{
  "best_global_step": 3000,
  "best_metric": 3.6086546316869916,
  "best_model_checkpoint": "checkpoints/gpt-2_seq1024_mla0-0-192/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": 2.0655815601348877,
      "learning_rate": 3.166666666666667e-05,
      "loss": 10.3751,
      "step": 20
    },
    {
      "epoch": 0.012056363499359506,
      "grad_norm": 1.6421316862106323,
      "learning_rate": 6.500000000000001e-05,
      "loss": 9.2194,
      "step": 40
    },
    {
      "epoch": 0.01808454524903926,
      "grad_norm": 0.7634080052375793,
      "learning_rate": 9.833333333333333e-05,
      "loss": 8.1618,
      "step": 60
    },
    {
      "epoch": 0.02411272699871901,
      "grad_norm": 0.5637308359146118,
      "learning_rate": 0.00013166666666666665,
      "loss": 7.4258,
      "step": 80
    },
    {
      "epoch": 0.030140908748398764,
      "grad_norm": 1.269168734550476,
      "learning_rate": 0.000165,
      "loss": 7.0598,
      "step": 100
    },
    {
      "epoch": 0.03616909049807852,
      "grad_norm": 0.9437507390975952,
      "learning_rate": 0.00019833333333333335,
      "loss": 6.7467,
      "step": 120
    },
    {
      "epoch": 0.04219727224775827,
      "grad_norm": 1.3413256406784058,
      "learning_rate": 0.00023166666666666667,
      "loss": 6.5024,
      "step": 140
    },
    {
      "epoch": 0.04822545399743802,
      "grad_norm": 0.8728963136672974,
      "learning_rate": 0.00026500000000000004,
      "loss": 6.318,
      "step": 160
    },
    {
      "epoch": 0.05425363574711778,
      "grad_norm": 0.7673031091690063,
      "learning_rate": 0.00029833333333333334,
      "loss": 6.1662,
      "step": 180
    },
    {
      "epoch": 0.06028181749679753,
      "grad_norm": 0.8941771388053894,
      "learning_rate": 0.0003316666666666667,
      "loss": 6.034,
      "step": 200
    },
    {
      "epoch": 0.06630999924647728,
      "grad_norm": 0.8988394737243652,
      "learning_rate": 0.000365,
      "loss": 5.8946,
      "step": 220
    },
    {
      "epoch": 0.07233818099615703,
      "grad_norm": 1.3544883728027344,
      "learning_rate": 0.00039833333333333333,
      "loss": 5.7744,
      "step": 240
    },
    {
      "epoch": 0.07836636274583679,
      "grad_norm": 0.7703800201416016,
      "learning_rate": 0.0004316666666666667,
      "loss": 5.6709,
      "step": 260
    },
    {
      "epoch": 0.08439454449551655,
      "grad_norm": 1.1792443990707397,
      "learning_rate": 0.000465,
      "loss": 5.5611,
      "step": 280
    },
    {
      "epoch": 0.09042272624519629,
      "grad_norm": 0.6350497007369995,
      "learning_rate": 0.0004983333333333334,
      "loss": 5.4642,
      "step": 300
    },
    {
      "epoch": 0.09042272624519629,
      "eval_loss": 5.3608598637309335,
      "eval_perplexity": 212.90793957896608,
      "eval_runtime": 112.3491,
      "eval_samples_per_second": 14.766,
      "eval_steps_per_second": 0.231,
      "step": 300
    },
    {
      "epoch": 0.09645090799487605,
      "grad_norm": 0.9472798109054565,
      "learning_rate": 0.0004964814814814814,
      "loss": 5.3503,
      "step": 320
    },
    {
      "epoch": 0.1024790897445558,
      "grad_norm": 0.5043119788169861,
      "learning_rate": 0.0004927777777777777,
      "loss": 5.2791,
      "step": 340
    },
    {
      "epoch": 0.10850727149423556,
      "grad_norm": 0.515604555606842,
      "learning_rate": 0.0004890740740740741,
      "loss": 5.166,
      "step": 360
    },
    {
      "epoch": 0.1145354532439153,
      "grad_norm": 1.0375034809112549,
      "learning_rate": 0.0004853703703703704,
      "loss": 5.0872,
      "step": 380
    },
    {
      "epoch": 0.12056363499359506,
      "grad_norm": 0.5367968678474426,
      "learning_rate": 0.0004816666666666667,
      "loss": 5.0222,
      "step": 400
    },
    {
      "epoch": 0.1265918167432748,
      "grad_norm": 0.598520040512085,
      "learning_rate": 0.00047796296296296297,
      "loss": 4.9135,
      "step": 420
    },
    {
      "epoch": 0.13261999849295456,
      "grad_norm": 1.0570532083511353,
      "learning_rate": 0.0004742592592592593,
      "loss": 4.8567,
      "step": 440
    },
    {
      "epoch": 0.1386481802426343,
      "grad_norm": 0.6489225625991821,
      "learning_rate": 0.00047055555555555555,
      "loss": 4.799,
      "step": 460
    },
    {
      "epoch": 0.14467636199231407,
      "grad_norm": 0.825153112411499,
      "learning_rate": 0.00046685185185185187,
      "loss": 4.7057,
      "step": 480
    },
    {
      "epoch": 0.15070454374199382,
      "grad_norm": 0.8194488883018494,
      "learning_rate": 0.00046314814814814813,
      "loss": 4.6468,
      "step": 500
    },
    {
      "epoch": 0.15673272549167358,
      "grad_norm": 0.609926700592041,
      "learning_rate": 0.00045944444444444445,
      "loss": 4.5995,
      "step": 520
    },
    {
      "epoch": 0.16276090724135334,
      "grad_norm": 0.9008710384368896,
      "learning_rate": 0.0004557407407407407,
      "loss": 4.5535,
      "step": 540
    },
    {
      "epoch": 0.1687890889910331,
      "grad_norm": 0.7430111765861511,
      "learning_rate": 0.00045203703703703703,
      "loss": 4.5071,
      "step": 560
    },
    {
      "epoch": 0.17481727074071282,
      "grad_norm": 0.4787514805793762,
      "learning_rate": 0.0004483333333333333,
      "loss": 4.4607,
      "step": 580
    },
    {
      "epoch": 0.18084545249039258,
      "grad_norm": 0.7504919767379761,
      "learning_rate": 0.00044462962962962967,
      "loss": 4.4368,
      "step": 600
    },
    {
      "epoch": 0.18084545249039258,
      "eval_loss": 4.38678531582523,
      "eval_perplexity": 80.3816017324782,
      "eval_runtime": 112.1043,
      "eval_samples_per_second": 14.799,
      "eval_steps_per_second": 0.232,
      "step": 600
    },
    {
      "epoch": 0.18687363424007233,
      "grad_norm": 0.6413805484771729,
      "learning_rate": 0.00044092592592592594,
      "loss": 4.4005,
      "step": 620
    },
    {
      "epoch": 0.1929018159897521,
      "grad_norm": 0.5979523658752441,
      "learning_rate": 0.00043722222222222225,
      "loss": 4.37,
      "step": 640
    },
    {
      "epoch": 0.19892999773943185,
      "grad_norm": 0.4198634624481201,
      "learning_rate": 0.0004335185185185185,
      "loss": 4.3388,
      "step": 660
    },
    {
      "epoch": 0.2049581794891116,
      "grad_norm": 0.5570924282073975,
      "learning_rate": 0.00042981481481481484,
      "loss": 4.3121,
      "step": 680
    },
    {
      "epoch": 0.21098636123879136,
      "grad_norm": 0.8379751443862915,
      "learning_rate": 0.0004261111111111111,
      "loss": 4.2963,
      "step": 700
    },
    {
      "epoch": 0.21701454298847112,
      "grad_norm": 0.7025314569473267,
      "learning_rate": 0.0004224074074074074,
      "loss": 4.2653,
      "step": 720
    },
    {
      "epoch": 0.22304272473815084,
      "grad_norm": 0.43065857887268066,
      "learning_rate": 0.0004187037037037037,
      "loss": 4.236,
      "step": 740
    },
    {
      "epoch": 0.2290709064878306,
      "grad_norm": 0.539214015007019,
      "learning_rate": 0.000415,
      "loss": 4.2176,
      "step": 760
    },
    {
      "epoch": 0.23509908823751036,
      "grad_norm": 0.696628749370575,
      "learning_rate": 0.00041129629629629627,
      "loss": 4.2025,
      "step": 780
    },
    {
      "epoch": 0.2411272699871901,
      "grad_norm": 0.75990891456604,
      "learning_rate": 0.00040759259259259264,
      "loss": 4.1898,
      "step": 800
    },
    {
      "epoch": 0.24715545173686987,
      "grad_norm": 0.48626840114593506,
      "learning_rate": 0.0004038888888888889,
      "loss": 4.1687,
      "step": 820
    },
    {
      "epoch": 0.2531836334865496,
      "grad_norm": 0.7186273336410522,
      "learning_rate": 0.0004001851851851852,
      "loss": 4.1505,
      "step": 840
    },
    {
      "epoch": 0.2592118152362294,
      "grad_norm": 0.45199310779571533,
      "learning_rate": 0.0003964814814814815,
      "loss": 4.1405,
      "step": 860
    },
    {
      "epoch": 0.2652399969859091,
      "grad_norm": 0.3754367530345917,
      "learning_rate": 0.0003927777777777778,
      "loss": 4.1194,
      "step": 880
    },
    {
      "epoch": 0.2712681787355889,
      "grad_norm": 0.7503505349159241,
      "learning_rate": 0.00038907407407407407,
      "loss": 4.1047,
      "step": 900
    },
    {
      "epoch": 0.2712681787355889,
      "eval_loss": 4.062260297662503,
      "eval_perplexity": 58.10549848421657,
      "eval_runtime": 112.298,
      "eval_samples_per_second": 14.773,
      "eval_steps_per_second": 0.232,
      "step": 900
    },
    {
      "epoch": 0.2772963604852686,
      "grad_norm": 0.4853602349758148,
      "learning_rate": 0.0003853703703703704,
      "loss": 4.0974,
      "step": 920
    },
    {
      "epoch": 0.2833245422349484,
      "grad_norm": 0.4983716905117035,
      "learning_rate": 0.00038166666666666666,
      "loss": 4.0792,
      "step": 940
    },
    {
      "epoch": 0.28935272398462814,
      "grad_norm": 0.5486569404602051,
      "learning_rate": 0.000377962962962963,
      "loss": 4.0617,
      "step": 960
    },
    {
      "epoch": 0.29538090573430786,
      "grad_norm": 0.6470081210136414,
      "learning_rate": 0.00037425925925925924,
      "loss": 4.0604,
      "step": 980
    },
    {
      "epoch": 0.30140908748398765,
      "grad_norm": 0.5754533410072327,
      "learning_rate": 0.0003705555555555556,
      "loss": 4.0451,
      "step": 1000
    },
    {
      "epoch": 0.3074372692336674,
      "grad_norm": 0.5240389108657837,
      "learning_rate": 0.0003668518518518519,
      "loss": 4.0378,
      "step": 1020
    },
    {
      "epoch": 0.31346545098334716,
      "grad_norm": 0.6442040801048279,
      "learning_rate": 0.0003631481481481482,
      "loss": 4.0167,
      "step": 1040
    },
    {
      "epoch": 0.3194936327330269,
      "grad_norm": 0.6846771836280823,
      "learning_rate": 0.00035944444444444446,
      "loss": 4.0201,
      "step": 1060
    },
    {
      "epoch": 0.3255218144827067,
      "grad_norm": 0.5063695311546326,
      "learning_rate": 0.0003557407407407408,
      "loss": 4.0032,
      "step": 1080
    },
    {
      "epoch": 0.3315499962323864,
      "grad_norm": 0.42528533935546875,
      "learning_rate": 0.00035203703703703704,
      "loss": 3.9892,
      "step": 1100
    },
    {
      "epoch": 0.3375781779820662,
      "grad_norm": 0.49422556161880493,
      "learning_rate": 0.00034833333333333336,
      "loss": 3.9772,
      "step": 1120
    },
    {
      "epoch": 0.3436063597317459,
      "grad_norm": 0.549915611743927,
      "learning_rate": 0.0003446296296296296,
      "loss": 3.9732,
      "step": 1140
    },
    {
      "epoch": 0.34963454148142564,
      "grad_norm": 0.47555869817733765,
      "learning_rate": 0.0003409259259259259,
      "loss": 3.9578,
      "step": 1160
    },
    {
      "epoch": 0.3556627232311054,
      "grad_norm": 0.6645114421844482,
      "learning_rate": 0.0003372222222222222,
      "loss": 3.9525,
      "step": 1180
    },
    {
      "epoch": 0.36169090498078516,
      "grad_norm": 0.46186399459838867,
      "learning_rate": 0.0003335185185185185,
      "loss": 3.955,
      "step": 1200
    },
    {
      "epoch": 0.36169090498078516,
      "eval_loss": 3.899267339070575,
      "eval_perplexity": 49.36626711743568,
      "eval_runtime": 112.2672,
      "eval_samples_per_second": 14.777,
      "eval_steps_per_second": 0.232,
      "step": 1200
    },
    {
      "epoch": 0.36771908673046494,
      "grad_norm": 0.41637852787971497,
      "learning_rate": 0.00032981481481481485,
      "loss": 3.9353,
      "step": 1220
    },
    {
      "epoch": 0.37374726848014467,
      "grad_norm": 0.743456244468689,
      "learning_rate": 0.0003261111111111111,
      "loss": 3.9322,
      "step": 1240
    },
    {
      "epoch": 0.37977545022982445,
      "grad_norm": 0.6069373488426208,
      "learning_rate": 0.00032240740740740743,
      "loss": 3.9222,
      "step": 1260
    },
    {
      "epoch": 0.3858036319795042,
      "grad_norm": 0.42226338386535645,
      "learning_rate": 0.0003187037037037037,
      "loss": 3.9151,
      "step": 1280
    },
    {
      "epoch": 0.3918318137291839,
      "grad_norm": 0.5269602537155151,
      "learning_rate": 0.000315,
      "loss": 3.9054,
      "step": 1300
    },
    {
      "epoch": 0.3978599954788637,
      "grad_norm": 0.49215298891067505,
      "learning_rate": 0.0003112962962962963,
      "loss": 3.9038,
      "step": 1320
    },
    {
      "epoch": 0.4038881772285434,
      "grad_norm": 0.497126042842865,
      "learning_rate": 0.0003075925925925926,
      "loss": 3.8952,
      "step": 1340
    },
    {
      "epoch": 0.4099163589782232,
      "grad_norm": 0.47206273674964905,
      "learning_rate": 0.00030388888888888886,
      "loss": 3.8891,
      "step": 1360
    },
    {
      "epoch": 0.41594454072790293,
      "grad_norm": 0.5514901876449585,
      "learning_rate": 0.0003001851851851852,
      "loss": 3.8832,
      "step": 1380
    },
    {
      "epoch": 0.4219727224775827,
      "grad_norm": 0.3839927613735199,
      "learning_rate": 0.00029648148148148144,
      "loss": 3.8766,
      "step": 1400
    },
    {
      "epoch": 0.42800090422726245,
      "grad_norm": 0.5974825620651245,
      "learning_rate": 0.0002927777777777778,
      "loss": 3.876,
      "step": 1420
    },
    {
      "epoch": 0.43402908597694223,
      "grad_norm": 0.49415451288223267,
      "learning_rate": 0.0002890740740740741,
      "loss": 3.8614,
      "step": 1440
    },
    {
      "epoch": 0.44005726772662196,
      "grad_norm": 0.404469758272171,
      "learning_rate": 0.0002853703703703704,
      "loss": 3.8663,
      "step": 1460
    },
    {
      "epoch": 0.4460854494763017,
      "grad_norm": 0.43800851702690125,
      "learning_rate": 0.00028166666666666666,
      "loss": 3.8557,
      "step": 1480
    },
    {
      "epoch": 0.45211363122598147,
      "grad_norm": 0.48329687118530273,
      "learning_rate": 0.000277962962962963,
      "loss": 3.851,
      "step": 1500
    },
    {
      "epoch": 0.45211363122598147,
      "eval_loss": 3.8007436296406283,
      "eval_perplexity": 44.73443798166865,
      "eval_runtime": 112.0663,
      "eval_samples_per_second": 14.804,
      "eval_steps_per_second": 0.232,
      "step": 1500
    },
    {
      "epoch": 0.4581418129756612,
      "grad_norm": 0.5079576373100281,
      "learning_rate": 0.00027425925925925925,
      "loss": 3.8448,
      "step": 1520
    },
    {
      "epoch": 0.464169994725341,
      "grad_norm": 0.5414606928825378,
      "learning_rate": 0.00027055555555555557,
      "loss": 3.841,
      "step": 1540
    },
    {
      "epoch": 0.4701981764750207,
      "grad_norm": 0.4043631851673126,
      "learning_rate": 0.00026685185185185183,
      "loss": 3.831,
      "step": 1560
    },
    {
      "epoch": 0.4762263582247005,
      "grad_norm": 0.6338221430778503,
      "learning_rate": 0.00026314814814814815,
      "loss": 3.8272,
      "step": 1580
    },
    {
      "epoch": 0.4822545399743802,
      "grad_norm": 0.47206878662109375,
      "learning_rate": 0.0002594444444444444,
      "loss": 3.8157,
      "step": 1600
    },
    {
      "epoch": 0.48828272172405995,
      "grad_norm": 0.46456432342529297,
      "learning_rate": 0.0002557407407407408,
      "loss": 3.8237,
      "step": 1620
    },
    {
      "epoch": 0.49431090347373974,
      "grad_norm": 0.6102868914604187,
      "learning_rate": 0.00025203703703703705,
      "loss": 3.8143,
      "step": 1640
    },
    {
      "epoch": 0.5003390852234195,
      "grad_norm": 0.48834484815597534,
      "learning_rate": 0.0002483333333333333,
      "loss": 3.8111,
      "step": 1660
    },
    {
      "epoch": 0.5063672669730992,
      "grad_norm": 0.4487571120262146,
      "learning_rate": 0.00024462962962962963,
      "loss": 3.8086,
      "step": 1680
    },
    {
      "epoch": 0.512395448722779,
      "grad_norm": 0.42693281173706055,
      "learning_rate": 0.00024092592592592593,
      "loss": 3.7953,
      "step": 1700
    },
    {
      "epoch": 0.5184236304724588,
      "grad_norm": 0.5261042714118958,
      "learning_rate": 0.00023722222222222222,
      "loss": 3.7971,
      "step": 1720
    },
    {
      "epoch": 0.5244518122221385,
      "grad_norm": 0.4606678783893585,
      "learning_rate": 0.0002335185185185185,
      "loss": 3.7957,
      "step": 1740
    },
    {
      "epoch": 0.5304799939718182,
      "grad_norm": 0.43703415989875793,
      "learning_rate": 0.0002298148148148148,
      "loss": 3.7935,
      "step": 1760
    },
    {
      "epoch": 0.536508175721498,
      "grad_norm": 0.5641158223152161,
      "learning_rate": 0.00022611111111111112,
      "loss": 3.7821,
      "step": 1780
    },
    {
      "epoch": 0.5425363574711778,
      "grad_norm": 0.4422275424003601,
      "learning_rate": 0.0002224074074074074,
      "loss": 3.7857,
      "step": 1800
    },
    {
      "epoch": 0.5425363574711778,
      "eval_loss": 3.7325062866163825,
      "eval_perplexity": 41.783698967520365,
      "eval_runtime": 112.2687,
      "eval_samples_per_second": 14.777,
      "eval_steps_per_second": 0.232,
      "step": 1800
    },
    {
      "epoch": 0.5485645392208575,
      "grad_norm": 0.4784490764141083,
      "learning_rate": 0.0002187037037037037,
      "loss": 3.7825,
      "step": 1820
    },
    {
      "epoch": 0.5545927209705372,
      "grad_norm": 0.403177946805954,
      "learning_rate": 0.000215,
      "loss": 3.7736,
      "step": 1840
    },
    {
      "epoch": 0.560620902720217,
      "grad_norm": 0.4561764597892761,
      "learning_rate": 0.00021129629629629629,
      "loss": 3.7674,
      "step": 1860
    },
    {
      "epoch": 0.5666490844698968,
      "grad_norm": 0.44008079171180725,
      "learning_rate": 0.0002075925925925926,
      "loss": 3.7669,
      "step": 1880
    },
    {
      "epoch": 0.5726772662195765,
      "grad_norm": 0.4967263638973236,
      "learning_rate": 0.0002038888888888889,
      "loss": 3.7591,
      "step": 1900
    },
    {
      "epoch": 0.5787054479692563,
      "grad_norm": 0.47853296995162964,
      "learning_rate": 0.0002001851851851852,
      "loss": 3.7614,
      "step": 1920
    },
    {
      "epoch": 0.584733629718936,
      "grad_norm": 0.39971593022346497,
      "learning_rate": 0.00019648148148148148,
      "loss": 3.7556,
      "step": 1940
    },
    {
      "epoch": 0.5907618114686157,
      "grad_norm": 0.37392178177833557,
      "learning_rate": 0.00019277777777777777,
      "loss": 3.7571,
      "step": 1960
    },
    {
      "epoch": 0.5967899932182955,
      "grad_norm": 0.45188137888908386,
      "learning_rate": 0.00018907407407407406,
      "loss": 3.7449,
      "step": 1980
    },
    {
      "epoch": 0.6028181749679753,
      "grad_norm": 0.42207691073417664,
      "learning_rate": 0.00018537037037037038,
      "loss": 3.7469,
      "step": 2000
    },
    {
      "epoch": 0.6088463567176551,
      "grad_norm": 0.3993569314479828,
      "learning_rate": 0.00018166666666666667,
      "loss": 3.7428,
      "step": 2020
    },
    {
      "epoch": 0.6148745384673348,
      "grad_norm": 0.4090699255466461,
      "learning_rate": 0.00017796296296296296,
      "loss": 3.7373,
      "step": 2040
    },
    {
      "epoch": 0.6209027202170145,
      "grad_norm": 0.4046180248260498,
      "learning_rate": 0.00017425925925925926,
      "loss": 3.7382,
      "step": 2060
    },
    {
      "epoch": 0.6269309019666943,
      "grad_norm": 0.38884371519088745,
      "learning_rate": 0.00017055555555555555,
      "loss": 3.7359,
      "step": 2080
    },
    {
      "epoch": 0.632959083716374,
      "grad_norm": 0.36972570419311523,
      "learning_rate": 0.00016685185185185187,
      "loss": 3.7313,
      "step": 2100
    },
    {
      "epoch": 0.632959083716374,
      "eval_loss": 3.6836323830824136,
      "eval_perplexity": 39.79066683201549,
      "eval_runtime": 112.1491,
      "eval_samples_per_second": 14.793,
      "eval_steps_per_second": 0.232,
      "step": 2100
    },
    {
      "epoch": 0.6389872654660538,
      "grad_norm": 0.3997746407985687,
      "learning_rate": 0.00016314814814814816,
      "loss": 3.7302,
      "step": 2120
    },
    {
      "epoch": 0.6450154472157336,
      "grad_norm": 0.45008304715156555,
      "learning_rate": 0.00015944444444444445,
      "loss": 3.7287,
      "step": 2140
    },
    {
      "epoch": 0.6510436289654133,
      "grad_norm": 0.3761305510997772,
      "learning_rate": 0.00015574074074074074,
      "loss": 3.7264,
      "step": 2160
    },
    {
      "epoch": 0.657071810715093,
      "grad_norm": 0.3815375566482544,
      "learning_rate": 0.00015203703703703703,
      "loss": 3.7249,
      "step": 2180
    },
    {
      "epoch": 0.6630999924647728,
      "grad_norm": 0.3969159722328186,
      "learning_rate": 0.00014833333333333335,
      "loss": 3.7203,
      "step": 2200
    },
    {
      "epoch": 0.6691281742144526,
      "grad_norm": 0.4421869218349457,
      "learning_rate": 0.00014462962962962964,
      "loss": 3.7182,
      "step": 2220
    },
    {
      "epoch": 0.6751563559641324,
      "grad_norm": 0.413221538066864,
      "learning_rate": 0.00014092592592592594,
      "loss": 3.7178,
      "step": 2240
    },
    {
      "epoch": 0.681184537713812,
      "grad_norm": 0.32278236746788025,
      "learning_rate": 0.00013722222222222223,
      "loss": 3.7043,
      "step": 2260
    },
    {
      "epoch": 0.6872127194634918,
      "grad_norm": 0.429057240486145,
      "learning_rate": 0.00013351851851851852,
      "loss": 3.7116,
      "step": 2280
    },
    {
      "epoch": 0.6932409012131716,
      "grad_norm": 0.37107640504837036,
      "learning_rate": 0.00012981481481481484,
      "loss": 3.7045,
      "step": 2300
    },
    {
      "epoch": 0.6992690829628513,
      "grad_norm": 0.3797242045402527,
      "learning_rate": 0.00012611111111111113,
      "loss": 3.6983,
      "step": 2320
    },
    {
      "epoch": 0.7052972647125311,
      "grad_norm": 0.3916177451610565,
      "learning_rate": 0.00012240740740740742,
      "loss": 3.7001,
      "step": 2340
    },
    {
      "epoch": 0.7113254464622109,
      "grad_norm": 0.3719088137149811,
      "learning_rate": 0.0001187037037037037,
      "loss": 3.6997,
      "step": 2360
    },
    {
      "epoch": 0.7173536282118906,
      "grad_norm": 0.38734784722328186,
      "learning_rate": 0.000115,
      "loss": 3.6908,
      "step": 2380
    },
    {
      "epoch": 0.7233818099615703,
      "grad_norm": 0.3510124385356903,
      "learning_rate": 0.0001112962962962963,
      "loss": 3.699,
      "step": 2400
    },
    {
      "epoch": 0.7233818099615703,
      "eval_loss": 3.647647414912115,
      "eval_perplexity": 38.38425751201017,
      "eval_runtime": 112.268,
      "eval_samples_per_second": 14.777,
      "eval_steps_per_second": 0.232,
      "step": 2400
    },
    {
      "epoch": 0.7294099917112501,
      "grad_norm": 0.36017531156539917,
      "learning_rate": 0.00010759259259259259,
      "loss": 3.6951,
      "step": 2420
    },
    {
      "epoch": 0.7354381734609299,
      "grad_norm": 0.3744637072086334,
      "learning_rate": 0.00010388888888888889,
      "loss": 3.6927,
      "step": 2440
    },
    {
      "epoch": 0.7414663552106096,
      "grad_norm": 0.38499677181243896,
      "learning_rate": 0.00010018518518518518,
      "loss": 3.6932,
      "step": 2460
    },
    {
      "epoch": 0.7474945369602893,
      "grad_norm": 0.3262474834918976,
      "learning_rate": 9.648148148148149e-05,
      "loss": 3.686,
      "step": 2480
    },
    {
      "epoch": 0.7535227187099691,
      "grad_norm": 0.332756906747818,
      "learning_rate": 9.277777777777778e-05,
      "loss": 3.6867,
      "step": 2500
    },
    {
      "epoch": 0.7595509004596489,
      "grad_norm": 0.32537171244621277,
      "learning_rate": 8.907407407407407e-05,
      "loss": 3.6851,
      "step": 2520
    },
    {
      "epoch": 0.7655790822093286,
      "grad_norm": 0.3477420508861542,
      "learning_rate": 8.537037037037038e-05,
      "loss": 3.685,
      "step": 2540
    },
    {
      "epoch": 0.7716072639590084,
      "grad_norm": 0.3186497390270233,
      "learning_rate": 8.166666666666667e-05,
      "loss": 3.68,
      "step": 2560
    },
    {
      "epoch": 0.7776354457086881,
      "grad_norm": 0.31295379996299744,
      "learning_rate": 7.796296296296296e-05,
      "loss": 3.6789,
      "step": 2580
    },
    {
      "epoch": 0.7836636274583678,
      "grad_norm": 0.3082525432109833,
      "learning_rate": 7.425925925925927e-05,
      "loss": 3.6722,
      "step": 2600
    },
    {
      "epoch": 0.7896918092080476,
      "grad_norm": 0.34101417660713196,
      "learning_rate": 7.055555555555556e-05,
      "loss": 3.6786,
      "step": 2620
    },
    {
      "epoch": 0.7957199909577274,
      "grad_norm": 0.3168187141418457,
      "learning_rate": 6.685185185185186e-05,
      "loss": 3.6757,
      "step": 2640
    },
    {
      "epoch": 0.8017481727074072,
      "grad_norm": 0.3118196129798889,
      "learning_rate": 6.314814814814815e-05,
      "loss": 3.6753,
      "step": 2660
    },
    {
      "epoch": 0.8077763544570868,
      "grad_norm": 0.2771850526332855,
      "learning_rate": 5.9444444444444445e-05,
      "loss": 3.6678,
      "step": 2680
    },
    {
      "epoch": 0.8138045362067666,
      "grad_norm": 0.2887312173843384,
      "learning_rate": 5.5740740740740744e-05,
      "loss": 3.6706,
      "step": 2700
    },
    {
      "epoch": 0.8138045362067666,
      "eval_loss": 3.6225326291998914,
      "eval_perplexity": 37.432249883239926,
      "eval_runtime": 112.1422,
      "eval_samples_per_second": 14.794,
      "eval_steps_per_second": 0.232,
      "step": 2700
    },
    {
      "epoch": 0.8198327179564464,
      "grad_norm": 0.2720380127429962,
      "learning_rate": 5.2037037037037035e-05,
      "loss": 3.6683,
      "step": 2720
    },
    {
      "epoch": 0.8258608997061262,
      "grad_norm": 0.3013540506362915,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 3.6709,
      "step": 2740
    },
    {
      "epoch": 0.8318890814558059,
      "grad_norm": 0.2793949246406555,
      "learning_rate": 4.462962962962963e-05,
      "loss": 3.6664,
      "step": 2760
    },
    {
      "epoch": 0.8379172632054857,
      "grad_norm": 0.25227662920951843,
      "learning_rate": 4.092592592592593e-05,
      "loss": 3.6646,
      "step": 2780
    },
    {
      "epoch": 0.8439454449551654,
      "grad_norm": 0.2647295296192169,
      "learning_rate": 3.722222222222222e-05,
      "loss": 3.6664,
      "step": 2800
    },
    {
      "epoch": 0.8499736267048451,
      "grad_norm": 0.2721077799797058,
      "learning_rate": 3.351851851851852e-05,
      "loss": 3.6689,
      "step": 2820
    },
    {
      "epoch": 0.8560018084545249,
      "grad_norm": 0.2462477684020996,
      "learning_rate": 2.9814814814814815e-05,
      "loss": 3.6609,
      "step": 2840
    },
    {
      "epoch": 0.8620299902042047,
      "grad_norm": 0.23483806848526,
      "learning_rate": 2.6111111111111114e-05,
      "loss": 3.6586,
      "step": 2860
    },
    {
      "epoch": 0.8680581719538845,
      "grad_norm": 0.24066658318042755,
      "learning_rate": 2.240740740740741e-05,
      "loss": 3.6624,
      "step": 2880
    },
    {
      "epoch": 0.8740863537035641,
      "grad_norm": 0.2355494350194931,
      "learning_rate": 1.8703703703703707e-05,
      "loss": 3.6576,
      "step": 2900
    },
    {
      "epoch": 0.8801145354532439,
      "grad_norm": 0.23151390254497528,
      "learning_rate": 1.5e-05,
      "loss": 3.6615,
      "step": 2920
    },
    {
      "epoch": 0.8861427172029237,
      "grad_norm": 0.21725036203861237,
      "learning_rate": 1.1296296296296295e-05,
      "loss": 3.6526,
      "step": 2940
    },
    {
      "epoch": 0.8921708989526034,
      "grad_norm": 0.21774446964263916,
      "learning_rate": 7.592592592592593e-06,
      "loss": 3.6491,
      "step": 2960
    },
    {
      "epoch": 0.8981990807022832,
      "grad_norm": 0.19765916466712952,
      "learning_rate": 3.888888888888889e-06,
      "loss": 3.6564,
      "step": 2980
    },
    {
      "epoch": 0.9042272624519629,
      "grad_norm": 0.20228278636932373,
      "learning_rate": 1.8518518518518518e-07,
      "loss": 3.6603,
      "step": 3000
    },
    {
      "epoch": 0.9042272624519629,
      "eval_loss": 3.6086546316869916,
      "eval_perplexity": 36.916353298282225,
      "eval_runtime": 112.2545,
      "eval_samples_per_second": 14.779,
      "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": 1.537800155430912e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}