diff --git "a/checkpoint-10000/trainer_state.json" "b/checkpoint-10000/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-10000/trainer_state.json"
@@ -0,0 +1,7033 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.2603918028660458,
+  "eval_steps": 500,
+  "global_step": 10000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0002603918028660458,
+      "grad_norm": 0.4500846266746521,
+      "learning_rate": 5.194805194805195e-06,
+      "loss": 1.0381,
+      "step": 10
+    },
+    {
+      "epoch": 0.0005207836057320916,
+      "grad_norm": 0.35188010334968567,
+      "learning_rate": 1.038961038961039e-05,
+      "loss": 1.0108,
+      "step": 20
+    },
+    {
+      "epoch": 0.0007811754085981374,
+      "grad_norm": 0.2300374060869217,
+      "learning_rate": 1.5584415584415583e-05,
+      "loss": 0.9668,
+      "step": 30
+    },
+    {
+      "epoch": 0.0010415672114641832,
+      "grad_norm": 0.16189467906951904,
+      "learning_rate": 2.077922077922078e-05,
+      "loss": 0.918,
+      "step": 40
+    },
+    {
+      "epoch": 0.001301959014330229,
+      "grad_norm": 0.18843211233615875,
+      "learning_rate": 2.5974025974025972e-05,
+      "loss": 0.9265,
+      "step": 50
+    },
+    {
+      "epoch": 0.0015623508171962747,
+      "grad_norm": 0.20334510505199432,
+      "learning_rate": 3.1168831168831166e-05,
+      "loss": 0.9234,
+      "step": 60
+    },
+    {
+      "epoch": 0.0018227426200623205,
+      "grad_norm": 0.1745327115058899,
+      "learning_rate": 3.6363636363636364e-05,
+      "loss": 0.881,
+      "step": 70
+    },
+    {
+      "epoch": 0.0020831344229283663,
+      "grad_norm": 0.18667331337928772,
+      "learning_rate": 4.155844155844156e-05,
+      "loss": 0.8592,
+      "step": 80
+    },
+    {
+      "epoch": 0.002343526225794412,
+      "grad_norm": 0.1848158985376358,
+      "learning_rate": 4.675324675324675e-05,
+      "loss": 0.8537,
+      "step": 90
+    },
+    {
+      "epoch": 0.002603918028660458,
+      "grad_norm": 0.17589879035949707,
+      "learning_rate": 5.1948051948051944e-05,
+      "loss": 0.8518,
+      "step": 100
+    },
+    {
+      "epoch": 0.0028643098315265037,
+      "grad_norm": 0.2132624089717865,
+      "learning_rate": 5.714285714285714e-05,
+      "loss": 0.8511,
+      "step": 110
+    },
+    {
+      "epoch": 0.0031247016343925495,
+      "grad_norm": 0.23070092499256134,
+      "learning_rate": 6.233766233766233e-05,
+      "loss": 0.7975,
+      "step": 120
+    },
+    {
+      "epoch": 0.0033850934372585953,
+      "grad_norm": 0.25368157029151917,
+      "learning_rate": 6.753246753246754e-05,
+      "loss": 0.8134,
+      "step": 130
+    },
+    {
+      "epoch": 0.003645485240124641,
+      "grad_norm": 0.22897231578826904,
+      "learning_rate": 7.272727272727273e-05,
+      "loss": 0.8322,
+      "step": 140
+    },
+    {
+      "epoch": 0.003905877042990687,
+      "grad_norm": 0.19932536780834198,
+      "learning_rate": 7.792207792207793e-05,
+      "loss": 0.7959,
+      "step": 150
+    },
+    {
+      "epoch": 0.004166268845856733,
+      "grad_norm": 0.21011792123317719,
+      "learning_rate": 8.311688311688312e-05,
+      "loss": 0.8102,
+      "step": 160
+    },
+    {
+      "epoch": 0.004426660648722778,
+      "grad_norm": 0.20594824850559235,
+      "learning_rate": 8.831168831168831e-05,
+      "loss": 0.8128,
+      "step": 170
+    },
+    {
+      "epoch": 0.004687052451588824,
+      "grad_norm": 0.20465536415576935,
+      "learning_rate": 9.35064935064935e-05,
+      "loss": 0.7989,
+      "step": 180
+    },
+    {
+      "epoch": 0.00494744425445487,
+      "grad_norm": 0.4109392762184143,
+      "learning_rate": 9.870129870129871e-05,
+      "loss": 0.8108,
+      "step": 190
+    },
+    {
+      "epoch": 0.005207836057320916,
+      "grad_norm": 0.4293076694011688,
+      "learning_rate": 0.00010389610389610389,
+      "loss": 0.8101,
+      "step": 200
+    },
+    {
+      "epoch": 0.005468227860186962,
+      "grad_norm": 0.31628963351249695,
+      "learning_rate": 0.00010909090909090909,
+      "loss": 0.7989,
+      "step": 210
+    },
+    {
+      "epoch": 0.005728619663053007,
+      "grad_norm": 0.24642810225486755,
+      "learning_rate": 0.00011428571428571428,
+      "loss": 0.7751,
+      "step": 220
+    },
+    {
+      "epoch": 0.005989011465919053,
+      "grad_norm": 0.3599106967449188,
+      "learning_rate": 0.00011948051948051949,
+      "loss": 0.8063,
+      "step": 230
+    },
+    {
+      "epoch": 0.006249403268785099,
+      "grad_norm": 0.17053447663784027,
+      "learning_rate": 0.00012467532467532467,
+      "loss": 0.7751,
+      "step": 240
+    },
+    {
+      "epoch": 0.006509795071651145,
+      "grad_norm": 0.17303769290447235,
+      "learning_rate": 0.00012987012987012987,
+      "loss": 0.7883,
+      "step": 250
+    },
+    {
+      "epoch": 0.0067701868745171905,
+      "grad_norm": 0.1815861016511917,
+      "learning_rate": 0.00013506493506493507,
+      "loss": 0.788,
+      "step": 260
+    },
+    {
+      "epoch": 0.007030578677383236,
+      "grad_norm": 0.24125365912914276,
+      "learning_rate": 0.00014025974025974028,
+      "loss": 0.8018,
+      "step": 270
+    },
+    {
+      "epoch": 0.007290970480249282,
+      "grad_norm": 0.19443446397781372,
+      "learning_rate": 0.00014545454545454546,
+      "loss": 0.7908,
+      "step": 280
+    },
+    {
+      "epoch": 0.007551362283115328,
+      "grad_norm": 0.17829768359661102,
+      "learning_rate": 0.00015064935064935066,
+      "loss": 0.8033,
+      "step": 290
+    },
+    {
+      "epoch": 0.007811754085981374,
+      "grad_norm": 0.19535653293132782,
+      "learning_rate": 0.00015584415584415587,
+      "loss": 0.7997,
+      "step": 300
+    },
+    {
+      "epoch": 0.008072145888847419,
+      "grad_norm": 0.19930541515350342,
+      "learning_rate": 0.00016103896103896104,
+      "loss": 0.7945,
+      "step": 310
+    },
+    {
+      "epoch": 0.008332537691713465,
+      "grad_norm": 0.2156297266483307,
+      "learning_rate": 0.00016623376623376625,
+      "loss": 0.8018,
+      "step": 320
+    },
+    {
+      "epoch": 0.00859292949457951,
+      "grad_norm": 0.1924206018447876,
+      "learning_rate": 0.00017142857142857143,
+      "loss": 0.7746,
+      "step": 330
+    },
+    {
+      "epoch": 0.008853321297445557,
+      "grad_norm": 0.2294880747795105,
+      "learning_rate": 0.00017662337662337663,
+      "loss": 0.8152,
+      "step": 340
+    },
+    {
+      "epoch": 0.009113713100311602,
+      "grad_norm": 0.16817067563533783,
+      "learning_rate": 0.00018181818181818183,
+      "loss": 0.7972,
+      "step": 350
+    },
+    {
+      "epoch": 0.009374104903177648,
+      "grad_norm": 0.18544812500476837,
+      "learning_rate": 0.000187012987012987,
+      "loss": 0.7801,
+      "step": 360
+    },
+    {
+      "epoch": 0.009634496706043693,
+      "grad_norm": 0.19597066938877106,
+      "learning_rate": 0.00019220779220779222,
+      "loss": 0.7706,
+      "step": 370
+    },
+    {
+      "epoch": 0.00989488850890974,
+      "grad_norm": 0.40291881561279297,
+      "learning_rate": 0.00019740259740259742,
+      "loss": 0.7911,
+      "step": 380
+    },
+    {
+      "epoch": 0.010155280311775785,
+      "grad_norm": 0.23841074109077454,
+      "learning_rate": 0.00019999996515752773,
+      "loss": 0.7861,
+      "step": 390
+    },
+    {
+      "epoch": 0.010415672114641832,
+      "grad_norm": 0.1675388514995575,
+      "learning_rate": 0.00019999968641789507,
+      "loss": 0.788,
+      "step": 400
+    },
+    {
+      "epoch": 0.010676063917507876,
+      "grad_norm": 1.8860758543014526,
+      "learning_rate": 0.0001999991289394067,
+      "loss": 0.7632,
+      "step": 410
+    },
+    {
+      "epoch": 0.010936455720373923,
+      "grad_norm": 0.17022117972373962,
+      "learning_rate": 0.00019999829272361654,
+      "loss": 0.784,
+      "step": 420
+    },
+    {
+      "epoch": 0.011196847523239968,
+      "grad_norm": 0.21460269391536713,
+      "learning_rate": 0.00019999717777285545,
+      "loss": 0.761,
+      "step": 430
+    },
+    {
+      "epoch": 0.011457239326106015,
+      "grad_norm": 0.19413785636425018,
+      "learning_rate": 0.00019999578409023126,
+      "loss": 0.7772,
+      "step": 440
+    },
+    {
+      "epoch": 0.01171763112897206,
+      "grad_norm": 0.20223405957221985,
+      "learning_rate": 0.00019999411167962868,
+      "loss": 0.7811,
+      "step": 450
+    },
+    {
+      "epoch": 0.011978022931838106,
+      "grad_norm": 0.15166303515434265,
+      "learning_rate": 0.00019999216054570942,
+      "loss": 0.7709,
+      "step": 460
+    },
+    {
+      "epoch": 0.012238414734704151,
+      "grad_norm": 0.16307081282138824,
+      "learning_rate": 0.00019998993069391205,
+      "loss": 0.7811,
+      "step": 470
+    },
+    {
+      "epoch": 0.012498806537570198,
+      "grad_norm": 0.15996049344539642,
+      "learning_rate": 0.00019998742213045206,
+      "loss": 0.7599,
+      "step": 480
+    },
+    {
+      "epoch": 0.012759198340436243,
+      "grad_norm": 0.17560279369354248,
+      "learning_rate": 0.00019998463486232179,
+      "loss": 0.7572,
+      "step": 490
+    },
+    {
+      "epoch": 0.01301959014330229,
+      "grad_norm": 0.17571642994880676,
+      "learning_rate": 0.0001999815688972905,
+      "loss": 0.7643,
+      "step": 500
+    },
+    {
+      "epoch": 0.013279981946168334,
+      "grad_norm": 0.17719799280166626,
+      "learning_rate": 0.00019997822424390422,
+      "loss": 0.7923,
+      "step": 510
+    },
+    {
+      "epoch": 0.013540373749034381,
+      "grad_norm": 0.19846616685390472,
+      "learning_rate": 0.00019997460091148586,
+      "loss": 0.7674,
+      "step": 520
+    },
+    {
+      "epoch": 0.013800765551900426,
+      "grad_norm": 0.2715558111667633,
+      "learning_rate": 0.00019997069891013503,
+      "loss": 0.7421,
+      "step": 530
+    },
+    {
+      "epoch": 0.014061157354766473,
+      "grad_norm": 0.1725197583436966,
+      "learning_rate": 0.00019996651825072826,
+      "loss": 0.7663,
+      "step": 540
+    },
+    {
+      "epoch": 0.014321549157632518,
+      "grad_norm": 0.15060502290725708,
+      "learning_rate": 0.00019996205894491856,
+      "loss": 0.7794,
+      "step": 550
+    },
+    {
+      "epoch": 0.014581940960498564,
+      "grad_norm": 0.16645808517932892,
+      "learning_rate": 0.00019995732100513592,
+      "loss": 0.752,
+      "step": 560
+    },
+    {
+      "epoch": 0.014842332763364609,
+      "grad_norm": 0.1736789345741272,
+      "learning_rate": 0.00019995230444458682,
+      "loss": 0.7788,
+      "step": 570
+    },
+    {
+      "epoch": 0.015102724566230656,
+      "grad_norm": 0.15416319668293,
+      "learning_rate": 0.0001999470092772544,
+      "loss": 0.7656,
+      "step": 580
+    },
+    {
+      "epoch": 0.0153631163690967,
+      "grad_norm": 0.16610187292099,
+      "learning_rate": 0.00019994143551789839,
+      "loss": 0.7676,
+      "step": 590
+    },
+    {
+      "epoch": 0.015623508171962747,
+      "grad_norm": 0.15843011438846588,
+      "learning_rate": 0.00019993558318205507,
+      "loss": 0.7746,
+      "step": 600
+    },
+    {
+      "epoch": 0.015883899974828794,
+      "grad_norm": 0.26837801933288574,
+      "learning_rate": 0.00019992945228603724,
+      "loss": 0.7617,
+      "step": 610
+    },
+    {
+      "epoch": 0.016144291777694837,
+      "grad_norm": 0.15099173784255981,
+      "learning_rate": 0.0001999230428469341,
+      "loss": 0.7601,
+      "step": 620
+    },
+    {
+      "epoch": 0.016404683580560884,
+      "grad_norm": 0.15511856973171234,
+      "learning_rate": 0.00019991635488261138,
+      "loss": 0.7647,
+      "step": 630
+    },
+    {
+      "epoch": 0.01666507538342693,
+      "grad_norm": 0.14919579029083252,
+      "learning_rate": 0.00019990938841171104,
+      "loss": 0.7692,
+      "step": 640
+    },
+    {
+      "epoch": 0.016925467186292977,
+      "grad_norm": 0.15838642418384552,
+      "learning_rate": 0.0001999021434536514,
+      "loss": 0.7763,
+      "step": 650
+    },
+    {
+      "epoch": 0.01718585898915902,
+      "grad_norm": 0.15956635773181915,
+      "learning_rate": 0.00019989462002862704,
+      "loss": 0.7598,
+      "step": 660
+    },
+    {
+      "epoch": 0.017446250792025067,
+      "grad_norm": 0.1499069333076477,
+      "learning_rate": 0.0001998868181576088,
+      "loss": 0.7626,
+      "step": 670
+    },
+    {
+      "epoch": 0.017706642594891114,
+      "grad_norm": 0.2170073390007019,
+      "learning_rate": 0.00019987873786234348,
+      "loss": 0.7569,
+      "step": 680
+    },
+    {
+      "epoch": 0.01796703439775716,
+      "grad_norm": 0.17841948568820953,
+      "learning_rate": 0.00019987037916535417,
+      "loss": 0.7494,
+      "step": 690
+    },
+    {
+      "epoch": 0.018227426200623204,
+      "grad_norm": 0.2066909372806549,
+      "learning_rate": 0.0001998617420899398,
+      "loss": 0.7609,
+      "step": 700
+    },
+    {
+      "epoch": 0.01848781800348925,
+      "grad_norm": 0.17015361785888672,
+      "learning_rate": 0.0001998528266601754,
+      "loss": 0.7761,
+      "step": 710
+    },
+    {
+      "epoch": 0.018748209806355297,
+      "grad_norm": 0.22166290879249573,
+      "learning_rate": 0.0001998436329009118,
+      "loss": 0.7573,
+      "step": 720
+    },
+    {
+      "epoch": 0.01900860160922134,
+      "grad_norm": 0.15084640681743622,
+      "learning_rate": 0.00019983416083777563,
+      "loss": 0.7775,
+      "step": 730
+    },
+    {
+      "epoch": 0.019268993412087387,
+      "grad_norm": 0.17800921201705933,
+      "learning_rate": 0.0001998244104971693,
+      "loss": 0.7359,
+      "step": 740
+    },
+    {
+      "epoch": 0.019529385214953433,
+      "grad_norm": 0.17354707419872284,
+      "learning_rate": 0.0001998143819062709,
+      "loss": 0.7415,
+      "step": 750
+    },
+    {
+      "epoch": 0.01978977701781948,
+      "grad_norm": 0.16408118605613708,
+      "learning_rate": 0.00019980407509303413,
+      "loss": 0.7708,
+      "step": 760
+    },
+    {
+      "epoch": 0.020050168820685523,
+      "grad_norm": 0.16820089519023895,
+      "learning_rate": 0.00019979349008618808,
+      "loss": 0.791,
+      "step": 770
+    },
+    {
+      "epoch": 0.02031056062355157,
+      "grad_norm": 0.15958388149738312,
+      "learning_rate": 0.00019978262691523743,
+      "loss": 0.7412,
+      "step": 780
+    },
+    {
+      "epoch": 0.020570952426417616,
+      "grad_norm": 0.1646542251110077,
+      "learning_rate": 0.00019977148561046217,
+      "loss": 0.7529,
+      "step": 790
+    },
+    {
+      "epoch": 0.020831344229283663,
+      "grad_norm": 0.17032025754451752,
+      "learning_rate": 0.0001997600662029175,
+      "loss": 0.7656,
+      "step": 800
+    },
+    {
+      "epoch": 0.021091736032149706,
+      "grad_norm": 0.17189227044582367,
+      "learning_rate": 0.00019974836872443388,
+      "loss": 0.7433,
+      "step": 810
+    },
+    {
+      "epoch": 0.021352127835015753,
+      "grad_norm": 0.16334249079227448,
+      "learning_rate": 0.0001997363932076168,
+      "loss": 0.7703,
+      "step": 820
+    },
+    {
+      "epoch": 0.0216125196378818,
+      "grad_norm": 0.1676424890756607,
+      "learning_rate": 0.00019972413968584682,
+      "loss": 0.7603,
+      "step": 830
+    },
+    {
+      "epoch": 0.021872911440747846,
+      "grad_norm": 0.16826209425926208,
+      "learning_rate": 0.0001997116081932793,
+      "loss": 0.7569,
+      "step": 840
+    },
+    {
+      "epoch": 0.02213330324361389,
+      "grad_norm": 0.1876436173915863,
+      "learning_rate": 0.0001996987987648446,
+      "loss": 0.7553,
+      "step": 850
+    },
+    {
+      "epoch": 0.022393695046479936,
+      "grad_norm": 0.17252250015735626,
+      "learning_rate": 0.0001996857114362476,
+      "loss": 0.7644,
+      "step": 860
+    },
+    {
+      "epoch": 0.022654086849345983,
+      "grad_norm": 0.1632252335548401,
+      "learning_rate": 0.00019967234624396793,
+      "loss": 0.7568,
+      "step": 870
+    },
+    {
+      "epoch": 0.02291447865221203,
+      "grad_norm": 0.1818259060382843,
+      "learning_rate": 0.00019965870322525965,
+      "loss": 0.7672,
+      "step": 880
+    },
+    {
+      "epoch": 0.023174870455078073,
+      "grad_norm": 0.15418195724487305,
+      "learning_rate": 0.0001996447824181513,
+      "loss": 0.7642,
+      "step": 890
+    },
+    {
+      "epoch": 0.02343526225794412,
+      "grad_norm": 0.17383505403995514,
+      "learning_rate": 0.0001996305838614457,
+      "loss": 0.7607,
+      "step": 900
+    },
+    {
+      "epoch": 0.023695654060810166,
+      "grad_norm": 0.17794272303581238,
+      "learning_rate": 0.00019961610759471984,
+      "loss": 0.7588,
+      "step": 910
+    },
+    {
+      "epoch": 0.023956045863676213,
+      "grad_norm": 0.1909121572971344,
+      "learning_rate": 0.00019960135365832486,
+      "loss": 0.7438,
+      "step": 920
+    },
+    {
+      "epoch": 0.024216437666542256,
+      "grad_norm": 0.17758873105049133,
+      "learning_rate": 0.00019958632209338587,
+      "loss": 0.7323,
+      "step": 930
+    },
+    {
+      "epoch": 0.024476829469408302,
+      "grad_norm": 0.15553662180900574,
+      "learning_rate": 0.00019957101294180174,
+      "loss": 0.7508,
+      "step": 940
+    },
+    {
+      "epoch": 0.02473722127227435,
+      "grad_norm": 0.15310749411582947,
+      "learning_rate": 0.00019955542624624522,
+      "loss": 0.7563,
+      "step": 950
+    },
+    {
+      "epoch": 0.024997613075140396,
+      "grad_norm": 0.1628728210926056,
+      "learning_rate": 0.00019953956205016256,
+      "loss": 0.7524,
+      "step": 960
+    },
+    {
+      "epoch": 0.02525800487800644,
+      "grad_norm": 0.16211454570293427,
+      "learning_rate": 0.00019952342039777362,
+      "loss": 0.7564,
+      "step": 970
+    },
+    {
+      "epoch": 0.025518396680872486,
+      "grad_norm": 0.15663012862205505,
+      "learning_rate": 0.00019950700133407163,
+      "loss": 0.7395,
+      "step": 980
+    },
+    {
+      "epoch": 0.025778788483738532,
+      "grad_norm": 0.1684863567352295,
+      "learning_rate": 0.00019949030490482296,
+      "loss": 0.753,
+      "step": 990
+    },
+    {
+      "epoch": 0.02603918028660458,
+      "grad_norm": 0.1561436653137207,
+      "learning_rate": 0.0001994733311565673,
+      "loss": 0.7409,
+      "step": 1000
+    },
+    {
+      "epoch": 0.026299572089470622,
+      "grad_norm": 0.1781485229730606,
+      "learning_rate": 0.0001994560801366171,
+      "loss": 0.762,
+      "step": 1010
+    },
+    {
+      "epoch": 0.02655996389233667,
+      "grad_norm": 0.15422071516513824,
+      "learning_rate": 0.00019943855189305792,
+      "loss": 0.7291,
+      "step": 1020
+    },
+    {
+      "epoch": 0.026820355695202715,
+      "grad_norm": 0.17980527877807617,
+      "learning_rate": 0.00019942074647474786,
+      "loss": 0.7732,
+      "step": 1030
+    },
+    {
+      "epoch": 0.027080747498068762,
+      "grad_norm": 0.15810626745224,
+      "learning_rate": 0.00019940266393131775,
+      "loss": 0.7764,
+      "step": 1040
+    },
+    {
+      "epoch": 0.027341139300934805,
+      "grad_norm": 0.16385480761528015,
+      "learning_rate": 0.00019938430431317081,
+      "loss": 0.7404,
+      "step": 1050
+    },
+    {
+      "epoch": 0.027601531103800852,
+      "grad_norm": 0.15134255588054657,
+      "learning_rate": 0.00019936566767148257,
+      "loss": 0.7506,
+      "step": 1060
+    },
+    {
+      "epoch": 0.0278619229066669,
+      "grad_norm": 0.1592187136411667,
+      "learning_rate": 0.00019934675405820077,
+      "loss": 0.73,
+      "step": 1070
+    },
+    {
+      "epoch": 0.028122314709532945,
+      "grad_norm": 0.16852422058582306,
+      "learning_rate": 0.00019932756352604515,
+      "loss": 0.7443,
+      "step": 1080
+    },
+    {
+      "epoch": 0.02838270651239899,
+      "grad_norm": 0.15741507709026337,
+      "learning_rate": 0.00019930809612850735,
+      "loss": 0.7377,
+      "step": 1090
+    },
+    {
+      "epoch": 0.028643098315265035,
+      "grad_norm": 0.22424879670143127,
+      "learning_rate": 0.00019928835191985076,
+      "loss": 0.7544,
+      "step": 1100
+    },
+    {
+      "epoch": 0.028903490118131082,
+      "grad_norm": 0.2047310769557953,
+      "learning_rate": 0.0001992683309551103,
+      "loss": 0.7441,
+      "step": 1110
+    },
+    {
+      "epoch": 0.02916388192099713,
+      "grad_norm": 0.16392463445663452,
+      "learning_rate": 0.00019924803329009243,
+      "loss": 0.7606,
+      "step": 1120
+    },
+    {
+      "epoch": 0.02942427372386317,
+      "grad_norm": 0.16227149963378906,
+      "learning_rate": 0.00019922745898137473,
+      "loss": 0.736,
+      "step": 1130
+    },
+    {
+      "epoch": 0.029684665526729218,
+      "grad_norm": 0.15652808547019958,
+      "learning_rate": 0.00019920660808630598,
+      "loss": 0.7513,
+      "step": 1140
+    },
+    {
+      "epoch": 0.029945057329595265,
+      "grad_norm": 0.15162768959999084,
+      "learning_rate": 0.00019918548066300592,
+      "loss": 0.7303,
+      "step": 1150
+    },
+    {
+      "epoch": 0.03020544913246131,
+      "grad_norm": 0.17650415003299713,
+      "learning_rate": 0.0001991640767703651,
+      "loss": 0.7254,
+      "step": 1160
+    },
+    {
+      "epoch": 0.030465840935327355,
+      "grad_norm": 0.1594468355178833,
+      "learning_rate": 0.00019914239646804462,
+      "loss": 0.741,
+      "step": 1170
+    },
+    {
+      "epoch": 0.0307262327381934,
+      "grad_norm": 0.17928367853164673,
+      "learning_rate": 0.00019912043981647616,
+      "loss": 0.7515,
+      "step": 1180
+    },
+    {
+      "epoch": 0.030986624541059448,
+      "grad_norm": 0.17009998857975006,
+      "learning_rate": 0.00019909820687686157,
+      "loss": 0.7539,
+      "step": 1190
+    },
+    {
+      "epoch": 0.031247016343925495,
+      "grad_norm": 0.16556763648986816,
+      "learning_rate": 0.0001990756977111729,
+      "loss": 0.7418,
+      "step": 1200
+    },
+    {
+      "epoch": 0.03150740814679154,
+      "grad_norm": 0.1561640352010727,
+      "learning_rate": 0.0001990529123821522,
+      "loss": 0.7465,
+      "step": 1210
+    },
+    {
+      "epoch": 0.03176779994965759,
+      "grad_norm": 0.15182287991046906,
+      "learning_rate": 0.00019902985095331113,
+      "loss": 0.7694,
+      "step": 1220
+    },
+    {
+      "epoch": 0.03202819175252363,
+      "grad_norm": 0.15173685550689697,
+      "learning_rate": 0.00019900651348893114,
+      "loss": 0.7519,
+      "step": 1230
+    },
+    {
+      "epoch": 0.032288583555389674,
+      "grad_norm": 0.16535787284374237,
+      "learning_rate": 0.00019898290005406296,
+      "loss": 0.7646,
+      "step": 1240
+    },
+    {
+      "epoch": 0.032548975358255725,
+      "grad_norm": 0.19272534549236298,
+      "learning_rate": 0.00019895901071452667,
+      "loss": 0.7655,
+      "step": 1250
+    },
+    {
+      "epoch": 0.03280936716112177,
+      "grad_norm": 0.1672705113887787,
+      "learning_rate": 0.0001989348455369113,
+      "loss": 0.7486,
+      "step": 1260
+    },
+    {
+      "epoch": 0.03306975896398781,
+      "grad_norm": 0.1525493860244751,
+      "learning_rate": 0.0001989104045885748,
+      "loss": 0.7546,
+      "step": 1270
+    },
+    {
+      "epoch": 0.03333015076685386,
+      "grad_norm": 0.16333037614822388,
+      "learning_rate": 0.00019888568793764385,
+      "loss": 0.7299,
+      "step": 1280
+    },
+    {
+      "epoch": 0.033590542569719904,
+      "grad_norm": 0.1590205729007721,
+      "learning_rate": 0.00019886069565301355,
+      "loss": 0.762,
+      "step": 1290
+    },
+    {
+      "epoch": 0.033850934372585954,
+      "grad_norm": 0.15006420016288757,
+      "learning_rate": 0.00019883542780434733,
+      "loss": 0.7531,
+      "step": 1300
+    },
+    {
+      "epoch": 0.034111326175452,
+      "grad_norm": 0.18390792608261108,
+      "learning_rate": 0.0001988098844620767,
+      "loss": 0.7621,
+      "step": 1310
+    },
+    {
+      "epoch": 0.03437171797831804,
+      "grad_norm": 0.17046166956424713,
+      "learning_rate": 0.0001987840656974011,
+      "loss": 0.7422,
+      "step": 1320
+    },
+    {
+      "epoch": 0.03463210978118409,
+      "grad_norm": 0.15121813118457794,
+      "learning_rate": 0.00019875797158228775,
+      "loss": 0.7555,
+      "step": 1330
+    },
+    {
+      "epoch": 0.034892501584050134,
+      "grad_norm": 0.16219307482242584,
+      "learning_rate": 0.00019873160218947125,
+      "loss": 0.7301,
+      "step": 1340
+    },
+    {
+      "epoch": 0.03515289338691618,
+      "grad_norm": 0.1779986321926117,
+      "learning_rate": 0.00019870495759245362,
+      "loss": 0.7356,
+      "step": 1350
+    },
+    {
+      "epoch": 0.03541328518978223,
+      "grad_norm": 0.16951359808444977,
+      "learning_rate": 0.0001986780378655039,
+      "loss": 0.7645,
+      "step": 1360
+    },
+    {
+      "epoch": 0.03567367699264827,
+      "grad_norm": 0.16620802879333496,
+      "learning_rate": 0.0001986508430836581,
+      "loss": 0.7331,
+      "step": 1370
+    },
+    {
+      "epoch": 0.03593406879551432,
+      "grad_norm": 0.1577858328819275,
+      "learning_rate": 0.0001986233733227188,
+      "loss": 0.7667,
+      "step": 1380
+    },
+    {
+      "epoch": 0.036194460598380364,
+      "grad_norm": 0.1637091338634491,
+      "learning_rate": 0.00019859562865925525,
+      "loss": 0.7521,
+      "step": 1390
+    },
+    {
+      "epoch": 0.03645485240124641,
+      "grad_norm": 0.15061691403388977,
+      "learning_rate": 0.00019856760917060277,
+      "loss": 0.744,
+      "step": 1400
+    },
+    {
+      "epoch": 0.03671524420411246,
+      "grad_norm": 0.15373477339744568,
+      "learning_rate": 0.00019853931493486287,
+      "loss": 0.7677,
+      "step": 1410
+    },
+    {
+      "epoch": 0.0369756360069785,
+      "grad_norm": 0.16468606889247894,
+      "learning_rate": 0.00019851074603090277,
+      "loss": 0.7179,
+      "step": 1420
+    },
+    {
+      "epoch": 0.037236027809844544,
+      "grad_norm": 0.16084876656532288,
+      "learning_rate": 0.00019848190253835536,
+      "loss": 0.749,
+      "step": 1430
+    },
+    {
+      "epoch": 0.037496419612710594,
+      "grad_norm": 0.16743004322052002,
+      "learning_rate": 0.00019845278453761896,
+      "loss": 0.7483,
+      "step": 1440
+    },
+    {
+      "epoch": 0.03775681141557664,
+      "grad_norm": 0.17335088551044464,
+      "learning_rate": 0.00019842339210985696,
+      "loss": 0.735,
+      "step": 1450
+    },
+    {
+      "epoch": 0.03801720321844268,
+      "grad_norm": 0.1546197235584259,
+      "learning_rate": 0.00019839372533699774,
+      "loss": 0.7549,
+      "step": 1460
+    },
+    {
+      "epoch": 0.03827759502130873,
+      "grad_norm": 0.16218656301498413,
+      "learning_rate": 0.00019836378430173438,
+      "loss": 0.7425,
+      "step": 1470
+    },
+    {
+      "epoch": 0.03853798682417477,
+      "grad_norm": 0.1712743639945984,
+      "learning_rate": 0.0001983335690875245,
+      "loss": 0.733,
+      "step": 1480
+    },
+    {
+      "epoch": 0.03879837862704082,
+      "grad_norm": 0.15490613877773285,
+      "learning_rate": 0.00019830307977858984,
+      "loss": 0.7265,
+      "step": 1490
+    },
+    {
+      "epoch": 0.03905877042990687,
+      "grad_norm": 0.1646670252084732,
+      "learning_rate": 0.00019827231645991623,
+      "loss": 0.7315,
+      "step": 1500
+    },
+    {
+      "epoch": 0.03931916223277291,
+      "grad_norm": 0.1599082201719284,
+      "learning_rate": 0.00019824127921725326,
+      "loss": 0.7293,
+      "step": 1510
+    },
+    {
+      "epoch": 0.03957955403563896,
+      "grad_norm": 0.1565747708082199,
+      "learning_rate": 0.00019820996813711407,
+      "loss": 0.7396,
+      "step": 1520
+    },
+    {
+      "epoch": 0.039839945838505,
+      "grad_norm": 0.154826357960701,
+      "learning_rate": 0.0001981783833067751,
+      "loss": 0.7217,
+      "step": 1530
+    },
+    {
+      "epoch": 0.040100337641371046,
+      "grad_norm": 0.16705222427845,
+      "learning_rate": 0.0001981465248142758,
+      "loss": 0.761,
+      "step": 1540
+    },
+    {
+      "epoch": 0.040360729444237096,
+      "grad_norm": 0.15651623904705048,
+      "learning_rate": 0.00019811439274841842,
+      "loss": 0.7565,
+      "step": 1550
+    },
+    {
+      "epoch": 0.04062112124710314,
+      "grad_norm": 0.16211090981960297,
+      "learning_rate": 0.00019808198719876782,
+      "loss": 0.7555,
+      "step": 1560
+    },
+    {
+      "epoch": 0.04088151304996919,
+      "grad_norm": 0.16856881976127625,
+      "learning_rate": 0.00019804930825565112,
+      "loss": 0.7567,
+      "step": 1570
+    },
+    {
+      "epoch": 0.04114190485283523,
+      "grad_norm": 0.1588718593120575,
+      "learning_rate": 0.00019801635601015752,
+      "loss": 0.729,
+      "step": 1580
+    },
+    {
+      "epoch": 0.041402296655701276,
+      "grad_norm": 0.17078711092472076,
+      "learning_rate": 0.00019798313055413808,
+      "loss": 0.7418,
+      "step": 1590
+    },
+    {
+      "epoch": 0.041662688458567326,
+      "grad_norm": 0.16652734577655792,
+      "learning_rate": 0.00019794963198020525,
+      "loss": 0.7341,
+      "step": 1600
+    },
+    {
+      "epoch": 0.04192308026143337,
+      "grad_norm": 0.15535488724708557,
+      "learning_rate": 0.00019791586038173296,
+      "loss": 0.7396,
+      "step": 1610
+    },
+    {
+      "epoch": 0.04218347206429941,
+      "grad_norm": 0.3506317734718323,
+      "learning_rate": 0.00019788181585285602,
+      "loss": 0.7345,
+      "step": 1620
+    },
+    {
+      "epoch": 0.04244386386716546,
+      "grad_norm": 0.16875872015953064,
+      "learning_rate": 0.00019784749848847003,
+      "loss": 0.7214,
+      "step": 1630
+    },
+    {
+      "epoch": 0.042704255670031506,
+      "grad_norm": 0.17675861716270447,
+      "learning_rate": 0.0001978129083842312,
+      "loss": 0.7431,
+      "step": 1640
+    },
+    {
+      "epoch": 0.042964647472897556,
+      "grad_norm": 0.15601837635040283,
+      "learning_rate": 0.00019777804563655583,
+      "loss": 0.7215,
+      "step": 1650
+    },
+    {
+      "epoch": 0.0432250392757636,
+      "grad_norm": 0.1874823123216629,
+      "learning_rate": 0.00019774291034262026,
+      "loss": 0.727,
+      "step": 1660
+    },
+    {
+      "epoch": 0.04348543107862964,
+      "grad_norm": 0.17005637288093567,
+      "learning_rate": 0.00019770750260036054,
+      "loss": 0.7446,
+      "step": 1670
+    },
+    {
+      "epoch": 0.04374582288149569,
+      "grad_norm": 0.17069579660892487,
+      "learning_rate": 0.00019767182250847207,
+      "loss": 0.7266,
+      "step": 1680
+    },
+    {
+      "epoch": 0.044006214684361736,
+      "grad_norm": 0.16133156418800354,
+      "learning_rate": 0.00019763587016640948,
+      "loss": 0.7568,
+      "step": 1690
+    },
+    {
+      "epoch": 0.04426660648722778,
+      "grad_norm": 0.16229428350925446,
+      "learning_rate": 0.00019759964567438623,
+      "loss": 0.7402,
+      "step": 1700
+    },
+    {
+      "epoch": 0.04452699829009383,
+      "grad_norm": 0.1622512936592102,
+      "learning_rate": 0.00019756314913337432,
+      "loss": 0.7536,
+      "step": 1710
+    },
+    {
+      "epoch": 0.04478739009295987,
+      "grad_norm": 0.2161218672990799,
+      "learning_rate": 0.00019752638064510415,
+      "loss": 0.723,
+      "step": 1720
+    },
+    {
+      "epoch": 0.04504778189582592,
+      "grad_norm": 0.154169961810112,
+      "learning_rate": 0.00019748934031206414,
+      "loss": 0.7441,
+      "step": 1730
+    },
+    {
+      "epoch": 0.045308173698691966,
+      "grad_norm": 0.15468057990074158,
+      "learning_rate": 0.00019745202823750034,
+      "loss": 0.7349,
+      "step": 1740
+    },
+    {
+      "epoch": 0.04556856550155801,
+      "grad_norm": 0.2015281468629837,
+      "learning_rate": 0.0001974144445254164,
+      "loss": 0.726,
+      "step": 1750
+    },
+    {
+      "epoch": 0.04582895730442406,
+      "grad_norm": 0.1931644082069397,
+      "learning_rate": 0.00019737658928057302,
+      "loss": 0.7604,
+      "step": 1760
+    },
+    {
+      "epoch": 0.0460893491072901,
+      "grad_norm": 0.1528482288122177,
+      "learning_rate": 0.00019733846260848776,
+      "loss": 0.7408,
+      "step": 1770
+    },
+    {
+      "epoch": 0.046349740910156145,
+      "grad_norm": 0.16370061039924622,
+      "learning_rate": 0.0001973000646154349,
+      "loss": 0.7647,
+      "step": 1780
+    },
+    {
+      "epoch": 0.046610132713022195,
+      "grad_norm": 0.16271348297595978,
+      "learning_rate": 0.00019726139540844484,
+      "loss": 0.7212,
+      "step": 1790
+    },
+    {
+      "epoch": 0.04687052451588824,
+      "grad_norm": 0.16218173503875732,
+      "learning_rate": 0.00019722245509530401,
+      "loss": 0.735,
+      "step": 1800
+    },
+    {
+      "epoch": 0.04713091631875429,
+      "grad_norm": 0.17063820362091064,
+      "learning_rate": 0.00019718324378455458,
+      "loss": 0.7311,
+      "step": 1810
+    },
+    {
+      "epoch": 0.04739130812162033,
+      "grad_norm": 0.1678459346294403,
+      "learning_rate": 0.00019714376158549404,
+      "loss": 0.7486,
+      "step": 1820
+    },
+    {
+      "epoch": 0.047651699924486375,
+      "grad_norm": 0.15926459431648254,
+      "learning_rate": 0.00019710400860817494,
+      "loss": 0.743,
+      "step": 1830
+    },
+    {
+      "epoch": 0.047912091727352425,
+      "grad_norm": 0.1775251179933548,
+      "learning_rate": 0.00019706398496340463,
+      "loss": 0.7512,
+      "step": 1840
+    },
+    {
+      "epoch": 0.04817248353021847,
+      "grad_norm": 0.1572408229112625,
+      "learning_rate": 0.00019702369076274494,
+      "loss": 0.733,
+      "step": 1850
+    },
+    {
+      "epoch": 0.04843287533308451,
+      "grad_norm": 0.29658186435699463,
+      "learning_rate": 0.0001969831261185118,
+      "loss": 0.7297,
+      "step": 1860
+    },
+    {
+      "epoch": 0.04869326713595056,
+      "grad_norm": 0.16520118713378906,
+      "learning_rate": 0.00019694229114377494,
+      "loss": 0.721,
+      "step": 1870
+    },
+    {
+      "epoch": 0.048953658938816605,
+      "grad_norm": 0.17762574553489685,
+      "learning_rate": 0.00019690118595235774,
+      "loss": 0.7304,
+      "step": 1880
+    },
+    {
+      "epoch": 0.049214050741682655,
+      "grad_norm": 0.16636615991592407,
+      "learning_rate": 0.00019685981065883663,
+      "loss": 0.7257,
+      "step": 1890
+    },
+    {
+      "epoch": 0.0494744425445487,
+      "grad_norm": 0.1622323989868164,
+      "learning_rate": 0.00019681816537854102,
+      "loss": 0.7353,
+      "step": 1900
+    },
+    {
+      "epoch": 0.04973483434741474,
+      "grad_norm": 0.17419832944869995,
+      "learning_rate": 0.00019677625022755289,
+      "loss": 0.7452,
+      "step": 1910
+    },
+    {
+      "epoch": 0.04999522615028079,
+      "grad_norm": 0.17460434138774872,
+      "learning_rate": 0.00019673406532270634,
+      "loss": 0.7391,
+      "step": 1920
+    },
+    {
+      "epoch": 0.050255617953146835,
+      "grad_norm": 0.15844550728797913,
+      "learning_rate": 0.00019669161078158753,
+      "loss": 0.7327,
+      "step": 1930
+    },
+    {
+      "epoch": 0.05051600975601288,
+      "grad_norm": 0.1638839989900589,
+      "learning_rate": 0.0001966488867225341,
+      "loss": 0.745,
+      "step": 1940
+    },
+    {
+      "epoch": 0.05077640155887893,
+      "grad_norm": 0.1587786227464676,
+      "learning_rate": 0.00019660589326463498,
+      "loss": 0.7476,
+      "step": 1950
+    },
+    {
+      "epoch": 0.05103679336174497,
+      "grad_norm": 0.15708380937576294,
+      "learning_rate": 0.00019656263052773002,
+      "loss": 0.7208,
+      "step": 1960
+    },
+    {
+      "epoch": 0.05129718516461102,
+      "grad_norm": 0.15816234052181244,
+      "learning_rate": 0.00019651909863240965,
+      "loss": 0.7262,
+      "step": 1970
+    },
+    {
+      "epoch": 0.051557576967477065,
+      "grad_norm": 0.16749270260334015,
+      "learning_rate": 0.00019647529770001456,
+      "loss": 0.7284,
+      "step": 1980
+    },
+    {
+      "epoch": 0.05181796877034311,
+      "grad_norm": 0.16943767666816711,
+      "learning_rate": 0.00019643122785263536,
+      "loss": 0.7225,
+      "step": 1990
+    },
+    {
+      "epoch": 0.05207836057320916,
+      "grad_norm": 0.42929205298423767,
+      "learning_rate": 0.00019638688921311224,
+      "loss": 0.7305,
+      "step": 2000
+    },
+    {
+      "epoch": 0.0523387523760752,
+      "grad_norm": 0.15851692855358124,
+      "learning_rate": 0.00019634228190503454,
+      "loss": 0.7344,
+      "step": 2010
+    },
+    {
+      "epoch": 0.052599144178941244,
+      "grad_norm": 0.16053883731365204,
+      "learning_rate": 0.00019629740605274062,
+      "loss": 0.7468,
+      "step": 2020
+    },
+    {
+      "epoch": 0.052859535981807294,
+      "grad_norm": 0.16504009068012238,
+      "learning_rate": 0.00019625226178131728,
+      "loss": 0.7375,
+      "step": 2030
+    },
+    {
+      "epoch": 0.05311992778467334,
+      "grad_norm": 0.1618044674396515,
+      "learning_rate": 0.00019620684921659953,
+      "loss": 0.7201,
+      "step": 2040
+    },
+    {
+      "epoch": 0.05338031958753939,
+      "grad_norm": 0.15512776374816895,
+      "learning_rate": 0.00019616116848517027,
+      "loss": 0.7355,
+      "step": 2050
+    },
+    {
+      "epoch": 0.05364071139040543,
+      "grad_norm": 0.17377036809921265,
+      "learning_rate": 0.00019611521971435979,
+      "loss": 0.7226,
+      "step": 2060
+    },
+    {
+      "epoch": 0.053901103193271474,
+      "grad_norm": 0.1685250997543335,
+      "learning_rate": 0.0001960690030322456,
+      "loss": 0.7483,
+      "step": 2070
+    },
+    {
+      "epoch": 0.054161494996137524,
+      "grad_norm": 0.18394522368907928,
+      "learning_rate": 0.00019602251856765194,
+      "loss": 0.7385,
+      "step": 2080
+    },
+    {
+      "epoch": 0.05442188679900357,
+      "grad_norm": 0.1753673106431961,
+      "learning_rate": 0.0001959757664501495,
+      "loss": 0.7378,
+      "step": 2090
+    },
+    {
+      "epoch": 0.05468227860186961,
+      "grad_norm": 0.1795465052127838,
+      "learning_rate": 0.000195928746810055,
+      "loss": 0.748,
+      "step": 2100
+    },
+    {
+      "epoch": 0.05494267040473566,
+      "grad_norm": 0.16327305138111115,
+      "learning_rate": 0.0001958814597784309,
+      "loss": 0.7306,
+      "step": 2110
+    },
+    {
+      "epoch": 0.055203062207601704,
+      "grad_norm": 0.15880291163921356,
+      "learning_rate": 0.00019583390548708486,
+      "loss": 0.7281,
+      "step": 2120
+    },
+    {
+      "epoch": 0.05546345401046775,
+      "grad_norm": 0.1702323853969574,
+      "learning_rate": 0.0001957860840685696,
+      "loss": 0.7407,
+      "step": 2130
+    },
+    {
+      "epoch": 0.0557238458133338,
+      "grad_norm": 0.16931670904159546,
+      "learning_rate": 0.0001957379956561825,
+      "loss": 0.7272,
+      "step": 2140
+    },
+    {
+      "epoch": 0.05598423761619984,
+      "grad_norm": 0.15455976128578186,
+      "learning_rate": 0.000195689640383965,
+      "loss": 0.7398,
+      "step": 2150
+    },
+    {
+      "epoch": 0.05624462941906589,
+      "grad_norm": 0.16061417758464813,
+      "learning_rate": 0.0001956410183867024,
+      "loss": 0.749,
+      "step": 2160
+    },
+    {
+      "epoch": 0.056505021221931934,
+      "grad_norm": 0.14933143556118011,
+      "learning_rate": 0.00019559212979992365,
+      "loss": 0.7418,
+      "step": 2170
+    },
+    {
+      "epoch": 0.05676541302479798,
+      "grad_norm": 0.1592816412448883,
+      "learning_rate": 0.00019554297475990058,
+      "loss": 0.7423,
+      "step": 2180
+    },
+    {
+      "epoch": 0.05702580482766403,
+      "grad_norm": 0.1677238792181015,
+      "learning_rate": 0.00019549355340364787,
+      "loss": 0.7101,
+      "step": 2190
+    },
+    {
+      "epoch": 0.05728619663053007,
+      "grad_norm": 0.3558599054813385,
+      "learning_rate": 0.00019544386586892238,
+      "loss": 0.725,
+      "step": 2200
+    },
+    {
+      "epoch": 0.05754658843339611,
+      "grad_norm": 0.1746376007795334,
+      "learning_rate": 0.00019539391229422313,
+      "loss": 0.7479,
+      "step": 2210
+    },
+    {
+      "epoch": 0.057806980236262163,
+      "grad_norm": 0.15979182720184326,
+      "learning_rate": 0.00019534369281879049,
+      "loss": 0.7352,
+      "step": 2220
+    },
+    {
+      "epoch": 0.05806737203912821,
+      "grad_norm": 0.16173166036605835,
+      "learning_rate": 0.0001952932075826061,
+      "loss": 0.7364,
+      "step": 2230
+    },
+    {
+      "epoch": 0.05832776384199426,
+      "grad_norm": 0.1514744907617569,
+      "learning_rate": 0.00019524245672639245,
+      "loss": 0.734,
+      "step": 2240
+    },
+    {
+      "epoch": 0.0585881556448603,
+      "grad_norm": 0.16860373318195343,
+      "learning_rate": 0.00019519144039161222,
+      "loss": 0.7098,
+      "step": 2250
+    },
+    {
+      "epoch": 0.05884854744772634,
+      "grad_norm": 0.16847743093967438,
+      "learning_rate": 0.00019514015872046833,
+      "loss": 0.7103,
+      "step": 2260
+    },
+    {
+      "epoch": 0.05910893925059239,
+      "grad_norm": 0.16181516647338867,
+      "learning_rate": 0.00019508861185590307,
+      "loss": 0.7561,
+      "step": 2270
+    },
+    {
+      "epoch": 0.059369331053458436,
+      "grad_norm": 0.16594484448432922,
+      "learning_rate": 0.0001950367999415981,
+      "loss": 0.7308,
+      "step": 2280
+    },
+    {
+      "epoch": 0.05962972285632448,
+      "grad_norm": 0.166441410779953,
+      "learning_rate": 0.00019498472312197375,
+      "loss": 0.735,
+      "step": 2290
+    },
+    {
+      "epoch": 0.05989011465919053,
+      "grad_norm": 0.16273920238018036,
+      "learning_rate": 0.00019493238154218886,
+      "loss": 0.7458,
+      "step": 2300
+    },
+    {
+      "epoch": 0.06015050646205657,
+      "grad_norm": 0.16227276623249054,
+      "learning_rate": 0.00019487977534814012,
+      "loss": 0.7143,
+      "step": 2310
+    },
+    {
+      "epoch": 0.06041089826492262,
+      "grad_norm": 0.1619606912136078,
+      "learning_rate": 0.000194826904686462,
+      "loss": 0.7285,
+      "step": 2320
+    },
+    {
+      "epoch": 0.060671290067788666,
+      "grad_norm": 0.1596045345067978,
+      "learning_rate": 0.00019477376970452603,
+      "loss": 0.7513,
+      "step": 2330
+    },
+    {
+      "epoch": 0.06093168187065471,
+      "grad_norm": 0.17504757642745972,
+      "learning_rate": 0.00019472037055044044,
+      "loss": 0.7376,
+      "step": 2340
+    },
+    {
+      "epoch": 0.06119207367352076,
+      "grad_norm": 0.1559167355298996,
+      "learning_rate": 0.00019466670737304992,
+      "loss": 0.7339,
+      "step": 2350
+    },
+    {
+      "epoch": 0.0614524654763868,
+      "grad_norm": 0.1624836027622223,
+      "learning_rate": 0.0001946127803219351,
+      "loss": 0.7258,
+      "step": 2360
+    },
+    {
+      "epoch": 0.061712857279252846,
+      "grad_norm": 0.17907138168811798,
+      "learning_rate": 0.00019455858954741206,
+      "loss": 0.72,
+      "step": 2370
+    },
+    {
+      "epoch": 0.061973249082118896,
+      "grad_norm": 0.15922705829143524,
+      "learning_rate": 0.00019450413520053202,
+      "loss": 0.7187,
+      "step": 2380
+    },
+    {
+      "epoch": 0.06223364088498494,
+      "grad_norm": 0.1552513986825943,
+      "learning_rate": 0.0001944494174330809,
+      "loss": 0.7183,
+      "step": 2390
+    },
+    {
+      "epoch": 0.06249403268785099,
+      "grad_norm": 0.16838514804840088,
+      "learning_rate": 0.00019439443639757885,
+      "loss": 0.7286,
+      "step": 2400
+    },
+    {
+      "epoch": 0.06275442449071703,
+      "grad_norm": 0.17352423071861267,
+      "learning_rate": 0.00019433919224727986,
+      "loss": 0.7436,
+      "step": 2410
+    },
+    {
+      "epoch": 0.06301481629358308,
+      "grad_norm": 0.17366603016853333,
+      "learning_rate": 0.0001942836851361713,
+      "loss": 0.7265,
+      "step": 2420
+    },
+    {
+      "epoch": 0.06327520809644913,
+      "grad_norm": 0.14833413064479828,
+      "learning_rate": 0.00019422791521897357,
+      "loss": 0.7234,
+      "step": 2430
+    },
+    {
+      "epoch": 0.06353559989931518,
+      "grad_norm": 0.16602723300457,
+      "learning_rate": 0.00019417188265113958,
+      "loss": 0.725,
+      "step": 2440
+    },
+    {
+      "epoch": 0.06379599170218121,
+      "grad_norm": 0.17290353775024414,
+      "learning_rate": 0.00019411558758885438,
+      "loss": 0.7174,
+      "step": 2450
+    },
+    {
+      "epoch": 0.06405638350504726,
+      "grad_norm": 0.16486665606498718,
+      "learning_rate": 0.0001940590301890346,
+      "loss": 0.7301,
+      "step": 2460
+    },
+    {
+      "epoch": 0.06431677530791331,
+      "grad_norm": 0.16255232691764832,
+      "learning_rate": 0.00019400221060932827,
+      "loss": 0.7462,
+      "step": 2470
+    },
+    {
+      "epoch": 0.06457716711077935,
+      "grad_norm": 0.16139757633209229,
+      "learning_rate": 0.0001939451290081141,
+      "loss": 0.7424,
+      "step": 2480
+    },
+    {
+      "epoch": 0.0648375589136454,
+      "grad_norm": 0.165597602725029,
+      "learning_rate": 0.00019388778554450117,
+      "loss": 0.7426,
+      "step": 2490
+    },
+    {
+      "epoch": 0.06509795071651145,
+      "grad_norm": 0.19819000363349915,
+      "learning_rate": 0.00019383018037832854,
+      "loss": 0.7356,
+      "step": 2500
+    },
+    {
+      "epoch": 0.06535834251937749,
+      "grad_norm": 0.16469696164131165,
+      "learning_rate": 0.00019377231367016467,
+      "loss": 0.718,
+      "step": 2510
+    },
+    {
+      "epoch": 0.06561873432224354,
+      "grad_norm": 0.1644965261220932,
+      "learning_rate": 0.00019371418558130702,
+      "loss": 0.7253,
+      "step": 2520
+    },
+    {
+      "epoch": 0.06587912612510959,
+      "grad_norm": 0.15347526967525482,
+      "learning_rate": 0.00019365579627378174,
+      "loss": 0.7214,
+      "step": 2530
+    },
+    {
+      "epoch": 0.06613951792797562,
+      "grad_norm": 0.1618672013282776,
+      "learning_rate": 0.00019359714591034302,
+      "loss": 0.7204,
+      "step": 2540
+    },
+    {
+      "epoch": 0.06639990973084167,
+      "grad_norm": 0.17043665051460266,
+      "learning_rate": 0.00019353823465447268,
+      "loss": 0.7278,
+      "step": 2550
+    },
+    {
+      "epoch": 0.06666030153370772,
+      "grad_norm": 0.15762579441070557,
+      "learning_rate": 0.00019347906267037983,
+      "loss": 0.7283,
+      "step": 2560
+    },
+    {
+      "epoch": 0.06692069333657376,
+      "grad_norm": 0.1622801572084427,
+      "learning_rate": 0.00019341963012300029,
+      "loss": 0.7193,
+      "step": 2570
+    },
+    {
+      "epoch": 0.06718108513943981,
+      "grad_norm": 0.16705769300460815,
+      "learning_rate": 0.00019335993717799617,
+      "loss": 0.7414,
+      "step": 2580
+    },
+    {
+      "epoch": 0.06744147694230586,
+      "grad_norm": 0.15886452794075012,
+      "learning_rate": 0.00019329998400175545,
+      "loss": 0.7242,
+      "step": 2590
+    },
+    {
+      "epoch": 0.06770186874517191,
+      "grad_norm": 0.17994090914726257,
+      "learning_rate": 0.00019323977076139142,
+      "loss": 0.7017,
+      "step": 2600
+    },
+    {
+      "epoch": 0.06796226054803794,
+      "grad_norm": 0.1609068214893341,
+      "learning_rate": 0.00019317929762474232,
+      "loss": 0.7352,
+      "step": 2610
+    },
+    {
+      "epoch": 0.068222652350904,
+      "grad_norm": 0.15605950355529785,
+      "learning_rate": 0.0001931185647603708,
+      "loss": 0.7249,
+      "step": 2620
+    },
+    {
+      "epoch": 0.06848304415377005,
+      "grad_norm": 0.16057750582695007,
+      "learning_rate": 0.00019305757233756352,
+      "loss": 0.7521,
+      "step": 2630
+    },
+    {
+      "epoch": 0.06874343595663608,
+      "grad_norm": 0.1703862100839615,
+      "learning_rate": 0.00019299632052633054,
+      "loss": 0.7245,
+      "step": 2640
+    },
+    {
+      "epoch": 0.06900382775950213,
+      "grad_norm": 0.16324444115161896,
+      "learning_rate": 0.00019293480949740505,
+      "loss": 0.7395,
+      "step": 2650
+    },
+    {
+      "epoch": 0.06926421956236818,
+      "grad_norm": 0.15283791720867157,
+      "learning_rate": 0.00019287303942224266,
+      "loss": 0.7158,
+      "step": 2660
+    },
+    {
+      "epoch": 0.06952461136523422,
+      "grad_norm": 0.1882282942533493,
+      "learning_rate": 0.00019281101047302114,
+      "loss": 0.724,
+      "step": 2670
+    },
+    {
+      "epoch": 0.06978500316810027,
+      "grad_norm": 0.16147953271865845,
+      "learning_rate": 0.00019274872282263984,
+      "loss": 0.7365,
+      "step": 2680
+    },
+    {
+      "epoch": 0.07004539497096632,
+      "grad_norm": 0.1614103466272354,
+      "learning_rate": 0.00019268617664471916,
+      "loss": 0.7206,
+      "step": 2690
+    },
+    {
+      "epoch": 0.07030578677383235,
+      "grad_norm": 0.16784432530403137,
+      "learning_rate": 0.00019262337211360016,
+      "loss": 0.7279,
+      "step": 2700
+    },
+    {
+      "epoch": 0.0705661785766984,
+      "grad_norm": 0.15966112911701202,
+      "learning_rate": 0.000192560309404344,
+      "loss": 0.7274,
+      "step": 2710
+    },
+    {
+      "epoch": 0.07082657037956445,
+      "grad_norm": 0.16970521211624146,
+      "learning_rate": 0.0001924969886927315,
+      "loss": 0.7038,
+      "step": 2720
+    },
+    {
+      "epoch": 0.07108696218243049,
+      "grad_norm": 0.16143856942653656,
+      "learning_rate": 0.00019243341015526272,
+      "loss": 0.7097,
+      "step": 2730
+    },
+    {
+      "epoch": 0.07134735398529654,
+      "grad_norm": 0.16041269898414612,
+      "learning_rate": 0.00019236957396915623,
+      "loss": 0.722,
+      "step": 2740
+    },
+    {
+      "epoch": 0.07160774578816259,
+      "grad_norm": 0.15845969319343567,
+      "learning_rate": 0.00019230548031234882,
+      "loss": 0.7238,
+      "step": 2750
+    },
+    {
+      "epoch": 0.07186813759102864,
+      "grad_norm": 0.14966030418872833,
+      "learning_rate": 0.00019224112936349502,
+      "loss": 0.7182,
+      "step": 2760
+    },
+    {
+      "epoch": 0.07212852939389468,
+      "grad_norm": 0.16525116562843323,
+      "learning_rate": 0.00019217652130196653,
+      "loss": 0.7397,
+      "step": 2770
+    },
+    {
+      "epoch": 0.07238892119676073,
+      "grad_norm": 0.18119119107723236,
+      "learning_rate": 0.0001921116563078516,
+      "loss": 0.7222,
+      "step": 2780
+    },
+    {
+      "epoch": 0.07264931299962678,
+      "grad_norm": 0.1709197610616684,
+      "learning_rate": 0.00019204653456195478,
+      "loss": 0.7068,
+      "step": 2790
+    },
+    {
+      "epoch": 0.07290970480249281,
+      "grad_norm": 0.16309161484241486,
+      "learning_rate": 0.00019198115624579625,
+      "loss": 0.7349,
+      "step": 2800
+    },
+    {
+      "epoch": 0.07317009660535886,
+      "grad_norm": 0.1736750453710556,
+      "learning_rate": 0.00019191552154161135,
+      "loss": 0.7445,
+      "step": 2810
+    },
+    {
+      "epoch": 0.07343048840822491,
+      "grad_norm": 0.15009112656116486,
+      "learning_rate": 0.00019184963063235006,
+      "loss": 0.7034,
+      "step": 2820
+    },
+    {
+      "epoch": 0.07369088021109095,
+      "grad_norm": 0.17244628071784973,
+      "learning_rate": 0.0001917834837016766,
+      "loss": 0.7285,
+      "step": 2830
+    },
+    {
+      "epoch": 0.073951272013957,
+      "grad_norm": 0.15991820394992828,
+      "learning_rate": 0.00019171708093396861,
+      "loss": 0.7096,
+      "step": 2840
+    },
+    {
+      "epoch": 0.07421166381682305,
+      "grad_norm": 0.17037667334079742,
+      "learning_rate": 0.0001916504225143171,
+      "loss": 0.7177,
+      "step": 2850
+    },
+    {
+      "epoch": 0.07447205561968909,
+      "grad_norm": 0.16700348258018494,
+      "learning_rate": 0.00019158350862852553,
+      "loss": 0.7453,
+      "step": 2860
+    },
+    {
+      "epoch": 0.07473244742255514,
+      "grad_norm": 0.17683659493923187,
+      "learning_rate": 0.00019151633946310948,
+      "loss": 0.7331,
+      "step": 2870
+    },
+    {
+      "epoch": 0.07499283922542119,
+      "grad_norm": 0.16364306211471558,
+      "learning_rate": 0.00019144891520529608,
+      "loss": 0.7347,
+      "step": 2880
+    },
+    {
+      "epoch": 0.07525323102828722,
+      "grad_norm": 0.1781424731016159,
+      "learning_rate": 0.00019138123604302355,
+      "loss": 0.7169,
+      "step": 2890
+    },
+    {
+      "epoch": 0.07551362283115327,
+      "grad_norm": 0.16007259488105774,
+      "learning_rate": 0.00019131330216494064,
+      "loss": 0.7269,
+      "step": 2900
+    },
+    {
+      "epoch": 0.07577401463401932,
+      "grad_norm": 0.1604921519756317,
+      "learning_rate": 0.00019124511376040598,
+      "loss": 0.7094,
+      "step": 2910
+    },
+    {
+      "epoch": 0.07603440643688536,
+      "grad_norm": 0.16649965941905975,
+      "learning_rate": 0.00019117667101948782,
+      "loss": 0.7271,
+      "step": 2920
+    },
+    {
+      "epoch": 0.07629479823975141,
+      "grad_norm": 0.16084066033363342,
+      "learning_rate": 0.0001911079741329632,
+      "loss": 0.7239,
+      "step": 2930
+    },
+    {
+      "epoch": 0.07655519004261746,
+      "grad_norm": 0.1651066243648529,
+      "learning_rate": 0.0001910390232923177,
+      "loss": 0.7304,
+      "step": 2940
+    },
+    {
+      "epoch": 0.07681558184548351,
+      "grad_norm": 0.1528957635164261,
+      "learning_rate": 0.00019096981868974467,
+      "loss": 0.7068,
+      "step": 2950
+    },
+    {
+      "epoch": 0.07707597364834955,
+      "grad_norm": 0.172830730676651,
+      "learning_rate": 0.00019090036051814483,
+      "loss": 0.7277,
+      "step": 2960
+    },
+    {
+      "epoch": 0.0773363654512156,
+      "grad_norm": 0.15909147262573242,
+      "learning_rate": 0.00019083064897112571,
+      "loss": 0.7135,
+      "step": 2970
+    },
+    {
+      "epoch": 0.07759675725408165,
+      "grad_norm": 0.16273066401481628,
+      "learning_rate": 0.0001907606842430011,
+      "loss": 0.7346,
+      "step": 2980
+    },
+    {
+      "epoch": 0.07785714905694768,
+      "grad_norm": 0.1595291793346405,
+      "learning_rate": 0.00019069046652879049,
+      "loss": 0.7377,
+      "step": 2990
+    },
+    {
+      "epoch": 0.07811754085981373,
+      "grad_norm": 0.15573470294475555,
+      "learning_rate": 0.0001906199960242185,
+      "loss": 0.7026,
+      "step": 3000
+    },
+    {
+      "epoch": 0.07837793266267978,
+      "grad_norm": 0.1670667678117752,
+      "learning_rate": 0.0001905492729257145,
+      "loss": 0.7231,
+      "step": 3010
+    },
+    {
+      "epoch": 0.07863832446554582,
+      "grad_norm": 0.17074571549892426,
+      "learning_rate": 0.00019047829743041184,
+      "loss": 0.7003,
+      "step": 3020
+    },
+    {
+      "epoch": 0.07889871626841187,
+      "grad_norm": 0.16979442536830902,
+      "learning_rate": 0.00019040706973614738,
+      "loss": 0.7217,
+      "step": 3030
+    },
+    {
+      "epoch": 0.07915910807127792,
+      "grad_norm": 0.15843816101551056,
+      "learning_rate": 0.00019033559004146103,
+      "loss": 0.7334,
+      "step": 3040
+    },
+    {
+      "epoch": 0.07941949987414396,
+      "grad_norm": 0.1607016921043396,
+      "learning_rate": 0.0001902638585455951,
+      "loss": 0.7271,
+      "step": 3050
+    },
+    {
+      "epoch": 0.07967989167701,
+      "grad_norm": 0.1619115173816681,
+      "learning_rate": 0.0001901918754484938,
+      "loss": 0.7144,
+      "step": 3060
+    },
+    {
+      "epoch": 0.07994028347987606,
+      "grad_norm": 0.1638360172510147,
+      "learning_rate": 0.00019011964095080254,
+      "loss": 0.7149,
+      "step": 3070
+    },
+    {
+      "epoch": 0.08020067528274209,
+      "grad_norm": 0.16503652930259705,
+      "learning_rate": 0.00019004715525386764,
+      "loss": 0.7011,
+      "step": 3080
+    },
+    {
+      "epoch": 0.08046106708560814,
+      "grad_norm": 0.16763822734355927,
+      "learning_rate": 0.00018997441855973552,
+      "loss": 0.7145,
+      "step": 3090
+    },
+    {
+      "epoch": 0.08072145888847419,
+      "grad_norm": 0.1621125340461731,
+      "learning_rate": 0.0001899014310711522,
+      "loss": 0.7318,
+      "step": 3100
+    },
+    {
+      "epoch": 0.08098185069134024,
+      "grad_norm": 0.16480112075805664,
+      "learning_rate": 0.0001898281929915629,
+      "loss": 0.7145,
+      "step": 3110
+    },
+    {
+      "epoch": 0.08124224249420628,
+      "grad_norm": 0.1805388629436493,
+      "learning_rate": 0.00018975470452511112,
+      "loss": 0.7102,
+      "step": 3120
+    },
+    {
+      "epoch": 0.08150263429707233,
+      "grad_norm": 0.1902652531862259,
+      "learning_rate": 0.00018968096587663853,
+      "loss": 0.7281,
+      "step": 3130
+    },
+    {
+      "epoch": 0.08176302609993838,
+      "grad_norm": 0.1732487976551056,
+      "learning_rate": 0.00018960697725168397,
+      "loss": 0.7434,
+      "step": 3140
+    },
+    {
+      "epoch": 0.08202341790280442,
+      "grad_norm": 0.1662171334028244,
+      "learning_rate": 0.00018953273885648314,
+      "loss": 0.716,
+      "step": 3150
+    },
+    {
+      "epoch": 0.08228380970567047,
+      "grad_norm": 0.16129222512245178,
+      "learning_rate": 0.00018945825089796797,
+      "loss": 0.7318,
+      "step": 3160
+    },
+    {
+      "epoch": 0.08254420150853652,
+      "grad_norm": 0.16837772727012634,
+      "learning_rate": 0.00018938351358376596,
+      "loss": 0.7137,
+      "step": 3170
+    },
+    {
+      "epoch": 0.08280459331140255,
+      "grad_norm": 0.1618524193763733,
+      "learning_rate": 0.00018930852712219974,
+      "loss": 0.7079,
+      "step": 3180
+    },
+    {
+      "epoch": 0.0830649851142686,
+      "grad_norm": 0.16333432495594025,
+      "learning_rate": 0.00018923329172228632,
+      "loss": 0.7062,
+      "step": 3190
+    },
+    {
+      "epoch": 0.08332537691713465,
+      "grad_norm": 0.15985700488090515,
+      "learning_rate": 0.00018915780759373672,
+      "loss": 0.7277,
+      "step": 3200
+    },
+    {
+      "epoch": 0.08358576872000069,
+      "grad_norm": 0.16181236505508423,
+      "learning_rate": 0.0001890820749469551,
+      "loss": 0.7048,
+      "step": 3210
+    },
+    {
+      "epoch": 0.08384616052286674,
+      "grad_norm": 0.1679672747850418,
+      "learning_rate": 0.00018900609399303853,
+      "loss": 0.73,
+      "step": 3220
+    },
+    {
+      "epoch": 0.08410655232573279,
+      "grad_norm": 0.16680286824703217,
+      "learning_rate": 0.00018892986494377606,
+      "loss": 0.7169,
+      "step": 3230
+    },
+    {
+      "epoch": 0.08436694412859883,
+      "grad_norm": 0.15980315208435059,
+      "learning_rate": 0.00018885338801164834,
+      "loss": 0.7346,
+      "step": 3240
+    },
+    {
+      "epoch": 0.08462733593146488,
+      "grad_norm": 0.16863352060317993,
+      "learning_rate": 0.00018877666340982695,
+      "loss": 0.7256,
+      "step": 3250
+    },
+    {
+      "epoch": 0.08488772773433093,
+      "grad_norm": 0.1657836139202118,
+      "learning_rate": 0.0001886996913521739,
+      "loss": 0.7219,
+      "step": 3260
+    },
+    {
+      "epoch": 0.08514811953719698,
+      "grad_norm": 0.1683470606803894,
+      "learning_rate": 0.00018862247205324087,
+      "loss": 0.7279,
+      "step": 3270
+    },
+    {
+      "epoch": 0.08540851134006301,
+      "grad_norm": 0.1686122715473175,
+      "learning_rate": 0.00018854500572826867,
+      "loss": 0.7178,
+      "step": 3280
+    },
+    {
+      "epoch": 0.08566890314292906,
+      "grad_norm": 0.16122782230377197,
+      "learning_rate": 0.00018846729259318682,
+      "loss": 0.7289,
+      "step": 3290
+    },
+    {
+      "epoch": 0.08592929494579511,
+      "grad_norm": 0.1920589804649353,
+      "learning_rate": 0.0001883893328646126,
+      "loss": 0.7264,
+      "step": 3300
+    },
+    {
+      "epoch": 0.08618968674866115,
+      "grad_norm": 0.17415335774421692,
+      "learning_rate": 0.00018831112675985083,
+      "loss": 0.7378,
+      "step": 3310
+    },
+    {
+      "epoch": 0.0864500785515272,
+      "grad_norm": 0.16903561353683472,
+      "learning_rate": 0.00018823267449689292,
+      "loss": 0.702,
+      "step": 3320
+    },
+    {
+      "epoch": 0.08671047035439325,
+      "grad_norm": 0.17158570885658264,
+      "learning_rate": 0.00018815397629441658,
+      "loss": 0.7264,
+      "step": 3330
+    },
+    {
+      "epoch": 0.08697086215725928,
+      "grad_norm": 0.17094087600708008,
+      "learning_rate": 0.0001880750323717849,
+      "loss": 0.741,
+      "step": 3340
+    },
+    {
+      "epoch": 0.08723125396012534,
+      "grad_norm": 0.1618223935365677,
+      "learning_rate": 0.000187995842949046,
+      "loss": 0.7084,
+      "step": 3350
+    },
+    {
+      "epoch": 0.08749164576299139,
+      "grad_norm": 0.1671626716852188,
+      "learning_rate": 0.0001879164082469322,
+      "loss": 0.7236,
+      "step": 3360
+    },
+    {
+      "epoch": 0.08775203756585742,
+      "grad_norm": 0.1681569665670395,
+      "learning_rate": 0.00018783672848685966,
+      "loss": 0.7382,
+      "step": 3370
+    },
+    {
+      "epoch": 0.08801242936872347,
+      "grad_norm": 0.16421955823898315,
+      "learning_rate": 0.00018775680389092748,
+      "loss": 0.717,
+      "step": 3380
+    },
+    {
+      "epoch": 0.08827282117158952,
+      "grad_norm": 0.174809530377388,
+      "learning_rate": 0.00018767663468191725,
+      "loss": 0.7225,
+      "step": 3390
+    },
+    {
+      "epoch": 0.08853321297445556,
+      "grad_norm": 0.16181902587413788,
+      "learning_rate": 0.00018759622108329243,
+      "loss": 0.7014,
+      "step": 3400
+    },
+    {
+      "epoch": 0.08879360477732161,
+      "grad_norm": 0.15579254925251007,
+      "learning_rate": 0.0001875155633191977,
+      "loss": 0.7125,
+      "step": 3410
+    },
+    {
+      "epoch": 0.08905399658018766,
+      "grad_norm": 0.16342496871948242,
+      "learning_rate": 0.00018743466161445823,
+      "loss": 0.7075,
+      "step": 3420
+    },
+    {
+      "epoch": 0.08931438838305371,
+      "grad_norm": 0.17215611040592194,
+      "learning_rate": 0.00018735351619457923,
+      "loss": 0.7331,
+      "step": 3430
+    },
+    {
+      "epoch": 0.08957478018591974,
+      "grad_norm": 0.1682904213666916,
+      "learning_rate": 0.00018727212728574522,
+      "loss": 0.734,
+      "step": 3440
+    },
+    {
+      "epoch": 0.0898351719887858,
+      "grad_norm": 0.16969889402389526,
+      "learning_rate": 0.00018719049511481948,
+      "loss": 0.7224,
+      "step": 3450
+    },
+    {
+      "epoch": 0.09009556379165184,
+      "grad_norm": 0.16607950627803802,
+      "learning_rate": 0.00018710861990934324,
+      "loss": 0.7218,
+      "step": 3460
+    },
+    {
+      "epoch": 0.09035595559451788,
+      "grad_norm": 0.16665585339069366,
+      "learning_rate": 0.00018702650189753525,
+      "loss": 0.7152,
+      "step": 3470
+    },
+    {
+      "epoch": 0.09061634739738393,
+      "grad_norm": 0.16812992095947266,
+      "learning_rate": 0.00018694414130829103,
+      "loss": 0.7097,
+      "step": 3480
+    },
+    {
+      "epoch": 0.09087673920024998,
+      "grad_norm": 0.16855508089065552,
+      "learning_rate": 0.00018686153837118224,
+      "loss": 0.7268,
+      "step": 3490
+    },
+    {
+      "epoch": 0.09113713100311602,
+      "grad_norm": 0.1634734570980072,
+      "learning_rate": 0.00018677869331645613,
+      "loss": 0.7485,
+      "step": 3500
+    },
+    {
+      "epoch": 0.09139752280598207,
+      "grad_norm": 0.17040792107582092,
+      "learning_rate": 0.00018669560637503478,
+      "loss": 0.7529,
+      "step": 3510
+    },
+    {
+      "epoch": 0.09165791460884812,
+      "grad_norm": 0.16575287282466888,
+      "learning_rate": 0.00018661227777851447,
+      "loss": 0.7053,
+      "step": 3520
+    },
+    {
+      "epoch": 0.09191830641171415,
+      "grad_norm": 0.16195647418498993,
+      "learning_rate": 0.00018652870775916515,
+      "loss": 0.7498,
+      "step": 3530
+    },
+    {
+      "epoch": 0.0921786982145802,
+      "grad_norm": 0.16672903299331665,
+      "learning_rate": 0.00018644489654992956,
+      "loss": 0.7168,
+      "step": 3540
+    },
+    {
+      "epoch": 0.09243909001744625,
+      "grad_norm": 0.16135342419147491,
+      "learning_rate": 0.00018636084438442298,
+      "loss": 0.6868,
+      "step": 3550
+    },
+    {
+      "epoch": 0.09269948182031229,
+      "grad_norm": 0.16677817702293396,
+      "learning_rate": 0.0001862765514969321,
+      "loss": 0.7128,
+      "step": 3560
+    },
+    {
+      "epoch": 0.09295987362317834,
+      "grad_norm": 0.16983947157859802,
+      "learning_rate": 0.00018619201812241474,
+      "loss": 0.7189,
+      "step": 3570
+    },
+    {
+      "epoch": 0.09322026542604439,
+      "grad_norm": 0.17054444551467896,
+      "learning_rate": 0.00018610724449649895,
+      "loss": 0.6916,
+      "step": 3580
+    },
+    {
+      "epoch": 0.09348065722891043,
+      "grad_norm": 0.1672518253326416,
+      "learning_rate": 0.0001860222308554825,
+      "loss": 0.7002,
+      "step": 3590
+    },
+    {
+      "epoch": 0.09374104903177648,
+      "grad_norm": 0.16726090013980865,
+      "learning_rate": 0.00018593697743633225,
+      "loss": 0.7153,
+      "step": 3600
+    },
+    {
+      "epoch": 0.09400144083464253,
+      "grad_norm": 0.16627268493175507,
+      "learning_rate": 0.0001858514844766833,
+      "loss": 0.7121,
+      "step": 3610
+    },
+    {
+      "epoch": 0.09426183263750858,
+      "grad_norm": 0.16322413086891174,
+      "learning_rate": 0.00018576575221483852,
+      "loss": 0.7018,
+      "step": 3620
+    },
+    {
+      "epoch": 0.09452222444037461,
+      "grad_norm": 0.1642254739999771,
+      "learning_rate": 0.00018567978088976776,
+      "loss": 0.706,
+      "step": 3630
+    },
+    {
+      "epoch": 0.09478261624324066,
+      "grad_norm": 0.17001990973949432,
+      "learning_rate": 0.00018559357074110727,
+      "loss": 0.7236,
+      "step": 3640
+    },
+    {
+      "epoch": 0.09504300804610671,
+      "grad_norm": 0.166624054312706,
+      "learning_rate": 0.00018550712200915897,
+      "loss": 0.7129,
+      "step": 3650
+    },
+    {
+      "epoch": 0.09530339984897275,
+      "grad_norm": 0.17228394746780396,
+      "learning_rate": 0.00018542043493488986,
+      "loss": 0.7096,
+      "step": 3660
+    },
+    {
+      "epoch": 0.0955637916518388,
+      "grad_norm": 0.17133943736553192,
+      "learning_rate": 0.0001853335097599312,
+      "loss": 0.7164,
+      "step": 3670
+    },
+    {
+      "epoch": 0.09582418345470485,
+      "grad_norm": 0.16100934147834778,
+      "learning_rate": 0.00018524634672657805,
+      "loss": 0.711,
+      "step": 3680
+    },
+    {
+      "epoch": 0.09608457525757089,
+      "grad_norm": 0.17711074650287628,
+      "learning_rate": 0.00018515894607778837,
+      "loss": 0.7153,
+      "step": 3690
+    },
+    {
+      "epoch": 0.09634496706043694,
+      "grad_norm": 0.20850953459739685,
+      "learning_rate": 0.00018507130805718252,
+      "loss": 0.7309,
+      "step": 3700
+    },
+    {
+      "epoch": 0.09660535886330299,
+      "grad_norm": 0.15805912017822266,
+      "learning_rate": 0.00018498343290904246,
+      "loss": 0.72,
+      "step": 3710
+    },
+    {
+      "epoch": 0.09686575066616902,
+      "grad_norm": 0.17850877344608307,
+      "learning_rate": 0.00018489532087831114,
+      "loss": 0.7235,
+      "step": 3720
+    },
+    {
+      "epoch": 0.09712614246903507,
+      "grad_norm": 0.16705553233623505,
+      "learning_rate": 0.0001848069722105918,
+      "loss": 0.7233,
+      "step": 3730
+    },
+    {
+      "epoch": 0.09738653427190112,
+      "grad_norm": 0.17024511098861694,
+      "learning_rate": 0.0001847183871521473,
+      "loss": 0.7127,
+      "step": 3740
+    },
+    {
+      "epoch": 0.09764692607476716,
+      "grad_norm": 0.16217222809791565,
+      "learning_rate": 0.0001846295659498994,
+      "loss": 0.6844,
+      "step": 3750
+    },
+    {
+      "epoch": 0.09790731787763321,
+      "grad_norm": 0.20443694293498993,
+      "learning_rate": 0.00018454050885142803,
+      "loss": 0.7327,
+      "step": 3760
+    },
+    {
+      "epoch": 0.09816770968049926,
+      "grad_norm": 0.16370315849781036,
+      "learning_rate": 0.00018445121610497072,
+      "loss": 0.7191,
+      "step": 3770
+    },
+    {
+      "epoch": 0.09842810148336531,
+      "grad_norm": 0.16387638449668884,
+      "learning_rate": 0.00018436168795942187,
+      "loss": 0.724,
+      "step": 3780
+    },
+    {
+      "epoch": 0.09868849328623135,
+      "grad_norm": 0.1664215326309204,
+      "learning_rate": 0.00018427192466433193,
+      "loss": 0.6955,
+      "step": 3790
+    },
+    {
+      "epoch": 0.0989488850890974,
+      "grad_norm": 0.17285983264446259,
+      "learning_rate": 0.0001841819264699069,
+      "loss": 0.7063,
+      "step": 3800
+    },
+    {
+      "epoch": 0.09920927689196345,
+      "grad_norm": 0.1637914627790451,
+      "learning_rate": 0.00018409169362700753,
+      "loss": 0.7211,
+      "step": 3810
+    },
+    {
+      "epoch": 0.09946966869482948,
+      "grad_norm": 0.1695072054862976,
+      "learning_rate": 0.00018400122638714855,
+      "loss": 0.7172,
+      "step": 3820
+    },
+    {
+      "epoch": 0.09973006049769553,
+      "grad_norm": 0.16097985208034515,
+      "learning_rate": 0.00018391052500249812,
+      "loss": 0.7073,
+      "step": 3830
+    },
+    {
+      "epoch": 0.09999045230056158,
+      "grad_norm": 0.16112865507602692,
+      "learning_rate": 0.00018381958972587707,
+      "loss": 0.7126,
+      "step": 3840
+    },
+    {
+      "epoch": 0.10025084410342762,
+      "grad_norm": 0.16681082546710968,
+      "learning_rate": 0.0001837284208107581,
+      "loss": 0.7331,
+      "step": 3850
+    },
+    {
+      "epoch": 0.10051123590629367,
+      "grad_norm": 0.16337643563747406,
+      "learning_rate": 0.0001836370185112652,
+      "loss": 0.7185,
+      "step": 3860
+    },
+    {
+      "epoch": 0.10077162770915972,
+      "grad_norm": 0.16530117392539978,
+      "learning_rate": 0.0001835453830821729,
+      "loss": 0.7058,
+      "step": 3870
+    },
+    {
+      "epoch": 0.10103201951202576,
+      "grad_norm": 0.1671489030122757,
+      "learning_rate": 0.00018345351477890554,
+      "loss": 0.7225,
+      "step": 3880
+    },
+    {
+      "epoch": 0.1012924113148918,
+      "grad_norm": 0.1631312221288681,
+      "learning_rate": 0.00018336141385753661,
+      "loss": 0.7189,
+      "step": 3890
+    },
+    {
+      "epoch": 0.10155280311775786,
+      "grad_norm": 0.16685089468955994,
+      "learning_rate": 0.00018326908057478795,
+      "loss": 0.7076,
+      "step": 3900
+    },
+    {
+      "epoch": 0.10181319492062389,
+      "grad_norm": 0.16232247650623322,
+      "learning_rate": 0.00018317651518802913,
+      "loss": 0.7168,
+      "step": 3910
+    },
+    {
+      "epoch": 0.10207358672348994,
+      "grad_norm": 0.16136577725410461,
+      "learning_rate": 0.0001830837179552766,
+      "loss": 0.7335,
+      "step": 3920
+    },
+    {
+      "epoch": 0.10233397852635599,
+      "grad_norm": 0.17820622026920319,
+      "learning_rate": 0.00018299068913519312,
+      "loss": 0.7071,
+      "step": 3930
+    },
+    {
+      "epoch": 0.10259437032922204,
+      "grad_norm": 0.15879027545452118,
+      "learning_rate": 0.00018289742898708707,
+      "loss": 0.7301,
+      "step": 3940
+    },
+    {
+      "epoch": 0.10285476213208808,
+      "grad_norm": 0.1569281965494156,
+      "learning_rate": 0.00018280393777091143,
+      "loss": 0.7018,
+      "step": 3950
+    },
+    {
+      "epoch": 0.10311515393495413,
+      "grad_norm": 0.15729983150959015,
+      "learning_rate": 0.00018271021574726336,
+      "loss": 0.717,
+      "step": 3960
+    },
+    {
+      "epoch": 0.10337554573782018,
+      "grad_norm": 0.1715734452009201,
+      "learning_rate": 0.00018261626317738348,
+      "loss": 0.7008,
+      "step": 3970
+    },
+    {
+      "epoch": 0.10363593754068622,
+      "grad_norm": 0.16048088669776917,
+      "learning_rate": 0.00018252208032315482,
+      "loss": 0.7108,
+      "step": 3980
+    },
+    {
+      "epoch": 0.10389632934355227,
+      "grad_norm": 0.1652391403913498,
+      "learning_rate": 0.00018242766744710248,
+      "loss": 0.7164,
+      "step": 3990
+    },
+    {
+      "epoch": 0.10415672114641832,
+      "grad_norm": 0.17617185413837433,
+      "learning_rate": 0.0001823330248123926,
+      "loss": 0.7152,
+      "step": 4000
+    },
+    {
+      "epoch": 0.10441711294928435,
+      "grad_norm": 0.16789917647838593,
+      "learning_rate": 0.00018223815268283186,
+      "loss": 0.742,
+      "step": 4010
+    },
+    {
+      "epoch": 0.1046775047521504,
+      "grad_norm": 0.15830789506435394,
+      "learning_rate": 0.00018214305132286656,
+      "loss": 0.7127,
+      "step": 4020
+    },
+    {
+      "epoch": 0.10493789655501645,
+      "grad_norm": 0.16791877150535583,
+      "learning_rate": 0.00018204772099758194,
+      "loss": 0.7173,
+      "step": 4030
+    },
+    {
+      "epoch": 0.10519828835788249,
+      "grad_norm": 0.1708458811044693,
+      "learning_rate": 0.00018195216197270156,
+      "loss": 0.7152,
+      "step": 4040
+    },
+    {
+      "epoch": 0.10545868016074854,
+      "grad_norm": 0.16406644880771637,
+      "learning_rate": 0.0001818563745145863,
+      "loss": 0.709,
+      "step": 4050
+    },
+    {
+      "epoch": 0.10571907196361459,
+      "grad_norm": 0.16491225361824036,
+      "learning_rate": 0.00018176035889023396,
+      "loss": 0.6957,
+      "step": 4060
+    },
+    {
+      "epoch": 0.10597946376648062,
+      "grad_norm": 0.18035203218460083,
+      "learning_rate": 0.00018166411536727812,
+      "loss": 0.7166,
+      "step": 4070
+    },
+    {
+      "epoch": 0.10623985556934668,
+      "grad_norm": 0.1827452927827835,
+      "learning_rate": 0.0001815676442139878,
+      "loss": 0.7359,
+      "step": 4080
+    },
+    {
+      "epoch": 0.10650024737221273,
+      "grad_norm": 0.17651152610778809,
+      "learning_rate": 0.00018147094569926644,
+      "loss": 0.7055,
+      "step": 4090
+    },
+    {
+      "epoch": 0.10676063917507878,
+      "grad_norm": 0.16017574071884155,
+      "learning_rate": 0.00018137402009265114,
+      "loss": 0.7175,
+      "step": 4100
+    },
+    {
+      "epoch": 0.10702103097794481,
+      "grad_norm": 0.16183608770370483,
+      "learning_rate": 0.00018127686766431213,
+      "loss": 0.7099,
+      "step": 4110
+    },
+    {
+      "epoch": 0.10728142278081086,
+      "grad_norm": 0.173137828707695,
+      "learning_rate": 0.00018117948868505182,
+      "loss": 0.7139,
+      "step": 4120
+    },
+    {
+      "epoch": 0.10754181458367691,
+      "grad_norm": 0.1736387312412262,
+      "learning_rate": 0.00018108188342630413,
+      "loss": 0.7219,
+      "step": 4130
+    },
+    {
+      "epoch": 0.10780220638654295,
+      "grad_norm": 0.1679358333349228,
+      "learning_rate": 0.0001809840521601337,
+      "loss": 0.7057,
+      "step": 4140
+    },
+    {
+      "epoch": 0.108062598189409,
+      "grad_norm": 0.16037985682487488,
+      "learning_rate": 0.00018088599515923503,
+      "loss": 0.7111,
+      "step": 4150
+    },
+    {
+      "epoch": 0.10832298999227505,
+      "grad_norm": 0.15554189682006836,
+      "learning_rate": 0.00018078771269693209,
+      "loss": 0.6978,
+      "step": 4160
+    },
+    {
+      "epoch": 0.10858338179514108,
+      "grad_norm": 0.1662927269935608,
+      "learning_rate": 0.00018068920504717704,
+      "loss": 0.7016,
+      "step": 4170
+    },
+    {
+      "epoch": 0.10884377359800713,
+      "grad_norm": 0.168484628200531,
+      "learning_rate": 0.0001805904724845499,
+      "loss": 0.695,
+      "step": 4180
+    },
+    {
+      "epoch": 0.10910416540087318,
+      "grad_norm": 0.18269601464271545,
+      "learning_rate": 0.0001804915152842575,
+      "loss": 0.7194,
+      "step": 4190
+    },
+    {
+      "epoch": 0.10936455720373922,
+      "grad_norm": 0.17641094326972961,
+      "learning_rate": 0.00018039233372213294,
+      "loss": 0.7209,
+      "step": 4200
+    },
+    {
+      "epoch": 0.10962494900660527,
+      "grad_norm": 0.18733248114585876,
+      "learning_rate": 0.00018029292807463456,
+      "loss": 0.7089,
+      "step": 4210
+    },
+    {
+      "epoch": 0.10988534080947132,
+      "grad_norm": 0.16925212740898132,
+      "learning_rate": 0.00018019329861884543,
+      "loss": 0.7097,
+      "step": 4220
+    },
+    {
+      "epoch": 0.11014573261233736,
+      "grad_norm": 0.15558676421642303,
+      "learning_rate": 0.00018009344563247245,
+      "loss": 0.6766,
+      "step": 4230
+    },
+    {
+      "epoch": 0.11040612441520341,
+      "grad_norm": 0.16530027985572815,
+      "learning_rate": 0.00017999336939384555,
+      "loss": 0.7141,
+      "step": 4240
+    },
+    {
+      "epoch": 0.11066651621806946,
+      "grad_norm": 0.16560573875904083,
+      "learning_rate": 0.00017989307018191693,
+      "loss": 0.6989,
+      "step": 4250
+    },
+    {
+      "epoch": 0.1109269080209355,
+      "grad_norm": 0.17466457188129425,
+      "learning_rate": 0.00017979254827626037,
+      "loss": 0.7206,
+      "step": 4260
+    },
+    {
+      "epoch": 0.11118729982380154,
+      "grad_norm": 0.18086989223957062,
+      "learning_rate": 0.00017969180395707038,
+      "loss": 0.714,
+      "step": 4270
+    },
+    {
+      "epoch": 0.1114476916266676,
+      "grad_norm": 1.1102476119995117,
+      "learning_rate": 0.00017959083750516137,
+      "loss": 0.723,
+      "step": 4280
+    },
+    {
+      "epoch": 0.11170808342953364,
+      "grad_norm": 0.19353629648685455,
+      "learning_rate": 0.00017948964920196697,
+      "loss": 0.718,
+      "step": 4290
+    },
+    {
+      "epoch": 0.11196847523239968,
+      "grad_norm": 0.2126789540052414,
+      "learning_rate": 0.00017938823932953915,
+      "loss": 0.7171,
+      "step": 4300
+    },
+    {
+      "epoch": 0.11222886703526573,
+      "grad_norm": 0.16679035127162933,
+      "learning_rate": 0.00017928660817054758,
+      "loss": 0.7113,
+      "step": 4310
+    },
+    {
+      "epoch": 0.11248925883813178,
+      "grad_norm": 0.1647113859653473,
+      "learning_rate": 0.00017918475600827861,
+      "loss": 0.7184,
+      "step": 4320
+    },
+    {
+      "epoch": 0.11274965064099782,
+      "grad_norm": 0.16823258996009827,
+      "learning_rate": 0.00017908268312663475,
+      "loss": 0.6876,
+      "step": 4330
+    },
+    {
+      "epoch": 0.11301004244386387,
+      "grad_norm": 0.17199194431304932,
+      "learning_rate": 0.0001789803898101336,
+      "loss": 0.6954,
+      "step": 4340
+    },
+    {
+      "epoch": 0.11327043424672992,
+      "grad_norm": 0.16541388630867004,
+      "learning_rate": 0.00017887787634390733,
+      "loss": 0.709,
+      "step": 4350
+    },
+    {
+      "epoch": 0.11353082604959595,
+      "grad_norm": 0.17238549888134003,
+      "learning_rate": 0.00017877514301370168,
+      "loss": 0.7056,
+      "step": 4360
+    },
+    {
+      "epoch": 0.113791217852462,
+      "grad_norm": 0.18374156951904297,
+      "learning_rate": 0.00017867219010587527,
+      "loss": 0.6951,
+      "step": 4370
+    },
+    {
+      "epoch": 0.11405160965532805,
+      "grad_norm": 0.15940402448177338,
+      "learning_rate": 0.00017856901790739874,
+      "loss": 0.6873,
+      "step": 4380
+    },
+    {
+      "epoch": 0.11431200145819409,
+      "grad_norm": 0.16984274983406067,
+      "learning_rate": 0.00017846562670585398,
+      "loss": 0.7065,
+      "step": 4390
+    },
+    {
+      "epoch": 0.11457239326106014,
+      "grad_norm": 0.9227787852287292,
+      "learning_rate": 0.0001783620167894334,
+      "loss": 0.7591,
+      "step": 4400
+    },
+    {
+      "epoch": 0.11483278506392619,
+      "grad_norm": 0.1881164163351059,
+      "learning_rate": 0.00017825818844693892,
+      "loss": 0.7133,
+      "step": 4410
+    },
+    {
+      "epoch": 0.11509317686679223,
+      "grad_norm": 0.20252032577991486,
+      "learning_rate": 0.00017815414196778147,
+      "loss": 0.7145,
+      "step": 4420
+    },
+    {
+      "epoch": 0.11535356866965828,
+      "grad_norm": 0.23380456864833832,
+      "learning_rate": 0.0001780498776419799,
+      "loss": 0.7232,
+      "step": 4430
+    },
+    {
+      "epoch": 0.11561396047252433,
+      "grad_norm": 0.1666252464056015,
+      "learning_rate": 0.00017794539576016028,
+      "loss": 0.7224,
+      "step": 4440
+    },
+    {
+      "epoch": 0.11587435227539038,
+      "grad_norm": 0.2421305775642395,
+      "learning_rate": 0.0001778406966135551,
+      "loss": 0.7031,
+      "step": 4450
+    },
+    {
+      "epoch": 0.11613474407825641,
+      "grad_norm": 0.17607644200325012,
+      "learning_rate": 0.00017773578049400256,
+      "loss": 0.7162,
+      "step": 4460
+    },
+    {
+      "epoch": 0.11639513588112246,
+      "grad_norm": 0.1905539333820343,
+      "learning_rate": 0.00017763064769394553,
+      "loss": 0.7087,
+      "step": 4470
+    },
+    {
+      "epoch": 0.11665552768398851,
+      "grad_norm": 0.17976076900959015,
+      "learning_rate": 0.00017752529850643088,
+      "loss": 0.7274,
+      "step": 4480
+    },
+    {
+      "epoch": 0.11691591948685455,
+      "grad_norm": 0.16786770522594452,
+      "learning_rate": 0.00017741973322510864,
+      "loss": 0.7003,
+      "step": 4490
+    },
+    {
+      "epoch": 0.1171763112897206,
+      "grad_norm": 0.1559310108423233,
+      "learning_rate": 0.00017731395214423126,
+      "loss": 0.7279,
+      "step": 4500
+    },
+    {
+      "epoch": 0.11743670309258665,
+      "grad_norm": 0.16374662518501282,
+      "learning_rate": 0.00017720795555865255,
+      "loss": 0.6924,
+      "step": 4510
+    },
+    {
+      "epoch": 0.11769709489545269,
+      "grad_norm": 0.1647823303937912,
+      "learning_rate": 0.00017710174376382712,
+      "loss": 0.7224,
+      "step": 4520
+    },
+    {
+      "epoch": 0.11795748669831874,
+      "grad_norm": 0.18070833384990692,
+      "learning_rate": 0.0001769953170558095,
+      "loss": 0.7223,
+      "step": 4530
+    },
+    {
+      "epoch": 0.11821787850118479,
+      "grad_norm": 0.16518941521644592,
+      "learning_rate": 0.0001768886757312532,
+      "loss": 0.7153,
+      "step": 4540
+    },
+    {
+      "epoch": 0.11847827030405082,
+      "grad_norm": 0.17081689834594727,
+      "learning_rate": 0.0001767818200874099,
+      "loss": 0.7197,
+      "step": 4550
+    },
+    {
+      "epoch": 0.11873866210691687,
+      "grad_norm": 0.16503626108169556,
+      "learning_rate": 0.00017667475042212874,
+      "loss": 0.7222,
+      "step": 4560
+    },
+    {
+      "epoch": 0.11899905390978292,
+      "grad_norm": 0.17310748994350433,
+      "learning_rate": 0.00017656746703385547,
+      "loss": 0.7057,
+      "step": 4570
+    },
+    {
+      "epoch": 0.11925944571264896,
+      "grad_norm": 0.1613416224718094,
+      "learning_rate": 0.0001764599702216315,
+      "loss": 0.7133,
+      "step": 4580
+    },
+    {
+      "epoch": 0.11951983751551501,
+      "grad_norm": 0.184098020195961,
+      "learning_rate": 0.00017635226028509313,
+      "loss": 0.72,
+      "step": 4590
+    },
+    {
+      "epoch": 0.11978022931838106,
+      "grad_norm": 0.1819763481616974,
+      "learning_rate": 0.00017624433752447076,
+      "loss": 0.7154,
+      "step": 4600
+    },
+    {
+      "epoch": 0.12004062112124711,
+      "grad_norm": 0.17078550159931183,
+      "learning_rate": 0.000176136202240588,
+      "loss": 0.7205,
+      "step": 4610
+    },
+    {
+      "epoch": 0.12030101292411315,
+      "grad_norm": 0.16961923241615295,
+      "learning_rate": 0.00017602785473486084,
+      "loss": 0.6997,
+      "step": 4620
+    },
+    {
+      "epoch": 0.1205614047269792,
+      "grad_norm": 0.18945324420928955,
+      "learning_rate": 0.00017591929530929684,
+      "loss": 0.7295,
+      "step": 4630
+    },
+    {
+      "epoch": 0.12082179652984525,
+      "grad_norm": 0.1690925508737564,
+      "learning_rate": 0.00017581052426649424,
+      "loss": 0.714,
+      "step": 4640
+    },
+    {
+      "epoch": 0.12108218833271128,
+      "grad_norm": 0.16739274561405182,
+      "learning_rate": 0.00017570154190964113,
+      "loss": 0.7018,
+      "step": 4650
+    },
+    {
+      "epoch": 0.12134258013557733,
+      "grad_norm": 0.17123165726661682,
+      "learning_rate": 0.0001755923485425147,
+      "loss": 0.7288,
+      "step": 4660
+    },
+    {
+      "epoch": 0.12160297193844338,
+      "grad_norm": 0.17006556689739227,
+      "learning_rate": 0.0001754829444694802,
+      "loss": 0.7185,
+      "step": 4670
+    },
+    {
+      "epoch": 0.12186336374130942,
+      "grad_norm": 0.17169633507728577,
+      "learning_rate": 0.00017537332999549023,
+      "loss": 0.7169,
+      "step": 4680
+    },
+    {
+      "epoch": 0.12212375554417547,
+      "grad_norm": 0.17851942777633667,
+      "learning_rate": 0.00017526350542608393,
+      "loss": 0.7138,
+      "step": 4690
+    },
+    {
+      "epoch": 0.12238414734704152,
+      "grad_norm": 0.16988542675971985,
+      "learning_rate": 0.00017515347106738597,
+      "loss": 0.7188,
+      "step": 4700
+    },
+    {
+      "epoch": 0.12264453914990756,
+      "grad_norm": 0.16778436303138733,
+      "learning_rate": 0.00017504322722610582,
+      "loss": 0.7034,
+      "step": 4710
+    },
+    {
+      "epoch": 0.1229049309527736,
+      "grad_norm": 0.1835828721523285,
+      "learning_rate": 0.00017493277420953688,
+      "loss": 0.7084,
+      "step": 4720
+    },
+    {
+      "epoch": 0.12316532275563966,
+      "grad_norm": 0.16590015590190887,
+      "learning_rate": 0.0001748221123255556,
+      "loss": 0.7023,
+      "step": 4730
+    },
+    {
+      "epoch": 0.12342571455850569,
+      "grad_norm": 0.16521744430065155,
+      "learning_rate": 0.00017471124188262062,
+      "loss": 0.7224,
+      "step": 4740
+    },
+    {
+      "epoch": 0.12368610636137174,
+      "grad_norm": 0.17280088365077972,
+      "learning_rate": 0.00017460016318977188,
+      "loss": 0.7087,
+      "step": 4750
+    },
+    {
+      "epoch": 0.12394649816423779,
+      "grad_norm": 0.17343877255916595,
+      "learning_rate": 0.00017448887655662984,
+      "loss": 0.7212,
+      "step": 4760
+    },
+    {
+      "epoch": 0.12420688996710384,
+      "grad_norm": 0.17116770148277283,
+      "learning_rate": 0.0001743773822933946,
+      "loss": 0.7289,
+      "step": 4770
+    },
+    {
+      "epoch": 0.12446728176996988,
+      "grad_norm": 0.17440655827522278,
+      "learning_rate": 0.0001742656807108449,
+      "loss": 0.7353,
+      "step": 4780
+    },
+    {
+      "epoch": 0.12472767357283593,
+      "grad_norm": 0.17304064333438873,
+      "learning_rate": 0.0001741537721203375,
+      "loss": 0.7322,
+      "step": 4790
+    },
+    {
+      "epoch": 0.12498806537570198,
+      "grad_norm": 0.17076118290424347,
+      "learning_rate": 0.00017404165683380604,
+      "loss": 0.7199,
+      "step": 4800
+    },
+    {
+      "epoch": 0.12524845717856803,
+      "grad_norm": 0.17371006309986115,
+      "learning_rate": 0.0001739293351637604,
+      "loss": 0.7035,
+      "step": 4810
+    },
+    {
+      "epoch": 0.12550884898143405,
+      "grad_norm": 0.16811051964759827,
+      "learning_rate": 0.0001738168074232857,
+      "loss": 0.7113,
+      "step": 4820
+    },
+    {
+      "epoch": 0.1257692407843001,
+      "grad_norm": 0.17408473789691925,
+      "learning_rate": 0.00017370407392604145,
+      "loss": 0.7151,
+      "step": 4830
+    },
+    {
+      "epoch": 0.12602963258716615,
+      "grad_norm": 0.16915513575077057,
+      "learning_rate": 0.0001735911349862607,
+      "loss": 0.7089,
+      "step": 4840
+    },
+    {
+      "epoch": 0.1262900243900322,
+      "grad_norm": 0.17281986773014069,
+      "learning_rate": 0.00017347799091874913,
+      "loss": 0.6935,
+      "step": 4850
+    },
+    {
+      "epoch": 0.12655041619289825,
+      "grad_norm": 0.16271603107452393,
+      "learning_rate": 0.00017336464203888426,
+      "loss": 0.7131,
+      "step": 4860
+    },
+    {
+      "epoch": 0.1268108079957643,
+      "grad_norm": 0.1661551147699356,
+      "learning_rate": 0.00017325108866261439,
+      "loss": 0.6899,
+      "step": 4870
+    },
+    {
+      "epoch": 0.12707119979863035,
+      "grad_norm": 0.16490161418914795,
+      "learning_rate": 0.00017313733110645792,
+      "loss": 0.7042,
+      "step": 4880
+    },
+    {
+      "epoch": 0.12733159160149637,
+      "grad_norm": 0.17232102155685425,
+      "learning_rate": 0.00017302336968750236,
+      "loss": 0.6965,
+      "step": 4890
+    },
+    {
+      "epoch": 0.12759198340436242,
+      "grad_norm": 0.17850111424922943,
+      "learning_rate": 0.00017290920472340347,
+      "loss": 0.7185,
+      "step": 4900
+    },
+    {
+      "epoch": 0.12785237520722847,
+      "grad_norm": 0.16296492516994476,
+      "learning_rate": 0.00017279483653238434,
+      "loss": 0.7179,
+      "step": 4910
+    },
+    {
+      "epoch": 0.12811276701009452,
+      "grad_norm": 0.1704237312078476,
+      "learning_rate": 0.00017268026543323454,
+      "loss": 0.7103,
+      "step": 4920
+    },
+    {
+      "epoch": 0.12837315881296057,
+      "grad_norm": 0.18265944719314575,
+      "learning_rate": 0.00017256549174530928,
+      "loss": 0.6988,
+      "step": 4930
+    },
+    {
+      "epoch": 0.12863355061582663,
+      "grad_norm": 0.17257574200630188,
+      "learning_rate": 0.0001724505157885284,
+      "loss": 0.7231,
+      "step": 4940
+    },
+    {
+      "epoch": 0.12889394241869265,
+      "grad_norm": 0.1703302413225174,
+      "learning_rate": 0.0001723353378833756,
+      "loss": 0.7192,
+      "step": 4950
+    },
+    {
+      "epoch": 0.1291543342215587,
+      "grad_norm": 0.16430538892745972,
+      "learning_rate": 0.00017221995835089743,
+      "loss": 0.7007,
+      "step": 4960
+    },
+    {
+      "epoch": 0.12941472602442475,
+      "grad_norm": 0.162199467420578,
+      "learning_rate": 0.0001721043775127025,
+      "loss": 0.7123,
+      "step": 4970
+    },
+    {
+      "epoch": 0.1296751178272908,
+      "grad_norm": 0.1680443435907364,
+      "learning_rate": 0.00017198859569096057,
+      "loss": 0.6976,
+      "step": 4980
+    },
+    {
+      "epoch": 0.12993550963015685,
+      "grad_norm": 0.16768567264080048,
+      "learning_rate": 0.0001718726132084015,
+      "loss": 0.6942,
+      "step": 4990
+    },
+    {
+      "epoch": 0.1301959014330229,
+      "grad_norm": 0.16759110987186432,
+      "learning_rate": 0.00017175643038831463,
+      "loss": 0.6995,
+      "step": 5000
+    },
+    {
+      "epoch": 0.13045629323588892,
+      "grad_norm": 0.17288459837436676,
+      "learning_rate": 0.00017164004755454754,
+      "loss": 0.7193,
+      "step": 5010
+    },
+    {
+      "epoch": 0.13071668503875497,
+      "grad_norm": 0.18251213431358337,
+      "learning_rate": 0.00017152346503150552,
+      "loss": 0.7156,
+      "step": 5020
+    },
+    {
+      "epoch": 0.13097707684162102,
+      "grad_norm": 0.17177079617977142,
+      "learning_rate": 0.0001714066831441503,
+      "loss": 0.7354,
+      "step": 5030
+    },
+    {
+      "epoch": 0.13123746864448707,
+      "grad_norm": 0.18522126972675323,
+      "learning_rate": 0.00017128970221799947,
+      "loss": 0.717,
+      "step": 5040
+    },
+    {
+      "epoch": 0.13149786044735312,
+      "grad_norm": 0.17539465427398682,
+      "learning_rate": 0.00017117252257912532,
+      "loss": 0.7204,
+      "step": 5050
+    },
+    {
+      "epoch": 0.13175825225021917,
+      "grad_norm": 0.16526655852794647,
+      "learning_rate": 0.00017105514455415403,
+      "loss": 0.7099,
+      "step": 5060
+    },
+    {
+      "epoch": 0.13201864405308522,
+      "grad_norm": 0.1709279865026474,
+      "learning_rate": 0.00017093756847026483,
+      "loss": 0.7205,
+      "step": 5070
+    },
+    {
+      "epoch": 0.13227903585595124,
+      "grad_norm": 0.16695177555084229,
+      "learning_rate": 0.00017081979465518896,
+      "loss": 0.7092,
+      "step": 5080
+    },
+    {
+      "epoch": 0.1325394276588173,
+      "grad_norm": 0.1795840859413147,
+      "learning_rate": 0.00017070182343720882,
+      "loss": 0.6983,
+      "step": 5090
+    },
+    {
+      "epoch": 0.13279981946168334,
+      "grad_norm": 0.17300280928611755,
+      "learning_rate": 0.0001705836551451571,
+      "loss": 0.7018,
+      "step": 5100
+    },
+    {
+      "epoch": 0.1330602112645494,
+      "grad_norm": 0.16966886818408966,
+      "learning_rate": 0.00017046529010841574,
+      "loss": 0.708,
+      "step": 5110
+    },
+    {
+      "epoch": 0.13332060306741544,
+      "grad_norm": 0.18426869809627533,
+      "learning_rate": 0.00017034672865691515,
+      "loss": 0.7118,
+      "step": 5120
+    },
+    {
+      "epoch": 0.1335809948702815,
+      "grad_norm": 0.1808805912733078,
+      "learning_rate": 0.00017022797112113314,
+      "loss": 0.7041,
+      "step": 5130
+    },
+    {
+      "epoch": 0.13384138667314752,
+      "grad_norm": 0.17788094282150269,
+      "learning_rate": 0.00017010901783209427,
+      "loss": 0.7032,
+      "step": 5140
+    },
+    {
+      "epoch": 0.13410177847601357,
+      "grad_norm": 0.17427201569080353,
+      "learning_rate": 0.00016998986912136852,
+      "loss": 0.7194,
+      "step": 5150
+    },
+    {
+      "epoch": 0.13436217027887962,
+      "grad_norm": 0.17783771455287933,
+      "learning_rate": 0.0001698705253210707,
+      "loss": 0.705,
+      "step": 5160
+    },
+    {
+      "epoch": 0.13462256208174567,
+      "grad_norm": 0.16358746588230133,
+      "learning_rate": 0.0001697509867638594,
+      "loss": 0.7118,
+      "step": 5170
+    },
+    {
+      "epoch": 0.13488295388461172,
+      "grad_norm": 0.17140185832977295,
+      "learning_rate": 0.0001696312537829361,
+      "loss": 0.6929,
+      "step": 5180
+    },
+    {
+      "epoch": 0.13514334568747777,
+      "grad_norm": 0.1690370887517929,
+      "learning_rate": 0.0001695113267120441,
+      "loss": 0.7044,
+      "step": 5190
+    },
+    {
+      "epoch": 0.13540373749034382,
+      "grad_norm": 0.16261689364910126,
+      "learning_rate": 0.0001693912058854679,
+      "loss": 0.6933,
+      "step": 5200
+    },
+    {
+      "epoch": 0.13566412929320984,
+      "grad_norm": 0.16243134438991547,
+      "learning_rate": 0.00016927089163803192,
+      "loss": 0.7009,
+      "step": 5210
+    },
+    {
+      "epoch": 0.1359245210960759,
+      "grad_norm": 0.17476530373096466,
+      "learning_rate": 0.00016915038430509978,
+      "loss": 0.7215,
+      "step": 5220
+    },
+    {
+      "epoch": 0.13618491289894194,
+      "grad_norm": 0.167569100856781,
+      "learning_rate": 0.00016902968422257325,
+      "loss": 0.6975,
+      "step": 5230
+    },
+    {
+      "epoch": 0.136445304701808,
+      "grad_norm": 0.1719943732023239,
+      "learning_rate": 0.00016890879172689146,
+      "loss": 0.7324,
+      "step": 5240
+    },
+    {
+      "epoch": 0.13670569650467404,
+      "grad_norm": 0.167982816696167,
+      "learning_rate": 0.00016878770715502984,
+      "loss": 0.6884,
+      "step": 5250
+    },
+    {
+      "epoch": 0.1369660883075401,
+      "grad_norm": 0.1711047738790512,
+      "learning_rate": 0.00016866643084449914,
+      "loss": 0.7164,
+      "step": 5260
+    },
+    {
+      "epoch": 0.1372264801104061,
+      "grad_norm": 0.17784222960472107,
+      "learning_rate": 0.00016854496313334466,
+      "loss": 0.6987,
+      "step": 5270
+    },
+    {
+      "epoch": 0.13748687191327216,
+      "grad_norm": 0.16513650119304657,
+      "learning_rate": 0.00016842330436014517,
+      "loss": 0.7078,
+      "step": 5280
+    },
+    {
+      "epoch": 0.1377472637161382,
+      "grad_norm": 0.17001712322235107,
+      "learning_rate": 0.000168301454864012,
+      "loss": 0.7053,
+      "step": 5290
+    },
+    {
+      "epoch": 0.13800765551900426,
+      "grad_norm": 0.1930648535490036,
+      "learning_rate": 0.0001681794149845881,
+      "loss": 0.6954,
+      "step": 5300
+    },
+    {
+      "epoch": 0.1382680473218703,
+      "grad_norm": 0.1823079138994217,
+      "learning_rate": 0.00016805718506204711,
+      "loss": 0.7325,
+      "step": 5310
+    },
+    {
+      "epoch": 0.13852843912473636,
+      "grad_norm": 0.16231094300746918,
+      "learning_rate": 0.00016793476543709238,
+      "loss": 0.7074,
+      "step": 5320
+    },
+    {
+      "epoch": 0.13878883092760239,
+      "grad_norm": 0.16010914742946625,
+      "learning_rate": 0.00016781215645095604,
+      "loss": 0.6946,
+      "step": 5330
+    },
+    {
+      "epoch": 0.13904922273046844,
+      "grad_norm": 0.17128457129001617,
+      "learning_rate": 0.00016768935844539804,
+      "loss": 0.7116,
+      "step": 5340
+    },
+    {
+      "epoch": 0.13930961453333449,
+      "grad_norm": 0.17700603604316711,
+      "learning_rate": 0.00016756637176270522,
+      "loss": 0.7276,
+      "step": 5350
+    },
+    {
+      "epoch": 0.13957000633620054,
+      "grad_norm": 0.16596612334251404,
+      "learning_rate": 0.00016744319674569033,
+      "loss": 0.6683,
+      "step": 5360
+    },
+    {
+      "epoch": 0.1398303981390666,
+      "grad_norm": 0.17127594351768494,
+      "learning_rate": 0.00016731983373769106,
+      "loss": 0.7053,
+      "step": 5370
+    },
+    {
+      "epoch": 0.14009078994193264,
+      "grad_norm": 0.17723415791988373,
+      "learning_rate": 0.00016719628308256915,
+      "loss": 0.7294,
+      "step": 5380
+    },
+    {
+      "epoch": 0.1403511817447987,
+      "grad_norm": 0.16194766759872437,
+      "learning_rate": 0.00016707254512470936,
+      "loss": 0.7164,
+      "step": 5390
+    },
+    {
+      "epoch": 0.1406115735476647,
+      "grad_norm": 0.17294222116470337,
+      "learning_rate": 0.00016694862020901852,
+      "loss": 0.7084,
+      "step": 5400
+    },
+    {
+      "epoch": 0.14087196535053076,
+      "grad_norm": 0.17684406042099,
+      "learning_rate": 0.00016682450868092463,
+      "loss": 0.7141,
+      "step": 5410
+    },
+    {
+      "epoch": 0.1411323571533968,
+      "grad_norm": 0.16746504604816437,
+      "learning_rate": 0.00016670021088637579,
+      "loss": 0.7336,
+      "step": 5420
+    },
+    {
+      "epoch": 0.14139274895626286,
+      "grad_norm": 0.2371237874031067,
+      "learning_rate": 0.00016657572717183937,
+      "loss": 0.7001,
+      "step": 5430
+    },
+    {
+      "epoch": 0.1416531407591289,
+      "grad_norm": 0.16818630695343018,
+      "learning_rate": 0.00016645105788430095,
+      "loss": 0.7124,
+      "step": 5440
+    },
+    {
+      "epoch": 0.14191353256199496,
+      "grad_norm": 0.1719949096441269,
+      "learning_rate": 0.00016632620337126333,
+      "loss": 0.7169,
+      "step": 5450
+    },
+    {
+      "epoch": 0.14217392436486098,
+      "grad_norm": 0.1754136085510254,
+      "learning_rate": 0.00016620116398074567,
+      "loss": 0.6966,
+      "step": 5460
+    },
+    {
+      "epoch": 0.14243431616772703,
+      "grad_norm": 0.1630551517009735,
+      "learning_rate": 0.0001660759400612824,
+      "loss": 0.695,
+      "step": 5470
+    },
+    {
+      "epoch": 0.14269470797059308,
+      "grad_norm": 0.16431647539138794,
+      "learning_rate": 0.00016595053196192234,
+      "loss": 0.6983,
+      "step": 5480
+    },
+    {
+      "epoch": 0.14295509977345913,
+      "grad_norm": 0.1706673949956894,
+      "learning_rate": 0.00016582494003222772,
+      "loss": 0.7025,
+      "step": 5490
+    },
+    {
+      "epoch": 0.14321549157632518,
+      "grad_norm": 0.1757555902004242,
+      "learning_rate": 0.00016569916462227312,
+      "loss": 0.7159,
+      "step": 5500
+    },
+    {
+      "epoch": 0.14347588337919123,
+      "grad_norm": 0.1697273999452591,
+      "learning_rate": 0.0001655732060826446,
+      "loss": 0.7112,
+      "step": 5510
+    },
+    {
+      "epoch": 0.14373627518205728,
+      "grad_norm": 0.17706693708896637,
+      "learning_rate": 0.00016544706476443862,
+      "loss": 0.746,
+      "step": 5520
+    },
+    {
+      "epoch": 0.1439966669849233,
+      "grad_norm": 0.17265696823596954,
+      "learning_rate": 0.00016532074101926117,
+      "loss": 0.7276,
+      "step": 5530
+    },
+    {
+      "epoch": 0.14425705878778936,
+      "grad_norm": 0.16773734986782074,
+      "learning_rate": 0.00016519423519922668,
+      "loss": 0.6944,
+      "step": 5540
+    },
+    {
+      "epoch": 0.1445174505906554,
+      "grad_norm": 0.17664338648319244,
+      "learning_rate": 0.0001650675476569572,
+      "loss": 0.6953,
+      "step": 5550
+    },
+    {
+      "epoch": 0.14477784239352146,
+      "grad_norm": 0.17331312596797943,
+      "learning_rate": 0.00016494067874558117,
+      "loss": 0.7113,
+      "step": 5560
+    },
+    {
+      "epoch": 0.1450382341963875,
+      "grad_norm": 0.1759510487318039,
+      "learning_rate": 0.00016481362881873272,
+      "loss": 0.71,
+      "step": 5570
+    },
+    {
+      "epoch": 0.14529862599925356,
+      "grad_norm": 0.18095999956130981,
+      "learning_rate": 0.00016468639823055044,
+      "loss": 0.7255,
+      "step": 5580
+    },
+    {
+      "epoch": 0.14555901780211958,
+      "grad_norm": 0.17434170842170715,
+      "learning_rate": 0.0001645589873356765,
+      "loss": 0.6989,
+      "step": 5590
+    },
+    {
+      "epoch": 0.14581940960498563,
+      "grad_norm": 0.16754081845283508,
+      "learning_rate": 0.00016443139648925572,
+      "loss": 0.7196,
+      "step": 5600
+    },
+    {
+      "epoch": 0.14607980140785168,
+      "grad_norm": 0.17215152084827423,
+      "learning_rate": 0.00016430362604693448,
+      "loss": 0.7104,
+      "step": 5610
+    },
+    {
+      "epoch": 0.14634019321071773,
+      "grad_norm": 0.17826251685619354,
+      "learning_rate": 0.00016417567636485973,
+      "loss": 0.7142,
+      "step": 5620
+    },
+    {
+      "epoch": 0.14660058501358378,
+      "grad_norm": 0.17078974843025208,
+      "learning_rate": 0.00016404754779967813,
+      "loss": 0.717,
+      "step": 5630
+    },
+    {
+      "epoch": 0.14686097681644983,
+      "grad_norm": 0.18085996806621552,
+      "learning_rate": 0.0001639192407085348,
+      "loss": 0.711,
+      "step": 5640
+    },
+    {
+      "epoch": 0.14712136861931585,
+      "grad_norm": 0.19176393747329712,
+      "learning_rate": 0.00016379075544907268,
+      "loss": 0.699,
+      "step": 5650
+    },
+    {
+      "epoch": 0.1473817604221819,
+      "grad_norm": 0.17717042565345764,
+      "learning_rate": 0.00016366209237943122,
+      "loss": 0.7069,
+      "step": 5660
+    },
+    {
+      "epoch": 0.14764215222504795,
+      "grad_norm": 0.17190402746200562,
+      "learning_rate": 0.0001635332518582455,
+      "loss": 0.6988,
+      "step": 5670
+    },
+    {
+      "epoch": 0.147902544027914,
+      "grad_norm": 0.18004834651947021,
+      "learning_rate": 0.00016340423424464516,
+      "loss": 0.7196,
+      "step": 5680
+    },
+    {
+      "epoch": 0.14816293583078005,
+      "grad_norm": 0.17457543313503265,
+      "learning_rate": 0.00016327503989825364,
+      "loss": 0.7109,
+      "step": 5690
+    },
+    {
+      "epoch": 0.1484233276336461,
+      "grad_norm": 0.1904648393392563,
+      "learning_rate": 0.00016314566917918693,
+      "loss": 0.6874,
+      "step": 5700
+    },
+    {
+      "epoch": 0.14868371943651215,
+      "grad_norm": 0.18096795678138733,
+      "learning_rate": 0.00016301612244805252,
+      "loss": 0.7154,
+      "step": 5710
+    },
+    {
+      "epoch": 0.14894411123937817,
+      "grad_norm": 0.17521658539772034,
+      "learning_rate": 0.00016288640006594874,
+      "loss": 0.703,
+      "step": 5720
+    },
+    {
+      "epoch": 0.14920450304224422,
+      "grad_norm": 0.180609330534935,
+      "learning_rate": 0.00016275650239446328,
+      "loss": 0.7033,
+      "step": 5730
+    },
+    {
+      "epoch": 0.14946489484511027,
+      "grad_norm": 0.19524458050727844,
+      "learning_rate": 0.00016262642979567268,
+      "loss": 0.6925,
+      "step": 5740
+    },
+    {
+      "epoch": 0.14972528664797632,
+      "grad_norm": 0.16790035367012024,
+      "learning_rate": 0.00016249618263214087,
+      "loss": 0.7104,
+      "step": 5750
+    },
+    {
+      "epoch": 0.14998567845084237,
+      "grad_norm": 0.16895075142383575,
+      "learning_rate": 0.00016236576126691843,
+      "loss": 0.7126,
+      "step": 5760
+    },
+    {
+      "epoch": 0.15024607025370842,
+      "grad_norm": 0.18354611098766327,
+      "learning_rate": 0.00016223516606354163,
+      "loss": 0.7143,
+      "step": 5770
+    },
+    {
+      "epoch": 0.15050646205657445,
+      "grad_norm": 0.1766786128282547,
+      "learning_rate": 0.00016210439738603108,
+      "loss": 0.7164,
+      "step": 5780
+    },
+    {
+      "epoch": 0.1507668538594405,
+      "grad_norm": 0.17794294655323029,
+      "learning_rate": 0.00016197345559889107,
+      "loss": 0.6919,
+      "step": 5790
+    },
+    {
+      "epoch": 0.15102724566230655,
+      "grad_norm": 0.1671450436115265,
+      "learning_rate": 0.00016184234106710838,
+      "loss": 0.7048,
+      "step": 5800
+    },
+    {
+      "epoch": 0.1512876374651726,
+      "grad_norm": 0.16732144355773926,
+      "learning_rate": 0.00016171105415615134,
+      "loss": 0.7168,
+      "step": 5810
+    },
+    {
+      "epoch": 0.15154802926803865,
+      "grad_norm": 0.16594409942626953,
+      "learning_rate": 0.00016157959523196866,
+      "loss": 0.6942,
+      "step": 5820
+    },
+    {
+      "epoch": 0.1518084210709047,
+      "grad_norm": 0.173036128282547,
+      "learning_rate": 0.00016144796466098865,
+      "loss": 0.7093,
+      "step": 5830
+    },
+    {
+      "epoch": 0.15206881287377072,
+      "grad_norm": 0.17697438597679138,
+      "learning_rate": 0.00016131616281011798,
+      "loss": 0.7122,
+      "step": 5840
+    },
+    {
+      "epoch": 0.15232920467663677,
+      "grad_norm": 0.1677410751581192,
+      "learning_rate": 0.0001611841900467408,
+      "loss": 0.7027,
+      "step": 5850
+    },
+    {
+      "epoch": 0.15258959647950282,
+      "grad_norm": 0.16899007558822632,
+      "learning_rate": 0.0001610520467387176,
+      "loss": 0.6949,
+      "step": 5860
+    },
+    {
+      "epoch": 0.15284998828236887,
+      "grad_norm": 0.17043690383434296,
+      "learning_rate": 0.00016091973325438428,
+      "loss": 0.7052,
+      "step": 5870
+    },
+    {
+      "epoch": 0.15311038008523492,
+      "grad_norm": 0.17432864010334015,
+      "learning_rate": 0.00016078724996255114,
+      "loss": 0.7026,
+      "step": 5880
+    },
+    {
+      "epoch": 0.15337077188810097,
+      "grad_norm": 0.17059588432312012,
+      "learning_rate": 0.00016065459723250164,
+      "loss": 0.7234,
+      "step": 5890
+    },
+    {
+      "epoch": 0.15363116369096702,
+      "grad_norm": 0.16677695512771606,
+      "learning_rate": 0.0001605217754339918,
+      "loss": 0.7281,
+      "step": 5900
+    },
+    {
+      "epoch": 0.15389155549383304,
+      "grad_norm": 0.16831070184707642,
+      "learning_rate": 0.00016038878493724858,
+      "loss": 0.6934,
+      "step": 5910
+    },
+    {
+      "epoch": 0.1541519472966991,
+      "grad_norm": 0.17344878613948822,
+      "learning_rate": 0.00016025562611296946,
+      "loss": 0.7096,
+      "step": 5920
+    },
+    {
+      "epoch": 0.15441233909956514,
+      "grad_norm": 0.16978445649147034,
+      "learning_rate": 0.0001601222993323209,
+      "loss": 0.6851,
+      "step": 5930
+    },
+    {
+      "epoch": 0.1546727309024312,
+      "grad_norm": 0.17398668825626373,
+      "learning_rate": 0.00015998880496693766,
+      "loss": 0.7022,
+      "step": 5940
+    },
+    {
+      "epoch": 0.15493312270529724,
+      "grad_norm": 0.16338910162448883,
+      "learning_rate": 0.00015985514338892154,
+      "loss": 0.7257,
+      "step": 5950
+    },
+    {
+      "epoch": 0.1551935145081633,
+      "grad_norm": 0.1754075437784195,
+      "learning_rate": 0.0001597213149708405,
+      "loss": 0.7012,
+      "step": 5960
+    },
+    {
+      "epoch": 0.15545390631102932,
+      "grad_norm": 0.16998423635959625,
+      "learning_rate": 0.00015958732008572744,
+      "loss": 0.7313,
+      "step": 5970
+    },
+    {
+      "epoch": 0.15571429811389537,
+      "grad_norm": 0.18673604726791382,
+      "learning_rate": 0.00015945315910707945,
+      "loss": 0.6923,
+      "step": 5980
+    },
+    {
+      "epoch": 0.15597468991676142,
+      "grad_norm": 0.17040878534317017,
+      "learning_rate": 0.0001593188324088564,
+      "loss": 0.7009,
+      "step": 5990
+    },
+    {
+      "epoch": 0.15623508171962747,
+      "grad_norm": 0.16569504141807556,
+      "learning_rate": 0.00015918434036548017,
+      "loss": 0.6919,
+      "step": 6000
+    },
+    {
+      "epoch": 0.15649547352249352,
+      "grad_norm": 0.16722472012043,
+      "learning_rate": 0.00015904968335183354,
+      "loss": 0.7028,
+      "step": 6010
+    },
+    {
+      "epoch": 0.15675586532535957,
+      "grad_norm": 0.16823537647724152,
+      "learning_rate": 0.00015891486174325912,
+      "loss": 0.6822,
+      "step": 6020
+    },
+    {
+      "epoch": 0.15701625712822562,
+      "grad_norm": 0.1750400960445404,
+      "learning_rate": 0.00015877987591555822,
+      "loss": 0.7146,
+      "step": 6030
+    },
+    {
+      "epoch": 0.15727664893109164,
+      "grad_norm": 0.1703154593706131,
+      "learning_rate": 0.00015864472624499008,
+      "loss": 0.7057,
+      "step": 6040
+    },
+    {
+      "epoch": 0.1575370407339577,
+      "grad_norm": 0.17317266762256622,
+      "learning_rate": 0.00015850941310827045,
+      "loss": 0.692,
+      "step": 6050
+    },
+    {
+      "epoch": 0.15779743253682374,
+      "grad_norm": 0.18270671367645264,
+      "learning_rate": 0.00015837393688257083,
+      "loss": 0.7279,
+      "step": 6060
+    },
+    {
+      "epoch": 0.1580578243396898,
+      "grad_norm": 0.1647195667028427,
+      "learning_rate": 0.00015823829794551724,
+      "loss": 0.6999,
+      "step": 6070
+    },
+    {
+      "epoch": 0.15831821614255584,
+      "grad_norm": 0.17559175193309784,
+      "learning_rate": 0.00015810249667518936,
+      "loss": 0.6952,
+      "step": 6080
+    },
+    {
+      "epoch": 0.1585786079454219,
+      "grad_norm": 0.168562114238739,
+      "learning_rate": 0.00015796653345011928,
+      "loss": 0.7013,
+      "step": 6090
+    },
+    {
+      "epoch": 0.1588389997482879,
+      "grad_norm": 0.18079742789268494,
+      "learning_rate": 0.00015783040864929043,
+      "loss": 0.7115,
+      "step": 6100
+    },
+    {
+      "epoch": 0.15909939155115396,
+      "grad_norm": 0.17427009344100952,
+      "learning_rate": 0.00015769412265213684,
+      "loss": 0.6955,
+      "step": 6110
+    },
+    {
+      "epoch": 0.15935978335402,
+      "grad_norm": 0.18136094510555267,
+      "learning_rate": 0.00015755767583854163,
+      "loss": 0.7129,
+      "step": 6120
+    },
+    {
+      "epoch": 0.15962017515688606,
+      "grad_norm": 0.17283126711845398,
+      "learning_rate": 0.00015742106858883633,
+      "loss": 0.6993,
+      "step": 6130
+    },
+    {
+      "epoch": 0.1598805669597521,
+      "grad_norm": 0.16747362911701202,
+      "learning_rate": 0.00015728430128379963,
+      "loss": 0.718,
+      "step": 6140
+    },
+    {
+      "epoch": 0.16014095876261816,
+      "grad_norm": 0.17267820239067078,
+      "learning_rate": 0.0001571473743046563,
+      "loss": 0.6946,
+      "step": 6150
+    },
+    {
+      "epoch": 0.16040135056548419,
+      "grad_norm": 0.17288938164710999,
+      "learning_rate": 0.00015701028803307625,
+      "loss": 0.7031,
+      "step": 6160
+    },
+    {
+      "epoch": 0.16066174236835024,
+      "grad_norm": 0.16363868117332458,
+      "learning_rate": 0.0001568730428511734,
+      "loss": 0.7038,
+      "step": 6170
+    },
+    {
+      "epoch": 0.16092213417121629,
+      "grad_norm": 0.17383398115634918,
+      "learning_rate": 0.0001567356391415046,
+      "loss": 0.695,
+      "step": 6180
+    },
+    {
+      "epoch": 0.16118252597408234,
+      "grad_norm": 0.15880119800567627,
+      "learning_rate": 0.00015659807728706854,
+      "loss": 0.6977,
+      "step": 6190
+    },
+    {
+      "epoch": 0.16144291777694839,
+      "grad_norm": 0.1778557300567627,
+      "learning_rate": 0.0001564603576713048,
+      "loss": 0.7319,
+      "step": 6200
+    },
+    {
+      "epoch": 0.16170330957981444,
+      "grad_norm": 0.1838517040014267,
+      "learning_rate": 0.00015632248067809265,
+      "loss": 0.6999,
+      "step": 6210
+    },
+    {
+      "epoch": 0.16196370138268049,
+      "grad_norm": 0.1765550673007965,
+      "learning_rate": 0.00015618444669175004,
+      "loss": 0.689,
+      "step": 6220
+    },
+    {
+      "epoch": 0.1622240931855465,
+      "grad_norm": 0.17367658019065857,
+      "learning_rate": 0.00015604625609703257,
+      "loss": 0.7028,
+      "step": 6230
+    },
+    {
+      "epoch": 0.16248448498841256,
+      "grad_norm": 0.1650581955909729,
+      "learning_rate": 0.0001559079092791323,
+      "loss": 0.6926,
+      "step": 6240
+    },
+    {
+      "epoch": 0.1627448767912786,
+      "grad_norm": 0.16974525153636932,
+      "learning_rate": 0.00015576940662367675,
+      "loss": 0.711,
+      "step": 6250
+    },
+    {
+      "epoch": 0.16300526859414466,
+      "grad_norm": 0.17168691754341125,
+      "learning_rate": 0.00015563074851672787,
+      "loss": 0.6949,
+      "step": 6260
+    },
+    {
+      "epoch": 0.1632656603970107,
+      "grad_norm": 0.17578692734241486,
+      "learning_rate": 0.00015549193534478094,
+      "loss": 0.6858,
+      "step": 6270
+    },
+    {
+      "epoch": 0.16352605219987676,
+      "grad_norm": 0.17432482540607452,
+      "learning_rate": 0.00015535296749476335,
+      "loss": 0.7025,
+      "step": 6280
+    },
+    {
+      "epoch": 0.16378644400274278,
+      "grad_norm": 0.17605715990066528,
+      "learning_rate": 0.00015521384535403376,
+      "loss": 0.7021,
+      "step": 6290
+    },
+    {
+      "epoch": 0.16404683580560883,
+      "grad_norm": 0.17800843715667725,
+      "learning_rate": 0.0001550745693103808,
+      "loss": 0.6998,
+      "step": 6300
+    },
+    {
+      "epoch": 0.16430722760847488,
+      "grad_norm": 0.17110416293144226,
+      "learning_rate": 0.0001549351397520222,
+      "loss": 0.7116,
+      "step": 6310
+    },
+    {
+      "epoch": 0.16456761941134093,
+      "grad_norm": 0.1723811775445938,
+      "learning_rate": 0.00015479555706760344,
+      "loss": 0.72,
+      "step": 6320
+    },
+    {
+      "epoch": 0.16482801121420698,
+      "grad_norm": 0.1752457171678543,
+      "learning_rate": 0.00015465582164619703,
+      "loss": 0.6866,
+      "step": 6330
+    },
+    {
+      "epoch": 0.16508840301707303,
+      "grad_norm": 0.1849747896194458,
+      "learning_rate": 0.00015451593387730105,
+      "loss": 0.7288,
+      "step": 6340
+    },
+    {
+      "epoch": 0.16534879481993905,
+      "grad_norm": 0.16762009263038635,
+      "learning_rate": 0.00015437589415083828,
+      "loss": 0.6919,
+      "step": 6350
+    },
+    {
+      "epoch": 0.1656091866228051,
+      "grad_norm": 0.17073680460453033,
+      "learning_rate": 0.00015423570285715506,
+      "loss": 0.709,
+      "step": 6360
+    },
+    {
+      "epoch": 0.16586957842567115,
+      "grad_norm": 0.16971205174922943,
+      "learning_rate": 0.00015409536038702028,
+      "loss": 0.7156,
+      "step": 6370
+    },
+    {
+      "epoch": 0.1661299702285372,
+      "grad_norm": 0.17233216762542725,
+      "learning_rate": 0.00015395486713162414,
+      "loss": 0.7187,
+      "step": 6380
+    },
+    {
+      "epoch": 0.16639036203140325,
+      "grad_norm": 0.17737407982349396,
+      "learning_rate": 0.00015381422348257715,
+      "loss": 0.71,
+      "step": 6390
+    },
+    {
+      "epoch": 0.1666507538342693,
+      "grad_norm": 0.17228469252586365,
+      "learning_rate": 0.00015367342983190906,
+      "loss": 0.7224,
+      "step": 6400
+    },
+    {
+      "epoch": 0.16691114563713536,
+      "grad_norm": 0.17020554840564728,
+      "learning_rate": 0.00015353248657206764,
+      "loss": 0.7052,
+      "step": 6410
+    },
+    {
+      "epoch": 0.16717153744000138,
+      "grad_norm": 0.1692919284105301,
+      "learning_rate": 0.00015339139409591783,
+      "loss": 0.7035,
+      "step": 6420
+    },
+    {
+      "epoch": 0.16743192924286743,
+      "grad_norm": 0.17089873552322388,
+      "learning_rate": 0.00015325015279674036,
+      "loss": 0.7043,
+      "step": 6430
+    },
+    {
+      "epoch": 0.16769232104573348,
+      "grad_norm": 0.16704905033111572,
+      "learning_rate": 0.00015310876306823092,
+      "loss": 0.6924,
+      "step": 6440
+    },
+    {
+      "epoch": 0.16795271284859953,
+      "grad_norm": 0.1821179836988449,
+      "learning_rate": 0.00015296722530449877,
+      "loss": 0.6996,
+      "step": 6450
+    },
+    {
+      "epoch": 0.16821310465146558,
+      "grad_norm": 0.17202268540859222,
+      "learning_rate": 0.000152825539900066,
+      "loss": 0.7164,
+      "step": 6460
+    },
+    {
+      "epoch": 0.16847349645433163,
+      "grad_norm": 0.16377121210098267,
+      "learning_rate": 0.00015268370724986601,
+      "loss": 0.699,
+      "step": 6470
+    },
+    {
+      "epoch": 0.16873388825719765,
+      "grad_norm": 0.17642685770988464,
+      "learning_rate": 0.00015254172774924277,
+      "loss": 0.7127,
+      "step": 6480
+    },
+    {
+      "epoch": 0.1689942800600637,
+      "grad_norm": 0.1717827171087265,
+      "learning_rate": 0.0001523996017939496,
+      "loss": 0.6927,
+      "step": 6490
+    },
+    {
+      "epoch": 0.16925467186292975,
+      "grad_norm": 0.1642349660396576,
+      "learning_rate": 0.000152257329780148,
+      "loss": 0.704,
+      "step": 6500
+    },
+    {
+      "epoch": 0.1695150636657958,
+      "grad_norm": 0.17559029161930084,
+      "learning_rate": 0.00015211491210440653,
+      "loss": 0.7236,
+      "step": 6510
+    },
+    {
+      "epoch": 0.16977545546866185,
+      "grad_norm": 0.1805913746356964,
+      "learning_rate": 0.00015197234916369996,
+      "loss": 0.6829,
+      "step": 6520
+    },
+    {
+      "epoch": 0.1700358472715279,
+      "grad_norm": 0.17583122849464417,
+      "learning_rate": 0.00015182964135540772,
+      "loss": 0.7045,
+      "step": 6530
+    },
+    {
+      "epoch": 0.17029623907439395,
+      "grad_norm": 0.1662655770778656,
+      "learning_rate": 0.00015168678907731323,
+      "loss": 0.7173,
+      "step": 6540
+    },
+    {
+      "epoch": 0.17055663087725997,
+      "grad_norm": 0.1719600409269333,
+      "learning_rate": 0.00015154379272760252,
+      "loss": 0.7082,
+      "step": 6550
+    },
+    {
+      "epoch": 0.17081702268012602,
+      "grad_norm": 0.18137727677822113,
+      "learning_rate": 0.00015140065270486324,
+      "loss": 0.7314,
+      "step": 6560
+    },
+    {
+      "epoch": 0.17107741448299207,
+      "grad_norm": 0.17727130651474,
+      "learning_rate": 0.00015125736940808351,
+      "loss": 0.7093,
+      "step": 6570
+    },
+    {
+      "epoch": 0.17133780628585812,
+      "grad_norm": 0.19269149005413055,
+      "learning_rate": 0.00015111394323665081,
+      "loss": 0.6859,
+      "step": 6580
+    },
+    {
+      "epoch": 0.17159819808872417,
+      "grad_norm": 0.18232578039169312,
+      "learning_rate": 0.00015097037459035084,
+      "loss": 0.726,
+      "step": 6590
+    },
+    {
+      "epoch": 0.17185858989159022,
+      "grad_norm": 0.16427166759967804,
+      "learning_rate": 0.0001508266638693665,
+      "loss": 0.6896,
+      "step": 6600
+    },
+    {
+      "epoch": 0.17211898169445625,
+      "grad_norm": 0.17567217350006104,
+      "learning_rate": 0.00015068281147427657,
+      "loss": 0.6968,
+      "step": 6610
+    },
+    {
+      "epoch": 0.1723793734973223,
+      "grad_norm": 0.17443399131298065,
+      "learning_rate": 0.00015053881780605495,
+      "loss": 0.7034,
+      "step": 6620
+    },
+    {
+      "epoch": 0.17263976530018835,
+      "grad_norm": 0.16696178913116455,
+      "learning_rate": 0.00015039468326606911,
+      "loss": 0.6897,
+      "step": 6630
+    },
+    {
+      "epoch": 0.1729001571030544,
+      "grad_norm": 0.1779368370771408,
+      "learning_rate": 0.00015025040825607935,
+      "loss": 0.7036,
+      "step": 6640
+    },
+    {
+      "epoch": 0.17316054890592045,
+      "grad_norm": 0.17863190174102783,
+      "learning_rate": 0.00015010599317823734,
+      "loss": 0.7032,
+      "step": 6650
+    },
+    {
+      "epoch": 0.1734209407087865,
+      "grad_norm": 0.1763063222169876,
+      "learning_rate": 0.00014996143843508538,
+      "loss": 0.6825,
+      "step": 6660
+    },
+    {
+      "epoch": 0.17368133251165252,
+      "grad_norm": 0.17159639298915863,
+      "learning_rate": 0.00014981674442955488,
+      "loss": 0.6988,
+      "step": 6670
+    },
+    {
+      "epoch": 0.17394172431451857,
+      "grad_norm": 0.17756308615207672,
+      "learning_rate": 0.00014967191156496564,
+      "loss": 0.7018,
+      "step": 6680
+    },
+    {
+      "epoch": 0.17420211611738462,
+      "grad_norm": 0.17641721665859222,
+      "learning_rate": 0.00014952694024502429,
+      "loss": 0.7321,
+      "step": 6690
+    },
+    {
+      "epoch": 0.17446250792025067,
+      "grad_norm": 0.17100387811660767,
+      "learning_rate": 0.00014938183087382354,
+      "loss": 0.7099,
+      "step": 6700
+    },
+    {
+      "epoch": 0.17472289972311672,
+      "grad_norm": 0.16952122747898102,
+      "learning_rate": 0.00014923658385584084,
+      "loss": 0.6802,
+      "step": 6710
+    },
+    {
+      "epoch": 0.17498329152598277,
+      "grad_norm": 0.1642334908246994,
+      "learning_rate": 0.00014909119959593734,
+      "loss": 0.7083,
+      "step": 6720
+    },
+    {
+      "epoch": 0.17524368332884882,
+      "grad_norm": 0.16929323971271515,
+      "learning_rate": 0.00014894567849935673,
+      "loss": 0.7144,
+      "step": 6730
+    },
+    {
+      "epoch": 0.17550407513171484,
+      "grad_norm": 0.180099755525589,
+      "learning_rate": 0.00014880002097172418,
+      "loss": 0.7043,
+      "step": 6740
+    },
+    {
+      "epoch": 0.1757644669345809,
+      "grad_norm": 0.17287197709083557,
+      "learning_rate": 0.000148654227419045,
+      "loss": 0.6891,
+      "step": 6750
+    },
+    {
+      "epoch": 0.17602485873744694,
+      "grad_norm": 0.17458143830299377,
+      "learning_rate": 0.00014850829824770382,
+      "loss": 0.6889,
+      "step": 6760
+    },
+    {
+      "epoch": 0.176285250540313,
+      "grad_norm": 0.1722840964794159,
+      "learning_rate": 0.00014836223386446317,
+      "loss": 0.7058,
+      "step": 6770
+    },
+    {
+      "epoch": 0.17654564234317904,
+      "grad_norm": 0.17535988986492157,
+      "learning_rate": 0.0001482160346764625,
+      "loss": 0.7148,
+      "step": 6780
+    },
+    {
+      "epoch": 0.1768060341460451,
+      "grad_norm": 0.16451425850391388,
+      "learning_rate": 0.00014806970109121709,
+      "loss": 0.6975,
+      "step": 6790
+    },
+    {
+      "epoch": 0.17706642594891112,
+      "grad_norm": 0.16712118685245514,
+      "learning_rate": 0.00014792323351661671,
+      "loss": 0.7148,
+      "step": 6800
+    },
+    {
+      "epoch": 0.17732681775177717,
+      "grad_norm": 0.17245186865329742,
+      "learning_rate": 0.0001477766323609248,
+      "loss": 0.7216,
+      "step": 6810
+    },
+    {
+      "epoch": 0.17758720955464322,
+      "grad_norm": 0.17335280776023865,
+      "learning_rate": 0.00014762989803277686,
+      "loss": 0.6949,
+      "step": 6820
+    },
+    {
+      "epoch": 0.17784760135750927,
+      "grad_norm": 0.16834279894828796,
+      "learning_rate": 0.0001474830309411799,
+      "loss": 0.6884,
+      "step": 6830
+    },
+    {
+      "epoch": 0.17810799316037532,
+      "grad_norm": 0.17499734461307526,
+      "learning_rate": 0.0001473360314955108,
+      "loss": 0.7065,
+      "step": 6840
+    },
+    {
+      "epoch": 0.17836838496324137,
+      "grad_norm": 0.17072665691375732,
+      "learning_rate": 0.00014718890010551543,
+      "loss": 0.7067,
+      "step": 6850
+    },
+    {
+      "epoch": 0.17862877676610742,
+      "grad_norm": 0.192660853266716,
+      "learning_rate": 0.00014704163718130735,
+      "loss": 0.7263,
+      "step": 6860
+    },
+    {
+      "epoch": 0.17888916856897344,
+      "grad_norm": 0.18734219670295715,
+      "learning_rate": 0.00014689424313336696,
+      "loss": 0.7044,
+      "step": 6870
+    },
+    {
+      "epoch": 0.1791495603718395,
+      "grad_norm": 0.17117907106876373,
+      "learning_rate": 0.00014674671837253998,
+      "loss": 0.7112,
+      "step": 6880
+    },
+    {
+      "epoch": 0.17940995217470554,
+      "grad_norm": 0.18425500392913818,
+      "learning_rate": 0.00014659906331003652,
+      "loss": 0.6925,
+      "step": 6890
+    },
+    {
+      "epoch": 0.1796703439775716,
+      "grad_norm": 0.1706390231847763,
+      "learning_rate": 0.00014645127835742988,
+      "loss": 0.6916,
+      "step": 6900
+    },
+    {
+      "epoch": 0.17993073578043764,
+      "grad_norm": 0.17749357223510742,
+      "learning_rate": 0.00014630336392665552,
+      "loss": 0.7011,
+      "step": 6910
+    },
+    {
+      "epoch": 0.1801911275833037,
+      "grad_norm": 0.16545963287353516,
+      "learning_rate": 0.00014615532043000965,
+      "loss": 0.7163,
+      "step": 6920
+    },
+    {
+      "epoch": 0.1804515193861697,
+      "grad_norm": 0.1756933480501175,
+      "learning_rate": 0.00014600714828014833,
+      "loss": 0.6954,
+      "step": 6930
+    },
+    {
+      "epoch": 0.18071191118903576,
+      "grad_norm": 0.16752491891384125,
+      "learning_rate": 0.0001458588478900862,
+      "loss": 0.6902,
+      "step": 6940
+    },
+    {
+      "epoch": 0.1809723029919018,
+      "grad_norm": 0.16998566687107086,
+      "learning_rate": 0.00014571041967319535,
+      "loss": 0.6906,
+      "step": 6950
+    },
+    {
+      "epoch": 0.18123269479476786,
+      "grad_norm": 0.1744604855775833,
+      "learning_rate": 0.00014556186404320428,
+      "loss": 0.6814,
+      "step": 6960
+    },
+    {
+      "epoch": 0.1814930865976339,
+      "grad_norm": 0.17596711218357086,
+      "learning_rate": 0.00014541318141419646,
+      "loss": 0.7141,
+      "step": 6970
+    },
+    {
+      "epoch": 0.18175347840049996,
+      "grad_norm": 0.16803975403308868,
+      "learning_rate": 0.00014526437220060948,
+      "loss": 0.6875,
+      "step": 6980
+    },
+    {
+      "epoch": 0.18201387020336598,
+      "grad_norm": 0.1845843344926834,
+      "learning_rate": 0.00014511543681723377,
+      "loss": 0.6921,
+      "step": 6990
+    },
+    {
+      "epoch": 0.18227426200623204,
+      "grad_norm": 0.1783183515071869,
+      "learning_rate": 0.00014496637567921135,
+      "loss": 0.6982,
+      "step": 7000
+    },
+    {
+      "epoch": 0.18253465380909809,
+      "grad_norm": 0.17463266849517822,
+      "learning_rate": 0.00014481718920203492,
+      "loss": 0.7094,
+      "step": 7010
+    },
+    {
+      "epoch": 0.18279504561196414,
+      "grad_norm": 0.17779332399368286,
+      "learning_rate": 0.00014466787780154636,
+      "loss": 0.6989,
+      "step": 7020
+    },
+    {
+      "epoch": 0.18305543741483019,
+      "grad_norm": 0.17536398768424988,
+      "learning_rate": 0.0001445184418939359,
+      "loss": 0.7009,
+      "step": 7030
+    },
+    {
+      "epoch": 0.18331582921769624,
+      "grad_norm": 0.17731131613254547,
+      "learning_rate": 0.00014436888189574084,
+      "loss": 0.6964,
+      "step": 7040
+    },
+    {
+      "epoch": 0.18357622102056229,
+      "grad_norm": 0.18129530549049377,
+      "learning_rate": 0.00014421919822384428,
+      "loss": 0.6822,
+      "step": 7050
+    },
+    {
+      "epoch": 0.1838366128234283,
+      "grad_norm": 0.16552576422691345,
+      "learning_rate": 0.00014406939129547406,
+      "loss": 0.7172,
+      "step": 7060
+    },
+    {
+      "epoch": 0.18409700462629436,
+      "grad_norm": 0.17690911889076233,
+      "learning_rate": 0.00014391946152820167,
+      "loss": 0.6883,
+      "step": 7070
+    },
+    {
+      "epoch": 0.1843573964291604,
+      "grad_norm": 0.1851797252893448,
+      "learning_rate": 0.00014376940933994082,
+      "loss": 0.6988,
+      "step": 7080
+    },
+    {
+      "epoch": 0.18461778823202646,
+      "grad_norm": 0.18322911858558655,
+      "learning_rate": 0.0001436192351489467,
+      "loss": 0.689,
+      "step": 7090
+    },
+    {
+      "epoch": 0.1848781800348925,
+      "grad_norm": 0.17606942355632782,
+      "learning_rate": 0.0001434689393738144,
+      "loss": 0.7034,
+      "step": 7100
+    },
+    {
+      "epoch": 0.18513857183775856,
+      "grad_norm": 0.1715836524963379,
+      "learning_rate": 0.0001433185224334779,
+      "loss": 0.6939,
+      "step": 7110
+    },
+    {
+      "epoch": 0.18539896364062458,
+      "grad_norm": 0.18052712082862854,
+      "learning_rate": 0.00014316798474720908,
+      "loss": 0.6858,
+      "step": 7120
+    },
+    {
+      "epoch": 0.18565935544349063,
+      "grad_norm": 0.17608265578746796,
+      "learning_rate": 0.00014301732673461617,
+      "loss": 0.7166,
+      "step": 7130
+    },
+    {
+      "epoch": 0.18591974724635668,
+      "grad_norm": 0.17561575770378113,
+      "learning_rate": 0.000142866548815643,
+      "loss": 0.7208,
+      "step": 7140
+    },
+    {
+      "epoch": 0.18618013904922273,
+      "grad_norm": 0.17250195145606995,
+      "learning_rate": 0.00014271565141056746,
+      "loss": 0.687,
+      "step": 7150
+    },
+    {
+      "epoch": 0.18644053085208878,
+      "grad_norm": 0.1769152581691742,
+      "learning_rate": 0.00014256463494000064,
+      "loss": 0.7007,
+      "step": 7160
+    },
+    {
+      "epoch": 0.18670092265495483,
+      "grad_norm": 0.16784866154193878,
+      "learning_rate": 0.0001424134998248854,
+      "loss": 0.6913,
+      "step": 7170
+    },
+    {
+      "epoch": 0.18696131445782085,
+      "grad_norm": 0.18374022841453552,
+      "learning_rate": 0.00014226224648649532,
+      "loss": 0.7048,
+      "step": 7180
+    },
+    {
+      "epoch": 0.1872217062606869,
+      "grad_norm": 0.173793226480484,
+      "learning_rate": 0.0001421108753464336,
+      "loss": 0.6877,
+      "step": 7190
+    },
+    {
+      "epoch": 0.18748209806355295,
+      "grad_norm": 0.17282545566558838,
+      "learning_rate": 0.00014195938682663175,
+      "loss": 0.7132,
+      "step": 7200
+    },
+    {
+      "epoch": 0.187742489866419,
+      "grad_norm": 0.16977694630622864,
+      "learning_rate": 0.00014180778134934843,
+      "loss": 0.6813,
+      "step": 7210
+    },
+    {
+      "epoch": 0.18800288166928505,
+      "grad_norm": 0.18771855533123016,
+      "learning_rate": 0.00014165605933716836,
+      "loss": 0.7037,
+      "step": 7220
+    },
+    {
+      "epoch": 0.1882632734721511,
+      "grad_norm": 0.17175500094890594,
+      "learning_rate": 0.00014150422121300105,
+      "loss": 0.6926,
+      "step": 7230
+    },
+    {
+      "epoch": 0.18852366527501715,
+      "grad_norm": 0.17340902984142303,
+      "learning_rate": 0.00014135226740007967,
+      "loss": 0.6987,
+      "step": 7240
+    },
+    {
+      "epoch": 0.18878405707788318,
+      "grad_norm": 0.16895218193531036,
+      "learning_rate": 0.00014120019832195986,
+      "loss": 0.6771,
+      "step": 7250
+    },
+    {
+      "epoch": 0.18904444888074923,
+      "grad_norm": 0.17422990500926971,
+      "learning_rate": 0.00014104801440251863,
+      "loss": 0.7027,
+      "step": 7260
+    },
+    {
+      "epoch": 0.18930484068361528,
+      "grad_norm": 0.1705746352672577,
+      "learning_rate": 0.00014089571606595294,
+      "loss": 0.6907,
+      "step": 7270
+    },
+    {
+      "epoch": 0.18956523248648133,
+      "grad_norm": 0.16771703958511353,
+      "learning_rate": 0.0001407433037367788,
+      "loss": 0.7105,
+      "step": 7280
+    },
+    {
+      "epoch": 0.18982562428934738,
+      "grad_norm": 0.1731031835079193,
+      "learning_rate": 0.00014059077783982993,
+      "loss": 0.6683,
+      "step": 7290
+    },
+    {
+      "epoch": 0.19008601609221343,
+      "grad_norm": 0.18010441958904266,
+      "learning_rate": 0.00014043813880025658,
+      "loss": 0.692,
+      "step": 7300
+    },
+    {
+      "epoch": 0.19034640789507945,
+      "grad_norm": 0.17735110223293304,
+      "learning_rate": 0.0001402853870435244,
+      "loss": 0.6639,
+      "step": 7310
+    },
+    {
+      "epoch": 0.1906067996979455,
+      "grad_norm": 0.16834424436092377,
+      "learning_rate": 0.00014013252299541323,
+      "loss": 0.6915,
+      "step": 7320
+    },
+    {
+      "epoch": 0.19086719150081155,
+      "grad_norm": 0.17520864307880402,
+      "learning_rate": 0.00013997954708201593,
+      "loss": 0.6907,
+      "step": 7330
+    },
+    {
+      "epoch": 0.1911275833036776,
+      "grad_norm": 0.18474625051021576,
+      "learning_rate": 0.00013982645972973715,
+      "loss": 0.7003,
+      "step": 7340
+    },
+    {
+      "epoch": 0.19138797510654365,
+      "grad_norm": 0.1774008870124817,
+      "learning_rate": 0.00013967326136529208,
+      "loss": 0.7133,
+      "step": 7350
+    },
+    {
+      "epoch": 0.1916483669094097,
+      "grad_norm": 0.17432044446468353,
+      "learning_rate": 0.00013951995241570552,
+      "loss": 0.7072,
+      "step": 7360
+    },
+    {
+      "epoch": 0.19190875871227575,
+      "grad_norm": 0.1843249648809433,
+      "learning_rate": 0.0001393665333083104,
+      "loss": 0.6928,
+      "step": 7370
+    },
+    {
+      "epoch": 0.19216915051514177,
+      "grad_norm": 0.17402757704257965,
+      "learning_rate": 0.00013921300447074667,
+      "loss": 0.7106,
+      "step": 7380
+    },
+    {
+      "epoch": 0.19242954231800782,
+      "grad_norm": 0.17242158949375153,
+      "learning_rate": 0.00013905936633096025,
+      "loss": 0.6903,
+      "step": 7390
+    },
+    {
+      "epoch": 0.19268993412087387,
+      "grad_norm": 0.18488599359989166,
+      "learning_rate": 0.00013890561931720167,
+      "loss": 0.688,
+      "step": 7400
+    },
+    {
+      "epoch": 0.19295032592373992,
+      "grad_norm": 0.17597128450870514,
+      "learning_rate": 0.0001387517638580249,
+      "loss": 0.7062,
+      "step": 7410
+    },
+    {
+      "epoch": 0.19321071772660597,
+      "grad_norm": 0.1805078685283661,
+      "learning_rate": 0.00013859780038228634,
+      "loss": 0.6999,
+      "step": 7420
+    },
+    {
+      "epoch": 0.19347110952947202,
+      "grad_norm": 0.17452898621559143,
+      "learning_rate": 0.00013844372931914325,
+      "loss": 0.6961,
+      "step": 7430
+    },
+    {
+      "epoch": 0.19373150133233805,
+      "grad_norm": 0.17718560993671417,
+      "learning_rate": 0.00013828955109805295,
+      "loss": 0.6909,
+      "step": 7440
+    },
+    {
+      "epoch": 0.1939918931352041,
+      "grad_norm": 0.1740763783454895,
+      "learning_rate": 0.00013813526614877138,
+      "loss": 0.7126,
+      "step": 7450
+    },
+    {
+      "epoch": 0.19425228493807015,
+      "grad_norm": 0.17878217995166779,
+      "learning_rate": 0.000137980874901352,
+      "loss": 0.7,
+      "step": 7460
+    },
+    {
+      "epoch": 0.1945126767409362,
+      "grad_norm": 0.17267367243766785,
+      "learning_rate": 0.0001378263777861445,
+      "loss": 0.6756,
+      "step": 7470
+    },
+    {
+      "epoch": 0.19477306854380225,
+      "grad_norm": 0.174414724111557,
+      "learning_rate": 0.00013767177523379382,
+      "loss": 0.6948,
+      "step": 7480
+    },
+    {
+      "epoch": 0.1950334603466683,
+      "grad_norm": 0.17386887967586517,
+      "learning_rate": 0.00013751706767523865,
+      "loss": 0.6848,
+      "step": 7490
+    },
+    {
+      "epoch": 0.19529385214953432,
+      "grad_norm": 0.1732412576675415,
+      "learning_rate": 0.0001373622555417104,
+      "loss": 0.6914,
+      "step": 7500
+    },
+    {
+      "epoch": 0.19555424395240037,
+      "grad_norm": 0.17556026577949524,
+      "learning_rate": 0.00013720733926473206,
+      "loss": 0.6866,
+      "step": 7510
+    },
+    {
+      "epoch": 0.19581463575526642,
+      "grad_norm": 0.1789248138666153,
+      "learning_rate": 0.00013705231927611675,
+      "loss": 0.7071,
+      "step": 7520
+    },
+    {
+      "epoch": 0.19607502755813247,
+      "grad_norm": 0.17923463881015778,
+      "learning_rate": 0.00013689719600796683,
+      "loss": 0.684,
+      "step": 7530
+    },
+    {
+      "epoch": 0.19633541936099852,
+      "grad_norm": 0.18686242401599884,
+      "learning_rate": 0.00013674196989267245,
+      "loss": 0.7003,
+      "step": 7540
+    },
+    {
+      "epoch": 0.19659581116386457,
+      "grad_norm": 0.1743983030319214,
+      "learning_rate": 0.00013658664136291046,
+      "loss": 0.7067,
+      "step": 7550
+    },
+    {
+      "epoch": 0.19685620296673062,
+      "grad_norm": 0.17865756154060364,
+      "learning_rate": 0.0001364312108516432,
+      "loss": 0.7005,
+      "step": 7560
+    },
+    {
+      "epoch": 0.19711659476959664,
+      "grad_norm": 0.17367780208587646,
+      "learning_rate": 0.0001362756787921173,
+      "loss": 0.6852,
+      "step": 7570
+    },
+    {
+      "epoch": 0.1973769865724627,
+      "grad_norm": 0.17140452563762665,
+      "learning_rate": 0.00013612004561786232,
+      "loss": 0.6904,
+      "step": 7580
+    },
+    {
+      "epoch": 0.19763737837532874,
+      "grad_norm": 0.17559275031089783,
+      "learning_rate": 0.00013596431176268978,
+      "loss": 0.7174,
+      "step": 7590
+    },
+    {
+      "epoch": 0.1978977701781948,
+      "grad_norm": 0.17835845053195953,
+      "learning_rate": 0.00013580847766069183,
+      "loss": 0.6827,
+      "step": 7600
+    },
+    {
+      "epoch": 0.19815816198106084,
+      "grad_norm": 0.1730433702468872,
+      "learning_rate": 0.00013565254374624,
+      "loss": 0.7099,
+      "step": 7610
+    },
+    {
+      "epoch": 0.1984185537839269,
+      "grad_norm": 0.1799464374780655,
+      "learning_rate": 0.00013549651045398407,
+      "loss": 0.6777,
+      "step": 7620
+    },
+    {
+      "epoch": 0.19867894558679292,
+      "grad_norm": 0.1781192421913147,
+      "learning_rate": 0.00013534037821885084,
+      "loss": 0.7009,
+      "step": 7630
+    },
+    {
+      "epoch": 0.19893933738965897,
+      "grad_norm": 0.17447307705879211,
+      "learning_rate": 0.00013518414747604285,
+      "loss": 0.7016,
+      "step": 7640
+    },
+    {
+      "epoch": 0.19919972919252502,
+      "grad_norm": 0.17640911042690277,
+      "learning_rate": 0.00013502781866103726,
+      "loss": 0.6878,
+      "step": 7650
+    },
+    {
+      "epoch": 0.19946012099539107,
+      "grad_norm": 0.17320814728736877,
+      "learning_rate": 0.00013487139220958458,
+      "loss": 0.6966,
+      "step": 7660
+    },
+    {
+      "epoch": 0.19972051279825712,
+      "grad_norm": 0.17917166650295258,
+      "learning_rate": 0.0001347148685577075,
+      "loss": 0.6858,
+      "step": 7670
+    },
+    {
+      "epoch": 0.19998090460112317,
+      "grad_norm": 0.1751902550458908,
+      "learning_rate": 0.00013455824814169958,
+      "loss": 0.7011,
+      "step": 7680
+    },
+    {
+      "epoch": 0.20024129640398922,
+      "grad_norm": 0.18323281407356262,
+      "learning_rate": 0.00013440153139812418,
+      "loss": 0.6794,
+      "step": 7690
+    },
+    {
+      "epoch": 0.20050168820685524,
+      "grad_norm": 0.1677102893590927,
+      "learning_rate": 0.00013424471876381311,
+      "loss": 0.6791,
+      "step": 7700
+    },
+    {
+      "epoch": 0.2007620800097213,
+      "grad_norm": 0.17628876864910126,
+      "learning_rate": 0.0001340878106758655,
+      "loss": 0.6987,
+      "step": 7710
+    },
+    {
+      "epoch": 0.20102247181258734,
+      "grad_norm": 0.1839548945426941,
+      "learning_rate": 0.00013393080757164653,
+      "loss": 0.6971,
+      "step": 7720
+    },
+    {
+      "epoch": 0.2012828636154534,
+      "grad_norm": 0.1783684343099594,
+      "learning_rate": 0.00013377370988878624,
+      "loss": 0.6881,
+      "step": 7730
+    },
+    {
+      "epoch": 0.20154325541831944,
+      "grad_norm": 0.17916792631149292,
+      "learning_rate": 0.00013361651806517824,
+      "loss": 0.698,
+      "step": 7740
+    },
+    {
+      "epoch": 0.2018036472211855,
+      "grad_norm": 0.17960959672927856,
+      "learning_rate": 0.00013345923253897866,
+      "loss": 0.6858,
+      "step": 7750
+    },
+    {
+      "epoch": 0.2020640390240515,
+      "grad_norm": 0.17584697902202606,
+      "learning_rate": 0.00013330185374860467,
+      "loss": 0.6835,
+      "step": 7760
+    },
+    {
+      "epoch": 0.20232443082691756,
+      "grad_norm": 0.17761898040771484,
+      "learning_rate": 0.00013314438213273355,
+      "loss": 0.701,
+      "step": 7770
+    },
+    {
+      "epoch": 0.2025848226297836,
+      "grad_norm": 0.16848579049110413,
+      "learning_rate": 0.00013298681813030122,
+      "loss": 0.7162,
+      "step": 7780
+    },
+    {
+      "epoch": 0.20284521443264966,
+      "grad_norm": 0.17633990943431854,
+      "learning_rate": 0.00013282916218050118,
+      "loss": 0.7054,
+      "step": 7790
+    },
+    {
+      "epoch": 0.2031056062355157,
+      "grad_norm": 0.1785901039838791,
+      "learning_rate": 0.00013267141472278316,
+      "loss": 0.7137,
+      "step": 7800
+    },
+    {
+      "epoch": 0.20336599803838176,
+      "grad_norm": 0.16447508335113525,
+      "learning_rate": 0.00013251357619685207,
+      "loss": 0.7167,
+      "step": 7810
+    },
+    {
+      "epoch": 0.20362638984124778,
+      "grad_norm": 0.17035554349422455,
+      "learning_rate": 0.00013235564704266653,
+      "loss": 0.6956,
+      "step": 7820
+    },
+    {
+      "epoch": 0.20388678164411383,
+      "grad_norm": 0.18053407967090607,
+      "learning_rate": 0.00013219762770043787,
+      "loss": 0.7016,
+      "step": 7830
+    },
+    {
+      "epoch": 0.20414717344697988,
+      "grad_norm": 0.18030588328838348,
+      "learning_rate": 0.00013203951861062873,
+      "loss": 0.7082,
+      "step": 7840
+    },
+    {
+      "epoch": 0.20440756524984594,
+      "grad_norm": 0.17645570635795593,
+      "learning_rate": 0.00013188132021395204,
+      "loss": 0.6894,
+      "step": 7850
+    },
+    {
+      "epoch": 0.20466795705271199,
+      "grad_norm": 0.1644545942544937,
+      "learning_rate": 0.00013172303295136948,
+      "loss": 0.6944,
+      "step": 7860
+    },
+    {
+      "epoch": 0.20492834885557804,
+      "grad_norm": 0.1747378408908844,
+      "learning_rate": 0.0001315646572640906,
+      "loss": 0.7009,
+      "step": 7870
+    },
+    {
+      "epoch": 0.20518874065844409,
+      "grad_norm": 0.17992566525936127,
+      "learning_rate": 0.00013140619359357139,
+      "loss": 0.7139,
+      "step": 7880
+    },
+    {
+      "epoch": 0.2054491324613101,
+      "grad_norm": 0.17254693806171417,
+      "learning_rate": 0.000131247642381513,
+      "loss": 0.6878,
+      "step": 7890
+    },
+    {
+      "epoch": 0.20570952426417616,
+      "grad_norm": 0.19822190701961517,
+      "learning_rate": 0.00013108900406986066,
+      "loss": 0.7181,
+      "step": 7900
+    },
+    {
+      "epoch": 0.2059699160670422,
+      "grad_norm": 0.18208132684230804,
+      "learning_rate": 0.0001309302791008024,
+      "loss": 0.7086,
+      "step": 7910
+    },
+    {
+      "epoch": 0.20623030786990826,
+      "grad_norm": 0.1721629798412323,
+      "learning_rate": 0.00013077146791676776,
+      "loss": 0.6874,
+      "step": 7920
+    },
+    {
+      "epoch": 0.2064906996727743,
+      "grad_norm": 0.1776220202445984,
+      "learning_rate": 0.00013061257096042654,
+      "loss": 0.6997,
+      "step": 7930
+    },
+    {
+      "epoch": 0.20675109147564036,
+      "grad_norm": 0.1767396479845047,
+      "learning_rate": 0.00013045358867468782,
+      "loss": 0.6839,
+      "step": 7940
+    },
+    {
+      "epoch": 0.20701148327850638,
+      "grad_norm": 0.17351138591766357,
+      "learning_rate": 0.00013029452150269827,
+      "loss": 0.7025,
+      "step": 7950
+    },
+    {
+      "epoch": 0.20727187508137243,
+      "grad_norm": 0.18614953756332397,
+      "learning_rate": 0.00013013536988784138,
+      "loss": 0.7054,
+      "step": 7960
+    },
+    {
+      "epoch": 0.20753226688423848,
+      "grad_norm": 0.17871534824371338,
+      "learning_rate": 0.00012997613427373586,
+      "loss": 0.6855,
+      "step": 7970
+    },
+    {
+      "epoch": 0.20779265868710453,
+      "grad_norm": 0.17882663011550903,
+      "learning_rate": 0.00012981681510423472,
+      "loss": 0.6973,
+      "step": 7980
+    },
+    {
+      "epoch": 0.20805305048997058,
+      "grad_norm": 0.17025555670261383,
+      "learning_rate": 0.0001296574128234237,
+      "loss": 0.6915,
+      "step": 7990
+    },
+    {
+      "epoch": 0.20831344229283663,
+      "grad_norm": 0.18960419297218323,
+      "learning_rate": 0.0001294979278756203,
+      "loss": 0.6961,
+      "step": 8000
+    },
+    {
+      "epoch": 0.20857383409570265,
+      "grad_norm": 0.18628227710723877,
+      "learning_rate": 0.00012933836070537247,
+      "loss": 0.6676,
+      "step": 8010
+    },
+    {
+      "epoch": 0.2088342258985687,
+      "grad_norm": 0.17117147147655487,
+      "learning_rate": 0.00012917871175745726,
+      "loss": 0.6929,
+      "step": 8020
+    },
+    {
+      "epoch": 0.20909461770143475,
+      "grad_norm": 0.17737753689289093,
+      "learning_rate": 0.00012901898147687975,
+      "loss": 0.7125,
+      "step": 8030
+    },
+    {
+      "epoch": 0.2093550095043008,
+      "grad_norm": 0.17105183005332947,
+      "learning_rate": 0.00012885917030887174,
+      "loss": 0.6709,
+      "step": 8040
+    },
+    {
+      "epoch": 0.20961540130716685,
+      "grad_norm": 0.17960920929908752,
+      "learning_rate": 0.00012869927869889032,
+      "loss": 0.6898,
+      "step": 8050
+    },
+    {
+      "epoch": 0.2098757931100329,
+      "grad_norm": 0.1776961088180542,
+      "learning_rate": 0.00012853930709261707,
+      "loss": 0.6987,
+      "step": 8060
+    },
+    {
+      "epoch": 0.21013618491289895,
+      "grad_norm": 0.16933076083660126,
+      "learning_rate": 0.00012837925593595626,
+      "loss": 0.6793,
+      "step": 8070
+    },
+    {
+      "epoch": 0.21039657671576498,
+      "grad_norm": 0.17447830736637115,
+      "learning_rate": 0.0001282191256750342,
+      "loss": 0.69,
+      "step": 8080
+    },
+    {
+      "epoch": 0.21065696851863103,
+      "grad_norm": 0.17908291518688202,
+      "learning_rate": 0.00012805891675619747,
+      "loss": 0.6805,
+      "step": 8090
+    },
+    {
+      "epoch": 0.21091736032149708,
+      "grad_norm": 0.17689751088619232,
+      "learning_rate": 0.00012789862962601201,
+      "loss": 0.6993,
+      "step": 8100
+    },
+    {
+      "epoch": 0.21117775212436313,
+      "grad_norm": 0.17400488257408142,
+      "learning_rate": 0.00012773826473126172,
+      "loss": 0.6836,
+      "step": 8110
+    },
+    {
+      "epoch": 0.21143814392722918,
+      "grad_norm": 0.17445240914821625,
+      "learning_rate": 0.00012757782251894727,
+      "loss": 0.7105,
+      "step": 8120
+    },
+    {
+      "epoch": 0.21169853573009523,
+      "grad_norm": 0.18754400312900543,
+      "learning_rate": 0.00012741730343628484,
+      "loss": 0.7093,
+      "step": 8130
+    },
+    {
+      "epoch": 0.21195892753296125,
+      "grad_norm": 0.1769934892654419,
+      "learning_rate": 0.0001272567079307049,
+      "loss": 0.6922,
+      "step": 8140
+    },
+    {
+      "epoch": 0.2122193193358273,
+      "grad_norm": 0.17582187056541443,
+      "learning_rate": 0.00012709603644985094,
+      "loss": 0.6955,
+      "step": 8150
+    },
+    {
+      "epoch": 0.21247971113869335,
+      "grad_norm": 0.1896626353263855,
+      "learning_rate": 0.00012693528944157818,
+      "loss": 0.706,
+      "step": 8160
+    },
+    {
+      "epoch": 0.2127401029415594,
+      "grad_norm": 0.17897091805934906,
+      "learning_rate": 0.0001267744673539524,
+      "loss": 0.6903,
+      "step": 8170
+    },
+    {
+      "epoch": 0.21300049474442545,
+      "grad_norm": 0.1837705820798874,
+      "learning_rate": 0.0001266135706352487,
+      "loss": 0.6972,
+      "step": 8180
+    },
+    {
+      "epoch": 0.2132608865472915,
+      "grad_norm": 0.17943723499774933,
+      "learning_rate": 0.0001264525997339501,
+      "loss": 0.6788,
+      "step": 8190
+    },
+    {
+      "epoch": 0.21352127835015755,
+      "grad_norm": 0.18042807281017303,
+      "learning_rate": 0.0001262915550987465,
+      "loss": 0.7139,
+      "step": 8200
+    },
+    {
+      "epoch": 0.21378167015302357,
+      "grad_norm": 0.1821472942829132,
+      "learning_rate": 0.00012613043717853324,
+      "loss": 0.6902,
+      "step": 8210
+    },
+    {
+      "epoch": 0.21404206195588962,
+      "grad_norm": 0.18218253552913666,
+      "learning_rate": 0.00012596924642241004,
+      "loss": 0.6861,
+      "step": 8220
+    },
+    {
+      "epoch": 0.21430245375875567,
+      "grad_norm": 0.18331663310527802,
+      "learning_rate": 0.00012580798327967952,
+      "loss": 0.6746,
+      "step": 8230
+    },
+    {
+      "epoch": 0.21456284556162172,
+      "grad_norm": 0.1842353492975235,
+      "learning_rate": 0.0001256466481998461,
+      "loss": 0.6932,
+      "step": 8240
+    },
+    {
+      "epoch": 0.21482323736448777,
+      "grad_norm": 0.16256186366081238,
+      "learning_rate": 0.00012548524163261485,
+      "loss": 0.6843,
+      "step": 8250
+    },
+    {
+      "epoch": 0.21508362916735382,
+      "grad_norm": 0.17319098114967346,
+      "learning_rate": 0.00012532376402788988,
+      "loss": 0.7022,
+      "step": 8260
+    },
+    {
+      "epoch": 0.21534402097021985,
+      "grad_norm": 0.1779172569513321,
+      "learning_rate": 0.00012516221583577355,
+      "loss": 0.7039,
+      "step": 8270
+    },
+    {
+      "epoch": 0.2156044127730859,
+      "grad_norm": 0.17087218165397644,
+      "learning_rate": 0.00012500059750656474,
+      "loss": 0.6917,
+      "step": 8280
+    },
+    {
+      "epoch": 0.21586480457595195,
+      "grad_norm": 0.16772834956645966,
+      "learning_rate": 0.00012483890949075801,
+      "loss": 0.6848,
+      "step": 8290
+    },
+    {
+      "epoch": 0.216125196378818,
+      "grad_norm": 0.17318886518478394,
+      "learning_rate": 0.0001246771522390421,
+      "loss": 0.6973,
+      "step": 8300
+    },
+    {
+      "epoch": 0.21638558818168405,
+      "grad_norm": 0.16938996315002441,
+      "learning_rate": 0.00012451532620229872,
+      "loss": 0.7016,
+      "step": 8310
+    },
+    {
+      "epoch": 0.2166459799845501,
+      "grad_norm": 0.17802828550338745,
+      "learning_rate": 0.0001243534318316013,
+      "loss": 0.6871,
+      "step": 8320
+    },
+    {
+      "epoch": 0.21690637178741612,
+      "grad_norm": 0.17595522105693817,
+      "learning_rate": 0.00012419146957821383,
+      "loss": 0.7055,
+      "step": 8330
+    },
+    {
+      "epoch": 0.21716676359028217,
+      "grad_norm": 0.1841190904378891,
+      "learning_rate": 0.0001240294398935894,
+      "loss": 0.7154,
+      "step": 8340
+    },
+    {
+      "epoch": 0.21742715539314822,
+      "grad_norm": 0.16829776763916016,
+      "learning_rate": 0.0001238673432293692,
+      "loss": 0.6925,
+      "step": 8350
+    },
+    {
+      "epoch": 0.21768754719601427,
+      "grad_norm": 0.1729755699634552,
+      "learning_rate": 0.00012370518003738091,
+      "loss": 0.7118,
+      "step": 8360
+    },
+    {
+      "epoch": 0.21794793899888032,
+      "grad_norm": 0.16530723869800568,
+      "learning_rate": 0.00012354295076963786,
+      "loss": 0.7029,
+      "step": 8370
+    },
+    {
+      "epoch": 0.21820833080174637,
+      "grad_norm": 0.17311285436153412,
+      "learning_rate": 0.00012338065587833748,
+      "loss": 0.7013,
+      "step": 8380
+    },
+    {
+      "epoch": 0.21846872260461242,
+      "grad_norm": 0.1771146059036255,
+      "learning_rate": 0.00012321829581586003,
+      "loss": 0.6997,
+      "step": 8390
+    },
+    {
+      "epoch": 0.21872911440747844,
+      "grad_norm": 0.1830943077802658,
+      "learning_rate": 0.0001230558710347676,
+      "loss": 0.6945,
+      "step": 8400
+    },
+    {
+      "epoch": 0.2189895062103445,
+      "grad_norm": 0.1760828197002411,
+      "learning_rate": 0.00012289338198780253,
+      "loss": 0.6888,
+      "step": 8410
+    },
+    {
+      "epoch": 0.21924989801321054,
+      "grad_norm": 0.1744048148393631,
+      "learning_rate": 0.00012273082912788637,
+      "loss": 0.676,
+      "step": 8420
+    },
+    {
+      "epoch": 0.2195102898160766,
+      "grad_norm": 0.17182762920856476,
+      "learning_rate": 0.00012256821290811855,
+      "loss": 0.707,
+      "step": 8430
+    },
+    {
+      "epoch": 0.21977068161894264,
+      "grad_norm": 0.1801503300666809,
+      "learning_rate": 0.00012240553378177504,
+      "loss": 0.6907,
+      "step": 8440
+    },
+    {
+      "epoch": 0.2200310734218087,
+      "grad_norm": 0.1753336787223816,
+      "learning_rate": 0.0001222427922023072,
+      "loss": 0.7034,
+      "step": 8450
+    },
+    {
+      "epoch": 0.22029146522467472,
+      "grad_norm": 0.180914506316185,
+      "learning_rate": 0.00012207998862334048,
+      "loss": 0.6773,
+      "step": 8460
+    },
+    {
+      "epoch": 0.22055185702754077,
+      "grad_norm": 0.17284120619297028,
+      "learning_rate": 0.00012191712349867315,
+      "loss": 0.6983,
+      "step": 8470
+    },
+    {
+      "epoch": 0.22081224883040682,
+      "grad_norm": 0.18915420770645142,
+      "learning_rate": 0.00012175419728227498,
+      "loss": 0.7106,
+      "step": 8480
+    },
+    {
+      "epoch": 0.22107264063327287,
+      "grad_norm": 0.178519144654274,
+      "learning_rate": 0.00012159121042828612,
+      "loss": 0.7054,
+      "step": 8490
+    },
+    {
+      "epoch": 0.22133303243613892,
+      "grad_norm": 0.17323723435401917,
+      "learning_rate": 0.00012142816339101562,
+      "loss": 0.6965,
+      "step": 8500
+    },
+    {
+      "epoch": 0.22159342423900497,
+      "grad_norm": 0.23265057802200317,
+      "learning_rate": 0.00012126505662494041,
+      "loss": 0.6914,
+      "step": 8510
+    },
+    {
+      "epoch": 0.221853816041871,
+      "grad_norm": 0.1790781468153,
+      "learning_rate": 0.00012110189058470382,
+      "loss": 0.6911,
+      "step": 8520
+    },
+    {
+      "epoch": 0.22211420784473704,
+      "grad_norm": 0.17768746614456177,
+      "learning_rate": 0.00012093866572511443,
+      "loss": 0.6829,
+      "step": 8530
+    },
+    {
+      "epoch": 0.2223745996476031,
+      "grad_norm": 0.17421717941761017,
+      "learning_rate": 0.00012077538250114478,
+      "loss": 0.6963,
+      "step": 8540
+    },
+    {
+      "epoch": 0.22263499145046914,
+      "grad_norm": 0.1880684494972229,
+      "learning_rate": 0.0001206120413679301,
+      "loss": 0.6946,
+      "step": 8550
+    },
+    {
+      "epoch": 0.2228953832533352,
+      "grad_norm": 0.17597487568855286,
+      "learning_rate": 0.00012044864278076698,
+      "loss": 0.7016,
+      "step": 8560
+    },
+    {
+      "epoch": 0.22315577505620124,
+      "grad_norm": 0.17353573441505432,
+      "learning_rate": 0.00012028518719511225,
+      "loss": 0.6771,
+      "step": 8570
+    },
+    {
+      "epoch": 0.2234161668590673,
+      "grad_norm": 0.18333446979522705,
+      "learning_rate": 0.00012012167506658156,
+      "loss": 0.6696,
+      "step": 8580
+    },
+    {
+      "epoch": 0.2236765586619333,
+      "grad_norm": 0.1803252398967743,
+      "learning_rate": 0.00011995810685094813,
+      "loss": 0.6988,
+      "step": 8590
+    },
+    {
+      "epoch": 0.22393695046479936,
+      "grad_norm": 0.17487098276615143,
+      "learning_rate": 0.0001197944830041416,
+      "loss": 0.7198,
+      "step": 8600
+    },
+    {
+      "epoch": 0.2241973422676654,
+      "grad_norm": 0.174115851521492,
+      "learning_rate": 0.00011963080398224661,
+      "loss": 0.6904,
+      "step": 8610
+    },
+    {
+      "epoch": 0.22445773407053146,
+      "grad_norm": 0.17489659786224365,
+      "learning_rate": 0.0001194670702415017,
+      "loss": 0.6738,
+      "step": 8620
+    },
+    {
+      "epoch": 0.2247181258733975,
+      "grad_norm": 0.17910204827785492,
+      "learning_rate": 0.00011930328223829779,
+      "loss": 0.7042,
+      "step": 8630
+    },
+    {
+      "epoch": 0.22497851767626356,
+      "grad_norm": 0.17602995038032532,
+      "learning_rate": 0.00011913944042917711,
+      "loss": 0.7068,
+      "step": 8640
+    },
+    {
+      "epoch": 0.22523890947912958,
+      "grad_norm": 0.18906055390834808,
+      "learning_rate": 0.0001189755452708319,
+      "loss": 0.7128,
+      "step": 8650
+    },
+    {
+      "epoch": 0.22549930128199563,
+      "grad_norm": 0.17224588990211487,
+      "learning_rate": 0.0001188115972201031,
+      "loss": 0.7158,
+      "step": 8660
+    },
+    {
+      "epoch": 0.22575969308486168,
+      "grad_norm": 0.18284164369106293,
+      "learning_rate": 0.00011864759673397901,
+      "loss": 0.6838,
+      "step": 8670
+    },
+    {
+      "epoch": 0.22602008488772773,
+      "grad_norm": 0.17522230744361877,
+      "learning_rate": 0.00011848354426959417,
+      "loss": 0.6832,
+      "step": 8680
+    },
+    {
+      "epoch": 0.22628047669059378,
+      "grad_norm": 0.17182756960391998,
+      "learning_rate": 0.00011831944028422798,
+      "loss": 0.6911,
+      "step": 8690
+    },
+    {
+      "epoch": 0.22654086849345983,
+      "grad_norm": 0.17457079887390137,
+      "learning_rate": 0.00011815528523530349,
+      "loss": 0.692,
+      "step": 8700
+    },
+    {
+      "epoch": 0.22680126029632589,
+      "grad_norm": 0.17242825031280518,
+      "learning_rate": 0.00011799107958038593,
+      "loss": 0.6949,
+      "step": 8710
+    },
+    {
+      "epoch": 0.2270616520991919,
+      "grad_norm": 0.18559910356998444,
+      "learning_rate": 0.0001178268237771818,
+      "loss": 0.6802,
+      "step": 8720
+    },
+    {
+      "epoch": 0.22732204390205796,
+      "grad_norm": 0.18594273924827576,
+      "learning_rate": 0.00011766251828353722,
+      "loss": 0.7008,
+      "step": 8730
+    },
+    {
+      "epoch": 0.227582435704924,
+      "grad_norm": 0.17495504021644592,
+      "learning_rate": 0.00011749816355743694,
+      "loss": 0.7119,
+      "step": 8740
+    },
+    {
+      "epoch": 0.22784282750779006,
+      "grad_norm": 0.17004092037677765,
+      "learning_rate": 0.00011733376005700283,
+      "loss": 0.685,
+      "step": 8750
+    },
+    {
+      "epoch": 0.2281032193106561,
+      "grad_norm": 0.17227309942245483,
+      "learning_rate": 0.00011716930824049274,
+      "loss": 0.6895,
+      "step": 8760
+    },
+    {
+      "epoch": 0.22836361111352216,
+      "grad_norm": 0.18226534128189087,
+      "learning_rate": 0.00011700480856629929,
+      "loss": 0.6874,
+      "step": 8770
+    },
+    {
+      "epoch": 0.22862400291638818,
+      "grad_norm": 0.17005032300949097,
+      "learning_rate": 0.0001168402614929484,
+      "loss": 0.693,
+      "step": 8780
+    },
+    {
+      "epoch": 0.22888439471925423,
+      "grad_norm": 0.1789228320121765,
+      "learning_rate": 0.0001166756674790981,
+      "loss": 0.6952,
+      "step": 8790
+    },
+    {
+      "epoch": 0.22914478652212028,
+      "grad_norm": 0.17560814321041107,
+      "learning_rate": 0.00011651102698353738,
+      "loss": 0.7052,
+      "step": 8800
+    },
+    {
+      "epoch": 0.22940517832498633,
+      "grad_norm": 0.18214547634124756,
+      "learning_rate": 0.0001163463404651846,
+      "loss": 0.6864,
+      "step": 8810
+    },
+    {
+      "epoch": 0.22966557012785238,
+      "grad_norm": 0.17090913653373718,
+      "learning_rate": 0.00011618160838308663,
+      "loss": 0.6847,
+      "step": 8820
+    },
+    {
+      "epoch": 0.22992596193071843,
+      "grad_norm": 0.1777815967798233,
+      "learning_rate": 0.00011601683119641714,
+      "loss": 0.7035,
+      "step": 8830
+    },
+    {
+      "epoch": 0.23018635373358445,
+      "grad_norm": 0.1708465963602066,
+      "learning_rate": 0.00011585200936447567,
+      "loss": 0.6939,
+      "step": 8840
+    },
+    {
+      "epoch": 0.2304467455364505,
+      "grad_norm": 0.1829666793346405,
+      "learning_rate": 0.00011568714334668612,
+      "loss": 0.6898,
+      "step": 8850
+    },
+    {
+      "epoch": 0.23070713733931655,
+      "grad_norm": 0.17379315197467804,
+      "learning_rate": 0.0001155222336025956,
+      "loss": 0.7003,
+      "step": 8860
+    },
+    {
+      "epoch": 0.2309675291421826,
+      "grad_norm": 0.17232435941696167,
+      "learning_rate": 0.00011535728059187307,
+      "loss": 0.6967,
+      "step": 8870
+    },
+    {
+      "epoch": 0.23122792094504865,
+      "grad_norm": 0.1840342879295349,
+      "learning_rate": 0.00011519228477430814,
+      "loss": 0.695,
+      "step": 8880
+    },
+    {
+      "epoch": 0.2314883127479147,
+      "grad_norm": 0.17518925666809082,
+      "learning_rate": 0.00011502724660980968,
+      "loss": 0.6982,
+      "step": 8890
+    },
+    {
+      "epoch": 0.23174870455078075,
+      "grad_norm": 0.18575233221054077,
+      "learning_rate": 0.00011486216655840463,
+      "loss": 0.722,
+      "step": 8900
+    },
+    {
+      "epoch": 0.23200909635364678,
+      "grad_norm": 0.18776971101760864,
+      "learning_rate": 0.00011469704508023667,
+      "loss": 0.7283,
+      "step": 8910
+    },
+    {
+      "epoch": 0.23226948815651283,
+      "grad_norm": 0.17691543698310852,
+      "learning_rate": 0.00011453188263556493,
+      "loss": 0.6796,
+      "step": 8920
+    },
+    {
+      "epoch": 0.23252987995937888,
+      "grad_norm": 0.17249678075313568,
+      "learning_rate": 0.00011436667968476284,
+      "loss": 0.6891,
+      "step": 8930
+    },
+    {
+      "epoch": 0.23279027176224493,
+      "grad_norm": 0.17315714061260223,
+      "learning_rate": 0.00011420143668831662,
+      "loss": 0.6844,
+      "step": 8940
+    },
+    {
+      "epoch": 0.23305066356511098,
+      "grad_norm": 0.17669366300106049,
+      "learning_rate": 0.00011403615410682416,
+      "loss": 0.6848,
+      "step": 8950
+    },
+    {
+      "epoch": 0.23331105536797703,
+      "grad_norm": 0.1768043041229248,
+      "learning_rate": 0.00011387083240099364,
+      "loss": 0.6887,
+      "step": 8960
+    },
+    {
+      "epoch": 0.23357144717084305,
+      "grad_norm": 0.17968913912773132,
+      "learning_rate": 0.00011370547203164237,
+      "loss": 0.7092,
+      "step": 8970
+    },
+    {
+      "epoch": 0.2338318389737091,
+      "grad_norm": 0.1745702028274536,
+      "learning_rate": 0.0001135400734596954,
+      "loss": 0.6812,
+      "step": 8980
+    },
+    {
+      "epoch": 0.23409223077657515,
+      "grad_norm": 0.1707821488380432,
+      "learning_rate": 0.00011337463714618423,
+      "loss": 0.6685,
+      "step": 8990
+    },
+    {
+      "epoch": 0.2343526225794412,
+      "grad_norm": 0.1789083033800125,
+      "learning_rate": 0.00011320916355224565,
+      "loss": 0.6746,
+      "step": 9000
+    },
+    {
+      "epoch": 0.23461301438230725,
+      "grad_norm": 0.17847011983394623,
+      "learning_rate": 0.00011304365313912028,
+      "loss": 0.695,
+      "step": 9010
+    },
+    {
+      "epoch": 0.2348734061851733,
+      "grad_norm": 0.18885819613933563,
+      "learning_rate": 0.00011287810636815139,
+      "loss": 0.698,
+      "step": 9020
+    },
+    {
+      "epoch": 0.23513379798803935,
+      "grad_norm": 0.17764124274253845,
+      "learning_rate": 0.00011271252370078363,
+      "loss": 0.6892,
+      "step": 9030
+    },
+    {
+      "epoch": 0.23539418979090537,
+      "grad_norm": 0.17720068991184235,
+      "learning_rate": 0.0001125469055985616,
+      "loss": 0.6926,
+      "step": 9040
+    },
+    {
+      "epoch": 0.23565458159377142,
+      "grad_norm": 0.17689958214759827,
+      "learning_rate": 0.00011238125252312888,
+      "loss": 0.6987,
+      "step": 9050
+    },
+    {
+      "epoch": 0.23591497339663747,
+      "grad_norm": 0.19373558461666107,
+      "learning_rate": 0.00011221556493622627,
+      "loss": 0.687,
+      "step": 9060
+    },
+    {
+      "epoch": 0.23617536519950352,
+      "grad_norm": 0.1779530942440033,
+      "learning_rate": 0.000112049843299691,
+      "loss": 0.6735,
+      "step": 9070
+    },
+    {
+      "epoch": 0.23643575700236957,
+      "grad_norm": 0.1853463500738144,
+      "learning_rate": 0.00011188408807545505,
+      "loss": 0.6765,
+      "step": 9080
+    },
+    {
+      "epoch": 0.23669614880523562,
+      "grad_norm": 0.1809043437242508,
+      "learning_rate": 0.00011171829972554415,
+      "loss": 0.6775,
+      "step": 9090
+    },
+    {
+      "epoch": 0.23695654060810165,
+      "grad_norm": 0.16961175203323364,
+      "learning_rate": 0.00011155247871207624,
+      "loss": 0.6947,
+      "step": 9100
+    },
+    {
+      "epoch": 0.2372169324109677,
+      "grad_norm": 0.1882382184267044,
+      "learning_rate": 0.0001113866254972604,
+      "loss": 0.6855,
+      "step": 9110
+    },
+    {
+      "epoch": 0.23747732421383375,
+      "grad_norm": 0.18256156146526337,
+      "learning_rate": 0.00011122074054339537,
+      "loss": 0.7131,
+      "step": 9120
+    },
+    {
+      "epoch": 0.2377377160166998,
+      "grad_norm": 0.18344970047473907,
+      "learning_rate": 0.00011105482431286852,
+      "loss": 0.6596,
+      "step": 9130
+    },
+    {
+      "epoch": 0.23799810781956585,
+      "grad_norm": 0.18037189543247223,
+      "learning_rate": 0.00011088887726815422,
+      "loss": 0.6853,
+      "step": 9140
+    },
+    {
+      "epoch": 0.2382584996224319,
+      "grad_norm": 0.17328627407550812,
+      "learning_rate": 0.00011072289987181288,
+      "loss": 0.6833,
+      "step": 9150
+    },
+    {
+      "epoch": 0.23851889142529792,
+      "grad_norm": 0.1846616417169571,
+      "learning_rate": 0.00011055689258648938,
+      "loss": 0.6809,
+      "step": 9160
+    },
+    {
+      "epoch": 0.23877928322816397,
+      "grad_norm": 0.17733514308929443,
+      "learning_rate": 0.00011039085587491203,
+      "loss": 0.7013,
+      "step": 9170
+    },
+    {
+      "epoch": 0.23903967503103002,
+      "grad_norm": 0.17087826132774353,
+      "learning_rate": 0.0001102247901998911,
+      "loss": 0.6808,
+      "step": 9180
+    },
+    {
+      "epoch": 0.23930006683389607,
+      "grad_norm": 0.17394234240055084,
+      "learning_rate": 0.00011005869602431758,
+      "loss": 0.6955,
+      "step": 9190
+    },
+    {
+      "epoch": 0.23956045863676212,
+      "grad_norm": 0.1748206615447998,
+      "learning_rate": 0.00010989257381116194,
+      "loss": 0.6944,
+      "step": 9200
+    },
+    {
+      "epoch": 0.23982085043962817,
+      "grad_norm": 0.1819637566804886,
+      "learning_rate": 0.00010972642402347277,
+      "loss": 0.6875,
+      "step": 9210
+    },
+    {
+      "epoch": 0.24008124224249422,
+      "grad_norm": 0.17814664542675018,
+      "learning_rate": 0.00010956024712437558,
+      "loss": 0.6864,
+      "step": 9220
+    },
+    {
+      "epoch": 0.24034163404536024,
+      "grad_norm": 0.17680172622203827,
+      "learning_rate": 0.00010939404357707141,
+      "loss": 0.6727,
+      "step": 9230
+    },
+    {
+      "epoch": 0.2406020258482263,
+      "grad_norm": 0.175360307097435,
+      "learning_rate": 0.00010922781384483553,
+      "loss": 0.6972,
+      "step": 9240
+    },
+    {
+      "epoch": 0.24086241765109234,
+      "grad_norm": 0.1745460033416748,
+      "learning_rate": 0.0001090615583910163,
+      "loss": 0.6933,
+      "step": 9250
+    },
+    {
+      "epoch": 0.2411228094539584,
+      "grad_norm": 0.1743789166212082,
+      "learning_rate": 0.00010889527767903366,
+      "loss": 0.6719,
+      "step": 9260
+    },
+    {
+      "epoch": 0.24138320125682444,
+      "grad_norm": 0.17672204971313477,
+      "learning_rate": 0.00010872897217237812,
+      "loss": 0.6705,
+      "step": 9270
+    },
+    {
+      "epoch": 0.2416435930596905,
+      "grad_norm": 0.16973887383937836,
+      "learning_rate": 0.00010856264233460907,
+      "loss": 0.6772,
+      "step": 9280
+    },
+    {
+      "epoch": 0.24190398486255651,
+      "grad_norm": 0.17453502118587494,
+      "learning_rate": 0.00010839628862935392,
+      "loss": 0.6856,
+      "step": 9290
+    },
+    {
+      "epoch": 0.24216437666542256,
+      "grad_norm": 0.17804546654224396,
+      "learning_rate": 0.00010822991152030655,
+      "loss": 0.6766,
+      "step": 9300
+    },
+    {
+      "epoch": 0.24242476846828862,
+      "grad_norm": 0.17404384911060333,
+      "learning_rate": 0.00010806351147122605,
+      "loss": 0.6914,
+      "step": 9310
+    },
+    {
+      "epoch": 0.24268516027115467,
+      "grad_norm": 0.17581741511821747,
+      "learning_rate": 0.00010789708894593547,
+      "loss": 0.6933,
+      "step": 9320
+    },
+    {
+      "epoch": 0.24294555207402072,
+      "grad_norm": 0.18238358199596405,
+      "learning_rate": 0.00010773064440832047,
+      "loss": 0.7241,
+      "step": 9330
+    },
+    {
+      "epoch": 0.24320594387688677,
+      "grad_norm": 0.17800027132034302,
+      "learning_rate": 0.00010756417832232819,
+      "loss": 0.6881,
+      "step": 9340
+    },
+    {
+      "epoch": 0.2434663356797528,
+      "grad_norm": 0.17522063851356506,
+      "learning_rate": 0.00010739769115196566,
+      "loss": 0.6916,
+      "step": 9350
+    },
+    {
+      "epoch": 0.24372672748261884,
+      "grad_norm": 0.17781999707221985,
+      "learning_rate": 0.00010723118336129878,
+      "loss": 0.6809,
+      "step": 9360
+    },
+    {
+      "epoch": 0.2439871192854849,
+      "grad_norm": 0.1767900288105011,
+      "learning_rate": 0.00010706465541445095,
+      "loss": 0.6966,
+      "step": 9370
+    },
+    {
+      "epoch": 0.24424751108835094,
+      "grad_norm": 0.19158516824245453,
+      "learning_rate": 0.00010689810777560174,
+      "loss": 0.6949,
+      "step": 9380
+    },
+    {
+      "epoch": 0.244507902891217,
+      "grad_norm": 0.1787518560886383,
+      "learning_rate": 0.00010673154090898552,
+      "loss": 0.6922,
+      "step": 9390
+    },
+    {
+      "epoch": 0.24476829469408304,
+      "grad_norm": 0.17122618854045868,
+      "learning_rate": 0.00010656495527889039,
+      "loss": 0.6878,
+      "step": 9400
+    },
+    {
+      "epoch": 0.2450286864969491,
+      "grad_norm": 0.18457964062690735,
+      "learning_rate": 0.00010639835134965664,
+      "loss": 0.6723,
+      "step": 9410
+    },
+    {
+      "epoch": 0.2452890782998151,
+      "grad_norm": 0.17373178899288177,
+      "learning_rate": 0.00010623172958567563,
+      "loss": 0.6954,
+      "step": 9420
+    },
+    {
+      "epoch": 0.24554947010268116,
+      "grad_norm": 0.18350104987621307,
+      "learning_rate": 0.0001060650904513884,
+      "loss": 0.6824,
+      "step": 9430
+    },
+    {
+      "epoch": 0.2458098619055472,
+      "grad_norm": 0.17673034965991974,
+      "learning_rate": 0.00010589843441128445,
+      "loss": 0.6968,
+      "step": 9440
+    },
+    {
+      "epoch": 0.24607025370841326,
+      "grad_norm": 0.17909649014472961,
+      "learning_rate": 0.00010573176192990032,
+      "loss": 0.7048,
+      "step": 9450
+    },
+    {
+      "epoch": 0.2463306455112793,
+      "grad_norm": 0.18223048746585846,
+      "learning_rate": 0.00010556507347181854,
+      "loss": 0.6939,
+      "step": 9460
+    },
+    {
+      "epoch": 0.24659103731414536,
+      "grad_norm": 0.17868250608444214,
+      "learning_rate": 0.00010539836950166595,
+      "loss": 0.7082,
+      "step": 9470
+    },
+    {
+      "epoch": 0.24685142911701138,
+      "grad_norm": 0.1783342808485031,
+      "learning_rate": 0.00010523165048411283,
+      "loss": 0.6903,
+      "step": 9480
+    },
+    {
+      "epoch": 0.24711182091987743,
+      "grad_norm": 0.17360354959964752,
+      "learning_rate": 0.00010506491688387127,
+      "loss": 0.6635,
+      "step": 9490
+    },
+    {
+      "epoch": 0.24737221272274348,
+      "grad_norm": 0.18421316146850586,
+      "learning_rate": 0.0001048981691656941,
+      "loss": 0.6848,
+      "step": 9500
+    },
+    {
+      "epoch": 0.24763260452560953,
+      "grad_norm": 0.1938525140285492,
+      "learning_rate": 0.00010473140779437341,
+      "loss": 0.7,
+      "step": 9510
+    },
+    {
+      "epoch": 0.24789299632847558,
+      "grad_norm": 0.17088712751865387,
+      "learning_rate": 0.00010456463323473942,
+      "loss": 0.6827,
+      "step": 9520
+    },
+    {
+      "epoch": 0.24815338813134163,
+      "grad_norm": 0.17812180519104004,
+      "learning_rate": 0.00010439784595165909,
+      "loss": 0.6662,
+      "step": 9530
+    },
+    {
+      "epoch": 0.24841377993420768,
+      "grad_norm": 0.18220531940460205,
+      "learning_rate": 0.00010423104641003484,
+      "loss": 0.6878,
+      "step": 9540
+    },
+    {
+      "epoch": 0.2486741717370737,
+      "grad_norm": 0.17387458682060242,
+      "learning_rate": 0.00010406423507480325,
+      "loss": 0.6896,
+      "step": 9550
+    },
+    {
+      "epoch": 0.24893456353993976,
+      "grad_norm": 0.16997171938419342,
+      "learning_rate": 0.00010389741241093382,
+      "loss": 0.6982,
+      "step": 9560
+    },
+    {
+      "epoch": 0.2491949553428058,
+      "grad_norm": 0.18163439631462097,
+      "learning_rate": 0.00010373057888342755,
+      "loss": 0.6769,
+      "step": 9570
+    },
+    {
+      "epoch": 0.24945534714567186,
+      "grad_norm": 0.1709483563899994,
+      "learning_rate": 0.00010356373495731579,
+      "loss": 0.6935,
+      "step": 9580
+    },
+    {
+      "epoch": 0.2497157389485379,
+      "grad_norm": 0.17501015961170197,
+      "learning_rate": 0.00010339688109765882,
+      "loss": 0.6961,
+      "step": 9590
+    },
+    {
+      "epoch": 0.24997613075140396,
+      "grad_norm": 0.177652046084404,
+      "learning_rate": 0.00010323001776954466,
+      "loss": 0.687,
+      "step": 9600
+    },
+    {
+      "epoch": 0.25023652255427,
+      "grad_norm": 0.16977092623710632,
+      "learning_rate": 0.00010306314543808768,
+      "loss": 0.6758,
+      "step": 9610
+    },
+    {
+      "epoch": 0.25049691435713606,
+      "grad_norm": 0.17998236417770386,
+      "learning_rate": 0.00010289626456842743,
+      "loss": 0.7022,
+      "step": 9620
+    },
+    {
+      "epoch": 0.2507573061600021,
+      "grad_norm": 0.17297130823135376,
+      "learning_rate": 0.00010272937562572714,
+      "loss": 0.6614,
+      "step": 9630
+    },
+    {
+      "epoch": 0.2510176979628681,
+      "grad_norm": 0.1776668131351471,
+      "learning_rate": 0.00010256247907517258,
+      "loss": 0.6723,
+      "step": 9640
+    },
+    {
+      "epoch": 0.2512780897657342,
+      "grad_norm": 0.17850397527217865,
+      "learning_rate": 0.00010239557538197077,
+      "loss": 0.6856,
+      "step": 9650
+    },
+    {
+      "epoch": 0.2515384815686002,
+      "grad_norm": 0.17861175537109375,
+      "learning_rate": 0.00010222866501134864,
+      "loss": 0.6912,
+      "step": 9660
+    },
+    {
+      "epoch": 0.2517988733714663,
+      "grad_norm": 0.18047182261943817,
+      "learning_rate": 0.00010206174842855167,
+      "loss": 0.7047,
+      "step": 9670
+    },
+    {
+      "epoch": 0.2520592651743323,
+      "grad_norm": 0.1728067547082901,
+      "learning_rate": 0.00010189482609884269,
+      "loss": 0.7061,
+      "step": 9680
+    },
+    {
+      "epoch": 0.2523196569771984,
+      "grad_norm": 0.1772705316543579,
+      "learning_rate": 0.00010172789848750057,
+      "loss": 0.6817,
+      "step": 9690
+    },
+    {
+      "epoch": 0.2525800487800644,
+      "grad_norm": 0.17873169481754303,
+      "learning_rate": 0.00010156096605981889,
+      "loss": 0.682,
+      "step": 9700
+    },
+    {
+      "epoch": 0.2528404405829304,
+      "grad_norm": 0.1827894151210785,
+      "learning_rate": 0.00010139402928110463,
+      "loss": 0.6636,
+      "step": 9710
+    },
+    {
+      "epoch": 0.2531008323857965,
+      "grad_norm": 0.16760335862636566,
+      "learning_rate": 0.00010122708861667689,
+      "loss": 0.6725,
+      "step": 9720
+    },
+    {
+      "epoch": 0.2533612241886625,
+      "grad_norm": 0.18811176717281342,
+      "learning_rate": 0.00010106014453186568,
+      "loss": 0.6856,
+      "step": 9730
+    },
+    {
+      "epoch": 0.2536216159915286,
+      "grad_norm": 0.1821964532136917,
+      "learning_rate": 0.00010089319749201043,
+      "loss": 0.6736,
+      "step": 9740
+    },
+    {
+      "epoch": 0.2538820077943946,
+      "grad_norm": 0.1711428463459015,
+      "learning_rate": 0.00010072624796245895,
+      "loss": 0.6811,
+      "step": 9750
+    },
+    {
+      "epoch": 0.2541423995972607,
+      "grad_norm": 0.1852736473083496,
+      "learning_rate": 0.00010055929640856583,
+      "loss": 0.703,
+      "step": 9760
+    },
+    {
+      "epoch": 0.2544027914001267,
+      "grad_norm": 0.18267115950584412,
+      "learning_rate": 0.00010039234329569145,
+      "loss": 0.6874,
+      "step": 9770
+    },
+    {
+      "epoch": 0.25466318320299275,
+      "grad_norm": 0.1899646520614624,
+      "learning_rate": 0.00010022538908920039,
+      "loss": 0.6655,
+      "step": 9780
+    },
+    {
+      "epoch": 0.2549235750058588,
+      "grad_norm": 0.18259014189243317,
+      "learning_rate": 0.00010005843425446042,
+      "loss": 0.6755,
+      "step": 9790
+    },
+    {
+      "epoch": 0.25518396680872485,
+      "grad_norm": 0.18706420063972473,
+      "learning_rate": 9.989147925684094e-05,
+      "loss": 0.678,
+      "step": 9800
+    },
+    {
+      "epoch": 0.2554443586115909,
+      "grad_norm": 0.17754283547401428,
+      "learning_rate": 9.97245245617119e-05,
+      "loss": 0.6793,
+      "step": 9810
+    },
+    {
+      "epoch": 0.25570475041445695,
+      "grad_norm": 0.16971617937088013,
+      "learning_rate": 9.955757063444235e-05,
+      "loss": 0.6864,
+      "step": 9820
+    },
+    {
+      "epoch": 0.25596514221732297,
+      "grad_norm": 0.1726657599210739,
+      "learning_rate": 9.939061794039922e-05,
+      "loss": 0.6751,
+      "step": 9830
+    },
+    {
+      "epoch": 0.25622553402018905,
+      "grad_norm": 0.18278145790100098,
+      "learning_rate": 9.922366694494602e-05,
+      "loss": 0.7024,
+      "step": 9840
+    },
+    {
+      "epoch": 0.25648592582305507,
+      "grad_norm": 0.1729104220867157,
+      "learning_rate": 9.905671811344141e-05,
+      "loss": 0.684,
+      "step": 9850
+    },
+    {
+      "epoch": 0.25674631762592115,
+      "grad_norm": 0.18848705291748047,
+      "learning_rate": 9.888977191123825e-05,
+      "loss": 0.6933,
+      "step": 9860
+    },
+    {
+      "epoch": 0.2570067094287872,
+      "grad_norm": 0.19198398292064667,
+      "learning_rate": 9.872282880368186e-05,
+      "loss": 0.6784,
+      "step": 9870
+    },
+    {
+      "epoch": 0.25726710123165325,
+      "grad_norm": 0.17954795062541962,
+      "learning_rate": 9.855588925610898e-05,
+      "loss": 0.6959,
+      "step": 9880
+    },
+    {
+      "epoch": 0.2575274930345193,
+      "grad_norm": 0.18434350192546844,
+      "learning_rate": 9.838895373384649e-05,
+      "loss": 0.7033,
+      "step": 9890
+    },
+    {
+      "epoch": 0.2577878848373853,
+      "grad_norm": 0.17057251930236816,
+      "learning_rate": 9.822202270221002e-05,
+      "loss": 0.674,
+      "step": 9900
+    },
+    {
+      "epoch": 0.2580482766402514,
+      "grad_norm": 0.18517877161502838,
+      "learning_rate": 9.805509662650268e-05,
+      "loss": 0.6922,
+      "step": 9910
+    },
+    {
+      "epoch": 0.2583086684431174,
+      "grad_norm": 0.1838076114654541,
+      "learning_rate": 9.788817597201373e-05,
+      "loss": 0.6831,
+      "step": 9920
+    },
+    {
+      "epoch": 0.2585690602459835,
+      "grad_norm": 0.17518235743045807,
+      "learning_rate": 9.772126120401734e-05,
+      "loss": 0.7101,
+      "step": 9930
+    },
+    {
+      "epoch": 0.2588294520488495,
+      "grad_norm": 0.18193618953227997,
+      "learning_rate": 9.755435278777136e-05,
+      "loss": 0.6702,
+      "step": 9940
+    },
+    {
+      "epoch": 0.2590898438517156,
+      "grad_norm": 0.1757705956697464,
+      "learning_rate": 9.738745118851581e-05,
+      "loss": 0.6949,
+      "step": 9950
+    },
+    {
+      "epoch": 0.2593502356545816,
+      "grad_norm": 0.1771358698606491,
+      "learning_rate": 9.722055687147174e-05,
+      "loss": 0.6654,
+      "step": 9960
+    },
+    {
+      "epoch": 0.2596106274574476,
+      "grad_norm": 0.17536123096942902,
+      "learning_rate": 9.705367030183992e-05,
+      "loss": 0.6797,
+      "step": 9970
+    },
+    {
+      "epoch": 0.2598710192603137,
+      "grad_norm": 0.1831720471382141,
+      "learning_rate": 9.688679194479955e-05,
+      "loss": 0.696,
+      "step": 9980
+    },
+    {
+      "epoch": 0.2601314110631797,
+      "grad_norm": 0.17759549617767334,
+      "learning_rate": 9.671992226550692e-05,
+      "loss": 0.7069,
+      "step": 9990
+    },
+    {
+      "epoch": 0.2603918028660458,
+      "grad_norm": 0.17348560690879822,
+      "learning_rate": 9.65530617290941e-05,
+      "loss": 0.6808,
+      "step": 10000
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 19202,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.6916733763584e+18,
+  "train_batch_size": 5,
+  "trial_name": null,
+  "trial_params": null
+}