rather my
 brother die b

 Decoded target: A:
 I am now going to resolve him: I had rather my
 brother die by
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2137274742126465
 [Seed 9] Step 0 --> val loss 4.185962200164795
 Step 10: loss 3.835308074951172
 Step 20: loss 3.320496082305908
 Step 30: loss 2.861642360687256
 Step 40: loss 2.647559404373169
 Step 50: loss 2.534616708755493
 Step 60: loss 2.560427665710449
 Step 70: loss 2.4538755416870117
 Step 80: loss 2.404590606689453
 Step 90: loss 2.362426519393921
 Step 100: loss 2.3570966720581055
 Step 110: loss 2.254436731338501
 Step 120: loss 2.3023509979248047
 Step 130: loss 2.2917613983154297
 Step 140: loss 2.2067956924438477
 Step 150: loss 2.148261070251465
 Step 160: loss 2.2518324851989746
 Step 170: loss 2.1321115493774414
 Step 180: loss 2.1812219619750977
 Step 190: loss 2.125553607940674
 Step 200: loss 1.9622933864593506
 [Seed 9] Step 200 --> val loss 2.308889150619507
 Step 210: loss 2.126617193222046
 Step 220: loss 2.125180959701538
 Step 230: loss 1.9793944358825684
 Step 240: loss 2.0984082221984863
 Step 250: loss 2.061891555786133
 Step 260: loss 2.023954391479492
 Step 270: loss 2.000436544418335
 Step 280: loss 2.0369744300842285
 Step 290: loss 2.034757614135742
 Step 300: loss 2.024995803833008
 Step 310: loss 2.088515281677246
 Step 320: loss 2.053431987762451
 Step 330: loss 2.115814208984375
 Step 340: loss 1.9756765365600586
 Step 350: loss 1.9790663719177246
 Step 360: loss 2.0131540298461914
 Step 370: loss 1.9551280736923218
 Step 380: loss 1.9833288192749023
 Step 390: loss 1.9769048690795898
 Step 400: loss 1.9808425903320312
 [Seed 9] Step 400 --> val loss 2.301460027694702
 Step 410: loss 2.0058248043060303
 Step 420: loss 1.9905614852905273
 Step 430: loss 1.9362552165985107
 Step 440: loss 2.04898738861084
 Step 450: loss 1.918461799621582
 Step 460: loss 1.9956778287887573
 Step 470: loss 1.9663000106811523
 Step 480: loss 1.9548225402832031
 Step 490: loss 1.9629650115966797
 Step 500: loss 2.0045816898345947
 Step 510: loss 2.110255718231201
 Step 520: loss 1.888152837753296
 Step 530: loss 1.9888709783554077
 Step 540: loss 1.9961535930633545
 Step 550: loss 2.0256307125091553
 Step 560: loss 1.8861608505249023
 Step 570: loss 2.113862991333008
 Step 580: loss 2.047541379928589
 Step 590: loss 2.063232898712158
 Step 600: loss 2.0698695182800293
 [Seed 9] Step 600 --> val loss 2.4240243434906006
 Step 610: loss 1.9838597774505615
 Step 620: loss 1.9891765117645264
 Step 630: loss 2.0315001010894775
 Step 640: loss 2.071030616760254
 Step 650: loss 2.052856922149658
 Step 660: loss 1.9669440984725952
 Step 670: loss 2.0275163650512695
 Step 680: loss 1.9675538539886475
 Step 690: loss 1.9995267391204834
 Step 700: loss 2.093799591064453
 Step 710: loss 2.0752573013305664
 Step 720: loss 2.0959372520446777
 Step 730: loss 1.9655702114105225
 Step 740: loss 2.068056583404541
 Step 750: loss 2.030311107635498
 Step 760: loss 1.9549145698547363
 Step 770: loss 1.9642056226730347
 Step 780: loss 1.890539526939392
 Step 790: loss 2.115111827850342
 Step 800: loss 1.9758069515228271
 [Seed 9] Step 800 --> val loss 2.3492777347564697
 Step 810: loss 2.0060009956359863
 Step 820: loss 2.1461451053619385
 Step 830: loss 2.0535035133361816
 Step 840: loss 2.1225531101226807
 Step 850: loss 2.052644968032837
 Step 860: loss 1.8942828178405762
 Step 870: loss 2.098362684249878
 Step 880: loss 2.1481003761291504
 Step 890: loss 1.9952421188354492
 Step 900: loss 2.045112371444702
 Step 910: loss 1.9422845840454102
 Step 920: loss 1.9921882152557373
 Step 930: loss 2.1305465698242188
 Step 940: loss 2.0923361778259277
 Step 950: loss 2.021538257598877
 Step 960: loss 2.1393017768859863
 Step 970: loss 1.9535062313079834
 Step 980: loss 1.9121804237365723
 Step 990: loss 1.9716036319732666
 Step 1000: loss 1.8683490753173828
 [Seed 9] Step 1000 --> val loss 2.3225059509277344
 Step 1010: loss 1.8803108930587769
 Step 1020: loss 2.0452327728271484
 Step 1030: loss 2.020733594894409
 Step 1040: loss 2.060506582260132
 Step 1050: loss 1.8657505512237549
 Step 1060: loss 2.042727470397949
 Step 1070: loss 2.1043248176574707
 Step 1080: loss 1.9983571767807007
 Step 1090: loss 1.9699933528900146
 Step 1100: loss 1.9667205810546875
 Step 1110: loss 1.967648983001709
 Step 1120: loss 2.065981388092041
 Step 1130: loss 1.9944159984588623
 Step 1140: loss 1.8862794637680054
 Step 1150: loss 1.8628009557724
 Step 1160: loss 2.1070947647094727
 Step 1170: loss 1.993842601776123
 Step 1180: loss 1.9853461980819702
 Step 1190: loss 2.0239548683166504
 Step 1200: loss 1.9137167930603027
 [Seed 9] Step 1200 --> val loss 2.2679924964904785
 Step 1210: loss 1.9754365682601929
 Step 1220: loss 2.0048022270202637
 Step 1230: loss 1.966747760772705
 Step 1240: loss 1.955816626548767
 Step 1250: loss 1.923715353012085
 Step 1260: loss 1.8237502574920654
 Step 1270: loss 1.9173338413238525
 Step 1280: loss 1.9548476934432983
 Step 1290: loss 1.9293410778045654
 Step 1300: loss 1.9791157245635986
 Step 1310: loss 1.9273080825805664
 Step 1320: loss 1.9800270795822144
 Step 1330: loss 1.9059374332427979
 Step 1340: loss 1.940953016281128
 Step 1350: loss 1.7764968872070312
 Step 1360: loss 1.783355474472046
 Step 1370: loss 1.8350404500961304
 Step 1380: loss 1.921083688735962
 Step 1390: loss 1.8292831182479858
 Step 1400: loss 1.842041015625
 [Seed 9] Step 1400 --> val loss 2.040717840194702
 Step 1410: loss 1.8610563278198242
 Step 1420: loss 1.877755880355835
 Step 1430: loss 1.8825712203979492
 Step 1440: loss 1.8250014781951904
 Step 1450: loss 1.858306884765625
 Step 1460: loss 1.7679141759872437
 Step 1470: loss 1.9878753423690796
 Step 1480: loss 1.8338468074798584
 Step 1490: loss 1.8789342641830444
 Step 1500: loss 1.8953652381896973
 Step 1510: loss 1.7510818243026733
 Step 1520: loss 1.8229045867919922
 Step 1530: loss 1.7778939008712769
 Step 1540: loss 1.8516347408294678
 Step 1550: loss 1.9546916484832764
 Step 1560: loss 1.9322936534881592
 Step 1570: loss 1.8255685567855835
 Step 1580: loss 1.950388789176941
 Step 1590: loss 1.8859584331512451
 Step 1600: loss 1.793022871017456
 [Seed 9] Step 1600 --> val loss 2.0551695823669434
 Step 1610: loss 1.8800883293151855
 Step 1620: loss 1.756182074546814
 Step 1630: loss 1.732055425643921
 Step 1640: loss 1.7870862483978271
 Step 1650: loss 1.7086000442504883
 Step 1660: loss 1.7728219032287598
 Step 1670: loss 1.7636398077011108
 Step 1680: loss 1.6981675624847412
 Step 1690: loss 1.9302176237106323
 Step 1700: loss 1.7933543920516968
 Step 1710: loss 1.7895123958587646
 Step 1720: loss 1.8067106008529663
 Step 1730: loss 1.7866500616073608
 Step 1740: loss 1.628284215927124
 Step 1750: loss 1.7660735845565796
 Step 1760: loss 1.7442954778671265
 Step 1770: loss 1.755476951599121
 Step 1780: loss 1.71309494972229
 Step 1790: loss 1.9540189504623413
 Step 1800: loss 1.795607328414917
 [Seed 9] Step 1800 --> val loss 1.8603471517562866
 Step 1810: loss 1.6846388578414917
 Step 1820: loss 1.803554892539978
 Step 1830: loss 1.682957410812378
 Step 1840: loss 1.8142297267913818
 Step 1850: loss 1.8515828847885132
 Step 1860: loss 1.7608096599578857
 Step 1870: loss 1.6625522375106812
 Step 1880: loss 1.7792396545410156
 Step 1890: loss 1.8427808284759521
 Step 1900: loss 1.7330842018127441
 Step 1910: loss 1.8666837215423584
 Step 1920: loss 1.692422866821289
 Step 1930: loss 1.729608416557312
 Step 1940: loss 1.7695386409759521
 Step 1950: loss 1.7937387228012085
 Step 1960: loss 1.6685651540756226
 Step 1970: loss 1.7302868366241455
 Step 1980: loss 1.759502649307251
 Step 1990: loss 1.7687599658966064
 Step 2000: loss 1.7391624450683594
 [Seed 9] Step 2000 --> val loss 1.8335838317871094
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [59 57 58 47 41 43  6  1 48 59] ...
 First target sequence: [57 58 47 41 43  6  1 48 59 57] ...

 Decoded input: ustice, justice!

 DUKE VINCENTIO:
 Relate your wrongs; in what? b

 Decoded target: stice, justice!

 DUKE VINCENTIO:
 Relate your wrongs; in what? by
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.206463813781738
 [Seed 10] Step 0 --> val loss 4.183530330657959
 Step 10: loss 3.840056896209717
 Step 20: loss 3.3385162353515625
 Step 30: loss 2.822721004486084
 Step 40: loss 2.664966344833374
 Step 50: loss 2.6242055892944336
 Step 60: loss 2.546250104904175
 Step 70: loss 2.371267795562744
 Step 80: loss 2.3880906105041504
 Step 90: loss 2.392817258834839
 Step 100: loss 2.210587501525879
 Step 110: loss 2.250582695007324
 Step 120: loss 2.226181983947754
 Step 130: loss 2.208202838897705
 Step 140: loss 2.152001142501831
 Step 150: loss 2.2219929695129395
 Step 160: loss 2.241886615753174
 Step 170: loss 2.1442718505859375
 Step 180: loss 2.0826056003570557
 Step 190: loss 2.083804130554199
 Step 200: loss 2.0933995246887207
 [Seed 10] Step 200 --> val loss 2.2898499965667725
 Step 210: loss 2.087355613708496
 Step 220: loss 2.005821704864502
 Step 230: loss 2.034088134765625
 Step 240: loss 2.1540563106536865
 Step 250: loss 2.1471879482269287
 Step 260: loss 2.1329705715179443
 Step 270: loss 2.108008861541748
 Step 280: loss 1.9990339279174805
 Step 290: loss 2.070310354232788
 Step 300: loss 1.9467405080795288
 Step 310: loss 2.0832693576812744
 Step 320: loss 1.9770030975341797
 Step 330: loss 2.1297764778137207
 Step 340: loss 2.130660057067871
 Step 350: loss 2.1001620292663574
 Step 360: loss 2.003575325012207
 Step 370: loss 1.9978641271591187
 Step 380: loss 2.025724411010742
 Step 390: loss 2.018624782562256
 Step 400: loss 2.0613276958465576
 [Seed 10] Step 400 --> val loss 2.228867769241333
 Step 410: loss 2.034121513366699
 Step 420: loss 1.9971380233764648
 Step 430: loss 2.047893524169922
 Step 440: loss 1.990696907043457
 Step 450: loss 2.106754779815674
 Step 460: loss 2.0297625064849854
 Step 470: loss 1.9723005294799805
 Step 480: loss 2.020629644393921
 Step 490: loss 2.0190811157226562
 Step 500: loss 1.9668381214141846
 Step 510: loss 2.031806707382202
 Step 520: loss 1.928107500076294
 Step 530: loss 2.1007814407348633
 Step 540: loss 2.1034371852874756
 Step 550: loss 1.9903056621551514
 Step 560: loss 2.0551156997680664
 Step 570: loss 2.052628517150879
 Step 580: loss 2.069850444793701
 Step 590: loss 2.100399971008301
 Step 600: loss 2.032658576965332
 [Seed 10] Step 600 --> val loss 2.2686033248901367
 Step 610: loss 2.026585102081299
 Step 620: loss 1.8735871315002441
 Step 630: loss 1.9659428596496582
 Step 640: loss 1.958484411239624
 Step 650: loss 2.042579174041748
 Step 660: loss 2.030822515487671
 Step 670: loss 2.051046371459961
 Step 680: loss 2.0810022354125977
 Step 690: loss 2.0010733604431152
 Step 700: loss 2.049419403076172
 Step 710: loss 2.163206100463867
 Step 720: loss 2.094696521759033
 Step 730: loss 2.0663537979125977
 Step 740: loss 2.046722888946533
 Step 750: loss 2.0491433143615723
 Step 760: loss 2.1250414848327637
 Step 770: loss 2.1169726848602295
 Step 780: loss 2.118317127227783
 Step 790: loss 1.9706707000732422
 Step 800: loss 2.1656627655029297
 [Seed 10] Step 800 --> val loss 2.3163726329803467
 Step 810: loss 1.9978671073913574
 Step 820: loss 1.9783824682235718
 Step 830: loss 2.0602457523345947
 Step 840: loss 2.1056296825408936
 Step 850: loss 2.075490713119507
 Step 860: loss 2.0734472274780273
 Step 870: loss 2.0877630710601807
 Step 880: loss 2.027794361114502
 Step 890: loss 1.9068241119384766
 Step 900: loss 2.086662530899048
 Step 910: loss 2.139049768447876
 Step 920: loss 2.0308234691619873
 Step 930: loss 2.1306896209716797
 Step 940: loss 2.116211414337158
 Step 950: loss 2.050549268722534
 Step 960: loss 2.124309778213501
 Step 970: loss 2.1711976528167725
 Step 980: loss 2.0695812702178955
 Step 990: loss 2.0491929054260254
 Step 1000: loss 2.068603515625
 [Seed 10] Step 1000 --> val loss 2.332378625869751
 Step 1010: loss 2.0117082595825195
 Step 1020: loss 2.1399741172790527
 Step 1030: loss 2.2030534744262695
 Step 1040: loss 2.1696953773498535
 Step 1050: loss 2.026958465576172
 Step 1060: loss 2.1387553215026855
 Step 1070: loss 2.0418033599853516
 Step 1080: loss 2.035010814666748
 Step 1090: loss 2.0306928157806396
 Step 1100: loss 2.1601831912994385
 Step 1110: loss 2.095064640045166
 Step 1120: loss 1.99961256980896
 Step 1130: loss 2.0994060039520264
 Step 1140: loss 2.0109705924987793
 Step 1150: loss 2.169227123260498
 Step 1160: loss 2.060337781906128
 Step 1170: loss 2.120373249053955
 Step 1180: loss 2.1068103313446045
 Step 1190: loss 2.0811524391174316
 Step 1200: loss 2.088437557220459
 [Seed 10] Step 1200 --> val loss 2.3870372772216797
 Step 1210: loss 1.969043493270874
 Step 1220: loss 2.084329843521118
 Step 1230: loss 2.0182266235351562
 Step 1240: loss 1.956784725189209
 Step 1250: loss 2.0920424461364746
 Step 1260: loss 2.086984157562256
 Step 1270: loss 2.130443811416626
 Step 1280: loss 2.095449924468994
 Step 1290: loss 1.9634923934936523
 Step 1300: loss 2.1126229763031006
 Step 1310: loss 1.9460422992706299
 Step 1320: loss 2.0732598304748535
 Step 1330: loss 2.11653995513916
 Step 1340: loss 2.000356674194336
 Step 1350: loss 2.0143227577209473
 Step 1360: loss 2.0295214653015137
 Step 1370: loss 1.9995057582855225
 Step 1380: loss 1.9661550521850586
 Step 1390: loss 2.0389273166656494
 Step 1400: loss 2.0906567573547363
 [Seed 10] Step 1400 --> val loss 2.28161883354187
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 10] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 10] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 10] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [45 59 51 43 52 58 57 11  0 13] ...
 First target sequence: [59 51 43 52 58 57 11  0 13 52] ...

 Decoded input: guments;
 And, if I fall not in my deep intent,
 Clarence hath not

 Decoded target: uments;
 And, if I fall not in my deep intent,
 Clarence hath not 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.210469722747803
 [Seed 11] Step 0 --> val loss 4.185428619384766
 Step 10: loss 3.864114761352539
 Step 20: loss 3.3361968994140625
 Step 30: loss 2.8760986328125
 Step 40: loss 2.682473659515381
 Step 50: loss 2.620961904525757
 Step 60: loss 2.507976531982422
 Step 70: loss 2.461561918258667
 Step 80: loss 2.37260103225708
 Step 90: loss 2.3256945610046387
 Step 100: loss 2.31980299949646
 Step 110: loss 2.283308506011963
 Step 120: loss 2.252631187438965
 Step 130: loss 2.1882169246673584
 Step 140: loss 2.2004990577697754
 Step 150: loss 2.171968936920166
 Step 160: loss 2.1490960121154785
 Step 170: loss 2.1934030055999756
 Step 180: loss 2.123075008392334
 Step 190: loss 2.076899528503418
 Step 200: loss 2.051532506942749
 [Seed 11] Step 200 --> val loss 2.4605181217193604
 Step 210: loss 2.0281548500061035
 Step 220: loss 2.0680294036865234
 Step 230: loss 1.9601930379867554
 Step 240: loss 2.172732353210449
 Step 250: loss 2.013998031616211
 Step 260: loss 2.1212034225463867
 Step 270: loss 2.0442252159118652
 Step 280: loss 2.0183472633361816
 Step 290: loss 2.039313316345215
 Step 300: loss 2.104043960571289
 Step 310: loss 2.0509753227233887
 Step 320: loss 2.0688188076019287
 Step 330: loss 1.9781694412231445
 Step 340: loss 2.1209094524383545
 Step 350: loss 2.006887674331665
 Step 360: loss 1.9781912565231323
 Step 370: loss 2.0122218132019043
 Step 380: loss 2.015000104904175
 Step 390: loss 2.02461576461792
 Step 400: loss 2.0510149002075195
 [Seed 11] Step 400 --> val loss 2.2818894386291504
 Step 410: loss 2.0573432445526123
 Step 420: loss 1.9513611793518066
 Step 430: loss 2.0842576026916504
 Step 440: loss 1.99560546875
 Step 450: loss 2.0604310035705566
 Step 460: loss 2.004312038421631
 Step 470: loss 2.0159401893615723
 Step 480: loss 1.9750041961669922
 Step 490: loss 2.10662841796875
 Step 500: loss 2.063169002532959
 Step 510: loss 1.9992197751998901
 Step 520: loss 2.026538848876953
 Step 530: loss 2.1136348247528076
 Step 540: loss 2.002758741378784
 Step 550: loss 1.975529432296753
 Step 560: loss 1.9498262405395508
 Step 570: loss 1.9011510610580444
 Step 580: loss 1.9333105087280273
 Step 590: loss 2.0509753227233887
 Step 600: loss 1.9807456731796265
 [Seed 11] Step 600 --> val loss 2.4093196392059326
 Step 610: loss 1.9887676239013672
 Step 620: loss 1.8947124481201172
 Step 630: loss 1.950779676437378
 Step 640: loss 2.052327871322632
 Step 650: loss 2.0770530700683594
 Step 660: loss 2.003446102142334
 Step 670: loss 1.9907493591308594
 Step 680: loss 1.98565673828125
 Step 690: loss 2.1021370887756348
 Step 700: loss 1.9304649829864502
 Step 710: loss 2.031313419342041
 Step 720: loss 2.0029733180999756
 Step 730: loss 1.9265035390853882
 Step 740: loss 2.0580410957336426
 Step 750: loss 1.9696621894836426
 Step 760: loss 2.07951283454895
 Step 770: loss 1.970069169998169
 Step 780: loss 1.9165023565292358
 Step 790: loss 1.9158704280853271
 Step 800: loss 1.9006555080413818
 [Seed 11] Step 800 --> val loss 2.236539602279663
 Step 810: loss 2.0405521392822266
 Step 820: loss 2.102752685546875
 Step 830: loss 2.01188588142395
 Step 840: loss 1.9569594860076904
 Step 850: loss 1.9715826511383057
 Step 860: loss 1.9035539627075195
 Step 870: loss 1.8425487279891968
 Step 880: loss 2.0434176921844482
 Step 890: loss 2.020329475402832
 Step 900: loss 2.0302295684814453
 Step 910: loss 2.0266876220703125
 Step 920: loss 1.9981210231781006
 Step 930: loss 1.9662322998046875
 Step 940: loss 1.912244200706482
 Step 950: loss 1.9508781433105469
 Step 960: loss 1.997481346130371
 Step 970: loss 1.8536094427108765
 Step 980: loss 1.9835014343261719
 Step 990: loss 1.8800609111785889
 Step 1000: loss 1.9148786067962646
 [Seed 11] Step 1000 --> val loss 2.186709403991699
 Step 1010: loss 1.9905219078063965
 Step 1020: loss 1.9335250854492188
 Step 1030: loss 1.9684207439422607
 Step 1040: loss 1.9791619777679443
 Step 1050: loss 1.9589815139770508
 Step 1060: loss 1.9705266952514648
 Step 1070: loss 1.9248980283737183
 Step 1080: loss 1.9047092199325562
 Step 1090: loss 1.9070475101470947
 Step 1100: loss 1.9777024984359741
 Step 1110: loss 1.9917399883270264
 Step 1120: loss 1.9372423887252808
 Step 1130: loss 1.9547653198242188
 Step 1140: loss 2.0053958892822266
 Step 1150: loss 1.9226592779159546
 Step 1160: loss 1.894850492477417
 Step 1170: loss 1.8917864561080933
 Step 1180: loss 1.9511274099349976
 Step 1190: loss 1.9311697483062744
 Step 1200: loss 1.8023662567138672
 [Seed 11] Step 1200 --> val loss 2.2839431762695312
 Step 1210: loss 1.9237827062606812
 Step 1220: loss 1.9840030670166016
 Step 1230: loss 1.8795708417892456
 Step 1240: loss 2.0151376724243164
 Step 1250: loss 2.000460386276245
 Step 1260: loss 2.010453462600708
 Step 1270: loss 1.994962453842163
 Step 1280: loss 1.8583788871765137
 Step 1290: loss 1.9158014059066772
 Step 1300: loss 1.9663684368133545
 Step 1310: loss 1.9278510808944702
 Step 1320: loss 1.9270477294921875
 Step 1330: loss 1.7990474700927734
 Step 1340: loss 1.876729965209961
 Step 1350: loss 1.8641796112060547
 Step 1360: loss 2.0842535495758057
 Step 1370: loss 1.8090074062347412
 Step 1380: loss 1.9278968572616577
 Step 1390: loss 1.8372526168823242
 Step 1400: loss 1.8740074634552002
 [Seed 11] Step 1400 --> val loss 2.1324760913848877
 Step 1410: loss 1.7783187627792358
 Step 1420: loss 1.754773497581482
 Step 1430: loss 1.8315002918243408
 Step 1440: loss 1.8171967267990112
 Step 1450: loss 1.7541837692260742
 Step 1460: loss 1.8534619808197021
 Step 1470: loss 2.003803014755249
 Step 1480: loss 1.8213237524032593
 Step 1490: loss 1.8520593643188477
 Step 1500: loss 1.7670717239379883
 Step 1510: loss 1.8213218450546265
 Step 1520: loss 1.8501085042953491
 Step 1530: loss 1.7034273147583008
 Step 1540: loss 1.930643081665039
 Step 1550: loss 1.7958590984344482
 Step 1560: loss 1.7206153869628906
 Step 1570: loss 1.8203034400939941
 Step 1580: loss 1.6394269466400146
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 11] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 11] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 11] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [43  1 42 59 49 43  1 61 47 50] ...
 First target sequence: [ 1 42 59 49 43  1 61 47 50 50] ...

 Decoded input: e duke will return no more; or you
 imagine me too unhurtful an o

 Decoded target:  duke will return no more; or you
 imagine me too unhurtful an op
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.218724727630615
 [Seed 12] Step 0 --> val loss 4.1834635734558105
 Step 10: loss 3.8534438610076904
 Step 20: loss 3.2904582023620605
 Step 30: loss 2.9023032188415527
 Step 40: loss 2.7139053344726562
 Step 50: loss 2.558980941772461
 Step 60: loss 2.4629087448120117
 Step 70: loss 2.474635601043701
 Step 80: loss 2.3033905029296875
 Step 90: loss 2.32216215133667
 Step 100: loss 2.252140998840332
 Step 110: loss 2.2550458908081055
 Step 120: loss 2.2178163528442383
 Step 130: loss 2.1775097846984863
 Step 140: loss 2.237761974334717
 Step 150: loss 2.135385036468506
 Step 160: loss 2.1106715202331543
 Step 170: loss 2.10347843170166
 Step 180: loss 2.2011332511901855
 Step 190: loss 2.1375620365142822
 Step 200: loss 2.1199755668640137
 [Seed 12] Step 200 --> val loss 2.2758030891418457
 Step 210: loss 2.147359848022461
 Step 220: loss 2.0193893909454346
 Step 230: loss 2.110912561416626
 Step 240: loss 2.0626585483551025
 Step 250: loss 2.0979180335998535
 Step 260: loss 2.0729117393493652
 Step 270: loss 1.964705228805542
 Step 280: loss 2.056720733642578
 Step 290: loss 2.067342519760132
 Step 300: loss 2.1458740234375
 Step 310: loss 2.0035321712493896
 Step 320: loss 1.9739904403686523
 Step 330: loss 2.0199344158172607
 Step 340: loss 2.103950262069702
 Step 350: loss 1.9952754974365234
 Step 360: loss 2.07157564163208
 Step 370: loss 2.043424367904663
 Step 380: loss 1.9709837436676025
 Step 390: loss 2.1073787212371826
 Step 400: loss 1.988936424255371
 [Seed 12] Step 400 --> val loss 2.2659549713134766
 Step 410: loss 2.0401206016540527
 Step 420: loss 2.0061111450195312
 Step 430: loss 1.9877393245697021
 Step 440: loss 2.037564992904663
 Step 450: loss 1.907461166381836
 Step 460: loss 2.031639575958252
 Step 470: loss 2.024646043777466
 Step 480: loss 2.030670166015625
 Step 490: loss 1.9655647277832031
 Step 500: loss 1.9535531997680664
 Step 510: loss 1.9534316062927246
 Step 520: loss 2.0904746055603027
 Step 530: loss 1.9928531646728516
 Step 540: loss 2.0712807178497314
 Step 550: loss 1.9309165477752686
 Step 560: loss 1.9621689319610596
 Step 570: loss 1.8777015209197998
 Step 580: loss 1.9114841222763062
 Step 590: loss 2.054169178009033
 Step 600: loss 1.997165560722351
 [Seed 12] Step 600 --> val loss 2.348665714263916
 Step 610: loss 2.0175795555114746
 Step 620: loss 2.1117072105407715
 Step 630: loss 1.9604600667953491
 Step 640: loss 1.9775454998016357
 Step 650: loss 2.107494354248047
 Step 660: loss 1.9149353504180908
 Step 670: loss 2.0372815132141113
 Step 680: loss 1.9887545108795166
 Step 690: loss 1.9732431173324585
 Step 700: loss 1.9805930852890015
 Step 710: loss 2.060826539993286
 Step 720: loss 1.9431512355804443
 Step 730: loss 1.9568367004394531
 Step 740: loss 1.928457498550415
 Step 750: loss 2.0571470260620117
 Step 760: loss 1.9877855777740479
 Step 770: loss 2.0734243392944336
 Step 780: loss 2.0295305252075195
 Step 790: loss 2.0821895599365234
 Step 800: loss 2.0150413513183594
 [Seed 12] Step 800 --> val loss 2.3185365200042725
 Step 810: loss 2.036541223526001
 Step 820: loss 1.9669404029846191
 Step 830: loss 1.9787571430206299
 Step 840: loss 2.0018696784973145
 Step 850: loss 1.9204705953598022
 Step 860: loss 1.9999229907989502
 Step 870: loss 2.145423650741577
 Step 880: loss 1.9994224309921265
 Step 890: loss 1.9313668012619019
 Step 900: loss 1.9742765426635742
 Step 910: loss 1.946262240409851
 Step 920: loss 1.8533143997192383
 Step 930: loss 2.0200343132019043
 Step 940: loss 1.862230658531189
 Step 950: loss 1.980712652206421
 Step 960: loss 1.9596059322357178
 Step 970: loss 1.976704478263855
 Step 980: loss 2.036360263824463
 Step 990: loss 1.880743384361267
 Step 1000: loss 1.9659643173217773
 [Seed 12] Step 1000 --> val loss 2.132054567337036
 Step 1010: loss 1.8946236371994019
 Step 1020: loss 1.8504045009613037
 Step 1030: loss 1.862898588180542
 Step 1040: loss 1.9859693050384521
 Step 1050: loss 1.9348316192626953
 Step 1060: loss 1.912575602531433
 Step 1070: loss 1.9349912405014038
 Step 1080: loss 1.891371250152588
 Step 1090: loss 1.9597890377044678
 Step 1100: loss 1.9035518169403076
 Step 1110: loss 1.9645580053329468
 Step 1120: loss 1.9922988414764404
 Step 1130: loss 2.0146913528442383
 Step 1140: loss 2.0565602779388428
 Step 1150: loss 2.044619560241699
 Step 1160: loss 2.0963685512542725
 Step 1170: loss 1.8409336805343628
 Step 1180: loss 1.9124960899353027
 Step 1190: loss 2.0543336868286133
 Step 1200: loss 1.8992164134979248
 [Seed 12] Step 1200 --> val loss 2.276977777481079
 Step 1210: loss 1.8322151899337769
 Step 1220: loss 2.0040042400360107
 Step 1230: loss 1.9218978881835938
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 12] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 12] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 12] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 12] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [50 50  1 58 46 43  1 58 56 39] ...
 First target sequence: [50  1 58 46 43  1 58 56 39 47] ...

 Decoded input: ll the traitor, in the high'st degree
 He hath abused your powers

 Decoded target: l the traitor, in the high'st degree
 He hath abused your powers.
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.228219985961914
 [Seed 13] Step 0 --> val loss 4.184866905212402
 Step 10: loss 3.854077100753784
 Step 20: loss 3.310561180114746
 Step 30: loss 2.8490843772888184
 Step 40: loss 2.6574370861053467
 Step 50: loss 2.58359432220459
 Step 60: loss 2.5253076553344727
 Step 70: loss 2.4379324913024902
 Step 80: loss 2.496168613433838
 Step 90: loss 2.256075859069824
 Step 100: loss 2.318638324737549
 Step 110: loss 2.2492599487304688
 Step 120: loss 2.2911250591278076
 Step 130: loss 2.2760190963745117
 Step 140: loss 2.234922170639038
 Step 150: loss 2.1557235717773438
 Step 160: loss 2.2065439224243164
 Step 170: loss 2.0444788932800293
 Step 180: loss 2.0086731910705566
 Step 190: loss 2.002027988433838
 Step 200: loss 1.9765914678573608
 [Seed 13] Step 200 --> val loss 2.3457515239715576
 Step 210: loss 2.0794196128845215
 Step 220: loss 2.130774974822998
 Step 230: loss 2.063762903213501
 Step 240: loss 2.1554057598114014
 Step 250: loss 2.0546605587005615
 Step 260: loss 2.0058481693267822
 Step 270: loss 2.0722508430480957
 Step 280: loss 2.101771831512451
 Step 290: loss 1.9916419982910156
 Step 300: loss 1.9635875225067139
 Step 310: loss 2.027451515197754
 Step 320: loss 2.029641628265381
 Step 330: loss 2.0243613719940186
 Step 340: loss 1.9692378044128418
 Step 350: loss 1.9833307266235352
 Step 360: loss 2.051652669906616
 Step 370: loss 2.0390539169311523
 Step 380: loss 1.9239814281463623
 Step 390: loss 2.000922679901123
 Step 400: loss 2.0860366821289062
 [Seed 13] Step 400 --> val loss 2.209822177886963
 Step 410: loss 2.0381016731262207
 Step 420: loss 1.9924745559692383
 Step 430: loss 1.9576200246810913
 Step 440: loss 1.9196547269821167
 Step 450: loss 2.091168165206909
 Step 460: loss 2.007999897003174
 Step 470: loss 2.022353172302246
 Step 480: loss 2.00490665435791
 Step 490: loss 2.0650382041931152
 Step 500: loss 2.0571887493133545
 Step 510: loss 2.0776076316833496
 Step 520: loss 1.938120722770691
 Step 530: loss 2.070878505706787
 Step 540: loss 1.9368042945861816
 Step 550: loss 2.0121984481811523
 Step 560: loss 2.148251533508301
 Step 570: loss 1.9734487533569336
 Step 580: loss 1.9245431423187256
 Step 590: loss 2.0377564430236816
 Step 600: loss 2.1144421100616455
 [Seed 13] Step 600 --> val loss 2.2586112022399902
 Step 610: loss 2.0314855575561523
 Step 620: loss 2.0260825157165527
 Step 630: loss 1.9671039581298828
 Step 640: loss 1.9954776763916016
 Step 650: loss 1.8763597011566162
 Step 660: loss 1.940698504447937
 Step 670: loss 2.0271756649017334
 Step 680: loss 1.9647550582885742
 Step 690: loss 1.8444856405258179
 Step 700: loss 2.0224521160125732
 Step 710: loss 2.0245301723480225
 Step 720: loss 1.9230564832687378
 Step 730: loss 2.0579192638397217
 Step 740: loss 1.9569458961486816
 Step 750: loss 2.0079853534698486
 Step 760: loss 1.9887818098068237
 Step 770: loss 1.8885576725006104
 Step 780: loss 2.0141677856445312
 Step 790: loss 1.9206702709197998
 Step 800: loss 1.9409464597702026
 [Seed 13] Step 800 --> val loss 2.1867923736572266
 Step 810: loss 1.9887022972106934
 Step 820: loss 2.013219118118286
 Step 830: loss 2.01131010055542
 Step 840: loss 1.9695756435394287
 Step 850: loss 2.021116256713867
 Step 860: loss 1.9385960102081299
 Step 870: loss 2.013826608657837
 Step 880: loss 1.9981210231781006
 Step 890: loss 1.959172010421753
 Step 900: loss 1.8224718570709229
 Step 910: loss 2.0122246742248535
 Step 920: loss 1.9592616558074951
 Step 930: loss 1.9573650360107422
 Step 940: loss 2.063533306121826
 Step 950: loss 1.9652059078216553
 Step 960: loss 1.869673728942871
 Step 970: loss 1.9412626028060913
 Step 980: loss 1.984804630279541
 Step 990: loss 1.8738079071044922
 Step 1000: loss 1.929560661315918
 [Seed 13] Step 1000 --> val loss 2.215526580810547
 Step 1010: loss 1.9505994319915771
 Step 1020: loss 2.046518087387085
 Step 1030: loss 1.9548224210739136
 Step 1040: loss 1.8995122909545898
 Step 1050: loss 1.8836629390716553
 Step 1060: loss 1.8263921737670898
 Step 1070: loss 1.878386378288269
 Step 1080: loss 1.8619236946105957
 Step 1090: loss 1.9628651142120361
 Step 1100: loss 1.8878517150878906
 Step 1110: loss 1.9625486135482788
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 13] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 13] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 13] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 13] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 13] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [57  1 46 39 58 46  1 58 46 43] ...
 First target sequence: [ 1 46 39 58 46  1 58 46 43  1] ...

 Decoded input: s hath the duke inferr'd;'
 But nothing spake in warrant from him

 Decoded target:  hath the duke inferr'd;'
 But nothing spake in warrant from hims
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.214322566986084
 [Seed 14] Step 0 --> val loss 4.183311939239502
 Step 10: loss 3.8586859703063965
 Step 20: loss 3.3267571926116943
 Step 30: loss 2.764962673187256
 Step 40: loss 2.698190689086914
 Step 50: loss 2.5705060958862305
 Step 60: loss 2.4922549724578857
 Step 70: loss 2.393315076828003
 Step 80: loss 2.3936023712158203
 Step 90: loss 2.304429769515991
 Step 100: loss 2.3080430030822754
 Step 110: loss 2.307769536972046
 Step 120: loss 2.1632275581359863
 Step 130: loss 2.282217502593994
 Step 140: loss 2.1745765209198
 Step 150: loss 2.1588985919952393
 Step 160: loss 2.140364646911621
 Step 170: loss 2.1777119636535645
 Step 180: loss 2.088848829269409
 Step 190: loss 2.1248207092285156
 Step 200: loss 2.188889503479004
 [Seed 14] Step 200 --> val loss 2.3974146842956543
 Step 210: loss 2.113356351852417
 Step 220: loss 2.1063270568847656
 Step 230: loss 2.063946485519409
 Step 240: loss 2.0702152252197266
 Step 250: loss 2.0166304111480713
 Step 260: loss 2.0377793312072754
 Step 270: loss 2.0519018173217773
 Step 280: loss 2.0744495391845703
 Step 290: loss 1.944283366203308
 Step 300: loss 2.070620536804199
 Step 310: loss 2.07479190826416
 Step 320: loss 2.08650279045105
 Step 330: loss 2.0422439575195312
 Step 340: loss 2.0622050762176514
 Step 350: loss 2.0028505325317383
 Step 360: loss 2.087881088256836
 Step 370: loss 1.9996356964111328
 Step 380: loss 2.009042263031006
 Step 390: loss 1.9949007034301758
 Step 400: loss 2.104245185852051
 [Seed 14] Step 400 --> val loss 2.2902073860168457
 Step 410: loss 2.090256452560425
 Step 420: loss 2.202742576599121
 Step 430: loss 1.9650993347167969
 Step 440: loss 2.0402441024780273
 Step 450: loss 2.0628983974456787
 Step 460: loss 1.9264129400253296
 Step 470: loss 1.9860042333602905
 Step 480: loss 2.1274900436401367
 Step 490: loss 2.06549334526062
 Step 500: loss 2.066545009613037
 Step 510: loss 1.9501804113388062
 Step 520: loss 1.9648617506027222
 Step 530: loss 1.9574246406555176
 Step 540: loss 2.0098447799682617
 Step 550: loss 2.1046628952026367
 Step 560: loss 1.9697364568710327
 Step 570: loss 2.0097806453704834
 Step 580: loss 2.0062572956085205
 Step 590: loss 2.0571043491363525
 Step 600: loss 1.9335708618164062
 [Seed 14] Step 600 --> val loss 2.244596481323242
 Step 610: loss 2.053191661834717
 Step 620: loss 1.9330191612243652
 Step 630: loss 1.9651899337768555
 Step 640: loss 1.9619840383529663
 Step 650: loss 2.0168850421905518
 Step 660: loss 2.00738525390625
 Step 670: loss 1.925236701965332
 Step 680: loss 1.9887237548828125
 Step 690: loss 2.0181899070739746
 Step 700: loss 1.9800901412963867
 Step 710: loss 2.0762009620666504
 Step 720: loss 1.9831079244613647
 Step 730: loss 2.0594594478607178
 Step 740: loss 2.0765552520751953
 Step 750: loss 1.9606668949127197
 Step 760: loss 2.0786991119384766
 Step 770: loss 2.095529317855835
 Step 780: loss 2.0483083724975586
 Step 790: loss 1.9845775365829468
 Step 800: loss 2.08445405960083
 [Seed 14] Step 800 --> val loss 2.3084192276000977
 Step 810: loss 2.043938398361206
 Step 820: loss 1.9754109382629395
 Step 830: loss 2.062103033065796
 Step 840: loss 1.9362192153930664
 Step 850: loss 1.9957650899887085
 Step 860: loss 2.0565085411071777
 Step 870: loss 1.8575503826141357
 Step 880: loss 2.0344629287719727
 Step 890: loss 2.0127482414245605
 Step 900: loss 1.9062564373016357
 Step 910: loss 1.8868975639343262
 Step 920: loss 1.9611200094223022
 Step 930: loss 1.8491071462631226
 Step 940: loss 1.9314355850219727
 Step 950: loss 1.9792102575302124
 Step 960: loss 1.9843658208847046
 Step 970: loss 2.110795021057129
 Step 980: loss 2.014063835144043
 Step 990: loss 1.9366058111190796
 Step 1000: loss 1.9118599891662598
 [Seed 14] Step 1000 --> val loss 2.1884570121765137
 Step 1010: loss 1.984410285949707
 Step 1020: loss 1.8815416097640991
 Step 1030: loss 1.9770963191986084
 Step 1040: loss 1.8546741008758545
 Step 1050: loss 1.8810367584228516
 Step 1060: loss 1.9040420055389404
 Step 1070: loss 2.073944091796875
 Step 1080: loss 1.9118441343307495
 Step 1090: loss 1.872460126876831
 Step 1100: loss 1.913826823234558
 Step 1110: loss 1.8065640926361084
 Step 1120: loss 1.7792829275131226
 Step 1130: loss 1.855804204940796
 Step 1140: loss 1.6772079467773438
 Step 1150: loss 1.9156665802001953
 Step 1160: loss 1.9874446392059326
 Step 1170: loss 1.8978943824768066
 Step 1180: loss 1.9088778495788574
 Step 1190: loss 1.9128730297088623
 Step 1200: loss 2.006446599960327
 [Seed 14] Step 1200 --> val loss 2.130993127822876
 Step 1210: loss 1.9459559917449951
 Step 1220: loss 1.8342232704162598
 Step 1230: loss 1.8507342338562012
 Step 1240: loss 1.836674451828003
 Step 1250: loss 1.859811782836914
 Step 1260: loss 1.921095848083496
 Step 1270: loss 1.8706554174423218
 Step 1280: loss 1.9765222072601318
 Step 1290: loss 1.7845070362091064
 Step 1300: loss 1.8620446920394897
 Step 1310: loss 1.9123543500900269
 Step 1320: loss 1.8215482234954834
 Step 1330: loss 1.9753503799438477
 Step 1340: loss 1.8688087463378906
 Step 1350: loss 1.7308930158615112
 Step 1360: loss 1.8665064573287964
 Step 1370: loss 1.7571513652801514
 Step 1380: loss 1.8997124433517456
 Step 1390: loss 1.9425463676452637
 Step 1400: loss 1.8839399814605713
 [Seed 14] Step 1400 --> val loss 2.019545078277588
 Step 1410: loss 1.7865408658981323
 Step 1420: loss 1.6897448301315308
 Step 1430: loss 1.8118464946746826
 Step 1440: loss 1.9482735395431519
 Step 1450: loss 1.6727803945541382
 Step 1460: loss 1.7642202377319336
 Step 1470: loss 1.8439258337020874
 Step 1480: loss 1.7966079711914062
 Step 1490: loss 1.833447813987732
 Step 1500: loss 1.7692183256149292
 Step 1510: loss 1.8190562725067139
 Step 1520: loss 1.9265108108520508
 Step 1530: loss 1.7244490385055542
 Step 1540: loss 1.8089520931243896
 Step 1550: loss 1.738590121269226
 Step 1560: loss 1.8967679738998413
 Step 1570: loss 1.756434440612793
 Step 1580: loss 1.8009586334228516
 Step 1590: loss 1.7913963794708252
 Step 1600: loss 1.887265920639038
 [Seed 14] Step 1600 --> val loss 2.035127639770508
 Step 1610: loss 1.8041938543319702
 Step 1620: loss 1.6592612266540527
 Step 1630: loss 1.7493302822113037
 Step 1640: loss 1.8478020429611206
 Step 1650: loss 1.7129478454589844
 Step 1660: loss 1.726773977279663
 Step 1670: loss 1.701298475265503
 Step 1680: loss 1.7482919692993164
 Step 1690: loss 1.7763617038726807
 Step 1700: loss 1.6260122060775757
 Step 1710: loss 1.8164088726043701
 Step 1720: loss 1.7138557434082031
 Step 1730: loss 1.771691083908081
 Step 1740: loss 1.7605724334716797
 Step 1750: loss 1.6452713012695312
 Step 1760: loss 1.7099370956420898
 Step 1770: loss 1.6417292356491089
 Step 1780: loss 1.7280381917953491
 Step 1790: loss 1.851511836051941
 Step 1800: loss 1.7705343961715698
 [Seed 14] Step 1800 --> val loss 1.92317795753479
 Step 1810: loss 1.8964691162109375
 Step 1820: loss 1.6496450901031494
 Step 1830: loss 1.7339789867401123
 Step 1840: loss 1.7008590698242188
 Step 1850: loss 1.7828099727630615
 Step 1860: loss 1.7321078777313232
 Step 1870: loss 1.7445523738861084
 Step 1880: loss 1.8075730800628662
 Step 1890: loss 1.6636136770248413
 Step 1900: loss 1.7787714004516602
 Step 1910: loss 1.6590030193328857
 Step 1920: loss 1.7205545902252197
 Step 1930: loss 1.6287736892700195
 Step 1940: loss 1.7377837896347046
 Step 1950: loss 1.6787885427474976
 Step 1960: loss 1.6503489017486572
 Step 1970: loss 1.768946647644043
 Step 1980: loss 1.6561931371688843
 Step 1990: loss 1.7704805135726929
 Step 2000: loss 1.5696346759796143
 [Seed 14] Step 2000 --> val loss 1.9000786542892456
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [43  1 50 53 56 42 57  6  0 25] ...
 First target sequence: [ 1 50 53 56 42 57  6  0 25 59] ...

 Decoded input: e lords,
 Must give this cur the lie: and his own notion--
 Who we

 Decoded target:  lords,
 Must give this cur the lie: and his own notion--
 Who wea
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.219054222106934
 [Seed 15] Step 0 --> val loss 4.184095859527588
 Step 10: loss 3.8447999954223633
 Step 20: loss 3.2786293029785156
 Step 30: loss 2.817779541015625
 Step 40: loss 2.722820281982422
 Step 50: loss 2.646718740463257
 Step 60: loss 2.5064585208892822
 Step 70: loss 2.368252754211426
 Step 80: loss 2.308896064758301
 Step 90: loss 2.220158576965332
 Step 100: loss 2.207674264907837
 Step 110: loss 2.3054089546203613
 Step 120: loss 2.2268226146698
 Step 130: loss 2.1403374671936035
 Step 140: loss 2.279939651489258
 Step 150: loss 2.1296119689941406
 Step 160: loss 2.0849967002868652
 Step 170: loss 2.283351421356201
 Step 180: loss 2.104495048522949
 Step 190: loss 2.0379257202148438
 Step 200: loss 2.168125867843628
 [Seed 15] Step 200 --> val loss 2.2489566802978516
 Step 210: loss 1.9903377294540405
 Step 220: loss 2.14359188079834
 Step 230: loss 2.04860258102417
 Step 240: loss 2.031703472137451
 Step 250: loss 2.1054258346557617
 Step 260: loss 2.091693878173828
 Step 270: loss 1.9968328475952148
 Step 280: loss 2.142895221710205
 Step 290: loss 2.2062206268310547
 Step 300: loss 1.9212067127227783
 Step 310: loss 1.9482645988464355
 Step 320: loss 2.0060746669769287
 Step 330: loss 1.9305461645126343
 Step 340: loss 2.148890972137451
 Step 350: loss 2.026111602783203
 Step 360: loss 1.9492881298065186
 Step 370: loss 1.9706714153289795
 Step 380: loss 2.0344600677490234
 Step 390: loss 2.028292417526245
 Step 400: loss 1.9669444561004639
 [Seed 15] Step 400 --> val loss 2.326310873031616
 Step 410: loss 2.1360220909118652
 Step 420: loss 1.8804588317871094
 Step 430: loss 1.9455623626708984
 Step 440: loss 2.1150360107421875
 Step 450: loss 1.84795081615448
 Step 460: loss 1.995789885520935
 Step 470: loss 2.027830123901367
 Step 480: loss 1.9715683460235596
 Step 490: loss 2.048245668411255
 Step 500: loss 2.111473798751831
 Step 510: loss 1.971390724182129
 Step 520: loss 2.030533790588379
 Step 530: loss 2.0727908611297607
 Step 540: loss 1.99965238571167
 Step 550: loss 1.9642360210418701
 Step 560: loss 1.8840636014938354
 Step 570: loss 1.9660353660583496
 Step 580: loss 2.0839831829071045
 Step 590: loss 2.1103832721710205
 Step 600: loss 2.019822597503662
 [Seed 15] Step 600 --> val loss 2.1955010890960693
 Step 610: loss 1.9761312007904053
 Step 620: loss 2.103778839111328
 Step 630: loss 2.0033211708068848
 Step 640: loss 1.9067219495773315
 Step 650: loss 2.004962921142578
 Step 660: loss 1.9633872509002686
 Step 670: loss 1.9936975240707397
 Step 680: loss 2.0794687271118164
 Step 690: loss 2.1355552673339844
 Step 700: loss 1.9582427740097046
 Step 710: loss 1.9303808212280273
 Step 720: loss 2.141989231109619
 Step 730: loss 2.0221991539001465
 Step 740: loss 1.9969695806503296
 Step 750: loss 1.9585368633270264
 Step 760: loss 1.8632667064666748
 Step 770: loss 2.0593175888061523
 Step 780: loss 2.036121368408203
 Step 790: loss 1.992119312286377
 Step 800: loss 1.932748794555664
 [Seed 15] Step 800 --> val loss 2.168508291244507
 Step 810: loss 1.9831103086471558
 Step 820: loss 1.9256954193115234
 Step 830: loss 2.001389741897583
 Step 840: loss 2.0237221717834473
 Step 850: loss 1.9398574829101562
 Step 860: loss 2.049471855163574
 Step 870: loss 1.9855053424835205
 Step 880: loss 2.0417087078094482
 Step 890: loss 1.9406659603118896
 Step 900: loss 1.9356813430786133
 Step 910: loss 1.9446004629135132
 Step 920: loss 1.9773119688034058
 Step 930: loss 1.902752161026001
 Step 940: loss 1.9908040761947632
 Step 950: loss 1.9080569744110107
 Step 960: loss 1.9912779331207275
 Step 970: loss 1.9896544218063354
 Step 980: loss 1.8943389654159546
 Step 990: loss 1.9307506084442139
 Step 1000: loss 1.8668181896209717
 [Seed 15] Step 1000 --> val loss 2.1483535766601562
 Step 1010: loss 2.0467681884765625
 Step 1020: loss 2.099368095397949
 Step 1030: loss 1.9739129543304443
 Step 1040: loss 1.8653873205184937
 Step 1050: loss 1.9239192008972168
 Step 1060: loss 1.9309719800949097
 Step 1070: loss 1.9410912990570068
 Step 1080: loss 1.8920539617538452
 Step 1090: loss 1.8425335884094238
 Step 1100: loss 1.977574348449707
 Step 1110: loss 2.0475687980651855
 Step 1120: loss 1.8263237476348877
 Step 1130: loss 1.9033622741699219
 Step 1140: loss 1.91852605342865
 Step 1150: loss 1.9106214046478271
 Step 1160: loss 1.9862674474716187
 Step 1170: loss 1.7957487106323242
 Step 1180: loss 1.8087104558944702
 Step 1190: loss 1.890737533569336
 Step 1200: loss 1.7616784572601318
 [Seed 15] Step 1200 --> val loss 1.9754136800765991
 Step 1210: loss 1.7712409496307373
 Step 1220: loss 1.9472084045410156
 Step 1230: loss 1.8663966655731201
 Step 1240: loss 1.9062068462371826
 Step 1250: loss 1.8314247131347656
 Step 1260: loss 1.6948007345199585
 Step 1270: loss 1.8077964782714844
 Step 1280: loss 1.7374744415283203
 Step 1290: loss 1.9204585552215576
 Step 1300: loss 1.7938802242279053
 Step 1310: loss 1.956791877746582
 Step 1320: loss 1.9651894569396973
 Step 1330: loss 1.8053805828094482
 Step 1340: loss 1.8675789833068848
 Step 1350: loss 1.7167861461639404
 Step 1360: loss 1.829533576965332
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 15] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 15] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 15] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 15] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [50  1 46 59 57 40 39 52 42 11] ...
 First target sequence: [ 1 46 59 57 40 39 52 42 11  0] ...

 Decoded input: l husband;
 The other for some while a friend.

 LEONTES:

 MAMILLI

 Decoded target:  husband;
 The other for some while a friend.

 LEONTES:

 MAMILLIU
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.215900897979736
 [Seed 16] Step 0 --> val loss 4.184566497802734
 Step 10: loss 3.8529555797576904
 Step 20: loss 3.3063712120056152
 Step 30: loss 2.908135414123535
 Step 40: loss 2.756844997406006
 Step 50: loss 2.5868184566497803
 Step 60: loss 2.574083089828491
 Step 70: loss 2.4522721767425537
 Step 80: loss 2.330817937850952
 Step 90: loss 2.3724982738494873
 Step 100: loss 2.2006993293762207
 Step 110: loss 2.180032253265381
 Step 120: loss 2.2619218826293945
 Step 130: loss 2.2343039512634277
 Step 140: loss 2.161010503768921
 Step 150: loss 2.116676092147827
 Step 160: loss 2.0418927669525146
 Step 170: loss 2.103837013244629
 Step 180: loss 2.179147481918335
 Step 190: loss 2.04935359954834
 Step 200: loss 2.1846933364868164
 [Seed 16] Step 200 --> val loss 2.3221375942230225
 Step 210: loss 2.1431305408477783
 Step 220: loss 2.081643581390381
 Step 230: loss 2.207257032394409
 Step 240: loss 1.9976972341537476
 Step 250: loss 2.128826141357422
 Step 260: loss 1.9667890071868896
 Step 270: loss 2.090775966644287
 Step 280: loss 2.0678536891937256
 Step 290: loss 2.0677108764648438
 Step 300: loss 2.146937847137451
 Step 310: loss 2.0337512493133545
 Step 320: loss 2.0274972915649414
 Step 330: loss 2.0158541202545166
 Step 340: loss 2.0318241119384766
 Step 350: loss 2.039836883544922
 Step 360: loss 1.9733495712280273
 Step 370: loss 2.1069722175598145
 Step 380: loss 2.0226962566375732
 Step 390: loss 1.9445785284042358
 Step 400: loss 1.9702495336532593
 [Seed 16] Step 400 --> val loss 2.302482843399048
 Step 410: loss 2.0398097038269043
 Step 420: loss 2.062375783920288
 Step 430: loss 2.088489055633545
 Step 440: loss 1.9611563682556152
 Step 450: loss 2.0501997470855713
 Step 460: loss 2.064523220062256
 Step 470: loss 2.0337300300598145
 Step 480: loss 2.1368393898010254
 Step 490: loss 1.9391858577728271
 Step 500: loss 2.0275940895080566
 Step 510: loss 2.0312423706054688
 Step 520: loss 2.0299196243286133
 Step 530: loss 1.9690015316009521
 Step 540: loss 2.024716854095459
 Step 550: loss 1.9718669652938843
 Step 560: loss 1.8632352352142334
 Step 570: loss 1.8917312622070312
 Step 580: loss 1.9590387344360352
 Step 590: loss 1.9665831327438354
 Step 600: loss 1.9396398067474365
 [Seed 16] Step 600 --> val loss 2.2504401206970215
 Step 610: loss 1.9222739934921265
 Step 620: loss 2.1069555282592773
 Step 630: loss 1.9405336380004883
 Step 640: loss 1.9532537460327148
 Step 650: loss 2.086456775665283
 Step 660: loss 1.9928996562957764
 Step 670: loss 1.9233601093292236
 Step 680: loss 1.8628103733062744
 Step 690: loss 1.994653582572937
 Step 700: loss 1.9432777166366577
 Step 710: loss 2.028238296508789
 Step 720: loss 2.0319249629974365
 Step 730: loss 2.162468910217285
 Step 740: loss 2.090270757675171
 Step 750: loss 2.115410804748535
 Step 760: loss 2.142357110977173
 Step 770: loss 2.013491630554199
 Step 780: loss 2.110280990600586
 Step 790: loss 2.0965030193328857
 Step 800: loss 2.090848922729492
 [Seed 16] Step 800 --> val loss 2.412700653076172
 Step 810: loss 2.1023054122924805
 Step 820: loss 1.9698171615600586
 Step 830: loss 2.0569896697998047
 Step 840: loss 2.006422281265259
 Step 850: loss 2.0505576133728027
 Step 860: loss 1.9573243856430054
 Step 870: loss 2.0146684646606445
 Step 880: loss 1.9679803848266602
 Step 890: loss 1.953711986541748
 Step 900: loss 2.0147929191589355
 Step 910: loss 2.1101231575012207
 Step 920: loss 2.173464775085449
 Step 930: loss 2.0453615188598633
 Step 940: loss 2.001171827316284
 Step 950: loss 2.1543500423431396
 Step 960: loss 2.1456210613250732
 Step 970: loss 2.072094440460205
 Step 980: loss 2.1501107215881348
 Step 990: loss 1.9119523763656616
 Step 1000: loss 2.0262584686279297
 [Seed 16] Step 1000 --> val loss 2.3128669261932373
 Step 1010: loss 2.0468311309814453
 Step 1020: loss 2.1691296100616455
 Step 1030: loss 2.1117262840270996
 Step 1040: loss 1.9812787771224976
 Step 1050: loss 2.0466909408569336
 Step 1060: loss 1.9229354858398438
 Step 1070: loss 2.034447431564331
 Step 1080: loss nan
 Step 1090: loss nan
 Step 1100: loss nan
 Step 1110: loss nan
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 16] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 16] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 16] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 16] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 16] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [43  1 40 50 43 57 57 43 42  1] ...
 First target sequence: [ 1 40 50 43 57 57 43 42  1 51] ...

 Decoded input: e blessed my rude hand.
 Did my heart love till now? forswear it,

 Decoded target:  blessed my rude hand.
 Did my heart love till now? forswear it, 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.207720756530762
 [Seed 17] Step 0 --> val loss 4.1822028160095215
 Step 10: loss 3.825212001800537
 Step 20: loss 3.2948994636535645
 Step 30: loss 2.9055514335632324
 Step 40: loss 2.6977481842041016
 Step 50: loss 2.583083152770996
 Step 60: loss 2.590320110321045
 Step 70: loss 2.445499897003174
 Step 80: loss 2.383923292160034
 Step 90: loss 2.3712596893310547
 Step 100: loss 2.2914435863494873
 Step 110: loss 2.317233085632324
 Step 120: loss 2.1332931518554688
 Step 130: loss 2.311633586883545
 Step 140: loss 2.2260043621063232
 Step 150: loss 2.139308452606201
 Step 160: loss 2.3105974197387695
 Step 170: loss 2.196526050567627
 Step 180: loss 2.039271593093872
 Step 190: loss 2.2769153118133545
 Step 200: loss 2.0873544216156006
 [Seed 17] Step 200 --> val loss 2.3895769119262695
 Step 210: loss 2.2046780586242676
 Step 220: loss 2.1913533210754395
 Step 230: loss 2.0344138145446777
 Step 240: loss 2.1117351055145264
 Step 250: loss 1.9906251430511475
 Step 260: loss 1.9361231327056885
 Step 270: loss 2.047081708908081
 Step 280: loss 2.0961718559265137
 Step 290: loss 2.054084300994873
 Step 300: loss 2.0859169960021973
 Step 310: loss 2.200026512145996
 Step 320: loss 1.9426263570785522
 Step 330: loss 2.150839328765869
 Step 340: loss 1.9088139533996582
 Step 350: loss 1.9062302112579346
 Step 360: loss 2.1500823497772217
 Step 370: loss 1.992842674255371
 Step 380: loss 2.004610538482666
 Step 390: loss 2.022249698638916
 Step 400: loss 2.0313210487365723
 [Seed 17] Step 400 --> val loss 2.267409086227417
 Step 410: loss 1.9963610172271729
 Step 420: loss 2.0957226753234863
 Step 430: loss 2.0189712047576904
 Step 440: loss 2.0167741775512695
 Step 450: loss 1.999338150024414
 Step 460: loss 1.9794286489486694
 Step 470: loss 1.9820375442504883
 Step 480: loss 2.0955848693847656
 Step 490: loss 1.9259297847747803
 Step 500: loss 2.129448652267456
 Step 510: loss 1.917560338973999
 Step 520: loss 2.003030776977539
 Step 530: loss 2.029334545135498
 Step 540: loss 1.9540141820907593
 Step 550: loss 2.017599582672119
 Step 560: loss 2.0592637062072754
 Step 570: loss 1.9851975440979004
 Step 580: loss 1.9138939380645752
 Step 590: loss 1.9644383192062378
 Step 600: loss 1.9688050746917725
 [Seed 17] Step 600 --> val loss 2.254171848297119
 Step 610: loss 1.9311211109161377
 Step 620: loss 1.9866225719451904
 Step 630: loss 2.0393738746643066
 Step 640: loss 1.926295280456543
 Step 650: loss 2.0670113563537598
 Step 660: loss 2.0857934951782227
 Step 670: loss 1.9996812343597412
 Step 680: loss 1.9782756567001343
 Step 690: loss 2.0176172256469727
 Step 700: loss 1.9938459396362305
 Step 710: loss 2.0194950103759766
 Step 720: loss 1.9227561950683594
 Step 730: loss 1.886537790298462
 Step 740: loss 1.850414752960205
 Step 750: loss 1.9083890914916992
 Step 760: loss 2.0123074054718018
 Step 770: loss 1.8850784301757812
 Step 780: loss 2.0568816661834717
 Step 790: loss 1.9844090938568115
 Step 800: loss 1.9662364721298218
 [Seed 17] Step 800 --> val loss 2.238698720932007
 Step 810: loss 2.084786891937256
 Step 820: loss 2.0281991958618164
 Step 830: loss 1.9649910926818848
 Step 840: loss 1.9756273031234741
 Step 850: loss 1.9650286436080933
 Step 860: loss 1.9232137203216553
 Step 870: loss 1.9615838527679443
 Step 880: loss 2.0611705780029297
 Step 890: loss 2.0956473350524902
 Step 900: loss 2.105666160583496
 Step 910: loss 2.132922649383545
 Step 920: loss 1.909757375717163
 Step 930: loss 1.9857099056243896
 Step 940: loss 1.922879934310913
 Step 950: loss 1.9881633520126343
 Step 960: loss 1.9928383827209473
 Step 970: loss 2.012345314025879
 Step 980: loss 1.9458558559417725
 Step 990: loss 1.912521243095398
 Step 1000: loss 1.9370540380477905
 [Seed 17] Step 1000 --> val loss 2.164346218109131
 Step 1010: loss 1.9257137775421143
 Step 1020: loss 1.886033058166504
 Step 1030: loss 1.9522618055343628
 Step 1040: loss 2.0662119388580322
 Step 1050: loss 1.9151023626327515
 Step 1060: loss 1.9545156955718994
 Step 1070: loss 1.9943584203720093
 Step 1080: loss 1.9762392044067383
 Step 1090: loss 1.9689730405807495
 Step 1100: loss 1.9296326637268066
 Step 1110: loss 1.9998102188110352
 Step 1120: loss 1.954620122909546
 Step 1130: loss 1.9591014385223389
 Step 1140: loss 2.034813404083252
 Step 1150: loss 2.009154796600342
 Step 1160: loss 1.9213190078735352
 Step 1170: loss 1.8467363119125366
 Step 1180: loss 2.0788733959198
 Step 1190: loss 1.959596872329712
 Step 1200: loss 1.9115724563598633
 [Seed 17] Step 1200 --> val loss 2.2623746395111084
 Step 1210: loss 1.9212665557861328
 Step 1220: loss 2.002070665359497
 Step 1230: loss 1.8688687086105347
 Step 1240: loss 2.0145516395568848
 Step 1250: loss 1.9723138809204102
 Step 1260: loss 1.802416205406189
 Step 1270: loss 1.9629335403442383
 Step 1280: loss 1.9323649406433105
 Step 1290: loss 1.8663880825042725
 Step 1300: loss 1.9194883108139038
 Step 1310: loss 1.9179186820983887
 Step 1320: loss 1.9252521991729736
 Step 1330: loss 1.9250774383544922
 Step 1340: loss 1.9127134084701538
 Step 1350: loss 1.9177453517913818
 Step 1360: loss 1.9066108465194702
 Step 1370: loss 1.8880800008773804
 Step 1380: loss 2.001667022705078
 Step 1390: loss 1.9688286781311035
 Step 1400: loss 1.8240983486175537
 [Seed 17] Step 1400 --> val loss 2.1693625450134277
 Step 1410: loss 1.8898452520370483
 Step 1420: loss 1.9547879695892334
 Step 1430: loss 1.889565348625183
 Step 1440: loss 1.9374606609344482
 Step 1450: loss 1.8774985074996948
 Step 1460: loss 1.9716180562973022
 Step 1470: loss 1.8670909404754639
 Step 1480: loss 1.9251817464828491
 Step 1490: loss 1.9163482189178467
 Step 1500: loss 1.9030592441558838
 Step 1510: loss 1.8466752767562866
 Step 1520: loss 1.8948814868927002
 Step 1530: loss 1.8878164291381836
 Step 1540: loss 1.836984634399414
 Step 1550: loss 1.9283679723739624
 Step 1560: loss 1.835489273071289
 Step 1570: loss 1.8240598440170288
 Step 1580: loss 1.8079074621200562
 Step 1590: loss 1.8641345500946045
 Step 1600: loss 1.814220905303955
 [Seed 17] Step 1600 --> val loss 1.9914804697036743
 Step 1610: loss 1.797644019126892
 Step 1620: loss 1.9551753997802734
 Step 1630: loss 1.9155569076538086
 Step 1640: loss 1.8471860885620117
 Step 1650: loss 1.759438395500183
 Step 1660: loss 1.8673007488250732
 Step 1670: loss 1.8294134140014648
 Step 1680: loss 1.8805975914001465
 Step 1690: loss 1.829643964767456
 Step 1700: loss 1.9089434146881104
 Step 1710: loss 1.8262237310409546
 Step 1720: loss 1.8202035427093506
 Step 1730: loss 1.606139898300171
 Step 1740: loss 1.8639873266220093
 Step 1750: loss 1.8055155277252197
 Step 1760: loss 1.918163537979126
 Step 1770: loss 1.9243135452270508
 Step 1780: loss 1.7998603582382202
 Step 1790: loss 1.7151626348495483
 Step 1800: loss 1.8691753149032593
 [Seed 17] Step 1800 --> val loss 1.9528471231460571
 Step 1810: loss 1.7679179906845093
 Step 1820: loss 1.7918493747711182
 Step 1830: loss 1.6963924169540405
 Step 1840: loss 1.691033124923706
 Step 1850: loss 1.7478742599487305
 Step 1860: loss 1.7547664642333984
 Step 1870: loss 1.764203429222107
 Step 1880: loss 1.627507209777832
 Step 1890: loss 1.7486480474472046
 Step 1900: loss 1.7224664688110352
 Step 1910: loss 1.818260908126831
 Step 1920: loss 1.725428581237793
 Step 1930: loss 1.7716745138168335
 Step 1940: loss 1.803930640220642
 Step 1950: loss 1.7124847173690796
 Step 1960: loss 1.7794530391693115
 Step 1970: loss 1.8656644821166992
 Step 1980: loss 1.8429573774337769
 Step 1990: loss 1.7111657857894897
 Step 2000: loss 1.8459148406982422
 [Seed 17] Step 2000 --> val loss 1.9089670181274414
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [30 16 10  0 31 53 50 42 47 43] ...
 First target sequence: [16 10  0 31 53 50 42 47 43 56] ...

 Decoded input: RD:
 Soldiers, away with him!

 Tutor:
 Ah, Clifford, murder not th

 Decoded target: D:
 Soldiers, away with him!

 Tutor:
 Ah, Clifford, murder not thi
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.201973915100098
 [Seed 18] Step 0 --> val loss 4.182651042938232
 Step 10: loss 3.8559279441833496
 Step 20: loss 3.371493339538574
 Step 30: loss 2.9571404457092285
 Step 40: loss 2.6700007915496826
 Step 50: loss 2.687440872192383
 Step 60: loss 2.453012704849243
 Step 70: loss 2.2739977836608887
 Step 80: loss 2.49465274810791
 Step 90: loss 2.368377447128296
 Step 100: loss 2.41637921333313
 Step 110: loss 2.3981614112854004
 Step 120: loss 2.224790573120117
 Step 130: loss 2.211216926574707
 Step 140: loss 2.2254867553710938
 Step 150: loss 2.214817762374878
 Step 160: loss 2.1422247886657715
 Step 170: loss 2.2079684734344482
 Step 180: loss 2.041510581970215
 Step 190: loss 2.084311008453369
 Step 200: loss 2.1088922023773193
 [Seed 18] Step 200 --> val loss 2.3377578258514404
 Step 210: loss 2.0856146812438965
 Step 220: loss 2.1398255825042725
 Step 230: loss 2.130211114883423
 Step 240: loss 2.1695661544799805
 Step 250: loss 2.174455165863037
 Step 260: loss 2.197031021118164
 Step 270: loss 2.0692028999328613
 Step 280: loss 2.058140516281128
 Step 290: loss 2.0942695140838623
 Step 300: loss 2.0929696559906006
 Step 310: loss 2.0078413486480713
 Step 320: loss 2.008110523223877
 Step 330: loss 2.0347702503204346
 Step 340: loss 2.16618013381958
 Step 350: loss 2.050126075744629
 Step 360: loss 1.9817025661468506
 Step 370: loss 2.0223441123962402
 Step 380: loss 2.133293390274048
 Step 390: loss 1.8931896686553955
 Step 400: loss 1.9531047344207764
 [Seed 18] Step 400 --> val loss 2.1994826793670654
 Step 410: loss 2.104633331298828
 Step 420: loss 1.9747791290283203
 Step 430: loss 1.931931972503662
 Step 440: loss 1.9094533920288086
 Step 450: loss 2.069643974304199
 Step 460: loss 2.005103588104248
 Step 470: loss 1.9722527265548706
 Step 480: loss 2.039590835571289
 Step 490: loss 2.0075438022613525
 Step 500: loss 2.1527791023254395
 Step 510: loss 1.9596565961837769
 Step 520: loss 1.9899511337280273
 Step 530: loss 1.9744396209716797
 Step 540: loss 1.9105186462402344
 Step 550: loss 2.1068384647369385
 Step 560: loss 2.001955509185791
 Step 570: loss 1.9025275707244873
 Step 580: loss 2.0731074810028076
 Step 590: loss 1.8785022497177124
 Step 600: loss 1.9631927013397217
 [Seed 18] Step 600 --> val loss 2.22438645362854
 Step 610: loss 1.9603132009506226
 Step 620: loss 1.9360617399215698
 Step 630: loss 2.012730598449707
 Step 640: loss 1.8892101049423218
 Step 650: loss 2.061884880065918
 Step 660: loss 1.958001971244812
 Step 670: loss 2.0233609676361084
 Step 680: loss 1.9945971965789795
 Step 690: loss 1.889641523361206
 Step 700: loss 1.910482406616211
 Step 710: loss 1.948589563369751
 Step 720: loss 1.920603632926941
 Step 730: loss 1.9803886413574219
 Step 740: loss 1.951027274131775
 Step 750: loss 2.026343822479248
 Step 760: loss 1.8734276294708252
 Step 770: loss 2.0138916969299316
 Step 780: loss 1.9477771520614624
 Step 790: loss 1.8516420125961304
 Step 800: loss 1.9162890911102295
 [Seed 18] Step 800 --> val loss 2.2685329914093018
 Step 810: loss 1.9187841415405273
 Step 820: loss 2.0331740379333496
 Step 830: loss 1.8361451625823975
 Step 840: loss 2.0071218013763428
 Step 850: loss 2.0641584396362305
 Step 860: loss 1.9664530754089355
 Step 870: loss 2.139329195022583
 Step 880: loss 1.9844239950180054
 Step 890: loss 2.002046823501587
 Step 900: loss 1.8659976720809937
 Step 910: loss 1.9971592426300049
 Step 920: loss 1.890920639038086
 Step 930: loss 2.0585858821868896
 Step 940: loss 1.9452584981918335
 Step 950: loss 1.9198708534240723
 Step 960: loss 1.837651014328003
 Step 970: loss 1.8137636184692383
 Step 980: loss 1.9494489431381226
 Step 990: loss 2.038963794708252
 Step 1000: loss 1.9588111639022827
 [Seed 18] Step 1000 --> val loss 2.2182388305664062
 Step 1010: loss 2.087707042694092
 Step 1020: loss 2.0987396240234375
 Step 1030: loss 1.9073634147644043
 Step 1040: loss 1.8535699844360352
 Step 1050: loss 1.8020007610321045
 Step 1060: loss 1.965314269065857
 Step 1070: loss 1.9774173498153687
 Step 1080: loss 2.011178493499756
 Step 1090: loss 2.0107991695404053
 Step 1100: loss 1.921987533569336
 Step 1110: loss 1.9323863983154297
 Step 1120: loss 1.9538347721099854
 Step 1130: loss 1.958434820175171
 Step 1140: loss 1.8304927349090576
 Step 1150: loss 1.8998825550079346
 Step 1160: loss 1.922600269317627
 Step 1170: loss 1.8881057500839233
 Step 1180: loss 1.9650771617889404
 Step 1190: loss 1.9862149953842163
 Step 1200: loss 1.794732928276062
 [Seed 18] Step 1200 --> val loss 2.176720142364502
 Step 1210: loss 1.8363604545593262
 Step 1220: loss 1.7649301290512085
 Step 1230: loss 1.928767204284668
 Step 1240: loss 1.8372963666915894
 Step 1250: loss 1.8227490186691284
 Step 1260: loss 1.9451279640197754
 Step 1270: loss 1.8931996822357178
 Step 1280: loss 1.9118554592132568
 Step 1290: loss 1.9436802864074707
 Step 1300: loss 1.7719835042953491
 Step 1310: loss 1.9058215618133545
 Step 1320: loss 1.9073219299316406
 Step 1330: loss 1.7552040815353394
 Step 1340: loss 1.9305486679077148
 Step 1350: loss 1.8504003286361694
 Step 1360: loss 1.8670856952667236
 Step 1370: loss 1.8894469738006592
 Step 1380: loss 1.9270904064178467
 Step 1390: loss 1.8058509826660156
 Step 1400: loss 1.9097938537597656
 [Seed 18] Step 1400 --> val loss 2.075174331665039
 Step 1410: loss 1.7904739379882812
 Step 1420: loss 1.8743696212768555
 Step 1430: loss 1.8045127391815186
 Step 1440: loss 1.8111566305160522
 Step 1450: loss 1.7252559661865234
 Step 1460: loss 1.8290952444076538
 Step 1470: loss 1.7644670009613037
 Step 1480: loss 1.9565579891204834
 Step 1490: loss 1.673047423362732
 Step 1500: loss 1.8532130718231201
 Step 1510: loss 1.839417815208435
 Step 1520: loss 1.7732090950012207
 Step 1530: loss 1.806388258934021
 Step 1540: loss 1.6884711980819702
 Step 1550: loss 1.808741807937622
 Step 1560: loss 1.8867104053497314
 Step 1570: loss 1.6758555173873901
 Step 1580: loss 1.7667983770370483
 Step 1590: loss 1.7601455450057983
 Step 1600: loss 1.7838033437728882
 [Seed 18] Step 1600 --> val loss 1.915321946144104
 Step 1610: loss 1.7948172092437744
 Step 1620: loss 1.7954375743865967
 Step 1630: loss 1.795530080795288
 Step 1640: loss 1.7713407278060913
 Step 1650: loss 1.8883661031723022
 Step 1660: loss 1.8365768194198608
 Step 1670: loss 1.7170066833496094
 Step 1680: loss 1.7513034343719482
 Step 1690: loss 1.7737417221069336
 Step 1700: loss 1.7510578632354736
 Step 1710: loss 1.627913475036621
 Step 1720: loss 1.756230115890503
 Step 1730: loss 1.6411700248718262
 Step 1740: loss 1.7595919370651245
 Step 1750: loss 1.7149779796600342
 Step 1760: loss 1.771960973739624
 Step 1770: loss 1.6450022459030151
 Step 1780: loss 1.748437762260437
 Step 1790: loss 1.7457375526428223
 Step 1800: loss 1.6532361507415771
 [Seed 18] Step 1800 --> val loss 1.840130090713501
 Step 1810: loss 1.7655004262924194
 Step 1820: loss 1.671708583831787
 Step 1830: loss 1.74502694606781
 Step 1840: loss 1.7797203063964844
 Step 1850: loss 1.6858251094818115
 Step 1860: loss 1.6646236181259155
 Step 1870: loss 1.6174286603927612
 Step 1880: loss 1.6855666637420654
 Step 1890: loss 1.744685173034668
 Step 1900: loss 1.6830857992172241
 Step 1910: loss 1.6286227703094482
 Step 1920: loss 1.7465176582336426
 Step 1930: loss 1.737027883529663
 Step 1940: loss 1.643978476524353
 Step 1950: loss 1.6980247497558594
 Step 1960: loss 1.6033920049667358
 Step 1970: loss 1.7040470838546753
 Step 1980: loss 1.8122987747192383
 Step 1990: loss 1.656650424003601
 Step 2000: loss 1.7209466695785522
 [Seed 18] Step 2000 --> val loss 1.7891883850097656
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [56 57 43  2  1 59 56 45 43  1] ...
 First target sequence: [57 43  2  1 59 56 45 43  1 42] ...

 Decoded input: rse! urge doubts to them that fear.

 LORD WILLOUGHBY:
 Hold out m

 Decoded target: se! urge doubts to them that fear.

 LORD WILLOUGHBY:
 Hold out my
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.228902816772461
 [Seed 19] Step 0 --> val loss 4.183101177215576
 Step 10: loss 3.857027769088745
 Step 20: loss 3.2869632244110107
 Step 30: loss 2.8034133911132812
 Step 40: loss 2.704606533050537
 Step 50: loss 2.5742719173431396
 Step 60: loss 2.5015971660614014
 Step 70: loss 2.382699966430664
 Step 80: loss 2.3677291870117188
 Step 90: loss 2.2841243743896484
 Step 100: loss 2.326866626739502
 Step 110: loss 2.253241777420044
 Step 120: loss 2.27793288230896
 Step 130: loss 2.145003318786621
 Step 140: loss 2.2450687885284424
 Step 150: loss 2.1568381786346436
 Step 160: loss 2.2351808547973633
 Step 170: loss 2.0940053462982178
 Step 180: loss 2.127530336380005
 Step 190: loss 2.2044291496276855
 Step 200: loss 2.1375620365142822
 [Seed 19] Step 200 --> val loss 2.3360910415649414
 Step 210: loss 2.164816379547119
 Step 220: loss 2.1131019592285156
 Step 230: loss 1.9668607711791992
 Step 240: loss 2.006270170211792
 Step 250: loss 2.031968116760254
 Step 260: loss 2.0675735473632812
 Step 270: loss 1.9747213125228882
 Step 280: loss 1.9816540479660034
 Step 290: loss 2.1002070903778076
 Step 300: loss 2.1028904914855957
 Step 310: loss 2.006779193878174
 Step 320: loss 2.106187582015991
 Step 330: loss 2.0273189544677734
 Step 340: loss 2.0348052978515625
 Step 350: loss 1.8540453910827637
 Step 360: loss 2.063291072845459
 Step 370: loss 2.062554121017456
 Step 380: loss 1.948988437652588
 Step 390: loss 2.0062785148620605
 Step 400: loss 2.1152257919311523
 [Seed 19] Step 400 --> val loss 2.281575918197632
 Step 410: loss 2.069434642791748
 Step 420: loss 2.0642952919006348
 Step 430: loss 2.0509984493255615
 Step 440: loss 1.9573768377304077
 Step 450: loss 1.9020789861679077
 Step 460: loss 2.047640562057495
 Step 470: loss 1.9930700063705444
 Step 480: loss 1.9402350187301636
 Step 490: loss 2.0050582885742188
 Step 500: loss 2.007336139678955
 Step 510: loss 1.9493526220321655
 Step 520: loss 2.0202417373657227
 Step 530: loss 2.02510929107666
 Step 540: loss 2.035984992980957
 Step 550: loss 2.0120837688446045
 Step 560: loss 2.0263900756835938
 Step 570: loss 1.9528272151947021
 Step 580: loss 2.010615348815918
 Step 590: loss 1.973067045211792
 Step 600: loss 2.024765729904175
 [Seed 19] Step 600 --> val loss 2.1155471801757812
 Step 610: loss 2.114974021911621
 Step 620: loss 1.9202378988265991
 Step 630: loss 2.024564266204834
 Step 640: loss 1.8874599933624268
 Step 650: loss 1.8926019668579102
 Step 660: loss 1.8711998462677002
 Step 670: loss 1.901045322418213
 Step 680: loss 1.9679133892059326
 Step 690: loss 1.9051283597946167
 Step 700: loss 1.9041951894760132
 Step 710: loss 1.9393043518066406
 Step 720: loss 1.994063138961792
 Step 730: loss 1.8176391124725342
 Step 740: loss 1.9367945194244385
 Step 750: loss 1.87576425075531
 Step 760: loss 2.0132176876068115
 Step 770: loss 1.8670543432235718
 Step 780: loss 1.9063539505004883
 Step 790: loss 1.9746145009994507
 Step 800: loss 1.9818700551986694
 [Seed 19] Step 800 --> val loss 2.2954888343811035
 Step 810: loss 1.9767574071884155
 Step 820: loss 2.0176563262939453
 Step 830: loss 2.0024516582489014
 Step 840: loss 1.93024480342865
 Step 850: loss 1.921354055404663
 Step 860: loss 1.8505661487579346
 Step 870: loss 1.953228235244751
 Step 880: loss 1.9232478141784668
 Step 890: loss 1.9488294124603271
 Step 900: loss 1.8983516693115234
 Step 910: loss 1.8675963878631592
 Step 920: loss 1.8672107458114624
 Step 930: loss 1.7940181493759155
 Step 940: loss 1.8774149417877197
 Step 950: loss 1.909509539604187
 Step 960: loss 1.8804137706756592
 Step 970: loss 1.7677005529403687
 Step 980: loss 1.9177954196929932
 Step 990: loss 1.8956111669540405
 Step 1000: loss 1.8695285320281982
 [Seed 19] Step 1000 --> val loss 2.250347852706909
 Step 1010: loss 1.9703807830810547
 Step 1020: loss 1.8856725692749023
 Step 1030: loss 1.966089129447937
 Step 1040: loss 1.9786288738250732
 Step 1050: loss 1.8978114128112793
 Step 1060: loss 1.9586331844329834
 Step 1070: loss 2.033046007156372
 Step 1080: loss 1.896769404411316
 Step 1090: loss 1.840436339378357
 Step 1100: loss 1.9333925247192383
 Step 1110: loss 1.9019745588302612
 Step 1120: loss 1.9107682704925537
 Step 1130: loss 1.9665277004241943
 Step 1140: loss 1.8743135929107666
 Step 1150: loss 1.836340308189392
 Step 1160: loss 1.944187879562378
 Step 1170: loss 1.8917319774627686
 Step 1180: loss 1.9258267879486084
 Step 1190: loss 1.8544895648956299
 Step 1200: loss 1.9524787664413452
 [Seed 19] Step 1200 --> val loss 2.218228578567505
 Step 1210: loss 1.8824537992477417
 Step 1220: loss 1.8615694046020508
 Step 1230: loss 1.8213329315185547
 Step 1240: loss 1.962876558303833
 Step 1250: loss 1.9040559530258179
 Step 1260: loss 1.8413670063018799
 Step 1270: loss 1.879677653312683
 Step 1280: loss 1.8362401723861694
 Step 1290: loss 1.8826450109481812
 Step 1300: loss 1.8195505142211914
 Step 1310: loss 1.8506078720092773
 Step 1320: loss 1.8485870361328125
 Step 1330: loss 1.8865537643432617
 Step 1340: loss 1.983703374862671
 Step 1350: loss 1.8383510112762451
 Step 1360: loss 1.8940210342407227
 Step 1370: loss 1.8833296298980713
 Step 1380: loss 1.817777156829834
 Step 1390: loss 1.7207520008087158
 Step 1400: loss 1.8218955993652344
 [Seed 19] Step 1400 --> val loss 2.0705363750457764
 Step 1410: loss 1.7929599285125732
 Step 1420: loss 1.8464314937591553
 Step 1430: loss 1.880921721458435
 Step 1440: loss 1.6977323293685913
 Step 1450: loss 1.8668272495269775
 Step 1460: loss 1.7985963821411133
 Step 1470: loss 1.8661073446273804
 Step 1480: loss 1.7190488576889038
 Step 1490: loss 1.8191083669662476
 Step 1500: loss 1.6479198932647705
 Step 1510: loss 1.8058745861053467
 Step 1520: loss 1.8343852758407593
 Step 1530: loss 1.7680294513702393
 Step 1540: loss 1.7260122299194336
 Step 1550: loss 1.8563728332519531
 Step 1560: loss 1.6671626567840576
 Step 1570: loss 1.8133624792099
 Step 1580: loss 1.7406920194625854
 Step 1590: loss 1.785808801651001
 Step 1600: loss 1.8991718292236328
 [Seed 19] Step 1600 --> val loss 1.861488938331604
 Step 1610: loss 1.5852231979370117
 Step 1620: loss 1.7703089714050293
 Step 1630: loss 1.7837321758270264
 Step 1640: loss 1.7545922994613647
 Step 1650: loss 1.774943232536316
 Step 1660: loss 1.6316341161727905
 Step 1670: loss 1.8214110136032104
 Step 1680: loss 1.7874469757080078
 Step 1690: loss 1.7467780113220215
 Step 1700: loss 1.708043098449707
 Step 1710: loss 1.583945631980896
 Step 1720: loss 1.6098417043685913
 Step 1730: loss 1.6890109777450562
 Step 1740: loss 1.7671630382537842
 Step 1750: loss 1.6482480764389038
 Step 1760: loss 1.6347541809082031
 Step 1770: loss 1.80223548412323
 Step 1780: loss 1.7399849891662598
 Step 1790: loss 1.6628321409225464
 Step 1800: loss 1.7714104652404785
 [Seed 19] Step 1800 --> val loss 1.818497657775879
 Step 1810: loss 1.6728146076202393
 Step 1820: loss 1.7558002471923828
 Step 1830: loss 1.6565284729003906
 Step 1840: loss 1.6646583080291748
 Step 1850: loss 1.7009962797164917
 Step 1860: loss 1.7051069736480713
 Step 1870: loss 1.6000735759735107
 Step 1880: loss 1.7262495756149292
 Step 1890: loss 1.6113831996917725
 Step 1900: loss 1.6445519924163818
 Step 1910: loss 1.682339072227478
 Step 1920: loss 1.6693767309188843
 Step 1930: loss 1.6996912956237793
 Step 1940: loss 1.6468117237091064
 Step 1950: loss 1.6264451742172241
 Step 1960: loss 1.726996660232544
 Step 1970: loss 1.6935155391693115
 Step 1980: loss 1.7492332458496094
 Step 1990: loss 1.6397756338119507
 Step 2000: loss 1.5989396572113037
 [Seed 19] Step 2000 --> val loss 1.7990573644638062
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [58 46 43 56  8  0  0 16 27 30] ...
 First target sequence: [46 43 56  8  0  0 16 27 30 15] ...

 Decoded input: ther.

 DORCAS:
 Thou hast sworn my love to be.

 MOPSA:
 Thou hast 

 Decoded target: her.

 DORCAS:
 Thou hast sworn my love to be.

 MOPSA:
 Thou hast s
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.192948818206787
 [Seed 20] Step 0 --> val loss 4.185471057891846
 Step 10: loss 3.833178997039795
 Step 20: loss 3.314063549041748
 Step 30: loss 2.880687713623047
 Step 40: loss 2.698762893676758
 Step 50: loss 2.5499134063720703
 Step 60: loss 2.537269115447998
 Step 70: loss 2.484443426132202
 Step 80: loss 2.3656229972839355
 Step 90: loss 2.3631787300109863
 Step 100: loss 2.374593734741211
 Step 110: loss 2.233882427215576
 Step 120: loss 2.258683204650879
 Step 130: loss 2.2500624656677246
 Step 140: loss 2.267636299133301
 Step 150: loss 2.137681007385254
 Step 160: loss 2.1589884757995605
 Step 170: loss 2.1020824909210205
 Step 180: loss 2.134371757507324
 Step 190: loss 2.1717867851257324
 Step 200: loss 2.068070411682129
 [Seed 20] Step 200 --> val loss 2.371032476425171
 Step 210: loss 2.095829486846924
 Step 220: loss 2.0954418182373047
 Step 230: loss 2.022787094116211
 Step 240: loss 1.97525155544281
 Step 250: loss 2.076366424560547
 Step 260: loss 2.1445226669311523
 Step 270: loss 2.079353094100952
 Step 280: loss 1.9375522136688232
 Step 290: loss 2.064016342163086
 Step 300: loss 2.0315356254577637
 Step 310: loss 2.0616586208343506
 Step 320: loss 2.1605734825134277
 Step 330: loss 1.9693034887313843
 Step 340: loss 2.025357723236084
 Step 350: loss 1.9782047271728516
 Step 360: loss 2.035069465637207
 Step 370: loss 1.994199275970459
 Step 380: loss 1.9936422109603882
 Step 390: loss 2.158062696456909
 Step 400: loss 2.0593504905700684
 [Seed 20] Step 400 --> val loss 2.2373604774475098
 Step 410: loss 1.9361493587493896
 Step 420: loss 2.0181283950805664
 Step 430: loss 2.116698741912842
 Step 440: loss 1.960148811340332
 Step 450: loss 1.910308599472046
 Step 460: loss 1.9789133071899414
 Step 470: loss 2.021726131439209
 Step 480: loss 1.9752992391586304
 Step 490: loss 1.954584002494812
 Step 500: loss 2.040011167526245
 Step 510: loss 1.9962396621704102
 Step 520: loss 1.996194839477539
 Step 530: loss 1.9831621646881104
 Step 540: loss 2.0618643760681152
 Step 550: loss 2.0236594676971436
 Step 560: loss 2.0403308868408203
 Step 570: loss 2.0716116428375244
 Step 580: loss 2.0549235343933105
 Step 590: loss 2.1221718788146973
 Step 600: loss 1.9096254110336304
 [Seed 20] Step 600 --> val loss 2.355912208557129
 Step 610: loss 2.0623059272766113
 Step 620: loss 2.018789291381836
 Step 630: loss 1.9338268041610718
 Step 640: loss 2.082034111022949
 Step 650: loss 2.043639659881592
 Step 660: loss 2.0794105529785156
 Step 670: loss 1.9518229961395264
 Step 680: loss 1.9318907260894775
 Step 690: loss 1.9601033926010132
 Step 700: loss 1.9574906826019287
 Step 710: loss 2.0269675254821777
 Step 720: loss 2.111104965209961
 Step 730: loss 2.0063867568969727
 Step 740: loss 2.0423200130462646
 Step 750: loss 1.9946458339691162
 Step 760: loss 2.0846598148345947
 Step 770: loss 2.072528600692749
 Step 780: loss 2.0069758892059326
 Step 790: loss 2.1151905059814453
 Step 800: loss 2.0473732948303223
 [Seed 20] Step 800 --> val loss 2.263268232345581
 Step 810: loss 1.9817233085632324
 Step 820: loss 1.9834916591644287
 Step 830: loss 1.9832656383514404
 Step 840: loss 2.1282877922058105
 Step 850: loss 1.9409496784210205
 Step 860: loss 2.0453615188598633
 Step 870: loss 1.9903687238693237
 Step 880: loss 2.03373122215271
 Step 890: loss 2.1471595764160156
 Step 900: loss 2.1843342781066895
 Step 910: loss 2.0989747047424316
 Step 920: loss 2.0775020122528076
 Step 930: loss 2.080778121948242
 Step 940: loss 2.161600112915039
 Step 950: loss 1.9695323705673218
 Step 960: loss 1.918921709060669
 Step 970: loss 2.0302090644836426
 Step 980: loss 1.9500659704208374
 Step 990: loss 2.012192487716675
 Step 1000: loss 1.9656057357788086
 [Seed 20] Step 1000 --> val loss 2.319161891937256
 Step 1010: loss 1.9580163955688477
 Step 1020: loss 1.8998637199401855
 Step 1030: loss 1.969194769859314
 Step 1040: loss 1.935187816619873
 Step 1050: loss 1.9625518321990967
 Step 1060: loss 1.9416699409484863
 Step 1070: loss 1.9686670303344727
 Step 1080: loss 1.9858648777008057
 Step 1090: loss 1.844529390335083
 Step 1100: loss 1.8800691366195679
 Step 1110: loss 1.9007930755615234
 Step 1120: loss 1.9812040328979492
 Step 1130: loss 1.8634603023529053
 Step 1140: loss 1.9002296924591064
 Step 1150: loss 1.8441455364227295
 Step 1160: loss 1.8598359823226929
 Step 1170: loss 1.8226983547210693
 Step 1180: loss 1.9391124248504639
 Step 1190: loss 1.841726303100586
 Step 1200: loss 1.8317584991455078
 [Seed 20] Step 1200 --> val loss 2.0371458530426025
 Step 1210: loss 1.8686645030975342
 Step 1220: loss 1.8360791206359863
 Step 1230: loss 1.998282790184021
 Step 1240: loss 1.7807912826538086
 Step 1250: loss 1.8328557014465332
 Step 1260: loss 1.8432121276855469
 Step 1270: loss 1.8679860830307007
 Step 1280: loss 1.819883108139038
 Step 1290: loss 1.7965211868286133
 Step 1300: loss 1.8111352920532227
 Step 1310: loss 1.8660478591918945
 Step 1320: loss 1.7358267307281494
 Step 1330: loss 1.8669981956481934
 Step 1340: loss 1.9581668376922607
 Step 1350: loss 1.8380489349365234
 Step 1360: loss 1.8021752834320068
 Step 1370: loss 1.827453851699829
 Step 1380: loss 1.8092668056488037
 Step 1390: loss 1.8060920238494873
 Step 1400: loss 1.8202693462371826
 [Seed 20] Step 1400 --> val loss 2.055490732192993
 Step 1410: loss 1.8084758520126343
 Step 1420: loss 1.7887554168701172
 Step 1430: loss 1.910176157951355
 Step 1440: loss 1.771327257156372
 Step 1450: loss 1.7891998291015625
 Step 1460: loss 1.8408448696136475
 Step 1470: loss 1.85977303981781
 Step 1480: loss 1.756716012954712
 Step 1490: loss 1.8204835653305054
 Step 1500: loss 1.9479312896728516
 Step 1510: loss 1.5762381553649902
 Step 1520: loss 1.831925392150879
 Step 1530: loss 1.8667411804199219
 Step 1540: loss 1.8103883266448975
 Step 1550: loss 1.7465472221374512
 Step 1560: loss 1.8010495901107788
 Step 1570: loss 1.788618564605713
 Step 1580: loss 1.7157472372055054
 Step 1590: loss 1.7549197673797607
 Step 1600: loss 1.790279746055603
 [Seed 20] Step 1600 --> val loss 1.9676817655563354
 Step 1610: loss 1.9475929737091064
 Step 1620: loss 1.7733075618743896
 Step 1630: loss 1.7903159856796265
 Step 1640: loss 1.705488920211792
 Step 1650: loss 1.808363914489746
 Step 1660: loss 1.836688756942749
 Step 1670: loss 1.7925981283187866
 Step 1680: loss 1.7522914409637451
 Step 1690: loss 1.7777774333953857
 Step 1700: loss 1.6655677556991577
 Step 1710: loss 1.6437017917633057
 Step 1720: loss 1.6740617752075195
 Step 1730: loss 1.6993826627731323
 Step 1740: loss 1.7712502479553223
 Step 1750: loss 1.6342109441757202
 Step 1760: loss 1.6677783727645874
 Step 1770: loss 1.675667643547058
 Step 1780: loss 1.8069121837615967
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 20] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 20] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 2  1 61 46 39 58  1 41 39 52] ...
 First target sequence: [ 1 61 46 39 58  1 41 39 52 52] ...

 Decoded input: ! what cannot be avoided
 'Twere childish weakness to lament or f

 Decoded target:  what cannot be avoided
 'Twere childish weakness to lament or fe
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.197821617126465
 [Seed 21] Step 0 --> val loss 4.18304967880249
 Step 10: loss 3.854599952697754
 Step 20: loss 3.322859287261963
 Step 30: loss 2.9111685752868652
 Step 40: loss 2.594132423400879
 Step 50: loss 2.5631473064422607
 Step 60: loss 2.453070640563965
 Step 70: loss 2.4598255157470703
 Step 80: loss 2.4456992149353027
 Step 90: loss 2.251591682434082
 Step 100: loss 2.3401618003845215
 Step 110: loss 2.344680070877075
 Step 120: loss 2.235297679901123
 Step 130: loss 2.223472833633423
 Step 140: loss 2.2088866233825684
 Step 150: loss 2.1781938076019287
 Step 160: loss 2.189465045928955
 Step 170: loss 2.1884689331054688
 Step 180: loss 2.1644206047058105
 Step 190: loss 2.129985809326172
 Step 200: loss 2.131906032562256
 [Seed 21] Step 200 --> val loss 2.289029598236084
 Step 210: loss 2.0974721908569336
 Step 220: loss 2.138789176940918
 Step 230: loss 2.116062641143799
 Step 240: loss 2.0908946990966797
 Step 250: loss 2.1564507484436035
 Step 260: loss 2.1214218139648438
 Step 270: loss 2.088474988937378
 Step 280: loss 1.9711320400238037
 Step 290: loss 2.058366298675537
 Step 300: loss 1.9361575841903687
 Step 310: loss 2.058335781097412
 Step 320: loss 2.073930263519287
 Step 330: loss 2.0723953247070312
 Step 340: loss 2.0669610500335693
 Step 350: loss 1.9968284368515015
 Step 360: loss 2.000171422958374
 Step 370: loss 1.9942612648010254
 Step 380: loss 2.0344173908233643
 Step 390: loss 1.995328426361084
 Step 400: loss 1.9990873336791992
 [Seed 21] Step 400 --> val loss 2.264442205429077
 Step 410: loss 2.0551888942718506
 Step 420: loss 2.011279821395874
 Step 430: loss 1.909097671508789
 Step 440: loss 2.1139955520629883
 Step 450: loss 1.8861626386642456
 Step 460: loss 1.8833073377609253
 Step 470: loss 1.9498142004013062
 Step 480: loss 2.0687055587768555
 Step 490: loss 1.9990363121032715
 Step 500: loss 2.0293638706207275
 Step 510: loss 1.9548405408859253
 Step 520: loss 2.076455593109131
 Step 530: loss 1.8200922012329102
 Step 540: loss 2.0478498935699463
 Step 550: loss 2.01128888130188
 Step 560: loss 2.0101280212402344
 Step 570: loss 1.882497787475586
 Step 580: loss 2.0712242126464844
 Step 590: loss 2.067772626876831
 Step 600: loss 1.9868030548095703
 [Seed 21] Step 600 --> val loss 2.302229642868042
 Step 610: loss 1.9617111682891846
 Step 620: loss 1.9555778503417969
 Step 630: loss 2.024235248565674
 Step 640: loss 1.9634298086166382
 Step 650: loss 1.9691059589385986
 Step 660: loss 2.0065507888793945
 Step 670: loss 2.00221586227417
 Step 680: loss 2.048290252685547
 Step 690: loss 1.9682215452194214
 Step 700: loss 1.9644845724105835
 Step 710: loss 1.9354922771453857
 Step 720: loss 1.9725245237350464
 Step 730: loss 1.943900465965271
 Step 740: loss 1.815183401107788
 Step 750: loss 1.9654663801193237
 Step 760: loss 1.9095897674560547
 Step 770: loss 2.040308952331543
 Step 780: loss 1.9216350317001343
 Step 790: loss 1.8903993368148804
 Step 800: loss 2.024090528488159
 [Seed 21] Step 800 --> val loss 2.2396607398986816
 Step 810: loss 1.9480633735656738
 Step 820: loss 1.992990493774414
 Step 830: loss 2.064680576324463
 Step 840: loss 2.0335941314697266
 Step 850: loss 1.971047282218933
 Step 860: loss 1.9517344236373901
 Step 870: loss 1.9372739791870117
 Step 880: loss 1.9748458862304688
 Step 890: loss 2.068589210510254
 Step 900: loss 1.95808744430542
 Step 910: loss 2.037060022354126
 Step 920: loss 1.9031380414962769
 Step 930: loss 1.9378316402435303
 Step 940: loss 1.8352006673812866
 Step 950: loss 1.9888851642608643
 Step 960: loss 1.9270141124725342
 Step 970: loss 1.9213393926620483
 Step 980: loss 2.010509967803955
 Step 990: loss 1.9302937984466553
 Step 1000: loss 1.8961386680603027
 [Seed 21] Step 1000 --> val loss 2.291008710861206
 Step 1010: loss 1.8619132041931152
 Step 1020: loss 1.9579970836639404
 Step 1030: loss 2.0365118980407715
 Step 1040: loss 1.9120874404907227
 Step 1050: loss 2.040639638900757
 Step 1060: loss 1.8731626272201538
 Step 1070: loss 1.9482154846191406
 Step 1080: loss 2.0876970291137695
 Step 1090: loss 1.9676313400268555
 Step 1100: loss 1.8026034832000732
 Step 1110: loss 1.950956106185913
 Step 1120: loss 1.933773398399353
 Step 1130: loss 1.8096873760223389
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 21] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 21] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 21] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 21] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 21] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [61 46 39 58  1 42 43 39 58 46] ...
 First target sequence: [46 39 58  1 42 43 39 58 46  1] ...

 Decoded input: what death is doing.

 LEONTES:
 Take her hence:
 Her heart is but 

 Decoded target: hat death is doing.

 LEONTES:
 Take her hence:
 Her heart is but o
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.209897041320801
 [Seed 22] Step 0 --> val loss 4.1852335929870605
 Step 10: loss 3.8337931632995605
 Step 20: loss 3.2824485301971436
 Step 30: loss 2.837507963180542
 Step 40: loss 2.6355419158935547
 Step 50: loss 2.570739269256592
 Step 60: loss 2.48781156539917
 Step 70: loss 2.497781753540039
 Step 80: loss 2.331033229827881
 Step 90: loss 2.286808967590332
 Step 100: loss 2.246453046798706
 Step 110: loss 2.297933340072632
 Step 120: loss 2.158897876739502
 Step 130: loss 2.1946206092834473
 Step 140: loss 2.0674071311950684
 Step 150: loss 2.1663475036621094
 Step 160: loss 2.160388708114624
 Step 170: loss 2.093724012374878
 Step 180: loss 2.08231782913208
 Step 190: loss 2.21420955657959
 Step 200: loss 2.261220932006836
 [Seed 22] Step 200 --> val loss 2.3396995067596436
 Step 210: loss 2.1964211463928223
 Step 220: loss 2.0835752487182617
 Step 230: loss 2.100456714630127
 Step 240: loss 1.9835495948791504
 Step 250: loss 2.124386787414551
 Step 260: loss 2.095249891281128
 Step 270: loss 2.041208028793335
 Step 280: loss 2.1359615325927734
 Step 290: loss 2.1368398666381836
 Step 300: loss 2.019376516342163
 Step 310: loss 2.016347646713257
 Step 320: loss 1.955849051475525
 Step 330: loss 1.9997987747192383
 Step 340: loss 2.1161179542541504
 Step 350: loss 1.999868392944336
 Step 360: loss 1.9378799200057983
 Step 370: loss 1.9562005996704102
 Step 380: loss 2.0107858180999756
 Step 390: loss 2.0204248428344727
 Step 400: loss 1.9485915899276733
 [Seed 22] Step 400 --> val loss 2.268825054168701
 Step 410: loss 1.9721002578735352
 Step 420: loss 1.9277809858322144
 Step 430: loss 2.039571762084961
 Step 440: loss 2.010998010635376
 Step 450: loss 2.031942844390869
 Step 460: loss 2.150153398513794
 Step 470: loss 2.0200672149658203
 Step 480: loss 2.0049147605895996
 Step 490: loss 1.9808251857757568
 Step 500: loss 1.9718049764633179
 Step 510: loss 1.9553083181381226
 Step 520: loss 1.8852765560150146
 Step 530: loss 1.9932483434677124
 Step 540: loss 1.9338983297348022
 Step 550: loss 2.132908344268799
 Step 560: loss 1.9481757879257202
 Step 570: loss 1.8913778066635132
 Step 580: loss 1.973149299621582
 Step 590: loss 1.8960916996002197
 Step 600: loss 2.002631187438965
 [Seed 22] Step 600 --> val loss 2.3007428646087646
 Step 610: loss 1.9438831806182861
 Step 620: loss 1.946113109588623
 Step 630: loss 1.9792718887329102
 Step 640: loss 1.9425815343856812
 Step 650: loss 2.052722692489624
 Step 660: loss 1.9222487211227417
 Step 670: loss 1.877441167831421
 Step 680: loss 1.9419620037078857
 Step 690: loss 2.0223026275634766
 Step 700: loss 1.9071893692016602
 Step 710: loss 2.0414559841156006
 Step 720: loss 1.97528076171875
 Step 730: loss 2.112973928451538
 Step 740: loss 2.012659788131714
 Step 750: loss 1.8827524185180664
 Step 760: loss 2.0255560874938965
 Step 770: loss 1.9554904699325562
 Step 780: loss 2.0583512783050537
 Step 790: loss 1.8797954320907593
 Step 800: loss 1.9612698554992676
 [Seed 22] Step 800 --> val loss 2.311732053756714
 Step 810: loss 2.0043206214904785
 Step 820: loss 2.04706072807312
 Step 830: loss 2.034139633178711
 Step 840: loss 1.9441421031951904
 Step 850: loss 1.9108333587646484
 Step 860: loss 1.9549939632415771
 Step 870: loss 1.9942611455917358
 Step 880: loss 1.833571195602417
 Step 890: loss 1.862609624862671
 Step 900: loss 1.9056308269500732
 Step 910: loss 1.7789969444274902
 Step 920: loss 2.0181660652160645
 Step 930: loss 2.0011508464813232
 Step 940: loss 1.864282488822937
 Step 950: loss 1.8881019353866577
 Step 960: loss 1.801431655883789
 Step 970: loss 1.8449857234954834
 Step 980: loss 1.849437952041626
 Step 990: loss 1.8144820928573608
 Step 1000: loss 1.9646180868148804
 [Seed 22] Step 1000 --> val loss 2.1647865772247314
 Step 1010: loss 2.023458242416382
 Step 1020: loss 1.9172706604003906
 Step 1030: loss 2.0369255542755127
 Step 1040: loss 2.0015788078308105
 Step 1050: loss 1.892240285873413
 Step 1060: loss 2.022869110107422
 Step 1070: loss 1.9157333374023438
 Step 1080: loss 1.8527196645736694
 Step 1090: loss 2.037898540496826
 Step 1100: loss 1.9590802192687988
 Step 1110: loss 2.0503392219543457
 Step 1120: loss 1.948500394821167
 Step 1130: loss 2.006967544555664
 Step 1140: loss 2.0194032192230225
 Step 1150: loss 1.8610570430755615
 Step 1160: loss 1.871907114982605
 Step 1170: loss 1.7893047332763672
 Step 1180: loss 1.9537467956542969
 Step 1190: loss 1.8434690237045288
 Step 1200: loss 1.9167263507843018
 [Seed 22] Step 1200 --> val loss 2.092834949493408
 Step 1210: loss 2.024885654449463
 Step 1220: loss 1.8425029516220093
 Step 1230: loss 1.922375202178955
 Step 1240: loss 1.92521333694458
 Step 1250: loss 1.9093854427337646
 Step 1260: loss 1.919032096862793
 Step 1270: loss 1.87713623046875
 Step 1280: loss 1.8364980220794678
 Step 1290: loss 1.965212106704712
 Step 1300: loss 1.9225099086761475
 Step 1310: loss 1.9872125387191772
 Step 1320: loss 1.855303168296814
 Step 1330: loss 1.8399572372436523
 Step 1340: loss 2.077082872390747
 Step 1350: loss 1.9421542882919312
 Step 1360: loss 1.8704299926757812
 Step 1370: loss 1.8375029563903809
 Step 1380: loss 1.9863784313201904
 Step 1390: loss 1.878409743309021
 Step 1400: loss 1.9494812488555908
 [Seed 22] Step 1400 --> val loss 2.0573196411132812
 Step 1410: loss 1.9172844886779785
 Step 1420: loss 1.7981996536254883
 Step 1430: loss 1.8910157680511475
 Step 1440: loss 1.879813551902771
 Step 1450: loss 1.8679847717285156
 Step 1460: loss 1.8160755634307861
 Step 1470: loss 1.8509492874145508
 Step 1480: loss 1.8158563375473022
 Step 1490: loss 1.845076560974121
 Step 1500: loss 1.8555212020874023
 Step 1510: loss 1.6872498989105225
 Step 1520: loss 1.8225455284118652
 Step 1530: loss 1.92992103099823
 Step 1540: loss 1.7379096746444702
 Step 1550: loss 1.7520358562469482
 Step 1560: loss 1.8082070350646973
 Step 1570: loss 1.6578357219696045
 Step 1580: loss 1.81086003780365
 Step 1590: loss 1.7533291578292847
 Step 1600: loss 1.8155559301376343
 [Seed 22] Step 1600 --> val loss 2.0089385509490967
 Step 1610: loss 1.6765574216842651
 Step 1620: loss 1.9429558515548706
 Step 1630: loss 1.7825517654418945
 Step 1640: loss 1.941828727722168
 Step 1650: loss 1.62376868724823
 Step 1660: loss 1.9157207012176514
 Step 1670: loss 1.7952340841293335
 Step 1680: loss 1.8231048583984375
 Step 1690: loss 1.7736289501190186
 Step 1700: loss 1.7057689428329468
 Step 1710: loss 1.8240575790405273
 Step 1720: loss 1.7228749990463257
 Step 1730: loss 1.714529037475586
 Step 1740: loss 1.7058436870574951
 Step 1750: loss 1.6710152626037598
 Step 1760: loss 1.5906403064727783
 Step 1770: loss 1.7206329107284546
 Step 1780: loss 1.655883550643921
 Step 1790: loss 1.6642580032348633
 Step 1800: loss 1.7690443992614746
 [Seed 22] Step 1800 --> val loss 1.9302656650543213
 Step 1810: loss 1.678815245628357
 Step 1820: loss 1.6842527389526367
 Step 1830: loss 1.7626934051513672
 Step 1840: loss 1.6849801540374756
 Step 1850: loss 1.676241397857666
 Step 1860: loss 1.640634298324585
 Step 1870: loss 1.7358214855194092
 Step 1880: loss 1.8073244094848633
 Step 1890: loss 1.7501949071884155
 Step 1900: loss 1.8037593364715576
 Step 1910: loss 1.6468838453292847
 Step 1920: loss 1.7404141426086426
 Step 1930: loss 1.6161344051361084
 Step 1940: loss 1.7649452686309814
 Step 1950: loss 1.7106908559799194
 Step 1960: loss 1.606760859489441
 Step 1970: loss 1.6707080602645874
 Step 1980: loss 1.6967287063598633
 Step 1990: loss 1.756812572479248
 Step 2000: loss 1.6441130638122559
 [Seed 22] Step 2000 --> val loss 1.832013726234436
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [19 14 30 27 23 17 10  0 20 39] ...
 First target sequence: [14 30 27 23 17 10  0 20 39 60] ...

 Decoded input: GBROKE:
 Have thy desire.

 DUKE OF YORK:

 HENRY BOLINGBROKE:
 Vill

 Decoded target: BROKE:
 Have thy desire.

 DUKE OF YORK:

 HENRY BOLINGBROKE:
 Villa
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.210999965667725
 [Seed 23] Step 0 --> val loss 4.185616970062256
 Step 10: loss 3.8470211029052734
 Step 20: loss 3.3525397777557373
 Step 30: loss 2.79919695854187
 Step 40: loss 2.598353147506714
 Step 50: loss 2.590519905090332
 Step 60: loss 2.525052785873413
 Step 70: loss 2.4222800731658936
 Step 80: loss 2.3531298637390137
 Step 90: loss 2.3287792205810547
 Step 100: loss 2.2922444343566895
 Step 110: loss 2.274582624435425
 Step 120: loss 2.258537769317627
 Step 130: loss 2.237440586090088
 Step 140: loss 2.1656718254089355
 Step 150: loss 2.2198486328125
 Step 160: loss 2.194366693496704
 Step 170: loss 2.2029895782470703
 Step 180: loss 2.205169200897217
 Step 190: loss 2.170797109603882
 Step 200: loss 2.081918239593506
 [Seed 23] Step 200 --> val loss 2.3934974670410156
 Step 210: loss 2.084749698638916
 Step 220: loss 2.066624164581299
 Step 230: loss 2.1372714042663574
 Step 240: loss 2.004089117050171
 Step 250: loss 2.0601327419281006
 Step 260: loss 2.080242156982422
 Step 270: loss 2.083458423614502
 Step 280: loss 2.127002716064453
 Step 290: loss 1.9972753524780273
 Step 300: loss 1.9244093894958496
 Step 310: loss 2.0632081031799316
 Step 320: loss 2.0261120796203613
 Step 330: loss 2.008251667022705
 Step 340: loss 1.999070167541504
 Step 350: loss 2.070890426635742
 Step 360: loss 1.917306900024414
 Step 370: loss 2.019423484802246
 Step 380: loss 2.0112099647521973
 Step 390: loss 2.178744316101074
 Step 400: loss 2.002751111984253
 [Seed 23] Step 400 --> val loss 2.297865629196167
 Step 410: loss 1.9945247173309326
 Step 420: loss 2.0191075801849365
 Step 430: loss 1.9474624395370483
 Step 440: loss 2.1025795936584473
 Step 450: loss 2.0137569904327393
 Step 460: loss 1.9284558296203613
 Step 470: loss 1.9775575399398804
 Step 480: loss 2.0354130268096924
 Step 490: loss 1.91366708278656
 Step 500: loss 2.0256123542785645
 Step 510: loss 1.8684618473052979
 Step 520: loss 1.9334192276000977
 Step 530: loss 1.9002350568771362
 Step 540: loss 2.0353684425354004
 Step 550: loss 1.9923880100250244
 Step 560: loss 1.9346126317977905
 Step 570: loss 2.0934667587280273
 Step 580: loss 1.8746103048324585
 Step 590: loss 1.9462950229644775
 Step 600: loss 1.9955004453659058
 [Seed 23] Step 600 --> val loss 2.3266444206237793
 Step 610: loss 2.004729747772217
 Step 620: loss 2.0021913051605225
 Step 630: loss 2.101853132247925
 Step 640: loss 1.908843755722046
 Step 650: loss 2.0189058780670166
 Step 660: loss 1.9394080638885498
 Step 670: loss 1.9548356533050537
 Step 680: loss 1.9515817165374756
 Step 690: loss 1.8877910375595093
 Step 700: loss 1.9255435466766357
 Step 710: loss 1.9494746923446655
 Step 720: loss 1.8715052604675293
 Step 730: loss 2.0418851375579834
 Step 740: loss 1.9151567220687866
 Step 750: loss 1.8919410705566406
 Step 760: loss 1.880433440208435
 Step 770: loss 1.9953813552856445
 Step 780: loss 1.8556216955184937
 Step 790: loss 1.886602759361267
 Step 800: loss 1.9109910726547241
 [Seed 23] Step 800 --> val loss 2.112401008605957
 Step 810: loss 1.9306354522705078
 Step 820: loss 1.941868543624878
 Step 830: loss 1.9578359127044678
 Step 840: loss 1.9256525039672852
 Step 850: loss 2.0557618141174316
 Step 860: loss 1.9572393894195557
 Step 870: loss 1.9407281875610352
 Step 880: loss 2.0301241874694824
 Step 890: loss 1.9139635562896729
 Step 900: loss 1.978981375694275
 Step 910: loss 1.8213098049163818
 Step 920: loss 1.879368782043457
 Step 930: loss 2.016589879989624
 Step 940: loss 1.9265550374984741
 Step 950: loss 1.8319424390792847
 Step 960: loss 1.9069045782089233
 Step 970: loss 1.8628734350204468
 Step 980: loss 1.852548360824585
 Step 990: loss 1.9659522771835327
 Step 1000: loss 1.958935260772705
 [Seed 23] Step 1000 --> val loss 2.1309497356414795
 Step 1010: loss 1.8213046789169312
 Step 1020: loss 1.8330581188201904
 Step 1030: loss 1.9251093864440918
 Step 1040: loss 1.8880693912506104
 Step 1050: loss 1.9852913618087769
 Step 1060: loss 1.9021530151367188
 Step 1070: loss 1.953508734703064
 Step 1080: loss 1.8984003067016602
 Step 1090: loss 1.8257355690002441
 Step 1100: loss 1.7791612148284912
 Step 1110: loss 1.8858201503753662
 Step 1120: loss 1.8774316310882568
 Step 1130: loss 1.733410358428955
 Step 1140: loss 1.9246604442596436
 Step 1150: loss 1.8856945037841797
 Step 1160: loss 1.8369731903076172
 Step 1170: loss 1.7835667133331299
 Step 1180: loss 1.9138168096542358
 Step 1190: loss 1.7921465635299683
 Step 1200: loss 1.8745296001434326
 [Seed 23] Step 1200 --> val loss 2.0255250930786133
 Step 1210: loss 1.8121857643127441
 Step 1220: loss 1.8291934728622437
 Step 1230: loss 1.7315268516540527
 Step 1240: loss 1.8195518255233765
 Step 1250: loss 1.815477728843689
 Step 1260: loss 1.7977495193481445
 Step 1270: loss 1.841017723083496
 Step 1280: loss 1.779291033744812
 Step 1290: loss 1.7723387479782104
 Step 1300: loss 1.8005484342575073
 Step 1310: loss 1.841124415397644
 Step 1320: loss 1.8640186786651611
 Step 1330: loss 1.8580703735351562
 Step 1340: loss 1.669132947921753
 Step 1350: loss 1.7041105031967163
 Step 1360: loss 1.6677502393722534
 Step 1370: loss 1.8179725408554077
 Step 1380: loss 1.7788041830062866
 Step 1390: loss 1.8701934814453125
 Step 1400: loss 1.7383337020874023
 [Seed 23] Step 1400 --> val loss 1.9826643466949463
 Step 1410: loss 1.670379877090454
 Step 1420: loss 1.8655380010604858
 Step 1430: loss 1.7857742309570312
 Step 1440: loss 1.929357886314392
 Step 1450: loss 1.8817453384399414
 Step 1460: loss 1.6505063772201538
 Step 1470: loss 1.9009194374084473
 Step 1480: loss 1.7475852966308594
 Step 1490: loss 1.7876102924346924
 Step 1500: loss 1.8344916105270386
 Step 1510: loss 1.758269190788269
 Step 1520: loss 1.7498515844345093
 Step 1530: loss 1.5831563472747803
 Step 1540: loss 1.6309516429901123
 Step 1550: loss 1.8482956886291504
 Step 1560: loss 1.7050504684448242
 Step 1570: loss 1.810903787612915
 Step 1580: loss 1.6740891933441162
 Step 1590: loss 1.728845477104187
 Step 1600: loss 1.7705963850021362
 [Seed 23] Step 1600 --> val loss 1.861127495765686
 Step 1610: loss 1.595495581626892
 Step 1620: loss 1.874147653579712
 Step 1630: loss 1.748369574546814
 Step 1640: loss 1.7556235790252686
 Step 1650: loss 1.8752667903900146
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 23] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 23] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [50 58  6  1 63 53 59  1 56 39] ...
 First target sequence: [58  6  1 63 53 59  1 56 39 58] ...

 Decoded input: lt, you rat-catcher, will you walk?

 TYBALT:
 What wouldst thou h

 Decoded target: t, you rat-catcher, will you walk?

 TYBALT:
 What wouldst thou ha
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.227328777313232
 [Seed 24] Step 0 --> val loss 4.1832275390625
 Step 10: loss 3.8557324409484863
 Step 20: loss 3.3705081939697266
 Step 30: loss 2.8702847957611084
 Step 40: loss 2.6259665489196777
 Step 50: loss 2.5425992012023926
 Step 60: loss 2.454204559326172
 Step 70: loss 2.448554515838623
 Step 80: loss 2.4383106231689453
 Step 90: loss 2.31857967376709
 Step 100: loss 2.344087839126587
 Step 110: loss 2.1934146881103516
 Step 120: loss 2.2169418334960938
 Step 130: loss 2.066413640975952
 Step 140: loss 2.1853227615356445
 Step 150: loss 2.1103925704956055
 Step 160: loss 2.1059653759002686
 Step 170: loss 2.2077276706695557
 Step 180: loss 2.1537835597991943
 Step 190: loss 2.1871337890625
 Step 200: loss 2.1646761894226074
 [Seed 24] Step 200 --> val loss 2.341813087463379
 Step 210: loss 2.0106377601623535
 Step 220: loss 2.1787590980529785
 Step 230: loss 2.1132984161376953
 Step 240: loss 2.0642993450164795
 Step 250: loss 2.110382556915283
 Step 260: loss 2.0533289909362793
 Step 270: loss 2.057389736175537
 Step 280: loss 2.001765012741089
 Step 290: loss 2.0051767826080322
 Step 300: loss 2.0314087867736816
 Step 310: loss 2.095661163330078
 Step 320: loss 2.045670986175537
 Step 330: loss 2.056156635284424
 Step 340: loss 1.960136890411377
 Step 350: loss 2.062166452407837
 Step 360: loss 2.069701671600342
 Step 370: loss 2.044065475463867
 Step 380: loss 1.9493904113769531
 Step 390: loss 2.047999858856201
 Step 400: loss 1.9139059782028198
 [Seed 24] Step 400 --> val loss 2.3030078411102295
 Step 410: loss 1.968409538269043
 Step 420: loss 1.9842767715454102
 Step 430: loss 2.067138671875
 Step 440: loss 2.048739433288574
 Step 450: loss 2.04941725730896
 Step 460: loss 2.143270969390869
 Step 470: loss 2.036701202392578
 Step 480: loss 1.995417594909668
 Step 490: loss 1.979750156402588
 Step 500: loss 2.075510025024414
 Step 510: loss 1.9822386503219604
 Step 520: loss 2.0434980392456055
 Step 530: loss 2.011091470718384
 Step 540: loss 2.034921646118164
 Step 550: loss 1.9787700176239014
 Step 560: loss 1.9833014011383057
 Step 570: loss 1.940734624862671
 Step 580: loss 2.0926754474639893
 Step 590: loss 2.0553624629974365
 Step 600: loss 1.9594533443450928
 [Seed 24] Step 600 --> val loss 2.290487051010132
 Step 610: loss 1.9169886112213135
 Step 620: loss 1.9722089767456055
 Step 630: loss 1.9742889404296875
 Step 640: loss 2.0781781673431396
 Step 650: loss 2.016706943511963
 Step 660: loss 2.0695807933807373
 Step 670: loss 1.9843806028366089
 Step 680: loss 2.0588371753692627
 Step 690: loss 2.0411314964294434
 Step 700: loss 1.865952491760254
 Step 710: loss 2.1241455078125
 Step 720: loss 2.095940589904785
 Step 730: loss 2.0504093170166016
 Step 740: loss 2.036778450012207
 Step 750: loss 2.059596538543701
 Step 760: loss 2.015780448913574
 Step 770: loss 2.0555529594421387
 Step 780: loss 2.1084792613983154
 Step 790: loss 1.9903943538665771
 Step 800: loss 2.024221658706665
 [Seed 24] Step 800 --> val loss 2.234285831451416
 Step 810: loss 1.989271879196167
 Step 820: loss 2.134443998336792
 Step 830: loss 2.055382251739502
 Step 840: loss 2.0049309730529785
 Step 850: loss 1.9746217727661133
 Step 860: loss 2.0676283836364746
 Step 870: loss 1.9383233785629272
 Step 880: loss 2.064387798309326
 Step 890: loss 1.9808019399642944
 Step 900: loss 2.066633701324463
 Step 910: loss 1.9528807401657104
 Step 920: loss 1.9571282863616943
 Step 930: loss 1.976144790649414
 Step 940: loss 2.0114972591400146
 Step 950: loss 2.076643228530884
 Step 960: loss 2.013967275619507
 Step 970: loss 2.0761303901672363
 Step 980: loss 2.128618001937866
 Step 990: loss 2.0532350540161133
 Step 1000: loss 2.0487747192382812
 [Seed 24] Step 1000 --> val loss 2.2193009853363037
 Step 1010: loss 1.979319453239441
 Step 1020: loss 1.9051287174224854
 Step 1030: loss 2.000452995300293
 Step 1040: loss 2.0201616287231445
 Step 1050: loss 2.057328939437866
 Step 1060: loss 1.839470624923706
 Step 1070: loss 1.8356058597564697
 Step 1080: loss 2.0920348167419434
 Step 1090: loss 1.9190043210983276
 Step 1100: loss 1.9642236232757568
 Step 1110: loss 1.8819622993469238
 Step 1120: loss 2.0569334030151367
 Step 1130: loss 2.0378623008728027
 Step 1140: loss 1.9370338916778564
 Step 1150: loss 1.935256838798523
 Step 1160: loss 1.8204450607299805
 Step 1170: loss 1.9584182500839233
 Step 1180: loss 1.9737918376922607
 Step 1190: loss 1.909332275390625
 Step 1200: loss 1.9389095306396484
 [Seed 24] Step 1200 --> val loss 2.09019136428833
 Step 1210: loss 1.9823930263519287
 Step 1220: loss 1.8629615306854248
 Step 1230: loss 1.9635400772094727
 Step 1240: loss 1.8114137649536133
 Step 1250: loss 2.0255415439605713
 Step 1260: loss 1.9674081802368164
 Step 1270: loss 1.9270474910736084
 Step 1280: loss 1.9639935493469238
 Step 1290: loss 1.8916901350021362
 Step 1300: loss 1.8753890991210938
 Step 1310: loss 1.8876680135726929
 Step 1320: loss 1.8034305572509766
 Step 1330: loss 1.969077467918396
 Step 1340: loss 1.9025882482528687
 Step 1350: loss 1.8099021911621094
 Step 1360: loss 1.8157200813293457
 Step 1370: loss 1.8701375722885132
 Step 1380: loss 1.8612264394760132
 Step 1390: loss 1.9338898658752441
 Step 1400: loss 1.9353183507919312
 [Seed 24] Step 1400 --> val loss 2.1557328701019287
 Step 1410: loss 1.8876707553863525
 Step 1420: loss 2.053651809692383
 Step 1430: loss 1.710097312927246
 Step 1440: loss 1.8137167692184448
 Step 1450: loss 1.8238756656646729
 Step 1460: loss 1.8468412160873413
 Step 1470: loss 1.7507776021957397
 Step 1480: loss 1.9759472608566284
 Step 1490: loss 1.7678627967834473
 Step 1500: loss 1.8271212577819824
 Step 1510: loss 1.954001784324646
 Step 1520: loss 1.994344711303711
 Step 1530: loss 1.803213119506836
 Step 1540: loss 1.8702702522277832
 Step 1550: loss 1.89218270778656
 Step 1560: loss 1.856878638267517
 Step 1570: loss 1.6679767370224
 Step 1580: loss 1.9049701690673828
 Step 1590: loss 1.8502429723739624
 Step 1600: loss 1.6295874118804932
 [Seed 24] Step 1600 --> val loss 2.0677993297576904
 Step 1610: loss 1.887918472290039
 Step 1620: loss 1.7150475978851318
 Step 1630: loss 1.8296092748641968
 Step 1640: loss 1.7192363739013672
 Step 1650: loss 1.8306803703308105
 Step 1660: loss 1.7886275053024292
 Step 1670: loss 1.8264278173446655
 Step 1680: loss 1.8171266317367554
 Step 1690: loss 1.6758220195770264
 Step 1700: loss 1.78376042842865
 Step 1710: loss 1.8203363418579102
 Step 1720: loss 1.7606909275054932
 Step 1730: loss 1.751622200012207
 Step 1740: loss 1.8346574306488037
 Step 1750: loss 1.5859912633895874
 Step 1760: loss 1.7247886657714844
 Step 1770: loss 1.7981595993041992
 Step 1780: loss 1.7972257137298584
 Step 1790: loss 1.7418245077133179
 Step 1800: loss 1.6913646459579468
 [Seed 24] Step 1800 --> val loss 1.9520677328109741
 Step 1810: loss 1.7086619138717651
 Step 1820: loss 1.6737817525863647
 Step 1830: loss 1.922864317893982
 Step 1840: loss 1.7860417366027832
 Step 1850: loss 1.606076955795288
 Step 1860: loss 1.6787395477294922
 Step 1870: loss 1.7085602283477783
 Step 1880: loss 1.6877222061157227
 Step 1890: loss 1.661240577697754
 Step 1900: loss 1.767000436782837
 Step 1910: loss 1.6081395149230957
 Step 1920: loss 1.7327302694320679
 Step 1930: loss 1.5394223928451538
 Step 1940: loss 1.6930021047592163
 Step 1950: loss 1.7285269498825073
 Step 1960: loss 1.7491897344589233
 Step 1970: loss 1.7782249450683594
 Step 1980: loss 1.6202610731124878
 Step 1990: loss 1.6982707977294922
 Step 2000: loss 1.648116111755371
 [Seed 24] Step 2000 --> val loss 1.8820613622665405
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 0 24 33 15 21 27 10  0 35 46] ...
 First target sequence: [24 33 15 21 27 10  0 35 46 63] ...

 Decoded input: 
 LUCIO:
 Why, how now, Claudio! whence comes this restraint?

 CLA

 Decoded target: LUCIO:
 Why, how now, Claudio! whence comes this restraint?

 CLAU
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.203446388244629
 [Seed 25] Step 0 --> val loss 4.180419445037842
 Step 10: loss 3.832669734954834
 Step 20: loss 3.371039867401123
 Step 30: loss 2.829942226409912
 Step 40: loss 2.6854827404022217
 Step 50: loss 2.5834360122680664
 Step 60: loss 2.5396130084991455
 Step 70: loss 2.4841232299804688
 Step 80: loss 2.3678195476531982
 Step 90: loss 2.3521318435668945
 Step 100: loss 2.261017084121704
 Step 110: loss 2.2089412212371826
 Step 120: loss 2.2126903533935547
 Step 130: loss 2.2173945903778076
 Step 140: loss 2.171266794204712
 Step 150: loss 2.2484869956970215
 Step 160: loss 2.1580142974853516
 Step 170: loss 2.221980094909668
 Step 180: loss 2.105517864227295
 Step 190: loss 2.0773534774780273
 Step 200: loss 2.135132312774658
 [Seed 25] Step 200 --> val loss 2.362248659133911
 Step 210: loss 2.0716423988342285
 Step 220: loss 2.134382486343384
 Step 230: loss 2.0510623455047607
 Step 240: loss 2.1665196418762207
 Step 250: loss 2.0185487270355225
 Step 260: loss 2.0620193481445312
 Step 270: loss 2.178515911102295
 Step 280: loss 2.014286994934082
 Step 290: loss 2.1422319412231445
 Step 300: loss 2.060587167739868
 Step 310: loss 1.9990234375
 Step 320: loss 1.935238242149353
 Step 330: loss 1.946629524230957
 Step 340: loss 2.170680522918701
 Step 350: loss 2.0413026809692383
 Step 360: loss 1.9272998571395874
 Step 370: loss 2.078235626220703
 Step 380: loss 2.12751841545105
 Step 390: loss 2.0635504722595215
 Step 400: loss 1.9003913402557373
 [Seed 25] Step 400 --> val loss 2.2873194217681885
 Step 410: loss 1.922658920288086
 Step 420: loss 2.018432855606079
 Step 430: loss 1.989898443222046
 Step 440: loss 2.0148541927337646
 Step 450: loss 2.0647828578948975
 Step 460: loss 1.9228681325912476
 Step 470: loss 1.9465234279632568
 Step 480: loss 2.0146284103393555
 Step 490: loss 1.9087709188461304
 Step 500: loss 1.9642963409423828
 Step 510: loss 1.9719500541687012
 Step 520: loss 1.9646470546722412
 Step 530: loss 2.017366409301758
 Step 540: loss 2.049811363220215
 Step 550: loss 2.025650978088379
 Step 560: loss 2.072504997253418
 Step 570: loss 2.082122325897217
 Step 580: loss 2.116715908050537
 Step 590: loss 2.038968563079834
 Step 600: loss 1.9995629787445068
 [Seed 25] Step 600 --> val loss 2.222946882247925
 Step 610: loss 1.9754728078842163
 Step 620: loss 2.0414366722106934
 Step 630: loss 2.090503692626953
 Step 640: loss 1.9868152141571045
 Step 650: loss 2.055717945098877
 Step 660: loss 2.0284643173217773
 Step 670: loss 1.9878309965133667
 Step 680: loss 1.8881070613861084
 Step 690: loss 1.9827775955200195
 Step 700: loss 1.9009685516357422
 Step 710: loss 1.9733232259750366
 Step 720: loss 1.9936420917510986
 Step 730: loss 1.9734386205673218
 Step 740: loss 1.8916234970092773
 Step 750: loss 1.916111946105957
 Step 760: loss 1.922687292098999
 Step 770: loss 2.0995640754699707
 Step 780: loss 2.0653395652770996
 Step 790: loss 1.902569055557251
 Step 800: loss 1.952967882156372
 [Seed 25] Step 800 --> val loss 2.2663936614990234
 Step 810: loss 1.9452838897705078
 Step 820: loss 1.9108823537826538
 Step 830: loss 1.8855912685394287
 Step 840: loss 1.9039390087127686
 Step 850: loss 1.9017060995101929
 Step 860: loss 1.9613382816314697
 Step 870: loss 1.8948091268539429
 Step 880: loss 1.9635004997253418
 Step 890: loss 1.8067902326583862
 Step 900: loss 1.910597324371338
 Step 910: loss 1.9960521459579468
 Step 920: loss 1.859910011291504
 Step 930: loss 1.9475469589233398
 Step 940: loss 1.9396427869796753
 Step 950: loss 1.9851807355880737
 Step 960: loss 1.9496434926986694
 Step 970: loss 1.8786604404449463
 Step 980: loss 1.8728227615356445
 Step 990: loss 1.9587867259979248
 Step 1000: loss 1.9488961696624756
 [Seed 25] Step 1000 --> val loss 2.143486738204956
 Step 1010: loss 1.9625179767608643
 Step 1020: loss 2.05454158782959
 Step 1030: loss 2.0108611583709717
 Step 1040: loss 2.0071158409118652
 Step 1050: loss 2.0526044368743896
 Step 1060: loss 1.880446434020996
 Step 1070: loss 1.9741482734680176
 Step 1080: loss 1.8982155323028564
 Step 1090: loss 1.9940292835235596
 Step 1100: loss 1.8825912475585938
 Step 1110: loss 1.892216682434082
 Step 1120: loss 1.7959083318710327
 Step 1130: loss 1.9665641784667969
 Step 1140: loss 1.869405746459961
 Step 1150: loss 1.8979742527008057
 Step 1160: loss 1.8614635467529297
 Step 1170: loss 1.9545323848724365
 Step 1180: loss 1.860836386680603
 Step 1190: loss 1.8317140340805054
 Step 1200: loss 1.687819242477417
 [Seed 25] Step 1200 --> val loss 2.0369575023651123
 Step 1210: loss 1.9428739547729492
 Step 1220: loss 1.7744028568267822
 Step 1230: loss 1.741201400756836
 Step 1240: loss 1.8118127584457397
 Step 1250: loss 1.7989834547042847
 Step 1260: loss 1.8555196523666382
 Step 1270: loss 1.9456591606140137
 Step 1280: loss 1.783024787902832
 Step 1290: loss 1.7684153318405151
 Step 1300: loss 1.8707798719406128
 Step 1310: loss 1.7920039892196655
 Step 1320: loss 1.738236665725708
 Step 1330: loss 1.7840855121612549
 Step 1340: loss 1.9808213710784912
 Step 1350: loss 1.9198668003082275
 Step 1360: loss 1.7808892726898193
 Step 1370: loss 1.7939119338989258
 Step 1380: loss 1.8408979177474976
 Step 1390: loss 1.8254883289337158
 Step 1400: loss 1.869821548461914
 [Seed 25] Step 1400 --> val loss 2.0108611583709717
 Step 1410: loss 1.8660691976547241
 Step 1420: loss 1.794743537902832
 Step 1430: loss 1.7983306646347046
 Step 1440: loss 1.8925063610076904
 Step 1450: loss 1.828599452972412
 Step 1460: loss 1.7590954303741455
 Step 1470: loss 1.6792668104171753
 Step 1480: loss 1.8493455648422241
 Step 1490: loss 1.797520399093628
 Step 1500: loss 1.7814433574676514
 Step 1510: loss 1.7792290449142456
 Step 1520: loss 1.9271268844604492
 Step 1530: loss 1.7006118297576904
 Step 1540: loss 1.8643028736114502
 Step 1550: loss 1.672804594039917
 Step 1560: loss 1.739499807357788
 Step 1570: loss 1.8033493757247925
 Step 1580: loss 1.8527883291244507
 Step 1590: loss 1.8450864553451538
 Step 1600: loss 1.7742241621017456
 [Seed 25] Step 1600 --> val loss 1.947564959526062
 Step 1610: loss 1.840269923210144
 Step 1620: loss 1.7029838562011719
 Step 1630: loss 1.7437905073165894
 Step 1640: loss 1.6830213069915771
 Step 1650: loss 1.6630096435546875
 Step 1660: loss 1.6461423635482788
 Step 1670: loss 1.6720775365829468
 Step 1680: loss 1.6994693279266357
 Step 1690: loss 1.7667465209960938
 Step 1700: loss 1.802525281906128
 Step 1710: loss 1.8152492046356201
 Step 1720: loss 1.7185125350952148
 Step 1730: loss 1.7960008382797241
 Step 1740: loss 1.7023332118988037
 Step 1750: loss 1.702906847000122
 Step 1760: loss 1.808915138244629
 Step 1770: loss 1.6599750518798828
 Step 1780: loss 1.6511247158050537
 Step 1790: loss 1.7875843048095703
 Step 1800: loss 1.7879219055175781
 [Seed 25] Step 1800 --> val loss 1.8257274627685547
 Step 1810: loss 1.722650408744812
 Step 1820: loss 1.748796820640564
 Step 1830: loss 1.7037980556488037
 Step 1840: loss 1.6632661819458008
 Step 1850: loss 1.7060878276824951
 Step 1860: loss 1.8091996908187866
 Step 1870: loss 1.694617748260498
 Step 1880: loss 1.711029052734375
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 25] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [43 56 43  1 58 46 63  1 52 59] ...
 First target sequence: [56 43  1 58 46 63  1 52 59 56] ...

 Decoded input: ere thy nurse, thy tongue to teach,
 'Pardon' should be the first

 Decoded target: re thy nurse, thy tongue to teach,
 'Pardon' should be the first 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.205352306365967
 [Seed 26] Step 0 --> val loss 4.187572002410889
 Step 10: loss 3.858971118927002
 Step 20: loss 3.290287733078003
 Step 30: loss 3.0079848766326904
 Step 40: loss 2.5868642330169678
 Step 50: loss 2.5956568717956543
 Step 60: loss 2.502533435821533
 Step 70: loss 2.4178404808044434
 Step 80: loss 2.356440544128418
 Step 90: loss 2.2908072471618652
 Step 100: loss 2.3592917919158936
 Step 110: loss 2.3399453163146973
 Step 120: loss 2.211378574371338
 Step 130: loss 2.1564290523529053
 Step 140: loss 2.2597548961639404
 Step 150: loss 2.2221741676330566
 Step 160: loss 2.017508029937744
 Step 170: loss 2.1797902584075928
 Step 180: loss 2.0611040592193604
 Step 190: loss 2.1868364810943604
 Step 200: loss 2.1364963054656982
 [Seed 26] Step 200 --> val loss 2.34561824798584
 Step 210: loss 2.156365156173706
 Step 220: loss 2.078681468963623
 Step 230: loss 2.1106879711151123
 Step 240: loss 2.0551629066467285
 Step 250: loss 2.1788101196289062
 Step 260: loss 1.962194800376892
 Step 270: loss 2.0947887897491455
 Step 280: loss 2.042825937271118
 Step 290: loss 2.139997959136963
 Step 300: loss 2.0431668758392334
 Step 310: loss 1.907578468322754
 Step 320: loss 1.9832520484924316
 Step 330: loss 2.0629208087921143
 Step 340: loss 2.042720317840576
 Step 350: loss 2.0371060371398926
 Step 360: loss 2.0422487258911133
 Step 370: loss 2.0211453437805176
 Step 380: loss 1.9125112295150757
 Step 390: loss 1.945665717124939
 Step 400: loss 2.156719207763672
 [Seed 26] Step 400 --> val loss 2.222773313522339
 Step 410: loss 2.0730061531066895
 Step 420: loss 2.0328996181488037
 Step 430: loss 1.9693007469177246
 Step 440: loss 1.9717211723327637
 Step 450: loss 2.016486406326294
 Step 460: loss 2.0774669647216797
 Step 470: loss 2.1035149097442627
 Step 480: loss 2.10139799118042
 Step 490: loss 2.057234048843384
 Step 500: loss 1.998369574546814
 Step 510: loss 2.0426483154296875
 Step 520: loss 1.984450101852417
 Step 530: loss 1.964082956314087
 Step 540: loss 1.9492172002792358
 Step 550: loss 1.9794254302978516
 Step 560: loss 2.0568103790283203
 Step 570: loss 2.0103840827941895
 Step 580: loss 1.995906114578247
 Step 590: loss 2.033870220184326
 Step 600: loss 1.972400426864624
 [Seed 26] Step 600 --> val loss 2.2642109394073486
 Step 610: loss 2.034930944442749
 Step 620: loss 2.038954019546509
 Step 630: loss 1.9234826564788818
 Step 640: loss 2.0373129844665527
 Step 650: loss 2.080162525177002
 Step 660: loss 2.06886887550354
 Step 670: loss 2.006488800048828
 Step 680: loss 1.9887664318084717
 Step 690: loss 2.099111318588257
 Step 700: loss 2.056795597076416
 Step 710: loss 1.9925212860107422
 Step 720: loss 2.136547565460205
 Step 730: loss 2.0755832195281982
 Step 740: loss 2.151003360748291
 Step 750: loss 1.9632456302642822
 Step 760: loss 2.0108840465545654
 Step 770: loss 2.061257839202881
 Step 780: loss 2.1635971069335938
 Step 790: loss 1.9972103834152222
 Step 800: loss 2.0962741374969482
 [Seed 26] Step 800 --> val loss 2.222162961959839
 Step 810: loss 2.0084075927734375
 Step 820: loss 2.0506393909454346
 Step 830: loss 2.0323734283447266
 Step 840: loss 2.0492539405822754
 Step 850: loss 2.054985761642456
 Step 860: loss 2.0048322677612305
 Step 870: loss 2.070601463317871
 Step 880: loss 2.0607898235321045
 Step 890: loss 1.9519777297973633
 Step 900: loss 2.0503995418548584
 Step 910: loss 2.1515519618988037
 Step 920: loss 2.0084924697875977
 Step 930: loss 2.027207851409912
 Step 940: loss 2.1384260654449463
 Step 950: loss 2.1412181854248047
 Step 960: loss 2.165182590484619
 Step 970: loss 2.1250100135803223
 Step 980: loss 2.0457348823547363
 Step 990: loss 2.17067813873291
 Step 1000: loss 2.0564661026000977
 [Seed 26] Step 1000 --> val loss 2.3106648921966553
 Step 1010: loss 2.1948273181915283
 Step 1020: loss 2.1204404830932617
 Step 1030: loss 2.081573247909546
 Step 1040: loss 2.109363317489624
 Step 1050: loss 2.0234487056732178
 Step 1060: loss 2.0754454135894775
 Step 1070: loss 2.111846923828125
 Step 1080: loss 2.1203315258026123
 Step 1090: loss 2.089280605316162
 Step 1100: loss 1.9988887310028076
 Step 1110: loss 1.9756991863250732
 Step 1120: loss 2.0538666248321533
 Step 1130: loss 2.0747551918029785
 Step 1140: loss 2.0043187141418457
 Step 1150: loss 2.1119072437286377
 Step 1160: loss 2.0244927406311035
 Step 1170: loss 1.8949182033538818
 Step 1180: loss 1.922776699066162
 Step 1190: loss 2.007376194000244
 Step 1200: loss 2.1185293197631836
 [Seed 26] Step 1200 --> val loss 2.2914843559265137
 Step 1210: loss 1.9753894805908203
 Step 1220: loss 2.0894691944122314
 Step 1230: loss 1.9278905391693115
 Step 1240: loss 2.003632068634033
 Step 1250: loss 1.9541863203048706
 Step 1260: loss 1.9664192199707031
 Step 1270: loss 1.9525952339172363
 Step 1280: loss 1.9531625509262085
 Step 1290: loss 2.073481321334839
 Step 1300: loss 1.9124524593353271
 Step 1310: loss 1.950287103652954
 Step 1320: loss 1.918532371520996
 Step 1330: loss 1.9440720081329346
 Step 1340: loss 1.9876840114593506
 Step 1350: loss 1.9429233074188232
 Step 1360: loss 1.8575280904769897
 Step 1370: loss 1.9619125127792358
 Step 1380: loss 1.910136103630066
 Step 1390: loss 1.9837039709091187
 Step 1400: loss 1.90640389919281
 [Seed 26] Step 1400 --> val loss 2.1450557708740234
 Step 1410: loss 1.8164327144622803
 Step 1420: loss 1.880955457687378
 Step 1430: loss 1.7558155059814453
 Step 1440: loss 1.8682835102081299
 Step 1450: loss 1.929983139038086
 Step 1460: loss 1.7753099203109741
 Step 1470: loss 1.798316240310669
 Step 1480: loss 1.9956258535385132
 Step 1490: loss 1.7516539096832275
 Step 1500: loss 1.9239219427108765
 Step 1510: loss 1.9936141967773438
 Step 1520: loss 1.8903535604476929
 Step 1530: loss 1.7894643545150757
 Step 1540: loss 1.8842103481292725
 Step 1550: loss 1.7608609199523926
 Step 1560: loss 1.8310446739196777
 Step 1570: loss 1.7399253845214844
 Step 1580: loss 1.8800694942474365
 Step 1590: loss 1.921575903892517
 Step 1600: loss 1.7814388275146484
 [Seed 26] Step 1600 --> val loss 2.1695926189422607
 Step 1610: loss 1.7758777141571045
 Step 1620: loss 1.8246095180511475
 Step 1630: loss 1.839306354522705
 Step 1640: loss 1.8735359907150269
 Step 1650: loss 1.7688722610473633
 Step 1660: loss 1.8934755325317383
 Step 1670: loss 1.8436076641082764
 Step 1680: loss 1.7563321590423584
 Step 1690: loss 1.8889250755310059
 Step 1700: loss 1.838402509689331
 Step 1710: loss 1.7908157110214233
 Step 1720: loss 1.9046611785888672
 Step 1730: loss 1.8144853115081787
 Step 1740: loss 1.7960361242294312
 Step 1750: loss 1.8098294734954834
 Step 1760: loss 1.8784632682800293
 Step 1770: loss 1.755631446838379
 Step 1780: loss 1.737112045288086
 Step 1790: loss 1.7640416622161865
 Step 1800: loss 1.8838685750961304
 [Seed 26] Step 1800 --> val loss 1.9748834371566772
 Step 1810: loss 1.8891462087631226
 Step 1820: loss 1.698992133140564
 Step 1830: loss 1.703143835067749
 Step 1840: loss 1.7302335500717163
 Step 1850: loss 1.7379730939865112
 Step 1860: loss 1.858915090560913
 Step 1870: loss 1.8281787633895874
 Step 1880: loss 1.779886245727539
 Step 1890: loss 1.80037260055542
 Step 1900: loss 1.7426433563232422
 Step 1910: loss 1.7003517150878906
 Step 1920: loss 1.891850471496582
 Step 1930: loss 1.757049322128296
 Step 1940: loss 1.853420615196228
 Step 1950: loss 1.8135778903961182
 Step 1960: loss 1.671008586883545
 Step 1970: loss 1.5801780223846436
 Step 1980: loss 1.729225516319275
 Step 1990: loss 1.6527690887451172
 Step 2000: loss 1.7384262084960938
 [Seed 26] Step 2000 --> val loss 1.9186407327651978
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [46 43  1 45 53 53 57 43  6  1] ...
 First target sequence: [43  1 45 53 53 57 43  6  1 54] ...

 Decoded input: he goose, proves thee far and wide a broad goose.

 MERCUTIO:
 Why

 Decoded target: e goose, proves thee far and wide a broad goose.

 MERCUTIO:
 Why,
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2090044021606445
 [Seed 27] Step 0 --> val loss 4.185011386871338
 Step 10: loss 3.863252639770508
 Step 20: loss 3.3218038082122803
 Step 30: loss 2.8164448738098145
 Step 40: loss 2.591707706451416
 Step 50: loss 2.5678935050964355
 Step 60: loss 2.4165172576904297
 Step 70: loss 2.3167600631713867
 Step 80: loss 2.3372387886047363
 Step 90: loss 2.344982862472534
 Step 100: loss 2.2478508949279785
 Step 110: loss 2.3103699684143066
 Step 120: loss 2.209019422531128
 Step 130: loss 2.1563096046447754
 Step 140: loss 2.210362195968628
 Step 150: loss 2.2137374877929688
 Step 160: loss 2.142164707183838
 Step 170: loss 2.0594186782836914
 Step 180: loss 2.1156864166259766
 Step 190: loss 2.2273592948913574
 Step 200: loss 2.107722520828247
 [Seed 27] Step 200 --> val loss 2.3143117427825928
 Step 210: loss 2.058497905731201
 Step 220: loss 2.1043736934661865
 Step 230: loss 2.015212059020996
 Step 240: loss 2.0030062198638916
 Step 250: loss 2.172133684158325
 Step 260: loss 2.095693588256836
 Step 270: loss 2.0081069469451904
 Step 280: loss 1.9744231700897217
 Step 290: loss 2.070091724395752
 Step 300: loss 2.0800063610076904
 Step 310: loss 2.150031566619873
 Step 320: loss 2.0918779373168945
 Step 330: loss 2.019097089767456
 Step 340: loss 2.107400894165039
 Step 350: loss 2.1053712368011475
 Step 360: loss 2.0060832500457764
 Step 370: loss 1.921250343322754
 Step 380: loss 2.025067090988159
 Step 390: loss 1.9375070333480835
 Step 400: loss 1.9254205226898193
 [Seed 27] Step 400 --> val loss 2.2053635120391846
 Step 410: loss 1.9450658559799194
 Step 420: loss 2.084196090698242
 Step 430: loss 2.105234146118164
 Step 440: loss 1.9510440826416016
 Step 450: loss 1.9586572647094727
 Step 460: loss 2.013935089111328
 Step 470: loss 2.163754940032959
 Step 480: loss 2.02586030960083
 Step 490: loss 2.023533344268799
 Step 500: loss 2.0272319316864014
 Step 510: loss 2.074697494506836
 Step 520: loss 1.988493800163269
 Step 530: loss 1.9649814367294312
 Step 540: loss 1.859243392944336
 Step 550: loss 2.0142524242401123
 Step 560: loss 1.9807956218719482
 Step 570: loss 1.9767913818359375
 Step 580: loss 1.9305908679962158
 Step 590: loss 2.0396728515625
 Step 600: loss 1.9702670574188232
 [Seed 27] Step 600 --> val loss 2.349879026412964
 Step 610: loss 2.048356294631958
 Step 620: loss 1.9839977025985718
 Step 630: loss 2.007061004638672
 Step 640: loss 1.9577430486679077
 Step 650: loss 1.9897582530975342
 Step 660: loss 1.9371886253356934
 Step 670: loss 1.977167010307312
 Step 680: loss 1.9885337352752686
 Step 690: loss 1.9837729930877686
 Step 700: loss 1.9343969821929932
 Step 710: loss 1.9174308776855469
 Step 720: loss 1.9790284633636475
 Step 730: loss 1.977947473526001
 Step 740: loss 2.144768714904785
 Step 750: loss 2.0114693641662598
 Step 760: loss 2.0061845779418945
 Step 770: loss 2.0261502265930176
 Step 780: loss 2.0100038051605225
 Step 790: loss 1.9043772220611572
 Step 800: loss 2.031428813934326
 [Seed 27] Step 800 --> val loss 2.2164649963378906
 Step 810: loss 2.0930886268615723
 Step 820: loss 1.9208030700683594
 Step 830: loss 2.0522525310516357
 Step 840: loss 2.071902275085449
 Step 850: loss 1.9557560682296753
 Step 860: loss 1.927661657333374
 Step 870: loss 2.0059709548950195
 Step 880: loss 1.9850778579711914
 Step 890: loss 2.0543248653411865
 Step 900: loss 1.870753288269043
 Step 910: loss 1.9029254913330078
 Step 920: loss 2.013622760772705
 Step 930: loss 2.109556198120117
 Step 940: loss 2.0354597568511963
 Step 950: loss 2.1526260375976562
 Step 960: loss 2.030660629272461
 Step 970: loss 2.025589942932129
 Step 980: loss 1.9167439937591553
 Step 990: loss 1.9292032718658447
 Step 1000: loss 1.974435567855835
 [Seed 27] Step 1000 --> val loss 2.2851614952087402
 Step 1010: loss 1.97239089012146
 Step 1020: loss 2.02032470703125
 Step 1030: loss 1.8424452543258667
 Step 1040: loss 2.021890640258789
 Step 1050: loss 1.978476881980896
 Step 1060: loss 1.8819894790649414
 Step 1070: loss 1.9283273220062256
 Step 1080: loss 1.9462642669677734
 Step 1090: loss 1.815739393234253
 Step 1100: loss 1.9022988080978394
 Step 1110: loss 2.0795934200286865
 Step 1120: loss 1.947683334350586
 Step 1130: loss 1.9665184020996094
 Step 1140: loss 1.879719614982605
 Step 1150: loss 1.9155492782592773
 Step 1160: loss 1.925445556640625
 Step 1170: loss 1.858028769493103
 Step 1180: loss 1.903560996055603
 Step 1190: loss 1.9249743223190308
 Step 1200: loss 1.8936071395874023
 [Seed 27] Step 1200 --> val loss 2.1484670639038086
 Step 1210: loss 1.8863314390182495
 Step 1220: loss 1.8811712265014648
 Step 1230: loss 1.9272642135620117
 Step 1240: loss 1.9374475479125977
 Step 1250: loss 1.799546241760254
 Step 1260: loss 1.8641653060913086
 Step 1270: loss 1.821472406387329
 Step 1280: loss 1.9370250701904297
 Step 1290: loss 1.9876172542572021
 Step 1300: loss 1.9782092571258545
 Step 1310: loss 1.8685230016708374
 Step 1320: loss 1.7928493022918701
 Step 1330: loss 1.8876330852508545
 Step 1340: loss 1.7765135765075684
 Step 1350: loss 1.8576959371566772
 Step 1360: loss 1.7836987972259521
 Step 1370: loss 1.9442150592803955
 Step 1380: loss 1.8089499473571777
 Step 1390: loss 1.8868680000305176
 Step 1400: loss 1.8579802513122559
 [Seed 27] Step 1400 --> val loss 1.9666826725006104
 Step 1410: loss 1.8486754894256592
 Step 1420: loss 1.7655811309814453
 Step 1430: loss 1.8407440185546875
 Step 1440: loss 1.8184741735458374
 Step 1450: loss 1.80032479763031
 Step 1460: loss 1.7004919052124023
 Step 1470: loss 1.8240712881088257
 Step 1480: loss 1.7714011669158936
 Step 1490: loss 1.7809624671936035
 Step 1500: loss 1.6899513006210327
 Step 1510: loss 1.7115716934204102
 Step 1520: loss 1.782310962677002
 Step 1530: loss 1.696616530418396
 Step 1540: loss 1.8836253881454468
 Step 1550: loss 1.8309259414672852
 Step 1560: loss 1.7356895208358765
 Step 1570: loss 1.7411956787109375
 Step 1580: loss 1.854804515838623
 Step 1590: loss 1.8025445938110352
 Step 1600: loss 1.709578275680542
 [Seed 27] Step 1600 --> val loss 1.8869737386703491
 Step 1610: loss 1.7103567123413086
 Step 1620: loss 1.7468832731246948
 Step 1630: loss 1.7957797050476074
 Step 1640: loss 1.8035142421722412
 Step 1650: loss 1.774122714996338
 Step 1660: loss 1.9030832052230835
 Step 1670: loss 1.7660167217254639
 Step 1680: loss 1.7649645805358887
 Step 1690: loss 1.6607882976531982
 Step 1700: loss 1.7178291082382202
 Step 1710: loss 1.7235352993011475
 Step 1720: loss 1.8559201955795288
 Step 1730: loss 1.777881383895874
 Step 1740: loss 1.716869831085205
 Step 1750: loss 1.6578912734985352
 Step 1760: loss 1.768640398979187
 Step 1770: loss 1.642340064048767
 Step 1780: loss 1.723536491394043
 Step 1790: loss 1.752956748008728
 Step 1800: loss 1.6012195348739624
 [Seed 27] Step 1800 --> val loss 1.8296762704849243
 Step 1810: loss 1.7330269813537598
 Step 1820: loss 1.7172400951385498
 Step 1830: loss 1.659358024597168
 Step 1840: loss 1.8508524894714355
 Step 1850: loss 1.6791484355926514
 Step 1860: loss 1.626320719718933
 Step 1870: loss 1.5914604663848877
 Step 1880: loss 1.7912739515304565
 Step 1890: loss 1.661092758178711
 Step 1900: loss 1.676452398300171
 Step 1910: loss 1.7042063474655151
 Step 1920: loss 1.6260408163070679
 Step 1930: loss 1.6991885900497437
 Step 1940: loss 1.6542768478393555
 Step 1950: loss 1.553863763809204
 Step 1960: loss 1.6829891204833984
 Step 1970: loss 1.523558497428894
 Step 1980: loss 1.708691120147705
 Step 1990: loss 1.7943530082702637
 Step 2000: loss 1.5672658681869507
 [Seed 27] Step 2000 --> val loss 1.797479510307312
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 0 20 39 42  1 58 47 51 43  1] ...
 First target sequence: [20 39 42  1 58 47 51 43  1 41] ...

 Decoded input: 
 Had time cohered with place or place with wishing,
 Or that the 

 Decoded target: Had time cohered with place or place with wishing,
 Or that the r
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.212194442749023
 [Seed 28] Step 0 --> val loss 4.186477184295654
 Step 10: loss 3.8271892070770264
 Step 20: loss 3.3682193756103516
 Step 30: loss 2.905200958251953
 Step 40: loss 2.68357253074646
 Step 50: loss 2.6107709407806396
 Step 60: loss 2.523205280303955
 Step 70: loss 2.5107927322387695
 Step 80: loss 2.3582496643066406
 Step 90: loss 2.3323702812194824
 Step 100: loss 2.2566604614257812
 Step 110: loss 2.351337194442749
 Step 120: loss 2.1733551025390625
 Step 130: loss 2.210057258605957
 Step 140: loss 2.2747602462768555
 Step 150: loss 2.200928211212158
 Step 160: loss 2.2828221321105957
 Step 170: loss 2.2268600463867188
 Step 180: loss 2.1855897903442383
 Step 190: loss 2.1549153327941895
 Step 200: loss 2.06827449798584
 [Seed 28] Step 200 --> val loss 2.3525357246398926
 Step 210: loss 2.0966858863830566
 Step 220: loss 2.181595802307129
 Step 230: loss 2.0936288833618164
 Step 240: loss 2.075310707092285
 Step 250: loss 2.1449122428894043
 Step 260: loss 2.0525894165039062
 Step 270: loss 2.1113650798797607
 Step 280: loss 2.1028878688812256
 Step 290: loss 2.140238046646118
 Step 300: loss 1.92418372631073
 Step 310: loss 2.0366947650909424
 Step 320: loss 2.0354702472686768
 Step 330: loss 2.026543617248535
 Step 340: loss 2.058107614517212
 Step 350: loss 2.018343448638916
 Step 360: loss 1.8539371490478516
 Step 370: loss 2.196335792541504
 Step 380: loss 2.0472524166107178
 Step 390: loss 2.011357307434082
 Step 400: loss 2.088545322418213
 [Seed 28] Step 400 --> val loss 2.332282304763794
 Step 410: loss 2.0288748741149902
 Step 420: loss 2.1781346797943115
 Step 430: loss 2.1183667182922363
 Step 440: loss 2.0629067420959473
 Step 450: loss 2.114144802093506
 Step 460: loss 2.0926270484924316
 Step 470: loss 1.9613778591156006
 Step 480: loss 2.038557529449463
 Step 490: loss 1.9704389572143555
 Step 500: loss 2.068593740463257
 Step 510: loss 2.033538341522217
 Step 520: loss 1.9800024032592773
 Step 530: loss 1.9741144180297852
 Step 540: loss 2.0659921169281006
 Step 550: loss 2.1133430004119873
 Step 560: loss 1.9878957271575928
 Step 570: loss 1.9955039024353027
 Step 580: loss 1.9843025207519531
 Step 590: loss 2.0162954330444336
 Step 600: loss 1.9650280475616455
 [Seed 28] Step 600 --> val loss 2.269752264022827
 Step 610: loss 2.06317138671875
 Step 620: loss 2.074580430984497
 Step 630: loss 2.1642093658447266
 Step 640: loss 1.92660653591156
 Step 650: loss 2.0576233863830566
 Step 660: loss 1.8901715278625488
 Step 670: loss 2.1042404174804688
 Step 680: loss 2.046201229095459
 Step 690: loss 2.096557140350342
 Step 700: loss 2.0194849967956543
 Step 710: loss 1.9966745376586914
 Step 720: loss 2.015648365020752
 Step 730: loss 1.9443587064743042
 Step 740: loss 2.005330801010132
 Step 750: loss 2.0231196880340576
 Step 760: loss 1.9511594772338867
 Step 770: loss 2.0782313346862793
 Step 780: loss 1.9671697616577148
 Step 790: loss 2.0910024642944336
 Step 800: loss 1.957948088645935
 [Seed 28] Step 800 --> val loss 2.29997181892395
 Step 810: loss 1.942856788635254
 Step 820: loss 2.0320682525634766
 Step 830: loss 2.1440787315368652
 Step 840: loss 2.0738766193389893
 Step 850: loss 2.0348167419433594
 Step 860: loss 1.996559500694275
 Step 870: loss 2.0779147148132324
 Step 880: loss 2.0822768211364746
 Step 890: loss 1.9615334272384644
 Step 900: loss 2.017174243927002
 Step 910: loss 1.9910645484924316
 Step 920: loss 1.9600437879562378
 Step 930: loss 2.06213641166687
 Step 940: loss 1.9926813840866089
 Step 950: loss 2.1480553150177
 Step 960: loss 2.042278289794922
 Step 970: loss nan
 Step 980: loss nan
 Step 990: loss nan
 Step 1000: loss nan
 [Seed 28] Step 1000 --> val loss nan
 Step 1010: loss nan
 Step 1020: loss nan
 Step 1030: loss nan
 Step 1040: loss nan
 Step 1050: loss nan
 Step 1060: loss nan
 Step 1070: loss nan
 Step 1080: loss nan
 Step 1090: loss nan
 Step 1100: loss nan
 Step 1110: loss nan
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 28] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 28] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 28] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 28] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 28] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 47 52  1 58 46 43  1 57 41] ...
 First target sequence: [47 52  1 58 46 43  1 57 41 43] ...

 Decoded input:  in the scene,
 He proved best man i' the field, and for his meed

 Decoded target: in the scene,
 He proved best man i' the field, and for his meed

 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.222440719604492
 [Seed 29] Step 0 --> val loss 4.187168121337891
 Step 10: loss 3.8496787548065186
 Step 20: loss 3.2865118980407715
 Step 30: loss 2.793802261352539
 Step 40: loss 2.756763219833374
 Step 50: loss 2.5823144912719727
 Step 60: loss 2.5589098930358887
 Step 70: loss 2.3899617195129395
 Step 80: loss 2.3273544311523438
 Step 90: loss 2.2159910202026367
 Step 100: loss 2.351608991622925
 Step 110: loss 2.284047842025757
 Step 120: loss 2.128950834274292
 Step 130: loss 2.1529922485351562
 Step 140: loss 2.246389389038086
 Step 150: loss 2.1458683013916016
 Step 160: loss 2.2478556632995605
 Step 170: loss 2.1087117195129395
 Step 180: loss 2.2054014205932617
 Step 190: loss 2.0834455490112305
 Step 200: loss 2.0347352027893066
 [Seed 29] Step 200 --> val loss 2.4197194576263428
 Step 210: loss 2.081969738006592
 Step 220: loss 2.125401020050049
 Step 230: loss 2.0578300952911377
 Step 240: loss 1.9932212829589844
 Step 250: loss 2.0449771881103516
 Step 260: loss 2.188220739364624
 Step 270: loss 2.0768208503723145
 Step 280: loss 2.0641510486602783
 Step 290: loss 2.0132524967193604
 Step 300: loss 1.9592840671539307
 Step 310: loss 1.9747003316879272
 Step 320: loss 2.0750393867492676
 Step 330: loss 2.0375823974609375
 Step 340: loss 2.033545970916748
 Step 350: loss 2.0786123275756836
 Step 360: loss 1.9756546020507812
 Step 370: loss 2.075695037841797
 Step 380: loss 2.084911346435547
 Step 390: loss 2.0371270179748535
 Step 400: loss 2.008077383041382
 [Seed 29] Step 400 --> val loss 2.252464771270752
 Step 410: loss 2.058969497680664
 Step 420: loss 1.9459656476974487
 Step 430: loss 2.007711887359619
 Step 440: loss 2.0262527465820312
 Step 450: loss 1.9324774742126465
 Step 460: loss 1.9083614349365234
 Step 470: loss 1.984288215637207
 Step 480: loss 1.9962105751037598
 Step 490: loss 1.9177095890045166
 Step 500: loss 1.8603417873382568
 Step 510: loss 2.0012388229370117
 Step 520: loss 1.9749724864959717
 Step 530: loss 2.0778541564941406
 Step 540: loss 2.041365146636963
 Step 550: loss 2.0265398025512695
 Step 560: loss 2.062518835067749
 Step 570: loss 1.9950783252716064
 Step 580: loss 1.9340156316757202
 Step 590: loss 2.032895803451538
 Step 600: loss 2.114243984222412
 [Seed 29] Step 600 --> val loss 2.248497247695923
 Step 610: loss 2.0241811275482178
 Step 620: loss 2.101454734802246
 Step 630: loss 1.9945297241210938
 Step 640: loss 2.0552022457122803
 Step 650: loss 1.9720606803894043
 Step 660: loss 1.9791940450668335
 Step 670: loss 1.9499914646148682
 Step 680: loss 1.939666986465454
 Step 690: loss 1.9081604480743408
 Step 700: loss 1.9724395275115967
 Step 710: loss 1.8633086681365967
 Step 720: loss 1.9879586696624756
 Step 730: loss 1.9705772399902344
 Step 740: loss 2.0174129009246826
 Step 750: loss 2.130439281463623
 Step 760: loss 1.92500901222229
 Step 770: loss 2.0310826301574707
 Step 780: loss 2.0478248596191406
 Step 790: loss 2.1147727966308594
 Step 800: loss 2.030001640319824
 [Seed 29] Step 800 --> val loss 2.2966933250427246
 Step 810: loss 2.0543277263641357
 Step 820: loss 1.9736531972885132
 Step 830: loss 2.041083574295044
 Step 840: loss 2.0672719478607178
 Step 850: loss 2.047515392303467
 Step 860: loss 1.9224833250045776
 Step 870: loss 2.0819690227508545
 Step 880: loss 2.045659303665161
 Step 890: loss 1.9414000511169434
 Step 900: loss 1.969834327697754
 Step 910: loss 2.1030726432800293
 Step 920: loss 1.9766645431518555
 Step 930: loss 1.9947807788848877
 Step 940: loss 1.9954986572265625
 Step 950: loss 1.9980252981185913
 Step 960: loss 1.8670015335083008
 Step 970: loss 1.9932775497436523
 Step 980: loss 1.848968505859375
 Step 990: loss 2.176095962524414
 Step 1000: loss 1.9696276187896729
 [Seed 29] Step 1000 --> val loss 2.2669780254364014
 Step 1010: loss 1.8789489269256592
 Step 1020: loss 1.96640145778656
 Step 1030: loss 2.023588180541992
 Step 1040: loss 1.9797487258911133
 Step 1050: loss 2.0043699741363525
 Step 1060: loss 2.0528526306152344
 Step 1070: loss 1.9656003713607788
 Step 1080: loss 2.0977258682250977
 Step 1090: loss 1.9773776531219482
 Step 1100: loss 2.0271248817443848
 Step 1110: loss 2.0482544898986816
 Step 1120: loss 1.9835402965545654
 Step 1130: loss 1.9554184675216675
 Step 1140: loss 1.9829702377319336
 Step 1150: loss 1.9960367679595947
 Step 1160: loss 1.9828084707260132
 Step 1170: loss 2.038904905319214
 Step 1180: loss 1.9833660125732422
 Step 1190: loss 1.9391878843307495
 Step 1200: loss 1.9962825775146484
 [Seed 29] Step 1200 --> val loss 2.297227382659912
 Step 1210: loss 2.130645751953125
 Step 1220: loss 2.067162275314331
 Step 1230: loss 1.9756401777267456
 Step 1240: loss 1.9351768493652344
 Step 1250: loss 1.9206998348236084
 Step 1260: loss 1.9118432998657227
 Step 1270: loss 2.008267879486084
 Step 1280: loss 2.1030404567718506
 Step 1290: loss 1.868298053741455
 Step 1300: loss 2.0072267055511475
 Step 1310: loss 2.058925151824951
 Step 1320: loss 1.995319128036499
 Step 1330: loss 2.0722196102142334
 Step 1340: loss 1.9782861471176147
 Step 1350: loss 2.077576160430908
 Step 1360: loss 1.9759315252304077
 Step 1370: loss 1.9864885807037354
 Step 1380: loss 1.9175422191619873
 Step 1390: loss 2.0092742443084717
 Step 1400: loss 1.9471938610076904
 [Seed 29] Step 1400 --> val loss 2.217451810836792
 Step 1410: loss 2.0206093788146973
 Step 1420: loss 1.9409898519515991
 Step 1430: loss 1.8398635387420654
 Step 1440: loss 2.0006308555603027
 Step 1450: loss 1.952970266342163
 Step 1460: loss 1.864903211593628
 Step 1470: loss 1.919324278831482
 Step 1480: loss 1.8156728744506836
 Step 1490: loss 1.9377562999725342
 Step 1500: loss 1.8151962757110596
 Step 1510: loss 1.9185168743133545
 Step 1520: loss 1.9179487228393555
 Step 1530: loss 1.8487966060638428
 Step 1540: loss 1.8549344539642334
 Step 1550: loss 1.8470637798309326
 Step 1560: loss 1.86884343624115
 Step 1570: loss 1.8461854457855225
 Step 1580: loss 1.803917407989502
 Step 1590: loss 1.8272883892059326
 Step 1600: loss 1.9483985900878906
 [Seed 29] Step 1600 --> val loss 2.080524444580078
 Step 1610: loss 1.8333206176757812
 Step 1620: loss 1.8195550441741943
 Step 1630: loss 1.815901517868042
 Step 1640: loss 1.8116779327392578
 Step 1650: loss 1.7146649360656738
 Step 1660: loss 1.7759058475494385
 Step 1670: loss 1.863523244857788
 Step 1680: loss 1.7506647109985352
 Step 1690: loss 1.8702764511108398
 Step 1700: loss 1.8950955867767334
 Step 1710: loss 1.8372786045074463
 Step 1720: loss 1.8683618307113647
 Step 1730: loss 1.83780038356781
 Step 1740: loss 1.815588355064392
 Step 1750: loss 1.8774511814117432
 Step 1760: loss 1.8335497379302979
 Step 1770: loss 1.969738245010376
 Step 1780: loss 1.7423732280731201
 Step 1790: loss 1.7203733921051025
 Step 1800: loss 1.7367799282073975
 [Seed 29] Step 1800 --> val loss 1.9944556951522827
 Step 1810: loss 1.8740241527557373
 Step 1820: loss 1.8627197742462158
 Step 1830: loss 1.8030831813812256
 Step 1840: loss 1.7922900915145874
 Step 1850: loss 1.659294843673706
 Step 1860: loss 1.872157096862793
 Step 1870: loss 1.8078086376190186
 Step 1880: loss 1.8942551612854004
 Step 1890: loss 1.8866593837738037
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 29] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 58 46 43  1 54 43 39 41 43] ...
 First target sequence: [58 46 43  1 54 43 39 41 43  8] ...

 Decoded input:  the peace.

 PARIS:
 Of honourable reckoning are you both;
 And pi

 Decoded target: the peace.

 PARIS:
 Of honourable reckoning are you both;
 And pit
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2110490798950195
 [Seed 30] Step 0 --> val loss 4.184312343597412
 Step 10: loss 3.8455560207366943
 Step 20: loss 3.3130006790161133
 Step 30: loss 2.7967185974121094
 Step 40: loss 2.6613388061523438
 Step 50: loss 2.47694730758667
 Step 60: loss 2.4576756954193115
 Step 70: loss 2.4911935329437256
 Step 80: loss 2.3759303092956543
 Step 90: loss 2.3841137886047363
 Step 100: loss 2.2910852432250977
 Step 110: loss 2.366567373275757
 Step 120: loss 2.2472052574157715
 Step 130: loss 2.215925693511963
 Step 140: loss 2.215348958969116
 Step 150: loss 2.1552846431732178
 Step 160: loss 2.053246259689331
 Step 170: loss 2.139148473739624
 Step 180: loss 2.195326805114746
 Step 190: loss 2.0097415447235107
 Step 200: loss 2.1283092498779297
 [Seed 30] Step 200 --> val loss 2.2971787452697754
 Step 210: loss 2.093669891357422
 Step 220: loss 2.027400016784668
 Step 230: loss 2.196803092956543
 Step 240: loss 2.129493474960327
 Step 250: loss 2.1439671516418457
 Step 260: loss 2.1603617668151855
 Step 270: loss 2.0300731658935547
 Step 280: loss 2.148127317428589
 Step 290: loss 2.045170307159424
 Step 300: loss 2.1563491821289062
 Step 310: loss 2.068775177001953
 Step 320: loss 2.1240384578704834
 Step 330: loss 2.106346368789673
 Step 340: loss 2.0389037132263184
 Step 350: loss 2.0376431941986084
 Step 360: loss 2.0652387142181396
 Step 370: loss 2.029134750366211
 Step 380: loss 1.9770716428756714
 Step 390: loss 1.9381145238876343
 Step 400: loss 1.9905017614364624
 [Seed 30] Step 400 --> val loss 2.2163138389587402
 Step 410: loss 1.9931976795196533
 Step 420: loss 1.9772000312805176
 Step 430: loss 1.914720058441162
 Step 440: loss 1.9979016780853271
 Step 450: loss 2.0208752155303955
 Step 460: loss 1.9334796667099
 Step 470: loss 2.1123273372650146
 Step 480: loss 2.002351760864258
 Step 490: loss 1.9797357320785522
 Step 500: loss 1.950326919555664
 Step 510: loss 2.1013994216918945
 Step 520: loss 2.0059728622436523
 Step 530: loss 1.9935258626937866
 Step 540: loss 2.082869291305542
 Step 550: loss 1.9808257818222046
 Step 560: loss 1.951914668083191
 Step 570: loss 1.9962263107299805
 Step 580: loss 1.8708720207214355
 Step 590: loss 1.9862592220306396
 Step 600: loss 2.1629791259765625
 [Seed 30] Step 600 --> val loss 2.136021375656128
 Step 610: loss 1.8457896709442139
 Step 620: loss 1.9151840209960938
 Step 630: loss 1.9868887662887573
 Step 640: loss 2.0378637313842773
 Step 650: loss 1.9234769344329834
 Step 660: loss 2.0277631282806396
 Step 670: loss 2.0268077850341797
 Step 680: loss 2.006774663925171
 Step 690: loss 1.94173264503479
 Step 700: loss 1.923385500907898
 Step 710: loss 2.041717290878296
 Step 720: loss 1.9470914602279663
 Step 730: loss 1.9355485439300537
 Step 740: loss 1.9620816707611084
 Step 750: loss 1.8921196460723877
 Step 760: loss 1.831763505935669
 Step 770: loss 2.089387893676758
 Step 780: loss 1.9182217121124268
 Step 790: loss 1.9831187725067139
 Step 800: loss 2.021672248840332
 [Seed 30] Step 800 --> val loss 2.2328317165374756
 Step 810: loss 2.0265698432922363
 Step 820: loss 2.02567982673645
 Step 830: loss 2.0402143001556396
 Step 840: loss 1.9981114864349365
 Step 850: loss 1.9693913459777832
 Step 860: loss 1.9288796186447144
 Step 870: loss 1.9743733406066895
 Step 880: loss 2.0397653579711914
 Step 890: loss 1.9602913856506348
 Step 900: loss 1.9728269577026367
 Step 910: loss 1.9455475807189941
 Step 920: loss 1.9495867490768433
 Step 930: loss 2.0033457279205322
 Step 940: loss 1.885713815689087
 Step 950: loss 1.8273935317993164
 Step 960: loss 1.8809881210327148
 Step 970: loss 1.9437940120697021
 Step 980: loss 1.8882701396942139
 Step 990: loss 1.8656518459320068
 Step 1000: loss 2.0664162635803223
 [Seed 30] Step 1000 --> val loss 2.1943485736846924
 Step 1010: loss 1.8642537593841553
 Step 1020: loss 1.9786405563354492
 Step 1030: loss 1.929714560508728
 Step 1040: loss 2.106363534927368
 Step 1050: loss 1.9857652187347412
 Step 1060: loss 2.0488202571868896
 Step 1070: loss 1.8822176456451416
 Step 1080: loss 1.8926501274108887
 Step 1090: loss 1.9686601161956787
 Step 1100: loss 1.9199155569076538
 Step 1110: loss 1.9089325666427612
 Step 1120: loss 1.976232886314392
 Step 1130: loss 1.922255277633667
 Step 1140: loss 1.9948973655700684
 Step 1150: loss 1.99433171749115
 Step 1160: loss 1.9438894987106323
 Step 1170: loss 1.8617022037506104
 Step 1180: loss 1.8398422002792358
 Step 1190: loss 1.7774319648742676
 Step 1200: loss 1.9246248006820679
 [Seed 30] Step 1200 --> val loss 2.072420835494995
 Step 1210: loss 1.8472070693969727
 Step 1220: loss 1.8232157230377197
 Step 1230: loss 1.9167354106903076
 Step 1240: loss 1.8957488536834717
 Step 1250: loss 1.9021533727645874
 Step 1260: loss 1.9415102005004883
 Step 1270: loss 1.904707908630371
 Step 1280: loss 1.9638214111328125
 Step 1290: loss 1.894486427307129
 Step 1300: loss 1.8235740661621094
 Step 1310: loss 1.9623054265975952
 Step 1320: loss 1.8778445720672607
 Step 1330: loss 1.7909506559371948
 Step 1340: loss 1.895042896270752
 Step 1350: loss 1.8733323812484741
 Step 1360: loss 1.7554450035095215
 Step 1370: loss 1.8540486097335815
 Step 1380: loss 1.7875165939331055
 Step 1390: loss 1.7839868068695068
 Step 1400: loss 1.898877501487732
 [Seed 30] Step 1400 --> val loss 1.9117746353149414
 Step 1410: loss 1.7560744285583496
 Step 1420: loss 1.8068279027938843
 Step 1430: loss 1.7533010244369507
 Step 1440: loss 1.7668222188949585
 Step 1450: loss 1.8262513875961304
 Step 1460: loss 1.7820667028427124
 Step 1470: loss 1.9087942838668823
 Step 1480: loss 1.6729267835617065
 Step 1490: loss 1.8570998907089233
 Step 1500: loss 1.752570629119873
 Step 1510: loss 1.7030954360961914
 Step 1520: loss 1.7704226970672607
 Step 1530: loss 1.732966423034668
 Step 1540: loss 1.61625337600708
 Step 1550: loss 1.7958793640136719
 Step 1560: loss 1.8557488918304443
 Step 1570: loss 1.6752078533172607
 Step 1580: loss 1.623639464378357
 Step 1590: loss 1.7364063262939453
 Step 1600: loss 1.7041786909103394
 [Seed 30] Step 1600 --> val loss 1.9659998416900635
 Step 1610: loss 1.7697995901107788
 Step 1620: loss 1.7780259847640991
 Step 1630: loss 1.7485822439193726
 Step 1640: loss 1.7693742513656616
 Step 1650: loss 1.764246940612793
 Step 1660: loss 1.6833770275115967
 Step 1670: loss 1.6958577632904053
 Step 1680: loss 1.654908537864685
 Step 1690: loss 1.7861096858978271
 Step 1700: loss 1.7143945693969727
 Step 1710: loss 1.6794589757919312
 Step 1720: loss 1.6882232427597046
 Step 1730: loss 1.738796591758728
 Step 1740: loss 1.737044095993042
 Step 1750: loss 1.654106616973877
 Step 1760: loss 1.7458057403564453
 Step 1770: loss 1.587013602256775
 Step 1780: loss 1.659614086151123
 Step 1790: loss 1.6275014877319336
 Step 1800: loss 1.7366665601730347
 [Seed 30] Step 1800 --> val loss 1.791337013244629
 Step 1810: loss 1.676041603088379
 Step 1820: loss 1.753009557723999
 Step 1830: loss 1.666666030883789
 Step 1840: loss 1.70638108253479
 Step 1850: loss 1.6330798864364624
 Step 1860: loss 1.6346687078475952
 Step 1870: loss 1.6745790243148804
 Step 1880: loss 1.6684801578521729
 Step 1890: loss 1.8592565059661865
 Step 1900: loss 1.6482570171356201
 Step 1910: loss 1.6197593212127686
 Step 1920: loss 1.6441495418548584
 Step 1930: loss 1.6303224563598633
 Step 1940: loss 1.6983027458190918
 Step 1950: loss 1.536661982536316
 Step 1960: loss 1.7917163372039795
 Step 1970: loss 1.6457421779632568
 Step 1980: loss 1.6943072080612183
 Step 1990: loss 1.6684545278549194
 Step 2000: loss 1.6778377294540405
 [Seed 30] Step 2000 --> val loss 1.7552546262741089
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 46 47 58 46 43 56  1 52 53] ...
 First target sequence: [46 47 58 46 43 56  1 52 53 61] ...

 Decoded input:  hither now to slaughter thee.

 CLARENCE:
 It cannot be; for when

 Decoded target: hither now to slaughter thee.

 CLARENCE:
 It cannot be; for when 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.206593036651611
 [Seed 31] Step 0 --> val loss 4.18508768081665
 Step 10: loss 3.8413772583007812
 Step 20: loss 3.321992874145508
 Step 30: loss 2.8796424865722656
 Step 40: loss 2.681732177734375
 Step 50: loss 2.6223835945129395
 Step 60: loss 2.515770435333252
 Step 70: loss 2.4306256771087646
 Step 80: loss 2.3830621242523193
 Step 90: loss 2.412374973297119
 Step 100: loss 2.259230375289917
 Step 110: loss 2.2609219551086426
 Step 120: loss 2.1932497024536133
 Step 130: loss 2.2368974685668945
 Step 140: loss 2.207101583480835
 Step 150: loss 2.219590663909912
 Step 160: loss 2.1421170234680176
 Step 170: loss 2.227947235107422
 Step 180: loss 2.175100803375244
 Step 190: loss 1.9730825424194336
 Step 200: loss 2.1257433891296387
 [Seed 31] Step 200 --> val loss 2.4299981594085693
 Step 210: loss 2.152315378189087
 Step 220: loss 2.1271848678588867
 Step 230: loss 2.139463424682617
 Step 240: loss 2.0306575298309326
 Step 250: loss 2.0862443447113037
 Step 260: loss 2.069854259490967
 Step 270: loss 2.1517715454101562
 Step 280: loss 2.195887327194214
 Step 290: loss 2.0481467247009277
 Step 300: loss 2.083200216293335
 Step 310: loss 2.1930465698242188
 Step 320: loss 2.154737949371338
 Step 330: loss 1.9429844617843628
 Step 340: loss 2.0156593322753906
 Step 350: loss 2.012990713119507
 Step 360: loss 2.130063772201538
 Step 370: loss 2.008801221847534
 Step 380: loss 2.0609545707702637
 Step 390: loss 2.085477352142334
 Step 400: loss 2.1656177043914795
 [Seed 31] Step 400 --> val loss 2.2424843311309814
 Step 410: loss 2.1301980018615723
 Step 420: loss 2.054080009460449
 Step 430: loss 2.081425666809082
 Step 440: loss 2.038395881652832
 Step 450: loss 2.060079574584961
 Step 460: loss 2.068061590194702
 Step 470: loss 1.896832823753357
 Step 480: loss 2.028688669204712
 Step 490: loss 1.966022253036499
 Step 500: loss 1.9923375844955444
 Step 510: loss 2.004887580871582
 Step 520: loss 2.1175172328948975
 Step 530: loss 2.063302516937256
 Step 540: loss 1.9938621520996094
 Step 550: loss 1.997549057006836
 Step 560: loss 1.892393946647644
 Step 570: loss 2.12886118888855
 Step 580: loss 2.0246105194091797
 Step 590: loss 2.040621757507324
 Step 600: loss 1.9562139511108398
 [Seed 31] Step 600 --> val loss 2.184394121170044
 Step 610: loss 2.0035157203674316
 Step 620: loss 1.9898953437805176
 Step 630: loss 1.9706122875213623
 Step 640: loss 2.0435304641723633
 Step 650: loss 2.030294418334961
 Step 660: loss 2.056196689605713
 Step 670: loss 2.0871798992156982
 Step 680: loss 2.0073084831237793
 Step 690: loss 1.9780186414718628
 Step 700: loss 2.1927196979522705
 Step 710: loss 2.0400800704956055
 Step 720: loss 1.8889198303222656
 Step 730: loss 1.960003137588501
 Step 740: loss 1.9258379936218262
 Step 750: loss 1.936946153640747
 Step 760: loss 2.083449125289917
 Step 770: loss 2.1005468368530273
 Step 780: loss 1.9751704931259155
 Step 790: loss 2.0876364707946777
 Step 800: loss 2.153634548187256
 [Seed 31] Step 800 --> val loss 2.170968770980835
 Step 810: loss 1.9527791738510132
 Step 820: loss 2.0946316719055176
 Step 830: loss 2.0712931156158447
 Step 840: loss 2.049715995788574
 Step 850: loss 2.066861152648926
 Step 860: loss 1.9349544048309326
 Step 870: loss 2.029510974884033
 Step 880: loss 2.0658416748046875
 Step 890: loss 2.011587619781494
 Step 900: loss 2.030813694000244
 Step 910: loss 2.0376129150390625
 Step 920: loss 1.8924598693847656
 Step 930: loss 2.0933074951171875
 Step 940: loss 2.0302329063415527
 Step 950: loss 2.0243024826049805
 Step 960: loss 2.122042417526245
 Step 970: loss 1.945787787437439
 Step 980: loss 2.0976860523223877
 Step 990: loss 2.005704641342163
 Step 1000: loss 2.093961238861084
 [Seed 31] Step 1000 --> val loss 2.2514829635620117
 Step 1010: loss 2.225630283355713
 Step 1020: loss 2.163806438446045
 Step 1030: loss 2.1654410362243652
 Step 1040: loss 2.092498540878296
 Step 1050: loss 2.1435251235961914
 Step 1060: loss 2.010371208190918
 Step 1070: loss 2.01899790763855
 Step 1080: loss 2.0618391036987305
 Step 1090: loss 2.128330945968628
 Step 1100: loss 1.9470109939575195
 Step 1110: loss 2.0606250762939453
 Step 1120: loss 2.130061626434326
 Step 1130: loss 1.996849775314331
 Step 1140: loss 2.0215961933135986
 Step 1150: loss 2.0812559127807617
 Step 1160: loss 2.05549693107605
 Step 1170: loss 2.0024843215942383
 Step 1180: loss 2.0278096199035645
 Step 1190: loss 2.02644681930542
 Step 1200: loss 2.1037840843200684
 [Seed 31] Step 1200 --> val loss 2.2554805278778076
 Step 1210: loss 1.9659702777862549
 Step 1220: loss 2.019937515258789
 Step 1230: loss 2.050835609436035
 Step 1240: loss 2.079394578933716
 Step 1250: loss 2.0458433628082275
 Step 1260: loss 2.0697708129882812
 Step 1270: loss 2.0326812267303467
 Step 1280: loss 1.998785376548767
 Step 1290: loss 1.977677345275879
 Step 1300: loss 1.8717154264450073
 Step 1310: loss 2.0365138053894043
 Step 1320: loss 1.9523528814315796
 Step 1330: loss 2.019315719604492
 Step 1340: loss 2.0103139877319336
 Step 1350: loss 1.7977697849273682
 Step 1360: loss 2.0239171981811523
 Step 1370: loss 1.980023741722107
 Step 1380: loss 2.0930137634277344
 Step 1390: loss 1.9765703678131104
 Step 1400: loss 2.0151333808898926
 [Seed 31] Step 1400 --> val loss 2.157376289367676
 Step 1410: loss 1.8481086492538452
 Step 1420: loss 1.9188778400421143
 Step 1430: loss 1.9748597145080566
 Step 1440: loss 1.8654966354370117
 Step 1450: loss 1.800511360168457
 Step 1460: loss 1.8834785223007202
 Step 1470: loss 1.8779642581939697
 Step 1480: loss 1.932478666305542
 Step 1490: loss 1.835633635520935
 Step 1500: loss 1.9251797199249268
 Step 1510: loss 1.9206674098968506
 Step 1520: loss 1.878190040588379
 Step 1530: loss 1.8801753520965576
 Step 1540: loss 1.943229079246521
 Step 1550: loss 1.938148856163025
 Step 1560: loss 1.9018704891204834
 Step 1570: loss 1.9010686874389648
 Step 1580: loss 2.036015272140503
 Step 1590: loss 1.8492090702056885
 Step 1600: loss 1.87030029296875
 [Seed 31] Step 1600 --> val loss 2.045478582382202
 Step 1610: loss 1.9051318168640137
 Step 1620: loss 1.7969131469726562
 Step 1630: loss 1.895001769065857
 Step 1640: loss 1.9142179489135742
 Step 1650: loss 1.9754791259765625
 Step 1660: loss 1.713505506515503
 Step 1670: loss 1.9221303462982178
 Step 1680: loss 2.010915517807007
 Step 1690: loss 1.806166648864746
 Step 1700: loss 1.8547242879867554
 Step 1710: loss 1.9023512601852417
 Step 1720: loss 1.9266307353973389
 Step 1730: loss 1.9717477560043335
 Step 1740: loss 1.8338003158569336
 Step 1750: loss 1.829959750175476
 Step 1760: loss 2.0960230827331543
 Step 1770: loss 1.9254539012908936
 Step 1780: loss 1.9177006483078003
 Step 1790: loss 1.7078819274902344
 Step 1800: loss 1.7325592041015625
 [Seed 31] Step 1800 --> val loss 2.107344388961792
 Step 1810: loss 1.7362935543060303
 Step 1820: loss 1.846743106842041
 Step 1830: loss 1.730191946029663
 Step 1840: loss 1.877624750137329
 Step 1850: loss 1.9216645956039429
 Step 1860: loss 1.9214935302734375
 Step 1870: loss 1.855900764465332
 Step 1880: loss 1.98788583278656
 Step 1890: loss 1.7906780242919922
 Step 1900: loss 1.8384320735931396
 Step 1910: loss 1.8716591596603394
 Step 1920: loss 1.8510520458221436
 Step 1930: loss 1.7733690738677979
 Step 1940: loss 1.7725520133972168
 Step 1950: loss 1.9210989475250244
 Step 1960: loss 1.7606487274169922
 Step 1970: loss 1.8459866046905518
 Step 1980: loss 1.7657159566879272
 Step 1990: loss 1.840493083000183
 Step 2000: loss 1.7680633068084717
 [Seed 31] Step 2000 --> val loss 2.0175182819366455
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 44 56 53 51  1 58 46 43  1] ...
 First target sequence: [44 56 53 51  1 58 46 43  1 51] ...

 Decoded input:  from the manacles
 Of the all-building law; and that there were


 Decoded target: from the manacles
 Of the all-building law; and that there were
 N
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.21236515045166
 [Seed 32] Step 0 --> val loss 4.183606147766113
 Step 10: loss 3.84712553024292
 Step 20: loss 3.3049750328063965
 Step 30: loss 2.8630566596984863
 Step 40: loss 2.7247350215911865
 Step 50: loss 2.6292877197265625
 Step 60: loss 2.456404209136963
 Step 70: loss 2.4236302375793457
 Step 80: loss 2.4849376678466797
 Step 90: loss 2.238617420196533
 Step 100: loss 2.244410991668701
 Step 110: loss 2.248016834259033
 Step 120: loss 2.2522759437561035
 Step 130: loss 2.2343931198120117
 Step 140: loss 2.1309423446655273
 Step 150: loss 2.23618221282959
 Step 160: loss 2.2646498680114746
 Step 170: loss 2.1334235668182373
 Step 180: loss 2.1334333419799805
 Step 190: loss 2.1912801265716553
 Step 200: loss 2.104227066040039
 [Seed 32] Step 200 --> val loss 2.356635093688965
 Step 210: loss 2.079256772994995
 Step 220: loss 2.0777177810668945
 Step 230: loss 2.126988410949707
 Step 240: loss 2.0340735912323
 Step 250: loss 2.0991628170013428
 Step 260: loss 2.063385009765625
 Step 270: loss 2.1169142723083496
 Step 280: loss 2.034369468688965
 Step 290: loss 2.110480308532715
 Step 300: loss 2.0578296184539795
 Step 310: loss 2.126715898513794
 Step 320: loss 2.078436851501465
 Step 330: loss 1.965358018875122
 Step 340: loss 2.048766851425171
 Step 350: loss 2.1084465980529785
 Step 360: loss 1.9784259796142578
 Step 370: loss 2.0278000831604004
 Step 380: loss 2.13989520072937
 Step 390: loss 2.126844644546509
 Step 400: loss 2.0048704147338867
 [Seed 32] Step 400 --> val loss 2.3650732040405273
 Step 410: loss 2.0211315155029297
 Step 420: loss 2.0753419399261475
 Step 430: loss 2.12861704826355
 Step 440: loss 1.9956064224243164
 Step 450: loss 2.0154178142547607
 Step 460: loss 1.8895130157470703
 Step 470: loss 2.0331788063049316
 Step 480: loss 2.0504097938537598
 Step 490: loss 2.0964548587799072
 Step 500: loss 2.151047706604004
 Step 510: loss 2.121124744415283
 Step 520: loss 2.0075318813323975
 Step 530: loss 2.0295233726501465
 Step 540: loss 1.9758003950119019
 Step 550: loss 2.0427582263946533
 Step 560: loss 1.9189996719360352
 Step 570: loss 2.066715955734253
 Step 580: loss 2.0409789085388184
 Step 590: loss 2.048555850982666
 Step 600: loss 2.0025458335876465
 [Seed 32] Step 600 --> val loss 2.3661868572235107
 Step 610: loss 1.9605224132537842
 Step 620: loss 2.046299457550049
 Step 630: loss 2.0113744735717773
 Step 640: loss 2.1794700622558594
 Step 650: loss 2.0743987560272217
 Step 660: loss 2.021699905395508
 Step 670: loss 1.9062869548797607
 Step 680: loss 1.9319944381713867
 Step 690: loss 2.0192418098449707
 Step 700: loss 2.0791563987731934
 Step 710: loss 2.155069351196289
 Step 720: loss 2.0715622901916504
 Step 730: loss 2.156352996826172
 Step 740: loss 2.1409549713134766
 Step 750: loss 2.0239627361297607
 Step 760: loss 2.082716941833496
 Step 770: loss 2.0669403076171875
 Step 780: loss 2.090745687484741
 Step 790: loss 2.0680043697357178
 Step 800: loss 1.964010238647461
 [Seed 32] Step 800 --> val loss 2.287419557571411
 Step 810: loss 2.0202722549438477
 Step 820: loss 2.110745429992676
 Step 830: loss 2.076493263244629
 Step 840: loss 2.0476770401000977
 Step 850: loss 2.1084985733032227
 Step 860: loss 1.9929227828979492
 Step 870: loss 2.052966594696045
 Step 880: loss 1.938802719116211
 Step 890: loss 1.9377063512802124
 Step 900: loss 1.9641330242156982
 Step 910: loss 1.9895875453948975
 Step 920: loss 2.0720598697662354
 Step 930: loss 2.0636041164398193
 Step 940: loss 2.0019643306732178
 Step 950: loss 1.9791593551635742
 Step 960: loss 1.9369134902954102
 Step 970: loss 1.8581390380859375
 Step 980: loss 2.075291872024536
 Step 990: loss 2.042048454284668
 Step 1000: loss 1.9619321823120117
 [Seed 32] Step 1000 --> val loss 2.2887027263641357
 Step 1010: loss 1.9365346431732178
 Step 1020: loss 2.1289567947387695
 Step 1030: loss 1.9494127035140991
 Step 1040: loss 1.9864780902862549
 Step 1050: loss 1.9525226354599
 Step 1060: loss 1.8404277563095093
 Step 1070: loss 2.0911872386932373
 Step 1080: loss 1.9807653427124023
 Step 1090: loss 1.9822415113449097
 Step 1100: loss 1.9837650060653687
 Step 1110: loss 2.0201528072357178
 Step 1120: loss 1.8995568752288818
 Step 1130: loss 2.0291690826416016
 Step 1140: loss 2.0585944652557373
 Step 1150: loss 1.9794577360153198
 Step 1160: loss 1.9697794914245605
 Step 1170: loss 2.0227320194244385
 Step 1180: loss 1.926918387413025
 Step 1190: loss 1.8688738346099854
 Step 1200: loss 2.049473762512207
 [Seed 32] Step 1200 --> val loss 2.2209012508392334
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 32] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 32] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 32] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 32] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [39 52 52 63  0 32 56 43 51 40] ...
 First target sequence: [52 52 63  0 32 56 43 51 40 50] ...

 Decoded input: anny
 Tremble at patience. You, my lord, best know,
 Who least wil

 Decoded target: nny
 Tremble at patience. You, my lord, best know,
 Who least will
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.204583168029785
 [Seed 33] Step 0 --> val loss 4.185637950897217
 Step 10: loss 3.841442584991455
 Step 20: loss 3.3419322967529297
 Step 30: loss 2.8267600536346436
 Step 40: loss 2.6405694484710693
 Step 50: loss 2.5945351123809814
 Step 60: loss 2.449289560317993
 Step 70: loss 2.4411396980285645
 Step 80: loss 2.410566806793213
 Step 90: loss 2.3832318782806396
 Step 100: loss 2.2271852493286133
 Step 110: loss 2.259866237640381
 Step 120: loss 2.203866720199585
 Step 130: loss 2.2215490341186523
 Step 140: loss 2.1134748458862305
 Step 150: loss 2.26835298538208
 Step 160: loss 2.1359126567840576
 Step 170: loss 2.1102957725524902
 Step 180: loss 2.2024097442626953
 Step 190: loss 2.105490207672119
 Step 200: loss 2.1044728755950928
 [Seed 33] Step 200 --> val loss 2.4074513912200928
 Step 210: loss 2.1061017513275146
 Step 220: loss 2.178560256958008
 Step 230: loss 2.1315600872039795
 Step 240: loss 2.044224262237549
 Step 250: loss 2.076420545578003
 Step 260: loss 2.0589373111724854
 Step 270: loss 2.0657124519348145
 Step 280: loss 2.090778350830078
 Step 290: loss 2.041318655014038
 Step 300: loss 2.11466383934021
 Step 310: loss 2.1591174602508545
 Step 320: loss 1.9567911624908447
 Step 330: loss 1.9055192470550537
 Step 340: loss 2.0471749305725098
 Step 350: loss 2.0762476921081543
 Step 360: loss 2.120894193649292
 Step 370: loss 2.047027349472046
 Step 380: loss 2.008544921875
 Step 390: loss 1.9705390930175781
 Step 400: loss 2.0130128860473633
 [Seed 33] Step 400 --> val loss 2.2921791076660156
 Step 410: loss 1.9894146919250488
 Step 420: loss 1.9913673400878906
 Step 430: loss 2.050868272781372
 Step 440: loss 2.0060019493103027
 Step 450: loss 2.036663055419922
 Step 460: loss 2.0761358737945557
 Step 470: loss 1.9867233037948608
 Step 480: loss 1.99540376663208
 Step 490: loss 1.9887217283248901
 Step 500: loss 2.075843334197998
 Step 510: loss 1.9908103942871094
 Step 520: loss 2.032259941101074
 Step 530: loss 2.0353596210479736
 Step 540: loss 2.1132659912109375
 Step 550: loss 2.09184193611145
 Step 560: loss 2.0296568870544434
 Step 570: loss 2.073944568634033
 Step 580: loss 2.0474908351898193
 Step 590: loss 2.062593936920166
 Step 600: loss 2.0073468685150146
 [Seed 33] Step 600 --> val loss 2.168912887573242
 Step 610: loss 2.0602617263793945
 Step 620: loss 1.9161670207977295
 Step 630: loss 2.0773611068725586
 Step 640: loss 2.0896050930023193
 Step 650: loss 2.1330535411834717
 Step 660: loss 2.147031307220459
 Step 670: loss 2.0978951454162598
 Step 680: loss 2.1278162002563477
 Step 690: loss 2.112734794616699
 Step 700: loss 2.014336585998535
 Step 710: loss 2.0261988639831543
 Step 720: loss 2.099311590194702
 Step 730: loss 2.0867209434509277
 Step 740: loss 2.0418105125427246
 Step 750: loss 2.1149778366088867
 Step 760: loss 2.10139536857605
 Step 770: loss 2.1296298503875732
 Step 780: loss 2.0618958473205566
 Step 790: loss 2.0518455505371094
 Step 800: loss 1.998706340789795
 [Seed 33] Step 800 --> val loss 2.384385824203491
 Step 810: loss 2.135911703109741
 Step 820: loss 2.03519344329834
 Step 830: loss 2.092179775238037
 Step 840: loss 2.1103172302246094
 Step 850: loss 2.183957099914551
 Step 860: loss 1.9728477001190186
 Step 870: loss 2.074995756149292
 Step 880: loss 2.1399049758911133
 Step 890: loss 2.1101701259613037
 Step 900: loss 2.0299274921417236
 Step 910: loss 2.029348850250244
 Step 920: loss 2.0307393074035645
 Step 930: loss 2.082916259765625
 Step 940: loss 1.98384690284729
 Step 950: loss 1.9989993572235107
 Step 960: loss 2.17463755607605
 Step 970: loss 1.8626861572265625
 Step 980: loss 2.0864455699920654
 Step 990: loss 1.8808865547180176
 Step 1000: loss 1.9834866523742676
 [Seed 33] Step 1000 --> val loss 2.158047914505005
 Step 1010: loss 2.118818998336792
 Step 1020: loss 2.0098133087158203
 Step 1030: loss 2.072169303894043
 Step 1040: loss 2.053090810775757
 Step 1050: loss 2.1954505443573
 Step 1060: loss 2.0961203575134277
 Step 1070: loss 2.051938533782959
 Step 1080: loss 1.974507212638855
 Step 1090: loss 2.0029120445251465
 Step 1100: loss 2.060884952545166
 Step 1110: loss 2.204561471939087
 Step 1120: loss 2.109652519226074
 Step 1130: loss 2.1742806434631348
 Step 1140: loss 2.174313545227051
 Step 1150: loss 2.1041171550750732
 Step 1160: loss 2.0733890533447266
 Step 1170: loss 2.1019680500030518
 Step 1180: loss 2.039708137512207
 Step 1190: loss 2.113900661468506
 Step 1200: loss 2.0376391410827637
 [Seed 33] Step 1200 --> val loss 2.4406259059906006
 Step 1210: loss 2.1568918228149414
 Step 1220: loss 2.0111958980560303
 Step 1230: loss 1.9832344055175781
 Step 1240: loss 2.055910110473633
 Step 1250: loss 2.1059763431549072
 Step 1260: loss 2.040152072906494
 Step 1270: loss 1.9648466110229492
 Step 1280: loss 2.2042484283447266
 Step 1290: loss 2.1077933311462402
 Step 1300: loss 2.0522632598876953
 Step 1310: loss 1.9198874235153198
 Step 1320: loss 2.0278677940368652
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 33] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 33] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 33] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 33] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 7  7  0  0 29 33 17 17 26  1] ...
 First target sequence: [ 7  0  0 29 33 17 17 26  1 17] ...

 Decoded input: --

 QUEEN ELIZABETH:
 Profaned, dishonour'd, and the third usurp'

 Decoded target: -

 QUEEN ELIZABETH:
 Profaned, dishonour'd, and the third usurp'd
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.205196857452393
 [Seed 34] Step 0 --> val loss 4.185079574584961
 Step 10: loss 3.865683078765869
 Step 20: loss 3.305929183959961
 Step 30: loss 2.7597408294677734
 Step 40: loss 2.6812381744384766
 Step 50: loss 2.603386402130127
 Step 60: loss 2.4838151931762695
 Step 70: loss 2.4858744144439697
 Step 80: loss 2.2736918926239014
 Step 90: loss 2.3499772548675537
 Step 100: loss 2.2946503162384033
 Step 110: loss 2.2476234436035156
 Step 120: loss 2.2495298385620117
 Step 130: loss 2.1899330615997314
 Step 140: loss 2.1386375427246094
 Step 150: loss 2.2310245037078857
 Step 160: loss 2.062187671661377
 Step 170: loss 2.105288028717041
 Step 180: loss 2.1151950359344482
 Step 190: loss 2.0928702354431152
 Step 200: loss 2.112466812133789
 [Seed 34] Step 200 --> val loss 2.3677477836608887
 Step 210: loss 2.127309560775757
 Step 220: loss 2.128159284591675
 Step 230: loss 2.0205159187316895
 Step 240: loss 2.1333508491516113
 Step 250: loss 2.005166530609131
 Step 260: loss 1.974428415298462
 Step 270: loss 1.9910796880722046
 Step 280: loss 2.0804545879364014
 Step 290: loss 1.9170410633087158
 Step 300: loss 2.1684305667877197
 Step 310: loss 2.086322784423828
 Step 320: loss 1.9470762014389038
 Step 330: loss 2.088090658187866
 Step 340: loss 1.9567055702209473
 Step 350: loss 2.0377089977264404
 Step 360: loss 2.0247879028320312
 Step 370: loss 2.1539783477783203
 Step 380: loss 2.014277458190918
 Step 390: loss 1.9829219579696655
 Step 400: loss 2.068551778793335
 [Seed 34] Step 400 --> val loss 2.205761671066284
 Step 410: loss 2.051896095275879
 Step 420: loss 2.132296562194824
 Step 430: loss 1.9120941162109375
 Step 440: loss 1.9070237874984741
 Step 450: loss 1.9776500463485718
 Step 460: loss 2.0076398849487305
 Step 470: loss 1.9326612949371338
 Step 480: loss 1.974426031112671
 Step 490: loss 1.9650886058807373
 Step 500: loss 2.078059673309326
 Step 510: loss 1.961134672164917
 Step 520: loss 1.980985403060913
 Step 530: loss 1.9215335845947266
 Step 540: loss 1.942044973373413
 Step 550: loss 2.0359909534454346
 Step 560: loss 2.0864996910095215
 Step 570: loss 2.072155475616455
 Step 580: loss 1.8976242542266846
 Step 590: loss 2.0612401962280273
 Step 600: loss 1.9931153059005737
 [Seed 34] Step 600 --> val loss 2.22318434715271
 Step 610: loss 2.023677110671997
 Step 620: loss 2.0472283363342285
 Step 630: loss 1.9358956813812256
 Step 640: loss 2.0129408836364746
 Step 650: loss 1.9317564964294434
 Step 660: loss 2.070164203643799
 Step 670: loss 1.8857921361923218
 Step 680: loss 1.9806321859359741
 Step 690: loss 2.0647902488708496
 Step 700: loss 1.899391531944275
 Step 710: loss 1.9453575611114502
 Step 720: loss 2.0339770317077637
 Step 730: loss 2.0317282676696777
 Step 740: loss 1.8152952194213867
 Step 750: loss 2.0047130584716797
 Step 760: loss 2.0098347663879395
 Step 770: loss 1.928613543510437
 Step 780: loss 1.9429280757904053
 Step 790: loss 1.9329713582992554
 Step 800: loss 1.9872688055038452
 [Seed 34] Step 800 --> val loss 2.1845908164978027
 Step 810: loss 1.9491500854492188
 Step 820: loss 1.9231603145599365
 Step 830: loss 2.001272201538086
 Step 840: loss 1.9752418994903564
 Step 850: loss 2.003509998321533
 Step 860: loss 1.9431933164596558
 Step 870: loss 1.9966135025024414
 Step 880: loss 2.0447633266448975
 Step 890: loss 2.0940911769866943
 Step 900: loss 2.1228175163269043
 Step 910: loss 2.157317638397217
 Step 920: loss 2.067315101623535
 Step 930: loss 1.890594482421875
 Step 940: loss 2.097446918487549
 Step 950: loss 2.1203553676605225
 Step 960: loss 1.9979604482650757
 Step 970: loss 2.0594611167907715
 Step 980: loss 2.072800874710083
 Step 990: loss 2.032179832458496
 Step 1000: loss 2.0245542526245117
 [Seed 34] Step 1000 --> val loss 2.2658402919769287
 Step 1010: loss 2.0555367469787598
 Step 1020: loss 2.0819907188415527
 Step 1030: loss 1.9357194900512695
 Step 1040: loss 2.021359920501709
 Step 1050: loss 1.939305067062378
 Step 1060: loss 1.9616425037384033
 Step 1070: loss 1.9475177526474
 Step 1080: loss 1.9028090238571167
 Step 1090: loss 1.9292364120483398
 Step 1100: loss 2.0074217319488525
 Step 1110: loss 2.010296583175659
 Step 1120: loss 1.9708284139633179
 Step 1130: loss 2.0352864265441895
 Step 1140: loss 1.9424371719360352
 Step 1150: loss 1.9918696880340576
 Step 1160: loss 1.9064323902130127
 Step 1170: loss 1.8971974849700928
 Step 1180: loss 1.9230735301971436
 Step 1190: loss 2.09136962890625
 Step 1200: loss 1.9154958724975586
 [Seed 34] Step 1200 --> val loss 2.188385009765625
 Step 1210: loss 1.8834975957870483
 Step 1220: loss 1.8975534439086914
 Step 1230: loss 1.8873047828674316
 Step 1240: loss 2.0798661708831787
 Step 1250: loss 1.8946740627288818
 Step 1260: loss 1.925931453704834
 Step 1270: loss 1.9412040710449219
 Step 1280: loss 1.9196033477783203
 Step 1290: loss 1.9798247814178467
 Step 1300: loss 1.9752219915390015
 Step 1310: loss 1.9501962661743164
 Step 1320: loss 1.9577178955078125
 Step 1330: loss 1.8031724691390991
 Step 1340: loss 1.9284815788269043
 Step 1350: loss 1.7970027923583984
 Step 1360: loss 1.9666321277618408
 Step 1370: loss 1.863726019859314
 Step 1380: loss 1.915360450744629
 Step 1390: loss 1.866952657699585
 Step 1400: loss 1.8770489692687988
 [Seed 34] Step 1400 --> val loss 2.1218326091766357
 Step 1410: loss 1.856318473815918
 Step 1420: loss 1.915824294090271
 Step 1430: loss 2.070479393005371
 Step 1440: loss 1.9034526348114014
 Step 1450: loss 1.863971471786499
 Step 1460: loss 1.9510282278060913
 Step 1470: loss 1.8020509481430054
 Step 1480: loss 1.903799057006836
 Step 1490: loss 1.8093488216400146
 Step 1500: loss 1.8358547687530518
 Step 1510: loss 1.8017829656600952
 Step 1520: loss 1.8803428411483765
 Step 1530: loss 1.8395345211029053
 Step 1540: loss 1.8370269536972046
 Step 1550: loss 1.9201916456222534
 Step 1560: loss 1.8961261510849
 Step 1570: loss 1.8368120193481445
 Step 1580: loss 1.8282155990600586
 Step 1590: loss 1.9704267978668213
 Step 1600: loss 1.8045518398284912
 [Seed 34] Step 1600 --> val loss 2.01058030128479
 Step 1610: loss 1.8956077098846436
 Step 1620: loss 1.7936408519744873
 Step 1630: loss 1.8321714401245117
 Step 1640: loss 1.6968421936035156
 Step 1650: loss 1.6880218982696533
 Step 1660: loss 1.6132047176361084
 Step 1670: loss 1.7688912153244019
 Step 1680: loss 1.6097438335418701
 Step 1690: loss 1.8530558347702026
 Step 1700: loss 1.7414913177490234
 Step 1710: loss 1.710569143295288
 Step 1720: loss 1.7971676588058472
 Step 1730: loss 1.7397663593292236
 Step 1740: loss 1.8362088203430176
 Step 1750: loss 1.7781403064727783
 Step 1760: loss 1.7298167943954468
 Step 1770: loss 1.7807550430297852
 Step 1780: loss 1.7212737798690796
 Step 1790: loss 1.8455753326416016
 Step 1800: loss 1.7310562133789062
 [Seed 34] Step 1800 --> val loss 1.999359369277954
 Step 1810: loss 1.7547824382781982
 Step 1820: loss 1.8560594320297241
 Step 1830: loss 1.815958023071289
 Step 1840: loss 1.8106086254119873
 Step 1850: loss 1.7805421352386475
 Step 1860: loss 1.7243192195892334
 Step 1870: loss 1.7050063610076904
 Step 1880: loss 1.7555904388427734
 Step 1890: loss 1.680886149406433
 Step 1900: loss 1.891716718673706
 Step 1910: loss 1.5947030782699585
 Step 1920: loss 1.597696304321289
 Step 1930: loss 1.7381585836410522
 Step 1940: loss 1.7726815938949585
 Step 1950: loss 1.7178481817245483
 Step 1960: loss 1.723562240600586
 Step 1970: loss 1.8017711639404297
 Step 1980: loss 1.7280795574188232
 Step 1990: loss 1.7315572500228882
 Step 2000: loss 1.7337602376937866
 [Seed 34] Step 2000 --> val loss 1.959747314453125
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [41 46  5 42  1 46 47 51  1 46] ...
 First target sequence: [46  5 42  1 46 47 51  1 46 53] ...

 Decoded input: ch'd him how he singled Clifford forth.
 Methought he bore him in

 Decoded target: h'd him how he singled Clifford forth.
 Methought he bore him in 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.211065292358398
 [Seed 35] Step 0 --> val loss 4.181889533996582
 Step 10: loss 3.858039379119873
 Step 20: loss 3.3372879028320312
 Step 30: loss 2.9021997451782227
 Step 40: loss 2.6885604858398438
 Step 50: loss 2.5347957611083984
 Step 60: loss 2.5808379650115967
 Step 70: loss 2.371779203414917
 Step 80: loss 2.433350086212158
 Step 90: loss 2.3564653396606445
 Step 100: loss 2.2796525955200195
 Step 110: loss 2.1926841735839844
 Step 120: loss 2.182882785797119
 Step 130: loss 2.1831066608428955
 Step 140: loss 2.1881673336029053
 Step 150: loss 2.1976141929626465
 Step 160: loss 2.0976061820983887
 Step 170: loss 2.158745288848877
 Step 180: loss 2.0606255531311035
 Step 190: loss 2.0496435165405273
 Step 200: loss 2.140064001083374
 [Seed 35] Step 200 --> val loss 2.4274089336395264
 Step 210: loss 2.1011366844177246
 Step 220: loss 2.0461010932922363
 Step 230: loss 2.1392195224761963
 Step 240: loss 2.014861583709717
 Step 250: loss 2.1155385971069336
 Step 260: loss 2.0706162452697754
 Step 270: loss 2.039768934249878
 Step 280: loss 2.094430446624756
 Step 290: loss 2.078939437866211
 Step 300: loss 2.0716400146484375
 Step 310: loss 2.053342342376709
 Step 320: loss 1.9765084981918335
 Step 330: loss 2.012510061264038
 Step 340: loss 1.97921884059906
 Step 350: loss 2.0974435806274414
 Step 360: loss 2.03116512298584
 Step 370: loss 1.958420753479004
 Step 380: loss 2.0966882705688477
 Step 390: loss 2.1155128479003906
 Step 400: loss 2.005091667175293
 [Seed 35] Step 400 --> val loss 2.2648236751556396
 Step 410: loss 1.9377527236938477
 Step 420: loss 2.0341405868530273
 Step 430: loss 2.0101864337921143
 Step 440: loss 1.9496320486068726
 Step 450: loss 2.00445556640625
 Step 460: loss 2.027010917663574
 Step 470: loss 1.9021828174591064
 Step 480: loss 1.9756253957748413
 Step 490: loss 1.95521879196167
 Step 500: loss 1.901845932006836
 Step 510: loss 2.0926625728607178
 Step 520: loss 1.9999117851257324
 Step 530: loss 2.1242008209228516
 Step 540: loss 2.084794521331787
 Step 550: loss 1.9833734035491943
 Step 560: loss 2.1045186519622803
 Step 570: loss 1.8557393550872803
 Step 580: loss 2.017508029937744
 Step 590: loss 2.129044532775879
 Step 600: loss 2.1055073738098145
 [Seed 35] Step 600 --> val loss 2.305534839630127
 Step 610: loss 1.9866020679473877
 Step 620: loss 1.9750322103500366
 Step 630: loss 2.091033935546875
 Step 640: loss 2.107553005218506
 Step 650: loss 2.0292961597442627
 Step 660: loss 2.000607967376709
 Step 670: loss 1.9297780990600586
 Step 680: loss 1.9477351903915405
 Step 690: loss 1.8637590408325195
 Step 700: loss 1.9793285131454468
 Step 710: loss 1.973036766052246
 Step 720: loss 2.046143054962158
 Step 730: loss 2.024731159210205
 Step 740: loss 2.10988450050354
 Step 750: loss 1.9806817770004272
 Step 760: loss 1.959822654724121
 Step 770: loss 2.024271011352539
 Step 780: loss 2.127689838409424
 Step 790: loss 1.9228070974349976
 Step 800: loss 2.0415141582489014
 [Seed 35] Step 800 --> val loss 2.3726205825805664
 Step 810: loss 1.9581010341644287
 Step 820: loss 2.0686492919921875
 Step 830: loss 2.0189743041992188
 Step 840: loss 2.0059821605682373
 Step 850: loss 2.0158581733703613
 Step 860: loss 2.079749584197998
 Step 870: loss 2.0622220039367676
 Step 880: loss 2.035611867904663
 Step 890: loss 1.948541522026062
 Step 900: loss 2.085465908050537
 Step 910: loss 2.1784770488739014
 Step 920: loss 1.947666883468628
 Step 930: loss 1.987065076828003
 Step 940: loss 1.9673106670379639
 Step 950: loss 1.937096118927002
 Step 960: loss 1.8748767375946045
 Step 970: loss 2.0485916137695312
 Step 980: loss 2.006260633468628
 Step 990: loss 1.9713857173919678
 Step 1000: loss 1.9568192958831787
 [Seed 35] Step 1000 --> val loss 2.137382984161377
 Step 1010: loss 1.9632062911987305
 Step 1020: loss 1.9052064418792725
 Step 1030: loss 1.8702939748764038
 Step 1040: loss 1.8920706510543823
 Step 1050: loss 1.896162509918213
 Step 1060: loss 1.9235481023788452
 Step 1070: loss 1.9649895429611206
 Step 1080: loss 2.033820867538452
 Step 1090: loss 1.9197444915771484
 Step 1100: loss 1.907631278038025
 Step 1110: loss 2.0115585327148438
 Step 1120: loss 2.009962558746338
 Step 1130: loss 1.9366090297698975
 Step 1140: loss 1.8553231954574585
 Step 1150: loss 1.9135884046554565
 Step 1160: loss 1.8668880462646484
 Step 1170: loss 1.894148826599121
 Step 1180: loss 1.8841791152954102
 Step 1190: loss 2.0162768363952637
 Step 1200: loss 1.7844294309616089
 [Seed 35] Step 1200 --> val loss 2.1578288078308105
 Step 1210: loss 1.9214181900024414
 Step 1220: loss 1.9633922576904297
 Step 1230: loss 1.9683682918548584
 Step 1240: loss 1.8166911602020264
 Step 1250: loss 1.8895492553710938
 Step 1260: loss 1.8990709781646729
 Step 1270: loss 1.9326860904693604
 Step 1280: loss 2.00520658493042
 Step 1290: loss 1.875756025314331
 Step 1300: loss 1.8404334783554077
 Step 1310: loss 1.8863112926483154
 Step 1320: loss 1.9360259771347046
 Step 1330: loss 1.9350517988204956
 Step 1340: loss 1.9016497135162354
 Step 1350: loss 1.9031262397766113
 Step 1360: loss 1.886537790298462
 Step 1370: loss 1.9289379119873047
 Step 1380: loss 1.911224126815796
 Step 1390: loss 1.8426671028137207
 Step 1400: loss 1.8744525909423828
 [Seed 35] Step 1400 --> val loss 2.0963456630706787
 Step 1410: loss 1.8259007930755615
 Step 1420: loss 1.7606180906295776
 Step 1430: loss 1.767310619354248
 Step 1440: loss 1.7963123321533203
 Step 1450: loss 2.0222630500793457
 Step 1460: loss 1.8619041442871094
 Step 1470: loss 1.8656749725341797
 Step 1480: loss 1.8846420049667358
 Step 1490: loss 1.7149910926818848
 Step 1500: loss 1.8059767484664917
 Step 1510: loss 1.8802711963653564
 Step 1520: loss 1.8130826950073242
 Step 1530: loss 1.7877616882324219
 Step 1540: loss 1.7929744720458984
 Step 1550: loss 1.677355170249939
 Step 1560: loss 1.8100087642669678
 Step 1570: loss 1.7534701824188232
 Step 1580: loss 1.8420352935791016
 Step 1590: loss 1.7042548656463623
 Step 1600: loss 1.8438804149627686
 [Seed 35] Step 1600 --> val loss 2.0319790840148926
 Step 1610: loss 1.8771789073944092
 Step 1620: loss 1.7447748184204102
 Step 1630: loss 1.7866485118865967
 Step 1640: loss 1.8294103145599365
 Step 1650: loss 1.771948218345642
 Step 1660: loss 1.931299090385437
 Step 1670: loss 1.8396615982055664
 Step 1680: loss 1.8504284620285034
 Step 1690: loss 1.7598727941513062
 Step 1700: loss 1.6335256099700928
 Step 1710: loss 1.7521326541900635
 Step 1720: loss 1.8195164203643799
 Step 1730: loss 1.7738621234893799
 Step 1740: loss 1.6564639806747437
 Step 1750: loss 1.8097742795944214
 Step 1760: loss 1.7619144916534424
 Step 1770: loss 1.6329258680343628
 Step 1780: loss 1.731483817100525
 Step 1790: loss 1.7959855794906616
 Step 1800: loss 1.7548584938049316
 [Seed 35] Step 1800 --> val loss 1.8874729871749878
 Step 1810: loss 1.7533868551254272
 Step 1820: loss 1.7448737621307373
 Step 1830: loss 1.6504745483398438
 Step 1840: loss 1.7528194189071655
 Step 1850: loss 1.7577073574066162
 Step 1860: loss 1.7887334823608398
 Step 1870: loss 1.7155059576034546
 Step 1880: loss 1.652329921722412
 Step 1890: loss 1.7451413869857788
 Step 1900: loss 1.8111042976379395
 Step 1910: loss 1.8572582006454468
 Step 1920: loss 1.727062702178955
 Step 1930: loss 1.7389525175094604
 Step 1940: loss 1.8029224872589111
 Step 1950: loss 1.7040050029754639
 Step 1960: loss 1.6361955404281616
 Step 1970: loss 1.7628700733184814
 Step 1980: loss 1.7326077222824097
 Step 1990: loss 1.8409087657928467
 Step 2000: loss 1.630143404006958
 [Seed 35] Step 2000 --> val loss 1.8269418478012085
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [35 21 31  1 36 21 10  0 14 59] ...
 First target sequence: [21 31  1 36 21 10  0 14 59 58] ...

 Decoded input: WIS XI:
 But, Warwick,
 Thou and Oxford, with five thousand men,
 S

 Decoded target: IS XI:
 But, Warwick,
 Thou and Oxford, with five thousand men,
 Sh
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.204963207244873
 [Seed 36] Step 0 --> val loss 4.186765193939209
 Step 10: loss 3.8673200607299805
 Step 20: loss 3.2987284660339355
 Step 30: loss 2.907601833343506
 Step 40: loss 2.7316367626190186
 Step 50: loss 2.5635268688201904
 Step 60: loss 2.4577748775482178
 Step 70: loss 2.413634777069092
 Step 80: loss 2.4709086418151855
 Step 90: loss 2.2586793899536133
 Step 100: loss 2.2803335189819336
 Step 110: loss 2.2563159465789795
 Step 120: loss 2.2313318252563477
 Step 130: loss 2.127084255218506
 Step 140: loss 2.2596499919891357
 Step 150: loss 2.2420966625213623
 Step 160: loss 2.0663113594055176
 Step 170: loss 2.0724568367004395
 Step 180: loss 2.112027645111084
 Step 190: loss 2.209118604660034
 Step 200: loss 2.0457053184509277
 [Seed 36] Step 200 --> val loss 2.3082053661346436
 Step 210: loss 1.9473750591278076
 Step 220: loss 2.063854694366455
 Step 230: loss 2.05618953704834
 Step 240: loss 2.036679744720459
 Step 250: loss 2.1895642280578613
 Step 260: loss 2.0171151161193848
 Step 270: loss 2.162898540496826
 Step 280: loss 2.052114486694336
 Step 290: loss 2.0975425243377686
 Step 300: loss 2.0641565322875977
 Step 310: loss 2.0621843338012695
 Step 320: loss 1.9523735046386719
 Step 330: loss 2.0939126014709473
 Step 340: loss 2.016839027404785
 Step 350: loss 2.1167449951171875
 Step 360: loss 1.9912700653076172
 Step 370: loss 2.026071071624756
 Step 380: loss 2.0048320293426514
 Step 390: loss 2.0064826011657715
 Step 400: loss 2.06797194480896
 [Seed 36] Step 400 --> val loss 2.4218218326568604
 Step 410: loss 2.0863780975341797
 Step 420: loss 1.9882644414901733
 Step 430: loss 1.98830246925354
 Step 440: loss 2.0053470134735107
 Step 450: loss 2.020979881286621
 Step 460: loss 1.9811921119689941
 Step 470: loss 2.077071189880371
 Step 480: loss 2.0649821758270264
 Step 490: loss 2.1394691467285156
 Step 500: loss 2.0536558628082275
 Step 510: loss 1.9402273893356323
 Step 520: loss 2.0106239318847656
 Step 530: loss 1.949586272239685
 Step 540: loss 2.0205273628234863
 Step 550: loss 2.0172643661499023
 Step 560: loss 2.0363240242004395
 Step 570: loss 2.153257369995117
 Step 580: loss 2.004876136779785
 Step 590: loss 2.0856237411499023
 Step 600: loss 2.0644166469573975
 [Seed 36] Step 600 --> val loss 2.2540841102600098
 Step 610: loss 2.0020978450775146
 Step 620: loss 2.1656227111816406
 Step 630: loss 1.9366114139556885
 Step 640: loss 2.05777907371521
 Step 650: loss 1.9818849563598633
 Step 660: loss 1.991579294204712
 Step 670: loss 1.938306212425232
 Step 680: loss 1.8663190603256226
 Step 690: loss 1.9031085968017578
 Step 700: loss 1.9418745040893555
 Step 710: loss 2.01241397857666
 Step 720: loss 2.0184497833251953
 Step 730: loss 2.0365993976593018
 Step 740: loss 2.0004937648773193
 Step 750: loss 1.9777143001556396
 Step 760: loss 2.019534111022949
 Step 770: loss 2.014392375946045
 Step 780: loss 2.0437569618225098
 Step 790: loss 1.8969616889953613
 Step 800: loss 1.9659909009933472
 [Seed 36] Step 800 --> val loss 2.1666553020477295
 Step 810: loss 1.960176944732666
 Step 820: loss 1.9557533264160156
 Step 830: loss 1.8593276739120483
 Step 840: loss 1.9235223531723022
 Step 850: loss 2.0263609886169434
 Step 860: loss 1.930355429649353
 Step 870: loss 1.8524160385131836
 Step 880: loss 2.0681700706481934
 Step 890: loss 1.9773731231689453
 Step 900: loss 1.9638326168060303
 Step 910: loss 1.872667670249939
 Step 920: loss 1.943030595779419
 Step 930: loss 1.9536241292953491
 Step 940: loss 1.9730169773101807
 Step 950: loss 1.9860152006149292
 Step 960: loss 2.082641124725342
 Step 970: loss 2.031541347503662
 Step 980: loss 1.9915704727172852
 Step 990: loss 1.9235401153564453
 Step 1000: loss 1.9246591329574585
 [Seed 36] Step 1000 --> val loss 2.2279868125915527
 Step 1010: loss 1.9120142459869385
 Step 1020: loss 1.9632338285446167
 Step 1030: loss 1.8337829113006592
 Step 1040: loss 1.878730058670044
 Step 1050: loss 1.9640816450119019
 Step 1060: loss 1.8152735233306885
 Step 1070: loss 1.8880770206451416
 Step 1080: loss 1.9829881191253662
 Step 1090: loss 1.8198916912078857
 Step 1100: loss 2.0160350799560547
 Step 1110: loss 1.889445185661316
 Step 1120: loss 2.0060691833496094
 Step 1130: loss 1.9655612707138062
 Step 1140: loss 2.0611982345581055
 Step 1150: loss 1.9377329349517822
 Step 1160: loss 1.9485173225402832
 Step 1170: loss 1.946332335472107
 Step 1180: loss 1.9108140468597412
 Step 1190: loss 1.9834330081939697
 Step 1200: loss nan
 [Seed 36] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 36] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 36] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 36] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 36] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [47 58 46 53 59 58  1 52 53 58] ...
 First target sequence: [58 46 53 59 58  1 52 53 58 43] ...

 Decoded input: ithout note, here's many else have done,--
 You shout me forth
 In

 Decoded target: thout note, here's many else have done,--
 You shout me forth
 In 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.205314636230469
 [Seed 37] Step 0 --> val loss 4.182346820831299
 Step 10: loss 3.8450021743774414
 Step 20: loss 3.299769878387451
 Step 30: loss 2.8713603019714355
 Step 40: loss 2.623690128326416
 Step 50: loss 2.636096477508545
 Step 60: loss 2.5021204948425293
 Step 70: loss 2.3728957176208496
 Step 80: loss 2.4089455604553223
 Step 90: loss 2.2074341773986816
 Step 100: loss 2.193382740020752
 Step 110: loss 2.205970048904419
 Step 120: loss 2.233898878097534
 Step 130: loss 2.2159335613250732
 Step 140: loss 2.2200710773468018
 Step 150: loss 2.1418681144714355
 Step 160: loss 2.2115955352783203
 Step 170: loss 2.220540761947632
 Step 180: loss 2.227849006652832
 Step 190: loss 2.064983367919922
 Step 200: loss 2.134476661682129
 [Seed 37] Step 200 --> val loss 2.361624240875244
 Step 210: loss 2.2035365104675293
 Step 220: loss 2.154491662979126
 Step 230: loss 2.0383501052856445
 Step 240: loss 2.186286687850952
 Step 250: loss 2.0623326301574707
 Step 260: loss 1.9441980123519897
 Step 270: loss 1.9645259380340576
 Step 280: loss 2.1588592529296875
 Step 290: loss 2.07076358795166
 Step 300: loss 2.1821794509887695
 Step 310: loss 2.0695862770080566
 Step 320: loss 1.9828221797943115
 Step 330: loss 2.152895212173462
 Step 340: loss 2.0366597175598145
 Step 350: loss 2.028379440307617
 Step 360: loss 2.050736427307129
 Step 370: loss 1.9869645833969116
 Step 380: loss 2.126276731491089
 Step 390: loss 1.9754059314727783
 Step 400: loss 2.169301986694336
 [Seed 37] Step 400 --> val loss 2.334056854248047
 Step 410: loss 1.9948322772979736
 Step 420: loss 2.1538596153259277
 Step 430: loss 2.0448203086853027
 Step 440: loss 2.040346145629883
 Step 450: loss 2.0148754119873047
 Step 460: loss 1.9416269063949585
 Step 470: loss 2.105583667755127
 Step 480: loss 2.108160972595215
 Step 490: loss 2.0432000160217285
 Step 500: loss 2.0768704414367676
 Step 510: loss 2.0110878944396973
 Step 520: loss 1.9946937561035156
 Step 530: loss 2.0550103187561035
 Step 540: loss 1.9690911769866943
 Step 550: loss 1.903674602508545
 Step 560: loss 2.027143955230713
 Step 570: loss 1.9617538452148438
 Step 580: loss 1.9553725719451904
 Step 590: loss 2.005333662033081
 Step 600: loss 2.0054588317871094
 [Seed 37] Step 600 --> val loss 2.3271262645721436
 Step 610: loss 1.9986670017242432
 Step 620: loss 1.950315237045288
 Step 630: loss 1.8866596221923828
 Step 640: loss 1.9775381088256836
 Step 650: loss 2.0093913078308105
 Step 660: loss 1.9969353675842285
 Step 670: loss 1.9526265859603882
 Step 680: loss 1.90451979637146
 Step 690: loss 2.0477681159973145
 Step 700: loss 1.996260166168213
 Step 710: loss 1.8734617233276367
 Step 720: loss 1.9586519002914429
 Step 730: loss 1.9656352996826172
 Step 740: loss 1.9715646505355835
 Step 750: loss 2.0461905002593994
 Step 760: loss 1.9927833080291748
 Step 770: loss 2.100788116455078
 Step 780: loss 1.7715991735458374
 Step 790: loss 1.9710869789123535
 Step 800: loss 1.9799020290374756
 [Seed 37] Step 800 --> val loss 2.2123076915740967
 Step 810: loss 1.9341933727264404
 Step 820: loss 1.9342758655548096
 Step 830: loss 2.141507863998413
 Step 840: loss 1.9480313062667847
 Step 850: loss 1.9503387212753296
 Step 860: loss 1.9571131467819214
 Step 870: loss 2.0088257789611816
 Step 880: loss 1.8803987503051758
 Step 890: loss 1.8840751647949219
 Step 900: loss 1.9271678924560547
 Step 910: loss 1.8553965091705322
 Step 920: loss 1.899087905883789
 Step 930: loss 1.8862180709838867
 Step 940: loss 1.973691701889038
 Step 950: loss 2.1119048595428467
 Step 960: loss 2.0323069095611572
 Step 970: loss 1.9134318828582764
 Step 980: loss 1.8975112438201904
 Step 990: loss 1.8968133926391602
 Step 1000: loss 2.0004053115844727
 [Seed 37] Step 1000 --> val loss 2.1771624088287354
 Step 1010: loss 1.9464788436889648
 Step 1020: loss 1.7974294424057007
 Step 1030: loss 1.8174980878829956
 Step 1040: loss 1.9474310874938965
 Step 1050: loss 1.943825602531433
 Step 1060: loss 1.9420493841171265
 Step 1070: loss 1.9089138507843018
 Step 1080: loss 1.917433738708496
 Step 1090: loss 1.8766759634017944
 Step 1100: loss 1.8847945928573608
 Step 1110: loss 1.9950687885284424
 Step 1120: loss 1.8382070064544678
 Step 1130: loss 1.959972620010376
 Step 1140: loss 2.0049006938934326
 Step 1150: loss 1.8909146785736084
 Step 1160: loss 1.8537191152572632
 Step 1170: loss 1.8253593444824219
 Step 1180: loss 1.963672399520874
 Step 1190: loss 1.9401830434799194
 Step 1200: loss 1.836747407913208
 [Seed 37] Step 1200 --> val loss 2.135309934616089
 Step 1210: loss 1.9579007625579834
 Step 1220: loss 1.918733835220337
 Step 1230: loss 1.8252156972885132
 Step 1240: loss 1.9540307521820068
 Step 1250: loss 1.9436798095703125
 Step 1260: loss 1.986755132675171
 Step 1270: loss 1.9335782527923584
 Step 1280: loss 1.9292943477630615
 Step 1290: loss 1.8356099128723145
 Step 1300: loss 1.896981954574585
 Step 1310: loss 1.9527089595794678
 Step 1320: loss 1.938136100769043
 Step 1330: loss 1.7722028493881226
 Step 1340: loss 1.8790442943572998
 Step 1350: loss 1.9194307327270508
 Step 1360: loss 1.837484359741211
 Step 1370: loss 1.838293433189392
 Step 1380: loss 1.8565537929534912
 Step 1390: loss 1.9755284786224365
 Step 1400: loss 1.9349321126937866
 [Seed 37] Step 1400 --> val loss 2.048255205154419
 Step 1410: loss 1.7867854833602905
 Step 1420: loss 1.8676637411117554
 Step 1430: loss 1.82876718044281
 Step 1440: loss 1.7590653896331787
 Step 1450: loss 1.778993010520935
 Step 1460: loss 1.7751520872116089
 Step 1470: loss 1.955068826675415
 Step 1480: loss 1.8165614604949951
 Step 1490: loss 1.808441162109375
 Step 1500: loss 1.8238017559051514
 Step 1510: loss 1.83665931224823
 Step 1520: loss 1.8707603216171265
 Step 1530: loss 1.754645586013794
 Step 1540: loss 1.8399102687835693
 Step 1550: loss 1.8407459259033203
 Step 1560: loss 1.8432058095932007
 Step 1570: loss 1.8209600448608398
 Step 1580: loss 1.7897083759307861
 Step 1590: loss 1.762208342552185
 Step 1600: loss 1.7840633392333984
 [Seed 37] Step 1600 --> val loss 2.050867795944214
 Step 1610: loss 1.709438681602478
 Step 1620: loss 1.8566029071807861
 Step 1630: loss 1.7283802032470703
 Step 1640: loss 1.908789038658142
 Step 1650: loss 1.75190007686615
 Step 1660: loss 1.7909824848175049
 Step 1670: loss 1.7785495519638062
 Step 1680: loss 1.7828893661499023
 Step 1690: loss 1.7373874187469482
 Step 1700: loss 1.7832460403442383
 Step 1710: loss 1.7656505107879639
 Step 1720: loss 1.7513163089752197
 Step 1730: loss 1.7910741567611694
 Step 1740: loss 1.751265287399292
 Step 1750: loss 1.7115423679351807
 Step 1760: loss 1.6485586166381836
 Step 1770: loss 1.833601713180542
 Step 1780: loss 1.6328492164611816
 Step 1790: loss 1.7054132223129272
 Step 1800: loss 1.6812604665756226
 [Seed 37] Step 1800 --> val loss 1.9559719562530518
 Step 1810: loss 1.6490778923034668
 Step 1820: loss 1.7107648849487305
 Step 1830: loss 1.6816567182540894
 Step 1840: loss 1.770329475402832
 Step 1850: loss 1.6170201301574707
 Step 1860: loss 1.7086713314056396
 Step 1870: loss 1.6797068119049072
 Step 1880: loss 1.7050211429595947
 Step 1890: loss 1.6453490257263184
 Step 1900: loss 1.7207305431365967
 Step 1910: loss 1.695660948753357
 Step 1920: loss 1.8237279653549194
 Step 1930: loss 1.7219561338424683
 Step 1940: loss 1.651543378829956
 Step 1950: loss 1.743043303489685
 Step 1960: loss 1.6508100032806396
 Step 1970: loss 1.7652482986450195
 Step 1980: loss 1.6069307327270508
 Step 1990: loss 1.5882329940795898
 Step 2000: loss 1.6543712615966797
 [Seed 37] Step 2000 --> val loss 1.8968876600265503
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [47 43  1 58 46 43 43  1 58 53] ...
 First target sequence: [43  1 58 46 43 43  1 58 53  1] ...

 Decoded input: ie thee to hell for shame, and leave the world,
 Thou cacodemon! 

 Decoded target: e thee to hell for shame, and leave the world,
 Thou cacodemon! t
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.214001178741455
 [Seed 38] Step 0 --> val loss 4.186831474304199
 Step 10: loss 3.874337673187256
 Step 20: loss 3.283510684967041
 Step 30: loss 2.828366756439209
 Step 40: loss 2.632326602935791
 Step 50: loss 2.540799856185913
 Step 60: loss 2.562077522277832
 Step 70: loss 2.359389066696167
 Step 80: loss 2.2639236450195312
 Step 90: loss 2.2995803356170654
 Step 100: loss 2.3477020263671875
 Step 110: loss 2.2771403789520264
 Step 120: loss 2.2167181968688965
 Step 130: loss 2.1890554428100586
 Step 140: loss 2.2709197998046875
 Step 150: loss 2.2076804637908936
 Step 160: loss 2.128499746322632
 Step 170: loss 1.9965267181396484
 Step 180: loss 2.107393264770508
 Step 190: loss 2.1424434185028076
 Step 200: loss 2.147223711013794
 [Seed 38] Step 200 --> val loss 2.3377087116241455
 Step 210: loss 2.1434450149536133
 Step 220: loss 2.1426455974578857
 Step 230: loss 2.2707884311676025
 Step 240: loss 2.069762706756592
 Step 250: loss 2.051135301589966
 Step 260: loss 2.1478021144866943
 Step 270: loss 2.040512800216675
 Step 280: loss 2.009219169616699
 Step 290: loss 2.0575952529907227
 Step 300: loss 2.0386648178100586
 Step 310: loss 2.032029151916504
 Step 320: loss 1.9156930446624756
 Step 330: loss 2.096222400665283
 Step 340: loss 1.8532147407531738
 Step 350: loss 2.057713508605957
 Step 360: loss 1.963519811630249
 Step 370: loss 2.016860008239746
 Step 380: loss 1.8211429119110107
 Step 390: loss 2.0906283855438232
 Step 400: loss 2.027251720428467
 [Seed 38] Step 400 --> val loss 2.3037614822387695
 Step 410: loss 2.0835399627685547
 Step 420: loss 2.1263766288757324
 Step 430: loss 1.8781700134277344
 Step 440: loss 2.0545003414154053
 Step 450: loss 1.985080599784851
 Step 460: loss 2.053778648376465
 Step 470: loss 2.022656202316284
 Step 480: loss 2.013538360595703
 Step 490: loss 1.9968494176864624
 Step 500: loss 1.989154577255249
 Step 510: loss 2.0611836910247803
 Step 520: loss 2.088843822479248
 Step 530: loss 1.996941328048706
 Step 540: loss 2.0404086112976074
 Step 550: loss 2.0396475791931152
 Step 560: loss 2.0743374824523926
 Step 570: loss 2.0310559272766113
 Step 580: loss 1.9608219861984253
 Step 590: loss 1.9800620079040527
 Step 600: loss 2.018343448638916
 [Seed 38] Step 600 --> val loss 2.3123350143432617
 Step 610: loss 1.9986350536346436
 Step 620: loss 1.974034309387207
 Step 630: loss 1.968158483505249
 Step 640: loss 2.064337730407715
 Step 650: loss 2.0594475269317627
 Step 660: loss 1.957127571105957
 Step 670: loss 1.9421018362045288
 Step 680: loss 2.0255112648010254
 Step 690: loss 1.8627431392669678
 Step 700: loss 1.9881752729415894
 Step 710: loss 1.9676053524017334
 Step 720: loss 1.8857457637786865
 Step 730: loss 2.033114433288574
 Step 740: loss 2.0180959701538086
 Step 750: loss 1.9485172033309937
 Step 760: loss 1.9286210536956787
 Step 770: loss 2.0752224922180176
 Step 780: loss 1.9436429738998413
 Step 790: loss 1.9859012365341187
 Step 800: loss 1.9897830486297607
 [Seed 38] Step 800 --> val loss 2.195774793624878
 Step 810: loss 1.9820393323898315
 Step 820: loss 2.0391178131103516
 Step 830: loss 1.9255225658416748
 Step 840: loss 1.8469586372375488
 Step 850: loss 1.9881123304367065
 Step 860: loss 1.8572362661361694
 Step 870: loss 1.9834861755371094
 Step 880: loss 1.8726900815963745
 Step 890: loss 2.0272538661956787
 Step 900: loss 1.9183738231658936
 Step 910: loss 2.0648488998413086
 Step 920: loss 1.819075345993042
 Step 930: loss 1.9313769340515137
 Step 940: loss 2.001579999923706
 Step 950: loss 1.9089131355285645
 Step 960: loss 2.0714478492736816
 Step 970: loss 1.9871807098388672
 Step 980: loss 1.9911117553710938
 Step 990: loss 1.8904268741607666
 Step 1000: loss 1.890181303024292
 [Seed 38] Step 1000 --> val loss 2.097627639770508
 Step 1010: loss 1.9143638610839844
 Step 1020: loss 1.9584193229675293
 Step 1030: loss 1.9705619812011719
 Step 1040: loss 1.8528127670288086
 Step 1050: loss 1.9526081085205078
 Step 1060: loss 2.044577121734619
 Step 1070: loss 1.883723497390747
 Step 1080: loss 1.878486156463623
 Step 1090: loss 1.8738765716552734
 Step 1100: loss 1.9040111303329468
 Step 1110: loss 1.9833943843841553
 Step 1120: loss 1.9353704452514648
 Step 1130: loss 1.970621109008789
 Step 1140: loss 1.7801949977874756
 Step 1150: loss 2.0519886016845703
 Step 1160: loss 1.8612315654754639
 Step 1170: loss 2.033757209777832
 Step 1180: loss 1.918898582458496
 Step 1190: loss 2.029775619506836
 Step 1200: loss 1.9861292839050293
 [Seed 38] Step 1200 --> val loss 2.077009916305542
 Step 1210: loss 1.9869284629821777
 Step 1220: loss 1.8598227500915527
 Step 1230: loss 1.6938352584838867
 Step 1240: loss 1.8711585998535156
 Step 1250: loss 1.8234399557113647
 Step 1260: loss 1.9620202779769897
 Step 1270: loss 1.7809404134750366
 Step 1280: loss 1.850314974784851
 Step 1290: loss 1.8409748077392578
 Step 1300: loss 1.9409291744232178
 Step 1310: loss 1.9759553670883179
 Step 1320: loss 1.964996099472046
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 38] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 38] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 38] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 38] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [56 42 57  1 46 43 56 43  1 39] ...
 First target sequence: [42 57  1 46 43 56 43  1 39 41] ...

 Decoded input: rds here accuse my zeal:
 'Tis not the trial of a woman's war,
 Th

 Decoded target: ds here accuse my zeal:
 'Tis not the trial of a woman's war,
 The
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.195012092590332
 [Seed 39] Step 0 --> val loss 4.185646057128906
 Step 10: loss 3.8462131023406982
 Step 20: loss 3.338012218475342
 Step 30: loss 2.805021286010742
 Step 40: loss 2.5944442749023438
 Step 50: loss 2.532290458679199
 Step 60: loss 2.5390543937683105
 Step 70: loss 2.317596197128296
 Step 80: loss 2.4432008266448975
 Step 90: loss 2.2623140811920166
 Step 100: loss 2.3724799156188965
 Step 110: loss 2.0985939502716064
 Step 120: loss 2.196463108062744
 Step 130: loss 2.2248055934906006
 Step 140: loss 2.191911220550537
 Step 150: loss 2.2480759620666504
 Step 160: loss 2.1888866424560547
 Step 170: loss 2.081291675567627
 Step 180: loss 2.1363840103149414
 Step 190: loss 2.145817279815674
 Step 200: loss 2.1242032051086426
 [Seed 39] Step 200 --> val loss 2.3406054973602295
 Step 210: loss 2.1645121574401855
 Step 220: loss 2.072186231613159
 Step 230: loss 1.9973459243774414
 Step 240: loss 2.0829591751098633
 Step 250: loss 1.994585394859314
 Step 260: loss 2.1123037338256836
 Step 270: loss 2.014224052429199
 Step 280: loss 2.0536162853240967
 Step 290: loss 2.0266575813293457
 Step 300: loss 1.9661859273910522
 Step 310: loss 2.0955638885498047
 Step 320: loss 2.0384347438812256
 Step 330: loss 2.0851399898529053
 Step 340: loss 2.0281858444213867
 Step 350: loss 2.1052377223968506
 Step 360: loss 2.0733542442321777
 Step 370: loss 1.9999573230743408
 Step 380: loss 2.083488941192627
 Step 390: loss 2.060652017593384
 Step 400: loss 2.02777099609375
 [Seed 39] Step 400 --> val loss 2.253967046737671
 Step 410: loss 1.977026343345642
 Step 420: loss 1.8817580938339233
 Step 430: loss 2.0463199615478516
 Step 440: loss 2.0312209129333496
 Step 450: loss 2.007706642150879
 Step 460: loss 1.9523465633392334
 Step 470: loss 1.9450643062591553
 Step 480: loss 2.0906081199645996
 Step 490: loss 1.9234765768051147
 Step 500: loss 1.9216563701629639
 Step 510: loss 1.884507179260254
 Step 520: loss 2.011962413787842
 Step 530: loss 2.017716884613037
 Step 540: loss 2.0028252601623535
 Step 550: loss 2.0017313957214355
 Step 560: loss 2.024484157562256
 Step 570: loss 2.0230042934417725
 Step 580: loss 2.1337525844573975
 Step 590: loss 1.9909127950668335
 Step 600: loss 2.0243802070617676
 [Seed 39] Step 600 --> val loss 2.264768362045288
 Step 610: loss 2.019836187362671
 Step 620: loss 1.9711570739746094
 Step 630: loss 1.9925134181976318
 Step 640: loss 1.8922580480575562
 Step 650: loss 1.915809988975525
 Step 660: loss 2.056302309036255
 Step 670: loss 1.9625575542449951
 Step 680: loss 2.0130462646484375
 Step 690: loss 2.082880973815918
 Step 700: loss 2.033392906188965
 Step 710: loss 2.0100934505462646
 Step 720: loss 1.9666751623153687
 Step 730: loss 2.040433645248413
 Step 740: loss 1.899837851524353
 Step 750: loss 1.9725335836410522
 Step 760: loss 1.887769103050232
 Step 770: loss 1.836885690689087
 Step 780: loss 1.9977481365203857
 Step 790: loss 1.9373393058776855
 Step 800: loss 2.023076295852661
 [Seed 39] Step 800 --> val loss 2.2137997150421143
 Step 810: loss 2.070364475250244
 Step 820: loss 1.937407374382019
 Step 830: loss 1.8424413204193115
 Step 840: loss 2.0137882232666016
 Step 850: loss 1.8695898056030273
 Step 860: loss 1.9427522420883179
 Step 870: loss 1.958620548248291
 Step 880: loss 1.9231245517730713
 Step 890: loss 1.9667538404464722
 Step 900: loss 1.8867319822311401
 Step 910: loss 1.9303834438323975
 Step 920: loss 1.9417293071746826
 Step 930: loss 1.790022850036621
 Step 940: loss 2.0560531616210938
 Step 950: loss 1.9301190376281738
 Step 960: loss 1.970546007156372
 Step 970: loss 1.9594649076461792
 Step 980: loss 1.8801339864730835
 Step 990: loss 1.941349744796753
 Step 1000: loss 2.034923553466797
 [Seed 39] Step 1000 --> val loss 2.2699999809265137
 Step 1010: loss 1.9300143718719482
 Step 1020: loss 1.9350330829620361
 Step 1030: loss 1.8713958263397217
 Step 1040: loss 1.897202968597412
 Step 1050: loss 1.9037078619003296
 Step 1060: loss 1.8225135803222656
 Step 1070: loss 1.9206079244613647
 Step 1080: loss 1.932844877243042
 Step 1090: loss 2.060133218765259
 Step 1100: loss 1.9544267654418945
 Step 1110: loss 1.8920176029205322
 Step 1120: loss 1.8469886779785156
 Step 1130: loss 1.8673099279403687
 Step 1140: loss 1.7971051931381226
 Step 1150: loss 1.7879724502563477
 Step 1160: loss 1.7594587802886963
 Step 1170: loss 1.9059922695159912
 Step 1180: loss 1.7365493774414062
 Step 1190: loss 1.7744383811950684
 Step 1200: loss 1.8528759479522705
 [Seed 39] Step 1200 --> val loss 2.091813564300537
 Step 1210: loss 1.810619831085205
 Step 1220: loss 1.7910627126693726
 Step 1230: loss 1.7849597930908203
 Step 1240: loss 1.8382353782653809
 Step 1250: loss 1.878157377243042
 Step 1260: loss 1.7856667041778564
 Step 1270: loss 1.792273998260498
 Step 1280: loss 1.8201794624328613
 Step 1290: loss 1.777848720550537
 Step 1300: loss 1.9257986545562744
 Step 1310: loss 1.7820978164672852
 Step 1320: loss 1.7261693477630615
 Step 1330: loss 1.8437427282333374
 Step 1340: loss 1.7250961065292358
 Step 1350: loss 1.739572286605835
 Step 1360: loss 1.687598466873169
 Step 1370: loss 1.7268009185791016
 Step 1380: loss 1.7268427610397339
 Step 1390: loss 1.973134994506836
 Step 1400: loss 1.6741217374801636
 [Seed 39] Step 1400 --> val loss 1.9797674417495728
 Step 1410: loss 1.731508731842041
 Step 1420: loss 1.8613471984863281
 Step 1430: loss 1.8674143552780151
 Step 1440: loss 1.8252347707748413
 Step 1450: loss 1.7531826496124268
 Step 1460: loss 1.8448255062103271
 Step 1470: loss 1.779123306274414
 Step 1480: loss 1.758890986442566
 Step 1490: loss 1.8328580856323242
 Step 1500: loss 1.8098253011703491
 Step 1510: loss 1.8225531578063965
 Step 1520: loss 1.8162474632263184
 Step 1530: loss 1.79820716381073
 Step 1540: loss 1.8234657049179077
 Step 1550: loss 1.895554542541504
 Step 1560: loss 1.7260363101959229
 Step 1570: loss 1.840258002281189
 Step 1580: loss 1.7928171157836914
 Step 1590: loss 1.781245231628418
 Step 1600: loss 1.7012689113616943
 [Seed 39] Step 1600 --> val loss 2.004221200942993
 Step 1610: loss 1.6616542339324951
 Step 1620: loss 1.7123539447784424
 Step 1630: loss 1.6838716268539429
 Step 1640: loss 1.6508145332336426
 Step 1650: loss 1.7989940643310547
 Step 1660: loss 1.8246406316757202
 Step 1670: loss 1.785886526107788
 Step 1680: loss 1.7261779308319092
 Step 1690: loss 1.6737284660339355
 Step 1700: loss 1.6614316701889038
 Step 1710: loss 1.60531747341156
 Step 1720: loss 1.6930421590805054
 Step 1730: loss 1.7708871364593506
 Step 1740: loss 1.678121566772461
 Step 1750: loss 1.6543140411376953
 Step 1760: loss 1.653564453125
 Step 1770: loss 1.772732138633728
 Step 1780: loss 1.5824440717697144
 Step 1790: loss 1.7086608409881592
 Step 1800: loss 1.7067923545837402
 [Seed 39] Step 1800 --> val loss 1.867618203163147
 Step 1810: loss 1.7112505435943604
 Step 1820: loss 1.667930006980896
 Step 1830: loss 1.7932204008102417
 Step 1840: loss 1.6042957305908203
 Step 1850: loss 1.6658916473388672
 Step 1860: loss 1.6853355169296265
 Step 1870: loss 1.6907014846801758
 Step 1880: loss 1.603124976158142
 Step 1890: loss 1.704846739768982
 Step 1900: loss 1.7983192205429077
 Step 1910: loss 1.8603847026824951
 Step 1920: loss 1.7069783210754395
 Step 1930: loss 1.6548480987548828
 Step 1940: loss 1.5707297325134277
 Step 1950: loss 1.6827361583709717
 Step 1960: loss 1.6859822273254395
 Step 1970: loss 1.668903112411499
 Step 1980: loss 1.6114556789398193
 Step 1990: loss 1.6938433647155762
 Step 2000: loss 1.7062525749206543
 [Seed 39] Step 2000 --> val loss 1.798439860343933
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [56 43 61 39 56 42 10  1 40 43] ...
 First target sequence: [43 61 39 56 42 10  1 40 43  1] ...

 Decoded input: reward: be gone.

 KING LEWIS XI:
 But, Warwick,
 Thou and Oxford, 

 Decoded target: eward: be gone.

 KING LEWIS XI:
 But, Warwick,
 Thou and Oxford, w
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.206743240356445
 [Seed 40] Step 0 --> val loss 4.186513900756836
 Step 10: loss 3.8225784301757812
 Step 20: loss 3.3829874992370605
 Step 30: loss 2.899400234222412
 Step 40: loss 2.713648796081543
 Step 50: loss 2.7012641429901123
 Step 60: loss 2.5848662853240967
 Step 70: loss 2.420698404312134
 Step 80: loss 2.306962490081787
 Step 90: loss 2.378787040710449
 Step 100: loss 2.378633499145508
 Step 110: loss 2.2275450229644775
 Step 120: loss 2.2238073348999023
 Step 130: loss 2.2330422401428223
 Step 140: loss 2.2099533081054688
 Step 150: loss 2.1810827255249023
 Step 160: loss 2.1048974990844727
 Step 170: loss 2.1031954288482666
 Step 180: loss 2.1181538105010986
 Step 190: loss 2.151257276535034
 Step 200: loss 2.133267879486084
 [Seed 40] Step 200 --> val loss 2.2679903507232666
 Step 210: loss 2.0816986560821533
 Step 220: loss 2.197951078414917
 Step 230: loss 2.118912696838379
 Step 240: loss 2.1027817726135254
 Step 250: loss 2.030893325805664
 Step 260: loss 2.0680644512176514
 Step 270: loss 2.037489891052246
 Step 280: loss 2.173046112060547
 Step 290: loss 2.069977283477783
 Step 300: loss 2.1251955032348633
 Step 310: loss 1.9479124546051025
 Step 320: loss 2.0166172981262207
 Step 330: loss 2.143616199493408
 Step 340: loss 1.9105050563812256
 Step 350: loss 2.0329482555389404
 Step 360: loss 2.107023000717163
 Step 370: loss 2.0832228660583496
 Step 380: loss 2.0597734451293945
 Step 390: loss 1.9544248580932617
 Step 400: loss 2.034217357635498
 [Seed 40] Step 400 --> val loss 2.352147340774536
 Step 410: loss 2.0748181343078613
 Step 420: loss 1.9086812734603882
 Step 430: loss 1.9498319625854492
 Step 440: loss 2.136320114135742
 Step 450: loss 1.957685947418213
 Step 460: loss 1.9941500425338745
 Step 470: loss 1.9404394626617432
 Step 480: loss 2.1241397857666016
 Step 490: loss 2.002711296081543
 Step 500: loss 2.0879340171813965
 Step 510: loss 2.0142412185668945
 Step 520: loss 1.9793428182601929
 Step 530: loss 1.9829156398773193
 Step 540: loss 1.9574973583221436
 Step 550: loss 1.993729591369629
 Step 560: loss 2.025587797164917
 Step 570: loss 1.9741079807281494
 Step 580: loss 1.948229193687439
 Step 590: loss 1.9277305603027344
 Step 600: loss 2.104407787322998
 [Seed 40] Step 600 --> val loss 2.205512762069702
 Step 610: loss 2.0004539489746094
 Step 620: loss 1.9881881475448608
 Step 630: loss 1.9137487411499023
 Step 640: loss 1.931522250175476
 Step 650: loss 2.0118799209594727
 Step 660: loss 2.0824365615844727
 Step 670: loss 2.0138702392578125
 Step 680: loss 2.063969135284424
 Step 690: loss 1.8849725723266602
 Step 700: loss 1.9761489629745483
 Step 710: loss 2.0462634563446045
 Step 720: loss 1.9813202619552612
 Step 730: loss 2.123389720916748
 Step 740: loss 1.9962457418441772
 Step 750: loss 1.9827508926391602
 Step 760: loss 2.0475244522094727
 Step 770: loss 1.9861879348754883
 Step 780: loss 1.9488475322723389
 Step 790: loss 2.010704278945923
 Step 800: loss 1.833980917930603
 [Seed 40] Step 800 --> val loss 2.319202184677124
 Step 810: loss 2.0571086406707764
 Step 820: loss 1.933293342590332
 Step 830: loss 2.0038061141967773
 Step 840: loss 1.8094881772994995
 Step 850: loss 1.9334945678710938
 Step 860: loss 1.980590581893921
 Step 870: loss 2.040942668914795
 Step 880: loss 2.015505313873291
 Step 890: loss 2.0053579807281494
 Step 900: loss 1.935703992843628
 Step 910: loss 2.000457763671875
 Step 920: loss 1.8993371725082397
 Step 930: loss 1.9698588848114014
 Step 940: loss 1.9676889181137085
 Step 950: loss 1.9992237091064453
 Step 960: loss 2.1402273178100586
 Step 970: loss 2.1396842002868652
 Step 980: loss 1.9000802040100098
 Step 990: loss 2.076329469680786
 Step 1000: loss 1.9822618961334229
 [Seed 40] Step 1000 --> val loss 2.246110677719116
 Step 1010: loss 2.038516044616699
 Step 1020: loss 1.9468162059783936
 Step 1030: loss 2.056180477142334
 Step 1040: loss 1.9411323070526123
 Step 1050: loss 2.074188709259033
 Step 1060: loss 1.984833002090454
 Step 1070: loss 1.8966789245605469
 Step 1080: loss 1.8920215368270874
 Step 1090: loss 1.8908920288085938
 Step 1100: loss 2.0768818855285645
 Step 1110: loss 1.9329814910888672
 Step 1120: loss 1.8809823989868164
 Step 1130: loss 2.0088577270507812
 Step 1140: loss 1.952594518661499
 Step 1150: loss 1.9151734113693237
 Step 1160: loss 2.022029161453247
 Step 1170: loss 1.9142086505889893
 Step 1180: loss 1.844923973083496
 Step 1190: loss 1.916886806488037
 Step 1200: loss 1.841306209564209
 [Seed 40] Step 1200 --> val loss 2.14652156829834
 Step 1210: loss 1.8708747625350952
 Step 1220: loss 2.0453028678894043
 Step 1230: loss 1.9864025115966797
 Step 1240: loss 1.9346027374267578
 Step 1250: loss 1.8158591985702515
 Step 1260: loss 1.8840651512145996
 Step 1270: loss 2.008507251739502
 Step 1280: loss 2.0240249633789062
 Step 1290: loss 2.0216565132141113
 Step 1300: loss 2.0148143768310547
 Step 1310: loss 1.9994724988937378
 Step 1320: loss 1.972227931022644
 Step 1330: loss 2.08443546295166
 Step 1340: loss 1.9533970355987549
 Step 1350: loss 1.8776037693023682
 Step 1360: loss 2.041106939315796
 Step 1370: loss 1.9101217985153198
 Step 1380: loss 1.821216106414795
 Step 1390: loss 1.8466157913208008
 Step 1400: loss 1.895359992980957
 [Seed 40] Step 1400 --> val loss 2.1960246562957764
 Step 1410: loss 2.012429714202881
 Step 1420: loss 1.8758296966552734
 Step 1430: loss 1.8587932586669922
 Step 1440: loss 1.8485667705535889
 Step 1450: loss 1.8719934225082397
 Step 1460: loss 1.8279794454574585
 Step 1470: loss 1.886012077331543
 Step 1480: loss 1.8925530910491943
 Step 1490: loss 1.731333613395691
 Step 1500: loss 1.9076154232025146
 Step 1510: loss 1.852025032043457
 Step 1520: loss 1.8871870040893555
 Step 1530: loss 1.9353322982788086
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 40] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 40] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 40] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 5 57  1 61 53 51 40  1 21  1] ...
 First target sequence: [57  1 61 53 51 40  1 21  1 40] ...

 Decoded input: 's womb I bury them:
 Where in that nest of spicery they shall br

 Decoded target: s womb I bury them:
 Where in that nest of spicery they shall bre
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.216192245483398
 [Seed 41] Step 0 --> val loss 4.187090873718262
 Step 10: loss 3.8366332054138184
 Step 20: loss 3.3294453620910645
 Step 30: loss 2.8536651134490967
 Step 40: loss 2.737334728240967
 Step 50: loss 2.5379204750061035
 Step 60: loss 2.4918839931488037
 Step 70: loss 2.3818986415863037
 Step 80: loss 2.4432430267333984
 Step 90: loss 2.33290958404541
 Step 100: loss 2.27559757232666
 Step 110: loss 2.2452917098999023
 Step 120: loss 2.1147775650024414
 Step 130: loss 2.1973471641540527
 Step 140: loss 2.2140159606933594
 Step 150: loss 2.1225194931030273
 Step 160: loss 2.218527317047119
 Step 170: loss 2.0946426391601562
 Step 180: loss 2.091954231262207
 Step 190: loss 2.1444201469421387
 Step 200: loss 2.119999408721924
 [Seed 41] Step 200 --> val loss 2.389002799987793
 Step 210: loss 2.0627145767211914
 Step 220: loss 2.144155502319336
 Step 230: loss 2.19741153717041
 Step 240: loss 1.9849178791046143
 Step 250: loss 2.1180341243743896
 Step 260: loss 2.161771297454834
 Step 270: loss 2.118574619293213
 Step 280: loss 2.048889636993408
 Step 290: loss 2.0667166709899902
 Step 300: loss 2.052614212036133
 Step 310: loss 2.0345234870910645
 Step 320: loss 2.008455276489258
 Step 330: loss 2.012047529220581
 Step 340: loss 2.0263681411743164
 Step 350: loss 1.9839417934417725
 Step 360: loss 1.9673923254013062
 Step 370: loss 1.9334663152694702
 Step 380: loss 2.06815242767334
 Step 390: loss 2.004436492919922
 Step 400: loss 2.056030750274658
 [Seed 41] Step 400 --> val loss 2.223027229309082
 Step 410: loss 1.9947532415390015
 Step 420: loss 2.146423101425171
 Step 430: loss 2.0258121490478516
 Step 440: loss 1.976215124130249
 Step 450: loss 2.019411087036133
 Step 460: loss 2.0944528579711914
 Step 470: loss 1.9991806745529175
 Step 480: loss 2.0023159980773926
 Step 490: loss 2.0293593406677246
 Step 500: loss 1.9535012245178223
 Step 510: loss 2.0193819999694824
 Step 520: loss 1.9523036479949951
 Step 530: loss 2.074523448944092
 Step 540: loss 2.0615127086639404
 Step 550: loss 2.0821304321289062
 Step 560: loss 1.974308729171753
 Step 570: loss 1.9372437000274658
 Step 580: loss 2.0274295806884766
 Step 590: loss 1.918137550354004
 Step 600: loss 1.9825671911239624
 [Seed 41] Step 600 --> val loss 2.163586378097534
 Step 610: loss 2.1580960750579834
 Step 620: loss 2.0040879249572754
 Step 630: loss 1.9506752490997314
 Step 640: loss 2.000253915786743
 Step 650: loss 1.9725507497787476
 Step 660: loss 2.0617918968200684
 Step 670: loss 1.9494514465332031
 Step 680: loss 2.011683225631714
 Step 690: loss 2.0443015098571777
 Step 700: loss 2.090188980102539
 Step 710: loss 2.014090061187744
 Step 720: loss 2.0093398094177246
 Step 730: loss 1.9579360485076904
 Step 740: loss 2.046182632446289
 Step 750: loss 2.067906379699707
 Step 760: loss 2.0087080001831055
 Step 770: loss 1.9217323064804077
 Step 780: loss 2.0469279289245605
 Step 790: loss 1.9870185852050781
 Step 800: loss 2.093621253967285
 [Seed 41] Step 800 --> val loss 2.2355167865753174
 Step 810: loss 2.0515871047973633
 Step 820: loss 2.0159099102020264
 Step 830: loss 1.992288589477539
 Step 840: loss 1.908416509628296
 Step 850: loss 1.95501708984375
 Step 860: loss 1.9320729970932007
 Step 870: loss 2.057610034942627
 Step 880: loss 2.0288331508636475
 Step 890: loss 2.0372776985168457
 Step 900: loss 2.1234116554260254
 Step 910: loss 2.0997185707092285
 Step 920: loss 2.1006853580474854
 Step 930: loss 1.9545409679412842
 Step 940: loss 2.042897939682007
 Step 950: loss 2.092611789703369
 Step 960: loss 2.0068106651306152
 Step 970: loss 1.9601119756698608
 Step 980: loss 1.9791560173034668
 Step 990: loss 1.9645905494689941
 Step 1000: loss 1.9899816513061523
 [Seed 41] Step 1000 --> val loss 2.315990686416626
 Step 1010: loss 2.03853702545166
 Step 1020: loss 2.0504651069641113
 Step 1030: loss 2.0727970600128174
 Step 1040: loss 2.075150489807129
 Step 1050: loss 1.9584568738937378
 Step 1060: loss 1.9706761837005615
 Step 1070: loss 1.9523468017578125
 Step 1080: loss 1.975675344467163
 Step 1090: loss 1.8706129789352417
 Step 1100: loss 1.99373459815979
 Step 1110: loss 2.076079845428467
 Step 1120: loss 2.0293920040130615
 Step 1130: loss 2.039733409881592
 Step 1140: loss 2.0312769412994385
 Step 1150: loss 2.1537604331970215
 Step 1160: loss 1.9821233749389648
 Step 1170: loss 1.9963722229003906
 Step 1180: loss 2.032954692840576
 Step 1190: loss 2.05825138092041
 Step 1200: loss 2.00665545463562
 [Seed 41] Step 1200 --> val loss 2.261110782623291
 Step 1210: loss 1.9365031719207764
 Step 1220: loss 1.9978854656219482
 Step 1230: loss 1.9756927490234375
 Step 1240: loss 2.1067237854003906
 Step 1250: loss 1.9685823917388916
 Step 1260: loss 1.989954948425293
 Step 1270: loss 2.104304790496826
 Step 1280: loss 2.0551726818084717
 Step 1290: loss 1.9702380895614624
 Step 1300: loss 1.8999996185302734
 Step 1310: loss 2.043440341949463
 Step 1320: loss 2.0345168113708496
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 41] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 41] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 41] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 41] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [40 59 58  1 53 52 43  1 57 58] ...
 First target sequence: [59 58  1 53 52 43  1 57 58 43] ...

 Decoded input: but one step below,
 Even of your mettle, of your very blood;
 Of 

 Decoded target: ut one step below,
 Even of your mettle, of your very blood;
 Of a
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.207557678222656
 [Seed 42] Step 0 --> val loss 4.183665752410889
 Step 10: loss 3.8443045616149902
 Step 20: loss 3.3240013122558594
 Step 30: loss 2.9074602127075195
 Step 40: loss 2.6803646087646484
 Step 50: loss 2.6237547397613525
 Step 60: loss 2.474132537841797
 Step 70: loss 2.37544322013855
 Step 80: loss 2.467573881149292
 Step 90: loss 2.36454176902771
 Step 100: loss 2.34686279296875
 Step 110: loss 2.233403205871582
 Step 120: loss 2.320678234100342
 Step 130: loss 2.235042095184326
 Step 140: loss 2.2252042293548584
 Step 150: loss 2.1995913982391357
 Step 160: loss 2.2639312744140625
 Step 170: loss 2.18391752243042
 Step 180: loss 2.1971054077148438
 Step 190: loss 2.1040666103363037
 Step 200: loss 2.074162006378174
 [Seed 42] Step 200 --> val loss 2.3951573371887207
 Step 210: loss 2.2688136100769043
 Step 220: loss 2.0169880390167236
 Step 230: loss 2.092374563217163
 Step 240: loss 2.1641461849212646
 Step 250: loss 2.125023603439331
 Step 260: loss 1.9702643156051636
 Step 270: loss 2.0599160194396973
 Step 280: loss 2.1051430702209473
 Step 290: loss 2.0587453842163086
 Step 300: loss 2.143782377243042
 Step 310: loss 1.9247245788574219
 Step 320: loss 2.0157134532928467
 Step 330: loss 2.0854997634887695
 Step 340: loss 2.0292651653289795
 Step 350: loss 2.006636619567871
 Step 360: loss 2.114684820175171
 Step 370: loss 2.0230846405029297
 Step 380: loss 1.9485113620758057
 Step 390: loss 1.989101767539978
 Step 400: loss 1.9893344640731812
 [Seed 42] Step 400 --> val loss 2.203493595123291
 Step 410: loss 2.0288314819335938
 Step 420: loss 2.083268642425537
 Step 430: loss 2.0056722164154053
 Step 440: loss 2.068469762802124
 Step 450: loss 2.0420994758605957
 Step 460: loss 2.0393214225769043
 Step 470: loss 1.921349287033081
 Step 480: loss 2.09005069732666
 Step 490: loss 1.9291112422943115
 Step 500: loss 1.9973386526107788
 Step 510: loss 2.156639575958252
 Step 520: loss 2.156325340270996
 Step 530: loss 2.06378173828125
 Step 540: loss 2.028883934020996
 Step 550: loss 1.9552115201950073
 Step 560: loss 1.9423243999481201
 Step 570: loss 1.979333758354187
 Step 580: loss 1.991106390953064
 Step 590: loss 2.0553815364837646
 Step 600: loss 2.0631237030029297
 [Seed 42] Step 600 --> val loss 2.171243667602539
 Step 610: loss 2.0710418224334717
 Step 620: loss 2.0201916694641113
 Step 630: loss 2.06239914894104
 Step 640: loss 2.029562473297119
 Step 650: loss 1.9019124507904053
 Step 660: loss 2.006082773208618
 Step 670: loss 2.0779170989990234
 Step 680: loss 1.9610028266906738
 Step 690: loss 2.017559766769409
 Step 700: loss 2.0431902408599854
 Step 710: loss 1.9710168838500977
 Step 720: loss 1.9833883047103882
 Step 730: loss 2.00152325630188
 Step 740: loss 2.0120410919189453
 Step 750: loss 1.988094687461853
 Step 760: loss 1.9754388332366943
 Step 770: loss 2.0443525314331055
 Step 780: loss 1.982949137687683
 Step 790: loss 2.0567588806152344
 Step 800: loss 1.9292213916778564
 [Seed 42] Step 800 --> val loss 2.205906629562378
 Step 810: loss 1.9756357669830322
 Step 820: loss 2.0695810317993164
 Step 830: loss 1.8962063789367676
 Step 840: loss 1.9837325811386108
 Step 850: loss 1.9252543449401855
 Step 860: loss 1.9568060636520386
 Step 870: loss 2.083667755126953
 Step 880: loss 2.003279209136963
 Step 890: loss 2.045210361480713
 Step 900: loss 1.8115878105163574
 Step 910: loss 1.9681916236877441
 Step 920: loss 1.9874275922775269
 Step 930: loss 1.9679651260375977
 Step 940: loss 1.9553518295288086
 Step 950: loss 1.9413269758224487
 Step 960: loss 2.0044312477111816
 Step 970: loss 1.95261549949646
 Step 980: loss 1.863532304763794
 Step 990: loss 1.967365026473999
 Step 1000: loss 1.9694907665252686
 [Seed 42] Step 1000 --> val loss 2.144049882888794
 Step 1010: loss 1.9818061590194702
 Step 1020: loss 1.9829318523406982
 Step 1030: loss 1.9952154159545898
 Step 1040: loss 1.962836742401123
 Step 1050: loss 2.02083683013916
 Step 1060: loss 1.9057247638702393
 Step 1070: loss 2.0221986770629883
 Step 1080: loss 2.035327911376953
 Step 1090: loss 1.9479215145111084
 Step 1100: loss 2.0350258350372314
 Step 1110: loss 1.7757055759429932
 Step 1120: loss 1.9642906188964844
 Step 1130: loss 1.928487777709961
 Step 1140: loss 1.9380344152450562
 Step 1150: loss 1.9566272497177124
 Step 1160: loss 1.8780975341796875
 Step 1170: loss 1.8441388607025146
 Step 1180: loss 1.88483464717865
 Step 1190: loss 1.93291175365448
 Step 1200: loss 1.994161605834961
 [Seed 42] Step 1200 --> val loss 2.172584056854248
 Step 1210: loss 1.9174139499664307
 Step 1220: loss 2.0373332500457764
 Step 1230: loss 1.8021440505981445
 Step 1240: loss 1.805044174194336
 Step 1250: loss 1.85636305809021
 Step 1260: loss 2.034318447113037
 Step 1270: loss 1.7635104656219482
 Step 1280: loss 1.9370111227035522
 Step 1290: loss 1.9187648296356201
 Step 1300: loss 1.8306422233581543
 Step 1310: loss 1.9929962158203125
 Step 1320: loss 1.9052741527557373
 Step 1330: loss 1.9189567565917969
 Step 1340: loss 1.843712329864502
 Step 1350: loss 1.8933427333831787
 Step 1360: loss 1.7477396726608276
 Step 1370: loss 1.9328761100769043
 Step 1380: loss 1.8397586345672607
 Step 1390: loss 1.8760347366333008
 Step 1400: loss 1.897133708000183
 [Seed 42] Step 1400 --> val loss 1.9480022192001343
 Step 1410: loss 1.9094972610473633
 Step 1420: loss 1.8715341091156006
 Step 1430: loss 1.8423455953598022
 Step 1440: loss 1.8567559719085693
 Step 1450: loss 1.8520257472991943
 Step 1460: loss 1.8873157501220703
 Step 1470: loss 1.8156373500823975
 Step 1480: loss 1.712869644165039
 Step 1490: loss 1.804749846458435
 Step 1500: loss 1.8239248991012573
 Step 1510: loss 1.7168967723846436
 Step 1520: loss 1.8383746147155762
 Step 1530: loss 1.827126383781433
 Step 1540: loss 1.9359054565429688
 Step 1550: loss 1.7427024841308594
 Step 1560: loss 1.7795346975326538
 Step 1570: loss 1.8610706329345703
 Step 1580: loss 1.6950128078460693
 Step 1590: loss 1.7426519393920898
 Step 1600: loss 1.7251546382904053
 [Seed 42] Step 1600 --> val loss 1.9648525714874268
 Step 1610: loss 1.7761586904525757
 Step 1620: loss 1.778570532798767
 Step 1630: loss 1.7897967100143433
 Step 1640: loss 1.703801155090332
 Step 1650: loss 1.7584764957427979
 Step 1660: loss 1.7816882133483887
 Step 1670: loss 1.7816728353500366
 Step 1680: loss 1.6016582250595093
 Step 1690: loss 1.6107547283172607
 Step 1700: loss 1.71640944480896
 Step 1710: loss 1.7325729131698608
 Step 1720: loss 1.7113574743270874
 Step 1730: loss 1.7149112224578857
 Step 1740: loss 1.7001694440841675
 Step 1750: loss 1.6782855987548828
 Step 1760: loss 1.5935108661651611
 Step 1770: loss 1.7172893285751343
 Step 1780: loss 1.776443362236023
 Step 1790: loss 1.599801778793335
 Step 1800: loss 1.7347981929779053
 [Seed 42] Step 1800 --> val loss 1.909358024597168
 Step 1810: loss 1.7749817371368408
 Step 1820: loss 1.6420905590057373
 Step 1830: loss 1.6719214916229248
 Step 1840: loss 1.5952528715133667
 Step 1850: loss 1.5370280742645264
 Step 1860: loss 1.6869949102401733
 Step 1870: loss 1.7948691844940186
 Step 1880: loss 1.6473958492279053
 Step 1890: loss 1.653080940246582
 Step 1900: loss 1.641263484954834
 Step 1910: loss 1.6334521770477295
 Step 1920: loss 1.6353305578231812
 Step 1930: loss 1.690993070602417
 Step 1940: loss 1.6468031406402588
 Step 1950: loss 1.7641446590423584
 Step 1960: loss 1.6441158056259155
 Step 1970: loss 1.7246110439300537
 Step 1980: loss 1.6963129043579102
 Step 1990: loss 1.5292022228240967
 Step 2000: loss 1.5905184745788574
 [Seed 42] Step 2000 --> val loss 1.8629509210586548
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [39 50 10  0 35 43 39 56  1 47] ...
 First target sequence: [50 10  0 35 43 39 56  1 47 58] ...

 Decoded input: al:
 Wear it, enjoy it, and make much of it.

 RICHMOND:
 Great God

 Decoded target: l:
 Wear it, enjoy it, and make much of it.

 RICHMOND:
 Great God 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.216355323791504
 [Seed 43] Step 0 --> val loss 4.18345308303833
 Step 10: loss 3.8222007751464844
 Step 20: loss 3.319751262664795
 Step 30: loss 2.86531138420105
 Step 40: loss 2.534066915512085
 Step 50: loss 2.5775012969970703
 Step 60: loss 2.46254301071167
 Step 70: loss 2.4788055419921875
 Step 80: loss 2.2660698890686035
 Step 90: loss 2.4047367572784424
 Step 100: loss 2.436094284057617
 Step 110: loss 2.342034101486206
 Step 120: loss 2.205740451812744
 Step 130: loss 2.278078079223633
 Step 140: loss 2.188905715942383
 Step 150: loss 2.1818535327911377
 Step 160: loss 2.019566297531128
 Step 170: loss 2.1521060466766357
 Step 180: loss 2.1297130584716797
 Step 190: loss 2.0746045112609863
 Step 200: loss 2.0164122581481934
 [Seed 43] Step 200 --> val loss 2.3382503986358643
 Step 210: loss 2.1026816368103027
 Step 220: loss 2.1798932552337646
 Step 230: loss 2.160311460494995
 Step 240: loss 2.0775837898254395
 Step 250: loss 2.075143814086914
 Step 260: loss 2.1708590984344482
 Step 270: loss 2.0335426330566406
 Step 280: loss 2.0829696655273438
 Step 290: loss 1.9825842380523682
 Step 300: loss 2.0543909072875977
 Step 310: loss 1.9079965353012085
 Step 320: loss 2.170928478240967
 Step 330: loss 1.9971487522125244
 Step 340: loss 2.0540361404418945
 Step 350: loss 2.0600316524505615
 Step 360: loss 2.0744104385375977
 Step 370: loss 2.1422832012176514
 Step 380: loss 2.0546956062316895
 Step 390: loss 2.037724494934082
 Step 400: loss 2.0842745304107666
 [Seed 43] Step 400 --> val loss 2.2528107166290283
 Step 410: loss 2.133085250854492
 Step 420: loss 2.025425910949707
 Step 430: loss 2.0422263145446777
 Step 440: loss 1.9591357707977295
 Step 450: loss 1.970635175704956
 Step 460: loss 2.02778697013855
 Step 470: loss 1.996337890625
 Step 480: loss 2.1313891410827637
 Step 490: loss 1.9677757024765015
 Step 500: loss 1.982812762260437
 Step 510: loss 1.9386180639266968
 Step 520: loss 1.963597297668457
 Step 530: loss 2.0453531742095947
 Step 540: loss 2.062730550765991
 Step 550: loss 2.0198118686676025
 Step 560: loss 1.9732850790023804
 Step 570: loss 2.0662002563476562
 Step 580: loss 1.9305429458618164
 Step 590: loss 1.9870433807373047
 Step 600: loss 1.9821033477783203
 [Seed 43] Step 600 --> val loss 2.2145814895629883
 Step 610: loss 2.0313167572021484
 Step 620: loss 2.111675262451172
 Step 630: loss 2.042668104171753
 Step 640: loss 2.0770061016082764
 Step 650: loss 1.979906439781189
 Step 660: loss 1.9791210889816284
 Step 670: loss 1.895825743675232
 Step 680: loss 1.9314645528793335
 Step 690: loss 1.9702212810516357
 Step 700: loss 2.0487570762634277
 Step 710: loss 2.0312535762786865
 Step 720: loss 2.001614570617676
 Step 730: loss 2.088381290435791
 Step 740: loss 2.0629069805145264
 Step 750: loss 2.0507960319519043
 Step 760: loss 2.037595748901367
 Step 770: loss 2.0709633827209473
 Step 780: loss 2.0958762168884277
 Step 790: loss 2.0205695629119873
 Step 800: loss 2.129021644592285
 [Seed 43] Step 800 --> val loss 2.2082114219665527
 Step 810: loss 2.029097080230713
 Step 820: loss 2.044588804244995
 Step 830: loss 2.027886152267456
 Step 840: loss 2.07073974609375
 Step 850: loss 2.016038417816162
 Step 860: loss 1.957260012626648
 Step 870: loss 1.9664491415023804
 Step 880: loss 2.120288372039795
 Step 890: loss 1.9148459434509277
 Step 900: loss 2.131145477294922
 Step 910: loss 2.08819580078125
 Step 920: loss 2.04868745803833
 Step 930: loss 2.143218994140625
 Step 940: loss 1.9842637777328491
 Step 950: loss 2.008809804916382
 Step 960: loss 1.9561254978179932
 Step 970: loss 1.9833320379257202
 Step 980: loss 1.8785251379013062
 Step 990: loss 2.011413097381592
 Step 1000: loss 2.0609261989593506
 [Seed 43] Step 1000 --> val loss 2.258387327194214
 Step 1010: loss 2.054399251937866
 Step 1020: loss 1.9495782852172852
 Step 1030: loss 2.101158618927002
 Step 1040: loss 2.0394253730773926
 Step 1050: loss 1.9470186233520508
 Step 1060: loss 1.9042015075683594
 Step 1070: loss 1.895254373550415
 Step 1080: loss 1.9385552406311035
 Step 1090: loss 2.010899543762207
 Step 1100: loss 1.9502124786376953
 Step 1110: loss 1.9410837888717651
 Step 1120: loss 1.87680184841156
 Step 1130: loss 1.903037428855896
 Step 1140: loss 1.8826583623886108
 Step 1150: loss 2.0951032638549805
 Step 1160: loss 1.9930499792099
 Step 1170: loss 1.9402318000793457
 Step 1180: loss 1.9125967025756836
 Step 1190: loss 1.960934042930603
 Step 1200: loss 1.9418729543685913
 [Seed 43] Step 1200 --> val loss 2.1448440551757812
 Step 1210: loss 1.7957448959350586
 Step 1220: loss 2.1122701168060303
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 43] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 43] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 43] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 43] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 61 47 58 46  1 63 53 59 12] ...
 First target sequence: [61 47 58 46  1 63 53 59 12  0] ...

 Decoded input:  with you?

 ABHORSON:
 Truly, sir, I would desire you to clap int

 Decoded target: with you?

 ABHORSON:
 Truly, sir, I would desire you to clap into
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.195560455322266
 [Seed 44] Step 0 --> val loss 4.184371471405029
 Step 10: loss 3.8602709770202637
 Step 20: loss 3.372891902923584
 Step 30: loss 2.8647685050964355
 Step 40: loss 2.6945853233337402
 Step 50: loss 2.474022388458252
 Step 60: loss 2.487584352493286
 Step 70: loss 2.3086907863616943
 Step 80: loss 2.2793779373168945
 Step 90: loss 2.3466098308563232
 Step 100: loss 2.3631300926208496
 Step 110: loss 2.4486615657806396
 Step 120: loss 2.281825542449951
 Step 130: loss 2.2506158351898193
 Step 140: loss 2.3153648376464844
 Step 150: loss 2.204676628112793
 Step 160: loss 2.191243886947632
 Step 170: loss 2.0496387481689453
 Step 180: loss 2.130788803100586
 Step 190: loss 2.1463325023651123
 Step 200: loss 2.1945133209228516
 [Seed 44] Step 200 --> val loss 2.3218963146209717
 Step 210: loss 2.1624419689178467
 Step 220: loss 2.268357753753662
 Step 230: loss 2.1522018909454346
 Step 240: loss 2.215087890625
 Step 250: loss 2.175788640975952
 Step 260: loss 1.999258041381836
 Step 270: loss 2.1173360347747803
 Step 280: loss 2.2131659984588623
 Step 290: loss 2.130007743835449
 Step 300: loss 2.0109031200408936
 Step 310: loss 1.946777582168579
 Step 320: loss 2.0338332653045654
 Step 330: loss 2.0497286319732666
 Step 340: loss 2.093892812728882
 Step 350: loss 1.8924498558044434
 Step 360: loss 1.9101495742797852
 Step 370: loss 2.1946332454681396
 Step 380: loss 2.0771031379699707
 Step 390: loss 1.986328363418579
 Step 400: loss 1.9559869766235352
 [Seed 44] Step 400 --> val loss 2.2015159130096436
 Step 410: loss 1.9566164016723633
 Step 420: loss 2.034881591796875
 Step 430: loss 2.0250906944274902
 Step 440: loss 1.8483929634094238
 Step 450: loss 2.032839775085449
 Step 460: loss 1.9607555866241455
 Step 470: loss 1.9976909160614014
 Step 480: loss 1.9956477880477905
 Step 490: loss 1.9547650814056396
 Step 500: loss 2.0121006965637207
 Step 510: loss 2.0188257694244385
 Step 520: loss 2.0416626930236816
 Step 530: loss 2.0255086421966553
 Step 540: loss 2.131267786026001
 Step 550: loss 1.9341920614242554
 Step 560: loss 1.9905160665512085
 Step 570: loss 2.0239827632904053
 Step 580: loss 1.9904067516326904
 Step 590: loss 2.128528118133545
 Step 600: loss 2.0318665504455566
 [Seed 44] Step 600 --> val loss 2.2773892879486084
 Step 610: loss 1.9138548374176025
 Step 620: loss 1.9575984477996826
 Step 630: loss 1.9591691493988037
 Step 640: loss 1.972815990447998
 Step 650: loss 1.9584717750549316
 Step 660: loss 1.9706480503082275
 Step 670: loss 2.0828142166137695
 Step 680: loss 2.0929977893829346
 Step 690: loss 1.9626119136810303
 Step 700: loss 2.005220413208008
 Step 710: loss 1.9962379932403564
 Step 720: loss 1.9677581787109375
 Step 730: loss 1.93310546875
 Step 740: loss 1.9268090724945068
 Step 750: loss 1.93989896774292
 Step 760: loss 1.8318575620651245
 Step 770: loss 1.9741524457931519
 Step 780: loss 1.9971169233322144
 Step 790: loss 1.9810703992843628
 Step 800: loss 1.9465720653533936
 [Seed 44] Step 800 --> val loss 2.298759937286377
 Step 810: loss 1.9719057083129883
 Step 820: loss 2.090426206588745
 Step 830: loss 1.9105638265609741
 Step 840: loss 1.9642388820648193
 Step 850: loss 1.9394086599349976
 Step 860: loss 1.9956016540527344
 Step 870: loss 1.9897470474243164
 Step 880: loss 2.063217878341675
 Step 890: loss 1.908010721206665
 Step 900: loss 1.9764997959136963
 Step 910: loss 2.0279486179351807
 Step 920: loss 2.040904998779297
 Step 930: loss 2.0331037044525146
 Step 940: loss 1.9543006420135498
 Step 950: loss 2.0940182209014893
 Step 960: loss 1.9304651021957397
 Step 970: loss 1.9949114322662354
 Step 980: loss 2.0073208808898926
 Step 990: loss 2.064988613128662
 Step 1000: loss 2.0260632038116455
 [Seed 44] Step 1000 --> val loss 2.252042770385742
 Step 1010: loss 2.118624210357666
 Step 1020: loss 2.026155471801758
 Step 1030: loss 2.0032100677490234
 Step 1040: loss 2.011857509613037
 Step 1050: loss 1.996565580368042
 Step 1060: loss 1.920041561126709
 Step 1070: loss 1.9879777431488037
 Step 1080: loss 1.9843251705169678
 Step 1090: loss 2.082341194152832
 Step 1100: loss 2.0804250240325928
 Step 1110: loss 1.9939775466918945
 Step 1120: loss 1.9755744934082031
 Step 1130: loss 1.957010269165039
 Step 1140: loss 1.96246337890625
 Step 1150: loss 1.9567453861236572
 Step 1160: loss 2.017517566680908
 Step 1170: loss 1.9382117986679077
 Step 1180: loss 1.9914350509643555
 Step 1190: loss 2.0653438568115234
 Step 1200: loss 2.0647761821746826
 [Seed 44] Step 1200 --> val loss 2.243245840072632
 Step 1210: loss 2.0897216796875
 Step 1220: loss 2.0050668716430664
 Step 1230: loss 1.9937751293182373
 Step 1240: loss 2.019833564758301
 Step 1250: loss 2.0469417572021484
 Step 1260: loss 1.927304983139038
 Step 1270: loss 2.0829379558563232
 Step 1280: loss 2.030764102935791
 Step 1290: loss 2.0734260082244873
 Step 1300: loss 1.9961116313934326
 Step 1310: loss 1.9332904815673828
 Step 1320: loss 2.1202988624572754
 Step 1330: loss 1.835750937461853
 Step 1340: loss 1.9837675094604492
 Step 1350: loss 1.9083369970321655
 Step 1360: loss 1.9967845678329468
 Step 1370: loss 1.865460753440857
 Step 1380: loss 1.9708794355392456
 Step 1390: loss 1.9049838781356812
 Step 1400: loss 1.9054386615753174
 [Seed 44] Step 1400 --> val loss 2.1728081703186035
 Step 1410: loss 1.9620246887207031
 Step 1420: loss 1.8433730602264404
 Step 1430: loss 1.8908822536468506
 Step 1440: loss 1.9317210912704468
 Step 1450: loss 1.937718391418457
 Step 1460: loss 1.9450204372406006
 Step 1470: loss 1.8535709381103516
 Step 1480: loss 1.9384870529174805
 Step 1490: loss 1.9729207754135132
 Step 1500: loss 1.933164358139038
 Step 1510: loss 1.9859713315963745
 Step 1520: loss 1.8590009212493896
 Step 1530: loss 1.8742518424987793
 Step 1540: loss 2.0886752605438232
 Step 1550: loss 1.874671220779419
 Step 1560: loss 2.029768943786621
 Step 1570: loss 1.8814195394515991
 Step 1580: loss 1.7410211563110352
 Step 1590: loss 1.8810933828353882
 Step 1600: loss 1.9071801900863647
 [Seed 44] Step 1600 --> val loss 2.0196917057037354
 Step 1610: loss 1.762007713317871
 Step 1620: loss 1.892970323562622
 Step 1630: loss 1.89072847366333
 Step 1640: loss 1.8839826583862305
 Step 1650: loss 1.885664701461792
 Step 1660: loss 1.8486170768737793
 Step 1670: loss 1.7612829208374023
 Step 1680: loss 1.7883870601654053
 Step 1690: loss 1.8708651065826416
 Step 1700: loss 1.8408327102661133
 Step 1710: loss 1.7045745849609375
 Step 1720: loss 1.913798213005066
 Step 1730: loss 1.7944767475128174
 Step 1740: loss 1.7894309759140015
 Step 1750: loss 1.7364702224731445
 Step 1760: loss 1.7891292572021484
 Step 1770: loss 1.8853328227996826
 Step 1780: loss 1.8883731365203857
 Step 1790: loss 1.8336257934570312
 Step 1800: loss 1.8384907245635986
 [Seed 44] Step 1800 --> val loss 1.9613771438598633
 Step 1810: loss 1.7672799825668335
 Step 1820: loss 1.783105731010437
 Step 1830: loss 1.8634542226791382
 Step 1840: loss 1.8523176908493042
 Step 1850: loss 1.7987146377563477
 Step 1860: loss 1.7039625644683838
 Step 1870: loss 1.7786086797714233
 Step 1880: loss 1.7480844259262085
 Step 1890: loss 1.8999003171920776
 Step 1900: loss 1.8652236461639404
 Step 1910: loss 1.656134009361267
 Step 1920: loss 1.7024104595184326
 Step 1930: loss 1.7984155416488647
 Step 1940: loss 1.8031597137451172
 Step 1950: loss 1.787292242050171
 Step 1960: loss 1.5782221555709839
 Step 1970: loss 1.6613452434539795
 Step 1980: loss 1.7515491247177124
 Step 1990: loss 1.7258180379867554
 Step 2000: loss 1.675703525543213
 [Seed 44] Step 2000 --> val loss 1.9264118671417236
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [57 54 43 39 49  5 57 58  1 58] ...
 First target sequence: [54 43 39 49  5 57 58  1 58 56] ...

 Decoded input: speak'st truth,
 Methinks thou speak'st not well.
 How long is't s

 Decoded target: peak'st truth,
 Methinks thou speak'st not well.
 How long is't si
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.218226432800293
 [Seed 45] Step 0 --> val loss 4.18561315536499
 Step 10: loss 3.845210552215576
 Step 20: loss 3.34033465385437
 Step 30: loss 2.8443431854248047
 Step 40: loss 2.637694835662842
 Step 50: loss 2.486351490020752
 Step 60: loss 2.4065113067626953
 Step 70: loss 2.3957040309906006
 Step 80: loss 2.373859405517578
 Step 90: loss 2.3781940937042236
 Step 100: loss 2.3237133026123047
 Step 110: loss 2.32839298248291
 Step 120: loss 2.168778896331787
 Step 130: loss 2.1936049461364746
 Step 140: loss 2.3138585090637207
 Step 150: loss 2.2116270065307617
 Step 160: loss 2.2375996112823486
 Step 170: loss 2.167539358139038
 Step 180: loss 2.1407127380371094
 Step 190: loss 2.114011287689209
 Step 200: loss 2.1909427642822266
 [Seed 45] Step 200 --> val loss 2.317364454269409
 Step 210: loss 2.1348328590393066
 Step 220: loss 2.07987642288208
 Step 230: loss 2.07606840133667
 Step 240: loss 2.1269335746765137
 Step 250: loss 2.0522308349609375
 Step 260: loss 2.0151522159576416
 Step 270: loss 2.07499098777771
 Step 280: loss 2.0858168601989746
 Step 290: loss 2.0294604301452637
 Step 300: loss 1.9766920804977417
 Step 310: loss 1.9406745433807373
 Step 320: loss 2.0456440448760986
 Step 330: loss 1.9764198064804077
 Step 340: loss 1.9258317947387695
 Step 350: loss 2.0560531616210938
 Step 360: loss 2.0103678703308105
 Step 370: loss 2.012200355529785
 Step 380: loss 2.100795269012451
 Step 390: loss 2.1147947311401367
 Step 400: loss 1.986870527267456
 [Seed 45] Step 400 --> val loss 2.383880138397217
 Step 410: loss 1.9927266836166382
 Step 420: loss 1.9038641452789307
 Step 430: loss 1.9903978109359741
 Step 440: loss 1.941037893295288
 Step 450: loss 2.0073347091674805
 Step 460: loss 1.9922746419906616
 Step 470: loss 2.118408679962158
 Step 480: loss 2.034350872039795
 Step 490: loss 2.063779354095459
 Step 500: loss 1.9072790145874023
 Step 510: loss 2.120375633239746
 Step 520: loss 2.054626941680908
 Step 530: loss 1.971335768699646
 Step 540: loss 1.9357637166976929
 Step 550: loss 2.0754618644714355
 Step 560: loss 1.9734610319137573
 Step 570: loss 2.061330795288086
 Step 580: loss 2.040374755859375
 Step 590: loss 2.0499768257141113
 Step 600: loss 2.0410733222961426
 [Seed 45] Step 600 --> val loss 2.188605785369873
 Step 610: loss 1.9174407720565796
 Step 620: loss 1.9416530132293701
 Step 630: loss 1.981580138206482
 Step 640: loss 2.0990772247314453
 Step 650: loss 2.056086540222168
 Step 660: loss 2.0272767543792725
 Step 670: loss 1.9240286350250244
 Step 680: loss 1.868708848953247
 Step 690: loss 1.9900058507919312
 Step 700: loss 1.9703967571258545
 Step 710: loss 1.902014136314392
 Step 720: loss 1.9844825267791748
 Step 730: loss 1.967558741569519
 Step 740: loss 1.9900931119918823
 Step 750: loss 1.9482486248016357
 Step 760: loss 1.9787667989730835
 Step 770: loss 1.8725385665893555
 Step 780: loss 2.0172908306121826
 Step 790: loss 1.9003009796142578
 Step 800: loss 2.035501480102539
 [Seed 45] Step 800 --> val loss 2.250246047973633
 Step 810: loss 2.013705015182495
 Step 820: loss 2.0203492641448975
 Step 830: loss 1.9879668951034546
 Step 840: loss 1.9553420543670654
 Step 850: loss 1.9153518676757812
 Step 860: loss 2.102235794067383
 Step 870: loss 2.011376142501831
 Step 880: loss 1.8769716024398804
 Step 890: loss 1.8930314779281616
 Step 900: loss 1.9467405080795288
 Step 910: loss 1.9421466588974
 Step 920: loss 1.9736897945404053
 Step 930: loss 1.9635744094848633
 Step 940: loss 1.8921031951904297
 Step 950: loss 1.8829600811004639
 Step 960: loss 1.95414137840271
 Step 970: loss 1.9290695190429688
 Step 980: loss 1.8269126415252686
 Step 990: loss 1.8678455352783203
 Step 1000: loss 2.0444507598876953
 [Seed 45] Step 1000 --> val loss 2.1412570476531982
 Step 1010: loss 1.8643393516540527
 Step 1020: loss 2.019320487976074
 Step 1030: loss 1.9910054206848145
 Step 1040: loss 1.9042916297912598
 Step 1050: loss 2.0391697883605957
 Step 1060: loss 1.8668922185897827
 Step 1070: loss 1.833942174911499
 Step 1080: loss 1.9831020832061768
 Step 1090: loss 2.0073792934417725
 Step 1100: loss 1.823965072631836
 Step 1110: loss 1.9019416570663452
 Step 1120: loss 1.8469918966293335
 Step 1130: loss 1.9166345596313477
 Step 1140: loss 1.8734021186828613
 Step 1150: loss 1.8540732860565186
 Step 1160: loss 1.8033597469329834
 Step 1170: loss 2.0537476539611816
 Step 1180: loss 1.9615514278411865
 Step 1190: loss 1.9183906316757202
 Step 1200: loss 1.9440631866455078
 [Seed 45] Step 1200 --> val loss 2.1424336433410645
 Step 1210: loss 1.859830617904663
 Step 1220: loss 1.9055957794189453
 Step 1230: loss 1.9028778076171875
 Step 1240: loss 1.938226580619812
 Step 1250: loss 1.9222228527069092
 Step 1260: loss 1.8620367050170898
 Step 1270: loss 1.8386777639389038
 Step 1280: loss 1.9376320838928223
 Step 1290: loss 1.7605781555175781
 Step 1300: loss 1.9143189191818237
 Step 1310: loss 1.8531219959259033
 Step 1320: loss 1.822697401046753
 Step 1330: loss 1.7860991954803467
 Step 1340: loss 1.9187393188476562
 Step 1350: loss 1.918561339378357
 Step 1360: loss 1.8708059787750244
 Step 1370: loss 1.8454669713974
 Step 1380: loss 1.781925916671753
 Step 1390: loss 1.8625946044921875
 Step 1400: loss 1.6960608959197998
 [Seed 45] Step 1400 --> val loss 2.0532844066619873
 Step 1410: loss 1.8810800313949585
 Step 1420: loss 1.9309190511703491
 Step 1430: loss 1.7379987239837646
 Step 1440: loss 1.9663410186767578
 Step 1450: loss 1.8813461065292358
 Step 1460: loss 1.7943332195281982
 Step 1470: loss 1.8865375518798828
 Step 1480: loss 1.816177248954773
 Step 1490: loss 1.8204244375228882
 Step 1500: loss 1.8890295028686523
 Step 1510: loss 1.8517379760742188
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 45] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 45] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 45] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 45 53  1 39 40 53 59 58  1] ...
 First target sequence: [45 53  1 39 40 53 59 58  1 58] ...

 Decoded input:  go about to make
 me the king's brother-in-law.

 Clown:
 Indeed, 

 Decoded target: go about to make
 me the king's brother-in-law.

 Clown:
 Indeed, b
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.199724197387695
 [Seed 46] Step 0 --> val loss 4.1883111000061035
 Step 10: loss 3.8505282402038574
 Step 20: loss 3.3170688152313232
 Step 30: loss 2.8458919525146484
 Step 40: loss 2.6147992610931396
 Step 50: loss 2.5784623622894287
 Step 60: loss 2.5098233222961426
 Step 70: loss 2.36934757232666
 Step 80: loss 2.426421880722046
 Step 90: loss 2.3388946056365967
 Step 100: loss 2.341273546218872
 Step 110: loss 2.2381253242492676
 Step 120: loss 2.1142537593841553
 Step 130: loss 2.123103380203247
 Step 140: loss 2.2301528453826904
 Step 150: loss 2.199735164642334
 Step 160: loss 2.1220614910125732
 Step 170: loss 2.052746295928955
 Step 180: loss 2.108912467956543
 Step 190: loss 2.048466920852661
 Step 200: loss 2.111416816711426
 [Seed 46] Step 200 --> val loss 2.2794017791748047
 Step 210: loss 2.2077980041503906
 Step 220: loss 2.124868392944336
 Step 230: loss 2.0513756275177
 Step 240: loss 2.0354528427124023
 Step 250: loss 2.075145959854126
 Step 260: loss 2.025876760482788
 Step 270: loss 1.9922454357147217
 Step 280: loss 2.037135124206543
 Step 290: loss 2.0380165576934814
 Step 300: loss 2.0645833015441895
 Step 310: loss 2.0094566345214844
 Step 320: loss 2.0458009243011475
 Step 330: loss 2.01358962059021
 Step 340: loss 1.9233882427215576
 Step 350: loss 1.9392282962799072
 Step 360: loss 1.9018192291259766
 Step 370: loss 2.1351094245910645
 Step 380: loss 1.9861053228378296
 Step 390: loss 1.901707649230957
 Step 400: loss 1.9834555387496948
 [Seed 46] Step 400 --> val loss 2.220329999923706
 Step 410: loss 1.989040732383728
 Step 420: loss 2.0440258979797363
 Step 430: loss 1.937217116355896
 Step 440: loss 1.906667709350586
 Step 450: loss 1.9113268852233887
 Step 460: loss 1.9790233373641968
 Step 470: loss 2.0408260822296143
 Step 480: loss 1.9581959247589111
 Step 490: loss 1.9920945167541504
 Step 500: loss 1.935844898223877
 Step 510: loss 1.963958501815796
 Step 520: loss 1.973436713218689
 Step 530: loss 1.9718749523162842
 Step 540: loss 1.9268605709075928
 Step 550: loss 2.0437521934509277
 Step 560: loss 2.0763673782348633
 Step 570: loss 1.9798777103424072
 Step 580: loss 1.901681661605835
 Step 590: loss 1.9873371124267578
 Step 600: loss 2.0086660385131836
 [Seed 46] Step 600 --> val loss 2.222862958908081
 Step 610: loss 2.057291269302368
 Step 620: loss 2.004983425140381
 Step 630: loss 1.9762182235717773
 Step 640: loss 1.8799831867218018
 Step 650: loss 2.012388229370117
 Step 660: loss 1.953185796737671
 Step 670: loss 1.943298101425171
 Step 680: loss 1.9479583501815796
 Step 690: loss 1.9451792240142822
 Step 700: loss 1.8310420513153076
 Step 710: loss 1.9589670896530151
 Step 720: loss 1.889360785484314
 Step 730: loss 1.9887475967407227
 Step 740: loss 1.973966360092163
 Step 750: loss 2.090125799179077
 Step 760: loss 1.9683603048324585
 Step 770: loss 1.9535071849822998
 Step 780: loss 1.974819540977478
 Step 790: loss 2.0433928966522217
 Step 800: loss 1.9540923833847046
 [Seed 46] Step 800 --> val loss 2.1581642627716064
 Step 810: loss 1.9519550800323486
 Step 820: loss 1.9424322843551636
 Step 830: loss 1.9277830123901367
 Step 840: loss 1.9870309829711914
 Step 850: loss 1.9400031566619873
 Step 860: loss 2.0300095081329346
 Step 870: loss 1.979515790939331
 Step 880: loss 1.932952642440796
 Step 890: loss 1.929987907409668
 Step 900: loss 1.8886383771896362
 Step 910: loss 1.9562129974365234
 Step 920: loss 1.9420350790023804
 Step 930: loss 1.9665101766586304
 Step 940: loss 1.9356532096862793
 Step 950: loss 1.8707504272460938
 Step 960: loss 1.970461368560791
 Step 970: loss 1.8999799489974976
 Step 980: loss 1.894364356994629
 Step 990: loss 1.9748098850250244
 Step 1000: loss 1.9280494451522827
 [Seed 46] Step 1000 --> val loss 2.1624226570129395
 Step 1010: loss 1.9818552732467651
 Step 1020: loss 1.916853904724121
 Step 1030: loss 1.9806392192840576
 Step 1040: loss 1.9629418849945068
 Step 1050: loss 1.978031873703003
 Step 1060: loss 1.9805831909179688
 Step 1070: loss 1.9286760091781616
 Step 1080: loss 1.9665632247924805
 Step 1090: loss 1.9014763832092285
 Step 1100: loss 1.8301976919174194
 Step 1110: loss 1.9991562366485596
 Step 1120: loss 1.9298452138900757
 Step 1130: loss 1.9364612102508545
 Step 1140: loss 1.9631117582321167
 Step 1150: loss 1.9109880924224854
 Step 1160: loss 2.090933322906494
 Step 1170: loss 1.9490001201629639
 Step 1180: loss 2.0465474128723145
 Step 1190: loss 2.0399460792541504
 Step 1200: loss 2.0026187896728516
 [Seed 46] Step 1200 --> val loss 2.2100718021392822
 Step 1210: loss 1.9534982442855835
 Step 1220: loss 1.971358060836792
 Step 1230: loss 1.9543265104293823
 Step 1240: loss 1.9201023578643799
 Step 1250: loss 1.8921706676483154
 Step 1260: loss 1.851694107055664
 Step 1270: loss 1.979395866394043
 Step 1280: loss 1.928715705871582
 Step 1290: loss 1.987770438194275
 Step 1300: loss 1.9222053289413452
 Step 1310: loss 1.8702707290649414
 Step 1320: loss 1.9164271354675293
 Step 1330: loss 1.8521792888641357
 Step 1340: loss 1.897587537765503
 Step 1350: loss 1.9422025680541992
 Step 1360: loss 1.8698794841766357
 Step 1370: loss 2.028672695159912
 Step 1380: loss 1.9089159965515137
 Step 1390: loss 2.005448341369629
 Step 1400: loss 1.9192737340927124
 [Seed 46] Step 1400 --> val loss 2.1573293209075928
 Step 1410: loss 1.8326276540756226
 Step 1420: loss 1.8879868984222412
 Step 1430: loss 1.829769492149353
 Step 1440: loss 1.7933788299560547
 Step 1450: loss 1.8495666980743408
 Step 1460: loss 1.893218994140625
 Step 1470: loss 1.8895611763000488
 Step 1480: loss 1.8296022415161133
 Step 1490: loss 1.7160991430282593
 Step 1500: loss 1.8688714504241943
 Step 1510: loss 1.8576821088790894
 Step 1520: loss 1.8551194667816162
 Step 1530: loss 1.7964180707931519
 Step 1540: loss 1.9483118057250977
 Step 1550: loss 1.927034854888916
 Step 1560: loss 1.8917853832244873
 Step 1570: loss 1.920890212059021
 Step 1580: loss 1.7931487560272217
 Step 1590: loss 1.9214122295379639
 Step 1600: loss 1.9009339809417725
 [Seed 46] Step 1600 --> val loss 2.0281856060028076
 Step 1610: loss 1.7990096807479858
 Step 1620: loss 1.8386647701263428
 Step 1630: loss 1.841322660446167
 Step 1640: loss 1.7727149724960327
 Step 1650: loss 1.8678977489471436
 Step 1660: loss 1.8094096183776855
 Step 1670: loss 1.8587884902954102
 Step 1680: loss 1.649344801902771
 Step 1690: loss 1.8494415283203125
 Step 1700: loss 1.8742389678955078
 Step 1710: loss 1.7801040410995483
 Step 1720: loss 1.7997956275939941
 Step 1730: loss 1.849406123161316
 Step 1740: loss 1.834718942642212
 Step 1750: loss 1.7856420278549194
 Step 1760: loss 1.8470253944396973
 Step 1770: loss 1.8204492330551147
 Step 1780: loss 1.6817363500595093
 Step 1790: loss 1.8544127941131592
 Step 1800: loss 1.7895450592041016
 [Seed 46] Step 1800 --> val loss 1.8601468801498413
 Step 1810: loss 1.712463140487671
 Step 1820: loss 1.8054145574569702
 Step 1830: loss 1.7038204669952393
 Step 1840: loss 1.71878182888031
 Step 1850: loss 1.7286324501037598
 Step 1860: loss 1.82432222366333
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 46] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 58 43 52  1 42 39 63 57  5] ...
 First target sequence: [58 43 52  1 42 39 63 57  5  1] ...

 Decoded input:  ten days' wonder at the least.

 CLARENCE:
 That's a day longer t

 Decoded target: ten days' wonder at the least.

 CLARENCE:
 That's a day longer th
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.216434001922607
 [Seed 47] Step 0 --> val loss 4.184873104095459
 Step 10: loss 3.872704029083252
 Step 20: loss 3.293999671936035
 Step 30: loss 3.0240988731384277
 Step 40: loss 2.719097852706909
 Step 50: loss 2.690526008605957
 Step 60: loss 2.559978485107422
 Step 70: loss 2.3804931640625
 Step 80: loss 2.438976764678955
 Step 90: loss 2.302849292755127
 Step 100: loss 2.309640407562256
 Step 110: loss 2.2721524238586426
 Step 120: loss 2.2559080123901367
 Step 130: loss 2.2807974815368652
 Step 140: loss 2.2365822792053223
 Step 150: loss 2.134075880050659
 Step 160: loss 2.183804988861084
 Step 170: loss 2.115598201751709
 Step 180: loss 2.1807219982147217
 Step 190: loss 2.140273094177246
 Step 200: loss 2.083378314971924
 [Seed 47] Step 200 --> val loss 2.3404059410095215
 Step 210: loss 2.0857295989990234
 Step 220: loss 2.139655590057373
 Step 230: loss 2.092264175415039
 Step 240: loss 2.075899124145508
 Step 250: loss 2.1548283100128174
 Step 260: loss 2.0573267936706543
 Step 270: loss 2.0320167541503906
 Step 280: loss 2.0035860538482666
 Step 290: loss 2.02405047416687
 Step 300: loss 2.0535614490509033
 Step 310: loss 2.143951892852783
 Step 320: loss 2.044571876525879
 Step 330: loss 2.1333329677581787
 Step 340: loss 2.0023317337036133
 Step 350: loss 2.049020767211914
 Step 360: loss 1.8966535329818726
 Step 370: loss 2.041576385498047
 Step 380: loss 2.0769641399383545
 Step 390: loss 2.047649383544922
 Step 400: loss 2.156049966812134
 [Seed 47] Step 400 --> val loss 2.281810998916626
 Step 410: loss 2.0363211631774902
 Step 420: loss 1.8464158773422241
 Step 430: loss 1.999803900718689
 Step 440: loss 2.046489715576172
 Step 450: loss 1.9996293783187866
 Step 460: loss 2.0273070335388184
 Step 470: loss 2.018932819366455
 Step 480: loss 2.033649444580078
 Step 490: loss 2.0211997032165527
 Step 500: loss 1.9505813121795654
 Step 510: loss 2.0968360900878906
 Step 520: loss 1.9794974327087402
 Step 530: loss 1.9758358001708984
 Step 540: loss 2.058182954788208
 Step 550: loss 1.9508146047592163
 Step 560: loss 1.9955228567123413
 Step 570: loss 1.9055322408676147
 Step 580: loss 2.028407573699951
 Step 590: loss 1.8939409255981445
 Step 600: loss 2.0134692192077637
 [Seed 47] Step 600 --> val loss 2.259728193283081
 Step 610: loss 1.9474034309387207
 Step 620: loss 2.020059585571289
 Step 630: loss 2.0282979011535645
 Step 640: loss 1.9721107482910156
 Step 650: loss 1.9222021102905273
 Step 660: loss 1.9407858848571777
 Step 670: loss 2.0295073986053467
 Step 680: loss 1.9297635555267334
 Step 690: loss 1.940359115600586
 Step 700: loss 2.020425796508789
 Step 710: loss 1.8559272289276123
 Step 720: loss 1.9456008672714233
 Step 730: loss 1.938002586364746
 Step 740: loss 1.9164910316467285
 Step 750: loss 1.9880870580673218
 Step 760: loss 2.018857955932617
 Step 770: loss 2.032376766204834
 Step 780: loss 1.8996648788452148
 Step 790: loss 1.9653851985931396
 Step 800: loss 2.031393051147461
 [Seed 47] Step 800 --> val loss 2.1801328659057617
 Step 810: loss 2.0246052742004395
 Step 820: loss 2.0866260528564453
 Step 830: loss 1.9603233337402344
 Step 840: loss 2.079561233520508
 Step 850: loss 1.925970435142517
 Step 860: loss 2.055623769760132
 Step 870: loss 2.004678726196289
 Step 880: loss 1.947453260421753
 Step 890: loss 2.0821123123168945
 Step 900: loss 1.9537627696990967
 Step 910: loss 2.027881622314453
 Step 920: loss 2.0473201274871826
 Step 930: loss 1.9633941650390625
 Step 940: loss 1.9771173000335693
 Step 950: loss 1.9635661840438843
 Step 960: loss 2.0221781730651855
 Step 970: loss 2.063213586807251
 Step 980: loss 1.9418416023254395
 Step 990: loss 1.9445905685424805
 Step 1000: loss 1.9408185482025146
 [Seed 47] Step 1000 --> val loss 2.3030643463134766
 Step 1010: loss 1.9998376369476318
 Step 1020: loss 2.019733190536499
 Step 1030: loss 2.078256368637085
 Step 1040: loss 1.9440028667449951
 Step 1050: loss 2.0823402404785156
 Step 1060: loss 2.042586326599121
 Step 1070: loss 2.0129895210266113
 Step 1080: loss 1.9671156406402588
 Step 1090: loss 1.9201468229293823
 Step 1100: loss 2.024097442626953
 Step 1110: loss 1.9680503606796265
 Step 1120: loss 1.9614022970199585
 Step 1130: loss 1.965146541595459
 Step 1140: loss 1.9169795513153076
 Step 1150: loss 1.984866976737976
 Step 1160: loss 1.9871912002563477
 Step 1170: loss 1.9690335988998413
 Step 1180: loss 1.9032480716705322
 Step 1190: loss 1.8956739902496338
 Step 1200: loss 1.9229964017868042
 [Seed 47] Step 1200 --> val loss 2.2465243339538574
 Step 1210: loss 1.9081964492797852
 Step 1220: loss 1.8104791641235352
 Step 1230: loss 1.8672285079956055
 Step 1240: loss 1.888859510421753
 Step 1250: loss 1.894723653793335
 Step 1260: loss 1.925215721130371
 Step 1270: loss 1.8494548797607422
 Step 1280: loss 1.7293155193328857
 Step 1290: loss 1.8753564357757568
 Step 1300: loss 1.8422918319702148
 Step 1310: loss 1.891197919845581
 Step 1320: loss 1.8345450162887573
 Step 1330: loss 1.9196593761444092
 Step 1340: loss 1.9153499603271484
 Step 1350: loss 1.8341339826583862
 Step 1360: loss 1.8515281677246094
 Step 1370: loss 1.8893671035766602
 Step 1380: loss 1.8096178770065308
 Step 1390: loss 1.8020458221435547
 Step 1400: loss 1.919959545135498
 [Seed 47] Step 1400 --> val loss 2.201455593109131
 Step 1410: loss 1.7926310300827026
 Step 1420: loss 1.7883027791976929
 Step 1430: loss 1.9547109603881836
 Step 1440: loss 1.9878486394882202
 Step 1450: loss 1.972311019897461
 Step 1460: loss 1.9236587285995483
 Step 1470: loss 1.8258154392242432
 Step 1480: loss 1.9710367918014526
 Step 1490: loss 1.849072813987732
 Step 1500: loss 1.8229742050170898
 Step 1510: loss 1.7808341979980469
 Step 1520: loss 1.8720815181732178
 Step 1530: loss 1.897419810295105
 Step 1540: loss 1.7667607069015503
 Step 1550: loss 1.933786153793335
 Step 1560: loss 1.8967387676239014
 Step 1570: loss 1.840998649597168
 Step 1580: loss 1.8013169765472412
 Step 1590: loss 1.8526605367660522
 Step 1600: loss 1.7991564273834229
 [Seed 47] Step 1600 --> val loss 2.0798776149749756
 Step 1610: loss 1.818068265914917
 Step 1620: loss 1.7106908559799194
 Step 1630: loss 1.8135802745819092
 Step 1640: loss 1.7287521362304688
 Step 1650: loss 1.849806785583496
 Step 1660: loss 1.8694355487823486
 Step 1670: loss 1.8439664840698242
 Step 1680: loss 1.6948325634002686
 Step 1690: loss 1.7003726959228516
 Step 1700: loss 1.8256282806396484
 Step 1710: loss 1.7500419616699219
 Step 1720: loss 1.6625096797943115
 Step 1730: loss 1.8120992183685303
 Step 1740: loss 1.7789167165756226
 Step 1750: loss 1.6481531858444214
 Step 1760: loss 1.7109698057174683
 Step 1770: loss 1.6812255382537842
 Step 1780: loss 1.839977741241455
 Step 1790: loss 1.8784888982772827
 Step 1800: loss 1.769364356994629
 [Seed 47] Step 1800 --> val loss 2.027337074279785
 Step 1810: loss 1.7026641368865967
 Step 1820: loss 1.8742755651474
 Step 1830: loss 1.7910264730453491
 Step 1840: loss 1.6271939277648926
 Step 1850: loss 1.7479327917099
 Step 1860: loss 1.7123606204986572
 Step 1870: loss 1.7950619459152222
 Step 1880: loss 1.7826316356658936
 Step 1890: loss 1.6639716625213623
 Step 1900: loss 1.7767428159713745
 Step 1910: loss 1.6646965742111206
 Step 1920: loss 1.696272850036621
 Step 1930: loss 1.842576026916504
 Step 1940: loss 1.7135347127914429
 Step 1950: loss 1.6935538053512573
 Step 1960: loss 1.8457143306732178
 Step 1970: loss 1.6830732822418213
 Step 1980: loss 1.605644702911377
 Step 1990: loss 1.839363932609558
 Step 2000: loss 1.7992370128631592
 [Seed 47] Step 2000 --> val loss 1.9535969495773315
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [52 41 43  8  0 32 46 53 59  1] ...
 First target sequence: [41 43  8  0 32 46 53 59  1 61] ...

 Decoded input: nce.
 Thou wouldst be fee'd, I see, to make me sport:
 York cannot

 Decoded target: ce.
 Thou wouldst be fee'd, I see, to make me sport:
 York cannot 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.215218544006348
 [Seed 48] Step 0 --> val loss 4.1832709312438965
 Step 10: loss 3.8332571983337402
 Step 20: loss 3.3551077842712402
 Step 30: loss 2.959308624267578
 Step 40: loss 2.6769638061523438
 Step 50: loss 2.6292741298675537
 Step 60: loss 2.4756810665130615
 Step 70: loss 2.5100667476654053
 Step 80: loss 2.371307134628296
 Step 90: loss 2.282254695892334
 Step 100: loss 2.3172144889831543
 Step 110: loss 2.29682993888855
 Step 120: loss 2.2377405166625977
 Step 130: loss 2.1881113052368164
 Step 140: loss 2.1800942420959473
 Step 150: loss 2.2347311973571777
 Step 160: loss 2.102729320526123
 Step 170: loss 2.144104480743408
 Step 180: loss 2.1143202781677246
 Step 190: loss 2.179717540740967
 Step 200: loss 2.1706905364990234
 [Seed 48] Step 200 --> val loss 2.3157832622528076
 Step 210: loss 2.1436123847961426
 Step 220: loss 2.0769803524017334
 Step 230: loss 2.117417573928833
 Step 240: loss 2.0556650161743164
 Step 250: loss 2.2115485668182373
 Step 260: loss 2.1601226329803467
 Step 270: loss 2.0429277420043945
 Step 280: loss 2.29725980758667
 Step 290: loss 1.9592012166976929
 Step 300: loss 2.0818090438842773
 Step 310: loss 2.0018649101257324
 Step 320: loss 2.1307287216186523
 Step 330: loss 2.0672781467437744
 Step 340: loss 2.0708751678466797
 Step 350: loss 2.0389275550842285
 Step 360: loss 1.9880160093307495
 Step 370: loss 2.0343613624572754
 Step 380: loss 2.0543293952941895
 Step 390: loss 1.973562240600586
 Step 400: loss 1.9860637187957764
 [Seed 48] Step 400 --> val loss 2.322876214981079
 Step 410: loss 1.9488134384155273
 Step 420: loss 2.0090317726135254
 Step 430: loss 2.0264713764190674
 Step 440: loss 2.151914596557617
 Step 450: loss 2.01737380027771
 Step 460: loss 2.0722274780273438
 Step 470: loss 1.9890129566192627
 Step 480: loss 1.9884828329086304
 Step 490: loss 1.988692045211792
 Step 500: loss 1.9107472896575928
 Step 510: loss 1.9479182958602905
 Step 520: loss 2.019926071166992
 Step 530: loss 2.0675277709960938
 Step 540: loss 2.113868236541748
 Step 550: loss 1.820814847946167
 Step 560: loss 1.9791358709335327
 Step 570: loss 1.9653234481811523
 Step 580: loss 1.9427493810653687
 Step 590: loss 2.03055739402771
 Step 600: loss 2.0719447135925293
 [Seed 48] Step 600 --> val loss 2.258206605911255
 Step 610: loss 2.0549817085266113
 Step 620: loss 2.022845983505249
 Step 630: loss 2.000579357147217
 Step 640: loss 2.0397162437438965
 Step 650: loss 1.9557359218597412
 Step 660: loss 1.9989194869995117
 Step 670: loss 2.0720958709716797
 Step 680: loss 2.1035265922546387
 Step 690: loss 2.166508674621582
 Step 700: loss 2.0504751205444336
 Step 710: loss 1.981621503829956
 Step 720: loss 2.0108394622802734
 Step 730: loss 1.9653009176254272
 Step 740: loss 1.9937455654144287
 Step 750: loss 2.0184457302093506
 Step 760: loss 1.9824464321136475
 Step 770: loss 2.0354204177856445
 Step 780: loss 1.964747428894043
 Step 790: loss 2.03283953666687
 Step 800: loss 2.02948260307312
 [Seed 48] Step 800 --> val loss 2.2502543926239014
 Step 810: loss 2.0406665802001953
 Step 820: loss 2.0149240493774414
 Step 830: loss 2.050914764404297
 Step 840: loss 2.1507906913757324
 Step 850: loss 2.1110830307006836
 Step 860: loss 2.0082311630249023
 Step 870: loss 1.986810564994812
 Step 880: loss 2.0035715103149414
 Step 890: loss 2.0450799465179443
 Step 900: loss 1.9052245616912842
 Step 910: loss 2.0992157459259033
 Step 920: loss 2.222781181335449
 Step 930: loss 2.057206869125366
 Step 940: loss 2.001159429550171
 Step 950: loss 2.0105865001678467
 Step 960: loss 2.097019910812378
 Step 970: loss 2.0153748989105225
 Step 980: loss 2.0194220542907715
 Step 990: loss 2.056030035018921
 Step 1000: loss 2.179018497467041
 [Seed 48] Step 1000 --> val loss 2.2901973724365234
 Step 1010: loss 1.9924180507659912
 Step 1020: loss 2.086024045944214
 Step 1030: loss 2.0032882690429688
 Step 1040: loss 1.9974216222763062
 Step 1050: loss 2.1020097732543945
 Step 1060: loss 1.9780833721160889
 Step 1070: loss 2.0748839378356934
 Step 1080: loss 1.9510762691497803
 Step 1090: loss 2.095731019973755
 Step 1100: loss 2.0490686893463135
 Step 1110: loss 1.9082651138305664
 Step 1120: loss 2.0814807415008545
 Step 1130: loss 2.022909641265869
 Step 1140: loss 1.971165418624878
 Step 1150: loss 2.0059995651245117
 Step 1160: loss 2.0090389251708984
 Step 1170: loss 1.9401715993881226
 Step 1180: loss 2.0387775897979736
 Step 1190: loss 1.9737308025360107
 Step 1200: loss 2.151883125305176
 [Seed 48] Step 1200 --> val loss 2.3051795959472656
 Step 1210: loss 2.1099276542663574
 Step 1220: loss 2.043872594833374
 Step 1230: loss 2.061579465866089
 Step 1240: loss 2.038919448852539
 Step 1250: loss 2.0864672660827637
 Step 1260: loss 2.0234603881835938
 Step 1270: loss 2.0587639808654785
 Step 1280: loss 1.9749505519866943
 Step 1290: loss 1.8444337844848633
 Step 1300: loss 2.001331090927124
 Step 1310: loss 1.9736909866333008
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 48] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 48] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 48] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 48] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 30 53 57 39 50 47 52 43 10] ...
 First target sequence: [30 53 57 39 50 47 52 43 10  0] ...

 Decoded input:  Rosaline:
 And art thou changed? pronounce this sentence then,
 W

 Decoded target: Rosaline:
 And art thou changed? pronounce this sentence then,
 Wo
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.210145473480225
 [Seed 49] Step 0 --> val loss 4.184634208679199
 Step 10: loss 3.8516902923583984
 Step 20: loss 3.3245322704315186
 Step 30: loss 2.920928955078125
 Step 40: loss 2.628051280975342
 Step 50: loss 2.5037591457366943
 Step 60: loss 2.5580215454101562
 Step 70: loss 2.3362181186676025
 Step 80: loss 2.4405455589294434
 Step 90: loss 2.402357578277588
 Step 100: loss 2.3066413402557373
 Step 110: loss 2.2345986366271973
 Step 120: loss 2.1772122383117676
 Step 130: loss 2.2963664531707764
 Step 140: loss 2.1829731464385986
 Step 150: loss 2.2286860942840576
 Step 160: loss 2.2186317443847656
 Step 170: loss 2.133603572845459
 Step 180: loss 2.1949377059936523
 Step 190: loss 2.0998854637145996
 Step 200: loss 2.060699939727783
 [Seed 49] Step 200 --> val loss 2.2977304458618164
 Step 210: loss 2.063060998916626
 Step 220: loss 2.1120381355285645
 Step 230: loss 2.187602996826172
 Step 240: loss 2.1487393379211426
 Step 250: loss 1.9699602127075195
 Step 260: loss 2.1355159282684326
 Step 270: loss 2.0489697456359863
 Step 280: loss 2.005800724029541
 Step 290: loss 2.117180347442627
 Step 300: loss 1.9877980947494507
 Step 310: loss 2.0948400497436523
 Step 320: loss 2.1598262786865234
 Step 330: loss 2.080080509185791
 Step 340: loss 1.9379165172576904
 Step 350: loss 1.9479923248291016
 Step 360: loss 2.103652000427246
 Step 370: loss 1.9314184188842773
 Step 380: loss 1.9377797842025757
 Step 390: loss 1.9383232593536377
 Step 400: loss 2.0026004314422607
 [Seed 49] Step 400 --> val loss 2.1622238159179688
 Step 410: loss 1.924030065536499
 Step 420: loss 2.0004372596740723
 Step 430: loss 2.07163143157959
 Step 440: loss 2.0975170135498047
 Step 450: loss 1.9647393226623535
 Step 460: loss 2.0509822368621826
 Step 470: loss 1.9047174453735352
 Step 480: loss 1.9012154340744019
 Step 490: loss 2.0489940643310547
 Step 500: loss 2.0320184230804443
 Step 510: loss 1.9996986389160156
 Step 520: loss 1.9075124263763428
 Step 530: loss 2.009246349334717
 Step 540: loss 1.9794669151306152
 Step 550: loss 2.021993637084961
 Step 560: loss 2.126995086669922
 Step 570: loss 2.013338804244995
 Step 580: loss 2.014000415802002
 Step 590: loss 1.9785585403442383
 Step 600: loss 2.095754623413086
 [Seed 49] Step 600 --> val loss 2.332104206085205
 Step 610: loss 1.8963751792907715
 Step 620: loss 1.988766074180603
 Step 630: loss 2.0084481239318848
 Step 640: loss 2.011143684387207
 Step 650: loss 2.115971088409424
 Step 660: loss 2.0159449577331543
 Step 670: loss 1.8946502208709717
 Step 680: loss 2.002150058746338
 Step 690: loss 1.9590320587158203
 Step 700: loss 1.9505622386932373
 Step 710: loss 1.9873822927474976
 Step 720: loss 2.0500152111053467
 Step 730: loss 1.958540916442871
 Step 740: loss 1.9876512289047241
 Step 750: loss 2.022406816482544
 Step 760: loss 1.9809945821762085
 Step 770: loss 2.000047206878662
 Step 780: loss 2.0886502265930176
 Step 790: loss 1.9725608825683594
 Step 800: loss 2.001845359802246
 [Seed 49] Step 800 --> val loss 2.269943952560425
 Step 810: loss 2.0553441047668457
 Step 820: loss 1.9222252368927002
 Step 830: loss 2.027566909790039
 Step 840: loss 2.0404624938964844
 Step 850: loss 2.0127172470092773
 Step 860: loss 2.055629253387451
 Step 870: loss 2.071216583251953
 Step 880: loss 2.015803337097168
 Step 890: loss 1.9301016330718994
 Step 900: loss 2.021364688873291
 Step 910: loss 2.09074330329895
 Step 920: loss 2.1055121421813965
 Step 930: loss 1.9795554876327515
 Step 940: loss 1.9568426609039307
 Step 950: loss 1.9157721996307373
 Step 960: loss 1.9423878192901611
 Step 970: loss 1.9087742567062378
 Step 980: loss 1.8992303609848022
 Step 990: loss 1.9465655088424683
 Step 1000: loss 1.8799827098846436
 [Seed 49] Step 1000 --> val loss 2.238337516784668
 Step 1010: loss 2.0155043601989746
 Step 1020: loss 1.9356136322021484
 Step 1030: loss 1.8301711082458496
 Step 1040: loss 1.8013396263122559
 Step 1050: loss 1.9649808406829834
 Step 1060: loss 2.0068469047546387
 Step 1070: loss 1.8253949880599976
 Step 1080: loss 1.967053771018982
 Step 1090: loss 1.9430809020996094
 Step 1100: loss 2.0020751953125
 Step 1110: loss 1.96964430809021
 Step 1120: loss 2.052320957183838
 Step 1130: loss 1.9920833110809326
 Step 1140: loss 1.981947422027588
 Step 1150: loss 1.9466567039489746
 Step 1160: loss 2.083534002304077
 Step 1170: loss 2.0127015113830566
 Step 1180: loss 2.0387182235717773
 Step 1190: loss 1.9622077941894531
 Step 1200: loss 2.015838146209717
 [Seed 49] Step 1200 --> val loss 2.2062318325042725
 Step 1210: loss 1.8846263885498047
 Step 1220: loss 2.0196001529693604
 Step 1230: loss 1.9237430095672607
 Step 1240: loss 2.139138698577881
 Step 1250: loss 1.9487414360046387
 Step 1260: loss 2.009563446044922
 Step 1270: loss 1.8554770946502686
 Step 1280: loss 1.9584382772445679
 Step 1290: loss 1.9482759237289429
 Step 1300: loss 1.8567723035812378
 Step 1310: loss 1.924760341644287
 Step 1320: loss 1.7926597595214844
 Step 1330: loss 1.968881607055664
 Step 1340: loss 1.9133062362670898
 Step 1350: loss 1.8926198482513428
 Step 1360: loss 1.9878820180892944
 Step 1370: loss 1.9933007955551147
 Step 1380: loss 1.9212650060653687
 Step 1390: loss 1.8185310363769531
 Step 1400: loss 1.8803188800811768
 [Seed 49] Step 1400 --> val loss 2.0787651538848877
 Step 1410: loss 1.87174391746521
 Step 1420: loss 1.8524847030639648
 Step 1430: loss 1.9119322299957275
 Step 1440: loss 1.9291532039642334
 Step 1450: loss 1.7967939376831055
 Step 1460: loss 1.6915745735168457
 Step 1470: loss 1.9239685535430908
 Step 1480: loss 1.8656399250030518
 Step 1490: loss 1.7893741130828857
 Step 1500: loss 1.9289138317108154
 Step 1510: loss 1.9155206680297852
 Step 1520: loss 1.8040088415145874
 Step 1530: loss 1.9104069471359253
 Step 1540: loss 1.7211863994598389
 Step 1550: loss 1.7734750509262085
 Step 1560: loss 1.8144457340240479
 Step 1570: loss 1.8106367588043213
 Step 1580: loss 1.827222228050232
 Step 1590: loss 1.7975057363510132
 Step 1600: loss 1.8766177892684937
 [Seed 49] Step 1600 --> val loss 2.007200241088867
 Step 1610: loss 1.886183738708496
 Step 1620: loss 1.9927020072937012
 Step 1630: loss 1.8683340549468994
 Step 1640: loss 1.7406162023544312
 Step 1650: loss 1.824507713317871
 Step 1660: loss 1.7432469129562378
 Step 1670: loss 1.833047866821289
 Step 1680: loss 1.7616934776306152
 Step 1690: loss 1.7798761129379272
 Step 1700: loss 1.7965402603149414
 Step 1710: loss 1.9582388401031494
 Step 1720: loss 1.7560546398162842
 Step 1730: loss 1.7491382360458374
 Step 1740: loss 1.8122190237045288
 Step 1750: loss 1.7981979846954346
 Step 1760: loss 1.6489522457122803
 Step 1770: loss 1.7894048690795898
 Step 1780: loss 1.7886979579925537
 Step 1790: loss 1.814859390258789
 Step 1800: loss 1.681382656097412
 [Seed 49] Step 1800 --> val loss 1.908881425857544
 Step 1810: loss 1.7057769298553467
 Step 1820: loss 1.7716052532196045
 Step 1830: loss 1.7723478078842163
 Step 1840: loss 1.6865310668945312
 Step 1850: loss 1.7688322067260742
 Step 1860: loss 1.780837893486023
 Step 1870: loss 1.6968425512313843
 Step 1880: loss 1.610372543334961
 Step 1890: loss 1.6704163551330566
 Step 1900: loss 1.6366868019104004
 Step 1910: loss 1.6726913452148438
 Step 1920: loss 1.8052945137023926
 Step 1930: loss 1.6551353931427002
 Step 1940: loss 1.724584937095642
 Step 1950: loss 1.780067801475525
 Step 1960: loss 1.6832565069198608
 Step 1970: loss 1.7546966075897217
 Step 1980: loss 1.8337221145629883
 Step 1990: loss 1.7579644918441772
 Step 2000: loss 1.6908591985702515
 [Seed 49] Step 2000 --> val loss 1.841861367225647
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [54 39 47 52 58 47 52 45  6  1] ...
 First target sequence: [39 47 52 58 47 52 45  6  1 42] ...

 Decoded input: painting, do prove my occupation a mystery:
 but what mystery the

 Decoded target: ainting, do prove my occupation a mystery:
 but what mystery ther
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.216989040374756
 [Seed 50] Step 0 --> val loss 4.186938285827637
 Step 10: loss 3.857044219970703
 Step 20: loss 3.356189250946045
 Step 30: loss 3.002901554107666
 Step 40: loss 2.6064882278442383
 Step 50: loss 2.561964511871338
 Step 60: loss 2.4737486839294434
 Step 70: loss 2.5213000774383545
 Step 80: loss 2.3416833877563477
 Step 90: loss 2.4193856716156006
 Step 100: loss 2.3356008529663086
 Step 110: loss 2.3446662425994873
 Step 120: loss 2.2992238998413086
 Step 130: loss 2.3081979751586914
 Step 140: loss 2.2922446727752686
 Step 150: loss 2.155353546142578
 Step 160: loss 2.145148515701294
 Step 170: loss 2.2893850803375244
 Step 180: loss 2.0446043014526367
 Step 190: loss 2.0618948936462402
 Step 200: loss 2.1618099212646484
 [Seed 50] Step 200 --> val loss 2.3386483192443848
 Step 210: loss 2.0726876258850098
 Step 220: loss 2.143655776977539
 Step 230: loss 2.056438446044922
 Step 240: loss 2.111255645751953
 Step 250: loss 2.1051769256591797
 Step 260: loss 2.0630064010620117
 Step 270: loss 2.0356674194335938
 Step 280: loss 2.025287628173828
 Step 290: loss 2.1221024990081787
 Step 300: loss 2.1153459548950195
 Step 310: loss 2.129551887512207
 Step 320: loss 2.0012688636779785
 Step 330: loss 2.1463980674743652
 Step 340: loss 2.0175957679748535
 Step 350: loss 2.046260118484497
 Step 360: loss 2.038907051086426
 Step 370: loss 1.9927582740783691
 Step 380: loss 2.005591630935669
 Step 390: loss 1.9141277074813843
 Step 400: loss 2.077479362487793
 [Seed 50] Step 400 --> val loss 2.25453519821167
 Step 410: loss 2.113725185394287
 Step 420: loss 2.09717059135437
 Step 430: loss 1.983730673789978
 Step 440: loss 2.1493287086486816
 Step 450: loss 1.9336531162261963
 Step 460: loss 2.1523795127868652
 Step 470: loss 2.0491652488708496
 Step 480: loss 2.062547206878662
 Step 490: loss 1.9559876918792725
 Step 500: loss 1.8628828525543213
 Step 510: loss 1.977758765220642
 Step 520: loss 1.9212874174118042
 Step 530: loss 2.05796217918396
 Step 540: loss 1.9189114570617676
 Step 550: loss 1.914346694946289
 Step 560: loss 2.061613082885742
 Step 570: loss 1.9809716939926147
 Step 580: loss 2.0087287425994873
 Step 590: loss 1.916187047958374
 Step 600: loss 1.955190658569336
 [Seed 50] Step 600 --> val loss 2.1508991718292236
 Step 610: loss 1.882171630859375
 Step 620: loss 2.0885300636291504
 Step 630: loss 1.9044396877288818
 Step 640: loss 2.021963119506836
 Step 650: loss 1.9771852493286133
 Step 660: loss 1.900599718093872
 Step 670: loss 1.939678430557251
 Step 680: loss 1.9901984930038452
 Step 690: loss 1.8330304622650146
 Step 700: loss 1.941857099533081
 Step 710: loss 1.978929877281189
 Step 720: loss 2.0520544052124023
 Step 730: loss 1.8859237432479858
 Step 740: loss 2.0032079219818115
 Step 750: loss 1.9549295902252197
 Step 760: loss 1.9338792562484741
 Step 770: loss 1.936497449874878
 Step 780: loss 1.85349440574646
 Step 790: loss 2.0330922603607178
 Step 800: loss 2.004906415939331
 [Seed 50] Step 800 --> val loss 2.217151641845703
 Step 810: loss 1.9622113704681396
 Step 820: loss 2.0162887573242188
 Step 830: loss 2.0055150985717773
 Step 840: loss 1.8680598735809326
 Step 850: loss 1.9567651748657227
 Step 860: loss 1.9611700773239136
 Step 870: loss 2.007000684738159
 Step 880: loss 1.9828364849090576
 Step 890: loss 1.925437331199646
 Step 900: loss 1.9799835681915283
 Step 910: loss 2.0410547256469727
 Step 920: loss 2.011727809906006
 Step 930: loss 1.9850964546203613
 Step 940: loss 1.9939162731170654
 Step 950: loss 1.9742449522018433
 Step 960: loss 1.9434655904769897
 Step 970: loss 1.9436254501342773
 Step 980: loss 2.065551280975342
 Step 990: loss 1.8665668964385986
 Step 1000: loss 1.9066208600997925
 [Seed 50] Step 1000 --> val loss 2.172682523727417
 Step 1010: loss 1.9136207103729248
 Step 1020: loss 1.908257246017456
 Step 1030: loss 2.0489277839660645
 Step 1040: loss 1.9416015148162842
 Step 1050: loss 1.9072351455688477
 Step 1060: loss 1.9114879369735718
 Step 1070: loss 1.9640605449676514
 Step 1080: loss 1.960803747177124
 Step 1090: loss 1.9188933372497559
 Step 1100: loss 1.9127767086029053
 Step 1110: loss 2.0105252265930176
 Step 1120: loss 1.933013916015625
 Step 1130: loss 1.9222663640975952
 Step 1140: loss 1.8952178955078125
 Step 1150: loss 2.057182550430298
 Step 1160: loss 1.9333312511444092
 Step 1170: loss 1.8627170324325562
 Step 1180: loss 1.8076698780059814
 Step 1190: loss 1.9906809329986572
 Step 1200: loss 1.9283193349838257
 [Seed 50] Step 1200 --> val loss 2.2088565826416016
 Step 1210: loss 1.946739673614502
 Step 1220: loss 1.8203144073486328
 Step 1230: loss 1.7820943593978882
 Step 1240: loss 1.9320902824401855
 Step 1250: loss 1.8110275268554688
 Step 1260: loss 1.8039497137069702
 Step 1270: loss 1.8412902355194092
 Step 1280: loss 1.8422050476074219
 Step 1290: loss 1.8278648853302002
 Step 1300: loss 1.841933250427246
 Step 1310: loss 1.9372925758361816
 Step 1320: loss 1.8253792524337769
 Step 1330: loss 1.8792810440063477
 Step 1340: loss 1.788591742515564
 Step 1350: loss 1.788437843322754
 Step 1360: loss 1.7998580932617188
 Step 1370: loss 1.8511995077133179
 Step 1380: loss 1.8624255657196045
 Step 1390: loss 1.7497209310531616
 Step 1400: loss 1.8635835647583008
 [Seed 50] Step 1400 --> val loss 2.0894172191619873
 Step 1410: loss 1.7710702419281006
 Step 1420: loss 1.824172019958496
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 50] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 50] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 50] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [53 51  1 51 63  1 51 53 59 58] ...
 First target sequence: [51  1 51 63  1 51 53 59 58 46] ...

 Decoded input: om my mouth the wish of happy years:
 As gentle and as jocund as 

 Decoded target: m my mouth the wish of happy years:
 As gentle and as jocund as t
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.205662250518799
 [Seed 51] Step 0 --> val loss 4.181329250335693
 Step 10: loss 3.8453750610351562
 Step 20: loss 3.410186290740967
 Step 30: loss 2.95589017868042
 Step 40: loss 2.6095943450927734
 Step 50: loss 2.5806374549865723
 Step 60: loss 2.4600605964660645
 Step 70: loss 2.3673577308654785
 Step 80: loss 2.3541736602783203
 Step 90: loss 2.335005760192871
 Step 100: loss 2.2641587257385254
 Step 110: loss 2.271697759628296
 Step 120: loss 2.1732213497161865
 Step 130: loss 2.264775037765503
 Step 140: loss 2.2158775329589844
 Step 150: loss 2.139573574066162
 Step 160: loss 2.1471734046936035
 Step 170: loss 2.121152639389038
 Step 180: loss 2.185314655303955
 Step 190: loss 2.222580909729004
 Step 200: loss 2.066528081893921
 [Seed 51] Step 200 --> val loss 2.3020482063293457
 Step 210: loss 2.0284321308135986
 Step 220: loss 1.9994906187057495
 Step 230: loss 2.145862579345703
 Step 240: loss 2.047761917114258
 Step 250: loss 2.00100040435791
 Step 260: loss 2.120265483856201
 Step 270: loss 2.028170347213745
 Step 280: loss 2.1249449253082275
 Step 290: loss 1.911353349685669
 Step 300: loss 2.1043307781219482
 Step 310: loss 2.111691474914551
 Step 320: loss 2.006025791168213
 Step 330: loss 2.008202314376831
 Step 340: loss 1.9793643951416016
 Step 350: loss 2.0251948833465576
 Step 360: loss 2.0111207962036133
 Step 370: loss 2.089597463607788
 Step 380: loss 2.0747575759887695
 Step 390: loss 1.8268518447875977
 Step 400: loss 1.8672045469284058
 [Seed 51] Step 400 --> val loss 2.300936222076416
 Step 410: loss 1.9457964897155762
 Step 420: loss 1.981529951095581
 Step 430: loss 2.0413401126861572
 Step 440: loss 1.9724647998809814
 Step 450: loss 2.0443358421325684
 Step 460: loss 1.9495190382003784
 Step 470: loss 1.9872560501098633
 Step 480: loss 2.0064139366149902
 Step 490: loss 2.0505847930908203
 Step 500: loss 1.939934492111206
 Step 510: loss 2.0257060527801514
 Step 520: loss 1.965965747833252
 Step 530: loss 1.9173047542572021
 Step 540: loss 1.8364403247833252
 Step 550: loss 2.054443359375
 Step 560: loss 1.9553964138031006
 Step 570: loss 1.9874374866485596
 Step 580: loss 2.0095558166503906
 Step 590: loss 1.9277195930480957
 Step 600: loss 1.998915433883667
 [Seed 51] Step 600 --> val loss 2.2820005416870117
 Step 610: loss 2.0733628273010254
 Step 620: loss 2.057237148284912
 Step 630: loss 1.9786866903305054
 Step 640: loss 1.9808452129364014
 Step 650: loss 1.950103521347046
 Step 660: loss 2.0313892364501953
 Step 670: loss 2.052985191345215
 Step 680: loss 1.9022592306137085
 Step 690: loss 2.018803119659424
 Step 700: loss 1.9777276515960693
 Step 710: loss 1.9194854497909546
 Step 720: loss 2.117187261581421
 Step 730: loss 2.1656370162963867
 Step 740: loss 1.9100291728973389
 Step 750: loss 1.9939897060394287
 Step 760: loss 1.9473400115966797
 Step 770: loss 2.0337533950805664
 Step 780: loss 1.8788942098617554
 Step 790: loss 1.8731015920639038
 Step 800: loss 1.9822766780853271
 [Seed 51] Step 800 --> val loss 2.1755549907684326
 Step 810: loss 1.9637022018432617
 Step 820: loss 1.8338680267333984
 Step 830: loss 1.9327396154403687
 Step 840: loss 1.9156866073608398
 Step 850: loss 2.0270490646362305
 Step 860: loss 2.0582075119018555
 Step 870: loss 1.936309576034546
 Step 880: loss 1.9242843389511108
 Step 890: loss 1.9251527786254883
 Step 900: loss 1.9477189779281616
 Step 910: loss 2.0019354820251465
 Step 920: loss 2.0085091590881348
 Step 930: loss 1.946429967880249
 Step 940: loss 1.9574933052062988
 Step 950: loss 1.9634647369384766
 Step 960: loss 1.896289587020874
 Step 970: loss 1.8842815160751343
 Step 980: loss 1.9581823348999023
 Step 990: loss 1.8431466817855835
 Step 1000: loss 1.844337821006775
 [Seed 51] Step 1000 --> val loss 2.1471917629241943
 Step 1010: loss 1.9163568019866943
 Step 1020: loss 1.970188856124878
 Step 1030: loss 2.04345965385437
 Step 1040: loss 1.9426101446151733
 Step 1050: loss 1.8461798429489136
 Step 1060: loss 1.9940400123596191
 Step 1070: loss 2.0068891048431396
 Step 1080: loss 1.8646934032440186
 Step 1090: loss 1.903527021408081
 Step 1100: loss 2.078352928161621
 Step 1110: loss 2.0607385635375977
 Step 1120: loss 1.9793119430541992
 Step 1130: loss 1.918501853942871
 Step 1140: loss 1.831473708152771
 Step 1150: loss 1.8602464199066162
 Step 1160: loss 1.9265210628509521
 Step 1170: loss 1.8861265182495117
 Step 1180: loss 1.9205793142318726
 Step 1190: loss 1.9263947010040283
 Step 1200: loss 1.9222419261932373
 [Seed 51] Step 1200 --> val loss 2.165976047515869
 Step 1210: loss 2.0795390605926514
 Step 1220: loss 2.1159303188323975
 Step 1230: loss 2.012232780456543
 Step 1240: loss 2.0267789363861084
 Step 1250: loss 1.9610623121261597
 Step 1260: loss 1.973801612854004
 Step 1270: loss 1.9224722385406494
 Step 1280: loss 1.9269769191741943
 Step 1290: loss 1.9854729175567627
 Step 1300: loss 2.031522750854492
 Step 1310: loss 1.8210047483444214
 Step 1320: loss 2.0026822090148926
 Step 1330: loss 2.026254415512085
 Step 1340: loss 1.769941806793213
 Step 1350: loss 1.8131892681121826
 Step 1360: loss 1.9501292705535889
 Step 1370: loss 1.9790608882904053
 Step 1380: loss 1.7782294750213623
 Step 1390: loss 2.0398130416870117
 Step 1400: loss 1.8183839321136475
 [Seed 51] Step 1400 --> val loss 2.1278328895568848
 Step 1410: loss 1.9398173093795776
 Step 1420: loss 1.869213342666626
 Step 1430: loss 1.8952655792236328
 Step 1440: loss 1.940775752067566
 Step 1450: loss 1.74210786819458
 Step 1460: loss 1.834441900253296
 Step 1470: loss 1.9010941982269287
 Step 1480: loss 1.8117384910583496
 Step 1490: loss 1.7877178192138672
 Step 1500: loss 1.8758586645126343
 Step 1510: loss 1.7895349264144897
 Step 1520: loss 1.9083597660064697
 Step 1530: loss 1.84981369972229
 Step 1540: loss 2.0040388107299805
 Step 1550: loss 1.865411639213562
 Step 1560: loss 1.8763622045516968
 Step 1570: loss 1.7666337490081787
 Step 1580: loss 1.9171078205108643
 Step 1590: loss 1.7195262908935547
 Step 1600: loss 1.6817048788070679
 [Seed 51] Step 1600 --> val loss 1.9888337850570679
 Step 1610: loss 1.7160208225250244
 Step 1620: loss 1.7400703430175781
 Step 1630: loss 1.680978536605835
 Step 1640: loss 1.783470869064331
 Step 1650: loss 1.754654884338379
 Step 1660: loss 1.7117891311645508
 Step 1670: loss 1.670940637588501
 Step 1680: loss 1.798198938369751
 Step 1690: loss 1.8053991794586182
 Step 1700: loss 1.9115670919418335
 Step 1710: loss 1.7755632400512695
 Step 1720: loss 1.7450177669525146
 Step 1730: loss 1.9164724349975586
 Step 1740: loss 1.754328727722168
 Step 1750: loss 1.7672395706176758
 Step 1760: loss 1.7499911785125732
 Step 1770: loss 1.7532498836517334
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 51] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 51] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 46 43 39 56 58 57  1 53 44] ...
 First target sequence: [46 43 39 56 58 57  1 53 44  1] ...

 Decoded input:  hearts of sorrow and your eyes of tears:
 Come home with me to s

 Decoded target: hearts of sorrow and your eyes of tears:
 Come home with me to su
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2081522941589355
 [Seed 52] Step 0 --> val loss 4.184819221496582
 Step 10: loss 3.84637451171875
 Step 20: loss 3.4035260677337646
 Step 30: loss 2.896317958831787
 Step 40: loss 2.682640790939331
 Step 50: loss 2.5758719444274902
 Step 60: loss 2.4647133350372314
 Step 70: loss 2.4997448921203613
 Step 80: loss 2.336277484893799
 Step 90: loss 2.391747236251831
 Step 100: loss 2.3895959854125977
 Step 110: loss 2.2620599269866943
 Step 120: loss 2.240445852279663
 Step 130: loss 2.1880311965942383
 Step 140: loss 2.178941011428833
 Step 150: loss 2.217566967010498
 Step 160: loss 2.274580955505371
 Step 170: loss 2.1405630111694336
 Step 180: loss 2.2527856826782227
 Step 190: loss 2.110858678817749
 Step 200: loss 2.0967297554016113
 [Seed 52] Step 200 --> val loss 2.3313992023468018
 Step 210: loss 2.03717303276062
 Step 220: loss 2.0369794368743896
 Step 230: loss 2.036515235900879
 Step 240: loss 2.054649829864502
 Step 250: loss 2.005659341812134
 Step 260: loss 2.0655829906463623
 Step 270: loss 2.0352296829223633
 Step 280: loss 2.0141241550445557
 Step 290: loss 2.0552239418029785
 Step 300: loss 2.0840864181518555
 Step 310: loss 2.067636251449585
 Step 320: loss 2.0333964824676514
 Step 330: loss 2.104246139526367
 Step 340: loss 2.1406664848327637
 Step 350: loss 1.9909255504608154
 Step 360: loss 2.0437960624694824
 Step 370: loss 2.0532665252685547
 Step 380: loss 1.990365982055664
 Step 390: loss 1.9712566137313843
 Step 400: loss 2.0113954544067383
 [Seed 52] Step 400 --> val loss 2.249138593673706
 Step 410: loss 2.0043373107910156
 Step 420: loss 1.9769384860992432
 Step 430: loss 1.9384946823120117
 Step 440: loss 2.0096001625061035
 Step 450: loss 1.9345935583114624
 Step 460: loss 2.0492424964904785
 Step 470: loss 2.0481297969818115
 Step 480: loss 1.921029806137085
 Step 490: loss 2.0240955352783203
 Step 500: loss 2.042661666870117
 Step 510: loss 2.0282483100891113
 Step 520: loss 2.042243480682373
 Step 530: loss 2.084019660949707
 Step 540: loss 1.9976372718811035
 Step 550: loss 2.1122710704803467
 Step 560: loss 1.9887584447860718
 Step 570: loss 2.0444183349609375
 Step 580: loss 2.006869077682495
 Step 590: loss 2.017601728439331
 Step 600: loss 1.9771760702133179
 [Seed 52] Step 600 --> val loss 2.384734630584717
 Step 610: loss 2.0536036491394043
 Step 620: loss 2.0496010780334473
 Step 630: loss 2.0163345336914062
 Step 640: loss 2.0354080200195312
 Step 650: loss 1.971122145652771
 Step 660: loss 1.9681833982467651
 Step 670: loss 2.1334760189056396
 Step 680: loss 1.843654990196228
 Step 690: loss 2.0278480052948
 Step 700: loss 2.0523197650909424
 Step 710: loss 1.9502885341644287
 Step 720: loss 1.9852039813995361
 Step 730: loss 1.993173599243164
 Step 740: loss 1.9944095611572266
 Step 750: loss 1.9964711666107178
 Step 760: loss 1.9508589506149292
 Step 770: loss 1.9701942205429077
 Step 780: loss 1.9534391164779663
 Step 790: loss 1.959383249282837
 Step 800: loss 2.095120906829834
 [Seed 52] Step 800 --> val loss 2.3616387844085693
 Step 810: loss 1.9925620555877686
 Step 820: loss 1.950005292892456
 Step 830: loss 2.049038887023926
 Step 840: loss 2.1343677043914795
 Step 850: loss 1.9668267965316772
 Step 860: loss 1.9005331993103027
 Step 870: loss 2.0037031173706055
 Step 880: loss 2.1623847484588623
 Step 890: loss 2.0024631023406982
 Step 900: loss 1.9870548248291016
 Step 910: loss 2.0418801307678223
 Step 920: loss 2.022348403930664
 Step 930: loss 2.058534860610962
 Step 940: loss 2.089524984359741
 Step 950: loss 1.9459346532821655
 Step 960: loss 2.0377554893493652
 Step 970: loss 1.9610437154769897
 Step 980: loss 1.974752426147461
 Step 990: loss 2.0304622650146484
 Step 1000: loss 1.888138771057129
 [Seed 52] Step 1000 --> val loss 2.219510793685913
 Step 1010: loss 2.0382461547851562
 Step 1020: loss 2.009032964706421
 Step 1030: loss 1.8980128765106201
 Step 1040: loss 1.9525425434112549
 Step 1050: loss 1.895873785018921
 Step 1060: loss 1.9357751607894897
 Step 1070: loss 1.952587366104126
 Step 1080: loss 1.9681825637817383
 Step 1090: loss 2.0713229179382324
 Step 1100: loss 1.949979543685913
 Step 1110: loss 1.94814133644104
 Step 1120: loss 1.9213721752166748
 Step 1130: loss 1.792630910873413
 Step 1140: loss 1.963173270225525
 Step 1150: loss 2.0036072731018066
 Step 1160: loss 1.9421952962875366
 Step 1170: loss 2.0154690742492676
 Step 1180: loss 1.8685630559921265
 Step 1190: loss 1.9352020025253296
 Step 1200: loss 1.86465585231781
 [Seed 52] Step 1200 --> val loss 2.1768834590911865
 Step 1210: loss 1.9918382167816162
 Step 1220: loss 1.7935994863510132
 Step 1230: loss 1.8944330215454102
 Step 1240: loss 1.8098094463348389
 Step 1250: loss 2.02068829536438
 Step 1260: loss 1.8375422954559326
 Step 1270: loss 1.8588513135910034
 Step 1280: loss 1.9519656896591187
 Step 1290: loss 1.9162172079086304
 Step 1300: loss 1.9009578227996826
 Step 1310: loss 1.8264459371566772
 Step 1320: loss 1.8727312088012695
 Step 1330: loss 1.9193909168243408
 Step 1340: loss 1.8896172046661377
 Step 1350: loss 1.9099820852279663
 Step 1360: loss 1.9366668462753296
 Step 1370: loss 1.89009690284729
 Step 1380: loss 1.9445035457611084
 Step 1390: loss 1.9022331237792969
 Step 1400: loss 1.9073724746704102
 [Seed 52] Step 1400 --> val loss 2.0935218334198
 Step 1410: loss 1.820908546447754
 Step 1420: loss 1.9379733800888062
 Step 1430: loss 1.9351356029510498
 Step 1440: loss 2.057404041290283
 Step 1450: loss 1.9272512197494507
 Step 1460: loss 1.850280523300171
 Step 1470: loss 1.9012712240219116
 Step 1480: loss 1.8513269424438477
 Step 1490: loss 1.8075798749923706
 Step 1500: loss 1.8498179912567139
 Step 1510: loss 1.919438362121582
 Step 1520: loss 1.8182554244995117
 Step 1530: loss 1.9023675918579102
 Step 1540: loss 1.9138753414154053
 Step 1550: loss 1.821602702140808
 Step 1560: loss 1.8099188804626465
 Step 1570: loss 1.8731286525726318
 Step 1580: loss 1.8554795980453491
 Step 1590: loss 1.8327088356018066
 Step 1600: loss 1.8156001567840576
 [Seed 52] Step 1600 --> val loss 2.0102698802948
 Step 1610: loss 1.7633049488067627
 Step 1620: loss 1.8302371501922607
 Step 1630: loss 1.762315273284912
 Step 1640: loss 1.740776538848877
 Step 1650: loss 1.7926251888275146
 Step 1660: loss 1.707646131515503
 Step 1670: loss 1.887110710144043
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 52] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 52] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [43  2  0 16 43 39 58 46  1 39] ...
 First target sequence: [ 2  0 16 43 39 58 46  1 39 52] ...

 Decoded input: e!
 Death and destruction dog thee at the heels;
 Thy mother's nam

 Decoded target: !
 Death and destruction dog thee at the heels;
 Thy mother's name
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.197164058685303
 [Seed 53] Step 0 --> val loss 4.182758331298828
 Step 10: loss 3.833538770675659
 Step 20: loss 3.356627941131592
 Step 30: loss 2.8236892223358154
 Step 40: loss 2.7589569091796875
 Step 50: loss 2.6390483379364014
 Step 60: loss 2.536550998687744
 Step 70: loss 2.492971897125244
 Step 80: loss 2.3440635204315186
 Step 90: loss 2.2728352546691895
 Step 100: loss 2.3230037689208984
 Step 110: loss 2.266845703125
 Step 120: loss 2.1748147010803223
 Step 130: loss 2.1734976768493652
 Step 140: loss 2.101102352142334
 Step 150: loss 2.080885410308838
 Step 160: loss 2.1943435668945312
 Step 170: loss 2.279364585876465
 Step 180: loss 2.095913887023926
 Step 190: loss 2.014892101287842
 Step 200: loss 2.0559513568878174
 [Seed 53] Step 200 --> val loss 2.3115437030792236
 Step 210: loss 2.0482378005981445
 Step 220: loss 2.1876368522644043
 Step 230: loss 2.0451629161834717
 Step 240: loss 2.064755916595459
 Step 250: loss 2.042607307434082
 Step 260: loss 2.1483943462371826
 Step 270: loss 2.0243146419525146
 Step 280: loss 2.049586057662964
 Step 290: loss 1.9996719360351562
 Step 300: loss 2.0058536529541016
 Step 310: loss 2.0355796813964844
 Step 320: loss 1.993581771850586
 Step 330: loss 2.098843812942505
 Step 340: loss 2.0251641273498535
 Step 350: loss 1.9889497756958008
 Step 360: loss 1.9230964183807373
 Step 370: loss 2.0600638389587402
 Step 380: loss 1.9266002178192139
 Step 390: loss 2.0162551403045654
 Step 400: loss 1.9988882541656494
 [Seed 53] Step 400 --> val loss 2.282153844833374
 Step 410: loss 1.9551633596420288
 Step 420: loss 2.0698037147521973
 Step 430: loss 2.0550222396850586
 Step 440: loss 1.9495929479599
 Step 450: loss 1.9927160739898682
 Step 460: loss 2.041313409805298
 Step 470: loss 1.9767489433288574
 Step 480: loss 1.9990496635437012
 Step 490: loss 2.0503621101379395
 Step 500: loss 1.972816824913025
 Step 510: loss 2.062849998474121
 Step 520: loss 1.8487138748168945
 Step 530: loss 1.9720830917358398
 Step 540: loss 1.950269103050232
 Step 550: loss 1.9659504890441895
 Step 560: loss 2.0262792110443115
 Step 570: loss 2.012691020965576
 Step 580: loss 2.1339592933654785
 Step 590: loss 2.0260961055755615
 Step 600: loss 2.086263656616211
 [Seed 53] Step 600 --> val loss 2.2080373764038086
 Step 610: loss 1.949092149734497
 Step 620: loss 1.9951636791229248
 Step 630: loss 1.9248183965682983
 Step 640: loss 1.8905375003814697
 Step 650: loss 2.0213065147399902
 Step 660: loss 2.081707000732422
 Step 670: loss 2.0673022270202637
 Step 680: loss 1.9700430631637573
 Step 690: loss 1.9890860319137573
 Step 700: loss 2.0131301879882812
 Step 710: loss 2.062678337097168
 Step 720: loss 2.0743298530578613
 Step 730: loss 2.0521492958068848
 Step 740: loss 2.038882255554199
 Step 750: loss 1.9783952236175537
 Step 760: loss 2.017195224761963
 Step 770: loss 1.9714174270629883
 Step 780: loss 2.106215000152588
 Step 790: loss 2.1771953105926514
 Step 800: loss 2.0754330158233643
 [Seed 53] Step 800 --> val loss 2.290121078491211
 Step 810: loss 2.0727081298828125
 Step 820: loss 2.052535057067871
 Step 830: loss 2.136873960494995
 Step 840: loss 2.0628273487091064
 Step 850: loss 2.033019781112671
 Step 860: loss 2.158926010131836
 Step 870: loss 2.042754650115967
 Step 880: loss 1.9962928295135498
 Step 890: loss 1.9211578369140625
 Step 900: loss 2.0996007919311523
 Step 910: loss 2.085020065307617
 Step 920: loss 2.037614583969116
 Step 930: loss 2.0689010620117188
 Step 940: loss 2.102869749069214
 Step 950: loss 1.9516938924789429
 Step 960: loss 2.067805290222168
 Step 970: loss 1.9953316450119019
 Step 980: loss 2.0696449279785156
 Step 990: loss 2.087221145629883
 Step 1000: loss 1.9394700527191162
 [Seed 53] Step 1000 --> val loss 2.2344682216644287
 Step 1010: loss 2.002458333969116
 Step 1020: loss 2.0782856941223145
 Step 1030: loss 2.0625131130218506
 Step 1040: loss 2.1782100200653076
 Step 1050: loss 2.005582332611084
 Step 1060: loss 2.118614673614502
 Step 1070: loss 2.0221829414367676
 Step 1080: loss 2.0558557510375977
 Step 1090: loss 2.1302552223205566
 Step 1100: loss 2.008507251739502
 Step 1110: loss 2.0869128704071045
 Step 1120: loss 2.0364418029785156
 Step 1130: loss 1.990370512008667
 Step 1140: loss 2.0579586029052734
 Step 1150: loss 1.9485286474227905
 Step 1160: loss 2.0804319381713867
 Step 1170: loss 1.9980733394622803
 Step 1180: loss 2.018880605697632
 Step 1190: loss 1.951934814453125
 Step 1200: loss 1.9906257390975952
 [Seed 53] Step 1200 --> val loss 2.2236392498016357
 Step 1210: loss 1.975743055343628
 Step 1220: loss 2.0687918663024902
 Step 1230: loss 1.991154670715332
 Step 1240: loss 2.150592803955078
 Step 1250: loss 2.1070058345794678
 Step 1260: loss 2.1077208518981934
 Step 1270: loss 1.9622557163238525
 Step 1280: loss 2.081852436065674
 Step 1290: loss 2.144397735595703
 Step 1300: loss 1.9216740131378174
 Step 1310: loss 2.1118576526641846
 Step 1320: loss 2.0841245651245117
 Step 1330: loss 2.0320510864257812
 Step 1340: loss 2.0198450088500977
 Step 1350: loss 1.958391547203064
 Step 1360: loss 1.9514214992523193
 Step 1370: loss 1.9860581159591675
 Step 1380: loss 2.002863883972168
 Step 1390: loss 1.9716352224349976
 Step 1400: loss 1.998112440109253
 [Seed 53] Step 1400 --> val loss 2.1440351009368896
 Step 1410: loss 1.9051741361618042
 Step 1420: loss 1.9042022228240967
 Step 1430: loss 1.9932483434677124
 Step 1440: loss 2.007216453552246
 Step 1450: loss 1.7525008916854858
 Step 1460: loss 1.9172462224960327
 Step 1470: loss 1.9395396709442139
 Step 1480: loss 1.9611141681671143
 Step 1490: loss 2.0388917922973633
 Step 1500: loss 1.955166220664978
 Step 1510: loss 1.8978537321090698
 Step 1520: loss 1.7895350456237793
 Step 1530: loss 1.94384765625
 Step 1540: loss 1.8790569305419922
 Step 1550: loss 1.8511872291564941
 Step 1560: loss 1.881394863128662
 Step 1570: loss 1.8856561183929443
 Step 1580: loss 1.9004429578781128
 Step 1590: loss 1.9050703048706055
 Step 1600: loss 1.9301025867462158
 [Seed 53] Step 1600 --> val loss 2.082455635070801
 Step 1610: loss 1.8378591537475586
 Step 1620: loss 2.006312370300293
 Step 1630: loss 2.0078234672546387
 Step 1640: loss 1.8190139532089233
 Step 1650: loss 1.823750376701355
 Step 1660: loss 1.853595495223999
 Step 1670: loss 1.8677324056625366
 Step 1680: loss 1.7744026184082031
 Step 1690: loss 1.963031530380249
 Step 1700: loss 1.7613223791122437
 Step 1710: loss 1.885270118713379
 Step 1720: loss 1.84086012840271
 Step 1730: loss 1.8784425258636475
 Step 1740: loss 1.8268582820892334
 Step 1750: loss 1.8339570760726929
 Step 1760: loss 1.7907260656356812
 Step 1770: loss 1.8277853727340698
 Step 1780: loss 1.8931875228881836
 Step 1790: loss 1.7917150259017944
 Step 1800: loss 1.8210837841033936
 [Seed 53] Step 1800 --> val loss 1.9106365442276
 Step 1810: loss 1.7870957851409912
 Step 1820: loss 1.9192527532577515
 Step 1830: loss 1.8038482666015625
 Step 1840: loss 1.91233229637146
 Step 1850: loss 1.7445253133773804
 Step 1860: loss 1.8645579814910889
 Step 1870: loss 1.7316348552703857
 Step 1880: loss 1.9404910802841187
 Step 1890: loss 1.715105414390564
 Step 1900: loss 1.8534797430038452
 Step 1910: loss 1.7718862295150757
 Step 1920: loss 1.652251124382019
 Step 1930: loss 1.6623601913452148
 Step 1940: loss 1.7585052251815796
 Step 1950: loss 1.7378143072128296
 Step 1960: loss 1.8532460927963257
 Step 1970: loss 1.753793716430664
 Step 1980: loss 1.7661117315292358
 Step 1990: loss 1.8662350177764893
 Step 2000: loss 1.6971895694732666
 [Seed 53] Step 2000 --> val loss 1.8942348957061768
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [41 43  1 58 46 39 58  1 21  1] ...
 First target sequence: [43  1 58 46 39 58  1 21  1 51] ...

 Decoded input: ce that I myself
 Find in myself no pity to myself?
 Methought the

 Decoded target: e that I myself
 Find in myself no pity to myself?
 Methought the 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.198422908782959
 [Seed 54] Step 0 --> val loss 4.183627605438232
 Step 10: loss 3.864470958709717
 Step 20: loss 3.38285493850708
 Step 30: loss 3.0047099590301514
 Step 40: loss 2.6616196632385254
 Step 50: loss 2.627671003341675
 Step 60: loss 2.506256103515625
 Step 70: loss 2.412795066833496
 Step 80: loss 2.3575668334960938
 Step 90: loss 2.33433198928833
 Step 100: loss 2.4061970710754395
 Step 110: loss 2.207747459411621
 Step 120: loss 2.20349383354187
 Step 130: loss 2.1337382793426514
 Step 140: loss 2.291430950164795
 Step 150: loss 2.343209743499756
 Step 160: loss 2.1696577072143555
 Step 170: loss 2.1218090057373047
 Step 180: loss 2.1213395595550537
 Step 190: loss 2.097069263458252
 Step 200: loss 2.1107757091522217
 [Seed 54] Step 200 --> val loss 2.3555636405944824
 Step 210: loss 2.088230609893799
 Step 220: loss 2.0581555366516113
 Step 230: loss 2.1072511672973633
 Step 240: loss 1.8950103521347046
 Step 250: loss 2.0960259437561035
 Step 260: loss 2.0837650299072266
 Step 270: loss 2.1218879222869873
 Step 280: loss 2.005850315093994
 Step 290: loss 2.0823516845703125
 Step 300: loss 1.954602599143982
 Step 310: loss 1.968041181564331
 Step 320: loss 2.035364866256714
 Step 330: loss 2.044153928756714
 Step 340: loss 1.9908406734466553
 Step 350: loss 2.0016372203826904
 Step 360: loss 2.049664258956909
 Step 370: loss 2.0338940620422363
 Step 380: loss 2.0302836894989014
 Step 390: loss 1.9511241912841797
 Step 400: loss 1.9926503896713257
 [Seed 54] Step 400 --> val loss 2.2601139545440674
 Step 410: loss 2.0269534587860107
 Step 420: loss 2.0936732292175293
 Step 430: loss 2.0570526123046875
 Step 440: loss 2.012734889984131
 Step 450: loss 2.099255084991455
 Step 460: loss 2.033134937286377
 Step 470: loss 2.0053000450134277
 Step 480: loss 2.010488271713257
 Step 490: loss 2.062631845474243
 Step 500: loss 2.0362110137939453
 Step 510: loss 2.09159517288208
 Step 520: loss 2.064533233642578
 Step 530: loss 1.9991391897201538
 Step 540: loss 2.1217517852783203
 Step 550: loss 2.00235652923584
 Step 560: loss 2.0032730102539062
 Step 570: loss 2.0751397609710693
 Step 580: loss 2.002427577972412
 Step 590: loss 2.035431385040283
 Step 600: loss 2.0263237953186035
 [Seed 54] Step 600 --> val loss 2.2896480560302734
 Step 610: loss 2.127229690551758
 Step 620: loss 1.8985023498535156
 Step 630: loss 2.04764461517334
 Step 640: loss 2.0592126846313477
 Step 650: loss 2.045757532119751
 Step 660: loss 2.0043201446533203
 Step 670: loss 2.0283989906311035
 Step 680: loss 1.957754373550415
 Step 690: loss 1.944993019104004
 Step 700: loss 2.0287623405456543
 Step 710: loss 2.0682904720306396
 Step 720: loss 2.1673152446746826
 Step 730: loss 2.0559329986572266
 Step 740: loss 1.9689044952392578
 Step 750: loss 1.9614471197128296
 Step 760: loss 1.9605430364608765
 Step 770: loss 1.9954519271850586
 Step 780: loss 2.0598161220550537
 Step 790: loss 1.9350800514221191
 Step 800: loss 1.962737798690796
 [Seed 54] Step 800 --> val loss 2.2820403575897217
 Step 810: loss 2.052170515060425
 Step 820: loss 2.0820212364196777
 Step 830: loss 2.0286402702331543
 Step 840: loss 1.9122505187988281
 Step 850: loss 2.1032814979553223
 Step 860: loss 1.905824065208435
 Step 870: loss 1.9553401470184326
 Step 880: loss 1.9939298629760742
 Step 890: loss 1.858039140701294
 Step 900: loss 2.0771121978759766
 Step 910: loss 2.0560920238494873
 Step 920: loss 1.9409725666046143
 Step 930: loss 1.972791075706482
 Step 940: loss 1.926920771598816
 Step 950: loss 1.8449938297271729
 Step 960: loss 2.0657858848571777
 Step 970: loss 1.9029498100280762
 Step 980: loss 1.9921209812164307
 Step 990: loss 1.9854249954223633
 Step 1000: loss 1.9277681112289429
 [Seed 54] Step 1000 --> val loss 2.200758218765259
 Step 1010: loss 1.956571102142334
 Step 1020: loss 1.8884862661361694
 Step 1030: loss 1.9762605428695679
 Step 1040: loss 1.988065481185913
 Step 1050: loss 1.9752485752105713
 Step 1060: loss 1.954403281211853
 Step 1070: loss 1.9608581066131592
 Step 1080: loss 1.9937427043914795
 Step 1090: loss 2.071732997894287
 Step 1100: loss 1.9080890417099
 Step 1110: loss 1.9351692199707031
 Step 1120: loss 1.948586344718933
 Step 1130: loss 1.9793059825897217
 Step 1140: loss 1.915144443511963
 Step 1150: loss 2.0485281944274902
 Step 1160: loss 1.9008710384368896
 Step 1170: loss 1.945744514465332
 Step 1180: loss 1.9309642314910889
 Step 1190: loss 1.909428358078003
 Step 1200: loss 1.9015908241271973
 [Seed 54] Step 1200 --> val loss 2.0601155757904053
 Step 1210: loss 1.8371987342834473
 Step 1220: loss 1.872201681137085
 Step 1230: loss 1.9324228763580322
 Step 1240: loss 1.9115169048309326
 Step 1250: loss 1.99901282787323
 Step 1260: loss 1.7908058166503906
 Step 1270: loss 1.9711021184921265
 Step 1280: loss 2.003248929977417
 Step 1290: loss 1.7758872509002686
 Step 1300: loss 1.92072331905365
 Step 1310: loss 1.9359797239303589
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 54] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 54] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 54] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 54] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [39 57 57  1 47 57  1 52 53  1] ...
 First target sequence: [57 57  1 47 57  1 52 53  1 40] ...

 Decoded input: ass is no bigger than thy head;
 And yet, incaged in so small a v

 Decoded target: ss is no bigger than thy head;
 And yet, incaged in so small a ve
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.214068412780762
 [Seed 55] Step 0 --> val loss 4.184071063995361
 Step 10: loss 3.835714340209961
 Step 20: loss 3.3208694458007812
 Step 30: loss 2.8281524181365967
 Step 40: loss 2.713836193084717
 Step 50: loss 2.5335068702697754
 Step 60: loss 2.4546632766723633
 Step 70: loss 2.3705685138702393
 Step 80: loss 2.3906497955322266
 Step 90: loss 2.342785596847534
 Step 100: loss 2.284160614013672
 Step 110: loss 2.19730281829834
 Step 120: loss 2.24118709564209
 Step 130: loss 2.249523639678955
 Step 140: loss 2.2548141479492188
 Step 150: loss 2.177161455154419
 Step 160: loss 2.113821029663086
 Step 170: loss 2.1717965602874756
 Step 180: loss 2.0739498138427734
 Step 190: loss 2.129024028778076
 Step 200: loss 2.2296037673950195
 [Seed 55] Step 200 --> val loss 2.280733823776245
 Step 210: loss 2.1440539360046387
 Step 220: loss 2.0863404273986816
 Step 230: loss 2.131446599960327
 Step 240: loss 2.111945390701294
 Step 250: loss 2.0484180450439453
 Step 260: loss 2.0790328979492188
 Step 270: loss 2.1156394481658936
 Step 280: loss 2.1268186569213867
 Step 290: loss 2.079218864440918
 Step 300: loss 2.0429415702819824
 Step 310: loss 2.1156814098358154
 Step 320: loss 2.083178997039795
 Step 330: loss 1.9351201057434082
 Step 340: loss 2.0081021785736084
 Step 350: loss 2.1121103763580322
 Step 360: loss 2.1840734481811523
 Step 370: loss 2.178563117980957
 Step 380: loss 2.1200199127197266
 Step 390: loss 2.059504747390747
 Step 400: loss 2.078427314758301
 [Seed 55] Step 400 --> val loss 2.3538122177124023
 Step 410: loss 2.1440954208374023
 Step 420: loss 2.0840134620666504
 Step 430: loss 2.010348081588745
 Step 440: loss 1.9727051258087158
 Step 450: loss 1.9515883922576904
 Step 460: loss 1.9870442152023315
 Step 470: loss 1.9806307554244995
 Step 480: loss 2.04325008392334
 Step 490: loss 2.0027737617492676
 Step 500: loss 2.0505571365356445
 Step 510: loss 2.0284218788146973
 Step 520: loss 2.0115816593170166
 Step 530: loss 1.9796751737594604
 Step 540: loss 2.020172119140625
 Step 550: loss 1.92411208152771
 Step 560: loss 1.9853671789169312
 Step 570: loss 2.0490620136260986
 Step 580: loss 1.9928783178329468
 Step 590: loss 1.9513554573059082
 Step 600: loss 2.056198835372925
 [Seed 55] Step 600 --> val loss 2.3442559242248535
 Step 610: loss 2.0991110801696777
 Step 620: loss 2.010464668273926
 Step 630: loss 1.9628734588623047
 Step 640: loss 1.9215635061264038
 Step 650: loss 2.0762996673583984
 Step 660: loss 1.9436848163604736
 Step 670: loss 2.103586435317993
 Step 680: loss 1.921151876449585
 Step 690: loss 2.0375900268554688
 Step 700: loss 1.873571753501892
 Step 710: loss 2.118499279022217
 Step 720: loss 1.95355224609375
 Step 730: loss 1.9709444046020508
 Step 740: loss 2.0648865699768066
 Step 750: loss 2.028935194015503
 Step 760: loss 1.9319638013839722
 Step 770: loss 2.07603120803833
 Step 780: loss 2.021026372909546
 Step 790: loss 1.9132531881332397
 Step 800: loss 2.0291240215301514
 [Seed 55] Step 800 --> val loss 2.2683422565460205
 Step 810: loss 2.0140607357025146
 Step 820: loss 2.0464324951171875
 Step 830: loss 2.061558246612549
 Step 840: loss 1.9138250350952148
 Step 850: loss 1.9109199047088623
 Step 860: loss 2.0160977840423584
 Step 870: loss 2.0081286430358887
 Step 880: loss 2.0914833545684814
 Step 890: loss 1.9227104187011719
 Step 900: loss 2.119997262954712
 Step 910: loss 1.9667638540267944
 Step 920: loss 2.018327474594116
 Step 930: loss 1.9813592433929443
 Step 940: loss 2.0502495765686035
 Step 950: loss 2.038937568664551
 Step 960: loss 2.0363106727600098
 Step 970: loss 1.8040099143981934
 Step 980: loss 2.0934500694274902
 Step 990: loss 1.9328417778015137
 Step 1000: loss 1.9521230459213257
 [Seed 55] Step 1000 --> val loss 2.1639139652252197
 Step 1010: loss 2.076568603515625
 Step 1020: loss 2.1109554767608643
 Step 1030: loss 1.991822600364685
 Step 1040: loss 1.909315824508667
 Step 1050: loss 1.8401974439620972
 Step 1060: loss 1.979593276977539
 Step 1070: loss 1.9154915809631348
 Step 1080: loss 1.8826704025268555
 Step 1090: loss 1.957955241203308
 Step 1100: loss 1.940119981765747
 Step 1110: loss 1.9216363430023193
 Step 1120: loss 1.9483674764633179
 Step 1130: loss 2.0222339630126953
 Step 1140: loss 1.8906352519989014
 Step 1150: loss 1.8759596347808838
 Step 1160: loss 1.8502414226531982
 Step 1170: loss 1.8848292827606201
 Step 1180: loss 1.7540416717529297
 Step 1190: loss 1.878257155418396
 Step 1200: loss 1.8931694030761719
 [Seed 55] Step 1200 --> val loss 2.0571677684783936
 Step 1210: loss 1.864607572555542
 Step 1220: loss 1.8975517749786377
 Step 1230: loss 1.8548444509506226
 Step 1240: loss 1.867293357849121
 Step 1250: loss 1.8799359798431396
 Step 1260: loss 1.802284598350525
 Step 1270: loss 1.855857253074646
 Step 1280: loss 1.8556010723114014
 Step 1290: loss 1.8297533988952637
 Step 1300: loss 1.90802800655365
 Step 1310: loss 1.7381999492645264
 Step 1320: loss 1.8133634328842163
 Step 1330: loss 1.870737075805664
 Step 1340: loss 1.7417596578598022
 Step 1350: loss 1.7661685943603516
 Step 1360: loss 1.8769574165344238
 Step 1370: loss 1.9088691473007202
 Step 1380: loss 1.8533780574798584
 Step 1390: loss 1.8128299713134766
 Step 1400: loss 1.8446712493896484
 [Seed 55] Step 1400 --> val loss 1.9938278198242188
 Step 1410: loss 1.9093254804611206
 Step 1420: loss 1.7925527095794678
 Step 1430: loss 1.8766076564788818
 Step 1440: loss 1.6619253158569336
 Step 1450: loss 1.8616727590560913
 Step 1460: loss 1.8557136058807373
 Step 1470: loss 1.7441282272338867
 Step 1480: loss 1.8864206075668335
 Step 1490: loss 1.756948471069336
 Step 1500: loss 1.8035539388656616
 Step 1510: loss 1.7858513593673706
 Step 1520: loss 1.7144349813461304
 Step 1530: loss 1.704816222190857
 Step 1540: loss 1.755213737487793
 Step 1550: loss 1.8485610485076904
 Step 1560: loss 1.8024194240570068
 Step 1570: loss 1.8372418880462646
 Step 1580: loss 1.7941482067108154
 Step 1590: loss 1.6933048963546753
 Step 1600: loss 1.7981619834899902
 [Seed 55] Step 1600 --> val loss 1.8329895734786987
 Step 1610: loss 1.813897967338562
 Step 1620: loss 1.772647738456726
 Step 1630: loss 1.7510452270507812
 Step 1640: loss 1.759766936302185
 Step 1650: loss 1.8122613430023193
 Step 1660: loss 1.6645636558532715
 Step 1670: loss 1.786373496055603
 Step 1680: loss 1.710770606994629
 Step 1690: loss 1.6438287496566772
 Step 1700: loss 1.6611065864562988
 Step 1710: loss 1.7287487983703613
 Step 1720: loss 1.8536241054534912
 Step 1730: loss 1.6479482650756836
 Step 1740: loss 1.6673482656478882
 Step 1750: loss 1.6970562934875488
 Step 1760: loss 1.73331618309021
 Step 1770: loss 1.631665587425232
 Step 1780: loss 1.8295034170150757
 Step 1790: loss 1.7055578231811523
 Step 1800: loss 1.7077735662460327
 [Seed 55] Step 1800 --> val loss 1.773346185684204
 Step 1810: loss 1.7804301977157593
 Step 1820: loss 1.674207091331482
 Step 1830: loss 1.7142612934112549
 Step 1840: loss 1.5566835403442383
 Step 1850: loss 1.6644401550292969
 Step 1860: loss 1.6947987079620361
 Step 1870: loss 1.6371378898620605
 Step 1880: loss 1.7654368877410889
 Step 1890: loss 1.6326773166656494
 Step 1900: loss 1.6815791130065918
 Step 1910: loss 1.7893489599227905
 Step 1920: loss 1.8606982231140137
 Step 1930: loss 1.6096290349960327
 Step 1940: loss 1.5861778259277344
 Step 1950: loss 1.6769421100616455
 Step 1960: loss 1.697847843170166
 Step 1970: loss 1.6898406744003296
 Step 1980: loss 1.664242148399353
 Step 1990: loss 1.6592572927474976
 Step 2000: loss 1.6961455345153809
 [Seed 55] Step 2000 --> val loss 1.738916039466858
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 40 63  1 58 46 43  1 47 52] ...
 First target sequence: [40 63  1 58 46 43  1 47 52 58] ...

 Decoded input:  by the interpretation of full time
 May show like all yourself.


 Decoded target: by the interpretation of full time
 May show like all yourself.


 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.208351135253906
 [Seed 56] Step 0 --> val loss 4.18482780456543
 Step 10: loss 3.8532519340515137
 Step 20: loss 3.3394923210144043
 Step 30: loss 2.9045214653015137
 Step 40: loss 2.59027361869812
 Step 50: loss 2.6342649459838867
 Step 60: loss 2.428541421890259
 Step 70: loss 2.431276798248291
 Step 80: loss 2.3243913650512695
 Step 90: loss 2.3786263465881348
 Step 100: loss 2.365133762359619
 Step 110: loss 2.1305270195007324
 Step 120: loss 2.1871659755706787
 Step 130: loss 2.1644134521484375
 Step 140: loss 2.2243990898132324
 Step 150: loss 2.163423538208008
 Step 160: loss 2.1327714920043945
 Step 170: loss 2.165614366531372
 Step 180: loss 2.183804988861084
 Step 190: loss 2.0829286575317383
 Step 200: loss 2.119264602661133
 [Seed 56] Step 200 --> val loss 2.271878957748413
 Step 210: loss 2.1537649631500244
 Step 220: loss 2.1275830268859863
 Step 230: loss 2.0726938247680664
 Step 240: loss 2.1752707958221436
 Step 250: loss 2.126953601837158
 Step 260: loss 1.9673042297363281
 Step 270: loss 1.9954428672790527
 Step 280: loss 2.1500120162963867
 Step 290: loss 2.0109715461730957
 Step 300: loss 1.9505765438079834
 Step 310: loss 1.9651167392730713
 Step 320: loss 1.9466899633407593
 Step 330: loss 2.101633071899414
 Step 340: loss 2.1508994102478027
 Step 350: loss 2.0137314796447754
 Step 360: loss 2.0915541648864746
 Step 370: loss 2.1114814281463623
 Step 380: loss 1.9875125885009766
 Step 390: loss 2.0399527549743652
 Step 400: loss 2.102614164352417
 [Seed 56] Step 400 --> val loss 2.2636775970458984
 Step 410: loss 2.0473294258117676
 Step 420: loss 2.0616488456726074
 Step 430: loss 2.130317449569702
 Step 440: loss 2.0333662033081055
 Step 450: loss 2.117372989654541
 Step 460: loss 2.0215911865234375
 Step 470: loss 1.9431235790252686
 Step 480: loss 1.9500603675842285
 Step 490: loss 1.9752877950668335
 Step 500: loss 2.0576109886169434
 Step 510: loss 1.9883017539978027
 Step 520: loss 2.090424060821533
 Step 530: loss 2.052684783935547
 Step 540: loss 2.0442728996276855
 Step 550: loss 1.8732519149780273
 Step 560: loss 1.9966199398040771
 Step 570: loss 2.0230355262756348
 Step 580: loss 2.0022287368774414
 Step 590: loss 2.0579755306243896
 Step 600: loss 2.0555243492126465
 [Seed 56] Step 600 --> val loss 2.244762897491455
 Step 610: loss 1.9742954969406128
 Step 620: loss 2.085479497909546
 Step 630: loss 1.9137473106384277
 Step 640: loss 2.021711826324463
 Step 650: loss 1.994530439376831
 Step 660: loss 1.9010013341903687
 Step 670: loss 2.0509305000305176
 Step 680: loss 1.945016622543335
 Step 690: loss 1.9125638008117676
 Step 700: loss 2.0437064170837402
 Step 710: loss 2.002641201019287
 Step 720: loss 1.97975754737854
 Step 730: loss 1.9653515815734863
 Step 740: loss 2.0222740173339844
 Step 750: loss 1.9423669576644897
 Step 760: loss 1.9377150535583496
 Step 770: loss 1.9501179456710815
 Step 780: loss 1.9949349164962769
 Step 790: loss 1.943039894104004
 Step 800: loss 2.0435688495635986
 [Seed 56] Step 800 --> val loss 2.261791467666626
 Step 810: loss 2.036912441253662
 Step 820: loss 2.0636887550354004
 Step 830: loss 1.856637954711914
 Step 840: loss 2.012299060821533
 Step 850: loss 2.030761241912842
 Step 860: loss 1.959153413772583
 Step 870: loss 2.137420177459717
 Step 880: loss 2.037424087524414
 Step 890: loss 2.0333683490753174
 Step 900: loss 2.0095767974853516
 Step 910: loss 2.0614449977874756
 Step 920: loss 2.072723150253296
 Step 930: loss 2.054527997970581
 Step 940: loss 2.0048987865448
 Step 950: loss 1.9233176708221436
 Step 960: loss 1.9626182317733765
 Step 970: loss 1.945749044418335
 Step 980: loss 2.038050889968872
 Step 990: loss 1.956508994102478
 Step 1000: loss 2.1266956329345703
 [Seed 56] Step 1000 --> val loss 2.2535107135772705
 Step 1010: loss 1.938957691192627
 Step 1020: loss 1.9859727621078491
 Step 1030: loss 1.9706494808197021
 Step 1040: loss 1.9597185850143433
 Step 1050: loss 1.931105613708496
 Step 1060: loss 2.0093860626220703
 Step 1070: loss 2.012545585632324
 Step 1080: loss 2.001704692840576
 Step 1090: loss 2.0593533515930176
 Step 1100: loss 1.9307491779327393
 Step 1110: loss 2.0119082927703857
 Step 1120: loss 1.9336504936218262
 Step 1130: loss 1.9678491353988647
 Step 1140: loss 1.9186424016952515
 Step 1150: loss 2.065539836883545
 Step 1160: loss 1.9889202117919922
 Step 1170: loss 2.056332588195801
 Step 1180: loss 1.9844921827316284
 Step 1190: loss 1.9886173009872437
 Step 1200: loss 1.9096120595932007
 [Seed 56] Step 1200 --> val loss 2.273320198059082
 Step 1210: loss 2.2109851837158203
 Step 1220: loss 1.8966444730758667
 Step 1230: loss 1.9647486209869385
 Step 1240: loss 1.989679217338562
 Step 1250: loss 1.9855053424835205
 Step 1260: loss 1.9044617414474487
 Step 1270: loss 1.8072452545166016
 Step 1280: loss 1.9106156826019287
 Step 1290: loss 1.9323924779891968
 Step 1300: loss 1.9382331371307373
 Step 1310: loss 1.9877209663391113
 Step 1320: loss 1.864498496055603
 Step 1330: loss 1.851836919784546
 Step 1340: loss 1.984978437423706
 Step 1350: loss 2.00968599319458
 Step 1360: loss 1.8666551113128662
 Step 1370: loss 1.8516957759857178
 Step 1380: loss 2.0189967155456543
 Step 1390: loss 1.8732414245605469
 Step 1400: loss 1.8100297451019287
 [Seed 56] Step 1400 --> val loss 2.174959182739258
 Step 1410: loss 1.8898342847824097
 Step 1420: loss 1.9152703285217285
 Step 1430: loss 1.9607770442962646
 Step 1440: loss 1.8104947805404663
 Step 1450: loss 1.8390653133392334
 Step 1460: loss 1.996671438217163
 Step 1470: loss 1.9627532958984375
 Step 1480: loss 2.057422637939453
 Step 1490: loss 1.767214298248291
 Step 1500: loss 1.9653370380401611
 Step 1510: loss 1.9198672771453857
 Step 1520: loss 1.857802152633667
 Step 1530: loss 1.9224700927734375
 Step 1540: loss 1.7991862297058105
 Step 1550: loss 1.9673470258712769
 Step 1560: loss 1.8151286840438843
 Step 1570: loss 1.8720557689666748
 Step 1580: loss 2.0579776763916016
 Step 1590: loss 1.9314420223236084
 Step 1600: loss 1.8624179363250732
 [Seed 56] Step 1600 --> val loss 2.213282823562622
 Step 1610: loss 1.8238170146942139
 Step 1620: loss 1.8981876373291016
 Step 1630: loss 1.8931958675384521
 Step 1640: loss 1.8896949291229248
 Step 1650: loss 1.8576620817184448
 Step 1660: loss 1.8206290006637573
 Step 1670: loss 1.816307544708252
 Step 1680: loss 1.7307517528533936
 Step 1690: loss 1.9506547451019287
 Step 1700: loss 1.8571422100067139
 Step 1710: loss 1.7940855026245117
 Step 1720: loss 1.8936504125595093
 Step 1730: loss 1.7183122634887695
 Step 1740: loss 1.837499976158142
 Step 1750: loss 1.7961688041687012
 Step 1760: loss 1.9085906744003296
 Step 1770: loss 1.908376693725586
 Step 1780: loss 1.7203927040100098
 Step 1790: loss 1.8594624996185303
 Step 1800: loss 1.916910171508789
 [Seed 56] Step 1800 --> val loss 2.0588529109954834
 Step 1810: loss 1.7316467761993408
 Step 1820: loss 1.7507121562957764
 Step 1830: loss 1.9475905895233154
 Step 1840: loss 1.7834275960922241
 Step 1850: loss 1.7376680374145508
 Step 1860: loss 1.8041142225265503
 Step 1870: loss 1.756161093711853
 Step 1880: loss 1.7887037992477417
 Step 1890: loss 1.7781248092651367
 Step 1900: loss 1.7881743907928467
 Step 1910: loss 1.827427864074707
 Step 1920: loss 1.636605978012085
 Step 1930: loss 1.8080508708953857
 Step 1940: loss 1.70222806930542
 Step 1950: loss 1.7512493133544922
 Step 1960: loss 1.7789673805236816
 Step 1970: loss 1.7422511577606201
 Step 1980: loss 1.641627311706543
 Step 1990: loss 1.7749398946762085
 Step 2000: loss 1.7990292310714722
 [Seed 56] Step 2000 --> val loss 1.953159213066101
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [59 58  1 47 52  1 46 43 56  1] ...
 First target sequence: [58  1 47 52  1 46 43 56  1 58] ...

 Decoded input: ut in her tender heart the aspiring flame
 Of golden sovereignty;

 Decoded target: t in her tender heart the aspiring flame
 Of golden sovereignty; 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.220479488372803
 [Seed 57] Step 0 --> val loss 4.186150074005127
 Step 10: loss 3.8266754150390625
 Step 20: loss 3.3364295959472656
 Step 30: loss 2.8747191429138184
 Step 40: loss 2.623500108718872
 Step 50: loss 2.4985978603363037
 Step 60: loss 2.5072309970855713
 Step 70: loss 2.387725591659546
 Step 80: loss 2.2605323791503906
 Step 90: loss 2.3945302963256836
 Step 100: loss 2.3303465843200684
 Step 110: loss 2.206254005432129
 Step 120: loss 2.1661295890808105
 Step 130: loss 2.301975727081299
 Step 140: loss 2.1738991737365723
 Step 150: loss 2.1622824668884277
 Step 160: loss 2.1442294120788574
 Step 170: loss 2.212143659591675
 Step 180: loss 2.2226967811584473
 Step 190: loss 2.0919463634490967
 Step 200: loss 2.202831268310547
 [Seed 57] Step 200 --> val loss 2.4278550148010254
 Step 210: loss 2.1057567596435547
 Step 220: loss 2.1001479625701904
 Step 230: loss 2.017381191253662
 Step 240: loss 2.08113169670105
 Step 250: loss 2.044924259185791
 Step 260: loss 2.012996196746826
 Step 270: loss 2.035796642303467
 Step 280: loss 2.0877161026000977
 Step 290: loss 2.0229997634887695
 Step 300: loss 2.0369820594787598
 Step 310: loss 2.0501575469970703
 Step 320: loss 2.0921683311462402
 Step 330: loss 1.9522662162780762
 Step 340: loss 2.0625829696655273
 Step 350: loss 2.026339292526245
 Step 360: loss 2.0289664268493652
 Step 370: loss 1.967358946800232
 Step 380: loss 1.9709687232971191
 Step 390: loss 2.069646120071411
 Step 400: loss 2.112030506134033
 [Seed 57] Step 400 --> val loss 2.2797040939331055
 Step 410: loss 2.0258476734161377
 Step 420: loss 2.0698466300964355
 Step 430: loss 1.9913783073425293
 Step 440: loss 1.9079599380493164
 Step 450: loss 1.999673843383789
 Step 460: loss 2.0004334449768066
 Step 470: loss 2.079237461090088
 Step 480: loss 2.059648275375366
 Step 490: loss 1.9451777935028076
 Step 500: loss 2.0406341552734375
 Step 510: loss 1.999314785003662
 Step 520: loss 2.069038152694702
 Step 530: loss 1.974230170249939
 Step 540: loss 2.0275630950927734
 Step 550: loss 1.934170126914978
 Step 560: loss 1.9920566082000732
 Step 570: loss 1.9745779037475586
 Step 580: loss 1.9606850147247314
 Step 590: loss 2.00199294090271
 Step 600: loss 1.9883430004119873
 [Seed 57] Step 600 --> val loss 2.1599202156066895
 Step 610: loss 2.0692801475524902
 Step 620: loss 2.01045298576355
 Step 630: loss 1.9511394500732422
 Step 640: loss 1.9647492170333862
 Step 650: loss 2.0538864135742188
 Step 660: loss 2.1101255416870117
 Step 670: loss 2.040412425994873
 Step 680: loss 2.146347999572754
 Step 690: loss 1.9773505926132202
 Step 700: loss 2.0163474082946777
 Step 710: loss 2.0501885414123535
 Step 720: loss 1.9482072591781616
 Step 730: loss 1.7612636089324951
 Step 740: loss 1.9577497243881226
 Step 750: loss 2.0080642700195312
 Step 760: loss 1.9943647384643555
 Step 770: loss 1.984410047531128
 Step 780: loss 1.9570891857147217
 Step 790: loss 1.9185924530029297
 Step 800: loss 1.9921789169311523
 [Seed 57] Step 800 --> val loss 2.192225217819214
 Step 810: loss 1.928545355796814
 Step 820: loss 2.1135013103485107
 Step 830: loss 2.029849052429199
 Step 840: loss 2.011028289794922
 Step 850: loss 2.0381574630737305
 Step 860: loss 1.910723090171814
 Step 870: loss 2.0760791301727295
 Step 880: loss 1.9725027084350586
 Step 890: loss 1.9426525831222534
 Step 900: loss 2.0405349731445312
 Step 910: loss 1.9941352605819702
 Step 920: loss 1.9061808586120605
 Step 930: loss 1.9828588962554932
 Step 940: loss 2.0204601287841797
 Step 950: loss 1.843881607055664
 Step 960: loss 1.9082353115081787
 Step 970: loss 1.9449447393417358
 Step 980: loss 1.9376730918884277
 Step 990: loss 1.9989500045776367
 Step 1000: loss 1.9580416679382324
 [Seed 57] Step 1000 --> val loss 2.130181074142456
 Step 1010: loss 1.9727758169174194
 Step 1020: loss 1.8993724584579468
 Step 1030: loss 2.0781846046447754
 Step 1040: loss 1.9838627576828003
 Step 1050: loss 1.9603271484375
 Step 1060: loss 2.103250026702881
 Step 1070: loss 1.8426952362060547
 Step 1080: loss 1.9928975105285645
 Step 1090: loss 2.0244078636169434
 Step 1100: loss 1.8054527044296265
 Step 1110: loss 2.1008687019348145
 Step 1120: loss 1.9977238178253174
 Step 1130: loss 2.102708578109741
 Step 1140: loss 1.9009116888046265
 Step 1150: loss 1.866058588027954
 Step 1160: loss 1.866465449333191
 Step 1170: loss 1.8297985792160034
 Step 1180: loss 1.941049337387085
 Step 1190: loss 1.9233545064926147
 Step 1200: loss 2.0600156784057617
 [Seed 57] Step 1200 --> val loss 2.2010602951049805
 Step 1210: loss 1.811372995376587
 Step 1220: loss 1.9571274518966675
 Step 1230: loss 1.9766547679901123
 Step 1240: loss 1.8657646179199219
 Step 1250: loss 1.9853004217147827
 Step 1260: loss 1.9929320812225342
 Step 1270: loss 1.9221076965332031
 Step 1280: loss 1.9353492259979248
 Step 1290: loss 1.954416036605835
 Step 1300: loss 2.001673698425293
 Step 1310: loss 1.9096921682357788
 Step 1320: loss 1.9362950325012207
 Step 1330: loss 1.913580298423767
 Step 1340: loss 1.9833177328109741
 Step 1350: loss 1.8566291332244873
 Step 1360: loss 1.8677613735198975
 Step 1370: loss 1.7836699485778809
 Step 1380: loss 1.9344520568847656
 Step 1390: loss 1.7488691806793213
 Step 1400: loss 1.8974647521972656
 [Seed 57] Step 1400 --> val loss 1.9883453845977783
 Step 1410: loss 1.82595956325531
 Step 1420: loss 1.6825779676437378
 Step 1430: loss 1.9121897220611572
 Step 1440: loss 1.721769094467163
 Step 1450: loss 1.713151454925537
 Step 1460: loss 1.8075532913208008
 Step 1470: loss 1.8212230205535889
 Step 1480: loss 1.8150293827056885
 Step 1490: loss 1.9626221656799316
 Step 1500: loss 1.9188512563705444
 Step 1510: loss 1.7766337394714355
 Step 1520: loss 1.7186086177825928
 Step 1530: loss 1.8672369718551636
 Step 1540: loss 1.7143515348434448
 Step 1550: loss 1.7324750423431396
 Step 1560: loss 1.7899049520492554
 Step 1570: loss 1.7685497999191284
 Step 1580: loss 1.8090732097625732
 Step 1590: loss 1.8317065238952637
 Step 1600: loss 1.7080307006835938
 [Seed 57] Step 1600 --> val loss 1.9502537250518799
 Step 1610: loss 1.8329633474349976
 Step 1620: loss 1.7513225078582764
 Step 1630: loss 1.6593549251556396
 Step 1640: loss 1.7441953420639038
 Step 1650: loss 1.7251918315887451
 Step 1660: loss 1.742188811302185
 Step 1670: loss 1.7796828746795654
 Step 1680: loss 1.730412483215332
 Step 1690: loss 1.7436716556549072
 Step 1700: loss 1.868433952331543
 Step 1710: loss 1.8713735342025757
 Step 1720: loss 1.729448676109314
 Step 1730: loss 1.7545223236083984
 Step 1740: loss 1.6490898132324219
 Step 1750: loss 1.6972253322601318
 Step 1760: loss 1.754692554473877
 Step 1770: loss 1.882635474205017
 Step 1780: loss 1.7845392227172852
 Step 1790: loss 1.57589852809906
 Step 1800: loss 1.6592611074447632
 [Seed 57] Step 1800 --> val loss 1.9354089498519897
 Step 1810: loss 1.7708680629730225
 Step 1820: loss 1.6363060474395752
 Step 1830: loss 1.6421189308166504
 Step 1840: loss 1.812972068786621
 Step 1850: loss 1.6113885641098022
 Step 1860: loss 1.6566494703292847
 Step 1870: loss 1.6538636684417725
 Step 1880: loss 1.6973025798797607
 Step 1890: loss 1.8571629524230957
 Step 1900: loss 1.6830967664718628
 Step 1910: loss 1.631624460220337
 Step 1920: loss 1.7175477743148804
 Step 1930: loss 1.6984590291976929
 Step 1940: loss 1.708767294883728
 Step 1950: loss 1.698455810546875
 Step 1960: loss 1.7005112171173096
 Step 1970: loss 1.6934406757354736
 Step 1980: loss 1.717228889465332
 Step 1990: loss 1.6239001750946045
 Step 2000: loss 1.77803373336792
 [Seed 57] Step 2000 --> val loss 1.8719733953475952
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [46 43 56 43 44 53 56 43  8  5] ...
 First target sequence: [43 56 43 44 53 56 43  8  5  1] ...

 Decoded input: herefore.' When we granted that,
 Here was 'I thank you for your 

 Decoded target: erefore.' When we granted that,
 Here was 'I thank you for your v
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.189874649047852
 [Seed 58] Step 0 --> val loss 4.185962200164795
 Step 10: loss 3.8477001190185547
 Step 20: loss 3.340463399887085
 Step 30: loss 2.8623769283294678
 Step 40: loss 2.6369285583496094
 Step 50: loss 2.597032308578491
 Step 60: loss 2.5018067359924316
 Step 70: loss 2.4521210193634033
 Step 80: loss 2.39026141166687
 Step 90: loss 2.3300561904907227
 Step 100: loss 2.2255020141601562
 Step 110: loss 2.278576374053955
 Step 120: loss 2.32623553276062
 Step 130: loss 2.244144916534424
 Step 140: loss 2.250633716583252
 Step 150: loss 2.2261385917663574
 Step 160: loss 2.164663076400757
 Step 170: loss 2.114356517791748
 Step 180: loss 2.130305051803589
 Step 190: loss 2.1149065494537354
 Step 200: loss 2.0340936183929443
 [Seed 58] Step 200 --> val loss 2.3668363094329834
 Step 210: loss 2.1124725341796875
 Step 220: loss 2.0442121028900146
 Step 230: loss 2.0173981189727783
 Step 240: loss 2.057800054550171
 Step 250: loss 1.9924733638763428
 Step 260: loss 2.043734550476074
 Step 270: loss 2.058215856552124
 Step 280: loss 2.072038412094116
 Step 290: loss 2.04012393951416
 Step 300: loss 1.9423749446868896
 Step 310: loss 2.0796010494232178
 Step 320: loss 2.0292012691497803
 Step 330: loss 2.009815216064453
 Step 340: loss 2.0618841648101807
 Step 350: loss 1.9778684377670288
 Step 360: loss 2.012971878051758
 Step 370: loss 2.1460399627685547
 Step 380: loss 2.001725673675537
 Step 390: loss 2.2438693046569824
 Step 400: loss 2.1184935569763184
 [Seed 58] Step 400 --> val loss 2.2597897052764893
 Step 410: loss 2.072936534881592
 Step 420: loss 1.8812086582183838
 Step 430: loss 2.025026798248291
 Step 440: loss 2.0568723678588867
 Step 450: loss 2.064265012741089
 Step 460: loss 1.9749755859375
 Step 470: loss 2.017575263977051
 Step 480: loss 1.9594848155975342
 Step 490: loss 2.006270408630371
 Step 500: loss 1.9248255491256714
 Step 510: loss 2.064814567565918
 Step 520: loss 2.134552478790283
 Step 530: loss 1.9963496923446655
 Step 540: loss 2.1138525009155273
 Step 550: loss 1.9601635932922363
 Step 560: loss 2.0775506496429443
 Step 570: loss 2.023733139038086
 Step 580: loss 2.0210700035095215
 Step 590: loss 2.0968575477600098
 Step 600: loss 2.042741060256958
 [Seed 58] Step 600 --> val loss 2.3233697414398193
 Step 610: loss 2.066922664642334
 Step 620: loss 2.074805736541748
 Step 630: loss 2.044658660888672
 Step 640: loss 2.0698633193969727
 Step 650: loss 2.016273021697998
 Step 660: loss 2.086785078048706
 Step 670: loss 2.100611925125122
 Step 680: loss 1.965540885925293
 Step 690: loss 1.9037630558013916
 Step 700: loss 1.9758232831954956
 Step 710: loss 2.055671453475952
 Step 720: loss 2.07462477684021
 Step 730: loss 2.137417793273926
 Step 740: loss 2.0181093215942383
 Step 750: loss 2.071150302886963
 Step 760: loss 2.0209057331085205
 Step 770: loss 1.9416418075561523
 Step 780: loss 1.96820068359375
 Step 790: loss 1.9230170249938965
 Step 800: loss 1.9985651969909668
 [Seed 58] Step 800 --> val loss 2.2687318325042725
 Step 810: loss 1.9559321403503418
 Step 820: loss 2.062934398651123
 Step 830: loss 2.028834104537964
 Step 840: loss 1.973679542541504
 Step 850: loss 1.9553145170211792
 Step 860: loss 1.9495518207550049
 Step 870: loss 2.0549283027648926
 Step 880: loss 2.0979814529418945
 Step 890: loss 1.894943118095398
 Step 900: loss 1.9581750631332397
 Step 910: loss 1.947608470916748
 Step 920: loss 2.105545997619629
 Step 930: loss 1.9250128269195557
 Step 940: loss 1.9160802364349365
 Step 950: loss 1.9350411891937256
 Step 960: loss 1.929907202720642
 Step 970: loss 2.0905115604400635
 Step 980: loss 2.020028829574585
 Step 990: loss 2.052967071533203
 Step 1000: loss 1.8594611883163452
 [Seed 58] Step 1000 --> val loss 2.12052321434021
 Step 1010: loss 1.9413923025131226
 Step 1020: loss 1.9725922346115112
 Step 1030: loss 2.0580036640167236
 Step 1040: loss 2.0705506801605225
 Step 1050: loss 2.0529420375823975
 Step 1060: loss 1.926648497581482
 Step 1070: loss 1.9064134359359741
 Step 1080: loss 2.0195109844207764
 Step 1090: loss 1.989715337753296
 Step 1100: loss 1.9811948537826538
 Step 1110: loss 1.8366140127182007
 Step 1120: loss 1.7972195148468018
 Step 1130: loss 1.9965283870697021
 Step 1140: loss 1.9007015228271484
 Step 1150: loss 1.909898042678833
 Step 1160: loss 1.867033839225769
 Step 1170: loss 1.91159188747406
 Step 1180: loss 1.8849620819091797
 Step 1190: loss 1.9608161449432373
 Step 1200: loss 1.9719823598861694
 [Seed 58] Step 1200 --> val loss 2.1654796600341797
 Step 1210: loss 1.9461991786956787
 Step 1220: loss 1.890038251876831
 Step 1230: loss 1.9204463958740234
 Step 1240: loss 1.9390122890472412
 Step 1250: loss 1.95974862575531
 Step 1260: loss 1.8840510845184326
 Step 1270: loss 1.9082502126693726
 Step 1280: loss 1.8912618160247803
 Step 1290: loss 1.9346531629562378
 Step 1300: loss 1.9079043865203857
 Step 1310: loss 2.002483367919922
 Step 1320: loss 1.976383090019226
 Step 1330: loss 1.8889694213867188
 Step 1340: loss 1.8768634796142578
 Step 1350: loss 1.8894097805023193
 Step 1360: loss 1.975775957107544
 Step 1370: loss 2.0236592292785645
 Step 1380: loss 1.9210350513458252
 Step 1390: loss 1.892653226852417
 Step 1400: loss 1.954952597618103
 [Seed 58] Step 1400 --> val loss 2.02497935295105
 Step 1410: loss 1.920853853225708
 Step 1420: loss 1.9098738431930542
 Step 1430: loss 2.074775457382202
 Step 1440: loss 1.9505977630615234
 Step 1450: loss 1.8108386993408203
 Step 1460: loss 1.9176876544952393
 Step 1470: loss 1.8716106414794922
 Step 1480: loss 1.8546867370605469
 Step 1490: loss 1.7918989658355713
 Step 1500: loss 1.845159888267517
 Step 1510: loss 1.8821338415145874
 Step 1520: loss 1.9287879467010498
 Step 1530: loss 1.844078779220581
 Step 1540: loss 1.7987512350082397
 Step 1550: loss 1.850661277770996
 Step 1560: loss 1.7394897937774658
 Step 1570: loss 1.7367929220199585
 Step 1580: loss 1.7251847982406616
 Step 1590: loss 1.7613656520843506
 Step 1600: loss 1.7865227460861206
 [Seed 58] Step 1600 --> val loss 2.0621793270111084
 Step 1610: loss 1.7807015180587769
 Step 1620: loss 1.6863963603973389
 Step 1630: loss 1.785793662071228
 Step 1640: loss 1.825382947921753
 Step 1650: loss 1.9519232511520386
 Step 1660: loss 1.7386443614959717
 Step 1670: loss 1.8654687404632568
 Step 1680: loss 1.6602433919906616
 Step 1690: loss 1.812319278717041
 Step 1700: loss 1.7227987051010132
 Step 1710: loss 1.7835919857025146
 Step 1720: loss 1.7231495380401611
 Step 1730: loss 1.7179235219955444
 Step 1740: loss 1.722051978111267
 Step 1750: loss 1.8274586200714111
 Step 1760: loss 1.7572468519210815
 Step 1770: loss 1.72350013256073
 Step 1780: loss 1.6900696754455566
 Step 1790: loss 1.7053030729293823
 Step 1800: loss 1.7937167882919312
 [Seed 58] Step 1800 --> val loss 1.961400032043457
 Step 1810: loss 1.814737319946289
 Step 1820: loss 1.7595481872558594
 Step 1830: loss 1.687117576599121
 Step 1840: loss 1.84684157371521
 Step 1850: loss 1.723592758178711
 Step 1860: loss 1.6815580129623413
 Step 1870: loss 1.839437484741211
 Step 1880: loss 1.8082029819488525
 Step 1890: loss 1.7274329662322998
 Step 1900: loss 1.6700654029846191
 Step 1910: loss 1.62626051902771
 Step 1920: loss 1.6400396823883057
 Step 1930: loss 1.693256139755249
 Step 1940: loss 1.63582181930542
 Step 1950: loss 1.6632635593414307
 Step 1960: loss 1.7339977025985718
 Step 1970: loss 1.677947998046875
 Step 1980: loss 1.7118066549301147
 Step 1990: loss 1.8255656957626343
 Step 2000: loss 1.616254210472107
 [Seed 58] Step 2000 --> val loss 1.844765543937683
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [59 57  8  0  0 32 46 47 56 42] ...
 First target sequence: [57  8  0  0 32 46 47 56 42  1] ...

 Decoded input: us.

 Third Watchman:
 Ay, but give me worship and quietness;
 I li

 Decoded target: s.

 Third Watchman:
 Ay, but give me worship and quietness;
 I lik
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.193964958190918
 [Seed 59] Step 0 --> val loss 4.183913707733154
 Step 10: loss 3.855520725250244
 Step 20: loss 3.3072280883789062
 Step 30: loss 2.89624285697937
 Step 40: loss 2.6455469131469727
 Step 50: loss 2.569359302520752
 Step 60: loss 2.443291425704956
 Step 70: loss 2.405095100402832
 Step 80: loss 2.4033632278442383
 Step 90: loss 2.2916855812072754
 Step 100: loss 2.3185219764709473
 Step 110: loss 2.314547061920166
 Step 120: loss 2.285552501678467
 Step 130: loss 2.2506933212280273
 Step 140: loss 2.114269256591797
 Step 150: loss 2.1404101848602295
 Step 160: loss 2.191723108291626
 Step 170: loss 2.168287754058838
 Step 180: loss 2.021272897720337
 Step 190: loss 2.1190025806427
 Step 200: loss 2.1036672592163086
 [Seed 59] Step 200 --> val loss 2.3833115100860596
 Step 210: loss 2.066776752471924
 Step 220: loss 2.0582542419433594
 Step 230: loss 2.087106227874756
 Step 240: loss 2.0487911701202393
 Step 250: loss 2.061964988708496
 Step 260: loss 1.9839215278625488
 Step 270: loss 2.053231716156006
 Step 280: loss 2.137523889541626
 Step 290: loss 2.108894109725952
 Step 300: loss 2.0561366081237793
 Step 310: loss 2.0484025478363037
 Step 320: loss 2.127519130706787
 Step 330: loss 2.003523111343384
 Step 340: loss 2.083217144012451
 Step 350: loss 1.9045909643173218
 Step 360: loss 2.0258853435516357
 Step 370: loss 2.0157134532928467
 Step 380: loss 2.039734363555908
 Step 390: loss 1.914854884147644
 Step 400: loss 1.990679144859314
 [Seed 59] Step 400 --> val loss 2.2157723903656006
 Step 410: loss 2.108921527862549
 Step 420: loss 2.0561208724975586
 Step 430: loss 1.9418315887451172
 Step 440: loss 1.9513120651245117
 Step 450: loss 2.013228416442871
 Step 460: loss 2.006742477416992
 Step 470: loss 2.029435396194458
 Step 480: loss 2.0315046310424805
 Step 490: loss 2.114192008972168
 Step 500: loss 2.014224052429199
 Step 510: loss 2.0330352783203125
 Step 520: loss 2.053384304046631
 Step 530: loss 1.9177378416061401
 Step 540: loss 2.023678779602051
 Step 550: loss 2.0464401245117188
 Step 560: loss 1.8982290029525757
 Step 570: loss 2.028707504272461
 Step 580: loss 2.0431013107299805
 Step 590: loss 1.950257658958435
 Step 600: loss 1.9291415214538574
 [Seed 59] Step 600 --> val loss 2.2678353786468506
 Step 610: loss 2.016608953475952
 Step 620: loss 2.0178158283233643
 Step 630: loss 2.0191636085510254
 Step 640: loss 1.9446866512298584
 Step 650: loss 1.9239301681518555
 Step 660: loss 2.065056562423706
 Step 670: loss 1.8465569019317627
 Step 680: loss 2.0369174480438232
 Step 690: loss 1.981512427330017
 Step 700: loss 2.0078582763671875
 Step 710: loss 1.926314353942871
 Step 720: loss 2.0094146728515625
 Step 730: loss 2.0926437377929688
 Step 740: loss 1.9709266424179077
 Step 750: loss 2.078444480895996
 Step 760: loss 1.9684786796569824
 Step 770: loss 2.018767833709717
 Step 780: loss 2.0807037353515625
 Step 790: loss 1.934300422668457
 Step 800: loss 2.021811008453369
 [Seed 59] Step 800 --> val loss 2.227600574493408
 Step 810: loss 1.9979066848754883
 Step 820: loss 2.010880947113037
 Step 830: loss 2.1965787410736084
 Step 840: loss 2.1128172874450684
 Step 850: loss 2.0415267944335938
 Step 860: loss 2.0013909339904785
 Step 870: loss 1.9257855415344238
 Step 880: loss 2.026089906692505
 Step 890: loss 2.0596346855163574
 Step 900: loss 1.9710315465927124
 Step 910: loss 2.1509299278259277
 Step 920: loss 2.322986602783203
 Step 930: loss 1.9962999820709229
 Step 940: loss 2.075946569442749
 Step 950: loss 2.0199756622314453
 Step 960: loss 1.997671127319336
 Step 970: loss 2.064120292663574
 Step 980: loss 2.008626937866211
 Step 990: loss 1.9254987239837646
 Step 1000: loss 1.9037336111068726
 [Seed 59] Step 1000 --> val loss 2.302280902862549
 Step 1010: loss 1.9709584712982178
 Step 1020: loss 1.9616355895996094
 Step 1030: loss 2.046766757965088
 Step 1040: loss 1.9839935302734375
 Step 1050: loss 2.0281317234039307
 Step 1060: loss 1.9649585485458374
 Step 1070: loss 1.938366174697876
 Step 1080: loss 1.9261367321014404
 Step 1090: loss 1.9023178815841675
 Step 1100: loss 1.9869440793991089
 Step 1110: loss 1.9929225444793701
 Step 1120: loss 1.9120275974273682
 Step 1130: loss 1.9694583415985107
 Step 1140: loss 1.9308702945709229
 Step 1150: loss 1.9867581129074097
 Step 1160: loss 1.8571317195892334
 Step 1170: loss 1.9575591087341309
 Step 1180: loss 1.898472785949707
 Step 1190: loss 1.9468636512756348
 Step 1200: loss 1.8839647769927979
 [Seed 59] Step 1200 --> val loss 2.135709047317505
 Step 1210: loss 1.879523754119873
 Step 1220: loss 1.850595474243164
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 59] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 59] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 59] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 59] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [49  1 51 43  1 39 52  1 53 50] ...
 First target sequence: [ 1 51 43  1 39 52  1 53 50 42] ...

 Decoded input: k me an old murderer,
 Now I have stain'd the childhood of our jo

 Decoded target:  me an old murderer,
 Now I have stain'd the childhood of our joy
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.209256172180176
 [Seed 60] Step 0 --> val loss 4.186305999755859
 Step 10: loss 3.835047483444214
 Step 20: loss 3.328996419906616
 Step 30: loss 2.8433375358581543
 Step 40: loss 2.707965850830078
 Step 50: loss 2.569307804107666
 Step 60: loss 2.5638647079467773
 Step 70: loss 2.4562723636627197
 Step 80: loss 2.3851537704467773
 Step 90: loss 2.3579280376434326
 Step 100: loss 2.32978892326355
 Step 110: loss 2.3288631439208984
 Step 120: loss 2.186829090118408
 Step 130: loss 2.1914775371551514
 Step 140: loss 2.146921157836914
 Step 150: loss 2.2473888397216797
 Step 160: loss 2.1548051834106445
 Step 170: loss 2.0893208980560303
 Step 180: loss 2.0759592056274414
 Step 190: loss 2.2016215324401855
 Step 200: loss 2.0168609619140625
 [Seed 60] Step 200 --> val loss 2.3699378967285156
 Step 210: loss 2.083047866821289
 Step 220: loss 2.0650641918182373
 Step 230: loss 2.0627269744873047
 Step 240: loss 2.034608840942383
 Step 250: loss 2.175340414047241
 Step 260: loss 2.110142707824707
 Step 270: loss 2.038696765899658
 Step 280: loss 2.048401117324829
 Step 290: loss 2.041512966156006
 Step 300: loss 1.9845613241195679
 Step 310: loss 2.021589994430542
 Step 320: loss 2.0602970123291016
 Step 330: loss 1.9640557765960693
 Step 340: loss 2.0048723220825195
 Step 350: loss 1.9441967010498047
 Step 360: loss 1.9348978996276855
 Step 370: loss 2.0618720054626465
 Step 380: loss 2.0844738483428955
 Step 390: loss 2.0513477325439453
 Step 400: loss 1.923802137374878
 [Seed 60] Step 400 --> val loss 2.3267924785614014
 Step 410: loss 2.0164954662323
 Step 420: loss 1.9967120885849
 Step 430: loss 1.978015661239624
 Step 440: loss 2.0097007751464844
 Step 450: loss 2.025036334991455
 Step 460: loss 2.028573513031006
 Step 470: loss 1.9993376731872559
 Step 480: loss 1.9058418273925781
 Step 490: loss 2.0050435066223145
 Step 500: loss 1.9911613464355469
 Step 510: loss 2.1323153972625732
 Step 520: loss 1.9966437816619873
 Step 530: loss 2.00689435005188
 Step 540: loss 2.004974126815796
 Step 550: loss 2.0083117485046387
 Step 560: loss 2.0490756034851074
 Step 570: loss 2.079594135284424
 Step 580: loss 2.0525403022766113
 Step 590: loss 1.873440146446228
 Step 600: loss 2.011237144470215
 [Seed 60] Step 600 --> val loss 2.2931010723114014
 Step 610: loss 1.948781967163086
 Step 620: loss 2.040522336959839
 Step 630: loss 2.0201005935668945
 Step 640: loss 2.0117032527923584
 Step 650: loss 2.0620899200439453
 Step 660: loss 2.1122822761535645
 Step 670: loss 2.1826376914978027
 Step 680: loss 1.9580552577972412
 Step 690: loss 1.9144262075424194
 Step 700: loss 2.035090208053589
 Step 710: loss 1.9628690481185913
 Step 720: loss 1.9943132400512695
 Step 730: loss 2.035891532897949
 Step 740: loss 1.8791165351867676
 Step 750: loss 1.9146636724472046
 Step 760: loss 2.00734806060791
 Step 770: loss 2.0122909545898438
 Step 780: loss 1.9621412754058838
 Step 790: loss 1.9899181127548218
 Step 800: loss 1.951037883758545
 [Seed 60] Step 800 --> val loss 2.305989980697632
 Step 810: loss 1.9495309591293335
 Step 820: loss 2.089874029159546
 Step 830: loss 1.9664087295532227
 Step 840: loss 2.085616111755371
 Step 850: loss 2.0065338611602783
 Step 860: loss 1.993945837020874
 Step 870: loss 2.0224647521972656
 Step 880: loss 1.9983361959457397
 Step 890: loss 1.946280837059021
 Step 900: loss 1.9259638786315918
 Step 910: loss 2.054349184036255
 Step 920: loss 1.9515695571899414
 Step 930: loss 2.0743978023529053
 Step 940: loss 1.9630554914474487
 Step 950: loss 1.9161008596420288
 Step 960: loss 2.1336188316345215
 Step 970: loss 1.9012587070465088
 Step 980: loss 1.9814872741699219
 Step 990: loss 1.9342117309570312
 Step 1000: loss 1.9211394786834717
 [Seed 60] Step 1000 --> val loss 2.1791388988494873
 Step 1010: loss 1.9906203746795654
 Step 1020: loss 2.0113439559936523
 Step 1030: loss 1.8922910690307617
 Step 1040: loss 2.024080276489258
 Step 1050: loss 2.005836248397827
 Step 1060: loss 1.9119501113891602
 Step 1070: loss 2.0440375804901123
 Step 1080: loss 1.916869878768921
 Step 1090: loss 1.9012712240219116
 Step 1100: loss 1.9655754566192627
 Step 1110: loss 1.9892926216125488
 Step 1120: loss 1.9387540817260742
 Step 1130: loss 1.8993220329284668
 Step 1140: loss 1.8903911113739014
 Step 1150: loss 1.961791753768921
 Step 1160: loss 1.8692684173583984
 Step 1170: loss 1.9145126342773438
 Step 1180: loss 1.9152252674102783
 Step 1190: loss 1.9583978652954102
 Step 1200: loss 1.7903430461883545
 [Seed 60] Step 1200 --> val loss 2.157116651535034
 Step 1210: loss 1.865409255027771
 Step 1220: loss 2.0164132118225098
 Step 1230: loss 1.9586913585662842
 Step 1240: loss 1.8006787300109863
 Step 1250: loss 1.9468584060668945
 Step 1260: loss 1.9435007572174072
 Step 1270: loss 1.9602760076522827
 Step 1280: loss 1.9914090633392334
 Step 1290: loss 1.8286222219467163
 Step 1300: loss 1.8551839590072632
 Step 1310: loss 1.9612767696380615
 Step 1320: loss 1.9453634023666382
 Step 1330: loss 1.9525256156921387
 Step 1340: loss 1.9031081199645996
 Step 1350: loss 1.7361940145492554
 Step 1360: loss 1.9081676006317139
 Step 1370: loss 1.8945128917694092
 Step 1380: loss 1.859232783317566
 Step 1390: loss 1.8675057888031006
 Step 1400: loss 1.9921433925628662
 [Seed 60] Step 1400 --> val loss 2.066171646118164
 Step 1410: loss 1.8134880065917969
 Step 1420: loss 1.7758934497833252
 Step 1430: loss 1.7759864330291748
 Step 1440: loss 1.7786991596221924
 Step 1450: loss 1.8239574432373047
 Step 1460: loss 1.901240587234497
 Step 1470: loss 1.874475359916687
 Step 1480: loss 1.7950958013534546
 Step 1490: loss 1.7889151573181152
 Step 1500: loss 1.885947585105896
 Step 1510: loss 1.7463496923446655
 Step 1520: loss 1.8592422008514404
 Step 1530: loss 1.8140777349472046
 Step 1540: loss 1.6853883266448975
 Step 1550: loss 1.8391984701156616
 Step 1560: loss 1.8012140989303589
 Step 1570: loss 1.8595259189605713
 Step 1580: loss 1.8265924453735352
 Step 1590: loss 1.6467007398605347
 Step 1600: loss 1.789665937423706
 [Seed 60] Step 1600 --> val loss 2.0315747261047363
 Step 1610: loss 1.9302239418029785
 Step 1620: loss 1.8150051832199097
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 60] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 60] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [44 44 11  0 35 46 43 52  1 54] ...
 First target sequence: [44 11  0 35 46 43 52  1 54 56] ...

 Decoded input: ff;
 When presently through all thy veins shall run
 A cold and dr

 Decoded target: f;
 When presently through all thy veins shall run
 A cold and dro
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.21489143371582
 [Seed 61] Step 0 --> val loss 4.181804656982422
 Step 10: loss 3.8344669342041016
 Step 20: loss 3.350628614425659
 Step 30: loss 2.909876585006714
 Step 40: loss 2.647015333175659
 Step 50: loss 2.6477012634277344
 Step 60: loss 2.5415446758270264
 Step 70: loss 2.4644675254821777
 Step 80: loss 2.3997302055358887
 Step 90: loss 2.309502363204956
 Step 100: loss 2.292618751525879
 Step 110: loss 2.3759841918945312
 Step 120: loss 2.220953941345215
 Step 130: loss 2.2021424770355225
 Step 140: loss 2.1787149906158447
 Step 150: loss 2.219069480895996
 Step 160: loss 2.188091993331909
 Step 170: loss 2.078385829925537
 Step 180: loss 2.227076530456543
 Step 190: loss 2.3660497665405273
 Step 200: loss 2.2172162532806396
 [Seed 61] Step 200 --> val loss 2.270681142807007
 Step 210: loss 2.125100612640381
 Step 220: loss 2.0510799884796143
 Step 230: loss 2.0581507682800293
 Step 240: loss 2.0657894611358643
 Step 250: loss 2.1014034748077393
 Step 260: loss 2.110276937484741
 Step 270: loss 2.122713327407837
 Step 280: loss 2.061009168624878
 Step 290: loss 2.087529182434082
 Step 300: loss 2.1596322059631348
 Step 310: loss 1.923214316368103
 Step 320: loss 2.034531831741333
 Step 330: loss 2.05930757522583
 Step 340: loss 2.010087490081787
 Step 350: loss 2.005796194076538
 Step 360: loss 2.0237817764282227
 Step 370: loss 2.1216909885406494
 Step 380: loss 2.1019177436828613
 Step 390: loss 2.09452748298645
 Step 400: loss 2.1006298065185547
 [Seed 61] Step 400 --> val loss 2.1940367221832275
 Step 410: loss 2.027341842651367
 Step 420: loss 2.0488345623016357
 Step 430: loss 2.010850429534912
 Step 440: loss 2.0672013759613037
 Step 450: loss 2.1012380123138428
 Step 460: loss 1.9720687866210938
 Step 470: loss 2.0313382148742676
 Step 480: loss 2.0621190071105957
 Step 490: loss 2.0604496002197266
 Step 500: loss 2.0189263820648193
 Step 510: loss 2.199683427810669
 Step 520: loss 2.0049588680267334
 Step 530: loss 2.0846028327941895
 Step 540: loss 2.0503087043762207
 Step 550: loss 1.972873568534851
 Step 560: loss 1.9144413471221924
 Step 570: loss 1.967120885848999
 Step 580: loss 1.9776562452316284
 Step 590: loss 2.0361709594726562
 Step 600: loss 2.050515651702881
 [Seed 61] Step 600 --> val loss 2.3012959957122803
 Step 610: loss 2.0312538146972656
 Step 620: loss 2.073059558868408
 Step 630: loss 1.9370572566986084
 Step 640: loss 2.0284080505371094
 Step 650: loss 2.104646682739258
 Step 660: loss 2.1379475593566895
 Step 670: loss 1.9314963817596436
 Step 680: loss 2.0554773807525635
 Step 690: loss 2.0347747802734375
 Step 700: loss 2.0008087158203125
 Step 710: loss 2.076662063598633
 Step 720: loss 2.0053720474243164
 Step 730: loss 2.0269827842712402
 Step 740: loss 2.073394536972046
 Step 750: loss 2.0130817890167236
 Step 760: loss 1.9995921850204468
 Step 770: loss 2.0548739433288574
 Step 780: loss 1.9612760543823242
 Step 790: loss 2.1809613704681396
 Step 800: loss 1.9309849739074707
 [Seed 61] Step 800 --> val loss 2.2416911125183105
 Step 810: loss 2.093449115753174
 Step 820: loss 2.0656909942626953
 Step 830: loss 1.9767770767211914
 Step 840: loss 1.958675742149353
 Step 850: loss 2.0017781257629395
 Step 860: loss 2.0879440307617188
 Step 870: loss 2.101682662963867
 Step 880: loss 1.932314395904541
 Step 890: loss 1.9813960790634155
 Step 900: loss 1.9944865703582764
 Step 910: loss 2.0572166442871094
 Step 920: loss 2.0646331310272217
 Step 930: loss 2.0183613300323486
 Step 940: loss 1.9667081832885742
 Step 950: loss 2.042936325073242
 Step 960: loss 1.960506796836853
 Step 970: loss 1.930920958518982
 Step 980: loss 1.972752332687378
 Step 990: loss 1.9954421520233154
 Step 1000: loss 2.034868001937866
 [Seed 61] Step 1000 --> val loss 2.2211191654205322
 Step 1010: loss 1.9787715673446655
 Step 1020: loss 1.9379913806915283
 Step 1030: loss 1.8839200735092163
 Step 1040: loss 2.012190341949463
 Step 1050: loss 2.0317625999450684
 Step 1060: loss 1.9974550008773804
 Step 1070: loss 1.9997437000274658
 Step 1080: loss 1.984764814376831
 Step 1090: loss 1.957726001739502
 Step 1100: loss 1.8835281133651733
 Step 1110: loss 1.9221961498260498
 Step 1120: loss 1.9319725036621094
 Step 1130: loss 1.9403579235076904
 Step 1140: loss 1.908608317375183
 Step 1150: loss 1.920509696006775
 Step 1160: loss 1.9126780033111572
 Step 1170: loss 1.911096453666687
 Step 1180: loss 1.7619142532348633
 Step 1190: loss 1.7780965566635132
 Step 1200: loss 1.9445350170135498
 [Seed 61] Step 1200 --> val loss 2.1838629245758057
 Step 1210: loss 2.0950536727905273
 Step 1220: loss 1.940901756286621
 Step 1230: loss 2.022251844406128
 Step 1240: loss 1.8267698287963867
 Step 1250: loss 1.8832695484161377
 Step 1260: loss 1.8638633489608765
 Step 1270: loss 1.8922536373138428
 Step 1280: loss 1.9582475423812866
 Step 1290: loss 1.873035192489624
 Step 1300: loss 1.8923838138580322
 Step 1310: loss 1.8438552618026733
 Step 1320: loss 1.888380765914917
 Step 1330: loss 1.9256775379180908
 Step 1340: loss 1.858807921409607
 Step 1350: loss 1.8900407552719116
 Step 1360: loss 1.8722423315048218
 Step 1370: loss 1.97149658203125
 Step 1380: loss 1.8533093929290771
 Step 1390: loss 1.8573784828186035
 Step 1400: loss 1.866349220275879
 [Seed 61] Step 1400 --> val loss 2.038153886795044
 Step 1410: loss 1.9268615245819092
 Step 1420: loss 1.8415627479553223
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 61] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 61] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 61] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [41 46  1 60 47 56 58 59 43  1] ...
 First target sequence: [46  1 60 47 56 58 59 43  1 40] ...

 Decoded input: ch virtue begs and virtue grants.

 KING EDWARD IV:
 No, by my tro

 Decoded target: h virtue begs and virtue grants.

 KING EDWARD IV:
 No, by my trot
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.188974380493164
 [Seed 62] Step 0 --> val loss 4.185174465179443
 Step 10: loss 3.8491551876068115
 Step 20: loss 3.344158172607422
 Step 30: loss 2.940509796142578
 Step 40: loss 2.6174352169036865
 Step 50: loss 2.5547165870666504
 Step 60: loss 2.4533472061157227
 Step 70: loss 2.45259952545166
 Step 80: loss 2.3264999389648438
 Step 90: loss 2.222599744796753
 Step 100: loss 2.2181804180145264
 Step 110: loss 2.272944688796997
 Step 120: loss 2.1431498527526855
 Step 130: loss 2.2062315940856934
 Step 140: loss 2.413623809814453
 Step 150: loss 2.109447956085205
 Step 160: loss 2.1996989250183105
 Step 170: loss 2.1021831035614014
 Step 180: loss 2.0370216369628906
 Step 190: loss 2.073979139328003
 Step 200: loss 2.072164297103882
 [Seed 62] Step 200 --> val loss 2.338395833969116
 Step 210: loss 2.066073417663574
 Step 220: loss 2.1937828063964844
 Step 230: loss 2.086089849472046
 Step 240: loss 2.03204345703125
 Step 250: loss 2.0442633628845215
 Step 260: loss 2.072937488555908
 Step 270: loss 2.1230435371398926
 Step 280: loss 2.061197280883789
 Step 290: loss 2.0529675483703613
 Step 300: loss 2.0501880645751953
 Step 310: loss 2.0461416244506836
 Step 320: loss 2.065481662750244
 Step 330: loss 1.9283193349838257
 Step 340: loss 1.9953057765960693
 Step 350: loss 1.930195689201355
 Step 360: loss 2.012481212615967
 Step 370: loss 2.058772087097168
 Step 380: loss 2.090137004852295
 Step 390: loss 2.0662856101989746
 Step 400: loss 2.0908148288726807
 [Seed 62] Step 400 --> val loss 2.223869562149048
 Step 410: loss 2.042907476425171
 Step 420: loss 1.984084129333496
 Step 430: loss 1.9844725131988525
 Step 440: loss 2.069263219833374
 Step 450: loss 2.050264835357666
 Step 460: loss 1.9778164625167847
 Step 470: loss 2.1106109619140625
 Step 480: loss 2.0107197761535645
 Step 490: loss 2.056361198425293
 Step 500: loss 1.9543671607971191
 Step 510: loss 2.016197681427002
 Step 520: loss 2.0628840923309326
 Step 530: loss 1.9693506956100464
 Step 540: loss 2.0513548851013184
 Step 550: loss 1.9727230072021484
 Step 560: loss 1.9421991109848022
 Step 570: loss 1.966334581375122
 Step 580: loss 2.115471363067627
 Step 590: loss 1.9869029521942139
 Step 600: loss 1.9907121658325195
 [Seed 62] Step 600 --> val loss 2.194397449493408
 Step 610: loss 2.0922746658325195
 Step 620: loss 1.9697483777999878
 Step 630: loss 1.9762805700302124
 Step 640: loss 2.0179014205932617
 Step 650: loss 2.0022459030151367
 Step 660: loss 2.0776009559631348
 Step 670: loss 1.957267165184021
 Step 680: loss 2.1502017974853516
 Step 690: loss 2.0271172523498535
 Step 700: loss 1.887143850326538
 Step 710: loss 2.0622923374176025
 Step 720: loss 1.9137113094329834
 Step 730: loss 2.022155523300171
 Step 740: loss 1.8538053035736084
 Step 750: loss 2.059170722961426
 Step 760: loss 1.9960211515426636
 Step 770: loss 1.947859764099121
 Step 780: loss 2.077780246734619
 Step 790: loss 2.035825729370117
 Step 800: loss 2.0508322715759277
 [Seed 62] Step 800 --> val loss 2.2095773220062256
 Step 810: loss 1.9411420822143555
 Step 820: loss 1.9510382413864136
 Step 830: loss 1.9032349586486816
 Step 840: loss 1.984480381011963
 Step 850: loss 1.974793791770935
 Step 860: loss 1.9076298475265503
 Step 870: loss 2.0643622875213623
 Step 880: loss 2.04522705078125
 Step 890: loss 1.9286407232284546
 Step 900: loss 1.8628143072128296
 Step 910: loss 2.0367610454559326
 Step 920: loss 2.0198962688446045
 Step 930: loss 1.9105823040008545
 Step 940: loss 2.0068531036376953
 Step 950: loss 1.7705711126327515
 Step 960: loss 1.9301135540008545
 Step 970: loss 2.012889862060547
 Step 980: loss 1.8997406959533691
 Step 990: loss 1.8931193351745605
 Step 1000: loss 2.0877299308776855
 [Seed 62] Step 1000 --> val loss 2.21985125541687
 Step 1010: loss 1.9409856796264648
 Step 1020: loss 1.981828212738037
 Step 1030: loss 1.8229612112045288
 Step 1040: loss 1.8418660163879395
 Step 1050: loss 2.0554637908935547
 Step 1060: loss 1.9309356212615967
 Step 1070: loss 2.003507614135742
 Step 1080: loss 1.889937162399292
 Step 1090: loss 2.024207592010498
 Step 1100: loss 1.9641637802124023
 Step 1110: loss 1.939812421798706
 Step 1120: loss 1.9700983762741089
 Step 1130: loss 2.017716884613037
 Step 1140: loss 2.0262601375579834
 Step 1150: loss 1.8997883796691895
 Step 1160: loss 1.8660337924957275
 Step 1170: loss 1.9747003316879272
 Step 1180: loss 1.9071760177612305
 Step 1190: loss 1.8558357954025269
 Step 1200: loss 2.033684730529785
 [Seed 62] Step 1200 --> val loss 2.1625986099243164
 Step 1210: loss 1.8698856830596924
 Step 1220: loss 1.8022745847702026
 Step 1230: loss 1.8806371688842773
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 62] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 62] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 62] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 62] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [56 42  6  1 51 63  1 51 53 58] ...
 First target sequence: [42  6  1 51 63  1 51 53 58 46] ...

 Decoded input: rd, my mother lives.

 BUCKINGHAM:
 Fear not, my lord, I'll play t

 Decoded target: d, my mother lives.

 BUCKINGHAM:
 Fear not, my lord, I'll play th
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.205653190612793
 [Seed 63] Step 0 --> val loss 4.185075759887695
 Step 10: loss 3.852994918823242
 Step 20: loss 3.3853836059570312
 Step 30: loss 2.87249755859375
 Step 40: loss 2.637958288192749
 Step 50: loss 2.5508925914764404
 Step 60: loss 2.452357053756714
 Step 70: loss 2.3750317096710205
 Step 80: loss 2.316270112991333
 Step 90: loss 2.306741714477539
 Step 100: loss 2.363693952560425
 Step 110: loss 2.321010112762451
 Step 120: loss 2.3310041427612305
 Step 130: loss 2.1495532989501953
 Step 140: loss 2.161283254623413
 Step 150: loss 2.1783289909362793
 Step 160: loss 2.1745223999023438
 Step 170: loss 2.1411681175231934
 Step 180: loss 2.147878885269165
 Step 190: loss 2.1190905570983887
 Step 200: loss 2.1410293579101562
 [Seed 63] Step 200 --> val loss 2.3762013912200928
 Step 210: loss 2.0432186126708984
 Step 220: loss 2.0399842262268066
 Step 230: loss 2.0440621376037598
 Step 240: loss 2.046902656555176
 Step 250: loss 2.0022568702697754
 Step 260: loss 2.153843879699707
 Step 270: loss 2.11344051361084
 Step 280: loss 2.0583674907684326
 Step 290: loss 2.1135647296905518
 Step 300: loss 2.0882136821746826
 Step 310: loss 2.064863681793213
 Step 320: loss 2.021303176879883
 Step 330: loss 2.084336996078491
 Step 340: loss 2.083495616912842
 Step 350: loss 1.9465856552124023
 Step 360: loss 2.137676954269409
 Step 370: loss 1.9526853561401367
 Step 380: loss 1.9396135807037354
 Step 390: loss 1.951637625694275
 Step 400: loss 2.068798065185547
 [Seed 63] Step 400 --> val loss 2.229541778564453
 Step 410: loss 2.0177667140960693
 Step 420: loss 2.072601318359375
 Step 430: loss 2.0192198753356934
 Step 440: loss 1.9929585456848145
 Step 450: loss 2.0695641040802
 Step 460: loss 2.055694580078125
 Step 470: loss 2.038508176803589
 Step 480: loss 1.9831979274749756
 Step 490: loss 2.0179147720336914
 Step 500: loss 1.9636011123657227
 Step 510: loss 1.9009225368499756
 Step 520: loss 1.8821014165878296
 Step 530: loss 2.019137382507324
 Step 540: loss 1.943739652633667
 Step 550: loss 1.92569100856781
 Step 560: loss 1.9798877239227295
 Step 570: loss 2.000901222229004
 Step 580: loss 2.001755714416504
 Step 590: loss 2.0501551628112793
 Step 600: loss 1.946264624595642
 [Seed 63] Step 600 --> val loss 2.2747178077697754
 Step 610: loss 2.068800926208496
 Step 620: loss 1.922116994857788
 Step 630: loss 1.9479904174804688
 Step 640: loss 1.9938466548919678
 Step 650: loss 2.10589861869812
 Step 660: loss 1.856872320175171
 Step 670: loss 2.030168294906616
 Step 680: loss 1.9678704738616943
 Step 690: loss 2.010592222213745
 Step 700: loss 2.041835308074951
 Step 710: loss 2.076026678085327
 Step 720: loss 2.0143840312957764
 Step 730: loss 2.0009703636169434
 Step 740: loss 1.9674198627471924
 Step 750: loss 1.9491147994995117
 Step 760: loss 1.9435456991195679
 Step 770: loss 1.9887948036193848
 Step 780: loss 2.067269802093506
 Step 790: loss 2.0854556560516357
 Step 800: loss 2.0100317001342773
 [Seed 63] Step 800 --> val loss 2.185495615005493
 Step 810: loss 2.081271171569824
 Step 820: loss 2.0452959537506104
 Step 830: loss 1.961918592453003
 Step 840: loss 2.0157198905944824
 Step 850: loss 1.9485763311386108
 Step 860: loss 2.0792312622070312
 Step 870: loss 2.0005199909210205
 Step 880: loss 2.0337073802948
 Step 890: loss 1.9053272008895874
 Step 900: loss 1.9936041831970215
 Step 910: loss 1.9803285598754883
 Step 920: loss 1.9650297164916992
 Step 930: loss 1.990739345550537
 Step 940: loss 1.9903558492660522
 Step 950: loss 2.0555338859558105
 Step 960: loss 1.8982295989990234
 Step 970: loss 1.96323561668396
 Step 980: loss 1.7901586294174194
 Step 990: loss 2.052018880844116
 Step 1000: loss 2.0016250610351562
 [Seed 63] Step 1000 --> val loss 2.2343735694885254
 Step 1010: loss 2.073291778564453
 Step 1020: loss 1.9595258235931396
 Step 1030: loss 1.956342101097107
 Step 1040: loss 1.923105001449585
 Step 1050: loss 1.821897268295288
 Step 1060: loss 1.9389560222625732
 Step 1070: loss 1.8741092681884766
 Step 1080: loss 1.9873197078704834
 Step 1090: loss 1.9977483749389648
 Step 1100: loss 1.8837661743164062
 Step 1110: loss 1.9497032165527344
 Step 1120: loss 1.8746933937072754
 Step 1130: loss 2.0179049968719482
 Step 1140: loss 1.8761496543884277
 Step 1150: loss 1.8517274856567383
 Step 1160: loss 1.86420476436615
 Step 1170: loss 1.8519738912582397
 Step 1180: loss 2.021261215209961
 Step 1190: loss 1.8776038885116577
 Step 1200: loss 1.8533380031585693
 [Seed 63] Step 1200 --> val loss 2.0276591777801514
 Step 1210: loss 1.7995423078536987
 Step 1220: loss 1.819949984550476
 Step 1230: loss 1.7789769172668457
 Step 1240: loss 1.798069715499878
 Step 1250: loss 1.8343629837036133
 Step 1260: loss 1.9180450439453125
 Step 1270: loss 1.8627532720565796
 Step 1280: loss 1.90854811668396
 Step 1290: loss 1.8273900747299194
 Step 1300: loss 1.8966712951660156
 Step 1310: loss 1.8484710454940796
 Step 1320: loss 1.9392719268798828
 Step 1330: loss 1.9035943746566772
 Step 1340: loss 1.8051084280014038
 Step 1350: loss 1.9094637632369995
 Step 1360: loss 1.7649157047271729
 Step 1370: loss 1.8042572736740112
 Step 1380: loss 2.0087857246398926
 Step 1390: loss 1.9578393697738647
 Step 1400: loss 1.8207343816757202
 [Seed 63] Step 1400 --> val loss 2.0044972896575928
 Step 1410: loss 1.8049486875534058
 Step 1420: loss 1.8604724407196045
 Step 1430: loss 1.9047054052352905
 Step 1440: loss 1.79205322265625
 Step 1450: loss 1.978193759918213
 Step 1460: loss 1.866002082824707
 Step 1470: loss 1.8315141201019287
 Step 1480: loss 1.9072345495224
 Step 1490: loss 1.7981027364730835
 Step 1500: loss 1.9649556875228882
 Step 1510: loss 1.7930015325546265
 Step 1520: loss 1.8094923496246338
 Step 1530: loss 1.9108279943466187
 Step 1540: loss 1.8317269086837769
 Step 1550: loss 1.8825187683105469
 Step 1560: loss 1.6978014707565308
 Step 1570: loss 1.8053205013275146
 Step 1580: loss 1.722365379333496
 Step 1590: loss 1.8727951049804688
 Step 1600: loss 1.793068289756775
 [Seed 63] Step 1600 --> val loss 1.8840303421020508
 Step 1610: loss 1.654345989227295
 Step 1620: loss 1.7704766988754272
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 63] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 63] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [42  1 39 50 50  1 58 46 43 47] ...
 First target sequence: [ 1 39 50 50  1 58 46 43 47 56] ...

 Decoded input: d all their ministers attend on him.

 GLOUCESTER:
 What doth she 

 Decoded target:  all their ministers attend on him.

 GLOUCESTER:
 What doth she s
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.223361968994141
 [Seed 64] Step 0 --> val loss 4.18109655380249
 Step 10: loss 3.8342909812927246
 Step 20: loss 3.3286635875701904
 Step 30: loss 2.837810516357422
 Step 40: loss 2.6136531829833984
 Step 50: loss 2.525017261505127
 Step 60: loss 2.424253225326538
 Step 70: loss 2.396982192993164
 Step 80: loss 2.508880138397217
 Step 90: loss 2.442974090576172
 Step 100: loss 2.2624118328094482
 Step 110: loss 2.2863879203796387
 Step 120: loss 2.2447519302368164
 Step 130: loss 2.2176735401153564
 Step 140: loss 2.044891595840454
 Step 150: loss 2.301677703857422
 Step 160: loss 2.2480173110961914
 Step 170: loss 2.0985922813415527
 Step 180: loss 2.0179388523101807
 Step 190: loss 2.1625468730926514
 Step 200: loss 2.0452635288238525
 [Seed 64] Step 200 --> val loss 2.459357500076294
 Step 210: loss 2.0980968475341797
 Step 220: loss 2.1007790565490723
 Step 230: loss 2.2012228965759277
 Step 240: loss 2.083534002304077
 Step 250: loss 2.0801196098327637
 Step 260: loss 2.068385124206543
 Step 270: loss 2.187506675720215
 Step 280: loss 1.975499153137207
 Step 290: loss 1.930687665939331
 Step 300: loss 1.9652278423309326
 Step 310: loss 2.08158016204834
 Step 320: loss 2.0925984382629395
 Step 330: loss 2.0814883708953857
 Step 340: loss 2.047267436981201
 Step 350: loss 2.0488405227661133
 Step 360: loss 2.0280063152313232
 Step 370: loss 1.972646713256836
 Step 380: loss 2.0179429054260254
 Step 390: loss 2.107724905014038
 Step 400: loss 2.0464208126068115
 [Seed 64] Step 400 --> val loss 2.345731735229492
 Step 410: loss 1.9271459579467773
 Step 420: loss 2.1090826988220215
 Step 430: loss 2.0221171379089355
 Step 440: loss 2.017150402069092
 Step 450: loss 2.087761878967285
 Step 460: loss 2.037309408187866
 Step 470: loss 2.0729990005493164
 Step 480: loss 1.9597309827804565
 Step 490: loss 2.0652894973754883
 Step 500: loss 2.0055198669433594
 Step 510: loss 1.9838981628417969
 Step 520: loss 2.003150224685669
 Step 530: loss 2.0572638511657715
 Step 540: loss 1.9999154806137085
 Step 550: loss 2.168735980987549
 Step 560: loss 2.023252487182617
 Step 570: loss 1.9543046951293945
 Step 580: loss 2.023556709289551
 Step 590: loss 1.9409507513046265
 Step 600: loss 1.918397307395935
 [Seed 64] Step 600 --> val loss 2.2082536220550537
 Step 610: loss 2.0958733558654785
 Step 620: loss 2.042320966720581
 Step 630: loss 1.9755758047103882
 Step 640: loss 2.0709972381591797
 Step 650: loss 2.025709867477417
 Step 660: loss 2.1220285892486572
 Step 670: loss 1.9093173742294312
 Step 680: loss 1.9048833847045898
 Step 690: loss 1.9663887023925781
 Step 700: loss 2.0875651836395264
 Step 710: loss 1.9616869688034058
 Step 720: loss 2.0493850708007812
 Step 730: loss 2.1766090393066406
 Step 740: loss 2.127317190170288
 Step 750: loss 2.031099319458008
 Step 760: loss 2.185314893722534
 Step 770: loss 1.9854391813278198
 Step 780: loss 2.0340933799743652
 Step 790: loss 2.0178489685058594
 Step 800: loss 2.008208751678467
 [Seed 64] Step 800 --> val loss 2.2628371715545654
 Step 810: loss 1.8945634365081787
 Step 820: loss 1.9382041692733765
 Step 830: loss 2.0020837783813477
 Step 840: loss 2.021806478500366
 Step 850: loss 1.914939284324646
 Step 860: loss 2.1095454692840576
 Step 870: loss 2.037703037261963
 Step 880: loss 1.9719901084899902
 Step 890: loss 2.023709535598755
 Step 900: loss 1.9783897399902344
 Step 910: loss 2.0354042053222656
 Step 920: loss 1.8618543148040771
 Step 930: loss 2.0021560192108154
 Step 940: loss 2.0801382064819336
 Step 950: loss 1.9211106300354004
 Step 960: loss 1.9932243824005127
 Step 970: loss 2.0008914470672607
 Step 980: loss 2.031428813934326
 Step 990: loss 2.102755069732666
 Step 1000: loss 1.9837005138397217
 [Seed 64] Step 1000 --> val loss 2.2523725032806396
 Step 1010: loss 2.027024745941162
 Step 1020: loss 1.9383246898651123
 Step 1030: loss 1.9241619110107422
 Step 1040: loss 1.8582919836044312
 Step 1050: loss 1.865465521812439
 Step 1060: loss 1.9249495267868042
 Step 1070: loss 1.9740270376205444
 Step 1080: loss 1.978073000907898
 Step 1090: loss 1.9934648275375366
 Step 1100: loss 1.991272211074829
 Step 1110: loss 1.9794578552246094
 Step 1120: loss 2.0168862342834473
 Step 1130: loss 1.9987859725952148
 Step 1140: loss 1.9361672401428223
 Step 1150: loss 1.9996163845062256
 Step 1160: loss 2.0340445041656494
 Step 1170: loss 2.0422513484954834
 Step 1180: loss 1.953749418258667
 Step 1190: loss 2.0533320903778076
 Step 1200: loss 1.8678849935531616
 [Seed 64] Step 1200 --> val loss 2.174508810043335
 Step 1210: loss 1.8528075218200684
 Step 1220: loss 1.9662961959838867
 Step 1230: loss 2.0128207206726074
 Step 1240: loss 2.0648279190063477
 Step 1250: loss 1.7859277725219727
 Step 1260: loss 1.997128963470459
 Step 1270: loss 1.9366061687469482
 Step 1280: loss 1.9584439992904663
 Step 1290: loss 1.9971578121185303
 Step 1300: loss 2.0988235473632812
 Step 1310: loss 1.8548660278320312
 Step 1320: loss 1.8052442073822021
 Step 1330: loss 1.9714149236679077
 Step 1340: loss 1.9741382598876953
 Step 1350: loss 1.9381049871444702
 Step 1360: loss 1.8826359510421753
 Step 1370: loss 1.8295215368270874
 Step 1380: loss 1.8047380447387695
 Step 1390: loss 1.8524866104125977
 Step 1400: loss 1.996201515197754
 [Seed 64] Step 1400 --> val loss 2.0853753089904785
 Step 1410: loss 1.9209814071655273
 Step 1420: loss 1.8275401592254639
 Step 1430: loss 1.9240983724594116
 Step 1440: loss 1.8599655628204346
 Step 1450: loss 2.01167631149292
 Step 1460: loss 1.881319522857666
 Step 1470: loss 1.9769821166992188
 Step 1480: loss 1.9122341871261597
 Step 1490: loss 1.9511773586273193
 Step 1500: loss 1.823055624961853
 Step 1510: loss 1.8223791122436523
 Step 1520: loss 1.8539018630981445
 Step 1530: loss 1.8396480083465576
 Step 1540: loss 1.8503177165985107
 Step 1550: loss 1.9229469299316406
 Step 1560: loss 1.9243590831756592
 Step 1570: loss 1.9791746139526367
 Step 1580: loss 1.908017635345459
 Step 1590: loss 1.935118556022644
 Step 1600: loss 1.9459452629089355
 [Seed 64] Step 1600 --> val loss 2.0704963207244873
 Step 1610: loss 1.7677353620529175
 Step 1620: loss 2.0489306449890137
 Step 1630: loss 1.9993059635162354
 Step 1640: loss 1.868176817893982
 Step 1650: loss 1.9410486221313477
 Step 1660: loss 1.7929599285125732
 Step 1670: loss 1.8808469772338867
 Step 1680: loss 1.8397245407104492
 Step 1690: loss 1.8620774745941162
 Step 1700: loss 1.7410354614257812
 Step 1710: loss 1.6992003917694092
 Step 1720: loss 1.8895677328109741
 Step 1730: loss 1.8476927280426025
 Step 1740: loss 1.8597934246063232
 Step 1750: loss 1.870345950126648
 Step 1760: loss 1.9164700508117676
 Step 1770: loss 1.7493776082992554
 Step 1780: loss 1.7058145999908447
 Step 1790: loss 1.8282790184020996
 Step 1800: loss 1.754301905632019
 [Seed 64] Step 1800 --> val loss 1.9406670331954956
 Step 1810: loss 1.777295470237732
 Step 1820: loss 1.7856574058532715
 Step 1830: loss 1.8073699474334717
 Step 1840: loss 1.735445261001587
 Step 1850: loss 1.7748407125473022
 Step 1860: loss 1.864162802696228
 Step 1870: loss 1.762041449546814
 Step 1880: loss 1.7699545621871948
 Step 1890: loss 1.773636817932129
 Step 1900: loss 1.8069286346435547
 Step 1910: loss 1.6980834007263184
 Step 1920: loss 1.671004056930542
 Step 1930: loss 1.7953386306762695
 Step 1940: loss 1.7071609497070312
 Step 1950: loss 1.717698097229004
 Step 1960: loss 1.6800354719161987
 Step 1970: loss 1.6859862804412842
 Step 1980: loss 1.7969346046447754
 Step 1990: loss 1.8893332481384277
 Step 2000: loss 1.7537190914154053
 [Seed 64] Step 2000 --> val loss 1.9283636808395386
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 58 46 53 59  1 46 39 57 58] ...
 First target sequence: [58 46 53 59  1 46 39 57 58  1] ...

 Decoded input:  thou hast wrought
 A deed of slander with thy fatal hand
 Upon my

 Decoded target: thou hast wrought
 A deed of slander with thy fatal hand
 Upon my 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.217906475067139
 [Seed 65] Step 0 --> val loss 4.1828765869140625
 Step 10: loss 3.8209891319274902
 Step 20: loss 3.428365468978882
 Step 30: loss 2.7499184608459473
 Step 40: loss 2.746339797973633
 Step 50: loss 2.587355613708496
 Step 60: loss 2.4532642364501953
 Step 70: loss 2.4481444358825684
 Step 80: loss 2.377474784851074
 Step 90: loss 2.4271883964538574
 Step 100: loss 2.246804714202881
 Step 110: loss 2.3047525882720947
 Step 120: loss 2.221851348876953
 Step 130: loss 2.26965069770813
 Step 140: loss 2.108309745788574
 Step 150: loss 2.1481451988220215
 Step 160: loss 2.0289711952209473
 Step 170: loss 2.1661272048950195
 Step 180: loss 2.1885452270507812
 Step 190: loss 2.179182767868042
 Step 200: loss 2.066821813583374
 [Seed 65] Step 200 --> val loss 2.32285213470459
 Step 210: loss 2.0928213596343994
 Step 220: loss 2.0921082496643066
 Step 230: loss 2.1414105892181396
 Step 240: loss 2.087435245513916
 Step 250: loss 2.0033838748931885
 Step 260: loss 2.01562237739563
 Step 270: loss 2.1153106689453125
 Step 280: loss 1.945448637008667
 Step 290: loss 2.1842527389526367
 Step 300: loss 2.0868988037109375
 Step 310: loss 1.9727485179901123
 Step 320: loss 2.0546212196350098
 Step 330: loss 2.0979976654052734
 Step 340: loss 2.068423271179199
 Step 350: loss 1.9679670333862305
 Step 360: loss 1.8820102214813232
 Step 370: loss 1.9844882488250732
 Step 380: loss 1.8156802654266357
 Step 390: loss 2.0196492671966553
 Step 400: loss 2.017829418182373
 [Seed 65] Step 400 --> val loss 2.378767728805542
 Step 410: loss 1.9962282180786133
 Step 420: loss 1.9701193571090698
 Step 430: loss 2.0944981575012207
 Step 440: loss 1.886681079864502
 Step 450: loss 1.9838752746582031
 Step 460: loss 2.078758478164673
 Step 470: loss 2.0199477672576904
 Step 480: loss 2.036701202392578
 Step 490: loss 1.9658730030059814
 Step 500: loss 1.9349740743637085
 Step 510: loss 2.1329240798950195
 Step 520: loss 2.012631893157959
 Step 530: loss 2.0275990962982178
 Step 540: loss 1.9674218893051147
 Step 550: loss 2.0326004028320312
 Step 560: loss 2.057112455368042
 Step 570: loss 1.9745585918426514
 Step 580: loss 1.9593708515167236
 Step 590: loss 2.0259594917297363
 Step 600: loss 2.0305333137512207
 [Seed 65] Step 600 --> val loss 2.207592725753784
 Step 610: loss 2.087667465209961
 Step 620: loss 1.95512855052948
 Step 630: loss 2.0097782611846924
 Step 640: loss 2.1565358638763428
 Step 650: loss 2.021760940551758
 Step 660: loss 2.0889875888824463
 Step 670: loss 1.880337119102478
 Step 680: loss 1.875753402709961
 Step 690: loss 1.9182240962982178
 Step 700: loss 1.912014365196228
 Step 710: loss 2.054438591003418
 Step 720: loss 1.9575040340423584
 Step 730: loss 1.9765790700912476
 Step 740: loss 2.10176420211792
 Step 750: loss 2.2217049598693848
 Step 760: loss 2.03202486038208
 Step 770: loss 2.036588191986084
 Step 780: loss 2.067077159881592
 Step 790: loss 2.0143885612487793
 Step 800: loss 2.0535221099853516
 [Seed 65] Step 800 --> val loss 2.1672511100769043
 Step 810: loss 1.9614286422729492
 Step 820: loss 1.903390645980835
 Step 830: loss 2.044548988342285
 Step 840: loss 1.881439447402954
 Step 850: loss 1.9064884185791016
 Step 860: loss 2.06478214263916
 Step 870: loss 2.012287139892578
 Step 880: loss 2.089324712753296
 Step 890: loss 1.980396032333374
 Step 900: loss 1.832571029663086
 Step 910: loss 1.9666566848754883
 Step 920: loss 2.159853219985962
 Step 930: loss 1.9623297452926636
 Step 940: loss 1.981981635093689
 Step 950: loss 2.115365505218506
 Step 960: loss 2.021879196166992
 Step 970: loss 1.9016237258911133
 Step 980: loss 2.026550531387329
 Step 990: loss 1.9518258571624756
 Step 1000: loss 2.127326011657715
 [Seed 65] Step 1000 --> val loss 2.138127326965332
 Step 1010: loss 1.9289665222167969
 Step 1020: loss 1.9016799926757812
 Step 1030: loss 2.010991096496582
 Step 1040: loss 2.041637420654297
 Step 1050: loss 1.9826182126998901
 Step 1060: loss 1.9919278621673584
 Step 1070: loss 2.0282657146453857
 Step 1080: loss 1.954741358757019
 Step 1090: loss nan
 Step 1100: loss nan
 Step 1110: loss nan
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 65] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 65] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 65] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 65] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 65] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [46 47 57  1 41 53 59 56 58  0] ...
 First target sequence: [47 57  1 41 53 59 56 58  0 20] ...

 Decoded input: his court
 Hides not his visage from our cottage but
 Looks on ali

 Decoded target: is court
 Hides not his visage from our cottage but
 Looks on alik
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.213371753692627
 [Seed 66] Step 0 --> val loss 4.18611478805542
 Step 10: loss 3.851254940032959
 Step 20: loss 3.295447826385498
 Step 30: loss 2.8711414337158203
 Step 40: loss 2.5801448822021484
 Step 50: loss 2.5331716537475586
 Step 60: loss 2.4547367095947266
 Step 70: loss 2.3779680728912354
 Step 80: loss 2.4663970470428467
 Step 90: loss 2.4130122661590576
 Step 100: loss 2.3310189247131348
 Step 110: loss 2.198301076889038
 Step 120: loss 2.1831278800964355
 Step 130: loss 2.2831382751464844
 Step 140: loss 2.0457100868225098
 Step 150: loss 2.15246844291687
 Step 160: loss 2.098982810974121
 Step 170: loss 2.1937432289123535
 Step 180: loss 2.1289045810699463
 Step 190: loss 2.069335699081421
 Step 200: loss 2.0786099433898926
 [Seed 66] Step 200 --> val loss 2.3370635509490967
 Step 210: loss 2.071420192718506
 Step 220: loss 2.031648874282837
 Step 230: loss 2.1389036178588867
 Step 240: loss 2.0553977489471436
 Step 250: loss 2.18304443359375
 Step 260: loss 2.0278122425079346
 Step 270: loss 2.1986043453216553
 Step 280: loss 1.8629392385482788
 Step 290: loss 2.0197219848632812
 Step 300: loss 1.9929873943328857
 Step 310: loss 2.052446126937866
 Step 320: loss 2.1370201110839844
 Step 330: loss 2.105818271636963
 Step 340: loss 2.024380683898926
 Step 350: loss 2.0374462604522705
 Step 360: loss 2.0108108520507812
 Step 370: loss 1.9557826519012451
 Step 380: loss 1.8130948543548584
 Step 390: loss 1.9668992757797241
 Step 400: loss 2.0733909606933594
 [Seed 66] Step 400 --> val loss 2.2496848106384277
 Step 410: loss 2.013867139816284
 Step 420: loss 1.9174010753631592
 Step 430: loss 2.1067705154418945
 Step 440: loss 1.9654605388641357
 Step 450: loss 1.995877981185913
 Step 460: loss 1.950432538986206
 Step 470: loss 2.076510190963745
 Step 480: loss 2.1541998386383057
 Step 490: loss 2.0130271911621094
 Step 500: loss 2.041722297668457
 Step 510: loss 2.030970573425293
 Step 520: loss 1.9158799648284912
 Step 530: loss 2.093728542327881
 Step 540: loss 1.9744844436645508
 Step 550: loss 2.014625310897827
 Step 560: loss 2.0036513805389404
 Step 570: loss 2.094777822494507
 Step 580: loss 2.0747251510620117
 Step 590: loss 1.9865964651107788
 Step 600: loss 1.9701991081237793
 [Seed 66] Step 600 --> val loss 2.339104652404785
 Step 610: loss 2.104769468307495
 Step 620: loss 2.137923002243042
 Step 630: loss 2.103148937225342
 Step 640: loss 1.9884660243988037
 Step 650: loss 1.9806703329086304
 Step 660: loss 2.044508457183838
 Step 670: loss 1.9965218305587769
 Step 680: loss 2.084455966949463
 Step 690: loss 2.0503926277160645
 Step 700: loss 1.902269721031189
 Step 710: loss 1.972590684890747
 Step 720: loss 2.0628623962402344
 Step 730: loss 2.0303971767425537
 Step 740: loss 1.9908422231674194
 Step 750: loss 1.9791152477264404
 Step 760: loss 1.942327857017517
 Step 770: loss 2.0751028060913086
 Step 780: loss 2.0185623168945312
 Step 790: loss 1.9402107000350952
 Step 800: loss 2.0226025581359863
 [Seed 66] Step 800 --> val loss 2.3263161182403564
 Step 810: loss 1.8609845638275146
 Step 820: loss 2.073108673095703
 Step 830: loss 1.9877240657806396
 Step 840: loss 2.043008804321289
 Step 850: loss 2.0370230674743652
 Step 860: loss 2.1227824687957764
 Step 870: loss 2.116497278213501
 Step 880: loss 1.991537094116211
 Step 890: loss 2.1363608837127686
 Step 900: loss 1.9928350448608398
 Step 910: loss 1.994571566581726
 Step 920: loss 1.9611599445343018
 Step 930: loss 1.975909948348999
 Step 940: loss 2.061475992202759
 Step 950: loss 2.0362353324890137
 Step 960: loss 1.9582514762878418
 Step 970: loss 2.0394797325134277
 Step 980: loss 2.1706933975219727
 Step 990: loss 1.8920167684555054
 Step 1000: loss 2.086171865463257
 [Seed 66] Step 1000 --> val loss 2.3327014446258545
 Step 1010: loss 2.013641357421875
 Step 1020: loss 2.122189998626709
 Step 1030: loss 2.0206305980682373
 Step 1040: loss 2.054409980773926
 Step 1050: loss 2.1060431003570557
 Step 1060: loss 1.9117873907089233
 Step 1070: loss 2.110088348388672
 Step 1080: loss 2.004378318786621
 Step 1090: loss 2.036238670349121
 Step 1100: loss 1.9707958698272705
 Step 1110: loss 2.0444467067718506
 Step 1120: loss 1.9362691640853882
 Step 1130: loss 1.99711012840271
 Step 1140: loss 1.9244837760925293
 Step 1150: loss 2.058849334716797
 Step 1160: loss 1.998671531677246
 Step 1170: loss 2.0121922492980957
 Step 1180: loss 1.9858033657073975
 Step 1190: loss 2.063274621963501
 Step 1200: loss 1.9341213703155518
 [Seed 66] Step 1200 --> val loss 2.139155149459839
 Step 1210: loss 1.908725619316101
 Step 1220: loss 1.9038326740264893
 Step 1230: loss 1.9714850187301636
 Step 1240: loss 2.047420024871826
 Step 1250: loss 2.058958053588867
 Step 1260: loss 1.8774921894073486
 Step 1270: loss 1.9076436758041382
 Step 1280: loss 1.9578866958618164
 Step 1290: loss 1.8515784740447998
 Step 1300: loss 1.891361951828003
 Step 1310: loss 1.9957126379013062
 Step 1320: loss 1.8947231769561768
 Step 1330: loss 1.9716460704803467
 Step 1340: loss 1.980950117111206
 Step 1350: loss 1.9698565006256104
 Step 1360: loss 1.9632160663604736
 Step 1370: loss 1.9966462850570679
 Step 1380: loss 1.9597837924957275
 Step 1390: loss 1.9270652532577515
 Step 1400: loss 1.915771484375
 [Seed 66] Step 1400 --> val loss 2.051543712615967
 Step 1410: loss 1.8883298635482788
 Step 1420: loss 1.918398380279541
 Step 1430: loss 1.9908822774887085
 Step 1440: loss 1.9919967651367188
 Step 1450: loss 1.9284188747406006
 Step 1460: loss 1.9085882902145386
 Step 1470: loss 1.9711298942565918
 Step 1480: loss 1.9493260383605957
 Step 1490: loss 1.8474200963974
 Step 1500: loss 1.9030792713165283
 Step 1510: loss 1.9138597249984741
 Step 1520: loss 1.829356074333191
 Step 1530: loss 1.9670376777648926
 Step 1540: loss 1.7900432348251343
 Step 1550: loss 1.6996257305145264
 Step 1560: loss 1.9329532384872437
 Step 1570: loss 1.965175986289978
 Step 1580: loss 1.8825044631958008
 Step 1590: loss 1.875516653060913
 Step 1600: loss 1.7875916957855225
 [Seed 66] Step 1600 --> val loss 2.05877947807312
 Step 1610: loss 1.928910732269287
 Step 1620: loss 1.8906011581420898
 Step 1630: loss 1.7845948934555054
 Step 1640: loss 1.8347762823104858
 Step 1650: loss 1.7890491485595703
 Step 1660: loss 1.8092126846313477
 Step 1670: loss 1.8168087005615234
 Step 1680: loss 1.9093592166900635
 Step 1690: loss 1.8314412832260132
 Step 1700: loss 1.9058631658554077
 Step 1710: loss 1.681776762008667
 Step 1720: loss 1.761878252029419
 Step 1730: loss 1.8204816579818726
 Step 1740: loss 1.803593397140503
 Step 1750: loss 1.8007529973983765
 Step 1760: loss 1.8159170150756836
 Step 1770: loss 1.9386271238327026
 Step 1780: loss 1.8270014524459839
 Step 1790: loss 1.785867691040039
 Step 1800: loss 1.7002291679382324
 [Seed 66] Step 1800 --> val loss 1.9884014129638672
 Step 1810: loss 1.8326289653778076
 Step 1820: loss 1.742725133895874
 Step 1830: loss 1.8103197813034058
 Step 1840: loss 1.7077956199645996
 Step 1850: loss 1.7930314540863037
 Step 1860: loss 1.6282377243041992
 Step 1870: loss 1.7978639602661133
 Step 1880: loss 1.7539048194885254
 Step 1890: loss 1.7481753826141357
 Step 1900: loss 1.6727584600448608
 Step 1910: loss 1.868695616722107
 Step 1920: loss 1.7783821821212769
 Step 1930: loss 1.7636399269104004
 Step 1940: loss 1.7614266872406006
 Step 1950: loss 1.7232611179351807
 Step 1960: loss 1.6708569526672363
 Step 1970: loss 1.7117713689804077
 Step 1980: loss 1.5580673217773438
 Step 1990: loss 1.8066948652267456
 Step 2000: loss 1.824476957321167
 [Seed 66] Step 2000 --> val loss 1.9348175525665283
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [59 56  1 56 43 51 43 51 40 56] ...
 First target sequence: [56  1 56 43 51 43 51 40 56 39] ...

 Decoded input: ur remembrances: but you have found,
 Scaling his present bearing

 Decoded target: r remembrances: but you have found,
 Scaling his present bearing 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.220024108886719
 [Seed 67] Step 0 --> val loss 4.184690952301025
 Step 10: loss 3.8529841899871826
 Step 20: loss 3.3398749828338623
 Step 30: loss 2.9402663707733154
 Step 40: loss 2.740999221801758
 Step 50: loss 2.5733277797698975
 Step 60: loss 2.449295997619629
 Step 70: loss 2.4338250160217285
 Step 80: loss 2.330308437347412
 Step 90: loss 2.337261915206909
 Step 100: loss 2.232053279876709
 Step 110: loss 2.198281764984131
 Step 120: loss 2.257370948791504
 Step 130: loss 2.156583309173584
 Step 140: loss 2.206456184387207
 Step 150: loss 2.2381601333618164
 Step 160: loss 2.173661231994629
 Step 170: loss 2.100656509399414
 Step 180: loss 2.086679697036743
 Step 190: loss 2.20849871635437
 Step 200: loss 1.9671077728271484
 [Seed 67] Step 200 --> val loss 2.3009448051452637
 Step 210: loss 2.098667621612549
 Step 220: loss 2.10178542137146
 Step 230: loss 2.0217974185943604
 Step 240: loss 2.0684986114501953
 Step 250: loss 2.100287914276123
 Step 260: loss 2.1175732612609863
 Step 270: loss 2.0771965980529785
 Step 280: loss 2.1244354248046875
 Step 290: loss 2.084906578063965
 Step 300: loss 2.0487024784088135
 Step 310: loss 1.9948680400848389
 Step 320: loss 2.092538833618164
 Step 330: loss 1.9169275760650635
 Step 340: loss 2.058804988861084
 Step 350: loss 2.0553014278411865
 Step 360: loss 1.8540618419647217
 Step 370: loss 2.06953763961792
 Step 380: loss 1.942975401878357
 Step 390: loss 2.077237606048584
 Step 400: loss 2.004135847091675
 [Seed 67] Step 400 --> val loss 2.208024263381958
 Step 410: loss 1.9132986068725586
 Step 420: loss 1.9849307537078857
 Step 430: loss 1.958890438079834
 Step 440: loss 1.9195268154144287
 Step 450: loss 2.0340566635131836
 Step 460: loss 2.11660099029541
 Step 470: loss 1.976999044418335
 Step 480: loss 1.9606897830963135
 Step 490: loss 2.0353639125823975
 Step 500: loss 2.041057825088501
 Step 510: loss 2.100058078765869
 Step 520: loss 2.146406650543213
 Step 530: loss 1.8905423879623413
 Step 540: loss 1.881969690322876
 Step 550: loss 2.0509159564971924
 Step 560: loss 1.8902801275253296
 Step 570: loss 1.8522084951400757
 Step 580: loss 2.0773727893829346
 Step 590: loss 2.1304755210876465
 Step 600: loss 1.9728182554244995
 [Seed 67] Step 600 --> val loss 2.285397529602051
 Step 610: loss 2.077526807785034
 Step 620: loss 2.0274624824523926
 Step 630: loss 2.0069308280944824
 Step 640: loss 1.9631894826889038
 Step 650: loss 1.8923585414886475
 Step 660: loss 1.8951520919799805
 Step 670: loss 1.97910475730896
 Step 680: loss 1.9091129302978516
 Step 690: loss 2.1543970108032227
 Step 700: loss 2.0488595962524414
 Step 710: loss 2.008016586303711
 Step 720: loss 1.9406394958496094
 Step 730: loss 1.9375872611999512
 Step 740: loss 2.0437262058258057
 Step 750: loss 1.9887136220932007
 Step 760: loss 2.0286078453063965
 Step 770: loss 1.9034976959228516
 Step 780: loss 2.0597352981567383
 Step 790: loss 2.0682358741760254
 Step 800: loss 2.0015742778778076
 [Seed 67] Step 800 --> val loss 2.258631944656372
 Step 810: loss 1.9502112865447998
 Step 820: loss 2.0429415702819824
 Step 830: loss 1.9755643606185913
 Step 840: loss 1.973737120628357
 Step 850: loss 1.991510272026062
 Step 860: loss 1.9702037572860718
 Step 870: loss 1.9498586654663086
 Step 880: loss 1.9954683780670166
 Step 890: loss 1.9649555683135986
 Step 900: loss 1.9717509746551514
 Step 910: loss 2.0065689086914062
 Step 920: loss 1.9999053478240967
 Step 930: loss 2.095465660095215
 Step 940: loss 2.0127501487731934
 Step 950: loss 2.1474406719207764
 Step 960: loss 2.0792579650878906
 Step 970: loss 2.0027780532836914
 Step 980: loss 2.0573031902313232
 Step 990: loss 2.072669744491577
 Step 1000: loss 2.0433263778686523
 [Seed 67] Step 1000 --> val loss 2.2721590995788574
 Step 1010: loss 2.0012753009796143
 Step 1020: loss 2.07326340675354
 Step 1030: loss 1.9062862396240234
 Step 1040: loss 2.014233112335205
 Step 1050: loss 1.9825185537338257
 Step 1060: loss 2.0023648738861084
 Step 1070: loss 1.990515112876892
 Step 1080: loss 2.004537582397461
 Step 1090: loss 1.9455384016036987
 Step 1100: loss 1.9363969564437866
 Step 1110: loss 1.952773094177246
 Step 1120: loss 1.9227657318115234
 Step 1130: loss 1.9350684881210327
 Step 1140: loss 2.085163116455078
 Step 1150: loss 2.036613941192627
 Step 1160: loss 2.0826663970947266
 Step 1170: loss 2.0213801860809326
 Step 1180: loss 1.9578273296356201
 Step 1190: loss 1.9215867519378662
 Step 1200: loss 1.852774977684021
 [Seed 67] Step 1200 --> val loss 2.158553123474121
 Step 1210: loss 2.0416557788848877
 Step 1220: loss 1.9204726219177246
 Step 1230: loss 1.8785346746444702
 Step 1240: loss 2.0090484619140625
 Step 1250: loss 1.951619029045105
 Step 1260: loss 1.962530493736267
 Step 1270: loss 1.9878947734832764
 Step 1280: loss 1.974125623703003
 Step 1290: loss 1.9456701278686523
 Step 1300: loss 1.9329421520233154
 Step 1310: loss 2.054732322692871
 Step 1320: loss 1.9657857418060303
 Step 1330: loss 1.9450445175170898
 Step 1340: loss 1.9124339818954468
 Step 1350: loss 1.953980803489685
 Step 1360: loss 1.7716290950775146
 Step 1370: loss 1.8909472227096558
 Step 1380: loss 1.9597270488739014
 Step 1390: loss 1.9789257049560547
 Step 1400: loss 1.8608245849609375
 [Seed 67] Step 1400 --> val loss 2.0433948040008545
 Step 1410: loss 1.8776277303695679
 Step 1420: loss 2.010867118835449
 Step 1430: loss 1.8671436309814453
 Step 1440: loss 1.9341940879821777
 Step 1450: loss 1.8084958791732788
 Step 1460: loss 1.8006720542907715
 Step 1470: loss 1.888555884361267
 Step 1480: loss 2.0052449703216553
 Step 1490: loss 1.867178201675415
 Step 1500: loss 1.9658095836639404
 Step 1510: loss 1.8979133367538452
 Step 1520: loss 1.8933045864105225
 Step 1530: loss 1.8687604665756226
 Step 1540: loss 1.8052241802215576
 Step 1550: loss 1.8384802341461182
 Step 1560: loss 1.9373000860214233
 Step 1570: loss 1.880476951599121
 Step 1580: loss 1.8497695922851562
 Step 1590: loss 1.9449846744537354
 Step 1600: loss 1.8237733840942383
 [Seed 67] Step 1600 --> val loss 2.025282621383667
 Step 1610: loss 1.9089345932006836
 Step 1620: loss 1.9403867721557617
 Step 1630: loss 1.8422110080718994
 Step 1640: loss 1.6841439008712769
 Step 1650: loss 1.8896405696868896
 Step 1660: loss 1.793825626373291
 Step 1670: loss 1.8595890998840332
 Step 1680: loss 1.7694181203842163
 Step 1690: loss 1.860985517501831
 Step 1700: loss 1.8069804906845093
 Step 1710: loss 1.868067741394043
 Step 1720: loss 1.8682498931884766
 Step 1730: loss 1.7552069425582886
 Step 1740: loss 1.7333464622497559
 Step 1750: loss 1.8535301685333252
 Step 1760: loss 1.7754229307174683
 Step 1770: loss 1.8402471542358398
 Step 1780: loss 1.659014105796814
 Step 1790: loss 1.6069245338439941
 Step 1800: loss 1.8342188596725464
 [Seed 67] Step 1800 --> val loss 1.9802058935165405
 Step 1810: loss 1.8314313888549805
 Step 1820: loss 1.8265074491500854
 Step 1830: loss 1.8630757331848145
 Step 1840: loss 1.6217772960662842
 Step 1850: loss 1.8233617544174194
 Step 1860: loss 1.8205041885375977
 Step 1870: loss 1.79868483543396
 Step 1880: loss 1.775492548942566
 Step 1890: loss 1.7869162559509277
 Step 1900: loss 1.7453464269638062
 Step 1910: loss 1.7771881818771362
 Step 1920: loss 1.767006278038025
 Step 1930: loss 1.6788333654403687
 Step 1940: loss 1.7819780111312866
 Step 1950: loss 1.744585394859314
 Step 1960: loss 1.6118993759155273
 Step 1970: loss 1.7273814678192139
 Step 1980: loss 1.6933279037475586
 Step 1990: loss 1.767526626586914
 Step 2000: loss 1.9577586650848389
 [Seed 67] Step 2000 --> val loss 1.8864034414291382
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [58  1 46 39 60 43  1 51 43  1] ...
 First target sequence: [ 1 46 39 60 43  1 51 43  1 61] ...

 Decoded input: t have me weep? why, now thou hast thy will:
 For raging wind blo

 Decoded target:  have me weep? why, now thou hast thy will:
 For raging wind blow
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2213287353515625
 [Seed 68] Step 0 --> val loss 4.182828426361084
 Step 10: loss 3.834897518157959
 Step 20: loss 3.3515048027038574
 Step 30: loss 2.800602674484253
 Step 40: loss 2.663729667663574
 Step 50: loss 2.5706846714019775
 Step 60: loss 2.4765121936798096
 Step 70: loss 2.402374267578125
 Step 80: loss 2.374157190322876
 Step 90: loss 2.3841264247894287
 Step 100: loss 2.3958144187927246
 Step 110: loss 2.3960909843444824
 Step 120: loss 2.1866044998168945
 Step 130: loss 2.3018083572387695
 Step 140: loss 2.177335262298584
 Step 150: loss 2.102296829223633
 Step 160: loss 2.017906665802002
 Step 170: loss 2.12910795211792
 Step 180: loss 2.1068665981292725
 Step 190: loss 2.2208986282348633
 Step 200: loss 2.1411566734313965
 [Seed 68] Step 200 --> val loss 2.3936898708343506
 Step 210: loss 2.1051337718963623
 Step 220: loss 2.0052647590637207
 Step 230: loss 1.9594342708587646
 Step 240: loss 2.0440804958343506
 Step 250: loss 2.0581111907958984
 Step 260: loss 2.063239574432373
 Step 270: loss 2.1362619400024414
 Step 280: loss 2.038158893585205
 Step 290: loss 2.0150957107543945
 Step 300: loss 2.0672733783721924
 Step 310: loss 2.014988899230957
 Step 320: loss 1.940262794494629
 Step 330: loss 2.0330042839050293
 Step 340: loss 1.962461233139038
 Step 350: loss 1.9287440776824951
 Step 360: loss 2.12104868888855
 Step 370: loss 1.9695508480072021
 Step 380: loss 1.9751689434051514
 Step 390: loss 1.9436213970184326
 Step 400: loss 1.9011157751083374
 [Seed 68] Step 400 --> val loss 2.239739418029785
 Step 410: loss 1.976099967956543
 Step 420: loss 2.1048407554626465
 Step 430: loss 1.9359302520751953
 Step 440: loss 2.0194883346557617
 Step 450: loss 1.9686803817749023
 Step 460: loss 2.029042959213257
 Step 470: loss 2.0165669918060303
 Step 480: loss 2.0540976524353027
 Step 490: loss 1.9894349575042725
 Step 500: loss 1.9880393743515015
 Step 510: loss 2.108618974685669
 Step 520: loss 2.037209987640381
 Step 530: loss 2.0791375637054443
 Step 540: loss 2.018024444580078
 Step 550: loss 2.027940273284912
 Step 560: loss 2.041682720184326
 Step 570: loss 2.0229156017303467
 Step 580: loss 2.0779924392700195
 Step 590: loss 1.954125165939331
 Step 600: loss 1.985662817955017
 [Seed 68] Step 600 --> val loss 2.3131792545318604
 Step 610: loss 1.95876145362854
 Step 620: loss 2.0151476860046387
 Step 630: loss 2.075279712677002
 Step 640: loss 2.0392589569091797
 Step 650: loss 2.0685625076293945
 Step 660: loss 2.0192346572875977
 Step 670: loss 2.0448577404022217
 Step 680: loss 2.0976786613464355
 Step 690: loss 1.9059739112854004
 Step 700: loss 1.9854927062988281
 Step 710: loss 2.029784917831421
 Step 720: loss 2.044703960418701
 Step 730: loss 2.034809112548828
 Step 740: loss 2.1814165115356445
 Step 750: loss 1.8245309591293335
 Step 760: loss 2.0915660858154297
 Step 770: loss 1.9476890563964844
 Step 780: loss 2.043562650680542
 Step 790: loss 2.0319032669067383
 Step 800: loss 2.007493019104004
 [Seed 68] Step 800 --> val loss 2.258225679397583
 Step 810: loss 2.0235402584075928
 Step 820: loss 2.06183123588562
 Step 830: loss 2.1308038234710693
 Step 840: loss 2.088886260986328
 Step 850: loss 2.0010154247283936
 Step 860: loss 2.0496742725372314
 Step 870: loss 1.941884994506836
 Step 880: loss 2.100160598754883
 Step 890: loss 1.985848307609558
 Step 900: loss 1.9512799978256226
 Step 910: loss 2.1302077770233154
 Step 920: loss 2.0131356716156006
 Step 930: loss 2.055661678314209
 Step 940: loss 2.1787667274475098
 Step 950: loss 2.0744376182556152
 Step 960: loss 1.916365385055542
 Step 970: loss 1.9533255100250244
 Step 980: loss 1.9484117031097412
 Step 990: loss 2.1080727577209473
 Step 1000: loss 2.0036869049072266
 [Seed 68] Step 1000 --> val loss 2.30940318107605
 Step 1010: loss 2.060149669647217
 Step 1020: loss 1.9923133850097656
 Step 1030: loss 1.9048666954040527
 Step 1040: loss 2.056096076965332
 Step 1050: loss 2.0071778297424316
 Step 1060: loss 2.035524845123291
 Step 1070: loss 1.8839175701141357
 Step 1080: loss 1.9804258346557617
 Step 1090: loss 1.9752507209777832
 Step 1100: loss 1.9267023801803589
 Step 1110: loss 2.0577468872070312
 Step 1120: loss 1.9872891902923584
 Step 1130: loss 1.9626240730285645
 Step 1140: loss 2.0247983932495117
 Step 1150: loss 1.9566221237182617
 Step 1160: loss 1.9698896408081055
 Step 1170: loss 1.9813272953033447
 Step 1180: loss 1.9675989151000977
 Step 1190: loss 1.9026955366134644
 Step 1200: loss 1.9809684753417969
 [Seed 68] Step 1200 --> val loss 2.2125792503356934
 Step 1210: loss 2.0622060298919678
 Step 1220: loss 1.921940803527832
 Step 1230: loss 1.8930768966674805
 Step 1240: loss 1.9491941928863525
 Step 1250: loss 1.9892282485961914
 Step 1260: loss 2.078537702560425
 Step 1270: loss 1.9003486633300781
 Step 1280: loss 1.8943194150924683
 Step 1290: loss 1.8741686344146729
 Step 1300: loss 2.1961655616760254
 Step 1310: loss 1.8676660060882568
 Step 1320: loss 1.976014494895935
 Step 1330: loss 1.9281563758850098
 Step 1340: loss 1.859097957611084
 Step 1350: loss 1.9611284732818604
 Step 1360: loss 1.850562572479248
 Step 1370: loss 1.8691319227218628
 Step 1380: loss 1.812075138092041
 Step 1390: loss 1.8608993291854858
 Step 1400: loss 1.8631566762924194
 [Seed 68] Step 1400 --> val loss 2.1573991775512695
 Step 1410: loss 1.9078291654586792
 Step 1420: loss 1.8985493183135986
 Step 1430: loss 1.9238300323486328
 Step 1440: loss 1.910536766052246
 Step 1450: loss 1.8509042263031006
 Step 1460: loss 1.9203827381134033
 Step 1470: loss 1.8870084285736084
 Step 1480: loss 1.8032106161117554
 Step 1490: loss 1.9752683639526367
 Step 1500: loss 1.8670591115951538
 Step 1510: loss 1.922690749168396
 Step 1520: loss 1.8181203603744507
 Step 1530: loss 1.8515007495880127
 Step 1540: loss 1.8663733005523682
 Step 1550: loss 1.7973874807357788
 Step 1560: loss 1.7849825620651245
 Step 1570: loss 1.9023122787475586
 Step 1580: loss 1.8381599187850952
 Step 1590: loss 1.8965535163879395
 Step 1600: loss 1.8285096883773804
 [Seed 68] Step 1600 --> val loss 2.08267879486084
 Step 1610: loss 1.85982084274292
 Step 1620: loss 1.8923228979110718
 Step 1630: loss 1.7279332876205444
 Step 1640: loss 1.867204189300537
 Step 1650: loss 1.8468263149261475
 Step 1660: loss 1.7199376821517944
 Step 1670: loss 1.7426939010620117
 Step 1680: loss 1.7624043226242065
 Step 1690: loss 1.655434012413025
 Step 1700: loss 1.568788766860962
 Step 1710: loss 1.8068712949752808
 Step 1720: loss 1.7572946548461914
 Step 1730: loss 1.6914262771606445
 Step 1740: loss 1.703956961631775
 Step 1750: loss 1.8427114486694336
 Step 1760: loss 1.717574119567871
 Step 1770: loss 1.7124443054199219
 Step 1780: loss 1.8254469633102417
 Step 1790: loss 1.7336289882659912
 Step 1800: loss 1.7366087436676025
 [Seed 68] Step 1800 --> val loss 2.0111682415008545
 Step 1810: loss 1.9277814626693726
 Step 1820: loss 1.723200798034668
 Step 1830: loss 1.837148666381836
 Step 1840: loss 1.7781555652618408
 Step 1850: loss 1.79052734375
 Step 1860: loss 1.6786832809448242
 Step 1870: loss 1.6725753545761108
 Step 1880: loss 1.7694591283798218
 Step 1890: loss 1.6494064331054688
 Step 1900: loss 1.7766952514648438
 Step 1910: loss 1.8283189535140991
 Step 1920: loss 1.737683653831482
 Step 1930: loss 1.6752389669418335
 Step 1940: loss 1.779161810874939
 Step 1950: loss 1.7059221267700195
 Step 1960: loss 1.7345218658447266
 Step 1970: loss 1.6934330463409424
 Step 1980: loss 1.6015775203704834
 Step 1990: loss 1.6448004245758057
 Step 2000: loss 1.5766067504882812
 [Seed 68] Step 2000 --> val loss 1.9657562971115112
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [39 51 54 50 43 56  1 57 58 56] ...
 First target sequence: [51 54 50 43 56  1 57 58 56 43] ...

 Decoded input: ampler strength indeed
 Than most have of his age.

 POLIXENES:
 By

 Decoded target: mpler strength indeed
 Than most have of his age.

 POLIXENES:
 By 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.212583065032959
 [Seed 69] Step 0 --> val loss 4.18471097946167
 Step 10: loss 3.8548789024353027
 Step 20: loss 3.338606595993042
 Step 30: loss 2.8400402069091797
 Step 40: loss 2.726663112640381
 Step 50: loss 2.590207815170288
 Step 60: loss 2.4584484100341797
 Step 70: loss 2.4921388626098633
 Step 80: loss 2.337197780609131
 Step 90: loss 2.1934659481048584
 Step 100: loss 2.2249526977539062
 Step 110: loss 2.2006635665893555
 Step 120: loss 2.1929588317871094
 Step 130: loss 2.1551976203918457
 Step 140: loss 2.108409881591797
 Step 150: loss 2.0858676433563232
 Step 160: loss 2.121710777282715
 Step 170: loss 2.152714729309082
 Step 180: loss 2.1335926055908203
 Step 190: loss 1.9833873510360718
 Step 200: loss 2.1322569847106934
 [Seed 69] Step 200 --> val loss 2.314197301864624
 Step 210: loss 2.1102426052093506
 Step 220: loss 2.1930184364318848
 Step 230: loss 2.1058664321899414
 Step 240: loss 2.0054569244384766
 Step 250: loss 2.151242733001709
 Step 260: loss 2.062338352203369
 Step 270: loss 2.0667569637298584
 Step 280: loss 2.081984519958496
 Step 290: loss 1.948075294494629
 Step 300: loss 2.0534920692443848
 Step 310: loss 2.060516834259033
 Step 320: loss 2.0660862922668457
 Step 330: loss 2.086667060852051
 Step 340: loss 2.1001651287078857
 Step 350: loss 1.991476058959961
 Step 360: loss 2.0175905227661133
 Step 370: loss 2.104642391204834
 Step 380: loss 2.094939947128296
 Step 390: loss 2.063948631286621
 Step 400: loss 2.159818649291992
 [Seed 69] Step 400 --> val loss 2.23774790763855
 Step 410: loss 1.9638125896453857
 Step 420: loss 2.0127792358398438
 Step 430: loss 2.023454189300537
 Step 440: loss 1.9445242881774902
 Step 450: loss 1.9800827503204346
 Step 460: loss 2.062561273574829
 Step 470: loss 1.9539318084716797
 Step 480: loss 1.9820785522460938
 Step 490: loss 2.080716609954834
 Step 500: loss 1.9415873289108276
 Step 510: loss 2.0903372764587402
 Step 520: loss 2.0457301139831543
 Step 530: loss 1.9923162460327148
 Step 540: loss 1.8410069942474365
 Step 550: loss 2.067819595336914
 Step 560: loss 2.079399585723877
 Step 570: loss 2.0860557556152344
 Step 580: loss 2.07837176322937
 Step 590: loss 1.9890050888061523
 Step 600: loss 1.9580949544906616
 [Seed 69] Step 600 --> val loss 2.2876837253570557
 Step 610: loss 2.054720640182495
 Step 620: loss 2.085482120513916
 Step 630: loss 2.0847249031066895
 Step 640: loss 2.1682586669921875
 Step 650: loss 1.9543771743774414
 Step 660: loss 2.0805563926696777
 Step 670: loss 2.104309320449829
 Step 680: loss 1.9811968803405762
 Step 690: loss 2.077530860900879
 Step 700: loss 1.9790618419647217
 Step 710: loss 2.19032621383667
 Step 720: loss 2.084811210632324
 Step 730: loss 2.120972156524658
 Step 740: loss 2.1199800968170166
 Step 750: loss 2.1233737468719482
 Step 760: loss 2.2005815505981445
 Step 770: loss 2.0100255012512207
 Step 780: loss 2.0898666381835938
 Step 790: loss 1.9855287075042725
 Step 800: loss 2.100393056869507
 [Seed 69] Step 800 --> val loss 2.3413894176483154
 Step 810: loss 2.1291885375976562
 Step 820: loss 2.0284218788146973
 Step 830: loss 2.0263590812683105
 Step 840: loss 2.0688798427581787
 Step 850: loss 2.0195529460906982
 Step 860: loss 2.053201913833618
 Step 870: loss 1.9349794387817383
 Step 880: loss 2.1008386611938477
 Step 890: loss 2.0787503719329834
 Step 900: loss 2.0914547443389893
 Step 910: loss 2.0276570320129395
 Step 920: loss 2.0225276947021484
 Step 930: loss 2.1498687267303467
 Step 940: loss 2.11678409576416
 Step 950: loss 2.1658935546875
 Step 960: loss 2.070380687713623
 Step 970: loss 2.1420397758483887
 Step 980: loss 2.052579402923584
 Step 990: loss 2.120105743408203
 Step 1000: loss 2.1544508934020996
 [Seed 69] Step 1000 --> val loss 2.386389970779419
 Step 1010: loss 2.086162567138672
 Step 1020: loss 2.0828585624694824
 Step 1030: loss 2.159221649169922
 Step 1040: loss 2.0604734420776367
 Step 1050: loss 1.9770967960357666
 Step 1060: loss 2.035449504852295
 Step 1070: loss 2.1092939376831055
 Step 1080: loss 2.015380620956421
 Step 1090: loss 2.083439826965332
 Step 1100: loss 1.9450263977050781
 Step 1110: loss 1.9597909450531006
 Step 1120: loss 1.929898977279663
 Step 1130: loss 2.003533363342285
 Step 1140: loss 1.9190845489501953
 Step 1150: loss 2.0511598587036133
 Step 1160: loss 1.9042799472808838
 Step 1170: loss 1.9363431930541992
 Step 1180: loss 1.9703843593597412
 Step 1190: loss 1.9714325666427612
 Step 1200: loss 2.070423126220703
 [Seed 69] Step 1200 --> val loss 2.229598045349121
 Step 1210: loss 1.9437847137451172
 Step 1220: loss 2.0507190227508545
 Step 1230: loss 1.9277472496032715
 Step 1240: loss 1.9620399475097656
 Step 1250: loss 1.935198187828064
 Step 1260: loss 1.9959462881088257
 Step 1270: loss 1.9645965099334717
 Step 1280: loss 1.9476203918457031
 Step 1290: loss 2.0685172080993652
 Step 1300: loss 1.9909608364105225
 Step 1310: loss 1.85542893409729
 Step 1320: loss 1.946947455406189
 Step 1330: loss 1.9032397270202637
 Step 1340: loss 1.937920093536377
 Step 1350: loss 2.0337514877319336
 Step 1360: loss 1.9908092021942139
 Step 1370: loss 2.009594440460205
 Step 1380: loss 2.1196625232696533
 Step 1390: loss 1.9614605903625488
 Step 1400: loss 1.9536349773406982
 [Seed 69] Step 1400 --> val loss 2.1267764568328857
 Step 1410: loss 2.0037660598754883
 Step 1420: loss 1.8984713554382324
 Step 1430: loss 1.9571268558502197
 Step 1440: loss 1.9721215963363647
 Step 1450: loss 2.0343666076660156
 Step 1460: loss 1.97472083568573
 Step 1470: loss 1.9885518550872803
 Step 1480: loss 2.0455336570739746
 Step 1490: loss 1.8789509534835815
 Step 1500: loss 1.963295340538025
 Step 1510: loss 1.9800591468811035
 Step 1520: loss 1.9742481708526611
 Step 1530: loss 1.9567580223083496
 Step 1540: loss 1.9590768814086914
 Step 1550: loss 2.0043387413024902
 Step 1560: loss 1.8123013973236084
 Step 1570: loss 1.8998607397079468
 Step 1580: loss 1.9513295888900757
 Step 1590: loss 1.8609120845794678
 Step 1600: loss 1.9024676084518433
 [Seed 69] Step 1600 --> val loss 2.019429922103882
 Step 1610: loss 1.8303604125976562
 Step 1620: loss 1.916712999343872
 Step 1630: loss 1.7968554496765137
 Step 1640: loss 1.8458986282348633
 Step 1650: loss 1.8827588558197021
 Step 1660: loss 1.9046363830566406
 Step 1670: loss 1.7680104970932007
 Step 1680: loss 1.8791062831878662
 Step 1690: loss 1.8965692520141602
 Step 1700: loss 1.9052855968475342
 Step 1710: loss 1.8024580478668213
 Step 1720: loss 1.9152708053588867
 Step 1730: loss 1.803525447845459
 Step 1740: loss 1.8841524124145508
 Step 1750: loss 1.8666890859603882
 Step 1760: loss 1.804266333580017
 Step 1770: loss 1.8799365758895874
 Step 1780: loss 1.7381799221038818
 Step 1790: loss 1.7382686138153076
 Step 1800: loss 1.758040189743042
 [Seed 69] Step 1800 --> val loss 2.016792058944702
 Step 1810: loss 1.708099126815796
 Step 1820: loss 1.8792054653167725
 Step 1830: loss 1.7996543645858765
 Step 1840: loss 1.887997031211853
 Step 1850: loss 1.7595529556274414
 Step 1860: loss 1.8303241729736328
 Step 1870: loss 1.7193751335144043
 Step 1880: loss 1.7218230962753296
 Step 1890: loss 1.8454097509384155
 Step 1900: loss 1.8421335220336914
 Step 1910: loss 1.6612975597381592
 Step 1920: loss 1.7441508769989014
 Step 1930: loss 1.848158836364746
 Step 1940: loss 1.8543933629989624
 Step 1950: loss 1.7142890691757202
 Step 1960: loss 1.809638500213623
 Step 1970: loss 1.888602614402771
 Step 1980: loss 1.7430371046066284
 Step 1990: loss 1.8503913879394531
 Step 2000: loss 1.6801170110702515
 [Seed 69] Step 2000 --> val loss 1.9793367385864258
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [57  0 18 53 56  1 57 47 52 49] ...
 First target sequence: [ 0 18 53 56  1 57 47 52 49 47] ...

 Decoded input: s
 For sinking under them.

 SICINIUS:
 This, as you say, suggested

 Decoded target: 
 For sinking under them.

 SICINIUS:
 This, as you say, suggested

 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.208529472351074
 [Seed 70] Step 0 --> val loss 4.183833599090576
 Step 10: loss 3.8595950603485107
 Step 20: loss 3.361332654953003
 Step 30: loss 2.891284465789795
 Step 40: loss 2.7654082775115967
 Step 50: loss 2.605894088745117
 Step 60: loss 2.4300951957702637
 Step 70: loss 2.426945209503174
 Step 80: loss 2.317964553833008
 Step 90: loss 2.3493735790252686
 Step 100: loss 2.287872314453125
 Step 110: loss 2.2009177207946777
 Step 120: loss 2.2257018089294434
 Step 130: loss 2.238964796066284
 Step 140: loss 2.202021598815918
 Step 150: loss 2.1523900032043457
 Step 160: loss 2.1161746978759766
 Step 170: loss 2.152573347091675
 Step 180: loss 2.1457040309906006
 Step 190: loss 2.125056266784668
 Step 200: loss 2.1674070358276367
 [Seed 70] Step 200 --> val loss 2.361920118331909
 Step 210: loss 1.9674080610275269
 Step 220: loss 1.9714972972869873
 Step 230: loss 2.0029401779174805
 Step 240: loss 2.1426055431365967
 Step 250: loss 1.9934872388839722
 Step 260: loss 2.1680479049682617
 Step 270: loss 2.1539297103881836
 Step 280: loss 1.9978604316711426
 Step 290: loss 2.0527119636535645
 Step 300: loss 1.8584951162338257
 Step 310: loss 2.120398998260498
 Step 320: loss 1.9549440145492554
 Step 330: loss 2.022120952606201
 Step 340: loss 2.0604968070983887
 Step 350: loss 2.095021963119507
 Step 360: loss 2.0459108352661133
 Step 370: loss 2.029381513595581
 Step 380: loss 2.0740861892700195
 Step 390: loss 2.0307698249816895
 Step 400: loss 1.9426902532577515
 [Seed 70] Step 400 --> val loss 2.177712917327881
 Step 410: loss 2.0948338508605957
 Step 420: loss 1.9992135763168335
 Step 430: loss 2.015181064605713
 Step 440: loss 1.9875297546386719
 Step 450: loss 2.094593048095703
 Step 460: loss 2.18404483795166
 Step 470: loss 1.9854798316955566
 Step 480: loss 2.066005229949951
 Step 490: loss 1.8570226430892944
 Step 500: loss 1.9559504985809326
 Step 510: loss 2.017988920211792
 Step 520: loss 2.00756573677063
 Step 530: loss 2.1223132610321045
 Step 540: loss 2.092578172683716
 Step 550: loss 1.9290432929992676
 Step 560: loss 1.9713290929794312
 Step 570: loss 1.9208691120147705
 Step 580: loss 1.9712016582489014
 Step 590: loss 1.924241065979004
 Step 600: loss 1.9709434509277344
 [Seed 70] Step 600 --> val loss 2.218740463256836
 Step 610: loss 1.8958057165145874
 Step 620: loss 1.995544195175171
 Step 630: loss 1.9693915843963623
 Step 640: loss 2.0945286750793457
 Step 650: loss 2.1516242027282715
 Step 660: loss 2.030587911605835
 Step 670: loss 1.9633684158325195
 Step 680: loss 1.8806238174438477
 Step 690: loss 1.9156564474105835
 Step 700: loss 1.9639699459075928
 Step 710: loss 1.9610155820846558
 Step 720: loss 1.9854720830917358
 Step 730: loss 2.054810047149658
 Step 740: loss 1.997262954711914
 Step 750: loss 2.032616138458252
 Step 760: loss 1.9829826354980469
 Step 770: loss 2.192471504211426
 Step 780: loss 2.037372350692749
 Step 790: loss 2.0267767906188965
 Step 800: loss 2.099213123321533
 [Seed 70] Step 800 --> val loss 2.2103676795959473
 Step 810: loss 1.9618747234344482
 Step 820: loss 2.1151986122131348
 Step 830: loss 1.9958304166793823
 Step 840: loss 2.0953292846679688
 Step 850: loss 1.9238786697387695
 Step 860: loss 1.935339331626892
 Step 870: loss 1.9306211471557617
 Step 880: loss 1.98093581199646
 Step 890: loss 1.96649169921875
 Step 900: loss 1.9509700536727905
 Step 910: loss 1.8863877058029175
 Step 920: loss 2.001175880432129
 Step 930: loss 1.9476855993270874
 Step 940: loss 1.9296445846557617
 Step 950: loss 2.0302236080169678
 Step 960: loss 2.0176262855529785
 Step 970: loss 2.067610025405884
 Step 980: loss 2.0183372497558594
 Step 990: loss 1.9503800868988037
 Step 1000: loss 2.0988965034484863
 [Seed 70] Step 1000 --> val loss 2.1496026515960693
 Step 1010: loss 2.0115156173706055
 Step 1020: loss 2.0033555030822754
 Step 1030: loss 1.9871561527252197
 Step 1040: loss 2.0805540084838867
 Step 1050: loss 1.986197829246521
 Step 1060: loss 1.9981870651245117
 Step 1070: loss 2.1147384643554688
 Step 1080: loss 2.0723321437835693
 Step 1090: loss 2.0067501068115234
 Step 1100: loss 1.9615367650985718
 Step 1110: loss 1.9975566864013672
 Step 1120: loss 2.0691077709198
 Step 1130: loss 1.9959449768066406
 Step 1140: loss 1.9867663383483887
 Step 1150: loss 1.9887921810150146
 Step 1160: loss 2.0035781860351562
 Step 1170: loss 1.970785140991211
 Step 1180: loss 1.891100287437439
 Step 1190: loss 2.2113006114959717
 Step 1200: loss 2.0020737648010254
 [Seed 70] Step 1200 --> val loss 2.1531448364257812
 Step 1210: loss 2.0683302879333496
 Step 1220: loss 2.173348903656006
 Step 1230: loss 1.9825456142425537
 Step 1240: loss 1.886303186416626
 Step 1250: loss 1.8871062994003296
 Step 1260: loss 2.228236675262451
 Step 1270: loss 2.0420212745666504
 Step 1280: loss 2.0550849437713623
 Step 1290: loss 2.136049747467041
 Step 1300: loss 1.971085548400879
 Step 1310: loss 1.977518916130066
 Step 1320: loss 2.046386480331421
 Step 1330: loss 2.1013035774230957
 Step 1340: loss 2.068268299102783
 Step 1350: loss 2.031247138977051
 Step 1360: loss 1.9923958778381348
 Step 1370: loss 1.988779902458191
 Step 1380: loss 1.9526557922363281
 Step 1390: loss 2.0709025859832764
 Step 1400: loss 2.105215072631836
 [Seed 70] Step 1400 --> val loss 2.1693930625915527
 Step 1410: loss 1.953971266746521
 Step 1420: loss 1.9931868314743042
 Step 1430: loss 1.9109667539596558
 Step 1440: loss 1.9421939849853516
 Step 1450: loss 1.9843286275863647
 Step 1460: loss 1.8943850994110107
 Step 1470: loss 2.0352237224578857
 Step 1480: loss 1.9396600723266602
 Step 1490: loss 1.904235601425171
 Step 1500: loss 1.817138433456421
 Step 1510: loss 1.9924272298812866
 Step 1520: loss 1.9646142721176147
 Step 1530: loss 1.979013204574585
 Step 1540: loss 1.977921962738037
 Step 1550: loss 1.9278581142425537
 Step 1560: loss 1.989549160003662
 Step 1570: loss 1.9207100868225098
 Step 1580: loss 1.8414548635482788
 Step 1590: loss 2.0108516216278076
 Step 1600: loss 1.9134992361068726
 [Seed 70] Step 1600 --> val loss 2.131948947906494
 Step 1610: loss 1.8560831546783447
 Step 1620: loss 1.8529984951019287
 Step 1630: loss 1.7508094310760498
 Step 1640: loss 1.8956482410430908
 Step 1650: loss 1.8947676420211792
 Step 1660: loss 1.8630106449127197
 Step 1670: loss 1.893821120262146
 Step 1680: loss 1.8499033451080322
 Step 1690: loss 1.8607194423675537
 Step 1700: loss 1.978082537651062
 Step 1710: loss 1.8650249242782593
 Step 1720: loss 1.9559434652328491
 Step 1730: loss 1.8678865432739258
 Step 1740: loss 1.8689498901367188
 Step 1750: loss 1.8002294301986694
 Step 1760: loss 1.7740801572799683
 Step 1770: loss 1.881590723991394
 Step 1780: loss 1.876776933670044
 Step 1790: loss 1.7495086193084717
 Step 1800: loss 1.873056411743164
 [Seed 70] Step 1800 --> val loss 2.05989146232605
 Step 1810: loss 1.8196208477020264
 Step 1820: loss 1.899446725845337
 Step 1830: loss 1.7766655683517456
 Step 1840: loss 1.893144965171814
 Step 1850: loss 1.9262903928756714
 Step 1860: loss 1.765425682067871
 Step 1870: loss 1.8528879880905151
 Step 1880: loss 1.7426297664642334
 Step 1890: loss 1.870457410812378
 Step 1900: loss 1.8649482727050781
 Step 1910: loss 1.8209114074707031
 Step 1920: loss 1.827390432357788
 Step 1930: loss 1.8236801624298096
 Step 1940: loss 1.8491871356964111
 Step 1950: loss 1.74004065990448
 Step 1960: loss 1.786434531211853
 Step 1970: loss 1.7708661556243896
 Step 1980: loss 1.761784553527832
 Step 1990: loss 1.7640703916549683
 Step 2000: loss 1.8157247304916382
 [Seed 70] Step 2000 --> val loss 2.008516311645508
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [54 50 39 41 43  6  1 61 43  1] ...
 First target sequence: [50 39 41 43  6  1 61 43  1 42] ...

 Decoded input: place, we did commend
 To your remembrances: but you have found,


 Decoded target: lace, we did commend
 To your remembrances: but you have found,
 S
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.220525741577148
 [Seed 71] Step 0 --> val loss 4.183578968048096
 Step 10: loss 3.8429760932922363
 Step 20: loss 3.303636074066162
 Step 30: loss 2.857189655303955
 Step 40: loss 2.6239538192749023
 Step 50: loss 2.6333389282226562
 Step 60: loss 2.552983522415161
 Step 70: loss 2.4348697662353516
 Step 80: loss 2.3759336471557617
 Step 90: loss 2.3995864391326904
 Step 100: loss 2.20977783203125
 Step 110: loss 2.28916072845459
 Step 120: loss 2.319326162338257
 Step 130: loss 2.2220261096954346
 Step 140: loss 2.1750612258911133
 Step 150: loss 2.209380865097046
 Step 160: loss 2.051255941390991
 Step 170: loss 2.182232141494751
 Step 180: loss 2.126488208770752
 Step 190: loss 2.048196792602539
 Step 200: loss 2.171659469604492
 [Seed 71] Step 200 --> val loss 2.4060213565826416
 Step 210: loss 2.208456516265869
 Step 220: loss 1.975532054901123
 Step 230: loss 2.070056676864624
 Step 240: loss 2.126389980316162
 Step 250: loss 2.127335548400879
 Step 260: loss 2.070753574371338
 Step 270: loss 2.052661180496216
 Step 280: loss 2.1182456016540527
 Step 290: loss 1.9857177734375
 Step 300: loss 2.098834753036499
 Step 310: loss 2.13273024559021
 Step 320: loss 1.9965451955795288
 Step 330: loss 2.1552534103393555
 Step 340: loss 2.097961664199829
 Step 350: loss 2.038039207458496
 Step 360: loss 1.9631032943725586
 Step 370: loss 1.9451591968536377
 Step 380: loss 1.9906319379806519
 Step 390: loss 2.0400609970092773
 Step 400: loss 1.9785149097442627
 [Seed 71] Step 400 --> val loss 2.318692445755005
 Step 410: loss 2.043400526046753
 Step 420: loss 2.1400606632232666
 Step 430: loss 2.0214784145355225
 Step 440: loss 2.044271230697632
 Step 450: loss 2.0258376598358154
 Step 460: loss 2.0161144733428955
 Step 470: loss 2.0604729652404785
 Step 480: loss 2.064823627471924
 Step 490: loss 1.8757623434066772
 Step 500: loss 2.0060834884643555
 Step 510: loss 2.0831165313720703
 Step 520: loss 2.017570734024048
 Step 530: loss 1.9496171474456787
 Step 540: loss 2.0352447032928467
 Step 550: loss 2.0116896629333496
 Step 560: loss 2.0800952911376953
 Step 570: loss 1.9158411026000977
 Step 580: loss 2.0220415592193604
 Step 590: loss 1.9899519681930542
 Step 600: loss 2.0434212684631348
 [Seed 71] Step 600 --> val loss 2.2799887657165527
 Step 610: loss 2.001124620437622
 Step 620: loss 2.019329309463501
 Step 630: loss 2.086310863494873
 Step 640: loss 2.0987067222595215
 Step 650: loss 2.128786563873291
 Step 660: loss 1.9667246341705322
 Step 670: loss 2.014819383621216
 Step 680: loss 2.0652740001678467
 Step 690: loss 1.993769645690918
 Step 700: loss 2.0096492767333984
 Step 710: loss 2.052685022354126
 Step 720: loss 2.0720949172973633
 Step 730: loss 2.0025596618652344
 Step 740: loss 2.0165836811065674
 Step 750: loss 2.2221217155456543
 Step 760: loss 2.1369504928588867
 Step 770: loss 2.024829864501953
 Step 780: loss 2.125617504119873
 Step 790: loss 2.09299373626709
 Step 800: loss 1.9832525253295898
 [Seed 71] Step 800 --> val loss 2.3480353355407715
 Step 810: loss 1.9358652830123901
 Step 820: loss 2.172022581100464
 Step 830: loss 1.9708070755004883
 Step 840: loss 2.1114015579223633
 Step 850: loss 2.092566967010498
 Step 860: loss 2.0093448162078857
 Step 870: loss 1.9708422422409058
 Step 880: loss 2.056542158126831
 Step 890: loss 2.0620875358581543
 Step 900: loss 2.21376371383667
 Step 910: loss 1.9457039833068848
 Step 920: loss 2.1802921295166016
 Step 930: loss 2.2079038619995117
 Step 940: loss 2.040628433227539
 Step 950: loss 2.1238176822662354
 Step 960: loss 2.032534122467041
 Step 970: loss 1.941583514213562
 Step 980: loss 1.891512155532837
 Step 990: loss 2.0170817375183105
 Step 1000: loss 2.2283272743225098
 [Seed 71] Step 1000 --> val loss 2.297518014907837
 Step 1010: loss 2.0623655319213867
 Step 1020: loss 1.9866867065429688
 Step 1030: loss 1.9884908199310303
 Step 1040: loss 2.00943660736084
 Step 1050: loss 2.091226577758789
 Step 1060: loss 1.9975008964538574
 Step 1070: loss 1.9781135320663452
 Step 1080: loss 2.126987934112549
 Step 1090: loss 1.995270013809204
 Step 1100: loss 2.0089197158813477
 Step 1110: loss 2.1417083740234375
 Step 1120: loss 1.9696725606918335
 Step 1130: loss 2.0098214149475098
 Step 1140: loss 2.0065789222717285
 Step 1150: loss 2.0605907440185547
 Step 1160: loss 2.116209030151367
 Step 1170: loss 1.932098388671875
 Step 1180: loss 1.977167010307312
 Step 1190: loss 1.9299663305282593
 Step 1200: loss 2.0409512519836426
 [Seed 71] Step 1200 --> val loss 2.2300097942352295
 Step 1210: loss 2.1312520503997803
 Step 1220: loss 1.9842475652694702
 Step 1230: loss 1.9232844114303589
 Step 1240: loss 1.9480876922607422
 Step 1250: loss 1.9302902221679688
 Step 1260: loss 1.969318151473999
 Step 1270: loss 2.0541443824768066
 Step 1280: loss 2.0424323081970215
 Step 1290: loss 1.9934748411178589
 Step 1300: loss 2.0784928798675537
 Step 1310: loss 1.8899410963058472
 Step 1320: loss 1.9973167181015015
 Step 1330: loss 1.998120665550232
 Step 1340: loss 1.8963778018951416
 Step 1350: loss 2.043792247772217
 Step 1360: loss 2.0034966468811035
 Step 1370: loss 2.0235750675201416
 Step 1380: loss 2.0395894050598145
 Step 1390: loss 2.0067100524902344
 Step 1400: loss 1.87882399559021
 [Seed 71] Step 1400 --> val loss 2.152381420135498
 Step 1410: loss 1.8650290966033936
 Step 1420: loss 1.9253718852996826
 Step 1430: loss 1.9763282537460327
 Step 1440: loss 2.0374698638916016
 Step 1450: loss 1.9782989025115967
 Step 1460: loss 2.0948519706726074
 Step 1470: loss 1.9779753684997559
 Step 1480: loss 1.8704293966293335
 Step 1490: loss 1.874054193496704
 Step 1500: loss 2.0543384552001953
 Step 1510: loss 1.9905204772949219
 Step 1520: loss 2.0289177894592285
 Step 1530: loss 1.8637208938598633
 Step 1540: loss 1.9472830295562744
 Step 1550: loss 1.8842748403549194
 Step 1560: loss 1.9156625270843506
 Step 1570: loss 1.8347443342208862
 Step 1580: loss 1.979992389678955
 Step 1590: loss 1.874847412109375
 Step 1600: loss 1.88694167137146
 [Seed 71] Step 1600 --> val loss 2.14544939994812
 Step 1610: loss 2.025394916534424
 Step 1620: loss 1.9215511083602905
 Step 1630: loss 1.9064805507659912
 Step 1640: loss 1.9106884002685547
 Step 1650: loss 1.8834304809570312
 Step 1660: loss 1.9120105504989624
 Step 1670: loss 1.878002643585205
 Step 1680: loss 1.9242957830429077
 Step 1690: loss 1.9157098531723022
 Step 1700: loss 1.9295599460601807
 Step 1710: loss 1.8290202617645264
 Step 1720: loss 1.9220807552337646
 Step 1730: loss 1.8294910192489624
 Step 1740: loss 1.9313342571258545
 Step 1750: loss 1.7780441045761108
 Step 1760: loss 1.7388856410980225
 Step 1770: loss 1.7873977422714233
 Step 1780: loss 1.8278167247772217
 Step 1790: loss 1.7279202938079834
 Step 1800: loss 1.8696820735931396
 [Seed 71] Step 1800 --> val loss 2.0345609188079834
 Step 1810: loss 1.86879563331604
 Step 1820: loss 1.7377914190292358
 Step 1830: loss 1.8704335689544678
 Step 1840: loss 1.732816457748413
 Step 1850: loss 1.698878288269043
 Step 1860: loss 1.7881488800048828
 Step 1870: loss 1.8414082527160645
 Step 1880: loss 1.845083475112915
 Step 1890: loss 1.7275948524475098
 Step 1900: loss 1.7392972707748413
 Step 1910: loss 1.8413231372833252
 Step 1920: loss 1.7622379064559937
 Step 1930: loss 1.9034662246704102
 Step 1940: loss 1.7261708974838257
 Step 1950: loss 1.839094877243042
 Step 1960: loss 1.7680346965789795
 Step 1970: loss 1.8196024894714355
 Step 1980: loss 1.8110889196395874
 Step 1990: loss 1.8521469831466675
 Step 2000: loss 1.659389853477478
 [Seed 71] Step 2000 --> val loss 1.982019305229187
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [42 60 43 52 58 59 56 43  1 47] ...
 First target sequence: [60 43 52 58 59 56 43  1 47 57] ...

 Decoded input: dventure is so early up,
 That calls our person from our morning'

 Decoded target: venture is so early up,
 That calls our person from our morning's
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.219250679016113
 [Seed 72] Step 0 --> val loss 4.184813499450684
 Step 10: loss 3.832524061203003
 Step 20: loss 3.323945999145508
 Step 30: loss 2.9202818870544434
 Step 40: loss 2.678628921508789
 Step 50: loss 2.523688793182373
 Step 60: loss 2.543621778488159
 Step 70: loss 2.3941311836242676
 Step 80: loss 2.3966867923736572
 Step 90: loss 2.3803043365478516
 Step 100: loss 2.29575252532959
 Step 110: loss 2.1727702617645264
 Step 120: loss 2.3088507652282715
 Step 130: loss 2.257140874862671
 Step 140: loss 2.2032790184020996
 Step 150: loss 2.1587624549865723
 Step 160: loss 2.2230124473571777
 Step 170: loss 2.1414761543273926
 Step 180: loss 2.181985855102539
 Step 190: loss 2.1262662410736084
 Step 200: loss 2.12575101852417
 [Seed 72] Step 200 --> val loss 2.3085899353027344
 Step 210: loss 2.1290438175201416
 Step 220: loss 2.19435977935791
 Step 230: loss 2.031794548034668
 Step 240: loss 2.0784640312194824
 Step 250: loss 2.1386189460754395
 Step 260: loss 2.106606960296631
 Step 270: loss 2.1244306564331055
 Step 280: loss 2.0499181747436523
 Step 290: loss 2.063781976699829
 Step 300: loss 2.1155807971954346
 Step 310: loss 2.1153087615966797
 Step 320: loss 2.0812888145446777
 Step 330: loss 2.011096477508545
 Step 340: loss 2.044713020324707
 Step 350: loss 2.140803098678589
 Step 360: loss 2.032139301300049
 Step 370: loss 2.1060266494750977
 Step 380: loss 2.1544418334960938
 Step 390: loss 2.0654077529907227
 Step 400: loss 1.9296820163726807
 [Seed 72] Step 400 --> val loss 2.2339131832122803
 Step 410: loss 2.020352840423584
 Step 420: loss 2.108128070831299
 Step 430: loss 2.0750300884246826
 Step 440: loss 2.0393917560577393
 Step 450: loss 2.089162588119507
 Step 460: loss 2.088582992553711
 Step 470: loss 1.9350241422653198
 Step 480: loss 1.9952071905136108
 Step 490: loss 1.952507734298706
 Step 500: loss 2.0206410884857178
 Step 510: loss 1.957519292831421
 Step 520: loss 2.0355520248413086
 Step 530: loss 1.899268627166748
 Step 540: loss 2.114089012145996
 Step 550: loss 1.9835741519927979
 Step 560: loss 2.0242161750793457
 Step 570: loss 1.8868736028671265
 Step 580: loss 1.9699081182479858
 Step 590: loss 1.929976463317871
 Step 600: loss 2.048722267150879
 [Seed 72] Step 600 --> val loss 2.1234359741210938
 Step 610: loss 1.9469902515411377
 Step 620: loss 2.0093955993652344
 Step 630: loss 1.938715934753418
 Step 640: loss 1.9581289291381836
 Step 650: loss 1.9750337600708008
 Step 660: loss 2.0340375900268555
 Step 670: loss 2.003854274749756
 Step 680: loss 1.8957500457763672
 Step 690: loss 1.9312894344329834
 Step 700: loss 2.1205811500549316
 Step 710: loss 1.9697580337524414
 Step 720: loss 2.0282418727874756
 Step 730: loss 2.05605411529541
 Step 740: loss 2.2504405975341797
 Step 750: loss 1.9980039596557617
 Step 760: loss 2.0004515647888184
 Step 770: loss 1.9477070569992065
 Step 780: loss 1.971724510192871
 Step 790: loss 1.9922302961349487
 Step 800: loss 1.9045170545578003
 [Seed 72] Step 800 --> val loss 2.2735955715179443
 Step 810: loss 1.9928979873657227
 Step 820: loss 2.058258056640625
 Step 830: loss 2.003190040588379
 Step 840: loss 2.0573997497558594
 Step 850: loss 2.0854172706604004
 Step 860: loss 2.050654172897339
 Step 870: loss 2.0456693172454834
 Step 880: loss 1.9543174505233765
 Step 890: loss 2.146959066390991
 Step 900: loss 2.0182766914367676
 Step 910: loss 2.0527071952819824
 Step 920: loss 1.944049596786499
 Step 930: loss 2.025329351425171
 Step 940: loss 1.869736671447754
 Step 950: loss 2.079348564147949
 Step 960: loss 2.013461112976074
 Step 970: loss 1.912137746810913
 Step 980: loss 1.9323713779449463
 Step 990: loss nan
 Step 1000: loss nan
 [Seed 72] Step 1000 --> val loss nan
 Step 1010: loss nan
 Step 1020: loss nan
 Step 1030: loss nan
 Step 1040: loss nan
 Step 1050: loss nan
 Step 1060: loss nan
 Step 1070: loss nan
 Step 1080: loss nan
 Step 1090: loss nan
 Step 1100: loss nan
 Step 1110: loss nan
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 72] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 72] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 72] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 72] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 72] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [53 52 45  8  0 21  1 61 39 57] ...
 First target sequence: [52 45  8  0 21  1 61 39 57  1] ...

 Decoded input: ong.
 I was too hot to do somebody good,
 That is too cold in thin

 Decoded target: ng.
 I was too hot to do somebody good,
 That is too cold in think
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.209516525268555
 [Seed 73] Step 0 --> val loss 4.184791088104248
 Step 10: loss 3.840376615524292
 Step 20: loss 3.3718810081481934
 Step 30: loss 2.824902057647705
 Step 40: loss 2.6360273361206055
 Step 50: loss 2.4840950965881348
 Step 60: loss 2.5252695083618164
 Step 70: loss 2.4323654174804688
 Step 80: loss 2.2640368938446045
 Step 90: loss 2.305546760559082
 Step 100: loss 2.3279128074645996
 Step 110: loss 2.292717695236206
 Step 120: loss 2.171461820602417
 Step 130: loss 2.230297088623047
 Step 140: loss 2.1070895195007324
 Step 150: loss 2.2194392681121826
 Step 160: loss 2.2013633251190186
 Step 170: loss 2.05029559135437
 Step 180: loss 2.2270665168762207
 Step 190: loss 2.0554304122924805
 Step 200: loss 2.0978801250457764
 [Seed 73] Step 200 --> val loss 2.361072301864624
 Step 210: loss 2.07576322555542
 Step 220: loss 2.0407934188842773
 Step 230: loss 2.0695834159851074
 Step 240: loss 2.011996030807495
 Step 250: loss 2.107058048248291
 Step 260: loss 2.180610179901123
 Step 270: loss 2.0552425384521484
 Step 280: loss 2.0440311431884766
 Step 290: loss 2.0792288780212402
 Step 300: loss 1.9934489727020264
 Step 310: loss 2.088120937347412
 Step 320: loss 2.0597054958343506
 Step 330: loss 2.0748486518859863
 Step 340: loss 2.0442330837249756
 Step 350: loss 1.9971603155136108
 Step 360: loss 1.9391570091247559
 Step 370: loss 1.994213342666626
 Step 380: loss 1.9409730434417725
 Step 390: loss 2.0604171752929688
 Step 400: loss 2.036039352416992
 [Seed 73] Step 400 --> val loss 2.2509372234344482
 Step 410: loss 2.1049604415893555
 Step 420: loss 2.0202276706695557
 Step 430: loss 2.0500924587249756
 Step 440: loss 2.0065929889678955
 Step 450: loss 1.9266488552093506
 Step 460: loss 1.9396770000457764
 Step 470: loss 1.9358603954315186
 Step 480: loss 2.133747100830078
 Step 490: loss 2.062051773071289
 Step 500: loss 2.0835788249969482
 Step 510: loss 2.194296360015869
 Step 520: loss 1.9842815399169922
 Step 530: loss 1.9483578205108643
 Step 540: loss 1.9731624126434326
 Step 550: loss 1.8749322891235352
 Step 560: loss 2.010420560836792
 Step 570: loss 2.14237642288208
 Step 580: loss 2.080021381378174
 Step 590: loss 1.8681997060775757
 Step 600: loss 2.0112791061401367
 [Seed 73] Step 600 --> val loss 2.3386526107788086
 Step 610: loss 2.100341320037842
 Step 620: loss 1.9032354354858398
 Step 630: loss 1.8811582326889038
 Step 640: loss 2.0732734203338623
 Step 650: loss 2.084709405899048
 Step 660: loss 2.0832676887512207
 Step 670: loss 1.9968169927597046
 Step 680: loss 1.9050053358078003
 Step 690: loss 1.9770677089691162
 Step 700: loss 2.0456886291503906
 Step 710: loss 1.98687744140625
 Step 720: loss 2.0221970081329346
 Step 730: loss 2.0170137882232666
 Step 740: loss 2.0503194332122803
 Step 750: loss 1.8645155429840088
 Step 760: loss 1.9569149017333984
 Step 770: loss 1.954991340637207
 Step 780: loss 2.060318946838379
 Step 790: loss 1.9726102352142334
 Step 800: loss 2.124452590942383
 [Seed 73] Step 800 --> val loss 2.305140972137451
 Step 810: loss 2.0299792289733887
 Step 820: loss 1.9710358381271362
 Step 830: loss 2.1258981227874756
 Step 840: loss 2.0019242763519287
 Step 850: loss 2.07668137550354
 Step 860: loss 1.974161982536316
 Step 870: loss 1.9395921230316162
 Step 880: loss 2.044931411743164
 Step 890: loss 1.9651225805282593
 Step 900: loss 1.9966548681259155
 Step 910: loss 1.9472821950912476
 Step 920: loss 1.8658655881881714
 Step 930: loss 1.9119691848754883
 Step 940: loss 1.9697203636169434
 Step 950: loss 2.0370407104492188
 Step 960: loss 2.045281171798706
 Step 970: loss 1.8625882863998413
 Step 980: loss 1.8004541397094727
 Step 990: loss 2.041127920150757
 Step 1000: loss 1.9192652702331543
 [Seed 73] Step 1000 --> val loss 2.175995349884033
 Step 1010: loss 1.8782222270965576
 Step 1020: loss 2.038294792175293
 Step 1030: loss 2.0294997692108154
 Step 1040: loss 1.874550700187683
 Step 1050: loss 2.009551525115967
 Step 1060: loss 1.8973426818847656
 Step 1070: loss 1.8848732709884644
 Step 1080: loss 1.989074945449829
 Step 1090: loss 1.8903306722640991
 Step 1100: loss 2.045062303543091
 Step 1110: loss 1.8810338973999023
 Step 1120: loss 1.9052999019622803
 Step 1130: loss 1.9003747701644897
 Step 1140: loss 1.9557474851608276
 Step 1150: loss 1.9594829082489014
 Step 1160: loss 2.074073076248169
 Step 1170: loss 1.9089133739471436
 Step 1180: loss 1.9842805862426758
 Step 1190: loss 1.8881996870040894
 Step 1200: loss 1.9382022619247437
 [Seed 73] Step 1200 --> val loss 2.0765769481658936
 Step 1210: loss 1.990260124206543
 Step 1220: loss 1.9943914413452148
 Step 1230: loss 2.04819655418396
 Step 1240: loss 1.915992021560669
 Step 1250: loss 1.9999440908432007
 Step 1260: loss 1.9016474485397339
 Step 1270: loss 1.8856432437896729
 Step 1280: loss 2.0756874084472656
 Step 1290: loss 2.0675463676452637
 Step 1300: loss 1.8431119918823242
 Step 1310: loss 1.9672034978866577
 Step 1320: loss 2.097801446914673
 Step 1330: loss 2.045295000076294
 Step 1340: loss 1.9408925771713257
 Step 1350: loss 2.056053638458252
 Step 1360: loss 1.8295094966888428
 Step 1370: loss 1.9511792659759521
 Step 1380: loss 1.886788249015808
 Step 1390: loss 1.9366626739501953
 Step 1400: loss 1.8615401983261108
 [Seed 73] Step 1400 --> val loss 2.0368831157684326
 Step 1410: loss 1.9319524765014648
 Step 1420: loss 2.0249321460723877
 Step 1430: loss 1.9659322500228882
 Step 1440: loss 2.0884041786193848
 Step 1450: loss 1.960479736328125
 Step 1460: loss 1.9681724309921265
 Step 1470: loss 1.9719798564910889
 Step 1480: loss 2.044858932495117
 Step 1490: loss 1.863542079925537
 Step 1500: loss 1.874680757522583
 Step 1510: loss 1.9688819646835327
 Step 1520: loss 1.9529142379760742
 Step 1530: loss 1.9136972427368164
 Step 1540: loss 1.8731207847595215
 Step 1550: loss 1.908494472503662
 Step 1560: loss 1.778131365776062
 Step 1570: loss 2.0134029388427734
 Step 1580: loss 1.9057104587554932
 Step 1590: loss 1.9218329191207886
 Step 1600: loss 1.8336255550384521
 [Seed 73] Step 1600 --> val loss 1.9594799280166626
 Step 1610: loss 1.7972577810287476
 Step 1620: loss 1.8319597244262695
 Step 1630: loss 1.7948122024536133
 Step 1640: loss 1.8524761199951172
 Step 1650: loss 1.810917615890503
 Step 1660: loss 1.8356513977050781
 Step 1670: loss 1.8685120344161987
 Step 1680: loss 1.7613967657089233
 Step 1690: loss 1.8670649528503418
 Step 1700: loss 1.9436801671981812
 Step 1710: loss 1.8070621490478516
 Step 1720: loss 1.8119256496429443
 Step 1730: loss 1.734818935394287
 Step 1740: loss 1.8362042903900146
 Step 1750: loss 1.8027138710021973
 Step 1760: loss 1.774881362915039
 Step 1770: loss 1.819732427597046
 Step 1780: loss 1.769866704940796
 Step 1790: loss 1.7611373662948608
 Step 1800: loss 1.8355201482772827
 [Seed 73] Step 1800 --> val loss 1.9577072858810425
 Step 1810: loss 1.9404805898666382
 Step 1820: loss 1.8670895099639893
 Step 1830: loss 1.7539982795715332
 Step 1840: loss 1.7533972263336182
 Step 1850: loss 1.714590072631836
 Step 1860: loss 1.747006893157959
 Step 1870: loss 1.818943977355957
 Step 1880: loss 1.7839035987854004
 Step 1890: loss 1.751795768737793
 Step 1900: loss 1.8645005226135254
 Step 1910: loss 1.7100319862365723
 Step 1920: loss 1.8630508184432983
 Step 1930: loss 1.8180291652679443
 Step 1940: loss 1.7755115032196045
 Step 1950: loss 1.6765598058700562
 Step 1960: loss 1.776482105255127
 Step 1970: loss 1.6764476299285889
 Step 1980: loss 1.6859996318817139
 Step 1990: loss 1.857465147972107
 Step 2000: loss 1.7988494634628296
 [Seed 73] Step 2000 --> val loss 1.9085012674331665
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 46 43  1 46 39 58 46  1 57] ...
 First target sequence: [46 43  1 46 39 58 46  1 57 58] ...

 Decoded input:  he hath stain'd?

 ISABELLA:
 Sir, believe this,
 I had rather giv

 Decoded target: he hath stain'd?

 ISABELLA:
 Sir, believe this,
 I had rather give
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.203078269958496
 [Seed 74] Step 0 --> val loss 4.18369197845459
 Step 10: loss 3.8531103134155273
 Step 20: loss 3.360748767852783
 Step 30: loss 2.8722171783447266
 Step 40: loss 2.6310038566589355
 Step 50: loss 2.5211992263793945
 Step 60: loss 2.5915918350219727
 Step 70: loss 2.344874858856201
 Step 80: loss 2.3589818477630615
 Step 90: loss 2.3133318424224854
 Step 100: loss 2.2865402698516846
 Step 110: loss 2.212341785430908
 Step 120: loss 2.2593023777008057
 Step 130: loss 2.3151345252990723
 Step 140: loss 2.1741302013397217
 Step 150: loss 2.154963254928589
 Step 160: loss 2.1223578453063965
 Step 170: loss 2.1041009426116943
 Step 180: loss 2.066518545150757
 Step 190: loss 2.0765533447265625
 Step 200: loss 2.0558230876922607
 [Seed 74] Step 200 --> val loss 2.4280364513397217
 Step 210: loss 2.1239116191864014
 Step 220: loss 2.118290901184082
 Step 230: loss 2.031881332397461
 Step 240: loss 2.1165270805358887
 Step 250: loss 2.0533456802368164
 Step 260: loss 2.0890047550201416
 Step 270: loss 2.1087095737457275
 Step 280: loss 2.0044167041778564
 Step 290: loss 1.9854954481124878
 Step 300: loss 2.079954147338867
 Step 310: loss 2.145658493041992
 Step 320: loss 2.0243794918060303
 Step 330: loss 2.0225954055786133
 Step 340: loss 2.0313937664031982
 Step 350: loss 2.0194931030273438
 Step 360: loss 1.999361515045166
 Step 370: loss 2.0724873542785645
 Step 380: loss 2.02811861038208
 Step 390: loss 1.941502332687378
 Step 400: loss 1.9168075323104858
 [Seed 74] Step 400 --> val loss 2.1483943462371826
 Step 410: loss 2.040675163269043
 Step 420: loss 2.0534024238586426
 Step 430: loss 2.0147664546966553
 Step 440: loss 1.9953317642211914
 Step 450: loss 1.9734323024749756
 Step 460: loss 2.020263195037842
 Step 470: loss 1.9250264167785645
 Step 480: loss 1.9410332441329956
 Step 490: loss 2.1623694896698
 Step 500: loss 1.9543565511703491
 Step 510: loss 2.0308310985565186
 Step 520: loss 1.9980103969573975
 Step 530: loss 2.005829334259033
 Step 540: loss 2.0652761459350586
 Step 550: loss 1.9612107276916504
 Step 560: loss 1.989149570465088
 Step 570: loss 1.9884469509124756
 Step 580: loss 2.0143442153930664
 Step 590: loss 1.9913806915283203
 Step 600: loss 2.0681140422821045
 [Seed 74] Step 600 --> val loss 2.2974154949188232
 Step 610: loss 1.9584002494812012
 Step 620: loss 1.940406084060669
 Step 630: loss 2.04276704788208
 Step 640: loss 1.943227767944336
 Step 650: loss 1.997358798980713
 Step 660: loss 1.9861459732055664
 Step 670: loss 2.012754201889038
 Step 680: loss 1.9653022289276123
 Step 690: loss 1.9144561290740967
 Step 700: loss 1.985587477684021
 Step 710: loss 2.062279462814331
 Step 720: loss 2.0110256671905518
 Step 730: loss 1.9689383506774902
 Step 740: loss 1.9973235130310059
 Step 750: loss 1.9755738973617554
 Step 760: loss 1.9485070705413818
 Step 770: loss 1.9615085124969482
 Step 780: loss 1.955541968345642
 Step 790: loss 2.003537178039551
 Step 800: loss 1.8974781036376953
 [Seed 74] Step 800 --> val loss 2.244565963745117
 Step 810: loss 1.9179737567901611
 Step 820: loss 1.9074935913085938
 Step 830: loss 2.0036301612854004
 Step 840: loss 1.8629225492477417
 Step 850: loss 1.9383175373077393
 Step 860: loss 1.932828426361084
 Step 870: loss 1.958770751953125
 Step 880: loss 1.9563950300216675
 Step 890: loss 1.9545201063156128
 Step 900: loss 1.961080551147461
 Step 910: loss 1.872743844985962
 Step 920: loss 1.938058614730835
 Step 930: loss 2.021129608154297
 Step 940: loss 2.04048752784729
 Step 950: loss 2.000359058380127
 Step 960: loss 1.9777568578720093
 Step 970: loss 1.9754753112792969
 Step 980: loss 1.9463942050933838
 Step 990: loss 2.0140342712402344
 Step 1000: loss 2.0395278930664062
 [Seed 74] Step 1000 --> val loss 2.0919570922851562
 Step 1010: loss 1.9016427993774414
 Step 1020: loss 1.9190891981124878
 Step 1030: loss 2.186666250228882
 Step 1040: loss 1.978808879852295
 Step 1050: loss 1.8123741149902344
 Step 1060: loss 1.934274673461914
 Step 1070: loss 2.0695695877075195
 Step 1080: loss 1.9751766920089722
 Step 1090: loss 1.9124587774276733
 Step 1100: loss 1.9227008819580078
 Step 1110: loss 1.9759258031845093
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 74] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 74] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 74] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 74] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 74] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [56 47 52 41 43 50 63  1 43 39] ...
 First target sequence: [47 52 41 43 50 63  1 43 39 45] ...

 Decoded input: rincely eagle,
 Under whose shade the ramping lion slept,
 Whose t

 Decoded target: incely eagle,
 Under whose shade the ramping lion slept,
 Whose to
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.202980041503906
 [Seed 75] Step 0 --> val loss 4.185405254364014
 Step 10: loss 3.8426833152770996
 Step 20: loss 3.3264102935791016
 Step 30: loss 2.8470306396484375
 Step 40: loss 2.6414990425109863
 Step 50: loss 2.5355331897735596
 Step 60: loss 2.462667465209961
 Step 70: loss 2.431809425354004
 Step 80: loss 2.3937649726867676
 Step 90: loss 2.3523178100585938
 Step 100: loss 2.2364718914031982
 Step 110: loss 2.232043981552124
 Step 120: loss 2.218404769897461
 Step 130: loss 2.2605791091918945
 Step 140: loss 2.1998772621154785
 Step 150: loss 2.1032841205596924
 Step 160: loss 2.2184128761291504
 Step 170: loss 2.101442813873291
 Step 180: loss 2.1852197647094727
 Step 190: loss 2.0698297023773193
 Step 200: loss 2.213230609893799
 [Seed 75] Step 200 --> val loss 2.2918906211853027
 Step 210: loss 2.0636954307556152
 Step 220: loss 2.0632288455963135
 Step 230: loss 2.145827531814575
 Step 240: loss 2.053752899169922
 Step 250: loss 2.079892873764038
 Step 260: loss 1.9572994709014893
 Step 270: loss 1.8922675848007202
 Step 280: loss 2.061187267303467
 Step 290: loss 1.963838815689087
 Step 300: loss 2.0567095279693604
 Step 310: loss 2.072580337524414
 Step 320: loss 2.08760142326355
 Step 330: loss 2.076835870742798
 Step 340: loss 2.0026683807373047
 Step 350: loss 1.9049098491668701
 Step 360: loss 2.0845141410827637
 Step 370: loss 1.9911308288574219
 Step 380: loss 2.0610530376434326
 Step 390: loss 1.977505087852478
 Step 400: loss 1.9766329526901245
 [Seed 75] Step 400 --> val loss 2.162444591522217
 Step 410: loss 1.9682679176330566
 Step 420: loss 2.0901269912719727
 Step 430: loss 1.9752519130706787
 Step 440: loss 1.901008129119873
 Step 450: loss 1.9923189878463745
 Step 460: loss 2.1293554306030273
 Step 470: loss 1.9739904403686523
 Step 480: loss 1.9543741941452026
 Step 490: loss 2.001368522644043
 Step 500: loss 2.0258541107177734
 Step 510: loss 1.9999828338623047
 Step 520: loss 2.0513875484466553
 Step 530: loss 2.018810272216797
 Step 540: loss 2.0626749992370605
 Step 550: loss 2.015626907348633
 Step 560: loss 1.8454713821411133
 Step 570: loss 2.1062088012695312
 Step 580: loss 2.0924530029296875
 Step 590: loss 1.947272777557373
 Step 600: loss 2.0863125324249268
 [Seed 75] Step 600 --> val loss 2.15583872795105
 Step 610: loss 1.9960530996322632
 Step 620: loss 2.019914388656616
 Step 630: loss 1.9935888051986694
 Step 640: loss 2.00309419631958
 Step 650: loss 2.094914197921753
 Step 660: loss 2.033331871032715
 Step 670: loss 1.9235872030258179
 Step 680: loss 2.0803143978118896
 Step 690: loss 2.0467028617858887
 Step 700: loss 1.9431638717651367
 Step 710: loss 2.0084586143493652
 Step 720: loss 2.031008243560791
 Step 730: loss 1.8988699913024902
 Step 740: loss 2.036912441253662
 Step 750: loss 1.9426240921020508
 Step 760: loss 1.86029851436615
 Step 770: loss 1.9435548782348633
 Step 780: loss 1.950116515159607
 Step 790: loss 1.9759738445281982
 Step 800: loss 1.917773723602295
 [Seed 75] Step 800 --> val loss 2.196629047393799
 Step 810: loss 1.9082359075546265
 Step 820: loss 1.9107086658477783
 Step 830: loss 2.044544219970703
 Step 840: loss 1.9339708089828491
 Step 850: loss 2.047177791595459
 Step 860: loss 1.8743410110473633
 Step 870: loss 1.9397437572479248
 Step 880: loss 1.8217881917953491
 Step 890: loss 2.001305341720581
 Step 900: loss 2.066270112991333
 Step 910: loss 1.9840507507324219
 Step 920: loss 1.9942662715911865
 Step 930: loss 1.906195878982544
 Step 940: loss 1.9513726234436035
 Step 950: loss 2.0425074100494385
 Step 960: loss 1.9921050071716309
 Step 970: loss 1.8480415344238281
 Step 980: loss 1.946312665939331
 Step 990: loss 1.944596290588379
 Step 1000: loss 1.9151322841644287
 [Seed 75] Step 1000 --> val loss 2.2458713054656982
 Step 1010: loss 2.1060068607330322
 Step 1020: loss 1.9182045459747314
 Step 1030: loss 1.8660995960235596
 Step 1040: loss 2.095294952392578
 Step 1050: loss 1.9819053411483765
 Step 1060: loss 2.065838575363159
 Step 1070: loss 1.9116184711456299
 Step 1080: loss 1.9040563106536865
 Step 1090: loss 2.055053472518921
 Step 1100: loss 1.9602718353271484
 Step 1110: loss 1.9393689632415771
 Step 1120: loss 1.9056692123413086
 Step 1130: loss 2.0554141998291016
 Step 1140: loss 1.900730013847351
 Step 1150: loss 2.048849582672119
 Step 1160: loss 2.025362968444824
 Step 1170: loss 1.951366901397705
 Step 1180: loss 2.083113193511963
 Step 1190: loss 1.9541468620300293
 Step 1200: loss 1.9756536483764648
 [Seed 75] Step 1200 --> val loss 2.24853777885437
 Step 1210: loss 2.0930612087249756
 Step 1220: loss 1.8768095970153809
 Step 1230: loss 1.9981725215911865
 Step 1240: loss 2.0795516967773438
 Step 1250: loss 1.8759372234344482
 Step 1260: loss 1.8853678703308105
 Step 1270: loss 1.8707956075668335
 Step 1280: loss 2.0769529342651367
 Step 1290: loss 2.058420181274414
 Step 1300: loss 1.9725730419158936
 Step 1310: loss 2.0129780769348145
 Step 1320: loss 1.9125632047653198
 Step 1330: loss 1.9599088430404663
 Step 1340: loss 1.880678415298462
 Step 1350: loss 1.9366645812988281
 Step 1360: loss 1.8538802862167358
 Step 1370: loss 1.897865653038025
 Step 1380: loss 1.8034242391586304
 Step 1390: loss 1.8289482593536377
 Step 1400: loss 2.0437543392181396
 [Seed 75] Step 1400 --> val loss 2.146895170211792
 Step 1410: loss 1.9237775802612305
 Step 1420: loss 1.9235399961471558
 Step 1430: loss 1.8289673328399658
 Step 1440: loss 1.7287282943725586
 Step 1450: loss 1.8493080139160156
 Step 1460: loss 1.8700774908065796
 Step 1470: loss 1.8438228368759155
 Step 1480: loss 1.7791194915771484
 Step 1490: loss 1.9278364181518555
 Step 1500: loss 1.859683871269226
 Step 1510: loss 1.756752371788025
 Step 1520: loss 1.8106639385223389
 Step 1530: loss 1.786077857017517
 Step 1540: loss 1.7726653814315796
 Step 1550: loss 1.8491220474243164
 Step 1560: loss 1.6312129497528076
 Step 1570: loss 1.7421653270721436
 Step 1580: loss 1.8995046615600586
 Step 1590: loss 1.91348397731781
 Step 1600: loss 1.8285939693450928
 [Seed 75] Step 1600 --> val loss 2.1463043689727783
 Step 1610: loss 1.792989730834961
 Step 1620: loss 1.8529508113861084
 Step 1630: loss 1.8416255712509155
 Step 1640: loss 1.7048892974853516
 Step 1650: loss 1.7691762447357178
 Step 1660: loss 1.7820152044296265
 Step 1670: loss 1.821311354637146
 Step 1680: loss 1.7727034091949463
 Step 1690: loss 1.643842101097107
 Step 1700: loss 1.8170764446258545
 Step 1710: loss 1.7602403163909912
 Step 1720: loss 1.8038761615753174
 Step 1730: loss 1.733489990234375
 Step 1740: loss 1.7931698560714722
 Step 1750: loss 1.8056001663208008
 Step 1760: loss 1.874758243560791
 Step 1770: loss 1.7373679876327515
 Step 1780: loss 1.7815525531768799
 Step 1790: loss 1.8054275512695312
 Step 1800: loss 1.8515609502792358
 [Seed 75] Step 1800 --> val loss 2.034456253051758
 Step 1810: loss 1.6581544876098633
 Step 1820: loss 1.8207266330718994
 Step 1830: loss 1.6769388914108276
 Step 1840: loss 1.7052340507507324
 Step 1850: loss 1.715968132019043
 Step 1860: loss 1.9304794073104858
 Step 1870: loss 1.7297385931015015
 Step 1880: loss 1.7887580394744873
 Step 1890: loss 1.6816190481185913
 Step 1900: loss 1.6252182722091675
 Step 1910: loss 1.7514855861663818
 Step 1920: loss 1.7823550701141357
 Step 1930: loss 1.7781667709350586
 Step 1940: loss 1.8806161880493164
 Step 1950: loss 1.687427282333374
 Step 1960: loss 1.7598880529403687
 Step 1970: loss 1.789585828781128
 Step 1980: loss 1.7783098220825195
 Step 1990: loss 1.7435033321380615
 Step 2000: loss 1.881624460220337
 [Seed 75] Step 2000 --> val loss 1.9464409351348877
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [42  1 63 47 43 50 42  1 59 57] ...
 First target sequence: [ 1 63 47 43 50 42  1 59 57  1] ...

 Decoded input: d yield us but the superfluity, while it were
 wholesome, we migh

 Decoded target:  yield us but the superfluity, while it were
 wholesome, we might
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.224676132202148
 [Seed 76] Step 0 --> val loss 4.183635234832764
 Step 10: loss 3.8651015758514404
 Step 20: loss 3.293449878692627
 Step 30: loss 2.8675408363342285
 Step 40: loss 2.6635680198669434
 Step 50: loss 2.541415214538574
 Step 60: loss 2.480069160461426
 Step 70: loss 2.4392194747924805
 Step 80: loss 2.454192638397217
 Step 90: loss 2.3195133209228516
 Step 100: loss 2.317138433456421
 Step 110: loss 2.323060989379883
 Step 120: loss 2.2083306312561035
 Step 130: loss 2.1374058723449707
 Step 140: loss 2.2550528049468994
 Step 150: loss 2.2193236351013184
 Step 160: loss 2.132669448852539
 Step 170: loss 2.088914632797241
 Step 180: loss 2.1552305221557617
 Step 190: loss 2.0527031421661377
 Step 200: loss 2.0214626789093018
 [Seed 76] Step 200 --> val loss 2.288837432861328
 Step 210: loss 2.172518253326416
 Step 220: loss 2.0892276763916016
 Step 230: loss 2.0429000854492188
 Step 240: loss 2.0379819869995117
 Step 250: loss 2.036121368408203
 Step 260: loss 2.0399224758148193
 Step 270: loss 1.9453933238983154
 Step 280: loss 1.9408448934555054
 Step 290: loss 2.0866007804870605
 Step 300: loss 2.127699375152588
 Step 310: loss 2.0616469383239746
 Step 320: loss 2.158742666244507
 Step 330: loss 2.0094332695007324
 Step 340: loss 1.989137053489685
 Step 350: loss 2.009795904159546
 Step 360: loss 2.0699965953826904
 Step 370: loss 2.0697710514068604
 Step 380: loss 1.9657310247421265
 Step 390: loss 2.1550519466400146
 Step 400: loss 2.0468709468841553
 [Seed 76] Step 400 --> val loss 2.390556573867798
 Step 410: loss 2.030712604522705
 Step 420: loss 1.9427282810211182
 Step 430: loss 1.9811134338378906
 Step 440: loss 2.1419520378112793
 Step 450: loss 2.059481143951416
 Step 460: loss 2.0298452377319336
 Step 470: loss 1.97433602809906
 Step 480: loss 2.1400585174560547
 Step 490: loss 1.997795581817627
 Step 500: loss 2.030529022216797
 Step 510: loss 2.0809285640716553
 Step 520: loss 1.9381908178329468
 Step 530: loss 1.9894025325775146
 Step 540: loss 2.105747938156128
 Step 550: loss 2.1085004806518555
 Step 560: loss 2.0427427291870117
 Step 570: loss 2.0912675857543945
 Step 580: loss 1.938534140586853
 Step 590: loss 2.169779062271118
 Step 600: loss 2.1248526573181152
 [Seed 76] Step 600 --> val loss 2.396723508834839
 Step 610: loss 1.9405641555786133
 Step 620: loss 2.0126538276672363
 Step 630: loss 2.002822160720825
 Step 640: loss 2.041585922241211
 Step 650: loss 2.0021510124206543
 Step 660: loss 2.052527904510498
 Step 670: loss 2.0871052742004395
 Step 680: loss 1.96974515914917
 Step 690: loss 1.9542968273162842
 Step 700: loss 2.07861328125
 Step 710: loss 1.9151829481124878
 Step 720: loss 2.0935072898864746
 Step 730: loss 2.0304341316223145
 Step 740: loss 2.074535608291626
 Step 750: loss 2.094954013824463
 Step 760: loss 1.9392106533050537
 Step 770: loss 1.9215134382247925
 Step 780: loss 2.023162841796875
 Step 790: loss 1.9964807033538818
 Step 800: loss 1.868669033050537
 [Seed 76] Step 800 --> val loss 2.192815065383911
 Step 810: loss 1.9428666830062866
 Step 820: loss 1.9773576259613037
 Step 830: loss 1.9960225820541382
 Step 840: loss 2.09175968170166
 Step 850: loss 2.0389885902404785
 Step 860: loss 2.142425298690796
 Step 870: loss 1.9960908889770508
 Step 880: loss 2.139551877975464
 Step 890: loss 1.8339011669158936
 Step 900: loss 1.8188680410385132
 Step 910: loss 1.9884308576583862
 Step 920: loss 2.026554584503174
 Step 930: loss 1.8980023860931396
 Step 940: loss 1.831610083580017
 Step 950: loss 2.013423442840576
 Step 960: loss 2.0867860317230225
 Step 970: loss 1.8888497352600098
 Step 980: loss 1.9557571411132812
 Step 990: loss 1.98264479637146
 Step 1000: loss 1.9583003520965576
 [Seed 76] Step 1000 --> val loss 2.191110372543335
 Step 1010: loss 2.0276944637298584
 Step 1020: loss 1.8841522932052612
 Step 1030: loss 2.0282235145568848
 Step 1040: loss 2.076082706451416
 Step 1050: loss 2.13249135017395
 Step 1060: loss 2.1277413368225098
 Step 1070: loss 2.126133918762207
 Step 1080: loss 1.9677962064743042
 Step 1090: loss 2.027169942855835
 Step 1100: loss 1.9202897548675537
 Step 1110: loss 2.076127052307129
 Step 1120: loss 2.0328457355499268
 Step 1130: loss 1.9559820890426636
 Step 1140: loss 1.9804171323776245
 Step 1150: loss 2.0054492950439453
 Step 1160: loss 1.8959630727767944
 Step 1170: loss 1.9812692403793335
 Step 1180: loss 1.9117131233215332
 Step 1190: loss 1.9303247928619385
 Step 1200: loss 1.954197883605957
 [Seed 76] Step 1200 --> val loss 2.1388251781463623
 Step 1210: loss 2.017371892929077
 Step 1220: loss 2.074422597885132
 Step 1230: loss 1.9748096466064453
 Step 1240: loss 1.9391324520111084
 Step 1250: loss 1.9673666954040527
 Step 1260: loss 1.8956550359725952
 Step 1270: loss 1.9656195640563965
 Step 1280: loss 1.8768004179000854
 Step 1290: loss 1.9173040390014648
 Step 1300: loss 1.944301962852478
 Step 1310: loss 1.9421151876449585
 Step 1320: loss 1.868882656097412
 Step 1330: loss 1.9985771179199219
 Step 1340: loss 1.8201221227645874
 Step 1350: loss 1.8206079006195068
 Step 1360: loss 1.9165608882904053
 Step 1370: loss 1.8570704460144043
 Step 1380: loss 1.964978575706482
 Step 1390: loss 1.9037590026855469
 Step 1400: loss 1.9961869716644287
 [Seed 76] Step 1400 --> val loss 2.1429555416107178
 Step 1410: loss 1.8645232915878296
 Step 1420: loss 1.9506330490112305
 Step 1430: loss 2.029172658920288
 Step 1440: loss 1.968318223953247
 Step 1450: loss 1.9017479419708252
 Step 1460: loss 1.9121990203857422
 Step 1470: loss 1.810443639755249
 Step 1480: loss 1.9425370693206787
 Step 1490: loss 1.8562145233154297
 Step 1500: loss 1.981635332107544
 Step 1510: loss 1.8981845378875732
 Step 1520: loss 1.8319728374481201
 Step 1530: loss 1.83499276638031
 Step 1540: loss 1.9173943996429443
 Step 1550: loss 1.8588651418685913
 Step 1560: loss 1.803648829460144
 Step 1570: loss 1.7973172664642334
 Step 1580: loss 1.826272964477539
 Step 1590: loss 1.8483951091766357
 Step 1600: loss 1.7481554746627808
 [Seed 76] Step 1600 --> val loss 2.0781588554382324
 Step 1610: loss 1.7520564794540405
 Step 1620: loss 1.8455450534820557
 Step 1630: loss 1.7024450302124023
 Step 1640: loss 1.835276484489441
 Step 1650: loss 1.8559907674789429
 Step 1660: loss 1.823209285736084
 Step 1670: loss 1.7717041969299316
 Step 1680: loss 1.7291572093963623
 Step 1690: loss 1.9179818630218506
 Step 1700: loss 1.8543939590454102
 Step 1710: loss 1.8984944820404053
 Step 1720: loss 1.8740956783294678
 Step 1730: loss 1.8695929050445557
 Step 1740: loss 1.673649549484253
 Step 1750: loss 1.7177817821502686
 Step 1760: loss 1.7256078720092773
 Step 1770: loss 1.8921293020248413
 Step 1780: loss 1.6831631660461426
 Step 1790: loss 1.7258328199386597
 Step 1800: loss 1.755737543106079
 [Seed 76] Step 1800 --> val loss 1.9089195728302002
 Step 1810: loss 1.7383651733398438
 Step 1820: loss 1.8045305013656616
 Step 1830: loss 1.7464393377304077
 Step 1840: loss 1.8598241806030273
 Step 1850: loss 1.773895502090454
 Step 1860: loss 1.7196459770202637
 Step 1870: loss 1.661226511001587
 Step 1880: loss 1.7489827871322632
 Step 1890: loss 1.8238390684127808
 Step 1900: loss 1.743800401687622
 Step 1910: loss 1.7332628965377808
 Step 1920: loss 1.7719786167144775
 Step 1930: loss 1.583399772644043
 Step 1940: loss 1.8056367635726929
 Step 1950: loss 1.737334966659546
 Step 1960: loss 1.658917784690857
 Step 1970: loss 1.7848796844482422
 Step 1980: loss 1.6224815845489502
 Step 1990: loss 1.5958516597747803
 Step 2000: loss 1.8280787467956543
 [Seed 76] Step 2000 --> val loss 1.8595117330551147
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [39 50  1 57 47 56  6  1 44 56] ...
 First target sequence: [50  1 57 47 56  6  1 44 56 53] ...

 Decoded input: al sir, from thence; from him, whose daughter
 His tears proclaim

 Decoded target: l sir, from thence; from him, whose daughter
 His tears proclaim'
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.200252056121826
 [Seed 77] Step 0 --> val loss 4.182332515716553
 Step 10: loss 3.8447396755218506
 Step 20: loss 3.358687162399292
 Step 30: loss 2.955019950866699
 Step 40: loss 2.694828987121582
 Step 50: loss 2.5067546367645264
 Step 60: loss 2.5895612239837646
 Step 70: loss 2.3982455730438232
 Step 80: loss 2.3296313285827637
 Step 90: loss 2.348132848739624
 Step 100: loss 2.299633264541626
 Step 110: loss 2.2838010787963867
 Step 120: loss 2.2550392150878906
 Step 130: loss 2.14217472076416
 Step 140: loss 2.2164688110351562
 Step 150: loss 2.144549608230591
 Step 160: loss 2.245239734649658
 Step 170: loss 2.0561296939849854
 Step 180: loss 2.1485509872436523
 Step 190: loss 2.074505567550659
 Step 200: loss 2.124880790710449
 [Seed 77] Step 200 --> val loss 2.403475284576416
 Step 210: loss 2.1332178115844727
 Step 220: loss 2.036865234375
 Step 230: loss 1.9854716062545776
 Step 240: loss 2.1568703651428223
 Step 250: loss 2.0811009407043457
 Step 260: loss 2.0543830394744873
 Step 270: loss 2.0913686752319336
 Step 280: loss 2.0713601112365723
 Step 290: loss 2.114765167236328
 Step 300: loss 2.0752925872802734
 Step 310: loss 2.0168278217315674
 Step 320: loss 2.0482070446014404
 Step 330: loss 2.0386404991149902
 Step 340: loss 1.9418638944625854
 Step 350: loss 2.047201633453369
 Step 360: loss 2.079185962677002
 Step 370: loss 2.1342954635620117
 Step 380: loss 2.0147788524627686
 Step 390: loss 2.0521533489227295
 Step 400: loss 2.011831283569336
 [Seed 77] Step 400 --> val loss 2.2420177459716797
 Step 410: loss 1.9843530654907227
 Step 420: loss 1.973672866821289
 Step 430: loss 1.9869369268417358
 Step 440: loss 2.0162172317504883
 Step 450: loss 1.9748934507369995
 Step 460: loss 1.9058318138122559
 Step 470: loss 2.0098867416381836
 Step 480: loss 2.089569091796875
 Step 490: loss 2.0029051303863525
 Step 500: loss 2.0828237533569336
 Step 510: loss 2.119074583053589
 Step 520: loss 2.043931722640991
 Step 530: loss 1.9798945188522339
 Step 540: loss 1.9938116073608398
 Step 550: loss 2.0097670555114746
 Step 560: loss 2.003283977508545
 Step 570: loss 1.85094153881073
 Step 580: loss 1.9796546697616577
 Step 590: loss 2.003131151199341
 Step 600: loss 2.0806069374084473
 [Seed 77] Step 600 --> val loss 2.265117645263672
 Step 610: loss 2.1049928665161133
 Step 620: loss 1.9937890768051147
 Step 630: loss 1.9970391988754272
 Step 640: loss 1.9524859189987183
 Step 650: loss 1.9398759603500366
 Step 660: loss 1.917825698852539
 Step 670: loss 2.033111572265625
 Step 680: loss 1.9885544776916504
 Step 690: loss 1.8781875371932983
 Step 700: loss 2.088637351989746
 Step 710: loss 1.9746336936950684
 Step 720: loss 1.9891307353973389
 Step 730: loss 2.0255627632141113
 Step 740: loss 1.9217256307601929
 Step 750: loss 2.102076530456543
 Step 760: loss 2.0349318981170654
 Step 770: loss 2.1265335083007812
 Step 780: loss 2.02821683883667
 Step 790: loss 2.0247273445129395
 Step 800: loss 1.9766114950180054
 [Seed 77] Step 800 --> val loss 2.2965378761291504
 Step 810: loss 2.142085075378418
 Step 820: loss 1.952413558959961
 Step 830: loss 2.0220274925231934
 Step 840: loss 1.9337037801742554
 Step 850: loss 2.0649185180664062
 Step 860: loss 2.047508716583252
 Step 870: loss 1.990613341331482
 Step 880: loss 1.880872130393982
 Step 890: loss 1.9735783338546753
 Step 900: loss 1.9746068716049194
 Step 910: loss 2.0823264122009277
 Step 920: loss 1.901580810546875
 Step 930: loss 1.8562507629394531
 Step 940: loss 2.0409038066864014
 Step 950: loss 2.042966604232788
 Step 960: loss 2.087692975997925
 Step 970: loss 2.0952157974243164
 Step 980: loss 1.9388809204101562
 Step 990: loss 1.981123447418213
 Step 1000: loss 1.9360264539718628
 [Seed 77] Step 1000 --> val loss 2.2994747161865234
 Step 1010: loss 1.9839799404144287
 Step 1020: loss 2.055361747741699
 Step 1030: loss 1.9973710775375366
 Step 1040: loss 2.0393290519714355
 Step 1050: loss 1.9535101652145386
 Step 1060: loss 1.8990477323532104
 Step 1070: loss 2.134021759033203
 Step 1080: loss 2.0206007957458496
 Step 1090: loss 1.9196162223815918
 Step 1100: loss 1.9810351133346558
 Step 1110: loss 1.953695297241211
 Step 1120: loss 1.923928141593933
 Step 1130: loss 1.8646583557128906
 Step 1140: loss 1.9157626628875732
 Step 1150: loss 2.0510501861572266
 Step 1160: loss 1.905414342880249
 Step 1170: loss 2.0449066162109375
 Step 1180: loss 2.0552735328674316
 Step 1190: loss 1.9346154928207397
 Step 1200: loss 1.9045270681381226
 [Seed 77] Step 1200 --> val loss 2.211951494216919
 Step 1210: loss 2.033010959625244
 Step 1220: loss 1.886514663696289
 Step 1230: loss 2.046381711959839
 Step 1240: loss 1.991769552230835
 Step 1250: loss 2.0662238597869873
 Step 1260: loss 2.007290840148926
 Step 1270: loss 2.0886383056640625
 Step 1280: loss 1.9519469738006592
 Step 1290: loss 2.0506911277770996
 Step 1300: loss 1.9450500011444092
 Step 1310: loss 1.8278529644012451
 Step 1320: loss 2.0171236991882324
 Step 1330: loss 1.9791641235351562
 Step 1340: loss 1.9141204357147217
 Step 1350: loss 1.8885821104049683
 Step 1360: loss 1.9785821437835693
 Step 1370: loss 2.0251853466033936
 Step 1380: loss 1.9582579135894775
 Step 1390: loss 1.955620288848877
 Step 1400: loss 1.912170648574829
 [Seed 77] Step 1400 --> val loss 2.3159332275390625
 Step 1410: loss 2.0631661415100098
 Step 1420: loss 1.9356677532196045
 Step 1430: loss 2.090998411178589
 Step 1440: loss 2.026996612548828
 Step 1450: loss 1.9979966878890991
 Step 1460: loss 1.945859670639038
 Step 1470: loss 1.8848199844360352
 Step 1480: loss 1.874837875366211
 Step 1490: loss 1.9098289012908936
 Step 1500: loss 1.9599031209945679
 Step 1510: loss 1.9049010276794434
 Step 1520: loss 2.0064945220947266
 Step 1530: loss 1.8862018585205078
 Step 1540: loss 1.94325852394104
 Step 1550: loss 1.9321095943450928
 Step 1560: loss 1.7603540420532227
 Step 1570: loss 1.957273244857788
 Step 1580: loss 1.8729747533798218
 Step 1590: loss 1.860842227935791
 Step 1600: loss 1.8441753387451172
 [Seed 77] Step 1600 --> val loss 2.148688316345215
 Step 1610: loss 2.0044612884521484
 Step 1620: loss 1.9667314291000366
 Step 1630: loss 1.7329901456832886
 Step 1640: loss 1.8748414516448975
 Step 1650: loss 1.8823695182800293
 Step 1660: loss 1.90365731716156
 Step 1670: loss 1.7903388738632202
 Step 1680: loss 1.8828151226043701
 Step 1690: loss 1.7718700170516968
 Step 1700: loss 1.845694899559021
 Step 1710: loss 1.8870835304260254
 Step 1720: loss 1.7741565704345703
 Step 1730: loss 1.8264049291610718
 Step 1740: loss 1.8435662984848022
 Step 1750: loss 1.9445271492004395
 Step 1760: loss 1.8523859977722168
 Step 1770: loss 1.7793605327606201
 Step 1780: loss 1.85501229763031
 Step 1790: loss 1.9152953624725342
 Step 1800: loss 1.6987403631210327
 [Seed 77] Step 1800 --> val loss 2.046043872833252
 Step 1810: loss 1.7444498538970947
 Step 1820: loss 1.866021990776062
 Step 1830: loss 1.7845433950424194
 Step 1840: loss 1.8471848964691162
 Step 1850: loss 1.8292028903961182
 Step 1860: loss 1.6583340167999268
 Step 1870: loss 1.8003928661346436
 Step 1880: loss 1.7917121648788452
 Step 1890: loss 1.837627649307251
 Step 1900: loss 1.6754435300827026
 Step 1910: loss 1.8062150478363037
 Step 1920: loss 1.8107566833496094
 Step 1930: loss 1.8704564571380615
 Step 1940: loss 1.692286491394043
 Step 1950: loss 1.7320327758789062
 Step 1960: loss 1.6999263763427734
 Step 1970: loss 1.708532691001892
 Step 1980: loss 1.7723281383514404
 Step 1990: loss 1.8139827251434326
 Step 2000: loss 1.6862553358078003
 [Seed 77] Step 2000 --> val loss 1.9886068105697632
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [43 47 56  1 50 47 60 43 57  1] ...
 First target sequence: [47 56  1 50 47 60 43 57  1 40] ...

 Decoded input: eir lives before:
 The setting sun, and music at the close,
 As th

 Decoded target: ir lives before:
 The setting sun, and music at the close,
 As the
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.198906898498535
 [Seed 78] Step 0 --> val loss 4.1847991943359375
 Step 10: loss 3.846385955810547
 Step 20: loss 3.295635223388672
 Step 30: loss 2.8110761642456055
 Step 40: loss 2.6981379985809326
 Step 50: loss 2.6466617584228516
 Step 60: loss 2.499514102935791
 Step 70: loss 2.4460315704345703
 Step 80: loss 2.356025218963623
 Step 90: loss 2.3439693450927734
 Step 100: loss 2.32150936126709
 Step 110: loss 2.243891716003418
 Step 120: loss 2.320338726043701
 Step 130: loss 2.2308168411254883
 Step 140: loss 2.160604476928711
 Step 150: loss 2.1657989025115967
 Step 160: loss 2.1052193641662598
 Step 170: loss 2.0880935192108154
 Step 180: loss 2.234389305114746
 Step 190: loss 2.2117202281951904
 Step 200: loss 2.1074013710021973
 [Seed 78] Step 200 --> val loss 2.4382030963897705
 Step 210: loss 2.212217092514038
 Step 220: loss 2.078232526779175
 Step 230: loss 2.03403902053833
 Step 240: loss 1.9930601119995117
 Step 250: loss 2.1384387016296387
 Step 260: loss 1.9696815013885498
 Step 270: loss 2.038912296295166
 Step 280: loss 2.102450132369995
 Step 290: loss 2.0943102836608887
 Step 300: loss 2.0844364166259766
 Step 310: loss 2.031684398651123
 Step 320: loss 2.0144195556640625
 Step 330: loss 2.0372378826141357
 Step 340: loss 2.0194220542907715
 Step 350: loss 1.9305229187011719
 Step 360: loss 1.988365888595581
 Step 370: loss 1.9661378860473633
 Step 380: loss 1.97316312789917
 Step 390: loss 2.044494152069092
 Step 400: loss 1.9951289892196655
 [Seed 78] Step 400 --> val loss 2.2551205158233643
 Step 410: loss 2.0175375938415527
 Step 420: loss 1.9180768728256226
 Step 430: loss 1.8973348140716553
 Step 440: loss 2.041095733642578
 Step 450: loss 2.0350632667541504
 Step 460: loss 2.076789140701294
 Step 470: loss 1.984352469444275
 Step 480: loss 2.0698413848876953
 Step 490: loss 1.9435285329818726
 Step 500: loss 2.130082607269287
 Step 510: loss 2.006744861602783
 Step 520: loss 1.9228456020355225
 Step 530: loss 2.0604546070098877
 Step 540: loss 1.9458011388778687
 Step 550: loss 1.9847867488861084
 Step 560: loss 1.9774903059005737
 Step 570: loss 1.9940065145492554
 Step 580: loss 1.9589030742645264
 Step 590: loss 2.05318021774292
 Step 600: loss 1.984629511833191
 [Seed 78] Step 600 --> val loss 2.2397093772888184
 Step 610: loss 2.1297121047973633
 Step 620: loss 1.8913593292236328
 Step 630: loss 2.034017324447632
 Step 640: loss 2.020071029663086
 Step 650: loss 1.9774179458618164
 Step 660: loss 2.020045042037964
 Step 670: loss 1.9263561964035034
 Step 680: loss 2.009296417236328
 Step 690: loss 1.9584510326385498
 Step 700: loss 2.079542875289917
 Step 710: loss 1.9110865592956543
 Step 720: loss 2.023859739303589
 Step 730: loss 2.076953887939453
 Step 740: loss 2.0046892166137695
 Step 750: loss 2.0182888507843018
 Step 760: loss 1.9430373907089233
 Step 770: loss 2.0016212463378906
 Step 780: loss 2.005746603012085
 Step 790: loss 2.009490728378296
 Step 800: loss 2.059674024581909
 [Seed 78] Step 800 --> val loss 2.2214856147766113
 Step 810: loss 2.0511252880096436
 Step 820: loss 1.9543607234954834
 Step 830: loss 2.024871349334717
 Step 840: loss 2.1882026195526123
 Step 850: loss 1.9413816928863525
 Step 860: loss 2.038119077682495
 Step 870: loss 1.8279670476913452
 Step 880: loss 1.9313125610351562
 Step 890: loss 1.9632673263549805
 Step 900: loss 1.9373053312301636
 Step 910: loss 1.8793747425079346
 Step 920: loss 1.9057201147079468
 Step 930: loss 1.9163435697555542
 Step 940: loss 1.9730377197265625
 Step 950: loss 1.9076926708221436
 Step 960: loss 1.9731677770614624
 Step 970: loss 1.9939439296722412
 Step 980: loss 1.9339139461517334
 Step 990: loss 1.8736317157745361
 Step 1000: loss 1.9378767013549805
 [Seed 78] Step 1000 --> val loss 2.0769870281219482
 Step 1010: loss 1.829817771911621
 Step 1020: loss 1.8964976072311401
 Step 1030: loss 1.948844313621521
 Step 1040: loss 1.9620341062545776
 Step 1050: loss 2.0332727432250977
 Step 1060: loss 2.0080976486206055
 Step 1070: loss 2.0180563926696777
 Step 1080: loss 1.9674972295761108
 Step 1090: loss 1.8550032377243042
 Step 1100: loss 1.889795184135437
 Step 1110: loss 1.9545258283615112
 Step 1120: loss 1.9597082138061523
 Step 1130: loss 1.8319590091705322
 Step 1140: loss 2.0484836101531982
 Step 1150: loss 1.8339488506317139
 Step 1160: loss 1.9115972518920898
 Step 1170: loss 1.8816213607788086
 Step 1180: loss 1.9468226432800293
 Step 1190: loss 1.8731273412704468
 Step 1200: loss 1.9418338537216187
 [Seed 78] Step 1200 --> val loss 2.16398549079895
 Step 1210: loss 1.8402974605560303
 Step 1220: loss 1.8615235090255737
 Step 1230: loss 1.9238927364349365
 Step 1240: loss 1.8504356145858765
 Step 1250: loss 1.8082503080368042
 Step 1260: loss 1.9118856191635132
 Step 1270: loss 1.837433099746704
 Step 1280: loss 1.901918888092041
 Step 1290: loss 1.8428714275360107
 Step 1300: loss 1.901437759399414
 Step 1310: loss 1.9522517919540405
 Step 1320: loss 1.9135879278182983
 Step 1330: loss 1.7887802124023438
 Step 1340: loss 1.9329332113265991
 Step 1350: loss 1.6946548223495483
 Step 1360: loss 1.811279058456421
 Step 1370: loss 1.9061102867126465
 Step 1380: loss 1.7925795316696167
 Step 1390: loss 1.8118836879730225
 Step 1400: loss 1.8205217123031616
 [Seed 78] Step 1400 --> val loss 1.9754161834716797
 Step 1410: loss 1.8712701797485352
 Step 1420: loss 1.720055103302002
 Step 1430: loss 1.9350249767303467
 Step 1440: loss 1.723811149597168
 Step 1450: loss 1.8315489292144775
 Step 1460: loss 1.9130713939666748
 Step 1470: loss 1.972121000289917
 Step 1480: loss 1.85652756690979
 Step 1490: loss 1.7892935276031494
 Step 1500: loss 1.8192682266235352
 Step 1510: loss 1.794363021850586
 Step 1520: loss 1.808228850364685
 Step 1530: loss 1.723382830619812
 Step 1540: loss 1.9366004467010498
 Step 1550: loss 1.874481201171875
 Step 1560: loss 1.7889301776885986
 Step 1570: loss 1.8974612951278687
 Step 1580: loss 1.7033333778381348
 Step 1590: loss 1.8062270879745483
 Step 1600: loss 1.8429453372955322
 [Seed 78] Step 1600 --> val loss 1.875950813293457
 Step 1610: loss 1.78515625
 Step 1620: loss 1.9111392498016357
 Step 1630: loss 1.7777093648910522
 Step 1640: loss 1.7042158842086792
 Step 1650: loss 1.8625268936157227
 Step 1660: loss 1.7484941482543945
 Step 1670: loss 1.7552107572555542
 Step 1680: loss 1.885878324508667
 Step 1690: loss 1.8260544538497925
 Step 1700: loss 1.7400321960449219
 Step 1710: loss 1.6772868633270264
 Step 1720: loss 1.7184418439865112
 Step 1730: loss 1.8187167644500732
 Step 1740: loss 1.747488260269165
 Step 1750: loss 1.5275466442108154
 Step 1760: loss 1.7166657447814941
 Step 1770: loss 1.6348541975021362
 Step 1780: loss 1.7805900573730469
 Step 1790: loss 1.6975069046020508
 Step 1800: loss 1.744732141494751
 [Seed 78] Step 1800 --> val loss 1.8563727140426636
 Step 1810: loss 1.7261359691619873
 Step 1820: loss 1.752084732055664
 Step 1830: loss 1.7959622144699097
 Step 1840: loss 1.7408298254013062
 Step 1850: loss 1.778021216392517
 Step 1860: loss 1.7466371059417725
 Step 1870: loss 1.7306621074676514
 Step 1880: loss 1.9680126905441284
 Step 1890: loss 1.8220503330230713
 Step 1900: loss 1.6806046962738037
 Step 1910: loss 1.6932872533798218
 Step 1920: loss 1.7165969610214233
 Step 1930: loss 1.6276592016220093
 Step 1940: loss 1.6854877471923828
 Step 1950: loss 1.5891554355621338
 Step 1960: loss 1.766082763671875
 Step 1970: loss 1.6553728580474854
 Step 1980: loss 1.6957356929779053
 Step 1990: loss 1.7018020153045654
 Step 2000: loss 1.6646140813827515
 [Seed 78] Step 2000 --> val loss 1.824994683265686
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [49 47 52 45  1 45 56 43 39 57] ...
 First target sequence: [47 52 45  1 45 56 43 39 57 63] ...

 Decoded input: king greasy caps in hooting at
 Coriolanus' exile. Now he's comin

 Decoded target: ing greasy caps in hooting at
 Coriolanus' exile. Now he's coming
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.211041450500488
 [Seed 79] Step 0 --> val loss 4.187140941619873
 Step 10: loss 3.863737106323242
 Step 20: loss 3.369731903076172
 Step 30: loss 2.8533859252929688
 Step 40: loss 2.747398853302002
 Step 50: loss 2.5686817169189453
 Step 60: loss 2.4826884269714355
 Step 70: loss 2.3859379291534424
 Step 80: loss 2.339049816131592
 Step 90: loss 2.2948498725891113
 Step 100: loss 2.306863307952881
 Step 110: loss 2.1815555095672607
 Step 120: loss 2.2007322311401367
 Step 130: loss 2.2397003173828125
 Step 140: loss 2.2380924224853516
 Step 150: loss 2.137423515319824
 Step 160: loss 2.0326781272888184
 Step 170: loss 2.1656620502471924
 Step 180: loss 2.044822931289673
 Step 190: loss 2.037074089050293
 Step 200: loss 2.1006393432617188
 [Seed 79] Step 200 --> val loss 2.3000428676605225
 Step 210: loss 2.144577980041504
 Step 220: loss 2.071931838989258
 Step 230: loss 2.133544921875
 Step 240: loss 2.0695948600769043
 Step 250: loss 2.049825668334961
 Step 260: loss 2.047691583633423
 Step 270: loss 2.0104446411132812
 Step 280: loss 1.9081082344055176
 Step 290: loss 2.053179979324341
 Step 300: loss 2.100318670272827
 Step 310: loss 1.9659241437911987
 Step 320: loss 2.0139386653900146
 Step 330: loss 1.9633920192718506
 Step 340: loss 2.024235486984253
 Step 350: loss 2.0778255462646484
 Step 360: loss 2.0163302421569824
 Step 370: loss 1.9787068367004395
 Step 380: loss 1.8876936435699463
 Step 390: loss 1.9437388181686401
 Step 400: loss 1.950143575668335
 [Seed 79] Step 400 --> val loss 2.2929863929748535
 Step 410: loss 2.075932025909424
 Step 420: loss 1.9905766248703003
 Step 430: loss 2.0374088287353516
 Step 440: loss 1.8775641918182373
 Step 450: loss 2.00618839263916
 Step 460: loss 2.0639846324920654
 Step 470: loss 1.9274804592132568
 Step 480: loss 1.9744315147399902
 Step 490: loss 2.074801445007324
 Step 500: loss 1.8552567958831787
 Step 510: loss 2.013519763946533
 Step 520: loss 1.9785264730453491
 Step 530: loss 1.9939059019088745
 Step 540: loss 2.0012078285217285
 Step 550: loss 2.0719966888427734
 Step 560: loss 2.0477519035339355
 Step 570: loss 1.9650176763534546
 Step 580: loss 1.91118323802948
 Step 590: loss 2.045917510986328
 Step 600: loss 1.96584153175354
 [Seed 79] Step 600 --> val loss 2.137956380844116
 Step 610: loss 2.0107860565185547
 Step 620: loss 1.9868663549423218
 Step 630: loss 2.011131763458252
 Step 640: loss 2.085308790206909
 Step 650: loss 1.9560706615447998
 Step 660: loss 1.99643874168396
 Step 670: loss 1.9303264617919922
 Step 680: loss 1.8459663391113281
 Step 690: loss 1.940706729888916
 Step 700: loss 1.9525537490844727
 Step 710: loss 2.0019731521606445
 Step 720: loss 1.9645049571990967
 Step 730: loss 1.9923803806304932
 Step 740: loss 2.0927886962890625
 Step 750: loss 2.085845470428467
 Step 760: loss 2.0884480476379395
 Step 770: loss 2.0318894386291504
 Step 780: loss 2.105351209640503
 Step 790: loss 1.949115514755249
 Step 800: loss 2.0502920150756836
 [Seed 79] Step 800 --> val loss 2.1604011058807373
 Step 810: loss 2.035769462585449
 Step 820: loss 2.050510883331299
 Step 830: loss 1.922978162765503
 Step 840: loss 2.072209358215332
 Step 850: loss 1.9804399013519287
 Step 860: loss 1.905731201171875
 Step 870: loss 2.0223751068115234
 Step 880: loss 1.9953887462615967
 Step 890: loss 1.990746259689331
 Step 900: loss 2.023134708404541
 Step 910: loss 2.034759998321533
 Step 920: loss 1.9717206954956055
 Step 930: loss 2.079010009765625
 Step 940: loss 2.035027265548706
 Step 950: loss 2.0119521617889404
 Step 960: loss 2.081010341644287
 Step 970: loss 1.924826979637146
 Step 980: loss 1.85829496383667
 Step 990: loss 1.942559003829956
 Step 1000: loss 1.8803865909576416
 [Seed 79] Step 1000 --> val loss 2.2022950649261475
 Step 1010: loss 1.9996745586395264
 Step 1020: loss 2.0402393341064453
 Step 1030: loss 2.013519763946533
 Step 1040: loss 2.0477983951568604
 Step 1050: loss 1.9068772792816162
 Step 1060: loss 1.9064607620239258
 Step 1070: loss 2.097031354904175
 Step 1080: loss 1.9587091207504272
 Step 1090: loss 2.0454981327056885
 Step 1100: loss 1.8505853414535522
 Step 1110: loss 2.0144717693328857
 Step 1120: loss 2.0056815147399902
 Step 1130: loss 2.0244269371032715
 Step 1140: loss 1.9794425964355469
 Step 1150: loss 2.0517067909240723
 Step 1160: loss 1.8797285556793213
 Step 1170: loss 1.9357560873031616
 Step 1180: loss 1.9858602285385132
 Step 1190: loss 1.9363855123519897
 Step 1200: loss 2.035330295562744
 [Seed 79] Step 1200 --> val loss 2.0888919830322266
 Step 1210: loss 2.0044002532958984
 Step 1220: loss 1.9063947200775146
 Step 1230: loss 1.9681193828582764
 Step 1240: loss 1.9456032514572144
 Step 1250: loss 2.055513858795166
 Step 1260: loss 1.9745311737060547
 Step 1270: loss 2.01359224319458
 Step 1280: loss 1.9434478282928467
 Step 1290: loss 1.9278725385665894
 Step 1300: loss 1.9597896337509155
 Step 1310: loss 1.971977949142456
 Step 1320: loss 1.8644664287567139
 Step 1330: loss 1.8705346584320068
 Step 1340: loss 1.892362356185913
 Step 1350: loss 1.854672908782959
 Step 1360: loss 1.835493803024292
 Step 1370: loss 2.0543313026428223
 Step 1380: loss 2.0695366859436035
 Step 1390: loss 1.809362769126892
 Step 1400: loss 1.9514020681381226
 [Seed 79] Step 1400 --> val loss 2.1066834926605225
 Step 1410: loss 1.9801740646362305
 Step 1420: loss 1.7998641729354858
 Step 1430: loss 2.0113987922668457
 Step 1440: loss 1.9650405645370483
 Step 1450: loss 1.957930564880371
 Step 1460: loss 1.8829377889633179
 Step 1470: loss 1.8933976888656616
 Step 1480: loss 1.9712083339691162
 Step 1490: loss 2.052433967590332
 Step 1500: loss 2.071227788925171
 Step 1510: loss 1.9504528045654297
 Step 1520: loss 1.917473554611206
 Step 1530: loss 1.9873679876327515
 Step 1540: loss 1.9228448867797852
 Step 1550: loss 1.898816466331482
 Step 1560: loss 1.9112046957015991
 Step 1570: loss 1.872094988822937
 Step 1580: loss 1.9267983436584473
 Step 1590: loss 1.8589200973510742
 Step 1600: loss 1.9451541900634766
 [Seed 79] Step 1600 --> val loss 2.034064531326294
 Step 1610: loss 1.7820004224777222
 Step 1620: loss 1.9048442840576172
 Step 1630: loss 1.8248989582061768
 Step 1640: loss 1.8918828964233398
 Step 1650: loss 1.8897866010665894
 Step 1660: loss 1.8486899137496948
 Step 1670: loss 1.9081997871398926
 Step 1680: loss 1.7648115158081055
 Step 1690: loss 1.8476009368896484
 Step 1700: loss 1.8974502086639404
 Step 1710: loss 1.9586288928985596
 Step 1720: loss 1.918530821800232
 Step 1730: loss 1.8103402853012085
 Step 1740: loss 1.8744395971298218
 Step 1750: loss 1.7416448593139648
 Step 1760: loss 1.7618401050567627
 Step 1770: loss 1.8796539306640625
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 79] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 79] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [45 39 47 52  1 57 46 39 50 50] ...
 First target sequence: [39 47 52  1 57 46 39 50 50  1] ...

 Decoded input: gain shall you be mother to a king,
 And all the ruins of distres

 Decoded target: ain shall you be mother to a king,
 And all the ruins of distress
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.206831455230713
 [Seed 80] Step 0 --> val loss 4.185006618499756
 Step 10: loss 3.850675344467163
 Step 20: loss 3.3042778968811035
 Step 30: loss 2.888476848602295
 Step 40: loss 2.743706226348877
 Step 50: loss 2.6416664123535156
 Step 60: loss 2.4373650550842285
 Step 70: loss 2.4284465312957764
 Step 80: loss 2.3593719005584717
 Step 90: loss 2.3368453979492188
 Step 100: loss 2.351700782775879
 Step 110: loss 2.2463531494140625
 Step 120: loss 2.2408714294433594
 Step 130: loss 2.2296831607818604
 Step 140: loss 2.2978196144104004
 Step 150: loss 2.184248208999634
 Step 160: loss 2.1216912269592285
 Step 170: loss 2.1181697845458984
 Step 180: loss 2.067714214324951
 Step 190: loss 2.2185466289520264
 Step 200: loss 2.198219060897827
 [Seed 80] Step 200 --> val loss 2.291472911834717
 Step 210: loss 2.085797071456909
 Step 220: loss 2.1006999015808105
 Step 230: loss 2.057856321334839
 Step 240: loss 2.141735076904297
 Step 250: loss 2.1448493003845215
 Step 260: loss 2.0998847484588623
 Step 270: loss 2.114098310470581
 Step 280: loss 2.084829807281494
 Step 290: loss 2.0640172958374023
 Step 300: loss 2.129506826400757
 Step 310: loss 2.0213944911956787
 Step 320: loss 2.0726399421691895
 Step 330: loss 2.002641439437866
 Step 340: loss 2.173043966293335
 Step 350: loss 2.0449724197387695
 Step 360: loss 2.0452537536621094
 Step 370: loss 2.034343719482422
 Step 380: loss 2.171433210372925
 Step 390: loss 2.0386910438537598
 Step 400: loss 2.006350040435791
 [Seed 80] Step 400 --> val loss 2.3849711418151855
 Step 410: loss 2.029789924621582
 Step 420: loss 1.9945504665374756
 Step 430: loss 2.064343214035034
 Step 440: loss 2.1253275871276855
 Step 450: loss 2.060121536254883
 Step 460: loss 1.9525388479232788
 Step 470: loss 2.136779546737671
 Step 480: loss 1.961089849472046
 Step 490: loss 1.984407901763916
 Step 500: loss 2.0029664039611816
 Step 510: loss 1.9966380596160889
 Step 520: loss 1.9923995733261108
 Step 530: loss 1.9500125646591187
 Step 540: loss 1.962120771408081
 Step 550: loss 2.0804691314697266
 Step 560: loss 1.9557616710662842
 Step 570: loss 1.9975438117980957
 Step 580: loss 1.9698556661605835
 Step 590: loss 1.9840803146362305
 Step 600: loss 2.0035006999969482
 [Seed 80] Step 600 --> val loss 2.1909167766571045
 Step 610: loss 1.9557015895843506
 Step 620: loss 1.9813201427459717
 Step 630: loss 2.0591230392456055
 Step 640: loss 1.9489073753356934
 Step 650: loss 1.956451416015625
 Step 660: loss 2.125211000442505
 Step 670: loss 1.8523039817810059
 Step 680: loss 2.0038986206054688
 Step 690: loss 1.9281229972839355
 Step 700: loss 1.9250472784042358
 Step 710: loss 2.0779919624328613
 Step 720: loss 2.0655741691589355
 Step 730: loss 1.9044955968856812
 Step 740: loss 1.951646327972412
 Step 750: loss 1.9376262426376343
 Step 760: loss 1.946065902709961
 Step 770: loss 1.9604251384735107
 Step 780: loss 1.7837984561920166
 Step 790: loss 2.068438768386841
 Step 800: loss 1.947993278503418
 [Seed 80] Step 800 --> val loss 2.2788453102111816
 Step 810: loss 1.986926555633545
 Step 820: loss 1.9464573860168457
 Step 830: loss 1.9745129346847534
 Step 840: loss 1.852465033531189
 Step 850: loss 1.9011573791503906
 Step 860: loss 1.8392475843429565
 Step 870: loss 1.9297494888305664
 Step 880: loss 1.9525903463363647
 Step 890: loss 1.9412224292755127
 Step 900: loss 1.8651107549667358
 Step 910: loss 2.000880241394043
 Step 920: loss 1.9474716186523438
 Step 930: loss 1.9453943967819214
 Step 940: loss 1.94724440574646
 Step 950: loss 1.9085029363632202
 Step 960: loss 1.8743155002593994
 Step 970: loss 1.998175024986267
 Step 980: loss 1.985371470451355
 Step 990: loss 2.0600147247314453
 Step 1000: loss 1.9815118312835693
 [Seed 80] Step 1000 --> val loss 2.212494134902954
 Step 1010: loss 1.9502055644989014
 Step 1020: loss 1.9992356300354004
 Step 1030: loss 1.9390015602111816
 Step 1040: loss 1.9716800451278687
 Step 1050: loss 1.9523983001708984
 Step 1060: loss 1.9742755889892578
 Step 1070: loss 2.0663022994995117
 Step 1080: loss 2.003021240234375
 Step 1090: loss 1.953234076499939
 Step 1100: loss 1.8505592346191406
 Step 1110: loss 1.8739955425262451
 Step 1120: loss 2.0627615451812744
 Step 1130: loss 1.9619505405426025
 Step 1140: loss 1.9679001569747925
 Step 1150: loss 1.9800103902816772
 Step 1160: loss 2.104184627532959
 Step 1170: loss 1.8705024719238281
 Step 1180: loss 1.9347617626190186
 Step 1190: loss 2.0493712425231934
 Step 1200: loss 1.8557212352752686
 [Seed 80] Step 1200 --> val loss 2.1698157787323
 Step 1210: loss 1.87381112575531
 Step 1220: loss 1.8736793994903564
 Step 1230: loss 1.8984315395355225
 Step 1240: loss 1.9477161169052124
 Step 1250: loss 1.9424928426742554
 Step 1260: loss 1.9758418798446655
 Step 1270: loss 1.8613733053207397
 Step 1280: loss 1.8101835250854492
 Step 1290: loss 1.8561744689941406
 Step 1300: loss 1.801294207572937
 Step 1310: loss 1.8551878929138184
 Step 1320: loss 1.9061040878295898
 Step 1330: loss 1.944164514541626
 Step 1340: loss 1.8734476566314697
 Step 1350: loss 1.8017349243164062
 Step 1360: loss 1.8803787231445312
 Step 1370: loss 1.8559739589691162
 Step 1380: loss 1.9152135848999023
 Step 1390: loss 1.8256416320800781
 Step 1400: loss 1.8445491790771484
 [Seed 80] Step 1400 --> val loss 2.019932270050049
 Step 1410: loss 1.7982953786849976
 Step 1420: loss 1.8538817167282104
 Step 1430: loss 1.8234885931015015
 Step 1440: loss 1.8062350749969482
 Step 1450: loss 1.762903094291687
 Step 1460: loss 1.8894637823104858
 Step 1470: loss 1.887450933456421
 Step 1480: loss 1.8185886144638062
 Step 1490: loss 1.9717426300048828
 Step 1500: loss 1.8229689598083496
 Step 1510: loss 1.813024878501892
 Step 1520: loss 1.7646617889404297
 Step 1530: loss 1.8224294185638428
 Step 1540: loss 1.8186540603637695
 Step 1550: loss 1.7341583967208862
 Step 1560: loss 1.8048375844955444
 Step 1570: loss 1.7659955024719238
 Step 1580: loss 1.8414058685302734
 Step 1590: loss 1.8875325918197632
 Step 1600: loss 1.6648170948028564
 [Seed 80] Step 1600 --> val loss 1.8634716272354126
 Step 1610: loss 1.7913626432418823
 Step 1620: loss 1.795885443687439
 Step 1630: loss 1.7720803022384644
 Step 1640: loss 1.7429031133651733
 Step 1650: loss 1.6484570503234863
 Step 1660: loss 1.7566474676132202
 Step 1670: loss 1.77055025100708
 Step 1680: loss 1.6303536891937256
 Step 1690: loss 1.7651361227035522
 Step 1700: loss 1.8006000518798828
 Step 1710: loss 1.7483372688293457
 Step 1720: loss 1.714181900024414
 Step 1730: loss 1.7836278676986694
 Step 1740: loss 1.7535290718078613
 Step 1750: loss 1.8028587102890015
 Step 1760: loss 1.7968316078186035
 Step 1770: loss 1.7122371196746826
 Step 1780: loss 1.720716953277588
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 80] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 80] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [52  0 39 56 43  1 58 46 43 63] ...
 First target sequence: [ 0 39 56 43  1 58 46 43 63  1] ...

 Decoded input: n
 are they gone, and there they intend to sup.

 Second Gentleman

 Decoded target: 
 are they gone, and there they intend to sup.

 Second Gentleman:
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.216590404510498
 [Seed 81] Step 0 --> val loss 4.183829307556152
 Step 10: loss 3.837881565093994
 Step 20: loss 3.3561782836914062
 Step 30: loss 2.9131107330322266
 Step 40: loss 2.785977840423584
 Step 50: loss 2.51660418510437
 Step 60: loss 2.508964776992798
 Step 70: loss 2.491838216781616
 Step 80: loss 2.4409289360046387
 Step 90: loss 2.2659053802490234
 Step 100: loss 2.2667787075042725
 Step 110: loss 2.310715675354004
 Step 120: loss 2.195713996887207
 Step 130: loss 2.278374433517456
 Step 140: loss 2.127751350402832
 Step 150: loss 2.19769287109375
 Step 160: loss 2.1621148586273193
 Step 170: loss 2.1593801975250244
 Step 180: loss 2.1247940063476562
 Step 190: loss 1.9328206777572632
 Step 200: loss 2.174811363220215
 [Seed 81] Step 200 --> val loss 2.40268874168396
 Step 210: loss 2.127901792526245
 Step 220: loss 2.1369481086730957
 Step 230: loss 2.0018250942230225
 Step 240: loss 2.118126392364502
 Step 250: loss 2.0431370735168457
 Step 260: loss 2.088432788848877
 Step 270: loss 1.9765676259994507
 Step 280: loss 2.0839478969573975
 Step 290: loss 2.014462947845459
 Step 300: loss 2.082432270050049
 Step 310: loss 2.101167678833008
 Step 320: loss 2.2179205417633057
 Step 330: loss 2.132449150085449
 Step 340: loss 2.0787875652313232
 Step 350: loss 2.0508480072021484
 Step 360: loss 1.9904861450195312
 Step 370: loss 1.9435901641845703
 Step 380: loss 1.9979842901229858
 Step 390: loss 2.0711278915405273
 Step 400: loss 2.041184425354004
 [Seed 81] Step 400 --> val loss 2.284074544906616
 Step 410: loss 2.006596326828003
 Step 420: loss 1.9767873287200928
 Step 430: loss 2.089907646179199
 Step 440: loss 2.108799695968628
 Step 450: loss 2.085677146911621
 Step 460: loss 2.0574910640716553
 Step 470: loss 2.155090570449829
 Step 480: loss 2.100649833679199
 Step 490: loss 1.8779186010360718
 Step 500: loss 1.9935646057128906
 Step 510: loss 2.172569751739502
 Step 520: loss 2.065584659576416
 Step 530: loss 2.078472852706909
 Step 540: loss 2.1045432090759277
 Step 550: loss 2.0225892066955566
 Step 560: loss 1.9684170484542847
 Step 570: loss 2.163813591003418
 Step 580: loss 1.9781172275543213
 Step 590: loss 1.9305100440979004
 Step 600: loss 2.083003520965576
 [Seed 81] Step 600 --> val loss 2.112138032913208
 Step 610: loss 1.9567617177963257
 Step 620: loss 1.9823168516159058
 Step 630: loss 2.026179075241089
 Step 640: loss 1.9801499843597412
 Step 650: loss 1.9525299072265625
 Step 660: loss 2.036294937133789
 Step 670: loss 1.9579318761825562
 Step 680: loss 2.013704776763916
 Step 690: loss 1.9143928289413452
 Step 700: loss 2.0704963207244873
 Step 710: loss 2.072890520095825
 Step 720: loss 2.0698370933532715
 Step 730: loss 1.9383115768432617
 Step 740: loss 1.9339053630828857
 Step 750: loss 2.030749559402466
 Step 760: loss 1.9955013990402222
 Step 770: loss 2.104043960571289
 Step 780: loss 2.1163554191589355
 Step 790: loss 2.0080809593200684
 Step 800: loss 2.049516439437866
 [Seed 81] Step 800 --> val loss 2.254798173904419
 Step 810: loss 1.9477097988128662
 Step 820: loss 2.0317437648773193
 Step 830: loss 2.068025827407837
 Step 840: loss 1.9380168914794922
 Step 850: loss 1.9721307754516602
 Step 860: loss 2.1604299545288086
 Step 870: loss 1.9198729991912842
 Step 880: loss 1.9395277500152588
 Step 890: loss 1.9565010070800781
 Step 900: loss 1.968942403793335
 Step 910: loss 2.0876235961914062
 Step 920: loss 2.1074445247650146
 Step 930: loss 2.0593113899230957
 Step 940: loss 1.9105943441390991
 Step 950: loss 2.000558853149414
 Step 960: loss 2.006321907043457
 Step 970: loss 2.0683765411376953
 Step 980: loss 2.0031232833862305
 Step 990: loss 2.0772838592529297
 Step 1000: loss 2.0190587043762207
 [Seed 81] Step 1000 --> val loss 2.0825507640838623
 Step 1010: loss 1.9212870597839355
 Step 1020: loss 1.9100589752197266
 Step 1030: loss 1.9457283020019531
 Step 1040: loss 1.9633076190948486
 Step 1050: loss 2.0399434566497803
 Step 1060: loss 1.903369665145874
 Step 1070: loss 1.916029691696167
 Step 1080: loss 1.970180869102478
 Step 1090: loss 1.942129135131836
 Step 1100: loss 1.866194486618042
 Step 1110: loss 2.015482187271118
 Step 1120: loss 1.9303492307662964
 Step 1130: loss 1.9357810020446777
 Step 1140: loss 1.9005188941955566
 Step 1150: loss 1.8820810317993164
 Step 1160: loss 1.8792874813079834
 Step 1170: loss 1.8769972324371338
 Step 1180: loss 1.9558935165405273
 Step 1190: loss 1.8861501216888428
 Step 1200: loss 1.8730056285858154
 [Seed 81] Step 1200 --> val loss 2.209883451461792
 Step 1210: loss 2.017765522003174
 Step 1220: loss 1.8536841869354248
 Step 1230: loss 1.9769432544708252
 Step 1240: loss 1.921461820602417
 Step 1250: loss 1.9520163536071777
 Step 1260: loss 1.8871272802352905
 Step 1270: loss 1.915984034538269
 Step 1280: loss 1.9856574535369873
 Step 1290: loss 1.9689111709594727
 Step 1300: loss 1.8610553741455078
 Step 1310: loss 1.7510846853256226
 Step 1320: loss 1.8984415531158447
 Step 1330: loss 1.9250620603561401
 Step 1340: loss 1.8671624660491943
 Step 1350: loss 1.9486918449401855
 Step 1360: loss 1.8610528707504272
 Step 1370: loss 1.9132957458496094
 Step 1380: loss 1.8663325309753418
 Step 1390: loss 1.9837427139282227
 Step 1400: loss 1.7783575057983398
 [Seed 81] Step 1400 --> val loss 2.1125357151031494
 Step 1410: loss 2.0115952491760254
 Step 1420: loss 1.826621413230896
 Step 1430: loss 1.8255114555358887
 Step 1440: loss 1.8423664569854736
 Step 1450: loss 1.927990436553955
 Step 1460: loss 1.888956069946289
 Step 1470: loss 1.9213465452194214
 Step 1480: loss 1.8898897171020508
 Step 1490: loss 2.0029783248901367
 Step 1500: loss 1.8591477870941162
 Step 1510: loss 1.8538885116577148
 Step 1520: loss 1.8405673503875732
 Step 1530: loss 1.959665298461914
 Step 1540: loss 1.8463799953460693
 Step 1550: loss 1.8826308250427246
 Step 1560: loss 1.865929365158081
 Step 1570: loss 1.8442208766937256
 Step 1580: loss 1.736783742904663
 Step 1590: loss 1.8103340864181519
 Step 1600: loss 1.8188289403915405
 [Seed 81] Step 1600 --> val loss 1.9783884286880493
 Step 1610: loss 1.7838020324707031
 Step 1620: loss 1.7980865240097046
 Step 1630: loss 1.8360399007797241
 Step 1640: loss 1.8419660329818726
 Step 1650: loss 1.939809799194336
 Step 1660: loss 1.718679428100586
 Step 1670: loss 1.8850629329681396
 Step 1680: loss 1.8928039073944092
 Step 1690: loss 1.7355531454086304
 Step 1700: loss 1.8115918636322021
 Step 1710: loss 1.6941890716552734
 Step 1720: loss 1.8064744472503662
 Step 1730: loss 1.7582404613494873
 Step 1740: loss 1.686516523361206
 Step 1750: loss 1.789748191833496
 Step 1760: loss 1.759622573852539
 Step 1770: loss 1.735068678855896
 Step 1780: loss 1.7797788381576538
 Step 1790: loss 1.680688738822937
 Step 1800: loss 1.8573451042175293
 [Seed 81] Step 1800 --> val loss 1.9137626886367798
 Step 1810: loss 1.7898238897323608
 Step 1820: loss 1.8271969556808472
 Step 1830: loss 1.5991125106811523
 Step 1840: loss 1.7889952659606934
 Step 1850: loss 1.6869806051254272
 Step 1860: loss 1.707701563835144
 Step 1870: loss 1.80535888671875
 Step 1880: loss 1.7868603467941284
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 81] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 52 53 58  1 39 58  1 63 53] ...
 First target sequence: [52 53 58  1 39 58  1 63 53 59] ...

 Decoded input:  not at your father's house these
 seven years
 Be born another su

 Decoded target: not at your father's house these
 seven years
 Be born another suc
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.211050987243652
 [Seed 82] Step 0 --> val loss 4.186384201049805
 Step 10: loss 3.8270812034606934
 Step 20: loss 3.2917895317077637
 Step 30: loss 2.8729934692382812
 Step 40: loss 2.723050594329834
 Step 50: loss 2.6240124702453613
 Step 60: loss 2.5082685947418213
 Step 70: loss 2.3645381927490234
 Step 80: loss 2.5036847591400146
 Step 90: loss 2.333146572113037
 Step 100: loss 2.3754262924194336
 Step 110: loss 2.1939988136291504
 Step 120: loss 2.249936103820801
 Step 130: loss 2.220360279083252
 Step 140: loss 2.297139883041382
 Step 150: loss 2.148557186126709
 Step 160: loss 2.2122700214385986
 Step 170: loss 2.132720708847046
 Step 180: loss 2.187612771987915
 Step 190: loss 2.0932698249816895
 Step 200: loss 2.079956531524658
 [Seed 82] Step 200 --> val loss 2.3640270233154297
 Step 210: loss 2.1756510734558105
 Step 220: loss 2.0301008224487305
 Step 230: loss 2.119682788848877
 Step 240: loss 2.175198554992676
 Step 250: loss 2.1156864166259766
 Step 260: loss 1.9828615188598633
 Step 270: loss 2.1628518104553223
 Step 280: loss 1.9453887939453125
 Step 290: loss 2.041618824005127
 Step 300: loss 1.9791045188903809
 Step 310: loss 2.0426642894744873
 Step 320: loss 2.0506856441497803
 Step 330: loss 2.0220861434936523
 Step 340: loss 2.0375163555145264
 Step 350: loss 2.1128735542297363
 Step 360: loss 2.0121641159057617
 Step 370: loss 2.017594575881958
 Step 380: loss 2.0549867153167725
 Step 390: loss 2.035861015319824
 Step 400: loss 2.0243778228759766
 [Seed 82] Step 400 --> val loss 2.2216546535491943
 Step 410: loss 2.058060646057129
 Step 420: loss 2.022008180618286
 Step 430: loss 1.8781604766845703
 Step 440: loss 1.952755331993103
 Step 450: loss 2.0070319175720215
 Step 460: loss 1.8610938787460327
 Step 470: loss 1.9885845184326172
 Step 480: loss 2.0978894233703613
 Step 490: loss 1.931633472442627
 Step 500: loss 2.125225782394409
 Step 510: loss 2.0641555786132812
 Step 520: loss 2.0163204669952393
 Step 530: loss 1.9196933507919312
 Step 540: loss 2.011556386947632
 Step 550: loss 2.060872793197632
 Step 560: loss 2.029510498046875
 Step 570: loss 2.0599331855773926
 Step 580: loss 2.0220887660980225
 Step 590: loss 2.0614993572235107
 Step 600: loss 2.1005682945251465
 [Seed 82] Step 600 --> val loss 2.2557666301727295
 Step 610: loss 2.1614155769348145
 Step 620: loss 2.1160073280334473
 Step 630: loss 2.153590679168701
 Step 640: loss 2.049426555633545
 Step 650: loss 2.076169013977051
 Step 660: loss 2.1560845375061035
 Step 670: loss 2.0143985748291016
 Step 680: loss 2.1618785858154297
 Step 690: loss 2.186947822570801
 Step 700: loss 2.176801919937134
 Step 710: loss 2.0308218002319336
 Step 720: loss 1.9854758977890015
 Step 730: loss 2.0603747367858887
 Step 740: loss 2.0693328380584717
 Step 750: loss 1.99216890335083
 Step 760: loss 2.016965866088867
 Step 770: loss 2.018576145172119
 Step 780: loss 2.0911011695861816
 Step 790: loss 1.9722447395324707
 Step 800: loss 1.9427493810653687
 [Seed 82] Step 800 --> val loss 2.322197675704956
 Step 810: loss 1.9496257305145264
 Step 820: loss 1.9324407577514648
 Step 830: loss 2.0084359645843506
 Step 840: loss 1.9789564609527588
 Step 850: loss 2.150177478790283
 Step 860: loss 2.0352187156677246
 Step 870: loss 2.0137813091278076
 Step 880: loss 2.046579122543335
 Step 890: loss 2.056612491607666
 Step 900: loss 2.068180561065674
 Step 910: loss 1.9746801853179932
 Step 920: loss 2.020500659942627
 Step 930: loss 1.9125837087631226
 Step 940: loss 2.0227150917053223
 Step 950: loss 2.010329008102417
 Step 960: loss 2.0973222255706787
 Step 970: loss 2.0193660259246826
 Step 980: loss 2.2361979484558105
 Step 990: loss 2.1410067081451416
 Step 1000: loss 2.0942232608795166
 [Seed 82] Step 1000 --> val loss 2.304361343383789
 Step 1010: loss 2.04528546333313
 Step 1020: loss 2.0164029598236084
 Step 1030: loss 2.124152660369873
 Step 1040: loss 1.9976314306259155
 Step 1050: loss 2.0214693546295166
 Step 1060: loss 1.940281629562378
 Step 1070: loss 1.966631293296814
 Step 1080: loss 2.0321366786956787
 Step 1090: loss 1.9729266166687012
 Step 1100: loss 1.9611008167266846
 Step 1110: loss 2.1749753952026367
 Step 1120: loss 2.060579776763916
 Step 1130: loss 1.9755454063415527
 Step 1140: loss 1.9928300380706787
 Step 1150: loss 1.9771894216537476
 Step 1160: loss 1.957721471786499
 Step 1170: loss 1.925614833831787
 Step 1180: loss 1.860335111618042
 Step 1190: loss 2.063279151916504
 Step 1200: loss 2.026320695877075
 [Seed 82] Step 1200 --> val loss 2.134077787399292
 Step 1210: loss 1.9254941940307617
 Step 1220: loss 2.0115017890930176
 Step 1230: loss 1.9418071508407593
 Step 1240: loss 2.0729997158050537
 Step 1250: loss 1.7950210571289062
 Step 1260: loss 1.8550286293029785
 Step 1270: loss 1.9097464084625244
 Step 1280: loss 1.9619004726409912
 Step 1290: loss 1.924125075340271
 Step 1300: loss 1.9365304708480835
 Step 1310: loss 1.8159832954406738
 Step 1320: loss 1.9523099660873413
 Step 1330: loss 1.9066970348358154
 Step 1340: loss 1.859262228012085
 Step 1350: loss 1.9531660079956055
 Step 1360: loss 1.908884048461914
 Step 1370: loss 1.9506096839904785
 Step 1380: loss 1.968550443649292
 Step 1390: loss 1.9420557022094727
 Step 1400: loss 1.7986429929733276
 [Seed 82] Step 1400 --> val loss 2.1443405151367188
 Step 1410: loss 2.0256316661834717
 Step 1420: loss 1.8931045532226562
 Step 1430: loss 1.9731662273406982
 Step 1440: loss 1.84528386592865
 Step 1450: loss 1.8636010885238647
 Step 1460: loss 1.8094290494918823
 Step 1470: loss 1.8804035186767578
 Step 1480: loss 1.8813713788986206
 Step 1490: loss 1.9150047302246094
 Step 1500: loss 1.923991084098816
 Step 1510: loss 1.7541418075561523
 Step 1520: loss 1.9095863103866577
 Step 1530: loss 1.8429865837097168
 Step 1540: loss 1.8039335012435913
 Step 1550: loss 1.8440265655517578
 Step 1560: loss 1.8635199069976807
 Step 1570: loss 1.8043346405029297
 Step 1580: loss 1.7415926456451416
 Step 1590: loss 1.7416670322418213
 Step 1600: loss 1.8791804313659668
 [Seed 82] Step 1600 --> val loss 1.9700720310211182
 Step 1610: loss 1.9150123596191406
 Step 1620: loss 1.7946813106536865
 Step 1630: loss 1.7757766246795654
 Step 1640: loss 1.7596986293792725
 Step 1650: loss 1.8178738355636597
 Step 1660: loss 1.832315444946289
 Step 1670: loss 1.8448675870895386
 Step 1680: loss 1.7692615985870361
 Step 1690: loss 1.858963966369629
 Step 1700: loss 1.7849042415618896
 Step 1710: loss 1.8733981847763062
 Step 1720: loss 1.8187720775604248
 Step 1730: loss 1.8013544082641602
 Step 1740: loss 1.799783706665039
 Step 1750: loss 1.8017969131469727
 Step 1760: loss 1.7540442943572998
 Step 1770: loss 1.795483946800232
 Step 1780: loss 1.714187502861023
 Step 1790: loss 1.795208215713501
 Step 1800: loss 1.708435297012329
 [Seed 82] Step 1800 --> val loss 1.9090036153793335
 Step 1810: loss 1.66752028465271
 Step 1820: loss 1.8794538974761963
 Step 1830: loss 1.8358547687530518
 Step 1840: loss 1.6918649673461914
 Step 1850: loss 1.812294840812683
 Step 1860: loss 1.6962745189666748
 Step 1870: loss 1.6958459615707397
 Step 1880: loss 1.7383012771606445
 Step 1890: loss 1.615039587020874
 Step 1900: loss 1.709671974182129
 Step 1910: loss 1.6240994930267334
 Step 1920: loss 1.7219007015228271
 Step 1930: loss 1.6274113655090332
 Step 1940: loss 1.7761800289154053
 Step 1950: loss 1.7322566509246826
 Step 1960: loss 1.70108163356781
 Step 1970: loss 1.7699029445648193
 Step 1980: loss 1.577773094177246
 Step 1990: loss 1.8316242694854736
 Step 2000: loss nan
 [Seed 82] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [50 39 57  6  1 54 53 53 56  1] ...
 First target sequence: [39 57  6  1 54 53 53 56  1 42] ...

 Decoded input: las, poor duke! the task he undertakes
 Is numbering sands and dr

 Decoded target: as, poor duke! the task he undertakes
 Is numbering sands and dri
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2214860916137695
 [Seed 83] Step 0 --> val loss 4.185660362243652
 Step 10: loss 3.8363616466522217
 Step 20: loss 3.3147177696228027
 Step 30: loss 2.7999300956726074
 Step 40: loss 2.7491912841796875
 Step 50: loss 2.558988094329834
 Step 60: loss 2.4729208946228027
 Step 70: loss 2.400697946548462
 Step 80: loss 2.4329209327697754
 Step 90: loss 2.395857810974121
 Step 100: loss 2.275256872177124
 Step 110: loss 2.38272762298584
 Step 120: loss 2.246445417404175
 Step 130: loss 2.204096794128418
 Step 140: loss 2.265981435775757
 Step 150: loss 2.1437149047851562
 Step 160: loss 2.251859188079834
 Step 170: loss 2.1286635398864746
 Step 180: loss 2.0967178344726562
 Step 190: loss 2.1785550117492676
 Step 200: loss 2.1807003021240234
 [Seed 83] Step 200 --> val loss 2.308696746826172
 Step 210: loss 2.1625380516052246
 Step 220: loss 2.200937509536743
 Step 230: loss 2.1559817790985107
 Step 240: loss 2.0932836532592773
 Step 250: loss 1.9895148277282715
 Step 260: loss 2.070904493331909
 Step 270: loss 2.0886330604553223
 Step 280: loss 2.0385220050811768
 Step 290: loss 2.093818187713623
 Step 300: loss 2.1186234951019287
 Step 310: loss 2.0184731483459473
 Step 320: loss 1.9230488538742065
 Step 330: loss 2.0159387588500977
 Step 340: loss 2.1399848461151123
 Step 350: loss 2.2504043579101562
 Step 360: loss 2.031109571456909
 Step 370: loss 2.0250120162963867
 Step 380: loss 2.0247654914855957
 Step 390: loss 1.966759443283081
 Step 400: loss 1.9798239469528198
 [Seed 83] Step 400 --> val loss 2.2902672290802
 Step 410: loss 1.9696846008300781
 Step 420: loss 1.9300611019134521
 Step 430: loss 2.0611534118652344
 Step 440: loss 1.9172770977020264
 Step 450: loss 1.9998741149902344
 Step 460: loss 2.034245729446411
 Step 470: loss 2.015038013458252
 Step 480: loss 1.9655351638793945
 Step 490: loss 2.0637192726135254
 Step 500: loss 2.00238299369812
 Step 510: loss 2.022864818572998
 Step 520: loss 2.0445587635040283
 Step 530: loss 2.1275863647460938
 Step 540: loss 2.152322769165039
 Step 550: loss 2.059638023376465
 Step 560: loss 2.12278413772583
 Step 570: loss 2.009140729904175
 Step 580: loss 1.9959609508514404
 Step 590: loss 1.9952802658081055
 Step 600: loss 1.9151209592819214
 [Seed 83] Step 600 --> val loss 2.2799603939056396
 Step 610: loss 2.080397367477417
 Step 620: loss 1.951867938041687
 Step 630: loss 1.9195140600204468
 Step 640: loss 1.9495656490325928
 Step 650: loss 2.0037965774536133
 Step 660: loss 2.053316116333008
 Step 670: loss 1.9683899879455566
 Step 680: loss 1.9168663024902344
 Step 690: loss 2.0953245162963867
 Step 700: loss 1.9817512035369873
 Step 710: loss 1.9639708995819092
 Step 720: loss 1.9871940612792969
 Step 730: loss 1.9451417922973633
 Step 740: loss 1.9205297231674194
 Step 750: loss 2.0185647010803223
 Step 760: loss 1.973301649093628
 Step 770: loss 1.8997281789779663
 Step 780: loss 1.9690450429916382
 Step 790: loss 2.0143961906433105
 Step 800: loss 2.047234058380127
 [Seed 83] Step 800 --> val loss 2.260263442993164
 Step 810: loss 1.8957427740097046
 Step 820: loss 1.9884618520736694
 Step 830: loss 2.007988452911377
 Step 840: loss 1.960188388824463
 Step 850: loss 2.0581369400024414
 Step 860: loss 2.0324230194091797
 Step 870: loss 2.0115251541137695
 Step 880: loss 1.9941219091415405
 Step 890: loss 2.0079731941223145
 Step 900: loss 2.005678653717041
 Step 910: loss 1.9684157371520996
 Step 920: loss 1.9194602966308594
 Step 930: loss 2.0049731731414795
 Step 940: loss 1.906482219696045
 Step 950: loss 2.0284910202026367
 Step 960: loss 1.8787931203842163
 Step 970: loss 1.95286226272583
 Step 980: loss 1.9113595485687256
 Step 990: loss 2.0699193477630615
 Step 1000: loss 1.9229501485824585
 [Seed 83] Step 1000 --> val loss 2.1689059734344482
 Step 1010: loss 1.9148447513580322
 Step 1020: loss 1.8930091857910156
 Step 1030: loss 1.8227651119232178
 Step 1040: loss 1.797892451286316
 Step 1050: loss 1.9626643657684326
 Step 1060: loss 1.9816187620162964
 Step 1070: loss 1.9864124059677124
 Step 1080: loss 1.9960241317749023
 Step 1090: loss 1.8325872421264648
 Step 1100: loss 1.836305022239685
 Step 1110: loss 1.7828443050384521
 Step 1120: loss 1.867564082145691
 Step 1130: loss 1.9356353282928467
 Step 1140: loss 1.934833288192749
 Step 1150: loss 1.8916358947753906
 Step 1160: loss 1.8863639831542969
 Step 1170: loss 1.952341079711914
 Step 1180: loss 2.0494906902313232
 Step 1190: loss 1.8907322883605957
 Step 1200: loss 1.8748674392700195
 [Seed 83] Step 1200 --> val loss 2.156548500061035
 Step 1210: loss 1.9010604619979858
 Step 1220: loss 1.888655662536621
 Step 1230: loss 2.0112433433532715
 Step 1240: loss 1.886068344116211
 Step 1250: loss 1.8218038082122803
 Step 1260: loss 1.8346433639526367
 Step 1270: loss 1.8764041662216187
 Step 1280: loss 1.92193603515625
 Step 1290: loss 1.856261134147644
 Step 1300: loss 1.9193214178085327
 Step 1310: loss 1.9944400787353516
 Step 1320: loss 1.8005053997039795
 Step 1330: loss 1.8995345830917358
 Step 1340: loss 1.794339895248413
 Step 1350: loss 1.8228037357330322
 Step 1360: loss 1.8407481908798218
 Step 1370: loss 1.8362860679626465
 Step 1380: loss 1.8610074520111084
 Step 1390: loss 1.673161506652832
 Step 1400: loss 1.8027465343475342
 [Seed 83] Step 1400 --> val loss 1.9964386224746704
 Step 1410: loss 1.8420665264129639
 Step 1420: loss 1.8120660781860352
 Step 1430: loss 1.7754243612289429
 Step 1440: loss 1.822453498840332
 Step 1450: loss 1.684941053390503
 Step 1460: loss 1.6878856420516968
 Step 1470: loss 1.8005969524383545
 Step 1480: loss 1.9323943853378296
 Step 1490: loss 1.855149745941162
 Step 1500: loss 1.7858505249023438
 Step 1510: loss 1.7165451049804688
 Step 1520: loss 1.7775866985321045
 Step 1530: loss 1.875501036643982
 Step 1540: loss 1.7744529247283936
 Step 1550: loss 1.692347764968872
 Step 1560: loss 1.8707884550094604
 Step 1570: loss 1.6772764921188354
 Step 1580: loss 1.7525599002838135
 Step 1590: loss 1.843366026878357
 Step 1600: loss 1.6903547048568726
 [Seed 83] Step 1600 --> val loss 1.9011366367340088
 Step 1610: loss 1.7782058715820312
 Step 1620: loss 1.7257325649261475
 Step 1630: loss 1.7710065841674805
 Step 1640: loss 1.8102582693099976
 Step 1650: loss 1.7337133884429932
 Step 1660: loss 1.7703313827514648
 Step 1670: loss 1.8597148656845093
 Step 1680: loss 1.7147033214569092
 Step 1690: loss 1.6532706022262573
 Step 1700: loss 1.6466279029846191
 Step 1710: loss 1.6639268398284912
 Step 1720: loss 1.8493030071258545
 Step 1730: loss 1.6027028560638428
 Step 1740: loss 1.5546821355819702
 Step 1750: loss 1.6561810970306396
 Step 1760: loss 1.7514294385910034
 Step 1770: loss 1.663142204284668
 Step 1780: loss 1.8283107280731201
 Step 1790: loss 1.687589406967163
 Step 1800: loss 1.684112548828125
 [Seed 83] Step 1800 --> val loss 1.8321555852890015
 Step 1810: loss 1.6122392416000366
 Step 1820: loss 1.678039789199829
 Step 1830: loss 1.6750752925872803
 Step 1840: loss 1.6617627143859863
 Step 1850: loss 1.6778967380523682
 Step 1860: loss 1.5885461568832397
 Step 1870: loss 1.6966474056243896
 Step 1880: loss 1.625704288482666
 Step 1890: loss 1.8062227964401245
 Step 1900: loss 1.6253236532211304
 Step 1910: loss 1.5350769758224487
 Step 1920: loss 1.6648811101913452
 Step 1930: loss 1.639531135559082
 Step 1940: loss 1.7656590938568115
 Step 1950: loss 1.6637485027313232
 Step 1960: loss 1.7316639423370361
 Step 1970: loss 1.6051925420761108
 Step 1980: loss 1.729274034500122
 Step 1990: loss 1.6165685653686523
 Step 2000: loss 1.7061750888824463
 [Seed 83] Step 2000 --> val loss 1.7689552307128906
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [46 43  1 40 43 58 58 43 56  6] ...
 First target sequence: [43  1 40 43 58 58 43 56  6  1] ...

 Decoded input: he better, given me by so holy a man.
 Hath yet the deputy sent m

 Decoded target: e better, given me by so holy a man.
 Hath yet the deputy sent my
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.222921371459961
 [Seed 84] Step 0 --> val loss 4.186336040496826
 Step 10: loss 3.8616576194763184
 Step 20: loss 3.313730239868164
 Step 30: loss 2.839474678039551
 Step 40: loss 2.7177348136901855
 Step 50: loss 2.502817153930664
 Step 60: loss 2.5075039863586426
 Step 70: loss 2.3709545135498047
 Step 80: loss 2.3402981758117676
 Step 90: loss 2.2560231685638428
 Step 100: loss 2.277042865753174
 Step 110: loss 2.2317912578582764
 Step 120: loss 2.1494522094726562
 Step 130: loss 2.1630311012268066
 Step 140: loss 2.2051987648010254
 Step 150: loss 2.149925470352173
 Step 160: loss 2.134883403778076
 Step 170: loss 2.156505823135376
 Step 180: loss 2.229250192642212
 Step 190: loss 2.05454158782959
 Step 200: loss 2.098776340484619
 [Seed 84] Step 200 --> val loss 2.373115301132202
 Step 210: loss 2.120854139328003
 Step 220: loss 2.0667028427124023
 Step 230: loss 2.1391115188598633
 Step 240: loss 2.191645622253418
 Step 250: loss 2.0154004096984863
 Step 260: loss 2.191377878189087
 Step 270: loss 1.9581022262573242
 Step 280: loss 2.1766164302825928
 Step 290: loss 2.0977094173431396
 Step 300: loss 2.152662992477417
 Step 310: loss 2.0088069438934326
 Step 320: loss 2.0764260292053223
 Step 330: loss 2.013127326965332
 Step 340: loss 2.0625529289245605
 Step 350: loss 1.847558856010437
 Step 360: loss 2.046807289123535
 Step 370: loss 2.039621353149414
 Step 380: loss 1.9582046270370483
 Step 390: loss 1.9931825399398804
 Step 400: loss 2.064150333404541
 [Seed 84] Step 400 --> val loss 2.3460428714752197
 Step 410: loss 2.02168607711792
 Step 420: loss 2.042240858078003
 Step 430: loss 2.0779566764831543
 Step 440: loss 2.125523090362549
 Step 450: loss 2.0371177196502686
 Step 460: loss 1.9858484268188477
 Step 470: loss 1.9646682739257812
 Step 480: loss 2.0757038593292236
 Step 490: loss 2.0158896446228027
 Step 500: loss 2.0311760902404785
 Step 510: loss 1.9380199909210205
 Step 520: loss 1.8963210582733154
 Step 530: loss 2.060718536376953
 Step 540: loss 1.971557855606079
 Step 550: loss 2.03163480758667
 Step 560: loss 2.006657600402832
 Step 570: loss 1.9340541362762451
 Step 580: loss 2.05452823638916
 Step 590: loss 2.099335193634033
 Step 600: loss 2.0687062740325928
 [Seed 84] Step 600 --> val loss 2.3367507457733154
 Step 610: loss 1.9383468627929688
 Step 620: loss 1.9954938888549805
 Step 630: loss 2.10614275932312
 Step 640: loss 2.0046255588531494
 Step 650: loss 2.0670509338378906
 Step 660: loss 2.0974931716918945
 Step 670: loss 2.039029598236084
 Step 680: loss 2.004055976867676
 Step 690: loss 2.0041134357452393
 Step 700: loss 1.9579932689666748
 Step 710: loss 2.133080005645752
 Step 720: loss 2.0961670875549316
 Step 730: loss 2.0522098541259766
 Step 740: loss 2.0217642784118652
 Step 750: loss 1.9520586729049683
 Step 760: loss 2.0268008708953857
 Step 770: loss 2.0626235008239746
 Step 780: loss 2.098705768585205
 Step 790: loss 1.9276341199874878
 Step 800: loss 2.0909667015075684
 [Seed 84] Step 800 --> val loss 2.242236852645874
 Step 810: loss 2.0590524673461914
 Step 820: loss 2.0169456005096436
 Step 830: loss 2.1087703704833984
 Step 840: loss 1.9767355918884277
 Step 850: loss 2.0995936393737793
 Step 860: loss 1.923233985900879
 Step 870: loss 1.9553842544555664
 Step 880: loss 2.0001916885375977
 Step 890: loss 2.016220808029175
 Step 900: loss 1.9482009410858154
 Step 910: loss 2.0960779190063477
 Step 920: loss 2.0029239654541016
 Step 930: loss 1.9773750305175781
 Step 940: loss 1.908625841140747
 Step 950: loss 1.9264037609100342
 Step 960: loss 1.9000606536865234
 Step 970: loss 1.9423093795776367
 Step 980: loss 1.9233611822128296
 Step 990: loss 1.9046424627304077
 Step 1000: loss 1.9908273220062256
 [Seed 84] Step 1000 --> val loss 2.198853015899658
 Step 1010: loss 2.0236737728118896
 Step 1020: loss 1.9734820127487183
 Step 1030: loss 1.949444055557251
 Step 1040: loss 2.076140880584717
 Step 1050: loss 2.1446456909179688
 Step 1060: loss 1.9742287397384644
 Step 1070: loss 1.9865907430648804
 Step 1080: loss 1.9698786735534668
 Step 1090: loss 2.066667079925537
 Step 1100: loss 2.0178468227386475
 Step 1110: loss 2.0141854286193848
 Step 1120: loss 2.0287888050079346
 Step 1130: loss 1.951659917831421
 Step 1140: loss 2.0786523818969727
 Step 1150: loss 2.0496883392333984
 Step 1160: loss 2.0142717361450195
 Step 1170: loss 1.8945391178131104
 Step 1180: loss 1.9566600322723389
 Step 1190: loss 1.8250240087509155
 Step 1200: loss 1.9487531185150146
 [Seed 84] Step 1200 --> val loss 2.1977553367614746
 Step 1210: loss 1.9040063619613647
 Step 1220: loss 1.9828437566757202
 Step 1230: loss 1.9183883666992188
 Step 1240: loss 1.9535640478134155
 Step 1250: loss 1.8666232824325562
 Step 1260: loss 1.9654353857040405
 Step 1270: loss 1.81184983253479
 Step 1280: loss 1.9211124181747437
 Step 1290: loss 1.9065358638763428
 Step 1300: loss 1.9330590963363647
 Step 1310: loss 1.872122049331665
 Step 1320: loss 2.069897413253784
 Step 1330: loss 1.8548390865325928
 Step 1340: loss 2.0399348735809326
 Step 1350: loss 1.8455886840820312
 Step 1360: loss 1.8593356609344482
 Step 1370: loss 1.7678303718566895
 Step 1380: loss 1.8703107833862305
 Step 1390: loss 1.7894628047943115
 Step 1400: loss 1.8961875438690186
 [Seed 84] Step 1400 --> val loss 2.078387498855591
 Step 1410: loss 1.9531633853912354
 Step 1420: loss 1.9448256492614746
 Step 1430: loss 1.868809700012207
 Step 1440: loss 1.974605917930603
 Step 1450: loss 1.9259707927703857
 Step 1460: loss 1.8616218566894531
 Step 1470: loss 1.815557837486267
 Step 1480: loss 1.693560004234314
 Step 1490: loss 1.8184607028961182
 Step 1500: loss 1.7604079246520996
 Step 1510: loss 1.8383302688598633
 Step 1520: loss 1.7339067459106445
 Step 1530: loss 1.8448504209518433
 Step 1540: loss 1.7468125820159912
 Step 1550: loss 1.8514665365219116
 Step 1560: loss 1.8157753944396973
 Step 1570: loss 1.8101470470428467
 Step 1580: loss 1.814260721206665
 Step 1590: loss 1.7534949779510498
 Step 1600: loss 1.8690297603607178
 [Seed 84] Step 1600 --> val loss 1.9778366088867188
 Step 1610: loss 1.749241828918457
 Step 1620: loss 1.858344554901123
 Step 1630: loss 1.7857797145843506
 Step 1640: loss 1.719590425491333
 Step 1650: loss 1.8757624626159668
 Step 1660: loss 1.7648980617523193
 Step 1670: loss 1.809877872467041
 Step 1680: loss 1.8691514730453491
 Step 1690: loss 1.9107158184051514
 Step 1700: loss 1.6959073543548584
 Step 1710: loss 1.7486995458602905
 Step 1720: loss 1.775376558303833
 Step 1730: loss 1.707879900932312
 Step 1740: loss 1.5909326076507568
 Step 1750: loss 1.7806333303451538
 Step 1760: loss 1.6752750873565674
 Step 1770: loss 1.8904006481170654
 Step 1780: loss 1.8028907775878906
 Step 1790: loss 1.705686330795288
 Step 1800: loss 1.775869369506836
 [Seed 84] Step 1800 --> val loss 1.9273185729980469
 Step 1810: loss 1.6976789236068726
 Step 1820: loss 1.727987289428711
 Step 1830: loss 1.7434133291244507
 Step 1840: loss 1.6369173526763916
 Step 1850: loss 1.6198251247406006
 Step 1860: loss 1.8352707624435425
 Step 1870: loss 1.6502487659454346
 Step 1880: loss 1.7712161540985107
 Step 1890: loss 1.7963861227035522
 Step 1900: loss 1.8284133672714233
 Step 1910: loss 1.643234133720398
 Step 1920: loss 1.67488431930542
 Step 1930: loss 1.8245396614074707
 Step 1940: loss 1.7055435180664062
 Step 1950: loss 1.6272152662277222
 Step 1960: loss 1.7614269256591797
 Step 1970: loss 1.6579084396362305
 Step 1980: loss 1.6587923765182495
 Step 1990: loss 1.720237135887146
 Step 2000: loss 1.6864389181137085
 [Seed 84] Step 2000 --> val loss 1.8707021474838257
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 6  1 46 47 57  1 39 54 54 39] ...
 First target sequence: [ 1 46 47 57  1 39 54 54 39 56] ...

 Decoded input: , his apparent open guilt omitted,
 I mean, his conversation with

 Decoded target:  his apparent open guilt omitted,
 I mean, his conversation with 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.215097427368164
 [Seed 85] Step 0 --> val loss 4.1848320960998535
 Step 10: loss 3.873277187347412
 Step 20: loss 3.343310832977295
 Step 30: loss 2.846310615539551
 Step 40: loss 2.8116493225097656
 Step 50: loss 2.6264896392822266
 Step 60: loss 2.4813249111175537
 Step 70: loss 2.4213199615478516
 Step 80: loss 2.3096163272857666
 Step 90: loss 2.429457902908325
 Step 100: loss 2.2967689037323
 Step 110: loss 2.2265689373016357
 Step 120: loss 2.212735176086426
 Step 130: loss 2.253760814666748
 Step 140: loss 2.225588083267212
 Step 150: loss 2.1933696269989014
 Step 160: loss 2.1209726333618164
 Step 170: loss 2.1831438541412354
 Step 180: loss 2.2539188861846924
 Step 190: loss 2.143113851547241
 Step 200: loss 2.0577754974365234
 [Seed 85] Step 200 --> val loss 2.3439574241638184
 Step 210: loss 2.0659873485565186
 Step 220: loss 2.0181078910827637
 Step 230: loss 1.9997785091400146
 Step 240: loss 2.096139907836914
 Step 250: loss 2.084864854812622
 Step 260: loss 2.020453929901123
 Step 270: loss 2.148324966430664
 Step 280: loss 2.0285520553588867
 Step 290: loss 2.122936725616455
 Step 300: loss 2.163303852081299
 Step 310: loss 2.084102153778076
 Step 320: loss 2.0016684532165527
 Step 330: loss 1.9207873344421387
 Step 340: loss 2.093111753463745
 Step 350: loss 2.1818318367004395
 Step 360: loss 2.126312017440796
 Step 370: loss 1.9833110570907593
 Step 380: loss 2.1107401847839355
 Step 390: loss 2.1742653846740723
 Step 400: loss 2.0569286346435547
 [Seed 85] Step 400 --> val loss 2.3264734745025635
 Step 410: loss 1.9909753799438477
 Step 420: loss 2.0015127658843994
 Step 430: loss 2.0850846767425537
 Step 440: loss 2.021728277206421
 Step 450: loss 1.9212589263916016
 Step 460: loss 2.0069761276245117
 Step 470: loss 2.0422072410583496
 Step 480: loss 2.0957751274108887
 Step 490: loss 2.030107259750366
 Step 500: loss 1.9299468994140625
 Step 510: loss 1.9580931663513184
 Step 520: loss 1.896472454071045
 Step 530: loss 1.9426921606063843
 Step 540: loss 2.0146920680999756
 Step 550: loss 2.0050787925720215
 Step 560: loss 1.8755958080291748
 Step 570: loss 2.03985333442688
 Step 580: loss 1.963439702987671
 Step 590: loss 1.9376068115234375
 Step 600: loss 1.9807822704315186
 [Seed 85] Step 600 --> val loss 2.241389513015747
 Step 610: loss 2.017089366912842
 Step 620: loss 1.9152482748031616
 Step 630: loss 1.9161689281463623
 Step 640: loss 1.9579766988754272
 Step 650: loss 1.9478873014450073
 Step 660: loss 2.0157060623168945
 Step 670: loss 2.0645923614501953
 Step 680: loss 1.9912031888961792
 Step 690: loss 1.9736799001693726
 Step 700: loss 1.9104652404785156
 Step 710: loss 1.9557456970214844
 Step 720: loss 2.096312999725342
 Step 730: loss 1.9346203804016113
 Step 740: loss 1.9396536350250244
 Step 750: loss 1.9684550762176514
 Step 760: loss 2.002856969833374
 Step 770: loss 2.132100820541382
 Step 780: loss 2.107124090194702
 Step 790: loss 1.9510091543197632
 Step 800: loss 2.1491246223449707
 [Seed 85] Step 800 --> val loss 2.3006083965301514
 Step 810: loss 2.227968454360962
 Step 820: loss 1.972337007522583
 Step 830: loss 2.1376843452453613
 Step 840: loss 2.043135166168213
 Step 850: loss 1.9069204330444336
 Step 860: loss 2.0640618801116943
 Step 870: loss 1.969130277633667
 Step 880: loss 1.8913596868515015
 Step 890: loss 2.1604485511779785
 Step 900: loss 2.0129199028015137
 Step 910: loss 2.0362155437469482
 Step 920: loss 1.8994916677474976
 Step 930: loss 2.051398992538452
 Step 940: loss 1.9599097967147827
 Step 950: loss 1.9066331386566162
 Step 960: loss 2.049823045730591
 Step 970: loss 1.9843626022338867
 Step 980: loss 1.9536556005477905
 Step 990: loss 1.9530138969421387
 Step 1000: loss 2.1054253578186035
 [Seed 85] Step 1000 --> val loss 2.3224916458129883
 Step 1010: loss 2.112253427505493
 Step 1020: loss 2.0012569427490234
 Step 1030: loss 1.8938636779785156
 Step 1040: loss 1.924926519393921
 Step 1050: loss 2.0303657054901123
 Step 1060: loss 2.0725321769714355
 Step 1070: loss 2.1094956398010254
 Step 1080: loss 1.9350141286849976
 Step 1090: loss 2.0244879722595215
 Step 1100: loss 1.9761922359466553
 Step 1110: loss 1.9969356060028076
 Step 1120: loss 1.9535462856292725
 Step 1130: loss 1.976876139640808
 Step 1140: loss 2.0071849822998047
 Step 1150: loss 2.0124435424804688
 Step 1160: loss 1.9690297842025757
 Step 1170: loss 2.0211563110351562
 Step 1180: loss 1.948956847190857
 Step 1190: loss 1.845879316329956
 Step 1200: loss 2.0707297325134277
 [Seed 85] Step 1200 --> val loss 2.204148530960083
 Step 1210: loss 1.9666955471038818
 Step 1220: loss 1.9994500875473022
 Step 1230: loss 1.9432218074798584
 Step 1240: loss 1.9868391752243042
 Step 1250: loss 2.033888816833496
 Step 1260: loss 1.9701223373413086
 Step 1270: loss 1.9574000835418701
 Step 1280: loss 1.9787006378173828
 Step 1290: loss 1.85528564453125
 Step 1300: loss 1.9739468097686768
 Step 1310: loss 1.9473216533660889
 Step 1320: loss 1.827328085899353
 Step 1330: loss 1.9212220907211304
 Step 1340: loss 1.951887607574463
 Step 1350: loss 1.9347333908081055
 Step 1360: loss 2.007725715637207
 Step 1370: loss 2.0509583950042725
 Step 1380: loss 1.9935967922210693
 Step 1390: loss 1.8837401866912842
 Step 1400: loss 2.010354995727539
 [Seed 85] Step 1400 --> val loss 2.182924747467041
 Step 1410: loss 1.993290901184082
 Step 1420: loss 1.962686538696289
 Step 1430: loss 1.8741134405136108
 Step 1440: loss 1.8515214920043945
 Step 1450: loss 1.9337409734725952
 Step 1460: loss 1.8577992916107178
 Step 1470: loss 1.925244688987732
 Step 1480: loss 1.8655250072479248
 Step 1490: loss 1.9535884857177734
 Step 1500: loss 1.9189561605453491
 Step 1510: loss 2.0593342781066895
 Step 1520: loss 1.7791459560394287
 Step 1530: loss 1.8926008939743042
 Step 1540: loss 1.891722559928894
 Step 1550: loss 1.969611644744873
 Step 1560: loss 1.841705322265625
 Step 1570: loss 1.894517421722412
 Step 1580: loss 1.8815773725509644
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 85] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 85] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 85] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [63 57 10  0 27  6  1 40 63  1] ...
 First target sequence: [57 10  0 27  6  1 40 63  1 58] ...

 Decoded input: ys:
 O, by this count I shall be much in years
 Ere I again behold

 Decoded target: s:
 O, by this count I shall be much in years
 Ere I again behold 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.197160243988037
 [Seed 86] Step 0 --> val loss 4.181693077087402
 Step 10: loss 3.850271224975586
 Step 20: loss 3.3061203956604004
 Step 30: loss 2.899120330810547
 Step 40: loss 2.68727970123291
 Step 50: loss 2.633476734161377
 Step 60: loss 2.348590612411499
 Step 70: loss 2.4132699966430664
 Step 80: loss 2.3758981227874756
 Step 90: loss 2.274954080581665
 Step 100: loss 2.315551519393921
 Step 110: loss 2.278630018234253
 Step 120: loss 2.235032081604004
 Step 130: loss 2.284672737121582
 Step 140: loss 2.162376880645752
 Step 150: loss 2.1557421684265137
 Step 160: loss 2.236520528793335
 Step 170: loss 2.1049599647521973
 Step 180: loss 2.0893707275390625
 Step 190: loss 2.0492849349975586
 Step 200: loss 2.1332497596740723
 [Seed 86] Step 200 --> val loss 2.2816553115844727
 Step 210: loss 2.0335025787353516
 Step 220: loss 2.065248489379883
 Step 230: loss 2.0079221725463867
 Step 240: loss 2.037874937057495
 Step 250: loss 2.0656068325042725
 Step 260: loss 1.9818238019943237
 Step 270: loss 2.0290770530700684
 Step 280: loss 1.9834508895874023
 Step 290: loss 2.148665189743042
 Step 300: loss 2.15592622756958
 Step 310: loss 2.0097098350524902
 Step 320: loss 2.0200247764587402
 Step 330: loss 1.9464973211288452
 Step 340: loss 1.9777244329452515
 Step 350: loss 2.0153045654296875
 Step 360: loss 1.9240678548812866
 Step 370: loss 1.9175677299499512
 Step 380: loss 1.9038922786712646
 Step 390: loss 1.9255447387695312
 Step 400: loss 2.0838701725006104
 [Seed 86] Step 400 --> val loss 2.2187554836273193
 Step 410: loss 1.9607200622558594
 Step 420: loss 1.9925334453582764
 Step 430: loss 2.1245169639587402
 Step 440: loss 2.016662359237671
 Step 450: loss 1.986652135848999
 Step 460: loss 1.9683717489242554
 Step 470: loss 2.041452407836914
 Step 480: loss 2.1471657752990723
 Step 490: loss 1.9613107442855835
 Step 500: loss 1.8652013540267944
 Step 510: loss 1.9819492101669312
 Step 520: loss 1.9965600967407227
 Step 530: loss 2.0253677368164062
 Step 540: loss 2.0478017330169678
 Step 550: loss 1.972249984741211
 Step 560: loss 1.9680761098861694
 Step 570: loss 1.9673824310302734
 Step 580: loss 2.044276714324951
 Step 590: loss 2.1823935508728027
 Step 600: loss 2.119778871536255
 [Seed 86] Step 600 --> val loss 2.2954728603363037
 Step 610: loss 2.025942802429199
 Step 620: loss 2.0138587951660156
 Step 630: loss 1.9762872457504272
 Step 640: loss 1.897199034690857
 Step 650: loss 1.9506549835205078
 Step 660: loss 2.0383574962615967
 Step 670: loss 1.9394052028656006
 Step 680: loss 2.0884721279144287
 Step 690: loss 2.242684841156006
 Step 700: loss 1.954502820968628
 Step 710: loss 2.002964973449707
 Step 720: loss 1.8899940252304077
 Step 730: loss 1.9116849899291992
 Step 740: loss 2.077085494995117
 Step 750: loss 2.017775297164917
 Step 760: loss 1.946812391281128
 Step 770: loss 1.99967360496521
 Step 780: loss 2.040997266769409
 Step 790: loss 1.8937921524047852
 Step 800: loss 1.8786786794662476
 [Seed 86] Step 800 --> val loss 2.1549956798553467
 Step 810: loss 2.0020291805267334
 Step 820: loss 2.0056395530700684
 Step 830: loss 1.9382250308990479
 Step 840: loss 1.9313796758651733
 Step 850: loss 1.9480032920837402
 Step 860: loss 1.984326720237732
 Step 870: loss 1.9475016593933105
 Step 880: loss 1.926500678062439
 Step 890: loss 1.9720478057861328
 Step 900: loss 1.963460922241211
 Step 910: loss 1.8934653997421265
 Step 920: loss 1.9881776571273804
 Step 930: loss 2.019371747970581
 Step 940: loss 1.9721893072128296
 Step 950: loss 1.9412250518798828
 Step 960: loss 2.050654411315918
 Step 970: loss 1.9930477142333984
 Step 980: loss 1.9177626371383667
 Step 990: loss 2.061819553375244
 Step 1000: loss 1.9616597890853882
 [Seed 86] Step 1000 --> val loss 2.183758497238159
 Step 1010: loss 1.9292798042297363
 Step 1020: loss 2.0125885009765625
 Step 1030: loss 2.107386827468872
 Step 1040: loss 1.9549291133880615
 Step 1050: loss 2.050865888595581
 Step 1060: loss 1.9517145156860352
 Step 1070: loss 1.9790728092193604
 Step 1080: loss 1.888514518737793
 Step 1090: loss 1.9409122467041016
 Step 1100: loss 1.9521682262420654
 Step 1110: loss 1.944112777709961
 Step 1120: loss 2.055830955505371
 Step 1130: loss 1.924150824546814
 Step 1140: loss 2.0612666606903076
 Step 1150: loss 1.9377362728118896
 Step 1160: loss 2.006378173828125
 Step 1170: loss 1.9951715469360352
 Step 1180: loss 1.9953107833862305
 Step 1190: loss 2.0106606483459473
 Step 1200: loss 2.144369125366211
 [Seed 86] Step 1200 --> val loss 2.117821216583252
 Step 1210: loss 2.0406341552734375
 Step 1220: loss 1.8947910070419312
 Step 1230: loss 1.9404611587524414
 Step 1240: loss 1.910711646080017
 Step 1250: loss 1.9466049671173096
 Step 1260: loss 1.734309196472168
 Step 1270: loss 2.013662815093994
 Step 1280: loss 2.1153485774993896
 Step 1290: loss 1.9579112529754639
 Step 1300: loss 1.8360642194747925
 Step 1310: loss 1.8931068181991577
 Step 1320: loss 2.0327341556549072
 Step 1330: loss 1.9369314908981323
 Step 1340: loss 2.011476516723633
 Step 1350: loss 1.971107840538025
 Step 1360: loss 1.98577880859375
 Step 1370: loss 1.9835948944091797
 Step 1380: loss 1.9524242877960205
 Step 1390: loss 2.0995802879333496
 Step 1400: loss 2.0754988193511963
 [Seed 86] Step 1400 --> val loss 2.132040023803711
 Step 1410: loss 1.9356520175933838
 Step 1420: loss 1.9094833135604858
 Step 1430: loss 1.983396291732788
 Step 1440: loss 1.9563201665878296
 Step 1450: loss 1.90404212474823
 Step 1460: loss 1.9052690267562866
 Step 1470: loss 1.8725332021713257
 Step 1480: loss 2.006953716278076
 Step 1490: loss 1.9914964437484741
 Step 1500: loss 1.9166721105575562
 Step 1510: loss 1.998986005783081
 Step 1520: loss 1.9216140508651733
 Step 1530: loss 1.8285526037216187
 Step 1540: loss 1.9401001930236816
 Step 1550: loss 1.8703808784484863
 Step 1560: loss 1.9701118469238281
 Step 1570: loss 1.908215045928955
 Step 1580: loss 1.9221972227096558
 Step 1590: loss 1.8901584148406982
 Step 1600: loss 1.8104383945465088
 [Seed 86] Step 1600 --> val loss 2.0858120918273926
 Step 1610: loss 1.8763859272003174
 Step 1620: loss 1.8873767852783203
 Step 1630: loss 1.9475314617156982
 Step 1640: loss 1.795419692993164
 Step 1650: loss 1.9035520553588867
 Step 1660: loss 1.905383825302124
 Step 1670: loss 1.8379955291748047
 Step 1680: loss 1.8001351356506348
 Step 1690: loss 1.9313462972640991
 Step 1700: loss 1.9290885925292969
 Step 1710: loss 1.8239452838897705
 Step 1720: loss 1.802322268486023
 Step 1730: loss 1.8459129333496094
 Step 1740: loss 1.9364408254623413
 Step 1750: loss 1.7626023292541504
 Step 1760: loss 1.9043081998825073
 Step 1770: loss 1.8142423629760742
 Step 1780: loss 1.8016698360443115
 Step 1790: loss 1.8307689428329468
 Step 1800: loss 1.7333078384399414
 [Seed 86] Step 1800 --> val loss 1.9909517765045166
 Step 1810: loss 1.8050754070281982
 Step 1820: loss 1.9276123046875
 Step 1830: loss 1.8974961042404175
 Step 1840: loss 1.8864673376083374
 Step 1850: loss 1.7929266691207886
 Step 1860: loss 1.8633458614349365
 Step 1870: loss 1.7367064952850342
 Step 1880: loss 1.8392837047576904
 Step 1890: loss 1.7683601379394531
 Step 1900: loss 1.795369029045105
 Step 1910: loss 1.7997283935546875
 Step 1920: loss 1.6847562789916992
 Step 1930: loss 1.8254674673080444
 Step 1940: loss 1.7403234243392944
 Step 1950: loss 1.6862988471984863
 Step 1960: loss 1.887657880783081
 Step 1970: loss 1.86590576171875
 Step 1980: loss 1.8472964763641357
 Step 1990: loss 1.8132762908935547
 Step 2000: loss 1.8244400024414062
 [Seed 86] Step 2000 --> val loss 1.9712644815444946
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [39 56 39 50 50 43 50  5 42  0] ...
 First target sequence: [56 39 50 50 43 50  5 42  0 17] ...

 Decoded input: arallel'd
 Even with the stroke and line of his great justice:
 He

 Decoded target: rallel'd
 Even with the stroke and line of his great justice:
 He 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.207216262817383
 [Seed 87] Step 0 --> val loss 4.183108329772949
 Step 10: loss 3.8511149883270264
 Step 20: loss 3.2748916149139404
 Step 30: loss 2.956115245819092
 Step 40: loss 2.637559413909912
 Step 50: loss 2.5753114223480225
 Step 60: loss 2.5493783950805664
 Step 70: loss 2.497476100921631
 Step 80: loss 2.330838680267334
 Step 90: loss 2.311826229095459
 Step 100: loss 2.3629612922668457
 Step 110: loss 2.216481924057007
 Step 120: loss 2.294459819793701
 Step 130: loss 2.06174898147583
 Step 140: loss 2.102069854736328
 Step 150: loss 2.2004213333129883
 Step 160: loss 2.2205886840820312
 Step 170: loss 2.1021180152893066
 Step 180: loss 2.20383620262146
 Step 190: loss 2.032770872116089
 Step 200: loss 2.12917423248291
 [Seed 87] Step 200 --> val loss 2.2983813285827637
 Step 210: loss 2.0799126625061035
 Step 220: loss 2.0810678005218506
 Step 230: loss 2.0426063537597656
 Step 240: loss 2.1416707038879395
 Step 250: loss 2.0871477127075195
 Step 260: loss 2.0162997245788574
 Step 270: loss 2.068253517150879
 Step 280: loss 2.096172332763672
 Step 290: loss 2.1815402507781982
 Step 300: loss 1.9981963634490967
 Step 310: loss 2.016326427459717
 Step 320: loss 2.004000663757324
 Step 330: loss 2.0563039779663086
 Step 340: loss 2.0072591304779053
 Step 350: loss 2.113765239715576
 Step 360: loss 1.9786174297332764
 Step 370: loss 2.0685582160949707
 Step 380: loss 2.1310434341430664
 Step 390: loss 2.020777940750122
 Step 400: loss 1.9066256284713745
 [Seed 87] Step 400 --> val loss 2.1914851665496826
 Step 410: loss 1.9377892017364502
 Step 420: loss 1.9318469762802124
 Step 430: loss 1.9484565258026123
 Step 440: loss 1.9947292804718018
 Step 450: loss 1.892035722732544
 Step 460: loss 2.0191361904144287
 Step 470: loss 1.9521074295043945
 Step 480: loss 1.9887754917144775
 Step 490: loss 1.896744966506958
 Step 500: loss 2.0301036834716797
 Step 510: loss 1.9922665357589722
 Step 520: loss 2.0120503902435303
 Step 530: loss 2.024919033050537
 Step 540: loss 2.017918109893799
 Step 550: loss 1.8898251056671143
 Step 560: loss 2.0600931644439697
 Step 570: loss 1.975248098373413
 Step 580: loss 2.0122406482696533
 Step 590: loss 1.8708322048187256
 Step 600: loss 2.0325427055358887
 [Seed 87] Step 600 --> val loss 2.229217290878296
 Step 610: loss 2.012542963027954
 Step 620: loss 2.0796356201171875
 Step 630: loss 2.0527923107147217
 Step 640: loss 2.0162911415100098
 Step 650: loss 2.0243782997131348
 Step 660: loss 2.011103630065918
 Step 670: loss 1.8178520202636719
 Step 680: loss 2.0235066413879395
 Step 690: loss 1.986128568649292
 Step 700: loss 1.990386724472046
 Step 710: loss 1.9432220458984375
 Step 720: loss 2.021782875061035
 Step 730: loss 1.9607949256896973
 Step 740: loss 1.9471073150634766
 Step 750: loss 2.0059289932250977
 Step 760: loss 1.906003713607788
 Step 770: loss 1.9267711639404297
 Step 780: loss 1.9064897298812866
 Step 790: loss 1.8708839416503906
 Step 800: loss 1.922723650932312
 [Seed 87] Step 800 --> val loss 2.0969836711883545
 Step 810: loss 2.0750272274017334
 Step 820: loss 1.9744991064071655
 Step 830: loss 1.9364451169967651
 Step 840: loss 1.9635270833969116
 Step 850: loss 2.016246795654297
 Step 860: loss 1.9141215085983276
 Step 870: loss 2.0905818939208984
 Step 880: loss 2.00799822807312
 Step 890: loss 2.133235216140747
 Step 900: loss 1.9907236099243164
 Step 910: loss 1.980513334274292
 Step 920: loss 2.034186601638794
 Step 930: loss 2.1198105812072754
 Step 940: loss 2.0682473182678223
 Step 950: loss 1.9479901790618896
 Step 960: loss 2.0820183753967285
 Step 970: loss 2.1181750297546387
 Step 980: loss 2.0000433921813965
 Step 990: loss 2.185967206954956
 Step 1000: loss 1.9183725118637085
 [Seed 87] Step 1000 --> val loss 2.243608236312866
 Step 1010: loss 2.1226558685302734
 Step 1020: loss 2.0050716400146484
 Step 1030: loss 2.0243115425109863
 Step 1040: loss 1.7943742275238037
 Step 1050: loss 2.0439701080322266
 Step 1060: loss 2.0870566368103027
 Step 1070: loss 1.920511245727539
 Step 1080: loss 2.0702524185180664
 Step 1090: loss 2.0255653858184814
 Step 1100: loss 2.0183303356170654
 Step 1110: loss 1.9046519994735718
 Step 1120: loss 1.8453540802001953
 Step 1130: loss 1.8422825336456299
 Step 1140: loss 1.9666255712509155
 Step 1150: loss 1.9558748006820679
 Step 1160: loss 1.8168491125106812
 Step 1170: loss 1.899139165878296
 Step 1180: loss 1.9427299499511719
 Step 1190: loss 1.9492524862289429
 Step 1200: loss 2.067591905593872
 [Seed 87] Step 1200 --> val loss 2.1323673725128174
 Step 1210: loss 1.8872740268707275
 Step 1220: loss 1.9706166982650757
 Step 1230: loss 2.008147954940796
 Step 1240: loss 1.9280498027801514
 Step 1250: loss 1.898777723312378
 Step 1260: loss 1.9690908193588257
 Step 1270: loss 2.0295591354370117
 Step 1280: loss 1.9643429517745972
 Step 1290: loss 1.9436275959014893
 Step 1300: loss 1.9365830421447754
 Step 1310: loss 1.871910810470581
 Step 1320: loss 1.9304497241973877
 Step 1330: loss 2.0385594367980957
 Step 1340: loss 1.9112911224365234
 Step 1350: loss 1.8591750860214233
 Step 1360: loss 1.8652223348617554
 Step 1370: loss 1.877262830734253
 Step 1380: loss 1.8821375370025635
 Step 1390: loss 1.990519404411316
 Step 1400: loss 1.8649773597717285
 [Seed 87] Step 1400 --> val loss 2.0398929119110107
 Step 1410: loss 1.9328221082687378
 Step 1420: loss 1.8582638502120972
 Step 1430: loss 1.9222577810287476
 Step 1440: loss 1.8657760620117188
 Step 1450: loss 1.989035964012146
 Step 1460: loss 1.9461216926574707
 Step 1470: loss 1.9939807653427124
 Step 1480: loss 2.044053077697754
 Step 1490: loss 1.912343978881836
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 87] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 87] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 87] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 1 57 53  1 42 47 43 12  0  0] ...
 First target sequence: [57 53  1 42 47 43 12  0  0 16] ...

 Decoded input:  so die?

 DUKE OF AUMERLE:
 No, good my lord; let's fight with ge

 Decoded target: so die?

 DUKE OF AUMERLE:
 No, good my lord; let's fight with gen
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.2090301513671875
 [Seed 88] Step 0 --> val loss 4.184083938598633
 Step 10: loss 3.848057508468628
 Step 20: loss 3.3302388191223145
 Step 30: loss 2.9449172019958496
 Step 40: loss 2.656895160675049
 Step 50: loss 2.58539080619812
 Step 60: loss 2.555633068084717
 Step 70: loss 2.4393625259399414
 Step 80: loss 2.3947322368621826
 Step 90: loss 2.351425886154175
 Step 100: loss 2.265377998352051
 Step 110: loss 2.2923145294189453
 Step 120: loss 2.2050747871398926
 Step 130: loss 2.2476067543029785
 Step 140: loss 2.1628801822662354
 Step 150: loss 2.238816738128662
 Step 160: loss 2.1839282512664795
 Step 170: loss 2.098771572113037
 Step 180: loss 2.136854887008667
 Step 190: loss 2.125054121017456
 Step 200: loss 2.119412899017334
 [Seed 88] Step 200 --> val loss 2.340258836746216
 Step 210: loss 2.217817783355713
 Step 220: loss 2.0154836177825928
 Step 230: loss 2.1335244178771973
 Step 240: loss 2.1711230278015137
 Step 250: loss 2.094048023223877
 Step 260: loss 2.0923914909362793
 Step 270: loss 2.0806117057800293
 Step 280: loss 2.048361301422119
 Step 290: loss 2.0267984867095947
 Step 300: loss 1.9977166652679443
 Step 310: loss 2.0504043102264404
 Step 320: loss 2.038407325744629
 Step 330: loss 2.052060604095459
 Step 340: loss 2.1494832038879395
 Step 350: loss 2.0222673416137695
 Step 360: loss 1.9816219806671143
 Step 370: loss 2.049288511276245
 Step 380: loss 1.9901142120361328
 Step 390: loss 2.0301852226257324
 Step 400: loss 1.9304678440093994
 [Seed 88] Step 400 --> val loss 2.2464733123779297
 Step 410: loss 2.0243349075317383
 Step 420: loss 1.9696824550628662
 Step 430: loss 1.9295036792755127
 Step 440: loss 1.9696038961410522
 Step 450: loss 1.9541505575180054
 Step 460: loss 1.9888417720794678
 Step 470: loss 1.8698420524597168
 Step 480: loss 1.9139560461044312
 Step 490: loss 1.9763354063034058
 Step 500: loss 1.9818428754806519
 Step 510: loss 1.9448671340942383
 Step 520: loss 2.067558765411377
 Step 530: loss 1.9730838537216187
 Step 540: loss 1.9548434019088745
 Step 550: loss 2.126418113708496
 Step 560: loss 1.9354490041732788
 Step 570: loss 1.9169915914535522
 Step 580: loss 1.8866785764694214
 Step 590: loss 1.9949312210083008
 Step 600: loss 2.0022497177124023
 [Seed 88] Step 600 --> val loss 2.234771251678467
 Step 610: loss 2.0314950942993164
 Step 620: loss 2.0717084407806396
 Step 630: loss 1.9764277935028076
 Step 640: loss 2.0937933921813965
 Step 650: loss 1.9235334396362305
 Step 660: loss 2.0365991592407227
 Step 670: loss 1.9792163372039795
 Step 680: loss 1.9064722061157227
 Step 690: loss 1.9694031476974487
 Step 700: loss 1.9227778911590576
 Step 710: loss 1.9352729320526123
 Step 720: loss 1.8518779277801514
 Step 730: loss 1.971766471862793
 Step 740: loss 1.971734881401062
 Step 750: loss 1.9466222524642944
 Step 760: loss 1.9635696411132812
 Step 770: loss 1.992537260055542
 Step 780: loss 2.02423095703125
 Step 790: loss 1.857985496520996
 Step 800: loss 1.9530978202819824
 [Seed 88] Step 800 --> val loss 2.1337015628814697
 Step 810: loss 1.9012402296066284
 Step 820: loss 1.9195246696472168
 Step 830: loss 1.874420404434204
 Step 840: loss 1.9244968891143799
 Step 850: loss 1.9383158683776855
 Step 860: loss 1.8402042388916016
 Step 870: loss 1.8651494979858398
 Step 880: loss 1.9640212059020996
 Step 890: loss 1.900329351425171
 Step 900: loss 1.9487680196762085
 Step 910: loss 1.8483951091766357
 Step 920: loss 1.9636693000793457
 Step 930: loss 1.927926778793335
 Step 940: loss 1.9288501739501953
 Step 950: loss 1.9229930639266968
 Step 960: loss 1.808429479598999
 Step 970: loss 2.024143695831299
 Step 980: loss 1.8245735168457031
 Step 990: loss 1.8492586612701416
 Step 1000: loss 1.9834264516830444
 [Seed 88] Step 1000 --> val loss 2.1983706951141357
 Step 1010: loss 1.8763211965560913
 Step 1020: loss 1.897143840789795
 Step 1030: loss 1.8802907466888428
 Step 1040: loss 1.9193100929260254
 Step 1050: loss 1.7989282608032227
 Step 1060: loss 1.8307454586029053
 Step 1070: loss 1.939956545829773
 Step 1080: loss 1.8519198894500732
 Step 1090: loss 1.9016460180282593
 Step 1100: loss 1.7820394039154053
 Step 1110: loss 1.8703546524047852
 Step 1120: loss 1.7767175436019897
 Step 1130: loss 1.7974278926849365
 Step 1140: loss 1.8101061582565308
 Step 1150: loss 1.8812564611434937
 Step 1160: loss 1.8187432289123535
 Step 1170: loss 1.916034460067749
 Step 1180: loss 1.9108247756958008
 Step 1190: loss 1.9053319692611694
 Step 1200: loss 1.8829829692840576
 [Seed 88] Step 1200 --> val loss 2.1651504039764404
 Step 1210: loss 1.7746338844299316
 Step 1220: loss 1.8941539525985718
 Step 1230: loss 1.868139624595642
 Step 1240: loss 1.8856110572814941
 Step 1250: loss 1.856170892715454
 Step 1260: loss 1.8451474905014038
 Step 1270: loss 1.8783921003341675
 Step 1280: loss 1.8764817714691162
 Step 1290: loss 1.774672508239746
 Step 1300: loss 1.7166999578475952
 Step 1310: loss 1.8171346187591553
 Step 1320: loss 1.812713623046875
 Step 1330: loss 1.9234298467636108
 Step 1340: loss 1.6904176473617554
 Step 1350: loss 1.7951663732528687
 Step 1360: loss 1.8208082914352417
 Step 1370: loss 1.8038976192474365
 Step 1380: loss 1.7718472480773926
 Step 1390: loss 1.7831289768218994
 Step 1400: loss 1.7512239217758179
 [Seed 88] Step 1400 --> val loss 2.046370267868042
 Step 1410: loss 1.759522795677185
 Step 1420: loss 1.7808372974395752
 Step 1430: loss 1.778296947479248
 Step 1440: loss 1.8424224853515625
 Step 1450: loss 1.8008941411972046
 Step 1460: loss 1.7571886777877808
 Step 1470: loss 1.8545165061950684
 Step 1480: loss 1.8376054763793945
 Step 1490: loss 1.678998351097107
 Step 1500: loss 1.6617279052734375
 Step 1510: loss 1.7183616161346436
 Step 1520: loss 1.7295417785644531
 Step 1530: loss 1.815209150314331
 Step 1540: loss 1.719238519668579
 Step 1550: loss 1.6502196788787842
 Step 1560: loss 1.7266368865966797
 Step 1570: loss 1.7207467555999756
 Step 1580: loss 1.7367477416992188
 Step 1590: loss 1.8818855285644531
 Step 1600: loss 1.7384179830551147
 [Seed 88] Step 1600 --> val loss 1.8939847946166992
 Step 1610: loss 1.7310097217559814
 Step 1620: loss 1.6403697729110718
 Step 1630: loss 1.8145112991333008
 Step 1640: loss 1.6859703063964844
 Step 1650: loss 1.7122701406478882
 Step 1660: loss 1.7691318988800049
 Step 1670: loss 1.6302284002304077
 Step 1680: loss 1.7540572881698608
 Step 1690: loss 1.6421455144882202
 Step 1700: loss 1.681396484375
 Step 1710: loss 1.719437599182129
 Step 1720: loss 1.696944236755371
 Step 1730: loss 1.640213966369629
 Step 1740: loss 1.637158751487732
 Step 1750: loss 1.5882813930511475
 Step 1760: loss 1.6747984886169434
 Step 1770: loss 1.7182821035385132
 Step 1780: loss 1.7358965873718262
 Step 1790: loss 1.6477140188217163
 Step 1800: loss 1.6553735733032227
 [Seed 88] Step 1800 --> val loss 1.8059180974960327
 Step 1810: loss 1.6867311000823975
 Step 1820: loss 1.6904916763305664
 Step 1830: loss 1.6878199577331543
 Step 1840: loss 1.6777236461639404
 Step 1850: loss 1.682607889175415
 Step 1860: loss 1.6738351583480835
 Step 1870: loss 1.6840476989746094
 Step 1880: loss 1.5684503316879272
 Step 1890: loss 1.6806421279907227
 Step 1900: loss 1.6322640180587769
 Step 1910: loss 1.6533689498901367
 Step 1920: loss 1.6980918645858765
 Step 1930: loss 1.706102728843689
 Step 1940: loss 1.6300424337387085
 Step 1950: loss 1.5757675170898438
 Step 1960: loss 1.6110776662826538
 Step 1970: loss 1.690338134765625
 Step 1980: loss 1.6478936672210693
 Step 1990: loss 1.7504074573516846
 Step 2000: loss 1.6642041206359863
 [Seed 88] Step 2000 --> val loss 1.751327395439148
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [46 43  1 56 43 57 58  1 61 47] ...
 First target sequence: [43  1 56 43 57 58  1 61 47 50] ...

 Decoded input: he rest will serve
 For a short holding: if we lose the field,
 We

 Decoded target: e rest will serve
 For a short holding: if we lose the field,
 We 
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.206527233123779
 [Seed 89] Step 0 --> val loss 4.182226181030273
 Step 10: loss 3.8556430339813232
 Step 20: loss 3.3238344192504883
 Step 30: loss 2.8429322242736816
 Step 40: loss 2.677852153778076
 Step 50: loss 2.6033663749694824
 Step 60: loss 2.4897711277008057
 Step 70: loss 2.490168571472168
 Step 80: loss 2.452519416809082
 Step 90: loss 2.319241523742676
 Step 100: loss 2.3526241779327393
 Step 110: loss 2.1748147010803223
 Step 120: loss 2.1864242553710938
 Step 130: loss 2.1789913177490234
 Step 140: loss 2.206925392150879
 Step 150: loss 2.262233257293701
 Step 160: loss 2.0416245460510254
 Step 170: loss 2.1477015018463135
 Step 180: loss 2.1708006858825684
 Step 190: loss 2.03425669670105
 Step 200: loss 2.0937671661376953
 [Seed 89] Step 200 --> val loss 2.2990691661834717
 Step 210: loss 2.180321455001831
 Step 220: loss 2.1795263290405273
 Step 230: loss 2.1152474880218506
 Step 240: loss 2.104125738143921
 Step 250: loss 2.0211987495422363
 Step 260: loss 2.027653217315674
 Step 270: loss 2.071876287460327
 Step 280: loss 2.1147055625915527
 Step 290: loss 2.0314111709594727
 Step 300: loss 2.0087671279907227
 Step 310: loss 2.058032751083374
 Step 320: loss 2.0125927925109863
 Step 330: loss 1.9971895217895508
 Step 340: loss 2.006599187850952
 Step 350: loss 1.9807538986206055
 Step 360: loss 2.057384729385376
 Step 370: loss 2.178737163543701
 Step 380: loss 1.9899020195007324
 Step 390: loss 2.004908561706543
 Step 400: loss 2.0291907787323
 [Seed 89] Step 400 --> val loss 2.2983169555664062
 Step 410: loss 2.004903554916382
 Step 420: loss 1.999109148979187
 Step 430: loss 2.030482053756714
 Step 440: loss 2.0894360542297363
 Step 450: loss 2.059593915939331
 Step 460: loss 2.1310532093048096
 Step 470: loss 2.068113327026367
 Step 480: loss 2.0413975715637207
 Step 490: loss 2.026577949523926
 Step 500: loss 1.9633296728134155
 Step 510: loss 1.957122802734375
 Step 520: loss 2.049685001373291
 Step 530: loss 2.0179061889648438
 Step 540: loss 2.0088815689086914
 Step 550: loss 2.0062663555145264
 Step 560: loss 1.9266809225082397
 Step 570: loss 2.1180291175842285
 Step 580: loss 1.9765623807907104
 Step 590: loss 2.0193824768066406
 Step 600: loss 1.9941208362579346
 [Seed 89] Step 600 --> val loss 2.2823855876922607
 Step 610: loss 1.9374854564666748
 Step 620: loss 1.950989007949829
 Step 630: loss 1.9649200439453125
 Step 640: loss 2.0311458110809326
 Step 650: loss 2.081217050552368
 Step 660: loss 1.8774055242538452
 Step 670: loss 2.057459831237793
 Step 680: loss 2.175260066986084
 Step 690: loss 2.0138092041015625
 Step 700: loss 2.027848720550537
 Step 710: loss 1.9559396505355835
 Step 720: loss 1.9719758033752441
 Step 730: loss 1.9606971740722656
 Step 740: loss 1.9221391677856445
 Step 750: loss 2.0806257724761963
 Step 760: loss 1.9600496292114258
 Step 770: loss 1.992767095565796
 Step 780: loss 2.0567708015441895
 Step 790: loss 2.022735595703125
 Step 800: loss 1.9391148090362549
 [Seed 89] Step 800 --> val loss 2.1600308418273926
 Step 810: loss 1.9069241285324097
 Step 820: loss 2.1311471462249756
 Step 830: loss 1.9658139944076538
 Step 840: loss 2.200305223464966
 Step 850: loss 1.963911533355713
 Step 860: loss 2.022883415222168
 Step 870: loss 1.9702908992767334
 Step 880: loss 1.9446643590927124
 Step 890: loss 1.9644778966903687
 Step 900: loss 1.8568665981292725
 Step 910: loss 2.1099154949188232
 Step 920: loss 2.0215413570404053
 Step 930: loss 2.055994987487793
 Step 940: loss 1.903571605682373
 Step 950: loss 2.0188403129577637
 Step 960: loss 1.998327612876892
 Step 970: loss 2.0960166454315186
 Step 980: loss 1.9634740352630615
 Step 990: loss 1.9293826818466187
 Step 1000: loss 2.0410971641540527
 [Seed 89] Step 1000 --> val loss 2.35439395904541
 Step 1010: loss 2.055784225463867
 Step 1020: loss 2.112036943435669
 Step 1030: loss 2.051079511642456
 Step 1040: loss 1.9734301567077637
 Step 1050: loss 2.0374228954315186
 Step 1060: loss 2.196659803390503
 Step 1070: loss 2.042820930480957
 Step 1080: loss 1.9329955577850342
 Step 1090: loss 2.1005465984344482
 Step 1100: loss 2.075467109680176
 Step 1110: loss 2.0535550117492676
 Step 1120: loss 2.0131611824035645
 Step 1130: loss 2.2087831497192383
 Step 1140: loss 2.177764415740967
 Step 1150: loss 2.0178542137145996
 Step 1160: loss 2.0178041458129883
 Step 1170: loss 2.078481674194336
 Step 1180: loss 1.9789097309112549
 Step 1190: loss 1.9881665706634521
 Step 1200: loss 2.0550920963287354
 [Seed 89] Step 1200 --> val loss 2.2621707916259766
 Step 1210: loss 2.161663055419922
 Step 1220: loss 2.040778160095215
 Step 1230: loss 2.057234525680542
 Step 1240: loss 1.872894048690796
 Step 1250: loss 1.9808168411254883
 Step 1260: loss 2.0500893592834473
 Step 1270: loss 2.077557325363159
 Step 1280: loss 1.9349559545516968
 Step 1290: loss 2.2352120876312256
 Step 1300: loss 2.095331907272339
 Step 1310: loss 2.0936472415924072
 Step 1320: loss 2.0726571083068848
 Step 1330: loss 2.0686583518981934
 Step 1340: loss 1.9940783977508545
 Step 1350: loss 2.034498691558838
 Step 1360: loss 2.0525214672088623
 Step 1370: loss 1.9828956127166748
 Step 1380: loss 1.9590139389038086
 Step 1390: loss 2.1794066429138184
 Step 1400: loss 1.989534616470337
 [Seed 89] Step 1400 --> val loss 2.2703940868377686
 Step 1410: loss 1.987283706665039
 Step 1420: loss 1.9598793983459473
 Step 1430: loss 2.0674352645874023
 Step 1440: loss 2.0501327514648438
 Step 1450: loss 2.0543124675750732
 Step 1460: loss 2.022019863128662
 Step 1470: loss 1.9661893844604492
 Step 1480: loss 1.893119215965271
 Step 1490: loss 1.9607200622558594
 Step 1500: loss 2.092149257659912
 Step 1510: loss 2.048460006713867
 Step 1520: loss 2.0240468978881836
 Step 1530: loss 1.967643141746521
 Step 1540: loss 1.8912334442138672
 Step 1550: loss 1.886991262435913
 Step 1560: loss 1.8630508184432983
 Step 1570: loss 1.8551750183105469
 Step 1580: loss 1.9401795864105225
 Step 1590: loss 1.8835809230804443
 Step 1600: loss 1.929457426071167
 [Seed 89] Step 1600 --> val loss 2.1359095573425293
 Step 1610: loss 1.8702787160873413
 Step 1620: loss 1.8100541830062866
 Step 1630: loss 1.910729169845581
 Step 1640: loss 1.955016851425171
 Step 1650: loss 1.780155062675476
 Step 1660: loss 1.8860782384872437
 Step 1670: loss 1.8793268203735352
 Step 1680: loss 1.8317363262176514
 Step 1690: loss 1.8124016523361206
 Step 1700: loss 1.9191560745239258
 Step 1710: loss 1.8751904964447021
 Step 1720: loss 1.9118436574935913
 Step 1730: loss 1.8124583959579468
 Step 1740: loss 1.9030482769012451
 Step 1750: loss 1.7846078872680664
 Step 1760: loss 1.860457420349121
 Step 1770: loss 1.8777952194213867
 Step 1780: loss 1.8806421756744385
 Step 1790: loss 1.8962981700897217
 Step 1800: loss 1.861546516418457
 [Seed 89] Step 1800 --> val loss 2.1009254455566406
 Step 1810: loss 1.8107764720916748
 Step 1820: loss 1.849408507347107
 Step 1830: loss 1.8715757131576538
 Step 1840: loss 1.9005111455917358
 Step 1850: loss 1.8635915517807007
 Step 1860: loss 1.872169852256775
 Step 1870: loss 1.8446391820907593
 Step 1880: loss 1.903198003768921
 Step 1890: loss 1.8251131772994995
 Step 1900: loss 1.898982048034668
 Step 1910: loss 1.7014148235321045
 Step 1920: loss 1.8094229698181152
 Step 1930: loss 1.8326966762542725
 Step 1940: loss 1.810787558555603
 Step 1950: loss 1.9557654857635498
 Step 1960: loss 1.7736839056015015
 Step 1970: loss 1.759879469871521
 Step 1980: loss 1.8925933837890625
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 89] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 0 32 53  1 51 53 57 58  1 58] ...
 First target sequence: [32 53  1 51 53 57 58  1 58 46] ...

 Decoded input: 
 To most that teach.

 PERDITA:
 Your pardon, sir; for this
 I'll b

 Decoded target: To most that teach.

 PERDITA:
 Your pardon, sir; for this
 I'll bl
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.202552795410156
 [Seed 90] Step 0 --> val loss 4.187029838562012
 Step 10: loss 3.882206439971924
 Step 20: loss 3.2883810997009277
 Step 30: loss 2.956338882446289
 Step 40: loss 2.7130794525146484
 Step 50: loss 2.625120162963867
 Step 60: loss 2.5462660789489746
 Step 70: loss 2.431349515914917
 Step 80: loss 2.344714641571045
 Step 90: loss 2.2807445526123047
 Step 100: loss 2.234774351119995
 Step 110: loss 2.3168904781341553
 Step 120: loss 2.239839553833008
 Step 130: loss 2.2153286933898926
 Step 140: loss 2.2634568214416504
 Step 150: loss 2.183112859725952
 Step 160: loss 2.150625228881836
 Step 170: loss 2.168769359588623
 Step 180: loss 2.1542444229125977
 Step 190: loss 2.017770290374756
 Step 200: loss 2.063058614730835
 [Seed 90] Step 200 --> val loss 2.284893274307251
 Step 210: loss 2.009884834289551
 Step 220: loss 2.142430543899536
 Step 230: loss 2.051994800567627
 Step 240: loss 2.0430333614349365
 Step 250: loss 1.9141954183578491
 Step 260: loss 2.143242597579956
 Step 270: loss 2.0616302490234375
 Step 280: loss 2.0030670166015625
 Step 290: loss 1.8793885707855225
 Step 300: loss 2.0694403648376465
 Step 310: loss 2.137974977493286
 Step 320: loss 2.037121534347534
 Step 330: loss 2.085371494293213
 Step 340: loss 2.0826029777526855
 Step 350: loss 2.1442744731903076
 Step 360: loss 1.908724069595337
 Step 370: loss 1.9603707790374756
 Step 380: loss 2.063812732696533
 Step 390: loss 2.0030033588409424
 Step 400: loss 1.9803826808929443
 [Seed 90] Step 400 --> val loss 2.1909289360046387
 Step 410: loss 2.027601480484009
 Step 420: loss 2.0028083324432373
 Step 430: loss 2.0689313411712646
 Step 440: loss 1.9910392761230469
 Step 450: loss 1.9297622442245483
 Step 460: loss 2.099538803100586
 Step 470: loss 2.030240535736084
 Step 480: loss 2.041886329650879
 Step 490: loss 2.049266815185547
 Step 500: loss 1.9141058921813965
 Step 510: loss 1.97641921043396
 Step 520: loss 2.088425636291504
 Step 530: loss 2.0779848098754883
 Step 540: loss 1.9845861196517944
 Step 550: loss 1.9907042980194092
 Step 560: loss 2.039874315261841
 Step 570: loss 2.0225419998168945
 Step 580: loss 1.9751348495483398
 Step 590: loss 2.0206570625305176
 Step 600: loss 2.056276798248291
 [Seed 90] Step 600 --> val loss 2.1452507972717285
 Step 610: loss 1.958902359008789
 Step 620: loss 2.0455260276794434
 Step 630: loss 1.946141242980957
 Step 640: loss 1.9533812999725342
 Step 650: loss 2.04988956451416
 Step 660: loss 2.022982120513916
 Step 670: loss 1.9574611186981201
 Step 680: loss 2.075164318084717
 Step 690: loss 2.043649196624756
 Step 700: loss 1.932849645614624
 Step 710: loss 1.9539417028427124
 Step 720: loss 2.150104522705078
 Step 730: loss 2.12391996383667
 Step 740: loss 2.1128268241882324
 Step 750: loss 2.097926139831543
 Step 760: loss 1.9629048109054565
 Step 770: loss 2.0854556560516357
 Step 780: loss 2.1117687225341797
 Step 790: loss 2.1483683586120605
 Step 800: loss 2.011237621307373
 [Seed 90] Step 800 --> val loss 2.392749071121216
 Step 810: loss 2.0617563724517822
 Step 820: loss 2.1305136680603027
 Step 830: loss 2.0594801902770996
 Step 840: loss 2.1324989795684814
 Step 850: loss 1.9123141765594482
 Step 860: loss 1.9620555639266968
 Step 870: loss 1.9037110805511475
 Step 880: loss 1.9799020290374756
 Step 890: loss 1.9695253372192383
 Step 900: loss 1.9687808752059937
 Step 910: loss 1.9265122413635254
 Step 920: loss 2.071131706237793
 Step 930: loss 1.9134223461151123
 Step 940: loss 2.0185391902923584
 Step 950: loss 1.9203161001205444
 Step 960: loss 2.013406276702881
 Step 970: loss 1.9401172399520874
 Step 980: loss 2.012338399887085
 Step 990: loss 1.9764692783355713
 Step 1000: loss 1.9676539897918701
 [Seed 90] Step 1000 --> val loss 2.2798163890838623
 Step 1010: loss 2.0210793018341064
 Step 1020: loss 1.9754550457000732
 Step 1030: loss 2.11181640625
 Step 1040: loss 2.0255584716796875
 Step 1050: loss 2.028364896774292
 Step 1060: loss 2.041015625
 Step 1070: loss 2.0201854705810547
 Step 1080: loss 1.9775407314300537
 Step 1090: loss 1.995153784751892
 Step 1100: loss 1.896005630493164
 Step 1110: loss 1.9424794912338257
 Step 1120: loss 1.9296420812606812
 Step 1130: loss 2.005998373031616
 Step 1140: loss 1.9100435972213745
 Step 1150: loss 1.9886382818222046
 Step 1160: loss 2.1901488304138184
 Step 1170: loss 1.9958398342132568
 Step 1180: loss 1.9814093112945557
 Step 1190: loss 1.7519540786743164
 Step 1200: loss 1.912339448928833
 [Seed 90] Step 1200 --> val loss 2.129246473312378
 Step 1210: loss 1.999957799911499
 Step 1220: loss 1.9632266759872437
 Step 1230: loss 1.988644003868103
 Step 1240: loss 1.8324315547943115
 Step 1250: loss 1.8874667882919312
 Step 1260: loss 1.84521484375
 Step 1270: loss 1.9483067989349365
 Step 1280: loss 1.9398486614227295
 Step 1290: loss 1.853940486907959
 Step 1300: loss 1.8825181722640991
 Step 1310: loss 1.9154266119003296
 Step 1320: loss 1.893390417098999
 Step 1330: loss 1.9075963497161865
 Step 1340: loss 1.9853366613388062
 Step 1350: loss 1.89476478099823
 Step 1360: loss 1.876847267150879
 Step 1370: loss 1.8571927547454834
 Step 1380: loss 1.8646349906921387
 Step 1390: loss 1.794474482536316
 Step 1400: loss 1.920617938041687
 [Seed 90] Step 1400 --> val loss 2.11704683303833
 Step 1410: loss 1.712249517440796
 Step 1420: loss 1.8685048818588257
 Step 1430: loss 1.947245478630066
 Step 1440: loss 2.022444248199463
 Step 1450: loss 1.8157601356506348
 Step 1460: loss 1.8897603750228882
 Step 1470: loss 1.848210096359253
 Step 1480: loss 1.8444021940231323
 Step 1490: loss 1.8732059001922607
 Step 1500: loss 1.7803200483322144
 Step 1510: loss 1.9500157833099365
 Step 1520: loss 1.7984431982040405
 Step 1530: loss 1.780112624168396
 Step 1540: loss 1.7205591201782227
 Step 1550: loss 1.8329904079437256
 Step 1560: loss 1.8404550552368164
 Step 1570: loss 1.9002645015716553
 Step 1580: loss 1.8534296751022339
 Step 1590: loss 1.845388412475586
 Step 1600: loss 1.8087466955184937
 [Seed 90] Step 1600 --> val loss 2.014726400375366
 Step 1610: loss 1.889878273010254
 Step 1620: loss 1.8659746646881104
 Step 1630: loss 1.697911024093628
 Step 1640: loss 1.7599449157714844
 Step 1650: loss 1.7806404829025269
 Step 1660: loss 1.7571325302124023
 Step 1670: loss 1.843559741973877
 Step 1680: loss 1.7870256900787354
 Step 1690: loss 1.8582665920257568
 Step 1700: loss 1.7191370725631714
 Step 1710: loss 1.7099336385726929
 Step 1720: loss 1.8076528310775757
 Step 1730: loss 1.901686668395996
 Step 1740: loss 1.8067266941070557
 Step 1750: loss 1.8347314596176147
 Step 1760: loss 1.8765990734100342
 Step 1770: loss 1.7772159576416016
 Step 1780: loss 1.8647370338439941
 Step 1790: loss 1.735784888267517
 Step 1800: loss 1.661874771118164
 [Seed 90] Step 1800 --> val loss 1.9120606184005737
 Step 1810: loss 1.709998369216919
 Step 1820: loss 1.772439956665039
 Step 1830: loss 1.810187578201294
 Step 1840: loss 1.7622997760772705
 Step 1850: loss 1.6401793956756592
 Step 1860: loss 1.7799499034881592
 Step 1870: loss 1.762162446975708
 Step 1880: loss 1.7619750499725342
 Step 1890: loss 1.862586259841919
 Step 1900: loss 1.79986572265625
 Step 1910: loss 1.8191895484924316
 Step 1920: loss 1.7153568267822266
 Step 1930: loss 1.7292178869247437
 Step 1940: loss 1.7864396572113037
 Step 1950: loss 1.6633124351501465
 Step 1960: loss 1.831115484237671
 Step 1970: loss 1.6400312185287476
 Step 1980: loss 1.7432664632797241
 Step 1990: loss 1.6320103406906128
 Step 2000: loss 1.766810417175293
 [Seed 90] Step 2000 --> val loss 1.8812755346298218
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [61 39 56 42 57 46 47 54 11  0] ...
 First target sequence: [39 56 42 57 46 47 54 11  0 18] ...

 Decoded input: wardship;
 For well we know, no hand of blood and bone
 Can gripe 

 Decoded target: ardship;
 For well we know, no hand of blood and bone
 Can gripe t
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.198585510253906
 [Seed 91] Step 0 --> val loss 4.181880474090576
 Step 10: loss 3.862649917602539
 Step 20: loss 3.294417381286621
 Step 30: loss 2.7458033561706543
 Step 40: loss 2.6780600547790527
 Step 50: loss 2.4631972312927246
 Step 60: loss 2.487180709838867
 Step 70: loss 2.3840491771698
 Step 80: loss 2.377269744873047
 Step 90: loss 2.3257205486297607
 Step 100: loss 2.3027701377868652
 Step 110: loss 2.253061294555664
 Step 120: loss 2.2256760597229004
 Step 130: loss 2.2230136394500732
 Step 140: loss 2.265120506286621
 Step 150: loss 1.9903771877288818
 Step 160: loss 2.1617238521575928
 Step 170: loss 2.1904520988464355
 Step 180: loss 2.1506552696228027
 Step 190: loss 2.052591562271118
 Step 200: loss 2.0259647369384766
 [Seed 91] Step 200 --> val loss 2.2662508487701416
 Step 210: loss 2.1121065616607666
 Step 220: loss 2.0591931343078613
 Step 230: loss 2.173556327819824
 Step 240: loss 1.9021610021591187
 Step 250: loss 2.0339534282684326
 Step 260: loss 2.0610785484313965
 Step 270: loss 1.9503110647201538
 Step 280: loss 2.129176378250122
 Step 290: loss 2.088778495788574
 Step 300: loss 2.02771258354187
 Step 310: loss 2.0119643211364746
 Step 320: loss 2.0492360591888428
 Step 330: loss 2.082674980163574
 Step 340: loss 2.034144163131714
 Step 350: loss 2.0873348712921143
 Step 360: loss 2.052917003631592
 Step 370: loss 2.0721473693847656
 Step 380: loss 2.101033926010132
 Step 390: loss 1.9695641994476318
 Step 400: loss 2.002622127532959
 [Seed 91] Step 400 --> val loss 2.2747316360473633
 Step 410: loss 2.0369670391082764
 Step 420: loss 2.0468688011169434
 Step 430: loss 2.014681816101074
 Step 440: loss 1.9865500926971436
 Step 450: loss 2.0766425132751465
 Step 460: loss 2.041687250137329
 Step 470: loss 2.1668426990509033
 Step 480: loss 1.973026990890503
 Step 490: loss 1.8854434490203857
 Step 500: loss 2.0718696117401123
 Step 510: loss 1.982447862625122
 Step 520: loss 2.0061137676239014
 Step 530: loss 2.0512611865997314
 Step 540: loss 2.026947498321533
 Step 550: loss 2.012845516204834
 Step 560: loss 2.021944999694824
 Step 570: loss 1.9966707229614258
 Step 580: loss 1.9785460233688354
 Step 590: loss 2.0193018913269043
 Step 600: loss 1.955306053161621
 [Seed 91] Step 600 --> val loss 2.3138046264648438
 Step 610: loss 2.0878896713256836
 Step 620: loss 2.0224523544311523
 Step 630: loss 1.9452590942382812
 Step 640: loss 1.9989112615585327
 Step 650: loss 1.9510501623153687
 Step 660: loss 1.9983036518096924
 Step 670: loss 2.070636510848999
 Step 680: loss 2.0153605937957764
 Step 690: loss 1.9958518743515015
 Step 700: loss 1.9800485372543335
 Step 710: loss 2.174133062362671
 Step 720: loss 2.05633544921875
 Step 730: loss 2.1127302646636963
 Step 740: loss 2.072744846343994
 Step 750: loss 2.0344016551971436
 Step 760: loss 1.9863042831420898
 Step 770: loss 1.9118531942367554
 Step 780: loss 2.012615442276001
 Step 790: loss 2.0751686096191406
 Step 800: loss 2.0218944549560547
 [Seed 91] Step 800 --> val loss 2.3248982429504395
 Step 810: loss 1.9905365705490112
 Step 820: loss 1.9217698574066162
 Step 830: loss 1.9693082571029663
 Step 840: loss 2.214832305908203
 Step 850: loss 2.0117645263671875
 Step 860: loss 2.0976500511169434
 Step 870: loss 2.0764546394348145
 Step 880: loss 2.15678334236145
 Step 890: loss 2.117870330810547
 Step 900: loss 1.997534155845642
 Step 910: loss 2.155703544616699
 Step 920: loss 2.1329097747802734
 Step 930: loss 2.0548713207244873
 Step 940: loss 2.177544116973877
 Step 950: loss 1.9355311393737793
 Step 960: loss 2.138680934906006
 Step 970: loss 2.124502658843994
 Step 980: loss 2.0360968112945557
 Step 990: loss 2.0007424354553223
 Step 1000: loss 2.0268161296844482
 [Seed 91] Step 1000 --> val loss 2.2740609645843506
 Step 1010: loss 1.9928438663482666
 Step 1020: loss 2.1003060340881348
 Step 1030: loss 2.088209629058838
 Step 1040: loss 2.2153825759887695
 Step 1050: loss 2.1304495334625244
 Step 1060: loss 2.191458225250244
 Step 1070: loss 2.2005982398986816
 Step 1080: loss 2.131027936935425
 Step 1090: loss 2.1208252906799316
 Step 1100: loss 2.0191125869750977
 Step 1110: loss 2.0511832237243652
 Step 1120: loss 1.9955451488494873
 Step 1130: loss 2.1415317058563232
 Step 1140: loss 2.1510276794433594
 Step 1150: loss 2.078866720199585
 Step 1160: loss 2.069051742553711
 Step 1170: loss 2.1138577461242676
 Step 1180: loss 2.093507766723633
 Step 1190: loss 2.1087403297424316
 Step 1200: loss 2.1241114139556885
 [Seed 91] Step 1200 --> val loss 2.342367649078369
 Step 1210: loss 2.019470691680908
 Step 1220: loss 2.125800848007202
 Step 1230: loss 2.059856414794922
 Step 1240: loss 2.0443367958068848
 Step 1250: loss 2.082338333129883
 Step 1260: loss 2.023347854614258
 Step 1270: loss 2.1056690216064453
 Step 1280: loss 2.012990713119507
 Step 1290: loss 2.130958318710327
 Step 1300: loss 2.0269598960876465
 Step 1310: loss 2.035975933074951
 Step 1320: loss 1.985560417175293
 Step 1330: loss 2.1000680923461914
 Step 1340: loss 2.0363526344299316
 Step 1350: loss 2.1039175987243652
 Step 1360: loss 2.1572041511535645
 Step 1370: loss 2.012418270111084
 Step 1380: loss 2.0274996757507324
 Step 1390: loss 2.165627956390381
 Step 1400: loss 1.9777412414550781
 [Seed 91] Step 1400 --> val loss 2.277043342590332
 Step 1410: loss 2.018663167953491
 Step 1420: loss 2.0239052772521973
 Step 1430: loss 2.170957565307617
 Step 1440: loss 2.0903260707855225
 Step 1450: loss 1.9910520315170288
 Step 1460: loss 1.9242773056030273
 Step 1470: loss 1.9289758205413818
 Step 1480: loss 2.022566795349121
 Step 1490: loss 2.071099042892456
 Step 1500: loss 1.974586009979248
 Step 1510: loss 2.057342052459717
 Step 1520: loss 2.0194051265716553
 Step 1530: loss 1.9447091817855835
 Step 1540: loss 2.082714557647705
 Step 1550: loss 1.8437244892120361
 Step 1560: loss 1.896227240562439
 Step 1570: loss 1.950218915939331
 Step 1580: loss 1.9277477264404297
 Step 1590: loss 1.9346727132797241
 Step 1600: loss 2.0539591312408447
 [Seed 91] Step 1600 --> val loss 2.196009874343872
 Step 1610: loss 1.9688196182250977
 Step 1620: loss 1.842734694480896
 Step 1630: loss 1.8829030990600586
 Step 1640: loss 1.9393079280853271
 Step 1650: loss 1.8914941549301147
 Step 1660: loss 1.8970019817352295
 Step 1670: loss 1.8631609678268433
 Step 1680: loss 2.0056025981903076
 Step 1690: loss 1.9736416339874268
 Step 1700: loss 1.8589643239974976
 Step 1710: loss 1.8610814809799194
 Step 1720: loss 1.7853387594223022
 Step 1730: loss 1.9244639873504639
 Step 1740: loss 1.8891581296920776
 Step 1750: loss 1.8652819395065308
 Step 1760: loss 1.8970658779144287
 Step 1770: loss 1.952149748802185
 Step 1780: loss 1.8326051235198975
 Step 1790: loss 1.7793712615966797
 Step 1800: loss 1.8589071035385132
 [Seed 91] Step 1800 --> val loss 2.0814297199249268
 Step 1810: loss 1.8926968574523926
 Step 1820: loss 1.8799118995666504
 Step 1830: loss 1.7975447177886963
 Step 1840: loss 1.780981421470642
 Step 1850: loss 1.976653814315796
 Step 1860: loss 1.7838866710662842
 Step 1870: loss 1.9153316020965576
 Step 1880: loss 1.795750379562378
 Step 1890: loss 1.762028455734253
 Step 1900: loss 1.84458327293396
 Step 1910: loss 1.8871288299560547
 Step 1920: loss 1.9522912502288818
 Step 1930: loss 1.8424315452575684
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 91] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [51 43 11  1 52 53 61  1 21  1] ...
 First target sequence: [43 11  1 52 53 61  1 21  1 61] ...

 Decoded input: me; now I will unmask.
 This is that face, thou cruel Angelo,
 Whi

 Decoded target: e; now I will unmask.
 This is that face, thou cruel Angelo,
 Whic
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.202316761016846
 [Seed 92] Step 0 --> val loss 4.183539390563965
 Step 10: loss 3.857377529144287
 Step 20: loss 3.3405110836029053
 Step 30: loss 2.8630175590515137
 Step 40: loss 2.741501808166504
 Step 50: loss 2.5725088119506836
 Step 60: loss 2.4666357040405273
 Step 70: loss 2.4297404289245605
 Step 80: loss 2.482165813446045
 Step 90: loss 2.349858283996582
 Step 100: loss 2.314014434814453
 Step 110: loss 2.2457494735717773
 Step 120: loss 2.157548427581787
 Step 130: loss 2.2336933612823486
 Step 140: loss 2.1894755363464355
 Step 150: loss 2.1780097484588623
 Step 160: loss 2.147953748703003
 Step 170: loss 2.1358351707458496
 Step 180: loss 2.0142483711242676
 Step 190: loss 2.007439136505127
 Step 200: loss 2.1057231426239014
 [Seed 92] Step 200 --> val loss 2.3803250789642334
 Step 210: loss 2.072418689727783
 Step 220: loss 2.09238338470459
 Step 230: loss 2.1326241493225098
 Step 240: loss 2.0519566535949707
 Step 250: loss 2.039501667022705
 Step 260: loss 2.023245334625244
 Step 270: loss 1.9612921476364136
 Step 280: loss 2.091460704803467
 Step 290: loss 2.10482120513916
 Step 300: loss 1.9998754262924194
 Step 310: loss 2.0073137283325195
 Step 320: loss 2.0636470317840576
 Step 330: loss 1.977957010269165
 Step 340: loss 1.972421646118164
 Step 350: loss 2.183696746826172
 Step 360: loss 2.0449376106262207
 Step 370: loss 1.9351314306259155
 Step 380: loss 2.012362241744995
 Step 390: loss 2.026707410812378
 Step 400: loss 2.0075132846832275
 [Seed 92] Step 400 --> val loss 2.2834725379943848
 Step 410: loss 2.079240322113037
 Step 420: loss 1.8939120769500732
 Step 430: loss 1.9334468841552734
 Step 440: loss 2.024183988571167
 Step 450: loss 1.9431915283203125
 Step 460: loss 1.9861133098602295
 Step 470: loss 2.0085670948028564
 Step 480: loss 1.8817098140716553
 Step 490: loss 2.056656837463379
 Step 500: loss 2.1040165424346924
 Step 510: loss 2.0723876953125
 Step 520: loss 1.9672845602035522
 Step 530: loss 1.9769103527069092
 Step 540: loss 2.0065953731536865
 Step 550: loss 1.9391343593597412
 Step 560: loss 1.9150991439819336
 Step 570: loss 1.994032621383667
 Step 580: loss 2.11372447013855
 Step 590: loss 2.1527934074401855
 Step 600: loss 2.0344159603118896
 [Seed 92] Step 600 --> val loss 2.2505404949188232
 Step 610: loss 2.099043369293213
 Step 620: loss 1.9972466230392456
 Step 630: loss 2.0821073055267334
 Step 640: loss 2.019725799560547
 Step 650: loss 2.0863287448883057
 Step 660: loss 2.0524346828460693
 Step 670: loss 1.9968079328536987
 Step 680: loss 2.1288583278656006
 Step 690: loss 2.04707407951355
 Step 700: loss 2.0792360305786133
 Step 710: loss 2.0808467864990234
 Step 720: loss 2.126800060272217
 Step 730: loss 2.065556287765503
 Step 740: loss 2.1404523849487305
 Step 750: loss 2.065192699432373
 Step 760: loss 2.0313825607299805
 Step 770: loss 2.0336742401123047
 Step 780: loss 2.010854959487915
 Step 790: loss 2.057483673095703
 Step 800: loss 1.970132827758789
 [Seed 92] Step 800 --> val loss 2.330493211746216
 Step 810: loss 2.072129964828491
 Step 820: loss 2.0292887687683105
 Step 830: loss 2.0704784393310547
 Step 840: loss 2.2132060527801514
 Step 850: loss 2.0486392974853516
 Step 860: loss 2.0680127143859863
 Step 870: loss 2.0259361267089844
 Step 880: loss 2.0463690757751465
 Step 890: loss 2.2276716232299805
 Step 900: loss 2.1521213054656982
 Step 910: loss 1.971837043762207
 Step 920: loss 2.1565184593200684
 Step 930: loss 2.135413646697998
 Step 940: loss 2.0165812969207764
 Step 950: loss 1.9411542415618896
 Step 960: loss 2.0772342681884766
 Step 970: loss 2.0803382396698
 Step 980: loss 2.0194060802459717
 Step 990: loss 2.0911660194396973
 Step 1000: loss 2.0341756343841553
 [Seed 92] Step 1000 --> val loss 2.2407066822052
 Step 1010: loss 2.069094657897949
 Step 1020: loss 2.002955913543701
 Step 1030: loss 2.0224149227142334
 Step 1040: loss 2.127516746520996
 Step 1050: loss 2.0812478065490723
 Step 1060: loss 2.089895486831665
 Step 1070: loss 2.031778335571289
 Step 1080: loss 2.020226001739502
 Step 1090: loss 2.0498530864715576
 Step 1100: loss 2.0730795860290527
 Step 1110: loss 1.966719388961792
 Step 1120: loss 2.0515990257263184
 Step 1130: loss 1.9646804332733154
 Step 1140: loss 2.018889904022217
 Step 1150: loss 2.120978355407715
 Step 1160: loss 2.0025970935821533
 Step 1170: loss 2.051259756088257
 Step 1180: loss 2.106748342514038
 Step 1190: loss 2.0156571865081787
 Step 1200: loss 2.0719499588012695
 [Seed 92] Step 1200 --> val loss 2.2483456134796143
 Step 1210: loss 2.1049747467041016
 Step 1220: loss 2.041104555130005
 Step 1230: loss 2.041029930114746
 Step 1240: loss 2.002908229827881
 Step 1250: loss 2.060784101486206
 Step 1260: loss 1.9736688137054443
 Step 1270: loss 2.0075595378875732
 Step 1280: loss 2.0503528118133545
 Step 1290: loss 2.0844953060150146
 Step 1300: loss 2.047313690185547
 Step 1310: loss 2.134300947189331
 Step 1320: loss 2.013956069946289
 Step 1330: loss 1.9078949689865112
 Step 1340: loss 2.0344650745391846
 Step 1350: loss 2.0335097312927246
 Step 1360: loss 2.0339765548706055
 Step 1370: loss 1.9186567068099976
 Step 1380: loss 2.039511203765869
 Step 1390: loss 1.9610576629638672
 Step 1400: loss 1.929027795791626
 [Seed 92] Step 1400 --> val loss 2.1517603397369385
 Step 1410: loss 2.0646770000457764
 Step 1420: loss 1.9771943092346191
 Step 1430: loss 2.037696599960327
 Step 1440: loss 2.129246234893799
 Step 1450: loss 1.9501436948776245
 Step 1460: loss 1.9339675903320312
 Step 1470: loss 1.8959472179412842
 Step 1480: loss 1.8500465154647827
 Step 1490: loss 1.8063805103302002
 Step 1500: loss 1.90216863155365
 Step 1510: loss 1.9119534492492676
 Step 1520: loss 1.8249783515930176
 Step 1530: loss 1.841657042503357
 Step 1540: loss 1.9748969078063965
 Step 1550: loss 2.027785539627075
 Step 1560: loss 1.9347212314605713
 Step 1570: loss 1.777875304222107
 Step 1580: loss 1.8887205123901367
 Step 1590: loss 1.816420316696167
 Step 1600: loss 1.7994718551635742
 [Seed 92] Step 1600 --> val loss 2.023589611053467
 Step 1610: loss 1.9353289604187012
 Step 1620: loss 1.9482924938201904
 Step 1630: loss 1.928807258605957
 Step 1640: loss 1.8923448324203491
 Step 1650: loss 1.851930022239685
 Step 1660: loss 1.840040922164917
 Step 1670: loss 1.8261536359786987
 Step 1680: loss 1.8388175964355469
 Step 1690: loss 1.890082597732544
 Step 1700: loss 1.7672423124313354
 Step 1710: loss 1.8162178993225098
 Step 1720: loss 1.8230332136154175
 Step 1730: loss 1.792851209640503
 Step 1740: loss 1.7784069776535034
 Step 1750: loss 1.8401451110839844
 Step 1760: loss 1.870269775390625
 Step 1770: loss 1.7883113622665405
 Step 1780: loss 1.8294854164123535
 Step 1790: loss 1.7811874151229858
 Step 1800: loss 1.8373353481292725
 [Seed 92] Step 1800 --> val loss 1.9966200590133667
 Step 1810: loss 2.0314557552337646
 Step 1820: loss 1.6357249021530151
 Step 1830: loss 1.7525876760482788
 Step 1840: loss 1.7832139730453491
 Step 1850: loss 1.8111227750778198
 Step 1860: loss 1.793921947479248
 Step 1870: loss 1.8447911739349365
 Step 1880: loss 1.9557437896728516
 Step 1890: loss 1.8893965482711792
 Step 1900: loss 1.7681115865707397
 Step 1910: loss 1.8632367849349976
 Step 1920: loss 1.6811304092407227
 Step 1930: loss 1.8534355163574219
 Step 1940: loss 1.8301233053207397
 Step 1950: loss 1.7118768692016602
 Step 1960: loss 1.7823758125305176
 Step 1970: loss 1.7287185192108154
 Step 1980: loss 1.8058788776397705
 Step 1990: loss 1.6643069982528687
 Step 2000: loss 1.8735597133636475
 [Seed 92] Step 2000 --> val loss 1.922793984413147
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [59 58  1 61 47 58 46  1 39  1] ...
 First target sequence: [58  1 61 47 58 46  1 39  1 56] ...

 Decoded input: ut with a rear-ward following Tybalt's death,
 'Romeo is banished

 Decoded target: t with a rear-ward following Tybalt's death,
 'Romeo is banished,
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.207757949829102
 [Seed 93] Step 0 --> val loss 4.183466911315918
 Step 10: loss 3.843233823776245
 Step 20: loss 3.2790775299072266
 Step 30: loss 2.8935227394104004
 Step 40: loss 2.8495640754699707
 Step 50: loss 2.5889041423797607
 Step 60: loss 2.490506649017334
 Step 70: loss 2.424783706665039
 Step 80: loss 2.3527917861938477
 Step 90: loss 2.344667673110962
 Step 100: loss 2.2754316329956055
 Step 110: loss 2.307854175567627
 Step 120: loss 2.2804222106933594
 Step 130: loss 2.1765589714050293
 Step 140: loss 2.2092225551605225
 Step 150: loss 2.251434326171875
 Step 160: loss 2.0885257720947266
 Step 170: loss 2.1532845497131348
 Step 180: loss 2.236985683441162
 Step 190: loss 2.0719614028930664
 Step 200: loss 2.180980682373047
 [Seed 93] Step 200 --> val loss 2.338381052017212
 Step 210: loss 2.0517282485961914
 Step 220: loss 2.115123987197876
 Step 230: loss 2.2478530406951904
 Step 240: loss 2.060244083404541
 Step 250: loss 2.1184468269348145
 Step 260: loss 2.177917718887329
 Step 270: loss 2.0705065727233887
 Step 280: loss 2.0676333904266357
 Step 290: loss 2.1164350509643555
 Step 300: loss 2.075535535812378
 Step 310: loss 1.9664421081542969
 Step 320: loss 2.0612645149230957
 Step 330: loss 1.9144294261932373
 Step 340: loss 2.0211780071258545
 Step 350: loss 2.108086347579956
 Step 360: loss 1.9105327129364014
 Step 370: loss 1.986061453819275
 Step 380: loss 1.9780501127243042
 Step 390: loss 2.107057571411133
 Step 400: loss 1.9321911334991455
 [Seed 93] Step 400 --> val loss 2.246919631958008
 Step 410: loss 2.157799482345581
 Step 420: loss 1.9933547973632812
 Step 430: loss 2.0315377712249756
 Step 440: loss 1.936431884765625
 Step 450: loss 2.0939111709594727
 Step 460: loss 1.9451364278793335
 Step 470: loss 2.1580810546875
 Step 480: loss 1.9641698598861694
 Step 490: loss 1.9175937175750732
 Step 500: loss 1.9907140731811523
 Step 510: loss 2.036865234375
 Step 520: loss 1.9583303928375244
 Step 530: loss 2.070464611053467
 Step 540: loss 2.083629608154297
 Step 550: loss 2.0786943435668945
 Step 560: loss 2.028554916381836
 Step 570: loss 2.0336246490478516
 Step 580: loss 2.0018692016601562
 Step 590: loss 1.93155038356781
 Step 600: loss 2.02317476272583
 [Seed 93] Step 600 --> val loss 2.2540769577026367
 Step 610: loss 2.035811185836792
 Step 620: loss 2.0552690029144287
 Step 630: loss 1.9726905822753906
 Step 640: loss 1.9354383945465088
 Step 650: loss 1.8819448947906494
 Step 660: loss 1.8407447338104248
 Step 670: loss 1.983443260192871
 Step 680: loss 1.8880891799926758
 Step 690: loss 1.9011338949203491
 Step 700: loss 1.9575088024139404
 Step 710: loss 2.065265417098999
 Step 720: loss 2.006423234939575
 Step 730: loss 1.9428033828735352
 Step 740: loss 1.9304733276367188
 Step 750: loss 1.9204190969467163
 Step 760: loss 2.0187125205993652
 Step 770: loss 1.9188750982284546
 Step 780: loss 2.0265700817108154
 Step 790: loss 2.015770196914673
 Step 800: loss 1.9669742584228516
 [Seed 93] Step 800 --> val loss 2.2541158199310303
 Step 810: loss 1.9741462469100952
 Step 820: loss 2.03364634513855
 Step 830: loss 1.9854810237884521
 Step 840: loss 1.9010385274887085
 Step 850: loss 1.9497478008270264
 Step 860: loss 1.9571819305419922
 Step 870: loss 1.9082655906677246
 Step 880: loss 1.8996942043304443
 Step 890: loss 2.027009963989258
 Step 900: loss 1.8986775875091553
 Step 910: loss 2.0114829540252686
 Step 920: loss 2.0752878189086914
 Step 930: loss 2.080608367919922
 Step 940: loss 2.0914158821105957
 Step 950: loss 1.9064873456954956
 Step 960: loss 1.9972623586654663
 Step 970: loss 2.0446536540985107
 Step 980: loss 1.9664462804794312
 Step 990: loss 2.159153461456299
 Step 1000: loss 2.024995803833008
 [Seed 93] Step 1000 --> val loss 2.2807185649871826
 Step 1010: loss 1.9759671688079834
 Step 1020: loss 1.9809627532958984
 Step 1030: loss 1.8256990909576416
 Step 1040: loss 1.863513708114624
 Step 1050: loss 1.9552339315414429
 Step 1060: loss 1.987021803855896
 Step 1070: loss 1.8288342952728271
 Step 1080: loss 1.9565467834472656
 Step 1090: loss 1.9009766578674316
 Step 1100: loss 1.9749540090560913
 Step 1110: loss 1.9029746055603027
 Step 1120: loss 1.9425472021102905
 Step 1130: loss 2.0149261951446533
 Step 1140: loss 1.8973824977874756
 Step 1150: loss 1.8406217098236084
 Step 1160: loss 1.8508384227752686
 Step 1170: loss 1.9047982692718506
 Step 1180: loss 1.9184579849243164
 Step 1190: loss 1.8611582517623901
 Step 1200: loss 1.8314164876937866
 [Seed 93] Step 1200 --> val loss 2.1573266983032227
 Step 1210: loss 1.8562663793563843
 Step 1220: loss 1.8446897268295288
 Step 1230: loss 1.8610131740570068
 Step 1240: loss 1.900527000427246
 Step 1250: loss 1.865227460861206
 Step 1260: loss 1.8432302474975586
 Step 1270: loss 1.8282194137573242
 Step 1280: loss 1.9018630981445312
 Step 1290: loss 1.9151958227157593
 Step 1300: loss 1.8379157781600952
 Step 1310: loss 1.838007926940918
 Step 1320: loss 1.7926357984542847
 Step 1330: loss 1.7551295757293701
 Step 1340: loss 1.8265808820724487
 Step 1350: loss 1.8451025485992432
 Step 1360: loss 1.8939387798309326
 Step 1370: loss 1.7381398677825928
 Step 1380: loss 1.9077064990997314
 Step 1390: loss 1.9394118785858154
 Step 1400: loss 1.7886613607406616
 [Seed 93] Step 1400 --> val loss 2.038219451904297
 Step 1410: loss 1.8528046607971191
 Step 1420: loss 1.82155442237854
 Step 1430: loss 1.9677374362945557
 Step 1440: loss 1.8402823209762573
 Step 1450: loss 1.9594745635986328
 Step 1460: loss 1.6289353370666504
 Step 1470: loss 1.7438815832138062
 Step 1480: loss 1.8000733852386475
 Step 1490: loss 1.8782991170883179
 Step 1500: loss 1.8291305303573608
 Step 1510: loss 1.8774551153182983
 Step 1520: loss 1.7528750896453857
 Step 1530: loss 1.8104429244995117
 Step 1540: loss 1.7237577438354492
 Step 1550: loss 1.8051180839538574
 Step 1560: loss 1.7752647399902344
 Step 1570: loss 1.7683250904083252
 Step 1580: loss 1.997291088104248
 Step 1590: loss 1.7608106136322021
 Step 1600: loss 1.732970118522644
 [Seed 93] Step 1600 --> val loss 1.9580458402633667
 Step 1610: loss 1.6558598279953003
 Step 1620: loss 1.7818961143493652
 Step 1630: loss 1.7684634923934937
 Step 1640: loss 1.605417013168335
 Step 1650: loss 1.709681510925293
 Step 1660: loss 1.6870088577270508
 Step 1670: loss 1.8339847326278687
 Step 1680: loss 1.7548727989196777
 Step 1690: loss 1.762265920639038
 Step 1700: loss 1.6768605709075928
 Step 1710: loss 1.8076817989349365
 Step 1720: loss 1.7015275955200195
 Step 1730: loss 1.6104692220687866
 Step 1740: loss 1.684229850769043
 Step 1750: loss 1.6226916313171387
 Step 1760: loss 1.6605799198150635
 Step 1770: loss 1.5878900289535522
 Step 1780: loss 1.6773788928985596
 Step 1790: loss 1.6377837657928467
 Step 1800: loss 1.632259726524353
 [Seed 93] Step 1800 --> val loss 1.9412342309951782
 Step 1810: loss 1.7413740158081055
 Step 1820: loss 1.609667181968689
 Step 1830: loss 1.7672513723373413
 Step 1840: loss 1.6921288967132568
 Step 1850: loss 1.7226860523223877
 Step 1860: loss 1.6584094762802124
 Step 1870: loss 1.8631929159164429
 Step 1880: loss 1.6336562633514404
 Step 1890: loss 1.6635780334472656
 Step 1900: loss 1.760242223739624
 Step 1910: loss 1.7281755208969116
 Step 1920: loss 1.6565923690795898
 Step 1930: loss 1.7105655670166016
 Step 1940: loss 1.5368221998214722
 Step 1950: loss 1.725172996520996
 Step 1960: loss 1.6413214206695557
 Step 1970: loss 1.7159560918807983
 Step 1980: loss 1.7489560842514038
 Step 1990: loss 1.7974328994750977
 Step 2000: loss 1.7020747661590576
 [Seed 93] Step 2000 --> val loss 1.9098514318466187
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [21  1 61 39 56 56 39 52 58  1] ...
 First target sequence: [ 1 61 39 56 56 39 52 58  1 58] ...

 Decoded input: I warrant thee.

 Second Murderer:
 Spoke like a tail fellow that 

 Decoded target:  warrant thee.

 Second Murderer:
 Spoke like a tail fellow that r
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.213118553161621
 [Seed 94] Step 0 --> val loss 4.185718536376953
 Step 10: loss 3.8565430641174316
 Step 20: loss 3.370089530944824
 Step 30: loss 2.801752805709839
 Step 40: loss 2.470850944519043
 Step 50: loss 2.5783982276916504
 Step 60: loss 2.501279830932617
 Step 70: loss 2.418307304382324
 Step 80: loss 2.3683528900146484
 Step 90: loss 2.2873055934906006
 Step 100: loss 2.429760694503784
 Step 110: loss 2.141413450241089
 Step 120: loss 2.286501884460449
 Step 130: loss 2.216561794281006
 Step 140: loss 2.156862258911133
 Step 150: loss 2.205227851867676
 Step 160: loss 2.1672821044921875
 Step 170: loss 2.0938379764556885
 Step 180: loss 2.15610408782959
 Step 190: loss 2.0385899543762207
 Step 200: loss 2.074552536010742
 [Seed 94] Step 200 --> val loss 2.1911427974700928
 Step 210: loss 2.1050548553466797
 Step 220: loss 2.208148956298828
 Step 230: loss 2.078805923461914
 Step 240: loss 2.0931107997894287
 Step 250: loss 2.166614532470703
 Step 260: loss 2.023735523223877
 Step 270: loss 2.071458101272583
 Step 280: loss 2.0289571285247803
 Step 290: loss 1.9521307945251465
 Step 300: loss 2.1061110496520996
 Step 310: loss 2.095724582672119
 Step 320: loss 1.9237010478973389
 Step 330: loss 2.0264196395874023
 Step 340: loss 1.9620269536972046
 Step 350: loss 2.0253732204437256
 Step 360: loss 2.0029191970825195
 Step 370: loss 1.9466657638549805
 Step 380: loss 1.9873082637786865
 Step 390: loss 2.101870059967041
 Step 400: loss 2.0149543285369873
 [Seed 94] Step 400 --> val loss 2.2603015899658203
 Step 410: loss 1.9604053497314453
 Step 420: loss 2.1066882610321045
 Step 430: loss 2.0262961387634277
 Step 440: loss 1.9078147411346436
 Step 450: loss 1.9396408796310425
 Step 460: loss 2.072279930114746
 Step 470: loss 1.956153154373169
 Step 480: loss 1.9695460796356201
 Step 490: loss 2.0293493270874023
 Step 500: loss 2.020911931991577
 Step 510: loss 1.921433925628662
 Step 520: loss 1.902317762374878
 Step 530: loss 1.9509227275848389
 Step 540: loss 1.9863297939300537
 Step 550: loss 1.9540953636169434
 Step 560: loss 2.101505756378174
 Step 570: loss 2.042853355407715
 Step 580: loss 2.0525593757629395
 Step 590: loss 2.092937469482422
 Step 600: loss 1.9439224004745483
 [Seed 94] Step 600 --> val loss 2.151036500930786
 Step 610: loss 2.0565881729125977
 Step 620: loss 2.008589267730713
 Step 630: loss 2.096839427947998
 Step 640: loss 2.015430450439453
 Step 650: loss 2.016521692276001
 Step 660: loss 1.9516873359680176
 Step 670: loss 2.1875882148742676
 Step 680: loss 1.972561240196228
 Step 690: loss 1.9467029571533203
 Step 700: loss 2.0279316902160645
 Step 710: loss 1.9681050777435303
 Step 720: loss 2.011013984680176
 Step 730: loss 1.9848506450653076
 Step 740: loss 1.9783809185028076
 Step 750: loss 1.9460376501083374
 Step 760: loss 2.1224236488342285
 Step 770: loss 1.9594792127609253
 Step 780: loss 2.030120611190796
 Step 790: loss 2.0675370693206787
 Step 800: loss 1.9210901260375977
 [Seed 94] Step 800 --> val loss 2.198075294494629
 Step 810: loss 2.058648109436035
 Step 820: loss 2.005932331085205
 Step 830: loss 2.0860185623168945
 Step 840: loss 1.9868700504302979
 Step 850: loss 1.9183166027069092
 Step 860: loss 1.9446958303451538
 Step 870: loss 2.013063430786133
 Step 880: loss 1.867156744003296
 Step 890: loss 1.8687275648117065
 Step 900: loss 2.0232579708099365
 Step 910: loss 2.1151461601257324
 Step 920: loss 1.9496574401855469
 Step 930: loss 2.055819511413574
 Step 940: loss 1.9666917324066162
 Step 950: loss 2.0434489250183105
 Step 960: loss 1.8753087520599365
 Step 970: loss 2.0942485332489014
 Step 980: loss 2.0762627124786377
 Step 990: loss 2.0920329093933105
 Step 1000: loss 2.0257480144500732
 [Seed 94] Step 1000 --> val loss 2.2583374977111816
 Step 1010: loss 1.8957003355026245
 Step 1020: loss 1.9535764455795288
 Step 1030: loss 1.980516791343689
 Step 1040: loss 1.946929931640625
 Step 1050: loss 1.9717525243759155
 Step 1060: loss 2.0356452465057373
 Step 1070: loss 1.945585012435913
 Step 1080: loss 2.0752713680267334
 Step 1090: loss 1.9501594305038452
 Step 1100: loss 2.0171611309051514
 Step 1110: loss 2.033924102783203
 Step 1120: loss 1.8813289403915405
 Step 1130: loss 1.852919578552246
 Step 1140: loss 1.857889175415039
 Step 1150: loss 1.9000242948532104
 Step 1160: loss 1.9716905355453491
 Step 1170: loss 1.9603503942489624
 Step 1180: loss 1.9178245067596436
 Step 1190: loss 1.9352638721466064
 Step 1200: loss 1.983835220336914
 [Seed 94] Step 1200 --> val loss 2.1485683917999268
 Step 1210: loss 1.9602116346359253
 Step 1220: loss 1.9912300109863281
 Step 1230: loss 1.9552958011627197
 Step 1240: loss 1.8629298210144043
 Step 1250: loss 2.0386369228363037
 Step 1260: loss 1.9569848775863647
 Step 1270: loss 1.9663634300231934
 Step 1280: loss 1.9035662412643433
 Step 1290: loss 1.942395567893982
 Step 1300: loss 1.9898546934127808
 Step 1310: loss 2.0072617530822754
 Step 1320: loss 1.9725269079208374
 Step 1330: loss 2.00203275680542
 Step 1340: loss 1.9209818840026855
 Step 1350: loss 1.9914891719818115
 Step 1360: loss 2.0985875129699707
 Step 1370: loss 2.012850284576416
 Step 1380: loss 2.042907238006592
 Step 1390: loss 2.0612404346466064
 Step 1400: loss 2.017167091369629
 [Seed 94] Step 1400 --> val loss 2.1311488151550293
 Step 1410: loss 1.9832959175109863
 Step 1420: loss 2.074005126953125
 Step 1430: loss 1.9611470699310303
 Step 1440: loss 1.8292993307113647
 Step 1450: loss 1.974198818206787
 Step 1460: loss 1.9494152069091797
 Step 1470: loss 1.9289743900299072
 Step 1480: loss 1.9144948720932007
 Step 1490: loss 1.9304978847503662
 Step 1500: loss 1.8871040344238281
 Step 1510: loss 1.9061607122421265
 Step 1520: loss 1.918169379234314
 Step 1530: loss 1.844081163406372
 Step 1540: loss 1.9485127925872803
 Step 1550: loss 1.8905205726623535
 Step 1560: loss 1.8578780889511108
 Step 1570: loss 1.8929526805877686
 Step 1580: loss 1.863168478012085
 Step 1590: loss 1.9225738048553467
 Step 1600: loss 1.903381586074829
 [Seed 94] Step 1600 --> val loss 1.9931299686431885
 Step 1610: loss 1.8931430578231812
 Step 1620: loss 1.8236427307128906
 Step 1630: loss 1.8338130712509155
 Step 1640: loss 1.8318908214569092
 Step 1650: loss 1.836098551750183
 Step 1660: loss 1.8854625225067139
 Step 1670: loss 1.853434443473816
 Step 1680: loss 1.8678849935531616
 Step 1690: loss 1.8655093908309937
 Step 1700: loss 1.8640470504760742
 Step 1710: loss 1.7596025466918945
 Step 1720: loss 1.7554502487182617
 Step 1730: loss 1.8771909475326538
 Step 1740: loss 1.8359742164611816
 Step 1750: loss 1.8066905736923218
 Step 1760: loss 1.9188640117645264
 Step 1770: loss 1.8633935451507568
 Step 1780: loss 1.7732585668563843
 Step 1790: loss 1.8980039358139038
 Step 1800: loss 1.747502088546753
 [Seed 94] Step 1800 --> val loss 1.9306801557540894
 Step 1810: loss 1.7381466627120972
 Step 1820: loss 1.7433032989501953
 Step 1830: loss 1.5994248390197754
 Step 1840: loss 1.7064387798309326
 Step 1850: loss 1.876413106918335
 Step 1860: loss 1.9288010597229004
 Step 1870: loss 1.7807573080062866
 Step 1880: loss 1.8370132446289062
 Step 1890: loss 1.7283363342285156
 Step 1900: loss 1.826354742050171
 Step 1910: loss 1.860719084739685
 Step 1920: loss 1.669980525970459
 Step 1930: loss 1.8051891326904297
 Step 1940: loss 1.7433419227600098
 Step 1950: loss 1.7696365118026733
 Step 1960: loss 1.7002618312835693
 Step 1970: loss 1.8848459720611572
 Step 1980: loss 1.7559919357299805
 Step 1990: loss 1.908553123474121
 Step 2000: loss 1.8232616186141968
 [Seed 94] Step 2000 --> val loss 1.8975162506103516
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [21  5 50 50  1 58 43 50 50  1] ...
 First target sequence: [ 5 50 50  1 58 43 50 50  1 63] ...

 Decoded input: I'll tell you news indifferent good for either.
 Here is a gentle

 Decoded target: 'll tell you news indifferent good for either.
 Here is a gentlem
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.210166931152344
 [Seed 95] Step 0 --> val loss 4.186617374420166
 Step 10: loss 3.8663644790649414
 Step 20: loss 3.321328639984131
 Step 30: loss 2.8759372234344482
 Step 40: loss 2.712019920349121
 Step 50: loss 2.700369358062744
 Step 60: loss 2.496060371398926
 Step 70: loss 2.4417853355407715
 Step 80: loss 2.409116506576538
 Step 90: loss 2.328205108642578
 Step 100: loss 2.3175151348114014
 Step 110: loss 2.2580325603485107
 Step 120: loss 2.356428623199463
 Step 130: loss 2.3021421432495117
 Step 140: loss 2.202930450439453
 Step 150: loss 2.0916175842285156
 Step 160: loss 2.2461087703704834
 Step 170: loss 2.1763267517089844
 Step 180: loss 2.1902732849121094
 Step 190: loss 2.115725040435791
 Step 200: loss 2.212024211883545
 [Seed 95] Step 200 --> val loss 2.212897539138794
 Step 210: loss 2.0335471630096436
 Step 220: loss 2.1057043075561523
 Step 230: loss 2.1098294258117676
 Step 240: loss 2.0176196098327637
 Step 250: loss 2.071167469024658
 Step 260: loss 1.9691550731658936
 Step 270: loss 1.9765937328338623
 Step 280: loss 2.130849599838257
 Step 290: loss 2.1315486431121826
 Step 300: loss 2.0972604751586914
 Step 310: loss 2.0401060581207275
 Step 320: loss 2.0205225944519043
 Step 330: loss 2.0330092906951904
 Step 340: loss 2.033114433288574
 Step 350: loss 1.9246177673339844
 Step 360: loss 2.033499002456665
 Step 370: loss 1.956716537475586
 Step 380: loss 1.9440802335739136
 Step 390: loss 1.838998794555664
 Step 400: loss 2.0890910625457764
 [Seed 95] Step 400 --> val loss 2.2164981365203857
 Step 410: loss 2.0341782569885254
 Step 420: loss 1.8660569190979004
 Step 430: loss 2.0172250270843506
 Step 440: loss 2.172483444213867
 Step 450: loss 1.9579353332519531
 Step 460: loss 1.9700044393539429
 Step 470: loss 2.0664234161376953
 Step 480: loss 2.0368027687072754
 Step 490: loss 1.966437816619873
 Step 500: loss 2.0302162170410156
 Step 510: loss 1.9374288320541382
 Step 520: loss 2.040243625640869
 Step 530: loss 2.0832018852233887
 Step 540: loss 1.906826376914978
 Step 550: loss 2.0438897609710693
 Step 560: loss 1.9515235424041748
 Step 570: loss 2.042302131652832
 Step 580: loss 2.0549402236938477
 Step 590: loss 2.006739616394043
 Step 600: loss 1.918378233909607
 [Seed 95] Step 600 --> val loss 2.2052500247955322
 Step 610: loss 1.8611886501312256
 Step 620: loss 1.9518537521362305
 Step 630: loss 1.9385802745819092
 Step 640: loss 1.9963107109069824
 Step 650: loss 2.0465681552886963
 Step 660: loss 2.0598418712615967
 Step 670: loss 2.0903844833374023
 Step 680: loss 1.9604848623275757
 Step 690: loss 1.9597346782684326
 Step 700: loss 1.9788917303085327
 Step 710: loss 1.8751376867294312
 Step 720: loss 1.9648089408874512
 Step 730: loss 1.88897705078125
 Step 740: loss 2.0788755416870117
 Step 750: loss 2.0087811946868896
 Step 760: loss 2.007267951965332
 Step 770: loss 2.073728561401367
 Step 780: loss 1.9305713176727295
 Step 790: loss 2.023059368133545
 Step 800: loss 1.924651026725769
 [Seed 95] Step 800 --> val loss 2.2333786487579346
 Step 810: loss 2.006894826889038
 Step 820: loss 2.0497987270355225
 Step 830: loss 2.1005096435546875
 Step 840: loss 2.053778648376465
 Step 850: loss 1.86142098903656
 Step 860: loss 2.0715272426605225
 Step 870: loss 1.87205970287323
 Step 880: loss 1.9949630498886108
 Step 890: loss 1.979041576385498
 Step 900: loss 1.9572066068649292
 Step 910: loss 2.118032217025757
 Step 920: loss 1.9941022396087646
 Step 930: loss 2.000635862350464
 Step 940: loss 2.1451210975646973
 Step 950: loss 1.959230661392212
 Step 960: loss 1.9489903450012207
 Step 970: loss 1.9796819686889648
 Step 980: loss 1.909872055053711
 Step 990: loss 1.962660551071167
 Step 1000: loss 1.839548945426941
 [Seed 95] Step 1000 --> val loss 2.1731557846069336
 Step 1010: loss 1.8349590301513672
 Step 1020: loss 2.0806262493133545
 Step 1030: loss 1.9463005065917969
 Step 1040: loss 1.8864487409591675
 Step 1050: loss 2.0575978755950928
 Step 1060: loss 1.934206485748291
 Step 1070: loss 1.8661162853240967
 Step 1080: loss 1.9694499969482422
 Step 1090: loss 1.8266429901123047
 Step 1100: loss 1.9082391262054443
 Step 1110: loss 1.9097875356674194
 Step 1120: loss 1.9351609945297241
 Step 1130: loss 1.8352168798446655
 Step 1140: loss 1.9112927913665771
 Step 1150: loss 1.9392859935760498
 Step 1160: loss 1.882310152053833
 Step 1170: loss 1.8860576152801514
 Step 1180: loss 1.8686892986297607
 Step 1190: loss 1.9074370861053467
 Step 1200: loss 1.9002293348312378
 [Seed 95] Step 1200 --> val loss 2.127829074859619
 Step 1210: loss 1.9099667072296143
 Step 1220: loss 1.8763421773910522
 Step 1230: loss 1.9639981985092163
 Step 1240: loss 1.801711082458496
 Step 1250: loss 1.8951568603515625
 Step 1260: loss 1.7792346477508545
 Step 1270: loss 1.851938009262085
 Step 1280: loss 1.8594568967819214
 Step 1290: loss 1.7554506063461304
 Step 1300: loss 1.9412152767181396
 Step 1310: loss 1.8063931465148926
 Step 1320: loss 1.8639061450958252
 Step 1330: loss 1.9124999046325684
 Step 1340: loss 1.8624128103256226
 Step 1350: loss 1.7567260265350342
 Step 1360: loss 1.8320815563201904
 Step 1370: loss 1.8058154582977295
 Step 1380: loss 1.8481690883636475
 Step 1390: loss 1.7711126804351807
 Step 1400: loss 1.9631578922271729
 [Seed 95] Step 1400 --> val loss 1.9367811679840088
 Step 1410: loss 1.6933894157409668
 Step 1420: loss 1.7423477172851562
 Step 1430: loss 1.8181670904159546
 Step 1440: loss 1.7790873050689697
 Step 1450: loss 1.7579944133758545
 Step 1460: loss 1.8952491283416748
 Step 1470: loss 1.8787769079208374
 Step 1480: loss 1.8320344686508179
 Step 1490: loss 1.7977159023284912
 Step 1500: loss 1.731541633605957
 Step 1510: loss 1.7454736232757568
 Step 1520: loss 1.7455748319625854
 Step 1530: loss 1.866398572921753
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 95] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 95] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 95] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [32 10  0 35 46 39 58  6  1 42] ...
 First target sequence: [10  0 35 46 39 58  6  1 42 53] ...

 Decoded input: T:
 What, dost thou scorn me for my gentle counsel?
 And soothe th

 Decoded target: :
 What, dost thou scorn me for my gentle counsel?
 And soothe the
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.200660228729248
 [Seed 96] Step 0 --> val loss 4.186720371246338
 Step 10: loss 3.849529981613159
 Step 20: loss 3.385849952697754
 Step 30: loss 2.9828498363494873
 Step 40: loss 2.6674556732177734
 Step 50: loss 2.6078333854675293
 Step 60: loss 2.442502975463867
 Step 70: loss 2.382772922515869
 Step 80: loss 2.2737157344818115
 Step 90: loss 2.347655773162842
 Step 100: loss 2.395373821258545
 Step 110: loss 2.2485570907592773
 Step 120: loss 2.2353391647338867
 Step 130: loss 2.2065043449401855
 Step 140: loss 2.26361083984375
 Step 150: loss 2.1200366020202637
 Step 160: loss 2.123183250427246
 Step 170: loss 2.169437885284424
 Step 180: loss 2.180903911590576
 Step 190: loss 2.0480499267578125
 Step 200: loss 2.0850210189819336
 [Seed 96] Step 200 --> val loss 2.3247272968292236
 Step 210: loss 2.1535823345184326
 Step 220: loss 2.054594039916992
 Step 230: loss 2.0777041912078857
 Step 240: loss 2.0371627807617188
 Step 250: loss 1.988249659538269
 Step 260: loss 2.090421438217163
 Step 270: loss 2.127628803253174
 Step 280: loss 2.0181148052215576
 Step 290: loss 2.02170991897583
 Step 300: loss 2.050739288330078
 Step 310: loss 2.095470905303955
 Step 320: loss 2.049635887145996
 Step 330: loss 2.101592540740967
 Step 340: loss 2.070821762084961
 Step 350: loss 2.0920658111572266
 Step 360: loss 2.0235610008239746
 Step 370: loss 2.0816245079040527
 Step 380: loss 2.142390251159668
 Step 390: loss 2.0482585430145264
 Step 400: loss 2.008817195892334
 [Seed 96] Step 400 --> val loss 2.2276618480682373
 Step 410: loss 2.070610284805298
 Step 420: loss 2.0166141986846924
 Step 430: loss 1.9819563627243042
 Step 440: loss 2.0165512561798096
 Step 450: loss 1.969985008239746
 Step 460: loss 1.9521210193634033
 Step 470: loss 1.9533822536468506
 Step 480: loss 1.9563664197921753
 Step 490: loss 2.0166897773742676
 Step 500: loss 2.071974515914917
 Step 510: loss 2.073199987411499
 Step 520: loss 1.9615968465805054
 Step 530: loss 2.106198787689209
 Step 540: loss 1.9128823280334473
 Step 550: loss 2.0201640129089355
 Step 560: loss 1.9950790405273438
 Step 570: loss 2.0390429496765137
 Step 580: loss 1.9328044652938843
 Step 590: loss 1.9536082744598389
 Step 600: loss 1.9301365613937378
 [Seed 96] Step 600 --> val loss 2.189206600189209
 Step 610: loss 1.9062070846557617
 Step 620: loss 1.9744212627410889
 Step 630: loss 1.9599330425262451
 Step 640: loss 2.004937171936035
 Step 650: loss 1.9963536262512207
 Step 660: loss 1.8857715129852295
 Step 670: loss 2.020561695098877
 Step 680: loss 1.9367121458053589
 Step 690: loss 2.061469316482544
 Step 700: loss 2.043138265609741
 Step 710: loss 1.9918159246444702
 Step 720: loss 1.9510741233825684
 Step 730: loss 1.8753451108932495
 Step 740: loss 1.9940218925476074
 Step 750: loss 1.954803228378296
 Step 760: loss 1.9341415166854858
 Step 770: loss 2.0866150856018066
 Step 780: loss 1.956256628036499
 Step 790: loss 1.9559314250946045
 Step 800: loss 1.9154398441314697
 [Seed 96] Step 800 --> val loss 2.264270544052124
 Step 810: loss 1.9627851247787476
 Step 820: loss 1.8464183807373047
 Step 830: loss 1.9515559673309326
 Step 840: loss 2.0167908668518066
 Step 850: loss 1.9915180206298828
 Step 860: loss 1.9940842390060425
 Step 870: loss 1.936036467552185
 Step 880: loss 1.9563519954681396
 Step 890: loss 2.0637705326080322
 Step 900: loss 2.0951812267303467
 Step 910: loss 1.9679464101791382
 Step 920: loss 2.0167291164398193
 Step 930: loss 2.0246198177337646
 Step 940: loss 1.9949617385864258
 Step 950: loss 1.925649881362915
 Step 960: loss 1.9507126808166504
 Step 970: loss 2.038297653198242
 Step 980: loss 1.917668104171753
 Step 990: loss 1.9868682622909546
 Step 1000: loss 1.8991153240203857
 [Seed 96] Step 1000 --> val loss 2.190521240234375
 Step 1010: loss 2.0321719646453857
 Step 1020: loss 1.9203609228134155
 Step 1030: loss 1.9614017009735107
 Step 1040: loss 1.7747629880905151
 Step 1050: loss 2.003960371017456
 Step 1060: loss 1.9892467260360718
 Step 1070: loss 1.8530076742172241
 Step 1080: loss 1.999100923538208
 Step 1090: loss 1.8115386962890625
 Step 1100: loss 1.8841575384140015
 Step 1110: loss 1.9390754699707031
 Step 1120: loss 1.979130744934082
 Step 1130: loss 1.8652633428573608
 Step 1140: loss 1.9298919439315796
 Step 1150: loss 1.7518585920333862
 Step 1160: loss 1.8777053356170654
 Step 1170: loss 1.8636735677719116
 Step 1180: loss 1.8054020404815674
 Step 1190: loss 1.945220708847046
 Step 1200: loss 1.9020612239837646
 [Seed 96] Step 1200 --> val loss 2.1405646800994873
 Step 1210: loss 1.9818549156188965
 Step 1220: loss 1.8600209951400757
 Step 1230: loss 1.9646940231323242
 Step 1240: loss 1.9272029399871826
 Step 1250: loss 1.9633439779281616
 Step 1260: loss 1.8600413799285889
 Step 1270: loss 1.9036887884140015
 Step 1280: loss 1.875390887260437
 Step 1290: loss 1.8968747854232788
 Step 1300: loss 1.980512022972107
 Step 1310: loss 1.9049959182739258
 Step 1320: loss 1.9929136037826538
 Step 1330: loss 1.8165616989135742
 Step 1340: loss 1.7685297727584839
 Step 1350: loss 1.9028165340423584
 Step 1360: loss 1.958606243133545
 Step 1370: loss 1.7840474843978882
 Step 1380: loss 1.9063398838043213
 Step 1390: loss 1.9299907684326172
 Step 1400: loss 1.846698522567749
 [Seed 96] Step 1400 --> val loss 2.0423781871795654
 Step 1410: loss 1.8845326900482178
 Step 1420: loss 1.8092588186264038
 Step 1430: loss 1.8477824926376343
 Step 1440: loss 1.85141921043396
 Step 1450: loss 1.805265188217163
 Step 1460: loss 1.7814722061157227
 Step 1470: loss 1.9010722637176514
 Step 1480: loss 1.894106149673462
 Step 1490: loss 1.9034456014633179
 Step 1500: loss 1.7305066585540771
 Step 1510: loss 1.897321105003357
 Step 1520: loss 1.820788860321045
 Step 1530: loss 1.7951622009277344
 Step 1540: loss 1.8183810710906982
 Step 1550: loss 1.8188340663909912
 Step 1560: loss 1.7727793455123901
 Step 1570: loss 1.7041501998901367
 Step 1580: loss 1.859292984008789
 Step 1590: loss 1.9144947528839111
 Step 1600: loss 1.8613293170928955
 [Seed 96] Step 1600 --> val loss 1.9030132293701172
 Step 1610: loss 1.7190698385238647
 Step 1620: loss 1.7474870681762695
 Step 1630: loss 1.7965114116668701
 Step 1640: loss 1.7026857137680054
 Step 1650: loss 1.785070776939392
 Step 1660: loss 1.8309507369995117
 Step 1670: loss 1.6462657451629639
 Step 1680: loss 1.8185288906097412
 Step 1690: loss 1.7343705892562866
 Step 1700: loss 1.8343620300292969
 Step 1710: loss 1.7658547163009644
 Step 1720: loss 1.68336820602417
 Step 1730: loss 1.7087898254394531
 Step 1740: loss 1.6195762157440186
 Step 1750: loss 1.5917954444885254
 Step 1760: loss 1.7145260572433472
 Step 1770: loss 1.7404555082321167
 Step 1780: loss 1.7846708297729492
 Step 1790: loss 1.7545679807662964
 Step 1800: loss 1.7468018531799316
 [Seed 96] Step 1800 --> val loss 1.8275173902511597
 Step 1810: loss 1.809969186782837
 Step 1820: loss 1.745018720626831
 Step 1830: loss 1.727360725402832
 Step 1840: loss 1.6853513717651367
 Step 1850: loss 1.6841089725494385
 Step 1860: loss 1.789790153503418
 Step 1870: loss 1.8139512538909912
 Step 1880: loss 1.7196376323699951
 Step 1890: loss 1.712446928024292
 Step 1900: loss 1.6829627752304077
 Step 1910: loss 1.7278857231140137
 Step 1920: loss 1.727701187133789
 Step 1930: loss 1.5938833951950073
 Step 1940: loss 1.6093223094940186
 Step 1950: loss 1.686253547668457
 Step 1960: loss 1.6466245651245117
 Step 1970: loss 1.6916555166244507
 Step 1980: loss 1.6587623357772827
 Step 1990: loss 1.693638563156128
 Step 2000: loss 1.6754250526428223
 [Seed 96] Step 2000 --> val loss 1.759362816810608
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [51 43 52  1 51 59 57 58  1 52] ...
 First target sequence: [43 52  1 51 59 57 58  1 52 43] ...

 Decoded input: men must needs abide;
 It boots not to resist both wind and tide.

 Decoded target: en must needs abide;
 It boots not to resist both wind and tide.

 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.212769508361816
 [Seed 97] Step 0 --> val loss 4.182321071624756
 Step 10: loss 3.83485746383667
 Step 20: loss 3.3591251373291016
 Step 30: loss 2.790229320526123
 Step 40: loss 2.6366848945617676
 Step 50: loss 2.619933605194092
 Step 60: loss 2.504088878631592
 Step 70: loss 2.5591278076171875
 Step 80: loss 2.340463638305664
 Step 90: loss 2.3494527339935303
 Step 100: loss 2.2437973022460938
 Step 110: loss 2.2342171669006348
 Step 120: loss 2.1771016120910645
 Step 130: loss 2.2815358638763428
 Step 140: loss 2.058685541152954
 Step 150: loss 2.1144332885742188
 Step 160: loss 2.161381483078003
 Step 170: loss 2.1700119972229004
 Step 180: loss 2.0770421028137207
 Step 190: loss 2.044410228729248
 Step 200: loss 2.1526894569396973
 [Seed 97] Step 200 --> val loss 2.2617347240448
 Step 210: loss 2.084763526916504
 Step 220: loss 2.2076797485351562
 Step 230: loss 2.1369614601135254
 Step 240: loss 2.169654130935669
 Step 250: loss 2.052018404006958
 Step 260: loss 1.967733383178711
 Step 270: loss 2.0023837089538574
 Step 280: loss 2.043745279312134
 Step 290: loss 1.9883016347885132
 Step 300: loss 2.149862289428711
 Step 310: loss 2.0817348957061768
 Step 320: loss 2.047548770904541
 Step 330: loss 1.97745680809021
 Step 340: loss 2.0093331336975098
 Step 350: loss 1.9889888763427734
 Step 360: loss 2.0229077339172363
 Step 370: loss 2.087850332260132
 Step 380: loss 1.9876880645751953
 Step 390: loss 1.941340684890747
 Step 400: loss 2.0125396251678467
 [Seed 97] Step 400 --> val loss 2.1931114196777344
 Step 410: loss 2.087123155593872
 Step 420: loss 2.021047353744507
 Step 430: loss 2.035924196243286
 Step 440: loss 1.9992072582244873
 Step 450: loss 2.0115668773651123
 Step 460: loss 1.9973371028900146
 Step 470: loss 1.935976266860962
 Step 480: loss 2.086073637008667
 Step 490: loss 2.016519069671631
 Step 500: loss 1.9937185049057007
 Step 510: loss 1.9832013845443726
 Step 520: loss 1.944655179977417
 Step 530: loss 2.0759902000427246
 Step 540: loss 1.9780545234680176
 Step 550: loss 2.008650064468384
 Step 560: loss 1.9650567770004272
 Step 570: loss 1.915686011314392
 Step 580: loss 1.9869321584701538
 Step 590: loss 2.0617423057556152
 Step 600: loss 2.028106212615967
 [Seed 97] Step 600 --> val loss 2.2573602199554443
 Step 610: loss 1.9809949398040771
 Step 620: loss 2.0960402488708496
 Step 630: loss 1.9727600812911987
 Step 640: loss 1.9747294187545776
 Step 650: loss 1.9701579809188843
 Step 660: loss 2.1061348915100098
 Step 670: loss 2.003044843673706
 Step 680: loss 1.8844497203826904
 Step 690: loss 1.9448394775390625
 Step 700: loss 1.9541953802108765
 Step 710: loss 2.0416746139526367
 Step 720: loss 2.0256991386413574
 Step 730: loss 1.9636037349700928
 Step 740: loss 1.8990329504013062
 Step 750: loss 2.059065580368042
 Step 760: loss 1.9853670597076416
 Step 770: loss 1.9426095485687256
 Step 780: loss 2.146148681640625
 Step 790: loss 1.9366600513458252
 Step 800: loss 2.093496799468994
 [Seed 97] Step 800 --> val loss 2.1934010982513428
 Step 810: loss 2.0913288593292236
 Step 820: loss 1.9607913494110107
 Step 830: loss 2.0381920337677
 Step 840: loss 2.0202503204345703
 Step 850: loss 2.0164856910705566
 Step 860: loss 1.8941600322723389
 Step 870: loss 2.0473716259002686
 Step 880: loss 1.9772428274154663
 Step 890: loss 1.9904950857162476
 Step 900: loss 2.0058343410491943
 Step 910: loss 2.049816131591797
 Step 920: loss 2.0067176818847656
 Step 930: loss 1.9485958814620972
 Step 940: loss 2.1493685245513916
 Step 950: loss 1.9571073055267334
 Step 960: loss 2.0089402198791504
 Step 970: loss 1.9460899829864502
 Step 980: loss 2.088862895965576
 Step 990: loss 1.9822949171066284
 Step 1000: loss 1.8569927215576172
 [Seed 97] Step 1000 --> val loss 2.134333372116089
 Step 1010: loss 1.9737129211425781
 Step 1020: loss 1.9929921627044678
 Step 1030: loss 2.035386323928833
 Step 1040: loss 1.978431224822998
 Step 1050: loss 2.056206226348877
 Step 1060: loss 2.025811195373535
 Step 1070: loss 1.907038688659668
 Step 1080: loss 1.9313244819641113
 Step 1090: loss 2.0536537170410156
 Step 1100: loss 2.0237247943878174
 Step 1110: loss 1.9034748077392578
 Step 1120: loss 2.086033821105957
 Step 1130: loss 1.8892310857772827
 Step 1140: loss 2.0142335891723633
 Step 1150: loss 2.0106000900268555
 Step 1160: loss 2.040745496749878
 Step 1170: loss 1.96803617477417
 Step 1180: loss 1.9879875183105469
 Step 1190: loss 1.9876635074615479
 Step 1200: loss 2.0694942474365234
 [Seed 97] Step 1200 --> val loss 2.1838929653167725
 Step 1210: loss 2.0320611000061035
 Step 1220: loss 1.9718083143234253
 Step 1230: loss 1.913862705230713
 Step 1240: loss 1.9343584775924683
 Step 1250: loss 1.8839895725250244
 Step 1260: loss 1.9942774772644043
 Step 1270: loss 1.9749915599822998
 Step 1280: loss 1.9424455165863037
 Step 1290: loss 1.9907376766204834
 Step 1300: loss 2.0587501525878906
 Step 1310: loss 2.0946130752563477
 Step 1320: loss 1.8137798309326172
 Step 1330: loss 1.9247214794158936
 Step 1340: loss 1.916783332824707
 Step 1350: loss 1.9782781600952148
 Step 1360: loss 2.07847261428833
 Step 1370: loss 2.0434048175811768
 Step 1380: loss 1.895404577255249
 Step 1390: loss 2.0267317295074463
 Step 1400: loss 1.957815170288086
 [Seed 97] Step 1400 --> val loss 2.179112195968628
 Step 1410: loss 1.8376058340072632
 Step 1420: loss 1.9458717107772827
 Step 1430: loss 1.9227924346923828
 Step 1440: loss 1.9717886447906494
 Step 1450: loss 1.8536171913146973
 Step 1460: loss 1.9474371671676636
 Step 1470: loss 2.0801761150360107
 Step 1480: loss 1.822113037109375
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 97] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 97] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 97] Step 2000 --> val loss nan
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [ 0 13 47 51  5 42  1 39 58  1] ...
 First target sequence: [13 47 51  5 42  1 39 58  1 63] ...

 Decoded input: 
 Aim'd at your highness, no inveterate malice.

 KING RICHARD II:

 Decoded target: Aim'd at your highness, no inveterate malice.

 KING RICHARD II:

 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.213230133056641
 [Seed 98] Step 0 --> val loss 4.183986186981201
 Step 10: loss 3.840423345565796
 Step 20: loss 3.3104429244995117
 Step 30: loss 2.821627378463745
 Step 40: loss 2.852658748626709
 Step 50: loss 2.5509982109069824
 Step 60: loss 2.4679980278015137
 Step 70: loss 2.443173408508301
 Step 80: loss 2.447871208190918
 Step 90: loss 2.489252805709839
 Step 100: loss 2.298065185546875
 Step 110: loss 2.388298273086548
 Step 120: loss 2.198972702026367
 Step 130: loss 2.1675944328308105
 Step 140: loss 2.2269091606140137
 Step 150: loss 2.1788501739501953
 Step 160: loss 2.127188205718994
 Step 170: loss 2.171058416366577
 Step 180: loss 2.162597417831421
 Step 190: loss 2.1033363342285156
 Step 200: loss 2.151337146759033
 [Seed 98] Step 200 --> val loss 2.3519668579101562
 Step 210: loss 2.234185218811035
 Step 220: loss 2.175053358078003
 Step 230: loss 2.1287620067596436
 Step 240: loss 2.1038074493408203
 Step 250: loss 2.0763187408447266
 Step 260: loss 2.170588493347168
 Step 270: loss 2.0568997859954834
 Step 280: loss 2.1765143871307373
 Step 290: loss 2.1976523399353027
 Step 300: loss 2.0654354095458984
 Step 310: loss 2.1585328578948975
 Step 320: loss 1.985334873199463
 Step 330: loss 2.0932488441467285
 Step 340: loss 1.982285499572754
 Step 350: loss 2.1380538940429688
 Step 360: loss 1.8935925960540771
 Step 370: loss 2.0470526218414307
 Step 380: loss 2.0908942222595215
 Step 390: loss 2.0768134593963623
 Step 400: loss 2.0227468013763428
 [Seed 98] Step 400 --> val loss 2.2236855030059814
 Step 410: loss 2.061800718307495
 Step 420: loss 2.022869110107422
 Step 430: loss 1.9674098491668701
 Step 440: loss 1.9985811710357666
 Step 450: loss 1.9921565055847168
 Step 460: loss 2.1202926635742188
 Step 470: loss 2.0678718090057373
 Step 480: loss 2.0316150188446045
 Step 490: loss 1.998166799545288
 Step 500: loss 1.93875253200531
 Step 510: loss 1.9795541763305664
 Step 520: loss 1.9024884700775146
 Step 530: loss 1.9439090490341187
 Step 540: loss 1.9653117656707764
 Step 550: loss 2.0454859733581543
 Step 560: loss 1.8812955617904663
 Step 570: loss 2.072631597518921
 Step 580: loss 2.06392240524292
 Step 590: loss 2.080695629119873
 Step 600: loss 1.9220222234725952
 [Seed 98] Step 600 --> val loss 2.239474058151245
 Step 610: loss 2.0010125637054443
 Step 620: loss 1.9427293539047241
 Step 630: loss 1.8498958349227905
 Step 640: loss 1.9449023008346558
 Step 650: loss 1.9887025356292725
 Step 660: loss 2.122729778289795
 Step 670: loss 1.9629020690917969
 Step 680: loss 1.94510817527771
 Step 690: loss 1.9285573959350586
 Step 700: loss 1.8463062047958374
 Step 710: loss 1.9021801948547363
 Step 720: loss 1.932626485824585
 Step 730: loss 1.9424151182174683
 Step 740: loss 1.9518356323242188
 Step 750: loss 1.8821020126342773
 Step 760: loss 1.9896745681762695
 Step 770: loss 1.9868600368499756
 Step 780: loss 1.9812908172607422
 Step 790: loss 1.9977186918258667
 Step 800: loss 1.9641849994659424
 [Seed 98] Step 800 --> val loss 2.1832618713378906
 Step 810: loss 1.9509508609771729
 Step 820: loss 1.9105582237243652
 Step 830: loss 2.099355459213257
 Step 840: loss 1.992061972618103
 Step 850: loss 2.017991542816162
 Step 860: loss 1.970638632774353
 Step 870: loss 1.8906501531600952
 Step 880: loss 2.0715198516845703
 Step 890: loss 2.02488374710083
 Step 900: loss 1.9990270137786865
 Step 910: loss 2.067139148712158
 Step 920: loss 1.9860481023788452
 Step 930: loss 2.0327110290527344
 Step 940: loss 2.040360450744629
 Step 950: loss 1.976737380027771
 Step 960: loss 2.1373064517974854
 Step 970: loss 1.8806076049804688
 Step 980: loss 1.9322185516357422
 Step 990: loss 1.952075719833374
 Step 1000: loss 1.947998046875
 [Seed 98] Step 1000 --> val loss 2.2187817096710205
 Step 1010: loss 2.041714668273926
 Step 1020: loss 1.9618041515350342
 Step 1030: loss 1.9094980955123901
 Step 1040: loss 2.0505728721618652
 Step 1050: loss 1.9665601253509521
 Step 1060: loss 1.896468162536621
 Step 1070: loss 1.949220895767212
 Step 1080: loss 1.9369434118270874
 Step 1090: loss 2.028921604156494
 Step 1100: loss 1.9177793264389038
 Step 1110: loss 1.9155480861663818
 Step 1120: loss 1.9139375686645508
 Step 1130: loss 1.9805092811584473
 Step 1140: loss 1.9329942464828491
 Step 1150: loss 1.921465277671814
 Step 1160: loss 1.9259452819824219
 Step 1170: loss 2.0583128929138184
 Step 1180: loss 1.9767932891845703
 Step 1190: loss 1.9484307765960693
 Step 1200: loss 1.9746173620224
 [Seed 98] Step 1200 --> val loss 2.2670114040374756
 Step 1210: loss 1.9063905477523804
 Step 1220: loss 1.8929333686828613
 Step 1230: loss 2.018064498901367
 Step 1240: loss 2.0414323806762695
 Step 1250: loss 1.981603980064392
 Step 1260: loss 1.967193603515625
 Step 1270: loss 1.9455783367156982
 Step 1280: loss 2.077899217605591
 Step 1290: loss 1.8419151306152344
 Step 1300: loss 1.9805107116699219
 Step 1310: loss 1.940140724182129
 Step 1320: loss 1.8630660772323608
 Step 1330: loss 1.954025149345398
 Step 1340: loss 1.9198530912399292
 Step 1350: loss 1.9335505962371826
 Step 1360: loss 1.807035207748413
 Step 1370: loss 1.8536872863769531
 Step 1380: loss 1.9569981098175049
 Step 1390: loss 1.9392328262329102
 Step 1400: loss 1.8927539587020874
 [Seed 98] Step 1400 --> val loss 2.0241317749023438
 Step 1410: loss 1.8833742141723633
 Step 1420: loss 1.866213321685791
 Step 1430: loss 1.9716652631759644
 Step 1440: loss 1.9288389682769775
 Step 1450: loss 1.880656123161316
 Step 1460: loss 1.984440565109253
 Step 1470: loss 1.8823325634002686
 Step 1480: loss 1.9334046840667725
 Step 1490: loss 1.8814573287963867
 Step 1500: loss 1.75408935546875
 Step 1510: loss 1.8348664045333862
 Step 1520: loss 1.870984435081482
 Step 1530: loss 1.7588059902191162
 Step 1540: loss 1.9401888847351074
 Step 1550: loss 1.9982023239135742
 Step 1560: loss 1.7533044815063477
 Step 1570: loss 1.8980495929718018
 Step 1580: loss 1.7916969060897827
 Step 1590: loss 1.8106274604797363
 Step 1600: loss 1.7872238159179688
 [Seed 98] Step 1600 --> val loss 2.0639591217041016
 Step 1610: loss 1.8421800136566162
 Step 1620: loss 1.7863248586654663
 Step 1630: loss 1.9293359518051147
 Step 1640: loss 1.8199151754379272
 Step 1650: loss 1.8414701223373413
 Step 1660: loss 1.8188533782958984
 Step 1670: loss 1.8988933563232422
 Step 1680: loss 1.7510589361190796
 Step 1690: loss 1.778497576713562
 Step 1700: loss 1.6971547603607178
 Step 1710: loss 1.8210357427597046
 Step 1720: loss 1.918211817741394
 Step 1730: loss 1.8060355186462402
 Step 1740: loss 1.8450340032577515
 Step 1750: loss 1.7725448608398438
 Step 1760: loss 1.901421070098877
 Step 1770: loss 1.6362460851669312
 Step 1780: loss 1.6945688724517822
 Step 1790: loss 1.8631269931793213
 Step 1800: loss 1.7896993160247803
 [Seed 98] Step 1800 --> val loss 1.937252402305603
 Step 1810: loss 1.732086181640625
 Step 1820: loss 1.7292051315307617
 Step 1830: loss 1.8019752502441406
 Step 1840: loss 1.7544111013412476
 Step 1850: loss 1.708195686340332
 Step 1860: loss 1.6032886505126953
 Step 1870: loss 1.7077586650848389
 Step 1880: loss 1.784113883972168
 Step 1890: loss 1.667280912399292
 Step 1900: loss 1.7722413539886475
 Step 1910: loss 1.8222893476486206
 Step 1920: loss 1.7106969356536865
 Step 1930: loss 1.7740178108215332
 Step 1940: loss 1.6964085102081299
 Step 1950: loss 1.7207019329071045
 Step 1960: loss 1.7350705862045288
 Step 1970: loss 1.61151123046875
 Step 1980: loss 1.7772762775421143
 Step 1990: loss 1.7188129425048828
 Step 2000: loss 1.8184865713119507
 [Seed 98] Step 2000 --> val loss 1.8575009107589722
 ====================================================================================================
 Input shape: (12, 64)
 Target shape: (12, 64)
 First input sequence: [27 24 21 27 10  0 26 53 58  1] ...
 First target sequence: [24 21 27 10  0 26 53 58  1 58] ...

 Decoded input: OLIO:
 Not to his father's; I spoke with his man.

 MERCUTIO:
 Ah, 

 Decoded target: LIO:
 Not to his father's; I spoke with his man.

 MERCUTIO:
 Ah, t
 CompositeModule
 ...consists of 26 atoms and 78 bonds
 ...non-smooth
 ...input sensitivity is 1.0
 ...contributes proportion 7.0 to feature learning of any supermodule
 Step 0: loss 4.221299171447754
 [Seed 99] Step 0 --> val loss 4.182692050933838
 Step 10: loss 3.8563060760498047
 Step 20: loss 3.3032500743865967
 Step 30: loss 2.9216976165771484
 Step 40: loss 2.6124188899993896
 Step 50: loss 2.5297722816467285
 Step 60: loss 2.4535155296325684
 Step 70: loss 2.5314173698425293
 Step 80: loss 2.362807273864746
 Step 90: loss 2.249311685562134
 Step 100: loss 2.2996912002563477
 Step 110: loss 2.229684829711914
 Step 120: loss 2.2906391620635986
 Step 130: loss 2.1215784549713135
 Step 140: loss 2.1365151405334473
 Step 150: loss 2.2214887142181396
 Step 160: loss 2.2403059005737305
 Step 170: loss 2.261080741882324
 Step 180: loss 2.185957431793213
 Step 190: loss 2.1285104751586914
 Step 200: loss 2.159724473953247
 [Seed 99] Step 200 --> val loss 2.340670347213745
 Step 210: loss 2.1039938926696777
 Step 220: loss 1.9503962993621826
 Step 230: loss 2.024515151977539
 Step 240: loss 2.106186628341675
 Step 250: loss 2.109154224395752
 Step 260: loss 2.0865304470062256
 Step 270: loss 2.133700370788574
 Step 280: loss 2.1083762645721436
 Step 290: loss 1.9946722984313965
 Step 300: loss 2.163235902786255
 Step 310: loss 2.0190443992614746
 Step 320: loss 1.943387508392334
 Step 330: loss 2.113252639770508
 Step 340: loss 2.184530258178711
 Step 350: loss 2.0407118797302246
 Step 360: loss 2.0311059951782227
 Step 370: loss 2.124037504196167
 Step 380: loss 1.9909805059432983
 Step 390: loss 2.017343521118164
 Step 400: loss 2.070819616317749
 [Seed 99] Step 400 --> val loss 2.226104497909546
 Step 410: loss 1.9687575101852417
 Step 420: loss 1.9556245803833008
 Step 430: loss 2.0761990547180176
 Step 440: loss 2.0775673389434814
 Step 450: loss 1.9830427169799805
 Step 460: loss 1.9707082509994507
 Step 470: loss 1.91650390625
 Step 480: loss 1.9896639585494995
 Step 490: loss 2.0204415321350098
 Step 500: loss 2.0391950607299805
 Step 510: loss 2.043853998184204
 Step 520: loss 2.0781521797180176
 Step 530: loss 2.072216033935547
 Step 540: loss 2.013058662414551
 Step 550: loss 2.0530543327331543
 Step 560: loss 1.8811407089233398
 Step 570: loss 1.957275390625
 Step 580: loss 1.9315168857574463
 Step 590: loss 1.962998390197754
 Step 600: loss 2.044961929321289
 [Seed 99] Step 600 --> val loss 2.159186363220215
 Step 610: loss 1.9662368297576904
 Step 620: loss 1.9603664875030518
 Step 630: loss 1.9998149871826172
 Step 640: loss 2.006333351135254
 Step 650: loss 1.8796005249023438
 Step 660: loss 2.0002455711364746
 Step 670: loss 2.014359951019287
 Step 680: loss 2.0276801586151123
 Step 690: loss 1.9359323978424072
 Step 700: loss 1.9395173788070679
 Step 710: loss 2.095719575881958
 Step 720: loss 1.9366462230682373
 Step 730: loss 1.8924894332885742
 Step 740: loss 2.041733741760254
 Step 750: loss 1.884415626525879
 Step 760: loss 1.9391998052597046
 Step 770: loss 1.937268614768982
 Step 780: loss 1.9632558822631836
 Step 790: loss 1.9253318309783936
 Step 800: loss 2.0495150089263916
 [Seed 99] Step 800 --> val loss 2.1791810989379883
 Step 810: loss 2.0498032569885254
 Step 820: loss 1.9190664291381836
 Step 830: loss 1.970568060874939
 Step 840: loss 2.149332284927368
 Step 850: loss 2.0298264026641846
 Step 860: loss 1.9861726760864258
 Step 870: loss 1.8677678108215332
 Step 880: loss 1.9365386962890625
 Step 890: loss 2.073042154312134
 Step 900: loss 2.0618515014648438
 Step 910: loss 1.9894322156906128
 Step 920: loss 1.944661021232605
 Step 930: loss 2.020565986633301
 Step 940: loss 1.9941753149032593
 Step 950: loss 1.8899967670440674
 Step 960: loss 1.9023692607879639
 Step 970: loss 1.9693782329559326
 Step 980: loss 1.947614073753357
 Step 990: loss 2.0031774044036865
 Step 1000: loss 2.0169761180877686
 [Seed 99] Step 1000 --> val loss 2.238865613937378
 Step 1010: loss nan
 Step 1020: loss nan
 Step 1030: loss nan
 Step 1040: loss nan
 Step 1050: loss nan
 Step 1060: loss nan
 Step 1070: loss nan
 Step 1080: loss nan
 Step 1090: loss nan
 Step 1100: loss nan
 Step 1110: loss nan
 Step 1120: loss nan
 Step 1130: loss nan
 Step 1140: loss nan
 Step 1150: loss nan
 Step 1160: loss nan
 Step 1170: loss nan
 Step 1180: loss nan
 Step 1190: loss nan
 Step 1200: loss nan
 [Seed 99] Step 1200 --> val loss nan
 Step 1210: loss nan
 Step 1220: loss nan
 Step 1230: loss nan
 Step 1240: loss nan
 Step 1250: loss nan
 Step 1260: loss nan
 Step 1270: loss nan
 Step 1280: loss nan
 Step 1290: loss nan
 Step 1300: loss nan
 Step 1310: loss nan
 Step 1320: loss nan
 Step 1330: loss nan
 Step 1340: loss nan
 Step 1350: loss nan
 Step 1360: loss nan
 Step 1370: loss nan
 Step 1380: loss nan
 Step 1390: loss nan
 Step 1400: loss nan
 [Seed 99] Step 1400 --> val loss nan
 Step 1410: loss nan
 Step 1420: loss nan
 Step 1430: loss nan
 Step 1440: loss nan
 Step 1450: loss nan
 Step 1460: loss nan
 Step 1470: loss nan
 Step 1480: loss nan
 Step 1490: loss nan
 Step 1500: loss nan
 Step 1510: loss nan
 Step 1520: loss nan
 Step 1530: loss nan
 Step 1540: loss nan
 Step 1550: loss nan
 Step 1560: loss nan
 Step 1570: loss nan
 Step 1580: loss nan
 Step 1590: loss nan
 Step 1600: loss nan
 [Seed 99] Step 1600 --> val loss nan
 Step 1610: loss nan
 Step 1620: loss nan
 Step 1630: loss nan
 Step 1640: loss nan
 Step 1650: loss nan
 Step 1660: loss nan
 Step 1670: loss nan
 Step 1680: loss nan
 Step 1690: loss nan
 Step 1700: loss nan
 Step 1710: loss nan
 Step 1720: loss nan
 Step 1730: loss nan
 Step 1740: loss nan
 Step 1750: loss nan
 Step 1760: loss nan
 Step 1770: loss nan
 Step 1780: loss nan
 Step 1790: loss nan
 Step 1800: loss nan
 [Seed 99] Step 1800 --> val loss nan
 Step 1810: loss nan
 Step 1820: loss nan
 Step 1830: loss nan
 Step 1840: loss nan
 Step 1850: loss nan
 Step 1860: loss nan
 Step 1870: loss nan
 Step 1880: loss nan
 Step 1890: loss nan
 Step 1900: loss nan
 Step 1910: loss nan
 Step 1920: loss nan
 Step 1930: loss nan
 Step 1940: loss nan
 Step 1950: loss nan
 Step 1960: loss nan
 Step 1970: loss nan
 Step 1980: loss nan
 Step 1990: loss nan
 Step 2000: loss nan
 [Seed 99] Step 2000 --> val loss nan
 ====================================================================================================