{
  "best_global_step": 3000,
  "best_metric": 3.602764652887152,
  "best_model_checkpoint": "checkpoints/gpt-2_seq1024_mla192-0-0/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": 2.0171725749969482,
      "learning_rate": 3.166666666666667e-05,
      "loss": 10.4028,
      "step": 20
    },
    {
      "epoch": 0.012056363499359506,
      "grad_norm": 1.66342031955719,
      "learning_rate": 6.500000000000001e-05,
      "loss": 9.2471,
      "step": 40
    },
    {
      "epoch": 0.01808454524903926,
      "grad_norm": 0.7821094989776611,
      "learning_rate": 9.833333333333333e-05,
      "loss": 8.2622,
      "step": 60
    },
    {
      "epoch": 0.02411272699871901,
      "grad_norm": 0.41097572445869446,
      "learning_rate": 0.00013166666666666665,
      "loss": 7.55,
      "step": 80
    },
    {
      "epoch": 0.030140908748398764,
      "grad_norm": 0.5691717863082886,
      "learning_rate": 0.000165,
      "loss": 7.2211,
      "step": 100
    },
    {
      "epoch": 0.03616909049807852,
      "grad_norm": 0.7314554452896118,
      "learning_rate": 0.00019833333333333335,
      "loss": 6.9285,
      "step": 120
    },
    {
      "epoch": 0.04219727224775827,
      "grad_norm": 1.0012000799179077,
      "learning_rate": 0.00023166666666666667,
      "loss": 6.6558,
      "step": 140
    },
    {
      "epoch": 0.04822545399743802,
      "grad_norm": 1.0241062641143799,
      "learning_rate": 0.00026500000000000004,
      "loss": 6.4608,
      "step": 160
    },
    {
      "epoch": 0.05425363574711778,
      "grad_norm": 0.9315752983093262,
      "learning_rate": 0.00029833333333333334,
      "loss": 6.2917,
      "step": 180
    },
    {
      "epoch": 0.06028181749679753,
      "grad_norm": 0.8718670606613159,
      "learning_rate": 0.0003316666666666667,
      "loss": 6.134,
      "step": 200
    },
    {
      "epoch": 0.06630999924647728,
      "grad_norm": 0.7919574975967407,
      "learning_rate": 0.000365,
      "loss": 5.9921,
      "step": 220
    },
    {
      "epoch": 0.07233818099615703,
      "grad_norm": 0.46674615144729614,
      "learning_rate": 0.00039833333333333333,
      "loss": 5.854,
      "step": 240
    },
    {
      "epoch": 0.07836636274583679,
      "grad_norm": 0.6191166043281555,
      "learning_rate": 0.0004316666666666667,
      "loss": 5.7443,
      "step": 260
    },
    {
      "epoch": 0.08439454449551655,
      "grad_norm": 0.6590830087661743,
      "learning_rate": 0.000465,
      "loss": 5.6087,
      "step": 280
    },
    {
      "epoch": 0.09042272624519629,
      "grad_norm": 1.0058612823486328,
      "learning_rate": 0.0004983333333333334,
      "loss": 5.5362,
      "step": 300
    },
    {
      "epoch": 0.09042272624519629,
      "eval_loss": 5.480001850152933,
      "eval_perplexity": 239.84715112775478,
      "eval_runtime": 112.6058,
      "eval_samples_per_second": 14.733,
      "eval_steps_per_second": 0.231,
      "step": 300
    },
    {
      "epoch": 0.09645090799487605,
      "grad_norm": 0.4994821846485138,
      "learning_rate": 0.0004964814814814814,
      "loss": 5.4092,
      "step": 320
    },
    {
      "epoch": 0.1024790897445558,
      "grad_norm": 0.4082713723182678,
      "learning_rate": 0.0004927777777777777,
      "loss": 5.3019,
      "step": 340
    },
    {
      "epoch": 0.10850727149423556,
      "grad_norm": 0.6187137961387634,
      "learning_rate": 0.0004890740740740741,
      "loss": 5.2061,
      "step": 360
    },
    {
      "epoch": 0.1145354532439153,
      "grad_norm": 0.9009650349617004,
      "learning_rate": 0.0004853703703703704,
      "loss": 5.1354,
      "step": 380
    },
    {
      "epoch": 0.12056363499359506,
      "grad_norm": 0.35602667927742004,
      "learning_rate": 0.0004816666666666667,
      "loss": 5.061,
      "step": 400
    },
    {
      "epoch": 0.1265918167432748,
      "grad_norm": 0.4166870713233948,
      "learning_rate": 0.00047796296296296297,
      "loss": 4.9749,
      "step": 420
    },
    {
      "epoch": 0.13261999849295456,
      "grad_norm": 0.5855613946914673,
      "learning_rate": 0.0004742592592592593,
      "loss": 4.9244,
      "step": 440
    },
    {
      "epoch": 0.1386481802426343,
      "grad_norm": 0.44319576025009155,
      "learning_rate": 0.00047055555555555555,
      "loss": 4.8589,
      "step": 460
    },
    {
      "epoch": 0.14467636199231407,
      "grad_norm": 0.36183497309684753,
      "learning_rate": 0.00046685185185185187,
      "loss": 4.8089,
      "step": 480
    },
    {
      "epoch": 0.15070454374199382,
      "grad_norm": 0.6022749543190002,
      "learning_rate": 0.00046314814814814813,
      "loss": 4.7438,
      "step": 500
    },
    {
      "epoch": 0.15673272549167358,
      "grad_norm": 0.5266714096069336,
      "learning_rate": 0.00045944444444444445,
      "loss": 4.6892,
      "step": 520
    },
    {
      "epoch": 0.16276090724135334,
      "grad_norm": 0.4757547676563263,
      "learning_rate": 0.0004557407407407407,
      "loss": 4.6401,
      "step": 540
    },
    {
      "epoch": 0.1687890889910331,
      "grad_norm": 0.516314685344696,
      "learning_rate": 0.00045203703703703703,
      "loss": 4.5807,
      "step": 560
    },
    {
      "epoch": 0.17481727074071282,
      "grad_norm": 0.4857338070869446,
      "learning_rate": 0.0004483333333333333,
      "loss": 4.5335,
      "step": 580
    },
    {
      "epoch": 0.18084545249039258,
      "grad_norm": 0.47939416766166687,
      "learning_rate": 0.00044462962962962967,
      "loss": 4.4956,
      "step": 600
    },
    {
      "epoch": 0.18084545249039258,
      "eval_loss": 4.426357781719664,
      "eval_perplexity": 83.62627640080112,
      "eval_runtime": 112.2391,
      "eval_samples_per_second": 14.781,
      "eval_steps_per_second": 0.232,
      "step": 600
    },
    {
      "epoch": 0.18687363424007233,
      "grad_norm": 0.4611107110977173,
      "learning_rate": 0.00044092592592592594,
      "loss": 4.4429,
      "step": 620
    },
    {
      "epoch": 0.1929018159897521,
      "grad_norm": 0.36294886469841003,
      "learning_rate": 0.00043722222222222225,
      "loss": 4.4097,
      "step": 640
    },
    {
      "epoch": 0.19892999773943185,
      "grad_norm": 0.4987587034702301,
      "learning_rate": 0.0004335185185185185,
      "loss": 4.3755,
      "step": 660
    },
    {
      "epoch": 0.2049581794891116,
      "grad_norm": 0.32254645228385925,
      "learning_rate": 0.00042981481481481484,
      "loss": 4.3561,
      "step": 680
    },
    {
      "epoch": 0.21098636123879136,
      "grad_norm": 0.6072115302085876,
      "learning_rate": 0.0004261111111111111,
      "loss": 4.3126,
      "step": 700
    },
    {
      "epoch": 0.21701454298847112,
      "grad_norm": 0.4066341519355774,
      "learning_rate": 0.0004224074074074074,
      "loss": 4.2923,
      "step": 720
    },
    {
      "epoch": 0.22304272473815084,
      "grad_norm": 0.3725150227546692,
      "learning_rate": 0.0004187037037037037,
      "loss": 4.263,
      "step": 740
    },
    {
      "epoch": 0.2290709064878306,
      "grad_norm": 0.37484949827194214,
      "learning_rate": 0.000415,
      "loss": 4.2421,
      "step": 760
    },
    {
      "epoch": 0.23509908823751036,
      "grad_norm": 0.431652694940567,
      "learning_rate": 0.00041129629629629627,
      "loss": 4.228,
      "step": 780
    },
    {
      "epoch": 0.2411272699871901,
      "grad_norm": 0.3584285080432892,
      "learning_rate": 0.00040759259259259264,
      "loss": 4.2054,
      "step": 800
    },
    {
      "epoch": 0.24715545173686987,
      "grad_norm": 0.4462144374847412,
      "learning_rate": 0.0004038888888888889,
      "loss": 4.1834,
      "step": 820
    },
    {
      "epoch": 0.2531836334865496,
      "grad_norm": 0.29847240447998047,
      "learning_rate": 0.0004001851851851852,
      "loss": 4.1624,
      "step": 840
    },
    {
      "epoch": 0.2592118152362294,
      "grad_norm": 0.4533553123474121,
      "learning_rate": 0.0003964814814814815,
      "loss": 4.1495,
      "step": 860
    },
    {
      "epoch": 0.2652399969859091,
      "grad_norm": 0.4561823010444641,
      "learning_rate": 0.0003927777777777778,
      "loss": 4.1357,
      "step": 880
    },
    {
      "epoch": 0.2712681787355889,
      "grad_norm": 0.33924582600593567,
      "learning_rate": 0.00038907407407407407,
      "loss": 4.1124,
      "step": 900
    },
    {
      "epoch": 0.2712681787355889,
      "eval_loss": 4.058111604067862,
      "eval_perplexity": 57.86493593004541,
      "eval_runtime": 112.261,
      "eval_samples_per_second": 14.778,
      "eval_steps_per_second": 0.232,
      "step": 900
    },
    {
      "epoch": 0.2772963604852686,
      "grad_norm": 0.5096554756164551,
      "learning_rate": 0.0003853703703703704,
      "loss": 4.1011,
      "step": 920
    },
    {
      "epoch": 0.2833245422349484,
      "grad_norm": 0.3418212831020355,
      "learning_rate": 0.00038166666666666666,
      "loss": 4.09,
      "step": 940
    },
    {
      "epoch": 0.28935272398462814,
      "grad_norm": 0.269795686006546,
      "learning_rate": 0.000377962962962963,
      "loss": 4.0681,
      "step": 960
    },
    {
      "epoch": 0.29538090573430786,
      "grad_norm": 0.35889971256256104,
      "learning_rate": 0.00037425925925925924,
      "loss": 4.0686,
      "step": 980
    },
    {
      "epoch": 0.30140908748398765,
      "grad_norm": 0.411249577999115,
      "learning_rate": 0.0003705555555555556,
      "loss": 4.0558,
      "step": 1000
    },
    {
      "epoch": 0.3074372692336674,
      "grad_norm": 0.3365558981895447,
      "learning_rate": 0.0003668518518518519,
      "loss": 4.0403,
      "step": 1020
    },
    {
      "epoch": 0.31346545098334716,
      "grad_norm": 0.35357165336608887,
      "learning_rate": 0.0003631481481481482,
      "loss": 4.0246,
      "step": 1040
    },
    {
      "epoch": 0.3194936327330269,
      "grad_norm": 0.25508153438568115,
      "learning_rate": 0.00035944444444444446,
      "loss": 4.0169,
      "step": 1060
    },
    {
      "epoch": 0.3255218144827067,
      "grad_norm": 0.3912709355354309,
      "learning_rate": 0.0003557407407407408,
      "loss": 4.0066,
      "step": 1080
    },
    {
      "epoch": 0.3315499962323864,
      "grad_norm": 0.5864457488059998,
      "learning_rate": 0.00035203703703703704,
      "loss": 3.9925,
      "step": 1100
    },
    {
      "epoch": 0.3375781779820662,
      "grad_norm": 0.2764241099357605,
      "learning_rate": 0.00034833333333333336,
      "loss": 3.9831,
      "step": 1120
    },
    {
      "epoch": 0.3436063597317459,
      "grad_norm": 0.330556720495224,
      "learning_rate": 0.0003446296296296296,
      "loss": 3.9741,
      "step": 1140
    },
    {
      "epoch": 0.34963454148142564,
      "grad_norm": 0.34419673681259155,
      "learning_rate": 0.0003409259259259259,
      "loss": 3.9597,
      "step": 1160
    },
    {
      "epoch": 0.3556627232311054,
      "grad_norm": 0.29691699147224426,
      "learning_rate": 0.0003372222222222222,
      "loss": 3.9525,
      "step": 1180
    },
    {
      "epoch": 0.36169090498078516,
      "grad_norm": 0.2951587438583374,
      "learning_rate": 0.0003335185185185185,
      "loss": 3.9489,
      "step": 1200
    },
    {
      "epoch": 0.36169090498078516,
      "eval_loss": 3.8972609229729485,
      "eval_perplexity": 49.26731714501557,
      "eval_runtime": 112.1012,
      "eval_samples_per_second": 14.799,
      "eval_steps_per_second": 0.232,
      "step": 1200
    },
    {
      "epoch": 0.36771908673046494,
      "grad_norm": 0.28599607944488525,
      "learning_rate": 0.00032981481481481485,
      "loss": 3.9341,
      "step": 1220
    },
    {
      "epoch": 0.37374726848014467,
      "grad_norm": 0.44271770119667053,
      "learning_rate": 0.0003261111111111111,
      "loss": 3.9357,
      "step": 1240
    },
    {
      "epoch": 0.37977545022982445,
      "grad_norm": 0.3313683271408081,
      "learning_rate": 0.00032240740740740743,
      "loss": 3.9205,
      "step": 1260
    },
    {
      "epoch": 0.3858036319795042,
      "grad_norm": 0.2592178285121918,
      "learning_rate": 0.0003187037037037037,
      "loss": 3.9143,
      "step": 1280
    },
    {
      "epoch": 0.3918318137291839,
      "grad_norm": 0.37822020053863525,
      "learning_rate": 0.000315,
      "loss": 3.9085,
      "step": 1300
    },
    {
      "epoch": 0.3978599954788637,
      "grad_norm": 0.33519598841667175,
      "learning_rate": 0.0003112962962962963,
      "loss": 3.8993,
      "step": 1320
    },
    {
      "epoch": 0.4038881772285434,
      "grad_norm": 0.36391812562942505,
      "learning_rate": 0.0003075925925925926,
      "loss": 3.8956,
      "step": 1340
    },
    {
      "epoch": 0.4099163589782232,
      "grad_norm": 0.2704394459724426,
      "learning_rate": 0.00030388888888888886,
      "loss": 3.889,
      "step": 1360
    },
    {
      "epoch": 0.41594454072790293,
      "grad_norm": 0.38525649905204773,
      "learning_rate": 0.0003001851851851852,
      "loss": 3.8811,
      "step": 1380
    },
    {
      "epoch": 0.4219727224775827,
      "grad_norm": 0.29260268807411194,
      "learning_rate": 0.00029648148148148144,
      "loss": 3.8735,
      "step": 1400
    },
    {
      "epoch": 0.42800090422726245,
      "grad_norm": 0.2522657811641693,
      "learning_rate": 0.0002927777777777778,
      "loss": 3.8714,
      "step": 1420
    },
    {
      "epoch": 0.43402908597694223,
      "grad_norm": 0.31784236431121826,
      "learning_rate": 0.0002890740740740741,
      "loss": 3.8596,
      "step": 1440
    },
    {
      "epoch": 0.44005726772662196,
      "grad_norm": 0.3017977476119995,
      "learning_rate": 0.0002853703703703704,
      "loss": 3.8624,
      "step": 1460
    },
    {
      "epoch": 0.4460854494763017,
      "grad_norm": 0.3564951717853546,
      "learning_rate": 0.00028166666666666666,
      "loss": 3.8571,
      "step": 1480
    },
    {
      "epoch": 0.45211363122598147,
      "grad_norm": 0.2593876123428345,
      "learning_rate": 0.000277962962962963,
      "loss": 3.8495,
      "step": 1500
    },
    {
      "epoch": 0.45211363122598147,
      "eval_loss": 3.795062538778086,
      "eval_perplexity": 44.481018107066156,
      "eval_runtime": 112.4309,
      "eval_samples_per_second": 14.756,
      "eval_steps_per_second": 0.231,
      "step": 1500
    },
    {
      "epoch": 0.4581418129756612,
      "grad_norm": 0.32840925455093384,
      "learning_rate": 0.00027425925925925925,
      "loss": 3.8425,
      "step": 1520
    },
    {
      "epoch": 0.464169994725341,
      "grad_norm": 0.3139461874961853,
      "learning_rate": 0.00027055555555555557,
      "loss": 3.839,
      "step": 1540
    },
    {
      "epoch": 0.4701981764750207,
      "grad_norm": 0.25834277272224426,
      "learning_rate": 0.00026685185185185183,
      "loss": 3.8272,
      "step": 1560
    },
    {
      "epoch": 0.4762263582247005,
      "grad_norm": 0.2804831266403198,
      "learning_rate": 0.00026314814814814815,
      "loss": 3.8248,
      "step": 1580
    },
    {
      "epoch": 0.4822545399743802,
      "grad_norm": 0.3130384087562561,
      "learning_rate": 0.0002594444444444444,
      "loss": 3.8114,
      "step": 1600
    },
    {
      "epoch": 0.48828272172405995,
      "grad_norm": 0.2702663242816925,
      "learning_rate": 0.0002557407407407408,
      "loss": 3.8205,
      "step": 1620
    },
    {
      "epoch": 0.49431090347373974,
      "grad_norm": 0.35775506496429443,
      "learning_rate": 0.00025203703703703705,
      "loss": 3.8103,
      "step": 1640
    },
    {
      "epoch": 0.5003390852234195,
      "grad_norm": 0.33006781339645386,
      "learning_rate": 0.0002483333333333333,
      "loss": 3.8069,
      "step": 1660
    },
    {
      "epoch": 0.5063672669730992,
      "grad_norm": 0.28608810901641846,
      "learning_rate": 0.00024462962962962963,
      "loss": 3.8046,
      "step": 1680
    },
    {
      "epoch": 0.512395448722779,
      "grad_norm": 0.2452249377965927,
      "learning_rate": 0.00024092592592592593,
      "loss": 3.7919,
      "step": 1700
    },
    {
      "epoch": 0.5184236304724588,
      "grad_norm": 0.29953446984291077,
      "learning_rate": 0.00023722222222222222,
      "loss": 3.7933,
      "step": 1720
    },
    {
      "epoch": 0.5244518122221385,
      "grad_norm": 0.31593525409698486,
      "learning_rate": 0.0002335185185185185,
      "loss": 3.7921,
      "step": 1740
    },
    {
      "epoch": 0.5304799939718182,
      "grad_norm": 0.37334874272346497,
      "learning_rate": 0.0002298148148148148,
      "loss": 3.7888,
      "step": 1760
    },
    {
      "epoch": 0.536508175721498,
      "grad_norm": 0.3065856695175171,
      "learning_rate": 0.00022611111111111112,
      "loss": 3.7784,
      "step": 1780
    },
    {
      "epoch": 0.5425363574711778,
      "grad_norm": 0.3357571065425873,
      "learning_rate": 0.0002224074074074074,
      "loss": 3.7812,
      "step": 1800
    },
    {
      "epoch": 0.5425363574711778,
      "eval_loss": 3.728826272039063,
      "eval_perplexity": 41.6302169275088,
      "eval_runtime": 112.1495,
      "eval_samples_per_second": 14.793,
      "eval_steps_per_second": 0.232,
      "step": 1800
    },
    {
      "epoch": 0.5485645392208575,
      "grad_norm": 0.2727905511856079,
      "learning_rate": 0.0002187037037037037,
      "loss": 3.7787,
      "step": 1820
    },
    {
      "epoch": 0.5545927209705372,
      "grad_norm": 0.25383642315864563,
      "learning_rate": 0.000215,
      "loss": 3.7697,
      "step": 1840
    },
    {
      "epoch": 0.560620902720217,
      "grad_norm": 0.2943139672279358,
      "learning_rate": 0.00021129629629629629,
      "loss": 3.7624,
      "step": 1860
    },
    {
      "epoch": 0.5666490844698968,
      "grad_norm": 0.2697204053401947,
      "learning_rate": 0.0002075925925925926,
      "loss": 3.7629,
      "step": 1880
    },
    {
      "epoch": 0.5726772662195765,
      "grad_norm": 0.2720482349395752,
      "learning_rate": 0.0002038888888888889,
      "loss": 3.7545,
      "step": 1900
    },
    {
      "epoch": 0.5787054479692563,
      "grad_norm": 0.3047272264957428,
      "learning_rate": 0.0002001851851851852,
      "loss": 3.7567,
      "step": 1920
    },
    {
      "epoch": 0.584733629718936,
      "grad_norm": 0.28646621108055115,
      "learning_rate": 0.00019648148148148148,
      "loss": 3.7521,
      "step": 1940
    },
    {
      "epoch": 0.5907618114686157,
      "grad_norm": 0.27155548334121704,
      "learning_rate": 0.00019277777777777777,
      "loss": 3.7528,
      "step": 1960
    },
    {
      "epoch": 0.5967899932182955,
      "grad_norm": 0.3533537685871124,
      "learning_rate": 0.00018907407407407406,
      "loss": 3.7403,
      "step": 1980
    },
    {
      "epoch": 0.6028181749679753,
      "grad_norm": 0.29706355929374695,
      "learning_rate": 0.00018537037037037038,
      "loss": 3.7428,
      "step": 2000
    },
    {
      "epoch": 0.6088463567176551,
      "grad_norm": 0.2581825852394104,
      "learning_rate": 0.00018166666666666667,
      "loss": 3.7382,
      "step": 2020
    },
    {
      "epoch": 0.6148745384673348,
      "grad_norm": 0.278071790933609,
      "learning_rate": 0.00017796296296296296,
      "loss": 3.7328,
      "step": 2040
    },
    {
      "epoch": 0.6209027202170145,
      "grad_norm": 0.2304176539182663,
      "learning_rate": 0.00017425925925925926,
      "loss": 3.7338,
      "step": 2060
    },
    {
      "epoch": 0.6269309019666943,
      "grad_norm": 0.2794628441333771,
      "learning_rate": 0.00017055555555555555,
      "loss": 3.7316,
      "step": 2080
    },
    {
      "epoch": 0.632959083716374,
      "grad_norm": 0.2320399135351181,
      "learning_rate": 0.00016685185185185187,
      "loss": 3.7267,
      "step": 2100
    },
    {
      "epoch": 0.632959083716374,
      "eval_loss": 3.6778735519165284,
      "eval_perplexity": 39.5621776465433,
      "eval_runtime": 112.1248,
      "eval_samples_per_second": 14.796,
      "eval_steps_per_second": 0.232,
      "step": 2100
    },
    {
      "epoch": 0.6389872654660538,
      "grad_norm": 0.26810985803604126,
      "learning_rate": 0.00016314814814814816,
      "loss": 3.7259,
      "step": 2120
    },
    {
      "epoch": 0.6450154472157336,
      "grad_norm": 0.28521475195884705,
      "learning_rate": 0.00015944444444444445,
      "loss": 3.7242,
      "step": 2140
    },
    {
      "epoch": 0.6510436289654133,
      "grad_norm": 0.24218100309371948,
      "learning_rate": 0.00015574074074074074,
      "loss": 3.7209,
      "step": 2160
    },
    {
      "epoch": 0.657071810715093,
      "grad_norm": 0.2587539553642273,
      "learning_rate": 0.00015203703703703703,
      "loss": 3.7198,
      "step": 2180
    },
    {
      "epoch": 0.6630999924647728,
      "grad_norm": 0.29283982515335083,
      "learning_rate": 0.00014833333333333335,
      "loss": 3.7162,
      "step": 2200
    },
    {
      "epoch": 0.6691281742144526,
      "grad_norm": 0.23661203682422638,
      "learning_rate": 0.00014462962962962964,
      "loss": 3.714,
      "step": 2220
    },
    {
      "epoch": 0.6751563559641324,
      "grad_norm": 0.2184392660856247,
      "learning_rate": 0.00014092592592592594,
      "loss": 3.7121,
      "step": 2240
    },
    {
      "epoch": 0.681184537713812,
      "grad_norm": 0.24910466372966766,
      "learning_rate": 0.00013722222222222223,
      "loss": 3.6996,
      "step": 2260
    },
    {
      "epoch": 0.6872127194634918,
      "grad_norm": 0.2523048222064972,
      "learning_rate": 0.00013351851851851852,
      "loss": 3.7068,
      "step": 2280
    },
    {
      "epoch": 0.6932409012131716,
      "grad_norm": 0.25991636514663696,
      "learning_rate": 0.00012981481481481484,
      "loss": 3.6997,
      "step": 2300
    },
    {
      "epoch": 0.6992690829628513,
      "grad_norm": 0.2645404636859894,
      "learning_rate": 0.00012611111111111113,
      "loss": 3.6931,
      "step": 2320
    },
    {
      "epoch": 0.7052972647125311,
      "grad_norm": 0.24797891080379486,
      "learning_rate": 0.00012240740740740742,
      "loss": 3.6951,
      "step": 2340
    },
    {
      "epoch": 0.7113254464622109,
      "grad_norm": 0.24686554074287415,
      "learning_rate": 0.0001187037037037037,
      "loss": 3.6944,
      "step": 2360
    },
    {
      "epoch": 0.7173536282118906,
      "grad_norm": 0.25571176409721375,
      "learning_rate": 0.000115,
      "loss": 3.6853,
      "step": 2380
    },
    {
      "epoch": 0.7233818099615703,
      "grad_norm": 0.23296213150024414,
      "learning_rate": 0.0001112962962962963,
      "loss": 3.6937,
      "step": 2400
    },
    {
      "epoch": 0.7233818099615703,
      "eval_loss": 3.6413854997651955,
      "eval_perplexity": 38.14464953385645,
      "eval_runtime": 112.0529,
      "eval_samples_per_second": 14.806,
      "eval_steps_per_second": 0.232,
      "step": 2400
    },
    {
      "epoch": 0.7294099917112501,
      "grad_norm": 0.23208530247211456,
      "learning_rate": 0.00010759259259259259,
      "loss": 3.6898,
      "step": 2420
    },
    {
      "epoch": 0.7354381734609299,
      "grad_norm": 0.2420375943183899,
      "learning_rate": 0.00010388888888888889,
      "loss": 3.6874,
      "step": 2440
    },
    {
      "epoch": 0.7414663552106096,
      "grad_norm": 0.2512781620025635,
      "learning_rate": 0.00010018518518518518,
      "loss": 3.6881,
      "step": 2460
    },
    {
      "epoch": 0.7474945369602893,
      "grad_norm": 0.20650093257427216,
      "learning_rate": 9.648148148148149e-05,
      "loss": 3.6808,
      "step": 2480
    },
    {
      "epoch": 0.7535227187099691,
      "grad_norm": 0.2324400097131729,
      "learning_rate": 9.277777777777778e-05,
      "loss": 3.6816,
      "step": 2500
    },
    {
      "epoch": 0.7595509004596489,
      "grad_norm": 0.21381932497024536,
      "learning_rate": 8.907407407407407e-05,
      "loss": 3.6802,
      "step": 2520
    },
    {
      "epoch": 0.7655790822093286,
      "grad_norm": 0.22228114306926727,
      "learning_rate": 8.537037037037038e-05,
      "loss": 3.68,
      "step": 2540
    },
    {
      "epoch": 0.7716072639590084,
      "grad_norm": 0.21108661592006683,
      "learning_rate": 8.166666666666667e-05,
      "loss": 3.6752,
      "step": 2560
    },
    {
      "epoch": 0.7776354457086881,
      "grad_norm": 0.20819643139839172,
      "learning_rate": 7.796296296296296e-05,
      "loss": 3.6737,
      "step": 2580
    },
    {
      "epoch": 0.7836636274583678,
      "grad_norm": 0.21848240494728088,
      "learning_rate": 7.425925925925927e-05,
      "loss": 3.6672,
      "step": 2600
    },
    {
      "epoch": 0.7896918092080476,
      "grad_norm": 0.2202238291501999,
      "learning_rate": 7.055555555555556e-05,
      "loss": 3.6735,
      "step": 2620
    },
    {
      "epoch": 0.7957199909577274,
      "grad_norm": 0.19879350066184998,
      "learning_rate": 6.685185185185186e-05,
      "loss": 3.6707,
      "step": 2640
    },
    {
      "epoch": 0.8017481727074072,
      "grad_norm": 0.204148530960083,
      "learning_rate": 6.314814814814815e-05,
      "loss": 3.6701,
      "step": 2660
    },
    {
      "epoch": 0.8077763544570868,
      "grad_norm": 0.2012234628200531,
      "learning_rate": 5.9444444444444445e-05,
      "loss": 3.6627,
      "step": 2680
    },
    {
      "epoch": 0.8138045362067666,
      "grad_norm": 0.1901106983423233,
      "learning_rate": 5.5740740740740744e-05,
      "loss": 3.6656,
      "step": 2700
    },
    {
      "epoch": 0.8138045362067666,
      "eval_loss": 3.6161044157228823,
      "eval_perplexity": 37.192399121937946,
      "eval_runtime": 112.241,
      "eval_samples_per_second": 14.781,
      "eval_steps_per_second": 0.232,
      "step": 2700
    },
    {
      "epoch": 0.8198327179564464,
      "grad_norm": 0.19361631572246552,
      "learning_rate": 5.2037037037037035e-05,
      "loss": 3.6636,
      "step": 2720
    },
    {
      "epoch": 0.8258608997061262,
      "grad_norm": 0.1905573159456253,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 3.6658,
      "step": 2740
    },
    {
      "epoch": 0.8318890814558059,
      "grad_norm": 0.18366636335849762,
      "learning_rate": 4.462962962962963e-05,
      "loss": 3.6613,
      "step": 2760
    },
    {
      "epoch": 0.8379172632054857,
      "grad_norm": 0.18855422735214233,
      "learning_rate": 4.092592592592593e-05,
      "loss": 3.6594,
      "step": 2780
    },
    {
      "epoch": 0.8439454449551654,
      "grad_norm": 0.18520458042621613,
      "learning_rate": 3.722222222222222e-05,
      "loss": 3.662,
      "step": 2800
    },
    {
      "epoch": 0.8499736267048451,
      "grad_norm": 0.17435666918754578,
      "learning_rate": 3.351851851851852e-05,
      "loss": 3.6639,
      "step": 2820
    },
    {
      "epoch": 0.8560018084545249,
      "grad_norm": 0.1787436455488205,
      "learning_rate": 2.9814814814814815e-05,
      "loss": 3.6559,
      "step": 2840
    },
    {
      "epoch": 0.8620299902042047,
      "grad_norm": 0.1714988499879837,
      "learning_rate": 2.6111111111111114e-05,
      "loss": 3.6534,
      "step": 2860
    },
    {
      "epoch": 0.8680581719538845,
      "grad_norm": 0.1794491559267044,
      "learning_rate": 2.240740740740741e-05,
      "loss": 3.6575,
      "step": 2880
    },
    {
      "epoch": 0.8740863537035641,
      "grad_norm": 0.17227113246917725,
      "learning_rate": 1.8703703703703707e-05,
      "loss": 3.6525,
      "step": 2900
    },
    {
      "epoch": 0.8801145354532439,
      "grad_norm": 0.16660287976264954,
      "learning_rate": 1.5e-05,
      "loss": 3.6569,
      "step": 2920
    },
    {
      "epoch": 0.8861427172029237,
      "grad_norm": 0.16066555678844452,
      "learning_rate": 1.1296296296296295e-05,
      "loss": 3.6477,
      "step": 2940
    },
    {
      "epoch": 0.8921708989526034,
      "grad_norm": 0.160615935921669,
      "learning_rate": 7.592592592592593e-06,
      "loss": 3.644,
      "step": 2960
    },
    {
      "epoch": 0.8981990807022832,
      "grad_norm": 0.15302099287509918,
      "learning_rate": 3.888888888888889e-06,
      "loss": 3.6517,
      "step": 2980
    },
    {
      "epoch": 0.9042272624519629,
      "grad_norm": 0.15698660910129547,
      "learning_rate": 1.8518518518518518e-07,
      "loss": 3.6558,
      "step": 3000
    },
    {
      "epoch": 0.9042272624519629,
      "eval_loss": 3.602764652887152,
      "eval_perplexity": 36.6995558529248,
      "eval_runtime": 112.1149,
      "eval_samples_per_second": 14.797,
      "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
}