{
  "best_global_step": 3000,
  "best_metric": 3.6842086777328205,
  "best_model_checkpoint": "checkpoints/gpt-2_seq1024_mla0-96-192/checkpoint-3000",
  "epoch": 0.9042272624519629,
  "eval_steps": 300,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.006028181749679753,
      "grad_norm": 2.0579254627227783,
      "learning_rate": 3.166666666666667e-05,
      "loss": 10.4161,
      "step": 20
    },
    {
      "epoch": 0.012056363499359506,
      "grad_norm": 1.7086316347122192,
      "learning_rate": 6.500000000000001e-05,
      "loss": 9.2798,
      "step": 40
    },
    {
      "epoch": 0.01808454524903926,
      "grad_norm": 0.830778181552887,
      "learning_rate": 9.833333333333333e-05,
      "loss": 8.2583,
      "step": 60
    },
    {
      "epoch": 0.02411272699871901,
      "grad_norm": 1.3369255065917969,
      "learning_rate": 0.00013166666666666665,
      "loss": 7.5659,
      "step": 80
    },
    {
      "epoch": 0.030140908748398764,
      "grad_norm": 0.7765972018241882,
      "learning_rate": 0.000165,
      "loss": 7.2475,
      "step": 100
    },
    {
      "epoch": 0.03616909049807852,
      "grad_norm": 0.576606273651123,
      "learning_rate": 0.00019833333333333335,
      "loss": 6.9385,
      "step": 120
    },
    {
      "epoch": 0.04219727224775827,
      "grad_norm": 1.2204625606536865,
      "learning_rate": 0.00023166666666666667,
      "loss": 6.6843,
      "step": 140
    },
    {
      "epoch": 0.04822545399743802,
      "grad_norm": 1.2481485605239868,
      "learning_rate": 0.00026500000000000004,
      "loss": 6.479,
      "step": 160
    },
    {
      "epoch": 0.05425363574711778,
      "grad_norm": 0.7792941927909851,
      "learning_rate": 0.00029833333333333334,
      "loss": 6.3156,
      "step": 180
    },
    {
      "epoch": 0.06028181749679753,
      "grad_norm": 1.0211087465286255,
      "learning_rate": 0.0003316666666666667,
      "loss": 6.1724,
      "step": 200
    },
    {
      "epoch": 0.06630999924647728,
      "grad_norm": 1.0964252948760986,
      "learning_rate": 0.000365,
      "loss": 6.025,
      "step": 220
    },
    {
      "epoch": 0.07233818099615703,
      "grad_norm": 1.1124714612960815,
      "learning_rate": 0.00039833333333333333,
      "loss": 5.9008,
      "step": 240
    },
    {
      "epoch": 0.07836636274583679,
      "grad_norm": 1.0428141355514526,
      "learning_rate": 0.0004316666666666667,
      "loss": 5.7927,
      "step": 260
    },
    {
      "epoch": 0.08439454449551655,
      "grad_norm": 0.7066069841384888,
      "learning_rate": 0.000465,
      "loss": 5.6734,
      "step": 280
    },
    {
      "epoch": 0.09042272624519629,
      "grad_norm": 1.2171207666397095,
      "learning_rate": 0.0004983333333333334,
      "loss": 5.589,
      "step": 300
    },
    {
      "epoch": 0.09042272624519629,
      "eval_loss": 5.521408836218453,
      "eval_perplexity": 249.9869799281079,
      "eval_runtime": 112.3146,
      "eval_samples_per_second": 14.771,
      "eval_steps_per_second": 0.231,
      "step": 300
    },
    {
      "epoch": 0.09645090799487605,
      "grad_norm": 0.8410649299621582,
      "learning_rate": 0.0004964814814814814,
      "loss": 5.4626,
      "step": 320
    },
    {
      "epoch": 0.1024790897445558,
      "grad_norm": 0.8579843640327454,
      "learning_rate": 0.0004927777777777777,
      "loss": 5.3854,
      "step": 340
    },
    {
      "epoch": 0.10850727149423556,
      "grad_norm": 0.8943135142326355,
      "learning_rate": 0.0004890740740740741,
      "loss": 5.3027,
      "step": 360
    },
    {
      "epoch": 0.1145354532439153,
      "grad_norm": 1.0680087804794312,
      "learning_rate": 0.0004853703703703704,
      "loss": 5.2049,
      "step": 380
    },
    {
      "epoch": 0.12056363499359506,
      "grad_norm": 0.8171730637550354,
      "learning_rate": 0.0004816666666666667,
      "loss": 5.1464,
      "step": 400
    },
    {
      "epoch": 0.1265918167432748,
      "grad_norm": 0.713478684425354,
      "learning_rate": 0.00047796296296296297,
      "loss": 5.0611,
      "step": 420
    },
    {
      "epoch": 0.13261999849295456,
      "grad_norm": 0.7907434701919556,
      "learning_rate": 0.0004742592592592593,
      "loss": 5.0196,
      "step": 440
    },
    {
      "epoch": 0.1386481802426343,
      "grad_norm": 0.5732178092002869,
      "learning_rate": 0.00047055555555555555,
      "loss": 4.9585,
      "step": 460
    },
    {
      "epoch": 0.14467636199231407,
      "grad_norm": 0.7607829570770264,
      "learning_rate": 0.00046685185185185187,
      "loss": 4.9042,
      "step": 480
    },
    {
      "epoch": 0.15070454374199382,
      "grad_norm": 0.7640454769134521,
      "learning_rate": 0.00046314814814814813,
      "loss": 4.8657,
      "step": 500
    },
    {
      "epoch": 0.15673272549167358,
      "grad_norm": 1.1241954565048218,
      "learning_rate": 0.00045944444444444445,
      "loss": 4.8114,
      "step": 520
    },
    {
      "epoch": 0.16276090724135334,
      "grad_norm": 0.7967432141304016,
      "learning_rate": 0.0004557407407407407,
      "loss": 4.766,
      "step": 540
    },
    {
      "epoch": 0.1687890889910331,
      "grad_norm": 0.7019729018211365,
      "learning_rate": 0.00045203703703703703,
      "loss": 4.7274,
      "step": 560
    },
    {
      "epoch": 0.17481727074071282,
      "grad_norm": 0.7308072447776794,
      "learning_rate": 0.0004483333333333333,
      "loss": 4.6837,
      "step": 580
    },
    {
      "epoch": 0.18084545249039258,
      "grad_norm": 0.776524007320404,
      "learning_rate": 0.00044462962962962967,
      "loss": 4.6383,
      "step": 600
    },
    {
      "epoch": 0.18084545249039258,
      "eval_loss": 4.580472151368435,
      "eval_perplexity": 97.56044663272338,
      "eval_runtime": 112.0483,
      "eval_samples_per_second": 14.806,
      "eval_steps_per_second": 0.232,
      "step": 600
    },
    {
      "epoch": 0.18687363424007233,
      "grad_norm": 0.9698750376701355,
      "learning_rate": 0.00044092592592592594,
      "loss": 4.601,
      "step": 620
    },
    {
      "epoch": 0.1929018159897521,
      "grad_norm": 0.9107672572135925,
      "learning_rate": 0.00043722222222222225,
      "loss": 4.5588,
      "step": 640
    },
    {
      "epoch": 0.19892999773943185,
      "grad_norm": 0.9225994944572449,
      "learning_rate": 0.0004335185185185185,
      "loss": 4.5222,
      "step": 660
    },
    {
      "epoch": 0.2049581794891116,
      "grad_norm": 0.6751540899276733,
      "learning_rate": 0.00042981481481481484,
      "loss": 4.4938,
      "step": 680
    },
    {
      "epoch": 0.21098636123879136,
      "grad_norm": 0.9872696995735168,
      "learning_rate": 0.0004261111111111111,
      "loss": 4.4514,
      "step": 700
    },
    {
      "epoch": 0.21701454298847112,
      "grad_norm": 0.8386589884757996,
      "learning_rate": 0.0004224074074074074,
      "loss": 4.4183,
      "step": 720
    },
    {
      "epoch": 0.22304272473815084,
      "grad_norm": 0.8140395879745483,
      "learning_rate": 0.0004187037037037037,
      "loss": 4.3853,
      "step": 740
    },
    {
      "epoch": 0.2290709064878306,
      "grad_norm": 0.86305171251297,
      "learning_rate": 0.000415,
      "loss": 4.3682,
      "step": 760
    },
    {
      "epoch": 0.23509908823751036,
      "grad_norm": 0.5293642282485962,
      "learning_rate": 0.00041129629629629627,
      "loss": 4.3428,
      "step": 780
    },
    {
      "epoch": 0.2411272699871901,
      "grad_norm": 0.9141278266906738,
      "learning_rate": 0.00040759259259259264,
      "loss": 4.3232,
      "step": 800
    },
    {
      "epoch": 0.24715545173686987,
      "grad_norm": 0.8114584684371948,
      "learning_rate": 0.0004038888888888889,
      "loss": 4.3026,
      "step": 820
    },
    {
      "epoch": 0.2531836334865496,
      "grad_norm": 0.642278790473938,
      "learning_rate": 0.0004001851851851852,
      "loss": 4.2781,
      "step": 840
    },
    {
      "epoch": 0.2592118152362294,
      "grad_norm": 0.7550613284111023,
      "learning_rate": 0.0003964814814814815,
      "loss": 4.2606,
      "step": 860
    },
    {
      "epoch": 0.2652399969859091,
      "grad_norm": 1.0664489269256592,
      "learning_rate": 0.0003927777777777778,
      "loss": 4.244,
      "step": 880
    },
    {
      "epoch": 0.2712681787355889,
      "grad_norm": 0.7011036276817322,
      "learning_rate": 0.00038907407407407407,
      "loss": 4.2329,
      "step": 900
    },
    {
      "epoch": 0.2712681787355889,
      "eval_loss": 4.179822730749129,
      "eval_perplexity": 65.35426688515841,
      "eval_runtime": 112.1531,
      "eval_samples_per_second": 14.792,
      "eval_steps_per_second": 0.232,
      "step": 900
    },
    {
      "epoch": 0.2772963604852686,
      "grad_norm": 0.6399344801902771,
      "learning_rate": 0.0003853703703703704,
      "loss": 4.2089,
      "step": 920
    },
    {
      "epoch": 0.2833245422349484,
      "grad_norm": 0.7618056535720825,
      "learning_rate": 0.00038166666666666666,
      "loss": 4.2015,
      "step": 940
    },
    {
      "epoch": 0.28935272398462814,
      "grad_norm": 0.6749306917190552,
      "learning_rate": 0.000377962962962963,
      "loss": 4.1759,
      "step": 960
    },
    {
      "epoch": 0.29538090573430786,
      "grad_norm": 0.8555495738983154,
      "learning_rate": 0.00037425925925925924,
      "loss": 4.175,
      "step": 980
    },
    {
      "epoch": 0.30140908748398765,
      "grad_norm": 1.014050841331482,
      "learning_rate": 0.0003705555555555556,
      "loss": 4.159,
      "step": 1000
    },
    {
      "epoch": 0.3074372692336674,
      "grad_norm": 0.569524884223938,
      "learning_rate": 0.0003668518518518519,
      "loss": 4.1444,
      "step": 1020
    },
    {
      "epoch": 0.31346545098334716,
      "grad_norm": 0.7568202018737793,
      "learning_rate": 0.0003631481481481482,
      "loss": 4.1294,
      "step": 1040
    },
    {
      "epoch": 0.3194936327330269,
      "grad_norm": 0.8762944340705872,
      "learning_rate": 0.00035944444444444446,
      "loss": 4.1241,
      "step": 1060
    },
    {
      "epoch": 0.3255218144827067,
      "grad_norm": 0.6150480508804321,
      "learning_rate": 0.0003557407407407408,
      "loss": 4.1143,
      "step": 1080
    },
    {
      "epoch": 0.3315499962323864,
      "grad_norm": 0.6287286877632141,
      "learning_rate": 0.00035203703703703704,
      "loss": 4.0968,
      "step": 1100
    },
    {
      "epoch": 0.3375781779820662,
      "grad_norm": 0.6983112096786499,
      "learning_rate": 0.00034833333333333336,
      "loss": 4.0863,
      "step": 1120
    },
    {
      "epoch": 0.3436063597317459,
      "grad_norm": 0.8278563618659973,
      "learning_rate": 0.0003446296296296296,
      "loss": 4.0776,
      "step": 1140
    },
    {
      "epoch": 0.34963454148142564,
      "grad_norm": 0.6279471516609192,
      "learning_rate": 0.0003409259259259259,
      "loss": 4.0651,
      "step": 1160
    },
    {
      "epoch": 0.3556627232311054,
      "grad_norm": 0.79290771484375,
      "learning_rate": 0.0003372222222222222,
      "loss": 4.058,
      "step": 1180
    },
    {
      "epoch": 0.36169090498078516,
      "grad_norm": 0.5708630681037903,
      "learning_rate": 0.0003335185185185185,
      "loss": 4.0533,
      "step": 1200
    },
    {
      "epoch": 0.36169090498078516,
      "eval_loss": 3.9993530909338384,
      "eval_perplexity": 54.562841416855505,
      "eval_runtime": 112.1178,
      "eval_samples_per_second": 14.797,
      "eval_steps_per_second": 0.232,
      "step": 1200
    },
    {
      "epoch": 0.36771908673046494,
      "grad_norm": 0.5408058166503906,
      "learning_rate": 0.00032981481481481485,
      "loss": 4.0346,
      "step": 1220
    },
    {
      "epoch": 0.37374726848014467,
      "grad_norm": 0.6973796486854553,
      "learning_rate": 0.0003261111111111111,
      "loss": 4.034,
      "step": 1240
    },
    {
      "epoch": 0.37977545022982445,
      "grad_norm": 0.5908318161964417,
      "learning_rate": 0.00032240740740740743,
      "loss": 4.0211,
      "step": 1260
    },
    {
      "epoch": 0.3858036319795042,
      "grad_norm": 0.7799516916275024,
      "learning_rate": 0.0003187037037037037,
      "loss": 4.0133,
      "step": 1280
    },
    {
      "epoch": 0.3918318137291839,
      "grad_norm": 0.6305871605873108,
      "learning_rate": 0.000315,
      "loss": 4.0094,
      "step": 1300
    },
    {
      "epoch": 0.3978599954788637,
      "grad_norm": 0.5868816375732422,
      "learning_rate": 0.0003112962962962963,
      "loss": 3.9963,
      "step": 1320
    },
    {
      "epoch": 0.4038881772285434,
      "grad_norm": 0.5462016463279724,
      "learning_rate": 0.0003075925925925926,
      "loss": 3.9959,
      "step": 1340
    },
    {
      "epoch": 0.4099163589782232,
      "grad_norm": 0.9148553609848022,
      "learning_rate": 0.00030388888888888886,
      "loss": 3.9841,
      "step": 1360
    },
    {
      "epoch": 0.41594454072790293,
      "grad_norm": 0.9339419007301331,
      "learning_rate": 0.0003001851851851852,
      "loss": 3.9774,
      "step": 1380
    },
    {
      "epoch": 0.4219727224775827,
      "grad_norm": 0.5173079371452332,
      "learning_rate": 0.00029648148148148144,
      "loss": 3.9714,
      "step": 1400
    },
    {
      "epoch": 0.42800090422726245,
      "grad_norm": 0.6499783992767334,
      "learning_rate": 0.0002927777777777778,
      "loss": 3.9704,
      "step": 1420
    },
    {
      "epoch": 0.43402908597694223,
      "grad_norm": 0.5502707362174988,
      "learning_rate": 0.0002890740740740741,
      "loss": 3.9525,
      "step": 1440
    },
    {
      "epoch": 0.44005726772662196,
      "grad_norm": 0.7076614499092102,
      "learning_rate": 0.0002853703703703704,
      "loss": 3.9593,
      "step": 1460
    },
    {
      "epoch": 0.4460854494763017,
      "grad_norm": 1.0200659036636353,
      "learning_rate": 0.00028166666666666666,
      "loss": 3.9477,
      "step": 1480
    },
    {
      "epoch": 0.45211363122598147,
      "grad_norm": 0.6163772344589233,
      "learning_rate": 0.000277962962962963,
      "loss": 3.9433,
      "step": 1500
    },
    {
      "epoch": 0.45211363122598147,
      "eval_loss": 3.8898620696995034,
      "eval_perplexity": 48.904140695693854,
      "eval_runtime": 112.2086,
      "eval_samples_per_second": 14.785,
      "eval_steps_per_second": 0.232,
      "step": 1500
    },
    {
      "epoch": 0.4581418129756612,
      "grad_norm": 0.5702603459358215,
      "learning_rate": 0.00027425925925925925,
      "loss": 3.9363,
      "step": 1520
    },
    {
      "epoch": 0.464169994725341,
      "grad_norm": 0.6020569205284119,
      "learning_rate": 0.00027055555555555557,
      "loss": 3.9312,
      "step": 1540
    },
    {
      "epoch": 0.4701981764750207,
      "grad_norm": 0.5328924655914307,
      "learning_rate": 0.00026685185185185183,
      "loss": 3.9189,
      "step": 1560
    },
    {
      "epoch": 0.4762263582247005,
      "grad_norm": 0.5905847549438477,
      "learning_rate": 0.00026314814814814815,
      "loss": 3.9174,
      "step": 1580
    },
    {
      "epoch": 0.4822545399743802,
      "grad_norm": 0.8177961707115173,
      "learning_rate": 0.0002594444444444444,
      "loss": 3.9032,
      "step": 1600
    },
    {
      "epoch": 0.48828272172405995,
      "grad_norm": 0.6691105961799622,
      "learning_rate": 0.0002557407407407408,
      "loss": 3.912,
      "step": 1620
    },
    {
      "epoch": 0.49431090347373974,
      "grad_norm": 0.6740971207618713,
      "learning_rate": 0.00025203703703703705,
      "loss": 3.9008,
      "step": 1640
    },
    {
      "epoch": 0.5003390852234195,
      "grad_norm": 0.47609272599220276,
      "learning_rate": 0.0002483333333333333,
      "loss": 3.8981,
      "step": 1660
    },
    {
      "epoch": 0.5063672669730992,
      "grad_norm": 0.575442373752594,
      "learning_rate": 0.00024462962962962963,
      "loss": 3.8951,
      "step": 1680
    },
    {
      "epoch": 0.512395448722779,
      "grad_norm": 0.7211048603057861,
      "learning_rate": 0.00024092592592592593,
      "loss": 3.8819,
      "step": 1700
    },
    {
      "epoch": 0.5184236304724588,
      "grad_norm": 0.7259971499443054,
      "learning_rate": 0.00023722222222222222,
      "loss": 3.8821,
      "step": 1720
    },
    {
      "epoch": 0.5244518122221385,
      "grad_norm": 0.670127809047699,
      "learning_rate": 0.0002335185185185185,
      "loss": 3.881,
      "step": 1740
    },
    {
      "epoch": 0.5304799939718182,
      "grad_norm": 0.5846367478370667,
      "learning_rate": 0.0002298148148148148,
      "loss": 3.8784,
      "step": 1760
    },
    {
      "epoch": 0.536508175721498,
      "grad_norm": 0.6691552996635437,
      "learning_rate": 0.00022611111111111112,
      "loss": 3.8665,
      "step": 1780
    },
    {
      "epoch": 0.5425363574711778,
      "grad_norm": 0.6708612442016602,
      "learning_rate": 0.0002224074074074074,
      "loss": 3.8692,
      "step": 1800
    },
    {
      "epoch": 0.5425363574711778,
      "eval_loss": 3.818516127264596,
      "eval_perplexity": 45.53658766213536,
      "eval_runtime": 112.2232,
      "eval_samples_per_second": 14.783,
      "eval_steps_per_second": 0.232,
      "step": 1800
    },
    {
      "epoch": 0.5485645392208575,
      "grad_norm": 0.6356279253959656,
      "learning_rate": 0.0002187037037037037,
      "loss": 3.8679,
      "step": 1820
    },
    {
      "epoch": 0.5545927209705372,
      "grad_norm": 0.60150545835495,
      "learning_rate": 0.000215,
      "loss": 3.8581,
      "step": 1840
    },
    {
      "epoch": 0.560620902720217,
      "grad_norm": 0.6865763664245605,
      "learning_rate": 0.00021129629629629629,
      "loss": 3.8506,
      "step": 1860
    },
    {
      "epoch": 0.5666490844698968,
      "grad_norm": 0.5327039361000061,
      "learning_rate": 0.0002075925925925926,
      "loss": 3.85,
      "step": 1880
    },
    {
      "epoch": 0.5726772662195765,
      "grad_norm": 0.6210691332817078,
      "learning_rate": 0.0002038888888888889,
      "loss": 3.8408,
      "step": 1900
    },
    {
      "epoch": 0.5787054479692563,
      "grad_norm": 0.5542188286781311,
      "learning_rate": 0.0002001851851851852,
      "loss": 3.844,
      "step": 1920
    },
    {
      "epoch": 0.584733629718936,
      "grad_norm": 0.5724864602088928,
      "learning_rate": 0.00019648148148148148,
      "loss": 3.8377,
      "step": 1940
    },
    {
      "epoch": 0.5907618114686157,
      "grad_norm": 0.5111119747161865,
      "learning_rate": 0.00019277777777777777,
      "loss": 3.8389,
      "step": 1960
    },
    {
      "epoch": 0.5967899932182955,
      "grad_norm": 0.6223249435424805,
      "learning_rate": 0.00018907407407407406,
      "loss": 3.8272,
      "step": 1980
    },
    {
      "epoch": 0.6028181749679753,
      "grad_norm": 0.5860840082168579,
      "learning_rate": 0.00018537037037037038,
      "loss": 3.8289,
      "step": 2000
    },
    {
      "epoch": 0.6088463567176551,
      "grad_norm": 0.5481979250907898,
      "learning_rate": 0.00018166666666666667,
      "loss": 3.824,
      "step": 2020
    },
    {
      "epoch": 0.6148745384673348,
      "grad_norm": 0.6854233741760254,
      "learning_rate": 0.00017796296296296296,
      "loss": 3.8203,
      "step": 2040
    },
    {
      "epoch": 0.6209027202170145,
      "grad_norm": 0.5089332461357117,
      "learning_rate": 0.00017425925925925926,
      "loss": 3.8201,
      "step": 2060
    },
    {
      "epoch": 0.6269309019666943,
      "grad_norm": 0.6237172484397888,
      "learning_rate": 0.00017055555555555555,
      "loss": 3.8169,
      "step": 2080
    },
    {
      "epoch": 0.632959083716374,
      "grad_norm": 0.551239013671875,
      "learning_rate": 0.00016685185185185187,
      "loss": 3.8122,
      "step": 2100
    },
    {
      "epoch": 0.632959083716374,
      "eval_loss": 3.7647549420147928,
      "eval_perplexity": 43.153129545026054,
      "eval_runtime": 112.0835,
      "eval_samples_per_second": 14.801,
      "eval_steps_per_second": 0.232,
      "step": 2100
    },
    {
      "epoch": 0.6389872654660538,
      "grad_norm": 0.4622141420841217,
      "learning_rate": 0.00016314814814814816,
      "loss": 3.8107,
      "step": 2120
    },
    {
      "epoch": 0.6450154472157336,
      "grad_norm": 0.5357872843742371,
      "learning_rate": 0.00015944444444444445,
      "loss": 3.8084,
      "step": 2140
    },
    {
      "epoch": 0.6510436289654133,
      "grad_norm": 0.4540112018585205,
      "learning_rate": 0.00015574074074074074,
      "loss": 3.8055,
      "step": 2160
    },
    {
      "epoch": 0.657071810715093,
      "grad_norm": 0.4942857325077057,
      "learning_rate": 0.00015203703703703703,
      "loss": 3.8048,
      "step": 2180
    },
    {
      "epoch": 0.6630999924647728,
      "grad_norm": 0.7828388214111328,
      "learning_rate": 0.00014833333333333335,
      "loss": 3.8011,
      "step": 2200
    },
    {
      "epoch": 0.6691281742144526,
      "grad_norm": 0.5057414174079895,
      "learning_rate": 0.00014462962962962964,
      "loss": 3.7976,
      "step": 2220
    },
    {
      "epoch": 0.6751563559641324,
      "grad_norm": 0.5598515272140503,
      "learning_rate": 0.00014092592592592594,
      "loss": 3.7964,
      "step": 2240
    },
    {
      "epoch": 0.681184537713812,
      "grad_norm": 0.46953949332237244,
      "learning_rate": 0.00013722222222222223,
      "loss": 3.7826,
      "step": 2260
    },
    {
      "epoch": 0.6872127194634918,
      "grad_norm": 0.5060424208641052,
      "learning_rate": 0.00013351851851851852,
      "loss": 3.7895,
      "step": 2280
    },
    {
      "epoch": 0.6932409012131716,
      "grad_norm": 0.5652453899383545,
      "learning_rate": 0.00012981481481481484,
      "loss": 3.7831,
      "step": 2300
    },
    {
      "epoch": 0.6992690829628513,
      "grad_norm": 0.5393581986427307,
      "learning_rate": 0.00012611111111111113,
      "loss": 3.7764,
      "step": 2320
    },
    {
      "epoch": 0.7052972647125311,
      "grad_norm": 0.5593135952949524,
      "learning_rate": 0.00012240740740740742,
      "loss": 3.7779,
      "step": 2340
    },
    {
      "epoch": 0.7113254464622109,
      "grad_norm": 0.5752879977226257,
      "learning_rate": 0.0001187037037037037,
      "loss": 3.7771,
      "step": 2360
    },
    {
      "epoch": 0.7173536282118906,
      "grad_norm": 0.5519211888313293,
      "learning_rate": 0.000115,
      "loss": 3.7681,
      "step": 2380
    },
    {
      "epoch": 0.7233818099615703,
      "grad_norm": 0.4668697714805603,
      "learning_rate": 0.0001112962962962963,
      "loss": 3.7757,
      "step": 2400
    },
    {
      "epoch": 0.7233818099615703,
      "eval_loss": 3.72596640661412,
      "eval_perplexity": 41.51133019054871,
      "eval_runtime": 112.0997,
      "eval_samples_per_second": 14.799,
      "eval_steps_per_second": 0.232,
      "step": 2400
    },
    {
      "epoch": 0.7294099917112501,
      "grad_norm": 0.4993445873260498,
      "learning_rate": 0.00010759259259259259,
      "loss": 3.7721,
      "step": 2420
    },
    {
      "epoch": 0.7354381734609299,
      "grad_norm": 0.4481184482574463,
      "learning_rate": 0.00010388888888888889,
      "loss": 3.7699,
      "step": 2440
    },
    {
      "epoch": 0.7414663552106096,
      "grad_norm": 0.4291679859161377,
      "learning_rate": 0.00010018518518518518,
      "loss": 3.7695,
      "step": 2460
    },
    {
      "epoch": 0.7474945369602893,
      "grad_norm": 0.45896485447883606,
      "learning_rate": 9.648148148148149e-05,
      "loss": 3.7629,
      "step": 2480
    },
    {
      "epoch": 0.7535227187099691,
      "grad_norm": 0.39767026901245117,
      "learning_rate": 9.277777777777778e-05,
      "loss": 3.763,
      "step": 2500
    },
    {
      "epoch": 0.7595509004596489,
      "grad_norm": 0.4772471785545349,
      "learning_rate": 8.907407407407407e-05,
      "loss": 3.7621,
      "step": 2520
    },
    {
      "epoch": 0.7655790822093286,
      "grad_norm": 0.4539508819580078,
      "learning_rate": 8.537037037037038e-05,
      "loss": 3.7618,
      "step": 2540
    },
    {
      "epoch": 0.7716072639590084,
      "grad_norm": 0.44643184542655945,
      "learning_rate": 8.166666666666667e-05,
      "loss": 3.7562,
      "step": 2560
    },
    {
      "epoch": 0.7776354457086881,
      "grad_norm": 0.4100569486618042,
      "learning_rate": 7.796296296296296e-05,
      "loss": 3.7546,
      "step": 2580
    },
    {
      "epoch": 0.7836636274583678,
      "grad_norm": 0.40634608268737793,
      "learning_rate": 7.425925925925927e-05,
      "loss": 3.7482,
      "step": 2600
    },
    {
      "epoch": 0.7896918092080476,
      "grad_norm": 0.4744322896003723,
      "learning_rate": 7.055555555555556e-05,
      "loss": 3.7542,
      "step": 2620
    },
    {
      "epoch": 0.7957199909577274,
      "grad_norm": 0.4235599935054779,
      "learning_rate": 6.685185185185186e-05,
      "loss": 3.7514,
      "step": 2640
    },
    {
      "epoch": 0.8017481727074072,
      "grad_norm": 0.4705050587654114,
      "learning_rate": 6.314814814814815e-05,
      "loss": 3.7511,
      "step": 2660
    },
    {
      "epoch": 0.8077763544570868,
      "grad_norm": 0.4208191931247711,
      "learning_rate": 5.9444444444444445e-05,
      "loss": 3.7434,
      "step": 2680
    },
    {
      "epoch": 0.8138045362067666,
      "grad_norm": 0.39097142219543457,
      "learning_rate": 5.5740740740740744e-05,
      "loss": 3.7461,
      "step": 2700
    },
    {
      "epoch": 0.8138045362067666,
      "eval_loss": 3.6990146794020826,
      "eval_perplexity": 40.40747042577191,
      "eval_runtime": 112.1899,
      "eval_samples_per_second": 14.787,
      "eval_steps_per_second": 0.232,
      "step": 2700
    },
    {
      "epoch": 0.8198327179564464,
      "grad_norm": 0.35043832659721375,
      "learning_rate": 5.2037037037037035e-05,
      "loss": 3.7433,
      "step": 2720
    },
    {
      "epoch": 0.8258608997061262,
      "grad_norm": 0.41040167212486267,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 3.7457,
      "step": 2740
    },
    {
      "epoch": 0.8318890814558059,
      "grad_norm": 0.38235458731651306,
      "learning_rate": 4.462962962962963e-05,
      "loss": 3.7413,
      "step": 2760
    },
    {
      "epoch": 0.8379172632054857,
      "grad_norm": 0.3330211937427521,
      "learning_rate": 4.092592592592593e-05,
      "loss": 3.7393,
      "step": 2780
    },
    {
      "epoch": 0.8439454449551654,
      "grad_norm": 0.3779599070549011,
      "learning_rate": 3.722222222222222e-05,
      "loss": 3.7412,
      "step": 2800
    },
    {
      "epoch": 0.8499736267048451,
      "grad_norm": 0.34843119978904724,
      "learning_rate": 3.351851851851852e-05,
      "loss": 3.7431,
      "step": 2820
    },
    {
      "epoch": 0.8560018084545249,
      "grad_norm": 0.3336455821990967,
      "learning_rate": 2.9814814814814815e-05,
      "loss": 3.7355,
      "step": 2840
    },
    {
      "epoch": 0.8620299902042047,
      "grad_norm": 0.3009018301963806,
      "learning_rate": 2.6111111111111114e-05,
      "loss": 3.7327,
      "step": 2860
    },
    {
      "epoch": 0.8680581719538845,
      "grad_norm": 0.31362465023994446,
      "learning_rate": 2.240740740740741e-05,
      "loss": 3.7366,
      "step": 2880
    },
    {
      "epoch": 0.8740863537035641,
      "grad_norm": 0.3083265423774719,
      "learning_rate": 1.8703703703703707e-05,
      "loss": 3.7313,
      "step": 2900
    },
    {
      "epoch": 0.8801145354532439,
      "grad_norm": 0.286391019821167,
      "learning_rate": 1.5e-05,
      "loss": 3.736,
      "step": 2920
    },
    {
      "epoch": 0.8861427172029237,
      "grad_norm": 0.27902212738990784,
      "learning_rate": 1.1296296296296295e-05,
      "loss": 3.7267,
      "step": 2940
    },
    {
      "epoch": 0.8921708989526034,
      "grad_norm": 0.27833643555641174,
      "learning_rate": 7.592592592592593e-06,
      "loss": 3.723,
      "step": 2960
    },
    {
      "epoch": 0.8981990807022832,
      "grad_norm": 0.25484341382980347,
      "learning_rate": 3.888888888888889e-06,
      "loss": 3.7308,
      "step": 2980
    },
    {
      "epoch": 0.9042272624519629,
      "grad_norm": 0.2665034532546997,
      "learning_rate": 1.8518518518518518e-07,
      "loss": 3.7341,
      "step": 3000
    },
    {
      "epoch": 0.9042272624519629,
      "eval_loss": 3.6842086777328205,
      "eval_perplexity": 39.813604589265466,
      "eval_runtime": 112.1577,
      "eval_samples_per_second": 14.792,
      "eval_steps_per_second": 0.232,
      "step": 3000
    }
  ],
  "logging_steps": 20,
  "max_steps": 3000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 300,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.320737071693824e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}