{
  "best_global_step": 3000,
  "best_metric": 3.614897001279198,
  "best_model_checkpoint": "checkpoints/gpt-2_seq1024_mla0-128-0/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": 2.0992038249969482,
      "learning_rate": 3.166666666666667e-05,
      "loss": 10.4611,
      "step": 20
    },
    {
      "epoch": 0.012056363499359506,
      "grad_norm": 1.4955384731292725,
      "learning_rate": 6.500000000000001e-05,
      "loss": 9.0904,
      "step": 40
    },
    {
      "epoch": 0.01808454524903926,
      "grad_norm": 1.5293831825256348,
      "learning_rate": 9.833333333333333e-05,
      "loss": 8.0384,
      "step": 60
    },
    {
      "epoch": 0.02411272699871901,
      "grad_norm": 0.5085040926933289,
      "learning_rate": 0.00013166666666666665,
      "loss": 7.2875,
      "step": 80
    },
    {
      "epoch": 0.030140908748398764,
      "grad_norm": 0.6033186912536621,
      "learning_rate": 0.000165,
      "loss": 6.8708,
      "step": 100
    },
    {
      "epoch": 0.03616909049807852,
      "grad_norm": 0.6809529066085815,
      "learning_rate": 0.00019833333333333335,
      "loss": 6.5803,
      "step": 120
    },
    {
      "epoch": 0.04219727224775827,
      "grad_norm": 0.7910824418067932,
      "learning_rate": 0.00023166666666666667,
      "loss": 6.3863,
      "step": 140
    },
    {
      "epoch": 0.04822545399743802,
      "grad_norm": 1.0070773363113403,
      "learning_rate": 0.00026500000000000004,
      "loss": 6.2393,
      "step": 160
    },
    {
      "epoch": 0.05425363574711778,
      "grad_norm": 0.7891514897346497,
      "learning_rate": 0.00029833333333333334,
      "loss": 6.0613,
      "step": 180
    },
    {
      "epoch": 0.06028181749679753,
      "grad_norm": 0.8121163249015808,
      "learning_rate": 0.0003316666666666667,
      "loss": 5.912,
      "step": 200
    },
    {
      "epoch": 0.06630999924647728,
      "grad_norm": 1.0883384943008423,
      "learning_rate": 0.000365,
      "loss": 5.7569,
      "step": 220
    },
    {
      "epoch": 0.07233818099615703,
      "grad_norm": 0.8521329164505005,
      "learning_rate": 0.00039833333333333333,
      "loss": 5.6241,
      "step": 240
    },
    {
      "epoch": 0.07836636274583679,
      "grad_norm": 0.9475963115692139,
      "learning_rate": 0.0004316666666666667,
      "loss": 5.5001,
      "step": 260
    },
    {
      "epoch": 0.08439454449551655,
      "grad_norm": 0.5242730379104614,
      "learning_rate": 0.000465,
      "loss": 5.372,
      "step": 280
    },
    {
      "epoch": 0.09042272624519629,
      "grad_norm": 0.599181592464447,
      "learning_rate": 0.0004983333333333334,
      "loss": 5.2824,
      "step": 300
    },
    {
      "epoch": 0.09042272624519629,
      "eval_loss": 5.1891092537402255,
      "eval_perplexity": 179.30876316619444,
      "eval_runtime": 112.5481,
      "eval_samples_per_second": 14.74,
      "eval_steps_per_second": 0.231,
      "step": 300
    },
    {
      "epoch": 0.09645090799487605,
      "grad_norm": 0.34896278381347656,
      "learning_rate": 0.0004964814814814814,
      "loss": 5.1596,
      "step": 320
    },
    {
      "epoch": 0.1024790897445558,
      "grad_norm": 0.8048242330551147,
      "learning_rate": 0.0004927777777777777,
      "loss": 5.0746,
      "step": 340
    },
    {
      "epoch": 0.10850727149423556,
      "grad_norm": 0.3908897042274475,
      "learning_rate": 0.0004890740740740741,
      "loss": 5.0077,
      "step": 360
    },
    {
      "epoch": 0.1145354532439153,
      "grad_norm": 0.7325732111930847,
      "learning_rate": 0.0004853703703703704,
      "loss": 4.9149,
      "step": 380
    },
    {
      "epoch": 0.12056363499359506,
      "grad_norm": 0.3958197832107544,
      "learning_rate": 0.0004816666666666667,
      "loss": 4.8885,
      "step": 400
    },
    {
      "epoch": 0.1265918167432748,
      "grad_norm": 0.3405342996120453,
      "learning_rate": 0.00047796296296296297,
      "loss": 4.8009,
      "step": 420
    },
    {
      "epoch": 0.13261999849295456,
      "grad_norm": 0.6988040804862976,
      "learning_rate": 0.0004742592592592593,
      "loss": 4.7684,
      "step": 440
    },
    {
      "epoch": 0.1386481802426343,
      "grad_norm": 0.298909455537796,
      "learning_rate": 0.00047055555555555555,
      "loss": 4.7265,
      "step": 460
    },
    {
      "epoch": 0.14467636199231407,
      "grad_norm": 0.5806158185005188,
      "learning_rate": 0.00046685185185185187,
      "loss": 4.6676,
      "step": 480
    },
    {
      "epoch": 0.15070454374199382,
      "grad_norm": 0.5041508078575134,
      "learning_rate": 0.00046314814814814813,
      "loss": 4.6359,
      "step": 500
    },
    {
      "epoch": 0.15673272549167358,
      "grad_norm": 0.3164711594581604,
      "learning_rate": 0.00045944444444444445,
      "loss": 4.6046,
      "step": 520
    },
    {
      "epoch": 0.16276090724135334,
      "grad_norm": 0.7678843140602112,
      "learning_rate": 0.0004557407407407407,
      "loss": 4.5528,
      "step": 540
    },
    {
      "epoch": 0.1687890889910331,
      "grad_norm": 0.351477712392807,
      "learning_rate": 0.00045203703703703703,
      "loss": 4.5292,
      "step": 560
    },
    {
      "epoch": 0.17481727074071282,
      "grad_norm": 0.6850087642669678,
      "learning_rate": 0.0004483333333333333,
      "loss": 4.4822,
      "step": 580
    },
    {
      "epoch": 0.18084545249039258,
      "grad_norm": 0.7454901337623596,
      "learning_rate": 0.00044462962962962967,
      "loss": 4.4581,
      "step": 600
    },
    {
      "epoch": 0.18084545249039258,
      "eval_loss": 4.409127496012449,
      "eval_perplexity": 82.1977143744931,
      "eval_runtime": 112.2901,
      "eval_samples_per_second": 14.774,
      "eval_steps_per_second": 0.232,
      "step": 600
    },
    {
      "epoch": 0.18687363424007233,
      "grad_norm": 0.3134106397628784,
      "learning_rate": 0.00044092592592592594,
      "loss": 4.4146,
      "step": 620
    },
    {
      "epoch": 0.1929018159897521,
      "grad_norm": 0.5055473446846008,
      "learning_rate": 0.00043722222222222225,
      "loss": 4.3757,
      "step": 640
    },
    {
      "epoch": 0.19892999773943185,
      "grad_norm": 0.5237579941749573,
      "learning_rate": 0.0004335185185185185,
      "loss": 4.3453,
      "step": 660
    },
    {
      "epoch": 0.2049581794891116,
      "grad_norm": 0.398079514503479,
      "learning_rate": 0.00042981481481481484,
      "loss": 4.3276,
      "step": 680
    },
    {
      "epoch": 0.21098636123879136,
      "grad_norm": 0.4433188736438751,
      "learning_rate": 0.0004261111111111111,
      "loss": 4.2892,
      "step": 700
    },
    {
      "epoch": 0.21701454298847112,
      "grad_norm": 0.42719537019729614,
      "learning_rate": 0.0004224074074074074,
      "loss": 4.2615,
      "step": 720
    },
    {
      "epoch": 0.22304272473815084,
      "grad_norm": 0.4303497076034546,
      "learning_rate": 0.0004187037037037037,
      "loss": 4.233,
      "step": 740
    },
    {
      "epoch": 0.2290709064878306,
      "grad_norm": 0.5092538595199585,
      "learning_rate": 0.000415,
      "loss": 4.2113,
      "step": 760
    },
    {
      "epoch": 0.23509908823751036,
      "grad_norm": 0.6650151014328003,
      "learning_rate": 0.00041129629629629627,
      "loss": 4.204,
      "step": 780
    },
    {
      "epoch": 0.2411272699871901,
      "grad_norm": 0.43139395117759705,
      "learning_rate": 0.00040759259259259264,
      "loss": 4.1842,
      "step": 800
    },
    {
      "epoch": 0.24715545173686987,
      "grad_norm": 0.39379051327705383,
      "learning_rate": 0.0004038888888888889,
      "loss": 4.157,
      "step": 820
    },
    {
      "epoch": 0.2531836334865496,
      "grad_norm": 0.5419613718986511,
      "learning_rate": 0.0004001851851851852,
      "loss": 4.1432,
      "step": 840
    },
    {
      "epoch": 0.2592118152362294,
      "grad_norm": 0.3231978416442871,
      "learning_rate": 0.0003964814814814815,
      "loss": 4.1259,
      "step": 860
    },
    {
      "epoch": 0.2652399969859091,
      "grad_norm": 0.3947347104549408,
      "learning_rate": 0.0003927777777777778,
      "loss": 4.1225,
      "step": 880
    },
    {
      "epoch": 0.2712681787355889,
      "grad_norm": 0.4849476218223572,
      "learning_rate": 0.00038907407407407407,
      "loss": 4.0945,
      "step": 900
    },
    {
      "epoch": 0.2712681787355889,
      "eval_loss": 4.043287290524684,
      "eval_perplexity": 57.01345488110377,
      "eval_runtime": 112.1778,
      "eval_samples_per_second": 14.789,
      "eval_steps_per_second": 0.232,
      "step": 900
    },
    {
      "epoch": 0.2772963604852686,
      "grad_norm": 0.46728333830833435,
      "learning_rate": 0.0003853703703703704,
      "loss": 4.0828,
      "step": 920
    },
    {
      "epoch": 0.2833245422349484,
      "grad_norm": 0.3571609854698181,
      "learning_rate": 0.00038166666666666666,
      "loss": 4.0707,
      "step": 940
    },
    {
      "epoch": 0.28935272398462814,
      "grad_norm": 0.36861199140548706,
      "learning_rate": 0.000377962962962963,
      "loss": 4.0545,
      "step": 960
    },
    {
      "epoch": 0.29538090573430786,
      "grad_norm": 0.4048626124858856,
      "learning_rate": 0.00037425925925925924,
      "loss": 4.0493,
      "step": 980
    },
    {
      "epoch": 0.30140908748398765,
      "grad_norm": 0.41902783513069153,
      "learning_rate": 0.0003705555555555556,
      "loss": 4.0363,
      "step": 1000
    },
    {
      "epoch": 0.3074372692336674,
      "grad_norm": 0.4772008955478668,
      "learning_rate": 0.0003668518518518519,
      "loss": 4.0277,
      "step": 1020
    },
    {
      "epoch": 0.31346545098334716,
      "grad_norm": 0.4841766953468323,
      "learning_rate": 0.0003631481481481482,
      "loss": 4.0073,
      "step": 1040
    },
    {
      "epoch": 0.3194936327330269,
      "grad_norm": 0.4445819854736328,
      "learning_rate": 0.00035944444444444446,
      "loss": 4.0099,
      "step": 1060
    },
    {
      "epoch": 0.3255218144827067,
      "grad_norm": 0.31589964032173157,
      "learning_rate": 0.0003557407407407408,
      "loss": 3.9954,
      "step": 1080
    },
    {
      "epoch": 0.3315499962323864,
      "grad_norm": 0.32809144258499146,
      "learning_rate": 0.00035203703703703704,
      "loss": 3.9842,
      "step": 1100
    },
    {
      "epoch": 0.3375781779820662,
      "grad_norm": 0.5585801005363464,
      "learning_rate": 0.00034833333333333336,
      "loss": 3.9731,
      "step": 1120
    },
    {
      "epoch": 0.3436063597317459,
      "grad_norm": 0.33329740166664124,
      "learning_rate": 0.0003446296296296296,
      "loss": 3.9664,
      "step": 1140
    },
    {
      "epoch": 0.34963454148142564,
      "grad_norm": 0.39245834946632385,
      "learning_rate": 0.0003409259259259259,
      "loss": 3.9531,
      "step": 1160
    },
    {
      "epoch": 0.3556627232311054,
      "grad_norm": 0.34083107113838196,
      "learning_rate": 0.0003372222222222222,
      "loss": 3.9481,
      "step": 1180
    },
    {
      "epoch": 0.36169090498078516,
      "grad_norm": 0.363365113735199,
      "learning_rate": 0.0003335185185185185,
      "loss": 3.9454,
      "step": 1200
    },
    {
      "epoch": 0.36169090498078516,
      "eval_loss": 3.891763558483393,
      "eval_perplexity": 48.997219837139,
      "eval_runtime": 112.3031,
      "eval_samples_per_second": 14.773,
      "eval_steps_per_second": 0.232,
      "step": 1200
    },
    {
      "epoch": 0.36771908673046494,
      "grad_norm": 0.34547457098960876,
      "learning_rate": 0.00032981481481481485,
      "loss": 3.9326,
      "step": 1220
    },
    {
      "epoch": 0.37374726848014467,
      "grad_norm": 0.3942719101905823,
      "learning_rate": 0.0003261111111111111,
      "loss": 3.931,
      "step": 1240
    },
    {
      "epoch": 0.37977545022982445,
      "grad_norm": 0.41379085183143616,
      "learning_rate": 0.00032240740740740743,
      "loss": 3.9149,
      "step": 1260
    },
    {
      "epoch": 0.3858036319795042,
      "grad_norm": 0.3897305428981781,
      "learning_rate": 0.0003187037037037037,
      "loss": 3.915,
      "step": 1280
    },
    {
      "epoch": 0.3918318137291839,
      "grad_norm": 0.35454675555229187,
      "learning_rate": 0.000315,
      "loss": 3.9041,
      "step": 1300
    },
    {
      "epoch": 0.3978599954788637,
      "grad_norm": 0.40520045161247253,
      "learning_rate": 0.0003112962962962963,
      "loss": 3.8982,
      "step": 1320
    },
    {
      "epoch": 0.4038881772285434,
      "grad_norm": 0.37290430068969727,
      "learning_rate": 0.0003075925925925926,
      "loss": 3.8943,
      "step": 1340
    },
    {
      "epoch": 0.4099163589782232,
      "grad_norm": 0.36008208990097046,
      "learning_rate": 0.00030388888888888886,
      "loss": 3.8885,
      "step": 1360
    },
    {
      "epoch": 0.41594454072790293,
      "grad_norm": 0.4695078730583191,
      "learning_rate": 0.0003001851851851852,
      "loss": 3.8825,
      "step": 1380
    },
    {
      "epoch": 0.4219727224775827,
      "grad_norm": 0.3244253993034363,
      "learning_rate": 0.00029648148148148144,
      "loss": 3.8749,
      "step": 1400
    },
    {
      "epoch": 0.42800090422726245,
      "grad_norm": 0.3059616684913635,
      "learning_rate": 0.0002927777777777778,
      "loss": 3.8748,
      "step": 1420
    },
    {
      "epoch": 0.43402908597694223,
      "grad_norm": 0.421289324760437,
      "learning_rate": 0.0002890740740740741,
      "loss": 3.8617,
      "step": 1440
    },
    {
      "epoch": 0.44005726772662196,
      "grad_norm": 0.31943073868751526,
      "learning_rate": 0.0002853703703703704,
      "loss": 3.8665,
      "step": 1460
    },
    {
      "epoch": 0.4460854494763017,
      "grad_norm": 0.38908112049102783,
      "learning_rate": 0.00028166666666666666,
      "loss": 3.857,
      "step": 1480
    },
    {
      "epoch": 0.45211363122598147,
      "grad_norm": 0.3740065395832062,
      "learning_rate": 0.000277962962962963,
      "loss": 3.8522,
      "step": 1500
    },
    {
      "epoch": 0.45211363122598147,
      "eval_loss": 3.7993524888239567,
      "eval_perplexity": 44.67224934568574,
      "eval_runtime": 112.3077,
      "eval_samples_per_second": 14.772,
      "eval_steps_per_second": 0.232,
      "step": 1500
    },
    {
      "epoch": 0.4581418129756612,
      "grad_norm": 0.3724636435508728,
      "learning_rate": 0.00027425925925925925,
      "loss": 3.8458,
      "step": 1520
    },
    {
      "epoch": 0.464169994725341,
      "grad_norm": 0.34661316871643066,
      "learning_rate": 0.00027055555555555557,
      "loss": 3.844,
      "step": 1540
    },
    {
      "epoch": 0.4701981764750207,
      "grad_norm": 0.2814076244831085,
      "learning_rate": 0.00026685185185185183,
      "loss": 3.8319,
      "step": 1560
    },
    {
      "epoch": 0.4762263582247005,
      "grad_norm": 0.3638046383857727,
      "learning_rate": 0.00026314814814814815,
      "loss": 3.8301,
      "step": 1580
    },
    {
      "epoch": 0.4822545399743802,
      "grad_norm": 0.37232160568237305,
      "learning_rate": 0.0002594444444444444,
      "loss": 3.8169,
      "step": 1600
    },
    {
      "epoch": 0.48828272172405995,
      "grad_norm": 0.4013717472553253,
      "learning_rate": 0.0002557407407407408,
      "loss": 3.8263,
      "step": 1620
    },
    {
      "epoch": 0.49431090347373974,
      "grad_norm": 0.4506720304489136,
      "learning_rate": 0.00025203703703703705,
      "loss": 3.8174,
      "step": 1640
    },
    {
      "epoch": 0.5003390852234195,
      "grad_norm": 0.2753377854824066,
      "learning_rate": 0.0002483333333333333,
      "loss": 3.8134,
      "step": 1660
    },
    {
      "epoch": 0.5063672669730992,
      "grad_norm": 0.3540724813938141,
      "learning_rate": 0.00024462962962962963,
      "loss": 3.8118,
      "step": 1680
    },
    {
      "epoch": 0.512395448722779,
      "grad_norm": 0.35745769739151,
      "learning_rate": 0.00024092592592592593,
      "loss": 3.7999,
      "step": 1700
    },
    {
      "epoch": 0.5184236304724588,
      "grad_norm": 0.3902524411678314,
      "learning_rate": 0.00023722222222222222,
      "loss": 3.8004,
      "step": 1720
    },
    {
      "epoch": 0.5244518122221385,
      "grad_norm": 0.3244744539260864,
      "learning_rate": 0.0002335185185185185,
      "loss": 3.7993,
      "step": 1740
    },
    {
      "epoch": 0.5304799939718182,
      "grad_norm": 0.34227174520492554,
      "learning_rate": 0.0002298148148148148,
      "loss": 3.7969,
      "step": 1760
    },
    {
      "epoch": 0.536508175721498,
      "grad_norm": 0.349168598651886,
      "learning_rate": 0.00022611111111111112,
      "loss": 3.7867,
      "step": 1780
    },
    {
      "epoch": 0.5425363574711778,
      "grad_norm": 0.46753671765327454,
      "learning_rate": 0.0002224074074074074,
      "loss": 3.7903,
      "step": 1800
    },
    {
      "epoch": 0.5425363574711778,
      "eval_loss": 3.7372125118654314,
      "eval_perplexity": 41.980805917649974,
      "eval_runtime": 112.2567,
      "eval_samples_per_second": 14.779,
      "eval_steps_per_second": 0.232,
      "step": 1800
    },
    {
      "epoch": 0.5485645392208575,
      "grad_norm": 0.34198659658432007,
      "learning_rate": 0.0002187037037037037,
      "loss": 3.7876,
      "step": 1820
    },
    {
      "epoch": 0.5545927209705372,
      "grad_norm": 0.37303704023361206,
      "learning_rate": 0.000215,
      "loss": 3.7787,
      "step": 1840
    },
    {
      "epoch": 0.560620902720217,
      "grad_norm": 0.3005342185497284,
      "learning_rate": 0.00021129629629629629,
      "loss": 3.7718,
      "step": 1860
    },
    {
      "epoch": 0.5666490844698968,
      "grad_norm": 0.29448869824409485,
      "learning_rate": 0.0002075925925925926,
      "loss": 3.7724,
      "step": 1880
    },
    {
      "epoch": 0.5726772662195765,
      "grad_norm": 0.29300233721733093,
      "learning_rate": 0.0002038888888888889,
      "loss": 3.7652,
      "step": 1900
    },
    {
      "epoch": 0.5787054479692563,
      "grad_norm": 0.3517768681049347,
      "learning_rate": 0.0002001851851851852,
      "loss": 3.7675,
      "step": 1920
    },
    {
      "epoch": 0.584733629718936,
      "grad_norm": 0.31965792179107666,
      "learning_rate": 0.00019648148148148148,
      "loss": 3.7615,
      "step": 1940
    },
    {
      "epoch": 0.5907618114686157,
      "grad_norm": 0.29577893018722534,
      "learning_rate": 0.00019277777777777777,
      "loss": 3.7634,
      "step": 1960
    },
    {
      "epoch": 0.5967899932182955,
      "grad_norm": 0.33181506395339966,
      "learning_rate": 0.00018907407407407406,
      "loss": 3.7513,
      "step": 1980
    },
    {
      "epoch": 0.6028181749679753,
      "grad_norm": 0.3156045377254486,
      "learning_rate": 0.00018537037037037038,
      "loss": 3.7533,
      "step": 2000
    },
    {
      "epoch": 0.6088463567176551,
      "grad_norm": 0.3246684670448303,
      "learning_rate": 0.00018166666666666667,
      "loss": 3.7491,
      "step": 2020
    },
    {
      "epoch": 0.6148745384673348,
      "grad_norm": 0.2846499979496002,
      "learning_rate": 0.00017796296296296296,
      "loss": 3.7443,
      "step": 2040
    },
    {
      "epoch": 0.6209027202170145,
      "grad_norm": 0.34447523951530457,
      "learning_rate": 0.00017425925925925926,
      "loss": 3.7454,
      "step": 2060
    },
    {
      "epoch": 0.6269309019666943,
      "grad_norm": 0.3255724310874939,
      "learning_rate": 0.00017055555555555555,
      "loss": 3.7437,
      "step": 2080
    },
    {
      "epoch": 0.632959083716374,
      "grad_norm": 0.2995988130569458,
      "learning_rate": 0.00016685185185185187,
      "loss": 3.7384,
      "step": 2100
    },
    {
      "epoch": 0.632959083716374,
      "eval_loss": 3.6879229315555366,
      "eval_perplexity": 39.96175739053913,
      "eval_runtime": 112.2917,
      "eval_samples_per_second": 14.774,
      "eval_steps_per_second": 0.232,
      "step": 2100
    },
    {
      "epoch": 0.6389872654660538,
      "grad_norm": 0.2952309846878052,
      "learning_rate": 0.00016314814814814816,
      "loss": 3.7376,
      "step": 2120
    },
    {
      "epoch": 0.6450154472157336,
      "grad_norm": 0.28378739953041077,
      "learning_rate": 0.00015944444444444445,
      "loss": 3.7361,
      "step": 2140
    },
    {
      "epoch": 0.6510436289654133,
      "grad_norm": 0.3023880124092102,
      "learning_rate": 0.00015574074074074074,
      "loss": 3.7336,
      "step": 2160
    },
    {
      "epoch": 0.657071810715093,
      "grad_norm": 0.2970915138721466,
      "learning_rate": 0.00015203703703703703,
      "loss": 3.7325,
      "step": 2180
    },
    {
      "epoch": 0.6630999924647728,
      "grad_norm": 0.3934483230113983,
      "learning_rate": 0.00014833333333333335,
      "loss": 3.7294,
      "step": 2200
    },
    {
      "epoch": 0.6691281742144526,
      "grad_norm": 0.3007951080799103,
      "learning_rate": 0.00014462962962962964,
      "loss": 3.7264,
      "step": 2220
    },
    {
      "epoch": 0.6751563559641324,
      "grad_norm": 0.2756880819797516,
      "learning_rate": 0.00014092592592592594,
      "loss": 3.7251,
      "step": 2240
    },
    {
      "epoch": 0.681184537713812,
      "grad_norm": 0.264085978269577,
      "learning_rate": 0.00013722222222222223,
      "loss": 3.7122,
      "step": 2260
    },
    {
      "epoch": 0.6872127194634918,
      "grad_norm": 0.2801743745803833,
      "learning_rate": 0.00013351851851851852,
      "loss": 3.7193,
      "step": 2280
    },
    {
      "epoch": 0.6932409012131716,
      "grad_norm": 0.32809194922447205,
      "learning_rate": 0.00012981481481481484,
      "loss": 3.7126,
      "step": 2300
    },
    {
      "epoch": 0.6992690829628513,
      "grad_norm": 0.27438148856163025,
      "learning_rate": 0.00012611111111111113,
      "loss": 3.7063,
      "step": 2320
    },
    {
      "epoch": 0.7052972647125311,
      "grad_norm": 0.25681811571121216,
      "learning_rate": 0.00012240740740740742,
      "loss": 3.7085,
      "step": 2340
    },
    {
      "epoch": 0.7113254464622109,
      "grad_norm": 0.2568652331829071,
      "learning_rate": 0.0001187037037037037,
      "loss": 3.7076,
      "step": 2360
    },
    {
      "epoch": 0.7173536282118906,
      "grad_norm": 0.2873578369617462,
      "learning_rate": 0.000115,
      "loss": 3.6987,
      "step": 2380
    },
    {
      "epoch": 0.7233818099615703,
      "grad_norm": 0.26023069024086,
      "learning_rate": 0.0001112962962962963,
      "loss": 3.707,
      "step": 2400
    },
    {
      "epoch": 0.7233818099615703,
      "eval_loss": 3.653235740343999,
      "eval_perplexity": 38.599361710783484,
      "eval_runtime": 112.4066,
      "eval_samples_per_second": 14.759,
      "eval_steps_per_second": 0.231,
      "step": 2400
    },
    {
      "epoch": 0.7294099917112501,
      "grad_norm": 0.2499583661556244,
      "learning_rate": 0.00010759259259259259,
      "loss": 3.7033,
      "step": 2420
    },
    {
      "epoch": 0.7354381734609299,
      "grad_norm": 0.2881416380405426,
      "learning_rate": 0.00010388888888888889,
      "loss": 3.7016,
      "step": 2440
    },
    {
      "epoch": 0.7414663552106096,
      "grad_norm": 0.2542083263397217,
      "learning_rate": 0.00010018518518518518,
      "loss": 3.7015,
      "step": 2460
    },
    {
      "epoch": 0.7474945369602893,
      "grad_norm": 0.2287365347146988,
      "learning_rate": 9.648148148148149e-05,
      "loss": 3.6942,
      "step": 2480
    },
    {
      "epoch": 0.7535227187099691,
      "grad_norm": 0.24961161613464355,
      "learning_rate": 9.277777777777778e-05,
      "loss": 3.6952,
      "step": 2500
    },
    {
      "epoch": 0.7595509004596489,
      "grad_norm": 0.26621654629707336,
      "learning_rate": 8.907407407407407e-05,
      "loss": 3.6942,
      "step": 2520
    },
    {
      "epoch": 0.7655790822093286,
      "grad_norm": 0.27608931064605713,
      "learning_rate": 8.537037037037038e-05,
      "loss": 3.6937,
      "step": 2540
    },
    {
      "epoch": 0.7716072639590084,
      "grad_norm": 0.23405340313911438,
      "learning_rate": 8.166666666666667e-05,
      "loss": 3.6889,
      "step": 2560
    },
    {
      "epoch": 0.7776354457086881,
      "grad_norm": 0.22344228625297546,
      "learning_rate": 7.796296296296296e-05,
      "loss": 3.6875,
      "step": 2580
    },
    {
      "epoch": 0.7836636274583678,
      "grad_norm": 0.25090643763542175,
      "learning_rate": 7.425925925925927e-05,
      "loss": 3.6819,
      "step": 2600
    },
    {
      "epoch": 0.7896918092080476,
      "grad_norm": 0.24225394427776337,
      "learning_rate": 7.055555555555556e-05,
      "loss": 3.6878,
      "step": 2620
    },
    {
      "epoch": 0.7957199909577274,
      "grad_norm": 0.25418150424957275,
      "learning_rate": 6.685185185185186e-05,
      "loss": 3.6849,
      "step": 2640
    },
    {
      "epoch": 0.8017481727074072,
      "grad_norm": 0.2390448898077011,
      "learning_rate": 6.314814814814815e-05,
      "loss": 3.6844,
      "step": 2660
    },
    {
      "epoch": 0.8077763544570868,
      "grad_norm": 0.2440630942583084,
      "learning_rate": 5.9444444444444445e-05,
      "loss": 3.6768,
      "step": 2680
    },
    {
      "epoch": 0.8138045362067666,
      "grad_norm": 0.22975711524486542,
      "learning_rate": 5.5740740740740744e-05,
      "loss": 3.6796,
      "step": 2700
    },
    {
      "epoch": 0.8138045362067666,
      "eval_loss": 3.628363446560925,
      "eval_perplexity": 37.6511480531639,
      "eval_runtime": 112.169,
      "eval_samples_per_second": 14.79,
      "eval_steps_per_second": 0.232,
      "step": 2700
    },
    {
      "epoch": 0.8198327179564464,
      "grad_norm": 0.21335391700267792,
      "learning_rate": 5.2037037037037035e-05,
      "loss": 3.6776,
      "step": 2720
    },
    {
      "epoch": 0.8258608997061262,
      "grad_norm": 0.2257867455482483,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 3.6801,
      "step": 2740
    },
    {
      "epoch": 0.8318890814558059,
      "grad_norm": 0.2070840299129486,
      "learning_rate": 4.462962962962963e-05,
      "loss": 3.6754,
      "step": 2760
    },
    {
      "epoch": 0.8379172632054857,
      "grad_norm": 0.20770543813705444,
      "learning_rate": 4.092592592592593e-05,
      "loss": 3.6732,
      "step": 2780
    },
    {
      "epoch": 0.8439454449551654,
      "grad_norm": 0.23169459402561188,
      "learning_rate": 3.722222222222222e-05,
      "loss": 3.6754,
      "step": 2800
    },
    {
      "epoch": 0.8499736267048451,
      "grad_norm": 0.1974688619375229,
      "learning_rate": 3.351851851851852e-05,
      "loss": 3.6777,
      "step": 2820
    },
    {
      "epoch": 0.8560018084545249,
      "grad_norm": 0.2034008651971817,
      "learning_rate": 2.9814814814814815e-05,
      "loss": 3.67,
      "step": 2840
    },
    {
      "epoch": 0.8620299902042047,
      "grad_norm": 0.2013922780752182,
      "learning_rate": 2.6111111111111114e-05,
      "loss": 3.6671,
      "step": 2860
    },
    {
      "epoch": 0.8680581719538845,
      "grad_norm": 0.2004835456609726,
      "learning_rate": 2.240740740740741e-05,
      "loss": 3.6717,
      "step": 2880
    },
    {
      "epoch": 0.8740863537035641,
      "grad_norm": 0.19306205213069916,
      "learning_rate": 1.8703703703703707e-05,
      "loss": 3.6664,
      "step": 2900
    },
    {
      "epoch": 0.8801145354532439,
      "grad_norm": 0.19328537583351135,
      "learning_rate": 1.5e-05,
      "loss": 3.6711,
      "step": 2920
    },
    {
      "epoch": 0.8861427172029237,
      "grad_norm": 0.18284295499324799,
      "learning_rate": 1.1296296296296295e-05,
      "loss": 3.6616,
      "step": 2940
    },
    {
      "epoch": 0.8921708989526034,
      "grad_norm": 0.18400640785694122,
      "learning_rate": 7.592592592592593e-06,
      "loss": 3.658,
      "step": 2960
    },
    {
      "epoch": 0.8981990807022832,
      "grad_norm": 0.1778123676776886,
      "learning_rate": 3.888888888888889e-06,
      "loss": 3.6655,
      "step": 2980
    },
    {
      "epoch": 0.9042272624519629,
      "grad_norm": 0.17845961451530457,
      "learning_rate": 1.8518518518518518e-07,
      "loss": 3.6693,
      "step": 3000
    },
    {
      "epoch": 0.9042272624519629,
      "eval_loss": 3.614897001279198,
      "eval_perplexity": 37.14751958159777,
      "eval_runtime": 112.2672,
      "eval_samples_per_second": 14.777,
      "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": 1.404194997141504e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}