Files
outputs/last-checkpoint/trainer_state.json
ModelHub XC 56547fe85e 初始化项目,由ModelHub XC社区提供模型
Model: Ba2han/outputs
Source: Original Platform
2026-08-03 14:39:21 +08:00

55507 lines
1.3 MiB

{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.96,
"eval_steps": 3300,
"global_step": 15840,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.00012121212121212121,
"grad_norm": 8.97504997253418,
"learning_rate": 4.848484848484849e-06,
"loss": 11.357827186584473,
"step": 2
},
{
"epoch": 0.00024242424242424242,
"grad_norm": 8.348938941955566,
"learning_rate": 1.4545454545454545e-05,
"loss": 11.106054306030273,
"step": 4
},
{
"epoch": 0.0003636363636363636,
"grad_norm": 7.870068550109863,
"learning_rate": 2.4242424242424244e-05,
"loss": 10.375286102294922,
"step": 6
},
{
"epoch": 0.00048484848484848484,
"grad_norm": 8.913647651672363,
"learning_rate": 3.3939393939393945e-05,
"loss": 9.937556266784668,
"step": 8
},
{
"epoch": 0.0006060606060606061,
"grad_norm": 2.8719382286071777,
"learning_rate": 4.3636363636363636e-05,
"loss": 9.630653381347656,
"step": 10
},
{
"epoch": 0.0007272727272727272,
"grad_norm": 2.4104228019714355,
"learning_rate": 5.333333333333333e-05,
"loss": 9.360559463500977,
"step": 12
},
{
"epoch": 0.0008484848484848485,
"grad_norm": 2.0712571144104004,
"learning_rate": 6.303030303030302e-05,
"loss": 9.180842399597168,
"step": 14
},
{
"epoch": 0.0009696969696969697,
"grad_norm": 1.9145631790161133,
"learning_rate": 7.272727272727273e-05,
"loss": 8.967535018920898,
"step": 16
},
{
"epoch": 0.001090909090909091,
"grad_norm": 2.62880802154541,
"learning_rate": 8.242424242424243e-05,
"loss": 8.75141716003418,
"step": 18
},
{
"epoch": 0.0012121212121212121,
"grad_norm": 1.3583670854568481,
"learning_rate": 9.212121212121214e-05,
"loss": 8.42448902130127,
"step": 20
},
{
"epoch": 0.0013333333333333333,
"grad_norm": 1.4527056217193604,
"learning_rate": 0.00010181818181818181,
"loss": 8.26899242401123,
"step": 22
},
{
"epoch": 0.0014545454545454545,
"grad_norm": 1.197034478187561,
"learning_rate": 0.00011151515151515153,
"loss": 8.070782661437988,
"step": 24
},
{
"epoch": 0.0015757575757575758,
"grad_norm": 0.6651339530944824,
"learning_rate": 0.00012121212121212122,
"loss": 7.898675918579102,
"step": 26
},
{
"epoch": 0.001696969696969697,
"grad_norm": 1.4606103897094727,
"learning_rate": 0.00013090909090909093,
"loss": 7.694046497344971,
"step": 28
},
{
"epoch": 0.0018181818181818182,
"grad_norm": 0.5734741687774658,
"learning_rate": 0.0001406060606060606,
"loss": 7.662841320037842,
"step": 30
},
{
"epoch": 0.0019393939393939393,
"grad_norm": 1.7686759233474731,
"learning_rate": 0.0001503030303030303,
"loss": 7.468966007232666,
"step": 32
},
{
"epoch": 0.0020606060606060605,
"grad_norm": 1.5994386672973633,
"learning_rate": 0.00016,
"loss": 7.472892761230469,
"step": 34
},
{
"epoch": 0.002181818181818182,
"grad_norm": 1.171338438987732,
"learning_rate": 0.00016969696969696972,
"loss": 7.422743797302246,
"step": 36
},
{
"epoch": 0.002303030303030303,
"grad_norm": 0.8250885605812073,
"learning_rate": 0.0001793939393939394,
"loss": 7.369729995727539,
"step": 38
},
{
"epoch": 0.0024242424242424242,
"grad_norm": 0.9469469785690308,
"learning_rate": 0.0001890909090909091,
"loss": 7.213651657104492,
"step": 40
},
{
"epoch": 0.0025454545454545456,
"grad_norm": 0.9908637404441833,
"learning_rate": 0.00019878787878787878,
"loss": 7.254331588745117,
"step": 42
},
{
"epoch": 0.0026666666666666666,
"grad_norm": 1.0451252460479736,
"learning_rate": 0.0002084848484848485,
"loss": 7.208287715911865,
"step": 44
},
{
"epoch": 0.002787878787878788,
"grad_norm": 1.5917110443115234,
"learning_rate": 0.00021818181818181818,
"loss": 7.121533393859863,
"step": 46
},
{
"epoch": 0.002909090909090909,
"grad_norm": 1.541863203048706,
"learning_rate": 0.0002278787878787879,
"loss": 7.107870101928711,
"step": 48
},
{
"epoch": 0.0030303030303030303,
"grad_norm": 0.8188634514808655,
"learning_rate": 0.0002375757575757576,
"loss": 7.091838359832764,
"step": 50
},
{
"epoch": 0.0031515151515151517,
"grad_norm": 0.975650429725647,
"learning_rate": 0.00024727272727272727,
"loss": 7.066593170166016,
"step": 52
},
{
"epoch": 0.0032727272727272726,
"grad_norm": 1.7995296716690063,
"learning_rate": 0.000256969696969697,
"loss": 6.922163486480713,
"step": 54
},
{
"epoch": 0.003393939393939394,
"grad_norm": 0.8872882127761841,
"learning_rate": 0.0002666666666666667,
"loss": 6.982857704162598,
"step": 56
},
{
"epoch": 0.003515151515151515,
"grad_norm": 1.2043501138687134,
"learning_rate": 0.0002763636363636364,
"loss": 6.882087707519531,
"step": 58
},
{
"epoch": 0.0036363636363636364,
"grad_norm": 1.221639633178711,
"learning_rate": 0.0002860606060606061,
"loss": 6.876432418823242,
"step": 60
},
{
"epoch": 0.0037575757575757577,
"grad_norm": 0.7717297673225403,
"learning_rate": 0.0002957575757575758,
"loss": 6.867863655090332,
"step": 62
},
{
"epoch": 0.0038787878787878787,
"grad_norm": 0.7375457882881165,
"learning_rate": 0.0003054545454545455,
"loss": 6.744990825653076,
"step": 64
},
{
"epoch": 0.004,
"grad_norm": 0.8509403467178345,
"learning_rate": 0.00031515151515151515,
"loss": 6.619245529174805,
"step": 66
},
{
"epoch": 0.004121212121212121,
"grad_norm": 1.6761586666107178,
"learning_rate": 0.00032484848484848485,
"loss": 6.734144687652588,
"step": 68
},
{
"epoch": 0.004242424242424243,
"grad_norm": 1.3128397464752197,
"learning_rate": 0.00033454545454545456,
"loss": 6.6777215003967285,
"step": 70
},
{
"epoch": 0.004363636363636364,
"grad_norm": 0.9473265409469604,
"learning_rate": 0.00034424242424242426,
"loss": 6.866725444793701,
"step": 72
},
{
"epoch": 0.004484848484848485,
"grad_norm": 0.7980499863624573,
"learning_rate": 0.00035393939393939396,
"loss": 6.664373874664307,
"step": 74
},
{
"epoch": 0.004606060606060606,
"grad_norm": 1.1570857763290405,
"learning_rate": 0.00036363636363636367,
"loss": 6.495471954345703,
"step": 76
},
{
"epoch": 0.0047272727272727275,
"grad_norm": 0.6680865287780762,
"learning_rate": 0.0003733333333333334,
"loss": 6.586875915527344,
"step": 78
},
{
"epoch": 0.0048484848484848485,
"grad_norm": 0.9470999240875244,
"learning_rate": 0.000383030303030303,
"loss": 6.360803127288818,
"step": 80
},
{
"epoch": 0.004969696969696969,
"grad_norm": 1.3528695106506348,
"learning_rate": 0.00039272727272727273,
"loss": 6.52389669418335,
"step": 82
},
{
"epoch": 0.005090909090909091,
"grad_norm": 1.2385674715042114,
"learning_rate": 0.0004024242424242425,
"loss": 6.367569446563721,
"step": 84
},
{
"epoch": 0.005212121212121212,
"grad_norm": 0.9044284224510193,
"learning_rate": 0.00041212121212121214,
"loss": 6.387779235839844,
"step": 86
},
{
"epoch": 0.005333333333333333,
"grad_norm": 1.0086190700531006,
"learning_rate": 0.0004218181818181818,
"loss": 6.384609222412109,
"step": 88
},
{
"epoch": 0.005454545454545455,
"grad_norm": 0.7946683764457703,
"learning_rate": 0.00043151515151515154,
"loss": 6.304538249969482,
"step": 90
},
{
"epoch": 0.005575757575757576,
"grad_norm": 0.9922563433647156,
"learning_rate": 0.00044121212121212125,
"loss": 6.326659679412842,
"step": 92
},
{
"epoch": 0.005696969696969697,
"grad_norm": 1.6856857538223267,
"learning_rate": 0.0004509090909090909,
"loss": 6.351809978485107,
"step": 94
},
{
"epoch": 0.005818181818181818,
"grad_norm": 1.00692880153656,
"learning_rate": 0.00046060606060606066,
"loss": 6.314554691314697,
"step": 96
},
{
"epoch": 0.00593939393939394,
"grad_norm": 0.7551726698875427,
"learning_rate": 0.0004703030303030303,
"loss": 6.178233623504639,
"step": 98
},
{
"epoch": 0.006060606060606061,
"grad_norm": 0.8453118801116943,
"learning_rate": 0.00048,
"loss": 6.081550598144531,
"step": 100
},
{
"epoch": 0.0061818181818181816,
"grad_norm": 0.8440249562263489,
"learning_rate": 0.0004896969696969697,
"loss": 6.138357162475586,
"step": 102
},
{
"epoch": 0.006303030303030303,
"grad_norm": 0.9946629405021667,
"learning_rate": 0.0004993939393939395,
"loss": 6.058218002319336,
"step": 104
},
{
"epoch": 0.006424242424242424,
"grad_norm": 1.0654397010803223,
"learning_rate": 0.0005090909090909091,
"loss": 6.076464653015137,
"step": 106
},
{
"epoch": 0.006545454545454545,
"grad_norm": 0.8686041235923767,
"learning_rate": 0.0005187878787878788,
"loss": 5.95872688293457,
"step": 108
},
{
"epoch": 0.006666666666666667,
"grad_norm": 0.6830413341522217,
"learning_rate": 0.0005284848484848485,
"loss": 5.902834415435791,
"step": 110
},
{
"epoch": 0.006787878787878788,
"grad_norm": 1.0477262735366821,
"learning_rate": 0.0005381818181818182,
"loss": 5.8460164070129395,
"step": 112
},
{
"epoch": 0.006909090909090909,
"grad_norm": 0.6564334630966187,
"learning_rate": 0.0005478787878787878,
"loss": 6.002228736877441,
"step": 114
},
{
"epoch": 0.00703030303030303,
"grad_norm": 0.7843162417411804,
"learning_rate": 0.0005575757575757576,
"loss": 5.780240535736084,
"step": 116
},
{
"epoch": 0.007151515151515152,
"grad_norm": 0.9642066359519958,
"learning_rate": 0.0005672727272727274,
"loss": 5.727415084838867,
"step": 118
},
{
"epoch": 0.007272727272727273,
"grad_norm": 1.2070485353469849,
"learning_rate": 0.000576969696969697,
"loss": 5.800960063934326,
"step": 120
},
{
"epoch": 0.007393939393939394,
"grad_norm": 0.8290711641311646,
"learning_rate": 0.0005866666666666667,
"loss": 5.794440269470215,
"step": 122
},
{
"epoch": 0.0075151515151515155,
"grad_norm": 0.8914417028427124,
"learning_rate": 0.0005963636363636364,
"loss": 5.611122131347656,
"step": 124
},
{
"epoch": 0.0076363636363636364,
"grad_norm": 0.7072174549102783,
"learning_rate": 0.0006060606060606061,
"loss": 5.646268367767334,
"step": 126
},
{
"epoch": 0.007757575757575757,
"grad_norm": 0.8764508962631226,
"learning_rate": 0.0006157575757575757,
"loss": 5.6723761558532715,
"step": 128
},
{
"epoch": 0.00787878787878788,
"grad_norm": 0.6534025073051453,
"learning_rate": 0.0006254545454545455,
"loss": 5.622500419616699,
"step": 130
},
{
"epoch": 0.008,
"grad_norm": 0.5357821583747864,
"learning_rate": 0.0006351515151515152,
"loss": 5.552145481109619,
"step": 132
},
{
"epoch": 0.008121212121212121,
"grad_norm": 0.6033462285995483,
"learning_rate": 0.0006448484848484849,
"loss": 5.5264177322387695,
"step": 134
},
{
"epoch": 0.008242424242424242,
"grad_norm": 0.8388949632644653,
"learning_rate": 0.0006545454545454546,
"loss": 5.494859218597412,
"step": 136
},
{
"epoch": 0.008363636363636363,
"grad_norm": 0.9601350426673889,
"learning_rate": 0.0006642424242424243,
"loss": 5.32069730758667,
"step": 138
},
{
"epoch": 0.008484848484848486,
"grad_norm": 1.2640959024429321,
"learning_rate": 0.0006739393939393939,
"loss": 5.488650798797607,
"step": 140
},
{
"epoch": 0.008606060606060607,
"grad_norm": 0.8532082438468933,
"learning_rate": 0.0006836363636363636,
"loss": 5.437753200531006,
"step": 142
},
{
"epoch": 0.008727272727272728,
"grad_norm": 1.0036379098892212,
"learning_rate": 0.0006933333333333333,
"loss": 5.435619831085205,
"step": 144
},
{
"epoch": 0.008848484848484849,
"grad_norm": 0.6930304169654846,
"learning_rate": 0.0007030303030303031,
"loss": 5.5524492263793945,
"step": 146
},
{
"epoch": 0.00896969696969697,
"grad_norm": 0.6947512626647949,
"learning_rate": 0.0007127272727272728,
"loss": 5.4725847244262695,
"step": 148
},
{
"epoch": 0.00909090909090909,
"grad_norm": 0.5082792639732361,
"learning_rate": 0.0007224242424242425,
"loss": 5.255496501922607,
"step": 150
},
{
"epoch": 0.009212121212121211,
"grad_norm": 0.5149092674255371,
"learning_rate": 0.0007321212121212122,
"loss": 5.35959005355835,
"step": 152
},
{
"epoch": 0.009333333333333334,
"grad_norm": 0.468622624874115,
"learning_rate": 0.0007418181818181818,
"loss": 5.325175762176514,
"step": 154
},
{
"epoch": 0.009454545454545455,
"grad_norm": 0.598091721534729,
"learning_rate": 0.0007515151515151516,
"loss": 5.322880268096924,
"step": 156
},
{
"epoch": 0.009575757575757576,
"grad_norm": 0.8230356574058533,
"learning_rate": 0.0007612121212121212,
"loss": 5.333863735198975,
"step": 158
},
{
"epoch": 0.009696969696969697,
"grad_norm": 1.1033756732940674,
"learning_rate": 0.000770909090909091,
"loss": 5.284386157989502,
"step": 160
},
{
"epoch": 0.009818181818181818,
"grad_norm": 0.650258481502533,
"learning_rate": 0.0007806060606060606,
"loss": 5.252943515777588,
"step": 162
},
{
"epoch": 0.009939393939393939,
"grad_norm": 0.8603656888008118,
"learning_rate": 0.0007903030303030304,
"loss": 5.090350151062012,
"step": 164
},
{
"epoch": 0.01006060606060606,
"grad_norm": 0.4447465240955353,
"learning_rate": 0.0008,
"loss": 5.219063758850098,
"step": 166
},
{
"epoch": 0.010181818181818183,
"grad_norm": 0.4416581988334656,
"learning_rate": 0.0008,
"loss": 5.028714179992676,
"step": 168
},
{
"epoch": 0.010303030303030303,
"grad_norm": 0.43629151582717896,
"learning_rate": 0.0008,
"loss": 4.954435348510742,
"step": 170
},
{
"epoch": 0.010424242424242424,
"grad_norm": 0.48155081272125244,
"learning_rate": 0.0008,
"loss": 4.948851108551025,
"step": 172
},
{
"epoch": 0.010545454545454545,
"grad_norm": 0.5557255148887634,
"learning_rate": 0.0008,
"loss": 5.076390743255615,
"step": 174
},
{
"epoch": 0.010666666666666666,
"grad_norm": 0.5931875705718994,
"learning_rate": 0.0008,
"loss": 5.092473030090332,
"step": 176
},
{
"epoch": 0.010787878787878787,
"grad_norm": 0.8179289102554321,
"learning_rate": 0.0008,
"loss": 4.882107257843018,
"step": 178
},
{
"epoch": 0.01090909090909091,
"grad_norm": 0.9967648983001709,
"learning_rate": 0.0008,
"loss": 4.923511028289795,
"step": 180
},
{
"epoch": 0.011030303030303031,
"grad_norm": 1.1743577718734741,
"learning_rate": 0.0008,
"loss": 5.101078987121582,
"step": 182
},
{
"epoch": 0.011151515151515152,
"grad_norm": 1.0063247680664062,
"learning_rate": 0.0008,
"loss": 5.137593746185303,
"step": 184
},
{
"epoch": 0.011272727272727273,
"grad_norm": 0.5801673531532288,
"learning_rate": 0.0008,
"loss": 5.013425827026367,
"step": 186
},
{
"epoch": 0.011393939393939394,
"grad_norm": 0.5073955655097961,
"learning_rate": 0.0008,
"loss": 4.989893913269043,
"step": 188
},
{
"epoch": 0.011515151515151515,
"grad_norm": 0.5057386159896851,
"learning_rate": 0.0008,
"loss": 4.878107070922852,
"step": 190
},
{
"epoch": 0.011636363636363636,
"grad_norm": 0.42275702953338623,
"learning_rate": 0.0008,
"loss": 4.894931316375732,
"step": 192
},
{
"epoch": 0.011757575757575758,
"grad_norm": 0.4581521451473236,
"learning_rate": 0.0008,
"loss": 4.830855846405029,
"step": 194
},
{
"epoch": 0.01187878787878788,
"grad_norm": 0.36250874400138855,
"learning_rate": 0.0008,
"loss": 4.795516490936279,
"step": 196
},
{
"epoch": 0.012,
"grad_norm": 0.3030204772949219,
"learning_rate": 0.0008,
"loss": 4.763975143432617,
"step": 198
},
{
"epoch": 0.012121212121212121,
"grad_norm": 0.5053560137748718,
"learning_rate": 0.0008,
"loss": 4.734469413757324,
"step": 200
},
{
"epoch": 0.012242424242424242,
"grad_norm": 0.6947022080421448,
"learning_rate": 0.0008,
"loss": 4.838320255279541,
"step": 202
},
{
"epoch": 0.012363636363636363,
"grad_norm": 0.7823297381401062,
"learning_rate": 0.0008,
"loss": 4.6456990242004395,
"step": 204
},
{
"epoch": 0.012484848484848484,
"grad_norm": 0.654838502407074,
"learning_rate": 0.0008,
"loss": 4.81408166885376,
"step": 206
},
{
"epoch": 0.012606060606060607,
"grad_norm": 0.45549893379211426,
"learning_rate": 0.0008,
"loss": 4.798559665679932,
"step": 208
},
{
"epoch": 0.012727272727272728,
"grad_norm": 0.4346763789653778,
"learning_rate": 0.0008,
"loss": 4.724524974822998,
"step": 210
},
{
"epoch": 0.012848484848484849,
"grad_norm": 0.3581501245498657,
"learning_rate": 0.0008,
"loss": 4.660583019256592,
"step": 212
},
{
"epoch": 0.01296969696969697,
"grad_norm": 0.29114705324172974,
"learning_rate": 0.0008,
"loss": 4.576566219329834,
"step": 214
},
{
"epoch": 0.01309090909090909,
"grad_norm": 0.2913273572921753,
"learning_rate": 0.0008,
"loss": 4.589745044708252,
"step": 216
},
{
"epoch": 0.013212121212121212,
"grad_norm": 0.26899755001068115,
"learning_rate": 0.0008,
"loss": 4.623720169067383,
"step": 218
},
{
"epoch": 0.013333333333333334,
"grad_norm": 0.34857481718063354,
"learning_rate": 0.0008,
"loss": 4.624195098876953,
"step": 220
},
{
"epoch": 0.013454545454545455,
"grad_norm": 0.505099892616272,
"learning_rate": 0.0008,
"loss": 4.5323486328125,
"step": 222
},
{
"epoch": 0.013575757575757576,
"grad_norm": 0.4094547629356384,
"learning_rate": 0.0008,
"loss": 4.524002552032471,
"step": 224
},
{
"epoch": 0.013696969696969697,
"grad_norm": 0.3158399760723114,
"learning_rate": 0.0008,
"loss": 4.578722953796387,
"step": 226
},
{
"epoch": 0.013818181818181818,
"grad_norm": 0.53526771068573,
"learning_rate": 0.0008,
"loss": 4.528520107269287,
"step": 228
},
{
"epoch": 0.013939393939393939,
"grad_norm": 0.5406535267829895,
"learning_rate": 0.0008,
"loss": 4.560442924499512,
"step": 230
},
{
"epoch": 0.01406060606060606,
"grad_norm": 0.44130751490592957,
"learning_rate": 0.0008,
"loss": 4.484059810638428,
"step": 232
},
{
"epoch": 0.014181818181818183,
"grad_norm": 0.5150421261787415,
"learning_rate": 0.0008,
"loss": 4.506237030029297,
"step": 234
},
{
"epoch": 0.014303030303030304,
"grad_norm": 0.41909053921699524,
"learning_rate": 0.0008,
"loss": 4.363324165344238,
"step": 236
},
{
"epoch": 0.014424242424242424,
"grad_norm": 0.3764134347438812,
"learning_rate": 0.0008,
"loss": 4.3746843338012695,
"step": 238
},
{
"epoch": 0.014545454545454545,
"grad_norm": 0.5060068368911743,
"learning_rate": 0.0008,
"loss": 4.42823600769043,
"step": 240
},
{
"epoch": 0.014666666666666666,
"grad_norm": 0.5175958871841431,
"learning_rate": 0.0008,
"loss": 4.42431640625,
"step": 242
},
{
"epoch": 0.014787878787878787,
"grad_norm": 0.5552281737327576,
"learning_rate": 0.0008,
"loss": 4.365012168884277,
"step": 244
},
{
"epoch": 0.014909090909090908,
"grad_norm": 0.30622056126594543,
"learning_rate": 0.0008,
"loss": 4.217746734619141,
"step": 246
},
{
"epoch": 0.015030303030303031,
"grad_norm": 0.27839550375938416,
"learning_rate": 0.0008,
"loss": 4.380131721496582,
"step": 248
},
{
"epoch": 0.015151515151515152,
"grad_norm": 0.4964343309402466,
"learning_rate": 0.0008,
"loss": 4.266903877258301,
"step": 250
},
{
"epoch": 0.015272727272727273,
"grad_norm": 0.38328754901885986,
"learning_rate": 0.0008,
"loss": 4.353453159332275,
"step": 252
},
{
"epoch": 0.015393939393939394,
"grad_norm": 0.2895657420158386,
"learning_rate": 0.0008,
"loss": 4.225037097930908,
"step": 254
},
{
"epoch": 0.015515151515151515,
"grad_norm": 0.3326258063316345,
"learning_rate": 0.0008,
"loss": 4.262819766998291,
"step": 256
},
{
"epoch": 0.015636363636363636,
"grad_norm": 0.2709735631942749,
"learning_rate": 0.0008,
"loss": 4.279838562011719,
"step": 258
},
{
"epoch": 0.01575757575757576,
"grad_norm": 0.28639206290245056,
"learning_rate": 0.0008,
"loss": 4.2228240966796875,
"step": 260
},
{
"epoch": 0.015878787878787878,
"grad_norm": 0.3204135298728943,
"learning_rate": 0.0008,
"loss": 4.2194037437438965,
"step": 262
},
{
"epoch": 0.016,
"grad_norm": 0.3845968544483185,
"learning_rate": 0.0008,
"loss": 4.192597389221191,
"step": 264
},
{
"epoch": 0.01612121212121212,
"grad_norm": 0.45057976245880127,
"learning_rate": 0.0008,
"loss": 4.149747371673584,
"step": 266
},
{
"epoch": 0.016242424242424242,
"grad_norm": 0.36388057470321655,
"learning_rate": 0.0008,
"loss": 4.067847728729248,
"step": 268
},
{
"epoch": 0.016363636363636365,
"grad_norm": 0.3418745994567871,
"learning_rate": 0.0008,
"loss": 4.1565022468566895,
"step": 270
},
{
"epoch": 0.016484848484848484,
"grad_norm": 0.3683988153934479,
"learning_rate": 0.0008,
"loss": 4.09769868850708,
"step": 272
},
{
"epoch": 0.016606060606060607,
"grad_norm": 0.3683854341506958,
"learning_rate": 0.0008,
"loss": 3.9879941940307617,
"step": 274
},
{
"epoch": 0.016727272727272726,
"grad_norm": 0.47024789452552795,
"learning_rate": 0.0008,
"loss": 4.045173645019531,
"step": 276
},
{
"epoch": 0.01684848484848485,
"grad_norm": 0.46615248918533325,
"learning_rate": 0.0008,
"loss": 4.0129547119140625,
"step": 278
},
{
"epoch": 0.01696969696969697,
"grad_norm": 0.3514328896999359,
"learning_rate": 0.0008,
"loss": 4.037949562072754,
"step": 280
},
{
"epoch": 0.01709090909090909,
"grad_norm": 0.3677462041378021,
"learning_rate": 0.0008,
"loss": 4.070645809173584,
"step": 282
},
{
"epoch": 0.017212121212121213,
"grad_norm": 0.30767402052879333,
"learning_rate": 0.0008,
"loss": 3.9893791675567627,
"step": 284
},
{
"epoch": 0.017333333333333333,
"grad_norm": 0.30136391520500183,
"learning_rate": 0.0008,
"loss": 3.9328291416168213,
"step": 286
},
{
"epoch": 0.017454545454545455,
"grad_norm": 0.2418648898601532,
"learning_rate": 0.0008,
"loss": 3.9587371349334717,
"step": 288
},
{
"epoch": 0.017575757575757574,
"grad_norm": 0.3279234766960144,
"learning_rate": 0.0008,
"loss": 3.858621835708618,
"step": 290
},
{
"epoch": 0.017696969696969697,
"grad_norm": 0.3606475293636322,
"learning_rate": 0.0008,
"loss": 3.8565807342529297,
"step": 292
},
{
"epoch": 0.01781818181818182,
"grad_norm": 0.3619609773159027,
"learning_rate": 0.0008,
"loss": 3.8307735919952393,
"step": 294
},
{
"epoch": 0.01793939393939394,
"grad_norm": 0.33409959077835083,
"learning_rate": 0.0008,
"loss": 3.918273687362671,
"step": 296
},
{
"epoch": 0.01806060606060606,
"grad_norm": 0.2992623746395111,
"learning_rate": 0.0008,
"loss": 3.858957529067993,
"step": 298
},
{
"epoch": 0.01818181818181818,
"grad_norm": 0.33072054386138916,
"learning_rate": 0.0008,
"loss": 3.748624801635742,
"step": 300
},
{
"epoch": 0.018303030303030304,
"grad_norm": 0.28332406282424927,
"learning_rate": 0.0008,
"loss": 3.7838051319122314,
"step": 302
},
{
"epoch": 0.018424242424242423,
"grad_norm": 0.30517423152923584,
"learning_rate": 0.0008,
"loss": 3.802793264389038,
"step": 304
},
{
"epoch": 0.018545454545454546,
"grad_norm": 0.3117048442363739,
"learning_rate": 0.0008,
"loss": 3.811725378036499,
"step": 306
},
{
"epoch": 0.018666666666666668,
"grad_norm": 0.2468871921300888,
"learning_rate": 0.0008,
"loss": 3.7850887775421143,
"step": 308
},
{
"epoch": 0.018787878787878787,
"grad_norm": 0.283168226480484,
"learning_rate": 0.0008,
"loss": 3.6664271354675293,
"step": 310
},
{
"epoch": 0.01890909090909091,
"grad_norm": 0.3218361437320709,
"learning_rate": 0.0008,
"loss": 3.701439380645752,
"step": 312
},
{
"epoch": 0.01903030303030303,
"grad_norm": 0.31093841791152954,
"learning_rate": 0.0008,
"loss": 3.7514190673828125,
"step": 314
},
{
"epoch": 0.019151515151515152,
"grad_norm": 0.31061142683029175,
"learning_rate": 0.0008,
"loss": 3.7566089630126953,
"step": 316
},
{
"epoch": 0.01927272727272727,
"grad_norm": 0.29709354043006897,
"learning_rate": 0.0008,
"loss": 3.6478185653686523,
"step": 318
},
{
"epoch": 0.019393939393939394,
"grad_norm": 0.2535860538482666,
"learning_rate": 0.0008,
"loss": 3.7081849575042725,
"step": 320
},
{
"epoch": 0.019515151515151517,
"grad_norm": 0.2756178677082062,
"learning_rate": 0.0008,
"loss": 3.6849632263183594,
"step": 322
},
{
"epoch": 0.019636363636363636,
"grad_norm": 0.2705349922180176,
"learning_rate": 0.0008,
"loss": 3.568624973297119,
"step": 324
},
{
"epoch": 0.01975757575757576,
"grad_norm": 0.31116461753845215,
"learning_rate": 0.0008,
"loss": 3.6941237449645996,
"step": 326
},
{
"epoch": 0.019878787878787878,
"grad_norm": 0.2481263279914856,
"learning_rate": 0.0008,
"loss": 3.684450387954712,
"step": 328
},
{
"epoch": 0.02,
"grad_norm": 0.24916408956050873,
"learning_rate": 0.0008,
"loss": 3.6361258029937744,
"step": 330
},
{
"epoch": 0.02012121212121212,
"grad_norm": 0.2090545892715454,
"learning_rate": 0.0008,
"loss": 3.6069369316101074,
"step": 332
},
{
"epoch": 0.020242424242424242,
"grad_norm": 0.20673981308937073,
"learning_rate": 0.0008,
"loss": 3.5468575954437256,
"step": 334
},
{
"epoch": 0.020363636363636365,
"grad_norm": 0.22687797248363495,
"learning_rate": 0.0008,
"loss": 3.442838191986084,
"step": 336
},
{
"epoch": 0.020484848484848484,
"grad_norm": 0.2661835253238678,
"learning_rate": 0.0008,
"loss": 3.5916757583618164,
"step": 338
},
{
"epoch": 0.020606060606060607,
"grad_norm": 0.302510529756546,
"learning_rate": 0.0008,
"loss": 3.5437164306640625,
"step": 340
},
{
"epoch": 0.020727272727272726,
"grad_norm": 0.3088708519935608,
"learning_rate": 0.0008,
"loss": 3.5581564903259277,
"step": 342
},
{
"epoch": 0.02084848484848485,
"grad_norm": 0.29519781470298767,
"learning_rate": 0.0008,
"loss": 3.51533842086792,
"step": 344
},
{
"epoch": 0.020969696969696968,
"grad_norm": 0.24684779345989227,
"learning_rate": 0.0008,
"loss": 3.6499595642089844,
"step": 346
},
{
"epoch": 0.02109090909090909,
"grad_norm": 0.21150065958499908,
"learning_rate": 0.0008,
"loss": 3.462801218032837,
"step": 348
},
{
"epoch": 0.021212121212121213,
"grad_norm": 0.22903914749622345,
"learning_rate": 0.0008,
"loss": 3.542144536972046,
"step": 350
},
{
"epoch": 0.021333333333333333,
"grad_norm": 0.26454150676727295,
"learning_rate": 0.0008,
"loss": 3.440065860748291,
"step": 352
},
{
"epoch": 0.021454545454545455,
"grad_norm": 0.2401806265115738,
"learning_rate": 0.0008,
"loss": 3.5220937728881836,
"step": 354
},
{
"epoch": 0.021575757575757575,
"grad_norm": 0.254448801279068,
"learning_rate": 0.0008,
"loss": 3.5693957805633545,
"step": 356
},
{
"epoch": 0.021696969696969697,
"grad_norm": 0.2928229570388794,
"learning_rate": 0.0008,
"loss": 3.4866178035736084,
"step": 358
},
{
"epoch": 0.02181818181818182,
"grad_norm": 0.21962758898735046,
"learning_rate": 0.0008,
"loss": 3.4612746238708496,
"step": 360
},
{
"epoch": 0.02193939393939394,
"grad_norm": 0.1974160522222519,
"learning_rate": 0.0008,
"loss": 3.3538498878479004,
"step": 362
},
{
"epoch": 0.022060606060606062,
"grad_norm": 0.20830541849136353,
"learning_rate": 0.0008,
"loss": 3.4757261276245117,
"step": 364
},
{
"epoch": 0.02218181818181818,
"grad_norm": 0.21993106603622437,
"learning_rate": 0.0008,
"loss": 3.4556262493133545,
"step": 366
},
{
"epoch": 0.022303030303030304,
"grad_norm": 0.24312728643417358,
"learning_rate": 0.0008,
"loss": 3.432683229446411,
"step": 368
},
{
"epoch": 0.022424242424242423,
"grad_norm": 0.20460182428359985,
"learning_rate": 0.0008,
"loss": 3.4205639362335205,
"step": 370
},
{
"epoch": 0.022545454545454546,
"grad_norm": 0.2298145741224289,
"learning_rate": 0.0008,
"loss": 3.408175468444824,
"step": 372
},
{
"epoch": 0.02266666666666667,
"grad_norm": 0.2694374918937683,
"learning_rate": 0.0008,
"loss": 3.3059000968933105,
"step": 374
},
{
"epoch": 0.022787878787878788,
"grad_norm": 0.2133701741695404,
"learning_rate": 0.0008,
"loss": 3.444082260131836,
"step": 376
},
{
"epoch": 0.02290909090909091,
"grad_norm": 0.256267786026001,
"learning_rate": 0.0008,
"loss": 3.3378713130950928,
"step": 378
},
{
"epoch": 0.02303030303030303,
"grad_norm": 0.2305241823196411,
"learning_rate": 0.0008,
"loss": 3.4456775188446045,
"step": 380
},
{
"epoch": 0.023151515151515152,
"grad_norm": 0.2567751407623291,
"learning_rate": 0.0008,
"loss": 3.3252758979797363,
"step": 382
},
{
"epoch": 0.02327272727272727,
"grad_norm": 0.2227526158094406,
"learning_rate": 0.0008,
"loss": 3.3726842403411865,
"step": 384
},
{
"epoch": 0.023393939393939394,
"grad_norm": 0.2189718782901764,
"learning_rate": 0.0008,
"loss": 3.453778028488159,
"step": 386
},
{
"epoch": 0.023515151515151517,
"grad_norm": 0.22577111423015594,
"learning_rate": 0.0008,
"loss": 3.3605237007141113,
"step": 388
},
{
"epoch": 0.023636363636363636,
"grad_norm": 0.19532853364944458,
"learning_rate": 0.0008,
"loss": 3.3928592205047607,
"step": 390
},
{
"epoch": 0.02375757575757576,
"grad_norm": 0.18626846373081207,
"learning_rate": 0.0008,
"loss": 3.4171979427337646,
"step": 392
},
{
"epoch": 0.023878787878787878,
"grad_norm": 0.1684773713350296,
"learning_rate": 0.0008,
"loss": 3.296658515930176,
"step": 394
},
{
"epoch": 0.024,
"grad_norm": 0.18544712662696838,
"learning_rate": 0.0008,
"loss": 3.2549309730529785,
"step": 396
},
{
"epoch": 0.02412121212121212,
"grad_norm": 0.18653690814971924,
"learning_rate": 0.0008,
"loss": 3.3225343227386475,
"step": 398
},
{
"epoch": 0.024242424242424242,
"grad_norm": 0.19594897329807281,
"learning_rate": 0.0008,
"loss": 3.319326877593994,
"step": 400
},
{
"epoch": 0.024363636363636365,
"grad_norm": 0.21079207956790924,
"learning_rate": 0.0008,
"loss": 3.212855339050293,
"step": 402
},
{
"epoch": 0.024484848484848484,
"grad_norm": 0.18678079545497894,
"learning_rate": 0.0008,
"loss": 3.305032730102539,
"step": 404
},
{
"epoch": 0.024606060606060607,
"grad_norm": 0.21345235407352448,
"learning_rate": 0.0008,
"loss": 3.4202210903167725,
"step": 406
},
{
"epoch": 0.024727272727272726,
"grad_norm": 0.24083738029003143,
"learning_rate": 0.0008,
"loss": 3.213761329650879,
"step": 408
},
{
"epoch": 0.02484848484848485,
"grad_norm": 0.20810891687870026,
"learning_rate": 0.0008,
"loss": 3.3248403072357178,
"step": 410
},
{
"epoch": 0.024969696969696968,
"grad_norm": 0.1987171173095703,
"learning_rate": 0.0008,
"loss": 3.2100441455841064,
"step": 412
},
{
"epoch": 0.02509090909090909,
"grad_norm": 0.18585024774074554,
"learning_rate": 0.0008,
"loss": 3.3468050956726074,
"step": 414
},
{
"epoch": 0.025212121212121213,
"grad_norm": 0.17276975512504578,
"learning_rate": 0.0008,
"loss": 3.2669124603271484,
"step": 416
},
{
"epoch": 0.025333333333333333,
"grad_norm": 0.18679523468017578,
"learning_rate": 0.0008,
"loss": 3.305856943130493,
"step": 418
},
{
"epoch": 0.025454545454545455,
"grad_norm": 0.1827041357755661,
"learning_rate": 0.0008,
"loss": 3.293316125869751,
"step": 420
},
{
"epoch": 0.025575757575757575,
"grad_norm": 0.19492028653621674,
"learning_rate": 0.0008,
"loss": 3.2414097785949707,
"step": 422
},
{
"epoch": 0.025696969696969697,
"grad_norm": 0.1917216181755066,
"learning_rate": 0.0008,
"loss": 3.224213123321533,
"step": 424
},
{
"epoch": 0.025818181818181817,
"grad_norm": 0.2075928896665573,
"learning_rate": 0.0008,
"loss": 3.283973217010498,
"step": 426
},
{
"epoch": 0.02593939393939394,
"grad_norm": 0.18258722126483917,
"learning_rate": 0.0008,
"loss": 3.2378711700439453,
"step": 428
},
{
"epoch": 0.026060606060606062,
"grad_norm": 0.17154109477996826,
"learning_rate": 0.0008,
"loss": 3.239980697631836,
"step": 430
},
{
"epoch": 0.02618181818181818,
"grad_norm": 0.18013250827789307,
"learning_rate": 0.0008,
"loss": 3.204636573791504,
"step": 432
},
{
"epoch": 0.026303030303030304,
"grad_norm": 0.17967544496059418,
"learning_rate": 0.0008,
"loss": 3.2176053524017334,
"step": 434
},
{
"epoch": 0.026424242424242423,
"grad_norm": 0.19593407213687897,
"learning_rate": 0.0008,
"loss": 3.2490882873535156,
"step": 436
},
{
"epoch": 0.026545454545454546,
"grad_norm": 0.21105563640594482,
"learning_rate": 0.0008,
"loss": 3.2512412071228027,
"step": 438
},
{
"epoch": 0.02666666666666667,
"grad_norm": 0.21172462403774261,
"learning_rate": 0.0008,
"loss": 3.2083756923675537,
"step": 440
},
{
"epoch": 0.026787878787878788,
"grad_norm": 0.20613734424114227,
"learning_rate": 0.0008,
"loss": 3.215622901916504,
"step": 442
},
{
"epoch": 0.02690909090909091,
"grad_norm": 0.1920866221189499,
"learning_rate": 0.0008,
"loss": 3.172032117843628,
"step": 444
},
{
"epoch": 0.02703030303030303,
"grad_norm": 0.19980253279209137,
"learning_rate": 0.0008,
"loss": 3.255258083343506,
"step": 446
},
{
"epoch": 0.027151515151515152,
"grad_norm": 0.18560081720352173,
"learning_rate": 0.0008,
"loss": 3.254140615463257,
"step": 448
},
{
"epoch": 0.02727272727272727,
"grad_norm": 0.18544770777225494,
"learning_rate": 0.0008,
"loss": 3.1986196041107178,
"step": 450
},
{
"epoch": 0.027393939393939394,
"grad_norm": 0.19149301946163177,
"learning_rate": 0.0008,
"loss": 3.2874796390533447,
"step": 452
},
{
"epoch": 0.027515151515151517,
"grad_norm": 0.19180330634117126,
"learning_rate": 0.0008,
"loss": 3.200110673904419,
"step": 454
},
{
"epoch": 0.027636363636363636,
"grad_norm": 0.18351376056671143,
"learning_rate": 0.0008,
"loss": 3.2116081714630127,
"step": 456
},
{
"epoch": 0.02775757575757576,
"grad_norm": 0.19180741906166077,
"learning_rate": 0.0008,
"loss": 3.124546766281128,
"step": 458
},
{
"epoch": 0.027878787878787878,
"grad_norm": 0.17548531293869019,
"learning_rate": 0.0008,
"loss": 3.128086805343628,
"step": 460
},
{
"epoch": 0.028,
"grad_norm": 0.18065159022808075,
"learning_rate": 0.0008,
"loss": 3.0144083499908447,
"step": 462
},
{
"epoch": 0.02812121212121212,
"grad_norm": 0.23563507199287415,
"learning_rate": 0.0008,
"loss": 3.134948968887329,
"step": 464
},
{
"epoch": 0.028242424242424242,
"grad_norm": 0.25060465931892395,
"learning_rate": 0.0008,
"loss": 3.220275640487671,
"step": 466
},
{
"epoch": 0.028363636363636365,
"grad_norm": 0.19759942591190338,
"learning_rate": 0.0008,
"loss": 3.1709189414978027,
"step": 468
},
{
"epoch": 0.028484848484848484,
"grad_norm": 0.17349790036678314,
"learning_rate": 0.0008,
"loss": 3.133767604827881,
"step": 470
},
{
"epoch": 0.028606060606060607,
"grad_norm": 0.16103310883045197,
"learning_rate": 0.0008,
"loss": 3.1511073112487793,
"step": 472
},
{
"epoch": 0.028727272727272726,
"grad_norm": 0.1536155343055725,
"learning_rate": 0.0008,
"loss": 3.153458833694458,
"step": 474
},
{
"epoch": 0.02884848484848485,
"grad_norm": 0.16233953833580017,
"learning_rate": 0.0008,
"loss": 3.103088855743408,
"step": 476
},
{
"epoch": 0.028969696969696968,
"grad_norm": 0.1503172218799591,
"learning_rate": 0.0008,
"loss": 3.1187267303466797,
"step": 478
},
{
"epoch": 0.02909090909090909,
"grad_norm": 0.15062548220157623,
"learning_rate": 0.0008,
"loss": 3.062068462371826,
"step": 480
},
{
"epoch": 0.029212121212121214,
"grad_norm": 0.14754259586334229,
"learning_rate": 0.0008,
"loss": 3.1466453075408936,
"step": 482
},
{
"epoch": 0.029333333333333333,
"grad_norm": 0.16204990446567535,
"learning_rate": 0.0008,
"loss": 3.0818166732788086,
"step": 484
},
{
"epoch": 0.029454545454545455,
"grad_norm": 0.2318359762430191,
"learning_rate": 0.0008,
"loss": 3.1093335151672363,
"step": 486
},
{
"epoch": 0.029575757575757575,
"grad_norm": 0.1841193288564682,
"learning_rate": 0.0008,
"loss": 3.2149367332458496,
"step": 488
},
{
"epoch": 0.029696969696969697,
"grad_norm": 0.2003197818994522,
"learning_rate": 0.0008,
"loss": 3.109757423400879,
"step": 490
},
{
"epoch": 0.029818181818181817,
"grad_norm": 0.1707594245672226,
"learning_rate": 0.0008,
"loss": 3.0052380561828613,
"step": 492
},
{
"epoch": 0.02993939393939394,
"grad_norm": 0.17520278692245483,
"learning_rate": 0.0008,
"loss": 3.1650912761688232,
"step": 494
},
{
"epoch": 0.030060606060606062,
"grad_norm": 0.18182702362537384,
"learning_rate": 0.0008,
"loss": 3.1383142471313477,
"step": 496
},
{
"epoch": 0.03018181818181818,
"grad_norm": 0.17171436548233032,
"learning_rate": 0.0008,
"loss": 3.0356242656707764,
"step": 498
},
{
"epoch": 0.030303030303030304,
"grad_norm": 0.17782188951969147,
"learning_rate": 0.0008,
"loss": 3.143406391143799,
"step": 500
},
{
"epoch": 0.030424242424242423,
"grad_norm": 0.2386026680469513,
"learning_rate": 0.0008,
"loss": 3.011464834213257,
"step": 502
},
{
"epoch": 0.030545454545454546,
"grad_norm": 0.16823378205299377,
"learning_rate": 0.0008,
"loss": 3.1168160438537598,
"step": 504
},
{
"epoch": 0.030666666666666665,
"grad_norm": 0.14969979226589203,
"learning_rate": 0.0008,
"loss": 3.0774800777435303,
"step": 506
},
{
"epoch": 0.030787878787878788,
"grad_norm": 0.16551975905895233,
"learning_rate": 0.0008,
"loss": 3.06292462348938,
"step": 508
},
{
"epoch": 0.03090909090909091,
"grad_norm": 0.217863991856575,
"learning_rate": 0.0008,
"loss": 3.0433037281036377,
"step": 510
},
{
"epoch": 0.03103030303030303,
"grad_norm": 0.16006837785243988,
"learning_rate": 0.0008,
"loss": 3.092416763305664,
"step": 512
},
{
"epoch": 0.031151515151515152,
"grad_norm": 0.1552363634109497,
"learning_rate": 0.0008,
"loss": 3.084879159927368,
"step": 514
},
{
"epoch": 0.03127272727272727,
"grad_norm": 0.14629562199115753,
"learning_rate": 0.0008,
"loss": 3.0902342796325684,
"step": 516
},
{
"epoch": 0.03139393939393939,
"grad_norm": 0.14411231875419617,
"learning_rate": 0.0008,
"loss": 2.986879348754883,
"step": 518
},
{
"epoch": 0.03151515151515152,
"grad_norm": 0.16847993433475494,
"learning_rate": 0.0008,
"loss": 3.0822856426239014,
"step": 520
},
{
"epoch": 0.031636363636363636,
"grad_norm": 0.149409681558609,
"learning_rate": 0.0008,
"loss": 3.018641948699951,
"step": 522
},
{
"epoch": 0.031757575757575755,
"grad_norm": 0.14600564539432526,
"learning_rate": 0.0008,
"loss": 3.0455117225646973,
"step": 524
},
{
"epoch": 0.03187878787878788,
"grad_norm": 0.1614578664302826,
"learning_rate": 0.0008,
"loss": 3.116377353668213,
"step": 526
},
{
"epoch": 0.032,
"grad_norm": 0.14525499939918518,
"learning_rate": 0.0008,
"loss": 2.985255718231201,
"step": 528
},
{
"epoch": 0.03212121212121212,
"grad_norm": 0.150455042719841,
"learning_rate": 0.0008,
"loss": 3.0385396480560303,
"step": 530
},
{
"epoch": 0.03224242424242424,
"grad_norm": 0.16066277027130127,
"learning_rate": 0.0008,
"loss": 3.034322738647461,
"step": 532
},
{
"epoch": 0.032363636363636365,
"grad_norm": 0.15391060709953308,
"learning_rate": 0.0008,
"loss": 3.0366103649139404,
"step": 534
},
{
"epoch": 0.032484848484848484,
"grad_norm": 0.14887331426143646,
"learning_rate": 0.0008,
"loss": 2.9216997623443604,
"step": 536
},
{
"epoch": 0.032606060606060604,
"grad_norm": 0.15154264867305756,
"learning_rate": 0.0008,
"loss": 2.999504327774048,
"step": 538
},
{
"epoch": 0.03272727272727273,
"grad_norm": 0.14853991568088531,
"learning_rate": 0.0008,
"loss": 2.987461566925049,
"step": 540
},
{
"epoch": 0.03284848484848485,
"grad_norm": 0.1701185405254364,
"learning_rate": 0.0008,
"loss": 3.038562059402466,
"step": 542
},
{
"epoch": 0.03296969696969697,
"grad_norm": 0.1607816368341446,
"learning_rate": 0.0008,
"loss": 2.990669012069702,
"step": 544
},
{
"epoch": 0.03309090909090909,
"grad_norm": 0.15095685422420502,
"learning_rate": 0.0008,
"loss": 3.0428266525268555,
"step": 546
},
{
"epoch": 0.033212121212121214,
"grad_norm": 0.14433546364307404,
"learning_rate": 0.0008,
"loss": 2.9407718181610107,
"step": 548
},
{
"epoch": 0.03333333333333333,
"grad_norm": 0.14404350519180298,
"learning_rate": 0.0008,
"loss": 2.9540135860443115,
"step": 550
},
{
"epoch": 0.03345454545454545,
"grad_norm": 0.16146287322044373,
"learning_rate": 0.0008,
"loss": 2.9493942260742188,
"step": 552
},
{
"epoch": 0.03357575757575758,
"grad_norm": 0.1664876937866211,
"learning_rate": 0.0008,
"loss": 2.9082322120666504,
"step": 554
},
{
"epoch": 0.0336969696969697,
"grad_norm": 0.15475095808506012,
"learning_rate": 0.0008,
"loss": 3.0174684524536133,
"step": 556
},
{
"epoch": 0.03381818181818182,
"grad_norm": 0.14952117204666138,
"learning_rate": 0.0008,
"loss": 3.0506439208984375,
"step": 558
},
{
"epoch": 0.03393939393939394,
"grad_norm": 0.16758497059345245,
"learning_rate": 0.0008,
"loss": 3.0673019886016846,
"step": 560
},
{
"epoch": 0.03406060606060606,
"grad_norm": 0.14651420712471008,
"learning_rate": 0.0008,
"loss": 2.920955181121826,
"step": 562
},
{
"epoch": 0.03418181818181818,
"grad_norm": 0.17143522202968597,
"learning_rate": 0.0008,
"loss": 3.0434558391571045,
"step": 564
},
{
"epoch": 0.0343030303030303,
"grad_norm": 0.1426212042570114,
"learning_rate": 0.0008,
"loss": 2.9456403255462646,
"step": 566
},
{
"epoch": 0.03442424242424243,
"grad_norm": 0.14558342099189758,
"learning_rate": 0.0008,
"loss": 2.9615488052368164,
"step": 568
},
{
"epoch": 0.034545454545454546,
"grad_norm": 0.1692395806312561,
"learning_rate": 0.0008,
"loss": 2.9330313205718994,
"step": 570
},
{
"epoch": 0.034666666666666665,
"grad_norm": 0.17397549748420715,
"learning_rate": 0.0008,
"loss": 2.9596598148345947,
"step": 572
},
{
"epoch": 0.03478787878787879,
"grad_norm": 0.16361340880393982,
"learning_rate": 0.0008,
"loss": 3.052358865737915,
"step": 574
},
{
"epoch": 0.03490909090909091,
"grad_norm": 0.16779865324497223,
"learning_rate": 0.0008,
"loss": 2.964066743850708,
"step": 576
},
{
"epoch": 0.03503030303030303,
"grad_norm": 0.13800302147865295,
"learning_rate": 0.0008,
"loss": 2.98765230178833,
"step": 578
},
{
"epoch": 0.03515151515151515,
"grad_norm": 0.13935010135173798,
"learning_rate": 0.0008,
"loss": 3.018916368484497,
"step": 580
},
{
"epoch": 0.035272727272727275,
"grad_norm": 0.14222946763038635,
"learning_rate": 0.0008,
"loss": 2.873932361602783,
"step": 582
},
{
"epoch": 0.035393939393939394,
"grad_norm": 0.15399911999702454,
"learning_rate": 0.0008,
"loss": 3.022982120513916,
"step": 584
},
{
"epoch": 0.035515151515151513,
"grad_norm": 0.14751893281936646,
"learning_rate": 0.0008,
"loss": 2.9483044147491455,
"step": 586
},
{
"epoch": 0.03563636363636364,
"grad_norm": 0.14612935483455658,
"learning_rate": 0.0008,
"loss": 3.0079457759857178,
"step": 588
},
{
"epoch": 0.03575757575757576,
"grad_norm": 0.14603056013584137,
"learning_rate": 0.0008,
"loss": 3.0186915397644043,
"step": 590
},
{
"epoch": 0.03587878787878788,
"grad_norm": 0.1394486427307129,
"learning_rate": 0.0008,
"loss": 2.955667495727539,
"step": 592
},
{
"epoch": 0.036,
"grad_norm": 0.1390790194272995,
"learning_rate": 0.0008,
"loss": 2.9570231437683105,
"step": 594
},
{
"epoch": 0.03612121212121212,
"grad_norm": 0.1442171335220337,
"learning_rate": 0.0008,
"loss": 2.97532057762146,
"step": 596
},
{
"epoch": 0.03624242424242424,
"grad_norm": 0.15127277374267578,
"learning_rate": 0.0008,
"loss": 2.910494327545166,
"step": 598
},
{
"epoch": 0.03636363636363636,
"grad_norm": 0.1394735723733902,
"learning_rate": 0.0008,
"loss": 2.924168109893799,
"step": 600
},
{
"epoch": 0.03648484848484849,
"grad_norm": 0.1540086269378662,
"learning_rate": 0.0008,
"loss": 2.9845988750457764,
"step": 602
},
{
"epoch": 0.03660606060606061,
"grad_norm": 0.17311856150627136,
"learning_rate": 0.0008,
"loss": 2.927744150161743,
"step": 604
},
{
"epoch": 0.036727272727272726,
"grad_norm": 0.14371098577976227,
"learning_rate": 0.0008,
"loss": 2.922285556793213,
"step": 606
},
{
"epoch": 0.036848484848484846,
"grad_norm": 0.13951198756694794,
"learning_rate": 0.0008,
"loss": 2.9370288848876953,
"step": 608
},
{
"epoch": 0.03696969696969697,
"grad_norm": 0.14434067904949188,
"learning_rate": 0.0008,
"loss": 2.9697577953338623,
"step": 610
},
{
"epoch": 0.03709090909090909,
"grad_norm": 0.14919859170913696,
"learning_rate": 0.0008,
"loss": 2.979060649871826,
"step": 612
},
{
"epoch": 0.03721212121212121,
"grad_norm": 0.14238537847995758,
"learning_rate": 0.0008,
"loss": 2.923292875289917,
"step": 614
},
{
"epoch": 0.037333333333333336,
"grad_norm": 0.13626030087471008,
"learning_rate": 0.0008,
"loss": 2.858288288116455,
"step": 616
},
{
"epoch": 0.037454545454545456,
"grad_norm": 0.13980725407600403,
"learning_rate": 0.0008,
"loss": 2.847784996032715,
"step": 618
},
{
"epoch": 0.037575757575757575,
"grad_norm": 0.14716844260692596,
"learning_rate": 0.0008,
"loss": 2.845705032348633,
"step": 620
},
{
"epoch": 0.037696969696969694,
"grad_norm": 0.1418210119009018,
"learning_rate": 0.0008,
"loss": 2.953434944152832,
"step": 622
},
{
"epoch": 0.03781818181818182,
"grad_norm": 0.13869202136993408,
"learning_rate": 0.0008,
"loss": 2.9007976055145264,
"step": 624
},
{
"epoch": 0.03793939393939394,
"grad_norm": 0.13616803288459778,
"learning_rate": 0.0008,
"loss": 2.94415545463562,
"step": 626
},
{
"epoch": 0.03806060606060606,
"grad_norm": 0.1417037397623062,
"learning_rate": 0.0008,
"loss": 2.923724412918091,
"step": 628
},
{
"epoch": 0.038181818181818185,
"grad_norm": 0.13439780473709106,
"learning_rate": 0.0008,
"loss": 2.9307355880737305,
"step": 630
},
{
"epoch": 0.038303030303030304,
"grad_norm": 0.13781346380710602,
"learning_rate": 0.0008,
"loss": 2.954785108566284,
"step": 632
},
{
"epoch": 0.03842424242424242,
"grad_norm": 0.14403529465198517,
"learning_rate": 0.0008,
"loss": 2.9103217124938965,
"step": 634
},
{
"epoch": 0.03854545454545454,
"grad_norm": 0.25868332386016846,
"learning_rate": 0.0008,
"loss": 2.9291200637817383,
"step": 636
},
{
"epoch": 0.03866666666666667,
"grad_norm": 0.13370473682880402,
"learning_rate": 0.0008,
"loss": 2.9023022651672363,
"step": 638
},
{
"epoch": 0.03878787878787879,
"grad_norm": 0.13921067118644714,
"learning_rate": 0.0008,
"loss": 2.8829033374786377,
"step": 640
},
{
"epoch": 0.03890909090909091,
"grad_norm": 0.14794135093688965,
"learning_rate": 0.0008,
"loss": 2.838698387145996,
"step": 642
},
{
"epoch": 0.03903030303030303,
"grad_norm": 0.1344001144170761,
"learning_rate": 0.0008,
"loss": 2.882659912109375,
"step": 644
},
{
"epoch": 0.03915151515151515,
"grad_norm": 0.1305534392595291,
"learning_rate": 0.0008,
"loss": 2.863396406173706,
"step": 646
},
{
"epoch": 0.03927272727272727,
"grad_norm": 0.1409977525472641,
"learning_rate": 0.0008,
"loss": 2.8610992431640625,
"step": 648
},
{
"epoch": 0.03939393939393939,
"grad_norm": 0.1957748830318451,
"learning_rate": 0.0008,
"loss": 2.8816120624542236,
"step": 650
},
{
"epoch": 0.03951515151515152,
"grad_norm": 0.12930573523044586,
"learning_rate": 0.0008,
"loss": 2.9270267486572266,
"step": 652
},
{
"epoch": 0.039636363636363636,
"grad_norm": 0.1313532143831253,
"learning_rate": 0.0008,
"loss": 2.8601999282836914,
"step": 654
},
{
"epoch": 0.039757575757575755,
"grad_norm": 0.1319667398929596,
"learning_rate": 0.0008,
"loss": 2.951965808868408,
"step": 656
},
{
"epoch": 0.03987878787878788,
"grad_norm": 0.14966745674610138,
"learning_rate": 0.0008,
"loss": 2.9171664714813232,
"step": 658
},
{
"epoch": 0.04,
"grad_norm": 0.15474537014961243,
"learning_rate": 0.0008,
"loss": 2.962900161743164,
"step": 660
},
{
"epoch": 0.04012121212121212,
"grad_norm": 0.1682961881160736,
"learning_rate": 0.0008,
"loss": 2.9665536880493164,
"step": 662
},
{
"epoch": 0.04024242424242424,
"grad_norm": 0.1524871587753296,
"learning_rate": 0.0008,
"loss": 2.963721513748169,
"step": 664
},
{
"epoch": 0.040363636363636365,
"grad_norm": 0.16089048981666565,
"learning_rate": 0.0008,
"loss": 2.929103136062622,
"step": 666
},
{
"epoch": 0.040484848484848485,
"grad_norm": 0.15341050922870636,
"learning_rate": 0.0008,
"loss": 2.8558969497680664,
"step": 668
},
{
"epoch": 0.040606060606060604,
"grad_norm": 0.13815748691558838,
"learning_rate": 0.0008,
"loss": 2.9092814922332764,
"step": 670
},
{
"epoch": 0.04072727272727273,
"grad_norm": 0.13577315211296082,
"learning_rate": 0.0008,
"loss": 2.821869373321533,
"step": 672
},
{
"epoch": 0.04084848484848485,
"grad_norm": 0.1391841471195221,
"learning_rate": 0.0008,
"loss": 2.8810579776763916,
"step": 674
},
{
"epoch": 0.04096969696969697,
"grad_norm": 0.14720574021339417,
"learning_rate": 0.0008,
"loss": 2.851976156234741,
"step": 676
},
{
"epoch": 0.04109090909090909,
"grad_norm": 0.21025408804416656,
"learning_rate": 0.0008,
"loss": 2.9023654460906982,
"step": 678
},
{
"epoch": 0.041212121212121214,
"grad_norm": 0.13861164450645447,
"learning_rate": 0.0008,
"loss": 2.8224081993103027,
"step": 680
},
{
"epoch": 0.04133333333333333,
"grad_norm": 0.12873832881450653,
"learning_rate": 0.0008,
"loss": 2.8723981380462646,
"step": 682
},
{
"epoch": 0.04145454545454545,
"grad_norm": 0.1253303587436676,
"learning_rate": 0.0008,
"loss": 2.810709238052368,
"step": 684
},
{
"epoch": 0.04157575757575758,
"grad_norm": 0.12730298936367035,
"learning_rate": 0.0008,
"loss": 2.9045820236206055,
"step": 686
},
{
"epoch": 0.0416969696969697,
"grad_norm": 0.1321014016866684,
"learning_rate": 0.0008,
"loss": 2.7892556190490723,
"step": 688
},
{
"epoch": 0.04181818181818182,
"grad_norm": 0.1333041787147522,
"learning_rate": 0.0008,
"loss": 2.772632122039795,
"step": 690
},
{
"epoch": 0.041939393939393936,
"grad_norm": 0.1299896240234375,
"learning_rate": 0.0008,
"loss": 2.806184768676758,
"step": 692
},
{
"epoch": 0.04206060606060606,
"grad_norm": 0.135824516415596,
"learning_rate": 0.0008,
"loss": 2.8782715797424316,
"step": 694
},
{
"epoch": 0.04218181818181818,
"grad_norm": 0.13341370224952698,
"learning_rate": 0.0008,
"loss": 2.875131607055664,
"step": 696
},
{
"epoch": 0.0423030303030303,
"grad_norm": 0.12718485295772552,
"learning_rate": 0.0008,
"loss": 2.9469480514526367,
"step": 698
},
{
"epoch": 0.04242424242424243,
"grad_norm": 0.1303737759590149,
"learning_rate": 0.0008,
"loss": 2.850902557373047,
"step": 700
},
{
"epoch": 0.042545454545454546,
"grad_norm": 0.13085296750068665,
"learning_rate": 0.0008,
"loss": 2.8760080337524414,
"step": 702
},
{
"epoch": 0.042666666666666665,
"grad_norm": 0.13041667640209198,
"learning_rate": 0.0008,
"loss": 2.801133155822754,
"step": 704
},
{
"epoch": 0.042787878787878784,
"grad_norm": 0.1376514583826065,
"learning_rate": 0.0008,
"loss": 2.8532888889312744,
"step": 706
},
{
"epoch": 0.04290909090909091,
"grad_norm": 0.13842840492725372,
"learning_rate": 0.0008,
"loss": 2.8160784244537354,
"step": 708
},
{
"epoch": 0.04303030303030303,
"grad_norm": 0.12489556521177292,
"learning_rate": 0.0008,
"loss": 2.7616965770721436,
"step": 710
},
{
"epoch": 0.04315151515151515,
"grad_norm": 0.12598678469657898,
"learning_rate": 0.0008,
"loss": 2.7892892360687256,
"step": 712
},
{
"epoch": 0.043272727272727275,
"grad_norm": 0.12128093093633652,
"learning_rate": 0.0008,
"loss": 2.8648624420166016,
"step": 714
},
{
"epoch": 0.043393939393939394,
"grad_norm": 0.12195690721273422,
"learning_rate": 0.0008,
"loss": 2.8440792560577393,
"step": 716
},
{
"epoch": 0.043515151515151514,
"grad_norm": 0.1399867683649063,
"learning_rate": 0.0008,
"loss": 2.8321645259857178,
"step": 718
},
{
"epoch": 0.04363636363636364,
"grad_norm": 0.13280552625656128,
"learning_rate": 0.0008,
"loss": 2.9339985847473145,
"step": 720
},
{
"epoch": 0.04375757575757576,
"grad_norm": 0.1302645355463028,
"learning_rate": 0.0008,
"loss": 2.823873996734619,
"step": 722
},
{
"epoch": 0.04387878787878788,
"grad_norm": 0.12091700732707977,
"learning_rate": 0.0008,
"loss": 2.8100738525390625,
"step": 724
},
{
"epoch": 0.044,
"grad_norm": 0.1299038976430893,
"learning_rate": 0.0008,
"loss": 2.814293146133423,
"step": 726
},
{
"epoch": 0.044121212121212124,
"grad_norm": 0.12724053859710693,
"learning_rate": 0.0008,
"loss": 2.795933723449707,
"step": 728
},
{
"epoch": 0.04424242424242424,
"grad_norm": 0.12532362341880798,
"learning_rate": 0.0008,
"loss": 2.7980096340179443,
"step": 730
},
{
"epoch": 0.04436363636363636,
"grad_norm": 0.12794388830661774,
"learning_rate": 0.0008,
"loss": 2.8418681621551514,
"step": 732
},
{
"epoch": 0.04448484848484849,
"grad_norm": 0.12713198363780975,
"learning_rate": 0.0008,
"loss": 2.835165023803711,
"step": 734
},
{
"epoch": 0.04460606060606061,
"grad_norm": 0.14070792496204376,
"learning_rate": 0.0008,
"loss": 2.846919298171997,
"step": 736
},
{
"epoch": 0.04472727272727273,
"grad_norm": 0.12815815210342407,
"learning_rate": 0.0008,
"loss": 2.8320844173431396,
"step": 738
},
{
"epoch": 0.044848484848484846,
"grad_norm": 0.12327364087104797,
"learning_rate": 0.0008,
"loss": 2.845261573791504,
"step": 740
},
{
"epoch": 0.04496969696969697,
"grad_norm": 0.1268470287322998,
"learning_rate": 0.0008,
"loss": 2.7880043983459473,
"step": 742
},
{
"epoch": 0.04509090909090909,
"grad_norm": 0.11996026337146759,
"learning_rate": 0.0008,
"loss": 2.8137447834014893,
"step": 744
},
{
"epoch": 0.04521212121212121,
"grad_norm": 0.12582260370254517,
"learning_rate": 0.0008,
"loss": 2.800065279006958,
"step": 746
},
{
"epoch": 0.04533333333333334,
"grad_norm": 0.11716218292713165,
"learning_rate": 0.0008,
"loss": 2.8071770668029785,
"step": 748
},
{
"epoch": 0.045454545454545456,
"grad_norm": 0.11834251135587692,
"learning_rate": 0.0008,
"loss": 2.8258886337280273,
"step": 750
},
{
"epoch": 0.045575757575757575,
"grad_norm": 0.13357476890087128,
"learning_rate": 0.0008,
"loss": 2.8557024002075195,
"step": 752
},
{
"epoch": 0.045696969696969694,
"grad_norm": 0.122672900557518,
"learning_rate": 0.0008,
"loss": 2.7642595767974854,
"step": 754
},
{
"epoch": 0.04581818181818182,
"grad_norm": 0.13104338943958282,
"learning_rate": 0.0008,
"loss": 2.7938239574432373,
"step": 756
},
{
"epoch": 0.04593939393939394,
"grad_norm": 0.12123287469148636,
"learning_rate": 0.0008,
"loss": 2.8204739093780518,
"step": 758
},
{
"epoch": 0.04606060606060606,
"grad_norm": 0.11743812263011932,
"learning_rate": 0.0008,
"loss": 2.8480348587036133,
"step": 760
},
{
"epoch": 0.046181818181818185,
"grad_norm": 0.1227068156003952,
"learning_rate": 0.0008,
"loss": 2.8580687046051025,
"step": 762
},
{
"epoch": 0.046303030303030304,
"grad_norm": 0.12315574288368225,
"learning_rate": 0.0008,
"loss": 2.8673183917999268,
"step": 764
},
{
"epoch": 0.04642424242424242,
"grad_norm": 0.123519167304039,
"learning_rate": 0.0008,
"loss": 2.796816825866699,
"step": 766
},
{
"epoch": 0.04654545454545454,
"grad_norm": 0.12256579846143723,
"learning_rate": 0.0008,
"loss": 2.869438409805298,
"step": 768
},
{
"epoch": 0.04666666666666667,
"grad_norm": 0.12488561868667603,
"learning_rate": 0.0008,
"loss": 2.72646427154541,
"step": 770
},
{
"epoch": 0.04678787878787879,
"grad_norm": 0.11663568764925003,
"learning_rate": 0.0008,
"loss": 2.8097939491271973,
"step": 772
},
{
"epoch": 0.04690909090909091,
"grad_norm": 0.12199313938617706,
"learning_rate": 0.0008,
"loss": 2.8485167026519775,
"step": 774
},
{
"epoch": 0.04703030303030303,
"grad_norm": 0.11423022300004959,
"learning_rate": 0.0008,
"loss": 2.832976818084717,
"step": 776
},
{
"epoch": 0.04715151515151515,
"grad_norm": 0.12107249349355698,
"learning_rate": 0.0008,
"loss": 2.755948543548584,
"step": 778
},
{
"epoch": 0.04727272727272727,
"grad_norm": 0.11858741194009781,
"learning_rate": 0.0008,
"loss": 2.862621307373047,
"step": 780
},
{
"epoch": 0.04739393939393939,
"grad_norm": 0.11461203545331955,
"learning_rate": 0.0008,
"loss": 2.733175039291382,
"step": 782
},
{
"epoch": 0.04751515151515152,
"grad_norm": 0.11895351856946945,
"learning_rate": 0.0008,
"loss": 2.708164691925049,
"step": 784
},
{
"epoch": 0.047636363636363636,
"grad_norm": 0.12406233698129654,
"learning_rate": 0.0008,
"loss": 2.725911855697632,
"step": 786
},
{
"epoch": 0.047757575757575756,
"grad_norm": 0.1257525086402893,
"learning_rate": 0.0008,
"loss": 2.876732587814331,
"step": 788
},
{
"epoch": 0.04787878787878788,
"grad_norm": 0.1256421059370041,
"learning_rate": 0.0008,
"loss": 2.7962048053741455,
"step": 790
},
{
"epoch": 0.048,
"grad_norm": 0.142322838306427,
"learning_rate": 0.0008,
"loss": 2.7867467403411865,
"step": 792
},
{
"epoch": 0.04812121212121212,
"grad_norm": 0.12080056965351105,
"learning_rate": 0.0008,
"loss": 2.728219747543335,
"step": 794
},
{
"epoch": 0.04824242424242424,
"grad_norm": 0.11979342252016068,
"learning_rate": 0.0008,
"loss": 2.723640203475952,
"step": 796
},
{
"epoch": 0.048363636363636366,
"grad_norm": 0.12700845301151276,
"learning_rate": 0.0008,
"loss": 2.8942372798919678,
"step": 798
},
{
"epoch": 0.048484848484848485,
"grad_norm": 0.12027334421873093,
"learning_rate": 0.0008,
"loss": 2.82177734375,
"step": 800
},
{
"epoch": 0.048606060606060604,
"grad_norm": 0.12404719740152359,
"learning_rate": 0.0008,
"loss": 2.803739070892334,
"step": 802
},
{
"epoch": 0.04872727272727273,
"grad_norm": 0.121683269739151,
"learning_rate": 0.0008,
"loss": 2.7722697257995605,
"step": 804
},
{
"epoch": 0.04884848484848485,
"grad_norm": 0.1171930730342865,
"learning_rate": 0.0008,
"loss": 2.8435051441192627,
"step": 806
},
{
"epoch": 0.04896969696969697,
"grad_norm": 0.11590387672185898,
"learning_rate": 0.0008,
"loss": 2.7791128158569336,
"step": 808
},
{
"epoch": 0.04909090909090909,
"grad_norm": 0.11669088900089264,
"learning_rate": 0.0008,
"loss": 2.815957546234131,
"step": 810
},
{
"epoch": 0.049212121212121214,
"grad_norm": 0.12032285332679749,
"learning_rate": 0.0008,
"loss": 2.7630574703216553,
"step": 812
},
{
"epoch": 0.04933333333333333,
"grad_norm": 0.11965234577655792,
"learning_rate": 0.0008,
"loss": 2.8138625621795654,
"step": 814
},
{
"epoch": 0.04945454545454545,
"grad_norm": 0.1195046529173851,
"learning_rate": 0.0008,
"loss": 2.842541217803955,
"step": 816
},
{
"epoch": 0.04957575757575758,
"grad_norm": 0.1200253888964653,
"learning_rate": 0.0008,
"loss": 2.776521921157837,
"step": 818
},
{
"epoch": 0.0496969696969697,
"grad_norm": 0.11728650331497192,
"learning_rate": 0.0008,
"loss": 2.8308160305023193,
"step": 820
},
{
"epoch": 0.04981818181818182,
"grad_norm": 0.12323697656393051,
"learning_rate": 0.0008,
"loss": 2.760850667953491,
"step": 822
},
{
"epoch": 0.049939393939393936,
"grad_norm": 0.12108653038740158,
"learning_rate": 0.0008,
"loss": 2.7815420627593994,
"step": 824
},
{
"epoch": 0.05006060606060606,
"grad_norm": 0.11688020080327988,
"learning_rate": 0.0008,
"loss": 2.8114840984344482,
"step": 826
},
{
"epoch": 0.05018181818181818,
"grad_norm": 0.12375954538583755,
"learning_rate": 0.0008,
"loss": 2.7183804512023926,
"step": 828
},
{
"epoch": 0.0503030303030303,
"grad_norm": 0.12011045962572098,
"learning_rate": 0.0008,
"loss": 2.8393380641937256,
"step": 830
},
{
"epoch": 0.05042424242424243,
"grad_norm": 0.1141817569732666,
"learning_rate": 0.0008,
"loss": 2.763029098510742,
"step": 832
},
{
"epoch": 0.050545454545454546,
"grad_norm": 0.12080072611570358,
"learning_rate": 0.0008,
"loss": 2.7412564754486084,
"step": 834
},
{
"epoch": 0.050666666666666665,
"grad_norm": 0.11752299219369888,
"learning_rate": 0.0008,
"loss": 2.831782102584839,
"step": 836
},
{
"epoch": 0.050787878787878785,
"grad_norm": 0.1883753538131714,
"learning_rate": 0.0008,
"loss": 2.801042318344116,
"step": 838
},
{
"epoch": 0.05090909090909091,
"grad_norm": 0.14781929552555084,
"learning_rate": 0.0008,
"loss": 2.726942777633667,
"step": 840
},
{
"epoch": 0.05103030303030303,
"grad_norm": 0.12364400923252106,
"learning_rate": 0.0008,
"loss": 2.7975711822509766,
"step": 842
},
{
"epoch": 0.05115151515151515,
"grad_norm": 0.11527018994092941,
"learning_rate": 0.0008,
"loss": 2.783881425857544,
"step": 844
},
{
"epoch": 0.051272727272727275,
"grad_norm": 0.12733705341815948,
"learning_rate": 0.0008,
"loss": 2.744743824005127,
"step": 846
},
{
"epoch": 0.051393939393939395,
"grad_norm": 0.11549273133277893,
"learning_rate": 0.0008,
"loss": 2.835817575454712,
"step": 848
},
{
"epoch": 0.051515151515151514,
"grad_norm": 0.11381974071264267,
"learning_rate": 0.0008,
"loss": 2.78991436958313,
"step": 850
},
{
"epoch": 0.05163636363636363,
"grad_norm": 0.11358799040317535,
"learning_rate": 0.0008,
"loss": 2.8159079551696777,
"step": 852
},
{
"epoch": 0.05175757575757576,
"grad_norm": 0.1236659586429596,
"learning_rate": 0.0008,
"loss": 2.7812955379486084,
"step": 854
},
{
"epoch": 0.05187878787878788,
"grad_norm": 0.12129868566989899,
"learning_rate": 0.0008,
"loss": 2.7351558208465576,
"step": 856
},
{
"epoch": 0.052,
"grad_norm": 0.11662942916154861,
"learning_rate": 0.0008,
"loss": 2.871737480163574,
"step": 858
},
{
"epoch": 0.052121212121212124,
"grad_norm": 0.1330779492855072,
"learning_rate": 0.0008,
"loss": 2.798617124557495,
"step": 860
},
{
"epoch": 0.05224242424242424,
"grad_norm": 0.11315251141786575,
"learning_rate": 0.0008,
"loss": 2.815581798553467,
"step": 862
},
{
"epoch": 0.05236363636363636,
"grad_norm": 0.11033479124307632,
"learning_rate": 0.0008,
"loss": 2.699082612991333,
"step": 864
},
{
"epoch": 0.05248484848484848,
"grad_norm": 0.10872790962457657,
"learning_rate": 0.0008,
"loss": 2.780463218688965,
"step": 866
},
{
"epoch": 0.05260606060606061,
"grad_norm": 0.11883790791034698,
"learning_rate": 0.0008,
"loss": 2.7807490825653076,
"step": 868
},
{
"epoch": 0.05272727272727273,
"grad_norm": 0.11434616148471832,
"learning_rate": 0.0008,
"loss": 2.812364101409912,
"step": 870
},
{
"epoch": 0.052848484848484846,
"grad_norm": 0.11327607184648514,
"learning_rate": 0.0008,
"loss": 2.7629072666168213,
"step": 872
},
{
"epoch": 0.05296969696969697,
"grad_norm": 0.11344680190086365,
"learning_rate": 0.0008,
"loss": 2.8015716075897217,
"step": 874
},
{
"epoch": 0.05309090909090909,
"grad_norm": 0.11428236961364746,
"learning_rate": 0.0008,
"loss": 2.824066638946533,
"step": 876
},
{
"epoch": 0.05321212121212121,
"grad_norm": 0.1124577671289444,
"learning_rate": 0.0008,
"loss": 2.8407106399536133,
"step": 878
},
{
"epoch": 0.05333333333333334,
"grad_norm": 0.1151081845164299,
"learning_rate": 0.0008,
"loss": 2.758446455001831,
"step": 880
},
{
"epoch": 0.053454545454545456,
"grad_norm": 0.11284763365983963,
"learning_rate": 0.0008,
"loss": 2.7593636512756348,
"step": 882
},
{
"epoch": 0.053575757575757575,
"grad_norm": 0.11293981224298477,
"learning_rate": 0.0008,
"loss": 2.769613027572632,
"step": 884
},
{
"epoch": 0.053696969696969694,
"grad_norm": 0.11765503883361816,
"learning_rate": 0.0008,
"loss": 2.9163100719451904,
"step": 886
},
{
"epoch": 0.05381818181818182,
"grad_norm": 0.11500603705644608,
"learning_rate": 0.0008,
"loss": 2.7763619422912598,
"step": 888
},
{
"epoch": 0.05393939393939394,
"grad_norm": 0.10888758301734924,
"learning_rate": 0.0008,
"loss": 2.7734885215759277,
"step": 890
},
{
"epoch": 0.05406060606060606,
"grad_norm": 0.11274725198745728,
"learning_rate": 0.0008,
"loss": 2.8008785247802734,
"step": 892
},
{
"epoch": 0.054181818181818185,
"grad_norm": 0.11535163968801498,
"learning_rate": 0.0008,
"loss": 2.8337900638580322,
"step": 894
},
{
"epoch": 0.054303030303030304,
"grad_norm": 0.11355795711278915,
"learning_rate": 0.0008,
"loss": 2.7661068439483643,
"step": 896
},
{
"epoch": 0.054424242424242424,
"grad_norm": 0.11904677003622055,
"learning_rate": 0.0008,
"loss": 2.850513219833374,
"step": 898
},
{
"epoch": 0.05454545454545454,
"grad_norm": 0.10995450615882874,
"learning_rate": 0.0008,
"loss": 2.751762628555298,
"step": 900
},
{
"epoch": 0.05466666666666667,
"grad_norm": 0.10874732583761215,
"learning_rate": 0.0008,
"loss": 2.7956020832061768,
"step": 902
},
{
"epoch": 0.05478787878787879,
"grad_norm": 0.11032379418611526,
"learning_rate": 0.0008,
"loss": 2.733192205429077,
"step": 904
},
{
"epoch": 0.05490909090909091,
"grad_norm": 0.11254703998565674,
"learning_rate": 0.0008,
"loss": 2.825544834136963,
"step": 906
},
{
"epoch": 0.055030303030303034,
"grad_norm": 0.10856771469116211,
"learning_rate": 0.0008,
"loss": 2.713992118835449,
"step": 908
},
{
"epoch": 0.05515151515151515,
"grad_norm": 0.11766117811203003,
"learning_rate": 0.0008,
"loss": 2.860283136367798,
"step": 910
},
{
"epoch": 0.05527272727272727,
"grad_norm": 0.1102495864033699,
"learning_rate": 0.0008,
"loss": 2.7138748168945312,
"step": 912
},
{
"epoch": 0.05539393939393939,
"grad_norm": 0.10665928572416306,
"learning_rate": 0.0008,
"loss": 2.7042555809020996,
"step": 914
},
{
"epoch": 0.05551515151515152,
"grad_norm": 0.6552644371986389,
"learning_rate": 0.0008,
"loss": 2.7432925701141357,
"step": 916
},
{
"epoch": 0.05563636363636364,
"grad_norm": 0.26266244053840637,
"learning_rate": 0.0008,
"loss": 2.812037229537964,
"step": 918
},
{
"epoch": 0.055757575757575756,
"grad_norm": 0.1375817358493805,
"learning_rate": 0.0008,
"loss": 2.7334041595458984,
"step": 920
},
{
"epoch": 0.05587878787878788,
"grad_norm": 0.10863844305276871,
"learning_rate": 0.0008,
"loss": 2.7840628623962402,
"step": 922
},
{
"epoch": 0.056,
"grad_norm": 0.11129338294267654,
"learning_rate": 0.0008,
"loss": 2.78403902053833,
"step": 924
},
{
"epoch": 0.05612121212121212,
"grad_norm": 0.10913985222578049,
"learning_rate": 0.0008,
"loss": 2.7725250720977783,
"step": 926
},
{
"epoch": 0.05624242424242424,
"grad_norm": 0.11481859534978867,
"learning_rate": 0.0008,
"loss": 2.7732160091400146,
"step": 928
},
{
"epoch": 0.056363636363636366,
"grad_norm": 0.11153978109359741,
"learning_rate": 0.0008,
"loss": 2.7629857063293457,
"step": 930
},
{
"epoch": 0.056484848484848485,
"grad_norm": 0.10774504393339157,
"learning_rate": 0.0008,
"loss": 2.7736735343933105,
"step": 932
},
{
"epoch": 0.056606060606060604,
"grad_norm": 0.10816671699285507,
"learning_rate": 0.0008,
"loss": 2.7727668285369873,
"step": 934
},
{
"epoch": 0.05672727272727273,
"grad_norm": 0.10808931291103363,
"learning_rate": 0.0008,
"loss": 2.780731439590454,
"step": 936
},
{
"epoch": 0.05684848484848485,
"grad_norm": 0.12034662067890167,
"learning_rate": 0.0008,
"loss": 2.7461488246917725,
"step": 938
},
{
"epoch": 0.05696969696969697,
"grad_norm": 0.11467177420854568,
"learning_rate": 0.0008,
"loss": 2.774670362472534,
"step": 940
},
{
"epoch": 0.05709090909090909,
"grad_norm": 0.12002246826887131,
"learning_rate": 0.0008,
"loss": 2.724019765853882,
"step": 942
},
{
"epoch": 0.057212121212121214,
"grad_norm": 0.11797193437814713,
"learning_rate": 0.0008,
"loss": 2.784608840942383,
"step": 944
},
{
"epoch": 0.05733333333333333,
"grad_norm": 0.11059360951185226,
"learning_rate": 0.0008,
"loss": 2.7857565879821777,
"step": 946
},
{
"epoch": 0.05745454545454545,
"grad_norm": 0.10681190341711044,
"learning_rate": 0.0008,
"loss": 2.77789044380188,
"step": 948
},
{
"epoch": 0.05757575757575758,
"grad_norm": 0.10746893286705017,
"learning_rate": 0.0008,
"loss": 2.7458720207214355,
"step": 950
},
{
"epoch": 0.0576969696969697,
"grad_norm": 0.13561666011810303,
"learning_rate": 0.0008,
"loss": 2.750920534133911,
"step": 952
},
{
"epoch": 0.05781818181818182,
"grad_norm": 0.11216607689857483,
"learning_rate": 0.0008,
"loss": 2.805328607559204,
"step": 954
},
{
"epoch": 0.057939393939393936,
"grad_norm": 0.11170454323291779,
"learning_rate": 0.0008,
"loss": 2.7973318099975586,
"step": 956
},
{
"epoch": 0.05806060606060606,
"grad_norm": 0.11469848453998566,
"learning_rate": 0.0008,
"loss": 2.8135976791381836,
"step": 958
},
{
"epoch": 0.05818181818181818,
"grad_norm": 0.11365678161382675,
"learning_rate": 0.0008,
"loss": 2.8213155269622803,
"step": 960
},
{
"epoch": 0.0583030303030303,
"grad_norm": 0.11559037864208221,
"learning_rate": 0.0008,
"loss": 2.7919974327087402,
"step": 962
},
{
"epoch": 0.05842424242424243,
"grad_norm": 0.10419357568025589,
"learning_rate": 0.0008,
"loss": 2.727445363998413,
"step": 964
},
{
"epoch": 0.058545454545454546,
"grad_norm": 0.10752372443675995,
"learning_rate": 0.0008,
"loss": 2.8075525760650635,
"step": 966
},
{
"epoch": 0.058666666666666666,
"grad_norm": 0.10548102110624313,
"learning_rate": 0.0008,
"loss": 2.7213451862335205,
"step": 968
},
{
"epoch": 0.058787878787878785,
"grad_norm": 0.1062842532992363,
"learning_rate": 0.0008,
"loss": 2.8679680824279785,
"step": 970
},
{
"epoch": 0.05890909090909091,
"grad_norm": 0.10604004561901093,
"learning_rate": 0.0008,
"loss": 2.7563412189483643,
"step": 972
},
{
"epoch": 0.05903030303030303,
"grad_norm": 0.10803446918725967,
"learning_rate": 0.0008,
"loss": 2.7682676315307617,
"step": 974
},
{
"epoch": 0.05915151515151515,
"grad_norm": 0.11371214687824249,
"learning_rate": 0.0008,
"loss": 2.747363567352295,
"step": 976
},
{
"epoch": 0.059272727272727276,
"grad_norm": 0.1047954261302948,
"learning_rate": 0.0008,
"loss": 2.787336587905884,
"step": 978
},
{
"epoch": 0.059393939393939395,
"grad_norm": 0.11213313043117523,
"learning_rate": 0.0008,
"loss": 2.7720463275909424,
"step": 980
},
{
"epoch": 0.059515151515151514,
"grad_norm": 0.10050637274980545,
"learning_rate": 0.0008,
"loss": 2.657055616378784,
"step": 982
},
{
"epoch": 0.05963636363636363,
"grad_norm": 0.10387534648180008,
"learning_rate": 0.0008,
"loss": 2.7530951499938965,
"step": 984
},
{
"epoch": 0.05975757575757576,
"grad_norm": 0.10494332015514374,
"learning_rate": 0.0008,
"loss": 2.697255849838257,
"step": 986
},
{
"epoch": 0.05987878787878788,
"grad_norm": 0.10190849751234055,
"learning_rate": 0.0008,
"loss": 2.7321763038635254,
"step": 988
},
{
"epoch": 0.06,
"grad_norm": 0.11387556791305542,
"learning_rate": 0.0008,
"loss": 2.853233814239502,
"step": 990
},
{
"epoch": 0.060121212121212124,
"grad_norm": 0.10946540534496307,
"learning_rate": 0.0008,
"loss": 2.683826208114624,
"step": 992
},
{
"epoch": 0.06024242424242424,
"grad_norm": 0.113526351749897,
"learning_rate": 0.0008,
"loss": 2.8229663372039795,
"step": 994
},
{
"epoch": 0.06036363636363636,
"grad_norm": 0.1097358763217926,
"learning_rate": 0.0008,
"loss": 2.7877047061920166,
"step": 996
},
{
"epoch": 0.06048484848484848,
"grad_norm": 0.105439193546772,
"learning_rate": 0.0008,
"loss": 2.750767469406128,
"step": 998
},
{
"epoch": 0.06060606060606061,
"grad_norm": 0.10947102308273315,
"learning_rate": 0.0008,
"loss": 2.7278988361358643,
"step": 1000
},
{
"epoch": 0.06072727272727273,
"grad_norm": 0.10375527292490005,
"learning_rate": 0.0008,
"loss": 2.7932498455047607,
"step": 1002
},
{
"epoch": 0.060848484848484846,
"grad_norm": 0.12052513659000397,
"learning_rate": 0.0008,
"loss": 2.8309338092803955,
"step": 1004
},
{
"epoch": 0.06096969696969697,
"grad_norm": 0.10666930675506592,
"learning_rate": 0.0008,
"loss": 2.7480032444000244,
"step": 1006
},
{
"epoch": 0.06109090909090909,
"grad_norm": 0.10063078999519348,
"learning_rate": 0.0008,
"loss": 2.6663639545440674,
"step": 1008
},
{
"epoch": 0.06121212121212121,
"grad_norm": 0.10531777143478394,
"learning_rate": 0.0008,
"loss": 2.807374954223633,
"step": 1010
},
{
"epoch": 0.06133333333333333,
"grad_norm": 0.09874039888381958,
"learning_rate": 0.0008,
"loss": 2.7494113445281982,
"step": 1012
},
{
"epoch": 0.061454545454545456,
"grad_norm": 0.10312879085540771,
"learning_rate": 0.0008,
"loss": 2.7591135501861572,
"step": 1014
},
{
"epoch": 0.061575757575757575,
"grad_norm": 0.10460477322340012,
"learning_rate": 0.0008,
"loss": 2.7682900428771973,
"step": 1016
},
{
"epoch": 0.061696969696969695,
"grad_norm": 0.11767061799764633,
"learning_rate": 0.0008,
"loss": 2.782888412475586,
"step": 1018
},
{
"epoch": 0.06181818181818182,
"grad_norm": 0.10390406101942062,
"learning_rate": 0.0008,
"loss": 2.7906060218811035,
"step": 1020
},
{
"epoch": 0.06193939393939394,
"grad_norm": 0.11261998862028122,
"learning_rate": 0.0008,
"loss": 2.845881462097168,
"step": 1022
},
{
"epoch": 0.06206060606060606,
"grad_norm": 0.11000007390975952,
"learning_rate": 0.0008,
"loss": 2.7572999000549316,
"step": 1024
},
{
"epoch": 0.06218181818181818,
"grad_norm": 0.10593010485172272,
"learning_rate": 0.0008,
"loss": 2.7648873329162598,
"step": 1026
},
{
"epoch": 0.062303030303030305,
"grad_norm": 0.11216188222169876,
"learning_rate": 0.0008,
"loss": 2.7386484146118164,
"step": 1028
},
{
"epoch": 0.062424242424242424,
"grad_norm": 0.12322531640529633,
"learning_rate": 0.0008,
"loss": 2.771615982055664,
"step": 1030
},
{
"epoch": 0.06254545454545454,
"grad_norm": 0.12855897843837738,
"learning_rate": 0.0008,
"loss": 2.7112789154052734,
"step": 1032
},
{
"epoch": 0.06266666666666666,
"grad_norm": 0.11324036121368408,
"learning_rate": 0.0008,
"loss": 2.731501579284668,
"step": 1034
},
{
"epoch": 0.06278787878787878,
"grad_norm": 0.11115721613168716,
"learning_rate": 0.0008,
"loss": 2.721006393432617,
"step": 1036
},
{
"epoch": 0.06290909090909091,
"grad_norm": 0.10267746448516846,
"learning_rate": 0.0008,
"loss": 2.733232259750366,
"step": 1038
},
{
"epoch": 0.06303030303030303,
"grad_norm": 0.10309632122516632,
"learning_rate": 0.0008,
"loss": 2.756570339202881,
"step": 1040
},
{
"epoch": 0.06315151515151515,
"grad_norm": 0.10219801217317581,
"learning_rate": 0.0008,
"loss": 2.7849576473236084,
"step": 1042
},
{
"epoch": 0.06327272727272727,
"grad_norm": 0.10706236213445663,
"learning_rate": 0.0008,
"loss": 2.7808432579040527,
"step": 1044
},
{
"epoch": 0.06339393939393939,
"grad_norm": 0.10435187816619873,
"learning_rate": 0.0008,
"loss": 2.7838220596313477,
"step": 1046
},
{
"epoch": 0.06351515151515151,
"grad_norm": 0.1006159633398056,
"learning_rate": 0.0008,
"loss": 2.595548629760742,
"step": 1048
},
{
"epoch": 0.06363636363636363,
"grad_norm": 0.09872297197580338,
"learning_rate": 0.0008,
"loss": 2.8014883995056152,
"step": 1050
},
{
"epoch": 0.06375757575757576,
"grad_norm": 0.09960950911045074,
"learning_rate": 0.0008,
"loss": 2.7912018299102783,
"step": 1052
},
{
"epoch": 0.06387878787878788,
"grad_norm": 0.1094994768500328,
"learning_rate": 0.0008,
"loss": 2.8203585147857666,
"step": 1054
},
{
"epoch": 0.064,
"grad_norm": 0.1039128452539444,
"learning_rate": 0.0008,
"loss": 2.7462637424468994,
"step": 1056
},
{
"epoch": 0.06412121212121212,
"grad_norm": 0.09868817031383514,
"learning_rate": 0.0008,
"loss": 2.7532553672790527,
"step": 1058
},
{
"epoch": 0.06424242424242424,
"grad_norm": 0.10448651760816574,
"learning_rate": 0.0008,
"loss": 2.7721524238586426,
"step": 1060
},
{
"epoch": 0.06436363636363636,
"grad_norm": 0.1073535680770874,
"learning_rate": 0.0008,
"loss": 2.798076868057251,
"step": 1062
},
{
"epoch": 0.06448484848484848,
"grad_norm": 0.10052069276571274,
"learning_rate": 0.0008,
"loss": 2.808983325958252,
"step": 1064
},
{
"epoch": 0.06460606060606061,
"grad_norm": 0.10386918485164642,
"learning_rate": 0.0008,
"loss": 2.810534715652466,
"step": 1066
},
{
"epoch": 0.06472727272727273,
"grad_norm": 0.10510435700416565,
"learning_rate": 0.0008,
"loss": 2.6844887733459473,
"step": 1068
},
{
"epoch": 0.06484848484848485,
"grad_norm": 0.1032533124089241,
"learning_rate": 0.0008,
"loss": 2.750452756881714,
"step": 1070
},
{
"epoch": 0.06496969696969697,
"grad_norm": 0.10727686434984207,
"learning_rate": 0.0008,
"loss": 2.7373037338256836,
"step": 1072
},
{
"epoch": 0.06509090909090909,
"grad_norm": 0.1024731993675232,
"learning_rate": 0.0008,
"loss": 2.6866695880889893,
"step": 1074
},
{
"epoch": 0.06521212121212121,
"grad_norm": 0.0970090702176094,
"learning_rate": 0.0008,
"loss": 2.6102054119110107,
"step": 1076
},
{
"epoch": 0.06533333333333333,
"grad_norm": 0.11017253994941711,
"learning_rate": 0.0008,
"loss": 2.8165056705474854,
"step": 1078
},
{
"epoch": 0.06545454545454546,
"grad_norm": 0.12739963829517365,
"learning_rate": 0.0008,
"loss": 2.8251709938049316,
"step": 1080
},
{
"epoch": 0.06557575757575758,
"grad_norm": 0.11254332214593887,
"learning_rate": 0.0008,
"loss": 2.7868316173553467,
"step": 1082
},
{
"epoch": 0.0656969696969697,
"grad_norm": 0.1105445921421051,
"learning_rate": 0.0008,
"loss": 2.7541580200195312,
"step": 1084
},
{
"epoch": 0.06581818181818182,
"grad_norm": 0.11152500659227371,
"learning_rate": 0.0008,
"loss": 2.814708709716797,
"step": 1086
},
{
"epoch": 0.06593939393939394,
"grad_norm": 0.09904300421476364,
"learning_rate": 0.0008,
"loss": 2.6963448524475098,
"step": 1088
},
{
"epoch": 0.06606060606060606,
"grad_norm": 0.10311632603406906,
"learning_rate": 0.0008,
"loss": 2.665494203567505,
"step": 1090
},
{
"epoch": 0.06618181818181817,
"grad_norm": 0.10524038970470428,
"learning_rate": 0.0008,
"loss": 2.778984308242798,
"step": 1092
},
{
"epoch": 0.06630303030303031,
"grad_norm": 0.10193421691656113,
"learning_rate": 0.0008,
"loss": 2.81954026222229,
"step": 1094
},
{
"epoch": 0.06642424242424243,
"grad_norm": 0.10093263536691666,
"learning_rate": 0.0008,
"loss": 2.6708261966705322,
"step": 1096
},
{
"epoch": 0.06654545454545455,
"grad_norm": 0.10057204216718674,
"learning_rate": 0.0008,
"loss": 2.662590503692627,
"step": 1098
},
{
"epoch": 0.06666666666666667,
"grad_norm": 0.102562814950943,
"learning_rate": 0.0008,
"loss": 2.7364935874938965,
"step": 1100
},
{
"epoch": 0.06678787878787878,
"grad_norm": 0.09902184456586838,
"learning_rate": 0.0008,
"loss": 2.822427272796631,
"step": 1102
},
{
"epoch": 0.0669090909090909,
"grad_norm": 0.10648692399263382,
"learning_rate": 0.0008,
"loss": 2.7379634380340576,
"step": 1104
},
{
"epoch": 0.06703030303030302,
"grad_norm": 0.10598955303430557,
"learning_rate": 0.0008,
"loss": 2.763761043548584,
"step": 1106
},
{
"epoch": 0.06715151515151516,
"grad_norm": 0.09697546809911728,
"learning_rate": 0.0008,
"loss": 2.6694371700286865,
"step": 1108
},
{
"epoch": 0.06727272727272728,
"grad_norm": 0.1014561876654625,
"learning_rate": 0.0008,
"loss": 2.7454357147216797,
"step": 1110
},
{
"epoch": 0.0673939393939394,
"grad_norm": 0.10657907277345657,
"learning_rate": 0.0008,
"loss": 2.7411441802978516,
"step": 1112
},
{
"epoch": 0.06751515151515151,
"grad_norm": 0.10003755986690521,
"learning_rate": 0.0008,
"loss": 2.746644973754883,
"step": 1114
},
{
"epoch": 0.06763636363636363,
"grad_norm": 0.10184390842914581,
"learning_rate": 0.0008,
"loss": 2.684945583343506,
"step": 1116
},
{
"epoch": 0.06775757575757575,
"grad_norm": 0.10358394682407379,
"learning_rate": 0.0008,
"loss": 2.791767120361328,
"step": 1118
},
{
"epoch": 0.06787878787878789,
"grad_norm": 0.09545230865478516,
"learning_rate": 0.0008,
"loss": 2.761082410812378,
"step": 1120
},
{
"epoch": 0.068,
"grad_norm": 0.09734773635864258,
"learning_rate": 0.0008,
"loss": 2.7488605976104736,
"step": 1122
},
{
"epoch": 0.06812121212121212,
"grad_norm": 0.10791043192148209,
"learning_rate": 0.0008,
"loss": 2.761838436126709,
"step": 1124
},
{
"epoch": 0.06824242424242424,
"grad_norm": 0.10146962851285934,
"learning_rate": 0.0008,
"loss": 2.737246036529541,
"step": 1126
},
{
"epoch": 0.06836363636363636,
"grad_norm": 0.10075950622558594,
"learning_rate": 0.0008,
"loss": 2.676727294921875,
"step": 1128
},
{
"epoch": 0.06848484848484848,
"grad_norm": 0.10153098404407501,
"learning_rate": 0.0008,
"loss": 2.6826674938201904,
"step": 1130
},
{
"epoch": 0.0686060606060606,
"grad_norm": 0.10679205507040024,
"learning_rate": 0.0008,
"loss": 2.6952431201934814,
"step": 1132
},
{
"epoch": 0.06872727272727273,
"grad_norm": 0.1378297656774521,
"learning_rate": 0.0008,
"loss": 2.7400901317596436,
"step": 1134
},
{
"epoch": 0.06884848484848485,
"grad_norm": 0.1125171035528183,
"learning_rate": 0.0008,
"loss": 2.683004856109619,
"step": 1136
},
{
"epoch": 0.06896969696969697,
"grad_norm": 0.10048860311508179,
"learning_rate": 0.0008,
"loss": 2.7399070262908936,
"step": 1138
},
{
"epoch": 0.06909090909090909,
"grad_norm": 0.10249553620815277,
"learning_rate": 0.0008,
"loss": 2.7360503673553467,
"step": 1140
},
{
"epoch": 0.06921212121212121,
"grad_norm": 0.09944004565477371,
"learning_rate": 0.0008,
"loss": 2.7137906551361084,
"step": 1142
},
{
"epoch": 0.06933333333333333,
"grad_norm": 0.1002887636423111,
"learning_rate": 0.0008,
"loss": 2.7368063926696777,
"step": 1144
},
{
"epoch": 0.06945454545454545,
"grad_norm": 0.09983127564191818,
"learning_rate": 0.0008,
"loss": 2.758765459060669,
"step": 1146
},
{
"epoch": 0.06957575757575758,
"grad_norm": 0.10275828838348389,
"learning_rate": 0.0008,
"loss": 2.717484951019287,
"step": 1148
},
{
"epoch": 0.0696969696969697,
"grad_norm": 0.09778661280870438,
"learning_rate": 0.0008,
"loss": 2.7768945693969727,
"step": 1150
},
{
"epoch": 0.06981818181818182,
"grad_norm": 0.09667395055294037,
"learning_rate": 0.0008,
"loss": 2.818690061569214,
"step": 1152
},
{
"epoch": 0.06993939393939394,
"grad_norm": 0.09618115425109863,
"learning_rate": 0.0008,
"loss": 2.772036075592041,
"step": 1154
},
{
"epoch": 0.07006060606060606,
"grad_norm": 0.09304889291524887,
"learning_rate": 0.0008,
"loss": 2.700634717941284,
"step": 1156
},
{
"epoch": 0.07018181818181818,
"grad_norm": 0.1032819077372551,
"learning_rate": 0.0008,
"loss": 2.7707149982452393,
"step": 1158
},
{
"epoch": 0.0703030303030303,
"grad_norm": 0.09858979284763336,
"learning_rate": 0.0008,
"loss": 2.755944013595581,
"step": 1160
},
{
"epoch": 0.07042424242424243,
"grad_norm": 0.09689060598611832,
"learning_rate": 0.0008,
"loss": 2.727018356323242,
"step": 1162
},
{
"epoch": 0.07054545454545455,
"grad_norm": 0.09509676694869995,
"learning_rate": 0.0008,
"loss": 2.714245557785034,
"step": 1164
},
{
"epoch": 0.07066666666666667,
"grad_norm": 0.1079745814204216,
"learning_rate": 0.0008,
"loss": 2.725033760070801,
"step": 1166
},
{
"epoch": 0.07078787878787879,
"grad_norm": 0.09512803703546524,
"learning_rate": 0.0008,
"loss": 2.772056818008423,
"step": 1168
},
{
"epoch": 0.07090909090909091,
"grad_norm": 0.10424716770648956,
"learning_rate": 0.0008,
"loss": 2.695141315460205,
"step": 1170
},
{
"epoch": 0.07103030303030303,
"grad_norm": 0.09547784924507141,
"learning_rate": 0.0008,
"loss": 2.6702280044555664,
"step": 1172
},
{
"epoch": 0.07115151515151515,
"grad_norm": 0.09844043850898743,
"learning_rate": 0.0008,
"loss": 2.758188247680664,
"step": 1174
},
{
"epoch": 0.07127272727272728,
"grad_norm": 0.09592889249324799,
"learning_rate": 0.0008,
"loss": 2.7358851432800293,
"step": 1176
},
{
"epoch": 0.0713939393939394,
"grad_norm": 0.09553863853216171,
"learning_rate": 0.0008,
"loss": 2.77636456489563,
"step": 1178
},
{
"epoch": 0.07151515151515152,
"grad_norm": 0.09952032566070557,
"learning_rate": 0.0008,
"loss": 2.7566757202148438,
"step": 1180
},
{
"epoch": 0.07163636363636364,
"grad_norm": 0.10154401510953903,
"learning_rate": 0.0008,
"loss": 2.6886110305786133,
"step": 1182
},
{
"epoch": 0.07175757575757576,
"grad_norm": 0.0977076068520546,
"learning_rate": 0.0008,
"loss": 2.658785820007324,
"step": 1184
},
{
"epoch": 0.07187878787878788,
"grad_norm": 0.09749935567378998,
"learning_rate": 0.0008,
"loss": 2.722980499267578,
"step": 1186
},
{
"epoch": 0.072,
"grad_norm": 0.09280367195606232,
"learning_rate": 0.0008,
"loss": 2.715665340423584,
"step": 1188
},
{
"epoch": 0.07212121212121213,
"grad_norm": 0.0969923734664917,
"learning_rate": 0.0008,
"loss": 2.7989604473114014,
"step": 1190
},
{
"epoch": 0.07224242424242425,
"grad_norm": 0.09864166378974915,
"learning_rate": 0.0008,
"loss": 2.771336078643799,
"step": 1192
},
{
"epoch": 0.07236363636363637,
"grad_norm": 0.09835527837276459,
"learning_rate": 0.0008,
"loss": 2.734208822250366,
"step": 1194
},
{
"epoch": 0.07248484848484849,
"grad_norm": 0.10177188366651535,
"learning_rate": 0.0008,
"loss": 2.725253105163574,
"step": 1196
},
{
"epoch": 0.0726060606060606,
"grad_norm": 0.09399539977312088,
"learning_rate": 0.0008,
"loss": 2.717209815979004,
"step": 1198
},
{
"epoch": 0.07272727272727272,
"grad_norm": 0.09672411531209946,
"learning_rate": 0.0008,
"loss": 2.6600453853607178,
"step": 1200
},
{
"epoch": 0.07284848484848484,
"grad_norm": 0.09553155303001404,
"learning_rate": 0.0008,
"loss": 2.6727542877197266,
"step": 1202
},
{
"epoch": 0.07296969696969698,
"grad_norm": 0.09705706685781479,
"learning_rate": 0.0008,
"loss": 2.692979574203491,
"step": 1204
},
{
"epoch": 0.0730909090909091,
"grad_norm": 0.0983552485704422,
"learning_rate": 0.0008,
"loss": 2.714634418487549,
"step": 1206
},
{
"epoch": 0.07321212121212121,
"grad_norm": 0.10290054976940155,
"learning_rate": 0.0008,
"loss": 2.7607944011688232,
"step": 1208
},
{
"epoch": 0.07333333333333333,
"grad_norm": 0.09246007353067398,
"learning_rate": 0.0008,
"loss": 2.5990138053894043,
"step": 1210
},
{
"epoch": 0.07345454545454545,
"grad_norm": 0.09491098672151566,
"learning_rate": 0.0008,
"loss": 2.741530656814575,
"step": 1212
},
{
"epoch": 0.07357575757575757,
"grad_norm": 0.09645241498947144,
"learning_rate": 0.0008,
"loss": 2.6070706844329834,
"step": 1214
},
{
"epoch": 0.07369696969696969,
"grad_norm": 0.0955623984336853,
"learning_rate": 0.0008,
"loss": 2.7080843448638916,
"step": 1216
},
{
"epoch": 0.07381818181818182,
"grad_norm": 0.10209999978542328,
"learning_rate": 0.0008,
"loss": 2.7507176399230957,
"step": 1218
},
{
"epoch": 0.07393939393939394,
"grad_norm": 0.0957312360405922,
"learning_rate": 0.0008,
"loss": 2.6857995986938477,
"step": 1220
},
{
"epoch": 0.07406060606060606,
"grad_norm": 0.10950547456741333,
"learning_rate": 0.0008,
"loss": 2.7433340549468994,
"step": 1222
},
{
"epoch": 0.07418181818181818,
"grad_norm": 0.0924009308218956,
"learning_rate": 0.0008,
"loss": 2.7433056831359863,
"step": 1224
},
{
"epoch": 0.0743030303030303,
"grad_norm": 0.0966748371720314,
"learning_rate": 0.0008,
"loss": 2.800767421722412,
"step": 1226
},
{
"epoch": 0.07442424242424242,
"grad_norm": 0.09927260130643845,
"learning_rate": 0.0008,
"loss": 2.703301191329956,
"step": 1228
},
{
"epoch": 0.07454545454545454,
"grad_norm": 0.09447076916694641,
"learning_rate": 0.0008,
"loss": 2.7371115684509277,
"step": 1230
},
{
"epoch": 0.07466666666666667,
"grad_norm": 0.0925452709197998,
"learning_rate": 0.0008,
"loss": 2.6516075134277344,
"step": 1232
},
{
"epoch": 0.07478787878787879,
"grad_norm": 0.09471046179533005,
"learning_rate": 0.0008,
"loss": 2.7764124870300293,
"step": 1234
},
{
"epoch": 0.07490909090909091,
"grad_norm": 0.09640388935804367,
"learning_rate": 0.0008,
"loss": 2.763026237487793,
"step": 1236
},
{
"epoch": 0.07503030303030303,
"grad_norm": 0.09387095272541046,
"learning_rate": 0.0008,
"loss": 2.729520797729492,
"step": 1238
},
{
"epoch": 0.07515151515151515,
"grad_norm": 0.09486036747694016,
"learning_rate": 0.0008,
"loss": 2.6838998794555664,
"step": 1240
},
{
"epoch": 0.07527272727272727,
"grad_norm": 0.09296433627605438,
"learning_rate": 0.0008,
"loss": 2.7472357749938965,
"step": 1242
},
{
"epoch": 0.07539393939393939,
"grad_norm": 0.09180768579244614,
"learning_rate": 0.0008,
"loss": 2.6851282119750977,
"step": 1244
},
{
"epoch": 0.07551515151515152,
"grad_norm": 0.0947239026427269,
"learning_rate": 0.0008,
"loss": 2.71765398979187,
"step": 1246
},
{
"epoch": 0.07563636363636364,
"grad_norm": 0.09402729570865631,
"learning_rate": 0.0008,
"loss": 2.7172491550445557,
"step": 1248
},
{
"epoch": 0.07575757575757576,
"grad_norm": 0.09778286516666412,
"learning_rate": 0.0008,
"loss": 2.7468907833099365,
"step": 1250
},
{
"epoch": 0.07587878787878788,
"grad_norm": 0.09188830852508545,
"learning_rate": 0.0008,
"loss": 2.635127305984497,
"step": 1252
},
{
"epoch": 0.076,
"grad_norm": 0.1003621518611908,
"learning_rate": 0.0008,
"loss": 2.692934274673462,
"step": 1254
},
{
"epoch": 0.07612121212121212,
"grad_norm": 0.13150876760482788,
"learning_rate": 0.0008,
"loss": 2.6976070404052734,
"step": 1256
},
{
"epoch": 0.07624242424242424,
"grad_norm": 0.10479919612407684,
"learning_rate": 0.0008,
"loss": 2.673322916030884,
"step": 1258
},
{
"epoch": 0.07636363636363637,
"grad_norm": 0.09724564850330353,
"learning_rate": 0.0008,
"loss": 2.5951178073883057,
"step": 1260
},
{
"epoch": 0.07648484848484849,
"grad_norm": 0.11176671832799911,
"learning_rate": 0.0008,
"loss": 2.708684206008911,
"step": 1262
},
{
"epoch": 0.07660606060606061,
"grad_norm": 0.10475602000951767,
"learning_rate": 0.0008,
"loss": 2.6999783515930176,
"step": 1264
},
{
"epoch": 0.07672727272727273,
"grad_norm": 0.09650661051273346,
"learning_rate": 0.0008,
"loss": 2.6722989082336426,
"step": 1266
},
{
"epoch": 0.07684848484848485,
"grad_norm": 0.09830686450004578,
"learning_rate": 0.0008,
"loss": 2.6973743438720703,
"step": 1268
},
{
"epoch": 0.07696969696969697,
"grad_norm": 0.09264106303453445,
"learning_rate": 0.0008,
"loss": 2.6686978340148926,
"step": 1270
},
{
"epoch": 0.07709090909090908,
"grad_norm": 0.09338618069887161,
"learning_rate": 0.0008,
"loss": 2.6830546855926514,
"step": 1272
},
{
"epoch": 0.07721212121212122,
"grad_norm": 0.09089746326208115,
"learning_rate": 0.0008,
"loss": 2.6994199752807617,
"step": 1274
},
{
"epoch": 0.07733333333333334,
"grad_norm": 0.09034706652164459,
"learning_rate": 0.0008,
"loss": 2.603339910507202,
"step": 1276
},
{
"epoch": 0.07745454545454546,
"grad_norm": 0.09117346256971359,
"learning_rate": 0.0008,
"loss": 2.6654796600341797,
"step": 1278
},
{
"epoch": 0.07757575757575758,
"grad_norm": 0.09081052243709564,
"learning_rate": 0.0008,
"loss": 2.732980489730835,
"step": 1280
},
{
"epoch": 0.0776969696969697,
"grad_norm": 0.09462390094995499,
"learning_rate": 0.0008,
"loss": 2.6725363731384277,
"step": 1282
},
{
"epoch": 0.07781818181818181,
"grad_norm": 0.09132064133882523,
"learning_rate": 0.0008,
"loss": 2.6259028911590576,
"step": 1284
},
{
"epoch": 0.07793939393939393,
"grad_norm": 0.09186138957738876,
"learning_rate": 0.0008,
"loss": 2.6995110511779785,
"step": 1286
},
{
"epoch": 0.07806060606060607,
"grad_norm": 0.09084296226501465,
"learning_rate": 0.0008,
"loss": 2.748821496963501,
"step": 1288
},
{
"epoch": 0.07818181818181819,
"grad_norm": 0.10343652218580246,
"learning_rate": 0.0008,
"loss": 2.7680506706237793,
"step": 1290
},
{
"epoch": 0.0783030303030303,
"grad_norm": 0.09251704066991806,
"learning_rate": 0.0008,
"loss": 2.645200490951538,
"step": 1292
},
{
"epoch": 0.07842424242424242,
"grad_norm": 0.09454711526632309,
"learning_rate": 0.0008,
"loss": 2.744016408920288,
"step": 1294
},
{
"epoch": 0.07854545454545454,
"grad_norm": 0.0972449779510498,
"learning_rate": 0.0008,
"loss": 2.7548274993896484,
"step": 1296
},
{
"epoch": 0.07866666666666666,
"grad_norm": 0.1028820276260376,
"learning_rate": 0.0008,
"loss": 2.7428367137908936,
"step": 1298
},
{
"epoch": 0.07878787878787878,
"grad_norm": 0.09608453512191772,
"learning_rate": 0.0008,
"loss": 2.7066102027893066,
"step": 1300
},
{
"epoch": 0.07890909090909091,
"grad_norm": 0.08984900265932083,
"learning_rate": 0.0008,
"loss": 2.6286916732788086,
"step": 1302
},
{
"epoch": 0.07903030303030303,
"grad_norm": 0.09002672880887985,
"learning_rate": 0.0008,
"loss": 2.5521903038024902,
"step": 1304
},
{
"epoch": 0.07915151515151515,
"grad_norm": 0.09350468218326569,
"learning_rate": 0.0008,
"loss": 2.6434006690979004,
"step": 1306
},
{
"epoch": 0.07927272727272727,
"grad_norm": 0.09358403086662292,
"learning_rate": 0.0008,
"loss": 2.6999473571777344,
"step": 1308
},
{
"epoch": 0.07939393939393939,
"grad_norm": 0.08885160088539124,
"learning_rate": 0.0008,
"loss": 2.638434410095215,
"step": 1310
},
{
"epoch": 0.07951515151515151,
"grad_norm": 0.09976061433553696,
"learning_rate": 0.0008,
"loss": 2.71301531791687,
"step": 1312
},
{
"epoch": 0.07963636363636363,
"grad_norm": 0.09363351762294769,
"learning_rate": 0.0008,
"loss": 2.7035441398620605,
"step": 1314
},
{
"epoch": 0.07975757575757576,
"grad_norm": 0.09178290516138077,
"learning_rate": 0.0008,
"loss": 2.7259907722473145,
"step": 1316
},
{
"epoch": 0.07987878787878788,
"grad_norm": 0.0942964255809784,
"learning_rate": 0.0008,
"loss": 2.659278631210327,
"step": 1318
},
{
"epoch": 0.08,
"grad_norm": 0.09538327157497406,
"learning_rate": 0.0008,
"loss": 2.725430488586426,
"step": 1320
},
{
"epoch": 0.08012121212121212,
"grad_norm": 0.09187775105237961,
"learning_rate": 0.0008,
"loss": 2.6551382541656494,
"step": 1322
},
{
"epoch": 0.08024242424242424,
"grad_norm": 0.09640644490718842,
"learning_rate": 0.0008,
"loss": 2.671532392501831,
"step": 1324
},
{
"epoch": 0.08036363636363636,
"grad_norm": 0.09292606264352798,
"learning_rate": 0.0008,
"loss": 2.681725025177002,
"step": 1326
},
{
"epoch": 0.08048484848484848,
"grad_norm": 0.09245065599679947,
"learning_rate": 0.0008,
"loss": 2.7174232006073,
"step": 1328
},
{
"epoch": 0.08060606060606061,
"grad_norm": 0.09761374443769455,
"learning_rate": 0.0008,
"loss": 2.6278254985809326,
"step": 1330
},
{
"epoch": 0.08072727272727273,
"grad_norm": 0.09434378147125244,
"learning_rate": 0.0008,
"loss": 2.6975269317626953,
"step": 1332
},
{
"epoch": 0.08084848484848485,
"grad_norm": 0.09773518145084381,
"learning_rate": 0.0008,
"loss": 2.6537961959838867,
"step": 1334
},
{
"epoch": 0.08096969696969697,
"grad_norm": 0.08928754180669785,
"learning_rate": 0.0008,
"loss": 2.7122628688812256,
"step": 1336
},
{
"epoch": 0.08109090909090909,
"grad_norm": 0.09773512184619904,
"learning_rate": 0.0008,
"loss": 2.621272087097168,
"step": 1338
},
{
"epoch": 0.08121212121212121,
"grad_norm": 0.09332410246133804,
"learning_rate": 0.0008,
"loss": 2.6689252853393555,
"step": 1340
},
{
"epoch": 0.08133333333333333,
"grad_norm": 0.09455086290836334,
"learning_rate": 0.0008,
"loss": 2.7028133869171143,
"step": 1342
},
{
"epoch": 0.08145454545454546,
"grad_norm": 0.0893181562423706,
"learning_rate": 0.0008,
"loss": 2.643312454223633,
"step": 1344
},
{
"epoch": 0.08157575757575758,
"grad_norm": 0.0894450917840004,
"learning_rate": 0.0008,
"loss": 2.689328908920288,
"step": 1346
},
{
"epoch": 0.0816969696969697,
"grad_norm": 0.08961638063192368,
"learning_rate": 0.0008,
"loss": 2.7342095375061035,
"step": 1348
},
{
"epoch": 0.08181818181818182,
"grad_norm": 0.09310819953680038,
"learning_rate": 0.0008,
"loss": 2.716865301132202,
"step": 1350
},
{
"epoch": 0.08193939393939394,
"grad_norm": 0.09002184122800827,
"learning_rate": 0.0008,
"loss": 2.664811849594116,
"step": 1352
},
{
"epoch": 0.08206060606060606,
"grad_norm": 0.08868520706892014,
"learning_rate": 0.0008,
"loss": 2.7112624645233154,
"step": 1354
},
{
"epoch": 0.08218181818181818,
"grad_norm": 0.08859609067440033,
"learning_rate": 0.0008,
"loss": 2.684385061264038,
"step": 1356
},
{
"epoch": 0.08230303030303031,
"grad_norm": 0.09288623183965683,
"learning_rate": 0.0008,
"loss": 2.6730711460113525,
"step": 1358
},
{
"epoch": 0.08242424242424243,
"grad_norm": 0.09359101206064224,
"learning_rate": 0.0008,
"loss": 2.7263965606689453,
"step": 1360
},
{
"epoch": 0.08254545454545455,
"grad_norm": 0.0897098258137703,
"learning_rate": 0.0008,
"loss": 2.6738288402557373,
"step": 1362
},
{
"epoch": 0.08266666666666667,
"grad_norm": 0.09183207899332047,
"learning_rate": 0.0008,
"loss": 2.651707410812378,
"step": 1364
},
{
"epoch": 0.08278787878787879,
"grad_norm": 0.09773950278759003,
"learning_rate": 0.0008,
"loss": 2.6941521167755127,
"step": 1366
},
{
"epoch": 0.0829090909090909,
"grad_norm": 0.09789338707923889,
"learning_rate": 0.0008,
"loss": 2.6938390731811523,
"step": 1368
},
{
"epoch": 0.08303030303030302,
"grad_norm": 0.09005718678236008,
"learning_rate": 0.0008,
"loss": 2.59061861038208,
"step": 1370
},
{
"epoch": 0.08315151515151516,
"grad_norm": 0.09538372606039047,
"learning_rate": 0.0008,
"loss": 2.6102356910705566,
"step": 1372
},
{
"epoch": 0.08327272727272728,
"grad_norm": 0.08766049146652222,
"learning_rate": 0.0008,
"loss": 2.6359150409698486,
"step": 1374
},
{
"epoch": 0.0833939393939394,
"grad_norm": 0.09231335669755936,
"learning_rate": 0.0008,
"loss": 2.7464466094970703,
"step": 1376
},
{
"epoch": 0.08351515151515151,
"grad_norm": 0.0906107947230339,
"learning_rate": 0.0008,
"loss": 2.5935513973236084,
"step": 1378
},
{
"epoch": 0.08363636363636363,
"grad_norm": 0.09378247708082199,
"learning_rate": 0.0008,
"loss": 2.702707529067993,
"step": 1380
},
{
"epoch": 0.08375757575757575,
"grad_norm": 0.08846460282802582,
"learning_rate": 0.0008,
"loss": 2.62131929397583,
"step": 1382
},
{
"epoch": 0.08387878787878787,
"grad_norm": 0.09060888737440109,
"learning_rate": 0.0008,
"loss": 2.6346817016601562,
"step": 1384
},
{
"epoch": 0.084,
"grad_norm": 0.09113235026597977,
"learning_rate": 0.0008,
"loss": 2.517608165740967,
"step": 1386
},
{
"epoch": 0.08412121212121212,
"grad_norm": 0.09208323061466217,
"learning_rate": 0.0008,
"loss": 2.693209648132324,
"step": 1388
},
{
"epoch": 0.08424242424242424,
"grad_norm": 0.09389518201351166,
"learning_rate": 0.0008,
"loss": 2.739779233932495,
"step": 1390
},
{
"epoch": 0.08436363636363636,
"grad_norm": 0.08875050395727158,
"learning_rate": 0.0008,
"loss": 2.601445436477661,
"step": 1392
},
{
"epoch": 0.08448484848484848,
"grad_norm": 0.09733282774686813,
"learning_rate": 0.0008,
"loss": 2.7066140174865723,
"step": 1394
},
{
"epoch": 0.0846060606060606,
"grad_norm": 0.09224621951580048,
"learning_rate": 0.0008,
"loss": 2.5514330863952637,
"step": 1396
},
{
"epoch": 0.08472727272727272,
"grad_norm": 0.0903393104672432,
"learning_rate": 0.0008,
"loss": 2.621084451675415,
"step": 1398
},
{
"epoch": 0.08484848484848485,
"grad_norm": 0.08832167834043503,
"learning_rate": 0.0008,
"loss": 2.6870267391204834,
"step": 1400
},
{
"epoch": 0.08496969696969697,
"grad_norm": 0.09335021674633026,
"learning_rate": 0.0008,
"loss": 2.652280330657959,
"step": 1402
},
{
"epoch": 0.08509090909090909,
"grad_norm": 0.09261874109506607,
"learning_rate": 0.0008,
"loss": 2.6201632022857666,
"step": 1404
},
{
"epoch": 0.08521212121212121,
"grad_norm": 0.09099006652832031,
"learning_rate": 0.0008,
"loss": 2.700554132461548,
"step": 1406
},
{
"epoch": 0.08533333333333333,
"grad_norm": 0.08844577521085739,
"learning_rate": 0.0008,
"loss": 2.647566795349121,
"step": 1408
},
{
"epoch": 0.08545454545454545,
"grad_norm": 0.09372097998857498,
"learning_rate": 0.0008,
"loss": 2.6363234519958496,
"step": 1410
},
{
"epoch": 0.08557575757575757,
"grad_norm": 0.08882050961256027,
"learning_rate": 0.0008,
"loss": 2.513490915298462,
"step": 1412
},
{
"epoch": 0.0856969696969697,
"grad_norm": 0.10105268657207489,
"learning_rate": 0.0008,
"loss": 2.567296266555786,
"step": 1414
},
{
"epoch": 0.08581818181818182,
"grad_norm": 0.08813564479351044,
"learning_rate": 0.0008,
"loss": 2.655867576599121,
"step": 1416
},
{
"epoch": 0.08593939393939394,
"grad_norm": 0.09019377082586288,
"learning_rate": 0.0008,
"loss": 2.7091310024261475,
"step": 1418
},
{
"epoch": 0.08606060606060606,
"grad_norm": 0.08781956881284714,
"learning_rate": 0.0008,
"loss": 2.6156423091888428,
"step": 1420
},
{
"epoch": 0.08618181818181818,
"grad_norm": 0.08940879255533218,
"learning_rate": 0.0008,
"loss": 2.680206298828125,
"step": 1422
},
{
"epoch": 0.0863030303030303,
"grad_norm": 0.08795444667339325,
"learning_rate": 0.0008,
"loss": 2.6787171363830566,
"step": 1424
},
{
"epoch": 0.08642424242424242,
"grad_norm": 0.08696664124727249,
"learning_rate": 0.0008,
"loss": 2.683709144592285,
"step": 1426
},
{
"epoch": 0.08654545454545455,
"grad_norm": 0.08849150687456131,
"learning_rate": 0.0008,
"loss": 2.6615357398986816,
"step": 1428
},
{
"epoch": 0.08666666666666667,
"grad_norm": 0.09533119201660156,
"learning_rate": 0.0008,
"loss": 2.6473400592803955,
"step": 1430
},
{
"epoch": 0.08678787878787879,
"grad_norm": 0.09041043370962143,
"learning_rate": 0.0008,
"loss": 2.601652145385742,
"step": 1432
},
{
"epoch": 0.08690909090909091,
"grad_norm": 0.08922062069177628,
"learning_rate": 0.0008,
"loss": 2.66264271736145,
"step": 1434
},
{
"epoch": 0.08703030303030303,
"grad_norm": 0.08709689229726791,
"learning_rate": 0.0008,
"loss": 2.6181814670562744,
"step": 1436
},
{
"epoch": 0.08715151515151515,
"grad_norm": 0.09013613313436508,
"learning_rate": 0.0008,
"loss": 2.647557258605957,
"step": 1438
},
{
"epoch": 0.08727272727272728,
"grad_norm": 0.08934540301561356,
"learning_rate": 0.0008,
"loss": 2.638791084289551,
"step": 1440
},
{
"epoch": 0.0873939393939394,
"grad_norm": 0.08995569497346878,
"learning_rate": 0.0008,
"loss": 2.674405813217163,
"step": 1442
},
{
"epoch": 0.08751515151515152,
"grad_norm": 0.09702397882938385,
"learning_rate": 0.0008,
"loss": 2.6402435302734375,
"step": 1444
},
{
"epoch": 0.08763636363636364,
"grad_norm": 0.09387005120515823,
"learning_rate": 0.0008,
"loss": 2.62923264503479,
"step": 1446
},
{
"epoch": 0.08775757575757576,
"grad_norm": 0.0910274013876915,
"learning_rate": 0.0008,
"loss": 2.6891958713531494,
"step": 1448
},
{
"epoch": 0.08787878787878788,
"grad_norm": 0.08927851170301437,
"learning_rate": 0.0008,
"loss": 2.721039056777954,
"step": 1450
},
{
"epoch": 0.088,
"grad_norm": 0.16199354827404022,
"learning_rate": 0.0008,
"loss": 2.628119945526123,
"step": 1452
},
{
"epoch": 0.08812121212121213,
"grad_norm": 0.09073691815137863,
"learning_rate": 0.0008,
"loss": 2.6373634338378906,
"step": 1454
},
{
"epoch": 0.08824242424242425,
"grad_norm": 0.09171473234891891,
"learning_rate": 0.0008,
"loss": 2.6497883796691895,
"step": 1456
},
{
"epoch": 0.08836363636363637,
"grad_norm": 0.08553361892700195,
"learning_rate": 0.0008,
"loss": 2.5857951641082764,
"step": 1458
},
{
"epoch": 0.08848484848484849,
"grad_norm": 0.08835607022047043,
"learning_rate": 0.0008,
"loss": 2.6065711975097656,
"step": 1460
},
{
"epoch": 0.0886060606060606,
"grad_norm": 0.08698023855686188,
"learning_rate": 0.0008,
"loss": 2.6533703804016113,
"step": 1462
},
{
"epoch": 0.08872727272727272,
"grad_norm": 0.5185900330543518,
"learning_rate": 0.0008,
"loss": 2.6406586170196533,
"step": 1464
},
{
"epoch": 0.08884848484848484,
"grad_norm": 0.09255213290452957,
"learning_rate": 0.0008,
"loss": 2.6533427238464355,
"step": 1466
},
{
"epoch": 0.08896969696969698,
"grad_norm": 0.09007678925991058,
"learning_rate": 0.0008,
"loss": 2.5949273109436035,
"step": 1468
},
{
"epoch": 0.0890909090909091,
"grad_norm": 0.08804433047771454,
"learning_rate": 0.0008,
"loss": 2.6567821502685547,
"step": 1470
},
{
"epoch": 0.08921212121212121,
"grad_norm": 0.1966586858034134,
"learning_rate": 0.0008,
"loss": 2.6652119159698486,
"step": 1472
},
{
"epoch": 0.08933333333333333,
"grad_norm": 0.09040851891040802,
"learning_rate": 0.0008,
"loss": 2.6056416034698486,
"step": 1474
},
{
"epoch": 0.08945454545454545,
"grad_norm": 0.09084872156381607,
"learning_rate": 0.0008,
"loss": 2.632899522781372,
"step": 1476
},
{
"epoch": 0.08957575757575757,
"grad_norm": 0.09229514747858047,
"learning_rate": 0.0008,
"loss": 2.7004611492156982,
"step": 1478
},
{
"epoch": 0.08969696969696969,
"grad_norm": 0.08851287513971329,
"learning_rate": 0.0008,
"loss": 2.609126329421997,
"step": 1480
},
{
"epoch": 0.08981818181818182,
"grad_norm": 0.0903470441699028,
"learning_rate": 0.0008,
"loss": 2.656648635864258,
"step": 1482
},
{
"epoch": 0.08993939393939394,
"grad_norm": 0.08800260722637177,
"learning_rate": 0.0008,
"loss": 2.7243072986602783,
"step": 1484
},
{
"epoch": 0.09006060606060606,
"grad_norm": 0.10308069735765457,
"learning_rate": 0.0008,
"loss": 2.6953489780426025,
"step": 1486
},
{
"epoch": 0.09018181818181818,
"grad_norm": 0.09400498867034912,
"learning_rate": 0.0008,
"loss": 2.6483876705169678,
"step": 1488
},
{
"epoch": 0.0903030303030303,
"grad_norm": 0.08710706233978271,
"learning_rate": 0.0008,
"loss": 2.64103102684021,
"step": 1490
},
{
"epoch": 0.09042424242424242,
"grad_norm": 0.09035059064626694,
"learning_rate": 0.0008,
"loss": 2.5844297409057617,
"step": 1492
},
{
"epoch": 0.09054545454545454,
"grad_norm": 0.08683605492115021,
"learning_rate": 0.0008,
"loss": 2.5937013626098633,
"step": 1494
},
{
"epoch": 0.09066666666666667,
"grad_norm": 0.08672159165143967,
"learning_rate": 0.0008,
"loss": 2.6233937740325928,
"step": 1496
},
{
"epoch": 0.09078787878787879,
"grad_norm": 0.0888688713312149,
"learning_rate": 0.0008,
"loss": 2.623565673828125,
"step": 1498
},
{
"epoch": 0.09090909090909091,
"grad_norm": 0.08680512011051178,
"learning_rate": 0.0008,
"loss": 2.638805866241455,
"step": 1500
},
{
"epoch": 0.09103030303030303,
"grad_norm": 0.09417331218719482,
"learning_rate": 0.0008,
"loss": 2.6118662357330322,
"step": 1502
},
{
"epoch": 0.09115151515151515,
"grad_norm": 0.08677076548337936,
"learning_rate": 0.0008,
"loss": 2.6401901245117188,
"step": 1504
},
{
"epoch": 0.09127272727272727,
"grad_norm": 0.08785993605852127,
"learning_rate": 0.0008,
"loss": 2.6606388092041016,
"step": 1506
},
{
"epoch": 0.09139393939393939,
"grad_norm": 0.08566335588693619,
"learning_rate": 0.0008,
"loss": 2.684577226638794,
"step": 1508
},
{
"epoch": 0.09151515151515152,
"grad_norm": 0.08592430502176285,
"learning_rate": 0.0008,
"loss": 2.6088056564331055,
"step": 1510
},
{
"epoch": 0.09163636363636364,
"grad_norm": 0.09170215576887131,
"learning_rate": 0.0008,
"loss": 2.6677052974700928,
"step": 1512
},
{
"epoch": 0.09175757575757576,
"grad_norm": 0.09531085193157196,
"learning_rate": 0.0008,
"loss": 2.628242254257202,
"step": 1514
},
{
"epoch": 0.09187878787878788,
"grad_norm": 0.08863271772861481,
"learning_rate": 0.0008,
"loss": 2.618424892425537,
"step": 1516
},
{
"epoch": 0.092,
"grad_norm": 0.09483732283115387,
"learning_rate": 0.0008,
"loss": 2.590444564819336,
"step": 1518
},
{
"epoch": 0.09212121212121212,
"grad_norm": 0.08690401911735535,
"learning_rate": 0.0008,
"loss": 2.574446201324463,
"step": 1520
},
{
"epoch": 0.09224242424242424,
"grad_norm": 0.08537572622299194,
"learning_rate": 0.0008,
"loss": 2.6364688873291016,
"step": 1522
},
{
"epoch": 0.09236363636363637,
"grad_norm": 0.08771361410617828,
"learning_rate": 0.0008,
"loss": 2.7112021446228027,
"step": 1524
},
{
"epoch": 0.09248484848484849,
"grad_norm": 0.09044937789440155,
"learning_rate": 0.0008,
"loss": 2.6020820140838623,
"step": 1526
},
{
"epoch": 0.09260606060606061,
"grad_norm": 0.08857179433107376,
"learning_rate": 0.0008,
"loss": 2.6600921154022217,
"step": 1528
},
{
"epoch": 0.09272727272727273,
"grad_norm": 0.11604785174131393,
"learning_rate": 0.0008,
"loss": 2.657456159591675,
"step": 1530
},
{
"epoch": 0.09284848484848485,
"grad_norm": 0.09160027652978897,
"learning_rate": 0.0008,
"loss": 2.6848342418670654,
"step": 1532
},
{
"epoch": 0.09296969696969697,
"grad_norm": 0.08830365538597107,
"learning_rate": 0.0008,
"loss": 2.6130943298339844,
"step": 1534
},
{
"epoch": 0.09309090909090909,
"grad_norm": 0.08683188259601593,
"learning_rate": 0.0008,
"loss": 2.619514226913452,
"step": 1536
},
{
"epoch": 0.09321212121212122,
"grad_norm": 0.08896386623382568,
"learning_rate": 0.0008,
"loss": 2.650501012802124,
"step": 1538
},
{
"epoch": 0.09333333333333334,
"grad_norm": 0.08683948963880539,
"learning_rate": 0.0008,
"loss": 2.5995771884918213,
"step": 1540
},
{
"epoch": 0.09345454545454546,
"grad_norm": 0.08958131074905396,
"learning_rate": 0.0008,
"loss": 2.6178417205810547,
"step": 1542
},
{
"epoch": 0.09357575757575758,
"grad_norm": 0.08749283105134964,
"learning_rate": 0.0008,
"loss": 2.659029006958008,
"step": 1544
},
{
"epoch": 0.0936969696969697,
"grad_norm": 0.08909732848405838,
"learning_rate": 0.0008,
"loss": 2.6066806316375732,
"step": 1546
},
{
"epoch": 0.09381818181818181,
"grad_norm": 0.0832602009177208,
"learning_rate": 0.0008,
"loss": 2.556025505065918,
"step": 1548
},
{
"epoch": 0.09393939393939393,
"grad_norm": 0.08735526353120804,
"learning_rate": 0.0008,
"loss": 2.6787126064300537,
"step": 1550
},
{
"epoch": 0.09406060606060607,
"grad_norm": 0.08316223323345184,
"learning_rate": 0.0008,
"loss": 2.7239785194396973,
"step": 1552
},
{
"epoch": 0.09418181818181819,
"grad_norm": 0.08262155950069427,
"learning_rate": 0.0008,
"loss": 2.5391643047332764,
"step": 1554
},
{
"epoch": 0.0943030303030303,
"grad_norm": 0.0838422179222107,
"learning_rate": 0.0008,
"loss": 2.6395416259765625,
"step": 1556
},
{
"epoch": 0.09442424242424242,
"grad_norm": 0.08517379313707352,
"learning_rate": 0.0008,
"loss": 2.624040365219116,
"step": 1558
},
{
"epoch": 0.09454545454545454,
"grad_norm": 0.08498326689004898,
"learning_rate": 0.0008,
"loss": 2.5834991931915283,
"step": 1560
},
{
"epoch": 0.09466666666666666,
"grad_norm": 0.08669628947973251,
"learning_rate": 0.0008,
"loss": 2.558928966522217,
"step": 1562
},
{
"epoch": 0.09478787878787878,
"grad_norm": 0.08625341951847076,
"learning_rate": 0.0008,
"loss": 2.7375762462615967,
"step": 1564
},
{
"epoch": 0.09490909090909092,
"grad_norm": 0.08421202003955841,
"learning_rate": 0.0008,
"loss": 2.572462797164917,
"step": 1566
},
{
"epoch": 0.09503030303030303,
"grad_norm": 0.09357159584760666,
"learning_rate": 0.0008,
"loss": 2.655160903930664,
"step": 1568
},
{
"epoch": 0.09515151515151515,
"grad_norm": 0.08775287866592407,
"learning_rate": 0.0008,
"loss": 2.577094078063965,
"step": 1570
},
{
"epoch": 0.09527272727272727,
"grad_norm": 0.09956718236207962,
"learning_rate": 0.0008,
"loss": 2.6473801136016846,
"step": 1572
},
{
"epoch": 0.09539393939393939,
"grad_norm": 0.09372375905513763,
"learning_rate": 0.0008,
"loss": 2.579684257507324,
"step": 1574
},
{
"epoch": 0.09551515151515151,
"grad_norm": 0.0853939950466156,
"learning_rate": 0.0008,
"loss": 2.654546022415161,
"step": 1576
},
{
"epoch": 0.09563636363636363,
"grad_norm": 0.0899270623922348,
"learning_rate": 0.0008,
"loss": 2.6330487728118896,
"step": 1578
},
{
"epoch": 0.09575757575757576,
"grad_norm": 0.09261959046125412,
"learning_rate": 0.0008,
"loss": 2.6766517162323,
"step": 1580
},
{
"epoch": 0.09587878787878788,
"grad_norm": 0.08990292996168137,
"learning_rate": 0.0008,
"loss": 2.657620668411255,
"step": 1582
},
{
"epoch": 0.096,
"grad_norm": 0.09065950661897659,
"learning_rate": 0.0008,
"loss": 2.646404981613159,
"step": 1584
},
{
"epoch": 0.09612121212121212,
"grad_norm": 0.08696164190769196,
"learning_rate": 0.0008,
"loss": 2.593493700027466,
"step": 1586
},
{
"epoch": 0.09624242424242424,
"grad_norm": 0.08450400829315186,
"learning_rate": 0.0008,
"loss": 2.6234257221221924,
"step": 1588
},
{
"epoch": 0.09636363636363636,
"grad_norm": 0.08389894664287567,
"learning_rate": 0.0008,
"loss": 2.626136302947998,
"step": 1590
},
{
"epoch": 0.09648484848484848,
"grad_norm": 0.08361047506332397,
"learning_rate": 0.0008,
"loss": 2.637864112854004,
"step": 1592
},
{
"epoch": 0.09660606060606061,
"grad_norm": 0.08342014998197556,
"learning_rate": 0.0008,
"loss": 2.6756787300109863,
"step": 1594
},
{
"epoch": 0.09672727272727273,
"grad_norm": 0.08585110306739807,
"learning_rate": 0.0008,
"loss": 2.6727747917175293,
"step": 1596
},
{
"epoch": 0.09684848484848485,
"grad_norm": 0.0895339846611023,
"learning_rate": 0.0008,
"loss": 2.585358142852783,
"step": 1598
},
{
"epoch": 0.09696969696969697,
"grad_norm": 0.08387647569179535,
"learning_rate": 0.0008,
"loss": 2.5264551639556885,
"step": 1600
},
{
"epoch": 0.09709090909090909,
"grad_norm": 0.0897347554564476,
"learning_rate": 0.0008,
"loss": 2.6327881813049316,
"step": 1602
},
{
"epoch": 0.09721212121212121,
"grad_norm": 0.08462958037853241,
"learning_rate": 0.0008,
"loss": 2.660599708557129,
"step": 1604
},
{
"epoch": 0.09733333333333333,
"grad_norm": 0.09181305021047592,
"learning_rate": 0.0008,
"loss": 2.6233842372894287,
"step": 1606
},
{
"epoch": 0.09745454545454546,
"grad_norm": 0.08996450155973434,
"learning_rate": 0.0008,
"loss": 2.659290313720703,
"step": 1608
},
{
"epoch": 0.09757575757575758,
"grad_norm": 0.08530374616384506,
"learning_rate": 0.0008,
"loss": 2.6582229137420654,
"step": 1610
},
{
"epoch": 0.0976969696969697,
"grad_norm": 0.08404190838336945,
"learning_rate": 0.0008,
"loss": 2.7057888507843018,
"step": 1612
},
{
"epoch": 0.09781818181818182,
"grad_norm": 0.08575439453125,
"learning_rate": 0.0008,
"loss": 2.6792948246002197,
"step": 1614
},
{
"epoch": 0.09793939393939394,
"grad_norm": 0.08285801857709885,
"learning_rate": 0.0008,
"loss": 2.6680548191070557,
"step": 1616
},
{
"epoch": 0.09806060606060606,
"grad_norm": 0.08758828043937683,
"learning_rate": 0.0008,
"loss": 2.646259307861328,
"step": 1618
},
{
"epoch": 0.09818181818181818,
"grad_norm": 0.08405006676912308,
"learning_rate": 0.0008,
"loss": 2.641662120819092,
"step": 1620
},
{
"epoch": 0.09830303030303031,
"grad_norm": 0.08208192884922028,
"learning_rate": 0.0008,
"loss": 2.5521953105926514,
"step": 1622
},
{
"epoch": 0.09842424242424243,
"grad_norm": 0.0825958326458931,
"learning_rate": 0.0008,
"loss": 2.614128828048706,
"step": 1624
},
{
"epoch": 0.09854545454545455,
"grad_norm": 0.08548508584499359,
"learning_rate": 0.0008,
"loss": 2.6462769508361816,
"step": 1626
},
{
"epoch": 0.09866666666666667,
"grad_norm": 0.08114834129810333,
"learning_rate": 0.0008,
"loss": 2.5936594009399414,
"step": 1628
},
{
"epoch": 0.09878787878787879,
"grad_norm": 0.08322920650243759,
"learning_rate": 0.0008,
"loss": 2.6053903102874756,
"step": 1630
},
{
"epoch": 0.0989090909090909,
"grad_norm": 0.08171942830085754,
"learning_rate": 0.0008,
"loss": 2.547373056411743,
"step": 1632
},
{
"epoch": 0.09903030303030302,
"grad_norm": 0.0825696587562561,
"learning_rate": 0.0008,
"loss": 2.6250126361846924,
"step": 1634
},
{
"epoch": 0.09915151515151516,
"grad_norm": 0.08161260187625885,
"learning_rate": 0.0008,
"loss": 2.611213207244873,
"step": 1636
},
{
"epoch": 0.09927272727272728,
"grad_norm": 0.08342213183641434,
"learning_rate": 0.0008,
"loss": 2.690593957901001,
"step": 1638
},
{
"epoch": 0.0993939393939394,
"grad_norm": 0.08369901031255722,
"learning_rate": 0.0008,
"loss": 2.5864529609680176,
"step": 1640
},
{
"epoch": 0.09951515151515151,
"grad_norm": 0.08894728869199753,
"learning_rate": 0.0008,
"loss": 2.6114554405212402,
"step": 1642
},
{
"epoch": 0.09963636363636363,
"grad_norm": 0.0820501521229744,
"learning_rate": 0.0008,
"loss": 2.5911145210266113,
"step": 1644
},
{
"epoch": 0.09975757575757575,
"grad_norm": 0.08264602720737457,
"learning_rate": 0.0008,
"loss": 2.5683865547180176,
"step": 1646
},
{
"epoch": 0.09987878787878787,
"grad_norm": 0.0817006379365921,
"learning_rate": 0.0008,
"loss": 2.68829607963562,
"step": 1648
},
{
"epoch": 0.1,
"grad_norm": 0.08306078612804413,
"learning_rate": 0.0008,
"loss": 2.6774704456329346,
"step": 1650
},
{
"epoch": 0.10012121212121212,
"grad_norm": 0.08262654393911362,
"learning_rate": 0.0008,
"loss": 2.531188726425171,
"step": 1652
},
{
"epoch": 0.10024242424242424,
"grad_norm": 0.08557119220495224,
"learning_rate": 0.0008,
"loss": 2.64438796043396,
"step": 1654
},
{
"epoch": 0.10036363636363636,
"grad_norm": 0.08177687972784042,
"learning_rate": 0.0008,
"loss": 2.5884757041931152,
"step": 1656
},
{
"epoch": 0.10048484848484848,
"grad_norm": 0.08427724987268448,
"learning_rate": 0.0008,
"loss": 2.581458568572998,
"step": 1658
},
{
"epoch": 0.1006060606060606,
"grad_norm": 0.08376149833202362,
"learning_rate": 0.0008,
"loss": 2.5682568550109863,
"step": 1660
},
{
"epoch": 0.10072727272727272,
"grad_norm": 0.08558733016252518,
"learning_rate": 0.0008,
"loss": 2.6483490467071533,
"step": 1662
},
{
"epoch": 0.10084848484848485,
"grad_norm": 0.08896568417549133,
"learning_rate": 0.0008,
"loss": 2.6307573318481445,
"step": 1664
},
{
"epoch": 0.10096969696969697,
"grad_norm": 0.08166155219078064,
"learning_rate": 0.0008,
"loss": 2.603933572769165,
"step": 1666
},
{
"epoch": 0.10109090909090909,
"grad_norm": 0.09330889582633972,
"learning_rate": 0.0008,
"loss": 2.5625405311584473,
"step": 1668
},
{
"epoch": 0.10121212121212121,
"grad_norm": 0.0874500647187233,
"learning_rate": 0.0008,
"loss": 2.5738942623138428,
"step": 1670
},
{
"epoch": 0.10133333333333333,
"grad_norm": 0.0817391648888588,
"learning_rate": 0.0008,
"loss": 2.5482842922210693,
"step": 1672
},
{
"epoch": 0.10145454545454545,
"grad_norm": 0.07998669892549515,
"learning_rate": 0.0008,
"loss": 2.492748260498047,
"step": 1674
},
{
"epoch": 0.10157575757575757,
"grad_norm": 0.08261477947235107,
"learning_rate": 0.0008,
"loss": 2.5435993671417236,
"step": 1676
},
{
"epoch": 0.1016969696969697,
"grad_norm": 0.08120805770158768,
"learning_rate": 0.0008,
"loss": 2.561574935913086,
"step": 1678
},
{
"epoch": 0.10181818181818182,
"grad_norm": 0.08402089774608612,
"learning_rate": 0.0008,
"loss": 2.507296085357666,
"step": 1680
},
{
"epoch": 0.10193939393939394,
"grad_norm": 0.08169030398130417,
"learning_rate": 0.0008,
"loss": 2.6275389194488525,
"step": 1682
},
{
"epoch": 0.10206060606060606,
"grad_norm": 0.0843028873205185,
"learning_rate": 0.0008,
"loss": 2.622335433959961,
"step": 1684
},
{
"epoch": 0.10218181818181818,
"grad_norm": 0.0838347002863884,
"learning_rate": 0.0008,
"loss": 2.6002395153045654,
"step": 1686
},
{
"epoch": 0.1023030303030303,
"grad_norm": 0.08948864787817001,
"learning_rate": 0.0008,
"loss": 2.5590603351593018,
"step": 1688
},
{
"epoch": 0.10242424242424242,
"grad_norm": 0.08389901369810104,
"learning_rate": 0.0008,
"loss": 2.64823055267334,
"step": 1690
},
{
"epoch": 0.10254545454545455,
"grad_norm": 0.08830294758081436,
"learning_rate": 0.0008,
"loss": 2.6561338901519775,
"step": 1692
},
{
"epoch": 0.10266666666666667,
"grad_norm": 0.08801761269569397,
"learning_rate": 0.0008,
"loss": 2.612487316131592,
"step": 1694
},
{
"epoch": 0.10278787878787879,
"grad_norm": 0.085148885846138,
"learning_rate": 0.0008,
"loss": 2.5871212482452393,
"step": 1696
},
{
"epoch": 0.10290909090909091,
"grad_norm": 0.08251719176769257,
"learning_rate": 0.0008,
"loss": 2.6397407054901123,
"step": 1698
},
{
"epoch": 0.10303030303030303,
"grad_norm": 0.08379548788070679,
"learning_rate": 0.0008,
"loss": 2.513746976852417,
"step": 1700
},
{
"epoch": 0.10315151515151515,
"grad_norm": 0.08554631471633911,
"learning_rate": 0.0008,
"loss": 2.626783609390259,
"step": 1702
},
{
"epoch": 0.10327272727272727,
"grad_norm": 0.08245866000652313,
"learning_rate": 0.0008,
"loss": 2.5849609375,
"step": 1704
},
{
"epoch": 0.1033939393939394,
"grad_norm": 0.08258315920829773,
"learning_rate": 0.0008,
"loss": 2.5623438358306885,
"step": 1706
},
{
"epoch": 0.10351515151515152,
"grad_norm": 0.08043941855430603,
"learning_rate": 0.0008,
"loss": 2.572948932647705,
"step": 1708
},
{
"epoch": 0.10363636363636364,
"grad_norm": 0.08376195281744003,
"learning_rate": 0.0008,
"loss": 2.6075985431671143,
"step": 1710
},
{
"epoch": 0.10375757575757576,
"grad_norm": 0.0853591337800026,
"learning_rate": 0.0008,
"loss": 2.592024087905884,
"step": 1712
},
{
"epoch": 0.10387878787878788,
"grad_norm": 0.08038382977247238,
"learning_rate": 0.0008,
"loss": 2.582634449005127,
"step": 1714
},
{
"epoch": 0.104,
"grad_norm": 0.07938387244939804,
"learning_rate": 0.0008,
"loss": 2.5241196155548096,
"step": 1716
},
{
"epoch": 0.10412121212121211,
"grad_norm": 0.08292094618082047,
"learning_rate": 0.0008,
"loss": 2.517606019973755,
"step": 1718
},
{
"epoch": 0.10424242424242425,
"grad_norm": 0.07931133359670639,
"learning_rate": 0.0008,
"loss": 2.5932164192199707,
"step": 1720
},
{
"epoch": 0.10436363636363637,
"grad_norm": 0.08004464209079742,
"learning_rate": 0.0008,
"loss": 2.6461195945739746,
"step": 1722
},
{
"epoch": 0.10448484848484849,
"grad_norm": 0.08412160724401474,
"learning_rate": 0.0008,
"loss": 2.6172235012054443,
"step": 1724
},
{
"epoch": 0.1046060606060606,
"grad_norm": 0.08084485679864883,
"learning_rate": 0.0008,
"loss": 2.616210460662842,
"step": 1726
},
{
"epoch": 0.10472727272727272,
"grad_norm": 0.08084439486265182,
"learning_rate": 0.0008,
"loss": 2.6154465675354004,
"step": 1728
},
{
"epoch": 0.10484848484848484,
"grad_norm": 0.08507845550775528,
"learning_rate": 0.0008,
"loss": 2.5784924030303955,
"step": 1730
},
{
"epoch": 0.10496969696969696,
"grad_norm": 0.08067142963409424,
"learning_rate": 0.0008,
"loss": 2.548933267593384,
"step": 1732
},
{
"epoch": 0.1050909090909091,
"grad_norm": 0.08325568586587906,
"learning_rate": 0.0008,
"loss": 2.511080741882324,
"step": 1734
},
{
"epoch": 0.10521212121212122,
"grad_norm": 0.08696416765451431,
"learning_rate": 0.0008,
"loss": 2.606433629989624,
"step": 1736
},
{
"epoch": 0.10533333333333333,
"grad_norm": 0.07870412617921829,
"learning_rate": 0.0008,
"loss": 2.6113502979278564,
"step": 1738
},
{
"epoch": 0.10545454545454545,
"grad_norm": 0.08291155844926834,
"learning_rate": 0.0008,
"loss": 2.53584361076355,
"step": 1740
},
{
"epoch": 0.10557575757575757,
"grad_norm": 0.0846715196967125,
"learning_rate": 0.0008,
"loss": 2.5767483711242676,
"step": 1742
},
{
"epoch": 0.10569696969696969,
"grad_norm": 0.0805196464061737,
"learning_rate": 0.0008,
"loss": 2.550333023071289,
"step": 1744
},
{
"epoch": 0.10581818181818183,
"grad_norm": 0.08009176701307297,
"learning_rate": 0.0008,
"loss": 2.590287923812866,
"step": 1746
},
{
"epoch": 0.10593939393939394,
"grad_norm": 0.08378510177135468,
"learning_rate": 0.0008,
"loss": 2.6424410343170166,
"step": 1748
},
{
"epoch": 0.10606060606060606,
"grad_norm": 0.08473794907331467,
"learning_rate": 0.0008,
"loss": 2.575641632080078,
"step": 1750
},
{
"epoch": 0.10618181818181818,
"grad_norm": 0.08306790888309479,
"learning_rate": 0.0008,
"loss": 2.5710952281951904,
"step": 1752
},
{
"epoch": 0.1063030303030303,
"grad_norm": 0.08381582051515579,
"learning_rate": 0.0008,
"loss": 2.6002063751220703,
"step": 1754
},
{
"epoch": 0.10642424242424242,
"grad_norm": 0.07890482991933823,
"learning_rate": 0.0008,
"loss": 2.591548442840576,
"step": 1756
},
{
"epoch": 0.10654545454545454,
"grad_norm": 0.08090407401323318,
"learning_rate": 0.0008,
"loss": 2.5472769737243652,
"step": 1758
},
{
"epoch": 0.10666666666666667,
"grad_norm": 0.07940313220024109,
"learning_rate": 0.0008,
"loss": 2.66999888420105,
"step": 1760
},
{
"epoch": 0.10678787878787879,
"grad_norm": 0.08196254074573517,
"learning_rate": 0.0008,
"loss": 2.535832405090332,
"step": 1762
},
{
"epoch": 0.10690909090909091,
"grad_norm": 0.08165888488292694,
"learning_rate": 0.0008,
"loss": 2.607508659362793,
"step": 1764
},
{
"epoch": 0.10703030303030303,
"grad_norm": 0.0800384134054184,
"learning_rate": 0.0008,
"loss": 2.56892991065979,
"step": 1766
},
{
"epoch": 0.10715151515151515,
"grad_norm": 0.08316487818956375,
"learning_rate": 0.0008,
"loss": 2.5711958408355713,
"step": 1768
},
{
"epoch": 0.10727272727272727,
"grad_norm": 0.08547675609588623,
"learning_rate": 0.0008,
"loss": 2.560065746307373,
"step": 1770
},
{
"epoch": 0.10739393939393939,
"grad_norm": 0.08764155209064484,
"learning_rate": 0.0008,
"loss": 2.6489295959472656,
"step": 1772
},
{
"epoch": 0.10751515151515152,
"grad_norm": 0.07988082617521286,
"learning_rate": 0.0008,
"loss": 2.6154465675354004,
"step": 1774
},
{
"epoch": 0.10763636363636364,
"grad_norm": 0.08359135687351227,
"learning_rate": 0.0008,
"loss": 2.6705987453460693,
"step": 1776
},
{
"epoch": 0.10775757575757576,
"grad_norm": 0.08470160514116287,
"learning_rate": 0.0008,
"loss": 2.6560919284820557,
"step": 1778
},
{
"epoch": 0.10787878787878788,
"grad_norm": 0.08035678416490555,
"learning_rate": 0.0008,
"loss": 2.6173012256622314,
"step": 1780
},
{
"epoch": 0.108,
"grad_norm": 0.08844577521085739,
"learning_rate": 0.0008,
"loss": 2.5728743076324463,
"step": 1782
},
{
"epoch": 0.10812121212121212,
"grad_norm": 0.07951035350561142,
"learning_rate": 0.0008,
"loss": 2.532910108566284,
"step": 1784
},
{
"epoch": 0.10824242424242424,
"grad_norm": 0.08250928670167923,
"learning_rate": 0.0008,
"loss": 2.5740959644317627,
"step": 1786
},
{
"epoch": 0.10836363636363637,
"grad_norm": 0.08284786343574524,
"learning_rate": 0.0008,
"loss": 2.5687453746795654,
"step": 1788
},
{
"epoch": 0.10848484848484849,
"grad_norm": 0.08517413586378098,
"learning_rate": 0.0008,
"loss": 2.5776591300964355,
"step": 1790
},
{
"epoch": 0.10860606060606061,
"grad_norm": 0.07955542206764221,
"learning_rate": 0.0008,
"loss": 2.5979530811309814,
"step": 1792
},
{
"epoch": 0.10872727272727273,
"grad_norm": 0.08014333248138428,
"learning_rate": 0.0008,
"loss": 2.5912444591522217,
"step": 1794
},
{
"epoch": 0.10884848484848485,
"grad_norm": 0.08296824246644974,
"learning_rate": 0.0008,
"loss": 2.562481641769409,
"step": 1796
},
{
"epoch": 0.10896969696969697,
"grad_norm": 0.07798575609922409,
"learning_rate": 0.0008,
"loss": 2.5949549674987793,
"step": 1798
},
{
"epoch": 0.10909090909090909,
"grad_norm": 0.0781145766377449,
"learning_rate": 0.0008,
"loss": 2.573051929473877,
"step": 1800
},
{
"epoch": 0.10921212121212122,
"grad_norm": 0.07924607396125793,
"learning_rate": 0.0008,
"loss": 2.4930264949798584,
"step": 1802
},
{
"epoch": 0.10933333333333334,
"grad_norm": 0.08180536329746246,
"learning_rate": 0.0008,
"loss": 2.5294201374053955,
"step": 1804
},
{
"epoch": 0.10945454545454546,
"grad_norm": 0.08046746253967285,
"learning_rate": 0.0008,
"loss": 2.5376110076904297,
"step": 1806
},
{
"epoch": 0.10957575757575758,
"grad_norm": 0.08170323073863983,
"learning_rate": 0.0008,
"loss": 2.5614559650421143,
"step": 1808
},
{
"epoch": 0.1096969696969697,
"grad_norm": 0.0811152532696724,
"learning_rate": 0.0008,
"loss": 2.5910582542419434,
"step": 1810
},
{
"epoch": 0.10981818181818181,
"grad_norm": 0.08472200483083725,
"learning_rate": 0.0008,
"loss": 2.595961093902588,
"step": 1812
},
{
"epoch": 0.10993939393939393,
"grad_norm": 0.07880979031324387,
"learning_rate": 0.0008,
"loss": 2.5567238330841064,
"step": 1814
},
{
"epoch": 0.11006060606060607,
"grad_norm": 0.0827631726861,
"learning_rate": 0.0008,
"loss": 2.613417387008667,
"step": 1816
},
{
"epoch": 0.11018181818181819,
"grad_norm": 0.0842319130897522,
"learning_rate": 0.0008,
"loss": 2.5092010498046875,
"step": 1818
},
{
"epoch": 0.1103030303030303,
"grad_norm": 0.07937268167734146,
"learning_rate": 0.0008,
"loss": 2.5757827758789062,
"step": 1820
},
{
"epoch": 0.11042424242424242,
"grad_norm": 0.08157451450824738,
"learning_rate": 0.0008,
"loss": 2.556257724761963,
"step": 1822
},
{
"epoch": 0.11054545454545454,
"grad_norm": 0.08298274874687195,
"learning_rate": 0.0008,
"loss": 2.5384204387664795,
"step": 1824
},
{
"epoch": 0.11066666666666666,
"grad_norm": 0.08375734090805054,
"learning_rate": 0.0008,
"loss": 2.518423080444336,
"step": 1826
},
{
"epoch": 0.11078787878787878,
"grad_norm": 0.08177173882722855,
"learning_rate": 0.0008,
"loss": 2.550194025039673,
"step": 1828
},
{
"epoch": 0.11090909090909092,
"grad_norm": 0.07926391810178757,
"learning_rate": 0.0008,
"loss": 2.547438144683838,
"step": 1830
},
{
"epoch": 0.11103030303030303,
"grad_norm": 0.08357267081737518,
"learning_rate": 0.0008,
"loss": 2.5919384956359863,
"step": 1832
},
{
"epoch": 0.11115151515151515,
"grad_norm": 0.07746285200119019,
"learning_rate": 0.0008,
"loss": 2.5227155685424805,
"step": 1834
},
{
"epoch": 0.11127272727272727,
"grad_norm": 0.07889530062675476,
"learning_rate": 0.0008,
"loss": 2.565220355987549,
"step": 1836
},
{
"epoch": 0.11139393939393939,
"grad_norm": 0.0776098296046257,
"learning_rate": 0.0008,
"loss": 2.524049758911133,
"step": 1838
},
{
"epoch": 0.11151515151515151,
"grad_norm": 0.07988344132900238,
"learning_rate": 0.0008,
"loss": 2.596508502960205,
"step": 1840
},
{
"epoch": 0.11163636363636363,
"grad_norm": 0.0809350311756134,
"learning_rate": 0.0008,
"loss": 2.6687557697296143,
"step": 1842
},
{
"epoch": 0.11175757575757576,
"grad_norm": 0.08160511404275894,
"learning_rate": 0.0008,
"loss": 2.658667802810669,
"step": 1844
},
{
"epoch": 0.11187878787878788,
"grad_norm": 0.07964735478162766,
"learning_rate": 0.0008,
"loss": 2.6374194622039795,
"step": 1846
},
{
"epoch": 0.112,
"grad_norm": 0.08100466430187225,
"learning_rate": 0.0008,
"loss": 2.4983513355255127,
"step": 1848
},
{
"epoch": 0.11212121212121212,
"grad_norm": 0.0864834114909172,
"learning_rate": 0.0008,
"loss": 2.620969533920288,
"step": 1850
},
{
"epoch": 0.11224242424242424,
"grad_norm": 0.172709658741951,
"learning_rate": 0.0008,
"loss": 2.564512252807617,
"step": 1852
},
{
"epoch": 0.11236363636363636,
"grad_norm": 0.0803733766078949,
"learning_rate": 0.0008,
"loss": 2.5546727180480957,
"step": 1854
},
{
"epoch": 0.11248484848484848,
"grad_norm": 0.08536186069250107,
"learning_rate": 0.0008,
"loss": 2.5879104137420654,
"step": 1856
},
{
"epoch": 0.11260606060606061,
"grad_norm": 0.08234191685914993,
"learning_rate": 0.0008,
"loss": 2.623807907104492,
"step": 1858
},
{
"epoch": 0.11272727272727273,
"grad_norm": 0.0786423608660698,
"learning_rate": 0.0008,
"loss": 2.5565335750579834,
"step": 1860
},
{
"epoch": 0.11284848484848485,
"grad_norm": 0.0816061720252037,
"learning_rate": 0.0008,
"loss": 2.647752523422241,
"step": 1862
},
{
"epoch": 0.11296969696969697,
"grad_norm": 0.07833302021026611,
"learning_rate": 0.0008,
"loss": 2.526172637939453,
"step": 1864
},
{
"epoch": 0.11309090909090909,
"grad_norm": 0.07706572860479355,
"learning_rate": 0.0008,
"loss": 2.595428466796875,
"step": 1866
},
{
"epoch": 0.11321212121212121,
"grad_norm": 0.0815252810716629,
"learning_rate": 0.0008,
"loss": 2.5555660724639893,
"step": 1868
},
{
"epoch": 0.11333333333333333,
"grad_norm": 0.07931073009967804,
"learning_rate": 0.0008,
"loss": 2.5754661560058594,
"step": 1870
},
{
"epoch": 0.11345454545454546,
"grad_norm": 0.07832778245210648,
"learning_rate": 0.0008,
"loss": 2.473487377166748,
"step": 1872
},
{
"epoch": 0.11357575757575758,
"grad_norm": 0.08451631665229797,
"learning_rate": 0.0008,
"loss": 2.46903920173645,
"step": 1874
},
{
"epoch": 0.1136969696969697,
"grad_norm": 0.08395351469516754,
"learning_rate": 0.0008,
"loss": 2.5439538955688477,
"step": 1876
},
{
"epoch": 0.11381818181818182,
"grad_norm": 0.08077947050333023,
"learning_rate": 0.0008,
"loss": 2.483140230178833,
"step": 1878
},
{
"epoch": 0.11393939393939394,
"grad_norm": 0.08051062375307083,
"learning_rate": 0.0008,
"loss": 2.5402326583862305,
"step": 1880
},
{
"epoch": 0.11406060606060606,
"grad_norm": 0.08310651034116745,
"learning_rate": 0.0008,
"loss": 2.5548527240753174,
"step": 1882
},
{
"epoch": 0.11418181818181818,
"grad_norm": 0.07922256737947464,
"learning_rate": 0.0008,
"loss": 2.5128228664398193,
"step": 1884
},
{
"epoch": 0.11430303030303031,
"grad_norm": 0.08044472336769104,
"learning_rate": 0.0008,
"loss": 2.5191962718963623,
"step": 1886
},
{
"epoch": 0.11442424242424243,
"grad_norm": 0.07645294070243835,
"learning_rate": 0.0008,
"loss": 2.5367119312286377,
"step": 1888
},
{
"epoch": 0.11454545454545455,
"grad_norm": 0.08101294189691544,
"learning_rate": 0.0008,
"loss": 2.601874589920044,
"step": 1890
},
{
"epoch": 0.11466666666666667,
"grad_norm": 0.07629749178886414,
"learning_rate": 0.0008,
"loss": 2.5014162063598633,
"step": 1892
},
{
"epoch": 0.11478787878787879,
"grad_norm": 0.07591025531291962,
"learning_rate": 0.0008,
"loss": 2.5347931385040283,
"step": 1894
},
{
"epoch": 0.1149090909090909,
"grad_norm": 0.07559724897146225,
"learning_rate": 0.0008,
"loss": 2.4728469848632812,
"step": 1896
},
{
"epoch": 0.11503030303030302,
"grad_norm": 0.07832006365060806,
"learning_rate": 0.0008,
"loss": 2.48507022857666,
"step": 1898
},
{
"epoch": 0.11515151515151516,
"grad_norm": 0.07801617681980133,
"learning_rate": 0.0008,
"loss": 2.557096242904663,
"step": 1900
},
{
"epoch": 0.11527272727272728,
"grad_norm": 0.10303416103124619,
"learning_rate": 0.0008,
"loss": 2.5679283142089844,
"step": 1902
},
{
"epoch": 0.1153939393939394,
"grad_norm": 0.08007843047380447,
"learning_rate": 0.0008,
"loss": 2.5529980659484863,
"step": 1904
},
{
"epoch": 0.11551515151515152,
"grad_norm": 0.07723647356033325,
"learning_rate": 0.0008,
"loss": 2.4967892169952393,
"step": 1906
},
{
"epoch": 0.11563636363636363,
"grad_norm": 0.07617611438035965,
"learning_rate": 0.0008,
"loss": 2.5565855503082275,
"step": 1908
},
{
"epoch": 0.11575757575757575,
"grad_norm": 0.08152547478675842,
"learning_rate": 0.0008,
"loss": 2.485872745513916,
"step": 1910
},
{
"epoch": 0.11587878787878787,
"grad_norm": 0.07907970994710922,
"learning_rate": 0.0008,
"loss": 2.615044355392456,
"step": 1912
},
{
"epoch": 0.116,
"grad_norm": 0.07603520154953003,
"learning_rate": 0.0008,
"loss": 2.5474045276641846,
"step": 1914
},
{
"epoch": 0.11612121212121213,
"grad_norm": 0.07645908743143082,
"learning_rate": 0.0008,
"loss": 2.5839340686798096,
"step": 1916
},
{
"epoch": 0.11624242424242424,
"grad_norm": 0.08038519322872162,
"learning_rate": 0.0008,
"loss": 2.6079697608947754,
"step": 1918
},
{
"epoch": 0.11636363636363636,
"grad_norm": 0.08025974780321121,
"learning_rate": 0.0008,
"loss": 2.582197904586792,
"step": 1920
},
{
"epoch": 0.11648484848484848,
"grad_norm": 0.0802643746137619,
"learning_rate": 0.0008,
"loss": 2.6118481159210205,
"step": 1922
},
{
"epoch": 0.1166060606060606,
"grad_norm": 0.0786847323179245,
"learning_rate": 0.0008,
"loss": 2.5025649070739746,
"step": 1924
},
{
"epoch": 0.11672727272727272,
"grad_norm": 0.07931985706090927,
"learning_rate": 0.0008,
"loss": 2.5947787761688232,
"step": 1926
},
{
"epoch": 0.11684848484848485,
"grad_norm": 0.07659479975700378,
"learning_rate": 0.0008,
"loss": 2.505234479904175,
"step": 1928
},
{
"epoch": 0.11696969696969697,
"grad_norm": 0.07724818587303162,
"learning_rate": 0.0008,
"loss": 2.513045072555542,
"step": 1930
},
{
"epoch": 0.11709090909090909,
"grad_norm": 0.07933937013149261,
"learning_rate": 0.0008,
"loss": 2.5808284282684326,
"step": 1932
},
{
"epoch": 0.11721212121212121,
"grad_norm": 0.0800996869802475,
"learning_rate": 0.0008,
"loss": 2.535442590713501,
"step": 1934
},
{
"epoch": 0.11733333333333333,
"grad_norm": 0.07885627448558807,
"learning_rate": 0.0008,
"loss": 2.5674121379852295,
"step": 1936
},
{
"epoch": 0.11745454545454545,
"grad_norm": 0.07888897508382797,
"learning_rate": 0.0008,
"loss": 2.572334051132202,
"step": 1938
},
{
"epoch": 0.11757575757575757,
"grad_norm": 0.07820527255535126,
"learning_rate": 0.0008,
"loss": 2.5956239700317383,
"step": 1940
},
{
"epoch": 0.1176969696969697,
"grad_norm": 0.07990499585866928,
"learning_rate": 0.0008,
"loss": 2.5474917888641357,
"step": 1942
},
{
"epoch": 0.11781818181818182,
"grad_norm": 0.08092857897281647,
"learning_rate": 0.0008,
"loss": 2.5225582122802734,
"step": 1944
},
{
"epoch": 0.11793939393939394,
"grad_norm": 0.07903316617012024,
"learning_rate": 0.0008,
"loss": 2.5358633995056152,
"step": 1946
},
{
"epoch": 0.11806060606060606,
"grad_norm": 0.08067017793655396,
"learning_rate": 0.0008,
"loss": 2.549748659133911,
"step": 1948
},
{
"epoch": 0.11818181818181818,
"grad_norm": 0.07513710856437683,
"learning_rate": 0.0008,
"loss": 2.464491367340088,
"step": 1950
},
{
"epoch": 0.1183030303030303,
"grad_norm": 0.07817662507295609,
"learning_rate": 0.0008,
"loss": 2.5247559547424316,
"step": 1952
},
{
"epoch": 0.11842424242424242,
"grad_norm": 0.07467043399810791,
"learning_rate": 0.0008,
"loss": 2.5231027603149414,
"step": 1954
},
{
"epoch": 0.11854545454545455,
"grad_norm": 0.07668938487768173,
"learning_rate": 0.0008,
"loss": 2.5834314823150635,
"step": 1956
},
{
"epoch": 0.11866666666666667,
"grad_norm": 0.07852034270763397,
"learning_rate": 0.0008,
"loss": 2.5520050525665283,
"step": 1958
},
{
"epoch": 0.11878787878787879,
"grad_norm": 0.07859796285629272,
"learning_rate": 0.0008,
"loss": 2.6046338081359863,
"step": 1960
},
{
"epoch": 0.11890909090909091,
"grad_norm": 0.07577870786190033,
"learning_rate": 0.0008,
"loss": 2.5782530307769775,
"step": 1962
},
{
"epoch": 0.11903030303030303,
"grad_norm": 0.08107993751764297,
"learning_rate": 0.0008,
"loss": 2.5706799030303955,
"step": 1964
},
{
"epoch": 0.11915151515151515,
"grad_norm": 0.08394039422273636,
"learning_rate": 0.0008,
"loss": 2.5240509510040283,
"step": 1966
},
{
"epoch": 0.11927272727272727,
"grad_norm": 0.07735666632652283,
"learning_rate": 0.0008,
"loss": 2.4988553524017334,
"step": 1968
},
{
"epoch": 0.1193939393939394,
"grad_norm": 0.07669788599014282,
"learning_rate": 0.0008,
"loss": 2.502908945083618,
"step": 1970
},
{
"epoch": 0.11951515151515152,
"grad_norm": 0.07579613476991653,
"learning_rate": 0.0008,
"loss": 2.496429920196533,
"step": 1972
},
{
"epoch": 0.11963636363636364,
"grad_norm": 0.07455839961767197,
"learning_rate": 0.0008,
"loss": 2.508849859237671,
"step": 1974
},
{
"epoch": 0.11975757575757576,
"grad_norm": 0.07721120864152908,
"learning_rate": 0.0008,
"loss": 2.499626398086548,
"step": 1976
},
{
"epoch": 0.11987878787878788,
"grad_norm": 0.07626793533563614,
"learning_rate": 0.0008,
"loss": 2.4832842350006104,
"step": 1978
},
{
"epoch": 0.12,
"grad_norm": 0.07914632558822632,
"learning_rate": 0.0008,
"loss": 2.574998617172241,
"step": 1980
},
{
"epoch": 0.12012121212121211,
"grad_norm": 0.0800415650010109,
"learning_rate": 0.0008,
"loss": 2.4994728565216064,
"step": 1982
},
{
"epoch": 0.12024242424242425,
"grad_norm": 0.07545595616102219,
"learning_rate": 0.0008,
"loss": 2.58841609954834,
"step": 1984
},
{
"epoch": 0.12036363636363637,
"grad_norm": 0.07582149654626846,
"learning_rate": 0.0008,
"loss": 2.5361456871032715,
"step": 1986
},
{
"epoch": 0.12048484848484849,
"grad_norm": 0.0746186152100563,
"learning_rate": 0.0008,
"loss": 2.600322961807251,
"step": 1988
},
{
"epoch": 0.1206060606060606,
"grad_norm": 0.07667519152164459,
"learning_rate": 0.0008,
"loss": 2.574183702468872,
"step": 1990
},
{
"epoch": 0.12072727272727272,
"grad_norm": 0.07411106675863266,
"learning_rate": 0.0008,
"loss": 2.4657859802246094,
"step": 1992
},
{
"epoch": 0.12084848484848484,
"grad_norm": 0.07614558190107346,
"learning_rate": 0.0008,
"loss": 2.563608169555664,
"step": 1994
},
{
"epoch": 0.12096969696969696,
"grad_norm": 0.07527037709951401,
"learning_rate": 0.0008,
"loss": 2.492877244949341,
"step": 1996
},
{
"epoch": 0.1210909090909091,
"grad_norm": 0.07498473674058914,
"learning_rate": 0.0008,
"loss": 2.577207088470459,
"step": 1998
},
{
"epoch": 0.12121212121212122,
"grad_norm": 0.08452074229717255,
"learning_rate": 0.0008,
"loss": 2.551314115524292,
"step": 2000
},
{
"epoch": 0.12133333333333333,
"grad_norm": 0.07714702188968658,
"learning_rate": 0.0008,
"loss": 2.480290651321411,
"step": 2002
},
{
"epoch": 0.12145454545454545,
"grad_norm": 0.0812622532248497,
"learning_rate": 0.0008,
"loss": 2.4882781505584717,
"step": 2004
},
{
"epoch": 0.12157575757575757,
"grad_norm": 0.1028304472565651,
"learning_rate": 0.0008,
"loss": 2.566134452819824,
"step": 2006
},
{
"epoch": 0.12169696969696969,
"grad_norm": 0.08645875006914139,
"learning_rate": 0.0008,
"loss": 2.4919838905334473,
"step": 2008
},
{
"epoch": 0.12181818181818181,
"grad_norm": 0.08240930736064911,
"learning_rate": 0.0008,
"loss": 2.5817999839782715,
"step": 2010
},
{
"epoch": 0.12193939393939394,
"grad_norm": 0.07395318895578384,
"learning_rate": 0.0008,
"loss": 2.5020995140075684,
"step": 2012
},
{
"epoch": 0.12206060606060606,
"grad_norm": 0.07520240545272827,
"learning_rate": 0.0008,
"loss": 2.523721694946289,
"step": 2014
},
{
"epoch": 0.12218181818181818,
"grad_norm": 0.07866991311311722,
"learning_rate": 0.0008,
"loss": 2.539628505706787,
"step": 2016
},
{
"epoch": 0.1223030303030303,
"grad_norm": 0.07618068158626556,
"learning_rate": 0.0008,
"loss": 2.5028769969940186,
"step": 2018
},
{
"epoch": 0.12242424242424242,
"grad_norm": 0.07841619104146957,
"learning_rate": 0.0008,
"loss": 2.559375762939453,
"step": 2020
},
{
"epoch": 0.12254545454545454,
"grad_norm": 0.08001592010259628,
"learning_rate": 0.0008,
"loss": 2.5575759410858154,
"step": 2022
},
{
"epoch": 0.12266666666666666,
"grad_norm": 0.07869841158390045,
"learning_rate": 0.0008,
"loss": 2.564208507537842,
"step": 2024
},
{
"epoch": 0.1227878787878788,
"grad_norm": 0.07790623605251312,
"learning_rate": 0.0008,
"loss": 2.6039628982543945,
"step": 2026
},
{
"epoch": 0.12290909090909091,
"grad_norm": 0.07511240988969803,
"learning_rate": 0.0008,
"loss": 2.595947742462158,
"step": 2028
},
{
"epoch": 0.12303030303030303,
"grad_norm": 0.0758950486779213,
"learning_rate": 0.0008,
"loss": 2.509847402572632,
"step": 2030
},
{
"epoch": 0.12315151515151515,
"grad_norm": 0.07532898336648941,
"learning_rate": 0.0008,
"loss": 2.5834481716156006,
"step": 2032
},
{
"epoch": 0.12327272727272727,
"grad_norm": 0.07525201886892319,
"learning_rate": 0.0008,
"loss": 2.4651682376861572,
"step": 2034
},
{
"epoch": 0.12339393939393939,
"grad_norm": 0.07575929164886475,
"learning_rate": 0.0008,
"loss": 2.578941822052002,
"step": 2036
},
{
"epoch": 0.12351515151515151,
"grad_norm": 0.07496031373739243,
"learning_rate": 0.0008,
"loss": 2.525634765625,
"step": 2038
},
{
"epoch": 0.12363636363636364,
"grad_norm": 0.07435885816812515,
"learning_rate": 0.0008,
"loss": 2.5471880435943604,
"step": 2040
},
{
"epoch": 0.12375757575757576,
"grad_norm": 0.0768258199095726,
"learning_rate": 0.0008,
"loss": 2.4974169731140137,
"step": 2042
},
{
"epoch": 0.12387878787878788,
"grad_norm": 0.07432639598846436,
"learning_rate": 0.0008,
"loss": 2.5387985706329346,
"step": 2044
},
{
"epoch": 0.124,
"grad_norm": 0.07835559546947479,
"learning_rate": 0.0008,
"loss": 2.5183753967285156,
"step": 2046
},
{
"epoch": 0.12412121212121212,
"grad_norm": 0.07599394023418427,
"learning_rate": 0.0008,
"loss": 2.5436673164367676,
"step": 2048
},
{
"epoch": 0.12424242424242424,
"grad_norm": 0.0745822861790657,
"learning_rate": 0.0008,
"loss": 2.4610037803649902,
"step": 2050
},
{
"epoch": 0.12436363636363636,
"grad_norm": 0.07971092313528061,
"learning_rate": 0.0008,
"loss": 2.602762222290039,
"step": 2052
},
{
"epoch": 0.12448484848484849,
"grad_norm": 0.07658050954341888,
"learning_rate": 0.0008,
"loss": 2.5141398906707764,
"step": 2054
},
{
"epoch": 0.12460606060606061,
"grad_norm": 0.09096317738294601,
"learning_rate": 0.0008,
"loss": 2.539989471435547,
"step": 2056
},
{
"epoch": 0.12472727272727273,
"grad_norm": 0.07740216702222824,
"learning_rate": 0.0008,
"loss": 2.479302167892456,
"step": 2058
},
{
"epoch": 0.12484848484848485,
"grad_norm": 0.07684116810560226,
"learning_rate": 0.0008,
"loss": 2.535499095916748,
"step": 2060
},
{
"epoch": 0.12496969696969697,
"grad_norm": 0.07898880541324615,
"learning_rate": 0.0008,
"loss": 2.5535802841186523,
"step": 2062
},
{
"epoch": 0.12509090909090909,
"grad_norm": 0.0799906849861145,
"learning_rate": 0.0008,
"loss": 2.4934959411621094,
"step": 2064
},
{
"epoch": 0.12521212121212122,
"grad_norm": 0.07639925181865692,
"learning_rate": 0.0008,
"loss": 2.4716477394104004,
"step": 2066
},
{
"epoch": 0.12533333333333332,
"grad_norm": 0.07716447114944458,
"learning_rate": 0.0008,
"loss": 2.4939522743225098,
"step": 2068
},
{
"epoch": 0.12545454545454546,
"grad_norm": 0.08624905347824097,
"learning_rate": 0.0008,
"loss": 2.4565269947052,
"step": 2070
},
{
"epoch": 0.12557575757575756,
"grad_norm": 0.08601079881191254,
"learning_rate": 0.0008,
"loss": 2.550020217895508,
"step": 2072
},
{
"epoch": 0.1256969696969697,
"grad_norm": 0.0825832188129425,
"learning_rate": 0.0008,
"loss": 2.533282995223999,
"step": 2074
},
{
"epoch": 0.12581818181818183,
"grad_norm": 0.08027687668800354,
"learning_rate": 0.0008,
"loss": 2.5443296432495117,
"step": 2076
},
{
"epoch": 0.12593939393939393,
"grad_norm": 0.07983439415693283,
"learning_rate": 0.0008,
"loss": 2.4913129806518555,
"step": 2078
},
{
"epoch": 0.12606060606060607,
"grad_norm": 0.08081002533435822,
"learning_rate": 0.0008,
"loss": 2.5032520294189453,
"step": 2080
},
{
"epoch": 0.12618181818181817,
"grad_norm": 0.07518228888511658,
"learning_rate": 0.0008,
"loss": 2.4649040699005127,
"step": 2082
},
{
"epoch": 0.1263030303030303,
"grad_norm": 0.0776754766702652,
"learning_rate": 0.0008,
"loss": 2.530956506729126,
"step": 2084
},
{
"epoch": 0.1264242424242424,
"grad_norm": 0.07367412000894547,
"learning_rate": 0.0008,
"loss": 2.5687143802642822,
"step": 2086
},
{
"epoch": 0.12654545454545454,
"grad_norm": 0.0767262652516365,
"learning_rate": 0.0008,
"loss": 2.5434491634368896,
"step": 2088
},
{
"epoch": 0.12666666666666668,
"grad_norm": 0.07795710861682892,
"learning_rate": 0.0008,
"loss": 2.5952775478363037,
"step": 2090
},
{
"epoch": 0.12678787878787878,
"grad_norm": 0.07448644936084747,
"learning_rate": 0.0008,
"loss": 2.470137357711792,
"step": 2092
},
{
"epoch": 0.12690909090909092,
"grad_norm": 0.07456477731466293,
"learning_rate": 0.0008,
"loss": 2.5189876556396484,
"step": 2094
},
{
"epoch": 0.12703030303030302,
"grad_norm": 0.07434865087270737,
"learning_rate": 0.0008,
"loss": 2.423478126525879,
"step": 2096
},
{
"epoch": 0.12715151515151515,
"grad_norm": 0.07545511424541473,
"learning_rate": 0.0008,
"loss": 2.4730913639068604,
"step": 2098
},
{
"epoch": 0.12727272727272726,
"grad_norm": 0.08423464745283127,
"learning_rate": 0.0008,
"loss": 2.6545050144195557,
"step": 2100
},
{
"epoch": 0.1273939393939394,
"grad_norm": 0.07716361433267593,
"learning_rate": 0.0008,
"loss": 2.554306983947754,
"step": 2102
},
{
"epoch": 0.12751515151515153,
"grad_norm": 0.07742869853973389,
"learning_rate": 0.0008,
"loss": 2.505958318710327,
"step": 2104
},
{
"epoch": 0.12763636363636363,
"grad_norm": 0.07398577779531479,
"learning_rate": 0.0008,
"loss": 2.520829200744629,
"step": 2106
},
{
"epoch": 0.12775757575757576,
"grad_norm": 0.0772351399064064,
"learning_rate": 0.0008,
"loss": 2.527123212814331,
"step": 2108
},
{
"epoch": 0.12787878787878787,
"grad_norm": 0.07603088766336441,
"learning_rate": 0.0008,
"loss": 2.5904555320739746,
"step": 2110
},
{
"epoch": 0.128,
"grad_norm": 0.07599996775388718,
"learning_rate": 0.0008,
"loss": 2.550854444503784,
"step": 2112
},
{
"epoch": 0.1281212121212121,
"grad_norm": 0.07388272136449814,
"learning_rate": 0.0008,
"loss": 2.5806689262390137,
"step": 2114
},
{
"epoch": 0.12824242424242424,
"grad_norm": 0.07664526253938675,
"learning_rate": 0.0008,
"loss": 2.4167685508728027,
"step": 2116
},
{
"epoch": 0.12836363636363637,
"grad_norm": 0.0726180225610733,
"learning_rate": 0.0008,
"loss": 2.4784727096557617,
"step": 2118
},
{
"epoch": 0.12848484848484848,
"grad_norm": 0.07451099902391434,
"learning_rate": 0.0008,
"loss": 2.520994186401367,
"step": 2120
},
{
"epoch": 0.1286060606060606,
"grad_norm": 0.07405152171850204,
"learning_rate": 0.0008,
"loss": 2.4232075214385986,
"step": 2122
},
{
"epoch": 0.12872727272727272,
"grad_norm": 0.07723712921142578,
"learning_rate": 0.0008,
"loss": 2.5508410930633545,
"step": 2124
},
{
"epoch": 0.12884848484848485,
"grad_norm": 0.0753454864025116,
"learning_rate": 0.0008,
"loss": 2.5554134845733643,
"step": 2126
},
{
"epoch": 0.12896969696969696,
"grad_norm": 0.07345808297395706,
"learning_rate": 0.0008,
"loss": 2.5491771697998047,
"step": 2128
},
{
"epoch": 0.1290909090909091,
"grad_norm": 0.07486522942781448,
"learning_rate": 0.0008,
"loss": 2.5477957725524902,
"step": 2130
},
{
"epoch": 0.12921212121212122,
"grad_norm": 0.08247523754835129,
"learning_rate": 0.0008,
"loss": 2.5828213691711426,
"step": 2132
},
{
"epoch": 0.12933333333333333,
"grad_norm": 0.07372749596834183,
"learning_rate": 0.0008,
"loss": 2.4696850776672363,
"step": 2134
},
{
"epoch": 0.12945454545454546,
"grad_norm": 0.07393962144851685,
"learning_rate": 0.0008,
"loss": 2.489823818206787,
"step": 2136
},
{
"epoch": 0.12957575757575757,
"grad_norm": 0.07611682265996933,
"learning_rate": 0.0008,
"loss": 2.4876794815063477,
"step": 2138
},
{
"epoch": 0.1296969696969697,
"grad_norm": 0.08126074075698853,
"learning_rate": 0.0008,
"loss": 2.5266056060791016,
"step": 2140
},
{
"epoch": 0.1298181818181818,
"grad_norm": 0.07929164171218872,
"learning_rate": 0.0008,
"loss": 2.51462721824646,
"step": 2142
},
{
"epoch": 0.12993939393939394,
"grad_norm": 0.07550403475761414,
"learning_rate": 0.0008,
"loss": 2.525212049484253,
"step": 2144
},
{
"epoch": 0.13006060606060607,
"grad_norm": 0.0771554484963417,
"learning_rate": 0.0008,
"loss": 2.488828659057617,
"step": 2146
},
{
"epoch": 0.13018181818181818,
"grad_norm": 0.07402404397726059,
"learning_rate": 0.0008,
"loss": 2.519540548324585,
"step": 2148
},
{
"epoch": 0.1303030303030303,
"grad_norm": 0.07653609663248062,
"learning_rate": 0.0008,
"loss": 2.5428454875946045,
"step": 2150
},
{
"epoch": 0.13042424242424241,
"grad_norm": 0.07732845097780228,
"learning_rate": 0.0008,
"loss": 2.4963057041168213,
"step": 2152
},
{
"epoch": 0.13054545454545455,
"grad_norm": 0.07697934657335281,
"learning_rate": 0.0008,
"loss": 2.587296485900879,
"step": 2154
},
{
"epoch": 0.13066666666666665,
"grad_norm": 0.07391789555549622,
"learning_rate": 0.0008,
"loss": 2.4827427864074707,
"step": 2156
},
{
"epoch": 0.1307878787878788,
"grad_norm": 0.07302044332027435,
"learning_rate": 0.0008,
"loss": 2.437274217605591,
"step": 2158
},
{
"epoch": 0.13090909090909092,
"grad_norm": 0.07194291800260544,
"learning_rate": 0.0008,
"loss": 2.4735991954803467,
"step": 2160
},
{
"epoch": 0.13103030303030302,
"grad_norm": 0.07477825880050659,
"learning_rate": 0.0008,
"loss": 2.5134358406066895,
"step": 2162
},
{
"epoch": 0.13115151515151516,
"grad_norm": 0.07711788266897202,
"learning_rate": 0.0008,
"loss": 2.524251699447632,
"step": 2164
},
{
"epoch": 0.13127272727272726,
"grad_norm": 0.07189943641424179,
"learning_rate": 0.0008,
"loss": 2.5326485633850098,
"step": 2166
},
{
"epoch": 0.1313939393939394,
"grad_norm": 0.07217628508806229,
"learning_rate": 0.0008,
"loss": 2.5515475273132324,
"step": 2168
},
{
"epoch": 0.1315151515151515,
"grad_norm": 0.07369779795408249,
"learning_rate": 0.0008,
"loss": 2.55915904045105,
"step": 2170
},
{
"epoch": 0.13163636363636363,
"grad_norm": 0.07374408841133118,
"learning_rate": 0.0008,
"loss": 2.5461792945861816,
"step": 2172
},
{
"epoch": 0.13175757575757577,
"grad_norm": 0.0731811597943306,
"learning_rate": 0.0008,
"loss": 2.4840707778930664,
"step": 2174
},
{
"epoch": 0.13187878787878787,
"grad_norm": 0.07581745088100433,
"learning_rate": 0.0008,
"loss": 2.5639140605926514,
"step": 2176
},
{
"epoch": 0.132,
"grad_norm": 0.07553887367248535,
"learning_rate": 0.0008,
"loss": 2.478267192840576,
"step": 2178
},
{
"epoch": 0.1321212121212121,
"grad_norm": 0.07477521896362305,
"learning_rate": 0.0008,
"loss": 2.571526288986206,
"step": 2180
},
{
"epoch": 0.13224242424242424,
"grad_norm": 0.07332218438386917,
"learning_rate": 0.0008,
"loss": 2.5956995487213135,
"step": 2182
},
{
"epoch": 0.13236363636363635,
"grad_norm": 0.07387803494930267,
"learning_rate": 0.0008,
"loss": 2.5755953788757324,
"step": 2184
},
{
"epoch": 0.13248484848484848,
"grad_norm": 0.0715310350060463,
"learning_rate": 0.0008,
"loss": 2.429337501525879,
"step": 2186
},
{
"epoch": 0.13260606060606062,
"grad_norm": 0.07476066052913666,
"learning_rate": 0.0008,
"loss": 2.5714337825775146,
"step": 2188
},
{
"epoch": 0.13272727272727272,
"grad_norm": 0.07082120329141617,
"learning_rate": 0.0008,
"loss": 2.392146587371826,
"step": 2190
},
{
"epoch": 0.13284848484848485,
"grad_norm": 0.07317781448364258,
"learning_rate": 0.0008,
"loss": 2.468334674835205,
"step": 2192
},
{
"epoch": 0.13296969696969696,
"grad_norm": 0.07456928491592407,
"learning_rate": 0.0008,
"loss": 2.5419490337371826,
"step": 2194
},
{
"epoch": 0.1330909090909091,
"grad_norm": 0.07272838056087494,
"learning_rate": 0.0008,
"loss": 2.524724245071411,
"step": 2196
},
{
"epoch": 0.1332121212121212,
"grad_norm": 0.07574515789747238,
"learning_rate": 0.0008,
"loss": 2.5079774856567383,
"step": 2198
},
{
"epoch": 0.13333333333333333,
"grad_norm": 0.07788367569446564,
"learning_rate": 0.0008,
"loss": 2.506497383117676,
"step": 2200
},
{
"epoch": 0.13345454545454546,
"grad_norm": 0.07362538576126099,
"learning_rate": 0.0008,
"loss": 2.5331404209136963,
"step": 2202
},
{
"epoch": 0.13357575757575757,
"grad_norm": 0.07488234341144562,
"learning_rate": 0.0008,
"loss": 2.5219509601593018,
"step": 2204
},
{
"epoch": 0.1336969696969697,
"grad_norm": 0.07592590898275375,
"learning_rate": 0.0008,
"loss": 2.537665367126465,
"step": 2206
},
{
"epoch": 0.1338181818181818,
"grad_norm": 0.07888289541006088,
"learning_rate": 0.0008,
"loss": 2.5398998260498047,
"step": 2208
},
{
"epoch": 0.13393939393939394,
"grad_norm": 0.07390788197517395,
"learning_rate": 0.0008,
"loss": 2.485811233520508,
"step": 2210
},
{
"epoch": 0.13406060606060605,
"grad_norm": 0.07291394472122192,
"learning_rate": 0.0008,
"loss": 2.504791498184204,
"step": 2212
},
{
"epoch": 0.13418181818181818,
"grad_norm": 0.0775604099035263,
"learning_rate": 0.0008,
"loss": 2.483065128326416,
"step": 2214
},
{
"epoch": 0.1343030303030303,
"grad_norm": 0.07665056735277176,
"learning_rate": 0.0008,
"loss": 2.5505192279815674,
"step": 2216
},
{
"epoch": 0.13442424242424242,
"grad_norm": 0.07134459167718887,
"learning_rate": 0.0008,
"loss": 2.4685513973236084,
"step": 2218
},
{
"epoch": 0.13454545454545455,
"grad_norm": 0.07204804569482803,
"learning_rate": 0.0008,
"loss": 2.5522944927215576,
"step": 2220
},
{
"epoch": 0.13466666666666666,
"grad_norm": 0.07351890951395035,
"learning_rate": 0.0008,
"loss": 2.478203296661377,
"step": 2222
},
{
"epoch": 0.1347878787878788,
"grad_norm": 0.07468769699335098,
"learning_rate": 0.0008,
"loss": 2.49233078956604,
"step": 2224
},
{
"epoch": 0.13490909090909092,
"grad_norm": 0.07314802706241608,
"learning_rate": 0.0008,
"loss": 2.4525957107543945,
"step": 2226
},
{
"epoch": 0.13503030303030303,
"grad_norm": 0.07337017357349396,
"learning_rate": 0.0008,
"loss": 2.5246360301971436,
"step": 2228
},
{
"epoch": 0.13515151515151516,
"grad_norm": 0.07445929199457169,
"learning_rate": 0.0008,
"loss": 2.492366075515747,
"step": 2230
},
{
"epoch": 0.13527272727272727,
"grad_norm": 0.0744335874915123,
"learning_rate": 0.0008,
"loss": 2.525787115097046,
"step": 2232
},
{
"epoch": 0.1353939393939394,
"grad_norm": 0.07010258734226227,
"learning_rate": 0.0008,
"loss": 2.506089925765991,
"step": 2234
},
{
"epoch": 0.1355151515151515,
"grad_norm": 0.07127749919891357,
"learning_rate": 0.0008,
"loss": 2.51304030418396,
"step": 2236
},
{
"epoch": 0.13563636363636364,
"grad_norm": 0.073407843708992,
"learning_rate": 0.0008,
"loss": 2.4755821228027344,
"step": 2238
},
{
"epoch": 0.13575757575757577,
"grad_norm": 0.07343313097953796,
"learning_rate": 0.0008,
"loss": 2.561004161834717,
"step": 2240
},
{
"epoch": 0.13587878787878788,
"grad_norm": 0.07378042489290237,
"learning_rate": 0.0008,
"loss": 2.4569945335388184,
"step": 2242
},
{
"epoch": 0.136,
"grad_norm": 0.07340779155492783,
"learning_rate": 0.0008,
"loss": 2.5376200675964355,
"step": 2244
},
{
"epoch": 0.13612121212121212,
"grad_norm": 0.07371710985898972,
"learning_rate": 0.0008,
"loss": 2.514014720916748,
"step": 2246
},
{
"epoch": 0.13624242424242425,
"grad_norm": 0.06963010132312775,
"learning_rate": 0.0008,
"loss": 2.4618935585021973,
"step": 2248
},
{
"epoch": 0.13636363636363635,
"grad_norm": 0.0722995176911354,
"learning_rate": 0.0008,
"loss": 2.4817774295806885,
"step": 2250
},
{
"epoch": 0.1364848484848485,
"grad_norm": 0.07103802263736725,
"learning_rate": 0.0008,
"loss": 2.4677326679229736,
"step": 2252
},
{
"epoch": 0.13660606060606062,
"grad_norm": 0.07134968042373657,
"learning_rate": 0.0008,
"loss": 2.549239158630371,
"step": 2254
},
{
"epoch": 0.13672727272727273,
"grad_norm": 0.07036776095628738,
"learning_rate": 0.0008,
"loss": 2.4914135932922363,
"step": 2256
},
{
"epoch": 0.13684848484848486,
"grad_norm": 0.07311167567968369,
"learning_rate": 0.0008,
"loss": 2.452558994293213,
"step": 2258
},
{
"epoch": 0.13696969696969696,
"grad_norm": 0.07037466019392014,
"learning_rate": 0.0008,
"loss": 2.513230323791504,
"step": 2260
},
{
"epoch": 0.1370909090909091,
"grad_norm": 0.07552643865346909,
"learning_rate": 0.0008,
"loss": 2.4842469692230225,
"step": 2262
},
{
"epoch": 0.1372121212121212,
"grad_norm": 0.07736295461654663,
"learning_rate": 0.0008,
"loss": 2.5280025005340576,
"step": 2264
},
{
"epoch": 0.13733333333333334,
"grad_norm": 0.0720943808555603,
"learning_rate": 0.0008,
"loss": 2.4819672107696533,
"step": 2266
},
{
"epoch": 0.13745454545454547,
"grad_norm": 0.07486707717180252,
"learning_rate": 0.0008,
"loss": 2.5478708744049072,
"step": 2268
},
{
"epoch": 0.13757575757575757,
"grad_norm": 0.0723852589726448,
"learning_rate": 0.0008,
"loss": 2.5189926624298096,
"step": 2270
},
{
"epoch": 0.1376969696969697,
"grad_norm": 0.07722058892250061,
"learning_rate": 0.0008,
"loss": 2.4988040924072266,
"step": 2272
},
{
"epoch": 0.1378181818181818,
"grad_norm": 0.07257533818483353,
"learning_rate": 0.0008,
"loss": 2.4370455741882324,
"step": 2274
},
{
"epoch": 0.13793939393939394,
"grad_norm": 0.0718521848320961,
"learning_rate": 0.0008,
"loss": 2.3745577335357666,
"step": 2276
},
{
"epoch": 0.13806060606060605,
"grad_norm": 0.07189729064702988,
"learning_rate": 0.0008,
"loss": 2.451524496078491,
"step": 2278
},
{
"epoch": 0.13818181818181818,
"grad_norm": 0.07054579257965088,
"learning_rate": 0.0008,
"loss": 2.4847376346588135,
"step": 2280
},
{
"epoch": 0.13830303030303032,
"grad_norm": 0.0712272897362709,
"learning_rate": 0.0008,
"loss": 2.4538259506225586,
"step": 2282
},
{
"epoch": 0.13842424242424242,
"grad_norm": 0.07186871767044067,
"learning_rate": 0.0008,
"loss": 2.4380297660827637,
"step": 2284
},
{
"epoch": 0.13854545454545455,
"grad_norm": 0.07121531665325165,
"learning_rate": 0.0008,
"loss": 2.519989013671875,
"step": 2286
},
{
"epoch": 0.13866666666666666,
"grad_norm": 0.07021085917949677,
"learning_rate": 0.0008,
"loss": 2.4404232501983643,
"step": 2288
},
{
"epoch": 0.1387878787878788,
"grad_norm": 0.07317303866147995,
"learning_rate": 0.0008,
"loss": 2.4517321586608887,
"step": 2290
},
{
"epoch": 0.1389090909090909,
"grad_norm": 0.0714523196220398,
"learning_rate": 0.0008,
"loss": 2.405569076538086,
"step": 2292
},
{
"epoch": 0.13903030303030303,
"grad_norm": 0.07365009933710098,
"learning_rate": 0.0008,
"loss": 2.5651137828826904,
"step": 2294
},
{
"epoch": 0.13915151515151516,
"grad_norm": 0.07251730561256409,
"learning_rate": 0.0008,
"loss": 2.4200851917266846,
"step": 2296
},
{
"epoch": 0.13927272727272727,
"grad_norm": 0.07165366411209106,
"learning_rate": 0.0008,
"loss": 2.600093126296997,
"step": 2298
},
{
"epoch": 0.1393939393939394,
"grad_norm": 0.07138187438249588,
"learning_rate": 0.0008,
"loss": 2.509793996810913,
"step": 2300
},
{
"epoch": 0.1395151515151515,
"grad_norm": 0.07166871428489685,
"learning_rate": 0.0008,
"loss": 2.4683752059936523,
"step": 2302
},
{
"epoch": 0.13963636363636364,
"grad_norm": 0.07308699935674667,
"learning_rate": 0.0008,
"loss": 2.475144863128662,
"step": 2304
},
{
"epoch": 0.13975757575757575,
"grad_norm": 0.07093770056962967,
"learning_rate": 0.0008,
"loss": 2.4853596687316895,
"step": 2306
},
{
"epoch": 0.13987878787878788,
"grad_norm": 0.06988460570573807,
"learning_rate": 0.0008,
"loss": 2.497823715209961,
"step": 2308
},
{
"epoch": 0.14,
"grad_norm": 0.07140327244997025,
"learning_rate": 0.0008,
"loss": 2.39980411529541,
"step": 2310
},
{
"epoch": 0.14012121212121212,
"grad_norm": 0.07309526205062866,
"learning_rate": 0.0008,
"loss": 2.5716552734375,
"step": 2312
},
{
"epoch": 0.14024242424242425,
"grad_norm": 0.07083943486213684,
"learning_rate": 0.0008,
"loss": 2.3660707473754883,
"step": 2314
},
{
"epoch": 0.14036363636363636,
"grad_norm": 0.07393694669008255,
"learning_rate": 0.0008,
"loss": 2.538632392883301,
"step": 2316
},
{
"epoch": 0.1404848484848485,
"grad_norm": 0.07225868105888367,
"learning_rate": 0.0008,
"loss": 2.427572250366211,
"step": 2318
},
{
"epoch": 0.1406060606060606,
"grad_norm": 0.069944828748703,
"learning_rate": 0.0008,
"loss": 2.4053213596343994,
"step": 2320
},
{
"epoch": 0.14072727272727273,
"grad_norm": 0.07577000558376312,
"learning_rate": 0.0008,
"loss": 2.51894474029541,
"step": 2322
},
{
"epoch": 0.14084848484848486,
"grad_norm": 0.07076112926006317,
"learning_rate": 0.0008,
"loss": 2.4552700519561768,
"step": 2324
},
{
"epoch": 0.14096969696969697,
"grad_norm": 0.07136660069227219,
"learning_rate": 0.0008,
"loss": 2.5086584091186523,
"step": 2326
},
{
"epoch": 0.1410909090909091,
"grad_norm": 0.0752607062458992,
"learning_rate": 0.0008,
"loss": 2.5506982803344727,
"step": 2328
},
{
"epoch": 0.1412121212121212,
"grad_norm": 0.07571206986904144,
"learning_rate": 0.0008,
"loss": 2.4211411476135254,
"step": 2330
},
{
"epoch": 0.14133333333333334,
"grad_norm": 0.07219422608613968,
"learning_rate": 0.0008,
"loss": 2.4734246730804443,
"step": 2332
},
{
"epoch": 0.14145454545454544,
"grad_norm": 0.0744132325053215,
"learning_rate": 0.0008,
"loss": 2.467674493789673,
"step": 2334
},
{
"epoch": 0.14157575757575758,
"grad_norm": 0.07347333431243896,
"learning_rate": 0.0008,
"loss": 2.455476999282837,
"step": 2336
},
{
"epoch": 0.1416969696969697,
"grad_norm": 0.07262501865625381,
"learning_rate": 0.0008,
"loss": 2.5193984508514404,
"step": 2338
},
{
"epoch": 0.14181818181818182,
"grad_norm": 0.07279674708843231,
"learning_rate": 0.0008,
"loss": 2.4847795963287354,
"step": 2340
},
{
"epoch": 0.14193939393939395,
"grad_norm": 0.0720701664686203,
"learning_rate": 0.0008,
"loss": 2.4036433696746826,
"step": 2342
},
{
"epoch": 0.14206060606060605,
"grad_norm": 0.07401404529809952,
"learning_rate": 0.0008,
"loss": 2.5127511024475098,
"step": 2344
},
{
"epoch": 0.1421818181818182,
"grad_norm": 0.07665159553289413,
"learning_rate": 0.0008,
"loss": 2.5259947776794434,
"step": 2346
},
{
"epoch": 0.1423030303030303,
"grad_norm": 0.07047493010759354,
"learning_rate": 0.0008,
"loss": 2.4933364391326904,
"step": 2348
},
{
"epoch": 0.14242424242424243,
"grad_norm": 0.07500440627336502,
"learning_rate": 0.0008,
"loss": 2.4921510219573975,
"step": 2350
},
{
"epoch": 0.14254545454545456,
"grad_norm": 0.07529532164335251,
"learning_rate": 0.0008,
"loss": 2.4778919219970703,
"step": 2352
},
{
"epoch": 0.14266666666666666,
"grad_norm": 0.07599107921123505,
"learning_rate": 0.0008,
"loss": 2.467775344848633,
"step": 2354
},
{
"epoch": 0.1427878787878788,
"grad_norm": 0.07355376332998276,
"learning_rate": 0.0008,
"loss": 2.5463266372680664,
"step": 2356
},
{
"epoch": 0.1429090909090909,
"grad_norm": 0.07381011545658112,
"learning_rate": 0.0008,
"loss": 2.531032085418701,
"step": 2358
},
{
"epoch": 0.14303030303030304,
"grad_norm": 0.06994525343179703,
"learning_rate": 0.0008,
"loss": 2.5959417819976807,
"step": 2360
},
{
"epoch": 0.14315151515151514,
"grad_norm": 0.07043395191431046,
"learning_rate": 0.0008,
"loss": 2.440598726272583,
"step": 2362
},
{
"epoch": 0.14327272727272727,
"grad_norm": 0.07186504453420639,
"learning_rate": 0.0008,
"loss": 2.523922920227051,
"step": 2364
},
{
"epoch": 0.1433939393939394,
"grad_norm": 0.07145015895366669,
"learning_rate": 0.0008,
"loss": 2.533161163330078,
"step": 2366
},
{
"epoch": 0.1435151515151515,
"grad_norm": 0.0677226111292839,
"learning_rate": 0.0008,
"loss": 2.4668498039245605,
"step": 2368
},
{
"epoch": 0.14363636363636365,
"grad_norm": 0.06843527406454086,
"learning_rate": 0.0008,
"loss": 2.3922762870788574,
"step": 2370
},
{
"epoch": 0.14375757575757575,
"grad_norm": 0.06960544735193253,
"learning_rate": 0.0008,
"loss": 2.4381422996520996,
"step": 2372
},
{
"epoch": 0.14387878787878788,
"grad_norm": 0.07168358564376831,
"learning_rate": 0.0008,
"loss": 2.4416918754577637,
"step": 2374
},
{
"epoch": 0.144,
"grad_norm": 0.07600297778844833,
"learning_rate": 0.0008,
"loss": 2.5091309547424316,
"step": 2376
},
{
"epoch": 0.14412121212121212,
"grad_norm": 0.07379657030105591,
"learning_rate": 0.0008,
"loss": 2.548192024230957,
"step": 2378
},
{
"epoch": 0.14424242424242426,
"grad_norm": 0.07520291954278946,
"learning_rate": 0.0008,
"loss": 2.454476833343506,
"step": 2380
},
{
"epoch": 0.14436363636363636,
"grad_norm": 0.0726599246263504,
"learning_rate": 0.0008,
"loss": 2.5277116298675537,
"step": 2382
},
{
"epoch": 0.1444848484848485,
"grad_norm": 0.07089592516422272,
"learning_rate": 0.0008,
"loss": 2.494361639022827,
"step": 2384
},
{
"epoch": 0.1446060606060606,
"grad_norm": 0.07306011766195297,
"learning_rate": 0.0008,
"loss": 2.5014586448669434,
"step": 2386
},
{
"epoch": 0.14472727272727273,
"grad_norm": 0.07222026586532593,
"learning_rate": 0.0008,
"loss": 2.5695860385894775,
"step": 2388
},
{
"epoch": 0.14484848484848484,
"grad_norm": 0.07006663829088211,
"learning_rate": 0.0008,
"loss": 2.4453513622283936,
"step": 2390
},
{
"epoch": 0.14496969696969697,
"grad_norm": 0.07333624362945557,
"learning_rate": 0.0008,
"loss": 2.4533238410949707,
"step": 2392
},
{
"epoch": 0.1450909090909091,
"grad_norm": 0.0818651020526886,
"learning_rate": 0.0008,
"loss": 2.4851746559143066,
"step": 2394
},
{
"epoch": 0.1452121212121212,
"grad_norm": 0.07136614620685577,
"learning_rate": 0.0008,
"loss": 2.508509397506714,
"step": 2396
},
{
"epoch": 0.14533333333333334,
"grad_norm": 0.06939436495304108,
"learning_rate": 0.0008,
"loss": 2.476149797439575,
"step": 2398
},
{
"epoch": 0.14545454545454545,
"grad_norm": 0.07145765423774719,
"learning_rate": 0.0008,
"loss": 2.418069362640381,
"step": 2400
},
{
"epoch": 0.14557575757575758,
"grad_norm": 0.07355836033821106,
"learning_rate": 0.0008,
"loss": 2.483440637588501,
"step": 2402
},
{
"epoch": 0.14569696969696969,
"grad_norm": 0.07323364168405533,
"learning_rate": 0.0008,
"loss": 2.555269718170166,
"step": 2404
},
{
"epoch": 0.14581818181818182,
"grad_norm": 0.07346950471401215,
"learning_rate": 0.0008,
"loss": 2.4819021224975586,
"step": 2406
},
{
"epoch": 0.14593939393939395,
"grad_norm": 0.07405179738998413,
"learning_rate": 0.0008,
"loss": 2.439404249191284,
"step": 2408
},
{
"epoch": 0.14606060606060606,
"grad_norm": 0.07432809472084045,
"learning_rate": 0.0008,
"loss": 2.466127872467041,
"step": 2410
},
{
"epoch": 0.1461818181818182,
"grad_norm": 0.07479830086231232,
"learning_rate": 0.0008,
"loss": 2.409163475036621,
"step": 2412
},
{
"epoch": 0.1463030303030303,
"grad_norm": 0.07066427171230316,
"learning_rate": 0.0008,
"loss": 2.484388589859009,
"step": 2414
},
{
"epoch": 0.14642424242424243,
"grad_norm": 0.07200539112091064,
"learning_rate": 0.0008,
"loss": 2.502199649810791,
"step": 2416
},
{
"epoch": 0.14654545454545453,
"grad_norm": 0.0723142996430397,
"learning_rate": 0.0008,
"loss": 2.3982980251312256,
"step": 2418
},
{
"epoch": 0.14666666666666667,
"grad_norm": 0.07178358733654022,
"learning_rate": 0.0008,
"loss": 2.4394493103027344,
"step": 2420
},
{
"epoch": 0.1467878787878788,
"grad_norm": 0.0762401670217514,
"learning_rate": 0.0008,
"loss": 2.447188377380371,
"step": 2422
},
{
"epoch": 0.1469090909090909,
"grad_norm": 0.0724373608827591,
"learning_rate": 0.0008,
"loss": 2.4016754627227783,
"step": 2424
},
{
"epoch": 0.14703030303030304,
"grad_norm": 0.07091701030731201,
"learning_rate": 0.0008,
"loss": 2.4485421180725098,
"step": 2426
},
{
"epoch": 0.14715151515151514,
"grad_norm": 0.07004357874393463,
"learning_rate": 0.0008,
"loss": 2.4469900131225586,
"step": 2428
},
{
"epoch": 0.14727272727272728,
"grad_norm": 0.06801891326904297,
"learning_rate": 0.0008,
"loss": 2.448176860809326,
"step": 2430
},
{
"epoch": 0.14739393939393938,
"grad_norm": 0.0710800513625145,
"learning_rate": 0.0008,
"loss": 2.5071828365325928,
"step": 2432
},
{
"epoch": 0.14751515151515152,
"grad_norm": 0.07216828316450119,
"learning_rate": 0.0008,
"loss": 2.5195682048797607,
"step": 2434
},
{
"epoch": 0.14763636363636365,
"grad_norm": 0.07173492014408112,
"learning_rate": 0.0008,
"loss": 2.4219794273376465,
"step": 2436
},
{
"epoch": 0.14775757575757575,
"grad_norm": 0.06855220347642899,
"learning_rate": 0.0008,
"loss": 2.4597597122192383,
"step": 2438
},
{
"epoch": 0.1478787878787879,
"grad_norm": 0.06821279227733612,
"learning_rate": 0.0008,
"loss": 2.4819204807281494,
"step": 2440
},
{
"epoch": 0.148,
"grad_norm": 0.06907371431589127,
"learning_rate": 0.0008,
"loss": 2.5156192779541016,
"step": 2442
},
{
"epoch": 0.14812121212121213,
"grad_norm": 0.06822527945041656,
"learning_rate": 0.0008,
"loss": 2.441620111465454,
"step": 2444
},
{
"epoch": 0.14824242424242423,
"grad_norm": 0.07177390903234482,
"learning_rate": 0.0008,
"loss": 2.43461275100708,
"step": 2446
},
{
"epoch": 0.14836363636363636,
"grad_norm": 0.06998538225889206,
"learning_rate": 0.0008,
"loss": 2.503647804260254,
"step": 2448
},
{
"epoch": 0.1484848484848485,
"grad_norm": 0.07186067849397659,
"learning_rate": 0.0008,
"loss": 2.425893783569336,
"step": 2450
},
{
"epoch": 0.1486060606060606,
"grad_norm": 0.07158838957548141,
"learning_rate": 0.0008,
"loss": 2.5321290493011475,
"step": 2452
},
{
"epoch": 0.14872727272727274,
"grad_norm": 0.07196220010519028,
"learning_rate": 0.0008,
"loss": 2.530115842819214,
"step": 2454
},
{
"epoch": 0.14884848484848484,
"grad_norm": 0.06851901113986969,
"learning_rate": 0.0008,
"loss": 2.4448037147521973,
"step": 2456
},
{
"epoch": 0.14896969696969697,
"grad_norm": 0.0719350203871727,
"learning_rate": 0.0008,
"loss": 2.479567527770996,
"step": 2458
},
{
"epoch": 0.14909090909090908,
"grad_norm": 0.07139955461025238,
"learning_rate": 0.0008,
"loss": 2.3601980209350586,
"step": 2460
},
{
"epoch": 0.1492121212121212,
"grad_norm": 0.0715802013874054,
"learning_rate": 0.0008,
"loss": 2.4668126106262207,
"step": 2462
},
{
"epoch": 0.14933333333333335,
"grad_norm": 0.07088276743888855,
"learning_rate": 0.0008,
"loss": 2.390610694885254,
"step": 2464
},
{
"epoch": 0.14945454545454545,
"grad_norm": 0.07396944612264633,
"learning_rate": 0.0008,
"loss": 2.51900053024292,
"step": 2466
},
{
"epoch": 0.14957575757575758,
"grad_norm": 0.06900956481695175,
"learning_rate": 0.0008,
"loss": 2.4021401405334473,
"step": 2468
},
{
"epoch": 0.1496969696969697,
"grad_norm": 0.06944061070680618,
"learning_rate": 0.0008,
"loss": 2.4413774013519287,
"step": 2470
},
{
"epoch": 0.14981818181818182,
"grad_norm": 0.06902553886175156,
"learning_rate": 0.0008,
"loss": 2.46260404586792,
"step": 2472
},
{
"epoch": 0.14993939393939393,
"grad_norm": 0.06915315985679626,
"learning_rate": 0.0008,
"loss": 2.4501993656158447,
"step": 2474
},
{
"epoch": 0.15006060606060606,
"grad_norm": 0.06937610357999802,
"learning_rate": 0.0008,
"loss": 2.5042691230773926,
"step": 2476
},
{
"epoch": 0.1501818181818182,
"grad_norm": 0.07008138298988342,
"learning_rate": 0.0008,
"loss": 2.5074563026428223,
"step": 2478
},
{
"epoch": 0.1503030303030303,
"grad_norm": 0.06869629770517349,
"learning_rate": 0.0008,
"loss": 2.506767988204956,
"step": 2480
},
{
"epoch": 0.15042424242424243,
"grad_norm": 0.06948976963758469,
"learning_rate": 0.0008,
"loss": 2.4299073219299316,
"step": 2482
},
{
"epoch": 0.15054545454545454,
"grad_norm": 0.07050379365682602,
"learning_rate": 0.0008,
"loss": 2.484818696975708,
"step": 2484
},
{
"epoch": 0.15066666666666667,
"grad_norm": 0.06942024827003479,
"learning_rate": 0.0008,
"loss": 2.466052770614624,
"step": 2486
},
{
"epoch": 0.15078787878787878,
"grad_norm": 0.07226774841547012,
"learning_rate": 0.0008,
"loss": 2.477065324783325,
"step": 2488
},
{
"epoch": 0.1509090909090909,
"grad_norm": 0.07058286666870117,
"learning_rate": 0.0008,
"loss": 2.5282013416290283,
"step": 2490
},
{
"epoch": 0.15103030303030304,
"grad_norm": 0.07089857012033463,
"learning_rate": 0.0008,
"loss": 2.4953269958496094,
"step": 2492
},
{
"epoch": 0.15115151515151515,
"grad_norm": 0.06824969500303268,
"learning_rate": 0.0008,
"loss": 2.4419403076171875,
"step": 2494
},
{
"epoch": 0.15127272727272728,
"grad_norm": 0.07112861424684525,
"learning_rate": 0.0008,
"loss": 2.5386030673980713,
"step": 2496
},
{
"epoch": 0.1513939393939394,
"grad_norm": 0.07213275134563446,
"learning_rate": 0.0008,
"loss": 2.4411914348602295,
"step": 2498
},
{
"epoch": 0.15151515151515152,
"grad_norm": 0.07535449415445328,
"learning_rate": 0.0008,
"loss": 2.499537467956543,
"step": 2500
},
{
"epoch": 0.15163636363636362,
"grad_norm": 0.0688779428601265,
"learning_rate": 0.0008,
"loss": 2.4332680702209473,
"step": 2502
},
{
"epoch": 0.15175757575757576,
"grad_norm": 0.07011019438505173,
"learning_rate": 0.0008,
"loss": 2.4603607654571533,
"step": 2504
},
{
"epoch": 0.1518787878787879,
"grad_norm": 0.0677572563290596,
"learning_rate": 0.0008,
"loss": 2.5020153522491455,
"step": 2506
},
{
"epoch": 0.152,
"grad_norm": 0.06833836436271667,
"learning_rate": 0.0008,
"loss": 2.439638376235962,
"step": 2508
},
{
"epoch": 0.15212121212121213,
"grad_norm": 0.06761213392019272,
"learning_rate": 0.0008,
"loss": 2.432396411895752,
"step": 2510
},
{
"epoch": 0.15224242424242423,
"grad_norm": 0.06982916593551636,
"learning_rate": 0.0008,
"loss": 2.505734920501709,
"step": 2512
},
{
"epoch": 0.15236363636363637,
"grad_norm": 0.07087282091379166,
"learning_rate": 0.0008,
"loss": 2.4637231826782227,
"step": 2514
},
{
"epoch": 0.15248484848484847,
"grad_norm": 0.07173216342926025,
"learning_rate": 0.0008,
"loss": 2.492630958557129,
"step": 2516
},
{
"epoch": 0.1526060606060606,
"grad_norm": 0.06881629675626755,
"learning_rate": 0.0008,
"loss": 2.437014102935791,
"step": 2518
},
{
"epoch": 0.15272727272727274,
"grad_norm": 0.06983078271150589,
"learning_rate": 0.0008,
"loss": 2.4408304691314697,
"step": 2520
},
{
"epoch": 0.15284848484848484,
"grad_norm": 0.06938973069190979,
"learning_rate": 0.0008,
"loss": 2.4200124740600586,
"step": 2522
},
{
"epoch": 0.15296969696969698,
"grad_norm": 0.0698440819978714,
"learning_rate": 0.0008,
"loss": 2.5287365913391113,
"step": 2524
},
{
"epoch": 0.15309090909090908,
"grad_norm": 0.06815673410892487,
"learning_rate": 0.0008,
"loss": 2.417501926422119,
"step": 2526
},
{
"epoch": 0.15321212121212122,
"grad_norm": 0.07132939994335175,
"learning_rate": 0.0008,
"loss": 2.4013030529022217,
"step": 2528
},
{
"epoch": 0.15333333333333332,
"grad_norm": 0.07225148379802704,
"learning_rate": 0.0008,
"loss": 2.427898406982422,
"step": 2530
},
{
"epoch": 0.15345454545454545,
"grad_norm": 0.07196836173534393,
"learning_rate": 0.0008,
"loss": 2.466188669204712,
"step": 2532
},
{
"epoch": 0.1535757575757576,
"grad_norm": 0.07452213764190674,
"learning_rate": 0.0008,
"loss": 2.4490408897399902,
"step": 2534
},
{
"epoch": 0.1536969696969697,
"grad_norm": 0.07136422395706177,
"learning_rate": 0.0008,
"loss": 2.4899814128875732,
"step": 2536
},
{
"epoch": 0.15381818181818183,
"grad_norm": 0.0711861327290535,
"learning_rate": 0.0008,
"loss": 2.4443297386169434,
"step": 2538
},
{
"epoch": 0.15393939393939393,
"grad_norm": 0.0667894035577774,
"learning_rate": 0.0008,
"loss": 2.5069639682769775,
"step": 2540
},
{
"epoch": 0.15406060606060606,
"grad_norm": 0.0689382553100586,
"learning_rate": 0.0008,
"loss": 2.5245916843414307,
"step": 2542
},
{
"epoch": 0.15418181818181817,
"grad_norm": 0.06836752593517303,
"learning_rate": 0.0008,
"loss": 2.4769952297210693,
"step": 2544
},
{
"epoch": 0.1543030303030303,
"grad_norm": 0.06621064990758896,
"learning_rate": 0.0008,
"loss": 2.350978374481201,
"step": 2546
},
{
"epoch": 0.15442424242424244,
"grad_norm": 0.06688772886991501,
"learning_rate": 0.0008,
"loss": 2.427025556564331,
"step": 2548
},
{
"epoch": 0.15454545454545454,
"grad_norm": 0.06703899800777435,
"learning_rate": 0.0008,
"loss": 2.373581886291504,
"step": 2550
},
{
"epoch": 0.15466666666666667,
"grad_norm": 0.0699489563703537,
"learning_rate": 0.0008,
"loss": 2.516940116882324,
"step": 2552
},
{
"epoch": 0.15478787878787878,
"grad_norm": 0.06721148639917374,
"learning_rate": 0.0008,
"loss": 2.4288089275360107,
"step": 2554
},
{
"epoch": 0.1549090909090909,
"grad_norm": 0.06832882016897202,
"learning_rate": 0.0008,
"loss": 2.4440484046936035,
"step": 2556
},
{
"epoch": 0.15503030303030302,
"grad_norm": 0.07132760435342789,
"learning_rate": 0.0008,
"loss": 2.417062282562256,
"step": 2558
},
{
"epoch": 0.15515151515151515,
"grad_norm": 0.0687774047255516,
"learning_rate": 0.0008,
"loss": 2.401379108428955,
"step": 2560
},
{
"epoch": 0.15527272727272728,
"grad_norm": 0.07053923606872559,
"learning_rate": 0.0008,
"loss": 2.3964309692382812,
"step": 2562
},
{
"epoch": 0.1553939393939394,
"grad_norm": 0.06890591233968735,
"learning_rate": 0.0008,
"loss": 2.45247220993042,
"step": 2564
},
{
"epoch": 0.15551515151515152,
"grad_norm": 0.06896795332431793,
"learning_rate": 0.0008,
"loss": 2.390634059906006,
"step": 2566
},
{
"epoch": 0.15563636363636363,
"grad_norm": 0.06625691056251526,
"learning_rate": 0.0008,
"loss": 2.487304925918579,
"step": 2568
},
{
"epoch": 0.15575757575757576,
"grad_norm": 0.07377240061759949,
"learning_rate": 0.0008,
"loss": 2.4417786598205566,
"step": 2570
},
{
"epoch": 0.15587878787878787,
"grad_norm": 0.06973312050104141,
"learning_rate": 0.0008,
"loss": 2.4549648761749268,
"step": 2572
},
{
"epoch": 0.156,
"grad_norm": 0.06808368861675262,
"learning_rate": 0.0008,
"loss": 2.381690263748169,
"step": 2574
},
{
"epoch": 0.15612121212121213,
"grad_norm": 0.06849049031734467,
"learning_rate": 0.0008,
"loss": 2.5121707916259766,
"step": 2576
},
{
"epoch": 0.15624242424242424,
"grad_norm": 0.06952201575040817,
"learning_rate": 0.0008,
"loss": 2.4710886478424072,
"step": 2578
},
{
"epoch": 0.15636363636363637,
"grad_norm": 0.06774258613586426,
"learning_rate": 0.0008,
"loss": 2.4576404094696045,
"step": 2580
},
{
"epoch": 0.15648484848484848,
"grad_norm": 0.06910566985607147,
"learning_rate": 0.0008,
"loss": 2.4034440517425537,
"step": 2582
},
{
"epoch": 0.1566060606060606,
"grad_norm": 0.07102058082818985,
"learning_rate": 0.0008,
"loss": 2.5657925605773926,
"step": 2584
},
{
"epoch": 0.15672727272727272,
"grad_norm": 0.06899001449346542,
"learning_rate": 0.0008,
"loss": 2.4725704193115234,
"step": 2586
},
{
"epoch": 0.15684848484848485,
"grad_norm": 0.06650899350643158,
"learning_rate": 0.0008,
"loss": 2.444744348526001,
"step": 2588
},
{
"epoch": 0.15696969696969698,
"grad_norm": 0.06645403802394867,
"learning_rate": 0.0008,
"loss": 2.428097724914551,
"step": 2590
},
{
"epoch": 0.1570909090909091,
"grad_norm": 0.0681406632065773,
"learning_rate": 0.0008,
"loss": 2.4638454914093018,
"step": 2592
},
{
"epoch": 0.15721212121212122,
"grad_norm": 0.06851319223642349,
"learning_rate": 0.0008,
"loss": 2.434051513671875,
"step": 2594
},
{
"epoch": 0.15733333333333333,
"grad_norm": 0.06931707262992859,
"learning_rate": 0.0008,
"loss": 2.489978313446045,
"step": 2596
},
{
"epoch": 0.15745454545454546,
"grad_norm": 0.06770271807909012,
"learning_rate": 0.0008,
"loss": 2.4493212699890137,
"step": 2598
},
{
"epoch": 0.15757575757575756,
"grad_norm": 0.06858024001121521,
"learning_rate": 0.0008,
"loss": 2.5097055435180664,
"step": 2600
},
{
"epoch": 0.1576969696969697,
"grad_norm": 0.0665673166513443,
"learning_rate": 0.0008,
"loss": 2.398186683654785,
"step": 2602
},
{
"epoch": 0.15781818181818183,
"grad_norm": 0.06999817490577698,
"learning_rate": 0.0008,
"loss": 2.4168081283569336,
"step": 2604
},
{
"epoch": 0.15793939393939394,
"grad_norm": 0.06844688951969147,
"learning_rate": 0.0008,
"loss": 2.438790798187256,
"step": 2606
},
{
"epoch": 0.15806060606060607,
"grad_norm": 0.06831025332212448,
"learning_rate": 0.0008,
"loss": 2.42495059967041,
"step": 2608
},
{
"epoch": 0.15818181818181817,
"grad_norm": 0.06954637169837952,
"learning_rate": 0.0008,
"loss": 2.457643508911133,
"step": 2610
},
{
"epoch": 0.1583030303030303,
"grad_norm": 0.06514674425125122,
"learning_rate": 0.0008,
"loss": 2.412978172302246,
"step": 2612
},
{
"epoch": 0.1584242424242424,
"grad_norm": 0.06806619465351105,
"learning_rate": 0.0008,
"loss": 2.4537808895111084,
"step": 2614
},
{
"epoch": 0.15854545454545454,
"grad_norm": 0.06631214916706085,
"learning_rate": 0.0008,
"loss": 2.3794069290161133,
"step": 2616
},
{
"epoch": 0.15866666666666668,
"grad_norm": 0.06817324459552765,
"learning_rate": 0.0008,
"loss": 2.409794807434082,
"step": 2618
},
{
"epoch": 0.15878787878787878,
"grad_norm": 0.07039792835712433,
"learning_rate": 0.0008,
"loss": 2.4840338230133057,
"step": 2620
},
{
"epoch": 0.15890909090909092,
"grad_norm": 0.07089892774820328,
"learning_rate": 0.0008,
"loss": 2.3720226287841797,
"step": 2622
},
{
"epoch": 0.15903030303030302,
"grad_norm": 0.06741703301668167,
"learning_rate": 0.0008,
"loss": 2.384026527404785,
"step": 2624
},
{
"epoch": 0.15915151515151515,
"grad_norm": 0.06524886190891266,
"learning_rate": 0.0008,
"loss": 2.4013168811798096,
"step": 2626
},
{
"epoch": 0.15927272727272726,
"grad_norm": 0.068948894739151,
"learning_rate": 0.0008,
"loss": 2.3487257957458496,
"step": 2628
},
{
"epoch": 0.1593939393939394,
"grad_norm": 0.06968192756175995,
"learning_rate": 0.0008,
"loss": 2.4308834075927734,
"step": 2630
},
{
"epoch": 0.15951515151515153,
"grad_norm": 0.07109712809324265,
"learning_rate": 0.0008,
"loss": 2.4372739791870117,
"step": 2632
},
{
"epoch": 0.15963636363636363,
"grad_norm": 0.06762351095676422,
"learning_rate": 0.0008,
"loss": 2.4803409576416016,
"step": 2634
},
{
"epoch": 0.15975757575757576,
"grad_norm": 0.0687972754240036,
"learning_rate": 0.0008,
"loss": 2.4845824241638184,
"step": 2636
},
{
"epoch": 0.15987878787878787,
"grad_norm": 0.06811901181936264,
"learning_rate": 0.0008,
"loss": 2.4968199729919434,
"step": 2638
},
{
"epoch": 0.16,
"grad_norm": 0.06784339249134064,
"learning_rate": 0.0008,
"loss": 2.4809930324554443,
"step": 2640
},
{
"epoch": 0.1601212121212121,
"grad_norm": 0.06575978547334671,
"learning_rate": 0.0008,
"loss": 2.4149250984191895,
"step": 2642
},
{
"epoch": 0.16024242424242424,
"grad_norm": 0.06628108024597168,
"learning_rate": 0.0008,
"loss": 2.4713969230651855,
"step": 2644
},
{
"epoch": 0.16036363636363637,
"grad_norm": 0.06582725793123245,
"learning_rate": 0.0008,
"loss": 2.435655117034912,
"step": 2646
},
{
"epoch": 0.16048484848484848,
"grad_norm": 0.06638836860656738,
"learning_rate": 0.0008,
"loss": 2.383162498474121,
"step": 2648
},
{
"epoch": 0.1606060606060606,
"grad_norm": 0.06683134287595749,
"learning_rate": 0.0008,
"loss": 2.405433177947998,
"step": 2650
},
{
"epoch": 0.16072727272727272,
"grad_norm": 0.06935524195432663,
"learning_rate": 0.0008,
"loss": 2.4789741039276123,
"step": 2652
},
{
"epoch": 0.16084848484848485,
"grad_norm": 0.06766800582408905,
"learning_rate": 0.0008,
"loss": 2.393683671951294,
"step": 2654
},
{
"epoch": 0.16096969696969696,
"grad_norm": 0.0706285908818245,
"learning_rate": 0.0008,
"loss": 2.415071487426758,
"step": 2656
},
{
"epoch": 0.1610909090909091,
"grad_norm": 0.06968796998262405,
"learning_rate": 0.0008,
"loss": 2.4460155963897705,
"step": 2658
},
{
"epoch": 0.16121212121212122,
"grad_norm": 0.06628693640232086,
"learning_rate": 0.0008,
"loss": 2.4601993560791016,
"step": 2660
},
{
"epoch": 0.16133333333333333,
"grad_norm": 0.06984694302082062,
"learning_rate": 0.0008,
"loss": 2.414761543273926,
"step": 2662
},
{
"epoch": 0.16145454545454546,
"grad_norm": 0.06819984316825867,
"learning_rate": 0.0008,
"loss": 2.4370851516723633,
"step": 2664
},
{
"epoch": 0.16157575757575757,
"grad_norm": 0.06843115389347076,
"learning_rate": 0.0008,
"loss": 2.4994258880615234,
"step": 2666
},
{
"epoch": 0.1616969696969697,
"grad_norm": 0.06630350649356842,
"learning_rate": 0.0008,
"loss": 2.4802801609039307,
"step": 2668
},
{
"epoch": 0.1618181818181818,
"grad_norm": 0.0668134018778801,
"learning_rate": 0.0008,
"loss": 2.365199089050293,
"step": 2670
},
{
"epoch": 0.16193939393939394,
"grad_norm": 0.06884768605232239,
"learning_rate": 0.0008,
"loss": 2.437446117401123,
"step": 2672
},
{
"epoch": 0.16206060606060607,
"grad_norm": 0.06900529563426971,
"learning_rate": 0.0008,
"loss": 2.528991222381592,
"step": 2674
},
{
"epoch": 0.16218181818181818,
"grad_norm": 0.06590142846107483,
"learning_rate": 0.0008,
"loss": 2.4282355308532715,
"step": 2676
},
{
"epoch": 0.1623030303030303,
"grad_norm": 0.06926210224628448,
"learning_rate": 0.0008,
"loss": 2.4806435108184814,
"step": 2678
},
{
"epoch": 0.16242424242424242,
"grad_norm": 0.06680769473314285,
"learning_rate": 0.0008,
"loss": 2.4718143939971924,
"step": 2680
},
{
"epoch": 0.16254545454545455,
"grad_norm": 0.06721188873052597,
"learning_rate": 0.0008,
"loss": 2.462372064590454,
"step": 2682
},
{
"epoch": 0.16266666666666665,
"grad_norm": 0.07184220105409622,
"learning_rate": 0.0008,
"loss": 2.4680075645446777,
"step": 2684
},
{
"epoch": 0.1627878787878788,
"grad_norm": 0.0698757991194725,
"learning_rate": 0.0008,
"loss": 2.560357093811035,
"step": 2686
},
{
"epoch": 0.16290909090909092,
"grad_norm": 0.07011677324771881,
"learning_rate": 0.0008,
"loss": 2.4097514152526855,
"step": 2688
},
{
"epoch": 0.16303030303030303,
"grad_norm": 0.0641055777668953,
"learning_rate": 0.0008,
"loss": 2.321916103363037,
"step": 2690
},
{
"epoch": 0.16315151515151516,
"grad_norm": 0.07526511698961258,
"learning_rate": 0.0008,
"loss": 2.458932399749756,
"step": 2692
},
{
"epoch": 0.16327272727272726,
"grad_norm": 0.07210209965705872,
"learning_rate": 0.0008,
"loss": 2.4629085063934326,
"step": 2694
},
{
"epoch": 0.1633939393939394,
"grad_norm": 0.08196789026260376,
"learning_rate": 0.0008,
"loss": 2.4669597148895264,
"step": 2696
},
{
"epoch": 0.1635151515151515,
"grad_norm": 0.07109034806489944,
"learning_rate": 0.0008,
"loss": 2.445246696472168,
"step": 2698
},
{
"epoch": 0.16363636363636364,
"grad_norm": 0.06589975208044052,
"learning_rate": 0.0008,
"loss": 2.386744260787964,
"step": 2700
},
{
"epoch": 0.16375757575757577,
"grad_norm": 0.06507444381713867,
"learning_rate": 0.0008,
"loss": 2.403676748275757,
"step": 2702
},
{
"epoch": 0.16387878787878787,
"grad_norm": 0.06939738988876343,
"learning_rate": 0.0008,
"loss": 2.3412413597106934,
"step": 2704
},
{
"epoch": 0.164,
"grad_norm": 0.06769145280122757,
"learning_rate": 0.0008,
"loss": 2.510087013244629,
"step": 2706
},
{
"epoch": 0.1641212121212121,
"grad_norm": 0.07131843268871307,
"learning_rate": 0.0008,
"loss": 2.4339752197265625,
"step": 2708
},
{
"epoch": 0.16424242424242425,
"grad_norm": 0.06944131851196289,
"learning_rate": 0.0008,
"loss": 2.4240288734436035,
"step": 2710
},
{
"epoch": 0.16436363636363635,
"grad_norm": 0.06872718781232834,
"learning_rate": 0.0008,
"loss": 2.4123001098632812,
"step": 2712
},
{
"epoch": 0.16448484848484848,
"grad_norm": 0.06868567317724228,
"learning_rate": 0.0008,
"loss": 2.393237590789795,
"step": 2714
},
{
"epoch": 0.16460606060606062,
"grad_norm": 0.06795956194400787,
"learning_rate": 0.0008,
"loss": 2.4679102897644043,
"step": 2716
},
{
"epoch": 0.16472727272727272,
"grad_norm": 0.06886272877454758,
"learning_rate": 0.0008,
"loss": 2.452667713165283,
"step": 2718
},
{
"epoch": 0.16484848484848486,
"grad_norm": 0.06851322203874588,
"learning_rate": 0.0008,
"loss": 2.470396041870117,
"step": 2720
},
{
"epoch": 0.16496969696969696,
"grad_norm": 0.07230063527822495,
"learning_rate": 0.0008,
"loss": 2.4776699542999268,
"step": 2722
},
{
"epoch": 0.1650909090909091,
"grad_norm": 0.06830786168575287,
"learning_rate": 0.0008,
"loss": 2.4418089389801025,
"step": 2724
},
{
"epoch": 0.1652121212121212,
"grad_norm": 0.07178404182195663,
"learning_rate": 0.0008,
"loss": 2.44338321685791,
"step": 2726
},
{
"epoch": 0.16533333333333333,
"grad_norm": 0.07089998573064804,
"learning_rate": 0.0008,
"loss": 2.4185640811920166,
"step": 2728
},
{
"epoch": 0.16545454545454547,
"grad_norm": 0.07188340276479721,
"learning_rate": 0.0008,
"loss": 2.4329147338867188,
"step": 2730
},
{
"epoch": 0.16557575757575757,
"grad_norm": 0.06913052499294281,
"learning_rate": 0.0008,
"loss": 2.4500579833984375,
"step": 2732
},
{
"epoch": 0.1656969696969697,
"grad_norm": 0.06571678072214127,
"learning_rate": 0.0008,
"loss": 2.4597060680389404,
"step": 2734
},
{
"epoch": 0.1658181818181818,
"grad_norm": 0.06630692631006241,
"learning_rate": 0.0008,
"loss": 2.5083377361297607,
"step": 2736
},
{
"epoch": 0.16593939393939394,
"grad_norm": 0.06588340550661087,
"learning_rate": 0.0008,
"loss": 2.402313232421875,
"step": 2738
},
{
"epoch": 0.16606060606060605,
"grad_norm": 0.06596973538398743,
"learning_rate": 0.0008,
"loss": 2.4195573329925537,
"step": 2740
},
{
"epoch": 0.16618181818181818,
"grad_norm": 0.06608486920595169,
"learning_rate": 0.0008,
"loss": 2.402195692062378,
"step": 2742
},
{
"epoch": 0.1663030303030303,
"grad_norm": 0.06636010855436325,
"learning_rate": 0.0008,
"loss": 2.4413065910339355,
"step": 2744
},
{
"epoch": 0.16642424242424242,
"grad_norm": 0.06706400960683823,
"learning_rate": 0.0008,
"loss": 2.338902235031128,
"step": 2746
},
{
"epoch": 0.16654545454545455,
"grad_norm": 0.06581086665391922,
"learning_rate": 0.0008,
"loss": 2.438366174697876,
"step": 2748
},
{
"epoch": 0.16666666666666666,
"grad_norm": 0.06734935194253922,
"learning_rate": 0.0008,
"loss": 2.402296304702759,
"step": 2750
},
{
"epoch": 0.1667878787878788,
"grad_norm": 0.06868048012256622,
"learning_rate": 0.0008,
"loss": 2.488098382949829,
"step": 2752
},
{
"epoch": 0.1669090909090909,
"grad_norm": 0.07135383784770966,
"learning_rate": 0.0008,
"loss": 2.4611079692840576,
"step": 2754
},
{
"epoch": 0.16703030303030303,
"grad_norm": 0.07041104137897491,
"learning_rate": 0.0008,
"loss": 2.38619327545166,
"step": 2756
},
{
"epoch": 0.16715151515151516,
"grad_norm": 0.06851822882890701,
"learning_rate": 0.0008,
"loss": 2.4367635250091553,
"step": 2758
},
{
"epoch": 0.16727272727272727,
"grad_norm": 0.06797768175601959,
"learning_rate": 0.0008,
"loss": 2.4093172550201416,
"step": 2760
},
{
"epoch": 0.1673939393939394,
"grad_norm": 0.06910546123981476,
"learning_rate": 0.0008,
"loss": 2.3963611125946045,
"step": 2762
},
{
"epoch": 0.1675151515151515,
"grad_norm": 0.07385605573654175,
"learning_rate": 0.0008,
"loss": 2.4435198307037354,
"step": 2764
},
{
"epoch": 0.16763636363636364,
"grad_norm": 0.08359528332948685,
"learning_rate": 0.0008,
"loss": 2.4756882190704346,
"step": 2766
},
{
"epoch": 0.16775757575757574,
"grad_norm": 0.0738753080368042,
"learning_rate": 0.0008,
"loss": 2.392726182937622,
"step": 2768
},
{
"epoch": 0.16787878787878788,
"grad_norm": 0.069158636033535,
"learning_rate": 0.0008,
"loss": 2.4304773807525635,
"step": 2770
},
{
"epoch": 0.168,
"grad_norm": 0.07120773196220398,
"learning_rate": 0.0008,
"loss": 2.4494550228118896,
"step": 2772
},
{
"epoch": 0.16812121212121212,
"grad_norm": 0.06754006445407867,
"learning_rate": 0.0008,
"loss": 2.428877830505371,
"step": 2774
},
{
"epoch": 0.16824242424242425,
"grad_norm": 0.06668662279844284,
"learning_rate": 0.0008,
"loss": 2.412306308746338,
"step": 2776
},
{
"epoch": 0.16836363636363635,
"grad_norm": 0.06821316480636597,
"learning_rate": 0.0008,
"loss": 2.441498279571533,
"step": 2778
},
{
"epoch": 0.1684848484848485,
"grad_norm": 0.06731689721345901,
"learning_rate": 0.0008,
"loss": 2.4642152786254883,
"step": 2780
},
{
"epoch": 0.1686060606060606,
"grad_norm": 0.06727799773216248,
"learning_rate": 0.0008,
"loss": 2.4763078689575195,
"step": 2782
},
{
"epoch": 0.16872727272727273,
"grad_norm": 0.06394544243812561,
"learning_rate": 0.0008,
"loss": 2.4704647064208984,
"step": 2784
},
{
"epoch": 0.16884848484848486,
"grad_norm": 0.06458068639039993,
"learning_rate": 0.0008,
"loss": 2.462096929550171,
"step": 2786
},
{
"epoch": 0.16896969696969696,
"grad_norm": 0.06515559554100037,
"learning_rate": 0.0008,
"loss": 2.421018362045288,
"step": 2788
},
{
"epoch": 0.1690909090909091,
"grad_norm": 0.06447729468345642,
"learning_rate": 0.0008,
"loss": 2.410299301147461,
"step": 2790
},
{
"epoch": 0.1692121212121212,
"grad_norm": 0.06744813919067383,
"learning_rate": 0.0008,
"loss": 2.4122607707977295,
"step": 2792
},
{
"epoch": 0.16933333333333334,
"grad_norm": 0.06712717562913895,
"learning_rate": 0.0008,
"loss": 2.369622230529785,
"step": 2794
},
{
"epoch": 0.16945454545454544,
"grad_norm": 0.06601685285568237,
"learning_rate": 0.0008,
"loss": 2.4234049320220947,
"step": 2796
},
{
"epoch": 0.16957575757575757,
"grad_norm": 0.06461860984563828,
"learning_rate": 0.0008,
"loss": 2.49368953704834,
"step": 2798
},
{
"epoch": 0.1696969696969697,
"grad_norm": 0.06852491199970245,
"learning_rate": 0.0008,
"loss": 2.384535789489746,
"step": 2800
},
{
"epoch": 0.1698181818181818,
"grad_norm": 0.07093381136655807,
"learning_rate": 0.0008,
"loss": 2.4853978157043457,
"step": 2802
},
{
"epoch": 0.16993939393939395,
"grad_norm": 0.06846247613430023,
"learning_rate": 0.0008,
"loss": 2.4758520126342773,
"step": 2804
},
{
"epoch": 0.17006060606060605,
"grad_norm": 0.06834618002176285,
"learning_rate": 0.0008,
"loss": 2.369992971420288,
"step": 2806
},
{
"epoch": 0.17018181818181818,
"grad_norm": 0.06877147406339645,
"learning_rate": 0.0008,
"loss": 2.405559778213501,
"step": 2808
},
{
"epoch": 0.1703030303030303,
"grad_norm": 0.06869377195835114,
"learning_rate": 0.0008,
"loss": 2.4338197708129883,
"step": 2810
},
{
"epoch": 0.17042424242424242,
"grad_norm": 0.06521221995353699,
"learning_rate": 0.0008,
"loss": 2.4598515033721924,
"step": 2812
},
{
"epoch": 0.17054545454545456,
"grad_norm": 0.06600851565599442,
"learning_rate": 0.0008,
"loss": 2.501908779144287,
"step": 2814
},
{
"epoch": 0.17066666666666666,
"grad_norm": 0.06640349328517914,
"learning_rate": 0.0008,
"loss": 2.4503116607666016,
"step": 2816
},
{
"epoch": 0.1707878787878788,
"grad_norm": 0.06792350113391876,
"learning_rate": 0.0008,
"loss": 2.4188594818115234,
"step": 2818
},
{
"epoch": 0.1709090909090909,
"grad_norm": 0.06706385314464569,
"learning_rate": 0.0008,
"loss": 2.522097110748291,
"step": 2820
},
{
"epoch": 0.17103030303030303,
"grad_norm": 0.06872203201055527,
"learning_rate": 0.0008,
"loss": 2.4838626384735107,
"step": 2822
},
{
"epoch": 0.17115151515151514,
"grad_norm": 0.06741680949926376,
"learning_rate": 0.0008,
"loss": 2.4176149368286133,
"step": 2824
},
{
"epoch": 0.17127272727272727,
"grad_norm": 0.06602033972740173,
"learning_rate": 0.0008,
"loss": 2.4836535453796387,
"step": 2826
},
{
"epoch": 0.1713939393939394,
"grad_norm": 0.06687402725219727,
"learning_rate": 0.0008,
"loss": 2.51054048538208,
"step": 2828
},
{
"epoch": 0.1715151515151515,
"grad_norm": 0.06729398667812347,
"learning_rate": 0.0008,
"loss": 2.5187313556671143,
"step": 2830
},
{
"epoch": 0.17163636363636364,
"grad_norm": 0.06923633068799973,
"learning_rate": 0.0008,
"loss": 2.4335169792175293,
"step": 2832
},
{
"epoch": 0.17175757575757575,
"grad_norm": 0.07403712719678879,
"learning_rate": 0.0008,
"loss": 2.476203203201294,
"step": 2834
},
{
"epoch": 0.17187878787878788,
"grad_norm": 0.06778639554977417,
"learning_rate": 0.0008,
"loss": 2.423624038696289,
"step": 2836
},
{
"epoch": 0.172,
"grad_norm": 0.06483469903469086,
"learning_rate": 0.0008,
"loss": 2.384229898452759,
"step": 2838
},
{
"epoch": 0.17212121212121212,
"grad_norm": 0.06469687074422836,
"learning_rate": 0.0008,
"loss": 2.4767961502075195,
"step": 2840
},
{
"epoch": 0.17224242424242425,
"grad_norm": 0.06500885635614395,
"learning_rate": 0.0008,
"loss": 2.3612453937530518,
"step": 2842
},
{
"epoch": 0.17236363636363636,
"grad_norm": 0.06984347850084305,
"learning_rate": 0.0008,
"loss": 2.391219139099121,
"step": 2844
},
{
"epoch": 0.1724848484848485,
"grad_norm": 0.0681329146027565,
"learning_rate": 0.0008,
"loss": 2.4565343856811523,
"step": 2846
},
{
"epoch": 0.1726060606060606,
"grad_norm": 0.06527169048786163,
"learning_rate": 0.0008,
"loss": 2.406852960586548,
"step": 2848
},
{
"epoch": 0.17272727272727273,
"grad_norm": 0.06671646982431412,
"learning_rate": 0.0008,
"loss": 2.438042163848877,
"step": 2850
},
{
"epoch": 0.17284848484848483,
"grad_norm": 0.0655691847205162,
"learning_rate": 0.0008,
"loss": 2.4252634048461914,
"step": 2852
},
{
"epoch": 0.17296969696969697,
"grad_norm": 0.06661248207092285,
"learning_rate": 0.0008,
"loss": 2.4958114624023438,
"step": 2854
},
{
"epoch": 0.1730909090909091,
"grad_norm": 0.0653623566031456,
"learning_rate": 0.0008,
"loss": 2.416302442550659,
"step": 2856
},
{
"epoch": 0.1732121212121212,
"grad_norm": 0.06340077519416809,
"learning_rate": 0.0008,
"loss": 2.4088048934936523,
"step": 2858
},
{
"epoch": 0.17333333333333334,
"grad_norm": 0.06746827811002731,
"learning_rate": 0.0008,
"loss": 2.3953733444213867,
"step": 2860
},
{
"epoch": 0.17345454545454544,
"grad_norm": 0.06604038178920746,
"learning_rate": 0.0008,
"loss": 2.4236557483673096,
"step": 2862
},
{
"epoch": 0.17357575757575758,
"grad_norm": 0.0648961141705513,
"learning_rate": 0.0008,
"loss": 2.407761335372925,
"step": 2864
},
{
"epoch": 0.1736969696969697,
"grad_norm": 0.06630862504243851,
"learning_rate": 0.0008,
"loss": 2.490525245666504,
"step": 2866
},
{
"epoch": 0.17381818181818182,
"grad_norm": 0.06421444565057755,
"learning_rate": 0.0008,
"loss": 2.3863799571990967,
"step": 2868
},
{
"epoch": 0.17393939393939395,
"grad_norm": 0.06982718408107758,
"learning_rate": 0.0008,
"loss": 2.4841771125793457,
"step": 2870
},
{
"epoch": 0.17406060606060605,
"grad_norm": 0.073214590549469,
"learning_rate": 0.0008,
"loss": 2.41642689704895,
"step": 2872
},
{
"epoch": 0.1741818181818182,
"grad_norm": 0.07575412094593048,
"learning_rate": 0.0008,
"loss": 2.4530348777770996,
"step": 2874
},
{
"epoch": 0.1743030303030303,
"grad_norm": 0.07427642494440079,
"learning_rate": 0.0008,
"loss": 2.371011972427368,
"step": 2876
},
{
"epoch": 0.17442424242424243,
"grad_norm": 0.06866831332445145,
"learning_rate": 0.0008,
"loss": 2.367356777191162,
"step": 2878
},
{
"epoch": 0.17454545454545456,
"grad_norm": 0.07062719017267227,
"learning_rate": 0.0008,
"loss": 2.4596896171569824,
"step": 2880
},
{
"epoch": 0.17466666666666666,
"grad_norm": 0.06616159528493881,
"learning_rate": 0.0008,
"loss": 2.5020017623901367,
"step": 2882
},
{
"epoch": 0.1747878787878788,
"grad_norm": 0.06751659512519836,
"learning_rate": 0.0008,
"loss": 2.457122325897217,
"step": 2884
},
{
"epoch": 0.1749090909090909,
"grad_norm": 0.06497512012720108,
"learning_rate": 0.0008,
"loss": 2.3616116046905518,
"step": 2886
},
{
"epoch": 0.17503030303030304,
"grad_norm": 0.06543174386024475,
"learning_rate": 0.0008,
"loss": 2.403066873550415,
"step": 2888
},
{
"epoch": 0.17515151515151514,
"grad_norm": 0.06579809635877609,
"learning_rate": 0.0008,
"loss": 2.4318294525146484,
"step": 2890
},
{
"epoch": 0.17527272727272727,
"grad_norm": 0.06662353128194809,
"learning_rate": 0.0008,
"loss": 2.4620985984802246,
"step": 2892
},
{
"epoch": 0.1753939393939394,
"grad_norm": 0.0651368722319603,
"learning_rate": 0.0008,
"loss": 2.42337965965271,
"step": 2894
},
{
"epoch": 0.1755151515151515,
"grad_norm": 0.07169242948293686,
"learning_rate": 0.0008,
"loss": 2.4599790573120117,
"step": 2896
},
{
"epoch": 0.17563636363636365,
"grad_norm": 0.0661090761423111,
"learning_rate": 0.0008,
"loss": 2.3978211879730225,
"step": 2898
},
{
"epoch": 0.17575757575757575,
"grad_norm": 0.06304091960191727,
"learning_rate": 0.0008,
"loss": 2.4329137802124023,
"step": 2900
},
{
"epoch": 0.17587878787878788,
"grad_norm": 0.0660262331366539,
"learning_rate": 0.0008,
"loss": 2.407559394836426,
"step": 2902
},
{
"epoch": 0.176,
"grad_norm": 0.06714970618486404,
"learning_rate": 0.0008,
"loss": 2.4342191219329834,
"step": 2904
},
{
"epoch": 0.17612121212121212,
"grad_norm": 0.06560352444648743,
"learning_rate": 0.0008,
"loss": 2.476201057434082,
"step": 2906
},
{
"epoch": 0.17624242424242426,
"grad_norm": 0.06390221416950226,
"learning_rate": 0.0008,
"loss": 2.3746371269226074,
"step": 2908
},
{
"epoch": 0.17636363636363636,
"grad_norm": 0.07328922301530838,
"learning_rate": 0.0008,
"loss": 2.423997402191162,
"step": 2910
},
{
"epoch": 0.1764848484848485,
"grad_norm": 0.068475641310215,
"learning_rate": 0.0008,
"loss": 2.509477376937866,
"step": 2912
},
{
"epoch": 0.1766060606060606,
"grad_norm": 0.06490171700716019,
"learning_rate": 0.0008,
"loss": 2.41855525970459,
"step": 2914
},
{
"epoch": 0.17672727272727273,
"grad_norm": 0.06773794442415237,
"learning_rate": 0.0008,
"loss": 2.401306629180908,
"step": 2916
},
{
"epoch": 0.17684848484848484,
"grad_norm": 0.06661373376846313,
"learning_rate": 0.0008,
"loss": 2.487328052520752,
"step": 2918
},
{
"epoch": 0.17696969696969697,
"grad_norm": 0.06855352967977524,
"learning_rate": 0.0008,
"loss": 2.494643449783325,
"step": 2920
},
{
"epoch": 0.1770909090909091,
"grad_norm": 0.06464157998561859,
"learning_rate": 0.0008,
"loss": 2.3970301151275635,
"step": 2922
},
{
"epoch": 0.1772121212121212,
"grad_norm": 0.06422387808561325,
"learning_rate": 0.0008,
"loss": 2.3771893978118896,
"step": 2924
},
{
"epoch": 0.17733333333333334,
"grad_norm": 0.06461814045906067,
"learning_rate": 0.0008,
"loss": 2.4436707496643066,
"step": 2926
},
{
"epoch": 0.17745454545454545,
"grad_norm": 0.06526704132556915,
"learning_rate": 0.0008,
"loss": 2.456667423248291,
"step": 2928
},
{
"epoch": 0.17757575757575758,
"grad_norm": 0.06408797949552536,
"learning_rate": 0.0008,
"loss": 2.4842896461486816,
"step": 2930
},
{
"epoch": 0.1776969696969697,
"grad_norm": 0.06361964344978333,
"learning_rate": 0.0008,
"loss": 2.4171180725097656,
"step": 2932
},
{
"epoch": 0.17781818181818182,
"grad_norm": 0.06551734358072281,
"learning_rate": 0.0008,
"loss": 2.400585174560547,
"step": 2934
},
{
"epoch": 0.17793939393939395,
"grad_norm": 0.06339745223522186,
"learning_rate": 0.0008,
"loss": 2.3861563205718994,
"step": 2936
},
{
"epoch": 0.17806060606060606,
"grad_norm": 0.06573393940925598,
"learning_rate": 0.0008,
"loss": 2.47640323638916,
"step": 2938
},
{
"epoch": 0.1781818181818182,
"grad_norm": 0.06413492560386658,
"learning_rate": 0.0008,
"loss": 2.30409836769104,
"step": 2940
},
{
"epoch": 0.1783030303030303,
"grad_norm": 0.0660824105143547,
"learning_rate": 0.0008,
"loss": 2.4268956184387207,
"step": 2942
},
{
"epoch": 0.17842424242424243,
"grad_norm": 0.06504757702350616,
"learning_rate": 0.0008,
"loss": 2.4115474224090576,
"step": 2944
},
{
"epoch": 0.17854545454545453,
"grad_norm": 0.06677333265542984,
"learning_rate": 0.0008,
"loss": 2.417860507965088,
"step": 2946
},
{
"epoch": 0.17866666666666667,
"grad_norm": 0.06536731123924255,
"learning_rate": 0.0008,
"loss": 2.4522511959075928,
"step": 2948
},
{
"epoch": 0.1787878787878788,
"grad_norm": 0.06371709704399109,
"learning_rate": 0.0008,
"loss": 2.3697257041931152,
"step": 2950
},
{
"epoch": 0.1789090909090909,
"grad_norm": 0.06604235619306564,
"learning_rate": 0.0008,
"loss": 2.4675302505493164,
"step": 2952
},
{
"epoch": 0.17903030303030304,
"grad_norm": 0.0691012442111969,
"learning_rate": 0.0008,
"loss": 2.4308841228485107,
"step": 2954
},
{
"epoch": 0.17915151515151514,
"grad_norm": 0.06627717614173889,
"learning_rate": 0.0008,
"loss": 2.386723279953003,
"step": 2956
},
{
"epoch": 0.17927272727272728,
"grad_norm": 0.06675651669502258,
"learning_rate": 0.0008,
"loss": 2.488985300064087,
"step": 2958
},
{
"epoch": 0.17939393939393938,
"grad_norm": 0.06861551105976105,
"learning_rate": 0.0008,
"loss": 2.503845453262329,
"step": 2960
},
{
"epoch": 0.17951515151515152,
"grad_norm": 0.06734225153923035,
"learning_rate": 0.0008,
"loss": 2.4475252628326416,
"step": 2962
},
{
"epoch": 0.17963636363636365,
"grad_norm": 0.06831453740596771,
"learning_rate": 0.0008,
"loss": 2.3445730209350586,
"step": 2964
},
{
"epoch": 0.17975757575757575,
"grad_norm": 0.06280075758695602,
"learning_rate": 0.0008,
"loss": 2.384721040725708,
"step": 2966
},
{
"epoch": 0.1798787878787879,
"grad_norm": 0.06443464010953903,
"learning_rate": 0.0008,
"loss": 2.402550458908081,
"step": 2968
},
{
"epoch": 0.18,
"grad_norm": 0.063672736287117,
"learning_rate": 0.0008,
"loss": 2.3694026470184326,
"step": 2970
},
{
"epoch": 0.18012121212121213,
"grad_norm": 0.06396728754043579,
"learning_rate": 0.0008,
"loss": 2.3973429203033447,
"step": 2972
},
{
"epoch": 0.18024242424242423,
"grad_norm": 0.0655377209186554,
"learning_rate": 0.0008,
"loss": 2.359654664993286,
"step": 2974
},
{
"epoch": 0.18036363636363636,
"grad_norm": 0.06725354492664337,
"learning_rate": 0.0008,
"loss": 2.4303712844848633,
"step": 2976
},
{
"epoch": 0.1804848484848485,
"grad_norm": 0.06575658917427063,
"learning_rate": 0.0008,
"loss": 2.356788158416748,
"step": 2978
},
{
"epoch": 0.1806060606060606,
"grad_norm": 0.06369442492723465,
"learning_rate": 0.0008,
"loss": 2.4433491230010986,
"step": 2980
},
{
"epoch": 0.18072727272727274,
"grad_norm": 0.06463176012039185,
"learning_rate": 0.0008,
"loss": 2.408127546310425,
"step": 2982
},
{
"epoch": 0.18084848484848484,
"grad_norm": 0.06508967280387878,
"learning_rate": 0.0008,
"loss": 2.4515345096588135,
"step": 2984
},
{
"epoch": 0.18096969696969697,
"grad_norm": 0.07074420154094696,
"learning_rate": 0.0008,
"loss": 2.3364691734313965,
"step": 2986
},
{
"epoch": 0.18109090909090908,
"grad_norm": 0.06833551079034805,
"learning_rate": 0.0008,
"loss": 2.4836556911468506,
"step": 2988
},
{
"epoch": 0.1812121212121212,
"grad_norm": 0.06307815760374069,
"learning_rate": 0.0008,
"loss": 2.4405529499053955,
"step": 2990
},
{
"epoch": 0.18133333333333335,
"grad_norm": 0.06525837630033493,
"learning_rate": 0.0008,
"loss": 2.4133706092834473,
"step": 2992
},
{
"epoch": 0.18145454545454545,
"grad_norm": 0.06315401196479797,
"learning_rate": 0.0008,
"loss": 2.4919486045837402,
"step": 2994
},
{
"epoch": 0.18157575757575758,
"grad_norm": 0.06396299600601196,
"learning_rate": 0.0008,
"loss": 2.438594102859497,
"step": 2996
},
{
"epoch": 0.1816969696969697,
"grad_norm": 0.06497132033109665,
"learning_rate": 0.0008,
"loss": 2.3589866161346436,
"step": 2998
},
{
"epoch": 0.18181818181818182,
"grad_norm": 0.063089519739151,
"learning_rate": 0.0008,
"loss": 2.3556575775146484,
"step": 3000
},
{
"epoch": 0.18193939393939393,
"grad_norm": 0.061228327453136444,
"learning_rate": 0.0008,
"loss": 2.3421101570129395,
"step": 3002
},
{
"epoch": 0.18206060606060606,
"grad_norm": 0.06395716965198517,
"learning_rate": 0.0008,
"loss": 2.4362633228302,
"step": 3004
},
{
"epoch": 0.1821818181818182,
"grad_norm": 0.0630573257803917,
"learning_rate": 0.0008,
"loss": 2.3572885990142822,
"step": 3006
},
{
"epoch": 0.1823030303030303,
"grad_norm": 0.06437784433364868,
"learning_rate": 0.0008,
"loss": 2.44697904586792,
"step": 3008
},
{
"epoch": 0.18242424242424243,
"grad_norm": 0.06221817806363106,
"learning_rate": 0.0008,
"loss": 2.3458499908447266,
"step": 3010
},
{
"epoch": 0.18254545454545454,
"grad_norm": 0.06503035128116608,
"learning_rate": 0.0008,
"loss": 2.35506534576416,
"step": 3012
},
{
"epoch": 0.18266666666666667,
"grad_norm": 0.06773760914802551,
"learning_rate": 0.0008,
"loss": 2.3458101749420166,
"step": 3014
},
{
"epoch": 0.18278787878787878,
"grad_norm": 0.06556382775306702,
"learning_rate": 0.0008,
"loss": 2.3702638149261475,
"step": 3016
},
{
"epoch": 0.1829090909090909,
"grad_norm": 0.06365402787923813,
"learning_rate": 0.0008,
"loss": 2.413487672805786,
"step": 3018
},
{
"epoch": 0.18303030303030304,
"grad_norm": 0.06358568370342255,
"learning_rate": 0.0008,
"loss": 2.3654000759124756,
"step": 3020
},
{
"epoch": 0.18315151515151515,
"grad_norm": 0.06623338907957077,
"learning_rate": 0.0008,
"loss": 2.4864466190338135,
"step": 3022
},
{
"epoch": 0.18327272727272728,
"grad_norm": 0.0661327987909317,
"learning_rate": 0.0008,
"loss": 2.388334274291992,
"step": 3024
},
{
"epoch": 0.1833939393939394,
"grad_norm": 0.06239744648337364,
"learning_rate": 0.0008,
"loss": 2.358694553375244,
"step": 3026
},
{
"epoch": 0.18351515151515152,
"grad_norm": 0.06309657543897629,
"learning_rate": 0.0008,
"loss": 2.4206435680389404,
"step": 3028
},
{
"epoch": 0.18363636363636363,
"grad_norm": 0.06468383967876434,
"learning_rate": 0.0008,
"loss": 2.4410994052886963,
"step": 3030
},
{
"epoch": 0.18375757575757576,
"grad_norm": 0.06330473721027374,
"learning_rate": 0.0008,
"loss": 2.386134386062622,
"step": 3032
},
{
"epoch": 0.1838787878787879,
"grad_norm": 0.06338783353567123,
"learning_rate": 0.0008,
"loss": 2.417536973953247,
"step": 3034
},
{
"epoch": 0.184,
"grad_norm": 0.06795301288366318,
"learning_rate": 0.0008,
"loss": 2.3483822345733643,
"step": 3036
},
{
"epoch": 0.18412121212121213,
"grad_norm": 0.0693645253777504,
"learning_rate": 0.0008,
"loss": 2.430048704147339,
"step": 3038
},
{
"epoch": 0.18424242424242424,
"grad_norm": 0.0688619464635849,
"learning_rate": 0.0008,
"loss": 2.3708560466766357,
"step": 3040
},
{
"epoch": 0.18436363636363637,
"grad_norm": 0.06377632170915604,
"learning_rate": 0.0008,
"loss": 2.3716514110565186,
"step": 3042
},
{
"epoch": 0.18448484848484847,
"grad_norm": 0.0634876936674118,
"learning_rate": 0.0008,
"loss": 2.4300529956817627,
"step": 3044
},
{
"epoch": 0.1846060606060606,
"grad_norm": 0.06335747241973877,
"learning_rate": 0.0008,
"loss": 2.4186036586761475,
"step": 3046
},
{
"epoch": 0.18472727272727274,
"grad_norm": 0.06232693791389465,
"learning_rate": 0.0008,
"loss": 2.3350799083709717,
"step": 3048
},
{
"epoch": 0.18484848484848485,
"grad_norm": 0.06333788484334946,
"learning_rate": 0.0008,
"loss": 2.3857357501983643,
"step": 3050
},
{
"epoch": 0.18496969696969698,
"grad_norm": 0.06505078822374344,
"learning_rate": 0.0008,
"loss": 2.3822238445281982,
"step": 3052
},
{
"epoch": 0.18509090909090908,
"grad_norm": 0.0699058249592781,
"learning_rate": 0.0008,
"loss": 2.430258274078369,
"step": 3054
},
{
"epoch": 0.18521212121212122,
"grad_norm": 0.07023857533931732,
"learning_rate": 0.0008,
"loss": 2.4610235691070557,
"step": 3056
},
{
"epoch": 0.18533333333333332,
"grad_norm": 0.07260244339704514,
"learning_rate": 0.0008,
"loss": 2.3353121280670166,
"step": 3058
},
{
"epoch": 0.18545454545454546,
"grad_norm": 0.06458577513694763,
"learning_rate": 0.0008,
"loss": 2.446903705596924,
"step": 3060
},
{
"epoch": 0.1855757575757576,
"grad_norm": 0.06479867547750473,
"learning_rate": 0.0008,
"loss": 2.332981824874878,
"step": 3062
},
{
"epoch": 0.1856969696969697,
"grad_norm": 0.0633944720029831,
"learning_rate": 0.0008,
"loss": 2.4449079036712646,
"step": 3064
},
{
"epoch": 0.18581818181818183,
"grad_norm": 0.06309222429990768,
"learning_rate": 0.0008,
"loss": 2.335327386856079,
"step": 3066
},
{
"epoch": 0.18593939393939393,
"grad_norm": 0.06408216059207916,
"learning_rate": 0.0008,
"loss": 2.4356696605682373,
"step": 3068
},
{
"epoch": 0.18606060606060607,
"grad_norm": 0.06471548974514008,
"learning_rate": 0.0008,
"loss": 2.444305419921875,
"step": 3070
},
{
"epoch": 0.18618181818181817,
"grad_norm": 0.06254447996616364,
"learning_rate": 0.0008,
"loss": 2.3471405506134033,
"step": 3072
},
{
"epoch": 0.1863030303030303,
"grad_norm": 0.06238701194524765,
"learning_rate": 0.0008,
"loss": 2.4049932956695557,
"step": 3074
},
{
"epoch": 0.18642424242424244,
"grad_norm": 0.06601325422525406,
"learning_rate": 0.0008,
"loss": 2.346419095993042,
"step": 3076
},
{
"epoch": 0.18654545454545454,
"grad_norm": 0.06371774524450302,
"learning_rate": 0.0008,
"loss": 2.4024994373321533,
"step": 3078
},
{
"epoch": 0.18666666666666668,
"grad_norm": 0.06827075779438019,
"learning_rate": 0.0008,
"loss": 2.3420512676239014,
"step": 3080
},
{
"epoch": 0.18678787878787878,
"grad_norm": 0.06593073159456253,
"learning_rate": 0.0008,
"loss": 2.381075382232666,
"step": 3082
},
{
"epoch": 0.1869090909090909,
"grad_norm": 0.06325356662273407,
"learning_rate": 0.0008,
"loss": 2.3802430629730225,
"step": 3084
},
{
"epoch": 0.18703030303030302,
"grad_norm": 0.06378749012947083,
"learning_rate": 0.0008,
"loss": 2.391240119934082,
"step": 3086
},
{
"epoch": 0.18715151515151515,
"grad_norm": 0.06422140449285507,
"learning_rate": 0.0008,
"loss": 2.3813891410827637,
"step": 3088
},
{
"epoch": 0.18727272727272729,
"grad_norm": 0.06184036284685135,
"learning_rate": 0.0008,
"loss": 2.3343183994293213,
"step": 3090
},
{
"epoch": 0.1873939393939394,
"grad_norm": 0.06425876170396805,
"learning_rate": 0.0008,
"loss": 2.4049947261810303,
"step": 3092
},
{
"epoch": 0.18751515151515152,
"grad_norm": 0.06367700546979904,
"learning_rate": 0.0008,
"loss": 2.4365158081054688,
"step": 3094
},
{
"epoch": 0.18763636363636363,
"grad_norm": 0.06330964714288712,
"learning_rate": 0.0008,
"loss": 2.3740439414978027,
"step": 3096
},
{
"epoch": 0.18775757575757576,
"grad_norm": 0.06318091601133347,
"learning_rate": 0.0008,
"loss": 2.3879053592681885,
"step": 3098
},
{
"epoch": 0.18787878787878787,
"grad_norm": 0.06569522619247437,
"learning_rate": 0.0008,
"loss": 2.3313913345336914,
"step": 3100
},
{
"epoch": 0.188,
"grad_norm": 0.06474056094884872,
"learning_rate": 0.0008,
"loss": 2.3862361907958984,
"step": 3102
},
{
"epoch": 0.18812121212121213,
"grad_norm": 0.06301421672105789,
"learning_rate": 0.0008,
"loss": 2.4065377712249756,
"step": 3104
},
{
"epoch": 0.18824242424242424,
"grad_norm": 0.06734290719032288,
"learning_rate": 0.0008,
"loss": 2.4801526069641113,
"step": 3106
},
{
"epoch": 0.18836363636363637,
"grad_norm": 0.07240349054336548,
"learning_rate": 0.0008,
"loss": 2.3244693279266357,
"step": 3108
},
{
"epoch": 0.18848484848484848,
"grad_norm": 0.07153523713350296,
"learning_rate": 0.0008,
"loss": 2.3961730003356934,
"step": 3110
},
{
"epoch": 0.1886060606060606,
"grad_norm": 0.07790029793977737,
"learning_rate": 0.0008,
"loss": 2.381458044052124,
"step": 3112
},
{
"epoch": 0.18872727272727272,
"grad_norm": 0.06602295488119125,
"learning_rate": 0.0008,
"loss": 2.4118356704711914,
"step": 3114
},
{
"epoch": 0.18884848484848485,
"grad_norm": 0.06292525678873062,
"learning_rate": 0.0008,
"loss": 2.3827216625213623,
"step": 3116
},
{
"epoch": 0.18896969696969698,
"grad_norm": 0.06467125564813614,
"learning_rate": 0.0008,
"loss": 2.414414882659912,
"step": 3118
},
{
"epoch": 0.1890909090909091,
"grad_norm": 0.06397843360900879,
"learning_rate": 0.0008,
"loss": 2.422454595565796,
"step": 3120
},
{
"epoch": 0.18921212121212122,
"grad_norm": 0.06532645225524902,
"learning_rate": 0.0008,
"loss": 2.4362432956695557,
"step": 3122
},
{
"epoch": 0.18933333333333333,
"grad_norm": 0.06214458867907524,
"learning_rate": 0.0008,
"loss": 2.3305087089538574,
"step": 3124
},
{
"epoch": 0.18945454545454546,
"grad_norm": 0.06185382232069969,
"learning_rate": 0.0008,
"loss": 2.355097770690918,
"step": 3126
},
{
"epoch": 0.18957575757575756,
"grad_norm": 0.06328699737787247,
"learning_rate": 0.0008,
"loss": 2.385023355484009,
"step": 3128
},
{
"epoch": 0.1896969696969697,
"grad_norm": 0.06175168976187706,
"learning_rate": 0.0008,
"loss": 2.375349998474121,
"step": 3130
},
{
"epoch": 0.18981818181818183,
"grad_norm": 0.061164215207099915,
"learning_rate": 0.0008,
"loss": 2.3967881202697754,
"step": 3132
},
{
"epoch": 0.18993939393939394,
"grad_norm": 0.06273940205574036,
"learning_rate": 0.0008,
"loss": 2.4550909996032715,
"step": 3134
},
{
"epoch": 0.19006060606060607,
"grad_norm": 0.06209006533026695,
"learning_rate": 0.0008,
"loss": 2.3875529766082764,
"step": 3136
},
{
"epoch": 0.19018181818181817,
"grad_norm": 0.062385592609643936,
"learning_rate": 0.0008,
"loss": 2.4529519081115723,
"step": 3138
},
{
"epoch": 0.1903030303030303,
"grad_norm": 0.060295820236206055,
"learning_rate": 0.0008,
"loss": 2.3687946796417236,
"step": 3140
},
{
"epoch": 0.1904242424242424,
"grad_norm": 0.06176496669650078,
"learning_rate": 0.0008,
"loss": 2.410369634628296,
"step": 3142
},
{
"epoch": 0.19054545454545455,
"grad_norm": 0.062257070094347,
"learning_rate": 0.0008,
"loss": 2.4153847694396973,
"step": 3144
},
{
"epoch": 0.19066666666666668,
"grad_norm": 0.06096712499856949,
"learning_rate": 0.0008,
"loss": 2.352400541305542,
"step": 3146
},
{
"epoch": 0.19078787878787878,
"grad_norm": 0.06203533336520195,
"learning_rate": 0.0008,
"loss": 2.419064521789551,
"step": 3148
},
{
"epoch": 0.19090909090909092,
"grad_norm": 0.06392648071050644,
"learning_rate": 0.0008,
"loss": 2.4934678077697754,
"step": 3150
},
{
"epoch": 0.19103030303030302,
"grad_norm": 0.06238826736807823,
"learning_rate": 0.0008,
"loss": 2.3837730884552,
"step": 3152
},
{
"epoch": 0.19115151515151516,
"grad_norm": 0.06620901077985764,
"learning_rate": 0.0008,
"loss": 2.3171088695526123,
"step": 3154
},
{
"epoch": 0.19127272727272726,
"grad_norm": 0.06229936331510544,
"learning_rate": 0.0008,
"loss": 2.400473117828369,
"step": 3156
},
{
"epoch": 0.1913939393939394,
"grad_norm": 0.0609317272901535,
"learning_rate": 0.0008,
"loss": 2.4552955627441406,
"step": 3158
},
{
"epoch": 0.19151515151515153,
"grad_norm": 0.06398898363113403,
"learning_rate": 0.0008,
"loss": 2.3514933586120605,
"step": 3160
},
{
"epoch": 0.19163636363636363,
"grad_norm": 0.06423084437847137,
"learning_rate": 0.0008,
"loss": 2.3731534481048584,
"step": 3162
},
{
"epoch": 0.19175757575757577,
"grad_norm": 0.06352158635854721,
"learning_rate": 0.0008,
"loss": 2.409132719039917,
"step": 3164
},
{
"epoch": 0.19187878787878787,
"grad_norm": 0.06730259209871292,
"learning_rate": 0.0008,
"loss": 2.4403772354125977,
"step": 3166
},
{
"epoch": 0.192,
"grad_norm": 0.06986167281866074,
"learning_rate": 0.0008,
"loss": 2.3875160217285156,
"step": 3168
},
{
"epoch": 0.1921212121212121,
"grad_norm": 0.06306920200586319,
"learning_rate": 0.0008,
"loss": 2.3901114463806152,
"step": 3170
},
{
"epoch": 0.19224242424242424,
"grad_norm": 0.06700063496828079,
"learning_rate": 0.0008,
"loss": 2.3452043533325195,
"step": 3172
},
{
"epoch": 0.19236363636363638,
"grad_norm": 0.06336411833763123,
"learning_rate": 0.0008,
"loss": 2.359001398086548,
"step": 3174
},
{
"epoch": 0.19248484848484848,
"grad_norm": 0.061670079827308655,
"learning_rate": 0.0008,
"loss": 2.4114975929260254,
"step": 3176
},
{
"epoch": 0.19260606060606061,
"grad_norm": 0.061906442046165466,
"learning_rate": 0.0008,
"loss": 2.3716981410980225,
"step": 3178
},
{
"epoch": 0.19272727272727272,
"grad_norm": 0.06208883970975876,
"learning_rate": 0.0008,
"loss": 2.424487829208374,
"step": 3180
},
{
"epoch": 0.19284848484848485,
"grad_norm": 0.06350213289260864,
"learning_rate": 0.0008,
"loss": 2.3474700450897217,
"step": 3182
},
{
"epoch": 0.19296969696969696,
"grad_norm": 0.06186451017856598,
"learning_rate": 0.0008,
"loss": 2.406841278076172,
"step": 3184
},
{
"epoch": 0.1930909090909091,
"grad_norm": 0.06617597490549088,
"learning_rate": 0.0008,
"loss": 2.4370031356811523,
"step": 3186
},
{
"epoch": 0.19321212121212122,
"grad_norm": 0.06790686398744583,
"learning_rate": 0.0008,
"loss": 2.401481866836548,
"step": 3188
},
{
"epoch": 0.19333333333333333,
"grad_norm": 0.06446363776922226,
"learning_rate": 0.0008,
"loss": 2.3419101238250732,
"step": 3190
},
{
"epoch": 0.19345454545454546,
"grad_norm": 0.06285624206066132,
"learning_rate": 0.0008,
"loss": 2.401165246963501,
"step": 3192
},
{
"epoch": 0.19357575757575757,
"grad_norm": 0.06348665058612823,
"learning_rate": 0.0008,
"loss": 2.40006160736084,
"step": 3194
},
{
"epoch": 0.1936969696969697,
"grad_norm": 0.06291841715574265,
"learning_rate": 0.0008,
"loss": 2.509073257446289,
"step": 3196
},
{
"epoch": 0.1938181818181818,
"grad_norm": 0.06314823776483536,
"learning_rate": 0.0008,
"loss": 2.361034870147705,
"step": 3198
},
{
"epoch": 0.19393939393939394,
"grad_norm": 0.06345036625862122,
"learning_rate": 0.0008,
"loss": 2.3938252925872803,
"step": 3200
},
{
"epoch": 0.19406060606060607,
"grad_norm": 0.06344062834978104,
"learning_rate": 0.0008,
"loss": 2.4154839515686035,
"step": 3202
},
{
"epoch": 0.19418181818181818,
"grad_norm": 0.06764448434114456,
"learning_rate": 0.0008,
"loss": 2.4613876342773438,
"step": 3204
},
{
"epoch": 0.1943030303030303,
"grad_norm": 0.06660337001085281,
"learning_rate": 0.0008,
"loss": 2.3489198684692383,
"step": 3206
},
{
"epoch": 0.19442424242424242,
"grad_norm": 0.06851381063461304,
"learning_rate": 0.0008,
"loss": 2.4193944931030273,
"step": 3208
},
{
"epoch": 0.19454545454545455,
"grad_norm": 0.06757763028144836,
"learning_rate": 0.0008,
"loss": 2.3870701789855957,
"step": 3210
},
{
"epoch": 0.19466666666666665,
"grad_norm": 0.06611692905426025,
"learning_rate": 0.0008,
"loss": 2.4861669540405273,
"step": 3212
},
{
"epoch": 0.1947878787878788,
"grad_norm": 0.06291025876998901,
"learning_rate": 0.0008,
"loss": 2.4252126216888428,
"step": 3214
},
{
"epoch": 0.19490909090909092,
"grad_norm": 0.062389444559812546,
"learning_rate": 0.0008,
"loss": 2.3808491230010986,
"step": 3216
},
{
"epoch": 0.19503030303030303,
"grad_norm": 0.06249437853693962,
"learning_rate": 0.0008,
"loss": 2.432171106338501,
"step": 3218
},
{
"epoch": 0.19515151515151516,
"grad_norm": 0.060542453080415726,
"learning_rate": 0.0008,
"loss": 2.390587091445923,
"step": 3220
},
{
"epoch": 0.19527272727272726,
"grad_norm": 0.0626068189740181,
"learning_rate": 0.0008,
"loss": 2.489795684814453,
"step": 3222
},
{
"epoch": 0.1953939393939394,
"grad_norm": 0.06425565481185913,
"learning_rate": 0.0008,
"loss": 2.408054828643799,
"step": 3224
},
{
"epoch": 0.1955151515151515,
"grad_norm": 0.06459380686283112,
"learning_rate": 0.0008,
"loss": 2.369596481323242,
"step": 3226
},
{
"epoch": 0.19563636363636364,
"grad_norm": 0.07072144746780396,
"learning_rate": 0.0008,
"loss": 2.3315045833587646,
"step": 3228
},
{
"epoch": 0.19575757575757577,
"grad_norm": 0.06522934883832932,
"learning_rate": 0.0008,
"loss": 2.412774085998535,
"step": 3230
},
{
"epoch": 0.19587878787878787,
"grad_norm": 0.06453763693571091,
"learning_rate": 0.0008,
"loss": 2.3289318084716797,
"step": 3232
},
{
"epoch": 0.196,
"grad_norm": 0.06257879734039307,
"learning_rate": 0.0008,
"loss": 2.366598129272461,
"step": 3234
},
{
"epoch": 0.1961212121212121,
"grad_norm": 0.06226776912808418,
"learning_rate": 0.0008,
"loss": 2.3690032958984375,
"step": 3236
},
{
"epoch": 0.19624242424242425,
"grad_norm": 0.06183931976556778,
"learning_rate": 0.0008,
"loss": 2.3198816776275635,
"step": 3238
},
{
"epoch": 0.19636363636363635,
"grad_norm": 0.06311545521020889,
"learning_rate": 0.0008,
"loss": 2.357520580291748,
"step": 3240
},
{
"epoch": 0.19648484848484848,
"grad_norm": 0.06399432569742203,
"learning_rate": 0.0008,
"loss": 2.3888821601867676,
"step": 3242
},
{
"epoch": 0.19660606060606062,
"grad_norm": 0.06572608649730682,
"learning_rate": 0.0008,
"loss": 2.396458387374878,
"step": 3244
},
{
"epoch": 0.19672727272727272,
"grad_norm": 0.06472118943929672,
"learning_rate": 0.0008,
"loss": 2.392313241958618,
"step": 3246
},
{
"epoch": 0.19684848484848486,
"grad_norm": 0.06490396708250046,
"learning_rate": 0.0008,
"loss": 2.387953996658325,
"step": 3248
},
{
"epoch": 0.19696969696969696,
"grad_norm": 0.062358733266592026,
"learning_rate": 0.0008,
"loss": 2.2817366123199463,
"step": 3250
},
{
"epoch": 0.1970909090909091,
"grad_norm": 0.06220576539635658,
"learning_rate": 0.0008,
"loss": 2.485719680786133,
"step": 3252
},
{
"epoch": 0.1972121212121212,
"grad_norm": 0.06013902276754379,
"learning_rate": 0.0008,
"loss": 2.3137056827545166,
"step": 3254
},
{
"epoch": 0.19733333333333333,
"grad_norm": 0.0608162097632885,
"learning_rate": 0.0008,
"loss": 2.3605339527130127,
"step": 3256
},
{
"epoch": 0.19745454545454547,
"grad_norm": 0.06299977749586105,
"learning_rate": 0.0008,
"loss": 2.4527487754821777,
"step": 3258
},
{
"epoch": 0.19757575757575757,
"grad_norm": 0.06210462376475334,
"learning_rate": 0.0008,
"loss": 2.3947136402130127,
"step": 3260
},
{
"epoch": 0.1976969696969697,
"grad_norm": 0.06282626837491989,
"learning_rate": 0.0008,
"loss": 2.2784852981567383,
"step": 3262
},
{
"epoch": 0.1978181818181818,
"grad_norm": 0.062332045286893845,
"learning_rate": 0.0008,
"loss": 2.3331308364868164,
"step": 3264
},
{
"epoch": 0.19793939393939394,
"grad_norm": 0.06282655149698257,
"learning_rate": 0.0008,
"loss": 2.39338755607605,
"step": 3266
},
{
"epoch": 0.19806060606060605,
"grad_norm": 0.062236566096544266,
"learning_rate": 0.0008,
"loss": 2.4421141147613525,
"step": 3268
},
{
"epoch": 0.19818181818181818,
"grad_norm": 0.06288332492113113,
"learning_rate": 0.0008,
"loss": 2.4447367191314697,
"step": 3270
},
{
"epoch": 0.19830303030303031,
"grad_norm": 0.0683501586318016,
"learning_rate": 0.0008,
"loss": 2.4148387908935547,
"step": 3272
},
{
"epoch": 0.19842424242424242,
"grad_norm": 0.06794796884059906,
"learning_rate": 0.0008,
"loss": 2.3373003005981445,
"step": 3274
},
{
"epoch": 0.19854545454545455,
"grad_norm": 0.062448542565107346,
"learning_rate": 0.0008,
"loss": 2.404829740524292,
"step": 3276
},
{
"epoch": 0.19866666666666666,
"grad_norm": 0.06303573399782181,
"learning_rate": 0.0008,
"loss": 2.412553548812866,
"step": 3278
},
{
"epoch": 0.1987878787878788,
"grad_norm": 0.06143725663423538,
"learning_rate": 0.0008,
"loss": 2.386976957321167,
"step": 3280
},
{
"epoch": 0.1989090909090909,
"grad_norm": 0.06411318480968475,
"learning_rate": 0.0008,
"loss": 2.3024871349334717,
"step": 3282
},
{
"epoch": 0.19903030303030303,
"grad_norm": 0.0699109211564064,
"learning_rate": 0.0008,
"loss": 2.3997457027435303,
"step": 3284
},
{
"epoch": 0.19915151515151516,
"grad_norm": 0.06363710761070251,
"learning_rate": 0.0008,
"loss": 2.3717827796936035,
"step": 3286
},
{
"epoch": 0.19927272727272727,
"grad_norm": 0.06265445798635483,
"learning_rate": 0.0008,
"loss": 2.36969256401062,
"step": 3288
},
{
"epoch": 0.1993939393939394,
"grad_norm": 0.06349951028823853,
"learning_rate": 0.0008,
"loss": 2.3728249073028564,
"step": 3290
},
{
"epoch": 0.1995151515151515,
"grad_norm": 0.061685748398303986,
"learning_rate": 0.0008,
"loss": 2.3721354007720947,
"step": 3292
},
{
"epoch": 0.19963636363636364,
"grad_norm": 0.06309618800878525,
"learning_rate": 0.0008,
"loss": 2.349907159805298,
"step": 3294
},
{
"epoch": 0.19975757575757574,
"grad_norm": 0.06090311333537102,
"learning_rate": 0.0008,
"loss": 2.357595443725586,
"step": 3296
},
{
"epoch": 0.19987878787878788,
"grad_norm": 0.06212380528450012,
"learning_rate": 0.0008,
"loss": 2.3869481086730957,
"step": 3298
},
{
"epoch": 0.2,
"grad_norm": 0.0660935565829277,
"learning_rate": 0.0008,
"loss": 2.3065407276153564,
"step": 3300
},
{
"epoch": 0.2,
"eval_loss": 2.3454651832580566,
"eval_runtime": 157.996,
"eval_samples_per_second": 1.582,
"eval_steps_per_second": 0.082,
"step": 3300
},
{
"epoch": 0.20012121212121212,
"grad_norm": 0.06644295901060104,
"learning_rate": 0.0008,
"loss": 2.343703031539917,
"step": 3302
},
{
"epoch": 0.20024242424242425,
"grad_norm": 0.06580781191587448,
"learning_rate": 0.0008,
"loss": 2.4586851596832275,
"step": 3304
},
{
"epoch": 0.20036363636363635,
"grad_norm": 0.06427513062953949,
"learning_rate": 0.0008,
"loss": 2.428772449493408,
"step": 3306
},
{
"epoch": 0.2004848484848485,
"grad_norm": 0.06291330605745316,
"learning_rate": 0.0008,
"loss": 2.383570909500122,
"step": 3308
},
{
"epoch": 0.2006060606060606,
"grad_norm": 0.06339685618877411,
"learning_rate": 0.0008,
"loss": 2.387653350830078,
"step": 3310
},
{
"epoch": 0.20072727272727273,
"grad_norm": 0.062389031052589417,
"learning_rate": 0.0008,
"loss": 2.2688257694244385,
"step": 3312
},
{
"epoch": 0.20084848484848486,
"grad_norm": 0.06149550527334213,
"learning_rate": 0.0008,
"loss": 2.420630693435669,
"step": 3314
},
{
"epoch": 0.20096969696969696,
"grad_norm": 0.06233922392129898,
"learning_rate": 0.0008,
"loss": 2.445103883743286,
"step": 3316
},
{
"epoch": 0.2010909090909091,
"grad_norm": 0.061441924422979355,
"learning_rate": 0.0008,
"loss": 2.3672516345977783,
"step": 3318
},
{
"epoch": 0.2012121212121212,
"grad_norm": 0.06120516359806061,
"learning_rate": 0.0008,
"loss": 2.3234353065490723,
"step": 3320
},
{
"epoch": 0.20133333333333334,
"grad_norm": 0.0591803677380085,
"learning_rate": 0.0008,
"loss": 2.358286142349243,
"step": 3322
},
{
"epoch": 0.20145454545454544,
"grad_norm": 0.06170762702822685,
"learning_rate": 0.0008,
"loss": 2.418170690536499,
"step": 3324
},
{
"epoch": 0.20157575757575757,
"grad_norm": 0.0659286230802536,
"learning_rate": 0.0008,
"loss": 2.358703136444092,
"step": 3326
},
{
"epoch": 0.2016969696969697,
"grad_norm": 0.06656909734010696,
"learning_rate": 0.0008,
"loss": 2.3758544921875,
"step": 3328
},
{
"epoch": 0.2018181818181818,
"grad_norm": 0.06442710012197495,
"learning_rate": 0.0008,
"loss": 2.370046377182007,
"step": 3330
},
{
"epoch": 0.20193939393939395,
"grad_norm": 0.059384673833847046,
"learning_rate": 0.0008,
"loss": 2.3198206424713135,
"step": 3332
},
{
"epoch": 0.20206060606060605,
"grad_norm": 0.060469869524240494,
"learning_rate": 0.0008,
"loss": 2.4109151363372803,
"step": 3334
},
{
"epoch": 0.20218181818181818,
"grad_norm": 0.06491401791572571,
"learning_rate": 0.0008,
"loss": 2.349246025085449,
"step": 3336
},
{
"epoch": 0.2023030303030303,
"grad_norm": 0.06123567372560501,
"learning_rate": 0.0008,
"loss": 2.4229648113250732,
"step": 3338
},
{
"epoch": 0.20242424242424242,
"grad_norm": 0.06035301089286804,
"learning_rate": 0.0008,
"loss": 2.382885217666626,
"step": 3340
},
{
"epoch": 0.20254545454545456,
"grad_norm": 0.06140818074345589,
"learning_rate": 0.0008,
"loss": 2.394896984100342,
"step": 3342
},
{
"epoch": 0.20266666666666666,
"grad_norm": 0.06633155047893524,
"learning_rate": 0.0008,
"loss": 2.40417742729187,
"step": 3344
},
{
"epoch": 0.2027878787878788,
"grad_norm": 0.06449693441390991,
"learning_rate": 0.0008,
"loss": 2.4040114879608154,
"step": 3346
},
{
"epoch": 0.2029090909090909,
"grad_norm": 0.0683647021651268,
"learning_rate": 0.0008,
"loss": 2.30930757522583,
"step": 3348
},
{
"epoch": 0.20303030303030303,
"grad_norm": 0.06314775347709656,
"learning_rate": 0.0008,
"loss": 2.4073078632354736,
"step": 3350
},
{
"epoch": 0.20315151515151514,
"grad_norm": 0.05942266434431076,
"learning_rate": 0.0008,
"loss": 2.364534616470337,
"step": 3352
},
{
"epoch": 0.20327272727272727,
"grad_norm": 0.06373094022274017,
"learning_rate": 0.0008,
"loss": 2.3898069858551025,
"step": 3354
},
{
"epoch": 0.2033939393939394,
"grad_norm": 0.06296821683645248,
"learning_rate": 0.0008,
"loss": 2.421342372894287,
"step": 3356
},
{
"epoch": 0.2035151515151515,
"grad_norm": 0.06376617401838303,
"learning_rate": 0.0008,
"loss": 2.426143169403076,
"step": 3358
},
{
"epoch": 0.20363636363636364,
"grad_norm": 0.06116616353392601,
"learning_rate": 0.0008,
"loss": 2.45633602142334,
"step": 3360
},
{
"epoch": 0.20375757575757575,
"grad_norm": 0.06398729979991913,
"learning_rate": 0.0008,
"loss": 2.3570268154144287,
"step": 3362
},
{
"epoch": 0.20387878787878788,
"grad_norm": 0.06505905091762543,
"learning_rate": 0.0008,
"loss": 2.3868179321289062,
"step": 3364
},
{
"epoch": 0.204,
"grad_norm": 0.06741669774055481,
"learning_rate": 0.0008,
"loss": 2.332385301589966,
"step": 3366
},
{
"epoch": 0.20412121212121212,
"grad_norm": 0.0672881007194519,
"learning_rate": 0.0008,
"loss": 2.4067134857177734,
"step": 3368
},
{
"epoch": 0.20424242424242425,
"grad_norm": 0.06542705744504929,
"learning_rate": 0.0008,
"loss": 2.387985944747925,
"step": 3370
},
{
"epoch": 0.20436363636363636,
"grad_norm": 0.06319617480039597,
"learning_rate": 0.0008,
"loss": 2.3548696041107178,
"step": 3372
},
{
"epoch": 0.2044848484848485,
"grad_norm": 0.06114068254828453,
"learning_rate": 0.0008,
"loss": 2.375096321105957,
"step": 3374
},
{
"epoch": 0.2046060606060606,
"grad_norm": 0.06301715970039368,
"learning_rate": 0.0008,
"loss": 2.398573637008667,
"step": 3376
},
{
"epoch": 0.20472727272727273,
"grad_norm": 0.06317971646785736,
"learning_rate": 0.0008,
"loss": 2.312035083770752,
"step": 3378
},
{
"epoch": 0.20484848484848484,
"grad_norm": 0.062296152114868164,
"learning_rate": 0.0008,
"loss": 2.341169595718384,
"step": 3380
},
{
"epoch": 0.20496969696969697,
"grad_norm": 0.062397509813308716,
"learning_rate": 0.0008,
"loss": 2.3179428577423096,
"step": 3382
},
{
"epoch": 0.2050909090909091,
"grad_norm": 0.06129693239927292,
"learning_rate": 0.0008,
"loss": 2.339761257171631,
"step": 3384
},
{
"epoch": 0.2052121212121212,
"grad_norm": 0.06106838211417198,
"learning_rate": 0.0008,
"loss": 2.3816041946411133,
"step": 3386
},
{
"epoch": 0.20533333333333334,
"grad_norm": 0.06205181032419205,
"learning_rate": 0.0008,
"loss": 2.3371403217315674,
"step": 3388
},
{
"epoch": 0.20545454545454545,
"grad_norm": 0.06108669564127922,
"learning_rate": 0.0008,
"loss": 2.3518929481506348,
"step": 3390
},
{
"epoch": 0.20557575757575758,
"grad_norm": 0.059581488370895386,
"learning_rate": 0.0008,
"loss": 2.2774853706359863,
"step": 3392
},
{
"epoch": 0.20569696969696968,
"grad_norm": 0.0615994930267334,
"learning_rate": 0.0008,
"loss": 2.4320669174194336,
"step": 3394
},
{
"epoch": 0.20581818181818182,
"grad_norm": 0.06516379117965698,
"learning_rate": 0.0008,
"loss": 2.335970640182495,
"step": 3396
},
{
"epoch": 0.20593939393939395,
"grad_norm": 0.0671282485127449,
"learning_rate": 0.0008,
"loss": 2.3610215187072754,
"step": 3398
},
{
"epoch": 0.20606060606060606,
"grad_norm": 0.06321428716182709,
"learning_rate": 0.0008,
"loss": 2.364454507827759,
"step": 3400
},
{
"epoch": 0.2061818181818182,
"grad_norm": 0.05995545908808708,
"learning_rate": 0.0008,
"loss": 2.3121426105499268,
"step": 3402
},
{
"epoch": 0.2063030303030303,
"grad_norm": 0.0600091926753521,
"learning_rate": 0.0008,
"loss": 2.398859977722168,
"step": 3404
},
{
"epoch": 0.20642424242424243,
"grad_norm": 0.06178819015622139,
"learning_rate": 0.0008,
"loss": 2.4045872688293457,
"step": 3406
},
{
"epoch": 0.20654545454545453,
"grad_norm": 0.06096385046839714,
"learning_rate": 0.0008,
"loss": 2.3972911834716797,
"step": 3408
},
{
"epoch": 0.20666666666666667,
"grad_norm": 0.06043785810470581,
"learning_rate": 0.0008,
"loss": 2.391327142715454,
"step": 3410
},
{
"epoch": 0.2067878787878788,
"grad_norm": 0.05988865718245506,
"learning_rate": 0.0008,
"loss": 2.321955919265747,
"step": 3412
},
{
"epoch": 0.2069090909090909,
"grad_norm": 0.06515010446310043,
"learning_rate": 0.0008,
"loss": 2.481890916824341,
"step": 3414
},
{
"epoch": 0.20703030303030304,
"grad_norm": 0.06396312266588211,
"learning_rate": 0.0008,
"loss": 2.440654754638672,
"step": 3416
},
{
"epoch": 0.20715151515151514,
"grad_norm": 0.07008164376020432,
"learning_rate": 0.0008,
"loss": 2.4312686920166016,
"step": 3418
},
{
"epoch": 0.20727272727272728,
"grad_norm": 0.06614779680967331,
"learning_rate": 0.0008,
"loss": 2.3492274284362793,
"step": 3420
},
{
"epoch": 0.20739393939393938,
"grad_norm": 0.06199490278959274,
"learning_rate": 0.0008,
"loss": 2.392008066177368,
"step": 3422
},
{
"epoch": 0.2075151515151515,
"grad_norm": 0.06075994297862053,
"learning_rate": 0.0008,
"loss": 2.336341619491577,
"step": 3424
},
{
"epoch": 0.20763636363636365,
"grad_norm": 0.0601169653236866,
"learning_rate": 0.0008,
"loss": 2.446683645248413,
"step": 3426
},
{
"epoch": 0.20775757575757575,
"grad_norm": 0.059829987585544586,
"learning_rate": 0.0008,
"loss": 2.310216188430786,
"step": 3428
},
{
"epoch": 0.20787878787878789,
"grad_norm": 0.06021615117788315,
"learning_rate": 0.0008,
"loss": 2.4164071083068848,
"step": 3430
},
{
"epoch": 0.208,
"grad_norm": 0.05985047668218613,
"learning_rate": 0.0008,
"loss": 2.3616135120391846,
"step": 3432
},
{
"epoch": 0.20812121212121212,
"grad_norm": 0.06018706411123276,
"learning_rate": 0.0008,
"loss": 2.3942294120788574,
"step": 3434
},
{
"epoch": 0.20824242424242423,
"grad_norm": 0.062283314764499664,
"learning_rate": 0.0008,
"loss": 2.2993602752685547,
"step": 3436
},
{
"epoch": 0.20836363636363636,
"grad_norm": 0.06245701387524605,
"learning_rate": 0.0008,
"loss": 2.4047412872314453,
"step": 3438
},
{
"epoch": 0.2084848484848485,
"grad_norm": 0.06061737239360809,
"learning_rate": 0.0008,
"loss": 2.374192714691162,
"step": 3440
},
{
"epoch": 0.2086060606060606,
"grad_norm": 0.05910682678222656,
"learning_rate": 0.0008,
"loss": 2.336188793182373,
"step": 3442
},
{
"epoch": 0.20872727272727273,
"grad_norm": 0.060264602303504944,
"learning_rate": 0.0008,
"loss": 2.381909132003784,
"step": 3444
},
{
"epoch": 0.20884848484848484,
"grad_norm": 0.06081404536962509,
"learning_rate": 0.0008,
"loss": 2.3856759071350098,
"step": 3446
},
{
"epoch": 0.20896969696969697,
"grad_norm": 0.06218332797288895,
"learning_rate": 0.0008,
"loss": 2.3362672328948975,
"step": 3448
},
{
"epoch": 0.20909090909090908,
"grad_norm": 0.06279107928276062,
"learning_rate": 0.0008,
"loss": 2.377014636993408,
"step": 3450
},
{
"epoch": 0.2092121212121212,
"grad_norm": 0.06299653649330139,
"learning_rate": 0.0008,
"loss": 2.338078498840332,
"step": 3452
},
{
"epoch": 0.20933333333333334,
"grad_norm": 0.061249881982803345,
"learning_rate": 0.0008,
"loss": 2.3899455070495605,
"step": 3454
},
{
"epoch": 0.20945454545454545,
"grad_norm": 0.060361023992300034,
"learning_rate": 0.0008,
"loss": 2.349741220474243,
"step": 3456
},
{
"epoch": 0.20957575757575758,
"grad_norm": 0.06222981959581375,
"learning_rate": 0.0008,
"loss": 2.4450809955596924,
"step": 3458
},
{
"epoch": 0.2096969696969697,
"grad_norm": 0.06303652375936508,
"learning_rate": 0.0008,
"loss": 2.306938886642456,
"step": 3460
},
{
"epoch": 0.20981818181818182,
"grad_norm": 0.06239501014351845,
"learning_rate": 0.0008,
"loss": 2.4098989963531494,
"step": 3462
},
{
"epoch": 0.20993939393939393,
"grad_norm": 0.061080511659383774,
"learning_rate": 0.0008,
"loss": 2.363349199295044,
"step": 3464
},
{
"epoch": 0.21006060606060606,
"grad_norm": 0.061735138297080994,
"learning_rate": 0.0008,
"loss": 2.335779905319214,
"step": 3466
},
{
"epoch": 0.2101818181818182,
"grad_norm": 0.06018690764904022,
"learning_rate": 0.0008,
"loss": 2.3641958236694336,
"step": 3468
},
{
"epoch": 0.2103030303030303,
"grad_norm": 0.06118885055184364,
"learning_rate": 0.0008,
"loss": 2.428173065185547,
"step": 3470
},
{
"epoch": 0.21042424242424243,
"grad_norm": 0.058638349175453186,
"learning_rate": 0.0008,
"loss": 2.368622064590454,
"step": 3472
},
{
"epoch": 0.21054545454545454,
"grad_norm": 0.05984420329332352,
"learning_rate": 0.0008,
"loss": 2.3924975395202637,
"step": 3474
},
{
"epoch": 0.21066666666666667,
"grad_norm": 0.062211260199546814,
"learning_rate": 0.0008,
"loss": 2.309882879257202,
"step": 3476
},
{
"epoch": 0.21078787878787877,
"grad_norm": 0.06024811416864395,
"learning_rate": 0.0008,
"loss": 2.332853317260742,
"step": 3478
},
{
"epoch": 0.2109090909090909,
"grad_norm": 0.06093132868409157,
"learning_rate": 0.0008,
"loss": 2.4030182361602783,
"step": 3480
},
{
"epoch": 0.21103030303030304,
"grad_norm": 0.05897024646401405,
"learning_rate": 0.0008,
"loss": 2.3691020011901855,
"step": 3482
},
{
"epoch": 0.21115151515151515,
"grad_norm": 0.06086757406592369,
"learning_rate": 0.0008,
"loss": 2.3034656047821045,
"step": 3484
},
{
"epoch": 0.21127272727272728,
"grad_norm": 0.0609210841357708,
"learning_rate": 0.0008,
"loss": 2.393033027648926,
"step": 3486
},
{
"epoch": 0.21139393939393938,
"grad_norm": 0.0612100251019001,
"learning_rate": 0.0008,
"loss": 2.4817583560943604,
"step": 3488
},
{
"epoch": 0.21151515151515152,
"grad_norm": 0.06567695736885071,
"learning_rate": 0.0008,
"loss": 2.3703057765960693,
"step": 3490
},
{
"epoch": 0.21163636363636365,
"grad_norm": 0.06198626756668091,
"learning_rate": 0.0008,
"loss": 2.3617489337921143,
"step": 3492
},
{
"epoch": 0.21175757575757576,
"grad_norm": 0.06281516700983047,
"learning_rate": 0.0008,
"loss": 2.43685245513916,
"step": 3494
},
{
"epoch": 0.2118787878787879,
"grad_norm": 0.06112358719110489,
"learning_rate": 0.0008,
"loss": 2.3176004886627197,
"step": 3496
},
{
"epoch": 0.212,
"grad_norm": 0.06005077064037323,
"learning_rate": 0.0008,
"loss": 2.3330771923065186,
"step": 3498
},
{
"epoch": 0.21212121212121213,
"grad_norm": 0.060638319700956345,
"learning_rate": 0.0008,
"loss": 2.4121053218841553,
"step": 3500
},
{
"epoch": 0.21224242424242423,
"grad_norm": 0.06295730173587799,
"learning_rate": 0.0008,
"loss": 2.3568124771118164,
"step": 3502
},
{
"epoch": 0.21236363636363637,
"grad_norm": 0.06820782274007797,
"learning_rate": 0.0008,
"loss": 2.419811725616455,
"step": 3504
},
{
"epoch": 0.2124848484848485,
"grad_norm": 0.0669749453663826,
"learning_rate": 0.0008,
"loss": 2.3697450160980225,
"step": 3506
},
{
"epoch": 0.2126060606060606,
"grad_norm": 0.05978759750723839,
"learning_rate": 0.0008,
"loss": 2.3736824989318848,
"step": 3508
},
{
"epoch": 0.21272727272727274,
"grad_norm": 0.06045331433415413,
"learning_rate": 0.0008,
"loss": 2.3797011375427246,
"step": 3510
},
{
"epoch": 0.21284848484848484,
"grad_norm": 0.059168849140405655,
"learning_rate": 0.0008,
"loss": 2.3515067100524902,
"step": 3512
},
{
"epoch": 0.21296969696969698,
"grad_norm": 0.058381225913763046,
"learning_rate": 0.0008,
"loss": 2.316814422607422,
"step": 3514
},
{
"epoch": 0.21309090909090908,
"grad_norm": 0.059535905718803406,
"learning_rate": 0.0008,
"loss": 2.2726504802703857,
"step": 3516
},
{
"epoch": 0.21321212121212121,
"grad_norm": 0.05905255302786827,
"learning_rate": 0.0008,
"loss": 2.432345151901245,
"step": 3518
},
{
"epoch": 0.21333333333333335,
"grad_norm": 0.05869733914732933,
"learning_rate": 0.0008,
"loss": 2.329862356185913,
"step": 3520
},
{
"epoch": 0.21345454545454545,
"grad_norm": 0.060022443532943726,
"learning_rate": 0.0008,
"loss": 2.3194291591644287,
"step": 3522
},
{
"epoch": 0.21357575757575759,
"grad_norm": 0.06119433045387268,
"learning_rate": 0.0008,
"loss": 2.3407065868377686,
"step": 3524
},
{
"epoch": 0.2136969696969697,
"grad_norm": 0.0638166293501854,
"learning_rate": 0.0008,
"loss": 2.3706393241882324,
"step": 3526
},
{
"epoch": 0.21381818181818182,
"grad_norm": 0.06201919913291931,
"learning_rate": 0.0008,
"loss": 2.4307913780212402,
"step": 3528
},
{
"epoch": 0.21393939393939393,
"grad_norm": 0.06076642870903015,
"learning_rate": 0.0008,
"loss": 2.3638548851013184,
"step": 3530
},
{
"epoch": 0.21406060606060606,
"grad_norm": 0.05945342034101486,
"learning_rate": 0.0008,
"loss": 2.357269287109375,
"step": 3532
},
{
"epoch": 0.2141818181818182,
"grad_norm": 0.05730807036161423,
"learning_rate": 0.0008,
"loss": 2.374481201171875,
"step": 3534
},
{
"epoch": 0.2143030303030303,
"grad_norm": 0.06174909323453903,
"learning_rate": 0.0008,
"loss": 2.435978889465332,
"step": 3536
},
{
"epoch": 0.21442424242424243,
"grad_norm": 0.06230989098548889,
"learning_rate": 0.0008,
"loss": 2.319591522216797,
"step": 3538
},
{
"epoch": 0.21454545454545454,
"grad_norm": 0.06090241298079491,
"learning_rate": 0.0008,
"loss": 2.355133533477783,
"step": 3540
},
{
"epoch": 0.21466666666666667,
"grad_norm": 0.061044320464134216,
"learning_rate": 0.0008,
"loss": 2.3594412803649902,
"step": 3542
},
{
"epoch": 0.21478787878787878,
"grad_norm": 0.05980578064918518,
"learning_rate": 0.0008,
"loss": 2.355220317840576,
"step": 3544
},
{
"epoch": 0.2149090909090909,
"grad_norm": 0.061101336032152176,
"learning_rate": 0.0008,
"loss": 2.3053674697875977,
"step": 3546
},
{
"epoch": 0.21503030303030304,
"grad_norm": 0.06387894600629807,
"learning_rate": 0.0008,
"loss": 2.360898494720459,
"step": 3548
},
{
"epoch": 0.21515151515151515,
"grad_norm": 0.06075090914964676,
"learning_rate": 0.0008,
"loss": 2.37306809425354,
"step": 3550
},
{
"epoch": 0.21527272727272728,
"grad_norm": 0.06056072190403938,
"learning_rate": 0.0008,
"loss": 2.375378370285034,
"step": 3552
},
{
"epoch": 0.2153939393939394,
"grad_norm": 0.06082633510231972,
"learning_rate": 0.0008,
"loss": 2.3431010246276855,
"step": 3554
},
{
"epoch": 0.21551515151515152,
"grad_norm": 0.060552991926670074,
"learning_rate": 0.0008,
"loss": 2.3898379802703857,
"step": 3556
},
{
"epoch": 0.21563636363636363,
"grad_norm": 0.06507236510515213,
"learning_rate": 0.0008,
"loss": 2.4007065296173096,
"step": 3558
},
{
"epoch": 0.21575757575757576,
"grad_norm": 0.06055418401956558,
"learning_rate": 0.0008,
"loss": 2.3419625759124756,
"step": 3560
},
{
"epoch": 0.2158787878787879,
"grad_norm": 0.06045026704668999,
"learning_rate": 0.0008,
"loss": 2.355135440826416,
"step": 3562
},
{
"epoch": 0.216,
"grad_norm": 0.06070132553577423,
"learning_rate": 0.0008,
"loss": 2.388695001602173,
"step": 3564
},
{
"epoch": 0.21612121212121213,
"grad_norm": 0.06095276027917862,
"learning_rate": 0.0008,
"loss": 2.3636395931243896,
"step": 3566
},
{
"epoch": 0.21624242424242424,
"grad_norm": 0.060285069048404694,
"learning_rate": 0.0008,
"loss": 2.365652084350586,
"step": 3568
},
{
"epoch": 0.21636363636363637,
"grad_norm": 0.059739843010902405,
"learning_rate": 0.0008,
"loss": 2.330418109893799,
"step": 3570
},
{
"epoch": 0.21648484848484847,
"grad_norm": 0.06000157818198204,
"learning_rate": 0.0008,
"loss": 2.369231939315796,
"step": 3572
},
{
"epoch": 0.2166060606060606,
"grad_norm": 0.0627000629901886,
"learning_rate": 0.0008,
"loss": 2.397778272628784,
"step": 3574
},
{
"epoch": 0.21672727272727274,
"grad_norm": 0.06301049143075943,
"learning_rate": 0.0008,
"loss": 2.3836798667907715,
"step": 3576
},
{
"epoch": 0.21684848484848485,
"grad_norm": 0.06147561967372894,
"learning_rate": 0.0008,
"loss": 2.3390555381774902,
"step": 3578
},
{
"epoch": 0.21696969696969698,
"grad_norm": 0.05926116555929184,
"learning_rate": 0.0008,
"loss": 2.378405809402466,
"step": 3580
},
{
"epoch": 0.21709090909090908,
"grad_norm": 0.059772055596113205,
"learning_rate": 0.0008,
"loss": 2.3297219276428223,
"step": 3582
},
{
"epoch": 0.21721212121212122,
"grad_norm": 0.06129893660545349,
"learning_rate": 0.0008,
"loss": 2.357309103012085,
"step": 3584
},
{
"epoch": 0.21733333333333332,
"grad_norm": 0.05956103652715683,
"learning_rate": 0.0008,
"loss": 2.312147617340088,
"step": 3586
},
{
"epoch": 0.21745454545454546,
"grad_norm": 0.05825521796941757,
"learning_rate": 0.0008,
"loss": 2.34063982963562,
"step": 3588
},
{
"epoch": 0.2175757575757576,
"grad_norm": 0.059185490012168884,
"learning_rate": 0.0008,
"loss": 2.317426919937134,
"step": 3590
},
{
"epoch": 0.2176969696969697,
"grad_norm": 0.060458336025476456,
"learning_rate": 0.0008,
"loss": 2.350567102432251,
"step": 3592
},
{
"epoch": 0.21781818181818183,
"grad_norm": 0.058740485459566116,
"learning_rate": 0.0008,
"loss": 2.327078104019165,
"step": 3594
},
{
"epoch": 0.21793939393939393,
"grad_norm": 0.05925673246383667,
"learning_rate": 0.0008,
"loss": 2.3151612281799316,
"step": 3596
},
{
"epoch": 0.21806060606060607,
"grad_norm": 0.06474731862545013,
"learning_rate": 0.0008,
"loss": 2.399787425994873,
"step": 3598
},
{
"epoch": 0.21818181818181817,
"grad_norm": 0.06320236623287201,
"learning_rate": 0.0008,
"loss": 2.4215986728668213,
"step": 3600
},
{
"epoch": 0.2183030303030303,
"grad_norm": 0.0596189945936203,
"learning_rate": 0.0008,
"loss": 2.4685134887695312,
"step": 3602
},
{
"epoch": 0.21842424242424244,
"grad_norm": 0.059221863746643066,
"learning_rate": 0.0008,
"loss": 2.363182306289673,
"step": 3604
},
{
"epoch": 0.21854545454545454,
"grad_norm": 0.05982910841703415,
"learning_rate": 0.0008,
"loss": 2.335524082183838,
"step": 3606
},
{
"epoch": 0.21866666666666668,
"grad_norm": 0.05940466374158859,
"learning_rate": 0.0008,
"loss": 2.356854200363159,
"step": 3608
},
{
"epoch": 0.21878787878787878,
"grad_norm": 0.06193922460079193,
"learning_rate": 0.0008,
"loss": 2.4274723529815674,
"step": 3610
},
{
"epoch": 0.21890909090909091,
"grad_norm": 0.06304489076137543,
"learning_rate": 0.0008,
"loss": 2.321931838989258,
"step": 3612
},
{
"epoch": 0.21903030303030302,
"grad_norm": 0.06336472183465958,
"learning_rate": 0.0008,
"loss": 2.4025306701660156,
"step": 3614
},
{
"epoch": 0.21915151515151515,
"grad_norm": 0.059730879962444305,
"learning_rate": 0.0008,
"loss": 2.349599599838257,
"step": 3616
},
{
"epoch": 0.21927272727272729,
"grad_norm": 0.0587596595287323,
"learning_rate": 0.0008,
"loss": 2.316051721572876,
"step": 3618
},
{
"epoch": 0.2193939393939394,
"grad_norm": 0.06056618690490723,
"learning_rate": 0.0008,
"loss": 2.374994993209839,
"step": 3620
},
{
"epoch": 0.21951515151515152,
"grad_norm": 0.0586276575922966,
"learning_rate": 0.0008,
"loss": 2.342954397201538,
"step": 3622
},
{
"epoch": 0.21963636363636363,
"grad_norm": 0.06325676292181015,
"learning_rate": 0.0008,
"loss": 2.405437469482422,
"step": 3624
},
{
"epoch": 0.21975757575757576,
"grad_norm": 0.058517422527074814,
"learning_rate": 0.0008,
"loss": 2.4241347312927246,
"step": 3626
},
{
"epoch": 0.21987878787878787,
"grad_norm": 0.0615883395075798,
"learning_rate": 0.0008,
"loss": 2.4417543411254883,
"step": 3628
},
{
"epoch": 0.22,
"grad_norm": 0.06099192425608635,
"learning_rate": 0.0008,
"loss": 2.4572947025299072,
"step": 3630
},
{
"epoch": 0.22012121212121213,
"grad_norm": 0.062163885682821274,
"learning_rate": 0.0008,
"loss": 2.409895420074463,
"step": 3632
},
{
"epoch": 0.22024242424242424,
"grad_norm": 0.061528280377388,
"learning_rate": 0.0008,
"loss": 2.33502459526062,
"step": 3634
},
{
"epoch": 0.22036363636363637,
"grad_norm": 0.0604330338537693,
"learning_rate": 0.0008,
"loss": 2.296618938446045,
"step": 3636
},
{
"epoch": 0.22048484848484848,
"grad_norm": 0.058347687125205994,
"learning_rate": 0.0008,
"loss": 2.4222660064697266,
"step": 3638
},
{
"epoch": 0.2206060606060606,
"grad_norm": 0.059658437967300415,
"learning_rate": 0.0008,
"loss": 2.408747911453247,
"step": 3640
},
{
"epoch": 0.22072727272727272,
"grad_norm": 0.05876055732369423,
"learning_rate": 0.0008,
"loss": 2.36826491355896,
"step": 3642
},
{
"epoch": 0.22084848484848485,
"grad_norm": 0.060310401022434235,
"learning_rate": 0.0008,
"loss": 2.3151445388793945,
"step": 3644
},
{
"epoch": 0.22096969696969698,
"grad_norm": 0.06790662556886673,
"learning_rate": 0.0008,
"loss": 2.317117214202881,
"step": 3646
},
{
"epoch": 0.2210909090909091,
"grad_norm": 0.06473617255687714,
"learning_rate": 0.0008,
"loss": 2.397817611694336,
"step": 3648
},
{
"epoch": 0.22121212121212122,
"grad_norm": 0.06439026445150375,
"learning_rate": 0.0008,
"loss": 2.3748483657836914,
"step": 3650
},
{
"epoch": 0.22133333333333333,
"grad_norm": 0.059371963143348694,
"learning_rate": 0.0008,
"loss": 2.353700876235962,
"step": 3652
},
{
"epoch": 0.22145454545454546,
"grad_norm": 0.05980909988284111,
"learning_rate": 0.0008,
"loss": 2.3805091381073,
"step": 3654
},
{
"epoch": 0.22157575757575756,
"grad_norm": 0.058304473757743835,
"learning_rate": 0.0008,
"loss": 2.3889524936676025,
"step": 3656
},
{
"epoch": 0.2216969696969697,
"grad_norm": 0.06019822508096695,
"learning_rate": 0.0008,
"loss": 2.255051374435425,
"step": 3658
},
{
"epoch": 0.22181818181818183,
"grad_norm": 0.060064543038606644,
"learning_rate": 0.0008,
"loss": 2.3320655822753906,
"step": 3660
},
{
"epoch": 0.22193939393939394,
"grad_norm": 0.05868018418550491,
"learning_rate": 0.0008,
"loss": 2.3182554244995117,
"step": 3662
},
{
"epoch": 0.22206060606060607,
"grad_norm": 0.061914388090372086,
"learning_rate": 0.0008,
"loss": 2.3971199989318848,
"step": 3664
},
{
"epoch": 0.22218181818181817,
"grad_norm": 0.06328081339597702,
"learning_rate": 0.0008,
"loss": 2.3444273471832275,
"step": 3666
},
{
"epoch": 0.2223030303030303,
"grad_norm": 0.06663263589143753,
"learning_rate": 0.0008,
"loss": 2.3806684017181396,
"step": 3668
},
{
"epoch": 0.2224242424242424,
"grad_norm": 0.07502009719610214,
"learning_rate": 0.0008,
"loss": 2.4248032569885254,
"step": 3670
},
{
"epoch": 0.22254545454545455,
"grad_norm": 0.06393028050661087,
"learning_rate": 0.0008,
"loss": 2.291872978210449,
"step": 3672
},
{
"epoch": 0.22266666666666668,
"grad_norm": 0.06751371920108795,
"learning_rate": 0.0008,
"loss": 2.3763504028320312,
"step": 3674
},
{
"epoch": 0.22278787878787878,
"grad_norm": 0.061670709401369095,
"learning_rate": 0.0008,
"loss": 2.3337697982788086,
"step": 3676
},
{
"epoch": 0.22290909090909092,
"grad_norm": 0.06985318660736084,
"learning_rate": 0.0008,
"loss": 2.3837108612060547,
"step": 3678
},
{
"epoch": 0.22303030303030302,
"grad_norm": 0.06124449148774147,
"learning_rate": 0.0008,
"loss": 2.2587385177612305,
"step": 3680
},
{
"epoch": 0.22315151515151516,
"grad_norm": 0.06146539747714996,
"learning_rate": 0.0008,
"loss": 2.2977778911590576,
"step": 3682
},
{
"epoch": 0.22327272727272726,
"grad_norm": 0.06092836707830429,
"learning_rate": 0.0008,
"loss": 2.3489439487457275,
"step": 3684
},
{
"epoch": 0.2233939393939394,
"grad_norm": 0.05877010151743889,
"learning_rate": 0.0008,
"loss": 2.3358330726623535,
"step": 3686
},
{
"epoch": 0.22351515151515153,
"grad_norm": 0.062047962099313736,
"learning_rate": 0.0008,
"loss": 2.2734766006469727,
"step": 3688
},
{
"epoch": 0.22363636363636363,
"grad_norm": 0.06308532506227493,
"learning_rate": 0.0008,
"loss": 2.313570976257324,
"step": 3690
},
{
"epoch": 0.22375757575757577,
"grad_norm": 0.05812095105648041,
"learning_rate": 0.0008,
"loss": 2.3644399642944336,
"step": 3692
},
{
"epoch": 0.22387878787878787,
"grad_norm": 0.0576254203915596,
"learning_rate": 0.0008,
"loss": 2.357138156890869,
"step": 3694
},
{
"epoch": 0.224,
"grad_norm": 0.06040069833397865,
"learning_rate": 0.0008,
"loss": 2.316075563430786,
"step": 3696
},
{
"epoch": 0.2241212121212121,
"grad_norm": 0.059217389672994614,
"learning_rate": 0.0008,
"loss": 2.325547456741333,
"step": 3698
},
{
"epoch": 0.22424242424242424,
"grad_norm": 0.058100275695323944,
"learning_rate": 0.0008,
"loss": 2.3026175498962402,
"step": 3700
},
{
"epoch": 0.22436363636363638,
"grad_norm": 0.060715481638908386,
"learning_rate": 0.0008,
"loss": 2.3769664764404297,
"step": 3702
},
{
"epoch": 0.22448484848484848,
"grad_norm": 0.061866600066423416,
"learning_rate": 0.0008,
"loss": 2.3021254539489746,
"step": 3704
},
{
"epoch": 0.22460606060606061,
"grad_norm": 0.07241630554199219,
"learning_rate": 0.0008,
"loss": 2.343722343444824,
"step": 3706
},
{
"epoch": 0.22472727272727272,
"grad_norm": 0.06259169429540634,
"learning_rate": 0.0008,
"loss": 2.399001121520996,
"step": 3708
},
{
"epoch": 0.22484848484848485,
"grad_norm": 0.05883322283625603,
"learning_rate": 0.0008,
"loss": 2.359848976135254,
"step": 3710
},
{
"epoch": 0.22496969696969696,
"grad_norm": 0.05811923369765282,
"learning_rate": 0.0008,
"loss": 2.392446517944336,
"step": 3712
},
{
"epoch": 0.2250909090909091,
"grad_norm": 0.05985423922538757,
"learning_rate": 0.0008,
"loss": 2.3331542015075684,
"step": 3714
},
{
"epoch": 0.22521212121212122,
"grad_norm": 0.058759406208992004,
"learning_rate": 0.0008,
"loss": 2.3251442909240723,
"step": 3716
},
{
"epoch": 0.22533333333333333,
"grad_norm": 0.0585932582616806,
"learning_rate": 0.0008,
"loss": 2.450383186340332,
"step": 3718
},
{
"epoch": 0.22545454545454546,
"grad_norm": 0.0594971589744091,
"learning_rate": 0.0008,
"loss": 2.3661599159240723,
"step": 3720
},
{
"epoch": 0.22557575757575757,
"grad_norm": 0.0601576492190361,
"learning_rate": 0.0008,
"loss": 2.3759279251098633,
"step": 3722
},
{
"epoch": 0.2256969696969697,
"grad_norm": 0.0583440363407135,
"learning_rate": 0.0008,
"loss": 2.3866426944732666,
"step": 3724
},
{
"epoch": 0.2258181818181818,
"grad_norm": 0.058524250984191895,
"learning_rate": 0.0008,
"loss": 2.309570789337158,
"step": 3726
},
{
"epoch": 0.22593939393939394,
"grad_norm": 0.05761159956455231,
"learning_rate": 0.0008,
"loss": 2.292820692062378,
"step": 3728
},
{
"epoch": 0.22606060606060607,
"grad_norm": 0.05933269113302231,
"learning_rate": 0.0008,
"loss": 2.287470579147339,
"step": 3730
},
{
"epoch": 0.22618181818181818,
"grad_norm": 0.0592174232006073,
"learning_rate": 0.0008,
"loss": 2.316711902618408,
"step": 3732
},
{
"epoch": 0.2263030303030303,
"grad_norm": 0.05722930654883385,
"learning_rate": 0.0008,
"loss": 2.2922117710113525,
"step": 3734
},
{
"epoch": 0.22642424242424242,
"grad_norm": 0.05949020758271217,
"learning_rate": 0.0008,
"loss": 2.347520351409912,
"step": 3736
},
{
"epoch": 0.22654545454545455,
"grad_norm": 0.057170093059539795,
"learning_rate": 0.0008,
"loss": 2.370473623275757,
"step": 3738
},
{
"epoch": 0.22666666666666666,
"grad_norm": 0.06258055567741394,
"learning_rate": 0.0008,
"loss": 2.361758232116699,
"step": 3740
},
{
"epoch": 0.2267878787878788,
"grad_norm": 0.058338914066553116,
"learning_rate": 0.0008,
"loss": 2.274831533432007,
"step": 3742
},
{
"epoch": 0.22690909090909092,
"grad_norm": 0.059533409774303436,
"learning_rate": 0.0008,
"loss": 2.287447452545166,
"step": 3744
},
{
"epoch": 0.22703030303030303,
"grad_norm": 0.06029726192355156,
"learning_rate": 0.0008,
"loss": 2.396071195602417,
"step": 3746
},
{
"epoch": 0.22715151515151516,
"grad_norm": 0.05982179939746857,
"learning_rate": 0.0008,
"loss": 2.377426862716675,
"step": 3748
},
{
"epoch": 0.22727272727272727,
"grad_norm": 0.05666445195674896,
"learning_rate": 0.0008,
"loss": 2.3247334957122803,
"step": 3750
},
{
"epoch": 0.2273939393939394,
"grad_norm": 0.0583440326154232,
"learning_rate": 0.0008,
"loss": 2.3328633308410645,
"step": 3752
},
{
"epoch": 0.2275151515151515,
"grad_norm": 0.05953441932797432,
"learning_rate": 0.0008,
"loss": 2.339372396469116,
"step": 3754
},
{
"epoch": 0.22763636363636364,
"grad_norm": 0.05980515480041504,
"learning_rate": 0.0008,
"loss": 2.3931870460510254,
"step": 3756
},
{
"epoch": 0.22775757575757577,
"grad_norm": 0.05760319158434868,
"learning_rate": 0.0008,
"loss": 2.361025333404541,
"step": 3758
},
{
"epoch": 0.22787878787878788,
"grad_norm": 0.059608712792396545,
"learning_rate": 0.0008,
"loss": 2.3359713554382324,
"step": 3760
},
{
"epoch": 0.228,
"grad_norm": 0.05781480297446251,
"learning_rate": 0.0008,
"loss": 2.4563286304473877,
"step": 3762
},
{
"epoch": 0.2281212121212121,
"grad_norm": 0.05885075777769089,
"learning_rate": 0.0008,
"loss": 2.3484930992126465,
"step": 3764
},
{
"epoch": 0.22824242424242425,
"grad_norm": 0.05794575437903404,
"learning_rate": 0.0008,
"loss": 2.4261159896850586,
"step": 3766
},
{
"epoch": 0.22836363636363635,
"grad_norm": 0.05755815654993057,
"learning_rate": 0.0008,
"loss": 2.2875118255615234,
"step": 3768
},
{
"epoch": 0.22848484848484849,
"grad_norm": 0.05794978886842728,
"learning_rate": 0.0008,
"loss": 2.3097119331359863,
"step": 3770
},
{
"epoch": 0.22860606060606062,
"grad_norm": 0.05857931450009346,
"learning_rate": 0.0008,
"loss": 2.2968499660491943,
"step": 3772
},
{
"epoch": 0.22872727272727272,
"grad_norm": 0.058345116674900055,
"learning_rate": 0.0008,
"loss": 2.3045918941497803,
"step": 3774
},
{
"epoch": 0.22884848484848486,
"grad_norm": 0.05901407077908516,
"learning_rate": 0.0008,
"loss": 2.3217978477478027,
"step": 3776
},
{
"epoch": 0.22896969696969696,
"grad_norm": 0.05986534059047699,
"learning_rate": 0.0008,
"loss": 2.416783332824707,
"step": 3778
},
{
"epoch": 0.2290909090909091,
"grad_norm": 0.20109300315380096,
"learning_rate": 0.0008,
"loss": 2.316739320755005,
"step": 3780
},
{
"epoch": 0.2292121212121212,
"grad_norm": 0.05921747907996178,
"learning_rate": 0.0008,
"loss": 2.355057716369629,
"step": 3782
},
{
"epoch": 0.22933333333333333,
"grad_norm": 0.05930429697036743,
"learning_rate": 0.0008,
"loss": 2.351317882537842,
"step": 3784
},
{
"epoch": 0.22945454545454547,
"grad_norm": 0.06075471639633179,
"learning_rate": 0.0008,
"loss": 2.335066318511963,
"step": 3786
},
{
"epoch": 0.22957575757575757,
"grad_norm": 0.06268128007650375,
"learning_rate": 0.0008,
"loss": 2.369474172592163,
"step": 3788
},
{
"epoch": 0.2296969696969697,
"grad_norm": 0.05948026478290558,
"learning_rate": 0.0008,
"loss": 2.341930866241455,
"step": 3790
},
{
"epoch": 0.2298181818181818,
"grad_norm": 0.0650133565068245,
"learning_rate": 0.0008,
"loss": 2.327554702758789,
"step": 3792
},
{
"epoch": 0.22993939393939394,
"grad_norm": 0.06232931837439537,
"learning_rate": 0.0008,
"loss": 2.3700950145721436,
"step": 3794
},
{
"epoch": 0.23006060606060605,
"grad_norm": 0.06412430852651596,
"learning_rate": 0.0008,
"loss": 2.385807752609253,
"step": 3796
},
{
"epoch": 0.23018181818181818,
"grad_norm": 0.061412785202264786,
"learning_rate": 0.0008,
"loss": 2.2746758460998535,
"step": 3798
},
{
"epoch": 0.23030303030303031,
"grad_norm": 0.059951893985271454,
"learning_rate": 0.0008,
"loss": 2.4383833408355713,
"step": 3800
},
{
"epoch": 0.23042424242424242,
"grad_norm": 0.061614494770765305,
"learning_rate": 0.0008,
"loss": 2.3045949935913086,
"step": 3802
},
{
"epoch": 0.23054545454545455,
"grad_norm": 0.057723693549633026,
"learning_rate": 0.0008,
"loss": 2.3085076808929443,
"step": 3804
},
{
"epoch": 0.23066666666666666,
"grad_norm": 0.05885179340839386,
"learning_rate": 0.0008,
"loss": 2.4058308601379395,
"step": 3806
},
{
"epoch": 0.2307878787878788,
"grad_norm": 0.058545246720314026,
"learning_rate": 0.0008,
"loss": 2.3356590270996094,
"step": 3808
},
{
"epoch": 0.2309090909090909,
"grad_norm": 0.05769583582878113,
"learning_rate": 0.0008,
"loss": 2.396768808364868,
"step": 3810
},
{
"epoch": 0.23103030303030303,
"grad_norm": 0.05468669906258583,
"learning_rate": 0.0008,
"loss": 2.3331267833709717,
"step": 3812
},
{
"epoch": 0.23115151515151516,
"grad_norm": 0.06243548169732094,
"learning_rate": 0.0008,
"loss": 2.330002546310425,
"step": 3814
},
{
"epoch": 0.23127272727272727,
"grad_norm": 0.06422052532434464,
"learning_rate": 0.0008,
"loss": 2.4068386554718018,
"step": 3816
},
{
"epoch": 0.2313939393939394,
"grad_norm": 0.05971241369843483,
"learning_rate": 0.0008,
"loss": 2.278559923171997,
"step": 3818
},
{
"epoch": 0.2315151515151515,
"grad_norm": 0.058539677411317825,
"learning_rate": 0.0008,
"loss": 2.371701955795288,
"step": 3820
},
{
"epoch": 0.23163636363636364,
"grad_norm": 0.05814191699028015,
"learning_rate": 0.0008,
"loss": 2.327314853668213,
"step": 3822
},
{
"epoch": 0.23175757575757575,
"grad_norm": 0.05799546092748642,
"learning_rate": 0.0008,
"loss": 2.3705921173095703,
"step": 3824
},
{
"epoch": 0.23187878787878788,
"grad_norm": 0.05907570943236351,
"learning_rate": 0.0008,
"loss": 2.4007577896118164,
"step": 3826
},
{
"epoch": 0.232,
"grad_norm": 0.056397564709186554,
"learning_rate": 0.0008,
"loss": 2.325538158416748,
"step": 3828
},
{
"epoch": 0.23212121212121212,
"grad_norm": 0.05815138667821884,
"learning_rate": 0.0008,
"loss": 2.2851619720458984,
"step": 3830
},
{
"epoch": 0.23224242424242425,
"grad_norm": 0.05725277587771416,
"learning_rate": 0.0008,
"loss": 2.283728837966919,
"step": 3832
},
{
"epoch": 0.23236363636363636,
"grad_norm": 0.056375280022621155,
"learning_rate": 0.0008,
"loss": 2.2371628284454346,
"step": 3834
},
{
"epoch": 0.2324848484848485,
"grad_norm": 0.05638980120420456,
"learning_rate": 0.0008,
"loss": 2.313784122467041,
"step": 3836
},
{
"epoch": 0.2326060606060606,
"grad_norm": 0.05666641891002655,
"learning_rate": 0.0008,
"loss": 2.27374267578125,
"step": 3838
},
{
"epoch": 0.23272727272727273,
"grad_norm": 0.0652182400226593,
"learning_rate": 0.0008,
"loss": 2.371403217315674,
"step": 3840
},
{
"epoch": 0.23284848484848486,
"grad_norm": 0.05796528607606888,
"learning_rate": 0.0008,
"loss": 2.410468101501465,
"step": 3842
},
{
"epoch": 0.23296969696969697,
"grad_norm": 0.05906786024570465,
"learning_rate": 0.0008,
"loss": 2.355583429336548,
"step": 3844
},
{
"epoch": 0.2330909090909091,
"grad_norm": 0.05942641943693161,
"learning_rate": 0.0008,
"loss": 2.3535244464874268,
"step": 3846
},
{
"epoch": 0.2332121212121212,
"grad_norm": 0.05831633508205414,
"learning_rate": 0.0008,
"loss": 2.3464303016662598,
"step": 3848
},
{
"epoch": 0.23333333333333334,
"grad_norm": 0.058287929743528366,
"learning_rate": 0.0008,
"loss": 2.3608744144439697,
"step": 3850
},
{
"epoch": 0.23345454545454544,
"grad_norm": 0.05863798037171364,
"learning_rate": 0.0008,
"loss": 2.2930095195770264,
"step": 3852
},
{
"epoch": 0.23357575757575758,
"grad_norm": 0.0570516362786293,
"learning_rate": 0.0008,
"loss": 2.2674875259399414,
"step": 3854
},
{
"epoch": 0.2336969696969697,
"grad_norm": 0.06633114069700241,
"learning_rate": 0.0008,
"loss": 2.3804638385772705,
"step": 3856
},
{
"epoch": 0.23381818181818181,
"grad_norm": 0.05837031826376915,
"learning_rate": 0.0008,
"loss": 2.2630417346954346,
"step": 3858
},
{
"epoch": 0.23393939393939395,
"grad_norm": 0.05885876342654228,
"learning_rate": 0.0008,
"loss": 2.3121442794799805,
"step": 3860
},
{
"epoch": 0.23406060606060605,
"grad_norm": 0.058823712170124054,
"learning_rate": 0.0008,
"loss": 2.321859359741211,
"step": 3862
},
{
"epoch": 0.23418181818181819,
"grad_norm": 0.05821113660931587,
"learning_rate": 0.0008,
"loss": 2.371182441711426,
"step": 3864
},
{
"epoch": 0.2343030303030303,
"grad_norm": 0.06509250402450562,
"learning_rate": 0.0008,
"loss": 2.351212978363037,
"step": 3866
},
{
"epoch": 0.23442424242424242,
"grad_norm": 0.059710919857025146,
"learning_rate": 0.0008,
"loss": 2.3224947452545166,
"step": 3868
},
{
"epoch": 0.23454545454545456,
"grad_norm": 0.059378352016210556,
"learning_rate": 0.0008,
"loss": 2.2942121028900146,
"step": 3870
},
{
"epoch": 0.23466666666666666,
"grad_norm": 0.058225587010383606,
"learning_rate": 0.0008,
"loss": 2.3027846813201904,
"step": 3872
},
{
"epoch": 0.2347878787878788,
"grad_norm": 0.0590234138071537,
"learning_rate": 0.0008,
"loss": 2.3752713203430176,
"step": 3874
},
{
"epoch": 0.2349090909090909,
"grad_norm": 0.06160789728164673,
"learning_rate": 0.0008,
"loss": 2.35866379737854,
"step": 3876
},
{
"epoch": 0.23503030303030303,
"grad_norm": 0.06144797429442406,
"learning_rate": 0.0008,
"loss": 2.3664562702178955,
"step": 3878
},
{
"epoch": 0.23515151515151514,
"grad_norm": 0.06430505961179733,
"learning_rate": 0.0008,
"loss": 2.250906229019165,
"step": 3880
},
{
"epoch": 0.23527272727272727,
"grad_norm": 0.06820724904537201,
"learning_rate": 0.0008,
"loss": 2.395662307739258,
"step": 3882
},
{
"epoch": 0.2353939393939394,
"grad_norm": 0.06850579380989075,
"learning_rate": 0.0008,
"loss": 2.2976865768432617,
"step": 3884
},
{
"epoch": 0.2355151515151515,
"grad_norm": 0.06636672466993332,
"learning_rate": 0.0008,
"loss": 2.3368399143218994,
"step": 3886
},
{
"epoch": 0.23563636363636364,
"grad_norm": 0.0672006830573082,
"learning_rate": 0.0008,
"loss": 2.3422114849090576,
"step": 3888
},
{
"epoch": 0.23575757575757575,
"grad_norm": 0.059673767536878586,
"learning_rate": 0.0008,
"loss": 2.3023509979248047,
"step": 3890
},
{
"epoch": 0.23587878787878788,
"grad_norm": 0.059761468321084976,
"learning_rate": 0.0008,
"loss": 2.375650644302368,
"step": 3892
},
{
"epoch": 0.236,
"grad_norm": 0.05827547237277031,
"learning_rate": 0.0008,
"loss": 2.3903191089630127,
"step": 3894
},
{
"epoch": 0.23612121212121212,
"grad_norm": 0.059061452746391296,
"learning_rate": 0.0008,
"loss": 2.330333709716797,
"step": 3896
},
{
"epoch": 0.23624242424242425,
"grad_norm": 0.058004897087812424,
"learning_rate": 0.0008,
"loss": 2.370435953140259,
"step": 3898
},
{
"epoch": 0.23636363636363636,
"grad_norm": 0.05900784954428673,
"learning_rate": 0.0008,
"loss": 2.3416755199432373,
"step": 3900
},
{
"epoch": 0.2364848484848485,
"grad_norm": 0.05726140737533569,
"learning_rate": 0.0008,
"loss": 2.3155927658081055,
"step": 3902
},
{
"epoch": 0.2366060606060606,
"grad_norm": 0.058020446449518204,
"learning_rate": 0.0008,
"loss": 2.3458967208862305,
"step": 3904
},
{
"epoch": 0.23672727272727273,
"grad_norm": 0.05846282094717026,
"learning_rate": 0.0008,
"loss": 2.329672336578369,
"step": 3906
},
{
"epoch": 0.23684848484848484,
"grad_norm": 0.05903160199522972,
"learning_rate": 0.0008,
"loss": 2.392167091369629,
"step": 3908
},
{
"epoch": 0.23696969696969697,
"grad_norm": 0.058208759874105453,
"learning_rate": 0.0008,
"loss": 2.336911201477051,
"step": 3910
},
{
"epoch": 0.2370909090909091,
"grad_norm": 0.05819345638155937,
"learning_rate": 0.0008,
"loss": 2.381307601928711,
"step": 3912
},
{
"epoch": 0.2372121212121212,
"grad_norm": 0.05838175490498543,
"learning_rate": 0.0008,
"loss": 2.3711512088775635,
"step": 3914
},
{
"epoch": 0.23733333333333334,
"grad_norm": 0.05862388014793396,
"learning_rate": 0.0008,
"loss": 2.29900860786438,
"step": 3916
},
{
"epoch": 0.23745454545454545,
"grad_norm": 0.058891668915748596,
"learning_rate": 0.0008,
"loss": 2.401496648788452,
"step": 3918
},
{
"epoch": 0.23757575757575758,
"grad_norm": 0.05935104191303253,
"learning_rate": 0.0008,
"loss": 2.299919843673706,
"step": 3920
},
{
"epoch": 0.23769696969696968,
"grad_norm": 0.0572003610432148,
"learning_rate": 0.0008,
"loss": 2.3073511123657227,
"step": 3922
},
{
"epoch": 0.23781818181818182,
"grad_norm": 0.06172458454966545,
"learning_rate": 0.0008,
"loss": 2.434468984603882,
"step": 3924
},
{
"epoch": 0.23793939393939395,
"grad_norm": 0.05826212838292122,
"learning_rate": 0.0008,
"loss": 2.306359052658081,
"step": 3926
},
{
"epoch": 0.23806060606060606,
"grad_norm": 0.056962959468364716,
"learning_rate": 0.0008,
"loss": 2.3367090225219727,
"step": 3928
},
{
"epoch": 0.2381818181818182,
"grad_norm": 0.06560514122247696,
"learning_rate": 0.0008,
"loss": 2.378406047821045,
"step": 3930
},
{
"epoch": 0.2383030303030303,
"grad_norm": 0.05808865278959274,
"learning_rate": 0.0008,
"loss": 2.3073678016662598,
"step": 3932
},
{
"epoch": 0.23842424242424243,
"grad_norm": 0.05970320478081703,
"learning_rate": 0.0008,
"loss": 2.3152074813842773,
"step": 3934
},
{
"epoch": 0.23854545454545453,
"grad_norm": 0.05879412218928337,
"learning_rate": 0.0008,
"loss": 2.2809810638427734,
"step": 3936
},
{
"epoch": 0.23866666666666667,
"grad_norm": 0.05620420724153519,
"learning_rate": 0.0008,
"loss": 2.221090793609619,
"step": 3938
},
{
"epoch": 0.2387878787878788,
"grad_norm": 0.05752316489815712,
"learning_rate": 0.0008,
"loss": 2.2999398708343506,
"step": 3940
},
{
"epoch": 0.2389090909090909,
"grad_norm": 0.057674650102853775,
"learning_rate": 0.0008,
"loss": 2.3606085777282715,
"step": 3942
},
{
"epoch": 0.23903030303030304,
"grad_norm": 0.05615035071969032,
"learning_rate": 0.0008,
"loss": 2.3004605770111084,
"step": 3944
},
{
"epoch": 0.23915151515151514,
"grad_norm": 0.05954557657241821,
"learning_rate": 0.0008,
"loss": 2.423476457595825,
"step": 3946
},
{
"epoch": 0.23927272727272728,
"grad_norm": 0.0631764680147171,
"learning_rate": 0.0008,
"loss": 2.362670421600342,
"step": 3948
},
{
"epoch": 0.23939393939393938,
"grad_norm": 0.06506195664405823,
"learning_rate": 0.0008,
"loss": 2.3454747200012207,
"step": 3950
},
{
"epoch": 0.23951515151515151,
"grad_norm": 0.058207977563142776,
"learning_rate": 0.0008,
"loss": 2.353426218032837,
"step": 3952
},
{
"epoch": 0.23963636363636365,
"grad_norm": 0.05796258524060249,
"learning_rate": 0.0008,
"loss": 2.2876508235931396,
"step": 3954
},
{
"epoch": 0.23975757575757575,
"grad_norm": 0.05863906443119049,
"learning_rate": 0.0008,
"loss": 2.2994260787963867,
"step": 3956
},
{
"epoch": 0.23987878787878789,
"grad_norm": 0.06288156658411026,
"learning_rate": 0.0008,
"loss": 2.3417880535125732,
"step": 3958
},
{
"epoch": 0.24,
"grad_norm": 0.06759846955537796,
"learning_rate": 0.0008,
"loss": 2.3592028617858887,
"step": 3960
},
{
"epoch": 0.24012121212121212,
"grad_norm": 0.058577585965394974,
"learning_rate": 0.0008,
"loss": 2.3341927528381348,
"step": 3962
},
{
"epoch": 0.24024242424242423,
"grad_norm": 0.057559359818696976,
"learning_rate": 0.0008,
"loss": 2.3697829246520996,
"step": 3964
},
{
"epoch": 0.24036363636363636,
"grad_norm": 0.05777288228273392,
"learning_rate": 0.0008,
"loss": 2.3558735847473145,
"step": 3966
},
{
"epoch": 0.2404848484848485,
"grad_norm": 0.056973520666360855,
"learning_rate": 0.0008,
"loss": 2.208238363265991,
"step": 3968
},
{
"epoch": 0.2406060606060606,
"grad_norm": 0.057529885321855545,
"learning_rate": 0.0008,
"loss": 2.3652331829071045,
"step": 3970
},
{
"epoch": 0.24072727272727273,
"grad_norm": 0.05713169276714325,
"learning_rate": 0.0008,
"loss": 2.3655126094818115,
"step": 3972
},
{
"epoch": 0.24084848484848484,
"grad_norm": 0.058684736490249634,
"learning_rate": 0.0008,
"loss": 2.367666006088257,
"step": 3974
},
{
"epoch": 0.24096969696969697,
"grad_norm": 0.058565668761730194,
"learning_rate": 0.0008,
"loss": 2.3013381958007812,
"step": 3976
},
{
"epoch": 0.24109090909090908,
"grad_norm": 0.057647403329610825,
"learning_rate": 0.0008,
"loss": 2.3404695987701416,
"step": 3978
},
{
"epoch": 0.2412121212121212,
"grad_norm": 0.05621769651770592,
"learning_rate": 0.0008,
"loss": 2.3322060108184814,
"step": 3980
},
{
"epoch": 0.24133333333333334,
"grad_norm": 0.055793240666389465,
"learning_rate": 0.0008,
"loss": 2.2863526344299316,
"step": 3982
},
{
"epoch": 0.24145454545454545,
"grad_norm": 0.055241961032152176,
"learning_rate": 0.0008,
"loss": 2.3187427520751953,
"step": 3984
},
{
"epoch": 0.24157575757575758,
"grad_norm": 0.05842605233192444,
"learning_rate": 0.0008,
"loss": 2.3507907390594482,
"step": 3986
},
{
"epoch": 0.2416969696969697,
"grad_norm": 0.05926887318491936,
"learning_rate": 0.0008,
"loss": 2.282949924468994,
"step": 3988
},
{
"epoch": 0.24181818181818182,
"grad_norm": 0.059015173465013504,
"learning_rate": 0.0008,
"loss": 2.30261492729187,
"step": 3990
},
{
"epoch": 0.24193939393939393,
"grad_norm": 0.05698901787400246,
"learning_rate": 0.0008,
"loss": 2.342902898788452,
"step": 3992
},
{
"epoch": 0.24206060606060606,
"grad_norm": 0.059526145458221436,
"learning_rate": 0.0008,
"loss": 2.3741798400878906,
"step": 3994
},
{
"epoch": 0.2421818181818182,
"grad_norm": 0.06330876052379608,
"learning_rate": 0.0008,
"loss": 2.3275294303894043,
"step": 3996
},
{
"epoch": 0.2423030303030303,
"grad_norm": 0.062432661652565,
"learning_rate": 0.0008,
"loss": 2.3211746215820312,
"step": 3998
},
{
"epoch": 0.24242424242424243,
"grad_norm": 0.05915740504860878,
"learning_rate": 0.0008,
"loss": 2.3302693367004395,
"step": 4000
},
{
"epoch": 0.24254545454545454,
"grad_norm": 0.05811610817909241,
"learning_rate": 0.0008,
"loss": 2.318605899810791,
"step": 4002
},
{
"epoch": 0.24266666666666667,
"grad_norm": 0.05992354452610016,
"learning_rate": 0.0008,
"loss": 2.3499338626861572,
"step": 4004
},
{
"epoch": 0.24278787878787877,
"grad_norm": 0.055885251611471176,
"learning_rate": 0.0008,
"loss": 2.295022964477539,
"step": 4006
},
{
"epoch": 0.2429090909090909,
"grad_norm": 0.05762479826807976,
"learning_rate": 0.0008,
"loss": 2.3521554470062256,
"step": 4008
},
{
"epoch": 0.24303030303030304,
"grad_norm": 0.056739915162324905,
"learning_rate": 0.0008,
"loss": 2.3270480632781982,
"step": 4010
},
{
"epoch": 0.24315151515151515,
"grad_norm": 0.06058281287550926,
"learning_rate": 0.0008,
"loss": 2.3155791759490967,
"step": 4012
},
{
"epoch": 0.24327272727272728,
"grad_norm": 0.06252892315387726,
"learning_rate": 0.0008,
"loss": 2.3584296703338623,
"step": 4014
},
{
"epoch": 0.24339393939393938,
"grad_norm": 0.06383760273456573,
"learning_rate": 0.0008,
"loss": 2.301232099533081,
"step": 4016
},
{
"epoch": 0.24351515151515152,
"grad_norm": 0.06355509907007217,
"learning_rate": 0.0008,
"loss": 2.2691781520843506,
"step": 4018
},
{
"epoch": 0.24363636363636362,
"grad_norm": 0.05729431286454201,
"learning_rate": 0.0008,
"loss": 2.3952081203460693,
"step": 4020
},
{
"epoch": 0.24375757575757576,
"grad_norm": 0.05660161375999451,
"learning_rate": 0.0008,
"loss": 2.306033134460449,
"step": 4022
},
{
"epoch": 0.2438787878787879,
"grad_norm": 0.05900964513421059,
"learning_rate": 0.0008,
"loss": 2.373225688934326,
"step": 4024
},
{
"epoch": 0.244,
"grad_norm": 0.05967697873711586,
"learning_rate": 0.0008,
"loss": 2.3713021278381348,
"step": 4026
},
{
"epoch": 0.24412121212121213,
"grad_norm": 0.05836792662739754,
"learning_rate": 0.0008,
"loss": 2.3546252250671387,
"step": 4028
},
{
"epoch": 0.24424242424242423,
"grad_norm": 0.0576455257833004,
"learning_rate": 0.0008,
"loss": 2.354295492172241,
"step": 4030
},
{
"epoch": 0.24436363636363637,
"grad_norm": 0.056558337062597275,
"learning_rate": 0.0008,
"loss": 2.329935073852539,
"step": 4032
},
{
"epoch": 0.24448484848484847,
"grad_norm": 0.054967015981674194,
"learning_rate": 0.0008,
"loss": 2.304921865463257,
"step": 4034
},
{
"epoch": 0.2446060606060606,
"grad_norm": 0.0567418672144413,
"learning_rate": 0.0008,
"loss": 2.330441474914551,
"step": 4036
},
{
"epoch": 0.24472727272727274,
"grad_norm": 0.05728789418935776,
"learning_rate": 0.0008,
"loss": 2.3416805267333984,
"step": 4038
},
{
"epoch": 0.24484848484848484,
"grad_norm": 0.056817974895238876,
"learning_rate": 0.0008,
"loss": 2.3149664402008057,
"step": 4040
},
{
"epoch": 0.24496969696969698,
"grad_norm": 0.0586363710463047,
"learning_rate": 0.0008,
"loss": 2.3302032947540283,
"step": 4042
},
{
"epoch": 0.24509090909090908,
"grad_norm": 0.06244336813688278,
"learning_rate": 0.0008,
"loss": 2.3016648292541504,
"step": 4044
},
{
"epoch": 0.24521212121212121,
"grad_norm": 0.058773599565029144,
"learning_rate": 0.0008,
"loss": 2.2876479625701904,
"step": 4046
},
{
"epoch": 0.24533333333333332,
"grad_norm": 0.0559917688369751,
"learning_rate": 0.0008,
"loss": 2.362435817718506,
"step": 4048
},
{
"epoch": 0.24545454545454545,
"grad_norm": 0.05505934730172157,
"learning_rate": 0.0008,
"loss": 2.340914249420166,
"step": 4050
},
{
"epoch": 0.2455757575757576,
"grad_norm": 0.056102629750967026,
"learning_rate": 0.0008,
"loss": 2.28023362159729,
"step": 4052
},
{
"epoch": 0.2456969696969697,
"grad_norm": 0.05614113807678223,
"learning_rate": 0.0008,
"loss": 2.3224358558654785,
"step": 4054
},
{
"epoch": 0.24581818181818182,
"grad_norm": 0.056151557713747025,
"learning_rate": 0.0008,
"loss": 2.3566386699676514,
"step": 4056
},
{
"epoch": 0.24593939393939393,
"grad_norm": 0.05578375980257988,
"learning_rate": 0.0008,
"loss": 2.345747470855713,
"step": 4058
},
{
"epoch": 0.24606060606060606,
"grad_norm": 0.0565609484910965,
"learning_rate": 0.0008,
"loss": 2.307706117630005,
"step": 4060
},
{
"epoch": 0.24618181818181817,
"grad_norm": 0.05739511549472809,
"learning_rate": 0.0008,
"loss": 2.351661205291748,
"step": 4062
},
{
"epoch": 0.2463030303030303,
"grad_norm": 0.05969545617699623,
"learning_rate": 0.0008,
"loss": 2.3809664249420166,
"step": 4064
},
{
"epoch": 0.24642424242424243,
"grad_norm": 0.055300820618867874,
"learning_rate": 0.0008,
"loss": 2.2858052253723145,
"step": 4066
},
{
"epoch": 0.24654545454545454,
"grad_norm": 0.05574730038642883,
"learning_rate": 0.0008,
"loss": 2.28066349029541,
"step": 4068
},
{
"epoch": 0.24666666666666667,
"grad_norm": 0.05715244635939598,
"learning_rate": 0.0008,
"loss": 2.3349177837371826,
"step": 4070
},
{
"epoch": 0.24678787878787878,
"grad_norm": 0.05595707893371582,
"learning_rate": 0.0008,
"loss": 2.289381504058838,
"step": 4072
},
{
"epoch": 0.2469090909090909,
"grad_norm": 0.057443633675575256,
"learning_rate": 0.0008,
"loss": 2.317105531692505,
"step": 4074
},
{
"epoch": 0.24703030303030302,
"grad_norm": 0.05788515880703926,
"learning_rate": 0.0008,
"loss": 2.2469866275787354,
"step": 4076
},
{
"epoch": 0.24715151515151515,
"grad_norm": 0.057550109922885895,
"learning_rate": 0.0008,
"loss": 2.2977888584136963,
"step": 4078
},
{
"epoch": 0.24727272727272728,
"grad_norm": 0.059939853847026825,
"learning_rate": 0.0008,
"loss": 2.4117014408111572,
"step": 4080
},
{
"epoch": 0.2473939393939394,
"grad_norm": 0.055290281772613525,
"learning_rate": 0.0008,
"loss": 2.3443610668182373,
"step": 4082
},
{
"epoch": 0.24751515151515152,
"grad_norm": 0.05711251497268677,
"learning_rate": 0.0008,
"loss": 2.3237903118133545,
"step": 4084
},
{
"epoch": 0.24763636363636363,
"grad_norm": 0.057593874633312225,
"learning_rate": 0.0008,
"loss": 2.3088464736938477,
"step": 4086
},
{
"epoch": 0.24775757575757576,
"grad_norm": 0.06026874855160713,
"learning_rate": 0.0008,
"loss": 2.3023717403411865,
"step": 4088
},
{
"epoch": 0.24787878787878787,
"grad_norm": 0.061730895191431046,
"learning_rate": 0.0008,
"loss": 2.368833541870117,
"step": 4090
},
{
"epoch": 0.248,
"grad_norm": 0.06465277820825577,
"learning_rate": 0.0008,
"loss": 2.3476903438568115,
"step": 4092
},
{
"epoch": 0.24812121212121213,
"grad_norm": 0.06601168215274811,
"learning_rate": 0.0008,
"loss": 2.2619576454162598,
"step": 4094
},
{
"epoch": 0.24824242424242424,
"grad_norm": 0.06376107782125473,
"learning_rate": 0.0008,
"loss": 2.334989070892334,
"step": 4096
},
{
"epoch": 0.24836363636363637,
"grad_norm": 0.05787290632724762,
"learning_rate": 0.0008,
"loss": 2.3401365280151367,
"step": 4098
},
{
"epoch": 0.24848484848484848,
"grad_norm": 0.05997396260499954,
"learning_rate": 0.0008,
"loss": 2.325411319732666,
"step": 4100
},
{
"epoch": 0.2486060606060606,
"grad_norm": 0.053739480674266815,
"learning_rate": 0.0008,
"loss": 2.343714714050293,
"step": 4102
},
{
"epoch": 0.2487272727272727,
"grad_norm": 0.05758555978536606,
"learning_rate": 0.0008,
"loss": 2.3331921100616455,
"step": 4104
},
{
"epoch": 0.24884848484848485,
"grad_norm": 0.05989546328783035,
"learning_rate": 0.0008,
"loss": 2.289428234100342,
"step": 4106
},
{
"epoch": 0.24896969696969698,
"grad_norm": 0.05948084220290184,
"learning_rate": 0.0008,
"loss": 2.351261615753174,
"step": 4108
},
{
"epoch": 0.24909090909090909,
"grad_norm": 0.05688918009400368,
"learning_rate": 0.0008,
"loss": 2.345147132873535,
"step": 4110
},
{
"epoch": 0.24921212121212122,
"grad_norm": 0.05545199662446976,
"learning_rate": 0.0008,
"loss": 2.323451042175293,
"step": 4112
},
{
"epoch": 0.24933333333333332,
"grad_norm": 0.05650079622864723,
"learning_rate": 0.0008,
"loss": 2.275995969772339,
"step": 4114
},
{
"epoch": 0.24945454545454546,
"grad_norm": 0.05652569606900215,
"learning_rate": 0.0008,
"loss": 2.3242151737213135,
"step": 4116
},
{
"epoch": 0.24957575757575756,
"grad_norm": 0.06021900475025177,
"learning_rate": 0.0008,
"loss": 2.279433250427246,
"step": 4118
},
{
"epoch": 0.2496969696969697,
"grad_norm": 0.056628305464982986,
"learning_rate": 0.0008,
"loss": 2.4045519828796387,
"step": 4120
},
{
"epoch": 0.24981818181818183,
"grad_norm": 0.057233311235904694,
"learning_rate": 0.0008,
"loss": 2.392268419265747,
"step": 4122
},
{
"epoch": 0.24993939393939393,
"grad_norm": 0.05779198929667473,
"learning_rate": 0.0008,
"loss": 2.2828595638275146,
"step": 4124
},
{
"epoch": 0.25006060606060604,
"grad_norm": 0.05635727569460869,
"learning_rate": 0.0008,
"loss": 2.3448638916015625,
"step": 4126
},
{
"epoch": 0.25018181818181817,
"grad_norm": 0.05565233156085014,
"learning_rate": 0.0008,
"loss": 2.3306944370269775,
"step": 4128
},
{
"epoch": 0.2503030303030303,
"grad_norm": 0.057450778782367706,
"learning_rate": 0.0008,
"loss": 2.3550944328308105,
"step": 4130
},
{
"epoch": 0.25042424242424244,
"grad_norm": 0.05571691319346428,
"learning_rate": 0.0008,
"loss": 2.329329252243042,
"step": 4132
},
{
"epoch": 0.25054545454545457,
"grad_norm": 0.05670097842812538,
"learning_rate": 0.0008,
"loss": 2.389962673187256,
"step": 4134
},
{
"epoch": 0.25066666666666665,
"grad_norm": 0.057332925498485565,
"learning_rate": 0.0008,
"loss": 2.364967107772827,
"step": 4136
},
{
"epoch": 0.2507878787878788,
"grad_norm": 0.05854044482111931,
"learning_rate": 0.0008,
"loss": 2.3457093238830566,
"step": 4138
},
{
"epoch": 0.2509090909090909,
"grad_norm": 0.05886657536029816,
"learning_rate": 0.0008,
"loss": 2.329021453857422,
"step": 4140
},
{
"epoch": 0.25103030303030305,
"grad_norm": 0.057533275336027145,
"learning_rate": 0.0008,
"loss": 2.2667837142944336,
"step": 4142
},
{
"epoch": 0.2511515151515151,
"grad_norm": 0.05617799982428551,
"learning_rate": 0.0008,
"loss": 2.3037116527557373,
"step": 4144
},
{
"epoch": 0.25127272727272726,
"grad_norm": 0.05456642061471939,
"learning_rate": 0.0008,
"loss": 2.2937614917755127,
"step": 4146
},
{
"epoch": 0.2513939393939394,
"grad_norm": 0.058536410331726074,
"learning_rate": 0.0008,
"loss": 2.3001251220703125,
"step": 4148
},
{
"epoch": 0.2515151515151515,
"grad_norm": 0.06260526180267334,
"learning_rate": 0.0008,
"loss": 2.264333486557007,
"step": 4150
},
{
"epoch": 0.25163636363636366,
"grad_norm": 0.05842176452279091,
"learning_rate": 0.0008,
"loss": 2.2603983879089355,
"step": 4152
},
{
"epoch": 0.25175757575757574,
"grad_norm": 0.05674688145518303,
"learning_rate": 0.0008,
"loss": 2.287539005279541,
"step": 4154
},
{
"epoch": 0.25187878787878787,
"grad_norm": 0.06014281138777733,
"learning_rate": 0.0008,
"loss": 2.324986219406128,
"step": 4156
},
{
"epoch": 0.252,
"grad_norm": 0.06347078084945679,
"learning_rate": 0.0008,
"loss": 2.3204336166381836,
"step": 4158
},
{
"epoch": 0.25212121212121213,
"grad_norm": 0.06463081389665604,
"learning_rate": 0.0008,
"loss": 2.340038537979126,
"step": 4160
},
{
"epoch": 0.25224242424242427,
"grad_norm": 0.05744042620062828,
"learning_rate": 0.0008,
"loss": 2.3520915508270264,
"step": 4162
},
{
"epoch": 0.25236363636363635,
"grad_norm": 0.057763926684856415,
"learning_rate": 0.0008,
"loss": 2.356677293777466,
"step": 4164
},
{
"epoch": 0.2524848484848485,
"grad_norm": 0.05633429065346718,
"learning_rate": 0.0008,
"loss": 2.369781494140625,
"step": 4166
},
{
"epoch": 0.2526060606060606,
"grad_norm": 0.0570233128964901,
"learning_rate": 0.0008,
"loss": 2.344925880432129,
"step": 4168
},
{
"epoch": 0.25272727272727274,
"grad_norm": 0.05592390522360802,
"learning_rate": 0.0008,
"loss": 2.3292970657348633,
"step": 4170
},
{
"epoch": 0.2528484848484848,
"grad_norm": 0.056636400520801544,
"learning_rate": 0.0008,
"loss": 2.256655216217041,
"step": 4172
},
{
"epoch": 0.25296969696969696,
"grad_norm": 0.059749506413936615,
"learning_rate": 0.0008,
"loss": 2.3522400856018066,
"step": 4174
},
{
"epoch": 0.2530909090909091,
"grad_norm": 0.05735812336206436,
"learning_rate": 0.0008,
"loss": 2.3366944789886475,
"step": 4176
},
{
"epoch": 0.2532121212121212,
"grad_norm": 0.05503764748573303,
"learning_rate": 0.0008,
"loss": 2.23744535446167,
"step": 4178
},
{
"epoch": 0.25333333333333335,
"grad_norm": 0.0545646995306015,
"learning_rate": 0.0008,
"loss": 2.2351043224334717,
"step": 4180
},
{
"epoch": 0.25345454545454543,
"grad_norm": 0.05866941809654236,
"learning_rate": 0.0008,
"loss": 2.3339765071868896,
"step": 4182
},
{
"epoch": 0.25357575757575757,
"grad_norm": 0.058561697602272034,
"learning_rate": 0.0008,
"loss": 2.283374547958374,
"step": 4184
},
{
"epoch": 0.2536969696969697,
"grad_norm": 0.060208242386579514,
"learning_rate": 0.0008,
"loss": 2.400604248046875,
"step": 4186
},
{
"epoch": 0.25381818181818183,
"grad_norm": 0.05761896073818207,
"learning_rate": 0.0008,
"loss": 2.32677960395813,
"step": 4188
},
{
"epoch": 0.25393939393939396,
"grad_norm": 0.056925542652606964,
"learning_rate": 0.0008,
"loss": 2.3617851734161377,
"step": 4190
},
{
"epoch": 0.25406060606060604,
"grad_norm": 0.05591524764895439,
"learning_rate": 0.0008,
"loss": 2.3383073806762695,
"step": 4192
},
{
"epoch": 0.2541818181818182,
"grad_norm": 0.05571106821298599,
"learning_rate": 0.0008,
"loss": 2.281059503555298,
"step": 4194
},
{
"epoch": 0.2543030303030303,
"grad_norm": 0.05675478279590607,
"learning_rate": 0.0008,
"loss": 2.3294458389282227,
"step": 4196
},
{
"epoch": 0.25442424242424244,
"grad_norm": 0.05888626351952553,
"learning_rate": 0.0008,
"loss": 2.3085741996765137,
"step": 4198
},
{
"epoch": 0.2545454545454545,
"grad_norm": 0.056861843913793564,
"learning_rate": 0.0008,
"loss": 2.3872368335723877,
"step": 4200
},
{
"epoch": 0.25466666666666665,
"grad_norm": 0.05430280789732933,
"learning_rate": 0.0008,
"loss": 2.279229164123535,
"step": 4202
},
{
"epoch": 0.2547878787878788,
"grad_norm": 0.05511534586548805,
"learning_rate": 0.0008,
"loss": 2.272477626800537,
"step": 4204
},
{
"epoch": 0.2549090909090909,
"grad_norm": 0.05574113503098488,
"learning_rate": 0.0008,
"loss": 2.3679065704345703,
"step": 4206
},
{
"epoch": 0.25503030303030305,
"grad_norm": 0.05980075150728226,
"learning_rate": 0.0008,
"loss": 2.3541781902313232,
"step": 4208
},
{
"epoch": 0.25515151515151513,
"grad_norm": 0.054730866104364395,
"learning_rate": 0.0008,
"loss": 2.277813673019409,
"step": 4210
},
{
"epoch": 0.25527272727272726,
"grad_norm": 0.05589761212468147,
"learning_rate": 0.0008,
"loss": 2.341710329055786,
"step": 4212
},
{
"epoch": 0.2553939393939394,
"grad_norm": 0.05805794149637222,
"learning_rate": 0.0008,
"loss": 2.3640384674072266,
"step": 4214
},
{
"epoch": 0.25551515151515153,
"grad_norm": 0.056781355291604996,
"learning_rate": 0.0008,
"loss": 2.3683981895446777,
"step": 4216
},
{
"epoch": 0.25563636363636366,
"grad_norm": 0.05414292588829994,
"learning_rate": 0.0008,
"loss": 2.321899890899658,
"step": 4218
},
{
"epoch": 0.25575757575757574,
"grad_norm": 0.055929284542798996,
"learning_rate": 0.0008,
"loss": 2.322610855102539,
"step": 4220
},
{
"epoch": 0.25587878787878787,
"grad_norm": 0.05690184235572815,
"learning_rate": 0.0008,
"loss": 2.3497962951660156,
"step": 4222
},
{
"epoch": 0.256,
"grad_norm": 0.056055840104818344,
"learning_rate": 0.0008,
"loss": 2.3252310752868652,
"step": 4224
},
{
"epoch": 0.25612121212121214,
"grad_norm": 0.0618160143494606,
"learning_rate": 0.0008,
"loss": 2.352323293685913,
"step": 4226
},
{
"epoch": 0.2562424242424242,
"grad_norm": 0.05740390717983246,
"learning_rate": 0.0008,
"loss": 2.2857789993286133,
"step": 4228
},
{
"epoch": 0.25636363636363635,
"grad_norm": 0.06569648534059525,
"learning_rate": 0.0008,
"loss": 2.3283746242523193,
"step": 4230
},
{
"epoch": 0.2564848484848485,
"grad_norm": 0.06229054555296898,
"learning_rate": 0.0008,
"loss": 2.3130173683166504,
"step": 4232
},
{
"epoch": 0.2566060606060606,
"grad_norm": 0.06264207512140274,
"learning_rate": 0.0008,
"loss": 2.301839590072632,
"step": 4234
},
{
"epoch": 0.25672727272727275,
"grad_norm": 0.05815274640917778,
"learning_rate": 0.0008,
"loss": 2.2874255180358887,
"step": 4236
},
{
"epoch": 0.2568484848484848,
"grad_norm": 0.055635035037994385,
"learning_rate": 0.0008,
"loss": 2.282374382019043,
"step": 4238
},
{
"epoch": 0.25696969696969696,
"grad_norm": 0.05759519711136818,
"learning_rate": 0.0008,
"loss": 2.3015522956848145,
"step": 4240
},
{
"epoch": 0.2570909090909091,
"grad_norm": 0.060060765594244,
"learning_rate": 0.0008,
"loss": 2.2561726570129395,
"step": 4242
},
{
"epoch": 0.2572121212121212,
"grad_norm": 0.06374011188745499,
"learning_rate": 0.0008,
"loss": 2.31095027923584,
"step": 4244
},
{
"epoch": 0.25733333333333336,
"grad_norm": 0.06497907638549805,
"learning_rate": 0.0008,
"loss": 2.3348801136016846,
"step": 4246
},
{
"epoch": 0.25745454545454544,
"grad_norm": 0.0697486624121666,
"learning_rate": 0.0008,
"loss": 2.3624074459075928,
"step": 4248
},
{
"epoch": 0.25757575757575757,
"grad_norm": 0.06896919757127762,
"learning_rate": 0.0008,
"loss": 2.3707239627838135,
"step": 4250
},
{
"epoch": 0.2576969696969697,
"grad_norm": 0.05961034074425697,
"learning_rate": 0.0008,
"loss": 2.2750244140625,
"step": 4252
},
{
"epoch": 0.25781818181818184,
"grad_norm": 0.0617859847843647,
"learning_rate": 0.0008,
"loss": 2.3342278003692627,
"step": 4254
},
{
"epoch": 0.2579393939393939,
"grad_norm": 0.06287611275911331,
"learning_rate": 0.0008,
"loss": 2.3110196590423584,
"step": 4256
},
{
"epoch": 0.25806060606060605,
"grad_norm": 0.0638209730386734,
"learning_rate": 0.0008,
"loss": 2.332153797149658,
"step": 4258
},
{
"epoch": 0.2581818181818182,
"grad_norm": 0.057309191673994064,
"learning_rate": 0.0008,
"loss": 2.4058730602264404,
"step": 4260
},
{
"epoch": 0.2583030303030303,
"grad_norm": 0.054415177553892136,
"learning_rate": 0.0008,
"loss": 2.29413104057312,
"step": 4262
},
{
"epoch": 0.25842424242424245,
"grad_norm": 0.055903203785419464,
"learning_rate": 0.0008,
"loss": 2.288361072540283,
"step": 4264
},
{
"epoch": 0.2585454545454545,
"grad_norm": 0.05618906393647194,
"learning_rate": 0.0008,
"loss": 2.270815849304199,
"step": 4266
},
{
"epoch": 0.25866666666666666,
"grad_norm": 0.05481981858611107,
"learning_rate": 0.0008,
"loss": 2.30847430229187,
"step": 4268
},
{
"epoch": 0.2587878787878788,
"grad_norm": 0.0547417588531971,
"learning_rate": 0.0008,
"loss": 2.3686296939849854,
"step": 4270
},
{
"epoch": 0.2589090909090909,
"grad_norm": 0.05604785680770874,
"learning_rate": 0.0008,
"loss": 2.340155601501465,
"step": 4272
},
{
"epoch": 0.25903030303030306,
"grad_norm": 0.05870114639401436,
"learning_rate": 0.0008,
"loss": 2.2731759548187256,
"step": 4274
},
{
"epoch": 0.25915151515151513,
"grad_norm": 0.057838961482048035,
"learning_rate": 0.0008,
"loss": 2.3712141513824463,
"step": 4276
},
{
"epoch": 0.25927272727272727,
"grad_norm": 0.05772295594215393,
"learning_rate": 0.0008,
"loss": 2.3474466800689697,
"step": 4278
},
{
"epoch": 0.2593939393939394,
"grad_norm": 0.0622720792889595,
"learning_rate": 0.0008,
"loss": 2.2890326976776123,
"step": 4280
},
{
"epoch": 0.25951515151515153,
"grad_norm": 0.05503302067518234,
"learning_rate": 0.0008,
"loss": 2.2934646606445312,
"step": 4282
},
{
"epoch": 0.2596363636363636,
"grad_norm": 0.05621762573719025,
"learning_rate": 0.0008,
"loss": 2.377910852432251,
"step": 4284
},
{
"epoch": 0.25975757575757574,
"grad_norm": 0.05381322652101517,
"learning_rate": 0.0008,
"loss": 2.3143434524536133,
"step": 4286
},
{
"epoch": 0.2598787878787879,
"grad_norm": 0.055681850761175156,
"learning_rate": 0.0008,
"loss": 2.29799747467041,
"step": 4288
},
{
"epoch": 0.26,
"grad_norm": 0.05553799867630005,
"learning_rate": 0.0008,
"loss": 2.3460168838500977,
"step": 4290
},
{
"epoch": 0.26012121212121214,
"grad_norm": 0.05441885069012642,
"learning_rate": 0.0008,
"loss": 2.3033447265625,
"step": 4292
},
{
"epoch": 0.2602424242424242,
"grad_norm": 0.05563442409038544,
"learning_rate": 0.0008,
"loss": 2.3364248275756836,
"step": 4294
},
{
"epoch": 0.26036363636363635,
"grad_norm": 0.06326998025178909,
"learning_rate": 0.0008,
"loss": 2.357935667037964,
"step": 4296
},
{
"epoch": 0.2604848484848485,
"grad_norm": 0.05490720644593239,
"learning_rate": 0.0008,
"loss": 2.3286972045898438,
"step": 4298
},
{
"epoch": 0.2606060606060606,
"grad_norm": 0.05577677860856056,
"learning_rate": 0.0008,
"loss": 2.236208915710449,
"step": 4300
},
{
"epoch": 0.26072727272727275,
"grad_norm": 0.05760166794061661,
"learning_rate": 0.0008,
"loss": 2.359118938446045,
"step": 4302
},
{
"epoch": 0.26084848484848483,
"grad_norm": 0.06259392946958542,
"learning_rate": 0.0008,
"loss": 2.2779173851013184,
"step": 4304
},
{
"epoch": 0.26096969696969696,
"grad_norm": 0.05766137316823006,
"learning_rate": 0.0008,
"loss": 2.335425615310669,
"step": 4306
},
{
"epoch": 0.2610909090909091,
"grad_norm": 0.0577441044151783,
"learning_rate": 0.0008,
"loss": 2.3527965545654297,
"step": 4308
},
{
"epoch": 0.26121212121212123,
"grad_norm": 0.05444977805018425,
"learning_rate": 0.0008,
"loss": 2.284224510192871,
"step": 4310
},
{
"epoch": 0.2613333333333333,
"grad_norm": 0.05591145530343056,
"learning_rate": 0.0008,
"loss": 2.322556734085083,
"step": 4312
},
{
"epoch": 0.26145454545454544,
"grad_norm": 0.055309537798166275,
"learning_rate": 0.0008,
"loss": 2.3614070415496826,
"step": 4314
},
{
"epoch": 0.2615757575757576,
"grad_norm": 0.056182000786066055,
"learning_rate": 0.0008,
"loss": 2.249640703201294,
"step": 4316
},
{
"epoch": 0.2616969696969697,
"grad_norm": 0.059864334762096405,
"learning_rate": 0.0008,
"loss": 2.2953872680664062,
"step": 4318
},
{
"epoch": 0.26181818181818184,
"grad_norm": 0.060440853238105774,
"learning_rate": 0.0008,
"loss": 2.4168572425842285,
"step": 4320
},
{
"epoch": 0.2619393939393939,
"grad_norm": 0.05850349739193916,
"learning_rate": 0.0008,
"loss": 2.339043378829956,
"step": 4322
},
{
"epoch": 0.26206060606060605,
"grad_norm": 0.06013832986354828,
"learning_rate": 0.0008,
"loss": 2.368027687072754,
"step": 4324
},
{
"epoch": 0.2621818181818182,
"grad_norm": 0.055758439004421234,
"learning_rate": 0.0008,
"loss": 2.296452522277832,
"step": 4326
},
{
"epoch": 0.2623030303030303,
"grad_norm": 0.05686220899224281,
"learning_rate": 0.0008,
"loss": 2.3284006118774414,
"step": 4328
},
{
"epoch": 0.26242424242424245,
"grad_norm": 0.05612155795097351,
"learning_rate": 0.0008,
"loss": 2.2454798221588135,
"step": 4330
},
{
"epoch": 0.2625454545454545,
"grad_norm": 0.05710737034678459,
"learning_rate": 0.0008,
"loss": 2.3864622116088867,
"step": 4332
},
{
"epoch": 0.26266666666666666,
"grad_norm": 0.0596865639090538,
"learning_rate": 0.0008,
"loss": 2.2644946575164795,
"step": 4334
},
{
"epoch": 0.2627878787878788,
"grad_norm": 0.057357363402843475,
"learning_rate": 0.0008,
"loss": 2.2979393005371094,
"step": 4336
},
{
"epoch": 0.2629090909090909,
"grad_norm": 0.05579742416739464,
"learning_rate": 0.0008,
"loss": 2.2894034385681152,
"step": 4338
},
{
"epoch": 0.263030303030303,
"grad_norm": 0.05662556737661362,
"learning_rate": 0.0008,
"loss": 2.3357486724853516,
"step": 4340
},
{
"epoch": 0.26315151515151514,
"grad_norm": 0.056189801543951035,
"learning_rate": 0.0008,
"loss": 2.2746496200561523,
"step": 4342
},
{
"epoch": 0.26327272727272727,
"grad_norm": 0.05607850104570389,
"learning_rate": 0.0008,
"loss": 2.2704083919525146,
"step": 4344
},
{
"epoch": 0.2633939393939394,
"grad_norm": 0.05562647059559822,
"learning_rate": 0.0008,
"loss": 2.2440199851989746,
"step": 4346
},
{
"epoch": 0.26351515151515154,
"grad_norm": 0.05765749514102936,
"learning_rate": 0.0008,
"loss": 2.260124921798706,
"step": 4348
},
{
"epoch": 0.2636363636363636,
"grad_norm": 0.05546452850103378,
"learning_rate": 0.0008,
"loss": 2.2718915939331055,
"step": 4350
},
{
"epoch": 0.26375757575757575,
"grad_norm": 0.05510798096656799,
"learning_rate": 0.0008,
"loss": 2.2904577255249023,
"step": 4352
},
{
"epoch": 0.2638787878787879,
"grad_norm": 0.055440399795770645,
"learning_rate": 0.0008,
"loss": 2.3170828819274902,
"step": 4354
},
{
"epoch": 0.264,
"grad_norm": 0.055160120129585266,
"learning_rate": 0.0008,
"loss": 2.2706222534179688,
"step": 4356
},
{
"epoch": 0.26412121212121215,
"grad_norm": 0.05435779318213463,
"learning_rate": 0.0008,
"loss": 2.2562472820281982,
"step": 4358
},
{
"epoch": 0.2642424242424242,
"grad_norm": 0.05511787533760071,
"learning_rate": 0.0008,
"loss": 2.299335479736328,
"step": 4360
},
{
"epoch": 0.26436363636363636,
"grad_norm": 0.055900104343891144,
"learning_rate": 0.0008,
"loss": 2.3356540203094482,
"step": 4362
},
{
"epoch": 0.2644848484848485,
"grad_norm": 0.05687551572918892,
"learning_rate": 0.0008,
"loss": 2.2783687114715576,
"step": 4364
},
{
"epoch": 0.2646060606060606,
"grad_norm": 0.055902667343616486,
"learning_rate": 0.0008,
"loss": 2.291616201400757,
"step": 4366
},
{
"epoch": 0.2647272727272727,
"grad_norm": 0.056907426565885544,
"learning_rate": 0.0008,
"loss": 2.2615129947662354,
"step": 4368
},
{
"epoch": 0.26484848484848483,
"grad_norm": 0.05750015750527382,
"learning_rate": 0.0008,
"loss": 2.2673869132995605,
"step": 4370
},
{
"epoch": 0.26496969696969697,
"grad_norm": 0.060904983431100845,
"learning_rate": 0.0008,
"loss": 2.3777527809143066,
"step": 4372
},
{
"epoch": 0.2650909090909091,
"grad_norm": 0.05451042577624321,
"learning_rate": 0.0008,
"loss": 2.2933707237243652,
"step": 4374
},
{
"epoch": 0.26521212121212123,
"grad_norm": 0.05863112211227417,
"learning_rate": 0.0008,
"loss": 2.284985303878784,
"step": 4376
},
{
"epoch": 0.2653333333333333,
"grad_norm": 0.05504077672958374,
"learning_rate": 0.0008,
"loss": 2.3280937671661377,
"step": 4378
},
{
"epoch": 0.26545454545454544,
"grad_norm": 0.057485394179821014,
"learning_rate": 0.0008,
"loss": 2.285851001739502,
"step": 4380
},
{
"epoch": 0.2655757575757576,
"grad_norm": 0.05482746660709381,
"learning_rate": 0.0008,
"loss": 2.299494743347168,
"step": 4382
},
{
"epoch": 0.2656969696969697,
"grad_norm": 0.05491326004266739,
"learning_rate": 0.0008,
"loss": 2.3160533905029297,
"step": 4384
},
{
"epoch": 0.26581818181818184,
"grad_norm": 0.05553668737411499,
"learning_rate": 0.0008,
"loss": 2.3788583278656006,
"step": 4386
},
{
"epoch": 0.2659393939393939,
"grad_norm": 0.05617121607065201,
"learning_rate": 0.0008,
"loss": 2.351022481918335,
"step": 4388
},
{
"epoch": 0.26606060606060605,
"grad_norm": 0.057379286736249924,
"learning_rate": 0.0008,
"loss": 2.3241209983825684,
"step": 4390
},
{
"epoch": 0.2661818181818182,
"grad_norm": 0.05544310063123703,
"learning_rate": 0.0008,
"loss": 2.335630178451538,
"step": 4392
},
{
"epoch": 0.2663030303030303,
"grad_norm": 0.05629166215658188,
"learning_rate": 0.0008,
"loss": 2.4173436164855957,
"step": 4394
},
{
"epoch": 0.2664242424242424,
"grad_norm": 0.06594951450824738,
"learning_rate": 0.0008,
"loss": 2.329554319381714,
"step": 4396
},
{
"epoch": 0.26654545454545453,
"grad_norm": 0.061623405665159225,
"learning_rate": 0.0008,
"loss": 2.2713334560394287,
"step": 4398
},
{
"epoch": 0.26666666666666666,
"grad_norm": 0.05813000351190567,
"learning_rate": 0.0008,
"loss": 2.311405658721924,
"step": 4400
},
{
"epoch": 0.2667878787878788,
"grad_norm": 0.0570761077105999,
"learning_rate": 0.0008,
"loss": 2.231105089187622,
"step": 4402
},
{
"epoch": 0.26690909090909093,
"grad_norm": 0.05629724636673927,
"learning_rate": 0.0008,
"loss": 2.3192861080169678,
"step": 4404
},
{
"epoch": 0.267030303030303,
"grad_norm": 0.05966106429696083,
"learning_rate": 0.0008,
"loss": 2.2758755683898926,
"step": 4406
},
{
"epoch": 0.26715151515151514,
"grad_norm": 0.0548529326915741,
"learning_rate": 0.0008,
"loss": 2.3425636291503906,
"step": 4408
},
{
"epoch": 0.2672727272727273,
"grad_norm": 0.05449633672833443,
"learning_rate": 0.0008,
"loss": 2.3448526859283447,
"step": 4410
},
{
"epoch": 0.2673939393939394,
"grad_norm": 0.056486498564481735,
"learning_rate": 0.0008,
"loss": 2.2922096252441406,
"step": 4412
},
{
"epoch": 0.26751515151515154,
"grad_norm": 0.056034453213214874,
"learning_rate": 0.0008,
"loss": 2.33062481880188,
"step": 4414
},
{
"epoch": 0.2676363636363636,
"grad_norm": 0.058028824627399445,
"learning_rate": 0.0008,
"loss": 2.286266565322876,
"step": 4416
},
{
"epoch": 0.26775757575757575,
"grad_norm": 0.061083100736141205,
"learning_rate": 0.0008,
"loss": 2.322835922241211,
"step": 4418
},
{
"epoch": 0.2678787878787879,
"grad_norm": 0.0690770074725151,
"learning_rate": 0.0008,
"loss": 2.333631992340088,
"step": 4420
},
{
"epoch": 0.268,
"grad_norm": 0.06654415279626846,
"learning_rate": 0.0008,
"loss": 2.2591192722320557,
"step": 4422
},
{
"epoch": 0.2681212121212121,
"grad_norm": 0.06089257076382637,
"learning_rate": 0.0008,
"loss": 2.323967695236206,
"step": 4424
},
{
"epoch": 0.2682424242424242,
"grad_norm": 0.06275535374879837,
"learning_rate": 0.0008,
"loss": 2.3018388748168945,
"step": 4426
},
{
"epoch": 0.26836363636363636,
"grad_norm": 0.05570146441459656,
"learning_rate": 0.0008,
"loss": 2.335879325866699,
"step": 4428
},
{
"epoch": 0.2684848484848485,
"grad_norm": 0.0574457123875618,
"learning_rate": 0.0008,
"loss": 2.263704299926758,
"step": 4430
},
{
"epoch": 0.2686060606060606,
"grad_norm": 0.05816537141799927,
"learning_rate": 0.0008,
"loss": 2.263063430786133,
"step": 4432
},
{
"epoch": 0.2687272727272727,
"grad_norm": 0.06037929654121399,
"learning_rate": 0.0008,
"loss": 2.3603522777557373,
"step": 4434
},
{
"epoch": 0.26884848484848484,
"grad_norm": 0.058139048516750336,
"learning_rate": 0.0008,
"loss": 2.276779890060425,
"step": 4436
},
{
"epoch": 0.26896969696969697,
"grad_norm": 0.06184757873415947,
"learning_rate": 0.0008,
"loss": 2.283045530319214,
"step": 4438
},
{
"epoch": 0.2690909090909091,
"grad_norm": 0.06046583503484726,
"learning_rate": 0.0008,
"loss": 2.3350141048431396,
"step": 4440
},
{
"epoch": 0.26921212121212124,
"grad_norm": 0.06265371292829514,
"learning_rate": 0.0008,
"loss": 2.340869665145874,
"step": 4442
},
{
"epoch": 0.2693333333333333,
"grad_norm": 0.05927056074142456,
"learning_rate": 0.0008,
"loss": 2.213787794113159,
"step": 4444
},
{
"epoch": 0.26945454545454545,
"grad_norm": 0.05632511153817177,
"learning_rate": 0.0008,
"loss": 2.2683205604553223,
"step": 4446
},
{
"epoch": 0.2695757575757576,
"grad_norm": 0.05658271908760071,
"learning_rate": 0.0008,
"loss": 2.202610731124878,
"step": 4448
},
{
"epoch": 0.2696969696969697,
"grad_norm": 0.0549091175198555,
"learning_rate": 0.0008,
"loss": 2.300187587738037,
"step": 4450
},
{
"epoch": 0.26981818181818185,
"grad_norm": 0.056642547249794006,
"learning_rate": 0.0008,
"loss": 2.2751898765563965,
"step": 4452
},
{
"epoch": 0.2699393939393939,
"grad_norm": 0.054272983223199844,
"learning_rate": 0.0008,
"loss": 2.285466432571411,
"step": 4454
},
{
"epoch": 0.27006060606060606,
"grad_norm": 0.055064063519239426,
"learning_rate": 0.0008,
"loss": 2.3249363899230957,
"step": 4456
},
{
"epoch": 0.2701818181818182,
"grad_norm": 0.05510306730866432,
"learning_rate": 0.0008,
"loss": 2.2728583812713623,
"step": 4458
},
{
"epoch": 0.2703030303030303,
"grad_norm": 0.05381877347826958,
"learning_rate": 0.0008,
"loss": 2.2859671115875244,
"step": 4460
},
{
"epoch": 0.2704242424242424,
"grad_norm": 0.0540739931166172,
"learning_rate": 0.0008,
"loss": 2.3315670490264893,
"step": 4462
},
{
"epoch": 0.27054545454545453,
"grad_norm": 0.05868004262447357,
"learning_rate": 0.0008,
"loss": 2.3146228790283203,
"step": 4464
},
{
"epoch": 0.27066666666666667,
"grad_norm": 0.055460650473833084,
"learning_rate": 0.0008,
"loss": 2.304168462753296,
"step": 4466
},
{
"epoch": 0.2707878787878788,
"grad_norm": 0.054302118718624115,
"learning_rate": 0.0008,
"loss": 2.266071319580078,
"step": 4468
},
{
"epoch": 0.27090909090909093,
"grad_norm": 0.05713856220245361,
"learning_rate": 0.0008,
"loss": 2.418792724609375,
"step": 4470
},
{
"epoch": 0.271030303030303,
"grad_norm": 0.054921284317970276,
"learning_rate": 0.0008,
"loss": 2.2935643196105957,
"step": 4472
},
{
"epoch": 0.27115151515151514,
"grad_norm": 0.05473969131708145,
"learning_rate": 0.0008,
"loss": 2.2383832931518555,
"step": 4474
},
{
"epoch": 0.2712727272727273,
"grad_norm": 0.055265892297029495,
"learning_rate": 0.0008,
"loss": 2.2663137912750244,
"step": 4476
},
{
"epoch": 0.2713939393939394,
"grad_norm": 0.0560644306242466,
"learning_rate": 0.0008,
"loss": 2.338989019393921,
"step": 4478
},
{
"epoch": 0.27151515151515154,
"grad_norm": 0.05848073586821556,
"learning_rate": 0.0008,
"loss": 2.259679079055786,
"step": 4480
},
{
"epoch": 0.2716363636363636,
"grad_norm": 0.05602727830410004,
"learning_rate": 0.0008,
"loss": 2.3460962772369385,
"step": 4482
},
{
"epoch": 0.27175757575757575,
"grad_norm": 0.05963916704058647,
"learning_rate": 0.0008,
"loss": 2.338251829147339,
"step": 4484
},
{
"epoch": 0.2718787878787879,
"grad_norm": 0.05473119020462036,
"learning_rate": 0.0008,
"loss": 2.2361812591552734,
"step": 4486
},
{
"epoch": 0.272,
"grad_norm": 0.05632244795560837,
"learning_rate": 0.0008,
"loss": 2.3219704627990723,
"step": 4488
},
{
"epoch": 0.2721212121212121,
"grad_norm": 0.05743682384490967,
"learning_rate": 0.0008,
"loss": 2.3507864475250244,
"step": 4490
},
{
"epoch": 0.27224242424242423,
"grad_norm": 0.05668110027909279,
"learning_rate": 0.0008,
"loss": 2.2996158599853516,
"step": 4492
},
{
"epoch": 0.27236363636363636,
"grad_norm": 0.05944516137242317,
"learning_rate": 0.0008,
"loss": 2.3043384552001953,
"step": 4494
},
{
"epoch": 0.2724848484848485,
"grad_norm": 0.05492531880736351,
"learning_rate": 0.0008,
"loss": 2.2570409774780273,
"step": 4496
},
{
"epoch": 0.27260606060606063,
"grad_norm": 0.05819449946284294,
"learning_rate": 0.0008,
"loss": 2.3615431785583496,
"step": 4498
},
{
"epoch": 0.2727272727272727,
"grad_norm": 0.052362531423568726,
"learning_rate": 0.0008,
"loss": 2.2644379138946533,
"step": 4500
},
{
"epoch": 0.27284848484848484,
"grad_norm": 0.05583372339606285,
"learning_rate": 0.0008,
"loss": 2.341691493988037,
"step": 4502
},
{
"epoch": 0.272969696969697,
"grad_norm": 0.05368338152766228,
"learning_rate": 0.0008,
"loss": 2.2677295207977295,
"step": 4504
},
{
"epoch": 0.2730909090909091,
"grad_norm": 0.05373888462781906,
"learning_rate": 0.0008,
"loss": 2.256135940551758,
"step": 4506
},
{
"epoch": 0.27321212121212124,
"grad_norm": 0.05435385927557945,
"learning_rate": 0.0008,
"loss": 2.289001226425171,
"step": 4508
},
{
"epoch": 0.2733333333333333,
"grad_norm": 0.05619831010699272,
"learning_rate": 0.0008,
"loss": 2.2866034507751465,
"step": 4510
},
{
"epoch": 0.27345454545454545,
"grad_norm": 0.054338257759809494,
"learning_rate": 0.0008,
"loss": 2.377121686935425,
"step": 4512
},
{
"epoch": 0.2735757575757576,
"grad_norm": 0.057151202112436295,
"learning_rate": 0.0008,
"loss": 2.318267583847046,
"step": 4514
},
{
"epoch": 0.2736969696969697,
"grad_norm": 0.055295269936323166,
"learning_rate": 0.0008,
"loss": 2.2654826641082764,
"step": 4516
},
{
"epoch": 0.2738181818181818,
"grad_norm": 0.055207423865795135,
"learning_rate": 0.0008,
"loss": 2.253031015396118,
"step": 4518
},
{
"epoch": 0.2739393939393939,
"grad_norm": 0.054739370942115784,
"learning_rate": 0.0008,
"loss": 2.2413949966430664,
"step": 4520
},
{
"epoch": 0.27406060606060606,
"grad_norm": 0.056064993143081665,
"learning_rate": 0.0008,
"loss": 2.259080410003662,
"step": 4522
},
{
"epoch": 0.2741818181818182,
"grad_norm": 0.05825633555650711,
"learning_rate": 0.0008,
"loss": 2.299001455307007,
"step": 4524
},
{
"epoch": 0.2743030303030303,
"grad_norm": 0.057286038994789124,
"learning_rate": 0.0008,
"loss": 2.282667398452759,
"step": 4526
},
{
"epoch": 0.2744242424242424,
"grad_norm": 0.05544722452759743,
"learning_rate": 0.0008,
"loss": 2.3126418590545654,
"step": 4528
},
{
"epoch": 0.27454545454545454,
"grad_norm": 0.056840721517801285,
"learning_rate": 0.0008,
"loss": 2.2945141792297363,
"step": 4530
},
{
"epoch": 0.27466666666666667,
"grad_norm": 0.06730905175209045,
"learning_rate": 0.0008,
"loss": 2.3873817920684814,
"step": 4532
},
{
"epoch": 0.2747878787878788,
"grad_norm": 0.05557776242494583,
"learning_rate": 0.0008,
"loss": 2.2681007385253906,
"step": 4534
},
{
"epoch": 0.27490909090909094,
"grad_norm": 0.05237122252583504,
"learning_rate": 0.0008,
"loss": 2.2651913166046143,
"step": 4536
},
{
"epoch": 0.275030303030303,
"grad_norm": 0.0551861971616745,
"learning_rate": 0.0008,
"loss": 2.3458220958709717,
"step": 4538
},
{
"epoch": 0.27515151515151515,
"grad_norm": 0.05319899320602417,
"learning_rate": 0.0008,
"loss": 2.2834768295288086,
"step": 4540
},
{
"epoch": 0.2752727272727273,
"grad_norm": 0.054759055376052856,
"learning_rate": 0.0008,
"loss": 2.290147542953491,
"step": 4542
},
{
"epoch": 0.2753939393939394,
"grad_norm": 0.05858103558421135,
"learning_rate": 0.0008,
"loss": 2.2598659992218018,
"step": 4544
},
{
"epoch": 0.2755151515151515,
"grad_norm": 0.05364903062582016,
"learning_rate": 0.0008,
"loss": 2.2834460735321045,
"step": 4546
},
{
"epoch": 0.2756363636363636,
"grad_norm": 0.05603271350264549,
"learning_rate": 0.0008,
"loss": 2.267739772796631,
"step": 4548
},
{
"epoch": 0.27575757575757576,
"grad_norm": 0.058283451944589615,
"learning_rate": 0.0008,
"loss": 2.380167007446289,
"step": 4550
},
{
"epoch": 0.2758787878787879,
"grad_norm": 0.057976607233285904,
"learning_rate": 0.0008,
"loss": 2.334597587585449,
"step": 4552
},
{
"epoch": 0.276,
"grad_norm": 0.057005710899829865,
"learning_rate": 0.0008,
"loss": 2.2801833152770996,
"step": 4554
},
{
"epoch": 0.2761212121212121,
"grad_norm": 0.056253500282764435,
"learning_rate": 0.0008,
"loss": 2.260707378387451,
"step": 4556
},
{
"epoch": 0.27624242424242423,
"grad_norm": 0.07027553021907806,
"learning_rate": 0.0008,
"loss": 2.31195330619812,
"step": 4558
},
{
"epoch": 0.27636363636363637,
"grad_norm": 0.06384358555078506,
"learning_rate": 0.0008,
"loss": 2.345432758331299,
"step": 4560
},
{
"epoch": 0.2764848484848485,
"grad_norm": 0.0585150420665741,
"learning_rate": 0.0008,
"loss": 2.2566909790039062,
"step": 4562
},
{
"epoch": 0.27660606060606063,
"grad_norm": 0.05975113809108734,
"learning_rate": 0.0008,
"loss": 2.274639844894409,
"step": 4564
},
{
"epoch": 0.2767272727272727,
"grad_norm": 0.055555231869220734,
"learning_rate": 0.0008,
"loss": 2.382713794708252,
"step": 4566
},
{
"epoch": 0.27684848484848484,
"grad_norm": 0.052927613258361816,
"learning_rate": 0.0008,
"loss": 2.2779908180236816,
"step": 4568
},
{
"epoch": 0.276969696969697,
"grad_norm": 0.05448170751333237,
"learning_rate": 0.0008,
"loss": 2.326643228530884,
"step": 4570
},
{
"epoch": 0.2770909090909091,
"grad_norm": 0.05488854646682739,
"learning_rate": 0.0008,
"loss": 2.28167462348938,
"step": 4572
},
{
"epoch": 0.2772121212121212,
"grad_norm": 0.05355045199394226,
"learning_rate": 0.0008,
"loss": 2.2580606937408447,
"step": 4574
},
{
"epoch": 0.2773333333333333,
"grad_norm": 0.05489703640341759,
"learning_rate": 0.0008,
"loss": 2.2866828441619873,
"step": 4576
},
{
"epoch": 0.27745454545454545,
"grad_norm": 0.05332602933049202,
"learning_rate": 0.0008,
"loss": 2.24198317527771,
"step": 4578
},
{
"epoch": 0.2775757575757576,
"grad_norm": 0.05336891487240791,
"learning_rate": 0.0008,
"loss": 2.2367184162139893,
"step": 4580
},
{
"epoch": 0.2776969696969697,
"grad_norm": 0.05534444749355316,
"learning_rate": 0.0008,
"loss": 2.3387296199798584,
"step": 4582
},
{
"epoch": 0.2778181818181818,
"grad_norm": 0.053866203874349594,
"learning_rate": 0.0008,
"loss": 2.3007869720458984,
"step": 4584
},
{
"epoch": 0.27793939393939393,
"grad_norm": 0.05586421117186546,
"learning_rate": 0.0008,
"loss": 2.3029415607452393,
"step": 4586
},
{
"epoch": 0.27806060606060606,
"grad_norm": 0.051511313766241074,
"learning_rate": 0.0008,
"loss": 2.3013534545898438,
"step": 4588
},
{
"epoch": 0.2781818181818182,
"grad_norm": 0.05458967015147209,
"learning_rate": 0.0008,
"loss": 2.2247138023376465,
"step": 4590
},
{
"epoch": 0.27830303030303033,
"grad_norm": 0.056070782244205475,
"learning_rate": 0.0008,
"loss": 2.365159273147583,
"step": 4592
},
{
"epoch": 0.2784242424242424,
"grad_norm": 0.056724853813648224,
"learning_rate": 0.0008,
"loss": 2.2961156368255615,
"step": 4594
},
{
"epoch": 0.27854545454545454,
"grad_norm": 0.054196182638406754,
"learning_rate": 0.0008,
"loss": 2.290943145751953,
"step": 4596
},
{
"epoch": 0.2786666666666667,
"grad_norm": 0.05384622514247894,
"learning_rate": 0.0008,
"loss": 2.2959678173065186,
"step": 4598
},
{
"epoch": 0.2787878787878788,
"grad_norm": 0.05365370213985443,
"learning_rate": 0.0008,
"loss": 2.2308952808380127,
"step": 4600
},
{
"epoch": 0.2789090909090909,
"grad_norm": 0.05660124123096466,
"learning_rate": 0.0008,
"loss": 2.3204615116119385,
"step": 4602
},
{
"epoch": 0.279030303030303,
"grad_norm": 0.05970025435090065,
"learning_rate": 0.0008,
"loss": 2.30749249458313,
"step": 4604
},
{
"epoch": 0.27915151515151515,
"grad_norm": 0.053729891777038574,
"learning_rate": 0.0008,
"loss": 2.269469976425171,
"step": 4606
},
{
"epoch": 0.2792727272727273,
"grad_norm": 0.05600248649716377,
"learning_rate": 0.0008,
"loss": 2.261716842651367,
"step": 4608
},
{
"epoch": 0.2793939393939394,
"grad_norm": 0.05390690639615059,
"learning_rate": 0.0008,
"loss": 2.2816827297210693,
"step": 4610
},
{
"epoch": 0.2795151515151515,
"grad_norm": 0.05523274838924408,
"learning_rate": 0.0008,
"loss": 2.2962806224823,
"step": 4612
},
{
"epoch": 0.2796363636363636,
"grad_norm": 0.054051924496889114,
"learning_rate": 0.0008,
"loss": 2.2770349979400635,
"step": 4614
},
{
"epoch": 0.27975757575757576,
"grad_norm": 0.05514691397547722,
"learning_rate": 0.0008,
"loss": 2.311007499694824,
"step": 4616
},
{
"epoch": 0.2798787878787879,
"grad_norm": 0.05605493485927582,
"learning_rate": 0.0008,
"loss": 2.349321126937866,
"step": 4618
},
{
"epoch": 0.28,
"grad_norm": 0.058174602687358856,
"learning_rate": 0.0008,
"loss": 2.277482509613037,
"step": 4620
},
{
"epoch": 0.2801212121212121,
"grad_norm": 0.054451826959848404,
"learning_rate": 0.0008,
"loss": 2.281430244445801,
"step": 4622
},
{
"epoch": 0.28024242424242424,
"grad_norm": 0.05681096017360687,
"learning_rate": 0.0008,
"loss": 2.2752134799957275,
"step": 4624
},
{
"epoch": 0.28036363636363637,
"grad_norm": 0.05428697541356087,
"learning_rate": 0.0008,
"loss": 2.3062386512756348,
"step": 4626
},
{
"epoch": 0.2804848484848485,
"grad_norm": 0.05642550066113472,
"learning_rate": 0.0008,
"loss": 2.2855379581451416,
"step": 4628
},
{
"epoch": 0.2806060606060606,
"grad_norm": 0.05335645750164986,
"learning_rate": 0.0008,
"loss": 2.2884914875030518,
"step": 4630
},
{
"epoch": 0.2807272727272727,
"grad_norm": 0.05774903669953346,
"learning_rate": 0.0008,
"loss": 2.2853777408599854,
"step": 4632
},
{
"epoch": 0.28084848484848485,
"grad_norm": 0.055051956325769424,
"learning_rate": 0.0008,
"loss": 2.314424991607666,
"step": 4634
},
{
"epoch": 0.280969696969697,
"grad_norm": 0.05437519773840904,
"learning_rate": 0.0008,
"loss": 2.212987184524536,
"step": 4636
},
{
"epoch": 0.2810909090909091,
"grad_norm": 0.05705779418349266,
"learning_rate": 0.0008,
"loss": 2.270890235900879,
"step": 4638
},
{
"epoch": 0.2812121212121212,
"grad_norm": 0.055935923010110855,
"learning_rate": 0.0008,
"loss": 2.293299436569214,
"step": 4640
},
{
"epoch": 0.2813333333333333,
"grad_norm": 0.05573229491710663,
"learning_rate": 0.0008,
"loss": 2.330489158630371,
"step": 4642
},
{
"epoch": 0.28145454545454546,
"grad_norm": 0.05469466745853424,
"learning_rate": 0.0008,
"loss": 2.280174732208252,
"step": 4644
},
{
"epoch": 0.2815757575757576,
"grad_norm": 0.055617764592170715,
"learning_rate": 0.0008,
"loss": 2.2556402683258057,
"step": 4646
},
{
"epoch": 0.2816969696969697,
"grad_norm": 0.056385431438684464,
"learning_rate": 0.0008,
"loss": 2.355471611022949,
"step": 4648
},
{
"epoch": 0.2818181818181818,
"grad_norm": 0.05487361550331116,
"learning_rate": 0.0008,
"loss": 2.377854824066162,
"step": 4650
},
{
"epoch": 0.28193939393939393,
"grad_norm": 0.05418841540813446,
"learning_rate": 0.0008,
"loss": 2.358874797821045,
"step": 4652
},
{
"epoch": 0.28206060606060607,
"grad_norm": 0.054295577108860016,
"learning_rate": 0.0008,
"loss": 2.300501585006714,
"step": 4654
},
{
"epoch": 0.2821818181818182,
"grad_norm": 0.055122483521699905,
"learning_rate": 0.0008,
"loss": 2.344276189804077,
"step": 4656
},
{
"epoch": 0.2823030303030303,
"grad_norm": 0.05558707192540169,
"learning_rate": 0.0008,
"loss": 2.2393605709075928,
"step": 4658
},
{
"epoch": 0.2824242424242424,
"grad_norm": 0.055209267884492874,
"learning_rate": 0.0008,
"loss": 2.2562649250030518,
"step": 4660
},
{
"epoch": 0.28254545454545454,
"grad_norm": 0.05391997843980789,
"learning_rate": 0.0008,
"loss": 2.2977449893951416,
"step": 4662
},
{
"epoch": 0.2826666666666667,
"grad_norm": 0.05468451604247093,
"learning_rate": 0.0008,
"loss": 2.2150542736053467,
"step": 4664
},
{
"epoch": 0.2827878787878788,
"grad_norm": 0.053801655769348145,
"learning_rate": 0.0008,
"loss": 2.23134183883667,
"step": 4666
},
{
"epoch": 0.2829090909090909,
"grad_norm": 0.054412227123975754,
"learning_rate": 0.0008,
"loss": 2.2484869956970215,
"step": 4668
},
{
"epoch": 0.283030303030303,
"grad_norm": 0.056372083723545074,
"learning_rate": 0.0008,
"loss": 2.2156550884246826,
"step": 4670
},
{
"epoch": 0.28315151515151515,
"grad_norm": 0.056354787200689316,
"learning_rate": 0.0008,
"loss": 2.2584102153778076,
"step": 4672
},
{
"epoch": 0.2832727272727273,
"grad_norm": 0.05433962121605873,
"learning_rate": 0.0008,
"loss": 2.245120048522949,
"step": 4674
},
{
"epoch": 0.2833939393939394,
"grad_norm": 0.05540841817855835,
"learning_rate": 0.0008,
"loss": 2.1810500621795654,
"step": 4676
},
{
"epoch": 0.2835151515151515,
"grad_norm": 0.05647530034184456,
"learning_rate": 0.0008,
"loss": 2.3001677989959717,
"step": 4678
},
{
"epoch": 0.28363636363636363,
"grad_norm": 0.05409097671508789,
"learning_rate": 0.0008,
"loss": 2.3189520835876465,
"step": 4680
},
{
"epoch": 0.28375757575757576,
"grad_norm": 0.05225584656000137,
"learning_rate": 0.0008,
"loss": 2.3566932678222656,
"step": 4682
},
{
"epoch": 0.2838787878787879,
"grad_norm": 0.060842115432024,
"learning_rate": 0.0008,
"loss": 2.3278191089630127,
"step": 4684
},
{
"epoch": 0.284,
"grad_norm": 0.05829113349318504,
"learning_rate": 0.0008,
"loss": 2.310246467590332,
"step": 4686
},
{
"epoch": 0.2841212121212121,
"grad_norm": 0.06196233630180359,
"learning_rate": 0.0008,
"loss": 2.2644834518432617,
"step": 4688
},
{
"epoch": 0.28424242424242424,
"grad_norm": 0.06200785934925079,
"learning_rate": 0.0008,
"loss": 2.2010104656219482,
"step": 4690
},
{
"epoch": 0.2843636363636364,
"grad_norm": 0.06878723949193954,
"learning_rate": 0.0008,
"loss": 2.2502923011779785,
"step": 4692
},
{
"epoch": 0.2844848484848485,
"grad_norm": 0.06470175832509995,
"learning_rate": 0.0008,
"loss": 2.1857707500457764,
"step": 4694
},
{
"epoch": 0.2846060606060606,
"grad_norm": 0.062308214604854584,
"learning_rate": 0.0008,
"loss": 2.1942501068115234,
"step": 4696
},
{
"epoch": 0.2847272727272727,
"grad_norm": 0.0531114861369133,
"learning_rate": 0.0008,
"loss": 2.296896457672119,
"step": 4698
},
{
"epoch": 0.28484848484848485,
"grad_norm": 0.05554616078734398,
"learning_rate": 0.0008,
"loss": 2.2936432361602783,
"step": 4700
},
{
"epoch": 0.284969696969697,
"grad_norm": 0.054572656750679016,
"learning_rate": 0.0008,
"loss": 2.2823069095611572,
"step": 4702
},
{
"epoch": 0.2850909090909091,
"grad_norm": 0.05597798153758049,
"learning_rate": 0.0008,
"loss": 2.2546253204345703,
"step": 4704
},
{
"epoch": 0.2852121212121212,
"grad_norm": 0.061293505132198334,
"learning_rate": 0.0008,
"loss": 2.2340636253356934,
"step": 4706
},
{
"epoch": 0.2853333333333333,
"grad_norm": 0.052963320165872574,
"learning_rate": 0.0008,
"loss": 2.262117385864258,
"step": 4708
},
{
"epoch": 0.28545454545454546,
"grad_norm": 0.05516951531171799,
"learning_rate": 0.0008,
"loss": 2.2496190071105957,
"step": 4710
},
{
"epoch": 0.2855757575757576,
"grad_norm": 0.05589240416884422,
"learning_rate": 0.0008,
"loss": 2.306231737136841,
"step": 4712
},
{
"epoch": 0.28569696969696967,
"grad_norm": 0.05905822291970253,
"learning_rate": 0.0008,
"loss": 2.2571423053741455,
"step": 4714
},
{
"epoch": 0.2858181818181818,
"grad_norm": 0.06786094605922699,
"learning_rate": 0.0008,
"loss": 2.269880533218384,
"step": 4716
},
{
"epoch": 0.28593939393939394,
"grad_norm": 0.06577903032302856,
"learning_rate": 0.0008,
"loss": 2.2611794471740723,
"step": 4718
},
{
"epoch": 0.28606060606060607,
"grad_norm": 0.059658233076334,
"learning_rate": 0.0008,
"loss": 2.341372013092041,
"step": 4720
},
{
"epoch": 0.2861818181818182,
"grad_norm": 0.05681830644607544,
"learning_rate": 0.0008,
"loss": 2.2807044982910156,
"step": 4722
},
{
"epoch": 0.2863030303030303,
"grad_norm": 0.05459382012486458,
"learning_rate": 0.0008,
"loss": 2.346247673034668,
"step": 4724
},
{
"epoch": 0.2864242424242424,
"grad_norm": 0.05237637832760811,
"learning_rate": 0.0008,
"loss": 2.281033992767334,
"step": 4726
},
{
"epoch": 0.28654545454545455,
"grad_norm": 0.05617569014430046,
"learning_rate": 0.0008,
"loss": 2.262033700942993,
"step": 4728
},
{
"epoch": 0.2866666666666667,
"grad_norm": 0.05935538560152054,
"learning_rate": 0.0008,
"loss": 2.307227373123169,
"step": 4730
},
{
"epoch": 0.2867878787878788,
"grad_norm": 0.05719233676791191,
"learning_rate": 0.0008,
"loss": 2.2527730464935303,
"step": 4732
},
{
"epoch": 0.2869090909090909,
"grad_norm": 0.05530914291739464,
"learning_rate": 0.0008,
"loss": 2.2869203090667725,
"step": 4734
},
{
"epoch": 0.287030303030303,
"grad_norm": 0.053372498601675034,
"learning_rate": 0.0008,
"loss": 2.257380723953247,
"step": 4736
},
{
"epoch": 0.28715151515151516,
"grad_norm": 0.05570876598358154,
"learning_rate": 0.0008,
"loss": 2.3500354290008545,
"step": 4738
},
{
"epoch": 0.2872727272727273,
"grad_norm": 0.05364924669265747,
"learning_rate": 0.0008,
"loss": 2.1910483837127686,
"step": 4740
},
{
"epoch": 0.28739393939393937,
"grad_norm": 0.060286711901426315,
"learning_rate": 0.0008,
"loss": 2.2812328338623047,
"step": 4742
},
{
"epoch": 0.2875151515151515,
"grad_norm": 0.05501750111579895,
"learning_rate": 0.0008,
"loss": 2.308928966522217,
"step": 4744
},
{
"epoch": 0.28763636363636363,
"grad_norm": 0.0573563277721405,
"learning_rate": 0.0008,
"loss": 2.3335440158843994,
"step": 4746
},
{
"epoch": 0.28775757575757577,
"grad_norm": 0.053601279854774475,
"learning_rate": 0.0008,
"loss": 2.3178937435150146,
"step": 4748
},
{
"epoch": 0.2878787878787879,
"grad_norm": 0.054257266223430634,
"learning_rate": 0.0008,
"loss": 2.2960715293884277,
"step": 4750
},
{
"epoch": 0.288,
"grad_norm": 0.05575403943657875,
"learning_rate": 0.0008,
"loss": 2.2941930294036865,
"step": 4752
},
{
"epoch": 0.2881212121212121,
"grad_norm": 0.05487275496125221,
"learning_rate": 0.0008,
"loss": 2.288525342941284,
"step": 4754
},
{
"epoch": 0.28824242424242424,
"grad_norm": 0.05387573689222336,
"learning_rate": 0.0008,
"loss": 2.3342533111572266,
"step": 4756
},
{
"epoch": 0.2883636363636364,
"grad_norm": 0.054857831448316574,
"learning_rate": 0.0008,
"loss": 2.2665278911590576,
"step": 4758
},
{
"epoch": 0.2884848484848485,
"grad_norm": 0.05643100291490555,
"learning_rate": 0.0008,
"loss": 2.1866962909698486,
"step": 4760
},
{
"epoch": 0.2886060606060606,
"grad_norm": 0.05491183325648308,
"learning_rate": 0.0008,
"loss": 2.310840129852295,
"step": 4762
},
{
"epoch": 0.2887272727272727,
"grad_norm": 0.05809604376554489,
"learning_rate": 0.0008,
"loss": 2.27667498588562,
"step": 4764
},
{
"epoch": 0.28884848484848485,
"grad_norm": 0.05596885830163956,
"learning_rate": 0.0008,
"loss": 2.2711737155914307,
"step": 4766
},
{
"epoch": 0.288969696969697,
"grad_norm": 0.053733885288238525,
"learning_rate": 0.0008,
"loss": 2.3843672275543213,
"step": 4768
},
{
"epoch": 0.28909090909090907,
"grad_norm": 0.05530449002981186,
"learning_rate": 0.0008,
"loss": 2.2390480041503906,
"step": 4770
},
{
"epoch": 0.2892121212121212,
"grad_norm": 0.05367610976099968,
"learning_rate": 0.0008,
"loss": 2.271472215652466,
"step": 4772
},
{
"epoch": 0.28933333333333333,
"grad_norm": 0.0550968237221241,
"learning_rate": 0.0008,
"loss": 2.268082618713379,
"step": 4774
},
{
"epoch": 0.28945454545454546,
"grad_norm": 0.05347451567649841,
"learning_rate": 0.0008,
"loss": 2.223788261413574,
"step": 4776
},
{
"epoch": 0.2895757575757576,
"grad_norm": 0.05908353254199028,
"learning_rate": 0.0008,
"loss": 2.2246227264404297,
"step": 4778
},
{
"epoch": 0.2896969696969697,
"grad_norm": 0.05593821406364441,
"learning_rate": 0.0008,
"loss": 2.2636520862579346,
"step": 4780
},
{
"epoch": 0.2898181818181818,
"grad_norm": 0.05654796585440636,
"learning_rate": 0.0008,
"loss": 2.254469871520996,
"step": 4782
},
{
"epoch": 0.28993939393939394,
"grad_norm": 0.0545928031206131,
"learning_rate": 0.0008,
"loss": 2.2249398231506348,
"step": 4784
},
{
"epoch": 0.2900606060606061,
"grad_norm": 0.052820634096860886,
"learning_rate": 0.0008,
"loss": 2.2484130859375,
"step": 4786
},
{
"epoch": 0.2901818181818182,
"grad_norm": 0.05386760085821152,
"learning_rate": 0.0008,
"loss": 2.2864763736724854,
"step": 4788
},
{
"epoch": 0.2903030303030303,
"grad_norm": 0.05664343759417534,
"learning_rate": 0.0008,
"loss": 2.2853448390960693,
"step": 4790
},
{
"epoch": 0.2904242424242424,
"grad_norm": 0.05446917191147804,
"learning_rate": 0.0008,
"loss": 2.349919319152832,
"step": 4792
},
{
"epoch": 0.29054545454545455,
"grad_norm": 0.05720677226781845,
"learning_rate": 0.0008,
"loss": 2.397477388381958,
"step": 4794
},
{
"epoch": 0.2906666666666667,
"grad_norm": 0.05324156954884529,
"learning_rate": 0.0008,
"loss": 2.2922027111053467,
"step": 4796
},
{
"epoch": 0.29078787878787876,
"grad_norm": 0.05390412360429764,
"learning_rate": 0.0008,
"loss": 2.236441135406494,
"step": 4798
},
{
"epoch": 0.2909090909090909,
"grad_norm": 0.05422819033265114,
"learning_rate": 0.0008,
"loss": 2.302670955657959,
"step": 4800
},
{
"epoch": 0.29103030303030303,
"grad_norm": 0.05637020245194435,
"learning_rate": 0.0008,
"loss": 2.254502534866333,
"step": 4802
},
{
"epoch": 0.29115151515151516,
"grad_norm": 0.05390185862779617,
"learning_rate": 0.0008,
"loss": 2.360191822052002,
"step": 4804
},
{
"epoch": 0.2912727272727273,
"grad_norm": 0.05615033581852913,
"learning_rate": 0.0008,
"loss": 2.4006574153900146,
"step": 4806
},
{
"epoch": 0.29139393939393937,
"grad_norm": 0.057728253304958344,
"learning_rate": 0.0008,
"loss": 2.338367462158203,
"step": 4808
},
{
"epoch": 0.2915151515151515,
"grad_norm": 0.057007912546396255,
"learning_rate": 0.0008,
"loss": 2.3011012077331543,
"step": 4810
},
{
"epoch": 0.29163636363636364,
"grad_norm": 0.054072968661785126,
"learning_rate": 0.0008,
"loss": 2.2724814414978027,
"step": 4812
},
{
"epoch": 0.29175757575757577,
"grad_norm": 0.0564652644097805,
"learning_rate": 0.0008,
"loss": 2.18826961517334,
"step": 4814
},
{
"epoch": 0.2918787878787879,
"grad_norm": 0.054671190679073334,
"learning_rate": 0.0008,
"loss": 2.2887320518493652,
"step": 4816
},
{
"epoch": 0.292,
"grad_norm": 0.05508033186197281,
"learning_rate": 0.0008,
"loss": 2.2871971130371094,
"step": 4818
},
{
"epoch": 0.2921212121212121,
"grad_norm": 0.05275227501988411,
"learning_rate": 0.0008,
"loss": 2.2408926486968994,
"step": 4820
},
{
"epoch": 0.29224242424242425,
"grad_norm": 0.05506299063563347,
"learning_rate": 0.0008,
"loss": 2.243800163269043,
"step": 4822
},
{
"epoch": 0.2923636363636364,
"grad_norm": 0.058799903839826584,
"learning_rate": 0.0008,
"loss": 2.1883175373077393,
"step": 4824
},
{
"epoch": 0.29248484848484846,
"grad_norm": 0.05424119532108307,
"learning_rate": 0.0008,
"loss": 2.373542547225952,
"step": 4826
},
{
"epoch": 0.2926060606060606,
"grad_norm": 0.0568314790725708,
"learning_rate": 0.0008,
"loss": 2.265829086303711,
"step": 4828
},
{
"epoch": 0.2927272727272727,
"grad_norm": 0.0527508370578289,
"learning_rate": 0.0008,
"loss": 2.271395206451416,
"step": 4830
},
{
"epoch": 0.29284848484848486,
"grad_norm": 0.05373500660061836,
"learning_rate": 0.0008,
"loss": 2.2753844261169434,
"step": 4832
},
{
"epoch": 0.292969696969697,
"grad_norm": 0.05158925801515579,
"learning_rate": 0.0008,
"loss": 2.2330262660980225,
"step": 4834
},
{
"epoch": 0.29309090909090907,
"grad_norm": 0.05250396952033043,
"learning_rate": 0.0008,
"loss": 2.229161024093628,
"step": 4836
},
{
"epoch": 0.2932121212121212,
"grad_norm": 0.05737446993589401,
"learning_rate": 0.0008,
"loss": 2.265381097793579,
"step": 4838
},
{
"epoch": 0.29333333333333333,
"grad_norm": 0.061819907277822495,
"learning_rate": 0.0008,
"loss": 2.302894115447998,
"step": 4840
},
{
"epoch": 0.29345454545454547,
"grad_norm": 0.05942108482122421,
"learning_rate": 0.0008,
"loss": 2.279719114303589,
"step": 4842
},
{
"epoch": 0.2935757575757576,
"grad_norm": 0.05644964799284935,
"learning_rate": 0.0008,
"loss": 2.228940486907959,
"step": 4844
},
{
"epoch": 0.2936969696969697,
"grad_norm": 0.05474438890814781,
"learning_rate": 0.0008,
"loss": 2.2454674243927,
"step": 4846
},
{
"epoch": 0.2938181818181818,
"grad_norm": 0.05287795886397362,
"learning_rate": 0.0008,
"loss": 2.2697131633758545,
"step": 4848
},
{
"epoch": 0.29393939393939394,
"grad_norm": 0.054473914206027985,
"learning_rate": 0.0008,
"loss": 2.260518789291382,
"step": 4850
},
{
"epoch": 0.2940606060606061,
"grad_norm": 0.056441549211740494,
"learning_rate": 0.0008,
"loss": 2.2611052989959717,
"step": 4852
},
{
"epoch": 0.29418181818181816,
"grad_norm": 0.05172049626708031,
"learning_rate": 0.0008,
"loss": 2.261385440826416,
"step": 4854
},
{
"epoch": 0.2943030303030303,
"grad_norm": 0.05301522836089134,
"learning_rate": 0.0008,
"loss": 2.240074396133423,
"step": 4856
},
{
"epoch": 0.2944242424242424,
"grad_norm": 0.053613126277923584,
"learning_rate": 0.0008,
"loss": 2.254446029663086,
"step": 4858
},
{
"epoch": 0.29454545454545455,
"grad_norm": 0.05630803853273392,
"learning_rate": 0.0008,
"loss": 2.2889537811279297,
"step": 4860
},
{
"epoch": 0.2946666666666667,
"grad_norm": 0.0523088276386261,
"learning_rate": 0.0008,
"loss": 2.209427833557129,
"step": 4862
},
{
"epoch": 0.29478787878787877,
"grad_norm": 0.054167717695236206,
"learning_rate": 0.0008,
"loss": 2.192564010620117,
"step": 4864
},
{
"epoch": 0.2949090909090909,
"grad_norm": 0.05412587895989418,
"learning_rate": 0.0008,
"loss": 2.233938217163086,
"step": 4866
},
{
"epoch": 0.29503030303030303,
"grad_norm": 0.052139054983854294,
"learning_rate": 0.0008,
"loss": 2.255357503890991,
"step": 4868
},
{
"epoch": 0.29515151515151516,
"grad_norm": 0.05194449797272682,
"learning_rate": 0.0008,
"loss": 2.2717137336730957,
"step": 4870
},
{
"epoch": 0.2952727272727273,
"grad_norm": 0.054211970418691635,
"learning_rate": 0.0008,
"loss": 2.1991188526153564,
"step": 4872
},
{
"epoch": 0.2953939393939394,
"grad_norm": 0.05371275544166565,
"learning_rate": 0.0008,
"loss": 2.2485504150390625,
"step": 4874
},
{
"epoch": 0.2955151515151515,
"grad_norm": 0.056722939014434814,
"learning_rate": 0.0008,
"loss": 2.2276387214660645,
"step": 4876
},
{
"epoch": 0.29563636363636364,
"grad_norm": 0.05243941396474838,
"learning_rate": 0.0008,
"loss": 2.342050790786743,
"step": 4878
},
{
"epoch": 0.2957575757575758,
"grad_norm": 0.05393067002296448,
"learning_rate": 0.0008,
"loss": 2.297621965408325,
"step": 4880
},
{
"epoch": 0.29587878787878785,
"grad_norm": 0.05563806742429733,
"learning_rate": 0.0008,
"loss": 2.291226863861084,
"step": 4882
},
{
"epoch": 0.296,
"grad_norm": 0.05408297851681709,
"learning_rate": 0.0008,
"loss": 2.2713119983673096,
"step": 4884
},
{
"epoch": 0.2961212121212121,
"grad_norm": 0.05365954339504242,
"learning_rate": 0.0008,
"loss": 2.330564498901367,
"step": 4886
},
{
"epoch": 0.29624242424242425,
"grad_norm": 0.056111399084329605,
"learning_rate": 0.0008,
"loss": 2.3294458389282227,
"step": 4888
},
{
"epoch": 0.2963636363636364,
"grad_norm": 0.05629744008183479,
"learning_rate": 0.0008,
"loss": 2.3094677925109863,
"step": 4890
},
{
"epoch": 0.29648484848484846,
"grad_norm": 0.05357637628912926,
"learning_rate": 0.0008,
"loss": 2.2358896732330322,
"step": 4892
},
{
"epoch": 0.2966060606060606,
"grad_norm": 0.05376078188419342,
"learning_rate": 0.0008,
"loss": 2.383877992630005,
"step": 4894
},
{
"epoch": 0.29672727272727273,
"grad_norm": 0.054002080112695694,
"learning_rate": 0.0008,
"loss": 2.274066686630249,
"step": 4896
},
{
"epoch": 0.29684848484848486,
"grad_norm": 0.05502447485923767,
"learning_rate": 0.0008,
"loss": 2.2798376083374023,
"step": 4898
},
{
"epoch": 0.296969696969697,
"grad_norm": 0.05441747233271599,
"learning_rate": 0.0008,
"loss": 2.2438488006591797,
"step": 4900
},
{
"epoch": 0.29709090909090907,
"grad_norm": 0.05227840691804886,
"learning_rate": 0.0008,
"loss": 2.233762264251709,
"step": 4902
},
{
"epoch": 0.2972121212121212,
"grad_norm": 0.0536375530064106,
"learning_rate": 0.0008,
"loss": 2.265068292617798,
"step": 4904
},
{
"epoch": 0.29733333333333334,
"grad_norm": 0.05492078512907028,
"learning_rate": 0.0008,
"loss": 2.2509372234344482,
"step": 4906
},
{
"epoch": 0.29745454545454547,
"grad_norm": 0.05377637967467308,
"learning_rate": 0.0008,
"loss": 2.2441213130950928,
"step": 4908
},
{
"epoch": 0.29757575757575755,
"grad_norm": 0.05574806407094002,
"learning_rate": 0.0008,
"loss": 2.3523788452148438,
"step": 4910
},
{
"epoch": 0.2976969696969697,
"grad_norm": 0.05469955503940582,
"learning_rate": 0.0008,
"loss": 2.347738027572632,
"step": 4912
},
{
"epoch": 0.2978181818181818,
"grad_norm": 0.061578117311000824,
"learning_rate": 0.0008,
"loss": 2.277494192123413,
"step": 4914
},
{
"epoch": 0.29793939393939395,
"grad_norm": 0.05529829114675522,
"learning_rate": 0.0008,
"loss": 2.2174062728881836,
"step": 4916
},
{
"epoch": 0.2980606060606061,
"grad_norm": 0.053765732795000076,
"learning_rate": 0.0008,
"loss": 2.3003671169281006,
"step": 4918
},
{
"epoch": 0.29818181818181816,
"grad_norm": 0.05339481309056282,
"learning_rate": 0.0008,
"loss": 2.1993398666381836,
"step": 4920
},
{
"epoch": 0.2983030303030303,
"grad_norm": 0.05606154724955559,
"learning_rate": 0.0008,
"loss": 2.2379393577575684,
"step": 4922
},
{
"epoch": 0.2984242424242424,
"grad_norm": 0.05595916509628296,
"learning_rate": 0.0008,
"loss": 2.2111332416534424,
"step": 4924
},
{
"epoch": 0.29854545454545456,
"grad_norm": 0.05740240216255188,
"learning_rate": 0.0008,
"loss": 2.3122992515563965,
"step": 4926
},
{
"epoch": 0.2986666666666667,
"grad_norm": 0.06078239902853966,
"learning_rate": 0.0008,
"loss": 2.295248031616211,
"step": 4928
},
{
"epoch": 0.29878787878787877,
"grad_norm": 0.057778362184762955,
"learning_rate": 0.0008,
"loss": 2.218975782394409,
"step": 4930
},
{
"epoch": 0.2989090909090909,
"grad_norm": 0.057347528636455536,
"learning_rate": 0.0008,
"loss": 2.262341260910034,
"step": 4932
},
{
"epoch": 0.29903030303030304,
"grad_norm": 0.053604986518621445,
"learning_rate": 0.0008,
"loss": 2.277545690536499,
"step": 4934
},
{
"epoch": 0.29915151515151517,
"grad_norm": 0.05310754477977753,
"learning_rate": 0.0008,
"loss": 2.2764534950256348,
"step": 4936
},
{
"epoch": 0.29927272727272725,
"grad_norm": 0.05433344841003418,
"learning_rate": 0.0008,
"loss": 2.335740804672241,
"step": 4938
},
{
"epoch": 0.2993939393939394,
"grad_norm": 0.05318735912442207,
"learning_rate": 0.0008,
"loss": 2.230260133743286,
"step": 4940
},
{
"epoch": 0.2995151515151515,
"grad_norm": 0.053097862750291824,
"learning_rate": 0.0008,
"loss": 2.2347874641418457,
"step": 4942
},
{
"epoch": 0.29963636363636365,
"grad_norm": 0.05405677109956741,
"learning_rate": 0.0008,
"loss": 2.2529425621032715,
"step": 4944
},
{
"epoch": 0.2997575757575758,
"grad_norm": 0.054108504205942154,
"learning_rate": 0.0008,
"loss": 2.316314935684204,
"step": 4946
},
{
"epoch": 0.29987878787878786,
"grad_norm": 0.06374972313642502,
"learning_rate": 0.0008,
"loss": 2.272117853164673,
"step": 4948
},
{
"epoch": 0.3,
"grad_norm": 0.058928124606609344,
"learning_rate": 0.0008,
"loss": 2.2701776027679443,
"step": 4950
},
{
"epoch": 0.3001212121212121,
"grad_norm": 0.0532425232231617,
"learning_rate": 0.0008,
"loss": 2.3056509494781494,
"step": 4952
},
{
"epoch": 0.30024242424242426,
"grad_norm": 0.052035823464393616,
"learning_rate": 0.0008,
"loss": 2.1708133220672607,
"step": 4954
},
{
"epoch": 0.3003636363636364,
"grad_norm": 0.0554540678858757,
"learning_rate": 0.0008,
"loss": 2.265753746032715,
"step": 4956
},
{
"epoch": 0.30048484848484847,
"grad_norm": 0.05161741003394127,
"learning_rate": 0.0008,
"loss": 2.178636312484741,
"step": 4958
},
{
"epoch": 0.3006060606060606,
"grad_norm": 0.05749468505382538,
"learning_rate": 0.0008,
"loss": 2.2840747833251953,
"step": 4960
},
{
"epoch": 0.30072727272727273,
"grad_norm": 0.05530589818954468,
"learning_rate": 0.0008,
"loss": 2.2059574127197266,
"step": 4962
},
{
"epoch": 0.30084848484848487,
"grad_norm": 0.06192709505558014,
"learning_rate": 0.0008,
"loss": 2.256516695022583,
"step": 4964
},
{
"epoch": 0.30096969696969694,
"grad_norm": 0.06051327660679817,
"learning_rate": 0.0008,
"loss": 2.2183141708374023,
"step": 4966
},
{
"epoch": 0.3010909090909091,
"grad_norm": 0.06127804145216942,
"learning_rate": 0.0008,
"loss": 2.300039291381836,
"step": 4968
},
{
"epoch": 0.3012121212121212,
"grad_norm": 0.055958908051252365,
"learning_rate": 0.0008,
"loss": 2.253392457962036,
"step": 4970
},
{
"epoch": 0.30133333333333334,
"grad_norm": 0.05586221069097519,
"learning_rate": 0.0008,
"loss": 2.243908405303955,
"step": 4972
},
{
"epoch": 0.3014545454545455,
"grad_norm": 0.05609317496418953,
"learning_rate": 0.0008,
"loss": 2.231538772583008,
"step": 4974
},
{
"epoch": 0.30157575757575755,
"grad_norm": 0.058466631919145584,
"learning_rate": 0.0008,
"loss": 2.2387049198150635,
"step": 4976
},
{
"epoch": 0.3016969696969697,
"grad_norm": 0.05503614619374275,
"learning_rate": 0.0008,
"loss": 2.2997922897338867,
"step": 4978
},
{
"epoch": 0.3018181818181818,
"grad_norm": 0.051841191947460175,
"learning_rate": 0.0008,
"loss": 2.168401002883911,
"step": 4980
},
{
"epoch": 0.30193939393939395,
"grad_norm": 0.05500571429729462,
"learning_rate": 0.0008,
"loss": 2.283348321914673,
"step": 4982
},
{
"epoch": 0.3020606060606061,
"grad_norm": 0.056693185120821,
"learning_rate": 0.0008,
"loss": 2.2652807235717773,
"step": 4984
},
{
"epoch": 0.30218181818181816,
"grad_norm": 0.05780131369829178,
"learning_rate": 0.0008,
"loss": 2.402055501937866,
"step": 4986
},
{
"epoch": 0.3023030303030303,
"grad_norm": 0.05481209605932236,
"learning_rate": 0.0008,
"loss": 2.252267360687256,
"step": 4988
},
{
"epoch": 0.30242424242424243,
"grad_norm": 0.054683033376932144,
"learning_rate": 0.0008,
"loss": 2.3153955936431885,
"step": 4990
},
{
"epoch": 0.30254545454545456,
"grad_norm": 0.0549120269715786,
"learning_rate": 0.0008,
"loss": 2.2285315990448,
"step": 4992
},
{
"epoch": 0.30266666666666664,
"grad_norm": 0.05461292713880539,
"learning_rate": 0.0008,
"loss": 2.279456377029419,
"step": 4994
},
{
"epoch": 0.3027878787878788,
"grad_norm": 0.05328740179538727,
"learning_rate": 0.0008,
"loss": 2.2267327308654785,
"step": 4996
},
{
"epoch": 0.3029090909090909,
"grad_norm": 0.05204177647829056,
"learning_rate": 0.0008,
"loss": 2.3238365650177,
"step": 4998
},
{
"epoch": 0.30303030303030304,
"grad_norm": 0.0535978302359581,
"learning_rate": 0.0008,
"loss": 2.334669351577759,
"step": 5000
},
{
"epoch": 0.30315151515151517,
"grad_norm": 0.053077831864356995,
"learning_rate": 0.0008,
"loss": 2.254889726638794,
"step": 5002
},
{
"epoch": 0.30327272727272725,
"grad_norm": 0.05378474295139313,
"learning_rate": 0.0008,
"loss": 2.311197519302368,
"step": 5004
},
{
"epoch": 0.3033939393939394,
"grad_norm": 0.05600333958864212,
"learning_rate": 0.0008,
"loss": 2.2842280864715576,
"step": 5006
},
{
"epoch": 0.3035151515151515,
"grad_norm": 0.05423963442444801,
"learning_rate": 0.0008,
"loss": 2.2714693546295166,
"step": 5008
},
{
"epoch": 0.30363636363636365,
"grad_norm": 0.05285424366593361,
"learning_rate": 0.0008,
"loss": 2.333055257797241,
"step": 5010
},
{
"epoch": 0.3037575757575758,
"grad_norm": 0.052807558327913284,
"learning_rate": 0.0008,
"loss": 2.2704017162323,
"step": 5012
},
{
"epoch": 0.30387878787878786,
"grad_norm": 0.05349867045879364,
"learning_rate": 0.0008,
"loss": 2.280057191848755,
"step": 5014
},
{
"epoch": 0.304,
"grad_norm": 0.05338842794299126,
"learning_rate": 0.0008,
"loss": 2.332848310470581,
"step": 5016
},
{
"epoch": 0.3041212121212121,
"grad_norm": 0.052908122539520264,
"learning_rate": 0.0008,
"loss": 2.248776435852051,
"step": 5018
},
{
"epoch": 0.30424242424242426,
"grad_norm": 0.05230122059583664,
"learning_rate": 0.0008,
"loss": 2.2155802249908447,
"step": 5020
},
{
"epoch": 0.30436363636363634,
"grad_norm": 0.05702662467956543,
"learning_rate": 0.0008,
"loss": 2.2990665435791016,
"step": 5022
},
{
"epoch": 0.30448484848484847,
"grad_norm": 0.053798168897628784,
"learning_rate": 0.0008,
"loss": 2.338630199432373,
"step": 5024
},
{
"epoch": 0.3046060606060606,
"grad_norm": 0.05323721840977669,
"learning_rate": 0.0008,
"loss": 2.2832024097442627,
"step": 5026
},
{
"epoch": 0.30472727272727274,
"grad_norm": 0.05317485332489014,
"learning_rate": 0.0008,
"loss": 2.235816478729248,
"step": 5028
},
{
"epoch": 0.30484848484848487,
"grad_norm": 0.054157670587301254,
"learning_rate": 0.0008,
"loss": 2.2120401859283447,
"step": 5030
},
{
"epoch": 0.30496969696969695,
"grad_norm": 0.05373736843466759,
"learning_rate": 0.0008,
"loss": 2.254624843597412,
"step": 5032
},
{
"epoch": 0.3050909090909091,
"grad_norm": 0.05406279116868973,
"learning_rate": 0.0008,
"loss": 2.297678232192993,
"step": 5034
},
{
"epoch": 0.3052121212121212,
"grad_norm": 0.053006626665592194,
"learning_rate": 0.0008,
"loss": 2.164088487625122,
"step": 5036
},
{
"epoch": 0.30533333333333335,
"grad_norm": 0.05513172969222069,
"learning_rate": 0.0008,
"loss": 2.3062992095947266,
"step": 5038
},
{
"epoch": 0.3054545454545455,
"grad_norm": 0.056745801120996475,
"learning_rate": 0.0008,
"loss": 2.2405099868774414,
"step": 5040
},
{
"epoch": 0.30557575757575756,
"grad_norm": 0.0590730756521225,
"learning_rate": 0.0008,
"loss": 2.265498638153076,
"step": 5042
},
{
"epoch": 0.3056969696969697,
"grad_norm": 0.053258445113897324,
"learning_rate": 0.0008,
"loss": 2.2991507053375244,
"step": 5044
},
{
"epoch": 0.3058181818181818,
"grad_norm": 0.0523897185921669,
"learning_rate": 0.0008,
"loss": 2.2384979724884033,
"step": 5046
},
{
"epoch": 0.30593939393939396,
"grad_norm": 0.0546933077275753,
"learning_rate": 0.0008,
"loss": 2.22880482673645,
"step": 5048
},
{
"epoch": 0.30606060606060603,
"grad_norm": 0.052717678248882294,
"learning_rate": 0.0008,
"loss": 2.3368165493011475,
"step": 5050
},
{
"epoch": 0.30618181818181817,
"grad_norm": 0.05550898239016533,
"learning_rate": 0.0008,
"loss": 2.310922861099243,
"step": 5052
},
{
"epoch": 0.3063030303030303,
"grad_norm": 0.05458122491836548,
"learning_rate": 0.0008,
"loss": 2.2277729511260986,
"step": 5054
},
{
"epoch": 0.30642424242424243,
"grad_norm": 0.05683701112866402,
"learning_rate": 0.0008,
"loss": 2.3476510047912598,
"step": 5056
},
{
"epoch": 0.30654545454545457,
"grad_norm": 0.055481042712926865,
"learning_rate": 0.0008,
"loss": 2.2605435848236084,
"step": 5058
},
{
"epoch": 0.30666666666666664,
"grad_norm": 0.05490657687187195,
"learning_rate": 0.0008,
"loss": 2.2554590702056885,
"step": 5060
},
{
"epoch": 0.3067878787878788,
"grad_norm": 0.0583437904715538,
"learning_rate": 0.0008,
"loss": 2.328857898712158,
"step": 5062
},
{
"epoch": 0.3069090909090909,
"grad_norm": 0.05138527974486351,
"learning_rate": 0.0008,
"loss": 2.1672849655151367,
"step": 5064
},
{
"epoch": 0.30703030303030304,
"grad_norm": 0.05599382147192955,
"learning_rate": 0.0008,
"loss": 2.3279693126678467,
"step": 5066
},
{
"epoch": 0.3071515151515152,
"grad_norm": 0.053198229521512985,
"learning_rate": 0.0008,
"loss": 2.2444393634796143,
"step": 5068
},
{
"epoch": 0.30727272727272725,
"grad_norm": 0.05641688033938408,
"learning_rate": 0.0008,
"loss": 2.278364658355713,
"step": 5070
},
{
"epoch": 0.3073939393939394,
"grad_norm": 0.05334009602665901,
"learning_rate": 0.0008,
"loss": 2.263478994369507,
"step": 5072
},
{
"epoch": 0.3075151515151515,
"grad_norm": 0.054407086223363876,
"learning_rate": 0.0008,
"loss": 2.264556884765625,
"step": 5074
},
{
"epoch": 0.30763636363636365,
"grad_norm": 0.05220150202512741,
"learning_rate": 0.0008,
"loss": 2.2518765926361084,
"step": 5076
},
{
"epoch": 0.3077575757575758,
"grad_norm": 0.056127119809389114,
"learning_rate": 0.0008,
"loss": 2.319220781326294,
"step": 5078
},
{
"epoch": 0.30787878787878786,
"grad_norm": 0.05199400335550308,
"learning_rate": 0.0008,
"loss": 2.231708288192749,
"step": 5080
},
{
"epoch": 0.308,
"grad_norm": 0.05307885259389877,
"learning_rate": 0.0008,
"loss": 2.305011749267578,
"step": 5082
},
{
"epoch": 0.30812121212121213,
"grad_norm": 0.052763596177101135,
"learning_rate": 0.0008,
"loss": 2.25946044921875,
"step": 5084
},
{
"epoch": 0.30824242424242426,
"grad_norm": 0.05106998234987259,
"learning_rate": 0.0008,
"loss": 2.2860519886016846,
"step": 5086
},
{
"epoch": 0.30836363636363634,
"grad_norm": 0.053581602871418,
"learning_rate": 0.0008,
"loss": 2.2750282287597656,
"step": 5088
},
{
"epoch": 0.3084848484848485,
"grad_norm": 0.052475620061159134,
"learning_rate": 0.0008,
"loss": 2.2096164226531982,
"step": 5090
},
{
"epoch": 0.3086060606060606,
"grad_norm": 0.05314183607697487,
"learning_rate": 0.0008,
"loss": 2.304520606994629,
"step": 5092
},
{
"epoch": 0.30872727272727274,
"grad_norm": 0.059590309858322144,
"learning_rate": 0.0008,
"loss": 2.3434953689575195,
"step": 5094
},
{
"epoch": 0.30884848484848487,
"grad_norm": 0.05446378514170647,
"learning_rate": 0.0008,
"loss": 2.2060909271240234,
"step": 5096
},
{
"epoch": 0.30896969696969695,
"grad_norm": 0.0550554133951664,
"learning_rate": 0.0008,
"loss": 2.2664926052093506,
"step": 5098
},
{
"epoch": 0.3090909090909091,
"grad_norm": 0.05286594480276108,
"learning_rate": 0.0008,
"loss": 2.248910427093506,
"step": 5100
},
{
"epoch": 0.3092121212121212,
"grad_norm": 0.052872080355882645,
"learning_rate": 0.0008,
"loss": 2.1852738857269287,
"step": 5102
},
{
"epoch": 0.30933333333333335,
"grad_norm": 0.051445167511701584,
"learning_rate": 0.0008,
"loss": 2.2851803302764893,
"step": 5104
},
{
"epoch": 0.3094545454545455,
"grad_norm": 0.0543697252869606,
"learning_rate": 0.0008,
"loss": 2.283940553665161,
"step": 5106
},
{
"epoch": 0.30957575757575756,
"grad_norm": 0.054856229573488235,
"learning_rate": 0.0008,
"loss": 2.286625623703003,
"step": 5108
},
{
"epoch": 0.3096969696969697,
"grad_norm": 0.05181875452399254,
"learning_rate": 0.0008,
"loss": 2.2789828777313232,
"step": 5110
},
{
"epoch": 0.3098181818181818,
"grad_norm": 0.051082905381917953,
"learning_rate": 0.0008,
"loss": 2.1700141429901123,
"step": 5112
},
{
"epoch": 0.30993939393939396,
"grad_norm": 0.052807554602622986,
"learning_rate": 0.0008,
"loss": 2.301614761352539,
"step": 5114
},
{
"epoch": 0.31006060606060604,
"grad_norm": 0.05252240598201752,
"learning_rate": 0.0008,
"loss": 2.2442057132720947,
"step": 5116
},
{
"epoch": 0.31018181818181817,
"grad_norm": 0.05466960370540619,
"learning_rate": 0.0008,
"loss": 2.2839367389678955,
"step": 5118
},
{
"epoch": 0.3103030303030303,
"grad_norm": 0.052947044372558594,
"learning_rate": 0.0008,
"loss": 2.213625192642212,
"step": 5120
},
{
"epoch": 0.31042424242424244,
"grad_norm": 0.05478903651237488,
"learning_rate": 0.0008,
"loss": 2.226149082183838,
"step": 5122
},
{
"epoch": 0.31054545454545457,
"grad_norm": 0.053755708038806915,
"learning_rate": 0.0008,
"loss": 2.218336582183838,
"step": 5124
},
{
"epoch": 0.31066666666666665,
"grad_norm": 0.05421191081404686,
"learning_rate": 0.0008,
"loss": 2.270271062850952,
"step": 5126
},
{
"epoch": 0.3107878787878788,
"grad_norm": 0.05245404690504074,
"learning_rate": 0.0008,
"loss": 2.2537052631378174,
"step": 5128
},
{
"epoch": 0.3109090909090909,
"grad_norm": 0.05168338492512703,
"learning_rate": 0.0008,
"loss": 2.2403016090393066,
"step": 5130
},
{
"epoch": 0.31103030303030305,
"grad_norm": 0.05370907858014107,
"learning_rate": 0.0008,
"loss": 2.3733503818511963,
"step": 5132
},
{
"epoch": 0.3111515151515152,
"grad_norm": 0.053792569786310196,
"learning_rate": 0.0008,
"loss": 2.345569133758545,
"step": 5134
},
{
"epoch": 0.31127272727272726,
"grad_norm": 0.05238606035709381,
"learning_rate": 0.0008,
"loss": 2.2127203941345215,
"step": 5136
},
{
"epoch": 0.3113939393939394,
"grad_norm": 0.05897875502705574,
"learning_rate": 0.0008,
"loss": 2.287423610687256,
"step": 5138
},
{
"epoch": 0.3115151515151515,
"grad_norm": 0.06137946993112564,
"learning_rate": 0.0008,
"loss": 2.2477710247039795,
"step": 5140
},
{
"epoch": 0.31163636363636366,
"grad_norm": 0.05962546169757843,
"learning_rate": 0.0008,
"loss": 2.2057816982269287,
"step": 5142
},
{
"epoch": 0.31175757575757573,
"grad_norm": 0.05552038550376892,
"learning_rate": 0.0008,
"loss": 2.208259344100952,
"step": 5144
},
{
"epoch": 0.31187878787878787,
"grad_norm": 0.05648447945713997,
"learning_rate": 0.0008,
"loss": 2.2516539096832275,
"step": 5146
},
{
"epoch": 0.312,
"grad_norm": 0.05407373979687691,
"learning_rate": 0.0008,
"loss": 2.263554811477661,
"step": 5148
},
{
"epoch": 0.31212121212121213,
"grad_norm": 0.055779192596673965,
"learning_rate": 0.0008,
"loss": 2.306926727294922,
"step": 5150
},
{
"epoch": 0.31224242424242427,
"grad_norm": 0.05822879821062088,
"learning_rate": 0.0008,
"loss": 2.2879910469055176,
"step": 5152
},
{
"epoch": 0.31236363636363634,
"grad_norm": 0.055379025638103485,
"learning_rate": 0.0008,
"loss": 2.222379446029663,
"step": 5154
},
{
"epoch": 0.3124848484848485,
"grad_norm": 0.0529063418507576,
"learning_rate": 0.0008,
"loss": 2.287830114364624,
"step": 5156
},
{
"epoch": 0.3126060606060606,
"grad_norm": 0.05322248116135597,
"learning_rate": 0.0008,
"loss": 2.2729780673980713,
"step": 5158
},
{
"epoch": 0.31272727272727274,
"grad_norm": 0.05397942662239075,
"learning_rate": 0.0008,
"loss": 2.2990798950195312,
"step": 5160
},
{
"epoch": 0.3128484848484849,
"grad_norm": 0.0514509454369545,
"learning_rate": 0.0008,
"loss": 2.238706588745117,
"step": 5162
},
{
"epoch": 0.31296969696969695,
"grad_norm": 0.054145876318216324,
"learning_rate": 0.0008,
"loss": 2.25433611869812,
"step": 5164
},
{
"epoch": 0.3130909090909091,
"grad_norm": 0.058178115636110306,
"learning_rate": 0.0008,
"loss": 2.2490108013153076,
"step": 5166
},
{
"epoch": 0.3132121212121212,
"grad_norm": 0.06320502609014511,
"learning_rate": 0.0008,
"loss": 2.303352117538452,
"step": 5168
},
{
"epoch": 0.31333333333333335,
"grad_norm": 0.05172431468963623,
"learning_rate": 0.0008,
"loss": 2.2509217262268066,
"step": 5170
},
{
"epoch": 0.31345454545454543,
"grad_norm": 0.05369063466787338,
"learning_rate": 0.0008,
"loss": 2.2455320358276367,
"step": 5172
},
{
"epoch": 0.31357575757575756,
"grad_norm": 0.05260274186730385,
"learning_rate": 0.0008,
"loss": 2.2076597213745117,
"step": 5174
},
{
"epoch": 0.3136969696969697,
"grad_norm": 0.05160117149353027,
"learning_rate": 0.0008,
"loss": 2.2287423610687256,
"step": 5176
},
{
"epoch": 0.31381818181818183,
"grad_norm": 0.056940916925668716,
"learning_rate": 0.0008,
"loss": 2.2457714080810547,
"step": 5178
},
{
"epoch": 0.31393939393939396,
"grad_norm": 0.05299292132258415,
"learning_rate": 0.0008,
"loss": 2.3148868083953857,
"step": 5180
},
{
"epoch": 0.31406060606060604,
"grad_norm": 0.05536327883601189,
"learning_rate": 0.0008,
"loss": 2.2653095722198486,
"step": 5182
},
{
"epoch": 0.3141818181818182,
"grad_norm": 0.05206378549337387,
"learning_rate": 0.0008,
"loss": 2.266627311706543,
"step": 5184
},
{
"epoch": 0.3143030303030303,
"grad_norm": 0.05323569476604462,
"learning_rate": 0.0008,
"loss": 2.2370221614837646,
"step": 5186
},
{
"epoch": 0.31442424242424244,
"grad_norm": 0.055650997906923294,
"learning_rate": 0.0008,
"loss": 2.267007827758789,
"step": 5188
},
{
"epoch": 0.3145454545454546,
"grad_norm": 0.053603529930114746,
"learning_rate": 0.0008,
"loss": 2.2553799152374268,
"step": 5190
},
{
"epoch": 0.31466666666666665,
"grad_norm": 0.05420469492673874,
"learning_rate": 0.0008,
"loss": 2.304157018661499,
"step": 5192
},
{
"epoch": 0.3147878787878788,
"grad_norm": 0.06082844361662865,
"learning_rate": 0.0008,
"loss": 2.255937099456787,
"step": 5194
},
{
"epoch": 0.3149090909090909,
"grad_norm": 0.060298558324575424,
"learning_rate": 0.0008,
"loss": 2.2109673023223877,
"step": 5196
},
{
"epoch": 0.31503030303030305,
"grad_norm": 0.05528339743614197,
"learning_rate": 0.0008,
"loss": 2.21468448638916,
"step": 5198
},
{
"epoch": 0.3151515151515151,
"grad_norm": 0.05354516580700874,
"learning_rate": 0.0008,
"loss": 2.2536542415618896,
"step": 5200
},
{
"epoch": 0.31527272727272726,
"grad_norm": 0.05291422829031944,
"learning_rate": 0.0008,
"loss": 2.289047956466675,
"step": 5202
},
{
"epoch": 0.3153939393939394,
"grad_norm": 0.05205340310931206,
"learning_rate": 0.0008,
"loss": 2.2639050483703613,
"step": 5204
},
{
"epoch": 0.3155151515151515,
"grad_norm": 0.05167258903384209,
"learning_rate": 0.0008,
"loss": 2.280545711517334,
"step": 5206
},
{
"epoch": 0.31563636363636366,
"grad_norm": 0.05644863471388817,
"learning_rate": 0.0008,
"loss": 2.3265373706817627,
"step": 5208
},
{
"epoch": 0.31575757575757574,
"grad_norm": 0.05484391748905182,
"learning_rate": 0.0008,
"loss": 2.2909352779388428,
"step": 5210
},
{
"epoch": 0.31587878787878787,
"grad_norm": 0.05280205234885216,
"learning_rate": 0.0008,
"loss": 2.26828670501709,
"step": 5212
},
{
"epoch": 0.316,
"grad_norm": 0.05580522492527962,
"learning_rate": 0.0008,
"loss": 2.2683770656585693,
"step": 5214
},
{
"epoch": 0.31612121212121214,
"grad_norm": 0.0523495227098465,
"learning_rate": 0.0008,
"loss": 2.283836603164673,
"step": 5216
},
{
"epoch": 0.31624242424242427,
"grad_norm": 0.05549922212958336,
"learning_rate": 0.0008,
"loss": 2.26760196685791,
"step": 5218
},
{
"epoch": 0.31636363636363635,
"grad_norm": 0.05449224263429642,
"learning_rate": 0.0008,
"loss": 2.3642630577087402,
"step": 5220
},
{
"epoch": 0.3164848484848485,
"grad_norm": 0.05219032242894173,
"learning_rate": 0.0008,
"loss": 2.2695727348327637,
"step": 5222
},
{
"epoch": 0.3166060606060606,
"grad_norm": 0.052773602306842804,
"learning_rate": 0.0008,
"loss": 2.3136188983917236,
"step": 5224
},
{
"epoch": 0.31672727272727275,
"grad_norm": 0.052060145884752274,
"learning_rate": 0.0008,
"loss": 2.263679027557373,
"step": 5226
},
{
"epoch": 0.3168484848484848,
"grad_norm": 0.05263189971446991,
"learning_rate": 0.0008,
"loss": 2.3192131519317627,
"step": 5228
},
{
"epoch": 0.31696969696969696,
"grad_norm": 0.054213911294937134,
"learning_rate": 0.0008,
"loss": 2.1985907554626465,
"step": 5230
},
{
"epoch": 0.3170909090909091,
"grad_norm": 0.052608538419008255,
"learning_rate": 0.0008,
"loss": 2.282184362411499,
"step": 5232
},
{
"epoch": 0.3172121212121212,
"grad_norm": 0.05043691396713257,
"learning_rate": 0.0008,
"loss": 2.211106777191162,
"step": 5234
},
{
"epoch": 0.31733333333333336,
"grad_norm": 0.05246850475668907,
"learning_rate": 0.0008,
"loss": 2.2285594940185547,
"step": 5236
},
{
"epoch": 0.31745454545454543,
"grad_norm": 0.055879075080156326,
"learning_rate": 0.0008,
"loss": 2.2168850898742676,
"step": 5238
},
{
"epoch": 0.31757575757575757,
"grad_norm": 0.05339336767792702,
"learning_rate": 0.0008,
"loss": 2.2591588497161865,
"step": 5240
},
{
"epoch": 0.3176969696969697,
"grad_norm": 0.054350439459085464,
"learning_rate": 0.0008,
"loss": 2.316257953643799,
"step": 5242
},
{
"epoch": 0.31781818181818183,
"grad_norm": 0.053266435861587524,
"learning_rate": 0.0008,
"loss": 2.3071250915527344,
"step": 5244
},
{
"epoch": 0.31793939393939397,
"grad_norm": 0.052477676421403885,
"learning_rate": 0.0008,
"loss": 2.297837734222412,
"step": 5246
},
{
"epoch": 0.31806060606060604,
"grad_norm": 0.0534704364836216,
"learning_rate": 0.0008,
"loss": 2.277111530303955,
"step": 5248
},
{
"epoch": 0.3181818181818182,
"grad_norm": 0.05562315508723259,
"learning_rate": 0.0008,
"loss": 2.26536226272583,
"step": 5250
},
{
"epoch": 0.3183030303030303,
"grad_norm": 0.05066218972206116,
"learning_rate": 0.0008,
"loss": 2.3144309520721436,
"step": 5252
},
{
"epoch": 0.31842424242424244,
"grad_norm": 0.05705937370657921,
"learning_rate": 0.0008,
"loss": 2.2704877853393555,
"step": 5254
},
{
"epoch": 0.3185454545454545,
"grad_norm": 0.0529039092361927,
"learning_rate": 0.0008,
"loss": 2.253664016723633,
"step": 5256
},
{
"epoch": 0.31866666666666665,
"grad_norm": 0.05326903983950615,
"learning_rate": 0.0008,
"loss": 2.225534439086914,
"step": 5258
},
{
"epoch": 0.3187878787878788,
"grad_norm": 0.05121243745088577,
"learning_rate": 0.0008,
"loss": 2.2570419311523438,
"step": 5260
},
{
"epoch": 0.3189090909090909,
"grad_norm": 0.056219182908535004,
"learning_rate": 0.0008,
"loss": 2.2673821449279785,
"step": 5262
},
{
"epoch": 0.31903030303030305,
"grad_norm": 0.05711553990840912,
"learning_rate": 0.0008,
"loss": 2.208186626434326,
"step": 5264
},
{
"epoch": 0.31915151515151513,
"grad_norm": 0.051490668207407,
"learning_rate": 0.0008,
"loss": 2.2180910110473633,
"step": 5266
},
{
"epoch": 0.31927272727272726,
"grad_norm": 0.05434102937579155,
"learning_rate": 0.0008,
"loss": 2.2338855266571045,
"step": 5268
},
{
"epoch": 0.3193939393939394,
"grad_norm": 0.05402636528015137,
"learning_rate": 0.0008,
"loss": 2.323922634124756,
"step": 5270
},
{
"epoch": 0.31951515151515153,
"grad_norm": 0.05529558286070824,
"learning_rate": 0.0008,
"loss": 2.1825671195983887,
"step": 5272
},
{
"epoch": 0.31963636363636366,
"grad_norm": 0.05041230842471123,
"learning_rate": 0.0008,
"loss": 2.283132791519165,
"step": 5274
},
{
"epoch": 0.31975757575757574,
"grad_norm": 0.05366094037890434,
"learning_rate": 0.0008,
"loss": 2.2340610027313232,
"step": 5276
},
{
"epoch": 0.3198787878787879,
"grad_norm": 0.05257079750299454,
"learning_rate": 0.0008,
"loss": 2.2663280963897705,
"step": 5278
},
{
"epoch": 0.32,
"grad_norm": 0.052391305565834045,
"learning_rate": 0.0008,
"loss": 2.2532975673675537,
"step": 5280
},
{
"epoch": 0.32012121212121214,
"grad_norm": 0.056173574179410934,
"learning_rate": 0.0008,
"loss": 2.1904492378234863,
"step": 5282
},
{
"epoch": 0.3202424242424242,
"grad_norm": 0.05472239851951599,
"learning_rate": 0.0008,
"loss": 2.262233257293701,
"step": 5284
},
{
"epoch": 0.32036363636363635,
"grad_norm": 0.051241565495729446,
"learning_rate": 0.0008,
"loss": 2.263385772705078,
"step": 5286
},
{
"epoch": 0.3204848484848485,
"grad_norm": 0.05140797793865204,
"learning_rate": 0.0008,
"loss": 2.2199037075042725,
"step": 5288
},
{
"epoch": 0.3206060606060606,
"grad_norm": 0.05274258553981781,
"learning_rate": 0.0008,
"loss": 2.226914644241333,
"step": 5290
},
{
"epoch": 0.32072727272727275,
"grad_norm": 0.05170126259326935,
"learning_rate": 0.0008,
"loss": 2.1817076206207275,
"step": 5292
},
{
"epoch": 0.3208484848484848,
"grad_norm": 0.0500362366437912,
"learning_rate": 0.0008,
"loss": 2.217071294784546,
"step": 5294
},
{
"epoch": 0.32096969696969696,
"grad_norm": 0.05077001079916954,
"learning_rate": 0.0008,
"loss": 2.2472894191741943,
"step": 5296
},
{
"epoch": 0.3210909090909091,
"grad_norm": 0.05097509175539017,
"learning_rate": 0.0008,
"loss": 2.2884068489074707,
"step": 5298
},
{
"epoch": 0.3212121212121212,
"grad_norm": 0.05051326006650925,
"learning_rate": 0.0008,
"loss": 2.2487690448760986,
"step": 5300
},
{
"epoch": 0.32133333333333336,
"grad_norm": 0.05673372745513916,
"learning_rate": 0.0008,
"loss": 2.1978023052215576,
"step": 5302
},
{
"epoch": 0.32145454545454544,
"grad_norm": 0.05627625063061714,
"learning_rate": 0.0008,
"loss": 2.2372593879699707,
"step": 5304
},
{
"epoch": 0.32157575757575757,
"grad_norm": 0.051625221967697144,
"learning_rate": 0.0008,
"loss": 2.2600319385528564,
"step": 5306
},
{
"epoch": 0.3216969696969697,
"grad_norm": 0.05276033654808998,
"learning_rate": 0.0008,
"loss": 2.312293529510498,
"step": 5308
},
{
"epoch": 0.32181818181818184,
"grad_norm": 0.051407549530267715,
"learning_rate": 0.0008,
"loss": 2.1728811264038086,
"step": 5310
},
{
"epoch": 0.3219393939393939,
"grad_norm": 0.05442081764340401,
"learning_rate": 0.0008,
"loss": 2.2043864727020264,
"step": 5312
},
{
"epoch": 0.32206060606060605,
"grad_norm": 0.05285586416721344,
"learning_rate": 0.0008,
"loss": 2.2359650135040283,
"step": 5314
},
{
"epoch": 0.3221818181818182,
"grad_norm": 0.05186446011066437,
"learning_rate": 0.0008,
"loss": 2.294674873352051,
"step": 5316
},
{
"epoch": 0.3223030303030303,
"grad_norm": 0.05560285225510597,
"learning_rate": 0.0008,
"loss": 2.178133964538574,
"step": 5318
},
{
"epoch": 0.32242424242424245,
"grad_norm": 0.0552506223320961,
"learning_rate": 0.0008,
"loss": 2.2930986881256104,
"step": 5320
},
{
"epoch": 0.3225454545454545,
"grad_norm": 0.054093871265649796,
"learning_rate": 0.0008,
"loss": 2.2221741676330566,
"step": 5322
},
{
"epoch": 0.32266666666666666,
"grad_norm": 0.05332799255847931,
"learning_rate": 0.0008,
"loss": 2.2977309226989746,
"step": 5324
},
{
"epoch": 0.3227878787878788,
"grad_norm": 0.053985197097063065,
"learning_rate": 0.0008,
"loss": 2.3668861389160156,
"step": 5326
},
{
"epoch": 0.3229090909090909,
"grad_norm": 0.05173453316092491,
"learning_rate": 0.0008,
"loss": 2.275434970855713,
"step": 5328
},
{
"epoch": 0.32303030303030306,
"grad_norm": 0.053992610424757004,
"learning_rate": 0.0008,
"loss": 2.2491555213928223,
"step": 5330
},
{
"epoch": 0.32315151515151513,
"grad_norm": 0.053788233548402786,
"learning_rate": 0.0008,
"loss": 2.342629909515381,
"step": 5332
},
{
"epoch": 0.32327272727272727,
"grad_norm": 0.05247446894645691,
"learning_rate": 0.0008,
"loss": 2.2054646015167236,
"step": 5334
},
{
"epoch": 0.3233939393939394,
"grad_norm": 0.05854100361466408,
"learning_rate": 0.0008,
"loss": 2.2749366760253906,
"step": 5336
},
{
"epoch": 0.32351515151515153,
"grad_norm": 0.06322918832302094,
"learning_rate": 0.0008,
"loss": 2.301295518875122,
"step": 5338
},
{
"epoch": 0.3236363636363636,
"grad_norm": 0.06053263321518898,
"learning_rate": 0.0008,
"loss": 2.3333513736724854,
"step": 5340
},
{
"epoch": 0.32375757575757574,
"grad_norm": 0.057731129229068756,
"learning_rate": 0.0008,
"loss": 2.2558937072753906,
"step": 5342
},
{
"epoch": 0.3238787878787879,
"grad_norm": 0.056044768542051315,
"learning_rate": 0.0008,
"loss": 2.1430654525756836,
"step": 5344
},
{
"epoch": 0.324,
"grad_norm": 0.05465024337172508,
"learning_rate": 0.0008,
"loss": 2.27260684967041,
"step": 5346
},
{
"epoch": 0.32412121212121214,
"grad_norm": 0.05361025780439377,
"learning_rate": 0.0008,
"loss": 2.2426557540893555,
"step": 5348
},
{
"epoch": 0.3242424242424242,
"grad_norm": 0.05234428867697716,
"learning_rate": 0.0008,
"loss": 2.3179750442504883,
"step": 5350
},
{
"epoch": 0.32436363636363635,
"grad_norm": 0.053809333592653275,
"learning_rate": 0.0008,
"loss": 2.2484073638916016,
"step": 5352
},
{
"epoch": 0.3244848484848485,
"grad_norm": 0.052514269948005676,
"learning_rate": 0.0008,
"loss": 2.333420753479004,
"step": 5354
},
{
"epoch": 0.3246060606060606,
"grad_norm": 0.05770666524767876,
"learning_rate": 0.0008,
"loss": 2.224508047103882,
"step": 5356
},
{
"epoch": 0.32472727272727275,
"grad_norm": 0.05597487464547157,
"learning_rate": 0.0008,
"loss": 2.2206103801727295,
"step": 5358
},
{
"epoch": 0.32484848484848483,
"grad_norm": 0.05357656255364418,
"learning_rate": 0.0008,
"loss": 2.3157355785369873,
"step": 5360
},
{
"epoch": 0.32496969696969696,
"grad_norm": 0.05336331948637962,
"learning_rate": 0.0008,
"loss": 2.27239990234375,
"step": 5362
},
{
"epoch": 0.3250909090909091,
"grad_norm": 0.050701577216386795,
"learning_rate": 0.0008,
"loss": 2.152679920196533,
"step": 5364
},
{
"epoch": 0.32521212121212123,
"grad_norm": 0.05092359706759453,
"learning_rate": 0.0008,
"loss": 2.249755382537842,
"step": 5366
},
{
"epoch": 0.3253333333333333,
"grad_norm": 0.05044754967093468,
"learning_rate": 0.0008,
"loss": 2.273810625076294,
"step": 5368
},
{
"epoch": 0.32545454545454544,
"grad_norm": 0.05252624303102493,
"learning_rate": 0.0008,
"loss": 2.253786325454712,
"step": 5370
},
{
"epoch": 0.3255757575757576,
"grad_norm": 0.056526634842157364,
"learning_rate": 0.0008,
"loss": 2.2265541553497314,
"step": 5372
},
{
"epoch": 0.3256969696969697,
"grad_norm": 0.055767931044101715,
"learning_rate": 0.0008,
"loss": 2.233996868133545,
"step": 5374
},
{
"epoch": 0.32581818181818184,
"grad_norm": 0.05808468535542488,
"learning_rate": 0.0008,
"loss": 2.2933177947998047,
"step": 5376
},
{
"epoch": 0.3259393939393939,
"grad_norm": 0.05862117186188698,
"learning_rate": 0.0008,
"loss": 2.2376554012298584,
"step": 5378
},
{
"epoch": 0.32606060606060605,
"grad_norm": 0.05928567796945572,
"learning_rate": 0.0008,
"loss": 2.1990482807159424,
"step": 5380
},
{
"epoch": 0.3261818181818182,
"grad_norm": 0.052516140043735504,
"learning_rate": 0.0008,
"loss": 2.207632541656494,
"step": 5382
},
{
"epoch": 0.3263030303030303,
"grad_norm": 0.05679621174931526,
"learning_rate": 0.0008,
"loss": 2.2769880294799805,
"step": 5384
},
{
"epoch": 0.32642424242424245,
"grad_norm": 0.051626767963171005,
"learning_rate": 0.0008,
"loss": 2.289910078048706,
"step": 5386
},
{
"epoch": 0.3265454545454545,
"grad_norm": 0.05135383456945419,
"learning_rate": 0.0008,
"loss": 2.260389566421509,
"step": 5388
},
{
"epoch": 0.32666666666666666,
"grad_norm": 0.05053543299436569,
"learning_rate": 0.0008,
"loss": 2.2299160957336426,
"step": 5390
},
{
"epoch": 0.3267878787878788,
"grad_norm": 0.050200603902339935,
"learning_rate": 0.0008,
"loss": 2.2810704708099365,
"step": 5392
},
{
"epoch": 0.3269090909090909,
"grad_norm": 0.05121053755283356,
"learning_rate": 0.0008,
"loss": 2.2809548377990723,
"step": 5394
},
{
"epoch": 0.327030303030303,
"grad_norm": 0.052041102200746536,
"learning_rate": 0.0008,
"loss": 2.2930891513824463,
"step": 5396
},
{
"epoch": 0.32715151515151514,
"grad_norm": 0.05310126394033432,
"learning_rate": 0.0008,
"loss": 2.266623020172119,
"step": 5398
},
{
"epoch": 0.32727272727272727,
"grad_norm": 0.052748315036296844,
"learning_rate": 0.0008,
"loss": 2.2476646900177,
"step": 5400
},
{
"epoch": 0.3273939393939394,
"grad_norm": 0.058248281478881836,
"learning_rate": 0.0008,
"loss": 2.2215776443481445,
"step": 5402
},
{
"epoch": 0.32751515151515154,
"grad_norm": 0.050300631672143936,
"learning_rate": 0.0008,
"loss": 2.1788711547851562,
"step": 5404
},
{
"epoch": 0.3276363636363636,
"grad_norm": 0.05315973982214928,
"learning_rate": 0.0008,
"loss": 2.288642406463623,
"step": 5406
},
{
"epoch": 0.32775757575757575,
"grad_norm": 0.05166429653763771,
"learning_rate": 0.0008,
"loss": 2.255889654159546,
"step": 5408
},
{
"epoch": 0.3278787878787879,
"grad_norm": 0.051731374114751816,
"learning_rate": 0.0008,
"loss": 2.25463604927063,
"step": 5410
},
{
"epoch": 0.328,
"grad_norm": 0.05051145330071449,
"learning_rate": 0.0008,
"loss": 2.240633487701416,
"step": 5412
},
{
"epoch": 0.32812121212121215,
"grad_norm": 0.052244387567043304,
"learning_rate": 0.0008,
"loss": 2.2220752239227295,
"step": 5414
},
{
"epoch": 0.3282424242424242,
"grad_norm": 0.05238129943609238,
"learning_rate": 0.0008,
"loss": 2.26833176612854,
"step": 5416
},
{
"epoch": 0.32836363636363636,
"grad_norm": 0.057548679411411285,
"learning_rate": 0.0008,
"loss": 2.189582586288452,
"step": 5418
},
{
"epoch": 0.3284848484848485,
"grad_norm": 0.05470182001590729,
"learning_rate": 0.0008,
"loss": 2.201359748840332,
"step": 5420
},
{
"epoch": 0.3286060606060606,
"grad_norm": 0.054357849061489105,
"learning_rate": 0.0008,
"loss": 2.235959529876709,
"step": 5422
},
{
"epoch": 0.3287272727272727,
"grad_norm": 0.05358314886689186,
"learning_rate": 0.0008,
"loss": 2.2660770416259766,
"step": 5424
},
{
"epoch": 0.32884848484848483,
"grad_norm": 0.052821580320596695,
"learning_rate": 0.0008,
"loss": 2.2931203842163086,
"step": 5426
},
{
"epoch": 0.32896969696969697,
"grad_norm": 0.05957585573196411,
"learning_rate": 0.0008,
"loss": 2.259779453277588,
"step": 5428
},
{
"epoch": 0.3290909090909091,
"grad_norm": 0.05704871937632561,
"learning_rate": 0.0008,
"loss": 2.315983772277832,
"step": 5430
},
{
"epoch": 0.32921212121212123,
"grad_norm": 0.05119642615318298,
"learning_rate": 0.0008,
"loss": 2.242826223373413,
"step": 5432
},
{
"epoch": 0.3293333333333333,
"grad_norm": 0.05293199047446251,
"learning_rate": 0.0008,
"loss": 2.250272750854492,
"step": 5434
},
{
"epoch": 0.32945454545454544,
"grad_norm": 0.060934025794267654,
"learning_rate": 0.0008,
"loss": 2.2740869522094727,
"step": 5436
},
{
"epoch": 0.3295757575757576,
"grad_norm": 0.05443413555622101,
"learning_rate": 0.0008,
"loss": 2.2066543102264404,
"step": 5438
},
{
"epoch": 0.3296969696969697,
"grad_norm": 0.05378193408250809,
"learning_rate": 0.0008,
"loss": 2.2750442028045654,
"step": 5440
},
{
"epoch": 0.32981818181818184,
"grad_norm": 0.05166340246796608,
"learning_rate": 0.0008,
"loss": 2.2438294887542725,
"step": 5442
},
{
"epoch": 0.3299393939393939,
"grad_norm": 0.05006808415055275,
"learning_rate": 0.0008,
"loss": 2.202366590499878,
"step": 5444
},
{
"epoch": 0.33006060606060605,
"grad_norm": 0.054871633648872375,
"learning_rate": 0.0008,
"loss": 2.3330419063568115,
"step": 5446
},
{
"epoch": 0.3301818181818182,
"grad_norm": 0.05251624435186386,
"learning_rate": 0.0008,
"loss": 2.2445366382598877,
"step": 5448
},
{
"epoch": 0.3303030303030303,
"grad_norm": 0.05195561796426773,
"learning_rate": 0.0008,
"loss": 2.2141711711883545,
"step": 5450
},
{
"epoch": 0.3304242424242424,
"grad_norm": 0.05249389633536339,
"learning_rate": 0.0008,
"loss": 2.3311309814453125,
"step": 5452
},
{
"epoch": 0.33054545454545453,
"grad_norm": 0.050009939819574356,
"learning_rate": 0.0008,
"loss": 2.19038724899292,
"step": 5454
},
{
"epoch": 0.33066666666666666,
"grad_norm": 0.0519530326128006,
"learning_rate": 0.0008,
"loss": 2.2466912269592285,
"step": 5456
},
{
"epoch": 0.3307878787878788,
"grad_norm": 0.05233702436089516,
"learning_rate": 0.0008,
"loss": 2.2587637901306152,
"step": 5458
},
{
"epoch": 0.33090909090909093,
"grad_norm": 0.05140082538127899,
"learning_rate": 0.0008,
"loss": 2.277503252029419,
"step": 5460
},
{
"epoch": 0.331030303030303,
"grad_norm": 0.053310099989175797,
"learning_rate": 0.0008,
"loss": 2.3261301517486572,
"step": 5462
},
{
"epoch": 0.33115151515151514,
"grad_norm": 0.05079266428947449,
"learning_rate": 0.0008,
"loss": 2.238736152648926,
"step": 5464
},
{
"epoch": 0.3312727272727273,
"grad_norm": 0.050818026065826416,
"learning_rate": 0.0008,
"loss": 2.1969966888427734,
"step": 5466
},
{
"epoch": 0.3313939393939394,
"grad_norm": 0.052852459251880646,
"learning_rate": 0.0008,
"loss": 2.2932655811309814,
"step": 5468
},
{
"epoch": 0.33151515151515154,
"grad_norm": 0.04987957328557968,
"learning_rate": 0.0008,
"loss": 2.2001166343688965,
"step": 5470
},
{
"epoch": 0.3316363636363636,
"grad_norm": 0.05112595856189728,
"learning_rate": 0.0008,
"loss": 2.3080954551696777,
"step": 5472
},
{
"epoch": 0.33175757575757575,
"grad_norm": 0.052628692239522934,
"learning_rate": 0.0008,
"loss": 2.3051798343658447,
"step": 5474
},
{
"epoch": 0.3318787878787879,
"grad_norm": 0.05683489143848419,
"learning_rate": 0.0008,
"loss": 2.2375428676605225,
"step": 5476
},
{
"epoch": 0.332,
"grad_norm": 0.06053069233894348,
"learning_rate": 0.0008,
"loss": 2.1841089725494385,
"step": 5478
},
{
"epoch": 0.3321212121212121,
"grad_norm": 0.05660552531480789,
"learning_rate": 0.0008,
"loss": 2.3046205043792725,
"step": 5480
},
{
"epoch": 0.33224242424242423,
"grad_norm": 0.054854802787303925,
"learning_rate": 0.0008,
"loss": 2.3218331336975098,
"step": 5482
},
{
"epoch": 0.33236363636363636,
"grad_norm": 0.052271246910095215,
"learning_rate": 0.0008,
"loss": 2.2081682682037354,
"step": 5484
},
{
"epoch": 0.3324848484848485,
"grad_norm": 0.05305660516023636,
"learning_rate": 0.0008,
"loss": 2.242936611175537,
"step": 5486
},
{
"epoch": 0.3326060606060606,
"grad_norm": 0.05313907191157341,
"learning_rate": 0.0008,
"loss": 2.21592116355896,
"step": 5488
},
{
"epoch": 0.3327272727272727,
"grad_norm": 0.04953081160783768,
"learning_rate": 0.0008,
"loss": 2.1898114681243896,
"step": 5490
},
{
"epoch": 0.33284848484848484,
"grad_norm": 0.052375223487615585,
"learning_rate": 0.0008,
"loss": 2.2779088020324707,
"step": 5492
},
{
"epoch": 0.33296969696969697,
"grad_norm": 0.05448406562209129,
"learning_rate": 0.0008,
"loss": 2.210096597671509,
"step": 5494
},
{
"epoch": 0.3330909090909091,
"grad_norm": 0.051673565059900284,
"learning_rate": 0.0008,
"loss": 2.1918656826019287,
"step": 5496
},
{
"epoch": 0.33321212121212124,
"grad_norm": 0.05259697139263153,
"learning_rate": 0.0008,
"loss": 2.3302395343780518,
"step": 5498
},
{
"epoch": 0.3333333333333333,
"grad_norm": 0.05146722495555878,
"learning_rate": 0.0008,
"loss": 2.2305734157562256,
"step": 5500
},
{
"epoch": 0.33345454545454545,
"grad_norm": 0.052488550543785095,
"learning_rate": 0.0008,
"loss": 2.24491810798645,
"step": 5502
},
{
"epoch": 0.3335757575757576,
"grad_norm": 0.05749073997139931,
"learning_rate": 0.0008,
"loss": 2.239663600921631,
"step": 5504
},
{
"epoch": 0.3336969696969697,
"grad_norm": 0.0584464929997921,
"learning_rate": 0.0008,
"loss": 2.244873523712158,
"step": 5506
},
{
"epoch": 0.3338181818181818,
"grad_norm": 0.05404992774128914,
"learning_rate": 0.0008,
"loss": 2.166145086288452,
"step": 5508
},
{
"epoch": 0.3339393939393939,
"grad_norm": 0.05260791629552841,
"learning_rate": 0.0008,
"loss": 2.219820976257324,
"step": 5510
},
{
"epoch": 0.33406060606060606,
"grad_norm": 0.05383313447237015,
"learning_rate": 0.0008,
"loss": 2.1880571842193604,
"step": 5512
},
{
"epoch": 0.3341818181818182,
"grad_norm": 0.05109872296452522,
"learning_rate": 0.0008,
"loss": 2.2922942638397217,
"step": 5514
},
{
"epoch": 0.3343030303030303,
"grad_norm": 0.05094429478049278,
"learning_rate": 0.0008,
"loss": 2.2411038875579834,
"step": 5516
},
{
"epoch": 0.3344242424242424,
"grad_norm": 0.06018686294555664,
"learning_rate": 0.0008,
"loss": 2.285975933074951,
"step": 5518
},
{
"epoch": 0.33454545454545453,
"grad_norm": 0.055872008204460144,
"learning_rate": 0.0008,
"loss": 2.287388801574707,
"step": 5520
},
{
"epoch": 0.33466666666666667,
"grad_norm": 0.05262831598520279,
"learning_rate": 0.0008,
"loss": 2.2292561531066895,
"step": 5522
},
{
"epoch": 0.3347878787878788,
"grad_norm": 0.05343739315867424,
"learning_rate": 0.0008,
"loss": 2.1840450763702393,
"step": 5524
},
{
"epoch": 0.33490909090909093,
"grad_norm": 0.05536837503314018,
"learning_rate": 0.0008,
"loss": 2.2548985481262207,
"step": 5526
},
{
"epoch": 0.335030303030303,
"grad_norm": 0.056267522275447845,
"learning_rate": 0.0008,
"loss": 2.2226741313934326,
"step": 5528
},
{
"epoch": 0.33515151515151514,
"grad_norm": 0.051190655678510666,
"learning_rate": 0.0008,
"loss": 2.1849353313446045,
"step": 5530
},
{
"epoch": 0.3352727272727273,
"grad_norm": 0.05475626140832901,
"learning_rate": 0.0008,
"loss": 2.2988247871398926,
"step": 5532
},
{
"epoch": 0.3353939393939394,
"grad_norm": 0.054103702306747437,
"learning_rate": 0.0008,
"loss": 2.2729806900024414,
"step": 5534
},
{
"epoch": 0.3355151515151515,
"grad_norm": 0.05178387090563774,
"learning_rate": 0.0008,
"loss": 2.2369167804718018,
"step": 5536
},
{
"epoch": 0.3356363636363636,
"grad_norm": 0.050831496715545654,
"learning_rate": 0.0008,
"loss": 2.3344740867614746,
"step": 5538
},
{
"epoch": 0.33575757575757575,
"grad_norm": 0.05350077524781227,
"learning_rate": 0.0008,
"loss": 2.2387633323669434,
"step": 5540
},
{
"epoch": 0.3358787878787879,
"grad_norm": 0.062117304652929306,
"learning_rate": 0.0008,
"loss": 2.319638967514038,
"step": 5542
},
{
"epoch": 0.336,
"grad_norm": 0.06296572089195251,
"learning_rate": 0.0008,
"loss": 2.238285541534424,
"step": 5544
},
{
"epoch": 0.3361212121212121,
"grad_norm": 0.05752076208591461,
"learning_rate": 0.0008,
"loss": 2.180166244506836,
"step": 5546
},
{
"epoch": 0.33624242424242423,
"grad_norm": 0.056342147290706635,
"learning_rate": 0.0008,
"loss": 2.253131151199341,
"step": 5548
},
{
"epoch": 0.33636363636363636,
"grad_norm": 0.05454893782734871,
"learning_rate": 0.0008,
"loss": 2.2599031925201416,
"step": 5550
},
{
"epoch": 0.3364848484848485,
"grad_norm": 0.053011614829301834,
"learning_rate": 0.0008,
"loss": 2.2702126502990723,
"step": 5552
},
{
"epoch": 0.33660606060606063,
"grad_norm": 0.05906348302960396,
"learning_rate": 0.0008,
"loss": 2.245781421661377,
"step": 5554
},
{
"epoch": 0.3367272727272727,
"grad_norm": 0.05737348645925522,
"learning_rate": 0.0008,
"loss": 2.225053548812866,
"step": 5556
},
{
"epoch": 0.33684848484848484,
"grad_norm": 0.05855663865804672,
"learning_rate": 0.0008,
"loss": 2.2524585723876953,
"step": 5558
},
{
"epoch": 0.336969696969697,
"grad_norm": 0.05519278347492218,
"learning_rate": 0.0008,
"loss": 2.2742860317230225,
"step": 5560
},
{
"epoch": 0.3370909090909091,
"grad_norm": 0.06135447323322296,
"learning_rate": 0.0008,
"loss": 2.253138542175293,
"step": 5562
},
{
"epoch": 0.3372121212121212,
"grad_norm": 0.05326376482844353,
"learning_rate": 0.0008,
"loss": 2.1953318119049072,
"step": 5564
},
{
"epoch": 0.3373333333333333,
"grad_norm": 0.05473532900214195,
"learning_rate": 0.0008,
"loss": 2.283531665802002,
"step": 5566
},
{
"epoch": 0.33745454545454545,
"grad_norm": 0.05501887947320938,
"learning_rate": 0.0008,
"loss": 2.259277582168579,
"step": 5568
},
{
"epoch": 0.3375757575757576,
"grad_norm": 0.055205363780260086,
"learning_rate": 0.0008,
"loss": 2.284001588821411,
"step": 5570
},
{
"epoch": 0.3376969696969697,
"grad_norm": 0.051685530692338943,
"learning_rate": 0.0008,
"loss": 2.2552876472473145,
"step": 5572
},
{
"epoch": 0.3378181818181818,
"grad_norm": 0.051783185452222824,
"learning_rate": 0.0008,
"loss": 2.23629093170166,
"step": 5574
},
{
"epoch": 0.33793939393939393,
"grad_norm": 0.050138819962739944,
"learning_rate": 0.0008,
"loss": 2.2314252853393555,
"step": 5576
},
{
"epoch": 0.33806060606060606,
"grad_norm": 0.05092225223779678,
"learning_rate": 0.0008,
"loss": 2.2135934829711914,
"step": 5578
},
{
"epoch": 0.3381818181818182,
"grad_norm": 0.051636673510074615,
"learning_rate": 0.0008,
"loss": 2.2159199714660645,
"step": 5580
},
{
"epoch": 0.3383030303030303,
"grad_norm": 0.05083484575152397,
"learning_rate": 0.0008,
"loss": 2.2281041145324707,
"step": 5582
},
{
"epoch": 0.3384242424242424,
"grad_norm": 0.05328470841050148,
"learning_rate": 0.0008,
"loss": 2.221622943878174,
"step": 5584
},
{
"epoch": 0.33854545454545454,
"grad_norm": 0.05103734880685806,
"learning_rate": 0.0008,
"loss": 2.2559919357299805,
"step": 5586
},
{
"epoch": 0.33866666666666667,
"grad_norm": 0.05345030501484871,
"learning_rate": 0.0008,
"loss": 2.2835206985473633,
"step": 5588
},
{
"epoch": 0.3387878787878788,
"grad_norm": 0.05394890904426575,
"learning_rate": 0.0008,
"loss": 2.2759265899658203,
"step": 5590
},
{
"epoch": 0.3389090909090909,
"grad_norm": 0.05111484229564667,
"learning_rate": 0.0008,
"loss": 2.2390730381011963,
"step": 5592
},
{
"epoch": 0.339030303030303,
"grad_norm": 0.05080556124448776,
"learning_rate": 0.0008,
"loss": 2.246476173400879,
"step": 5594
},
{
"epoch": 0.33915151515151515,
"grad_norm": 0.05149397253990173,
"learning_rate": 0.0008,
"loss": 2.243220090866089,
"step": 5596
},
{
"epoch": 0.3392727272727273,
"grad_norm": 0.053153883665800095,
"learning_rate": 0.0008,
"loss": 2.147970199584961,
"step": 5598
},
{
"epoch": 0.3393939393939394,
"grad_norm": 0.052676983177661896,
"learning_rate": 0.0008,
"loss": 2.2288496494293213,
"step": 5600
},
{
"epoch": 0.3395151515151515,
"grad_norm": 0.05586696416139603,
"learning_rate": 0.0008,
"loss": 2.2745883464813232,
"step": 5602
},
{
"epoch": 0.3396363636363636,
"grad_norm": 0.05252278596162796,
"learning_rate": 0.0008,
"loss": 2.188666820526123,
"step": 5604
},
{
"epoch": 0.33975757575757576,
"grad_norm": 0.050591301172971725,
"learning_rate": 0.0008,
"loss": 2.205155372619629,
"step": 5606
},
{
"epoch": 0.3398787878787879,
"grad_norm": 0.05311284586787224,
"learning_rate": 0.0008,
"loss": 2.1965644359588623,
"step": 5608
},
{
"epoch": 0.34,
"grad_norm": 0.051164645701646805,
"learning_rate": 0.0008,
"loss": 2.2636332511901855,
"step": 5610
},
{
"epoch": 0.3401212121212121,
"grad_norm": 0.05294632539153099,
"learning_rate": 0.0008,
"loss": 2.2326548099517822,
"step": 5612
},
{
"epoch": 0.34024242424242424,
"grad_norm": 0.052880410104990005,
"learning_rate": 0.0008,
"loss": 2.2943077087402344,
"step": 5614
},
{
"epoch": 0.34036363636363637,
"grad_norm": 0.05139737203717232,
"learning_rate": 0.0008,
"loss": 2.271604061126709,
"step": 5616
},
{
"epoch": 0.3404848484848485,
"grad_norm": 0.056922804564237595,
"learning_rate": 0.0008,
"loss": 2.2445549964904785,
"step": 5618
},
{
"epoch": 0.3406060606060606,
"grad_norm": 0.052218250930309296,
"learning_rate": 0.0008,
"loss": 2.225914478302002,
"step": 5620
},
{
"epoch": 0.3407272727272727,
"grad_norm": 0.049959104508161545,
"learning_rate": 0.0008,
"loss": 2.2073702812194824,
"step": 5622
},
{
"epoch": 0.34084848484848485,
"grad_norm": 0.05028089880943298,
"learning_rate": 0.0008,
"loss": 2.196833610534668,
"step": 5624
},
{
"epoch": 0.340969696969697,
"grad_norm": 0.05181104689836502,
"learning_rate": 0.0008,
"loss": 2.2469422817230225,
"step": 5626
},
{
"epoch": 0.3410909090909091,
"grad_norm": 0.05140150710940361,
"learning_rate": 0.0008,
"loss": 2.2469913959503174,
"step": 5628
},
{
"epoch": 0.3412121212121212,
"grad_norm": 0.05028458684682846,
"learning_rate": 0.0008,
"loss": 2.2474212646484375,
"step": 5630
},
{
"epoch": 0.3413333333333333,
"grad_norm": 0.05348304659128189,
"learning_rate": 0.0008,
"loss": 2.2134902477264404,
"step": 5632
},
{
"epoch": 0.34145454545454546,
"grad_norm": 0.050895705819129944,
"learning_rate": 0.0008,
"loss": 2.251155138015747,
"step": 5634
},
{
"epoch": 0.3415757575757576,
"grad_norm": 0.05823320522904396,
"learning_rate": 0.0008,
"loss": 2.223558187484741,
"step": 5636
},
{
"epoch": 0.3416969696969697,
"grad_norm": 0.05139965936541557,
"learning_rate": 0.0008,
"loss": 2.19826078414917,
"step": 5638
},
{
"epoch": 0.3418181818181818,
"grad_norm": 0.06512259691953659,
"learning_rate": 0.0008,
"loss": 2.2999610900878906,
"step": 5640
},
{
"epoch": 0.34193939393939393,
"grad_norm": 0.06221131607890129,
"learning_rate": 0.0008,
"loss": 2.187690496444702,
"step": 5642
},
{
"epoch": 0.34206060606060606,
"grad_norm": 0.055910490453243256,
"learning_rate": 0.0008,
"loss": 2.2651331424713135,
"step": 5644
},
{
"epoch": 0.3421818181818182,
"grad_norm": 0.053876809775829315,
"learning_rate": 0.0008,
"loss": 2.2660346031188965,
"step": 5646
},
{
"epoch": 0.3423030303030303,
"grad_norm": 0.05157052353024483,
"learning_rate": 0.0008,
"loss": 2.2087631225585938,
"step": 5648
},
{
"epoch": 0.3424242424242424,
"grad_norm": 0.05026421695947647,
"learning_rate": 0.0008,
"loss": 2.1985573768615723,
"step": 5650
},
{
"epoch": 0.34254545454545454,
"grad_norm": 0.05134473741054535,
"learning_rate": 0.0008,
"loss": 2.2175328731536865,
"step": 5652
},
{
"epoch": 0.3426666666666667,
"grad_norm": 0.05091085284948349,
"learning_rate": 0.0008,
"loss": 2.2355494499206543,
"step": 5654
},
{
"epoch": 0.3427878787878788,
"grad_norm": 0.05094783008098602,
"learning_rate": 0.0008,
"loss": 2.2526400089263916,
"step": 5656
},
{
"epoch": 0.3429090909090909,
"grad_norm": 0.053957074880599976,
"learning_rate": 0.0008,
"loss": 2.2229161262512207,
"step": 5658
},
{
"epoch": 0.343030303030303,
"grad_norm": 0.061695367097854614,
"learning_rate": 0.0008,
"loss": 2.274773120880127,
"step": 5660
},
{
"epoch": 0.34315151515151515,
"grad_norm": 0.05691468343138695,
"learning_rate": 0.0008,
"loss": 2.2531604766845703,
"step": 5662
},
{
"epoch": 0.3432727272727273,
"grad_norm": 0.051768917590379715,
"learning_rate": 0.0008,
"loss": 2.203748941421509,
"step": 5664
},
{
"epoch": 0.3433939393939394,
"grad_norm": 0.05583003908395767,
"learning_rate": 0.0008,
"loss": 2.3067643642425537,
"step": 5666
},
{
"epoch": 0.3435151515151515,
"grad_norm": 0.052506305277347565,
"learning_rate": 0.0008,
"loss": 2.2596216201782227,
"step": 5668
},
{
"epoch": 0.34363636363636363,
"grad_norm": 0.05397767201066017,
"learning_rate": 0.0008,
"loss": 2.2311086654663086,
"step": 5670
},
{
"epoch": 0.34375757575757576,
"grad_norm": 0.05574968829751015,
"learning_rate": 0.0008,
"loss": 2.2764668464660645,
"step": 5672
},
{
"epoch": 0.3438787878787879,
"grad_norm": 0.054686810821294785,
"learning_rate": 0.0008,
"loss": 2.217184066772461,
"step": 5674
},
{
"epoch": 0.344,
"grad_norm": 0.05302365496754646,
"learning_rate": 0.0008,
"loss": 2.2561700344085693,
"step": 5676
},
{
"epoch": 0.3441212121212121,
"grad_norm": 0.051208045333623886,
"learning_rate": 0.0008,
"loss": 2.2501933574676514,
"step": 5678
},
{
"epoch": 0.34424242424242424,
"grad_norm": 0.05107516795396805,
"learning_rate": 0.0008,
"loss": 2.2198641300201416,
"step": 5680
},
{
"epoch": 0.34436363636363637,
"grad_norm": 0.051407620310783386,
"learning_rate": 0.0008,
"loss": 2.205711841583252,
"step": 5682
},
{
"epoch": 0.3444848484848485,
"grad_norm": 0.05669273063540459,
"learning_rate": 0.0008,
"loss": 2.203049421310425,
"step": 5684
},
{
"epoch": 0.3446060606060606,
"grad_norm": 0.056729819625616074,
"learning_rate": 0.0008,
"loss": 2.253950357437134,
"step": 5686
},
{
"epoch": 0.3447272727272727,
"grad_norm": 0.05480062961578369,
"learning_rate": 0.0008,
"loss": 2.2955684661865234,
"step": 5688
},
{
"epoch": 0.34484848484848485,
"grad_norm": 0.05240105837583542,
"learning_rate": 0.0008,
"loss": 2.189633369445801,
"step": 5690
},
{
"epoch": 0.344969696969697,
"grad_norm": 0.05187736451625824,
"learning_rate": 0.0008,
"loss": 2.2975025177001953,
"step": 5692
},
{
"epoch": 0.3450909090909091,
"grad_norm": 0.051142241805791855,
"learning_rate": 0.0008,
"loss": 2.3187248706817627,
"step": 5694
},
{
"epoch": 0.3452121212121212,
"grad_norm": 0.05110097676515579,
"learning_rate": 0.0008,
"loss": 2.2451281547546387,
"step": 5696
},
{
"epoch": 0.3453333333333333,
"grad_norm": 0.050787296146154404,
"learning_rate": 0.0008,
"loss": 2.2377519607543945,
"step": 5698
},
{
"epoch": 0.34545454545454546,
"grad_norm": 0.051549993455410004,
"learning_rate": 0.0008,
"loss": 2.226004123687744,
"step": 5700
},
{
"epoch": 0.3455757575757576,
"grad_norm": 0.05160880833864212,
"learning_rate": 0.0008,
"loss": 2.200514793395996,
"step": 5702
},
{
"epoch": 0.34569696969696967,
"grad_norm": 0.05211353302001953,
"learning_rate": 0.0008,
"loss": 2.287576913833618,
"step": 5704
},
{
"epoch": 0.3458181818181818,
"grad_norm": 0.05288524180650711,
"learning_rate": 0.0008,
"loss": 2.1471059322357178,
"step": 5706
},
{
"epoch": 0.34593939393939394,
"grad_norm": 0.05379557982087135,
"learning_rate": 0.0008,
"loss": 2.210874080657959,
"step": 5708
},
{
"epoch": 0.34606060606060607,
"grad_norm": 0.05126970633864403,
"learning_rate": 0.0008,
"loss": 2.243093729019165,
"step": 5710
},
{
"epoch": 0.3461818181818182,
"grad_norm": 0.05132444202899933,
"learning_rate": 0.0008,
"loss": 2.2574462890625,
"step": 5712
},
{
"epoch": 0.3463030303030303,
"grad_norm": 0.050138939172029495,
"learning_rate": 0.0008,
"loss": 2.2761781215667725,
"step": 5714
},
{
"epoch": 0.3464242424242424,
"grad_norm": 0.04876798763871193,
"learning_rate": 0.0008,
"loss": 2.154876708984375,
"step": 5716
},
{
"epoch": 0.34654545454545455,
"grad_norm": 0.058654773980379105,
"learning_rate": 0.0008,
"loss": 2.2656900882720947,
"step": 5718
},
{
"epoch": 0.3466666666666667,
"grad_norm": 0.05316869542002678,
"learning_rate": 0.0008,
"loss": 2.167518377304077,
"step": 5720
},
{
"epoch": 0.3467878787878788,
"grad_norm": 0.05404474213719368,
"learning_rate": 0.0008,
"loss": 2.280177593231201,
"step": 5722
},
{
"epoch": 0.3469090909090909,
"grad_norm": 0.05998310446739197,
"learning_rate": 0.0008,
"loss": 2.253704786300659,
"step": 5724
},
{
"epoch": 0.347030303030303,
"grad_norm": 0.05305926874279976,
"learning_rate": 0.0008,
"loss": 2.2147562503814697,
"step": 5726
},
{
"epoch": 0.34715151515151516,
"grad_norm": 0.05259893462061882,
"learning_rate": 0.0008,
"loss": 2.2170541286468506,
"step": 5728
},
{
"epoch": 0.3472727272727273,
"grad_norm": 0.051350027322769165,
"learning_rate": 0.0008,
"loss": 2.1836183071136475,
"step": 5730
},
{
"epoch": 0.3473939393939394,
"grad_norm": 0.05210857093334198,
"learning_rate": 0.0008,
"loss": 2.2656142711639404,
"step": 5732
},
{
"epoch": 0.3475151515151515,
"grad_norm": 0.051797155290842056,
"learning_rate": 0.0008,
"loss": 2.198730707168579,
"step": 5734
},
{
"epoch": 0.34763636363636363,
"grad_norm": 0.05294670909643173,
"learning_rate": 0.0008,
"loss": 2.3125829696655273,
"step": 5736
},
{
"epoch": 0.34775757575757577,
"grad_norm": 0.05324368178844452,
"learning_rate": 0.0008,
"loss": 2.264216661453247,
"step": 5738
},
{
"epoch": 0.3478787878787879,
"grad_norm": 0.05208118259906769,
"learning_rate": 0.0008,
"loss": 2.2259418964385986,
"step": 5740
},
{
"epoch": 0.348,
"grad_norm": 0.0505206398665905,
"learning_rate": 0.0008,
"loss": 2.2952849864959717,
"step": 5742
},
{
"epoch": 0.3481212121212121,
"grad_norm": 0.05104375630617142,
"learning_rate": 0.0008,
"loss": 2.3190646171569824,
"step": 5744
},
{
"epoch": 0.34824242424242424,
"grad_norm": 0.05111270397901535,
"learning_rate": 0.0008,
"loss": 2.260364532470703,
"step": 5746
},
{
"epoch": 0.3483636363636364,
"grad_norm": 0.057908400893211365,
"learning_rate": 0.0008,
"loss": 2.2589805126190186,
"step": 5748
},
{
"epoch": 0.3484848484848485,
"grad_norm": 0.062286969274282455,
"learning_rate": 0.0008,
"loss": 2.3042736053466797,
"step": 5750
},
{
"epoch": 0.3486060606060606,
"grad_norm": 0.05679725855588913,
"learning_rate": 0.0008,
"loss": 2.2226476669311523,
"step": 5752
},
{
"epoch": 0.3487272727272727,
"grad_norm": 0.05512818694114685,
"learning_rate": 0.0008,
"loss": 2.2947964668273926,
"step": 5754
},
{
"epoch": 0.34884848484848485,
"grad_norm": 0.05124705284833908,
"learning_rate": 0.0008,
"loss": 2.1772305965423584,
"step": 5756
},
{
"epoch": 0.348969696969697,
"grad_norm": 0.050652876496315,
"learning_rate": 0.0008,
"loss": 2.2506725788116455,
"step": 5758
},
{
"epoch": 0.3490909090909091,
"grad_norm": 0.05045401304960251,
"learning_rate": 0.0008,
"loss": 2.280979633331299,
"step": 5760
},
{
"epoch": 0.3492121212121212,
"grad_norm": 0.0495368130505085,
"learning_rate": 0.0008,
"loss": 2.1787960529327393,
"step": 5762
},
{
"epoch": 0.34933333333333333,
"grad_norm": 0.06129473075270653,
"learning_rate": 0.0008,
"loss": 2.1736268997192383,
"step": 5764
},
{
"epoch": 0.34945454545454546,
"grad_norm": 0.05075656995177269,
"learning_rate": 0.0008,
"loss": 2.2095906734466553,
"step": 5766
},
{
"epoch": 0.3495757575757576,
"grad_norm": 0.052088845521211624,
"learning_rate": 0.0008,
"loss": 2.230311393737793,
"step": 5768
},
{
"epoch": 0.3496969696969697,
"grad_norm": 0.05480481684207916,
"learning_rate": 0.0008,
"loss": 2.3177056312561035,
"step": 5770
},
{
"epoch": 0.3498181818181818,
"grad_norm": 0.060376718640327454,
"learning_rate": 0.0008,
"loss": 2.1911089420318604,
"step": 5772
},
{
"epoch": 0.34993939393939394,
"grad_norm": 0.057974208146333694,
"learning_rate": 0.0008,
"loss": 2.2164061069488525,
"step": 5774
},
{
"epoch": 0.35006060606060607,
"grad_norm": 0.051224347203969955,
"learning_rate": 0.0008,
"loss": 2.2568395137786865,
"step": 5776
},
{
"epoch": 0.3501818181818182,
"grad_norm": 0.04876340925693512,
"learning_rate": 0.0008,
"loss": 2.217000961303711,
"step": 5778
},
{
"epoch": 0.3503030303030303,
"grad_norm": 0.04899138584733009,
"learning_rate": 0.0008,
"loss": 2.151881217956543,
"step": 5780
},
{
"epoch": 0.3504242424242424,
"grad_norm": 0.052509456872940063,
"learning_rate": 0.0008,
"loss": 2.2933924198150635,
"step": 5782
},
{
"epoch": 0.35054545454545455,
"grad_norm": 0.0507633201777935,
"learning_rate": 0.0008,
"loss": 2.27144455909729,
"step": 5784
},
{
"epoch": 0.3506666666666667,
"grad_norm": 0.04968249052762985,
"learning_rate": 0.0008,
"loss": 2.210165500640869,
"step": 5786
},
{
"epoch": 0.3507878787878788,
"grad_norm": 0.05343864858150482,
"learning_rate": 0.0008,
"loss": 2.2131335735321045,
"step": 5788
},
{
"epoch": 0.3509090909090909,
"grad_norm": 0.05604918673634529,
"learning_rate": 0.0008,
"loss": 2.1325955390930176,
"step": 5790
},
{
"epoch": 0.351030303030303,
"grad_norm": 0.055685244500637054,
"learning_rate": 0.0008,
"loss": 2.128321886062622,
"step": 5792
},
{
"epoch": 0.35115151515151516,
"grad_norm": 0.05331191048026085,
"learning_rate": 0.0008,
"loss": 2.268179178237915,
"step": 5794
},
{
"epoch": 0.3512727272727273,
"grad_norm": 0.05296797305345535,
"learning_rate": 0.0008,
"loss": 2.240161180496216,
"step": 5796
},
{
"epoch": 0.35139393939393937,
"grad_norm": 0.05467648059129715,
"learning_rate": 0.0008,
"loss": 2.2356066703796387,
"step": 5798
},
{
"epoch": 0.3515151515151515,
"grad_norm": 0.05018698796629906,
"learning_rate": 0.0008,
"loss": 2.197781562805176,
"step": 5800
},
{
"epoch": 0.35163636363636364,
"grad_norm": 0.05200861394405365,
"learning_rate": 0.0008,
"loss": 2.237682580947876,
"step": 5802
},
{
"epoch": 0.35175757575757577,
"grad_norm": 0.059084370732307434,
"learning_rate": 0.0008,
"loss": 2.1893844604492188,
"step": 5804
},
{
"epoch": 0.3518787878787879,
"grad_norm": 0.05222151800990105,
"learning_rate": 0.0008,
"loss": 2.1846673488616943,
"step": 5806
},
{
"epoch": 0.352,
"grad_norm": 0.05128961428999901,
"learning_rate": 0.0008,
"loss": 2.3106870651245117,
"step": 5808
},
{
"epoch": 0.3521212121212121,
"grad_norm": 0.0522160679101944,
"learning_rate": 0.0008,
"loss": 2.146926164627075,
"step": 5810
},
{
"epoch": 0.35224242424242425,
"grad_norm": 0.04965825751423836,
"learning_rate": 0.0008,
"loss": 2.264050006866455,
"step": 5812
},
{
"epoch": 0.3523636363636364,
"grad_norm": 0.04947194457054138,
"learning_rate": 0.0008,
"loss": 2.238386631011963,
"step": 5814
},
{
"epoch": 0.3524848484848485,
"grad_norm": 0.05019660294055939,
"learning_rate": 0.0008,
"loss": 2.1913392543792725,
"step": 5816
},
{
"epoch": 0.3526060606060606,
"grad_norm": 0.05032924935221672,
"learning_rate": 0.0008,
"loss": 2.1960580348968506,
"step": 5818
},
{
"epoch": 0.3527272727272727,
"grad_norm": 0.0504014827311039,
"learning_rate": 0.0008,
"loss": 2.278998851776123,
"step": 5820
},
{
"epoch": 0.35284848484848486,
"grad_norm": 0.0505659356713295,
"learning_rate": 0.0008,
"loss": 2.2288665771484375,
"step": 5822
},
{
"epoch": 0.352969696969697,
"grad_norm": 0.0499691404402256,
"learning_rate": 0.0008,
"loss": 2.2447080612182617,
"step": 5824
},
{
"epoch": 0.35309090909090907,
"grad_norm": 0.05170738324522972,
"learning_rate": 0.0008,
"loss": 2.2230801582336426,
"step": 5826
},
{
"epoch": 0.3532121212121212,
"grad_norm": 0.04958108812570572,
"learning_rate": 0.0008,
"loss": 2.2766242027282715,
"step": 5828
},
{
"epoch": 0.35333333333333333,
"grad_norm": 0.05137578770518303,
"learning_rate": 0.0008,
"loss": 2.225234270095825,
"step": 5830
},
{
"epoch": 0.35345454545454547,
"grad_norm": 0.05264224857091904,
"learning_rate": 0.0008,
"loss": 2.2996184825897217,
"step": 5832
},
{
"epoch": 0.3535757575757576,
"grad_norm": 0.049507930874824524,
"learning_rate": 0.0008,
"loss": 2.305128812789917,
"step": 5834
},
{
"epoch": 0.3536969696969697,
"grad_norm": 0.049255434423685074,
"learning_rate": 0.0008,
"loss": 2.222486972808838,
"step": 5836
},
{
"epoch": 0.3538181818181818,
"grad_norm": 0.052102603018283844,
"learning_rate": 0.0008,
"loss": 2.235384702682495,
"step": 5838
},
{
"epoch": 0.35393939393939394,
"grad_norm": 0.051388151943683624,
"learning_rate": 0.0008,
"loss": 2.172370195388794,
"step": 5840
},
{
"epoch": 0.3540606060606061,
"grad_norm": 0.05260121822357178,
"learning_rate": 0.0008,
"loss": 2.225534439086914,
"step": 5842
},
{
"epoch": 0.3541818181818182,
"grad_norm": 0.05242770165205002,
"learning_rate": 0.0008,
"loss": 2.281484603881836,
"step": 5844
},
{
"epoch": 0.3543030303030303,
"grad_norm": 0.05004158988595009,
"learning_rate": 0.0008,
"loss": 2.2251474857330322,
"step": 5846
},
{
"epoch": 0.3544242424242424,
"grad_norm": 0.0497896745800972,
"learning_rate": 0.0008,
"loss": 2.303356885910034,
"step": 5848
},
{
"epoch": 0.35454545454545455,
"grad_norm": 0.052752017974853516,
"learning_rate": 0.0008,
"loss": 2.254265308380127,
"step": 5850
},
{
"epoch": 0.3546666666666667,
"grad_norm": 0.04928575083613396,
"learning_rate": 0.0008,
"loss": 2.1982831954956055,
"step": 5852
},
{
"epoch": 0.35478787878787876,
"grad_norm": 0.0529162622988224,
"learning_rate": 0.0008,
"loss": 2.2204995155334473,
"step": 5854
},
{
"epoch": 0.3549090909090909,
"grad_norm": 0.054794199764728546,
"learning_rate": 0.0008,
"loss": 2.206209421157837,
"step": 5856
},
{
"epoch": 0.35503030303030303,
"grad_norm": 0.05097115412354469,
"learning_rate": 0.0008,
"loss": 2.1855432987213135,
"step": 5858
},
{
"epoch": 0.35515151515151516,
"grad_norm": 0.05035323277115822,
"learning_rate": 0.0008,
"loss": 2.217373847961426,
"step": 5860
},
{
"epoch": 0.3552727272727273,
"grad_norm": 0.051241178065538406,
"learning_rate": 0.0008,
"loss": 2.1739895343780518,
"step": 5862
},
{
"epoch": 0.3553939393939394,
"grad_norm": 0.05091148987412453,
"learning_rate": 0.0008,
"loss": 2.23093843460083,
"step": 5864
},
{
"epoch": 0.3555151515151515,
"grad_norm": 0.054157521575689316,
"learning_rate": 0.0008,
"loss": 2.2372655868530273,
"step": 5866
},
{
"epoch": 0.35563636363636364,
"grad_norm": 0.05099915340542793,
"learning_rate": 0.0008,
"loss": 2.2815659046173096,
"step": 5868
},
{
"epoch": 0.3557575757575758,
"grad_norm": 0.05384884402155876,
"learning_rate": 0.0008,
"loss": 2.1772937774658203,
"step": 5870
},
{
"epoch": 0.3558787878787879,
"grad_norm": 0.05591750144958496,
"learning_rate": 0.0008,
"loss": 2.215122699737549,
"step": 5872
},
{
"epoch": 0.356,
"grad_norm": 0.05606465041637421,
"learning_rate": 0.0008,
"loss": 2.274747371673584,
"step": 5874
},
{
"epoch": 0.3561212121212121,
"grad_norm": 0.05729743838310242,
"learning_rate": 0.0008,
"loss": 2.244946241378784,
"step": 5876
},
{
"epoch": 0.35624242424242425,
"grad_norm": 0.06857825815677643,
"learning_rate": 0.0008,
"loss": 2.242947816848755,
"step": 5878
},
{
"epoch": 0.3563636363636364,
"grad_norm": 0.059537917375564575,
"learning_rate": 0.0008,
"loss": 2.1654393672943115,
"step": 5880
},
{
"epoch": 0.35648484848484846,
"grad_norm": 0.05120176076889038,
"learning_rate": 0.0008,
"loss": 2.1654577255249023,
"step": 5882
},
{
"epoch": 0.3566060606060606,
"grad_norm": 0.05275990813970566,
"learning_rate": 0.0008,
"loss": 2.280895233154297,
"step": 5884
},
{
"epoch": 0.3567272727272727,
"grad_norm": 0.050827737897634506,
"learning_rate": 0.0008,
"loss": 2.2094056606292725,
"step": 5886
},
{
"epoch": 0.35684848484848486,
"grad_norm": 0.052801698446273804,
"learning_rate": 0.0008,
"loss": 2.216451406478882,
"step": 5888
},
{
"epoch": 0.356969696969697,
"grad_norm": 0.0586533360183239,
"learning_rate": 0.0008,
"loss": 2.333327054977417,
"step": 5890
},
{
"epoch": 0.35709090909090907,
"grad_norm": 0.05421488732099533,
"learning_rate": 0.0008,
"loss": 2.2010035514831543,
"step": 5892
},
{
"epoch": 0.3572121212121212,
"grad_norm": 0.05009719729423523,
"learning_rate": 0.0008,
"loss": 2.1992385387420654,
"step": 5894
},
{
"epoch": 0.35733333333333334,
"grad_norm": 0.05126538872718811,
"learning_rate": 0.0008,
"loss": 2.2807118892669678,
"step": 5896
},
{
"epoch": 0.35745454545454547,
"grad_norm": 0.04860067740082741,
"learning_rate": 0.0008,
"loss": 2.1818220615386963,
"step": 5898
},
{
"epoch": 0.3575757575757576,
"grad_norm": 0.04948527738451958,
"learning_rate": 0.0008,
"loss": 2.2051446437835693,
"step": 5900
},
{
"epoch": 0.3576969696969697,
"grad_norm": 0.055695392191410065,
"learning_rate": 0.0008,
"loss": 2.249216318130493,
"step": 5902
},
{
"epoch": 0.3578181818181818,
"grad_norm": 0.054224688559770584,
"learning_rate": 0.0008,
"loss": 2.214036226272583,
"step": 5904
},
{
"epoch": 0.35793939393939395,
"grad_norm": 0.05558338761329651,
"learning_rate": 0.0008,
"loss": 2.224785804748535,
"step": 5906
},
{
"epoch": 0.3580606060606061,
"grad_norm": 0.05193831026554108,
"learning_rate": 0.0008,
"loss": 2.1324052810668945,
"step": 5908
},
{
"epoch": 0.35818181818181816,
"grad_norm": 0.050966426730155945,
"learning_rate": 0.0008,
"loss": 2.219444513320923,
"step": 5910
},
{
"epoch": 0.3583030303030303,
"grad_norm": 0.05016708746552467,
"learning_rate": 0.0008,
"loss": 2.2592897415161133,
"step": 5912
},
{
"epoch": 0.3584242424242424,
"grad_norm": 0.05058800056576729,
"learning_rate": 0.0008,
"loss": 2.2776989936828613,
"step": 5914
},
{
"epoch": 0.35854545454545456,
"grad_norm": 0.04920389875769615,
"learning_rate": 0.0008,
"loss": 2.1414108276367188,
"step": 5916
},
{
"epoch": 0.3586666666666667,
"grad_norm": 0.05084595829248428,
"learning_rate": 0.0008,
"loss": 2.165642738342285,
"step": 5918
},
{
"epoch": 0.35878787878787877,
"grad_norm": 0.05304887890815735,
"learning_rate": 0.0008,
"loss": 2.293698310852051,
"step": 5920
},
{
"epoch": 0.3589090909090909,
"grad_norm": 0.05216508358716965,
"learning_rate": 0.0008,
"loss": 2.1753621101379395,
"step": 5922
},
{
"epoch": 0.35903030303030303,
"grad_norm": 0.048683639615774155,
"learning_rate": 0.0008,
"loss": 2.2475831508636475,
"step": 5924
},
{
"epoch": 0.35915151515151517,
"grad_norm": 0.05030354484915733,
"learning_rate": 0.0008,
"loss": 2.300729513168335,
"step": 5926
},
{
"epoch": 0.3592727272727273,
"grad_norm": 0.05001804605126381,
"learning_rate": 0.0008,
"loss": 2.2447357177734375,
"step": 5928
},
{
"epoch": 0.3593939393939394,
"grad_norm": 0.05204688012599945,
"learning_rate": 0.0008,
"loss": 2.165228843688965,
"step": 5930
},
{
"epoch": 0.3595151515151515,
"grad_norm": 0.05048573762178421,
"learning_rate": 0.0008,
"loss": 2.124971628189087,
"step": 5932
},
{
"epoch": 0.35963636363636364,
"grad_norm": 0.05659837648272514,
"learning_rate": 0.0008,
"loss": 2.234976053237915,
"step": 5934
},
{
"epoch": 0.3597575757575758,
"grad_norm": 0.06276922672986984,
"learning_rate": 0.0008,
"loss": 2.1909615993499756,
"step": 5936
},
{
"epoch": 0.35987878787878785,
"grad_norm": 0.05822509154677391,
"learning_rate": 0.0008,
"loss": 2.200023651123047,
"step": 5938
},
{
"epoch": 0.36,
"grad_norm": 0.05549249425530434,
"learning_rate": 0.0008,
"loss": 2.194127082824707,
"step": 5940
},
{
"epoch": 0.3601212121212121,
"grad_norm": 0.05211513862013817,
"learning_rate": 0.0008,
"loss": 2.2123031616210938,
"step": 5942
},
{
"epoch": 0.36024242424242425,
"grad_norm": 0.052284125238657,
"learning_rate": 0.0008,
"loss": 2.2585718631744385,
"step": 5944
},
{
"epoch": 0.3603636363636364,
"grad_norm": 0.055511146783828735,
"learning_rate": 0.0008,
"loss": 2.158719062805176,
"step": 5946
},
{
"epoch": 0.36048484848484846,
"grad_norm": 0.062119320034980774,
"learning_rate": 0.0008,
"loss": 2.1786537170410156,
"step": 5948
},
{
"epoch": 0.3606060606060606,
"grad_norm": 0.06525465101003647,
"learning_rate": 0.0008,
"loss": 2.2519137859344482,
"step": 5950
},
{
"epoch": 0.36072727272727273,
"grad_norm": 0.060715898871421814,
"learning_rate": 0.0008,
"loss": 2.153093099594116,
"step": 5952
},
{
"epoch": 0.36084848484848486,
"grad_norm": 0.055444374680519104,
"learning_rate": 0.0008,
"loss": 2.273353338241577,
"step": 5954
},
{
"epoch": 0.360969696969697,
"grad_norm": 0.05603004992008209,
"learning_rate": 0.0008,
"loss": 2.217111110687256,
"step": 5956
},
{
"epoch": 0.3610909090909091,
"grad_norm": 0.054905492812395096,
"learning_rate": 0.0008,
"loss": 2.284252405166626,
"step": 5958
},
{
"epoch": 0.3612121212121212,
"grad_norm": 0.06032660976052284,
"learning_rate": 0.0008,
"loss": 2.1924009323120117,
"step": 5960
},
{
"epoch": 0.36133333333333334,
"grad_norm": 0.06775419414043427,
"learning_rate": 0.0008,
"loss": 2.190206289291382,
"step": 5962
},
{
"epoch": 0.3614545454545455,
"grad_norm": 0.0616842545568943,
"learning_rate": 0.0008,
"loss": 2.2234349250793457,
"step": 5964
},
{
"epoch": 0.36157575757575755,
"grad_norm": 0.05436982959508896,
"learning_rate": 0.0008,
"loss": 2.27829647064209,
"step": 5966
},
{
"epoch": 0.3616969696969697,
"grad_norm": 0.06430494785308838,
"learning_rate": 0.0008,
"loss": 2.203202247619629,
"step": 5968
},
{
"epoch": 0.3618181818181818,
"grad_norm": 0.05552124232053757,
"learning_rate": 0.0008,
"loss": 2.2188446521759033,
"step": 5970
},
{
"epoch": 0.36193939393939395,
"grad_norm": 0.05218339338898659,
"learning_rate": 0.0008,
"loss": 2.1821482181549072,
"step": 5972
},
{
"epoch": 0.3620606060606061,
"grad_norm": 0.052748244255781174,
"learning_rate": 0.0008,
"loss": 2.1574618816375732,
"step": 5974
},
{
"epoch": 0.36218181818181816,
"grad_norm": 0.052279677242040634,
"learning_rate": 0.0008,
"loss": 2.2395737171173096,
"step": 5976
},
{
"epoch": 0.3623030303030303,
"grad_norm": 0.05026300624012947,
"learning_rate": 0.0008,
"loss": 2.229888916015625,
"step": 5978
},
{
"epoch": 0.3624242424242424,
"grad_norm": 0.050100427120923996,
"learning_rate": 0.0008,
"loss": 2.2425882816314697,
"step": 5980
},
{
"epoch": 0.36254545454545456,
"grad_norm": 0.05422943830490112,
"learning_rate": 0.0008,
"loss": 2.2463479042053223,
"step": 5982
},
{
"epoch": 0.3626666666666667,
"grad_norm": 0.048990268260240555,
"learning_rate": 0.0008,
"loss": 2.205371141433716,
"step": 5984
},
{
"epoch": 0.36278787878787877,
"grad_norm": 0.05055132880806923,
"learning_rate": 0.0008,
"loss": 2.242563247680664,
"step": 5986
},
{
"epoch": 0.3629090909090909,
"grad_norm": 0.049552228301763535,
"learning_rate": 0.0008,
"loss": 2.3041598796844482,
"step": 5988
},
{
"epoch": 0.36303030303030304,
"grad_norm": 0.05168351158499718,
"learning_rate": 0.0008,
"loss": 2.3070425987243652,
"step": 5990
},
{
"epoch": 0.36315151515151517,
"grad_norm": 0.051634881645441055,
"learning_rate": 0.0008,
"loss": 2.214949369430542,
"step": 5992
},
{
"epoch": 0.36327272727272725,
"grad_norm": 0.06110695004463196,
"learning_rate": 0.0008,
"loss": 2.2856338024139404,
"step": 5994
},
{
"epoch": 0.3633939393939394,
"grad_norm": 0.05230925977230072,
"learning_rate": 0.0008,
"loss": 2.245950222015381,
"step": 5996
},
{
"epoch": 0.3635151515151515,
"grad_norm": 0.05315594747662544,
"learning_rate": 0.0008,
"loss": 2.219942331314087,
"step": 5998
},
{
"epoch": 0.36363636363636365,
"grad_norm": 0.05147039517760277,
"learning_rate": 0.0008,
"loss": 2.2160837650299072,
"step": 6000
},
{
"epoch": 0.3637575757575758,
"grad_norm": 0.05168053135275841,
"learning_rate": 0.0008,
"loss": 2.2122273445129395,
"step": 6002
},
{
"epoch": 0.36387878787878786,
"grad_norm": 0.050486356019973755,
"learning_rate": 0.0008,
"loss": 2.2154979705810547,
"step": 6004
},
{
"epoch": 0.364,
"grad_norm": 0.052653323858976364,
"learning_rate": 0.0008,
"loss": 2.2347280979156494,
"step": 6006
},
{
"epoch": 0.3641212121212121,
"grad_norm": 0.05572618544101715,
"learning_rate": 0.0008,
"loss": 2.190516710281372,
"step": 6008
},
{
"epoch": 0.36424242424242426,
"grad_norm": 0.050541702657938004,
"learning_rate": 0.0008,
"loss": 2.1981019973754883,
"step": 6010
},
{
"epoch": 0.3643636363636364,
"grad_norm": 0.050268176943063736,
"learning_rate": 0.0008,
"loss": 2.2077043056488037,
"step": 6012
},
{
"epoch": 0.36448484848484847,
"grad_norm": 0.04831693693995476,
"learning_rate": 0.0008,
"loss": 2.2322463989257812,
"step": 6014
},
{
"epoch": 0.3646060606060606,
"grad_norm": 0.04972399026155472,
"learning_rate": 0.0008,
"loss": 2.2406694889068604,
"step": 6016
},
{
"epoch": 0.36472727272727273,
"grad_norm": 0.05195030942559242,
"learning_rate": 0.0008,
"loss": 2.2685656547546387,
"step": 6018
},
{
"epoch": 0.36484848484848487,
"grad_norm": 0.050141334533691406,
"learning_rate": 0.0008,
"loss": 2.2527332305908203,
"step": 6020
},
{
"epoch": 0.36496969696969694,
"grad_norm": 0.053204260766506195,
"learning_rate": 0.0008,
"loss": 2.236208200454712,
"step": 6022
},
{
"epoch": 0.3650909090909091,
"grad_norm": 0.049010857939720154,
"learning_rate": 0.0008,
"loss": 2.2001149654388428,
"step": 6024
},
{
"epoch": 0.3652121212121212,
"grad_norm": 0.0561552457511425,
"learning_rate": 0.0008,
"loss": 2.2045228481292725,
"step": 6026
},
{
"epoch": 0.36533333333333334,
"grad_norm": 0.05037800595164299,
"learning_rate": 0.0008,
"loss": 2.2655913829803467,
"step": 6028
},
{
"epoch": 0.3654545454545455,
"grad_norm": 0.058155834674835205,
"learning_rate": 0.0008,
"loss": 2.176403522491455,
"step": 6030
},
{
"epoch": 0.36557575757575755,
"grad_norm": 0.05279046297073364,
"learning_rate": 0.0008,
"loss": 2.2391929626464844,
"step": 6032
},
{
"epoch": 0.3656969696969697,
"grad_norm": 0.05214262008666992,
"learning_rate": 0.0008,
"loss": 2.25041127204895,
"step": 6034
},
{
"epoch": 0.3658181818181818,
"grad_norm": 0.05131208524107933,
"learning_rate": 0.0008,
"loss": 2.2285661697387695,
"step": 6036
},
{
"epoch": 0.36593939393939395,
"grad_norm": 0.05090207979083061,
"learning_rate": 0.0008,
"loss": 2.199195623397827,
"step": 6038
},
{
"epoch": 0.3660606060606061,
"grad_norm": 0.05313055217266083,
"learning_rate": 0.0008,
"loss": 2.267254590988159,
"step": 6040
},
{
"epoch": 0.36618181818181816,
"grad_norm": 0.0630984753370285,
"learning_rate": 0.0008,
"loss": 2.1780216693878174,
"step": 6042
},
{
"epoch": 0.3663030303030303,
"grad_norm": 0.053812675178050995,
"learning_rate": 0.0008,
"loss": 2.255065679550171,
"step": 6044
},
{
"epoch": 0.36642424242424243,
"grad_norm": 0.06651737540960312,
"learning_rate": 0.0008,
"loss": 2.1683738231658936,
"step": 6046
},
{
"epoch": 0.36654545454545456,
"grad_norm": 0.056994449347257614,
"learning_rate": 0.0008,
"loss": 2.2058045864105225,
"step": 6048
},
{
"epoch": 0.36666666666666664,
"grad_norm": 0.0571809820830822,
"learning_rate": 0.0008,
"loss": 2.28831148147583,
"step": 6050
},
{
"epoch": 0.3667878787878788,
"grad_norm": 0.06408073753118515,
"learning_rate": 0.0008,
"loss": 2.2582314014434814,
"step": 6052
},
{
"epoch": 0.3669090909090909,
"grad_norm": 0.05994915962219238,
"learning_rate": 0.0008,
"loss": 2.2138938903808594,
"step": 6054
},
{
"epoch": 0.36703030303030304,
"grad_norm": 0.05550333857536316,
"learning_rate": 0.0008,
"loss": 2.2994329929351807,
"step": 6056
},
{
"epoch": 0.3671515151515152,
"grad_norm": 0.066077321767807,
"learning_rate": 0.0008,
"loss": 2.2848453521728516,
"step": 6058
},
{
"epoch": 0.36727272727272725,
"grad_norm": 0.06589488685131073,
"learning_rate": 0.0008,
"loss": 2.2980730533599854,
"step": 6060
},
{
"epoch": 0.3673939393939394,
"grad_norm": 0.05329607427120209,
"learning_rate": 0.0008,
"loss": 2.223531484603882,
"step": 6062
},
{
"epoch": 0.3675151515151515,
"grad_norm": 0.049316152930259705,
"learning_rate": 0.0008,
"loss": 2.165224075317383,
"step": 6064
},
{
"epoch": 0.36763636363636365,
"grad_norm": 0.049191903322935104,
"learning_rate": 0.0008,
"loss": 2.1829960346221924,
"step": 6066
},
{
"epoch": 0.3677575757575758,
"grad_norm": 0.0492112897336483,
"learning_rate": 0.0008,
"loss": 2.167872905731201,
"step": 6068
},
{
"epoch": 0.36787878787878786,
"grad_norm": 0.05084263160824776,
"learning_rate": 0.0008,
"loss": 2.19299578666687,
"step": 6070
},
{
"epoch": 0.368,
"grad_norm": 0.051008276641368866,
"learning_rate": 0.0008,
"loss": 2.207911729812622,
"step": 6072
},
{
"epoch": 0.3681212121212121,
"grad_norm": 0.05193276330828667,
"learning_rate": 0.0008,
"loss": 2.2523374557495117,
"step": 6074
},
{
"epoch": 0.36824242424242426,
"grad_norm": 0.050271302461624146,
"learning_rate": 0.0008,
"loss": 2.155559778213501,
"step": 6076
},
{
"epoch": 0.36836363636363634,
"grad_norm": 0.05088774114847183,
"learning_rate": 0.0008,
"loss": 2.287292242050171,
"step": 6078
},
{
"epoch": 0.36848484848484847,
"grad_norm": 0.05108281224966049,
"learning_rate": 0.0008,
"loss": 2.2243053913116455,
"step": 6080
},
{
"epoch": 0.3686060606060606,
"grad_norm": 0.04886860027909279,
"learning_rate": 0.0008,
"loss": 2.2227718830108643,
"step": 6082
},
{
"epoch": 0.36872727272727274,
"grad_norm": 0.049365099519491196,
"learning_rate": 0.0008,
"loss": 2.28366756439209,
"step": 6084
},
{
"epoch": 0.36884848484848487,
"grad_norm": 0.050788506865501404,
"learning_rate": 0.0008,
"loss": 2.2324438095092773,
"step": 6086
},
{
"epoch": 0.36896969696969695,
"grad_norm": 0.0511053130030632,
"learning_rate": 0.0008,
"loss": 2.3203349113464355,
"step": 6088
},
{
"epoch": 0.3690909090909091,
"grad_norm": 0.054228320717811584,
"learning_rate": 0.0008,
"loss": 2.2516839504241943,
"step": 6090
},
{
"epoch": 0.3692121212121212,
"grad_norm": 0.05526077747344971,
"learning_rate": 0.0008,
"loss": 2.253662109375,
"step": 6092
},
{
"epoch": 0.36933333333333335,
"grad_norm": 0.04964616894721985,
"learning_rate": 0.0008,
"loss": 2.1954116821289062,
"step": 6094
},
{
"epoch": 0.3694545454545455,
"grad_norm": 0.05059455335140228,
"learning_rate": 0.0008,
"loss": 2.2482478618621826,
"step": 6096
},
{
"epoch": 0.36957575757575756,
"grad_norm": 0.049772847443819046,
"learning_rate": 0.0008,
"loss": 2.169504165649414,
"step": 6098
},
{
"epoch": 0.3696969696969697,
"grad_norm": 0.054206687957048416,
"learning_rate": 0.0008,
"loss": 2.1710801124572754,
"step": 6100
},
{
"epoch": 0.3698181818181818,
"grad_norm": 0.054807450622320175,
"learning_rate": 0.0008,
"loss": 2.2272894382476807,
"step": 6102
},
{
"epoch": 0.36993939393939396,
"grad_norm": 0.049473002552986145,
"learning_rate": 0.0008,
"loss": 2.2298576831817627,
"step": 6104
},
{
"epoch": 0.37006060606060603,
"grad_norm": 0.051481377333402634,
"learning_rate": 0.0008,
"loss": 2.286006212234497,
"step": 6106
},
{
"epoch": 0.37018181818181817,
"grad_norm": 0.05448209121823311,
"learning_rate": 0.0008,
"loss": 2.2608487606048584,
"step": 6108
},
{
"epoch": 0.3703030303030303,
"grad_norm": 0.0503140427172184,
"learning_rate": 0.0008,
"loss": 2.2563023567199707,
"step": 6110
},
{
"epoch": 0.37042424242424243,
"grad_norm": 0.04971175640821457,
"learning_rate": 0.0008,
"loss": 2.170903444290161,
"step": 6112
},
{
"epoch": 0.37054545454545457,
"grad_norm": 0.04971685633063316,
"learning_rate": 0.0008,
"loss": 2.2403600215911865,
"step": 6114
},
{
"epoch": 0.37066666666666664,
"grad_norm": 0.051253896206617355,
"learning_rate": 0.0008,
"loss": 2.2610373497009277,
"step": 6116
},
{
"epoch": 0.3707878787878788,
"grad_norm": 0.04890467971563339,
"learning_rate": 0.0008,
"loss": 2.2180116176605225,
"step": 6118
},
{
"epoch": 0.3709090909090909,
"grad_norm": 0.05040240287780762,
"learning_rate": 0.0008,
"loss": 2.215742826461792,
"step": 6120
},
{
"epoch": 0.37103030303030304,
"grad_norm": 0.05265432596206665,
"learning_rate": 0.0008,
"loss": 2.286012887954712,
"step": 6122
},
{
"epoch": 0.3711515151515152,
"grad_norm": 0.05061691254377365,
"learning_rate": 0.0008,
"loss": 2.2005960941314697,
"step": 6124
},
{
"epoch": 0.37127272727272725,
"grad_norm": 0.050920989364385605,
"learning_rate": 0.0008,
"loss": 2.188429117202759,
"step": 6126
},
{
"epoch": 0.3713939393939394,
"grad_norm": 0.04863970726728439,
"learning_rate": 0.0008,
"loss": 2.2019481658935547,
"step": 6128
},
{
"epoch": 0.3715151515151515,
"grad_norm": 0.05246885493397713,
"learning_rate": 0.0008,
"loss": 2.2535645961761475,
"step": 6130
},
{
"epoch": 0.37163636363636365,
"grad_norm": 0.05000101029872894,
"learning_rate": 0.0008,
"loss": 2.228530168533325,
"step": 6132
},
{
"epoch": 0.37175757575757573,
"grad_norm": 0.05094544589519501,
"learning_rate": 0.0008,
"loss": 2.2263295650482178,
"step": 6134
},
{
"epoch": 0.37187878787878786,
"grad_norm": 0.05280892178416252,
"learning_rate": 0.0008,
"loss": 2.1958420276641846,
"step": 6136
},
{
"epoch": 0.372,
"grad_norm": 0.05082041397690773,
"learning_rate": 0.0008,
"loss": 2.254847288131714,
"step": 6138
},
{
"epoch": 0.37212121212121213,
"grad_norm": 0.051992569118738174,
"learning_rate": 0.0008,
"loss": 2.256202459335327,
"step": 6140
},
{
"epoch": 0.37224242424242426,
"grad_norm": 0.05474456399679184,
"learning_rate": 0.0008,
"loss": 2.2923784255981445,
"step": 6142
},
{
"epoch": 0.37236363636363634,
"grad_norm": 0.05078866332769394,
"learning_rate": 0.0008,
"loss": 2.269453525543213,
"step": 6144
},
{
"epoch": 0.3724848484848485,
"grad_norm": 0.0496428944170475,
"learning_rate": 0.0008,
"loss": 2.2083449363708496,
"step": 6146
},
{
"epoch": 0.3726060606060606,
"grad_norm": 0.04930863529443741,
"learning_rate": 0.0008,
"loss": 2.2248282432556152,
"step": 6148
},
{
"epoch": 0.37272727272727274,
"grad_norm": 0.04964311048388481,
"learning_rate": 0.0008,
"loss": 2.262681484222412,
"step": 6150
},
{
"epoch": 0.3728484848484849,
"grad_norm": 0.048174336552619934,
"learning_rate": 0.0008,
"loss": 2.211256980895996,
"step": 6152
},
{
"epoch": 0.37296969696969695,
"grad_norm": 0.04927174374461174,
"learning_rate": 0.0008,
"loss": 2.2057371139526367,
"step": 6154
},
{
"epoch": 0.3730909090909091,
"grad_norm": 0.05739181861281395,
"learning_rate": 0.0008,
"loss": 2.247854232788086,
"step": 6156
},
{
"epoch": 0.3732121212121212,
"grad_norm": 0.055204372853040695,
"learning_rate": 0.0008,
"loss": 2.2670583724975586,
"step": 6158
},
{
"epoch": 0.37333333333333335,
"grad_norm": 0.05920034274458885,
"learning_rate": 0.0008,
"loss": 2.1954143047332764,
"step": 6160
},
{
"epoch": 0.37345454545454543,
"grad_norm": 0.05637278035283089,
"learning_rate": 0.0008,
"loss": 2.1965529918670654,
"step": 6162
},
{
"epoch": 0.37357575757575756,
"grad_norm": 0.05422032251954079,
"learning_rate": 0.0008,
"loss": 2.211569309234619,
"step": 6164
},
{
"epoch": 0.3736969696969697,
"grad_norm": 0.05029464140534401,
"learning_rate": 0.0008,
"loss": 2.217831611633301,
"step": 6166
},
{
"epoch": 0.3738181818181818,
"grad_norm": 0.04886966943740845,
"learning_rate": 0.0008,
"loss": 2.2573680877685547,
"step": 6168
},
{
"epoch": 0.37393939393939396,
"grad_norm": 0.05174904689192772,
"learning_rate": 0.0008,
"loss": 2.210780143737793,
"step": 6170
},
{
"epoch": 0.37406060606060604,
"grad_norm": 0.05044989660382271,
"learning_rate": 0.0008,
"loss": 2.2110049724578857,
"step": 6172
},
{
"epoch": 0.37418181818181817,
"grad_norm": 0.05268499627709389,
"learning_rate": 0.0008,
"loss": 2.2577080726623535,
"step": 6174
},
{
"epoch": 0.3743030303030303,
"grad_norm": 0.05021002143621445,
"learning_rate": 0.0008,
"loss": 2.236021041870117,
"step": 6176
},
{
"epoch": 0.37442424242424244,
"grad_norm": 0.049366891384124756,
"learning_rate": 0.0008,
"loss": 2.234419107437134,
"step": 6178
},
{
"epoch": 0.37454545454545457,
"grad_norm": 0.05181357264518738,
"learning_rate": 0.0008,
"loss": 2.240183115005493,
"step": 6180
},
{
"epoch": 0.37466666666666665,
"grad_norm": 0.04892997443675995,
"learning_rate": 0.0008,
"loss": 2.197430372238159,
"step": 6182
},
{
"epoch": 0.3747878787878788,
"grad_norm": 0.051324665546417236,
"learning_rate": 0.0008,
"loss": 2.2826414108276367,
"step": 6184
},
{
"epoch": 0.3749090909090909,
"grad_norm": 0.052252065390348434,
"learning_rate": 0.0008,
"loss": 2.251877546310425,
"step": 6186
},
{
"epoch": 0.37503030303030305,
"grad_norm": 0.05598190426826477,
"learning_rate": 0.0008,
"loss": 2.226550817489624,
"step": 6188
},
{
"epoch": 0.3751515151515151,
"grad_norm": 0.05833970010280609,
"learning_rate": 0.0008,
"loss": 2.2451794147491455,
"step": 6190
},
{
"epoch": 0.37527272727272726,
"grad_norm": 0.049111515283584595,
"learning_rate": 0.0008,
"loss": 2.2056543827056885,
"step": 6192
},
{
"epoch": 0.3753939393939394,
"grad_norm": 0.04770319163799286,
"learning_rate": 0.0008,
"loss": 2.16849684715271,
"step": 6194
},
{
"epoch": 0.3755151515151515,
"grad_norm": 0.053695205599069595,
"learning_rate": 0.0008,
"loss": 2.2576370239257812,
"step": 6196
},
{
"epoch": 0.37563636363636366,
"grad_norm": 0.061715610325336456,
"learning_rate": 0.0008,
"loss": 2.289677381515503,
"step": 6198
},
{
"epoch": 0.37575757575757573,
"grad_norm": 0.07048241794109344,
"learning_rate": 0.0008,
"loss": 2.226992607116699,
"step": 6200
},
{
"epoch": 0.37587878787878787,
"grad_norm": 0.0841790959239006,
"learning_rate": 0.0008,
"loss": 2.2150533199310303,
"step": 6202
},
{
"epoch": 0.376,
"grad_norm": 0.08274037390947342,
"learning_rate": 0.0008,
"loss": 2.2749903202056885,
"step": 6204
},
{
"epoch": 0.37612121212121213,
"grad_norm": 0.08678814768791199,
"learning_rate": 0.0008,
"loss": 2.2010300159454346,
"step": 6206
},
{
"epoch": 0.37624242424242427,
"grad_norm": 0.07580674439668655,
"learning_rate": 0.0008,
"loss": 2.240554094314575,
"step": 6208
},
{
"epoch": 0.37636363636363634,
"grad_norm": 0.058067578822374344,
"learning_rate": 0.0008,
"loss": 2.224766492843628,
"step": 6210
},
{
"epoch": 0.3764848484848485,
"grad_norm": 0.049939341843128204,
"learning_rate": 0.0008,
"loss": 2.2053000926971436,
"step": 6212
},
{
"epoch": 0.3766060606060606,
"grad_norm": 0.05718298628926277,
"learning_rate": 0.0008,
"loss": 2.2760422229766846,
"step": 6214
},
{
"epoch": 0.37672727272727274,
"grad_norm": 0.07346386462450027,
"learning_rate": 0.0008,
"loss": 2.1909706592559814,
"step": 6216
},
{
"epoch": 0.3768484848484848,
"grad_norm": 0.06454505026340485,
"learning_rate": 0.0008,
"loss": 2.1972525119781494,
"step": 6218
},
{
"epoch": 0.37696969696969695,
"grad_norm": 0.054347068071365356,
"learning_rate": 0.0008,
"loss": 2.1655826568603516,
"step": 6220
},
{
"epoch": 0.3770909090909091,
"grad_norm": 0.05038495734333992,
"learning_rate": 0.0008,
"loss": 2.2374496459960938,
"step": 6222
},
{
"epoch": 0.3772121212121212,
"grad_norm": 0.047640617936849594,
"learning_rate": 0.0008,
"loss": 2.163755178451538,
"step": 6224
},
{
"epoch": 0.37733333333333335,
"grad_norm": 0.05077016353607178,
"learning_rate": 0.0008,
"loss": 2.302488088607788,
"step": 6226
},
{
"epoch": 0.37745454545454543,
"grad_norm": 0.05165749415755272,
"learning_rate": 0.0008,
"loss": 2.2601168155670166,
"step": 6228
},
{
"epoch": 0.37757575757575756,
"grad_norm": 0.04949644207954407,
"learning_rate": 0.0008,
"loss": 2.169654130935669,
"step": 6230
},
{
"epoch": 0.3776969696969697,
"grad_norm": 0.04993593320250511,
"learning_rate": 0.0008,
"loss": 2.2594282627105713,
"step": 6232
},
{
"epoch": 0.37781818181818183,
"grad_norm": 0.05190431699156761,
"learning_rate": 0.0008,
"loss": 2.1318650245666504,
"step": 6234
},
{
"epoch": 0.37793939393939396,
"grad_norm": 0.05481202155351639,
"learning_rate": 0.0008,
"loss": 2.2028281688690186,
"step": 6236
},
{
"epoch": 0.37806060606060604,
"grad_norm": 0.061077363789081573,
"learning_rate": 0.0008,
"loss": 2.222536325454712,
"step": 6238
},
{
"epoch": 0.3781818181818182,
"grad_norm": 0.053542282432317734,
"learning_rate": 0.0008,
"loss": 2.1759932041168213,
"step": 6240
},
{
"epoch": 0.3783030303030303,
"grad_norm": 0.05176479369401932,
"learning_rate": 0.0008,
"loss": 2.1703450679779053,
"step": 6242
},
{
"epoch": 0.37842424242424244,
"grad_norm": 0.0537683442234993,
"learning_rate": 0.0008,
"loss": 2.228435754776001,
"step": 6244
},
{
"epoch": 0.3785454545454545,
"grad_norm": 0.05257045105099678,
"learning_rate": 0.0008,
"loss": 2.2556185722351074,
"step": 6246
},
{
"epoch": 0.37866666666666665,
"grad_norm": 0.06341647356748581,
"learning_rate": 0.0008,
"loss": 2.2022132873535156,
"step": 6248
},
{
"epoch": 0.3787878787878788,
"grad_norm": 0.05227024480700493,
"learning_rate": 0.0008,
"loss": 2.1921825408935547,
"step": 6250
},
{
"epoch": 0.3789090909090909,
"grad_norm": 0.05117889493703842,
"learning_rate": 0.0008,
"loss": 2.231248617172241,
"step": 6252
},
{
"epoch": 0.37903030303030305,
"grad_norm": 0.04905521497130394,
"learning_rate": 0.0008,
"loss": 2.2654106616973877,
"step": 6254
},
{
"epoch": 0.37915151515151513,
"grad_norm": 0.05297402665019035,
"learning_rate": 0.0008,
"loss": 2.207658290863037,
"step": 6256
},
{
"epoch": 0.37927272727272726,
"grad_norm": 0.051013462245464325,
"learning_rate": 0.0008,
"loss": 2.1639931201934814,
"step": 6258
},
{
"epoch": 0.3793939393939394,
"grad_norm": 0.04918723925948143,
"learning_rate": 0.0008,
"loss": 2.167147159576416,
"step": 6260
},
{
"epoch": 0.3795151515151515,
"grad_norm": 0.049193043261766434,
"learning_rate": 0.0008,
"loss": 2.142843723297119,
"step": 6262
},
{
"epoch": 0.37963636363636366,
"grad_norm": 0.04916848614811897,
"learning_rate": 0.0008,
"loss": 2.261798143386841,
"step": 6264
},
{
"epoch": 0.37975757575757574,
"grad_norm": 0.04940469190478325,
"learning_rate": 0.0008,
"loss": 2.2236576080322266,
"step": 6266
},
{
"epoch": 0.37987878787878787,
"grad_norm": 0.049694325774908066,
"learning_rate": 0.0008,
"loss": 2.1856071949005127,
"step": 6268
},
{
"epoch": 0.38,
"grad_norm": 0.049508631229400635,
"learning_rate": 0.0008,
"loss": 2.2739763259887695,
"step": 6270
},
{
"epoch": 0.38012121212121214,
"grad_norm": 0.05809696763753891,
"learning_rate": 0.0008,
"loss": 2.2481510639190674,
"step": 6272
},
{
"epoch": 0.3802424242424242,
"grad_norm": 0.05417311191558838,
"learning_rate": 0.0008,
"loss": 2.2409439086914062,
"step": 6274
},
{
"epoch": 0.38036363636363635,
"grad_norm": 0.054346419870853424,
"learning_rate": 0.0008,
"loss": 2.2318496704101562,
"step": 6276
},
{
"epoch": 0.3804848484848485,
"grad_norm": 0.049952294677495956,
"learning_rate": 0.0008,
"loss": 2.269801616668701,
"step": 6278
},
{
"epoch": 0.3806060606060606,
"grad_norm": 0.05076064169406891,
"learning_rate": 0.0008,
"loss": 2.1728017330169678,
"step": 6280
},
{
"epoch": 0.38072727272727275,
"grad_norm": 0.051997460424900055,
"learning_rate": 0.0008,
"loss": 2.1790432929992676,
"step": 6282
},
{
"epoch": 0.3808484848484848,
"grad_norm": 0.054925303906202316,
"learning_rate": 0.0008,
"loss": 2.2384307384490967,
"step": 6284
},
{
"epoch": 0.38096969696969696,
"grad_norm": 0.061243779957294464,
"learning_rate": 0.0008,
"loss": 2.1593363285064697,
"step": 6286
},
{
"epoch": 0.3810909090909091,
"grad_norm": 0.050257738679647446,
"learning_rate": 0.0008,
"loss": 2.185220241546631,
"step": 6288
},
{
"epoch": 0.3812121212121212,
"grad_norm": 0.05040682852268219,
"learning_rate": 0.0008,
"loss": 2.225395441055298,
"step": 6290
},
{
"epoch": 0.38133333333333336,
"grad_norm": 0.05608825758099556,
"learning_rate": 0.0008,
"loss": 2.2375683784484863,
"step": 6292
},
{
"epoch": 0.38145454545454544,
"grad_norm": 0.04969557747244835,
"learning_rate": 0.0008,
"loss": 2.2272183895111084,
"step": 6294
},
{
"epoch": 0.38157575757575757,
"grad_norm": 0.0490078330039978,
"learning_rate": 0.0008,
"loss": 2.3075642585754395,
"step": 6296
},
{
"epoch": 0.3816969696969697,
"grad_norm": 0.04936710372567177,
"learning_rate": 0.0008,
"loss": 2.2744226455688477,
"step": 6298
},
{
"epoch": 0.38181818181818183,
"grad_norm": 0.05108270049095154,
"learning_rate": 0.0008,
"loss": 2.254473924636841,
"step": 6300
},
{
"epoch": 0.3819393939393939,
"grad_norm": 0.0483689084649086,
"learning_rate": 0.0008,
"loss": 2.2168774604797363,
"step": 6302
},
{
"epoch": 0.38206060606060605,
"grad_norm": 0.05109141021966934,
"learning_rate": 0.0008,
"loss": 2.2301535606384277,
"step": 6304
},
{
"epoch": 0.3821818181818182,
"grad_norm": 0.05004499480128288,
"learning_rate": 0.0008,
"loss": 2.180492401123047,
"step": 6306
},
{
"epoch": 0.3823030303030303,
"grad_norm": 0.05054722726345062,
"learning_rate": 0.0008,
"loss": 2.2253944873809814,
"step": 6308
},
{
"epoch": 0.38242424242424244,
"grad_norm": 0.05035874992609024,
"learning_rate": 0.0008,
"loss": 2.2361648082733154,
"step": 6310
},
{
"epoch": 0.3825454545454545,
"grad_norm": 0.05376135930418968,
"learning_rate": 0.0008,
"loss": 2.1787891387939453,
"step": 6312
},
{
"epoch": 0.38266666666666665,
"grad_norm": 0.052339114248752594,
"learning_rate": 0.0008,
"loss": 2.1741912364959717,
"step": 6314
},
{
"epoch": 0.3827878787878788,
"grad_norm": 0.04956478998064995,
"learning_rate": 0.0008,
"loss": 2.240274429321289,
"step": 6316
},
{
"epoch": 0.3829090909090909,
"grad_norm": 0.05031091347336769,
"learning_rate": 0.0008,
"loss": 2.192474603652954,
"step": 6318
},
{
"epoch": 0.38303030303030305,
"grad_norm": 0.04969346523284912,
"learning_rate": 0.0008,
"loss": 2.1712119579315186,
"step": 6320
},
{
"epoch": 0.38315151515151513,
"grad_norm": 0.054034024477005005,
"learning_rate": 0.0008,
"loss": 2.212092161178589,
"step": 6322
},
{
"epoch": 0.38327272727272726,
"grad_norm": 0.06079046055674553,
"learning_rate": 0.0008,
"loss": 2.2803564071655273,
"step": 6324
},
{
"epoch": 0.3833939393939394,
"grad_norm": 0.0592387355864048,
"learning_rate": 0.0008,
"loss": 2.1799449920654297,
"step": 6326
},
{
"epoch": 0.38351515151515153,
"grad_norm": 0.059609390795230865,
"learning_rate": 0.0008,
"loss": 2.215428352355957,
"step": 6328
},
{
"epoch": 0.3836363636363636,
"grad_norm": 0.06657528877258301,
"learning_rate": 0.0008,
"loss": 2.2607271671295166,
"step": 6330
},
{
"epoch": 0.38375757575757574,
"grad_norm": 0.06488171219825745,
"learning_rate": 0.0008,
"loss": 2.1818366050720215,
"step": 6332
},
{
"epoch": 0.3838787878787879,
"grad_norm": 0.06056765466928482,
"learning_rate": 0.0008,
"loss": 2.2213244438171387,
"step": 6334
},
{
"epoch": 0.384,
"grad_norm": 0.05593273788690567,
"learning_rate": 0.0008,
"loss": 2.228307008743286,
"step": 6336
},
{
"epoch": 0.38412121212121214,
"grad_norm": 0.06115461513400078,
"learning_rate": 0.0008,
"loss": 2.208317995071411,
"step": 6338
},
{
"epoch": 0.3842424242424242,
"grad_norm": 0.049011390656232834,
"learning_rate": 0.0008,
"loss": 2.2096173763275146,
"step": 6340
},
{
"epoch": 0.38436363636363635,
"grad_norm": 0.05031735077500343,
"learning_rate": 0.0008,
"loss": 2.1890251636505127,
"step": 6342
},
{
"epoch": 0.3844848484848485,
"grad_norm": 0.04905756562948227,
"learning_rate": 0.0008,
"loss": 2.15727162361145,
"step": 6344
},
{
"epoch": 0.3846060606060606,
"grad_norm": 0.0514572374522686,
"learning_rate": 0.0008,
"loss": 2.216355085372925,
"step": 6346
},
{
"epoch": 0.38472727272727275,
"grad_norm": 0.05677633360028267,
"learning_rate": 0.0008,
"loss": 2.2253055572509766,
"step": 6348
},
{
"epoch": 0.38484848484848483,
"grad_norm": 0.06390445679426193,
"learning_rate": 0.0008,
"loss": 2.250971794128418,
"step": 6350
},
{
"epoch": 0.38496969696969696,
"grad_norm": 0.07571981102228165,
"learning_rate": 0.0008,
"loss": 2.244896411895752,
"step": 6352
},
{
"epoch": 0.3850909090909091,
"grad_norm": 0.058065660297870636,
"learning_rate": 0.0008,
"loss": 2.2012336254119873,
"step": 6354
},
{
"epoch": 0.38521212121212123,
"grad_norm": 0.05337086319923401,
"learning_rate": 0.0008,
"loss": 2.209510564804077,
"step": 6356
},
{
"epoch": 0.38533333333333336,
"grad_norm": 0.04829394072294235,
"learning_rate": 0.0008,
"loss": 2.184954881668091,
"step": 6358
},
{
"epoch": 0.38545454545454544,
"grad_norm": 0.04974988475441933,
"learning_rate": 0.0008,
"loss": 2.2517805099487305,
"step": 6360
},
{
"epoch": 0.38557575757575757,
"grad_norm": 0.05004474148154259,
"learning_rate": 0.0008,
"loss": 2.2252962589263916,
"step": 6362
},
{
"epoch": 0.3856969696969697,
"grad_norm": 0.04782189801335335,
"learning_rate": 0.0008,
"loss": 2.2429087162017822,
"step": 6364
},
{
"epoch": 0.38581818181818184,
"grad_norm": 0.05035564675927162,
"learning_rate": 0.0008,
"loss": 2.2658426761627197,
"step": 6366
},
{
"epoch": 0.3859393939393939,
"grad_norm": 0.048933595418930054,
"learning_rate": 0.0008,
"loss": 2.1446847915649414,
"step": 6368
},
{
"epoch": 0.38606060606060605,
"grad_norm": 0.04990734905004501,
"learning_rate": 0.0008,
"loss": 2.2122485637664795,
"step": 6370
},
{
"epoch": 0.3861818181818182,
"grad_norm": 0.0479319803416729,
"learning_rate": 0.0008,
"loss": 2.2556724548339844,
"step": 6372
},
{
"epoch": 0.3863030303030303,
"grad_norm": 0.04805348441004753,
"learning_rate": 0.0008,
"loss": 2.21808123588562,
"step": 6374
},
{
"epoch": 0.38642424242424245,
"grad_norm": 0.05676881968975067,
"learning_rate": 0.0008,
"loss": 2.264660596847534,
"step": 6376
},
{
"epoch": 0.3865454545454545,
"grad_norm": 0.053430572152137756,
"learning_rate": 0.0008,
"loss": 2.164060592651367,
"step": 6378
},
{
"epoch": 0.38666666666666666,
"grad_norm": 0.050887081772089005,
"learning_rate": 0.0008,
"loss": 2.2611799240112305,
"step": 6380
},
{
"epoch": 0.3867878787878788,
"grad_norm": 0.05261319875717163,
"learning_rate": 0.0008,
"loss": 2.2947843074798584,
"step": 6382
},
{
"epoch": 0.3869090909090909,
"grad_norm": 0.049361035227775574,
"learning_rate": 0.0008,
"loss": 2.223756790161133,
"step": 6384
},
{
"epoch": 0.38703030303030306,
"grad_norm": 0.050372201949357986,
"learning_rate": 0.0008,
"loss": 2.201510429382324,
"step": 6386
},
{
"epoch": 0.38715151515151514,
"grad_norm": 0.05341761186718941,
"learning_rate": 0.0008,
"loss": 2.2166991233825684,
"step": 6388
},
{
"epoch": 0.38727272727272727,
"grad_norm": 0.05008848011493683,
"learning_rate": 0.0008,
"loss": 2.2103078365325928,
"step": 6390
},
{
"epoch": 0.3873939393939394,
"grad_norm": 0.05050990730524063,
"learning_rate": 0.0008,
"loss": 2.2913365364074707,
"step": 6392
},
{
"epoch": 0.38751515151515153,
"grad_norm": 0.05057825893163681,
"learning_rate": 0.0008,
"loss": 2.2233805656433105,
"step": 6394
},
{
"epoch": 0.3876363636363636,
"grad_norm": 0.047817353159189224,
"learning_rate": 0.0008,
"loss": 2.2088310718536377,
"step": 6396
},
{
"epoch": 0.38775757575757575,
"grad_norm": 0.05160114914178848,
"learning_rate": 0.0008,
"loss": 2.3090710639953613,
"step": 6398
},
{
"epoch": 0.3878787878787879,
"grad_norm": 0.04895404353737831,
"learning_rate": 0.0008,
"loss": 2.246213674545288,
"step": 6400
},
{
"epoch": 0.388,
"grad_norm": 0.050217095762491226,
"learning_rate": 0.0008,
"loss": 2.207036256790161,
"step": 6402
},
{
"epoch": 0.38812121212121214,
"grad_norm": 0.04855349659919739,
"learning_rate": 0.0008,
"loss": 2.224496841430664,
"step": 6404
},
{
"epoch": 0.3882424242424242,
"grad_norm": 0.04890432581305504,
"learning_rate": 0.0008,
"loss": 2.212674617767334,
"step": 6406
},
{
"epoch": 0.38836363636363636,
"grad_norm": 0.04979965090751648,
"learning_rate": 0.0008,
"loss": 2.2038590908050537,
"step": 6408
},
{
"epoch": 0.3884848484848485,
"grad_norm": 0.05360913649201393,
"learning_rate": 0.0008,
"loss": 2.297715663909912,
"step": 6410
},
{
"epoch": 0.3886060606060606,
"grad_norm": 0.04997607693076134,
"learning_rate": 0.0008,
"loss": 2.1938047409057617,
"step": 6412
},
{
"epoch": 0.38872727272727275,
"grad_norm": 0.05007851496338844,
"learning_rate": 0.0008,
"loss": 2.198521852493286,
"step": 6414
},
{
"epoch": 0.38884848484848483,
"grad_norm": 0.04870527982711792,
"learning_rate": 0.0008,
"loss": 2.233105182647705,
"step": 6416
},
{
"epoch": 0.38896969696969697,
"grad_norm": 0.04975539445877075,
"learning_rate": 0.0008,
"loss": 2.2931761741638184,
"step": 6418
},
{
"epoch": 0.3890909090909091,
"grad_norm": 0.052925147116184235,
"learning_rate": 0.0008,
"loss": 2.2939984798431396,
"step": 6420
},
{
"epoch": 0.38921212121212123,
"grad_norm": 0.04909331724047661,
"learning_rate": 0.0008,
"loss": 2.180140733718872,
"step": 6422
},
{
"epoch": 0.3893333333333333,
"grad_norm": 0.05284643545746803,
"learning_rate": 0.0008,
"loss": 2.156909465789795,
"step": 6424
},
{
"epoch": 0.38945454545454544,
"grad_norm": 0.05343027785420418,
"learning_rate": 0.0008,
"loss": 2.23162579536438,
"step": 6426
},
{
"epoch": 0.3895757575757576,
"grad_norm": 0.053300272673368454,
"learning_rate": 0.0008,
"loss": 2.2059149742126465,
"step": 6428
},
{
"epoch": 0.3896969696969697,
"grad_norm": 0.04904617369174957,
"learning_rate": 0.0008,
"loss": 2.2786495685577393,
"step": 6430
},
{
"epoch": 0.38981818181818184,
"grad_norm": 0.051839299499988556,
"learning_rate": 0.0008,
"loss": 2.1941874027252197,
"step": 6432
},
{
"epoch": 0.3899393939393939,
"grad_norm": 0.056227363646030426,
"learning_rate": 0.0008,
"loss": 2.28531551361084,
"step": 6434
},
{
"epoch": 0.39006060606060605,
"grad_norm": 0.06748947501182556,
"learning_rate": 0.0008,
"loss": 2.322948694229126,
"step": 6436
},
{
"epoch": 0.3901818181818182,
"grad_norm": 0.05277034640312195,
"learning_rate": 0.0008,
"loss": 2.193542003631592,
"step": 6438
},
{
"epoch": 0.3903030303030303,
"grad_norm": 0.05266221985220909,
"learning_rate": 0.0008,
"loss": 2.273836135864258,
"step": 6440
},
{
"epoch": 0.39042424242424245,
"grad_norm": 0.04864291846752167,
"learning_rate": 0.0008,
"loss": 2.249382734298706,
"step": 6442
},
{
"epoch": 0.39054545454545453,
"grad_norm": 0.050327349454164505,
"learning_rate": 0.0008,
"loss": 2.2686824798583984,
"step": 6444
},
{
"epoch": 0.39066666666666666,
"grad_norm": 0.04822613298892975,
"learning_rate": 0.0008,
"loss": 2.211977243423462,
"step": 6446
},
{
"epoch": 0.3907878787878788,
"grad_norm": 0.04914354160428047,
"learning_rate": 0.0008,
"loss": 2.2218399047851562,
"step": 6448
},
{
"epoch": 0.39090909090909093,
"grad_norm": 0.050543010234832764,
"learning_rate": 0.0008,
"loss": 2.2831199169158936,
"step": 6450
},
{
"epoch": 0.391030303030303,
"grad_norm": 0.05098201334476471,
"learning_rate": 0.0008,
"loss": 2.3202908039093018,
"step": 6452
},
{
"epoch": 0.39115151515151514,
"grad_norm": 0.05131610110402107,
"learning_rate": 0.0008,
"loss": 2.24808931350708,
"step": 6454
},
{
"epoch": 0.39127272727272727,
"grad_norm": 0.05023717135190964,
"learning_rate": 0.0008,
"loss": 2.274474859237671,
"step": 6456
},
{
"epoch": 0.3913939393939394,
"grad_norm": 0.05194488540291786,
"learning_rate": 0.0008,
"loss": 2.236548900604248,
"step": 6458
},
{
"epoch": 0.39151515151515154,
"grad_norm": 0.0487709604203701,
"learning_rate": 0.0008,
"loss": 2.185894250869751,
"step": 6460
},
{
"epoch": 0.3916363636363636,
"grad_norm": 0.05163448676466942,
"learning_rate": 0.0008,
"loss": 2.2043347358703613,
"step": 6462
},
{
"epoch": 0.39175757575757575,
"grad_norm": 0.0495079830288887,
"learning_rate": 0.0008,
"loss": 2.2601466178894043,
"step": 6464
},
{
"epoch": 0.3918787878787879,
"grad_norm": 0.05631306394934654,
"learning_rate": 0.0008,
"loss": 2.17828631401062,
"step": 6466
},
{
"epoch": 0.392,
"grad_norm": 0.055871475487947464,
"learning_rate": 0.0008,
"loss": 2.202779531478882,
"step": 6468
},
{
"epoch": 0.39212121212121215,
"grad_norm": 0.04914116486907005,
"learning_rate": 0.0008,
"loss": 2.2536051273345947,
"step": 6470
},
{
"epoch": 0.3922424242424242,
"grad_norm": 0.05032382532954216,
"learning_rate": 0.0008,
"loss": 2.2673299312591553,
"step": 6472
},
{
"epoch": 0.39236363636363636,
"grad_norm": 0.049736201763153076,
"learning_rate": 0.0008,
"loss": 2.250762701034546,
"step": 6474
},
{
"epoch": 0.3924848484848485,
"grad_norm": 0.049901120364665985,
"learning_rate": 0.0008,
"loss": 2.1691651344299316,
"step": 6476
},
{
"epoch": 0.3926060606060606,
"grad_norm": 0.0499703474342823,
"learning_rate": 0.0008,
"loss": 2.149733066558838,
"step": 6478
},
{
"epoch": 0.3927272727272727,
"grad_norm": 0.05135257542133331,
"learning_rate": 0.0008,
"loss": 2.2155306339263916,
"step": 6480
},
{
"epoch": 0.39284848484848484,
"grad_norm": 0.05229981988668442,
"learning_rate": 0.0008,
"loss": 2.1990163326263428,
"step": 6482
},
{
"epoch": 0.39296969696969697,
"grad_norm": 0.05177270993590355,
"learning_rate": 0.0008,
"loss": 2.2972846031188965,
"step": 6484
},
{
"epoch": 0.3930909090909091,
"grad_norm": 0.056560300290584564,
"learning_rate": 0.0008,
"loss": 2.253481149673462,
"step": 6486
},
{
"epoch": 0.39321212121212124,
"grad_norm": 0.04903853312134743,
"learning_rate": 0.0008,
"loss": 2.258401870727539,
"step": 6488
},
{
"epoch": 0.3933333333333333,
"grad_norm": 0.0483318492770195,
"learning_rate": 0.0008,
"loss": 2.2456719875335693,
"step": 6490
},
{
"epoch": 0.39345454545454545,
"grad_norm": 0.05128072202205658,
"learning_rate": 0.0008,
"loss": 2.2595059871673584,
"step": 6492
},
{
"epoch": 0.3935757575757576,
"grad_norm": 0.048993296921253204,
"learning_rate": 0.0008,
"loss": 2.3007662296295166,
"step": 6494
},
{
"epoch": 0.3936969696969697,
"grad_norm": 0.050383422523736954,
"learning_rate": 0.0008,
"loss": 2.276503324508667,
"step": 6496
},
{
"epoch": 0.39381818181818184,
"grad_norm": 0.0493372417986393,
"learning_rate": 0.0008,
"loss": 2.2360477447509766,
"step": 6498
},
{
"epoch": 0.3939393939393939,
"grad_norm": 0.0497635155916214,
"learning_rate": 0.0008,
"loss": 2.226583480834961,
"step": 6500
},
{
"epoch": 0.39406060606060606,
"grad_norm": 0.04942116141319275,
"learning_rate": 0.0008,
"loss": 2.2736189365386963,
"step": 6502
},
{
"epoch": 0.3941818181818182,
"grad_norm": 0.05127545818686485,
"learning_rate": 0.0008,
"loss": 2.259828567504883,
"step": 6504
},
{
"epoch": 0.3943030303030303,
"grad_norm": 0.050768181681632996,
"learning_rate": 0.0008,
"loss": 2.198474168777466,
"step": 6506
},
{
"epoch": 0.3944242424242424,
"grad_norm": 0.05422455817461014,
"learning_rate": 0.0008,
"loss": 2.1971657276153564,
"step": 6508
},
{
"epoch": 0.39454545454545453,
"grad_norm": 0.06475596874952316,
"learning_rate": 0.0008,
"loss": 2.21722674369812,
"step": 6510
},
{
"epoch": 0.39466666666666667,
"grad_norm": 0.07344599813222885,
"learning_rate": 0.0008,
"loss": 2.2495920658111572,
"step": 6512
},
{
"epoch": 0.3947878787878788,
"grad_norm": 0.059748511761426926,
"learning_rate": 0.0008,
"loss": 2.2041866779327393,
"step": 6514
},
{
"epoch": 0.39490909090909093,
"grad_norm": 0.06446664035320282,
"learning_rate": 0.0008,
"loss": 2.320448875427246,
"step": 6516
},
{
"epoch": 0.395030303030303,
"grad_norm": 0.05999482050538063,
"learning_rate": 0.0008,
"loss": 2.201706647872925,
"step": 6518
},
{
"epoch": 0.39515151515151514,
"grad_norm": 0.06591591984033585,
"learning_rate": 0.0008,
"loss": 2.292060375213623,
"step": 6520
},
{
"epoch": 0.3952727272727273,
"grad_norm": 0.055587079375982285,
"learning_rate": 0.0008,
"loss": 2.294059991836548,
"step": 6522
},
{
"epoch": 0.3953939393939394,
"grad_norm": 0.06446746736764908,
"learning_rate": 0.0008,
"loss": 2.248838424682617,
"step": 6524
},
{
"epoch": 0.39551515151515154,
"grad_norm": 0.061107996851205826,
"learning_rate": 0.0008,
"loss": 2.1758627891540527,
"step": 6526
},
{
"epoch": 0.3956363636363636,
"grad_norm": 0.05465291067957878,
"learning_rate": 0.0008,
"loss": 2.1700990200042725,
"step": 6528
},
{
"epoch": 0.39575757575757575,
"grad_norm": 0.051772139966487885,
"learning_rate": 0.0008,
"loss": 2.2223782539367676,
"step": 6530
},
{
"epoch": 0.3958787878787879,
"grad_norm": 0.05172950401902199,
"learning_rate": 0.0008,
"loss": 2.2226130962371826,
"step": 6532
},
{
"epoch": 0.396,
"grad_norm": 0.05341942980885506,
"learning_rate": 0.0008,
"loss": 2.214386463165283,
"step": 6534
},
{
"epoch": 0.3961212121212121,
"grad_norm": 0.04965854063630104,
"learning_rate": 0.0008,
"loss": 2.140707492828369,
"step": 6536
},
{
"epoch": 0.39624242424242423,
"grad_norm": 0.04925324022769928,
"learning_rate": 0.0008,
"loss": 2.2285547256469727,
"step": 6538
},
{
"epoch": 0.39636363636363636,
"grad_norm": 0.04928707331418991,
"learning_rate": 0.0008,
"loss": 2.2021400928497314,
"step": 6540
},
{
"epoch": 0.3964848484848485,
"grad_norm": 0.048753753304481506,
"learning_rate": 0.0008,
"loss": 2.1719553470611572,
"step": 6542
},
{
"epoch": 0.39660606060606063,
"grad_norm": 0.04959544911980629,
"learning_rate": 0.0008,
"loss": 2.1610748767852783,
"step": 6544
},
{
"epoch": 0.3967272727272727,
"grad_norm": 0.05592402443289757,
"learning_rate": 0.0008,
"loss": 2.2980222702026367,
"step": 6546
},
{
"epoch": 0.39684848484848484,
"grad_norm": 0.0558185875415802,
"learning_rate": 0.0008,
"loss": 2.2639613151550293,
"step": 6548
},
{
"epoch": 0.396969696969697,
"grad_norm": 0.05702419951558113,
"learning_rate": 0.0008,
"loss": 2.220937490463257,
"step": 6550
},
{
"epoch": 0.3970909090909091,
"grad_norm": 0.04961099848151207,
"learning_rate": 0.0008,
"loss": 2.2272822856903076,
"step": 6552
},
{
"epoch": 0.39721212121212124,
"grad_norm": 0.06308319419622421,
"learning_rate": 0.0008,
"loss": 2.248635768890381,
"step": 6554
},
{
"epoch": 0.3973333333333333,
"grad_norm": 0.06727393716573715,
"learning_rate": 0.0008,
"loss": 2.16605544090271,
"step": 6556
},
{
"epoch": 0.39745454545454545,
"grad_norm": 0.05149586871266365,
"learning_rate": 0.0008,
"loss": 2.2403573989868164,
"step": 6558
},
{
"epoch": 0.3975757575757576,
"grad_norm": 0.04841329902410507,
"learning_rate": 0.0008,
"loss": 2.2777092456817627,
"step": 6560
},
{
"epoch": 0.3976969696969697,
"grad_norm": 0.05147114023566246,
"learning_rate": 0.0008,
"loss": 2.2010538578033447,
"step": 6562
},
{
"epoch": 0.3978181818181818,
"grad_norm": 0.051930565387010574,
"learning_rate": 0.0008,
"loss": 2.2024362087249756,
"step": 6564
},
{
"epoch": 0.3979393939393939,
"grad_norm": 0.05351851508021355,
"learning_rate": 0.0008,
"loss": 2.1489217281341553,
"step": 6566
},
{
"epoch": 0.39806060606060606,
"grad_norm": 0.04970606788992882,
"learning_rate": 0.0008,
"loss": 2.3312978744506836,
"step": 6568
},
{
"epoch": 0.3981818181818182,
"grad_norm": 0.05212164670228958,
"learning_rate": 0.0008,
"loss": 2.1623449325561523,
"step": 6570
},
{
"epoch": 0.3983030303030303,
"grad_norm": 0.0540316142141819,
"learning_rate": 0.0008,
"loss": 2.2452831268310547,
"step": 6572
},
{
"epoch": 0.3984242424242424,
"grad_norm": 0.04947185516357422,
"learning_rate": 0.0008,
"loss": 2.20965838432312,
"step": 6574
},
{
"epoch": 0.39854545454545454,
"grad_norm": 0.04798305034637451,
"learning_rate": 0.0008,
"loss": 2.184826612472534,
"step": 6576
},
{
"epoch": 0.39866666666666667,
"grad_norm": 0.049425236880779266,
"learning_rate": 0.0008,
"loss": 2.2700953483581543,
"step": 6578
},
{
"epoch": 0.3987878787878788,
"grad_norm": 0.04994813725352287,
"learning_rate": 0.0008,
"loss": 2.215829610824585,
"step": 6580
},
{
"epoch": 0.39890909090909094,
"grad_norm": 0.05152971297502518,
"learning_rate": 0.0008,
"loss": 2.2008848190307617,
"step": 6582
},
{
"epoch": 0.399030303030303,
"grad_norm": 0.05011424422264099,
"learning_rate": 0.0008,
"loss": 2.239699363708496,
"step": 6584
},
{
"epoch": 0.39915151515151515,
"grad_norm": 0.04904955253005028,
"learning_rate": 0.0008,
"loss": 2.180354118347168,
"step": 6586
},
{
"epoch": 0.3992727272727273,
"grad_norm": 0.048919204622507095,
"learning_rate": 0.0008,
"loss": 2.1776137351989746,
"step": 6588
},
{
"epoch": 0.3993939393939394,
"grad_norm": 0.04815306514501572,
"learning_rate": 0.0008,
"loss": 2.2180964946746826,
"step": 6590
},
{
"epoch": 0.3995151515151515,
"grad_norm": 0.051992014050483704,
"learning_rate": 0.0008,
"loss": 2.16782546043396,
"step": 6592
},
{
"epoch": 0.3996363636363636,
"grad_norm": 0.047688283026218414,
"learning_rate": 0.0008,
"loss": 2.171884059906006,
"step": 6594
},
{
"epoch": 0.39975757575757576,
"grad_norm": 0.04819457232952118,
"learning_rate": 0.0008,
"loss": 2.2159719467163086,
"step": 6596
},
{
"epoch": 0.3998787878787879,
"grad_norm": 0.05394309386610985,
"learning_rate": 0.0008,
"loss": 2.2110755443573,
"step": 6598
},
{
"epoch": 0.4,
"grad_norm": 0.04968157038092613,
"learning_rate": 0.0008,
"loss": 2.2448298931121826,
"step": 6600
},
{
"epoch": 0.4,
"eval_loss": 2.165179967880249,
"eval_runtime": 154.3744,
"eval_samples_per_second": 1.619,
"eval_steps_per_second": 0.084,
"step": 6600
},
{
"epoch": 0.4001212121212121,
"grad_norm": 0.05033470690250397,
"learning_rate": 0.0008,
"loss": 2.2561984062194824,
"step": 6602
},
{
"epoch": 0.40024242424242423,
"grad_norm": 0.04861438274383545,
"learning_rate": 0.0008,
"loss": 2.1583597660064697,
"step": 6604
},
{
"epoch": 0.40036363636363637,
"grad_norm": 0.05001963675022125,
"learning_rate": 0.0008,
"loss": 2.165323257446289,
"step": 6606
},
{
"epoch": 0.4004848484848485,
"grad_norm": 0.05441609397530556,
"learning_rate": 0.0008,
"loss": 2.2913262844085693,
"step": 6608
},
{
"epoch": 0.40060606060606063,
"grad_norm": 0.04825204238295555,
"learning_rate": 0.0008,
"loss": 2.2187788486480713,
"step": 6610
},
{
"epoch": 0.4007272727272727,
"grad_norm": 0.04992261901497841,
"learning_rate": 0.0008,
"loss": 2.161912441253662,
"step": 6612
},
{
"epoch": 0.40084848484848484,
"grad_norm": 0.055454473942518234,
"learning_rate": 0.0008,
"loss": 2.2088582515716553,
"step": 6614
},
{
"epoch": 0.400969696969697,
"grad_norm": 0.05924883484840393,
"learning_rate": 0.0008,
"loss": 2.235761880874634,
"step": 6616
},
{
"epoch": 0.4010909090909091,
"grad_norm": 0.07106713950634003,
"learning_rate": 0.0008,
"loss": 2.2250680923461914,
"step": 6618
},
{
"epoch": 0.4012121212121212,
"grad_norm": 0.05976840853691101,
"learning_rate": 0.0008,
"loss": 2.1292293071746826,
"step": 6620
},
{
"epoch": 0.4013333333333333,
"grad_norm": 0.05857114866375923,
"learning_rate": 0.0008,
"loss": 2.208348274230957,
"step": 6622
},
{
"epoch": 0.40145454545454545,
"grad_norm": 0.04921579360961914,
"learning_rate": 0.0008,
"loss": 2.1743576526641846,
"step": 6624
},
{
"epoch": 0.4015757575757576,
"grad_norm": 0.04746944084763527,
"learning_rate": 0.0008,
"loss": 2.2404894828796387,
"step": 6626
},
{
"epoch": 0.4016969696969697,
"grad_norm": 0.04956137761473656,
"learning_rate": 0.0008,
"loss": 2.1956605911254883,
"step": 6628
},
{
"epoch": 0.4018181818181818,
"grad_norm": 0.049748942255973816,
"learning_rate": 0.0008,
"loss": 2.2261972427368164,
"step": 6630
},
{
"epoch": 0.40193939393939393,
"grad_norm": 0.049373283982276917,
"learning_rate": 0.0008,
"loss": 2.2448768615722656,
"step": 6632
},
{
"epoch": 0.40206060606060606,
"grad_norm": 0.05029839277267456,
"learning_rate": 0.0008,
"loss": 2.1759326457977295,
"step": 6634
},
{
"epoch": 0.4021818181818182,
"grad_norm": 0.05401972681283951,
"learning_rate": 0.0008,
"loss": 2.1374213695526123,
"step": 6636
},
{
"epoch": 0.40230303030303033,
"grad_norm": 0.04874705895781517,
"learning_rate": 0.0008,
"loss": 2.181469202041626,
"step": 6638
},
{
"epoch": 0.4024242424242424,
"grad_norm": 0.04813924804329872,
"learning_rate": 0.0008,
"loss": 2.164128303527832,
"step": 6640
},
{
"epoch": 0.40254545454545454,
"grad_norm": 0.04859054461121559,
"learning_rate": 0.0008,
"loss": 2.196756601333618,
"step": 6642
},
{
"epoch": 0.4026666666666667,
"grad_norm": 0.051346175372600555,
"learning_rate": 0.0008,
"loss": 2.1205358505249023,
"step": 6644
},
{
"epoch": 0.4027878787878788,
"grad_norm": 0.04908078908920288,
"learning_rate": 0.0008,
"loss": 2.1769587993621826,
"step": 6646
},
{
"epoch": 0.4029090909090909,
"grad_norm": 0.050437506288290024,
"learning_rate": 0.0008,
"loss": 2.2283859252929688,
"step": 6648
},
{
"epoch": 0.403030303030303,
"grad_norm": 0.05037881061434746,
"learning_rate": 0.0008,
"loss": 2.218318462371826,
"step": 6650
},
{
"epoch": 0.40315151515151515,
"grad_norm": 0.04966982826590538,
"learning_rate": 0.0008,
"loss": 2.1954360008239746,
"step": 6652
},
{
"epoch": 0.4032727272727273,
"grad_norm": 0.04849361628293991,
"learning_rate": 0.0008,
"loss": 2.1769771575927734,
"step": 6654
},
{
"epoch": 0.4033939393939394,
"grad_norm": 0.049819134175777435,
"learning_rate": 0.0008,
"loss": 2.220332384109497,
"step": 6656
},
{
"epoch": 0.4035151515151515,
"grad_norm": 0.04931815341114998,
"learning_rate": 0.0008,
"loss": 2.1983160972595215,
"step": 6658
},
{
"epoch": 0.4036363636363636,
"grad_norm": 0.04879777878522873,
"learning_rate": 0.0008,
"loss": 2.1458897590637207,
"step": 6660
},
{
"epoch": 0.40375757575757576,
"grad_norm": 0.05018557608127594,
"learning_rate": 0.0008,
"loss": 2.1299965381622314,
"step": 6662
},
{
"epoch": 0.4038787878787879,
"grad_norm": 0.049003034830093384,
"learning_rate": 0.0008,
"loss": 2.218534231185913,
"step": 6664
},
{
"epoch": 0.404,
"grad_norm": 0.048276010900735855,
"learning_rate": 0.0008,
"loss": 2.1567959785461426,
"step": 6666
},
{
"epoch": 0.4041212121212121,
"grad_norm": 0.049996547400951385,
"learning_rate": 0.0008,
"loss": 2.1973862648010254,
"step": 6668
},
{
"epoch": 0.40424242424242424,
"grad_norm": 0.04950267821550369,
"learning_rate": 0.0008,
"loss": 2.1911001205444336,
"step": 6670
},
{
"epoch": 0.40436363636363637,
"grad_norm": 0.05044705048203468,
"learning_rate": 0.0008,
"loss": 2.165182113647461,
"step": 6672
},
{
"epoch": 0.4044848484848485,
"grad_norm": 0.05097615346312523,
"learning_rate": 0.0008,
"loss": 2.160419225692749,
"step": 6674
},
{
"epoch": 0.4046060606060606,
"grad_norm": 0.05123843252658844,
"learning_rate": 0.0008,
"loss": 2.152824878692627,
"step": 6676
},
{
"epoch": 0.4047272727272727,
"grad_norm": 0.050082627683877945,
"learning_rate": 0.0008,
"loss": 2.237433433532715,
"step": 6678
},
{
"epoch": 0.40484848484848485,
"grad_norm": 0.05176863074302673,
"learning_rate": 0.0008,
"loss": 2.189969539642334,
"step": 6680
},
{
"epoch": 0.404969696969697,
"grad_norm": 0.04890067130327225,
"learning_rate": 0.0008,
"loss": 2.207070827484131,
"step": 6682
},
{
"epoch": 0.4050909090909091,
"grad_norm": 0.049618322402238846,
"learning_rate": 0.0008,
"loss": 2.180570363998413,
"step": 6684
},
{
"epoch": 0.4052121212121212,
"grad_norm": 0.058986712247133255,
"learning_rate": 0.0008,
"loss": 2.182086944580078,
"step": 6686
},
{
"epoch": 0.4053333333333333,
"grad_norm": 0.05055321380496025,
"learning_rate": 0.0008,
"loss": 2.189790725708008,
"step": 6688
},
{
"epoch": 0.40545454545454546,
"grad_norm": 0.049196820706129074,
"learning_rate": 0.0008,
"loss": 2.1627442836761475,
"step": 6690
},
{
"epoch": 0.4055757575757576,
"grad_norm": 0.04924066737294197,
"learning_rate": 0.0008,
"loss": 2.1923508644104004,
"step": 6692
},
{
"epoch": 0.4056969696969697,
"grad_norm": 0.04895932972431183,
"learning_rate": 0.0008,
"loss": 2.157562732696533,
"step": 6694
},
{
"epoch": 0.4058181818181818,
"grad_norm": 0.049744199961423874,
"learning_rate": 0.0008,
"loss": 2.2231147289276123,
"step": 6696
},
{
"epoch": 0.40593939393939393,
"grad_norm": 0.046628668904304504,
"learning_rate": 0.0008,
"loss": 2.163891315460205,
"step": 6698
},
{
"epoch": 0.40606060606060607,
"grad_norm": 0.04959949851036072,
"learning_rate": 0.0008,
"loss": 2.1832470893859863,
"step": 6700
},
{
"epoch": 0.4061818181818182,
"grad_norm": 0.049962736666202545,
"learning_rate": 0.0008,
"loss": 2.1648552417755127,
"step": 6702
},
{
"epoch": 0.4063030303030303,
"grad_norm": 0.04877236485481262,
"learning_rate": 0.0008,
"loss": 2.1724445819854736,
"step": 6704
},
{
"epoch": 0.4064242424242424,
"grad_norm": 0.048129238188266754,
"learning_rate": 0.0008,
"loss": 2.1567747592926025,
"step": 6706
},
{
"epoch": 0.40654545454545454,
"grad_norm": 0.049137189984321594,
"learning_rate": 0.0008,
"loss": 2.1751821041107178,
"step": 6708
},
{
"epoch": 0.4066666666666667,
"grad_norm": 0.04982771351933479,
"learning_rate": 0.0008,
"loss": 2.1731081008911133,
"step": 6710
},
{
"epoch": 0.4067878787878788,
"grad_norm": 0.04888566583395004,
"learning_rate": 0.0008,
"loss": 2.2346363067626953,
"step": 6712
},
{
"epoch": 0.4069090909090909,
"grad_norm": 0.04934757202863693,
"learning_rate": 0.0008,
"loss": 2.2218456268310547,
"step": 6714
},
{
"epoch": 0.407030303030303,
"grad_norm": 0.04846339300274849,
"learning_rate": 0.0008,
"loss": 2.169480800628662,
"step": 6716
},
{
"epoch": 0.40715151515151515,
"grad_norm": 0.05147343873977661,
"learning_rate": 0.0008,
"loss": 2.1769092082977295,
"step": 6718
},
{
"epoch": 0.4072727272727273,
"grad_norm": 0.05345636233687401,
"learning_rate": 0.0008,
"loss": 2.223159074783325,
"step": 6720
},
{
"epoch": 0.4073939393939394,
"grad_norm": 0.05984744429588318,
"learning_rate": 0.0008,
"loss": 2.095165252685547,
"step": 6722
},
{
"epoch": 0.4075151515151515,
"grad_norm": 0.05674438551068306,
"learning_rate": 0.0008,
"loss": 2.194789171218872,
"step": 6724
},
{
"epoch": 0.40763636363636363,
"grad_norm": 0.054544657468795776,
"learning_rate": 0.0008,
"loss": 2.225322961807251,
"step": 6726
},
{
"epoch": 0.40775757575757576,
"grad_norm": 0.05077916383743286,
"learning_rate": 0.0008,
"loss": 2.1967031955718994,
"step": 6728
},
{
"epoch": 0.4078787878787879,
"grad_norm": 0.05233580991625786,
"learning_rate": 0.0008,
"loss": 2.236421585083008,
"step": 6730
},
{
"epoch": 0.408,
"grad_norm": 0.053125180304050446,
"learning_rate": 0.0008,
"loss": 2.0913562774658203,
"step": 6732
},
{
"epoch": 0.4081212121212121,
"grad_norm": 0.051827095448970795,
"learning_rate": 0.0008,
"loss": 2.1589245796203613,
"step": 6734
},
{
"epoch": 0.40824242424242424,
"grad_norm": 0.04943408817052841,
"learning_rate": 0.0008,
"loss": 2.1879734992980957,
"step": 6736
},
{
"epoch": 0.4083636363636364,
"grad_norm": 0.05261505767703056,
"learning_rate": 0.0008,
"loss": 2.1996777057647705,
"step": 6738
},
{
"epoch": 0.4084848484848485,
"grad_norm": 0.04883308708667755,
"learning_rate": 0.0008,
"loss": 2.166642665863037,
"step": 6740
},
{
"epoch": 0.4086060606060606,
"grad_norm": 0.05081753060221672,
"learning_rate": 0.0008,
"loss": 2.2195403575897217,
"step": 6742
},
{
"epoch": 0.4087272727272727,
"grad_norm": 0.05544598773121834,
"learning_rate": 0.0008,
"loss": 2.2020623683929443,
"step": 6744
},
{
"epoch": 0.40884848484848485,
"grad_norm": 0.057212382555007935,
"learning_rate": 0.0008,
"loss": 2.1479341983795166,
"step": 6746
},
{
"epoch": 0.408969696969697,
"grad_norm": 0.05178205296397209,
"learning_rate": 0.0008,
"loss": 2.1902902126312256,
"step": 6748
},
{
"epoch": 0.4090909090909091,
"grad_norm": 0.04850571230053902,
"learning_rate": 0.0008,
"loss": 2.196218252182007,
"step": 6750
},
{
"epoch": 0.4092121212121212,
"grad_norm": 0.05097746476531029,
"learning_rate": 0.0008,
"loss": 2.2823338508605957,
"step": 6752
},
{
"epoch": 0.4093333333333333,
"grad_norm": 0.05016610398888588,
"learning_rate": 0.0008,
"loss": 2.2527146339416504,
"step": 6754
},
{
"epoch": 0.40945454545454546,
"grad_norm": 0.05075736716389656,
"learning_rate": 0.0008,
"loss": 2.1527726650238037,
"step": 6756
},
{
"epoch": 0.4095757575757576,
"grad_norm": 0.04847331345081329,
"learning_rate": 0.0008,
"loss": 2.183933973312378,
"step": 6758
},
{
"epoch": 0.40969696969696967,
"grad_norm": 0.0507415272295475,
"learning_rate": 0.0008,
"loss": 2.2809643745422363,
"step": 6760
},
{
"epoch": 0.4098181818181818,
"grad_norm": 0.050324540585279465,
"learning_rate": 0.0008,
"loss": 2.1675057411193848,
"step": 6762
},
{
"epoch": 0.40993939393939394,
"grad_norm": 0.04886079207062721,
"learning_rate": 0.0008,
"loss": 2.2101340293884277,
"step": 6764
},
{
"epoch": 0.41006060606060607,
"grad_norm": 0.05086323246359825,
"learning_rate": 0.0008,
"loss": 2.1484920978546143,
"step": 6766
},
{
"epoch": 0.4101818181818182,
"grad_norm": 0.05093774572014809,
"learning_rate": 0.0008,
"loss": 2.272383689880371,
"step": 6768
},
{
"epoch": 0.4103030303030303,
"grad_norm": 0.049736905843019485,
"learning_rate": 0.0008,
"loss": 2.1516451835632324,
"step": 6770
},
{
"epoch": 0.4104242424242424,
"grad_norm": 0.049852725118398666,
"learning_rate": 0.0008,
"loss": 2.1497695446014404,
"step": 6772
},
{
"epoch": 0.41054545454545455,
"grad_norm": 0.04796461760997772,
"learning_rate": 0.0008,
"loss": 2.1974828243255615,
"step": 6774
},
{
"epoch": 0.4106666666666667,
"grad_norm": 0.05082154646515846,
"learning_rate": 0.0008,
"loss": 2.1764678955078125,
"step": 6776
},
{
"epoch": 0.4107878787878788,
"grad_norm": 0.0564785972237587,
"learning_rate": 0.0008,
"loss": 2.1687817573547363,
"step": 6778
},
{
"epoch": 0.4109090909090909,
"grad_norm": 0.05041922628879547,
"learning_rate": 0.0008,
"loss": 2.0973827838897705,
"step": 6780
},
{
"epoch": 0.411030303030303,
"grad_norm": 0.05083362013101578,
"learning_rate": 0.0008,
"loss": 2.1436822414398193,
"step": 6782
},
{
"epoch": 0.41115151515151516,
"grad_norm": 0.049126360565423965,
"learning_rate": 0.0008,
"loss": 2.1327812671661377,
"step": 6784
},
{
"epoch": 0.4112727272727273,
"grad_norm": 0.05113162100315094,
"learning_rate": 0.0008,
"loss": 2.1640820503234863,
"step": 6786
},
{
"epoch": 0.41139393939393937,
"grad_norm": 0.04845524951815605,
"learning_rate": 0.0008,
"loss": 2.1353087425231934,
"step": 6788
},
{
"epoch": 0.4115151515151515,
"grad_norm": 0.050916753709316254,
"learning_rate": 0.0008,
"loss": 2.186110258102417,
"step": 6790
},
{
"epoch": 0.41163636363636363,
"grad_norm": 0.04916217178106308,
"learning_rate": 0.0008,
"loss": 2.1318697929382324,
"step": 6792
},
{
"epoch": 0.41175757575757577,
"grad_norm": 0.05460149422287941,
"learning_rate": 0.0008,
"loss": 2.2133777141571045,
"step": 6794
},
{
"epoch": 0.4118787878787879,
"grad_norm": 0.052100226283073425,
"learning_rate": 0.0008,
"loss": 2.2555267810821533,
"step": 6796
},
{
"epoch": 0.412,
"grad_norm": 0.050578948110342026,
"learning_rate": 0.0008,
"loss": 2.2183170318603516,
"step": 6798
},
{
"epoch": 0.4121212121212121,
"grad_norm": 0.04876306280493736,
"learning_rate": 0.0008,
"loss": 2.152484893798828,
"step": 6800
},
{
"epoch": 0.41224242424242424,
"grad_norm": 0.04958086460828781,
"learning_rate": 0.0008,
"loss": 2.2432451248168945,
"step": 6802
},
{
"epoch": 0.4123636363636364,
"grad_norm": 0.04917731508612633,
"learning_rate": 0.0008,
"loss": 2.1452081203460693,
"step": 6804
},
{
"epoch": 0.4124848484848485,
"grad_norm": 0.04767732694745064,
"learning_rate": 0.0008,
"loss": 2.194037437438965,
"step": 6806
},
{
"epoch": 0.4126060606060606,
"grad_norm": 0.04971034452319145,
"learning_rate": 0.0008,
"loss": 2.178405523300171,
"step": 6808
},
{
"epoch": 0.4127272727272727,
"grad_norm": 0.04825672134757042,
"learning_rate": 0.0008,
"loss": 2.2015812397003174,
"step": 6810
},
{
"epoch": 0.41284848484848485,
"grad_norm": 0.04941675439476967,
"learning_rate": 0.0008,
"loss": 2.171182155609131,
"step": 6812
},
{
"epoch": 0.412969696969697,
"grad_norm": 0.049516983330249786,
"learning_rate": 0.0008,
"loss": 2.1636850833892822,
"step": 6814
},
{
"epoch": 0.41309090909090906,
"grad_norm": 0.051900118589401245,
"learning_rate": 0.0008,
"loss": 2.244971513748169,
"step": 6816
},
{
"epoch": 0.4132121212121212,
"grad_norm": 0.04778072237968445,
"learning_rate": 0.0008,
"loss": 2.1866538524627686,
"step": 6818
},
{
"epoch": 0.41333333333333333,
"grad_norm": 0.05586099624633789,
"learning_rate": 0.0008,
"loss": 2.158273220062256,
"step": 6820
},
{
"epoch": 0.41345454545454546,
"grad_norm": 0.05432557314634323,
"learning_rate": 0.0008,
"loss": 2.1621086597442627,
"step": 6822
},
{
"epoch": 0.4135757575757576,
"grad_norm": 0.05383209511637688,
"learning_rate": 0.0008,
"loss": 2.188413381576538,
"step": 6824
},
{
"epoch": 0.4136969696969697,
"grad_norm": 0.05685052275657654,
"learning_rate": 0.0008,
"loss": 2.176307439804077,
"step": 6826
},
{
"epoch": 0.4138181818181818,
"grad_norm": 0.0575178861618042,
"learning_rate": 0.0008,
"loss": 2.126272201538086,
"step": 6828
},
{
"epoch": 0.41393939393939394,
"grad_norm": 0.0533238984644413,
"learning_rate": 0.0008,
"loss": 2.2228353023529053,
"step": 6830
},
{
"epoch": 0.4140606060606061,
"grad_norm": 0.049409013241529465,
"learning_rate": 0.0008,
"loss": 2.2039854526519775,
"step": 6832
},
{
"epoch": 0.4141818181818182,
"grad_norm": 0.04796241223812103,
"learning_rate": 0.0008,
"loss": 2.11824893951416,
"step": 6834
},
{
"epoch": 0.4143030303030303,
"grad_norm": 0.04862634092569351,
"learning_rate": 0.0008,
"loss": 2.249976873397827,
"step": 6836
},
{
"epoch": 0.4144242424242424,
"grad_norm": 0.04881352186203003,
"learning_rate": 0.0008,
"loss": 2.1636173725128174,
"step": 6838
},
{
"epoch": 0.41454545454545455,
"grad_norm": 0.048817142844200134,
"learning_rate": 0.0008,
"loss": 2.1321334838867188,
"step": 6840
},
{
"epoch": 0.4146666666666667,
"grad_norm": 0.04873786121606827,
"learning_rate": 0.0008,
"loss": 2.245145797729492,
"step": 6842
},
{
"epoch": 0.41478787878787876,
"grad_norm": 0.05186041072010994,
"learning_rate": 0.0008,
"loss": 2.1745917797088623,
"step": 6844
},
{
"epoch": 0.4149090909090909,
"grad_norm": 0.05241953581571579,
"learning_rate": 0.0008,
"loss": 2.1453747749328613,
"step": 6846
},
{
"epoch": 0.415030303030303,
"grad_norm": 0.04888249561190605,
"learning_rate": 0.0008,
"loss": 2.1555044651031494,
"step": 6848
},
{
"epoch": 0.41515151515151516,
"grad_norm": 0.04902350902557373,
"learning_rate": 0.0008,
"loss": 2.152402639389038,
"step": 6850
},
{
"epoch": 0.4152727272727273,
"grad_norm": 0.04817958548665047,
"learning_rate": 0.0008,
"loss": 2.149080753326416,
"step": 6852
},
{
"epoch": 0.41539393939393937,
"grad_norm": 0.051156237721443176,
"learning_rate": 0.0008,
"loss": 2.1346163749694824,
"step": 6854
},
{
"epoch": 0.4155151515151515,
"grad_norm": 0.05757680535316467,
"learning_rate": 0.0008,
"loss": 2.1885669231414795,
"step": 6856
},
{
"epoch": 0.41563636363636364,
"grad_norm": 0.05539647862315178,
"learning_rate": 0.0008,
"loss": 2.2006585597991943,
"step": 6858
},
{
"epoch": 0.41575757575757577,
"grad_norm": 0.057499032467603683,
"learning_rate": 0.0008,
"loss": 2.195388078689575,
"step": 6860
},
{
"epoch": 0.4158787878787879,
"grad_norm": 0.0538918599486351,
"learning_rate": 0.0008,
"loss": 2.1944897174835205,
"step": 6862
},
{
"epoch": 0.416,
"grad_norm": 0.0534750297665596,
"learning_rate": 0.0008,
"loss": 2.1705079078674316,
"step": 6864
},
{
"epoch": 0.4161212121212121,
"grad_norm": 0.05027841776609421,
"learning_rate": 0.0008,
"loss": 2.1547489166259766,
"step": 6866
},
{
"epoch": 0.41624242424242425,
"grad_norm": 0.052188728004693985,
"learning_rate": 0.0008,
"loss": 2.1519293785095215,
"step": 6868
},
{
"epoch": 0.4163636363636364,
"grad_norm": 0.048092860728502274,
"learning_rate": 0.0008,
"loss": 2.154111385345459,
"step": 6870
},
{
"epoch": 0.41648484848484846,
"grad_norm": 0.054826218634843826,
"learning_rate": 0.0008,
"loss": 2.130009174346924,
"step": 6872
},
{
"epoch": 0.4166060606060606,
"grad_norm": 0.04611746966838837,
"learning_rate": 0.0008,
"loss": 2.1045422554016113,
"step": 6874
},
{
"epoch": 0.4167272727272727,
"grad_norm": 0.04876801744103432,
"learning_rate": 0.0008,
"loss": 2.136885404586792,
"step": 6876
},
{
"epoch": 0.41684848484848486,
"grad_norm": 0.04656074568629265,
"learning_rate": 0.0008,
"loss": 2.122108221054077,
"step": 6878
},
{
"epoch": 0.416969696969697,
"grad_norm": 0.0485246479511261,
"learning_rate": 0.0008,
"loss": 2.2068347930908203,
"step": 6880
},
{
"epoch": 0.41709090909090907,
"grad_norm": 0.04891160503029823,
"learning_rate": 0.0008,
"loss": 2.181889772415161,
"step": 6882
},
{
"epoch": 0.4172121212121212,
"grad_norm": 0.05183199793100357,
"learning_rate": 0.0008,
"loss": 2.1384220123291016,
"step": 6884
},
{
"epoch": 0.41733333333333333,
"grad_norm": 0.055727165192365646,
"learning_rate": 0.0008,
"loss": 2.135946750640869,
"step": 6886
},
{
"epoch": 0.41745454545454547,
"grad_norm": 0.07114315778017044,
"learning_rate": 0.0008,
"loss": 2.1944873332977295,
"step": 6888
},
{
"epoch": 0.4175757575757576,
"grad_norm": 0.05409567803144455,
"learning_rate": 0.0008,
"loss": 2.125129461288452,
"step": 6890
},
{
"epoch": 0.4176969696969697,
"grad_norm": 0.04847820848226547,
"learning_rate": 0.0008,
"loss": 2.1284914016723633,
"step": 6892
},
{
"epoch": 0.4178181818181818,
"grad_norm": 0.04931463673710823,
"learning_rate": 0.0008,
"loss": 2.1243062019348145,
"step": 6894
},
{
"epoch": 0.41793939393939394,
"grad_norm": 0.047714680433273315,
"learning_rate": 0.0008,
"loss": 2.1505115032196045,
"step": 6896
},
{
"epoch": 0.4180606060606061,
"grad_norm": 0.046377070248126984,
"learning_rate": 0.0008,
"loss": 2.124580144882202,
"step": 6898
},
{
"epoch": 0.41818181818181815,
"grad_norm": 0.05034399405121803,
"learning_rate": 0.0008,
"loss": 2.1956419944763184,
"step": 6900
},
{
"epoch": 0.4183030303030303,
"grad_norm": 0.04749593511223793,
"learning_rate": 0.0008,
"loss": 2.1574292182922363,
"step": 6902
},
{
"epoch": 0.4184242424242424,
"grad_norm": 0.04989727959036827,
"learning_rate": 0.0008,
"loss": 2.1445891857147217,
"step": 6904
},
{
"epoch": 0.41854545454545455,
"grad_norm": 0.04897862672805786,
"learning_rate": 0.0008,
"loss": 2.113037586212158,
"step": 6906
},
{
"epoch": 0.4186666666666667,
"grad_norm": 0.049909140914678574,
"learning_rate": 0.0008,
"loss": 2.1616947650909424,
"step": 6908
},
{
"epoch": 0.41878787878787876,
"grad_norm": 0.04857470095157623,
"learning_rate": 0.0008,
"loss": 2.1462595462799072,
"step": 6910
},
{
"epoch": 0.4189090909090909,
"grad_norm": 0.05178428068757057,
"learning_rate": 0.0008,
"loss": 2.1410181522369385,
"step": 6912
},
{
"epoch": 0.41903030303030303,
"grad_norm": 0.05295518785715103,
"learning_rate": 0.0008,
"loss": 2.1839396953582764,
"step": 6914
},
{
"epoch": 0.41915151515151516,
"grad_norm": 0.054640114307403564,
"learning_rate": 0.0008,
"loss": 2.126459836959839,
"step": 6916
},
{
"epoch": 0.4192727272727273,
"grad_norm": 0.05394330620765686,
"learning_rate": 0.0008,
"loss": 2.138474702835083,
"step": 6918
},
{
"epoch": 0.4193939393939394,
"grad_norm": 0.056985512375831604,
"learning_rate": 0.0008,
"loss": 2.2265923023223877,
"step": 6920
},
{
"epoch": 0.4195151515151515,
"grad_norm": 0.051369842141866684,
"learning_rate": 0.0008,
"loss": 2.1662065982818604,
"step": 6922
},
{
"epoch": 0.41963636363636364,
"grad_norm": 0.05260578170418739,
"learning_rate": 0.0008,
"loss": 2.1985034942626953,
"step": 6924
},
{
"epoch": 0.4197575757575758,
"grad_norm": 0.05403503403067589,
"learning_rate": 0.0008,
"loss": 2.216421604156494,
"step": 6926
},
{
"epoch": 0.41987878787878785,
"grad_norm": 0.05372398719191551,
"learning_rate": 0.0008,
"loss": 2.1867544651031494,
"step": 6928
},
{
"epoch": 0.42,
"grad_norm": 0.05662621557712555,
"learning_rate": 0.0008,
"loss": 2.1638970375061035,
"step": 6930
},
{
"epoch": 0.4201212121212121,
"grad_norm": 0.04819444566965103,
"learning_rate": 0.0008,
"loss": 2.1828129291534424,
"step": 6932
},
{
"epoch": 0.42024242424242425,
"grad_norm": 0.050052985548973083,
"learning_rate": 0.0008,
"loss": 2.150217056274414,
"step": 6934
},
{
"epoch": 0.4203636363636364,
"grad_norm": 0.048632796853780746,
"learning_rate": 0.0008,
"loss": 2.1623644828796387,
"step": 6936
},
{
"epoch": 0.42048484848484846,
"grad_norm": 0.051047272980213165,
"learning_rate": 0.0008,
"loss": 2.181708574295044,
"step": 6938
},
{
"epoch": 0.4206060606060606,
"grad_norm": 0.055711232125759125,
"learning_rate": 0.0008,
"loss": 2.130218505859375,
"step": 6940
},
{
"epoch": 0.4207272727272727,
"grad_norm": 0.06395269930362701,
"learning_rate": 0.0008,
"loss": 2.1548678874969482,
"step": 6942
},
{
"epoch": 0.42084848484848486,
"grad_norm": 0.06358631700277328,
"learning_rate": 0.0008,
"loss": 2.153724193572998,
"step": 6944
},
{
"epoch": 0.420969696969697,
"grad_norm": 0.06011927127838135,
"learning_rate": 0.0008,
"loss": 2.1884474754333496,
"step": 6946
},
{
"epoch": 0.42109090909090907,
"grad_norm": 0.050227049738168716,
"learning_rate": 0.0008,
"loss": 2.1193971633911133,
"step": 6948
},
{
"epoch": 0.4212121212121212,
"grad_norm": 0.04814944788813591,
"learning_rate": 0.0008,
"loss": 2.139760732650757,
"step": 6950
},
{
"epoch": 0.42133333333333334,
"grad_norm": 0.047623880207538605,
"learning_rate": 0.0008,
"loss": 2.1118526458740234,
"step": 6952
},
{
"epoch": 0.42145454545454547,
"grad_norm": 0.0508355014026165,
"learning_rate": 0.0008,
"loss": 2.1213388442993164,
"step": 6954
},
{
"epoch": 0.42157575757575755,
"grad_norm": 0.04744875058531761,
"learning_rate": 0.0008,
"loss": 2.148028612136841,
"step": 6956
},
{
"epoch": 0.4216969696969697,
"grad_norm": 0.048865195363759995,
"learning_rate": 0.0008,
"loss": 2.1649320125579834,
"step": 6958
},
{
"epoch": 0.4218181818181818,
"grad_norm": 0.04893174394965172,
"learning_rate": 0.0008,
"loss": 2.0986764430999756,
"step": 6960
},
{
"epoch": 0.42193939393939395,
"grad_norm": 0.04702275991439819,
"learning_rate": 0.0008,
"loss": 2.164762258529663,
"step": 6962
},
{
"epoch": 0.4220606060606061,
"grad_norm": 0.047821447253227234,
"learning_rate": 0.0008,
"loss": 2.165928602218628,
"step": 6964
},
{
"epoch": 0.42218181818181816,
"grad_norm": 0.04695947468280792,
"learning_rate": 0.0008,
"loss": 2.0964536666870117,
"step": 6966
},
{
"epoch": 0.4223030303030303,
"grad_norm": 0.04873279482126236,
"learning_rate": 0.0008,
"loss": 2.1499862670898438,
"step": 6968
},
{
"epoch": 0.4224242424242424,
"grad_norm": 0.04853171855211258,
"learning_rate": 0.0008,
"loss": 2.1500933170318604,
"step": 6970
},
{
"epoch": 0.42254545454545456,
"grad_norm": 0.04966359958052635,
"learning_rate": 0.0008,
"loss": 2.190601110458374,
"step": 6972
},
{
"epoch": 0.4226666666666667,
"grad_norm": 0.04920261725783348,
"learning_rate": 0.0008,
"loss": 2.179267406463623,
"step": 6974
},
{
"epoch": 0.42278787878787877,
"grad_norm": 0.04995420575141907,
"learning_rate": 0.0008,
"loss": 2.169517993927002,
"step": 6976
},
{
"epoch": 0.4229090909090909,
"grad_norm": 0.05531139671802521,
"learning_rate": 0.0008,
"loss": 2.165985107421875,
"step": 6978
},
{
"epoch": 0.42303030303030303,
"grad_norm": 0.05139917880296707,
"learning_rate": 0.0008,
"loss": 2.1444098949432373,
"step": 6980
},
{
"epoch": 0.42315151515151517,
"grad_norm": 0.05776160582900047,
"learning_rate": 0.0008,
"loss": 2.0911638736724854,
"step": 6982
},
{
"epoch": 0.4232727272727273,
"grad_norm": 0.05017286539077759,
"learning_rate": 0.0008,
"loss": 2.2101078033447266,
"step": 6984
},
{
"epoch": 0.4233939393939394,
"grad_norm": 0.05587420612573624,
"learning_rate": 0.0008,
"loss": 2.140604257583618,
"step": 6986
},
{
"epoch": 0.4235151515151515,
"grad_norm": 0.05709005892276764,
"learning_rate": 0.0008,
"loss": 2.0998501777648926,
"step": 6988
},
{
"epoch": 0.42363636363636364,
"grad_norm": 0.0545685812830925,
"learning_rate": 0.0008,
"loss": 2.126401901245117,
"step": 6990
},
{
"epoch": 0.4237575757575758,
"grad_norm": 0.05490932986140251,
"learning_rate": 0.0008,
"loss": 2.153693914413452,
"step": 6992
},
{
"epoch": 0.42387878787878785,
"grad_norm": 0.059728290885686874,
"learning_rate": 0.0008,
"loss": 2.117865562438965,
"step": 6994
},
{
"epoch": 0.424,
"grad_norm": 0.05074365809559822,
"learning_rate": 0.0008,
"loss": 2.1418540477752686,
"step": 6996
},
{
"epoch": 0.4241212121212121,
"grad_norm": 0.04911286011338234,
"learning_rate": 0.0008,
"loss": 2.232640027999878,
"step": 6998
},
{
"epoch": 0.42424242424242425,
"grad_norm": 0.046422429382801056,
"learning_rate": 0.0008,
"loss": 2.131110429763794,
"step": 7000
},
{
"epoch": 0.4243636363636364,
"grad_norm": 0.04848906397819519,
"learning_rate": 0.0008,
"loss": 2.1270244121551514,
"step": 7002
},
{
"epoch": 0.42448484848484846,
"grad_norm": 0.05340702086687088,
"learning_rate": 0.0008,
"loss": 2.174980878829956,
"step": 7004
},
{
"epoch": 0.4246060606060606,
"grad_norm": 0.04799322783946991,
"learning_rate": 0.0008,
"loss": 2.1420509815216064,
"step": 7006
},
{
"epoch": 0.42472727272727273,
"grad_norm": 0.04889779910445213,
"learning_rate": 0.0008,
"loss": 2.096865653991699,
"step": 7008
},
{
"epoch": 0.42484848484848486,
"grad_norm": 0.05287853628396988,
"learning_rate": 0.0008,
"loss": 2.1722629070281982,
"step": 7010
},
{
"epoch": 0.424969696969697,
"grad_norm": 0.05052591487765312,
"learning_rate": 0.0008,
"loss": 2.1228721141815186,
"step": 7012
},
{
"epoch": 0.4250909090909091,
"grad_norm": 0.049848657101392746,
"learning_rate": 0.0008,
"loss": 2.220724582672119,
"step": 7014
},
{
"epoch": 0.4252121212121212,
"grad_norm": 0.04826965555548668,
"learning_rate": 0.0008,
"loss": 2.198415756225586,
"step": 7016
},
{
"epoch": 0.42533333333333334,
"grad_norm": 0.052980516105890274,
"learning_rate": 0.0008,
"loss": 2.1200366020202637,
"step": 7018
},
{
"epoch": 0.4254545454545455,
"grad_norm": 0.05239897966384888,
"learning_rate": 0.0008,
"loss": 2.169330358505249,
"step": 7020
},
{
"epoch": 0.42557575757575755,
"grad_norm": 0.0505121685564518,
"learning_rate": 0.0008,
"loss": 2.1951663494110107,
"step": 7022
},
{
"epoch": 0.4256969696969697,
"grad_norm": 0.052635155618190765,
"learning_rate": 0.0008,
"loss": 2.1597795486450195,
"step": 7024
},
{
"epoch": 0.4258181818181818,
"grad_norm": 0.056303177028894424,
"learning_rate": 0.0008,
"loss": 2.168071985244751,
"step": 7026
},
{
"epoch": 0.42593939393939395,
"grad_norm": 0.05253642797470093,
"learning_rate": 0.0008,
"loss": 2.2003462314605713,
"step": 7028
},
{
"epoch": 0.4260606060606061,
"grad_norm": 0.05197736620903015,
"learning_rate": 0.0008,
"loss": 2.1266629695892334,
"step": 7030
},
{
"epoch": 0.42618181818181816,
"grad_norm": 0.047049157321453094,
"learning_rate": 0.0008,
"loss": 2.1252880096435547,
"step": 7032
},
{
"epoch": 0.4263030303030303,
"grad_norm": 0.05090818554162979,
"learning_rate": 0.0008,
"loss": 2.156839609146118,
"step": 7034
},
{
"epoch": 0.42642424242424243,
"grad_norm": 0.049216050654649734,
"learning_rate": 0.0008,
"loss": 2.0821549892425537,
"step": 7036
},
{
"epoch": 0.42654545454545456,
"grad_norm": 0.047388866543769836,
"learning_rate": 0.0008,
"loss": 2.233478307723999,
"step": 7038
},
{
"epoch": 0.4266666666666667,
"grad_norm": 0.04752982407808304,
"learning_rate": 0.0008,
"loss": 2.142542839050293,
"step": 7040
},
{
"epoch": 0.42678787878787877,
"grad_norm": 0.047738876193761826,
"learning_rate": 0.0008,
"loss": 2.0996921062469482,
"step": 7042
},
{
"epoch": 0.4269090909090909,
"grad_norm": 0.04929357022047043,
"learning_rate": 0.0008,
"loss": 2.1080644130706787,
"step": 7044
},
{
"epoch": 0.42703030303030304,
"grad_norm": 0.05023511126637459,
"learning_rate": 0.0008,
"loss": 2.1577610969543457,
"step": 7046
},
{
"epoch": 0.42715151515151517,
"grad_norm": 0.05407708138227463,
"learning_rate": 0.0008,
"loss": 2.1213061809539795,
"step": 7048
},
{
"epoch": 0.42727272727272725,
"grad_norm": 0.0488930381834507,
"learning_rate": 0.0008,
"loss": 2.1296122074127197,
"step": 7050
},
{
"epoch": 0.4273939393939394,
"grad_norm": 0.0503632128238678,
"learning_rate": 0.0008,
"loss": 2.1352412700653076,
"step": 7052
},
{
"epoch": 0.4275151515151515,
"grad_norm": 0.05190669372677803,
"learning_rate": 0.0008,
"loss": 2.1181130409240723,
"step": 7054
},
{
"epoch": 0.42763636363636365,
"grad_norm": 0.05347536504268646,
"learning_rate": 0.0008,
"loss": 2.138056993484497,
"step": 7056
},
{
"epoch": 0.4277575757575758,
"grad_norm": 0.04818965867161751,
"learning_rate": 0.0008,
"loss": 2.136913299560547,
"step": 7058
},
{
"epoch": 0.42787878787878786,
"grad_norm": 0.05448780953884125,
"learning_rate": 0.0008,
"loss": 2.179333209991455,
"step": 7060
},
{
"epoch": 0.428,
"grad_norm": 0.04867913946509361,
"learning_rate": 0.0008,
"loss": 2.116222381591797,
"step": 7062
},
{
"epoch": 0.4281212121212121,
"grad_norm": 0.04998209699988365,
"learning_rate": 0.0008,
"loss": 2.167692184448242,
"step": 7064
},
{
"epoch": 0.42824242424242426,
"grad_norm": 0.0475957989692688,
"learning_rate": 0.0008,
"loss": 2.1213982105255127,
"step": 7066
},
{
"epoch": 0.4283636363636364,
"grad_norm": 0.04908333718776703,
"learning_rate": 0.0008,
"loss": 2.1772470474243164,
"step": 7068
},
{
"epoch": 0.42848484848484847,
"grad_norm": 0.054474957287311554,
"learning_rate": 0.0008,
"loss": 2.193469524383545,
"step": 7070
},
{
"epoch": 0.4286060606060606,
"grad_norm": 0.0504494346678257,
"learning_rate": 0.0008,
"loss": 2.1101324558258057,
"step": 7072
},
{
"epoch": 0.42872727272727273,
"grad_norm": 0.04683234170079231,
"learning_rate": 0.0008,
"loss": 2.120560646057129,
"step": 7074
},
{
"epoch": 0.42884848484848487,
"grad_norm": 0.04861340671777725,
"learning_rate": 0.0008,
"loss": 2.0612895488739014,
"step": 7076
},
{
"epoch": 0.42896969696969695,
"grad_norm": 0.04787250980734825,
"learning_rate": 0.0008,
"loss": 2.1513333320617676,
"step": 7078
},
{
"epoch": 0.4290909090909091,
"grad_norm": 0.05461634695529938,
"learning_rate": 0.0008,
"loss": 2.127601146697998,
"step": 7080
},
{
"epoch": 0.4292121212121212,
"grad_norm": 0.04775075241923332,
"learning_rate": 0.0008,
"loss": 2.1121859550476074,
"step": 7082
},
{
"epoch": 0.42933333333333334,
"grad_norm": 0.04932071268558502,
"learning_rate": 0.0008,
"loss": 2.169273853302002,
"step": 7084
},
{
"epoch": 0.4294545454545455,
"grad_norm": 0.04829373583197594,
"learning_rate": 0.0008,
"loss": 2.117488384246826,
"step": 7086
},
{
"epoch": 0.42957575757575756,
"grad_norm": 0.049531903117895126,
"learning_rate": 0.0008,
"loss": 2.1261284351348877,
"step": 7088
},
{
"epoch": 0.4296969696969697,
"grad_norm": 0.04776924103498459,
"learning_rate": 0.0008,
"loss": 2.170743465423584,
"step": 7090
},
{
"epoch": 0.4298181818181818,
"grad_norm": 0.05091337114572525,
"learning_rate": 0.0008,
"loss": 2.1898131370544434,
"step": 7092
},
{
"epoch": 0.42993939393939395,
"grad_norm": 0.04854974150657654,
"learning_rate": 0.0008,
"loss": 2.192401885986328,
"step": 7094
},
{
"epoch": 0.4300606060606061,
"grad_norm": 0.04941518232226372,
"learning_rate": 0.0008,
"loss": 2.1383419036865234,
"step": 7096
},
{
"epoch": 0.43018181818181817,
"grad_norm": 0.047440022230148315,
"learning_rate": 0.0008,
"loss": 2.185673236846924,
"step": 7098
},
{
"epoch": 0.4303030303030303,
"grad_norm": 0.04944285377860069,
"learning_rate": 0.0008,
"loss": 2.249866247177124,
"step": 7100
},
{
"epoch": 0.43042424242424243,
"grad_norm": 0.04907004162669182,
"learning_rate": 0.0008,
"loss": 2.129054307937622,
"step": 7102
},
{
"epoch": 0.43054545454545456,
"grad_norm": 0.04919368773698807,
"learning_rate": 0.0008,
"loss": 2.136550188064575,
"step": 7104
},
{
"epoch": 0.43066666666666664,
"grad_norm": 0.05332620069384575,
"learning_rate": 0.0008,
"loss": 2.111649990081787,
"step": 7106
},
{
"epoch": 0.4307878787878788,
"grad_norm": 0.057944405823946,
"learning_rate": 0.0008,
"loss": 2.148608684539795,
"step": 7108
},
{
"epoch": 0.4309090909090909,
"grad_norm": 0.06441311538219452,
"learning_rate": 0.0008,
"loss": 2.1609952449798584,
"step": 7110
},
{
"epoch": 0.43103030303030304,
"grad_norm": 0.05305487662553787,
"learning_rate": 0.0008,
"loss": 2.188563346862793,
"step": 7112
},
{
"epoch": 0.4311515151515152,
"grad_norm": 0.04996899142861366,
"learning_rate": 0.0008,
"loss": 2.1432573795318604,
"step": 7114
},
{
"epoch": 0.43127272727272725,
"grad_norm": 0.04828493669629097,
"learning_rate": 0.0008,
"loss": 2.128063678741455,
"step": 7116
},
{
"epoch": 0.4313939393939394,
"grad_norm": 0.04738388955593109,
"learning_rate": 0.0008,
"loss": 2.1120970249176025,
"step": 7118
},
{
"epoch": 0.4315151515151515,
"grad_norm": 0.04912789538502693,
"learning_rate": 0.0008,
"loss": 2.097579002380371,
"step": 7120
},
{
"epoch": 0.43163636363636365,
"grad_norm": 0.04773698002099991,
"learning_rate": 0.0008,
"loss": 2.0663232803344727,
"step": 7122
},
{
"epoch": 0.4317575757575758,
"grad_norm": 0.04765781760215759,
"learning_rate": 0.0008,
"loss": 2.1123692989349365,
"step": 7124
},
{
"epoch": 0.43187878787878786,
"grad_norm": 0.04934055730700493,
"learning_rate": 0.0008,
"loss": 2.129359722137451,
"step": 7126
},
{
"epoch": 0.432,
"grad_norm": 0.05003395676612854,
"learning_rate": 0.0008,
"loss": 2.1457393169403076,
"step": 7128
},
{
"epoch": 0.43212121212121213,
"grad_norm": 0.04842570796608925,
"learning_rate": 0.0008,
"loss": 2.1448631286621094,
"step": 7130
},
{
"epoch": 0.43224242424242426,
"grad_norm": 0.04860939085483551,
"learning_rate": 0.0008,
"loss": 2.052734136581421,
"step": 7132
},
{
"epoch": 0.43236363636363634,
"grad_norm": 0.04977916181087494,
"learning_rate": 0.0008,
"loss": 2.14056134223938,
"step": 7134
},
{
"epoch": 0.43248484848484847,
"grad_norm": 0.05023197457194328,
"learning_rate": 0.0008,
"loss": 2.125458240509033,
"step": 7136
},
{
"epoch": 0.4326060606060606,
"grad_norm": 0.0488293394446373,
"learning_rate": 0.0008,
"loss": 2.2001967430114746,
"step": 7138
},
{
"epoch": 0.43272727272727274,
"grad_norm": 0.05970047041773796,
"learning_rate": 0.0008,
"loss": 2.084284543991089,
"step": 7140
},
{
"epoch": 0.43284848484848487,
"grad_norm": 0.05421001836657524,
"learning_rate": 0.0008,
"loss": 2.152266502380371,
"step": 7142
},
{
"epoch": 0.43296969696969695,
"grad_norm": 0.04911796748638153,
"learning_rate": 0.0008,
"loss": 2.1324784755706787,
"step": 7144
},
{
"epoch": 0.4330909090909091,
"grad_norm": 0.050306666642427444,
"learning_rate": 0.0008,
"loss": 2.1728267669677734,
"step": 7146
},
{
"epoch": 0.4332121212121212,
"grad_norm": 0.05297704041004181,
"learning_rate": 0.0008,
"loss": 2.163111925125122,
"step": 7148
},
{
"epoch": 0.43333333333333335,
"grad_norm": 0.07730308920145035,
"learning_rate": 0.0008,
"loss": 2.122028350830078,
"step": 7150
},
{
"epoch": 0.4334545454545455,
"grad_norm": 0.07211358845233917,
"learning_rate": 0.0008,
"loss": 2.1187095642089844,
"step": 7152
},
{
"epoch": 0.43357575757575756,
"grad_norm": 0.06381864845752716,
"learning_rate": 0.0008,
"loss": 2.09898042678833,
"step": 7154
},
{
"epoch": 0.4336969696969697,
"grad_norm": 0.06412708014249802,
"learning_rate": 0.0008,
"loss": 2.1431519985198975,
"step": 7156
},
{
"epoch": 0.4338181818181818,
"grad_norm": 0.055843286216259,
"learning_rate": 0.0008,
"loss": 2.1282434463500977,
"step": 7158
},
{
"epoch": 0.43393939393939396,
"grad_norm": 0.057258229702711105,
"learning_rate": 0.0008,
"loss": 2.151549816131592,
"step": 7160
},
{
"epoch": 0.43406060606060604,
"grad_norm": 0.05388399586081505,
"learning_rate": 0.0008,
"loss": 2.1576504707336426,
"step": 7162
},
{
"epoch": 0.43418181818181817,
"grad_norm": 0.04919697716832161,
"learning_rate": 0.0008,
"loss": 2.0766940116882324,
"step": 7164
},
{
"epoch": 0.4343030303030303,
"grad_norm": 0.05069831758737564,
"learning_rate": 0.0008,
"loss": 2.158115863800049,
"step": 7166
},
{
"epoch": 0.43442424242424243,
"grad_norm": 0.055250540375709534,
"learning_rate": 0.0008,
"loss": 2.1066038608551025,
"step": 7168
},
{
"epoch": 0.43454545454545457,
"grad_norm": 0.04930012673139572,
"learning_rate": 0.0008,
"loss": 2.114520311355591,
"step": 7170
},
{
"epoch": 0.43466666666666665,
"grad_norm": 0.04619089141488075,
"learning_rate": 0.0008,
"loss": 2.1834781169891357,
"step": 7172
},
{
"epoch": 0.4347878787878788,
"grad_norm": 0.047460559755563736,
"learning_rate": 0.0008,
"loss": 2.129467010498047,
"step": 7174
},
{
"epoch": 0.4349090909090909,
"grad_norm": 0.04738704487681389,
"learning_rate": 0.0008,
"loss": 2.0887503623962402,
"step": 7176
},
{
"epoch": 0.43503030303030304,
"grad_norm": 0.05242237448692322,
"learning_rate": 0.0008,
"loss": 2.112459421157837,
"step": 7178
},
{
"epoch": 0.4351515151515152,
"grad_norm": 0.055275846272706985,
"learning_rate": 0.0008,
"loss": 2.155482769012451,
"step": 7180
},
{
"epoch": 0.43527272727272726,
"grad_norm": 0.05692269280552864,
"learning_rate": 0.0008,
"loss": 2.1761960983276367,
"step": 7182
},
{
"epoch": 0.4353939393939394,
"grad_norm": 0.05420577526092529,
"learning_rate": 0.0008,
"loss": 2.1336662769317627,
"step": 7184
},
{
"epoch": 0.4355151515151515,
"grad_norm": 0.04992717131972313,
"learning_rate": 0.0008,
"loss": 2.1362040042877197,
"step": 7186
},
{
"epoch": 0.43563636363636365,
"grad_norm": 0.049077462404966354,
"learning_rate": 0.0008,
"loss": 2.1310200691223145,
"step": 7188
},
{
"epoch": 0.43575757575757573,
"grad_norm": 0.0501585491001606,
"learning_rate": 0.0008,
"loss": 2.1516475677490234,
"step": 7190
},
{
"epoch": 0.43587878787878787,
"grad_norm": 0.04790109395980835,
"learning_rate": 0.0008,
"loss": 2.114185094833374,
"step": 7192
},
{
"epoch": 0.436,
"grad_norm": 0.05266379937529564,
"learning_rate": 0.0008,
"loss": 2.119473934173584,
"step": 7194
},
{
"epoch": 0.43612121212121213,
"grad_norm": 0.050536975264549255,
"learning_rate": 0.0008,
"loss": 2.08667254447937,
"step": 7196
},
{
"epoch": 0.43624242424242426,
"grad_norm": 0.05683806166052818,
"learning_rate": 0.0008,
"loss": 2.139087677001953,
"step": 7198
},
{
"epoch": 0.43636363636363634,
"grad_norm": 0.05037148296833038,
"learning_rate": 0.0008,
"loss": 2.1466915607452393,
"step": 7200
},
{
"epoch": 0.4364848484848485,
"grad_norm": 0.04866168648004532,
"learning_rate": 0.0008,
"loss": 2.132878541946411,
"step": 7202
},
{
"epoch": 0.4366060606060606,
"grad_norm": 0.046966541558504105,
"learning_rate": 0.0008,
"loss": 2.1299853324890137,
"step": 7204
},
{
"epoch": 0.43672727272727274,
"grad_norm": 0.052686192095279694,
"learning_rate": 0.0008,
"loss": 2.1604721546173096,
"step": 7206
},
{
"epoch": 0.4368484848484849,
"grad_norm": 0.047334395349025726,
"learning_rate": 0.0008,
"loss": 2.1120526790618896,
"step": 7208
},
{
"epoch": 0.43696969696969695,
"grad_norm": 0.047630466520786285,
"learning_rate": 0.0008,
"loss": 2.057694911956787,
"step": 7210
},
{
"epoch": 0.4370909090909091,
"grad_norm": 0.05002722516655922,
"learning_rate": 0.0008,
"loss": 2.198768138885498,
"step": 7212
},
{
"epoch": 0.4372121212121212,
"grad_norm": 0.04819141700863838,
"learning_rate": 0.0008,
"loss": 2.114989757537842,
"step": 7214
},
{
"epoch": 0.43733333333333335,
"grad_norm": 0.05294293910264969,
"learning_rate": 0.0008,
"loss": 2.1160006523132324,
"step": 7216
},
{
"epoch": 0.43745454545454543,
"grad_norm": 0.05041459575295448,
"learning_rate": 0.0008,
"loss": 2.0874216556549072,
"step": 7218
},
{
"epoch": 0.43757575757575756,
"grad_norm": 0.05009935796260834,
"learning_rate": 0.0008,
"loss": 2.102025032043457,
"step": 7220
},
{
"epoch": 0.4376969696969697,
"grad_norm": 0.05356008931994438,
"learning_rate": 0.0008,
"loss": 2.142519474029541,
"step": 7222
},
{
"epoch": 0.43781818181818183,
"grad_norm": 0.052798546850681305,
"learning_rate": 0.0008,
"loss": 2.138226270675659,
"step": 7224
},
{
"epoch": 0.43793939393939396,
"grad_norm": 0.050941117107868195,
"learning_rate": 0.0008,
"loss": 2.112008571624756,
"step": 7226
},
{
"epoch": 0.43806060606060604,
"grad_norm": 0.04814717918634415,
"learning_rate": 0.0008,
"loss": 2.0734941959381104,
"step": 7228
},
{
"epoch": 0.4381818181818182,
"grad_norm": 0.04676767438650131,
"learning_rate": 0.0008,
"loss": 2.0693817138671875,
"step": 7230
},
{
"epoch": 0.4383030303030303,
"grad_norm": 0.0498138964176178,
"learning_rate": 0.0008,
"loss": 2.090667724609375,
"step": 7232
},
{
"epoch": 0.43842424242424244,
"grad_norm": 0.04806691035628319,
"learning_rate": 0.0008,
"loss": 2.0654807090759277,
"step": 7234
},
{
"epoch": 0.43854545454545457,
"grad_norm": 0.04899058863520622,
"learning_rate": 0.0008,
"loss": 2.150169849395752,
"step": 7236
},
{
"epoch": 0.43866666666666665,
"grad_norm": 0.04754219576716423,
"learning_rate": 0.0008,
"loss": 2.0982861518859863,
"step": 7238
},
{
"epoch": 0.4387878787878788,
"grad_norm": 0.049298159778118134,
"learning_rate": 0.0008,
"loss": 2.106947898864746,
"step": 7240
},
{
"epoch": 0.4389090909090909,
"grad_norm": 0.04817667230963707,
"learning_rate": 0.0008,
"loss": 2.100059986114502,
"step": 7242
},
{
"epoch": 0.43903030303030305,
"grad_norm": 0.04893858730792999,
"learning_rate": 0.0008,
"loss": 2.1018989086151123,
"step": 7244
},
{
"epoch": 0.4391515151515151,
"grad_norm": 0.048234932124614716,
"learning_rate": 0.0008,
"loss": 2.0961875915527344,
"step": 7246
},
{
"epoch": 0.43927272727272726,
"grad_norm": 0.0507332943379879,
"learning_rate": 0.0008,
"loss": 2.129467725753784,
"step": 7248
},
{
"epoch": 0.4393939393939394,
"grad_norm": 0.0578002855181694,
"learning_rate": 0.0008,
"loss": 2.035194158554077,
"step": 7250
},
{
"epoch": 0.4395151515151515,
"grad_norm": 0.046848099678754807,
"learning_rate": 0.0008,
"loss": 2.118279218673706,
"step": 7252
},
{
"epoch": 0.43963636363636366,
"grad_norm": 0.048638470470905304,
"learning_rate": 0.0008,
"loss": 2.149688959121704,
"step": 7254
},
{
"epoch": 0.43975757575757574,
"grad_norm": 0.048163462430238724,
"learning_rate": 0.0008,
"loss": 2.076887845993042,
"step": 7256
},
{
"epoch": 0.43987878787878787,
"grad_norm": 0.045823946595191956,
"learning_rate": 0.0008,
"loss": 2.1308224201202393,
"step": 7258
},
{
"epoch": 0.44,
"grad_norm": 0.05151960998773575,
"learning_rate": 0.0008,
"loss": 2.1327762603759766,
"step": 7260
},
{
"epoch": 0.44012121212121214,
"grad_norm": 0.04812181740999222,
"learning_rate": 0.0008,
"loss": 2.0947959423065186,
"step": 7262
},
{
"epoch": 0.44024242424242427,
"grad_norm": 0.04835138097405434,
"learning_rate": 0.0008,
"loss": 2.0963985919952393,
"step": 7264
},
{
"epoch": 0.44036363636363635,
"grad_norm": 0.05196968466043472,
"learning_rate": 0.0008,
"loss": 2.202014684677124,
"step": 7266
},
{
"epoch": 0.4404848484848485,
"grad_norm": 0.047693293541669846,
"learning_rate": 0.0008,
"loss": 2.066214084625244,
"step": 7268
},
{
"epoch": 0.4406060606060606,
"grad_norm": 0.05092163756489754,
"learning_rate": 0.0008,
"loss": 2.182917594909668,
"step": 7270
},
{
"epoch": 0.44072727272727275,
"grad_norm": 0.047995954751968384,
"learning_rate": 0.0008,
"loss": 2.0584218502044678,
"step": 7272
},
{
"epoch": 0.4408484848484848,
"grad_norm": 0.04904264211654663,
"learning_rate": 0.0008,
"loss": 2.135213851928711,
"step": 7274
},
{
"epoch": 0.44096969696969696,
"grad_norm": 0.046371594071388245,
"learning_rate": 0.0008,
"loss": 2.087268829345703,
"step": 7276
},
{
"epoch": 0.4410909090909091,
"grad_norm": 0.05019066855311394,
"learning_rate": 0.0008,
"loss": 2.166855573654175,
"step": 7278
},
{
"epoch": 0.4412121212121212,
"grad_norm": 0.05015747621655464,
"learning_rate": 0.0008,
"loss": 2.0607025623321533,
"step": 7280
},
{
"epoch": 0.44133333333333336,
"grad_norm": 0.052481506019830704,
"learning_rate": 0.0008,
"loss": 2.0828633308410645,
"step": 7282
},
{
"epoch": 0.44145454545454543,
"grad_norm": 0.050084419548511505,
"learning_rate": 0.0008,
"loss": 2.106311559677124,
"step": 7284
},
{
"epoch": 0.44157575757575757,
"grad_norm": 0.04697716608643532,
"learning_rate": 0.0008,
"loss": 2.101006031036377,
"step": 7286
},
{
"epoch": 0.4416969696969697,
"grad_norm": 0.04826412722468376,
"learning_rate": 0.0008,
"loss": 2.086379289627075,
"step": 7288
},
{
"epoch": 0.44181818181818183,
"grad_norm": 0.047705989331007004,
"learning_rate": 0.0008,
"loss": 2.1583006381988525,
"step": 7290
},
{
"epoch": 0.44193939393939397,
"grad_norm": 0.04886586591601372,
"learning_rate": 0.0008,
"loss": 2.0743906497955322,
"step": 7292
},
{
"epoch": 0.44206060606060604,
"grad_norm": 0.05008908361196518,
"learning_rate": 0.0008,
"loss": 2.164811849594116,
"step": 7294
},
{
"epoch": 0.4421818181818182,
"grad_norm": 0.04708297923207283,
"learning_rate": 0.0008,
"loss": 2.1539385318756104,
"step": 7296
},
{
"epoch": 0.4423030303030303,
"grad_norm": 0.056358661502599716,
"learning_rate": 0.0008,
"loss": 2.150622606277466,
"step": 7298
},
{
"epoch": 0.44242424242424244,
"grad_norm": 0.048985980451107025,
"learning_rate": 0.0008,
"loss": 2.1009562015533447,
"step": 7300
},
{
"epoch": 0.4425454545454545,
"grad_norm": 0.04857241362333298,
"learning_rate": 0.0008,
"loss": 2.0838119983673096,
"step": 7302
},
{
"epoch": 0.44266666666666665,
"grad_norm": 0.04753390699625015,
"learning_rate": 0.0008,
"loss": 2.0967915058135986,
"step": 7304
},
{
"epoch": 0.4427878787878788,
"grad_norm": 0.05063021555542946,
"learning_rate": 0.0008,
"loss": 2.105006217956543,
"step": 7306
},
{
"epoch": 0.4429090909090909,
"grad_norm": 0.05946565791964531,
"learning_rate": 0.0008,
"loss": 2.089315176010132,
"step": 7308
},
{
"epoch": 0.44303030303030305,
"grad_norm": 0.050313275307416916,
"learning_rate": 0.0008,
"loss": 2.106616973876953,
"step": 7310
},
{
"epoch": 0.44315151515151513,
"grad_norm": 0.050277236849069595,
"learning_rate": 0.0008,
"loss": 2.120623826980591,
"step": 7312
},
{
"epoch": 0.44327272727272726,
"grad_norm": 0.05090756341814995,
"learning_rate": 0.0008,
"loss": 2.2010304927825928,
"step": 7314
},
{
"epoch": 0.4433939393939394,
"grad_norm": 0.04745107889175415,
"learning_rate": 0.0008,
"loss": 2.027700662612915,
"step": 7316
},
{
"epoch": 0.44351515151515153,
"grad_norm": 0.050340376794338226,
"learning_rate": 0.0008,
"loss": 2.0686168670654297,
"step": 7318
},
{
"epoch": 0.44363636363636366,
"grad_norm": 0.049739908427000046,
"learning_rate": 0.0008,
"loss": 2.1550707817077637,
"step": 7320
},
{
"epoch": 0.44375757575757574,
"grad_norm": 0.05017562583088875,
"learning_rate": 0.0008,
"loss": 2.0938611030578613,
"step": 7322
},
{
"epoch": 0.4438787878787879,
"grad_norm": 0.048576731234788895,
"learning_rate": 0.0008,
"loss": 2.099262237548828,
"step": 7324
},
{
"epoch": 0.444,
"grad_norm": 0.048441726714372635,
"learning_rate": 0.0008,
"loss": 2.130103826522827,
"step": 7326
},
{
"epoch": 0.44412121212121214,
"grad_norm": 0.047003425657749176,
"learning_rate": 0.0008,
"loss": 2.141730546951294,
"step": 7328
},
{
"epoch": 0.4442424242424242,
"grad_norm": 0.04881230741739273,
"learning_rate": 0.0008,
"loss": 2.111814498901367,
"step": 7330
},
{
"epoch": 0.44436363636363635,
"grad_norm": 0.046707432717084885,
"learning_rate": 0.0008,
"loss": 2.115523338317871,
"step": 7332
},
{
"epoch": 0.4444848484848485,
"grad_norm": 0.04799891635775566,
"learning_rate": 0.0008,
"loss": 2.119828701019287,
"step": 7334
},
{
"epoch": 0.4446060606060606,
"grad_norm": 0.04941687732934952,
"learning_rate": 0.0008,
"loss": 2.1184165477752686,
"step": 7336
},
{
"epoch": 0.44472727272727275,
"grad_norm": 0.053010016679763794,
"learning_rate": 0.0008,
"loss": 2.173741102218628,
"step": 7338
},
{
"epoch": 0.4448484848484848,
"grad_norm": 0.047652434557676315,
"learning_rate": 0.0008,
"loss": 2.1075165271759033,
"step": 7340
},
{
"epoch": 0.44496969696969696,
"grad_norm": 0.050543755292892456,
"learning_rate": 0.0008,
"loss": 2.102993965148926,
"step": 7342
},
{
"epoch": 0.4450909090909091,
"grad_norm": 0.05157260224223137,
"learning_rate": 0.0008,
"loss": 2.0628254413604736,
"step": 7344
},
{
"epoch": 0.4452121212121212,
"grad_norm": 0.05085534602403641,
"learning_rate": 0.0008,
"loss": 2.118671417236328,
"step": 7346
},
{
"epoch": 0.44533333333333336,
"grad_norm": 0.05177607014775276,
"learning_rate": 0.0008,
"loss": 2.1235885620117188,
"step": 7348
},
{
"epoch": 0.44545454545454544,
"grad_norm": 0.04830687865614891,
"learning_rate": 0.0008,
"loss": 2.0651087760925293,
"step": 7350
},
{
"epoch": 0.44557575757575757,
"grad_norm": 0.05020773410797119,
"learning_rate": 0.0008,
"loss": 2.0680787563323975,
"step": 7352
},
{
"epoch": 0.4456969696969697,
"grad_norm": 0.047404639422893524,
"learning_rate": 0.0008,
"loss": 2.0782065391540527,
"step": 7354
},
{
"epoch": 0.44581818181818184,
"grad_norm": 0.048843830823898315,
"learning_rate": 0.0008,
"loss": 2.140713930130005,
"step": 7356
},
{
"epoch": 0.4459393939393939,
"grad_norm": 0.048814088106155396,
"learning_rate": 0.0008,
"loss": 2.07389497756958,
"step": 7358
},
{
"epoch": 0.44606060606060605,
"grad_norm": 0.048142097890377045,
"learning_rate": 0.0008,
"loss": 2.07733154296875,
"step": 7360
},
{
"epoch": 0.4461818181818182,
"grad_norm": 0.048051510006189346,
"learning_rate": 0.0008,
"loss": 2.10280704498291,
"step": 7362
},
{
"epoch": 0.4463030303030303,
"grad_norm": 0.04953155294060707,
"learning_rate": 0.0008,
"loss": 2.10071063041687,
"step": 7364
},
{
"epoch": 0.44642424242424245,
"grad_norm": 0.04858984425663948,
"learning_rate": 0.0008,
"loss": 2.0747013092041016,
"step": 7366
},
{
"epoch": 0.4465454545454545,
"grad_norm": 0.04883934184908867,
"learning_rate": 0.0008,
"loss": 2.172410726547241,
"step": 7368
},
{
"epoch": 0.44666666666666666,
"grad_norm": 0.05000294744968414,
"learning_rate": 0.0008,
"loss": 2.0984535217285156,
"step": 7370
},
{
"epoch": 0.4467878787878788,
"grad_norm": 0.05588582530617714,
"learning_rate": 0.0008,
"loss": 2.1355481147766113,
"step": 7372
},
{
"epoch": 0.4469090909090909,
"grad_norm": 0.053124766796827316,
"learning_rate": 0.0008,
"loss": 2.1605565547943115,
"step": 7374
},
{
"epoch": 0.44703030303030306,
"grad_norm": 0.04898274317383766,
"learning_rate": 0.0008,
"loss": 2.134938955307007,
"step": 7376
},
{
"epoch": 0.44715151515151513,
"grad_norm": 0.04843109846115112,
"learning_rate": 0.0008,
"loss": 2.1441221237182617,
"step": 7378
},
{
"epoch": 0.44727272727272727,
"grad_norm": 0.04634997248649597,
"learning_rate": 0.0008,
"loss": 2.0942740440368652,
"step": 7380
},
{
"epoch": 0.4473939393939394,
"grad_norm": 0.04760165885090828,
"learning_rate": 0.0008,
"loss": 2.101223945617676,
"step": 7382
},
{
"epoch": 0.44751515151515153,
"grad_norm": 0.048087961971759796,
"learning_rate": 0.0008,
"loss": 2.0647706985473633,
"step": 7384
},
{
"epoch": 0.4476363636363636,
"grad_norm": 0.05823522061109543,
"learning_rate": 0.0008,
"loss": 2.13316011428833,
"step": 7386
},
{
"epoch": 0.44775757575757574,
"grad_norm": 0.06863464415073395,
"learning_rate": 0.0008,
"loss": 2.2145910263061523,
"step": 7388
},
{
"epoch": 0.4478787878787879,
"grad_norm": 0.05954680219292641,
"learning_rate": 0.0008,
"loss": 2.0454633235931396,
"step": 7390
},
{
"epoch": 0.448,
"grad_norm": 0.05587686598300934,
"learning_rate": 0.0008,
"loss": 2.190138578414917,
"step": 7392
},
{
"epoch": 0.44812121212121214,
"grad_norm": 0.05758091062307358,
"learning_rate": 0.0008,
"loss": 2.1205968856811523,
"step": 7394
},
{
"epoch": 0.4482424242424242,
"grad_norm": 0.05075783655047417,
"learning_rate": 0.0008,
"loss": 2.0945122241973877,
"step": 7396
},
{
"epoch": 0.44836363636363635,
"grad_norm": 0.05520037189126015,
"learning_rate": 0.0008,
"loss": 2.1364288330078125,
"step": 7398
},
{
"epoch": 0.4484848484848485,
"grad_norm": 0.054477326571941376,
"learning_rate": 0.0008,
"loss": 2.105804443359375,
"step": 7400
},
{
"epoch": 0.4486060606060606,
"grad_norm": 0.05589365214109421,
"learning_rate": 0.0008,
"loss": 2.0841872692108154,
"step": 7402
},
{
"epoch": 0.44872727272727275,
"grad_norm": 0.04680480808019638,
"learning_rate": 0.0008,
"loss": 2.1067445278167725,
"step": 7404
},
{
"epoch": 0.44884848484848483,
"grad_norm": 0.0474330335855484,
"learning_rate": 0.0008,
"loss": 2.047431230545044,
"step": 7406
},
{
"epoch": 0.44896969696969696,
"grad_norm": 0.048224885016679764,
"learning_rate": 0.0008,
"loss": 2.131682872772217,
"step": 7408
},
{
"epoch": 0.4490909090909091,
"grad_norm": 0.04770480468869209,
"learning_rate": 0.0008,
"loss": 2.06789231300354,
"step": 7410
},
{
"epoch": 0.44921212121212123,
"grad_norm": 0.05122453346848488,
"learning_rate": 0.0008,
"loss": 2.089057207107544,
"step": 7412
},
{
"epoch": 0.4493333333333333,
"grad_norm": 0.047482289373874664,
"learning_rate": 0.0008,
"loss": 2.0681450366973877,
"step": 7414
},
{
"epoch": 0.44945454545454544,
"grad_norm": 0.04671156406402588,
"learning_rate": 0.0008,
"loss": 2.1224365234375,
"step": 7416
},
{
"epoch": 0.4495757575757576,
"grad_norm": 0.05121256411075592,
"learning_rate": 0.0008,
"loss": 2.1112418174743652,
"step": 7418
},
{
"epoch": 0.4496969696969697,
"grad_norm": 0.04588034003973007,
"learning_rate": 0.0008,
"loss": 2.036555290222168,
"step": 7420
},
{
"epoch": 0.44981818181818184,
"grad_norm": 0.04854871332645416,
"learning_rate": 0.0008,
"loss": 2.087265968322754,
"step": 7422
},
{
"epoch": 0.4499393939393939,
"grad_norm": 0.04731190577149391,
"learning_rate": 0.0008,
"loss": 2.135136127471924,
"step": 7424
},
{
"epoch": 0.45006060606060605,
"grad_norm": 0.04748288542032242,
"learning_rate": 0.0008,
"loss": 2.133690595626831,
"step": 7426
},
{
"epoch": 0.4501818181818182,
"grad_norm": 0.048541102558374405,
"learning_rate": 0.0008,
"loss": 2.055795669555664,
"step": 7428
},
{
"epoch": 0.4503030303030303,
"grad_norm": 0.0480864979326725,
"learning_rate": 0.0008,
"loss": 2.125208854675293,
"step": 7430
},
{
"epoch": 0.45042424242424245,
"grad_norm": 0.05218435451388359,
"learning_rate": 0.0008,
"loss": 2.160724639892578,
"step": 7432
},
{
"epoch": 0.4505454545454545,
"grad_norm": 0.04822815954685211,
"learning_rate": 0.0008,
"loss": 2.126342296600342,
"step": 7434
},
{
"epoch": 0.45066666666666666,
"grad_norm": 0.04760834202170372,
"learning_rate": 0.0008,
"loss": 2.092278003692627,
"step": 7436
},
{
"epoch": 0.4507878787878788,
"grad_norm": 0.04630446061491966,
"learning_rate": 0.0008,
"loss": 2.079991102218628,
"step": 7438
},
{
"epoch": 0.4509090909090909,
"grad_norm": 0.04610174521803856,
"learning_rate": 0.0008,
"loss": 2.056349039077759,
"step": 7440
},
{
"epoch": 0.451030303030303,
"grad_norm": 0.04851452633738518,
"learning_rate": 0.0008,
"loss": 2.135129690170288,
"step": 7442
},
{
"epoch": 0.45115151515151514,
"grad_norm": 0.046956323087215424,
"learning_rate": 0.0008,
"loss": 2.063232183456421,
"step": 7444
},
{
"epoch": 0.45127272727272727,
"grad_norm": 0.0474533848464489,
"learning_rate": 0.0008,
"loss": 2.1185524463653564,
"step": 7446
},
{
"epoch": 0.4513939393939394,
"grad_norm": 0.04765245318412781,
"learning_rate": 0.0008,
"loss": 2.142416000366211,
"step": 7448
},
{
"epoch": 0.45151515151515154,
"grad_norm": 0.04774774610996246,
"learning_rate": 0.0008,
"loss": 2.15474009513855,
"step": 7450
},
{
"epoch": 0.4516363636363636,
"grad_norm": 0.048135463148355484,
"learning_rate": 0.0008,
"loss": 2.1069586277008057,
"step": 7452
},
{
"epoch": 0.45175757575757575,
"grad_norm": 0.054653801023960114,
"learning_rate": 0.0008,
"loss": 2.1577935218811035,
"step": 7454
},
{
"epoch": 0.4518787878787879,
"grad_norm": 0.054024141281843185,
"learning_rate": 0.0008,
"loss": 2.1097981929779053,
"step": 7456
},
{
"epoch": 0.452,
"grad_norm": 0.04655935615301132,
"learning_rate": 0.0008,
"loss": 2.1055362224578857,
"step": 7458
},
{
"epoch": 0.45212121212121215,
"grad_norm": 0.04921692982316017,
"learning_rate": 0.0008,
"loss": 2.0664255619049072,
"step": 7460
},
{
"epoch": 0.4522424242424242,
"grad_norm": 0.0491446778178215,
"learning_rate": 0.0008,
"loss": 2.1529574394226074,
"step": 7462
},
{
"epoch": 0.45236363636363636,
"grad_norm": 0.048092860728502274,
"learning_rate": 0.0008,
"loss": 2.160560131072998,
"step": 7464
},
{
"epoch": 0.4524848484848485,
"grad_norm": 0.04785016179084778,
"learning_rate": 0.0008,
"loss": 2.0908498764038086,
"step": 7466
},
{
"epoch": 0.4526060606060606,
"grad_norm": 0.04789261519908905,
"learning_rate": 0.0008,
"loss": 2.05668306350708,
"step": 7468
},
{
"epoch": 0.4527272727272727,
"grad_norm": 0.047095540910959244,
"learning_rate": 0.0008,
"loss": 2.0993738174438477,
"step": 7470
},
{
"epoch": 0.45284848484848483,
"grad_norm": 0.0471954382956028,
"learning_rate": 0.0008,
"loss": 2.125953197479248,
"step": 7472
},
{
"epoch": 0.45296969696969697,
"grad_norm": 0.05001875385642052,
"learning_rate": 0.0008,
"loss": 2.140200138092041,
"step": 7474
},
{
"epoch": 0.4530909090909091,
"grad_norm": 0.0545443519949913,
"learning_rate": 0.0008,
"loss": 2.098280429840088,
"step": 7476
},
{
"epoch": 0.45321212121212123,
"grad_norm": 0.04948710650205612,
"learning_rate": 0.0008,
"loss": 2.1384308338165283,
"step": 7478
},
{
"epoch": 0.4533333333333333,
"grad_norm": 0.05152368173003197,
"learning_rate": 0.0008,
"loss": 2.175987482070923,
"step": 7480
},
{
"epoch": 0.45345454545454544,
"grad_norm": 0.04728783667087555,
"learning_rate": 0.0008,
"loss": 2.086487293243408,
"step": 7482
},
{
"epoch": 0.4535757575757576,
"grad_norm": 0.047158095985651016,
"learning_rate": 0.0008,
"loss": 2.115760326385498,
"step": 7484
},
{
"epoch": 0.4536969696969697,
"grad_norm": 0.04761200770735741,
"learning_rate": 0.0008,
"loss": 2.072856903076172,
"step": 7486
},
{
"epoch": 0.45381818181818184,
"grad_norm": 0.05041646212339401,
"learning_rate": 0.0008,
"loss": 2.065770149230957,
"step": 7488
},
{
"epoch": 0.4539393939393939,
"grad_norm": 0.04841908439993858,
"learning_rate": 0.0008,
"loss": 2.137028694152832,
"step": 7490
},
{
"epoch": 0.45406060606060605,
"grad_norm": 0.048900168389081955,
"learning_rate": 0.0008,
"loss": 2.158881425857544,
"step": 7492
},
{
"epoch": 0.4541818181818182,
"grad_norm": 0.04665938764810562,
"learning_rate": 0.0008,
"loss": 2.0666909217834473,
"step": 7494
},
{
"epoch": 0.4543030303030303,
"grad_norm": 0.04539509117603302,
"learning_rate": 0.0008,
"loss": 2.0404577255249023,
"step": 7496
},
{
"epoch": 0.4544242424242424,
"grad_norm": 0.04724154993891716,
"learning_rate": 0.0008,
"loss": 2.080249309539795,
"step": 7498
},
{
"epoch": 0.45454545454545453,
"grad_norm": 0.04866427183151245,
"learning_rate": 0.0008,
"loss": 2.12614369392395,
"step": 7500
},
{
"epoch": 0.45466666666666666,
"grad_norm": 0.04632652923464775,
"learning_rate": 0.0008,
"loss": 2.03940486907959,
"step": 7502
},
{
"epoch": 0.4547878787878788,
"grad_norm": 0.04563645273447037,
"learning_rate": 0.0008,
"loss": 2.087388277053833,
"step": 7504
},
{
"epoch": 0.45490909090909093,
"grad_norm": 0.0492389015853405,
"learning_rate": 0.0008,
"loss": 2.034419536590576,
"step": 7506
},
{
"epoch": 0.455030303030303,
"grad_norm": 0.050164662301540375,
"learning_rate": 0.0008,
"loss": 2.2095398902893066,
"step": 7508
},
{
"epoch": 0.45515151515151514,
"grad_norm": 0.05091225728392601,
"learning_rate": 0.0008,
"loss": 2.1542577743530273,
"step": 7510
},
{
"epoch": 0.4552727272727273,
"grad_norm": 0.046323105692863464,
"learning_rate": 0.0008,
"loss": 2.1046574115753174,
"step": 7512
},
{
"epoch": 0.4553939393939394,
"grad_norm": 0.050830647349357605,
"learning_rate": 0.0008,
"loss": 2.1260132789611816,
"step": 7514
},
{
"epoch": 0.45551515151515154,
"grad_norm": 0.04947786405682564,
"learning_rate": 0.0008,
"loss": 2.119473695755005,
"step": 7516
},
{
"epoch": 0.4556363636363636,
"grad_norm": 0.05142691358923912,
"learning_rate": 0.0008,
"loss": 2.0841848850250244,
"step": 7518
},
{
"epoch": 0.45575757575757575,
"grad_norm": 0.05312613025307655,
"learning_rate": 0.0008,
"loss": 2.1724741458892822,
"step": 7520
},
{
"epoch": 0.4558787878787879,
"grad_norm": 0.05401022359728813,
"learning_rate": 0.0008,
"loss": 2.093640089035034,
"step": 7522
},
{
"epoch": 0.456,
"grad_norm": 0.04742462933063507,
"learning_rate": 0.0008,
"loss": 2.029794454574585,
"step": 7524
},
{
"epoch": 0.4561212121212121,
"grad_norm": 0.04728186875581741,
"learning_rate": 0.0008,
"loss": 2.10357403755188,
"step": 7526
},
{
"epoch": 0.4562424242424242,
"grad_norm": 0.04748276621103287,
"learning_rate": 0.0008,
"loss": 2.085740566253662,
"step": 7528
},
{
"epoch": 0.45636363636363636,
"grad_norm": 0.048909734934568405,
"learning_rate": 0.0008,
"loss": 2.1611757278442383,
"step": 7530
},
{
"epoch": 0.4564848484848485,
"grad_norm": 0.04775683581829071,
"learning_rate": 0.0008,
"loss": 2.117316722869873,
"step": 7532
},
{
"epoch": 0.4566060606060606,
"grad_norm": 0.06255496293306351,
"learning_rate": 0.0008,
"loss": 2.0965731143951416,
"step": 7534
},
{
"epoch": 0.4567272727272727,
"grad_norm": 0.059664759784936905,
"learning_rate": 0.0008,
"loss": 2.0393142700195312,
"step": 7536
},
{
"epoch": 0.45684848484848484,
"grad_norm": 0.05820098891854286,
"learning_rate": 0.0008,
"loss": 2.13781476020813,
"step": 7538
},
{
"epoch": 0.45696969696969697,
"grad_norm": 0.05842194706201553,
"learning_rate": 0.0008,
"loss": 2.0919718742370605,
"step": 7540
},
{
"epoch": 0.4570909090909091,
"grad_norm": 0.05291518196463585,
"learning_rate": 0.0008,
"loss": 2.118253707885742,
"step": 7542
},
{
"epoch": 0.45721212121212124,
"grad_norm": 0.04654444009065628,
"learning_rate": 0.0008,
"loss": 2.0196242332458496,
"step": 7544
},
{
"epoch": 0.4573333333333333,
"grad_norm": 0.04749660938978195,
"learning_rate": 0.0008,
"loss": 2.110804796218872,
"step": 7546
},
{
"epoch": 0.45745454545454545,
"grad_norm": 0.05107355862855911,
"learning_rate": 0.0008,
"loss": 2.058840036392212,
"step": 7548
},
{
"epoch": 0.4575757575757576,
"grad_norm": 0.04937640577554703,
"learning_rate": 0.0008,
"loss": 2.1323766708374023,
"step": 7550
},
{
"epoch": 0.4576969696969697,
"grad_norm": 0.04788240045309067,
"learning_rate": 0.0008,
"loss": 2.106238603591919,
"step": 7552
},
{
"epoch": 0.4578181818181818,
"grad_norm": 0.05887771025300026,
"learning_rate": 0.0008,
"loss": 2.112138032913208,
"step": 7554
},
{
"epoch": 0.4579393939393939,
"grad_norm": 0.05976273491978645,
"learning_rate": 0.0008,
"loss": 2.115326166152954,
"step": 7556
},
{
"epoch": 0.45806060606060606,
"grad_norm": 0.06427808851003647,
"learning_rate": 0.0008,
"loss": 2.1589040756225586,
"step": 7558
},
{
"epoch": 0.4581818181818182,
"grad_norm": 0.06619606167078018,
"learning_rate": 0.0008,
"loss": 2.0671229362487793,
"step": 7560
},
{
"epoch": 0.4583030303030303,
"grad_norm": 0.06546876579523087,
"learning_rate": 0.0008,
"loss": 2.147027015686035,
"step": 7562
},
{
"epoch": 0.4584242424242424,
"grad_norm": 0.05641326680779457,
"learning_rate": 0.0008,
"loss": 2.099329948425293,
"step": 7564
},
{
"epoch": 0.45854545454545453,
"grad_norm": 0.05222805216908455,
"learning_rate": 0.0008,
"loss": 2.180464744567871,
"step": 7566
},
{
"epoch": 0.45866666666666667,
"grad_norm": 0.051673028618097305,
"learning_rate": 0.0008,
"loss": 2.1070027351379395,
"step": 7568
},
{
"epoch": 0.4587878787878788,
"grad_norm": 0.05040234327316284,
"learning_rate": 0.0008,
"loss": 2.1427807807922363,
"step": 7570
},
{
"epoch": 0.45890909090909093,
"grad_norm": 0.0472881942987442,
"learning_rate": 0.0008,
"loss": 2.1372458934783936,
"step": 7572
},
{
"epoch": 0.459030303030303,
"grad_norm": 0.04719431325793266,
"learning_rate": 0.0008,
"loss": 2.0121612548828125,
"step": 7574
},
{
"epoch": 0.45915151515151514,
"grad_norm": 0.049041129648685455,
"learning_rate": 0.0008,
"loss": 2.17842960357666,
"step": 7576
},
{
"epoch": 0.4592727272727273,
"grad_norm": 0.04618363454937935,
"learning_rate": 0.0008,
"loss": 2.171497106552124,
"step": 7578
},
{
"epoch": 0.4593939393939394,
"grad_norm": 0.048294827342033386,
"learning_rate": 0.0008,
"loss": 2.213261365890503,
"step": 7580
},
{
"epoch": 0.4595151515151515,
"grad_norm": 0.04733685776591301,
"learning_rate": 0.0008,
"loss": 2.1308882236480713,
"step": 7582
},
{
"epoch": 0.4596363636363636,
"grad_norm": 0.047278132289648056,
"learning_rate": 0.0008,
"loss": 2.0860157012939453,
"step": 7584
},
{
"epoch": 0.45975757575757575,
"grad_norm": 0.047056518495082855,
"learning_rate": 0.0008,
"loss": 2.1041195392608643,
"step": 7586
},
{
"epoch": 0.4598787878787879,
"grad_norm": 0.04911809042096138,
"learning_rate": 0.0008,
"loss": 2.080740451812744,
"step": 7588
},
{
"epoch": 0.46,
"grad_norm": 0.04814844951033592,
"learning_rate": 0.0008,
"loss": 2.1114141941070557,
"step": 7590
},
{
"epoch": 0.4601212121212121,
"grad_norm": 0.046983975917100906,
"learning_rate": 0.0008,
"loss": 2.1417794227600098,
"step": 7592
},
{
"epoch": 0.46024242424242423,
"grad_norm": 0.04844740778207779,
"learning_rate": 0.0008,
"loss": 2.1814465522766113,
"step": 7594
},
{
"epoch": 0.46036363636363636,
"grad_norm": 0.04905096814036369,
"learning_rate": 0.0008,
"loss": 2.112128973007202,
"step": 7596
},
{
"epoch": 0.4604848484848485,
"grad_norm": 0.05014410614967346,
"learning_rate": 0.0008,
"loss": 2.1078805923461914,
"step": 7598
},
{
"epoch": 0.46060606060606063,
"grad_norm": 0.04779994860291481,
"learning_rate": 0.0008,
"loss": 2.0902342796325684,
"step": 7600
},
{
"epoch": 0.4607272727272727,
"grad_norm": 0.04870864748954773,
"learning_rate": 0.0008,
"loss": 2.1678943634033203,
"step": 7602
},
{
"epoch": 0.46084848484848484,
"grad_norm": 0.051615260541439056,
"learning_rate": 0.0008,
"loss": 2.0678837299346924,
"step": 7604
},
{
"epoch": 0.460969696969697,
"grad_norm": 0.049924734979867935,
"learning_rate": 0.0008,
"loss": 2.099114418029785,
"step": 7606
},
{
"epoch": 0.4610909090909091,
"grad_norm": 0.053907155990600586,
"learning_rate": 0.0008,
"loss": 2.07541561126709,
"step": 7608
},
{
"epoch": 0.4612121212121212,
"grad_norm": 0.046316079795360565,
"learning_rate": 0.0008,
"loss": 2.098017692565918,
"step": 7610
},
{
"epoch": 0.4613333333333333,
"grad_norm": 0.04889779910445213,
"learning_rate": 0.0008,
"loss": 2.1116440296173096,
"step": 7612
},
{
"epoch": 0.46145454545454545,
"grad_norm": 0.049767106771469116,
"learning_rate": 0.0008,
"loss": 2.1062846183776855,
"step": 7614
},
{
"epoch": 0.4615757575757576,
"grad_norm": 0.04530864581465721,
"learning_rate": 0.0008,
"loss": 2.063171863555908,
"step": 7616
},
{
"epoch": 0.4616969696969697,
"grad_norm": 0.04718030244112015,
"learning_rate": 0.0008,
"loss": 2.1044156551361084,
"step": 7618
},
{
"epoch": 0.4618181818181818,
"grad_norm": 0.04808547720313072,
"learning_rate": 0.0008,
"loss": 2.1620566844940186,
"step": 7620
},
{
"epoch": 0.4619393939393939,
"grad_norm": 0.046241093426942825,
"learning_rate": 0.0008,
"loss": 2.1145522594451904,
"step": 7622
},
{
"epoch": 0.46206060606060606,
"grad_norm": 0.0461094044148922,
"learning_rate": 0.0008,
"loss": 2.144455671310425,
"step": 7624
},
{
"epoch": 0.4621818181818182,
"grad_norm": 0.04753754287958145,
"learning_rate": 0.0008,
"loss": 2.087293863296509,
"step": 7626
},
{
"epoch": 0.4623030303030303,
"grad_norm": 0.05131247267127037,
"learning_rate": 0.0008,
"loss": 2.151812791824341,
"step": 7628
},
{
"epoch": 0.4624242424242424,
"grad_norm": 0.048498645424842834,
"learning_rate": 0.0008,
"loss": 2.108394145965576,
"step": 7630
},
{
"epoch": 0.46254545454545454,
"grad_norm": 0.04701276123523712,
"learning_rate": 0.0008,
"loss": 2.1857662200927734,
"step": 7632
},
{
"epoch": 0.46266666666666667,
"grad_norm": 0.04785503074526787,
"learning_rate": 0.0008,
"loss": 2.12408709526062,
"step": 7634
},
{
"epoch": 0.4627878787878788,
"grad_norm": 0.04961546137928963,
"learning_rate": 0.0008,
"loss": 2.095566987991333,
"step": 7636
},
{
"epoch": 0.46290909090909094,
"grad_norm": 0.0463014654815197,
"learning_rate": 0.0008,
"loss": 2.1292126178741455,
"step": 7638
},
{
"epoch": 0.463030303030303,
"grad_norm": 0.04931877180933952,
"learning_rate": 0.0008,
"loss": 2.1352901458740234,
"step": 7640
},
{
"epoch": 0.46315151515151515,
"grad_norm": 0.05236108973622322,
"learning_rate": 0.0008,
"loss": 2.1170663833618164,
"step": 7642
},
{
"epoch": 0.4632727272727273,
"grad_norm": 0.05628889426589012,
"learning_rate": 0.0008,
"loss": 2.104698657989502,
"step": 7644
},
{
"epoch": 0.4633939393939394,
"grad_norm": 0.05381162837147713,
"learning_rate": 0.0008,
"loss": 2.120851516723633,
"step": 7646
},
{
"epoch": 0.4635151515151515,
"grad_norm": 0.04946468397974968,
"learning_rate": 0.0008,
"loss": 2.152106285095215,
"step": 7648
},
{
"epoch": 0.4636363636363636,
"grad_norm": 0.04655197635293007,
"learning_rate": 0.0008,
"loss": 2.075639486312866,
"step": 7650
},
{
"epoch": 0.46375757575757576,
"grad_norm": 0.04973439499735832,
"learning_rate": 0.0008,
"loss": 2.203655958175659,
"step": 7652
},
{
"epoch": 0.4638787878787879,
"grad_norm": 0.04819870740175247,
"learning_rate": 0.0008,
"loss": 2.1141364574432373,
"step": 7654
},
{
"epoch": 0.464,
"grad_norm": 0.05483425408601761,
"learning_rate": 0.0008,
"loss": 2.1490678787231445,
"step": 7656
},
{
"epoch": 0.4641212121212121,
"grad_norm": 0.053899019956588745,
"learning_rate": 0.0008,
"loss": 2.1072120666503906,
"step": 7658
},
{
"epoch": 0.46424242424242423,
"grad_norm": 0.053938742727041245,
"learning_rate": 0.0008,
"loss": 2.148496389389038,
"step": 7660
},
{
"epoch": 0.46436363636363637,
"grad_norm": 0.052833396941423416,
"learning_rate": 0.0008,
"loss": 2.0321550369262695,
"step": 7662
},
{
"epoch": 0.4644848484848485,
"grad_norm": 0.056913189589977264,
"learning_rate": 0.0008,
"loss": 2.0975191593170166,
"step": 7664
},
{
"epoch": 0.46460606060606063,
"grad_norm": 0.06311430782079697,
"learning_rate": 0.0008,
"loss": 2.142453193664551,
"step": 7666
},
{
"epoch": 0.4647272727272727,
"grad_norm": 0.05416390672326088,
"learning_rate": 0.0008,
"loss": 2.0819923877716064,
"step": 7668
},
{
"epoch": 0.46484848484848484,
"grad_norm": 0.055941712111234665,
"learning_rate": 0.0008,
"loss": 2.1338939666748047,
"step": 7670
},
{
"epoch": 0.464969696969697,
"grad_norm": 0.05262612923979759,
"learning_rate": 0.0008,
"loss": 2.1096866130828857,
"step": 7672
},
{
"epoch": 0.4650909090909091,
"grad_norm": 0.052230916917324066,
"learning_rate": 0.0008,
"loss": 2.144395589828491,
"step": 7674
},
{
"epoch": 0.4652121212121212,
"grad_norm": 0.04831041023135185,
"learning_rate": 0.0008,
"loss": 2.1559977531433105,
"step": 7676
},
{
"epoch": 0.4653333333333333,
"grad_norm": 0.047479476779699326,
"learning_rate": 0.0008,
"loss": 2.1632111072540283,
"step": 7678
},
{
"epoch": 0.46545454545454545,
"grad_norm": 0.047473058104515076,
"learning_rate": 0.0008,
"loss": 2.056450843811035,
"step": 7680
},
{
"epoch": 0.4655757575757576,
"grad_norm": 0.04743990674614906,
"learning_rate": 0.0008,
"loss": 2.1572916507720947,
"step": 7682
},
{
"epoch": 0.4656969696969697,
"grad_norm": 0.04639030247926712,
"learning_rate": 0.0008,
"loss": 2.13230037689209,
"step": 7684
},
{
"epoch": 0.4658181818181818,
"grad_norm": 0.047521863132715225,
"learning_rate": 0.0008,
"loss": 2.1784491539001465,
"step": 7686
},
{
"epoch": 0.46593939393939393,
"grad_norm": 0.04986390471458435,
"learning_rate": 0.0008,
"loss": 2.209078788757324,
"step": 7688
},
{
"epoch": 0.46606060606060606,
"grad_norm": 0.048058025538921356,
"learning_rate": 0.0008,
"loss": 2.1271729469299316,
"step": 7690
},
{
"epoch": 0.4661818181818182,
"grad_norm": 0.048516057431697845,
"learning_rate": 0.0008,
"loss": 2.098435878753662,
"step": 7692
},
{
"epoch": 0.46630303030303033,
"grad_norm": 0.053064409643411636,
"learning_rate": 0.0008,
"loss": 2.079401969909668,
"step": 7694
},
{
"epoch": 0.4664242424242424,
"grad_norm": 0.05535154417157173,
"learning_rate": 0.0008,
"loss": 2.1457059383392334,
"step": 7696
},
{
"epoch": 0.46654545454545454,
"grad_norm": 0.0560523085296154,
"learning_rate": 0.0008,
"loss": 2.0516417026519775,
"step": 7698
},
{
"epoch": 0.4666666666666667,
"grad_norm": 0.0626177117228508,
"learning_rate": 0.0008,
"loss": 2.126251697540283,
"step": 7700
},
{
"epoch": 0.4667878787878788,
"grad_norm": 0.06692148745059967,
"learning_rate": 0.0008,
"loss": 2.137068033218384,
"step": 7702
},
{
"epoch": 0.4669090909090909,
"grad_norm": 0.055504221469163895,
"learning_rate": 0.0008,
"loss": 2.1791481971740723,
"step": 7704
},
{
"epoch": 0.467030303030303,
"grad_norm": 0.05313727259635925,
"learning_rate": 0.0008,
"loss": 2.187209367752075,
"step": 7706
},
{
"epoch": 0.46715151515151515,
"grad_norm": 0.050593357533216476,
"learning_rate": 0.0008,
"loss": 2.133718252182007,
"step": 7708
},
{
"epoch": 0.4672727272727273,
"grad_norm": 0.05548588186502457,
"learning_rate": 0.0008,
"loss": 2.140845775604248,
"step": 7710
},
{
"epoch": 0.4673939393939394,
"grad_norm": 0.05020551756024361,
"learning_rate": 0.0008,
"loss": 2.1493687629699707,
"step": 7712
},
{
"epoch": 0.4675151515151515,
"grad_norm": 0.05388456583023071,
"learning_rate": 0.0008,
"loss": 2.107923984527588,
"step": 7714
},
{
"epoch": 0.46763636363636363,
"grad_norm": 0.060066431760787964,
"learning_rate": 0.0008,
"loss": 2.182857036590576,
"step": 7716
},
{
"epoch": 0.46775757575757576,
"grad_norm": 0.048560865223407745,
"learning_rate": 0.0008,
"loss": 2.1714446544647217,
"step": 7718
},
{
"epoch": 0.4678787878787879,
"grad_norm": 0.048549897968769073,
"learning_rate": 0.0008,
"loss": 2.1300153732299805,
"step": 7720
},
{
"epoch": 0.468,
"grad_norm": 0.047158870846033096,
"learning_rate": 0.0008,
"loss": 2.1113686561584473,
"step": 7722
},
{
"epoch": 0.4681212121212121,
"grad_norm": 0.04859399050474167,
"learning_rate": 0.0008,
"loss": 2.0611348152160645,
"step": 7724
},
{
"epoch": 0.46824242424242424,
"grad_norm": 0.04683173820376396,
"learning_rate": 0.0008,
"loss": 2.1530580520629883,
"step": 7726
},
{
"epoch": 0.46836363636363637,
"grad_norm": 0.05119595304131508,
"learning_rate": 0.0008,
"loss": 2.074812650680542,
"step": 7728
},
{
"epoch": 0.4684848484848485,
"grad_norm": 0.05134538188576698,
"learning_rate": 0.0008,
"loss": 2.134648323059082,
"step": 7730
},
{
"epoch": 0.4686060606060606,
"grad_norm": 0.04927036166191101,
"learning_rate": 0.0008,
"loss": 2.0895955562591553,
"step": 7732
},
{
"epoch": 0.4687272727272727,
"grad_norm": 0.05107812583446503,
"learning_rate": 0.0008,
"loss": 2.1063473224639893,
"step": 7734
},
{
"epoch": 0.46884848484848485,
"grad_norm": 0.0467974878847599,
"learning_rate": 0.0008,
"loss": 2.1186671257019043,
"step": 7736
},
{
"epoch": 0.468969696969697,
"grad_norm": 0.04699726775288582,
"learning_rate": 0.0008,
"loss": 2.100695848464966,
"step": 7738
},
{
"epoch": 0.4690909090909091,
"grad_norm": 0.04609157145023346,
"learning_rate": 0.0008,
"loss": 2.128302574157715,
"step": 7740
},
{
"epoch": 0.4692121212121212,
"grad_norm": 0.052142996340990067,
"learning_rate": 0.0008,
"loss": 2.159088134765625,
"step": 7742
},
{
"epoch": 0.4693333333333333,
"grad_norm": 0.04935440048575401,
"learning_rate": 0.0008,
"loss": 2.1979434490203857,
"step": 7744
},
{
"epoch": 0.46945454545454546,
"grad_norm": 0.04624375328421593,
"learning_rate": 0.0008,
"loss": 2.0792617797851562,
"step": 7746
},
{
"epoch": 0.4695757575757576,
"grad_norm": 0.04524421691894531,
"learning_rate": 0.0008,
"loss": 2.111048936843872,
"step": 7748
},
{
"epoch": 0.4696969696969697,
"grad_norm": 0.04965274780988693,
"learning_rate": 0.0008,
"loss": 2.150052309036255,
"step": 7750
},
{
"epoch": 0.4698181818181818,
"grad_norm": 0.05734262242913246,
"learning_rate": 0.0008,
"loss": 2.114938259124756,
"step": 7752
},
{
"epoch": 0.46993939393939393,
"grad_norm": 0.05163584649562836,
"learning_rate": 0.0008,
"loss": 2.1804885864257812,
"step": 7754
},
{
"epoch": 0.47006060606060607,
"grad_norm": 0.053141988813877106,
"learning_rate": 0.0008,
"loss": 2.1107120513916016,
"step": 7756
},
{
"epoch": 0.4701818181818182,
"grad_norm": 0.05395261198282242,
"learning_rate": 0.0008,
"loss": 2.1411752700805664,
"step": 7758
},
{
"epoch": 0.4703030303030303,
"grad_norm": 0.054602380841970444,
"learning_rate": 0.0008,
"loss": 2.2198328971862793,
"step": 7760
},
{
"epoch": 0.4704242424242424,
"grad_norm": 0.049698665738105774,
"learning_rate": 0.0008,
"loss": 2.2441580295562744,
"step": 7762
},
{
"epoch": 0.47054545454545454,
"grad_norm": 0.04899683967232704,
"learning_rate": 0.0008,
"loss": 2.161566972732544,
"step": 7764
},
{
"epoch": 0.4706666666666667,
"grad_norm": 0.04801814630627632,
"learning_rate": 0.0008,
"loss": 2.0886361598968506,
"step": 7766
},
{
"epoch": 0.4707878787878788,
"grad_norm": 0.04720612242817879,
"learning_rate": 0.0008,
"loss": 2.0627472400665283,
"step": 7768
},
{
"epoch": 0.4709090909090909,
"grad_norm": 0.0481283962726593,
"learning_rate": 0.0008,
"loss": 2.084184169769287,
"step": 7770
},
{
"epoch": 0.471030303030303,
"grad_norm": 0.04973241314291954,
"learning_rate": 0.0008,
"loss": 2.1606147289276123,
"step": 7772
},
{
"epoch": 0.47115151515151515,
"grad_norm": 0.04914570599794388,
"learning_rate": 0.0008,
"loss": 2.0816752910614014,
"step": 7774
},
{
"epoch": 0.4712727272727273,
"grad_norm": 0.0501277893781662,
"learning_rate": 0.0008,
"loss": 2.171379566192627,
"step": 7776
},
{
"epoch": 0.4713939393939394,
"grad_norm": 0.054322950541973114,
"learning_rate": 0.0008,
"loss": 2.1811516284942627,
"step": 7778
},
{
"epoch": 0.4715151515151515,
"grad_norm": 0.05051146075129509,
"learning_rate": 0.0008,
"loss": 2.0741701126098633,
"step": 7780
},
{
"epoch": 0.47163636363636363,
"grad_norm": 0.04586479067802429,
"learning_rate": 0.0008,
"loss": 2.111088514328003,
"step": 7782
},
{
"epoch": 0.47175757575757576,
"grad_norm": 0.0474519319832325,
"learning_rate": 0.0008,
"loss": 2.1311535835266113,
"step": 7784
},
{
"epoch": 0.4718787878787879,
"grad_norm": 0.04756001755595207,
"learning_rate": 0.0008,
"loss": 2.1129307746887207,
"step": 7786
},
{
"epoch": 0.472,
"grad_norm": 0.05132107064127922,
"learning_rate": 0.0008,
"loss": 2.080934524536133,
"step": 7788
},
{
"epoch": 0.4721212121212121,
"grad_norm": 0.049793656915426254,
"learning_rate": 0.0008,
"loss": 2.0702879428863525,
"step": 7790
},
{
"epoch": 0.47224242424242424,
"grad_norm": 0.04897426813840866,
"learning_rate": 0.0008,
"loss": 2.139272451400757,
"step": 7792
},
{
"epoch": 0.4723636363636364,
"grad_norm": 0.054162174463272095,
"learning_rate": 0.0008,
"loss": 2.139761209487915,
"step": 7794
},
{
"epoch": 0.4724848484848485,
"grad_norm": 0.06623450666666031,
"learning_rate": 0.0008,
"loss": 2.169760227203369,
"step": 7796
},
{
"epoch": 0.4726060606060606,
"grad_norm": 0.05541841685771942,
"learning_rate": 0.0008,
"loss": 2.1518800258636475,
"step": 7798
},
{
"epoch": 0.4727272727272727,
"grad_norm": 0.05782613158226013,
"learning_rate": 0.0008,
"loss": 2.1647908687591553,
"step": 7800
},
{
"epoch": 0.47284848484848485,
"grad_norm": 0.05477684363722801,
"learning_rate": 0.0008,
"loss": 2.1423285007476807,
"step": 7802
},
{
"epoch": 0.472969696969697,
"grad_norm": 0.052356112748384476,
"learning_rate": 0.0008,
"loss": 2.172147274017334,
"step": 7804
},
{
"epoch": 0.4730909090909091,
"grad_norm": 0.05273795127868652,
"learning_rate": 0.0008,
"loss": 2.1792378425598145,
"step": 7806
},
{
"epoch": 0.4732121212121212,
"grad_norm": 0.04733181372284889,
"learning_rate": 0.0008,
"loss": 2.210275888442993,
"step": 7808
},
{
"epoch": 0.47333333333333333,
"grad_norm": 0.04708237200975418,
"learning_rate": 0.0008,
"loss": 2.1415843963623047,
"step": 7810
},
{
"epoch": 0.47345454545454546,
"grad_norm": 0.04973042383790016,
"learning_rate": 0.0008,
"loss": 2.1726772785186768,
"step": 7812
},
{
"epoch": 0.4735757575757576,
"grad_norm": 0.058485113084316254,
"learning_rate": 0.0008,
"loss": 2.1601686477661133,
"step": 7814
},
{
"epoch": 0.47369696969696967,
"grad_norm": 0.052819639444351196,
"learning_rate": 0.0008,
"loss": 2.1721363067626953,
"step": 7816
},
{
"epoch": 0.4738181818181818,
"grad_norm": 0.047417815774679184,
"learning_rate": 0.0008,
"loss": 2.040755033493042,
"step": 7818
},
{
"epoch": 0.47393939393939394,
"grad_norm": 0.047737088054418564,
"learning_rate": 0.0008,
"loss": 2.1780011653900146,
"step": 7820
},
{
"epoch": 0.47406060606060607,
"grad_norm": 0.04836716502904892,
"learning_rate": 0.0008,
"loss": 2.106632709503174,
"step": 7822
},
{
"epoch": 0.4741818181818182,
"grad_norm": 0.04749545827507973,
"learning_rate": 0.0008,
"loss": 2.129652500152588,
"step": 7824
},
{
"epoch": 0.4743030303030303,
"grad_norm": 0.0461200475692749,
"learning_rate": 0.0008,
"loss": 2.1432957649230957,
"step": 7826
},
{
"epoch": 0.4744242424242424,
"grad_norm": 0.04649371653795242,
"learning_rate": 0.0008,
"loss": 2.1268410682678223,
"step": 7828
},
{
"epoch": 0.47454545454545455,
"grad_norm": 0.04604218900203705,
"learning_rate": 0.0008,
"loss": 2.0585973262786865,
"step": 7830
},
{
"epoch": 0.4746666666666667,
"grad_norm": 0.04769732803106308,
"learning_rate": 0.0008,
"loss": 2.1329891681671143,
"step": 7832
},
{
"epoch": 0.4747878787878788,
"grad_norm": 0.047733161598443985,
"learning_rate": 0.0008,
"loss": 2.1028199195861816,
"step": 7834
},
{
"epoch": 0.4749090909090909,
"grad_norm": 0.048652637749910355,
"learning_rate": 0.0008,
"loss": 2.082977294921875,
"step": 7836
},
{
"epoch": 0.475030303030303,
"grad_norm": 0.0453946627676487,
"learning_rate": 0.0008,
"loss": 2.147714138031006,
"step": 7838
},
{
"epoch": 0.47515151515151516,
"grad_norm": 0.04920551925897598,
"learning_rate": 0.0008,
"loss": 2.1078643798828125,
"step": 7840
},
{
"epoch": 0.4752727272727273,
"grad_norm": 0.058100901544094086,
"learning_rate": 0.0008,
"loss": 2.1616594791412354,
"step": 7842
},
{
"epoch": 0.47539393939393937,
"grad_norm": 0.07178092002868652,
"learning_rate": 0.0008,
"loss": 2.1851024627685547,
"step": 7844
},
{
"epoch": 0.4755151515151515,
"grad_norm": 0.06142118573188782,
"learning_rate": 0.0008,
"loss": 2.1309378147125244,
"step": 7846
},
{
"epoch": 0.47563636363636363,
"grad_norm": 0.051846470683813095,
"learning_rate": 0.0008,
"loss": 2.0791690349578857,
"step": 7848
},
{
"epoch": 0.47575757575757577,
"grad_norm": 0.05302877724170685,
"learning_rate": 0.0008,
"loss": 2.177856922149658,
"step": 7850
},
{
"epoch": 0.4758787878787879,
"grad_norm": 0.04779786244034767,
"learning_rate": 0.0008,
"loss": 2.131756544113159,
"step": 7852
},
{
"epoch": 0.476,
"grad_norm": 0.0551488995552063,
"learning_rate": 0.0008,
"loss": 2.1308631896972656,
"step": 7854
},
{
"epoch": 0.4761212121212121,
"grad_norm": 0.05077655613422394,
"learning_rate": 0.0008,
"loss": 2.16190242767334,
"step": 7856
},
{
"epoch": 0.47624242424242424,
"grad_norm": 0.048700492829084396,
"learning_rate": 0.0008,
"loss": 2.2228965759277344,
"step": 7858
},
{
"epoch": 0.4763636363636364,
"grad_norm": 0.04932093992829323,
"learning_rate": 0.0008,
"loss": 2.189732551574707,
"step": 7860
},
{
"epoch": 0.4764848484848485,
"grad_norm": 0.047903090715408325,
"learning_rate": 0.0008,
"loss": 2.142238140106201,
"step": 7862
},
{
"epoch": 0.4766060606060606,
"grad_norm": 0.0464843325316906,
"learning_rate": 0.0008,
"loss": 2.1494863033294678,
"step": 7864
},
{
"epoch": 0.4767272727272727,
"grad_norm": 0.0470871701836586,
"learning_rate": 0.0008,
"loss": 2.0988240242004395,
"step": 7866
},
{
"epoch": 0.47684848484848485,
"grad_norm": 0.0481981560587883,
"learning_rate": 0.0008,
"loss": 2.1299610137939453,
"step": 7868
},
{
"epoch": 0.476969696969697,
"grad_norm": 0.0452624186873436,
"learning_rate": 0.0008,
"loss": 2.1494200229644775,
"step": 7870
},
{
"epoch": 0.47709090909090907,
"grad_norm": 0.0462813638150692,
"learning_rate": 0.0008,
"loss": 2.100618839263916,
"step": 7872
},
{
"epoch": 0.4772121212121212,
"grad_norm": 0.050225067883729935,
"learning_rate": 0.0008,
"loss": 2.169041872024536,
"step": 7874
},
{
"epoch": 0.47733333333333333,
"grad_norm": 0.0475248321890831,
"learning_rate": 0.0008,
"loss": 2.1589109897613525,
"step": 7876
},
{
"epoch": 0.47745454545454546,
"grad_norm": 0.04625628888607025,
"learning_rate": 0.0008,
"loss": 2.138144016265869,
"step": 7878
},
{
"epoch": 0.4775757575757576,
"grad_norm": 0.05648026615381241,
"learning_rate": 0.0008,
"loss": 2.1396164894104004,
"step": 7880
},
{
"epoch": 0.4776969696969697,
"grad_norm": 0.04969198256731033,
"learning_rate": 0.0008,
"loss": 2.1569714546203613,
"step": 7882
},
{
"epoch": 0.4778181818181818,
"grad_norm": 0.04833647236227989,
"learning_rate": 0.0008,
"loss": 2.1282126903533936,
"step": 7884
},
{
"epoch": 0.47793939393939394,
"grad_norm": 0.04670590162277222,
"learning_rate": 0.0008,
"loss": 2.112946033477783,
"step": 7886
},
{
"epoch": 0.4780606060606061,
"grad_norm": 0.04878044128417969,
"learning_rate": 0.0008,
"loss": 2.125394344329834,
"step": 7888
},
{
"epoch": 0.4781818181818182,
"grad_norm": 0.047552332282066345,
"learning_rate": 0.0008,
"loss": 2.0626561641693115,
"step": 7890
},
{
"epoch": 0.4783030303030303,
"grad_norm": 0.04979480430483818,
"learning_rate": 0.0008,
"loss": 2.2040255069732666,
"step": 7892
},
{
"epoch": 0.4784242424242424,
"grad_norm": 0.051708392798900604,
"learning_rate": 0.0008,
"loss": 2.099243640899658,
"step": 7894
},
{
"epoch": 0.47854545454545455,
"grad_norm": 0.05382946878671646,
"learning_rate": 0.0008,
"loss": 2.160111904144287,
"step": 7896
},
{
"epoch": 0.4786666666666667,
"grad_norm": 0.047110654413700104,
"learning_rate": 0.0008,
"loss": 2.154275417327881,
"step": 7898
},
{
"epoch": 0.47878787878787876,
"grad_norm": 0.046473823487758636,
"learning_rate": 0.0008,
"loss": 2.183300256729126,
"step": 7900
},
{
"epoch": 0.4789090909090909,
"grad_norm": 0.045401014387607574,
"learning_rate": 0.0008,
"loss": 2.096517562866211,
"step": 7902
},
{
"epoch": 0.47903030303030303,
"grad_norm": 0.04684897139668465,
"learning_rate": 0.0008,
"loss": 2.091083288192749,
"step": 7904
},
{
"epoch": 0.47915151515151516,
"grad_norm": 0.048833683133125305,
"learning_rate": 0.0008,
"loss": 2.160156488418579,
"step": 7906
},
{
"epoch": 0.4792727272727273,
"grad_norm": 0.048005953431129456,
"learning_rate": 0.0008,
"loss": 2.2454333305358887,
"step": 7908
},
{
"epoch": 0.4793939393939394,
"grad_norm": 0.046593666076660156,
"learning_rate": 0.0008,
"loss": 2.179857015609741,
"step": 7910
},
{
"epoch": 0.4795151515151515,
"grad_norm": 0.04776585474610329,
"learning_rate": 0.0008,
"loss": 2.114001512527466,
"step": 7912
},
{
"epoch": 0.47963636363636364,
"grad_norm": 0.04525863751769066,
"learning_rate": 0.0008,
"loss": 2.0724501609802246,
"step": 7914
},
{
"epoch": 0.47975757575757577,
"grad_norm": 0.06028628721833229,
"learning_rate": 0.0008,
"loss": 2.1502861976623535,
"step": 7916
},
{
"epoch": 0.4798787878787879,
"grad_norm": 0.06398729234933853,
"learning_rate": 0.0008,
"loss": 2.212368965148926,
"step": 7918
},
{
"epoch": 0.48,
"grad_norm": 0.06953177601099014,
"learning_rate": 0.0008,
"loss": 2.1673521995544434,
"step": 7920
},
{
"epoch": 0.4801212121212121,
"grad_norm": 0.05715607479214668,
"learning_rate": 0.0008,
"loss": 2.0598833560943604,
"step": 7922
},
{
"epoch": 0.48024242424242425,
"grad_norm": 0.04697805643081665,
"learning_rate": 0.0008,
"loss": 2.1252596378326416,
"step": 7924
},
{
"epoch": 0.4803636363636364,
"grad_norm": 0.04683684557676315,
"learning_rate": 0.0008,
"loss": 2.1027157306671143,
"step": 7926
},
{
"epoch": 0.48048484848484846,
"grad_norm": 0.049796536564826965,
"learning_rate": 0.0008,
"loss": 2.193052053451538,
"step": 7928
},
{
"epoch": 0.4806060606060606,
"grad_norm": 0.047527506947517395,
"learning_rate": 0.0008,
"loss": 2.0814361572265625,
"step": 7930
},
{
"epoch": 0.4807272727272727,
"grad_norm": 0.04576047137379646,
"learning_rate": 0.0008,
"loss": 2.1170480251312256,
"step": 7932
},
{
"epoch": 0.48084848484848486,
"grad_norm": 0.04638442024588585,
"learning_rate": 0.0008,
"loss": 2.056560754776001,
"step": 7934
},
{
"epoch": 0.480969696969697,
"grad_norm": 0.048837676644325256,
"learning_rate": 0.0008,
"loss": 2.152461051940918,
"step": 7936
},
{
"epoch": 0.48109090909090907,
"grad_norm": 0.0549081452190876,
"learning_rate": 0.0008,
"loss": 2.173325300216675,
"step": 7938
},
{
"epoch": 0.4812121212121212,
"grad_norm": 0.055167924612760544,
"learning_rate": 0.0008,
"loss": 2.1231517791748047,
"step": 7940
},
{
"epoch": 0.48133333333333334,
"grad_norm": 0.05610305815935135,
"learning_rate": 0.0008,
"loss": 2.1356730461120605,
"step": 7942
},
{
"epoch": 0.48145454545454547,
"grad_norm": 0.05421973019838333,
"learning_rate": 0.0008,
"loss": 2.120123863220215,
"step": 7944
},
{
"epoch": 0.4815757575757576,
"grad_norm": 0.04841269925236702,
"learning_rate": 0.0008,
"loss": 2.123396396636963,
"step": 7946
},
{
"epoch": 0.4816969696969697,
"grad_norm": 0.04628857225179672,
"learning_rate": 0.0008,
"loss": 2.1597063541412354,
"step": 7948
},
{
"epoch": 0.4818181818181818,
"grad_norm": 0.04735277220606804,
"learning_rate": 0.0008,
"loss": 2.162797689437866,
"step": 7950
},
{
"epoch": 0.48193939393939395,
"grad_norm": 0.04584914818406105,
"learning_rate": 0.0008,
"loss": 2.133094549179077,
"step": 7952
},
{
"epoch": 0.4820606060606061,
"grad_norm": 0.0480254702270031,
"learning_rate": 0.0008,
"loss": 2.2479312419891357,
"step": 7954
},
{
"epoch": 0.48218181818181816,
"grad_norm": 0.04556788131594658,
"learning_rate": 0.0008,
"loss": 2.1938912868499756,
"step": 7956
},
{
"epoch": 0.4823030303030303,
"grad_norm": 0.04473787546157837,
"learning_rate": 0.0008,
"loss": 2.087282657623291,
"step": 7958
},
{
"epoch": 0.4824242424242424,
"grad_norm": 0.04727059602737427,
"learning_rate": 0.0008,
"loss": 2.1773219108581543,
"step": 7960
},
{
"epoch": 0.48254545454545456,
"grad_norm": 0.0471804216504097,
"learning_rate": 0.0008,
"loss": 2.2470622062683105,
"step": 7962
},
{
"epoch": 0.4826666666666667,
"grad_norm": 0.0463046133518219,
"learning_rate": 0.0008,
"loss": 2.182440757751465,
"step": 7964
},
{
"epoch": 0.48278787878787877,
"grad_norm": 0.04832308739423752,
"learning_rate": 0.0008,
"loss": 2.1898748874664307,
"step": 7966
},
{
"epoch": 0.4829090909090909,
"grad_norm": 0.04742458462715149,
"learning_rate": 0.0008,
"loss": 2.0950472354888916,
"step": 7968
},
{
"epoch": 0.48303030303030303,
"grad_norm": 0.04693712666630745,
"learning_rate": 0.0008,
"loss": 2.1353418827056885,
"step": 7970
},
{
"epoch": 0.48315151515151517,
"grad_norm": 0.04560433328151703,
"learning_rate": 0.0008,
"loss": 2.100820541381836,
"step": 7972
},
{
"epoch": 0.4832727272727273,
"grad_norm": 0.04588646441698074,
"learning_rate": 0.0008,
"loss": 2.17008113861084,
"step": 7974
},
{
"epoch": 0.4833939393939394,
"grad_norm": 0.05156596004962921,
"learning_rate": 0.0008,
"loss": 2.180776596069336,
"step": 7976
},
{
"epoch": 0.4835151515151515,
"grad_norm": 0.05276881158351898,
"learning_rate": 0.0008,
"loss": 2.1485612392425537,
"step": 7978
},
{
"epoch": 0.48363636363636364,
"grad_norm": 0.05519138649106026,
"learning_rate": 0.0008,
"loss": 2.197990894317627,
"step": 7980
},
{
"epoch": 0.4837575757575758,
"grad_norm": 0.049400582909584045,
"learning_rate": 0.0008,
"loss": 2.1576011180877686,
"step": 7982
},
{
"epoch": 0.48387878787878785,
"grad_norm": 0.052837494760751724,
"learning_rate": 0.0008,
"loss": 2.1409919261932373,
"step": 7984
},
{
"epoch": 0.484,
"grad_norm": 0.058008115738630295,
"learning_rate": 0.0008,
"loss": 2.1156907081604004,
"step": 7986
},
{
"epoch": 0.4841212121212121,
"grad_norm": 0.0614786334335804,
"learning_rate": 0.0008,
"loss": 2.1968023777008057,
"step": 7988
},
{
"epoch": 0.48424242424242425,
"grad_norm": 0.05824822932481766,
"learning_rate": 0.0008,
"loss": 2.092435121536255,
"step": 7990
},
{
"epoch": 0.4843636363636364,
"grad_norm": 0.060352906584739685,
"learning_rate": 0.0008,
"loss": 2.1590638160705566,
"step": 7992
},
{
"epoch": 0.48448484848484846,
"grad_norm": 0.05581668019294739,
"learning_rate": 0.0008,
"loss": 2.1576030254364014,
"step": 7994
},
{
"epoch": 0.4846060606060606,
"grad_norm": 0.05193927884101868,
"learning_rate": 0.0008,
"loss": 2.14442777633667,
"step": 7996
},
{
"epoch": 0.48472727272727273,
"grad_norm": 0.048746995627880096,
"learning_rate": 0.0008,
"loss": 2.1728620529174805,
"step": 7998
},
{
"epoch": 0.48484848484848486,
"grad_norm": 0.045741356909275055,
"learning_rate": 0.0008,
"loss": 2.079671859741211,
"step": 8000
},
{
"epoch": 0.484969696969697,
"grad_norm": 0.04755518585443497,
"learning_rate": 0.0008,
"loss": 2.1981284618377686,
"step": 8002
},
{
"epoch": 0.4850909090909091,
"grad_norm": 0.05018129572272301,
"learning_rate": 0.0008,
"loss": 2.2168023586273193,
"step": 8004
},
{
"epoch": 0.4852121212121212,
"grad_norm": 0.04779853671789169,
"learning_rate": 0.0008,
"loss": 2.1361589431762695,
"step": 8006
},
{
"epoch": 0.48533333333333334,
"grad_norm": 0.04702983424067497,
"learning_rate": 0.0008,
"loss": 2.1300199031829834,
"step": 8008
},
{
"epoch": 0.48545454545454547,
"grad_norm": 0.05032796785235405,
"learning_rate": 0.0008,
"loss": 2.0925590991973877,
"step": 8010
},
{
"epoch": 0.48557575757575755,
"grad_norm": 0.04834022745490074,
"learning_rate": 0.0008,
"loss": 2.2065885066986084,
"step": 8012
},
{
"epoch": 0.4856969696969697,
"grad_norm": 0.048035845160484314,
"learning_rate": 0.0008,
"loss": 2.1623475551605225,
"step": 8014
},
{
"epoch": 0.4858181818181818,
"grad_norm": 0.0492490753531456,
"learning_rate": 0.0008,
"loss": 2.11659574508667,
"step": 8016
},
{
"epoch": 0.48593939393939395,
"grad_norm": 0.050389207899570465,
"learning_rate": 0.0008,
"loss": 2.1764402389526367,
"step": 8018
},
{
"epoch": 0.4860606060606061,
"grad_norm": 0.04670332744717598,
"learning_rate": 0.0008,
"loss": 2.195909261703491,
"step": 8020
},
{
"epoch": 0.48618181818181816,
"grad_norm": 0.047224149107933044,
"learning_rate": 0.0008,
"loss": 2.1715056896209717,
"step": 8022
},
{
"epoch": 0.4863030303030303,
"grad_norm": 0.05065590888261795,
"learning_rate": 0.0008,
"loss": 2.1839547157287598,
"step": 8024
},
{
"epoch": 0.4864242424242424,
"grad_norm": 0.04880532622337341,
"learning_rate": 0.0008,
"loss": 2.1532344818115234,
"step": 8026
},
{
"epoch": 0.48654545454545456,
"grad_norm": 0.04704447090625763,
"learning_rate": 0.0008,
"loss": 2.1869595050811768,
"step": 8028
},
{
"epoch": 0.4866666666666667,
"grad_norm": 0.0473904050886631,
"learning_rate": 0.0008,
"loss": 2.1304478645324707,
"step": 8030
},
{
"epoch": 0.48678787878787877,
"grad_norm": 0.048481885343790054,
"learning_rate": 0.0008,
"loss": 2.1630771160125732,
"step": 8032
},
{
"epoch": 0.4869090909090909,
"grad_norm": 0.046491678804159164,
"learning_rate": 0.0008,
"loss": 2.20939040184021,
"step": 8034
},
{
"epoch": 0.48703030303030304,
"grad_norm": 0.04588380455970764,
"learning_rate": 0.0008,
"loss": 2.191132068634033,
"step": 8036
},
{
"epoch": 0.48715151515151517,
"grad_norm": 0.04920245334506035,
"learning_rate": 0.0008,
"loss": 2.1452081203460693,
"step": 8038
},
{
"epoch": 0.48727272727272725,
"grad_norm": 0.06036463379859924,
"learning_rate": 0.0008,
"loss": 2.125612258911133,
"step": 8040
},
{
"epoch": 0.4873939393939394,
"grad_norm": 0.06885825097560883,
"learning_rate": 0.0008,
"loss": 2.14691424369812,
"step": 8042
},
{
"epoch": 0.4875151515151515,
"grad_norm": 0.06516905128955841,
"learning_rate": 0.0008,
"loss": 2.1530356407165527,
"step": 8044
},
{
"epoch": 0.48763636363636365,
"grad_norm": 0.05860385298728943,
"learning_rate": 0.0008,
"loss": 2.147707939147949,
"step": 8046
},
{
"epoch": 0.4877575757575758,
"grad_norm": 0.059222131967544556,
"learning_rate": 0.0008,
"loss": 2.194840431213379,
"step": 8048
},
{
"epoch": 0.48787878787878786,
"grad_norm": 0.0511152520775795,
"learning_rate": 0.0008,
"loss": 2.204620838165283,
"step": 8050
},
{
"epoch": 0.488,
"grad_norm": 0.04738345369696617,
"learning_rate": 0.0008,
"loss": 2.167820453643799,
"step": 8052
},
{
"epoch": 0.4881212121212121,
"grad_norm": 0.04710760712623596,
"learning_rate": 0.0008,
"loss": 2.168243646621704,
"step": 8054
},
{
"epoch": 0.48824242424242426,
"grad_norm": 0.05177466198801994,
"learning_rate": 0.0008,
"loss": 2.1572201251983643,
"step": 8056
},
{
"epoch": 0.4883636363636364,
"grad_norm": 0.04869115352630615,
"learning_rate": 0.0008,
"loss": 2.075375556945801,
"step": 8058
},
{
"epoch": 0.48848484848484847,
"grad_norm": 0.05025269463658333,
"learning_rate": 0.0008,
"loss": 2.1828317642211914,
"step": 8060
},
{
"epoch": 0.4886060606060606,
"grad_norm": 0.04819400608539581,
"learning_rate": 0.0008,
"loss": 2.2058727741241455,
"step": 8062
},
{
"epoch": 0.48872727272727273,
"grad_norm": 0.04893394559621811,
"learning_rate": 0.0008,
"loss": 2.250335454940796,
"step": 8064
},
{
"epoch": 0.48884848484848487,
"grad_norm": 0.050121236592531204,
"learning_rate": 0.0008,
"loss": 2.2132654190063477,
"step": 8066
},
{
"epoch": 0.48896969696969694,
"grad_norm": 0.0483601875603199,
"learning_rate": 0.0008,
"loss": 2.1698122024536133,
"step": 8068
},
{
"epoch": 0.4890909090909091,
"grad_norm": 0.04785027354955673,
"learning_rate": 0.0008,
"loss": 2.1873414516448975,
"step": 8070
},
{
"epoch": 0.4892121212121212,
"grad_norm": 0.04718153923749924,
"learning_rate": 0.0008,
"loss": 2.2780964374542236,
"step": 8072
},
{
"epoch": 0.48933333333333334,
"grad_norm": 0.04798334464430809,
"learning_rate": 0.0008,
"loss": 2.101529598236084,
"step": 8074
},
{
"epoch": 0.4894545454545455,
"grad_norm": 0.04611857235431671,
"learning_rate": 0.0008,
"loss": 2.167292833328247,
"step": 8076
},
{
"epoch": 0.48957575757575755,
"grad_norm": 0.045211050659418106,
"learning_rate": 0.0008,
"loss": 2.1878604888916016,
"step": 8078
},
{
"epoch": 0.4896969696969697,
"grad_norm": 0.04678706079721451,
"learning_rate": 0.0008,
"loss": 2.160085439682007,
"step": 8080
},
{
"epoch": 0.4898181818181818,
"grad_norm": 0.04661682993173599,
"learning_rate": 0.0008,
"loss": 2.1523172855377197,
"step": 8082
},
{
"epoch": 0.48993939393939395,
"grad_norm": 0.0516045056283474,
"learning_rate": 0.0008,
"loss": 2.1481988430023193,
"step": 8084
},
{
"epoch": 0.4900606060606061,
"grad_norm": 0.049808651208877563,
"learning_rate": 0.0008,
"loss": 2.088008165359497,
"step": 8086
},
{
"epoch": 0.49018181818181816,
"grad_norm": 0.06430309265851974,
"learning_rate": 0.0008,
"loss": 2.2354788780212402,
"step": 8088
},
{
"epoch": 0.4903030303030303,
"grad_norm": 0.06837473809719086,
"learning_rate": 0.0008,
"loss": 2.2027931213378906,
"step": 8090
},
{
"epoch": 0.49042424242424243,
"grad_norm": 0.04776355251669884,
"learning_rate": 0.0008,
"loss": 2.2230489253997803,
"step": 8092
},
{
"epoch": 0.49054545454545456,
"grad_norm": 0.050362974405288696,
"learning_rate": 0.0008,
"loss": 2.1656877994537354,
"step": 8094
},
{
"epoch": 0.49066666666666664,
"grad_norm": 0.050989679992198944,
"learning_rate": 0.0008,
"loss": 2.111717939376831,
"step": 8096
},
{
"epoch": 0.4907878787878788,
"grad_norm": 0.047923292964696884,
"learning_rate": 0.0008,
"loss": 2.100940227508545,
"step": 8098
},
{
"epoch": 0.4909090909090909,
"grad_norm": 0.04554542526602745,
"learning_rate": 0.0008,
"loss": 2.149564743041992,
"step": 8100
},
{
"epoch": 0.49103030303030304,
"grad_norm": 0.04567817226052284,
"learning_rate": 0.0008,
"loss": 2.1128101348876953,
"step": 8102
},
{
"epoch": 0.4911515151515152,
"grad_norm": 0.046744536608457565,
"learning_rate": 0.0008,
"loss": 2.2363498210906982,
"step": 8104
},
{
"epoch": 0.49127272727272725,
"grad_norm": 0.046018172055482864,
"learning_rate": 0.0008,
"loss": 2.1862754821777344,
"step": 8106
},
{
"epoch": 0.4913939393939394,
"grad_norm": 0.048875629901885986,
"learning_rate": 0.0008,
"loss": 2.177825927734375,
"step": 8108
},
{
"epoch": 0.4915151515151515,
"grad_norm": 0.0472327396273613,
"learning_rate": 0.0008,
"loss": 2.130298376083374,
"step": 8110
},
{
"epoch": 0.49163636363636365,
"grad_norm": 0.04676814749836922,
"learning_rate": 0.0008,
"loss": 2.1980721950531006,
"step": 8112
},
{
"epoch": 0.4917575757575758,
"grad_norm": 0.04516500607132912,
"learning_rate": 0.0008,
"loss": 2.1468393802642822,
"step": 8114
},
{
"epoch": 0.49187878787878786,
"grad_norm": 0.04904119297862053,
"learning_rate": 0.0008,
"loss": 2.1239147186279297,
"step": 8116
},
{
"epoch": 0.492,
"grad_norm": 0.0500326082110405,
"learning_rate": 0.0008,
"loss": 2.138835906982422,
"step": 8118
},
{
"epoch": 0.4921212121212121,
"grad_norm": 0.05328955873847008,
"learning_rate": 0.0008,
"loss": 2.1513001918792725,
"step": 8120
},
{
"epoch": 0.49224242424242426,
"grad_norm": 0.05241120606660843,
"learning_rate": 0.0008,
"loss": 2.232006788253784,
"step": 8122
},
{
"epoch": 0.49236363636363634,
"grad_norm": 0.04845290631055832,
"learning_rate": 0.0008,
"loss": 2.1838676929473877,
"step": 8124
},
{
"epoch": 0.49248484848484847,
"grad_norm": 0.053707510232925415,
"learning_rate": 0.0008,
"loss": 2.1833393573760986,
"step": 8126
},
{
"epoch": 0.4926060606060606,
"grad_norm": 0.05237961933016777,
"learning_rate": 0.0008,
"loss": 2.142829418182373,
"step": 8128
},
{
"epoch": 0.49272727272727274,
"grad_norm": 0.05672919750213623,
"learning_rate": 0.0008,
"loss": 2.0982344150543213,
"step": 8130
},
{
"epoch": 0.49284848484848487,
"grad_norm": 0.06768239289522171,
"learning_rate": 0.0008,
"loss": 2.1772124767303467,
"step": 8132
},
{
"epoch": 0.49296969696969695,
"grad_norm": 0.06101243570446968,
"learning_rate": 0.0008,
"loss": 2.1232857704162598,
"step": 8134
},
{
"epoch": 0.4930909090909091,
"grad_norm": 0.04699873924255371,
"learning_rate": 0.0008,
"loss": 2.0971736907958984,
"step": 8136
},
{
"epoch": 0.4932121212121212,
"grad_norm": 0.04717868193984032,
"learning_rate": 0.0008,
"loss": 2.128848075866699,
"step": 8138
},
{
"epoch": 0.49333333333333335,
"grad_norm": 0.04809015616774559,
"learning_rate": 0.0008,
"loss": 2.2503280639648438,
"step": 8140
},
{
"epoch": 0.4934545454545455,
"grad_norm": 0.0454794205725193,
"learning_rate": 0.0008,
"loss": 2.215733766555786,
"step": 8142
},
{
"epoch": 0.49357575757575756,
"grad_norm": 0.045128341764211655,
"learning_rate": 0.0008,
"loss": 2.1557118892669678,
"step": 8144
},
{
"epoch": 0.4936969696969697,
"grad_norm": 0.046841610223054886,
"learning_rate": 0.0008,
"loss": 2.1252994537353516,
"step": 8146
},
{
"epoch": 0.4938181818181818,
"grad_norm": 0.04633751139044762,
"learning_rate": 0.0008,
"loss": 2.205721855163574,
"step": 8148
},
{
"epoch": 0.49393939393939396,
"grad_norm": 0.04508954659104347,
"learning_rate": 0.0008,
"loss": 2.2272467613220215,
"step": 8150
},
{
"epoch": 0.49406060606060603,
"grad_norm": 0.047399669885635376,
"learning_rate": 0.0008,
"loss": 2.1831653118133545,
"step": 8152
},
{
"epoch": 0.49418181818181817,
"grad_norm": 0.04831470921635628,
"learning_rate": 0.0008,
"loss": 2.1266767978668213,
"step": 8154
},
{
"epoch": 0.4943030303030303,
"grad_norm": 0.04635467007756233,
"learning_rate": 0.0008,
"loss": 2.1595776081085205,
"step": 8156
},
{
"epoch": 0.49442424242424243,
"grad_norm": 0.048109591007232666,
"learning_rate": 0.0008,
"loss": 2.1448564529418945,
"step": 8158
},
{
"epoch": 0.49454545454545457,
"grad_norm": 0.04712845757603645,
"learning_rate": 0.0008,
"loss": 2.2030115127563477,
"step": 8160
},
{
"epoch": 0.49466666666666664,
"grad_norm": 0.048840463161468506,
"learning_rate": 0.0008,
"loss": 2.1227118968963623,
"step": 8162
},
{
"epoch": 0.4947878787878788,
"grad_norm": 0.045379240065813065,
"learning_rate": 0.0008,
"loss": 2.1675143241882324,
"step": 8164
},
{
"epoch": 0.4949090909090909,
"grad_norm": 0.04535629227757454,
"learning_rate": 0.0008,
"loss": 2.135972023010254,
"step": 8166
},
{
"epoch": 0.49503030303030304,
"grad_norm": 0.0460146889090538,
"learning_rate": 0.0008,
"loss": 2.1254985332489014,
"step": 8168
},
{
"epoch": 0.4951515151515152,
"grad_norm": 0.052786629647016525,
"learning_rate": 0.0008,
"loss": 2.1938858032226562,
"step": 8170
},
{
"epoch": 0.49527272727272725,
"grad_norm": 0.05849597603082657,
"learning_rate": 0.0008,
"loss": 2.1848714351654053,
"step": 8172
},
{
"epoch": 0.4953939393939394,
"grad_norm": 0.05627073347568512,
"learning_rate": 0.0008,
"loss": 2.177468776702881,
"step": 8174
},
{
"epoch": 0.4955151515151515,
"grad_norm": 0.05701513588428497,
"learning_rate": 0.0008,
"loss": 2.138036012649536,
"step": 8176
},
{
"epoch": 0.49563636363636365,
"grad_norm": 0.04620877653360367,
"learning_rate": 0.0008,
"loss": 2.140601396560669,
"step": 8178
},
{
"epoch": 0.49575757575757573,
"grad_norm": 0.04556256905198097,
"learning_rate": 0.0008,
"loss": 2.147451877593994,
"step": 8180
},
{
"epoch": 0.49587878787878786,
"grad_norm": 0.044421665370464325,
"learning_rate": 0.0008,
"loss": 2.1688339710235596,
"step": 8182
},
{
"epoch": 0.496,
"grad_norm": 0.04711424186825752,
"learning_rate": 0.0008,
"loss": 2.1958963871002197,
"step": 8184
},
{
"epoch": 0.49612121212121213,
"grad_norm": 0.045987579971551895,
"learning_rate": 0.0008,
"loss": 2.186655282974243,
"step": 8186
},
{
"epoch": 0.49624242424242426,
"grad_norm": 0.04943123832345009,
"learning_rate": 0.0008,
"loss": 2.1217451095581055,
"step": 8188
},
{
"epoch": 0.49636363636363634,
"grad_norm": 0.04740197956562042,
"learning_rate": 0.0008,
"loss": 2.191524028778076,
"step": 8190
},
{
"epoch": 0.4964848484848485,
"grad_norm": 0.04600048065185547,
"learning_rate": 0.0008,
"loss": 2.187910795211792,
"step": 8192
},
{
"epoch": 0.4966060606060606,
"grad_norm": 0.049969155341386795,
"learning_rate": 0.0008,
"loss": 2.153129816055298,
"step": 8194
},
{
"epoch": 0.49672727272727274,
"grad_norm": 0.04763882979750633,
"learning_rate": 0.0008,
"loss": 2.1623642444610596,
"step": 8196
},
{
"epoch": 0.4968484848484849,
"grad_norm": 0.04595251753926277,
"learning_rate": 0.0008,
"loss": 2.185467004776001,
"step": 8198
},
{
"epoch": 0.49696969696969695,
"grad_norm": 0.052502721548080444,
"learning_rate": 0.0008,
"loss": 2.156458616256714,
"step": 8200
},
{
"epoch": 0.4970909090909091,
"grad_norm": 0.05706549808382988,
"learning_rate": 0.0008,
"loss": 2.2539138793945312,
"step": 8202
},
{
"epoch": 0.4972121212121212,
"grad_norm": 0.054811831563711166,
"learning_rate": 0.0008,
"loss": 2.2606303691864014,
"step": 8204
},
{
"epoch": 0.49733333333333335,
"grad_norm": 0.05520012602210045,
"learning_rate": 0.0008,
"loss": 2.0689024925231934,
"step": 8206
},
{
"epoch": 0.4974545454545454,
"grad_norm": 0.04858216643333435,
"learning_rate": 0.0008,
"loss": 2.1212246417999268,
"step": 8208
},
{
"epoch": 0.49757575757575756,
"grad_norm": 0.054647352546453476,
"learning_rate": 0.0008,
"loss": 2.1998848915100098,
"step": 8210
},
{
"epoch": 0.4976969696969697,
"grad_norm": 0.04862206056714058,
"learning_rate": 0.0008,
"loss": 2.192903518676758,
"step": 8212
},
{
"epoch": 0.4978181818181818,
"grad_norm": 0.053280871361494064,
"learning_rate": 0.0008,
"loss": 2.1320762634277344,
"step": 8214
},
{
"epoch": 0.49793939393939396,
"grad_norm": 0.04882730543613434,
"learning_rate": 0.0008,
"loss": 2.1743366718292236,
"step": 8216
},
{
"epoch": 0.49806060606060604,
"grad_norm": 0.045933403074741364,
"learning_rate": 0.0008,
"loss": 2.170347213745117,
"step": 8218
},
{
"epoch": 0.49818181818181817,
"grad_norm": 0.045763734728097916,
"learning_rate": 0.0008,
"loss": 2.146291971206665,
"step": 8220
},
{
"epoch": 0.4983030303030303,
"grad_norm": 0.04588675871491432,
"learning_rate": 0.0008,
"loss": 2.128345489501953,
"step": 8222
},
{
"epoch": 0.49842424242424244,
"grad_norm": 0.04707220941781998,
"learning_rate": 0.0008,
"loss": 2.1400463581085205,
"step": 8224
},
{
"epoch": 0.49854545454545457,
"grad_norm": 0.04686984419822693,
"learning_rate": 0.0008,
"loss": 2.129045248031616,
"step": 8226
},
{
"epoch": 0.49866666666666665,
"grad_norm": 0.04595641791820526,
"learning_rate": 0.0008,
"loss": 2.133805274963379,
"step": 8228
},
{
"epoch": 0.4987878787878788,
"grad_norm": 0.04642193019390106,
"learning_rate": 0.0008,
"loss": 2.1419248580932617,
"step": 8230
},
{
"epoch": 0.4989090909090909,
"grad_norm": 0.04766688868403435,
"learning_rate": 0.0008,
"loss": 2.163240909576416,
"step": 8232
},
{
"epoch": 0.49903030303030305,
"grad_norm": 0.04558701068162918,
"learning_rate": 0.0008,
"loss": 2.1784377098083496,
"step": 8234
},
{
"epoch": 0.4991515151515151,
"grad_norm": 0.0485723577439785,
"learning_rate": 0.0008,
"loss": 2.188000440597534,
"step": 8236
},
{
"epoch": 0.49927272727272726,
"grad_norm": 0.050375428050756454,
"learning_rate": 0.0008,
"loss": 2.1444485187530518,
"step": 8238
},
{
"epoch": 0.4993939393939394,
"grad_norm": 0.0477910079061985,
"learning_rate": 0.0008,
"loss": 2.0946812629699707,
"step": 8240
},
{
"epoch": 0.4995151515151515,
"grad_norm": 0.049106497317552567,
"learning_rate": 0.0008,
"loss": 2.163905143737793,
"step": 8242
},
{
"epoch": 0.49963636363636366,
"grad_norm": 0.04727230221033096,
"learning_rate": 0.0008,
"loss": 2.0582895278930664,
"step": 8244
},
{
"epoch": 0.49975757575757573,
"grad_norm": 0.04703858494758606,
"learning_rate": 0.0008,
"loss": 2.1509952545166016,
"step": 8246
},
{
"epoch": 0.49987878787878787,
"grad_norm": 0.04940459504723549,
"learning_rate": 0.0008,
"loss": 2.1422626972198486,
"step": 8248
},
{
"epoch": 0.5,
"grad_norm": 0.04618372395634651,
"learning_rate": 0.0008,
"loss": 2.126713752746582,
"step": 8250
},
{
"epoch": 0.5001212121212121,
"grad_norm": 0.04753788933157921,
"learning_rate": 0.0008,
"loss": 2.0747878551483154,
"step": 8252
},
{
"epoch": 0.5002424242424243,
"grad_norm": 0.04675167798995972,
"learning_rate": 0.0008,
"loss": 2.2127225399017334,
"step": 8254
},
{
"epoch": 0.5003636363636363,
"grad_norm": 0.046687301248311996,
"learning_rate": 0.0008,
"loss": 2.1687121391296387,
"step": 8256
},
{
"epoch": 0.5004848484848485,
"grad_norm": 0.05368051305413246,
"learning_rate": 0.0008,
"loss": 2.1707875728607178,
"step": 8258
},
{
"epoch": 0.5006060606060606,
"grad_norm": 0.0520273819565773,
"learning_rate": 0.0008,
"loss": 2.1292724609375,
"step": 8260
},
{
"epoch": 0.5007272727272727,
"grad_norm": 0.050026532262563705,
"learning_rate": 0.0008,
"loss": 2.1943674087524414,
"step": 8262
},
{
"epoch": 0.5008484848484849,
"grad_norm": 0.052951693534851074,
"learning_rate": 0.0008,
"loss": 2.148890256881714,
"step": 8264
},
{
"epoch": 0.500969696969697,
"grad_norm": 0.057563286274671555,
"learning_rate": 0.0008,
"loss": 2.1275806427001953,
"step": 8266
},
{
"epoch": 0.5010909090909091,
"grad_norm": 0.06776952743530273,
"learning_rate": 0.0008,
"loss": 2.157639503479004,
"step": 8268
},
{
"epoch": 0.5012121212121212,
"grad_norm": 0.05548914521932602,
"learning_rate": 0.0008,
"loss": 2.1497583389282227,
"step": 8270
},
{
"epoch": 0.5013333333333333,
"grad_norm": 0.06335851550102234,
"learning_rate": 0.0008,
"loss": 2.122023105621338,
"step": 8272
},
{
"epoch": 0.5014545454545455,
"grad_norm": 0.06312880665063858,
"learning_rate": 0.0008,
"loss": 2.1484220027923584,
"step": 8274
},
{
"epoch": 0.5015757575757576,
"grad_norm": 0.049535755068063736,
"learning_rate": 0.0008,
"loss": 2.166900634765625,
"step": 8276
},
{
"epoch": 0.5016969696969698,
"grad_norm": 0.0472804494202137,
"learning_rate": 0.0008,
"loss": 2.1180500984191895,
"step": 8278
},
{
"epoch": 0.5018181818181818,
"grad_norm": 0.04706694558262825,
"learning_rate": 0.0008,
"loss": 2.1438212394714355,
"step": 8280
},
{
"epoch": 0.5019393939393939,
"grad_norm": 0.046209972351789474,
"learning_rate": 0.0008,
"loss": 2.1080408096313477,
"step": 8282
},
{
"epoch": 0.5020606060606061,
"grad_norm": 0.045565180480480194,
"learning_rate": 0.0008,
"loss": 2.2248342037200928,
"step": 8284
},
{
"epoch": 0.5021818181818182,
"grad_norm": 0.04442087188363075,
"learning_rate": 0.0008,
"loss": 2.1944468021392822,
"step": 8286
},
{
"epoch": 0.5023030303030303,
"grad_norm": 0.04569323733448982,
"learning_rate": 0.0008,
"loss": 2.1491992473602295,
"step": 8288
},
{
"epoch": 0.5024242424242424,
"grad_norm": 0.04603752866387367,
"learning_rate": 0.0008,
"loss": 2.1619515419006348,
"step": 8290
},
{
"epoch": 0.5025454545454545,
"grad_norm": 0.045195575803518295,
"learning_rate": 0.0008,
"loss": 2.137420177459717,
"step": 8292
},
{
"epoch": 0.5026666666666667,
"grad_norm": 0.04743070527911186,
"learning_rate": 0.0008,
"loss": 2.1941967010498047,
"step": 8294
},
{
"epoch": 0.5027878787878788,
"grad_norm": 0.04482589289546013,
"learning_rate": 0.0008,
"loss": 2.1450107097625732,
"step": 8296
},
{
"epoch": 0.5029090909090909,
"grad_norm": 0.04417955502867699,
"learning_rate": 0.0008,
"loss": 2.0869414806365967,
"step": 8298
},
{
"epoch": 0.503030303030303,
"grad_norm": 0.04728525131940842,
"learning_rate": 0.0008,
"loss": 2.1566834449768066,
"step": 8300
},
{
"epoch": 0.5031515151515151,
"grad_norm": 0.04557187482714653,
"learning_rate": 0.0008,
"loss": 2.138300657272339,
"step": 8302
},
{
"epoch": 0.5032727272727273,
"grad_norm": 0.04922821745276451,
"learning_rate": 0.0008,
"loss": 2.156038999557495,
"step": 8304
},
{
"epoch": 0.5033939393939394,
"grad_norm": 0.04671449959278107,
"learning_rate": 0.0008,
"loss": 2.2207746505737305,
"step": 8306
},
{
"epoch": 0.5035151515151515,
"grad_norm": 0.0474519282579422,
"learning_rate": 0.0008,
"loss": 2.220104217529297,
"step": 8308
},
{
"epoch": 0.5036363636363637,
"grad_norm": 0.046981893479824066,
"learning_rate": 0.0008,
"loss": 2.129133939743042,
"step": 8310
},
{
"epoch": 0.5037575757575757,
"grad_norm": 0.04772515967488289,
"learning_rate": 0.0008,
"loss": 2.1690475940704346,
"step": 8312
},
{
"epoch": 0.5038787878787879,
"grad_norm": 0.04585590958595276,
"learning_rate": 0.0008,
"loss": 2.136798620223999,
"step": 8314
},
{
"epoch": 0.504,
"grad_norm": 0.04713081195950508,
"learning_rate": 0.0008,
"loss": 2.16054105758667,
"step": 8316
},
{
"epoch": 0.5041212121212121,
"grad_norm": 0.04808669537305832,
"learning_rate": 0.0008,
"loss": 2.2000579833984375,
"step": 8318
},
{
"epoch": 0.5042424242424243,
"grad_norm": 0.05108082294464111,
"learning_rate": 0.0008,
"loss": 2.1388683319091797,
"step": 8320
},
{
"epoch": 0.5043636363636363,
"grad_norm": 0.054240282624959946,
"learning_rate": 0.0008,
"loss": 2.0851006507873535,
"step": 8322
},
{
"epoch": 0.5044848484848485,
"grad_norm": 0.046349018812179565,
"learning_rate": 0.0008,
"loss": 2.1698405742645264,
"step": 8324
},
{
"epoch": 0.5046060606060606,
"grad_norm": 0.04833495616912842,
"learning_rate": 0.0008,
"loss": 2.1538515090942383,
"step": 8326
},
{
"epoch": 0.5047272727272727,
"grad_norm": 0.05482051521539688,
"learning_rate": 0.0008,
"loss": 2.148196220397949,
"step": 8328
},
{
"epoch": 0.5048484848484849,
"grad_norm": 0.04909079521894455,
"learning_rate": 0.0008,
"loss": 2.0988783836364746,
"step": 8330
},
{
"epoch": 0.504969696969697,
"grad_norm": 0.05174148082733154,
"learning_rate": 0.0008,
"loss": 2.189291000366211,
"step": 8332
},
{
"epoch": 0.5050909090909091,
"grad_norm": 0.05886248126626015,
"learning_rate": 0.0008,
"loss": 2.1858556270599365,
"step": 8334
},
{
"epoch": 0.5052121212121212,
"grad_norm": 0.05439787730574608,
"learning_rate": 0.0008,
"loss": 2.1881017684936523,
"step": 8336
},
{
"epoch": 0.5053333333333333,
"grad_norm": 0.06349267065525055,
"learning_rate": 0.0008,
"loss": 2.135181188583374,
"step": 8338
},
{
"epoch": 0.5054545454545455,
"grad_norm": 0.05695410445332527,
"learning_rate": 0.0008,
"loss": 2.1856887340545654,
"step": 8340
},
{
"epoch": 0.5055757575757576,
"grad_norm": 0.05183325707912445,
"learning_rate": 0.0008,
"loss": 2.1277432441711426,
"step": 8342
},
{
"epoch": 0.5056969696969696,
"grad_norm": 0.055816568434238434,
"learning_rate": 0.0008,
"loss": 2.199383497238159,
"step": 8344
},
{
"epoch": 0.5058181818181818,
"grad_norm": 0.05534963309764862,
"learning_rate": 0.0008,
"loss": 2.1434249877929688,
"step": 8346
},
{
"epoch": 0.5059393939393939,
"grad_norm": 0.051075492054224014,
"learning_rate": 0.0008,
"loss": 2.2040655612945557,
"step": 8348
},
{
"epoch": 0.5060606060606061,
"grad_norm": 0.04613843932747841,
"learning_rate": 0.0008,
"loss": 2.173184871673584,
"step": 8350
},
{
"epoch": 0.5061818181818182,
"grad_norm": 0.04540320485830307,
"learning_rate": 0.0008,
"loss": 2.069232940673828,
"step": 8352
},
{
"epoch": 0.5063030303030303,
"grad_norm": 0.04711904749274254,
"learning_rate": 0.0008,
"loss": 2.1974644660949707,
"step": 8354
},
{
"epoch": 0.5064242424242424,
"grad_norm": 0.04584847018122673,
"learning_rate": 0.0008,
"loss": 2.125601291656494,
"step": 8356
},
{
"epoch": 0.5065454545454545,
"grad_norm": 0.04686010256409645,
"learning_rate": 0.0008,
"loss": 2.131800651550293,
"step": 8358
},
{
"epoch": 0.5066666666666667,
"grad_norm": 0.0487171933054924,
"learning_rate": 0.0008,
"loss": 2.2304675579071045,
"step": 8360
},
{
"epoch": 0.5067878787878788,
"grad_norm": 0.05850938335061073,
"learning_rate": 0.0008,
"loss": 2.122990369796753,
"step": 8362
},
{
"epoch": 0.5069090909090909,
"grad_norm": 0.052166882902383804,
"learning_rate": 0.0008,
"loss": 2.1749730110168457,
"step": 8364
},
{
"epoch": 0.507030303030303,
"grad_norm": 0.04772840440273285,
"learning_rate": 0.0008,
"loss": 2.193629503250122,
"step": 8366
},
{
"epoch": 0.5071515151515151,
"grad_norm": 0.04916509613394737,
"learning_rate": 0.0008,
"loss": 2.203782081604004,
"step": 8368
},
{
"epoch": 0.5072727272727273,
"grad_norm": 0.04665343090891838,
"learning_rate": 0.0008,
"loss": 2.1856369972229004,
"step": 8370
},
{
"epoch": 0.5073939393939394,
"grad_norm": 0.047499917447566986,
"learning_rate": 0.0008,
"loss": 2.1621389389038086,
"step": 8372
},
{
"epoch": 0.5075151515151515,
"grad_norm": 0.047155335545539856,
"learning_rate": 0.0008,
"loss": 2.1648263931274414,
"step": 8374
},
{
"epoch": 0.5076363636363637,
"grad_norm": 0.04917951300740242,
"learning_rate": 0.0008,
"loss": 2.160984992980957,
"step": 8376
},
{
"epoch": 0.5077575757575757,
"grad_norm": 0.05143747106194496,
"learning_rate": 0.0008,
"loss": 2.1361734867095947,
"step": 8378
},
{
"epoch": 0.5078787878787879,
"grad_norm": 0.04854264110326767,
"learning_rate": 0.0008,
"loss": 2.162557601928711,
"step": 8380
},
{
"epoch": 0.508,
"grad_norm": 0.048763252794742584,
"learning_rate": 0.0008,
"loss": 2.0927484035491943,
"step": 8382
},
{
"epoch": 0.5081212121212121,
"grad_norm": 0.045139893889427185,
"learning_rate": 0.0008,
"loss": 2.223177433013916,
"step": 8384
},
{
"epoch": 0.5082424242424243,
"grad_norm": 0.046081844717264175,
"learning_rate": 0.0008,
"loss": 2.1395139694213867,
"step": 8386
},
{
"epoch": 0.5083636363636364,
"grad_norm": 0.04526335746049881,
"learning_rate": 0.0008,
"loss": 2.1028833389282227,
"step": 8388
},
{
"epoch": 0.5084848484848485,
"grad_norm": 0.04607897251844406,
"learning_rate": 0.0008,
"loss": 2.160555839538574,
"step": 8390
},
{
"epoch": 0.5086060606060606,
"grad_norm": 0.046668194234371185,
"learning_rate": 0.0008,
"loss": 2.1911563873291016,
"step": 8392
},
{
"epoch": 0.5087272727272727,
"grad_norm": 0.046017199754714966,
"learning_rate": 0.0008,
"loss": 2.1131012439727783,
"step": 8394
},
{
"epoch": 0.5088484848484849,
"grad_norm": 0.04757123813033104,
"learning_rate": 0.0008,
"loss": 2.181053638458252,
"step": 8396
},
{
"epoch": 0.508969696969697,
"grad_norm": 0.05007690191268921,
"learning_rate": 0.0008,
"loss": 2.1062660217285156,
"step": 8398
},
{
"epoch": 0.509090909090909,
"grad_norm": 0.05519925057888031,
"learning_rate": 0.0008,
"loss": 2.216257333755493,
"step": 8400
},
{
"epoch": 0.5092121212121212,
"grad_norm": 0.04499412328004837,
"learning_rate": 0.0008,
"loss": 2.134782075881958,
"step": 8402
},
{
"epoch": 0.5093333333333333,
"grad_norm": 0.047390181571245193,
"learning_rate": 0.0008,
"loss": 2.1485912799835205,
"step": 8404
},
{
"epoch": 0.5094545454545455,
"grad_norm": 0.04757282882928848,
"learning_rate": 0.0008,
"loss": 2.182154893875122,
"step": 8406
},
{
"epoch": 0.5095757575757576,
"grad_norm": 0.04614488035440445,
"learning_rate": 0.0008,
"loss": 2.200502872467041,
"step": 8408
},
{
"epoch": 0.5096969696969696,
"grad_norm": 0.052768465131521225,
"learning_rate": 0.0008,
"loss": 2.172635793685913,
"step": 8410
},
{
"epoch": 0.5098181818181818,
"grad_norm": 0.054344117641448975,
"learning_rate": 0.0008,
"loss": 2.120060443878174,
"step": 8412
},
{
"epoch": 0.5099393939393939,
"grad_norm": 0.052818186581134796,
"learning_rate": 0.0008,
"loss": 2.1689207553863525,
"step": 8414
},
{
"epoch": 0.5100606060606061,
"grad_norm": 0.04738256335258484,
"learning_rate": 0.0008,
"loss": 2.1231932640075684,
"step": 8416
},
{
"epoch": 0.5101818181818182,
"grad_norm": 0.04628175124526024,
"learning_rate": 0.0008,
"loss": 2.0969176292419434,
"step": 8418
},
{
"epoch": 0.5103030303030303,
"grad_norm": 0.0444316640496254,
"learning_rate": 0.0008,
"loss": 2.0992579460144043,
"step": 8420
},
{
"epoch": 0.5104242424242424,
"grad_norm": 0.04930891841650009,
"learning_rate": 0.0008,
"loss": 2.1576640605926514,
"step": 8422
},
{
"epoch": 0.5105454545454545,
"grad_norm": 0.05840415507555008,
"learning_rate": 0.0008,
"loss": 2.1917362213134766,
"step": 8424
},
{
"epoch": 0.5106666666666667,
"grad_norm": 0.053547557443380356,
"learning_rate": 0.0008,
"loss": 2.2495715618133545,
"step": 8426
},
{
"epoch": 0.5107878787878788,
"grad_norm": 0.04541819915175438,
"learning_rate": 0.0008,
"loss": 2.2255442142486572,
"step": 8428
},
{
"epoch": 0.5109090909090909,
"grad_norm": 0.04512858763337135,
"learning_rate": 0.0008,
"loss": 2.188781976699829,
"step": 8430
},
{
"epoch": 0.5110303030303031,
"grad_norm": 0.046277862042188644,
"learning_rate": 0.0008,
"loss": 2.2492177486419678,
"step": 8432
},
{
"epoch": 0.5111515151515151,
"grad_norm": 0.04600614681839943,
"learning_rate": 0.0008,
"loss": 2.1495940685272217,
"step": 8434
},
{
"epoch": 0.5112727272727273,
"grad_norm": 0.04568784311413765,
"learning_rate": 0.0008,
"loss": 2.173769950866699,
"step": 8436
},
{
"epoch": 0.5113939393939394,
"grad_norm": 0.052489541471004486,
"learning_rate": 0.0008,
"loss": 2.1669960021972656,
"step": 8438
},
{
"epoch": 0.5115151515151515,
"grad_norm": 0.044003888964653015,
"learning_rate": 0.0008,
"loss": 2.0443081855773926,
"step": 8440
},
{
"epoch": 0.5116363636363637,
"grad_norm": 0.045515984296798706,
"learning_rate": 0.0008,
"loss": 2.1745829582214355,
"step": 8442
},
{
"epoch": 0.5117575757575757,
"grad_norm": 0.053701017051935196,
"learning_rate": 0.0008,
"loss": 2.11906099319458,
"step": 8444
},
{
"epoch": 0.5118787878787879,
"grad_norm": 0.05366675928235054,
"learning_rate": 0.0008,
"loss": 2.1765103340148926,
"step": 8446
},
{
"epoch": 0.512,
"grad_norm": 0.0498976968228817,
"learning_rate": 0.0008,
"loss": 2.2087323665618896,
"step": 8448
},
{
"epoch": 0.5121212121212121,
"grad_norm": 0.04663627967238426,
"learning_rate": 0.0008,
"loss": 2.172914981842041,
"step": 8450
},
{
"epoch": 0.5122424242424243,
"grad_norm": 0.05596553906798363,
"learning_rate": 0.0008,
"loss": 2.0802512168884277,
"step": 8452
},
{
"epoch": 0.5123636363636364,
"grad_norm": 0.0593256838619709,
"learning_rate": 0.0008,
"loss": 2.158299684524536,
"step": 8454
},
{
"epoch": 0.5124848484848484,
"grad_norm": 0.05325794219970703,
"learning_rate": 0.0008,
"loss": 2.188397169113159,
"step": 8456
},
{
"epoch": 0.5126060606060606,
"grad_norm": 0.049928318709135056,
"learning_rate": 0.0008,
"loss": 2.234545946121216,
"step": 8458
},
{
"epoch": 0.5127272727272727,
"grad_norm": 0.04948364943265915,
"learning_rate": 0.0008,
"loss": 2.198573589324951,
"step": 8460
},
{
"epoch": 0.5128484848484849,
"grad_norm": 0.04699978604912758,
"learning_rate": 0.0008,
"loss": 2.18436861038208,
"step": 8462
},
{
"epoch": 0.512969696969697,
"grad_norm": 0.0454506017267704,
"learning_rate": 0.0008,
"loss": 2.1416680812835693,
"step": 8464
},
{
"epoch": 0.513090909090909,
"grad_norm": 0.04455338791012764,
"learning_rate": 0.0008,
"loss": 2.0984182357788086,
"step": 8466
},
{
"epoch": 0.5132121212121212,
"grad_norm": 0.04663264751434326,
"learning_rate": 0.0008,
"loss": 2.1529643535614014,
"step": 8468
},
{
"epoch": 0.5133333333333333,
"grad_norm": 0.04517247900366783,
"learning_rate": 0.0008,
"loss": 2.2263691425323486,
"step": 8470
},
{
"epoch": 0.5134545454545455,
"grad_norm": 0.05333007127046585,
"learning_rate": 0.0008,
"loss": 2.108067274093628,
"step": 8472
},
{
"epoch": 0.5135757575757576,
"grad_norm": 0.05820532515645027,
"learning_rate": 0.0008,
"loss": 2.1061363220214844,
"step": 8474
},
{
"epoch": 0.5136969696969697,
"grad_norm": 0.04732801020145416,
"learning_rate": 0.0008,
"loss": 2.1895670890808105,
"step": 8476
},
{
"epoch": 0.5138181818181818,
"grad_norm": 0.04764959588646889,
"learning_rate": 0.0008,
"loss": 2.116126537322998,
"step": 8478
},
{
"epoch": 0.5139393939393939,
"grad_norm": 0.05399865657091141,
"learning_rate": 0.0008,
"loss": 2.1652774810791016,
"step": 8480
},
{
"epoch": 0.5140606060606061,
"grad_norm": 0.05341809242963791,
"learning_rate": 0.0008,
"loss": 2.187131881713867,
"step": 8482
},
{
"epoch": 0.5141818181818182,
"grad_norm": 0.0535990335047245,
"learning_rate": 0.0008,
"loss": 2.177508592605591,
"step": 8484
},
{
"epoch": 0.5143030303030303,
"grad_norm": 0.04730105400085449,
"learning_rate": 0.0008,
"loss": 2.117436647415161,
"step": 8486
},
{
"epoch": 0.5144242424242425,
"grad_norm": 0.047370634973049164,
"learning_rate": 0.0008,
"loss": 2.181685209274292,
"step": 8488
},
{
"epoch": 0.5145454545454545,
"grad_norm": 0.048775266855955124,
"learning_rate": 0.0008,
"loss": 2.177309274673462,
"step": 8490
},
{
"epoch": 0.5146666666666667,
"grad_norm": 0.04493803158402443,
"learning_rate": 0.0008,
"loss": 2.160421133041382,
"step": 8492
},
{
"epoch": 0.5147878787878788,
"grad_norm": 0.04507630690932274,
"learning_rate": 0.0008,
"loss": 2.1621315479278564,
"step": 8494
},
{
"epoch": 0.5149090909090909,
"grad_norm": 0.04609949141740799,
"learning_rate": 0.0008,
"loss": 2.1323323249816895,
"step": 8496
},
{
"epoch": 0.5150303030303031,
"grad_norm": 0.050645023584365845,
"learning_rate": 0.0008,
"loss": 2.0798287391662598,
"step": 8498
},
{
"epoch": 0.5151515151515151,
"grad_norm": 0.04513709619641304,
"learning_rate": 0.0008,
"loss": 2.163363456726074,
"step": 8500
},
{
"epoch": 0.5152727272727273,
"grad_norm": 0.04790177568793297,
"learning_rate": 0.0008,
"loss": 2.121560573577881,
"step": 8502
},
{
"epoch": 0.5153939393939394,
"grad_norm": 0.045462410897016525,
"learning_rate": 0.0008,
"loss": 2.1058592796325684,
"step": 8504
},
{
"epoch": 0.5155151515151515,
"grad_norm": 0.05221061408519745,
"learning_rate": 0.0008,
"loss": 2.2150495052337646,
"step": 8506
},
{
"epoch": 0.5156363636363637,
"grad_norm": 0.045630890876054764,
"learning_rate": 0.0008,
"loss": 2.1593387126922607,
"step": 8508
},
{
"epoch": 0.5157575757575757,
"grad_norm": 0.04750077426433563,
"learning_rate": 0.0008,
"loss": 2.1736526489257812,
"step": 8510
},
{
"epoch": 0.5158787878787878,
"grad_norm": 0.045027174055576324,
"learning_rate": 0.0008,
"loss": 2.177401542663574,
"step": 8512
},
{
"epoch": 0.516,
"grad_norm": 0.057939112186431885,
"learning_rate": 0.0008,
"loss": 2.1279850006103516,
"step": 8514
},
{
"epoch": 0.5161212121212121,
"grad_norm": 0.05399900674819946,
"learning_rate": 0.0008,
"loss": 2.1284899711608887,
"step": 8516
},
{
"epoch": 0.5162424242424243,
"grad_norm": 0.059826627373695374,
"learning_rate": 0.0008,
"loss": 2.1333189010620117,
"step": 8518
},
{
"epoch": 0.5163636363636364,
"grad_norm": 0.07290604710578918,
"learning_rate": 0.0008,
"loss": 2.133570671081543,
"step": 8520
},
{
"epoch": 0.5164848484848484,
"grad_norm": 0.05476842448115349,
"learning_rate": 0.0008,
"loss": 2.19606351852417,
"step": 8522
},
{
"epoch": 0.5166060606060606,
"grad_norm": 0.05136002600193024,
"learning_rate": 0.0008,
"loss": 2.14558482170105,
"step": 8524
},
{
"epoch": 0.5167272727272727,
"grad_norm": 0.06669149547815323,
"learning_rate": 0.0008,
"loss": 2.1894819736480713,
"step": 8526
},
{
"epoch": 0.5168484848484849,
"grad_norm": 0.06364791840314865,
"learning_rate": 0.0008,
"loss": 2.193425178527832,
"step": 8528
},
{
"epoch": 0.516969696969697,
"grad_norm": 0.062456633895635605,
"learning_rate": 0.0008,
"loss": 2.154658317565918,
"step": 8530
},
{
"epoch": 0.517090909090909,
"grad_norm": 0.06306169927120209,
"learning_rate": 0.0008,
"loss": 2.1670215129852295,
"step": 8532
},
{
"epoch": 0.5172121212121212,
"grad_norm": 0.05986541137099266,
"learning_rate": 0.0008,
"loss": 2.0977962017059326,
"step": 8534
},
{
"epoch": 0.5173333333333333,
"grad_norm": 0.0536867119371891,
"learning_rate": 0.0008,
"loss": 2.1175646781921387,
"step": 8536
},
{
"epoch": 0.5174545454545455,
"grad_norm": 0.06026748567819595,
"learning_rate": 0.0008,
"loss": 2.1687917709350586,
"step": 8538
},
{
"epoch": 0.5175757575757576,
"grad_norm": 0.059326834976673126,
"learning_rate": 0.0008,
"loss": 2.1408584117889404,
"step": 8540
},
{
"epoch": 0.5176969696969697,
"grad_norm": 0.05075798183679581,
"learning_rate": 0.0008,
"loss": 2.175293445587158,
"step": 8542
},
{
"epoch": 0.5178181818181818,
"grad_norm": 0.04999053105711937,
"learning_rate": 0.0008,
"loss": 2.1908087730407715,
"step": 8544
},
{
"epoch": 0.5179393939393939,
"grad_norm": 0.044635429978370667,
"learning_rate": 0.0008,
"loss": 2.1118810176849365,
"step": 8546
},
{
"epoch": 0.5180606060606061,
"grad_norm": 0.04613838344812393,
"learning_rate": 0.0008,
"loss": 2.2061879634857178,
"step": 8548
},
{
"epoch": 0.5181818181818182,
"grad_norm": 0.04996630921959877,
"learning_rate": 0.0008,
"loss": 2.1704750061035156,
"step": 8550
},
{
"epoch": 0.5183030303030303,
"grad_norm": 0.04486759752035141,
"learning_rate": 0.0008,
"loss": 2.1862940788269043,
"step": 8552
},
{
"epoch": 0.5184242424242425,
"grad_norm": 0.04621780663728714,
"learning_rate": 0.0008,
"loss": 2.24623966217041,
"step": 8554
},
{
"epoch": 0.5185454545454545,
"grad_norm": 0.05570010468363762,
"learning_rate": 0.0008,
"loss": 2.110278844833374,
"step": 8556
},
{
"epoch": 0.5186666666666667,
"grad_norm": 0.06315582245588303,
"learning_rate": 0.0008,
"loss": 2.166139602661133,
"step": 8558
},
{
"epoch": 0.5187878787878788,
"grad_norm": 0.0669611245393753,
"learning_rate": 0.0008,
"loss": 2.124897003173828,
"step": 8560
},
{
"epoch": 0.5189090909090909,
"grad_norm": 0.0786876454949379,
"learning_rate": 0.0008,
"loss": 2.1735262870788574,
"step": 8562
},
{
"epoch": 0.5190303030303031,
"grad_norm": 0.08753079175949097,
"learning_rate": 0.0008,
"loss": 2.1843996047973633,
"step": 8564
},
{
"epoch": 0.5191515151515151,
"grad_norm": 0.06962619721889496,
"learning_rate": 0.0008,
"loss": 2.151451826095581,
"step": 8566
},
{
"epoch": 0.5192727272727272,
"grad_norm": 0.06650211662054062,
"learning_rate": 0.0008,
"loss": 2.114278554916382,
"step": 8568
},
{
"epoch": 0.5193939393939394,
"grad_norm": 0.054574355483055115,
"learning_rate": 0.0008,
"loss": 2.1254260540008545,
"step": 8570
},
{
"epoch": 0.5195151515151515,
"grad_norm": 0.05264878273010254,
"learning_rate": 0.0008,
"loss": 2.2087299823760986,
"step": 8572
},
{
"epoch": 0.5196363636363637,
"grad_norm": 0.05008414015173912,
"learning_rate": 0.0008,
"loss": 2.1561832427978516,
"step": 8574
},
{
"epoch": 0.5197575757575758,
"grad_norm": 0.046195924282073975,
"learning_rate": 0.0008,
"loss": 2.1468000411987305,
"step": 8576
},
{
"epoch": 0.5198787878787878,
"grad_norm": 0.046108946204185486,
"learning_rate": 0.0008,
"loss": 2.1146886348724365,
"step": 8578
},
{
"epoch": 0.52,
"grad_norm": 0.047432564198970795,
"learning_rate": 0.0008,
"loss": 2.1746394634246826,
"step": 8580
},
{
"epoch": 0.5201212121212121,
"grad_norm": 0.04539509862661362,
"learning_rate": 0.0008,
"loss": 2.225393772125244,
"step": 8582
},
{
"epoch": 0.5202424242424243,
"grad_norm": 0.0449226014316082,
"learning_rate": 0.0008,
"loss": 2.1708765029907227,
"step": 8584
},
{
"epoch": 0.5203636363636364,
"grad_norm": 0.044190652668476105,
"learning_rate": 0.0008,
"loss": 2.1315791606903076,
"step": 8586
},
{
"epoch": 0.5204848484848484,
"grad_norm": 0.04661904647946358,
"learning_rate": 0.0008,
"loss": 2.193117380142212,
"step": 8588
},
{
"epoch": 0.5206060606060606,
"grad_norm": 0.04620092734694481,
"learning_rate": 0.0008,
"loss": 2.1691153049468994,
"step": 8590
},
{
"epoch": 0.5207272727272727,
"grad_norm": 0.045683689415454865,
"learning_rate": 0.0008,
"loss": 2.0833261013031006,
"step": 8592
},
{
"epoch": 0.5208484848484849,
"grad_norm": 0.04585098475217819,
"learning_rate": 0.0008,
"loss": 2.1540794372558594,
"step": 8594
},
{
"epoch": 0.520969696969697,
"grad_norm": 0.047720082104206085,
"learning_rate": 0.0008,
"loss": 2.17844820022583,
"step": 8596
},
{
"epoch": 0.521090909090909,
"grad_norm": 0.04936341568827629,
"learning_rate": 0.0008,
"loss": 2.1814894676208496,
"step": 8598
},
{
"epoch": 0.5212121212121212,
"grad_norm": 0.048220645636320114,
"learning_rate": 0.0008,
"loss": 2.127140522003174,
"step": 8600
},
{
"epoch": 0.5213333333333333,
"grad_norm": 0.0472579263150692,
"learning_rate": 0.0008,
"loss": 2.1229825019836426,
"step": 8602
},
{
"epoch": 0.5214545454545455,
"grad_norm": 0.04390188306570053,
"learning_rate": 0.0008,
"loss": 2.1254186630249023,
"step": 8604
},
{
"epoch": 0.5215757575757576,
"grad_norm": 0.04476621747016907,
"learning_rate": 0.0008,
"loss": 2.1118087768554688,
"step": 8606
},
{
"epoch": 0.5216969696969697,
"grad_norm": 0.045748285949230194,
"learning_rate": 0.0008,
"loss": 2.1635637283325195,
"step": 8608
},
{
"epoch": 0.5218181818181818,
"grad_norm": 0.046533409506082535,
"learning_rate": 0.0008,
"loss": 2.1834588050842285,
"step": 8610
},
{
"epoch": 0.5219393939393939,
"grad_norm": 0.053135115653276443,
"learning_rate": 0.0008,
"loss": 2.240097999572754,
"step": 8612
},
{
"epoch": 0.5220606060606061,
"grad_norm": 0.05270520970225334,
"learning_rate": 0.0008,
"loss": 2.1147916316986084,
"step": 8614
},
{
"epoch": 0.5221818181818182,
"grad_norm": 0.0487690232694149,
"learning_rate": 0.0008,
"loss": 2.146523952484131,
"step": 8616
},
{
"epoch": 0.5223030303030303,
"grad_norm": 0.04681255295872688,
"learning_rate": 0.0008,
"loss": 2.1656246185302734,
"step": 8618
},
{
"epoch": 0.5224242424242425,
"grad_norm": 0.049087826162576675,
"learning_rate": 0.0008,
"loss": 2.187962055206299,
"step": 8620
},
{
"epoch": 0.5225454545454545,
"grad_norm": 0.05348682031035423,
"learning_rate": 0.0008,
"loss": 2.2151730060577393,
"step": 8622
},
{
"epoch": 0.5226666666666666,
"grad_norm": 0.04622306674718857,
"learning_rate": 0.0008,
"loss": 2.173006057739258,
"step": 8624
},
{
"epoch": 0.5227878787878788,
"grad_norm": 0.04601255804300308,
"learning_rate": 0.0008,
"loss": 2.216400146484375,
"step": 8626
},
{
"epoch": 0.5229090909090909,
"grad_norm": 0.046211037784814835,
"learning_rate": 0.0008,
"loss": 2.1282763481140137,
"step": 8628
},
{
"epoch": 0.5230303030303031,
"grad_norm": 0.04699113592505455,
"learning_rate": 0.0008,
"loss": 2.2003555297851562,
"step": 8630
},
{
"epoch": 0.5231515151515151,
"grad_norm": 0.048487551510334015,
"learning_rate": 0.0008,
"loss": 2.148366928100586,
"step": 8632
},
{
"epoch": 0.5232727272727272,
"grad_norm": 0.047762077301740646,
"learning_rate": 0.0008,
"loss": 2.126533269882202,
"step": 8634
},
{
"epoch": 0.5233939393939394,
"grad_norm": 0.055928051471710205,
"learning_rate": 0.0008,
"loss": 2.1211905479431152,
"step": 8636
},
{
"epoch": 0.5235151515151515,
"grad_norm": 0.053136032074689865,
"learning_rate": 0.0008,
"loss": 2.1787593364715576,
"step": 8638
},
{
"epoch": 0.5236363636363637,
"grad_norm": 0.06587613373994827,
"learning_rate": 0.0008,
"loss": 2.150348663330078,
"step": 8640
},
{
"epoch": 0.5237575757575758,
"grad_norm": 0.06883400678634644,
"learning_rate": 0.0008,
"loss": 2.1492419242858887,
"step": 8642
},
{
"epoch": 0.5238787878787878,
"grad_norm": 0.05619535222649574,
"learning_rate": 0.0008,
"loss": 2.1583967208862305,
"step": 8644
},
{
"epoch": 0.524,
"grad_norm": 0.04664282500743866,
"learning_rate": 0.0008,
"loss": 2.1727635860443115,
"step": 8646
},
{
"epoch": 0.5241212121212121,
"grad_norm": 0.04457813501358032,
"learning_rate": 0.0008,
"loss": 2.117239475250244,
"step": 8648
},
{
"epoch": 0.5242424242424243,
"grad_norm": 0.045408573001623154,
"learning_rate": 0.0008,
"loss": 2.1837689876556396,
"step": 8650
},
{
"epoch": 0.5243636363636364,
"grad_norm": 0.04601973295211792,
"learning_rate": 0.0008,
"loss": 2.1479737758636475,
"step": 8652
},
{
"epoch": 0.5244848484848484,
"grad_norm": 0.04619995877146721,
"learning_rate": 0.0008,
"loss": 2.160665988922119,
"step": 8654
},
{
"epoch": 0.5246060606060606,
"grad_norm": 0.04568478465080261,
"learning_rate": 0.0008,
"loss": 2.1632237434387207,
"step": 8656
},
{
"epoch": 0.5247272727272727,
"grad_norm": 0.0454363189637661,
"learning_rate": 0.0008,
"loss": 2.2191905975341797,
"step": 8658
},
{
"epoch": 0.5248484848484849,
"grad_norm": 0.045400504022836685,
"learning_rate": 0.0008,
"loss": 2.194936752319336,
"step": 8660
},
{
"epoch": 0.524969696969697,
"grad_norm": 0.04602816700935364,
"learning_rate": 0.0008,
"loss": 2.191707134246826,
"step": 8662
},
{
"epoch": 0.525090909090909,
"grad_norm": 0.04758608341217041,
"learning_rate": 0.0008,
"loss": 2.1289024353027344,
"step": 8664
},
{
"epoch": 0.5252121212121212,
"grad_norm": 0.051739491522312164,
"learning_rate": 0.0008,
"loss": 2.1386172771453857,
"step": 8666
},
{
"epoch": 0.5253333333333333,
"grad_norm": 0.05346198379993439,
"learning_rate": 0.0008,
"loss": 2.1896023750305176,
"step": 8668
},
{
"epoch": 0.5254545454545455,
"grad_norm": 0.05432479456067085,
"learning_rate": 0.0008,
"loss": 2.140143871307373,
"step": 8670
},
{
"epoch": 0.5255757575757576,
"grad_norm": 0.05434626713395119,
"learning_rate": 0.0008,
"loss": 2.176504611968994,
"step": 8672
},
{
"epoch": 0.5256969696969697,
"grad_norm": 0.05408091098070145,
"learning_rate": 0.0008,
"loss": 2.186129570007324,
"step": 8674
},
{
"epoch": 0.5258181818181819,
"grad_norm": 0.0464819073677063,
"learning_rate": 0.0008,
"loss": 2.2182095050811768,
"step": 8676
},
{
"epoch": 0.5259393939393939,
"grad_norm": 0.051499031484127045,
"learning_rate": 0.0008,
"loss": 2.1956238746643066,
"step": 8678
},
{
"epoch": 0.526060606060606,
"grad_norm": 0.0540907196700573,
"learning_rate": 0.0008,
"loss": 2.1950438022613525,
"step": 8680
},
{
"epoch": 0.5261818181818182,
"grad_norm": 0.0464242622256279,
"learning_rate": 0.0008,
"loss": 2.1734321117401123,
"step": 8682
},
{
"epoch": 0.5263030303030303,
"grad_norm": 0.047390636056661606,
"learning_rate": 0.0008,
"loss": 2.1411497592926025,
"step": 8684
},
{
"epoch": 0.5264242424242425,
"grad_norm": 0.04947170987725258,
"learning_rate": 0.0008,
"loss": 2.1326913833618164,
"step": 8686
},
{
"epoch": 0.5265454545454545,
"grad_norm": 0.04789343103766441,
"learning_rate": 0.0008,
"loss": 2.200929641723633,
"step": 8688
},
{
"epoch": 0.5266666666666666,
"grad_norm": 0.049071185290813446,
"learning_rate": 0.0008,
"loss": 2.1352365016937256,
"step": 8690
},
{
"epoch": 0.5267878787878788,
"grad_norm": 0.05098288133740425,
"learning_rate": 0.0008,
"loss": 2.09945011138916,
"step": 8692
},
{
"epoch": 0.5269090909090909,
"grad_norm": 0.04962518438696861,
"learning_rate": 0.0008,
"loss": 2.184572458267212,
"step": 8694
},
{
"epoch": 0.5270303030303031,
"grad_norm": 0.046823885291814804,
"learning_rate": 0.0008,
"loss": 2.1044015884399414,
"step": 8696
},
{
"epoch": 0.5271515151515151,
"grad_norm": 0.04716826230287552,
"learning_rate": 0.0008,
"loss": 2.125134229660034,
"step": 8698
},
{
"epoch": 0.5272727272727272,
"grad_norm": 0.05268719419836998,
"learning_rate": 0.0008,
"loss": 2.1537535190582275,
"step": 8700
},
{
"epoch": 0.5273939393939394,
"grad_norm": 0.05879351496696472,
"learning_rate": 0.0008,
"loss": 2.1584339141845703,
"step": 8702
},
{
"epoch": 0.5275151515151515,
"grad_norm": 0.044291574507951736,
"learning_rate": 0.0008,
"loss": 2.106015205383301,
"step": 8704
},
{
"epoch": 0.5276363636363637,
"grad_norm": 0.05333657190203667,
"learning_rate": 0.0008,
"loss": 2.1595706939697266,
"step": 8706
},
{
"epoch": 0.5277575757575758,
"grad_norm": 0.046298861503601074,
"learning_rate": 0.0008,
"loss": 2.135265827178955,
"step": 8708
},
{
"epoch": 0.5278787878787878,
"grad_norm": 0.04891844838857651,
"learning_rate": 0.0008,
"loss": 2.1753838062286377,
"step": 8710
},
{
"epoch": 0.528,
"grad_norm": 0.04929834604263306,
"learning_rate": 0.0008,
"loss": 2.217649459838867,
"step": 8712
},
{
"epoch": 0.5281212121212121,
"grad_norm": 0.04454046115279198,
"learning_rate": 0.0008,
"loss": 2.1148681640625,
"step": 8714
},
{
"epoch": 0.5282424242424243,
"grad_norm": 0.05580737069249153,
"learning_rate": 0.0008,
"loss": 2.073085308074951,
"step": 8716
},
{
"epoch": 0.5283636363636364,
"grad_norm": 0.04711587354540825,
"learning_rate": 0.0008,
"loss": 2.193030834197998,
"step": 8718
},
{
"epoch": 0.5284848484848484,
"grad_norm": 0.04549514874815941,
"learning_rate": 0.0008,
"loss": 2.185446262359619,
"step": 8720
},
{
"epoch": 0.5286060606060606,
"grad_norm": 0.048432450741529465,
"learning_rate": 0.0008,
"loss": 2.1270334720611572,
"step": 8722
},
{
"epoch": 0.5287272727272727,
"grad_norm": 0.0474231019616127,
"learning_rate": 0.0008,
"loss": 2.1626033782958984,
"step": 8724
},
{
"epoch": 0.5288484848484849,
"grad_norm": 0.047864291816949844,
"learning_rate": 0.0008,
"loss": 2.242114305496216,
"step": 8726
},
{
"epoch": 0.528969696969697,
"grad_norm": 0.04878392443060875,
"learning_rate": 0.0008,
"loss": 2.116807222366333,
"step": 8728
},
{
"epoch": 0.5290909090909091,
"grad_norm": 0.05730767548084259,
"learning_rate": 0.0008,
"loss": 2.15409517288208,
"step": 8730
},
{
"epoch": 0.5292121212121212,
"grad_norm": 0.04608892276883125,
"learning_rate": 0.0008,
"loss": 2.164428949356079,
"step": 8732
},
{
"epoch": 0.5293333333333333,
"grad_norm": 0.046024344861507416,
"learning_rate": 0.0008,
"loss": 2.1836047172546387,
"step": 8734
},
{
"epoch": 0.5294545454545454,
"grad_norm": 0.04849056154489517,
"learning_rate": 0.0008,
"loss": 2.1660633087158203,
"step": 8736
},
{
"epoch": 0.5295757575757576,
"grad_norm": 0.04650299623608589,
"learning_rate": 0.0008,
"loss": 2.1669390201568604,
"step": 8738
},
{
"epoch": 0.5296969696969697,
"grad_norm": 0.04644303768873215,
"learning_rate": 0.0008,
"loss": 2.1236822605133057,
"step": 8740
},
{
"epoch": 0.5298181818181819,
"grad_norm": 0.048778679221868515,
"learning_rate": 0.0008,
"loss": 2.158953905105591,
"step": 8742
},
{
"epoch": 0.5299393939393939,
"grad_norm": 0.04868125915527344,
"learning_rate": 0.0008,
"loss": 2.151559352874756,
"step": 8744
},
{
"epoch": 0.530060606060606,
"grad_norm": 0.04691349342465401,
"learning_rate": 0.0008,
"loss": 2.099644422531128,
"step": 8746
},
{
"epoch": 0.5301818181818182,
"grad_norm": 0.04618412256240845,
"learning_rate": 0.0008,
"loss": 2.2123451232910156,
"step": 8748
},
{
"epoch": 0.5303030303030303,
"grad_norm": 0.046770308166742325,
"learning_rate": 0.0008,
"loss": 2.176469087600708,
"step": 8750
},
{
"epoch": 0.5304242424242425,
"grad_norm": 0.04693131893873215,
"learning_rate": 0.0008,
"loss": 2.17159366607666,
"step": 8752
},
{
"epoch": 0.5305454545454545,
"grad_norm": 0.04776100814342499,
"learning_rate": 0.0008,
"loss": 2.145416259765625,
"step": 8754
},
{
"epoch": 0.5306666666666666,
"grad_norm": 0.047048501670360565,
"learning_rate": 0.0008,
"loss": 2.170748710632324,
"step": 8756
},
{
"epoch": 0.5307878787878788,
"grad_norm": 0.04620970040559769,
"learning_rate": 0.0008,
"loss": 2.132807493209839,
"step": 8758
},
{
"epoch": 0.5309090909090909,
"grad_norm": 0.04703725874423981,
"learning_rate": 0.0008,
"loss": 2.158921003341675,
"step": 8760
},
{
"epoch": 0.5310303030303031,
"grad_norm": 0.04579084366559982,
"learning_rate": 0.0008,
"loss": 2.074666976928711,
"step": 8762
},
{
"epoch": 0.5311515151515152,
"grad_norm": 0.04763205721974373,
"learning_rate": 0.0008,
"loss": 2.2122974395751953,
"step": 8764
},
{
"epoch": 0.5312727272727272,
"grad_norm": 0.05401817709207535,
"learning_rate": 0.0008,
"loss": 2.180164337158203,
"step": 8766
},
{
"epoch": 0.5313939393939394,
"grad_norm": 0.051120173186063766,
"learning_rate": 0.0008,
"loss": 2.162339448928833,
"step": 8768
},
{
"epoch": 0.5315151515151515,
"grad_norm": 0.046210139989852905,
"learning_rate": 0.0008,
"loss": 2.098797082901001,
"step": 8770
},
{
"epoch": 0.5316363636363637,
"grad_norm": 0.046538837254047394,
"learning_rate": 0.0008,
"loss": 2.165800094604492,
"step": 8772
},
{
"epoch": 0.5317575757575758,
"grad_norm": 0.04622175544500351,
"learning_rate": 0.0008,
"loss": 2.2072551250457764,
"step": 8774
},
{
"epoch": 0.5318787878787878,
"grad_norm": 0.04747419059276581,
"learning_rate": 0.0008,
"loss": 2.132028579711914,
"step": 8776
},
{
"epoch": 0.532,
"grad_norm": 0.047398436814546585,
"learning_rate": 0.0008,
"loss": 2.0946714878082275,
"step": 8778
},
{
"epoch": 0.5321212121212121,
"grad_norm": 0.05177504941821098,
"learning_rate": 0.0008,
"loss": 2.188328504562378,
"step": 8780
},
{
"epoch": 0.5322424242424243,
"grad_norm": 0.05261954665184021,
"learning_rate": 0.0008,
"loss": 2.1437742710113525,
"step": 8782
},
{
"epoch": 0.5323636363636364,
"grad_norm": 0.057136327028274536,
"learning_rate": 0.0008,
"loss": 2.1142046451568604,
"step": 8784
},
{
"epoch": 0.5324848484848484,
"grad_norm": 0.05783483386039734,
"learning_rate": 0.0008,
"loss": 2.135657787322998,
"step": 8786
},
{
"epoch": 0.5326060606060606,
"grad_norm": 0.060534052550792694,
"learning_rate": 0.0008,
"loss": 2.103593111038208,
"step": 8788
},
{
"epoch": 0.5327272727272727,
"grad_norm": 0.05502758547663689,
"learning_rate": 0.0008,
"loss": 2.1260552406311035,
"step": 8790
},
{
"epoch": 0.5328484848484848,
"grad_norm": 0.05236747860908508,
"learning_rate": 0.0008,
"loss": 2.1368932723999023,
"step": 8792
},
{
"epoch": 0.532969696969697,
"grad_norm": 0.04659409448504448,
"learning_rate": 0.0008,
"loss": 2.1016457080841064,
"step": 8794
},
{
"epoch": 0.5330909090909091,
"grad_norm": 0.048025138676166534,
"learning_rate": 0.0008,
"loss": 2.122283458709717,
"step": 8796
},
{
"epoch": 0.5332121212121212,
"grad_norm": 0.046175118535757065,
"learning_rate": 0.0008,
"loss": 2.171455144882202,
"step": 8798
},
{
"epoch": 0.5333333333333333,
"grad_norm": 0.04858355224132538,
"learning_rate": 0.0008,
"loss": 2.1423144340515137,
"step": 8800
},
{
"epoch": 0.5334545454545454,
"grad_norm": 0.04491902515292168,
"learning_rate": 0.0008,
"loss": 2.198267698287964,
"step": 8802
},
{
"epoch": 0.5335757575757576,
"grad_norm": 0.04706721007823944,
"learning_rate": 0.0008,
"loss": 2.0998363494873047,
"step": 8804
},
{
"epoch": 0.5336969696969697,
"grad_norm": 0.04476979002356529,
"learning_rate": 0.0008,
"loss": 2.1169915199279785,
"step": 8806
},
{
"epoch": 0.5338181818181819,
"grad_norm": 0.04749346897006035,
"learning_rate": 0.0008,
"loss": 2.1570396423339844,
"step": 8808
},
{
"epoch": 0.5339393939393939,
"grad_norm": 0.04556579142808914,
"learning_rate": 0.0008,
"loss": 2.169663906097412,
"step": 8810
},
{
"epoch": 0.534060606060606,
"grad_norm": 0.04652049019932747,
"learning_rate": 0.0008,
"loss": 2.1502110958099365,
"step": 8812
},
{
"epoch": 0.5341818181818182,
"grad_norm": 0.04460356384515762,
"learning_rate": 0.0008,
"loss": 2.153834104537964,
"step": 8814
},
{
"epoch": 0.5343030303030303,
"grad_norm": 0.050916627049446106,
"learning_rate": 0.0008,
"loss": 2.165104866027832,
"step": 8816
},
{
"epoch": 0.5344242424242425,
"grad_norm": 0.050401680171489716,
"learning_rate": 0.0008,
"loss": 2.1718997955322266,
"step": 8818
},
{
"epoch": 0.5345454545454545,
"grad_norm": 0.045983053743839264,
"learning_rate": 0.0008,
"loss": 2.1749138832092285,
"step": 8820
},
{
"epoch": 0.5346666666666666,
"grad_norm": 0.047589175403118134,
"learning_rate": 0.0008,
"loss": 2.1375908851623535,
"step": 8822
},
{
"epoch": 0.5347878787878788,
"grad_norm": 0.047274306416511536,
"learning_rate": 0.0008,
"loss": 2.122556686401367,
"step": 8824
},
{
"epoch": 0.5349090909090909,
"grad_norm": 0.047303207218647,
"learning_rate": 0.0008,
"loss": 2.1552608013153076,
"step": 8826
},
{
"epoch": 0.5350303030303031,
"grad_norm": 0.0492565780878067,
"learning_rate": 0.0008,
"loss": 2.2120087146759033,
"step": 8828
},
{
"epoch": 0.5351515151515152,
"grad_norm": 0.04627785086631775,
"learning_rate": 0.0008,
"loss": 2.1519179344177246,
"step": 8830
},
{
"epoch": 0.5352727272727272,
"grad_norm": 0.05050015076994896,
"learning_rate": 0.0008,
"loss": 2.194899082183838,
"step": 8832
},
{
"epoch": 0.5353939393939394,
"grad_norm": 0.04626123979687691,
"learning_rate": 0.0008,
"loss": 2.168220281600952,
"step": 8834
},
{
"epoch": 0.5355151515151515,
"grad_norm": 0.05076048523187637,
"learning_rate": 0.0008,
"loss": 2.140307903289795,
"step": 8836
},
{
"epoch": 0.5356363636363637,
"grad_norm": 0.047656670212745667,
"learning_rate": 0.0008,
"loss": 2.1521198749542236,
"step": 8838
},
{
"epoch": 0.5357575757575758,
"grad_norm": 0.04721148684620857,
"learning_rate": 0.0008,
"loss": 2.083895444869995,
"step": 8840
},
{
"epoch": 0.5358787878787878,
"grad_norm": 0.04511282965540886,
"learning_rate": 0.0008,
"loss": 2.116098165512085,
"step": 8842
},
{
"epoch": 0.536,
"grad_norm": 0.045850180089473724,
"learning_rate": 0.0008,
"loss": 2.0764386653900146,
"step": 8844
},
{
"epoch": 0.5361212121212121,
"grad_norm": 0.04931901395320892,
"learning_rate": 0.0008,
"loss": 2.0957887172698975,
"step": 8846
},
{
"epoch": 0.5362424242424242,
"grad_norm": 0.05549366772174835,
"learning_rate": 0.0008,
"loss": 2.101243257522583,
"step": 8848
},
{
"epoch": 0.5363636363636364,
"grad_norm": 0.05132679641246796,
"learning_rate": 0.0008,
"loss": 2.07131290435791,
"step": 8850
},
{
"epoch": 0.5364848484848485,
"grad_norm": 0.04890340194106102,
"learning_rate": 0.0008,
"loss": 2.1260290145874023,
"step": 8852
},
{
"epoch": 0.5366060606060606,
"grad_norm": 0.045809224247932434,
"learning_rate": 0.0008,
"loss": 2.1493983268737793,
"step": 8854
},
{
"epoch": 0.5367272727272727,
"grad_norm": 0.0473540723323822,
"learning_rate": 0.0008,
"loss": 2.1264986991882324,
"step": 8856
},
{
"epoch": 0.5368484848484848,
"grad_norm": 0.05196674168109894,
"learning_rate": 0.0008,
"loss": 2.081768035888672,
"step": 8858
},
{
"epoch": 0.536969696969697,
"grad_norm": 0.04728655144572258,
"learning_rate": 0.0008,
"loss": 2.1650478839874268,
"step": 8860
},
{
"epoch": 0.5370909090909091,
"grad_norm": 0.04660249128937721,
"learning_rate": 0.0008,
"loss": 2.1655242443084717,
"step": 8862
},
{
"epoch": 0.5372121212121213,
"grad_norm": 0.045961543917655945,
"learning_rate": 0.0008,
"loss": 2.0785036087036133,
"step": 8864
},
{
"epoch": 0.5373333333333333,
"grad_norm": 0.044937461614608765,
"learning_rate": 0.0008,
"loss": 2.119596004486084,
"step": 8866
},
{
"epoch": 0.5374545454545454,
"grad_norm": 0.04533024877309799,
"learning_rate": 0.0008,
"loss": 2.1213040351867676,
"step": 8868
},
{
"epoch": 0.5375757575757576,
"grad_norm": 0.04991013929247856,
"learning_rate": 0.0008,
"loss": 2.081308603286743,
"step": 8870
},
{
"epoch": 0.5376969696969697,
"grad_norm": 0.05056043714284897,
"learning_rate": 0.0008,
"loss": 2.154758930206299,
"step": 8872
},
{
"epoch": 0.5378181818181819,
"grad_norm": 0.053088754415512085,
"learning_rate": 0.0008,
"loss": 2.0644798278808594,
"step": 8874
},
{
"epoch": 0.5379393939393939,
"grad_norm": 0.04704603552818298,
"learning_rate": 0.0008,
"loss": 2.1108431816101074,
"step": 8876
},
{
"epoch": 0.538060606060606,
"grad_norm": 0.05661095306277275,
"learning_rate": 0.0008,
"loss": 2.1406500339508057,
"step": 8878
},
{
"epoch": 0.5381818181818182,
"grad_norm": 0.04812159016728401,
"learning_rate": 0.0008,
"loss": 2.1560072898864746,
"step": 8880
},
{
"epoch": 0.5383030303030303,
"grad_norm": 0.05266579985618591,
"learning_rate": 0.0008,
"loss": 2.158108711242676,
"step": 8882
},
{
"epoch": 0.5384242424242425,
"grad_norm": 0.04805731773376465,
"learning_rate": 0.0008,
"loss": 2.0633790493011475,
"step": 8884
},
{
"epoch": 0.5385454545454545,
"grad_norm": 0.046101320534944534,
"learning_rate": 0.0008,
"loss": 2.1507153511047363,
"step": 8886
},
{
"epoch": 0.5386666666666666,
"grad_norm": 0.045733995735645294,
"learning_rate": 0.0008,
"loss": 2.1607666015625,
"step": 8888
},
{
"epoch": 0.5387878787878788,
"grad_norm": 0.049697477370500565,
"learning_rate": 0.0008,
"loss": 2.044422149658203,
"step": 8890
},
{
"epoch": 0.5389090909090909,
"grad_norm": 0.046755723655223846,
"learning_rate": 0.0008,
"loss": 2.1414575576782227,
"step": 8892
},
{
"epoch": 0.5390303030303031,
"grad_norm": 0.04731722176074982,
"learning_rate": 0.0008,
"loss": 2.1559994220733643,
"step": 8894
},
{
"epoch": 0.5391515151515152,
"grad_norm": 0.04687732830643654,
"learning_rate": 0.0008,
"loss": 2.1816132068634033,
"step": 8896
},
{
"epoch": 0.5392727272727272,
"grad_norm": 0.05190713703632355,
"learning_rate": 0.0008,
"loss": 2.1026086807250977,
"step": 8898
},
{
"epoch": 0.5393939393939394,
"grad_norm": 0.06283974647521973,
"learning_rate": 0.0008,
"loss": 2.20451021194458,
"step": 8900
},
{
"epoch": 0.5395151515151515,
"grad_norm": 0.06881430000066757,
"learning_rate": 0.0008,
"loss": 2.2274889945983887,
"step": 8902
},
{
"epoch": 0.5396363636363637,
"grad_norm": 0.06196245178580284,
"learning_rate": 0.0008,
"loss": 2.14373779296875,
"step": 8904
},
{
"epoch": 0.5397575757575758,
"grad_norm": 0.049798622727394104,
"learning_rate": 0.0008,
"loss": 2.200401544570923,
"step": 8906
},
{
"epoch": 0.5398787878787878,
"grad_norm": 0.04722832888364792,
"learning_rate": 0.0008,
"loss": 2.1154305934906006,
"step": 8908
},
{
"epoch": 0.54,
"grad_norm": 0.060799602419137955,
"learning_rate": 0.0008,
"loss": 2.1369125843048096,
"step": 8910
},
{
"epoch": 0.5401212121212121,
"grad_norm": 0.08700034022331238,
"learning_rate": 0.0008,
"loss": 2.1902871131896973,
"step": 8912
},
{
"epoch": 0.5402424242424242,
"grad_norm": 0.08934146910905838,
"learning_rate": 0.0008,
"loss": 2.1112468242645264,
"step": 8914
},
{
"epoch": 0.5403636363636364,
"grad_norm": 0.06665695458650589,
"learning_rate": 0.0008,
"loss": 2.1320605278015137,
"step": 8916
},
{
"epoch": 0.5404848484848485,
"grad_norm": 0.07137511670589447,
"learning_rate": 0.0008,
"loss": 2.1838674545288086,
"step": 8918
},
{
"epoch": 0.5406060606060606,
"grad_norm": 0.04981014132499695,
"learning_rate": 0.0008,
"loss": 2.170440912246704,
"step": 8920
},
{
"epoch": 0.5407272727272727,
"grad_norm": 0.047387924045324326,
"learning_rate": 0.0008,
"loss": 2.1310172080993652,
"step": 8922
},
{
"epoch": 0.5408484848484848,
"grad_norm": 0.04612470790743828,
"learning_rate": 0.0008,
"loss": 2.073934316635132,
"step": 8924
},
{
"epoch": 0.540969696969697,
"grad_norm": 0.04689427465200424,
"learning_rate": 0.0008,
"loss": 2.1366608142852783,
"step": 8926
},
{
"epoch": 0.5410909090909091,
"grad_norm": 0.05093684047460556,
"learning_rate": 0.0008,
"loss": 2.0522303581237793,
"step": 8928
},
{
"epoch": 0.5412121212121213,
"grad_norm": 0.0452854223549366,
"learning_rate": 0.0008,
"loss": 2.018732786178589,
"step": 8930
},
{
"epoch": 0.5413333333333333,
"grad_norm": 0.048446524888277054,
"learning_rate": 0.0008,
"loss": 2.156745433807373,
"step": 8932
},
{
"epoch": 0.5414545454545454,
"grad_norm": 0.049191150814294815,
"learning_rate": 0.0008,
"loss": 2.1368021965026855,
"step": 8934
},
{
"epoch": 0.5415757575757576,
"grad_norm": 0.05045134946703911,
"learning_rate": 0.0008,
"loss": 2.165477752685547,
"step": 8936
},
{
"epoch": 0.5416969696969697,
"grad_norm": 0.047132380306720734,
"learning_rate": 0.0008,
"loss": 2.1028943061828613,
"step": 8938
},
{
"epoch": 0.5418181818181819,
"grad_norm": 0.04545298591256142,
"learning_rate": 0.0008,
"loss": 2.1500632762908936,
"step": 8940
},
{
"epoch": 0.5419393939393939,
"grad_norm": 0.0498824343085289,
"learning_rate": 0.0008,
"loss": 2.1539061069488525,
"step": 8942
},
{
"epoch": 0.542060606060606,
"grad_norm": 0.048333872109651566,
"learning_rate": 0.0008,
"loss": 2.162977457046509,
"step": 8944
},
{
"epoch": 0.5421818181818182,
"grad_norm": 0.059710543602705,
"learning_rate": 0.0008,
"loss": 2.1688475608825684,
"step": 8946
},
{
"epoch": 0.5423030303030303,
"grad_norm": 0.059325266629457474,
"learning_rate": 0.0008,
"loss": 2.026805877685547,
"step": 8948
},
{
"epoch": 0.5424242424242425,
"grad_norm": 0.05524768680334091,
"learning_rate": 0.0008,
"loss": 2.1523067951202393,
"step": 8950
},
{
"epoch": 0.5425454545454546,
"grad_norm": 0.06731913238763809,
"learning_rate": 0.0008,
"loss": 2.134692668914795,
"step": 8952
},
{
"epoch": 0.5426666666666666,
"grad_norm": 0.04913710057735443,
"learning_rate": 0.0008,
"loss": 2.103654623031616,
"step": 8954
},
{
"epoch": 0.5427878787878788,
"grad_norm": 0.046949710696935654,
"learning_rate": 0.0008,
"loss": 2.120983600616455,
"step": 8956
},
{
"epoch": 0.5429090909090909,
"grad_norm": 0.047160230576992035,
"learning_rate": 0.0008,
"loss": 2.1268742084503174,
"step": 8958
},
{
"epoch": 0.5430303030303031,
"grad_norm": 0.05085734650492668,
"learning_rate": 0.0008,
"loss": 2.1441729068756104,
"step": 8960
},
{
"epoch": 0.5431515151515152,
"grad_norm": 0.048603206872940063,
"learning_rate": 0.0008,
"loss": 2.1923563480377197,
"step": 8962
},
{
"epoch": 0.5432727272727272,
"grad_norm": 0.054621390998363495,
"learning_rate": 0.0008,
"loss": 2.1528408527374268,
"step": 8964
},
{
"epoch": 0.5433939393939394,
"grad_norm": 0.05402825027704239,
"learning_rate": 0.0008,
"loss": 2.0985047817230225,
"step": 8966
},
{
"epoch": 0.5435151515151515,
"grad_norm": 0.047704823315143585,
"learning_rate": 0.0008,
"loss": 2.073065996170044,
"step": 8968
},
{
"epoch": 0.5436363636363636,
"grad_norm": 0.04797424003481865,
"learning_rate": 0.0008,
"loss": 2.249933958053589,
"step": 8970
},
{
"epoch": 0.5437575757575758,
"grad_norm": 0.04514982923865318,
"learning_rate": 0.0008,
"loss": 2.1603736877441406,
"step": 8972
},
{
"epoch": 0.5438787878787879,
"grad_norm": 0.04583803564310074,
"learning_rate": 0.0008,
"loss": 2.101763963699341,
"step": 8974
},
{
"epoch": 0.544,
"grad_norm": 0.053082194179296494,
"learning_rate": 0.0008,
"loss": 2.1229066848754883,
"step": 8976
},
{
"epoch": 0.5441212121212121,
"grad_norm": 0.05294787138700485,
"learning_rate": 0.0008,
"loss": 2.098785161972046,
"step": 8978
},
{
"epoch": 0.5442424242424242,
"grad_norm": 0.05375227704644203,
"learning_rate": 0.0008,
"loss": 2.171412944793701,
"step": 8980
},
{
"epoch": 0.5443636363636364,
"grad_norm": 0.04901891574263573,
"learning_rate": 0.0008,
"loss": 2.0789294242858887,
"step": 8982
},
{
"epoch": 0.5444848484848485,
"grad_norm": 0.05311161279678345,
"learning_rate": 0.0008,
"loss": 2.1414237022399902,
"step": 8984
},
{
"epoch": 0.5446060606060606,
"grad_norm": 0.05000828579068184,
"learning_rate": 0.0008,
"loss": 2.0777628421783447,
"step": 8986
},
{
"epoch": 0.5447272727272727,
"grad_norm": 0.04529724270105362,
"learning_rate": 0.0008,
"loss": 2.2521002292633057,
"step": 8988
},
{
"epoch": 0.5448484848484848,
"grad_norm": 0.046573787927627563,
"learning_rate": 0.0008,
"loss": 2.1269357204437256,
"step": 8990
},
{
"epoch": 0.544969696969697,
"grad_norm": 0.05146729201078415,
"learning_rate": 0.0008,
"loss": 2.069091558456421,
"step": 8992
},
{
"epoch": 0.5450909090909091,
"grad_norm": 0.048249002546072006,
"learning_rate": 0.0008,
"loss": 2.1051254272460938,
"step": 8994
},
{
"epoch": 0.5452121212121213,
"grad_norm": 0.04833785817027092,
"learning_rate": 0.0008,
"loss": 2.0906589031219482,
"step": 8996
},
{
"epoch": 0.5453333333333333,
"grad_norm": 0.05760408565402031,
"learning_rate": 0.0008,
"loss": 2.1003575325012207,
"step": 8998
},
{
"epoch": 0.5454545454545454,
"grad_norm": 0.054709166288375854,
"learning_rate": 0.0008,
"loss": 2.111342668533325,
"step": 9000
},
{
"epoch": 0.5455757575757576,
"grad_norm": 0.050709113478660583,
"learning_rate": 0.0008,
"loss": 2.2271289825439453,
"step": 9002
},
{
"epoch": 0.5456969696969697,
"grad_norm": 0.05555157735943794,
"learning_rate": 0.0008,
"loss": 2.0897562503814697,
"step": 9004
},
{
"epoch": 0.5458181818181819,
"grad_norm": 0.0538334995508194,
"learning_rate": 0.0008,
"loss": 2.1033008098602295,
"step": 9006
},
{
"epoch": 0.545939393939394,
"grad_norm": 0.05081198364496231,
"learning_rate": 0.0008,
"loss": 2.085883855819702,
"step": 9008
},
{
"epoch": 0.546060606060606,
"grad_norm": 0.04642565920948982,
"learning_rate": 0.0008,
"loss": 2.0406246185302734,
"step": 9010
},
{
"epoch": 0.5461818181818182,
"grad_norm": 0.05984676629304886,
"learning_rate": 0.0008,
"loss": 2.1954848766326904,
"step": 9012
},
{
"epoch": 0.5463030303030303,
"grad_norm": 0.0630536824464798,
"learning_rate": 0.0008,
"loss": 2.1239211559295654,
"step": 9014
},
{
"epoch": 0.5464242424242425,
"grad_norm": 0.05366978049278259,
"learning_rate": 0.0008,
"loss": 2.0811562538146973,
"step": 9016
},
{
"epoch": 0.5465454545454546,
"grad_norm": 0.04443780705332756,
"learning_rate": 0.0008,
"loss": 2.128652334213257,
"step": 9018
},
{
"epoch": 0.5466666666666666,
"grad_norm": 0.04632680490612984,
"learning_rate": 0.0008,
"loss": 2.1366660594940186,
"step": 9020
},
{
"epoch": 0.5467878787878788,
"grad_norm": 0.05580275133252144,
"learning_rate": 0.0008,
"loss": 2.0981552600860596,
"step": 9022
},
{
"epoch": 0.5469090909090909,
"grad_norm": 0.051582228392362595,
"learning_rate": 0.0008,
"loss": 2.121314287185669,
"step": 9024
},
{
"epoch": 0.547030303030303,
"grad_norm": 0.04500115290284157,
"learning_rate": 0.0008,
"loss": 2.2101376056671143,
"step": 9026
},
{
"epoch": 0.5471515151515152,
"grad_norm": 0.047117818146944046,
"learning_rate": 0.0008,
"loss": 2.077725887298584,
"step": 9028
},
{
"epoch": 0.5472727272727272,
"grad_norm": 0.04584265872836113,
"learning_rate": 0.0008,
"loss": 2.230353355407715,
"step": 9030
},
{
"epoch": 0.5473939393939394,
"grad_norm": 0.0483592227101326,
"learning_rate": 0.0008,
"loss": 2.1140193939208984,
"step": 9032
},
{
"epoch": 0.5475151515151515,
"grad_norm": 0.04497748985886574,
"learning_rate": 0.0008,
"loss": 2.114037036895752,
"step": 9034
},
{
"epoch": 0.5476363636363636,
"grad_norm": 0.04548939689993858,
"learning_rate": 0.0008,
"loss": 2.1338272094726562,
"step": 9036
},
{
"epoch": 0.5477575757575758,
"grad_norm": 0.045067865401506424,
"learning_rate": 0.0008,
"loss": 2.1243600845336914,
"step": 9038
},
{
"epoch": 0.5478787878787879,
"grad_norm": 0.048513129353523254,
"learning_rate": 0.0008,
"loss": 2.0842125415802,
"step": 9040
},
{
"epoch": 0.548,
"grad_norm": 0.04637809842824936,
"learning_rate": 0.0008,
"loss": 2.1147987842559814,
"step": 9042
},
{
"epoch": 0.5481212121212121,
"grad_norm": 0.045672934502363205,
"learning_rate": 0.0008,
"loss": 2.1256155967712402,
"step": 9044
},
{
"epoch": 0.5482424242424242,
"grad_norm": 0.04629864543676376,
"learning_rate": 0.0008,
"loss": 2.1600983142852783,
"step": 9046
},
{
"epoch": 0.5483636363636364,
"grad_norm": 0.051019612699747086,
"learning_rate": 0.0008,
"loss": 2.064417839050293,
"step": 9048
},
{
"epoch": 0.5484848484848485,
"grad_norm": 0.053016819059848785,
"learning_rate": 0.0008,
"loss": 2.0985910892486572,
"step": 9050
},
{
"epoch": 0.5486060606060607,
"grad_norm": 0.05159096419811249,
"learning_rate": 0.0008,
"loss": 2.138340711593628,
"step": 9052
},
{
"epoch": 0.5487272727272727,
"grad_norm": 0.047770705074071884,
"learning_rate": 0.0008,
"loss": 2.1293609142303467,
"step": 9054
},
{
"epoch": 0.5488484848484848,
"grad_norm": 0.04540763795375824,
"learning_rate": 0.0008,
"loss": 2.049954652786255,
"step": 9056
},
{
"epoch": 0.548969696969697,
"grad_norm": 0.04587807133793831,
"learning_rate": 0.0008,
"loss": 2.062554121017456,
"step": 9058
},
{
"epoch": 0.5490909090909091,
"grad_norm": 0.043385881930589676,
"learning_rate": 0.0008,
"loss": 2.0893473625183105,
"step": 9060
},
{
"epoch": 0.5492121212121213,
"grad_norm": 0.04990014061331749,
"learning_rate": 0.0008,
"loss": 2.0623068809509277,
"step": 9062
},
{
"epoch": 0.5493333333333333,
"grad_norm": 0.049697041511535645,
"learning_rate": 0.0008,
"loss": 2.09887957572937,
"step": 9064
},
{
"epoch": 0.5494545454545454,
"grad_norm": 0.046254608780145645,
"learning_rate": 0.0008,
"loss": 2.1867525577545166,
"step": 9066
},
{
"epoch": 0.5495757575757576,
"grad_norm": 0.04975316300988197,
"learning_rate": 0.0008,
"loss": 2.1888012886047363,
"step": 9068
},
{
"epoch": 0.5496969696969697,
"grad_norm": 0.04570948705077171,
"learning_rate": 0.0008,
"loss": 2.123342275619507,
"step": 9070
},
{
"epoch": 0.5498181818181819,
"grad_norm": 0.04581759497523308,
"learning_rate": 0.0008,
"loss": 2.0548317432403564,
"step": 9072
},
{
"epoch": 0.549939393939394,
"grad_norm": 0.04806012660264969,
"learning_rate": 0.0008,
"loss": 2.0488665103912354,
"step": 9074
},
{
"epoch": 0.550060606060606,
"grad_norm": 0.055171966552734375,
"learning_rate": 0.0008,
"loss": 2.1415958404541016,
"step": 9076
},
{
"epoch": 0.5501818181818182,
"grad_norm": 0.060776203870773315,
"learning_rate": 0.0008,
"loss": 2.120274782180786,
"step": 9078
},
{
"epoch": 0.5503030303030303,
"grad_norm": 0.05132869631052017,
"learning_rate": 0.0008,
"loss": 2.1179440021514893,
"step": 9080
},
{
"epoch": 0.5504242424242424,
"grad_norm": 0.06765241175889969,
"learning_rate": 0.0008,
"loss": 2.0611135959625244,
"step": 9082
},
{
"epoch": 0.5505454545454546,
"grad_norm": 0.05524882301688194,
"learning_rate": 0.0008,
"loss": 2.1199498176574707,
"step": 9084
},
{
"epoch": 0.5506666666666666,
"grad_norm": 0.055927351117134094,
"learning_rate": 0.0008,
"loss": 2.041346549987793,
"step": 9086
},
{
"epoch": 0.5507878787878788,
"grad_norm": 0.05457756668329239,
"learning_rate": 0.0008,
"loss": 2.2084319591522217,
"step": 9088
},
{
"epoch": 0.5509090909090909,
"grad_norm": 0.0559866800904274,
"learning_rate": 0.0008,
"loss": 2.08538818359375,
"step": 9090
},
{
"epoch": 0.551030303030303,
"grad_norm": 0.04753634333610535,
"learning_rate": 0.0008,
"loss": 2.1796016693115234,
"step": 9092
},
{
"epoch": 0.5511515151515152,
"grad_norm": 0.05510437488555908,
"learning_rate": 0.0008,
"loss": 2.141375780105591,
"step": 9094
},
{
"epoch": 0.5512727272727272,
"grad_norm": 0.06048452854156494,
"learning_rate": 0.0008,
"loss": 2.1404531002044678,
"step": 9096
},
{
"epoch": 0.5513939393939394,
"grad_norm": 0.06613709777593613,
"learning_rate": 0.0008,
"loss": 2.0671660900115967,
"step": 9098
},
{
"epoch": 0.5515151515151515,
"grad_norm": 0.05610032007098198,
"learning_rate": 0.0008,
"loss": 2.050607681274414,
"step": 9100
},
{
"epoch": 0.5516363636363636,
"grad_norm": 0.05413105711340904,
"learning_rate": 0.0008,
"loss": 2.1087160110473633,
"step": 9102
},
{
"epoch": 0.5517575757575758,
"grad_norm": 0.04332798346877098,
"learning_rate": 0.0008,
"loss": 2.0802125930786133,
"step": 9104
},
{
"epoch": 0.5518787878787879,
"grad_norm": 0.046087246388196945,
"learning_rate": 0.0008,
"loss": 2.085374355316162,
"step": 9106
},
{
"epoch": 0.552,
"grad_norm": 0.047814857214689255,
"learning_rate": 0.0008,
"loss": 2.14816951751709,
"step": 9108
},
{
"epoch": 0.5521212121212121,
"grad_norm": 0.04943268373608589,
"learning_rate": 0.0008,
"loss": 2.061154842376709,
"step": 9110
},
{
"epoch": 0.5522424242424242,
"grad_norm": 0.04540696740150452,
"learning_rate": 0.0008,
"loss": 2.1216001510620117,
"step": 9112
},
{
"epoch": 0.5523636363636364,
"grad_norm": 0.044582631438970566,
"learning_rate": 0.0008,
"loss": 2.1892967224121094,
"step": 9114
},
{
"epoch": 0.5524848484848485,
"grad_norm": 0.045067619532346725,
"learning_rate": 0.0008,
"loss": 2.1067497730255127,
"step": 9116
},
{
"epoch": 0.5526060606060607,
"grad_norm": 0.044802311807870865,
"learning_rate": 0.0008,
"loss": 2.0255801677703857,
"step": 9118
},
{
"epoch": 0.5527272727272727,
"grad_norm": 0.04699620231986046,
"learning_rate": 0.0008,
"loss": 2.0865631103515625,
"step": 9120
},
{
"epoch": 0.5528484848484848,
"grad_norm": 0.04443885385990143,
"learning_rate": 0.0008,
"loss": 2.047379732131958,
"step": 9122
},
{
"epoch": 0.552969696969697,
"grad_norm": 0.04696553945541382,
"learning_rate": 0.0008,
"loss": 2.0765247344970703,
"step": 9124
},
{
"epoch": 0.5530909090909091,
"grad_norm": 0.04516385495662689,
"learning_rate": 0.0008,
"loss": 2.0895400047302246,
"step": 9126
},
{
"epoch": 0.5532121212121213,
"grad_norm": 0.04376617446541786,
"learning_rate": 0.0008,
"loss": 2.0721609592437744,
"step": 9128
},
{
"epoch": 0.5533333333333333,
"grad_norm": 0.050105299800634384,
"learning_rate": 0.0008,
"loss": 2.112190008163452,
"step": 9130
},
{
"epoch": 0.5534545454545454,
"grad_norm": 0.05035905912518501,
"learning_rate": 0.0008,
"loss": 2.1040327548980713,
"step": 9132
},
{
"epoch": 0.5535757575757576,
"grad_norm": 0.04767022654414177,
"learning_rate": 0.0008,
"loss": 2.0822911262512207,
"step": 9134
},
{
"epoch": 0.5536969696969697,
"grad_norm": 0.05032365396618843,
"learning_rate": 0.0008,
"loss": 2.063143730163574,
"step": 9136
},
{
"epoch": 0.5538181818181818,
"grad_norm": 0.044426292181015015,
"learning_rate": 0.0008,
"loss": 2.099061965942383,
"step": 9138
},
{
"epoch": 0.553939393939394,
"grad_norm": 0.045002974569797516,
"learning_rate": 0.0008,
"loss": 2.1299548149108887,
"step": 9140
},
{
"epoch": 0.554060606060606,
"grad_norm": 0.04684218764305115,
"learning_rate": 0.0008,
"loss": 2.069319486618042,
"step": 9142
},
{
"epoch": 0.5541818181818182,
"grad_norm": 0.044668130576610565,
"learning_rate": 0.0008,
"loss": 2.147143602371216,
"step": 9144
},
{
"epoch": 0.5543030303030303,
"grad_norm": 0.04496139660477638,
"learning_rate": 0.0008,
"loss": 2.0905826091766357,
"step": 9146
},
{
"epoch": 0.5544242424242424,
"grad_norm": 0.046754613518714905,
"learning_rate": 0.0008,
"loss": 2.0660927295684814,
"step": 9148
},
{
"epoch": 0.5545454545454546,
"grad_norm": 0.04497559741139412,
"learning_rate": 0.0008,
"loss": 2.0726559162139893,
"step": 9150
},
{
"epoch": 0.5546666666666666,
"grad_norm": 0.043929096311330795,
"learning_rate": 0.0008,
"loss": 2.0814602375030518,
"step": 9152
},
{
"epoch": 0.5547878787878788,
"grad_norm": 0.04850081726908684,
"learning_rate": 0.0008,
"loss": 2.0875277519226074,
"step": 9154
},
{
"epoch": 0.5549090909090909,
"grad_norm": 0.04314367473125458,
"learning_rate": 0.0008,
"loss": 2.0092082023620605,
"step": 9156
},
{
"epoch": 0.555030303030303,
"grad_norm": 0.05242739990353584,
"learning_rate": 0.0008,
"loss": 2.054948568344116,
"step": 9158
},
{
"epoch": 0.5551515151515152,
"grad_norm": 0.05444403737783432,
"learning_rate": 0.0008,
"loss": 2.0998003482818604,
"step": 9160
},
{
"epoch": 0.5552727272727273,
"grad_norm": 0.05883711576461792,
"learning_rate": 0.0008,
"loss": 2.1133108139038086,
"step": 9162
},
{
"epoch": 0.5553939393939394,
"grad_norm": 0.043979398906230927,
"learning_rate": 0.0008,
"loss": 2.040067195892334,
"step": 9164
},
{
"epoch": 0.5555151515151515,
"grad_norm": 0.04663465917110443,
"learning_rate": 0.0008,
"loss": 2.015568733215332,
"step": 9166
},
{
"epoch": 0.5556363636363636,
"grad_norm": 0.051379770040512085,
"learning_rate": 0.0008,
"loss": 2.059941053390503,
"step": 9168
},
{
"epoch": 0.5557575757575758,
"grad_norm": 0.05317751318216324,
"learning_rate": 0.0008,
"loss": 2.0342442989349365,
"step": 9170
},
{
"epoch": 0.5558787878787879,
"grad_norm": 0.046021196991205215,
"learning_rate": 0.0008,
"loss": 2.0519282817840576,
"step": 9172
},
{
"epoch": 0.556,
"grad_norm": 0.05019720643758774,
"learning_rate": 0.0008,
"loss": 2.055069923400879,
"step": 9174
},
{
"epoch": 0.5561212121212121,
"grad_norm": 0.05257274582982063,
"learning_rate": 0.0008,
"loss": 2.1381337642669678,
"step": 9176
},
{
"epoch": 0.5562424242424242,
"grad_norm": 0.04516459256410599,
"learning_rate": 0.0008,
"loss": 2.0941240787506104,
"step": 9178
},
{
"epoch": 0.5563636363636364,
"grad_norm": 0.05317554250359535,
"learning_rate": 0.0008,
"loss": 2.149589776992798,
"step": 9180
},
{
"epoch": 0.5564848484848485,
"grad_norm": 0.046958886086940765,
"learning_rate": 0.0008,
"loss": 2.0573503971099854,
"step": 9182
},
{
"epoch": 0.5566060606060607,
"grad_norm": 0.04726814106106758,
"learning_rate": 0.0008,
"loss": 2.1396384239196777,
"step": 9184
},
{
"epoch": 0.5567272727272727,
"grad_norm": 0.0451323576271534,
"learning_rate": 0.0008,
"loss": 2.0250906944274902,
"step": 9186
},
{
"epoch": 0.5568484848484848,
"grad_norm": 0.04568498209118843,
"learning_rate": 0.0008,
"loss": 2.0839667320251465,
"step": 9188
},
{
"epoch": 0.556969696969697,
"grad_norm": 0.044302694499492645,
"learning_rate": 0.0008,
"loss": 2.053900957107544,
"step": 9190
},
{
"epoch": 0.5570909090909091,
"grad_norm": 0.04425807669758797,
"learning_rate": 0.0008,
"loss": 2.066148281097412,
"step": 9192
},
{
"epoch": 0.5572121212121212,
"grad_norm": 0.04532415792346001,
"learning_rate": 0.0008,
"loss": 2.0736727714538574,
"step": 9194
},
{
"epoch": 0.5573333333333333,
"grad_norm": 0.046504467725753784,
"learning_rate": 0.0008,
"loss": 2.143587827682495,
"step": 9196
},
{
"epoch": 0.5574545454545454,
"grad_norm": 0.04544759914278984,
"learning_rate": 0.0008,
"loss": 2.0887162685394287,
"step": 9198
},
{
"epoch": 0.5575757575757576,
"grad_norm": 0.0474565215408802,
"learning_rate": 0.0008,
"loss": 2.0657405853271484,
"step": 9200
},
{
"epoch": 0.5576969696969697,
"grad_norm": 0.05991155281662941,
"learning_rate": 0.0008,
"loss": 2.1192314624786377,
"step": 9202
},
{
"epoch": 0.5578181818181818,
"grad_norm": 0.05379079282283783,
"learning_rate": 0.0008,
"loss": 2.078022003173828,
"step": 9204
},
{
"epoch": 0.557939393939394,
"grad_norm": 0.05468444153666496,
"learning_rate": 0.0008,
"loss": 2.0673468112945557,
"step": 9206
},
{
"epoch": 0.558060606060606,
"grad_norm": 0.05786069482564926,
"learning_rate": 0.0008,
"loss": 2.161686897277832,
"step": 9208
},
{
"epoch": 0.5581818181818182,
"grad_norm": 0.04980327561497688,
"learning_rate": 0.0008,
"loss": 2.0578629970550537,
"step": 9210
},
{
"epoch": 0.5583030303030303,
"grad_norm": 0.051359374076128006,
"learning_rate": 0.0008,
"loss": 1.993133783340454,
"step": 9212
},
{
"epoch": 0.5584242424242424,
"grad_norm": 0.05020865797996521,
"learning_rate": 0.0008,
"loss": 2.050947904586792,
"step": 9214
},
{
"epoch": 0.5585454545454546,
"grad_norm": 0.047544095665216446,
"learning_rate": 0.0008,
"loss": 2.1037135124206543,
"step": 9216
},
{
"epoch": 0.5586666666666666,
"grad_norm": 0.045268941670656204,
"learning_rate": 0.0008,
"loss": 2.0421366691589355,
"step": 9218
},
{
"epoch": 0.5587878787878788,
"grad_norm": 0.049381908029317856,
"learning_rate": 0.0008,
"loss": 2.0985920429229736,
"step": 9220
},
{
"epoch": 0.5589090909090909,
"grad_norm": 0.048116426914930344,
"learning_rate": 0.0008,
"loss": 2.0460855960845947,
"step": 9222
},
{
"epoch": 0.559030303030303,
"grad_norm": 0.045693542808294296,
"learning_rate": 0.0008,
"loss": 2.060994863510132,
"step": 9224
},
{
"epoch": 0.5591515151515152,
"grad_norm": 0.04747629538178444,
"learning_rate": 0.0008,
"loss": 2.086820125579834,
"step": 9226
},
{
"epoch": 0.5592727272727273,
"grad_norm": 0.04425165802240372,
"learning_rate": 0.0008,
"loss": 2.0330567359924316,
"step": 9228
},
{
"epoch": 0.5593939393939394,
"grad_norm": 0.04462664574384689,
"learning_rate": 0.0008,
"loss": 2.0729990005493164,
"step": 9230
},
{
"epoch": 0.5595151515151515,
"grad_norm": 0.045130155980587006,
"learning_rate": 0.0008,
"loss": 2.0324418544769287,
"step": 9232
},
{
"epoch": 0.5596363636363636,
"grad_norm": 0.043741438537836075,
"learning_rate": 0.0008,
"loss": 2.067614793777466,
"step": 9234
},
{
"epoch": 0.5597575757575758,
"grad_norm": 0.044662363827228546,
"learning_rate": 0.0008,
"loss": 2.1249043941497803,
"step": 9236
},
{
"epoch": 0.5598787878787879,
"grad_norm": 0.045371223241090775,
"learning_rate": 0.0008,
"loss": 2.1626718044281006,
"step": 9238
},
{
"epoch": 0.56,
"grad_norm": 0.05174648016691208,
"learning_rate": 0.0008,
"loss": 2.0892393589019775,
"step": 9240
},
{
"epoch": 0.5601212121212121,
"grad_norm": 0.052778102457523346,
"learning_rate": 0.0008,
"loss": 2.0340352058410645,
"step": 9242
},
{
"epoch": 0.5602424242424242,
"grad_norm": 0.04608987271785736,
"learning_rate": 0.0008,
"loss": 2.0898470878601074,
"step": 9244
},
{
"epoch": 0.5603636363636364,
"grad_norm": 0.056317660957574844,
"learning_rate": 0.0008,
"loss": 2.0595877170562744,
"step": 9246
},
{
"epoch": 0.5604848484848485,
"grad_norm": 0.05253066495060921,
"learning_rate": 0.0008,
"loss": 2.0317318439483643,
"step": 9248
},
{
"epoch": 0.5606060606060606,
"grad_norm": 0.05629071220755577,
"learning_rate": 0.0008,
"loss": 2.06006121635437,
"step": 9250
},
{
"epoch": 0.5607272727272727,
"grad_norm": 0.056347161531448364,
"learning_rate": 0.0008,
"loss": 1.9976879358291626,
"step": 9252
},
{
"epoch": 0.5608484848484848,
"grad_norm": 0.06382175534963608,
"learning_rate": 0.0008,
"loss": 1.9801692962646484,
"step": 9254
},
{
"epoch": 0.560969696969697,
"grad_norm": 0.06156039610505104,
"learning_rate": 0.0008,
"loss": 2.082552433013916,
"step": 9256
},
{
"epoch": 0.5610909090909091,
"grad_norm": 0.05812281742691994,
"learning_rate": 0.0008,
"loss": 2.1298747062683105,
"step": 9258
},
{
"epoch": 0.5612121212121212,
"grad_norm": 0.054538294672966,
"learning_rate": 0.0008,
"loss": 2.1035408973693848,
"step": 9260
},
{
"epoch": 0.5613333333333334,
"grad_norm": 0.047348473221063614,
"learning_rate": 0.0008,
"loss": 2.0225956439971924,
"step": 9262
},
{
"epoch": 0.5614545454545454,
"grad_norm": 0.045593224465847015,
"learning_rate": 0.0008,
"loss": 2.1611766815185547,
"step": 9264
},
{
"epoch": 0.5615757575757576,
"grad_norm": 0.04511889070272446,
"learning_rate": 0.0008,
"loss": 2.050574779510498,
"step": 9266
},
{
"epoch": 0.5616969696969697,
"grad_norm": 0.05786488577723503,
"learning_rate": 0.0008,
"loss": 2.030881881713867,
"step": 9268
},
{
"epoch": 0.5618181818181818,
"grad_norm": 0.06706758588552475,
"learning_rate": 0.0008,
"loss": 2.0403637886047363,
"step": 9270
},
{
"epoch": 0.561939393939394,
"grad_norm": 0.05471501499414444,
"learning_rate": 0.0008,
"loss": 2.055309772491455,
"step": 9272
},
{
"epoch": 0.562060606060606,
"grad_norm": 0.06061052903532982,
"learning_rate": 0.0008,
"loss": 2.03935170173645,
"step": 9274
},
{
"epoch": 0.5621818181818182,
"grad_norm": 0.06825697422027588,
"learning_rate": 0.0008,
"loss": 2.061898946762085,
"step": 9276
},
{
"epoch": 0.5623030303030303,
"grad_norm": 0.06225455924868584,
"learning_rate": 0.0008,
"loss": 2.054572582244873,
"step": 9278
},
{
"epoch": 0.5624242424242424,
"grad_norm": 0.04830459877848625,
"learning_rate": 0.0008,
"loss": 1.9316598176956177,
"step": 9280
},
{
"epoch": 0.5625454545454546,
"grad_norm": 0.044505294412374496,
"learning_rate": 0.0008,
"loss": 2.0283069610595703,
"step": 9282
},
{
"epoch": 0.5626666666666666,
"grad_norm": 0.0438636913895607,
"learning_rate": 0.0008,
"loss": 2.068791389465332,
"step": 9284
},
{
"epoch": 0.5627878787878788,
"grad_norm": 0.04344771429896355,
"learning_rate": 0.0008,
"loss": 2.1390764713287354,
"step": 9286
},
{
"epoch": 0.5629090909090909,
"grad_norm": 0.04410022497177124,
"learning_rate": 0.0008,
"loss": 2.042423725128174,
"step": 9288
},
{
"epoch": 0.563030303030303,
"grad_norm": 0.04477225989103317,
"learning_rate": 0.0008,
"loss": 2.0285933017730713,
"step": 9290
},
{
"epoch": 0.5631515151515152,
"grad_norm": 0.04522207751870155,
"learning_rate": 0.0008,
"loss": 2.088583469390869,
"step": 9292
},
{
"epoch": 0.5632727272727273,
"grad_norm": 0.04431043937802315,
"learning_rate": 0.0008,
"loss": 2.084260940551758,
"step": 9294
},
{
"epoch": 0.5633939393939394,
"grad_norm": 0.04300132021307945,
"learning_rate": 0.0008,
"loss": 1.9689749479293823,
"step": 9296
},
{
"epoch": 0.5635151515151515,
"grad_norm": 0.04343254864215851,
"learning_rate": 0.0008,
"loss": 2.063758373260498,
"step": 9298
},
{
"epoch": 0.5636363636363636,
"grad_norm": 0.04221072420477867,
"learning_rate": 0.0008,
"loss": 2.070366382598877,
"step": 9300
},
{
"epoch": 0.5637575757575758,
"grad_norm": 0.044405095279216766,
"learning_rate": 0.0008,
"loss": 2.056730270385742,
"step": 9302
},
{
"epoch": 0.5638787878787879,
"grad_norm": 0.042156342417001724,
"learning_rate": 0.0008,
"loss": 2.0968611240386963,
"step": 9304
},
{
"epoch": 0.564,
"grad_norm": 0.04349891096353531,
"learning_rate": 0.0008,
"loss": 2.07193660736084,
"step": 9306
},
{
"epoch": 0.5641212121212121,
"grad_norm": 0.043549858033657074,
"learning_rate": 0.0008,
"loss": 2.037937879562378,
"step": 9308
},
{
"epoch": 0.5642424242424242,
"grad_norm": 0.042007699608802795,
"learning_rate": 0.0008,
"loss": 2.0484440326690674,
"step": 9310
},
{
"epoch": 0.5643636363636364,
"grad_norm": 0.0462723970413208,
"learning_rate": 0.0008,
"loss": 2.0688881874084473,
"step": 9312
},
{
"epoch": 0.5644848484848485,
"grad_norm": 0.04345852881669998,
"learning_rate": 0.0008,
"loss": 2.0790762901306152,
"step": 9314
},
{
"epoch": 0.5646060606060606,
"grad_norm": 0.04463418200612068,
"learning_rate": 0.0008,
"loss": 2.091142416000366,
"step": 9316
},
{
"epoch": 0.5647272727272727,
"grad_norm": 0.043843772262334824,
"learning_rate": 0.0008,
"loss": 2.0229506492614746,
"step": 9318
},
{
"epoch": 0.5648484848484848,
"grad_norm": 0.04786458611488342,
"learning_rate": 0.0008,
"loss": 2.0353899002075195,
"step": 9320
},
{
"epoch": 0.564969696969697,
"grad_norm": 0.04248730465769768,
"learning_rate": 0.0008,
"loss": 2.056833505630493,
"step": 9322
},
{
"epoch": 0.5650909090909091,
"grad_norm": 0.043540000915527344,
"learning_rate": 0.0008,
"loss": 1.9939087629318237,
"step": 9324
},
{
"epoch": 0.5652121212121212,
"grad_norm": 0.04419425129890442,
"learning_rate": 0.0008,
"loss": 2.043074131011963,
"step": 9326
},
{
"epoch": 0.5653333333333334,
"grad_norm": 0.04274079203605652,
"learning_rate": 0.0008,
"loss": 2.072272300720215,
"step": 9328
},
{
"epoch": 0.5654545454545454,
"grad_norm": 0.04236466437578201,
"learning_rate": 0.0008,
"loss": 1.9979774951934814,
"step": 9330
},
{
"epoch": 0.5655757575757576,
"grad_norm": 0.05156737193465233,
"learning_rate": 0.0008,
"loss": 1.9493056535720825,
"step": 9332
},
{
"epoch": 0.5656969696969697,
"grad_norm": 0.05050985515117645,
"learning_rate": 0.0008,
"loss": 1.9983114004135132,
"step": 9334
},
{
"epoch": 0.5658181818181818,
"grad_norm": 0.05730649456381798,
"learning_rate": 0.0008,
"loss": 2.0674500465393066,
"step": 9336
},
{
"epoch": 0.565939393939394,
"grad_norm": 0.04453360289335251,
"learning_rate": 0.0008,
"loss": 2.0326926708221436,
"step": 9338
},
{
"epoch": 0.566060606060606,
"grad_norm": 0.046399835497140884,
"learning_rate": 0.0008,
"loss": 2.068956136703491,
"step": 9340
},
{
"epoch": 0.5661818181818182,
"grad_norm": 0.043777573853731155,
"learning_rate": 0.0008,
"loss": 2.086794137954712,
"step": 9342
},
{
"epoch": 0.5663030303030303,
"grad_norm": 0.042066540569067,
"learning_rate": 0.0008,
"loss": 1.9840457439422607,
"step": 9344
},
{
"epoch": 0.5664242424242424,
"grad_norm": 0.0422976016998291,
"learning_rate": 0.0008,
"loss": 1.9965252876281738,
"step": 9346
},
{
"epoch": 0.5665454545454546,
"grad_norm": 0.04300987347960472,
"learning_rate": 0.0008,
"loss": 2.0669736862182617,
"step": 9348
},
{
"epoch": 0.5666666666666667,
"grad_norm": 0.04161219671368599,
"learning_rate": 0.0008,
"loss": 2.0106444358825684,
"step": 9350
},
{
"epoch": 0.5667878787878788,
"grad_norm": 0.04497823491692543,
"learning_rate": 0.0008,
"loss": 2.0162527561187744,
"step": 9352
},
{
"epoch": 0.5669090909090909,
"grad_norm": 0.042093660682439804,
"learning_rate": 0.0008,
"loss": 2.0719852447509766,
"step": 9354
},
{
"epoch": 0.567030303030303,
"grad_norm": 0.044764041900634766,
"learning_rate": 0.0008,
"loss": 2.0678203105926514,
"step": 9356
},
{
"epoch": 0.5671515151515152,
"grad_norm": 0.04558722302317619,
"learning_rate": 0.0008,
"loss": 2.061964750289917,
"step": 9358
},
{
"epoch": 0.5672727272727273,
"grad_norm": 0.04297410696744919,
"learning_rate": 0.0008,
"loss": 2.1086959838867188,
"step": 9360
},
{
"epoch": 0.5673939393939393,
"grad_norm": 0.04305315390229225,
"learning_rate": 0.0008,
"loss": 2.0932390689849854,
"step": 9362
},
{
"epoch": 0.5675151515151515,
"grad_norm": 0.04632577300071716,
"learning_rate": 0.0008,
"loss": 2.10782790184021,
"step": 9364
},
{
"epoch": 0.5676363636363636,
"grad_norm": 0.04314867779612541,
"learning_rate": 0.0008,
"loss": 2.10473895072937,
"step": 9366
},
{
"epoch": 0.5677575757575758,
"grad_norm": 0.044099584221839905,
"learning_rate": 0.0008,
"loss": 2.068610191345215,
"step": 9368
},
{
"epoch": 0.5678787878787879,
"grad_norm": 0.045903317630290985,
"learning_rate": 0.0008,
"loss": 2.0952868461608887,
"step": 9370
},
{
"epoch": 0.568,
"grad_norm": 0.04349956661462784,
"learning_rate": 0.0008,
"loss": 2.0613903999328613,
"step": 9372
},
{
"epoch": 0.5681212121212121,
"grad_norm": 0.043020620942115784,
"learning_rate": 0.0008,
"loss": 2.0418777465820312,
"step": 9374
},
{
"epoch": 0.5682424242424242,
"grad_norm": 0.04476744681596756,
"learning_rate": 0.0008,
"loss": 2.0358803272247314,
"step": 9376
},
{
"epoch": 0.5683636363636364,
"grad_norm": 0.0542496033012867,
"learning_rate": 0.0008,
"loss": 2.0389416217803955,
"step": 9378
},
{
"epoch": 0.5684848484848485,
"grad_norm": 0.053819119930267334,
"learning_rate": 0.0008,
"loss": 2.0099620819091797,
"step": 9380
},
{
"epoch": 0.5686060606060606,
"grad_norm": 0.04276927933096886,
"learning_rate": 0.0008,
"loss": 1.9767725467681885,
"step": 9382
},
{
"epoch": 0.5687272727272727,
"grad_norm": 0.044380031526088715,
"learning_rate": 0.0008,
"loss": 2.039132833480835,
"step": 9384
},
{
"epoch": 0.5688484848484848,
"grad_norm": 0.047349706292152405,
"learning_rate": 0.0008,
"loss": 2.082087278366089,
"step": 9386
},
{
"epoch": 0.568969696969697,
"grad_norm": 0.04568793997168541,
"learning_rate": 0.0008,
"loss": 2.0551111698150635,
"step": 9388
},
{
"epoch": 0.5690909090909091,
"grad_norm": 0.04721604660153389,
"learning_rate": 0.0008,
"loss": 2.052335262298584,
"step": 9390
},
{
"epoch": 0.5692121212121212,
"grad_norm": 0.04871431738138199,
"learning_rate": 0.0008,
"loss": 2.0606529712677,
"step": 9392
},
{
"epoch": 0.5693333333333334,
"grad_norm": 0.05865452066063881,
"learning_rate": 0.0008,
"loss": 1.998281717300415,
"step": 9394
},
{
"epoch": 0.5694545454545454,
"grad_norm": 0.04417392984032631,
"learning_rate": 0.0008,
"loss": 1.9789668321609497,
"step": 9396
},
{
"epoch": 0.5695757575757576,
"grad_norm": 0.04373488202691078,
"learning_rate": 0.0008,
"loss": 2.104863405227661,
"step": 9398
},
{
"epoch": 0.5696969696969697,
"grad_norm": 0.05360971391201019,
"learning_rate": 0.0008,
"loss": 2.00209903717041,
"step": 9400
},
{
"epoch": 0.5698181818181818,
"grad_norm": 0.05002213269472122,
"learning_rate": 0.0008,
"loss": 2.024155378341675,
"step": 9402
},
{
"epoch": 0.569939393939394,
"grad_norm": 0.04739605262875557,
"learning_rate": 0.0008,
"loss": 1.998848557472229,
"step": 9404
},
{
"epoch": 0.570060606060606,
"grad_norm": 0.047317370772361755,
"learning_rate": 0.0008,
"loss": 2.078568935394287,
"step": 9406
},
{
"epoch": 0.5701818181818182,
"grad_norm": 0.043966054916381836,
"learning_rate": 0.0008,
"loss": 2.017091751098633,
"step": 9408
},
{
"epoch": 0.5703030303030303,
"grad_norm": 0.060607511550188065,
"learning_rate": 0.0008,
"loss": 1.993346095085144,
"step": 9410
},
{
"epoch": 0.5704242424242424,
"grad_norm": 0.05267557129263878,
"learning_rate": 0.0008,
"loss": 2.063122034072876,
"step": 9412
},
{
"epoch": 0.5705454545454546,
"grad_norm": 0.05445079877972603,
"learning_rate": 0.0008,
"loss": 2.0728232860565186,
"step": 9414
},
{
"epoch": 0.5706666666666667,
"grad_norm": 0.05232739448547363,
"learning_rate": 0.0008,
"loss": 2.0359811782836914,
"step": 9416
},
{
"epoch": 0.5707878787878787,
"grad_norm": 0.044943612068891525,
"learning_rate": 0.0008,
"loss": 1.9807909727096558,
"step": 9418
},
{
"epoch": 0.5709090909090909,
"grad_norm": 0.04189455881714821,
"learning_rate": 0.0008,
"loss": 1.987267255783081,
"step": 9420
},
{
"epoch": 0.571030303030303,
"grad_norm": 0.0416729710996151,
"learning_rate": 0.0008,
"loss": 2.0115957260131836,
"step": 9422
},
{
"epoch": 0.5711515151515152,
"grad_norm": 0.04375241696834564,
"learning_rate": 0.0008,
"loss": 2.0352461338043213,
"step": 9424
},
{
"epoch": 0.5712727272727273,
"grad_norm": 0.043386295437812805,
"learning_rate": 0.0008,
"loss": 2.0749433040618896,
"step": 9426
},
{
"epoch": 0.5713939393939393,
"grad_norm": 0.04206737130880356,
"learning_rate": 0.0008,
"loss": 1.9775725603103638,
"step": 9428
},
{
"epoch": 0.5715151515151515,
"grad_norm": 0.04310179874300957,
"learning_rate": 0.0008,
"loss": 1.9849467277526855,
"step": 9430
},
{
"epoch": 0.5716363636363636,
"grad_norm": 0.04170616716146469,
"learning_rate": 0.0008,
"loss": 1.9882615804672241,
"step": 9432
},
{
"epoch": 0.5717575757575758,
"grad_norm": 0.042559247463941574,
"learning_rate": 0.0008,
"loss": 2.0011870861053467,
"step": 9434
},
{
"epoch": 0.5718787878787879,
"grad_norm": 0.044815294444561005,
"learning_rate": 0.0008,
"loss": 2.0524654388427734,
"step": 9436
},
{
"epoch": 0.572,
"grad_norm": 0.04772842675447464,
"learning_rate": 0.0008,
"loss": 2.08249831199646,
"step": 9438
},
{
"epoch": 0.5721212121212121,
"grad_norm": 0.05073649063706398,
"learning_rate": 0.0008,
"loss": 2.0823895931243896,
"step": 9440
},
{
"epoch": 0.5722424242424242,
"grad_norm": 0.05055997520685196,
"learning_rate": 0.0008,
"loss": 2.0181961059570312,
"step": 9442
},
{
"epoch": 0.5723636363636364,
"grad_norm": 0.04555310308933258,
"learning_rate": 0.0008,
"loss": 2.0268056392669678,
"step": 9444
},
{
"epoch": 0.5724848484848485,
"grad_norm": 0.04373890534043312,
"learning_rate": 0.0008,
"loss": 2.0834012031555176,
"step": 9446
},
{
"epoch": 0.5726060606060606,
"grad_norm": 0.042812660336494446,
"learning_rate": 0.0008,
"loss": 2.078707218170166,
"step": 9448
},
{
"epoch": 0.5727272727272728,
"grad_norm": 0.046797994524240494,
"learning_rate": 0.0008,
"loss": 2.0658888816833496,
"step": 9450
},
{
"epoch": 0.5728484848484848,
"grad_norm": 0.054780732840299606,
"learning_rate": 0.0008,
"loss": 2.0231618881225586,
"step": 9452
},
{
"epoch": 0.572969696969697,
"grad_norm": 0.04568479582667351,
"learning_rate": 0.0008,
"loss": 2.069403648376465,
"step": 9454
},
{
"epoch": 0.5730909090909091,
"grad_norm": 0.042833246290683746,
"learning_rate": 0.0008,
"loss": 1.9737075567245483,
"step": 9456
},
{
"epoch": 0.5732121212121212,
"grad_norm": 0.044294558465480804,
"learning_rate": 0.0008,
"loss": 2.0791738033294678,
"step": 9458
},
{
"epoch": 0.5733333333333334,
"grad_norm": 0.043068304657936096,
"learning_rate": 0.0008,
"loss": 2.100771903991699,
"step": 9460
},
{
"epoch": 0.5734545454545454,
"grad_norm": 0.050348203629255295,
"learning_rate": 0.0008,
"loss": 2.129054307937622,
"step": 9462
},
{
"epoch": 0.5735757575757576,
"grad_norm": 0.04462176188826561,
"learning_rate": 0.0008,
"loss": 2.0503954887390137,
"step": 9464
},
{
"epoch": 0.5736969696969697,
"grad_norm": 0.04174726456403732,
"learning_rate": 0.0008,
"loss": 1.991123914718628,
"step": 9466
},
{
"epoch": 0.5738181818181818,
"grad_norm": 0.043989889323711395,
"learning_rate": 0.0008,
"loss": 2.071826457977295,
"step": 9468
},
{
"epoch": 0.573939393939394,
"grad_norm": 0.04269234091043472,
"learning_rate": 0.0008,
"loss": 1.994990587234497,
"step": 9470
},
{
"epoch": 0.574060606060606,
"grad_norm": 0.04638609290122986,
"learning_rate": 0.0008,
"loss": 2.0250661373138428,
"step": 9472
},
{
"epoch": 0.5741818181818181,
"grad_norm": 0.0433814600110054,
"learning_rate": 0.0008,
"loss": 2.0293500423431396,
"step": 9474
},
{
"epoch": 0.5743030303030303,
"grad_norm": 0.05210094898939133,
"learning_rate": 0.0008,
"loss": 2.100003957748413,
"step": 9476
},
{
"epoch": 0.5744242424242424,
"grad_norm": 0.04710717126727104,
"learning_rate": 0.0008,
"loss": 2.031658172607422,
"step": 9478
},
{
"epoch": 0.5745454545454546,
"grad_norm": 0.045682456344366074,
"learning_rate": 0.0008,
"loss": 2.082179307937622,
"step": 9480
},
{
"epoch": 0.5746666666666667,
"grad_norm": 0.04489285871386528,
"learning_rate": 0.0008,
"loss": 2.0468668937683105,
"step": 9482
},
{
"epoch": 0.5747878787878787,
"grad_norm": 0.04231232777237892,
"learning_rate": 0.0008,
"loss": 2.0962231159210205,
"step": 9484
},
{
"epoch": 0.5749090909090909,
"grad_norm": 0.04353361576795578,
"learning_rate": 0.0008,
"loss": 2.025364637374878,
"step": 9486
},
{
"epoch": 0.575030303030303,
"grad_norm": 0.047301314771175385,
"learning_rate": 0.0008,
"loss": 2.0638504028320312,
"step": 9488
},
{
"epoch": 0.5751515151515152,
"grad_norm": 0.043390050530433655,
"learning_rate": 0.0008,
"loss": 1.947954773902893,
"step": 9490
},
{
"epoch": 0.5752727272727273,
"grad_norm": 0.0434449166059494,
"learning_rate": 0.0008,
"loss": 2.001662015914917,
"step": 9492
},
{
"epoch": 0.5753939393939393,
"grad_norm": 0.04240219295024872,
"learning_rate": 0.0008,
"loss": 1.9934643507003784,
"step": 9494
},
{
"epoch": 0.5755151515151515,
"grad_norm": 0.04436623677611351,
"learning_rate": 0.0008,
"loss": 2.0386290550231934,
"step": 9496
},
{
"epoch": 0.5756363636363636,
"grad_norm": 0.0458361990749836,
"learning_rate": 0.0008,
"loss": 2.04075026512146,
"step": 9498
},
{
"epoch": 0.5757575757575758,
"grad_norm": 0.042982134968042374,
"learning_rate": 0.0008,
"loss": 2.0211679935455322,
"step": 9500
},
{
"epoch": 0.5758787878787879,
"grad_norm": 0.04677689075469971,
"learning_rate": 0.0008,
"loss": 2.0564236640930176,
"step": 9502
},
{
"epoch": 0.576,
"grad_norm": 0.04695945605635643,
"learning_rate": 0.0008,
"loss": 2.063922643661499,
"step": 9504
},
{
"epoch": 0.5761212121212121,
"grad_norm": 0.043071165680885315,
"learning_rate": 0.0008,
"loss": 1.9933022260665894,
"step": 9506
},
{
"epoch": 0.5762424242424242,
"grad_norm": 0.04110961779952049,
"learning_rate": 0.0008,
"loss": 2.090001344680786,
"step": 9508
},
{
"epoch": 0.5763636363636364,
"grad_norm": 0.04160899668931961,
"learning_rate": 0.0008,
"loss": 1.9885045289993286,
"step": 9510
},
{
"epoch": 0.5764848484848485,
"grad_norm": 0.041946299374103546,
"learning_rate": 0.0008,
"loss": 2.055490493774414,
"step": 9512
},
{
"epoch": 0.5766060606060606,
"grad_norm": 0.04259805008769035,
"learning_rate": 0.0008,
"loss": 2.019848346710205,
"step": 9514
},
{
"epoch": 0.5767272727272728,
"grad_norm": 0.04254680126905441,
"learning_rate": 0.0008,
"loss": 2.0497219562530518,
"step": 9516
},
{
"epoch": 0.5768484848484848,
"grad_norm": 0.042381320148706436,
"learning_rate": 0.0008,
"loss": 2.0705583095550537,
"step": 9518
},
{
"epoch": 0.576969696969697,
"grad_norm": 0.043551623821258545,
"learning_rate": 0.0008,
"loss": 2.099128246307373,
"step": 9520
},
{
"epoch": 0.5770909090909091,
"grad_norm": 0.04293908551335335,
"learning_rate": 0.0008,
"loss": 1.9853276014328003,
"step": 9522
},
{
"epoch": 0.5772121212121212,
"grad_norm": 0.04705501347780228,
"learning_rate": 0.0008,
"loss": 2.020448684692383,
"step": 9524
},
{
"epoch": 0.5773333333333334,
"grad_norm": 0.043071597814559937,
"learning_rate": 0.0008,
"loss": 2.0605149269104004,
"step": 9526
},
{
"epoch": 0.5774545454545454,
"grad_norm": 0.04256204888224602,
"learning_rate": 0.0008,
"loss": 2.06734561920166,
"step": 9528
},
{
"epoch": 0.5775757575757576,
"grad_norm": 0.04568665474653244,
"learning_rate": 0.0008,
"loss": 2.0641095638275146,
"step": 9530
},
{
"epoch": 0.5776969696969697,
"grad_norm": 0.04708019644021988,
"learning_rate": 0.0008,
"loss": 2.049644947052002,
"step": 9532
},
{
"epoch": 0.5778181818181818,
"grad_norm": 0.04305180162191391,
"learning_rate": 0.0008,
"loss": 2.0466816425323486,
"step": 9534
},
{
"epoch": 0.577939393939394,
"grad_norm": 0.0492180660367012,
"learning_rate": 0.0008,
"loss": 2.0051047801971436,
"step": 9536
},
{
"epoch": 0.578060606060606,
"grad_norm": 0.05676954612135887,
"learning_rate": 0.0008,
"loss": 2.0399789810180664,
"step": 9538
},
{
"epoch": 0.5781818181818181,
"grad_norm": 0.0460364930331707,
"learning_rate": 0.0008,
"loss": 2.0246195793151855,
"step": 9540
},
{
"epoch": 0.5783030303030303,
"grad_norm": 0.048556700348854065,
"learning_rate": 0.0008,
"loss": 1.9755362272262573,
"step": 9542
},
{
"epoch": 0.5784242424242424,
"grad_norm": 0.044494807720184326,
"learning_rate": 0.0008,
"loss": 2.0079855918884277,
"step": 9544
},
{
"epoch": 0.5785454545454546,
"grad_norm": 0.04164673015475273,
"learning_rate": 0.0008,
"loss": 2.0278685092926025,
"step": 9546
},
{
"epoch": 0.5786666666666667,
"grad_norm": 0.04225900396704674,
"learning_rate": 0.0008,
"loss": 1.997248888015747,
"step": 9548
},
{
"epoch": 0.5787878787878787,
"grad_norm": 0.04143327474594116,
"learning_rate": 0.0008,
"loss": 2.040236711502075,
"step": 9550
},
{
"epoch": 0.5789090909090909,
"grad_norm": 0.048590097576379776,
"learning_rate": 0.0008,
"loss": 2.016940116882324,
"step": 9552
},
{
"epoch": 0.579030303030303,
"grad_norm": 0.048966627568006516,
"learning_rate": 0.0008,
"loss": 2.072272539138794,
"step": 9554
},
{
"epoch": 0.5791515151515152,
"grad_norm": 0.04815008491277695,
"learning_rate": 0.0008,
"loss": 2.0104897022247314,
"step": 9556
},
{
"epoch": 0.5792727272727273,
"grad_norm": 0.0462520495057106,
"learning_rate": 0.0008,
"loss": 2.061641216278076,
"step": 9558
},
{
"epoch": 0.5793939393939394,
"grad_norm": 0.04513523727655411,
"learning_rate": 0.0008,
"loss": 1.952101230621338,
"step": 9560
},
{
"epoch": 0.5795151515151515,
"grad_norm": 0.04108847305178642,
"learning_rate": 0.0008,
"loss": 2.017068386077881,
"step": 9562
},
{
"epoch": 0.5796363636363636,
"grad_norm": 0.04362650215625763,
"learning_rate": 0.0008,
"loss": 2.0426976680755615,
"step": 9564
},
{
"epoch": 0.5797575757575758,
"grad_norm": 0.040708769112825394,
"learning_rate": 0.0008,
"loss": 2.0407392978668213,
"step": 9566
},
{
"epoch": 0.5798787878787879,
"grad_norm": 0.04211296886205673,
"learning_rate": 0.0008,
"loss": 2.0756402015686035,
"step": 9568
},
{
"epoch": 0.58,
"grad_norm": 0.04062624275684357,
"learning_rate": 0.0008,
"loss": 2.0053062438964844,
"step": 9570
},
{
"epoch": 0.5801212121212121,
"grad_norm": 0.04043220728635788,
"learning_rate": 0.0008,
"loss": 2.071294069290161,
"step": 9572
},
{
"epoch": 0.5802424242424242,
"grad_norm": 0.04836738854646683,
"learning_rate": 0.0008,
"loss": 2.0054965019226074,
"step": 9574
},
{
"epoch": 0.5803636363636364,
"grad_norm": 0.044159241020679474,
"learning_rate": 0.0008,
"loss": 2.053959608078003,
"step": 9576
},
{
"epoch": 0.5804848484848485,
"grad_norm": 0.042257800698280334,
"learning_rate": 0.0008,
"loss": 2.062903642654419,
"step": 9578
},
{
"epoch": 0.5806060606060606,
"grad_norm": 0.04298780858516693,
"learning_rate": 0.0008,
"loss": 2.0252685546875,
"step": 9580
},
{
"epoch": 0.5807272727272728,
"grad_norm": 0.04229258745908737,
"learning_rate": 0.0008,
"loss": 1.9949640035629272,
"step": 9582
},
{
"epoch": 0.5808484848484848,
"grad_norm": 0.04200108349323273,
"learning_rate": 0.0008,
"loss": 2.1047110557556152,
"step": 9584
},
{
"epoch": 0.580969696969697,
"grad_norm": 0.04912323132157326,
"learning_rate": 0.0008,
"loss": 2.0734550952911377,
"step": 9586
},
{
"epoch": 0.5810909090909091,
"grad_norm": 0.04357565939426422,
"learning_rate": 0.0008,
"loss": 2.0510356426239014,
"step": 9588
},
{
"epoch": 0.5812121212121212,
"grad_norm": 0.047415222972631454,
"learning_rate": 0.0008,
"loss": 1.9920896291732788,
"step": 9590
},
{
"epoch": 0.5813333333333334,
"grad_norm": 0.043416548520326614,
"learning_rate": 0.0008,
"loss": 1.9793879985809326,
"step": 9592
},
{
"epoch": 0.5814545454545454,
"grad_norm": 0.0584261529147625,
"learning_rate": 0.0008,
"loss": 2.050306558609009,
"step": 9594
},
{
"epoch": 0.5815757575757575,
"grad_norm": 0.04963228851556778,
"learning_rate": 0.0008,
"loss": 2.0105717182159424,
"step": 9596
},
{
"epoch": 0.5816969696969697,
"grad_norm": 0.044069789350032806,
"learning_rate": 0.0008,
"loss": 2.017712354660034,
"step": 9598
},
{
"epoch": 0.5818181818181818,
"grad_norm": 0.04271155595779419,
"learning_rate": 0.0008,
"loss": 2.0414721965789795,
"step": 9600
},
{
"epoch": 0.581939393939394,
"grad_norm": 0.04194570332765579,
"learning_rate": 0.0008,
"loss": 1.967421531677246,
"step": 9602
},
{
"epoch": 0.5820606060606061,
"grad_norm": 0.04179735481739044,
"learning_rate": 0.0008,
"loss": 2.017085313796997,
"step": 9604
},
{
"epoch": 0.5821818181818181,
"grad_norm": 0.052576642483472824,
"learning_rate": 0.0008,
"loss": 2.1044604778289795,
"step": 9606
},
{
"epoch": 0.5823030303030303,
"grad_norm": 0.04098612442612648,
"learning_rate": 0.0008,
"loss": 2.0146379470825195,
"step": 9608
},
{
"epoch": 0.5824242424242424,
"grad_norm": 0.042695410549640656,
"learning_rate": 0.0008,
"loss": 2.0501232147216797,
"step": 9610
},
{
"epoch": 0.5825454545454546,
"grad_norm": 0.04244145378470421,
"learning_rate": 0.0008,
"loss": 2.119382381439209,
"step": 9612
},
{
"epoch": 0.5826666666666667,
"grad_norm": 0.042726293206214905,
"learning_rate": 0.0008,
"loss": 2.0093307495117188,
"step": 9614
},
{
"epoch": 0.5827878787878787,
"grad_norm": 0.04199203848838806,
"learning_rate": 0.0008,
"loss": 2.050337314605713,
"step": 9616
},
{
"epoch": 0.5829090909090909,
"grad_norm": 0.05191028490662575,
"learning_rate": 0.0008,
"loss": 2.0717151165008545,
"step": 9618
},
{
"epoch": 0.583030303030303,
"grad_norm": 0.0406339131295681,
"learning_rate": 0.0008,
"loss": 2.004110097885132,
"step": 9620
},
{
"epoch": 0.5831515151515152,
"grad_norm": 0.04079688712954521,
"learning_rate": 0.0008,
"loss": 2.0856363773345947,
"step": 9622
},
{
"epoch": 0.5832727272727273,
"grad_norm": 0.042057570070028305,
"learning_rate": 0.0008,
"loss": 1.9967947006225586,
"step": 9624
},
{
"epoch": 0.5833939393939394,
"grad_norm": 0.04477166384458542,
"learning_rate": 0.0008,
"loss": 2.0531005859375,
"step": 9626
},
{
"epoch": 0.5835151515151515,
"grad_norm": 0.042878370732069016,
"learning_rate": 0.0008,
"loss": 2.022923707962036,
"step": 9628
},
{
"epoch": 0.5836363636363636,
"grad_norm": 0.04324834421277046,
"learning_rate": 0.0008,
"loss": 2.112607955932617,
"step": 9630
},
{
"epoch": 0.5837575757575758,
"grad_norm": 0.04741819202899933,
"learning_rate": 0.0008,
"loss": 2.0385403633117676,
"step": 9632
},
{
"epoch": 0.5838787878787879,
"grad_norm": 0.04259010776877403,
"learning_rate": 0.0008,
"loss": 2.046346664428711,
"step": 9634
},
{
"epoch": 0.584,
"grad_norm": 0.050935547798871994,
"learning_rate": 0.0008,
"loss": 2.051588535308838,
"step": 9636
},
{
"epoch": 0.5841212121212122,
"grad_norm": 0.05044984444975853,
"learning_rate": 0.0008,
"loss": 2.0657410621643066,
"step": 9638
},
{
"epoch": 0.5842424242424242,
"grad_norm": 0.04158443957567215,
"learning_rate": 0.0008,
"loss": 2.022091865539551,
"step": 9640
},
{
"epoch": 0.5843636363636364,
"grad_norm": 0.04499749839305878,
"learning_rate": 0.0008,
"loss": 2.106610059738159,
"step": 9642
},
{
"epoch": 0.5844848484848485,
"grad_norm": 0.04176574945449829,
"learning_rate": 0.0008,
"loss": 1.9731956720352173,
"step": 9644
},
{
"epoch": 0.5846060606060606,
"grad_norm": 0.04849352687597275,
"learning_rate": 0.0008,
"loss": 2.034907341003418,
"step": 9646
},
{
"epoch": 0.5847272727272728,
"grad_norm": 0.04335574433207512,
"learning_rate": 0.0008,
"loss": 2.0194711685180664,
"step": 9648
},
{
"epoch": 0.5848484848484848,
"grad_norm": 0.0415458008646965,
"learning_rate": 0.0008,
"loss": 1.993955135345459,
"step": 9650
},
{
"epoch": 0.5849696969696969,
"grad_norm": 0.04239674657583237,
"learning_rate": 0.0008,
"loss": 1.9564099311828613,
"step": 9652
},
{
"epoch": 0.5850909090909091,
"grad_norm": 0.04545285552740097,
"learning_rate": 0.0008,
"loss": 2.016866683959961,
"step": 9654
},
{
"epoch": 0.5852121212121212,
"grad_norm": 0.04217096045613289,
"learning_rate": 0.0008,
"loss": 2.05489444732666,
"step": 9656
},
{
"epoch": 0.5853333333333334,
"grad_norm": 0.04557541012763977,
"learning_rate": 0.0008,
"loss": 2.095128059387207,
"step": 9658
},
{
"epoch": 0.5854545454545454,
"grad_norm": 0.044309262186288834,
"learning_rate": 0.0008,
"loss": 2.0893232822418213,
"step": 9660
},
{
"epoch": 0.5855757575757575,
"grad_norm": 0.045952651649713516,
"learning_rate": 0.0008,
"loss": 1.9862627983093262,
"step": 9662
},
{
"epoch": 0.5856969696969697,
"grad_norm": 0.041518718004226685,
"learning_rate": 0.0008,
"loss": 1.971205472946167,
"step": 9664
},
{
"epoch": 0.5858181818181818,
"grad_norm": 0.04701179638504982,
"learning_rate": 0.0008,
"loss": 2.073457956314087,
"step": 9666
},
{
"epoch": 0.585939393939394,
"grad_norm": 0.05289539322257042,
"learning_rate": 0.0008,
"loss": 1.9877381324768066,
"step": 9668
},
{
"epoch": 0.5860606060606061,
"grad_norm": 0.04916290566325188,
"learning_rate": 0.0008,
"loss": 1.9656710624694824,
"step": 9670
},
{
"epoch": 0.5861818181818181,
"grad_norm": 0.05116012319922447,
"learning_rate": 0.0008,
"loss": 2.1118507385253906,
"step": 9672
},
{
"epoch": 0.5863030303030303,
"grad_norm": 0.04666770249605179,
"learning_rate": 0.0008,
"loss": 2.066478729248047,
"step": 9674
},
{
"epoch": 0.5864242424242424,
"grad_norm": 0.0504104308784008,
"learning_rate": 0.0008,
"loss": 2.0609753131866455,
"step": 9676
},
{
"epoch": 0.5865454545454546,
"grad_norm": 0.0649639144539833,
"learning_rate": 0.0008,
"loss": 2.1089491844177246,
"step": 9678
},
{
"epoch": 0.5866666666666667,
"grad_norm": 0.07170930504798889,
"learning_rate": 0.0008,
"loss": 2.088178873062134,
"step": 9680
},
{
"epoch": 0.5867878787878787,
"grad_norm": 0.06233922392129898,
"learning_rate": 0.0008,
"loss": 2.0857036113739014,
"step": 9682
},
{
"epoch": 0.5869090909090909,
"grad_norm": 0.06017894670367241,
"learning_rate": 0.0008,
"loss": 1.9936378002166748,
"step": 9684
},
{
"epoch": 0.587030303030303,
"grad_norm": 0.06288101524114609,
"learning_rate": 0.0008,
"loss": 2.009002208709717,
"step": 9686
},
{
"epoch": 0.5871515151515152,
"grad_norm": 0.05411646515130997,
"learning_rate": 0.0008,
"loss": 2.051553726196289,
"step": 9688
},
{
"epoch": 0.5872727272727273,
"grad_norm": 0.05095900222659111,
"learning_rate": 0.0008,
"loss": 2.0437939167022705,
"step": 9690
},
{
"epoch": 0.5873939393939394,
"grad_norm": 0.04790781810879707,
"learning_rate": 0.0008,
"loss": 2.0590178966522217,
"step": 9692
},
{
"epoch": 0.5875151515151515,
"grad_norm": 0.047608789056539536,
"learning_rate": 0.0008,
"loss": 2.0241949558258057,
"step": 9694
},
{
"epoch": 0.5876363636363636,
"grad_norm": 0.04575788974761963,
"learning_rate": 0.0008,
"loss": 2.057347536087036,
"step": 9696
},
{
"epoch": 0.5877575757575758,
"grad_norm": 0.05414172634482384,
"learning_rate": 0.0008,
"loss": 2.037694215774536,
"step": 9698
},
{
"epoch": 0.5878787878787879,
"grad_norm": 0.05658893287181854,
"learning_rate": 0.0008,
"loss": 2.013685464859009,
"step": 9700
},
{
"epoch": 0.588,
"grad_norm": 0.05206071212887764,
"learning_rate": 0.0008,
"loss": 2.0941174030303955,
"step": 9702
},
{
"epoch": 0.5881212121212122,
"grad_norm": 0.049880895763635635,
"learning_rate": 0.0008,
"loss": 2.0491907596588135,
"step": 9704
},
{
"epoch": 0.5882424242424242,
"grad_norm": 0.04149235785007477,
"learning_rate": 0.0008,
"loss": 2.076054811477661,
"step": 9706
},
{
"epoch": 0.5883636363636363,
"grad_norm": 0.04372473433613777,
"learning_rate": 0.0008,
"loss": 2.0679500102996826,
"step": 9708
},
{
"epoch": 0.5884848484848485,
"grad_norm": 0.04173668101429939,
"learning_rate": 0.0008,
"loss": 2.044388771057129,
"step": 9710
},
{
"epoch": 0.5886060606060606,
"grad_norm": 0.042364321649074554,
"learning_rate": 0.0008,
"loss": 2.0516231060028076,
"step": 9712
},
{
"epoch": 0.5887272727272728,
"grad_norm": 0.041862402111291885,
"learning_rate": 0.0008,
"loss": 2.003577947616577,
"step": 9714
},
{
"epoch": 0.5888484848484848,
"grad_norm": 0.042283594608306885,
"learning_rate": 0.0008,
"loss": 2.1050567626953125,
"step": 9716
},
{
"epoch": 0.5889696969696969,
"grad_norm": 0.049776963889598846,
"learning_rate": 0.0008,
"loss": 2.006413698196411,
"step": 9718
},
{
"epoch": 0.5890909090909091,
"grad_norm": 0.04789600148797035,
"learning_rate": 0.0008,
"loss": 1.98939049243927,
"step": 9720
},
{
"epoch": 0.5892121212121212,
"grad_norm": 0.04150210693478584,
"learning_rate": 0.0008,
"loss": 2.0585920810699463,
"step": 9722
},
{
"epoch": 0.5893333333333334,
"grad_norm": 0.040252141654491425,
"learning_rate": 0.0008,
"loss": 2.1206445693969727,
"step": 9724
},
{
"epoch": 0.5894545454545455,
"grad_norm": 0.04338102042675018,
"learning_rate": 0.0008,
"loss": 1.9981904029846191,
"step": 9726
},
{
"epoch": 0.5895757575757575,
"grad_norm": 0.043370891362428665,
"learning_rate": 0.0008,
"loss": 2.09861159324646,
"step": 9728
},
{
"epoch": 0.5896969696969697,
"grad_norm": 0.040715042501688004,
"learning_rate": 0.0008,
"loss": 2.002056837081909,
"step": 9730
},
{
"epoch": 0.5898181818181818,
"grad_norm": 0.04180682450532913,
"learning_rate": 0.0008,
"loss": 2.067734956741333,
"step": 9732
},
{
"epoch": 0.589939393939394,
"grad_norm": 0.042770255357027054,
"learning_rate": 0.0008,
"loss": 2.06711483001709,
"step": 9734
},
{
"epoch": 0.5900606060606061,
"grad_norm": 0.04303314909338951,
"learning_rate": 0.0008,
"loss": 2.0428242683410645,
"step": 9736
},
{
"epoch": 0.5901818181818181,
"grad_norm": 0.04391847550868988,
"learning_rate": 0.0008,
"loss": 2.0542705059051514,
"step": 9738
},
{
"epoch": 0.5903030303030303,
"grad_norm": 0.04399926960468292,
"learning_rate": 0.0008,
"loss": 1.9527674913406372,
"step": 9740
},
{
"epoch": 0.5904242424242424,
"grad_norm": 0.040764763951301575,
"learning_rate": 0.0008,
"loss": 2.071773052215576,
"step": 9742
},
{
"epoch": 0.5905454545454546,
"grad_norm": 0.04192975163459778,
"learning_rate": 0.0008,
"loss": 2.087380886077881,
"step": 9744
},
{
"epoch": 0.5906666666666667,
"grad_norm": 0.049099527299404144,
"learning_rate": 0.0008,
"loss": 2.0372729301452637,
"step": 9746
},
{
"epoch": 0.5907878787878788,
"grad_norm": 0.05963549017906189,
"learning_rate": 0.0008,
"loss": 2.0028958320617676,
"step": 9748
},
{
"epoch": 0.5909090909090909,
"grad_norm": 0.05643559992313385,
"learning_rate": 0.0008,
"loss": 2.045304536819458,
"step": 9750
},
{
"epoch": 0.591030303030303,
"grad_norm": 0.05542367696762085,
"learning_rate": 0.0008,
"loss": 2.0606637001037598,
"step": 9752
},
{
"epoch": 0.5911515151515152,
"grad_norm": 0.054836172610521317,
"learning_rate": 0.0008,
"loss": 2.0195300579071045,
"step": 9754
},
{
"epoch": 0.5912727272727273,
"grad_norm": 0.06857011467218399,
"learning_rate": 0.0008,
"loss": 2.0540268421173096,
"step": 9756
},
{
"epoch": 0.5913939393939394,
"grad_norm": 0.05692942813038826,
"learning_rate": 0.0008,
"loss": 2.071040630340576,
"step": 9758
},
{
"epoch": 0.5915151515151515,
"grad_norm": 0.05699079856276512,
"learning_rate": 0.0008,
"loss": 2.027756929397583,
"step": 9760
},
{
"epoch": 0.5916363636363636,
"grad_norm": 0.05637818202376366,
"learning_rate": 0.0008,
"loss": 2.0720977783203125,
"step": 9762
},
{
"epoch": 0.5917575757575757,
"grad_norm": 0.05399816483259201,
"learning_rate": 0.0008,
"loss": 2.046163320541382,
"step": 9764
},
{
"epoch": 0.5918787878787879,
"grad_norm": 0.05548634007573128,
"learning_rate": 0.0008,
"loss": 2.0405776500701904,
"step": 9766
},
{
"epoch": 0.592,
"grad_norm": 0.053244948387145996,
"learning_rate": 0.0008,
"loss": 2.0591750144958496,
"step": 9768
},
{
"epoch": 0.5921212121212122,
"grad_norm": 0.04729587212204933,
"learning_rate": 0.0008,
"loss": 2.0562686920166016,
"step": 9770
},
{
"epoch": 0.5922424242424242,
"grad_norm": 0.049359094351530075,
"learning_rate": 0.0008,
"loss": 2.0442662239074707,
"step": 9772
},
{
"epoch": 0.5923636363636363,
"grad_norm": 0.04695166274905205,
"learning_rate": 0.0008,
"loss": 2.0776524543762207,
"step": 9774
},
{
"epoch": 0.5924848484848485,
"grad_norm": 0.05447431281208992,
"learning_rate": 0.0008,
"loss": 2.0487935543060303,
"step": 9776
},
{
"epoch": 0.5926060606060606,
"grad_norm": 0.04477718099951744,
"learning_rate": 0.0008,
"loss": 1.9919934272766113,
"step": 9778
},
{
"epoch": 0.5927272727272728,
"grad_norm": 0.04094836115837097,
"learning_rate": 0.0008,
"loss": 2.0207533836364746,
"step": 9780
},
{
"epoch": 0.5928484848484848,
"grad_norm": 0.04160090163350105,
"learning_rate": 0.0008,
"loss": 2.010892391204834,
"step": 9782
},
{
"epoch": 0.5929696969696969,
"grad_norm": 0.04261619970202446,
"learning_rate": 0.0008,
"loss": 2.0163183212280273,
"step": 9784
},
{
"epoch": 0.5930909090909091,
"grad_norm": 0.042786166071891785,
"learning_rate": 0.0008,
"loss": 2.034501791000366,
"step": 9786
},
{
"epoch": 0.5932121212121212,
"grad_norm": 0.041133616119623184,
"learning_rate": 0.0008,
"loss": 2.023550033569336,
"step": 9788
},
{
"epoch": 0.5933333333333334,
"grad_norm": 0.04214499890804291,
"learning_rate": 0.0008,
"loss": 2.028643846511841,
"step": 9790
},
{
"epoch": 0.5934545454545455,
"grad_norm": 0.04118794575333595,
"learning_rate": 0.0008,
"loss": 2.0059218406677246,
"step": 9792
},
{
"epoch": 0.5935757575757575,
"grad_norm": 0.04559580609202385,
"learning_rate": 0.0008,
"loss": 2.0441017150878906,
"step": 9794
},
{
"epoch": 0.5936969696969697,
"grad_norm": 0.04639052227139473,
"learning_rate": 0.0008,
"loss": 2.078183174133301,
"step": 9796
},
{
"epoch": 0.5938181818181818,
"grad_norm": 0.04781967028975487,
"learning_rate": 0.0008,
"loss": 2.0931568145751953,
"step": 9798
},
{
"epoch": 0.593939393939394,
"grad_norm": 0.04122455045580864,
"learning_rate": 0.0008,
"loss": 2.040112257003784,
"step": 9800
},
{
"epoch": 0.5940606060606061,
"grad_norm": 0.041246362030506134,
"learning_rate": 0.0008,
"loss": 2.1053621768951416,
"step": 9802
},
{
"epoch": 0.5941818181818181,
"grad_norm": 0.0428580641746521,
"learning_rate": 0.0008,
"loss": 2.0720279216766357,
"step": 9804
},
{
"epoch": 0.5943030303030303,
"grad_norm": 0.047497425228357315,
"learning_rate": 0.0008,
"loss": 2.011019229888916,
"step": 9806
},
{
"epoch": 0.5944242424242424,
"grad_norm": 0.042364660650491714,
"learning_rate": 0.0008,
"loss": 2.0990777015686035,
"step": 9808
},
{
"epoch": 0.5945454545454546,
"grad_norm": 0.04529521241784096,
"learning_rate": 0.0008,
"loss": 2.025174379348755,
"step": 9810
},
{
"epoch": 0.5946666666666667,
"grad_norm": 0.04137659817934036,
"learning_rate": 0.0008,
"loss": 2.00514554977417,
"step": 9812
},
{
"epoch": 0.5947878787878788,
"grad_norm": 0.04561348631978035,
"learning_rate": 0.0008,
"loss": 2.024024724960327,
"step": 9814
},
{
"epoch": 0.5949090909090909,
"grad_norm": 0.04185249283909798,
"learning_rate": 0.0008,
"loss": 2.0130198001861572,
"step": 9816
},
{
"epoch": 0.595030303030303,
"grad_norm": 0.041212234646081924,
"learning_rate": 0.0008,
"loss": 2.1089518070220947,
"step": 9818
},
{
"epoch": 0.5951515151515151,
"grad_norm": 0.04174741357564926,
"learning_rate": 0.0008,
"loss": 2.0315439701080322,
"step": 9820
},
{
"epoch": 0.5952727272727273,
"grad_norm": 0.040285710245370865,
"learning_rate": 0.0008,
"loss": 2.0919578075408936,
"step": 9822
},
{
"epoch": 0.5953939393939394,
"grad_norm": 0.04116203263401985,
"learning_rate": 0.0008,
"loss": 2.0245232582092285,
"step": 9824
},
{
"epoch": 0.5955151515151516,
"grad_norm": 0.04203838109970093,
"learning_rate": 0.0008,
"loss": 2.0785129070281982,
"step": 9826
},
{
"epoch": 0.5956363636363636,
"grad_norm": 0.04074010252952576,
"learning_rate": 0.0008,
"loss": 2.040881633758545,
"step": 9828
},
{
"epoch": 0.5957575757575757,
"grad_norm": 0.04425198584794998,
"learning_rate": 0.0008,
"loss": 2.045701265335083,
"step": 9830
},
{
"epoch": 0.5958787878787879,
"grad_norm": 0.0483483150601387,
"learning_rate": 0.0008,
"loss": 2.005821943283081,
"step": 9832
},
{
"epoch": 0.596,
"grad_norm": 0.04125728830695152,
"learning_rate": 0.0008,
"loss": 2.0040123462677,
"step": 9834
},
{
"epoch": 0.5961212121212122,
"grad_norm": 0.04030376672744751,
"learning_rate": 0.0008,
"loss": 1.9945884943008423,
"step": 9836
},
{
"epoch": 0.5962424242424242,
"grad_norm": 0.041009895503520966,
"learning_rate": 0.0008,
"loss": 2.008159875869751,
"step": 9838
},
{
"epoch": 0.5963636363636363,
"grad_norm": 0.04065202549099922,
"learning_rate": 0.0008,
"loss": 2.0870144367218018,
"step": 9840
},
{
"epoch": 0.5964848484848485,
"grad_norm": 0.04376637563109398,
"learning_rate": 0.0008,
"loss": 2.0757391452789307,
"step": 9842
},
{
"epoch": 0.5966060606060606,
"grad_norm": 0.051688410341739655,
"learning_rate": 0.0008,
"loss": 2.1274473667144775,
"step": 9844
},
{
"epoch": 0.5967272727272728,
"grad_norm": 0.06391502916812897,
"learning_rate": 0.0008,
"loss": 2.031921625137329,
"step": 9846
},
{
"epoch": 0.5968484848484849,
"grad_norm": 0.04927889257669449,
"learning_rate": 0.0008,
"loss": 2.031177520751953,
"step": 9848
},
{
"epoch": 0.5969696969696969,
"grad_norm": 0.04778888821601868,
"learning_rate": 0.0008,
"loss": 2.0500237941741943,
"step": 9850
},
{
"epoch": 0.5970909090909091,
"grad_norm": 0.04378962889313698,
"learning_rate": 0.0008,
"loss": 2.0953898429870605,
"step": 9852
},
{
"epoch": 0.5972121212121212,
"grad_norm": 0.040774013847112656,
"learning_rate": 0.0008,
"loss": 2.052186965942383,
"step": 9854
},
{
"epoch": 0.5973333333333334,
"grad_norm": 0.04104777052998543,
"learning_rate": 0.0008,
"loss": 2.0072500705718994,
"step": 9856
},
{
"epoch": 0.5974545454545455,
"grad_norm": 0.042439695447683334,
"learning_rate": 0.0008,
"loss": 2.0291974544525146,
"step": 9858
},
{
"epoch": 0.5975757575757575,
"grad_norm": 0.04685891792178154,
"learning_rate": 0.0008,
"loss": 2.12577223777771,
"step": 9860
},
{
"epoch": 0.5976969696969697,
"grad_norm": 0.041731420904397964,
"learning_rate": 0.0008,
"loss": 2.0449376106262207,
"step": 9862
},
{
"epoch": 0.5978181818181818,
"grad_norm": 0.04092064127326012,
"learning_rate": 0.0008,
"loss": 2.0443637371063232,
"step": 9864
},
{
"epoch": 0.597939393939394,
"grad_norm": 0.04143928363919258,
"learning_rate": 0.0008,
"loss": 1.9971065521240234,
"step": 9866
},
{
"epoch": 0.5980606060606061,
"grad_norm": 0.044705670326948166,
"learning_rate": 0.0008,
"loss": 2.0796234607696533,
"step": 9868
},
{
"epoch": 0.5981818181818181,
"grad_norm": 0.043604664504528046,
"learning_rate": 0.0008,
"loss": 1.9689793586730957,
"step": 9870
},
{
"epoch": 0.5983030303030303,
"grad_norm": 0.043577760457992554,
"learning_rate": 0.0008,
"loss": 2.0133635997772217,
"step": 9872
},
{
"epoch": 0.5984242424242424,
"grad_norm": 0.04909781739115715,
"learning_rate": 0.0008,
"loss": 2.0716493129730225,
"step": 9874
},
{
"epoch": 0.5985454545454545,
"grad_norm": 0.04157146438956261,
"learning_rate": 0.0008,
"loss": 2.029487133026123,
"step": 9876
},
{
"epoch": 0.5986666666666667,
"grad_norm": 0.04353266581892967,
"learning_rate": 0.0008,
"loss": 2.0698940753936768,
"step": 9878
},
{
"epoch": 0.5987878787878788,
"grad_norm": 0.04292137920856476,
"learning_rate": 0.0008,
"loss": 1.9904496669769287,
"step": 9880
},
{
"epoch": 0.598909090909091,
"grad_norm": 0.045555032789707184,
"learning_rate": 0.0008,
"loss": 2.0880837440490723,
"step": 9882
},
{
"epoch": 0.599030303030303,
"grad_norm": 0.04245288297533989,
"learning_rate": 0.0008,
"loss": 2.0751445293426514,
"step": 9884
},
{
"epoch": 0.5991515151515151,
"grad_norm": 0.04103011637926102,
"learning_rate": 0.0008,
"loss": 2.04484486579895,
"step": 9886
},
{
"epoch": 0.5992727272727273,
"grad_norm": 0.04467086121439934,
"learning_rate": 0.0008,
"loss": 1.9803831577301025,
"step": 9888
},
{
"epoch": 0.5993939393939394,
"grad_norm": 0.048204973340034485,
"learning_rate": 0.0008,
"loss": 2.042189836502075,
"step": 9890
},
{
"epoch": 0.5995151515151516,
"grad_norm": 0.05364532396197319,
"learning_rate": 0.0008,
"loss": 2.0298237800598145,
"step": 9892
},
{
"epoch": 0.5996363636363636,
"grad_norm": 0.04522913694381714,
"learning_rate": 0.0008,
"loss": 2.0523722171783447,
"step": 9894
},
{
"epoch": 0.5997575757575757,
"grad_norm": 0.047304876148700714,
"learning_rate": 0.0008,
"loss": 2.1379940509796143,
"step": 9896
},
{
"epoch": 0.5998787878787879,
"grad_norm": 0.041847508400678635,
"learning_rate": 0.0008,
"loss": 2.082672119140625,
"step": 9898
},
{
"epoch": 0.6,
"grad_norm": 0.04546612873673439,
"learning_rate": 0.0008,
"loss": 2.044926404953003,
"step": 9900
},
{
"epoch": 0.6,
"eval_loss": 2.038447380065918,
"eval_runtime": 151.871,
"eval_samples_per_second": 1.646,
"eval_steps_per_second": 0.086,
"step": 9900
},
{
"epoch": 0.6001212121212122,
"grad_norm": 0.04761418700218201,
"learning_rate": 0.0008,
"loss": 2.079401731491089,
"step": 9902
},
{
"epoch": 0.6002424242424242,
"grad_norm": 0.04361937195062637,
"learning_rate": 0.0008,
"loss": 2.0450499057769775,
"step": 9904
},
{
"epoch": 0.6003636363636363,
"grad_norm": 0.04171966016292572,
"learning_rate": 0.0008,
"loss": 2.0545783042907715,
"step": 9906
},
{
"epoch": 0.6004848484848485,
"grad_norm": 0.04353747144341469,
"learning_rate": 0.0008,
"loss": 1.981941819190979,
"step": 9908
},
{
"epoch": 0.6006060606060606,
"grad_norm": 0.041466984897851944,
"learning_rate": 0.0008,
"loss": 2.0795345306396484,
"step": 9910
},
{
"epoch": 0.6007272727272728,
"grad_norm": 0.04318410903215408,
"learning_rate": 0.0008,
"loss": 2.047868251800537,
"step": 9912
},
{
"epoch": 0.6008484848484849,
"grad_norm": 0.04567323997616768,
"learning_rate": 0.0008,
"loss": 2.032005548477173,
"step": 9914
},
{
"epoch": 0.6009696969696969,
"grad_norm": 0.04290834069252014,
"learning_rate": 0.0008,
"loss": 2.089354991912842,
"step": 9916
},
{
"epoch": 0.6010909090909091,
"grad_norm": 0.041721682995557785,
"learning_rate": 0.0008,
"loss": 2.0152502059936523,
"step": 9918
},
{
"epoch": 0.6012121212121212,
"grad_norm": 0.04186512157320976,
"learning_rate": 0.0008,
"loss": 2.043116569519043,
"step": 9920
},
{
"epoch": 0.6013333333333334,
"grad_norm": 0.04443183168768883,
"learning_rate": 0.0008,
"loss": 2.0640792846679688,
"step": 9922
},
{
"epoch": 0.6014545454545455,
"grad_norm": 0.05031351000070572,
"learning_rate": 0.0008,
"loss": 2.064042091369629,
"step": 9924
},
{
"epoch": 0.6015757575757575,
"grad_norm": 0.047969311475753784,
"learning_rate": 0.0008,
"loss": 2.0339505672454834,
"step": 9926
},
{
"epoch": 0.6016969696969697,
"grad_norm": 0.04337681829929352,
"learning_rate": 0.0008,
"loss": 2.027583360671997,
"step": 9928
},
{
"epoch": 0.6018181818181818,
"grad_norm": 0.0434490405023098,
"learning_rate": 0.0008,
"loss": 1.98932945728302,
"step": 9930
},
{
"epoch": 0.6019393939393939,
"grad_norm": 0.04192259535193443,
"learning_rate": 0.0008,
"loss": 2.0375208854675293,
"step": 9932
},
{
"epoch": 0.6020606060606061,
"grad_norm": 0.04512995481491089,
"learning_rate": 0.0008,
"loss": 2.00233793258667,
"step": 9934
},
{
"epoch": 0.6021818181818182,
"grad_norm": 0.05068090558052063,
"learning_rate": 0.0008,
"loss": 2.0741147994995117,
"step": 9936
},
{
"epoch": 0.6023030303030303,
"grad_norm": 0.041162457317113876,
"learning_rate": 0.0008,
"loss": 2.0658161640167236,
"step": 9938
},
{
"epoch": 0.6024242424242424,
"grad_norm": 0.040891822427511215,
"learning_rate": 0.0008,
"loss": 2.0302751064300537,
"step": 9940
},
{
"epoch": 0.6025454545454545,
"grad_norm": 0.042870961129665375,
"learning_rate": 0.0008,
"loss": 2.1005043983459473,
"step": 9942
},
{
"epoch": 0.6026666666666667,
"grad_norm": 0.04919340834021568,
"learning_rate": 0.0008,
"loss": 2.014187812805176,
"step": 9944
},
{
"epoch": 0.6027878787878788,
"grad_norm": 0.042087968438863754,
"learning_rate": 0.0008,
"loss": 2.093320846557617,
"step": 9946
},
{
"epoch": 0.602909090909091,
"grad_norm": 0.0415801927447319,
"learning_rate": 0.0008,
"loss": 2.061854124069214,
"step": 9948
},
{
"epoch": 0.603030303030303,
"grad_norm": 0.04362623766064644,
"learning_rate": 0.0008,
"loss": 2.0270543098449707,
"step": 9950
},
{
"epoch": 0.6031515151515151,
"grad_norm": 0.04184890538454056,
"learning_rate": 0.0008,
"loss": 2.0461277961730957,
"step": 9952
},
{
"epoch": 0.6032727272727273,
"grad_norm": 0.043058186769485474,
"learning_rate": 0.0008,
"loss": 2.141096353530884,
"step": 9954
},
{
"epoch": 0.6033939393939394,
"grad_norm": 0.04449135437607765,
"learning_rate": 0.0008,
"loss": 2.007575035095215,
"step": 9956
},
{
"epoch": 0.6035151515151516,
"grad_norm": 0.0455276258289814,
"learning_rate": 0.0008,
"loss": 2.081150770187378,
"step": 9958
},
{
"epoch": 0.6036363636363636,
"grad_norm": 0.04144424572587013,
"learning_rate": 0.0008,
"loss": 2.041144847869873,
"step": 9960
},
{
"epoch": 0.6037575757575757,
"grad_norm": 0.041155390441417694,
"learning_rate": 0.0008,
"loss": 2.033494710922241,
"step": 9962
},
{
"epoch": 0.6038787878787879,
"grad_norm": 0.04171587899327278,
"learning_rate": 0.0008,
"loss": 2.0908045768737793,
"step": 9964
},
{
"epoch": 0.604,
"grad_norm": 0.04237671196460724,
"learning_rate": 0.0008,
"loss": 2.02854323387146,
"step": 9966
},
{
"epoch": 0.6041212121212122,
"grad_norm": 0.04118337482213974,
"learning_rate": 0.0008,
"loss": 2.0893406867980957,
"step": 9968
},
{
"epoch": 0.6042424242424242,
"grad_norm": 0.041070085018873215,
"learning_rate": 0.0008,
"loss": 1.998357892036438,
"step": 9970
},
{
"epoch": 0.6043636363636363,
"grad_norm": 0.04157288372516632,
"learning_rate": 0.0008,
"loss": 2.0152859687805176,
"step": 9972
},
{
"epoch": 0.6044848484848485,
"grad_norm": 0.04033197462558746,
"learning_rate": 0.0008,
"loss": 2.0470244884490967,
"step": 9974
},
{
"epoch": 0.6046060606060606,
"grad_norm": 0.04007977619767189,
"learning_rate": 0.0008,
"loss": 2.0565884113311768,
"step": 9976
},
{
"epoch": 0.6047272727272728,
"grad_norm": 0.04429515451192856,
"learning_rate": 0.0008,
"loss": 2.084118604660034,
"step": 9978
},
{
"epoch": 0.6048484848484849,
"grad_norm": 0.04474706947803497,
"learning_rate": 0.0008,
"loss": 2.1360580921173096,
"step": 9980
},
{
"epoch": 0.6049696969696969,
"grad_norm": 0.041404955089092255,
"learning_rate": 0.0008,
"loss": 1.9839935302734375,
"step": 9982
},
{
"epoch": 0.6050909090909091,
"grad_norm": 0.0417676605284214,
"learning_rate": 0.0008,
"loss": 2.032862424850464,
"step": 9984
},
{
"epoch": 0.6052121212121212,
"grad_norm": 0.040533293038606644,
"learning_rate": 0.0008,
"loss": 2.082766056060791,
"step": 9986
},
{
"epoch": 0.6053333333333333,
"grad_norm": 0.04095756262540817,
"learning_rate": 0.0008,
"loss": 1.9988423585891724,
"step": 9988
},
{
"epoch": 0.6054545454545455,
"grad_norm": 0.04029757156968117,
"learning_rate": 0.0008,
"loss": 2.072321653366089,
"step": 9990
},
{
"epoch": 0.6055757575757575,
"grad_norm": 0.042986057698726654,
"learning_rate": 0.0008,
"loss": 2.101795196533203,
"step": 9992
},
{
"epoch": 0.6056969696969697,
"grad_norm": 0.04269036278128624,
"learning_rate": 0.0008,
"loss": 2.0463778972625732,
"step": 9994
},
{
"epoch": 0.6058181818181818,
"grad_norm": 0.040230561047792435,
"learning_rate": 0.0008,
"loss": 2.0527801513671875,
"step": 9996
},
{
"epoch": 0.6059393939393939,
"grad_norm": 0.04750789701938629,
"learning_rate": 0.0008,
"loss": 2.0966179370880127,
"step": 9998
},
{
"epoch": 0.6060606060606061,
"grad_norm": 0.05250247195363045,
"learning_rate": 0.0008,
"loss": 2.0886247158050537,
"step": 10000
},
{
"epoch": 0.6061818181818182,
"grad_norm": 0.0442047193646431,
"learning_rate": 0.0008,
"loss": 2.094456911087036,
"step": 10002
},
{
"epoch": 0.6063030303030303,
"grad_norm": 0.042288877069950104,
"learning_rate": 0.0008,
"loss": 2.0955448150634766,
"step": 10004
},
{
"epoch": 0.6064242424242424,
"grad_norm": 0.04061783477663994,
"learning_rate": 0.0008,
"loss": 2.0408987998962402,
"step": 10006
},
{
"epoch": 0.6065454545454545,
"grad_norm": 0.04272190481424332,
"learning_rate": 0.0008,
"loss": 2.062671184539795,
"step": 10008
},
{
"epoch": 0.6066666666666667,
"grad_norm": 0.04089139401912689,
"learning_rate": 0.0008,
"loss": 2.002155303955078,
"step": 10010
},
{
"epoch": 0.6067878787878788,
"grad_norm": 0.04046671465039253,
"learning_rate": 0.0008,
"loss": 2.036890745162964,
"step": 10012
},
{
"epoch": 0.606909090909091,
"grad_norm": 0.03968726471066475,
"learning_rate": 0.0008,
"loss": 2.098478317260742,
"step": 10014
},
{
"epoch": 0.607030303030303,
"grad_norm": 0.04034325107932091,
"learning_rate": 0.0008,
"loss": 2.0409975051879883,
"step": 10016
},
{
"epoch": 0.6071515151515151,
"grad_norm": 0.041690386831760406,
"learning_rate": 0.0008,
"loss": 2.033432722091675,
"step": 10018
},
{
"epoch": 0.6072727272727273,
"grad_norm": 0.04065422713756561,
"learning_rate": 0.0008,
"loss": 2.0782084465026855,
"step": 10020
},
{
"epoch": 0.6073939393939394,
"grad_norm": 0.04225682094693184,
"learning_rate": 0.0008,
"loss": 2.045529365539551,
"step": 10022
},
{
"epoch": 0.6075151515151516,
"grad_norm": 0.04336206987500191,
"learning_rate": 0.0008,
"loss": 2.0780999660491943,
"step": 10024
},
{
"epoch": 0.6076363636363636,
"grad_norm": 0.05207006260752678,
"learning_rate": 0.0008,
"loss": 2.034261703491211,
"step": 10026
},
{
"epoch": 0.6077575757575757,
"grad_norm": 0.06479660421609879,
"learning_rate": 0.0008,
"loss": 2.0315654277801514,
"step": 10028
},
{
"epoch": 0.6078787878787879,
"grad_norm": 0.05585022643208504,
"learning_rate": 0.0008,
"loss": 2.0321226119995117,
"step": 10030
},
{
"epoch": 0.608,
"grad_norm": 0.04448631778359413,
"learning_rate": 0.0008,
"loss": 2.067575454711914,
"step": 10032
},
{
"epoch": 0.6081212121212122,
"grad_norm": 0.042228080332279205,
"learning_rate": 0.0008,
"loss": 2.049743890762329,
"step": 10034
},
{
"epoch": 0.6082424242424243,
"grad_norm": 0.04091258347034454,
"learning_rate": 0.0008,
"loss": 2.0499496459960938,
"step": 10036
},
{
"epoch": 0.6083636363636363,
"grad_norm": 0.045665912330150604,
"learning_rate": 0.0008,
"loss": 2.087883949279785,
"step": 10038
},
{
"epoch": 0.6084848484848485,
"grad_norm": 0.043186724185943604,
"learning_rate": 0.0008,
"loss": 2.0912392139434814,
"step": 10040
},
{
"epoch": 0.6086060606060606,
"grad_norm": 0.045545294880867004,
"learning_rate": 0.0008,
"loss": 2.086946964263916,
"step": 10042
},
{
"epoch": 0.6087272727272727,
"grad_norm": 0.04484128579497337,
"learning_rate": 0.0008,
"loss": 1.995355486869812,
"step": 10044
},
{
"epoch": 0.6088484848484849,
"grad_norm": 0.03926258534193039,
"learning_rate": 0.0008,
"loss": 2.090721368789673,
"step": 10046
},
{
"epoch": 0.6089696969696969,
"grad_norm": 0.04187706485390663,
"learning_rate": 0.0008,
"loss": 2.0198628902435303,
"step": 10048
},
{
"epoch": 0.6090909090909091,
"grad_norm": 0.042292993515729904,
"learning_rate": 0.0008,
"loss": 2.0414326190948486,
"step": 10050
},
{
"epoch": 0.6092121212121212,
"grad_norm": 0.04174136742949486,
"learning_rate": 0.0008,
"loss": 2.0292654037475586,
"step": 10052
},
{
"epoch": 0.6093333333333333,
"grad_norm": 0.040093451738357544,
"learning_rate": 0.0008,
"loss": 2.0646090507507324,
"step": 10054
},
{
"epoch": 0.6094545454545455,
"grad_norm": 0.04752276837825775,
"learning_rate": 0.0008,
"loss": 2.00658917427063,
"step": 10056
},
{
"epoch": 0.6095757575757575,
"grad_norm": 0.04058754816651344,
"learning_rate": 0.0008,
"loss": 2.050006866455078,
"step": 10058
},
{
"epoch": 0.6096969696969697,
"grad_norm": 0.04149288684129715,
"learning_rate": 0.0008,
"loss": 2.09098744392395,
"step": 10060
},
{
"epoch": 0.6098181818181818,
"grad_norm": 0.03969769924879074,
"learning_rate": 0.0008,
"loss": 2.0217552185058594,
"step": 10062
},
{
"epoch": 0.6099393939393939,
"grad_norm": 0.0440235435962677,
"learning_rate": 0.0008,
"loss": 2.054797649383545,
"step": 10064
},
{
"epoch": 0.6100606060606061,
"grad_norm": 0.04636158049106598,
"learning_rate": 0.0008,
"loss": 2.0066237449645996,
"step": 10066
},
{
"epoch": 0.6101818181818182,
"grad_norm": 0.0403071865439415,
"learning_rate": 0.0008,
"loss": 2.0785927772521973,
"step": 10068
},
{
"epoch": 0.6103030303030303,
"grad_norm": 0.05158194899559021,
"learning_rate": 0.0008,
"loss": 2.0543315410614014,
"step": 10070
},
{
"epoch": 0.6104242424242424,
"grad_norm": 0.0709422379732132,
"learning_rate": 0.0008,
"loss": 2.118565559387207,
"step": 10072
},
{
"epoch": 0.6105454545454545,
"grad_norm": 0.05569835379719734,
"learning_rate": 0.0008,
"loss": 2.044252872467041,
"step": 10074
},
{
"epoch": 0.6106666666666667,
"grad_norm": 0.047366801649332047,
"learning_rate": 0.0008,
"loss": 2.057725667953491,
"step": 10076
},
{
"epoch": 0.6107878787878788,
"grad_norm": 0.04149840772151947,
"learning_rate": 0.0008,
"loss": 2.055138111114502,
"step": 10078
},
{
"epoch": 0.610909090909091,
"grad_norm": 0.04389996826648712,
"learning_rate": 0.0008,
"loss": 2.0952348709106445,
"step": 10080
},
{
"epoch": 0.611030303030303,
"grad_norm": 0.042512863874435425,
"learning_rate": 0.0008,
"loss": 1.9981913566589355,
"step": 10082
},
{
"epoch": 0.6111515151515151,
"grad_norm": 0.04143204540014267,
"learning_rate": 0.0008,
"loss": 2.0540387630462646,
"step": 10084
},
{
"epoch": 0.6112727272727273,
"grad_norm": 0.043302297592163086,
"learning_rate": 0.0008,
"loss": 2.0286073684692383,
"step": 10086
},
{
"epoch": 0.6113939393939394,
"grad_norm": 0.04768359288573265,
"learning_rate": 0.0008,
"loss": 2.0622451305389404,
"step": 10088
},
{
"epoch": 0.6115151515151516,
"grad_norm": 0.04860081151127815,
"learning_rate": 0.0008,
"loss": 2.071643590927124,
"step": 10090
},
{
"epoch": 0.6116363636363636,
"grad_norm": 0.04829791560769081,
"learning_rate": 0.0008,
"loss": 2.1087687015533447,
"step": 10092
},
{
"epoch": 0.6117575757575757,
"grad_norm": 0.04974350705742836,
"learning_rate": 0.0008,
"loss": 2.099496603012085,
"step": 10094
},
{
"epoch": 0.6118787878787879,
"grad_norm": 0.04548562690615654,
"learning_rate": 0.0008,
"loss": 2.000756025314331,
"step": 10096
},
{
"epoch": 0.612,
"grad_norm": 0.04679716005921364,
"learning_rate": 0.0008,
"loss": 2.0448923110961914,
"step": 10098
},
{
"epoch": 0.6121212121212121,
"grad_norm": 0.041808806359767914,
"learning_rate": 0.0008,
"loss": 1.9892853498458862,
"step": 10100
},
{
"epoch": 0.6122424242424243,
"grad_norm": 0.04448048397898674,
"learning_rate": 0.0008,
"loss": 2.1180667877197266,
"step": 10102
},
{
"epoch": 0.6123636363636363,
"grad_norm": 0.05378001928329468,
"learning_rate": 0.0008,
"loss": 2.0682032108306885,
"step": 10104
},
{
"epoch": 0.6124848484848485,
"grad_norm": 0.046297211199998856,
"learning_rate": 0.0008,
"loss": 2.0087287425994873,
"step": 10106
},
{
"epoch": 0.6126060606060606,
"grad_norm": 0.04921703040599823,
"learning_rate": 0.0008,
"loss": 2.0583903789520264,
"step": 10108
},
{
"epoch": 0.6127272727272727,
"grad_norm": 0.047016508877277374,
"learning_rate": 0.0008,
"loss": 2.023056983947754,
"step": 10110
},
{
"epoch": 0.6128484848484849,
"grad_norm": 0.04029259830713272,
"learning_rate": 0.0008,
"loss": 2.1234467029571533,
"step": 10112
},
{
"epoch": 0.6129696969696969,
"grad_norm": 0.04926621541380882,
"learning_rate": 0.0008,
"loss": 2.062235116958618,
"step": 10114
},
{
"epoch": 0.6130909090909091,
"grad_norm": 0.0414152666926384,
"learning_rate": 0.0008,
"loss": 2.039320468902588,
"step": 10116
},
{
"epoch": 0.6132121212121212,
"grad_norm": 0.04020259901881218,
"learning_rate": 0.0008,
"loss": 2.0737664699554443,
"step": 10118
},
{
"epoch": 0.6133333333333333,
"grad_norm": 0.039703719317913055,
"learning_rate": 0.0008,
"loss": 2.0766546726226807,
"step": 10120
},
{
"epoch": 0.6134545454545455,
"grad_norm": 0.04445591941475868,
"learning_rate": 0.0008,
"loss": 2.0475170612335205,
"step": 10122
},
{
"epoch": 0.6135757575757576,
"grad_norm": 0.042032089084386826,
"learning_rate": 0.0008,
"loss": 2.063563585281372,
"step": 10124
},
{
"epoch": 0.6136969696969697,
"grad_norm": 0.0407901406288147,
"learning_rate": 0.0008,
"loss": 1.968110203742981,
"step": 10126
},
{
"epoch": 0.6138181818181818,
"grad_norm": 0.04041441157460213,
"learning_rate": 0.0008,
"loss": 2.0927882194519043,
"step": 10128
},
{
"epoch": 0.6139393939393939,
"grad_norm": 0.04385813698172569,
"learning_rate": 0.0008,
"loss": 2.1213037967681885,
"step": 10130
},
{
"epoch": 0.6140606060606061,
"grad_norm": 0.04149911180138588,
"learning_rate": 0.0008,
"loss": 2.108860969543457,
"step": 10132
},
{
"epoch": 0.6141818181818182,
"grad_norm": 0.04066208750009537,
"learning_rate": 0.0008,
"loss": 2.0105414390563965,
"step": 10134
},
{
"epoch": 0.6143030303030304,
"grad_norm": 0.04346756637096405,
"learning_rate": 0.0008,
"loss": 2.0897727012634277,
"step": 10136
},
{
"epoch": 0.6144242424242424,
"grad_norm": 0.04515507072210312,
"learning_rate": 0.0008,
"loss": 2.0217511653900146,
"step": 10138
},
{
"epoch": 0.6145454545454545,
"grad_norm": 0.04749459773302078,
"learning_rate": 0.0008,
"loss": 2.004225730895996,
"step": 10140
},
{
"epoch": 0.6146666666666667,
"grad_norm": 0.05469020828604698,
"learning_rate": 0.0008,
"loss": 2.0392167568206787,
"step": 10142
},
{
"epoch": 0.6147878787878788,
"grad_norm": 0.04892660677433014,
"learning_rate": 0.0008,
"loss": 2.029155969619751,
"step": 10144
},
{
"epoch": 0.614909090909091,
"grad_norm": 0.047131508588790894,
"learning_rate": 0.0008,
"loss": 2.1125712394714355,
"step": 10146
},
{
"epoch": 0.615030303030303,
"grad_norm": 0.04185932129621506,
"learning_rate": 0.0008,
"loss": 2.0732576847076416,
"step": 10148
},
{
"epoch": 0.6151515151515151,
"grad_norm": 0.04206562787294388,
"learning_rate": 0.0008,
"loss": 2.075420379638672,
"step": 10150
},
{
"epoch": 0.6152727272727273,
"grad_norm": 0.047590505331754684,
"learning_rate": 0.0008,
"loss": 2.067148208618164,
"step": 10152
},
{
"epoch": 0.6153939393939394,
"grad_norm": 0.04908321052789688,
"learning_rate": 0.0008,
"loss": 2.0065910816192627,
"step": 10154
},
{
"epoch": 0.6155151515151516,
"grad_norm": 0.04788610711693764,
"learning_rate": 0.0008,
"loss": 2.0565695762634277,
"step": 10156
},
{
"epoch": 0.6156363636363636,
"grad_norm": 0.049474652856588364,
"learning_rate": 0.0008,
"loss": 2.058699369430542,
"step": 10158
},
{
"epoch": 0.6157575757575757,
"grad_norm": 0.06210954114794731,
"learning_rate": 0.0008,
"loss": 2.06677508354187,
"step": 10160
},
{
"epoch": 0.6158787878787879,
"grad_norm": 0.05026642605662346,
"learning_rate": 0.0008,
"loss": 2.0461320877075195,
"step": 10162
},
{
"epoch": 0.616,
"grad_norm": 0.046704281121492386,
"learning_rate": 0.0008,
"loss": 2.082803249359131,
"step": 10164
},
{
"epoch": 0.6161212121212121,
"grad_norm": 0.04569169133901596,
"learning_rate": 0.0008,
"loss": 2.0361838340759277,
"step": 10166
},
{
"epoch": 0.6162424242424243,
"grad_norm": 0.04236941784620285,
"learning_rate": 0.0008,
"loss": 2.122480630874634,
"step": 10168
},
{
"epoch": 0.6163636363636363,
"grad_norm": 0.04226727783679962,
"learning_rate": 0.0008,
"loss": 2.1178503036499023,
"step": 10170
},
{
"epoch": 0.6164848484848485,
"grad_norm": 0.04123258963227272,
"learning_rate": 0.0008,
"loss": 2.0550315380096436,
"step": 10172
},
{
"epoch": 0.6166060606060606,
"grad_norm": 0.04081227257847786,
"learning_rate": 0.0008,
"loss": 2.019474744796753,
"step": 10174
},
{
"epoch": 0.6167272727272727,
"grad_norm": 0.04398832097649574,
"learning_rate": 0.0008,
"loss": 2.0544750690460205,
"step": 10176
},
{
"epoch": 0.6168484848484849,
"grad_norm": 0.04241466522216797,
"learning_rate": 0.0008,
"loss": 2.062290906906128,
"step": 10178
},
{
"epoch": 0.616969696969697,
"grad_norm": 0.0437067449092865,
"learning_rate": 0.0008,
"loss": 2.0494778156280518,
"step": 10180
},
{
"epoch": 0.6170909090909091,
"grad_norm": 0.04332631453871727,
"learning_rate": 0.0008,
"loss": 2.0049898624420166,
"step": 10182
},
{
"epoch": 0.6172121212121212,
"grad_norm": 0.04576027765870094,
"learning_rate": 0.0008,
"loss": 2.06680965423584,
"step": 10184
},
{
"epoch": 0.6173333333333333,
"grad_norm": 0.049179669469594955,
"learning_rate": 0.0008,
"loss": 2.066101551055908,
"step": 10186
},
{
"epoch": 0.6174545454545455,
"grad_norm": 0.05491883307695389,
"learning_rate": 0.0008,
"loss": 2.0469937324523926,
"step": 10188
},
{
"epoch": 0.6175757575757576,
"grad_norm": 0.04904977232217789,
"learning_rate": 0.0008,
"loss": 2.0541627407073975,
"step": 10190
},
{
"epoch": 0.6176969696969697,
"grad_norm": 0.0421115942299366,
"learning_rate": 0.0008,
"loss": 2.0754499435424805,
"step": 10192
},
{
"epoch": 0.6178181818181818,
"grad_norm": 0.04124470055103302,
"learning_rate": 0.0008,
"loss": 2.1043691635131836,
"step": 10194
},
{
"epoch": 0.6179393939393939,
"grad_norm": 0.04123154282569885,
"learning_rate": 0.0008,
"loss": 2.0041720867156982,
"step": 10196
},
{
"epoch": 0.6180606060606061,
"grad_norm": 0.053508564829826355,
"learning_rate": 0.0008,
"loss": 2.0463316440582275,
"step": 10198
},
{
"epoch": 0.6181818181818182,
"grad_norm": 0.04503655061125755,
"learning_rate": 0.0008,
"loss": 2.0800468921661377,
"step": 10200
},
{
"epoch": 0.6183030303030304,
"grad_norm": 0.04404851794242859,
"learning_rate": 0.0008,
"loss": 2.018942356109619,
"step": 10202
},
{
"epoch": 0.6184242424242424,
"grad_norm": 0.058089837431907654,
"learning_rate": 0.0008,
"loss": 2.085179090499878,
"step": 10204
},
{
"epoch": 0.6185454545454545,
"grad_norm": 0.05678946524858475,
"learning_rate": 0.0008,
"loss": 2.0642576217651367,
"step": 10206
},
{
"epoch": 0.6186666666666667,
"grad_norm": 0.04838884621858597,
"learning_rate": 0.0008,
"loss": 2.007716655731201,
"step": 10208
},
{
"epoch": 0.6187878787878788,
"grad_norm": 0.03952307254076004,
"learning_rate": 0.0008,
"loss": 2.042813301086426,
"step": 10210
},
{
"epoch": 0.618909090909091,
"grad_norm": 0.04357096925377846,
"learning_rate": 0.0008,
"loss": 2.1363070011138916,
"step": 10212
},
{
"epoch": 0.619030303030303,
"grad_norm": 0.042068157345056534,
"learning_rate": 0.0008,
"loss": 2.018315315246582,
"step": 10214
},
{
"epoch": 0.6191515151515151,
"grad_norm": 0.04382069408893585,
"learning_rate": 0.0008,
"loss": 2.03249454498291,
"step": 10216
},
{
"epoch": 0.6192727272727273,
"grad_norm": 0.040362950414419174,
"learning_rate": 0.0008,
"loss": 2.0420455932617188,
"step": 10218
},
{
"epoch": 0.6193939393939394,
"grad_norm": 0.04301993176341057,
"learning_rate": 0.0008,
"loss": 2.0546464920043945,
"step": 10220
},
{
"epoch": 0.6195151515151515,
"grad_norm": 0.04228587821125984,
"learning_rate": 0.0008,
"loss": 1.9748436212539673,
"step": 10222
},
{
"epoch": 0.6196363636363637,
"grad_norm": 0.04204791039228439,
"learning_rate": 0.0008,
"loss": 2.051501512527466,
"step": 10224
},
{
"epoch": 0.6197575757575757,
"grad_norm": 0.04131440445780754,
"learning_rate": 0.0008,
"loss": 1.988959550857544,
"step": 10226
},
{
"epoch": 0.6198787878787879,
"grad_norm": 0.04281692951917648,
"learning_rate": 0.0008,
"loss": 2.077925682067871,
"step": 10228
},
{
"epoch": 0.62,
"grad_norm": 0.04242538660764694,
"learning_rate": 0.0008,
"loss": 2.05753755569458,
"step": 10230
},
{
"epoch": 0.6201212121212121,
"grad_norm": 0.03981756791472435,
"learning_rate": 0.0008,
"loss": 2.0021681785583496,
"step": 10232
},
{
"epoch": 0.6202424242424243,
"grad_norm": 0.04266967996954918,
"learning_rate": 0.0008,
"loss": 2.1205687522888184,
"step": 10234
},
{
"epoch": 0.6203636363636363,
"grad_norm": 0.05250680819153786,
"learning_rate": 0.0008,
"loss": 2.0818686485290527,
"step": 10236
},
{
"epoch": 0.6204848484848485,
"grad_norm": 0.07078196853399277,
"learning_rate": 0.0008,
"loss": 2.0633296966552734,
"step": 10238
},
{
"epoch": 0.6206060606060606,
"grad_norm": 0.05430322885513306,
"learning_rate": 0.0008,
"loss": 2.069234848022461,
"step": 10240
},
{
"epoch": 0.6207272727272727,
"grad_norm": 0.047631315886974335,
"learning_rate": 0.0008,
"loss": 1.983152151107788,
"step": 10242
},
{
"epoch": 0.6208484848484849,
"grad_norm": 0.04390088841319084,
"learning_rate": 0.0008,
"loss": 2.112992763519287,
"step": 10244
},
{
"epoch": 0.620969696969697,
"grad_norm": 0.047156888991594315,
"learning_rate": 0.0008,
"loss": 2.1097261905670166,
"step": 10246
},
{
"epoch": 0.6210909090909091,
"grad_norm": 0.04867114871740341,
"learning_rate": 0.0008,
"loss": 2.060737133026123,
"step": 10248
},
{
"epoch": 0.6212121212121212,
"grad_norm": 0.05077095702290535,
"learning_rate": 0.0008,
"loss": 2.0953612327575684,
"step": 10250
},
{
"epoch": 0.6213333333333333,
"grad_norm": 0.05959615856409073,
"learning_rate": 0.0008,
"loss": 2.051717758178711,
"step": 10252
},
{
"epoch": 0.6214545454545455,
"grad_norm": 0.052309807389974594,
"learning_rate": 0.0008,
"loss": 2.1026647090911865,
"step": 10254
},
{
"epoch": 0.6215757575757576,
"grad_norm": 0.047404058277606964,
"learning_rate": 0.0008,
"loss": 2.080410957336426,
"step": 10256
},
{
"epoch": 0.6216969696969697,
"grad_norm": 0.05455075576901436,
"learning_rate": 0.0008,
"loss": 2.0268568992614746,
"step": 10258
},
{
"epoch": 0.6218181818181818,
"grad_norm": 0.053322311490774155,
"learning_rate": 0.0008,
"loss": 2.0241076946258545,
"step": 10260
},
{
"epoch": 0.6219393939393939,
"grad_norm": 0.047652557492256165,
"learning_rate": 0.0008,
"loss": 2.081092357635498,
"step": 10262
},
{
"epoch": 0.6220606060606061,
"grad_norm": 0.04553017020225525,
"learning_rate": 0.0008,
"loss": 2.0657002925872803,
"step": 10264
},
{
"epoch": 0.6221818181818182,
"grad_norm": 0.04054577276110649,
"learning_rate": 0.0008,
"loss": 2.0543935298919678,
"step": 10266
},
{
"epoch": 0.6223030303030304,
"grad_norm": 0.04168444126844406,
"learning_rate": 0.0008,
"loss": 2.0518665313720703,
"step": 10268
},
{
"epoch": 0.6224242424242424,
"grad_norm": 0.04001783952116966,
"learning_rate": 0.0008,
"loss": 1.976689338684082,
"step": 10270
},
{
"epoch": 0.6225454545454545,
"grad_norm": 0.04802733287215233,
"learning_rate": 0.0008,
"loss": 2.0653159618377686,
"step": 10272
},
{
"epoch": 0.6226666666666667,
"grad_norm": 0.047018226236104965,
"learning_rate": 0.0008,
"loss": 2.134535789489746,
"step": 10274
},
{
"epoch": 0.6227878787878788,
"grad_norm": 0.048215191811323166,
"learning_rate": 0.0008,
"loss": 2.05714750289917,
"step": 10276
},
{
"epoch": 0.6229090909090909,
"grad_norm": 0.05343217775225639,
"learning_rate": 0.0008,
"loss": 2.060046672821045,
"step": 10278
},
{
"epoch": 0.623030303030303,
"grad_norm": 0.04313361644744873,
"learning_rate": 0.0008,
"loss": 2.006789207458496,
"step": 10280
},
{
"epoch": 0.6231515151515151,
"grad_norm": 0.04114801809191704,
"learning_rate": 0.0008,
"loss": 2.061086893081665,
"step": 10282
},
{
"epoch": 0.6232727272727273,
"grad_norm": 0.042347390204668045,
"learning_rate": 0.0008,
"loss": 1.9885125160217285,
"step": 10284
},
{
"epoch": 0.6233939393939394,
"grad_norm": 0.04199928045272827,
"learning_rate": 0.0008,
"loss": 2.036904811859131,
"step": 10286
},
{
"epoch": 0.6235151515151515,
"grad_norm": 0.039942845702171326,
"learning_rate": 0.0008,
"loss": 2.0284602642059326,
"step": 10288
},
{
"epoch": 0.6236363636363637,
"grad_norm": 0.03932390362024307,
"learning_rate": 0.0008,
"loss": 1.993929386138916,
"step": 10290
},
{
"epoch": 0.6237575757575757,
"grad_norm": 0.04483873024582863,
"learning_rate": 0.0008,
"loss": 2.0938498973846436,
"step": 10292
},
{
"epoch": 0.6238787878787879,
"grad_norm": 0.051386959850788116,
"learning_rate": 0.0008,
"loss": 2.049576759338379,
"step": 10294
},
{
"epoch": 0.624,
"grad_norm": 0.04598652198910713,
"learning_rate": 0.0008,
"loss": 2.0979793071746826,
"step": 10296
},
{
"epoch": 0.6241212121212121,
"grad_norm": 0.04094807058572769,
"learning_rate": 0.0008,
"loss": 2.0111401081085205,
"step": 10298
},
{
"epoch": 0.6242424242424243,
"grad_norm": 0.05136030539870262,
"learning_rate": 0.0008,
"loss": 2.014101028442383,
"step": 10300
},
{
"epoch": 0.6243636363636363,
"grad_norm": 0.05406546592712402,
"learning_rate": 0.0008,
"loss": 2.024404287338257,
"step": 10302
},
{
"epoch": 0.6244848484848485,
"grad_norm": 0.0553971491754055,
"learning_rate": 0.0008,
"loss": 2.071190118789673,
"step": 10304
},
{
"epoch": 0.6246060606060606,
"grad_norm": 0.05652173236012459,
"learning_rate": 0.0008,
"loss": 2.0946414470672607,
"step": 10306
},
{
"epoch": 0.6247272727272727,
"grad_norm": 0.046243440359830856,
"learning_rate": 0.0008,
"loss": 2.1182353496551514,
"step": 10308
},
{
"epoch": 0.6248484848484849,
"grad_norm": 0.039480313658714294,
"learning_rate": 0.0008,
"loss": 2.044050931930542,
"step": 10310
},
{
"epoch": 0.624969696969697,
"grad_norm": 0.043322592973709106,
"learning_rate": 0.0008,
"loss": 1.9984006881713867,
"step": 10312
},
{
"epoch": 0.6250909090909091,
"grad_norm": 0.053910594433546066,
"learning_rate": 0.0008,
"loss": 1.9675636291503906,
"step": 10314
},
{
"epoch": 0.6252121212121212,
"grad_norm": 0.05897307023406029,
"learning_rate": 0.0008,
"loss": 2.0506772994995117,
"step": 10316
},
{
"epoch": 0.6253333333333333,
"grad_norm": 0.056205835193395615,
"learning_rate": 0.0008,
"loss": 2.071389675140381,
"step": 10318
},
{
"epoch": 0.6254545454545455,
"grad_norm": 0.06144187971949577,
"learning_rate": 0.0008,
"loss": 2.031966209411621,
"step": 10320
},
{
"epoch": 0.6255757575757576,
"grad_norm": 0.06324569880962372,
"learning_rate": 0.0008,
"loss": 2.0946481227874756,
"step": 10322
},
{
"epoch": 0.6256969696969698,
"grad_norm": 0.05682316794991493,
"learning_rate": 0.0008,
"loss": 2.00164532661438,
"step": 10324
},
{
"epoch": 0.6258181818181818,
"grad_norm": 0.04654648154973984,
"learning_rate": 0.0008,
"loss": 2.0782480239868164,
"step": 10326
},
{
"epoch": 0.6259393939393939,
"grad_norm": 0.040731292217969894,
"learning_rate": 0.0008,
"loss": 2.0746824741363525,
"step": 10328
},
{
"epoch": 0.6260606060606061,
"grad_norm": 0.04014768451452255,
"learning_rate": 0.0008,
"loss": 2.0356733798980713,
"step": 10330
},
{
"epoch": 0.6261818181818182,
"grad_norm": 0.04069908335804939,
"learning_rate": 0.0008,
"loss": 2.012195110321045,
"step": 10332
},
{
"epoch": 0.6263030303030303,
"grad_norm": 0.04133635386824608,
"learning_rate": 0.0008,
"loss": 2.0724036693573,
"step": 10334
},
{
"epoch": 0.6264242424242424,
"grad_norm": 0.04036537557840347,
"learning_rate": 0.0008,
"loss": 2.0124800205230713,
"step": 10336
},
{
"epoch": 0.6265454545454545,
"grad_norm": 0.040761057287454605,
"learning_rate": 0.0008,
"loss": 2.031153917312622,
"step": 10338
},
{
"epoch": 0.6266666666666667,
"grad_norm": 0.039212748408317566,
"learning_rate": 0.0008,
"loss": 2.0603315830230713,
"step": 10340
},
{
"epoch": 0.6267878787878788,
"grad_norm": 0.04087004438042641,
"learning_rate": 0.0008,
"loss": 2.073761224746704,
"step": 10342
},
{
"epoch": 0.6269090909090909,
"grad_norm": 0.03977564722299576,
"learning_rate": 0.0008,
"loss": 2.029789447784424,
"step": 10344
},
{
"epoch": 0.627030303030303,
"grad_norm": 0.05178411677479744,
"learning_rate": 0.0008,
"loss": 2.138108730316162,
"step": 10346
},
{
"epoch": 0.6271515151515151,
"grad_norm": 0.052128296345472336,
"learning_rate": 0.0008,
"loss": 2.0604095458984375,
"step": 10348
},
{
"epoch": 0.6272727272727273,
"grad_norm": 0.05153246968984604,
"learning_rate": 0.0008,
"loss": 2.0511763095855713,
"step": 10350
},
{
"epoch": 0.6273939393939394,
"grad_norm": 0.04532146453857422,
"learning_rate": 0.0008,
"loss": 2.044877290725708,
"step": 10352
},
{
"epoch": 0.6275151515151515,
"grad_norm": 0.04986152797937393,
"learning_rate": 0.0008,
"loss": 2.0453522205352783,
"step": 10354
},
{
"epoch": 0.6276363636363637,
"grad_norm": 0.05014058202505112,
"learning_rate": 0.0008,
"loss": 2.008852958679199,
"step": 10356
},
{
"epoch": 0.6277575757575757,
"grad_norm": 0.05649475380778313,
"learning_rate": 0.0008,
"loss": 2.0148820877075195,
"step": 10358
},
{
"epoch": 0.6278787878787879,
"grad_norm": 0.05017043277621269,
"learning_rate": 0.0008,
"loss": 1.9873493909835815,
"step": 10360
},
{
"epoch": 0.628,
"grad_norm": 0.04581909626722336,
"learning_rate": 0.0008,
"loss": 2.050694465637207,
"step": 10362
},
{
"epoch": 0.6281212121212121,
"grad_norm": 0.04151628911495209,
"learning_rate": 0.0008,
"loss": 2.091447114944458,
"step": 10364
},
{
"epoch": 0.6282424242424243,
"grad_norm": 0.044248465448617935,
"learning_rate": 0.0008,
"loss": 2.0530717372894287,
"step": 10366
},
{
"epoch": 0.6283636363636363,
"grad_norm": 0.04162420704960823,
"learning_rate": 0.0008,
"loss": 2.0046520233154297,
"step": 10368
},
{
"epoch": 0.6284848484848485,
"grad_norm": 0.044357385486364365,
"learning_rate": 0.0008,
"loss": 2.0735747814178467,
"step": 10370
},
{
"epoch": 0.6286060606060606,
"grad_norm": 0.043578267097473145,
"learning_rate": 0.0008,
"loss": 2.0462958812713623,
"step": 10372
},
{
"epoch": 0.6287272727272727,
"grad_norm": 0.04049228876829147,
"learning_rate": 0.0008,
"loss": 2.1037776470184326,
"step": 10374
},
{
"epoch": 0.6288484848484849,
"grad_norm": 0.04146990552544594,
"learning_rate": 0.0008,
"loss": 2.0906312465667725,
"step": 10376
},
{
"epoch": 0.628969696969697,
"grad_norm": 0.0400114506483078,
"learning_rate": 0.0008,
"loss": 2.0808608531951904,
"step": 10378
},
{
"epoch": 0.6290909090909091,
"grad_norm": 0.043969038873910904,
"learning_rate": 0.0008,
"loss": 2.0222926139831543,
"step": 10380
},
{
"epoch": 0.6292121212121212,
"grad_norm": 0.04336601123213768,
"learning_rate": 0.0008,
"loss": 2.072545051574707,
"step": 10382
},
{
"epoch": 0.6293333333333333,
"grad_norm": 0.04032933712005615,
"learning_rate": 0.0008,
"loss": 2.048393964767456,
"step": 10384
},
{
"epoch": 0.6294545454545455,
"grad_norm": 0.05023579299449921,
"learning_rate": 0.0008,
"loss": 2.0785717964172363,
"step": 10386
},
{
"epoch": 0.6295757575757576,
"grad_norm": 0.045858778059482574,
"learning_rate": 0.0008,
"loss": 2.0833356380462646,
"step": 10388
},
{
"epoch": 0.6296969696969696,
"grad_norm": 0.04417332261800766,
"learning_rate": 0.0008,
"loss": 2.012096405029297,
"step": 10390
},
{
"epoch": 0.6298181818181818,
"grad_norm": 0.04218036308884621,
"learning_rate": 0.0008,
"loss": 2.014789342880249,
"step": 10392
},
{
"epoch": 0.6299393939393939,
"grad_norm": 0.04185286536812782,
"learning_rate": 0.0008,
"loss": 2.0657505989074707,
"step": 10394
},
{
"epoch": 0.6300606060606061,
"grad_norm": 0.044435054063797,
"learning_rate": 0.0008,
"loss": 2.043280839920044,
"step": 10396
},
{
"epoch": 0.6301818181818182,
"grad_norm": 0.04470399022102356,
"learning_rate": 0.0008,
"loss": 2.050792932510376,
"step": 10398
},
{
"epoch": 0.6303030303030303,
"grad_norm": 0.04289347305893898,
"learning_rate": 0.0008,
"loss": 2.0847342014312744,
"step": 10400
},
{
"epoch": 0.6304242424242424,
"grad_norm": 0.04229894280433655,
"learning_rate": 0.0008,
"loss": 2.0417721271514893,
"step": 10402
},
{
"epoch": 0.6305454545454545,
"grad_norm": 0.04059493541717529,
"learning_rate": 0.0008,
"loss": 2.0057213306427,
"step": 10404
},
{
"epoch": 0.6306666666666667,
"grad_norm": 0.04267735406756401,
"learning_rate": 0.0008,
"loss": 2.0471932888031006,
"step": 10406
},
{
"epoch": 0.6307878787878788,
"grad_norm": 0.04085344821214676,
"learning_rate": 0.0008,
"loss": 2.034759759902954,
"step": 10408
},
{
"epoch": 0.6309090909090909,
"grad_norm": 0.04208699241280556,
"learning_rate": 0.0008,
"loss": 2.0552818775177,
"step": 10410
},
{
"epoch": 0.631030303030303,
"grad_norm": 0.043050721287727356,
"learning_rate": 0.0008,
"loss": 2.075732707977295,
"step": 10412
},
{
"epoch": 0.6311515151515151,
"grad_norm": 0.04211968928575516,
"learning_rate": 0.0008,
"loss": 2.039583921432495,
"step": 10414
},
{
"epoch": 0.6312727272727273,
"grad_norm": 0.048401422798633575,
"learning_rate": 0.0008,
"loss": 2.049607038497925,
"step": 10416
},
{
"epoch": 0.6313939393939394,
"grad_norm": 0.041056036949157715,
"learning_rate": 0.0008,
"loss": 2.0128369331359863,
"step": 10418
},
{
"epoch": 0.6315151515151515,
"grad_norm": 0.03967892751097679,
"learning_rate": 0.0008,
"loss": 1.9962308406829834,
"step": 10420
},
{
"epoch": 0.6316363636363637,
"grad_norm": 0.04715515673160553,
"learning_rate": 0.0008,
"loss": 1.9973461627960205,
"step": 10422
},
{
"epoch": 0.6317575757575757,
"grad_norm": 0.04020143672823906,
"learning_rate": 0.0008,
"loss": 2.0090718269348145,
"step": 10424
},
{
"epoch": 0.6318787878787879,
"grad_norm": 0.04361402615904808,
"learning_rate": 0.0008,
"loss": 2.0035400390625,
"step": 10426
},
{
"epoch": 0.632,
"grad_norm": 0.053887441754341125,
"learning_rate": 0.0008,
"loss": 2.074716567993164,
"step": 10428
},
{
"epoch": 0.6321212121212121,
"grad_norm": 0.04071863368153572,
"learning_rate": 0.0008,
"loss": 2.054772138595581,
"step": 10430
},
{
"epoch": 0.6322424242424243,
"grad_norm": 0.04080454260110855,
"learning_rate": 0.0008,
"loss": 2.0216917991638184,
"step": 10432
},
{
"epoch": 0.6323636363636363,
"grad_norm": 0.040770869702100754,
"learning_rate": 0.0008,
"loss": 2.0736541748046875,
"step": 10434
},
{
"epoch": 0.6324848484848485,
"grad_norm": 0.03995240479707718,
"learning_rate": 0.0008,
"loss": 1.9814461469650269,
"step": 10436
},
{
"epoch": 0.6326060606060606,
"grad_norm": 0.04040864109992981,
"learning_rate": 0.0008,
"loss": 2.0131471157073975,
"step": 10438
},
{
"epoch": 0.6327272727272727,
"grad_norm": 0.04035896062850952,
"learning_rate": 0.0008,
"loss": 2.032036781311035,
"step": 10440
},
{
"epoch": 0.6328484848484849,
"grad_norm": 0.040777143090963364,
"learning_rate": 0.0008,
"loss": 2.0429306030273438,
"step": 10442
},
{
"epoch": 0.632969696969697,
"grad_norm": 0.040577035397291183,
"learning_rate": 0.0008,
"loss": 2.0277600288391113,
"step": 10444
},
{
"epoch": 0.633090909090909,
"grad_norm": 0.0410136878490448,
"learning_rate": 0.0008,
"loss": 2.074518918991089,
"step": 10446
},
{
"epoch": 0.6332121212121212,
"grad_norm": 0.04067283496260643,
"learning_rate": 0.0008,
"loss": 2.0221283435821533,
"step": 10448
},
{
"epoch": 0.6333333333333333,
"grad_norm": 0.04046730324625969,
"learning_rate": 0.0008,
"loss": 2.017331600189209,
"step": 10450
},
{
"epoch": 0.6334545454545455,
"grad_norm": 0.040669649839401245,
"learning_rate": 0.0008,
"loss": 2.0722079277038574,
"step": 10452
},
{
"epoch": 0.6335757575757576,
"grad_norm": 0.04521651566028595,
"learning_rate": 0.0008,
"loss": 2.0221681594848633,
"step": 10454
},
{
"epoch": 0.6336969696969696,
"grad_norm": 0.04483189433813095,
"learning_rate": 0.0008,
"loss": 2.038266658782959,
"step": 10456
},
{
"epoch": 0.6338181818181818,
"grad_norm": 0.04013039916753769,
"learning_rate": 0.0008,
"loss": 2.0422534942626953,
"step": 10458
},
{
"epoch": 0.6339393939393939,
"grad_norm": 0.040793631225824356,
"learning_rate": 0.0008,
"loss": 2.0246450901031494,
"step": 10460
},
{
"epoch": 0.6340606060606061,
"grad_norm": 0.04105642810463905,
"learning_rate": 0.0008,
"loss": 1.9963150024414062,
"step": 10462
},
{
"epoch": 0.6341818181818182,
"grad_norm": 0.043382737785577774,
"learning_rate": 0.0008,
"loss": 2.1359646320343018,
"step": 10464
},
{
"epoch": 0.6343030303030303,
"grad_norm": 0.04372112825512886,
"learning_rate": 0.0008,
"loss": 2.0451903343200684,
"step": 10466
},
{
"epoch": 0.6344242424242424,
"grad_norm": 0.051353346556425095,
"learning_rate": 0.0008,
"loss": 2.077895402908325,
"step": 10468
},
{
"epoch": 0.6345454545454545,
"grad_norm": 0.050360966473817825,
"learning_rate": 0.0008,
"loss": 2.0203046798706055,
"step": 10470
},
{
"epoch": 0.6346666666666667,
"grad_norm": 0.04951340705156326,
"learning_rate": 0.0008,
"loss": 2.104799270629883,
"step": 10472
},
{
"epoch": 0.6347878787878788,
"grad_norm": 0.04104296863079071,
"learning_rate": 0.0008,
"loss": 2.0573976039886475,
"step": 10474
},
{
"epoch": 0.6349090909090909,
"grad_norm": 0.049726858735084534,
"learning_rate": 0.0008,
"loss": 2.0523719787597656,
"step": 10476
},
{
"epoch": 0.6350303030303031,
"grad_norm": 0.06320419907569885,
"learning_rate": 0.0008,
"loss": 2.090744972229004,
"step": 10478
},
{
"epoch": 0.6351515151515151,
"grad_norm": 0.06110332906246185,
"learning_rate": 0.0008,
"loss": 2.039963960647583,
"step": 10480
},
{
"epoch": 0.6352727272727273,
"grad_norm": 0.0478023998439312,
"learning_rate": 0.0008,
"loss": 2.0328664779663086,
"step": 10482
},
{
"epoch": 0.6353939393939394,
"grad_norm": 0.04198424890637398,
"learning_rate": 0.0008,
"loss": 2.0324063301086426,
"step": 10484
},
{
"epoch": 0.6355151515151515,
"grad_norm": 0.04134237393736839,
"learning_rate": 0.0008,
"loss": 2.0285158157348633,
"step": 10486
},
{
"epoch": 0.6356363636363637,
"grad_norm": 0.04849648475646973,
"learning_rate": 0.0008,
"loss": 2.0316240787506104,
"step": 10488
},
{
"epoch": 0.6357575757575757,
"grad_norm": 0.040633369237184525,
"learning_rate": 0.0008,
"loss": 2.0052003860473633,
"step": 10490
},
{
"epoch": 0.6358787878787879,
"grad_norm": 0.04894213378429413,
"learning_rate": 0.0008,
"loss": 2.0214970111846924,
"step": 10492
},
{
"epoch": 0.636,
"grad_norm": 0.047110069543123245,
"learning_rate": 0.0008,
"loss": 2.0231375694274902,
"step": 10494
},
{
"epoch": 0.6361212121212121,
"grad_norm": 0.04740598797798157,
"learning_rate": 0.0008,
"loss": 2.0768368244171143,
"step": 10496
},
{
"epoch": 0.6362424242424243,
"grad_norm": 0.0397944450378418,
"learning_rate": 0.0008,
"loss": 2.0048537254333496,
"step": 10498
},
{
"epoch": 0.6363636363636364,
"grad_norm": 0.04116680845618248,
"learning_rate": 0.0008,
"loss": 2.0941646099090576,
"step": 10500
},
{
"epoch": 0.6364848484848484,
"grad_norm": 0.040642205625772476,
"learning_rate": 0.0008,
"loss": 2.044212818145752,
"step": 10502
},
{
"epoch": 0.6366060606060606,
"grad_norm": 0.04026055708527565,
"learning_rate": 0.0008,
"loss": 2.0710158348083496,
"step": 10504
},
{
"epoch": 0.6367272727272727,
"grad_norm": 0.0451180562376976,
"learning_rate": 0.0008,
"loss": 2.101998805999756,
"step": 10506
},
{
"epoch": 0.6368484848484849,
"grad_norm": 0.04860169067978859,
"learning_rate": 0.0008,
"loss": 2.025620937347412,
"step": 10508
},
{
"epoch": 0.636969696969697,
"grad_norm": 0.04406809061765671,
"learning_rate": 0.0008,
"loss": 2.1199123859405518,
"step": 10510
},
{
"epoch": 0.637090909090909,
"grad_norm": 0.04272427782416344,
"learning_rate": 0.0008,
"loss": 2.073854446411133,
"step": 10512
},
{
"epoch": 0.6372121212121212,
"grad_norm": 0.03928150609135628,
"learning_rate": 0.0008,
"loss": 2.006014823913574,
"step": 10514
},
{
"epoch": 0.6373333333333333,
"grad_norm": 0.04123329371213913,
"learning_rate": 0.0008,
"loss": 2.076625347137451,
"step": 10516
},
{
"epoch": 0.6374545454545455,
"grad_norm": 0.04198009893298149,
"learning_rate": 0.0008,
"loss": 2.0214662551879883,
"step": 10518
},
{
"epoch": 0.6375757575757576,
"grad_norm": 0.041921645402908325,
"learning_rate": 0.0008,
"loss": 2.0892016887664795,
"step": 10520
},
{
"epoch": 0.6376969696969697,
"grad_norm": 0.04215230047702789,
"learning_rate": 0.0008,
"loss": 2.010708808898926,
"step": 10522
},
{
"epoch": 0.6378181818181818,
"grad_norm": 0.04735802486538887,
"learning_rate": 0.0008,
"loss": 2.0564725399017334,
"step": 10524
},
{
"epoch": 0.6379393939393939,
"grad_norm": 0.046282537281513214,
"learning_rate": 0.0008,
"loss": 2.065124750137329,
"step": 10526
},
{
"epoch": 0.6380606060606061,
"grad_norm": 0.04069935530424118,
"learning_rate": 0.0008,
"loss": 1.997698187828064,
"step": 10528
},
{
"epoch": 0.6381818181818182,
"grad_norm": 0.04554257541894913,
"learning_rate": 0.0008,
"loss": 2.110703706741333,
"step": 10530
},
{
"epoch": 0.6383030303030303,
"grad_norm": 0.04396204277873039,
"learning_rate": 0.0008,
"loss": 2.0596840381622314,
"step": 10532
},
{
"epoch": 0.6384242424242424,
"grad_norm": 0.051701538264751434,
"learning_rate": 0.0008,
"loss": 2.0847578048706055,
"step": 10534
},
{
"epoch": 0.6385454545454545,
"grad_norm": 0.044198017567396164,
"learning_rate": 0.0008,
"loss": 2.0478246212005615,
"step": 10536
},
{
"epoch": 0.6386666666666667,
"grad_norm": 0.0415523536503315,
"learning_rate": 0.0008,
"loss": 2.0444235801696777,
"step": 10538
},
{
"epoch": 0.6387878787878788,
"grad_norm": 0.04069177433848381,
"learning_rate": 0.0008,
"loss": 2.066434383392334,
"step": 10540
},
{
"epoch": 0.6389090909090909,
"grad_norm": 0.046909790486097336,
"learning_rate": 0.0008,
"loss": 2.006826162338257,
"step": 10542
},
{
"epoch": 0.6390303030303031,
"grad_norm": 0.04509374126791954,
"learning_rate": 0.0008,
"loss": 2.038069009780884,
"step": 10544
},
{
"epoch": 0.6391515151515151,
"grad_norm": 0.054905716329813004,
"learning_rate": 0.0008,
"loss": 2.0947513580322266,
"step": 10546
},
{
"epoch": 0.6392727272727273,
"grad_norm": 0.04359889775514603,
"learning_rate": 0.0008,
"loss": 2.111706018447876,
"step": 10548
},
{
"epoch": 0.6393939393939394,
"grad_norm": 0.04081260412931442,
"learning_rate": 0.0008,
"loss": 2.043316125869751,
"step": 10550
},
{
"epoch": 0.6395151515151515,
"grad_norm": 0.040786296129226685,
"learning_rate": 0.0008,
"loss": 2.026217460632324,
"step": 10552
},
{
"epoch": 0.6396363636363637,
"grad_norm": 0.0402446947991848,
"learning_rate": 0.0008,
"loss": 2.0145158767700195,
"step": 10554
},
{
"epoch": 0.6397575757575757,
"grad_norm": 0.040364816784858704,
"learning_rate": 0.0008,
"loss": 2.026563882827759,
"step": 10556
},
{
"epoch": 0.6398787878787878,
"grad_norm": 0.04293091595172882,
"learning_rate": 0.0008,
"loss": 2.094867706298828,
"step": 10558
},
{
"epoch": 0.64,
"grad_norm": 0.06870633363723755,
"learning_rate": 0.0008,
"loss": 1.9887374639511108,
"step": 10560
},
{
"epoch": 0.6401212121212121,
"grad_norm": 0.06681699305772781,
"learning_rate": 0.0008,
"loss": 2.0443031787872314,
"step": 10562
},
{
"epoch": 0.6402424242424243,
"grad_norm": 0.05120959132909775,
"learning_rate": 0.0008,
"loss": 2.0040924549102783,
"step": 10564
},
{
"epoch": 0.6403636363636364,
"grad_norm": 0.04873562231659889,
"learning_rate": 0.0008,
"loss": 2.034087896347046,
"step": 10566
},
{
"epoch": 0.6404848484848484,
"grad_norm": 0.04063502699136734,
"learning_rate": 0.0008,
"loss": 2.0617947578430176,
"step": 10568
},
{
"epoch": 0.6406060606060606,
"grad_norm": 0.04736337065696716,
"learning_rate": 0.0008,
"loss": 2.033449649810791,
"step": 10570
},
{
"epoch": 0.6407272727272727,
"grad_norm": 0.040957260876894,
"learning_rate": 0.0008,
"loss": 2.0974881649017334,
"step": 10572
},
{
"epoch": 0.6408484848484849,
"grad_norm": 0.04228580743074417,
"learning_rate": 0.0008,
"loss": 2.0218095779418945,
"step": 10574
},
{
"epoch": 0.640969696969697,
"grad_norm": 0.039312638342380524,
"learning_rate": 0.0008,
"loss": 2.078134298324585,
"step": 10576
},
{
"epoch": 0.641090909090909,
"grad_norm": 0.04164254665374756,
"learning_rate": 0.0008,
"loss": 2.052442789077759,
"step": 10578
},
{
"epoch": 0.6412121212121212,
"grad_norm": 0.03966309502720833,
"learning_rate": 0.0008,
"loss": 2.0294086933135986,
"step": 10580
},
{
"epoch": 0.6413333333333333,
"grad_norm": 0.040486548095941544,
"learning_rate": 0.0008,
"loss": 2.0435965061187744,
"step": 10582
},
{
"epoch": 0.6414545454545455,
"grad_norm": 0.0395013689994812,
"learning_rate": 0.0008,
"loss": 2.032428741455078,
"step": 10584
},
{
"epoch": 0.6415757575757576,
"grad_norm": 0.03989096358418465,
"learning_rate": 0.0008,
"loss": 2.0504279136657715,
"step": 10586
},
{
"epoch": 0.6416969696969697,
"grad_norm": 0.04125724732875824,
"learning_rate": 0.0008,
"loss": 2.0641634464263916,
"step": 10588
},
{
"epoch": 0.6418181818181818,
"grad_norm": 0.04226572439074516,
"learning_rate": 0.0008,
"loss": 2.0425403118133545,
"step": 10590
},
{
"epoch": 0.6419393939393939,
"grad_norm": 0.04594234749674797,
"learning_rate": 0.0008,
"loss": 2.010281562805176,
"step": 10592
},
{
"epoch": 0.6420606060606061,
"grad_norm": 0.048436328768730164,
"learning_rate": 0.0008,
"loss": 2.0432422161102295,
"step": 10594
},
{
"epoch": 0.6421818181818182,
"grad_norm": 0.045835092663764954,
"learning_rate": 0.0008,
"loss": 2.060971736907959,
"step": 10596
},
{
"epoch": 0.6423030303030303,
"grad_norm": 0.04199066758155823,
"learning_rate": 0.0008,
"loss": 2.048128366470337,
"step": 10598
},
{
"epoch": 0.6424242424242425,
"grad_norm": 0.04122932627797127,
"learning_rate": 0.0008,
"loss": 2.0205471515655518,
"step": 10600
},
{
"epoch": 0.6425454545454545,
"grad_norm": 0.040305737406015396,
"learning_rate": 0.0008,
"loss": 2.0065131187438965,
"step": 10602
},
{
"epoch": 0.6426666666666667,
"grad_norm": 0.04080755263566971,
"learning_rate": 0.0008,
"loss": 2.0436816215515137,
"step": 10604
},
{
"epoch": 0.6427878787878788,
"grad_norm": 0.04232707619667053,
"learning_rate": 0.0008,
"loss": 2.1188485622406006,
"step": 10606
},
{
"epoch": 0.6429090909090909,
"grad_norm": 0.04404500126838684,
"learning_rate": 0.0008,
"loss": 2.01149320602417,
"step": 10608
},
{
"epoch": 0.6430303030303031,
"grad_norm": 0.044932153075933456,
"learning_rate": 0.0008,
"loss": 2.0597550868988037,
"step": 10610
},
{
"epoch": 0.6431515151515151,
"grad_norm": 0.04333998262882233,
"learning_rate": 0.0008,
"loss": 2.08803129196167,
"step": 10612
},
{
"epoch": 0.6432727272727272,
"grad_norm": 0.04068513959646225,
"learning_rate": 0.0008,
"loss": 2.0351433753967285,
"step": 10614
},
{
"epoch": 0.6433939393939394,
"grad_norm": 0.04331589862704277,
"learning_rate": 0.0008,
"loss": 2.0490384101867676,
"step": 10616
},
{
"epoch": 0.6435151515151515,
"grad_norm": 0.04019073769450188,
"learning_rate": 0.0008,
"loss": 2.072335720062256,
"step": 10618
},
{
"epoch": 0.6436363636363637,
"grad_norm": 0.03898182511329651,
"learning_rate": 0.0008,
"loss": 2.0801119804382324,
"step": 10620
},
{
"epoch": 0.6437575757575758,
"grad_norm": 0.044556234031915665,
"learning_rate": 0.0008,
"loss": 2.0423271656036377,
"step": 10622
},
{
"epoch": 0.6438787878787878,
"grad_norm": 0.04151058569550514,
"learning_rate": 0.0008,
"loss": 2.0774474143981934,
"step": 10624
},
{
"epoch": 0.644,
"grad_norm": 0.040732499212026596,
"learning_rate": 0.0008,
"loss": 2.096959114074707,
"step": 10626
},
{
"epoch": 0.6441212121212121,
"grad_norm": 0.04785393178462982,
"learning_rate": 0.0008,
"loss": 1.9779300689697266,
"step": 10628
},
{
"epoch": 0.6442424242424243,
"grad_norm": 0.04361748322844505,
"learning_rate": 0.0008,
"loss": 2.0504791736602783,
"step": 10630
},
{
"epoch": 0.6443636363636364,
"grad_norm": 0.051367320120334625,
"learning_rate": 0.0008,
"loss": 2.0278961658477783,
"step": 10632
},
{
"epoch": 0.6444848484848484,
"grad_norm": 0.04323229566216469,
"learning_rate": 0.0008,
"loss": 1.9616633653640747,
"step": 10634
},
{
"epoch": 0.6446060606060606,
"grad_norm": 0.04810146987438202,
"learning_rate": 0.0008,
"loss": 2.0341176986694336,
"step": 10636
},
{
"epoch": 0.6447272727272727,
"grad_norm": 0.04686879739165306,
"learning_rate": 0.0008,
"loss": 2.024221420288086,
"step": 10638
},
{
"epoch": 0.6448484848484849,
"grad_norm": 0.04736946150660515,
"learning_rate": 0.0008,
"loss": 1.9682576656341553,
"step": 10640
},
{
"epoch": 0.644969696969697,
"grad_norm": 0.041783131659030914,
"learning_rate": 0.0008,
"loss": 2.004380941390991,
"step": 10642
},
{
"epoch": 0.645090909090909,
"grad_norm": 0.04173954948782921,
"learning_rate": 0.0008,
"loss": 2.0568971633911133,
"step": 10644
},
{
"epoch": 0.6452121212121212,
"grad_norm": 0.04018475115299225,
"learning_rate": 0.0008,
"loss": 2.084078788757324,
"step": 10646
},
{
"epoch": 0.6453333333333333,
"grad_norm": 0.0415162667632103,
"learning_rate": 0.0008,
"loss": 2.125229597091675,
"step": 10648
},
{
"epoch": 0.6454545454545455,
"grad_norm": 0.04296228289604187,
"learning_rate": 0.0008,
"loss": 1.9890377521514893,
"step": 10650
},
{
"epoch": 0.6455757575757576,
"grad_norm": 0.04390754923224449,
"learning_rate": 0.0008,
"loss": 2.044459819793701,
"step": 10652
},
{
"epoch": 0.6456969696969697,
"grad_norm": 0.0432104617357254,
"learning_rate": 0.0008,
"loss": 2.0392658710479736,
"step": 10654
},
{
"epoch": 0.6458181818181818,
"grad_norm": 0.04281314089894295,
"learning_rate": 0.0008,
"loss": 2.040313243865967,
"step": 10656
},
{
"epoch": 0.6459393939393939,
"grad_norm": 0.04530719667673111,
"learning_rate": 0.0008,
"loss": 2.0593550205230713,
"step": 10658
},
{
"epoch": 0.6460606060606061,
"grad_norm": 0.05081390216946602,
"learning_rate": 0.0008,
"loss": 2.025420665740967,
"step": 10660
},
{
"epoch": 0.6461818181818182,
"grad_norm": 0.04082050919532776,
"learning_rate": 0.0008,
"loss": 2.054703712463379,
"step": 10662
},
{
"epoch": 0.6463030303030303,
"grad_norm": 0.043140824884176254,
"learning_rate": 0.0008,
"loss": 2.113339424133301,
"step": 10664
},
{
"epoch": 0.6464242424242425,
"grad_norm": 0.04111466184258461,
"learning_rate": 0.0008,
"loss": 2.0728836059570312,
"step": 10666
},
{
"epoch": 0.6465454545454545,
"grad_norm": 0.041501838713884354,
"learning_rate": 0.0008,
"loss": 2.0841102600097656,
"step": 10668
},
{
"epoch": 0.6466666666666666,
"grad_norm": 0.040834300220012665,
"learning_rate": 0.0008,
"loss": 2.0179288387298584,
"step": 10670
},
{
"epoch": 0.6467878787878788,
"grad_norm": 0.04044702649116516,
"learning_rate": 0.0008,
"loss": 2.055972099304199,
"step": 10672
},
{
"epoch": 0.6469090909090909,
"grad_norm": 0.055238936096429825,
"learning_rate": 0.0008,
"loss": 2.070812940597534,
"step": 10674
},
{
"epoch": 0.6470303030303031,
"grad_norm": 0.04445519670844078,
"learning_rate": 0.0008,
"loss": 2.0933890342712402,
"step": 10676
},
{
"epoch": 0.6471515151515151,
"grad_norm": 0.04189300164580345,
"learning_rate": 0.0008,
"loss": 2.0503575801849365,
"step": 10678
},
{
"epoch": 0.6472727272727272,
"grad_norm": 0.03971027955412865,
"learning_rate": 0.0008,
"loss": 2.007481336593628,
"step": 10680
},
{
"epoch": 0.6473939393939394,
"grad_norm": 0.03989313915371895,
"learning_rate": 0.0008,
"loss": 2.0354316234588623,
"step": 10682
},
{
"epoch": 0.6475151515151515,
"grad_norm": 0.04243423044681549,
"learning_rate": 0.0008,
"loss": 2.1067638397216797,
"step": 10684
},
{
"epoch": 0.6476363636363637,
"grad_norm": 0.042581263929605484,
"learning_rate": 0.0008,
"loss": 2.0249550342559814,
"step": 10686
},
{
"epoch": 0.6477575757575758,
"grad_norm": 0.04414933919906616,
"learning_rate": 0.0008,
"loss": 2.064206838607788,
"step": 10688
},
{
"epoch": 0.6478787878787878,
"grad_norm": 0.04132945463061333,
"learning_rate": 0.0008,
"loss": 1.9564241170883179,
"step": 10690
},
{
"epoch": 0.648,
"grad_norm": 0.04267419874668121,
"learning_rate": 0.0008,
"loss": 2.002647638320923,
"step": 10692
},
{
"epoch": 0.6481212121212121,
"grad_norm": 0.0418211854994297,
"learning_rate": 0.0008,
"loss": 2.038219928741455,
"step": 10694
},
{
"epoch": 0.6482424242424243,
"grad_norm": 0.039422739297151566,
"learning_rate": 0.0008,
"loss": 2.077955961227417,
"step": 10696
},
{
"epoch": 0.6483636363636364,
"grad_norm": 0.05431586503982544,
"learning_rate": 0.0008,
"loss": 2.023139476776123,
"step": 10698
},
{
"epoch": 0.6484848484848484,
"grad_norm": 0.049942225217819214,
"learning_rate": 0.0008,
"loss": 1.9339450597763062,
"step": 10700
},
{
"epoch": 0.6486060606060606,
"grad_norm": 0.04302513599395752,
"learning_rate": 0.0008,
"loss": 1.9879167079925537,
"step": 10702
},
{
"epoch": 0.6487272727272727,
"grad_norm": 0.04620015248656273,
"learning_rate": 0.0008,
"loss": 2.045673131942749,
"step": 10704
},
{
"epoch": 0.6488484848484849,
"grad_norm": 0.03986641392111778,
"learning_rate": 0.0008,
"loss": 2.050510883331299,
"step": 10706
},
{
"epoch": 0.648969696969697,
"grad_norm": 0.03950010985136032,
"learning_rate": 0.0008,
"loss": 2.021761417388916,
"step": 10708
},
{
"epoch": 0.649090909090909,
"grad_norm": 0.040867533534765244,
"learning_rate": 0.0008,
"loss": 2.0447840690612793,
"step": 10710
},
{
"epoch": 0.6492121212121212,
"grad_norm": 0.04158800095319748,
"learning_rate": 0.0008,
"loss": 1.9998259544372559,
"step": 10712
},
{
"epoch": 0.6493333333333333,
"grad_norm": 0.04581955820322037,
"learning_rate": 0.0008,
"loss": 1.9862772226333618,
"step": 10714
},
{
"epoch": 0.6494545454545455,
"grad_norm": 0.05773434415459633,
"learning_rate": 0.0008,
"loss": 2.0100011825561523,
"step": 10716
},
{
"epoch": 0.6495757575757576,
"grad_norm": 0.049918439239263535,
"learning_rate": 0.0008,
"loss": 2.0204014778137207,
"step": 10718
},
{
"epoch": 0.6496969696969697,
"grad_norm": 0.03981964290142059,
"learning_rate": 0.0008,
"loss": 2.0142264366149902,
"step": 10720
},
{
"epoch": 0.6498181818181819,
"grad_norm": 0.04074487090110779,
"learning_rate": 0.0008,
"loss": 2.0849318504333496,
"step": 10722
},
{
"epoch": 0.6499393939393939,
"grad_norm": 0.03994906321167946,
"learning_rate": 0.0008,
"loss": 2.028566598892212,
"step": 10724
},
{
"epoch": 0.650060606060606,
"grad_norm": 0.04633011296391487,
"learning_rate": 0.0008,
"loss": 2.0987675189971924,
"step": 10726
},
{
"epoch": 0.6501818181818182,
"grad_norm": 0.04370083659887314,
"learning_rate": 0.0008,
"loss": 1.997671365737915,
"step": 10728
},
{
"epoch": 0.6503030303030303,
"grad_norm": 0.04155488312244415,
"learning_rate": 0.0008,
"loss": 2.030498743057251,
"step": 10730
},
{
"epoch": 0.6504242424242425,
"grad_norm": 0.039951831102371216,
"learning_rate": 0.0008,
"loss": 2.034471273422241,
"step": 10732
},
{
"epoch": 0.6505454545454545,
"grad_norm": 0.04621196165680885,
"learning_rate": 0.0008,
"loss": 2.0115795135498047,
"step": 10734
},
{
"epoch": 0.6506666666666666,
"grad_norm": 0.04057440906763077,
"learning_rate": 0.0008,
"loss": 2.1017959117889404,
"step": 10736
},
{
"epoch": 0.6507878787878788,
"grad_norm": 0.039708029478788376,
"learning_rate": 0.0008,
"loss": 2.0384087562561035,
"step": 10738
},
{
"epoch": 0.6509090909090909,
"grad_norm": 0.038830503821372986,
"learning_rate": 0.0008,
"loss": 2.005049705505371,
"step": 10740
},
{
"epoch": 0.6510303030303031,
"grad_norm": 0.04296210780739784,
"learning_rate": 0.0008,
"loss": 2.0065579414367676,
"step": 10742
},
{
"epoch": 0.6511515151515151,
"grad_norm": 0.04391971975564957,
"learning_rate": 0.0008,
"loss": 2.040971040725708,
"step": 10744
},
{
"epoch": 0.6512727272727272,
"grad_norm": 0.05948328971862793,
"learning_rate": 0.0008,
"loss": 2.081846237182617,
"step": 10746
},
{
"epoch": 0.6513939393939394,
"grad_norm": 0.04808294028043747,
"learning_rate": 0.0008,
"loss": 2.0431559085845947,
"step": 10748
},
{
"epoch": 0.6515151515151515,
"grad_norm": 0.050048600882291794,
"learning_rate": 0.0008,
"loss": 2.0733182430267334,
"step": 10750
},
{
"epoch": 0.6516363636363637,
"grad_norm": 0.05559591203927994,
"learning_rate": 0.0008,
"loss": 2.044760227203369,
"step": 10752
},
{
"epoch": 0.6517575757575758,
"grad_norm": 0.04782276973128319,
"learning_rate": 0.0008,
"loss": 2.0343923568725586,
"step": 10754
},
{
"epoch": 0.6518787878787878,
"grad_norm": 0.05471215024590492,
"learning_rate": 0.0008,
"loss": 2.059448003768921,
"step": 10756
},
{
"epoch": 0.652,
"grad_norm": 0.04446481168270111,
"learning_rate": 0.0008,
"loss": 2.0740795135498047,
"step": 10758
},
{
"epoch": 0.6521212121212121,
"grad_norm": 0.04116393253207207,
"learning_rate": 0.0008,
"loss": 2.057861804962158,
"step": 10760
},
{
"epoch": 0.6522424242424243,
"grad_norm": 0.039303794503211975,
"learning_rate": 0.0008,
"loss": 2.017775058746338,
"step": 10762
},
{
"epoch": 0.6523636363636364,
"grad_norm": 0.04098054766654968,
"learning_rate": 0.0008,
"loss": 1.9899159669876099,
"step": 10764
},
{
"epoch": 0.6524848484848484,
"grad_norm": 0.0403890386223793,
"learning_rate": 0.0008,
"loss": 2.0145535469055176,
"step": 10766
},
{
"epoch": 0.6526060606060606,
"grad_norm": 0.04087986424565315,
"learning_rate": 0.0008,
"loss": 2.047701597213745,
"step": 10768
},
{
"epoch": 0.6527272727272727,
"grad_norm": 0.03990672528743744,
"learning_rate": 0.0008,
"loss": 2.048370122909546,
"step": 10770
},
{
"epoch": 0.6528484848484849,
"grad_norm": 0.050464652478694916,
"learning_rate": 0.0008,
"loss": 2.046814203262329,
"step": 10772
},
{
"epoch": 0.652969696969697,
"grad_norm": 0.04444744065403938,
"learning_rate": 0.0008,
"loss": 2.0539350509643555,
"step": 10774
},
{
"epoch": 0.653090909090909,
"grad_norm": 0.052913419902324677,
"learning_rate": 0.0008,
"loss": 2.018432855606079,
"step": 10776
},
{
"epoch": 0.6532121212121212,
"grad_norm": 0.0515458919107914,
"learning_rate": 0.0008,
"loss": 2.0259180068969727,
"step": 10778
},
{
"epoch": 0.6533333333333333,
"grad_norm": 0.049490995705127716,
"learning_rate": 0.0008,
"loss": 2.04024076461792,
"step": 10780
},
{
"epoch": 0.6534545454545454,
"grad_norm": 0.04688310623168945,
"learning_rate": 0.0008,
"loss": 1.9782994985580444,
"step": 10782
},
{
"epoch": 0.6535757575757576,
"grad_norm": 0.04161975532770157,
"learning_rate": 0.0008,
"loss": 2.0940661430358887,
"step": 10784
},
{
"epoch": 0.6536969696969697,
"grad_norm": 0.044266942888498306,
"learning_rate": 0.0008,
"loss": 2.0579662322998047,
"step": 10786
},
{
"epoch": 0.6538181818181819,
"grad_norm": 0.04917491227388382,
"learning_rate": 0.0008,
"loss": 2.037888288497925,
"step": 10788
},
{
"epoch": 0.6539393939393939,
"grad_norm": 0.041098956018686295,
"learning_rate": 0.0008,
"loss": 2.1123814582824707,
"step": 10790
},
{
"epoch": 0.654060606060606,
"grad_norm": 0.041467055678367615,
"learning_rate": 0.0008,
"loss": 2.041828155517578,
"step": 10792
},
{
"epoch": 0.6541818181818182,
"grad_norm": 0.040117256343364716,
"learning_rate": 0.0008,
"loss": 2.052750587463379,
"step": 10794
},
{
"epoch": 0.6543030303030303,
"grad_norm": 0.04312734678387642,
"learning_rate": 0.0008,
"loss": 1.9649348258972168,
"step": 10796
},
{
"epoch": 0.6544242424242425,
"grad_norm": 0.04347332939505577,
"learning_rate": 0.0008,
"loss": 2.0498383045196533,
"step": 10798
},
{
"epoch": 0.6545454545454545,
"grad_norm": 0.04807496815919876,
"learning_rate": 0.0008,
"loss": 1.9988189935684204,
"step": 10800
},
{
"epoch": 0.6546666666666666,
"grad_norm": 0.04069000110030174,
"learning_rate": 0.0008,
"loss": 2.0007851123809814,
"step": 10802
},
{
"epoch": 0.6547878787878788,
"grad_norm": 0.042377639561891556,
"learning_rate": 0.0008,
"loss": 2.01989483833313,
"step": 10804
},
{
"epoch": 0.6549090909090909,
"grad_norm": 0.04787394404411316,
"learning_rate": 0.0008,
"loss": 2.03580379486084,
"step": 10806
},
{
"epoch": 0.6550303030303031,
"grad_norm": 0.051159609109163284,
"learning_rate": 0.0008,
"loss": 2.0287280082702637,
"step": 10808
},
{
"epoch": 0.6551515151515152,
"grad_norm": 0.05814361572265625,
"learning_rate": 0.0008,
"loss": 2.032425880432129,
"step": 10810
},
{
"epoch": 0.6552727272727272,
"grad_norm": 0.05335332825779915,
"learning_rate": 0.0008,
"loss": 2.028049945831299,
"step": 10812
},
{
"epoch": 0.6553939393939394,
"grad_norm": 0.04725004360079765,
"learning_rate": 0.0008,
"loss": 2.0480637550354004,
"step": 10814
},
{
"epoch": 0.6555151515151515,
"grad_norm": 0.04685915634036064,
"learning_rate": 0.0008,
"loss": 1.9826489686965942,
"step": 10816
},
{
"epoch": 0.6556363636363637,
"grad_norm": 0.04343322664499283,
"learning_rate": 0.0008,
"loss": 2.0558595657348633,
"step": 10818
},
{
"epoch": 0.6557575757575758,
"grad_norm": 0.04040123522281647,
"learning_rate": 0.0008,
"loss": 1.9916883707046509,
"step": 10820
},
{
"epoch": 0.6558787878787878,
"grad_norm": 0.041586559265851974,
"learning_rate": 0.0008,
"loss": 1.9251823425292969,
"step": 10822
},
{
"epoch": 0.656,
"grad_norm": 0.04132063314318657,
"learning_rate": 0.0008,
"loss": 2.001948356628418,
"step": 10824
},
{
"epoch": 0.6561212121212121,
"grad_norm": 0.04148130863904953,
"learning_rate": 0.0008,
"loss": 2.080517530441284,
"step": 10826
},
{
"epoch": 0.6562424242424243,
"grad_norm": 0.04291995242238045,
"learning_rate": 0.0008,
"loss": 2.0091629028320312,
"step": 10828
},
{
"epoch": 0.6563636363636364,
"grad_norm": 0.04585760086774826,
"learning_rate": 0.0008,
"loss": 2.000497341156006,
"step": 10830
},
{
"epoch": 0.6564848484848484,
"grad_norm": 0.040265459567308426,
"learning_rate": 0.0008,
"loss": 1.998470664024353,
"step": 10832
},
{
"epoch": 0.6566060606060606,
"grad_norm": 0.04750536382198334,
"learning_rate": 0.0008,
"loss": 2.1071889400482178,
"step": 10834
},
{
"epoch": 0.6567272727272727,
"grad_norm": 0.04662739858031273,
"learning_rate": 0.0008,
"loss": 2.069239616394043,
"step": 10836
},
{
"epoch": 0.6568484848484849,
"grad_norm": 0.04183643311262131,
"learning_rate": 0.0008,
"loss": 2.0273702144622803,
"step": 10838
},
{
"epoch": 0.656969696969697,
"grad_norm": 0.04290969669818878,
"learning_rate": 0.0008,
"loss": 2.0333034992218018,
"step": 10840
},
{
"epoch": 0.6570909090909091,
"grad_norm": 0.04519156366586685,
"learning_rate": 0.0008,
"loss": 2.001948118209839,
"step": 10842
},
{
"epoch": 0.6572121212121212,
"grad_norm": 0.04118402674794197,
"learning_rate": 0.0008,
"loss": 2.016139268875122,
"step": 10844
},
{
"epoch": 0.6573333333333333,
"grad_norm": 0.04067740589380264,
"learning_rate": 0.0008,
"loss": 2.0941314697265625,
"step": 10846
},
{
"epoch": 0.6574545454545454,
"grad_norm": 0.04047587141394615,
"learning_rate": 0.0008,
"loss": 2.0311012268066406,
"step": 10848
},
{
"epoch": 0.6575757575757576,
"grad_norm": 0.04020274430513382,
"learning_rate": 0.0008,
"loss": 1.9940794706344604,
"step": 10850
},
{
"epoch": 0.6576969696969697,
"grad_norm": 0.04009728506207466,
"learning_rate": 0.0008,
"loss": 1.9516959190368652,
"step": 10852
},
{
"epoch": 0.6578181818181819,
"grad_norm": 0.04066215083003044,
"learning_rate": 0.0008,
"loss": 2.014478921890259,
"step": 10854
},
{
"epoch": 0.6579393939393939,
"grad_norm": 0.0406780019402504,
"learning_rate": 0.0008,
"loss": 2.0091662406921387,
"step": 10856
},
{
"epoch": 0.658060606060606,
"grad_norm": 0.04166564717888832,
"learning_rate": 0.0008,
"loss": 2.03719425201416,
"step": 10858
},
{
"epoch": 0.6581818181818182,
"grad_norm": 0.040419746190309525,
"learning_rate": 0.0008,
"loss": 2.007493019104004,
"step": 10860
},
{
"epoch": 0.6583030303030303,
"grad_norm": 0.04218980297446251,
"learning_rate": 0.0008,
"loss": 2.030339241027832,
"step": 10862
},
{
"epoch": 0.6584242424242425,
"grad_norm": 0.054844584316015244,
"learning_rate": 0.0008,
"loss": 2.0588619709014893,
"step": 10864
},
{
"epoch": 0.6585454545454545,
"grad_norm": 0.041454002261161804,
"learning_rate": 0.0008,
"loss": 2.1113626956939697,
"step": 10866
},
{
"epoch": 0.6586666666666666,
"grad_norm": 0.04272955656051636,
"learning_rate": 0.0008,
"loss": 2.075204849243164,
"step": 10868
},
{
"epoch": 0.6587878787878788,
"grad_norm": 0.04071085527539253,
"learning_rate": 0.0008,
"loss": 2.0712575912475586,
"step": 10870
},
{
"epoch": 0.6589090909090909,
"grad_norm": 0.039184924215078354,
"learning_rate": 0.0008,
"loss": 1.9795068502426147,
"step": 10872
},
{
"epoch": 0.6590303030303031,
"grad_norm": 0.04285260662436485,
"learning_rate": 0.0008,
"loss": 2.1024229526519775,
"step": 10874
},
{
"epoch": 0.6591515151515152,
"grad_norm": 0.039120499044656754,
"learning_rate": 0.0008,
"loss": 1.9766298532485962,
"step": 10876
},
{
"epoch": 0.6592727272727272,
"grad_norm": 0.04377676919102669,
"learning_rate": 0.0008,
"loss": 2.03830623626709,
"step": 10878
},
{
"epoch": 0.6593939393939394,
"grad_norm": 0.04031244292855263,
"learning_rate": 0.0008,
"loss": 2.0210888385772705,
"step": 10880
},
{
"epoch": 0.6595151515151515,
"grad_norm": 0.04080049321055412,
"learning_rate": 0.0008,
"loss": 1.9913820028305054,
"step": 10882
},
{
"epoch": 0.6596363636363637,
"grad_norm": 0.04067091643810272,
"learning_rate": 0.0008,
"loss": 2.0245163440704346,
"step": 10884
},
{
"epoch": 0.6597575757575758,
"grad_norm": 0.04470154270529747,
"learning_rate": 0.0008,
"loss": 2.0059282779693604,
"step": 10886
},
{
"epoch": 0.6598787878787878,
"grad_norm": 0.04656121879816055,
"learning_rate": 0.0008,
"loss": 1.9944771528244019,
"step": 10888
},
{
"epoch": 0.66,
"grad_norm": 0.04032089561223984,
"learning_rate": 0.0008,
"loss": 2.00691556930542,
"step": 10890
},
{
"epoch": 0.6601212121212121,
"grad_norm": 0.040887895971536636,
"learning_rate": 0.0008,
"loss": 2.052168607711792,
"step": 10892
},
{
"epoch": 0.6602424242424243,
"grad_norm": 0.039522625505924225,
"learning_rate": 0.0008,
"loss": 2.0616023540496826,
"step": 10894
},
{
"epoch": 0.6603636363636364,
"grad_norm": 0.03912054002285004,
"learning_rate": 0.0008,
"loss": 2.0307090282440186,
"step": 10896
},
{
"epoch": 0.6604848484848485,
"grad_norm": 0.04143983870744705,
"learning_rate": 0.0008,
"loss": 2.025470018386841,
"step": 10898
},
{
"epoch": 0.6606060606060606,
"grad_norm": 0.04118170216679573,
"learning_rate": 0.0008,
"loss": 2.01487398147583,
"step": 10900
},
{
"epoch": 0.6607272727272727,
"grad_norm": 0.042949896305799484,
"learning_rate": 0.0008,
"loss": 1.9856985807418823,
"step": 10902
},
{
"epoch": 0.6608484848484848,
"grad_norm": 0.04617304354906082,
"learning_rate": 0.0008,
"loss": 2.068222761154175,
"step": 10904
},
{
"epoch": 0.660969696969697,
"grad_norm": 0.04603615775704384,
"learning_rate": 0.0008,
"loss": 2.0357656478881836,
"step": 10906
},
{
"epoch": 0.6610909090909091,
"grad_norm": 0.060139790177345276,
"learning_rate": 0.0008,
"loss": 2.0588741302490234,
"step": 10908
},
{
"epoch": 0.6612121212121213,
"grad_norm": 0.04298872873187065,
"learning_rate": 0.0008,
"loss": 2.0485482215881348,
"step": 10910
},
{
"epoch": 0.6613333333333333,
"grad_norm": 0.04325520247220993,
"learning_rate": 0.0008,
"loss": 1.9987165927886963,
"step": 10912
},
{
"epoch": 0.6614545454545454,
"grad_norm": 0.04252750054001808,
"learning_rate": 0.0008,
"loss": 2.0140461921691895,
"step": 10914
},
{
"epoch": 0.6615757575757576,
"grad_norm": 0.04449233040213585,
"learning_rate": 0.0008,
"loss": 2.0634655952453613,
"step": 10916
},
{
"epoch": 0.6616969696969697,
"grad_norm": 0.048924773931503296,
"learning_rate": 0.0008,
"loss": 2.0199391841888428,
"step": 10918
},
{
"epoch": 0.6618181818181819,
"grad_norm": 0.039753399789333344,
"learning_rate": 0.0008,
"loss": 2.0231871604919434,
"step": 10920
},
{
"epoch": 0.6619393939393939,
"grad_norm": 0.04270562529563904,
"learning_rate": 0.0008,
"loss": 2.1061863899230957,
"step": 10922
},
{
"epoch": 0.662060606060606,
"grad_norm": 0.04507264867424965,
"learning_rate": 0.0008,
"loss": 2.085556983947754,
"step": 10924
},
{
"epoch": 0.6621818181818182,
"grad_norm": 0.04054282233119011,
"learning_rate": 0.0008,
"loss": 2.0065135955810547,
"step": 10926
},
{
"epoch": 0.6623030303030303,
"grad_norm": 0.04424315690994263,
"learning_rate": 0.0008,
"loss": 2.0174946784973145,
"step": 10928
},
{
"epoch": 0.6624242424242425,
"grad_norm": 0.058525342494249344,
"learning_rate": 0.0008,
"loss": 2.042369842529297,
"step": 10930
},
{
"epoch": 0.6625454545454545,
"grad_norm": 0.04673735797405243,
"learning_rate": 0.0008,
"loss": 2.0127978324890137,
"step": 10932
},
{
"epoch": 0.6626666666666666,
"grad_norm": 0.042371198534965515,
"learning_rate": 0.0008,
"loss": 2.036818742752075,
"step": 10934
},
{
"epoch": 0.6627878787878788,
"grad_norm": 0.04471875727176666,
"learning_rate": 0.0008,
"loss": 1.9588984251022339,
"step": 10936
},
{
"epoch": 0.6629090909090909,
"grad_norm": 0.04103919491171837,
"learning_rate": 0.0008,
"loss": 2.0237951278686523,
"step": 10938
},
{
"epoch": 0.6630303030303031,
"grad_norm": 0.041774895042181015,
"learning_rate": 0.0008,
"loss": 2.083514451980591,
"step": 10940
},
{
"epoch": 0.6631515151515152,
"grad_norm": 0.042551953345537186,
"learning_rate": 0.0008,
"loss": 1.9825395345687866,
"step": 10942
},
{
"epoch": 0.6632727272727272,
"grad_norm": 0.04136151820421219,
"learning_rate": 0.0008,
"loss": 2.0181610584259033,
"step": 10944
},
{
"epoch": 0.6633939393939394,
"grad_norm": 0.04275251179933548,
"learning_rate": 0.0008,
"loss": 2.028324842453003,
"step": 10946
},
{
"epoch": 0.6635151515151515,
"grad_norm": 0.04658668115735054,
"learning_rate": 0.0008,
"loss": 2.0803375244140625,
"step": 10948
},
{
"epoch": 0.6636363636363637,
"grad_norm": 0.040637195110321045,
"learning_rate": 0.0008,
"loss": 2.026167631149292,
"step": 10950
},
{
"epoch": 0.6637575757575758,
"grad_norm": 0.0492960587143898,
"learning_rate": 0.0008,
"loss": 2.084536552429199,
"step": 10952
},
{
"epoch": 0.6638787878787878,
"grad_norm": 0.04276944324374199,
"learning_rate": 0.0008,
"loss": 1.9986047744750977,
"step": 10954
},
{
"epoch": 0.664,
"grad_norm": 0.04245857894420624,
"learning_rate": 0.0008,
"loss": 2.0778160095214844,
"step": 10956
},
{
"epoch": 0.6641212121212121,
"grad_norm": 0.041922200471162796,
"learning_rate": 0.0008,
"loss": 1.997550129890442,
"step": 10958
},
{
"epoch": 0.6642424242424242,
"grad_norm": 0.04006035625934601,
"learning_rate": 0.0008,
"loss": 2.063826560974121,
"step": 10960
},
{
"epoch": 0.6643636363636364,
"grad_norm": 0.04036014899611473,
"learning_rate": 0.0008,
"loss": 2.0003902912139893,
"step": 10962
},
{
"epoch": 0.6644848484848485,
"grad_norm": 0.046745527535676956,
"learning_rate": 0.0008,
"loss": 2.0098187923431396,
"step": 10964
},
{
"epoch": 0.6646060606060606,
"grad_norm": 0.04261269420385361,
"learning_rate": 0.0008,
"loss": 1.9778540134429932,
"step": 10966
},
{
"epoch": 0.6647272727272727,
"grad_norm": 0.040851738303899765,
"learning_rate": 0.0008,
"loss": 1.958568811416626,
"step": 10968
},
{
"epoch": 0.6648484848484848,
"grad_norm": 0.040406737476587296,
"learning_rate": 0.0008,
"loss": 2.0198469161987305,
"step": 10970
},
{
"epoch": 0.664969696969697,
"grad_norm": 0.04007527604699135,
"learning_rate": 0.0008,
"loss": 2.0779905319213867,
"step": 10972
},
{
"epoch": 0.6650909090909091,
"grad_norm": 0.04341789335012436,
"learning_rate": 0.0008,
"loss": 2.082392454147339,
"step": 10974
},
{
"epoch": 0.6652121212121213,
"grad_norm": 0.04594157263636589,
"learning_rate": 0.0008,
"loss": 2.0570666790008545,
"step": 10976
},
{
"epoch": 0.6653333333333333,
"grad_norm": 0.0439421646296978,
"learning_rate": 0.0008,
"loss": 2.072160243988037,
"step": 10978
},
{
"epoch": 0.6654545454545454,
"grad_norm": 0.04070933908224106,
"learning_rate": 0.0008,
"loss": 2.058485269546509,
"step": 10980
},
{
"epoch": 0.6655757575757576,
"grad_norm": 0.0393497534096241,
"learning_rate": 0.0008,
"loss": 1.9837759733200073,
"step": 10982
},
{
"epoch": 0.6656969696969697,
"grad_norm": 0.03906195983290672,
"learning_rate": 0.0008,
"loss": 1.9683446884155273,
"step": 10984
},
{
"epoch": 0.6658181818181819,
"grad_norm": 0.041105762124061584,
"learning_rate": 0.0008,
"loss": 2.013895273208618,
"step": 10986
},
{
"epoch": 0.6659393939393939,
"grad_norm": 0.03981298953294754,
"learning_rate": 0.0008,
"loss": 2.021681070327759,
"step": 10988
},
{
"epoch": 0.666060606060606,
"grad_norm": 0.04099587723612785,
"learning_rate": 0.0008,
"loss": 2.032620429992676,
"step": 10990
},
{
"epoch": 0.6661818181818182,
"grad_norm": 0.04721679911017418,
"learning_rate": 0.0008,
"loss": 2.057673215866089,
"step": 10992
},
{
"epoch": 0.6663030303030303,
"grad_norm": 0.03910978510975838,
"learning_rate": 0.0008,
"loss": 2.0460333824157715,
"step": 10994
},
{
"epoch": 0.6664242424242425,
"grad_norm": 0.040756095200777054,
"learning_rate": 0.0008,
"loss": 1.973179817199707,
"step": 10996
},
{
"epoch": 0.6665454545454546,
"grad_norm": 0.0417574979364872,
"learning_rate": 0.0008,
"loss": 2.021510362625122,
"step": 10998
},
{
"epoch": 0.6666666666666666,
"grad_norm": 0.044432178139686584,
"learning_rate": 0.0008,
"loss": 2.0282280445098877,
"step": 11000
},
{
"epoch": 0.6667878787878788,
"grad_norm": 0.04066716507077217,
"learning_rate": 0.0008,
"loss": 1.9848284721374512,
"step": 11002
},
{
"epoch": 0.6669090909090909,
"grad_norm": 0.04544346034526825,
"learning_rate": 0.0008,
"loss": 2.0421011447906494,
"step": 11004
},
{
"epoch": 0.6670303030303031,
"grad_norm": 0.04027484357357025,
"learning_rate": 0.0008,
"loss": 2.0200603008270264,
"step": 11006
},
{
"epoch": 0.6671515151515152,
"grad_norm": 0.04079143702983856,
"learning_rate": 0.0008,
"loss": 2.033811330795288,
"step": 11008
},
{
"epoch": 0.6672727272727272,
"grad_norm": 0.04092727228999138,
"learning_rate": 0.0008,
"loss": 2.011979818344116,
"step": 11010
},
{
"epoch": 0.6673939393939394,
"grad_norm": 0.04424966126680374,
"learning_rate": 0.0008,
"loss": 2.0702295303344727,
"step": 11012
},
{
"epoch": 0.6675151515151515,
"grad_norm": 0.04204338788986206,
"learning_rate": 0.0008,
"loss": 2.037323236465454,
"step": 11014
},
{
"epoch": 0.6676363636363636,
"grad_norm": 0.05556361377239227,
"learning_rate": 0.0008,
"loss": 2.0236854553222656,
"step": 11016
},
{
"epoch": 0.6677575757575758,
"grad_norm": 0.06354600936174393,
"learning_rate": 0.0008,
"loss": 2.08034086227417,
"step": 11018
},
{
"epoch": 0.6678787878787878,
"grad_norm": 0.06018998101353645,
"learning_rate": 0.0008,
"loss": 2.0985190868377686,
"step": 11020
},
{
"epoch": 0.668,
"grad_norm": 0.04941658675670624,
"learning_rate": 0.0008,
"loss": 2.039332866668701,
"step": 11022
},
{
"epoch": 0.6681212121212121,
"grad_norm": 0.04280613362789154,
"learning_rate": 0.0008,
"loss": 2.0975897312164307,
"step": 11024
},
{
"epoch": 0.6682424242424242,
"grad_norm": 0.04590749740600586,
"learning_rate": 0.0008,
"loss": 2.0228142738342285,
"step": 11026
},
{
"epoch": 0.6683636363636364,
"grad_norm": 0.042043786495923996,
"learning_rate": 0.0008,
"loss": 2.036153554916382,
"step": 11028
},
{
"epoch": 0.6684848484848485,
"grad_norm": 0.043300919234752655,
"learning_rate": 0.0008,
"loss": 2.058351516723633,
"step": 11030
},
{
"epoch": 0.6686060606060606,
"grad_norm": 0.039242010563611984,
"learning_rate": 0.0008,
"loss": 2.066991090774536,
"step": 11032
},
{
"epoch": 0.6687272727272727,
"grad_norm": 0.045957308262586594,
"learning_rate": 0.0008,
"loss": 1.9603499174118042,
"step": 11034
},
{
"epoch": 0.6688484848484848,
"grad_norm": 0.042770642787218094,
"learning_rate": 0.0008,
"loss": 2.005218982696533,
"step": 11036
},
{
"epoch": 0.668969696969697,
"grad_norm": 0.05434314161539078,
"learning_rate": 0.0008,
"loss": 2.0221214294433594,
"step": 11038
},
{
"epoch": 0.6690909090909091,
"grad_norm": 0.07544242590665817,
"learning_rate": 0.0008,
"loss": 1.992893099784851,
"step": 11040
},
{
"epoch": 0.6692121212121213,
"grad_norm": 0.06005619838833809,
"learning_rate": 0.0008,
"loss": 2.0028481483459473,
"step": 11042
},
{
"epoch": 0.6693333333333333,
"grad_norm": 0.06455206125974655,
"learning_rate": 0.0008,
"loss": 2.0674564838409424,
"step": 11044
},
{
"epoch": 0.6694545454545454,
"grad_norm": 0.07924848049879074,
"learning_rate": 0.0008,
"loss": 2.0387954711914062,
"step": 11046
},
{
"epoch": 0.6695757575757576,
"grad_norm": 0.08137133717536926,
"learning_rate": 0.0008,
"loss": 2.0575501918792725,
"step": 11048
},
{
"epoch": 0.6696969696969697,
"grad_norm": 0.07442744076251984,
"learning_rate": 0.0008,
"loss": 2.0211551189422607,
"step": 11050
},
{
"epoch": 0.6698181818181819,
"grad_norm": 0.060952406376600266,
"learning_rate": 0.0008,
"loss": 2.000669479370117,
"step": 11052
},
{
"epoch": 0.669939393939394,
"grad_norm": 0.04153238981962204,
"learning_rate": 0.0008,
"loss": 2.074366807937622,
"step": 11054
},
{
"epoch": 0.670060606060606,
"grad_norm": 0.041037801653146744,
"learning_rate": 0.0008,
"loss": 2.01877760887146,
"step": 11056
},
{
"epoch": 0.6701818181818182,
"grad_norm": 0.03962745890021324,
"learning_rate": 0.0008,
"loss": 2.0404775142669678,
"step": 11058
},
{
"epoch": 0.6703030303030303,
"grad_norm": 0.05148822441697121,
"learning_rate": 0.0008,
"loss": 2.016575574874878,
"step": 11060
},
{
"epoch": 0.6704242424242425,
"grad_norm": 0.050873853266239166,
"learning_rate": 0.0008,
"loss": 2.0862529277801514,
"step": 11062
},
{
"epoch": 0.6705454545454546,
"grad_norm": 0.04268103837966919,
"learning_rate": 0.0008,
"loss": 2.048017978668213,
"step": 11064
},
{
"epoch": 0.6706666666666666,
"grad_norm": 0.04235207289457321,
"learning_rate": 0.0008,
"loss": 2.0247867107391357,
"step": 11066
},
{
"epoch": 0.6707878787878788,
"grad_norm": 0.042169246822595596,
"learning_rate": 0.0008,
"loss": 1.9738373756408691,
"step": 11068
},
{
"epoch": 0.6709090909090909,
"grad_norm": 0.045375943183898926,
"learning_rate": 0.0008,
"loss": 1.939339518547058,
"step": 11070
},
{
"epoch": 0.671030303030303,
"grad_norm": 0.040882982313632965,
"learning_rate": 0.0008,
"loss": 2.061352252960205,
"step": 11072
},
{
"epoch": 0.6711515151515152,
"grad_norm": 0.03959853574633598,
"learning_rate": 0.0008,
"loss": 2.0246775150299072,
"step": 11074
},
{
"epoch": 0.6712727272727272,
"grad_norm": 0.04140223190188408,
"learning_rate": 0.0008,
"loss": 2.075302839279175,
"step": 11076
},
{
"epoch": 0.6713939393939394,
"grad_norm": 0.03912576660513878,
"learning_rate": 0.0008,
"loss": 2.038529396057129,
"step": 11078
},
{
"epoch": 0.6715151515151515,
"grad_norm": 0.040821392089128494,
"learning_rate": 0.0008,
"loss": 2.0469188690185547,
"step": 11080
},
{
"epoch": 0.6716363636363636,
"grad_norm": 0.04798780754208565,
"learning_rate": 0.0008,
"loss": 2.0002686977386475,
"step": 11082
},
{
"epoch": 0.6717575757575758,
"grad_norm": 0.054577358067035675,
"learning_rate": 0.0008,
"loss": 2.105081081390381,
"step": 11084
},
{
"epoch": 0.6718787878787879,
"grad_norm": 0.055143970996141434,
"learning_rate": 0.0008,
"loss": 1.9685285091400146,
"step": 11086
},
{
"epoch": 0.672,
"grad_norm": 0.05139290541410446,
"learning_rate": 0.0008,
"loss": 1.9984408617019653,
"step": 11088
},
{
"epoch": 0.6721212121212121,
"grad_norm": 0.05116335302591324,
"learning_rate": 0.0008,
"loss": 2.0052194595336914,
"step": 11090
},
{
"epoch": 0.6722424242424242,
"grad_norm": 0.058949898928403854,
"learning_rate": 0.0008,
"loss": 2.017319679260254,
"step": 11092
},
{
"epoch": 0.6723636363636364,
"grad_norm": 0.04878431186079979,
"learning_rate": 0.0008,
"loss": 2.090590000152588,
"step": 11094
},
{
"epoch": 0.6724848484848485,
"grad_norm": 0.0435309037566185,
"learning_rate": 0.0008,
"loss": 1.9992071390151978,
"step": 11096
},
{
"epoch": 0.6726060606060607,
"grad_norm": 0.0400727353990078,
"learning_rate": 0.0008,
"loss": 1.9950473308563232,
"step": 11098
},
{
"epoch": 0.6727272727272727,
"grad_norm": 0.04067965969443321,
"learning_rate": 0.0008,
"loss": 2.004293441772461,
"step": 11100
},
{
"epoch": 0.6728484848484848,
"grad_norm": 0.04320705682039261,
"learning_rate": 0.0008,
"loss": 2.095142126083374,
"step": 11102
},
{
"epoch": 0.672969696969697,
"grad_norm": 0.04174032434821129,
"learning_rate": 0.0008,
"loss": 2.0315935611724854,
"step": 11104
},
{
"epoch": 0.6730909090909091,
"grad_norm": 0.042101290076971054,
"learning_rate": 0.0008,
"loss": 1.976998209953308,
"step": 11106
},
{
"epoch": 0.6732121212121213,
"grad_norm": 0.038578785955905914,
"learning_rate": 0.0008,
"loss": 2.0273118019104004,
"step": 11108
},
{
"epoch": 0.6733333333333333,
"grad_norm": 0.04163231700658798,
"learning_rate": 0.0008,
"loss": 2.042557954788208,
"step": 11110
},
{
"epoch": 0.6734545454545454,
"grad_norm": 0.06099005043506622,
"learning_rate": 0.0008,
"loss": 2.0406885147094727,
"step": 11112
},
{
"epoch": 0.6735757575757576,
"grad_norm": 0.05986581742763519,
"learning_rate": 0.0008,
"loss": 2.044405698776245,
"step": 11114
},
{
"epoch": 0.6736969696969697,
"grad_norm": 0.04928968846797943,
"learning_rate": 0.0008,
"loss": 1.9767111539840698,
"step": 11116
},
{
"epoch": 0.6738181818181819,
"grad_norm": 0.04589634761214256,
"learning_rate": 0.0008,
"loss": 2.1036431789398193,
"step": 11118
},
{
"epoch": 0.673939393939394,
"grad_norm": 0.04385237395763397,
"learning_rate": 0.0008,
"loss": 2.036595106124878,
"step": 11120
},
{
"epoch": 0.674060606060606,
"grad_norm": 0.03992724418640137,
"learning_rate": 0.0008,
"loss": 2.009917974472046,
"step": 11122
},
{
"epoch": 0.6741818181818182,
"grad_norm": 0.042954329401254654,
"learning_rate": 0.0008,
"loss": 2.0617904663085938,
"step": 11124
},
{
"epoch": 0.6743030303030303,
"grad_norm": 0.041542667895555496,
"learning_rate": 0.0008,
"loss": 1.9464854001998901,
"step": 11126
},
{
"epoch": 0.6744242424242424,
"grad_norm": 0.042650721967220306,
"learning_rate": 0.0008,
"loss": 2.0234835147857666,
"step": 11128
},
{
"epoch": 0.6745454545454546,
"grad_norm": 0.03841077536344528,
"learning_rate": 0.0008,
"loss": 2.0541274547576904,
"step": 11130
},
{
"epoch": 0.6746666666666666,
"grad_norm": 0.04099129140377045,
"learning_rate": 0.0008,
"loss": 2.0375638008117676,
"step": 11132
},
{
"epoch": 0.6747878787878788,
"grad_norm": 0.04196937754750252,
"learning_rate": 0.0008,
"loss": 2.015850305557251,
"step": 11134
},
{
"epoch": 0.6749090909090909,
"grad_norm": 0.040151458233594894,
"learning_rate": 0.0008,
"loss": 2.076544761657715,
"step": 11136
},
{
"epoch": 0.675030303030303,
"grad_norm": 0.04073752835392952,
"learning_rate": 0.0008,
"loss": 2.010023355484009,
"step": 11138
},
{
"epoch": 0.6751515151515152,
"grad_norm": 0.040370646864175797,
"learning_rate": 0.0008,
"loss": 2.0443899631500244,
"step": 11140
},
{
"epoch": 0.6752727272727272,
"grad_norm": 0.04113660380244255,
"learning_rate": 0.0008,
"loss": 2.0431270599365234,
"step": 11142
},
{
"epoch": 0.6753939393939394,
"grad_norm": 0.05372660979628563,
"learning_rate": 0.0008,
"loss": 1.9357813596725464,
"step": 11144
},
{
"epoch": 0.6755151515151515,
"grad_norm": 0.06788382679224014,
"learning_rate": 0.0008,
"loss": 2.0059027671813965,
"step": 11146
},
{
"epoch": 0.6756363636363636,
"grad_norm": 0.05359559506177902,
"learning_rate": 0.0008,
"loss": 2.030993938446045,
"step": 11148
},
{
"epoch": 0.6757575757575758,
"grad_norm": 0.04934390261769295,
"learning_rate": 0.0008,
"loss": 2.0346639156341553,
"step": 11150
},
{
"epoch": 0.6758787878787879,
"grad_norm": 0.05721955746412277,
"learning_rate": 0.0008,
"loss": 2.0726418495178223,
"step": 11152
},
{
"epoch": 0.676,
"grad_norm": 0.055272553116083145,
"learning_rate": 0.0008,
"loss": 1.9854755401611328,
"step": 11154
},
{
"epoch": 0.6761212121212121,
"grad_norm": 0.051482439041137695,
"learning_rate": 0.0008,
"loss": 1.9973862171173096,
"step": 11156
},
{
"epoch": 0.6762424242424242,
"grad_norm": 0.043016690760850906,
"learning_rate": 0.0008,
"loss": 2.058199882507324,
"step": 11158
},
{
"epoch": 0.6763636363636364,
"grad_norm": 0.042179953306913376,
"learning_rate": 0.0008,
"loss": 2.041982889175415,
"step": 11160
},
{
"epoch": 0.6764848484848485,
"grad_norm": 0.04187849536538124,
"learning_rate": 0.0008,
"loss": 2.0262370109558105,
"step": 11162
},
{
"epoch": 0.6766060606060607,
"grad_norm": 0.045001376420259476,
"learning_rate": 0.0008,
"loss": 2.062448740005493,
"step": 11164
},
{
"epoch": 0.6767272727272727,
"grad_norm": 0.041484858840703964,
"learning_rate": 0.0008,
"loss": 2.0610804557800293,
"step": 11166
},
{
"epoch": 0.6768484848484848,
"grad_norm": 0.043292086571455,
"learning_rate": 0.0008,
"loss": 2.0246641635894775,
"step": 11168
},
{
"epoch": 0.676969696969697,
"grad_norm": 0.045185089111328125,
"learning_rate": 0.0008,
"loss": 2.025740623474121,
"step": 11170
},
{
"epoch": 0.6770909090909091,
"grad_norm": 0.039957668632268906,
"learning_rate": 0.0008,
"loss": 2.094499111175537,
"step": 11172
},
{
"epoch": 0.6772121212121213,
"grad_norm": 0.038854826241731644,
"learning_rate": 0.0008,
"loss": 2.046625852584839,
"step": 11174
},
{
"epoch": 0.6773333333333333,
"grad_norm": 0.04013057053089142,
"learning_rate": 0.0008,
"loss": 2.0167412757873535,
"step": 11176
},
{
"epoch": 0.6774545454545454,
"grad_norm": 0.05085596814751625,
"learning_rate": 0.0008,
"loss": 1.9909160137176514,
"step": 11178
},
{
"epoch": 0.6775757575757576,
"grad_norm": 0.041447583585977554,
"learning_rate": 0.0008,
"loss": 1.974941372871399,
"step": 11180
},
{
"epoch": 0.6776969696969697,
"grad_norm": 0.04783628135919571,
"learning_rate": 0.0008,
"loss": 2.011061191558838,
"step": 11182
},
{
"epoch": 0.6778181818181818,
"grad_norm": 0.04320615157485008,
"learning_rate": 0.0008,
"loss": 1.9351414442062378,
"step": 11184
},
{
"epoch": 0.677939393939394,
"grad_norm": 0.040527619421482086,
"learning_rate": 0.0008,
"loss": 2.0472068786621094,
"step": 11186
},
{
"epoch": 0.678060606060606,
"grad_norm": 0.04048457741737366,
"learning_rate": 0.0008,
"loss": 2.0137083530426025,
"step": 11188
},
{
"epoch": 0.6781818181818182,
"grad_norm": 0.04002455249428749,
"learning_rate": 0.0008,
"loss": 1.9925742149353027,
"step": 11190
},
{
"epoch": 0.6783030303030303,
"grad_norm": 0.040721647441387177,
"learning_rate": 0.0008,
"loss": 2.0002450942993164,
"step": 11192
},
{
"epoch": 0.6784242424242424,
"grad_norm": 0.05126996710896492,
"learning_rate": 0.0008,
"loss": 2.0212249755859375,
"step": 11194
},
{
"epoch": 0.6785454545454546,
"grad_norm": 0.04560898616909981,
"learning_rate": 0.0008,
"loss": 2.03940749168396,
"step": 11196
},
{
"epoch": 0.6786666666666666,
"grad_norm": 0.056464292109012604,
"learning_rate": 0.0008,
"loss": 2.030423402786255,
"step": 11198
},
{
"epoch": 0.6787878787878788,
"grad_norm": 0.06508172303438187,
"learning_rate": 0.0008,
"loss": 2.0350072383880615,
"step": 11200
},
{
"epoch": 0.6789090909090909,
"grad_norm": 0.07504697144031525,
"learning_rate": 0.0008,
"loss": 1.9933651685714722,
"step": 11202
},
{
"epoch": 0.679030303030303,
"grad_norm": 0.06959469616413116,
"learning_rate": 0.0008,
"loss": 2.0146775245666504,
"step": 11204
},
{
"epoch": 0.6791515151515152,
"grad_norm": 0.07346043735742569,
"learning_rate": 0.0008,
"loss": 1.9896413087844849,
"step": 11206
},
{
"epoch": 0.6792727272727273,
"grad_norm": 0.07482420653104782,
"learning_rate": 0.0008,
"loss": 2.046022415161133,
"step": 11208
},
{
"epoch": 0.6793939393939394,
"grad_norm": 0.06833110749721527,
"learning_rate": 0.0008,
"loss": 2.0335121154785156,
"step": 11210
},
{
"epoch": 0.6795151515151515,
"grad_norm": 0.05245792865753174,
"learning_rate": 0.0008,
"loss": 1.9401689767837524,
"step": 11212
},
{
"epoch": 0.6796363636363636,
"grad_norm": 0.05001083016395569,
"learning_rate": 0.0008,
"loss": 2.0655369758605957,
"step": 11214
},
{
"epoch": 0.6797575757575758,
"grad_norm": 0.040184952318668365,
"learning_rate": 0.0008,
"loss": 2.070976972579956,
"step": 11216
},
{
"epoch": 0.6798787878787879,
"grad_norm": 0.04091757908463478,
"learning_rate": 0.0008,
"loss": 2.0698258876800537,
"step": 11218
},
{
"epoch": 0.68,
"grad_norm": 0.04607166349887848,
"learning_rate": 0.0008,
"loss": 2.068307399749756,
"step": 11220
},
{
"epoch": 0.6801212121212121,
"grad_norm": 0.04333265870809555,
"learning_rate": 0.0008,
"loss": 1.9937046766281128,
"step": 11222
},
{
"epoch": 0.6802424242424242,
"grad_norm": 0.052056729793548584,
"learning_rate": 0.0008,
"loss": 2.0311079025268555,
"step": 11224
},
{
"epoch": 0.6803636363636364,
"grad_norm": 0.04005021974444389,
"learning_rate": 0.0008,
"loss": 2.0030624866485596,
"step": 11226
},
{
"epoch": 0.6804848484848485,
"grad_norm": 0.044494081288576126,
"learning_rate": 0.0008,
"loss": 2.075261354446411,
"step": 11228
},
{
"epoch": 0.6806060606060607,
"grad_norm": 0.042870402336120605,
"learning_rate": 0.0008,
"loss": 1.9548594951629639,
"step": 11230
},
{
"epoch": 0.6807272727272727,
"grad_norm": 0.04566153883934021,
"learning_rate": 0.0008,
"loss": 2.065077543258667,
"step": 11232
},
{
"epoch": 0.6808484848484848,
"grad_norm": 0.04084378480911255,
"learning_rate": 0.0008,
"loss": 2.0468356609344482,
"step": 11234
},
{
"epoch": 0.680969696969697,
"grad_norm": 0.04395538941025734,
"learning_rate": 0.0008,
"loss": 2.06256103515625,
"step": 11236
},
{
"epoch": 0.6810909090909091,
"grad_norm": 0.05339490622282028,
"learning_rate": 0.0008,
"loss": 1.9922065734863281,
"step": 11238
},
{
"epoch": 0.6812121212121212,
"grad_norm": 0.057842496782541275,
"learning_rate": 0.0008,
"loss": 2.065967082977295,
"step": 11240
},
{
"epoch": 0.6813333333333333,
"grad_norm": 0.04510696232318878,
"learning_rate": 0.0008,
"loss": 1.9964497089385986,
"step": 11242
},
{
"epoch": 0.6814545454545454,
"grad_norm": 0.04903124272823334,
"learning_rate": 0.0008,
"loss": 2.093423843383789,
"step": 11244
},
{
"epoch": 0.6815757575757576,
"grad_norm": 0.04208313673734665,
"learning_rate": 0.0008,
"loss": 1.9968985319137573,
"step": 11246
},
{
"epoch": 0.6816969696969697,
"grad_norm": 0.040752485394477844,
"learning_rate": 0.0008,
"loss": 2.0151526927948,
"step": 11248
},
{
"epoch": 0.6818181818181818,
"grad_norm": 0.04042941704392433,
"learning_rate": 0.0008,
"loss": 1.9503984451293945,
"step": 11250
},
{
"epoch": 0.681939393939394,
"grad_norm": 0.03957477584481239,
"learning_rate": 0.0008,
"loss": 2.026596784591675,
"step": 11252
},
{
"epoch": 0.682060606060606,
"grad_norm": 0.04965978115797043,
"learning_rate": 0.0008,
"loss": 2.018749713897705,
"step": 11254
},
{
"epoch": 0.6821818181818182,
"grad_norm": 0.05360209941864014,
"learning_rate": 0.0008,
"loss": 2.033102512359619,
"step": 11256
},
{
"epoch": 0.6823030303030303,
"grad_norm": 0.04952362924814224,
"learning_rate": 0.0008,
"loss": 2.0597023963928223,
"step": 11258
},
{
"epoch": 0.6824242424242424,
"grad_norm": 0.048359863460063934,
"learning_rate": 0.0008,
"loss": 2.0180258750915527,
"step": 11260
},
{
"epoch": 0.6825454545454546,
"grad_norm": 0.04386774078011513,
"learning_rate": 0.0008,
"loss": 2.048546314239502,
"step": 11262
},
{
"epoch": 0.6826666666666666,
"grad_norm": 0.04150533303618431,
"learning_rate": 0.0008,
"loss": 2.0648510456085205,
"step": 11264
},
{
"epoch": 0.6827878787878788,
"grad_norm": 0.03939434513449669,
"learning_rate": 0.0008,
"loss": 2.0273044109344482,
"step": 11266
},
{
"epoch": 0.6829090909090909,
"grad_norm": 0.040085166692733765,
"learning_rate": 0.0008,
"loss": 2.069491147994995,
"step": 11268
},
{
"epoch": 0.683030303030303,
"grad_norm": 0.04018412530422211,
"learning_rate": 0.0008,
"loss": 1.956043004989624,
"step": 11270
},
{
"epoch": 0.6831515151515152,
"grad_norm": 0.04158221557736397,
"learning_rate": 0.0008,
"loss": 2.0626766681671143,
"step": 11272
},
{
"epoch": 0.6832727272727273,
"grad_norm": 0.0402897372841835,
"learning_rate": 0.0008,
"loss": 1.9543805122375488,
"step": 11274
},
{
"epoch": 0.6833939393939394,
"grad_norm": 0.03959556296467781,
"learning_rate": 0.0008,
"loss": 1.9800984859466553,
"step": 11276
},
{
"epoch": 0.6835151515151515,
"grad_norm": 0.041134320199489594,
"learning_rate": 0.0008,
"loss": 2.080040454864502,
"step": 11278
},
{
"epoch": 0.6836363636363636,
"grad_norm": 0.04031941667199135,
"learning_rate": 0.0008,
"loss": 2.0380005836486816,
"step": 11280
},
{
"epoch": 0.6837575757575758,
"grad_norm": 0.0409538708627224,
"learning_rate": 0.0008,
"loss": 2.0118467807769775,
"step": 11282
},
{
"epoch": 0.6838787878787879,
"grad_norm": 0.04608134180307388,
"learning_rate": 0.0008,
"loss": 2.0321531295776367,
"step": 11284
},
{
"epoch": 0.684,
"grad_norm": 0.043322909623384476,
"learning_rate": 0.0008,
"loss": 2.0241501331329346,
"step": 11286
},
{
"epoch": 0.6841212121212121,
"grad_norm": 0.04046950861811638,
"learning_rate": 0.0008,
"loss": 2.0120251178741455,
"step": 11288
},
{
"epoch": 0.6842424242424242,
"grad_norm": 0.04007450491189957,
"learning_rate": 0.0008,
"loss": 1.9518996477127075,
"step": 11290
},
{
"epoch": 0.6843636363636364,
"grad_norm": 0.03881354257464409,
"learning_rate": 0.0008,
"loss": 2.025144338607788,
"step": 11292
},
{
"epoch": 0.6844848484848485,
"grad_norm": 0.045541491359472275,
"learning_rate": 0.0008,
"loss": 2.0882694721221924,
"step": 11294
},
{
"epoch": 0.6846060606060606,
"grad_norm": 0.04416464641690254,
"learning_rate": 0.0008,
"loss": 2.0047788619995117,
"step": 11296
},
{
"epoch": 0.6847272727272727,
"grad_norm": 0.0440937876701355,
"learning_rate": 0.0008,
"loss": 2.032676935195923,
"step": 11298
},
{
"epoch": 0.6848484848484848,
"grad_norm": 0.0462893508374691,
"learning_rate": 0.0008,
"loss": 2.0081570148468018,
"step": 11300
},
{
"epoch": 0.684969696969697,
"grad_norm": 0.039611898362636566,
"learning_rate": 0.0008,
"loss": 2.0227651596069336,
"step": 11302
},
{
"epoch": 0.6850909090909091,
"grad_norm": 0.039169907569885254,
"learning_rate": 0.0008,
"loss": 2.06927752494812,
"step": 11304
},
{
"epoch": 0.6852121212121212,
"grad_norm": 0.04039622098207474,
"learning_rate": 0.0008,
"loss": 2.043088674545288,
"step": 11306
},
{
"epoch": 0.6853333333333333,
"grad_norm": 0.0477239154279232,
"learning_rate": 0.0008,
"loss": 2.0150649547576904,
"step": 11308
},
{
"epoch": 0.6854545454545454,
"grad_norm": 0.06262925267219543,
"learning_rate": 0.0008,
"loss": 2.004969835281372,
"step": 11310
},
{
"epoch": 0.6855757575757576,
"grad_norm": 0.04971245676279068,
"learning_rate": 0.0008,
"loss": 1.9651507139205933,
"step": 11312
},
{
"epoch": 0.6856969696969697,
"grad_norm": 0.046067044138908386,
"learning_rate": 0.0008,
"loss": 2.0113329887390137,
"step": 11314
},
{
"epoch": 0.6858181818181818,
"grad_norm": 0.05082911625504494,
"learning_rate": 0.0008,
"loss": 2.0575666427612305,
"step": 11316
},
{
"epoch": 0.685939393939394,
"grad_norm": 0.046013958752155304,
"learning_rate": 0.0008,
"loss": 1.9717183113098145,
"step": 11318
},
{
"epoch": 0.686060606060606,
"grad_norm": 0.04110277071595192,
"learning_rate": 0.0008,
"loss": 1.9905884265899658,
"step": 11320
},
{
"epoch": 0.6861818181818182,
"grad_norm": 0.042738527059555054,
"learning_rate": 0.0008,
"loss": 1.9926677942276,
"step": 11322
},
{
"epoch": 0.6863030303030303,
"grad_norm": 0.042675089091062546,
"learning_rate": 0.0008,
"loss": 1.9815571308135986,
"step": 11324
},
{
"epoch": 0.6864242424242424,
"grad_norm": 0.04054078832268715,
"learning_rate": 0.0008,
"loss": 2.0430846214294434,
"step": 11326
},
{
"epoch": 0.6865454545454546,
"grad_norm": 0.040280669927597046,
"learning_rate": 0.0008,
"loss": 1.9908064603805542,
"step": 11328
},
{
"epoch": 0.6866666666666666,
"grad_norm": 0.04064139723777771,
"learning_rate": 0.0008,
"loss": 2.0116114616394043,
"step": 11330
},
{
"epoch": 0.6867878787878788,
"grad_norm": 0.039250172674655914,
"learning_rate": 0.0008,
"loss": 2.012089252471924,
"step": 11332
},
{
"epoch": 0.6869090909090909,
"grad_norm": 0.04079850763082504,
"learning_rate": 0.0008,
"loss": 2.0682153701782227,
"step": 11334
},
{
"epoch": 0.687030303030303,
"grad_norm": 0.04164876043796539,
"learning_rate": 0.0008,
"loss": 2.0095365047454834,
"step": 11336
},
{
"epoch": 0.6871515151515152,
"grad_norm": 0.04839440435171127,
"learning_rate": 0.0008,
"loss": 2.113734722137451,
"step": 11338
},
{
"epoch": 0.6872727272727273,
"grad_norm": 0.04746574908494949,
"learning_rate": 0.0008,
"loss": 2.0253348350524902,
"step": 11340
},
{
"epoch": 0.6873939393939394,
"grad_norm": 0.04252607375383377,
"learning_rate": 0.0008,
"loss": 2.0307364463806152,
"step": 11342
},
{
"epoch": 0.6875151515151515,
"grad_norm": 0.06058284267783165,
"learning_rate": 0.0008,
"loss": 2.0018041133880615,
"step": 11344
},
{
"epoch": 0.6876363636363636,
"grad_norm": 0.05485445633530617,
"learning_rate": 0.0008,
"loss": 2.0099315643310547,
"step": 11346
},
{
"epoch": 0.6877575757575758,
"grad_norm": 0.05028192326426506,
"learning_rate": 0.0008,
"loss": 2.011496067047119,
"step": 11348
},
{
"epoch": 0.6878787878787879,
"grad_norm": 0.051163531839847565,
"learning_rate": 0.0008,
"loss": 2.040003776550293,
"step": 11350
},
{
"epoch": 0.688,
"grad_norm": 0.04551416635513306,
"learning_rate": 0.0008,
"loss": 2.066185474395752,
"step": 11352
},
{
"epoch": 0.6881212121212121,
"grad_norm": 0.04327185079455376,
"learning_rate": 0.0008,
"loss": 2.0824596881866455,
"step": 11354
},
{
"epoch": 0.6882424242424242,
"grad_norm": 0.05074770376086235,
"learning_rate": 0.0008,
"loss": 1.979599118232727,
"step": 11356
},
{
"epoch": 0.6883636363636364,
"grad_norm": 0.04598614200949669,
"learning_rate": 0.0008,
"loss": 2.065037727355957,
"step": 11358
},
{
"epoch": 0.6884848484848485,
"grad_norm": 0.0421559140086174,
"learning_rate": 0.0008,
"loss": 2.070807933807373,
"step": 11360
},
{
"epoch": 0.6886060606060606,
"grad_norm": 0.041516754776239395,
"learning_rate": 0.0008,
"loss": 2.046823501586914,
"step": 11362
},
{
"epoch": 0.6887272727272727,
"grad_norm": 0.03899841010570526,
"learning_rate": 0.0008,
"loss": 2.02215838432312,
"step": 11364
},
{
"epoch": 0.6888484848484848,
"grad_norm": 0.04019280895590782,
"learning_rate": 0.0008,
"loss": 2.0649847984313965,
"step": 11366
},
{
"epoch": 0.688969696969697,
"grad_norm": 0.04192037135362625,
"learning_rate": 0.0008,
"loss": 2.011662721633911,
"step": 11368
},
{
"epoch": 0.6890909090909091,
"grad_norm": 0.052331894636154175,
"learning_rate": 0.0008,
"loss": 2.0772619247436523,
"step": 11370
},
{
"epoch": 0.6892121212121212,
"grad_norm": 0.04527489095926285,
"learning_rate": 0.0008,
"loss": 2.003801107406616,
"step": 11372
},
{
"epoch": 0.6893333333333334,
"grad_norm": 0.04368693009018898,
"learning_rate": 0.0008,
"loss": 2.0848565101623535,
"step": 11374
},
{
"epoch": 0.6894545454545454,
"grad_norm": 0.04128141328692436,
"learning_rate": 0.0008,
"loss": 2.062854528427124,
"step": 11376
},
{
"epoch": 0.6895757575757576,
"grad_norm": 0.038360778242349625,
"learning_rate": 0.0008,
"loss": 2.0371487140655518,
"step": 11378
},
{
"epoch": 0.6896969696969697,
"grad_norm": 0.03854220360517502,
"learning_rate": 0.0008,
"loss": 1.9937664270401,
"step": 11380
},
{
"epoch": 0.6898181818181818,
"grad_norm": 0.04003984108567238,
"learning_rate": 0.0008,
"loss": 2.006711721420288,
"step": 11382
},
{
"epoch": 0.689939393939394,
"grad_norm": 0.041554007679224014,
"learning_rate": 0.0008,
"loss": 2.0144360065460205,
"step": 11384
},
{
"epoch": 0.690060606060606,
"grad_norm": 0.043545156717300415,
"learning_rate": 0.0008,
"loss": 2.005117654800415,
"step": 11386
},
{
"epoch": 0.6901818181818182,
"grad_norm": 0.045737553387880325,
"learning_rate": 0.0008,
"loss": 2.0670552253723145,
"step": 11388
},
{
"epoch": 0.6903030303030303,
"grad_norm": 0.04337381571531296,
"learning_rate": 0.0008,
"loss": 2.0096945762634277,
"step": 11390
},
{
"epoch": 0.6904242424242424,
"grad_norm": 0.048890359699726105,
"learning_rate": 0.0008,
"loss": 2.058213949203491,
"step": 11392
},
{
"epoch": 0.6905454545454546,
"grad_norm": 0.04562138766050339,
"learning_rate": 0.0008,
"loss": 1.968894362449646,
"step": 11394
},
{
"epoch": 0.6906666666666667,
"grad_norm": 0.04224153980612755,
"learning_rate": 0.0008,
"loss": 1.9641021490097046,
"step": 11396
},
{
"epoch": 0.6907878787878788,
"grad_norm": 0.0413205511868,
"learning_rate": 0.0008,
"loss": 2.0177695751190186,
"step": 11398
},
{
"epoch": 0.6909090909090909,
"grad_norm": 0.03905868902802467,
"learning_rate": 0.0008,
"loss": 1.950446367263794,
"step": 11400
},
{
"epoch": 0.691030303030303,
"grad_norm": 0.040962036699056625,
"learning_rate": 0.0008,
"loss": 2.019986867904663,
"step": 11402
},
{
"epoch": 0.6911515151515152,
"grad_norm": 0.041869740933179855,
"learning_rate": 0.0008,
"loss": 2.00740385055542,
"step": 11404
},
{
"epoch": 0.6912727272727273,
"grad_norm": 0.042743559926748276,
"learning_rate": 0.0008,
"loss": 2.039161443710327,
"step": 11406
},
{
"epoch": 0.6913939393939393,
"grad_norm": 0.040737032890319824,
"learning_rate": 0.0008,
"loss": 2.028386354446411,
"step": 11408
},
{
"epoch": 0.6915151515151515,
"grad_norm": 0.04722309857606888,
"learning_rate": 0.0008,
"loss": 1.972355842590332,
"step": 11410
},
{
"epoch": 0.6916363636363636,
"grad_norm": 0.039777543395757675,
"learning_rate": 0.0008,
"loss": 2.0567374229431152,
"step": 11412
},
{
"epoch": 0.6917575757575758,
"grad_norm": 0.041065145283937454,
"learning_rate": 0.0008,
"loss": 1.96239173412323,
"step": 11414
},
{
"epoch": 0.6918787878787879,
"grad_norm": 0.04054342582821846,
"learning_rate": 0.0008,
"loss": 2.0441975593566895,
"step": 11416
},
{
"epoch": 0.692,
"grad_norm": 0.040299609303474426,
"learning_rate": 0.0008,
"loss": 1.9604262113571167,
"step": 11418
},
{
"epoch": 0.6921212121212121,
"grad_norm": 0.04077598452568054,
"learning_rate": 0.0008,
"loss": 2.078935146331787,
"step": 11420
},
{
"epoch": 0.6922424242424242,
"grad_norm": 0.04094502329826355,
"learning_rate": 0.0008,
"loss": 2.0231406688690186,
"step": 11422
},
{
"epoch": 0.6923636363636364,
"grad_norm": 0.04438091069459915,
"learning_rate": 0.0008,
"loss": 1.9821597337722778,
"step": 11424
},
{
"epoch": 0.6924848484848485,
"grad_norm": 0.042240966111421585,
"learning_rate": 0.0008,
"loss": 1.9379018545150757,
"step": 11426
},
{
"epoch": 0.6926060606060606,
"grad_norm": 0.040022920817136765,
"learning_rate": 0.0008,
"loss": 2.0199925899505615,
"step": 11428
},
{
"epoch": 0.6927272727272727,
"grad_norm": 0.0396309457719326,
"learning_rate": 0.0008,
"loss": 2.0281529426574707,
"step": 11430
},
{
"epoch": 0.6928484848484848,
"grad_norm": 0.04298890009522438,
"learning_rate": 0.0008,
"loss": 2.0546228885650635,
"step": 11432
},
{
"epoch": 0.692969696969697,
"grad_norm": 0.046377941966056824,
"learning_rate": 0.0008,
"loss": 1.9972702264785767,
"step": 11434
},
{
"epoch": 0.6930909090909091,
"grad_norm": 0.05198214203119278,
"learning_rate": 0.0008,
"loss": 2.0617735385894775,
"step": 11436
},
{
"epoch": 0.6932121212121212,
"grad_norm": 0.0464705266058445,
"learning_rate": 0.0008,
"loss": 1.9564118385314941,
"step": 11438
},
{
"epoch": 0.6933333333333334,
"grad_norm": 0.03918071836233139,
"learning_rate": 0.0008,
"loss": 2.024742841720581,
"step": 11440
},
{
"epoch": 0.6934545454545454,
"grad_norm": 0.040172554552555084,
"learning_rate": 0.0008,
"loss": 2.0712943077087402,
"step": 11442
},
{
"epoch": 0.6935757575757576,
"grad_norm": 0.040077418088912964,
"learning_rate": 0.0008,
"loss": 2.0637800693511963,
"step": 11444
},
{
"epoch": 0.6936969696969697,
"grad_norm": 0.04030665010213852,
"learning_rate": 0.0008,
"loss": 2.0861101150512695,
"step": 11446
},
{
"epoch": 0.6938181818181818,
"grad_norm": 0.04300270974636078,
"learning_rate": 0.0008,
"loss": 2.0077064037323,
"step": 11448
},
{
"epoch": 0.693939393939394,
"grad_norm": 0.05237225070595741,
"learning_rate": 0.0008,
"loss": 2.055795431137085,
"step": 11450
},
{
"epoch": 0.694060606060606,
"grad_norm": 0.05002260580658913,
"learning_rate": 0.0008,
"loss": 1.9867198467254639,
"step": 11452
},
{
"epoch": 0.6941818181818182,
"grad_norm": 0.04716549813747406,
"learning_rate": 0.0008,
"loss": 1.9866098165512085,
"step": 11454
},
{
"epoch": 0.6943030303030303,
"grad_norm": 0.05266338959336281,
"learning_rate": 0.0008,
"loss": 2.0569968223571777,
"step": 11456
},
{
"epoch": 0.6944242424242424,
"grad_norm": 0.051802754402160645,
"learning_rate": 0.0008,
"loss": 1.99575674533844,
"step": 11458
},
{
"epoch": 0.6945454545454546,
"grad_norm": 0.06791353970766068,
"learning_rate": 0.0008,
"loss": 2.031116485595703,
"step": 11460
},
{
"epoch": 0.6946666666666667,
"grad_norm": 0.0508418045938015,
"learning_rate": 0.0008,
"loss": 1.9998607635498047,
"step": 11462
},
{
"epoch": 0.6947878787878788,
"grad_norm": 0.04190883785486221,
"learning_rate": 0.0008,
"loss": 2.024711847305298,
"step": 11464
},
{
"epoch": 0.6949090909090909,
"grad_norm": 0.0437631830573082,
"learning_rate": 0.0008,
"loss": 2.067957878112793,
"step": 11466
},
{
"epoch": 0.695030303030303,
"grad_norm": 0.04544564709067345,
"learning_rate": 0.0008,
"loss": 2.0349621772766113,
"step": 11468
},
{
"epoch": 0.6951515151515152,
"grad_norm": 0.049205075949430466,
"learning_rate": 0.0008,
"loss": 2.057220697402954,
"step": 11470
},
{
"epoch": 0.6952727272727273,
"grad_norm": 0.051448386162519455,
"learning_rate": 0.0008,
"loss": 1.9830214977264404,
"step": 11472
},
{
"epoch": 0.6953939393939393,
"grad_norm": 0.049788907170295715,
"learning_rate": 0.0008,
"loss": 2.0438413619995117,
"step": 11474
},
{
"epoch": 0.6955151515151515,
"grad_norm": 0.055096570402383804,
"learning_rate": 0.0008,
"loss": 2.042998790740967,
"step": 11476
},
{
"epoch": 0.6956363636363636,
"grad_norm": 0.05942985787987709,
"learning_rate": 0.0008,
"loss": 2.0180890560150146,
"step": 11478
},
{
"epoch": 0.6957575757575758,
"grad_norm": 0.06546644121408463,
"learning_rate": 0.0008,
"loss": 2.0254178047180176,
"step": 11480
},
{
"epoch": 0.6958787878787879,
"grad_norm": 0.05387546494603157,
"learning_rate": 0.0008,
"loss": 1.970753788948059,
"step": 11482
},
{
"epoch": 0.696,
"grad_norm": 0.06026993691921234,
"learning_rate": 0.0008,
"loss": 2.104635715484619,
"step": 11484
},
{
"epoch": 0.6961212121212121,
"grad_norm": 0.06006631627678871,
"learning_rate": 0.0008,
"loss": 1.9572052955627441,
"step": 11486
},
{
"epoch": 0.6962424242424242,
"grad_norm": 0.0732954889535904,
"learning_rate": 0.0008,
"loss": 1.9575635194778442,
"step": 11488
},
{
"epoch": 0.6963636363636364,
"grad_norm": 0.06702104210853577,
"learning_rate": 0.0008,
"loss": 2.030181646347046,
"step": 11490
},
{
"epoch": 0.6964848484848485,
"grad_norm": 0.06729593873023987,
"learning_rate": 0.0008,
"loss": 2.002420425415039,
"step": 11492
},
{
"epoch": 0.6966060606060606,
"grad_norm": 0.06406145542860031,
"learning_rate": 0.0008,
"loss": 2.0216405391693115,
"step": 11494
},
{
"epoch": 0.6967272727272728,
"grad_norm": 0.05460098013281822,
"learning_rate": 0.0008,
"loss": 2.0620245933532715,
"step": 11496
},
{
"epoch": 0.6968484848484848,
"grad_norm": 0.0474989227950573,
"learning_rate": 0.0008,
"loss": 1.9690637588500977,
"step": 11498
},
{
"epoch": 0.696969696969697,
"grad_norm": 0.04214438050985336,
"learning_rate": 0.0008,
"loss": 2.060678005218506,
"step": 11500
},
{
"epoch": 0.6970909090909091,
"grad_norm": 0.04195312783122063,
"learning_rate": 0.0008,
"loss": 2.0220015048980713,
"step": 11502
},
{
"epoch": 0.6972121212121212,
"grad_norm": 0.04053686931729317,
"learning_rate": 0.0008,
"loss": 2.0222489833831787,
"step": 11504
},
{
"epoch": 0.6973333333333334,
"grad_norm": 0.03806217759847641,
"learning_rate": 0.0008,
"loss": 1.9452850818634033,
"step": 11506
},
{
"epoch": 0.6974545454545454,
"grad_norm": 0.0444975271821022,
"learning_rate": 0.0008,
"loss": 2.0256409645080566,
"step": 11508
},
{
"epoch": 0.6975757575757576,
"grad_norm": 0.04766411334276199,
"learning_rate": 0.0008,
"loss": 2.004164457321167,
"step": 11510
},
{
"epoch": 0.6976969696969697,
"grad_norm": 0.046215225011110306,
"learning_rate": 0.0008,
"loss": 2.0290515422821045,
"step": 11512
},
{
"epoch": 0.6978181818181818,
"grad_norm": 0.042238395661115646,
"learning_rate": 0.0008,
"loss": 2.0604805946350098,
"step": 11514
},
{
"epoch": 0.697939393939394,
"grad_norm": 0.04859485104680061,
"learning_rate": 0.0008,
"loss": 1.9890841245651245,
"step": 11516
},
{
"epoch": 0.698060606060606,
"grad_norm": 0.04851385951042175,
"learning_rate": 0.0008,
"loss": 2.012197732925415,
"step": 11518
},
{
"epoch": 0.6981818181818182,
"grad_norm": 0.04202970862388611,
"learning_rate": 0.0008,
"loss": 1.9375687837600708,
"step": 11520
},
{
"epoch": 0.6983030303030303,
"grad_norm": 0.04183860868215561,
"learning_rate": 0.0008,
"loss": 1.9976341724395752,
"step": 11522
},
{
"epoch": 0.6984242424242424,
"grad_norm": 0.04162517189979553,
"learning_rate": 0.0008,
"loss": 1.987486481666565,
"step": 11524
},
{
"epoch": 0.6985454545454546,
"grad_norm": 0.03960360214114189,
"learning_rate": 0.0008,
"loss": 2.0816290378570557,
"step": 11526
},
{
"epoch": 0.6986666666666667,
"grad_norm": 0.04026215523481369,
"learning_rate": 0.0008,
"loss": 2.016914129257202,
"step": 11528
},
{
"epoch": 0.6987878787878787,
"grad_norm": 0.044670891016721725,
"learning_rate": 0.0008,
"loss": 2.042571544647217,
"step": 11530
},
{
"epoch": 0.6989090909090909,
"grad_norm": 0.04287100210785866,
"learning_rate": 0.0008,
"loss": 2.063115119934082,
"step": 11532
},
{
"epoch": 0.699030303030303,
"grad_norm": 0.038978539407253265,
"learning_rate": 0.0008,
"loss": 2.0010745525360107,
"step": 11534
},
{
"epoch": 0.6991515151515152,
"grad_norm": 0.04091154411435127,
"learning_rate": 0.0008,
"loss": 2.000256299972534,
"step": 11536
},
{
"epoch": 0.6992727272727273,
"grad_norm": 0.04536037892103195,
"learning_rate": 0.0008,
"loss": 2.013428211212158,
"step": 11538
},
{
"epoch": 0.6993939393939393,
"grad_norm": 0.04320112615823746,
"learning_rate": 0.0008,
"loss": 2.043651819229126,
"step": 11540
},
{
"epoch": 0.6995151515151515,
"grad_norm": 0.04143396019935608,
"learning_rate": 0.0008,
"loss": 1.956825852394104,
"step": 11542
},
{
"epoch": 0.6996363636363636,
"grad_norm": 0.040108274668455124,
"learning_rate": 0.0008,
"loss": 1.9751650094985962,
"step": 11544
},
{
"epoch": 0.6997575757575758,
"grad_norm": 0.03954119607806206,
"learning_rate": 0.0008,
"loss": 2.0277023315429688,
"step": 11546
},
{
"epoch": 0.6998787878787879,
"grad_norm": 0.03936706855893135,
"learning_rate": 0.0008,
"loss": 2.0641367435455322,
"step": 11548
},
{
"epoch": 0.7,
"grad_norm": 0.03993086516857147,
"learning_rate": 0.0008,
"loss": 1.9722362756729126,
"step": 11550
},
{
"epoch": 0.7001212121212121,
"grad_norm": 0.04314535856246948,
"learning_rate": 0.0008,
"loss": 1.979189157485962,
"step": 11552
},
{
"epoch": 0.7002424242424242,
"grad_norm": 0.043606169521808624,
"learning_rate": 0.0008,
"loss": 2.047373056411743,
"step": 11554
},
{
"epoch": 0.7003636363636364,
"grad_norm": 0.052421968430280685,
"learning_rate": 0.0008,
"loss": 2.0453274250030518,
"step": 11556
},
{
"epoch": 0.7004848484848485,
"grad_norm": 0.0396440289914608,
"learning_rate": 0.0008,
"loss": 2.004427433013916,
"step": 11558
},
{
"epoch": 0.7006060606060606,
"grad_norm": 0.04110633581876755,
"learning_rate": 0.0008,
"loss": 2.0502665042877197,
"step": 11560
},
{
"epoch": 0.7007272727272728,
"grad_norm": 0.04236676171422005,
"learning_rate": 0.0008,
"loss": 2.060058355331421,
"step": 11562
},
{
"epoch": 0.7008484848484848,
"grad_norm": 0.04324449226260185,
"learning_rate": 0.0008,
"loss": 2.0113911628723145,
"step": 11564
},
{
"epoch": 0.700969696969697,
"grad_norm": 0.040605705231428146,
"learning_rate": 0.0008,
"loss": 2.0628135204315186,
"step": 11566
},
{
"epoch": 0.7010909090909091,
"grad_norm": 0.0395781546831131,
"learning_rate": 0.0008,
"loss": 2.011240243911743,
"step": 11568
},
{
"epoch": 0.7012121212121212,
"grad_norm": 0.03934726119041443,
"learning_rate": 0.0008,
"loss": 2.0190343856811523,
"step": 11570
},
{
"epoch": 0.7013333333333334,
"grad_norm": 0.042977623641490936,
"learning_rate": 0.0008,
"loss": 2.0520567893981934,
"step": 11572
},
{
"epoch": 0.7014545454545454,
"grad_norm": 0.041656337678432465,
"learning_rate": 0.0008,
"loss": 2.0116593837738037,
"step": 11574
},
{
"epoch": 0.7015757575757576,
"grad_norm": 0.03952982276678085,
"learning_rate": 0.0008,
"loss": 1.983279824256897,
"step": 11576
},
{
"epoch": 0.7016969696969697,
"grad_norm": 0.042592767626047134,
"learning_rate": 0.0008,
"loss": 2.0209434032440186,
"step": 11578
},
{
"epoch": 0.7018181818181818,
"grad_norm": 0.03878522291779518,
"learning_rate": 0.0008,
"loss": 2.0647809505462646,
"step": 11580
},
{
"epoch": 0.701939393939394,
"grad_norm": 0.04598168283700943,
"learning_rate": 0.0008,
"loss": 2.036573648452759,
"step": 11582
},
{
"epoch": 0.702060606060606,
"grad_norm": 0.03954634070396423,
"learning_rate": 0.0008,
"loss": 2.0777134895324707,
"step": 11584
},
{
"epoch": 0.7021818181818181,
"grad_norm": 0.04160306230187416,
"learning_rate": 0.0008,
"loss": 2.0252349376678467,
"step": 11586
},
{
"epoch": 0.7023030303030303,
"grad_norm": 0.04306407645344734,
"learning_rate": 0.0008,
"loss": 2.0309078693389893,
"step": 11588
},
{
"epoch": 0.7024242424242424,
"grad_norm": 0.049818120896816254,
"learning_rate": 0.0008,
"loss": 2.0266735553741455,
"step": 11590
},
{
"epoch": 0.7025454545454546,
"grad_norm": 0.03791007399559021,
"learning_rate": 0.0008,
"loss": 2.013657808303833,
"step": 11592
},
{
"epoch": 0.7026666666666667,
"grad_norm": 0.04051339998841286,
"learning_rate": 0.0008,
"loss": 2.0570266246795654,
"step": 11594
},
{
"epoch": 0.7027878787878787,
"grad_norm": 0.040157251060009,
"learning_rate": 0.0008,
"loss": 2.0171008110046387,
"step": 11596
},
{
"epoch": 0.7029090909090909,
"grad_norm": 0.04263655096292496,
"learning_rate": 0.0008,
"loss": 1.9039450883865356,
"step": 11598
},
{
"epoch": 0.703030303030303,
"grad_norm": 0.04033757746219635,
"learning_rate": 0.0008,
"loss": 2.0026330947875977,
"step": 11600
},
{
"epoch": 0.7031515151515152,
"grad_norm": 0.038081929087638855,
"learning_rate": 0.0008,
"loss": 1.9242817163467407,
"step": 11602
},
{
"epoch": 0.7032727272727273,
"grad_norm": 0.04202253371477127,
"learning_rate": 0.0008,
"loss": 2.0312325954437256,
"step": 11604
},
{
"epoch": 0.7033939393939393,
"grad_norm": 0.05188232660293579,
"learning_rate": 0.0008,
"loss": 1.9624131917953491,
"step": 11606
},
{
"epoch": 0.7035151515151515,
"grad_norm": 0.039285797625780106,
"learning_rate": 0.0008,
"loss": 2.0106546878814697,
"step": 11608
},
{
"epoch": 0.7036363636363636,
"grad_norm": 0.04898590222001076,
"learning_rate": 0.0008,
"loss": 2.0632660388946533,
"step": 11610
},
{
"epoch": 0.7037575757575758,
"grad_norm": 0.0560162216424942,
"learning_rate": 0.0008,
"loss": 1.986149549484253,
"step": 11612
},
{
"epoch": 0.7038787878787879,
"grad_norm": 0.058755941689014435,
"learning_rate": 0.0008,
"loss": 2.0571203231811523,
"step": 11614
},
{
"epoch": 0.704,
"grad_norm": 0.062092628329992294,
"learning_rate": 0.0008,
"loss": 2.021599531173706,
"step": 11616
},
{
"epoch": 0.7041212121212121,
"grad_norm": 0.05463102459907532,
"learning_rate": 0.0008,
"loss": 2.018702507019043,
"step": 11618
},
{
"epoch": 0.7042424242424242,
"grad_norm": 0.057836927473545074,
"learning_rate": 0.0008,
"loss": 2.047999858856201,
"step": 11620
},
{
"epoch": 0.7043636363636364,
"grad_norm": 0.04316433519124985,
"learning_rate": 0.0008,
"loss": 2.04414963722229,
"step": 11622
},
{
"epoch": 0.7044848484848485,
"grad_norm": 0.04047482833266258,
"learning_rate": 0.0008,
"loss": 1.9999268054962158,
"step": 11624
},
{
"epoch": 0.7046060606060606,
"grad_norm": 0.04221399500966072,
"learning_rate": 0.0008,
"loss": 2.0887045860290527,
"step": 11626
},
{
"epoch": 0.7047272727272728,
"grad_norm": 0.04545561596751213,
"learning_rate": 0.0008,
"loss": 2.0435433387756348,
"step": 11628
},
{
"epoch": 0.7048484848484848,
"grad_norm": 0.051864538341760635,
"learning_rate": 0.0008,
"loss": 2.0219290256500244,
"step": 11630
},
{
"epoch": 0.704969696969697,
"grad_norm": 0.04249655082821846,
"learning_rate": 0.0008,
"loss": 2.026632070541382,
"step": 11632
},
{
"epoch": 0.7050909090909091,
"grad_norm": 0.03859639912843704,
"learning_rate": 0.0008,
"loss": 2.004472255706787,
"step": 11634
},
{
"epoch": 0.7052121212121212,
"grad_norm": 0.04040413349866867,
"learning_rate": 0.0008,
"loss": 1.998222827911377,
"step": 11636
},
{
"epoch": 0.7053333333333334,
"grad_norm": 0.041748445481061935,
"learning_rate": 0.0008,
"loss": 1.9923875331878662,
"step": 11638
},
{
"epoch": 0.7054545454545454,
"grad_norm": 0.04146488010883331,
"learning_rate": 0.0008,
"loss": 2.051471471786499,
"step": 11640
},
{
"epoch": 0.7055757575757575,
"grad_norm": 0.04135383293032646,
"learning_rate": 0.0008,
"loss": 2.0140421390533447,
"step": 11642
},
{
"epoch": 0.7056969696969697,
"grad_norm": 0.04213410243391991,
"learning_rate": 0.0008,
"loss": 2.0171563625335693,
"step": 11644
},
{
"epoch": 0.7058181818181818,
"grad_norm": 0.041156210005283356,
"learning_rate": 0.0008,
"loss": 1.9871984720230103,
"step": 11646
},
{
"epoch": 0.705939393939394,
"grad_norm": 0.04247326776385307,
"learning_rate": 0.0008,
"loss": 2.0058929920196533,
"step": 11648
},
{
"epoch": 0.706060606060606,
"grad_norm": 0.0419560931622982,
"learning_rate": 0.0008,
"loss": 2.0025951862335205,
"step": 11650
},
{
"epoch": 0.7061818181818181,
"grad_norm": 0.04083414375782013,
"learning_rate": 0.0008,
"loss": 2.0641071796417236,
"step": 11652
},
{
"epoch": 0.7063030303030303,
"grad_norm": 0.03981854021549225,
"learning_rate": 0.0008,
"loss": 2.0461554527282715,
"step": 11654
},
{
"epoch": 0.7064242424242424,
"grad_norm": 0.042074717581272125,
"learning_rate": 0.0008,
"loss": 2.012944221496582,
"step": 11656
},
{
"epoch": 0.7065454545454546,
"grad_norm": 0.04420943185687065,
"learning_rate": 0.0008,
"loss": 2.030197858810425,
"step": 11658
},
{
"epoch": 0.7066666666666667,
"grad_norm": 0.04400624334812164,
"learning_rate": 0.0008,
"loss": 2.0675857067108154,
"step": 11660
},
{
"epoch": 0.7067878787878787,
"grad_norm": 0.04093965142965317,
"learning_rate": 0.0008,
"loss": 2.0414295196533203,
"step": 11662
},
{
"epoch": 0.7069090909090909,
"grad_norm": 0.04141984134912491,
"learning_rate": 0.0008,
"loss": 1.989594578742981,
"step": 11664
},
{
"epoch": 0.707030303030303,
"grad_norm": 0.038820087909698486,
"learning_rate": 0.0008,
"loss": 1.9478448629379272,
"step": 11666
},
{
"epoch": 0.7071515151515152,
"grad_norm": 0.04175204411149025,
"learning_rate": 0.0008,
"loss": 2.064452648162842,
"step": 11668
},
{
"epoch": 0.7072727272727273,
"grad_norm": 0.039604030549526215,
"learning_rate": 0.0008,
"loss": 1.9993891716003418,
"step": 11670
},
{
"epoch": 0.7073939393939394,
"grad_norm": 0.04145537689328194,
"learning_rate": 0.0008,
"loss": 2.027179718017578,
"step": 11672
},
{
"epoch": 0.7075151515151515,
"grad_norm": 0.038695357739925385,
"learning_rate": 0.0008,
"loss": 1.9951882362365723,
"step": 11674
},
{
"epoch": 0.7076363636363636,
"grad_norm": 0.0398096963763237,
"learning_rate": 0.0008,
"loss": 1.9410920143127441,
"step": 11676
},
{
"epoch": 0.7077575757575758,
"grad_norm": 0.04159040004014969,
"learning_rate": 0.0008,
"loss": 2.036567211151123,
"step": 11678
},
{
"epoch": 0.7078787878787879,
"grad_norm": 0.05784895271062851,
"learning_rate": 0.0008,
"loss": 2.024534225463867,
"step": 11680
},
{
"epoch": 0.708,
"grad_norm": 0.05531792715191841,
"learning_rate": 0.0008,
"loss": 2.0456390380859375,
"step": 11682
},
{
"epoch": 0.7081212121212122,
"grad_norm": 0.04146944731473923,
"learning_rate": 0.0008,
"loss": 1.9570547342300415,
"step": 11684
},
{
"epoch": 0.7082424242424242,
"grad_norm": 0.04569248855113983,
"learning_rate": 0.0008,
"loss": 1.9785244464874268,
"step": 11686
},
{
"epoch": 0.7083636363636364,
"grad_norm": 0.04044776409864426,
"learning_rate": 0.0008,
"loss": 2.051442861557007,
"step": 11688
},
{
"epoch": 0.7084848484848485,
"grad_norm": 0.038437530398368835,
"learning_rate": 0.0008,
"loss": 2.0191853046417236,
"step": 11690
},
{
"epoch": 0.7086060606060606,
"grad_norm": 0.042696788907051086,
"learning_rate": 0.0008,
"loss": 1.9945294857025146,
"step": 11692
},
{
"epoch": 0.7087272727272728,
"grad_norm": 0.042495839297771454,
"learning_rate": 0.0008,
"loss": 1.9865034818649292,
"step": 11694
},
{
"epoch": 0.7088484848484848,
"grad_norm": 0.03970561549067497,
"learning_rate": 0.0008,
"loss": 2.0560030937194824,
"step": 11696
},
{
"epoch": 0.7089696969696969,
"grad_norm": 0.040114790201187134,
"learning_rate": 0.0008,
"loss": 2.025988817214966,
"step": 11698
},
{
"epoch": 0.7090909090909091,
"grad_norm": 0.04027203470468521,
"learning_rate": 0.0008,
"loss": 2.0374202728271484,
"step": 11700
},
{
"epoch": 0.7092121212121212,
"grad_norm": 0.039050158113241196,
"learning_rate": 0.0008,
"loss": 1.942068099975586,
"step": 11702
},
{
"epoch": 0.7093333333333334,
"grad_norm": 0.04008448123931885,
"learning_rate": 0.0008,
"loss": 1.9597030878067017,
"step": 11704
},
{
"epoch": 0.7094545454545454,
"grad_norm": 0.03955308720469475,
"learning_rate": 0.0008,
"loss": 2.0228521823883057,
"step": 11706
},
{
"epoch": 0.7095757575757575,
"grad_norm": 0.04657505080103874,
"learning_rate": 0.0008,
"loss": 1.9945229291915894,
"step": 11708
},
{
"epoch": 0.7096969696969697,
"grad_norm": 0.03983425721526146,
"learning_rate": 0.0008,
"loss": 2.0733797550201416,
"step": 11710
},
{
"epoch": 0.7098181818181818,
"grad_norm": 0.03918801620602608,
"learning_rate": 0.0008,
"loss": 2.0247623920440674,
"step": 11712
},
{
"epoch": 0.709939393939394,
"grad_norm": 0.051804233342409134,
"learning_rate": 0.0008,
"loss": 1.986845850944519,
"step": 11714
},
{
"epoch": 0.7100606060606061,
"grad_norm": 0.05602971464395523,
"learning_rate": 0.0008,
"loss": 2.0143682956695557,
"step": 11716
},
{
"epoch": 0.7101818181818181,
"grad_norm": 0.05103546008467674,
"learning_rate": 0.0008,
"loss": 2.0768063068389893,
"step": 11718
},
{
"epoch": 0.7103030303030303,
"grad_norm": 0.04390063136816025,
"learning_rate": 0.0008,
"loss": 1.9270809888839722,
"step": 11720
},
{
"epoch": 0.7104242424242424,
"grad_norm": 0.04039713740348816,
"learning_rate": 0.0008,
"loss": 2.014831781387329,
"step": 11722
},
{
"epoch": 0.7105454545454546,
"grad_norm": 0.04024764150381088,
"learning_rate": 0.0008,
"loss": 2.008779764175415,
"step": 11724
},
{
"epoch": 0.7106666666666667,
"grad_norm": 0.03950544074177742,
"learning_rate": 0.0008,
"loss": 2.063570499420166,
"step": 11726
},
{
"epoch": 0.7107878787878787,
"grad_norm": 0.040657974779605865,
"learning_rate": 0.0008,
"loss": 2.0255305767059326,
"step": 11728
},
{
"epoch": 0.7109090909090909,
"grad_norm": 0.04000889137387276,
"learning_rate": 0.0008,
"loss": 2.0158398151397705,
"step": 11730
},
{
"epoch": 0.711030303030303,
"grad_norm": 0.05153869092464447,
"learning_rate": 0.0008,
"loss": 1.9817789793014526,
"step": 11732
},
{
"epoch": 0.7111515151515152,
"grad_norm": 0.04415399208664894,
"learning_rate": 0.0008,
"loss": 2.0230631828308105,
"step": 11734
},
{
"epoch": 0.7112727272727273,
"grad_norm": 0.04020826891064644,
"learning_rate": 0.0008,
"loss": 2.0424227714538574,
"step": 11736
},
{
"epoch": 0.7113939393939394,
"grad_norm": 0.039231378585100174,
"learning_rate": 0.0008,
"loss": 2.0067949295043945,
"step": 11738
},
{
"epoch": 0.7115151515151515,
"grad_norm": 0.04120717570185661,
"learning_rate": 0.0008,
"loss": 1.9848850965499878,
"step": 11740
},
{
"epoch": 0.7116363636363636,
"grad_norm": 0.03874882310628891,
"learning_rate": 0.0008,
"loss": 1.9692614078521729,
"step": 11742
},
{
"epoch": 0.7117575757575758,
"grad_norm": 0.04026180878281593,
"learning_rate": 0.0008,
"loss": 1.9789340496063232,
"step": 11744
},
{
"epoch": 0.7118787878787879,
"grad_norm": 0.04016873612999916,
"learning_rate": 0.0008,
"loss": 2.0006392002105713,
"step": 11746
},
{
"epoch": 0.712,
"grad_norm": 0.042569778859615326,
"learning_rate": 0.0008,
"loss": 2.0121099948883057,
"step": 11748
},
{
"epoch": 0.7121212121212122,
"grad_norm": 0.048033565282821655,
"learning_rate": 0.0008,
"loss": 2.0522091388702393,
"step": 11750
},
{
"epoch": 0.7122424242424242,
"grad_norm": 0.04117279872298241,
"learning_rate": 0.0008,
"loss": 2.0106003284454346,
"step": 11752
},
{
"epoch": 0.7123636363636363,
"grad_norm": 0.047025084495544434,
"learning_rate": 0.0008,
"loss": 1.9875423908233643,
"step": 11754
},
{
"epoch": 0.7124848484848485,
"grad_norm": 0.046967096626758575,
"learning_rate": 0.0008,
"loss": 1.9749512672424316,
"step": 11756
},
{
"epoch": 0.7126060606060606,
"grad_norm": 0.05590936914086342,
"learning_rate": 0.0008,
"loss": 2.0102522373199463,
"step": 11758
},
{
"epoch": 0.7127272727272728,
"grad_norm": 0.06367897242307663,
"learning_rate": 0.0008,
"loss": 1.999449372291565,
"step": 11760
},
{
"epoch": 0.7128484848484848,
"grad_norm": 0.04587361589074135,
"learning_rate": 0.0008,
"loss": 1.9415181875228882,
"step": 11762
},
{
"epoch": 0.7129696969696969,
"grad_norm": 0.04299506917595863,
"learning_rate": 0.0008,
"loss": 2.0316028594970703,
"step": 11764
},
{
"epoch": 0.7130909090909091,
"grad_norm": 0.039693985134363174,
"learning_rate": 0.0008,
"loss": 2.0502567291259766,
"step": 11766
},
{
"epoch": 0.7132121212121212,
"grad_norm": 0.0396740697324276,
"learning_rate": 0.0008,
"loss": 1.9957141876220703,
"step": 11768
},
{
"epoch": 0.7133333333333334,
"grad_norm": 0.03926324471831322,
"learning_rate": 0.0008,
"loss": 2.0168285369873047,
"step": 11770
},
{
"epoch": 0.7134545454545455,
"grad_norm": 0.047267645597457886,
"learning_rate": 0.0008,
"loss": 2.0041093826293945,
"step": 11772
},
{
"epoch": 0.7135757575757575,
"grad_norm": 0.049777157604694366,
"learning_rate": 0.0008,
"loss": 2.044621229171753,
"step": 11774
},
{
"epoch": 0.7136969696969697,
"grad_norm": 0.06057944893836975,
"learning_rate": 0.0008,
"loss": 2.0922482013702393,
"step": 11776
},
{
"epoch": 0.7138181818181818,
"grad_norm": 0.053692542016506195,
"learning_rate": 0.0008,
"loss": 1.9963828325271606,
"step": 11778
},
{
"epoch": 0.713939393939394,
"grad_norm": 0.04637736827135086,
"learning_rate": 0.0008,
"loss": 1.974694848060608,
"step": 11780
},
{
"epoch": 0.7140606060606061,
"grad_norm": 0.05613093450665474,
"learning_rate": 0.0008,
"loss": 2.159933090209961,
"step": 11782
},
{
"epoch": 0.7141818181818181,
"grad_norm": 0.04919698089361191,
"learning_rate": 0.0008,
"loss": 1.9447885751724243,
"step": 11784
},
{
"epoch": 0.7143030303030303,
"grad_norm": 0.04443551227450371,
"learning_rate": 0.0008,
"loss": 2.030714273452759,
"step": 11786
},
{
"epoch": 0.7144242424242424,
"grad_norm": 0.04681827500462532,
"learning_rate": 0.0008,
"loss": 1.9962621927261353,
"step": 11788
},
{
"epoch": 0.7145454545454546,
"grad_norm": 0.04206976294517517,
"learning_rate": 0.0008,
"loss": 2.047022581100464,
"step": 11790
},
{
"epoch": 0.7146666666666667,
"grad_norm": 0.04969754442572594,
"learning_rate": 0.0008,
"loss": 2.0354464054107666,
"step": 11792
},
{
"epoch": 0.7147878787878787,
"grad_norm": 0.054267555475234985,
"learning_rate": 0.0008,
"loss": 2.031723976135254,
"step": 11794
},
{
"epoch": 0.7149090909090909,
"grad_norm": 0.06087062507867813,
"learning_rate": 0.0008,
"loss": 2.019244432449341,
"step": 11796
},
{
"epoch": 0.715030303030303,
"grad_norm": 0.050790559500455856,
"learning_rate": 0.0008,
"loss": 2.0429282188415527,
"step": 11798
},
{
"epoch": 0.7151515151515152,
"grad_norm": 0.047034889459609985,
"learning_rate": 0.0008,
"loss": 1.9966493844985962,
"step": 11800
},
{
"epoch": 0.7152727272727273,
"grad_norm": 0.04773225635290146,
"learning_rate": 0.0008,
"loss": 1.9979679584503174,
"step": 11802
},
{
"epoch": 0.7153939393939394,
"grad_norm": 0.043585944920778275,
"learning_rate": 0.0008,
"loss": 2.047583818435669,
"step": 11804
},
{
"epoch": 0.7155151515151515,
"grad_norm": 0.04305750131607056,
"learning_rate": 0.0008,
"loss": 1.9991358518600464,
"step": 11806
},
{
"epoch": 0.7156363636363636,
"grad_norm": 0.03903680294752121,
"learning_rate": 0.0008,
"loss": 1.9835494756698608,
"step": 11808
},
{
"epoch": 0.7157575757575757,
"grad_norm": 0.03937274217605591,
"learning_rate": 0.0008,
"loss": 1.985821008682251,
"step": 11810
},
{
"epoch": 0.7158787878787879,
"grad_norm": 0.04267155006527901,
"learning_rate": 0.0008,
"loss": 1.903796911239624,
"step": 11812
},
{
"epoch": 0.716,
"grad_norm": 0.04378246143460274,
"learning_rate": 0.0008,
"loss": 1.9906953573226929,
"step": 11814
},
{
"epoch": 0.7161212121212122,
"grad_norm": 0.05035695806145668,
"learning_rate": 0.0008,
"loss": 2.024592161178589,
"step": 11816
},
{
"epoch": 0.7162424242424242,
"grad_norm": 0.04668338969349861,
"learning_rate": 0.0008,
"loss": 1.9945372343063354,
"step": 11818
},
{
"epoch": 0.7163636363636363,
"grad_norm": 0.041450150310993195,
"learning_rate": 0.0008,
"loss": 1.9900164604187012,
"step": 11820
},
{
"epoch": 0.7164848484848485,
"grad_norm": 0.03888806328177452,
"learning_rate": 0.0008,
"loss": 1.9772547483444214,
"step": 11822
},
{
"epoch": 0.7166060606060606,
"grad_norm": 0.04108624532818794,
"learning_rate": 0.0008,
"loss": 2.038947343826294,
"step": 11824
},
{
"epoch": 0.7167272727272728,
"grad_norm": 0.03909878805279732,
"learning_rate": 0.0008,
"loss": 2.0113322734832764,
"step": 11826
},
{
"epoch": 0.7168484848484848,
"grad_norm": 0.03995625302195549,
"learning_rate": 0.0008,
"loss": 2.0384395122528076,
"step": 11828
},
{
"epoch": 0.7169696969696969,
"grad_norm": 0.039622705429792404,
"learning_rate": 0.0008,
"loss": 2.0556821823120117,
"step": 11830
},
{
"epoch": 0.7170909090909091,
"grad_norm": 0.039065584540367126,
"learning_rate": 0.0008,
"loss": 2.032176971435547,
"step": 11832
},
{
"epoch": 0.7172121212121212,
"grad_norm": 0.04497751221060753,
"learning_rate": 0.0008,
"loss": 2.0210015773773193,
"step": 11834
},
{
"epoch": 0.7173333333333334,
"grad_norm": 0.0481233112514019,
"learning_rate": 0.0008,
"loss": 2.025909900665283,
"step": 11836
},
{
"epoch": 0.7174545454545455,
"grad_norm": 0.05319056659936905,
"learning_rate": 0.0008,
"loss": 2.0077381134033203,
"step": 11838
},
{
"epoch": 0.7175757575757575,
"grad_norm": 0.04401149973273277,
"learning_rate": 0.0008,
"loss": 1.9971418380737305,
"step": 11840
},
{
"epoch": 0.7176969696969697,
"grad_norm": 0.039941418915987015,
"learning_rate": 0.0008,
"loss": 1.9686123132705688,
"step": 11842
},
{
"epoch": 0.7178181818181818,
"grad_norm": 0.04380321875214577,
"learning_rate": 0.0008,
"loss": 1.9706971645355225,
"step": 11844
},
{
"epoch": 0.717939393939394,
"grad_norm": 0.03985712304711342,
"learning_rate": 0.0008,
"loss": 2.0321593284606934,
"step": 11846
},
{
"epoch": 0.7180606060606061,
"grad_norm": 0.03858083486557007,
"learning_rate": 0.0008,
"loss": 1.9013307094573975,
"step": 11848
},
{
"epoch": 0.7181818181818181,
"grad_norm": 0.03808318451046944,
"learning_rate": 0.0008,
"loss": 1.970610499382019,
"step": 11850
},
{
"epoch": 0.7183030303030303,
"grad_norm": 0.04122079908847809,
"learning_rate": 0.0008,
"loss": 2.006354331970215,
"step": 11852
},
{
"epoch": 0.7184242424242424,
"grad_norm": 0.03876466676592827,
"learning_rate": 0.0008,
"loss": 2.0076985359191895,
"step": 11854
},
{
"epoch": 0.7185454545454546,
"grad_norm": 0.040456246584653854,
"learning_rate": 0.0008,
"loss": 1.9918912649154663,
"step": 11856
},
{
"epoch": 0.7186666666666667,
"grad_norm": 0.04619381204247475,
"learning_rate": 0.0008,
"loss": 2.0553383827209473,
"step": 11858
},
{
"epoch": 0.7187878787878788,
"grad_norm": 0.0685865506529808,
"learning_rate": 0.0008,
"loss": 2.001451253890991,
"step": 11860
},
{
"epoch": 0.7189090909090909,
"grad_norm": 0.07290156185626984,
"learning_rate": 0.0008,
"loss": 1.9888399839401245,
"step": 11862
},
{
"epoch": 0.719030303030303,
"grad_norm": 0.05712306872010231,
"learning_rate": 0.0008,
"loss": 1.980115532875061,
"step": 11864
},
{
"epoch": 0.7191515151515151,
"grad_norm": 0.06645742803812027,
"learning_rate": 0.0008,
"loss": 1.955422282218933,
"step": 11866
},
{
"epoch": 0.7192727272727273,
"grad_norm": 0.08258972316980362,
"learning_rate": 0.0008,
"loss": 2.0127480030059814,
"step": 11868
},
{
"epoch": 0.7193939393939394,
"grad_norm": 0.06754591315984726,
"learning_rate": 0.0008,
"loss": 1.969568133354187,
"step": 11870
},
{
"epoch": 0.7195151515151516,
"grad_norm": 0.04590057209134102,
"learning_rate": 0.0008,
"loss": 1.9647531509399414,
"step": 11872
},
{
"epoch": 0.7196363636363636,
"grad_norm": 0.0396968275308609,
"learning_rate": 0.0008,
"loss": 2.039576768875122,
"step": 11874
},
{
"epoch": 0.7197575757575757,
"grad_norm": 0.04823590815067291,
"learning_rate": 0.0008,
"loss": 2.042100191116333,
"step": 11876
},
{
"epoch": 0.7198787878787879,
"grad_norm": 0.04580932855606079,
"learning_rate": 0.0008,
"loss": 2.0629007816314697,
"step": 11878
},
{
"epoch": 0.72,
"grad_norm": 0.05541167035698891,
"learning_rate": 0.0008,
"loss": 1.9770904779434204,
"step": 11880
},
{
"epoch": 0.7201212121212122,
"grad_norm": 0.05730587989091873,
"learning_rate": 0.0008,
"loss": 2.0098915100097656,
"step": 11882
},
{
"epoch": 0.7202424242424242,
"grad_norm": 0.059354279190301895,
"learning_rate": 0.0008,
"loss": 2.007465362548828,
"step": 11884
},
{
"epoch": 0.7203636363636363,
"grad_norm": 0.040373545140028,
"learning_rate": 0.0008,
"loss": 2.0376038551330566,
"step": 11886
},
{
"epoch": 0.7204848484848485,
"grad_norm": 0.04074762016534805,
"learning_rate": 0.0008,
"loss": 1.9510397911071777,
"step": 11888
},
{
"epoch": 0.7206060606060606,
"grad_norm": 0.04245731979608536,
"learning_rate": 0.0008,
"loss": 2.0499978065490723,
"step": 11890
},
{
"epoch": 0.7207272727272728,
"grad_norm": 0.03991222754120827,
"learning_rate": 0.0008,
"loss": 2.028775453567505,
"step": 11892
},
{
"epoch": 0.7208484848484848,
"grad_norm": 0.04262569546699524,
"learning_rate": 0.0008,
"loss": 2.0024616718292236,
"step": 11894
},
{
"epoch": 0.7209696969696969,
"grad_norm": 0.04175490513443947,
"learning_rate": 0.0008,
"loss": 2.026740312576294,
"step": 11896
},
{
"epoch": 0.7210909090909091,
"grad_norm": 0.0458141565322876,
"learning_rate": 0.0008,
"loss": 1.9899674654006958,
"step": 11898
},
{
"epoch": 0.7212121212121212,
"grad_norm": 0.040830034762620926,
"learning_rate": 0.0008,
"loss": 1.927998423576355,
"step": 11900
},
{
"epoch": 0.7213333333333334,
"grad_norm": 0.042841874063014984,
"learning_rate": 0.0008,
"loss": 2.0259385108947754,
"step": 11902
},
{
"epoch": 0.7214545454545455,
"grad_norm": 0.040621016174554825,
"learning_rate": 0.0008,
"loss": 1.9664815664291382,
"step": 11904
},
{
"epoch": 0.7215757575757575,
"grad_norm": 0.04153507202863693,
"learning_rate": 0.0008,
"loss": 1.9392836093902588,
"step": 11906
},
{
"epoch": 0.7216969696969697,
"grad_norm": 0.05582691356539726,
"learning_rate": 0.0008,
"loss": 2.004307985305786,
"step": 11908
},
{
"epoch": 0.7218181818181818,
"grad_norm": 0.054351069033145905,
"learning_rate": 0.0008,
"loss": 2.065708875656128,
"step": 11910
},
{
"epoch": 0.721939393939394,
"grad_norm": 0.043170493096113205,
"learning_rate": 0.0008,
"loss": 1.9482616186141968,
"step": 11912
},
{
"epoch": 0.7220606060606061,
"grad_norm": 0.04211672022938728,
"learning_rate": 0.0008,
"loss": 1.968494176864624,
"step": 11914
},
{
"epoch": 0.7221818181818181,
"grad_norm": 0.04559300094842911,
"learning_rate": 0.0008,
"loss": 1.9890764951705933,
"step": 11916
},
{
"epoch": 0.7223030303030303,
"grad_norm": 0.04509004205465317,
"learning_rate": 0.0008,
"loss": 2.0479540824890137,
"step": 11918
},
{
"epoch": 0.7224242424242424,
"grad_norm": 0.04323495924472809,
"learning_rate": 0.0008,
"loss": 2.0431249141693115,
"step": 11920
},
{
"epoch": 0.7225454545454545,
"grad_norm": 0.03895367309451103,
"learning_rate": 0.0008,
"loss": 1.9634543657302856,
"step": 11922
},
{
"epoch": 0.7226666666666667,
"grad_norm": 0.043289799243211746,
"learning_rate": 0.0008,
"loss": 1.9945626258850098,
"step": 11924
},
{
"epoch": 0.7227878787878788,
"grad_norm": 0.041133813560009,
"learning_rate": 0.0008,
"loss": 2.019050121307373,
"step": 11926
},
{
"epoch": 0.722909090909091,
"grad_norm": 0.04063062742352486,
"learning_rate": 0.0008,
"loss": 1.9912891387939453,
"step": 11928
},
{
"epoch": 0.723030303030303,
"grad_norm": 0.04038848727941513,
"learning_rate": 0.0008,
"loss": 1.993392825126648,
"step": 11930
},
{
"epoch": 0.7231515151515151,
"grad_norm": 0.04633088409900665,
"learning_rate": 0.0008,
"loss": 2.034642457962036,
"step": 11932
},
{
"epoch": 0.7232727272727273,
"grad_norm": 0.055921316146850586,
"learning_rate": 0.0008,
"loss": 2.018842935562134,
"step": 11934
},
{
"epoch": 0.7233939393939394,
"grad_norm": 0.0415317565202713,
"learning_rate": 0.0008,
"loss": 1.9120007753372192,
"step": 11936
},
{
"epoch": 0.7235151515151516,
"grad_norm": 0.04122428223490715,
"learning_rate": 0.0008,
"loss": 1.9723854064941406,
"step": 11938
},
{
"epoch": 0.7236363636363636,
"grad_norm": 0.041066113859415054,
"learning_rate": 0.0008,
"loss": 2.00606632232666,
"step": 11940
},
{
"epoch": 0.7237575757575757,
"grad_norm": 0.040599070489406586,
"learning_rate": 0.0008,
"loss": 1.9601995944976807,
"step": 11942
},
{
"epoch": 0.7238787878787879,
"grad_norm": 0.039607901126146317,
"learning_rate": 0.0008,
"loss": 1.9590165615081787,
"step": 11944
},
{
"epoch": 0.724,
"grad_norm": 0.04305874556303024,
"learning_rate": 0.0008,
"loss": 2.0160133838653564,
"step": 11946
},
{
"epoch": 0.7241212121212122,
"grad_norm": 0.04319291189312935,
"learning_rate": 0.0008,
"loss": 1.9599109888076782,
"step": 11948
},
{
"epoch": 0.7242424242424242,
"grad_norm": 0.0438007190823555,
"learning_rate": 0.0008,
"loss": 1.9512510299682617,
"step": 11950
},
{
"epoch": 0.7243636363636363,
"grad_norm": 0.04174793139100075,
"learning_rate": 0.0008,
"loss": 1.9362409114837646,
"step": 11952
},
{
"epoch": 0.7244848484848485,
"grad_norm": 0.040995657444000244,
"learning_rate": 0.0008,
"loss": 2.0290184020996094,
"step": 11954
},
{
"epoch": 0.7246060606060606,
"grad_norm": 0.040257129818201065,
"learning_rate": 0.0008,
"loss": 1.9885430335998535,
"step": 11956
},
{
"epoch": 0.7247272727272728,
"grad_norm": 0.040332335978746414,
"learning_rate": 0.0008,
"loss": 1.9471869468688965,
"step": 11958
},
{
"epoch": 0.7248484848484849,
"grad_norm": 0.05850379541516304,
"learning_rate": 0.0008,
"loss": 1.99061918258667,
"step": 11960
},
{
"epoch": 0.7249696969696969,
"grad_norm": 0.0577886700630188,
"learning_rate": 0.0008,
"loss": 2.005554676055908,
"step": 11962
},
{
"epoch": 0.7250909090909091,
"grad_norm": 0.03931594640016556,
"learning_rate": 0.0008,
"loss": 1.9609719514846802,
"step": 11964
},
{
"epoch": 0.7252121212121212,
"grad_norm": 0.040886323899030685,
"learning_rate": 0.0008,
"loss": 1.9769319295883179,
"step": 11966
},
{
"epoch": 0.7253333333333334,
"grad_norm": 0.03984948992729187,
"learning_rate": 0.0008,
"loss": 1.9672023057937622,
"step": 11968
},
{
"epoch": 0.7254545454545455,
"grad_norm": 0.04086470231413841,
"learning_rate": 0.0008,
"loss": 1.9666377305984497,
"step": 11970
},
{
"epoch": 0.7255757575757575,
"grad_norm": 0.03946239873766899,
"learning_rate": 0.0008,
"loss": 1.947134256362915,
"step": 11972
},
{
"epoch": 0.7256969696969697,
"grad_norm": 0.04143080115318298,
"learning_rate": 0.0008,
"loss": 2.0255613327026367,
"step": 11974
},
{
"epoch": 0.7258181818181818,
"grad_norm": 0.04720653221011162,
"learning_rate": 0.0008,
"loss": 2.0132408142089844,
"step": 11976
},
{
"epoch": 0.7259393939393939,
"grad_norm": 0.05285007134079933,
"learning_rate": 0.0008,
"loss": 1.9894310235977173,
"step": 11978
},
{
"epoch": 0.7260606060606061,
"grad_norm": 0.046671804040670395,
"learning_rate": 0.0008,
"loss": 1.9330554008483887,
"step": 11980
},
{
"epoch": 0.7261818181818182,
"grad_norm": 0.04439868405461311,
"learning_rate": 0.0008,
"loss": 2.0320887565612793,
"step": 11982
},
{
"epoch": 0.7263030303030303,
"grad_norm": 0.04846612364053726,
"learning_rate": 0.0008,
"loss": 2.024696111679077,
"step": 11984
},
{
"epoch": 0.7264242424242424,
"grad_norm": 0.04778246954083443,
"learning_rate": 0.0008,
"loss": 2.006805419921875,
"step": 11986
},
{
"epoch": 0.7265454545454545,
"grad_norm": 0.04174488037824631,
"learning_rate": 0.0008,
"loss": 1.9990912675857544,
"step": 11988
},
{
"epoch": 0.7266666666666667,
"grad_norm": 0.042433250695466995,
"learning_rate": 0.0008,
"loss": 2.0000369548797607,
"step": 11990
},
{
"epoch": 0.7267878787878788,
"grad_norm": 0.04470886290073395,
"learning_rate": 0.0008,
"loss": 1.956964373588562,
"step": 11992
},
{
"epoch": 0.726909090909091,
"grad_norm": 0.04675665125250816,
"learning_rate": 0.0008,
"loss": 1.9486496448516846,
"step": 11994
},
{
"epoch": 0.727030303030303,
"grad_norm": 0.045978792011737823,
"learning_rate": 0.0008,
"loss": 1.9549237489700317,
"step": 11996
},
{
"epoch": 0.7271515151515151,
"grad_norm": 0.04545130580663681,
"learning_rate": 0.0008,
"loss": 1.9683167934417725,
"step": 11998
},
{
"epoch": 0.7272727272727273,
"grad_norm": 0.0496344231069088,
"learning_rate": 0.0008,
"loss": 2.0214059352874756,
"step": 12000
},
{
"epoch": 0.7273939393939394,
"grad_norm": 0.05226263031363487,
"learning_rate": 0.0008,
"loss": 1.9597829580307007,
"step": 12002
},
{
"epoch": 0.7275151515151516,
"grad_norm": 0.04573979601264,
"learning_rate": 0.0008,
"loss": 1.965517282485962,
"step": 12004
},
{
"epoch": 0.7276363636363636,
"grad_norm": 0.04502628371119499,
"learning_rate": 0.0008,
"loss": 2.0153748989105225,
"step": 12006
},
{
"epoch": 0.7277575757575757,
"grad_norm": 0.05450456216931343,
"learning_rate": 0.0008,
"loss": 2.019644021987915,
"step": 12008
},
{
"epoch": 0.7278787878787879,
"grad_norm": 0.054990608245134354,
"learning_rate": 0.0008,
"loss": 1.9994972944259644,
"step": 12010
},
{
"epoch": 0.728,
"grad_norm": 0.04142531380057335,
"learning_rate": 0.0008,
"loss": 1.9765281677246094,
"step": 12012
},
{
"epoch": 0.7281212121212122,
"grad_norm": 0.04082568734884262,
"learning_rate": 0.0008,
"loss": 1.9717575311660767,
"step": 12014
},
{
"epoch": 0.7282424242424242,
"grad_norm": 0.040222037583589554,
"learning_rate": 0.0008,
"loss": 1.9327293634414673,
"step": 12016
},
{
"epoch": 0.7283636363636363,
"grad_norm": 0.039895374327898026,
"learning_rate": 0.0008,
"loss": 2.0427463054656982,
"step": 12018
},
{
"epoch": 0.7284848484848485,
"grad_norm": 0.041340991854667664,
"learning_rate": 0.0008,
"loss": 1.9427956342697144,
"step": 12020
},
{
"epoch": 0.7286060606060606,
"grad_norm": 0.04247135668992996,
"learning_rate": 0.0008,
"loss": 2.001692056655884,
"step": 12022
},
{
"epoch": 0.7287272727272728,
"grad_norm": 0.04239148274064064,
"learning_rate": 0.0008,
"loss": 1.9965946674346924,
"step": 12024
},
{
"epoch": 0.7288484848484849,
"grad_norm": 0.04122056066989899,
"learning_rate": 0.0008,
"loss": 2.0432868003845215,
"step": 12026
},
{
"epoch": 0.7289696969696969,
"grad_norm": 0.0415252186357975,
"learning_rate": 0.0008,
"loss": 1.9728518724441528,
"step": 12028
},
{
"epoch": 0.7290909090909091,
"grad_norm": 0.04597119241952896,
"learning_rate": 0.0008,
"loss": 1.9267873764038086,
"step": 12030
},
{
"epoch": 0.7292121212121212,
"grad_norm": 0.04975138232111931,
"learning_rate": 0.0008,
"loss": 1.9934945106506348,
"step": 12032
},
{
"epoch": 0.7293333333333333,
"grad_norm": 0.059103887528181076,
"learning_rate": 0.0008,
"loss": 1.9341458082199097,
"step": 12034
},
{
"epoch": 0.7294545454545455,
"grad_norm": 0.051832158118486404,
"learning_rate": 0.0008,
"loss": 1.931018352508545,
"step": 12036
},
{
"epoch": 0.7295757575757575,
"grad_norm": 0.060718003660440445,
"learning_rate": 0.0008,
"loss": 1.9647879600524902,
"step": 12038
},
{
"epoch": 0.7296969696969697,
"grad_norm": 0.041673898696899414,
"learning_rate": 0.0008,
"loss": 1.9922820329666138,
"step": 12040
},
{
"epoch": 0.7298181818181818,
"grad_norm": 0.04040820151567459,
"learning_rate": 0.0008,
"loss": 1.9623812437057495,
"step": 12042
},
{
"epoch": 0.7299393939393939,
"grad_norm": 0.04323505610227585,
"learning_rate": 0.0008,
"loss": 1.9821383953094482,
"step": 12044
},
{
"epoch": 0.7300606060606061,
"grad_norm": 0.039691075682640076,
"learning_rate": 0.0008,
"loss": 1.9889841079711914,
"step": 12046
},
{
"epoch": 0.7301818181818182,
"grad_norm": 0.03991839662194252,
"learning_rate": 0.0008,
"loss": 2.027676582336426,
"step": 12048
},
{
"epoch": 0.7303030303030303,
"grad_norm": 0.04226832464337349,
"learning_rate": 0.0008,
"loss": 1.948804497718811,
"step": 12050
},
{
"epoch": 0.7304242424242424,
"grad_norm": 0.040316320955753326,
"learning_rate": 0.0008,
"loss": 1.9932067394256592,
"step": 12052
},
{
"epoch": 0.7305454545454545,
"grad_norm": 0.03936856612563133,
"learning_rate": 0.0008,
"loss": 1.9450358152389526,
"step": 12054
},
{
"epoch": 0.7306666666666667,
"grad_norm": 0.0453290157020092,
"learning_rate": 0.0008,
"loss": 1.9358327388763428,
"step": 12056
},
{
"epoch": 0.7307878787878788,
"grad_norm": 0.04353158548474312,
"learning_rate": 0.0008,
"loss": 1.9583920240402222,
"step": 12058
},
{
"epoch": 0.730909090909091,
"grad_norm": 0.053481575101614,
"learning_rate": 0.0008,
"loss": 1.9572644233703613,
"step": 12060
},
{
"epoch": 0.731030303030303,
"grad_norm": 0.04531782865524292,
"learning_rate": 0.0008,
"loss": 1.9820594787597656,
"step": 12062
},
{
"epoch": 0.7311515151515151,
"grad_norm": 0.05459542199969292,
"learning_rate": 0.0008,
"loss": 1.9690849781036377,
"step": 12064
},
{
"epoch": 0.7312727272727273,
"grad_norm": 0.04748455807566643,
"learning_rate": 0.0008,
"loss": 1.9694935083389282,
"step": 12066
},
{
"epoch": 0.7313939393939394,
"grad_norm": 0.03941575437784195,
"learning_rate": 0.0008,
"loss": 1.9930355548858643,
"step": 12068
},
{
"epoch": 0.7315151515151516,
"grad_norm": 0.04187697544693947,
"learning_rate": 0.0008,
"loss": 1.9798120260238647,
"step": 12070
},
{
"epoch": 0.7316363636363636,
"grad_norm": 0.040173210203647614,
"learning_rate": 0.0008,
"loss": 1.9375624656677246,
"step": 12072
},
{
"epoch": 0.7317575757575757,
"grad_norm": 0.03915499150753021,
"learning_rate": 0.0008,
"loss": 1.9253424406051636,
"step": 12074
},
{
"epoch": 0.7318787878787879,
"grad_norm": 0.041569579392671585,
"learning_rate": 0.0008,
"loss": 2.01029896736145,
"step": 12076
},
{
"epoch": 0.732,
"grad_norm": 0.04001252353191376,
"learning_rate": 0.0008,
"loss": 1.9717305898666382,
"step": 12078
},
{
"epoch": 0.7321212121212122,
"grad_norm": 0.04034058377146721,
"learning_rate": 0.0008,
"loss": 1.9090557098388672,
"step": 12080
},
{
"epoch": 0.7322424242424243,
"grad_norm": 0.03951490297913551,
"learning_rate": 0.0008,
"loss": 1.898497223854065,
"step": 12082
},
{
"epoch": 0.7323636363636363,
"grad_norm": 0.04109441488981247,
"learning_rate": 0.0008,
"loss": 1.9728502035140991,
"step": 12084
},
{
"epoch": 0.7324848484848485,
"grad_norm": 0.04774521663784981,
"learning_rate": 0.0008,
"loss": 1.9922876358032227,
"step": 12086
},
{
"epoch": 0.7326060606060606,
"grad_norm": 0.0429692342877388,
"learning_rate": 0.0008,
"loss": 1.978742241859436,
"step": 12088
},
{
"epoch": 0.7327272727272728,
"grad_norm": 0.048786845058202744,
"learning_rate": 0.0008,
"loss": 1.9585273265838623,
"step": 12090
},
{
"epoch": 0.7328484848484849,
"grad_norm": 0.04348132759332657,
"learning_rate": 0.0008,
"loss": 2.0412166118621826,
"step": 12092
},
{
"epoch": 0.7329696969696969,
"grad_norm": 0.045670777559280396,
"learning_rate": 0.0008,
"loss": 1.9981297254562378,
"step": 12094
},
{
"epoch": 0.7330909090909091,
"grad_norm": 0.04181252792477608,
"learning_rate": 0.0008,
"loss": 1.9724414348602295,
"step": 12096
},
{
"epoch": 0.7332121212121212,
"grad_norm": 0.0402420312166214,
"learning_rate": 0.0008,
"loss": 1.9012384414672852,
"step": 12098
},
{
"epoch": 0.7333333333333333,
"grad_norm": 0.04539228230714798,
"learning_rate": 0.0008,
"loss": 2.007127285003662,
"step": 12100
},
{
"epoch": 0.7334545454545455,
"grad_norm": 0.04242068529129028,
"learning_rate": 0.0008,
"loss": 2.0146334171295166,
"step": 12102
},
{
"epoch": 0.7335757575757575,
"grad_norm": 0.04039992019534111,
"learning_rate": 0.0008,
"loss": 1.9919602870941162,
"step": 12104
},
{
"epoch": 0.7336969696969697,
"grad_norm": 0.040238868445158005,
"learning_rate": 0.0008,
"loss": 2.024505853652954,
"step": 12106
},
{
"epoch": 0.7338181818181818,
"grad_norm": 0.042125627398490906,
"learning_rate": 0.0008,
"loss": 2.0177836418151855,
"step": 12108
},
{
"epoch": 0.7339393939393939,
"grad_norm": 0.04346601292490959,
"learning_rate": 0.0008,
"loss": 1.9961092472076416,
"step": 12110
},
{
"epoch": 0.7340606060606061,
"grad_norm": 0.045353781431913376,
"learning_rate": 0.0008,
"loss": 1.9533544778823853,
"step": 12112
},
{
"epoch": 0.7341818181818182,
"grad_norm": 0.05673907697200775,
"learning_rate": 0.0008,
"loss": 1.9213250875473022,
"step": 12114
},
{
"epoch": 0.7343030303030303,
"grad_norm": 0.049236807972192764,
"learning_rate": 0.0008,
"loss": 1.9284976720809937,
"step": 12116
},
{
"epoch": 0.7344242424242424,
"grad_norm": 0.056413259357213974,
"learning_rate": 0.0008,
"loss": 1.9118622541427612,
"step": 12118
},
{
"epoch": 0.7345454545454545,
"grad_norm": 0.04394591227173805,
"learning_rate": 0.0008,
"loss": 1.911201000213623,
"step": 12120
},
{
"epoch": 0.7346666666666667,
"grad_norm": 0.04192680865526199,
"learning_rate": 0.0008,
"loss": 1.9481943845748901,
"step": 12122
},
{
"epoch": 0.7347878787878788,
"grad_norm": 0.04155374690890312,
"learning_rate": 0.0008,
"loss": 1.8898277282714844,
"step": 12124
},
{
"epoch": 0.734909090909091,
"grad_norm": 0.040960926562547684,
"learning_rate": 0.0008,
"loss": 1.8867526054382324,
"step": 12126
},
{
"epoch": 0.735030303030303,
"grad_norm": 0.041485514491796494,
"learning_rate": 0.0008,
"loss": 1.972066879272461,
"step": 12128
},
{
"epoch": 0.7351515151515151,
"grad_norm": 0.04194037243723869,
"learning_rate": 0.0008,
"loss": 1.9216762781143188,
"step": 12130
},
{
"epoch": 0.7352727272727273,
"grad_norm": 0.04956291988492012,
"learning_rate": 0.0008,
"loss": 1.9285253286361694,
"step": 12132
},
{
"epoch": 0.7353939393939394,
"grad_norm": 0.04393472149968147,
"learning_rate": 0.0008,
"loss": 1.9645600318908691,
"step": 12134
},
{
"epoch": 0.7355151515151516,
"grad_norm": 0.041283369064331055,
"learning_rate": 0.0008,
"loss": 1.9633026123046875,
"step": 12136
},
{
"epoch": 0.7356363636363636,
"grad_norm": 0.04467035457491875,
"learning_rate": 0.0008,
"loss": 1.9529874324798584,
"step": 12138
},
{
"epoch": 0.7357575757575757,
"grad_norm": 0.03954613581299782,
"learning_rate": 0.0008,
"loss": 1.9031537771224976,
"step": 12140
},
{
"epoch": 0.7358787878787879,
"grad_norm": 0.04291243851184845,
"learning_rate": 0.0008,
"loss": 1.9523775577545166,
"step": 12142
},
{
"epoch": 0.736,
"grad_norm": 0.046416569501161575,
"learning_rate": 0.0008,
"loss": 1.9486794471740723,
"step": 12144
},
{
"epoch": 0.7361212121212122,
"grad_norm": 0.0416623130440712,
"learning_rate": 0.0008,
"loss": 1.9068057537078857,
"step": 12146
},
{
"epoch": 0.7362424242424243,
"grad_norm": 0.05110824853181839,
"learning_rate": 0.0008,
"loss": 1.9255406856536865,
"step": 12148
},
{
"epoch": 0.7363636363636363,
"grad_norm": 0.047163043171167374,
"learning_rate": 0.0008,
"loss": 1.913894772529602,
"step": 12150
},
{
"epoch": 0.7364848484848485,
"grad_norm": 0.04198639467358589,
"learning_rate": 0.0008,
"loss": 2.016226053237915,
"step": 12152
},
{
"epoch": 0.7366060606060606,
"grad_norm": 0.03925275057554245,
"learning_rate": 0.0008,
"loss": 2.0308940410614014,
"step": 12154
},
{
"epoch": 0.7367272727272727,
"grad_norm": 0.04380973428487778,
"learning_rate": 0.0008,
"loss": 1.945684552192688,
"step": 12156
},
{
"epoch": 0.7368484848484849,
"grad_norm": 0.05188988521695137,
"learning_rate": 0.0008,
"loss": 1.9195125102996826,
"step": 12158
},
{
"epoch": 0.7369696969696969,
"grad_norm": 0.05531122535467148,
"learning_rate": 0.0008,
"loss": 1.9109445810317993,
"step": 12160
},
{
"epoch": 0.7370909090909091,
"grad_norm": 0.07270162552595139,
"learning_rate": 0.0008,
"loss": 1.9860447645187378,
"step": 12162
},
{
"epoch": 0.7372121212121212,
"grad_norm": 0.053600847721099854,
"learning_rate": 0.0008,
"loss": 1.9060413837432861,
"step": 12164
},
{
"epoch": 0.7373333333333333,
"grad_norm": 0.04889136925339699,
"learning_rate": 0.0008,
"loss": 1.9559426307678223,
"step": 12166
},
{
"epoch": 0.7374545454545455,
"grad_norm": 0.04513783007860184,
"learning_rate": 0.0008,
"loss": 1.97987961769104,
"step": 12168
},
{
"epoch": 0.7375757575757576,
"grad_norm": 0.04952410236001015,
"learning_rate": 0.0008,
"loss": 1.8827651739120483,
"step": 12170
},
{
"epoch": 0.7376969696969697,
"grad_norm": 0.041803028434515,
"learning_rate": 0.0008,
"loss": 1.9816347360610962,
"step": 12172
},
{
"epoch": 0.7378181818181818,
"grad_norm": 0.04897923022508621,
"learning_rate": 0.0008,
"loss": 2.0044713020324707,
"step": 12174
},
{
"epoch": 0.7379393939393939,
"grad_norm": 0.049475353211164474,
"learning_rate": 0.0008,
"loss": 1.9297679662704468,
"step": 12176
},
{
"epoch": 0.7380606060606061,
"grad_norm": 0.04954821988940239,
"learning_rate": 0.0008,
"loss": 1.953007459640503,
"step": 12178
},
{
"epoch": 0.7381818181818182,
"grad_norm": 0.04240415617823601,
"learning_rate": 0.0008,
"loss": 1.913155436515808,
"step": 12180
},
{
"epoch": 0.7383030303030303,
"grad_norm": 0.04357443377375603,
"learning_rate": 0.0008,
"loss": 1.96974515914917,
"step": 12182
},
{
"epoch": 0.7384242424242424,
"grad_norm": 0.0403817743062973,
"learning_rate": 0.0008,
"loss": 1.90036940574646,
"step": 12184
},
{
"epoch": 0.7385454545454545,
"grad_norm": 0.042505715042352676,
"learning_rate": 0.0008,
"loss": 1.9014513492584229,
"step": 12186
},
{
"epoch": 0.7386666666666667,
"grad_norm": 0.04608554020524025,
"learning_rate": 0.0008,
"loss": 1.8753280639648438,
"step": 12188
},
{
"epoch": 0.7387878787878788,
"grad_norm": 0.04676125571131706,
"learning_rate": 0.0008,
"loss": 1.9641886949539185,
"step": 12190
},
{
"epoch": 0.738909090909091,
"grad_norm": 0.04725400730967522,
"learning_rate": 0.0008,
"loss": 1.9325441122055054,
"step": 12192
},
{
"epoch": 0.739030303030303,
"grad_norm": 0.041792094707489014,
"learning_rate": 0.0008,
"loss": 1.9255062341690063,
"step": 12194
},
{
"epoch": 0.7391515151515151,
"grad_norm": 0.0416441448032856,
"learning_rate": 0.0008,
"loss": 1.9139028787612915,
"step": 12196
},
{
"epoch": 0.7392727272727273,
"grad_norm": 0.040078844875097275,
"learning_rate": 0.0008,
"loss": 1.8411473035812378,
"step": 12198
},
{
"epoch": 0.7393939393939394,
"grad_norm": 0.04046814516186714,
"learning_rate": 0.0008,
"loss": 1.9732271432876587,
"step": 12200
},
{
"epoch": 0.7395151515151516,
"grad_norm": 0.0411464087665081,
"learning_rate": 0.0008,
"loss": 1.8755697011947632,
"step": 12202
},
{
"epoch": 0.7396363636363636,
"grad_norm": 0.041602324694395065,
"learning_rate": 0.0008,
"loss": 1.9441581964492798,
"step": 12204
},
{
"epoch": 0.7397575757575757,
"grad_norm": 0.03926650434732437,
"learning_rate": 0.0008,
"loss": 1.9468858242034912,
"step": 12206
},
{
"epoch": 0.7398787878787879,
"grad_norm": 0.0403447225689888,
"learning_rate": 0.0008,
"loss": 1.9598771333694458,
"step": 12208
},
{
"epoch": 0.74,
"grad_norm": 0.041523173451423645,
"learning_rate": 0.0008,
"loss": 1.9470566511154175,
"step": 12210
},
{
"epoch": 0.7401212121212121,
"grad_norm": 0.04156642034649849,
"learning_rate": 0.0008,
"loss": 1.9746148586273193,
"step": 12212
},
{
"epoch": 0.7402424242424243,
"grad_norm": 0.040530454367399216,
"learning_rate": 0.0008,
"loss": 1.9651297330856323,
"step": 12214
},
{
"epoch": 0.7403636363636363,
"grad_norm": 0.04262766242027283,
"learning_rate": 0.0008,
"loss": 1.945491075515747,
"step": 12216
},
{
"epoch": 0.7404848484848485,
"grad_norm": 0.04373732954263687,
"learning_rate": 0.0008,
"loss": 1.9080889225006104,
"step": 12218
},
{
"epoch": 0.7406060606060606,
"grad_norm": 0.03906963765621185,
"learning_rate": 0.0008,
"loss": 1.8880341053009033,
"step": 12220
},
{
"epoch": 0.7407272727272727,
"grad_norm": 0.039417557418346405,
"learning_rate": 0.0008,
"loss": 1.8887381553649902,
"step": 12222
},
{
"epoch": 0.7408484848484849,
"grad_norm": 0.04517851024866104,
"learning_rate": 0.0008,
"loss": 1.9479358196258545,
"step": 12224
},
{
"epoch": 0.740969696969697,
"grad_norm": 0.040142327547073364,
"learning_rate": 0.0008,
"loss": 1.942565679550171,
"step": 12226
},
{
"epoch": 0.7410909090909091,
"grad_norm": 0.03928595036268234,
"learning_rate": 0.0008,
"loss": 1.9283130168914795,
"step": 12228
},
{
"epoch": 0.7412121212121212,
"grad_norm": 0.05274701863527298,
"learning_rate": 0.0008,
"loss": 1.9219660758972168,
"step": 12230
},
{
"epoch": 0.7413333333333333,
"grad_norm": 0.06899772584438324,
"learning_rate": 0.0008,
"loss": 1.904919147491455,
"step": 12232
},
{
"epoch": 0.7414545454545455,
"grad_norm": 0.042389098554849625,
"learning_rate": 0.0008,
"loss": 1.851293683052063,
"step": 12234
},
{
"epoch": 0.7415757575757576,
"grad_norm": 0.04122532904148102,
"learning_rate": 0.0008,
"loss": 1.8873701095581055,
"step": 12236
},
{
"epoch": 0.7416969696969697,
"grad_norm": 0.041384778916835785,
"learning_rate": 0.0008,
"loss": 1.9711087942123413,
"step": 12238
},
{
"epoch": 0.7418181818181818,
"grad_norm": 0.0407080352306366,
"learning_rate": 0.0008,
"loss": 2.0289978981018066,
"step": 12240
},
{
"epoch": 0.7419393939393939,
"grad_norm": 0.042396508157253265,
"learning_rate": 0.0008,
"loss": 1.9209803342819214,
"step": 12242
},
{
"epoch": 0.7420606060606061,
"grad_norm": 0.04180603474378586,
"learning_rate": 0.0008,
"loss": 1.9352049827575684,
"step": 12244
},
{
"epoch": 0.7421818181818182,
"grad_norm": 0.04147474467754364,
"learning_rate": 0.0008,
"loss": 1.9456499814987183,
"step": 12246
},
{
"epoch": 0.7423030303030304,
"grad_norm": 0.04077368974685669,
"learning_rate": 0.0008,
"loss": 1.950557827949524,
"step": 12248
},
{
"epoch": 0.7424242424242424,
"grad_norm": 0.04155569151043892,
"learning_rate": 0.0008,
"loss": 1.9119054079055786,
"step": 12250
},
{
"epoch": 0.7425454545454545,
"grad_norm": 0.04192004352807999,
"learning_rate": 0.0008,
"loss": 1.8969000577926636,
"step": 12252
},
{
"epoch": 0.7426666666666667,
"grad_norm": 0.041822027415037155,
"learning_rate": 0.0008,
"loss": 1.891443133354187,
"step": 12254
},
{
"epoch": 0.7427878787878788,
"grad_norm": 0.046521417796611786,
"learning_rate": 0.0008,
"loss": 1.9958277940750122,
"step": 12256
},
{
"epoch": 0.742909090909091,
"grad_norm": 0.04029463231563568,
"learning_rate": 0.0008,
"loss": 1.9549375772476196,
"step": 12258
},
{
"epoch": 0.743030303030303,
"grad_norm": 0.04060797020792961,
"learning_rate": 0.0008,
"loss": 1.9204145669937134,
"step": 12260
},
{
"epoch": 0.7431515151515151,
"grad_norm": 0.0410165973007679,
"learning_rate": 0.0008,
"loss": 1.9212898015975952,
"step": 12262
},
{
"epoch": 0.7432727272727273,
"grad_norm": 0.03944450989365578,
"learning_rate": 0.0008,
"loss": 1.8742789030075073,
"step": 12264
},
{
"epoch": 0.7433939393939394,
"grad_norm": 0.046286385506391525,
"learning_rate": 0.0008,
"loss": 1.9605052471160889,
"step": 12266
},
{
"epoch": 0.7435151515151515,
"grad_norm": 0.04094153270125389,
"learning_rate": 0.0008,
"loss": 1.9407663345336914,
"step": 12268
},
{
"epoch": 0.7436363636363637,
"grad_norm": 0.04492645338177681,
"learning_rate": 0.0008,
"loss": 1.903891682624817,
"step": 12270
},
{
"epoch": 0.7437575757575757,
"grad_norm": 0.04228735342621803,
"learning_rate": 0.0008,
"loss": 1.9257431030273438,
"step": 12272
},
{
"epoch": 0.7438787878787879,
"grad_norm": 0.04061284288764,
"learning_rate": 0.0008,
"loss": 1.9282711744308472,
"step": 12274
},
{
"epoch": 0.744,
"grad_norm": 0.039965368807315826,
"learning_rate": 0.0008,
"loss": 1.974097728729248,
"step": 12276
},
{
"epoch": 0.7441212121212121,
"grad_norm": 0.04117044806480408,
"learning_rate": 0.0008,
"loss": 1.8966151475906372,
"step": 12278
},
{
"epoch": 0.7442424242424243,
"grad_norm": 0.05033106356859207,
"learning_rate": 0.0008,
"loss": 1.9542388916015625,
"step": 12280
},
{
"epoch": 0.7443636363636363,
"grad_norm": 0.05322505533695221,
"learning_rate": 0.0008,
"loss": 1.9465372562408447,
"step": 12282
},
{
"epoch": 0.7444848484848485,
"grad_norm": 0.04658607393503189,
"learning_rate": 0.0008,
"loss": 1.9531958103179932,
"step": 12284
},
{
"epoch": 0.7446060606060606,
"grad_norm": 0.05351344868540764,
"learning_rate": 0.0008,
"loss": 1.9371416568756104,
"step": 12286
},
{
"epoch": 0.7447272727272727,
"grad_norm": 0.044670719653367996,
"learning_rate": 0.0008,
"loss": 1.934341311454773,
"step": 12288
},
{
"epoch": 0.7448484848484849,
"grad_norm": 0.05179738253355026,
"learning_rate": 0.0008,
"loss": 1.9792276620864868,
"step": 12290
},
{
"epoch": 0.744969696969697,
"grad_norm": 0.045296456664800644,
"learning_rate": 0.0008,
"loss": 1.9317454099655151,
"step": 12292
},
{
"epoch": 0.7450909090909091,
"grad_norm": 0.04503370821475983,
"learning_rate": 0.0008,
"loss": 1.9118783473968506,
"step": 12294
},
{
"epoch": 0.7452121212121212,
"grad_norm": 0.04079233109951019,
"learning_rate": 0.0008,
"loss": 1.8691902160644531,
"step": 12296
},
{
"epoch": 0.7453333333333333,
"grad_norm": 0.05107549950480461,
"learning_rate": 0.0008,
"loss": 1.9558959007263184,
"step": 12298
},
{
"epoch": 0.7454545454545455,
"grad_norm": 0.046532899141311646,
"learning_rate": 0.0008,
"loss": 1.894897699356079,
"step": 12300
},
{
"epoch": 0.7455757575757576,
"grad_norm": 0.04050144553184509,
"learning_rate": 0.0008,
"loss": 1.9764057397842407,
"step": 12302
},
{
"epoch": 0.7456969696969697,
"grad_norm": 0.04237145557999611,
"learning_rate": 0.0008,
"loss": 1.9206000566482544,
"step": 12304
},
{
"epoch": 0.7458181818181818,
"grad_norm": 0.04348461702466011,
"learning_rate": 0.0008,
"loss": 1.956388235092163,
"step": 12306
},
{
"epoch": 0.7459393939393939,
"grad_norm": 0.041260477155447006,
"learning_rate": 0.0008,
"loss": 1.9610286951065063,
"step": 12308
},
{
"epoch": 0.7460606060606061,
"grad_norm": 0.041825070977211,
"learning_rate": 0.0008,
"loss": 1.8976531028747559,
"step": 12310
},
{
"epoch": 0.7461818181818182,
"grad_norm": 0.03973943367600441,
"learning_rate": 0.0008,
"loss": 1.8873381614685059,
"step": 12312
},
{
"epoch": 0.7463030303030304,
"grad_norm": 0.04058300331234932,
"learning_rate": 0.0008,
"loss": 1.8993664979934692,
"step": 12314
},
{
"epoch": 0.7464242424242424,
"grad_norm": 0.04060116410255432,
"learning_rate": 0.0008,
"loss": 1.9527009725570679,
"step": 12316
},
{
"epoch": 0.7465454545454545,
"grad_norm": 0.04818102344870567,
"learning_rate": 0.0008,
"loss": 1.9136301279067993,
"step": 12318
},
{
"epoch": 0.7466666666666667,
"grad_norm": 0.05494755506515503,
"learning_rate": 0.0008,
"loss": 1.9189568758010864,
"step": 12320
},
{
"epoch": 0.7467878787878788,
"grad_norm": 0.04097892716526985,
"learning_rate": 0.0008,
"loss": 1.9139842987060547,
"step": 12322
},
{
"epoch": 0.7469090909090909,
"grad_norm": 0.041015900671482086,
"learning_rate": 0.0008,
"loss": 1.884967565536499,
"step": 12324
},
{
"epoch": 0.747030303030303,
"grad_norm": 0.04775072634220123,
"learning_rate": 0.0008,
"loss": 1.8924553394317627,
"step": 12326
},
{
"epoch": 0.7471515151515151,
"grad_norm": 0.05017458274960518,
"learning_rate": 0.0008,
"loss": 1.9673182964324951,
"step": 12328
},
{
"epoch": 0.7472727272727273,
"grad_norm": 0.05333387106657028,
"learning_rate": 0.0008,
"loss": 1.9149526357650757,
"step": 12330
},
{
"epoch": 0.7473939393939394,
"grad_norm": 0.06312637776136398,
"learning_rate": 0.0008,
"loss": 1.9371182918548584,
"step": 12332
},
{
"epoch": 0.7475151515151515,
"grad_norm": 0.05349436774849892,
"learning_rate": 0.0008,
"loss": 1.9065301418304443,
"step": 12334
},
{
"epoch": 0.7476363636363637,
"grad_norm": 0.0535939559340477,
"learning_rate": 0.0008,
"loss": 1.9337109327316284,
"step": 12336
},
{
"epoch": 0.7477575757575757,
"grad_norm": 0.054625123739242554,
"learning_rate": 0.0008,
"loss": 1.9370628595352173,
"step": 12338
},
{
"epoch": 0.7478787878787879,
"grad_norm": 0.0413084402680397,
"learning_rate": 0.0008,
"loss": 2.014044761657715,
"step": 12340
},
{
"epoch": 0.748,
"grad_norm": 0.04525915905833244,
"learning_rate": 0.0008,
"loss": 1.9253066778182983,
"step": 12342
},
{
"epoch": 0.7481212121212121,
"grad_norm": 0.046556733548641205,
"learning_rate": 0.0008,
"loss": 1.9690871238708496,
"step": 12344
},
{
"epoch": 0.7482424242424243,
"grad_norm": 0.04782700538635254,
"learning_rate": 0.0008,
"loss": 1.9266886711120605,
"step": 12346
},
{
"epoch": 0.7483636363636363,
"grad_norm": 0.05460929870605469,
"learning_rate": 0.0008,
"loss": 1.922353982925415,
"step": 12348
},
{
"epoch": 0.7484848484848485,
"grad_norm": 0.052135854959487915,
"learning_rate": 0.0008,
"loss": 1.9245554208755493,
"step": 12350
},
{
"epoch": 0.7486060606060606,
"grad_norm": 0.04769505187869072,
"learning_rate": 0.0008,
"loss": 1.8674814701080322,
"step": 12352
},
{
"epoch": 0.7487272727272727,
"grad_norm": 0.04303627088665962,
"learning_rate": 0.0008,
"loss": 1.8684985637664795,
"step": 12354
},
{
"epoch": 0.7488484848484849,
"grad_norm": 0.04147672280669212,
"learning_rate": 0.0008,
"loss": 1.9356356859207153,
"step": 12356
},
{
"epoch": 0.748969696969697,
"grad_norm": 0.04027855768799782,
"learning_rate": 0.0008,
"loss": 1.939682960510254,
"step": 12358
},
{
"epoch": 0.7490909090909091,
"grad_norm": 0.03998201712965965,
"learning_rate": 0.0008,
"loss": 1.9771558046340942,
"step": 12360
},
{
"epoch": 0.7492121212121212,
"grad_norm": 0.04026809334754944,
"learning_rate": 0.0008,
"loss": 1.9299170970916748,
"step": 12362
},
{
"epoch": 0.7493333333333333,
"grad_norm": 0.03963082283735275,
"learning_rate": 0.0008,
"loss": 1.9303760528564453,
"step": 12364
},
{
"epoch": 0.7494545454545455,
"grad_norm": 0.04689796268939972,
"learning_rate": 0.0008,
"loss": 1.9214311838150024,
"step": 12366
},
{
"epoch": 0.7495757575757576,
"grad_norm": 0.041698701679706573,
"learning_rate": 0.0008,
"loss": 1.9598876237869263,
"step": 12368
},
{
"epoch": 0.7496969696969698,
"grad_norm": 0.039739321917295456,
"learning_rate": 0.0008,
"loss": 1.8897398710250854,
"step": 12370
},
{
"epoch": 0.7498181818181818,
"grad_norm": 0.04015382379293442,
"learning_rate": 0.0008,
"loss": 1.907642126083374,
"step": 12372
},
{
"epoch": 0.7499393939393939,
"grad_norm": 0.0411757230758667,
"learning_rate": 0.0008,
"loss": 1.8794636726379395,
"step": 12374
},
{
"epoch": 0.7500606060606061,
"grad_norm": 0.0425344854593277,
"learning_rate": 0.0008,
"loss": 1.919438362121582,
"step": 12376
},
{
"epoch": 0.7501818181818182,
"grad_norm": 0.0391160286962986,
"learning_rate": 0.0008,
"loss": 1.903922438621521,
"step": 12378
},
{
"epoch": 0.7503030303030302,
"grad_norm": 0.03957950696349144,
"learning_rate": 0.0008,
"loss": 1.9609332084655762,
"step": 12380
},
{
"epoch": 0.7504242424242424,
"grad_norm": 0.04061266407370567,
"learning_rate": 0.0008,
"loss": 1.9355077743530273,
"step": 12382
},
{
"epoch": 0.7505454545454545,
"grad_norm": 0.04159567132592201,
"learning_rate": 0.0008,
"loss": 1.936100721359253,
"step": 12384
},
{
"epoch": 0.7506666666666667,
"grad_norm": 0.040454622358083725,
"learning_rate": 0.0008,
"loss": 1.9224562644958496,
"step": 12386
},
{
"epoch": 0.7507878787878788,
"grad_norm": 0.04397231340408325,
"learning_rate": 0.0008,
"loss": 1.9190309047698975,
"step": 12388
},
{
"epoch": 0.7509090909090909,
"grad_norm": 0.04284265637397766,
"learning_rate": 0.0008,
"loss": 1.8991312980651855,
"step": 12390
},
{
"epoch": 0.751030303030303,
"grad_norm": 0.042032357305288315,
"learning_rate": 0.0008,
"loss": 1.8615827560424805,
"step": 12392
},
{
"epoch": 0.7511515151515151,
"grad_norm": 0.04201758652925491,
"learning_rate": 0.0008,
"loss": 1.9364423751831055,
"step": 12394
},
{
"epoch": 0.7512727272727273,
"grad_norm": 0.04195248335599899,
"learning_rate": 0.0008,
"loss": 1.9397224187850952,
"step": 12396
},
{
"epoch": 0.7513939393939394,
"grad_norm": 0.042408160865306854,
"learning_rate": 0.0008,
"loss": 1.8700059652328491,
"step": 12398
},
{
"epoch": 0.7515151515151515,
"grad_norm": 0.04476973041892052,
"learning_rate": 0.0008,
"loss": 1.941730260848999,
"step": 12400
},
{
"epoch": 0.7516363636363637,
"grad_norm": 0.04097463935613632,
"learning_rate": 0.0008,
"loss": 1.8662835359573364,
"step": 12402
},
{
"epoch": 0.7517575757575757,
"grad_norm": 0.04379766434431076,
"learning_rate": 0.0008,
"loss": 1.921968936920166,
"step": 12404
},
{
"epoch": 0.7518787878787879,
"grad_norm": 0.05306120216846466,
"learning_rate": 0.0008,
"loss": 1.9623997211456299,
"step": 12406
},
{
"epoch": 0.752,
"grad_norm": 0.041039515286684036,
"learning_rate": 0.0008,
"loss": 1.8784147500991821,
"step": 12408
},
{
"epoch": 0.7521212121212121,
"grad_norm": 0.039580605924129486,
"learning_rate": 0.0008,
"loss": 1.9188328981399536,
"step": 12410
},
{
"epoch": 0.7522424242424243,
"grad_norm": 0.04109292849898338,
"learning_rate": 0.0008,
"loss": 1.9561936855316162,
"step": 12412
},
{
"epoch": 0.7523636363636363,
"grad_norm": 0.04563502222299576,
"learning_rate": 0.0008,
"loss": 1.8869822025299072,
"step": 12414
},
{
"epoch": 0.7524848484848485,
"grad_norm": 0.041725438088178635,
"learning_rate": 0.0008,
"loss": 1.8920483589172363,
"step": 12416
},
{
"epoch": 0.7526060606060606,
"grad_norm": 0.04310137778520584,
"learning_rate": 0.0008,
"loss": 1.892112374305725,
"step": 12418
},
{
"epoch": 0.7527272727272727,
"grad_norm": 0.05249340459704399,
"learning_rate": 0.0008,
"loss": 1.9380022287368774,
"step": 12420
},
{
"epoch": 0.7528484848484849,
"grad_norm": 0.04349524527788162,
"learning_rate": 0.0008,
"loss": 1.9122140407562256,
"step": 12422
},
{
"epoch": 0.752969696969697,
"grad_norm": 0.04598882421851158,
"learning_rate": 0.0008,
"loss": 1.859177589416504,
"step": 12424
},
{
"epoch": 0.7530909090909091,
"grad_norm": 0.041458312422037125,
"learning_rate": 0.0008,
"loss": 1.9014281034469604,
"step": 12426
},
{
"epoch": 0.7532121212121212,
"grad_norm": 0.039004553109407425,
"learning_rate": 0.0008,
"loss": 1.8781837224960327,
"step": 12428
},
{
"epoch": 0.7533333333333333,
"grad_norm": 0.0437355600297451,
"learning_rate": 0.0008,
"loss": 1.8867950439453125,
"step": 12430
},
{
"epoch": 0.7534545454545455,
"grad_norm": 0.04183635115623474,
"learning_rate": 0.0008,
"loss": 1.8934451341629028,
"step": 12432
},
{
"epoch": 0.7535757575757576,
"grad_norm": 0.041300833225250244,
"learning_rate": 0.0008,
"loss": 1.95620858669281,
"step": 12434
},
{
"epoch": 0.7536969696969696,
"grad_norm": 0.05058787390589714,
"learning_rate": 0.0008,
"loss": 1.9007691144943237,
"step": 12436
},
{
"epoch": 0.7538181818181818,
"grad_norm": 0.06207270175218582,
"learning_rate": 0.0008,
"loss": 1.9066078662872314,
"step": 12438
},
{
"epoch": 0.7539393939393939,
"grad_norm": 0.04556097462773323,
"learning_rate": 0.0008,
"loss": 1.9757392406463623,
"step": 12440
},
{
"epoch": 0.7540606060606061,
"grad_norm": 0.04745195433497429,
"learning_rate": 0.0008,
"loss": 1.880365014076233,
"step": 12442
},
{
"epoch": 0.7541818181818182,
"grad_norm": 0.04044004902243614,
"learning_rate": 0.0008,
"loss": 1.9088897705078125,
"step": 12444
},
{
"epoch": 0.7543030303030303,
"grad_norm": 0.04719934239983559,
"learning_rate": 0.0008,
"loss": 1.9348244667053223,
"step": 12446
},
{
"epoch": 0.7544242424242424,
"grad_norm": 0.05637798085808754,
"learning_rate": 0.0008,
"loss": 1.8710517883300781,
"step": 12448
},
{
"epoch": 0.7545454545454545,
"grad_norm": 0.047449368983507156,
"learning_rate": 0.0008,
"loss": 1.8995680809020996,
"step": 12450
},
{
"epoch": 0.7546666666666667,
"grad_norm": 0.04883674159646034,
"learning_rate": 0.0008,
"loss": 1.8500092029571533,
"step": 12452
},
{
"epoch": 0.7547878787878788,
"grad_norm": 0.0469537228345871,
"learning_rate": 0.0008,
"loss": 1.8536643981933594,
"step": 12454
},
{
"epoch": 0.7549090909090909,
"grad_norm": 0.042527616024017334,
"learning_rate": 0.0008,
"loss": 1.8685930967330933,
"step": 12456
},
{
"epoch": 0.755030303030303,
"grad_norm": 0.044008005410432816,
"learning_rate": 0.0008,
"loss": 1.9145790338516235,
"step": 12458
},
{
"epoch": 0.7551515151515151,
"grad_norm": 0.04029818996787071,
"learning_rate": 0.0008,
"loss": 1.8761041164398193,
"step": 12460
},
{
"epoch": 0.7552727272727273,
"grad_norm": 0.03993184119462967,
"learning_rate": 0.0008,
"loss": 1.9144967794418335,
"step": 12462
},
{
"epoch": 0.7553939393939394,
"grad_norm": 0.041082222014665604,
"learning_rate": 0.0008,
"loss": 1.8798047304153442,
"step": 12464
},
{
"epoch": 0.7555151515151515,
"grad_norm": 0.04202492535114288,
"learning_rate": 0.0008,
"loss": 1.8857797384262085,
"step": 12466
},
{
"epoch": 0.7556363636363637,
"grad_norm": 0.04616932198405266,
"learning_rate": 0.0008,
"loss": 1.9386630058288574,
"step": 12468
},
{
"epoch": 0.7557575757575757,
"grad_norm": 0.06223063915967941,
"learning_rate": 0.0008,
"loss": 1.8964784145355225,
"step": 12470
},
{
"epoch": 0.7558787878787879,
"grad_norm": 0.06291936337947845,
"learning_rate": 0.0008,
"loss": 1.8849889039993286,
"step": 12472
},
{
"epoch": 0.756,
"grad_norm": 0.07074002176523209,
"learning_rate": 0.0008,
"loss": 1.919274091720581,
"step": 12474
},
{
"epoch": 0.7561212121212121,
"grad_norm": 0.06419520825147629,
"learning_rate": 0.0008,
"loss": 2.0070557594299316,
"step": 12476
},
{
"epoch": 0.7562424242424243,
"grad_norm": 0.04597286507487297,
"learning_rate": 0.0008,
"loss": 1.8809528350830078,
"step": 12478
},
{
"epoch": 0.7563636363636363,
"grad_norm": 0.04435879737138748,
"learning_rate": 0.0008,
"loss": 1.9050508737564087,
"step": 12480
},
{
"epoch": 0.7564848484848485,
"grad_norm": 0.04802217707037926,
"learning_rate": 0.0008,
"loss": 1.9153412580490112,
"step": 12482
},
{
"epoch": 0.7566060606060606,
"grad_norm": 0.05901272967457771,
"learning_rate": 0.0008,
"loss": 1.933168888092041,
"step": 12484
},
{
"epoch": 0.7567272727272727,
"grad_norm": 0.050420235842466354,
"learning_rate": 0.0008,
"loss": 1.9016897678375244,
"step": 12486
},
{
"epoch": 0.7568484848484849,
"grad_norm": 0.04131673276424408,
"learning_rate": 0.0008,
"loss": 1.9164897203445435,
"step": 12488
},
{
"epoch": 0.756969696969697,
"grad_norm": 0.04352710396051407,
"learning_rate": 0.0008,
"loss": 1.9316494464874268,
"step": 12490
},
{
"epoch": 0.757090909090909,
"grad_norm": 0.04312160611152649,
"learning_rate": 0.0008,
"loss": 1.8902578353881836,
"step": 12492
},
{
"epoch": 0.7572121212121212,
"grad_norm": 0.04919472336769104,
"learning_rate": 0.0008,
"loss": 1.9300537109375,
"step": 12494
},
{
"epoch": 0.7573333333333333,
"grad_norm": 0.047493986785411835,
"learning_rate": 0.0008,
"loss": 1.8853490352630615,
"step": 12496
},
{
"epoch": 0.7574545454545455,
"grad_norm": 0.04893500730395317,
"learning_rate": 0.0008,
"loss": 1.9015085697174072,
"step": 12498
},
{
"epoch": 0.7575757575757576,
"grad_norm": 0.0465058758854866,
"learning_rate": 0.0008,
"loss": 1.9932453632354736,
"step": 12500
},
{
"epoch": 0.7576969696969696,
"grad_norm": 0.04403264820575714,
"learning_rate": 0.0008,
"loss": 1.8591182231903076,
"step": 12502
},
{
"epoch": 0.7578181818181818,
"grad_norm": 0.04409673437476158,
"learning_rate": 0.0008,
"loss": 1.9207223653793335,
"step": 12504
},
{
"epoch": 0.7579393939393939,
"grad_norm": 0.039403509348630905,
"learning_rate": 0.0008,
"loss": 1.931728482246399,
"step": 12506
},
{
"epoch": 0.7580606060606061,
"grad_norm": 0.04270164668560028,
"learning_rate": 0.0008,
"loss": 1.9592889547348022,
"step": 12508
},
{
"epoch": 0.7581818181818182,
"grad_norm": 0.04047751799225807,
"learning_rate": 0.0008,
"loss": 1.8297315835952759,
"step": 12510
},
{
"epoch": 0.7583030303030303,
"grad_norm": 0.04091103747487068,
"learning_rate": 0.0008,
"loss": 1.8626033067703247,
"step": 12512
},
{
"epoch": 0.7584242424242424,
"grad_norm": 0.041250117123126984,
"learning_rate": 0.0008,
"loss": 1.8053828477859497,
"step": 12514
},
{
"epoch": 0.7585454545454545,
"grad_norm": 0.04978077858686447,
"learning_rate": 0.0008,
"loss": 1.8834515810012817,
"step": 12516
},
{
"epoch": 0.7586666666666667,
"grad_norm": 0.04206831380724907,
"learning_rate": 0.0008,
"loss": 1.9484175443649292,
"step": 12518
},
{
"epoch": 0.7587878787878788,
"grad_norm": 0.044600121676921844,
"learning_rate": 0.0008,
"loss": 1.8809443712234497,
"step": 12520
},
{
"epoch": 0.7589090909090909,
"grad_norm": 0.04046646133065224,
"learning_rate": 0.0008,
"loss": 1.8794641494750977,
"step": 12522
},
{
"epoch": 0.759030303030303,
"grad_norm": 0.04319394379854202,
"learning_rate": 0.0008,
"loss": 1.8876198530197144,
"step": 12524
},
{
"epoch": 0.7591515151515151,
"grad_norm": 0.043096017092466354,
"learning_rate": 0.0008,
"loss": 1.9373538494110107,
"step": 12526
},
{
"epoch": 0.7592727272727273,
"grad_norm": 0.042504001408815384,
"learning_rate": 0.0008,
"loss": 1.9288936853408813,
"step": 12528
},
{
"epoch": 0.7593939393939394,
"grad_norm": 0.04411248117685318,
"learning_rate": 0.0008,
"loss": 1.882681131362915,
"step": 12530
},
{
"epoch": 0.7595151515151515,
"grad_norm": 0.04517735168337822,
"learning_rate": 0.0008,
"loss": 1.9494891166687012,
"step": 12532
},
{
"epoch": 0.7596363636363637,
"grad_norm": 0.042022302746772766,
"learning_rate": 0.0008,
"loss": 1.8580163717269897,
"step": 12534
},
{
"epoch": 0.7597575757575757,
"grad_norm": 0.04121841862797737,
"learning_rate": 0.0008,
"loss": 1.9425431489944458,
"step": 12536
},
{
"epoch": 0.7598787878787879,
"grad_norm": 0.03963598608970642,
"learning_rate": 0.0008,
"loss": 1.908063292503357,
"step": 12538
},
{
"epoch": 0.76,
"grad_norm": 0.04273993521928787,
"learning_rate": 0.0008,
"loss": 1.9308580160140991,
"step": 12540
},
{
"epoch": 0.7601212121212121,
"grad_norm": 0.04483119770884514,
"learning_rate": 0.0007997979797979798,
"loss": 2.0346462726593018,
"step": 12542
},
{
"epoch": 0.7602424242424243,
"grad_norm": 0.040438830852508545,
"learning_rate": 0.0007993939393939395,
"loss": 1.979292392730713,
"step": 12544
},
{
"epoch": 0.7603636363636364,
"grad_norm": 0.04930206760764122,
"learning_rate": 0.000798989898989899,
"loss": 2.0374951362609863,
"step": 12546
},
{
"epoch": 0.7604848484848484,
"grad_norm": 0.04073861986398697,
"learning_rate": 0.0007985858585858586,
"loss": 1.9666820764541626,
"step": 12548
},
{
"epoch": 0.7606060606060606,
"grad_norm": 0.04052956774830818,
"learning_rate": 0.0007981818181818182,
"loss": 1.9964886903762817,
"step": 12550
},
{
"epoch": 0.7607272727272727,
"grad_norm": 0.038861118257045746,
"learning_rate": 0.0007977777777777778,
"loss": 2.013064384460449,
"step": 12552
},
{
"epoch": 0.7608484848484849,
"grad_norm": 0.04100293666124344,
"learning_rate": 0.0007973737373737374,
"loss": 2.008998155593872,
"step": 12554
},
{
"epoch": 0.760969696969697,
"grad_norm": 0.042416807264089584,
"learning_rate": 0.000796969696969697,
"loss": 1.9424998760223389,
"step": 12556
},
{
"epoch": 0.761090909090909,
"grad_norm": 0.040263108909130096,
"learning_rate": 0.0007965656565656566,
"loss": 2.018378973007202,
"step": 12558
},
{
"epoch": 0.7612121212121212,
"grad_norm": 0.04443824291229248,
"learning_rate": 0.0007961616161616162,
"loss": 2.00608491897583,
"step": 12560
},
{
"epoch": 0.7613333333333333,
"grad_norm": 0.0416722372174263,
"learning_rate": 0.0007957575757575759,
"loss": 2.014683246612549,
"step": 12562
},
{
"epoch": 0.7614545454545455,
"grad_norm": 0.042002227157354355,
"learning_rate": 0.0007953535353535354,
"loss": 1.984579086303711,
"step": 12564
},
{
"epoch": 0.7615757575757576,
"grad_norm": 0.0397065132856369,
"learning_rate": 0.000794949494949495,
"loss": 1.9998563528060913,
"step": 12566
},
{
"epoch": 0.7616969696969697,
"grad_norm": 0.04016793519258499,
"learning_rate": 0.0007945454545454546,
"loss": 2.0625665187835693,
"step": 12568
},
{
"epoch": 0.7618181818181818,
"grad_norm": 0.04011380672454834,
"learning_rate": 0.0007941414141414142,
"loss": 2.0061252117156982,
"step": 12570
},
{
"epoch": 0.7619393939393939,
"grad_norm": 0.05220978707075119,
"learning_rate": 0.0007937373737373737,
"loss": 2.0691137313842773,
"step": 12572
},
{
"epoch": 0.7620606060606061,
"grad_norm": 0.047436244785785675,
"learning_rate": 0.0007933333333333334,
"loss": 2.090104818344116,
"step": 12574
},
{
"epoch": 0.7621818181818182,
"grad_norm": 0.0454322025179863,
"learning_rate": 0.000792929292929293,
"loss": 2.0175178050994873,
"step": 12576
},
{
"epoch": 0.7623030303030303,
"grad_norm": 0.04178448021411896,
"learning_rate": 0.0007925252525252525,
"loss": 2.065824270248413,
"step": 12578
},
{
"epoch": 0.7624242424242424,
"grad_norm": 0.04005878046154976,
"learning_rate": 0.0007921212121212121,
"loss": 2.117156982421875,
"step": 12580
},
{
"epoch": 0.7625454545454545,
"grad_norm": 0.04271911829710007,
"learning_rate": 0.0007917171717171718,
"loss": 2.119072675704956,
"step": 12582
},
{
"epoch": 0.7626666666666667,
"grad_norm": 0.041054822504520416,
"learning_rate": 0.0007913131313131314,
"loss": 2.0002260208129883,
"step": 12584
},
{
"epoch": 0.7627878787878788,
"grad_norm": 0.04326825216412544,
"learning_rate": 0.0007909090909090909,
"loss": 2.062181234359741,
"step": 12586
},
{
"epoch": 0.7629090909090909,
"grad_norm": 0.04097198694944382,
"learning_rate": 0.0007905050505050505,
"loss": 2.1092638969421387,
"step": 12588
},
{
"epoch": 0.7630303030303031,
"grad_norm": 0.04077252373099327,
"learning_rate": 0.0007901010101010102,
"loss": 2.071500539779663,
"step": 12590
},
{
"epoch": 0.7631515151515151,
"grad_norm": 0.043130528181791306,
"learning_rate": 0.0007896969696969697,
"loss": 2.1335792541503906,
"step": 12592
},
{
"epoch": 0.7632727272727273,
"grad_norm": 0.043429844081401825,
"learning_rate": 0.0007892929292929293,
"loss": 2.0660293102264404,
"step": 12594
},
{
"epoch": 0.7633939393939394,
"grad_norm": 0.04609135910868645,
"learning_rate": 0.000788888888888889,
"loss": 2.182990550994873,
"step": 12596
},
{
"epoch": 0.7635151515151515,
"grad_norm": 0.04338185489177704,
"learning_rate": 0.0007884848484848486,
"loss": 2.155475378036499,
"step": 12598
},
{
"epoch": 0.7636363636363637,
"grad_norm": 0.042257003486156464,
"learning_rate": 0.0007880808080808081,
"loss": 2.0997626781463623,
"step": 12600
},
{
"epoch": 0.7637575757575757,
"grad_norm": 0.040354736149311066,
"learning_rate": 0.0007876767676767677,
"loss": 2.129809856414795,
"step": 12602
},
{
"epoch": 0.7638787878787878,
"grad_norm": 0.04482565447688103,
"learning_rate": 0.0007872727272727274,
"loss": 2.1358137130737305,
"step": 12604
},
{
"epoch": 0.764,
"grad_norm": 0.04288671910762787,
"learning_rate": 0.0007868686868686868,
"loss": 2.142606735229492,
"step": 12606
},
{
"epoch": 0.7641212121212121,
"grad_norm": 0.043324846774339676,
"learning_rate": 0.0007864646464646465,
"loss": 2.175297498703003,
"step": 12608
},
{
"epoch": 0.7642424242424243,
"grad_norm": 0.04380409047007561,
"learning_rate": 0.0007860606060606061,
"loss": 2.071028232574463,
"step": 12610
},
{
"epoch": 0.7643636363636364,
"grad_norm": 0.04719284176826477,
"learning_rate": 0.0007856565656565657,
"loss": 2.1706879138946533,
"step": 12612
},
{
"epoch": 0.7644848484848484,
"grad_norm": 0.052769482135772705,
"learning_rate": 0.0007852525252525252,
"loss": 2.2532317638397217,
"step": 12614
},
{
"epoch": 0.7646060606060606,
"grad_norm": 0.0526919886469841,
"learning_rate": 0.0007848484848484849,
"loss": 2.216707944869995,
"step": 12616
},
{
"epoch": 0.7647272727272727,
"grad_norm": 0.05870664864778519,
"learning_rate": 0.0007844444444444445,
"loss": 2.2171475887298584,
"step": 12618
},
{
"epoch": 0.7648484848484849,
"grad_norm": 0.04490828141570091,
"learning_rate": 0.0007840404040404041,
"loss": 2.140615940093994,
"step": 12620
},
{
"epoch": 0.764969696969697,
"grad_norm": 0.04558146744966507,
"learning_rate": 0.0007836363636363636,
"loss": 2.2159841060638428,
"step": 12622
},
{
"epoch": 0.765090909090909,
"grad_norm": 0.06300142407417297,
"learning_rate": 0.0007832323232323233,
"loss": 2.253589391708374,
"step": 12624
},
{
"epoch": 0.7652121212121212,
"grad_norm": 0.056389063596725464,
"learning_rate": 0.0007828282828282828,
"loss": 2.1611223220825195,
"step": 12626
},
{
"epoch": 0.7653333333333333,
"grad_norm": 0.05626522749662399,
"learning_rate": 0.0007824242424242425,
"loss": 2.1815221309661865,
"step": 12628
},
{
"epoch": 0.7654545454545455,
"grad_norm": 0.04235878214240074,
"learning_rate": 0.0007820202020202021,
"loss": 2.205601692199707,
"step": 12630
},
{
"epoch": 0.7655757575757576,
"grad_norm": 0.0455097034573555,
"learning_rate": 0.0007816161616161616,
"loss": 2.1855411529541016,
"step": 12632
},
{
"epoch": 0.7656969696969697,
"grad_norm": 0.042771559208631516,
"learning_rate": 0.0007812121212121213,
"loss": 2.126619577407837,
"step": 12634
},
{
"epoch": 0.7658181818181818,
"grad_norm": 0.047885436564683914,
"learning_rate": 0.0007808080808080808,
"loss": 2.224959373474121,
"step": 12636
},
{
"epoch": 0.7659393939393939,
"grad_norm": 0.044399600476026535,
"learning_rate": 0.0007804040404040405,
"loss": 2.187551498413086,
"step": 12638
},
{
"epoch": 0.7660606060606061,
"grad_norm": 0.04438615217804909,
"learning_rate": 0.00078,
"loss": 2.261439561843872,
"step": 12640
},
{
"epoch": 0.7661818181818182,
"grad_norm": 0.06503322720527649,
"learning_rate": 0.0007795959595959597,
"loss": 2.2749171257019043,
"step": 12642
},
{
"epoch": 0.7663030303030303,
"grad_norm": 0.0585797056555748,
"learning_rate": 0.0007791919191919192,
"loss": 2.209646701812744,
"step": 12644
},
{
"epoch": 0.7664242424242425,
"grad_norm": 0.04714217782020569,
"learning_rate": 0.0007787878787878788,
"loss": 2.196881055831909,
"step": 12646
},
{
"epoch": 0.7665454545454545,
"grad_norm": 0.041872140020132065,
"learning_rate": 0.0007783838383838384,
"loss": 2.188270330429077,
"step": 12648
},
{
"epoch": 0.7666666666666667,
"grad_norm": 0.04126197099685669,
"learning_rate": 0.0007779797979797981,
"loss": 2.2716634273529053,
"step": 12650
},
{
"epoch": 0.7667878787878788,
"grad_norm": 0.04184809327125549,
"learning_rate": 0.0007775757575757576,
"loss": 2.245332717895508,
"step": 12652
},
{
"epoch": 0.7669090909090909,
"grad_norm": 0.04115922003984451,
"learning_rate": 0.0007771717171717172,
"loss": 2.2641806602478027,
"step": 12654
},
{
"epoch": 0.7670303030303031,
"grad_norm": 0.046709779649972916,
"learning_rate": 0.0007767676767676768,
"loss": 2.31485652923584,
"step": 12656
},
{
"epoch": 0.7671515151515151,
"grad_norm": 0.04386667162179947,
"learning_rate": 0.0007763636363636364,
"loss": 2.274606227874756,
"step": 12658
},
{
"epoch": 0.7672727272727272,
"grad_norm": 0.04100430756807327,
"learning_rate": 0.0007759595959595959,
"loss": 2.2658705711364746,
"step": 12660
},
{
"epoch": 0.7673939393939394,
"grad_norm": 0.04348524287343025,
"learning_rate": 0.0007755555555555556,
"loss": 2.2559895515441895,
"step": 12662
},
{
"epoch": 0.7675151515151515,
"grad_norm": 0.044867586344480515,
"learning_rate": 0.0007751515151515153,
"loss": 2.3376922607421875,
"step": 12664
},
{
"epoch": 0.7676363636363637,
"grad_norm": 0.04329673945903778,
"learning_rate": 0.0007747474747474747,
"loss": 2.2448296546936035,
"step": 12666
},
{
"epoch": 0.7677575757575757,
"grad_norm": 0.0431828536093235,
"learning_rate": 0.0007743434343434344,
"loss": 2.2692770957946777,
"step": 12668
},
{
"epoch": 0.7678787878787878,
"grad_norm": 0.045329418033361435,
"learning_rate": 0.000773939393939394,
"loss": 2.297987222671509,
"step": 12670
},
{
"epoch": 0.768,
"grad_norm": 0.05611593648791313,
"learning_rate": 0.0007735353535353536,
"loss": 2.3240864276885986,
"step": 12672
},
{
"epoch": 0.7681212121212121,
"grad_norm": 0.04655918478965759,
"learning_rate": 0.0007731313131313131,
"loss": 2.343942642211914,
"step": 12674
},
{
"epoch": 0.7682424242424243,
"grad_norm": 0.04309716075658798,
"learning_rate": 0.0007727272727272728,
"loss": 2.3185503482818604,
"step": 12676
},
{
"epoch": 0.7683636363636364,
"grad_norm": 0.04363667964935303,
"learning_rate": 0.0007723232323232324,
"loss": 2.269561290740967,
"step": 12678
},
{
"epoch": 0.7684848484848484,
"grad_norm": 0.04190044850111008,
"learning_rate": 0.0007719191919191919,
"loss": 2.331587314605713,
"step": 12680
},
{
"epoch": 0.7686060606060606,
"grad_norm": 0.04826276749372482,
"learning_rate": 0.0007715151515151515,
"loss": 2.2813880443573,
"step": 12682
},
{
"epoch": 0.7687272727272727,
"grad_norm": 0.04747844859957695,
"learning_rate": 0.0007711111111111112,
"loss": 2.3475375175476074,
"step": 12684
},
{
"epoch": 0.7688484848484849,
"grad_norm": 0.052638374269008636,
"learning_rate": 0.0007707070707070708,
"loss": 2.321424961090088,
"step": 12686
},
{
"epoch": 0.768969696969697,
"grad_norm": 0.06004733592271805,
"learning_rate": 0.0007703030303030303,
"loss": 2.3726699352264404,
"step": 12688
},
{
"epoch": 0.769090909090909,
"grad_norm": 0.05712248384952545,
"learning_rate": 0.0007698989898989899,
"loss": 2.3676395416259766,
"step": 12690
},
{
"epoch": 0.7692121212121212,
"grad_norm": 0.0554472990334034,
"learning_rate": 0.0007694949494949496,
"loss": 2.3520894050598145,
"step": 12692
},
{
"epoch": 0.7693333333333333,
"grad_norm": 0.05357610434293747,
"learning_rate": 0.000769090909090909,
"loss": 2.3988537788391113,
"step": 12694
},
{
"epoch": 0.7694545454545455,
"grad_norm": 0.05547565594315529,
"learning_rate": 0.0007686868686868687,
"loss": 2.3897652626037598,
"step": 12696
},
{
"epoch": 0.7695757575757576,
"grad_norm": 0.04544910416007042,
"learning_rate": 0.0007682828282828284,
"loss": 2.364074230194092,
"step": 12698
},
{
"epoch": 0.7696969696969697,
"grad_norm": 0.045219916850328445,
"learning_rate": 0.0007678787878787879,
"loss": 2.371718406677246,
"step": 12700
},
{
"epoch": 0.7698181818181818,
"grad_norm": 0.04224534332752228,
"learning_rate": 0.0007674747474747475,
"loss": 2.4037485122680664,
"step": 12702
},
{
"epoch": 0.7699393939393939,
"grad_norm": 0.04393499717116356,
"learning_rate": 0.0007670707070707071,
"loss": 2.4616916179656982,
"step": 12704
},
{
"epoch": 0.7700606060606061,
"grad_norm": 0.04540430009365082,
"learning_rate": 0.0007666666666666667,
"loss": 2.45333194732666,
"step": 12706
},
{
"epoch": 0.7701818181818182,
"grad_norm": 0.043596651405096054,
"learning_rate": 0.0007662626262626263,
"loss": 2.4061942100524902,
"step": 12708
},
{
"epoch": 0.7703030303030303,
"grad_norm": 0.043312471359968185,
"learning_rate": 0.0007658585858585859,
"loss": 2.3886594772338867,
"step": 12710
},
{
"epoch": 0.7704242424242425,
"grad_norm": 0.04348092898726463,
"learning_rate": 0.0007654545454545455,
"loss": 2.3782169818878174,
"step": 12712
},
{
"epoch": 0.7705454545454545,
"grad_norm": 0.05074289068579674,
"learning_rate": 0.0007650505050505051,
"loss": 2.415584087371826,
"step": 12714
},
{
"epoch": 0.7706666666666667,
"grad_norm": 0.045656416565179825,
"learning_rate": 0.0007646464646464647,
"loss": 2.4495229721069336,
"step": 12716
},
{
"epoch": 0.7707878787878788,
"grad_norm": 0.06331637501716614,
"learning_rate": 0.0007642424242424243,
"loss": 2.4427340030670166,
"step": 12718
},
{
"epoch": 0.7709090909090909,
"grad_norm": 0.05128614977002144,
"learning_rate": 0.0007638383838383838,
"loss": 2.433253526687622,
"step": 12720
},
{
"epoch": 0.7710303030303031,
"grad_norm": 0.04832685738801956,
"learning_rate": 0.0007634343434343435,
"loss": 2.4685308933258057,
"step": 12722
},
{
"epoch": 0.7711515151515151,
"grad_norm": 0.04649219289422035,
"learning_rate": 0.000763030303030303,
"loss": 2.4227824211120605,
"step": 12724
},
{
"epoch": 0.7712727272727272,
"grad_norm": 0.04685407131910324,
"learning_rate": 0.0007626262626262626,
"loss": 2.3481926918029785,
"step": 12726
},
{
"epoch": 0.7713939393939394,
"grad_norm": 0.04710349813103676,
"learning_rate": 0.0007622222222222222,
"loss": 2.443406105041504,
"step": 12728
},
{
"epoch": 0.7715151515151515,
"grad_norm": 0.04605654254555702,
"learning_rate": 0.0007618181818181819,
"loss": 2.4351346492767334,
"step": 12730
},
{
"epoch": 0.7716363636363637,
"grad_norm": 0.04289453849196434,
"learning_rate": 0.0007614141414141415,
"loss": 2.4550678730010986,
"step": 12732
},
{
"epoch": 0.7717575757575758,
"grad_norm": 0.04204421117901802,
"learning_rate": 0.000761010101010101,
"loss": 2.4629085063934326,
"step": 12734
},
{
"epoch": 0.7718787878787878,
"grad_norm": 0.044543780386447906,
"learning_rate": 0.0007606060606060607,
"loss": 2.4982309341430664,
"step": 12736
},
{
"epoch": 0.772,
"grad_norm": 0.05362275242805481,
"learning_rate": 0.0007602020202020203,
"loss": 2.470094919204712,
"step": 12738
},
{
"epoch": 0.7721212121212121,
"grad_norm": 0.05270489677786827,
"learning_rate": 0.0007597979797979798,
"loss": 2.4776763916015625,
"step": 12740
},
{
"epoch": 0.7722424242424243,
"grad_norm": 0.04878697916865349,
"learning_rate": 0.0007593939393939394,
"loss": 2.4092442989349365,
"step": 12742
},
{
"epoch": 0.7723636363636364,
"grad_norm": 0.04614836722612381,
"learning_rate": 0.0007589898989898991,
"loss": 2.4621574878692627,
"step": 12744
},
{
"epoch": 0.7724848484848484,
"grad_norm": 0.04521850869059563,
"learning_rate": 0.0007585858585858586,
"loss": 2.499734401702881,
"step": 12746
},
{
"epoch": 0.7726060606060606,
"grad_norm": 0.04390499368309975,
"learning_rate": 0.0007581818181818182,
"loss": 2.501100540161133,
"step": 12748
},
{
"epoch": 0.7727272727272727,
"grad_norm": 0.045842792838811874,
"learning_rate": 0.0007577777777777778,
"loss": 2.475497007369995,
"step": 12750
},
{
"epoch": 0.7728484848484849,
"grad_norm": 0.04886357858777046,
"learning_rate": 0.0007573737373737375,
"loss": 2.445254325866699,
"step": 12752
},
{
"epoch": 0.772969696969697,
"grad_norm": 0.0511283203959465,
"learning_rate": 0.0007569696969696969,
"loss": 2.480238914489746,
"step": 12754
},
{
"epoch": 0.773090909090909,
"grad_norm": 0.04500778019428253,
"learning_rate": 0.0007565656565656566,
"loss": 2.4958903789520264,
"step": 12756
},
{
"epoch": 0.7732121212121212,
"grad_norm": 0.047281090170145035,
"learning_rate": 0.0007561616161616162,
"loss": 2.448146343231201,
"step": 12758
},
{
"epoch": 0.7733333333333333,
"grad_norm": 0.044455934315919876,
"learning_rate": 0.0007557575757575758,
"loss": 2.518587350845337,
"step": 12760
},
{
"epoch": 0.7734545454545455,
"grad_norm": 0.043076545000076294,
"learning_rate": 0.0007553535353535353,
"loss": 2.510380744934082,
"step": 12762
},
{
"epoch": 0.7735757575757576,
"grad_norm": 0.04863133653998375,
"learning_rate": 0.000754949494949495,
"loss": 2.552694320678711,
"step": 12764
},
{
"epoch": 0.7736969696969697,
"grad_norm": 0.04575743526220322,
"learning_rate": 0.0007545454545454546,
"loss": 2.482844352722168,
"step": 12766
},
{
"epoch": 0.7738181818181818,
"grad_norm": 0.04459463059902191,
"learning_rate": 0.0007541414141414141,
"loss": 2.4896492958068848,
"step": 12768
},
{
"epoch": 0.7739393939393939,
"grad_norm": 0.04890776425600052,
"learning_rate": 0.0007537373737373738,
"loss": 2.4572486877441406,
"step": 12770
},
{
"epoch": 0.7740606060606061,
"grad_norm": 0.04549407213926315,
"learning_rate": 0.0007533333333333334,
"loss": 2.491011381149292,
"step": 12772
},
{
"epoch": 0.7741818181818182,
"grad_norm": 0.04394189640879631,
"learning_rate": 0.000752929292929293,
"loss": 2.494438886642456,
"step": 12774
},
{
"epoch": 0.7743030303030303,
"grad_norm": 0.04909466952085495,
"learning_rate": 0.0007525252525252525,
"loss": 2.511922836303711,
"step": 12776
},
{
"epoch": 0.7744242424242425,
"grad_norm": 0.04470570385456085,
"learning_rate": 0.0007521212121212122,
"loss": 2.533024787902832,
"step": 12778
},
{
"epoch": 0.7745454545454545,
"grad_norm": 0.04376324266195297,
"learning_rate": 0.0007517171717171717,
"loss": 2.528724193572998,
"step": 12780
},
{
"epoch": 0.7746666666666666,
"grad_norm": 0.042851898819208145,
"learning_rate": 0.0007513131313131314,
"loss": 2.4977169036865234,
"step": 12782
},
{
"epoch": 0.7747878787878788,
"grad_norm": 0.047850750386714935,
"learning_rate": 0.0007509090909090909,
"loss": 2.643157482147217,
"step": 12784
},
{
"epoch": 0.7749090909090909,
"grad_norm": 0.0448644682765007,
"learning_rate": 0.0007505050505050506,
"loss": 2.5524837970733643,
"step": 12786
},
{
"epoch": 0.7750303030303031,
"grad_norm": 0.04567348212003708,
"learning_rate": 0.0007501010101010101,
"loss": 2.577564001083374,
"step": 12788
},
{
"epoch": 0.7751515151515151,
"grad_norm": 0.046540822833776474,
"learning_rate": 0.0007496969696969697,
"loss": 2.5711677074432373,
"step": 12790
},
{
"epoch": 0.7752727272727272,
"grad_norm": 0.047974999994039536,
"learning_rate": 0.0007492929292929293,
"loss": 2.4955170154571533,
"step": 12792
},
{
"epoch": 0.7753939393939394,
"grad_norm": 0.05598282068967819,
"learning_rate": 0.0007488888888888889,
"loss": 2.5107104778289795,
"step": 12794
},
{
"epoch": 0.7755151515151515,
"grad_norm": 0.06310189515352249,
"learning_rate": 0.0007484848484848485,
"loss": 2.542616844177246,
"step": 12796
},
{
"epoch": 0.7756363636363637,
"grad_norm": 0.07566085457801819,
"learning_rate": 0.0007480808080808081,
"loss": 2.652275800704956,
"step": 12798
},
{
"epoch": 0.7757575757575758,
"grad_norm": 0.07600195705890656,
"learning_rate": 0.0007476767676767677,
"loss": 2.6317834854125977,
"step": 12800
},
{
"epoch": 0.7758787878787878,
"grad_norm": 0.07576204091310501,
"learning_rate": 0.0007472727272727273,
"loss": 2.628864049911499,
"step": 12802
},
{
"epoch": 0.776,
"grad_norm": 0.05260759964585304,
"learning_rate": 0.000746868686868687,
"loss": 2.575460910797119,
"step": 12804
},
{
"epoch": 0.7761212121212121,
"grad_norm": 0.0445241704583168,
"learning_rate": 0.0007464646464646465,
"loss": 2.5369937419891357,
"step": 12806
},
{
"epoch": 0.7762424242424243,
"grad_norm": 0.04242302477359772,
"learning_rate": 0.0007460606060606061,
"loss": 2.58801531791687,
"step": 12808
},
{
"epoch": 0.7763636363636364,
"grad_norm": 0.042774222791194916,
"learning_rate": 0.0007456565656565657,
"loss": 2.5607545375823975,
"step": 12810
},
{
"epoch": 0.7764848484848484,
"grad_norm": 0.043507445603609085,
"learning_rate": 0.0007452525252525254,
"loss": 2.580716371536255,
"step": 12812
},
{
"epoch": 0.7766060606060606,
"grad_norm": 0.04378197714686394,
"learning_rate": 0.0007448484848484848,
"loss": 2.5691936016082764,
"step": 12814
},
{
"epoch": 0.7767272727272727,
"grad_norm": 0.04529546946287155,
"learning_rate": 0.0007444444444444445,
"loss": 2.5750393867492676,
"step": 12816
},
{
"epoch": 0.7768484848484849,
"grad_norm": 0.053495798259973526,
"learning_rate": 0.0007440404040404041,
"loss": 2.5900967121124268,
"step": 12818
},
{
"epoch": 0.776969696969697,
"grad_norm": 0.05101901665329933,
"learning_rate": 0.0007436363636363636,
"loss": 2.5482840538024902,
"step": 12820
},
{
"epoch": 0.777090909090909,
"grad_norm": 0.060583245009183884,
"learning_rate": 0.0007432323232323232,
"loss": 2.5953850746154785,
"step": 12822
},
{
"epoch": 0.7772121212121212,
"grad_norm": 0.052583303302526474,
"learning_rate": 0.0007428282828282829,
"loss": 2.6629562377929688,
"step": 12824
},
{
"epoch": 0.7773333333333333,
"grad_norm": 0.04834415391087532,
"learning_rate": 0.0007424242424242425,
"loss": 2.5758848190307617,
"step": 12826
},
{
"epoch": 0.7774545454545455,
"grad_norm": 0.04479902610182762,
"learning_rate": 0.000742020202020202,
"loss": 2.547616720199585,
"step": 12828
},
{
"epoch": 0.7775757575757576,
"grad_norm": 0.05078183859586716,
"learning_rate": 0.0007416161616161616,
"loss": 2.612529993057251,
"step": 12830
},
{
"epoch": 0.7776969696969697,
"grad_norm": 0.05115080997347832,
"learning_rate": 0.0007412121212121213,
"loss": 2.6378161907196045,
"step": 12832
},
{
"epoch": 0.7778181818181819,
"grad_norm": 0.052601709961891174,
"learning_rate": 0.0007408080808080808,
"loss": 2.6748812198638916,
"step": 12834
},
{
"epoch": 0.7779393939393939,
"grad_norm": 0.04368458688259125,
"learning_rate": 0.0007404040404040404,
"loss": 2.5678372383117676,
"step": 12836
},
{
"epoch": 0.778060606060606,
"grad_norm": 0.05059303715825081,
"learning_rate": 0.0007400000000000001,
"loss": 2.6174116134643555,
"step": 12838
},
{
"epoch": 0.7781818181818182,
"grad_norm": 0.04612511396408081,
"learning_rate": 0.0007395959595959597,
"loss": 2.6149260997772217,
"step": 12840
},
{
"epoch": 0.7783030303030303,
"grad_norm": 0.049865275621414185,
"learning_rate": 0.0007391919191919192,
"loss": 2.6045048236846924,
"step": 12842
},
{
"epoch": 0.7784242424242425,
"grad_norm": 0.05206015333533287,
"learning_rate": 0.0007387878787878788,
"loss": 2.6025266647338867,
"step": 12844
},
{
"epoch": 0.7785454545454545,
"grad_norm": 0.05972357839345932,
"learning_rate": 0.0007383838383838385,
"loss": 2.615468978881836,
"step": 12846
},
{
"epoch": 0.7786666666666666,
"grad_norm": 0.05425636097788811,
"learning_rate": 0.000737979797979798,
"loss": 2.635451078414917,
"step": 12848
},
{
"epoch": 0.7787878787878788,
"grad_norm": 0.04588595777750015,
"learning_rate": 0.0007375757575757576,
"loss": 2.616811990737915,
"step": 12850
},
{
"epoch": 0.7789090909090909,
"grad_norm": 0.05082259699702263,
"learning_rate": 0.0007371717171717172,
"loss": 2.6473841667175293,
"step": 12852
},
{
"epoch": 0.7790303030303031,
"grad_norm": 0.04821845144033432,
"learning_rate": 0.0007367676767676768,
"loss": 2.7158071994781494,
"step": 12854
},
{
"epoch": 0.7791515151515152,
"grad_norm": 0.04695912078022957,
"learning_rate": 0.0007363636363636363,
"loss": 2.6932218074798584,
"step": 12856
},
{
"epoch": 0.7792727272727272,
"grad_norm": 0.04561370238661766,
"learning_rate": 0.000735959595959596,
"loss": 2.6690220832824707,
"step": 12858
},
{
"epoch": 0.7793939393939394,
"grad_norm": 0.05558645725250244,
"learning_rate": 0.0007355555555555556,
"loss": 2.628401279449463,
"step": 12860
},
{
"epoch": 0.7795151515151515,
"grad_norm": 0.04310929402709007,
"learning_rate": 0.0007351515151515152,
"loss": 2.7098588943481445,
"step": 12862
},
{
"epoch": 0.7796363636363637,
"grad_norm": 0.04445216432213783,
"learning_rate": 0.0007347474747474747,
"loss": 2.6680805683135986,
"step": 12864
},
{
"epoch": 0.7797575757575758,
"grad_norm": 0.04591483995318413,
"learning_rate": 0.0007343434343434344,
"loss": 2.673837184906006,
"step": 12866
},
{
"epoch": 0.7798787878787878,
"grad_norm": 0.046467218548059464,
"learning_rate": 0.000733939393939394,
"loss": 2.6872410774230957,
"step": 12868
},
{
"epoch": 0.78,
"grad_norm": 0.044584546238183975,
"learning_rate": 0.0007335353535353536,
"loss": 2.7282626628875732,
"step": 12870
},
{
"epoch": 0.7801212121212121,
"grad_norm": 0.04937903583049774,
"learning_rate": 0.0007331313131313132,
"loss": 2.7435553073883057,
"step": 12872
},
{
"epoch": 0.7802424242424243,
"grad_norm": 0.05095362663269043,
"learning_rate": 0.0007327272727272727,
"loss": 2.765939474105835,
"step": 12874
},
{
"epoch": 0.7803636363636364,
"grad_norm": 0.04567648470401764,
"learning_rate": 0.0007323232323232324,
"loss": 2.699469566345215,
"step": 12876
},
{
"epoch": 0.7804848484848484,
"grad_norm": 0.05411459878087044,
"learning_rate": 0.000731919191919192,
"loss": 2.704542875289917,
"step": 12878
},
{
"epoch": 0.7806060606060606,
"grad_norm": 0.05779287591576576,
"learning_rate": 0.0007315151515151516,
"loss": 2.643942356109619,
"step": 12880
},
{
"epoch": 0.7807272727272727,
"grad_norm": 0.045056987553834915,
"learning_rate": 0.0007311111111111111,
"loss": 2.724759817123413,
"step": 12882
},
{
"epoch": 0.7808484848484849,
"grad_norm": 0.06000228971242905,
"learning_rate": 0.0007307070707070708,
"loss": 2.685663938522339,
"step": 12884
},
{
"epoch": 0.780969696969697,
"grad_norm": 0.04651818796992302,
"learning_rate": 0.0007303030303030303,
"loss": 2.6892802715301514,
"step": 12886
},
{
"epoch": 0.7810909090909091,
"grad_norm": 0.047202080488204956,
"learning_rate": 0.0007298989898989899,
"loss": 2.738321542739868,
"step": 12888
},
{
"epoch": 0.7812121212121212,
"grad_norm": 0.05054826661944389,
"learning_rate": 0.0007294949494949495,
"loss": 2.6726486682891846,
"step": 12890
},
{
"epoch": 0.7813333333333333,
"grad_norm": 0.044609908014535904,
"learning_rate": 0.0007290909090909092,
"loss": 2.6633214950561523,
"step": 12892
},
{
"epoch": 0.7814545454545454,
"grad_norm": 0.045361876487731934,
"learning_rate": 0.0007286868686868687,
"loss": 2.72220516204834,
"step": 12894
},
{
"epoch": 0.7815757575757576,
"grad_norm": 0.045131269842386246,
"learning_rate": 0.0007282828282828283,
"loss": 2.7291667461395264,
"step": 12896
},
{
"epoch": 0.7816969696969697,
"grad_norm": 0.047776106745004654,
"learning_rate": 0.0007278787878787879,
"loss": 2.70430326461792,
"step": 12898
},
{
"epoch": 0.7818181818181819,
"grad_norm": 0.04864230379462242,
"learning_rate": 0.0007274747474747476,
"loss": 2.7283806800842285,
"step": 12900
},
{
"epoch": 0.7819393939393939,
"grad_norm": 0.049562204629182816,
"learning_rate": 0.0007270707070707071,
"loss": 2.6835734844207764,
"step": 12902
},
{
"epoch": 0.782060606060606,
"grad_norm": 0.048225220292806625,
"learning_rate": 0.0007266666666666667,
"loss": 2.7117843627929688,
"step": 12904
},
{
"epoch": 0.7821818181818182,
"grad_norm": 0.06183773651719093,
"learning_rate": 0.0007262626262626264,
"loss": 2.7601511478424072,
"step": 12906
},
{
"epoch": 0.7823030303030303,
"grad_norm": 0.04668540507555008,
"learning_rate": 0.0007258585858585858,
"loss": 2.6988728046417236,
"step": 12908
},
{
"epoch": 0.7824242424242425,
"grad_norm": 0.04506373405456543,
"learning_rate": 0.0007254545454545455,
"loss": 2.7016801834106445,
"step": 12910
},
{
"epoch": 0.7825454545454545,
"grad_norm": 0.045195143669843674,
"learning_rate": 0.0007250505050505051,
"loss": 2.713531970977783,
"step": 12912
},
{
"epoch": 0.7826666666666666,
"grad_norm": 0.04459849372506142,
"learning_rate": 0.0007246464646464647,
"loss": 2.6549806594848633,
"step": 12914
},
{
"epoch": 0.7827878787878788,
"grad_norm": 0.0514596551656723,
"learning_rate": 0.0007242424242424242,
"loss": 2.721348762512207,
"step": 12916
},
{
"epoch": 0.7829090909090909,
"grad_norm": 0.06073107197880745,
"learning_rate": 0.0007238383838383839,
"loss": 2.7245302200317383,
"step": 12918
},
{
"epoch": 0.7830303030303031,
"grad_norm": 0.060770634561777115,
"learning_rate": 0.0007234343434343435,
"loss": 2.7223246097564697,
"step": 12920
},
{
"epoch": 0.7831515151515152,
"grad_norm": 0.054707907140254974,
"learning_rate": 0.000723030303030303,
"loss": 2.698256492614746,
"step": 12922
},
{
"epoch": 0.7832727272727272,
"grad_norm": 0.05258503556251526,
"learning_rate": 0.0007226262626262626,
"loss": 2.720781087875366,
"step": 12924
},
{
"epoch": 0.7833939393939394,
"grad_norm": 0.04706183075904846,
"learning_rate": 0.0007222222222222223,
"loss": 2.7440528869628906,
"step": 12926
},
{
"epoch": 0.7835151515151515,
"grad_norm": 0.04530370607972145,
"learning_rate": 0.0007218181818181819,
"loss": 2.704131841659546,
"step": 12928
},
{
"epoch": 0.7836363636363637,
"grad_norm": 0.04579959437251091,
"learning_rate": 0.0007214141414141414,
"loss": 2.718264579772949,
"step": 12930
},
{
"epoch": 0.7837575757575758,
"grad_norm": 0.04477856308221817,
"learning_rate": 0.000721010101010101,
"loss": 2.698681592941284,
"step": 12932
},
{
"epoch": 0.7838787878787878,
"grad_norm": 0.05169568210840225,
"learning_rate": 0.0007206060606060607,
"loss": 2.6752841472625732,
"step": 12934
},
{
"epoch": 0.784,
"grad_norm": 0.047062329947948456,
"learning_rate": 0.0007202020202020203,
"loss": 2.7047150135040283,
"step": 12936
},
{
"epoch": 0.7841212121212121,
"grad_norm": 0.045735280960798264,
"learning_rate": 0.0007197979797979798,
"loss": 2.726062059402466,
"step": 12938
},
{
"epoch": 0.7842424242424243,
"grad_norm": 0.05146170035004616,
"learning_rate": 0.0007193939393939395,
"loss": 2.7242558002471924,
"step": 12940
},
{
"epoch": 0.7843636363636364,
"grad_norm": 0.04407952353358269,
"learning_rate": 0.000718989898989899,
"loss": 2.7610719203948975,
"step": 12942
},
{
"epoch": 0.7844848484848485,
"grad_norm": 0.052351050078868866,
"learning_rate": 0.0007185858585858587,
"loss": 2.705827236175537,
"step": 12944
},
{
"epoch": 0.7846060606060606,
"grad_norm": 0.05580949783325195,
"learning_rate": 0.0007181818181818182,
"loss": 2.7750582695007324,
"step": 12946
},
{
"epoch": 0.7847272727272727,
"grad_norm": 0.0536104217171669,
"learning_rate": 0.0007177777777777778,
"loss": 2.7427170276641846,
"step": 12948
},
{
"epoch": 0.7848484848484848,
"grad_norm": 0.05571214109659195,
"learning_rate": 0.0007173737373737374,
"loss": 2.7508697509765625,
"step": 12950
},
{
"epoch": 0.784969696969697,
"grad_norm": 0.05746801570057869,
"learning_rate": 0.000716969696969697,
"loss": 2.673309564590454,
"step": 12952
},
{
"epoch": 0.7850909090909091,
"grad_norm": 0.06027154251933098,
"learning_rate": 0.0007165656565656566,
"loss": 2.695446014404297,
"step": 12954
},
{
"epoch": 0.7852121212121213,
"grad_norm": 0.049843501299619675,
"learning_rate": 0.0007161616161616162,
"loss": 2.782951831817627,
"step": 12956
},
{
"epoch": 0.7853333333333333,
"grad_norm": 0.050677720457315445,
"learning_rate": 0.0007157575757575758,
"loss": 2.726875066757202,
"step": 12958
},
{
"epoch": 0.7854545454545454,
"grad_norm": 0.04752598702907562,
"learning_rate": 0.0007153535353535354,
"loss": 2.756880044937134,
"step": 12960
},
{
"epoch": 0.7855757575757576,
"grad_norm": 0.05594700574874878,
"learning_rate": 0.0007149494949494949,
"loss": 2.7204201221466064,
"step": 12962
},
{
"epoch": 0.7856969696969697,
"grad_norm": 0.046631257981061935,
"learning_rate": 0.0007145454545454546,
"loss": 2.7649831771850586,
"step": 12964
},
{
"epoch": 0.7858181818181819,
"grad_norm": 0.05859227851033211,
"learning_rate": 0.0007141414141414141,
"loss": 2.7316203117370605,
"step": 12966
},
{
"epoch": 0.7859393939393939,
"grad_norm": 0.059217147529125214,
"learning_rate": 0.0007137373737373737,
"loss": 2.6934635639190674,
"step": 12968
},
{
"epoch": 0.786060606060606,
"grad_norm": 0.06383583694696426,
"learning_rate": 0.0007133333333333334,
"loss": 2.75429630279541,
"step": 12970
},
{
"epoch": 0.7861818181818182,
"grad_norm": 0.06373012810945511,
"learning_rate": 0.000712929292929293,
"loss": 2.7576754093170166,
"step": 12972
},
{
"epoch": 0.7863030303030303,
"grad_norm": 0.05767196789383888,
"learning_rate": 0.0007125252525252526,
"loss": 2.7992441654205322,
"step": 12974
},
{
"epoch": 0.7864242424242425,
"grad_norm": 0.050360869616270065,
"learning_rate": 0.0007121212121212121,
"loss": 2.73905086517334,
"step": 12976
},
{
"epoch": 0.7865454545454545,
"grad_norm": 0.04965699091553688,
"learning_rate": 0.0007117171717171718,
"loss": 2.7936692237854004,
"step": 12978
},
{
"epoch": 0.7866666666666666,
"grad_norm": 0.051724161952733994,
"learning_rate": 0.0007113131313131314,
"loss": 2.749941110610962,
"step": 12980
},
{
"epoch": 0.7867878787878788,
"grad_norm": 0.04734734445810318,
"learning_rate": 0.0007109090909090909,
"loss": 2.729030132293701,
"step": 12982
},
{
"epoch": 0.7869090909090909,
"grad_norm": 0.050445009022951126,
"learning_rate": 0.0007105050505050505,
"loss": 2.7496376037597656,
"step": 12984
},
{
"epoch": 0.7870303030303031,
"grad_norm": 0.06057169288396835,
"learning_rate": 0.0007101010101010102,
"loss": 2.76599383354187,
"step": 12986
},
{
"epoch": 0.7871515151515152,
"grad_norm": 0.055337272584438324,
"learning_rate": 0.0007096969696969698,
"loss": 2.7353909015655518,
"step": 12988
},
{
"epoch": 0.7872727272727272,
"grad_norm": 0.046721674501895905,
"learning_rate": 0.0007092929292929293,
"loss": 2.7717623710632324,
"step": 12990
},
{
"epoch": 0.7873939393939394,
"grad_norm": 0.0477215014398098,
"learning_rate": 0.0007088888888888889,
"loss": 2.747539520263672,
"step": 12992
},
{
"epoch": 0.7875151515151515,
"grad_norm": 0.05087020993232727,
"learning_rate": 0.0007084848484848486,
"loss": 2.7480525970458984,
"step": 12994
},
{
"epoch": 0.7876363636363637,
"grad_norm": 0.05453432723879814,
"learning_rate": 0.000708080808080808,
"loss": 2.738396644592285,
"step": 12996
},
{
"epoch": 0.7877575757575758,
"grad_norm": 0.055683087557554245,
"learning_rate": 0.0007076767676767677,
"loss": 2.6789801120758057,
"step": 12998
},
{
"epoch": 0.7878787878787878,
"grad_norm": 0.04935922846198082,
"learning_rate": 0.0007072727272727273,
"loss": 2.7556052207946777,
"step": 13000
},
{
"epoch": 0.788,
"grad_norm": 0.04959288239479065,
"learning_rate": 0.0007068686868686869,
"loss": 2.7819857597351074,
"step": 13002
},
{
"epoch": 0.7881212121212121,
"grad_norm": 0.05278705060482025,
"learning_rate": 0.0007064646464646465,
"loss": 2.7619662284851074,
"step": 13004
},
{
"epoch": 0.7882424242424242,
"grad_norm": 0.0550333634018898,
"learning_rate": 0.0007060606060606061,
"loss": 2.7777469158172607,
"step": 13006
},
{
"epoch": 0.7883636363636364,
"grad_norm": 0.06331103295087814,
"learning_rate": 0.0007056565656565657,
"loss": 2.776226043701172,
"step": 13008
},
{
"epoch": 0.7884848484848485,
"grad_norm": 0.05263159051537514,
"learning_rate": 0.0007052525252525252,
"loss": 2.773317813873291,
"step": 13010
},
{
"epoch": 0.7886060606060606,
"grad_norm": 0.04408787563443184,
"learning_rate": 0.0007048484848484849,
"loss": 2.7556824684143066,
"step": 13012
},
{
"epoch": 0.7887272727272727,
"grad_norm": 0.04878704622387886,
"learning_rate": 0.0007044444444444445,
"loss": 2.780618667602539,
"step": 13014
},
{
"epoch": 0.7888484848484848,
"grad_norm": 0.0501292422413826,
"learning_rate": 0.0007040404040404041,
"loss": 2.8232169151306152,
"step": 13016
},
{
"epoch": 0.788969696969697,
"grad_norm": 0.04706967994570732,
"learning_rate": 0.0007036363636363636,
"loss": 2.763132095336914,
"step": 13018
},
{
"epoch": 0.7890909090909091,
"grad_norm": 0.04455282539129257,
"learning_rate": 0.0007032323232323233,
"loss": 2.777132511138916,
"step": 13020
},
{
"epoch": 0.7892121212121213,
"grad_norm": 0.057565778493881226,
"learning_rate": 0.0007028282828282829,
"loss": 2.766697406768799,
"step": 13022
},
{
"epoch": 0.7893333333333333,
"grad_norm": 0.057665493339300156,
"learning_rate": 0.0007024242424242425,
"loss": 2.7381362915039062,
"step": 13024
},
{
"epoch": 0.7894545454545454,
"grad_norm": 0.054578278213739395,
"learning_rate": 0.000702020202020202,
"loss": 2.7788350582122803,
"step": 13026
},
{
"epoch": 0.7895757575757576,
"grad_norm": 0.053561292588710785,
"learning_rate": 0.0007016161616161617,
"loss": 2.8209919929504395,
"step": 13028
},
{
"epoch": 0.7896969696969697,
"grad_norm": 0.048120979219675064,
"learning_rate": 0.0007012121212121212,
"loss": 2.7841873168945312,
"step": 13030
},
{
"epoch": 0.7898181818181819,
"grad_norm": 0.045977674424648285,
"learning_rate": 0.0007008080808080809,
"loss": 2.7455508708953857,
"step": 13032
},
{
"epoch": 0.7899393939393939,
"grad_norm": 0.05141545459628105,
"learning_rate": 0.0007004040404040404,
"loss": 2.7507729530334473,
"step": 13034
},
{
"epoch": 0.790060606060606,
"grad_norm": 0.045574966818094254,
"learning_rate": 0.0007,
"loss": 2.7910284996032715,
"step": 13036
},
{
"epoch": 0.7901818181818182,
"grad_norm": 0.052448056638240814,
"learning_rate": 0.0006995959595959597,
"loss": 2.754842758178711,
"step": 13038
},
{
"epoch": 0.7903030303030303,
"grad_norm": 0.07158570736646652,
"learning_rate": 0.0006991919191919192,
"loss": 2.804783344268799,
"step": 13040
},
{
"epoch": 0.7904242424242425,
"grad_norm": 0.0506352074444294,
"learning_rate": 0.0006987878787878788,
"loss": 2.7546157836914062,
"step": 13042
},
{
"epoch": 0.7905454545454546,
"grad_norm": 0.05178079381585121,
"learning_rate": 0.0006983838383838384,
"loss": 2.7702338695526123,
"step": 13044
},
{
"epoch": 0.7906666666666666,
"grad_norm": 0.047850269824266434,
"learning_rate": 0.0006979797979797981,
"loss": 2.670844793319702,
"step": 13046
},
{
"epoch": 0.7907878787878788,
"grad_norm": 0.050300486385822296,
"learning_rate": 0.0006975757575757576,
"loss": 2.7778849601745605,
"step": 13048
},
{
"epoch": 0.7909090909090909,
"grad_norm": 0.0465228408575058,
"learning_rate": 0.0006971717171717172,
"loss": 2.771880865097046,
"step": 13050
},
{
"epoch": 0.7910303030303031,
"grad_norm": 0.04728807508945465,
"learning_rate": 0.0006967676767676768,
"loss": 2.7457480430603027,
"step": 13052
},
{
"epoch": 0.7911515151515152,
"grad_norm": 0.048885516822338104,
"learning_rate": 0.0006963636363636365,
"loss": 2.757521152496338,
"step": 13054
},
{
"epoch": 0.7912727272727272,
"grad_norm": 0.04819686338305473,
"learning_rate": 0.0006959595959595959,
"loss": 2.7572150230407715,
"step": 13056
},
{
"epoch": 0.7913939393939394,
"grad_norm": 0.04992295056581497,
"learning_rate": 0.0006955555555555556,
"loss": 2.8354079723358154,
"step": 13058
},
{
"epoch": 0.7915151515151515,
"grad_norm": 0.04702826961874962,
"learning_rate": 0.0006951515151515152,
"loss": 2.690727710723877,
"step": 13060
},
{
"epoch": 0.7916363636363636,
"grad_norm": 0.04663178324699402,
"learning_rate": 0.0006947474747474747,
"loss": 2.788396120071411,
"step": 13062
},
{
"epoch": 0.7917575757575758,
"grad_norm": 0.04557148367166519,
"learning_rate": 0.0006943434343434343,
"loss": 2.7726669311523438,
"step": 13064
},
{
"epoch": 0.7918787878787878,
"grad_norm": 0.0482805073261261,
"learning_rate": 0.000693939393939394,
"loss": 2.8176770210266113,
"step": 13066
},
{
"epoch": 0.792,
"grad_norm": 0.04674425721168518,
"learning_rate": 0.0006935353535353536,
"loss": 2.786742925643921,
"step": 13068
},
{
"epoch": 0.7921212121212121,
"grad_norm": 0.04674660041928291,
"learning_rate": 0.0006931313131313131,
"loss": 2.857231616973877,
"step": 13070
},
{
"epoch": 0.7922424242424242,
"grad_norm": 0.05158160254359245,
"learning_rate": 0.0006927272727272728,
"loss": 2.8024277687072754,
"step": 13072
},
{
"epoch": 0.7923636363636364,
"grad_norm": 0.05631386861205101,
"learning_rate": 0.0006923232323232324,
"loss": 2.7656500339508057,
"step": 13074
},
{
"epoch": 0.7924848484848485,
"grad_norm": 0.045749370008707047,
"learning_rate": 0.000691919191919192,
"loss": 2.819755792617798,
"step": 13076
},
{
"epoch": 0.7926060606060606,
"grad_norm": 0.04596208035945892,
"learning_rate": 0.0006915151515151515,
"loss": 2.8225109577178955,
"step": 13078
},
{
"epoch": 0.7927272727272727,
"grad_norm": 0.04624848812818527,
"learning_rate": 0.0006911111111111112,
"loss": 2.732177734375,
"step": 13080
},
{
"epoch": 0.7928484848484848,
"grad_norm": 0.047418806701898575,
"learning_rate": 0.0006907070707070708,
"loss": 2.793968915939331,
"step": 13082
},
{
"epoch": 0.792969696969697,
"grad_norm": 0.047092847526073456,
"learning_rate": 0.0006903030303030303,
"loss": 2.773519277572632,
"step": 13084
},
{
"epoch": 0.7930909090909091,
"grad_norm": 0.04617905616760254,
"learning_rate": 0.0006898989898989899,
"loss": 2.819995641708374,
"step": 13086
},
{
"epoch": 0.7932121212121213,
"grad_norm": 0.0469779446721077,
"learning_rate": 0.0006894949494949496,
"loss": 2.8254201412200928,
"step": 13088
},
{
"epoch": 0.7933333333333333,
"grad_norm": 0.06597650051116943,
"learning_rate": 0.0006890909090909091,
"loss": 2.7560386657714844,
"step": 13090
},
{
"epoch": 0.7934545454545454,
"grad_norm": 0.04744812101125717,
"learning_rate": 0.0006886868686868687,
"loss": 2.815908432006836,
"step": 13092
},
{
"epoch": 0.7935757575757576,
"grad_norm": 0.04491783678531647,
"learning_rate": 0.0006882828282828283,
"loss": 2.729473829269409,
"step": 13094
},
{
"epoch": 0.7936969696969697,
"grad_norm": 0.057662151753902435,
"learning_rate": 0.0006878787878787879,
"loss": 2.8109512329101562,
"step": 13096
},
{
"epoch": 0.7938181818181819,
"grad_norm": 0.05767659842967987,
"learning_rate": 0.0006874747474747474,
"loss": 2.8007898330688477,
"step": 13098
},
{
"epoch": 0.793939393939394,
"grad_norm": 0.04468579590320587,
"learning_rate": 0.0006870707070707071,
"loss": 2.799636125564575,
"step": 13100
},
{
"epoch": 0.794060606060606,
"grad_norm": 0.050391267985105515,
"learning_rate": 0.0006866666666666667,
"loss": 2.770794153213501,
"step": 13102
},
{
"epoch": 0.7941818181818182,
"grad_norm": 0.045197244733572006,
"learning_rate": 0.0006862626262626263,
"loss": 2.7634739875793457,
"step": 13104
},
{
"epoch": 0.7943030303030303,
"grad_norm": 0.04555029422044754,
"learning_rate": 0.000685858585858586,
"loss": 2.809053897857666,
"step": 13106
},
{
"epoch": 0.7944242424242425,
"grad_norm": 0.046583376824855804,
"learning_rate": 0.0006854545454545455,
"loss": 2.801868200302124,
"step": 13108
},
{
"epoch": 0.7945454545454546,
"grad_norm": 0.04825208708643913,
"learning_rate": 0.0006850505050505051,
"loss": 2.779825210571289,
"step": 13110
},
{
"epoch": 0.7946666666666666,
"grad_norm": 0.046051450073719025,
"learning_rate": 0.0006846464646464647,
"loss": 2.8168609142303467,
"step": 13112
},
{
"epoch": 0.7947878787878788,
"grad_norm": 0.04784306511282921,
"learning_rate": 0.0006842424242424243,
"loss": 2.7126646041870117,
"step": 13114
},
{
"epoch": 0.7949090909090909,
"grad_norm": 0.04582807421684265,
"learning_rate": 0.0006838383838383839,
"loss": 2.75772762298584,
"step": 13116
},
{
"epoch": 0.795030303030303,
"grad_norm": 0.04639701917767525,
"learning_rate": 0.0006834343434343435,
"loss": 2.79938006401062,
"step": 13118
},
{
"epoch": 0.7951515151515152,
"grad_norm": 0.047561246901750565,
"learning_rate": 0.000683030303030303,
"loss": 2.790745496749878,
"step": 13120
},
{
"epoch": 0.7952727272727272,
"grad_norm": 0.043681301176548004,
"learning_rate": 0.0006826262626262627,
"loss": 2.8214879035949707,
"step": 13122
},
{
"epoch": 0.7953939393939394,
"grad_norm": 0.04868364334106445,
"learning_rate": 0.0006822222222222222,
"loss": 2.8022592067718506,
"step": 13124
},
{
"epoch": 0.7955151515151515,
"grad_norm": 0.05575468763709068,
"learning_rate": 0.0006818181818181819,
"loss": 2.805053234100342,
"step": 13126
},
{
"epoch": 0.7956363636363636,
"grad_norm": 0.045058466494083405,
"learning_rate": 0.0006814141414141414,
"loss": 2.814293384552002,
"step": 13128
},
{
"epoch": 0.7957575757575758,
"grad_norm": 0.04958244040608406,
"learning_rate": 0.000681010101010101,
"loss": 2.8165037631988525,
"step": 13130
},
{
"epoch": 0.7958787878787879,
"grad_norm": 0.0470423586666584,
"learning_rate": 0.0006806060606060606,
"loss": 2.779778480529785,
"step": 13132
},
{
"epoch": 0.796,
"grad_norm": 0.04629089683294296,
"learning_rate": 0.0006802020202020203,
"loss": 2.8176965713500977,
"step": 13134
},
{
"epoch": 0.7961212121212121,
"grad_norm": 0.05449686944484711,
"learning_rate": 0.0006797979797979798,
"loss": 2.858067512512207,
"step": 13136
},
{
"epoch": 0.7962424242424242,
"grad_norm": 0.051575928926467896,
"learning_rate": 0.0006793939393939394,
"loss": 2.790954351425171,
"step": 13138
},
{
"epoch": 0.7963636363636364,
"grad_norm": 0.04488060995936394,
"learning_rate": 0.0006789898989898991,
"loss": 2.7668046951293945,
"step": 13140
},
{
"epoch": 0.7964848484848485,
"grad_norm": 0.053057342767715454,
"learning_rate": 0.0006785858585858587,
"loss": 2.8132989406585693,
"step": 13142
},
{
"epoch": 0.7966060606060607,
"grad_norm": 0.04662083461880684,
"learning_rate": 0.0006781818181818182,
"loss": 2.793585777282715,
"step": 13144
},
{
"epoch": 0.7967272727272727,
"grad_norm": 0.05255010351538658,
"learning_rate": 0.0006777777777777778,
"loss": 2.867196798324585,
"step": 13146
},
{
"epoch": 0.7968484848484848,
"grad_norm": 0.05707961320877075,
"learning_rate": 0.0006773737373737375,
"loss": 2.7568771839141846,
"step": 13148
},
{
"epoch": 0.796969696969697,
"grad_norm": 0.06333516538143158,
"learning_rate": 0.0006769696969696969,
"loss": 2.7879064083099365,
"step": 13150
},
{
"epoch": 0.7970909090909091,
"grad_norm": 0.060131702572107315,
"learning_rate": 0.0006765656565656566,
"loss": 2.7454588413238525,
"step": 13152
},
{
"epoch": 0.7972121212121213,
"grad_norm": 0.04718490689992905,
"learning_rate": 0.0006761616161616162,
"loss": 2.8699684143066406,
"step": 13154
},
{
"epoch": 0.7973333333333333,
"grad_norm": 0.04780982434749603,
"learning_rate": 0.0006757575757575758,
"loss": 2.8111112117767334,
"step": 13156
},
{
"epoch": 0.7974545454545454,
"grad_norm": 0.05249437689781189,
"learning_rate": 0.0006753535353535353,
"loss": 2.8280692100524902,
"step": 13158
},
{
"epoch": 0.7975757575757576,
"grad_norm": 0.052157800644636154,
"learning_rate": 0.000674949494949495,
"loss": 2.838442802429199,
"step": 13160
},
{
"epoch": 0.7976969696969697,
"grad_norm": 0.05081898719072342,
"learning_rate": 0.0006745454545454546,
"loss": 2.803860902786255,
"step": 13162
},
{
"epoch": 0.7978181818181819,
"grad_norm": 0.046473365277051926,
"learning_rate": 0.0006741414141414142,
"loss": 2.730713129043579,
"step": 13164
},
{
"epoch": 0.797939393939394,
"grad_norm": 0.0503675602376461,
"learning_rate": 0.0006737373737373737,
"loss": 2.8000776767730713,
"step": 13166
},
{
"epoch": 0.798060606060606,
"grad_norm": 0.047450922429561615,
"learning_rate": 0.0006733333333333334,
"loss": 2.81809401512146,
"step": 13168
},
{
"epoch": 0.7981818181818182,
"grad_norm": 0.04956042766571045,
"learning_rate": 0.000672929292929293,
"loss": 2.8566439151763916,
"step": 13170
},
{
"epoch": 0.7983030303030303,
"grad_norm": 0.04845551401376724,
"learning_rate": 0.0006725252525252525,
"loss": 2.835970401763916,
"step": 13172
},
{
"epoch": 0.7984242424242424,
"grad_norm": 0.05379324033856392,
"learning_rate": 0.0006721212121212122,
"loss": 2.786409854888916,
"step": 13174
},
{
"epoch": 0.7985454545454546,
"grad_norm": 0.05436549708247185,
"learning_rate": 0.0006717171717171718,
"loss": 2.796386480331421,
"step": 13176
},
{
"epoch": 0.7986666666666666,
"grad_norm": 0.046918828040361404,
"learning_rate": 0.0006713131313131314,
"loss": 2.7975411415100098,
"step": 13178
},
{
"epoch": 0.7987878787878788,
"grad_norm": 0.04895201325416565,
"learning_rate": 0.0006709090909090909,
"loss": 2.8221256732940674,
"step": 13180
},
{
"epoch": 0.7989090909090909,
"grad_norm": 0.048175204545259476,
"learning_rate": 0.0006705050505050506,
"loss": 2.7906148433685303,
"step": 13182
},
{
"epoch": 0.799030303030303,
"grad_norm": 0.046021219342947006,
"learning_rate": 0.0006701010101010101,
"loss": 2.7953836917877197,
"step": 13184
},
{
"epoch": 0.7991515151515152,
"grad_norm": 0.057057954370975494,
"learning_rate": 0.0006696969696969698,
"loss": 2.799076557159424,
"step": 13186
},
{
"epoch": 0.7992727272727272,
"grad_norm": 0.05215838551521301,
"learning_rate": 0.0006692929292929293,
"loss": 2.765577793121338,
"step": 13188
},
{
"epoch": 0.7993939393939394,
"grad_norm": 0.051222022622823715,
"learning_rate": 0.0006688888888888889,
"loss": 2.825721502304077,
"step": 13190
},
{
"epoch": 0.7995151515151515,
"grad_norm": 0.0453946590423584,
"learning_rate": 0.0006684848484848485,
"loss": 2.825640916824341,
"step": 13192
},
{
"epoch": 0.7996363636363636,
"grad_norm": 0.04847656935453415,
"learning_rate": 0.0006680808080808081,
"loss": 2.814901113510132,
"step": 13194
},
{
"epoch": 0.7997575757575758,
"grad_norm": 0.047098420560359955,
"learning_rate": 0.0006676767676767677,
"loss": 2.8131229877471924,
"step": 13196
},
{
"epoch": 0.7998787878787879,
"grad_norm": 0.04574795812368393,
"learning_rate": 0.0006672727272727273,
"loss": 2.8427302837371826,
"step": 13198
},
{
"epoch": 0.8,
"grad_norm": 0.05004130303859711,
"learning_rate": 0.0006668686868686869,
"loss": 2.796891212463379,
"step": 13200
},
{
"epoch": 0.8,
"eval_loss": 2.0081608295440674,
"eval_runtime": 148.9358,
"eval_samples_per_second": 1.679,
"eval_steps_per_second": 0.087,
"step": 13200
},
{
"epoch": 0.8001212121212121,
"grad_norm": 0.04972343519330025,
"learning_rate": 0.0006664646464646465,
"loss": 2.842597723007202,
"step": 13202
},
{
"epoch": 0.8002424242424242,
"grad_norm": 0.053594693541526794,
"learning_rate": 0.000666060606060606,
"loss": 2.834697961807251,
"step": 13204
},
{
"epoch": 0.8003636363636364,
"grad_norm": 0.06678860634565353,
"learning_rate": 0.0006656565656565657,
"loss": 2.8270435333251953,
"step": 13206
},
{
"epoch": 0.8004848484848485,
"grad_norm": 0.054577652364969254,
"learning_rate": 0.0006652525252525254,
"loss": 2.84419584274292,
"step": 13208
},
{
"epoch": 0.8006060606060607,
"grad_norm": 0.0529949776828289,
"learning_rate": 0.0006648484848484849,
"loss": 2.8259127140045166,
"step": 13210
},
{
"epoch": 0.8007272727272727,
"grad_norm": 0.060698483139276505,
"learning_rate": 0.0006644444444444445,
"loss": 2.8131980895996094,
"step": 13212
},
{
"epoch": 0.8008484848484848,
"grad_norm": 0.0827564224600792,
"learning_rate": 0.0006640404040404041,
"loss": 2.815143585205078,
"step": 13214
},
{
"epoch": 0.800969696969697,
"grad_norm": 0.08452309668064117,
"learning_rate": 0.0006636363636363638,
"loss": 2.7902421951293945,
"step": 13216
},
{
"epoch": 0.8010909090909091,
"grad_norm": 0.06322544813156128,
"learning_rate": 0.0006632323232323232,
"loss": 2.7816340923309326,
"step": 13218
},
{
"epoch": 0.8012121212121213,
"grad_norm": 0.04708153381943703,
"learning_rate": 0.0006628282828282829,
"loss": 2.815849781036377,
"step": 13220
},
{
"epoch": 0.8013333333333333,
"grad_norm": 0.05056323483586311,
"learning_rate": 0.0006624242424242425,
"loss": 2.78957462310791,
"step": 13222
},
{
"epoch": 0.8014545454545454,
"grad_norm": 0.052627723664045334,
"learning_rate": 0.000662020202020202,
"loss": 2.741135358810425,
"step": 13224
},
{
"epoch": 0.8015757575757576,
"grad_norm": 0.06479090452194214,
"learning_rate": 0.0006616161616161616,
"loss": 2.785752296447754,
"step": 13226
},
{
"epoch": 0.8016969696969697,
"grad_norm": 0.052118171006441116,
"learning_rate": 0.0006612121212121213,
"loss": 2.7871615886688232,
"step": 13228
},
{
"epoch": 0.8018181818181818,
"grad_norm": 0.0590946190059185,
"learning_rate": 0.0006608080808080809,
"loss": 2.829843282699585,
"step": 13230
},
{
"epoch": 0.801939393939394,
"grad_norm": 0.05718967318534851,
"learning_rate": 0.0006604040404040404,
"loss": 2.7966768741607666,
"step": 13232
},
{
"epoch": 0.802060606060606,
"grad_norm": 0.0460471510887146,
"learning_rate": 0.00066,
"loss": 2.8125481605529785,
"step": 13234
},
{
"epoch": 0.8021818181818182,
"grad_norm": 0.04622102901339531,
"learning_rate": 0.0006595959595959597,
"loss": 2.813819646835327,
"step": 13236
},
{
"epoch": 0.8023030303030303,
"grad_norm": 0.0448002927005291,
"learning_rate": 0.0006591919191919191,
"loss": 2.8254404067993164,
"step": 13238
},
{
"epoch": 0.8024242424242424,
"grad_norm": 0.04853089898824692,
"learning_rate": 0.0006587878787878788,
"loss": 2.8386130332946777,
"step": 13240
},
{
"epoch": 0.8025454545454546,
"grad_norm": 0.045443255454301834,
"learning_rate": 0.0006583838383838385,
"loss": 2.796541452407837,
"step": 13242
},
{
"epoch": 0.8026666666666666,
"grad_norm": 0.04561065882444382,
"learning_rate": 0.000657979797979798,
"loss": 2.7967734336853027,
"step": 13244
},
{
"epoch": 0.8027878787878788,
"grad_norm": 0.04827597364783287,
"learning_rate": 0.0006575757575757576,
"loss": 2.827510356903076,
"step": 13246
},
{
"epoch": 0.8029090909090909,
"grad_norm": 0.05161137878894806,
"learning_rate": 0.0006571717171717172,
"loss": 2.8710222244262695,
"step": 13248
},
{
"epoch": 0.803030303030303,
"grad_norm": 0.05577323958277702,
"learning_rate": 0.0006567676767676768,
"loss": 2.8459384441375732,
"step": 13250
},
{
"epoch": 0.8031515151515152,
"grad_norm": 0.05504683777689934,
"learning_rate": 0.0006563636363636364,
"loss": 2.8546500205993652,
"step": 13252
},
{
"epoch": 0.8032727272727272,
"grad_norm": 0.05809084326028824,
"learning_rate": 0.000655959595959596,
"loss": 2.8477871417999268,
"step": 13254
},
{
"epoch": 0.8033939393939394,
"grad_norm": 0.05423776060342789,
"learning_rate": 0.0006555555555555556,
"loss": 2.779170513153076,
"step": 13256
},
{
"epoch": 0.8035151515151515,
"grad_norm": 0.04468144103884697,
"learning_rate": 0.0006551515151515152,
"loss": 2.7995781898498535,
"step": 13258
},
{
"epoch": 0.8036363636363636,
"grad_norm": 0.05302496999502182,
"learning_rate": 0.0006547474747474747,
"loss": 2.832709789276123,
"step": 13260
},
{
"epoch": 0.8037575757575758,
"grad_norm": 0.047064587473869324,
"learning_rate": 0.0006543434343434344,
"loss": 2.800546884536743,
"step": 13262
},
{
"epoch": 0.8038787878787879,
"grad_norm": 0.04695127531886101,
"learning_rate": 0.000653939393939394,
"loss": 2.760993719100952,
"step": 13264
},
{
"epoch": 0.804,
"grad_norm": 0.04819091036915779,
"learning_rate": 0.0006535353535353536,
"loss": 2.8231041431427,
"step": 13266
},
{
"epoch": 0.8041212121212121,
"grad_norm": 0.04612186178565025,
"learning_rate": 0.0006531313131313131,
"loss": 2.773386001586914,
"step": 13268
},
{
"epoch": 0.8042424242424242,
"grad_norm": 0.049235135316848755,
"learning_rate": 0.0006527272727272728,
"loss": 2.7915542125701904,
"step": 13270
},
{
"epoch": 0.8043636363636364,
"grad_norm": 0.048958562314510345,
"learning_rate": 0.0006523232323232323,
"loss": 2.7697601318359375,
"step": 13272
},
{
"epoch": 0.8044848484848485,
"grad_norm": 0.04942447692155838,
"learning_rate": 0.000651919191919192,
"loss": 2.815126419067383,
"step": 13274
},
{
"epoch": 0.8046060606060607,
"grad_norm": 0.045611534267663956,
"learning_rate": 0.0006515151515151516,
"loss": 2.828047037124634,
"step": 13276
},
{
"epoch": 0.8047272727272727,
"grad_norm": 0.04904593527317047,
"learning_rate": 0.0006511111111111111,
"loss": 2.8176112174987793,
"step": 13278
},
{
"epoch": 0.8048484848484848,
"grad_norm": 0.05835844203829765,
"learning_rate": 0.0006507070707070708,
"loss": 2.7873711585998535,
"step": 13280
},
{
"epoch": 0.804969696969697,
"grad_norm": 0.05890290439128876,
"learning_rate": 0.0006503030303030303,
"loss": 2.8325226306915283,
"step": 13282
},
{
"epoch": 0.8050909090909091,
"grad_norm": 0.05861283466219902,
"learning_rate": 0.0006498989898989899,
"loss": 2.8114562034606934,
"step": 13284
},
{
"epoch": 0.8052121212121212,
"grad_norm": 0.0690818578004837,
"learning_rate": 0.0006494949494949495,
"loss": 2.8586788177490234,
"step": 13286
},
{
"epoch": 0.8053333333333333,
"grad_norm": 0.06437402218580246,
"learning_rate": 0.0006490909090909092,
"loss": 2.878419876098633,
"step": 13288
},
{
"epoch": 0.8054545454545454,
"grad_norm": 0.062086500227451324,
"learning_rate": 0.0006486868686868687,
"loss": 2.848719835281372,
"step": 13290
},
{
"epoch": 0.8055757575757576,
"grad_norm": 0.05085587874054909,
"learning_rate": 0.0006482828282828283,
"loss": 2.826549530029297,
"step": 13292
},
{
"epoch": 0.8056969696969697,
"grad_norm": 0.0458126999437809,
"learning_rate": 0.0006478787878787879,
"loss": 2.7532296180725098,
"step": 13294
},
{
"epoch": 0.8058181818181818,
"grad_norm": 0.04713047295808792,
"learning_rate": 0.0006474747474747476,
"loss": 2.8490493297576904,
"step": 13296
},
{
"epoch": 0.805939393939394,
"grad_norm": 0.04516933113336563,
"learning_rate": 0.000647070707070707,
"loss": 2.8068130016326904,
"step": 13298
},
{
"epoch": 0.806060606060606,
"grad_norm": 0.04681646078824997,
"learning_rate": 0.0006466666666666667,
"loss": 2.894982099533081,
"step": 13300
},
{
"epoch": 0.8061818181818182,
"grad_norm": 0.05127902701497078,
"learning_rate": 0.0006462626262626263,
"loss": 2.8475232124328613,
"step": 13302
},
{
"epoch": 0.8063030303030303,
"grad_norm": 0.07913738489151001,
"learning_rate": 0.0006458585858585858,
"loss": 2.853806495666504,
"step": 13304
},
{
"epoch": 0.8064242424242424,
"grad_norm": 0.049732889980077744,
"learning_rate": 0.0006454545454545454,
"loss": 2.848273754119873,
"step": 13306
},
{
"epoch": 0.8065454545454546,
"grad_norm": 0.05056318640708923,
"learning_rate": 0.0006450505050505051,
"loss": 2.8529787063598633,
"step": 13308
},
{
"epoch": 0.8066666666666666,
"grad_norm": 0.04908832907676697,
"learning_rate": 0.0006446464646464648,
"loss": 2.837106943130493,
"step": 13310
},
{
"epoch": 0.8067878787878788,
"grad_norm": 0.05091864615678787,
"learning_rate": 0.0006442424242424242,
"loss": 2.8518669605255127,
"step": 13312
},
{
"epoch": 0.8069090909090909,
"grad_norm": 0.04552064463496208,
"learning_rate": 0.0006438383838383839,
"loss": 2.8565785884857178,
"step": 13314
},
{
"epoch": 0.807030303030303,
"grad_norm": 0.04663857817649841,
"learning_rate": 0.0006434343434343435,
"loss": 2.9210078716278076,
"step": 13316
},
{
"epoch": 0.8071515151515152,
"grad_norm": 0.04617827758193016,
"learning_rate": 0.0006430303030303031,
"loss": 2.8313655853271484,
"step": 13318
},
{
"epoch": 0.8072727272727273,
"grad_norm": 0.047103896737098694,
"learning_rate": 0.0006426262626262626,
"loss": 2.851684808731079,
"step": 13320
},
{
"epoch": 0.8073939393939394,
"grad_norm": 0.0647672787308693,
"learning_rate": 0.0006422222222222223,
"loss": 2.8696982860565186,
"step": 13322
},
{
"epoch": 0.8075151515151515,
"grad_norm": 0.06326740980148315,
"learning_rate": 0.0006418181818181819,
"loss": 2.8032450675964355,
"step": 13324
},
{
"epoch": 0.8076363636363636,
"grad_norm": 0.06980804353952408,
"learning_rate": 0.0006414141414141414,
"loss": 2.841914415359497,
"step": 13326
},
{
"epoch": 0.8077575757575758,
"grad_norm": 0.06639118492603302,
"learning_rate": 0.000641010101010101,
"loss": 2.832143545150757,
"step": 13328
},
{
"epoch": 0.8078787878787879,
"grad_norm": 0.04689180478453636,
"learning_rate": 0.0006406060606060607,
"loss": 2.910076379776001,
"step": 13330
},
{
"epoch": 0.808,
"grad_norm": 0.04802010953426361,
"learning_rate": 0.0006402020202020202,
"loss": 2.8614161014556885,
"step": 13332
},
{
"epoch": 0.8081212121212121,
"grad_norm": 0.04749307036399841,
"learning_rate": 0.0006397979797979798,
"loss": 2.8640153408050537,
"step": 13334
},
{
"epoch": 0.8082424242424242,
"grad_norm": 0.048458635807037354,
"learning_rate": 0.0006393939393939394,
"loss": 2.8898651599884033,
"step": 13336
},
{
"epoch": 0.8083636363636364,
"grad_norm": 0.054579444229602814,
"learning_rate": 0.000638989898989899,
"loss": 2.8555750846862793,
"step": 13338
},
{
"epoch": 0.8084848484848485,
"grad_norm": 0.04965498298406601,
"learning_rate": 0.0006385858585858586,
"loss": 2.9503564834594727,
"step": 13340
},
{
"epoch": 0.8086060606060607,
"grad_norm": 0.05925993621349335,
"learning_rate": 0.0006381818181818182,
"loss": 2.9134037494659424,
"step": 13342
},
{
"epoch": 0.8087272727272727,
"grad_norm": 0.04877155274152756,
"learning_rate": 0.0006377777777777778,
"loss": 2.895134210586548,
"step": 13344
},
{
"epoch": 0.8088484848484848,
"grad_norm": 0.061736468225717545,
"learning_rate": 0.0006373737373737374,
"loss": 2.9100494384765625,
"step": 13346
},
{
"epoch": 0.808969696969697,
"grad_norm": 0.05526930093765259,
"learning_rate": 0.000636969696969697,
"loss": 2.9315638542175293,
"step": 13348
},
{
"epoch": 0.8090909090909091,
"grad_norm": 0.05464410036802292,
"learning_rate": 0.0006365656565656566,
"loss": 2.8236544132232666,
"step": 13350
},
{
"epoch": 0.8092121212121212,
"grad_norm": 0.05092160776257515,
"learning_rate": 0.0006361616161616162,
"loss": 2.8571701049804688,
"step": 13352
},
{
"epoch": 0.8093333333333333,
"grad_norm": 0.04927399381995201,
"learning_rate": 0.0006357575757575758,
"loss": 2.926161289215088,
"step": 13354
},
{
"epoch": 0.8094545454545454,
"grad_norm": 0.053003761917352676,
"learning_rate": 0.0006353535353535354,
"loss": 2.8849143981933594,
"step": 13356
},
{
"epoch": 0.8095757575757576,
"grad_norm": 0.05122891440987587,
"learning_rate": 0.000634949494949495,
"loss": 2.8966636657714844,
"step": 13358
},
{
"epoch": 0.8096969696969697,
"grad_norm": 0.04717274010181427,
"learning_rate": 0.0006345454545454546,
"loss": 2.894322633743286,
"step": 13360
},
{
"epoch": 0.8098181818181818,
"grad_norm": 0.047096967697143555,
"learning_rate": 0.0006341414141414142,
"loss": 2.9066710472106934,
"step": 13362
},
{
"epoch": 0.809939393939394,
"grad_norm": 0.04685286805033684,
"learning_rate": 0.0006337373737373738,
"loss": 2.8879613876342773,
"step": 13364
},
{
"epoch": 0.810060606060606,
"grad_norm": 0.06396210193634033,
"learning_rate": 0.0006333333333333333,
"loss": 2.9027326107025146,
"step": 13366
},
{
"epoch": 0.8101818181818182,
"grad_norm": 0.06790357083082199,
"learning_rate": 0.000632929292929293,
"loss": 2.869399070739746,
"step": 13368
},
{
"epoch": 0.8103030303030303,
"grad_norm": 0.07943769544363022,
"learning_rate": 0.0006325252525252525,
"loss": 2.8806538581848145,
"step": 13370
},
{
"epoch": 0.8104242424242424,
"grad_norm": 0.07490754127502441,
"learning_rate": 0.0006321212121212121,
"loss": 2.9181466102600098,
"step": 13372
},
{
"epoch": 0.8105454545454546,
"grad_norm": 0.04938077926635742,
"learning_rate": 0.0006317171717171717,
"loss": 2.9328792095184326,
"step": 13374
},
{
"epoch": 0.8106666666666666,
"grad_norm": 0.056151729077100754,
"learning_rate": 0.0006313131313131314,
"loss": 2.9134767055511475,
"step": 13376
},
{
"epoch": 0.8107878787878788,
"grad_norm": 0.04637347161769867,
"learning_rate": 0.0006309090909090909,
"loss": 2.8496155738830566,
"step": 13378
},
{
"epoch": 0.8109090909090909,
"grad_norm": 0.0474344901740551,
"learning_rate": 0.0006305050505050505,
"loss": 2.956608295440674,
"step": 13380
},
{
"epoch": 0.811030303030303,
"grad_norm": 0.04615860432386398,
"learning_rate": 0.0006301010101010102,
"loss": 2.8911640644073486,
"step": 13382
},
{
"epoch": 0.8111515151515152,
"grad_norm": 0.04846186563372612,
"learning_rate": 0.0006296969696969698,
"loss": 2.9187045097351074,
"step": 13384
},
{
"epoch": 0.8112727272727273,
"grad_norm": 0.057555533945560455,
"learning_rate": 0.0006292929292929293,
"loss": 2.9486401081085205,
"step": 13386
},
{
"epoch": 0.8113939393939394,
"grad_norm": 0.04940567910671234,
"learning_rate": 0.0006288888888888889,
"loss": 2.9071247577667236,
"step": 13388
},
{
"epoch": 0.8115151515151515,
"grad_norm": 0.05859570577740669,
"learning_rate": 0.0006284848484848486,
"loss": 2.8872697353363037,
"step": 13390
},
{
"epoch": 0.8116363636363636,
"grad_norm": 0.058793727308511734,
"learning_rate": 0.000628080808080808,
"loss": 2.9299094676971436,
"step": 13392
},
{
"epoch": 0.8117575757575758,
"grad_norm": 0.05904874578118324,
"learning_rate": 0.0006276767676767677,
"loss": 2.958874225616455,
"step": 13394
},
{
"epoch": 0.8118787878787879,
"grad_norm": 0.04633975028991699,
"learning_rate": 0.0006272727272727273,
"loss": 2.950331449508667,
"step": 13396
},
{
"epoch": 0.812,
"grad_norm": 0.04696591570973396,
"learning_rate": 0.0006268686868686869,
"loss": 2.9658002853393555,
"step": 13398
},
{
"epoch": 0.8121212121212121,
"grad_norm": 0.06830781698226929,
"learning_rate": 0.0006264646464646464,
"loss": 2.930807113647461,
"step": 13400
},
{
"epoch": 0.8122424242424242,
"grad_norm": 0.08038277924060822,
"learning_rate": 0.0006260606060606061,
"loss": 2.921006917953491,
"step": 13402
},
{
"epoch": 0.8123636363636364,
"grad_norm": 0.06898420304059982,
"learning_rate": 0.0006256565656565657,
"loss": 2.938656806945801,
"step": 13404
},
{
"epoch": 0.8124848484848485,
"grad_norm": 0.06661428511142731,
"learning_rate": 0.0006252525252525253,
"loss": 2.8790817260742188,
"step": 13406
},
{
"epoch": 0.8126060606060606,
"grad_norm": 0.05617070570588112,
"learning_rate": 0.0006248484848484848,
"loss": 2.905383586883545,
"step": 13408
},
{
"epoch": 0.8127272727272727,
"grad_norm": 0.049249645322561264,
"learning_rate": 0.0006244444444444445,
"loss": 2.944227457046509,
"step": 13410
},
{
"epoch": 0.8128484848484848,
"grad_norm": 0.04623255133628845,
"learning_rate": 0.0006240404040404041,
"loss": 2.870915651321411,
"step": 13412
},
{
"epoch": 0.812969696969697,
"grad_norm": 0.047330521047115326,
"learning_rate": 0.0006236363636363636,
"loss": 2.9853007793426514,
"step": 13414
},
{
"epoch": 0.8130909090909091,
"grad_norm": 0.04683776572346687,
"learning_rate": 0.0006232323232323233,
"loss": 2.990028142929077,
"step": 13416
},
{
"epoch": 0.8132121212121212,
"grad_norm": 0.046488143503665924,
"learning_rate": 0.0006228282828282829,
"loss": 2.955552577972412,
"step": 13418
},
{
"epoch": 0.8133333333333334,
"grad_norm": 0.04680723696947098,
"learning_rate": 0.0006224242424242425,
"loss": 2.9346518516540527,
"step": 13420
},
{
"epoch": 0.8134545454545454,
"grad_norm": 0.04793750122189522,
"learning_rate": 0.000622020202020202,
"loss": 2.991544008255005,
"step": 13422
},
{
"epoch": 0.8135757575757576,
"grad_norm": 0.055474959313869476,
"learning_rate": 0.0006216161616161617,
"loss": 2.934875965118408,
"step": 13424
},
{
"epoch": 0.8136969696969697,
"grad_norm": 0.05480336397886276,
"learning_rate": 0.0006212121212121212,
"loss": 2.989673614501953,
"step": 13426
},
{
"epoch": 0.8138181818181818,
"grad_norm": 0.0472421832382679,
"learning_rate": 0.0006208080808080809,
"loss": 2.9330434799194336,
"step": 13428
},
{
"epoch": 0.813939393939394,
"grad_norm": 0.04560041427612305,
"learning_rate": 0.0006204040404040404,
"loss": 2.9712090492248535,
"step": 13430
},
{
"epoch": 0.814060606060606,
"grad_norm": 0.04755638167262077,
"learning_rate": 0.00062,
"loss": 2.901801109313965,
"step": 13432
},
{
"epoch": 0.8141818181818182,
"grad_norm": 0.047088492661714554,
"learning_rate": 0.0006195959595959596,
"loss": 2.961364984512329,
"step": 13434
},
{
"epoch": 0.8143030303030303,
"grad_norm": 0.05138353258371353,
"learning_rate": 0.0006191919191919192,
"loss": 2.9592652320861816,
"step": 13436
},
{
"epoch": 0.8144242424242424,
"grad_norm": 0.053339216858148575,
"learning_rate": 0.0006187878787878788,
"loss": 2.9192113876342773,
"step": 13438
},
{
"epoch": 0.8145454545454546,
"grad_norm": 0.04640669748187065,
"learning_rate": 0.0006183838383838384,
"loss": 2.914261817932129,
"step": 13440
},
{
"epoch": 0.8146666666666667,
"grad_norm": 0.05631566792726517,
"learning_rate": 0.000617979797979798,
"loss": 2.988940954208374,
"step": 13442
},
{
"epoch": 0.8147878787878788,
"grad_norm": 0.06299835443496704,
"learning_rate": 0.0006175757575757576,
"loss": 2.950742244720459,
"step": 13444
},
{
"epoch": 0.8149090909090909,
"grad_norm": 0.05129518732428551,
"learning_rate": 0.0006171717171717172,
"loss": 2.955496072769165,
"step": 13446
},
{
"epoch": 0.815030303030303,
"grad_norm": 0.04723842069506645,
"learning_rate": 0.0006167676767676768,
"loss": 3.0095932483673096,
"step": 13448
},
{
"epoch": 0.8151515151515152,
"grad_norm": 0.05296776443719864,
"learning_rate": 0.0006163636363636365,
"loss": 3.0234873294830322,
"step": 13450
},
{
"epoch": 0.8152727272727273,
"grad_norm": 0.047097098082304,
"learning_rate": 0.000615959595959596,
"loss": 2.9220590591430664,
"step": 13452
},
{
"epoch": 0.8153939393939394,
"grad_norm": 0.04692884534597397,
"learning_rate": 0.0006155555555555556,
"loss": 2.969944477081299,
"step": 13454
},
{
"epoch": 0.8155151515151515,
"grad_norm": 0.047690149396657944,
"learning_rate": 0.0006151515151515152,
"loss": 2.9248740673065186,
"step": 13456
},
{
"epoch": 0.8156363636363636,
"grad_norm": 0.048905208706855774,
"learning_rate": 0.0006147474747474749,
"loss": 2.9581215381622314,
"step": 13458
},
{
"epoch": 0.8157575757575758,
"grad_norm": 0.047247640788555145,
"learning_rate": 0.0006143434343434343,
"loss": 2.9666407108306885,
"step": 13460
},
{
"epoch": 0.8158787878787879,
"grad_norm": 0.04641908034682274,
"learning_rate": 0.000613939393939394,
"loss": 2.9526333808898926,
"step": 13462
},
{
"epoch": 0.816,
"grad_norm": 0.05195852741599083,
"learning_rate": 0.0006135353535353536,
"loss": 2.967937707901001,
"step": 13464
},
{
"epoch": 0.8161212121212121,
"grad_norm": 0.05330153927206993,
"learning_rate": 0.0006131313131313131,
"loss": 2.9285640716552734,
"step": 13466
},
{
"epoch": 0.8162424242424242,
"grad_norm": 0.04722687229514122,
"learning_rate": 0.0006127272727272727,
"loss": 2.9428138732910156,
"step": 13468
},
{
"epoch": 0.8163636363636364,
"grad_norm": 0.05013379082083702,
"learning_rate": 0.0006123232323232324,
"loss": 2.9819719791412354,
"step": 13470
},
{
"epoch": 0.8164848484848485,
"grad_norm": 0.055104076862335205,
"learning_rate": 0.000611919191919192,
"loss": 2.93699049949646,
"step": 13472
},
{
"epoch": 0.8166060606060606,
"grad_norm": 0.04959428682923317,
"learning_rate": 0.0006115151515151515,
"loss": 3.0081541538238525,
"step": 13474
},
{
"epoch": 0.8167272727272727,
"grad_norm": 0.04693244770169258,
"learning_rate": 0.0006111111111111111,
"loss": 2.9718401432037354,
"step": 13476
},
{
"epoch": 0.8168484848484848,
"grad_norm": 0.04801761731505394,
"learning_rate": 0.0006107070707070708,
"loss": 2.981546401977539,
"step": 13478
},
{
"epoch": 0.816969696969697,
"grad_norm": 0.04796745255589485,
"learning_rate": 0.0006103030303030303,
"loss": 2.9974210262298584,
"step": 13480
},
{
"epoch": 0.8170909090909091,
"grad_norm": 0.0463077649474144,
"learning_rate": 0.0006098989898989899,
"loss": 2.951465606689453,
"step": 13482
},
{
"epoch": 0.8172121212121212,
"grad_norm": 0.04606413468718529,
"learning_rate": 0.0006094949494949496,
"loss": 2.9737911224365234,
"step": 13484
},
{
"epoch": 0.8173333333333334,
"grad_norm": 0.045160915702581406,
"learning_rate": 0.0006090909090909091,
"loss": 2.985938549041748,
"step": 13486
},
{
"epoch": 0.8174545454545454,
"grad_norm": 0.04978189989924431,
"learning_rate": 0.0006086868686868687,
"loss": 2.9661450386047363,
"step": 13488
},
{
"epoch": 0.8175757575757576,
"grad_norm": 0.050027038902044296,
"learning_rate": 0.0006082828282828283,
"loss": 3.0324676036834717,
"step": 13490
},
{
"epoch": 0.8176969696969697,
"grad_norm": 0.05199417471885681,
"learning_rate": 0.0006078787878787879,
"loss": 2.9807841777801514,
"step": 13492
},
{
"epoch": 0.8178181818181818,
"grad_norm": 0.04673049971461296,
"learning_rate": 0.0006074747474747475,
"loss": 3.0134613513946533,
"step": 13494
},
{
"epoch": 0.817939393939394,
"grad_norm": 0.05560171231627464,
"learning_rate": 0.0006070707070707071,
"loss": 2.940178632736206,
"step": 13496
},
{
"epoch": 0.818060606060606,
"grad_norm": 0.06095551699399948,
"learning_rate": 0.0006066666666666667,
"loss": 3.0344314575195312,
"step": 13498
},
{
"epoch": 0.8181818181818182,
"grad_norm": 0.05998795107007027,
"learning_rate": 0.0006062626262626263,
"loss": 2.9435455799102783,
"step": 13500
},
{
"epoch": 0.8183030303030303,
"grad_norm": 0.04748881980776787,
"learning_rate": 0.0006058585858585858,
"loss": 3.024442672729492,
"step": 13502
},
{
"epoch": 0.8184242424242424,
"grad_norm": 0.04832036793231964,
"learning_rate": 0.0006054545454545455,
"loss": 2.975292682647705,
"step": 13504
},
{
"epoch": 0.8185454545454546,
"grad_norm": 0.04811418429017067,
"learning_rate": 0.0006050505050505051,
"loss": 3.000211715698242,
"step": 13506
},
{
"epoch": 0.8186666666666667,
"grad_norm": 0.04817682132124901,
"learning_rate": 0.0006046464646464647,
"loss": 2.9785091876983643,
"step": 13508
},
{
"epoch": 0.8187878787878788,
"grad_norm": 0.05719006806612015,
"learning_rate": 0.0006042424242424242,
"loss": 3.0197064876556396,
"step": 13510
},
{
"epoch": 0.8189090909090909,
"grad_norm": 0.055795006453990936,
"learning_rate": 0.0006038383838383839,
"loss": 3.022819757461548,
"step": 13512
},
{
"epoch": 0.819030303030303,
"grad_norm": 0.0750897154211998,
"learning_rate": 0.0006034343434343435,
"loss": 3.0694408416748047,
"step": 13514
},
{
"epoch": 0.8191515151515152,
"grad_norm": 0.059763144701719284,
"learning_rate": 0.0006030303030303031,
"loss": 2.9911508560180664,
"step": 13516
},
{
"epoch": 0.8192727272727273,
"grad_norm": 0.0555020272731781,
"learning_rate": 0.0006026262626262627,
"loss": 2.993999719619751,
"step": 13518
},
{
"epoch": 0.8193939393939393,
"grad_norm": 0.06366230547428131,
"learning_rate": 0.0006022222222222222,
"loss": 3.009535312652588,
"step": 13520
},
{
"epoch": 0.8195151515151515,
"grad_norm": 0.05488186702132225,
"learning_rate": 0.0006018181818181819,
"loss": 2.954522132873535,
"step": 13522
},
{
"epoch": 0.8196363636363636,
"grad_norm": 0.04696402698755264,
"learning_rate": 0.0006014141414141414,
"loss": 3.010896682739258,
"step": 13524
},
{
"epoch": 0.8197575757575758,
"grad_norm": 0.057019468396902084,
"learning_rate": 0.000601010101010101,
"loss": 2.9565823078155518,
"step": 13526
},
{
"epoch": 0.8198787878787879,
"grad_norm": 0.05920756608247757,
"learning_rate": 0.0006006060606060606,
"loss": 3.000840187072754,
"step": 13528
},
{
"epoch": 0.82,
"grad_norm": 0.04865526407957077,
"learning_rate": 0.0006002020202020203,
"loss": 2.9985291957855225,
"step": 13530
},
{
"epoch": 0.8201212121212121,
"grad_norm": 0.04949101805686951,
"learning_rate": 0.0005997979797979798,
"loss": 3.027608633041382,
"step": 13532
},
{
"epoch": 0.8202424242424242,
"grad_norm": 0.0544966384768486,
"learning_rate": 0.0005993939393939394,
"loss": 3.022199869155884,
"step": 13534
},
{
"epoch": 0.8203636363636364,
"grad_norm": 0.04840617999434471,
"learning_rate": 0.000598989898989899,
"loss": 3.030252695083618,
"step": 13536
},
{
"epoch": 0.8204848484848485,
"grad_norm": 0.04821043834090233,
"learning_rate": 0.0005985858585858587,
"loss": 3.0278420448303223,
"step": 13538
},
{
"epoch": 0.8206060606060606,
"grad_norm": 0.05129171162843704,
"learning_rate": 0.0005981818181818181,
"loss": 2.9838619232177734,
"step": 13540
},
{
"epoch": 0.8207272727272727,
"grad_norm": 0.04930254444479942,
"learning_rate": 0.0005977777777777778,
"loss": 3.0106685161590576,
"step": 13542
},
{
"epoch": 0.8208484848484848,
"grad_norm": 0.05797601118683815,
"learning_rate": 0.0005973737373737374,
"loss": 2.9912991523742676,
"step": 13544
},
{
"epoch": 0.820969696969697,
"grad_norm": 0.06508145481348038,
"learning_rate": 0.000596969696969697,
"loss": 3.040030002593994,
"step": 13546
},
{
"epoch": 0.8210909090909091,
"grad_norm": 0.04764100909233093,
"learning_rate": 0.0005965656565656565,
"loss": 3.0220181941986084,
"step": 13548
},
{
"epoch": 0.8212121212121212,
"grad_norm": 0.054651957005262375,
"learning_rate": 0.0005961616161616162,
"loss": 3.0701799392700195,
"step": 13550
},
{
"epoch": 0.8213333333333334,
"grad_norm": 0.049908265471458435,
"learning_rate": 0.0005957575757575759,
"loss": 3.005467176437378,
"step": 13552
},
{
"epoch": 0.8214545454545454,
"grad_norm": 0.06047055870294571,
"learning_rate": 0.0005953535353535353,
"loss": 3.0351979732513428,
"step": 13554
},
{
"epoch": 0.8215757575757576,
"grad_norm": 0.06231583282351494,
"learning_rate": 0.000594949494949495,
"loss": 3.0363848209381104,
"step": 13556
},
{
"epoch": 0.8216969696969697,
"grad_norm": 0.058464959263801575,
"learning_rate": 0.0005945454545454546,
"loss": 3.0360662937164307,
"step": 13558
},
{
"epoch": 0.8218181818181818,
"grad_norm": 0.05971399322152138,
"learning_rate": 0.0005941414141414142,
"loss": 3.05120849609375,
"step": 13560
},
{
"epoch": 0.821939393939394,
"grad_norm": 0.0498645044863224,
"learning_rate": 0.0005937373737373737,
"loss": 3.0323102474212646,
"step": 13562
},
{
"epoch": 0.822060606060606,
"grad_norm": 0.04821431264281273,
"learning_rate": 0.0005933333333333334,
"loss": 3.019207239151001,
"step": 13564
},
{
"epoch": 0.8221818181818182,
"grad_norm": 0.048460543155670166,
"learning_rate": 0.000592929292929293,
"loss": 3.047628402709961,
"step": 13566
},
{
"epoch": 0.8223030303030303,
"grad_norm": 0.057887885719537735,
"learning_rate": 0.0005925252525252525,
"loss": 3.0030999183654785,
"step": 13568
},
{
"epoch": 0.8224242424242424,
"grad_norm": 0.05793848633766174,
"learning_rate": 0.0005921212121212121,
"loss": 3.0625665187835693,
"step": 13570
},
{
"epoch": 0.8225454545454546,
"grad_norm": 0.060549888759851456,
"learning_rate": 0.0005917171717171718,
"loss": 3.058459997177124,
"step": 13572
},
{
"epoch": 0.8226666666666667,
"grad_norm": 0.06365973502397537,
"learning_rate": 0.0005913131313131313,
"loss": 3.00738787651062,
"step": 13574
},
{
"epoch": 0.8227878787878787,
"grad_norm": 0.049028679728507996,
"learning_rate": 0.0005909090909090909,
"loss": 3.0627217292785645,
"step": 13576
},
{
"epoch": 0.8229090909090909,
"grad_norm": 0.0478411503136158,
"learning_rate": 0.0005905050505050505,
"loss": 3.0615110397338867,
"step": 13578
},
{
"epoch": 0.823030303030303,
"grad_norm": 0.051691774278879166,
"learning_rate": 0.0005901010101010101,
"loss": 3.033698081970215,
"step": 13580
},
{
"epoch": 0.8231515151515152,
"grad_norm": 0.05496274679899216,
"learning_rate": 0.0005896969696969697,
"loss": 3.021897554397583,
"step": 13582
},
{
"epoch": 0.8232727272727273,
"grad_norm": 0.0618591345846653,
"learning_rate": 0.0005892929292929293,
"loss": 3.044389009475708,
"step": 13584
},
{
"epoch": 0.8233939393939393,
"grad_norm": 0.06198738515377045,
"learning_rate": 0.0005888888888888889,
"loss": 3.0169591903686523,
"step": 13586
},
{
"epoch": 0.8235151515151515,
"grad_norm": 0.04764118790626526,
"learning_rate": 0.0005884848484848485,
"loss": 3.0570521354675293,
"step": 13588
},
{
"epoch": 0.8236363636363636,
"grad_norm": 0.05214489996433258,
"learning_rate": 0.0005880808080808082,
"loss": 3.0601789951324463,
"step": 13590
},
{
"epoch": 0.8237575757575758,
"grad_norm": 0.04733531177043915,
"learning_rate": 0.0005876767676767677,
"loss": 3.038083076477051,
"step": 13592
},
{
"epoch": 0.8238787878787879,
"grad_norm": 0.047226469963788986,
"learning_rate": 0.0005872727272727273,
"loss": 3.0380382537841797,
"step": 13594
},
{
"epoch": 0.824,
"grad_norm": 0.05643172189593315,
"learning_rate": 0.0005868686868686869,
"loss": 3.0314667224884033,
"step": 13596
},
{
"epoch": 0.8241212121212121,
"grad_norm": 0.059464432299137115,
"learning_rate": 0.0005864646464646465,
"loss": 3.0160417556762695,
"step": 13598
},
{
"epoch": 0.8242424242424242,
"grad_norm": 0.0753321647644043,
"learning_rate": 0.0005860606060606061,
"loss": 3.04836368560791,
"step": 13600
},
{
"epoch": 0.8243636363636364,
"grad_norm": 0.05770951136946678,
"learning_rate": 0.0005856565656565657,
"loss": 3.059128522872925,
"step": 13602
},
{
"epoch": 0.8244848484848485,
"grad_norm": 0.04923202469944954,
"learning_rate": 0.0005852525252525253,
"loss": 3.0162129402160645,
"step": 13604
},
{
"epoch": 0.8246060606060606,
"grad_norm": 0.049564167857170105,
"learning_rate": 0.0005848484848484849,
"loss": 3.0348286628723145,
"step": 13606
},
{
"epoch": 0.8247272727272728,
"grad_norm": 0.048757314682006836,
"learning_rate": 0.0005844444444444444,
"loss": 3.06526780128479,
"step": 13608
},
{
"epoch": 0.8248484848484848,
"grad_norm": 0.047182597219944,
"learning_rate": 0.0005840404040404041,
"loss": 3.056626558303833,
"step": 13610
},
{
"epoch": 0.824969696969697,
"grad_norm": 0.04671064019203186,
"learning_rate": 0.0005836363636363636,
"loss": 3.076991558074951,
"step": 13612
},
{
"epoch": 0.8250909090909091,
"grad_norm": 0.0465199239552021,
"learning_rate": 0.0005832323232323232,
"loss": 3.052119016647339,
"step": 13614
},
{
"epoch": 0.8252121212121212,
"grad_norm": 0.048488881438970566,
"learning_rate": 0.0005828282828282828,
"loss": 2.9932432174682617,
"step": 13616
},
{
"epoch": 0.8253333333333334,
"grad_norm": 0.055040281265974045,
"learning_rate": 0.0005824242424242425,
"loss": 3.0289714336395264,
"step": 13618
},
{
"epoch": 0.8254545454545454,
"grad_norm": 0.050386812537908554,
"learning_rate": 0.000582020202020202,
"loss": 3.061964750289917,
"step": 13620
},
{
"epoch": 0.8255757575757576,
"grad_norm": 0.055467892438173294,
"learning_rate": 0.0005816161616161616,
"loss": 3.0382306575775146,
"step": 13622
},
{
"epoch": 0.8256969696969697,
"grad_norm": 0.05555763840675354,
"learning_rate": 0.0005812121212121213,
"loss": 3.0695927143096924,
"step": 13624
},
{
"epoch": 0.8258181818181818,
"grad_norm": 0.0581030547618866,
"learning_rate": 0.0005808080808080809,
"loss": 3.0819971561431885,
"step": 13626
},
{
"epoch": 0.825939393939394,
"grad_norm": 0.053008537739515305,
"learning_rate": 0.0005804040404040404,
"loss": 3.063852071762085,
"step": 13628
},
{
"epoch": 0.826060606060606,
"grad_norm": 0.046426922082901,
"learning_rate": 0.00058,
"loss": 3.057105779647827,
"step": 13630
},
{
"epoch": 0.8261818181818181,
"grad_norm": 0.04602977633476257,
"learning_rate": 0.0005795959595959597,
"loss": 3.0511960983276367,
"step": 13632
},
{
"epoch": 0.8263030303030303,
"grad_norm": 0.047359760850667953,
"learning_rate": 0.0005791919191919191,
"loss": 3.089928150177002,
"step": 13634
},
{
"epoch": 0.8264242424242424,
"grad_norm": 0.05011024326086044,
"learning_rate": 0.0005787878787878788,
"loss": 3.1084020137786865,
"step": 13636
},
{
"epoch": 0.8265454545454546,
"grad_norm": 0.059670958667993546,
"learning_rate": 0.0005783838383838384,
"loss": 3.0786447525024414,
"step": 13638
},
{
"epoch": 0.8266666666666667,
"grad_norm": 0.05205462500452995,
"learning_rate": 0.0005779797979797981,
"loss": 3.089325428009033,
"step": 13640
},
{
"epoch": 0.8267878787878787,
"grad_norm": 0.05038244649767876,
"learning_rate": 0.0005775757575757575,
"loss": 3.0458412170410156,
"step": 13642
},
{
"epoch": 0.8269090909090909,
"grad_norm": 0.049855928868055344,
"learning_rate": 0.0005771717171717172,
"loss": 3.0736687183380127,
"step": 13644
},
{
"epoch": 0.827030303030303,
"grad_norm": 0.050245579332113266,
"learning_rate": 0.0005767676767676768,
"loss": 3.0994112491607666,
"step": 13646
},
{
"epoch": 0.8271515151515152,
"grad_norm": 0.04745440557599068,
"learning_rate": 0.0005763636363636364,
"loss": 3.056340217590332,
"step": 13648
},
{
"epoch": 0.8272727272727273,
"grad_norm": 0.049330420792102814,
"learning_rate": 0.0005759595959595959,
"loss": 3.0747358798980713,
"step": 13650
},
{
"epoch": 0.8273939393939393,
"grad_norm": 0.04851084202528,
"learning_rate": 0.0005755555555555556,
"loss": 3.0975215435028076,
"step": 13652
},
{
"epoch": 0.8275151515151515,
"grad_norm": 0.0608077198266983,
"learning_rate": 0.0005751515151515152,
"loss": 3.085906505584717,
"step": 13654
},
{
"epoch": 0.8276363636363636,
"grad_norm": 0.059960536658763885,
"learning_rate": 0.0005747474747474747,
"loss": 3.071129322052002,
"step": 13656
},
{
"epoch": 0.8277575757575758,
"grad_norm": 0.06676787883043289,
"learning_rate": 0.0005743434343434344,
"loss": 3.0777430534362793,
"step": 13658
},
{
"epoch": 0.8278787878787879,
"grad_norm": 0.06070912629365921,
"learning_rate": 0.000573939393939394,
"loss": 3.0826685428619385,
"step": 13660
},
{
"epoch": 0.828,
"grad_norm": 0.05517510697245598,
"learning_rate": 0.0005735353535353536,
"loss": 3.0801005363464355,
"step": 13662
},
{
"epoch": 0.8281212121212121,
"grad_norm": 0.051419906318187714,
"learning_rate": 0.0005731313131313131,
"loss": 3.0406131744384766,
"step": 13664
},
{
"epoch": 0.8282424242424242,
"grad_norm": 0.047955386340618134,
"learning_rate": 0.0005727272727272728,
"loss": 3.0642735958099365,
"step": 13666
},
{
"epoch": 0.8283636363636364,
"grad_norm": 0.048349566757678986,
"learning_rate": 0.0005723232323232323,
"loss": 3.0864741802215576,
"step": 13668
},
{
"epoch": 0.8284848484848485,
"grad_norm": 0.052033063024282455,
"learning_rate": 0.000571919191919192,
"loss": 3.052867889404297,
"step": 13670
},
{
"epoch": 0.8286060606060606,
"grad_norm": 0.04955743998289108,
"learning_rate": 0.0005715151515151515,
"loss": 3.0061380863189697,
"step": 13672
},
{
"epoch": 0.8287272727272728,
"grad_norm": 0.04854714497923851,
"learning_rate": 0.0005711111111111111,
"loss": 3.0143041610717773,
"step": 13674
},
{
"epoch": 0.8288484848484848,
"grad_norm": 0.049232542514801025,
"learning_rate": 0.0005707070707070707,
"loss": 3.049494504928589,
"step": 13676
},
{
"epoch": 0.828969696969697,
"grad_norm": 0.04694598913192749,
"learning_rate": 0.0005703030303030304,
"loss": 3.099536895751953,
"step": 13678
},
{
"epoch": 0.8290909090909091,
"grad_norm": 0.04602203518152237,
"learning_rate": 0.0005698989898989899,
"loss": 3.0516152381896973,
"step": 13680
},
{
"epoch": 0.8292121212121212,
"grad_norm": 0.048387590795755386,
"learning_rate": 0.0005694949494949495,
"loss": 3.058864116668701,
"step": 13682
},
{
"epoch": 0.8293333333333334,
"grad_norm": 0.04948975518345833,
"learning_rate": 0.0005690909090909091,
"loss": 3.0600273609161377,
"step": 13684
},
{
"epoch": 0.8294545454545454,
"grad_norm": 0.046487126499414444,
"learning_rate": 0.0005686868686868687,
"loss": 3.0601253509521484,
"step": 13686
},
{
"epoch": 0.8295757575757575,
"grad_norm": 0.048680033534765244,
"learning_rate": 0.0005682828282828283,
"loss": 3.0528223514556885,
"step": 13688
},
{
"epoch": 0.8296969696969697,
"grad_norm": 0.05114282667636871,
"learning_rate": 0.0005678787878787879,
"loss": 3.0353734493255615,
"step": 13690
},
{
"epoch": 0.8298181818181818,
"grad_norm": 0.04705335199832916,
"learning_rate": 0.0005674747474747476,
"loss": 3.042607069015503,
"step": 13692
},
{
"epoch": 0.829939393939394,
"grad_norm": 0.04883192107081413,
"learning_rate": 0.0005670707070707071,
"loss": 3.0757248401641846,
"step": 13694
},
{
"epoch": 0.830060606060606,
"grad_norm": 0.05167026445269585,
"learning_rate": 0.0005666666666666667,
"loss": 3.0898683071136475,
"step": 13696
},
{
"epoch": 0.8301818181818181,
"grad_norm": 0.046664923429489136,
"learning_rate": 0.0005662626262626263,
"loss": 3.0808730125427246,
"step": 13698
},
{
"epoch": 0.8303030303030303,
"grad_norm": 0.04716499522328377,
"learning_rate": 0.000565858585858586,
"loss": 3.0168039798736572,
"step": 13700
},
{
"epoch": 0.8304242424242424,
"grad_norm": 0.04711788520216942,
"learning_rate": 0.0005654545454545454,
"loss": 3.052604913711548,
"step": 13702
},
{
"epoch": 0.8305454545454546,
"grad_norm": 0.04815881326794624,
"learning_rate": 0.0005650505050505051,
"loss": 3.066643476486206,
"step": 13704
},
{
"epoch": 0.8306666666666667,
"grad_norm": 0.04559068754315376,
"learning_rate": 0.0005646464646464647,
"loss": 3.103578805923462,
"step": 13706
},
{
"epoch": 0.8307878787878787,
"grad_norm": 0.049779124557971954,
"learning_rate": 0.0005642424242424242,
"loss": 3.0376272201538086,
"step": 13708
},
{
"epoch": 0.8309090909090909,
"grad_norm": 0.04982629418373108,
"learning_rate": 0.0005638383838383838,
"loss": 3.0102570056915283,
"step": 13710
},
{
"epoch": 0.831030303030303,
"grad_norm": 0.0476880818605423,
"learning_rate": 0.0005634343434343435,
"loss": 3.056259870529175,
"step": 13712
},
{
"epoch": 0.8311515151515152,
"grad_norm": 0.049839697778224945,
"learning_rate": 0.0005630303030303031,
"loss": 3.0409231185913086,
"step": 13714
},
{
"epoch": 0.8312727272727273,
"grad_norm": 0.05717024952173233,
"learning_rate": 0.0005626262626262626,
"loss": 3.0458390712738037,
"step": 13716
},
{
"epoch": 0.8313939393939394,
"grad_norm": 0.05510488152503967,
"learning_rate": 0.0005622222222222222,
"loss": 3.0439844131469727,
"step": 13718
},
{
"epoch": 0.8315151515151515,
"grad_norm": 0.04678842052817345,
"learning_rate": 0.0005618181818181819,
"loss": 3.079713821411133,
"step": 13720
},
{
"epoch": 0.8316363636363636,
"grad_norm": 0.048351287841796875,
"learning_rate": 0.0005614141414141415,
"loss": 3.0602211952209473,
"step": 13722
},
{
"epoch": 0.8317575757575758,
"grad_norm": 0.05447176843881607,
"learning_rate": 0.000561010101010101,
"loss": 3.0050430297851562,
"step": 13724
},
{
"epoch": 0.8318787878787879,
"grad_norm": 0.05773351714015007,
"learning_rate": 0.0005606060606060607,
"loss": 3.08632755279541,
"step": 13726
},
{
"epoch": 0.832,
"grad_norm": 0.06283583492040634,
"learning_rate": 0.0005602020202020202,
"loss": 3.067247152328491,
"step": 13728
},
{
"epoch": 0.8321212121212122,
"grad_norm": 0.06690982729196548,
"learning_rate": 0.0005597979797979798,
"loss": 3.0980517864227295,
"step": 13730
},
{
"epoch": 0.8322424242424242,
"grad_norm": 0.0470271036028862,
"learning_rate": 0.0005593939393939394,
"loss": 3.047401189804077,
"step": 13732
},
{
"epoch": 0.8323636363636364,
"grad_norm": 0.04734386131167412,
"learning_rate": 0.0005589898989898991,
"loss": 3.0217180252075195,
"step": 13734
},
{
"epoch": 0.8324848484848485,
"grad_norm": 0.04914359003305435,
"learning_rate": 0.0005585858585858586,
"loss": 3.0471484661102295,
"step": 13736
},
{
"epoch": 0.8326060606060606,
"grad_norm": 0.047278519719839096,
"learning_rate": 0.0005581818181818182,
"loss": 3.0098021030426025,
"step": 13738
},
{
"epoch": 0.8327272727272728,
"grad_norm": 0.054294221103191376,
"learning_rate": 0.0005577777777777778,
"loss": 3.0408811569213867,
"step": 13740
},
{
"epoch": 0.8328484848484848,
"grad_norm": 0.04731460288167,
"learning_rate": 0.0005573737373737374,
"loss": 3.0162904262542725,
"step": 13742
},
{
"epoch": 0.8329696969696969,
"grad_norm": 0.05961072817444801,
"learning_rate": 0.000556969696969697,
"loss": 3.0318355560302734,
"step": 13744
},
{
"epoch": 0.8330909090909091,
"grad_norm": 0.05554961413145065,
"learning_rate": 0.0005565656565656566,
"loss": 3.0365543365478516,
"step": 13746
},
{
"epoch": 0.8332121212121212,
"grad_norm": 0.046842753887176514,
"learning_rate": 0.0005561616161616162,
"loss": 3.0382730960845947,
"step": 13748
},
{
"epoch": 0.8333333333333334,
"grad_norm": 0.05024043843150139,
"learning_rate": 0.0005557575757575758,
"loss": 3.0115773677825928,
"step": 13750
},
{
"epoch": 0.8334545454545454,
"grad_norm": 0.0481439083814621,
"learning_rate": 0.0005553535353535353,
"loss": 3.0894508361816406,
"step": 13752
},
{
"epoch": 0.8335757575757575,
"grad_norm": 0.056520309299230576,
"learning_rate": 0.000554949494949495,
"loss": 3.0022785663604736,
"step": 13754
},
{
"epoch": 0.8336969696969697,
"grad_norm": 0.05741679295897484,
"learning_rate": 0.0005545454545454546,
"loss": 2.976047992706299,
"step": 13756
},
{
"epoch": 0.8338181818181818,
"grad_norm": 0.04889359697699547,
"learning_rate": 0.0005541414141414142,
"loss": 3.085017681121826,
"step": 13758
},
{
"epoch": 0.833939393939394,
"grad_norm": 0.04671713337302208,
"learning_rate": 0.0005537373737373738,
"loss": 3.035235643386841,
"step": 13760
},
{
"epoch": 0.8340606060606061,
"grad_norm": 0.05103049427270889,
"learning_rate": 0.0005533333333333333,
"loss": 3.0306262969970703,
"step": 13762
},
{
"epoch": 0.8341818181818181,
"grad_norm": 0.04742493852972984,
"learning_rate": 0.000552929292929293,
"loss": 3.0735573768615723,
"step": 13764
},
{
"epoch": 0.8343030303030303,
"grad_norm": 0.046561770141124725,
"learning_rate": 0.0005525252525252526,
"loss": 3.0619070529937744,
"step": 13766
},
{
"epoch": 0.8344242424242424,
"grad_norm": 0.06087931618094444,
"learning_rate": 0.0005521212121212121,
"loss": 3.051612615585327,
"step": 13768
},
{
"epoch": 0.8345454545454546,
"grad_norm": 0.059457141906023026,
"learning_rate": 0.0005517171717171717,
"loss": 3.0369129180908203,
"step": 13770
},
{
"epoch": 0.8346666666666667,
"grad_norm": 0.05577906593680382,
"learning_rate": 0.0005513131313131314,
"loss": 3.0835278034210205,
"step": 13772
},
{
"epoch": 0.8347878787878787,
"grad_norm": 0.05445067584514618,
"learning_rate": 0.0005509090909090909,
"loss": 3.039630889892578,
"step": 13774
},
{
"epoch": 0.8349090909090909,
"grad_norm": 0.05301586166024208,
"learning_rate": 0.0005505050505050505,
"loss": 3.05086612701416,
"step": 13776
},
{
"epoch": 0.835030303030303,
"grad_norm": 0.05030181631445885,
"learning_rate": 0.0005501010101010101,
"loss": 3.08060622215271,
"step": 13778
},
{
"epoch": 0.8351515151515152,
"grad_norm": 0.04720666632056236,
"learning_rate": 0.0005496969696969698,
"loss": 3.02995228767395,
"step": 13780
},
{
"epoch": 0.8352727272727273,
"grad_norm": 0.04680046811699867,
"learning_rate": 0.0005492929292929292,
"loss": 3.026705741882324,
"step": 13782
},
{
"epoch": 0.8353939393939394,
"grad_norm": 0.05016390606760979,
"learning_rate": 0.0005488888888888889,
"loss": 3.0518457889556885,
"step": 13784
},
{
"epoch": 0.8355151515151515,
"grad_norm": 0.05727597326040268,
"learning_rate": 0.0005484848484848485,
"loss": 3.042036533355713,
"step": 13786
},
{
"epoch": 0.8356363636363636,
"grad_norm": 0.05192696303129196,
"learning_rate": 0.0005480808080808082,
"loss": 3.089231491088867,
"step": 13788
},
{
"epoch": 0.8357575757575758,
"grad_norm": 0.04974263533949852,
"learning_rate": 0.0005476767676767677,
"loss": 3.0327229499816895,
"step": 13790
},
{
"epoch": 0.8358787878787879,
"grad_norm": 0.061510536819696426,
"learning_rate": 0.0005472727272727273,
"loss": 2.991811990737915,
"step": 13792
},
{
"epoch": 0.836,
"grad_norm": 0.04944981262087822,
"learning_rate": 0.000546868686868687,
"loss": 3.0929198265075684,
"step": 13794
},
{
"epoch": 0.8361212121212122,
"grad_norm": 0.04830443486571312,
"learning_rate": 0.0005464646464646464,
"loss": 3.0534021854400635,
"step": 13796
},
{
"epoch": 0.8362424242424242,
"grad_norm": 0.05445937439799309,
"learning_rate": 0.0005460606060606061,
"loss": 3.056471586227417,
"step": 13798
},
{
"epoch": 0.8363636363636363,
"grad_norm": 0.056354038417339325,
"learning_rate": 0.0005456565656565657,
"loss": 3.0638790130615234,
"step": 13800
},
{
"epoch": 0.8364848484848485,
"grad_norm": 0.05875447764992714,
"learning_rate": 0.0005452525252525253,
"loss": 2.998694658279419,
"step": 13802
},
{
"epoch": 0.8366060606060606,
"grad_norm": 0.048524726182222366,
"learning_rate": 0.0005448484848484848,
"loss": 3.0309295654296875,
"step": 13804
},
{
"epoch": 0.8367272727272728,
"grad_norm": 0.05044050142168999,
"learning_rate": 0.0005444444444444445,
"loss": 3.0302743911743164,
"step": 13806
},
{
"epoch": 0.8368484848484848,
"grad_norm": 0.04807687923312187,
"learning_rate": 0.0005440404040404041,
"loss": 3.0619211196899414,
"step": 13808
},
{
"epoch": 0.8369696969696969,
"grad_norm": 0.046919554471969604,
"learning_rate": 0.0005436363636363637,
"loss": 3.0366976261138916,
"step": 13810
},
{
"epoch": 0.8370909090909091,
"grad_norm": 0.04730372130870819,
"learning_rate": 0.0005432323232323232,
"loss": 3.030794858932495,
"step": 13812
},
{
"epoch": 0.8372121212121212,
"grad_norm": 0.05022171884775162,
"learning_rate": 0.0005428282828282829,
"loss": 3.033234119415283,
"step": 13814
},
{
"epoch": 0.8373333333333334,
"grad_norm": 0.048432640731334686,
"learning_rate": 0.0005424242424242424,
"loss": 3.0277280807495117,
"step": 13816
},
{
"epoch": 0.8374545454545455,
"grad_norm": 0.05732167512178421,
"learning_rate": 0.000542020202020202,
"loss": 3.053408622741699,
"step": 13818
},
{
"epoch": 0.8375757575757575,
"grad_norm": 0.05478091537952423,
"learning_rate": 0.0005416161616161616,
"loss": 2.968904495239258,
"step": 13820
},
{
"epoch": 0.8376969696969697,
"grad_norm": 0.05327370762825012,
"learning_rate": 0.0005412121212121212,
"loss": 3.015718936920166,
"step": 13822
},
{
"epoch": 0.8378181818181818,
"grad_norm": 0.049561142921447754,
"learning_rate": 0.0005408080808080809,
"loss": 3.0685226917266846,
"step": 13824
},
{
"epoch": 0.837939393939394,
"grad_norm": 0.0500648207962513,
"learning_rate": 0.0005404040404040404,
"loss": 3.0447182655334473,
"step": 13826
},
{
"epoch": 0.8380606060606061,
"grad_norm": 0.05689282342791557,
"learning_rate": 0.0005400000000000001,
"loss": 3.041506767272949,
"step": 13828
},
{
"epoch": 0.8381818181818181,
"grad_norm": 0.05567318946123123,
"learning_rate": 0.0005395959595959596,
"loss": 3.045689821243286,
"step": 13830
},
{
"epoch": 0.8383030303030303,
"grad_norm": 0.04984410107135773,
"learning_rate": 0.0005391919191919193,
"loss": 3.0567610263824463,
"step": 13832
},
{
"epoch": 0.8384242424242424,
"grad_norm": 0.050773609429597855,
"learning_rate": 0.0005387878787878788,
"loss": 3.034923553466797,
"step": 13834
},
{
"epoch": 0.8385454545454546,
"grad_norm": 0.0500849112868309,
"learning_rate": 0.0005383838383838384,
"loss": 3.08501935005188,
"step": 13836
},
{
"epoch": 0.8386666666666667,
"grad_norm": 0.04951319843530655,
"learning_rate": 0.000537979797979798,
"loss": 3.019508123397827,
"step": 13838
},
{
"epoch": 0.8387878787878787,
"grad_norm": 0.05574405938386917,
"learning_rate": 0.0005375757575757576,
"loss": 2.9969377517700195,
"step": 13840
},
{
"epoch": 0.8389090909090909,
"grad_norm": 0.0614066943526268,
"learning_rate": 0.0005371717171717172,
"loss": 3.0168240070343018,
"step": 13842
},
{
"epoch": 0.839030303030303,
"grad_norm": 0.06263849884271622,
"learning_rate": 0.0005367676767676768,
"loss": 3.0727617740631104,
"step": 13844
},
{
"epoch": 0.8391515151515152,
"grad_norm": 0.0540604367852211,
"learning_rate": 0.0005363636363636364,
"loss": 3.0312938690185547,
"step": 13846
},
{
"epoch": 0.8392727272727273,
"grad_norm": 0.050298500806093216,
"learning_rate": 0.000535959595959596,
"loss": 3.020310640335083,
"step": 13848
},
{
"epoch": 0.8393939393939394,
"grad_norm": 0.04745107144117355,
"learning_rate": 0.0005355555555555555,
"loss": 3.05973744392395,
"step": 13850
},
{
"epoch": 0.8395151515151515,
"grad_norm": 0.04761083051562309,
"learning_rate": 0.0005351515151515152,
"loss": 3.0068914890289307,
"step": 13852
},
{
"epoch": 0.8396363636363636,
"grad_norm": 0.04821622371673584,
"learning_rate": 0.0005347474747474748,
"loss": 3.029892683029175,
"step": 13854
},
{
"epoch": 0.8397575757575757,
"grad_norm": 0.04855246841907501,
"learning_rate": 0.0005343434343434343,
"loss": 3.066859722137451,
"step": 13856
},
{
"epoch": 0.8398787878787879,
"grad_norm": 0.04855259507894516,
"learning_rate": 0.000533939393939394,
"loss": 2.971670389175415,
"step": 13858
},
{
"epoch": 0.84,
"grad_norm": 0.04930008202791214,
"learning_rate": 0.0005335353535353536,
"loss": 3.0219099521636963,
"step": 13860
},
{
"epoch": 0.8401212121212122,
"grad_norm": 0.060206733644008636,
"learning_rate": 0.0005331313131313131,
"loss": 3.0090506076812744,
"step": 13862
},
{
"epoch": 0.8402424242424242,
"grad_norm": 0.0677548497915268,
"learning_rate": 0.0005327272727272727,
"loss": 3.0603888034820557,
"step": 13864
},
{
"epoch": 0.8403636363636363,
"grad_norm": 0.050809986889362335,
"learning_rate": 0.0005323232323232324,
"loss": 2.999768018722534,
"step": 13866
},
{
"epoch": 0.8404848484848485,
"grad_norm": 0.04931553453207016,
"learning_rate": 0.000531919191919192,
"loss": 3.055536985397339,
"step": 13868
},
{
"epoch": 0.8406060606060606,
"grad_norm": 0.04742056876420975,
"learning_rate": 0.0005315151515151515,
"loss": 3.0081264972686768,
"step": 13870
},
{
"epoch": 0.8407272727272728,
"grad_norm": 0.04999512806534767,
"learning_rate": 0.0005311111111111111,
"loss": 3.001218557357788,
"step": 13872
},
{
"epoch": 0.8408484848484848,
"grad_norm": 0.05122840031981468,
"learning_rate": 0.0005307070707070708,
"loss": 3.035796642303467,
"step": 13874
},
{
"epoch": 0.8409696969696969,
"grad_norm": 0.05963926389813423,
"learning_rate": 0.0005303030303030302,
"loss": 3.0293893814086914,
"step": 13876
},
{
"epoch": 0.8410909090909091,
"grad_norm": 0.05520037189126015,
"learning_rate": 0.0005298989898989899,
"loss": 3.0764362812042236,
"step": 13878
},
{
"epoch": 0.8412121212121212,
"grad_norm": 0.049331460148096085,
"learning_rate": 0.0005294949494949495,
"loss": 3.04189133644104,
"step": 13880
},
{
"epoch": 0.8413333333333334,
"grad_norm": 0.05320977047085762,
"learning_rate": 0.0005290909090909092,
"loss": 3.0606558322906494,
"step": 13882
},
{
"epoch": 0.8414545454545455,
"grad_norm": 0.053117770701646805,
"learning_rate": 0.0005286868686868686,
"loss": 3.0294430255889893,
"step": 13884
},
{
"epoch": 0.8415757575757575,
"grad_norm": 0.0496649295091629,
"learning_rate": 0.0005282828282828283,
"loss": 3.013505697250366,
"step": 13886
},
{
"epoch": 0.8416969696969697,
"grad_norm": 0.05130770802497864,
"learning_rate": 0.0005278787878787879,
"loss": 3.054457664489746,
"step": 13888
},
{
"epoch": 0.8418181818181818,
"grad_norm": 0.06711860746145248,
"learning_rate": 0.0005274747474747475,
"loss": 3.0813028812408447,
"step": 13890
},
{
"epoch": 0.841939393939394,
"grad_norm": 0.06655155122280121,
"learning_rate": 0.0005270707070707071,
"loss": 3.034228801727295,
"step": 13892
},
{
"epoch": 0.8420606060606061,
"grad_norm": 0.051579371094703674,
"learning_rate": 0.0005266666666666667,
"loss": 3.045003890991211,
"step": 13894
},
{
"epoch": 0.8421818181818181,
"grad_norm": 0.05564428120851517,
"learning_rate": 0.0005262626262626263,
"loss": 3.0330519676208496,
"step": 13896
},
{
"epoch": 0.8423030303030303,
"grad_norm": 0.06053447723388672,
"learning_rate": 0.0005258585858585859,
"loss": 3.022475004196167,
"step": 13898
},
{
"epoch": 0.8424242424242424,
"grad_norm": 0.0713488832116127,
"learning_rate": 0.0005254545454545455,
"loss": 3.0458285808563232,
"step": 13900
},
{
"epoch": 0.8425454545454546,
"grad_norm": 0.053415168076753616,
"learning_rate": 0.0005250505050505051,
"loss": 3.0429694652557373,
"step": 13902
},
{
"epoch": 0.8426666666666667,
"grad_norm": 0.05544647201895714,
"learning_rate": 0.0005246464646464647,
"loss": 3.0336296558380127,
"step": 13904
},
{
"epoch": 0.8427878787878788,
"grad_norm": 0.05216856300830841,
"learning_rate": 0.0005242424242424242,
"loss": 3.0323164463043213,
"step": 13906
},
{
"epoch": 0.8429090909090909,
"grad_norm": 0.06013714522123337,
"learning_rate": 0.0005238383838383839,
"loss": 3.0079143047332764,
"step": 13908
},
{
"epoch": 0.843030303030303,
"grad_norm": 0.054876312613487244,
"learning_rate": 0.0005234343434343434,
"loss": 2.9822559356689453,
"step": 13910
},
{
"epoch": 0.8431515151515151,
"grad_norm": 0.053966645151376724,
"learning_rate": 0.0005230303030303031,
"loss": 3.002213954925537,
"step": 13912
},
{
"epoch": 0.8432727272727273,
"grad_norm": 0.04974038898944855,
"learning_rate": 0.0005226262626262626,
"loss": 3.002073049545288,
"step": 13914
},
{
"epoch": 0.8433939393939394,
"grad_norm": 0.048805564641952515,
"learning_rate": 0.0005222222222222222,
"loss": 3.0352234840393066,
"step": 13916
},
{
"epoch": 0.8435151515151516,
"grad_norm": 0.047877177596092224,
"learning_rate": 0.0005218181818181818,
"loss": 3.02665376663208,
"step": 13918
},
{
"epoch": 0.8436363636363636,
"grad_norm": 0.05774440988898277,
"learning_rate": 0.0005214141414141415,
"loss": 2.988978862762451,
"step": 13920
},
{
"epoch": 0.8437575757575757,
"grad_norm": 0.055178746581077576,
"learning_rate": 0.000521010101010101,
"loss": 3.071418523788452,
"step": 13922
},
{
"epoch": 0.8438787878787879,
"grad_norm": 0.0495782271027565,
"learning_rate": 0.0005206060606060606,
"loss": 3.0313236713409424,
"step": 13924
},
{
"epoch": 0.844,
"grad_norm": 0.0478297621011734,
"learning_rate": 0.0005202020202020203,
"loss": 3.0190091133117676,
"step": 13926
},
{
"epoch": 0.8441212121212122,
"grad_norm": 0.05165902525186539,
"learning_rate": 0.0005197979797979798,
"loss": 3.015874147415161,
"step": 13928
},
{
"epoch": 0.8442424242424242,
"grad_norm": 0.048919107764959335,
"learning_rate": 0.0005193939393939394,
"loss": 2.99009370803833,
"step": 13930
},
{
"epoch": 0.8443636363636363,
"grad_norm": 0.05016336962580681,
"learning_rate": 0.000518989898989899,
"loss": 3.0263006687164307,
"step": 13932
},
{
"epoch": 0.8444848484848485,
"grad_norm": 0.049689408391714096,
"learning_rate": 0.0005185858585858587,
"loss": 3.0245766639709473,
"step": 13934
},
{
"epoch": 0.8446060606060606,
"grad_norm": 0.047218695282936096,
"learning_rate": 0.0005181818181818182,
"loss": 3.0240907669067383,
"step": 13936
},
{
"epoch": 0.8447272727272728,
"grad_norm": 0.04817831143736839,
"learning_rate": 0.0005177777777777778,
"loss": 3.0310702323913574,
"step": 13938
},
{
"epoch": 0.8448484848484848,
"grad_norm": 0.047452278435230255,
"learning_rate": 0.0005173737373737374,
"loss": 2.9910662174224854,
"step": 13940
},
{
"epoch": 0.8449696969696969,
"grad_norm": 0.047477394342422485,
"learning_rate": 0.0005169696969696971,
"loss": 2.9907705783843994,
"step": 13942
},
{
"epoch": 0.8450909090909091,
"grad_norm": 0.04790354520082474,
"learning_rate": 0.0005165656565656565,
"loss": 2.9963150024414062,
"step": 13944
},
{
"epoch": 0.8452121212121212,
"grad_norm": 0.05080801993608475,
"learning_rate": 0.0005161616161616162,
"loss": 3.0027341842651367,
"step": 13946
},
{
"epoch": 0.8453333333333334,
"grad_norm": 0.054028015583753586,
"learning_rate": 0.0005157575757575758,
"loss": 2.9812042713165283,
"step": 13948
},
{
"epoch": 0.8454545454545455,
"grad_norm": 0.0485629066824913,
"learning_rate": 0.0005153535353535353,
"loss": 3.0353312492370605,
"step": 13950
},
{
"epoch": 0.8455757575757575,
"grad_norm": 0.04844731092453003,
"learning_rate": 0.0005149494949494949,
"loss": 2.9926657676696777,
"step": 13952
},
{
"epoch": 0.8456969696969697,
"grad_norm": 0.04839122295379639,
"learning_rate": 0.0005145454545454546,
"loss": 3.0281929969787598,
"step": 13954
},
{
"epoch": 0.8458181818181818,
"grad_norm": 0.05084831267595291,
"learning_rate": 0.0005141414141414142,
"loss": 3.0646982192993164,
"step": 13956
},
{
"epoch": 0.845939393939394,
"grad_norm": 0.049079302698373795,
"learning_rate": 0.0005137373737373737,
"loss": 3.0055272579193115,
"step": 13958
},
{
"epoch": 0.8460606060606061,
"grad_norm": 0.04796086996793747,
"learning_rate": 0.0005133333333333334,
"loss": 3.0288302898406982,
"step": 13960
},
{
"epoch": 0.8461818181818181,
"grad_norm": 0.05112894997000694,
"learning_rate": 0.000512929292929293,
"loss": 3.0249147415161133,
"step": 13962
},
{
"epoch": 0.8463030303030303,
"grad_norm": 0.04895438998937607,
"learning_rate": 0.0005125252525252526,
"loss": 3.003056287765503,
"step": 13964
},
{
"epoch": 0.8464242424242424,
"grad_norm": 0.04853765293955803,
"learning_rate": 0.0005121212121212121,
"loss": 3.0377604961395264,
"step": 13966
},
{
"epoch": 0.8465454545454546,
"grad_norm": 0.04999469220638275,
"learning_rate": 0.0005117171717171718,
"loss": 2.9664928913116455,
"step": 13968
},
{
"epoch": 0.8466666666666667,
"grad_norm": 0.04872550070285797,
"learning_rate": 0.0005113131313131313,
"loss": 3.007789373397827,
"step": 13970
},
{
"epoch": 0.8467878787878788,
"grad_norm": 0.055588625371456146,
"learning_rate": 0.000510909090909091,
"loss": 2.9732730388641357,
"step": 13972
},
{
"epoch": 0.846909090909091,
"grad_norm": 0.06512139737606049,
"learning_rate": 0.0005105050505050505,
"loss": 3.0316338539123535,
"step": 13974
},
{
"epoch": 0.847030303030303,
"grad_norm": 0.056886978447437286,
"learning_rate": 0.0005101010101010102,
"loss": 3.00374174118042,
"step": 13976
},
{
"epoch": 0.8471515151515151,
"grad_norm": 0.0493588000535965,
"learning_rate": 0.0005096969696969697,
"loss": 3.023202657699585,
"step": 13978
},
{
"epoch": 0.8472727272727273,
"grad_norm": 0.050261881202459335,
"learning_rate": 0.0005092929292929293,
"loss": 3.019454002380371,
"step": 13980
},
{
"epoch": 0.8473939393939394,
"grad_norm": 0.0498807467520237,
"learning_rate": 0.0005088888888888889,
"loss": 3.001552104949951,
"step": 13982
},
{
"epoch": 0.8475151515151516,
"grad_norm": 0.05647711083292961,
"learning_rate": 0.0005084848484848485,
"loss": 3.0330817699432373,
"step": 13984
},
{
"epoch": 0.8476363636363636,
"grad_norm": 0.04805181547999382,
"learning_rate": 0.000508080808080808,
"loss": 2.9641313552856445,
"step": 13986
},
{
"epoch": 0.8477575757575757,
"grad_norm": 0.04861879348754883,
"learning_rate": 0.0005076767676767677,
"loss": 2.9951932430267334,
"step": 13988
},
{
"epoch": 0.8478787878787879,
"grad_norm": 0.050768885761499405,
"learning_rate": 0.0005072727272727273,
"loss": 3.016080617904663,
"step": 13990
},
{
"epoch": 0.848,
"grad_norm": 0.05295084789395332,
"learning_rate": 0.0005068686868686869,
"loss": 3.009296178817749,
"step": 13992
},
{
"epoch": 0.8481212121212122,
"grad_norm": 0.056748777627944946,
"learning_rate": 0.0005064646464646466,
"loss": 3.0444531440734863,
"step": 13994
},
{
"epoch": 0.8482424242424242,
"grad_norm": 0.048173677176237106,
"learning_rate": 0.0005060606060606061,
"loss": 2.996213674545288,
"step": 13996
},
{
"epoch": 0.8483636363636363,
"grad_norm": 0.050079476088285446,
"learning_rate": 0.0005056565656565657,
"loss": 3.0293474197387695,
"step": 13998
},
{
"epoch": 0.8484848484848485,
"grad_norm": 0.047164469957351685,
"learning_rate": 0.0005052525252525253,
"loss": 2.9913687705993652,
"step": 14000
},
{
"epoch": 0.8486060606060606,
"grad_norm": 0.049793630838394165,
"learning_rate": 0.0005048484848484849,
"loss": 3.042208194732666,
"step": 14002
},
{
"epoch": 0.8487272727272728,
"grad_norm": 0.04782430827617645,
"learning_rate": 0.0005044444444444444,
"loss": 3.001535654067993,
"step": 14004
},
{
"epoch": 0.8488484848484849,
"grad_norm": 0.05401867628097534,
"learning_rate": 0.0005040404040404041,
"loss": 2.963942527770996,
"step": 14006
},
{
"epoch": 0.8489696969696969,
"grad_norm": 0.0527375191450119,
"learning_rate": 0.0005036363636363637,
"loss": 2.9883830547332764,
"step": 14008
},
{
"epoch": 0.8490909090909091,
"grad_norm": 0.04820151627063751,
"learning_rate": 0.0005032323232323232,
"loss": 3.0116159915924072,
"step": 14010
},
{
"epoch": 0.8492121212121212,
"grad_norm": 0.054287273436784744,
"learning_rate": 0.0005028282828282828,
"loss": 2.9679229259490967,
"step": 14012
},
{
"epoch": 0.8493333333333334,
"grad_norm": 0.04920995980501175,
"learning_rate": 0.0005024242424242425,
"loss": 2.9676830768585205,
"step": 14014
},
{
"epoch": 0.8494545454545455,
"grad_norm": 0.048981696367263794,
"learning_rate": 0.000502020202020202,
"loss": 2.9688947200775146,
"step": 14016
},
{
"epoch": 0.8495757575757575,
"grad_norm": 0.05256883054971695,
"learning_rate": 0.0005016161616161616,
"loss": 2.9390501976013184,
"step": 14018
},
{
"epoch": 0.8496969696969697,
"grad_norm": 0.049232400953769684,
"learning_rate": 0.0005012121212121212,
"loss": 2.999311685562134,
"step": 14020
},
{
"epoch": 0.8498181818181818,
"grad_norm": 0.051130183041095734,
"learning_rate": 0.0005008080808080809,
"loss": 3.014505386352539,
"step": 14022
},
{
"epoch": 0.849939393939394,
"grad_norm": 0.04917095601558685,
"learning_rate": 0.0005004040404040404,
"loss": 3.024186849594116,
"step": 14024
},
{
"epoch": 0.8500606060606061,
"grad_norm": 0.0490611232817173,
"learning_rate": 0.0005,
"loss": 3.0153932571411133,
"step": 14026
},
{
"epoch": 0.8501818181818181,
"grad_norm": 0.04779370501637459,
"learning_rate": 0.0004995959595959597,
"loss": 3.0111708641052246,
"step": 14028
},
{
"epoch": 0.8503030303030303,
"grad_norm": 0.05518573522567749,
"learning_rate": 0.0004991919191919193,
"loss": 2.996034860610962,
"step": 14030
},
{
"epoch": 0.8504242424242424,
"grad_norm": 0.0587821863591671,
"learning_rate": 0.0004987878787878788,
"loss": 2.9882311820983887,
"step": 14032
},
{
"epoch": 0.8505454545454545,
"grad_norm": 0.05010369420051575,
"learning_rate": 0.0004983838383838384,
"loss": 2.990635871887207,
"step": 14034
},
{
"epoch": 0.8506666666666667,
"grad_norm": 0.05044036731123924,
"learning_rate": 0.0004979797979797981,
"loss": 2.937511920928955,
"step": 14036
},
{
"epoch": 0.8507878787878788,
"grad_norm": 0.0488973893225193,
"learning_rate": 0.0004975757575757575,
"loss": 2.973806142807007,
"step": 14038
},
{
"epoch": 0.850909090909091,
"grad_norm": 0.051765523850917816,
"learning_rate": 0.0004971717171717172,
"loss": 2.945634603500366,
"step": 14040
},
{
"epoch": 0.851030303030303,
"grad_norm": 0.04776325449347496,
"learning_rate": 0.0004967676767676768,
"loss": 2.9843668937683105,
"step": 14042
},
{
"epoch": 0.8511515151515151,
"grad_norm": 0.049970243126153946,
"learning_rate": 0.0004963636363636364,
"loss": 3.034269332885742,
"step": 14044
},
{
"epoch": 0.8512727272727273,
"grad_norm": 0.0617811381816864,
"learning_rate": 0.0004959595959595959,
"loss": 2.9974045753479004,
"step": 14046
},
{
"epoch": 0.8513939393939394,
"grad_norm": 0.05318734422326088,
"learning_rate": 0.0004955555555555556,
"loss": 2.965582847595215,
"step": 14048
},
{
"epoch": 0.8515151515151516,
"grad_norm": 0.04931372404098511,
"learning_rate": 0.0004951515151515152,
"loss": 3.006769895553589,
"step": 14050
},
{
"epoch": 0.8516363636363636,
"grad_norm": 0.05901532620191574,
"learning_rate": 0.0004947474747474748,
"loss": 2.99212384223938,
"step": 14052
},
{
"epoch": 0.8517575757575757,
"grad_norm": 0.04830297455191612,
"learning_rate": 0.0004943434343434343,
"loss": 2.984041929244995,
"step": 14054
},
{
"epoch": 0.8518787878787879,
"grad_norm": 0.0513981431722641,
"learning_rate": 0.000493939393939394,
"loss": 3.092832326889038,
"step": 14056
},
{
"epoch": 0.852,
"grad_norm": 0.04877471551299095,
"learning_rate": 0.0004935353535353535,
"loss": 3.0178439617156982,
"step": 14058
},
{
"epoch": 0.8521212121212122,
"grad_norm": 0.05026483163237572,
"learning_rate": 0.0004931313131313131,
"loss": 3.0826704502105713,
"step": 14060
},
{
"epoch": 0.8522424242424242,
"grad_norm": 0.05427132919430733,
"learning_rate": 0.0004927272727272728,
"loss": 3.055943250656128,
"step": 14062
},
{
"epoch": 0.8523636363636363,
"grad_norm": 0.0530308373272419,
"learning_rate": 0.0004923232323232323,
"loss": 2.9746642112731934,
"step": 14064
},
{
"epoch": 0.8524848484848485,
"grad_norm": 0.0535893589258194,
"learning_rate": 0.000491919191919192,
"loss": 3.0271270275115967,
"step": 14066
},
{
"epoch": 0.8526060606060606,
"grad_norm": 0.05707880109548569,
"learning_rate": 0.0004915151515151515,
"loss": 3.045470714569092,
"step": 14068
},
{
"epoch": 0.8527272727272728,
"grad_norm": 0.04784010723233223,
"learning_rate": 0.0004911111111111112,
"loss": 3.0347962379455566,
"step": 14070
},
{
"epoch": 0.8528484848484849,
"grad_norm": 0.05448231101036072,
"learning_rate": 0.0004907070707070707,
"loss": 3.011704683303833,
"step": 14072
},
{
"epoch": 0.8529696969696969,
"grad_norm": 0.07736387848854065,
"learning_rate": 0.0004903030303030304,
"loss": 2.9790704250335693,
"step": 14074
},
{
"epoch": 0.8530909090909091,
"grad_norm": 0.07826820015907288,
"learning_rate": 0.0004898989898989899,
"loss": 2.96661376953125,
"step": 14076
},
{
"epoch": 0.8532121212121212,
"grad_norm": 0.049271490424871445,
"learning_rate": 0.0004894949494949495,
"loss": 3.042283296585083,
"step": 14078
},
{
"epoch": 0.8533333333333334,
"grad_norm": 0.057372674345970154,
"learning_rate": 0.0004890909090909091,
"loss": 3.000248670578003,
"step": 14080
},
{
"epoch": 0.8534545454545455,
"grad_norm": 0.05690908059477806,
"learning_rate": 0.0004886868686868688,
"loss": 3.034821033477783,
"step": 14082
},
{
"epoch": 0.8535757575757575,
"grad_norm": 0.05391664430499077,
"learning_rate": 0.0004882828282828283,
"loss": 3.0639889240264893,
"step": 14084
},
{
"epoch": 0.8536969696969697,
"grad_norm": 0.04808659851551056,
"learning_rate": 0.0004878787878787879,
"loss": 3.0581259727478027,
"step": 14086
},
{
"epoch": 0.8538181818181818,
"grad_norm": 0.04908743128180504,
"learning_rate": 0.00048747474747474746,
"loss": 3.033648729324341,
"step": 14088
},
{
"epoch": 0.8539393939393939,
"grad_norm": 0.04890848323702812,
"learning_rate": 0.0004870707070707071,
"loss": 3.049720525741577,
"step": 14090
},
{
"epoch": 0.8540606060606061,
"grad_norm": 0.048842381685972214,
"learning_rate": 0.00048666666666666666,
"loss": 3.030374050140381,
"step": 14092
},
{
"epoch": 0.8541818181818182,
"grad_norm": 0.05370254069566727,
"learning_rate": 0.0004862626262626263,
"loss": 3.021219253540039,
"step": 14094
},
{
"epoch": 0.8543030303030303,
"grad_norm": 0.05150708928704262,
"learning_rate": 0.0004858585858585859,
"loss": 3.02569317817688,
"step": 14096
},
{
"epoch": 0.8544242424242424,
"grad_norm": 0.05052366107702255,
"learning_rate": 0.0004854545454545455,
"loss": 2.9965853691101074,
"step": 14098
},
{
"epoch": 0.8545454545454545,
"grad_norm": 0.050053875893354416,
"learning_rate": 0.0004850505050505051,
"loss": 3.020806312561035,
"step": 14100
},
{
"epoch": 0.8546666666666667,
"grad_norm": 0.055880654603242874,
"learning_rate": 0.0004846464646464647,
"loss": 3.0215981006622314,
"step": 14102
},
{
"epoch": 0.8547878787878788,
"grad_norm": 0.07194066047668457,
"learning_rate": 0.0004842424242424243,
"loss": 3.0102107524871826,
"step": 14104
},
{
"epoch": 0.854909090909091,
"grad_norm": 0.059224944561719894,
"learning_rate": 0.00048383838383838387,
"loss": 3.0431272983551025,
"step": 14106
},
{
"epoch": 0.855030303030303,
"grad_norm": 0.05259707570075989,
"learning_rate": 0.0004834343434343435,
"loss": 2.9785783290863037,
"step": 14108
},
{
"epoch": 0.8551515151515151,
"grad_norm": 0.06625409424304962,
"learning_rate": 0.000483030303030303,
"loss": 3.0354232788085938,
"step": 14110
},
{
"epoch": 0.8552727272727273,
"grad_norm": 0.054187219589948654,
"learning_rate": 0.0004826262626262627,
"loss": 3.088959217071533,
"step": 14112
},
{
"epoch": 0.8553939393939394,
"grad_norm": 0.04868467152118683,
"learning_rate": 0.0004822222222222222,
"loss": 3.0511274337768555,
"step": 14114
},
{
"epoch": 0.8555151515151516,
"grad_norm": 0.05694824457168579,
"learning_rate": 0.00048181818181818184,
"loss": 3.006056070327759,
"step": 14116
},
{
"epoch": 0.8556363636363636,
"grad_norm": 0.0860932469367981,
"learning_rate": 0.0004814141414141414,
"loss": 3.024282693862915,
"step": 14118
},
{
"epoch": 0.8557575757575757,
"grad_norm": 0.06356607377529144,
"learning_rate": 0.00048101010101010103,
"loss": 3.052719831466675,
"step": 14120
},
{
"epoch": 0.8558787878787879,
"grad_norm": 0.05089312791824341,
"learning_rate": 0.0004806060606060606,
"loss": 3.054290294647217,
"step": 14122
},
{
"epoch": 0.856,
"grad_norm": 0.047726523131132126,
"learning_rate": 0.0004802020202020202,
"loss": 3.0510847568511963,
"step": 14124
},
{
"epoch": 0.8561212121212122,
"grad_norm": 0.047966498881578445,
"learning_rate": 0.0004797979797979798,
"loss": 3.0386931896209717,
"step": 14126
},
{
"epoch": 0.8562424242424242,
"grad_norm": 0.048524100333452225,
"learning_rate": 0.0004793939393939394,
"loss": 3.0365102291107178,
"step": 14128
},
{
"epoch": 0.8563636363636363,
"grad_norm": 0.048219870775938034,
"learning_rate": 0.00047898989898989905,
"loss": 3.050222396850586,
"step": 14130
},
{
"epoch": 0.8564848484848485,
"grad_norm": 0.04684511199593544,
"learning_rate": 0.0004785858585858586,
"loss": 3.0290489196777344,
"step": 14132
},
{
"epoch": 0.8566060606060606,
"grad_norm": 0.04703833907842636,
"learning_rate": 0.00047818181818181824,
"loss": 3.0699644088745117,
"step": 14134
},
{
"epoch": 0.8567272727272728,
"grad_norm": 0.048106733709573746,
"learning_rate": 0.0004777777777777778,
"loss": 3.0796475410461426,
"step": 14136
},
{
"epoch": 0.8568484848484849,
"grad_norm": 0.049467865377664566,
"learning_rate": 0.00047737373737373744,
"loss": 3.0939600467681885,
"step": 14138
},
{
"epoch": 0.8569696969696969,
"grad_norm": 0.05323538929224014,
"learning_rate": 0.00047696969696969696,
"loss": 3.0743114948272705,
"step": 14140
},
{
"epoch": 0.8570909090909091,
"grad_norm": 0.049413397908210754,
"learning_rate": 0.00047656565656565664,
"loss": 3.037111759185791,
"step": 14142
},
{
"epoch": 0.8572121212121212,
"grad_norm": 0.04853815957903862,
"learning_rate": 0.00047616161616161615,
"loss": 3.0570976734161377,
"step": 14144
},
{
"epoch": 0.8573333333333333,
"grad_norm": 0.05639440938830376,
"learning_rate": 0.0004757575757575758,
"loss": 3.0710415840148926,
"step": 14146
},
{
"epoch": 0.8574545454545455,
"grad_norm": 0.048022348433732986,
"learning_rate": 0.00047535353535353535,
"loss": 3.0675065517425537,
"step": 14148
},
{
"epoch": 0.8575757575757575,
"grad_norm": 0.0487670861184597,
"learning_rate": 0.00047494949494949497,
"loss": 3.0555672645568848,
"step": 14150
},
{
"epoch": 0.8576969696969697,
"grad_norm": 0.05253853276371956,
"learning_rate": 0.00047454545454545454,
"loss": 3.043731927871704,
"step": 14152
},
{
"epoch": 0.8578181818181818,
"grad_norm": 0.06073673069477081,
"learning_rate": 0.00047414141414141417,
"loss": 3.0875244140625,
"step": 14154
},
{
"epoch": 0.8579393939393939,
"grad_norm": 0.04937475174665451,
"learning_rate": 0.00047373737373737374,
"loss": 3.0840706825256348,
"step": 14156
},
{
"epoch": 0.8580606060606061,
"grad_norm": 0.051458872854709625,
"learning_rate": 0.00047333333333333336,
"loss": 3.044126033782959,
"step": 14158
},
{
"epoch": 0.8581818181818182,
"grad_norm": 0.04917445033788681,
"learning_rate": 0.00047292929292929293,
"loss": 3.0815622806549072,
"step": 14160
},
{
"epoch": 0.8583030303030303,
"grad_norm": 0.05183489993214607,
"learning_rate": 0.00047252525252525256,
"loss": 3.0489208698272705,
"step": 14162
},
{
"epoch": 0.8584242424242424,
"grad_norm": 0.05423115938901901,
"learning_rate": 0.0004721212121212122,
"loss": 3.072639226913452,
"step": 14164
},
{
"epoch": 0.8585454545454545,
"grad_norm": 0.04930400848388672,
"learning_rate": 0.00047171717171717176,
"loss": 3.0347485542297363,
"step": 14166
},
{
"epoch": 0.8586666666666667,
"grad_norm": 0.04806545376777649,
"learning_rate": 0.0004713131313131314,
"loss": 3.0974247455596924,
"step": 14168
},
{
"epoch": 0.8587878787878788,
"grad_norm": 0.0506218746304512,
"learning_rate": 0.0004709090909090909,
"loss": 3.055525064468384,
"step": 14170
},
{
"epoch": 0.858909090909091,
"grad_norm": 0.04794139787554741,
"learning_rate": 0.0004705050505050506,
"loss": 3.0689570903778076,
"step": 14172
},
{
"epoch": 0.859030303030303,
"grad_norm": 0.04874967411160469,
"learning_rate": 0.0004701010101010101,
"loss": 3.079979181289673,
"step": 14174
},
{
"epoch": 0.8591515151515151,
"grad_norm": 0.050874173641204834,
"learning_rate": 0.00046969696969696977,
"loss": 3.057882070541382,
"step": 14176
},
{
"epoch": 0.8592727272727273,
"grad_norm": 0.05018075928092003,
"learning_rate": 0.0004692929292929293,
"loss": 3.0484304428100586,
"step": 14178
},
{
"epoch": 0.8593939393939394,
"grad_norm": 0.050635166466236115,
"learning_rate": 0.0004688888888888889,
"loss": 3.0616257190704346,
"step": 14180
},
{
"epoch": 0.8595151515151516,
"grad_norm": 0.05116913840174675,
"learning_rate": 0.0004684848484848485,
"loss": 3.0463883876800537,
"step": 14182
},
{
"epoch": 0.8596363636363636,
"grad_norm": 0.055978842079639435,
"learning_rate": 0.0004680808080808081,
"loss": 3.0257303714752197,
"step": 14184
},
{
"epoch": 0.8597575757575757,
"grad_norm": 0.04804922640323639,
"learning_rate": 0.0004676767676767677,
"loss": 3.0374298095703125,
"step": 14186
},
{
"epoch": 0.8598787878787879,
"grad_norm": 0.05176619812846184,
"learning_rate": 0.0004672727272727273,
"loss": 3.086068868637085,
"step": 14188
},
{
"epoch": 0.86,
"grad_norm": 0.048618387430906296,
"learning_rate": 0.0004668686868686869,
"loss": 3.0548171997070312,
"step": 14190
},
{
"epoch": 0.8601212121212122,
"grad_norm": 0.048329371958971024,
"learning_rate": 0.0004664646464646465,
"loss": 3.047515392303467,
"step": 14192
},
{
"epoch": 0.8602424242424243,
"grad_norm": 0.048230309039354324,
"learning_rate": 0.000466060606060606,
"loss": 3.065047025680542,
"step": 14194
},
{
"epoch": 0.8603636363636363,
"grad_norm": 0.048555806279182434,
"learning_rate": 0.0004656565656565657,
"loss": 3.0464670658111572,
"step": 14196
},
{
"epoch": 0.8604848484848485,
"grad_norm": 0.0473727248609066,
"learning_rate": 0.0004652525252525253,
"loss": 3.0668671131134033,
"step": 14198
},
{
"epoch": 0.8606060606060606,
"grad_norm": 0.04749167710542679,
"learning_rate": 0.0004648484848484849,
"loss": 3.0775859355926514,
"step": 14200
},
{
"epoch": 0.8607272727272727,
"grad_norm": 0.0483962707221508,
"learning_rate": 0.0004644444444444445,
"loss": 3.106336832046509,
"step": 14202
},
{
"epoch": 0.8608484848484849,
"grad_norm": 0.047952305525541306,
"learning_rate": 0.00046404040404040403,
"loss": 3.0785071849823,
"step": 14204
},
{
"epoch": 0.8609696969696969,
"grad_norm": 0.04857010766863823,
"learning_rate": 0.0004636363636363637,
"loss": 3.0775985717773438,
"step": 14206
},
{
"epoch": 0.8610909090909091,
"grad_norm": 0.05767454206943512,
"learning_rate": 0.00046323232323232323,
"loss": 3.0829851627349854,
"step": 14208
},
{
"epoch": 0.8612121212121212,
"grad_norm": 0.07035728543996811,
"learning_rate": 0.00046282828282828286,
"loss": 3.0622942447662354,
"step": 14210
},
{
"epoch": 0.8613333333333333,
"grad_norm": 0.06269580125808716,
"learning_rate": 0.0004624242424242424,
"loss": 3.0741078853607178,
"step": 14212
},
{
"epoch": 0.8614545454545455,
"grad_norm": 0.052601754665374756,
"learning_rate": 0.00046202020202020205,
"loss": 3.0830729007720947,
"step": 14214
},
{
"epoch": 0.8615757575757576,
"grad_norm": 0.04960833489894867,
"learning_rate": 0.0004616161616161616,
"loss": 3.0892605781555176,
"step": 14216
},
{
"epoch": 0.8616969696969697,
"grad_norm": 0.06609844416379929,
"learning_rate": 0.00046121212121212125,
"loss": 3.0906708240509033,
"step": 14218
},
{
"epoch": 0.8618181818181818,
"grad_norm": 0.06039416044950485,
"learning_rate": 0.0004608080808080808,
"loss": 3.058279275894165,
"step": 14220
},
{
"epoch": 0.8619393939393939,
"grad_norm": 0.04888422042131424,
"learning_rate": 0.00046040404040404044,
"loss": 3.0818934440612793,
"step": 14222
},
{
"epoch": 0.8620606060606061,
"grad_norm": 0.05765349045395851,
"learning_rate": 0.00045999999999999996,
"loss": 3.0840189456939697,
"step": 14224
},
{
"epoch": 0.8621818181818182,
"grad_norm": 0.05035440996289253,
"learning_rate": 0.00045959595959595964,
"loss": 3.0892622470855713,
"step": 14226
},
{
"epoch": 0.8623030303030303,
"grad_norm": 0.04852138087153435,
"learning_rate": 0.00045919191919191916,
"loss": 3.1231110095977783,
"step": 14228
},
{
"epoch": 0.8624242424242424,
"grad_norm": 0.04890000447630882,
"learning_rate": 0.00045878787878787883,
"loss": 3.0768234729766846,
"step": 14230
},
{
"epoch": 0.8625454545454545,
"grad_norm": 0.05034164711833,
"learning_rate": 0.00045838383838383846,
"loss": 3.084923505783081,
"step": 14232
},
{
"epoch": 0.8626666666666667,
"grad_norm": 0.04946516826748848,
"learning_rate": 0.000457979797979798,
"loss": 3.08817458152771,
"step": 14234
},
{
"epoch": 0.8627878787878788,
"grad_norm": 0.04944796860218048,
"learning_rate": 0.00045757575757575766,
"loss": 3.0817508697509766,
"step": 14236
},
{
"epoch": 0.862909090909091,
"grad_norm": 0.048757076263427734,
"learning_rate": 0.00045717171717171717,
"loss": 3.1233668327331543,
"step": 14238
},
{
"epoch": 0.863030303030303,
"grad_norm": 0.054858945310115814,
"learning_rate": 0.0004567676767676768,
"loss": 3.0657949447631836,
"step": 14240
},
{
"epoch": 0.8631515151515151,
"grad_norm": 0.05709901079535484,
"learning_rate": 0.00045636363636363637,
"loss": 3.0783495903015137,
"step": 14242
},
{
"epoch": 0.8632727272727273,
"grad_norm": 0.05102752894163132,
"learning_rate": 0.000455959595959596,
"loss": 3.0596976280212402,
"step": 14244
},
{
"epoch": 0.8633939393939394,
"grad_norm": 0.05339802801609039,
"learning_rate": 0.00045555555555555556,
"loss": 3.0908467769622803,
"step": 14246
},
{
"epoch": 0.8635151515151516,
"grad_norm": 0.048612337559461594,
"learning_rate": 0.0004551515151515152,
"loss": 3.0650694370269775,
"step": 14248
},
{
"epoch": 0.8636363636363636,
"grad_norm": 0.046396542340517044,
"learning_rate": 0.00045474747474747476,
"loss": 3.085505485534668,
"step": 14250
},
{
"epoch": 0.8637575757575757,
"grad_norm": 0.05285479128360748,
"learning_rate": 0.0004543434343434344,
"loss": 3.1239333152770996,
"step": 14252
},
{
"epoch": 0.8638787878787879,
"grad_norm": 0.06442953646183014,
"learning_rate": 0.00045393939393939396,
"loss": 3.0927836894989014,
"step": 14254
},
{
"epoch": 0.864,
"grad_norm": 0.06545699387788773,
"learning_rate": 0.0004535353535353536,
"loss": 3.0815954208374023,
"step": 14256
},
{
"epoch": 0.8641212121212121,
"grad_norm": 0.05061618238687515,
"learning_rate": 0.0004531313131313131,
"loss": 3.105544328689575,
"step": 14258
},
{
"epoch": 0.8642424242424243,
"grad_norm": 0.04854736849665642,
"learning_rate": 0.0004527272727272728,
"loss": 3.0938878059387207,
"step": 14260
},
{
"epoch": 0.8643636363636363,
"grad_norm": 0.05057436600327492,
"learning_rate": 0.0004523232323232323,
"loss": 3.105086088180542,
"step": 14262
},
{
"epoch": 0.8644848484848485,
"grad_norm": 0.06721238791942596,
"learning_rate": 0.0004519191919191919,
"loss": 3.095231533050537,
"step": 14264
},
{
"epoch": 0.8646060606060606,
"grad_norm": 0.08103131502866745,
"learning_rate": 0.0004515151515151516,
"loss": 3.075529098510742,
"step": 14266
},
{
"epoch": 0.8647272727272727,
"grad_norm": 0.06079642474651337,
"learning_rate": 0.0004511111111111111,
"loss": 3.1172571182250977,
"step": 14268
},
{
"epoch": 0.8648484848484849,
"grad_norm": 0.047984909266233444,
"learning_rate": 0.0004507070707070708,
"loss": 3.0895543098449707,
"step": 14270
},
{
"epoch": 0.8649696969696969,
"grad_norm": 0.049492012709379196,
"learning_rate": 0.0004503030303030303,
"loss": 3.064901351928711,
"step": 14272
},
{
"epoch": 0.8650909090909091,
"grad_norm": 0.056201715022325516,
"learning_rate": 0.00044989898989898993,
"loss": 3.0901715755462646,
"step": 14274
},
{
"epoch": 0.8652121212121212,
"grad_norm": 0.04709451273083687,
"learning_rate": 0.0004494949494949495,
"loss": 3.0738351345062256,
"step": 14276
},
{
"epoch": 0.8653333333333333,
"grad_norm": 0.05263134092092514,
"learning_rate": 0.00044909090909090913,
"loss": 3.0654287338256836,
"step": 14278
},
{
"epoch": 0.8654545454545455,
"grad_norm": 0.05986723303794861,
"learning_rate": 0.0004486868686868687,
"loss": 3.104398012161255,
"step": 14280
},
{
"epoch": 0.8655757575757576,
"grad_norm": 0.05638425797224045,
"learning_rate": 0.0004482828282828283,
"loss": 3.13173508644104,
"step": 14282
},
{
"epoch": 0.8656969696969697,
"grad_norm": 0.05201071500778198,
"learning_rate": 0.0004478787878787879,
"loss": 3.138103485107422,
"step": 14284
},
{
"epoch": 0.8658181818181818,
"grad_norm": 0.04846996068954468,
"learning_rate": 0.0004474747474747475,
"loss": 3.0718882083892822,
"step": 14286
},
{
"epoch": 0.8659393939393939,
"grad_norm": 0.06036893278360367,
"learning_rate": 0.00044707070707070704,
"loss": 3.1248111724853516,
"step": 14288
},
{
"epoch": 0.8660606060606061,
"grad_norm": 0.04696693643927574,
"learning_rate": 0.0004466666666666667,
"loss": 3.078082799911499,
"step": 14290
},
{
"epoch": 0.8661818181818182,
"grad_norm": 0.0481564886868,
"learning_rate": 0.00044626262626262623,
"loss": 3.106235980987549,
"step": 14292
},
{
"epoch": 0.8663030303030304,
"grad_norm": 0.05229068920016289,
"learning_rate": 0.00044585858585858586,
"loss": 3.121110439300537,
"step": 14294
},
{
"epoch": 0.8664242424242424,
"grad_norm": 0.048244211822748184,
"learning_rate": 0.00044545454545454543,
"loss": 3.0963797569274902,
"step": 14296
},
{
"epoch": 0.8665454545454545,
"grad_norm": 0.04844045639038086,
"learning_rate": 0.00044505050505050506,
"loss": 3.0875866413116455,
"step": 14298
},
{
"epoch": 0.8666666666666667,
"grad_norm": 0.04815573990345001,
"learning_rate": 0.00044464646464646473,
"loss": 3.1057851314544678,
"step": 14300
},
{
"epoch": 0.8667878787878788,
"grad_norm": 0.04779226332902908,
"learning_rate": 0.00044424242424242425,
"loss": 3.081930160522461,
"step": 14302
},
{
"epoch": 0.866909090909091,
"grad_norm": 0.04932711645960808,
"learning_rate": 0.0004438383838383839,
"loss": 3.0604248046875,
"step": 14304
},
{
"epoch": 0.867030303030303,
"grad_norm": 0.05191284045577049,
"learning_rate": 0.00044343434343434345,
"loss": 3.1290481090545654,
"step": 14306
},
{
"epoch": 0.8671515151515151,
"grad_norm": 0.05566756799817085,
"learning_rate": 0.00044303030303030307,
"loss": 3.1240789890289307,
"step": 14308
},
{
"epoch": 0.8672727272727273,
"grad_norm": 0.05124182254076004,
"learning_rate": 0.00044262626262626264,
"loss": 3.0563087463378906,
"step": 14310
},
{
"epoch": 0.8673939393939394,
"grad_norm": 0.0487561896443367,
"learning_rate": 0.00044222222222222227,
"loss": 3.109678030014038,
"step": 14312
},
{
"epoch": 0.8675151515151515,
"grad_norm": 0.05559564754366875,
"learning_rate": 0.00044181818181818184,
"loss": 3.1110265254974365,
"step": 14314
},
{
"epoch": 0.8676363636363637,
"grad_norm": 0.06434163451194763,
"learning_rate": 0.00044141414141414146,
"loss": 3.0438990592956543,
"step": 14316
},
{
"epoch": 0.8677575757575757,
"grad_norm": 0.05155486240983009,
"learning_rate": 0.000441010101010101,
"loss": 3.1306979656219482,
"step": 14318
},
{
"epoch": 0.8678787878787879,
"grad_norm": 0.04756059870123863,
"learning_rate": 0.00044060606060606066,
"loss": 3.096900701522827,
"step": 14320
},
{
"epoch": 0.868,
"grad_norm": 0.05606486648321152,
"learning_rate": 0.0004402020202020202,
"loss": 3.113553524017334,
"step": 14322
},
{
"epoch": 0.8681212121212121,
"grad_norm": 0.05720949172973633,
"learning_rate": 0.00043979797979797986,
"loss": 3.103661060333252,
"step": 14324
},
{
"epoch": 0.8682424242424243,
"grad_norm": 0.047940995544195175,
"learning_rate": 0.00043939393939393937,
"loss": 3.1127705574035645,
"step": 14326
},
{
"epoch": 0.8683636363636363,
"grad_norm": 0.05042030289769173,
"learning_rate": 0.000438989898989899,
"loss": 3.120417356491089,
"step": 14328
},
{
"epoch": 0.8684848484848485,
"grad_norm": 0.06065693497657776,
"learning_rate": 0.00043858585858585857,
"loss": 3.1182596683502197,
"step": 14330
},
{
"epoch": 0.8686060606060606,
"grad_norm": 0.05794747918844223,
"learning_rate": 0.0004381818181818182,
"loss": 3.1336615085601807,
"step": 14332
},
{
"epoch": 0.8687272727272727,
"grad_norm": 0.04888366907835007,
"learning_rate": 0.0004377777777777778,
"loss": 3.098433256149292,
"step": 14334
},
{
"epoch": 0.8688484848484849,
"grad_norm": 0.06989632546901703,
"learning_rate": 0.0004373737373737374,
"loss": 3.112722158432007,
"step": 14336
},
{
"epoch": 0.868969696969697,
"grad_norm": 0.07578034698963165,
"learning_rate": 0.000436969696969697,
"loss": 3.1008095741271973,
"step": 14338
},
{
"epoch": 0.8690909090909091,
"grad_norm": 0.05043797567486763,
"learning_rate": 0.0004365656565656566,
"loss": 3.1012699604034424,
"step": 14340
},
{
"epoch": 0.8692121212121212,
"grad_norm": 0.049519263207912445,
"learning_rate": 0.0004361616161616162,
"loss": 3.125669002532959,
"step": 14342
},
{
"epoch": 0.8693333333333333,
"grad_norm": 0.04965291917324066,
"learning_rate": 0.0004357575757575758,
"loss": 3.0972909927368164,
"step": 14344
},
{
"epoch": 0.8694545454545455,
"grad_norm": 0.0476110465824604,
"learning_rate": 0.0004353535353535354,
"loss": 3.0835630893707275,
"step": 14346
},
{
"epoch": 0.8695757575757576,
"grad_norm": 0.047395046800374985,
"learning_rate": 0.000434949494949495,
"loss": 3.0941946506500244,
"step": 14348
},
{
"epoch": 0.8696969696969697,
"grad_norm": 0.04748532921075821,
"learning_rate": 0.0004345454545454546,
"loss": 3.1063482761383057,
"step": 14350
},
{
"epoch": 0.8698181818181818,
"grad_norm": 0.051559459418058395,
"learning_rate": 0.0004341414141414141,
"loss": 3.1141090393066406,
"step": 14352
},
{
"epoch": 0.8699393939393939,
"grad_norm": 0.053282372653484344,
"learning_rate": 0.0004337373737373738,
"loss": 3.0965707302093506,
"step": 14354
},
{
"epoch": 0.8700606060606061,
"grad_norm": 0.049667343497276306,
"learning_rate": 0.0004333333333333333,
"loss": 3.1317977905273438,
"step": 14356
},
{
"epoch": 0.8701818181818182,
"grad_norm": 0.04885130375623703,
"learning_rate": 0.00043292929292929294,
"loss": 3.084899663925171,
"step": 14358
},
{
"epoch": 0.8703030303030304,
"grad_norm": 0.05884130671620369,
"learning_rate": 0.0004325252525252525,
"loss": 3.0836400985717773,
"step": 14360
},
{
"epoch": 0.8704242424242424,
"grad_norm": 0.055444955825805664,
"learning_rate": 0.00043212121212121213,
"loss": 3.09928035736084,
"step": 14362
},
{
"epoch": 0.8705454545454545,
"grad_norm": 0.049598708748817444,
"learning_rate": 0.0004317171717171717,
"loss": 3.1215620040893555,
"step": 14364
},
{
"epoch": 0.8706666666666667,
"grad_norm": 0.04817265272140503,
"learning_rate": 0.00043131313131313133,
"loss": 3.107572317123413,
"step": 14366
},
{
"epoch": 0.8707878787878788,
"grad_norm": 0.050719328224658966,
"learning_rate": 0.00043090909090909096,
"loss": 3.1385934352874756,
"step": 14368
},
{
"epoch": 0.8709090909090909,
"grad_norm": 0.04804671183228493,
"learning_rate": 0.0004305050505050505,
"loss": 3.1016740798950195,
"step": 14370
},
{
"epoch": 0.871030303030303,
"grad_norm": 0.05319105461239815,
"learning_rate": 0.00043010101010101015,
"loss": 3.1105642318725586,
"step": 14372
},
{
"epoch": 0.8711515151515151,
"grad_norm": 0.05394924059510231,
"learning_rate": 0.0004296969696969697,
"loss": 3.133030414581299,
"step": 14374
},
{
"epoch": 0.8712727272727273,
"grad_norm": 0.04975232854485512,
"learning_rate": 0.00042929292929292935,
"loss": 3.157564878463745,
"step": 14376
},
{
"epoch": 0.8713939393939394,
"grad_norm": 0.04949374869465828,
"learning_rate": 0.0004288888888888889,
"loss": 3.1454038619995117,
"step": 14378
},
{
"epoch": 0.8715151515151515,
"grad_norm": 0.051104962825775146,
"learning_rate": 0.00042848484848484854,
"loss": 3.1243438720703125,
"step": 14380
},
{
"epoch": 0.8716363636363637,
"grad_norm": 0.052867595106363297,
"learning_rate": 0.00042808080808080806,
"loss": 3.112565755844116,
"step": 14382
},
{
"epoch": 0.8717575757575757,
"grad_norm": 0.05279192328453064,
"learning_rate": 0.00042767676767676774,
"loss": 3.1319780349731445,
"step": 14384
},
{
"epoch": 0.8718787878787879,
"grad_norm": 0.05061368644237518,
"learning_rate": 0.00042727272727272726,
"loss": 3.132542610168457,
"step": 14386
},
{
"epoch": 0.872,
"grad_norm": 0.04877034202218056,
"learning_rate": 0.0004268686868686869,
"loss": 3.0952179431915283,
"step": 14388
},
{
"epoch": 0.8721212121212121,
"grad_norm": 0.04686637967824936,
"learning_rate": 0.00042646464646464645,
"loss": 3.1430771350860596,
"step": 14390
},
{
"epoch": 0.8722424242424243,
"grad_norm": 0.04827306419610977,
"learning_rate": 0.0004260606060606061,
"loss": 3.0911355018615723,
"step": 14392
},
{
"epoch": 0.8723636363636363,
"grad_norm": 0.04818299040198326,
"learning_rate": 0.00042565656565656565,
"loss": 3.116142988204956,
"step": 14394
},
{
"epoch": 0.8724848484848485,
"grad_norm": 0.056792519986629486,
"learning_rate": 0.00042525252525252527,
"loss": 3.0860862731933594,
"step": 14396
},
{
"epoch": 0.8726060606060606,
"grad_norm": 0.05939554050564766,
"learning_rate": 0.00042484848484848484,
"loss": 3.0970396995544434,
"step": 14398
},
{
"epoch": 0.8727272727272727,
"grad_norm": 0.059091705828905106,
"learning_rate": 0.00042444444444444447,
"loss": 3.1142578125,
"step": 14400
},
{
"epoch": 0.8728484848484849,
"grad_norm": 0.05388859659433365,
"learning_rate": 0.0004240404040404041,
"loss": 3.0991039276123047,
"step": 14402
},
{
"epoch": 0.872969696969697,
"grad_norm": 0.05153131112456322,
"learning_rate": 0.00042363636363636366,
"loss": 3.076970100402832,
"step": 14404
},
{
"epoch": 0.8730909090909091,
"grad_norm": 0.05043292045593262,
"learning_rate": 0.0004232323232323233,
"loss": 3.1175081729888916,
"step": 14406
},
{
"epoch": 0.8732121212121212,
"grad_norm": 0.05058962106704712,
"learning_rate": 0.00042282828282828286,
"loss": 3.107888698577881,
"step": 14408
},
{
"epoch": 0.8733333333333333,
"grad_norm": 0.0482368990778923,
"learning_rate": 0.0004224242424242425,
"loss": 3.0797479152679443,
"step": 14410
},
{
"epoch": 0.8734545454545455,
"grad_norm": 0.04754696413874626,
"learning_rate": 0.000422020202020202,
"loss": 3.124319553375244,
"step": 14412
},
{
"epoch": 0.8735757575757576,
"grad_norm": 0.04978350177407265,
"learning_rate": 0.0004216161616161617,
"loss": 3.16721773147583,
"step": 14414
},
{
"epoch": 0.8736969696969697,
"grad_norm": 0.04862650856375694,
"learning_rate": 0.0004212121212121212,
"loss": 3.131146192550659,
"step": 14416
},
{
"epoch": 0.8738181818181818,
"grad_norm": 0.04800015315413475,
"learning_rate": 0.0004208080808080809,
"loss": 3.1178646087646484,
"step": 14418
},
{
"epoch": 0.8739393939393939,
"grad_norm": 0.05150672793388367,
"learning_rate": 0.0004204040404040404,
"loss": 3.1286840438842773,
"step": 14420
},
{
"epoch": 0.8740606060606061,
"grad_norm": 0.04756385087966919,
"learning_rate": 0.00042,
"loss": 3.1113884449005127,
"step": 14422
},
{
"epoch": 0.8741818181818182,
"grad_norm": 0.05197514593601227,
"learning_rate": 0.0004195959595959596,
"loss": 3.099468946456909,
"step": 14424
},
{
"epoch": 0.8743030303030302,
"grad_norm": 0.05098631605505943,
"learning_rate": 0.0004191919191919192,
"loss": 3.0918476581573486,
"step": 14426
},
{
"epoch": 0.8744242424242424,
"grad_norm": 0.05679531767964363,
"learning_rate": 0.0004187878787878788,
"loss": 3.1330161094665527,
"step": 14428
},
{
"epoch": 0.8745454545454545,
"grad_norm": 0.05435401573777199,
"learning_rate": 0.0004183838383838384,
"loss": 3.118414878845215,
"step": 14430
},
{
"epoch": 0.8746666666666667,
"grad_norm": 0.04748748987913132,
"learning_rate": 0.000417979797979798,
"loss": 3.099403142929077,
"step": 14432
},
{
"epoch": 0.8747878787878788,
"grad_norm": 0.0512390099465847,
"learning_rate": 0.0004175757575757576,
"loss": 3.143224000930786,
"step": 14434
},
{
"epoch": 0.8749090909090909,
"grad_norm": 0.050826698541641235,
"learning_rate": 0.00041717171717171723,
"loss": 3.145782947540283,
"step": 14436
},
{
"epoch": 0.875030303030303,
"grad_norm": 0.05362631380558014,
"learning_rate": 0.0004167676767676768,
"loss": 3.1673269271850586,
"step": 14438
},
{
"epoch": 0.8751515151515151,
"grad_norm": 0.06050999090075493,
"learning_rate": 0.0004163636363636364,
"loss": 3.0874602794647217,
"step": 14440
},
{
"epoch": 0.8752727272727273,
"grad_norm": 0.04988927021622658,
"learning_rate": 0.000415959595959596,
"loss": 3.107578754425049,
"step": 14442
},
{
"epoch": 0.8753939393939394,
"grad_norm": 0.05334986373782158,
"learning_rate": 0.0004155555555555556,
"loss": 3.12825083732605,
"step": 14444
},
{
"epoch": 0.8755151515151515,
"grad_norm": 0.07173621654510498,
"learning_rate": 0.00041515151515151514,
"loss": 3.143873929977417,
"step": 14446
},
{
"epoch": 0.8756363636363637,
"grad_norm": 0.0505019910633564,
"learning_rate": 0.0004147474747474748,
"loss": 3.101364850997925,
"step": 14448
},
{
"epoch": 0.8757575757575757,
"grad_norm": 0.04748190566897392,
"learning_rate": 0.00041434343434343433,
"loss": 3.123584508895874,
"step": 14450
},
{
"epoch": 0.8758787878787879,
"grad_norm": 0.05395134538412094,
"learning_rate": 0.00041393939393939396,
"loss": 3.1204617023468018,
"step": 14452
},
{
"epoch": 0.876,
"grad_norm": 0.05916650593280792,
"learning_rate": 0.00041353535353535353,
"loss": 3.1013619899749756,
"step": 14454
},
{
"epoch": 0.8761212121212121,
"grad_norm": 0.05472126230597496,
"learning_rate": 0.00041313131313131315,
"loss": 3.1222281455993652,
"step": 14456
},
{
"epoch": 0.8762424242424243,
"grad_norm": 0.050620146095752716,
"learning_rate": 0.0004127272727272727,
"loss": 3.152968406677246,
"step": 14458
},
{
"epoch": 0.8763636363636363,
"grad_norm": 0.06137436628341675,
"learning_rate": 0.00041232323232323235,
"loss": 3.1556973457336426,
"step": 14460
},
{
"epoch": 0.8764848484848485,
"grad_norm": 0.056075066328048706,
"learning_rate": 0.0004119191919191919,
"loss": 3.124094009399414,
"step": 14462
},
{
"epoch": 0.8766060606060606,
"grad_norm": 0.048165950924158096,
"learning_rate": 0.00041151515151515155,
"loss": 3.1225314140319824,
"step": 14464
},
{
"epoch": 0.8767272727272727,
"grad_norm": 0.05003613233566284,
"learning_rate": 0.0004111111111111111,
"loss": 3.1077566146850586,
"step": 14466
},
{
"epoch": 0.8768484848484849,
"grad_norm": 0.06120872497558594,
"learning_rate": 0.00041070707070707074,
"loss": 3.117555856704712,
"step": 14468
},
{
"epoch": 0.876969696969697,
"grad_norm": 0.05743727087974548,
"learning_rate": 0.00041030303030303037,
"loss": 3.100450277328491,
"step": 14470
},
{
"epoch": 0.8770909090909091,
"grad_norm": 0.04792417958378792,
"learning_rate": 0.00040989898989898994,
"loss": 3.137329578399658,
"step": 14472
},
{
"epoch": 0.8772121212121212,
"grad_norm": 0.05935823544859886,
"learning_rate": 0.00040949494949494956,
"loss": 3.116584300994873,
"step": 14474
},
{
"epoch": 0.8773333333333333,
"grad_norm": 0.047442998737096786,
"learning_rate": 0.0004090909090909091,
"loss": 3.1433331966400146,
"step": 14476
},
{
"epoch": 0.8774545454545455,
"grad_norm": 0.0534931980073452,
"learning_rate": 0.00040868686868686876,
"loss": 3.0997474193573,
"step": 14478
},
{
"epoch": 0.8775757575757576,
"grad_norm": 0.05021868646144867,
"learning_rate": 0.0004082828282828283,
"loss": 3.1593403816223145,
"step": 14480
},
{
"epoch": 0.8776969696969696,
"grad_norm": 0.051341790705919266,
"learning_rate": 0.0004078787878787879,
"loss": 3.1253855228424072,
"step": 14482
},
{
"epoch": 0.8778181818181818,
"grad_norm": 0.048263732343912125,
"learning_rate": 0.00040747474747474747,
"loss": 3.132117986679077,
"step": 14484
},
{
"epoch": 0.8779393939393939,
"grad_norm": 0.048316583037376404,
"learning_rate": 0.0004070707070707071,
"loss": 3.1419975757598877,
"step": 14486
},
{
"epoch": 0.8780606060606061,
"grad_norm": 0.04913982376456261,
"learning_rate": 0.00040666666666666667,
"loss": 3.137758731842041,
"step": 14488
},
{
"epoch": 0.8781818181818182,
"grad_norm": 0.05202285200357437,
"learning_rate": 0.0004062626262626263,
"loss": 3.1318445205688477,
"step": 14490
},
{
"epoch": 0.8783030303030303,
"grad_norm": 0.05039643868803978,
"learning_rate": 0.00040585858585858586,
"loss": 3.1420741081237793,
"step": 14492
},
{
"epoch": 0.8784242424242424,
"grad_norm": 0.04848403483629227,
"learning_rate": 0.0004054545454545455,
"loss": 3.127901077270508,
"step": 14494
},
{
"epoch": 0.8785454545454545,
"grad_norm": 0.05021706596016884,
"learning_rate": 0.00040505050505050506,
"loss": 3.1265969276428223,
"step": 14496
},
{
"epoch": 0.8786666666666667,
"grad_norm": 0.048826102167367935,
"learning_rate": 0.0004046464646464647,
"loss": 3.1101958751678467,
"step": 14498
},
{
"epoch": 0.8787878787878788,
"grad_norm": 0.04896646365523338,
"learning_rate": 0.0004042424242424242,
"loss": 3.1430907249450684,
"step": 14500
},
{
"epoch": 0.8789090909090909,
"grad_norm": 0.05035965517163277,
"learning_rate": 0.0004038383838383839,
"loss": 3.127858877182007,
"step": 14502
},
{
"epoch": 0.879030303030303,
"grad_norm": 0.0482577309012413,
"learning_rate": 0.0004034343434343435,
"loss": 3.152294158935547,
"step": 14504
},
{
"epoch": 0.8791515151515151,
"grad_norm": 0.05237266793847084,
"learning_rate": 0.000403030303030303,
"loss": 3.1277031898498535,
"step": 14506
},
{
"epoch": 0.8792727272727273,
"grad_norm": 0.07222101092338562,
"learning_rate": 0.0004026262626262627,
"loss": 3.085183620452881,
"step": 14508
},
{
"epoch": 0.8793939393939394,
"grad_norm": 0.04892655089497566,
"learning_rate": 0.0004022222222222222,
"loss": 3.1333861351013184,
"step": 14510
},
{
"epoch": 0.8795151515151515,
"grad_norm": 0.055462729185819626,
"learning_rate": 0.0004018181818181819,
"loss": 3.1259708404541016,
"step": 14512
},
{
"epoch": 0.8796363636363637,
"grad_norm": 0.07406949251890182,
"learning_rate": 0.0004014141414141414,
"loss": 3.1538639068603516,
"step": 14514
},
{
"epoch": 0.8797575757575757,
"grad_norm": 0.05291714519262314,
"learning_rate": 0.00040101010101010104,
"loss": 3.1534805297851562,
"step": 14516
},
{
"epoch": 0.8798787878787879,
"grad_norm": 0.05150260776281357,
"learning_rate": 0.0004006060606060606,
"loss": 3.1182162761688232,
"step": 14518
},
{
"epoch": 0.88,
"grad_norm": 0.050346627831459045,
"learning_rate": 0.00040020202020202023,
"loss": 3.1227304935455322,
"step": 14520
},
{
"epoch": 0.8801212121212121,
"grad_norm": 0.046950727701187134,
"learning_rate": 0.00039979797979797986,
"loss": 3.142589569091797,
"step": 14522
},
{
"epoch": 0.8802424242424243,
"grad_norm": 0.05172426998615265,
"learning_rate": 0.00039939393939393943,
"loss": 3.113618850708008,
"step": 14524
},
{
"epoch": 0.8803636363636363,
"grad_norm": 0.05403667315840721,
"learning_rate": 0.000398989898989899,
"loss": 3.1570541858673096,
"step": 14526
},
{
"epoch": 0.8804848484848485,
"grad_norm": 0.05905061960220337,
"learning_rate": 0.0003985858585858586,
"loss": 3.150233030319214,
"step": 14528
},
{
"epoch": 0.8806060606060606,
"grad_norm": 0.04921606555581093,
"learning_rate": 0.0003981818181818182,
"loss": 3.166177272796631,
"step": 14530
},
{
"epoch": 0.8807272727272727,
"grad_norm": 0.05499904602766037,
"learning_rate": 0.0003977777777777778,
"loss": 3.1458287239074707,
"step": 14532
},
{
"epoch": 0.8808484848484849,
"grad_norm": 0.06083129346370697,
"learning_rate": 0.0003973737373737374,
"loss": 3.130254030227661,
"step": 14534
},
{
"epoch": 0.880969696969697,
"grad_norm": 0.04747677594423294,
"learning_rate": 0.000396969696969697,
"loss": 3.132225513458252,
"step": 14536
},
{
"epoch": 0.881090909090909,
"grad_norm": 0.057498782873153687,
"learning_rate": 0.0003965656565656566,
"loss": 3.1393702030181885,
"step": 14538
},
{
"epoch": 0.8812121212121212,
"grad_norm": 0.055975161492824554,
"learning_rate": 0.00039616161616161616,
"loss": 3.1185452938079834,
"step": 14540
},
{
"epoch": 0.8813333333333333,
"grad_norm": 0.04917009174823761,
"learning_rate": 0.0003957575757575758,
"loss": 3.1341288089752197,
"step": 14542
},
{
"epoch": 0.8814545454545455,
"grad_norm": 0.06865264475345612,
"learning_rate": 0.00039535353535353535,
"loss": 3.1472861766815186,
"step": 14544
},
{
"epoch": 0.8815757575757576,
"grad_norm": 0.0667961984872818,
"learning_rate": 0.000394949494949495,
"loss": 3.123800039291382,
"step": 14546
},
{
"epoch": 0.8816969696969696,
"grad_norm": 0.05587344989180565,
"learning_rate": 0.00039454545454545455,
"loss": 3.087278127670288,
"step": 14548
},
{
"epoch": 0.8818181818181818,
"grad_norm": 0.048479706048965454,
"learning_rate": 0.0003941414141414141,
"loss": 3.15456485748291,
"step": 14550
},
{
"epoch": 0.8819393939393939,
"grad_norm": 0.05084102228283882,
"learning_rate": 0.00039373737373737375,
"loss": 3.1480343341827393,
"step": 14552
},
{
"epoch": 0.8820606060606061,
"grad_norm": 0.04854101315140724,
"learning_rate": 0.0003933333333333333,
"loss": 3.157419204711914,
"step": 14554
},
{
"epoch": 0.8821818181818182,
"grad_norm": 0.0480605773627758,
"learning_rate": 0.000392929292929293,
"loss": 3.1582796573638916,
"step": 14556
},
{
"epoch": 0.8823030303030303,
"grad_norm": 0.04829663410782814,
"learning_rate": 0.00039252525252525257,
"loss": 3.133721113204956,
"step": 14558
},
{
"epoch": 0.8824242424242424,
"grad_norm": 0.04908953234553337,
"learning_rate": 0.00039212121212121214,
"loss": 3.1482441425323486,
"step": 14560
},
{
"epoch": 0.8825454545454545,
"grad_norm": 0.05022772401571274,
"learning_rate": 0.00039171717171717176,
"loss": 3.135152816772461,
"step": 14562
},
{
"epoch": 0.8826666666666667,
"grad_norm": 0.04821702092885971,
"learning_rate": 0.00039131313131313133,
"loss": 3.1416797637939453,
"step": 14564
},
{
"epoch": 0.8827878787878788,
"grad_norm": 0.048291392624378204,
"learning_rate": 0.00039090909090909096,
"loss": 3.130188226699829,
"step": 14566
},
{
"epoch": 0.8829090909090909,
"grad_norm": 0.04851248115301132,
"learning_rate": 0.00039050505050505053,
"loss": 3.148695707321167,
"step": 14568
},
{
"epoch": 0.883030303030303,
"grad_norm": 0.052591472864151,
"learning_rate": 0.0003901010101010101,
"loss": 3.1682283878326416,
"step": 14570
},
{
"epoch": 0.8831515151515151,
"grad_norm": 0.0491965152323246,
"learning_rate": 0.0003896969696969697,
"loss": 3.143075704574585,
"step": 14572
},
{
"epoch": 0.8832727272727273,
"grad_norm": 0.053249720484018326,
"learning_rate": 0.0003892929292929293,
"loss": 3.12931227684021,
"step": 14574
},
{
"epoch": 0.8833939393939394,
"grad_norm": 0.05727974325418472,
"learning_rate": 0.0003888888888888889,
"loss": 3.1473965644836426,
"step": 14576
},
{
"epoch": 0.8835151515151515,
"grad_norm": 0.0491100512444973,
"learning_rate": 0.0003884848484848485,
"loss": 3.1295621395111084,
"step": 14578
},
{
"epoch": 0.8836363636363637,
"grad_norm": 0.05141565203666687,
"learning_rate": 0.00038808080808080806,
"loss": 3.0885863304138184,
"step": 14580
},
{
"epoch": 0.8837575757575757,
"grad_norm": 0.05118267610669136,
"learning_rate": 0.0003876767676767677,
"loss": 3.1629018783569336,
"step": 14582
},
{
"epoch": 0.8838787878787879,
"grad_norm": 0.04757444187998772,
"learning_rate": 0.00038727272727272726,
"loss": 3.1122748851776123,
"step": 14584
},
{
"epoch": 0.884,
"grad_norm": 0.05244014412164688,
"learning_rate": 0.0003868686868686869,
"loss": 3.1470580101013184,
"step": 14586
},
{
"epoch": 0.8841212121212121,
"grad_norm": 0.051347341388463974,
"learning_rate": 0.00038646464646464645,
"loss": 3.1450552940368652,
"step": 14588
},
{
"epoch": 0.8842424242424243,
"grad_norm": 0.0487225167453289,
"learning_rate": 0.0003860606060606061,
"loss": 3.1283743381500244,
"step": 14590
},
{
"epoch": 0.8843636363636364,
"grad_norm": 0.05083749443292618,
"learning_rate": 0.0003856565656565657,
"loss": 3.1366159915924072,
"step": 14592
},
{
"epoch": 0.8844848484848484,
"grad_norm": 0.05819185450673103,
"learning_rate": 0.0003852525252525253,
"loss": 3.138800621032715,
"step": 14594
},
{
"epoch": 0.8846060606060606,
"grad_norm": 0.05237838998436928,
"learning_rate": 0.0003848484848484849,
"loss": 3.154013156890869,
"step": 14596
},
{
"epoch": 0.8847272727272727,
"grad_norm": 0.0722350999712944,
"learning_rate": 0.00038444444444444447,
"loss": 3.1253814697265625,
"step": 14598
},
{
"epoch": 0.8848484848484849,
"grad_norm": 0.04862460121512413,
"learning_rate": 0.00038404040404040404,
"loss": 3.124443769454956,
"step": 14600
},
{
"epoch": 0.884969696969697,
"grad_norm": 0.05141684412956238,
"learning_rate": 0.00038363636363636367,
"loss": 3.1371066570281982,
"step": 14602
},
{
"epoch": 0.885090909090909,
"grad_norm": 0.05905822664499283,
"learning_rate": 0.00038323232323232324,
"loss": 3.1154849529266357,
"step": 14604
},
{
"epoch": 0.8852121212121212,
"grad_norm": 0.048293586820364,
"learning_rate": 0.00038282828282828286,
"loss": 3.1478354930877686,
"step": 14606
},
{
"epoch": 0.8853333333333333,
"grad_norm": 0.05992274731397629,
"learning_rate": 0.00038242424242424243,
"loss": 3.16186785697937,
"step": 14608
},
{
"epoch": 0.8854545454545455,
"grad_norm": 0.07360861450433731,
"learning_rate": 0.00038202020202020206,
"loss": 3.1198389530181885,
"step": 14610
},
{
"epoch": 0.8855757575757576,
"grad_norm": 0.048296067863702774,
"learning_rate": 0.00038161616161616163,
"loss": 3.115377426147461,
"step": 14612
},
{
"epoch": 0.8856969696969696,
"grad_norm": 0.06008489429950714,
"learning_rate": 0.0003812121212121212,
"loss": 3.1448230743408203,
"step": 14614
},
{
"epoch": 0.8858181818181818,
"grad_norm": 0.059179384261369705,
"learning_rate": 0.0003808080808080808,
"loss": 3.131953716278076,
"step": 14616
},
{
"epoch": 0.8859393939393939,
"grad_norm": 0.050360653549432755,
"learning_rate": 0.0003804040404040404,
"loss": 3.135280132293701,
"step": 14618
},
{
"epoch": 0.8860606060606061,
"grad_norm": 0.05520104616880417,
"learning_rate": 0.00038,
"loss": 3.124777317047119,
"step": 14620
},
{
"epoch": 0.8861818181818182,
"grad_norm": 0.0484369620680809,
"learning_rate": 0.0003795959595959596,
"loss": 3.1471331119537354,
"step": 14622
},
{
"epoch": 0.8863030303030303,
"grad_norm": 0.056895241141319275,
"learning_rate": 0.0003791919191919192,
"loss": 3.121798276901245,
"step": 14624
},
{
"epoch": 0.8864242424242424,
"grad_norm": 0.04914778098464012,
"learning_rate": 0.00037878787878787884,
"loss": 3.1694371700286865,
"step": 14626
},
{
"epoch": 0.8865454545454545,
"grad_norm": 0.04891003295779228,
"learning_rate": 0.0003783838383838384,
"loss": 3.150101661682129,
"step": 14628
},
{
"epoch": 0.8866666666666667,
"grad_norm": 0.052677564322948456,
"learning_rate": 0.00037797979797979804,
"loss": 3.1375200748443604,
"step": 14630
},
{
"epoch": 0.8867878787878788,
"grad_norm": 0.048272211104631424,
"learning_rate": 0.0003775757575757576,
"loss": 3.104398250579834,
"step": 14632
},
{
"epoch": 0.8869090909090909,
"grad_norm": 0.04832189157605171,
"learning_rate": 0.0003771717171717172,
"loss": 3.1449246406555176,
"step": 14634
},
{
"epoch": 0.8870303030303031,
"grad_norm": 0.04905533790588379,
"learning_rate": 0.0003767676767676768,
"loss": 3.1707398891448975,
"step": 14636
},
{
"epoch": 0.8871515151515151,
"grad_norm": 0.05004631355404854,
"learning_rate": 0.0003763636363636364,
"loss": 3.1472184658050537,
"step": 14638
},
{
"epoch": 0.8872727272727273,
"grad_norm": 0.048555128276348114,
"learning_rate": 0.000375959595959596,
"loss": 3.199464797973633,
"step": 14640
},
{
"epoch": 0.8873939393939394,
"grad_norm": 0.04904160648584366,
"learning_rate": 0.00037555555555555557,
"loss": 3.169489622116089,
"step": 14642
},
{
"epoch": 0.8875151515151515,
"grad_norm": 0.04741702601313591,
"learning_rate": 0.00037515151515151514,
"loss": 3.1289541721343994,
"step": 14644
},
{
"epoch": 0.8876363636363637,
"grad_norm": 0.0469234399497509,
"learning_rate": 0.00037474747474747477,
"loss": 3.1238415241241455,
"step": 14646
},
{
"epoch": 0.8877575757575757,
"grad_norm": 0.056924015283584595,
"learning_rate": 0.00037434343434343434,
"loss": 3.175884246826172,
"step": 14648
},
{
"epoch": 0.8878787878787879,
"grad_norm": 0.04910162836313248,
"learning_rate": 0.00037393939393939396,
"loss": 3.160053014755249,
"step": 14650
},
{
"epoch": 0.888,
"grad_norm": 0.04891703650355339,
"learning_rate": 0.00037353535353535353,
"loss": 3.1315040588378906,
"step": 14652
},
{
"epoch": 0.8881212121212121,
"grad_norm": 0.05271385237574577,
"learning_rate": 0.00037313131313131316,
"loss": 3.1403205394744873,
"step": 14654
},
{
"epoch": 0.8882424242424243,
"grad_norm": 0.05186305567622185,
"learning_rate": 0.00037272727272727273,
"loss": 3.132459878921509,
"step": 14656
},
{
"epoch": 0.8883636363636364,
"grad_norm": 0.050538260489702225,
"learning_rate": 0.00037232323232323235,
"loss": 3.1314585208892822,
"step": 14658
},
{
"epoch": 0.8884848484848484,
"grad_norm": 0.04907791316509247,
"learning_rate": 0.000371919191919192,
"loss": 3.1390717029571533,
"step": 14660
},
{
"epoch": 0.8886060606060606,
"grad_norm": 0.04969751834869385,
"learning_rate": 0.00037151515151515155,
"loss": 3.1516451835632324,
"step": 14662
},
{
"epoch": 0.8887272727272727,
"grad_norm": 0.04764828458428383,
"learning_rate": 0.0003711111111111111,
"loss": 3.105785608291626,
"step": 14664
},
{
"epoch": 0.8888484848484849,
"grad_norm": 0.06800170242786407,
"learning_rate": 0.00037070707070707075,
"loss": 3.156597852706909,
"step": 14666
},
{
"epoch": 0.888969696969697,
"grad_norm": 0.06164667755365372,
"learning_rate": 0.0003703030303030303,
"loss": 3.1290793418884277,
"step": 14668
},
{
"epoch": 0.889090909090909,
"grad_norm": 0.0488666296005249,
"learning_rate": 0.00036989898989898994,
"loss": 3.1106202602386475,
"step": 14670
},
{
"epoch": 0.8892121212121212,
"grad_norm": 0.04789494350552559,
"learning_rate": 0.0003694949494949495,
"loss": 3.1284635066986084,
"step": 14672
},
{
"epoch": 0.8893333333333333,
"grad_norm": 0.04961742088198662,
"learning_rate": 0.0003690909090909091,
"loss": 3.1384754180908203,
"step": 14674
},
{
"epoch": 0.8894545454545455,
"grad_norm": 0.04810045287013054,
"learning_rate": 0.0003686868686868687,
"loss": 3.1665546894073486,
"step": 14676
},
{
"epoch": 0.8895757575757576,
"grad_norm": 0.0478670634329319,
"learning_rate": 0.0003682828282828283,
"loss": 3.148791790008545,
"step": 14678
},
{
"epoch": 0.8896969696969697,
"grad_norm": 0.04977035149931908,
"learning_rate": 0.0003678787878787879,
"loss": 3.1468820571899414,
"step": 14680
},
{
"epoch": 0.8898181818181818,
"grad_norm": 0.059104740619659424,
"learning_rate": 0.0003674747474747475,
"loss": 3.152097225189209,
"step": 14682
},
{
"epoch": 0.8899393939393939,
"grad_norm": 0.048677340149879456,
"learning_rate": 0.0003670707070707071,
"loss": 3.1145591735839844,
"step": 14684
},
{
"epoch": 0.8900606060606061,
"grad_norm": 0.053980402648448944,
"learning_rate": 0.00036666666666666667,
"loss": 3.1200554370880127,
"step": 14686
},
{
"epoch": 0.8901818181818182,
"grad_norm": 0.053141605108976364,
"learning_rate": 0.00036626262626262624,
"loss": 3.1547632217407227,
"step": 14688
},
{
"epoch": 0.8903030303030303,
"grad_norm": 0.04921063780784607,
"learning_rate": 0.00036585858585858587,
"loss": 3.179555892944336,
"step": 14690
},
{
"epoch": 0.8904242424242425,
"grad_norm": 0.051950618624687195,
"learning_rate": 0.0003654545454545455,
"loss": 3.111675977706909,
"step": 14692
},
{
"epoch": 0.8905454545454545,
"grad_norm": 0.0495004765689373,
"learning_rate": 0.00036505050505050506,
"loss": 3.1178464889526367,
"step": 14694
},
{
"epoch": 0.8906666666666667,
"grad_norm": 0.04816634580492973,
"learning_rate": 0.0003646464646464647,
"loss": 3.156708002090454,
"step": 14696
},
{
"epoch": 0.8907878787878788,
"grad_norm": 0.05400200933218002,
"learning_rate": 0.00036424242424242426,
"loss": 3.141239881515503,
"step": 14698
},
{
"epoch": 0.8909090909090909,
"grad_norm": 0.05182170495390892,
"learning_rate": 0.0003638383838383839,
"loss": 3.1469502449035645,
"step": 14700
},
{
"epoch": 0.8910303030303031,
"grad_norm": 0.048622891306877136,
"learning_rate": 0.00036343434343434345,
"loss": 3.1774539947509766,
"step": 14702
},
{
"epoch": 0.8911515151515151,
"grad_norm": 0.050085969269275665,
"learning_rate": 0.0003630303030303031,
"loss": 3.1324546337127686,
"step": 14704
},
{
"epoch": 0.8912727272727273,
"grad_norm": 0.051776353269815445,
"learning_rate": 0.00036262626262626265,
"loss": 3.1446077823638916,
"step": 14706
},
{
"epoch": 0.8913939393939394,
"grad_norm": 0.048235274851322174,
"learning_rate": 0.0003622222222222222,
"loss": 3.18117094039917,
"step": 14708
},
{
"epoch": 0.8915151515151515,
"grad_norm": 0.04790253937244415,
"learning_rate": 0.00036181818181818185,
"loss": 3.1617863178253174,
"step": 14710
},
{
"epoch": 0.8916363636363637,
"grad_norm": 0.05286921188235283,
"learning_rate": 0.0003614141414141414,
"loss": 3.137791395187378,
"step": 14712
},
{
"epoch": 0.8917575757575757,
"grad_norm": 0.05688919499516487,
"learning_rate": 0.00036101010101010104,
"loss": 3.0899250507354736,
"step": 14714
},
{
"epoch": 0.8918787878787878,
"grad_norm": 0.05270593985915184,
"learning_rate": 0.0003606060606060606,
"loss": 3.1618034839630127,
"step": 14716
},
{
"epoch": 0.892,
"grad_norm": 0.048660025000572205,
"learning_rate": 0.0003602020202020202,
"loss": 3.169304370880127,
"step": 14718
},
{
"epoch": 0.8921212121212121,
"grad_norm": 0.04864850267767906,
"learning_rate": 0.0003597979797979798,
"loss": 3.1723883152008057,
"step": 14720
},
{
"epoch": 0.8922424242424243,
"grad_norm": 0.04903855174779892,
"learning_rate": 0.0003593939393939394,
"loss": 3.1555261611938477,
"step": 14722
},
{
"epoch": 0.8923636363636364,
"grad_norm": 0.04860149323940277,
"learning_rate": 0.000358989898989899,
"loss": 3.168954610824585,
"step": 14724
},
{
"epoch": 0.8924848484848484,
"grad_norm": 0.05015260726213455,
"learning_rate": 0.00035858585858585863,
"loss": 3.147505521774292,
"step": 14726
},
{
"epoch": 0.8926060606060606,
"grad_norm": 0.048208922147750854,
"learning_rate": 0.0003581818181818182,
"loss": 3.1359670162200928,
"step": 14728
},
{
"epoch": 0.8927272727272727,
"grad_norm": 0.05157000571489334,
"learning_rate": 0.0003577777777777778,
"loss": 3.120028018951416,
"step": 14730
},
{
"epoch": 0.8928484848484849,
"grad_norm": 0.051783133298158646,
"learning_rate": 0.0003573737373737374,
"loss": 3.1221063137054443,
"step": 14732
},
{
"epoch": 0.892969696969697,
"grad_norm": 0.05023888871073723,
"learning_rate": 0.000356969696969697,
"loss": 3.1645193099975586,
"step": 14734
},
{
"epoch": 0.893090909090909,
"grad_norm": 0.04996729642152786,
"learning_rate": 0.0003565656565656566,
"loss": 3.1292057037353516,
"step": 14736
},
{
"epoch": 0.8932121212121212,
"grad_norm": 0.048303622752428055,
"learning_rate": 0.00035616161616161616,
"loss": 3.1818056106567383,
"step": 14738
},
{
"epoch": 0.8933333333333333,
"grad_norm": 0.050007015466690063,
"learning_rate": 0.0003557575757575758,
"loss": 3.12282395362854,
"step": 14740
},
{
"epoch": 0.8934545454545455,
"grad_norm": 0.04808290675282478,
"learning_rate": 0.00035535353535353536,
"loss": 3.1345090866088867,
"step": 14742
},
{
"epoch": 0.8935757575757576,
"grad_norm": 0.04938118904829025,
"learning_rate": 0.000354949494949495,
"loss": 3.1716716289520264,
"step": 14744
},
{
"epoch": 0.8936969696969697,
"grad_norm": 0.04974251613020897,
"learning_rate": 0.00035454545454545455,
"loss": 3.122323513031006,
"step": 14746
},
{
"epoch": 0.8938181818181818,
"grad_norm": 0.05006527155637741,
"learning_rate": 0.0003541414141414142,
"loss": 3.1561777591705322,
"step": 14748
},
{
"epoch": 0.8939393939393939,
"grad_norm": 0.052018482238054276,
"learning_rate": 0.00035373737373737375,
"loss": 3.1707513332366943,
"step": 14750
},
{
"epoch": 0.8940606060606061,
"grad_norm": 0.048764124512672424,
"learning_rate": 0.0003533333333333333,
"loss": 3.139634847640991,
"step": 14752
},
{
"epoch": 0.8941818181818182,
"grad_norm": 0.04988965392112732,
"learning_rate": 0.00035292929292929295,
"loss": 3.146343946456909,
"step": 14754
},
{
"epoch": 0.8943030303030303,
"grad_norm": 0.05299023538827896,
"learning_rate": 0.0003525252525252525,
"loss": 3.1288204193115234,
"step": 14756
},
{
"epoch": 0.8944242424242425,
"grad_norm": 0.05778324231505394,
"learning_rate": 0.00035212121212121214,
"loss": 3.1384453773498535,
"step": 14758
},
{
"epoch": 0.8945454545454545,
"grad_norm": 0.050090864300727844,
"learning_rate": 0.00035171717171717177,
"loss": 3.147407293319702,
"step": 14760
},
{
"epoch": 0.8946666666666667,
"grad_norm": 0.051470328122377396,
"learning_rate": 0.00035131313131313134,
"loss": 3.122305393218994,
"step": 14762
},
{
"epoch": 0.8947878787878788,
"grad_norm": 0.053397320210933685,
"learning_rate": 0.00035090909090909096,
"loss": 3.1127467155456543,
"step": 14764
},
{
"epoch": 0.8949090909090909,
"grad_norm": 0.05050213634967804,
"learning_rate": 0.00035050505050505053,
"loss": 3.1266815662384033,
"step": 14766
},
{
"epoch": 0.8950303030303031,
"grad_norm": 0.05320972204208374,
"learning_rate": 0.0003501010101010101,
"loss": 3.1443276405334473,
"step": 14768
},
{
"epoch": 0.8951515151515151,
"grad_norm": 0.04682239145040512,
"learning_rate": 0.00034969696969696973,
"loss": 3.1656289100646973,
"step": 14770
},
{
"epoch": 0.8952727272727272,
"grad_norm": 0.04868459701538086,
"learning_rate": 0.0003492929292929293,
"loss": 3.1677634716033936,
"step": 14772
},
{
"epoch": 0.8953939393939394,
"grad_norm": 0.05588613823056221,
"learning_rate": 0.0003488888888888889,
"loss": 3.1221566200256348,
"step": 14774
},
{
"epoch": 0.8955151515151515,
"grad_norm": 0.05028705671429634,
"learning_rate": 0.0003484848484848485,
"loss": 3.1178321838378906,
"step": 14776
},
{
"epoch": 0.8956363636363637,
"grad_norm": 0.05022305250167847,
"learning_rate": 0.0003480808080808081,
"loss": 3.1491329669952393,
"step": 14778
},
{
"epoch": 0.8957575757575758,
"grad_norm": 0.04966117814183235,
"learning_rate": 0.0003476767676767677,
"loss": 3.139516830444336,
"step": 14780
},
{
"epoch": 0.8958787878787878,
"grad_norm": 0.04961990565061569,
"learning_rate": 0.00034727272727272726,
"loss": 3.1458778381347656,
"step": 14782
},
{
"epoch": 0.896,
"grad_norm": 0.04894484952092171,
"learning_rate": 0.0003468686868686869,
"loss": 3.1363425254821777,
"step": 14784
},
{
"epoch": 0.8961212121212121,
"grad_norm": 0.048975154757499695,
"learning_rate": 0.00034646464646464646,
"loss": 3.1577069759368896,
"step": 14786
},
{
"epoch": 0.8962424242424243,
"grad_norm": 0.05894167348742485,
"learning_rate": 0.0003460606060606061,
"loss": 3.119248151779175,
"step": 14788
},
{
"epoch": 0.8963636363636364,
"grad_norm": 0.053202610462903976,
"learning_rate": 0.00034565656565656565,
"loss": 3.1482574939727783,
"step": 14790
},
{
"epoch": 0.8964848484848484,
"grad_norm": 0.04980561137199402,
"learning_rate": 0.0003452525252525252,
"loss": 3.085305690765381,
"step": 14792
},
{
"epoch": 0.8966060606060606,
"grad_norm": 0.05013350769877434,
"learning_rate": 0.00034484848484848485,
"loss": 3.1652987003326416,
"step": 14794
},
{
"epoch": 0.8967272727272727,
"grad_norm": 0.05057762563228607,
"learning_rate": 0.0003444444444444445,
"loss": 3.162158489227295,
"step": 14796
},
{
"epoch": 0.8968484848484849,
"grad_norm": 0.05737732723355293,
"learning_rate": 0.0003440404040404041,
"loss": 3.1060080528259277,
"step": 14798
},
{
"epoch": 0.896969696969697,
"grad_norm": 0.07190729677677155,
"learning_rate": 0.00034363636363636367,
"loss": 3.1696269512176514,
"step": 14800
},
{
"epoch": 0.897090909090909,
"grad_norm": 0.06058016046881676,
"learning_rate": 0.00034323232323232324,
"loss": 3.1403374671936035,
"step": 14802
},
{
"epoch": 0.8972121212121212,
"grad_norm": 0.053367145359516144,
"learning_rate": 0.00034282828282828287,
"loss": 3.129279613494873,
"step": 14804
},
{
"epoch": 0.8973333333333333,
"grad_norm": 0.06218753382563591,
"learning_rate": 0.00034242424242424244,
"loss": 3.133208990097046,
"step": 14806
},
{
"epoch": 0.8974545454545455,
"grad_norm": 0.048050880432128906,
"learning_rate": 0.00034202020202020206,
"loss": 3.14495587348938,
"step": 14808
},
{
"epoch": 0.8975757575757576,
"grad_norm": 0.05881669744849205,
"learning_rate": 0.00034161616161616163,
"loss": 3.1054811477661133,
"step": 14810
},
{
"epoch": 0.8976969696969697,
"grad_norm": 0.055520765483379364,
"learning_rate": 0.0003412121212121212,
"loss": 3.1296653747558594,
"step": 14812
},
{
"epoch": 0.8978181818181818,
"grad_norm": 0.04711022973060608,
"learning_rate": 0.00034080808080808083,
"loss": 3.1564278602600098,
"step": 14814
},
{
"epoch": 0.8979393939393939,
"grad_norm": 0.056624773889780045,
"learning_rate": 0.0003404040404040404,
"loss": 3.110520601272583,
"step": 14816
},
{
"epoch": 0.8980606060606061,
"grad_norm": 0.04828416556119919,
"learning_rate": 0.00034,
"loss": 3.121507406234741,
"step": 14818
},
{
"epoch": 0.8981818181818182,
"grad_norm": 0.047909706830978394,
"learning_rate": 0.0003395959595959596,
"loss": 3.1311981678009033,
"step": 14820
},
{
"epoch": 0.8983030303030303,
"grad_norm": 0.04755779355764389,
"learning_rate": 0.0003391919191919192,
"loss": 3.1193759441375732,
"step": 14822
},
{
"epoch": 0.8984242424242425,
"grad_norm": 0.04891900345683098,
"learning_rate": 0.0003387878787878788,
"loss": 3.101379632949829,
"step": 14824
},
{
"epoch": 0.8985454545454545,
"grad_norm": 0.04839817434549332,
"learning_rate": 0.00033838383838383836,
"loss": 3.142587184906006,
"step": 14826
},
{
"epoch": 0.8986666666666666,
"grad_norm": 0.047052230685949326,
"learning_rate": 0.000337979797979798,
"loss": 3.119938373565674,
"step": 14828
},
{
"epoch": 0.8987878787878788,
"grad_norm": 0.048072606325149536,
"learning_rate": 0.0003375757575757576,
"loss": 3.103869676589966,
"step": 14830
},
{
"epoch": 0.8989090909090909,
"grad_norm": 0.047518957406282425,
"learning_rate": 0.0003371717171717172,
"loss": 3.1065526008605957,
"step": 14832
},
{
"epoch": 0.8990303030303031,
"grad_norm": 0.05075611546635628,
"learning_rate": 0.0003367676767676768,
"loss": 3.103808641433716,
"step": 14834
},
{
"epoch": 0.8991515151515151,
"grad_norm": 0.0482490137219429,
"learning_rate": 0.0003363636363636364,
"loss": 3.0947978496551514,
"step": 14836
},
{
"epoch": 0.8992727272727272,
"grad_norm": 0.05345182493329048,
"learning_rate": 0.000335959595959596,
"loss": 3.092360496520996,
"step": 14838
},
{
"epoch": 0.8993939393939394,
"grad_norm": 0.05291071534156799,
"learning_rate": 0.0003355555555555556,
"loss": 3.1102120876312256,
"step": 14840
},
{
"epoch": 0.8995151515151515,
"grad_norm": 0.04746268689632416,
"learning_rate": 0.00033515151515151515,
"loss": 3.1237640380859375,
"step": 14842
},
{
"epoch": 0.8996363636363637,
"grad_norm": 0.05320015549659729,
"learning_rate": 0.00033474747474747477,
"loss": 3.1246447563171387,
"step": 14844
},
{
"epoch": 0.8997575757575758,
"grad_norm": 0.05049571767449379,
"learning_rate": 0.00033434343434343434,
"loss": 3.096954107284546,
"step": 14846
},
{
"epoch": 0.8998787878787878,
"grad_norm": 0.048775240778923035,
"learning_rate": 0.00033393939393939397,
"loss": 3.085236072540283,
"step": 14848
},
{
"epoch": 0.9,
"grad_norm": 0.05154202878475189,
"learning_rate": 0.00033353535353535354,
"loss": 3.132136583328247,
"step": 14850
},
{
"epoch": 0.9001212121212121,
"grad_norm": 0.058824945241212845,
"learning_rate": 0.00033313131313131316,
"loss": 3.115528106689453,
"step": 14852
},
{
"epoch": 0.9002424242424243,
"grad_norm": 0.04966680705547333,
"learning_rate": 0.00033272727272727273,
"loss": 3.0529439449310303,
"step": 14854
},
{
"epoch": 0.9003636363636364,
"grad_norm": 0.0595027431845665,
"learning_rate": 0.0003323232323232323,
"loss": 3.0600781440734863,
"step": 14856
},
{
"epoch": 0.9004848484848484,
"grad_norm": 0.054943013936281204,
"learning_rate": 0.00033191919191919193,
"loss": 3.138410806655884,
"step": 14858
},
{
"epoch": 0.9006060606060606,
"grad_norm": 0.04661399871110916,
"learning_rate": 0.0003315151515151515,
"loss": 3.1090784072875977,
"step": 14860
},
{
"epoch": 0.9007272727272727,
"grad_norm": 0.058183711022138596,
"learning_rate": 0.0003311111111111111,
"loss": 3.086672306060791,
"step": 14862
},
{
"epoch": 0.9008484848484849,
"grad_norm": 0.05238233134150505,
"learning_rate": 0.00033070707070707075,
"loss": 3.1117947101593018,
"step": 14864
},
{
"epoch": 0.900969696969697,
"grad_norm": 0.048204757273197174,
"learning_rate": 0.0003303030303030303,
"loss": 3.07173228263855,
"step": 14866
},
{
"epoch": 0.901090909090909,
"grad_norm": 0.052000049501657486,
"learning_rate": 0.00032989898989898995,
"loss": 3.055920124053955,
"step": 14868
},
{
"epoch": 0.9012121212121212,
"grad_norm": 0.04882850497961044,
"learning_rate": 0.0003294949494949495,
"loss": 3.0585289001464844,
"step": 14870
},
{
"epoch": 0.9013333333333333,
"grad_norm": 0.05513302981853485,
"learning_rate": 0.00032909090909090914,
"loss": 3.0966548919677734,
"step": 14872
},
{
"epoch": 0.9014545454545455,
"grad_norm": 0.05091359093785286,
"learning_rate": 0.0003286868686868687,
"loss": 3.0789103507995605,
"step": 14874
},
{
"epoch": 0.9015757575757576,
"grad_norm": 0.04993242025375366,
"learning_rate": 0.0003282828282828283,
"loss": 3.0773961544036865,
"step": 14876
},
{
"epoch": 0.9016969696969697,
"grad_norm": 0.049429137259721756,
"learning_rate": 0.0003278787878787879,
"loss": 3.091400623321533,
"step": 14878
},
{
"epoch": 0.9018181818181819,
"grad_norm": 0.049758970737457275,
"learning_rate": 0.0003274747474747475,
"loss": 3.078472375869751,
"step": 14880
},
{
"epoch": 0.9019393939393939,
"grad_norm": 0.048321060836315155,
"learning_rate": 0.0003270707070707071,
"loss": 3.094616174697876,
"step": 14882
},
{
"epoch": 0.902060606060606,
"grad_norm": 0.050122033804655075,
"learning_rate": 0.0003266666666666667,
"loss": 3.141958713531494,
"step": 14884
},
{
"epoch": 0.9021818181818182,
"grad_norm": 0.04973853752017021,
"learning_rate": 0.00032626262626262624,
"loss": 3.063685655593872,
"step": 14886
},
{
"epoch": 0.9023030303030303,
"grad_norm": 0.05172710120677948,
"learning_rate": 0.00032585858585858587,
"loss": 3.0635793209075928,
"step": 14888
},
{
"epoch": 0.9024242424242425,
"grad_norm": 0.060692887753248215,
"learning_rate": 0.00032545454545454544,
"loss": 3.071237087249756,
"step": 14890
},
{
"epoch": 0.9025454545454545,
"grad_norm": 0.048152320086956024,
"learning_rate": 0.00032505050505050507,
"loss": 3.0557422637939453,
"step": 14892
},
{
"epoch": 0.9026666666666666,
"grad_norm": 0.056237295269966125,
"learning_rate": 0.00032464646464646464,
"loss": 3.0518040657043457,
"step": 14894
},
{
"epoch": 0.9027878787878788,
"grad_norm": 0.047257374972105026,
"learning_rate": 0.00032424242424242426,
"loss": 3.0600006580352783,
"step": 14896
},
{
"epoch": 0.9029090909090909,
"grad_norm": 0.061731044203042984,
"learning_rate": 0.0003238383838383839,
"loss": 3.0919981002807617,
"step": 14898
},
{
"epoch": 0.9030303030303031,
"grad_norm": 0.059665776789188385,
"learning_rate": 0.00032343434343434346,
"loss": 3.0763723850250244,
"step": 14900
},
{
"epoch": 0.9031515151515151,
"grad_norm": 0.0526990182697773,
"learning_rate": 0.0003230303030303031,
"loss": 3.095054864883423,
"step": 14902
},
{
"epoch": 0.9032727272727272,
"grad_norm": 0.04921716824173927,
"learning_rate": 0.00032262626262626265,
"loss": 3.0694658756256104,
"step": 14904
},
{
"epoch": 0.9033939393939394,
"grad_norm": 0.04852139577269554,
"learning_rate": 0.0003222222222222222,
"loss": 3.0684821605682373,
"step": 14906
},
{
"epoch": 0.9035151515151515,
"grad_norm": 0.05066565424203873,
"learning_rate": 0.00032181818181818185,
"loss": 3.033263921737671,
"step": 14908
},
{
"epoch": 0.9036363636363637,
"grad_norm": 0.05055539309978485,
"learning_rate": 0.0003214141414141414,
"loss": 3.049187183380127,
"step": 14910
},
{
"epoch": 0.9037575757575758,
"grad_norm": 0.054011739790439606,
"learning_rate": 0.00032101010101010104,
"loss": 3.071004867553711,
"step": 14912
},
{
"epoch": 0.9038787878787878,
"grad_norm": 0.04756789281964302,
"learning_rate": 0.0003206060606060606,
"loss": 3.080389976501465,
"step": 14914
},
{
"epoch": 0.904,
"grad_norm": 0.051860153675079346,
"learning_rate": 0.00032020202020202024,
"loss": 3.083324432373047,
"step": 14916
},
{
"epoch": 0.9041212121212121,
"grad_norm": 0.0473032109439373,
"learning_rate": 0.0003197979797979798,
"loss": 3.063307523727417,
"step": 14918
},
{
"epoch": 0.9042424242424243,
"grad_norm": 0.04940163344144821,
"learning_rate": 0.0003193939393939394,
"loss": 3.1022658348083496,
"step": 14920
},
{
"epoch": 0.9043636363636364,
"grad_norm": 0.046806175261735916,
"learning_rate": 0.000318989898989899,
"loss": 3.06487774848938,
"step": 14922
},
{
"epoch": 0.9044848484848484,
"grad_norm": 0.0524437315762043,
"learning_rate": 0.0003185858585858586,
"loss": 3.0934104919433594,
"step": 14924
},
{
"epoch": 0.9046060606060606,
"grad_norm": 0.048468489199876785,
"learning_rate": 0.0003181818181818182,
"loss": 3.0505168437957764,
"step": 14926
},
{
"epoch": 0.9047272727272727,
"grad_norm": 0.04842652752995491,
"learning_rate": 0.0003177777777777778,
"loss": 3.0219154357910156,
"step": 14928
},
{
"epoch": 0.9048484848484849,
"grad_norm": 0.05972776561975479,
"learning_rate": 0.00031737373737373734,
"loss": 3.0650787353515625,
"step": 14930
},
{
"epoch": 0.904969696969697,
"grad_norm": 0.053563788533210754,
"learning_rate": 0.000316969696969697,
"loss": 3.0455336570739746,
"step": 14932
},
{
"epoch": 0.9050909090909091,
"grad_norm": 0.0473889522254467,
"learning_rate": 0.0003165656565656566,
"loss": 3.049173593521118,
"step": 14934
},
{
"epoch": 0.9052121212121212,
"grad_norm": 0.05055326968431473,
"learning_rate": 0.00031616161616161617,
"loss": 3.079641103744507,
"step": 14936
},
{
"epoch": 0.9053333333333333,
"grad_norm": 0.04693346843123436,
"learning_rate": 0.0003157575757575758,
"loss": 3.034654140472412,
"step": 14938
},
{
"epoch": 0.9054545454545454,
"grad_norm": 0.05107806995511055,
"learning_rate": 0.00031535353535353536,
"loss": 3.1061806678771973,
"step": 14940
},
{
"epoch": 0.9055757575757576,
"grad_norm": 0.04929112270474434,
"learning_rate": 0.000314949494949495,
"loss": 3.074596405029297,
"step": 14942
},
{
"epoch": 0.9056969696969697,
"grad_norm": 0.0467042438685894,
"learning_rate": 0.00031454545454545456,
"loss": 3.069230079650879,
"step": 14944
},
{
"epoch": 0.9058181818181819,
"grad_norm": 0.04994487017393112,
"learning_rate": 0.0003141414141414142,
"loss": 3.0619499683380127,
"step": 14946
},
{
"epoch": 0.9059393939393939,
"grad_norm": 0.04997584968805313,
"learning_rate": 0.00031373737373737375,
"loss": 3.016120433807373,
"step": 14948
},
{
"epoch": 0.906060606060606,
"grad_norm": 0.0470002219080925,
"learning_rate": 0.0003133333333333333,
"loss": 3.022606611251831,
"step": 14950
},
{
"epoch": 0.9061818181818182,
"grad_norm": 0.05784716084599495,
"learning_rate": 0.00031292929292929295,
"loss": 3.112872838973999,
"step": 14952
},
{
"epoch": 0.9063030303030303,
"grad_norm": 0.049129828810691833,
"learning_rate": 0.0003125252525252525,
"loss": 3.061007022857666,
"step": 14954
},
{
"epoch": 0.9064242424242425,
"grad_norm": 0.047473739832639694,
"learning_rate": 0.00031212121212121214,
"loss": 2.997954845428467,
"step": 14956
},
{
"epoch": 0.9065454545454545,
"grad_norm": 0.05446930602192879,
"learning_rate": 0.0003117171717171717,
"loss": 3.110673189163208,
"step": 14958
},
{
"epoch": 0.9066666666666666,
"grad_norm": 0.04714999720454216,
"learning_rate": 0.0003113131313131313,
"loss": 3.0466716289520264,
"step": 14960
},
{
"epoch": 0.9067878787878788,
"grad_norm": 0.047948941588401794,
"learning_rate": 0.0003109090909090909,
"loss": 3.061476707458496,
"step": 14962
},
{
"epoch": 0.9069090909090909,
"grad_norm": 0.0473313108086586,
"learning_rate": 0.0003105050505050505,
"loss": 3.0429117679595947,
"step": 14964
},
{
"epoch": 0.9070303030303031,
"grad_norm": 0.049296434968709946,
"learning_rate": 0.00031010101010101016,
"loss": 3.0572829246520996,
"step": 14966
},
{
"epoch": 0.9071515151515152,
"grad_norm": 0.04767167940735817,
"learning_rate": 0.00030969696969696973,
"loss": 2.997082233428955,
"step": 14968
},
{
"epoch": 0.9072727272727272,
"grad_norm": 0.04733441770076752,
"learning_rate": 0.0003092929292929293,
"loss": 3.0527710914611816,
"step": 14970
},
{
"epoch": 0.9073939393939394,
"grad_norm": 0.04740438982844353,
"learning_rate": 0.00030888888888888893,
"loss": 3.0416367053985596,
"step": 14972
},
{
"epoch": 0.9075151515151515,
"grad_norm": 0.04941270872950554,
"learning_rate": 0.0003084848484848485,
"loss": 3.033539056777954,
"step": 14974
},
{
"epoch": 0.9076363636363637,
"grad_norm": 0.047885309904813766,
"learning_rate": 0.0003080808080808081,
"loss": 3.04938006401062,
"step": 14976
},
{
"epoch": 0.9077575757575758,
"grad_norm": 0.05416514724493027,
"learning_rate": 0.0003076767676767677,
"loss": 3.062807083129883,
"step": 14978
},
{
"epoch": 0.9078787878787878,
"grad_norm": 0.05082724243402481,
"learning_rate": 0.00030727272727272727,
"loss": 3.0251924991607666,
"step": 14980
},
{
"epoch": 0.908,
"grad_norm": 0.04759107530117035,
"learning_rate": 0.0003068686868686869,
"loss": 3.0336859226226807,
"step": 14982
},
{
"epoch": 0.9081212121212121,
"grad_norm": 0.04717792198061943,
"learning_rate": 0.00030646464646464646,
"loss": 3.041724920272827,
"step": 14984
},
{
"epoch": 0.9082424242424243,
"grad_norm": 0.048763323575258255,
"learning_rate": 0.0003060606060606061,
"loss": 3.0576844215393066,
"step": 14986
},
{
"epoch": 0.9083636363636364,
"grad_norm": 0.04772476479411125,
"learning_rate": 0.00030565656565656566,
"loss": 3.0206258296966553,
"step": 14988
},
{
"epoch": 0.9084848484848485,
"grad_norm": 0.05593089386820793,
"learning_rate": 0.0003052525252525253,
"loss": 3.021843910217285,
"step": 14990
},
{
"epoch": 0.9086060606060606,
"grad_norm": 0.04903830215334892,
"learning_rate": 0.00030484848484848485,
"loss": 3.0275931358337402,
"step": 14992
},
{
"epoch": 0.9087272727272727,
"grad_norm": 0.04874236881732941,
"learning_rate": 0.0003044444444444444,
"loss": 3.059288263320923,
"step": 14994
},
{
"epoch": 0.9088484848484848,
"grad_norm": 0.04857790842652321,
"learning_rate": 0.00030404040404040405,
"loss": 3.002352714538574,
"step": 14996
},
{
"epoch": 0.908969696969697,
"grad_norm": 0.0479942187666893,
"learning_rate": 0.0003036363636363636,
"loss": 2.9991049766540527,
"step": 14998
},
{
"epoch": 0.9090909090909091,
"grad_norm": 0.04842197895050049,
"learning_rate": 0.00030323232323232324,
"loss": 3.0747945308685303,
"step": 15000
},
{
"epoch": 0.9092121212121212,
"grad_norm": 0.04676712304353714,
"learning_rate": 0.00030282828282828287,
"loss": 2.9882025718688965,
"step": 15002
},
{
"epoch": 0.9093333333333333,
"grad_norm": 0.04951465502381325,
"learning_rate": 0.00030242424242424244,
"loss": 3.0186614990234375,
"step": 15004
},
{
"epoch": 0.9094545454545454,
"grad_norm": 0.05095890909433365,
"learning_rate": 0.00030202020202020207,
"loss": 3.0411832332611084,
"step": 15006
},
{
"epoch": 0.9095757575757576,
"grad_norm": 0.04776334762573242,
"learning_rate": 0.00030161616161616164,
"loss": 3.014949321746826,
"step": 15008
},
{
"epoch": 0.9096969696969697,
"grad_norm": 0.049257487058639526,
"learning_rate": 0.00030121212121212126,
"loss": 3.0532987117767334,
"step": 15010
},
{
"epoch": 0.9098181818181819,
"grad_norm": 0.04872163385152817,
"learning_rate": 0.00030080808080808083,
"loss": 3.0100114345550537,
"step": 15012
},
{
"epoch": 0.9099393939393939,
"grad_norm": 0.052923042327165604,
"learning_rate": 0.0003004040404040404,
"loss": 3.0519964694976807,
"step": 15014
},
{
"epoch": 0.910060606060606,
"grad_norm": 0.05132339149713516,
"learning_rate": 0.00030000000000000003,
"loss": 2.975999355316162,
"step": 15016
},
{
"epoch": 0.9101818181818182,
"grad_norm": 0.04828552529215813,
"learning_rate": 0.0002995959595959596,
"loss": 3.0606484413146973,
"step": 15018
},
{
"epoch": 0.9103030303030303,
"grad_norm": 0.054970402270555496,
"learning_rate": 0.0002991919191919192,
"loss": 2.9798662662506104,
"step": 15020
},
{
"epoch": 0.9104242424242425,
"grad_norm": 0.04717693477869034,
"learning_rate": 0.0002987878787878788,
"loss": 3.0172953605651855,
"step": 15022
},
{
"epoch": 0.9105454545454545,
"grad_norm": 0.05804942175745964,
"learning_rate": 0.00029838383838383837,
"loss": 3.029543161392212,
"step": 15024
},
{
"epoch": 0.9106666666666666,
"grad_norm": 0.048496782779693604,
"learning_rate": 0.000297979797979798,
"loss": 3.037346839904785,
"step": 15026
},
{
"epoch": 0.9107878787878788,
"grad_norm": 0.058821901679039,
"learning_rate": 0.00029757575757575756,
"loss": 3.0337815284729004,
"step": 15028
},
{
"epoch": 0.9109090909090909,
"grad_norm": 0.05058184266090393,
"learning_rate": 0.0002971717171717172,
"loss": 2.9844722747802734,
"step": 15030
},
{
"epoch": 0.9110303030303031,
"grad_norm": 0.058179114013910294,
"learning_rate": 0.00029676767676767676,
"loss": 3.0310075283050537,
"step": 15032
},
{
"epoch": 0.9111515151515152,
"grad_norm": 0.05036747083067894,
"learning_rate": 0.0002963636363636364,
"loss": 3.0208184719085693,
"step": 15034
},
{
"epoch": 0.9112727272727272,
"grad_norm": 0.047799378633499146,
"learning_rate": 0.000295959595959596,
"loss": 2.9850637912750244,
"step": 15036
},
{
"epoch": 0.9113939393939394,
"grad_norm": 0.057837631553411484,
"learning_rate": 0.0002955555555555556,
"loss": 2.9939279556274414,
"step": 15038
},
{
"epoch": 0.9115151515151515,
"grad_norm": 0.052373167127370834,
"learning_rate": 0.0002951515151515152,
"loss": 2.9720492362976074,
"step": 15040
},
{
"epoch": 0.9116363636363637,
"grad_norm": 0.05779006704688072,
"learning_rate": 0.0002947474747474748,
"loss": 2.986496686935425,
"step": 15042
},
{
"epoch": 0.9117575757575758,
"grad_norm": 0.04746294021606445,
"learning_rate": 0.00029434343434343434,
"loss": 3.007261037826538,
"step": 15044
},
{
"epoch": 0.9118787878787878,
"grad_norm": 0.04667294770479202,
"learning_rate": 0.00029393939393939397,
"loss": 3.0700912475585938,
"step": 15046
},
{
"epoch": 0.912,
"grad_norm": 0.0476558692753315,
"learning_rate": 0.00029353535353535354,
"loss": 3.0087509155273438,
"step": 15048
},
{
"epoch": 0.9121212121212121,
"grad_norm": 0.04732242599129677,
"learning_rate": 0.00029313131313131317,
"loss": 2.9987525939941406,
"step": 15050
},
{
"epoch": 0.9122424242424242,
"grad_norm": 0.04878328740596771,
"learning_rate": 0.00029272727272727274,
"loss": 2.983346700668335,
"step": 15052
},
{
"epoch": 0.9123636363636364,
"grad_norm": 0.04571862146258354,
"learning_rate": 0.0002923232323232323,
"loss": 3.0590264797210693,
"step": 15054
},
{
"epoch": 0.9124848484848485,
"grad_norm": 0.047181569039821625,
"learning_rate": 0.00029191919191919193,
"loss": 3.0224626064300537,
"step": 15056
},
{
"epoch": 0.9126060606060606,
"grad_norm": 0.0487745963037014,
"learning_rate": 0.0002915151515151515,
"loss": 3.035682201385498,
"step": 15058
},
{
"epoch": 0.9127272727272727,
"grad_norm": 0.04687568545341492,
"learning_rate": 0.00029111111111111113,
"loss": 3.024343252182007,
"step": 15060
},
{
"epoch": 0.9128484848484848,
"grad_norm": 0.048384543508291245,
"learning_rate": 0.0002907070707070707,
"loss": 2.9565834999084473,
"step": 15062
},
{
"epoch": 0.912969696969697,
"grad_norm": 0.04562794417142868,
"learning_rate": 0.0002903030303030303,
"loss": 2.9811742305755615,
"step": 15064
},
{
"epoch": 0.9130909090909091,
"grad_norm": 0.049166418612003326,
"learning_rate": 0.0002898989898989899,
"loss": 3.012622594833374,
"step": 15066
},
{
"epoch": 0.9132121212121213,
"grad_norm": 0.047454312443733215,
"learning_rate": 0.0002894949494949495,
"loss": 3.0380759239196777,
"step": 15068
},
{
"epoch": 0.9133333333333333,
"grad_norm": 0.0496814101934433,
"learning_rate": 0.00028909090909090914,
"loss": 3.0237767696380615,
"step": 15070
},
{
"epoch": 0.9134545454545454,
"grad_norm": 0.04831012710928917,
"learning_rate": 0.0002886868686868687,
"loss": 3.036762237548828,
"step": 15072
},
{
"epoch": 0.9135757575757576,
"grad_norm": 0.04668761044740677,
"learning_rate": 0.0002882828282828283,
"loss": 2.9863333702087402,
"step": 15074
},
{
"epoch": 0.9136969696969697,
"grad_norm": 0.04959672689437866,
"learning_rate": 0.0002878787878787879,
"loss": 2.986877202987671,
"step": 15076
},
{
"epoch": 0.9138181818181819,
"grad_norm": 0.04918094351887703,
"learning_rate": 0.0002874747474747475,
"loss": 2.976311206817627,
"step": 15078
},
{
"epoch": 0.9139393939393939,
"grad_norm": 0.04907088354229927,
"learning_rate": 0.0002870707070707071,
"loss": 2.9733030796051025,
"step": 15080
},
{
"epoch": 0.914060606060606,
"grad_norm": 0.04764211177825928,
"learning_rate": 0.0002866666666666667,
"loss": 2.9812636375427246,
"step": 15082
},
{
"epoch": 0.9141818181818182,
"grad_norm": 0.04676363989710808,
"learning_rate": 0.0002862626262626263,
"loss": 2.9789204597473145,
"step": 15084
},
{
"epoch": 0.9143030303030303,
"grad_norm": 0.04584423452615738,
"learning_rate": 0.0002858585858585859,
"loss": 2.979576826095581,
"step": 15086
},
{
"epoch": 0.9144242424242425,
"grad_norm": 0.05043371766805649,
"learning_rate": 0.00028545454545454544,
"loss": 2.9541687965393066,
"step": 15088
},
{
"epoch": 0.9145454545454546,
"grad_norm": 0.04901915043592453,
"learning_rate": 0.00028505050505050507,
"loss": 3.0485849380493164,
"step": 15090
},
{
"epoch": 0.9146666666666666,
"grad_norm": 0.0479789599776268,
"learning_rate": 0.00028464646464646464,
"loss": 2.9890599250793457,
"step": 15092
},
{
"epoch": 0.9147878787878788,
"grad_norm": 0.05677196756005287,
"learning_rate": 0.00028424242424242427,
"loss": 2.995720863342285,
"step": 15094
},
{
"epoch": 0.9149090909090909,
"grad_norm": 0.04789744317531586,
"learning_rate": 0.00028383838383838384,
"loss": 3.0160889625549316,
"step": 15096
},
{
"epoch": 0.9150303030303031,
"grad_norm": 0.05017987638711929,
"learning_rate": 0.0002834343434343434,
"loss": 2.9951679706573486,
"step": 15098
},
{
"epoch": 0.9151515151515152,
"grad_norm": 0.05381270498037338,
"learning_rate": 0.00028303030303030303,
"loss": 3.026564836502075,
"step": 15100
},
{
"epoch": 0.9152727272727272,
"grad_norm": 0.04739942029118538,
"learning_rate": 0.00028262626262626266,
"loss": 2.98760724067688,
"step": 15102
},
{
"epoch": 0.9153939393939394,
"grad_norm": 0.04690462723374367,
"learning_rate": 0.00028222222222222223,
"loss": 3.033282518386841,
"step": 15104
},
{
"epoch": 0.9155151515151515,
"grad_norm": 0.049353741109371185,
"learning_rate": 0.00028181818181818185,
"loss": 2.9691948890686035,
"step": 15106
},
{
"epoch": 0.9156363636363636,
"grad_norm": 0.04740137234330177,
"learning_rate": 0.0002814141414141414,
"loss": 2.973548650741577,
"step": 15108
},
{
"epoch": 0.9157575757575758,
"grad_norm": 0.050828561186790466,
"learning_rate": 0.00028101010101010105,
"loss": 3.017289161682129,
"step": 15110
},
{
"epoch": 0.9158787878787878,
"grad_norm": 0.047698013484478,
"learning_rate": 0.0002806060606060606,
"loss": 3.052917718887329,
"step": 15112
},
{
"epoch": 0.916,
"grad_norm": 0.04771312698721886,
"learning_rate": 0.00028020202020202024,
"loss": 3.0056533813476562,
"step": 15114
},
{
"epoch": 0.9161212121212121,
"grad_norm": 0.05024534836411476,
"learning_rate": 0.0002797979797979798,
"loss": 3.0437700748443604,
"step": 15116
},
{
"epoch": 0.9162424242424242,
"grad_norm": 0.04682096838951111,
"learning_rate": 0.0002793939393939394,
"loss": 2.969006061553955,
"step": 15118
},
{
"epoch": 0.9163636363636364,
"grad_norm": 0.0539851114153862,
"learning_rate": 0.000278989898989899,
"loss": 3.0597620010375977,
"step": 15120
},
{
"epoch": 0.9164848484848485,
"grad_norm": 0.05101285129785538,
"learning_rate": 0.0002785858585858586,
"loss": 2.9954166412353516,
"step": 15122
},
{
"epoch": 0.9166060606060606,
"grad_norm": 0.046719055622816086,
"learning_rate": 0.0002781818181818182,
"loss": 2.980445384979248,
"step": 15124
},
{
"epoch": 0.9167272727272727,
"grad_norm": 0.048412349075078964,
"learning_rate": 0.0002777777777777778,
"loss": 2.998853921890259,
"step": 15126
},
{
"epoch": 0.9168484848484848,
"grad_norm": 0.047937534749507904,
"learning_rate": 0.00027737373737373735,
"loss": 2.980308771133423,
"step": 15128
},
{
"epoch": 0.916969696969697,
"grad_norm": 0.046945493668317795,
"learning_rate": 0.000276969696969697,
"loss": 2.943295955657959,
"step": 15130
},
{
"epoch": 0.9170909090909091,
"grad_norm": 0.0489501953125,
"learning_rate": 0.00027656565656565654,
"loss": 2.985241413116455,
"step": 15132
},
{
"epoch": 0.9172121212121213,
"grad_norm": 0.052563924342393875,
"learning_rate": 0.00027616161616161617,
"loss": 2.9544925689697266,
"step": 15134
},
{
"epoch": 0.9173333333333333,
"grad_norm": 0.050009991973638535,
"learning_rate": 0.0002757575757575758,
"loss": 2.968353033065796,
"step": 15136
},
{
"epoch": 0.9174545454545454,
"grad_norm": 0.04881761595606804,
"learning_rate": 0.00027535353535353537,
"loss": 2.9597651958465576,
"step": 15138
},
{
"epoch": 0.9175757575757576,
"grad_norm": 0.059286829084157944,
"learning_rate": 0.000274949494949495,
"loss": 3.00923490524292,
"step": 15140
},
{
"epoch": 0.9176969696969697,
"grad_norm": 0.05409527197480202,
"learning_rate": 0.00027454545454545456,
"loss": 2.938490390777588,
"step": 15142
},
{
"epoch": 0.9178181818181819,
"grad_norm": 0.0527014285326004,
"learning_rate": 0.0002741414141414142,
"loss": 2.99735951423645,
"step": 15144
},
{
"epoch": 0.9179393939393939,
"grad_norm": 0.04541246220469475,
"learning_rate": 0.00027373737373737376,
"loss": 2.9610166549682617,
"step": 15146
},
{
"epoch": 0.918060606060606,
"grad_norm": 0.04901331663131714,
"learning_rate": 0.00027333333333333333,
"loss": 2.9742374420166016,
"step": 15148
},
{
"epoch": 0.9181818181818182,
"grad_norm": 0.04809056967496872,
"learning_rate": 0.00027292929292929295,
"loss": 2.9854633808135986,
"step": 15150
},
{
"epoch": 0.9183030303030303,
"grad_norm": 0.046987056732177734,
"learning_rate": 0.0002725252525252525,
"loss": 2.9715576171875,
"step": 15152
},
{
"epoch": 0.9184242424242425,
"grad_norm": 0.04981524869799614,
"learning_rate": 0.00027212121212121215,
"loss": 3.0201287269592285,
"step": 15154
},
{
"epoch": 0.9185454545454546,
"grad_norm": 0.047623902559280396,
"learning_rate": 0.0002717171717171717,
"loss": 3.0115416049957275,
"step": 15156
},
{
"epoch": 0.9186666666666666,
"grad_norm": 0.04729556292295456,
"learning_rate": 0.00027131313131313134,
"loss": 2.9959282875061035,
"step": 15158
},
{
"epoch": 0.9187878787878788,
"grad_norm": 0.04597991332411766,
"learning_rate": 0.0002709090909090909,
"loss": 3.037254810333252,
"step": 15160
},
{
"epoch": 0.9189090909090909,
"grad_norm": 0.048215601593256,
"learning_rate": 0.0002705050505050505,
"loss": 2.986534357070923,
"step": 15162
},
{
"epoch": 0.919030303030303,
"grad_norm": 0.04775414615869522,
"learning_rate": 0.0002701010101010101,
"loss": 2.969139337539673,
"step": 15164
},
{
"epoch": 0.9191515151515152,
"grad_norm": 0.05045792832970619,
"learning_rate": 0.0002696969696969697,
"loss": 2.9676644802093506,
"step": 15166
},
{
"epoch": 0.9192727272727272,
"grad_norm": 0.04817252233624458,
"learning_rate": 0.0002692929292929293,
"loss": 2.998842477798462,
"step": 15168
},
{
"epoch": 0.9193939393939394,
"grad_norm": 0.046549759805202484,
"learning_rate": 0.00026888888888888893,
"loss": 3.0248074531555176,
"step": 15170
},
{
"epoch": 0.9195151515151515,
"grad_norm": 0.0482671745121479,
"learning_rate": 0.0002684848484848485,
"loss": 2.9725282192230225,
"step": 15172
},
{
"epoch": 0.9196363636363636,
"grad_norm": 0.04823495075106621,
"learning_rate": 0.00026808080808080813,
"loss": 2.962048053741455,
"step": 15174
},
{
"epoch": 0.9197575757575758,
"grad_norm": 0.04840700700879097,
"learning_rate": 0.0002676767676767677,
"loss": 3.014925718307495,
"step": 15176
},
{
"epoch": 0.9198787878787879,
"grad_norm": 0.05046495050191879,
"learning_rate": 0.0002672727272727273,
"loss": 3.040780544281006,
"step": 15178
},
{
"epoch": 0.92,
"grad_norm": 0.05346962809562683,
"learning_rate": 0.0002668686868686869,
"loss": 3.026644229888916,
"step": 15180
},
{
"epoch": 0.9201212121212121,
"grad_norm": 0.047228019684553146,
"learning_rate": 0.00026646464646464646,
"loss": 2.9886648654937744,
"step": 15182
},
{
"epoch": 0.9202424242424242,
"grad_norm": 0.04819478467106819,
"learning_rate": 0.0002660606060606061,
"loss": 2.9803049564361572,
"step": 15184
},
{
"epoch": 0.9203636363636364,
"grad_norm": 0.05041540786623955,
"learning_rate": 0.00026565656565656566,
"loss": 2.9912827014923096,
"step": 15186
},
{
"epoch": 0.9204848484848485,
"grad_norm": 0.05379606783390045,
"learning_rate": 0.0002652525252525253,
"loss": 3.0327024459838867,
"step": 15188
},
{
"epoch": 0.9206060606060606,
"grad_norm": 0.04872296005487442,
"learning_rate": 0.00026484848484848486,
"loss": 3.034039258956909,
"step": 15190
},
{
"epoch": 0.9207272727272727,
"grad_norm": 0.05528697744011879,
"learning_rate": 0.00026444444444444443,
"loss": 2.954509735107422,
"step": 15192
},
{
"epoch": 0.9208484848484848,
"grad_norm": 0.049840137362480164,
"learning_rate": 0.00026404040404040405,
"loss": 3.013262987136841,
"step": 15194
},
{
"epoch": 0.920969696969697,
"grad_norm": 0.04598007723689079,
"learning_rate": 0.0002636363636363636,
"loss": 3.0502090454101562,
"step": 15196
},
{
"epoch": 0.9210909090909091,
"grad_norm": 0.04628410190343857,
"learning_rate": 0.00026323232323232325,
"loss": 3.0407967567443848,
"step": 15198
},
{
"epoch": 0.9212121212121213,
"grad_norm": 0.05356813594698906,
"learning_rate": 0.0002628282828282828,
"loss": 3.0016002655029297,
"step": 15200
},
{
"epoch": 0.9213333333333333,
"grad_norm": 0.05797095224261284,
"learning_rate": 0.00026242424242424244,
"loss": 2.997265100479126,
"step": 15202
},
{
"epoch": 0.9214545454545454,
"grad_norm": 0.048060767352581024,
"learning_rate": 0.00026202020202020207,
"loss": 3.0207021236419678,
"step": 15204
},
{
"epoch": 0.9215757575757576,
"grad_norm": 0.05373968556523323,
"learning_rate": 0.00026161616161616164,
"loss": 3.00209903717041,
"step": 15206
},
{
"epoch": 0.9216969696969697,
"grad_norm": 0.052732985466718674,
"learning_rate": 0.00026121212121212126,
"loss": 3.001075506210327,
"step": 15208
},
{
"epoch": 0.9218181818181819,
"grad_norm": 0.052968792617321014,
"learning_rate": 0.00026080808080808084,
"loss": 3.001727342605591,
"step": 15210
},
{
"epoch": 0.921939393939394,
"grad_norm": 0.04739104583859444,
"learning_rate": 0.0002604040404040404,
"loss": 2.997205972671509,
"step": 15212
},
{
"epoch": 0.922060606060606,
"grad_norm": 0.05933232232928276,
"learning_rate": 0.00026000000000000003,
"loss": 3.032184362411499,
"step": 15214
},
{
"epoch": 0.9221818181818182,
"grad_norm": 0.04867679998278618,
"learning_rate": 0.0002595959595959596,
"loss": 3.0752642154693604,
"step": 15216
},
{
"epoch": 0.9223030303030303,
"grad_norm": 0.04845423996448517,
"learning_rate": 0.00025919191919191923,
"loss": 3.0263795852661133,
"step": 15218
},
{
"epoch": 0.9224242424242424,
"grad_norm": 0.05050051584839821,
"learning_rate": 0.0002587878787878788,
"loss": 3.0147366523742676,
"step": 15220
},
{
"epoch": 0.9225454545454546,
"grad_norm": 0.04731815680861473,
"learning_rate": 0.00025838383838383837,
"loss": 3.06956148147583,
"step": 15222
},
{
"epoch": 0.9226666666666666,
"grad_norm": 0.04742210730910301,
"learning_rate": 0.000257979797979798,
"loss": 2.99751353263855,
"step": 15224
},
{
"epoch": 0.9227878787878788,
"grad_norm": 0.048089079558849335,
"learning_rate": 0.00025757575757575756,
"loss": 2.970702648162842,
"step": 15226
},
{
"epoch": 0.9229090909090909,
"grad_norm": 0.049308281391859055,
"learning_rate": 0.0002571717171717172,
"loss": 3.0644397735595703,
"step": 15228
},
{
"epoch": 0.923030303030303,
"grad_norm": 0.04764939472079277,
"learning_rate": 0.00025676767676767676,
"loss": 3.059779644012451,
"step": 15230
},
{
"epoch": 0.9231515151515152,
"grad_norm": 0.04847784340381622,
"learning_rate": 0.0002563636363636364,
"loss": 3.0484306812286377,
"step": 15232
},
{
"epoch": 0.9232727272727272,
"grad_norm": 0.0526236928999424,
"learning_rate": 0.00025595959595959596,
"loss": 3.0095198154449463,
"step": 15234
},
{
"epoch": 0.9233939393939394,
"grad_norm": 0.05197611078619957,
"learning_rate": 0.00025555555555555553,
"loss": 3.085571050643921,
"step": 15236
},
{
"epoch": 0.9235151515151515,
"grad_norm": 0.050345152616500854,
"learning_rate": 0.0002551515151515152,
"loss": 3.0545616149902344,
"step": 15238
},
{
"epoch": 0.9236363636363636,
"grad_norm": 0.05184698477387428,
"learning_rate": 0.0002547474747474748,
"loss": 3.0053529739379883,
"step": 15240
},
{
"epoch": 0.9237575757575758,
"grad_norm": 0.04677793011069298,
"learning_rate": 0.00025434343434343435,
"loss": 3.06418776512146,
"step": 15242
},
{
"epoch": 0.9238787878787879,
"grad_norm": 0.048886727541685104,
"learning_rate": 0.000253939393939394,
"loss": 3.037677526473999,
"step": 15244
},
{
"epoch": 0.924,
"grad_norm": 0.05041050538420677,
"learning_rate": 0.00025353535353535354,
"loss": 3.036999225616455,
"step": 15246
},
{
"epoch": 0.9241212121212121,
"grad_norm": 0.048749301582574844,
"learning_rate": 0.00025313131313131317,
"loss": 3.015958786010742,
"step": 15248
},
{
"epoch": 0.9242424242424242,
"grad_norm": 0.05030302703380585,
"learning_rate": 0.00025272727272727274,
"loss": 3.0140326023101807,
"step": 15250
},
{
"epoch": 0.9243636363636364,
"grad_norm": 0.053868625313043594,
"learning_rate": 0.00025232323232323236,
"loss": 3.0658063888549805,
"step": 15252
},
{
"epoch": 0.9244848484848485,
"grad_norm": 0.05038420110940933,
"learning_rate": 0.00025191919191919194,
"loss": 3.070720911026001,
"step": 15254
},
{
"epoch": 0.9246060606060607,
"grad_norm": 0.047590795904397964,
"learning_rate": 0.0002515151515151515,
"loss": 3.056928873062134,
"step": 15256
},
{
"epoch": 0.9247272727272727,
"grad_norm": 0.05170606076717377,
"learning_rate": 0.00025111111111111113,
"loss": 3.060220956802368,
"step": 15258
},
{
"epoch": 0.9248484848484848,
"grad_norm": 0.05282248929142952,
"learning_rate": 0.0002507070707070707,
"loss": 3.0586512088775635,
"step": 15260
},
{
"epoch": 0.924969696969697,
"grad_norm": 0.05099230259656906,
"learning_rate": 0.0002503030303030303,
"loss": 2.974452018737793,
"step": 15262
},
{
"epoch": 0.9250909090909091,
"grad_norm": 0.04809005931019783,
"learning_rate": 0.0002498989898989899,
"loss": 3.0251190662384033,
"step": 15264
},
{
"epoch": 0.9252121212121213,
"grad_norm": 0.04923873767256737,
"learning_rate": 0.00024949494949494947,
"loss": 3.0753602981567383,
"step": 15266
},
{
"epoch": 0.9253333333333333,
"grad_norm": 0.047278109937906265,
"learning_rate": 0.0002490909090909091,
"loss": 3.0312376022338867,
"step": 15268
},
{
"epoch": 0.9254545454545454,
"grad_norm": 0.053930480033159256,
"learning_rate": 0.00024868686868686866,
"loss": 3.020648956298828,
"step": 15270
},
{
"epoch": 0.9255757575757576,
"grad_norm": 0.049685489386320114,
"learning_rate": 0.0002482828282828283,
"loss": 2.998873233795166,
"step": 15272
},
{
"epoch": 0.9256969696969697,
"grad_norm": 0.05520882084965706,
"learning_rate": 0.0002478787878787879,
"loss": 3.058461904525757,
"step": 15274
},
{
"epoch": 0.9258181818181819,
"grad_norm": 0.04783989489078522,
"learning_rate": 0.0002474747474747475,
"loss": 2.9944565296173096,
"step": 15276
},
{
"epoch": 0.925939393939394,
"grad_norm": 0.05035402253270149,
"learning_rate": 0.0002470707070707071,
"loss": 2.998457670211792,
"step": 15278
},
{
"epoch": 0.926060606060606,
"grad_norm": 0.04756399616599083,
"learning_rate": 0.0002466666666666667,
"loss": 3.059016704559326,
"step": 15280
},
{
"epoch": 0.9261818181818182,
"grad_norm": 0.046713534742593765,
"learning_rate": 0.0002462626262626263,
"loss": 3.0394210815429688,
"step": 15282
},
{
"epoch": 0.9263030303030303,
"grad_norm": 0.04686173424124718,
"learning_rate": 0.0002458585858585859,
"loss": 3.045168399810791,
"step": 15284
},
{
"epoch": 0.9264242424242424,
"grad_norm": 0.047304607927799225,
"learning_rate": 0.00024545454545454545,
"loss": 3.0148980617523193,
"step": 15286
},
{
"epoch": 0.9265454545454546,
"grad_norm": 0.04716654494404793,
"learning_rate": 0.00024505050505050507,
"loss": 3.025423765182495,
"step": 15288
},
{
"epoch": 0.9266666666666666,
"grad_norm": 0.04873797297477722,
"learning_rate": 0.00024464646464646464,
"loss": 3.0100975036621094,
"step": 15290
},
{
"epoch": 0.9267878787878788,
"grad_norm": 0.0538412444293499,
"learning_rate": 0.00024424242424242427,
"loss": 2.9873123168945312,
"step": 15292
},
{
"epoch": 0.9269090909090909,
"grad_norm": 0.0496443472802639,
"learning_rate": 0.00024383838383838384,
"loss": 3.0609822273254395,
"step": 15294
},
{
"epoch": 0.927030303030303,
"grad_norm": 0.063149094581604,
"learning_rate": 0.00024343434343434344,
"loss": 3.008270502090454,
"step": 15296
},
{
"epoch": 0.9271515151515152,
"grad_norm": 0.05574154853820801,
"learning_rate": 0.00024303030303030304,
"loss": 3.0382161140441895,
"step": 15298
},
{
"epoch": 0.9272727272727272,
"grad_norm": 0.05069867521524429,
"learning_rate": 0.00024262626262626263,
"loss": 3.023195266723633,
"step": 15300
},
{
"epoch": 0.9273939393939394,
"grad_norm": 0.048660460859537125,
"learning_rate": 0.0002422222222222222,
"loss": 3.0100245475769043,
"step": 15302
},
{
"epoch": 0.9275151515151515,
"grad_norm": 0.04848002642393112,
"learning_rate": 0.0002418181818181818,
"loss": 3.0627706050872803,
"step": 15304
},
{
"epoch": 0.9276363636363636,
"grad_norm": 0.04914858564734459,
"learning_rate": 0.0002414141414141414,
"loss": 3.026715040206909,
"step": 15306
},
{
"epoch": 0.9277575757575758,
"grad_norm": 0.049361877143383026,
"learning_rate": 0.00024101010101010105,
"loss": 3.0934410095214844,
"step": 15308
},
{
"epoch": 0.9278787878787879,
"grad_norm": 0.048982419073581696,
"learning_rate": 0.00024060606060606062,
"loss": 3.031724214553833,
"step": 15310
},
{
"epoch": 0.928,
"grad_norm": 0.04893830418586731,
"learning_rate": 0.00024020202020202022,
"loss": 3.046769380569458,
"step": 15312
},
{
"epoch": 0.9281212121212121,
"grad_norm": 0.048483893275260925,
"learning_rate": 0.00023979797979797982,
"loss": 3.0229108333587646,
"step": 15314
},
{
"epoch": 0.9282424242424242,
"grad_norm": 0.04734852537512779,
"learning_rate": 0.00023939393939393942,
"loss": 3.0378520488739014,
"step": 15316
},
{
"epoch": 0.9283636363636364,
"grad_norm": 0.05064778029918671,
"learning_rate": 0.00023898989898989901,
"loss": 3.0349087715148926,
"step": 15318
},
{
"epoch": 0.9284848484848485,
"grad_norm": 0.048299457877874374,
"learning_rate": 0.0002385858585858586,
"loss": 3.02449107170105,
"step": 15320
},
{
"epoch": 0.9286060606060607,
"grad_norm": 0.054908037185668945,
"learning_rate": 0.00023818181818181818,
"loss": 3.0441524982452393,
"step": 15322
},
{
"epoch": 0.9287272727272727,
"grad_norm": 0.057252589613199234,
"learning_rate": 0.00023777777777777778,
"loss": 3.018800973892212,
"step": 15324
},
{
"epoch": 0.9288484848484848,
"grad_norm": 0.04773838818073273,
"learning_rate": 0.00023737373737373738,
"loss": 3.0674197673797607,
"step": 15326
},
{
"epoch": 0.928969696969697,
"grad_norm": 0.05324708670377731,
"learning_rate": 0.00023696969696969698,
"loss": 3.025228500366211,
"step": 15328
},
{
"epoch": 0.9290909090909091,
"grad_norm": 0.052584752440452576,
"learning_rate": 0.00023656565656565657,
"loss": 3.0662190914154053,
"step": 15330
},
{
"epoch": 0.9292121212121213,
"grad_norm": 0.04914461448788643,
"learning_rate": 0.00023616161616161617,
"loss": 3.0336921215057373,
"step": 15332
},
{
"epoch": 0.9293333333333333,
"grad_norm": 0.04678596928715706,
"learning_rate": 0.00023575757575757574,
"loss": 3.0516581535339355,
"step": 15334
},
{
"epoch": 0.9294545454545454,
"grad_norm": 0.05105559900403023,
"learning_rate": 0.00023535353535353534,
"loss": 3.0353193283081055,
"step": 15336
},
{
"epoch": 0.9295757575757576,
"grad_norm": 0.05281532555818558,
"learning_rate": 0.00023494949494949494,
"loss": 3.0382118225097656,
"step": 15338
},
{
"epoch": 0.9296969696969697,
"grad_norm": 0.0475974939763546,
"learning_rate": 0.00023454545454545454,
"loss": 3.005657434463501,
"step": 15340
},
{
"epoch": 0.9298181818181818,
"grad_norm": 0.04936264082789421,
"learning_rate": 0.00023414141414141416,
"loss": 3.044907808303833,
"step": 15342
},
{
"epoch": 0.929939393939394,
"grad_norm": 0.048163238912820816,
"learning_rate": 0.00023373737373737376,
"loss": 3.0071632862091064,
"step": 15344
},
{
"epoch": 0.930060606060606,
"grad_norm": 0.05633603408932686,
"learning_rate": 0.00023333333333333336,
"loss": 3.020859956741333,
"step": 15346
},
{
"epoch": 0.9301818181818182,
"grad_norm": 0.05004741996526718,
"learning_rate": 0.00023292929292929296,
"loss": 3.0235211849212646,
"step": 15348
},
{
"epoch": 0.9303030303030303,
"grad_norm": 0.049934227019548416,
"learning_rate": 0.00023252525252525255,
"loss": 3.036956548690796,
"step": 15350
},
{
"epoch": 0.9304242424242424,
"grad_norm": 0.049190279096364975,
"learning_rate": 0.00023212121212121215,
"loss": 3.025552749633789,
"step": 15352
},
{
"epoch": 0.9305454545454546,
"grad_norm": 0.05435819551348686,
"learning_rate": 0.00023171717171717172,
"loss": 2.99115252494812,
"step": 15354
},
{
"epoch": 0.9306666666666666,
"grad_norm": 0.04900391399860382,
"learning_rate": 0.00023131313131313132,
"loss": 3.0508484840393066,
"step": 15356
},
{
"epoch": 0.9307878787878788,
"grad_norm": 0.051562950015068054,
"learning_rate": 0.00023090909090909092,
"loss": 3.045416831970215,
"step": 15358
},
{
"epoch": 0.9309090909090909,
"grad_norm": 0.04933543503284454,
"learning_rate": 0.00023050505050505052,
"loss": 3.006898880004883,
"step": 15360
},
{
"epoch": 0.931030303030303,
"grad_norm": 0.05090056359767914,
"learning_rate": 0.00023010101010101011,
"loss": 2.9832661151885986,
"step": 15362
},
{
"epoch": 0.9311515151515152,
"grad_norm": 0.04939740151166916,
"learning_rate": 0.0002296969696969697,
"loss": 3.0564041137695312,
"step": 15364
},
{
"epoch": 0.9312727272727273,
"grad_norm": 0.05001511052250862,
"learning_rate": 0.00022929292929292928,
"loss": 3.053158760070801,
"step": 15366
},
{
"epoch": 0.9313939393939394,
"grad_norm": 0.05077577754855156,
"learning_rate": 0.00022888888888888888,
"loss": 2.972395896911621,
"step": 15368
},
{
"epoch": 0.9315151515151515,
"grad_norm": 0.04800216108560562,
"learning_rate": 0.00022848484848484848,
"loss": 3.0438199043273926,
"step": 15370
},
{
"epoch": 0.9316363636363636,
"grad_norm": 0.05949001759290695,
"learning_rate": 0.00022808080808080808,
"loss": 3.070096731185913,
"step": 15372
},
{
"epoch": 0.9317575757575758,
"grad_norm": 0.048368051648139954,
"learning_rate": 0.00022767676767676767,
"loss": 3.044050455093384,
"step": 15374
},
{
"epoch": 0.9318787878787879,
"grad_norm": 0.055126748979091644,
"learning_rate": 0.0002272727272727273,
"loss": 2.989521026611328,
"step": 15376
},
{
"epoch": 0.932,
"grad_norm": 0.049378860741853714,
"learning_rate": 0.0002268686868686869,
"loss": 3.0819180011749268,
"step": 15378
},
{
"epoch": 0.9321212121212121,
"grad_norm": 0.052349235862493515,
"learning_rate": 0.0002264646464646465,
"loss": 3.013503074645996,
"step": 15380
},
{
"epoch": 0.9322424242424242,
"grad_norm": 0.04820221662521362,
"learning_rate": 0.0002260606060606061,
"loss": 3.0219874382019043,
"step": 15382
},
{
"epoch": 0.9323636363636364,
"grad_norm": 0.05067289620637894,
"learning_rate": 0.00022565656565656566,
"loss": 3.045125961303711,
"step": 15384
},
{
"epoch": 0.9324848484848485,
"grad_norm": 0.048089075833559036,
"learning_rate": 0.00022525252525252526,
"loss": 3.0080020427703857,
"step": 15386
},
{
"epoch": 0.9326060606060607,
"grad_norm": 0.052010297775268555,
"learning_rate": 0.00022484848484848486,
"loss": 3.0249948501586914,
"step": 15388
},
{
"epoch": 0.9327272727272727,
"grad_norm": 0.049142975360155106,
"learning_rate": 0.00022444444444444446,
"loss": 3.065917730331421,
"step": 15390
},
{
"epoch": 0.9328484848484848,
"grad_norm": 0.046916522085666656,
"learning_rate": 0.00022404040404040406,
"loss": 3.052931785583496,
"step": 15392
},
{
"epoch": 0.932969696969697,
"grad_norm": 0.04994691535830498,
"learning_rate": 0.00022363636363636365,
"loss": 3.014647960662842,
"step": 15394
},
{
"epoch": 0.9330909090909091,
"grad_norm": 0.04775870218873024,
"learning_rate": 0.00022323232323232322,
"loss": 3.017942190170288,
"step": 15396
},
{
"epoch": 0.9332121212121212,
"grad_norm": 0.04726095125079155,
"learning_rate": 0.00022282828282828282,
"loss": 3.0403971672058105,
"step": 15398
},
{
"epoch": 0.9333333333333333,
"grad_norm": 0.04900322109460831,
"learning_rate": 0.00022242424242424242,
"loss": 2.9968631267547607,
"step": 15400
},
{
"epoch": 0.9334545454545454,
"grad_norm": 0.049017518758773804,
"learning_rate": 0.00022202020202020202,
"loss": 3.014045476913452,
"step": 15402
},
{
"epoch": 0.9335757575757576,
"grad_norm": 0.051660433411598206,
"learning_rate": 0.00022161616161616162,
"loss": 3.028594732284546,
"step": 15404
},
{
"epoch": 0.9336969696969697,
"grad_norm": 0.050443023443222046,
"learning_rate": 0.00022121212121212121,
"loss": 3.022115707397461,
"step": 15406
},
{
"epoch": 0.9338181818181818,
"grad_norm": 0.058351535350084305,
"learning_rate": 0.00022080808080808078,
"loss": 3.019472599029541,
"step": 15408
},
{
"epoch": 0.933939393939394,
"grad_norm": 0.05187256634235382,
"learning_rate": 0.00022040404040404044,
"loss": 3.0418455600738525,
"step": 15410
},
{
"epoch": 0.934060606060606,
"grad_norm": 0.06027960032224655,
"learning_rate": 0.00022000000000000003,
"loss": 3.056389331817627,
"step": 15412
},
{
"epoch": 0.9341818181818182,
"grad_norm": 0.05715470016002655,
"learning_rate": 0.00021959595959595963,
"loss": 3.0256123542785645,
"step": 15414
},
{
"epoch": 0.9343030303030303,
"grad_norm": 0.057363491505384445,
"learning_rate": 0.0002191919191919192,
"loss": 3.0516445636749268,
"step": 15416
},
{
"epoch": 0.9344242424242424,
"grad_norm": 0.07863470911979675,
"learning_rate": 0.0002187878787878788,
"loss": 3.014742374420166,
"step": 15418
},
{
"epoch": 0.9345454545454546,
"grad_norm": 0.05683382973074913,
"learning_rate": 0.0002183838383838384,
"loss": 3.038984537124634,
"step": 15420
},
{
"epoch": 0.9346666666666666,
"grad_norm": 0.05901367589831352,
"learning_rate": 0.000217979797979798,
"loss": 2.999816656112671,
"step": 15422
},
{
"epoch": 0.9347878787878788,
"grad_norm": 0.051490239799022675,
"learning_rate": 0.0002175757575757576,
"loss": 3.0703554153442383,
"step": 15424
},
{
"epoch": 0.9349090909090909,
"grad_norm": 0.048576731234788895,
"learning_rate": 0.0002171717171717172,
"loss": 2.999354600906372,
"step": 15426
},
{
"epoch": 0.935030303030303,
"grad_norm": 0.057345952838659286,
"learning_rate": 0.00021676767676767676,
"loss": 3.001673460006714,
"step": 15428
},
{
"epoch": 0.9351515151515152,
"grad_norm": 0.0488053634762764,
"learning_rate": 0.00021636363636363636,
"loss": 3.0274736881256104,
"step": 15430
},
{
"epoch": 0.9352727272727273,
"grad_norm": 0.061473146080970764,
"learning_rate": 0.00021595959595959596,
"loss": 3.0239806175231934,
"step": 15432
},
{
"epoch": 0.9353939393939394,
"grad_norm": 0.04858576878905296,
"learning_rate": 0.00021555555555555556,
"loss": 3.029201030731201,
"step": 15434
},
{
"epoch": 0.9355151515151515,
"grad_norm": 0.056817322969436646,
"learning_rate": 0.00021515151515151516,
"loss": 3.02553653717041,
"step": 15436
},
{
"epoch": 0.9356363636363636,
"grad_norm": 0.0488731786608696,
"learning_rate": 0.00021474747474747475,
"loss": 3.06377911567688,
"step": 15438
},
{
"epoch": 0.9357575757575758,
"grad_norm": 0.059253159910440445,
"learning_rate": 0.00021434343434343432,
"loss": 3.0484466552734375,
"step": 15440
},
{
"epoch": 0.9358787878787879,
"grad_norm": 0.04934869706630707,
"learning_rate": 0.00021393939393939392,
"loss": 3.0966053009033203,
"step": 15442
},
{
"epoch": 0.936,
"grad_norm": 0.05286795273423195,
"learning_rate": 0.00021353535353535357,
"loss": 3.067439317703247,
"step": 15444
},
{
"epoch": 0.9361212121212121,
"grad_norm": 0.04865637049078941,
"learning_rate": 0.00021313131313131317,
"loss": 3.0359151363372803,
"step": 15446
},
{
"epoch": 0.9362424242424242,
"grad_norm": 0.04912964999675751,
"learning_rate": 0.00021272727272727274,
"loss": 3.038516044616699,
"step": 15448
},
{
"epoch": 0.9363636363636364,
"grad_norm": 0.05106290802359581,
"learning_rate": 0.00021232323232323234,
"loss": 3.0596604347229004,
"step": 15450
},
{
"epoch": 0.9364848484848485,
"grad_norm": 0.048987556248903275,
"learning_rate": 0.00021191919191919194,
"loss": 3.029167413711548,
"step": 15452
},
{
"epoch": 0.9366060606060606,
"grad_norm": 0.05086161941289902,
"learning_rate": 0.00021151515151515154,
"loss": 3.0369575023651123,
"step": 15454
},
{
"epoch": 0.9367272727272727,
"grad_norm": 0.05616499483585358,
"learning_rate": 0.00021111111111111113,
"loss": 3.037477493286133,
"step": 15456
},
{
"epoch": 0.9368484848484848,
"grad_norm": 0.04882238432765007,
"learning_rate": 0.00021070707070707073,
"loss": 3.042635202407837,
"step": 15458
},
{
"epoch": 0.936969696969697,
"grad_norm": 0.04921192303299904,
"learning_rate": 0.0002103030303030303,
"loss": 3.0063631534576416,
"step": 15460
},
{
"epoch": 0.9370909090909091,
"grad_norm": 0.054276689887046814,
"learning_rate": 0.0002098989898989899,
"loss": 2.989459991455078,
"step": 15462
},
{
"epoch": 0.9372121212121212,
"grad_norm": 0.050851497799158096,
"learning_rate": 0.0002094949494949495,
"loss": 3.0287585258483887,
"step": 15464
},
{
"epoch": 0.9373333333333334,
"grad_norm": 0.05902981758117676,
"learning_rate": 0.0002090909090909091,
"loss": 3.0534214973449707,
"step": 15466
},
{
"epoch": 0.9374545454545454,
"grad_norm": 0.05460544675588608,
"learning_rate": 0.0002086868686868687,
"loss": 3.0622527599334717,
"step": 15468
},
{
"epoch": 0.9375757575757576,
"grad_norm": 0.05116083845496178,
"learning_rate": 0.0002082828282828283,
"loss": 3.0016589164733887,
"step": 15470
},
{
"epoch": 0.9376969696969697,
"grad_norm": 0.05332960933446884,
"learning_rate": 0.00020787878787878786,
"loss": 3.038156270980835,
"step": 15472
},
{
"epoch": 0.9378181818181818,
"grad_norm": 0.051887400448322296,
"learning_rate": 0.00020747474747474746,
"loss": 3.0635199546813965,
"step": 15474
},
{
"epoch": 0.937939393939394,
"grad_norm": 0.05284327641129494,
"learning_rate": 0.00020707070707070706,
"loss": 2.9723293781280518,
"step": 15476
},
{
"epoch": 0.938060606060606,
"grad_norm": 0.05430933088064194,
"learning_rate": 0.00020666666666666668,
"loss": 3.017566680908203,
"step": 15478
},
{
"epoch": 0.9381818181818182,
"grad_norm": 0.049333009868860245,
"learning_rate": 0.00020626262626262628,
"loss": 3.047358512878418,
"step": 15480
},
{
"epoch": 0.9383030303030303,
"grad_norm": 0.04948105663061142,
"learning_rate": 0.00020585858585858588,
"loss": 2.9835963249206543,
"step": 15482
},
{
"epoch": 0.9384242424242424,
"grad_norm": 0.04893963411450386,
"learning_rate": 0.00020545454545454548,
"loss": 3.0200445652008057,
"step": 15484
},
{
"epoch": 0.9385454545454546,
"grad_norm": 0.04996220022439957,
"learning_rate": 0.00020505050505050508,
"loss": 3.012974262237549,
"step": 15486
},
{
"epoch": 0.9386666666666666,
"grad_norm": 0.04869063198566437,
"learning_rate": 0.00020464646464646467,
"loss": 2.975890874862671,
"step": 15488
},
{
"epoch": 0.9387878787878788,
"grad_norm": 0.0509868785738945,
"learning_rate": 0.00020424242424242425,
"loss": 3.030564546585083,
"step": 15490
},
{
"epoch": 0.9389090909090909,
"grad_norm": 0.04893852770328522,
"learning_rate": 0.00020383838383838384,
"loss": 3.0349881649017334,
"step": 15492
},
{
"epoch": 0.939030303030303,
"grad_norm": 0.0536225251853466,
"learning_rate": 0.00020343434343434344,
"loss": 3.03829026222229,
"step": 15494
},
{
"epoch": 0.9391515151515152,
"grad_norm": 0.06121872365474701,
"learning_rate": 0.00020303030303030304,
"loss": 3.0408029556274414,
"step": 15496
},
{
"epoch": 0.9392727272727273,
"grad_norm": 0.05058020353317261,
"learning_rate": 0.00020262626262626264,
"loss": 3.004079818725586,
"step": 15498
},
{
"epoch": 0.9393939393939394,
"grad_norm": 0.052557215094566345,
"learning_rate": 0.00020222222222222223,
"loss": 3.034640073776245,
"step": 15500
},
{
"epoch": 0.9395151515151515,
"grad_norm": 0.04807470738887787,
"learning_rate": 0.0002018181818181818,
"loss": 3.004066228866577,
"step": 15502
},
{
"epoch": 0.9396363636363636,
"grad_norm": 0.050174351781606674,
"learning_rate": 0.0002014141414141414,
"loss": 2.9793128967285156,
"step": 15504
},
{
"epoch": 0.9397575757575758,
"grad_norm": 0.049994777888059616,
"learning_rate": 0.000201010101010101,
"loss": 2.9896397590637207,
"step": 15506
},
{
"epoch": 0.9398787878787879,
"grad_norm": 0.04861712455749512,
"learning_rate": 0.0002006060606060606,
"loss": 3.0374317169189453,
"step": 15508
},
{
"epoch": 0.94,
"grad_norm": 0.04972008988261223,
"learning_rate": 0.0002002020202020202,
"loss": 3.0536744594573975,
"step": 15510
},
{
"epoch": 0.9401212121212121,
"grad_norm": 0.04871795326471329,
"learning_rate": 0.0001997979797979798,
"loss": 3.011591672897339,
"step": 15512
},
{
"epoch": 0.9402424242424242,
"grad_norm": 0.05662468448281288,
"learning_rate": 0.0001993939393939394,
"loss": 3.0120842456817627,
"step": 15514
},
{
"epoch": 0.9403636363636364,
"grad_norm": 0.05033624917268753,
"learning_rate": 0.000198989898989899,
"loss": 3.0425589084625244,
"step": 15516
},
{
"epoch": 0.9404848484848485,
"grad_norm": 0.04844122752547264,
"learning_rate": 0.0001985858585858586,
"loss": 3.0392324924468994,
"step": 15518
},
{
"epoch": 0.9406060606060606,
"grad_norm": 0.04987867549061775,
"learning_rate": 0.00019818181818181821,
"loss": 3.04712176322937,
"step": 15520
},
{
"epoch": 0.9407272727272727,
"grad_norm": 0.04952312633395195,
"learning_rate": 0.00019777777777777778,
"loss": 3.0573787689208984,
"step": 15522
},
{
"epoch": 0.9408484848484848,
"grad_norm": 0.048636384308338165,
"learning_rate": 0.00019737373737373738,
"loss": 3.0145039558410645,
"step": 15524
},
{
"epoch": 0.940969696969697,
"grad_norm": 0.05189495533704758,
"learning_rate": 0.00019696969696969698,
"loss": 3.03045916557312,
"step": 15526
},
{
"epoch": 0.9410909090909091,
"grad_norm": 0.048595838248729706,
"learning_rate": 0.00019656565656565658,
"loss": 3.004059314727783,
"step": 15528
},
{
"epoch": 0.9412121212121212,
"grad_norm": 0.0512409508228302,
"learning_rate": 0.00019616161616161618,
"loss": 3.001244068145752,
"step": 15530
},
{
"epoch": 0.9413333333333334,
"grad_norm": 0.05690300837159157,
"learning_rate": 0.00019575757575757577,
"loss": 3.03310227394104,
"step": 15532
},
{
"epoch": 0.9414545454545454,
"grad_norm": 0.05050172656774521,
"learning_rate": 0.00019535353535353534,
"loss": 3.04647159576416,
"step": 15534
},
{
"epoch": 0.9415757575757576,
"grad_norm": 0.05274985358119011,
"learning_rate": 0.00019494949494949494,
"loss": 3.04443359375,
"step": 15536
},
{
"epoch": 0.9416969696969697,
"grad_norm": 0.05387469381093979,
"learning_rate": 0.00019454545454545457,
"loss": 3.0400564670562744,
"step": 15538
},
{
"epoch": 0.9418181818181818,
"grad_norm": 0.04906125366687775,
"learning_rate": 0.00019414141414141417,
"loss": 3.0327491760253906,
"step": 15540
},
{
"epoch": 0.941939393939394,
"grad_norm": 0.04946596175432205,
"learning_rate": 0.00019373737373737376,
"loss": 3.0425045490264893,
"step": 15542
},
{
"epoch": 0.942060606060606,
"grad_norm": 0.04992404952645302,
"learning_rate": 0.00019333333333333333,
"loss": 3.027269124984741,
"step": 15544
},
{
"epoch": 0.9421818181818182,
"grad_norm": 0.0529591403901577,
"learning_rate": 0.00019292929292929293,
"loss": 3.0505707263946533,
"step": 15546
},
{
"epoch": 0.9423030303030303,
"grad_norm": 0.051120344549417496,
"learning_rate": 0.00019252525252525253,
"loss": 3.057781219482422,
"step": 15548
},
{
"epoch": 0.9424242424242424,
"grad_norm": 0.0478610061109066,
"learning_rate": 0.00019212121212121213,
"loss": 2.998856782913208,
"step": 15550
},
{
"epoch": 0.9425454545454546,
"grad_norm": 0.05590425431728363,
"learning_rate": 0.00019171717171717173,
"loss": 3.0110368728637695,
"step": 15552
},
{
"epoch": 0.9426666666666667,
"grad_norm": 0.04849468171596527,
"learning_rate": 0.00019131313131313132,
"loss": 2.9929254055023193,
"step": 15554
},
{
"epoch": 0.9427878787878788,
"grad_norm": 0.05134075507521629,
"learning_rate": 0.00019090909090909092,
"loss": 3.020890235900879,
"step": 15556
},
{
"epoch": 0.9429090909090909,
"grad_norm": 0.04853273555636406,
"learning_rate": 0.00019050505050505052,
"loss": 3.007770538330078,
"step": 15558
},
{
"epoch": 0.943030303030303,
"grad_norm": 0.05042204260826111,
"learning_rate": 0.00019010101010101012,
"loss": 3.064668893814087,
"step": 15560
},
{
"epoch": 0.9431515151515152,
"grad_norm": 0.04948095977306366,
"learning_rate": 0.00018969696969696972,
"loss": 3.0182790756225586,
"step": 15562
},
{
"epoch": 0.9432727272727273,
"grad_norm": 0.04921668767929077,
"learning_rate": 0.0001892929292929293,
"loss": 3.009406566619873,
"step": 15564
},
{
"epoch": 0.9433939393939393,
"grad_norm": 0.0481036901473999,
"learning_rate": 0.00018888888888888888,
"loss": 2.981588363647461,
"step": 15566
},
{
"epoch": 0.9435151515151515,
"grad_norm": 0.048717956990003586,
"learning_rate": 0.00018848484848484848,
"loss": 3.024655818939209,
"step": 15568
},
{
"epoch": 0.9436363636363636,
"grad_norm": 0.049209363758563995,
"learning_rate": 0.00018808080808080808,
"loss": 3.032827377319336,
"step": 15570
},
{
"epoch": 0.9437575757575758,
"grad_norm": 0.049004729837179184,
"learning_rate": 0.0001876767676767677,
"loss": 3.0418193340301514,
"step": 15572
},
{
"epoch": 0.9438787878787879,
"grad_norm": 0.04952102154493332,
"learning_rate": 0.00018727272727272728,
"loss": 3.0262014865875244,
"step": 15574
},
{
"epoch": 0.944,
"grad_norm": 0.04889124259352684,
"learning_rate": 0.00018686868686868687,
"loss": 3.0542666912078857,
"step": 15576
},
{
"epoch": 0.9441212121212121,
"grad_norm": 0.048753343522548676,
"learning_rate": 0.00018646464646464647,
"loss": 2.9826924800872803,
"step": 15578
},
{
"epoch": 0.9442424242424242,
"grad_norm": 0.05071985721588135,
"learning_rate": 0.00018606060606060607,
"loss": 3.0763566493988037,
"step": 15580
},
{
"epoch": 0.9443636363636364,
"grad_norm": 0.049242135137319565,
"learning_rate": 0.00018565656565656567,
"loss": 3.0639493465423584,
"step": 15582
},
{
"epoch": 0.9444848484848485,
"grad_norm": 0.05701543018221855,
"learning_rate": 0.00018525252525252527,
"loss": 3.0447423458099365,
"step": 15584
},
{
"epoch": 0.9446060606060606,
"grad_norm": 0.05647512525320053,
"learning_rate": 0.00018484848484848484,
"loss": 3.072296380996704,
"step": 15586
},
{
"epoch": 0.9447272727272727,
"grad_norm": 0.05138588324189186,
"learning_rate": 0.00018444444444444446,
"loss": 3.007584810256958,
"step": 15588
},
{
"epoch": 0.9448484848484848,
"grad_norm": 0.049444276839494705,
"learning_rate": 0.00018404040404040406,
"loss": 3.0514116287231445,
"step": 15590
},
{
"epoch": 0.944969696969697,
"grad_norm": 0.04912831634283066,
"learning_rate": 0.00018363636363636366,
"loss": 3.0397298336029053,
"step": 15592
},
{
"epoch": 0.9450909090909091,
"grad_norm": 0.053070224821567535,
"learning_rate": 0.00018323232323232326,
"loss": 3.0223476886749268,
"step": 15594
},
{
"epoch": 0.9452121212121212,
"grad_norm": 0.05408964306116104,
"learning_rate": 0.00018282828282828283,
"loss": 3.0384228229522705,
"step": 15596
},
{
"epoch": 0.9453333333333334,
"grad_norm": 0.05109323188662529,
"learning_rate": 0.00018242424242424242,
"loss": 3.0653445720672607,
"step": 15598
},
{
"epoch": 0.9454545454545454,
"grad_norm": 0.05126390978693962,
"learning_rate": 0.00018202020202020202,
"loss": 3.090996503829956,
"step": 15600
},
{
"epoch": 0.9455757575757576,
"grad_norm": 0.0491819754242897,
"learning_rate": 0.00018161616161616162,
"loss": 3.0530757904052734,
"step": 15602
},
{
"epoch": 0.9456969696969697,
"grad_norm": 0.050144605338573456,
"learning_rate": 0.00018121212121212122,
"loss": 3.0149037837982178,
"step": 15604
},
{
"epoch": 0.9458181818181818,
"grad_norm": 0.04924320802092552,
"learning_rate": 0.00018080808080808082,
"loss": 3.067002296447754,
"step": 15606
},
{
"epoch": 0.945939393939394,
"grad_norm": 0.04983403533697128,
"learning_rate": 0.0001804040404040404,
"loss": 3.019683361053467,
"step": 15608
},
{
"epoch": 0.946060606060606,
"grad_norm": 0.05141449719667435,
"learning_rate": 0.00018,
"loss": 3.0369372367858887,
"step": 15610
},
{
"epoch": 0.9461818181818182,
"grad_norm": 0.04863531515002251,
"learning_rate": 0.0001795959595959596,
"loss": 3.0308358669281006,
"step": 15612
},
{
"epoch": 0.9463030303030303,
"grad_norm": 0.049230001866817474,
"learning_rate": 0.0001791919191919192,
"loss": 3.0470571517944336,
"step": 15614
},
{
"epoch": 0.9464242424242424,
"grad_norm": 0.048427753150463104,
"learning_rate": 0.0001787878787878788,
"loss": 3.0499167442321777,
"step": 15616
},
{
"epoch": 0.9465454545454546,
"grad_norm": 0.048432476818561554,
"learning_rate": 0.00017838383838383838,
"loss": 3.043076992034912,
"step": 15618
},
{
"epoch": 0.9466666666666667,
"grad_norm": 0.04924627020955086,
"learning_rate": 0.00017797979797979797,
"loss": 3.0535717010498047,
"step": 15620
},
{
"epoch": 0.9467878787878787,
"grad_norm": 0.051176052540540695,
"learning_rate": 0.0001775757575757576,
"loss": 3.048149585723877,
"step": 15622
},
{
"epoch": 0.9469090909090909,
"grad_norm": 0.04843096435070038,
"learning_rate": 0.0001771717171717172,
"loss": 3.042633295059204,
"step": 15624
},
{
"epoch": 0.947030303030303,
"grad_norm": 0.04827112704515457,
"learning_rate": 0.0001767676767676768,
"loss": 3.014803409576416,
"step": 15626
},
{
"epoch": 0.9471515151515152,
"grad_norm": 0.05004272237420082,
"learning_rate": 0.00017636363636363637,
"loss": 3.0407724380493164,
"step": 15628
},
{
"epoch": 0.9472727272727273,
"grad_norm": 0.04909525811672211,
"learning_rate": 0.00017595959595959596,
"loss": 3.035121202468872,
"step": 15630
},
{
"epoch": 0.9473939393939393,
"grad_norm": 0.05698690935969353,
"learning_rate": 0.00017555555555555556,
"loss": 3.038562059402466,
"step": 15632
},
{
"epoch": 0.9475151515151515,
"grad_norm": 0.04843836650252342,
"learning_rate": 0.00017515151515151516,
"loss": 3.0429539680480957,
"step": 15634
},
{
"epoch": 0.9476363636363636,
"grad_norm": 0.049286969006061554,
"learning_rate": 0.00017474747474747476,
"loss": 3.086639881134033,
"step": 15636
},
{
"epoch": 0.9477575757575758,
"grad_norm": 0.04797523096203804,
"learning_rate": 0.00017434343434343435,
"loss": 3.0544376373291016,
"step": 15638
},
{
"epoch": 0.9478787878787879,
"grad_norm": 0.05267041176557541,
"learning_rate": 0.00017393939393939395,
"loss": 3.0665767192840576,
"step": 15640
},
{
"epoch": 0.948,
"grad_norm": 0.04922143369913101,
"learning_rate": 0.00017353535353535355,
"loss": 3.023818254470825,
"step": 15642
},
{
"epoch": 0.9481212121212121,
"grad_norm": 0.048744723200798035,
"learning_rate": 0.00017313131313131315,
"loss": 3.0475008487701416,
"step": 15644
},
{
"epoch": 0.9482424242424242,
"grad_norm": 0.048895835876464844,
"learning_rate": 0.00017272727272727275,
"loss": 3.0317225456237793,
"step": 15646
},
{
"epoch": 0.9483636363636364,
"grad_norm": 0.04969105124473572,
"learning_rate": 0.00017232323232323232,
"loss": 3.0166869163513184,
"step": 15648
},
{
"epoch": 0.9484848484848485,
"grad_norm": 0.04833272099494934,
"learning_rate": 0.00017191919191919192,
"loss": 3.034177303314209,
"step": 15650
},
{
"epoch": 0.9486060606060606,
"grad_norm": 0.05050016567111015,
"learning_rate": 0.0001715151515151515,
"loss": 3.0504894256591797,
"step": 15652
},
{
"epoch": 0.9487272727272728,
"grad_norm": 0.05115874484181404,
"learning_rate": 0.0001711111111111111,
"loss": 3.0455822944641113,
"step": 15654
},
{
"epoch": 0.9488484848484848,
"grad_norm": 0.049892958253622055,
"learning_rate": 0.0001707070707070707,
"loss": 3.1053671836853027,
"step": 15656
},
{
"epoch": 0.948969696969697,
"grad_norm": 0.051689811050891876,
"learning_rate": 0.0001703030303030303,
"loss": 3.044862747192383,
"step": 15658
},
{
"epoch": 0.9490909090909091,
"grad_norm": 0.050586894154548645,
"learning_rate": 0.0001698989898989899,
"loss": 3.032987117767334,
"step": 15660
},
{
"epoch": 0.9492121212121212,
"grad_norm": 0.04895954579114914,
"learning_rate": 0.0001694949494949495,
"loss": 3.06919264793396,
"step": 15662
},
{
"epoch": 0.9493333333333334,
"grad_norm": 0.05009319260716438,
"learning_rate": 0.0001690909090909091,
"loss": 3.071161985397339,
"step": 15664
},
{
"epoch": 0.9494545454545454,
"grad_norm": 0.051976896822452545,
"learning_rate": 0.0001686868686868687,
"loss": 3.0612051486968994,
"step": 15666
},
{
"epoch": 0.9495757575757576,
"grad_norm": 0.05031817778944969,
"learning_rate": 0.0001682828282828283,
"loss": 3.0508081912994385,
"step": 15668
},
{
"epoch": 0.9496969696969697,
"grad_norm": 0.048356011509895325,
"learning_rate": 0.00016787878787878787,
"loss": 3.0699217319488525,
"step": 15670
},
{
"epoch": 0.9498181818181818,
"grad_norm": 0.05238514021039009,
"learning_rate": 0.00016747474747474747,
"loss": 3.0066373348236084,
"step": 15672
},
{
"epoch": 0.949939393939394,
"grad_norm": 0.04901914298534393,
"learning_rate": 0.0001670707070707071,
"loss": 3.0517160892486572,
"step": 15674
},
{
"epoch": 0.950060606060606,
"grad_norm": 0.05048520490527153,
"learning_rate": 0.0001666666666666667,
"loss": 3.0555028915405273,
"step": 15676
},
{
"epoch": 0.9501818181818181,
"grad_norm": 0.05010529235005379,
"learning_rate": 0.00016626262626262629,
"loss": 3.0781137943267822,
"step": 15678
},
{
"epoch": 0.9503030303030303,
"grad_norm": 0.05323826149106026,
"learning_rate": 0.00016585858585858586,
"loss": 3.1238059997558594,
"step": 15680
},
{
"epoch": 0.9504242424242424,
"grad_norm": 0.053634583950042725,
"learning_rate": 0.00016545454545454545,
"loss": 3.034567356109619,
"step": 15682
},
{
"epoch": 0.9505454545454546,
"grad_norm": 0.04938892275094986,
"learning_rate": 0.00016505050505050505,
"loss": 3.068929433822632,
"step": 15684
},
{
"epoch": 0.9506666666666667,
"grad_norm": 0.04895763844251633,
"learning_rate": 0.00016464646464646465,
"loss": 3.060349464416504,
"step": 15686
},
{
"epoch": 0.9507878787878787,
"grad_norm": 0.0484452024102211,
"learning_rate": 0.00016424242424242425,
"loss": 3.071992874145508,
"step": 15688
},
{
"epoch": 0.9509090909090909,
"grad_norm": 0.05171800032258034,
"learning_rate": 0.00016383838383838385,
"loss": 3.048373222351074,
"step": 15690
},
{
"epoch": 0.951030303030303,
"grad_norm": 0.049861203879117966,
"learning_rate": 0.00016343434343434344,
"loss": 3.1033411026000977,
"step": 15692
},
{
"epoch": 0.9511515151515152,
"grad_norm": 0.050121549516916275,
"learning_rate": 0.00016303030303030304,
"loss": 3.063870906829834,
"step": 15694
},
{
"epoch": 0.9512727272727273,
"grad_norm": 0.0517585463821888,
"learning_rate": 0.00016262626262626264,
"loss": 3.1038029193878174,
"step": 15696
},
{
"epoch": 0.9513939393939393,
"grad_norm": 0.05002948269248009,
"learning_rate": 0.00016222222222222224,
"loss": 3.060966730117798,
"step": 15698
},
{
"epoch": 0.9515151515151515,
"grad_norm": 0.04929853975772858,
"learning_rate": 0.00016181818181818184,
"loss": 3.042423725128174,
"step": 15700
},
{
"epoch": 0.9516363636363636,
"grad_norm": 0.04830727353692055,
"learning_rate": 0.0001614141414141414,
"loss": 3.0788261890411377,
"step": 15702
},
{
"epoch": 0.9517575757575758,
"grad_norm": 0.04835592210292816,
"learning_rate": 0.000161010101010101,
"loss": 3.041840076446533,
"step": 15704
},
{
"epoch": 0.9518787878787879,
"grad_norm": 0.04859842732548714,
"learning_rate": 0.0001606060606060606,
"loss": 3.039973735809326,
"step": 15706
},
{
"epoch": 0.952,
"grad_norm": 0.049100473523139954,
"learning_rate": 0.00016020202020202023,
"loss": 3.0315284729003906,
"step": 15708
},
{
"epoch": 0.9521212121212121,
"grad_norm": 0.050308942794799805,
"learning_rate": 0.00015979797979797983,
"loss": 3.0698208808898926,
"step": 15710
},
{
"epoch": 0.9522424242424242,
"grad_norm": 0.05116778984665871,
"learning_rate": 0.0001593939393939394,
"loss": 3.0448875427246094,
"step": 15712
},
{
"epoch": 0.9523636363636364,
"grad_norm": 0.0494735985994339,
"learning_rate": 0.000158989898989899,
"loss": 3.0764763355255127,
"step": 15714
},
{
"epoch": 0.9524848484848485,
"grad_norm": 0.060145922005176544,
"learning_rate": 0.0001585858585858586,
"loss": 3.077226400375366,
"step": 15716
},
{
"epoch": 0.9526060606060606,
"grad_norm": 0.04869908466935158,
"learning_rate": 0.0001581818181818182,
"loss": 3.072455406188965,
"step": 15718
},
{
"epoch": 0.9527272727272728,
"grad_norm": 0.05627526342868805,
"learning_rate": 0.0001577777777777778,
"loss": 3.0303144454956055,
"step": 15720
},
{
"epoch": 0.9528484848484848,
"grad_norm": 0.04811324551701546,
"learning_rate": 0.00015737373737373739,
"loss": 3.0413079261779785,
"step": 15722
},
{
"epoch": 0.952969696969697,
"grad_norm": 0.051033999770879745,
"learning_rate": 0.00015696969696969696,
"loss": 3.095437526702881,
"step": 15724
},
{
"epoch": 0.9530909090909091,
"grad_norm": 0.04885207861661911,
"learning_rate": 0.00015656565656565658,
"loss": 3.08821439743042,
"step": 15726
},
{
"epoch": 0.9532121212121212,
"grad_norm": 0.05047230049967766,
"learning_rate": 0.00015616161616161618,
"loss": 3.063981533050537,
"step": 15728
},
{
"epoch": 0.9533333333333334,
"grad_norm": 0.04993906244635582,
"learning_rate": 0.00015575757575757578,
"loss": 3.0649547576904297,
"step": 15730
},
{
"epoch": 0.9534545454545454,
"grad_norm": 0.04901989549398422,
"learning_rate": 0.00015535353535353538,
"loss": 3.0588221549987793,
"step": 15732
},
{
"epoch": 0.9535757575757575,
"grad_norm": 0.05232423171401024,
"learning_rate": 0.00015494949494949495,
"loss": 3.0712108612060547,
"step": 15734
},
{
"epoch": 0.9536969696969697,
"grad_norm": 0.051649320870637894,
"learning_rate": 0.00015454545454545454,
"loss": 3.0475387573242188,
"step": 15736
},
{
"epoch": 0.9538181818181818,
"grad_norm": 0.05005862936377525,
"learning_rate": 0.00015414141414141414,
"loss": 3.088308095932007,
"step": 15738
},
{
"epoch": 0.953939393939394,
"grad_norm": 0.049746960401535034,
"learning_rate": 0.00015373737373737374,
"loss": 3.082504987716675,
"step": 15740
},
{
"epoch": 0.954060606060606,
"grad_norm": 0.04855158552527428,
"learning_rate": 0.00015333333333333334,
"loss": 3.0292022228240967,
"step": 15742
},
{
"epoch": 0.9541818181818181,
"grad_norm": 0.04768714681267738,
"learning_rate": 0.00015292929292929294,
"loss": 3.0590527057647705,
"step": 15744
},
{
"epoch": 0.9543030303030303,
"grad_norm": 0.04855417087674141,
"learning_rate": 0.00015252525252525253,
"loss": 3.1062064170837402,
"step": 15746
},
{
"epoch": 0.9544242424242424,
"grad_norm": 0.05530768632888794,
"learning_rate": 0.00015212121212121213,
"loss": 3.082184314727783,
"step": 15748
},
{
"epoch": 0.9545454545454546,
"grad_norm": 0.04959571734070778,
"learning_rate": 0.00015171717171717173,
"loss": 3.034513473510742,
"step": 15750
},
{
"epoch": 0.9546666666666667,
"grad_norm": 0.05376448482275009,
"learning_rate": 0.00015131313131313133,
"loss": 3.058668851852417,
"step": 15752
},
{
"epoch": 0.9547878787878787,
"grad_norm": 0.04869585484266281,
"learning_rate": 0.0001509090909090909,
"loss": 3.086904764175415,
"step": 15754
},
{
"epoch": 0.9549090909090909,
"grad_norm": 0.051713403314352036,
"learning_rate": 0.0001505050505050505,
"loss": 3.0616326332092285,
"step": 15756
},
{
"epoch": 0.955030303030303,
"grad_norm": 0.05046727880835533,
"learning_rate": 0.0001501010101010101,
"loss": 3.0670979022979736,
"step": 15758
},
{
"epoch": 0.9551515151515152,
"grad_norm": 0.04920874163508415,
"learning_rate": 0.00014969696969696972,
"loss": 3.0991592407226562,
"step": 15760
},
{
"epoch": 0.9552727272727273,
"grad_norm": 0.05251651257276535,
"learning_rate": 0.00014929292929292932,
"loss": 3.0664758682250977,
"step": 15762
},
{
"epoch": 0.9553939393939394,
"grad_norm": 0.049125295132398605,
"learning_rate": 0.0001488888888888889,
"loss": 3.0915982723236084,
"step": 15764
},
{
"epoch": 0.9555151515151515,
"grad_norm": 0.048347122967243195,
"learning_rate": 0.00014848484848484849,
"loss": 3.070382595062256,
"step": 15766
},
{
"epoch": 0.9556363636363636,
"grad_norm": 0.04870107024908066,
"learning_rate": 0.00014808080808080808,
"loss": 3.075108051300049,
"step": 15768
},
{
"epoch": 0.9557575757575758,
"grad_norm": 0.0488167405128479,
"learning_rate": 0.00014767676767676768,
"loss": 3.0865166187286377,
"step": 15770
},
{
"epoch": 0.9558787878787879,
"grad_norm": 0.05521353706717491,
"learning_rate": 0.00014727272727272728,
"loss": 3.079735040664673,
"step": 15772
},
{
"epoch": 0.956,
"grad_norm": 0.04917971417307854,
"learning_rate": 0.00014686868686868688,
"loss": 3.0745558738708496,
"step": 15774
},
{
"epoch": 0.9561212121212121,
"grad_norm": 0.05129864066839218,
"learning_rate": 0.00014646464646464648,
"loss": 3.0585780143737793,
"step": 15776
},
{
"epoch": 0.9562424242424242,
"grad_norm": 0.05071466416120529,
"learning_rate": 0.00014606060606060607,
"loss": 3.1265082359313965,
"step": 15778
},
{
"epoch": 0.9563636363636364,
"grad_norm": 0.05265762656927109,
"learning_rate": 0.00014565656565656567,
"loss": 3.0922024250030518,
"step": 15780
},
{
"epoch": 0.9564848484848485,
"grad_norm": 0.05322057753801346,
"learning_rate": 0.00014525252525252527,
"loss": 3.0842103958129883,
"step": 15782
},
{
"epoch": 0.9566060606060606,
"grad_norm": 0.04821353778243065,
"learning_rate": 0.00014484848484848487,
"loss": 3.0852701663970947,
"step": 15784
},
{
"epoch": 0.9567272727272728,
"grad_norm": 0.04867221415042877,
"learning_rate": 0.00014444444444444444,
"loss": 3.073138475418091,
"step": 15786
},
{
"epoch": 0.9568484848484848,
"grad_norm": 0.05084449425339699,
"learning_rate": 0.00014404040404040404,
"loss": 3.09576416015625,
"step": 15788
},
{
"epoch": 0.9569696969696969,
"grad_norm": 0.04853491112589836,
"learning_rate": 0.00014363636363636363,
"loss": 3.089974880218506,
"step": 15790
},
{
"epoch": 0.9570909090909091,
"grad_norm": 0.048672761768102646,
"learning_rate": 0.00014323232323232323,
"loss": 3.050147771835327,
"step": 15792
},
{
"epoch": 0.9572121212121212,
"grad_norm": 0.0500684417784214,
"learning_rate": 0.00014282828282828286,
"loss": 3.0851659774780273,
"step": 15794
},
{
"epoch": 0.9573333333333334,
"grad_norm": 0.051558975130319595,
"learning_rate": 0.00014242424242424243,
"loss": 3.1109042167663574,
"step": 15796
},
{
"epoch": 0.9574545454545454,
"grad_norm": 0.05559660121798515,
"learning_rate": 0.00014202020202020203,
"loss": 3.1125175952911377,
"step": 15798
},
{
"epoch": 0.9575757575757575,
"grad_norm": 0.05025303363800049,
"learning_rate": 0.00014161616161616162,
"loss": 3.0782811641693115,
"step": 15800
},
{
"epoch": 0.9576969696969697,
"grad_norm": 0.05206725373864174,
"learning_rate": 0.00014121212121212122,
"loss": 3.1086833477020264,
"step": 15802
},
{
"epoch": 0.9578181818181818,
"grad_norm": 0.05180590599775314,
"learning_rate": 0.00014080808080808082,
"loss": 3.0979292392730713,
"step": 15804
},
{
"epoch": 0.957939393939394,
"grad_norm": 0.057155441492795944,
"learning_rate": 0.00014040404040404042,
"loss": 3.0437703132629395,
"step": 15806
},
{
"epoch": 0.9580606060606061,
"grad_norm": 0.04766825586557388,
"learning_rate": 0.00014,
"loss": 3.0877349376678467,
"step": 15808
},
{
"epoch": 0.9581818181818181,
"grad_norm": 0.04803939536213875,
"learning_rate": 0.0001395959595959596,
"loss": 3.0672221183776855,
"step": 15810
},
{
"epoch": 0.9583030303030303,
"grad_norm": 0.0549965463578701,
"learning_rate": 0.0001391919191919192,
"loss": 3.0717720985412598,
"step": 15812
},
{
"epoch": 0.9584242424242424,
"grad_norm": 0.0514734722673893,
"learning_rate": 0.0001387878787878788,
"loss": 3.0707602500915527,
"step": 15814
},
{
"epoch": 0.9585454545454546,
"grad_norm": 0.04876880720257759,
"learning_rate": 0.0001383838383838384,
"loss": 3.0840048789978027,
"step": 15816
},
{
"epoch": 0.9586666666666667,
"grad_norm": 0.055530328303575516,
"learning_rate": 0.00013797979797979798,
"loss": 3.1017704010009766,
"step": 15818
},
{
"epoch": 0.9587878787878787,
"grad_norm": 0.0503402054309845,
"learning_rate": 0.00013757575757575758,
"loss": 3.0865015983581543,
"step": 15820
},
{
"epoch": 0.9589090909090909,
"grad_norm": 0.051434144377708435,
"learning_rate": 0.00013717171717171717,
"loss": 3.088186740875244,
"step": 15822
},
{
"epoch": 0.959030303030303,
"grad_norm": 0.053344182670116425,
"learning_rate": 0.00013676767676767677,
"loss": 3.0878305435180664,
"step": 15824
},
{
"epoch": 0.9591515151515152,
"grad_norm": 0.04888543114066124,
"learning_rate": 0.00013636363636363637,
"loss": 3.078613758087158,
"step": 15826
},
{
"epoch": 0.9592727272727273,
"grad_norm": 0.05574370175600052,
"learning_rate": 0.00013595959595959597,
"loss": 3.1451833248138428,
"step": 15828
},
{
"epoch": 0.9593939393939394,
"grad_norm": 0.055157892405986786,
"learning_rate": 0.00013555555555555556,
"loss": 3.088019609451294,
"step": 15830
},
{
"epoch": 0.9595151515151515,
"grad_norm": 0.04848414286971092,
"learning_rate": 0.00013515151515151516,
"loss": 3.117337465286255,
"step": 15832
},
{
"epoch": 0.9596363636363636,
"grad_norm": 0.059998758137226105,
"learning_rate": 0.00013474747474747476,
"loss": 3.1215593814849854,
"step": 15834
},
{
"epoch": 0.9597575757575758,
"grad_norm": 0.049959585070610046,
"learning_rate": 0.00013434343434343436,
"loss": 3.1201400756835938,
"step": 15836
},
{
"epoch": 0.9598787878787879,
"grad_norm": 0.04953428730368614,
"learning_rate": 0.00013393939393939393,
"loss": 3.060267448425293,
"step": 15838
},
{
"epoch": 0.96,
"grad_norm": 0.04891553148627281,
"learning_rate": 0.00013353535353535353,
"loss": 3.091520071029663,
"step": 15840
}
],
"logging_steps": 2,
"max_steps": 16500,
"num_input_tokens_seen": 0,
"num_train_epochs": 9223372036854775807,
"save_steps": 660,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 3.7619745465690006e+19,
"train_batch_size": 20,
"trial_name": null,
"trial_params": null
}