{
  "best_global_step": 3000,
  "best_metric": 3.6425010418158132,
  "best_model_checkpoint": "checkpoints/gpt-2_seq1024_mla192-96-0/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": 2.1480085849761963,
      "learning_rate": 3.166666666666667e-05,
      "loss": 10.5105,
      "step": 20
    },
    {
      "epoch": 0.012056363499359506,
      "grad_norm": 1.516770839691162,
      "learning_rate": 6.500000000000001e-05,
      "loss": 9.1318,
      "step": 40
    },
    {
      "epoch": 0.01808454524903926,
      "grad_norm": 0.9571565985679626,
      "learning_rate": 9.833333333333333e-05,
      "loss": 8.0347,
      "step": 60
    },
    {
      "epoch": 0.02411272699871901,
      "grad_norm": 0.6516666412353516,
      "learning_rate": 0.00013166666666666665,
      "loss": 7.227,
      "step": 80
    },
    {
      "epoch": 0.030140908748398764,
      "grad_norm": 0.42714038491249084,
      "learning_rate": 0.000165,
      "loss": 6.8261,
      "step": 100
    },
    {
      "epoch": 0.03616909049807852,
      "grad_norm": 0.6521334052085876,
      "learning_rate": 0.00019833333333333335,
      "loss": 6.5542,
      "step": 120
    },
    {
      "epoch": 0.04219727224775827,
      "grad_norm": 0.7042223215103149,
      "learning_rate": 0.00023166666666666667,
      "loss": 6.3763,
      "step": 140
    },
    {
      "epoch": 0.04822545399743802,
      "grad_norm": 0.6393540501594543,
      "learning_rate": 0.00026500000000000004,
      "loss": 6.2425,
      "step": 160
    },
    {
      "epoch": 0.05425363574711778,
      "grad_norm": 0.6561896204948425,
      "learning_rate": 0.00029833333333333334,
      "loss": 6.1282,
      "step": 180
    },
    {
      "epoch": 0.06028181749679753,
      "grad_norm": 1.1607221364974976,
      "learning_rate": 0.0003316666666666667,
      "loss": 5.9763,
      "step": 200
    },
    {
      "epoch": 0.06630999924647728,
      "grad_norm": 1.0097905397415161,
      "learning_rate": 0.000365,
      "loss": 5.8254,
      "step": 220
    },
    {
      "epoch": 0.07233818099615703,
      "grad_norm": 0.4747564494609833,
      "learning_rate": 0.00039833333333333333,
      "loss": 5.6952,
      "step": 240
    },
    {
      "epoch": 0.07836636274583679,
      "grad_norm": 0.6741572022438049,
      "learning_rate": 0.0004316666666666667,
      "loss": 5.561,
      "step": 260
    },
    {
      "epoch": 0.08439454449551655,
      "grad_norm": 0.8652853965759277,
      "learning_rate": 0.000465,
      "loss": 5.4164,
      "step": 280
    },
    {
      "epoch": 0.09042272624519629,
      "grad_norm": 0.29904869198799133,
      "learning_rate": 0.0004983333333333334,
      "loss": 5.3297,
      "step": 300
    },
    {
      "epoch": 0.09042272624519629,
      "eval_loss": 5.215795573862642,
      "eval_perplexity": 184.15827422429723,
      "eval_runtime": 112.8243,
      "eval_samples_per_second": 14.704,
      "eval_steps_per_second": 0.23,
      "step": 300
    },
    {
      "epoch": 0.09645090799487605,
      "grad_norm": 0.46344780921936035,
      "learning_rate": 0.0004964814814814814,
      "loss": 5.2213,
      "step": 320
    },
    {
      "epoch": 0.1024790897445558,
      "grad_norm": 0.4897235035896301,
      "learning_rate": 0.0004927777777777777,
      "loss": 5.1118,
      "step": 340
    },
    {
      "epoch": 0.10850727149423556,
      "grad_norm": 0.8371343612670898,
      "learning_rate": 0.0004890740740740741,
      "loss": 5.0311,
      "step": 360
    },
    {
      "epoch": 0.1145354532439153,
      "grad_norm": 0.40067142248153687,
      "learning_rate": 0.0004853703703703704,
      "loss": 4.9694,
      "step": 380
    },
    {
      "epoch": 0.12056363499359506,
      "grad_norm": 0.4551117718219757,
      "learning_rate": 0.0004816666666666667,
      "loss": 4.8826,
      "step": 400
    },
    {
      "epoch": 0.1265918167432748,
      "grad_norm": 0.5077718496322632,
      "learning_rate": 0.00047796296296296297,
      "loss": 4.8351,
      "step": 420
    },
    {
      "epoch": 0.13261999849295456,
      "grad_norm": 0.3099275529384613,
      "learning_rate": 0.0004742592592592593,
      "loss": 4.7924,
      "step": 440
    },
    {
      "epoch": 0.1386481802426343,
      "grad_norm": 0.6599226593971252,
      "learning_rate": 0.00047055555555555555,
      "loss": 4.7417,
      "step": 460
    },
    {
      "epoch": 0.14467636199231407,
      "grad_norm": 0.2840152680873871,
      "learning_rate": 0.00046685185185185187,
      "loss": 4.7107,
      "step": 480
    },
    {
      "epoch": 0.15070454374199382,
      "grad_norm": 0.3753263056278229,
      "learning_rate": 0.00046314814814814813,
      "loss": 4.6587,
      "step": 500
    },
    {
      "epoch": 0.15673272549167358,
      "grad_norm": 0.6300585865974426,
      "learning_rate": 0.00045944444444444445,
      "loss": 4.6299,
      "step": 520
    },
    {
      "epoch": 0.16276090724135334,
      "grad_norm": 0.35931190848350525,
      "learning_rate": 0.0004557407407407407,
      "loss": 4.5934,
      "step": 540
    },
    {
      "epoch": 0.1687890889910331,
      "grad_norm": 0.4792696535587311,
      "learning_rate": 0.00045203703703703703,
      "loss": 4.5573,
      "step": 560
    },
    {
      "epoch": 0.17481727074071282,
      "grad_norm": 0.3047443926334381,
      "learning_rate": 0.0004483333333333333,
      "loss": 4.5383,
      "step": 580
    },
    {
      "epoch": 0.18084545249039258,
      "grad_norm": 0.42478007078170776,
      "learning_rate": 0.00044462962962962967,
      "loss": 4.5029,
      "step": 600
    },
    {
      "epoch": 0.18084545249039258,
      "eval_loss": 4.455365447186089,
      "eval_perplexity": 86.0876055704807,
      "eval_runtime": 112.8601,
      "eval_samples_per_second": 14.7,
      "eval_steps_per_second": 0.23,
      "step": 600
    },
    {
      "epoch": 0.18687363424007233,
      "grad_norm": 0.5473248958587646,
      "learning_rate": 0.00044092592592592594,
      "loss": 4.473,
      "step": 620
    },
    {
      "epoch": 0.1929018159897521,
      "grad_norm": 0.2810865044593811,
      "learning_rate": 0.00043722222222222225,
      "loss": 4.452,
      "step": 640
    },
    {
      "epoch": 0.19892999773943185,
      "grad_norm": 0.3900161683559418,
      "learning_rate": 0.0004335185185185185,
      "loss": 4.4294,
      "step": 660
    },
    {
      "epoch": 0.2049581794891116,
      "grad_norm": 0.4204959273338318,
      "learning_rate": 0.00042981481481481484,
      "loss": 4.41,
      "step": 680
    },
    {
      "epoch": 0.21098636123879136,
      "grad_norm": 0.3920799791812897,
      "learning_rate": 0.0004261111111111111,
      "loss": 4.3853,
      "step": 700
    },
    {
      "epoch": 0.21701454298847112,
      "grad_norm": 0.5068658590316772,
      "learning_rate": 0.0004224074074074074,
      "loss": 4.3615,
      "step": 720
    },
    {
      "epoch": 0.22304272473815084,
      "grad_norm": 0.35288578271865845,
      "learning_rate": 0.0004187037037037037,
      "loss": 4.3377,
      "step": 740
    },
    {
      "epoch": 0.2290709064878306,
      "grad_norm": 0.3630707561969757,
      "learning_rate": 0.000415,
      "loss": 4.3168,
      "step": 760
    },
    {
      "epoch": 0.23509908823751036,
      "grad_norm": 0.36306232213974,
      "learning_rate": 0.00041129629629629627,
      "loss": 4.3039,
      "step": 780
    },
    {
      "epoch": 0.2411272699871901,
      "grad_norm": 0.3880031704902649,
      "learning_rate": 0.00040759259259259264,
      "loss": 4.2825,
      "step": 800
    },
    {
      "epoch": 0.24715545173686987,
      "grad_norm": 0.3516429662704468,
      "learning_rate": 0.0004038888888888889,
      "loss": 4.2604,
      "step": 820
    },
    {
      "epoch": 0.2531836334865496,
      "grad_norm": 0.4158916473388672,
      "learning_rate": 0.0004001851851851852,
      "loss": 4.2444,
      "step": 840
    },
    {
      "epoch": 0.2592118152362294,
      "grad_norm": 0.3786226511001587,
      "learning_rate": 0.0003964814814814815,
      "loss": 4.2234,
      "step": 860
    },
    {
      "epoch": 0.2652399969859091,
      "grad_norm": 0.39826563000679016,
      "learning_rate": 0.0003927777777777778,
      "loss": 4.2069,
      "step": 880
    },
    {
      "epoch": 0.2712681787355889,
      "grad_norm": 0.3681701123714447,
      "learning_rate": 0.00038907407407407407,
      "loss": 4.1836,
      "step": 900
    },
    {
      "epoch": 0.2712681787355889,
      "eval_loss": 4.133429888189484,
      "eval_perplexity": 62.3915524114605,
      "eval_runtime": 112.5243,
      "eval_samples_per_second": 14.743,
      "eval_steps_per_second": 0.231,
      "step": 900
    },
    {
      "epoch": 0.2772963604852686,
      "grad_norm": 0.5110707879066467,
      "learning_rate": 0.0003853703703703704,
      "loss": 4.1709,
      "step": 920
    },
    {
      "epoch": 0.2833245422349484,
      "grad_norm": 0.3312435746192932,
      "learning_rate": 0.00038166666666666666,
      "loss": 4.1599,
      "step": 940
    },
    {
      "epoch": 0.28935272398462814,
      "grad_norm": 0.3452323079109192,
      "learning_rate": 0.000377962962962963,
      "loss": 4.1342,
      "step": 960
    },
    {
      "epoch": 0.29538090573430786,
      "grad_norm": 0.3328576385974884,
      "learning_rate": 0.00037425925925925924,
      "loss": 4.1219,
      "step": 980
    },
    {
      "epoch": 0.30140908748398765,
      "grad_norm": 0.46105054020881653,
      "learning_rate": 0.0003705555555555556,
      "loss": 4.118,
      "step": 1000
    },
    {
      "epoch": 0.3074372692336674,
      "grad_norm": 0.4100669026374817,
      "learning_rate": 0.0003668518518518519,
      "loss": 4.0937,
      "step": 1020
    },
    {
      "epoch": 0.31346545098334716,
      "grad_norm": 0.45084893703460693,
      "learning_rate": 0.0003631481481481482,
      "loss": 4.0767,
      "step": 1040
    },
    {
      "epoch": 0.3194936327330269,
      "grad_norm": 0.3935777544975281,
      "learning_rate": 0.00035944444444444446,
      "loss": 4.0719,
      "step": 1060
    },
    {
      "epoch": 0.3255218144827067,
      "grad_norm": 0.3900182545185089,
      "learning_rate": 0.0003557407407407408,
      "loss": 4.0622,
      "step": 1080
    },
    {
      "epoch": 0.3315499962323864,
      "grad_norm": 0.340773344039917,
      "learning_rate": 0.00035203703703703704,
      "loss": 4.0458,
      "step": 1100
    },
    {
      "epoch": 0.3375781779820662,
      "grad_norm": 0.32752472162246704,
      "learning_rate": 0.00034833333333333336,
      "loss": 4.0316,
      "step": 1120
    },
    {
      "epoch": 0.3436063597317459,
      "grad_norm": 0.40199995040893555,
      "learning_rate": 0.0003446296296296296,
      "loss": 4.0227,
      "step": 1140
    },
    {
      "epoch": 0.34963454148142564,
      "grad_norm": 0.35206139087677,
      "learning_rate": 0.0003409259259259259,
      "loss": 4.0116,
      "step": 1160
    },
    {
      "epoch": 0.3556627232311054,
      "grad_norm": 0.3591882884502411,
      "learning_rate": 0.0003372222222222222,
      "loss": 4.0005,
      "step": 1180
    },
    {
      "epoch": 0.36169090498078516,
      "grad_norm": 0.28603240847587585,
      "learning_rate": 0.0003335185185185185,
      "loss": 3.9962,
      "step": 1200
    },
    {
      "epoch": 0.36169090498078516,
      "eval_loss": 3.943046739341145,
      "eval_perplexity": 51.57549926821255,
      "eval_runtime": 112.6141,
      "eval_samples_per_second": 14.732,
      "eval_steps_per_second": 0.231,
      "step": 1200
    },
    {
      "epoch": 0.36771908673046494,
      "grad_norm": 0.39475950598716736,
      "learning_rate": 0.00032981481481481485,
      "loss": 3.9885,
      "step": 1220
    },
    {
      "epoch": 0.37374726848014467,
      "grad_norm": 0.29654407501220703,
      "learning_rate": 0.0003261111111111111,
      "loss": 3.9779,
      "step": 1240
    },
    {
      "epoch": 0.37977545022982445,
      "grad_norm": 0.37094104290008545,
      "learning_rate": 0.00032240740740740743,
      "loss": 3.9649,
      "step": 1260
    },
    {
      "epoch": 0.3858036319795042,
      "grad_norm": 0.35296326875686646,
      "learning_rate": 0.0003187037037037037,
      "loss": 3.9592,
      "step": 1280
    },
    {
      "epoch": 0.3918318137291839,
      "grad_norm": 0.38376501202583313,
      "learning_rate": 0.000315,
      "loss": 3.9532,
      "step": 1300
    },
    {
      "epoch": 0.3978599954788637,
      "grad_norm": 0.367652028799057,
      "learning_rate": 0.0003112962962962963,
      "loss": 3.9428,
      "step": 1320
    },
    {
      "epoch": 0.4038881772285434,
      "grad_norm": 0.3195006847381592,
      "learning_rate": 0.0003075925925925926,
      "loss": 3.9389,
      "step": 1340
    },
    {
      "epoch": 0.4099163589782232,
      "grad_norm": 0.3439401686191559,
      "learning_rate": 0.00030388888888888886,
      "loss": 3.9316,
      "step": 1360
    },
    {
      "epoch": 0.41594454072790293,
      "grad_norm": 0.33015045523643494,
      "learning_rate": 0.0003001851851851852,
      "loss": 3.9261,
      "step": 1380
    },
    {
      "epoch": 0.4219727224775827,
      "grad_norm": 0.40396958589553833,
      "learning_rate": 0.00029648148148148144,
      "loss": 3.9188,
      "step": 1400
    },
    {
      "epoch": 0.42800090422726245,
      "grad_norm": 0.3784024119377136,
      "learning_rate": 0.0002927777777777778,
      "loss": 3.9163,
      "step": 1420
    },
    {
      "epoch": 0.43402908597694223,
      "grad_norm": 0.40174204111099243,
      "learning_rate": 0.0002890740740740741,
      "loss": 3.9013,
      "step": 1440
    },
    {
      "epoch": 0.44005726772662196,
      "grad_norm": 0.34229812026023865,
      "learning_rate": 0.0002853703703703704,
      "loss": 3.9055,
      "step": 1460
    },
    {
      "epoch": 0.4460854494763017,
      "grad_norm": 0.3183324635028839,
      "learning_rate": 0.00028166666666666666,
      "loss": 3.8986,
      "step": 1480
    },
    {
      "epoch": 0.45211363122598147,
      "grad_norm": 0.3185831308364868,
      "learning_rate": 0.000277962962962963,
      "loss": 3.8919,
      "step": 1500
    },
    {
      "epoch": 0.45211363122598147,
      "eval_loss": 3.8387107361752624,
      "eval_perplexity": 46.46552947940307,
      "eval_runtime": 112.6834,
      "eval_samples_per_second": 14.723,
      "eval_steps_per_second": 0.231,
      "step": 1500
    },
    {
      "epoch": 0.4581418129756612,
      "grad_norm": 0.33784908056259155,
      "learning_rate": 0.00027425925925925925,
      "loss": 3.8844,
      "step": 1520
    },
    {
      "epoch": 0.464169994725341,
      "grad_norm": 0.3474332392215729,
      "learning_rate": 0.00027055555555555557,
      "loss": 3.8814,
      "step": 1540
    },
    {
      "epoch": 0.4701981764750207,
      "grad_norm": 0.34041211009025574,
      "learning_rate": 0.00026685185185185183,
      "loss": 3.8697,
      "step": 1560
    },
    {
      "epoch": 0.4762263582247005,
      "grad_norm": 0.317830890417099,
      "learning_rate": 0.00026314814814814815,
      "loss": 3.8683,
      "step": 1580
    },
    {
      "epoch": 0.4822545399743802,
      "grad_norm": 0.33679041266441345,
      "learning_rate": 0.0002594444444444444,
      "loss": 3.8539,
      "step": 1600
    },
    {
      "epoch": 0.48828272172405995,
      "grad_norm": 0.31559520959854126,
      "learning_rate": 0.0002557407407407408,
      "loss": 3.8625,
      "step": 1620
    },
    {
      "epoch": 0.49431090347373974,
      "grad_norm": 0.3042216897010803,
      "learning_rate": 0.00025203703703703705,
      "loss": 3.8522,
      "step": 1640
    },
    {
      "epoch": 0.5003390852234195,
      "grad_norm": 0.30980879068374634,
      "learning_rate": 0.0002483333333333333,
      "loss": 3.8493,
      "step": 1660
    },
    {
      "epoch": 0.5063672669730992,
      "grad_norm": 0.3831731677055359,
      "learning_rate": 0.00024462962962962963,
      "loss": 3.847,
      "step": 1680
    },
    {
      "epoch": 0.512395448722779,
      "grad_norm": 0.4048588275909424,
      "learning_rate": 0.00024092592592592593,
      "loss": 3.8349,
      "step": 1700
    },
    {
      "epoch": 0.5184236304724588,
      "grad_norm": 0.2709442377090454,
      "learning_rate": 0.00023722222222222222,
      "loss": 3.8346,
      "step": 1720
    },
    {
      "epoch": 0.5244518122221385,
      "grad_norm": 0.36984342336654663,
      "learning_rate": 0.0002335185185185185,
      "loss": 3.8332,
      "step": 1740
    },
    {
      "epoch": 0.5304799939718182,
      "grad_norm": 0.32902956008911133,
      "learning_rate": 0.0002298148148148148,
      "loss": 3.8301,
      "step": 1760
    },
    {
      "epoch": 0.536508175721498,
      "grad_norm": 0.3468938171863556,
      "learning_rate": 0.00022611111111111112,
      "loss": 3.8199,
      "step": 1780
    },
    {
      "epoch": 0.5425363574711778,
      "grad_norm": 0.329237163066864,
      "learning_rate": 0.0002224074074074074,
      "loss": 3.8233,
      "step": 1800
    },
    {
      "epoch": 0.5425363574711778,
      "eval_loss": 3.770818455511187,
      "eval_perplexity": 43.415584022446424,
      "eval_runtime": 112.6023,
      "eval_samples_per_second": 14.733,
      "eval_steps_per_second": 0.231,
      "step": 1800
    },
    {
      "epoch": 0.5485645392208575,
      "grad_norm": 0.39494046568870544,
      "learning_rate": 0.0002187037037037037,
      "loss": 3.8195,
      "step": 1820
    },
    {
      "epoch": 0.5545927209705372,
      "grad_norm": 0.338908314704895,
      "learning_rate": 0.000215,
      "loss": 3.8122,
      "step": 1840
    },
    {
      "epoch": 0.560620902720217,
      "grad_norm": 0.3257177472114563,
      "learning_rate": 0.00021129629629629629,
      "loss": 3.8042,
      "step": 1860
    },
    {
      "epoch": 0.5666490844698968,
      "grad_norm": 0.3085819184780121,
      "learning_rate": 0.0002075925925925926,
      "loss": 3.8046,
      "step": 1880
    },
    {
      "epoch": 0.5726772662195765,
      "grad_norm": 0.3287888169288635,
      "learning_rate": 0.0002038888888888889,
      "loss": 3.7973,
      "step": 1900
    },
    {
      "epoch": 0.5787054479692563,
      "grad_norm": 0.32657161355018616,
      "learning_rate": 0.0002001851851851852,
      "loss": 3.7999,
      "step": 1920
    },
    {
      "epoch": 0.584733629718936,
      "grad_norm": 0.35923653841018677,
      "learning_rate": 0.00019648148148148148,
      "loss": 3.7944,
      "step": 1940
    },
    {
      "epoch": 0.5907618114686157,
      "grad_norm": 0.29656362533569336,
      "learning_rate": 0.00019277777777777777,
      "loss": 3.7943,
      "step": 1960
    },
    {
      "epoch": 0.5967899932182955,
      "grad_norm": 0.29753783345222473,
      "learning_rate": 0.00018907407407407406,
      "loss": 3.7822,
      "step": 1980
    },
    {
      "epoch": 0.6028181749679753,
      "grad_norm": 0.4390193521976471,
      "learning_rate": 0.00018537037037037038,
      "loss": 3.7844,
      "step": 2000
    },
    {
      "epoch": 0.6088463567176551,
      "grad_norm": 0.30851641297340393,
      "learning_rate": 0.00018166666666666667,
      "loss": 3.7798,
      "step": 2020
    },
    {
      "epoch": 0.6148745384673348,
      "grad_norm": 0.30867844820022583,
      "learning_rate": 0.00017796296296296296,
      "loss": 3.775,
      "step": 2040
    },
    {
      "epoch": 0.6209027202170145,
      "grad_norm": 0.27897268533706665,
      "learning_rate": 0.00017425925925925926,
      "loss": 3.7755,
      "step": 2060
    },
    {
      "epoch": 0.6269309019666943,
      "grad_norm": 0.30619484186172485,
      "learning_rate": 0.00017055555555555555,
      "loss": 3.7732,
      "step": 2080
    },
    {
      "epoch": 0.632959083716374,
      "grad_norm": 0.3682888150215149,
      "learning_rate": 0.00016685185185185187,
      "loss": 3.7683,
      "step": 2100
    },
    {
      "epoch": 0.632959083716374,
      "eval_loss": 3.721253386914116,
      "eval_perplexity": 41.31614678649486,
      "eval_runtime": 112.4817,
      "eval_samples_per_second": 14.749,
      "eval_steps_per_second": 0.231,
      "step": 2100
    },
    {
      "epoch": 0.6389872654660538,
      "grad_norm": 0.2890338599681854,
      "learning_rate": 0.00016314814814814816,
      "loss": 3.7673,
      "step": 2120
    },
    {
      "epoch": 0.6450154472157336,
      "grad_norm": 0.31391507387161255,
      "learning_rate": 0.00015944444444444445,
      "loss": 3.7657,
      "step": 2140
    },
    {
      "epoch": 0.6510436289654133,
      "grad_norm": 0.4431641697883606,
      "learning_rate": 0.00015574074074074074,
      "loss": 3.7629,
      "step": 2160
    },
    {
      "epoch": 0.657071810715093,
      "grad_norm": 0.26458096504211426,
      "learning_rate": 0.00015203703703703703,
      "loss": 3.7619,
      "step": 2180
    },
    {
      "epoch": 0.6630999924647728,
      "grad_norm": 0.3415466547012329,
      "learning_rate": 0.00014833333333333335,
      "loss": 3.7578,
      "step": 2200
    },
    {
      "epoch": 0.6691281742144526,
      "grad_norm": 0.39624544978141785,
      "learning_rate": 0.00014462962962962964,
      "loss": 3.7545,
      "step": 2220
    },
    {
      "epoch": 0.6751563559641324,
      "grad_norm": 0.2807711362838745,
      "learning_rate": 0.00014092592592592594,
      "loss": 3.7538,
      "step": 2240
    },
    {
      "epoch": 0.681184537713812,
      "grad_norm": 0.2499433159828186,
      "learning_rate": 0.00013722222222222223,
      "loss": 3.7406,
      "step": 2260
    },
    {
      "epoch": 0.6872127194634918,
      "grad_norm": 0.268779993057251,
      "learning_rate": 0.00013351851851851852,
      "loss": 3.7475,
      "step": 2280
    },
    {
      "epoch": 0.6932409012131716,
      "grad_norm": 0.35491231083869934,
      "learning_rate": 0.00012981481481481484,
      "loss": 3.7418,
      "step": 2300
    },
    {
      "epoch": 0.6992690829628513,
      "grad_norm": 0.27142778038978577,
      "learning_rate": 0.00012611111111111113,
      "loss": 3.7349,
      "step": 2320
    },
    {
      "epoch": 0.7052972647125311,
      "grad_norm": 0.4189681112766266,
      "learning_rate": 0.00012240740740740742,
      "loss": 3.7365,
      "step": 2340
    },
    {
      "epoch": 0.7113254464622109,
      "grad_norm": 0.3451978266239166,
      "learning_rate": 0.0001187037037037037,
      "loss": 3.7363,
      "step": 2360
    },
    {
      "epoch": 0.7173536282118906,
      "grad_norm": 0.30949023365974426,
      "learning_rate": 0.000115,
      "loss": 3.7269,
      "step": 2380
    },
    {
      "epoch": 0.7233818099615703,
      "grad_norm": 0.26108455657958984,
      "learning_rate": 0.0001112962962962963,
      "loss": 3.7345,
      "step": 2400
    },
    {
      "epoch": 0.7233818099615703,
      "eval_loss": 3.682990132542246,
      "eval_perplexity": 39.765119459534034,
      "eval_runtime": 112.6093,
      "eval_samples_per_second": 14.732,
      "eval_steps_per_second": 0.231,
      "step": 2400
    },
    {
      "epoch": 0.7294099917112501,
      "grad_norm": 0.24509066343307495,
      "learning_rate": 0.00010759259259259259,
      "loss": 3.7306,
      "step": 2420
    },
    {
      "epoch": 0.7354381734609299,
      "grad_norm": 0.24950160086154938,
      "learning_rate": 0.00010388888888888889,
      "loss": 3.7287,
      "step": 2440
    },
    {
      "epoch": 0.7414663552106096,
      "grad_norm": 0.25740230083465576,
      "learning_rate": 0.00010018518518518518,
      "loss": 3.7295,
      "step": 2460
    },
    {
      "epoch": 0.7474945369602893,
      "grad_norm": 0.2500341832637787,
      "learning_rate": 9.648148148148149e-05,
      "loss": 3.7216,
      "step": 2480
    },
    {
      "epoch": 0.7535227187099691,
      "grad_norm": 0.3777545392513275,
      "learning_rate": 9.277777777777778e-05,
      "loss": 3.722,
      "step": 2500
    },
    {
      "epoch": 0.7595509004596489,
      "grad_norm": 0.2761068344116211,
      "learning_rate": 8.907407407407407e-05,
      "loss": 3.721,
      "step": 2520
    },
    {
      "epoch": 0.7655790822093286,
      "grad_norm": 0.27542102336883545,
      "learning_rate": 8.537037037037038e-05,
      "loss": 3.7207,
      "step": 2540
    },
    {
      "epoch": 0.7716072639590084,
      "grad_norm": 0.23626072704792023,
      "learning_rate": 8.166666666666667e-05,
      "loss": 3.7152,
      "step": 2560
    },
    {
      "epoch": 0.7776354457086881,
      "grad_norm": 0.24672257900238037,
      "learning_rate": 7.796296296296296e-05,
      "loss": 3.7139,
      "step": 2580
    },
    {
      "epoch": 0.7836636274583678,
      "grad_norm": 0.23259390890598297,
      "learning_rate": 7.425925925925927e-05,
      "loss": 3.7085,
      "step": 2600
    },
    {
      "epoch": 0.7896918092080476,
      "grad_norm": 0.24034681916236877,
      "learning_rate": 7.055555555555556e-05,
      "loss": 3.7146,
      "step": 2620
    },
    {
      "epoch": 0.7957199909577274,
      "grad_norm": 0.2229200005531311,
      "learning_rate": 6.685185185185186e-05,
      "loss": 3.7112,
      "step": 2640
    },
    {
      "epoch": 0.8017481727074072,
      "grad_norm": 0.23400990664958954,
      "learning_rate": 6.314814814814815e-05,
      "loss": 3.7109,
      "step": 2660
    },
    {
      "epoch": 0.8077763544570868,
      "grad_norm": 0.24455134570598602,
      "learning_rate": 5.9444444444444445e-05,
      "loss": 3.7033,
      "step": 2680
    },
    {
      "epoch": 0.8138045362067666,
      "grad_norm": 0.22538278996944427,
      "learning_rate": 5.5740740740740744e-05,
      "loss": 3.706,
      "step": 2700
    },
    {
      "epoch": 0.8138045362067666,
      "eval_loss": 3.6566177565628872,
      "eval_perplexity": 38.730126377599234,
      "eval_runtime": 112.791,
      "eval_samples_per_second": 14.709,
      "eval_steps_per_second": 0.231,
      "step": 2700
    },
    {
      "epoch": 0.8198327179564464,
      "grad_norm": 0.22014465928077698,
      "learning_rate": 5.2037037037037035e-05,
      "loss": 3.7037,
      "step": 2720
    },
    {
      "epoch": 0.8258608997061262,
      "grad_norm": 0.2345183938741684,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 3.7059,
      "step": 2740
    },
    {
      "epoch": 0.8318890814558059,
      "grad_norm": 0.21457970142364502,
      "learning_rate": 4.462962962962963e-05,
      "loss": 3.7008,
      "step": 2760
    },
    {
      "epoch": 0.8379172632054857,
      "grad_norm": 0.21294698119163513,
      "learning_rate": 4.092592592592593e-05,
      "loss": 3.699,
      "step": 2780
    },
    {
      "epoch": 0.8439454449551654,
      "grad_norm": 0.22244296967983246,
      "learning_rate": 3.722222222222222e-05,
      "loss": 3.7013,
      "step": 2800
    },
    {
      "epoch": 0.8499736267048451,
      "grad_norm": 0.2127576470375061,
      "learning_rate": 3.351851851851852e-05,
      "loss": 3.7033,
      "step": 2820
    },
    {
      "epoch": 0.8560018084545249,
      "grad_norm": 0.2008371353149414,
      "learning_rate": 2.9814814814814815e-05,
      "loss": 3.696,
      "step": 2840
    },
    {
      "epoch": 0.8620299902042047,
      "grad_norm": 0.22562889754772186,
      "learning_rate": 2.6111111111111114e-05,
      "loss": 3.6925,
      "step": 2860
    },
    {
      "epoch": 0.8680581719538845,
      "grad_norm": 0.19617195427417755,
      "learning_rate": 2.240740740740741e-05,
      "loss": 3.6972,
      "step": 2880
    },
    {
      "epoch": 0.8740863537035641,
      "grad_norm": 0.1990758776664734,
      "learning_rate": 1.8703703703703707e-05,
      "loss": 3.692,
      "step": 2900
    },
    {
      "epoch": 0.8801145354532439,
      "grad_norm": 0.20326583087444305,
      "learning_rate": 1.5e-05,
      "loss": 3.6965,
      "step": 2920
    },
    {
      "epoch": 0.8861427172029237,
      "grad_norm": 0.186484694480896,
      "learning_rate": 1.1296296296296295e-05,
      "loss": 3.6869,
      "step": 2940
    },
    {
      "epoch": 0.8921708989526034,
      "grad_norm": 0.1895342916250229,
      "learning_rate": 7.592592592592593e-06,
      "loss": 3.6835,
      "step": 2960
    },
    {
      "epoch": 0.8981990807022832,
      "grad_norm": 0.17950916290283203,
      "learning_rate": 3.888888888888889e-06,
      "loss": 3.6912,
      "step": 2980
    },
    {
      "epoch": 0.9042272624519629,
      "grad_norm": 0.18661454319953918,
      "learning_rate": 1.8518518518518518e-07,
      "loss": 3.6943,
      "step": 3000
    },
    {
      "epoch": 0.9042272624519629,
      "eval_loss": 3.6425010418158132,
      "eval_perplexity": 38.18722523749619,
      "eval_runtime": 112.6033,
      "eval_samples_per_second": 14.733,
      "eval_steps_per_second": 0.231,
      "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
}