{
  "best_metric": 0.9400918591493044,
  "best_model_checkpoint": "vit-msn-small-wbc-classifier-cells-separated-dataset-agregates-25/checkpoint-1556",
  "epoch": 24.843423799582464,
  "eval_steps": 500,
  "global_step": 2975,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08350730688935282,
      "grad_norm": 23.778831481933594,
      "learning_rate": 1.6778523489932886e-06,
      "loss": 1.6751,
      "step": 10
    },
    {
      "epoch": 0.16701461377870563,
      "grad_norm": 10.125561714172363,
      "learning_rate": 3.3557046979865773e-06,
      "loss": 1.2395,
      "step": 20
    },
    {
      "epoch": 0.25052192066805845,
      "grad_norm": 7.331182479858398,
      "learning_rate": 5.033557046979865e-06,
      "loss": 0.9635,
      "step": 30
    },
    {
      "epoch": 0.33402922755741127,
      "grad_norm": 6.259941577911377,
      "learning_rate": 6.7114093959731546e-06,
      "loss": 0.7296,
      "step": 40
    },
    {
      "epoch": 0.4175365344467641,
      "grad_norm": 9.791149139404297,
      "learning_rate": 8.389261744966444e-06,
      "loss": 0.582,
      "step": 50
    },
    {
      "epoch": 0.5010438413361169,
      "grad_norm": 8.130660057067871,
      "learning_rate": 1.006711409395973e-05,
      "loss": 0.4722,
      "step": 60
    },
    {
      "epoch": 0.5845511482254697,
      "grad_norm": 6.79920768737793,
      "learning_rate": 1.174496644295302e-05,
      "loss": 0.4658,
      "step": 70
    },
    {
      "epoch": 0.6680584551148225,
      "grad_norm": 6.699377536773682,
      "learning_rate": 1.3422818791946309e-05,
      "loss": 0.394,
      "step": 80
    },
    {
      "epoch": 0.7515657620041754,
      "grad_norm": 18.482280731201172,
      "learning_rate": 1.51006711409396e-05,
      "loss": 0.3988,
      "step": 90
    },
    {
      "epoch": 0.8350730688935282,
      "grad_norm": 14.189501762390137,
      "learning_rate": 1.6778523489932888e-05,
      "loss": 0.3955,
      "step": 100
    },
    {
      "epoch": 0.918580375782881,
      "grad_norm": 9.649164199829102,
      "learning_rate": 1.8456375838926178e-05,
      "loss": 0.351,
      "step": 110
    },
    {
      "epoch": 0.9937369519832986,
      "eval_accuracy": 0.9151301337948479,
      "eval_loss": 0.2522774636745453,
      "eval_runtime": 47.6631,
      "eval_samples_per_second": 315.191,
      "eval_steps_per_second": 4.93,
      "step": 119
    },
    {
      "epoch": 1.0020876826722338,
      "grad_norm": 8.503198623657227,
      "learning_rate": 2.013422818791946e-05,
      "loss": 0.377,
      "step": 120
    },
    {
      "epoch": 1.0855949895615866,
      "grad_norm": 6.6437249183654785,
      "learning_rate": 2.181208053691275e-05,
      "loss": 0.3533,
      "step": 130
    },
    {
      "epoch": 1.1691022964509394,
      "grad_norm": 5.940201282501221,
      "learning_rate": 2.348993288590604e-05,
      "loss": 0.3391,
      "step": 140
    },
    {
      "epoch": 1.2526096033402923,
      "grad_norm": 5.3783135414123535,
      "learning_rate": 2.516778523489933e-05,
      "loss": 0.3416,
      "step": 150
    },
    {
      "epoch": 1.336116910229645,
      "grad_norm": 7.877007484436035,
      "learning_rate": 2.6845637583892618e-05,
      "loss": 0.3356,
      "step": 160
    },
    {
      "epoch": 1.4196242171189979,
      "grad_norm": 8.07664680480957,
      "learning_rate": 2.8523489932885905e-05,
      "loss": 0.3246,
      "step": 170
    },
    {
      "epoch": 1.5031315240083507,
      "grad_norm": 10.465314865112305,
      "learning_rate": 3.02013422818792e-05,
      "loss": 0.3957,
      "step": 180
    },
    {
      "epoch": 1.5866388308977035,
      "grad_norm": 10.284217834472656,
      "learning_rate": 3.1879194630872485e-05,
      "loss": 0.3753,
      "step": 190
    },
    {
      "epoch": 1.6701461377870563,
      "grad_norm": 8.860455513000488,
      "learning_rate": 3.3557046979865775e-05,
      "loss": 0.3678,
      "step": 200
    },
    {
      "epoch": 1.7536534446764092,
      "grad_norm": 6.609280586242676,
      "learning_rate": 3.523489932885906e-05,
      "loss": 0.3449,
      "step": 210
    },
    {
      "epoch": 1.837160751565762,
      "grad_norm": 5.114744186401367,
      "learning_rate": 3.6912751677852356e-05,
      "loss": 0.3281,
      "step": 220
    },
    {
      "epoch": 1.9206680584551148,
      "grad_norm": 5.134997844696045,
      "learning_rate": 3.859060402684564e-05,
      "loss": 0.3364,
      "step": 230
    },
    {
      "epoch": 1.9958246346555324,
      "eval_accuracy": 0.9195233974572322,
      "eval_loss": 0.23547479510307312,
      "eval_runtime": 47.2123,
      "eval_samples_per_second": 318.201,
      "eval_steps_per_second": 4.978,
      "step": 239
    },
    {
      "epoch": 2.0041753653444676,
      "grad_norm": 7.969848155975342,
      "learning_rate": 4.026845637583892e-05,
      "loss": 0.3517,
      "step": 240
    },
    {
      "epoch": 2.0876826722338206,
      "grad_norm": 7.928933143615723,
      "learning_rate": 4.194630872483222e-05,
      "loss": 0.3229,
      "step": 250
    },
    {
      "epoch": 2.1711899791231732,
      "grad_norm": 5.597458839416504,
      "learning_rate": 4.36241610738255e-05,
      "loss": 0.3295,
      "step": 260
    },
    {
      "epoch": 2.2546972860125263,
      "grad_norm": 5.401355266571045,
      "learning_rate": 4.530201342281879e-05,
      "loss": 0.3238,
      "step": 270
    },
    {
      "epoch": 2.338204592901879,
      "grad_norm": 7.695993423461914,
      "learning_rate": 4.697986577181208e-05,
      "loss": 0.3138,
      "step": 280
    },
    {
      "epoch": 2.421711899791232,
      "grad_norm": 8.675772666931152,
      "learning_rate": 4.865771812080537e-05,
      "loss": 0.3461,
      "step": 290
    },
    {
      "epoch": 2.5052192066805845,
      "grad_norm": 9.49290657043457,
      "learning_rate": 4.99626447515876e-05,
      "loss": 0.3524,
      "step": 300
    },
    {
      "epoch": 2.588726513569937,
      "grad_norm": 3.6922707557678223,
      "learning_rate": 4.977586850952559e-05,
      "loss": 0.3145,
      "step": 310
    },
    {
      "epoch": 2.67223382045929,
      "grad_norm": 7.227968215942383,
      "learning_rate": 4.958909226746358e-05,
      "loss": 0.3054,
      "step": 320
    },
    {
      "epoch": 2.755741127348643,
      "grad_norm": 5.04173469543457,
      "learning_rate": 4.940231602540157e-05,
      "loss": 0.3404,
      "step": 330
    },
    {
      "epoch": 2.8392484342379958,
      "grad_norm": 5.110103607177734,
      "learning_rate": 4.9215539783339556e-05,
      "loss": 0.2992,
      "step": 340
    },
    {
      "epoch": 2.9227557411273484,
      "grad_norm": 5.469402313232422,
      "learning_rate": 4.902876354127755e-05,
      "loss": 0.2999,
      "step": 350
    },
    {
      "epoch": 2.997912317327766,
      "eval_accuracy": 0.9169273780203687,
      "eval_loss": 0.23837865889072418,
      "eval_runtime": 46.7529,
      "eval_samples_per_second": 321.327,
      "eval_steps_per_second": 5.026,
      "step": 359
    },
    {
      "epoch": 3.0062630480167014,
      "grad_norm": 5.160531997680664,
      "learning_rate": 4.884198729921554e-05,
      "loss": 0.2911,
      "step": 360
    },
    {
      "epoch": 3.0897703549060545,
      "grad_norm": 5.856198310852051,
      "learning_rate": 4.865521105715353e-05,
      "loss": 0.2959,
      "step": 370
    },
    {
      "epoch": 3.173277661795407,
      "grad_norm": 3.725022792816162,
      "learning_rate": 4.846843481509152e-05,
      "loss": 0.2956,
      "step": 380
    },
    {
      "epoch": 3.25678496868476,
      "grad_norm": 5.262429237365723,
      "learning_rate": 4.828165857302951e-05,
      "loss": 0.3044,
      "step": 390
    },
    {
      "epoch": 3.3402922755741127,
      "grad_norm": 5.595754623413086,
      "learning_rate": 4.80948823309675e-05,
      "loss": 0.3188,
      "step": 400
    },
    {
      "epoch": 3.4237995824634657,
      "grad_norm": 5.292787075042725,
      "learning_rate": 4.790810608890549e-05,
      "loss": 0.3001,
      "step": 410
    },
    {
      "epoch": 3.5073068893528183,
      "grad_norm": 4.534820079803467,
      "learning_rate": 4.772132984684348e-05,
      "loss": 0.3011,
      "step": 420
    },
    {
      "epoch": 3.5908141962421714,
      "grad_norm": 11.022297859191895,
      "learning_rate": 4.753455360478147e-05,
      "loss": 0.2965,
      "step": 430
    },
    {
      "epoch": 3.674321503131524,
      "grad_norm": 5.595705986022949,
      "learning_rate": 4.734777736271946e-05,
      "loss": 0.3172,
      "step": 440
    },
    {
      "epoch": 3.757828810020877,
      "grad_norm": 5.017602443695068,
      "learning_rate": 4.716100112065745e-05,
      "loss": 0.326,
      "step": 450
    },
    {
      "epoch": 3.8413361169102296,
      "grad_norm": 3.5311007499694824,
      "learning_rate": 4.697422487859545e-05,
      "loss": 0.2831,
      "step": 460
    },
    {
      "epoch": 3.9248434237995826,
      "grad_norm": 4.687069892883301,
      "learning_rate": 4.678744863653344e-05,
      "loss": 0.2861,
      "step": 470
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9341010450642349,
      "eval_loss": 0.190222829580307,
      "eval_runtime": 46.9219,
      "eval_samples_per_second": 320.17,
      "eval_steps_per_second": 5.008,
      "step": 479
    },
    {
      "epoch": 4.008350730688935,
      "grad_norm": 4.544097900390625,
      "learning_rate": 4.660067239447142e-05,
      "loss": 0.305,
      "step": 480
    },
    {
      "epoch": 4.091858037578288,
      "grad_norm": 4.896882057189941,
      "learning_rate": 4.641389615240941e-05,
      "loss": 0.2827,
      "step": 490
    },
    {
      "epoch": 4.175365344467641,
      "grad_norm": 3.459932804107666,
      "learning_rate": 4.622711991034741e-05,
      "loss": 0.2754,
      "step": 500
    },
    {
      "epoch": 4.258872651356993,
      "grad_norm": 4.049550533294678,
      "learning_rate": 4.60403436682854e-05,
      "loss": 0.3175,
      "step": 510
    },
    {
      "epoch": 4.3423799582463465,
      "grad_norm": 5.197091579437256,
      "learning_rate": 4.585356742622339e-05,
      "loss": 0.2696,
      "step": 520
    },
    {
      "epoch": 4.4258872651356995,
      "grad_norm": 3.382556915283203,
      "learning_rate": 4.566679118416138e-05,
      "loss": 0.2625,
      "step": 530
    },
    {
      "epoch": 4.509394572025053,
      "grad_norm": 5.480315208435059,
      "learning_rate": 4.548001494209937e-05,
      "loss": 0.3011,
      "step": 540
    },
    {
      "epoch": 4.592901878914405,
      "grad_norm": 4.315974712371826,
      "learning_rate": 4.529323870003736e-05,
      "loss": 0.2671,
      "step": 550
    },
    {
      "epoch": 4.676409185803758,
      "grad_norm": 4.109740734100342,
      "learning_rate": 4.510646245797535e-05,
      "loss": 0.2873,
      "step": 560
    },
    {
      "epoch": 4.759916492693111,
      "grad_norm": 4.254931449890137,
      "learning_rate": 4.491968621591334e-05,
      "loss": 0.2989,
      "step": 570
    },
    {
      "epoch": 4.843423799582464,
      "grad_norm": 2.6343159675598145,
      "learning_rate": 4.473290997385133e-05,
      "loss": 0.2851,
      "step": 580
    },
    {
      "epoch": 4.926931106471816,
      "grad_norm": 4.918463230133057,
      "learning_rate": 4.454613373178932e-05,
      "loss": 0.3014,
      "step": 590
    },
    {
      "epoch": 4.993736951983299,
      "eval_accuracy": 0.9289755707914531,
      "eval_loss": 0.21536433696746826,
      "eval_runtime": 46.4537,
      "eval_samples_per_second": 323.397,
      "eval_steps_per_second": 5.059,
      "step": 598
    },
    {
      "epoch": 5.010438413361169,
      "grad_norm": 4.048580646514893,
      "learning_rate": 4.435935748972731e-05,
      "loss": 0.3089,
      "step": 600
    },
    {
      "epoch": 5.093945720250522,
      "grad_norm": 2.6534504890441895,
      "learning_rate": 4.4172581247665304e-05,
      "loss": 0.2496,
      "step": 610
    },
    {
      "epoch": 5.177453027139875,
      "grad_norm": 5.901323318481445,
      "learning_rate": 4.398580500560329e-05,
      "loss": 0.3034,
      "step": 620
    },
    {
      "epoch": 5.260960334029227,
      "grad_norm": 4.444337844848633,
      "learning_rate": 4.379902876354128e-05,
      "loss": 0.2669,
      "step": 630
    },
    {
      "epoch": 5.34446764091858,
      "grad_norm": 6.911639213562012,
      "learning_rate": 4.361225252147927e-05,
      "loss": 0.26,
      "step": 640
    },
    {
      "epoch": 5.427974947807933,
      "grad_norm": 4.877162933349609,
      "learning_rate": 4.342547627941726e-05,
      "loss": 0.2732,
      "step": 650
    },
    {
      "epoch": 5.511482254697286,
      "grad_norm": 3.9637715816497803,
      "learning_rate": 4.3238700037355254e-05,
      "loss": 0.2957,
      "step": 660
    },
    {
      "epoch": 5.5949895615866385,
      "grad_norm": 4.107086181640625,
      "learning_rate": 4.3051923795293244e-05,
      "loss": 0.2746,
      "step": 670
    },
    {
      "epoch": 5.6784968684759916,
      "grad_norm": 4.021276950836182,
      "learning_rate": 4.2865147553231234e-05,
      "loss": 0.2755,
      "step": 680
    },
    {
      "epoch": 5.762004175365345,
      "grad_norm": 3.1536874771118164,
      "learning_rate": 4.2678371311169224e-05,
      "loss": 0.2781,
      "step": 690
    },
    {
      "epoch": 5.845511482254698,
      "grad_norm": 4.305199146270752,
      "learning_rate": 4.249159506910721e-05,
      "loss": 0.2536,
      "step": 700
    },
    {
      "epoch": 5.92901878914405,
      "grad_norm": 3.679614305496216,
      "learning_rate": 4.2304818827045204e-05,
      "loss": 0.292,
      "step": 710
    },
    {
      "epoch": 5.995824634655532,
      "eval_accuracy": 0.9382946149237835,
      "eval_loss": 0.176395446062088,
      "eval_runtime": 47.1844,
      "eval_samples_per_second": 318.389,
      "eval_steps_per_second": 4.98,
      "step": 718
    },
    {
      "epoch": 6.012526096033403,
      "grad_norm": 4.32383394241333,
      "learning_rate": 4.2118042584983194e-05,
      "loss": 0.2623,
      "step": 720
    },
    {
      "epoch": 6.096033402922756,
      "grad_norm": 3.5262672901153564,
      "learning_rate": 4.1931266342921183e-05,
      "loss": 0.2638,
      "step": 730
    },
    {
      "epoch": 6.179540709812109,
      "grad_norm": 4.600202560424805,
      "learning_rate": 4.1744490100859173e-05,
      "loss": 0.2747,
      "step": 740
    },
    {
      "epoch": 6.263048016701461,
      "grad_norm": 3.81154465675354,
      "learning_rate": 4.155771385879716e-05,
      "loss": 0.2528,
      "step": 750
    },
    {
      "epoch": 6.346555323590814,
      "grad_norm": 4.463883399963379,
      "learning_rate": 4.137093761673515e-05,
      "loss": 0.2588,
      "step": 760
    },
    {
      "epoch": 6.430062630480167,
      "grad_norm": 3.8353664875030518,
      "learning_rate": 4.118416137467314e-05,
      "loss": 0.2618,
      "step": 770
    },
    {
      "epoch": 6.51356993736952,
      "grad_norm": 3.2617974281311035,
      "learning_rate": 4.099738513261113e-05,
      "loss": 0.2581,
      "step": 780
    },
    {
      "epoch": 6.597077244258872,
      "grad_norm": 6.762134075164795,
      "learning_rate": 4.081060889054912e-05,
      "loss": 0.288,
      "step": 790
    },
    {
      "epoch": 6.680584551148225,
      "grad_norm": 3.397378444671631,
      "learning_rate": 4.062383264848711e-05,
      "loss": 0.2972,
      "step": 800
    },
    {
      "epoch": 6.764091858037578,
      "grad_norm": 3.1325230598449707,
      "learning_rate": 4.04370564064251e-05,
      "loss": 0.2783,
      "step": 810
    },
    {
      "epoch": 6.847599164926931,
      "grad_norm": 4.460423946380615,
      "learning_rate": 4.02502801643631e-05,
      "loss": 0.2532,
      "step": 820
    },
    {
      "epoch": 6.931106471816284,
      "grad_norm": 4.6303839683532715,
      "learning_rate": 4.006350392230109e-05,
      "loss": 0.2441,
      "step": 830
    },
    {
      "epoch": 6.997912317327766,
      "eval_accuracy": 0.9348332556746323,
      "eval_loss": 0.1894187480211258,
      "eval_runtime": 47.4308,
      "eval_samples_per_second": 316.735,
      "eval_steps_per_second": 4.955,
      "step": 838
    },
    {
      "epoch": 7.014613778705637,
      "grad_norm": 3.667642593383789,
      "learning_rate": 3.987672768023907e-05,
      "loss": 0.2548,
      "step": 840
    },
    {
      "epoch": 7.09812108559499,
      "grad_norm": 3.9511961936950684,
      "learning_rate": 3.968995143817706e-05,
      "loss": 0.2458,
      "step": 850
    },
    {
      "epoch": 7.181628392484343,
      "grad_norm": 4.884164810180664,
      "learning_rate": 3.950317519611505e-05,
      "loss": 0.2311,
      "step": 860
    },
    {
      "epoch": 7.265135699373695,
      "grad_norm": 5.010523796081543,
      "learning_rate": 3.931639895405305e-05,
      "loss": 0.2719,
      "step": 870
    },
    {
      "epoch": 7.348643006263048,
      "grad_norm": 4.1294636726379395,
      "learning_rate": 3.912962271199104e-05,
      "loss": 0.2674,
      "step": 880
    },
    {
      "epoch": 7.432150313152401,
      "grad_norm": 4.736259460449219,
      "learning_rate": 3.894284646992903e-05,
      "loss": 0.2591,
      "step": 890
    },
    {
      "epoch": 7.515657620041754,
      "grad_norm": 4.864129066467285,
      "learning_rate": 3.875607022786702e-05,
      "loss": 0.2645,
      "step": 900
    },
    {
      "epoch": 7.599164926931106,
      "grad_norm": 3.61091685295105,
      "learning_rate": 3.8569293985805e-05,
      "loss": 0.2505,
      "step": 910
    },
    {
      "epoch": 7.682672233820459,
      "grad_norm": 4.3915863037109375,
      "learning_rate": 3.8382517743743e-05,
      "loss": 0.2629,
      "step": 920
    },
    {
      "epoch": 7.766179540709812,
      "grad_norm": 4.1510419845581055,
      "learning_rate": 3.819574150168099e-05,
      "loss": 0.2463,
      "step": 930
    },
    {
      "epoch": 7.849686847599165,
      "grad_norm": 3.3655436038970947,
      "learning_rate": 3.800896525961898e-05,
      "loss": 0.2588,
      "step": 940
    },
    {
      "epoch": 7.933194154488517,
      "grad_norm": 3.585494041442871,
      "learning_rate": 3.782218901755697e-05,
      "loss": 0.2416,
      "step": 950
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9348998202755775,
      "eval_loss": 0.19133438169956207,
      "eval_runtime": 47.4532,
      "eval_samples_per_second": 316.585,
      "eval_steps_per_second": 4.952,
      "step": 958
    },
    {
      "epoch": 8.01670146137787,
      "grad_norm": 4.570364475250244,
      "learning_rate": 3.763541277549496e-05,
      "loss": 0.2512,
      "step": 960
    },
    {
      "epoch": 8.100208768267223,
      "grad_norm": 3.438917398452759,
      "learning_rate": 3.744863653343295e-05,
      "loss": 0.2376,
      "step": 970
    },
    {
      "epoch": 8.183716075156577,
      "grad_norm": 6.080543518066406,
      "learning_rate": 3.726186029137094e-05,
      "loss": 0.2552,
      "step": 980
    },
    {
      "epoch": 8.267223382045929,
      "grad_norm": 4.013462543487549,
      "learning_rate": 3.707508404930893e-05,
      "loss": 0.2494,
      "step": 990
    },
    {
      "epoch": 8.350730688935283,
      "grad_norm": 3.6214938163757324,
      "learning_rate": 3.688830780724692e-05,
      "loss": 0.2411,
      "step": 1000
    },
    {
      "epoch": 8.434237995824635,
      "grad_norm": 3.5283772945404053,
      "learning_rate": 3.670153156518491e-05,
      "loss": 0.2266,
      "step": 1010
    },
    {
      "epoch": 8.517745302713987,
      "grad_norm": 6.508725166320801,
      "learning_rate": 3.65147553231229e-05,
      "loss": 0.2463,
      "step": 1020
    },
    {
      "epoch": 8.60125260960334,
      "grad_norm": 3.3739099502563477,
      "learning_rate": 3.6327979081060895e-05,
      "loss": 0.2359,
      "step": 1030
    },
    {
      "epoch": 8.684759916492693,
      "grad_norm": 3.321066379547119,
      "learning_rate": 3.6141202838998885e-05,
      "loss": 0.2445,
      "step": 1040
    },
    {
      "epoch": 8.768267223382045,
      "grad_norm": 3.418613910675049,
      "learning_rate": 3.595442659693687e-05,
      "loss": 0.2588,
      "step": 1050
    },
    {
      "epoch": 8.851774530271399,
      "grad_norm": 3.841691255569458,
      "learning_rate": 3.576765035487486e-05,
      "loss": 0.2431,
      "step": 1060
    },
    {
      "epoch": 8.935281837160751,
      "grad_norm": 4.421708583831787,
      "learning_rate": 3.558087411281285e-05,
      "loss": 0.2642,
      "step": 1070
    },
    {
      "epoch": 8.993736951983298,
      "eval_accuracy": 0.9384943087266192,
      "eval_loss": 0.17376160621643066,
      "eval_runtime": 47.0041,
      "eval_samples_per_second": 319.611,
      "eval_steps_per_second": 5.0,
      "step": 1077
    },
    {
      "epoch": 9.018789144050105,
      "grad_norm": 3.4606330394744873,
      "learning_rate": 3.5394097870750844e-05,
      "loss": 0.2577,
      "step": 1080
    },
    {
      "epoch": 9.102296450939457,
      "grad_norm": 2.89786434173584,
      "learning_rate": 3.5207321628688834e-05,
      "loss": 0.2562,
      "step": 1090
    },
    {
      "epoch": 9.18580375782881,
      "grad_norm": 4.433104515075684,
      "learning_rate": 3.5020545386626824e-05,
      "loss": 0.2368,
      "step": 1100
    },
    {
      "epoch": 9.269311064718163,
      "grad_norm": 3.6675984859466553,
      "learning_rate": 3.4833769144564814e-05,
      "loss": 0.2429,
      "step": 1110
    },
    {
      "epoch": 9.352818371607516,
      "grad_norm": 3.072791814804077,
      "learning_rate": 3.46469929025028e-05,
      "loss": 0.2117,
      "step": 1120
    },
    {
      "epoch": 9.436325678496868,
      "grad_norm": 3.7456765174865723,
      "learning_rate": 3.4460216660440794e-05,
      "loss": 0.2368,
      "step": 1130
    },
    {
      "epoch": 9.519832985386222,
      "grad_norm": 5.997681140899658,
      "learning_rate": 3.4273440418378784e-05,
      "loss": 0.2513,
      "step": 1140
    },
    {
      "epoch": 9.603340292275574,
      "grad_norm": 4.1110124588012695,
      "learning_rate": 3.4086664176316774e-05,
      "loss": 0.2602,
      "step": 1150
    },
    {
      "epoch": 9.686847599164928,
      "grad_norm": 3.402402639389038,
      "learning_rate": 3.3899887934254764e-05,
      "loss": 0.2323,
      "step": 1160
    },
    {
      "epoch": 9.77035490605428,
      "grad_norm": 3.7895450592041016,
      "learning_rate": 3.3713111692192754e-05,
      "loss": 0.2339,
      "step": 1170
    },
    {
      "epoch": 9.853862212943632,
      "grad_norm": 2.2243807315826416,
      "learning_rate": 3.3526335450130744e-05,
      "loss": 0.2496,
      "step": 1180
    },
    {
      "epoch": 9.937369519832986,
      "grad_norm": 4.900872707366943,
      "learning_rate": 3.333955920806874e-05,
      "loss": 0.2482,
      "step": 1190
    },
    {
      "epoch": 9.995824634655532,
      "eval_accuracy": 0.9370964521067696,
      "eval_loss": 0.19109387695789337,
      "eval_runtime": 47.2977,
      "eval_samples_per_second": 317.626,
      "eval_steps_per_second": 4.969,
      "step": 1197
    },
    {
      "epoch": 10.020876826722338,
      "grad_norm": 5.6255998611450195,
      "learning_rate": 3.3152782966006724e-05,
      "loss": 0.2399,
      "step": 1200
    },
    {
      "epoch": 10.10438413361169,
      "grad_norm": 3.8569297790527344,
      "learning_rate": 3.2966006723944714e-05,
      "loss": 0.2342,
      "step": 1210
    },
    {
      "epoch": 10.187891440501044,
      "grad_norm": 3.362445831298828,
      "learning_rate": 3.2779230481882703e-05,
      "loss": 0.2104,
      "step": 1220
    },
    {
      "epoch": 10.271398747390396,
      "grad_norm": 3.918388605117798,
      "learning_rate": 3.2592454239820693e-05,
      "loss": 0.2319,
      "step": 1230
    },
    {
      "epoch": 10.35490605427975,
      "grad_norm": 4.792023658752441,
      "learning_rate": 3.240567799775869e-05,
      "loss": 0.2151,
      "step": 1240
    },
    {
      "epoch": 10.438413361169102,
      "grad_norm": 3.7443833351135254,
      "learning_rate": 3.221890175569668e-05,
      "loss": 0.2096,
      "step": 1250
    },
    {
      "epoch": 10.521920668058454,
      "grad_norm": 3.4863789081573486,
      "learning_rate": 3.203212551363467e-05,
      "loss": 0.2439,
      "step": 1260
    },
    {
      "epoch": 10.605427974947808,
      "grad_norm": 3.4557056427001953,
      "learning_rate": 3.184534927157265e-05,
      "loss": 0.2258,
      "step": 1270
    },
    {
      "epoch": 10.68893528183716,
      "grad_norm": 3.1916327476501465,
      "learning_rate": 3.165857302951064e-05,
      "loss": 0.2314,
      "step": 1280
    },
    {
      "epoch": 10.772442588726513,
      "grad_norm": 3.4237937927246094,
      "learning_rate": 3.147179678744864e-05,
      "loss": 0.2217,
      "step": 1290
    },
    {
      "epoch": 10.855949895615867,
      "grad_norm": 4.48100471496582,
      "learning_rate": 3.128502054538663e-05,
      "loss": 0.2268,
      "step": 1300
    },
    {
      "epoch": 10.939457202505219,
      "grad_norm": 4.131446361541748,
      "learning_rate": 3.109824430332462e-05,
      "loss": 0.2279,
      "step": 1310
    },
    {
      "epoch": 10.997912317327767,
      "eval_accuracy": 0.9380949211209478,
      "eval_loss": 0.1867293417453766,
      "eval_runtime": 47.3757,
      "eval_samples_per_second": 317.104,
      "eval_steps_per_second": 4.96,
      "step": 1317
    },
    {
      "epoch": 11.022964509394573,
      "grad_norm": 3.454545736312866,
      "learning_rate": 3.091146806126261e-05,
      "loss": 0.2134,
      "step": 1320
    },
    {
      "epoch": 11.106471816283925,
      "grad_norm": 5.787585258483887,
      "learning_rate": 3.07246918192006e-05,
      "loss": 0.2072,
      "step": 1330
    },
    {
      "epoch": 11.189979123173277,
      "grad_norm": 3.804635763168335,
      "learning_rate": 3.053791557713859e-05,
      "loss": 0.2113,
      "step": 1340
    },
    {
      "epoch": 11.273486430062631,
      "grad_norm": 2.5503077507019043,
      "learning_rate": 3.035113933507658e-05,
      "loss": 0.2299,
      "step": 1350
    },
    {
      "epoch": 11.356993736951983,
      "grad_norm": 3.30344295501709,
      "learning_rate": 3.016436309301457e-05,
      "loss": 0.2037,
      "step": 1360
    },
    {
      "epoch": 11.440501043841335,
      "grad_norm": 3.5026683807373047,
      "learning_rate": 2.997758685095256e-05,
      "loss": 0.2242,
      "step": 1370
    },
    {
      "epoch": 11.52400835073069,
      "grad_norm": 4.991878032684326,
      "learning_rate": 2.9790810608890552e-05,
      "loss": 0.2234,
      "step": 1380
    },
    {
      "epoch": 11.607515657620041,
      "grad_norm": 5.6365437507629395,
      "learning_rate": 2.9604034366828542e-05,
      "loss": 0.2238,
      "step": 1390
    },
    {
      "epoch": 11.691022964509395,
      "grad_norm": 3.200317144393921,
      "learning_rate": 2.9417258124766532e-05,
      "loss": 0.213,
      "step": 1400
    },
    {
      "epoch": 11.774530271398747,
      "grad_norm": 4.857986927032471,
      "learning_rate": 2.923048188270452e-05,
      "loss": 0.2279,
      "step": 1410
    },
    {
      "epoch": 11.8580375782881,
      "grad_norm": 3.4726953506469727,
      "learning_rate": 2.904370564064251e-05,
      "loss": 0.2346,
      "step": 1420
    },
    {
      "epoch": 11.941544885177453,
      "grad_norm": 4.036787986755371,
      "learning_rate": 2.8856929398580502e-05,
      "loss": 0.2331,
      "step": 1430
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9388936963322905,
      "eval_loss": 0.1814269721508026,
      "eval_runtime": 46.6546,
      "eval_samples_per_second": 322.004,
      "eval_steps_per_second": 5.037,
      "step": 1437
    },
    {
      "epoch": 12.025052192066806,
      "grad_norm": 6.564273834228516,
      "learning_rate": 2.8670153156518492e-05,
      "loss": 0.2032,
      "step": 1440
    },
    {
      "epoch": 12.108559498956158,
      "grad_norm": 2.8712363243103027,
      "learning_rate": 2.8483376914456482e-05,
      "loss": 0.2101,
      "step": 1450
    },
    {
      "epoch": 12.192066805845512,
      "grad_norm": 4.087766647338867,
      "learning_rate": 2.8296600672394475e-05,
      "loss": 0.2016,
      "step": 1460
    },
    {
      "epoch": 12.275574112734864,
      "grad_norm": 4.2068867683410645,
      "learning_rate": 2.8109824430332465e-05,
      "loss": 0.2072,
      "step": 1470
    },
    {
      "epoch": 12.359081419624218,
      "grad_norm": 2.1417343616485596,
      "learning_rate": 2.7923048188270452e-05,
      "loss": 0.2167,
      "step": 1480
    },
    {
      "epoch": 12.44258872651357,
      "grad_norm": 4.0305256843566895,
      "learning_rate": 2.773627194620844e-05,
      "loss": 0.2179,
      "step": 1490
    },
    {
      "epoch": 12.526096033402922,
      "grad_norm": 3.431574821472168,
      "learning_rate": 2.754949570414643e-05,
      "loss": 0.2102,
      "step": 1500
    },
    {
      "epoch": 12.609603340292276,
      "grad_norm": 3.1864852905273438,
      "learning_rate": 2.7362719462084425e-05,
      "loss": 0.21,
      "step": 1510
    },
    {
      "epoch": 12.693110647181628,
      "grad_norm": 4.349884033203125,
      "learning_rate": 2.7175943220022415e-05,
      "loss": 0.2339,
      "step": 1520
    },
    {
      "epoch": 12.776617954070982,
      "grad_norm": 3.7044005393981934,
      "learning_rate": 2.6989166977960405e-05,
      "loss": 0.2364,
      "step": 1530
    },
    {
      "epoch": 12.860125260960334,
      "grad_norm": 4.139816761016846,
      "learning_rate": 2.6802390735898398e-05,
      "loss": 0.1971,
      "step": 1540
    },
    {
      "epoch": 12.943632567849686,
      "grad_norm": 3.852142333984375,
      "learning_rate": 2.661561449383638e-05,
      "loss": 0.2208,
      "step": 1550
    },
    {
      "epoch": 12.993736951983298,
      "eval_accuracy": 0.9400918591493044,
      "eval_loss": 0.17901407182216644,
      "eval_runtime": 47.4241,
      "eval_samples_per_second": 316.78,
      "eval_steps_per_second": 4.955,
      "step": 1556
    },
    {
      "epoch": 13.02713987473904,
      "grad_norm": 3.2813127040863037,
      "learning_rate": 2.6428838251774375e-05,
      "loss": 0.1947,
      "step": 1560
    },
    {
      "epoch": 13.110647181628392,
      "grad_norm": 3.5853235721588135,
      "learning_rate": 2.6242062009712364e-05,
      "loss": 0.2054,
      "step": 1570
    },
    {
      "epoch": 13.194154488517745,
      "grad_norm": 5.702735900878906,
      "learning_rate": 2.6055285767650354e-05,
      "loss": 0.2423,
      "step": 1580
    },
    {
      "epoch": 13.277661795407099,
      "grad_norm": 2.9345734119415283,
      "learning_rate": 2.5868509525588348e-05,
      "loss": 0.2051,
      "step": 1590
    },
    {
      "epoch": 13.36116910229645,
      "grad_norm": 3.577324390411377,
      "learning_rate": 2.5681733283526338e-05,
      "loss": 0.2146,
      "step": 1600
    },
    {
      "epoch": 13.444676409185803,
      "grad_norm": 4.566287994384766,
      "learning_rate": 2.5494957041464328e-05,
      "loss": 0.1991,
      "step": 1610
    },
    {
      "epoch": 13.528183716075157,
      "grad_norm": 3.031125545501709,
      "learning_rate": 2.5308180799402314e-05,
      "loss": 0.2154,
      "step": 1620
    },
    {
      "epoch": 13.611691022964509,
      "grad_norm": 4.158732891082764,
      "learning_rate": 2.5121404557340304e-05,
      "loss": 0.1946,
      "step": 1630
    },
    {
      "epoch": 13.695198329853863,
      "grad_norm": 4.315579414367676,
      "learning_rate": 2.4934628315278297e-05,
      "loss": 0.1917,
      "step": 1640
    },
    {
      "epoch": 13.778705636743215,
      "grad_norm": 4.10720157623291,
      "learning_rate": 2.4747852073216287e-05,
      "loss": 0.1939,
      "step": 1650
    },
    {
      "epoch": 13.862212943632567,
      "grad_norm": 3.212385892868042,
      "learning_rate": 2.4561075831154277e-05,
      "loss": 0.2117,
      "step": 1660
    },
    {
      "epoch": 13.945720250521921,
      "grad_norm": 6.1147260665893555,
      "learning_rate": 2.437429958909227e-05,
      "loss": 0.2326,
      "step": 1670
    },
    {
      "epoch": 13.995824634655532,
      "eval_accuracy": 0.9366304999001531,
      "eval_loss": 0.19255150854587555,
      "eval_runtime": 46.9856,
      "eval_samples_per_second": 319.736,
      "eval_steps_per_second": 5.002,
      "step": 1676
    },
    {
      "epoch": 14.029227557411273,
      "grad_norm": 3.787531852722168,
      "learning_rate": 2.4187523347030257e-05,
      "loss": 0.185,
      "step": 1680
    },
    {
      "epoch": 14.112734864300627,
      "grad_norm": 3.1980652809143066,
      "learning_rate": 2.400074710496825e-05,
      "loss": 0.2099,
      "step": 1690
    },
    {
      "epoch": 14.19624217118998,
      "grad_norm": 5.701913356781006,
      "learning_rate": 2.381397086290624e-05,
      "loss": 0.1836,
      "step": 1700
    },
    {
      "epoch": 14.279749478079331,
      "grad_norm": 3.619966745376587,
      "learning_rate": 2.362719462084423e-05,
      "loss": 0.2131,
      "step": 1710
    },
    {
      "epoch": 14.363256784968685,
      "grad_norm": 2.987546443939209,
      "learning_rate": 2.344041837878222e-05,
      "loss": 0.1942,
      "step": 1720
    },
    {
      "epoch": 14.446764091858038,
      "grad_norm": 3.103140115737915,
      "learning_rate": 2.325364213672021e-05,
      "loss": 0.1909,
      "step": 1730
    },
    {
      "epoch": 14.53027139874739,
      "grad_norm": 3.544508695602417,
      "learning_rate": 2.3066865894658203e-05,
      "loss": 0.2042,
      "step": 1740
    },
    {
      "epoch": 14.613778705636744,
      "grad_norm": 4.177450656890869,
      "learning_rate": 2.288008965259619e-05,
      "loss": 0.1984,
      "step": 1750
    },
    {
      "epoch": 14.697286012526096,
      "grad_norm": 3.7388668060302734,
      "learning_rate": 2.269331341053418e-05,
      "loss": 0.2138,
      "step": 1760
    },
    {
      "epoch": 14.780793319415448,
      "grad_norm": 2.8315608501434326,
      "learning_rate": 2.2506537168472173e-05,
      "loss": 0.2081,
      "step": 1770
    },
    {
      "epoch": 14.864300626304802,
      "grad_norm": 3.59485125541687,
      "learning_rate": 2.231976092641016e-05,
      "loss": 0.1898,
      "step": 1780
    },
    {
      "epoch": 14.947807933194154,
      "grad_norm": 4.520532608032227,
      "learning_rate": 2.2132984684348153e-05,
      "loss": 0.1899,
      "step": 1790
    },
    {
      "epoch": 14.997912317327767,
      "eval_accuracy": 0.9371630167077148,
      "eval_loss": 0.19751960039138794,
      "eval_runtime": 47.3817,
      "eval_samples_per_second": 317.063,
      "eval_steps_per_second": 4.96,
      "step": 1796
    },
    {
      "epoch": 15.031315240083508,
      "grad_norm": 4.287206172943115,
      "learning_rate": 2.1946208442286143e-05,
      "loss": 0.1856,
      "step": 1800
    },
    {
      "epoch": 15.11482254697286,
      "grad_norm": 3.0820515155792236,
      "learning_rate": 2.1759432200224133e-05,
      "loss": 0.1809,
      "step": 1810
    },
    {
      "epoch": 15.198329853862212,
      "grad_norm": 4.431981086730957,
      "learning_rate": 2.1572655958162123e-05,
      "loss": 0.1904,
      "step": 1820
    },
    {
      "epoch": 15.281837160751566,
      "grad_norm": 4.667430400848389,
      "learning_rate": 2.1385879716100113e-05,
      "loss": 0.1854,
      "step": 1830
    },
    {
      "epoch": 15.365344467640918,
      "grad_norm": 3.1775588989257812,
      "learning_rate": 2.1199103474038103e-05,
      "loss": 0.206,
      "step": 1840
    },
    {
      "epoch": 15.448851774530272,
      "grad_norm": 4.10621976852417,
      "learning_rate": 2.1012327231976096e-05,
      "loss": 0.1997,
      "step": 1850
    },
    {
      "epoch": 15.532359081419624,
      "grad_norm": 3.4218993186950684,
      "learning_rate": 2.0825550989914083e-05,
      "loss": 0.1791,
      "step": 1860
    },
    {
      "epoch": 15.615866388308977,
      "grad_norm": 3.493351936340332,
      "learning_rate": 2.0638774747852076e-05,
      "loss": 0.1951,
      "step": 1870
    },
    {
      "epoch": 15.69937369519833,
      "grad_norm": 4.011974811553955,
      "learning_rate": 2.0451998505790066e-05,
      "loss": 0.1871,
      "step": 1880
    },
    {
      "epoch": 15.782881002087683,
      "grad_norm": 5.0797576904296875,
      "learning_rate": 2.0265222263728052e-05,
      "loss": 0.197,
      "step": 1890
    },
    {
      "epoch": 15.866388308977035,
      "grad_norm": 4.040988922119141,
      "learning_rate": 2.0078446021666046e-05,
      "loss": 0.1797,
      "step": 1900
    },
    {
      "epoch": 15.949895615866389,
      "grad_norm": 3.886463165283203,
      "learning_rate": 1.9891669779604036e-05,
      "loss": 0.1822,
      "step": 1910
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9351660786793583,
      "eval_loss": 0.20523545145988464,
      "eval_runtime": 46.5997,
      "eval_samples_per_second": 322.384,
      "eval_steps_per_second": 5.043,
      "step": 1916
    },
    {
      "epoch": 16.03340292275574,
      "grad_norm": 3.4292867183685303,
      "learning_rate": 1.9704893537542025e-05,
      "loss": 0.1948,
      "step": 1920
    },
    {
      "epoch": 16.116910229645093,
      "grad_norm": 4.07982063293457,
      "learning_rate": 1.9518117295480015e-05,
      "loss": 0.2022,
      "step": 1930
    },
    {
      "epoch": 16.200417536534445,
      "grad_norm": 3.2541186809539795,
      "learning_rate": 1.9331341053418005e-05,
      "loss": 0.1607,
      "step": 1940
    },
    {
      "epoch": 16.2839248434238,
      "grad_norm": 4.076241493225098,
      "learning_rate": 1.9144564811356e-05,
      "loss": 0.1579,
      "step": 1950
    },
    {
      "epoch": 16.367432150313153,
      "grad_norm": 3.7536261081695557,
      "learning_rate": 1.8957788569293985e-05,
      "loss": 0.1899,
      "step": 1960
    },
    {
      "epoch": 16.450939457202505,
      "grad_norm": 3.6366031169891357,
      "learning_rate": 1.8771012327231975e-05,
      "loss": 0.1818,
      "step": 1970
    },
    {
      "epoch": 16.534446764091857,
      "grad_norm": 3.176820755004883,
      "learning_rate": 1.858423608516997e-05,
      "loss": 0.1808,
      "step": 1980
    },
    {
      "epoch": 16.61795407098121,
      "grad_norm": 2.9712882041931152,
      "learning_rate": 1.839745984310796e-05,
      "loss": 0.1535,
      "step": 1990
    },
    {
      "epoch": 16.701461377870565,
      "grad_norm": 5.626895904541016,
      "learning_rate": 1.8210683601045948e-05,
      "loss": 0.1861,
      "step": 2000
    },
    {
      "epoch": 16.784968684759917,
      "grad_norm": 3.897157907485962,
      "learning_rate": 1.8023907358983938e-05,
      "loss": 0.1849,
      "step": 2010
    },
    {
      "epoch": 16.86847599164927,
      "grad_norm": 3.6972334384918213,
      "learning_rate": 1.7837131116921928e-05,
      "loss": 0.1562,
      "step": 2020
    },
    {
      "epoch": 16.95198329853862,
      "grad_norm": 6.524864673614502,
      "learning_rate": 1.7650354874859918e-05,
      "loss": 0.1837,
      "step": 2030
    },
    {
      "epoch": 16.993736951983298,
      "eval_accuracy": 0.9363642414963722,
      "eval_loss": 0.2078283280134201,
      "eval_runtime": 47.5905,
      "eval_samples_per_second": 315.672,
      "eval_steps_per_second": 4.938,
      "step": 2035
    },
    {
      "epoch": 17.035490605427974,
      "grad_norm": 4.618429183959961,
      "learning_rate": 1.7463578632797908e-05,
      "loss": 0.2028,
      "step": 2040
    },
    {
      "epoch": 17.11899791231733,
      "grad_norm": 4.595980167388916,
      "learning_rate": 1.7276802390735898e-05,
      "loss": 0.1687,
      "step": 2050
    },
    {
      "epoch": 17.20250521920668,
      "grad_norm": 2.575201988220215,
      "learning_rate": 1.709002614867389e-05,
      "loss": 0.1714,
      "step": 2060
    },
    {
      "epoch": 17.286012526096034,
      "grad_norm": 4.105851650238037,
      "learning_rate": 1.6903249906611878e-05,
      "loss": 0.1723,
      "step": 2070
    },
    {
      "epoch": 17.369519832985386,
      "grad_norm": 4.654282569885254,
      "learning_rate": 1.671647366454987e-05,
      "loss": 0.1891,
      "step": 2080
    },
    {
      "epoch": 17.453027139874738,
      "grad_norm": 3.8776493072509766,
      "learning_rate": 1.652969742248786e-05,
      "loss": 0.1662,
      "step": 2090
    },
    {
      "epoch": 17.53653444676409,
      "grad_norm": 7.09526252746582,
      "learning_rate": 1.634292118042585e-05,
      "loss": 0.1683,
      "step": 2100
    },
    {
      "epoch": 17.620041753653446,
      "grad_norm": 5.3857035636901855,
      "learning_rate": 1.615614493836384e-05,
      "loss": 0.1727,
      "step": 2110
    },
    {
      "epoch": 17.703549060542798,
      "grad_norm": 3.283949136734009,
      "learning_rate": 1.596936869630183e-05,
      "loss": 0.1903,
      "step": 2120
    },
    {
      "epoch": 17.78705636743215,
      "grad_norm": 5.982066631317139,
      "learning_rate": 1.578259245423982e-05,
      "loss": 0.1929,
      "step": 2130
    },
    {
      "epoch": 17.870563674321502,
      "grad_norm": 3.7765073776245117,
      "learning_rate": 1.559581621217781e-05,
      "loss": 0.1703,
      "step": 2140
    },
    {
      "epoch": 17.954070981210855,
      "grad_norm": 2.5590310096740723,
      "learning_rate": 1.54090399701158e-05,
      "loss": 0.1712,
      "step": 2150
    },
    {
      "epoch": 17.995824634655534,
      "eval_accuracy": 0.9288424415895626,
      "eval_loss": 0.23452672362327576,
      "eval_runtime": 47.2955,
      "eval_samples_per_second": 317.641,
      "eval_steps_per_second": 4.969,
      "step": 2155
    },
    {
      "epoch": 18.03757828810021,
      "grad_norm": 3.893446683883667,
      "learning_rate": 1.5222263728053792e-05,
      "loss": 0.1684,
      "step": 2160
    },
    {
      "epoch": 18.121085594989562,
      "grad_norm": 3.8617587089538574,
      "learning_rate": 1.5035487485991784e-05,
      "loss": 0.1913,
      "step": 2170
    },
    {
      "epoch": 18.204592901878915,
      "grad_norm": 3.266658306121826,
      "learning_rate": 1.4848711243929772e-05,
      "loss": 0.1699,
      "step": 2180
    },
    {
      "epoch": 18.288100208768267,
      "grad_norm": 3.2592689990997314,
      "learning_rate": 1.4661935001867764e-05,
      "loss": 0.1803,
      "step": 2190
    },
    {
      "epoch": 18.37160751565762,
      "grad_norm": 5.846588611602783,
      "learning_rate": 1.4475158759805754e-05,
      "loss": 0.1673,
      "step": 2200
    },
    {
      "epoch": 18.455114822546975,
      "grad_norm": 3.6083693504333496,
      "learning_rate": 1.4288382517743742e-05,
      "loss": 0.1631,
      "step": 2210
    },
    {
      "epoch": 18.538622129436327,
      "grad_norm": 3.4165074825286865,
      "learning_rate": 1.4101606275681733e-05,
      "loss": 0.152,
      "step": 2220
    },
    {
      "epoch": 18.62212943632568,
      "grad_norm": 3.8259122371673584,
      "learning_rate": 1.3914830033619725e-05,
      "loss": 0.1642,
      "step": 2230
    },
    {
      "epoch": 18.70563674321503,
      "grad_norm": 4.859554767608643,
      "learning_rate": 1.3728053791557715e-05,
      "loss": 0.1596,
      "step": 2240
    },
    {
      "epoch": 18.789144050104383,
      "grad_norm": 2.8771893978118896,
      "learning_rate": 1.3541277549495703e-05,
      "loss": 0.1484,
      "step": 2250
    },
    {
      "epoch": 18.872651356993735,
      "grad_norm": 4.107051849365234,
      "learning_rate": 1.3354501307433695e-05,
      "loss": 0.2161,
      "step": 2260
    },
    {
      "epoch": 18.95615866388309,
      "grad_norm": 5.493374347686768,
      "learning_rate": 1.3167725065371686e-05,
      "loss": 0.1715,
      "step": 2270
    },
    {
      "epoch": 18.997912317327767,
      "eval_accuracy": 0.9368301937029887,
      "eval_loss": 0.21558375656604767,
      "eval_runtime": 47.5388,
      "eval_samples_per_second": 316.016,
      "eval_steps_per_second": 4.943,
      "step": 2275
    },
    {
      "epoch": 19.039665970772443,
      "grad_norm": 2.9170358180999756,
      "learning_rate": 1.2980948823309675e-05,
      "loss": 0.1595,
      "step": 2280
    },
    {
      "epoch": 19.123173277661795,
      "grad_norm": 5.0754804611206055,
      "learning_rate": 1.2794172581247665e-05,
      "loss": 0.1671,
      "step": 2290
    },
    {
      "epoch": 19.206680584551147,
      "grad_norm": 4.310447692871094,
      "learning_rate": 1.2607396339185656e-05,
      "loss": 0.1759,
      "step": 2300
    },
    {
      "epoch": 19.2901878914405,
      "grad_norm": 5.011526584625244,
      "learning_rate": 1.2420620097123646e-05,
      "loss": 0.16,
      "step": 2310
    },
    {
      "epoch": 19.373695198329855,
      "grad_norm": 3.7505524158477783,
      "learning_rate": 1.2233843855061638e-05,
      "loss": 0.1622,
      "step": 2320
    },
    {
      "epoch": 19.457202505219207,
      "grad_norm": 4.531961917877197,
      "learning_rate": 1.2047067612999626e-05,
      "loss": 0.1648,
      "step": 2330
    },
    {
      "epoch": 19.54070981210856,
      "grad_norm": 2.6007046699523926,
      "learning_rate": 1.1860291370937618e-05,
      "loss": 0.1625,
      "step": 2340
    },
    {
      "epoch": 19.62421711899791,
      "grad_norm": 4.619992256164551,
      "learning_rate": 1.1673515128875608e-05,
      "loss": 0.1493,
      "step": 2350
    },
    {
      "epoch": 19.707724425887264,
      "grad_norm": 3.1437652111053467,
      "learning_rate": 1.1486738886813597e-05,
      "loss": 0.15,
      "step": 2360
    },
    {
      "epoch": 19.79123173277662,
      "grad_norm": 3.8565125465393066,
      "learning_rate": 1.1299962644751589e-05,
      "loss": 0.1433,
      "step": 2370
    },
    {
      "epoch": 19.87473903966597,
      "grad_norm": 4.988433837890625,
      "learning_rate": 1.1113186402689577e-05,
      "loss": 0.1663,
      "step": 2380
    },
    {
      "epoch": 19.958246346555324,
      "grad_norm": 4.064280033111572,
      "learning_rate": 1.0926410160627569e-05,
      "loss": 0.1516,
      "step": 2390
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9368301937029887,
      "eval_loss": 0.22790031135082245,
      "eval_runtime": 47.6095,
      "eval_samples_per_second": 315.546,
      "eval_steps_per_second": 4.936,
      "step": 2395
    },
    {
      "epoch": 20.041753653444676,
      "grad_norm": 4.115837097167969,
      "learning_rate": 1.0739633918565559e-05,
      "loss": 0.1424,
      "step": 2400
    },
    {
      "epoch": 20.12526096033403,
      "grad_norm": 4.192341327667236,
      "learning_rate": 1.055285767650355e-05,
      "loss": 0.155,
      "step": 2410
    },
    {
      "epoch": 20.20876826722338,
      "grad_norm": 4.125138282775879,
      "learning_rate": 1.0366081434441539e-05,
      "loss": 0.1576,
      "step": 2420
    },
    {
      "epoch": 20.292275574112736,
      "grad_norm": 3.7545394897460938,
      "learning_rate": 1.017930519237953e-05,
      "loss": 0.168,
      "step": 2430
    },
    {
      "epoch": 20.37578288100209,
      "grad_norm": 4.370765209197998,
      "learning_rate": 9.99252895031752e-06,
      "loss": 0.1566,
      "step": 2440
    },
    {
      "epoch": 20.45929018789144,
      "grad_norm": 2.987387180328369,
      "learning_rate": 9.80575270825551e-06,
      "loss": 0.1399,
      "step": 2450
    },
    {
      "epoch": 20.542797494780793,
      "grad_norm": 4.639241695404053,
      "learning_rate": 9.6189764661935e-06,
      "loss": 0.1447,
      "step": 2460
    },
    {
      "epoch": 20.626304801670145,
      "grad_norm": 2.296684741973877,
      "learning_rate": 9.43220022413149e-06,
      "loss": 0.1392,
      "step": 2470
    },
    {
      "epoch": 20.7098121085595,
      "grad_norm": 2.6289243698120117,
      "learning_rate": 9.245423982069482e-06,
      "loss": 0.1488,
      "step": 2480
    },
    {
      "epoch": 20.793319415448853,
      "grad_norm": 3.7014503479003906,
      "learning_rate": 9.058647740007472e-06,
      "loss": 0.1495,
      "step": 2490
    },
    {
      "epoch": 20.876826722338205,
      "grad_norm": 3.3644683361053467,
      "learning_rate": 8.871871497945462e-06,
      "loss": 0.1512,
      "step": 2500
    },
    {
      "epoch": 20.960334029227557,
      "grad_norm": 3.517514705657959,
      "learning_rate": 8.685095255883451e-06,
      "loss": 0.1504,
      "step": 2510
    },
    {
      "epoch": 20.993736951983298,
      "eval_accuracy": 0.9381614857218931,
      "eval_loss": 0.22127582132816315,
      "eval_runtime": 46.9949,
      "eval_samples_per_second": 319.673,
      "eval_steps_per_second": 5.001,
      "step": 2514
    },
    {
      "epoch": 21.04384133611691,
      "grad_norm": 2.8736536502838135,
      "learning_rate": 8.498319013821441e-06,
      "loss": 0.1439,
      "step": 2520
    },
    {
      "epoch": 21.127348643006265,
      "grad_norm": 3.371739387512207,
      "learning_rate": 8.311542771759433e-06,
      "loss": 0.1465,
      "step": 2530
    },
    {
      "epoch": 21.210855949895617,
      "grad_norm": 4.341720104217529,
      "learning_rate": 8.124766529697423e-06,
      "loss": 0.1272,
      "step": 2540
    },
    {
      "epoch": 21.29436325678497,
      "grad_norm": 4.213035583496094,
      "learning_rate": 7.937990287635413e-06,
      "loss": 0.1372,
      "step": 2550
    },
    {
      "epoch": 21.37787056367432,
      "grad_norm": 4.820446491241455,
      "learning_rate": 7.751214045573403e-06,
      "loss": 0.1413,
      "step": 2560
    },
    {
      "epoch": 21.461377870563673,
      "grad_norm": 5.027878761291504,
      "learning_rate": 7.564437803511394e-06,
      "loss": 0.1342,
      "step": 2570
    },
    {
      "epoch": 21.544885177453025,
      "grad_norm": 3.3291218280792236,
      "learning_rate": 7.3776615614493835e-06,
      "loss": 0.1385,
      "step": 2580
    },
    {
      "epoch": 21.62839248434238,
      "grad_norm": 4.18372106552124,
      "learning_rate": 7.190885319387375e-06,
      "loss": 0.1597,
      "step": 2590
    },
    {
      "epoch": 21.711899791231733,
      "grad_norm": 3.464853048324585,
      "learning_rate": 7.004109077325364e-06,
      "loss": 0.1516,
      "step": 2600
    },
    {
      "epoch": 21.795407098121085,
      "grad_norm": 3.7913620471954346,
      "learning_rate": 6.817332835263354e-06,
      "loss": 0.1407,
      "step": 2610
    },
    {
      "epoch": 21.878914405010438,
      "grad_norm": 3.104400873184204,
      "learning_rate": 6.630556593201345e-06,
      "loss": 0.1548,
      "step": 2620
    },
    {
      "epoch": 21.96242171189979,
      "grad_norm": 3.58847975730896,
      "learning_rate": 6.443780351139335e-06,
      "loss": 0.139,
      "step": 2630
    },
    {
      "epoch": 21.995824634655534,
      "eval_accuracy": 0.9370298875058244,
      "eval_loss": 0.2247340828180313,
      "eval_runtime": 47.4958,
      "eval_samples_per_second": 316.302,
      "eval_steps_per_second": 4.948,
      "step": 2634
    },
    {
      "epoch": 22.045929018789145,
      "grad_norm": 2.6478941440582275,
      "learning_rate": 6.257004109077326e-06,
      "loss": 0.1507,
      "step": 2640
    },
    {
      "epoch": 22.129436325678498,
      "grad_norm": 3.125650644302368,
      "learning_rate": 6.0702278670153155e-06,
      "loss": 0.1503,
      "step": 2650
    },
    {
      "epoch": 22.21294363256785,
      "grad_norm": 2.9791300296783447,
      "learning_rate": 5.883451624953306e-06,
      "loss": 0.1317,
      "step": 2660
    },
    {
      "epoch": 22.296450939457202,
      "grad_norm": 3.1157150268554688,
      "learning_rate": 5.696675382891296e-06,
      "loss": 0.1584,
      "step": 2670
    },
    {
      "epoch": 22.379958246346554,
      "grad_norm": 3.433903217315674,
      "learning_rate": 5.509899140829287e-06,
      "loss": 0.1424,
      "step": 2680
    },
    {
      "epoch": 22.46346555323591,
      "grad_norm": 2.764221429824829,
      "learning_rate": 5.323122898767277e-06,
      "loss": 0.135,
      "step": 2690
    },
    {
      "epoch": 22.546972860125262,
      "grad_norm": 6.406319618225098,
      "learning_rate": 5.136346656705268e-06,
      "loss": 0.1307,
      "step": 2700
    },
    {
      "epoch": 22.630480167014614,
      "grad_norm": 2.927151918411255,
      "learning_rate": 4.949570414643258e-06,
      "loss": 0.1259,
      "step": 2710
    },
    {
      "epoch": 22.713987473903966,
      "grad_norm": 2.335951089859009,
      "learning_rate": 4.762794172581248e-06,
      "loss": 0.138,
      "step": 2720
    },
    {
      "epoch": 22.79749478079332,
      "grad_norm": 3.3406760692596436,
      "learning_rate": 4.5760179305192375e-06,
      "loss": 0.1295,
      "step": 2730
    },
    {
      "epoch": 22.88100208768267,
      "grad_norm": 4.615182399749756,
      "learning_rate": 4.389241688457228e-06,
      "loss": 0.146,
      "step": 2740
    },
    {
      "epoch": 22.964509394572026,
      "grad_norm": 3.860175848007202,
      "learning_rate": 4.202465446395218e-06,
      "loss": 0.1264,
      "step": 2750
    },
    {
      "epoch": 22.997912317327767,
      "eval_accuracy": 0.938427744125674,
      "eval_loss": 0.23570792376995087,
      "eval_runtime": 47.1554,
      "eval_samples_per_second": 318.585,
      "eval_steps_per_second": 4.984,
      "step": 2754
    },
    {
      "epoch": 23.04801670146138,
      "grad_norm": 3.362682580947876,
      "learning_rate": 4.015689204333209e-06,
      "loss": 0.1557,
      "step": 2760
    },
    {
      "epoch": 23.13152400835073,
      "grad_norm": 4.399153709411621,
      "learning_rate": 3.8289129622712e-06,
      "loss": 0.1472,
      "step": 2770
    },
    {
      "epoch": 23.215031315240083,
      "grad_norm": 3.276092052459717,
      "learning_rate": 3.6421367202091897e-06,
      "loss": 0.1266,
      "step": 2780
    },
    {
      "epoch": 23.298538622129435,
      "grad_norm": 3.5454556941986084,
      "learning_rate": 3.45536047814718e-06,
      "loss": 0.1237,
      "step": 2790
    },
    {
      "epoch": 23.38204592901879,
      "grad_norm": 4.570245742797852,
      "learning_rate": 3.2685842360851704e-06,
      "loss": 0.138,
      "step": 2800
    },
    {
      "epoch": 23.465553235908143,
      "grad_norm": 4.357471466064453,
      "learning_rate": 3.0818079940231603e-06,
      "loss": 0.1675,
      "step": 2810
    },
    {
      "epoch": 23.549060542797495,
      "grad_norm": 2.96999192237854,
      "learning_rate": 2.8950317519611506e-06,
      "loss": 0.1432,
      "step": 2820
    },
    {
      "epoch": 23.632567849686847,
      "grad_norm": 2.6909894943237305,
      "learning_rate": 2.708255509899141e-06,
      "loss": 0.1306,
      "step": 2830
    },
    {
      "epoch": 23.7160751565762,
      "grad_norm": 2.8530972003936768,
      "learning_rate": 2.5214792678371313e-06,
      "loss": 0.1418,
      "step": 2840
    },
    {
      "epoch": 23.799582463465555,
      "grad_norm": 3.2527079582214355,
      "learning_rate": 2.3347030257751217e-06,
      "loss": 0.1367,
      "step": 2850
    },
    {
      "epoch": 23.883089770354907,
      "grad_norm": 4.381546497344971,
      "learning_rate": 2.147926783713112e-06,
      "loss": 0.1324,
      "step": 2860
    },
    {
      "epoch": 23.96659707724426,
      "grad_norm": 3.197265863418579,
      "learning_rate": 1.9611505416511024e-06,
      "loss": 0.1266,
      "step": 2870
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.9380949211209478,
      "eval_loss": 0.23597599565982819,
      "eval_runtime": 47.4629,
      "eval_samples_per_second": 316.521,
      "eval_steps_per_second": 4.951,
      "step": 2874
    },
    {
      "epoch": 24.05010438413361,
      "grad_norm": 4.494243621826172,
      "learning_rate": 1.7743742995890923e-06,
      "loss": 0.1277,
      "step": 2880
    },
    {
      "epoch": 24.133611691022963,
      "grad_norm": 2.9121129512786865,
      "learning_rate": 1.5875980575270827e-06,
      "loss": 0.1119,
      "step": 2890
    },
    {
      "epoch": 24.217118997912316,
      "grad_norm": 4.936645984649658,
      "learning_rate": 1.400821815465073e-06,
      "loss": 0.1321,
      "step": 2900
    },
    {
      "epoch": 24.30062630480167,
      "grad_norm": 3.3100907802581787,
      "learning_rate": 1.2140455734030631e-06,
      "loss": 0.1347,
      "step": 2910
    },
    {
      "epoch": 24.384133611691023,
      "grad_norm": 3.270296096801758,
      "learning_rate": 1.0272693313410535e-06,
      "loss": 0.1339,
      "step": 2920
    },
    {
      "epoch": 24.467640918580376,
      "grad_norm": 5.395968437194824,
      "learning_rate": 8.404930892790437e-07,
      "loss": 0.1348,
      "step": 2930
    },
    {
      "epoch": 24.551148225469728,
      "grad_norm": 3.4229140281677246,
      "learning_rate": 6.53716847217034e-07,
      "loss": 0.1362,
      "step": 2940
    },
    {
      "epoch": 24.63465553235908,
      "grad_norm": 3.543994188308716,
      "learning_rate": 4.669406051550243e-07,
      "loss": 0.1277,
      "step": 2950
    },
    {
      "epoch": 24.718162839248436,
      "grad_norm": 3.7220823764801025,
      "learning_rate": 2.801643630930146e-07,
      "loss": 0.1366,
      "step": 2960
    },
    {
      "epoch": 24.801670146137788,
      "grad_norm": 3.1788253784179688,
      "learning_rate": 9.338812103100486e-08,
      "loss": 0.1144,
      "step": 2970
    },
    {
      "epoch": 24.843423799582464,
      "eval_accuracy": 0.9374958397124409,
      "eval_loss": 0.23700778186321259,
      "eval_runtime": 47.4891,
      "eval_samples_per_second": 316.346,
      "eval_steps_per_second": 4.948,
      "step": 2975
    },
    {
      "epoch": 24.843423799582464,
      "step": 2975,
      "total_flos": 1.488814196353273e+19,
      "train_loss": 0.23466382104809544,
      "train_runtime": 6976.8228,
      "train_samples_per_second": 109.728,
      "train_steps_per_second": 0.426
    }
  ],
  "logging_steps": 10,
  "max_steps": 2975,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 25,
  "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": 1.488814196353273e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}