{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.0625,
  "eval_steps": 250,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.000125,
      "grad_norm": 2.4410972595214844,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 1.3861,
      "loss/crossentropy": 2.3799004554748535,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.20445775985717773,
      "loss/reg": 0.00019296666141599417,
      "step": 1
    },
    {
      "epoch": 0.00025,
      "grad_norm": 5.245810508728027,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 1.8002,
      "loss/crossentropy": 3.2071006298065186,
      "loss/hidden": 1.5234375,
      "loss/logits": 0.27488040924072266,
      "loss/reg": 0.00019296666141599417,
      "step": 2
    },
    {
      "epoch": 0.000375,
      "grad_norm": 2.093414545059204,
      "learning_rate": 3e-06,
      "loss": 1.1295,
      "loss/crossentropy": 2.474642753601074,
      "loss/hidden": 0.98046875,
      "loss/logits": 0.14706741273403168,
      "loss/reg": 0.00019296648679301143,
      "step": 3
    },
    {
      "epoch": 0.0005,
      "grad_norm": 2.3313233852386475,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.1829,
      "loss/crossentropy": 2.6595563888549805,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.15751230716705322,
      "loss/reg": 0.00019296605023555458,
      "step": 4
    },
    {
      "epoch": 0.000625,
      "grad_norm": 2.8557937145233154,
      "learning_rate": 5e-06,
      "loss": 1.2347,
      "loss/crossentropy": 2.4212286472320557,
      "loss/hidden": 1.0546875,
      "loss/logits": 0.1781189739704132,
      "loss/reg": 0.00019296558457426727,
      "step": 5
    },
    {
      "epoch": 0.00075,
      "grad_norm": 3.0438766479492188,
      "learning_rate": 6e-06,
      "loss": 1.3852,
      "loss/crossentropy": 2.5387792587280273,
      "loss/hidden": 1.1953125,
      "loss/logits": 0.1879885196685791,
      "loss/reg": 0.00019296500249765813,
      "step": 6
    },
    {
      "epoch": 0.000875,
      "grad_norm": 2.6279566287994385,
      "learning_rate": 7.000000000000001e-06,
      "loss": 1.3298,
      "loss/crossentropy": 2.4866859912872314,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.17939803004264832,
      "loss/reg": 0.00019296453683637083,
      "step": 7
    },
    {
      "epoch": 0.001,
      "grad_norm": 4.1743645668029785,
      "learning_rate": 8.000000000000001e-06,
      "loss": 1.3617,
      "loss/crossentropy": 2.6065773963928223,
      "loss/hidden": 1.15625,
      "loss/logits": 0.20351174473762512,
      "loss/reg": 0.00019296388200018555,
      "step": 8
    },
    {
      "epoch": 0.001125,
      "grad_norm": 2.4077372550964355,
      "learning_rate": 9e-06,
      "loss": 1.3601,
      "loss/crossentropy": 2.5489792823791504,
      "loss/hidden": 1.15625,
      "loss/logits": 0.20188993215560913,
      "loss/reg": 0.00019296332902740687,
      "step": 9
    },
    {
      "epoch": 0.00125,
      "grad_norm": 2.785814046859741,
      "learning_rate": 1e-05,
      "loss": 1.3588,
      "loss/crossentropy": 2.892237424850464,
      "loss/hidden": 1.1640625,
      "loss/logits": 0.19282634556293488,
      "loss/reg": 0.0001929624268086627,
      "step": 10
    },
    {
      "epoch": 0.001375,
      "grad_norm": 2.223435163497925,
      "learning_rate": 1.1000000000000001e-05,
      "loss": 1.3473,
      "loss/crossentropy": 2.6144230365753174,
      "loss/hidden": 1.140625,
      "loss/logits": 0.20472508668899536,
      "loss/reg": 0.0001929616992129013,
      "step": 11
    },
    {
      "epoch": 0.0015,
      "grad_norm": 2.2602744102478027,
      "learning_rate": 1.2e-05,
      "loss": 1.4537,
      "loss/crossentropy": 2.445138931274414,
      "loss/hidden": 1.234375,
      "loss/logits": 0.21741002798080444,
      "loss/reg": 0.0001929609279613942,
      "step": 12
    },
    {
      "epoch": 0.001625,
      "grad_norm": 2.167941093444824,
      "learning_rate": 1.3000000000000001e-05,
      "loss": 1.3285,
      "loss/crossentropy": 2.4828319549560547,
      "loss/hidden": 1.140625,
      "loss/logits": 0.18590743839740753,
      "loss/reg": 0.00019296009850222617,
      "step": 13
    },
    {
      "epoch": 0.00175,
      "grad_norm": 3.135204315185547,
      "learning_rate": 1.4000000000000001e-05,
      "loss": 1.2881,
      "loss/crossentropy": 2.485008955001831,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.16897538304328918,
      "loss/reg": 0.00019296004029456526,
      "step": 14
    },
    {
      "epoch": 0.001875,
      "grad_norm": 1.9516435861587524,
      "learning_rate": 1.5e-05,
      "loss": 1.3793,
      "loss/crossentropy": 2.6037957668304443,
      "loss/hidden": 1.15625,
      "loss/logits": 0.22108706831932068,
      "loss/reg": 0.0001929590798681602,
      "step": 15
    },
    {
      "epoch": 0.002,
      "grad_norm": 2.4786994457244873,
      "grad_norm_var": 0.7268455688420417,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 1.4938,
      "loss/crossentropy": 2.6338419914245605,
      "loss/hidden": 1.2578125,
      "loss/logits": 0.23409520089626312,
      "loss/reg": 0.00019295798847451806,
      "step": 16
    },
    {
      "epoch": 0.002125,
      "grad_norm": 1.935854196548462,
      "grad_norm_var": 0.7645541886139611,
      "learning_rate": 1.7000000000000003e-05,
      "loss": 1.1905,
      "loss/crossentropy": 2.5458219051361084,
      "loss/hidden": 1.015625,
      "loss/logits": 0.17296427488327026,
      "loss/reg": 0.00019295624224469066,
      "step": 17
    },
    {
      "epoch": 0.00225,
      "grad_norm": 4.175808906555176,
      "grad_norm_var": 0.4775368463766857,
      "learning_rate": 1.8e-05,
      "loss": 1.4189,
      "loss/crossentropy": 2.745898485183716,
      "loss/hidden": 1.21875,
      "loss/logits": 0.1982022523880005,
      "loss/reg": 0.00019295603851787746,
      "step": 18
    },
    {
      "epoch": 0.002375,
      "grad_norm": 2.6384661197662354,
      "grad_norm_var": 0.45452375883730595,
      "learning_rate": 1.9e-05,
      "loss": 1.546,
      "loss/crossentropy": 2.5159096717834473,
      "loss/hidden": 1.3203125,
      "loss/logits": 0.22378988564014435,
      "loss/reg": 0.00019295531092211604,
      "step": 19
    },
    {
      "epoch": 0.0025,
      "grad_norm": 1.930158019065857,
      "grad_norm_var": 0.48428273913252134,
      "learning_rate": 2e-05,
      "loss": 1.1151,
      "loss/crossentropy": 2.6261353492736816,
      "loss/hidden": 0.9609375,
      "loss/logits": 0.1521872580051422,
      "loss/reg": 0.00019295411766506732,
      "step": 20
    },
    {
      "epoch": 0.002625,
      "grad_norm": 2.75396728515625,
      "grad_norm_var": 0.48247025151936307,
      "learning_rate": 2.1e-05,
      "loss": 1.4758,
      "loss/crossentropy": 2.448762893676758,
      "loss/hidden": 1.2578125,
      "loss/logits": 0.21609561145305634,
      "loss/reg": 0.00019295368110761046,
      "step": 21
    },
    {
      "epoch": 0.00275,
      "grad_norm": 2.8601644039154053,
      "grad_norm_var": 0.475377454219718,
      "learning_rate": 2.2000000000000003e-05,
      "loss": 1.5402,
      "loss/crossentropy": 2.4616007804870605,
      "loss/hidden": 1.328125,
      "loss/logits": 0.21012833714485168,
      "loss/reg": 0.00019295132369734347,
      "step": 22
    },
    {
      "epoch": 0.002875,
      "grad_norm": 2.119006395339966,
      "grad_norm_var": 0.493518604142709,
      "learning_rate": 2.3000000000000003e-05,
      "loss": 1.3082,
      "loss/crossentropy": 2.633119821548462,
      "loss/hidden": 1.125,
      "loss/logits": 0.18130342662334442,
      "loss/reg": 0.0001929480058606714,
      "step": 23
    },
    {
      "epoch": 0.003,
      "grad_norm": 2.163001537322998,
      "grad_norm_var": 0.3308316653990078,
      "learning_rate": 2.4e-05,
      "loss": 1.2387,
      "loss/crossentropy": 2.494175672531128,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.16641545295715332,
      "loss/reg": 0.00019294496451038867,
      "step": 24
    },
    {
      "epoch": 0.003125,
      "grad_norm": 2.0063726902008057,
      "grad_norm_var": 0.34579458432486,
      "learning_rate": 2.5e-05,
      "loss": 1.122,
      "loss/crossentropy": 2.4526851177215576,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.1435042917728424,
      "loss/reg": 0.00019294198136776686,
      "step": 25
    },
    {
      "epoch": 0.00325,
      "grad_norm": 2.8159995079040527,
      "grad_norm_var": 0.34710604301850645,
      "learning_rate": 2.6000000000000002e-05,
      "loss": 1.4971,
      "loss/crossentropy": 2.3794209957122803,
      "loss/hidden": 1.28125,
      "loss/logits": 0.2139449566602707,
      "loss/reg": 0.00019293944933451712,
      "step": 26
    },
    {
      "epoch": 0.003375,
      "grad_norm": 4.617008209228516,
      "grad_norm_var": 0.6245762786757856,
      "learning_rate": 2.7000000000000002e-05,
      "loss": 1.5303,
      "loss/crossentropy": 2.374497413635254,
      "loss/hidden": 1.3046875,
      "loss/logits": 0.22372832894325256,
      "loss/reg": 0.0001929359568748623,
      "step": 27
    },
    {
      "epoch": 0.0035,
      "grad_norm": 3.328347682952881,
      "grad_norm_var": 0.6438493937520643,
      "learning_rate": 2.8000000000000003e-05,
      "loss": 1.5591,
      "loss/crossentropy": 2.732736587524414,
      "loss/hidden": 1.3203125,
      "loss/logits": 0.23687124252319336,
      "loss/reg": 0.00019293361401651055,
      "step": 28
    },
    {
      "epoch": 0.003625,
      "grad_norm": 2.3474817276000977,
      "grad_norm_var": 0.6333103119315846,
      "learning_rate": 2.9e-05,
      "loss": 1.2983,
      "loss/crossentropy": 2.3827714920043945,
      "loss/hidden": 1.09375,
      "loss/logits": 0.20258934795856476,
      "loss/reg": 0.00019292977231089026,
      "step": 29
    },
    {
      "epoch": 0.00375,
      "grad_norm": 2.560765504837036,
      "grad_norm_var": 0.6208746981103727,
      "learning_rate": 3e-05,
      "loss": 1.2558,
      "loss/crossentropy": 2.5275678634643555,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.16795387864112854,
      "loss/reg": 0.00019292706565465778,
      "step": 30
    },
    {
      "epoch": 0.003875,
      "grad_norm": 2.21565580368042,
      "grad_norm_var": 0.6000257496388314,
      "learning_rate": 3.1e-05,
      "loss": 1.1396,
      "loss/crossentropy": 2.991466999053955,
      "loss/hidden": 1.0,
      "loss/logits": 0.13766203820705414,
      "loss/reg": 0.00019292418437544256,
      "step": 31
    },
    {
      "epoch": 0.004,
      "grad_norm": 3.0804481506347656,
      "grad_norm_var": 0.6061713539146301,
      "learning_rate": 3.2000000000000005e-05,
      "loss": 1.4502,
      "loss/crossentropy": 2.5434651374816895,
      "loss/hidden": 1.234375,
      "loss/logits": 0.21392151713371277,
      "loss/reg": 0.00019291977514512837,
      "step": 32
    },
    {
      "epoch": 0.004125,
      "grad_norm": 3.4683053493499756,
      "grad_norm_var": 0.5923607081005333,
      "learning_rate": 3.3e-05,
      "loss": 1.586,
      "loss/crossentropy": 2.4229018688201904,
      "loss/hidden": 1.359375,
      "loss/logits": 0.2246609479188919,
      "loss/reg": 0.0001929169666254893,
      "step": 33
    },
    {
      "epoch": 0.00425,
      "grad_norm": 2.0332539081573486,
      "grad_norm_var": 0.4912531320085567,
      "learning_rate": 3.4000000000000007e-05,
      "loss": 1.1758,
      "loss/crossentropy": 2.4137802124023438,
      "loss/hidden": 1.015625,
      "loss/logits": 0.15825411677360535,
      "loss/reg": 0.000192913124919869,
      "step": 34
    },
    {
      "epoch": 0.004375,
      "grad_norm": 2.101600408554077,
      "grad_norm_var": 0.5125015485684291,
      "learning_rate": 3.5e-05,
      "loss": 1.2872,
      "loss/crossentropy": 2.5844082832336426,
      "loss/hidden": 1.09375,
      "loss/logits": 0.19150416553020477,
      "loss/reg": 0.0001929093268699944,
      "step": 35
    },
    {
      "epoch": 0.0045,
      "grad_norm": 3.8385353088378906,
      "grad_norm_var": 0.556932092742056,
      "learning_rate": 3.6e-05,
      "loss": 1.1458,
      "loss/crossentropy": 2.294477701187134,
      "loss/hidden": 0.98828125,
      "loss/logits": 0.15560123324394226,
      "loss/reg": 0.0001929069694597274,
      "step": 36
    },
    {
      "epoch": 0.004625,
      "grad_norm": 1.8931406736373901,
      "grad_norm_var": 0.6050138278151498,
      "learning_rate": 3.7e-05,
      "loss": 1.0789,
      "loss/crossentropy": 2.5690038204193115,
      "loss/hidden": 0.93359375,
      "loss/logits": 0.14334949851036072,
      "loss/reg": 0.00019290446653030813,
      "step": 37
    },
    {
      "epoch": 0.00475,
      "grad_norm": 1.720080018043518,
      "grad_norm_var": 0.6642705659226048,
      "learning_rate": 3.8e-05,
      "loss": 1.2069,
      "loss/crossentropy": 2.289663791656494,
      "loss/hidden": 1.046875,
      "loss/logits": 0.158083975315094,
      "loss/reg": 0.0001929002464748919,
      "step": 38
    },
    {
      "epoch": 0.004875,
      "grad_norm": 1.7338802814483643,
      "grad_norm_var": 0.7005152543709433,
      "learning_rate": 3.9000000000000006e-05,
      "loss": 1.0516,
      "loss/crossentropy": 2.6059696674346924,
      "loss/hidden": 0.921875,
      "loss/logits": 0.1277628391981125,
      "loss/reg": 0.00019289416377432644,
      "step": 39
    },
    {
      "epoch": 0.005,
      "grad_norm": 2.491640329360962,
      "grad_norm_var": 0.6872298635287296,
      "learning_rate": 4e-05,
      "loss": 1.4446,
      "loss/crossentropy": 2.4038562774658203,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.2316935658454895,
      "loss/reg": 0.0001928891142597422,
      "step": 40
    },
    {
      "epoch": 0.005125,
      "grad_norm": 2.42004656791687,
      "grad_norm_var": 0.6629334231954789,
      "learning_rate": 4.1e-05,
      "loss": 1.2198,
      "loss/crossentropy": 2.4321351051330566,
      "loss/hidden": 1.0625,
      "loss/logits": 0.15540093183517456,
      "loss/reg": 0.00019288310431875288,
      "step": 41
    },
    {
      "epoch": 0.00525,
      "grad_norm": 2.5629024505615234,
      "grad_norm_var": 0.6618966221430126,
      "learning_rate": 4.2e-05,
      "loss": 1.5254,
      "loss/crossentropy": 2.23368501663208,
      "loss/hidden": 1.3046875,
      "loss/logits": 0.21881133317947388,
      "loss/reg": 0.0001928747951751575,
      "step": 42
    },
    {
      "epoch": 0.005375,
      "grad_norm": 2.2626430988311768,
      "grad_norm_var": 0.3911191161730377,
      "learning_rate": 4.3e-05,
      "loss": 1.4362,
      "loss/crossentropy": 2.2880733013153076,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.22328704595565796,
      "loss/reg": 0.0001928645942825824,
      "step": 43
    },
    {
      "epoch": 0.0055,
      "grad_norm": 2.928407907485962,
      "grad_norm_var": 0.3571399417370721,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.1088,
      "loss/crossentropy": 2.837071418762207,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.1420312374830246,
      "loss/reg": 0.00019285624148324132,
      "step": 44
    },
    {
      "epoch": 0.005625,
      "grad_norm": 2.7014455795288086,
      "grad_norm_var": 0.35877893903101093,
      "learning_rate": 4.5e-05,
      "loss": 1.3821,
      "loss/crossentropy": 2.2957558631896973,
      "loss/hidden": 1.1953125,
      "loss/logits": 0.18486037850379944,
      "loss/reg": 0.0001928448909893632,
      "step": 45
    },
    {
      "epoch": 0.00575,
      "grad_norm": 1.8899732828140259,
      "grad_norm_var": 0.38153805228543447,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.0258,
      "loss/crossentropy": 2.3359339237213135,
      "loss/hidden": 0.90625,
      "loss/logits": 0.11766321957111359,
      "loss/reg": 0.00019283634901512414,
      "step": 46
    },
    {
      "epoch": 0.005875,
      "grad_norm": 2.2188658714294434,
      "grad_norm_var": 0.3814345973993899,
      "learning_rate": 4.7e-05,
      "loss": 1.0929,
      "loss/crossentropy": 2.6772139072418213,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.12616246938705444,
      "loss/reg": 0.00019282741413917392,
      "step": 47
    },
    {
      "epoch": 0.006,
      "grad_norm": 2.1530601978302,
      "grad_norm_var": 0.35835352199148307,
      "learning_rate": 4.8e-05,
      "loss": 1.2993,
      "loss/crossentropy": 2.678344249725342,
      "loss/hidden": 1.109375,
      "loss/logits": 0.18795417249202728,
      "loss/reg": 0.00019281756249256432,
      "step": 48
    },
    {
      "epoch": 0.006125,
      "grad_norm": 2.0299501419067383,
      "grad_norm_var": 0.28299041785001716,
      "learning_rate": 4.9e-05,
      "loss": 1.1433,
      "loss/crossentropy": 2.6076881885528564,
      "loss/hidden": 0.9921875,
      "loss/logits": 0.14920485019683838,
      "loss/reg": 0.00019280907872598618,
      "step": 49
    },
    {
      "epoch": 0.00625,
      "grad_norm": 2.6268253326416016,
      "grad_norm_var": 0.28301229188597427,
      "learning_rate": 5e-05,
      "loss": 1.057,
      "loss/crossentropy": 2.625675678253174,
      "loss/hidden": 0.91796875,
      "loss/logits": 0.13711729645729065,
      "loss/reg": 0.00019279817934148014,
      "step": 50
    },
    {
      "epoch": 0.006375,
      "grad_norm": 1.8832714557647705,
      "grad_norm_var": 0.29317342698340093,
      "learning_rate": 5.1000000000000006e-05,
      "loss": 1.2667,
      "loss/crossentropy": 2.4522390365600586,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.17878666520118713,
      "loss/reg": 0.00019278968102298677,
      "step": 51
    },
    {
      "epoch": 0.0065,
      "grad_norm": 2.7895314693450928,
      "grad_norm_var": 0.15160714498306468,
      "learning_rate": 5.2000000000000004e-05,
      "loss": 1.38,
      "loss/crossentropy": 2.143043041229248,
      "loss/hidden": 1.21875,
      "loss/logits": 0.15935076773166656,
      "loss/reg": 0.00019278022227808833,
      "step": 52
    },
    {
      "epoch": 0.006625,
      "grad_norm": 3.2793004512786865,
      "grad_norm_var": 0.20221103833160992,
      "learning_rate": 5.300000000000001e-05,
      "loss": 1.5103,
      "loss/crossentropy": 2.9073903560638428,
      "loss/hidden": 1.25,
      "loss/logits": 0.2583482563495636,
      "loss/reg": 0.00019277248065918684,
      "step": 53
    },
    {
      "epoch": 0.00675,
      "grad_norm": 2.013625144958496,
      "grad_norm_var": 0.18271730407283251,
      "learning_rate": 5.4000000000000005e-05,
      "loss": 1.0809,
      "loss/crossentropy": 2.50508975982666,
      "loss/hidden": 0.9453125,
      "loss/logits": 0.1336328089237213,
      "loss/reg": 0.00019276094099041075,
      "step": 54
    },
    {
      "epoch": 0.006875,
      "grad_norm": 2.584568738937378,
      "grad_norm_var": 0.15533136257730426,
      "learning_rate": 5.500000000000001e-05,
      "loss": 1.2027,
      "loss/crossentropy": 2.4901885986328125,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1539323329925537,
      "loss/reg": 0.00019275395607110113,
      "step": 55
    },
    {
      "epoch": 0.007,
      "grad_norm": 2.0262043476104736,
      "grad_norm_var": 0.16487505994896826,
      "learning_rate": 5.6000000000000006e-05,
      "loss": 1.3553,
      "loss/crossentropy": 2.299118757247925,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.20495867729187012,
      "loss/reg": 0.0001927466510096565,
      "step": 56
    },
    {
      "epoch": 0.007125,
      "grad_norm": 2.4767560958862305,
      "grad_norm_var": 0.16524151904890463,
      "learning_rate": 5.6999999999999996e-05,
      "loss": 1.2841,
      "loss/crossentropy": 2.5054197311401367,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.1650199145078659,
      "loss/reg": 0.00019273380166850984,
      "step": 57
    },
    {
      "epoch": 0.00725,
      "grad_norm": 1.8864933252334595,
      "grad_norm_var": 0.17929933439751622,
      "learning_rate": 5.8e-05,
      "loss": 1.0831,
      "loss/crossentropy": 2.6685423851013184,
      "loss/hidden": 0.95703125,
      "loss/logits": 0.1241319477558136,
      "loss/reg": 0.00019271954079158604,
      "step": 58
    },
    {
      "epoch": 0.007375,
      "grad_norm": 1.927100658416748,
      "grad_norm_var": 0.19066639705673746,
      "learning_rate": 5.9e-05,
      "loss": 1.2188,
      "loss/crossentropy": 2.4530575275421143,
      "loss/hidden": 1.0546875,
      "loss/logits": 0.1622183918952942,
      "loss/reg": 0.0001927079283632338,
      "step": 59
    },
    {
      "epoch": 0.0075,
      "grad_norm": 1.9683936834335327,
      "grad_norm_var": 0.17275381294839126,
      "learning_rate": 6e-05,
      "loss": 1.1956,
      "loss/crossentropy": 2.699939250946045,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.15460126101970673,
      "loss/reg": 0.00019269227050244808,
      "step": 60
    },
    {
      "epoch": 0.007625,
      "grad_norm": 2.430546998977661,
      "grad_norm_var": 0.1620622944705805,
      "learning_rate": 6.1e-05,
      "loss": 1.2826,
      "loss/crossentropy": 2.2341370582580566,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.16349473595619202,
      "loss/reg": 0.0001926780241774395,
      "step": 61
    },
    {
      "epoch": 0.00775,
      "grad_norm": 2.28157377243042,
      "grad_norm_var": 0.15224653123686924,
      "learning_rate": 6.2e-05,
      "loss": 1.0961,
      "loss/crossentropy": 2.3603291511535645,
      "loss/hidden": 0.953125,
      "loss/logits": 0.14104235172271729,
      "loss/reg": 0.00019266277377028018,
      "step": 62
    },
    {
      "epoch": 0.007875,
      "grad_norm": 1.9362468719482422,
      "grad_norm_var": 0.1597685683605616,
      "learning_rate": 6.3e-05,
      "loss": 1.1251,
      "loss/crossentropy": 2.6113460063934326,
      "loss/hidden": 0.984375,
      "loss/logits": 0.1387864351272583,
      "loss/reg": 0.00019264982256572694,
      "step": 63
    },
    {
      "epoch": 0.008,
      "grad_norm": 3.4527783393859863,
      "grad_norm_var": 0.245370177212871,
      "learning_rate": 6.400000000000001e-05,
      "loss": 1.5699,
      "loss/crossentropy": 2.3186392784118652,
      "loss/hidden": 1.3359375,
      "loss/logits": 0.2319989800453186,
      "loss/reg": 0.00019263311696704477,
      "step": 64
    },
    {
      "epoch": 0.008125,
      "grad_norm": 2.002469062805176,
      "grad_norm_var": 0.24658852169075185,
      "learning_rate": 6.500000000000001e-05,
      "loss": 1.2089,
      "loss/crossentropy": 2.1945531368255615,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1600569784641266,
      "loss/reg": 0.0001926190307131037,
      "step": 65
    },
    {
      "epoch": 0.00825,
      "grad_norm": 1.8942357301712036,
      "grad_norm_var": 0.25288209179575216,
      "learning_rate": 6.6e-05,
      "loss": 1.1301,
      "loss/crossentropy": 2.3987529277801514,
      "loss/hidden": 0.98046875,
      "loss/logits": 0.1477484107017517,
      "loss/reg": 0.0001926012773765251,
      "step": 66
    },
    {
      "epoch": 0.008375,
      "grad_norm": 2.9109678268432617,
      "grad_norm_var": 0.2615059196420252,
      "learning_rate": 6.7e-05,
      "loss": 1.2255,
      "loss/crossentropy": 2.5337891578674316,
      "loss/hidden": 1.0625,
      "loss/logits": 0.16106070578098297,
      "loss/reg": 0.00019258313113823533,
      "step": 67
    },
    {
      "epoch": 0.0085,
      "grad_norm": 2.1384077072143555,
      "grad_norm_var": 0.25126003810038394,
      "learning_rate": 6.800000000000001e-05,
      "loss": 1.3278,
      "loss/crossentropy": 2.577103614807129,
      "loss/hidden": 1.125,
      "loss/logits": 0.20084746181964874,
      "loss/reg": 0.00019256297673564404,
      "step": 68
    },
    {
      "epoch": 0.008625,
      "grad_norm": 2.1597537994384766,
      "grad_norm_var": 0.18723560405016518,
      "learning_rate": 6.9e-05,
      "loss": 1.1538,
      "loss/crossentropy": 2.6212854385375977,
      "loss/hidden": 1.0,
      "loss/logits": 0.15191976726055145,
      "loss/reg": 0.00019254189101047814,
      "step": 69
    },
    {
      "epoch": 0.00875,
      "grad_norm": 1.8894274234771729,
      "grad_norm_var": 0.19220724163081773,
      "learning_rate": 7e-05,
      "loss": 1.1419,
      "loss/crossentropy": 2.439669132232666,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.17508690059185028,
      "loss/reg": 0.00019252618949394673,
      "step": 70
    },
    {
      "epoch": 0.008875,
      "grad_norm": 1.9923198223114014,
      "grad_norm_var": 0.1875417585129336,
      "learning_rate": 7.1e-05,
      "loss": 1.159,
      "loss/crossentropy": 2.593405246734619,
      "loss/hidden": 0.99609375,
      "loss/logits": 0.16099245846271515,
      "loss/reg": 0.00019250869809184223,
      "step": 71
    },
    {
      "epoch": 0.009,
      "grad_norm": 2.1258373260498047,
      "grad_norm_var": 0.18570921033151055,
      "learning_rate": 7.2e-05,
      "loss": 1.2886,
      "loss/crossentropy": 2.31851863861084,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.16945742070674896,
      "loss/reg": 0.00019249116303399205,
      "step": 72
    },
    {
      "epoch": 0.009125,
      "grad_norm": 2.2625362873077393,
      "grad_norm_var": 0.1811603588675789,
      "learning_rate": 7.3e-05,
      "loss": 1.0782,
      "loss/crossentropy": 2.4330179691314697,
      "loss/hidden": 0.93359375,
      "loss/logits": 0.14271824061870575,
      "loss/reg": 0.0001924755924846977,
      "step": 73
    },
    {
      "epoch": 0.00925,
      "grad_norm": 2.4275050163269043,
      "grad_norm_var": 0.17657254479349263,
      "learning_rate": 7.4e-05,
      "loss": 1.2584,
      "loss/crossentropy": 2.5500473976135254,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.186170294880867,
      "loss/reg": 0.0001924607204273343,
      "step": 74
    },
    {
      "epoch": 0.009375,
      "grad_norm": 2.333041191101074,
      "grad_norm_var": 0.17007094778421958,
      "learning_rate": 7.500000000000001e-05,
      "loss": 1.0767,
      "loss/crossentropy": 2.6195261478424072,
      "loss/hidden": 0.94921875,
      "loss/logits": 0.12556512653827667,
      "loss/reg": 0.0001924492244143039,
      "step": 75
    },
    {
      "epoch": 0.0095,
      "grad_norm": 2.6031246185302734,
      "grad_norm_var": 0.1703287548027977,
      "learning_rate": 7.6e-05,
      "loss": 1.2179,
      "loss/crossentropy": 2.488292932510376,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.1925877332687378,
      "loss/reg": 0.000192438907106407,
      "step": 76
    },
    {
      "epoch": 0.009625,
      "grad_norm": 2.139622211456299,
      "grad_norm_var": 0.17065351345722057,
      "learning_rate": 7.7e-05,
      "loss": 1.0607,
      "loss/crossentropy": 2.934234619140625,
      "loss/hidden": 0.921875,
      "loss/logits": 0.1368916928768158,
      "loss/reg": 0.0001924296229844913,
      "step": 77
    },
    {
      "epoch": 0.00975,
      "grad_norm": 1.9406142234802246,
      "grad_norm_var": 0.17804626450119793,
      "learning_rate": 7.800000000000001e-05,
      "loss": 1.2752,
      "loss/crossentropy": 2.2727177143096924,
      "loss/hidden": 1.1015625,
      "loss/logits": 0.17168399691581726,
      "loss/reg": 0.0001924206007970497,
      "step": 78
    },
    {
      "epoch": 0.009875,
      "grad_norm": 3.055265426635742,
      "grad_norm_var": 0.20754827159903322,
      "learning_rate": 7.900000000000001e-05,
      "loss": 1.1607,
      "loss/crossentropy": 2.6755881309509277,
      "loss/hidden": 1.0,
      "loss/logits": 0.15880770981311798,
      "loss/reg": 0.00019241031259298325,
      "step": 79
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.129188060760498,
      "grad_norm_var": 0.11942340663251176,
      "learning_rate": 8e-05,
      "loss": 1.2684,
      "loss/crossentropy": 2.5891458988189697,
      "loss/hidden": 1.078125,
      "loss/logits": 0.18836861848831177,
      "loss/reg": 0.00019238927052356303,
      "step": 80
    },
    {
      "epoch": 0.010125,
      "grad_norm": 3.2390594482421875,
      "grad_norm_var": 0.17413858607061497,
      "learning_rate": 8.1e-05,
      "loss": 2.3912,
      "loss/crossentropy": 2.566899061203003,
      "loss/hidden": 1.7265625,
      "loss/logits": 0.6627247333526611,
      "loss/reg": 0.00019236840307712555,
      "step": 81
    },
    {
      "epoch": 0.01025,
      "grad_norm": 2.6307613849639893,
      "grad_norm_var": 0.16548936874203926,
      "learning_rate": 8.2e-05,
      "loss": 1.2744,
      "loss/crossentropy": 1.8344066143035889,
      "loss/hidden": 1.140625,
      "loss/logits": 0.1318967193365097,
      "loss/reg": 0.00019235462241340429,
      "step": 82
    },
    {
      "epoch": 0.010375,
      "grad_norm": 3.545522928237915,
      "grad_norm_var": 0.23612178547081322,
      "learning_rate": 8.3e-05,
      "loss": 1.2902,
      "loss/crossentropy": 2.7840187549591064,
      "loss/hidden": 1.109375,
      "loss/logits": 0.1788717359304428,
      "loss/reg": 0.00019232937484048307,
      "step": 83
    },
    {
      "epoch": 0.0105,
      "grad_norm": 2.2964553833007812,
      "grad_norm_var": 0.23189123641267292,
      "learning_rate": 8.4e-05,
      "loss": 1.4269,
      "loss/crossentropy": 2.6213274002075195,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.19838152825832367,
      "loss/reg": 0.00019230577163398266,
      "step": 84
    },
    {
      "epoch": 0.010625,
      "grad_norm": 2.2252533435821533,
      "grad_norm_var": 0.229859261969087,
      "learning_rate": 8.5e-05,
      "loss": 1.1939,
      "loss/crossentropy": 2.493159055709839,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.16855394840240479,
      "loss/reg": 0.00019228595192544162,
      "step": 85
    },
    {
      "epoch": 0.01075,
      "grad_norm": 2.2857768535614014,
      "grad_norm_var": 0.21125701567141184,
      "learning_rate": 8.6e-05,
      "loss": 1.2406,
      "loss/crossentropy": 2.486485719680786,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1917887181043625,
      "loss/reg": 0.00019227097800467163,
      "step": 86
    },
    {
      "epoch": 0.010875,
      "grad_norm": 4.509547710418701,
      "grad_norm_var": 0.4530040889277906,
      "learning_rate": 8.7e-05,
      "loss": 1.1073,
      "loss/crossentropy": 2.630571126937866,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.14057384431362152,
      "loss/reg": 0.00019225555297452956,
      "step": 87
    },
    {
      "epoch": 0.011,
      "grad_norm": 2.0168137550354004,
      "grad_norm_var": 0.4607750991685954,
      "learning_rate": 8.800000000000001e-05,
      "loss": 1.053,
      "loss/crossentropy": 2.3603971004486084,
      "loss/hidden": 0.9140625,
      "loss/logits": 0.13705039024353027,
      "loss/reg": 0.0001922310038935393,
      "step": 88
    },
    {
      "epoch": 0.011125,
      "grad_norm": 2.73807430267334,
      "grad_norm_var": 0.45335285375272844,
      "learning_rate": 8.900000000000001e-05,
      "loss": 1.3178,
      "loss/crossentropy": 2.93936824798584,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.19865703582763672,
      "loss/reg": 0.00019220533431507647,
      "step": 89
    },
    {
      "epoch": 0.01125,
      "grad_norm": 2.305682420730591,
      "grad_norm_var": 0.4576056958578417,
      "learning_rate": 9e-05,
      "loss": 1.0605,
      "loss/crossentropy": 2.4470906257629395,
      "loss/hidden": 0.921875,
      "loss/logits": 0.13670536875724792,
      "loss/reg": 0.00019217943190596998,
      "step": 90
    },
    {
      "epoch": 0.011375,
      "grad_norm": 2.333709239959717,
      "grad_norm_var": 0.45757975254874145,
      "learning_rate": 9.1e-05,
      "loss": 1.1485,
      "loss/crossentropy": 2.3914265632629395,
      "loss/hidden": 1.0,
      "loss/logits": 0.14661094546318054,
      "loss/reg": 0.00019215639622416347,
      "step": 91
    },
    {
      "epoch": 0.0115,
      "grad_norm": 3.1051077842712402,
      "grad_norm_var": 0.4718879306887779,
      "learning_rate": 9.200000000000001e-05,
      "loss": 1.1946,
      "loss/crossentropy": 2.3768973350524902,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1458442062139511,
      "loss/reg": 0.00019213555788155645,
      "step": 92
    },
    {
      "epoch": 0.011625,
      "grad_norm": 1.9250227212905884,
      "grad_norm_var": 0.48954230695473144,
      "learning_rate": 9.300000000000001e-05,
      "loss": 1.0554,
      "loss/crossentropy": 2.565377712249756,
      "loss/hidden": 0.9140625,
      "loss/logits": 0.13942840695381165,
      "loss/reg": 0.00019211515609640628,
      "step": 93
    },
    {
      "epoch": 0.01175,
      "grad_norm": 2.176602840423584,
      "grad_norm_var": 0.4709343827101543,
      "learning_rate": 9.4e-05,
      "loss": 1.1256,
      "loss/crossentropy": 2.4452078342437744,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.1470910608768463,
      "loss/reg": 0.00019209457968827337,
      "step": 94
    },
    {
      "epoch": 0.011875,
      "grad_norm": 3.14528489112854,
      "grad_norm_var": 0.4762166867826877,
      "learning_rate": 9.5e-05,
      "loss": 1.6721,
      "loss/crossentropy": 2.0026628971099854,
      "loss/hidden": 1.421875,
      "loss/logits": 0.24830523133277893,
      "loss/reg": 0.0001920641807373613,
      "step": 95
    },
    {
      "epoch": 0.012,
      "grad_norm": 2.5958688259124756,
      "grad_norm_var": 0.4566131842781544,
      "learning_rate": 9.6e-05,
      "loss": 1.2013,
      "loss/crossentropy": 2.620640993118286,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.1759262979030609,
      "loss/reg": 0.00019204463751520962,
      "step": 96
    },
    {
      "epoch": 0.012125,
      "grad_norm": 2.195146322250366,
      "grad_norm_var": 0.4486006387079304,
      "learning_rate": 9.7e-05,
      "loss": 1.1398,
      "loss/crossentropy": 2.547090768814087,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.16133451461791992,
      "loss/reg": 0.0001920342183439061,
      "step": 97
    },
    {
      "epoch": 0.01225,
      "grad_norm": 2.252614974975586,
      "grad_norm_var": 0.4573438457489293,
      "learning_rate": 9.8e-05,
      "loss": 1.1583,
      "loss/crossentropy": 2.4780287742614746,
      "loss/hidden": 1.0,
      "loss/logits": 0.15635529160499573,
      "loss/reg": 0.00019200837414246053,
      "step": 98
    },
    {
      "epoch": 0.012375,
      "grad_norm": 2.421096086502075,
      "grad_norm_var": 0.3951004366779675,
      "learning_rate": 9.900000000000001e-05,
      "loss": 1.1038,
      "loss/crossentropy": 2.2537543773651123,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.13705003261566162,
      "loss/reg": 0.0001919834321597591,
      "step": 99
    },
    {
      "epoch": 0.0125,
      "grad_norm": 2.174814462661743,
      "grad_norm_var": 0.3998617443443601,
      "learning_rate": 0.0001,
      "loss": 1.1691,
      "loss/crossentropy": 2.458591938018799,
      "loss/hidden": 0.99609375,
      "loss/logits": 0.17106780409812927,
      "loss/reg": 0.0001919578353408724,
      "step": 100
    },
    {
      "epoch": 0.012625,
      "grad_norm": 2.622562885284424,
      "grad_norm_var": 0.39382746835866644,
      "learning_rate": 0.0001,
      "loss": 1.1769,
      "loss/crossentropy": 2.5160534381866455,
      "loss/hidden": 1.0078125,
      "loss/logits": 0.16712725162506104,
      "loss/reg": 0.00019193820480722934,
      "step": 101
    },
    {
      "epoch": 0.01275,
      "grad_norm": 2.24092960357666,
      "grad_norm_var": 0.3955345231673798,
      "learning_rate": 0.0001,
      "loss": 1.2354,
      "loss/crossentropy": 2.1451447010040283,
      "loss/hidden": 1.078125,
      "loss/logits": 0.1553465873003006,
      "loss/reg": 0.00019191819592379034,
      "step": 102
    },
    {
      "epoch": 0.012875,
      "grad_norm": 2.344604730606079,
      "grad_norm_var": 0.12208757192350952,
      "learning_rate": 0.0001,
      "loss": 1.4627,
      "loss/crossentropy": 2.641226053237915,
      "loss/hidden": 1.203125,
      "loss/logits": 0.2576836943626404,
      "loss/reg": 0.00019189789600204676,
      "step": 103
    },
    {
      "epoch": 0.013,
      "grad_norm": 2.6583104133605957,
      "grad_norm_var": 0.1139956751841869,
      "learning_rate": 0.0001,
      "loss": 1.2235,
      "loss/crossentropy": 2.2702269554138184,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1746675968170166,
      "loss/reg": 0.00019186925783287734,
      "step": 104
    },
    {
      "epoch": 0.013125,
      "grad_norm": 2.452503204345703,
      "grad_norm_var": 0.10820816494096158,
      "learning_rate": 0.0001,
      "loss": 1.2496,
      "loss/crossentropy": 2.344975709915161,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.1617823839187622,
      "loss/reg": 0.000191839542821981,
      "step": 105
    },
    {
      "epoch": 0.01325,
      "grad_norm": 2.3220486640930176,
      "grad_norm_var": 0.1079440961702573,
      "learning_rate": 0.0001,
      "loss": 1.3184,
      "loss/crossentropy": 2.382570505142212,
      "loss/hidden": 1.140625,
      "loss/logits": 0.1758350431919098,
      "loss/reg": 0.00019180966774001718,
      "step": 106
    },
    {
      "epoch": 0.013375,
      "grad_norm": 2.275015115737915,
      "grad_norm_var": 0.1089551443983727,
      "learning_rate": 0.0001,
      "loss": 1.1398,
      "loss/crossentropy": 2.288180112838745,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.1613542139530182,
      "loss/reg": 0.00019177970534656197,
      "step": 107
    },
    {
      "epoch": 0.0135,
      "grad_norm": 2.281081199645996,
      "grad_norm_var": 0.0774087174098435,
      "learning_rate": 0.0001,
      "loss": 1.1199,
      "loss/crossentropy": 2.566871404647827,
      "loss/hidden": 0.96875,
      "loss/logits": 0.14925265312194824,
      "loss/reg": 0.00019174529006704688,
      "step": 108
    },
    {
      "epoch": 0.013625,
      "grad_norm": 2.281104564666748,
      "grad_norm_var": 0.06372173379328742,
      "learning_rate": 0.0001,
      "loss": 1.1916,
      "loss/crossentropy": 2.335575580596924,
      "loss/hidden": 1.015625,
      "loss/logits": 0.1740630865097046,
      "loss/reg": 0.00019171558960806578,
      "step": 109
    },
    {
      "epoch": 0.01375,
      "grad_norm": 3.2676873207092285,
      "grad_norm_var": 0.10526650532868657,
      "learning_rate": 0.0001,
      "loss": 1.4065,
      "loss/crossentropy": 2.33091402053833,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.2248753160238266,
      "loss/reg": 0.00019168361905030906,
      "step": 110
    },
    {
      "epoch": 0.013875,
      "grad_norm": 2.448396921157837,
      "grad_norm_var": 0.07293540299188876,
      "learning_rate": 0.0001,
      "loss": 1.1487,
      "loss/crossentropy": 2.759523391723633,
      "loss/hidden": 0.98828125,
      "loss/logits": 0.1585177183151245,
      "loss/reg": 0.0001916515757329762,
      "step": 111
    },
    {
      "epoch": 0.014,
      "grad_norm": 2.745311975479126,
      "grad_norm_var": 0.07769384665264147,
      "learning_rate": 0.0001,
      "loss": 1.3054,
      "loss/crossentropy": 2.6752800941467285,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.17071780562400818,
      "loss/reg": 0.00019161647651344538,
      "step": 112
    },
    {
      "epoch": 0.014125,
      "grad_norm": 6.7866058349609375,
      "grad_norm_var": 1.2475617279497555,
      "learning_rate": 0.0001,
      "loss": 1.527,
      "loss/crossentropy": 2.536259174346924,
      "loss/hidden": 1.328125,
      "loss/logits": 0.19696056842803955,
      "loss/reg": 0.0001915794564411044,
      "step": 113
    },
    {
      "epoch": 0.01425,
      "grad_norm": 2.572187900543213,
      "grad_norm_var": 1.2338838698080574,
      "learning_rate": 0.0001,
      "loss": 1.0705,
      "loss/crossentropy": 3.121675968170166,
      "loss/hidden": 0.92578125,
      "loss/logits": 0.1427587866783142,
      "loss/reg": 0.0001915483589982614,
      "step": 114
    },
    {
      "epoch": 0.014375,
      "grad_norm": 2.3879568576812744,
      "grad_norm_var": 1.2353765898385585,
      "learning_rate": 0.0001,
      "loss": 1.1787,
      "loss/crossentropy": 2.59924578666687,
      "loss/hidden": 1.015625,
      "loss/logits": 0.16120225191116333,
      "loss/reg": 0.0001915154862217605,
      "step": 115
    },
    {
      "epoch": 0.0145,
      "grad_norm": 2.207024097442627,
      "grad_norm_var": 1.2330085058190583,
      "learning_rate": 0.0001,
      "loss": 1.1257,
      "loss/crossentropy": 2.3122026920318604,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.15894125401973724,
      "loss/reg": 0.0001914859312819317,
      "step": 116
    },
    {
      "epoch": 0.014625,
      "grad_norm": 2.5943527221679688,
      "grad_norm_var": 1.233512504208524,
      "learning_rate": 0.0001,
      "loss": 1.0578,
      "loss/crossentropy": 2.4962315559387207,
      "loss/hidden": 0.9296875,
      "loss/logits": 0.12624415755271912,
      "loss/reg": 0.0001914564927574247,
      "step": 117
    },
    {
      "epoch": 0.01475,
      "grad_norm": 2.5308315753936768,
      "grad_norm_var": 1.2194136468208996,
      "learning_rate": 0.0001,
      "loss": 1.3378,
      "loss/crossentropy": 2.2406933307647705,
      "loss/hidden": 1.15625,
      "loss/logits": 0.17962361872196198,
      "loss/reg": 0.00019143095414619893,
      "step": 118
    },
    {
      "epoch": 0.014875,
      "grad_norm": 3.8317177295684814,
      "grad_norm_var": 1.2753290966219148,
      "learning_rate": 0.0001,
      "loss": 1.5298,
      "loss/crossentropy": 2.7381229400634766,
      "loss/hidden": 1.296875,
      "loss/logits": 0.2309999018907547,
      "loss/reg": 0.00019139735377393663,
      "step": 119
    },
    {
      "epoch": 0.015,
      "grad_norm": 2.552171468734741,
      "grad_norm_var": 1.2787832219082094,
      "learning_rate": 0.0001,
      "loss": 1.1698,
      "loss/crossentropy": 2.678725481033325,
      "loss/hidden": 1.0078125,
      "loss/logits": 0.16004806756973267,
      "loss/reg": 0.0001913599990075454,
      "step": 120
    },
    {
      "epoch": 0.015125,
      "grad_norm": 2.930230140686035,
      "grad_norm_var": 1.2679826365318356,
      "learning_rate": 0.0001,
      "loss": 1.3195,
      "loss/crossentropy": 2.5161564350128174,
      "loss/hidden": 1.1015625,
      "loss/logits": 0.21604904532432556,
      "loss/reg": 0.00019133626483380795,
      "step": 121
    },
    {
      "epoch": 0.01525,
      "grad_norm": 2.184554100036621,
      "grad_norm_var": 1.2793169490082976,
      "learning_rate": 0.0001,
      "loss": 1.3258,
      "loss/crossentropy": 2.1544501781463623,
      "loss/hidden": 1.140625,
      "loss/logits": 0.18321675062179565,
      "loss/reg": 0.0001913021260406822,
      "step": 122
    },
    {
      "epoch": 0.015375,
      "grad_norm": 2.3296663761138916,
      "grad_norm_var": 1.2751879992777064,
      "learning_rate": 0.0001,
      "loss": 1.2465,
      "loss/crossentropy": 2.3694839477539062,
      "loss/hidden": 1.078125,
      "loss/logits": 0.16649125516414642,
      "loss/reg": 0.00019128025451209396,
      "step": 123
    },
    {
      "epoch": 0.0155,
      "grad_norm": 2.683415174484253,
      "grad_norm_var": 1.2536762853317933,
      "learning_rate": 0.0001,
      "loss": 1.3152,
      "loss/crossentropy": 2.480109930038452,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.1961439996957779,
      "loss/reg": 0.00019124921527691185,
      "step": 124
    },
    {
      "epoch": 0.015625,
      "grad_norm": 2.2008211612701416,
      "grad_norm_var": 1.2606593807518112,
      "learning_rate": 0.0001,
      "loss": 1.2683,
      "loss/crossentropy": 2.3611254692077637,
      "loss/hidden": 1.09375,
      "loss/logits": 0.17260757088661194,
      "loss/reg": 0.00019121899094898254,
      "step": 125
    },
    {
      "epoch": 0.01575,
      "grad_norm": 2.5672714710235596,
      "grad_norm_var": 1.2561244980458308,
      "learning_rate": 0.0001,
      "loss": 1.1256,
      "loss/crossentropy": 2.4654054641723633,
      "loss/hidden": 0.97265625,
      "loss/logits": 0.1510714888572693,
      "loss/reg": 0.00019119179341942072,
      "step": 126
    },
    {
      "epoch": 0.015875,
      "grad_norm": 5.104769706726074,
      "grad_norm_var": 1.5559544106053274,
      "learning_rate": 0.0001,
      "loss": 1.7662,
      "loss/crossentropy": 2.8167941570281982,
      "loss/hidden": 1.46875,
      "loss/logits": 0.2955778241157532,
      "loss/reg": 0.00019115611212328076,
      "step": 127
    },
    {
      "epoch": 0.016,
      "grad_norm": 2.780869483947754,
      "grad_norm_var": 1.5547640591921663,
      "learning_rate": 0.0001,
      "loss": 1.3813,
      "loss/crossentropy": 2.4478983879089355,
      "loss/hidden": 1.1875,
      "loss/logits": 0.19183939695358276,
      "loss/reg": 0.00019112625159323215,
      "step": 128
    },
    {
      "epoch": 0.016125,
      "grad_norm": 2.753885269165039,
      "grad_norm_var": 0.5433630068432194,
      "learning_rate": 0.0001,
      "loss": 1.3723,
      "loss/crossentropy": 2.6112186908721924,
      "loss/hidden": 1.15625,
      "loss/logits": 0.2141554355621338,
      "loss/reg": 0.0001910965656861663,
      "step": 129
    },
    {
      "epoch": 0.01625,
      "grad_norm": 2.9186818599700928,
      "grad_norm_var": 0.5420405140966994,
      "learning_rate": 0.0001,
      "loss": 1.4122,
      "loss/crossentropy": 2.4329094886779785,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.199338898062706,
      "loss/reg": 0.00019106207764707506,
      "step": 130
    },
    {
      "epoch": 0.016375,
      "grad_norm": 3.3282015323638916,
      "grad_norm_var": 0.5475325270303432,
      "learning_rate": 0.0001,
      "loss": 1.3236,
      "loss/crossentropy": 2.808492660522461,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.20454144477844238,
      "loss/reg": 0.00019102977239526808,
      "step": 131
    },
    {
      "epoch": 0.0165,
      "grad_norm": 2.5251357555389404,
      "grad_norm_var": 0.5268546307130834,
      "learning_rate": 0.0001,
      "loss": 1.0806,
      "loss/crossentropy": 2.606909990310669,
      "loss/hidden": 0.94140625,
      "loss/logits": 0.1372692584991455,
      "loss/reg": 0.00019099873316008598,
      "step": 132
    },
    {
      "epoch": 0.016625,
      "grad_norm": 2.695373773574829,
      "grad_norm_var": 0.5238667023797914,
      "learning_rate": 0.0001,
      "loss": 1.1349,
      "loss/crossentropy": 2.5870134830474854,
      "loss/hidden": 0.97265625,
      "loss/logits": 0.16037797927856445,
      "loss/reg": 0.00019096150936093181,
      "step": 133
    },
    {
      "epoch": 0.01675,
      "grad_norm": 2.967905282974243,
      "grad_norm_var": 0.5160494986535296,
      "learning_rate": 0.0001,
      "loss": 1.3901,
      "loss/crossentropy": 2.3568673133850098,
      "loss/hidden": 1.1875,
      "loss/logits": 0.20067735016345978,
      "loss/reg": 0.0001909265120048076,
      "step": 134
    },
    {
      "epoch": 0.016875,
      "grad_norm": 2.248267412185669,
      "grad_norm_var": 0.4754480378524304,
      "learning_rate": 0.0001,
      "loss": 1.2149,
      "loss/crossentropy": 2.513349771499634,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1660783886909485,
      "loss/reg": 0.00019089688430540264,
      "step": 135
    },
    {
      "epoch": 0.017,
      "grad_norm": 2.6341240406036377,
      "grad_norm_var": 0.4731794320985157,
      "learning_rate": 0.0001,
      "loss": 1.4788,
      "loss/crossentropy": 2.307194709777832,
      "loss/hidden": 1.2421875,
      "loss/logits": 0.23469889163970947,
      "loss/reg": 0.00019086015527136624,
      "step": 136
    },
    {
      "epoch": 0.017125,
      "grad_norm": 26.939998626708984,
      "grad_norm_var": 36.90875808790156,
      "learning_rate": 0.0001,
      "loss": 1.9603,
      "loss/crossentropy": 2.4554264545440674,
      "loss/hidden": 1.7265625,
      "loss/logits": 0.23182228207588196,
      "loss/reg": 0.00019083071674685925,
      "step": 137
    },
    {
      "epoch": 0.01725,
      "grad_norm": 2.5422139167785645,
      "grad_norm_var": 36.81568419391185,
      "learning_rate": 0.0001,
      "loss": 1.0683,
      "loss/crossentropy": 2.4824135303497314,
      "loss/hidden": 0.921875,
      "loss/logits": 0.14447075128555298,
      "loss/reg": 0.00019080075435340405,
      "step": 138
    },
    {
      "epoch": 0.017375,
      "grad_norm": 8.87843132019043,
      "grad_norm_var": 37.75269230148678,
      "learning_rate": 0.0001,
      "loss": 2.5968,
      "loss/crossentropy": 3.632392406463623,
      "loss/hidden": 2.0,
      "loss/logits": 0.594857931137085,
      "loss/reg": 0.0001907727710204199,
      "step": 139
    },
    {
      "epoch": 0.0175,
      "grad_norm": 2.6119391918182373,
      "grad_norm_var": 37.77256905325698,
      "learning_rate": 0.0001,
      "loss": 1.2831,
      "loss/crossentropy": 2.694307804107666,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.19521979987621307,
      "loss/reg": 0.00019074990996159613,
      "step": 140
    },
    {
      "epoch": 0.017625,
      "grad_norm": 2.630040407180786,
      "grad_norm_var": 37.63927642256101,
      "learning_rate": 0.0001,
      "loss": 1.2854,
      "loss/crossentropy": 2.7011289596557617,
      "loss/hidden": 1.109375,
      "loss/logits": 0.17411190271377563,
      "loss/reg": 0.0001907304977066815,
      "step": 141
    },
    {
      "epoch": 0.01775,
      "grad_norm": 2.6480321884155273,
      "grad_norm_var": 37.616094691169124,
      "learning_rate": 0.0001,
      "loss": 1.5096,
      "loss/crossentropy": 2.260585069656372,
      "loss/hidden": 1.2890625,
      "loss/logits": 0.218610018491745,
      "loss/reg": 0.00019069462723564357,
      "step": 142
    },
    {
      "epoch": 0.017875,
      "grad_norm": 6.334838390350342,
      "grad_norm_var": 37.766716198289636,
      "learning_rate": 0.0001,
      "loss": 1.4905,
      "loss/crossentropy": 2.702542543411255,
      "loss/hidden": 1.265625,
      "loss/logits": 0.22292816638946533,
      "loss/reg": 0.00019065497326664627,
      "step": 143
    },
    {
      "epoch": 0.018,
      "grad_norm": 2.5124268531799316,
      "grad_norm_var": 37.84491654864667,
      "learning_rate": 0.0001,
      "loss": 1.4406,
      "loss/crossentropy": 2.3640494346618652,
      "loss/hidden": 1.234375,
      "loss/logits": 0.20432478189468384,
      "loss/reg": 0.00019061053171753883,
      "step": 144
    },
    {
      "epoch": 0.018125,
      "grad_norm": 2.5983355045318604,
      "grad_norm_var": 37.889344095265606,
      "learning_rate": 0.0001,
      "loss": 1.2585,
      "loss/crossentropy": 2.386502265930176,
      "loss/hidden": 1.109375,
      "loss/logits": 0.1472403109073639,
      "loss/reg": 0.0001905607496155426,
      "step": 145
    },
    {
      "epoch": 0.01825,
      "grad_norm": 2.3283042907714844,
      "grad_norm_var": 38.06027251189639,
      "learning_rate": 0.0001,
      "loss": 1.0838,
      "loss/crossentropy": 2.589846134185791,
      "loss/hidden": 0.94140625,
      "loss/logits": 0.14048275351524353,
      "loss/reg": 0.00019053251889999956,
      "step": 146
    },
    {
      "epoch": 0.018375,
      "grad_norm": 3.144895315170288,
      "grad_norm_var": 38.09776954094634,
      "learning_rate": 0.0001,
      "loss": 1.44,
      "loss/crossentropy": 2.7745416164398193,
      "loss/hidden": 1.234375,
      "loss/logits": 0.2037464678287506,
      "loss/reg": 0.00019048065587412566,
      "step": 147
    },
    {
      "epoch": 0.0185,
      "grad_norm": 26.648195266723633,
      "grad_norm_var": 67.26352470044017,
      "learning_rate": 0.0001,
      "loss": 1.5186,
      "loss/crossentropy": 2.64324951171875,
      "loss/hidden": 1.3125,
      "loss/logits": 0.20418909192085266,
      "loss/reg": 0.00019045177032239735,
      "step": 148
    },
    {
      "epoch": 0.018625,
      "grad_norm": 3.2434163093566895,
      "grad_norm_var": 67.02089246655065,
      "learning_rate": 0.0001,
      "loss": 1.4708,
      "loss/crossentropy": 2.3727641105651855,
      "loss/hidden": 1.25,
      "loss/logits": 0.2188955843448639,
      "loss/reg": 0.0001904223026940599,
      "step": 149
    },
    {
      "epoch": 0.01875,
      "grad_norm": 2.80842661857605,
      "grad_norm_var": 67.09348312744588,
      "learning_rate": 0.0001,
      "loss": 1.4085,
      "loss/crossentropy": 2.262383460998535,
      "loss/hidden": 1.21875,
      "loss/logits": 0.18781118094921112,
      "loss/reg": 0.0001903773081721738,
      "step": 150
    },
    {
      "epoch": 0.018875,
      "grad_norm": 2.683870553970337,
      "grad_norm_var": 66.87019083886788,
      "learning_rate": 0.0001,
      "loss": 1.2521,
      "loss/crossentropy": 2.5074141025543213,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.16421037912368774,
      "loss/reg": 0.00019033256103284657,
      "step": 151
    },
    {
      "epoch": 0.019,
      "grad_norm": 2.229814291000366,
      "grad_norm_var": 67.0793329518605,
      "learning_rate": 0.0001,
      "loss": 1.1402,
      "loss/crossentropy": 2.541126012802124,
      "loss/hidden": 1.0,
      "loss/logits": 0.1382928341627121,
      "loss/reg": 0.00019028309907298535,
      "step": 152
    },
    {
      "epoch": 0.019125,
      "grad_norm": 3.54891300201416,
      "grad_norm_var": 36.90022117788916,
      "learning_rate": 0.0001,
      "loss": 1.6011,
      "loss/crossentropy": 2.307600975036621,
      "loss/hidden": 1.3671875,
      "loss/logits": 0.23196424543857574,
      "loss/reg": 0.000190236751222983,
      "step": 153
    },
    {
      "epoch": 0.01925,
      "grad_norm": 2.360466718673706,
      "grad_norm_var": 36.95789528091398,
      "learning_rate": 0.0001,
      "loss": 1.1651,
      "loss/crossentropy": 2.8502321243286133,
      "loss/hidden": 0.99609375,
      "loss/logits": 0.16714423894882202,
      "loss/reg": 0.00019020496984012425,
      "step": 154
    },
    {
      "epoch": 0.019375,
      "grad_norm": 3.1953799724578857,
      "grad_norm_var": 35.90550452702474,
      "learning_rate": 0.0001,
      "loss": 1.0914,
      "loss/crossentropy": 2.809048652648926,
      "loss/hidden": 0.9453125,
      "loss/logits": 0.14417339861392975,
      "loss/reg": 0.00019016550504602492,
      "step": 155
    },
    {
      "epoch": 0.0195,
      "grad_norm": 2.6717631816864014,
      "grad_norm_var": 35.890903690685896,
      "learning_rate": 0.0001,
      "loss": 1.2632,
      "loss/crossentropy": 2.7067086696624756,
      "loss/hidden": 1.046875,
      "loss/logits": 0.21447248756885529,
      "loss/reg": 0.00019013263226952404,
      "step": 156
    },
    {
      "epoch": 0.019625,
      "grad_norm": 2.3536930084228516,
      "grad_norm_var": 35.96362699082104,
      "learning_rate": 0.0001,
      "loss": 1.1418,
      "loss/crossentropy": 2.5986528396606445,
      "loss/hidden": 0.98046875,
      "loss/logits": 0.15946394205093384,
      "loss/reg": 0.00019009722745977342,
      "step": 157
    },
    {
      "epoch": 0.01975,
      "grad_norm": 2.0697736740112305,
      "grad_norm_var": 36.1239934744858,
      "learning_rate": 0.0001,
      "loss": 1.1946,
      "loss/crossentropy": 2.4766931533813477,
      "loss/hidden": 1.03125,
      "loss/logits": 0.16143161058425903,
      "loss/reg": 0.0001900633069453761,
      "step": 158
    },
    {
      "epoch": 0.019875,
      "grad_norm": 2.4079439640045166,
      "grad_norm_var": 36.08560176253508,
      "learning_rate": 0.0001,
      "loss": 1.2239,
      "loss/crossentropy": 2.5432755947113037,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.18291215598583221,
      "loss/reg": 0.00019003944180440158,
      "step": 159
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.9956291913986206,
      "grad_norm_var": 36.21688030379835,
      "learning_rate": 0.0001,
      "loss": 1.2381,
      "loss/crossentropy": 2.6434080600738525,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1736569106578827,
      "loss/reg": 0.0001900054921861738,
      "step": 160
    },
    {
      "epoch": 0.020125,
      "grad_norm": 2.3660035133361816,
      "grad_norm_var": 36.268105524765524,
      "learning_rate": 0.0001,
      "loss": 1.3402,
      "loss/crossentropy": 2.7457640171051025,
      "loss/hidden": 1.140625,
      "loss/logits": 0.19771495461463928,
      "loss/reg": 0.00018996208382304758,
      "step": 161
    },
    {
      "epoch": 0.02025,
      "grad_norm": 2.2741127014160156,
      "grad_norm_var": 36.28129668661176,
      "learning_rate": 0.0001,
      "loss": 1.1122,
      "loss/crossentropy": 2.694679021835327,
      "loss/hidden": 0.9453125,
      "loss/logits": 0.16500751674175262,
      "loss/reg": 0.00018992851255461574,
      "step": 162
    },
    {
      "epoch": 0.020375,
      "grad_norm": 4.299161911010742,
      "grad_norm_var": 36.213705020452686,
      "learning_rate": 0.0001,
      "loss": 1.5997,
      "loss/crossentropy": 2.41153621673584,
      "loss/hidden": 1.3515625,
      "loss/logits": 0.24626833200454712,
      "loss/reg": 0.00018988759256899357,
      "step": 163
    },
    {
      "epoch": 0.0205,
      "grad_norm": 2.6989567279815674,
      "grad_norm_var": 0.37062173740110177,
      "learning_rate": 0.0001,
      "loss": 1.4101,
      "loss/crossentropy": 2.6252450942993164,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.2285255789756775,
      "loss/reg": 0.00018985987117048353,
      "step": 164
    },
    {
      "epoch": 0.020625,
      "grad_norm": 3.6266448497772217,
      "grad_norm_var": 0.40754436908909636,
      "learning_rate": 0.0001,
      "loss": 1.3041,
      "loss/crossentropy": 2.768099308013916,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.18500390648841858,
      "loss/reg": 0.00018983366317115724,
      "step": 165
    },
    {
      "epoch": 0.02075,
      "grad_norm": 3.0785396099090576,
      "grad_norm_var": 0.4151303111429129,
      "learning_rate": 0.0001,
      "loss": 1.4192,
      "loss/crossentropy": 2.621722459793091,
      "loss/hidden": 1.203125,
      "loss/logits": 0.21420907974243164,
      "loss/reg": 0.00018979469314217567,
      "step": 166
    },
    {
      "epoch": 0.020875,
      "grad_norm": 2.837559938430786,
      "grad_norm_var": 0.4154299188334017,
      "learning_rate": 0.0001,
      "loss": 1.3458,
      "loss/crossentropy": 2.430216073989868,
      "loss/hidden": 1.1640625,
      "loss/logits": 0.17987295985221863,
      "loss/reg": 0.00018976339197251946,
      "step": 167
    },
    {
      "epoch": 0.021,
      "grad_norm": 2.8974390029907227,
      "grad_norm_var": 0.396902541608811,
      "learning_rate": 0.0001,
      "loss": 1.2952,
      "loss/crossentropy": 2.8876166343688965,
      "loss/hidden": 1.09375,
      "loss/logits": 0.19960089027881622,
      "loss/reg": 0.00018972392717842013,
      "step": 168
    },
    {
      "epoch": 0.021125,
      "grad_norm": 2.533341646194458,
      "grad_norm_var": 0.3589553633283116,
      "learning_rate": 0.0001,
      "loss": 1.2505,
      "loss/crossentropy": 2.7849295139312744,
      "loss/hidden": 1.078125,
      "loss/logits": 0.17050783336162567,
      "loss/reg": 0.0001896892354125157,
      "step": 169
    },
    {
      "epoch": 0.02125,
      "grad_norm": 2.5643837451934814,
      "grad_norm_var": 0.35153012514085774,
      "learning_rate": 0.0001,
      "loss": 1.4785,
      "loss/crossentropy": 2.497753381729126,
      "loss/hidden": 1.234375,
      "loss/logits": 0.2422376275062561,
      "loss/reg": 0.00018963789625559002,
      "step": 170
    },
    {
      "epoch": 0.021375,
      "grad_norm": 2.6845455169677734,
      "grad_norm_var": 0.3369522102595737,
      "learning_rate": 0.0001,
      "loss": 1.2292,
      "loss/crossentropy": 2.5292482376098633,
      "loss/hidden": 1.046875,
      "loss/logits": 0.18040287494659424,
      "loss/reg": 0.00018960374291054904,
      "step": 171
    },
    {
      "epoch": 0.0215,
      "grad_norm": 2.3379464149475098,
      "grad_norm_var": 0.3456172785279794,
      "learning_rate": 0.0001,
      "loss": 1.3626,
      "loss/crossentropy": 2.5603489875793457,
      "loss/hidden": 1.15625,
      "loss/logits": 0.20441675186157227,
      "loss/reg": 0.00018956881831400096,
      "step": 172
    },
    {
      "epoch": 0.021625,
      "grad_norm": 2.594144582748413,
      "grad_norm_var": 0.33847746883165203,
      "learning_rate": 0.0001,
      "loss": 1.2032,
      "loss/crossentropy": 2.6486992835998535,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.17790886759757996,
      "loss/reg": 0.00018953454855363816,
      "step": 173
    },
    {
      "epoch": 0.02175,
      "grad_norm": 2.9685492515563965,
      "grad_norm_var": 0.3129452666235446,
      "learning_rate": 0.0001,
      "loss": 1.4619,
      "loss/crossentropy": 2.455124855041504,
      "loss/hidden": 1.25,
      "loss/logits": 0.21000754833221436,
      "loss/reg": 0.0001895053283078596,
      "step": 174
    },
    {
      "epoch": 0.021875,
      "grad_norm": 2.533724308013916,
      "grad_norm_var": 0.3080246907592235,
      "learning_rate": 0.0001,
      "loss": 1.3506,
      "loss/crossentropy": 2.4606072902679443,
      "loss/hidden": 1.171875,
      "loss/logits": 0.1768466681241989,
      "loss/reg": 0.0001894600281957537,
      "step": 175
    },
    {
      "epoch": 0.022,
      "grad_norm": 2.629793643951416,
      "grad_norm_var": 0.2678377821192363,
      "learning_rate": 0.0001,
      "loss": 1.3554,
      "loss/crossentropy": 2.50494647026062,
      "loss/hidden": 1.140625,
      "loss/logits": 0.2129271924495697,
      "loss/reg": 0.0001894147862913087,
      "step": 176
    },
    {
      "epoch": 0.022125,
      "grad_norm": 2.426023006439209,
      "grad_norm_var": 0.2645273844934496,
      "learning_rate": 0.0001,
      "loss": 1.2205,
      "loss/crossentropy": 2.4932215213775635,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1717696189880371,
      "loss/reg": 0.0001893793960334733,
      "step": 177
    },
    {
      "epoch": 0.02225,
      "grad_norm": 12.438695907592773,
      "grad_norm_var": 5.993566887433531,
      "learning_rate": 0.0001,
      "loss": 1.4933,
      "loss/crossentropy": 2.6111667156219482,
      "loss/hidden": 1.3125,
      "loss/logits": 0.17892810702323914,
      "loss/reg": 0.00018933985847979784,
      "step": 178
    },
    {
      "epoch": 0.022375,
      "grad_norm": 2.2316975593566895,
      "grad_norm_var": 6.025764924701703,
      "learning_rate": 0.0001,
      "loss": 1.2337,
      "loss/crossentropy": 2.452789306640625,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.19278019666671753,
      "loss/reg": 0.0001893048029160127,
      "step": 179
    },
    {
      "epoch": 0.0225,
      "grad_norm": 2.285240650177002,
      "grad_norm_var": 6.070589505634923,
      "learning_rate": 0.0001,
      "loss": 1.3731,
      "loss/crossentropy": 2.4663660526275635,
      "loss/hidden": 1.1875,
      "loss/logits": 0.18373644351959229,
      "loss/reg": 0.0001892724831122905,
      "step": 180
    },
    {
      "epoch": 0.022625,
      "grad_norm": 2.4717960357666016,
      "grad_norm_var": 6.102379780965068,
      "learning_rate": 0.0001,
      "loss": 1.1602,
      "loss/crossentropy": 2.7595460414886475,
      "loss/hidden": 0.9921875,
      "loss/logits": 0.16613003611564636,
      "loss/reg": 0.00018924209871329367,
      "step": 181
    },
    {
      "epoch": 0.02275,
      "grad_norm": 2.4991321563720703,
      "grad_norm_var": 6.134258503662537,
      "learning_rate": 0.0001,
      "loss": 1.1269,
      "loss/crossentropy": 2.270359516143799,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.14847519993782043,
      "loss/reg": 0.00018921452283393592,
      "step": 182
    },
    {
      "epoch": 0.022875,
      "grad_norm": 2.3847415447235107,
      "grad_norm_var": 6.167952691299849,
      "learning_rate": 0.0001,
      "loss": 1.2154,
      "loss/crossentropy": 2.49910831451416,
      "loss/hidden": 1.0546875,
      "loss/logits": 0.158855140209198,
      "loss/reg": 0.00018917533452622592,
      "step": 183
    },
    {
      "epoch": 0.023,
      "grad_norm": 2.5306761264801025,
      "grad_norm_var": 6.188958706490437,
      "learning_rate": 0.0001,
      "loss": 1.4351,
      "loss/crossentropy": 2.3558568954467773,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.2066374272108078,
      "loss/reg": 0.00018914024985861033,
      "step": 184
    },
    {
      "epoch": 0.023125,
      "grad_norm": 2.2277228832244873,
      "grad_norm_var": 6.219197407448101,
      "learning_rate": 0.0001,
      "loss": 1.0945,
      "loss/crossentropy": 2.6959574222564697,
      "loss/hidden": 0.9375,
      "loss/logits": 0.1550855189561844,
      "loss/reg": 0.00018909583741333336,
      "step": 185
    },
    {
      "epoch": 0.02325,
      "grad_norm": 3.3319711685180664,
      "grad_norm_var": 6.199868483198406,
      "learning_rate": 0.0001,
      "loss": 1.2316,
      "loss/crossentropy": 2.3768630027770996,
      "loss/hidden": 1.078125,
      "loss/logits": 0.15154394507408142,
      "loss/reg": 0.00018906217883341014,
      "step": 186
    },
    {
      "epoch": 0.023375,
      "grad_norm": 2.5634615421295166,
      "grad_norm_var": 6.208477354320166,
      "learning_rate": 0.0001,
      "loss": 1.2211,
      "loss/crossentropy": 2.679447650909424,
      "loss/hidden": 1.03125,
      "loss/logits": 0.18797855079174042,
      "loss/reg": 0.00018902822921518236,
      "step": 187
    },
    {
      "epoch": 0.0235,
      "grad_norm": 2.460848093032837,
      "grad_norm_var": 6.196057718240725,
      "learning_rate": 0.0001,
      "loss": 1.2703,
      "loss/crossentropy": 2.6185357570648193,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.18242943286895752,
      "loss/reg": 0.000188985766726546,
      "step": 188
    },
    {
      "epoch": 0.023625,
      "grad_norm": 2.464927911758423,
      "grad_norm_var": 6.206869955671459,
      "learning_rate": 0.0001,
      "loss": 1.1848,
      "loss/crossentropy": 2.4472947120666504,
      "loss/hidden": 1.015625,
      "loss/logits": 0.16724231839179993,
      "loss/reg": 0.00018895140965469182,
      "step": 189
    },
    {
      "epoch": 0.02375,
      "grad_norm": 2.6658761501312256,
      "grad_norm_var": 6.220041941034168,
      "learning_rate": 0.0001,
      "loss": 1.2639,
      "loss/crossentropy": 2.3520965576171875,
      "loss/hidden": 1.09375,
      "loss/logits": 0.1682794690132141,
      "loss/reg": 0.0001889086706796661,
      "step": 190
    },
    {
      "epoch": 0.023875,
      "grad_norm": 2.450937509536743,
      "grad_norm_var": 6.227097887980031,
      "learning_rate": 0.0001,
      "loss": 1.3416,
      "loss/crossentropy": 2.4653656482696533,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.19132134318351746,
      "loss/reg": 0.00018886124598793685,
      "step": 191
    },
    {
      "epoch": 0.024,
      "grad_norm": 2.5486769676208496,
      "grad_norm_var": 6.232908017729262,
      "learning_rate": 0.0001,
      "loss": 1.408,
      "loss/crossentropy": 2.4828147888183594,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.22640305757522583,
      "loss/reg": 0.00018882614676840603,
      "step": 192
    },
    {
      "epoch": 0.024125,
      "grad_norm": 2.733919858932495,
      "grad_norm_var": 6.210183098557901,
      "learning_rate": 0.0001,
      "loss": 1.3418,
      "loss/crossentropy": 2.7277894020080566,
      "loss/hidden": 1.1640625,
      "loss/logits": 0.17583458125591278,
      "loss/reg": 0.00018878061382565647,
      "step": 193
    },
    {
      "epoch": 0.02425,
      "grad_norm": 4.0347490310668945,
      "grad_norm_var": 0.20841963510717557,
      "learning_rate": 0.0001,
      "loss": 1.2188,
      "loss/crossentropy": 2.671088933944702,
      "loss/hidden": 1.0546875,
      "loss/logits": 0.16220757365226746,
      "loss/reg": 0.00018873742374125868,
      "step": 194
    },
    {
      "epoch": 0.024375,
      "grad_norm": 2.1080739498138428,
      "grad_norm_var": 0.21574061631260597,
      "learning_rate": 0.0001,
      "loss": 1.0983,
      "loss/crossentropy": 2.465228319168091,
      "loss/hidden": 0.953125,
      "loss/logits": 0.1432487666606903,
      "loss/reg": 0.00018869389896281064,
      "step": 195
    },
    {
      "epoch": 0.0245,
      "grad_norm": 3.3122878074645996,
      "grad_norm_var": 0.23717126048259918,
      "learning_rate": 0.0001,
      "loss": 1.4535,
      "loss/crossentropy": 2.543299913406372,
      "loss/hidden": 1.2421875,
      "loss/logits": 0.2094225287437439,
      "loss/reg": 0.00018865260062739253,
      "step": 196
    },
    {
      "epoch": 0.024625,
      "grad_norm": 3.8393805027008057,
      "grad_norm_var": 0.3171273295177985,
      "learning_rate": 0.0001,
      "loss": 1.4855,
      "loss/crossentropy": 2.7230522632598877,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.25702908635139465,
      "loss/reg": 0.00018861188436858356,
      "step": 197
    },
    {
      "epoch": 0.02475,
      "grad_norm": 2.3128864765167236,
      "grad_norm_var": 0.32576930180668173,
      "learning_rate": 0.0001,
      "loss": 1.2876,
      "loss/crossentropy": 2.470360517501831,
      "loss/hidden": 1.09375,
      "loss/logits": 0.19192585349082947,
      "loss/reg": 0.0001885706151369959,
      "step": 198
    },
    {
      "epoch": 0.024875,
      "grad_norm": 2.222193956375122,
      "grad_norm_var": 0.33529781396605357,
      "learning_rate": 0.0001,
      "loss": 1.2726,
      "loss/crossentropy": 2.4833052158355713,
      "loss/hidden": 1.078125,
      "loss/logits": 0.1925581395626068,
      "loss/reg": 0.00018852519860956818,
      "step": 199
    },
    {
      "epoch": 0.025,
      "grad_norm": 2.148888111114502,
      "grad_norm_var": 0.35496365745480013,
      "learning_rate": 0.0001,
      "loss": 1.1089,
      "loss/crossentropy": 2.445056915283203,
      "loss/hidden": 0.95703125,
      "loss/logits": 0.14994922280311584,
      "loss/reg": 0.00018848305626306683,
      "step": 200
    },
    {
      "epoch": 0.025125,
      "grad_norm": 3.009246587753296,
      "grad_norm_var": 0.34244750319667006,
      "learning_rate": 0.0001,
      "loss": 1.2831,
      "loss/crossentropy": 2.5913355350494385,
      "loss/hidden": 1.09375,
      "loss/logits": 0.18746912479400635,
      "loss/reg": 0.00018844057922251523,
      "step": 201
    },
    {
      "epoch": 0.02525,
      "grad_norm": 2.6826765537261963,
      "grad_norm_var": 0.31954091153954683,
      "learning_rate": 0.0001,
      "loss": 1.2207,
      "loss/crossentropy": 2.2942862510681152,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.14849132299423218,
      "loss/reg": 0.00018840315169654787,
      "step": 202
    },
    {
      "epoch": 0.025375,
      "grad_norm": 2.089634418487549,
      "grad_norm_var": 0.34361665903956173,
      "learning_rate": 0.0001,
      "loss": 1.2152,
      "loss/crossentropy": 2.565154790878296,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.17425358295440674,
      "loss/reg": 0.00018837135576177388,
      "step": 203
    },
    {
      "epoch": 0.0255,
      "grad_norm": 2.1764748096466064,
      "grad_norm_var": 0.35746666647812214,
      "learning_rate": 0.0001,
      "loss": 1.3562,
      "loss/crossentropy": 2.24812912940979,
      "loss/hidden": 1.140625,
      "loss/logits": 0.21368308365345,
      "loss/reg": 0.00018833015928976238,
      "step": 204
    },
    {
      "epoch": 0.025625,
      "grad_norm": 2.8530831336975098,
      "grad_norm_var": 0.35600843248713165,
      "learning_rate": 0.0001,
      "loss": 1.475,
      "loss/crossentropy": 2.0817508697509766,
      "loss/hidden": 1.2421875,
      "loss/logits": 0.23092345893383026,
      "loss/reg": 0.00018830793851520866,
      "step": 205
    },
    {
      "epoch": 0.02575,
      "grad_norm": 2.663485050201416,
      "grad_norm_var": 0.3560194494934476,
      "learning_rate": 0.0001,
      "loss": 1.1357,
      "loss/crossentropy": 2.7369191646575928,
      "loss/hidden": 0.97265625,
      "loss/logits": 0.1611793339252472,
      "loss/reg": 0.0001882883079815656,
      "step": 206
    },
    {
      "epoch": 0.025875,
      "grad_norm": 2.7267813682556152,
      "grad_norm_var": 0.35164556437612193,
      "learning_rate": 0.0001,
      "loss": 1.4018,
      "loss/crossentropy": 2.521235704421997,
      "loss/hidden": 1.1875,
      "loss/logits": 0.21241608262062073,
      "loss/reg": 0.0001882474316516891,
      "step": 207
    },
    {
      "epoch": 0.026,
      "grad_norm": 3.3730874061584473,
      "grad_norm_var": 0.37568723584622227,
      "learning_rate": 0.0001,
      "loss": 1.6168,
      "loss/crossentropy": 2.67789626121521,
      "loss/hidden": 1.375,
      "loss/logits": 0.2398722767829895,
      "loss/reg": 0.0001882062351796776,
      "step": 208
    },
    {
      "epoch": 0.026125,
      "grad_norm": 2.58087158203125,
      "grad_norm_var": 0.37784520807643934,
      "learning_rate": 0.0001,
      "loss": 1.4245,
      "loss/crossentropy": 2.64697003364563,
      "loss/hidden": 1.21875,
      "loss/logits": 0.2038642168045044,
      "loss/reg": 0.0001881623174995184,
      "step": 209
    },
    {
      "epoch": 0.02625,
      "grad_norm": 2.222271680831909,
      "grad_norm_var": 0.27470612970490343,
      "learning_rate": 0.0001,
      "loss": 1.3277,
      "loss/crossentropy": 2.5434417724609375,
      "loss/hidden": 1.140625,
      "loss/logits": 0.18515345454216003,
      "loss/reg": 0.00018812257621902972,
      "step": 210
    },
    {
      "epoch": 0.026375,
      "grad_norm": 1.9148452281951904,
      "grad_norm_var": 0.29087511560338725,
      "learning_rate": 0.0001,
      "loss": 1.097,
      "loss/crossentropy": 2.769158363342285,
      "loss/hidden": 0.94921875,
      "loss/logits": 0.14593489468097687,
      "loss/reg": 0.0001880890631582588,
      "step": 211
    },
    {
      "epoch": 0.0265,
      "grad_norm": 2.2211508750915527,
      "grad_norm_var": 0.26646107901443655,
      "learning_rate": 0.0001,
      "loss": 1.3446,
      "loss/crossentropy": 2.4616775512695312,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.19431401789188385,
      "loss/reg": 0.00018806415027938783,
      "step": 212
    },
    {
      "epoch": 0.026625,
      "grad_norm": 2.763000965118408,
      "grad_norm_var": 0.15595023444909822,
      "learning_rate": 0.0001,
      "loss": 1.3695,
      "loss/crossentropy": 2.598646640777588,
      "loss/hidden": 1.140625,
      "loss/logits": 0.22699865698814392,
      "loss/reg": 0.00018803446437232196,
      "step": 213
    },
    {
      "epoch": 0.02675,
      "grad_norm": 2.321552276611328,
      "grad_norm_var": 0.15574157634795635,
      "learning_rate": 0.0001,
      "loss": 1.3229,
      "loss/crossentropy": 2.574349880218506,
      "loss/hidden": 1.140625,
      "loss/logits": 0.18040841817855835,
      "loss/reg": 0.00018799355893861502,
      "step": 214
    },
    {
      "epoch": 0.026875,
      "grad_norm": 2.252286195755005,
      "grad_norm_var": 0.15469124462205552,
      "learning_rate": 0.0001,
      "loss": 1.2042,
      "loss/crossentropy": 2.391897439956665,
      "loss/hidden": 1.03125,
      "loss/logits": 0.17103314399719238,
      "loss/reg": 0.00018794478091876954,
      "step": 215
    },
    {
      "epoch": 0.027,
      "grad_norm": 2.9572079181671143,
      "grad_norm_var": 0.15769059669121527,
      "learning_rate": 0.0001,
      "loss": 1.4114,
      "loss/crossentropy": 2.4063875675201416,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.1986089050769806,
      "loss/reg": 0.00018789219029713422,
      "step": 216
    },
    {
      "epoch": 0.027125,
      "grad_norm": 2.3973214626312256,
      "grad_norm_var": 0.14366297343363474,
      "learning_rate": 0.0001,
      "loss": 1.2443,
      "loss/crossentropy": 2.5074877738952637,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1798933446407318,
      "loss/reg": 0.00018785115389619023,
      "step": 217
    },
    {
      "epoch": 0.02725,
      "grad_norm": 3.122760534286499,
      "grad_norm_var": 0.16576884575759981,
      "learning_rate": 0.0001,
      "loss": 1.3388,
      "loss/crossentropy": 2.428725242614746,
      "loss/hidden": 1.140625,
      "loss/logits": 0.19626876711845398,
      "loss/reg": 0.00018780773098114878,
      "step": 218
    },
    {
      "epoch": 0.027375,
      "grad_norm": 2.321108341217041,
      "grad_norm_var": 0.15522596127473184,
      "learning_rate": 0.0001,
      "loss": 1.2505,
      "loss/crossentropy": 2.409878969192505,
      "loss/hidden": 1.046875,
      "loss/logits": 0.2017081379890442,
      "loss/reg": 0.00018775733769871294,
      "step": 219
    },
    {
      "epoch": 0.0275,
      "grad_norm": 2.365797281265259,
      "grad_norm_var": 0.14793109297261797,
      "learning_rate": 0.0001,
      "loss": 1.1702,
      "loss/crossentropy": 2.588578224182129,
      "loss/hidden": 1.015625,
      "loss/logits": 0.15268871188163757,
      "loss/reg": 0.00018771766917780042,
      "step": 220
    },
    {
      "epoch": 0.027625,
      "grad_norm": 2.1108896732330322,
      "grad_norm_var": 0.153953573032076,
      "learning_rate": 0.0001,
      "loss": 1.255,
      "loss/crossentropy": 2.527043581008911,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.16717128455638885,
      "loss/reg": 0.00018766772700473666,
      "step": 221
    },
    {
      "epoch": 0.02775,
      "grad_norm": 2.5774085521698,
      "grad_norm_var": 0.15276588289228218,
      "learning_rate": 0.0001,
      "loss": 1.2479,
      "loss/crossentropy": 2.713797092437744,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1835174411535263,
      "loss/reg": 0.00018762832041829824,
      "step": 222
    },
    {
      "epoch": 0.027875,
      "grad_norm": 2.253450393676758,
      "grad_norm_var": 0.1533568435494087,
      "learning_rate": 0.0001,
      "loss": 1.4177,
      "loss/crossentropy": 2.6281967163085938,
      "loss/hidden": 1.1953125,
      "loss/logits": 0.22053499519824982,
      "loss/reg": 0.0001875831076176837,
      "step": 223
    },
    {
      "epoch": 0.028,
      "grad_norm": 2.454873561859131,
      "grad_norm_var": 0.09728623528139314,
      "learning_rate": 0.0001,
      "loss": 1.4309,
      "loss/crossentropy": 2.572938919067383,
      "loss/hidden": 1.21875,
      "loss/logits": 0.21031853556632996,
      "loss/reg": 0.00018754607299342752,
      "step": 224
    },
    {
      "epoch": 0.028125,
      "grad_norm": 2.402453899383545,
      "grad_norm_var": 0.0956224663481746,
      "learning_rate": 0.0001,
      "loss": 1.2624,
      "loss/crossentropy": 2.50075101852417,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.17460693418979645,
      "loss/reg": 0.0001875244197435677,
      "step": 225
    },
    {
      "epoch": 0.02825,
      "grad_norm": 1.8176156282424927,
      "grad_norm_var": 0.1163170905904891,
      "learning_rate": 0.0001,
      "loss": 1.203,
      "loss/crossentropy": 2.2213640213012695,
      "loss/hidden": 1.03125,
      "loss/logits": 0.16985741257667542,
      "loss/reg": 0.00018748667207546532,
      "step": 226
    },
    {
      "epoch": 0.028375,
      "grad_norm": 2.240485906600952,
      "grad_norm_var": 0.10227683752628369,
      "learning_rate": 0.0001,
      "loss": 1.2129,
      "loss/crossentropy": 2.7936031818389893,
      "loss/hidden": 1.03125,
      "loss/logits": 0.17974941432476044,
      "loss/reg": 0.00018747476860880852,
      "step": 227
    },
    {
      "epoch": 0.0285,
      "grad_norm": 2.478255033493042,
      "grad_norm_var": 0.09989290718763275,
      "learning_rate": 0.0001,
      "loss": 1.3474,
      "loss/crossentropy": 2.237700939178467,
      "loss/hidden": 1.15625,
      "loss/logits": 0.18926523625850677,
      "loss/reg": 0.00018743629334494472,
      "step": 228
    },
    {
      "epoch": 0.028625,
      "grad_norm": 2.2553858757019043,
      "grad_norm_var": 0.09327515190840353,
      "learning_rate": 0.0001,
      "loss": 1.302,
      "loss/crossentropy": 2.5399932861328125,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.1829543262720108,
      "loss/reg": 0.0001873974542832002,
      "step": 229
    },
    {
      "epoch": 0.02875,
      "grad_norm": 2.5666871070861816,
      "grad_norm_var": 0.09461214816090115,
      "learning_rate": 0.0001,
      "loss": 1.1668,
      "loss/crossentropy": 2.5342864990234375,
      "loss/hidden": 1.0,
      "loss/logits": 0.16493582725524902,
      "loss/reg": 0.0001873665169114247,
      "step": 230
    },
    {
      "epoch": 0.028875,
      "grad_norm": 2.3809237480163574,
      "grad_norm_var": 0.09292632453379727,
      "learning_rate": 0.0001,
      "loss": 1.2757,
      "loss/crossentropy": 2.565258026123047,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.2034991830587387,
      "loss/reg": 0.00018733744218479842,
      "step": 231
    },
    {
      "epoch": 0.029,
      "grad_norm": 11.484387397766113,
      "grad_norm_var": 5.249492807099011,
      "learning_rate": 0.0001,
      "loss": 1.2536,
      "loss/crossentropy": 2.610589027404785,
      "loss/hidden": 1.09375,
      "loss/logits": 0.15799307823181152,
      "loss/reg": 0.00018729745352175087,
      "step": 232
    },
    {
      "epoch": 0.029125,
      "grad_norm": 2.276120901107788,
      "grad_norm_var": 5.259372334728978,
      "learning_rate": 0.0001,
      "loss": 1.2413,
      "loss/crossentropy": 2.6497323513031006,
      "loss/hidden": 1.078125,
      "loss/logits": 0.16128413379192352,
      "loss/reg": 0.0001872599241323769,
      "step": 233
    },
    {
      "epoch": 0.02925,
      "grad_norm": 2.5592195987701416,
      "grad_norm_var": 5.265810753770302,
      "learning_rate": 0.0001,
      "loss": 1.3953,
      "loss/crossentropy": 2.6213786602020264,
      "loss/hidden": 1.171875,
      "loss/logits": 0.22152823209762573,
      "loss/reg": 0.0001872186257969588,
      "step": 234
    },
    {
      "epoch": 0.029375,
      "grad_norm": 2.91487455368042,
      "grad_norm_var": 5.241297695369635,
      "learning_rate": 0.0001,
      "loss": 1.4322,
      "loss/crossentropy": 2.8353567123413086,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.2038136124610901,
      "loss/reg": 0.0001871798885986209,
      "step": 235
    },
    {
      "epoch": 0.0295,
      "grad_norm": 2.8779571056365967,
      "grad_norm_var": 5.218058981409527,
      "learning_rate": 0.0001,
      "loss": 1.1964,
      "loss/crossentropy": 2.818354606628418,
      "loss/hidden": 1.03125,
      "loss/logits": 0.16329507529735565,
      "loss/reg": 0.0001871388085419312,
      "step": 236
    },
    {
      "epoch": 0.029625,
      "grad_norm": 2.4912526607513428,
      "grad_norm_var": 5.183116200958803,
      "learning_rate": 0.0001,
      "loss": 1.1612,
      "loss/crossentropy": 2.485246419906616,
      "loss/hidden": 1.0,
      "loss/logits": 0.15928974747657776,
      "loss/reg": 0.0001870945852715522,
      "step": 237
    },
    {
      "epoch": 0.02975,
      "grad_norm": 2.1257100105285645,
      "grad_norm_var": 5.221437379820037,
      "learning_rate": 0.0001,
      "loss": 1.1011,
      "loss/crossentropy": 2.5041284561157227,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.1343885362148285,
      "loss/reg": 0.00018704126705415547,
      "step": 238
    },
    {
      "epoch": 0.029875,
      "grad_norm": 2.609513998031616,
      "grad_norm_var": 5.195165909077186,
      "learning_rate": 0.0001,
      "loss": 1.2685,
      "loss/crossentropy": 2.5717175006866455,
      "loss/hidden": 1.09375,
      "loss/logits": 0.1728888750076294,
      "loss/reg": 0.00018699871725402772,
      "step": 239
    },
    {
      "epoch": 0.03,
      "grad_norm": 3.054771900177002,
      "grad_norm_var": 5.174376919782516,
      "learning_rate": 0.0001,
      "loss": 1.3059,
      "loss/crossentropy": 2.3587000370025635,
      "loss/hidden": 1.109375,
      "loss/logits": 0.19463828206062317,
      "loss/reg": 0.00018695260223466903,
      "step": 240
    },
    {
      "epoch": 0.030125,
      "grad_norm": 4.069124221801758,
      "grad_norm_var": 5.207761360833089,
      "learning_rate": 0.0001,
      "loss": 1.5327,
      "loss/crossentropy": 2.519747257232666,
      "loss/hidden": 1.3046875,
      "loss/logits": 0.22609561681747437,
      "loss/reg": 0.00018690834986045957,
      "step": 241
    },
    {
      "epoch": 0.03025,
      "grad_norm": 2.6919775009155273,
      "grad_norm_var": 5.101652290115754,
      "learning_rate": 0.0001,
      "loss": 1.4243,
      "loss/crossentropy": 2.5264573097229004,
      "loss/hidden": 1.203125,
      "loss/logits": 0.21935243904590607,
      "loss/reg": 0.00018685661780182272,
      "step": 242
    },
    {
      "epoch": 0.030375,
      "grad_norm": 2.9397759437561035,
      "grad_norm_var": 5.043470206735567,
      "learning_rate": 0.0001,
      "loss": 1.1593,
      "loss/crossentropy": 2.649984836578369,
      "loss/hidden": 0.9921875,
      "loss/logits": 0.16523152589797974,
      "loss/reg": 0.00018679779896046966,
      "step": 243
    },
    {
      "epoch": 0.0305,
      "grad_norm": 2.597238063812256,
      "grad_norm_var": 5.032333906433272,
      "learning_rate": 0.0001,
      "loss": 1.3144,
      "loss/crossentropy": 2.7007710933685303,
      "loss/hidden": 1.109375,
      "loss/logits": 0.20317299664020538,
      "loss/reg": 0.00018673852900974452,
      "step": 244
    },
    {
      "epoch": 0.030625,
      "grad_norm": 4.779845237731934,
      "grad_norm_var": 5.0980686958710555,
      "learning_rate": 0.0001,
      "loss": 1.9743,
      "loss/crossentropy": 2.7656006813049316,
      "loss/hidden": 1.5625,
      "loss/logits": 0.4099583327770233,
      "loss/reg": 0.0001866959355538711,
      "step": 245
    },
    {
      "epoch": 0.03075,
      "grad_norm": 3.3142945766448975,
      "grad_norm_var": 5.049814806516395,
      "learning_rate": 0.0001,
      "loss": 1.3097,
      "loss/crossentropy": 2.7472550868988037,
      "loss/hidden": 1.125,
      "loss/logits": 0.1828756332397461,
      "loss/reg": 0.00018664947128854692,
      "step": 246
    },
    {
      "epoch": 0.030875,
      "grad_norm": 2.6288223266601562,
      "grad_norm_var": 5.01838753008573,
      "learning_rate": 0.0001,
      "loss": 1.2715,
      "loss/crossentropy": 2.5212960243225098,
      "loss/hidden": 1.1015625,
      "loss/logits": 0.16810457408428192,
      "loss/reg": 0.00018659804482012987,
      "step": 247
    },
    {
      "epoch": 0.031,
      "grad_norm": 2.1690521240234375,
      "grad_norm_var": 0.4794672993561553,
      "learning_rate": 0.0001,
      "loss": 1.081,
      "loss/crossentropy": 2.517190933227539,
      "loss/hidden": 0.94140625,
      "loss/logits": 0.13774770498275757,
      "loss/reg": 0.0001865396770881489,
      "step": 248
    },
    {
      "epoch": 0.031125,
      "grad_norm": 2.642812490463257,
      "grad_norm_var": 0.45828649220525125,
      "learning_rate": 0.0001,
      "loss": 1.2944,
      "loss/crossentropy": 2.411734104156494,
      "loss/hidden": 1.125,
      "loss/logits": 0.16756606101989746,
      "loss/reg": 0.00018649944104254246,
      "step": 249
    },
    {
      "epoch": 0.03125,
      "grad_norm": 2.553938627243042,
      "grad_norm_var": 0.4585311039907272,
      "learning_rate": 0.0001,
      "loss": 1.1273,
      "loss/crossentropy": 2.9807546138763428,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.1488645374774933,
      "loss/reg": 0.00018645105592440814,
      "step": 250
    },
    {
      "epoch": 0.031375,
      "grad_norm": 2.010117292404175,
      "grad_norm_var": 0.5083579557676783,
      "learning_rate": 0.0001,
      "loss": 1.2306,
      "loss/crossentropy": 2.562990427017212,
      "loss/hidden": 1.046875,
      "loss/logits": 0.18183979392051697,
      "loss/reg": 0.00018639408517628908,
      "step": 251
    },
    {
      "epoch": 0.0315,
      "grad_norm": 2.712961196899414,
      "grad_norm_var": 0.5093841749170441,
      "learning_rate": 0.0001,
      "loss": 1.4763,
      "loss/crossentropy": 2.573420524597168,
      "loss/hidden": 1.2578125,
      "loss/logits": 0.21662938594818115,
      "loss/reg": 0.0001863273064373061,
      "step": 252
    },
    {
      "epoch": 0.031625,
      "grad_norm": 2.6070237159729004,
      "grad_norm_var": 0.504885617842933,
      "learning_rate": 0.0001,
      "loss": 1.1148,
      "loss/crossentropy": 2.421952486038208,
      "loss/hidden": 0.98046875,
      "loss/logits": 0.13251450657844543,
      "loss/reg": 0.00018625622033141553,
      "step": 253
    },
    {
      "epoch": 0.03175,
      "grad_norm": 2.6277847290039062,
      "grad_norm_var": 0.4725433925882648,
      "learning_rate": 0.0001,
      "loss": 1.3581,
      "loss/crossentropy": 2.6136345863342285,
      "loss/hidden": 1.15625,
      "loss/logits": 0.1999516785144806,
      "loss/reg": 0.00018621186609379947,
      "step": 254
    },
    {
      "epoch": 0.031875,
      "grad_norm": 4.751771926879883,
      "grad_norm_var": 0.683379142444816,
      "learning_rate": 0.0001,
      "loss": 1.5569,
      "loss/crossentropy": 2.437589645385742,
      "loss/hidden": 1.3125,
      "loss/logits": 0.2425535023212433,
      "loss/reg": 0.00018616624583955854,
      "step": 255
    },
    {
      "epoch": 0.032,
      "grad_norm": 2.601804494857788,
      "grad_norm_var": 0.693466035829209,
      "learning_rate": 0.0001,
      "loss": 1.477,
      "loss/crossentropy": 2.3009605407714844,
      "loss/hidden": 1.25,
      "loss/logits": 0.22516345977783203,
      "loss/reg": 0.00018612167332321405,
      "step": 256
    },
    {
      "epoch": 0.032125,
      "grad_norm": 2.594728708267212,
      "grad_norm_var": 0.6154499118248385,
      "learning_rate": 0.0001,
      "loss": 1.3481,
      "loss/crossentropy": 2.5490729808807373,
      "loss/hidden": 1.15625,
      "loss/logits": 0.1899745911359787,
      "loss/reg": 0.0001860785996541381,
      "step": 257
    },
    {
      "epoch": 0.03225,
      "grad_norm": 1.9762139320373535,
      "grad_norm_var": 0.666272320547633,
      "learning_rate": 0.0001,
      "loss": 1.1268,
      "loss/crossentropy": 2.6095380783081055,
      "loss/hidden": 0.96875,
      "loss/logits": 0.15619653463363647,
      "loss/reg": 0.00018601951887831092,
      "step": 258
    },
    {
      "epoch": 0.032375,
      "grad_norm": 2.5357439517974854,
      "grad_norm_var": 0.6713294887447486,
      "learning_rate": 0.0001,
      "loss": 1.1937,
      "loss/crossentropy": 2.7029552459716797,
      "loss/hidden": 1.03125,
      "loss/logits": 0.1605946123600006,
      "loss/reg": 0.00018597528105601668,
      "step": 259
    },
    {
      "epoch": 0.0325,
      "grad_norm": 2.8872361183166504,
      "grad_norm_var": 0.6680105601783904,
      "learning_rate": 0.0001,
      "loss": 1.1618,
      "loss/crossentropy": 2.5793302059173584,
      "loss/hidden": 1.0,
      "loss/logits": 0.15989942848682404,
      "loss/reg": 0.00018592287960927933,
      "step": 260
    },
    {
      "epoch": 0.032625,
      "grad_norm": 2.6888515949249268,
      "grad_norm_var": 0.39965034448394665,
      "learning_rate": 0.0001,
      "loss": 1.1163,
      "loss/crossentropy": 2.501629114151001,
      "loss/hidden": 0.9609375,
      "loss/logits": 0.15351587533950806,
      "loss/reg": 0.00018586948863230646,
      "step": 261
    },
    {
      "epoch": 0.03275,
      "grad_norm": 2.794921398162842,
      "grad_norm_var": 0.3744163537172878,
      "learning_rate": 0.0001,
      "loss": 1.3868,
      "loss/crossentropy": 2.473355293273926,
      "loss/hidden": 1.1875,
      "loss/logits": 0.19742509722709656,
      "loss/reg": 0.00018582609482109547,
      "step": 262
    },
    {
      "epoch": 0.032875,
      "grad_norm": 11.51866340637207,
      "grad_norm_var": 5.260212315476766,
      "learning_rate": 0.0001,
      "loss": 1.9065,
      "loss/crossentropy": 2.5873212814331055,
      "loss/hidden": 1.4765625,
      "loss/logits": 0.4280346632003784,
      "loss/reg": 0.00018578370509203523,
      "step": 263
    },
    {
      "epoch": 0.033,
      "grad_norm": 3.3071396350860596,
      "grad_norm_var": 5.180231931586934,
      "learning_rate": 0.0001,
      "loss": 1.237,
      "loss/crossentropy": 2.6701910495758057,
      "loss/hidden": 1.0625,
      "loss/logits": 0.17262707650661469,
      "loss/reg": 0.0001857366441981867,
      "step": 264
    },
    {
      "epoch": 0.033125,
      "grad_norm": 2.6719601154327393,
      "grad_norm_var": 5.177728124810292,
      "learning_rate": 0.0001,
      "loss": 1.4876,
      "loss/crossentropy": 2.741184711456299,
      "loss/hidden": 1.265625,
      "loss/logits": 0.22007906436920166,
      "loss/reg": 0.00018568705127108842,
      "step": 265
    },
    {
      "epoch": 0.03325,
      "grad_norm": 4.4371514320373535,
      "grad_norm_var": 5.211410221157409,
      "learning_rate": 0.0001,
      "loss": 1.8731,
      "loss/crossentropy": 2.4437432289123535,
      "loss/hidden": 1.6015625,
      "loss/logits": 0.269656777381897,
      "loss/reg": 0.00018562644254416227,
      "step": 266
    },
    {
      "epoch": 0.033375,
      "grad_norm": 3.256206750869751,
      "grad_norm_var": 5.074168773112569,
      "learning_rate": 0.0001,
      "loss": 1.3223,
      "loss/crossentropy": 2.6366071701049805,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.1875854730606079,
      "loss/reg": 0.00018558187002781779,
      "step": 267
    },
    {
      "epoch": 0.0335,
      "grad_norm": 2.422538995742798,
      "grad_norm_var": 5.109844600456275,
      "learning_rate": 0.0001,
      "loss": 1.3075,
      "loss/crossentropy": 2.5296108722686768,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.1884431093931198,
      "loss/reg": 0.00018553413974586874,
      "step": 268
    },
    {
      "epoch": 0.033625,
      "grad_norm": 4.04595422744751,
      "grad_norm_var": 5.071768309380559,
      "learning_rate": 0.0001,
      "loss": 1.8971,
      "loss/crossentropy": 2.3316025733947754,
      "loss/hidden": 1.625,
      "loss/logits": 0.2702447175979614,
      "loss/reg": 0.00018548894149716944,
      "step": 269
    },
    {
      "epoch": 0.03375,
      "grad_norm": 3.397867441177368,
      "grad_norm_var": 5.012096554664663,
      "learning_rate": 0.0001,
      "loss": 1.43,
      "loss/crossentropy": 2.2132034301757812,
      "loss/hidden": 1.234375,
      "loss/logits": 0.19377078115940094,
      "loss/reg": 0.00018543725309427828,
      "step": 270
    },
    {
      "epoch": 0.033875,
      "grad_norm": 2.360687494277954,
      "grad_norm_var": 5.007982625032073,
      "learning_rate": 0.0001,
      "loss": 1.2015,
      "loss/crossentropy": 2.6397757530212402,
      "loss/hidden": 1.03125,
      "loss/logits": 0.1683761179447174,
      "loss/reg": 0.00018539318989496678,
      "step": 271
    },
    {
      "epoch": 0.034,
      "grad_norm": 3.795145273208618,
      "grad_norm_var": 4.95906816389108,
      "learning_rate": 0.0001,
      "loss": 1.7341,
      "loss/crossentropy": 2.2904622554779053,
      "loss/hidden": 1.5,
      "loss/logits": 0.23220419883728027,
      "loss/reg": 0.000185342927579768,
      "step": 272
    },
    {
      "epoch": 0.034125,
      "grad_norm": 3.19511079788208,
      "grad_norm_var": 4.905671754488361,
      "learning_rate": 0.0001,
      "loss": 1.4552,
      "loss/crossentropy": 2.4893336296081543,
      "loss/hidden": 1.25,
      "loss/logits": 0.20336630940437317,
      "loss/reg": 0.00018529384396970272,
      "step": 273
    },
    {
      "epoch": 0.03425,
      "grad_norm": 2.818743944168091,
      "grad_norm_var": 4.769792764968277,
      "learning_rate": 0.0001,
      "loss": 1.1492,
      "loss/crossentropy": 2.5468082427978516,
      "loss/hidden": 0.9921875,
      "loss/logits": 0.1551104187965393,
      "loss/reg": 0.00018524785991758108,
      "step": 274
    },
    {
      "epoch": 0.034375,
      "grad_norm": 2.7294118404388428,
      "grad_norm_var": 4.743793641432837,
      "learning_rate": 0.0001,
      "loss": 1.2939,
      "loss/crossentropy": 2.7086288928985596,
      "loss/hidden": 1.109375,
      "loss/logits": 0.18263299763202667,
      "loss/reg": 0.00018520389858167619,
      "step": 275
    },
    {
      "epoch": 0.0345,
      "grad_norm": 4.2614545822143555,
      "grad_norm_var": 4.722892075276445,
      "learning_rate": 0.0001,
      "loss": 2.152,
      "loss/crossentropy": 2.3208584785461426,
      "loss/hidden": 1.7578125,
      "loss/logits": 0.3923119604587555,
      "loss/reg": 0.0001851657871156931,
      "step": 276
    },
    {
      "epoch": 0.034625,
      "grad_norm": 3.1579630374908447,
      "grad_norm_var": 4.671438964356158,
      "learning_rate": 0.0001,
      "loss": 1.5186,
      "loss/crossentropy": 2.52303409576416,
      "loss/hidden": 1.28125,
      "loss/logits": 0.23553629219532013,
      "loss/reg": 0.00018512809765525162,
      "step": 277
    },
    {
      "epoch": 0.03475,
      "grad_norm": 2.9455530643463135,
      "grad_norm_var": 4.653460522047108,
      "learning_rate": 0.0001,
      "loss": 1.2198,
      "loss/crossentropy": 2.64249324798584,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1554383635520935,
      "loss/reg": 0.00018509359506424516,
      "step": 278
    },
    {
      "epoch": 0.034875,
      "grad_norm": 2.588207721710205,
      "grad_norm_var": 0.4115949243025471,
      "learning_rate": 0.0001,
      "loss": 1.3877,
      "loss/crossentropy": 2.829423427581787,
      "loss/hidden": 1.171875,
      "loss/logits": 0.2140059471130371,
      "loss/reg": 0.00018506502965465188,
      "step": 279
    },
    {
      "epoch": 0.035,
      "grad_norm": 3.4081485271453857,
      "grad_norm_var": 0.41351468625660626,
      "learning_rate": 0.0001,
      "loss": 1.436,
      "loss/crossentropy": 2.624629259109497,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.2076188027858734,
      "loss/reg": 0.00018502252351026982,
      "step": 280
    },
    {
      "epoch": 0.035125,
      "grad_norm": 3.1615681648254395,
      "grad_norm_var": 0.39283411950296265,
      "learning_rate": 0.0001,
      "loss": 1.3441,
      "loss/crossentropy": 2.465009927749634,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.19385108351707458,
      "loss/reg": 0.00018499059660825878,
      "step": 281
    },
    {
      "epoch": 0.03525,
      "grad_norm": 3.2208797931671143,
      "grad_norm_var": 0.2925862508398,
      "learning_rate": 0.0001,
      "loss": 1.9994,
      "loss/crossentropy": 2.230750322341919,
      "loss/hidden": 1.578125,
      "loss/logits": 0.41937941312789917,
      "loss/reg": 0.00018494235700927675,
      "step": 282
    },
    {
      "epoch": 0.035375,
      "grad_norm": 3.2623984813690186,
      "grad_norm_var": 0.292657471443624,
      "learning_rate": 0.0001,
      "loss": 1.5126,
      "loss/crossentropy": 2.5379672050476074,
      "loss/hidden": 1.2734375,
      "loss/logits": 0.23736032843589783,
      "loss/reg": 0.00018488496425561607,
      "step": 283
    },
    {
      "epoch": 0.0355,
      "grad_norm": 2.368196487426758,
      "grad_norm_var": 0.2982812772165668,
      "learning_rate": 0.0001,
      "loss": 1.3005,
      "loss/crossentropy": 2.516019344329834,
      "loss/hidden": 1.109375,
      "loss/logits": 0.18929743766784668,
      "loss/reg": 0.0001848430110840127,
      "step": 284
    },
    {
      "epoch": 0.035625,
      "grad_norm": 3.191399335861206,
      "grad_norm_var": 0.24409669271128276,
      "learning_rate": 0.0001,
      "loss": 1.2823,
      "loss/crossentropy": 2.751084089279175,
      "loss/hidden": 1.109375,
      "loss/logits": 0.17108574509620667,
      "loss/reg": 0.00018479586287867278,
      "step": 285
    },
    {
      "epoch": 0.03575,
      "grad_norm": 2.592566967010498,
      "grad_norm_var": 0.2544086356402346,
      "learning_rate": 0.0001,
      "loss": 1.3514,
      "loss/crossentropy": 2.5993878841400146,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.23240423202514648,
      "loss/reg": 0.0001847518578870222,
      "step": 286
    },
    {
      "epoch": 0.035875,
      "grad_norm": 3.8178586959838867,
      "grad_norm_var": 0.2500656389811674,
      "learning_rate": 0.0001,
      "loss": 1.3704,
      "loss/crossentropy": 2.688004970550537,
      "loss/hidden": 1.1875,
      "loss/logits": 0.18104243278503418,
      "loss/reg": 0.00018470516079105437,
      "step": 287
    },
    {
      "epoch": 0.036,
      "grad_norm": 2.0925023555755615,
      "grad_norm_var": 0.2864185440912022,
      "learning_rate": 0.0001,
      "loss": 1.2565,
      "loss/crossentropy": 2.4902563095092773,
      "loss/hidden": 1.09375,
      "loss/logits": 0.16092851758003235,
      "loss/reg": 0.0001846601371653378,
      "step": 288
    },
    {
      "epoch": 0.036125,
      "grad_norm": 2.821599245071411,
      "grad_norm_var": 0.28794847130561624,
      "learning_rate": 0.0001,
      "loss": 1.5127,
      "loss/crossentropy": 2.331846237182617,
      "loss/hidden": 1.2578125,
      "loss/logits": 0.25302040576934814,
      "loss/reg": 0.00018461896979715675,
      "step": 289
    },
    {
      "epoch": 0.03625,
      "grad_norm": 2.387387275695801,
      "grad_norm_var": 0.31157863588131224,
      "learning_rate": 0.0001,
      "loss": 1.2607,
      "loss/crossentropy": 2.76906681060791,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.18852418661117554,
      "loss/reg": 0.00018456355610396713,
      "step": 290
    },
    {
      "epoch": 0.036375,
      "grad_norm": 3.449770450592041,
      "grad_norm_var": 0.3179789227700856,
      "learning_rate": 0.0001,
      "loss": 1.7591,
      "loss/crossentropy": 1.8213207721710205,
      "loss/hidden": 1.5390625,
      "loss/logits": 0.21822357177734375,
      "loss/reg": 0.00018452556105330586,
      "step": 291
    },
    {
      "epoch": 0.0365,
      "grad_norm": 2.499556303024292,
      "grad_norm_var": 0.2263369840310491,
      "learning_rate": 0.0001,
      "loss": 1.2925,
      "loss/crossentropy": 2.6276936531066895,
      "loss/hidden": 1.109375,
      "loss/logits": 0.18123921751976013,
      "loss/reg": 0.0001844725920818746,
      "step": 292
    },
    {
      "epoch": 0.036625,
      "grad_norm": 2.115734815597534,
      "grad_norm_var": 0.2632914348592083,
      "learning_rate": 0.0001,
      "loss": 1.2089,
      "loss/crossentropy": 2.6672511100769043,
      "loss/hidden": 1.03125,
      "loss/logits": 0.1757870316505432,
      "loss/reg": 0.0001844176003942266,
      "step": 293
    },
    {
      "epoch": 0.03675,
      "grad_norm": 2.684098958969116,
      "grad_norm_var": 0.2649372545619843,
      "learning_rate": 0.0001,
      "loss": 1.2995,
      "loss/crossentropy": 2.5663037300109863,
      "loss/hidden": 1.125,
      "loss/logits": 0.172622948884964,
      "loss/reg": 0.00018438571714796126,
      "step": 294
    },
    {
      "epoch": 0.036875,
      "grad_norm": 7.823673248291016,
      "grad_norm_var": 1.792621724898212,
      "learning_rate": 0.0001,
      "loss": 1.3821,
      "loss/crossentropy": 2.6714284420013428,
      "loss/hidden": 1.15625,
      "loss/logits": 0.22403180599212646,
      "loss/reg": 0.00018433824880048633,
      "step": 295
    },
    {
      "epoch": 0.037,
      "grad_norm": 2.6135120391845703,
      "grad_norm_var": 1.808029304785425,
      "learning_rate": 0.0001,
      "loss": 1.2104,
      "loss/crossentropy": 2.68971848487854,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.16948801279067993,
      "loss/reg": 0.00018430246564093977,
      "step": 296
    },
    {
      "epoch": 0.037125,
      "grad_norm": 8.026154518127441,
      "grad_norm_var": 3.3065969805558173,
      "learning_rate": 0.0001,
      "loss": 2.1847,
      "loss/crossentropy": 2.3800106048583984,
      "loss/hidden": 1.9453125,
      "loss/logits": 0.2375316321849823,
      "loss/reg": 0.0001842674391809851,
      "step": 297
    },
    {
      "epoch": 0.03725,
      "grad_norm": 2.896569013595581,
      "grad_norm_var": 3.3224491377570446,
      "learning_rate": 0.0001,
      "loss": 1.3748,
      "loss/crossentropy": 2.3903732299804688,
      "loss/hidden": 1.1953125,
      "loss/logits": 0.17761950194835663,
      "loss/reg": 0.00018421628919895738,
      "step": 298
    },
    {
      "epoch": 0.037375,
      "grad_norm": 3.096365451812744,
      "grad_norm_var": 3.327554446166753,
      "learning_rate": 0.0001,
      "loss": 1.4918,
      "loss/crossentropy": 2.355905532836914,
      "loss/hidden": 1.2734375,
      "loss/logits": 0.2165393829345703,
      "loss/reg": 0.00018417388491798192,
      "step": 299
    },
    {
      "epoch": 0.0375,
      "grad_norm": 5.26882266998291,
      "grad_norm_var": 3.452496381081827,
      "learning_rate": 0.0001,
      "loss": 1.9899,
      "loss/crossentropy": 2.9516849517822266,
      "loss/hidden": 1.6875,
      "loss/logits": 0.3005185127258301,
      "loss/reg": 0.0001841239572968334,
      "step": 300
    },
    {
      "epoch": 0.037625,
      "grad_norm": 3.068439483642578,
      "grad_norm_var": 3.4599122750924143,
      "learning_rate": 0.0001,
      "loss": 1.3886,
      "loss/crossentropy": 2.6546504497528076,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.20704253017902374,
      "loss/reg": 0.00018408475443720818,
      "step": 301
    },
    {
      "epoch": 0.03775,
      "grad_norm": 2.2956972122192383,
      "grad_norm_var": 3.504442894615314,
      "learning_rate": 0.0001,
      "loss": 1.1094,
      "loss/crossentropy": 2.4977052211761475,
      "loss/hidden": 0.96875,
      "loss/logits": 0.13884802162647247,
      "loss/reg": 0.00018404283036943525,
      "step": 302
    },
    {
      "epoch": 0.037875,
      "grad_norm": 2.5937392711639404,
      "grad_norm_var": 3.5559874858295015,
      "learning_rate": 0.0001,
      "loss": 1.3949,
      "loss/crossentropy": 2.379765033721924,
      "loss/hidden": 1.21875,
      "loss/logits": 0.17428620159626007,
      "loss/reg": 0.00018399336840957403,
      "step": 303
    },
    {
      "epoch": 0.038,
      "grad_norm": 3.512799024581909,
      "grad_norm_var": 3.4186760491291084,
      "learning_rate": 0.0001,
      "loss": 1.381,
      "loss/crossentropy": 2.4689581394195557,
      "loss/hidden": 1.1640625,
      "loss/logits": 0.2151029407978058,
      "loss/reg": 0.00018394750077277422,
      "step": 304
    },
    {
      "epoch": 0.038125,
      "grad_norm": 2.610177755355835,
      "grad_norm_var": 3.442626566538619,
      "learning_rate": 0.0001,
      "loss": 1.2278,
      "loss/crossentropy": 2.4508001804351807,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1634678691625595,
      "loss/reg": 0.00018390185141470283,
      "step": 305
    },
    {
      "epoch": 0.03825,
      "grad_norm": 2.4446496963500977,
      "grad_norm_var": 3.433886969311868,
      "learning_rate": 0.0001,
      "loss": 1.5023,
      "loss/crossentropy": 2.110161066055298,
      "loss/hidden": 1.3125,
      "loss/logits": 0.18798431754112244,
      "loss/reg": 0.00018385711882729083,
      "step": 306
    },
    {
      "epoch": 0.038375,
      "grad_norm": 2.9203567504882812,
      "grad_norm_var": 3.4593607482629065,
      "learning_rate": 0.0001,
      "loss": 1.2248,
      "loss/crossentropy": 2.6589579582214355,
      "loss/hidden": 1.0625,
      "loss/logits": 0.16043730080127716,
      "loss/reg": 0.00018381779955234379,
      "step": 307
    },
    {
      "epoch": 0.0385,
      "grad_norm": 3.625650644302368,
      "grad_norm_var": 3.3839899608280857,
      "learning_rate": 0.0001,
      "loss": 1.3888,
      "loss/crossentropy": 2.7281792163848877,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.269817590713501,
      "loss/reg": 0.00018376816296949983,
      "step": 308
    },
    {
      "epoch": 0.038625,
      "grad_norm": 2.889061689376831,
      "grad_norm_var": 3.268347098659014,
      "learning_rate": 0.0001,
      "loss": 1.1888,
      "loss/crossentropy": 2.523449420928955,
      "loss/hidden": 1.0078125,
      "loss/logits": 0.17915961146354675,
      "loss/reg": 0.00018371775513514876,
      "step": 309
    },
    {
      "epoch": 0.03875,
      "grad_norm": 2.768836736679077,
      "grad_norm_var": 3.257904120325863,
      "learning_rate": 0.0001,
      "loss": 1.3947,
      "loss/crossentropy": 2.3002350330352783,
      "loss/hidden": 1.1875,
      "loss/logits": 0.20538243651390076,
      "loss/reg": 0.00018366229778621346,
      "step": 310
    },
    {
      "epoch": 0.038875,
      "grad_norm": 2.4714229106903076,
      "grad_norm_var": 2.0722741056598655,
      "learning_rate": 0.0001,
      "loss": 1.3588,
      "loss/crossentropy": 2.5036730766296387,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.2085750699043274,
      "loss/reg": 0.00018361561524216086,
      "step": 311
    },
    {
      "epoch": 0.039,
      "grad_norm": 2.3803367614746094,
      "grad_norm_var": 2.0976025308533433,
      "learning_rate": 0.0001,
      "loss": 1.3446,
      "loss/crossentropy": 2.5774199962615967,
      "loss/hidden": 1.15625,
      "loss/logits": 0.18654365837574005,
      "loss/reg": 0.00018356091459281743,
      "step": 312
    },
    {
      "epoch": 0.039125,
      "grad_norm": 2.1344940662384033,
      "grad_norm_var": 0.557820051408968,
      "learning_rate": 0.0001,
      "loss": 1.1236,
      "loss/crossentropy": 2.492638349533081,
      "loss/hidden": 0.96875,
      "loss/logits": 0.15305551886558533,
      "loss/reg": 0.00018350353639107198,
      "step": 313
    },
    {
      "epoch": 0.03925,
      "grad_norm": 2.3380954265594482,
      "grad_norm_var": 0.5802561079704179,
      "learning_rate": 0.0001,
      "loss": 1.2665,
      "loss/crossentropy": 2.538328170776367,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.1786879152059555,
      "loss/reg": 0.00018345742137171328,
      "step": 314
    },
    {
      "epoch": 0.039375,
      "grad_norm": 2.4094135761260986,
      "grad_norm_var": 0.5918726782285414,
      "learning_rate": 0.0001,
      "loss": 1.3272,
      "loss/crossentropy": 2.4620587825775146,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.1925506591796875,
      "loss/reg": 0.00018340005772188306,
      "step": 315
    },
    {
      "epoch": 0.0395,
      "grad_norm": 2.4094653129577637,
      "grad_norm_var": 0.18384264866255365,
      "learning_rate": 0.0001,
      "loss": 1.6038,
      "loss/crossentropy": 2.343863010406494,
      "loss/hidden": 1.3125,
      "loss/logits": 0.28949666023254395,
      "loss/reg": 0.00018334249034523964,
      "step": 316
    },
    {
      "epoch": 0.039625,
      "grad_norm": 2.327509880065918,
      "grad_norm_var": 0.17973404957112868,
      "learning_rate": 0.0001,
      "loss": 1.2969,
      "loss/crossentropy": 2.6910533905029297,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.17789438366889954,
      "loss/reg": 0.0001832786510931328,
      "step": 317
    },
    {
      "epoch": 0.03975,
      "grad_norm": 1.9041651487350464,
      "grad_norm_var": 0.2069358760498758,
      "learning_rate": 0.0001,
      "loss": 1.1054,
      "loss/crossentropy": 2.636301279067993,
      "loss/hidden": 0.96875,
      "loss/logits": 0.13483907282352448,
      "loss/reg": 0.00018322949472349137,
      "step": 318
    },
    {
      "epoch": 0.039875,
      "grad_norm": 2.493028163909912,
      "grad_norm_var": 0.2077715093556349,
      "learning_rate": 0.0001,
      "loss": 1.3427,
      "loss/crossentropy": 2.3860602378845215,
      "loss/hidden": 1.15625,
      "loss/logits": 0.18460425734519958,
      "loss/reg": 0.00018318284128326923,
      "step": 319
    },
    {
      "epoch": 0.04,
      "grad_norm": 2.566598892211914,
      "grad_norm_var": 0.14887985654726912,
      "learning_rate": 0.0001,
      "loss": 1.1799,
      "loss/crossentropy": 2.562800884246826,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.15464608371257782,
      "loss/reg": 0.00018314005865249783,
      "step": 320
    },
    {
      "epoch": 0.040125,
      "grad_norm": 2.3247175216674805,
      "grad_norm_var": 0.15142847186754452,
      "learning_rate": 0.0001,
      "loss": 1.138,
      "loss/crossentropy": 2.6008517742156982,
      "loss/hidden": 0.97265625,
      "loss/logits": 0.16352114081382751,
      "loss/reg": 0.00018309340521227568,
      "step": 321
    },
    {
      "epoch": 0.04025,
      "grad_norm": 2.961496114730835,
      "grad_norm_var": 0.1625533330376787,
      "learning_rate": 0.0001,
      "loss": 1.4689,
      "loss/crossentropy": 1.9118317365646362,
      "loss/hidden": 1.296875,
      "loss/logits": 0.17024032771587372,
      "loss/reg": 0.00018305043340660632,
      "step": 322
    },
    {
      "epoch": 0.040375,
      "grad_norm": 3.0667834281921387,
      "grad_norm_var": 0.17097196220394337,
      "learning_rate": 0.0001,
      "loss": 1.7235,
      "loss/crossentropy": 2.053375482559204,
      "loss/hidden": 1.421875,
      "loss/logits": 0.2998199462890625,
      "loss/reg": 0.00018301077943760902,
      "step": 323
    },
    {
      "epoch": 0.0405,
      "grad_norm": 2.930929183959961,
      "grad_norm_var": 0.10306917410381727,
      "learning_rate": 0.0001,
      "loss": 1.2964,
      "loss/crossentropy": 2.5874621868133545,
      "loss/hidden": 1.1015625,
      "loss/logits": 0.19297093152999878,
      "loss/reg": 0.00018295719928573817,
      "step": 324
    },
    {
      "epoch": 0.040625,
      "grad_norm": 2.871757745742798,
      "grad_norm_var": 0.1022445182394837,
      "learning_rate": 0.0001,
      "loss": 1.2449,
      "loss/crossentropy": 2.6085433959960938,
      "loss/hidden": 1.078125,
      "loss/logits": 0.16499356925487518,
      "loss/reg": 0.0001828953973017633,
      "step": 325
    },
    {
      "epoch": 0.04075,
      "grad_norm": 2.489987850189209,
      "grad_norm_var": 0.09794334325425678,
      "learning_rate": 0.0001,
      "loss": 1.3676,
      "loss/crossentropy": 2.4801414012908936,
      "loss/hidden": 1.171875,
      "loss/logits": 0.1939350813627243,
      "loss/reg": 0.00018283820827491581,
      "step": 326
    },
    {
      "epoch": 0.040875,
      "grad_norm": 2.536003828048706,
      "grad_norm_var": 0.09791477775173399,
      "learning_rate": 0.0001,
      "loss": 1.3227,
      "loss/crossentropy": 2.7262425422668457,
      "loss/hidden": 1.109375,
      "loss/logits": 0.21148526668548584,
      "loss/reg": 0.0001827785454224795,
      "step": 327
    },
    {
      "epoch": 0.041,
      "grad_norm": 2.0945851802825928,
      "grad_norm_var": 0.10792211144253508,
      "learning_rate": 0.0001,
      "loss": 1.1186,
      "loss/crossentropy": 2.4119601249694824,
      "loss/hidden": 0.97265625,
      "loss/logits": 0.1441338062286377,
      "loss/reg": 0.0001827288360800594,
      "step": 328
    },
    {
      "epoch": 0.041125,
      "grad_norm": 2.836939811706543,
      "grad_norm_var": 0.10535360002502277,
      "learning_rate": 0.0001,
      "loss": 1.4351,
      "loss/crossentropy": 2.3377010822296143,
      "loss/hidden": 1.21875,
      "loss/logits": 0.21447591483592987,
      "loss/reg": 0.00018267397535964847,
      "step": 329
    },
    {
      "epoch": 0.04125,
      "grad_norm": 2.326241970062256,
      "grad_norm_var": 0.10567372742739559,
      "learning_rate": 0.0001,
      "loss": 1.1899,
      "loss/crossentropy": 2.7874836921691895,
      "loss/hidden": 1.015625,
      "loss/logits": 0.17246156930923462,
      "loss/reg": 0.00018262714729644358,
      "step": 330
    },
    {
      "epoch": 0.041375,
      "grad_norm": 2.6195242404937744,
      "grad_norm_var": 0.10493277845914116,
      "learning_rate": 0.0001,
      "loss": 1.4583,
      "loss/crossentropy": 2.6836705207824707,
      "loss/hidden": 1.25,
      "loss/logits": 0.20644301176071167,
      "loss/reg": 0.0001825743674999103,
      "step": 331
    },
    {
      "epoch": 0.0415,
      "grad_norm": 2.837834119796753,
      "grad_norm_var": 0.10851849947722905,
      "learning_rate": 0.0001,
      "loss": 1.3683,
      "loss/crossentropy": 2.336994171142578,
      "loss/hidden": 1.15625,
      "loss/logits": 0.21025338768959045,
      "loss/reg": 0.0001825322542572394,
      "step": 332
    },
    {
      "epoch": 0.041625,
      "grad_norm": 2.7226598262786865,
      "grad_norm_var": 0.10527721486086339,
      "learning_rate": 0.0001,
      "loss": 1.1225,
      "loss/crossentropy": 2.621798276901245,
      "loss/hidden": 0.97265625,
      "loss/logits": 0.1480618715286255,
      "loss/reg": 0.00018248235573992133,
      "step": 333
    },
    {
      "epoch": 0.04175,
      "grad_norm": 1.9014396667480469,
      "grad_norm_var": 0.10553016347722419,
      "learning_rate": 0.0001,
      "loss": 1.121,
      "loss/crossentropy": 2.642059087753296,
      "loss/hidden": 0.96875,
      "loss/logits": 0.15041938424110413,
      "loss/reg": 0.00018242868827655911,
      "step": 334
    },
    {
      "epoch": 0.041875,
      "grad_norm": 2.2809882164001465,
      "grad_norm_var": 0.11133012136604989,
      "learning_rate": 0.0001,
      "loss": 1.3509,
      "loss/crossentropy": 2.59220027923584,
      "loss/hidden": 1.140625,
      "loss/logits": 0.20842507481575012,
      "loss/reg": 0.00018237272161059082,
      "step": 335
    },
    {
      "epoch": 0.042,
      "grad_norm": 2.722670793533325,
      "grad_norm_var": 0.11245856535336524,
      "learning_rate": 0.0001,
      "loss": 1.3534,
      "loss/crossentropy": 2.7409770488739014,
      "loss/hidden": 1.15625,
      "loss/logits": 0.195334792137146,
      "loss/reg": 0.00018231497961096466,
      "step": 336
    },
    {
      "epoch": 0.042125,
      "grad_norm": 2.7385120391845703,
      "grad_norm_var": 0.1082322741184413,
      "learning_rate": 0.0001,
      "loss": 1.2894,
      "loss/crossentropy": 2.5753767490386963,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.17039254307746887,
      "loss/reg": 0.00018226687097921968,
      "step": 337
    },
    {
      "epoch": 0.04225,
      "grad_norm": 2.278662919998169,
      "grad_norm_var": 0.106386719047498,
      "learning_rate": 0.0001,
      "loss": 1.4188,
      "loss/crossentropy": 2.519120693206787,
      "loss/hidden": 1.203125,
      "loss/logits": 0.21385297179222107,
      "loss/reg": 0.00018221234495285898,
      "step": 338
    },
    {
      "epoch": 0.042375,
      "grad_norm": 2.7543537616729736,
      "grad_norm_var": 0.09214567617970199,
      "learning_rate": 0.0001,
      "loss": 1.2684,
      "loss/crossentropy": 2.615980625152588,
      "loss/hidden": 1.09375,
      "loss/logits": 0.17285585403442383,
      "loss/reg": 0.00018216087482869625,
      "step": 339
    },
    {
      "epoch": 0.0425,
      "grad_norm": 2.740844249725342,
      "grad_norm_var": 0.08497608623963113,
      "learning_rate": 0.0001,
      "loss": 1.3717,
      "loss/crossentropy": 2.3156495094299316,
      "loss/hidden": 1.1875,
      "loss/logits": 0.18235422670841217,
      "loss/reg": 0.000182102681719698,
      "step": 340
    },
    {
      "epoch": 0.042625,
      "grad_norm": 2.183875799179077,
      "grad_norm_var": 0.08476970381204461,
      "learning_rate": 0.0001,
      "loss": 1.2376,
      "loss/crossentropy": 2.5132007598876953,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1733236014842987,
      "loss/reg": 0.00018205813830718398,
      "step": 341
    },
    {
      "epoch": 0.04275,
      "grad_norm": 2.5386435985565186,
      "grad_norm_var": 0.0848263064399248,
      "learning_rate": 0.0001,
      "loss": 1.3462,
      "loss/crossentropy": 2.4127066135406494,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.16472230851650238,
      "loss/reg": 0.00018201676721218973,
      "step": 342
    },
    {
      "epoch": 0.042875,
      "grad_norm": 2.3057498931884766,
      "grad_norm_var": 0.08725284383438421,
      "learning_rate": 0.0001,
      "loss": 1.2497,
      "loss/crossentropy": 2.184678316116333,
      "loss/hidden": 1.0625,
      "loss/logits": 0.18537552654743195,
      "loss/reg": 0.0001819680182961747,
      "step": 343
    },
    {
      "epoch": 0.043,
      "grad_norm": 2.218367099761963,
      "grad_norm_var": 0.0816395413206275,
      "learning_rate": 0.0001,
      "loss": 1.1513,
      "loss/crossentropy": 2.449817419052124,
      "loss/hidden": 0.99609375,
      "loss/logits": 0.15338429808616638,
      "loss/reg": 0.00018192335846833885,
      "step": 344
    },
    {
      "epoch": 0.043125,
      "grad_norm": 2.2711985111236572,
      "grad_norm_var": 0.07626184388881233,
      "learning_rate": 0.0001,
      "loss": 1.0171,
      "loss/crossentropy": 2.509692907333374,
      "loss/hidden": 0.89453125,
      "loss/logits": 0.12079230695962906,
      "loss/reg": 0.00018188220565207303,
      "step": 345
    },
    {
      "epoch": 0.04325,
      "grad_norm": 2.787581443786621,
      "grad_norm_var": 0.08102267837075464,
      "learning_rate": 0.0001,
      "loss": 1.2428,
      "loss/crossentropy": 2.3301100730895996,
      "loss/hidden": 1.0546875,
      "loss/logits": 0.18630483746528625,
      "loss/reg": 0.00018185042426921427,
      "step": 346
    },
    {
      "epoch": 0.043375,
      "grad_norm": 2.0392682552337646,
      "grad_norm_var": 0.09234946001928937,
      "learning_rate": 0.0001,
      "loss": 1.0241,
      "loss/crossentropy": 2.588428020477295,
      "loss/hidden": 0.89453125,
      "loss/logits": 0.1277570128440857,
      "loss/reg": 0.00018179781909566373,
      "step": 347
    },
    {
      "epoch": 0.0435,
      "grad_norm": 2.7076616287231445,
      "grad_norm_var": 0.08681018440338353,
      "learning_rate": 0.0001,
      "loss": 1.4478,
      "loss/crossentropy": 2.6002542972564697,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.2193761169910431,
      "loss/reg": 0.0001817511219996959,
      "step": 348
    },
    {
      "epoch": 0.043625,
      "grad_norm": 2.353895425796509,
      "grad_norm_var": 0.08187996873415007,
      "learning_rate": 0.0001,
      "loss": 1.2846,
      "loss/crossentropy": 2.441397190093994,
      "loss/hidden": 1.09375,
      "loss/logits": 0.18899187445640564,
      "loss/reg": 0.0001816989533836022,
      "step": 349
    },
    {
      "epoch": 0.04375,
      "grad_norm": 2.078312397003174,
      "grad_norm_var": 0.07145312501918302,
      "learning_rate": 0.0001,
      "loss": 1.2337,
      "loss/crossentropy": 2.320770263671875,
      "loss/hidden": 1.0625,
      "loss/logits": 0.1694108545780182,
      "loss/reg": 0.00018164912762586027,
      "step": 350
    },
    {
      "epoch": 0.043875,
      "grad_norm": 2.455012321472168,
      "grad_norm_var": 0.06971347306554586,
      "learning_rate": 0.0001,
      "loss": 1.3011,
      "loss/crossentropy": 2.6782124042510986,
      "loss/hidden": 1.109375,
      "loss/logits": 0.1899489462375641,
      "loss/reg": 0.00018160381296183914,
      "step": 351
    },
    {
      "epoch": 0.044,
      "grad_norm": 3.1685619354248047,
      "grad_norm_var": 0.09844486312005239,
      "learning_rate": 0.0001,
      "loss": 1.3515,
      "loss/crossentropy": 2.8243625164031982,
      "loss/hidden": 1.140625,
      "loss/logits": 0.20903226733207703,
      "loss/reg": 0.00018155867292080075,
      "step": 352
    },
    {
      "epoch": 0.044125,
      "grad_norm": 3.2489728927612305,
      "grad_norm_var": 0.13257830736299236,
      "learning_rate": 0.0001,
      "loss": 1.5585,
      "loss/crossentropy": 2.8111252784729004,
      "loss/hidden": 1.3125,
      "loss/logits": 0.24414658546447754,
      "loss/reg": 0.0001815159630496055,
      "step": 353
    },
    {
      "epoch": 0.04425,
      "grad_norm": 3.3230886459350586,
      "grad_norm_var": 0.16879235535394155,
      "learning_rate": 0.0001,
      "loss": 1.4469,
      "loss/crossentropy": 2.5517845153808594,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.23417417705059052,
      "loss/reg": 0.00018148007802665234,
      "step": 354
    },
    {
      "epoch": 0.044375,
      "grad_norm": 2.2562191486358643,
      "grad_norm_var": 0.17228650926600247,
      "learning_rate": 0.0001,
      "loss": 1.1829,
      "loss/crossentropy": 2.422496795654297,
      "loss/hidden": 1.015625,
      "loss/logits": 0.16546528041362762,
      "loss/reg": 0.00018143345369026065,
      "step": 355
    },
    {
      "epoch": 0.0445,
      "grad_norm": 2.7640106678009033,
      "grad_norm_var": 0.17293323899421179,
      "learning_rate": 0.0001,
      "loss": 1.167,
      "loss/crossentropy": 2.828662157058716,
      "loss/hidden": 1.0,
      "loss/logits": 0.16519883275032043,
      "loss/reg": 0.00018138332234229892,
      "step": 356
    },
    {
      "epoch": 0.044625,
      "grad_norm": 2.3983893394470215,
      "grad_norm_var": 0.16551544063244136,
      "learning_rate": 0.0001,
      "loss": 1.1707,
      "loss/crossentropy": 2.605727434158325,
      "loss/hidden": 1.0078125,
      "loss/logits": 0.16108599305152893,
      "loss/reg": 0.0001813305716495961,
      "step": 357
    },
    {
      "epoch": 0.04475,
      "grad_norm": 3.6282835006713867,
      "grad_norm_var": 0.2370290852634845,
      "learning_rate": 0.0001,
      "loss": 1.3762,
      "loss/crossentropy": 2.6275181770324707,
      "loss/hidden": 1.15625,
      "loss/logits": 0.21812118589878082,
      "loss/reg": 0.00018128403462469578,
      "step": 358
    },
    {
      "epoch": 0.044875,
      "grad_norm": 3.763258934020996,
      "grad_norm_var": 0.30770300622119723,
      "learning_rate": 0.0001,
      "loss": 1.5858,
      "loss/crossentropy": 2.6596243381500244,
      "loss/hidden": 1.2890625,
      "loss/logits": 0.29489603638648987,
      "loss/reg": 0.00018124622874893248,
      "step": 359
    },
    {
      "epoch": 0.045,
      "grad_norm": 2.6321170330047607,
      "grad_norm_var": 0.29092860453977115,
      "learning_rate": 0.0001,
      "loss": 1.1651,
      "loss/crossentropy": 2.679645299911499,
      "loss/hidden": 1.015625,
      "loss/logits": 0.1476357877254486,
      "loss/reg": 0.00018119592277798802,
      "step": 360
    },
    {
      "epoch": 0.045125,
      "grad_norm": 2.794433832168579,
      "grad_norm_var": 0.27517751652292094,
      "learning_rate": 0.0001,
      "loss": 1.291,
      "loss/crossentropy": 2.46774959564209,
      "loss/hidden": 1.09375,
      "loss/logits": 0.1954488754272461,
      "loss/reg": 0.00018115356215275824,
      "step": 361
    },
    {
      "epoch": 0.04525,
      "grad_norm": 3.043492078781128,
      "grad_norm_var": 0.2797019428924839,
      "learning_rate": 0.0001,
      "loss": 1.5628,
      "loss/crossentropy": 2.5124197006225586,
      "loss/hidden": 1.328125,
      "loss/logits": 0.23281781375408173,
      "loss/reg": 0.00018111053213942796,
      "step": 362
    },
    {
      "epoch": 0.045375,
      "grad_norm": 2.155243158340454,
      "grad_norm_var": 0.2689192978759799,
      "learning_rate": 0.0001,
      "loss": 1.2775,
      "loss/crossentropy": 2.5151214599609375,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.1897062063217163,
      "loss/reg": 0.00018105284834746271,
      "step": 363
    },
    {
      "epoch": 0.0455,
      "grad_norm": 2.609884023666382,
      "grad_norm_var": 0.2706969753359004,
      "learning_rate": 0.0001,
      "loss": 1.3514,
      "loss/crossentropy": 2.506584882736206,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.20117563009262085,
      "loss/reg": 0.00018099797307513654,
      "step": 364
    },
    {
      "epoch": 0.045625,
      "grad_norm": 2.342069387435913,
      "grad_norm_var": 0.2713966376478879,
      "learning_rate": 0.0001,
      "loss": 1.3665,
      "loss/crossentropy": 2.7815566062927246,
      "loss/hidden": 1.15625,
      "loss/logits": 0.20847788453102112,
      "loss/reg": 0.00018094982078764588,
      "step": 365
    },
    {
      "epoch": 0.04575,
      "grad_norm": 3.0326218605041504,
      "grad_norm_var": 0.23758998657840114,
      "learning_rate": 0.0001,
      "loss": 1.3364,
      "loss/crossentropy": 2.523134469985962,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.20178905129432678,
      "loss/reg": 0.0001809141831472516,
      "step": 366
    },
    {
      "epoch": 0.045875,
      "grad_norm": 10.035235404968262,
      "grad_norm_var": 3.4286245913836826,
      "learning_rate": 0.0001,
      "loss": 1.7972,
      "loss/crossentropy": 2.548880100250244,
      "loss/hidden": 1.453125,
      "loss/logits": 0.34224259853363037,
      "loss/reg": 0.0001808816014090553,
      "step": 367
    },
    {
      "epoch": 0.046,
      "grad_norm": 2.35805606842041,
      "grad_norm_var": 3.4865601240502655,
      "learning_rate": 0.0001,
      "loss": 1.3183,
      "loss/crossentropy": 2.6983084678649902,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.1837082803249359,
      "loss/reg": 0.00018084091425407678,
      "step": 368
    },
    {
      "epoch": 0.046125,
      "grad_norm": 2.228501319885254,
      "grad_norm_var": 3.555062224897286,
      "learning_rate": 0.0001,
      "loss": 1.182,
      "loss/crossentropy": 2.2725465297698975,
      "loss/hidden": 1.015625,
      "loss/logits": 0.16455045342445374,
      "loss/reg": 0.0001807970111258328,
      "step": 369
    },
    {
      "epoch": 0.04625,
      "grad_norm": 3.036968469619751,
      "grad_norm_var": 3.555876206735041,
      "learning_rate": 0.0001,
      "loss": 1.3473,
      "loss/crossentropy": 2.462448835372925,
      "loss/hidden": 1.171875,
      "loss/logits": 0.1736328899860382,
      "loss/reg": 0.00018076066044159234,
      "step": 370
    },
    {
      "epoch": 0.046375,
      "grad_norm": 2.4165427684783936,
      "grad_norm_var": 3.5374699186157517,
      "learning_rate": 0.0001,
      "loss": 1.3649,
      "loss/crossentropy": 2.7596888542175293,
      "loss/hidden": 1.15625,
      "loss/logits": 0.20686297118663788,
      "loss/reg": 0.00018071448721457273,
      "step": 371
    },
    {
      "epoch": 0.0465,
      "grad_norm": 3.086303949356079,
      "grad_norm_var": 3.525121419257785,
      "learning_rate": 0.0001,
      "loss": 1.5493,
      "loss/crossentropy": 2.500018835067749,
      "loss/hidden": 1.28125,
      "loss/logits": 0.26621347665786743,
      "loss/reg": 0.00018066992925014347,
      "step": 372
    },
    {
      "epoch": 0.046625,
      "grad_norm": 2.4726622104644775,
      "grad_norm_var": 3.5173041221135075,
      "learning_rate": 0.0001,
      "loss": 1.136,
      "loss/crossentropy": 2.5688652992248535,
      "loss/hidden": 0.9921875,
      "loss/logits": 0.1419672667980194,
      "loss/reg": 0.00018063545576296747,
      "step": 373
    },
    {
      "epoch": 0.04675,
      "grad_norm": 2.6667373180389404,
      "grad_norm_var": 3.523672237020999,
      "learning_rate": 0.0001,
      "loss": 1.361,
      "loss/crossentropy": 2.5163731575012207,
      "loss/hidden": 1.140625,
      "loss/logits": 0.21853341162204742,
      "loss/reg": 0.0001806024374673143,
      "step": 374
    },
    {
      "epoch": 0.046875,
      "grad_norm": 2.3773033618927,
      "grad_norm_var": 3.5335662465775757,
      "learning_rate": 0.0001,
      "loss": 1.1721,
      "loss/crossentropy": 2.8756470680236816,
      "loss/hidden": 1.0078125,
      "loss/logits": 0.16248445212841034,
      "loss/reg": 0.00018057989655062556,
      "step": 375
    },
    {
      "epoch": 0.047,
      "grad_norm": 2.7833797931671143,
      "grad_norm_var": 3.5259529031790064,
      "learning_rate": 0.0001,
      "loss": 1.3487,
      "loss/crossentropy": 2.7149651050567627,
      "loss/hidden": 1.140625,
      "loss/logits": 0.20630021393299103,
      "loss/reg": 0.0001805323117878288,
      "step": 376
    },
    {
      "epoch": 0.047125,
      "grad_norm": 2.4678053855895996,
      "grad_norm_var": 3.545491291634365,
      "learning_rate": 0.0001,
      "loss": 1.4444,
      "loss/crossentropy": 2.2349612712860107,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.21606594324111938,
      "loss/reg": 0.00018048338824883103,
      "step": 377
    },
    {
      "epoch": 0.04725,
      "grad_norm": 2.368433952331543,
      "grad_norm_var": 3.5763182105236933,
      "learning_rate": 0.0001,
      "loss": 1.2423,
      "loss/crossentropy": 2.3095428943634033,
      "loss/hidden": 1.078125,
      "loss/logits": 0.1624036729335785,
      "loss/reg": 0.00018044162425212562,
      "step": 378
    },
    {
      "epoch": 0.047375,
      "grad_norm": 4.887925148010254,
      "grad_norm_var": 3.7252780043467792,
      "learning_rate": 0.0001,
      "loss": 1.357,
      "loss/crossentropy": 2.2361862659454346,
      "loss/hidden": 1.15625,
      "loss/logits": 0.1989278942346573,
      "loss/reg": 0.0001804078638087958,
      "step": 379
    },
    {
      "epoch": 0.0475,
      "grad_norm": 2.6829822063446045,
      "grad_norm_var": 3.7198784549945154,
      "learning_rate": 0.0001,
      "loss": 1.3576,
      "loss/crossentropy": 2.5164873600006104,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.20736883580684662,
      "loss/reg": 0.00018035774701274931,
      "step": 380
    },
    {
      "epoch": 0.047625,
      "grad_norm": 2.9376659393310547,
      "grad_norm_var": 3.673702627279779,
      "learning_rate": 0.0001,
      "loss": 1.2767,
      "loss/crossentropy": 2.5471227169036865,
      "loss/hidden": 1.09375,
      "loss/logits": 0.18117238581180573,
      "loss/reg": 0.000180321978405118,
      "step": 381
    },
    {
      "epoch": 0.04775,
      "grad_norm": 3.1138343811035156,
      "grad_norm_var": 3.671869876252352,
      "learning_rate": 0.0001,
      "loss": 1.5706,
      "loss/crossentropy": 2.5659494400024414,
      "loss/hidden": 1.28125,
      "loss/logits": 0.2875041365623474,
      "loss/reg": 0.00018027149781119078,
      "step": 382
    },
    {
      "epoch": 0.047875,
      "grad_norm": 2.51397705078125,
      "grad_norm_var": 0.3979920239186678,
      "learning_rate": 0.0001,
      "loss": 1.3404,
      "loss/crossentropy": 2.6721858978271484,
      "loss/hidden": 1.140625,
      "loss/logits": 0.19799375534057617,
      "loss/reg": 0.0001802304177545011,
      "step": 383
    },
    {
      "epoch": 0.048,
      "grad_norm": 2.6631741523742676,
      "grad_norm_var": 0.3868506457320857,
      "learning_rate": 0.0001,
      "loss": 1.1959,
      "loss/crossentropy": 2.8570642471313477,
      "loss/hidden": 1.046875,
      "loss/logits": 0.1472700834274292,
      "loss/reg": 0.00018018792616203427,
      "step": 384
    },
    {
      "epoch": 0.048125,
      "grad_norm": 2.225167989730835,
      "grad_norm_var": 0.38710267816574984,
      "learning_rate": 0.0001,
      "loss": 1.3387,
      "loss/crossentropy": 2.449850082397461,
      "loss/hidden": 1.140625,
      "loss/logits": 0.19622622430324554,
      "loss/reg": 0.00018013913359027356,
      "step": 385
    },
    {
      "epoch": 0.04825,
      "grad_norm": 3.642057418823242,
      "grad_norm_var": 0.429604118678219,
      "learning_rate": 0.0001,
      "loss": 1.5255,
      "loss/crossentropy": 3.060318946838379,
      "loss/hidden": 1.296875,
      "loss/logits": 0.2267942726612091,
      "loss/reg": 0.00018009122868534178,
      "step": 386
    },
    {
      "epoch": 0.048375,
      "grad_norm": 2.4205105304718018,
      "grad_norm_var": 0.4293855111770417,
      "learning_rate": 0.0001,
      "loss": 1.4919,
      "loss/crossentropy": 2.5439326763153076,
      "loss/hidden": 1.2578125,
      "loss/logits": 0.23233534395694733,
      "loss/reg": 0.0001800364989321679,
      "step": 387
    },
    {
      "epoch": 0.0485,
      "grad_norm": 3.090855598449707,
      "grad_norm_var": 0.429541218532165,
      "learning_rate": 0.0001,
      "loss": 1.3946,
      "loss/crossentropy": 2.3761212825775146,
      "loss/hidden": 1.1875,
      "loss/logits": 0.20534038543701172,
      "loss/reg": 0.0001799749006750062,
      "step": 388
    },
    {
      "epoch": 0.048625,
      "grad_norm": 2.829037666320801,
      "grad_norm_var": 0.4203970366893562,
      "learning_rate": 0.0001,
      "loss": 1.3953,
      "loss/crossentropy": 2.8099021911621094,
      "loss/hidden": 1.203125,
      "loss/logits": 0.19039469957351685,
      "loss/reg": 0.0001799268211470917,
      "step": 389
    },
    {
      "epoch": 0.04875,
      "grad_norm": 2.4180996417999268,
      "grad_norm_var": 0.43048309318027406,
      "learning_rate": 0.0001,
      "loss": 1.2337,
      "loss/crossentropy": 2.5344197750091553,
      "loss/hidden": 1.046875,
      "loss/logits": 0.18504786491394043,
      "loss/reg": 0.0001798665034584701,
      "step": 390
    },
    {
      "epoch": 0.048875,
      "grad_norm": 6.686630725860596,
      "grad_norm_var": 1.3259110009600303,
      "learning_rate": 0.0001,
      "loss": 1.4048,
      "loss/crossentropy": 1.59429132938385,
      "loss/hidden": 1.234375,
      "loss/logits": 0.16859778761863708,
      "loss/reg": 0.00017980851407628506,
      "step": 391
    },
    {
      "epoch": 0.049,
      "grad_norm": 2.47430682182312,
      "grad_norm_var": 1.3452680046498242,
      "learning_rate": 0.0001,
      "loss": 1.2505,
      "loss/crossentropy": 2.4976658821105957,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.17843782901763916,
      "loss/reg": 0.00017974060028791428,
      "step": 392
    },
    {
      "epoch": 0.049125,
      "grad_norm": 2.3827688694000244,
      "grad_norm_var": 1.3527620972999594,
      "learning_rate": 0.0001,
      "loss": 1.2697,
      "loss/crossentropy": 2.6137402057647705,
      "loss/hidden": 1.09375,
      "loss/logits": 0.17413964867591858,
      "loss/reg": 0.00017966754967346787,
      "step": 393
    },
    {
      "epoch": 0.04925,
      "grad_norm": 3.5299346446990967,
      "grad_norm_var": 1.3263260544293964,
      "learning_rate": 0.0001,
      "loss": 1.6096,
      "loss/crossentropy": 2.1338679790496826,
      "loss/hidden": 1.359375,
      "loss/logits": 0.24847310781478882,
      "loss/reg": 0.00017959915567189455,
      "step": 394
    },
    {
      "epoch": 0.049375,
      "grad_norm": 2.1859142780303955,
      "grad_norm_var": 1.1587385123986338,
      "learning_rate": 0.0001,
      "loss": 1.1247,
      "loss/crossentropy": 2.686222791671753,
      "loss/hidden": 0.96875,
      "loss/logits": 0.15412843227386475,
      "loss/reg": 0.00017955088696908206,
      "step": 395
    },
    {
      "epoch": 0.0495,
      "grad_norm": 2.372267484664917,
      "grad_norm_var": 1.1773802642484006,
      "learning_rate": 0.0001,
      "loss": 1.1605,
      "loss/crossentropy": 2.648873805999756,
      "loss/hidden": 0.9921875,
      "loss/logits": 0.16648846864700317,
      "loss/reg": 0.00017949036555364728,
      "step": 396
    },
    {
      "epoch": 0.049625,
      "grad_norm": 2.343231439590454,
      "grad_norm_var": 1.2018601019134032,
      "learning_rate": 0.0001,
      "loss": 1.1771,
      "loss/crossentropy": 2.3791165351867676,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.15182971954345703,
      "loss/reg": 0.00017943643615581095,
      "step": 397
    },
    {
      "epoch": 0.04975,
      "grad_norm": 2.7274091243743896,
      "grad_norm_var": 1.2017590131362377,
      "learning_rate": 0.0001,
      "loss": 1.324,
      "loss/crossentropy": 2.57863712310791,
      "loss/hidden": 1.140625,
      "loss/logits": 0.18160942196846008,
      "loss/reg": 0.00017938419478014112,
      "step": 398
    },
    {
      "epoch": 0.049875,
      "grad_norm": 4.155604362487793,
      "grad_norm_var": 1.2842575464972394,
      "learning_rate": 0.0001,
      "loss": 1.5653,
      "loss/crossentropy": 2.892331123352051,
      "loss/hidden": 1.34375,
      "loss/logits": 0.2197520136833191,
      "loss/reg": 0.0001793357077986002,
      "step": 399
    },
    {
      "epoch": 0.05,
      "grad_norm": 2.741466522216797,
      "grad_norm_var": 1.2810286441991583,
      "learning_rate": 0.0001,
      "loss": 1.3348,
      "loss/crossentropy": 2.546194314956665,
      "loss/hidden": 1.140625,
      "loss/logits": 0.19236674904823303,
      "loss/reg": 0.00017928793386090547,
      "step": 400
    },
    {
      "epoch": 0.050125,
      "grad_norm": 3.055189847946167,
      "grad_norm_var": 1.236778717086356,
      "learning_rate": 0.0001,
      "loss": 1.3152,
      "loss/crossentropy": 2.6719818115234375,
      "loss/hidden": 1.078125,
      "loss/logits": 0.23531492054462433,
      "loss/reg": 0.00017924165877047926,
      "step": 401
    },
    {
      "epoch": 0.05025,
      "grad_norm": 3.778313636779785,
      "grad_norm_var": 1.2484054094760106,
      "learning_rate": 0.0001,
      "loss": 1.5425,
      "loss/crossentropy": 2.7033426761627197,
      "loss/hidden": 1.28125,
      "loss/logits": 0.25942352414131165,
      "loss/reg": 0.00017917431250680238,
      "step": 402
    },
    {
      "epoch": 0.050375,
      "grad_norm": 3.0340867042541504,
      "grad_norm_var": 1.2184345071185567,
      "learning_rate": 0.0001,
      "loss": 1.455,
      "loss/crossentropy": 2.5875117778778076,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.22662264108657837,
      "loss/reg": 0.00017911198665387928,
      "step": 403
    },
    {
      "epoch": 0.0505,
      "grad_norm": 2.5151872634887695,
      "grad_norm_var": 1.2408325162170568,
      "learning_rate": 0.0001,
      "loss": 1.253,
      "loss/crossentropy": 2.6171019077301025,
      "loss/hidden": 1.078125,
      "loss/logits": 0.17306920886039734,
      "loss/reg": 0.0001790501846699044,
      "step": 404
    },
    {
      "epoch": 0.050625,
      "grad_norm": 2.4294838905334473,
      "grad_norm_var": 1.2640116286061014,
      "learning_rate": 0.0001,
      "loss": 1.1984,
      "loss/crossentropy": 2.4346368312835693,
      "loss/hidden": 1.015625,
      "loss/logits": 0.18100124597549438,
      "loss/reg": 0.00017898838268592954,
      "step": 405
    },
    {
      "epoch": 0.05075,
      "grad_norm": 2.7509233951568604,
      "grad_norm_var": 1.242810414819528,
      "learning_rate": 0.0001,
      "loss": 1.62,
      "loss/crossentropy": 2.047776937484741,
      "loss/hidden": 1.421875,
      "loss/logits": 0.19628939032554626,
      "loss/reg": 0.00017893253243528306,
      "step": 406
    },
    {
      "epoch": 0.050875,
      "grad_norm": 2.4834845066070557,
      "grad_norm_var": 0.3216287157750268,
      "learning_rate": 0.0001,
      "loss": 1.3514,
      "loss/crossentropy": 2.6643404960632324,
      "loss/hidden": 1.15625,
      "loss/logits": 0.1933945268392563,
      "loss/reg": 0.00017888467118609697,
      "step": 407
    },
    {
      "epoch": 0.051,
      "grad_norm": 2.417241334915161,
      "grad_norm_var": 0.32438624126894616,
      "learning_rate": 0.0001,
      "loss": 1.3021,
      "loss/crossentropy": 2.600691556930542,
      "loss/hidden": 1.109375,
      "loss/logits": 0.19094930589199066,
      "loss/reg": 0.00017882059910334647,
      "step": 408
    },
    {
      "epoch": 0.051125,
      "grad_norm": 2.3977980613708496,
      "grad_norm_var": 0.32355143397302466,
      "learning_rate": 0.0001,
      "loss": 1.4318,
      "loss/crossentropy": 2.6039021015167236,
      "loss/hidden": 1.1875,
      "loss/logits": 0.24255430698394775,
      "loss/reg": 0.00017876985657494515,
      "step": 409
    },
    {
      "epoch": 0.05125,
      "grad_norm": 2.205808639526367,
      "grad_norm_var": 0.3055601722416668,
      "learning_rate": 0.0001,
      "loss": 1.1411,
      "loss/crossentropy": 2.5921945571899414,
      "loss/hidden": 0.98828125,
      "loss/logits": 0.1510002166032791,
      "loss/reg": 0.0001787317160051316,
      "step": 410
    },
    {
      "epoch": 0.051375,
      "grad_norm": 3.166574478149414,
      "grad_norm_var": 0.2952319363017691,
      "learning_rate": 0.0001,
      "loss": 1.3997,
      "loss/crossentropy": 2.4906835556030273,
      "loss/hidden": 1.1796875,
      "loss/logits": 0.21822378039360046,
      "loss/reg": 0.00017868747818283737,
      "step": 411
    },
    {
      "epoch": 0.0515,
      "grad_norm": 2.248802661895752,
      "grad_norm_var": 0.30299352883237796,
      "learning_rate": 0.0001,
      "loss": 1.0924,
      "loss/crossentropy": 2.599449396133423,
      "loss/hidden": 0.9453125,
      "loss/logits": 0.14533644914627075,
      "loss/reg": 0.0001786567154340446,
      "step": 412
    },
    {
      "epoch": 0.051625,
      "grad_norm": 3.7585434913635254,
      "grad_norm_var": 0.3461126328200045,
      "learning_rate": 0.0001,
      "loss": 1.4199,
      "loss/crossentropy": 2.7979090213775635,
      "loss/hidden": 1.1875,
      "loss/logits": 0.23058170080184937,
      "loss/reg": 0.00017861124069895595,
      "step": 413
    },
    {
      "epoch": 0.05175,
      "grad_norm": 2.615172863006592,
      "grad_norm_var": 0.34898320978636455,
      "learning_rate": 0.0001,
      "loss": 1.3951,
      "loss/crossentropy": 1.9655026197433472,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.18238556385040283,
      "loss/reg": 0.00017857542843557894,
      "step": 414
    },
    {
      "epoch": 0.051875,
      "grad_norm": 2.38726806640625,
      "grad_norm_var": 0.2388532010949414,
      "learning_rate": 0.0001,
      "loss": 1.0998,
      "loss/crossentropy": 2.581796646118164,
      "loss/hidden": 0.94921875,
      "loss/logits": 0.1488036811351776,
      "loss/reg": 0.00017854152247309685,
      "step": 415
    },
    {
      "epoch": 0.052,
      "grad_norm": 2.1067628860473633,
      "grad_norm_var": 0.264675897864124,
      "learning_rate": 0.0001,
      "loss": 1.3556,
      "loss/crossentropy": 2.202812910079956,
      "loss/hidden": 1.171875,
      "loss/logits": 0.18191415071487427,
      "loss/reg": 0.00017848903371486813,
      "step": 416
    },
    {
      "epoch": 0.052125,
      "grad_norm": 2.8035178184509277,
      "grad_norm_var": 0.25703166277479733,
      "learning_rate": 0.0001,
      "loss": 1.367,
      "loss/crossentropy": 2.37416934967041,
      "loss/hidden": 1.125,
      "loss/logits": 0.24018406867980957,
      "loss/reg": 0.00017843768000602722,
      "step": 417
    },
    {
      "epoch": 0.05225,
      "grad_norm": 2.9603078365325928,
      "grad_norm_var": 0.1805549031603429,
      "learning_rate": 0.0001,
      "loss": 1.6557,
      "loss/crossentropy": 2.2720189094543457,
      "loss/hidden": 1.390625,
      "loss/logits": 0.2632881999015808,
      "loss/reg": 0.0001783809857442975,
      "step": 418
    },
    {
      "epoch": 0.052375,
      "grad_norm": 2.5229651927948,
      "grad_norm_var": 0.17020038194862058,
      "learning_rate": 0.0001,
      "loss": 1.4756,
      "loss/crossentropy": 2.63793683052063,
      "loss/hidden": 1.2265625,
      "loss/logits": 0.24721495807170868,
      "loss/reg": 0.00017833360470831394,
      "step": 419
    },
    {
      "epoch": 0.0525,
      "grad_norm": 2.355572462081909,
      "grad_norm_var": 0.17382358098597403,
      "learning_rate": 0.0001,
      "loss": 1.2231,
      "loss/crossentropy": 2.549771308898926,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.1823006570339203,
      "loss/reg": 0.0001782874605851248,
      "step": 420
    },
    {
      "epoch": 0.052625,
      "grad_norm": 2.9690279960632324,
      "grad_norm_var": 0.17970504092186654,
      "learning_rate": 0.0001,
      "loss": 1.3476,
      "loss/crossentropy": 3.0549476146698,
      "loss/hidden": 1.1484375,
      "loss/logits": 0.19741559028625488,
      "loss/reg": 0.00017824411042965949,
      "step": 421
    },
    {
      "epoch": 0.05275,
      "grad_norm": 2.880521059036255,
      "grad_norm_var": 0.18276892961250724,
      "learning_rate": 0.0001,
      "loss": 1.1749,
      "loss/crossentropy": 2.4630181789398193,
      "loss/hidden": 1.0078125,
      "loss/logits": 0.16526931524276733,
      "loss/reg": 0.00017819386266637594,
      "step": 422
    },
    {
      "epoch": 0.052875,
      "grad_norm": 2.373995542526245,
      "grad_norm_var": 0.18583898600397694,
      "learning_rate": 0.0001,
      "loss": 1.1957,
      "loss/crossentropy": 2.5825746059417725,
      "loss/hidden": 1.015625,
      "loss/logits": 0.17832687497138977,
      "loss/reg": 0.00017814231978263706,
      "step": 423
    },
    {
      "epoch": 0.053,
      "grad_norm": 2.8790440559387207,
      "grad_norm_var": 0.18572161644163973,
      "learning_rate": 0.0001,
      "loss": 1.3801,
      "loss/crossentropy": 2.7846927642822266,
      "loss/hidden": 1.15625,
      "loss/logits": 0.22202245891094208,
      "loss/reg": 0.00017809156270232052,
      "step": 424
    },
    {
      "epoch": 0.053125,
      "grad_norm": 2.737457275390625,
      "grad_norm_var": 0.18085466780064482,
      "learning_rate": 0.0001,
      "loss": 1.0736,
      "loss/crossentropy": 2.3963334560394287,
      "loss/hidden": 0.9375,
      "loss/logits": 0.1343117207288742,
      "loss/reg": 0.00017803694936446846,
      "step": 425
    },
    {
      "epoch": 0.05325,
      "grad_norm": 2.428834915161133,
      "grad_norm_var": 0.16969274721350278,
      "learning_rate": 0.0001,
      "loss": 1.157,
      "loss/crossentropy": 2.741081476211548,
      "loss/hidden": 0.99609375,
      "loss/logits": 0.15910392999649048,
      "loss/reg": 0.00017797992040868849,
      "step": 426
    },
    {
      "epoch": 0.053375,
      "grad_norm": 2.404265880584717,
      "grad_norm_var": 0.1585534584039614,
      "learning_rate": 0.0001,
      "loss": 1.2847,
      "loss/crossentropy": 2.89072322845459,
      "loss/hidden": 1.09375,
      "loss/logits": 0.1891355812549591,
      "loss/reg": 0.00017792356084100902,
      "step": 427
    },
    {
      "epoch": 0.0535,
      "grad_norm": 4.136691093444824,
      "grad_norm_var": 0.27981797299989897,
      "learning_rate": 0.0001,
      "loss": 1.4735,
      "loss/crossentropy": 2.6474928855895996,
      "loss/hidden": 1.21875,
      "loss/logits": 0.2529890537261963,
      "loss/reg": 0.0001778633304638788,
      "step": 428
    },
    {
      "epoch": 0.053625,
      "grad_norm": 2.81740140914917,
      "grad_norm_var": 0.21112886078796553,
      "learning_rate": 0.0001,
      "loss": 1.4668,
      "loss/crossentropy": 2.431791067123413,
      "loss/hidden": 1.21875,
      "loss/logits": 0.2462722361087799,
      "loss/reg": 0.00017780056805349886,
      "step": 429
    },
    {
      "epoch": 0.05375,
      "grad_norm": 3.5630476474761963,
      "grad_norm_var": 0.2551499062270144,
      "learning_rate": 0.0001,
      "loss": 1.4714,
      "loss/crossentropy": 2.573620557785034,
      "loss/hidden": 1.2421875,
      "loss/logits": 0.22747303545475006,
      "loss/reg": 0.00017774660955183208,
      "step": 430
    },
    {
      "epoch": 0.053875,
      "grad_norm": 27.927080154418945,
      "grad_norm_var": 39.71803281932496,
      "learning_rate": 0.0001,
      "loss": 1.2788,
      "loss/crossentropy": 2.723278522491455,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.191043421626091,
      "loss/reg": 0.00017769775877241045,
      "step": 431
    },
    {
      "epoch": 0.054,
      "grad_norm": 2.35017991065979,
      "grad_norm_var": 39.64838987108033,
      "learning_rate": 0.0001,
      "loss": 1.131,
      "loss/crossentropy": 2.7018210887908936,
      "loss/hidden": 0.98828125,
      "loss/logits": 0.14092248678207397,
      "loss/reg": 0.00017764477524906397,
      "step": 432
    },
    {
      "epoch": 0.054125,
      "grad_norm": 4.339812278747559,
      "grad_norm_var": 39.472594042649305,
      "learning_rate": 0.0001,
      "loss": 1.348,
      "loss/crossentropy": 2.808999538421631,
      "loss/hidden": 1.171875,
      "loss/logits": 0.17429979145526886,
      "loss/reg": 0.0001775865093804896,
      "step": 433
    },
    {
      "epoch": 0.05425,
      "grad_norm": 2.577270030975342,
      "grad_norm_var": 39.559269314754324,
      "learning_rate": 0.0001,
      "loss": 1.3842,
      "loss/crossentropy": 2.3989908695220947,
      "loss/hidden": 1.1875,
      "loss/logits": 0.19493895769119263,
      "loss/reg": 0.00017753323481883854,
      "step": 434
    },
    {
      "epoch": 0.054375,
      "grad_norm": 3.111717462539673,
      "grad_norm_var": 39.42935091258095,
      "learning_rate": 0.0001,
      "loss": 1.4152,
      "loss/crossentropy": 2.464742660522461,
      "loss/hidden": 1.203125,
      "loss/logits": 0.2103499472141266,
      "loss/reg": 0.00017748030950315297,
      "step": 435
    },
    {
      "epoch": 0.0545,
      "grad_norm": 2.0744073390960693,
      "grad_norm_var": 39.51433658135766,
      "learning_rate": 0.0001,
      "loss": 1.2571,
      "loss/crossentropy": 2.4160804748535156,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.18496504426002502,
      "loss/reg": 0.00017741357441991568,
      "step": 436
    },
    {
      "epoch": 0.054625,
      "grad_norm": 2.2014148235321045,
      "grad_norm_var": 39.705110235168064,
      "learning_rate": 0.0001,
      "loss": 1.2757,
      "loss/crossentropy": 2.558213710784912,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.20358391106128693,
      "loss/reg": 0.00017734503489919007,
      "step": 437
    },
    {
      "epoch": 0.05475,
      "grad_norm": 2.42464017868042,
      "grad_norm_var": 39.81199116769601,
      "learning_rate": 0.0001,
      "loss": 1.2475,
      "loss/crossentropy": 2.4901251792907715,
      "loss/hidden": 1.078125,
      "loss/logits": 0.16764749586582184,
      "loss/reg": 0.00017729295359458774,
      "step": 438
    },
    {
      "epoch": 0.054875,
      "grad_norm": 2.230916976928711,
      "grad_norm_var": 39.85185812679954,
      "learning_rate": 0.0001,
      "loss": 1.0936,
      "loss/crossentropy": 2.447075366973877,
      "loss/hidden": 0.9453125,
      "loss/logits": 0.1465335488319397,
      "loss/reg": 0.0001772290706867352,
      "step": 439
    },
    {
      "epoch": 0.055,
      "grad_norm": 2.749454975128174,
      "grad_norm_var": 39.878976148047535,
      "learning_rate": 0.0001,
      "loss": 1.4814,
      "loss/crossentropy": 2.2224342823028564,
      "loss/hidden": 1.2734375,
      "loss/logits": 0.20615623891353607,
      "loss/reg": 0.00017717515584081411,
      "step": 440
    },
    {
      "epoch": 0.055125,
      "grad_norm": 3.0442757606506348,
      "grad_norm_var": 39.81767857726663,
      "learning_rate": 0.0001,
      "loss": 1.8732,
      "loss/crossentropy": 2.110996723175049,
      "loss/hidden": 1.515625,
      "loss/logits": 0.35577309131622314,
      "loss/reg": 0.0001771111856214702,
      "step": 441
    },
    {
      "epoch": 0.05525,
      "grad_norm": 4.601929187774658,
      "grad_norm_var": 39.54202437298279,
      "learning_rate": 0.0001,
      "loss": 1.6329,
      "loss/crossentropy": 2.708014965057373,
      "loss/hidden": 1.3203125,
      "loss/logits": 0.3108658790588379,
      "loss/reg": 0.00017704560013953596,
      "step": 442
    },
    {
      "epoch": 0.055375,
      "grad_norm": 3.026305913925171,
      "grad_norm_var": 39.3895159629985,
      "learning_rate": 0.0001,
      "loss": 1.4032,
      "loss/crossentropy": 2.738539934158325,
      "loss/hidden": 1.1875,
      "loss/logits": 0.21392673254013062,
      "loss/reg": 0.00017697943258099258,
      "step": 443
    },
    {
      "epoch": 0.0555,
      "grad_norm": 3.2805826663970947,
      "grad_norm_var": 39.48518822606245,
      "learning_rate": 0.0001,
      "loss": 1.4055,
      "loss/crossentropy": 2.6368248462677,
      "loss/hidden": 1.203125,
      "loss/logits": 0.2005615532398224,
      "loss/reg": 0.0001769265509210527,
      "step": 444
    },
    {
      "epoch": 0.055625,
      "grad_norm": 2.9948861598968506,
      "grad_norm_var": 39.44686501090276,
      "learning_rate": 0.0001,
      "loss": 1.1039,
      "loss/crossentropy": 2.6170451641082764,
      "loss/hidden": 0.96484375,
      "loss/logits": 0.13730549812316895,
      "loss/reg": 0.00017686416686046869,
      "step": 445
    },
    {
      "epoch": 0.05575,
      "grad_norm": 2.2163240909576416,
      "grad_norm_var": 39.734049160677394,
      "learning_rate": 0.0001,
      "loss": 1.2018,
      "loss/crossentropy": 2.454167127609253,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.16098245978355408,
      "loss/reg": 0.00017681345343589783,
      "step": 446
    },
    {
      "epoch": 0.055875,
      "grad_norm": 2.6603283882141113,
      "grad_norm_var": 0.5323792649670357,
      "learning_rate": 0.0001,
      "loss": 1.3646,
      "loss/crossentropy": 2.499635696411133,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.22998638451099396,
      "loss/reg": 0.00017675201524980366,
      "step": 447
    },
    {
      "epoch": 0.056,
      "grad_norm": 2.228543758392334,
      "grad_norm_var": 0.5416984580164314,
      "learning_rate": 0.0001,
      "loss": 1.1798,
      "loss/crossentropy": 2.4845035076141357,
      "loss/hidden": 1.015625,
      "loss/logits": 0.16240708529949188,
      "loss/reg": 0.00017670769011601806,
      "step": 448
    },
    {
      "epoch": 0.056125,
      "grad_norm": 2.2279977798461914,
      "grad_norm_var": 0.4038044026013947,
      "learning_rate": 0.0001,
      "loss": 1.325,
      "loss/crossentropy": 2.3559677600860596,
      "loss/hidden": 1.140625,
      "loss/logits": 0.18264812231063843,
      "loss/reg": 0.00017664962797425687,
      "step": 449
    },
    {
      "epoch": 0.05625,
      "grad_norm": 2.7545721530914307,
      "grad_norm_var": 0.4022014302476805,
      "learning_rate": 0.0001,
      "loss": 1.4368,
      "loss/crossentropy": 2.374678373336792,
      "loss/hidden": 1.234375,
      "loss/logits": 0.20067429542541504,
      "loss/reg": 0.0001765888009686023,
      "step": 450
    },
    {
      "epoch": 0.056375,
      "grad_norm": 2.0431902408599854,
      "grad_norm_var": 0.4204979320743064,
      "learning_rate": 0.0001,
      "loss": 1.0879,
      "loss/crossentropy": 2.464282751083374,
      "loss/hidden": 0.94140625,
      "loss/logits": 0.1447741687297821,
      "loss/reg": 0.00017654395196586847,
      "step": 451
    },
    {
      "epoch": 0.0565,
      "grad_norm": 2.6901564598083496,
      "grad_norm_var": 0.39509245912944924,
      "learning_rate": 0.0001,
      "loss": 1.2069,
      "loss/crossentropy": 2.819383382797241,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.16609331965446472,
      "loss/reg": 0.00017649627989158034,
      "step": 452
    },
    {
      "epoch": 0.056625,
      "grad_norm": 2.1076624393463135,
      "grad_norm_var": 0.402011404785177,
      "learning_rate": 0.0001,
      "loss": 1.2556,
      "loss/crossentropy": 2.4384539127349854,
      "loss/hidden": 1.09375,
      "loss/logits": 0.16008606553077698,
      "loss/reg": 0.0001764398330124095,
      "step": 453
    },
    {
      "epoch": 0.05675,
      "grad_norm": 5.258527755737305,
      "grad_norm_var": 0.7979676690528414,
      "learning_rate": 0.0001,
      "loss": 1.6846,
      "loss/crossentropy": 2.5452864170074463,
      "loss/hidden": 1.4140625,
      "loss/logits": 0.26873037219047546,
      "loss/reg": 0.00017638294957578182,
      "step": 454
    },
    {
      "epoch": 0.056875,
      "grad_norm": 2.660983085632324,
      "grad_norm_var": 0.7721798756654529,
      "learning_rate": 0.0001,
      "loss": 1.4255,
      "loss/crossentropy": 2.585622549057007,
      "loss/hidden": 1.1875,
      "loss/logits": 0.23628488183021545,
      "loss/reg": 0.00017633216339163482,
      "step": 455
    },
    {
      "epoch": 0.057,
      "grad_norm": 2.7722721099853516,
      "grad_norm_var": 0.7717267059376826,
      "learning_rate": 0.0001,
      "loss": 1.2022,
      "loss/crossentropy": 2.800078868865967,
      "loss/hidden": 1.046875,
      "loss/logits": 0.15360021591186523,
      "loss/reg": 0.0001762784959282726,
      "step": 456
    },
    {
      "epoch": 0.057125,
      "grad_norm": 2.398487091064453,
      "grad_norm_var": 0.7862760060851559,
      "learning_rate": 0.0001,
      "loss": 1.6049,
      "loss/crossentropy": 2.1760308742523193,
      "loss/hidden": 1.328125,
      "loss/logits": 0.2749764025211334,
      "loss/reg": 0.00017622820450924337,
      "step": 457
    },
    {
      "epoch": 0.05725,
      "grad_norm": 2.1820383071899414,
      "grad_norm_var": 0.5935128198963845,
      "learning_rate": 0.0001,
      "loss": 1.2135,
      "loss/crossentropy": 2.6835989952087402,
      "loss/hidden": 1.046875,
      "loss/logits": 0.16489288210868835,
      "loss/reg": 0.00017618124547880143,
      "step": 458
    },
    {
      "epoch": 0.057375,
      "grad_norm": 2.3541500568389893,
      "grad_norm_var": 0.5942025229741127,
      "learning_rate": 0.0001,
      "loss": 1.3202,
      "loss/crossentropy": 2.6355140209198,
      "loss/hidden": 1.1328125,
      "loss/logits": 0.1856687068939209,
      "loss/reg": 0.00017612463852856308,
      "step": 459
    },
    {
      "epoch": 0.0575,
      "grad_norm": 2.4668965339660645,
      "grad_norm_var": 0.5700904660282996,
      "learning_rate": 0.0001,
      "loss": 1.3106,
      "loss/crossentropy": 2.656648874282837,
      "loss/hidden": 1.125,
      "loss/logits": 0.18380066752433777,
      "loss/reg": 0.00017607423069421202,
      "step": 460
    },
    {
      "epoch": 0.057625,
      "grad_norm": 2.6230361461639404,
      "grad_norm_var": 0.5604462661929044,
      "learning_rate": 0.0001,
      "loss": 1.2208,
      "loss/crossentropy": 2.62496280670166,
      "loss/hidden": 1.0546875,
      "loss/logits": 0.16433754563331604,
      "loss/reg": 0.0001760145096341148,
      "step": 461
    },
    {
      "epoch": 0.05775,
      "grad_norm": 2.7273848056793213,
      "grad_norm_var": 0.5504336260767483,
      "learning_rate": 0.0001,
      "loss": 1.4464,
      "loss/crossentropy": 2.558682441711426,
      "loss/hidden": 1.25,
      "loss/logits": 0.19467273354530334,
      "loss/reg": 0.00017597198893781751,
      "step": 462
    },
    {
      "epoch": 0.057875,
      "grad_norm": 2.6174404621124268,
      "grad_norm_var": 0.5504024009310662,
      "learning_rate": 0.0001,
      "loss": 1.4354,
      "loss/crossentropy": 2.1391024589538574,
      "loss/hidden": 1.234375,
      "loss/logits": 0.19923442602157593,
      "loss/reg": 0.00017592695076018572,
      "step": 463
    },
    {
      "epoch": 0.058,
      "grad_norm": 2.1421875953674316,
      "grad_norm_var": 0.5555149090661641,
      "learning_rate": 0.0001,
      "loss": 1.2089,
      "loss/crossentropy": 2.5548229217529297,
      "loss/hidden": 1.03125,
      "loss/logits": 0.17589880526065826,
      "loss/reg": 0.00017588076298125088,
      "step": 464
    },
    {
      "epoch": 0.058125,
      "grad_norm": 2.4386532306671143,
      "grad_norm_var": 0.5470902662726682,
      "learning_rate": 0.0001,
      "loss": 1.0766,
      "loss/crossentropy": 2.741363286972046,
      "loss/hidden": 0.94140625,
      "loss/logits": 0.13345757126808167,
      "loss/reg": 0.00017582789587322623,
      "step": 465
    },
    {
      "epoch": 0.05825,
      "grad_norm": 2.7671115398406982,
      "grad_norm_var": 0.5472918955756455,
      "learning_rate": 0.0001,
      "loss": 1.2597,
      "loss/crossentropy": 2.585339307785034,
      "loss/hidden": 1.09375,
      "loss/logits": 0.1642206907272339,
      "loss/reg": 0.00017577498510945588,
      "step": 466
    },
    {
      "epoch": 0.058375,
      "grad_norm": 2.075833797454834,
      "grad_norm_var": 0.5447581279211853,
      "learning_rate": 0.0001,
      "loss": 1.1296,
      "loss/crossentropy": 2.4353983402252197,
      "loss/hidden": 0.9765625,
      "loss/logits": 0.1513109803199768,
      "loss/reg": 0.00017573049990460277,
      "step": 467
    },
    {
      "epoch": 0.0585,
      "grad_norm": 2.181549072265625,
      "grad_norm_var": 0.557705888030068,
      "learning_rate": 0.0001,
      "loss": 1.234,
      "loss/crossentropy": 2.389277696609497,
      "loss/hidden": 1.0625,
      "loss/logits": 0.16970399022102356,
      "loss/reg": 0.0001756956335157156,
      "step": 468
    },
    {
      "epoch": 0.058625,
      "grad_norm": 2.239053249359131,
      "grad_norm_var": 0.5499689577837131,
      "learning_rate": 0.0001,
      "loss": 1.3819,
      "loss/crossentropy": 2.655245780944824,
      "loss/hidden": 1.15625,
      "loss/logits": 0.2238529622554779,
      "loss/reg": 0.00017564196605235338,
      "step": 469
    },
    {
      "epoch": 0.05875,
      "grad_norm": 2.3605029582977295,
      "grad_norm_var": 0.05499469594172955,
      "learning_rate": 0.0001,
      "loss": 1.3717,
      "loss/crossentropy": 2.3876919746398926,
      "loss/hidden": 1.171875,
      "loss/logits": 0.19806072115898132,
      "loss/reg": 0.00017558463150635362,
      "step": 470
    },
    {
      "epoch": 0.058875,
      "grad_norm": 2.848848581314087,
      "grad_norm_var": 0.06278663740608768,
      "learning_rate": 0.0001,
      "loss": 1.5331,
      "loss/crossentropy": 2.2159836292266846,
      "loss/hidden": 1.2890625,
      "loss/logits": 0.24231435358524323,
      "loss/reg": 0.00017553169163875282,
      "step": 471
    },
    {
      "epoch": 0.059,
      "grad_norm": 3.3035221099853516,
      "grad_norm_var": 0.10327356833769556,
      "learning_rate": 0.0001,
      "loss": 1.3096,
      "loss/crossentropy": 3.1982433795928955,
      "loss/hidden": 1.140625,
      "loss/logits": 0.16723856329917908,
      "loss/reg": 0.0001754647382767871,
      "step": 472
    },
    {
      "epoch": 0.059125,
      "grad_norm": 2.1994223594665527,
      "grad_norm_var": 0.10799121596538726,
      "learning_rate": 0.0001,
      "loss": 1.2996,
      "loss/crossentropy": 2.478044033050537,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.18064665794372559,
      "loss/reg": 0.0001753960968926549,
      "step": 473
    },
    {
      "epoch": 0.05925,
      "grad_norm": 2.7087080478668213,
      "grad_norm_var": 0.10507261048413336,
      "learning_rate": 0.0001,
      "loss": 1.2304,
      "loss/crossentropy": 2.533445119857788,
      "loss/hidden": 1.03125,
      "loss/logits": 0.19739562273025513,
      "loss/reg": 0.0001753264368744567,
      "step": 474
    },
    {
      "epoch": 0.059375,
      "grad_norm": 2.4171252250671387,
      "grad_norm_var": 0.1040673242944185,
      "learning_rate": 0.0001,
      "loss": 1.2977,
      "loss/crossentropy": 2.5889570713043213,
      "loss/hidden": 1.0859375,
      "loss/logits": 0.21002550423145294,
      "loss/reg": 0.00017525417206343263,
      "step": 475
    },
    {
      "epoch": 0.0595,
      "grad_norm": 3.3070313930511475,
      "grad_norm_var": 0.14365224039723495,
      "learning_rate": 0.0001,
      "loss": 1.7167,
      "loss/crossentropy": 2.7916061878204346,
      "loss/hidden": 1.4453125,
      "loss/logits": 0.2696676552295685,
      "loss/reg": 0.0001752021926222369,
      "step": 476
    },
    {
      "epoch": 0.059625,
      "grad_norm": 3.215808391571045,
      "grad_norm_var": 0.17060835871623775,
      "learning_rate": 0.0001,
      "loss": 1.519,
      "loss/crossentropy": 2.4236576557159424,
      "loss/hidden": 1.2734375,
      "loss/logits": 0.24385260045528412,
      "loss/reg": 0.0001751292875269428,
      "step": 477
    },
    {
      "epoch": 0.05975,
      "grad_norm": 2.4063854217529297,
      "grad_norm_var": 0.17146307657458593,
      "learning_rate": 0.0001,
      "loss": 1.0966,
      "loss/crossentropy": 2.4138169288635254,
      "loss/hidden": 0.94921875,
      "loss/logits": 0.14562630653381348,
      "loss/reg": 0.00017507674056105316,
      "step": 478
    },
    {
      "epoch": 0.059875,
      "grad_norm": 2.100240707397461,
      "grad_norm_var": 0.18538063838473515,
      "learning_rate": 0.0001,
      "loss": 1.2543,
      "loss/crossentropy": 2.656923770904541,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.18223264813423157,
      "loss/reg": 0.0001750182273099199,
      "step": 479
    },
    {
      "epoch": 0.06,
      "grad_norm": 2.416749954223633,
      "grad_norm_var": 0.17536422723811237,
      "learning_rate": 0.0001,
      "loss": 1.3935,
      "loss/crossentropy": 2.5389764308929443,
      "loss/hidden": 1.171875,
      "loss/logits": 0.21990099549293518,
      "loss/reg": 0.00017496509826742113,
      "step": 480
    },
    {
      "epoch": 0.060125,
      "grad_norm": 2.2483956813812256,
      "grad_norm_var": 0.18074697157593392,
      "learning_rate": 0.0001,
      "loss": 1.3504,
      "loss/crossentropy": 2.7335619926452637,
      "loss/hidden": 1.15625,
      "loss/logits": 0.19241070747375488,
      "loss/reg": 0.00017491589824203402,
      "step": 481
    },
    {
      "epoch": 0.06025,
      "grad_norm": 2.1877689361572266,
      "grad_norm_var": 0.1849354900853349,
      "learning_rate": 0.0001,
      "loss": 1.248,
      "loss/crossentropy": 2.66105318069458,
      "loss/hidden": 1.0625,
      "loss/logits": 0.18372279405593872,
      "loss/reg": 0.00017485868011135608,
      "step": 482
    },
    {
      "epoch": 0.060375,
      "grad_norm": 2.092740297317505,
      "grad_norm_var": 0.18396663403457297,
      "learning_rate": 0.0001,
      "loss": 1.2386,
      "loss/crossentropy": 2.7327380180358887,
      "loss/hidden": 1.0390625,
      "loss/logits": 0.19782081246376038,
      "loss/reg": 0.00017478555673733354,
      "step": 483
    },
    {
      "epoch": 0.0605,
      "grad_norm": 2.3183560371398926,
      "grad_norm_var": 0.17906094719213855,
      "learning_rate": 0.0001,
      "loss": 1.2824,
      "loss/crossentropy": 2.4746274948120117,
      "loss/hidden": 1.09375,
      "loss/logits": 0.1869126260280609,
      "loss/reg": 0.0001747341302689165,
      "step": 484
    },
    {
      "epoch": 0.060625,
      "grad_norm": 3.5248806476593018,
      "grad_norm_var": 0.23368608955632125,
      "learning_rate": 0.0001,
      "loss": 1.1942,
      "loss/crossentropy": 2.6376962661743164,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.1689874827861786,
      "loss/reg": 0.00017466919962316751,
      "step": 485
    },
    {
      "epoch": 0.06075,
      "grad_norm": 2.8418338298797607,
      "grad_norm_var": 0.23256916977224643,
      "learning_rate": 0.0001,
      "loss": 1.4452,
      "loss/crossentropy": 2.3843576908111572,
      "loss/hidden": 1.1953125,
      "loss/logits": 0.24809806048870087,
      "loss/reg": 0.00017459361697547138,
      "step": 486
    },
    {
      "epoch": 0.060875,
      "grad_norm": 2.697395086288452,
      "grad_norm_var": 0.22965639284835385,
      "learning_rate": 0.0001,
      "loss": 1.1924,
      "loss/crossentropy": 2.4852230548858643,
      "loss/hidden": 1.03125,
      "loss/logits": 0.15941958129405975,
      "loss/reg": 0.000174528788193129,
      "step": 487
    },
    {
      "epoch": 0.061,
      "grad_norm": 3.3923656940460205,
      "grad_norm_var": 0.23819745706471546,
      "learning_rate": 0.0001,
      "loss": 1.2613,
      "loss/crossentropy": 2.665536880493164,
      "loss/hidden": 1.078125,
      "loss/logits": 0.18143823742866516,
      "loss/reg": 0.00017445418052375317,
      "step": 488
    },
    {
      "epoch": 0.061125,
      "grad_norm": 3.820457935333252,
      "grad_norm_var": 0.30943274234138396,
      "learning_rate": 0.0001,
      "loss": 1.4177,
      "loss/crossentropy": 2.529301881790161,
      "loss/hidden": 1.203125,
      "loss/logits": 0.21287089586257935,
      "loss/reg": 0.00017437619681004435,
      "step": 489
    },
    {
      "epoch": 0.06125,
      "grad_norm": 2.441828966140747,
      "grad_norm_var": 0.3146780452696741,
      "learning_rate": 0.0001,
      "loss": 1.1788,
      "loss/crossentropy": 2.6879920959472656,
      "loss/hidden": 1.015625,
      "loss/logits": 0.1613830029964447,
      "loss/reg": 0.00017432142340112478,
      "step": 490
    },
    {
      "epoch": 0.061375,
      "grad_norm": 2.8247528076171875,
      "grad_norm_var": 0.30890959275739743,
      "learning_rate": 0.0001,
      "loss": 1.4539,
      "loss/crossentropy": 2.2834181785583496,
      "loss/hidden": 1.234375,
      "loss/logits": 0.21779246628284454,
      "loss/reg": 0.00017424933321308345,
      "step": 491
    },
    {
      "epoch": 0.0615,
      "grad_norm": 2.4493589401245117,
      "grad_norm_var": 0.2900195920917798,
      "learning_rate": 0.0001,
      "loss": 1.1726,
      "loss/crossentropy": 2.2889974117279053,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.14741873741149902,
      "loss/reg": 0.00017419550567865372,
      "step": 492
    },
    {
      "epoch": 0.061625,
      "grad_norm": 2.667480707168579,
      "grad_norm_var": 0.2700917314036502,
      "learning_rate": 0.0001,
      "loss": 1.4045,
      "loss/crossentropy": 2.404547929763794,
      "loss/hidden": 1.171875,
      "loss/logits": 0.23084591329097748,
      "loss/reg": 0.0001741291634971276,
      "step": 493
    },
    {
      "epoch": 0.06175,
      "grad_norm": 3.452826499938965,
      "grad_norm_var": 0.3042709664857275,
      "learning_rate": 0.0001,
      "loss": 1.3829,
      "loss/crossentropy": 2.9492647647857666,
      "loss/hidden": 1.1875,
      "loss/logits": 0.19365137815475464,
      "loss/reg": 0.00017407909035682678,
      "step": 494
    },
    {
      "epoch": 0.061875,
      "grad_norm": 2.6484220027923584,
      "grad_norm_var": 0.2779481152143487,
      "learning_rate": 0.0001,
      "loss": 1.4393,
      "loss/crossentropy": 2.4739902019500732,
      "loss/hidden": 1.2109375,
      "loss/logits": 0.22663193941116333,
      "loss/reg": 0.00017401424702256918,
      "step": 495
    },
    {
      "epoch": 0.062,
      "grad_norm": 2.3823630809783936,
      "grad_norm_var": 0.2795572822758951,
      "learning_rate": 0.0001,
      "loss": 1.4423,
      "loss/crossentropy": 2.1366026401519775,
      "loss/hidden": 1.25,
      "loss/logits": 0.1905536651611328,
      "loss/reg": 0.00017397114424966276,
      "step": 496
    },
    {
      "epoch": 0.062125,
      "grad_norm": 3.4144351482391357,
      "grad_norm_var": 0.28663513944516883,
      "learning_rate": 0.0001,
      "loss": 1.3353,
      "loss/crossentropy": 2.4802052974700928,
      "loss/hidden": 1.1171875,
      "loss/logits": 0.21641235053539276,
      "loss/reg": 0.00017391364963259548,
      "step": 497
    },
    {
      "epoch": 0.06225,
      "grad_norm": 2.339780330657959,
      "grad_norm_var": 0.2752179712740135,
      "learning_rate": 0.0001,
      "loss": 1.3462,
      "loss/crossentropy": 2.416666030883789,
      "loss/hidden": 1.140625,
      "loss/logits": 0.20387353003025055,
      "loss/reg": 0.00017384960665367544,
      "step": 498
    },
    {
      "epoch": 0.062375,
      "grad_norm": 2.3220226764678955,
      "grad_norm_var": 0.2559089262690113,
      "learning_rate": 0.0001,
      "loss": 1.1988,
      "loss/crossentropy": 2.668480634689331,
      "loss/hidden": 1.0234375,
      "loss/logits": 0.1736161708831787,
      "loss/reg": 0.00017379365453962237,
      "step": 499
    },
    {
      "epoch": 0.0625,
      "grad_norm": 2.35937762260437,
      "grad_norm_var": 0.25312725190736174,
      "learning_rate": 0.0001,
      "loss": 1.2363,
      "loss/crossentropy": 2.5724241733551025,
      "loss/hidden": 1.0703125,
      "loss/logits": 0.16428104043006897,
      "loss/reg": 0.00017374279559589922,
      "step": 500
    }
  ],
  "logging_steps": 1,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": true,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3.2202930782208e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}