{
  "best_global_step": null,
  "best_metric": 0.3671196699142456,
  "best_model_checkpoint": null,
  "epoch": 4.951838553918535,
  "eval_steps": 500,
  "global_step": 20000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002476166893648632,
      "grad_norm": 0.48706236481666565,
      "learning_rate": 0.0,
      "loss": 3.0743,
      "step": 1
    },
    {
      "epoch": 0.02476166893648632,
      "grad_norm": 0.6543678045272827,
      "learning_rate": 9.900000000000002e-06,
      "loss": 2.8614,
      "step": 100
    },
    {
      "epoch": 0.04952333787297264,
      "grad_norm": 0.5323843359947205,
      "learning_rate": 1.9900000000000003e-05,
      "loss": 1.8779,
      "step": 200
    },
    {
      "epoch": 0.07428500680945896,
      "grad_norm": 0.9197348952293396,
      "learning_rate": 2.9900000000000002e-05,
      "loss": 1.2173,
      "step": 300
    },
    {
      "epoch": 0.09904667574594528,
      "grad_norm": 1.113196849822998,
      "learning_rate": 3.99e-05,
      "loss": 0.9351,
      "step": 400
    },
    {
      "epoch": 0.1238083446824316,
      "grad_norm": 1.1514325141906738,
      "learning_rate": 4.99e-05,
      "loss": 0.8237,
      "step": 500
    },
    {
      "epoch": 0.1238083446824316,
      "eval_loss": 0.6713243126869202,
      "eval_runtime": 303.9033,
      "eval_samples_per_second": 47.249,
      "eval_steps_per_second": 2.955,
      "step": 500
    },
    {
      "epoch": 0.14857001361891792,
      "grad_norm": 1.1924649477005005,
      "learning_rate": 4.974866717440975e-05,
      "loss": 0.7509,
      "step": 600
    },
    {
      "epoch": 0.17333168255540424,
      "grad_norm": 1.0876811742782593,
      "learning_rate": 4.9494795633409495e-05,
      "loss": 0.6963,
      "step": 700
    },
    {
      "epoch": 0.19809335149189056,
      "grad_norm": 1.4002293348312378,
      "learning_rate": 4.924092409240924e-05,
      "loss": 0.6837,
      "step": 800
    },
    {
      "epoch": 0.22285502042837688,
      "grad_norm": 1.2242807149887085,
      "learning_rate": 4.898705255140899e-05,
      "loss": 0.6686,
      "step": 900
    },
    {
      "epoch": 0.2476166893648632,
      "grad_norm": 1.3704895973205566,
      "learning_rate": 4.873318101040874e-05,
      "loss": 0.6462,
      "step": 1000
    },
    {
      "epoch": 0.2476166893648632,
      "eval_loss": 0.5437047481536865,
      "eval_runtime": 309.4401,
      "eval_samples_per_second": 46.403,
      "eval_steps_per_second": 2.902,
      "step": 1000
    },
    {
      "epoch": 0.2723783583013495,
      "grad_norm": 1.2968043088912964,
      "learning_rate": 4.847930946940848e-05,
      "loss": 0.6301,
      "step": 1100
    },
    {
      "epoch": 0.29714002723783584,
      "grad_norm": 1.4131654500961304,
      "learning_rate": 4.822543792840823e-05,
      "loss": 0.6256,
      "step": 1200
    },
    {
      "epoch": 0.3219016961743221,
      "grad_norm": 1.3621524572372437,
      "learning_rate": 4.7971566387407974e-05,
      "loss": 0.6179,
      "step": 1300
    },
    {
      "epoch": 0.3466633651108085,
      "grad_norm": 1.4688136577606201,
      "learning_rate": 4.771769484640772e-05,
      "loss": 0.5963,
      "step": 1400
    },
    {
      "epoch": 0.37142503404729477,
      "grad_norm": 1.3867251873016357,
      "learning_rate": 4.746382330540747e-05,
      "loss": 0.5867,
      "step": 1500
    },
    {
      "epoch": 0.37142503404729477,
      "eval_loss": 0.5053889155387878,
      "eval_runtime": 301.1479,
      "eval_samples_per_second": 47.681,
      "eval_steps_per_second": 2.982,
      "step": 1500
    },
    {
      "epoch": 0.3961867029837811,
      "grad_norm": 1.018188714981079,
      "learning_rate": 4.720995176440721e-05,
      "loss": 0.5854,
      "step": 1600
    },
    {
      "epoch": 0.4209483719202674,
      "grad_norm": 1.3094606399536133,
      "learning_rate": 4.6956080223406954e-05,
      "loss": 0.567,
      "step": 1700
    },
    {
      "epoch": 0.44571004085675375,
      "grad_norm": 1.3288716077804565,
      "learning_rate": 4.6702208682406704e-05,
      "loss": 0.5629,
      "step": 1800
    },
    {
      "epoch": 0.47047170979324004,
      "grad_norm": 1.2919931411743164,
      "learning_rate": 4.6448337141406453e-05,
      "loss": 0.5568,
      "step": 1900
    },
    {
      "epoch": 0.4952333787297264,
      "grad_norm": 1.161441445350647,
      "learning_rate": 4.6194465600406197e-05,
      "loss": 0.5574,
      "step": 2000
    },
    {
      "epoch": 0.4952333787297264,
      "eval_loss": 0.4834918975830078,
      "eval_runtime": 308.7731,
      "eval_samples_per_second": 46.503,
      "eval_steps_per_second": 2.908,
      "step": 2000
    },
    {
      "epoch": 0.5199950476662127,
      "grad_norm": 1.0919381380081177,
      "learning_rate": 4.5940594059405946e-05,
      "loss": 0.5525,
      "step": 2100
    },
    {
      "epoch": 0.544756716602699,
      "grad_norm": 1.302264928817749,
      "learning_rate": 4.568672251840569e-05,
      "loss": 0.5486,
      "step": 2200
    },
    {
      "epoch": 0.5695183855391853,
      "grad_norm": 1.3316988945007324,
      "learning_rate": 4.543285097740543e-05,
      "loss": 0.5435,
      "step": 2300
    },
    {
      "epoch": 0.5942800544756717,
      "grad_norm": 1.16013765335083,
      "learning_rate": 4.517897943640518e-05,
      "loss": 0.537,
      "step": 2400
    },
    {
      "epoch": 0.619041723412158,
      "grad_norm": 1.0240144729614258,
      "learning_rate": 4.4925107895404926e-05,
      "loss": 0.5351,
      "step": 2500
    },
    {
      "epoch": 0.619041723412158,
      "eval_loss": 0.46248629689216614,
      "eval_runtime": 298.9584,
      "eval_samples_per_second": 48.03,
      "eval_steps_per_second": 3.004,
      "step": 2500
    },
    {
      "epoch": 0.6438033923486443,
      "grad_norm": 1.403138279914856,
      "learning_rate": 4.467123635440467e-05,
      "loss": 0.529,
      "step": 2600
    },
    {
      "epoch": 0.6685650612851306,
      "grad_norm": 1.0086416006088257,
      "learning_rate": 4.441736481340442e-05,
      "loss": 0.5179,
      "step": 2700
    },
    {
      "epoch": 0.693326730221617,
      "grad_norm": 1.0099183320999146,
      "learning_rate": 4.416349327240417e-05,
      "loss": 0.5159,
      "step": 2800
    },
    {
      "epoch": 0.7180883991581033,
      "grad_norm": 1.066780686378479,
      "learning_rate": 4.390962173140391e-05,
      "loss": 0.5184,
      "step": 2900
    },
    {
      "epoch": 0.7428500680945895,
      "grad_norm": 1.1880680322647095,
      "learning_rate": 4.365575019040366e-05,
      "loss": 0.5132,
      "step": 3000
    },
    {
      "epoch": 0.7428500680945895,
      "eval_loss": 0.4524896740913391,
      "eval_runtime": 309.4036,
      "eval_samples_per_second": 46.409,
      "eval_steps_per_second": 2.902,
      "step": 3000
    },
    {
      "epoch": 0.7676117370310759,
      "grad_norm": 1.128923773765564,
      "learning_rate": 4.3401878649403405e-05,
      "loss": 0.5132,
      "step": 3100
    },
    {
      "epoch": 0.7923734059675622,
      "grad_norm": 1.0441526174545288,
      "learning_rate": 4.314800710840315e-05,
      "loss": 0.5024,
      "step": 3200
    },
    {
      "epoch": 0.8171350749040486,
      "grad_norm": 1.3435834646224976,
      "learning_rate": 4.28941355674029e-05,
      "loss": 0.4999,
      "step": 3300
    },
    {
      "epoch": 0.8418967438405348,
      "grad_norm": 1.0047725439071655,
      "learning_rate": 4.264026402640264e-05,
      "loss": 0.5025,
      "step": 3400
    },
    {
      "epoch": 0.8666584127770212,
      "grad_norm": 1.1918917894363403,
      "learning_rate": 4.238639248540239e-05,
      "loss": 0.5002,
      "step": 3500
    },
    {
      "epoch": 0.8666584127770212,
      "eval_loss": 0.44193407893180847,
      "eval_runtime": 307.4102,
      "eval_samples_per_second": 46.71,
      "eval_steps_per_second": 2.921,
      "step": 3500
    },
    {
      "epoch": 0.8914200817135075,
      "grad_norm": 1.3518158197402954,
      "learning_rate": 4.2132520944402134e-05,
      "loss": 0.4972,
      "step": 3600
    },
    {
      "epoch": 0.9161817506499939,
      "grad_norm": 1.2649939060211182,
      "learning_rate": 4.1878649403401884e-05,
      "loss": 0.4935,
      "step": 3700
    },
    {
      "epoch": 0.9409434195864801,
      "grad_norm": 1.1180412769317627,
      "learning_rate": 4.162477786240163e-05,
      "loss": 0.5021,
      "step": 3800
    },
    {
      "epoch": 0.9657050885229664,
      "grad_norm": 0.9952785968780518,
      "learning_rate": 4.137090632140137e-05,
      "loss": 0.4995,
      "step": 3900
    },
    {
      "epoch": 0.9904667574594528,
      "grad_norm": 1.1129059791564941,
      "learning_rate": 4.111703478040112e-05,
      "loss": 0.4944,
      "step": 4000
    },
    {
      "epoch": 0.9904667574594528,
      "eval_loss": 0.4326762557029724,
      "eval_runtime": 305.7402,
      "eval_samples_per_second": 46.965,
      "eval_steps_per_second": 2.937,
      "step": 4000
    },
    {
      "epoch": 1.0151046180512566,
      "grad_norm": 1.2399227619171143,
      "learning_rate": 4.086316323940086e-05,
      "loss": 0.4828,
      "step": 4100
    },
    {
      "epoch": 1.0398662869877429,
      "grad_norm": 1.2356698513031006,
      "learning_rate": 4.0609291698400606e-05,
      "loss": 0.4818,
      "step": 4200
    },
    {
      "epoch": 1.0646279559242293,
      "grad_norm": 1.016559362411499,
      "learning_rate": 4.0355420157400356e-05,
      "loss": 0.4777,
      "step": 4300
    },
    {
      "epoch": 1.0893896248607156,
      "grad_norm": 1.1003488302230835,
      "learning_rate": 4.0101548616400106e-05,
      "loss": 0.4818,
      "step": 4400
    },
    {
      "epoch": 1.114151293797202,
      "grad_norm": 1.0257238149642944,
      "learning_rate": 3.984767707539985e-05,
      "loss": 0.4804,
      "step": 4500
    },
    {
      "epoch": 1.114151293797202,
      "eval_loss": 0.4317300021648407,
      "eval_runtime": 301.8975,
      "eval_samples_per_second": 47.563,
      "eval_steps_per_second": 2.975,
      "step": 4500
    },
    {
      "epoch": 1.1389129627336882,
      "grad_norm": 0.9789583683013916,
      "learning_rate": 3.95938055343996e-05,
      "loss": 0.4765,
      "step": 4600
    },
    {
      "epoch": 1.1636746316701745,
      "grad_norm": 1.2558050155639648,
      "learning_rate": 3.933993399339934e-05,
      "loss": 0.481,
      "step": 4700
    },
    {
      "epoch": 1.188436300606661,
      "grad_norm": 1.0205676555633545,
      "learning_rate": 3.9086062452399085e-05,
      "loss": 0.4826,
      "step": 4800
    },
    {
      "epoch": 1.2131979695431472,
      "grad_norm": 1.0989702939987183,
      "learning_rate": 3.8832190911398835e-05,
      "loss": 0.4716,
      "step": 4900
    },
    {
      "epoch": 1.2379596384796336,
      "grad_norm": 1.031387448310852,
      "learning_rate": 3.857831937039858e-05,
      "loss": 0.4704,
      "step": 5000
    },
    {
      "epoch": 1.2379596384796336,
      "eval_loss": 0.4228954017162323,
      "eval_runtime": 306.4121,
      "eval_samples_per_second": 46.862,
      "eval_steps_per_second": 2.931,
      "step": 5000
    },
    {
      "epoch": 1.2627213074161199,
      "grad_norm": 1.123007893562317,
      "learning_rate": 3.832444782939832e-05,
      "loss": 0.4688,
      "step": 5100
    },
    {
      "epoch": 1.2874829763526061,
      "grad_norm": 1.0141407251358032,
      "learning_rate": 3.807057628839807e-05,
      "loss": 0.4678,
      "step": 5200
    },
    {
      "epoch": 1.3122446452890926,
      "grad_norm": 1.0771005153656006,
      "learning_rate": 3.781670474739782e-05,
      "loss": 0.4659,
      "step": 5300
    },
    {
      "epoch": 1.3370063142255788,
      "grad_norm": 1.2410191297531128,
      "learning_rate": 3.7562833206397564e-05,
      "loss": 0.4725,
      "step": 5400
    },
    {
      "epoch": 1.361767983162065,
      "grad_norm": 0.9323300123214722,
      "learning_rate": 3.7308961665397314e-05,
      "loss": 0.4662,
      "step": 5500
    },
    {
      "epoch": 1.361767983162065,
      "eval_loss": 0.418260782957077,
      "eval_runtime": 305.893,
      "eval_samples_per_second": 46.941,
      "eval_steps_per_second": 2.936,
      "step": 5500
    },
    {
      "epoch": 1.3865296520985515,
      "grad_norm": 0.9075552225112915,
      "learning_rate": 3.705509012439706e-05,
      "loss": 0.469,
      "step": 5600
    },
    {
      "epoch": 1.4112913210350377,
      "grad_norm": 1.095397710800171,
      "learning_rate": 3.68012185833968e-05,
      "loss": 0.4696,
      "step": 5700
    },
    {
      "epoch": 1.436052989971524,
      "grad_norm": 1.2005417346954346,
      "learning_rate": 3.654734704239655e-05,
      "loss": 0.4553,
      "step": 5800
    },
    {
      "epoch": 1.4608146589080104,
      "grad_norm": 1.0341594219207764,
      "learning_rate": 3.6293475501396294e-05,
      "loss": 0.4632,
      "step": 5900
    },
    {
      "epoch": 1.4855763278444967,
      "grad_norm": 1.0710721015930176,
      "learning_rate": 3.603960396039604e-05,
      "loss": 0.4676,
      "step": 6000
    },
    {
      "epoch": 1.4855763278444967,
      "eval_loss": 0.41258370876312256,
      "eval_runtime": 333.4873,
      "eval_samples_per_second": 43.057,
      "eval_steps_per_second": 2.693,
      "step": 6000
    },
    {
      "epoch": 1.510337996780983,
      "grad_norm": 1.0697613954544067,
      "learning_rate": 3.5785732419395787e-05,
      "loss": 0.4653,
      "step": 6100
    },
    {
      "epoch": 1.5350996657174694,
      "grad_norm": 0.9342092275619507,
      "learning_rate": 3.5531860878395536e-05,
      "loss": 0.4584,
      "step": 6200
    },
    {
      "epoch": 1.5598613346539558,
      "grad_norm": 1.18372642993927,
      "learning_rate": 3.527798933739528e-05,
      "loss": 0.4646,
      "step": 6300
    },
    {
      "epoch": 1.5846230035904418,
      "grad_norm": 0.9926603436470032,
      "learning_rate": 3.502411779639503e-05,
      "loss": 0.4523,
      "step": 6400
    },
    {
      "epoch": 1.6093846725269283,
      "grad_norm": 0.8520728945732117,
      "learning_rate": 3.477024625539477e-05,
      "loss": 0.4508,
      "step": 6500
    },
    {
      "epoch": 1.6093846725269283,
      "eval_loss": 0.4122717082500458,
      "eval_runtime": 337.4066,
      "eval_samples_per_second": 42.557,
      "eval_steps_per_second": 2.661,
      "step": 6500
    },
    {
      "epoch": 1.6341463414634148,
      "grad_norm": 1.0143983364105225,
      "learning_rate": 3.4516374714394516e-05,
      "loss": 0.4612,
      "step": 6600
    },
    {
      "epoch": 1.658908010399901,
      "grad_norm": 1.0106840133666992,
      "learning_rate": 3.426250317339426e-05,
      "loss": 0.4684,
      "step": 6700
    },
    {
      "epoch": 1.6836696793363872,
      "grad_norm": 1.0055443048477173,
      "learning_rate": 3.400863163239401e-05,
      "loss": 0.4568,
      "step": 6800
    },
    {
      "epoch": 1.7084313482728737,
      "grad_norm": 0.9551578760147095,
      "learning_rate": 3.375476009139376e-05,
      "loss": 0.4604,
      "step": 6900
    },
    {
      "epoch": 1.73319301720936,
      "grad_norm": 0.9932017922401428,
      "learning_rate": 3.35008885503935e-05,
      "loss": 0.4472,
      "step": 7000
    },
    {
      "epoch": 1.73319301720936,
      "eval_loss": 0.40953782200813293,
      "eval_runtime": 334.3565,
      "eval_samples_per_second": 42.945,
      "eval_steps_per_second": 2.686,
      "step": 7000
    },
    {
      "epoch": 1.7579546861458462,
      "grad_norm": 1.0452265739440918,
      "learning_rate": 3.324701700939325e-05,
      "loss": 0.4604,
      "step": 7100
    },
    {
      "epoch": 1.7827163550823326,
      "grad_norm": 1.0083004236221313,
      "learning_rate": 3.2993145468392995e-05,
      "loss": 0.4507,
      "step": 7200
    },
    {
      "epoch": 1.8074780240188189,
      "grad_norm": 0.9387707710266113,
      "learning_rate": 3.273927392739274e-05,
      "loss": 0.4516,
      "step": 7300
    },
    {
      "epoch": 1.832239692955305,
      "grad_norm": 0.9475572109222412,
      "learning_rate": 3.248540238639249e-05,
      "loss": 0.4444,
      "step": 7400
    },
    {
      "epoch": 1.8570013618917915,
      "grad_norm": 1.093817949295044,
      "learning_rate": 3.223153084539223e-05,
      "loss": 0.456,
      "step": 7500
    },
    {
      "epoch": 1.8570013618917915,
      "eval_loss": 0.4029190242290497,
      "eval_runtime": 292.0631,
      "eval_samples_per_second": 49.164,
      "eval_steps_per_second": 3.075,
      "step": 7500
    },
    {
      "epoch": 1.8817630308282778,
      "grad_norm": 0.9310826063156128,
      "learning_rate": 3.1977659304391974e-05,
      "loss": 0.4552,
      "step": 7600
    },
    {
      "epoch": 1.906524699764764,
      "grad_norm": 1.0708472728729248,
      "learning_rate": 3.1723787763391724e-05,
      "loss": 0.4442,
      "step": 7700
    },
    {
      "epoch": 1.9312863687012505,
      "grad_norm": 0.9762539267539978,
      "learning_rate": 3.1469916222391474e-05,
      "loss": 0.438,
      "step": 7800
    },
    {
      "epoch": 1.956048037637737,
      "grad_norm": 1.1194584369659424,
      "learning_rate": 3.121604468139122e-05,
      "loss": 0.4475,
      "step": 7900
    },
    {
      "epoch": 1.980809706574223,
      "grad_norm": 1.001720905303955,
      "learning_rate": 3.096217314039097e-05,
      "loss": 0.4355,
      "step": 8000
    },
    {
      "epoch": 1.980809706574223,
      "eval_loss": 0.40090999007225037,
      "eval_runtime": 290.1583,
      "eval_samples_per_second": 49.487,
      "eval_steps_per_second": 3.095,
      "step": 8000
    },
    {
      "epoch": 2.005447567166027,
      "grad_norm": 1.2274198532104492,
      "learning_rate": 3.070830159939071e-05,
      "loss": 0.4395,
      "step": 8100
    },
    {
      "epoch": 2.0302092361025132,
      "grad_norm": 1.008782148361206,
      "learning_rate": 3.0454430058390453e-05,
      "loss": 0.4386,
      "step": 8200
    },
    {
      "epoch": 2.0549709050389997,
      "grad_norm": 1.110425591468811,
      "learning_rate": 3.0200558517390203e-05,
      "loss": 0.4427,
      "step": 8300
    },
    {
      "epoch": 2.0797325739754857,
      "grad_norm": 0.944615364074707,
      "learning_rate": 2.994668697638995e-05,
      "loss": 0.4367,
      "step": 8400
    },
    {
      "epoch": 2.104494242911972,
      "grad_norm": 1.0133123397827148,
      "learning_rate": 2.9692815435389693e-05,
      "loss": 0.4426,
      "step": 8500
    },
    {
      "epoch": 2.104494242911972,
      "eval_loss": 0.3965131342411041,
      "eval_runtime": 295.1479,
      "eval_samples_per_second": 48.65,
      "eval_steps_per_second": 3.043,
      "step": 8500
    },
    {
      "epoch": 2.1292559118484586,
      "grad_norm": 0.7603266835212708,
      "learning_rate": 2.9438943894389443e-05,
      "loss": 0.4388,
      "step": 8600
    },
    {
      "epoch": 2.154017580784945,
      "grad_norm": 0.928263783454895,
      "learning_rate": 2.9185072353389186e-05,
      "loss": 0.4386,
      "step": 8700
    },
    {
      "epoch": 2.178779249721431,
      "grad_norm": 1.0597220659255981,
      "learning_rate": 2.8931200812388932e-05,
      "loss": 0.4281,
      "step": 8800
    },
    {
      "epoch": 2.2035409186579176,
      "grad_norm": 1.0807472467422485,
      "learning_rate": 2.8677329271388682e-05,
      "loss": 0.4312,
      "step": 8900
    },
    {
      "epoch": 2.228302587594404,
      "grad_norm": 1.0931870937347412,
      "learning_rate": 2.8423457730388425e-05,
      "loss": 0.4369,
      "step": 9000
    },
    {
      "epoch": 2.228302587594404,
      "eval_loss": 0.39540666341781616,
      "eval_runtime": 292.4821,
      "eval_samples_per_second": 49.094,
      "eval_steps_per_second": 3.07,
      "step": 9000
    },
    {
      "epoch": 2.25306425653089,
      "grad_norm": 0.9188655614852905,
      "learning_rate": 2.816958618938817e-05,
      "loss": 0.4404,
      "step": 9100
    },
    {
      "epoch": 2.2778259254673765,
      "grad_norm": 1.0746827125549316,
      "learning_rate": 2.7915714648387918e-05,
      "loss": 0.4395,
      "step": 9200
    },
    {
      "epoch": 2.302587594403863,
      "grad_norm": 0.9879487156867981,
      "learning_rate": 2.7661843107387665e-05,
      "loss": 0.4209,
      "step": 9300
    },
    {
      "epoch": 2.327349263340349,
      "grad_norm": 1.0951037406921387,
      "learning_rate": 2.7407971566387408e-05,
      "loss": 0.431,
      "step": 9400
    },
    {
      "epoch": 2.3521109322768354,
      "grad_norm": 1.1300798654556274,
      "learning_rate": 2.7154100025387158e-05,
      "loss": 0.4286,
      "step": 9500
    },
    {
      "epoch": 2.3521109322768354,
      "eval_loss": 0.3942405581474304,
      "eval_runtime": 294.7824,
      "eval_samples_per_second": 48.711,
      "eval_steps_per_second": 3.046,
      "step": 9500
    },
    {
      "epoch": 2.376872601213322,
      "grad_norm": 1.2054657936096191,
      "learning_rate": 2.69002284843869e-05,
      "loss": 0.433,
      "step": 9600
    },
    {
      "epoch": 2.401634270149808,
      "grad_norm": 1.1896750926971436,
      "learning_rate": 2.6646356943386647e-05,
      "loss": 0.4295,
      "step": 9700
    },
    {
      "epoch": 2.4263959390862944,
      "grad_norm": 1.0336116552352905,
      "learning_rate": 2.639248540238639e-05,
      "loss": 0.4345,
      "step": 9800
    },
    {
      "epoch": 2.451157608022781,
      "grad_norm": 0.9962900876998901,
      "learning_rate": 2.613861386138614e-05,
      "loss": 0.4219,
      "step": 9900
    },
    {
      "epoch": 2.4759192769592673,
      "grad_norm": 1.2592706680297852,
      "learning_rate": 2.5884742320385884e-05,
      "loss": 0.44,
      "step": 10000
    },
    {
      "epoch": 2.4759192769592673,
      "eval_loss": 0.39048588275909424,
      "eval_runtime": 294.6469,
      "eval_samples_per_second": 48.733,
      "eval_steps_per_second": 3.048,
      "step": 10000
    },
    {
      "epoch": 2.5006809458957533,
      "grad_norm": 1.25287663936615,
      "learning_rate": 2.563087077938563e-05,
      "loss": 0.4331,
      "step": 10100
    },
    {
      "epoch": 2.5254426148322398,
      "grad_norm": 1.2010307312011719,
      "learning_rate": 2.537699923838538e-05,
      "loss": 0.427,
      "step": 10200
    },
    {
      "epoch": 2.5502042837687258,
      "grad_norm": 1.06023108959198,
      "learning_rate": 2.5123127697385123e-05,
      "loss": 0.4281,
      "step": 10300
    },
    {
      "epoch": 2.5749659527052122,
      "grad_norm": 1.094394326210022,
      "learning_rate": 2.486925615638487e-05,
      "loss": 0.4237,
      "step": 10400
    },
    {
      "epoch": 2.5997276216416987,
      "grad_norm": 1.039994716644287,
      "learning_rate": 2.461538461538462e-05,
      "loss": 0.4222,
      "step": 10500
    },
    {
      "epoch": 2.5997276216416987,
      "eval_loss": 0.3886333405971527,
      "eval_runtime": 298.9241,
      "eval_samples_per_second": 48.036,
      "eval_steps_per_second": 3.004,
      "step": 10500
    },
    {
      "epoch": 2.624489290578185,
      "grad_norm": 1.0885295867919922,
      "learning_rate": 2.4361513074384363e-05,
      "loss": 0.4297,
      "step": 10600
    },
    {
      "epoch": 2.649250959514671,
      "grad_norm": 0.9978443384170532,
      "learning_rate": 2.410764153338411e-05,
      "loss": 0.4251,
      "step": 10700
    },
    {
      "epoch": 2.6740126284511576,
      "grad_norm": 0.8032694458961487,
      "learning_rate": 2.3853769992383852e-05,
      "loss": 0.4321,
      "step": 10800
    },
    {
      "epoch": 2.698774297387644,
      "grad_norm": 1.1382070779800415,
      "learning_rate": 2.3599898451383602e-05,
      "loss": 0.4296,
      "step": 10900
    },
    {
      "epoch": 2.72353596632413,
      "grad_norm": 0.9287987947463989,
      "learning_rate": 2.334602691038335e-05,
      "loss": 0.417,
      "step": 11000
    },
    {
      "epoch": 2.72353596632413,
      "eval_loss": 0.3885912001132965,
      "eval_runtime": 299.1382,
      "eval_samples_per_second": 48.001,
      "eval_steps_per_second": 3.002,
      "step": 11000
    },
    {
      "epoch": 2.7482976352606165,
      "grad_norm": 1.0950629711151123,
      "learning_rate": 2.3092155369383092e-05,
      "loss": 0.4253,
      "step": 11100
    },
    {
      "epoch": 2.773059304197103,
      "grad_norm": 1.0151338577270508,
      "learning_rate": 2.2838283828382838e-05,
      "loss": 0.4311,
      "step": 11200
    },
    {
      "epoch": 2.7978209731335895,
      "grad_norm": 1.0384260416030884,
      "learning_rate": 2.2584412287382585e-05,
      "loss": 0.4277,
      "step": 11300
    },
    {
      "epoch": 2.8225826420700755,
      "grad_norm": 1.0595632791519165,
      "learning_rate": 2.233054074638233e-05,
      "loss": 0.4174,
      "step": 11400
    },
    {
      "epoch": 2.847344311006562,
      "grad_norm": 1.081904411315918,
      "learning_rate": 2.2076669205382078e-05,
      "loss": 0.4244,
      "step": 11500
    },
    {
      "epoch": 2.847344311006562,
      "eval_loss": 0.3848166763782501,
      "eval_runtime": 298.8323,
      "eval_samples_per_second": 48.05,
      "eval_steps_per_second": 3.005,
      "step": 11500
    },
    {
      "epoch": 2.872105979943048,
      "grad_norm": 1.3593488931655884,
      "learning_rate": 2.1822797664381824e-05,
      "loss": 0.4236,
      "step": 11600
    },
    {
      "epoch": 2.8968676488795344,
      "grad_norm": 0.9958997368812561,
      "learning_rate": 2.156892612338157e-05,
      "loss": 0.4272,
      "step": 11700
    },
    {
      "epoch": 2.921629317816021,
      "grad_norm": 0.9125483632087708,
      "learning_rate": 2.1315054582381317e-05,
      "loss": 0.4195,
      "step": 11800
    },
    {
      "epoch": 2.9463909867525073,
      "grad_norm": 1.026393175125122,
      "learning_rate": 2.1061183041381064e-05,
      "loss": 0.4216,
      "step": 11900
    },
    {
      "epoch": 2.9711526556889933,
      "grad_norm": 0.984633207321167,
      "learning_rate": 2.0807311500380807e-05,
      "loss": 0.4197,
      "step": 12000
    },
    {
      "epoch": 2.9711526556889933,
      "eval_loss": 0.3836701214313507,
      "eval_runtime": 299.7207,
      "eval_samples_per_second": 47.908,
      "eval_steps_per_second": 2.996,
      "step": 12000
    },
    {
      "epoch": 2.99591432462548,
      "grad_norm": 1.1759393215179443,
      "learning_rate": 2.0553439959380553e-05,
      "loss": 0.4182,
      "step": 12100
    },
    {
      "epoch": 3.0205521852172836,
      "grad_norm": 0.9976751208305359,
      "learning_rate": 2.02995684183803e-05,
      "loss": 0.4176,
      "step": 12200
    },
    {
      "epoch": 3.04531385415377,
      "grad_norm": 0.9128820896148682,
      "learning_rate": 2.0045696877380047e-05,
      "loss": 0.4115,
      "step": 12300
    },
    {
      "epoch": 3.070075523090256,
      "grad_norm": 0.8917282819747925,
      "learning_rate": 1.9791825336379793e-05,
      "loss": 0.4068,
      "step": 12400
    },
    {
      "epoch": 3.0948371920267426,
      "grad_norm": 1.04193115234375,
      "learning_rate": 1.9537953795379536e-05,
      "loss": 0.4138,
      "step": 12500
    },
    {
      "epoch": 3.0948371920267426,
      "eval_loss": 0.3802282214164734,
      "eval_runtime": 303.6383,
      "eval_samples_per_second": 47.29,
      "eval_steps_per_second": 2.957,
      "step": 12500
    },
    {
      "epoch": 3.119598860963229,
      "grad_norm": 1.0125975608825684,
      "learning_rate": 1.9284082254379286e-05,
      "loss": 0.4074,
      "step": 12600
    },
    {
      "epoch": 3.144360529899715,
      "grad_norm": 1.0177041292190552,
      "learning_rate": 1.9030210713379033e-05,
      "loss": 0.4179,
      "step": 12700
    },
    {
      "epoch": 3.1691221988362015,
      "grad_norm": 1.2167311906814575,
      "learning_rate": 1.8776339172378776e-05,
      "loss": 0.415,
      "step": 12800
    },
    {
      "epoch": 3.193883867772688,
      "grad_norm": 1.1083834171295166,
      "learning_rate": 1.8522467631378522e-05,
      "loss": 0.417,
      "step": 12900
    },
    {
      "epoch": 3.2186455367091744,
      "grad_norm": 1.3251594305038452,
      "learning_rate": 1.8268596090378272e-05,
      "loss": 0.4096,
      "step": 13000
    },
    {
      "epoch": 3.2186455367091744,
      "eval_loss": 0.37971043586730957,
      "eval_runtime": 307.1968,
      "eval_samples_per_second": 46.742,
      "eval_steps_per_second": 2.923,
      "step": 13000
    },
    {
      "epoch": 3.2434072056456604,
      "grad_norm": 1.1293381452560425,
      "learning_rate": 1.8014724549378015e-05,
      "loss": 0.4033,
      "step": 13100
    },
    {
      "epoch": 3.268168874582147,
      "grad_norm": 0.8698910474777222,
      "learning_rate": 1.7760853008377762e-05,
      "loss": 0.414,
      "step": 13200
    },
    {
      "epoch": 3.2929305435186333,
      "grad_norm": 0.8111106753349304,
      "learning_rate": 1.7506981467377508e-05,
      "loss": 0.4205,
      "step": 13300
    },
    {
      "epoch": 3.3176922124551194,
      "grad_norm": 1.0510135889053345,
      "learning_rate": 1.7253109926377255e-05,
      "loss": 0.4193,
      "step": 13400
    },
    {
      "epoch": 3.342453881391606,
      "grad_norm": 0.9042615294456482,
      "learning_rate": 1.6999238385377e-05,
      "loss": 0.4111,
      "step": 13500
    },
    {
      "epoch": 3.342453881391606,
      "eval_loss": 0.38024312257766724,
      "eval_runtime": 334.2385,
      "eval_samples_per_second": 42.96,
      "eval_steps_per_second": 2.687,
      "step": 13500
    },
    {
      "epoch": 3.3672155503280923,
      "grad_norm": 0.9368965029716492,
      "learning_rate": 1.6745366844376748e-05,
      "loss": 0.4124,
      "step": 13600
    },
    {
      "epoch": 3.3919772192645783,
      "grad_norm": 1.1354291439056396,
      "learning_rate": 1.649149530337649e-05,
      "loss": 0.4144,
      "step": 13700
    },
    {
      "epoch": 3.4167388882010647,
      "grad_norm": 0.862594723701477,
      "learning_rate": 1.6237623762376237e-05,
      "loss": 0.4039,
      "step": 13800
    },
    {
      "epoch": 3.441500557137551,
      "grad_norm": 1.2077276706695557,
      "learning_rate": 1.5983752221375984e-05,
      "loss": 0.4163,
      "step": 13900
    },
    {
      "epoch": 3.4662622260740372,
      "grad_norm": 1.1101346015930176,
      "learning_rate": 1.572988068037573e-05,
      "loss": 0.4103,
      "step": 14000
    },
    {
      "epoch": 3.4662622260740372,
      "eval_loss": 0.37751057744026184,
      "eval_runtime": 344.048,
      "eval_samples_per_second": 41.735,
      "eval_steps_per_second": 2.61,
      "step": 14000
    },
    {
      "epoch": 3.4910238950105237,
      "grad_norm": 0.9908251166343689,
      "learning_rate": 1.5476009139375477e-05,
      "loss": 0.4116,
      "step": 14100
    },
    {
      "epoch": 3.51578556394701,
      "grad_norm": 1.144132375717163,
      "learning_rate": 1.5222137598375222e-05,
      "loss": 0.41,
      "step": 14200
    },
    {
      "epoch": 3.5405472328834966,
      "grad_norm": 1.0292690992355347,
      "learning_rate": 1.4968266057374968e-05,
      "loss": 0.4157,
      "step": 14300
    },
    {
      "epoch": 3.5653089018199826,
      "grad_norm": 1.2643660306930542,
      "learning_rate": 1.4714394516374716e-05,
      "loss": 0.4081,
      "step": 14400
    },
    {
      "epoch": 3.590070570756469,
      "grad_norm": 1.3031563758850098,
      "learning_rate": 1.446052297537446e-05,
      "loss": 0.4098,
      "step": 14500
    },
    {
      "epoch": 3.590070570756469,
      "eval_loss": 0.3778412640094757,
      "eval_runtime": 330.1121,
      "eval_samples_per_second": 43.497,
      "eval_steps_per_second": 2.72,
      "step": 14500
    },
    {
      "epoch": 3.614832239692955,
      "grad_norm": 1.0413382053375244,
      "learning_rate": 1.4206651434374208e-05,
      "loss": 0.4075,
      "step": 14600
    },
    {
      "epoch": 3.6395939086294415,
      "grad_norm": 1.1236363649368286,
      "learning_rate": 1.3952779893373954e-05,
      "loss": 0.4049,
      "step": 14700
    },
    {
      "epoch": 3.664355577565928,
      "grad_norm": 0.9670480489730835,
      "learning_rate": 1.3698908352373699e-05,
      "loss": 0.4066,
      "step": 14800
    },
    {
      "epoch": 3.6891172465024145,
      "grad_norm": 0.8969297409057617,
      "learning_rate": 1.3445036811373446e-05,
      "loss": 0.4156,
      "step": 14900
    },
    {
      "epoch": 3.7138789154389005,
      "grad_norm": 1.0945312976837158,
      "learning_rate": 1.3191165270373194e-05,
      "loss": 0.4133,
      "step": 15000
    },
    {
      "epoch": 3.7138789154389005,
      "eval_loss": 0.3742968738079071,
      "eval_runtime": 331.5339,
      "eval_samples_per_second": 43.311,
      "eval_steps_per_second": 2.709,
      "step": 15000
    },
    {
      "epoch": 3.738640584375387,
      "grad_norm": 1.036347508430481,
      "learning_rate": 1.2937293729372937e-05,
      "loss": 0.4099,
      "step": 15100
    },
    {
      "epoch": 3.763402253311873,
      "grad_norm": 1.0111111402511597,
      "learning_rate": 1.2683422188372685e-05,
      "loss": 0.4115,
      "step": 15200
    },
    {
      "epoch": 3.7881639222483594,
      "grad_norm": 1.1293095350265503,
      "learning_rate": 1.242955064737243e-05,
      "loss": 0.4075,
      "step": 15300
    },
    {
      "epoch": 3.812925591184846,
      "grad_norm": 1.1474101543426514,
      "learning_rate": 1.2175679106372177e-05,
      "loss": 0.4182,
      "step": 15400
    },
    {
      "epoch": 3.8376872601213323,
      "grad_norm": 1.1107712984085083,
      "learning_rate": 1.1921807565371923e-05,
      "loss": 0.4081,
      "step": 15500
    },
    {
      "epoch": 3.8376872601213323,
      "eval_loss": 0.37404417991638184,
      "eval_runtime": 315.6431,
      "eval_samples_per_second": 45.491,
      "eval_steps_per_second": 2.845,
      "step": 15500
    },
    {
      "epoch": 3.8624489290578183,
      "grad_norm": 0.9492923021316528,
      "learning_rate": 1.166793602437167e-05,
      "loss": 0.3983,
      "step": 15600
    },
    {
      "epoch": 3.887210597994305,
      "grad_norm": 1.1327282190322876,
      "learning_rate": 1.1414064483371414e-05,
      "loss": 0.4043,
      "step": 15700
    },
    {
      "epoch": 3.9119722669307913,
      "grad_norm": 1.0399820804595947,
      "learning_rate": 1.1160192942371161e-05,
      "loss": 0.406,
      "step": 15800
    },
    {
      "epoch": 3.9367339358672773,
      "grad_norm": 1.1408737897872925,
      "learning_rate": 1.0906321401370907e-05,
      "loss": 0.4065,
      "step": 15900
    },
    {
      "epoch": 3.9614956048037637,
      "grad_norm": 0.9349102973937988,
      "learning_rate": 1.0652449860370652e-05,
      "loss": 0.4056,
      "step": 16000
    },
    {
      "epoch": 3.9614956048037637,
      "eval_loss": 0.37279391288757324,
      "eval_runtime": 294.1835,
      "eval_samples_per_second": 48.81,
      "eval_steps_per_second": 3.053,
      "step": 16000
    },
    {
      "epoch": 3.98625727374025,
      "grad_norm": 0.951724112033844,
      "learning_rate": 1.0398578319370399e-05,
      "loss": 0.4056,
      "step": 16100
    },
    {
      "epoch": 4.010895134332054,
      "grad_norm": 1.131766438484192,
      "learning_rate": 1.0144706778370145e-05,
      "loss": 0.3989,
      "step": 16200
    },
    {
      "epoch": 4.03565680326854,
      "grad_norm": 1.0067100524902344,
      "learning_rate": 9.890835237369892e-06,
      "loss": 0.4026,
      "step": 16300
    },
    {
      "epoch": 4.0604184722050265,
      "grad_norm": 1.0153299570083618,
      "learning_rate": 9.636963696369637e-06,
      "loss": 0.4013,
      "step": 16400
    },
    {
      "epoch": 4.085180141141513,
      "grad_norm": 1.2241690158843994,
      "learning_rate": 9.383092155369383e-06,
      "loss": 0.4,
      "step": 16500
    },
    {
      "epoch": 4.085180141141513,
      "eval_loss": 0.3727540373802185,
      "eval_runtime": 296.4387,
      "eval_samples_per_second": 48.438,
      "eval_steps_per_second": 3.029,
      "step": 16500
    },
    {
      "epoch": 4.109941810077999,
      "grad_norm": 0.9304781556129456,
      "learning_rate": 9.12922061436913e-06,
      "loss": 0.3969,
      "step": 16600
    },
    {
      "epoch": 4.134703479014486,
      "grad_norm": 1.1999999284744263,
      "learning_rate": 8.875349073368876e-06,
      "loss": 0.3921,
      "step": 16700
    },
    {
      "epoch": 4.159465147950971,
      "grad_norm": 1.0765935182571411,
      "learning_rate": 8.621477532368621e-06,
      "loss": 0.409,
      "step": 16800
    },
    {
      "epoch": 4.184226816887458,
      "grad_norm": 0.886246383190155,
      "learning_rate": 8.367605991368369e-06,
      "loss": 0.4019,
      "step": 16900
    },
    {
      "epoch": 4.208988485823944,
      "grad_norm": 1.0984737873077393,
      "learning_rate": 8.113734450368114e-06,
      "loss": 0.3991,
      "step": 17000
    },
    {
      "epoch": 4.208988485823944,
      "eval_loss": 0.3709471821784973,
      "eval_runtime": 298.2727,
      "eval_samples_per_second": 48.141,
      "eval_steps_per_second": 3.011,
      "step": 17000
    },
    {
      "epoch": 4.233750154760431,
      "grad_norm": 1.0152359008789062,
      "learning_rate": 7.85986290936786e-06,
      "loss": 0.3962,
      "step": 17100
    },
    {
      "epoch": 4.258511823696917,
      "grad_norm": 1.1482372283935547,
      "learning_rate": 7.605991368367606e-06,
      "loss": 0.4,
      "step": 17200
    },
    {
      "epoch": 4.283273492633404,
      "grad_norm": 0.9904143810272217,
      "learning_rate": 7.352119827367353e-06,
      "loss": 0.3918,
      "step": 17300
    },
    {
      "epoch": 4.30803516156989,
      "grad_norm": 1.0268852710723877,
      "learning_rate": 7.098248286367098e-06,
      "loss": 0.4049,
      "step": 17400
    },
    {
      "epoch": 4.332796830506376,
      "grad_norm": 1.005434274673462,
      "learning_rate": 6.844376745366844e-06,
      "loss": 0.4009,
      "step": 17500
    },
    {
      "epoch": 4.332796830506376,
      "eval_loss": 0.3704555034637451,
      "eval_runtime": 298.2902,
      "eval_samples_per_second": 48.138,
      "eval_steps_per_second": 3.01,
      "step": 17500
    },
    {
      "epoch": 4.357558499442862,
      "grad_norm": 1.2603648900985718,
      "learning_rate": 6.590505204366591e-06,
      "loss": 0.4073,
      "step": 17600
    },
    {
      "epoch": 4.382320168379349,
      "grad_norm": 1.231177568435669,
      "learning_rate": 6.336633663366337e-06,
      "loss": 0.4009,
      "step": 17700
    },
    {
      "epoch": 4.407081837315835,
      "grad_norm": 1.0092039108276367,
      "learning_rate": 6.082762122366083e-06,
      "loss": 0.3984,
      "step": 17800
    },
    {
      "epoch": 4.431843506252322,
      "grad_norm": 1.1094638109207153,
      "learning_rate": 5.828890581365829e-06,
      "loss": 0.4037,
      "step": 17900
    },
    {
      "epoch": 4.456605175188808,
      "grad_norm": 1.2728139162063599,
      "learning_rate": 5.575019040365576e-06,
      "loss": 0.3981,
      "step": 18000
    },
    {
      "epoch": 4.456605175188808,
      "eval_loss": 0.3691825568675995,
      "eval_runtime": 300.7735,
      "eval_samples_per_second": 47.74,
      "eval_steps_per_second": 2.986,
      "step": 18000
    },
    {
      "epoch": 4.481366844125294,
      "grad_norm": 0.9797173738479614,
      "learning_rate": 5.321147499365321e-06,
      "loss": 0.3858,
      "step": 18100
    },
    {
      "epoch": 4.50612851306178,
      "grad_norm": 1.1960067749023438,
      "learning_rate": 5.067275958365068e-06,
      "loss": 0.4028,
      "step": 18200
    },
    {
      "epoch": 4.5308901819982665,
      "grad_norm": 1.067478895187378,
      "learning_rate": 4.8134044173648135e-06,
      "loss": 0.4064,
      "step": 18300
    },
    {
      "epoch": 4.555651850934753,
      "grad_norm": 1.0445499420166016,
      "learning_rate": 4.55953287636456e-06,
      "loss": 0.3965,
      "step": 18400
    },
    {
      "epoch": 4.5804135198712395,
      "grad_norm": 1.035979986190796,
      "learning_rate": 4.3056613353643065e-06,
      "loss": 0.3909,
      "step": 18500
    },
    {
      "epoch": 4.5804135198712395,
      "eval_loss": 0.3688410818576813,
      "eval_runtime": 305.0552,
      "eval_samples_per_second": 47.07,
      "eval_steps_per_second": 2.944,
      "step": 18500
    },
    {
      "epoch": 4.605175188807726,
      "grad_norm": 0.9728735685348511,
      "learning_rate": 4.051789794364052e-06,
      "loss": 0.3943,
      "step": 18600
    },
    {
      "epoch": 4.629936857744212,
      "grad_norm": 1.086190938949585,
      "learning_rate": 3.7979182533637982e-06,
      "loss": 0.3992,
      "step": 18700
    },
    {
      "epoch": 4.654698526680698,
      "grad_norm": 1.103594183921814,
      "learning_rate": 3.544046712363544e-06,
      "loss": 0.3891,
      "step": 18800
    },
    {
      "epoch": 4.679460195617184,
      "grad_norm": 1.0151804685592651,
      "learning_rate": 3.2901751713632904e-06,
      "loss": 0.4009,
      "step": 18900
    },
    {
      "epoch": 4.704221864553671,
      "grad_norm": 1.141107201576233,
      "learning_rate": 3.0363036303630365e-06,
      "loss": 0.4028,
      "step": 19000
    },
    {
      "epoch": 4.704221864553671,
      "eval_loss": 0.36849796772003174,
      "eval_runtime": 296.4027,
      "eval_samples_per_second": 48.444,
      "eval_steps_per_second": 3.03,
      "step": 19000
    },
    {
      "epoch": 4.728983533490157,
      "grad_norm": 1.0555781126022339,
      "learning_rate": 2.7824320893627826e-06,
      "loss": 0.4033,
      "step": 19100
    },
    {
      "epoch": 4.753745202426644,
      "grad_norm": 1.1756597757339478,
      "learning_rate": 2.5285605483625287e-06,
      "loss": 0.4022,
      "step": 19200
    },
    {
      "epoch": 4.77850687136313,
      "grad_norm": 0.9200191497802734,
      "learning_rate": 2.2746890073622748e-06,
      "loss": 0.3987,
      "step": 19300
    },
    {
      "epoch": 4.803268540299616,
      "grad_norm": 1.1391457319259644,
      "learning_rate": 2.0208174663620213e-06,
      "loss": 0.3961,
      "step": 19400
    },
    {
      "epoch": 4.828030209236102,
      "grad_norm": 1.0987917184829712,
      "learning_rate": 1.7669459253617672e-06,
      "loss": 0.3903,
      "step": 19500
    },
    {
      "epoch": 4.828030209236102,
      "eval_loss": 0.367659330368042,
      "eval_runtime": 300.2959,
      "eval_samples_per_second": 47.816,
      "eval_steps_per_second": 2.99,
      "step": 19500
    },
    {
      "epoch": 4.852791878172589,
      "grad_norm": 0.9229596257209778,
      "learning_rate": 1.5130743843615132e-06,
      "loss": 0.3976,
      "step": 19600
    },
    {
      "epoch": 4.877553547109075,
      "grad_norm": 0.9962015151977539,
      "learning_rate": 1.2592028433612593e-06,
      "loss": 0.3957,
      "step": 19700
    },
    {
      "epoch": 4.902315216045562,
      "grad_norm": 1.4311672449111938,
      "learning_rate": 1.0053313023610054e-06,
      "loss": 0.3918,
      "step": 19800
    },
    {
      "epoch": 4.927076884982048,
      "grad_norm": 1.1852339506149292,
      "learning_rate": 7.514597613607515e-07,
      "loss": 0.3989,
      "step": 19900
    },
    {
      "epoch": 4.951838553918535,
      "grad_norm": 1.072943091392517,
      "learning_rate": 4.975882203604976e-07,
      "loss": 0.393,
      "step": 20000
    },
    {
      "epoch": 4.951838553918535,
      "eval_loss": 0.3671196699142456,
      "eval_runtime": 304.303,
      "eval_samples_per_second": 47.187,
      "eval_steps_per_second": 2.951,
      "step": 20000
    }
  ],
  "logging_steps": 100,
  "max_steps": 20195,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7.046592561946024e+21,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}