{
  "best_metric": 0.5842863321304321,
  "best_model_checkpoint": "miner_id_24/checkpoint-200",
  "epoch": 0.7554296506137866,
  "eval_steps": 50,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.003777148253068933,
      "grad_norm": 5.144922733306885,
      "learning_rate": 5e-06,
      "loss": 1.4055,
      "step": 1
    },
    {
      "epoch": 0.003777148253068933,
      "eval_loss": 2.09171986579895,
      "eval_runtime": 10.8476,
      "eval_samples_per_second": 41.115,
      "eval_steps_per_second": 20.558,
      "step": 1
    },
    {
      "epoch": 0.007554296506137866,
      "grad_norm": 5.170313358306885,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 2
    },
    {
      "epoch": 0.0113314447592068,
      "grad_norm": 4.963709354400635,
      "learning_rate": 1.5e-05,
      "loss": 1.5649,
      "step": 3
    },
    {
      "epoch": 0.015108593012275733,
      "grad_norm": 5.641475200653076,
      "learning_rate": 2e-05,
      "loss": 1.4524,
      "step": 4
    },
    {
      "epoch": 0.018885741265344664,
      "grad_norm": 3.8712124824523926,
      "learning_rate": 2.5e-05,
      "loss": 1.3812,
      "step": 5
    },
    {
      "epoch": 0.0226628895184136,
      "grad_norm": 3.4170196056365967,
      "learning_rate": 3e-05,
      "loss": 1.3778,
      "step": 6
    },
    {
      "epoch": 0.02644003777148253,
      "grad_norm": 3.4598617553710938,
      "learning_rate": 3.5e-05,
      "loss": 1.1023,
      "step": 7
    },
    {
      "epoch": 0.030217186024551465,
      "grad_norm": 3.242964744567871,
      "learning_rate": 4e-05,
      "loss": 1.1547,
      "step": 8
    },
    {
      "epoch": 0.0339943342776204,
      "grad_norm": 3.2598581314086914,
      "learning_rate": 4.5e-05,
      "loss": 1.1588,
      "step": 9
    },
    {
      "epoch": 0.03777148253068933,
      "grad_norm": 3.4516100883483887,
      "learning_rate": 5e-05,
      "loss": 1.1894,
      "step": 10
    },
    {
      "epoch": 0.04154863078375826,
      "grad_norm": 2.8747804164886475,
      "learning_rate": 5.500000000000001e-05,
      "loss": 1.083,
      "step": 11
    },
    {
      "epoch": 0.0453257790368272,
      "grad_norm": 2.639749526977539,
      "learning_rate": 6e-05,
      "loss": 1.0069,
      "step": 12
    },
    {
      "epoch": 0.049102927289896126,
      "grad_norm": 2.042722225189209,
      "learning_rate": 6.500000000000001e-05,
      "loss": 0.9725,
      "step": 13
    },
    {
      "epoch": 0.05288007554296506,
      "grad_norm": 2.2866015434265137,
      "learning_rate": 7e-05,
      "loss": 1.0499,
      "step": 14
    },
    {
      "epoch": 0.056657223796033995,
      "grad_norm": 2.047022819519043,
      "learning_rate": 7.500000000000001e-05,
      "loss": 1.0347,
      "step": 15
    },
    {
      "epoch": 0.06043437204910293,
      "grad_norm": 2.0791587829589844,
      "learning_rate": 8e-05,
      "loss": 1.0018,
      "step": 16
    },
    {
      "epoch": 0.06421152030217187,
      "grad_norm": 2.2220687866210938,
      "learning_rate": 8.5e-05,
      "loss": 0.9715,
      "step": 17
    },
    {
      "epoch": 0.0679886685552408,
      "grad_norm": 2.0520265102386475,
      "learning_rate": 9e-05,
      "loss": 0.927,
      "step": 18
    },
    {
      "epoch": 0.07176581680830972,
      "grad_norm": 2.1305389404296875,
      "learning_rate": 9.5e-05,
      "loss": 0.9662,
      "step": 19
    },
    {
      "epoch": 0.07554296506137866,
      "grad_norm": 1.899276614189148,
      "learning_rate": 0.0001,
      "loss": 0.9462,
      "step": 20
    },
    {
      "epoch": 0.07932011331444759,
      "grad_norm": 1.833707571029663,
      "learning_rate": 9.999238475781957e-05,
      "loss": 0.9486,
      "step": 21
    },
    {
      "epoch": 0.08309726156751653,
      "grad_norm": 1.9456089735031128,
      "learning_rate": 9.99695413509548e-05,
      "loss": 0.9008,
      "step": 22
    },
    {
      "epoch": 0.08687440982058546,
      "grad_norm": 1.9707703590393066,
      "learning_rate": 9.99314767377287e-05,
      "loss": 0.8909,
      "step": 23
    },
    {
      "epoch": 0.0906515580736544,
      "grad_norm": 2.10565447807312,
      "learning_rate": 9.987820251299122e-05,
      "loss": 1.0223,
      "step": 24
    },
    {
      "epoch": 0.09442870632672333,
      "grad_norm": 2.1001415252685547,
      "learning_rate": 9.980973490458728e-05,
      "loss": 0.9073,
      "step": 25
    },
    {
      "epoch": 0.09820585457979225,
      "grad_norm": 1.8849313259124756,
      "learning_rate": 9.972609476841367e-05,
      "loss": 0.9276,
      "step": 26
    },
    {
      "epoch": 0.10198300283286119,
      "grad_norm": 1.757330298423767,
      "learning_rate": 9.962730758206611e-05,
      "loss": 0.7808,
      "step": 27
    },
    {
      "epoch": 0.10576015108593012,
      "grad_norm": 1.9129984378814697,
      "learning_rate": 9.951340343707852e-05,
      "loss": 0.9231,
      "step": 28
    },
    {
      "epoch": 0.10953729933899906,
      "grad_norm": 2.1026272773742676,
      "learning_rate": 9.938441702975689e-05,
      "loss": 0.8499,
      "step": 29
    },
    {
      "epoch": 0.11331444759206799,
      "grad_norm": 2.4626994132995605,
      "learning_rate": 9.924038765061042e-05,
      "loss": 0.7585,
      "step": 30
    },
    {
      "epoch": 0.11709159584513693,
      "grad_norm": 2.149193525314331,
      "learning_rate": 9.908135917238321e-05,
      "loss": 0.8706,
      "step": 31
    },
    {
      "epoch": 0.12086874409820586,
      "grad_norm": 2.2951858043670654,
      "learning_rate": 9.890738003669029e-05,
      "loss": 0.8401,
      "step": 32
    },
    {
      "epoch": 0.12464589235127478,
      "grad_norm": 2.2280688285827637,
      "learning_rate": 9.871850323926177e-05,
      "loss": 0.8547,
      "step": 33
    },
    {
      "epoch": 0.12842304060434373,
      "grad_norm": 2.183703660964966,
      "learning_rate": 9.851478631379982e-05,
      "loss": 0.8295,
      "step": 34
    },
    {
      "epoch": 0.13220018885741266,
      "grad_norm": 2.0195577144622803,
      "learning_rate": 9.829629131445342e-05,
      "loss": 0.8436,
      "step": 35
    },
    {
      "epoch": 0.1359773371104816,
      "grad_norm": 2.405426025390625,
      "learning_rate": 9.806308479691595e-05,
      "loss": 0.8423,
      "step": 36
    },
    {
      "epoch": 0.1397544853635505,
      "grad_norm": 2.1901025772094727,
      "learning_rate": 9.781523779815179e-05,
      "loss": 0.8917,
      "step": 37
    },
    {
      "epoch": 0.14353163361661944,
      "grad_norm": 1.9201372861862183,
      "learning_rate": 9.755282581475769e-05,
      "loss": 0.7195,
      "step": 38
    },
    {
      "epoch": 0.14730878186968838,
      "grad_norm": 2.105783700942993,
      "learning_rate": 9.727592877996585e-05,
      "loss": 0.7851,
      "step": 39
    },
    {
      "epoch": 0.1510859301227573,
      "grad_norm": 2.496196985244751,
      "learning_rate": 9.698463103929542e-05,
      "loss": 0.9378,
      "step": 40
    },
    {
      "epoch": 0.15486307837582625,
      "grad_norm": 2.0118231773376465,
      "learning_rate": 9.667902132486009e-05,
      "loss": 0.6978,
      "step": 41
    },
    {
      "epoch": 0.15864022662889518,
      "grad_norm": 2.645458698272705,
      "learning_rate": 9.635919272833938e-05,
      "loss": 1.0023,
      "step": 42
    },
    {
      "epoch": 0.16241737488196412,
      "grad_norm": 2.71828556060791,
      "learning_rate": 9.602524267262203e-05,
      "loss": 1.0042,
      "step": 43
    },
    {
      "epoch": 0.16619452313503305,
      "grad_norm": 2.814579486846924,
      "learning_rate": 9.567727288213005e-05,
      "loss": 1.05,
      "step": 44
    },
    {
      "epoch": 0.16997167138810199,
      "grad_norm": 3.425931930541992,
      "learning_rate": 9.53153893518325e-05,
      "loss": 1.3135,
      "step": 45
    },
    {
      "epoch": 0.17374881964117092,
      "grad_norm": 3.859901189804077,
      "learning_rate": 9.493970231495835e-05,
      "loss": 1.2467,
      "step": 46
    },
    {
      "epoch": 0.17752596789423986,
      "grad_norm": 3.8892781734466553,
      "learning_rate": 9.45503262094184e-05,
      "loss": 1.2551,
      "step": 47
    },
    {
      "epoch": 0.1813031161473088,
      "grad_norm": 3.965714931488037,
      "learning_rate": 9.414737964294636e-05,
      "loss": 1.1796,
      "step": 48
    },
    {
      "epoch": 0.18508026440037773,
      "grad_norm": 3.65535569190979,
      "learning_rate": 9.373098535696979e-05,
      "loss": 1.2059,
      "step": 49
    },
    {
      "epoch": 0.18885741265344666,
      "grad_norm": 3.6969974040985107,
      "learning_rate": 9.330127018922194e-05,
      "loss": 1.054,
      "step": 50
    },
    {
      "epoch": 0.18885741265344666,
      "eval_loss": 1.0890626907348633,
      "eval_runtime": 10.8094,
      "eval_samples_per_second": 41.26,
      "eval_steps_per_second": 20.63,
      "step": 50
    },
    {
      "epoch": 0.19263456090651557,
      "grad_norm": 4.26871395111084,
      "learning_rate": 9.285836503510562e-05,
      "loss": 1.0492,
      "step": 51
    },
    {
      "epoch": 0.1964117091595845,
      "grad_norm": 4.4510064125061035,
      "learning_rate": 9.24024048078213e-05,
      "loss": 1.1435,
      "step": 52
    },
    {
      "epoch": 0.20018885741265344,
      "grad_norm": 3.34430193901062,
      "learning_rate": 9.193352839727121e-05,
      "loss": 1.0126,
      "step": 53
    },
    {
      "epoch": 0.20396600566572237,
      "grad_norm": 2.6075751781463623,
      "learning_rate": 9.145187862775209e-05,
      "loss": 0.9598,
      "step": 54
    },
    {
      "epoch": 0.2077431539187913,
      "grad_norm": 1.8077794313430786,
      "learning_rate": 9.09576022144496e-05,
      "loss": 0.8542,
      "step": 55
    },
    {
      "epoch": 0.21152030217186024,
      "grad_norm": 1.69106125831604,
      "learning_rate": 9.045084971874738e-05,
      "loss": 0.8221,
      "step": 56
    },
    {
      "epoch": 0.21529745042492918,
      "grad_norm": 1.7321957349777222,
      "learning_rate": 8.993177550236464e-05,
      "loss": 0.8262,
      "step": 57
    },
    {
      "epoch": 0.2190745986779981,
      "grad_norm": 1.8129096031188965,
      "learning_rate": 8.940053768033609e-05,
      "loss": 0.8381,
      "step": 58
    },
    {
      "epoch": 0.22285174693106705,
      "grad_norm": 1.7923287153244019,
      "learning_rate": 8.885729807284856e-05,
      "loss": 0.7784,
      "step": 59
    },
    {
      "epoch": 0.22662889518413598,
      "grad_norm": 1.7900950908660889,
      "learning_rate": 8.83022221559489e-05,
      "loss": 0.7865,
      "step": 60
    },
    {
      "epoch": 0.23040604343720492,
      "grad_norm": 1.5210798978805542,
      "learning_rate": 8.773547901113862e-05,
      "loss": 0.6971,
      "step": 61
    },
    {
      "epoch": 0.23418319169027385,
      "grad_norm": 1.5672999620437622,
      "learning_rate": 8.715724127386972e-05,
      "loss": 0.6601,
      "step": 62
    },
    {
      "epoch": 0.23796033994334279,
      "grad_norm": 1.5031838417053223,
      "learning_rate": 8.656768508095853e-05,
      "loss": 0.7785,
      "step": 63
    },
    {
      "epoch": 0.24173748819641172,
      "grad_norm": 1.4216283559799194,
      "learning_rate": 8.596699001693255e-05,
      "loss": 0.7241,
      "step": 64
    },
    {
      "epoch": 0.24551463644948066,
      "grad_norm": 1.938550353050232,
      "learning_rate": 8.535533905932738e-05,
      "loss": 0.7643,
      "step": 65
    },
    {
      "epoch": 0.24929178470254956,
      "grad_norm": 1.538443922996521,
      "learning_rate": 8.473291852294987e-05,
      "loss": 0.7225,
      "step": 66
    },
    {
      "epoch": 0.2530689329556185,
      "grad_norm": 1.6690013408660889,
      "learning_rate": 8.409991800312493e-05,
      "loss": 0.7407,
      "step": 67
    },
    {
      "epoch": 0.25684608120868746,
      "grad_norm": 1.4403491020202637,
      "learning_rate": 8.345653031794292e-05,
      "loss": 0.629,
      "step": 68
    },
    {
      "epoch": 0.26062322946175637,
      "grad_norm": 1.4733555316925049,
      "learning_rate": 8.280295144952536e-05,
      "loss": 0.5778,
      "step": 69
    },
    {
      "epoch": 0.26440037771482533,
      "grad_norm": 1.6242821216583252,
      "learning_rate": 8.213938048432697e-05,
      "loss": 0.7126,
      "step": 70
    },
    {
      "epoch": 0.26817752596789424,
      "grad_norm": 1.6202547550201416,
      "learning_rate": 8.146601955249188e-05,
      "loss": 0.7169,
      "step": 71
    },
    {
      "epoch": 0.2719546742209632,
      "grad_norm": 1.4979926347732544,
      "learning_rate": 8.07830737662829e-05,
      "loss": 0.6462,
      "step": 72
    },
    {
      "epoch": 0.2757318224740321,
      "grad_norm": 1.7165087461471558,
      "learning_rate": 8.009075115760243e-05,
      "loss": 0.6811,
      "step": 73
    },
    {
      "epoch": 0.279508970727101,
      "grad_norm": 1.5448169708251953,
      "learning_rate": 7.938926261462366e-05,
      "loss": 0.6585,
      "step": 74
    },
    {
      "epoch": 0.28328611898017,
      "grad_norm": 1.682450532913208,
      "learning_rate": 7.86788218175523e-05,
      "loss": 0.7064,
      "step": 75
    },
    {
      "epoch": 0.2870632672332389,
      "grad_norm": 1.5866179466247559,
      "learning_rate": 7.795964517353735e-05,
      "loss": 0.6464,
      "step": 76
    },
    {
      "epoch": 0.29084041548630785,
      "grad_norm": 1.8173009157180786,
      "learning_rate": 7.723195175075136e-05,
      "loss": 0.6584,
      "step": 77
    },
    {
      "epoch": 0.29461756373937675,
      "grad_norm": 1.7022247314453125,
      "learning_rate": 7.649596321166024e-05,
      "loss": 0.7146,
      "step": 78
    },
    {
      "epoch": 0.2983947119924457,
      "grad_norm": 1.561222791671753,
      "learning_rate": 7.575190374550272e-05,
      "loss": 0.6302,
      "step": 79
    },
    {
      "epoch": 0.3021718602455146,
      "grad_norm": 1.5313177108764648,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.6386,
      "step": 80
    },
    {
      "epoch": 0.3059490084985836,
      "grad_norm": 1.6224055290222168,
      "learning_rate": 7.424048101231686e-05,
      "loss": 0.6354,
      "step": 81
    },
    {
      "epoch": 0.3097261567516525,
      "grad_norm": 1.8029823303222656,
      "learning_rate": 7.347357813929454e-05,
      "loss": 0.6993,
      "step": 82
    },
    {
      "epoch": 0.31350330500472146,
      "grad_norm": 1.8876993656158447,
      "learning_rate": 7.269952498697734e-05,
      "loss": 0.6369,
      "step": 83
    },
    {
      "epoch": 0.31728045325779036,
      "grad_norm": 1.8419164419174194,
      "learning_rate": 7.191855733945387e-05,
      "loss": 0.562,
      "step": 84
    },
    {
      "epoch": 0.3210576015108593,
      "grad_norm": 1.7774418592453003,
      "learning_rate": 7.113091308703498e-05,
      "loss": 0.6443,
      "step": 85
    },
    {
      "epoch": 0.32483474976392823,
      "grad_norm": 1.7081546783447266,
      "learning_rate": 7.033683215379002e-05,
      "loss": 0.5524,
      "step": 86
    },
    {
      "epoch": 0.3286118980169972,
      "grad_norm": 1.6573454141616821,
      "learning_rate": 6.953655642446368e-05,
      "loss": 0.5105,
      "step": 87
    },
    {
      "epoch": 0.3323890462700661,
      "grad_norm": 1.8373266458511353,
      "learning_rate": 6.873032967079561e-05,
      "loss": 0.6277,
      "step": 88
    },
    {
      "epoch": 0.336166194523135,
      "grad_norm": 1.8988851308822632,
      "learning_rate": 6.7918397477265e-05,
      "loss": 0.6998,
      "step": 89
    },
    {
      "epoch": 0.33994334277620397,
      "grad_norm": 1.8741487264633179,
      "learning_rate": 6.710100716628344e-05,
      "loss": 0.681,
      "step": 90
    },
    {
      "epoch": 0.3437204910292729,
      "grad_norm": 1.905774712562561,
      "learning_rate": 6.627840772285784e-05,
      "loss": 0.6405,
      "step": 91
    },
    {
      "epoch": 0.34749763928234184,
      "grad_norm": 2.430171489715576,
      "learning_rate": 6.545084971874738e-05,
      "loss": 0.7597,
      "step": 92
    },
    {
      "epoch": 0.35127478753541075,
      "grad_norm": 2.2742724418640137,
      "learning_rate": 6.461858523613684e-05,
      "loss": 0.7883,
      "step": 93
    },
    {
      "epoch": 0.3550519357884797,
      "grad_norm": 2.2646121978759766,
      "learning_rate": 6.378186779084995e-05,
      "loss": 0.7239,
      "step": 94
    },
    {
      "epoch": 0.3588290840415486,
      "grad_norm": 2.812591791152954,
      "learning_rate": 6.294095225512603e-05,
      "loss": 1.0306,
      "step": 95
    },
    {
      "epoch": 0.3626062322946176,
      "grad_norm": 3.1954076290130615,
      "learning_rate": 6.209609477998338e-05,
      "loss": 0.9472,
      "step": 96
    },
    {
      "epoch": 0.3663833805476865,
      "grad_norm": 3.4073708057403564,
      "learning_rate": 6.124755271719325e-05,
      "loss": 0.9748,
      "step": 97
    },
    {
      "epoch": 0.37016052880075545,
      "grad_norm": 3.6358604431152344,
      "learning_rate": 6.0395584540887963e-05,
      "loss": 1.0534,
      "step": 98
    },
    {
      "epoch": 0.37393767705382436,
      "grad_norm": 3.7679121494293213,
      "learning_rate": 5.9540449768827246e-05,
      "loss": 1.0016,
      "step": 99
    },
    {
      "epoch": 0.3777148253068933,
      "grad_norm": 4.0353312492370605,
      "learning_rate": 5.868240888334653e-05,
      "loss": 1.2329,
      "step": 100
    },
    {
      "epoch": 0.3777148253068933,
      "eval_loss": 0.7517095804214478,
      "eval_runtime": 10.8231,
      "eval_samples_per_second": 41.208,
      "eval_steps_per_second": 20.604,
      "step": 100
    },
    {
      "epoch": 0.3814919735599622,
      "grad_norm": 2.111731767654419,
      "learning_rate": 5.782172325201155e-05,
      "loss": 0.7728,
      "step": 101
    },
    {
      "epoch": 0.38526912181303113,
      "grad_norm": 1.9944093227386475,
      "learning_rate": 5.695865504800327e-05,
      "loss": 0.7565,
      "step": 102
    },
    {
      "epoch": 0.3890462700661001,
      "grad_norm": 2.0362370014190674,
      "learning_rate": 5.6093467170257374e-05,
      "loss": 0.7602,
      "step": 103
    },
    {
      "epoch": 0.392823418319169,
      "grad_norm": 2.199453115463257,
      "learning_rate": 5.522642316338268e-05,
      "loss": 0.7571,
      "step": 104
    },
    {
      "epoch": 0.39660056657223797,
      "grad_norm": 1.6437965631484985,
      "learning_rate": 5.435778713738292e-05,
      "loss": 0.726,
      "step": 105
    },
    {
      "epoch": 0.4003777148253069,
      "grad_norm": 1.5133771896362305,
      "learning_rate": 5.348782368720626e-05,
      "loss": 0.6105,
      "step": 106
    },
    {
      "epoch": 0.40415486307837584,
      "grad_norm": 1.2506144046783447,
      "learning_rate": 5.26167978121472e-05,
      "loss": 0.584,
      "step": 107
    },
    {
      "epoch": 0.40793201133144474,
      "grad_norm": 1.2938867807388306,
      "learning_rate": 5.174497483512506e-05,
      "loss": 0.5689,
      "step": 108
    },
    {
      "epoch": 0.4117091595845137,
      "grad_norm": 1.3909262418746948,
      "learning_rate": 5.0872620321864185e-05,
      "loss": 0.691,
      "step": 109
    },
    {
      "epoch": 0.4154863078375826,
      "grad_norm": 1.446152687072754,
      "learning_rate": 5e-05,
      "loss": 0.6143,
      "step": 110
    },
    {
      "epoch": 0.4192634560906516,
      "grad_norm": 1.4228076934814453,
      "learning_rate": 4.912737967813583e-05,
      "loss": 0.5271,
      "step": 111
    },
    {
      "epoch": 0.4230406043437205,
      "grad_norm": 1.578036904335022,
      "learning_rate": 4.825502516487497e-05,
      "loss": 0.5497,
      "step": 112
    },
    {
      "epoch": 0.42681775259678945,
      "grad_norm": 1.4575614929199219,
      "learning_rate": 4.738320218785281e-05,
      "loss": 0.5396,
      "step": 113
    },
    {
      "epoch": 0.43059490084985835,
      "grad_norm": 1.4147710800170898,
      "learning_rate": 4.6512176312793736e-05,
      "loss": 0.5129,
      "step": 114
    },
    {
      "epoch": 0.4343720491029273,
      "grad_norm": 1.4170805215835571,
      "learning_rate": 4.564221286261709e-05,
      "loss": 0.5948,
      "step": 115
    },
    {
      "epoch": 0.4381491973559962,
      "grad_norm": 1.3644853830337524,
      "learning_rate": 4.477357683661734e-05,
      "loss": 0.556,
      "step": 116
    },
    {
      "epoch": 0.44192634560906513,
      "grad_norm": 1.5001084804534912,
      "learning_rate": 4.390653282974264e-05,
      "loss": 0.5968,
      "step": 117
    },
    {
      "epoch": 0.4457034938621341,
      "grad_norm": 1.5384044647216797,
      "learning_rate": 4.3041344951996746e-05,
      "loss": 0.6292,
      "step": 118
    },
    {
      "epoch": 0.449480642115203,
      "grad_norm": 1.4583220481872559,
      "learning_rate": 4.2178276747988446e-05,
      "loss": 0.5902,
      "step": 119
    },
    {
      "epoch": 0.45325779036827196,
      "grad_norm": 1.6002942323684692,
      "learning_rate": 4.131759111665349e-05,
      "loss": 0.5365,
      "step": 120
    },
    {
      "epoch": 0.45703493862134087,
      "grad_norm": 1.4325772523880005,
      "learning_rate": 4.045955023117276e-05,
      "loss": 0.5213,
      "step": 121
    },
    {
      "epoch": 0.46081208687440983,
      "grad_norm": 1.4905633926391602,
      "learning_rate": 3.960441545911204e-05,
      "loss": 0.5081,
      "step": 122
    },
    {
      "epoch": 0.46458923512747874,
      "grad_norm": 1.4398701190948486,
      "learning_rate": 3.875244728280676e-05,
      "loss": 0.4873,
      "step": 123
    },
    {
      "epoch": 0.4683663833805477,
      "grad_norm": 1.4782718420028687,
      "learning_rate": 3.790390522001662e-05,
      "loss": 0.478,
      "step": 124
    },
    {
      "epoch": 0.4721435316336166,
      "grad_norm": 1.7875652313232422,
      "learning_rate": 3.705904774487396e-05,
      "loss": 0.6598,
      "step": 125
    },
    {
      "epoch": 0.47592067988668557,
      "grad_norm": 1.6041302680969238,
      "learning_rate": 3.6218132209150045e-05,
      "loss": 0.541,
      "step": 126
    },
    {
      "epoch": 0.4796978281397545,
      "grad_norm": 1.499057412147522,
      "learning_rate": 3.5381414763863166e-05,
      "loss": 0.5327,
      "step": 127
    },
    {
      "epoch": 0.48347497639282344,
      "grad_norm": 1.6440989971160889,
      "learning_rate": 3.4549150281252636e-05,
      "loss": 0.5406,
      "step": 128
    },
    {
      "epoch": 0.48725212464589235,
      "grad_norm": 1.5986462831497192,
      "learning_rate": 3.372159227714218e-05,
      "loss": 0.4991,
      "step": 129
    },
    {
      "epoch": 0.4910292728989613,
      "grad_norm": 1.5090078115463257,
      "learning_rate": 3.289899283371657e-05,
      "loss": 0.5107,
      "step": 130
    },
    {
      "epoch": 0.4948064211520302,
      "grad_norm": 1.6454864740371704,
      "learning_rate": 3.2081602522734986e-05,
      "loss": 0.5222,
      "step": 131
    },
    {
      "epoch": 0.4985835694050991,
      "grad_norm": 1.7272270917892456,
      "learning_rate": 3.12696703292044e-05,
      "loss": 0.5258,
      "step": 132
    },
    {
      "epoch": 0.502360717658168,
      "grad_norm": 1.7388275861740112,
      "learning_rate": 3.046344357553632e-05,
      "loss": 0.5442,
      "step": 133
    },
    {
      "epoch": 0.506137865911237,
      "grad_norm": 2.111438035964966,
      "learning_rate": 2.9663167846209998e-05,
      "loss": 0.6797,
      "step": 134
    },
    {
      "epoch": 0.509915014164306,
      "grad_norm": 1.8713425397872925,
      "learning_rate": 2.886908691296504e-05,
      "loss": 0.5494,
      "step": 135
    },
    {
      "epoch": 0.5136921624173749,
      "grad_norm": 1.8569631576538086,
      "learning_rate": 2.8081442660546125e-05,
      "loss": 0.5905,
      "step": 136
    },
    {
      "epoch": 0.5174693106704438,
      "grad_norm": 1.817326307296753,
      "learning_rate": 2.7300475013022663e-05,
      "loss": 0.6306,
      "step": 137
    },
    {
      "epoch": 0.5212464589235127,
      "grad_norm": 2.2332284450531006,
      "learning_rate": 2.6526421860705473e-05,
      "loss": 0.618,
      "step": 138
    },
    {
      "epoch": 0.5250236071765817,
      "grad_norm": 1.9101030826568604,
      "learning_rate": 2.575951898768315e-05,
      "loss": 0.6246,
      "step": 139
    },
    {
      "epoch": 0.5288007554296507,
      "grad_norm": 2.1395204067230225,
      "learning_rate": 2.500000000000001e-05,
      "loss": 0.7272,
      "step": 140
    },
    {
      "epoch": 0.5325779036827195,
      "grad_norm": 2.0194380283355713,
      "learning_rate": 2.4248096254497288e-05,
      "loss": 0.5592,
      "step": 141
    },
    {
      "epoch": 0.5363550519357885,
      "grad_norm": 2.1775777339935303,
      "learning_rate": 2.350403678833976e-05,
      "loss": 0.6584,
      "step": 142
    },
    {
      "epoch": 0.5401322001888574,
      "grad_norm": 2.397078514099121,
      "learning_rate": 2.2768048249248648e-05,
      "loss": 0.737,
      "step": 143
    },
    {
      "epoch": 0.5439093484419264,
      "grad_norm": 2.925100564956665,
      "learning_rate": 2.2040354826462668e-05,
      "loss": 1.1214,
      "step": 144
    },
    {
      "epoch": 0.5476864966949953,
      "grad_norm": 3.194314956665039,
      "learning_rate": 2.132117818244771e-05,
      "loss": 0.8998,
      "step": 145
    },
    {
      "epoch": 0.5514636449480642,
      "grad_norm": 3.310019016265869,
      "learning_rate": 2.061073738537635e-05,
      "loss": 0.9178,
      "step": 146
    },
    {
      "epoch": 0.5552407932011332,
      "grad_norm": 3.959937810897827,
      "learning_rate": 1.9909248842397584e-05,
      "loss": 1.039,
      "step": 147
    },
    {
      "epoch": 0.559017941454202,
      "grad_norm": 3.4178574085235596,
      "learning_rate": 1.9216926233717085e-05,
      "loss": 0.9191,
      "step": 148
    },
    {
      "epoch": 0.562795089707271,
      "grad_norm": 3.7735092639923096,
      "learning_rate": 1.8533980447508137e-05,
      "loss": 1.0148,
      "step": 149
    },
    {
      "epoch": 0.56657223796034,
      "grad_norm": 3.873361825942993,
      "learning_rate": 1.7860619515673033e-05,
      "loss": 0.9154,
      "step": 150
    },
    {
      "epoch": 0.56657223796034,
      "eval_loss": 0.6154063940048218,
      "eval_runtime": 10.8157,
      "eval_samples_per_second": 41.236,
      "eval_steps_per_second": 20.618,
      "step": 150
    },
    {
      "epoch": 0.5703493862134089,
      "grad_norm": 1.4619814157485962,
      "learning_rate": 1.7197048550474643e-05,
      "loss": 0.6626,
      "step": 151
    },
    {
      "epoch": 0.5741265344664778,
      "grad_norm": 1.577013373374939,
      "learning_rate": 1.6543469682057106e-05,
      "loss": 0.6268,
      "step": 152
    },
    {
      "epoch": 0.5779036827195467,
      "grad_norm": 1.6228933334350586,
      "learning_rate": 1.5900081996875083e-05,
      "loss": 0.6897,
      "step": 153
    },
    {
      "epoch": 0.5816808309726157,
      "grad_norm": 1.4395544528961182,
      "learning_rate": 1.526708147705013e-05,
      "loss": 0.5543,
      "step": 154
    },
    {
      "epoch": 0.5854579792256847,
      "grad_norm": 1.3829885721206665,
      "learning_rate": 1.4644660940672627e-05,
      "loss": 0.5733,
      "step": 155
    },
    {
      "epoch": 0.5892351274787535,
      "grad_norm": 1.574398159980774,
      "learning_rate": 1.4033009983067452e-05,
      "loss": 0.5928,
      "step": 156
    },
    {
      "epoch": 0.5930122757318225,
      "grad_norm": 1.5374937057495117,
      "learning_rate": 1.3432314919041478e-05,
      "loss": 0.5972,
      "step": 157
    },
    {
      "epoch": 0.5967894239848914,
      "grad_norm": 1.3958810567855835,
      "learning_rate": 1.2842758726130283e-05,
      "loss": 0.5215,
      "step": 158
    },
    {
      "epoch": 0.6005665722379604,
      "grad_norm": 1.3954333066940308,
      "learning_rate": 1.22645209888614e-05,
      "loss": 0.5054,
      "step": 159
    },
    {
      "epoch": 0.6043437204910292,
      "grad_norm": 1.3704341650009155,
      "learning_rate": 1.1697777844051105e-05,
      "loss": 0.5053,
      "step": 160
    },
    {
      "epoch": 0.6081208687440982,
      "grad_norm": 1.5128916501998901,
      "learning_rate": 1.1142701927151456e-05,
      "loss": 0.541,
      "step": 161
    },
    {
      "epoch": 0.6118980169971672,
      "grad_norm": 1.3563505411148071,
      "learning_rate": 1.0599462319663905e-05,
      "loss": 0.5203,
      "step": 162
    },
    {
      "epoch": 0.615675165250236,
      "grad_norm": 1.4357309341430664,
      "learning_rate": 1.006822449763537e-05,
      "loss": 0.5797,
      "step": 163
    },
    {
      "epoch": 0.619452313503305,
      "grad_norm": 1.450360894203186,
      "learning_rate": 9.549150281252633e-06,
      "loss": 0.4857,
      "step": 164
    },
    {
      "epoch": 0.623229461756374,
      "grad_norm": 1.3640488386154175,
      "learning_rate": 9.042397785550405e-06,
      "loss": 0.4458,
      "step": 165
    },
    {
      "epoch": 0.6270066100094429,
      "grad_norm": 1.3531254529953003,
      "learning_rate": 8.548121372247918e-06,
      "loss": 0.4801,
      "step": 166
    },
    {
      "epoch": 0.6307837582625118,
      "grad_norm": 1.5654860734939575,
      "learning_rate": 8.066471602728803e-06,
      "loss": 0.5205,
      "step": 167
    },
    {
      "epoch": 0.6345609065155807,
      "grad_norm": 1.317825436592102,
      "learning_rate": 7.597595192178702e-06,
      "loss": 0.4803,
      "step": 168
    },
    {
      "epoch": 0.6383380547686497,
      "grad_norm": 1.5071754455566406,
      "learning_rate": 7.1416349648943894e-06,
      "loss": 0.5241,
      "step": 169
    },
    {
      "epoch": 0.6421152030217187,
      "grad_norm": 1.5320953130722046,
      "learning_rate": 6.698729810778065e-06,
      "loss": 0.5702,
      "step": 170
    },
    {
      "epoch": 0.6458923512747875,
      "grad_norm": 1.5714699029922485,
      "learning_rate": 6.269014643030213e-06,
      "loss": 0.5699,
      "step": 171
    },
    {
      "epoch": 0.6496694995278565,
      "grad_norm": 1.6827765703201294,
      "learning_rate": 5.852620357053651e-06,
      "loss": 0.5662,
      "step": 172
    },
    {
      "epoch": 0.6534466477809254,
      "grad_norm": 1.4517920017242432,
      "learning_rate": 5.449673790581611e-06,
      "loss": 0.4732,
      "step": 173
    },
    {
      "epoch": 0.6572237960339944,
      "grad_norm": 1.5523104667663574,
      "learning_rate": 5.060297685041659e-06,
      "loss": 0.4938,
      "step": 174
    },
    {
      "epoch": 0.6610009442870632,
      "grad_norm": 1.5553407669067383,
      "learning_rate": 4.684610648167503e-06,
      "loss": 0.5367,
      "step": 175
    },
    {
      "epoch": 0.6647780925401322,
      "grad_norm": 1.4565545320510864,
      "learning_rate": 4.322727117869951e-06,
      "loss": 0.4542,
      "step": 176
    },
    {
      "epoch": 0.6685552407932012,
      "grad_norm": 1.441739797592163,
      "learning_rate": 3.974757327377981e-06,
      "loss": 0.4465,
      "step": 177
    },
    {
      "epoch": 0.67233238904627,
      "grad_norm": 1.5486420392990112,
      "learning_rate": 3.6408072716606346e-06,
      "loss": 0.4657,
      "step": 178
    },
    {
      "epoch": 0.676109537299339,
      "grad_norm": 1.7560228109359741,
      "learning_rate": 3.3209786751399187e-06,
      "loss": 0.6235,
      "step": 179
    },
    {
      "epoch": 0.6798866855524079,
      "grad_norm": 1.4693523645401,
      "learning_rate": 3.0153689607045845e-06,
      "loss": 0.4095,
      "step": 180
    },
    {
      "epoch": 0.6836638338054769,
      "grad_norm": 1.4747397899627686,
      "learning_rate": 2.724071220034158e-06,
      "loss": 0.4201,
      "step": 181
    },
    {
      "epoch": 0.6874409820585458,
      "grad_norm": 1.7705806493759155,
      "learning_rate": 2.4471741852423237e-06,
      "loss": 0.5636,
      "step": 182
    },
    {
      "epoch": 0.6912181303116147,
      "grad_norm": 1.8058946132659912,
      "learning_rate": 2.1847622018482283e-06,
      "loss": 0.5317,
      "step": 183
    },
    {
      "epoch": 0.6949952785646837,
      "grad_norm": 1.6872100830078125,
      "learning_rate": 1.9369152030840556e-06,
      "loss": 0.5318,
      "step": 184
    },
    {
      "epoch": 0.6987724268177526,
      "grad_norm": 1.9177495241165161,
      "learning_rate": 1.70370868554659e-06,
      "loss": 0.5182,
      "step": 185
    },
    {
      "epoch": 0.7025495750708215,
      "grad_norm": 1.7883721590042114,
      "learning_rate": 1.4852136862001764e-06,
      "loss": 0.5542,
      "step": 186
    },
    {
      "epoch": 0.7063267233238905,
      "grad_norm": 2.057225465774536,
      "learning_rate": 1.2814967607382432e-06,
      "loss": 0.6009,
      "step": 187
    },
    {
      "epoch": 0.7101038715769594,
      "grad_norm": 1.8016092777252197,
      "learning_rate": 1.0926199633097157e-06,
      "loss": 0.5909,
      "step": 188
    },
    {
      "epoch": 0.7138810198300283,
      "grad_norm": 1.9157992601394653,
      "learning_rate": 9.186408276168013e-07,
      "loss": 0.5691,
      "step": 189
    },
    {
      "epoch": 0.7176581680830972,
      "grad_norm": 1.9489667415618896,
      "learning_rate": 7.596123493895991e-07,
      "loss": 0.5426,
      "step": 190
    },
    {
      "epoch": 0.7214353163361662,
      "grad_norm": 2.052964925765991,
      "learning_rate": 6.15582970243117e-07,
      "loss": 0.5927,
      "step": 191
    },
    {
      "epoch": 0.7252124645892352,
      "grad_norm": 2.0972838401794434,
      "learning_rate": 4.865965629214819e-07,
      "loss": 0.5993,
      "step": 192
    },
    {
      "epoch": 0.728989612842304,
      "grad_norm": 2.2699410915374756,
      "learning_rate": 3.7269241793390085e-07,
      "loss": 0.7367,
      "step": 193
    },
    {
      "epoch": 0.732766761095373,
      "grad_norm": 3.033625602722168,
      "learning_rate": 2.7390523158633554e-07,
      "loss": 0.8378,
      "step": 194
    },
    {
      "epoch": 0.7365439093484419,
      "grad_norm": 3.547144651412964,
      "learning_rate": 1.9026509541272275e-07,
      "loss": 0.9025,
      "step": 195
    },
    {
      "epoch": 0.7403210576015109,
      "grad_norm": 3.2008984088897705,
      "learning_rate": 1.2179748700879012e-07,
      "loss": 0.8758,
      "step": 196
    },
    {
      "epoch": 0.7440982058545798,
      "grad_norm": 3.484177827835083,
      "learning_rate": 6.852326227130834e-08,
      "loss": 0.9472,
      "step": 197
    },
    {
      "epoch": 0.7478753541076487,
      "grad_norm": 4.3841400146484375,
      "learning_rate": 3.04586490452119e-08,
      "loss": 1.0913,
      "step": 198
    },
    {
      "epoch": 0.7516525023607177,
      "grad_norm": 4.56240177154541,
      "learning_rate": 7.615242180436522e-09,
      "loss": 1.1496,
      "step": 199
    },
    {
      "epoch": 0.7554296506137866,
      "grad_norm": 4.790390491485596,
      "learning_rate": 0.0,
      "loss": 0.9284,
      "step": 200
    },
    {
      "epoch": 0.7554296506137866,
      "eval_loss": 0.5842863321304321,
      "eval_runtime": 10.7782,
      "eval_samples_per_second": 41.38,
      "eval_steps_per_second": 20.69,
      "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": 5,
        "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": 1.5476286069669888e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}