{
  "best_metric": 2.411439895629883,
  "best_model_checkpoint": "miner_id_24/checkpoint-200",
  "epoch": 0.03192338387869114,
  "eval_steps": 50,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0001596169193934557,
      "grad_norm": 68.28546905517578,
      "learning_rate": 8.000000000000001e-06,
      "loss": 15.393,
      "step": 1
    },
    {
      "epoch": 0.0001596169193934557,
      "eval_loss": 4.64123010635376,
      "eval_runtime": 260.6717,
      "eval_samples_per_second": 40.48,
      "eval_steps_per_second": 10.12,
      "step": 1
    },
    {
      "epoch": 0.0003192338387869114,
      "grad_norm": 29.318359375,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 17.0128,
      "step": 2
    },
    {
      "epoch": 0.00047885075818036713,
      "grad_norm": 28.234766006469727,
      "learning_rate": 2.4e-05,
      "loss": 16.7677,
      "step": 3
    },
    {
      "epoch": 0.0006384676775738228,
      "grad_norm": 27.14069938659668,
      "learning_rate": 3.2000000000000005e-05,
      "loss": 17.0856,
      "step": 4
    },
    {
      "epoch": 0.0007980845969672786,
      "grad_norm": 25.58334732055664,
      "learning_rate": 4e-05,
      "loss": 15.872,
      "step": 5
    },
    {
      "epoch": 0.0009577015163607343,
      "grad_norm": 23.222209930419922,
      "learning_rate": 4.8e-05,
      "loss": 15.4543,
      "step": 6
    },
    {
      "epoch": 0.0011173184357541898,
      "grad_norm": 18.329936981201172,
      "learning_rate": 5.6e-05,
      "loss": 14.7361,
      "step": 7
    },
    {
      "epoch": 0.0012769353551476455,
      "grad_norm": 14.219144821166992,
      "learning_rate": 6.400000000000001e-05,
      "loss": 13.0282,
      "step": 8
    },
    {
      "epoch": 0.0014365522745411015,
      "grad_norm": 14.171404838562012,
      "learning_rate": 7.2e-05,
      "loss": 13.8074,
      "step": 9
    },
    {
      "epoch": 0.0015961691939345571,
      "grad_norm": 15.184477806091309,
      "learning_rate": 8e-05,
      "loss": 13.6936,
      "step": 10
    },
    {
      "epoch": 0.0017557861133280128,
      "grad_norm": 15.673888206481934,
      "learning_rate": 7.999453219969877e-05,
      "loss": 13.8008,
      "step": 11
    },
    {
      "epoch": 0.0019154030327214685,
      "grad_norm": 16.34881019592285,
      "learning_rate": 7.997813029363704e-05,
      "loss": 13.7255,
      "step": 12
    },
    {
      "epoch": 0.0020750199521149242,
      "grad_norm": 14.588293075561523,
      "learning_rate": 7.99507987659322e-05,
      "loss": 12.3483,
      "step": 13
    },
    {
      "epoch": 0.0022346368715083797,
      "grad_norm": 13.352673530578613,
      "learning_rate": 7.991254508875098e-05,
      "loss": 12.6878,
      "step": 14
    },
    {
      "epoch": 0.0023942537909018356,
      "grad_norm": 14.82767105102539,
      "learning_rate": 7.98633797202668e-05,
      "loss": 13.222,
      "step": 15
    },
    {
      "epoch": 0.002553870710295291,
      "grad_norm": 12.547225952148438,
      "learning_rate": 7.980331610180046e-05,
      "loss": 12.7402,
      "step": 16
    },
    {
      "epoch": 0.002713487629688747,
      "grad_norm": 12.519619941711426,
      "learning_rate": 7.973237065414553e-05,
      "loss": 11.6347,
      "step": 17
    },
    {
      "epoch": 0.002873104549082203,
      "grad_norm": 13.17307186126709,
      "learning_rate": 7.965056277307902e-05,
      "loss": 12.6509,
      "step": 18
    },
    {
      "epoch": 0.0030327214684756584,
      "grad_norm": 12.776773452758789,
      "learning_rate": 7.955791482405875e-05,
      "loss": 11.2834,
      "step": 19
    },
    {
      "epoch": 0.0031923383878691143,
      "grad_norm": 12.828487396240234,
      "learning_rate": 7.94544521361089e-05,
      "loss": 11.4479,
      "step": 20
    },
    {
      "epoch": 0.0033519553072625698,
      "grad_norm": 13.230494499206543,
      "learning_rate": 7.93402029948953e-05,
      "loss": 11.7996,
      "step": 21
    },
    {
      "epoch": 0.0035115722266560257,
      "grad_norm": 13.53295612335205,
      "learning_rate": 7.921519863499239e-05,
      "loss": 12.8945,
      "step": 22
    },
    {
      "epoch": 0.003671189146049481,
      "grad_norm": 12.507688522338867,
      "learning_rate": 7.907947323134398e-05,
      "loss": 11.3031,
      "step": 23
    },
    {
      "epoch": 0.003830806065442937,
      "grad_norm": 12.64974594116211,
      "learning_rate": 7.893306388992023e-05,
      "loss": 11.6757,
      "step": 24
    },
    {
      "epoch": 0.0039904229848363925,
      "grad_norm": 11.937615394592285,
      "learning_rate": 7.877601063757323e-05,
      "loss": 11.6896,
      "step": 25
    },
    {
      "epoch": 0.0041500399042298484,
      "grad_norm": 12.201370239257812,
      "learning_rate": 7.860835641109395e-05,
      "loss": 11.497,
      "step": 26
    },
    {
      "epoch": 0.004309656823623304,
      "grad_norm": 12.877632141113281,
      "learning_rate": 7.843014704547393e-05,
      "loss": 11.4197,
      "step": 27
    },
    {
      "epoch": 0.004469273743016759,
      "grad_norm": 13.948447227478027,
      "learning_rate": 7.824143126137431e-05,
      "loss": 12.0349,
      "step": 28
    },
    {
      "epoch": 0.004628890662410215,
      "grad_norm": 11.921418190002441,
      "learning_rate": 7.804226065180615e-05,
      "loss": 10.664,
      "step": 29
    },
    {
      "epoch": 0.004788507581803671,
      "grad_norm": 14.005268096923828,
      "learning_rate": 7.783268966802539e-05,
      "loss": 12.3096,
      "step": 30
    },
    {
      "epoch": 0.004948124501197127,
      "grad_norm": 13.047632217407227,
      "learning_rate": 7.761277560464645e-05,
      "loss": 11.9234,
      "step": 31
    },
    {
      "epoch": 0.005107741420590582,
      "grad_norm": 12.718923568725586,
      "learning_rate": 7.738257858397844e-05,
      "loss": 10.5216,
      "step": 32
    },
    {
      "epoch": 0.005267358339984038,
      "grad_norm": 13.36783504486084,
      "learning_rate": 7.71421615395883e-05,
      "loss": 12.4396,
      "step": 33
    },
    {
      "epoch": 0.005426975259377494,
      "grad_norm": 13.024250984191895,
      "learning_rate": 7.68915901990954e-05,
      "loss": 9.8002,
      "step": 34
    },
    {
      "epoch": 0.00558659217877095,
      "grad_norm": 13.97553539276123,
      "learning_rate": 7.663093306620231e-05,
      "loss": 10.7099,
      "step": 35
    },
    {
      "epoch": 0.005746209098164406,
      "grad_norm": 14.469559669494629,
      "learning_rate": 7.636026140196651e-05,
      "loss": 11.3811,
      "step": 36
    },
    {
      "epoch": 0.005905826017557861,
      "grad_norm": 13.011373519897461,
      "learning_rate": 7.607964920531837e-05,
      "loss": 9.1886,
      "step": 37
    },
    {
      "epoch": 0.006065442936951317,
      "grad_norm": 14.455774307250977,
      "learning_rate": 7.578917319283055e-05,
      "loss": 11.2866,
      "step": 38
    },
    {
      "epoch": 0.006225059856344773,
      "grad_norm": 15.614158630371094,
      "learning_rate": 7.548891277774448e-05,
      "loss": 10.5648,
      "step": 39
    },
    {
      "epoch": 0.006384676775738229,
      "grad_norm": 15.963403701782227,
      "learning_rate": 7.517895004825956e-05,
      "loss": 10.5252,
      "step": 40
    },
    {
      "epoch": 0.006544293695131684,
      "grad_norm": 16.46887969970703,
      "learning_rate": 7.48593697450911e-05,
      "loss": 10.1787,
      "step": 41
    },
    {
      "epoch": 0.0067039106145251395,
      "grad_norm": 15.700961112976074,
      "learning_rate": 7.453025923830296e-05,
      "loss": 9.1733,
      "step": 42
    },
    {
      "epoch": 0.0068635275339185954,
      "grad_norm": 18.935749053955078,
      "learning_rate": 7.419170850342156e-05,
      "loss": 11.4134,
      "step": 43
    },
    {
      "epoch": 0.007023144453312051,
      "grad_norm": 20.495908737182617,
      "learning_rate": 7.384381009683742e-05,
      "loss": 11.4342,
      "step": 44
    },
    {
      "epoch": 0.007182761372705506,
      "grad_norm": 19.723838806152344,
      "learning_rate": 7.348665913050115e-05,
      "loss": 10.7341,
      "step": 45
    },
    {
      "epoch": 0.007342378292098962,
      "grad_norm": 21.278959274291992,
      "learning_rate": 7.312035324592081e-05,
      "loss": 10.5839,
      "step": 46
    },
    {
      "epoch": 0.007501995211492418,
      "grad_norm": 21.758159637451172,
      "learning_rate": 7.274499258746771e-05,
      "loss": 9.8437,
      "step": 47
    },
    {
      "epoch": 0.007661612130885874,
      "grad_norm": 21.18351936340332,
      "learning_rate": 7.236067977499791e-05,
      "loss": 10.0888,
      "step": 48
    },
    {
      "epoch": 0.00782122905027933,
      "grad_norm": 26.086111068725586,
      "learning_rate": 7.196751987579699e-05,
      "loss": 10.447,
      "step": 49
    },
    {
      "epoch": 0.007980845969672785,
      "grad_norm": 27.390546798706055,
      "learning_rate": 7.156562037585576e-05,
      "loss": 11.6472,
      "step": 50
    },
    {
      "epoch": 0.007980845969672785,
      "eval_loss": 2.9036827087402344,
      "eval_runtime": 260.6485,
      "eval_samples_per_second": 40.484,
      "eval_steps_per_second": 10.121,
      "step": 50
    },
    {
      "epoch": 0.008140462889066242,
      "grad_norm": 40.494049072265625,
      "learning_rate": 7.11550911504845e-05,
      "loss": 13.9192,
      "step": 51
    },
    {
      "epoch": 0.008300079808459697,
      "grad_norm": 30.184207916259766,
      "learning_rate": 7.073604443427437e-05,
      "loss": 13.2728,
      "step": 52
    },
    {
      "epoch": 0.008459696727853152,
      "grad_norm": 20.91021156311035,
      "learning_rate": 7.03085947904134e-05,
      "loss": 12.3766,
      "step": 53
    },
    {
      "epoch": 0.008619313647246609,
      "grad_norm": 15.238024711608887,
      "learning_rate": 6.987285907936617e-05,
      "loss": 11.4377,
      "step": 54
    },
    {
      "epoch": 0.008778930566640064,
      "grad_norm": 12.432860374450684,
      "learning_rate": 6.942895642692527e-05,
      "loss": 11.3605,
      "step": 55
    },
    {
      "epoch": 0.008938547486033519,
      "grad_norm": 12.549322128295898,
      "learning_rate": 6.897700819164357e-05,
      "loss": 11.8533,
      "step": 56
    },
    {
      "epoch": 0.009098164405426976,
      "grad_norm": 10.934309005737305,
      "learning_rate": 6.851713793165589e-05,
      "loss": 10.6387,
      "step": 57
    },
    {
      "epoch": 0.00925778132482043,
      "grad_norm": 11.76457405090332,
      "learning_rate": 6.804947137089955e-05,
      "loss": 11.8281,
      "step": 58
    },
    {
      "epoch": 0.009417398244213887,
      "grad_norm": 12.140151023864746,
      "learning_rate": 6.757413636474263e-05,
      "loss": 10.7103,
      "step": 59
    },
    {
      "epoch": 0.009577015163607342,
      "grad_norm": 12.146190643310547,
      "learning_rate": 6.709126286502965e-05,
      "loss": 11.8088,
      "step": 60
    },
    {
      "epoch": 0.009736632083000797,
      "grad_norm": 10.072521209716797,
      "learning_rate": 6.660098288455393e-05,
      "loss": 9.4006,
      "step": 61
    },
    {
      "epoch": 0.009896249002394254,
      "grad_norm": 11.379202842712402,
      "learning_rate": 6.610343046096674e-05,
      "loss": 9.9289,
      "step": 62
    },
    {
      "epoch": 0.01005586592178771,
      "grad_norm": 11.258583068847656,
      "learning_rate": 6.559874162013267e-05,
      "loss": 10.5815,
      "step": 63
    },
    {
      "epoch": 0.010215482841181164,
      "grad_norm": 11.229110717773438,
      "learning_rate": 6.508705433894149e-05,
      "loss": 10.3872,
      "step": 64
    },
    {
      "epoch": 0.010375099760574621,
      "grad_norm": 10.608044624328613,
      "learning_rate": 6.456850850758673e-05,
      "loss": 10.0493,
      "step": 65
    },
    {
      "epoch": 0.010534716679968076,
      "grad_norm": 11.73746395111084,
      "learning_rate": 6.404324589132101e-05,
      "loss": 10.4021,
      "step": 66
    },
    {
      "epoch": 0.010694333599361533,
      "grad_norm": 11.122593879699707,
      "learning_rate": 6.351141009169893e-05,
      "loss": 9.8319,
      "step": 67
    },
    {
      "epoch": 0.010853950518754988,
      "grad_norm": 12.417786598205566,
      "learning_rate": 6.297314650731775e-05,
      "loss": 10.5185,
      "step": 68
    },
    {
      "epoch": 0.011013567438148443,
      "grad_norm": 12.157649040222168,
      "learning_rate": 6.242860229406692e-05,
      "loss": 10.6265,
      "step": 69
    },
    {
      "epoch": 0.0111731843575419,
      "grad_norm": 12.59546184539795,
      "learning_rate": 6.18779263248971e-05,
      "loss": 10.6863,
      "step": 70
    },
    {
      "epoch": 0.011332801276935355,
      "grad_norm": 12.086047172546387,
      "learning_rate": 6.132126914911976e-05,
      "loss": 10.7338,
      "step": 71
    },
    {
      "epoch": 0.011492418196328812,
      "grad_norm": 12.330252647399902,
      "learning_rate": 6.075878295124861e-05,
      "loss": 10.9993,
      "step": 72
    },
    {
      "epoch": 0.011652035115722267,
      "grad_norm": 12.059850692749023,
      "learning_rate": 6.019062150939376e-05,
      "loss": 10.3537,
      "step": 73
    },
    {
      "epoch": 0.011811652035115722,
      "grad_norm": 12.277752876281738,
      "learning_rate": 5.9616940153220336e-05,
      "loss": 10.5273,
      "step": 74
    },
    {
      "epoch": 0.011971268954509178,
      "grad_norm": 12.278406143188477,
      "learning_rate": 5.903789572148295e-05,
      "loss": 9.6356,
      "step": 75
    },
    {
      "epoch": 0.012130885873902634,
      "grad_norm": 12.826895713806152,
      "learning_rate": 5.845364651914752e-05,
      "loss": 10.0302,
      "step": 76
    },
    {
      "epoch": 0.012290502793296089,
      "grad_norm": 13.393589973449707,
      "learning_rate": 5.786435227411227e-05,
      "loss": 10.6069,
      "step": 77
    },
    {
      "epoch": 0.012450119712689545,
      "grad_norm": 14.5519437789917,
      "learning_rate": 5.727017409353971e-05,
      "loss": 10.3158,
      "step": 78
    },
    {
      "epoch": 0.012609736632083,
      "grad_norm": 13.863892555236816,
      "learning_rate": 5.667127441981162e-05,
      "loss": 10.9629,
      "step": 79
    },
    {
      "epoch": 0.012769353551476457,
      "grad_norm": 13.034372329711914,
      "learning_rate": 5.606781698611879e-05,
      "loss": 10.0744,
      "step": 80
    },
    {
      "epoch": 0.012928970470869912,
      "grad_norm": 14.546815872192383,
      "learning_rate": 5.5459966771698096e-05,
      "loss": 10.2987,
      "step": 81
    },
    {
      "epoch": 0.013088587390263367,
      "grad_norm": 13.958318710327148,
      "learning_rate": 5.4847889956728834e-05,
      "loss": 10.9028,
      "step": 82
    },
    {
      "epoch": 0.013248204309656824,
      "grad_norm": 13.90129280090332,
      "learning_rate": 5.423175387690067e-05,
      "loss": 9.723,
      "step": 83
    },
    {
      "epoch": 0.013407821229050279,
      "grad_norm": 17.18065071105957,
      "learning_rate": 5.361172697766573e-05,
      "loss": 11.1042,
      "step": 84
    },
    {
      "epoch": 0.013567438148443736,
      "grad_norm": 14.979351043701172,
      "learning_rate": 5.298797876818735e-05,
      "loss": 11.4626,
      "step": 85
    },
    {
      "epoch": 0.013727055067837191,
      "grad_norm": 15.190160751342773,
      "learning_rate": 5.23606797749979e-05,
      "loss": 10.5648,
      "step": 86
    },
    {
      "epoch": 0.013886671987230646,
      "grad_norm": 15.130168914794922,
      "learning_rate": 5.17300014953786e-05,
      "loss": 11.0131,
      "step": 87
    },
    {
      "epoch": 0.014046288906624103,
      "grad_norm": 16.177791595458984,
      "learning_rate": 5.109611635047379e-05,
      "loss": 9.2829,
      "step": 88
    },
    {
      "epoch": 0.014205905826017558,
      "grad_norm": 15.867997169494629,
      "learning_rate": 5.04591976381528e-05,
      "loss": 10.1047,
      "step": 89
    },
    {
      "epoch": 0.014365522745411013,
      "grad_norm": 16.378780364990234,
      "learning_rate": 4.981941948563197e-05,
      "loss": 10.1567,
      "step": 90
    },
    {
      "epoch": 0.01452513966480447,
      "grad_norm": 15.733001708984375,
      "learning_rate": 4.9176956801870065e-05,
      "loss": 9.6687,
      "step": 91
    },
    {
      "epoch": 0.014684756584197925,
      "grad_norm": 16.034528732299805,
      "learning_rate": 4.853198522974988e-05,
      "loss": 9.8548,
      "step": 92
    },
    {
      "epoch": 0.014844373503591381,
      "grad_norm": 16.750699996948242,
      "learning_rate": 4.788468109805921e-05,
      "loss": 9.8802,
      "step": 93
    },
    {
      "epoch": 0.015003990422984836,
      "grad_norm": 16.891605377197266,
      "learning_rate": 4.7235221373284407e-05,
      "loss": 9.8121,
      "step": 94
    },
    {
      "epoch": 0.015163607342378291,
      "grad_norm": 20.089872360229492,
      "learning_rate": 4.658378361122936e-05,
      "loss": 11.083,
      "step": 95
    },
    {
      "epoch": 0.015323224261771748,
      "grad_norm": 18.97275733947754,
      "learning_rate": 4.593054590847368e-05,
      "loss": 8.2899,
      "step": 96
    },
    {
      "epoch": 0.015482841181165203,
      "grad_norm": 24.740419387817383,
      "learning_rate": 4.5275686853682765e-05,
      "loss": 9.7659,
      "step": 97
    },
    {
      "epoch": 0.01564245810055866,
      "grad_norm": 22.62858009338379,
      "learning_rate": 4.4619385478783456e-05,
      "loss": 8.7437,
      "step": 98
    },
    {
      "epoch": 0.015802075019952115,
      "grad_norm": 24.447660446166992,
      "learning_rate": 4.396182121001852e-05,
      "loss": 9.3997,
      "step": 99
    },
    {
      "epoch": 0.01596169193934557,
      "grad_norm": 25.942359924316406,
      "learning_rate": 4.33031738188933e-05,
      "loss": 9.2371,
      "step": 100
    },
    {
      "epoch": 0.01596169193934557,
      "eval_loss": 2.5352368354797363,
      "eval_runtime": 260.7157,
      "eval_samples_per_second": 40.473,
      "eval_steps_per_second": 10.118,
      "step": 100
    },
    {
      "epoch": 0.016121308858739025,
      "grad_norm": 11.645723342895508,
      "learning_rate": 4.264362337302798e-05,
      "loss": 11.301,
      "step": 101
    },
    {
      "epoch": 0.016280925778132484,
      "grad_norm": 10.529091835021973,
      "learning_rate": 4.1983350186928894e-05,
      "loss": 11.0128,
      "step": 102
    },
    {
      "epoch": 0.01644054269752594,
      "grad_norm": 11.569469451904297,
      "learning_rate": 4.132253477269233e-05,
      "loss": 10.7212,
      "step": 103
    },
    {
      "epoch": 0.016600159616919394,
      "grad_norm": 10.697254180908203,
      "learning_rate": 4.0661357790654345e-05,
      "loss": 11.4474,
      "step": 104
    },
    {
      "epoch": 0.01675977653631285,
      "grad_norm": 11.047499656677246,
      "learning_rate": 4e-05,
      "loss": 10.3108,
      "step": 105
    },
    {
      "epoch": 0.016919393455706304,
      "grad_norm": 10.388757705688477,
      "learning_rate": 3.933864220934566e-05,
      "loss": 9.421,
      "step": 106
    },
    {
      "epoch": 0.01707901037509976,
      "grad_norm": 10.334525108337402,
      "learning_rate": 3.8677465227307676e-05,
      "loss": 10.0589,
      "step": 107
    },
    {
      "epoch": 0.017238627294493217,
      "grad_norm": 10.174186706542969,
      "learning_rate": 3.8016649813071106e-05,
      "loss": 9.8046,
      "step": 108
    },
    {
      "epoch": 0.017398244213886672,
      "grad_norm": 11.022933006286621,
      "learning_rate": 3.735637662697203e-05,
      "loss": 9.9081,
      "step": 109
    },
    {
      "epoch": 0.017557861133280128,
      "grad_norm": 10.218436241149902,
      "learning_rate": 3.669682618110671e-05,
      "loss": 10.1829,
      "step": 110
    },
    {
      "epoch": 0.017717478052673583,
      "grad_norm": 10.797944068908691,
      "learning_rate": 3.6038178789981494e-05,
      "loss": 10.5076,
      "step": 111
    },
    {
      "epoch": 0.017877094972067038,
      "grad_norm": 10.90060043334961,
      "learning_rate": 3.538061452121656e-05,
      "loss": 9.8916,
      "step": 112
    },
    {
      "epoch": 0.018036711891460496,
      "grad_norm": 11.15035629272461,
      "learning_rate": 3.472431314631724e-05,
      "loss": 9.9848,
      "step": 113
    },
    {
      "epoch": 0.01819632881085395,
      "grad_norm": 10.41242790222168,
      "learning_rate": 3.406945409152632e-05,
      "loss": 9.4813,
      "step": 114
    },
    {
      "epoch": 0.018355945730247406,
      "grad_norm": 11.707329750061035,
      "learning_rate": 3.341621638877064e-05,
      "loss": 10.4817,
      "step": 115
    },
    {
      "epoch": 0.01851556264964086,
      "grad_norm": 11.070030212402344,
      "learning_rate": 3.276477862671562e-05,
      "loss": 8.8669,
      "step": 116
    },
    {
      "epoch": 0.018675179569034316,
      "grad_norm": 11.447722434997559,
      "learning_rate": 3.21153189019408e-05,
      "loss": 8.9964,
      "step": 117
    },
    {
      "epoch": 0.018834796488427775,
      "grad_norm": 12.059677124023438,
      "learning_rate": 3.146801477025013e-05,
      "loss": 9.8978,
      "step": 118
    },
    {
      "epoch": 0.01899441340782123,
      "grad_norm": 11.897377967834473,
      "learning_rate": 3.082304319812994e-05,
      "loss": 9.2506,
      "step": 119
    },
    {
      "epoch": 0.019154030327214685,
      "grad_norm": 12.78510570526123,
      "learning_rate": 3.0180580514368037e-05,
      "loss": 9.3649,
      "step": 120
    },
    {
      "epoch": 0.01931364724660814,
      "grad_norm": 11.492258071899414,
      "learning_rate": 2.9540802361847212e-05,
      "loss": 8.8457,
      "step": 121
    },
    {
      "epoch": 0.019473264166001595,
      "grad_norm": 11.497536659240723,
      "learning_rate": 2.890388364952623e-05,
      "loss": 10.1121,
      "step": 122
    },
    {
      "epoch": 0.019632881085395053,
      "grad_norm": 12.456028938293457,
      "learning_rate": 2.8269998504621416e-05,
      "loss": 10.9649,
      "step": 123
    },
    {
      "epoch": 0.01979249800478851,
      "grad_norm": 13.163850784301758,
      "learning_rate": 2.7639320225002108e-05,
      "loss": 9.9934,
      "step": 124
    },
    {
      "epoch": 0.019952114924181964,
      "grad_norm": 13.581376075744629,
      "learning_rate": 2.7012021231812666e-05,
      "loss": 10.2623,
      "step": 125
    },
    {
      "epoch": 0.02011173184357542,
      "grad_norm": 12.48007583618164,
      "learning_rate": 2.638827302233428e-05,
      "loss": 10.1343,
      "step": 126
    },
    {
      "epoch": 0.020271348762968874,
      "grad_norm": 13.67141342163086,
      "learning_rate": 2.576824612309934e-05,
      "loss": 9.6456,
      "step": 127
    },
    {
      "epoch": 0.02043096568236233,
      "grad_norm": 13.115945816040039,
      "learning_rate": 2.5152110043271166e-05,
      "loss": 9.8097,
      "step": 128
    },
    {
      "epoch": 0.020590582601755787,
      "grad_norm": 12.952895164489746,
      "learning_rate": 2.454003322830192e-05,
      "loss": 9.1189,
      "step": 129
    },
    {
      "epoch": 0.020750199521149242,
      "grad_norm": 14.22287654876709,
      "learning_rate": 2.393218301388123e-05,
      "loss": 9.1401,
      "step": 130
    },
    {
      "epoch": 0.020909816440542697,
      "grad_norm": 13.032074928283691,
      "learning_rate": 2.3328725580188395e-05,
      "loss": 9.0543,
      "step": 131
    },
    {
      "epoch": 0.021069433359936152,
      "grad_norm": 14.65915584564209,
      "learning_rate": 2.272982590646029e-05,
      "loss": 9.9715,
      "step": 132
    },
    {
      "epoch": 0.021229050279329607,
      "grad_norm": 16.45696449279785,
      "learning_rate": 2.2135647725887744e-05,
      "loss": 9.1353,
      "step": 133
    },
    {
      "epoch": 0.021388667198723066,
      "grad_norm": 15.06648063659668,
      "learning_rate": 2.1546353480852495e-05,
      "loss": 9.5186,
      "step": 134
    },
    {
      "epoch": 0.02154828411811652,
      "grad_norm": 14.118624687194824,
      "learning_rate": 2.096210427851706e-05,
      "loss": 8.2845,
      "step": 135
    },
    {
      "epoch": 0.021707901037509976,
      "grad_norm": 17.657686233520508,
      "learning_rate": 2.038305984677969e-05,
      "loss": 10.4191,
      "step": 136
    },
    {
      "epoch": 0.02186751795690343,
      "grad_norm": 17.25397300720215,
      "learning_rate": 1.9809378490606264e-05,
      "loss": 10.6506,
      "step": 137
    },
    {
      "epoch": 0.022027134876296886,
      "grad_norm": 16.477792739868164,
      "learning_rate": 1.9241217048751406e-05,
      "loss": 10.411,
      "step": 138
    },
    {
      "epoch": 0.022186751795690345,
      "grad_norm": 17.870357513427734,
      "learning_rate": 1.867873085088026e-05,
      "loss": 9.313,
      "step": 139
    },
    {
      "epoch": 0.0223463687150838,
      "grad_norm": 17.268783569335938,
      "learning_rate": 1.8122073675102935e-05,
      "loss": 9.4042,
      "step": 140
    },
    {
      "epoch": 0.022505985634477255,
      "grad_norm": 17.178604125976562,
      "learning_rate": 1.75713977059331e-05,
      "loss": 8.4997,
      "step": 141
    },
    {
      "epoch": 0.02266560255387071,
      "grad_norm": 52.73872375488281,
      "learning_rate": 1.702685349268226e-05,
      "loss": 8.9395,
      "step": 142
    },
    {
      "epoch": 0.022825219473264165,
      "grad_norm": 20.91663932800293,
      "learning_rate": 1.648858990830108e-05,
      "loss": 9.0917,
      "step": 143
    },
    {
      "epoch": 0.022984836392657623,
      "grad_norm": 18.2515811920166,
      "learning_rate": 1.5956754108678996e-05,
      "loss": 8.2123,
      "step": 144
    },
    {
      "epoch": 0.02314445331205108,
      "grad_norm": 22.17852020263672,
      "learning_rate": 1.5431491492413288e-05,
      "loss": 11.5065,
      "step": 145
    },
    {
      "epoch": 0.023304070231444533,
      "grad_norm": 22.831722259521484,
      "learning_rate": 1.491294566105852e-05,
      "loss": 9.9597,
      "step": 146
    },
    {
      "epoch": 0.02346368715083799,
      "grad_norm": 19.835552215576172,
      "learning_rate": 1.4401258379867335e-05,
      "loss": 8.9841,
      "step": 147
    },
    {
      "epoch": 0.023623304070231443,
      "grad_norm": 25.251991271972656,
      "learning_rate": 1.3896569539033253e-05,
      "loss": 8.8135,
      "step": 148
    },
    {
      "epoch": 0.023782920989624902,
      "grad_norm": 21.608013153076172,
      "learning_rate": 1.3399017115446067e-05,
      "loss": 8.2115,
      "step": 149
    },
    {
      "epoch": 0.023942537909018357,
      "grad_norm": 24.518632888793945,
      "learning_rate": 1.2908737134970367e-05,
      "loss": 6.6956,
      "step": 150
    },
    {
      "epoch": 0.023942537909018357,
      "eval_loss": 2.428995370864868,
      "eval_runtime": 260.9923,
      "eval_samples_per_second": 40.43,
      "eval_steps_per_second": 10.108,
      "step": 150
    },
    {
      "epoch": 0.024102154828411812,
      "grad_norm": 10.971107482910156,
      "learning_rate": 1.242586363525737e-05,
      "loss": 11.044,
      "step": 151
    },
    {
      "epoch": 0.024261771747805267,
      "grad_norm": 10.344849586486816,
      "learning_rate": 1.1950528629100457e-05,
      "loss": 9.7862,
      "step": 152
    },
    {
      "epoch": 0.024421388667198722,
      "grad_norm": 10.37899398803711,
      "learning_rate": 1.1482862068344121e-05,
      "loss": 10.5911,
      "step": 153
    },
    {
      "epoch": 0.024581005586592177,
      "grad_norm": 10.244215965270996,
      "learning_rate": 1.1022991808356442e-05,
      "loss": 10.5494,
      "step": 154
    },
    {
      "epoch": 0.024740622505985636,
      "grad_norm": 10.632125854492188,
      "learning_rate": 1.0571043573074737e-05,
      "loss": 9.6689,
      "step": 155
    },
    {
      "epoch": 0.02490023942537909,
      "grad_norm": 10.224246978759766,
      "learning_rate": 1.0127140920633857e-05,
      "loss": 9.5835,
      "step": 156
    },
    {
      "epoch": 0.025059856344772546,
      "grad_norm": 10.334432601928711,
      "learning_rate": 9.69140520958662e-06,
      "loss": 9.6411,
      "step": 157
    },
    {
      "epoch": 0.025219473264166,
      "grad_norm": 10.643871307373047,
      "learning_rate": 9.263955565725648e-06,
      "loss": 9.7923,
      "step": 158
    },
    {
      "epoch": 0.025379090183559456,
      "grad_norm": 11.414630889892578,
      "learning_rate": 8.844908849515509e-06,
      "loss": 10.6889,
      "step": 159
    },
    {
      "epoch": 0.025538707102952914,
      "grad_norm": 9.992051124572754,
      "learning_rate": 8.434379624144261e-06,
      "loss": 9.7891,
      "step": 160
    },
    {
      "epoch": 0.02569832402234637,
      "grad_norm": 11.766764640808105,
      "learning_rate": 8.032480124203013e-06,
      "loss": 10.3812,
      "step": 161
    },
    {
      "epoch": 0.025857940941739824,
      "grad_norm": 10.601237297058105,
      "learning_rate": 7.639320225002106e-06,
      "loss": 9.8553,
      "step": 162
    },
    {
      "epoch": 0.02601755786113328,
      "grad_norm": 11.593340873718262,
      "learning_rate": 7.255007412532307e-06,
      "loss": 9.2675,
      "step": 163
    },
    {
      "epoch": 0.026177174780526734,
      "grad_norm": 11.817658424377441,
      "learning_rate": 6.8796467540791986e-06,
      "loss": 9.4211,
      "step": 164
    },
    {
      "epoch": 0.026336791699920193,
      "grad_norm": 11.031808853149414,
      "learning_rate": 6.513340869498859e-06,
      "loss": 9.8802,
      "step": 165
    },
    {
      "epoch": 0.026496408619313648,
      "grad_norm": 12.100275993347168,
      "learning_rate": 6.1561899031625794e-06,
      "loss": 10.855,
      "step": 166
    },
    {
      "epoch": 0.026656025538707103,
      "grad_norm": 12.345802307128906,
      "learning_rate": 5.808291496578435e-06,
      "loss": 11.3456,
      "step": 167
    },
    {
      "epoch": 0.026815642458100558,
      "grad_norm": 11.653558731079102,
      "learning_rate": 5.469740761697044e-06,
      "loss": 8.7237,
      "step": 168
    },
    {
      "epoch": 0.026975259377494013,
      "grad_norm": 11.88528060913086,
      "learning_rate": 5.140630254908905e-06,
      "loss": 9.1543,
      "step": 169
    },
    {
      "epoch": 0.02713487629688747,
      "grad_norm": 11.614215850830078,
      "learning_rate": 4.821049951740442e-06,
      "loss": 8.7752,
      "step": 170
    },
    {
      "epoch": 0.027294493216280927,
      "grad_norm": 12.890314102172852,
      "learning_rate": 4.511087222255528e-06,
      "loss": 9.7857,
      "step": 171
    },
    {
      "epoch": 0.027454110135674382,
      "grad_norm": 12.401043891906738,
      "learning_rate": 4.2108268071694616e-06,
      "loss": 9.5022,
      "step": 172
    },
    {
      "epoch": 0.027613727055067837,
      "grad_norm": 13.726554870605469,
      "learning_rate": 3.9203507946816445e-06,
      "loss": 10.8293,
      "step": 173
    },
    {
      "epoch": 0.027773343974461292,
      "grad_norm": 12.232309341430664,
      "learning_rate": 3.6397385980335e-06,
      "loss": 9.8193,
      "step": 174
    },
    {
      "epoch": 0.027932960893854747,
      "grad_norm": 12.865934371948242,
      "learning_rate": 3.3690669337977e-06,
      "loss": 9.4224,
      "step": 175
    },
    {
      "epoch": 0.028092577813248205,
      "grad_norm": 13.326064109802246,
      "learning_rate": 3.1084098009046106e-06,
      "loss": 9.7565,
      "step": 176
    },
    {
      "epoch": 0.02825219473264166,
      "grad_norm": 13.985200881958008,
      "learning_rate": 2.8578384604117217e-06,
      "loss": 9.6235,
      "step": 177
    },
    {
      "epoch": 0.028411811652035115,
      "grad_norm": 13.09310531616211,
      "learning_rate": 2.6174214160215704e-06,
      "loss": 10.4877,
      "step": 178
    },
    {
      "epoch": 0.02857142857142857,
      "grad_norm": 13.141727447509766,
      "learning_rate": 2.3872243953535535e-06,
      "loss": 8.9112,
      "step": 179
    },
    {
      "epoch": 0.028731045490822026,
      "grad_norm": 13.407176971435547,
      "learning_rate": 2.1673103319746146e-06,
      "loss": 8.7778,
      "step": 180
    },
    {
      "epoch": 0.028890662410215484,
      "grad_norm": 14.593409538269043,
      "learning_rate": 1.957739348193859e-06,
      "loss": 9.7216,
      "step": 181
    },
    {
      "epoch": 0.02905027932960894,
      "grad_norm": 13.661775588989258,
      "learning_rate": 1.7585687386256944e-06,
      "loss": 9.5762,
      "step": 182
    },
    {
      "epoch": 0.029209896249002394,
      "grad_norm": 15.88332748413086,
      "learning_rate": 1.5698529545260744e-06,
      "loss": 9.5316,
      "step": 183
    },
    {
      "epoch": 0.02936951316839585,
      "grad_norm": 14.10777473449707,
      "learning_rate": 1.3916435889060575e-06,
      "loss": 9.6112,
      "step": 184
    },
    {
      "epoch": 0.029529130087789304,
      "grad_norm": 14.599610328674316,
      "learning_rate": 1.2239893624267852e-06,
      "loss": 9.3918,
      "step": 185
    },
    {
      "epoch": 0.029688747007182763,
      "grad_norm": 17.34296989440918,
      "learning_rate": 1.0669361100797704e-06,
      "loss": 10.6652,
      "step": 186
    },
    {
      "epoch": 0.029848363926576218,
      "grad_norm": 15.431779861450195,
      "learning_rate": 9.205267686560293e-07,
      "loss": 8.9332,
      "step": 187
    },
    {
      "epoch": 0.030007980845969673,
      "grad_norm": 15.42230224609375,
      "learning_rate": 7.848013650076258e-07,
      "loss": 9.0331,
      "step": 188
    },
    {
      "epoch": 0.030167597765363128,
      "grad_norm": 18.1710147857666,
      "learning_rate": 6.597970051047053e-07,
      "loss": 9.5894,
      "step": 189
    },
    {
      "epoch": 0.030327214684756583,
      "grad_norm": 17.977142333984375,
      "learning_rate": 5.455478638911071e-07,
      "loss": 9.2745,
      "step": 190
    },
    {
      "epoch": 0.03048683160415004,
      "grad_norm": 17.78211784362793,
      "learning_rate": 4.420851759412603e-07,
      "loss": 9.203,
      "step": 191
    },
    {
      "epoch": 0.030646448523543496,
      "grad_norm": 18.773420333862305,
      "learning_rate": 3.4943722692099224e-07,
      "loss": 9.5305,
      "step": 192
    },
    {
      "epoch": 0.03080606544293695,
      "grad_norm": 18.938011169433594,
      "learning_rate": 2.676293458544743e-07,
      "loss": 9.2699,
      "step": 193
    },
    {
      "epoch": 0.030965682362330407,
      "grad_norm": 18.520278930664062,
      "learning_rate": 1.9668389819954338e-07,
      "loss": 9.0121,
      "step": 194
    },
    {
      "epoch": 0.03112529928172386,
      "grad_norm": 18.693401336669922,
      "learning_rate": 1.3662027973320614e-07,
      "loss": 8.5485,
      "step": 195
    },
    {
      "epoch": 0.03128491620111732,
      "grad_norm": 21.85112762451172,
      "learning_rate": 8.745491124901861e-08,
      "loss": 9.7759,
      "step": 196
    },
    {
      "epoch": 0.03144453312051077,
      "grad_norm": 21.148921966552734,
      "learning_rate": 4.920123406781052e-08,
      "loss": 10.5365,
      "step": 197
    },
    {
      "epoch": 0.03160415003990423,
      "grad_norm": 26.282615661621094,
      "learning_rate": 2.1869706362958044e-08,
      "loss": 8.095,
      "step": 198
    },
    {
      "epoch": 0.03176376695929769,
      "grad_norm": 38.542179107666016,
      "learning_rate": 5.467800301239834e-09,
      "loss": 9.9331,
      "step": 199
    },
    {
      "epoch": 0.03192338387869114,
      "grad_norm": 29.25096321105957,
      "learning_rate": 0.0,
      "loss": 10.4418,
      "step": 200
    },
    {
      "epoch": 0.03192338387869114,
      "eval_loss": 2.411439895629883,
      "eval_runtime": 260.9844,
      "eval_samples_per_second": 40.432,
      "eval_steps_per_second": 10.108,
      "step": 200
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 4,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.2893106147950592e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}