{
  "best_metric": 0.08559587597846985,
  "best_model_checkpoint": "./vit-base-chest-xray/checkpoint-500",
  "epoch": 4.0,
  "eval_steps": 100,
  "global_step": 3060,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.013071895424836602,
      "grad_norm": 1.1339586973190308,
      "learning_rate": 0.0001993464052287582,
      "loss": 0.4681,
      "step": 10
    },
    {
      "epoch": 0.026143790849673203,
      "grad_norm": 3.4900288581848145,
      "learning_rate": 0.00019869281045751635,
      "loss": 0.2909,
      "step": 20
    },
    {
      "epoch": 0.0392156862745098,
      "grad_norm": 4.287875652313232,
      "learning_rate": 0.00019803921568627454,
      "loss": 0.2672,
      "step": 30
    },
    {
      "epoch": 0.05228758169934641,
      "grad_norm": 0.31267789006233215,
      "learning_rate": 0.0001973856209150327,
      "loss": 0.1288,
      "step": 40
    },
    {
      "epoch": 0.06535947712418301,
      "grad_norm": 0.20445282757282257,
      "learning_rate": 0.00019673202614379085,
      "loss": 0.2329,
      "step": 50
    },
    {
      "epoch": 0.0784313725490196,
      "grad_norm": 1.763299584388733,
      "learning_rate": 0.000196078431372549,
      "loss": 0.1666,
      "step": 60
    },
    {
      "epoch": 0.0915032679738562,
      "grad_norm": 4.610985279083252,
      "learning_rate": 0.0001954248366013072,
      "loss": 0.1897,
      "step": 70
    },
    {
      "epoch": 0.10457516339869281,
      "grad_norm": 0.2082081139087677,
      "learning_rate": 0.00019477124183006535,
      "loss": 0.1638,
      "step": 80
    },
    {
      "epoch": 0.11764705882352941,
      "grad_norm": 2.0618653297424316,
      "learning_rate": 0.00019411764705882354,
      "loss": 0.1161,
      "step": 90
    },
    {
      "epoch": 0.13071895424836602,
      "grad_norm": 2.996493101119995,
      "learning_rate": 0.0001934640522875817,
      "loss": 0.1891,
      "step": 100
    },
    {
      "epoch": 0.13071895424836602,
      "eval_accuracy": 0.9665236051502146,
      "eval_loss": 0.10277536511421204,
      "eval_runtime": 24.9127,
      "eval_samples_per_second": 46.763,
      "eval_steps_per_second": 5.86,
      "step": 100
    },
    {
      "epoch": 0.1437908496732026,
      "grad_norm": 1.4748905897140503,
      "learning_rate": 0.00019281045751633988,
      "loss": 0.181,
      "step": 110
    },
    {
      "epoch": 0.1568627450980392,
      "grad_norm": 2.811959743499756,
      "learning_rate": 0.00019215686274509807,
      "loss": 0.2343,
      "step": 120
    },
    {
      "epoch": 0.16993464052287582,
      "grad_norm": 1.9155172109603882,
      "learning_rate": 0.00019150326797385623,
      "loss": 0.1792,
      "step": 130
    },
    {
      "epoch": 0.1830065359477124,
      "grad_norm": 0.3125442564487457,
      "learning_rate": 0.0001908496732026144,
      "loss": 0.1898,
      "step": 140
    },
    {
      "epoch": 0.19607843137254902,
      "grad_norm": 0.10497033596038818,
      "learning_rate": 0.00019019607843137254,
      "loss": 0.0737,
      "step": 150
    },
    {
      "epoch": 0.20915032679738563,
      "grad_norm": 0.12402518093585968,
      "learning_rate": 0.00018954248366013073,
      "loss": 0.0448,
      "step": 160
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": 3.1576356887817383,
      "learning_rate": 0.00018888888888888888,
      "loss": 0.2302,
      "step": 170
    },
    {
      "epoch": 0.23529411764705882,
      "grad_norm": 0.15159478783607483,
      "learning_rate": 0.00018823529411764707,
      "loss": 0.094,
      "step": 180
    },
    {
      "epoch": 0.24836601307189543,
      "grad_norm": 0.0732378214597702,
      "learning_rate": 0.00018758169934640523,
      "loss": 0.1228,
      "step": 190
    },
    {
      "epoch": 0.26143790849673204,
      "grad_norm": 3.1713297367095947,
      "learning_rate": 0.0001869281045751634,
      "loss": 0.2123,
      "step": 200
    },
    {
      "epoch": 0.26143790849673204,
      "eval_accuracy": 0.9562231759656652,
      "eval_loss": 0.12540382146835327,
      "eval_runtime": 15.5185,
      "eval_samples_per_second": 75.072,
      "eval_steps_per_second": 9.408,
      "step": 200
    },
    {
      "epoch": 0.27450980392156865,
      "grad_norm": 0.42199578881263733,
      "learning_rate": 0.00018633986928104577,
      "loss": 0.1039,
      "step": 210
    },
    {
      "epoch": 0.2875816993464052,
      "grad_norm": 1.0121151208877563,
      "learning_rate": 0.00018568627450980392,
      "loss": 0.146,
      "step": 220
    },
    {
      "epoch": 0.3006535947712418,
      "grad_norm": 0.16740572452545166,
      "learning_rate": 0.0001850326797385621,
      "loss": 0.1232,
      "step": 230
    },
    {
      "epoch": 0.3137254901960784,
      "grad_norm": 0.7064481973648071,
      "learning_rate": 0.0001843790849673203,
      "loss": 0.2105,
      "step": 240
    },
    {
      "epoch": 0.32679738562091504,
      "grad_norm": 0.23291558027267456,
      "learning_rate": 0.00018372549019607842,
      "loss": 0.1186,
      "step": 250
    },
    {
      "epoch": 0.33986928104575165,
      "grad_norm": 0.3717576861381531,
      "learning_rate": 0.0001830718954248366,
      "loss": 0.1671,
      "step": 260
    },
    {
      "epoch": 0.35294117647058826,
      "grad_norm": 2.2546300888061523,
      "learning_rate": 0.00018241830065359477,
      "loss": 0.1799,
      "step": 270
    },
    {
      "epoch": 0.3660130718954248,
      "grad_norm": 1.6236902475357056,
      "learning_rate": 0.00018176470588235295,
      "loss": 0.5579,
      "step": 280
    },
    {
      "epoch": 0.3790849673202614,
      "grad_norm": 0.9936646819114685,
      "learning_rate": 0.0001811111111111111,
      "loss": 0.1809,
      "step": 290
    },
    {
      "epoch": 0.39215686274509803,
      "grad_norm": 4.444874286651611,
      "learning_rate": 0.0001804575163398693,
      "loss": 0.0536,
      "step": 300
    },
    {
      "epoch": 0.39215686274509803,
      "eval_accuracy": 0.9690987124463519,
      "eval_loss": 0.11418166011571884,
      "eval_runtime": 14.394,
      "eval_samples_per_second": 80.937,
      "eval_steps_per_second": 10.143,
      "step": 300
    },
    {
      "epoch": 0.40522875816993464,
      "grad_norm": 0.05227786302566528,
      "learning_rate": 0.00017980392156862745,
      "loss": 0.0896,
      "step": 310
    },
    {
      "epoch": 0.41830065359477125,
      "grad_norm": 0.06666386872529984,
      "learning_rate": 0.00017915032679738564,
      "loss": 0.0992,
      "step": 320
    },
    {
      "epoch": 0.43137254901960786,
      "grad_norm": 0.2057139277458191,
      "learning_rate": 0.0001784967320261438,
      "loss": 0.1225,
      "step": 330
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 1.088228702545166,
      "learning_rate": 0.00017784313725490195,
      "loss": 0.0892,
      "step": 340
    },
    {
      "epoch": 0.45751633986928103,
      "grad_norm": 0.7964250445365906,
      "learning_rate": 0.00017718954248366014,
      "loss": 0.1447,
      "step": 350
    },
    {
      "epoch": 0.47058823529411764,
      "grad_norm": 0.7459284663200378,
      "learning_rate": 0.0001765359477124183,
      "loss": 0.1589,
      "step": 360
    },
    {
      "epoch": 0.48366013071895425,
      "grad_norm": 0.40795373916625977,
      "learning_rate": 0.00017588235294117648,
      "loss": 0.2556,
      "step": 370
    },
    {
      "epoch": 0.49673202614379086,
      "grad_norm": 2.7890219688415527,
      "learning_rate": 0.00017522875816993464,
      "loss": 0.1933,
      "step": 380
    },
    {
      "epoch": 0.5098039215686274,
      "grad_norm": 1.0104824304580688,
      "learning_rate": 0.00017457516339869283,
      "loss": 0.1709,
      "step": 390
    },
    {
      "epoch": 0.5228758169934641,
      "grad_norm": 0.8147939443588257,
      "learning_rate": 0.00017392156862745098,
      "loss": 0.0799,
      "step": 400
    },
    {
      "epoch": 0.5228758169934641,
      "eval_accuracy": 0.9648068669527897,
      "eval_loss": 0.1173361986875534,
      "eval_runtime": 14.4858,
      "eval_samples_per_second": 80.423,
      "eval_steps_per_second": 10.079,
      "step": 400
    },
    {
      "epoch": 0.5359477124183006,
      "grad_norm": 0.25372856855392456,
      "learning_rate": 0.00017326797385620917,
      "loss": 0.1718,
      "step": 410
    },
    {
      "epoch": 0.5490196078431373,
      "grad_norm": 0.6395456194877625,
      "learning_rate": 0.00017261437908496733,
      "loss": 0.1003,
      "step": 420
    },
    {
      "epoch": 0.5620915032679739,
      "grad_norm": 1.1706054210662842,
      "learning_rate": 0.0001719607843137255,
      "loss": 0.2023,
      "step": 430
    },
    {
      "epoch": 0.5751633986928104,
      "grad_norm": 1.8857362270355225,
      "learning_rate": 0.00017130718954248367,
      "loss": 0.1288,
      "step": 440
    },
    {
      "epoch": 0.5882352941176471,
      "grad_norm": 0.07465291023254395,
      "learning_rate": 0.00017065359477124183,
      "loss": 0.0458,
      "step": 450
    },
    {
      "epoch": 0.6013071895424836,
      "grad_norm": 0.05449380353093147,
      "learning_rate": 0.00017,
      "loss": 0.112,
      "step": 460
    },
    {
      "epoch": 0.6143790849673203,
      "grad_norm": 2.8383517265319824,
      "learning_rate": 0.00016934640522875817,
      "loss": 0.1434,
      "step": 470
    },
    {
      "epoch": 0.6274509803921569,
      "grad_norm": 2.2419204711914062,
      "learning_rate": 0.00016869281045751636,
      "loss": 0.1108,
      "step": 480
    },
    {
      "epoch": 0.6405228758169934,
      "grad_norm": 1.1179159879684448,
      "learning_rate": 0.0001680392156862745,
      "loss": 0.0392,
      "step": 490
    },
    {
      "epoch": 0.6535947712418301,
      "grad_norm": 1.6863864660263062,
      "learning_rate": 0.0001673856209150327,
      "loss": 0.0537,
      "step": 500
    },
    {
      "epoch": 0.6535947712418301,
      "eval_accuracy": 0.9742489270386266,
      "eval_loss": 0.08559587597846985,
      "eval_runtime": 15.4,
      "eval_samples_per_second": 75.649,
      "eval_steps_per_second": 9.481,
      "step": 500
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 2.383035659790039,
      "learning_rate": 0.00016673202614379086,
      "loss": 0.2613,
      "step": 510
    },
    {
      "epoch": 0.6797385620915033,
      "grad_norm": 0.652946412563324,
      "learning_rate": 0.00016607843137254904,
      "loss": 0.1499,
      "step": 520
    },
    {
      "epoch": 0.6928104575163399,
      "grad_norm": 1.1430766582489014,
      "learning_rate": 0.0001654248366013072,
      "loss": 0.082,
      "step": 530
    },
    {
      "epoch": 0.7058823529411765,
      "grad_norm": 0.6167986989021301,
      "learning_rate": 0.00016477124183006536,
      "loss": 0.017,
      "step": 540
    },
    {
      "epoch": 0.7189542483660131,
      "grad_norm": 0.06140501797199249,
      "learning_rate": 0.00016411764705882354,
      "loss": 0.0802,
      "step": 550
    },
    {
      "epoch": 0.7320261437908496,
      "grad_norm": 0.047364696860313416,
      "learning_rate": 0.0001634640522875817,
      "loss": 0.1183,
      "step": 560
    },
    {
      "epoch": 0.7450980392156863,
      "grad_norm": 0.4764581024646759,
      "learning_rate": 0.00016281045751633989,
      "loss": 0.0451,
      "step": 570
    },
    {
      "epoch": 0.7581699346405228,
      "grad_norm": 0.05738784372806549,
      "learning_rate": 0.00016215686274509804,
      "loss": 0.0137,
      "step": 580
    },
    {
      "epoch": 0.7712418300653595,
      "grad_norm": 2.4237992763519287,
      "learning_rate": 0.00016150326797385623,
      "loss": 0.0707,
      "step": 590
    },
    {
      "epoch": 0.7843137254901961,
      "grad_norm": 0.10018886625766754,
      "learning_rate": 0.00016084967320261439,
      "loss": 0.0911,
      "step": 600
    },
    {
      "epoch": 0.7843137254901961,
      "eval_accuracy": 0.9424892703862661,
      "eval_loss": 0.20054155588150024,
      "eval_runtime": 14.614,
      "eval_samples_per_second": 79.718,
      "eval_steps_per_second": 9.99,
      "step": 600
    },
    {
      "epoch": 0.7973856209150327,
      "grad_norm": 0.5596518516540527,
      "learning_rate": 0.00016019607843137257,
      "loss": 0.0513,
      "step": 610
    },
    {
      "epoch": 0.8104575163398693,
      "grad_norm": 0.052388403564691544,
      "learning_rate": 0.00015954248366013073,
      "loss": 0.0463,
      "step": 620
    },
    {
      "epoch": 0.8235294117647058,
      "grad_norm": 0.567359983921051,
      "learning_rate": 0.0001588888888888889,
      "loss": 0.0966,
      "step": 630
    },
    {
      "epoch": 0.8366013071895425,
      "grad_norm": 2.3820440769195557,
      "learning_rate": 0.00015823529411764707,
      "loss": 0.0887,
      "step": 640
    },
    {
      "epoch": 0.8496732026143791,
      "grad_norm": 0.05958991125226021,
      "learning_rate": 0.00015758169934640523,
      "loss": 0.1156,
      "step": 650
    },
    {
      "epoch": 0.8627450980392157,
      "grad_norm": 0.09057465195655823,
      "learning_rate": 0.00015692810457516342,
      "loss": 0.0545,
      "step": 660
    },
    {
      "epoch": 0.8758169934640523,
      "grad_norm": 0.6337572932243347,
      "learning_rate": 0.00015627450980392157,
      "loss": 0.0771,
      "step": 670
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 0.7565958499908447,
      "learning_rate": 0.00015562091503267976,
      "loss": 0.1179,
      "step": 680
    },
    {
      "epoch": 0.9019607843137255,
      "grad_norm": 3.6693177223205566,
      "learning_rate": 0.00015496732026143792,
      "loss": 0.2026,
      "step": 690
    },
    {
      "epoch": 0.9150326797385621,
      "grad_norm": 0.13483676314353943,
      "learning_rate": 0.0001543137254901961,
      "loss": 0.1027,
      "step": 700
    },
    {
      "epoch": 0.9150326797385621,
      "eval_accuracy": 0.9708154506437768,
      "eval_loss": 0.08694365620613098,
      "eval_runtime": 14.2568,
      "eval_samples_per_second": 81.715,
      "eval_steps_per_second": 10.241,
      "step": 700
    },
    {
      "epoch": 0.9281045751633987,
      "grad_norm": 0.8454955220222473,
      "learning_rate": 0.00015366013071895426,
      "loss": 0.0757,
      "step": 710
    },
    {
      "epoch": 0.9411764705882353,
      "grad_norm": 4.228348255157471,
      "learning_rate": 0.00015300653594771242,
      "loss": 0.0729,
      "step": 720
    },
    {
      "epoch": 0.954248366013072,
      "grad_norm": 0.35448652505874634,
      "learning_rate": 0.00015235294117647057,
      "loss": 0.0605,
      "step": 730
    },
    {
      "epoch": 0.9673202614379085,
      "grad_norm": 0.04271356761455536,
      "learning_rate": 0.00015169934640522876,
      "loss": 0.0747,
      "step": 740
    },
    {
      "epoch": 0.9803921568627451,
      "grad_norm": 0.11471935361623764,
      "learning_rate": 0.00015104575163398694,
      "loss": 0.0345,
      "step": 750
    },
    {
      "epoch": 0.9934640522875817,
      "grad_norm": 0.3417186439037323,
      "learning_rate": 0.0001503921568627451,
      "loss": 0.1179,
      "step": 760
    },
    {
      "epoch": 1.0065359477124183,
      "grad_norm": 0.07955777645111084,
      "learning_rate": 0.0001497385620915033,
      "loss": 0.0091,
      "step": 770
    },
    {
      "epoch": 1.0196078431372548,
      "grad_norm": 0.07078742980957031,
      "learning_rate": 0.00014908496732026145,
      "loss": 0.0229,
      "step": 780
    },
    {
      "epoch": 1.0326797385620916,
      "grad_norm": 0.11356745660305023,
      "learning_rate": 0.00014843137254901963,
      "loss": 0.0351,
      "step": 790
    },
    {
      "epoch": 1.0457516339869282,
      "grad_norm": 1.0611521005630493,
      "learning_rate": 0.0001477777777777778,
      "loss": 0.1011,
      "step": 800
    },
    {
      "epoch": 1.0457516339869282,
      "eval_accuracy": 0.9630901287553648,
      "eval_loss": 0.10628381371498108,
      "eval_runtime": 14.4937,
      "eval_samples_per_second": 80.38,
      "eval_steps_per_second": 10.073,
      "step": 800
    },
    {
      "epoch": 1.0588235294117647,
      "grad_norm": 0.5016999244689941,
      "learning_rate": 0.00014712418300653597,
      "loss": 0.0779,
      "step": 810
    },
    {
      "epoch": 1.0718954248366013,
      "grad_norm": 0.2654229402542114,
      "learning_rate": 0.0001464705882352941,
      "loss": 0.1233,
      "step": 820
    },
    {
      "epoch": 1.0849673202614378,
      "grad_norm": 0.16079501807689667,
      "learning_rate": 0.0001458169934640523,
      "loss": 0.07,
      "step": 830
    },
    {
      "epoch": 1.0980392156862746,
      "grad_norm": 0.08982982486486435,
      "learning_rate": 0.00014516339869281045,
      "loss": 0.0525,
      "step": 840
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 0.029190940782427788,
      "learning_rate": 0.00014450980392156863,
      "loss": 0.0844,
      "step": 850
    },
    {
      "epoch": 1.1241830065359477,
      "grad_norm": 0.035515930503606796,
      "learning_rate": 0.0001438562091503268,
      "loss": 0.0231,
      "step": 860
    },
    {
      "epoch": 1.1372549019607843,
      "grad_norm": 0.19514258205890656,
      "learning_rate": 0.00014320261437908498,
      "loss": 0.0435,
      "step": 870
    },
    {
      "epoch": 1.1503267973856208,
      "grad_norm": 0.04219866171479225,
      "learning_rate": 0.00014254901960784316,
      "loss": 0.0517,
      "step": 880
    },
    {
      "epoch": 1.1633986928104576,
      "grad_norm": 0.0360812209546566,
      "learning_rate": 0.00014189542483660132,
      "loss": 0.0311,
      "step": 890
    },
    {
      "epoch": 1.1764705882352942,
      "grad_norm": 0.4207318425178528,
      "learning_rate": 0.0001412418300653595,
      "loss": 0.0717,
      "step": 900
    },
    {
      "epoch": 1.1764705882352942,
      "eval_accuracy": 0.9587982832618026,
      "eval_loss": 0.14242781698703766,
      "eval_runtime": 15.8491,
      "eval_samples_per_second": 73.506,
      "eval_steps_per_second": 9.212,
      "step": 900
    },
    {
      "epoch": 1.1895424836601307,
      "grad_norm": 2.1643173694610596,
      "learning_rate": 0.00014058823529411763,
      "loss": 0.0269,
      "step": 910
    },
    {
      "epoch": 1.2026143790849673,
      "grad_norm": 0.03318187966942787,
      "learning_rate": 0.00013993464052287582,
      "loss": 0.0272,
      "step": 920
    },
    {
      "epoch": 1.215686274509804,
      "grad_norm": 0.7758973240852356,
      "learning_rate": 0.00013928104575163398,
      "loss": 0.0633,
      "step": 930
    },
    {
      "epoch": 1.2287581699346406,
      "grad_norm": 1.6854296922683716,
      "learning_rate": 0.00013862745098039216,
      "loss": 0.0678,
      "step": 940
    },
    {
      "epoch": 1.2418300653594772,
      "grad_norm": 3.3588085174560547,
      "learning_rate": 0.00013797385620915032,
      "loss": 0.1088,
      "step": 950
    },
    {
      "epoch": 1.2549019607843137,
      "grad_norm": 0.2561059892177582,
      "learning_rate": 0.0001373202614379085,
      "loss": 0.0657,
      "step": 960
    },
    {
      "epoch": 1.2679738562091503,
      "grad_norm": 1.3019442558288574,
      "learning_rate": 0.00013666666666666666,
      "loss": 0.0405,
      "step": 970
    },
    {
      "epoch": 1.2810457516339868,
      "grad_norm": 0.23876748979091644,
      "learning_rate": 0.00013601307189542485,
      "loss": 0.0277,
      "step": 980
    },
    {
      "epoch": 1.2941176470588236,
      "grad_norm": 0.21840398013591766,
      "learning_rate": 0.00013535947712418303,
      "loss": 0.0356,
      "step": 990
    },
    {
      "epoch": 1.3071895424836601,
      "grad_norm": 0.05377864092588425,
      "learning_rate": 0.0001347058823529412,
      "loss": 0.0605,
      "step": 1000
    },
    {
      "epoch": 1.3071895424836601,
      "eval_accuracy": 0.9648068669527897,
      "eval_loss": 0.1524645835161209,
      "eval_runtime": 14.9355,
      "eval_samples_per_second": 78.002,
      "eval_steps_per_second": 9.775,
      "step": 1000
    },
    {
      "epoch": 1.3202614379084967,
      "grad_norm": 0.13815192878246307,
      "learning_rate": 0.00013405228758169935,
      "loss": 0.0715,
      "step": 1010
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 3.1915783882141113,
      "learning_rate": 0.0001333986928104575,
      "loss": 0.0762,
      "step": 1020
    },
    {
      "epoch": 1.34640522875817,
      "grad_norm": 0.12432447075843811,
      "learning_rate": 0.0001327450980392157,
      "loss": 0.0514,
      "step": 1030
    },
    {
      "epoch": 1.3594771241830066,
      "grad_norm": 0.029387619346380234,
      "learning_rate": 0.00013209150326797385,
      "loss": 0.0094,
      "step": 1040
    },
    {
      "epoch": 1.3725490196078431,
      "grad_norm": 0.02788469009101391,
      "learning_rate": 0.00013143790849673204,
      "loss": 0.0128,
      "step": 1050
    },
    {
      "epoch": 1.3856209150326797,
      "grad_norm": 0.38828742504119873,
      "learning_rate": 0.0001307843137254902,
      "loss": 0.1181,
      "step": 1060
    },
    {
      "epoch": 1.3986928104575163,
      "grad_norm": 0.029526783153414726,
      "learning_rate": 0.00013013071895424838,
      "loss": 0.0701,
      "step": 1070
    },
    {
      "epoch": 1.4117647058823528,
      "grad_norm": 8.863903999328613,
      "learning_rate": 0.00012947712418300654,
      "loss": 0.0349,
      "step": 1080
    },
    {
      "epoch": 1.4248366013071896,
      "grad_norm": 0.05002022534608841,
      "learning_rate": 0.00012882352941176472,
      "loss": 0.0111,
      "step": 1090
    },
    {
      "epoch": 1.4379084967320261,
      "grad_norm": 0.04968211427330971,
      "learning_rate": 0.00012816993464052288,
      "loss": 0.0573,
      "step": 1100
    },
    {
      "epoch": 1.4379084967320261,
      "eval_accuracy": 0.9699570815450643,
      "eval_loss": 0.09700144827365875,
      "eval_runtime": 14.9347,
      "eval_samples_per_second": 78.006,
      "eval_steps_per_second": 9.776,
      "step": 1100
    },
    {
      "epoch": 1.4509803921568627,
      "grad_norm": 0.021018337458372116,
      "learning_rate": 0.00012751633986928104,
      "loss": 0.0338,
      "step": 1110
    },
    {
      "epoch": 1.4640522875816995,
      "grad_norm": 0.05960996448993683,
      "learning_rate": 0.00012686274509803922,
      "loss": 0.0577,
      "step": 1120
    },
    {
      "epoch": 1.477124183006536,
      "grad_norm": 0.02618427947163582,
      "learning_rate": 0.00012620915032679738,
      "loss": 0.055,
      "step": 1130
    },
    {
      "epoch": 1.4901960784313726,
      "grad_norm": 0.05656788498163223,
      "learning_rate": 0.00012555555555555557,
      "loss": 0.1561,
      "step": 1140
    },
    {
      "epoch": 1.5032679738562091,
      "grad_norm": 0.046731386333703995,
      "learning_rate": 0.00012490196078431372,
      "loss": 0.0314,
      "step": 1150
    },
    {
      "epoch": 1.5163398692810457,
      "grad_norm": 0.06606756895780563,
      "learning_rate": 0.0001242483660130719,
      "loss": 0.0359,
      "step": 1160
    },
    {
      "epoch": 1.5294117647058822,
      "grad_norm": 0.065465547144413,
      "learning_rate": 0.00012359477124183007,
      "loss": 0.0312,
      "step": 1170
    },
    {
      "epoch": 1.5424836601307188,
      "grad_norm": 0.05767370015382767,
      "learning_rate": 0.00012294117647058825,
      "loss": 0.0548,
      "step": 1180
    },
    {
      "epoch": 1.5555555555555556,
      "grad_norm": 0.02790243923664093,
      "learning_rate": 0.0001222875816993464,
      "loss": 0.0483,
      "step": 1190
    },
    {
      "epoch": 1.5686274509803921,
      "grad_norm": 0.03534754365682602,
      "learning_rate": 0.00012163398692810457,
      "loss": 0.024,
      "step": 1200
    },
    {
      "epoch": 1.5686274509803921,
      "eval_accuracy": 0.975107296137339,
      "eval_loss": 0.08669988811016083,
      "eval_runtime": 16.5277,
      "eval_samples_per_second": 70.488,
      "eval_steps_per_second": 8.834,
      "step": 1200
    },
    {
      "epoch": 1.581699346405229,
      "grad_norm": 0.03731567785143852,
      "learning_rate": 0.00012098039215686274,
      "loss": 0.0143,
      "step": 1210
    },
    {
      "epoch": 1.5947712418300655,
      "grad_norm": 0.05173814296722412,
      "learning_rate": 0.00012032679738562091,
      "loss": 0.0096,
      "step": 1220
    },
    {
      "epoch": 1.607843137254902,
      "grad_norm": 0.07455731183290482,
      "learning_rate": 0.00011967320261437908,
      "loss": 0.0251,
      "step": 1230
    },
    {
      "epoch": 1.6209150326797386,
      "grad_norm": 0.13736297190189362,
      "learning_rate": 0.00011901960784313725,
      "loss": 0.011,
      "step": 1240
    },
    {
      "epoch": 1.6339869281045751,
      "grad_norm": 2.5091629028320312,
      "learning_rate": 0.00011836601307189544,
      "loss": 0.0372,
      "step": 1250
    },
    {
      "epoch": 1.6470588235294117,
      "grad_norm": 0.048637136816978455,
      "learning_rate": 0.00011771241830065361,
      "loss": 0.0373,
      "step": 1260
    },
    {
      "epoch": 1.6601307189542482,
      "grad_norm": 0.019804753363132477,
      "learning_rate": 0.00011705882352941178,
      "loss": 0.0171,
      "step": 1270
    },
    {
      "epoch": 1.673202614379085,
      "grad_norm": 2.910708427429199,
      "learning_rate": 0.00011640522875816995,
      "loss": 0.0118,
      "step": 1280
    },
    {
      "epoch": 1.6862745098039216,
      "grad_norm": 0.02344467304646969,
      "learning_rate": 0.00011575163398692812,
      "loss": 0.0259,
      "step": 1290
    },
    {
      "epoch": 1.6993464052287581,
      "grad_norm": 0.0272055696696043,
      "learning_rate": 0.00011509803921568627,
      "loss": 0.0056,
      "step": 1300
    },
    {
      "epoch": 1.6993464052287581,
      "eval_accuracy": 0.9759656652360515,
      "eval_loss": 0.08883309364318848,
      "eval_runtime": 14.3413,
      "eval_samples_per_second": 81.234,
      "eval_steps_per_second": 10.18,
      "step": 1300
    },
    {
      "epoch": 1.712418300653595,
      "grad_norm": 0.12058671563863754,
      "learning_rate": 0.00011444444444444444,
      "loss": 0.0078,
      "step": 1310
    },
    {
      "epoch": 1.7254901960784315,
      "grad_norm": 0.012682071886956692,
      "learning_rate": 0.00011379084967320261,
      "loss": 0.014,
      "step": 1320
    },
    {
      "epoch": 1.738562091503268,
      "grad_norm": 0.011953895911574364,
      "learning_rate": 0.00011313725490196078,
      "loss": 0.0195,
      "step": 1330
    },
    {
      "epoch": 1.7516339869281046,
      "grad_norm": 4.999667167663574,
      "learning_rate": 0.00011248366013071895,
      "loss": 0.0472,
      "step": 1340
    },
    {
      "epoch": 1.7647058823529411,
      "grad_norm": 0.011710714548826218,
      "learning_rate": 0.00011183006535947713,
      "loss": 0.0014,
      "step": 1350
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 2.8777430057525635,
      "learning_rate": 0.0001111764705882353,
      "loss": 0.0422,
      "step": 1360
    },
    {
      "epoch": 1.7908496732026142,
      "grad_norm": 0.15594060719013214,
      "learning_rate": 0.00011052287581699348,
      "loss": 0.0046,
      "step": 1370
    },
    {
      "epoch": 1.803921568627451,
      "grad_norm": 0.01740155927836895,
      "learning_rate": 0.00010986928104575165,
      "loss": 0.015,
      "step": 1380
    },
    {
      "epoch": 1.8169934640522876,
      "grad_norm": 0.010606693103909492,
      "learning_rate": 0.0001092156862745098,
      "loss": 0.022,
      "step": 1390
    },
    {
      "epoch": 1.8300653594771243,
      "grad_norm": 0.02794831059873104,
      "learning_rate": 0.00010856209150326797,
      "loss": 0.0051,
      "step": 1400
    },
    {
      "epoch": 1.8300653594771243,
      "eval_accuracy": 0.976824034334764,
      "eval_loss": 0.10543067008256912,
      "eval_runtime": 16.6018,
      "eval_samples_per_second": 70.173,
      "eval_steps_per_second": 8.794,
      "step": 1400
    },
    {
      "epoch": 1.843137254901961,
      "grad_norm": 0.038915157318115234,
      "learning_rate": 0.00010790849673202614,
      "loss": 0.0068,
      "step": 1410
    },
    {
      "epoch": 1.8562091503267975,
      "grad_norm": 0.01105284970253706,
      "learning_rate": 0.00010725490196078431,
      "loss": 0.0011,
      "step": 1420
    },
    {
      "epoch": 1.869281045751634,
      "grad_norm": 0.00978641677647829,
      "learning_rate": 0.00010660130718954248,
      "loss": 0.0286,
      "step": 1430
    },
    {
      "epoch": 1.8823529411764706,
      "grad_norm": 2.092639923095703,
      "learning_rate": 0.00010594771241830066,
      "loss": 0.0618,
      "step": 1440
    },
    {
      "epoch": 1.8954248366013071,
      "grad_norm": 0.11766193807125092,
      "learning_rate": 0.00010529411764705883,
      "loss": 0.0048,
      "step": 1450
    },
    {
      "epoch": 1.9084967320261437,
      "grad_norm": 0.013952341862022877,
      "learning_rate": 0.000104640522875817,
      "loss": 0.0041,
      "step": 1460
    },
    {
      "epoch": 1.9215686274509802,
      "grad_norm": 0.019551385194063187,
      "learning_rate": 0.00010398692810457517,
      "loss": 0.0297,
      "step": 1470
    },
    {
      "epoch": 1.934640522875817,
      "grad_norm": 0.020047994330525398,
      "learning_rate": 0.00010333333333333334,
      "loss": 0.0029,
      "step": 1480
    },
    {
      "epoch": 1.9477124183006536,
      "grad_norm": 1.58188796043396,
      "learning_rate": 0.0001026797385620915,
      "loss": 0.0436,
      "step": 1490
    },
    {
      "epoch": 1.9607843137254903,
      "grad_norm": 5.982904434204102,
      "learning_rate": 0.00010202614379084967,
      "loss": 0.063,
      "step": 1500
    },
    {
      "epoch": 1.9607843137254903,
      "eval_accuracy": 0.9570815450643777,
      "eval_loss": 0.18960346281528473,
      "eval_runtime": 14.9121,
      "eval_samples_per_second": 78.124,
      "eval_steps_per_second": 9.791,
      "step": 1500
    },
    {
      "epoch": 1.973856209150327,
      "grad_norm": 0.1691681146621704,
      "learning_rate": 0.00010137254901960784,
      "loss": 0.0505,
      "step": 1510
    },
    {
      "epoch": 1.9869281045751634,
      "grad_norm": 0.16387327015399933,
      "learning_rate": 0.00010071895424836601,
      "loss": 0.0139,
      "step": 1520
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.01866912469267845,
      "learning_rate": 0.00010006535947712419,
      "loss": 0.0318,
      "step": 1530
    },
    {
      "epoch": 2.0130718954248366,
      "grad_norm": 0.43450427055358887,
      "learning_rate": 9.941176470588236e-05,
      "loss": 0.0491,
      "step": 1540
    },
    {
      "epoch": 2.026143790849673,
      "grad_norm": 0.12364810705184937,
      "learning_rate": 9.875816993464053e-05,
      "loss": 0.0289,
      "step": 1550
    },
    {
      "epoch": 2.0392156862745097,
      "grad_norm": 0.008522446267306805,
      "learning_rate": 9.810457516339869e-05,
      "loss": 0.0225,
      "step": 1560
    },
    {
      "epoch": 2.052287581699346,
      "grad_norm": 0.047463927417993546,
      "learning_rate": 9.745098039215686e-05,
      "loss": 0.0038,
      "step": 1570
    },
    {
      "epoch": 2.065359477124183,
      "grad_norm": 0.008024625480175018,
      "learning_rate": 9.679738562091504e-05,
      "loss": 0.0014,
      "step": 1580
    },
    {
      "epoch": 2.0784313725490198,
      "grad_norm": 0.008238660171627998,
      "learning_rate": 9.614379084967322e-05,
      "loss": 0.0084,
      "step": 1590
    },
    {
      "epoch": 2.0915032679738563,
      "grad_norm": 0.007536654360592365,
      "learning_rate": 9.549019607843139e-05,
      "loss": 0.002,
      "step": 1600
    },
    {
      "epoch": 2.0915032679738563,
      "eval_accuracy": 0.9587982832618026,
      "eval_loss": 0.18859457969665527,
      "eval_runtime": 15.741,
      "eval_samples_per_second": 74.01,
      "eval_steps_per_second": 9.275,
      "step": 1600
    },
    {
      "epoch": 2.104575163398693,
      "grad_norm": 0.012532038614153862,
      "learning_rate": 9.483660130718954e-05,
      "loss": 0.0028,
      "step": 1610
    },
    {
      "epoch": 2.1176470588235294,
      "grad_norm": 0.006995632313191891,
      "learning_rate": 9.418300653594772e-05,
      "loss": 0.0012,
      "step": 1620
    },
    {
      "epoch": 2.130718954248366,
      "grad_norm": 0.007673881947994232,
      "learning_rate": 9.352941176470589e-05,
      "loss": 0.0008,
      "step": 1630
    },
    {
      "epoch": 2.1437908496732025,
      "grad_norm": 0.008022445254027843,
      "learning_rate": 9.287581699346406e-05,
      "loss": 0.0625,
      "step": 1640
    },
    {
      "epoch": 2.156862745098039,
      "grad_norm": 0.008978744968771935,
      "learning_rate": 9.222222222222223e-05,
      "loss": 0.0328,
      "step": 1650
    },
    {
      "epoch": 2.1699346405228757,
      "grad_norm": 0.008565125055611134,
      "learning_rate": 9.156862745098039e-05,
      "loss": 0.0429,
      "step": 1660
    },
    {
      "epoch": 2.183006535947712,
      "grad_norm": 0.009053260087966919,
      "learning_rate": 9.091503267973856e-05,
      "loss": 0.0227,
      "step": 1670
    },
    {
      "epoch": 2.196078431372549,
      "grad_norm": 0.008361272513866425,
      "learning_rate": 9.026143790849673e-05,
      "loss": 0.0011,
      "step": 1680
    },
    {
      "epoch": 2.2091503267973858,
      "grad_norm": 0.008024726063013077,
      "learning_rate": 8.96078431372549e-05,
      "loss": 0.001,
      "step": 1690
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 0.00964928325265646,
      "learning_rate": 8.895424836601307e-05,
      "loss": 0.005,
      "step": 1700
    },
    {
      "epoch": 2.2222222222222223,
      "eval_accuracy": 0.9733905579399141,
      "eval_loss": 0.11838679015636444,
      "eval_runtime": 15.5963,
      "eval_samples_per_second": 74.697,
      "eval_steps_per_second": 9.361,
      "step": 1700
    },
    {
      "epoch": 2.235294117647059,
      "grad_norm": 0.00880915205925703,
      "learning_rate": 8.830065359477125e-05,
      "loss": 0.0011,
      "step": 1710
    },
    {
      "epoch": 2.2483660130718954,
      "grad_norm": 0.655288577079773,
      "learning_rate": 8.764705882352942e-05,
      "loss": 0.0019,
      "step": 1720
    },
    {
      "epoch": 2.261437908496732,
      "grad_norm": 0.0070320917293429375,
      "learning_rate": 8.699346405228759e-05,
      "loss": 0.0009,
      "step": 1730
    },
    {
      "epoch": 2.2745098039215685,
      "grad_norm": 0.006975400261580944,
      "learning_rate": 8.633986928104576e-05,
      "loss": 0.0017,
      "step": 1740
    },
    {
      "epoch": 2.287581699346405,
      "grad_norm": 0.007141157519072294,
      "learning_rate": 8.568627450980392e-05,
      "loss": 0.001,
      "step": 1750
    },
    {
      "epoch": 2.3006535947712417,
      "grad_norm": 0.007537210825830698,
      "learning_rate": 8.503267973856209e-05,
      "loss": 0.0012,
      "step": 1760
    },
    {
      "epoch": 2.313725490196078,
      "grad_norm": 0.764282763004303,
      "learning_rate": 8.437908496732026e-05,
      "loss": 0.009,
      "step": 1770
    },
    {
      "epoch": 2.326797385620915,
      "grad_norm": 0.009738568216562271,
      "learning_rate": 8.372549019607843e-05,
      "loss": 0.0899,
      "step": 1780
    },
    {
      "epoch": 2.3398692810457518,
      "grad_norm": 0.05287899821996689,
      "learning_rate": 8.30718954248366e-05,
      "loss": 0.0027,
      "step": 1790
    },
    {
      "epoch": 2.3529411764705883,
      "grad_norm": 0.0178971029818058,
      "learning_rate": 8.241830065359478e-05,
      "loss": 0.0083,
      "step": 1800
    },
    {
      "epoch": 2.3529411764705883,
      "eval_accuracy": 0.9759656652360515,
      "eval_loss": 0.10839741677045822,
      "eval_runtime": 14.7686,
      "eval_samples_per_second": 78.884,
      "eval_steps_per_second": 9.886,
      "step": 1800
    },
    {
      "epoch": 2.366013071895425,
      "grad_norm": 0.01301596313714981,
      "learning_rate": 8.176470588235295e-05,
      "loss": 0.0016,
      "step": 1810
    },
    {
      "epoch": 2.3790849673202614,
      "grad_norm": 0.01324276439845562,
      "learning_rate": 8.111111111111112e-05,
      "loss": 0.0135,
      "step": 1820
    },
    {
      "epoch": 2.392156862745098,
      "grad_norm": 0.013326325453817844,
      "learning_rate": 8.045751633986929e-05,
      "loss": 0.0334,
      "step": 1830
    },
    {
      "epoch": 2.4052287581699345,
      "grad_norm": 0.011139987967908382,
      "learning_rate": 7.980392156862746e-05,
      "loss": 0.0014,
      "step": 1840
    },
    {
      "epoch": 2.418300653594771,
      "grad_norm": 0.0077317566610872746,
      "learning_rate": 7.915032679738562e-05,
      "loss": 0.0011,
      "step": 1850
    },
    {
      "epoch": 2.431372549019608,
      "grad_norm": 0.03215256333351135,
      "learning_rate": 7.849673202614379e-05,
      "loss": 0.0354,
      "step": 1860
    },
    {
      "epoch": 2.4444444444444446,
      "grad_norm": 0.009134942665696144,
      "learning_rate": 7.784313725490196e-05,
      "loss": 0.0018,
      "step": 1870
    },
    {
      "epoch": 2.457516339869281,
      "grad_norm": 0.008007184602320194,
      "learning_rate": 7.718954248366013e-05,
      "loss": 0.001,
      "step": 1880
    },
    {
      "epoch": 2.4705882352941178,
      "grad_norm": 0.015561921522021294,
      "learning_rate": 7.653594771241829e-05,
      "loss": 0.0009,
      "step": 1890
    },
    {
      "epoch": 2.4836601307189543,
      "grad_norm": 0.04056130349636078,
      "learning_rate": 7.588235294117648e-05,
      "loss": 0.0013,
      "step": 1900
    },
    {
      "epoch": 2.4836601307189543,
      "eval_accuracy": 0.9776824034334763,
      "eval_loss": 0.0902954488992691,
      "eval_runtime": 14.9792,
      "eval_samples_per_second": 77.774,
      "eval_steps_per_second": 9.747,
      "step": 1900
    },
    {
      "epoch": 2.496732026143791,
      "grad_norm": 0.00714436499401927,
      "learning_rate": 7.522875816993465e-05,
      "loss": 0.0434,
      "step": 1910
    },
    {
      "epoch": 2.5098039215686274,
      "grad_norm": 0.009256112389266491,
      "learning_rate": 7.457516339869282e-05,
      "loss": 0.001,
      "step": 1920
    },
    {
      "epoch": 2.522875816993464,
      "grad_norm": 0.016420746222138405,
      "learning_rate": 7.392156862745099e-05,
      "loss": 0.0282,
      "step": 1930
    },
    {
      "epoch": 2.5359477124183005,
      "grad_norm": 0.013019098900258541,
      "learning_rate": 7.326797385620915e-05,
      "loss": 0.0017,
      "step": 1940
    },
    {
      "epoch": 2.549019607843137,
      "grad_norm": 3.201817274093628,
      "learning_rate": 7.261437908496732e-05,
      "loss": 0.0274,
      "step": 1950
    },
    {
      "epoch": 2.5620915032679736,
      "grad_norm": 0.009792659431695938,
      "learning_rate": 7.196078431372549e-05,
      "loss": 0.0011,
      "step": 1960
    },
    {
      "epoch": 2.57516339869281,
      "grad_norm": 0.010890874080359936,
      "learning_rate": 7.130718954248366e-05,
      "loss": 0.0329,
      "step": 1970
    },
    {
      "epoch": 2.588235294117647,
      "grad_norm": 0.02582782320678234,
      "learning_rate": 7.065359477124184e-05,
      "loss": 0.0024,
      "step": 1980
    },
    {
      "epoch": 2.6013071895424837,
      "grad_norm": 4.812190055847168,
      "learning_rate": 7e-05,
      "loss": 0.0134,
      "step": 1990
    },
    {
      "epoch": 2.6143790849673203,
      "grad_norm": 0.05194231495261192,
      "learning_rate": 6.934640522875817e-05,
      "loss": 0.0298,
      "step": 2000
    },
    {
      "epoch": 2.6143790849673203,
      "eval_accuracy": 0.9733905579399141,
      "eval_loss": 0.10234156996011734,
      "eval_runtime": 15.3135,
      "eval_samples_per_second": 76.077,
      "eval_steps_per_second": 9.534,
      "step": 2000
    },
    {
      "epoch": 2.627450980392157,
      "grad_norm": 0.01054445281624794,
      "learning_rate": 6.869281045751634e-05,
      "loss": 0.0021,
      "step": 2010
    },
    {
      "epoch": 2.6405228758169934,
      "grad_norm": 0.010878126136958599,
      "learning_rate": 6.803921568627452e-05,
      "loss": 0.0015,
      "step": 2020
    },
    {
      "epoch": 2.65359477124183,
      "grad_norm": 0.032522134482860565,
      "learning_rate": 6.73856209150327e-05,
      "loss": 0.035,
      "step": 2030
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 0.02457094006240368,
      "learning_rate": 6.673202614379085e-05,
      "loss": 0.0032,
      "step": 2040
    },
    {
      "epoch": 2.6797385620915035,
      "grad_norm": 0.014886660501360893,
      "learning_rate": 6.607843137254902e-05,
      "loss": 0.003,
      "step": 2050
    },
    {
      "epoch": 2.69281045751634,
      "grad_norm": 0.020414328202605247,
      "learning_rate": 6.54248366013072e-05,
      "loss": 0.0017,
      "step": 2060
    },
    {
      "epoch": 2.7058823529411766,
      "grad_norm": 0.009341062046587467,
      "learning_rate": 6.477124183006537e-05,
      "loss": 0.0013,
      "step": 2070
    },
    {
      "epoch": 2.718954248366013,
      "grad_norm": 0.0061281765811145306,
      "learning_rate": 6.411764705882354e-05,
      "loss": 0.0009,
      "step": 2080
    },
    {
      "epoch": 2.7320261437908497,
      "grad_norm": 0.009815551340579987,
      "learning_rate": 6.34640522875817e-05,
      "loss": 0.0009,
      "step": 2090
    },
    {
      "epoch": 2.7450980392156863,
      "grad_norm": 0.0064156195148825645,
      "learning_rate": 6.281045751633987e-05,
      "loss": 0.0008,
      "step": 2100
    },
    {
      "epoch": 2.7450980392156863,
      "eval_accuracy": 0.976824034334764,
      "eval_loss": 0.110395647585392,
      "eval_runtime": 15.8536,
      "eval_samples_per_second": 73.485,
      "eval_steps_per_second": 9.209,
      "step": 2100
    },
    {
      "epoch": 2.758169934640523,
      "grad_norm": 0.009114363230764866,
      "learning_rate": 6.215686274509804e-05,
      "loss": 0.0008,
      "step": 2110
    },
    {
      "epoch": 2.7712418300653594,
      "grad_norm": 0.012108026072382927,
      "learning_rate": 6.150326797385621e-05,
      "loss": 0.0425,
      "step": 2120
    },
    {
      "epoch": 2.784313725490196,
      "grad_norm": 0.02400955744087696,
      "learning_rate": 6.0849673202614375e-05,
      "loss": 0.0011,
      "step": 2130
    },
    {
      "epoch": 2.7973856209150325,
      "grad_norm": 0.006131911184638739,
      "learning_rate": 6.0196078431372546e-05,
      "loss": 0.001,
      "step": 2140
    },
    {
      "epoch": 2.810457516339869,
      "grad_norm": 0.01375108677893877,
      "learning_rate": 5.9542483660130724e-05,
      "loss": 0.0014,
      "step": 2150
    },
    {
      "epoch": 2.8235294117647056,
      "grad_norm": 0.27725741267204285,
      "learning_rate": 5.8888888888888896e-05,
      "loss": 0.001,
      "step": 2160
    },
    {
      "epoch": 2.8366013071895426,
      "grad_norm": 0.005803197622299194,
      "learning_rate": 5.823529411764707e-05,
      "loss": 0.0012,
      "step": 2170
    },
    {
      "epoch": 2.849673202614379,
      "grad_norm": 0.005719684064388275,
      "learning_rate": 5.7581699346405225e-05,
      "loss": 0.0007,
      "step": 2180
    },
    {
      "epoch": 2.8627450980392157,
      "grad_norm": 0.049487028270959854,
      "learning_rate": 5.69281045751634e-05,
      "loss": 0.0013,
      "step": 2190
    },
    {
      "epoch": 2.8758169934640523,
      "grad_norm": 0.005333769600838423,
      "learning_rate": 5.627450980392157e-05,
      "loss": 0.0011,
      "step": 2200
    },
    {
      "epoch": 2.8758169934640523,
      "eval_accuracy": 0.9785407725321889,
      "eval_loss": 0.11278601735830307,
      "eval_runtime": 15.3235,
      "eval_samples_per_second": 76.027,
      "eval_steps_per_second": 9.528,
      "step": 2200
    },
    {
      "epoch": 2.888888888888889,
      "grad_norm": 0.013878793455660343,
      "learning_rate": 5.5620915032679746e-05,
      "loss": 0.0191,
      "step": 2210
    },
    {
      "epoch": 2.9019607843137254,
      "grad_norm": 0.005247410852462053,
      "learning_rate": 5.496732026143792e-05,
      "loss": 0.0008,
      "step": 2220
    },
    {
      "epoch": 2.915032679738562,
      "grad_norm": 0.010724911466240883,
      "learning_rate": 5.4313725490196076e-05,
      "loss": 0.0006,
      "step": 2230
    },
    {
      "epoch": 2.928104575163399,
      "grad_norm": 0.014705362729728222,
      "learning_rate": 5.366013071895425e-05,
      "loss": 0.0006,
      "step": 2240
    },
    {
      "epoch": 2.9411764705882355,
      "grad_norm": 0.007574844174087048,
      "learning_rate": 5.300653594771242e-05,
      "loss": 0.0007,
      "step": 2250
    },
    {
      "epoch": 2.954248366013072,
      "grad_norm": 0.004934164695441723,
      "learning_rate": 5.235294117647059e-05,
      "loss": 0.0015,
      "step": 2260
    },
    {
      "epoch": 2.9673202614379086,
      "grad_norm": 0.005061435047537088,
      "learning_rate": 5.169934640522877e-05,
      "loss": 0.0006,
      "step": 2270
    },
    {
      "epoch": 2.980392156862745,
      "grad_norm": 0.005057324655354023,
      "learning_rate": 5.104575163398693e-05,
      "loss": 0.0006,
      "step": 2280
    },
    {
      "epoch": 2.9934640522875817,
      "grad_norm": 0.004673287738114595,
      "learning_rate": 5.045751633986928e-05,
      "loss": 0.0127,
      "step": 2290
    },
    {
      "epoch": 3.0065359477124183,
      "grad_norm": 0.004593541845679283,
      "learning_rate": 4.980392156862745e-05,
      "loss": 0.0006,
      "step": 2300
    },
    {
      "epoch": 3.0065359477124183,
      "eval_accuracy": 0.9733905579399141,
      "eval_loss": 0.139481782913208,
      "eval_runtime": 14.0319,
      "eval_samples_per_second": 83.025,
      "eval_steps_per_second": 10.405,
      "step": 2300
    },
    {
      "epoch": 3.019607843137255,
      "grad_norm": 0.004796142224222422,
      "learning_rate": 4.915032679738562e-05,
      "loss": 0.0006,
      "step": 2310
    },
    {
      "epoch": 3.0326797385620914,
      "grad_norm": 0.005047705490142107,
      "learning_rate": 4.8496732026143794e-05,
      "loss": 0.0006,
      "step": 2320
    },
    {
      "epoch": 3.045751633986928,
      "grad_norm": 0.004708459600806236,
      "learning_rate": 4.7843137254901966e-05,
      "loss": 0.0005,
      "step": 2330
    },
    {
      "epoch": 3.0588235294117645,
      "grad_norm": 0.004462133627384901,
      "learning_rate": 4.718954248366013e-05,
      "loss": 0.0006,
      "step": 2340
    },
    {
      "epoch": 3.0718954248366015,
      "grad_norm": 0.004772055894136429,
      "learning_rate": 4.65359477124183e-05,
      "loss": 0.0009,
      "step": 2350
    },
    {
      "epoch": 3.084967320261438,
      "grad_norm": 0.004368507768958807,
      "learning_rate": 4.588235294117647e-05,
      "loss": 0.0005,
      "step": 2360
    },
    {
      "epoch": 3.0980392156862746,
      "grad_norm": 0.004448315594345331,
      "learning_rate": 4.5228758169934645e-05,
      "loss": 0.0005,
      "step": 2370
    },
    {
      "epoch": 3.111111111111111,
      "grad_norm": 0.004565922077745199,
      "learning_rate": 4.4575163398692816e-05,
      "loss": 0.0173,
      "step": 2380
    },
    {
      "epoch": 3.1241830065359477,
      "grad_norm": 0.004459939897060394,
      "learning_rate": 4.392156862745098e-05,
      "loss": 0.0069,
      "step": 2390
    },
    {
      "epoch": 3.1372549019607843,
      "grad_norm": 0.004509753547608852,
      "learning_rate": 4.326797385620915e-05,
      "loss": 0.0059,
      "step": 2400
    },
    {
      "epoch": 3.1372549019607843,
      "eval_accuracy": 0.9725321888412017,
      "eval_loss": 0.14194443821907043,
      "eval_runtime": 13.8686,
      "eval_samples_per_second": 84.002,
      "eval_steps_per_second": 10.527,
      "step": 2400
    },
    {
      "epoch": 3.150326797385621,
      "grad_norm": 0.004385701846331358,
      "learning_rate": 4.2614379084967324e-05,
      "loss": 0.0005,
      "step": 2410
    },
    {
      "epoch": 3.1633986928104574,
      "grad_norm": 0.5638505220413208,
      "learning_rate": 4.1960784313725496e-05,
      "loss": 0.0234,
      "step": 2420
    },
    {
      "epoch": 3.176470588235294,
      "grad_norm": 0.004813672509044409,
      "learning_rate": 4.130718954248366e-05,
      "loss": 0.0005,
      "step": 2430
    },
    {
      "epoch": 3.189542483660131,
      "grad_norm": 0.0054300627671182156,
      "learning_rate": 4.065359477124183e-05,
      "loss": 0.004,
      "step": 2440
    },
    {
      "epoch": 3.2026143790849675,
      "grad_norm": 0.004265918862074614,
      "learning_rate": 4e-05,
      "loss": 0.0005,
      "step": 2450
    },
    {
      "epoch": 3.215686274509804,
      "grad_norm": 0.005545614752918482,
      "learning_rate": 3.934640522875817e-05,
      "loss": 0.0005,
      "step": 2460
    },
    {
      "epoch": 3.2287581699346406,
      "grad_norm": 0.004381006117910147,
      "learning_rate": 3.8692810457516346e-05,
      "loss": 0.0005,
      "step": 2470
    },
    {
      "epoch": 3.241830065359477,
      "grad_norm": 0.004227633588016033,
      "learning_rate": 3.803921568627451e-05,
      "loss": 0.0005,
      "step": 2480
    },
    {
      "epoch": 3.2549019607843137,
      "grad_norm": 0.016775546595454216,
      "learning_rate": 3.738562091503268e-05,
      "loss": 0.0005,
      "step": 2490
    },
    {
      "epoch": 3.2679738562091503,
      "grad_norm": 0.024223702028393745,
      "learning_rate": 3.6732026143790854e-05,
      "loss": 0.0005,
      "step": 2500
    },
    {
      "epoch": 3.2679738562091503,
      "eval_accuracy": 0.9776824034334763,
      "eval_loss": 0.13354463875293732,
      "eval_runtime": 14.2248,
      "eval_samples_per_second": 81.899,
      "eval_steps_per_second": 10.264,
      "step": 2500
    },
    {
      "epoch": 3.281045751633987,
      "grad_norm": 0.0041252183727920055,
      "learning_rate": 3.607843137254902e-05,
      "loss": 0.0005,
      "step": 2510
    },
    {
      "epoch": 3.2941176470588234,
      "grad_norm": 0.004066206980496645,
      "learning_rate": 3.542483660130719e-05,
      "loss": 0.0008,
      "step": 2520
    },
    {
      "epoch": 3.30718954248366,
      "grad_norm": 0.004196736495941877,
      "learning_rate": 3.477124183006536e-05,
      "loss": 0.0036,
      "step": 2530
    },
    {
      "epoch": 3.3202614379084965,
      "grad_norm": 0.004051230847835541,
      "learning_rate": 3.411764705882353e-05,
      "loss": 0.0005,
      "step": 2540
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 0.004058164078742266,
      "learning_rate": 3.34640522875817e-05,
      "loss": 0.0175,
      "step": 2550
    },
    {
      "epoch": 3.34640522875817,
      "grad_norm": 0.0038333344273269176,
      "learning_rate": 3.281045751633987e-05,
      "loss": 0.0005,
      "step": 2560
    },
    {
      "epoch": 3.3594771241830066,
      "grad_norm": 0.004068716894835234,
      "learning_rate": 3.215686274509804e-05,
      "loss": 0.0005,
      "step": 2570
    },
    {
      "epoch": 3.372549019607843,
      "grad_norm": 0.004003152251243591,
      "learning_rate": 3.150326797385621e-05,
      "loss": 0.001,
      "step": 2580
    },
    {
      "epoch": 3.3856209150326797,
      "grad_norm": 0.0039999475702643394,
      "learning_rate": 3.0849673202614384e-05,
      "loss": 0.0005,
      "step": 2590
    },
    {
      "epoch": 3.3986928104575163,
      "grad_norm": 0.003976929467171431,
      "learning_rate": 3.019607843137255e-05,
      "loss": 0.0005,
      "step": 2600
    },
    {
      "epoch": 3.3986928104575163,
      "eval_accuracy": 0.976824034334764,
      "eval_loss": 0.12492600083351135,
      "eval_runtime": 13.7805,
      "eval_samples_per_second": 84.54,
      "eval_steps_per_second": 10.595,
      "step": 2600
    },
    {
      "epoch": 3.411764705882353,
      "grad_norm": 0.0038215164095163345,
      "learning_rate": 2.954248366013072e-05,
      "loss": 0.0004,
      "step": 2610
    },
    {
      "epoch": 3.4248366013071894,
      "grad_norm": 0.0037740464322268963,
      "learning_rate": 2.8888888888888888e-05,
      "loss": 0.0004,
      "step": 2620
    },
    {
      "epoch": 3.4379084967320264,
      "grad_norm": 0.003947110380977392,
      "learning_rate": 2.823529411764706e-05,
      "loss": 0.0005,
      "step": 2630
    },
    {
      "epoch": 3.450980392156863,
      "grad_norm": 0.003939367830753326,
      "learning_rate": 2.758169934640523e-05,
      "loss": 0.0004,
      "step": 2640
    },
    {
      "epoch": 3.4640522875816995,
      "grad_norm": 0.0037511338014155626,
      "learning_rate": 2.69281045751634e-05,
      "loss": 0.0004,
      "step": 2650
    },
    {
      "epoch": 3.477124183006536,
      "grad_norm": 0.004019754473119974,
      "learning_rate": 2.627450980392157e-05,
      "loss": 0.0367,
      "step": 2660
    },
    {
      "epoch": 3.4901960784313726,
      "grad_norm": 0.009803508408367634,
      "learning_rate": 2.562091503267974e-05,
      "loss": 0.0005,
      "step": 2670
    },
    {
      "epoch": 3.503267973856209,
      "grad_norm": 0.013077986426651478,
      "learning_rate": 2.496732026143791e-05,
      "loss": 0.0007,
      "step": 2680
    },
    {
      "epoch": 3.5163398692810457,
      "grad_norm": 0.007836658507585526,
      "learning_rate": 2.431372549019608e-05,
      "loss": 0.0008,
      "step": 2690
    },
    {
      "epoch": 3.5294117647058822,
      "grad_norm": 0.004085828084498644,
      "learning_rate": 2.366013071895425e-05,
      "loss": 0.0007,
      "step": 2700
    },
    {
      "epoch": 3.5294117647058822,
      "eval_accuracy": 0.9776824034334763,
      "eval_loss": 0.1157020702958107,
      "eval_runtime": 14.8164,
      "eval_samples_per_second": 78.629,
      "eval_steps_per_second": 9.854,
      "step": 2700
    },
    {
      "epoch": 3.542483660130719,
      "grad_norm": 0.004049411043524742,
      "learning_rate": 2.3006535947712418e-05,
      "loss": 0.0005,
      "step": 2710
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 0.015368789434432983,
      "learning_rate": 2.235294117647059e-05,
      "loss": 0.0006,
      "step": 2720
    },
    {
      "epoch": 3.568627450980392,
      "grad_norm": 0.003844345221295953,
      "learning_rate": 2.169934640522876e-05,
      "loss": 0.0008,
      "step": 2730
    },
    {
      "epoch": 3.581699346405229,
      "grad_norm": 0.004043512977659702,
      "learning_rate": 2.104575163398693e-05,
      "loss": 0.0005,
      "step": 2740
    },
    {
      "epoch": 3.5947712418300655,
      "grad_norm": 0.003938615787774324,
      "learning_rate": 2.0392156862745097e-05,
      "loss": 0.0005,
      "step": 2750
    },
    {
      "epoch": 3.607843137254902,
      "grad_norm": 0.008323022164404392,
      "learning_rate": 1.973856209150327e-05,
      "loss": 0.0005,
      "step": 2760
    },
    {
      "epoch": 3.6209150326797386,
      "grad_norm": 0.004000243730843067,
      "learning_rate": 1.9084967320261437e-05,
      "loss": 0.0005,
      "step": 2770
    },
    {
      "epoch": 3.633986928104575,
      "grad_norm": 0.004011669661849737,
      "learning_rate": 1.843137254901961e-05,
      "loss": 0.0005,
      "step": 2780
    },
    {
      "epoch": 3.6470588235294117,
      "grad_norm": 0.01243202667683363,
      "learning_rate": 1.777777777777778e-05,
      "loss": 0.0139,
      "step": 2790
    },
    {
      "epoch": 3.6601307189542482,
      "grad_norm": 0.0176347978413105,
      "learning_rate": 1.7124183006535948e-05,
      "loss": 0.0005,
      "step": 2800
    },
    {
      "epoch": 3.6601307189542482,
      "eval_accuracy": 0.9785407725321889,
      "eval_loss": 0.12019691616296768,
      "eval_runtime": 14.517,
      "eval_samples_per_second": 80.251,
      "eval_steps_per_second": 10.057,
      "step": 2800
    },
    {
      "epoch": 3.6732026143790852,
      "grad_norm": 0.0037740224506706,
      "learning_rate": 1.647058823529412e-05,
      "loss": 0.0005,
      "step": 2810
    },
    {
      "epoch": 3.686274509803922,
      "grad_norm": 0.004195007495582104,
      "learning_rate": 1.5816993464052288e-05,
      "loss": 0.0008,
      "step": 2820
    },
    {
      "epoch": 3.6993464052287583,
      "grad_norm": 0.0038069712463766336,
      "learning_rate": 1.5163398692810458e-05,
      "loss": 0.0007,
      "step": 2830
    },
    {
      "epoch": 3.712418300653595,
      "grad_norm": 0.010378316044807434,
      "learning_rate": 1.4509803921568629e-05,
      "loss": 0.0006,
      "step": 2840
    },
    {
      "epoch": 3.7254901960784315,
      "grad_norm": 0.003901832504197955,
      "learning_rate": 1.3856209150326799e-05,
      "loss": 0.0005,
      "step": 2850
    },
    {
      "epoch": 3.738562091503268,
      "grad_norm": 0.003902917029336095,
      "learning_rate": 1.3202614379084969e-05,
      "loss": 0.0005,
      "step": 2860
    },
    {
      "epoch": 3.7516339869281046,
      "grad_norm": 0.004125463310629129,
      "learning_rate": 1.2549019607843138e-05,
      "loss": 0.001,
      "step": 2870
    },
    {
      "epoch": 3.764705882352941,
      "grad_norm": 0.003765885718166828,
      "learning_rate": 1.1895424836601307e-05,
      "loss": 0.0006,
      "step": 2880
    },
    {
      "epoch": 3.7777777777777777,
      "grad_norm": 0.0036050116177648306,
      "learning_rate": 1.1241830065359478e-05,
      "loss": 0.001,
      "step": 2890
    },
    {
      "epoch": 3.7908496732026142,
      "grad_norm": 0.003705250099301338,
      "learning_rate": 1.0588235294117648e-05,
      "loss": 0.001,
      "step": 2900
    },
    {
      "epoch": 3.7908496732026142,
      "eval_accuracy": 0.9776824034334763,
      "eval_loss": 0.12394391000270844,
      "eval_runtime": 14.3108,
      "eval_samples_per_second": 81.407,
      "eval_steps_per_second": 10.202,
      "step": 2900
    },
    {
      "epoch": 3.803921568627451,
      "grad_norm": 0.01087539829313755,
      "learning_rate": 9.934640522875818e-06,
      "loss": 0.0005,
      "step": 2910
    },
    {
      "epoch": 3.8169934640522873,
      "grad_norm": 0.004286112263798714,
      "learning_rate": 9.281045751633987e-06,
      "loss": 0.0004,
      "step": 2920
    },
    {
      "epoch": 3.8300653594771243,
      "grad_norm": 0.005555053241550922,
      "learning_rate": 8.627450980392157e-06,
      "loss": 0.0006,
      "step": 2930
    },
    {
      "epoch": 3.843137254901961,
      "grad_norm": 0.05337900295853615,
      "learning_rate": 7.973856209150327e-06,
      "loss": 0.0007,
      "step": 2940
    },
    {
      "epoch": 3.8562091503267975,
      "grad_norm": 0.0040720016695559025,
      "learning_rate": 7.320261437908498e-06,
      "loss": 0.0004,
      "step": 2950
    },
    {
      "epoch": 3.869281045751634,
      "grad_norm": 0.0036330316215753555,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.0006,
      "step": 2960
    },
    {
      "epoch": 3.8823529411764706,
      "grad_norm": 0.00944702047854662,
      "learning_rate": 6.013071895424837e-06,
      "loss": 0.0009,
      "step": 2970
    },
    {
      "epoch": 3.895424836601307,
      "grad_norm": 0.0034994047600775957,
      "learning_rate": 5.359477124183006e-06,
      "loss": 0.0006,
      "step": 2980
    },
    {
      "epoch": 3.9084967320261437,
      "grad_norm": 0.030153293162584305,
      "learning_rate": 4.705882352941177e-06,
      "loss": 0.0007,
      "step": 2990
    },
    {
      "epoch": 3.9215686274509802,
      "grad_norm": 0.0036808131262660027,
      "learning_rate": 4.052287581699347e-06,
      "loss": 0.0004,
      "step": 3000
    },
    {
      "epoch": 3.9215686274509802,
      "eval_accuracy": 0.976824034334764,
      "eval_loss": 0.12309978902339935,
      "eval_runtime": 13.6907,
      "eval_samples_per_second": 85.094,
      "eval_steps_per_second": 10.664,
      "step": 3000
    },
    {
      "epoch": 3.9346405228758172,
      "grad_norm": 0.0034610480070114136,
      "learning_rate": 3.398692810457516e-06,
      "loss": 0.0005,
      "step": 3010
    },
    {
      "epoch": 3.947712418300654,
      "grad_norm": 0.0036175919231027365,
      "learning_rate": 2.7450980392156863e-06,
      "loss": 0.0005,
      "step": 3020
    },
    {
      "epoch": 3.9607843137254903,
      "grad_norm": 0.003583298297598958,
      "learning_rate": 2.091503267973856e-06,
      "loss": 0.0005,
      "step": 3030
    },
    {
      "epoch": 3.973856209150327,
      "grad_norm": 0.003492480143904686,
      "learning_rate": 1.4379084967320261e-06,
      "loss": 0.0005,
      "step": 3040
    },
    {
      "epoch": 3.9869281045751634,
      "grad_norm": 0.003486211644485593,
      "learning_rate": 7.843137254901962e-07,
      "loss": 0.0004,
      "step": 3050
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.0035035167820751667,
      "learning_rate": 1.30718954248366e-07,
      "loss": 0.0006,
      "step": 3060
    },
    {
      "epoch": 4.0,
      "step": 3060,
      "total_flos": 3.7909081319458406e+18,
      "train_loss": 0.047429592626662374,
      "train_runtime": 1590.6048,
      "train_samples_per_second": 30.756,
      "train_steps_per_second": 1.924
    }
  ],
  "logging_steps": 10,
  "max_steps": 3060,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 3.7909081319458406e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}