{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.999983338054218,
  "eval_steps": 50,
  "global_step": 3751,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002665911325124548,
      "grad_norm": 25.460365315883454,
      "learning_rate": 2.6595744680851066e-07,
      "loss": 0.5957,
      "step": 10
    },
    {
      "epoch": 0.005331822650249096,
      "grad_norm": 7.786416251497864,
      "learning_rate": 5.319148936170213e-07,
      "loss": 0.5163,
      "step": 20
    },
    {
      "epoch": 0.007997733975373645,
      "grad_norm": 6.545796555174715,
      "learning_rate": 7.97872340425532e-07,
      "loss": 0.4334,
      "step": 30
    },
    {
      "epoch": 0.010663645300498192,
      "grad_norm": 5.109112355076126,
      "learning_rate": 1.0638297872340427e-06,
      "loss": 0.3824,
      "step": 40
    },
    {
      "epoch": 0.01332955662562274,
      "grad_norm": 5.129086050533378,
      "learning_rate": 1.3297872340425533e-06,
      "loss": 0.3821,
      "step": 50
    },
    {
      "epoch": 0.01332955662562274,
      "eval_loss": 0.4735161364078522,
      "eval_runtime": 263.1387,
      "eval_samples_per_second": 37.239,
      "eval_steps_per_second": 4.655,
      "step": 50
    },
    {
      "epoch": 0.01599546795074729,
      "grad_norm": 3.9633957758228795,
      "learning_rate": 1.595744680851064e-06,
      "loss": 0.3431,
      "step": 60
    },
    {
      "epoch": 0.018661379275871838,
      "grad_norm": 5.099960208152394,
      "learning_rate": 1.8617021276595745e-06,
      "loss": 0.3327,
      "step": 70
    },
    {
      "epoch": 0.021327290600996383,
      "grad_norm": 4.6877091477318995,
      "learning_rate": 2.1276595744680853e-06,
      "loss": 0.3079,
      "step": 80
    },
    {
      "epoch": 0.023993201926120932,
      "grad_norm": 4.2158979353952,
      "learning_rate": 2.393617021276596e-06,
      "loss": 0.306,
      "step": 90
    },
    {
      "epoch": 0.02665911325124548,
      "grad_norm": 4.111674187143645,
      "learning_rate": 2.6595744680851065e-06,
      "loss": 0.302,
      "step": 100
    },
    {
      "epoch": 0.02665911325124548,
      "eval_loss": 0.3177907168865204,
      "eval_runtime": 263.1879,
      "eval_samples_per_second": 37.232,
      "eval_steps_per_second": 4.654,
      "step": 100
    },
    {
      "epoch": 0.02932502457637003,
      "grad_norm": 3.796268052824722,
      "learning_rate": 2.9255319148936174e-06,
      "loss": 0.2978,
      "step": 110
    },
    {
      "epoch": 0.03199093590149458,
      "grad_norm": 4.080365457232974,
      "learning_rate": 3.191489361702128e-06,
      "loss": 0.3046,
      "step": 120
    },
    {
      "epoch": 0.034656847226619124,
      "grad_norm": 3.570907286511261,
      "learning_rate": 3.457446808510639e-06,
      "loss": 0.3074,
      "step": 130
    },
    {
      "epoch": 0.037322758551743676,
      "grad_norm": 3.6056579252874896,
      "learning_rate": 3.723404255319149e-06,
      "loss": 0.3147,
      "step": 140
    },
    {
      "epoch": 0.03998866987686822,
      "grad_norm": 3.3791442973244163,
      "learning_rate": 3.98936170212766e-06,
      "loss": 0.2988,
      "step": 150
    },
    {
      "epoch": 0.03998866987686822,
      "eval_loss": 0.32532769441604614,
      "eval_runtime": 263.0021,
      "eval_samples_per_second": 37.258,
      "eval_steps_per_second": 4.658,
      "step": 150
    },
    {
      "epoch": 0.04265458120199277,
      "grad_norm": 3.7568359111958665,
      "learning_rate": 4.255319148936171e-06,
      "loss": 0.3049,
      "step": 160
    },
    {
      "epoch": 0.04532049252711732,
      "grad_norm": 3.568899075332689,
      "learning_rate": 4.521276595744681e-06,
      "loss": 0.3077,
      "step": 170
    },
    {
      "epoch": 0.047986403852241864,
      "grad_norm": 3.446482818546644,
      "learning_rate": 4.787234042553192e-06,
      "loss": 0.2957,
      "step": 180
    },
    {
      "epoch": 0.05065231517736641,
      "grad_norm": 3.3569986066448223,
      "learning_rate": 5.053191489361703e-06,
      "loss": 0.2993,
      "step": 190
    },
    {
      "epoch": 0.05331822650249096,
      "grad_norm": 3.502984902631741,
      "learning_rate": 5.319148936170213e-06,
      "loss": 0.3054,
      "step": 200
    },
    {
      "epoch": 0.05331822650249096,
      "eval_loss": 0.3250272572040558,
      "eval_runtime": 263.1161,
      "eval_samples_per_second": 37.242,
      "eval_steps_per_second": 4.656,
      "step": 200
    },
    {
      "epoch": 0.05598413782761551,
      "grad_norm": 2.8396576424885387,
      "learning_rate": 5.5851063829787235e-06,
      "loss": 0.3117,
      "step": 210
    },
    {
      "epoch": 0.05865004915274006,
      "grad_norm": 2.8315864264900372,
      "learning_rate": 5.851063829787235e-06,
      "loss": 0.3102,
      "step": 220
    },
    {
      "epoch": 0.061315960477864605,
      "grad_norm": 3.068071736618938,
      "learning_rate": 6.117021276595745e-06,
      "loss": 0.3096,
      "step": 230
    },
    {
      "epoch": 0.06398187180298916,
      "grad_norm": 3.2824497787656357,
      "learning_rate": 6.382978723404256e-06,
      "loss": 0.3046,
      "step": 240
    },
    {
      "epoch": 0.0666477831281137,
      "grad_norm": 2.750006857001217,
      "learning_rate": 6.648936170212767e-06,
      "loss": 0.2967,
      "step": 250
    },
    {
      "epoch": 0.0666477831281137,
      "eval_loss": 0.3232134282588959,
      "eval_runtime": 263.0112,
      "eval_samples_per_second": 37.257,
      "eval_steps_per_second": 4.658,
      "step": 250
    },
    {
      "epoch": 0.06931369445323825,
      "grad_norm": 2.7213269114514205,
      "learning_rate": 6.914893617021278e-06,
      "loss": 0.3112,
      "step": 260
    },
    {
      "epoch": 0.07197960577836279,
      "grad_norm": 2.5062744105701653,
      "learning_rate": 7.1808510638297875e-06,
      "loss": 0.314,
      "step": 270
    },
    {
      "epoch": 0.07464551710348735,
      "grad_norm": 2.541718637398392,
      "learning_rate": 7.446808510638298e-06,
      "loss": 0.3041,
      "step": 280
    },
    {
      "epoch": 0.0773114284286119,
      "grad_norm": 3.203380938887303,
      "learning_rate": 7.71276595744681e-06,
      "loss": 0.3136,
      "step": 290
    },
    {
      "epoch": 0.07997733975373644,
      "grad_norm": 2.7982203558961016,
      "learning_rate": 7.97872340425532e-06,
      "loss": 0.3137,
      "step": 300
    },
    {
      "epoch": 0.07997733975373644,
      "eval_loss": 0.3207298815250397,
      "eval_runtime": 262.9784,
      "eval_samples_per_second": 37.262,
      "eval_steps_per_second": 4.658,
      "step": 300
    },
    {
      "epoch": 0.08264325107886099,
      "grad_norm": 2.4586173652717216,
      "learning_rate": 8.24468085106383e-06,
      "loss": 0.3092,
      "step": 310
    },
    {
      "epoch": 0.08530916240398553,
      "grad_norm": 2.470341507162159,
      "learning_rate": 8.510638297872341e-06,
      "loss": 0.3255,
      "step": 320
    },
    {
      "epoch": 0.08797507372911008,
      "grad_norm": 2.531961920922859,
      "learning_rate": 8.776595744680852e-06,
      "loss": 0.3219,
      "step": 330
    },
    {
      "epoch": 0.09064098505423464,
      "grad_norm": 2.218215799424983,
      "learning_rate": 9.042553191489362e-06,
      "loss": 0.3093,
      "step": 340
    },
    {
      "epoch": 0.09330689637935918,
      "grad_norm": 2.5602707890452896,
      "learning_rate": 9.308510638297872e-06,
      "loss": 0.3221,
      "step": 350
    },
    {
      "epoch": 0.09330689637935918,
      "eval_loss": 0.3211060166358948,
      "eval_runtime": 262.9608,
      "eval_samples_per_second": 37.264,
      "eval_steps_per_second": 4.658,
      "step": 350
    },
    {
      "epoch": 0.09597280770448373,
      "grad_norm": 2.4918966800165228,
      "learning_rate": 9.574468085106385e-06,
      "loss": 0.3209,
      "step": 360
    },
    {
      "epoch": 0.09863871902960827,
      "grad_norm": 2.7550293721808603,
      "learning_rate": 9.840425531914895e-06,
      "loss": 0.3188,
      "step": 370
    },
    {
      "epoch": 0.10130463035473282,
      "grad_norm": 2.086538785606296,
      "learning_rate": 9.999965341346946e-06,
      "loss": 0.3202,
      "step": 380
    },
    {
      "epoch": 0.10397054167985738,
      "grad_norm": 2.291915706021129,
      "learning_rate": 9.999575437018172e-06,
      "loss": 0.3143,
      "step": 390
    },
    {
      "epoch": 0.10663645300498192,
      "grad_norm": 2.5734029639653406,
      "learning_rate": 9.998752338940612e-06,
      "loss": 0.3188,
      "step": 400
    },
    {
      "epoch": 0.10663645300498192,
      "eval_loss": 0.3204084634780884,
      "eval_runtime": 263.0611,
      "eval_samples_per_second": 37.25,
      "eval_steps_per_second": 4.657,
      "step": 400
    },
    {
      "epoch": 0.10930236433010647,
      "grad_norm": 1.9778855134324143,
      "learning_rate": 9.997496118432509e-06,
      "loss": 0.3026,
      "step": 410
    },
    {
      "epoch": 0.11196827565523101,
      "grad_norm": 2.2135770753044395,
      "learning_rate": 9.995806884340483e-06,
      "loss": 0.3192,
      "step": 420
    },
    {
      "epoch": 0.11463418698035556,
      "grad_norm": 2.3593510882812634,
      "learning_rate": 9.99368478303009e-06,
      "loss": 0.3143,
      "step": 430
    },
    {
      "epoch": 0.11730009830548012,
      "grad_norm": 2.2314058999493054,
      "learning_rate": 9.991129998373145e-06,
      "loss": 0.3189,
      "step": 440
    },
    {
      "epoch": 0.11996600963060466,
      "grad_norm": 2.1809009277203493,
      "learning_rate": 9.988142751731797e-06,
      "loss": 0.308,
      "step": 450
    },
    {
      "epoch": 0.11996600963060466,
      "eval_loss": 0.31486886739730835,
      "eval_runtime": 263.0029,
      "eval_samples_per_second": 37.258,
      "eval_steps_per_second": 4.658,
      "step": 450
    },
    {
      "epoch": 0.12263192095572921,
      "grad_norm": 2.009644779324099,
      "learning_rate": 9.984723301939337e-06,
      "loss": 0.3172,
      "step": 460
    },
    {
      "epoch": 0.12529783228085375,
      "grad_norm": 2.2024418868068314,
      "learning_rate": 9.980871945277777e-06,
      "loss": 0.3147,
      "step": 470
    },
    {
      "epoch": 0.12796374360597831,
      "grad_norm": 2.076529594327517,
      "learning_rate": 9.976589015452178e-06,
      "loss": 0.3211,
      "step": 480
    },
    {
      "epoch": 0.13062965493110285,
      "grad_norm": 1.921579954198359,
      "learning_rate": 9.97187488356174e-06,
      "loss": 0.3109,
      "step": 490
    },
    {
      "epoch": 0.1332955662562274,
      "grad_norm": 2.298024577351239,
      "learning_rate": 9.966729958067638e-06,
      "loss": 0.3123,
      "step": 500
    },
    {
      "epoch": 0.1332955662562274,
      "eval_loss": 0.3106406331062317,
      "eval_runtime": 263.0069,
      "eval_samples_per_second": 37.258,
      "eval_steps_per_second": 4.658,
      "step": 500
    },
    {
      "epoch": 0.13596147758135196,
      "grad_norm": 2.169882121277794,
      "learning_rate": 9.961154684757636e-06,
      "loss": 0.3032,
      "step": 510
    },
    {
      "epoch": 0.1386273889064765,
      "grad_norm": 2.101861158344262,
      "learning_rate": 9.955149546707465e-06,
      "loss": 0.3048,
      "step": 520
    },
    {
      "epoch": 0.14129330023160105,
      "grad_norm": 2.048031635489159,
      "learning_rate": 9.948715064238956e-06,
      "loss": 0.2947,
      "step": 530
    },
    {
      "epoch": 0.14395921155672559,
      "grad_norm": 2.188308108036795,
      "learning_rate": 9.941851794874969e-06,
      "loss": 0.2998,
      "step": 540
    },
    {
      "epoch": 0.14662512288185015,
      "grad_norm": 2.0056114881626472,
      "learning_rate": 9.934560333291077e-06,
      "loss": 0.3138,
      "step": 550
    },
    {
      "epoch": 0.14662512288185015,
      "eval_loss": 0.3050285875797272,
      "eval_runtime": 263.0851,
      "eval_samples_per_second": 37.247,
      "eval_steps_per_second": 4.656,
      "step": 550
    },
    {
      "epoch": 0.1492910342069747,
      "grad_norm": 1.9352426910596738,
      "learning_rate": 9.926841311264037e-06,
      "loss": 0.3014,
      "step": 560
    },
    {
      "epoch": 0.15195694553209924,
      "grad_norm": 2.2043603905292497,
      "learning_rate": 9.918695397617064e-06,
      "loss": 0.3076,
      "step": 570
    },
    {
      "epoch": 0.1546228568572238,
      "grad_norm": 1.896607728698459,
      "learning_rate": 9.91012329816186e-06,
      "loss": 0.3013,
      "step": 580
    },
    {
      "epoch": 0.15728876818234833,
      "grad_norm": 2.5158079295704776,
      "learning_rate": 9.901125755637473e-06,
      "loss": 0.3017,
      "step": 590
    },
    {
      "epoch": 0.15995467950747289,
      "grad_norm": 2.2139727973208747,
      "learning_rate": 9.89170354964594e-06,
      "loss": 0.3032,
      "step": 600
    },
    {
      "epoch": 0.15995467950747289,
      "eval_loss": 0.3046238422393799,
      "eval_runtime": 263.0944,
      "eval_samples_per_second": 37.245,
      "eval_steps_per_second": 4.656,
      "step": 600
    },
    {
      "epoch": 0.16262059083259742,
      "grad_norm": 2.0311233806900666,
      "learning_rate": 9.881857496584726e-06,
      "loss": 0.3027,
      "step": 610
    },
    {
      "epoch": 0.16528650215772198,
      "grad_norm": 1.8044187965587122,
      "learning_rate": 9.871588449575999e-06,
      "loss": 0.296,
      "step": 620
    },
    {
      "epoch": 0.16795241348284654,
      "grad_norm": 1.9287672543745633,
      "learning_rate": 9.860897298392712e-06,
      "loss": 0.3031,
      "step": 630
    },
    {
      "epoch": 0.17061832480797107,
      "grad_norm": 1.9418604281132057,
      "learning_rate": 9.849784969381488e-06,
      "loss": 0.3006,
      "step": 640
    },
    {
      "epoch": 0.17328423613309563,
      "grad_norm": 1.7607874454195112,
      "learning_rate": 9.83825242538238e-06,
      "loss": 0.2827,
      "step": 650
    },
    {
      "epoch": 0.17328423613309563,
      "eval_loss": 0.30168309807777405,
      "eval_runtime": 263.0754,
      "eval_samples_per_second": 37.248,
      "eval_steps_per_second": 4.656,
      "step": 650
    },
    {
      "epoch": 0.17595014745822016,
      "grad_norm": 1.6217183317438102,
      "learning_rate": 9.826300665645432e-06,
      "loss": 0.2814,
      "step": 660
    },
    {
      "epoch": 0.17861605878334472,
      "grad_norm": 1.8544819560192451,
      "learning_rate": 9.813930725744095e-06,
      "loss": 0.2852,
      "step": 670
    },
    {
      "epoch": 0.18128197010846928,
      "grad_norm": 1.7254305229833091,
      "learning_rate": 9.801143677485509e-06,
      "loss": 0.3016,
      "step": 680
    },
    {
      "epoch": 0.1839478814335938,
      "grad_norm": 1.76092048982559,
      "learning_rate": 9.787940628817627e-06,
      "loss": 0.3023,
      "step": 690
    },
    {
      "epoch": 0.18661379275871837,
      "grad_norm": 1.5976117922079591,
      "learning_rate": 9.774322723733216e-06,
      "loss": 0.2953,
      "step": 700
    },
    {
      "epoch": 0.18661379275871837,
      "eval_loss": 0.2970203757286072,
      "eval_runtime": 263.005,
      "eval_samples_per_second": 37.258,
      "eval_steps_per_second": 4.658,
      "step": 700
    },
    {
      "epoch": 0.1892797040838429,
      "grad_norm": 2.0364062908857505,
      "learning_rate": 9.760291142170739e-06,
      "loss": 0.2819,
      "step": 710
    },
    {
      "epoch": 0.19194561540896746,
      "grad_norm": 1.7860594933309035,
      "learning_rate": 9.745847099912116e-06,
      "loss": 0.2922,
      "step": 720
    },
    {
      "epoch": 0.19461152673409202,
      "grad_norm": 1.724488742261424,
      "learning_rate": 9.73099184847738e-06,
      "loss": 0.2936,
      "step": 730
    },
    {
      "epoch": 0.19727743805921655,
      "grad_norm": 2.0236324189364763,
      "learning_rate": 9.715726675016238e-06,
      "loss": 0.2852,
      "step": 740
    },
    {
      "epoch": 0.1999433493843411,
      "grad_norm": 1.5757420943087048,
      "learning_rate": 9.700052902196541e-06,
      "loss": 0.2854,
      "step": 750
    },
    {
      "epoch": 0.1999433493843411,
      "eval_loss": 0.292442262172699,
      "eval_runtime": 263.0027,
      "eval_samples_per_second": 37.258,
      "eval_steps_per_second": 4.658,
      "step": 750
    },
    {
      "epoch": 0.20260926070946564,
      "grad_norm": 1.6109796234381852,
      "learning_rate": 9.68397188808969e-06,
      "loss": 0.2911,
      "step": 760
    },
    {
      "epoch": 0.2052751720345902,
      "grad_norm": 1.7705435745788891,
      "learning_rate": 9.667485026052956e-06,
      "loss": 0.3003,
      "step": 770
    },
    {
      "epoch": 0.20794108335971476,
      "grad_norm": 1.8227445232808335,
      "learning_rate": 9.650593744608754e-06,
      "loss": 0.2954,
      "step": 780
    },
    {
      "epoch": 0.2106069946848393,
      "grad_norm": 1.6678622155458844,
      "learning_rate": 9.633299507320862e-06,
      "loss": 0.2921,
      "step": 790
    },
    {
      "epoch": 0.21327290600996385,
      "grad_norm": 1.7265641158944443,
      "learning_rate": 9.615603812667618e-06,
      "loss": 0.2872,
      "step": 800
    },
    {
      "epoch": 0.21327290600996385,
      "eval_loss": 0.2895732522010803,
      "eval_runtime": 263.041,
      "eval_samples_per_second": 37.253,
      "eval_steps_per_second": 4.657,
      "step": 800
    },
    {
      "epoch": 0.21593881733508838,
      "grad_norm": 1.7182998913074186,
      "learning_rate": 9.597508193912077e-06,
      "loss": 0.2907,
      "step": 810
    },
    {
      "epoch": 0.21860472866021294,
      "grad_norm": 1.5880777723000974,
      "learning_rate": 9.579014218969158e-06,
      "loss": 0.2867,
      "step": 820
    },
    {
      "epoch": 0.2212706399853375,
      "grad_norm": 1.7866067793206908,
      "learning_rate": 9.560123490269795e-06,
      "loss": 0.2942,
      "step": 830
    },
    {
      "epoch": 0.22393655131046203,
      "grad_norm": 1.5877926190764673,
      "learning_rate": 9.540837644622091e-06,
      "loss": 0.2832,
      "step": 840
    },
    {
      "epoch": 0.2266024626355866,
      "grad_norm": 1.9217023055863738,
      "learning_rate": 9.521158353069494e-06,
      "loss": 0.2866,
      "step": 850
    },
    {
      "epoch": 0.2266024626355866,
      "eval_loss": 0.28362876176834106,
      "eval_runtime": 263.2009,
      "eval_samples_per_second": 37.23,
      "eval_steps_per_second": 4.654,
      "step": 850
    },
    {
      "epoch": 0.22926837396071112,
      "grad_norm": 1.7253135194855898,
      "learning_rate": 9.501087320746007e-06,
      "loss": 0.2877,
      "step": 860
    },
    {
      "epoch": 0.23193428528583568,
      "grad_norm": 1.6645542489658331,
      "learning_rate": 9.480626286728445e-06,
      "loss": 0.2857,
      "step": 870
    },
    {
      "epoch": 0.23460019661096024,
      "grad_norm": 1.511024578102627,
      "learning_rate": 9.459777023885754e-06,
      "loss": 0.2839,
      "step": 880
    },
    {
      "epoch": 0.23726610793608477,
      "grad_norm": 1.4824802506716586,
      "learning_rate": 9.438541338725397e-06,
      "loss": 0.2833,
      "step": 890
    },
    {
      "epoch": 0.23993201926120933,
      "grad_norm": 1.564232445580215,
      "learning_rate": 9.416921071236821e-06,
      "loss": 0.2925,
      "step": 900
    },
    {
      "epoch": 0.23993201926120933,
      "eval_loss": 0.2794356942176819,
      "eval_runtime": 263.1403,
      "eval_samples_per_second": 37.239,
      "eval_steps_per_second": 4.655,
      "step": 900
    },
    {
      "epoch": 0.24259793058633386,
      "grad_norm": 1.812013202524589,
      "learning_rate": 9.394918094732044e-06,
      "loss": 0.2846,
      "step": 910
    },
    {
      "epoch": 0.24526384191145842,
      "grad_norm": 1.7480374081646133,
      "learning_rate": 9.37253431568332e-06,
      "loss": 0.2826,
      "step": 920
    },
    {
      "epoch": 0.24792975323658298,
      "grad_norm": 1.71680256902317,
      "learning_rate": 9.349771673557966e-06,
      "loss": 0.2856,
      "step": 930
    },
    {
      "epoch": 0.2505956645617075,
      "grad_norm": 1.6078777895847485,
      "learning_rate": 9.326632140650311e-06,
      "loss": 0.2886,
      "step": 940
    },
    {
      "epoch": 0.25326157588683207,
      "grad_norm": 1.4898994133826529,
      "learning_rate": 9.303117721910801e-06,
      "loss": 0.2843,
      "step": 950
    },
    {
      "epoch": 0.25326157588683207,
      "eval_loss": 0.28232333064079285,
      "eval_runtime": 262.9546,
      "eval_samples_per_second": 37.265,
      "eval_steps_per_second": 4.659,
      "step": 950
    },
    {
      "epoch": 0.25592748721195663,
      "grad_norm": 1.818300085431669,
      "learning_rate": 9.279230454772282e-06,
      "loss": 0.2758,
      "step": 960
    },
    {
      "epoch": 0.2585933985370812,
      "grad_norm": 1.5128664817372184,
      "learning_rate": 9.25497240897346e-06,
      "loss": 0.2789,
      "step": 970
    },
    {
      "epoch": 0.2612593098622057,
      "grad_norm": 1.6097431612586455,
      "learning_rate": 9.23034568637957e-06,
      "loss": 0.2854,
      "step": 980
    },
    {
      "epoch": 0.26392522118733025,
      "grad_norm": 1.4100117130863956,
      "learning_rate": 9.205352420800253e-06,
      "loss": 0.2857,
      "step": 990
    },
    {
      "epoch": 0.2665911325124548,
      "grad_norm": 1.6675059930867078,
      "learning_rate": 9.179994777804677e-06,
      "loss": 0.292,
      "step": 1000
    },
    {
      "epoch": 0.2665911325124548,
      "eval_loss": 0.27887627482414246,
      "eval_runtime": 262.9281,
      "eval_samples_per_second": 37.269,
      "eval_steps_per_second": 4.659,
      "step": 1000
    },
    {
      "epoch": 0.26925704383757937,
      "grad_norm": 1.644767282455157,
      "learning_rate": 9.154274954533895e-06,
      "loss": 0.2863,
      "step": 1010
    },
    {
      "epoch": 0.2719229551627039,
      "grad_norm": 1.6790096248660116,
      "learning_rate": 9.128195179510466e-06,
      "loss": 0.279,
      "step": 1020
    },
    {
      "epoch": 0.27458886648782843,
      "grad_norm": 1.583877941328967,
      "learning_rate": 9.101757712445369e-06,
      "loss": 0.2743,
      "step": 1030
    },
    {
      "epoch": 0.277254777812953,
      "grad_norm": 1.626963776268032,
      "learning_rate": 9.07496484404221e-06,
      "loss": 0.2757,
      "step": 1040
    },
    {
      "epoch": 0.27992068913807755,
      "grad_norm": 1.5034004223963462,
      "learning_rate": 9.04781889579873e-06,
      "loss": 0.2775,
      "step": 1050
    },
    {
      "epoch": 0.27992068913807755,
      "eval_loss": 0.27627232670783997,
      "eval_runtime": 262.9028,
      "eval_samples_per_second": 37.272,
      "eval_steps_per_second": 4.66,
      "step": 1050
    },
    {
      "epoch": 0.2825866004632021,
      "grad_norm": 1.4195880755569912,
      "learning_rate": 9.020322219805674e-06,
      "loss": 0.2803,
      "step": 1060
    },
    {
      "epoch": 0.28525251178832667,
      "grad_norm": 1.7563946494767981,
      "learning_rate": 8.99247719854297e-06,
      "loss": 0.2831,
      "step": 1070
    },
    {
      "epoch": 0.28791842311345117,
      "grad_norm": 1.4975000328635686,
      "learning_rate": 8.964286244673315e-06,
      "loss": 0.2794,
      "step": 1080
    },
    {
      "epoch": 0.29058433443857573,
      "grad_norm": 1.6808608884719483,
      "learning_rate": 8.935751800833117e-06,
      "loss": 0.2728,
      "step": 1090
    },
    {
      "epoch": 0.2932502457637003,
      "grad_norm": 1.5692764321165331,
      "learning_rate": 8.906876339420851e-06,
      "loss": 0.2652,
      "step": 1100
    },
    {
      "epoch": 0.2932502457637003,
      "eval_loss": 0.27173659205436707,
      "eval_runtime": 263.0328,
      "eval_samples_per_second": 37.254,
      "eval_steps_per_second": 4.657,
      "step": 1100
    },
    {
      "epoch": 0.29591615708882485,
      "grad_norm": 1.6549948875336085,
      "learning_rate": 8.877662362382844e-06,
      "loss": 0.2731,
      "step": 1110
    },
    {
      "epoch": 0.2985820684139494,
      "grad_norm": 1.6301740740445887,
      "learning_rate": 8.848112400996473e-06,
      "loss": 0.2669,
      "step": 1120
    },
    {
      "epoch": 0.3012479797390739,
      "grad_norm": 1.5648628679336365,
      "learning_rate": 8.818229015650862e-06,
      "loss": 0.274,
      "step": 1130
    },
    {
      "epoch": 0.30391389106419847,
      "grad_norm": 1.5356188950454304,
      "learning_rate": 8.788014795625018e-06,
      "loss": 0.275,
      "step": 1140
    },
    {
      "epoch": 0.30657980238932303,
      "grad_norm": 1.5817185861901346,
      "learning_rate": 8.757472358863481e-06,
      "loss": 0.27,
      "step": 1150
    },
    {
      "epoch": 0.30657980238932303,
      "eval_loss": 0.2712327241897583,
      "eval_runtime": 262.9142,
      "eval_samples_per_second": 37.271,
      "eval_steps_per_second": 4.659,
      "step": 1150
    },
    {
      "epoch": 0.3092457137144476,
      "grad_norm": 1.5194661669884653,
      "learning_rate": 8.726604351749503e-06,
      "loss": 0.2676,
      "step": 1160
    },
    {
      "epoch": 0.3119116250395721,
      "grad_norm": 1.7914357199240956,
      "learning_rate": 8.69541344887573e-06,
      "loss": 0.2779,
      "step": 1170
    },
    {
      "epoch": 0.31457753636469665,
      "grad_norm": 1.7310134858417519,
      "learning_rate": 8.66390235281248e-06,
      "loss": 0.2693,
      "step": 1180
    },
    {
      "epoch": 0.3172434476898212,
      "grad_norm": 1.5630957567521082,
      "learning_rate": 8.632073793873548e-06,
      "loss": 0.2709,
      "step": 1190
    },
    {
      "epoch": 0.31990935901494577,
      "grad_norm": 1.6939032554276248,
      "learning_rate": 8.599930529879669e-06,
      "loss": 0.277,
      "step": 1200
    },
    {
      "epoch": 0.31990935901494577,
      "eval_loss": 0.2749159336090088,
      "eval_runtime": 262.911,
      "eval_samples_per_second": 37.271,
      "eval_steps_per_second": 4.659,
      "step": 1200
    },
    {
      "epoch": 0.32257527034007033,
      "grad_norm": 1.5095047855589963,
      "learning_rate": 8.567475345919532e-06,
      "loss": 0.2629,
      "step": 1210
    },
    {
      "epoch": 0.32524118166519483,
      "grad_norm": 1.6080169624954084,
      "learning_rate": 8.534711054108487e-06,
      "loss": 0.2747,
      "step": 1220
    },
    {
      "epoch": 0.3279070929903194,
      "grad_norm": 1.3736048672777732,
      "learning_rate": 8.501640493344866e-06,
      "loss": 0.2644,
      "step": 1230
    },
    {
      "epoch": 0.33057300431544395,
      "grad_norm": 1.6744533289724879,
      "learning_rate": 8.468266529064025e-06,
      "loss": 0.263,
      "step": 1240
    },
    {
      "epoch": 0.3332389156405685,
      "grad_norm": 1.5519489984406099,
      "learning_rate": 8.434592052990044e-06,
      "loss": 0.2681,
      "step": 1250
    },
    {
      "epoch": 0.3332389156405685,
      "eval_loss": 0.2708764672279358,
      "eval_runtime": 262.898,
      "eval_samples_per_second": 37.273,
      "eval_steps_per_second": 4.66,
      "step": 1250
    },
    {
      "epoch": 0.33590482696569307,
      "grad_norm": 1.457281050568458,
      "learning_rate": 8.400619982885183e-06,
      "loss": 0.272,
      "step": 1260
    },
    {
      "epoch": 0.3385707382908176,
      "grad_norm": 1.5382553500774654,
      "learning_rate": 8.366353262297069e-06,
      "loss": 0.2713,
      "step": 1270
    },
    {
      "epoch": 0.34123664961594213,
      "grad_norm": 1.5058276908141497,
      "learning_rate": 8.331794860303644e-06,
      "loss": 0.2792,
      "step": 1280
    },
    {
      "epoch": 0.3439025609410667,
      "grad_norm": 1.5771918814364942,
      "learning_rate": 8.296947771255905e-06,
      "loss": 0.2628,
      "step": 1290
    },
    {
      "epoch": 0.34656847226619125,
      "grad_norm": 1.657853360665925,
      "learning_rate": 8.261815014518465e-06,
      "loss": 0.2699,
      "step": 1300
    },
    {
      "epoch": 0.34656847226619125,
      "eval_loss": 0.27181142568588257,
      "eval_runtime": 262.7648,
      "eval_samples_per_second": 37.292,
      "eval_steps_per_second": 4.662,
      "step": 1300
    },
    {
      "epoch": 0.3492343835913158,
      "grad_norm": 1.5883441926045438,
      "learning_rate": 8.226399634207929e-06,
      "loss": 0.2734,
      "step": 1310
    },
    {
      "epoch": 0.3519002949164403,
      "grad_norm": 1.5300996089830505,
      "learning_rate": 8.190704698929128e-06,
      "loss": 0.2638,
      "step": 1320
    },
    {
      "epoch": 0.3545662062415649,
      "grad_norm": 1.5262209102418558,
      "learning_rate": 8.154733301509249e-06,
      "loss": 0.2595,
      "step": 1330
    },
    {
      "epoch": 0.35723211756668943,
      "grad_norm": 1.6115009330913592,
      "learning_rate": 8.118488558729846e-06,
      "loss": 0.2663,
      "step": 1340
    },
    {
      "epoch": 0.359898028891814,
      "grad_norm": 1.5554372772477991,
      "learning_rate": 8.081973611056784e-06,
      "loss": 0.2682,
      "step": 1350
    },
    {
      "epoch": 0.359898028891814,
      "eval_loss": 0.267572283744812,
      "eval_runtime": 262.9095,
      "eval_samples_per_second": 37.271,
      "eval_steps_per_second": 4.659,
      "step": 1350
    },
    {
      "epoch": 0.36256394021693855,
      "grad_norm": 1.427915406398141,
      "learning_rate": 8.045191622368128e-06,
      "loss": 0.2593,
      "step": 1360
    },
    {
      "epoch": 0.36522985154206306,
      "grad_norm": 1.5334699622941028,
      "learning_rate": 8.008145779680011e-06,
      "loss": 0.2639,
      "step": 1370
    },
    {
      "epoch": 0.3678957628671876,
      "grad_norm": 1.4259972418658429,
      "learning_rate": 7.970839292870488e-06,
      "loss": 0.2669,
      "step": 1380
    },
    {
      "epoch": 0.3705616741923122,
      "grad_norm": 1.5749048798265741,
      "learning_rate": 7.933275394401407e-06,
      "loss": 0.2558,
      "step": 1390
    },
    {
      "epoch": 0.37322758551743673,
      "grad_norm": 1.4336549906080218,
      "learning_rate": 7.89545733903834e-06,
      "loss": 0.2668,
      "step": 1400
    },
    {
      "epoch": 0.37322758551743673,
      "eval_loss": 0.26621583104133606,
      "eval_runtime": 262.8653,
      "eval_samples_per_second": 37.278,
      "eval_steps_per_second": 4.66,
      "step": 1400
    },
    {
      "epoch": 0.3758934968425613,
      "grad_norm": 1.252003265223711,
      "learning_rate": 7.857388403568564e-06,
      "loss": 0.2562,
      "step": 1410
    },
    {
      "epoch": 0.3785594081676858,
      "grad_norm": 1.6246783307920207,
      "learning_rate": 7.819071886517134e-06,
      "loss": 0.2641,
      "step": 1420
    },
    {
      "epoch": 0.38122531949281036,
      "grad_norm": 1.4696921191370331,
      "learning_rate": 7.780511107861095e-06,
      "loss": 0.2574,
      "step": 1430
    },
    {
      "epoch": 0.3838912308179349,
      "grad_norm": 1.4689804757172065,
      "learning_rate": 7.741709408741804e-06,
      "loss": 0.2523,
      "step": 1440
    },
    {
      "epoch": 0.3865571421430595,
      "grad_norm": 1.464412667954816,
      "learning_rate": 7.702670151175435e-06,
      "loss": 0.2615,
      "step": 1450
    },
    {
      "epoch": 0.3865571421430595,
      "eval_loss": 0.26893940567970276,
      "eval_runtime": 262.8305,
      "eval_samples_per_second": 37.283,
      "eval_steps_per_second": 4.661,
      "step": 1450
    },
    {
      "epoch": 0.38922305346818403,
      "grad_norm": 1.4939931694247328,
      "learning_rate": 7.663396717761687e-06,
      "loss": 0.2598,
      "step": 1460
    },
    {
      "epoch": 0.39188896479330854,
      "grad_norm": 1.3658418367677023,
      "learning_rate": 7.6238925113906715e-06,
      "loss": 0.2595,
      "step": 1470
    },
    {
      "epoch": 0.3945548761184331,
      "grad_norm": 1.402377150997183,
      "learning_rate": 7.5841609549480854e-06,
      "loss": 0.2706,
      "step": 1480
    },
    {
      "epoch": 0.39722078744355765,
      "grad_norm": 1.6251726176323527,
      "learning_rate": 7.544205491018626e-06,
      "loss": 0.252,
      "step": 1490
    },
    {
      "epoch": 0.3998866987686822,
      "grad_norm": 1.4760120447972431,
      "learning_rate": 7.5040295815877e-06,
      "loss": 0.2501,
      "step": 1500
    },
    {
      "epoch": 0.3998866987686822,
      "eval_loss": 0.2582685649394989,
      "eval_runtime": 262.8368,
      "eval_samples_per_second": 37.282,
      "eval_steps_per_second": 4.661,
      "step": 1500
    },
    {
      "epoch": 0.4025526100938068,
      "grad_norm": 1.4743205374828412,
      "learning_rate": 7.463636707741458e-06,
      "loss": 0.2604,
      "step": 1510
    },
    {
      "epoch": 0.4052185214189313,
      "grad_norm": 1.3923700543379627,
      "learning_rate": 7.423030369365175e-06,
      "loss": 0.264,
      "step": 1520
    },
    {
      "epoch": 0.40788443274405584,
      "grad_norm": 1.4356363473684057,
      "learning_rate": 7.382214084839993e-06,
      "loss": 0.2565,
      "step": 1530
    },
    {
      "epoch": 0.4105503440691804,
      "grad_norm": 1.4051425897192589,
      "learning_rate": 7.341191390738073e-06,
      "loss": 0.2638,
      "step": 1540
    },
    {
      "epoch": 0.41321625539430495,
      "grad_norm": 1.3088487920660643,
      "learning_rate": 7.299965841516164e-06,
      "loss": 0.2545,
      "step": 1550
    },
    {
      "epoch": 0.41321625539430495,
      "eval_loss": 0.25675299763679504,
      "eval_runtime": 262.9203,
      "eval_samples_per_second": 37.27,
      "eval_steps_per_second": 4.659,
      "step": 1550
    },
    {
      "epoch": 0.4158821667194295,
      "grad_norm": 1.5688903433049168,
      "learning_rate": 7.2585410092076154e-06,
      "loss": 0.2637,
      "step": 1560
    },
    {
      "epoch": 0.418548078044554,
      "grad_norm": 1.431976384484085,
      "learning_rate": 7.216920483112886e-06,
      "loss": 0.2489,
      "step": 1570
    },
    {
      "epoch": 0.4212139893696786,
      "grad_norm": 1.531858850303692,
      "learning_rate": 7.175107869488539e-06,
      "loss": 0.2545,
      "step": 1580
    },
    {
      "epoch": 0.42387990069480314,
      "grad_norm": 1.4865553973552155,
      "learning_rate": 7.133106791234771e-06,
      "loss": 0.2644,
      "step": 1590
    },
    {
      "epoch": 0.4265458120199277,
      "grad_norm": 1.66021566236256,
      "learning_rate": 7.090920887581507e-06,
      "loss": 0.2618,
      "step": 1600
    },
    {
      "epoch": 0.4265458120199277,
      "eval_loss": 0.25225213170051575,
      "eval_runtime": 262.8963,
      "eval_samples_per_second": 37.273,
      "eval_steps_per_second": 4.66,
      "step": 1600
    },
    {
      "epoch": 0.42921172334505225,
      "grad_norm": 1.6133520399654793,
      "learning_rate": 7.048553813773075e-06,
      "loss": 0.246,
      "step": 1610
    },
    {
      "epoch": 0.43187763467017676,
      "grad_norm": 1.496263689865299,
      "learning_rate": 7.006009240751488e-06,
      "loss": 0.248,
      "step": 1620
    },
    {
      "epoch": 0.4345435459953013,
      "grad_norm": 1.529592079575296,
      "learning_rate": 6.963290854838376e-06,
      "loss": 0.2592,
      "step": 1630
    },
    {
      "epoch": 0.4372094573204259,
      "grad_norm": 1.3359842663344557,
      "learning_rate": 6.920402357415582e-06,
      "loss": 0.2502,
      "step": 1640
    },
    {
      "epoch": 0.43987536864555044,
      "grad_norm": 1.5115966603020272,
      "learning_rate": 6.877347464604446e-06,
      "loss": 0.2615,
      "step": 1650
    },
    {
      "epoch": 0.43987536864555044,
      "eval_loss": 0.2549818754196167,
      "eval_runtime": 263.068,
      "eval_samples_per_second": 37.249,
      "eval_steps_per_second": 4.657,
      "step": 1650
    },
    {
      "epoch": 0.442541279970675,
      "grad_norm": 1.7762192007454982,
      "learning_rate": 6.834129906943822e-06,
      "loss": 0.2468,
      "step": 1660
    },
    {
      "epoch": 0.4452071912957995,
      "grad_norm": 1.510769947566683,
      "learning_rate": 6.790753429066838e-06,
      "loss": 0.2522,
      "step": 1670
    },
    {
      "epoch": 0.44787310262092406,
      "grad_norm": 1.287630606180817,
      "learning_rate": 6.7472217893764465e-06,
      "loss": 0.2468,
      "step": 1680
    },
    {
      "epoch": 0.4505390139460486,
      "grad_norm": 1.5134316444240614,
      "learning_rate": 6.70353875971976e-06,
      "loss": 0.2525,
      "step": 1690
    },
    {
      "epoch": 0.4532049252711732,
      "grad_norm": 1.4296879650147967,
      "learning_rate": 6.659708125061242e-06,
      "loss": 0.2512,
      "step": 1700
    },
    {
      "epoch": 0.4532049252711732,
      "eval_loss": 0.24882382154464722,
      "eval_runtime": 262.8695,
      "eval_samples_per_second": 37.277,
      "eval_steps_per_second": 4.66,
      "step": 1700
    },
    {
      "epoch": 0.45587083659629773,
      "grad_norm": 1.4884981725217645,
      "learning_rate": 6.615733683154762e-06,
      "loss": 0.2522,
      "step": 1710
    },
    {
      "epoch": 0.45853674792142224,
      "grad_norm": 1.4903981491956688,
      "learning_rate": 6.571619244214521e-06,
      "loss": 0.2505,
      "step": 1720
    },
    {
      "epoch": 0.4612026592465468,
      "grad_norm": 1.3786999091258259,
      "learning_rate": 6.527368630584919e-06,
      "loss": 0.2596,
      "step": 1730
    },
    {
      "epoch": 0.46386857057167136,
      "grad_norm": 1.588650909339789,
      "learning_rate": 6.482985676409368e-06,
      "loss": 0.2494,
      "step": 1740
    },
    {
      "epoch": 0.4665344818967959,
      "grad_norm": 1.4631431734629796,
      "learning_rate": 6.438474227298065e-06,
      "loss": 0.245,
      "step": 1750
    },
    {
      "epoch": 0.4665344818967959,
      "eval_loss": 0.25038692355155945,
      "eval_runtime": 262.8028,
      "eval_samples_per_second": 37.287,
      "eval_steps_per_second": 4.661,
      "step": 1750
    },
    {
      "epoch": 0.4692003932219205,
      "grad_norm": 1.5412537109524729,
      "learning_rate": 6.393838139994797e-06,
      "loss": 0.2595,
      "step": 1760
    },
    {
      "epoch": 0.471866304547045,
      "grad_norm": 1.5353759046724331,
      "learning_rate": 6.349081282042768e-06,
      "loss": 0.2516,
      "step": 1770
    },
    {
      "epoch": 0.47453221587216954,
      "grad_norm": 1.783528374337059,
      "learning_rate": 6.304207531449486e-06,
      "loss": 0.2469,
      "step": 1780
    },
    {
      "epoch": 0.4771981271972941,
      "grad_norm": 1.5629443542703378,
      "learning_rate": 6.259220776350746e-06,
      "loss": 0.2354,
      "step": 1790
    },
    {
      "epoch": 0.47986403852241866,
      "grad_norm": 1.5839504852392476,
      "learning_rate": 6.2141249146737545e-06,
      "loss": 0.2503,
      "step": 1800
    },
    {
      "epoch": 0.47986403852241866,
      "eval_loss": 0.24808603525161743,
      "eval_runtime": 262.9156,
      "eval_samples_per_second": 37.271,
      "eval_steps_per_second": 4.659,
      "step": 1800
    },
    {
      "epoch": 0.4825299498475432,
      "grad_norm": 1.2856913499546363,
      "learning_rate": 6.168923853799369e-06,
      "loss": 0.2466,
      "step": 1810
    },
    {
      "epoch": 0.4851958611726677,
      "grad_norm": 1.5335554959397042,
      "learning_rate": 6.123621510223552e-06,
      "loss": 0.2467,
      "step": 1820
    },
    {
      "epoch": 0.4878617724977923,
      "grad_norm": 1.4473834266997454,
      "learning_rate": 6.0782218092180164e-06,
      "loss": 0.2523,
      "step": 1830
    },
    {
      "epoch": 0.49052768382291684,
      "grad_norm": 1.394637034197444,
      "learning_rate": 6.032728684490118e-06,
      "loss": 0.2428,
      "step": 1840
    },
    {
      "epoch": 0.4931935951480414,
      "grad_norm": 1.468907580360217,
      "learning_rate": 5.987146077842015e-06,
      "loss": 0.2402,
      "step": 1850
    },
    {
      "epoch": 0.4931935951480414,
      "eval_loss": 0.24496783316135406,
      "eval_runtime": 262.8614,
      "eval_samples_per_second": 37.278,
      "eval_steps_per_second": 4.66,
      "step": 1850
    },
    {
      "epoch": 0.49585950647316596,
      "grad_norm": 1.3904834622458775,
      "learning_rate": 5.941477938829126e-06,
      "loss": 0.2526,
      "step": 1860
    },
    {
      "epoch": 0.49852541779829046,
      "grad_norm": 1.4889823952197347,
      "learning_rate": 5.8957282244179125e-06,
      "loss": 0.2462,
      "step": 1870
    },
    {
      "epoch": 0.501191329123415,
      "grad_norm": 1.5388317154311413,
      "learning_rate": 5.84990089864303e-06,
      "loss": 0.2343,
      "step": 1880
    },
    {
      "epoch": 0.5038572404485396,
      "grad_norm": 1.4480277186406976,
      "learning_rate": 5.803999932263859e-06,
      "loss": 0.2403,
      "step": 1890
    },
    {
      "epoch": 0.5065231517736641,
      "grad_norm": 1.386625092855129,
      "learning_rate": 5.7580293024204455e-06,
      "loss": 0.2346,
      "step": 1900
    },
    {
      "epoch": 0.5065231517736641,
      "eval_loss": 0.24397991597652435,
      "eval_runtime": 263.0578,
      "eval_samples_per_second": 37.25,
      "eval_steps_per_second": 4.657,
      "step": 1900
    },
    {
      "epoch": 0.5091890630987886,
      "grad_norm": 1.4313341683411982,
      "learning_rate": 5.7119929922889065e-06,
      "loss": 0.251,
      "step": 1910
    },
    {
      "epoch": 0.5118549744239133,
      "grad_norm": 1.4840350228738965,
      "learning_rate": 5.665894990736301e-06,
      "loss": 0.2443,
      "step": 1920
    },
    {
      "epoch": 0.5145208857490378,
      "grad_norm": 1.4825583514024858,
      "learning_rate": 5.6197392919750095e-06,
      "loss": 0.2492,
      "step": 1930
    },
    {
      "epoch": 0.5171867970741624,
      "grad_norm": 1.309952322214834,
      "learning_rate": 5.573529895216648e-06,
      "loss": 0.2472,
      "step": 1940
    },
    {
      "epoch": 0.5198527083992869,
      "grad_norm": 1.4272349945023302,
      "learning_rate": 5.5272708043255605e-06,
      "loss": 0.2413,
      "step": 1950
    },
    {
      "epoch": 0.5198527083992869,
      "eval_loss": 0.24250540137290955,
      "eval_runtime": 262.9015,
      "eval_samples_per_second": 37.273,
      "eval_steps_per_second": 4.66,
      "step": 1950
    },
    {
      "epoch": 0.5225186197244114,
      "grad_norm": 1.3899478056194363,
      "learning_rate": 5.480966027471889e-06,
      "loss": 0.237,
      "step": 1960
    },
    {
      "epoch": 0.525184531049536,
      "grad_norm": 1.4329413761949321,
      "learning_rate": 5.434619576784288e-06,
      "loss": 0.2449,
      "step": 1970
    },
    {
      "epoch": 0.5278504423746605,
      "grad_norm": 1.4472221948239792,
      "learning_rate": 5.388235468002286e-06,
      "loss": 0.2237,
      "step": 1980
    },
    {
      "epoch": 0.5305163536997851,
      "grad_norm": 1.4195674353945604,
      "learning_rate": 5.341817720128344e-06,
      "loss": 0.2454,
      "step": 1990
    },
    {
      "epoch": 0.5331822650249096,
      "grad_norm": 1.3103614120039029,
      "learning_rate": 5.295370355079615e-06,
      "loss": 0.24,
      "step": 2000
    },
    {
      "epoch": 0.5331822650249096,
      "eval_loss": 0.2383483648300171,
      "eval_runtime": 262.9776,
      "eval_samples_per_second": 37.262,
      "eval_steps_per_second": 4.658,
      "step": 2000
    },
    {
      "epoch": 0.5358481763500341,
      "grad_norm": 1.3133683565273986,
      "learning_rate": 5.248897397339462e-06,
      "loss": 0.2373,
      "step": 2010
    },
    {
      "epoch": 0.5385140876751587,
      "grad_norm": 1.531592053829392,
      "learning_rate": 5.202402873608763e-06,
      "loss": 0.2355,
      "step": 2020
    },
    {
      "epoch": 0.5411799990002832,
      "grad_norm": 1.3614359816256922,
      "learning_rate": 5.155890812457e-06,
      "loss": 0.2415,
      "step": 2030
    },
    {
      "epoch": 0.5438459103254079,
      "grad_norm": 1.3385718906961956,
      "learning_rate": 5.1093652439732036e-06,
      "loss": 0.2319,
      "step": 2040
    },
    {
      "epoch": 0.5465118216505324,
      "grad_norm": 1.3010476401905142,
      "learning_rate": 5.062830199416764e-06,
      "loss": 0.2398,
      "step": 2050
    },
    {
      "epoch": 0.5465118216505324,
      "eval_loss": 0.24081052839756012,
      "eval_runtime": 262.8874,
      "eval_samples_per_second": 37.275,
      "eval_steps_per_second": 4.66,
      "step": 2050
    },
    {
      "epoch": 0.5491777329756569,
      "grad_norm": 1.364305498532448,
      "learning_rate": 5.0162897108681375e-06,
      "loss": 0.2393,
      "step": 2060
    },
    {
      "epoch": 0.5518436443007815,
      "grad_norm": 1.500947573673284,
      "learning_rate": 4.969747810879478e-06,
      "loss": 0.2447,
      "step": 2070
    },
    {
      "epoch": 0.554509555625906,
      "grad_norm": 1.320594568969342,
      "learning_rate": 4.923208532125235e-06,
      "loss": 0.2342,
      "step": 2080
    },
    {
      "epoch": 0.5571754669510306,
      "grad_norm": 1.5324196826217082,
      "learning_rate": 4.876675907052745e-06,
      "loss": 0.2355,
      "step": 2090
    },
    {
      "epoch": 0.5598413782761551,
      "grad_norm": 1.4300989151801904,
      "learning_rate": 4.8301539675328205e-06,
      "loss": 0.2473,
      "step": 2100
    },
    {
      "epoch": 0.5598413782761551,
      "eval_loss": 0.23838448524475098,
      "eval_runtime": 262.9492,
      "eval_samples_per_second": 37.266,
      "eval_steps_per_second": 4.659,
      "step": 2100
    },
    {
      "epoch": 0.5625072896012796,
      "grad_norm": 1.4958278476823126,
      "learning_rate": 4.783646744510416e-06,
      "loss": 0.2202,
      "step": 2110
    },
    {
      "epoch": 0.5651732009264042,
      "grad_norm": 1.4255485000398498,
      "learning_rate": 4.737158267655358e-06,
      "loss": 0.2367,
      "step": 2120
    },
    {
      "epoch": 0.5678391122515287,
      "grad_norm": 1.320727510403474,
      "learning_rate": 4.690692565013193e-06,
      "loss": 0.2355,
      "step": 2130
    },
    {
      "epoch": 0.5705050235766533,
      "grad_norm": 1.3566414156014617,
      "learning_rate": 4.644253662656167e-06,
      "loss": 0.2306,
      "step": 2140
    },
    {
      "epoch": 0.5731709349017778,
      "grad_norm": 1.3929922145408924,
      "learning_rate": 4.597845584334387e-06,
      "loss": 0.2423,
      "step": 2150
    },
    {
      "epoch": 0.5731709349017778,
      "eval_loss": 0.234757199883461,
      "eval_runtime": 262.9076,
      "eval_samples_per_second": 37.272,
      "eval_steps_per_second": 4.659,
      "step": 2150
    },
    {
      "epoch": 0.5758368462269023,
      "grad_norm": 1.374609479412353,
      "learning_rate": 4.55147235112718e-06,
      "loss": 0.2273,
      "step": 2160
    },
    {
      "epoch": 0.578502757552027,
      "grad_norm": 1.4785794464229693,
      "learning_rate": 4.505137981094675e-06,
      "loss": 0.2343,
      "step": 2170
    },
    {
      "epoch": 0.5811686688771515,
      "grad_norm": 1.4795260893327034,
      "learning_rate": 4.458846488929663e-06,
      "loss": 0.2324,
      "step": 2180
    },
    {
      "epoch": 0.5838345802022761,
      "grad_norm": 1.401523269373386,
      "learning_rate": 4.41260188560973e-06,
      "loss": 0.2345,
      "step": 2190
    },
    {
      "epoch": 0.5865004915274006,
      "grad_norm": 1.476962877773437,
      "learning_rate": 4.366408178049728e-06,
      "loss": 0.2294,
      "step": 2200
    },
    {
      "epoch": 0.5865004915274006,
      "eval_loss": 0.23114009201526642,
      "eval_runtime": 262.8457,
      "eval_samples_per_second": 37.28,
      "eval_steps_per_second": 4.661,
      "step": 2200
    },
    {
      "epoch": 0.5891664028525251,
      "grad_norm": 1.4302372289930234,
      "learning_rate": 4.32026936875459e-06,
      "loss": 0.2276,
      "step": 2210
    },
    {
      "epoch": 0.5918323141776497,
      "grad_norm": 1.4473961518449265,
      "learning_rate": 4.274189455472529e-06,
      "loss": 0.2273,
      "step": 2220
    },
    {
      "epoch": 0.5944982255027742,
      "grad_norm": 1.4008641287788706,
      "learning_rate": 4.228172430848645e-06,
      "loss": 0.2288,
      "step": 2230
    },
    {
      "epoch": 0.5971641368278988,
      "grad_norm": 1.4894210929250196,
      "learning_rate": 4.182222282078983e-06,
      "loss": 0.2357,
      "step": 2240
    },
    {
      "epoch": 0.5998300481530233,
      "grad_norm": 1.306703863087604,
      "learning_rate": 4.136342990565055e-06,
      "loss": 0.2403,
      "step": 2250
    },
    {
      "epoch": 0.5998300481530233,
      "eval_loss": 0.23227499425411224,
      "eval_runtime": 262.8907,
      "eval_samples_per_second": 37.274,
      "eval_steps_per_second": 4.66,
      "step": 2250
    },
    {
      "epoch": 0.6024959594781478,
      "grad_norm": 1.4013172843098054,
      "learning_rate": 4.090538531568867e-06,
      "loss": 0.2296,
      "step": 2260
    },
    {
      "epoch": 0.6051618708032724,
      "grad_norm": 1.3209487720250683,
      "learning_rate": 4.0448128738684775e-06,
      "loss": 0.2328,
      "step": 2270
    },
    {
      "epoch": 0.6078277821283969,
      "grad_norm": 1.4716940731508306,
      "learning_rate": 3.999169979414123e-06,
      "loss": 0.2325,
      "step": 2280
    },
    {
      "epoch": 0.6104936934535216,
      "grad_norm": 1.4656992929067083,
      "learning_rate": 3.9536138029849244e-06,
      "loss": 0.2168,
      "step": 2290
    },
    {
      "epoch": 0.6131596047786461,
      "grad_norm": 1.3738056017255034,
      "learning_rate": 3.908148291846225e-06,
      "loss": 0.2319,
      "step": 2300
    },
    {
      "epoch": 0.6131596047786461,
      "eval_loss": 0.22970426082611084,
      "eval_runtime": 262.999,
      "eval_samples_per_second": 37.259,
      "eval_steps_per_second": 4.658,
      "step": 2300
    },
    {
      "epoch": 0.6158255161037706,
      "grad_norm": 1.5466991422842942,
      "learning_rate": 3.862777385407569e-06,
      "loss": 0.2341,
      "step": 2310
    },
    {
      "epoch": 0.6184914274288952,
      "grad_norm": 1.3112376666812904,
      "learning_rate": 3.817505014881378e-06,
      "loss": 0.2233,
      "step": 2320
    },
    {
      "epoch": 0.6211573387540197,
      "grad_norm": 1.2953029619867271,
      "learning_rate": 3.7723351029423143e-06,
      "loss": 0.2267,
      "step": 2330
    },
    {
      "epoch": 0.6238232500791442,
      "grad_norm": 1.2975930501539998,
      "learning_rate": 3.72727156338741e-06,
      "loss": 0.2217,
      "step": 2340
    },
    {
      "epoch": 0.6264891614042688,
      "grad_norm": 1.450217813101867,
      "learning_rate": 3.6823183007969375e-06,
      "loss": 0.222,
      "step": 2350
    },
    {
      "epoch": 0.6264891614042688,
      "eval_loss": 0.22879666090011597,
      "eval_runtime": 262.998,
      "eval_samples_per_second": 37.259,
      "eval_steps_per_second": 4.658,
      "step": 2350
    },
    {
      "epoch": 0.6291550727293933,
      "grad_norm": 1.4772827053597228,
      "learning_rate": 3.637479210196102e-06,
      "loss": 0.224,
      "step": 2360
    },
    {
      "epoch": 0.6318209840545179,
      "grad_norm": 1.3820541090280336,
      "learning_rate": 3.59275817671755e-06,
      "loss": 0.2324,
      "step": 2370
    },
    {
      "epoch": 0.6344868953796424,
      "grad_norm": 1.6023213257482982,
      "learning_rate": 3.548159075264738e-06,
      "loss": 0.2402,
      "step": 2380
    },
    {
      "epoch": 0.6371528067047669,
      "grad_norm": 1.3905766577987069,
      "learning_rate": 3.5036857701761857e-06,
      "loss": 0.2296,
      "step": 2390
    },
    {
      "epoch": 0.6398187180298915,
      "grad_norm": 1.1561354043825967,
      "learning_rate": 3.4593421148906523e-06,
      "loss": 0.2193,
      "step": 2400
    },
    {
      "epoch": 0.6398187180298915,
      "eval_loss": 0.2302982062101364,
      "eval_runtime": 262.8211,
      "eval_samples_per_second": 37.284,
      "eval_steps_per_second": 4.661,
      "step": 2400
    },
    {
      "epoch": 0.642484629355016,
      "grad_norm": 1.597294433505113,
      "learning_rate": 3.4151319516132414e-06,
      "loss": 0.2241,
      "step": 2410
    },
    {
      "epoch": 0.6451505406801407,
      "grad_norm": 1.3695543102988872,
      "learning_rate": 3.3710591109824954e-06,
      "loss": 0.228,
      "step": 2420
    },
    {
      "epoch": 0.6478164520052652,
      "grad_norm": 1.4718036858879708,
      "learning_rate": 3.3271274117384834e-06,
      "loss": 0.2248,
      "step": 2430
    },
    {
      "epoch": 0.6504823633303897,
      "grad_norm": 1.4542440170222004,
      "learning_rate": 3.2833406603919243e-06,
      "loss": 0.2255,
      "step": 2440
    },
    {
      "epoch": 0.6531482746555143,
      "grad_norm": 1.4323637810184542,
      "learning_rate": 3.239702650894364e-06,
      "loss": 0.2252,
      "step": 2450
    },
    {
      "epoch": 0.6531482746555143,
      "eval_loss": 0.22466669976711273,
      "eval_runtime": 263.0427,
      "eval_samples_per_second": 37.253,
      "eval_steps_per_second": 4.657,
      "step": 2450
    },
    {
      "epoch": 0.6558141859806388,
      "grad_norm": 1.3508621042110618,
      "learning_rate": 3.1962171643094474e-06,
      "loss": 0.2262,
      "step": 2460
    },
    {
      "epoch": 0.6584800973057634,
      "grad_norm": 1.5246150137269523,
      "learning_rate": 3.152887968485303e-06,
      "loss": 0.2268,
      "step": 2470
    },
    {
      "epoch": 0.6611460086308879,
      "grad_norm": 1.383892905949888,
      "learning_rate": 3.1097188177280735e-06,
      "loss": 0.2307,
      "step": 2480
    },
    {
      "epoch": 0.6638119199560124,
      "grad_norm": 1.4496544374013476,
      "learning_rate": 3.0667134524766173e-06,
      "loss": 0.2306,
      "step": 2490
    },
    {
      "epoch": 0.666477831281137,
      "grad_norm": 1.3608828390672776,
      "learning_rate": 3.023875598978419e-06,
      "loss": 0.2304,
      "step": 2500
    },
    {
      "epoch": 0.666477831281137,
      "eval_loss": 0.22111880779266357,
      "eval_runtime": 263.0181,
      "eval_samples_per_second": 37.256,
      "eval_steps_per_second": 4.657,
      "step": 2500
    },
    {
      "epoch": 0.6691437426062615,
      "grad_norm": 1.4266456422805869,
      "learning_rate": 2.981208968966721e-06,
      "loss": 0.2308,
      "step": 2510
    },
    {
      "epoch": 0.6718096539313861,
      "grad_norm": 1.4406082869964068,
      "learning_rate": 2.9387172593389148e-06,
      "loss": 0.2279,
      "step": 2520
    },
    {
      "epoch": 0.6744755652565106,
      "grad_norm": 1.3971739179858862,
      "learning_rate": 2.896404151836227e-06,
      "loss": 0.2269,
      "step": 2530
    },
    {
      "epoch": 0.6771414765816351,
      "grad_norm": 1.4152700853128788,
      "learning_rate": 2.8542733127247024e-06,
      "loss": 0.2302,
      "step": 2540
    },
    {
      "epoch": 0.6798073879067598,
      "grad_norm": 1.6872576664694825,
      "learning_rate": 2.812328392477536e-06,
      "loss": 0.2139,
      "step": 2550
    },
    {
      "epoch": 0.6798073879067598,
      "eval_loss": 0.2199471890926361,
      "eval_runtime": 263.0108,
      "eval_samples_per_second": 37.257,
      "eval_steps_per_second": 4.658,
      "step": 2550
    },
    {
      "epoch": 0.6824732992318843,
      "grad_norm": 1.7126358001487447,
      "learning_rate": 2.7705730254587802e-06,
      "loss": 0.2237,
      "step": 2560
    },
    {
      "epoch": 0.6851392105570089,
      "grad_norm": 1.3640091375946775,
      "learning_rate": 2.729010829608442e-06,
      "loss": 0.2166,
      "step": 2570
    },
    {
      "epoch": 0.6878051218821334,
      "grad_norm": 1.3205123742802625,
      "learning_rate": 2.6876454061289892e-06,
      "loss": 0.2291,
      "step": 2580
    },
    {
      "epoch": 0.6904710332072579,
      "grad_norm": 1.41857906965664,
      "learning_rate": 2.646480339173337e-06,
      "loss": 0.2226,
      "step": 2590
    },
    {
      "epoch": 0.6931369445323825,
      "grad_norm": 1.4353906946512758,
      "learning_rate": 2.6055191955342886e-06,
      "loss": 0.2186,
      "step": 2600
    },
    {
      "epoch": 0.6931369445323825,
      "eval_loss": 0.219247967004776,
      "eval_runtime": 262.8123,
      "eval_samples_per_second": 37.285,
      "eval_steps_per_second": 4.661,
      "step": 2600
    },
    {
      "epoch": 0.695802855857507,
      "grad_norm": 1.509318119256484,
      "learning_rate": 2.564765524335478e-06,
      "loss": 0.2249,
      "step": 2610
    },
    {
      "epoch": 0.6984687671826316,
      "grad_norm": 1.5006689440254304,
      "learning_rate": 2.524222856723869e-06,
      "loss": 0.2244,
      "step": 2620
    },
    {
      "epoch": 0.7011346785077561,
      "grad_norm": 1.4381458956509503,
      "learning_rate": 2.483894705563778e-06,
      "loss": 0.2168,
      "step": 2630
    },
    {
      "epoch": 0.7038005898328806,
      "grad_norm": 1.4917726610546065,
      "learning_rate": 2.4437845651325116e-06,
      "loss": 0.2175,
      "step": 2640
    },
    {
      "epoch": 0.7064665011580052,
      "grad_norm": 1.3982582754588275,
      "learning_rate": 2.403895910817593e-06,
      "loss": 0.2156,
      "step": 2650
    },
    {
      "epoch": 0.7064665011580052,
      "eval_loss": 0.21830175817012787,
      "eval_runtime": 262.8854,
      "eval_samples_per_second": 37.275,
      "eval_steps_per_second": 4.66,
      "step": 2650
    },
    {
      "epoch": 0.7091324124831297,
      "grad_norm": 1.5108046870677676,
      "learning_rate": 2.364232198815638e-06,
      "loss": 0.2111,
      "step": 2660
    },
    {
      "epoch": 0.7117983238082544,
      "grad_norm": 1.3667591912176449,
      "learning_rate": 2.3247968658328825e-06,
      "loss": 0.2224,
      "step": 2670
    },
    {
      "epoch": 0.7144642351333789,
      "grad_norm": 1.4017600569153348,
      "learning_rate": 2.285593328787414e-06,
      "loss": 0.213,
      "step": 2680
    },
    {
      "epoch": 0.7171301464585034,
      "grad_norm": 1.440902705270559,
      "learning_rate": 2.246624984513099e-06,
      "loss": 0.2283,
      "step": 2690
    },
    {
      "epoch": 0.719796057783628,
      "grad_norm": 1.236716519182564,
      "learning_rate": 2.2078952094652705e-06,
      "loss": 0.2187,
      "step": 2700
    },
    {
      "epoch": 0.719796057783628,
      "eval_loss": 0.21585261821746826,
      "eval_runtime": 262.7547,
      "eval_samples_per_second": 37.293,
      "eval_steps_per_second": 4.662,
      "step": 2700
    },
    {
      "epoch": 0.7224619691087525,
      "grad_norm": 1.3538906092368184,
      "learning_rate": 2.1694073594281663e-06,
      "loss": 0.2221,
      "step": 2710
    },
    {
      "epoch": 0.7251278804338771,
      "grad_norm": 1.5194150460697764,
      "learning_rate": 2.131164769224164e-06,
      "loss": 0.2192,
      "step": 2720
    },
    {
      "epoch": 0.7277937917590016,
      "grad_norm": 1.3384474695172988,
      "learning_rate": 2.0931707524248268e-06,
      "loss": 0.2249,
      "step": 2730
    },
    {
      "epoch": 0.7304597030841261,
      "grad_norm": 1.5838731189086506,
      "learning_rate": 2.0554286010638076e-06,
      "loss": 0.2112,
      "step": 2740
    },
    {
      "epoch": 0.7331256144092507,
      "grad_norm": 1.569671558314134,
      "learning_rate": 2.017941585351591e-06,
      "loss": 0.222,
      "step": 2750
    },
    {
      "epoch": 0.7331256144092507,
      "eval_loss": 0.21738137304782867,
      "eval_runtime": 262.8306,
      "eval_samples_per_second": 37.283,
      "eval_steps_per_second": 4.661,
      "step": 2750
    },
    {
      "epoch": 0.7357915257343752,
      "grad_norm": 1.3524416472787477,
      "learning_rate": 1.98071295339216e-06,
      "loss": 0.2091,
      "step": 2760
    },
    {
      "epoch": 0.7384574370594998,
      "grad_norm": 1.360747048278261,
      "learning_rate": 1.9437459309015426e-06,
      "loss": 0.2133,
      "step": 2770
    },
    {
      "epoch": 0.7411233483846243,
      "grad_norm": 1.378087319514889,
      "learning_rate": 1.9070437209283304e-06,
      "loss": 0.2102,
      "step": 2780
    },
    {
      "epoch": 0.7437892597097489,
      "grad_norm": 1.3569822983158015,
      "learning_rate": 1.8706095035761418e-06,
      "loss": 0.2138,
      "step": 2790
    },
    {
      "epoch": 0.7464551710348735,
      "grad_norm": 1.3617876179506314,
      "learning_rate": 1.8344464357280722e-06,
      "loss": 0.2162,
      "step": 2800
    },
    {
      "epoch": 0.7464551710348735,
      "eval_loss": 0.21530871093273163,
      "eval_runtime": 262.8758,
      "eval_samples_per_second": 37.276,
      "eval_steps_per_second": 4.66,
      "step": 2800
    },
    {
      "epoch": 0.749121082359998,
      "grad_norm": 1.3465206929111626,
      "learning_rate": 1.7985576507731744e-06,
      "loss": 0.2245,
      "step": 2810
    },
    {
      "epoch": 0.7517869936851226,
      "grad_norm": 1.4990944824718522,
      "learning_rate": 1.762946258334951e-06,
      "loss": 0.2196,
      "step": 2820
    },
    {
      "epoch": 0.7544529050102471,
      "grad_norm": 1.2637615882614495,
      "learning_rate": 1.727615344001926e-06,
      "loss": 0.2119,
      "step": 2830
    },
    {
      "epoch": 0.7571188163353716,
      "grad_norm": 1.3954340785154413,
      "learning_rate": 1.6925679690602876e-06,
      "loss": 0.2106,
      "step": 2840
    },
    {
      "epoch": 0.7597847276604962,
      "grad_norm": 1.390096178394802,
      "learning_rate": 1.6578071702286396e-06,
      "loss": 0.2253,
      "step": 2850
    },
    {
      "epoch": 0.7597847276604962,
      "eval_loss": 0.21321707963943481,
      "eval_runtime": 263.0359,
      "eval_samples_per_second": 37.253,
      "eval_steps_per_second": 4.657,
      "step": 2850
    },
    {
      "epoch": 0.7624506389856207,
      "grad_norm": 1.3401955453773768,
      "learning_rate": 1.6233359593948777e-06,
      "loss": 0.22,
      "step": 2860
    },
    {
      "epoch": 0.7651165503107453,
      "grad_norm": 1.287854168515364,
      "learning_rate": 1.5891573233552315e-06,
      "loss": 0.2133,
      "step": 2870
    },
    {
      "epoch": 0.7677824616358698,
      "grad_norm": 1.3415313412375607,
      "learning_rate": 1.5552742235554551e-06,
      "loss": 0.2161,
      "step": 2880
    },
    {
      "epoch": 0.7704483729609943,
      "grad_norm": 1.4355286266779297,
      "learning_rate": 1.521689595834246e-06,
      "loss": 0.2129,
      "step": 2890
    },
    {
      "epoch": 0.773114284286119,
      "grad_norm": 1.3269614788833008,
      "learning_rate": 1.4884063501688539e-06,
      "loss": 0.2066,
      "step": 2900
    },
    {
      "epoch": 0.773114284286119,
      "eval_loss": 0.21342259645462036,
      "eval_runtime": 262.8386,
      "eval_samples_per_second": 37.281,
      "eval_steps_per_second": 4.661,
      "step": 2900
    },
    {
      "epoch": 0.7757801956112435,
      "grad_norm": 1.4582052603751765,
      "learning_rate": 1.4554273704229494e-06,
      "loss": 0.2244,
      "step": 2910
    },
    {
      "epoch": 0.7784461069363681,
      "grad_norm": 1.1948855303191914,
      "learning_rate": 1.4227555140967402e-06,
      "loss": 0.2156,
      "step": 2920
    },
    {
      "epoch": 0.7811120182614926,
      "grad_norm": 1.5589564581113802,
      "learning_rate": 1.3903936120793926e-06,
      "loss": 0.2192,
      "step": 2930
    },
    {
      "epoch": 0.7837779295866171,
      "grad_norm": 1.3051867211369865,
      "learning_rate": 1.3583444684037312e-06,
      "loss": 0.2156,
      "step": 2940
    },
    {
      "epoch": 0.7864438409117417,
      "grad_norm": 1.3423994247273423,
      "learning_rate": 1.3266108600032928e-06,
      "loss": 0.2113,
      "step": 2950
    },
    {
      "epoch": 0.7864438409117417,
      "eval_loss": 0.2107125222682953,
      "eval_runtime": 263.0418,
      "eval_samples_per_second": 37.253,
      "eval_steps_per_second": 4.657,
      "step": 2950
    },
    {
      "epoch": 0.7891097522368662,
      "grad_norm": 1.3712137564595037,
      "learning_rate": 1.2951955364717116e-06,
      "loss": 0.2143,
      "step": 2960
    },
    {
      "epoch": 0.7917756635619908,
      "grad_norm": 1.3657856665947792,
      "learning_rate": 1.2641012198244718e-06,
      "loss": 0.2065,
      "step": 2970
    },
    {
      "epoch": 0.7944415748871153,
      "grad_norm": 1.21692203839909,
      "learning_rate": 1.2333306042630672e-06,
      "loss": 0.2201,
      "step": 2980
    },
    {
      "epoch": 0.7971074862122398,
      "grad_norm": 1.4562111256731463,
      "learning_rate": 1.202886355941546e-06,
      "loss": 0.2097,
      "step": 2990
    },
    {
      "epoch": 0.7997733975373644,
      "grad_norm": 1.42678698336257,
      "learning_rate": 1.1727711127355118e-06,
      "loss": 0.2107,
      "step": 3000
    },
    {
      "epoch": 0.7997733975373644,
      "eval_loss": 0.20849083364009857,
      "eval_runtime": 262.8245,
      "eval_samples_per_second": 37.283,
      "eval_steps_per_second": 4.661,
      "step": 3000
    },
    {
      "epoch": 0.8024393088624889,
      "grad_norm": 1.6939593357963476,
      "learning_rate": 1.1429874840135492e-06,
      "loss": 0.2126,
      "step": 3010
    },
    {
      "epoch": 0.8051052201876135,
      "grad_norm": 1.365257481621389,
      "learning_rate": 1.1135380504111476e-06,
      "loss": 0.2024,
      "step": 3020
    },
    {
      "epoch": 0.807771131512738,
      "grad_norm": 1.4085575069668772,
      "learning_rate": 1.0844253636070805e-06,
      "loss": 0.2103,
      "step": 3030
    },
    {
      "epoch": 0.8104370428378626,
      "grad_norm": 1.4997581763477477,
      "learning_rate": 1.0556519461023301e-06,
      "loss": 0.2093,
      "step": 3040
    },
    {
      "epoch": 0.8131029541629872,
      "grad_norm": 1.384763329857118,
      "learning_rate": 1.0272202910015083e-06,
      "loss": 0.2055,
      "step": 3050
    },
    {
      "epoch": 0.8131029541629872,
      "eval_loss": 0.2096840888261795,
      "eval_runtime": 262.8819,
      "eval_samples_per_second": 37.275,
      "eval_steps_per_second": 4.66,
      "step": 3050
    },
    {
      "epoch": 0.8157688654881117,
      "grad_norm": 1.4019463660561169,
      "learning_rate": 9.991328617968509e-07,
      "loss": 0.2084,
      "step": 3060
    },
    {
      "epoch": 0.8184347768132363,
      "grad_norm": 1.4903762495669879,
      "learning_rate": 9.713920921547532e-07,
      "loss": 0.2076,
      "step": 3070
    },
    {
      "epoch": 0.8211006881383608,
      "grad_norm": 1.4296247263023294,
      "learning_rate": 9.440003857049173e-07,
      "loss": 0.2072,
      "step": 3080
    },
    {
      "epoch": 0.8237665994634853,
      "grad_norm": 1.498715554303158,
      "learning_rate": 9.169601158320707e-07,
      "loss": 0.2252,
      "step": 3090
    },
    {
      "epoch": 0.8264325107886099,
      "grad_norm": 1.4851333098682147,
      "learning_rate": 8.902736254703347e-07,
      "loss": 0.2045,
      "step": 3100
    },
    {
      "epoch": 0.8264325107886099,
      "eval_loss": 0.20747578144073486,
      "eval_runtime": 262.8395,
      "eval_samples_per_second": 37.281,
      "eval_steps_per_second": 4.661,
      "step": 3100
    },
    {
      "epoch": 0.8290984221137344,
      "grad_norm": 1.4703270343823225,
      "learning_rate": 8.639432269002102e-07,
      "loss": 0.2038,
      "step": 3110
    },
    {
      "epoch": 0.831764333438859,
      "grad_norm": 1.277311132392977,
      "learning_rate": 8.379712015482333e-07,
      "loss": 0.2071,
      "step": 3120
    },
    {
      "epoch": 0.8344302447639835,
      "grad_norm": 1.3393459507656948,
      "learning_rate": 8.123597997892918e-07,
      "loss": 0.1988,
      "step": 3130
    },
    {
      "epoch": 0.837096156089108,
      "grad_norm": 1.4461311558504941,
      "learning_rate": 7.871112407516474e-07,
      "loss": 0.2181,
      "step": 3140
    },
    {
      "epoch": 0.8397620674142326,
      "grad_norm": 1.435791157531731,
      "learning_rate": 7.622277121246513e-07,
      "loss": 0.2172,
      "step": 3150
    },
    {
      "epoch": 0.8397620674142326,
      "eval_loss": 0.2062389850616455,
      "eval_runtime": 262.9168,
      "eval_samples_per_second": 37.27,
      "eval_steps_per_second": 4.659,
      "step": 3150
    },
    {
      "epoch": 0.8424279787393572,
      "grad_norm": 1.3532701657995663,
      "learning_rate": 7.377113699691879e-07,
      "loss": 0.2045,
      "step": 3160
    },
    {
      "epoch": 0.8450938900644818,
      "grad_norm": 1.3493480728612797,
      "learning_rate": 7.135643385308677e-07,
      "loss": 0.2162,
      "step": 3170
    },
    {
      "epoch": 0.8477598013896063,
      "grad_norm": 1.5269560977358674,
      "learning_rate": 6.897887100559608e-07,
      "loss": 0.2089,
      "step": 3180
    },
    {
      "epoch": 0.8504257127147308,
      "grad_norm": 1.29264776120369,
      "learning_rate": 6.663865446101192e-07,
      "loss": 0.216,
      "step": 3190
    },
    {
      "epoch": 0.8530916240398554,
      "grad_norm": 1.6285538029089162,
      "learning_rate": 6.433598698998766e-07,
      "loss": 0.2138,
      "step": 3200
    },
    {
      "epoch": 0.8530916240398554,
      "eval_loss": 0.2074633538722992,
      "eval_runtime": 262.9966,
      "eval_samples_per_second": 37.259,
      "eval_steps_per_second": 4.658,
      "step": 3200
    },
    {
      "epoch": 0.8557575353649799,
      "grad_norm": 1.3678851449133813,
      "learning_rate": 6.207106810969577e-07,
      "loss": 0.2121,
      "step": 3210
    },
    {
      "epoch": 0.8584234466901045,
      "grad_norm": 1.535513254522266,
      "learning_rate": 5.98440940665399e-07,
      "loss": 0.2141,
      "step": 3220
    },
    {
      "epoch": 0.861089358015229,
      "grad_norm": 1.4232188919597053,
      "learning_rate": 5.765525781915172e-07,
      "loss": 0.1917,
      "step": 3230
    },
    {
      "epoch": 0.8637552693403535,
      "grad_norm": 1.379241907620819,
      "learning_rate": 5.550474902167091e-07,
      "loss": 0.213,
      "step": 3240
    },
    {
      "epoch": 0.8664211806654781,
      "grad_norm": 1.4866740177549789,
      "learning_rate": 5.339275400731331e-07,
      "loss": 0.194,
      "step": 3250
    },
    {
      "epoch": 0.8664211806654781,
      "eval_loss": 0.2050597220659256,
      "eval_runtime": 262.896,
      "eval_samples_per_second": 37.273,
      "eval_steps_per_second": 4.66,
      "step": 3250
    },
    {
      "epoch": 0.8690870919906026,
      "grad_norm": 1.3230806315640078,
      "learning_rate": 5.131945577222485e-07,
      "loss": 0.2031,
      "step": 3260
    },
    {
      "epoch": 0.8717530033157272,
      "grad_norm": 1.4976380022642137,
      "learning_rate": 4.92850339596268e-07,
      "loss": 0.2061,
      "step": 3270
    },
    {
      "epoch": 0.8744189146408518,
      "grad_norm": 1.421070272722705,
      "learning_rate": 4.728966484424913e-07,
      "loss": 0.2148,
      "step": 3280
    },
    {
      "epoch": 0.8770848259659763,
      "grad_norm": 1.321758338936035,
      "learning_rate": 4.5333521317058207e-07,
      "loss": 0.2181,
      "step": 3290
    },
    {
      "epoch": 0.8797507372911009,
      "grad_norm": 1.4222679966153535,
      "learning_rate": 4.3416772870275295e-07,
      "loss": 0.2133,
      "step": 3300
    },
    {
      "epoch": 0.8797507372911009,
      "eval_loss": 0.20511919260025024,
      "eval_runtime": 263.0571,
      "eval_samples_per_second": 37.25,
      "eval_steps_per_second": 4.657,
      "step": 3300
    },
    {
      "epoch": 0.8824166486162254,
      "grad_norm": 1.4283343273273894,
      "learning_rate": 4.153958558269189e-07,
      "loss": 0.1987,
      "step": 3310
    },
    {
      "epoch": 0.88508255994135,
      "grad_norm": 1.5675323422816119,
      "learning_rate": 3.9702122105278405e-07,
      "loss": 0.2,
      "step": 3320
    },
    {
      "epoch": 0.8877484712664745,
      "grad_norm": 1.3013063484388072,
      "learning_rate": 3.7904541647092506e-07,
      "loss": 0.1995,
      "step": 3330
    },
    {
      "epoch": 0.890414382591599,
      "grad_norm": 1.354917224230714,
      "learning_rate": 3.614699996148285e-07,
      "loss": 0.1988,
      "step": 3340
    },
    {
      "epoch": 0.8930802939167236,
      "grad_norm": 1.2832901778967167,
      "learning_rate": 3.442964933259474e-07,
      "loss": 0.2025,
      "step": 3350
    },
    {
      "epoch": 0.8930802939167236,
      "eval_loss": 0.20472171902656555,
      "eval_runtime": 263.0226,
      "eval_samples_per_second": 37.255,
      "eval_steps_per_second": 4.657,
      "step": 3350
    },
    {
      "epoch": 0.8957462052418481,
      "grad_norm": 1.510961640653365,
      "learning_rate": 3.275263856217442e-07,
      "loss": 0.2066,
      "step": 3360
    },
    {
      "epoch": 0.8984121165669727,
      "grad_norm": 1.4157593597386668,
      "learning_rate": 3.1116112956677045e-07,
      "loss": 0.2041,
      "step": 3370
    },
    {
      "epoch": 0.9010780278920972,
      "grad_norm": 1.3311731607352142,
      "learning_rate": 2.952021431467522e-07,
      "loss": 0.2075,
      "step": 3380
    },
    {
      "epoch": 0.9037439392172217,
      "grad_norm": 1.4936420011473173,
      "learning_rate": 2.7965080914573786e-07,
      "loss": 0.2102,
      "step": 3390
    },
    {
      "epoch": 0.9064098505423464,
      "grad_norm": 1.5419179206834437,
      "learning_rate": 2.6450847502627883e-07,
      "loss": 0.2088,
      "step": 3400
    },
    {
      "epoch": 0.9064098505423464,
      "eval_loss": 0.20498156547546387,
      "eval_runtime": 262.97,
      "eval_samples_per_second": 37.263,
      "eval_steps_per_second": 4.658,
      "step": 3400
    },
    {
      "epoch": 0.9090757618674709,
      "grad_norm": 1.4557819209662033,
      "learning_rate": 2.497764528126778e-07,
      "loss": 0.2127,
      "step": 3410
    },
    {
      "epoch": 0.9117416731925955,
      "grad_norm": 1.3483821589972327,
      "learning_rate": 2.3545601897731085e-07,
      "loss": 0.2087,
      "step": 3420
    },
    {
      "epoch": 0.91440758451772,
      "grad_norm": 1.3949209578658839,
      "learning_rate": 2.2154841433002062e-07,
      "loss": 0.208,
      "step": 3430
    },
    {
      "epoch": 0.9170734958428445,
      "grad_norm": 1.3985557189628612,
      "learning_rate": 2.0805484391061003e-07,
      "loss": 0.197,
      "step": 3440
    },
    {
      "epoch": 0.9197394071679691,
      "grad_norm": 1.6011076878070405,
      "learning_rate": 1.9497647688442478e-07,
      "loss": 0.204,
      "step": 3450
    },
    {
      "epoch": 0.9197394071679691,
      "eval_loss": 0.20443060994148254,
      "eval_runtime": 262.9479,
      "eval_samples_per_second": 37.266,
      "eval_steps_per_second": 4.659,
      "step": 3450
    },
    {
      "epoch": 0.9224053184930936,
      "grad_norm": 1.4046196100578223,
      "learning_rate": 1.8231444644105755e-07,
      "loss": 0.2063,
      "step": 3460
    },
    {
      "epoch": 0.9250712298182182,
      "grad_norm": 1.41779845252418,
      "learning_rate": 1.7006984969615226e-07,
      "loss": 0.2063,
      "step": 3470
    },
    {
      "epoch": 0.9277371411433427,
      "grad_norm": 1.3987321445452827,
      "learning_rate": 1.5824374759635165e-07,
      "loss": 0.2103,
      "step": 3480
    },
    {
      "epoch": 0.9304030524684672,
      "grad_norm": 1.2896847102113629,
      "learning_rate": 1.4683716482736364e-07,
      "loss": 0.1999,
      "step": 3490
    },
    {
      "epoch": 0.9330689637935918,
      "grad_norm": 1.4530276844098788,
      "learning_rate": 1.358510897251808e-07,
      "loss": 0.2059,
      "step": 3500
    },
    {
      "epoch": 0.9330689637935918,
      "eval_loss": 0.20389488339424133,
      "eval_runtime": 262.985,
      "eval_samples_per_second": 37.261,
      "eval_steps_per_second": 4.658,
      "step": 3500
    },
    {
      "epoch": 0.9357348751187163,
      "grad_norm": 1.376435217170314,
      "learning_rate": 1.252864741904425e-07,
      "loss": 0.2041,
      "step": 3510
    },
    {
      "epoch": 0.938400786443841,
      "grad_norm": 1.63745742682993,
      "learning_rate": 1.1514423360595939e-07,
      "loss": 0.2071,
      "step": 3520
    },
    {
      "epoch": 0.9410666977689655,
      "grad_norm": 1.352875159457686,
      "learning_rate": 1.0542524675739407e-07,
      "loss": 0.197,
      "step": 3530
    },
    {
      "epoch": 0.94373260909409,
      "grad_norm": 1.4239967793966135,
      "learning_rate": 9.613035575712304e-08,
      "loss": 0.2051,
      "step": 3540
    },
    {
      "epoch": 0.9463985204192146,
      "grad_norm": 1.3958904189251933,
      "learning_rate": 8.726036597126619e-08,
      "loss": 0.2103,
      "step": 3550
    },
    {
      "epoch": 0.9463985204192146,
      "eval_loss": 0.20391830801963806,
      "eval_runtime": 262.9962,
      "eval_samples_per_second": 37.259,
      "eval_steps_per_second": 4.658,
      "step": 3550
    },
    {
      "epoch": 0.9490644317443391,
      "grad_norm": 1.3789213423803108,
      "learning_rate": 7.881604594990932e-08,
      "loss": 0.2028,
      "step": 3560
    },
    {
      "epoch": 0.9517303430694637,
      "grad_norm": 1.4337723414549706,
      "learning_rate": 7.079812736050906e-08,
      "loss": 0.2141,
      "step": 3570
    },
    {
      "epoch": 0.9543962543945882,
      "grad_norm": 1.2816441548148045,
      "learning_rate": 6.3207304924498e-08,
      "loss": 0.2008,
      "step": 3580
    },
    {
      "epoch": 0.9570621657197127,
      "grad_norm": 1.3587442586245593,
      "learning_rate": 5.604423635709122e-08,
      "loss": 0.2088,
      "step": 3590
    },
    {
      "epoch": 0.9597280770448373,
      "grad_norm": 1.398655956796932,
      "learning_rate": 4.93095423102935e-08,
      "loss": 0.2102,
      "step": 3600
    },
    {
      "epoch": 0.9597280770448373,
      "eval_loss": 0.20385359227657318,
      "eval_runtime": 262.9737,
      "eval_samples_per_second": 37.262,
      "eval_steps_per_second": 4.658,
      "step": 3600
    },
    {
      "epoch": 0.9623939883699618,
      "grad_norm": 1.3545794976233818,
      "learning_rate": 4.3003806319127376e-08,
      "loss": 0.1974,
      "step": 3610
    },
    {
      "epoch": 0.9650598996950864,
      "grad_norm": 1.472940446548246,
      "learning_rate": 3.712757475106854e-08,
      "loss": 0.2038,
      "step": 3620
    },
    {
      "epoch": 0.9677258110202109,
      "grad_norm": 1.4688819140937053,
      "learning_rate": 3.168135675870654e-08,
      "loss": 0.2033,
      "step": 3630
    },
    {
      "epoch": 0.9703917223453354,
      "grad_norm": 1.559082216657787,
      "learning_rate": 2.6665624235629463e-08,
      "loss": 0.2066,
      "step": 3640
    },
    {
      "epoch": 0.97305763367046,
      "grad_norm": 1.4442615348656187,
      "learning_rate": 2.2080811775535006e-08,
      "loss": 0.2051,
      "step": 3650
    },
    {
      "epoch": 0.97305763367046,
      "eval_loss": 0.20379623770713806,
      "eval_runtime": 262.9438,
      "eval_samples_per_second": 37.267,
      "eval_steps_per_second": 4.659,
      "step": 3650
    },
    {
      "epoch": 0.9757235449955846,
      "grad_norm": 1.4302312819140066,
      "learning_rate": 1.7927316634573368e-08,
      "loss": 0.2005,
      "step": 3660
    },
    {
      "epoch": 0.9783894563207092,
      "grad_norm": 1.5231097086128733,
      "learning_rate": 1.4205498696930332e-08,
      "loss": 0.198,
      "step": 3670
    },
    {
      "epoch": 0.9810553676458337,
      "grad_norm": 1.3232525431455444,
      "learning_rate": 1.0915680443641663e-08,
      "loss": 0.1995,
      "step": 3680
    },
    {
      "epoch": 0.9837212789709582,
      "grad_norm": 1.373155468144249,
      "learning_rate": 8.058146924651012e-09,
      "loss": 0.2048,
      "step": 3690
    },
    {
      "epoch": 0.9863871902960828,
      "grad_norm": 1.4110519682344211,
      "learning_rate": 5.633145734114665e-09,
      "loss": 0.2017,
      "step": 3700
    },
    {
      "epoch": 0.9863871902960828,
      "eval_loss": 0.203701451420784,
      "eval_runtime": 263.031,
      "eval_samples_per_second": 37.254,
      "eval_steps_per_second": 4.657,
      "step": 3700
    },
    {
      "epoch": 0.9890531016212073,
      "grad_norm": 1.400340000512079,
      "learning_rate": 3.640886988945935e-09,
      "loss": 0.2069,
      "step": 3710
    },
    {
      "epoch": 0.9917190129463319,
      "grad_norm": 1.5211040678505572,
      "learning_rate": 2.0815433106080583e-09,
      "loss": 0.2021,
      "step": 3720
    },
    {
      "epoch": 0.9943849242714564,
      "grad_norm": 1.4375248921695807,
      "learning_rate": 9.55249810161152e-10,
      "loss": 0.2134,
      "step": 3730
    },
    {
      "epoch": 0.9970508355965809,
      "grad_norm": 1.492474671725836,
      "learning_rate": 2.6210407655269386e-10,
      "loss": 0.211,
      "step": 3740
    },
    {
      "epoch": 0.9997167469217055,
      "grad_norm": 1.6278639410883167,
      "learning_rate": 2.1661681620654963e-12,
      "loss": 0.2088,
      "step": 3750
    },
    {
      "epoch": 0.9997167469217055,
      "eval_loss": 0.2036588490009308,
      "eval_runtime": 262.7967,
      "eval_samples_per_second": 37.287,
      "eval_steps_per_second": 4.661,
      "step": 3750
    },
    {
      "epoch": 0.999983338054218,
      "step": 3751,
      "total_flos": 1934604910592000.0,
      "train_loss": 0.2527618171993302,
      "train_runtime": 72440.3327,
      "train_samples_per_second": 6.628,
      "train_steps_per_second": 0.052
    }
  ],
  "logging_steps": 10,
  "max_steps": 3751,
  "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1934604910592000.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}