{
  "best_global_step": 100000,
  "best_metric": 3.2714831829071045,
  "best_model_checkpoint": "/scratch/cl5625/exceptions/models/high_2000_495/checkpoint-100000",
  "epoch": 37.675333763996555,
  "eval_steps": 1000,
  "global_step": 140000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.013458225667527994,
      "grad_norm": 1.3015865087509155,
      "learning_rate": 0.000294,
      "loss": 8.5932,
      "step": 50
    },
    {
      "epoch": 0.026916451335055987,
      "grad_norm": 2.1779768466949463,
      "learning_rate": 0.0005939999999999999,
      "loss": 6.8393,
      "step": 100
    },
    {
      "epoch": 0.04037467700258398,
      "grad_norm": 1.2906817197799683,
      "learning_rate": 0.0005998020735155513,
      "loss": 6.3487,
      "step": 150
    },
    {
      "epoch": 0.053832902670111975,
      "grad_norm": 1.1974900960922241,
      "learning_rate": 0.0005996001077150935,
      "loss": 6.0858,
      "step": 200
    },
    {
      "epoch": 0.06729112833763996,
      "grad_norm": 1.116457462310791,
      "learning_rate": 0.0005993981419146358,
      "loss": 5.8927,
      "step": 250
    },
    {
      "epoch": 0.08074935400516796,
      "grad_norm": 1.3353101015090942,
      "learning_rate": 0.0005991961761141779,
      "loss": 5.7266,
      "step": 300
    },
    {
      "epoch": 0.09420757967269595,
      "grad_norm": 1.5416333675384521,
      "learning_rate": 0.0005989942103137202,
      "loss": 5.6088,
      "step": 350
    },
    {
      "epoch": 0.10766580534022395,
      "grad_norm": 1.8905576467514038,
      "learning_rate": 0.0005987922445132624,
      "loss": 5.477,
      "step": 400
    },
    {
      "epoch": 0.12112403100775193,
      "grad_norm": 1.7314372062683105,
      "learning_rate": 0.0005985902787128047,
      "loss": 5.361,
      "step": 450
    },
    {
      "epoch": 0.13458225667527993,
      "grad_norm": 1.5808533430099487,
      "learning_rate": 0.0005983883129123468,
      "loss": 5.2388,
      "step": 500
    },
    {
      "epoch": 0.14804048234280792,
      "grad_norm": 1.0360816717147827,
      "learning_rate": 0.000598186347111889,
      "loss": 5.15,
      "step": 550
    },
    {
      "epoch": 0.16149870801033592,
      "grad_norm": 1.282380223274231,
      "learning_rate": 0.0005979843813114312,
      "loss": 5.0461,
      "step": 600
    },
    {
      "epoch": 0.1749569336778639,
      "grad_norm": 0.9232556819915771,
      "learning_rate": 0.0005977824155109734,
      "loss": 4.9641,
      "step": 650
    },
    {
      "epoch": 0.1884151593453919,
      "grad_norm": 2.3980114459991455,
      "learning_rate": 0.0005975804497105157,
      "loss": 4.9106,
      "step": 700
    },
    {
      "epoch": 0.2018733850129199,
      "grad_norm": 0.9494748711585999,
      "learning_rate": 0.0005973784839100578,
      "loss": 4.8683,
      "step": 750
    },
    {
      "epoch": 0.2153316106804479,
      "grad_norm": 1.0844972133636475,
      "learning_rate": 0.0005971765181096001,
      "loss": 4.7836,
      "step": 800
    },
    {
      "epoch": 0.2287898363479759,
      "grad_norm": 1.1658976078033447,
      "learning_rate": 0.0005969745523091422,
      "loss": 4.7453,
      "step": 850
    },
    {
      "epoch": 0.24224806201550386,
      "grad_norm": 1.3042316436767578,
      "learning_rate": 0.0005967725865086845,
      "loss": 4.6924,
      "step": 900
    },
    {
      "epoch": 0.2557062876830319,
      "grad_norm": 1.0487701892852783,
      "learning_rate": 0.0005965706207082267,
      "loss": 4.6401,
      "step": 950
    },
    {
      "epoch": 0.26916451335055985,
      "grad_norm": 0.9355751872062683,
      "learning_rate": 0.0005963686549077689,
      "loss": 4.5772,
      "step": 1000
    },
    {
      "epoch": 0.26916451335055985,
      "eval_accuracy": 0.268002783967027,
      "eval_loss": 4.510490894317627,
      "eval_runtime": 84.3043,
      "eval_samples_per_second": 213.655,
      "eval_steps_per_second": 13.356,
      "step": 1000
    },
    {
      "epoch": 0.2826227390180879,
      "grad_norm": 0.6489158868789673,
      "learning_rate": 0.0005961666891073111,
      "loss": 4.5474,
      "step": 1050
    },
    {
      "epoch": 0.29608096468561584,
      "grad_norm": 1.284758448600769,
      "learning_rate": 0.0005959647233068533,
      "loss": 4.5052,
      "step": 1100
    },
    {
      "epoch": 0.30953919035314387,
      "grad_norm": 0.9760456681251526,
      "learning_rate": 0.0005957627575063955,
      "loss": 4.4691,
      "step": 1150
    },
    {
      "epoch": 0.32299741602067183,
      "grad_norm": 0.868435800075531,
      "learning_rate": 0.0005955607917059378,
      "loss": 4.4288,
      "step": 1200
    },
    {
      "epoch": 0.3364556416881998,
      "grad_norm": 0.847154974937439,
      "learning_rate": 0.00059535882590548,
      "loss": 4.4004,
      "step": 1250
    },
    {
      "epoch": 0.3499138673557278,
      "grad_norm": 0.7761279940605164,
      "learning_rate": 0.0005951568601050221,
      "loss": 4.3585,
      "step": 1300
    },
    {
      "epoch": 0.3633720930232558,
      "grad_norm": 0.7720751762390137,
      "learning_rate": 0.0005949548943045644,
      "loss": 4.3319,
      "step": 1350
    },
    {
      "epoch": 0.3768303186907838,
      "grad_norm": 0.8076860904693604,
      "learning_rate": 0.0005947529285041066,
      "loss": 4.3024,
      "step": 1400
    },
    {
      "epoch": 0.3902885443583118,
      "grad_norm": 1.030658483505249,
      "learning_rate": 0.0005945509627036488,
      "loss": 4.2728,
      "step": 1450
    },
    {
      "epoch": 0.4037467700258398,
      "grad_norm": 0.7001477479934692,
      "learning_rate": 0.000594348996903191,
      "loss": 4.2453,
      "step": 1500
    },
    {
      "epoch": 0.4172049956933678,
      "grad_norm": 0.8518731594085693,
      "learning_rate": 0.0005941470311027332,
      "loss": 4.2371,
      "step": 1550
    },
    {
      "epoch": 0.4306632213608958,
      "grad_norm": 0.9659989476203918,
      "learning_rate": 0.0005939450653022754,
      "loss": 4.2158,
      "step": 1600
    },
    {
      "epoch": 0.44412144702842377,
      "grad_norm": 0.7416383624076843,
      "learning_rate": 0.0005937430995018177,
      "loss": 4.1908,
      "step": 1650
    },
    {
      "epoch": 0.4575796726959518,
      "grad_norm": 0.6823920011520386,
      "learning_rate": 0.0005935411337013598,
      "loss": 4.1779,
      "step": 1700
    },
    {
      "epoch": 0.47103789836347976,
      "grad_norm": 0.7570599913597107,
      "learning_rate": 0.0005933391679009021,
      "loss": 4.1583,
      "step": 1750
    },
    {
      "epoch": 0.4844961240310077,
      "grad_norm": 0.7106193900108337,
      "learning_rate": 0.0005931372021004442,
      "loss": 4.1437,
      "step": 1800
    },
    {
      "epoch": 0.49795434969853575,
      "grad_norm": 0.8265014290809631,
      "learning_rate": 0.0005929352362999865,
      "loss": 4.1209,
      "step": 1850
    },
    {
      "epoch": 0.5114125753660638,
      "grad_norm": 0.7051160335540771,
      "learning_rate": 0.0005927332704995287,
      "loss": 4.1152,
      "step": 1900
    },
    {
      "epoch": 0.5248708010335917,
      "grad_norm": 0.7301550507545471,
      "learning_rate": 0.0005925313046990709,
      "loss": 4.098,
      "step": 1950
    },
    {
      "epoch": 0.5383290267011197,
      "grad_norm": 0.6782756447792053,
      "learning_rate": 0.0005923293388986131,
      "loss": 4.0956,
      "step": 2000
    },
    {
      "epoch": 0.5383290267011197,
      "eval_accuracy": 0.31626056644993794,
      "eval_loss": 4.0294365882873535,
      "eval_runtime": 83.1348,
      "eval_samples_per_second": 216.66,
      "eval_steps_per_second": 13.544,
      "step": 2000
    },
    {
      "epoch": 0.5517872523686477,
      "grad_norm": 0.6445055603981018,
      "learning_rate": 0.0005921273730981553,
      "loss": 4.0778,
      "step": 2050
    },
    {
      "epoch": 0.5652454780361758,
      "grad_norm": 0.6702537536621094,
      "learning_rate": 0.0005919254072976975,
      "loss": 4.0567,
      "step": 2100
    },
    {
      "epoch": 0.5787037037037037,
      "grad_norm": 0.7342326641082764,
      "learning_rate": 0.0005917234414972398,
      "loss": 4.0573,
      "step": 2150
    },
    {
      "epoch": 0.5921619293712317,
      "grad_norm": 0.8136904239654541,
      "learning_rate": 0.000591521475696782,
      "loss": 4.0324,
      "step": 2200
    },
    {
      "epoch": 0.6056201550387597,
      "grad_norm": 0.5917749404907227,
      "learning_rate": 0.0005913195098963241,
      "loss": 4.0329,
      "step": 2250
    },
    {
      "epoch": 0.6190783807062877,
      "grad_norm": 0.706507682800293,
      "learning_rate": 0.0005911175440958664,
      "loss": 4.0086,
      "step": 2300
    },
    {
      "epoch": 0.6325366063738157,
      "grad_norm": 0.5580946803092957,
      "learning_rate": 0.0005909155782954086,
      "loss": 4.0081,
      "step": 2350
    },
    {
      "epoch": 0.6459948320413437,
      "grad_norm": 0.5534974336624146,
      "learning_rate": 0.0005907136124949508,
      "loss": 4.0013,
      "step": 2400
    },
    {
      "epoch": 0.6594530577088716,
      "grad_norm": 0.7119669914245605,
      "learning_rate": 0.000590511646694493,
      "loss": 3.9868,
      "step": 2450
    },
    {
      "epoch": 0.6729112833763996,
      "grad_norm": 0.603784441947937,
      "learning_rate": 0.0005903096808940352,
      "loss": 3.977,
      "step": 2500
    },
    {
      "epoch": 0.6863695090439277,
      "grad_norm": 0.5871282815933228,
      "learning_rate": 0.0005901077150935774,
      "loss": 3.9565,
      "step": 2550
    },
    {
      "epoch": 0.6998277347114557,
      "grad_norm": 0.5962157845497131,
      "learning_rate": 0.0005899057492931197,
      "loss": 3.9566,
      "step": 2600
    },
    {
      "epoch": 0.7132859603789836,
      "grad_norm": 0.5701152682304382,
      "learning_rate": 0.0005897037834926618,
      "loss": 3.9548,
      "step": 2650
    },
    {
      "epoch": 0.7267441860465116,
      "grad_norm": 0.5659810304641724,
      "learning_rate": 0.0005895018176922041,
      "loss": 3.9526,
      "step": 2700
    },
    {
      "epoch": 0.7402024117140397,
      "grad_norm": 0.5718296766281128,
      "learning_rate": 0.0005892998518917462,
      "loss": 3.9246,
      "step": 2750
    },
    {
      "epoch": 0.7536606373815676,
      "grad_norm": 0.5351820588111877,
      "learning_rate": 0.0005890978860912885,
      "loss": 3.9399,
      "step": 2800
    },
    {
      "epoch": 0.7671188630490956,
      "grad_norm": 0.4873688220977783,
      "learning_rate": 0.0005888959202908307,
      "loss": 3.9238,
      "step": 2850
    },
    {
      "epoch": 0.7805770887166236,
      "grad_norm": 0.530052125453949,
      "learning_rate": 0.0005886939544903729,
      "loss": 3.9127,
      "step": 2900
    },
    {
      "epoch": 0.7940353143841516,
      "grad_norm": 0.49622371792793274,
      "learning_rate": 0.0005884919886899151,
      "loss": 3.9003,
      "step": 2950
    },
    {
      "epoch": 0.8074935400516796,
      "grad_norm": 0.5243384838104248,
      "learning_rate": 0.0005882900228894573,
      "loss": 3.8948,
      "step": 3000
    },
    {
      "epoch": 0.8074935400516796,
      "eval_accuracy": 0.3333123644902094,
      "eval_loss": 3.844620943069458,
      "eval_runtime": 83.1619,
      "eval_samples_per_second": 216.59,
      "eval_steps_per_second": 13.54,
      "step": 3000
    },
    {
      "epoch": 0.8209517657192076,
      "grad_norm": 0.4867589473724365,
      "learning_rate": 0.0005880880570889996,
      "loss": 3.909,
      "step": 3050
    },
    {
      "epoch": 0.8344099913867356,
      "grad_norm": 0.5779995322227478,
      "learning_rate": 0.0005878860912885418,
      "loss": 3.8856,
      "step": 3100
    },
    {
      "epoch": 0.8478682170542635,
      "grad_norm": 0.4581291973590851,
      "learning_rate": 0.000587684125488084,
      "loss": 3.8827,
      "step": 3150
    },
    {
      "epoch": 0.8613264427217916,
      "grad_norm": 0.5086135864257812,
      "learning_rate": 0.0005874821596876261,
      "loss": 3.8693,
      "step": 3200
    },
    {
      "epoch": 0.8747846683893196,
      "grad_norm": 0.5387232899665833,
      "learning_rate": 0.0005872801938871684,
      "loss": 3.8791,
      "step": 3250
    },
    {
      "epoch": 0.8882428940568475,
      "grad_norm": 0.4933193325996399,
      "learning_rate": 0.0005870782280867106,
      "loss": 3.8571,
      "step": 3300
    },
    {
      "epoch": 0.9017011197243755,
      "grad_norm": 0.49710357189178467,
      "learning_rate": 0.0005868762622862528,
      "loss": 3.8508,
      "step": 3350
    },
    {
      "epoch": 0.9151593453919036,
      "grad_norm": 0.47787806391716003,
      "learning_rate": 0.000586674296485795,
      "loss": 3.843,
      "step": 3400
    },
    {
      "epoch": 0.9286175710594315,
      "grad_norm": 0.5095878839492798,
      "learning_rate": 0.0005864723306853373,
      "loss": 3.8431,
      "step": 3450
    },
    {
      "epoch": 0.9420757967269595,
      "grad_norm": 0.4359574615955353,
      "learning_rate": 0.0005862703648848794,
      "loss": 3.8346,
      "step": 3500
    },
    {
      "epoch": 0.9555340223944875,
      "grad_norm": 0.44693535566329956,
      "learning_rate": 0.0005860683990844217,
      "loss": 3.8396,
      "step": 3550
    },
    {
      "epoch": 0.9689922480620154,
      "grad_norm": 0.45087817311286926,
      "learning_rate": 0.0005858664332839638,
      "loss": 3.8303,
      "step": 3600
    },
    {
      "epoch": 0.9824504737295435,
      "grad_norm": 0.42239823937416077,
      "learning_rate": 0.0005856644674835061,
      "loss": 3.8243,
      "step": 3650
    },
    {
      "epoch": 0.9959086993970715,
      "grad_norm": 0.4339318573474884,
      "learning_rate": 0.0005854625016830483,
      "loss": 3.8261,
      "step": 3700
    },
    {
      "epoch": 1.009151593453919,
      "grad_norm": 0.4108366370201111,
      "learning_rate": 0.0005852605358825905,
      "loss": 3.77,
      "step": 3750
    },
    {
      "epoch": 1.022609819121447,
      "grad_norm": 0.4174451231956482,
      "learning_rate": 0.0005850585700821327,
      "loss": 3.7508,
      "step": 3800
    },
    {
      "epoch": 1.036068044788975,
      "grad_norm": 0.49247533082962036,
      "learning_rate": 0.0005848566042816749,
      "loss": 3.7516,
      "step": 3850
    },
    {
      "epoch": 1.049526270456503,
      "grad_norm": 0.4506925344467163,
      "learning_rate": 0.0005846546384812171,
      "loss": 3.7506,
      "step": 3900
    },
    {
      "epoch": 1.062984496124031,
      "grad_norm": 0.4733441174030304,
      "learning_rate": 0.0005844526726807594,
      "loss": 3.7456,
      "step": 3950
    },
    {
      "epoch": 1.076442721791559,
      "grad_norm": 0.4067009687423706,
      "learning_rate": 0.0005842507068803016,
      "loss": 3.744,
      "step": 4000
    },
    {
      "epoch": 1.076442721791559,
      "eval_accuracy": 0.34432709135418743,
      "eval_loss": 3.734100341796875,
      "eval_runtime": 83.2662,
      "eval_samples_per_second": 216.318,
      "eval_steps_per_second": 13.523,
      "step": 4000
    },
    {
      "epoch": 1.089900947459087,
      "grad_norm": 0.4497452974319458,
      "learning_rate": 0.0005840487410798437,
      "loss": 3.7317,
      "step": 4050
    },
    {
      "epoch": 1.103359173126615,
      "grad_norm": 0.46182072162628174,
      "learning_rate": 0.000583846775279386,
      "loss": 3.7267,
      "step": 4100
    },
    {
      "epoch": 1.116817398794143,
      "grad_norm": 0.46391478180885315,
      "learning_rate": 0.0005836448094789282,
      "loss": 3.7289,
      "step": 4150
    },
    {
      "epoch": 1.130275624461671,
      "grad_norm": 0.4440414607524872,
      "learning_rate": 0.0005834428436784704,
      "loss": 3.7305,
      "step": 4200
    },
    {
      "epoch": 1.143733850129199,
      "grad_norm": 0.45285120606422424,
      "learning_rate": 0.0005832408778780126,
      "loss": 3.7352,
      "step": 4250
    },
    {
      "epoch": 1.157192075796727,
      "grad_norm": 0.447615385055542,
      "learning_rate": 0.0005830389120775548,
      "loss": 3.7227,
      "step": 4300
    },
    {
      "epoch": 1.1706503014642549,
      "grad_norm": 0.5164969563484192,
      "learning_rate": 0.000582836946277097,
      "loss": 3.7114,
      "step": 4350
    },
    {
      "epoch": 1.1841085271317828,
      "grad_norm": 0.43993330001831055,
      "learning_rate": 0.0005826349804766393,
      "loss": 3.7249,
      "step": 4400
    },
    {
      "epoch": 1.197566752799311,
      "grad_norm": 0.4666627049446106,
      "learning_rate": 0.0005824330146761814,
      "loss": 3.7117,
      "step": 4450
    },
    {
      "epoch": 1.211024978466839,
      "grad_norm": 0.4921836853027344,
      "learning_rate": 0.0005822310488757237,
      "loss": 3.7116,
      "step": 4500
    },
    {
      "epoch": 1.224483204134367,
      "grad_norm": 0.45889997482299805,
      "learning_rate": 0.0005820290830752658,
      "loss": 3.708,
      "step": 4550
    },
    {
      "epoch": 1.237941429801895,
      "grad_norm": 0.4475299119949341,
      "learning_rate": 0.0005818271172748081,
      "loss": 3.7069,
      "step": 4600
    },
    {
      "epoch": 1.251399655469423,
      "grad_norm": 0.4270066022872925,
      "learning_rate": 0.0005816251514743503,
      "loss": 3.7052,
      "step": 4650
    },
    {
      "epoch": 1.2648578811369509,
      "grad_norm": 0.44844046235084534,
      "learning_rate": 0.0005814231856738925,
      "loss": 3.7004,
      "step": 4700
    },
    {
      "epoch": 1.2783161068044788,
      "grad_norm": 0.4396093189716339,
      "learning_rate": 0.0005812212198734347,
      "loss": 3.6912,
      "step": 4750
    },
    {
      "epoch": 1.291774332472007,
      "grad_norm": 0.41136953234672546,
      "learning_rate": 0.0005810192540729769,
      "loss": 3.6972,
      "step": 4800
    },
    {
      "epoch": 1.3052325581395348,
      "grad_norm": 0.44841718673706055,
      "learning_rate": 0.0005808172882725192,
      "loss": 3.6882,
      "step": 4850
    },
    {
      "epoch": 1.318690783807063,
      "grad_norm": 0.4548879563808441,
      "learning_rate": 0.0005806153224720614,
      "loss": 3.6951,
      "step": 4900
    },
    {
      "epoch": 1.332149009474591,
      "grad_norm": 0.42023366689682007,
      "learning_rate": 0.0005804133566716036,
      "loss": 3.6777,
      "step": 4950
    },
    {
      "epoch": 1.3456072351421189,
      "grad_norm": 0.4213683605194092,
      "learning_rate": 0.0005802113908711457,
      "loss": 3.6945,
      "step": 5000
    },
    {
      "epoch": 1.3456072351421189,
      "eval_accuracy": 0.3510410324406473,
      "eval_loss": 3.663517713546753,
      "eval_runtime": 83.5912,
      "eval_samples_per_second": 215.477,
      "eval_steps_per_second": 13.47,
      "step": 5000
    },
    {
      "epoch": 1.3590654608096469,
      "grad_norm": 0.42015111446380615,
      "learning_rate": 0.000580009425070688,
      "loss": 3.6753,
      "step": 5050
    },
    {
      "epoch": 1.3725236864771748,
      "grad_norm": 0.4121742844581604,
      "learning_rate": 0.0005798074592702302,
      "loss": 3.6832,
      "step": 5100
    },
    {
      "epoch": 1.3859819121447028,
      "grad_norm": 0.4617924094200134,
      "learning_rate": 0.0005796054934697724,
      "loss": 3.6709,
      "step": 5150
    },
    {
      "epoch": 1.3994401378122308,
      "grad_norm": 0.4551539123058319,
      "learning_rate": 0.0005794035276693146,
      "loss": 3.6847,
      "step": 5200
    },
    {
      "epoch": 1.412898363479759,
      "grad_norm": 0.3819875121116638,
      "learning_rate": 0.0005792015618688568,
      "loss": 3.6788,
      "step": 5250
    },
    {
      "epoch": 1.4263565891472867,
      "grad_norm": 0.4178910553455353,
      "learning_rate": 0.000578999596068399,
      "loss": 3.6683,
      "step": 5300
    },
    {
      "epoch": 1.4398148148148149,
      "grad_norm": 0.4413163363933563,
      "learning_rate": 0.0005787976302679413,
      "loss": 3.6667,
      "step": 5350
    },
    {
      "epoch": 1.4532730404823428,
      "grad_norm": 0.4710264801979065,
      "learning_rate": 0.0005785956644674834,
      "loss": 3.6721,
      "step": 5400
    },
    {
      "epoch": 1.4667312661498708,
      "grad_norm": 0.3894467055797577,
      "learning_rate": 0.0005783936986670257,
      "loss": 3.6537,
      "step": 5450
    },
    {
      "epoch": 1.4801894918173988,
      "grad_norm": 0.367308109998703,
      "learning_rate": 0.0005781917328665678,
      "loss": 3.6552,
      "step": 5500
    },
    {
      "epoch": 1.4936477174849268,
      "grad_norm": 0.4133712947368622,
      "learning_rate": 0.0005779897670661101,
      "loss": 3.6636,
      "step": 5550
    },
    {
      "epoch": 1.507105943152455,
      "grad_norm": 0.37521886825561523,
      "learning_rate": 0.0005777878012656523,
      "loss": 3.6566,
      "step": 5600
    },
    {
      "epoch": 1.5205641688199827,
      "grad_norm": 0.38877028226852417,
      "learning_rate": 0.0005775858354651945,
      "loss": 3.6613,
      "step": 5650
    },
    {
      "epoch": 1.5340223944875109,
      "grad_norm": 0.3612709939479828,
      "learning_rate": 0.0005773838696647367,
      "loss": 3.6545,
      "step": 5700
    },
    {
      "epoch": 1.5474806201550386,
      "grad_norm": 0.4424472153186798,
      "learning_rate": 0.000577181903864279,
      "loss": 3.6617,
      "step": 5750
    },
    {
      "epoch": 1.5609388458225668,
      "grad_norm": 0.4011629521846771,
      "learning_rate": 0.0005769799380638212,
      "loss": 3.6456,
      "step": 5800
    },
    {
      "epoch": 1.5743970714900948,
      "grad_norm": 0.373329758644104,
      "learning_rate": 0.0005767779722633633,
      "loss": 3.647,
      "step": 5850
    },
    {
      "epoch": 1.5878552971576227,
      "grad_norm": 0.37653112411499023,
      "learning_rate": 0.0005765760064629056,
      "loss": 3.6312,
      "step": 5900
    },
    {
      "epoch": 1.6013135228251507,
      "grad_norm": 0.3729380965232849,
      "learning_rate": 0.0005763740406624477,
      "loss": 3.6448,
      "step": 5950
    },
    {
      "epoch": 1.6147717484926787,
      "grad_norm": 0.4230109453201294,
      "learning_rate": 0.00057617207486199,
      "loss": 3.6366,
      "step": 6000
    },
    {
      "epoch": 1.6147717484926787,
      "eval_accuracy": 0.3560429163771228,
      "eval_loss": 3.611274480819702,
      "eval_runtime": 83.3296,
      "eval_samples_per_second": 216.154,
      "eval_steps_per_second": 13.513,
      "step": 6000
    },
    {
      "epoch": 1.6282299741602069,
      "grad_norm": 0.43084651231765747,
      "learning_rate": 0.0005759701090615322,
      "loss": 3.6301,
      "step": 6050
    },
    {
      "epoch": 1.6416881998277346,
      "grad_norm": 0.41188517212867737,
      "learning_rate": 0.0005757681432610744,
      "loss": 3.6277,
      "step": 6100
    },
    {
      "epoch": 1.6551464254952628,
      "grad_norm": 0.37851613759994507,
      "learning_rate": 0.0005755661774606166,
      "loss": 3.6166,
      "step": 6150
    },
    {
      "epoch": 1.6686046511627906,
      "grad_norm": 0.39076900482177734,
      "learning_rate": 0.0005753642116601588,
      "loss": 3.625,
      "step": 6200
    },
    {
      "epoch": 1.6820628768303187,
      "grad_norm": 0.3747098743915558,
      "learning_rate": 0.000575162245859701,
      "loss": 3.6228,
      "step": 6250
    },
    {
      "epoch": 1.6955211024978467,
      "grad_norm": 0.41028639674186707,
      "learning_rate": 0.0005749602800592433,
      "loss": 3.6209,
      "step": 6300
    },
    {
      "epoch": 1.7089793281653747,
      "grad_norm": 0.40048956871032715,
      "learning_rate": 0.0005747583142587854,
      "loss": 3.6224,
      "step": 6350
    },
    {
      "epoch": 1.7224375538329026,
      "grad_norm": 0.36761754751205444,
      "learning_rate": 0.0005745563484583277,
      "loss": 3.625,
      "step": 6400
    },
    {
      "epoch": 1.7358957795004306,
      "grad_norm": 0.45381850004196167,
      "learning_rate": 0.0005743543826578698,
      "loss": 3.6201,
      "step": 6450
    },
    {
      "epoch": 1.7493540051679588,
      "grad_norm": 0.4151131808757782,
      "learning_rate": 0.0005741524168574121,
      "loss": 3.616,
      "step": 6500
    },
    {
      "epoch": 1.7628122308354865,
      "grad_norm": 0.37309277057647705,
      "learning_rate": 0.0005739504510569543,
      "loss": 3.6148,
      "step": 6550
    },
    {
      "epoch": 1.7762704565030147,
      "grad_norm": 0.376087486743927,
      "learning_rate": 0.0005737484852564966,
      "loss": 3.6054,
      "step": 6600
    },
    {
      "epoch": 1.7897286821705425,
      "grad_norm": 0.39143335819244385,
      "learning_rate": 0.0005735465194560387,
      "loss": 3.6151,
      "step": 6650
    },
    {
      "epoch": 1.8031869078380707,
      "grad_norm": 0.383264422416687,
      "learning_rate": 0.000573344553655581,
      "loss": 3.6206,
      "step": 6700
    },
    {
      "epoch": 1.8166451335055986,
      "grad_norm": 0.3560148775577545,
      "learning_rate": 0.0005731425878551232,
      "loss": 3.6161,
      "step": 6750
    },
    {
      "epoch": 1.8301033591731266,
      "grad_norm": 0.3778025507926941,
      "learning_rate": 0.0005729406220546653,
      "loss": 3.6042,
      "step": 6800
    },
    {
      "epoch": 1.8435615848406546,
      "grad_norm": 0.389453649520874,
      "learning_rate": 0.0005727386562542076,
      "loss": 3.6072,
      "step": 6850
    },
    {
      "epoch": 1.8570198105081825,
      "grad_norm": 0.3847193717956543,
      "learning_rate": 0.0005725366904537497,
      "loss": 3.5988,
      "step": 6900
    },
    {
      "epoch": 1.8704780361757107,
      "grad_norm": 0.3852189779281616,
      "learning_rate": 0.000572334724653292,
      "loss": 3.6107,
      "step": 6950
    },
    {
      "epoch": 1.8839362618432385,
      "grad_norm": 0.36440926790237427,
      "learning_rate": 0.0005721327588528342,
      "loss": 3.6008,
      "step": 7000
    },
    {
      "epoch": 1.8839362618432385,
      "eval_accuracy": 0.36019833266189577,
      "eval_loss": 3.5670289993286133,
      "eval_runtime": 83.3424,
      "eval_samples_per_second": 216.12,
      "eval_steps_per_second": 13.511,
      "step": 7000
    },
    {
      "epoch": 1.8973944875107667,
      "grad_norm": 0.35855185985565186,
      "learning_rate": 0.0005719307930523764,
      "loss": 3.5864,
      "step": 7050
    },
    {
      "epoch": 1.9108527131782944,
      "grad_norm": 0.37212860584259033,
      "learning_rate": 0.0005717288272519186,
      "loss": 3.5949,
      "step": 7100
    },
    {
      "epoch": 1.9243109388458226,
      "grad_norm": 0.34701022505760193,
      "learning_rate": 0.0005715268614514608,
      "loss": 3.5855,
      "step": 7150
    },
    {
      "epoch": 1.9377691645133506,
      "grad_norm": 0.36765119433403015,
      "learning_rate": 0.000571324895651003,
      "loss": 3.5943,
      "step": 7200
    },
    {
      "epoch": 1.9512273901808785,
      "grad_norm": 0.3625155985355377,
      "learning_rate": 0.0005711229298505453,
      "loss": 3.5861,
      "step": 7250
    },
    {
      "epoch": 1.9646856158484065,
      "grad_norm": 0.37622615694999695,
      "learning_rate": 0.0005709209640500874,
      "loss": 3.5874,
      "step": 7300
    },
    {
      "epoch": 1.9781438415159345,
      "grad_norm": 0.3395003378391266,
      "learning_rate": 0.0005707189982496297,
      "loss": 3.5798,
      "step": 7350
    },
    {
      "epoch": 1.9916020671834627,
      "grad_norm": 0.39717552065849304,
      "learning_rate": 0.0005705170324491719,
      "loss": 3.5818,
      "step": 7400
    },
    {
      "epoch": 2.00484496124031,
      "grad_norm": 0.3736705780029297,
      "learning_rate": 0.0005703150666487141,
      "loss": 3.5485,
      "step": 7450
    },
    {
      "epoch": 2.018303186907838,
      "grad_norm": 0.37968873977661133,
      "learning_rate": 0.0005701131008482563,
      "loss": 3.498,
      "step": 7500
    },
    {
      "epoch": 2.0317614125753662,
      "grad_norm": 0.34327080845832825,
      "learning_rate": 0.0005699111350477986,
      "loss": 3.5057,
      "step": 7550
    },
    {
      "epoch": 2.045219638242894,
      "grad_norm": 0.38864439725875854,
      "learning_rate": 0.0005697091692473408,
      "loss": 3.5015,
      "step": 7600
    },
    {
      "epoch": 2.058677863910422,
      "grad_norm": 0.40195029973983765,
      "learning_rate": 0.000569507203446883,
      "loss": 3.4928,
      "step": 7650
    },
    {
      "epoch": 2.07213608957795,
      "grad_norm": 0.3901396691799164,
      "learning_rate": 0.0005693052376464252,
      "loss": 3.5042,
      "step": 7700
    },
    {
      "epoch": 2.085594315245478,
      "grad_norm": 0.361380010843277,
      "learning_rate": 0.0005691032718459673,
      "loss": 3.5116,
      "step": 7750
    },
    {
      "epoch": 2.099052540913006,
      "grad_norm": 0.3697785437107086,
      "learning_rate": 0.0005689013060455096,
      "loss": 3.5011,
      "step": 7800
    },
    {
      "epoch": 2.112510766580534,
      "grad_norm": 0.3811769485473633,
      "learning_rate": 0.0005686993402450518,
      "loss": 3.502,
      "step": 7850
    },
    {
      "epoch": 2.125968992248062,
      "grad_norm": 0.33674320578575134,
      "learning_rate": 0.000568497374444594,
      "loss": 3.5095,
      "step": 7900
    },
    {
      "epoch": 2.13942721791559,
      "grad_norm": 0.39026930928230286,
      "learning_rate": 0.0005682954086441362,
      "loss": 3.5109,
      "step": 7950
    },
    {
      "epoch": 2.152885443583118,
      "grad_norm": 0.3860568404197693,
      "learning_rate": 0.0005680934428436784,
      "loss": 3.5165,
      "step": 8000
    },
    {
      "epoch": 2.152885443583118,
      "eval_accuracy": 0.36374728219890806,
      "eval_loss": 3.535961151123047,
      "eval_runtime": 83.2838,
      "eval_samples_per_second": 216.273,
      "eval_steps_per_second": 13.52,
      "step": 8000
    },
    {
      "epoch": 2.166343669250646,
      "grad_norm": 0.3673901855945587,
      "learning_rate": 0.0005678914770432206,
      "loss": 3.5063,
      "step": 8050
    },
    {
      "epoch": 2.179801894918174,
      "grad_norm": 0.3342771828174591,
      "learning_rate": 0.0005676895112427629,
      "loss": 3.5001,
      "step": 8100
    },
    {
      "epoch": 2.193260120585702,
      "grad_norm": 0.39416754245758057,
      "learning_rate": 0.000567487545442305,
      "loss": 3.5069,
      "step": 8150
    },
    {
      "epoch": 2.20671834625323,
      "grad_norm": 0.3709253966808319,
      "learning_rate": 0.0005672855796418473,
      "loss": 3.5057,
      "step": 8200
    },
    {
      "epoch": 2.220176571920758,
      "grad_norm": 0.40354976058006287,
      "learning_rate": 0.0005670836138413894,
      "loss": 3.4977,
      "step": 8250
    },
    {
      "epoch": 2.233634797588286,
      "grad_norm": 0.3623318672180176,
      "learning_rate": 0.0005668816480409317,
      "loss": 3.5063,
      "step": 8300
    },
    {
      "epoch": 2.2470930232558137,
      "grad_norm": 0.3478250503540039,
      "learning_rate": 0.0005666796822404739,
      "loss": 3.5046,
      "step": 8350
    },
    {
      "epoch": 2.260551248923342,
      "grad_norm": 0.37609925866127014,
      "learning_rate": 0.0005664777164400162,
      "loss": 3.4981,
      "step": 8400
    },
    {
      "epoch": 2.27400947459087,
      "grad_norm": 0.36902108788490295,
      "learning_rate": 0.0005662757506395583,
      "loss": 3.5006,
      "step": 8450
    },
    {
      "epoch": 2.287467700258398,
      "grad_norm": 0.349367618560791,
      "learning_rate": 0.0005660737848391006,
      "loss": 3.5046,
      "step": 8500
    },
    {
      "epoch": 2.300925925925926,
      "grad_norm": 0.37903445959091187,
      "learning_rate": 0.0005658718190386428,
      "loss": 3.5065,
      "step": 8550
    },
    {
      "epoch": 2.314384151593454,
      "grad_norm": 0.35075587034225464,
      "learning_rate": 0.000565669853238185,
      "loss": 3.4978,
      "step": 8600
    },
    {
      "epoch": 2.327842377260982,
      "grad_norm": 0.3737734854221344,
      "learning_rate": 0.0005654678874377272,
      "loss": 3.4991,
      "step": 8650
    },
    {
      "epoch": 2.3413006029285097,
      "grad_norm": 0.36533161997795105,
      "learning_rate": 0.0005652659216372693,
      "loss": 3.4981,
      "step": 8700
    },
    {
      "epoch": 2.354758828596038,
      "grad_norm": 0.38023629784584045,
      "learning_rate": 0.0005650639558368116,
      "loss": 3.5,
      "step": 8750
    },
    {
      "epoch": 2.3682170542635657,
      "grad_norm": 0.3384852409362793,
      "learning_rate": 0.0005648619900363538,
      "loss": 3.4945,
      "step": 8800
    },
    {
      "epoch": 2.381675279931094,
      "grad_norm": 0.35505008697509766,
      "learning_rate": 0.000564660024235896,
      "loss": 3.4965,
      "step": 8850
    },
    {
      "epoch": 2.395133505598622,
      "grad_norm": 0.38940027356147766,
      "learning_rate": 0.0005644580584354382,
      "loss": 3.4902,
      "step": 8900
    },
    {
      "epoch": 2.40859173126615,
      "grad_norm": 0.3613036274909973,
      "learning_rate": 0.0005642560926349804,
      "loss": 3.4838,
      "step": 8950
    },
    {
      "epoch": 2.422049956933678,
      "grad_norm": 0.3268532454967499,
      "learning_rate": 0.0005640541268345226,
      "loss": 3.5007,
      "step": 9000
    },
    {
      "epoch": 2.422049956933678,
      "eval_accuracy": 0.3667059533696388,
      "eval_loss": 3.5101945400238037,
      "eval_runtime": 83.2572,
      "eval_samples_per_second": 216.342,
      "eval_steps_per_second": 13.524,
      "step": 9000
    },
    {
      "epoch": 2.4355081826012057,
      "grad_norm": 0.3675191104412079,
      "learning_rate": 0.0005638521610340649,
      "loss": 3.5029,
      "step": 9050
    },
    {
      "epoch": 2.448966408268734,
      "grad_norm": 0.38351163268089294,
      "learning_rate": 0.000563650195233607,
      "loss": 3.4891,
      "step": 9100
    },
    {
      "epoch": 2.4624246339362617,
      "grad_norm": 0.34924453496932983,
      "learning_rate": 0.0005634482294331493,
      "loss": 3.4984,
      "step": 9150
    },
    {
      "epoch": 2.47588285960379,
      "grad_norm": 0.33259910345077515,
      "learning_rate": 0.0005632462636326914,
      "loss": 3.4922,
      "step": 9200
    },
    {
      "epoch": 2.4893410852713176,
      "grad_norm": 0.3981625735759735,
      "learning_rate": 0.0005630442978322337,
      "loss": 3.4899,
      "step": 9250
    },
    {
      "epoch": 2.502799310938846,
      "grad_norm": 0.3715845048427582,
      "learning_rate": 0.0005628423320317759,
      "loss": 3.4999,
      "step": 9300
    },
    {
      "epoch": 2.516257536606374,
      "grad_norm": 0.3902766704559326,
      "learning_rate": 0.0005626403662313182,
      "loss": 3.4926,
      "step": 9350
    },
    {
      "epoch": 2.5297157622739017,
      "grad_norm": 0.3570214807987213,
      "learning_rate": 0.0005624384004308603,
      "loss": 3.4801,
      "step": 9400
    },
    {
      "epoch": 2.54317398794143,
      "grad_norm": 0.3234241008758545,
      "learning_rate": 0.0005622364346304026,
      "loss": 3.4905,
      "step": 9450
    },
    {
      "epoch": 2.5566322136089576,
      "grad_norm": 0.35051244497299194,
      "learning_rate": 0.0005620344688299448,
      "loss": 3.4848,
      "step": 9500
    },
    {
      "epoch": 2.570090439276486,
      "grad_norm": 0.38920581340789795,
      "learning_rate": 0.0005618325030294869,
      "loss": 3.4857,
      "step": 9550
    },
    {
      "epoch": 2.583548664944014,
      "grad_norm": 0.37221840023994446,
      "learning_rate": 0.0005616305372290292,
      "loss": 3.4898,
      "step": 9600
    },
    {
      "epoch": 2.5970068906115418,
      "grad_norm": 0.3550560474395752,
      "learning_rate": 0.0005614285714285713,
      "loss": 3.4904,
      "step": 9650
    },
    {
      "epoch": 2.6104651162790695,
      "grad_norm": 0.37625744938850403,
      "learning_rate": 0.0005612266056281136,
      "loss": 3.4827,
      "step": 9700
    },
    {
      "epoch": 2.6239233419465977,
      "grad_norm": 0.3668668568134308,
      "learning_rate": 0.0005610246398276558,
      "loss": 3.4934,
      "step": 9750
    },
    {
      "epoch": 2.637381567614126,
      "grad_norm": 0.39287450909614563,
      "learning_rate": 0.000560822674027198,
      "loss": 3.4924,
      "step": 9800
    },
    {
      "epoch": 2.6508397932816536,
      "grad_norm": 0.332601934671402,
      "learning_rate": 0.0005606207082267402,
      "loss": 3.4869,
      "step": 9850
    },
    {
      "epoch": 2.664298018949182,
      "grad_norm": 0.3659941256046295,
      "learning_rate": 0.0005604187424262824,
      "loss": 3.4888,
      "step": 9900
    },
    {
      "epoch": 2.6777562446167096,
      "grad_norm": 0.3574960231781006,
      "learning_rate": 0.0005602167766258246,
      "loss": 3.4793,
      "step": 9950
    },
    {
      "epoch": 2.6912144702842378,
      "grad_norm": 0.36079293489456177,
      "learning_rate": 0.0005600148108253669,
      "loss": 3.4864,
      "step": 10000
    },
    {
      "epoch": 2.6912144702842378,
      "eval_accuracy": 0.3689281075064982,
      "eval_loss": 3.484949827194214,
      "eval_runtime": 83.331,
      "eval_samples_per_second": 216.15,
      "eval_steps_per_second": 13.512,
      "step": 10000
    },
    {
      "epoch": 2.704672695951766,
      "grad_norm": 0.3438493311405182,
      "learning_rate": 0.000559812845024909,
      "loss": 3.4899,
      "step": 10050
    },
    {
      "epoch": 2.7181309216192937,
      "grad_norm": 0.3419351875782013,
      "learning_rate": 0.0005596108792244513,
      "loss": 3.4786,
      "step": 10100
    },
    {
      "epoch": 2.7315891472868215,
      "grad_norm": 0.3579529821872711,
      "learning_rate": 0.0005594089134239935,
      "loss": 3.4743,
      "step": 10150
    },
    {
      "epoch": 2.7450473729543496,
      "grad_norm": 0.3665620982646942,
      "learning_rate": 0.0005592069476235358,
      "loss": 3.4809,
      "step": 10200
    },
    {
      "epoch": 2.758505598621878,
      "grad_norm": 0.36939823627471924,
      "learning_rate": 0.0005590049818230779,
      "loss": 3.4781,
      "step": 10250
    },
    {
      "epoch": 2.7719638242894056,
      "grad_norm": 0.3218006193637848,
      "learning_rate": 0.0005588030160226202,
      "loss": 3.4781,
      "step": 10300
    },
    {
      "epoch": 2.7854220499569338,
      "grad_norm": 0.3628097474575043,
      "learning_rate": 0.0005586010502221623,
      "loss": 3.4731,
      "step": 10350
    },
    {
      "epoch": 2.7988802756244615,
      "grad_norm": 0.3420850336551666,
      "learning_rate": 0.0005583990844217045,
      "loss": 3.4734,
      "step": 10400
    },
    {
      "epoch": 2.8123385012919897,
      "grad_norm": 0.335952490568161,
      "learning_rate": 0.0005581971186212468,
      "loss": 3.4771,
      "step": 10450
    },
    {
      "epoch": 2.825796726959518,
      "grad_norm": 0.33602234721183777,
      "learning_rate": 0.0005579951528207889,
      "loss": 3.4699,
      "step": 10500
    },
    {
      "epoch": 2.8392549526270456,
      "grad_norm": 0.34759023785591125,
      "learning_rate": 0.0005577931870203312,
      "loss": 3.4713,
      "step": 10550
    },
    {
      "epoch": 2.8527131782945734,
      "grad_norm": 0.35125479102134705,
      "learning_rate": 0.0005575912212198733,
      "loss": 3.4733,
      "step": 10600
    },
    {
      "epoch": 2.8661714039621016,
      "grad_norm": 0.347562700510025,
      "learning_rate": 0.0005573892554194156,
      "loss": 3.4875,
      "step": 10650
    },
    {
      "epoch": 2.8796296296296298,
      "grad_norm": 0.35739463567733765,
      "learning_rate": 0.0005571872896189578,
      "loss": 3.4606,
      "step": 10700
    },
    {
      "epoch": 2.8930878552971575,
      "grad_norm": 0.3398447632789612,
      "learning_rate": 0.0005569853238185,
      "loss": 3.4724,
      "step": 10750
    },
    {
      "epoch": 2.9065460809646857,
      "grad_norm": 0.33381515741348267,
      "learning_rate": 0.0005567833580180422,
      "loss": 3.4632,
      "step": 10800
    },
    {
      "epoch": 2.9200043066322134,
      "grad_norm": 0.36585623025894165,
      "learning_rate": 0.0005565813922175844,
      "loss": 3.4718,
      "step": 10850
    },
    {
      "epoch": 2.9334625322997416,
      "grad_norm": 0.3554155230522156,
      "learning_rate": 0.0005563794264171266,
      "loss": 3.4629,
      "step": 10900
    },
    {
      "epoch": 2.94692075796727,
      "grad_norm": 0.3801786005496979,
      "learning_rate": 0.0005561774606166689,
      "loss": 3.4655,
      "step": 10950
    },
    {
      "epoch": 2.9603789836347976,
      "grad_norm": 0.3750773072242737,
      "learning_rate": 0.000555975494816211,
      "loss": 3.464,
      "step": 11000
    },
    {
      "epoch": 2.9603789836347976,
      "eval_accuracy": 0.37141242650583456,
      "eval_loss": 3.462761878967285,
      "eval_runtime": 83.2607,
      "eval_samples_per_second": 216.333,
      "eval_steps_per_second": 13.524,
      "step": 11000
    },
    {
      "epoch": 2.9738372093023253,
      "grad_norm": 0.3510102927684784,
      "learning_rate": 0.0005557735290157534,
      "loss": 3.4702,
      "step": 11050
    },
    {
      "epoch": 2.9872954349698535,
      "grad_norm": 0.31275537610054016,
      "learning_rate": 0.0005555715632152955,
      "loss": 3.4635,
      "step": 11100
    },
    {
      "epoch": 3.000538329026701,
      "grad_norm": 0.34717661142349243,
      "learning_rate": 0.0005553695974148378,
      "loss": 3.4519,
      "step": 11150
    },
    {
      "epoch": 3.013996554694229,
      "grad_norm": 0.33590519428253174,
      "learning_rate": 0.0005551676316143799,
      "loss": 3.3721,
      "step": 11200
    },
    {
      "epoch": 3.027454780361757,
      "grad_norm": 0.35407522320747375,
      "learning_rate": 0.0005549656658139222,
      "loss": 3.3917,
      "step": 11250
    },
    {
      "epoch": 3.0409130060292853,
      "grad_norm": 0.35328567028045654,
      "learning_rate": 0.0005547637000134644,
      "loss": 3.3702,
      "step": 11300
    },
    {
      "epoch": 3.054371231696813,
      "grad_norm": 0.3492799699306488,
      "learning_rate": 0.0005545617342130065,
      "loss": 3.3861,
      "step": 11350
    },
    {
      "epoch": 3.067829457364341,
      "grad_norm": 0.3478204905986786,
      "learning_rate": 0.0005543597684125488,
      "loss": 3.3798,
      "step": 11400
    },
    {
      "epoch": 3.081287683031869,
      "grad_norm": 0.3717453181743622,
      "learning_rate": 0.0005541578026120909,
      "loss": 3.3916,
      "step": 11450
    },
    {
      "epoch": 3.094745908699397,
      "grad_norm": 0.33747634291648865,
      "learning_rate": 0.0005539558368116332,
      "loss": 3.3797,
      "step": 11500
    },
    {
      "epoch": 3.108204134366925,
      "grad_norm": 0.35223788022994995,
      "learning_rate": 0.0005537538710111754,
      "loss": 3.3857,
      "step": 11550
    },
    {
      "epoch": 3.121662360034453,
      "grad_norm": 0.34729883074760437,
      "learning_rate": 0.0005535519052107176,
      "loss": 3.3857,
      "step": 11600
    },
    {
      "epoch": 3.135120585701981,
      "grad_norm": 0.34449636936187744,
      "learning_rate": 0.0005533499394102598,
      "loss": 3.3853,
      "step": 11650
    },
    {
      "epoch": 3.148578811369509,
      "grad_norm": 0.4149003028869629,
      "learning_rate": 0.000553147973609802,
      "loss": 3.3923,
      "step": 11700
    },
    {
      "epoch": 3.162037037037037,
      "grad_norm": 0.33877599239349365,
      "learning_rate": 0.0005529460078093442,
      "loss": 3.3921,
      "step": 11750
    },
    {
      "epoch": 3.175495262704565,
      "grad_norm": 0.3686054050922394,
      "learning_rate": 0.0005527440420088865,
      "loss": 3.3865,
      "step": 11800
    },
    {
      "epoch": 3.188953488372093,
      "grad_norm": 0.3487144708633423,
      "learning_rate": 0.0005525420762084286,
      "loss": 3.3863,
      "step": 11850
    },
    {
      "epoch": 3.202411714039621,
      "grad_norm": 0.35557788610458374,
      "learning_rate": 0.0005523401104079709,
      "loss": 3.3975,
      "step": 11900
    },
    {
      "epoch": 3.215869939707149,
      "grad_norm": 0.32155174016952515,
      "learning_rate": 0.0005521381446075131,
      "loss": 3.3909,
      "step": 11950
    },
    {
      "epoch": 3.229328165374677,
      "grad_norm": 0.34835678339004517,
      "learning_rate": 0.0005519361788070554,
      "loss": 3.3924,
      "step": 12000
    },
    {
      "epoch": 3.229328165374677,
      "eval_accuracy": 0.3729031700110342,
      "eval_loss": 3.4515469074249268,
      "eval_runtime": 83.4051,
      "eval_samples_per_second": 215.958,
      "eval_steps_per_second": 13.5,
      "step": 12000
    },
    {
      "epoch": 3.242786391042205,
      "grad_norm": 0.33047643303871155,
      "learning_rate": 0.0005517342130065975,
      "loss": 3.3891,
      "step": 12050
    },
    {
      "epoch": 3.2562446167097328,
      "grad_norm": 0.3518250584602356,
      "learning_rate": 0.0005515322472061398,
      "loss": 3.3996,
      "step": 12100
    },
    {
      "epoch": 3.269702842377261,
      "grad_norm": 0.3466777205467224,
      "learning_rate": 0.0005513302814056819,
      "loss": 3.387,
      "step": 12150
    },
    {
      "epoch": 3.283161068044789,
      "grad_norm": 0.34431418776512146,
      "learning_rate": 0.0005511283156052242,
      "loss": 3.3861,
      "step": 12200
    },
    {
      "epoch": 3.296619293712317,
      "grad_norm": 0.3356941044330597,
      "learning_rate": 0.0005509263498047664,
      "loss": 3.3872,
      "step": 12250
    },
    {
      "epoch": 3.310077519379845,
      "grad_norm": 0.33507728576660156,
      "learning_rate": 0.0005507243840043085,
      "loss": 3.4042,
      "step": 12300
    },
    {
      "epoch": 3.323535745047373,
      "grad_norm": 0.34784412384033203,
      "learning_rate": 0.0005505224182038508,
      "loss": 3.3947,
      "step": 12350
    },
    {
      "epoch": 3.336993970714901,
      "grad_norm": 0.35030630230903625,
      "learning_rate": 0.0005503204524033929,
      "loss": 3.3957,
      "step": 12400
    },
    {
      "epoch": 3.3504521963824287,
      "grad_norm": 0.35533133149147034,
      "learning_rate": 0.0005501184866029352,
      "loss": 3.392,
      "step": 12450
    },
    {
      "epoch": 3.363910422049957,
      "grad_norm": 0.3376750648021698,
      "learning_rate": 0.0005499165208024774,
      "loss": 3.3933,
      "step": 12500
    },
    {
      "epoch": 3.3773686477174847,
      "grad_norm": 0.31374868750572205,
      "learning_rate": 0.0005497145550020196,
      "loss": 3.3934,
      "step": 12550
    },
    {
      "epoch": 3.390826873385013,
      "grad_norm": 0.3525221347808838,
      "learning_rate": 0.0005495125892015618,
      "loss": 3.4027,
      "step": 12600
    },
    {
      "epoch": 3.404285099052541,
      "grad_norm": 0.33895227313041687,
      "learning_rate": 0.000549310623401104,
      "loss": 3.3979,
      "step": 12650
    },
    {
      "epoch": 3.417743324720069,
      "grad_norm": 0.36806583404541016,
      "learning_rate": 0.0005491086576006462,
      "loss": 3.409,
      "step": 12700
    },
    {
      "epoch": 3.431201550387597,
      "grad_norm": 0.3475854694843292,
      "learning_rate": 0.0005489066918001885,
      "loss": 3.3958,
      "step": 12750
    },
    {
      "epoch": 3.4446597760551247,
      "grad_norm": 0.3557698726654053,
      "learning_rate": 0.0005487047259997306,
      "loss": 3.4124,
      "step": 12800
    },
    {
      "epoch": 3.458118001722653,
      "grad_norm": 0.34914955496788025,
      "learning_rate": 0.0005485027601992729,
      "loss": 3.3851,
      "step": 12850
    },
    {
      "epoch": 3.471576227390181,
      "grad_norm": 0.34068846702575684,
      "learning_rate": 0.0005483007943988151,
      "loss": 3.4001,
      "step": 12900
    },
    {
      "epoch": 3.485034453057709,
      "grad_norm": 0.3358888030052185,
      "learning_rate": 0.0005480988285983574,
      "loss": 3.399,
      "step": 12950
    },
    {
      "epoch": 3.4984926787252366,
      "grad_norm": 0.3400687277317047,
      "learning_rate": 0.0005478968627978995,
      "loss": 3.4073,
      "step": 13000
    },
    {
      "epoch": 3.4984926787252366,
      "eval_accuracy": 0.3747549470172744,
      "eval_loss": 3.4357733726501465,
      "eval_runtime": 83.1927,
      "eval_samples_per_second": 216.509,
      "eval_steps_per_second": 13.535,
      "step": 13000
    },
    {
      "epoch": 3.511950904392765,
      "grad_norm": 0.3225702941417694,
      "learning_rate": 0.0005476948969974418,
      "loss": 3.4025,
      "step": 13050
    },
    {
      "epoch": 3.525409130060293,
      "grad_norm": 0.36123523116111755,
      "learning_rate": 0.0005474929311969839,
      "loss": 3.4019,
      "step": 13100
    },
    {
      "epoch": 3.5388673557278207,
      "grad_norm": 0.33827775716781616,
      "learning_rate": 0.0005472909653965261,
      "loss": 3.4012,
      "step": 13150
    },
    {
      "epoch": 3.552325581395349,
      "grad_norm": 0.35949254035949707,
      "learning_rate": 0.0005470889995960684,
      "loss": 3.3957,
      "step": 13200
    },
    {
      "epoch": 3.5657838070628767,
      "grad_norm": 0.3394739031791687,
      "learning_rate": 0.0005468870337956105,
      "loss": 3.3955,
      "step": 13250
    },
    {
      "epoch": 3.579242032730405,
      "grad_norm": 0.34938886761665344,
      "learning_rate": 0.0005466850679951528,
      "loss": 3.4026,
      "step": 13300
    },
    {
      "epoch": 3.592700258397933,
      "grad_norm": 0.35778287053108215,
      "learning_rate": 0.0005464831021946949,
      "loss": 3.3849,
      "step": 13350
    },
    {
      "epoch": 3.606158484065461,
      "grad_norm": 0.32485082745552063,
      "learning_rate": 0.0005462811363942372,
      "loss": 3.3937,
      "step": 13400
    },
    {
      "epoch": 3.6196167097329885,
      "grad_norm": 0.3334294557571411,
      "learning_rate": 0.0005460791705937794,
      "loss": 3.414,
      "step": 13450
    },
    {
      "epoch": 3.6330749354005167,
      "grad_norm": 0.31381288170814514,
      "learning_rate": 0.0005458772047933216,
      "loss": 3.3917,
      "step": 13500
    },
    {
      "epoch": 3.646533161068045,
      "grad_norm": 0.3137832283973694,
      "learning_rate": 0.0005456752389928638,
      "loss": 3.3962,
      "step": 13550
    },
    {
      "epoch": 3.6599913867355727,
      "grad_norm": 0.3500538468360901,
      "learning_rate": 0.000545473273192406,
      "loss": 3.3848,
      "step": 13600
    },
    {
      "epoch": 3.673449612403101,
      "grad_norm": 0.3487536311149597,
      "learning_rate": 0.0005452713073919482,
      "loss": 3.391,
      "step": 13650
    },
    {
      "epoch": 3.6869078380706286,
      "grad_norm": 0.3269287943840027,
      "learning_rate": 0.0005450693415914905,
      "loss": 3.3899,
      "step": 13700
    },
    {
      "epoch": 3.700366063738157,
      "grad_norm": 0.33050355315208435,
      "learning_rate": 0.0005448673757910327,
      "loss": 3.4014,
      "step": 13750
    },
    {
      "epoch": 3.713824289405685,
      "grad_norm": 0.34464776515960693,
      "learning_rate": 0.0005446654099905749,
      "loss": 3.389,
      "step": 13800
    },
    {
      "epoch": 3.7272825150732127,
      "grad_norm": 0.32913777232170105,
      "learning_rate": 0.0005444634441901171,
      "loss": 3.404,
      "step": 13850
    },
    {
      "epoch": 3.7407407407407405,
      "grad_norm": 0.3600791394710541,
      "learning_rate": 0.0005442614783896594,
      "loss": 3.3951,
      "step": 13900
    },
    {
      "epoch": 3.7541989664082687,
      "grad_norm": 0.37241896986961365,
      "learning_rate": 0.0005440595125892015,
      "loss": 3.3954,
      "step": 13950
    },
    {
      "epoch": 3.767657192075797,
      "grad_norm": 0.337048202753067,
      "learning_rate": 0.0005438575467887438,
      "loss": 3.3945,
      "step": 14000
    },
    {
      "epoch": 3.767657192075797,
      "eval_accuracy": 0.376091303340717,
      "eval_loss": 3.420311689376831,
      "eval_runtime": 83.3301,
      "eval_samples_per_second": 216.152,
      "eval_steps_per_second": 13.513,
      "step": 14000
    },
    {
      "epoch": 3.7811154177433246,
      "grad_norm": 0.323229044675827,
      "learning_rate": 0.0005436555809882859,
      "loss": 3.3956,
      "step": 14050
    },
    {
      "epoch": 3.794573643410853,
      "grad_norm": 0.3261372148990631,
      "learning_rate": 0.0005434536151878281,
      "loss": 3.3882,
      "step": 14100
    },
    {
      "epoch": 3.8080318690783805,
      "grad_norm": 0.33051732182502747,
      "learning_rate": 0.0005432516493873704,
      "loss": 3.3895,
      "step": 14150
    },
    {
      "epoch": 3.8214900947459087,
      "grad_norm": 0.3285205066204071,
      "learning_rate": 0.0005430496835869125,
      "loss": 3.3882,
      "step": 14200
    },
    {
      "epoch": 3.834948320413437,
      "grad_norm": 0.33567163348197937,
      "learning_rate": 0.0005428477177864548,
      "loss": 3.3995,
      "step": 14250
    },
    {
      "epoch": 3.8484065460809647,
      "grad_norm": 0.35562512278556824,
      "learning_rate": 0.0005426457519859969,
      "loss": 3.4081,
      "step": 14300
    },
    {
      "epoch": 3.8618647717484924,
      "grad_norm": 0.33773207664489746,
      "learning_rate": 0.0005424437861855392,
      "loss": 3.3771,
      "step": 14350
    },
    {
      "epoch": 3.8753229974160206,
      "grad_norm": 0.3264458477497101,
      "learning_rate": 0.0005422418203850814,
      "loss": 3.3909,
      "step": 14400
    },
    {
      "epoch": 3.888781223083549,
      "grad_norm": 0.3677447438240051,
      "learning_rate": 0.0005420398545846236,
      "loss": 3.3982,
      "step": 14450
    },
    {
      "epoch": 3.9022394487510765,
      "grad_norm": 0.3357374370098114,
      "learning_rate": 0.0005418378887841658,
      "loss": 3.39,
      "step": 14500
    },
    {
      "epoch": 3.9156976744186047,
      "grad_norm": 0.3447486162185669,
      "learning_rate": 0.000541635922983708,
      "loss": 3.3946,
      "step": 14550
    },
    {
      "epoch": 3.9291559000861325,
      "grad_norm": 0.34921061992645264,
      "learning_rate": 0.0005414339571832503,
      "loss": 3.3972,
      "step": 14600
    },
    {
      "epoch": 3.9426141257536607,
      "grad_norm": 0.3274434208869934,
      "learning_rate": 0.0005412319913827925,
      "loss": 3.3876,
      "step": 14650
    },
    {
      "epoch": 3.956072351421189,
      "grad_norm": 0.39226579666137695,
      "learning_rate": 0.0005410300255823347,
      "loss": 3.3963,
      "step": 14700
    },
    {
      "epoch": 3.9695305770887166,
      "grad_norm": 0.3527931571006775,
      "learning_rate": 0.0005408280597818769,
      "loss": 3.3882,
      "step": 14750
    },
    {
      "epoch": 3.9829888027562443,
      "grad_norm": 0.31162092089653015,
      "learning_rate": 0.0005406260939814191,
      "loss": 3.3977,
      "step": 14800
    },
    {
      "epoch": 3.9964470284237725,
      "grad_norm": 0.3350144028663635,
      "learning_rate": 0.0005404241281809614,
      "loss": 3.3872,
      "step": 14850
    },
    {
      "epoch": 4.00968992248062,
      "grad_norm": 0.3405536115169525,
      "learning_rate": 0.0005402221623805035,
      "loss": 3.3227,
      "step": 14900
    },
    {
      "epoch": 4.023148148148148,
      "grad_norm": 0.335857093334198,
      "learning_rate": 0.0005400201965800457,
      "loss": 3.2953,
      "step": 14950
    },
    {
      "epoch": 4.036606373815676,
      "grad_norm": 0.343741238117218,
      "learning_rate": 0.000539818230779588,
      "loss": 3.294,
      "step": 15000
    },
    {
      "epoch": 4.036606373815676,
      "eval_accuracy": 0.37734302376367485,
      "eval_loss": 3.4138002395629883,
      "eval_runtime": 83.2945,
      "eval_samples_per_second": 216.245,
      "eval_steps_per_second": 13.518,
      "step": 15000
    },
    {
      "epoch": 4.050064599483204,
      "grad_norm": 0.37379246950149536,
      "learning_rate": 0.0005396162649791301,
      "loss": 3.2958,
      "step": 15050
    },
    {
      "epoch": 4.0635228251507325,
      "grad_norm": 0.3613829016685486,
      "learning_rate": 0.0005394142991786724,
      "loss": 3.3019,
      "step": 15100
    },
    {
      "epoch": 4.07698105081826,
      "grad_norm": 0.3490220606327057,
      "learning_rate": 0.0005392123333782145,
      "loss": 3.2915,
      "step": 15150
    },
    {
      "epoch": 4.090439276485788,
      "grad_norm": 0.333156019449234,
      "learning_rate": 0.0005390103675777568,
      "loss": 3.3009,
      "step": 15200
    },
    {
      "epoch": 4.103897502153316,
      "grad_norm": 0.3495680093765259,
      "learning_rate": 0.000538808401777299,
      "loss": 3.3082,
      "step": 15250
    },
    {
      "epoch": 4.117355727820844,
      "grad_norm": 0.30972912907600403,
      "learning_rate": 0.0005386064359768412,
      "loss": 3.2972,
      "step": 15300
    },
    {
      "epoch": 4.1308139534883725,
      "grad_norm": 0.3500850200653076,
      "learning_rate": 0.0005384044701763834,
      "loss": 3.3145,
      "step": 15350
    },
    {
      "epoch": 4.1442721791559,
      "grad_norm": 0.39197248220443726,
      "learning_rate": 0.0005382025043759256,
      "loss": 3.326,
      "step": 15400
    },
    {
      "epoch": 4.157730404823428,
      "grad_norm": 0.370394766330719,
      "learning_rate": 0.0005380005385754678,
      "loss": 3.3082,
      "step": 15450
    },
    {
      "epoch": 4.171188630490956,
      "grad_norm": 0.32483676075935364,
      "learning_rate": 0.0005377985727750101,
      "loss": 3.3111,
      "step": 15500
    },
    {
      "epoch": 4.184646856158484,
      "grad_norm": 0.3517882823944092,
      "learning_rate": 0.0005375966069745523,
      "loss": 3.3164,
      "step": 15550
    },
    {
      "epoch": 4.198105081826012,
      "grad_norm": 0.3727375268936157,
      "learning_rate": 0.0005373946411740945,
      "loss": 3.3181,
      "step": 15600
    },
    {
      "epoch": 4.21156330749354,
      "grad_norm": 0.3322087228298187,
      "learning_rate": 0.0005371926753736367,
      "loss": 3.3179,
      "step": 15650
    },
    {
      "epoch": 4.225021533161068,
      "grad_norm": 0.35543593764305115,
      "learning_rate": 0.000536990709573179,
      "loss": 3.3253,
      "step": 15700
    },
    {
      "epoch": 4.238479758828596,
      "grad_norm": 0.35299551486968994,
      "learning_rate": 0.0005367887437727211,
      "loss": 3.3201,
      "step": 15750
    },
    {
      "epoch": 4.251937984496124,
      "grad_norm": 0.3296690583229065,
      "learning_rate": 0.0005365867779722634,
      "loss": 3.324,
      "step": 15800
    },
    {
      "epoch": 4.265396210163652,
      "grad_norm": 0.31453442573547363,
      "learning_rate": 0.0005363848121718055,
      "loss": 3.3297,
      "step": 15850
    },
    {
      "epoch": 4.27885443583118,
      "grad_norm": 0.36801278591156006,
      "learning_rate": 0.0005361828463713477,
      "loss": 3.3303,
      "step": 15900
    },
    {
      "epoch": 4.292312661498708,
      "grad_norm": 0.3610571324825287,
      "learning_rate": 0.00053598088057089,
      "loss": 3.3194,
      "step": 15950
    },
    {
      "epoch": 4.305770887166236,
      "grad_norm": 0.3473931849002838,
      "learning_rate": 0.0005357789147704321,
      "loss": 3.33,
      "step": 16000
    },
    {
      "epoch": 4.305770887166236,
      "eval_accuracy": 0.378296834508675,
      "eval_loss": 3.406858205795288,
      "eval_runtime": 83.3803,
      "eval_samples_per_second": 216.022,
      "eval_steps_per_second": 13.504,
      "step": 16000
    },
    {
      "epoch": 4.319229112833764,
      "grad_norm": 0.32994207739830017,
      "learning_rate": 0.0005355769489699744,
      "loss": 3.3245,
      "step": 16050
    },
    {
      "epoch": 4.332687338501292,
      "grad_norm": 0.3246585428714752,
      "learning_rate": 0.0005353749831695165,
      "loss": 3.34,
      "step": 16100
    },
    {
      "epoch": 4.34614556416882,
      "grad_norm": 0.3300127387046814,
      "learning_rate": 0.0005351730173690588,
      "loss": 3.3242,
      "step": 16150
    },
    {
      "epoch": 4.359603789836348,
      "grad_norm": 0.3596031367778778,
      "learning_rate": 0.000534971051568601,
      "loss": 3.3374,
      "step": 16200
    },
    {
      "epoch": 4.373062015503876,
      "grad_norm": 0.32846489548683167,
      "learning_rate": 0.0005347690857681432,
      "loss": 3.324,
      "step": 16250
    },
    {
      "epoch": 4.386520241171404,
      "grad_norm": 0.34727606177330017,
      "learning_rate": 0.0005345671199676854,
      "loss": 3.3291,
      "step": 16300
    },
    {
      "epoch": 4.399978466838932,
      "grad_norm": 0.3296259641647339,
      "learning_rate": 0.0005343651541672276,
      "loss": 3.3273,
      "step": 16350
    },
    {
      "epoch": 4.41343669250646,
      "grad_norm": 0.3439517915248871,
      "learning_rate": 0.0005341631883667699,
      "loss": 3.3356,
      "step": 16400
    },
    {
      "epoch": 4.426894918173988,
      "grad_norm": 0.36822691559791565,
      "learning_rate": 0.0005339612225663121,
      "loss": 3.3355,
      "step": 16450
    },
    {
      "epoch": 4.440353143841516,
      "grad_norm": 0.33213186264038086,
      "learning_rate": 0.0005337592567658543,
      "loss": 3.3285,
      "step": 16500
    },
    {
      "epoch": 4.453811369509044,
      "grad_norm": 0.35229599475860596,
      "learning_rate": 0.0005335572909653965,
      "loss": 3.3319,
      "step": 16550
    },
    {
      "epoch": 4.467269595176572,
      "grad_norm": 0.3325550854206085,
      "learning_rate": 0.0005333553251649387,
      "loss": 3.3338,
      "step": 16600
    },
    {
      "epoch": 4.4807278208441,
      "grad_norm": 0.3523412346839905,
      "learning_rate": 0.000533153359364481,
      "loss": 3.3333,
      "step": 16650
    },
    {
      "epoch": 4.4941860465116275,
      "grad_norm": 0.33977991342544556,
      "learning_rate": 0.0005329513935640231,
      "loss": 3.332,
      "step": 16700
    },
    {
      "epoch": 4.507644272179156,
      "grad_norm": 0.3187093734741211,
      "learning_rate": 0.0005327494277635654,
      "loss": 3.3392,
      "step": 16750
    },
    {
      "epoch": 4.521102497846684,
      "grad_norm": 0.3350430130958557,
      "learning_rate": 0.0005325474619631075,
      "loss": 3.3313,
      "step": 16800
    },
    {
      "epoch": 4.534560723514212,
      "grad_norm": 0.33022499084472656,
      "learning_rate": 0.0005323454961626497,
      "loss": 3.3347,
      "step": 16850
    },
    {
      "epoch": 4.54801894918174,
      "grad_norm": 0.35957691073417664,
      "learning_rate": 0.000532143530362192,
      "loss": 3.3388,
      "step": 16900
    },
    {
      "epoch": 4.5614771748492675,
      "grad_norm": 0.3328224718570709,
      "learning_rate": 0.0005319415645617341,
      "loss": 3.3298,
      "step": 16950
    },
    {
      "epoch": 4.574935400516796,
      "grad_norm": 0.3368069529533386,
      "learning_rate": 0.0005317395987612764,
      "loss": 3.3293,
      "step": 17000
    },
    {
      "epoch": 4.574935400516796,
      "eval_accuracy": 0.3796052685902375,
      "eval_loss": 3.392096996307373,
      "eval_runtime": 83.3396,
      "eval_samples_per_second": 216.128,
      "eval_steps_per_second": 13.511,
      "step": 17000
    },
    {
      "epoch": 4.588393626184324,
      "grad_norm": 0.33241188526153564,
      "learning_rate": 0.0005315376329608185,
      "loss": 3.3326,
      "step": 17050
    },
    {
      "epoch": 4.601851851851852,
      "grad_norm": 0.35584133863449097,
      "learning_rate": 0.0005313356671603608,
      "loss": 3.3369,
      "step": 17100
    },
    {
      "epoch": 4.61531007751938,
      "grad_norm": 0.3345257639884949,
      "learning_rate": 0.000531133701359903,
      "loss": 3.3363,
      "step": 17150
    },
    {
      "epoch": 4.628768303186908,
      "grad_norm": 0.31600186228752136,
      "learning_rate": 0.0005309317355594452,
      "loss": 3.3374,
      "step": 17200
    },
    {
      "epoch": 4.642226528854436,
      "grad_norm": 0.3724959194660187,
      "learning_rate": 0.0005307297697589874,
      "loss": 3.3262,
      "step": 17250
    },
    {
      "epoch": 4.655684754521964,
      "grad_norm": 0.3523881435394287,
      "learning_rate": 0.0005305278039585297,
      "loss": 3.3349,
      "step": 17300
    },
    {
      "epoch": 4.669142980189492,
      "grad_norm": 0.3505227565765381,
      "learning_rate": 0.0005303258381580719,
      "loss": 3.3379,
      "step": 17350
    },
    {
      "epoch": 4.682601205857019,
      "grad_norm": 0.33256712555885315,
      "learning_rate": 0.0005301238723576141,
      "loss": 3.3394,
      "step": 17400
    },
    {
      "epoch": 4.696059431524548,
      "grad_norm": 0.32822030782699585,
      "learning_rate": 0.0005299219065571563,
      "loss": 3.3375,
      "step": 17450
    },
    {
      "epoch": 4.709517657192076,
      "grad_norm": 0.32785138487815857,
      "learning_rate": 0.0005297199407566985,
      "loss": 3.3305,
      "step": 17500
    },
    {
      "epoch": 4.722975882859604,
      "grad_norm": 0.32746267318725586,
      "learning_rate": 0.0005295179749562407,
      "loss": 3.3403,
      "step": 17550
    },
    {
      "epoch": 4.736434108527131,
      "grad_norm": 0.34052562713623047,
      "learning_rate": 0.000529316009155783,
      "loss": 3.3407,
      "step": 17600
    },
    {
      "epoch": 4.7498923341946595,
      "grad_norm": 0.3175918459892273,
      "learning_rate": 0.0005291140433553251,
      "loss": 3.3359,
      "step": 17650
    },
    {
      "epoch": 4.763350559862188,
      "grad_norm": 0.3336562216281891,
      "learning_rate": 0.0005289120775548673,
      "loss": 3.3457,
      "step": 17700
    },
    {
      "epoch": 4.776808785529716,
      "grad_norm": 0.3395072817802429,
      "learning_rate": 0.0005287101117544095,
      "loss": 3.3386,
      "step": 17750
    },
    {
      "epoch": 4.790267011197244,
      "grad_norm": 0.3338567018508911,
      "learning_rate": 0.0005285081459539517,
      "loss": 3.3311,
      "step": 17800
    },
    {
      "epoch": 4.803725236864771,
      "grad_norm": 0.346660852432251,
      "learning_rate": 0.000528306180153494,
      "loss": 3.3325,
      "step": 17850
    },
    {
      "epoch": 4.8171834625323,
      "grad_norm": 0.3293061852455139,
      "learning_rate": 0.0005281042143530361,
      "loss": 3.3392,
      "step": 17900
    },
    {
      "epoch": 4.830641688199828,
      "grad_norm": 0.3239416778087616,
      "learning_rate": 0.0005279022485525784,
      "loss": 3.3344,
      "step": 17950
    },
    {
      "epoch": 4.844099913867356,
      "grad_norm": 0.3364400267601013,
      "learning_rate": 0.0005277002827521205,
      "loss": 3.3378,
      "step": 18000
    },
    {
      "epoch": 4.844099913867356,
      "eval_accuracy": 0.3809082703290218,
      "eval_loss": 3.3793487548828125,
      "eval_runtime": 83.4144,
      "eval_samples_per_second": 215.934,
      "eval_steps_per_second": 13.499,
      "step": 18000
    },
    {
      "epoch": 4.857558139534884,
      "grad_norm": 0.33611395955085754,
      "learning_rate": 0.0005274983169516628,
      "loss": 3.3295,
      "step": 18050
    },
    {
      "epoch": 4.871016365202411,
      "grad_norm": 0.34581753611564636,
      "learning_rate": 0.000527296351151205,
      "loss": 3.3516,
      "step": 18100
    },
    {
      "epoch": 4.88447459086994,
      "grad_norm": 0.33185499906539917,
      "learning_rate": 0.0005270943853507472,
      "loss": 3.3341,
      "step": 18150
    },
    {
      "epoch": 4.897932816537468,
      "grad_norm": 0.32314810156822205,
      "learning_rate": 0.0005268924195502894,
      "loss": 3.3294,
      "step": 18200
    },
    {
      "epoch": 4.911391042204996,
      "grad_norm": 0.3456299901008606,
      "learning_rate": 0.0005266904537498317,
      "loss": 3.3339,
      "step": 18250
    },
    {
      "epoch": 4.924849267872523,
      "grad_norm": 0.3304615616798401,
      "learning_rate": 0.0005264884879493739,
      "loss": 3.3424,
      "step": 18300
    },
    {
      "epoch": 4.9383074935400515,
      "grad_norm": 0.3197613060474396,
      "learning_rate": 0.0005262865221489161,
      "loss": 3.3403,
      "step": 18350
    },
    {
      "epoch": 4.95176571920758,
      "grad_norm": 0.32757633924484253,
      "learning_rate": 0.0005260845563484583,
      "loss": 3.3357,
      "step": 18400
    },
    {
      "epoch": 4.965223944875108,
      "grad_norm": 0.35092049837112427,
      "learning_rate": 0.0005258825905480005,
      "loss": 3.3331,
      "step": 18450
    },
    {
      "epoch": 4.978682170542635,
      "grad_norm": 0.3141130208969116,
      "learning_rate": 0.0005256806247475427,
      "loss": 3.3371,
      "step": 18500
    },
    {
      "epoch": 4.992140396210163,
      "grad_norm": 0.30893778800964355,
      "learning_rate": 0.000525478658947085,
      "loss": 3.3414,
      "step": 18550
    },
    {
      "epoch": 5.0053832902670115,
      "grad_norm": 0.33693405985832214,
      "learning_rate": 0.0005252766931466271,
      "loss": 3.2945,
      "step": 18600
    },
    {
      "epoch": 5.018841515934539,
      "grad_norm": 0.3547412157058716,
      "learning_rate": 0.0005250747273461693,
      "loss": 3.2212,
      "step": 18650
    },
    {
      "epoch": 5.032299741602067,
      "grad_norm": 0.3842681050300598,
      "learning_rate": 0.0005248727615457115,
      "loss": 3.246,
      "step": 18700
    },
    {
      "epoch": 5.045757967269595,
      "grad_norm": 0.3256308138370514,
      "learning_rate": 0.0005246707957452537,
      "loss": 3.2392,
      "step": 18750
    },
    {
      "epoch": 5.059216192937123,
      "grad_norm": 0.30408331751823425,
      "learning_rate": 0.000524468829944796,
      "loss": 3.2556,
      "step": 18800
    },
    {
      "epoch": 5.0726744186046515,
      "grad_norm": 0.3305226266384125,
      "learning_rate": 0.0005242668641443381,
      "loss": 3.2611,
      "step": 18850
    },
    {
      "epoch": 5.086132644272179,
      "grad_norm": 0.32760271430015564,
      "learning_rate": 0.0005240648983438804,
      "loss": 3.251,
      "step": 18900
    },
    {
      "epoch": 5.099590869939707,
      "grad_norm": 0.342307448387146,
      "learning_rate": 0.0005238629325434225,
      "loss": 3.2572,
      "step": 18950
    },
    {
      "epoch": 5.113049095607235,
      "grad_norm": 0.3348807394504547,
      "learning_rate": 0.0005236609667429648,
      "loss": 3.2494,
      "step": 19000
    },
    {
      "epoch": 5.113049095607235,
      "eval_accuracy": 0.3813553521396694,
      "eval_loss": 3.3809144496917725,
      "eval_runtime": 83.3305,
      "eval_samples_per_second": 216.151,
      "eval_steps_per_second": 13.512,
      "step": 19000
    },
    {
      "epoch": 5.126507321274763,
      "grad_norm": 0.3393591046333313,
      "learning_rate": 0.000523459000942507,
      "loss": 3.2522,
      "step": 19050
    },
    {
      "epoch": 5.139965546942291,
      "grad_norm": 0.3312690556049347,
      "learning_rate": 0.0005232570351420493,
      "loss": 3.2522,
      "step": 19100
    },
    {
      "epoch": 5.153423772609819,
      "grad_norm": 0.3400869369506836,
      "learning_rate": 0.0005230550693415915,
      "loss": 3.2667,
      "step": 19150
    },
    {
      "epoch": 5.166881998277347,
      "grad_norm": 0.34973442554473877,
      "learning_rate": 0.0005228531035411337,
      "loss": 3.2705,
      "step": 19200
    },
    {
      "epoch": 5.180340223944875,
      "grad_norm": 0.3322692811489105,
      "learning_rate": 0.0005226511377406759,
      "loss": 3.252,
      "step": 19250
    },
    {
      "epoch": 5.1937984496124034,
      "grad_norm": 0.3838331997394562,
      "learning_rate": 0.0005224491719402181,
      "loss": 3.2625,
      "step": 19300
    },
    {
      "epoch": 5.207256675279931,
      "grad_norm": 0.304482102394104,
      "learning_rate": 0.0005222472061397603,
      "loss": 3.2596,
      "step": 19350
    },
    {
      "epoch": 5.220714900947459,
      "grad_norm": 0.3591323494911194,
      "learning_rate": 0.0005220452403393026,
      "loss": 3.2618,
      "step": 19400
    },
    {
      "epoch": 5.234173126614987,
      "grad_norm": 0.31714892387390137,
      "learning_rate": 0.0005218432745388447,
      "loss": 3.2704,
      "step": 19450
    },
    {
      "epoch": 5.247631352282515,
      "grad_norm": 0.33631953597068787,
      "learning_rate": 0.000521641308738387,
      "loss": 3.2827,
      "step": 19500
    },
    {
      "epoch": 5.2610895779500435,
      "grad_norm": 0.376729816198349,
      "learning_rate": 0.0005214393429379291,
      "loss": 3.2736,
      "step": 19550
    },
    {
      "epoch": 5.274547803617571,
      "grad_norm": 0.3226640820503235,
      "learning_rate": 0.0005212373771374713,
      "loss": 3.2638,
      "step": 19600
    },
    {
      "epoch": 5.288006029285099,
      "grad_norm": 0.3198661506175995,
      "learning_rate": 0.0005210354113370136,
      "loss": 3.2762,
      "step": 19650
    },
    {
      "epoch": 5.301464254952627,
      "grad_norm": 0.33013176918029785,
      "learning_rate": 0.0005208334455365557,
      "loss": 3.2894,
      "step": 19700
    },
    {
      "epoch": 5.314922480620155,
      "grad_norm": 0.3522643744945526,
      "learning_rate": 0.000520631479736098,
      "loss": 3.2791,
      "step": 19750
    },
    {
      "epoch": 5.328380706287683,
      "grad_norm": 0.30938246846199036,
      "learning_rate": 0.0005204295139356401,
      "loss": 3.2719,
      "step": 19800
    },
    {
      "epoch": 5.341838931955211,
      "grad_norm": 0.3250071108341217,
      "learning_rate": 0.0005202275481351824,
      "loss": 3.2873,
      "step": 19850
    },
    {
      "epoch": 5.355297157622739,
      "grad_norm": 0.3527717888355255,
      "learning_rate": 0.0005200255823347246,
      "loss": 3.2943,
      "step": 19900
    },
    {
      "epoch": 5.368755383290267,
      "grad_norm": 0.356894314289093,
      "learning_rate": 0.0005198236165342669,
      "loss": 3.2854,
      "step": 19950
    },
    {
      "epoch": 5.3822136089577945,
      "grad_norm": 0.3325164020061493,
      "learning_rate": 0.000519621650733809,
      "loss": 3.2682,
      "step": 20000
    },
    {
      "epoch": 5.3822136089577945,
      "eval_accuracy": 0.38204449914451466,
      "eval_loss": 3.374255895614624,
      "eval_runtime": 83.233,
      "eval_samples_per_second": 216.404,
      "eval_steps_per_second": 13.528,
      "step": 20000
    },
    {
      "epoch": 5.395671834625323,
      "grad_norm": 0.34308794140815735,
      "learning_rate": 0.0005194196849333513,
      "loss": 3.2809,
      "step": 20050
    },
    {
      "epoch": 5.409130060292851,
      "grad_norm": 0.33125001192092896,
      "learning_rate": 0.0005192177191328935,
      "loss": 3.2757,
      "step": 20100
    },
    {
      "epoch": 5.422588285960379,
      "grad_norm": 0.33310365676879883,
      "learning_rate": 0.0005190157533324357,
      "loss": 3.2892,
      "step": 20150
    },
    {
      "epoch": 5.436046511627907,
      "grad_norm": 0.3338058292865753,
      "learning_rate": 0.0005188137875319779,
      "loss": 3.2688,
      "step": 20200
    },
    {
      "epoch": 5.449504737295435,
      "grad_norm": 0.31025946140289307,
      "learning_rate": 0.00051861182173152,
      "loss": 3.2795,
      "step": 20250
    },
    {
      "epoch": 5.462962962962963,
      "grad_norm": 0.3378399908542633,
      "learning_rate": 0.0005184098559310623,
      "loss": 3.2739,
      "step": 20300
    },
    {
      "epoch": 5.476421188630491,
      "grad_norm": 0.3298187851905823,
      "learning_rate": 0.0005182078901306046,
      "loss": 3.2742,
      "step": 20350
    },
    {
      "epoch": 5.489879414298019,
      "grad_norm": 0.3323591947555542,
      "learning_rate": 0.0005180059243301467,
      "loss": 3.2859,
      "step": 20400
    },
    {
      "epoch": 5.503337639965547,
      "grad_norm": 0.34503909945487976,
      "learning_rate": 0.000517803958529689,
      "loss": 3.2828,
      "step": 20450
    },
    {
      "epoch": 5.516795865633075,
      "grad_norm": 0.3266506791114807,
      "learning_rate": 0.0005176019927292311,
      "loss": 3.2819,
      "step": 20500
    },
    {
      "epoch": 5.530254091300603,
      "grad_norm": 0.34373152256011963,
      "learning_rate": 0.0005174000269287733,
      "loss": 3.2834,
      "step": 20550
    },
    {
      "epoch": 5.543712316968131,
      "grad_norm": 0.3344741463661194,
      "learning_rate": 0.0005171980611283156,
      "loss": 3.2743,
      "step": 20600
    },
    {
      "epoch": 5.557170542635659,
      "grad_norm": 0.32738304138183594,
      "learning_rate": 0.0005169960953278577,
      "loss": 3.2818,
      "step": 20650
    },
    {
      "epoch": 5.5706287683031865,
      "grad_norm": 0.342244416475296,
      "learning_rate": 0.0005167941295274,
      "loss": 3.275,
      "step": 20700
    },
    {
      "epoch": 5.584086993970715,
      "grad_norm": 0.320925772190094,
      "learning_rate": 0.0005165921637269421,
      "loss": 3.2831,
      "step": 20750
    },
    {
      "epoch": 5.597545219638243,
      "grad_norm": 0.33367154002189636,
      "learning_rate": 0.0005163901979264844,
      "loss": 3.2941,
      "step": 20800
    },
    {
      "epoch": 5.611003445305771,
      "grad_norm": 0.33265218138694763,
      "learning_rate": 0.0005161882321260266,
      "loss": 3.2839,
      "step": 20850
    },
    {
      "epoch": 5.624461670973298,
      "grad_norm": 0.32520756125450134,
      "learning_rate": 0.0005159862663255689,
      "loss": 3.2789,
      "step": 20900
    },
    {
      "epoch": 5.637919896640827,
      "grad_norm": 0.33062663674354553,
      "learning_rate": 0.000515784300525111,
      "loss": 3.2907,
      "step": 20950
    },
    {
      "epoch": 5.651378122308355,
      "grad_norm": 0.32425010204315186,
      "learning_rate": 0.0005155823347246533,
      "loss": 3.3043,
      "step": 21000
    },
    {
      "epoch": 5.651378122308355,
      "eval_accuracy": 0.38322027541543296,
      "eval_loss": 3.3639509677886963,
      "eval_runtime": 83.4012,
      "eval_samples_per_second": 215.968,
      "eval_steps_per_second": 13.501,
      "step": 21000
    },
    {
      "epoch": 5.664836347975883,
      "grad_norm": 0.3426586389541626,
      "learning_rate": 0.0005153803689241955,
      "loss": 3.2875,
      "step": 21050
    },
    {
      "epoch": 5.678294573643411,
      "grad_norm": 0.3360244929790497,
      "learning_rate": 0.0005151784031237377,
      "loss": 3.2877,
      "step": 21100
    },
    {
      "epoch": 5.6917527993109385,
      "grad_norm": 0.3354634642601013,
      "learning_rate": 0.0005149764373232799,
      "loss": 3.2855,
      "step": 21150
    },
    {
      "epoch": 5.705211024978467,
      "grad_norm": 0.353573739528656,
      "learning_rate": 0.000514774471522822,
      "loss": 3.3003,
      "step": 21200
    },
    {
      "epoch": 5.718669250645995,
      "grad_norm": 0.34776267409324646,
      "learning_rate": 0.0005145725057223643,
      "loss": 3.2981,
      "step": 21250
    },
    {
      "epoch": 5.732127476313523,
      "grad_norm": 0.32677602767944336,
      "learning_rate": 0.0005143705399219066,
      "loss": 3.2921,
      "step": 21300
    },
    {
      "epoch": 5.745585701981051,
      "grad_norm": 0.3657623529434204,
      "learning_rate": 0.0005141685741214487,
      "loss": 3.2846,
      "step": 21350
    },
    {
      "epoch": 5.7590439276485785,
      "grad_norm": 0.33666494488716125,
      "learning_rate": 0.0005139666083209909,
      "loss": 3.2821,
      "step": 21400
    },
    {
      "epoch": 5.772502153316107,
      "grad_norm": 0.33352965116500854,
      "learning_rate": 0.0005137646425205331,
      "loss": 3.2904,
      "step": 21450
    },
    {
      "epoch": 5.785960378983635,
      "grad_norm": 0.3327900469303131,
      "learning_rate": 0.0005135626767200753,
      "loss": 3.2937,
      "step": 21500
    },
    {
      "epoch": 5.799418604651163,
      "grad_norm": 0.3650052845478058,
      "learning_rate": 0.0005133607109196176,
      "loss": 3.2969,
      "step": 21550
    },
    {
      "epoch": 5.81287683031869,
      "grad_norm": 0.318183034658432,
      "learning_rate": 0.0005131587451191597,
      "loss": 3.3042,
      "step": 21600
    },
    {
      "epoch": 5.826335055986219,
      "grad_norm": 0.31363117694854736,
      "learning_rate": 0.000512956779318702,
      "loss": 3.2965,
      "step": 21650
    },
    {
      "epoch": 5.839793281653747,
      "grad_norm": 0.31825101375579834,
      "learning_rate": 0.0005127548135182441,
      "loss": 3.2942,
      "step": 21700
    },
    {
      "epoch": 5.853251507321275,
      "grad_norm": 0.30818694829940796,
      "learning_rate": 0.0005125528477177865,
      "loss": 3.2831,
      "step": 21750
    },
    {
      "epoch": 5.866709732988802,
      "grad_norm": 0.34016096591949463,
      "learning_rate": 0.0005123508819173286,
      "loss": 3.298,
      "step": 21800
    },
    {
      "epoch": 5.8801679586563305,
      "grad_norm": 0.3365114629268646,
      "learning_rate": 0.0005121489161168709,
      "loss": 3.2912,
      "step": 21850
    },
    {
      "epoch": 5.893626184323859,
      "grad_norm": 0.3417421281337738,
      "learning_rate": 0.000511946950316413,
      "loss": 3.2923,
      "step": 21900
    },
    {
      "epoch": 5.907084409991387,
      "grad_norm": 0.33967578411102295,
      "learning_rate": 0.0005117449845159553,
      "loss": 3.2958,
      "step": 21950
    },
    {
      "epoch": 5.920542635658915,
      "grad_norm": 0.33003053069114685,
      "learning_rate": 0.0005115430187154975,
      "loss": 3.2926,
      "step": 22000
    },
    {
      "epoch": 5.920542635658915,
      "eval_accuracy": 0.3837755694942228,
      "eval_loss": 3.3562772274017334,
      "eval_runtime": 83.7297,
      "eval_samples_per_second": 215.121,
      "eval_steps_per_second": 13.448,
      "step": 22000
    },
    {
      "epoch": 5.934000861326442,
      "grad_norm": 0.34383735060691833,
      "learning_rate": 0.0005113410529150397,
      "loss": 3.2996,
      "step": 22050
    },
    {
      "epoch": 5.9474590869939705,
      "grad_norm": 0.3453098237514496,
      "learning_rate": 0.0005111390871145819,
      "loss": 3.2978,
      "step": 22100
    },
    {
      "epoch": 5.960917312661499,
      "grad_norm": 0.3298710882663727,
      "learning_rate": 0.000510937121314124,
      "loss": 3.2927,
      "step": 22150
    },
    {
      "epoch": 5.974375538329027,
      "grad_norm": 0.3096718490123749,
      "learning_rate": 0.0005107351555136663,
      "loss": 3.2974,
      "step": 22200
    },
    {
      "epoch": 5.987833763996555,
      "grad_norm": 0.36883923411369324,
      "learning_rate": 0.0005105331897132085,
      "loss": 3.3032,
      "step": 22250
    },
    {
      "epoch": 6.001076658053402,
      "grad_norm": 0.3556123673915863,
      "learning_rate": 0.0005103312239127507,
      "loss": 3.2894,
      "step": 22300
    },
    {
      "epoch": 6.0145348837209305,
      "grad_norm": 0.3620011508464813,
      "learning_rate": 0.0005101292581122929,
      "loss": 3.2015,
      "step": 22350
    },
    {
      "epoch": 6.027993109388458,
      "grad_norm": 0.32757192850112915,
      "learning_rate": 0.0005099272923118351,
      "loss": 3.1983,
      "step": 22400
    },
    {
      "epoch": 6.041451335055986,
      "grad_norm": 0.3457019031047821,
      "learning_rate": 0.0005097253265113773,
      "loss": 3.1999,
      "step": 22450
    },
    {
      "epoch": 6.054909560723514,
      "grad_norm": 0.33914270997047424,
      "learning_rate": 0.0005095233607109196,
      "loss": 3.2075,
      "step": 22500
    },
    {
      "epoch": 6.068367786391042,
      "grad_norm": 0.34044554829597473,
      "learning_rate": 0.0005093213949104617,
      "loss": 3.2107,
      "step": 22550
    },
    {
      "epoch": 6.0818260120585705,
      "grad_norm": 0.3726364076137543,
      "learning_rate": 0.000509119429110004,
      "loss": 3.2079,
      "step": 22600
    },
    {
      "epoch": 6.095284237726098,
      "grad_norm": 0.3230644762516022,
      "learning_rate": 0.0005089174633095462,
      "loss": 3.2059,
      "step": 22650
    },
    {
      "epoch": 6.108742463393626,
      "grad_norm": 0.33999621868133545,
      "learning_rate": 0.0005087154975090885,
      "loss": 3.2112,
      "step": 22700
    },
    {
      "epoch": 6.122200689061154,
      "grad_norm": 0.3561903238296509,
      "learning_rate": 0.0005085135317086306,
      "loss": 3.2035,
      "step": 22750
    },
    {
      "epoch": 6.135658914728682,
      "grad_norm": 0.3232370615005493,
      "learning_rate": 0.0005083115659081729,
      "loss": 3.2179,
      "step": 22800
    },
    {
      "epoch": 6.149117140396211,
      "grad_norm": 0.31362807750701904,
      "learning_rate": 0.0005081096001077151,
      "loss": 3.215,
      "step": 22850
    },
    {
      "epoch": 6.162575366063738,
      "grad_norm": 0.33420008420944214,
      "learning_rate": 0.0005079076343072573,
      "loss": 3.2268,
      "step": 22900
    },
    {
      "epoch": 6.176033591731266,
      "grad_norm": 0.35188084840774536,
      "learning_rate": 0.0005077056685067995,
      "loss": 3.2116,
      "step": 22950
    },
    {
      "epoch": 6.189491817398794,
      "grad_norm": 0.30291420221328735,
      "learning_rate": 0.0005075037027063417,
      "loss": 3.2291,
      "step": 23000
    },
    {
      "epoch": 6.189491817398794,
      "eval_accuracy": 0.38398612710030666,
      "eval_loss": 3.3611526489257812,
      "eval_runtime": 83.7063,
      "eval_samples_per_second": 215.181,
      "eval_steps_per_second": 13.452,
      "step": 23000
    },
    {
      "epoch": 6.2029500430663225,
      "grad_norm": 0.36683255434036255,
      "learning_rate": 0.0005073017369058839,
      "loss": 3.2227,
      "step": 23050
    },
    {
      "epoch": 6.21640826873385,
      "grad_norm": 0.33583587408065796,
      "learning_rate": 0.0005070997711054262,
      "loss": 3.2167,
      "step": 23100
    },
    {
      "epoch": 6.229866494401378,
      "grad_norm": 0.32098889350891113,
      "learning_rate": 0.0005068978053049683,
      "loss": 3.2377,
      "step": 23150
    },
    {
      "epoch": 6.243324720068906,
      "grad_norm": 0.331304132938385,
      "learning_rate": 0.0005066958395045105,
      "loss": 3.2282,
      "step": 23200
    },
    {
      "epoch": 6.256782945736434,
      "grad_norm": 0.33850225806236267,
      "learning_rate": 0.0005064938737040527,
      "loss": 3.2336,
      "step": 23250
    },
    {
      "epoch": 6.270241171403962,
      "grad_norm": 0.3568723797798157,
      "learning_rate": 0.0005062919079035949,
      "loss": 3.2383,
      "step": 23300
    },
    {
      "epoch": 6.28369939707149,
      "grad_norm": 0.33676791191101074,
      "learning_rate": 0.0005060899421031372,
      "loss": 3.2305,
      "step": 23350
    },
    {
      "epoch": 6.297157622739018,
      "grad_norm": 0.3623455762863159,
      "learning_rate": 0.0005058879763026793,
      "loss": 3.2383,
      "step": 23400
    },
    {
      "epoch": 6.310615848406546,
      "grad_norm": 0.3631848990917206,
      "learning_rate": 0.0005056860105022216,
      "loss": 3.2421,
      "step": 23450
    },
    {
      "epoch": 6.324074074074074,
      "grad_norm": 0.321790486574173,
      "learning_rate": 0.0005054840447017637,
      "loss": 3.2416,
      "step": 23500
    },
    {
      "epoch": 6.337532299741602,
      "grad_norm": 0.3293253183364868,
      "learning_rate": 0.0005052820789013061,
      "loss": 3.2288,
      "step": 23550
    },
    {
      "epoch": 6.35099052540913,
      "grad_norm": 0.35676512122154236,
      "learning_rate": 0.0005050801131008482,
      "loss": 3.2387,
      "step": 23600
    },
    {
      "epoch": 6.364448751076658,
      "grad_norm": 0.3277584910392761,
      "learning_rate": 0.0005048781473003905,
      "loss": 3.2351,
      "step": 23650
    },
    {
      "epoch": 6.377906976744186,
      "grad_norm": 0.35402724146842957,
      "learning_rate": 0.0005046761814999326,
      "loss": 3.2363,
      "step": 23700
    },
    {
      "epoch": 6.3913652024117145,
      "grad_norm": 0.331562876701355,
      "learning_rate": 0.0005044742156994749,
      "loss": 3.2462,
      "step": 23750
    },
    {
      "epoch": 6.404823428079242,
      "grad_norm": 0.35490426421165466,
      "learning_rate": 0.0005042722498990171,
      "loss": 3.2334,
      "step": 23800
    },
    {
      "epoch": 6.41828165374677,
      "grad_norm": 0.3440868854522705,
      "learning_rate": 0.0005040702840985593,
      "loss": 3.2463,
      "step": 23850
    },
    {
      "epoch": 6.431739879414298,
      "grad_norm": 0.31536388397216797,
      "learning_rate": 0.0005038683182981015,
      "loss": 3.2343,
      "step": 23900
    },
    {
      "epoch": 6.445198105081826,
      "grad_norm": 0.3300188183784485,
      "learning_rate": 0.0005036663524976436,
      "loss": 3.2491,
      "step": 23950
    },
    {
      "epoch": 6.458656330749354,
      "grad_norm": 0.32236894965171814,
      "learning_rate": 0.0005034643866971859,
      "loss": 3.2504,
      "step": 24000
    },
    {
      "epoch": 6.458656330749354,
      "eval_accuracy": 0.38445091834841133,
      "eval_loss": 3.3558194637298584,
      "eval_runtime": 83.7063,
      "eval_samples_per_second": 215.181,
      "eval_steps_per_second": 13.452,
      "step": 24000
    },
    {
      "epoch": 6.472114556416882,
      "grad_norm": 0.3504488468170166,
      "learning_rate": 0.0005032624208967281,
      "loss": 3.2413,
      "step": 24050
    },
    {
      "epoch": 6.48557278208441,
      "grad_norm": 0.34342992305755615,
      "learning_rate": 0.0005030604550962703,
      "loss": 3.2317,
      "step": 24100
    },
    {
      "epoch": 6.499031007751938,
      "grad_norm": 0.3460155427455902,
      "learning_rate": 0.0005028584892958125,
      "loss": 3.2461,
      "step": 24150
    },
    {
      "epoch": 6.5124892334194655,
      "grad_norm": 0.33641359210014343,
      "learning_rate": 0.0005026565234953547,
      "loss": 3.2575,
      "step": 24200
    },
    {
      "epoch": 6.525947459086994,
      "grad_norm": 0.3465987741947174,
      "learning_rate": 0.0005024545576948969,
      "loss": 3.2487,
      "step": 24250
    },
    {
      "epoch": 6.539405684754522,
      "grad_norm": 0.342723548412323,
      "learning_rate": 0.0005022525918944392,
      "loss": 3.2503,
      "step": 24300
    },
    {
      "epoch": 6.55286391042205,
      "grad_norm": 0.3434038460254669,
      "learning_rate": 0.0005020506260939813,
      "loss": 3.2549,
      "step": 24350
    },
    {
      "epoch": 6.566322136089578,
      "grad_norm": 0.3509742319583893,
      "learning_rate": 0.0005018486602935236,
      "loss": 3.2467,
      "step": 24400
    },
    {
      "epoch": 6.579780361757106,
      "grad_norm": 0.32056328654289246,
      "learning_rate": 0.0005016466944930658,
      "loss": 3.25,
      "step": 24450
    },
    {
      "epoch": 6.593238587424634,
      "grad_norm": 0.3654623329639435,
      "learning_rate": 0.0005014447286926081,
      "loss": 3.2399,
      "step": 24500
    },
    {
      "epoch": 6.606696813092162,
      "grad_norm": 0.33851301670074463,
      "learning_rate": 0.0005012427628921502,
      "loss": 3.2485,
      "step": 24550
    },
    {
      "epoch": 6.62015503875969,
      "grad_norm": 0.3233584761619568,
      "learning_rate": 0.0005010407970916925,
      "loss": 3.2415,
      "step": 24600
    },
    {
      "epoch": 6.633613264427218,
      "grad_norm": 0.32978555560112,
      "learning_rate": 0.0005008388312912346,
      "loss": 3.243,
      "step": 24650
    },
    {
      "epoch": 6.647071490094746,
      "grad_norm": 0.3322186768054962,
      "learning_rate": 0.0005006368654907769,
      "loss": 3.2461,
      "step": 24700
    },
    {
      "epoch": 6.660529715762274,
      "grad_norm": 0.31724002957344055,
      "learning_rate": 0.0005004348996903191,
      "loss": 3.2496,
      "step": 24750
    },
    {
      "epoch": 6.673987941429802,
      "grad_norm": 0.3331548273563385,
      "learning_rate": 0.0005002329338898613,
      "loss": 3.2599,
      "step": 24800
    },
    {
      "epoch": 6.68744616709733,
      "grad_norm": 0.36503592133522034,
      "learning_rate": 0.0005000309680894035,
      "loss": 3.2475,
      "step": 24850
    },
    {
      "epoch": 6.7009043927648575,
      "grad_norm": 0.3537766933441162,
      "learning_rate": 0.0004998290022889456,
      "loss": 3.2478,
      "step": 24900
    },
    {
      "epoch": 6.714362618432386,
      "grad_norm": 0.3297971189022064,
      "learning_rate": 0.0004996270364884879,
      "loss": 3.2596,
      "step": 24950
    },
    {
      "epoch": 6.727820844099914,
      "grad_norm": 0.3298289179801941,
      "learning_rate": 0.0004994250706880301,
      "loss": 3.268,
      "step": 25000
    },
    {
      "epoch": 6.727820844099914,
      "eval_accuracy": 0.38563908036086403,
      "eval_loss": 3.344057083129883,
      "eval_runtime": 83.7717,
      "eval_samples_per_second": 215.013,
      "eval_steps_per_second": 13.441,
      "step": 25000
    },
    {
      "epoch": 6.741279069767442,
      "grad_norm": 0.3281354010105133,
      "learning_rate": 0.0004992231048875723,
      "loss": 3.2595,
      "step": 25050
    },
    {
      "epoch": 6.754737295434969,
      "grad_norm": 0.32790902256965637,
      "learning_rate": 0.0004990211390871145,
      "loss": 3.2463,
      "step": 25100
    },
    {
      "epoch": 6.768195521102498,
      "grad_norm": 0.3453257083892822,
      "learning_rate": 0.0004988191732866567,
      "loss": 3.2652,
      "step": 25150
    },
    {
      "epoch": 6.781653746770026,
      "grad_norm": 0.3118387758731842,
      "learning_rate": 0.0004986172074861989,
      "loss": 3.2694,
      "step": 25200
    },
    {
      "epoch": 6.795111972437554,
      "grad_norm": 0.3285624384880066,
      "learning_rate": 0.0004984152416857412,
      "loss": 3.2546,
      "step": 25250
    },
    {
      "epoch": 6.808570198105082,
      "grad_norm": 0.3390444219112396,
      "learning_rate": 0.0004982132758852834,
      "loss": 3.2541,
      "step": 25300
    },
    {
      "epoch": 6.822028423772609,
      "grad_norm": 0.3274039924144745,
      "learning_rate": 0.0004980113100848256,
      "loss": 3.2581,
      "step": 25350
    },
    {
      "epoch": 6.835486649440138,
      "grad_norm": 0.3342575132846832,
      "learning_rate": 0.0004978093442843678,
      "loss": 3.2586,
      "step": 25400
    },
    {
      "epoch": 6.848944875107666,
      "grad_norm": 0.34611400961875916,
      "learning_rate": 0.0004976073784839101,
      "loss": 3.2508,
      "step": 25450
    },
    {
      "epoch": 6.862403100775194,
      "grad_norm": 0.34767094254493713,
      "learning_rate": 0.0004974054126834522,
      "loss": 3.2579,
      "step": 25500
    },
    {
      "epoch": 6.875861326442722,
      "grad_norm": 0.3290567100048065,
      "learning_rate": 0.0004972034468829945,
      "loss": 3.2484,
      "step": 25550
    },
    {
      "epoch": 6.8893195521102495,
      "grad_norm": 0.347561240196228,
      "learning_rate": 0.0004970014810825366,
      "loss": 3.2566,
      "step": 25600
    },
    {
      "epoch": 6.902777777777778,
      "grad_norm": 0.32147741317749023,
      "learning_rate": 0.0004967995152820789,
      "loss": 3.2559,
      "step": 25650
    },
    {
      "epoch": 6.916236003445306,
      "grad_norm": 0.35263851284980774,
      "learning_rate": 0.0004965975494816211,
      "loss": 3.2615,
      "step": 25700
    },
    {
      "epoch": 6.929694229112834,
      "grad_norm": 0.3382377624511719,
      "learning_rate": 0.0004963955836811633,
      "loss": 3.2644,
      "step": 25750
    },
    {
      "epoch": 6.943152454780362,
      "grad_norm": 0.32801884412765503,
      "learning_rate": 0.0004961936178807055,
      "loss": 3.2482,
      "step": 25800
    },
    {
      "epoch": 6.9566106804478895,
      "grad_norm": 0.3700309097766876,
      "learning_rate": 0.0004959916520802476,
      "loss": 3.265,
      "step": 25850
    },
    {
      "epoch": 6.970068906115418,
      "grad_norm": 0.3657413721084595,
      "learning_rate": 0.0004957896862797899,
      "loss": 3.2475,
      "step": 25900
    },
    {
      "epoch": 6.983527131782946,
      "grad_norm": 0.3408574163913727,
      "learning_rate": 0.0004955877204793321,
      "loss": 3.2587,
      "step": 25950
    },
    {
      "epoch": 6.996985357450473,
      "grad_norm": 0.35131752490997314,
      "learning_rate": 0.0004953857546788743,
      "loss": 3.2672,
      "step": 26000
    },
    {
      "epoch": 6.996985357450473,
      "eval_accuracy": 0.3864859825999888,
      "eval_loss": 3.335435152053833,
      "eval_runtime": 83.908,
      "eval_samples_per_second": 214.664,
      "eval_steps_per_second": 13.419,
      "step": 26000
    },
    {
      "epoch": 7.010228251507321,
      "grad_norm": 0.3232435882091522,
      "learning_rate": 0.0004951837888784165,
      "loss": 3.1775,
      "step": 26050
    },
    {
      "epoch": 7.0236864771748495,
      "grad_norm": 0.32738637924194336,
      "learning_rate": 0.0004949818230779587,
      "loss": 3.1609,
      "step": 26100
    },
    {
      "epoch": 7.037144702842378,
      "grad_norm": 0.3246043026447296,
      "learning_rate": 0.0004947798572775009,
      "loss": 3.1585,
      "step": 26150
    },
    {
      "epoch": 7.050602928509905,
      "grad_norm": 0.34505170583724976,
      "learning_rate": 0.0004945778914770432,
      "loss": 3.1678,
      "step": 26200
    },
    {
      "epoch": 7.064061154177433,
      "grad_norm": 0.3476838767528534,
      "learning_rate": 0.0004943759256765854,
      "loss": 3.1757,
      "step": 26250
    },
    {
      "epoch": 7.077519379844961,
      "grad_norm": 0.3383033573627472,
      "learning_rate": 0.0004941739598761276,
      "loss": 3.1767,
      "step": 26300
    },
    {
      "epoch": 7.09097760551249,
      "grad_norm": 0.3575884997844696,
      "learning_rate": 0.0004939719940756698,
      "loss": 3.1746,
      "step": 26350
    },
    {
      "epoch": 7.104435831180017,
      "grad_norm": 0.3479248881340027,
      "learning_rate": 0.0004937700282752121,
      "loss": 3.174,
      "step": 26400
    },
    {
      "epoch": 7.117894056847545,
      "grad_norm": 0.3662038743495941,
      "learning_rate": 0.0004935680624747542,
      "loss": 3.1734,
      "step": 26450
    },
    {
      "epoch": 7.131352282515073,
      "grad_norm": 0.3241896331310272,
      "learning_rate": 0.0004933660966742965,
      "loss": 3.1778,
      "step": 26500
    },
    {
      "epoch": 7.144810508182601,
      "grad_norm": 0.3458649218082428,
      "learning_rate": 0.0004931641308738386,
      "loss": 3.1855,
      "step": 26550
    },
    {
      "epoch": 7.15826873385013,
      "grad_norm": 0.3444419503211975,
      "learning_rate": 0.0004929621650733809,
      "loss": 3.1863,
      "step": 26600
    },
    {
      "epoch": 7.171726959517657,
      "grad_norm": 0.4012559950351715,
      "learning_rate": 0.0004927601992729231,
      "loss": 3.1897,
      "step": 26650
    },
    {
      "epoch": 7.185185185185185,
      "grad_norm": 0.3804725110530853,
      "learning_rate": 0.0004925582334724652,
      "loss": 3.1842,
      "step": 26700
    },
    {
      "epoch": 7.198643410852713,
      "grad_norm": 0.3514227569103241,
      "learning_rate": 0.0004923562676720075,
      "loss": 3.1924,
      "step": 26750
    },
    {
      "epoch": 7.2121016365202415,
      "grad_norm": 0.34808680415153503,
      "learning_rate": 0.0004921543018715497,
      "loss": 3.1951,
      "step": 26800
    },
    {
      "epoch": 7.225559862187769,
      "grad_norm": 0.3160098195075989,
      "learning_rate": 0.0004919523360710919,
      "loss": 3.1972,
      "step": 26850
    },
    {
      "epoch": 7.239018087855297,
      "grad_norm": 0.3309127986431122,
      "learning_rate": 0.0004917503702706341,
      "loss": 3.1872,
      "step": 26900
    },
    {
      "epoch": 7.252476313522825,
      "grad_norm": 0.3276006877422333,
      "learning_rate": 0.0004915484044701763,
      "loss": 3.2009,
      "step": 26950
    },
    {
      "epoch": 7.265934539190353,
      "grad_norm": 0.3357968032360077,
      "learning_rate": 0.0004913464386697185,
      "loss": 3.2013,
      "step": 27000
    },
    {
      "epoch": 7.265934539190353,
      "eval_accuracy": 0.386140702893005,
      "eval_loss": 3.3429081439971924,
      "eval_runtime": 83.6238,
      "eval_samples_per_second": 215.393,
      "eval_steps_per_second": 13.465,
      "step": 27000
    },
    {
      "epoch": 7.279392764857882,
      "grad_norm": 0.3478224277496338,
      "learning_rate": 0.0004911444728692608,
      "loss": 3.1965,
      "step": 27050
    },
    {
      "epoch": 7.292850990525409,
      "grad_norm": 0.3382481336593628,
      "learning_rate": 0.000490942507068803,
      "loss": 3.1978,
      "step": 27100
    },
    {
      "epoch": 7.306309216192937,
      "grad_norm": 0.36683815717697144,
      "learning_rate": 0.0004907405412683452,
      "loss": 3.2029,
      "step": 27150
    },
    {
      "epoch": 7.319767441860465,
      "grad_norm": 0.3402043581008911,
      "learning_rate": 0.0004905385754678874,
      "loss": 3.1898,
      "step": 27200
    },
    {
      "epoch": 7.333225667527993,
      "grad_norm": 0.3558410108089447,
      "learning_rate": 0.0004903366096674297,
      "loss": 3.211,
      "step": 27250
    },
    {
      "epoch": 7.346683893195521,
      "grad_norm": 0.34050697088241577,
      "learning_rate": 0.0004901346438669718,
      "loss": 3.1997,
      "step": 27300
    },
    {
      "epoch": 7.360142118863049,
      "grad_norm": 0.36426642537117004,
      "learning_rate": 0.0004899326780665141,
      "loss": 3.2192,
      "step": 27350
    },
    {
      "epoch": 7.373600344530577,
      "grad_norm": 0.33061543107032776,
      "learning_rate": 0.0004897307122660562,
      "loss": 3.2031,
      "step": 27400
    },
    {
      "epoch": 7.387058570198105,
      "grad_norm": 0.353532999753952,
      "learning_rate": 0.0004895287464655985,
      "loss": 3.2152,
      "step": 27450
    },
    {
      "epoch": 7.4005167958656335,
      "grad_norm": 0.34685155749320984,
      "learning_rate": 0.0004893267806651407,
      "loss": 3.2083,
      "step": 27500
    },
    {
      "epoch": 7.413975021533161,
      "grad_norm": 0.36312010884284973,
      "learning_rate": 0.0004891248148646829,
      "loss": 3.2145,
      "step": 27550
    },
    {
      "epoch": 7.427433247200689,
      "grad_norm": 0.3602902591228485,
      "learning_rate": 0.0004889228490642251,
      "loss": 3.2141,
      "step": 27600
    },
    {
      "epoch": 7.440891472868217,
      "grad_norm": 0.3237648606300354,
      "learning_rate": 0.0004887208832637672,
      "loss": 3.2109,
      "step": 27650
    },
    {
      "epoch": 7.454349698535745,
      "grad_norm": 0.3332701623439789,
      "learning_rate": 0.0004885189174633095,
      "loss": 3.2202,
      "step": 27700
    },
    {
      "epoch": 7.467807924203273,
      "grad_norm": 0.3481634557247162,
      "learning_rate": 0.0004883169516628517,
      "loss": 3.2054,
      "step": 27750
    },
    {
      "epoch": 7.481266149870801,
      "grad_norm": 0.3382105827331543,
      "learning_rate": 0.0004881149858623939,
      "loss": 3.2138,
      "step": 27800
    },
    {
      "epoch": 7.494724375538329,
      "grad_norm": 0.34605178236961365,
      "learning_rate": 0.00048791302006193614,
      "loss": 3.2098,
      "step": 27850
    },
    {
      "epoch": 7.508182601205857,
      "grad_norm": 0.33626753091812134,
      "learning_rate": 0.00048771105426147833,
      "loss": 3.2179,
      "step": 27900
    },
    {
      "epoch": 7.521640826873385,
      "grad_norm": 0.34346142411231995,
      "learning_rate": 0.00048750908846102053,
      "loss": 3.2201,
      "step": 27950
    },
    {
      "epoch": 7.535099052540913,
      "grad_norm": 0.34637099504470825,
      "learning_rate": 0.0004873071226605628,
      "loss": 3.22,
      "step": 28000
    },
    {
      "epoch": 7.535099052540913,
      "eval_accuracy": 0.3869608780056609,
      "eval_loss": 3.33520770072937,
      "eval_runtime": 83.6506,
      "eval_samples_per_second": 215.324,
      "eval_steps_per_second": 13.461,
      "step": 28000
    },
    {
      "epoch": 7.548557278208441,
      "grad_norm": 0.3463466465473175,
      "learning_rate": 0.00048710515686010503,
      "loss": 3.2042,
      "step": 28050
    },
    {
      "epoch": 7.562015503875969,
      "grad_norm": 0.3281874358654022,
      "learning_rate": 0.0004869031910596472,
      "loss": 3.223,
      "step": 28100
    },
    {
      "epoch": 7.575473729543497,
      "grad_norm": 0.34871819615364075,
      "learning_rate": 0.0004867012252591894,
      "loss": 3.2171,
      "step": 28150
    },
    {
      "epoch": 7.588931955211025,
      "grad_norm": 0.3180304169654846,
      "learning_rate": 0.0004864992594587316,
      "loss": 3.2252,
      "step": 28200
    },
    {
      "epoch": 7.602390180878553,
      "grad_norm": 0.3471381962299347,
      "learning_rate": 0.0004862972936582738,
      "loss": 3.2369,
      "step": 28250
    },
    {
      "epoch": 7.615848406546081,
      "grad_norm": 0.3487418591976166,
      "learning_rate": 0.00048609532785781606,
      "loss": 3.2188,
      "step": 28300
    },
    {
      "epoch": 7.629306632213609,
      "grad_norm": 0.350359171628952,
      "learning_rate": 0.00048589336205735826,
      "loss": 3.2196,
      "step": 28350
    },
    {
      "epoch": 7.6427648578811365,
      "grad_norm": 0.341923326253891,
      "learning_rate": 0.00048569139625690046,
      "loss": 3.2175,
      "step": 28400
    },
    {
      "epoch": 7.656223083548665,
      "grad_norm": 0.33095625042915344,
      "learning_rate": 0.00048548943045644265,
      "loss": 3.2147,
      "step": 28450
    },
    {
      "epoch": 7.669681309216193,
      "grad_norm": 0.3263479471206665,
      "learning_rate": 0.00048528746465598485,
      "loss": 3.2136,
      "step": 28500
    },
    {
      "epoch": 7.683139534883721,
      "grad_norm": 0.3331003785133362,
      "learning_rate": 0.0004850854988555271,
      "loss": 3.2203,
      "step": 28550
    },
    {
      "epoch": 7.696597760551249,
      "grad_norm": 0.3453432321548462,
      "learning_rate": 0.0004848835330550693,
      "loss": 3.2268,
      "step": 28600
    },
    {
      "epoch": 7.7100559862187765,
      "grad_norm": 0.3308742344379425,
      "learning_rate": 0.0004846815672546115,
      "loss": 3.2208,
      "step": 28650
    },
    {
      "epoch": 7.723514211886305,
      "grad_norm": 0.3838219940662384,
      "learning_rate": 0.0004844796014541537,
      "loss": 3.2193,
      "step": 28700
    },
    {
      "epoch": 7.736972437553833,
      "grad_norm": 0.3169751465320587,
      "learning_rate": 0.0004842776356536959,
      "loss": 3.2247,
      "step": 28750
    },
    {
      "epoch": 7.750430663221361,
      "grad_norm": 0.3663519322872162,
      "learning_rate": 0.00048407566985323813,
      "loss": 3.2296,
      "step": 28800
    },
    {
      "epoch": 7.763888888888889,
      "grad_norm": 0.32905447483062744,
      "learning_rate": 0.0004838737040527804,
      "loss": 3.2233,
      "step": 28850
    },
    {
      "epoch": 7.777347114556417,
      "grad_norm": 0.368269145488739,
      "learning_rate": 0.0004836717382523226,
      "loss": 3.2176,
      "step": 28900
    },
    {
      "epoch": 7.790805340223945,
      "grad_norm": 0.36298301815986633,
      "learning_rate": 0.0004834697724518648,
      "loss": 3.2218,
      "step": 28950
    },
    {
      "epoch": 7.804263565891473,
      "grad_norm": 0.34526821970939636,
      "learning_rate": 0.000483267806651407,
      "loss": 3.2162,
      "step": 29000
    },
    {
      "epoch": 7.804263565891473,
      "eval_accuracy": 0.38757734026413354,
      "eval_loss": 3.32964825630188,
      "eval_runtime": 83.8197,
      "eval_samples_per_second": 214.89,
      "eval_steps_per_second": 13.434,
      "step": 29000
    },
    {
      "epoch": 7.817721791559001,
      "grad_norm": 0.3341210186481476,
      "learning_rate": 0.0004830658408509492,
      "loss": 3.2275,
      "step": 29050
    },
    {
      "epoch": 7.831180017226529,
      "grad_norm": 0.3847617208957672,
      "learning_rate": 0.0004828638750504914,
      "loss": 3.234,
      "step": 29100
    },
    {
      "epoch": 7.844638242894057,
      "grad_norm": 0.334783136844635,
      "learning_rate": 0.0004826619092500336,
      "loss": 3.2251,
      "step": 29150
    },
    {
      "epoch": 7.858096468561585,
      "grad_norm": 0.3237602710723877,
      "learning_rate": 0.0004824599434495758,
      "loss": 3.2253,
      "step": 29200
    },
    {
      "epoch": 7.871554694229113,
      "grad_norm": 0.33085930347442627,
      "learning_rate": 0.00048225797764911806,
      "loss": 3.2314,
      "step": 29250
    },
    {
      "epoch": 7.885012919896641,
      "grad_norm": 0.34472253918647766,
      "learning_rate": 0.00048205601184866026,
      "loss": 3.2365,
      "step": 29300
    },
    {
      "epoch": 7.8984711455641685,
      "grad_norm": 0.33944690227508545,
      "learning_rate": 0.00048185404604820245,
      "loss": 3.2324,
      "step": 29350
    },
    {
      "epoch": 7.911929371231697,
      "grad_norm": 0.33092230558395386,
      "learning_rate": 0.00048165208024774465,
      "loss": 3.2388,
      "step": 29400
    },
    {
      "epoch": 7.925387596899225,
      "grad_norm": 0.3233077824115753,
      "learning_rate": 0.00048145011444728685,
      "loss": 3.236,
      "step": 29450
    },
    {
      "epoch": 7.938845822566753,
      "grad_norm": 0.3199717402458191,
      "learning_rate": 0.0004812481486468291,
      "loss": 3.2202,
      "step": 29500
    },
    {
      "epoch": 7.95230404823428,
      "grad_norm": 0.32801392674446106,
      "learning_rate": 0.0004810461828463713,
      "loss": 3.2317,
      "step": 29550
    },
    {
      "epoch": 7.965762273901809,
      "grad_norm": 0.3065549433231354,
      "learning_rate": 0.0004808442170459135,
      "loss": 3.2256,
      "step": 29600
    },
    {
      "epoch": 7.979220499569337,
      "grad_norm": 0.34186145663261414,
      "learning_rate": 0.0004806422512454557,
      "loss": 3.2237,
      "step": 29650
    },
    {
      "epoch": 7.992678725236865,
      "grad_norm": 0.3473663628101349,
      "learning_rate": 0.00048044028544499793,
      "loss": 3.2209,
      "step": 29700
    },
    {
      "epoch": 8.005921619293712,
      "grad_norm": 0.31494632363319397,
      "learning_rate": 0.0004802383196445402,
      "loss": 3.1834,
      "step": 29750
    },
    {
      "epoch": 8.01937984496124,
      "grad_norm": 0.32476192712783813,
      "learning_rate": 0.0004800363538440824,
      "loss": 3.1408,
      "step": 29800
    },
    {
      "epoch": 8.032838070628769,
      "grad_norm": 0.3453425467014313,
      "learning_rate": 0.0004798343880436246,
      "loss": 3.1313,
      "step": 29850
    },
    {
      "epoch": 8.046296296296296,
      "grad_norm": 0.3348669111728668,
      "learning_rate": 0.00047963242224316683,
      "loss": 3.1316,
      "step": 29900
    },
    {
      "epoch": 8.059754521963825,
      "grad_norm": 0.3372720777988434,
      "learning_rate": 0.000479430456442709,
      "loss": 3.1433,
      "step": 29950
    },
    {
      "epoch": 8.073212747631352,
      "grad_norm": 0.35592925548553467,
      "learning_rate": 0.0004792284906422512,
      "loss": 3.1399,
      "step": 30000
    },
    {
      "epoch": 8.073212747631352,
      "eval_accuracy": 0.38754007439267496,
      "eval_loss": 3.334799289703369,
      "eval_runtime": 83.5558,
      "eval_samples_per_second": 215.568,
      "eval_steps_per_second": 13.476,
      "step": 30000
    },
    {
      "epoch": 8.08667097329888,
      "grad_norm": 0.33823978900909424,
      "learning_rate": 0.0004790265248417934,
      "loss": 3.1519,
      "step": 30050
    },
    {
      "epoch": 8.100129198966409,
      "grad_norm": 0.32208123803138733,
      "learning_rate": 0.0004788245590413356,
      "loss": 3.1346,
      "step": 30100
    },
    {
      "epoch": 8.113587424633936,
      "grad_norm": 0.35270413756370544,
      "learning_rate": 0.00047862259324087786,
      "loss": 3.1424,
      "step": 30150
    },
    {
      "epoch": 8.127045650301465,
      "grad_norm": 0.340955913066864,
      "learning_rate": 0.00047842062744042006,
      "loss": 3.1539,
      "step": 30200
    },
    {
      "epoch": 8.140503875968992,
      "grad_norm": 0.35465332865715027,
      "learning_rate": 0.00047821866163996225,
      "loss": 3.1458,
      "step": 30250
    },
    {
      "epoch": 8.15396210163652,
      "grad_norm": 0.3508666157722473,
      "learning_rate": 0.00047801669583950445,
      "loss": 3.1626,
      "step": 30300
    },
    {
      "epoch": 8.167420327304049,
      "grad_norm": 0.34346407651901245,
      "learning_rate": 0.00047781473003904665,
      "loss": 3.1608,
      "step": 30350
    },
    {
      "epoch": 8.180878552971576,
      "grad_norm": 0.33316704630851746,
      "learning_rate": 0.0004776127642385889,
      "loss": 3.1583,
      "step": 30400
    },
    {
      "epoch": 8.194336778639105,
      "grad_norm": 0.3661840260028839,
      "learning_rate": 0.0004774107984381311,
      "loss": 3.1624,
      "step": 30450
    },
    {
      "epoch": 8.207795004306632,
      "grad_norm": 0.35324305295944214,
      "learning_rate": 0.0004772088326376733,
      "loss": 3.1525,
      "step": 30500
    },
    {
      "epoch": 8.22125322997416,
      "grad_norm": 0.3420873284339905,
      "learning_rate": 0.0004770068668372155,
      "loss": 3.1666,
      "step": 30550
    },
    {
      "epoch": 8.234711455641689,
      "grad_norm": 0.3589732050895691,
      "learning_rate": 0.0004768049010367577,
      "loss": 3.1777,
      "step": 30600
    },
    {
      "epoch": 8.248169681309216,
      "grad_norm": 0.3344517946243286,
      "learning_rate": 0.0004766029352363,
      "loss": 3.1749,
      "step": 30650
    },
    {
      "epoch": 8.261627906976745,
      "grad_norm": 0.34635987877845764,
      "learning_rate": 0.0004764009694358422,
      "loss": 3.1671,
      "step": 30700
    },
    {
      "epoch": 8.275086132644272,
      "grad_norm": 0.357856810092926,
      "learning_rate": 0.0004761990036353844,
      "loss": 3.168,
      "step": 30750
    },
    {
      "epoch": 8.2885443583118,
      "grad_norm": 0.3461187481880188,
      "learning_rate": 0.0004759970378349266,
      "loss": 3.1693,
      "step": 30800
    },
    {
      "epoch": 8.302002583979329,
      "grad_norm": 0.32600918412208557,
      "learning_rate": 0.0004757950720344688,
      "loss": 3.1826,
      "step": 30850
    },
    {
      "epoch": 8.315460809646856,
      "grad_norm": 0.33008185029029846,
      "learning_rate": 0.000475593106234011,
      "loss": 3.1647,
      "step": 30900
    },
    {
      "epoch": 8.328919035314383,
      "grad_norm": 0.3357200026512146,
      "learning_rate": 0.0004753911404335532,
      "loss": 3.1762,
      "step": 30950
    },
    {
      "epoch": 8.342377260981912,
      "grad_norm": 0.3322162926197052,
      "learning_rate": 0.0004751891746330954,
      "loss": 3.1783,
      "step": 31000
    },
    {
      "epoch": 8.342377260981912,
      "eval_accuracy": 0.3883333050851509,
      "eval_loss": 3.328897476196289,
      "eval_runtime": 83.7436,
      "eval_samples_per_second": 215.085,
      "eval_steps_per_second": 13.446,
      "step": 31000
    },
    {
      "epoch": 8.35583548664944,
      "grad_norm": 0.3404552936553955,
      "learning_rate": 0.0004749872088326376,
      "loss": 3.1839,
      "step": 31050
    },
    {
      "epoch": 8.369293712316969,
      "grad_norm": 0.3331236243247986,
      "learning_rate": 0.00047478524303217986,
      "loss": 3.1768,
      "step": 31100
    },
    {
      "epoch": 8.382751937984496,
      "grad_norm": 0.36006081104278564,
      "learning_rate": 0.00047458327723172206,
      "loss": 3.1749,
      "step": 31150
    },
    {
      "epoch": 8.396210163652023,
      "grad_norm": 0.33484339714050293,
      "learning_rate": 0.00047438131143126425,
      "loss": 3.1632,
      "step": 31200
    },
    {
      "epoch": 8.409668389319553,
      "grad_norm": 0.3836902678012848,
      "learning_rate": 0.00047417934563080645,
      "loss": 3.1913,
      "step": 31250
    },
    {
      "epoch": 8.42312661498708,
      "grad_norm": 0.3145540952682495,
      "learning_rate": 0.00047397737983034864,
      "loss": 3.1766,
      "step": 31300
    },
    {
      "epoch": 8.436584840654609,
      "grad_norm": 0.3409503400325775,
      "learning_rate": 0.0004737754140298909,
      "loss": 3.1753,
      "step": 31350
    },
    {
      "epoch": 8.450043066322136,
      "grad_norm": 0.33025360107421875,
      "learning_rate": 0.0004735734482294331,
      "loss": 3.191,
      "step": 31400
    },
    {
      "epoch": 8.463501291989663,
      "grad_norm": 0.3531903326511383,
      "learning_rate": 0.0004733714824289753,
      "loss": 3.184,
      "step": 31450
    },
    {
      "epoch": 8.476959517657193,
      "grad_norm": 0.33644282817840576,
      "learning_rate": 0.0004731695166285175,
      "loss": 3.1842,
      "step": 31500
    },
    {
      "epoch": 8.49041774332472,
      "grad_norm": 0.34396588802337646,
      "learning_rate": 0.0004729675508280598,
      "loss": 3.1943,
      "step": 31550
    },
    {
      "epoch": 8.503875968992247,
      "grad_norm": 0.3307275176048279,
      "learning_rate": 0.000472765585027602,
      "loss": 3.1885,
      "step": 31600
    },
    {
      "epoch": 8.517334194659776,
      "grad_norm": 0.34120792150497437,
      "learning_rate": 0.0004725636192271442,
      "loss": 3.1979,
      "step": 31650
    },
    {
      "epoch": 8.530792420327304,
      "grad_norm": 0.3874061107635498,
      "learning_rate": 0.0004723616534266864,
      "loss": 3.186,
      "step": 31700
    },
    {
      "epoch": 8.544250645994833,
      "grad_norm": 0.32064998149871826,
      "learning_rate": 0.0004721596876262286,
      "loss": 3.187,
      "step": 31750
    },
    {
      "epoch": 8.55770887166236,
      "grad_norm": 0.3350982367992401,
      "learning_rate": 0.0004719577218257708,
      "loss": 3.1965,
      "step": 31800
    },
    {
      "epoch": 8.571167097329887,
      "grad_norm": 0.3344133496284485,
      "learning_rate": 0.000471755756025313,
      "loss": 3.1927,
      "step": 31850
    },
    {
      "epoch": 8.584625322997416,
      "grad_norm": 0.3439996838569641,
      "learning_rate": 0.0004715537902248552,
      "loss": 3.1899,
      "step": 31900
    },
    {
      "epoch": 8.598083548664944,
      "grad_norm": 0.3472214937210083,
      "learning_rate": 0.0004713518244243974,
      "loss": 3.185,
      "step": 31950
    },
    {
      "epoch": 8.611541774332473,
      "grad_norm": 0.3920539617538452,
      "learning_rate": 0.00047114985862393966,
      "loss": 3.1991,
      "step": 32000
    },
    {
      "epoch": 8.611541774332473,
      "eval_accuracy": 0.38856254995039186,
      "eval_loss": 3.322324752807617,
      "eval_runtime": 83.7446,
      "eval_samples_per_second": 215.082,
      "eval_steps_per_second": 13.446,
      "step": 32000
    },
    {
      "epoch": 8.625,
      "grad_norm": 0.36540019512176514,
      "learning_rate": 0.00047094789282348186,
      "loss": 3.1948,
      "step": 32050
    },
    {
      "epoch": 8.638458225667527,
      "grad_norm": 0.35658854246139526,
      "learning_rate": 0.00047074592702302405,
      "loss": 3.1902,
      "step": 32100
    },
    {
      "epoch": 8.651916451335056,
      "grad_norm": 0.3959910273551941,
      "learning_rate": 0.00047054396122256625,
      "loss": 3.1995,
      "step": 32150
    },
    {
      "epoch": 8.665374677002584,
      "grad_norm": 0.33812761306762695,
      "learning_rate": 0.00047034199542210844,
      "loss": 3.1969,
      "step": 32200
    },
    {
      "epoch": 8.678832902670113,
      "grad_norm": 0.3247467577457428,
      "learning_rate": 0.0004701400296216507,
      "loss": 3.1863,
      "step": 32250
    },
    {
      "epoch": 8.69229112833764,
      "grad_norm": 0.3646974265575409,
      "learning_rate": 0.0004699380638211929,
      "loss": 3.1965,
      "step": 32300
    },
    {
      "epoch": 8.705749354005167,
      "grad_norm": 0.3399631381034851,
      "learning_rate": 0.0004697360980207351,
      "loss": 3.2021,
      "step": 32350
    },
    {
      "epoch": 8.719207579672696,
      "grad_norm": 0.36749687790870667,
      "learning_rate": 0.0004695341322202773,
      "loss": 3.1908,
      "step": 32400
    },
    {
      "epoch": 8.732665805340224,
      "grad_norm": 0.3429027199745178,
      "learning_rate": 0.0004693321664198196,
      "loss": 3.1982,
      "step": 32450
    },
    {
      "epoch": 8.746124031007753,
      "grad_norm": 0.3495529294013977,
      "learning_rate": 0.0004691302006193618,
      "loss": 3.1977,
      "step": 32500
    },
    {
      "epoch": 8.75958225667528,
      "grad_norm": 0.3478645086288452,
      "learning_rate": 0.000468928234818904,
      "loss": 3.2109,
      "step": 32550
    },
    {
      "epoch": 8.773040482342807,
      "grad_norm": 0.33168649673461914,
      "learning_rate": 0.0004687262690184462,
      "loss": 3.1961,
      "step": 32600
    },
    {
      "epoch": 8.786498708010337,
      "grad_norm": 0.33454975485801697,
      "learning_rate": 0.00046852430321798837,
      "loss": 3.2003,
      "step": 32650
    },
    {
      "epoch": 8.799956933677864,
      "grad_norm": 0.36655351519584656,
      "learning_rate": 0.0004683223374175306,
      "loss": 3.1986,
      "step": 32700
    },
    {
      "epoch": 8.813415159345391,
      "grad_norm": 0.3267432451248169,
      "learning_rate": 0.0004681203716170728,
      "loss": 3.2074,
      "step": 32750
    },
    {
      "epoch": 8.82687338501292,
      "grad_norm": 0.36144155263900757,
      "learning_rate": 0.000467918405816615,
      "loss": 3.197,
      "step": 32800
    },
    {
      "epoch": 8.840331610680447,
      "grad_norm": 0.33456239104270935,
      "learning_rate": 0.0004677164400161572,
      "loss": 3.194,
      "step": 32850
    },
    {
      "epoch": 8.853789836347977,
      "grad_norm": 0.3679030239582062,
      "learning_rate": 0.0004675144742156994,
      "loss": 3.2053,
      "step": 32900
    },
    {
      "epoch": 8.867248062015504,
      "grad_norm": 0.3380067050457001,
      "learning_rate": 0.00046731250841524166,
      "loss": 3.2055,
      "step": 32950
    },
    {
      "epoch": 8.880706287683031,
      "grad_norm": 0.3196934759616852,
      "learning_rate": 0.00046711054261478385,
      "loss": 3.2035,
      "step": 33000
    },
    {
      "epoch": 8.880706287683031,
      "eval_accuracy": 0.38922181906995684,
      "eval_loss": 3.3164241313934326,
      "eval_runtime": 83.6278,
      "eval_samples_per_second": 215.383,
      "eval_steps_per_second": 13.464,
      "step": 33000
    },
    {
      "epoch": 8.89416451335056,
      "grad_norm": 0.3437865972518921,
      "learning_rate": 0.00046690857681432605,
      "loss": 3.2039,
      "step": 33050
    },
    {
      "epoch": 8.907622739018088,
      "grad_norm": 0.3278834819793701,
      "learning_rate": 0.00046670661101386825,
      "loss": 3.2033,
      "step": 33100
    },
    {
      "epoch": 8.921080964685617,
      "grad_norm": 0.3331878185272217,
      "learning_rate": 0.00046650464521341044,
      "loss": 3.1912,
      "step": 33150
    },
    {
      "epoch": 8.934539190353144,
      "grad_norm": 0.33546411991119385,
      "learning_rate": 0.0004663026794129527,
      "loss": 3.2041,
      "step": 33200
    },
    {
      "epoch": 8.947997416020671,
      "grad_norm": 0.34766891598701477,
      "learning_rate": 0.0004661007136124949,
      "loss": 3.2119,
      "step": 33250
    },
    {
      "epoch": 8.9614556416882,
      "grad_norm": 0.3531126081943512,
      "learning_rate": 0.00046589874781203714,
      "loss": 3.2098,
      "step": 33300
    },
    {
      "epoch": 8.974913867355728,
      "grad_norm": 0.33179959654808044,
      "learning_rate": 0.0004656967820115794,
      "loss": 3.2111,
      "step": 33350
    },
    {
      "epoch": 8.988372093023255,
      "grad_norm": 0.3321326971054077,
      "learning_rate": 0.0004654948162111216,
      "loss": 3.2076,
      "step": 33400
    },
    {
      "epoch": 9.001614987080103,
      "grad_norm": 0.3816087543964386,
      "learning_rate": 0.0004652928504106638,
      "loss": 3.1958,
      "step": 33450
    },
    {
      "epoch": 9.015073212747632,
      "grad_norm": 0.3416728675365448,
      "learning_rate": 0.000465090884610206,
      "loss": 3.1112,
      "step": 33500
    },
    {
      "epoch": 9.02853143841516,
      "grad_norm": 0.3327523171901703,
      "learning_rate": 0.00046488891880974817,
      "loss": 3.1027,
      "step": 33550
    },
    {
      "epoch": 9.041989664082687,
      "grad_norm": 0.34036681056022644,
      "learning_rate": 0.0004646869530092904,
      "loss": 3.1179,
      "step": 33600
    },
    {
      "epoch": 9.055447889750216,
      "grad_norm": 0.31692638993263245,
      "learning_rate": 0.0004644849872088326,
      "loss": 3.1116,
      "step": 33650
    },
    {
      "epoch": 9.068906115417743,
      "grad_norm": 0.335548996925354,
      "learning_rate": 0.0004642830214083748,
      "loss": 3.1215,
      "step": 33700
    },
    {
      "epoch": 9.082364341085272,
      "grad_norm": 0.3582102060317993,
      "learning_rate": 0.000464081055607917,
      "loss": 3.1104,
      "step": 33750
    },
    {
      "epoch": 9.0958225667528,
      "grad_norm": 0.3529094457626343,
      "learning_rate": 0.0004638790898074592,
      "loss": 3.1318,
      "step": 33800
    },
    {
      "epoch": 9.109280792420327,
      "grad_norm": 0.36817148327827454,
      "learning_rate": 0.00046367712400700146,
      "loss": 3.124,
      "step": 33850
    },
    {
      "epoch": 9.122739018087856,
      "grad_norm": 0.3662441372871399,
      "learning_rate": 0.00046347515820654365,
      "loss": 3.1243,
      "step": 33900
    },
    {
      "epoch": 9.136197243755383,
      "grad_norm": 0.3498477637767792,
      "learning_rate": 0.00046327319240608585,
      "loss": 3.1238,
      "step": 33950
    },
    {
      "epoch": 9.14965546942291,
      "grad_norm": 0.3608578145503998,
      "learning_rate": 0.00046307122660562805,
      "loss": 3.1348,
      "step": 34000
    },
    {
      "epoch": 9.14965546942291,
      "eval_accuracy": 0.3886457734417542,
      "eval_loss": 3.326436758041382,
      "eval_runtime": 84.0101,
      "eval_samples_per_second": 214.403,
      "eval_steps_per_second": 13.403,
      "step": 34000
    },
    {
      "epoch": 9.16311369509044,
      "grad_norm": 0.3743852376937866,
      "learning_rate": 0.00046286926080517024,
      "loss": 3.126,
      "step": 34050
    },
    {
      "epoch": 9.176571920757967,
      "grad_norm": 0.34315603971481323,
      "learning_rate": 0.0004626672950047125,
      "loss": 3.1372,
      "step": 34100
    },
    {
      "epoch": 9.190030146425496,
      "grad_norm": 0.3354349136352539,
      "learning_rate": 0.0004624653292042547,
      "loss": 3.1435,
      "step": 34150
    },
    {
      "epoch": 9.203488372093023,
      "grad_norm": 0.37850311398506165,
      "learning_rate": 0.00046226336340379694,
      "loss": 3.1351,
      "step": 34200
    },
    {
      "epoch": 9.21694659776055,
      "grad_norm": 0.361669659614563,
      "learning_rate": 0.00046206139760333913,
      "loss": 3.1408,
      "step": 34250
    },
    {
      "epoch": 9.23040482342808,
      "grad_norm": 0.34540659189224243,
      "learning_rate": 0.0004618594318028814,
      "loss": 3.1451,
      "step": 34300
    },
    {
      "epoch": 9.243863049095607,
      "grad_norm": 0.4110398590564728,
      "learning_rate": 0.0004616574660024236,
      "loss": 3.148,
      "step": 34350
    },
    {
      "epoch": 9.257321274763136,
      "grad_norm": 0.34811997413635254,
      "learning_rate": 0.0004614555002019658,
      "loss": 3.1365,
      "step": 34400
    },
    {
      "epoch": 9.270779500430663,
      "grad_norm": 0.3769685924053192,
      "learning_rate": 0.000461253534401508,
      "loss": 3.1491,
      "step": 34450
    },
    {
      "epoch": 9.28423772609819,
      "grad_norm": 0.3319723308086395,
      "learning_rate": 0.00046105156860105017,
      "loss": 3.1463,
      "step": 34500
    },
    {
      "epoch": 9.29769595176572,
      "grad_norm": 0.334638774394989,
      "learning_rate": 0.0004608496028005924,
      "loss": 3.1539,
      "step": 34550
    },
    {
      "epoch": 9.311154177433247,
      "grad_norm": 0.3548935055732727,
      "learning_rate": 0.0004606476370001346,
      "loss": 3.1466,
      "step": 34600
    },
    {
      "epoch": 9.324612403100776,
      "grad_norm": 0.3433622419834137,
      "learning_rate": 0.0004604456711996768,
      "loss": 3.153,
      "step": 34650
    },
    {
      "epoch": 9.338070628768303,
      "grad_norm": 0.36048346757888794,
      "learning_rate": 0.000460243705399219,
      "loss": 3.1502,
      "step": 34700
    },
    {
      "epoch": 9.35152885443583,
      "grad_norm": 0.3432656526565552,
      "learning_rate": 0.0004600417395987612,
      "loss": 3.1535,
      "step": 34750
    },
    {
      "epoch": 9.36498708010336,
      "grad_norm": 0.36674490571022034,
      "learning_rate": 0.00045983977379830345,
      "loss": 3.1586,
      "step": 34800
    },
    {
      "epoch": 9.378445305770887,
      "grad_norm": 0.35690009593963623,
      "learning_rate": 0.00045963780799784565,
      "loss": 3.1596,
      "step": 34850
    },
    {
      "epoch": 9.391903531438416,
      "grad_norm": 0.3510376513004303,
      "learning_rate": 0.00045943584219738785,
      "loss": 3.155,
      "step": 34900
    },
    {
      "epoch": 9.405361757105943,
      "grad_norm": 0.3628822863101959,
      "learning_rate": 0.00045923387639693004,
      "loss": 3.1663,
      "step": 34950
    },
    {
      "epoch": 9.41881998277347,
      "grad_norm": 0.35595691204071045,
      "learning_rate": 0.00045903191059647224,
      "loss": 3.1709,
      "step": 35000
    },
    {
      "epoch": 9.41881998277347,
      "eval_accuracy": 0.38921845101743435,
      "eval_loss": 3.3197622299194336,
      "eval_runtime": 83.5077,
      "eval_samples_per_second": 215.693,
      "eval_steps_per_second": 13.484,
      "step": 35000
    },
    {
      "epoch": 9.432278208441,
      "grad_norm": 0.36002326011657715,
      "learning_rate": 0.0004588299447960145,
      "loss": 3.162,
      "step": 35050
    },
    {
      "epoch": 9.445736434108527,
      "grad_norm": 0.3294300138950348,
      "learning_rate": 0.00045862797899555674,
      "loss": 3.1472,
      "step": 35100
    },
    {
      "epoch": 9.459194659776054,
      "grad_norm": 0.33297887444496155,
      "learning_rate": 0.00045842601319509894,
      "loss": 3.1584,
      "step": 35150
    },
    {
      "epoch": 9.472652885443583,
      "grad_norm": 0.339891642332077,
      "learning_rate": 0.00045822404739464113,
      "loss": 3.1563,
      "step": 35200
    },
    {
      "epoch": 9.48611111111111,
      "grad_norm": 0.3429630696773529,
      "learning_rate": 0.0004580220815941834,
      "loss": 3.1585,
      "step": 35250
    },
    {
      "epoch": 9.49956933677864,
      "grad_norm": 0.35991042852401733,
      "learning_rate": 0.0004578201157937256,
      "loss": 3.1528,
      "step": 35300
    },
    {
      "epoch": 9.513027562446167,
      "grad_norm": 0.34086963534355164,
      "learning_rate": 0.0004576181499932678,
      "loss": 3.1604,
      "step": 35350
    },
    {
      "epoch": 9.526485788113694,
      "grad_norm": 0.3438695967197418,
      "learning_rate": 0.00045741618419280997,
      "loss": 3.1602,
      "step": 35400
    },
    {
      "epoch": 9.539944013781223,
      "grad_norm": 0.33108091354370117,
      "learning_rate": 0.0004572142183923522,
      "loss": 3.1582,
      "step": 35450
    },
    {
      "epoch": 9.55340223944875,
      "grad_norm": 0.3514120280742645,
      "learning_rate": 0.0004570122525918944,
      "loss": 3.1702,
      "step": 35500
    },
    {
      "epoch": 9.56686046511628,
      "grad_norm": 0.3507814407348633,
      "learning_rate": 0.0004568102867914366,
      "loss": 3.1684,
      "step": 35550
    },
    {
      "epoch": 9.580318690783807,
      "grad_norm": 0.34335893392562866,
      "learning_rate": 0.0004566083209909788,
      "loss": 3.1613,
      "step": 35600
    },
    {
      "epoch": 9.593776916451334,
      "grad_norm": 0.36005842685699463,
      "learning_rate": 0.000456406355190521,
      "loss": 3.184,
      "step": 35650
    },
    {
      "epoch": 9.607235142118864,
      "grad_norm": 0.34467431902885437,
      "learning_rate": 0.00045620438939006326,
      "loss": 3.1672,
      "step": 35700
    },
    {
      "epoch": 9.62069336778639,
      "grad_norm": 0.3669510781764984,
      "learning_rate": 0.00045600242358960545,
      "loss": 3.1825,
      "step": 35750
    },
    {
      "epoch": 9.634151593453918,
      "grad_norm": 0.39685648679733276,
      "learning_rate": 0.00045580045778914765,
      "loss": 3.1673,
      "step": 35800
    },
    {
      "epoch": 9.647609819121447,
      "grad_norm": 0.33608683943748474,
      "learning_rate": 0.00045559849198868984,
      "loss": 3.1671,
      "step": 35850
    },
    {
      "epoch": 9.661068044788975,
      "grad_norm": 0.36422276496887207,
      "learning_rate": 0.00045539652618823204,
      "loss": 3.1734,
      "step": 35900
    },
    {
      "epoch": 9.674526270456504,
      "grad_norm": 0.3638077974319458,
      "learning_rate": 0.0004551945603877743,
      "loss": 3.1725,
      "step": 35950
    },
    {
      "epoch": 9.687984496124031,
      "grad_norm": 0.36884841322898865,
      "learning_rate": 0.00045499259458731654,
      "loss": 3.1825,
      "step": 36000
    },
    {
      "epoch": 9.687984496124031,
      "eval_accuracy": 0.38990781531599067,
      "eval_loss": 3.3118937015533447,
      "eval_runtime": 83.2636,
      "eval_samples_per_second": 216.325,
      "eval_steps_per_second": 13.523,
      "step": 36000
    },
    {
      "epoch": 9.701442721791558,
      "grad_norm": 0.34827592968940735,
      "learning_rate": 0.00045479062878685874,
      "loss": 3.1795,
      "step": 36050
    },
    {
      "epoch": 9.714900947459087,
      "grad_norm": 0.36406606435775757,
      "learning_rate": 0.00045458866298640093,
      "loss": 3.173,
      "step": 36100
    },
    {
      "epoch": 9.728359173126615,
      "grad_norm": 0.3397294580936432,
      "learning_rate": 0.0004543866971859432,
      "loss": 3.1715,
      "step": 36150
    },
    {
      "epoch": 9.741817398794144,
      "grad_norm": 0.35161644220352173,
      "learning_rate": 0.0004541847313854854,
      "loss": 3.1744,
      "step": 36200
    },
    {
      "epoch": 9.755275624461671,
      "grad_norm": 0.3627067804336548,
      "learning_rate": 0.0004539827655850276,
      "loss": 3.1744,
      "step": 36250
    },
    {
      "epoch": 9.768733850129198,
      "grad_norm": 0.342001736164093,
      "learning_rate": 0.00045378079978456977,
      "loss": 3.1737,
      "step": 36300
    },
    {
      "epoch": 9.782192075796727,
      "grad_norm": 0.34132522344589233,
      "learning_rate": 0.00045357883398411197,
      "loss": 3.1688,
      "step": 36350
    },
    {
      "epoch": 9.795650301464255,
      "grad_norm": 0.3432362377643585,
      "learning_rate": 0.0004533768681836542,
      "loss": 3.18,
      "step": 36400
    },
    {
      "epoch": 9.809108527131784,
      "grad_norm": 0.32423079013824463,
      "learning_rate": 0.0004531749023831964,
      "loss": 3.1771,
      "step": 36450
    },
    {
      "epoch": 9.822566752799311,
      "grad_norm": 0.3321049213409424,
      "learning_rate": 0.0004529729365827386,
      "loss": 3.1845,
      "step": 36500
    },
    {
      "epoch": 9.836024978466838,
      "grad_norm": 0.3416637182235718,
      "learning_rate": 0.0004527709707822808,
      "loss": 3.1786,
      "step": 36550
    },
    {
      "epoch": 9.849483204134367,
      "grad_norm": 0.3366742730140686,
      "learning_rate": 0.000452569004981823,
      "loss": 3.1658,
      "step": 36600
    },
    {
      "epoch": 9.862941429801895,
      "grad_norm": 0.35643571615219116,
      "learning_rate": 0.00045236703918136525,
      "loss": 3.1828,
      "step": 36650
    },
    {
      "epoch": 9.876399655469424,
      "grad_norm": 0.3542114198207855,
      "learning_rate": 0.00045216507338090745,
      "loss": 3.1869,
      "step": 36700
    },
    {
      "epoch": 9.889857881136951,
      "grad_norm": 0.3500431180000305,
      "learning_rate": 0.00045196310758044964,
      "loss": 3.1756,
      "step": 36750
    },
    {
      "epoch": 9.903316106804478,
      "grad_norm": 0.3726409077644348,
      "learning_rate": 0.00045176114177999184,
      "loss": 3.1752,
      "step": 36800
    },
    {
      "epoch": 9.916774332472007,
      "grad_norm": 0.36989736557006836,
      "learning_rate": 0.00045155917597953404,
      "loss": 3.1844,
      "step": 36850
    },
    {
      "epoch": 9.930232558139535,
      "grad_norm": 0.3382627069950104,
      "learning_rate": 0.00045135721017907634,
      "loss": 3.1749,
      "step": 36900
    },
    {
      "epoch": 9.943690783807062,
      "grad_norm": 0.3394027650356293,
      "learning_rate": 0.00045115524437861854,
      "loss": 3.1767,
      "step": 36950
    },
    {
      "epoch": 9.957149009474591,
      "grad_norm": 0.37442541122436523,
      "learning_rate": 0.00045095327857816073,
      "loss": 3.1852,
      "step": 37000
    },
    {
      "epoch": 9.957149009474591,
      "eval_accuracy": 0.39029503270922233,
      "eval_loss": 3.3084702491760254,
      "eval_runtime": 83.1841,
      "eval_samples_per_second": 216.532,
      "eval_steps_per_second": 13.536,
      "step": 37000
    },
    {
      "epoch": 9.970607235142118,
      "grad_norm": 0.3456587493419647,
      "learning_rate": 0.00045075131277770293,
      "loss": 3.1771,
      "step": 37050
    },
    {
      "epoch": 9.984065460809648,
      "grad_norm": 0.36226093769073486,
      "learning_rate": 0.0004505493469772452,
      "loss": 3.1948,
      "step": 37100
    },
    {
      "epoch": 9.997523686477175,
      "grad_norm": 0.34133097529411316,
      "learning_rate": 0.0004503473811767874,
      "loss": 3.1837,
      "step": 37150
    },
    {
      "epoch": 10.010766580534023,
      "grad_norm": 0.3589644134044647,
      "learning_rate": 0.00045014541537632957,
      "loss": 3.0998,
      "step": 37200
    },
    {
      "epoch": 10.02422480620155,
      "grad_norm": 0.38088276982307434,
      "learning_rate": 0.00044994344957587177,
      "loss": 3.0851,
      "step": 37250
    },
    {
      "epoch": 10.037683031869078,
      "grad_norm": 0.35852065682411194,
      "learning_rate": 0.00044974148377541396,
      "loss": 3.0978,
      "step": 37300
    },
    {
      "epoch": 10.051141257536607,
      "grad_norm": 0.3353229761123657,
      "learning_rate": 0.0004495395179749562,
      "loss": 3.0815,
      "step": 37350
    },
    {
      "epoch": 10.064599483204134,
      "grad_norm": 0.3414810001850128,
      "learning_rate": 0.0004493375521744984,
      "loss": 3.0942,
      "step": 37400
    },
    {
      "epoch": 10.078057708871663,
      "grad_norm": 0.34384095668792725,
      "learning_rate": 0.0004491355863740406,
      "loss": 3.1024,
      "step": 37450
    },
    {
      "epoch": 10.09151593453919,
      "grad_norm": 0.3510953187942505,
      "learning_rate": 0.0004489336205735828,
      "loss": 3.0944,
      "step": 37500
    },
    {
      "epoch": 10.104974160206718,
      "grad_norm": 0.3401368260383606,
      "learning_rate": 0.00044873165477312505,
      "loss": 3.0962,
      "step": 37550
    },
    {
      "epoch": 10.118432385874247,
      "grad_norm": 0.34181979298591614,
      "learning_rate": 0.00044852968897266725,
      "loss": 3.1068,
      "step": 37600
    },
    {
      "epoch": 10.131890611541774,
      "grad_norm": 0.3376041054725647,
      "learning_rate": 0.00044832772317220945,
      "loss": 3.1013,
      "step": 37650
    },
    {
      "epoch": 10.145348837209303,
      "grad_norm": 0.3769124746322632,
      "learning_rate": 0.00044812575737175164,
      "loss": 3.1242,
      "step": 37700
    },
    {
      "epoch": 10.15880706287683,
      "grad_norm": 0.3543618619441986,
      "learning_rate": 0.00044792379157129384,
      "loss": 3.1099,
      "step": 37750
    },
    {
      "epoch": 10.172265288544358,
      "grad_norm": 0.37354061007499695,
      "learning_rate": 0.00044772182577083614,
      "loss": 3.1135,
      "step": 37800
    },
    {
      "epoch": 10.185723514211887,
      "grad_norm": 0.3754305839538574,
      "learning_rate": 0.00044751985997037834,
      "loss": 3.1076,
      "step": 37850
    },
    {
      "epoch": 10.199181739879414,
      "grad_norm": 0.387029767036438,
      "learning_rate": 0.00044731789416992053,
      "loss": 3.1208,
      "step": 37900
    },
    {
      "epoch": 10.212639965546943,
      "grad_norm": 0.34069687128067017,
      "learning_rate": 0.00044711592836946273,
      "loss": 3.1255,
      "step": 37950
    },
    {
      "epoch": 10.22609819121447,
      "grad_norm": 0.33347445726394653,
      "learning_rate": 0.000446913962569005,
      "loss": 3.1129,
      "step": 38000
    },
    {
      "epoch": 10.22609819121447,
      "eval_accuracy": 0.3900587257983697,
      "eval_loss": 3.3169922828674316,
      "eval_runtime": 83.2755,
      "eval_samples_per_second": 216.294,
      "eval_steps_per_second": 13.521,
      "step": 38000
    },
    {
      "epoch": 10.239556416881998,
      "grad_norm": 0.3639187812805176,
      "learning_rate": 0.0004467119967685472,
      "loss": 3.1134,
      "step": 38050
    },
    {
      "epoch": 10.253014642549527,
      "grad_norm": 0.3886873722076416,
      "learning_rate": 0.00044651003096808937,
      "loss": 3.1089,
      "step": 38100
    },
    {
      "epoch": 10.266472868217054,
      "grad_norm": 0.3527437448501587,
      "learning_rate": 0.00044630806516763157,
      "loss": 3.1274,
      "step": 38150
    },
    {
      "epoch": 10.279931093884581,
      "grad_norm": 0.37784016132354736,
      "learning_rate": 0.00044610609936717376,
      "loss": 3.1311,
      "step": 38200
    },
    {
      "epoch": 10.29338931955211,
      "grad_norm": 0.3572711944580078,
      "learning_rate": 0.000445904133566716,
      "loss": 3.1176,
      "step": 38250
    },
    {
      "epoch": 10.306847545219638,
      "grad_norm": 0.3570372462272644,
      "learning_rate": 0.0004457021677662582,
      "loss": 3.1291,
      "step": 38300
    },
    {
      "epoch": 10.320305770887167,
      "grad_norm": 0.36011946201324463,
      "learning_rate": 0.0004455002019658004,
      "loss": 3.128,
      "step": 38350
    },
    {
      "epoch": 10.333763996554694,
      "grad_norm": 0.3254188299179077,
      "learning_rate": 0.0004452982361653426,
      "loss": 3.1255,
      "step": 38400
    },
    {
      "epoch": 10.347222222222221,
      "grad_norm": 0.40706273913383484,
      "learning_rate": 0.0004450962703648848,
      "loss": 3.1329,
      "step": 38450
    },
    {
      "epoch": 10.36068044788975,
      "grad_norm": 0.3689863979816437,
      "learning_rate": 0.00044489430456442705,
      "loss": 3.1328,
      "step": 38500
    },
    {
      "epoch": 10.374138673557278,
      "grad_norm": 0.36276525259017944,
      "learning_rate": 0.00044469233876396925,
      "loss": 3.1378,
      "step": 38550
    },
    {
      "epoch": 10.387596899224807,
      "grad_norm": 0.3926582336425781,
      "learning_rate": 0.00044449037296351144,
      "loss": 3.1307,
      "step": 38600
    },
    {
      "epoch": 10.401055124892334,
      "grad_norm": 0.3865697979927063,
      "learning_rate": 0.0004442884071630537,
      "loss": 3.1317,
      "step": 38650
    },
    {
      "epoch": 10.414513350559861,
      "grad_norm": 0.358347088098526,
      "learning_rate": 0.00044408644136259594,
      "loss": 3.1362,
      "step": 38700
    },
    {
      "epoch": 10.42797157622739,
      "grad_norm": 0.3477441668510437,
      "learning_rate": 0.00044388447556213814,
      "loss": 3.1324,
      "step": 38750
    },
    {
      "epoch": 10.441429801894918,
      "grad_norm": 0.3335910737514496,
      "learning_rate": 0.00044368250976168033,
      "loss": 3.142,
      "step": 38800
    },
    {
      "epoch": 10.454888027562447,
      "grad_norm": 0.3818318843841553,
      "learning_rate": 0.00044348054396122253,
      "loss": 3.15,
      "step": 38850
    },
    {
      "epoch": 10.468346253229974,
      "grad_norm": 0.33925795555114746,
      "learning_rate": 0.0004432785781607647,
      "loss": 3.1392,
      "step": 38900
    },
    {
      "epoch": 10.481804478897502,
      "grad_norm": 0.3682875633239746,
      "learning_rate": 0.000443076612360307,
      "loss": 3.144,
      "step": 38950
    },
    {
      "epoch": 10.49526270456503,
      "grad_norm": 0.3501853346824646,
      "learning_rate": 0.0004428746465598492,
      "loss": 3.1422,
      "step": 39000
    },
    {
      "epoch": 10.49526270456503,
      "eval_accuracy": 0.390544051302176,
      "eval_loss": 3.3120977878570557,
      "eval_runtime": 83.2216,
      "eval_samples_per_second": 216.434,
      "eval_steps_per_second": 13.53,
      "step": 39000
    },
    {
      "epoch": 10.508720930232558,
      "grad_norm": 0.3623600900173187,
      "learning_rate": 0.00044267268075939137,
      "loss": 3.1456,
      "step": 39050
    },
    {
      "epoch": 10.522179155900087,
      "grad_norm": 0.3507665991783142,
      "learning_rate": 0.00044247071495893357,
      "loss": 3.1465,
      "step": 39100
    },
    {
      "epoch": 10.535637381567614,
      "grad_norm": 0.33959007263183594,
      "learning_rate": 0.00044226874915847576,
      "loss": 3.1376,
      "step": 39150
    },
    {
      "epoch": 10.549095607235142,
      "grad_norm": 0.3561212718486786,
      "learning_rate": 0.000442066783358018,
      "loss": 3.1553,
      "step": 39200
    },
    {
      "epoch": 10.56255383290267,
      "grad_norm": 0.38178831338882446,
      "learning_rate": 0.0004418648175575602,
      "loss": 3.153,
      "step": 39250
    },
    {
      "epoch": 10.576012058570198,
      "grad_norm": 0.35409092903137207,
      "learning_rate": 0.0004416628517571024,
      "loss": 3.144,
      "step": 39300
    },
    {
      "epoch": 10.589470284237725,
      "grad_norm": 0.36173319816589355,
      "learning_rate": 0.0004414608859566446,
      "loss": 3.1461,
      "step": 39350
    },
    {
      "epoch": 10.602928509905254,
      "grad_norm": 0.3296486735343933,
      "learning_rate": 0.00044125892015618685,
      "loss": 3.1455,
      "step": 39400
    },
    {
      "epoch": 10.616386735572782,
      "grad_norm": 0.3699939548969269,
      "learning_rate": 0.00044105695435572905,
      "loss": 3.1415,
      "step": 39450
    },
    {
      "epoch": 10.62984496124031,
      "grad_norm": 0.34273818135261536,
      "learning_rate": 0.00044085498855527124,
      "loss": 3.154,
      "step": 39500
    },
    {
      "epoch": 10.643303186907838,
      "grad_norm": 0.36061105132102966,
      "learning_rate": 0.0004406530227548135,
      "loss": 3.1566,
      "step": 39550
    },
    {
      "epoch": 10.656761412575365,
      "grad_norm": 0.33657217025756836,
      "learning_rate": 0.00044045105695435574,
      "loss": 3.1596,
      "step": 39600
    },
    {
      "epoch": 10.670219638242894,
      "grad_norm": 0.3700537085533142,
      "learning_rate": 0.00044024909115389794,
      "loss": 3.1563,
      "step": 39650
    },
    {
      "epoch": 10.683677863910422,
      "grad_norm": 0.35576534271240234,
      "learning_rate": 0.00044004712535344014,
      "loss": 3.1522,
      "step": 39700
    },
    {
      "epoch": 10.69713608957795,
      "grad_norm": 0.3644106686115265,
      "learning_rate": 0.00043984515955298233,
      "loss": 3.156,
      "step": 39750
    },
    {
      "epoch": 10.710594315245478,
      "grad_norm": 0.3310992121696472,
      "learning_rate": 0.00043964319375252453,
      "loss": 3.1531,
      "step": 39800
    },
    {
      "epoch": 10.724052540913005,
      "grad_norm": 0.35329774022102356,
      "learning_rate": 0.0004394412279520668,
      "loss": 3.1495,
      "step": 39850
    },
    {
      "epoch": 10.737510766580534,
      "grad_norm": 0.3860771059989929,
      "learning_rate": 0.000439239262151609,
      "loss": 3.1532,
      "step": 39900
    },
    {
      "epoch": 10.750968992248062,
      "grad_norm": 0.36905908584594727,
      "learning_rate": 0.00043903729635115117,
      "loss": 3.1744,
      "step": 39950
    },
    {
      "epoch": 10.764427217915589,
      "grad_norm": 0.35525012016296387,
      "learning_rate": 0.00043883533055069337,
      "loss": 3.1647,
      "step": 40000
    },
    {
      "epoch": 10.764427217915589,
      "eval_accuracy": 0.3909727717942333,
      "eval_loss": 3.3064162731170654,
      "eval_runtime": 83.3228,
      "eval_samples_per_second": 216.171,
      "eval_steps_per_second": 13.514,
      "step": 40000
    },
    {
      "epoch": 10.777885443583118,
      "grad_norm": 0.3454327881336212,
      "learning_rate": 0.00043863336475023556,
      "loss": 3.1446,
      "step": 40050
    },
    {
      "epoch": 10.791343669250645,
      "grad_norm": 0.3812251687049866,
      "learning_rate": 0.0004384313989497778,
      "loss": 3.1556,
      "step": 40100
    },
    {
      "epoch": 10.804801894918175,
      "grad_norm": 0.37816113233566284,
      "learning_rate": 0.00043822943314932,
      "loss": 3.1512,
      "step": 40150
    },
    {
      "epoch": 10.818260120585702,
      "grad_norm": 0.3591747283935547,
      "learning_rate": 0.0004380274673488622,
      "loss": 3.1608,
      "step": 40200
    },
    {
      "epoch": 10.83171834625323,
      "grad_norm": 0.3561844527721405,
      "learning_rate": 0.0004378255015484044,
      "loss": 3.1572,
      "step": 40250
    },
    {
      "epoch": 10.845176571920758,
      "grad_norm": 0.3455813229084015,
      "learning_rate": 0.0004376235357479466,
      "loss": 3.1681,
      "step": 40300
    },
    {
      "epoch": 10.858634797588286,
      "grad_norm": 0.32350650429725647,
      "learning_rate": 0.00043742156994748885,
      "loss": 3.1613,
      "step": 40350
    },
    {
      "epoch": 10.872093023255815,
      "grad_norm": 0.331007182598114,
      "learning_rate": 0.00043721960414703104,
      "loss": 3.1567,
      "step": 40400
    },
    {
      "epoch": 10.885551248923342,
      "grad_norm": 0.34472721815109253,
      "learning_rate": 0.0004370176383465733,
      "loss": 3.157,
      "step": 40450
    },
    {
      "epoch": 10.89900947459087,
      "grad_norm": 0.35679876804351807,
      "learning_rate": 0.0004368156725461155,
      "loss": 3.1592,
      "step": 40500
    },
    {
      "epoch": 10.912467700258398,
      "grad_norm": 0.3619687855243683,
      "learning_rate": 0.00043661370674565774,
      "loss": 3.1653,
      "step": 40550
    },
    {
      "epoch": 10.925925925925926,
      "grad_norm": 0.3671156167984009,
      "learning_rate": 0.00043641174094519994,
      "loss": 3.1635,
      "step": 40600
    },
    {
      "epoch": 10.939384151593455,
      "grad_norm": 0.3705694079399109,
      "learning_rate": 0.00043620977514474213,
      "loss": 3.1688,
      "step": 40650
    },
    {
      "epoch": 10.952842377260982,
      "grad_norm": 0.33977562189102173,
      "learning_rate": 0.00043600780934428433,
      "loss": 3.1576,
      "step": 40700
    },
    {
      "epoch": 10.96630060292851,
      "grad_norm": 0.3485282361507416,
      "learning_rate": 0.0004358058435438265,
      "loss": 3.1626,
      "step": 40750
    },
    {
      "epoch": 10.979758828596038,
      "grad_norm": 0.3598802983760834,
      "learning_rate": 0.0004356038777433688,
      "loss": 3.1606,
      "step": 40800
    },
    {
      "epoch": 10.993217054263566,
      "grad_norm": 0.354419469833374,
      "learning_rate": 0.00043540191194291097,
      "loss": 3.161,
      "step": 40850
    },
    {
      "epoch": 11.006459948320414,
      "grad_norm": 0.3764330744743347,
      "learning_rate": 0.00043519994614245317,
      "loss": 3.1103,
      "step": 40900
    },
    {
      "epoch": 11.019918173987941,
      "grad_norm": 0.350726455450058,
      "learning_rate": 0.00043499798034199536,
      "loss": 3.0629,
      "step": 40950
    },
    {
      "epoch": 11.03337639965547,
      "grad_norm": 0.36444082856178284,
      "learning_rate": 0.00043479601454153756,
      "loss": 3.0573,
      "step": 41000
    },
    {
      "epoch": 11.03337639965547,
      "eval_accuracy": 0.39095093377626483,
      "eval_loss": 3.311933994293213,
      "eval_runtime": 83.3115,
      "eval_samples_per_second": 216.201,
      "eval_steps_per_second": 13.516,
      "step": 41000
    },
    {
      "epoch": 11.046834625322997,
      "grad_norm": 0.3607778251171112,
      "learning_rate": 0.0004345940487410798,
      "loss": 3.0807,
      "step": 41050
    },
    {
      "epoch": 11.060292850990525,
      "grad_norm": 0.3665819764137268,
      "learning_rate": 0.000434392082940622,
      "loss": 3.086,
      "step": 41100
    },
    {
      "epoch": 11.073751076658054,
      "grad_norm": 0.34571734070777893,
      "learning_rate": 0.0004341901171401642,
      "loss": 3.0815,
      "step": 41150
    },
    {
      "epoch": 11.087209302325581,
      "grad_norm": 0.3720223307609558,
      "learning_rate": 0.0004339881513397064,
      "loss": 3.0645,
      "step": 41200
    },
    {
      "epoch": 11.10066752799311,
      "grad_norm": 0.3657797873020172,
      "learning_rate": 0.0004337861855392486,
      "loss": 3.0753,
      "step": 41250
    },
    {
      "epoch": 11.114125753660637,
      "grad_norm": 0.3682266175746918,
      "learning_rate": 0.00043358421973879084,
      "loss": 3.0825,
      "step": 41300
    },
    {
      "epoch": 11.127583979328165,
      "grad_norm": 0.3343811333179474,
      "learning_rate": 0.0004333822539383331,
      "loss": 3.0915,
      "step": 41350
    },
    {
      "epoch": 11.141042204995694,
      "grad_norm": 0.3661707937717438,
      "learning_rate": 0.0004331802881378753,
      "loss": 3.083,
      "step": 41400
    },
    {
      "epoch": 11.154500430663221,
      "grad_norm": 0.3465404510498047,
      "learning_rate": 0.00043297832233741754,
      "loss": 3.0877,
      "step": 41450
    },
    {
      "epoch": 11.16795865633075,
      "grad_norm": 0.39990806579589844,
      "learning_rate": 0.00043277635653695974,
      "loss": 3.0979,
      "step": 41500
    },
    {
      "epoch": 11.181416881998278,
      "grad_norm": 0.37263742089271545,
      "learning_rate": 0.00043257439073650193,
      "loss": 3.0996,
      "step": 41550
    },
    {
      "epoch": 11.194875107665805,
      "grad_norm": 0.35715845227241516,
      "learning_rate": 0.00043237242493604413,
      "loss": 3.0926,
      "step": 41600
    },
    {
      "epoch": 11.208333333333334,
      "grad_norm": 0.3820556402206421,
      "learning_rate": 0.0004321704591355863,
      "loss": 3.0983,
      "step": 41650
    },
    {
      "epoch": 11.221791559000861,
      "grad_norm": 0.35530298948287964,
      "learning_rate": 0.0004319684933351286,
      "loss": 3.0958,
      "step": 41700
    },
    {
      "epoch": 11.235249784668389,
      "grad_norm": 0.3512575626373291,
      "learning_rate": 0.00043176652753467077,
      "loss": 3.0999,
      "step": 41750
    },
    {
      "epoch": 11.248708010335918,
      "grad_norm": 0.34998154640197754,
      "learning_rate": 0.00043156456173421297,
      "loss": 3.1018,
      "step": 41800
    },
    {
      "epoch": 11.262166236003445,
      "grad_norm": 0.36305251717567444,
      "learning_rate": 0.00043136259593375516,
      "loss": 3.1103,
      "step": 41850
    },
    {
      "epoch": 11.275624461670974,
      "grad_norm": 0.3878055214881897,
      "learning_rate": 0.00043116063013329736,
      "loss": 3.1015,
      "step": 41900
    },
    {
      "epoch": 11.289082687338501,
      "grad_norm": 0.3685509264469147,
      "learning_rate": 0.0004309586643328396,
      "loss": 3.1088,
      "step": 41950
    },
    {
      "epoch": 11.302540913006029,
      "grad_norm": 0.37419840693473816,
      "learning_rate": 0.0004307566985323818,
      "loss": 3.1104,
      "step": 42000
    },
    {
      "epoch": 11.302540913006029,
      "eval_accuracy": 0.39106740320542993,
      "eval_loss": 3.3091750144958496,
      "eval_runtime": 83.0795,
      "eval_samples_per_second": 216.804,
      "eval_steps_per_second": 13.553,
      "step": 42000
    },
    {
      "epoch": 11.315999138673558,
      "grad_norm": 0.3501582443714142,
      "learning_rate": 0.000430554732731924,
      "loss": 3.1132,
      "step": 42050
    },
    {
      "epoch": 11.329457364341085,
      "grad_norm": 0.3571588695049286,
      "learning_rate": 0.0004303527669314662,
      "loss": 3.1121,
      "step": 42100
    },
    {
      "epoch": 11.342915590008614,
      "grad_norm": 0.35604408383369446,
      "learning_rate": 0.0004301508011310084,
      "loss": 3.1148,
      "step": 42150
    },
    {
      "epoch": 11.356373815676141,
      "grad_norm": 0.37641966342926025,
      "learning_rate": 0.00042994883533055065,
      "loss": 3.1037,
      "step": 42200
    },
    {
      "epoch": 11.369832041343669,
      "grad_norm": 0.3817889094352722,
      "learning_rate": 0.0004297468695300929,
      "loss": 3.1095,
      "step": 42250
    },
    {
      "epoch": 11.383290267011198,
      "grad_norm": 0.36300888657569885,
      "learning_rate": 0.0004295449037296351,
      "loss": 3.1139,
      "step": 42300
    },
    {
      "epoch": 11.396748492678725,
      "grad_norm": 0.37002626061439514,
      "learning_rate": 0.0004293429379291773,
      "loss": 3.1168,
      "step": 42350
    },
    {
      "epoch": 11.410206718346252,
      "grad_norm": 0.36631953716278076,
      "learning_rate": 0.00042914097212871954,
      "loss": 3.1098,
      "step": 42400
    },
    {
      "epoch": 11.423664944013781,
      "grad_norm": 0.34359294176101685,
      "learning_rate": 0.00042893900632826173,
      "loss": 3.1227,
      "step": 42450
    },
    {
      "epoch": 11.437123169681309,
      "grad_norm": 0.3452107608318329,
      "learning_rate": 0.00042873704052780393,
      "loss": 3.1276,
      "step": 42500
    },
    {
      "epoch": 11.450581395348838,
      "grad_norm": 0.34046146273612976,
      "learning_rate": 0.0004285350747273461,
      "loss": 3.1289,
      "step": 42550
    },
    {
      "epoch": 11.464039621016365,
      "grad_norm": 0.40337055921554565,
      "learning_rate": 0.0004283331089268883,
      "loss": 3.1135,
      "step": 42600
    },
    {
      "epoch": 11.477497846683892,
      "grad_norm": 0.39826324582099915,
      "learning_rate": 0.00042813114312643057,
      "loss": 3.1159,
      "step": 42650
    },
    {
      "epoch": 11.490956072351421,
      "grad_norm": 0.36773282289505005,
      "learning_rate": 0.00042792917732597277,
      "loss": 3.1186,
      "step": 42700
    },
    {
      "epoch": 11.504414298018949,
      "grad_norm": 0.376511812210083,
      "learning_rate": 0.00042772721152551496,
      "loss": 3.1233,
      "step": 42750
    },
    {
      "epoch": 11.517872523686478,
      "grad_norm": 0.3432180881500244,
      "learning_rate": 0.00042752524572505716,
      "loss": 3.1248,
      "step": 42800
    },
    {
      "epoch": 11.531330749354005,
      "grad_norm": 0.35417458415031433,
      "learning_rate": 0.00042732327992459936,
      "loss": 3.1205,
      "step": 42850
    },
    {
      "epoch": 11.544788975021532,
      "grad_norm": 0.3793041408061981,
      "learning_rate": 0.0004271213141241416,
      "loss": 3.1321,
      "step": 42900
    },
    {
      "epoch": 11.558247200689062,
      "grad_norm": 0.3891814649105072,
      "learning_rate": 0.0004269193483236838,
      "loss": 3.1265,
      "step": 42950
    },
    {
      "epoch": 11.571705426356589,
      "grad_norm": 0.3784371018409729,
      "learning_rate": 0.000426717382523226,
      "loss": 3.1398,
      "step": 43000
    },
    {
      "epoch": 11.571705426356589,
      "eval_accuracy": 0.3913007766511823,
      "eval_loss": 3.304755210876465,
      "eval_runtime": 83.3052,
      "eval_samples_per_second": 216.217,
      "eval_steps_per_second": 13.517,
      "step": 43000
    },
    {
      "epoch": 11.585163652024118,
      "grad_norm": 0.38434332609176636,
      "learning_rate": 0.0004265154167227682,
      "loss": 3.1301,
      "step": 43050
    },
    {
      "epoch": 11.598621877691645,
      "grad_norm": 0.3575095534324646,
      "learning_rate": 0.0004263134509223105,
      "loss": 3.1369,
      "step": 43100
    },
    {
      "epoch": 11.612080103359173,
      "grad_norm": 0.36131447553634644,
      "learning_rate": 0.0004261114851218527,
      "loss": 3.1245,
      "step": 43150
    },
    {
      "epoch": 11.625538329026702,
      "grad_norm": 0.41072115302085876,
      "learning_rate": 0.0004259095193213949,
      "loss": 3.13,
      "step": 43200
    },
    {
      "epoch": 11.638996554694229,
      "grad_norm": 0.349194198846817,
      "learning_rate": 0.0004257075535209371,
      "loss": 3.1322,
      "step": 43250
    },
    {
      "epoch": 11.652454780361758,
      "grad_norm": 0.38299086689949036,
      "learning_rate": 0.00042550558772047934,
      "loss": 3.1349,
      "step": 43300
    },
    {
      "epoch": 11.665913006029285,
      "grad_norm": 0.3737049400806427,
      "learning_rate": 0.00042530362192002153,
      "loss": 3.1413,
      "step": 43350
    },
    {
      "epoch": 11.679371231696813,
      "grad_norm": 0.3575431704521179,
      "learning_rate": 0.00042510165611956373,
      "loss": 3.14,
      "step": 43400
    },
    {
      "epoch": 11.692829457364342,
      "grad_norm": 0.3510909676551819,
      "learning_rate": 0.0004248996903191059,
      "loss": 3.1312,
      "step": 43450
    },
    {
      "epoch": 11.706287683031869,
      "grad_norm": 0.36683255434036255,
      "learning_rate": 0.0004246977245186481,
      "loss": 3.1301,
      "step": 43500
    },
    {
      "epoch": 11.719745908699396,
      "grad_norm": 0.3639971613883972,
      "learning_rate": 0.0004244957587181904,
      "loss": 3.1429,
      "step": 43550
    },
    {
      "epoch": 11.733204134366925,
      "grad_norm": 0.34593474864959717,
      "learning_rate": 0.00042429379291773257,
      "loss": 3.132,
      "step": 43600
    },
    {
      "epoch": 11.746662360034453,
      "grad_norm": 0.37789031863212585,
      "learning_rate": 0.00042409182711727477,
      "loss": 3.1334,
      "step": 43650
    },
    {
      "epoch": 11.760120585701982,
      "grad_norm": 0.3411363959312439,
      "learning_rate": 0.00042388986131681696,
      "loss": 3.1378,
      "step": 43700
    },
    {
      "epoch": 11.773578811369509,
      "grad_norm": 0.35004791617393494,
      "learning_rate": 0.00042368789551635916,
      "loss": 3.1346,
      "step": 43750
    },
    {
      "epoch": 11.787037037037036,
      "grad_norm": 0.37684744596481323,
      "learning_rate": 0.0004234859297159014,
      "loss": 3.1419,
      "step": 43800
    },
    {
      "epoch": 11.800495262704565,
      "grad_norm": 0.37968793511390686,
      "learning_rate": 0.0004232839639154436,
      "loss": 3.137,
      "step": 43850
    },
    {
      "epoch": 11.813953488372093,
      "grad_norm": 0.36071285605430603,
      "learning_rate": 0.0004230819981149858,
      "loss": 3.1438,
      "step": 43900
    },
    {
      "epoch": 11.827411714039622,
      "grad_norm": 0.3493013381958008,
      "learning_rate": 0.000422880032314528,
      "loss": 3.1298,
      "step": 43950
    },
    {
      "epoch": 11.840869939707149,
      "grad_norm": 0.3998880982398987,
      "learning_rate": 0.0004226780665140703,
      "loss": 3.1452,
      "step": 44000
    },
    {
      "epoch": 11.840869939707149,
      "eval_accuracy": 0.39200632933121776,
      "eval_loss": 3.2981667518615723,
      "eval_runtime": 83.3539,
      "eval_samples_per_second": 216.091,
      "eval_steps_per_second": 13.509,
      "step": 44000
    },
    {
      "epoch": 11.854328165374676,
      "grad_norm": 0.3362621068954468,
      "learning_rate": 0.0004224761007136125,
      "loss": 3.1378,
      "step": 44050
    },
    {
      "epoch": 11.867786391042205,
      "grad_norm": 0.3660929799079895,
      "learning_rate": 0.0004222741349131547,
      "loss": 3.1507,
      "step": 44100
    },
    {
      "epoch": 11.881244616709733,
      "grad_norm": 0.374482661485672,
      "learning_rate": 0.0004220721691126969,
      "loss": 3.1418,
      "step": 44150
    },
    {
      "epoch": 11.89470284237726,
      "grad_norm": 0.3572467565536499,
      "learning_rate": 0.0004218702033122391,
      "loss": 3.1428,
      "step": 44200
    },
    {
      "epoch": 11.90816106804479,
      "grad_norm": 0.35921573638916016,
      "learning_rate": 0.00042166823751178134,
      "loss": 3.1455,
      "step": 44250
    },
    {
      "epoch": 11.921619293712316,
      "grad_norm": 0.3823048770427704,
      "learning_rate": 0.00042146627171132353,
      "loss": 3.1385,
      "step": 44300
    },
    {
      "epoch": 11.935077519379846,
      "grad_norm": 0.3629566431045532,
      "learning_rate": 0.00042126430591086573,
      "loss": 3.1483,
      "step": 44350
    },
    {
      "epoch": 11.948535745047373,
      "grad_norm": 0.3452162444591522,
      "learning_rate": 0.0004210623401104079,
      "loss": 3.1416,
      "step": 44400
    },
    {
      "epoch": 11.9619939707149,
      "grad_norm": 0.3631886839866638,
      "learning_rate": 0.0004208603743099501,
      "loss": 3.1482,
      "step": 44450
    },
    {
      "epoch": 11.97545219638243,
      "grad_norm": 0.3417701721191406,
      "learning_rate": 0.00042065840850949237,
      "loss": 3.1457,
      "step": 44500
    },
    {
      "epoch": 11.988910422049956,
      "grad_norm": 0.35143524408340454,
      "learning_rate": 0.00042045644270903457,
      "loss": 3.1419,
      "step": 44550
    },
    {
      "epoch": 12.002153316106805,
      "grad_norm": 0.37438926100730896,
      "learning_rate": 0.00042025447690857676,
      "loss": 3.1208,
      "step": 44600
    },
    {
      "epoch": 12.015611541774332,
      "grad_norm": 0.34138888120651245,
      "learning_rate": 0.00042005251110811896,
      "loss": 3.0635,
      "step": 44650
    },
    {
      "epoch": 12.029069767441861,
      "grad_norm": 0.3567388355731964,
      "learning_rate": 0.00041985054530766115,
      "loss": 3.0451,
      "step": 44700
    },
    {
      "epoch": 12.042527993109388,
      "grad_norm": 0.3751562237739563,
      "learning_rate": 0.0004196485795072034,
      "loss": 3.0526,
      "step": 44750
    },
    {
      "epoch": 12.055986218776916,
      "grad_norm": 0.3859095275402069,
      "learning_rate": 0.0004194466137067456,
      "loss": 3.0532,
      "step": 44800
    },
    {
      "epoch": 12.069444444444445,
      "grad_norm": 0.3581213057041168,
      "learning_rate": 0.0004192446479062878,
      "loss": 3.0709,
      "step": 44850
    },
    {
      "epoch": 12.082902670111972,
      "grad_norm": 0.373753160238266,
      "learning_rate": 0.00041904268210583005,
      "loss": 3.0633,
      "step": 44900
    },
    {
      "epoch": 12.096360895779501,
      "grad_norm": 0.3698385953903198,
      "learning_rate": 0.0004188407163053723,
      "loss": 3.061,
      "step": 44950
    },
    {
      "epoch": 12.109819121447028,
      "grad_norm": 0.3444227874279022,
      "learning_rate": 0.0004186387505049145,
      "loss": 3.0687,
      "step": 45000
    },
    {
      "epoch": 12.109819121447028,
      "eval_accuracy": 0.39163551761317633,
      "eval_loss": 3.309868574142456,
      "eval_runtime": 83.3252,
      "eval_samples_per_second": 216.165,
      "eval_steps_per_second": 13.513,
      "step": 45000
    },
    {
      "epoch": 12.123277347114556,
      "grad_norm": 0.38895121216773987,
      "learning_rate": 0.0004184367847044567,
      "loss": 3.0686,
      "step": 45050
    },
    {
      "epoch": 12.136735572782085,
      "grad_norm": 0.36655595898628235,
      "learning_rate": 0.0004182348189039989,
      "loss": 3.0702,
      "step": 45100
    },
    {
      "epoch": 12.150193798449612,
      "grad_norm": 0.3856962323188782,
      "learning_rate": 0.00041803285310354114,
      "loss": 3.0804,
      "step": 45150
    },
    {
      "epoch": 12.163652024117141,
      "grad_norm": 0.3753387928009033,
      "learning_rate": 0.00041783088730308333,
      "loss": 3.0739,
      "step": 45200
    },
    {
      "epoch": 12.177110249784668,
      "grad_norm": 0.3965890407562256,
      "learning_rate": 0.00041762892150262553,
      "loss": 3.0779,
      "step": 45250
    },
    {
      "epoch": 12.190568475452196,
      "grad_norm": 0.3758774399757385,
      "learning_rate": 0.0004174269557021677,
      "loss": 3.0698,
      "step": 45300
    },
    {
      "epoch": 12.204026701119725,
      "grad_norm": 0.3307397961616516,
      "learning_rate": 0.0004172249899017099,
      "loss": 3.0729,
      "step": 45350
    },
    {
      "epoch": 12.217484926787252,
      "grad_norm": 0.3396947979927063,
      "learning_rate": 0.00041702302410125217,
      "loss": 3.0872,
      "step": 45400
    },
    {
      "epoch": 12.230943152454781,
      "grad_norm": 0.39316675066947937,
      "learning_rate": 0.00041682105830079437,
      "loss": 3.0874,
      "step": 45450
    },
    {
      "epoch": 12.244401378122308,
      "grad_norm": 0.36940768361091614,
      "learning_rate": 0.00041661909250033656,
      "loss": 3.0771,
      "step": 45500
    },
    {
      "epoch": 12.257859603789836,
      "grad_norm": 0.3521488606929779,
      "learning_rate": 0.00041641712669987876,
      "loss": 3.0887,
      "step": 45550
    },
    {
      "epoch": 12.271317829457365,
      "grad_norm": 0.3894931674003601,
      "learning_rate": 0.00041621516089942096,
      "loss": 3.0862,
      "step": 45600
    },
    {
      "epoch": 12.284776055124892,
      "grad_norm": 0.356293648481369,
      "learning_rate": 0.0004160131950989632,
      "loss": 3.096,
      "step": 45650
    },
    {
      "epoch": 12.298234280792421,
      "grad_norm": 0.3514855206012726,
      "learning_rate": 0.0004158112292985054,
      "loss": 3.0934,
      "step": 45700
    },
    {
      "epoch": 12.311692506459949,
      "grad_norm": 0.3516969084739685,
      "learning_rate": 0.0004156092634980476,
      "loss": 3.09,
      "step": 45750
    },
    {
      "epoch": 12.325150732127476,
      "grad_norm": 0.33617040514945984,
      "learning_rate": 0.00041540729769758985,
      "loss": 3.0993,
      "step": 45800
    },
    {
      "epoch": 12.338608957795005,
      "grad_norm": 0.41160762310028076,
      "learning_rate": 0.0004152053318971321,
      "loss": 3.0943,
      "step": 45850
    },
    {
      "epoch": 12.352067183462532,
      "grad_norm": 0.3748043477535248,
      "learning_rate": 0.0004150033660966743,
      "loss": 3.0894,
      "step": 45900
    },
    {
      "epoch": 12.36552540913006,
      "grad_norm": 0.3719998300075531,
      "learning_rate": 0.0004148014002962165,
      "loss": 3.1037,
      "step": 45950
    },
    {
      "epoch": 12.378983634797589,
      "grad_norm": 0.3646189272403717,
      "learning_rate": 0.0004145994344957587,
      "loss": 3.098,
      "step": 46000
    },
    {
      "epoch": 12.378983634797589,
      "eval_accuracy": 0.3917944679628671,
      "eval_loss": 3.3045859336853027,
      "eval_runtime": 83.258,
      "eval_samples_per_second": 216.339,
      "eval_steps_per_second": 13.524,
      "step": 46000
    },
    {
      "epoch": 12.392441860465116,
      "grad_norm": 0.378118634223938,
      "learning_rate": 0.0004143974686953009,
      "loss": 3.0976,
      "step": 46050
    },
    {
      "epoch": 12.405900086132645,
      "grad_norm": 0.3856581449508667,
      "learning_rate": 0.00041419550289484313,
      "loss": 3.0917,
      "step": 46100
    },
    {
      "epoch": 12.419358311800172,
      "grad_norm": 0.36028414964675903,
      "learning_rate": 0.00041399353709438533,
      "loss": 3.0849,
      "step": 46150
    },
    {
      "epoch": 12.4328165374677,
      "grad_norm": 0.36001455783843994,
      "learning_rate": 0.0004137915712939275,
      "loss": 3.1052,
      "step": 46200
    },
    {
      "epoch": 12.446274763135229,
      "grad_norm": 0.3475915193557739,
      "learning_rate": 0.0004135896054934697,
      "loss": 3.0963,
      "step": 46250
    },
    {
      "epoch": 12.459732988802756,
      "grad_norm": 0.37853172421455383,
      "learning_rate": 0.0004133876396930119,
      "loss": 3.105,
      "step": 46300
    },
    {
      "epoch": 12.473191214470285,
      "grad_norm": 0.3637268841266632,
      "learning_rate": 0.00041318567389255417,
      "loss": 3.1004,
      "step": 46350
    },
    {
      "epoch": 12.486649440137812,
      "grad_norm": 0.37405914068222046,
      "learning_rate": 0.00041298370809209636,
      "loss": 3.1036,
      "step": 46400
    },
    {
      "epoch": 12.50010766580534,
      "grad_norm": 0.3894309997558594,
      "learning_rate": 0.00041278174229163856,
      "loss": 3.1077,
      "step": 46450
    },
    {
      "epoch": 12.513565891472869,
      "grad_norm": 0.35571929812431335,
      "learning_rate": 0.00041257977649118076,
      "loss": 3.1117,
      "step": 46500
    },
    {
      "epoch": 12.527024117140396,
      "grad_norm": 0.35282158851623535,
      "learning_rate": 0.00041237781069072295,
      "loss": 3.1152,
      "step": 46550
    },
    {
      "epoch": 12.540482342807923,
      "grad_norm": 0.352897584438324,
      "learning_rate": 0.0004121758448902652,
      "loss": 3.1095,
      "step": 46600
    },
    {
      "epoch": 12.553940568475452,
      "grad_norm": 0.40007737278938293,
      "learning_rate": 0.0004119738790898074,
      "loss": 3.1118,
      "step": 46650
    },
    {
      "epoch": 12.56739879414298,
      "grad_norm": 0.3525993824005127,
      "learning_rate": 0.00041177191328934965,
      "loss": 3.1094,
      "step": 46700
    },
    {
      "epoch": 12.580857019810509,
      "grad_norm": 0.3826589286327362,
      "learning_rate": 0.00041156994748889185,
      "loss": 3.1186,
      "step": 46750
    },
    {
      "epoch": 12.594315245478036,
      "grad_norm": 0.37100452184677124,
      "learning_rate": 0.0004113679816884341,
      "loss": 3.1146,
      "step": 46800
    },
    {
      "epoch": 12.607773471145563,
      "grad_norm": 0.359927237033844,
      "learning_rate": 0.0004111660158879763,
      "loss": 3.1039,
      "step": 46850
    },
    {
      "epoch": 12.621231696813092,
      "grad_norm": 0.3640861511230469,
      "learning_rate": 0.0004109640500875185,
      "loss": 3.112,
      "step": 46900
    },
    {
      "epoch": 12.63468992248062,
      "grad_norm": 0.3482762277126312,
      "learning_rate": 0.0004107620842870607,
      "loss": 3.1083,
      "step": 46950
    },
    {
      "epoch": 12.648148148148149,
      "grad_norm": 0.3688249886035919,
      "learning_rate": 0.0004105601184866029,
      "loss": 3.1211,
      "step": 47000
    },
    {
      "epoch": 12.648148148148149,
      "eval_accuracy": 0.3923484582793902,
      "eval_loss": 3.297767400741577,
      "eval_runtime": 83.1536,
      "eval_samples_per_second": 216.611,
      "eval_steps_per_second": 13.541,
      "step": 47000
    },
    {
      "epoch": 12.661606373815676,
      "grad_norm": 0.38570636510849,
      "learning_rate": 0.00041035815268614513,
      "loss": 3.1164,
      "step": 47050
    },
    {
      "epoch": 12.675064599483203,
      "grad_norm": 0.35626137256622314,
      "learning_rate": 0.0004101561868856873,
      "loss": 3.123,
      "step": 47100
    },
    {
      "epoch": 12.688522825150732,
      "grad_norm": 0.37804028391838074,
      "learning_rate": 0.0004099542210852295,
      "loss": 3.1272,
      "step": 47150
    },
    {
      "epoch": 12.70198105081826,
      "grad_norm": 0.3478531539440155,
      "learning_rate": 0.0004097522552847717,
      "loss": 3.1131,
      "step": 47200
    },
    {
      "epoch": 12.715439276485789,
      "grad_norm": 0.3422720432281494,
      "learning_rate": 0.00040955028948431397,
      "loss": 3.1185,
      "step": 47250
    },
    {
      "epoch": 12.728897502153316,
      "grad_norm": 0.3598884046077728,
      "learning_rate": 0.00040934832368385616,
      "loss": 3.1103,
      "step": 47300
    },
    {
      "epoch": 12.742355727820843,
      "grad_norm": 0.38301265239715576,
      "learning_rate": 0.00040914635788339836,
      "loss": 3.1153,
      "step": 47350
    },
    {
      "epoch": 12.755813953488373,
      "grad_norm": 0.3430884778499603,
      "learning_rate": 0.00040894439208294056,
      "loss": 3.1138,
      "step": 47400
    },
    {
      "epoch": 12.7692721791559,
      "grad_norm": 0.3731801509857178,
      "learning_rate": 0.00040874242628248275,
      "loss": 3.1254,
      "step": 47450
    },
    {
      "epoch": 12.782730404823429,
      "grad_norm": 0.3847680389881134,
      "learning_rate": 0.000408540460482025,
      "loss": 3.1205,
      "step": 47500
    },
    {
      "epoch": 12.796188630490956,
      "grad_norm": 0.3496476113796234,
      "learning_rate": 0.00040833849468156725,
      "loss": 3.1254,
      "step": 47550
    },
    {
      "epoch": 12.809646856158484,
      "grad_norm": 0.3567728102207184,
      "learning_rate": 0.00040813652888110945,
      "loss": 3.1204,
      "step": 47600
    },
    {
      "epoch": 12.823105081826013,
      "grad_norm": 0.3505919575691223,
      "learning_rate": 0.00040793456308065165,
      "loss": 3.1252,
      "step": 47650
    },
    {
      "epoch": 12.83656330749354,
      "grad_norm": 0.34096086025238037,
      "learning_rate": 0.0004077325972801939,
      "loss": 3.1211,
      "step": 47700
    },
    {
      "epoch": 12.850021533161069,
      "grad_norm": 0.348812460899353,
      "learning_rate": 0.0004075306314797361,
      "loss": 3.1146,
      "step": 47750
    },
    {
      "epoch": 12.863479758828596,
      "grad_norm": 0.38171035051345825,
      "learning_rate": 0.0004073286656792783,
      "loss": 3.1294,
      "step": 47800
    },
    {
      "epoch": 12.876937984496124,
      "grad_norm": 0.35344094038009644,
      "learning_rate": 0.0004071266998788205,
      "loss": 3.1284,
      "step": 47850
    },
    {
      "epoch": 12.890396210163653,
      "grad_norm": 0.35678407549858093,
      "learning_rate": 0.0004069247340783627,
      "loss": 3.124,
      "step": 47900
    },
    {
      "epoch": 12.90385443583118,
      "grad_norm": 0.40017467737197876,
      "learning_rate": 0.00040672276827790493,
      "loss": 3.1238,
      "step": 47950
    },
    {
      "epoch": 12.917312661498707,
      "grad_norm": 0.353305846452713,
      "learning_rate": 0.0004065208024774471,
      "loss": 3.1252,
      "step": 48000
    },
    {
      "epoch": 12.917312661498707,
      "eval_accuracy": 0.39289973242452414,
      "eval_loss": 3.291661024093628,
      "eval_runtime": 83.3851,
      "eval_samples_per_second": 216.01,
      "eval_steps_per_second": 13.504,
      "step": 48000
    },
    {
      "epoch": 12.930770887166236,
      "grad_norm": 0.35048526525497437,
      "learning_rate": 0.0004063188366769893,
      "loss": 3.141,
      "step": 48050
    },
    {
      "epoch": 12.944229112833764,
      "grad_norm": 0.355907678604126,
      "learning_rate": 0.0004061168708765315,
      "loss": 3.1355,
      "step": 48100
    },
    {
      "epoch": 12.957687338501293,
      "grad_norm": 0.36536088585853577,
      "learning_rate": 0.0004059149050760737,
      "loss": 3.1293,
      "step": 48150
    },
    {
      "epoch": 12.97114556416882,
      "grad_norm": 0.3732836842536926,
      "learning_rate": 0.00040571293927561597,
      "loss": 3.1195,
      "step": 48200
    },
    {
      "epoch": 12.984603789836347,
      "grad_norm": 0.37110257148742676,
      "learning_rate": 0.00040551097347515816,
      "loss": 3.1322,
      "step": 48250
    },
    {
      "epoch": 12.998062015503876,
      "grad_norm": 0.3679368197917938,
      "learning_rate": 0.00040530900767470036,
      "loss": 3.1277,
      "step": 48300
    },
    {
      "epoch": 13.011304909560723,
      "grad_norm": 0.36868977546691895,
      "learning_rate": 0.00040510704187424255,
      "loss": 3.046,
      "step": 48350
    },
    {
      "epoch": 13.024763135228252,
      "grad_norm": 0.3860374689102173,
      "learning_rate": 0.00040490507607378475,
      "loss": 3.0256,
      "step": 48400
    },
    {
      "epoch": 13.038221360895779,
      "grad_norm": 0.36192137002944946,
      "learning_rate": 0.00040470311027332705,
      "loss": 3.0316,
      "step": 48450
    },
    {
      "epoch": 13.051679586563308,
      "grad_norm": 0.3843608796596527,
      "learning_rate": 0.00040450114447286925,
      "loss": 3.039,
      "step": 48500
    },
    {
      "epoch": 13.065137812230835,
      "grad_norm": 0.37431687116622925,
      "learning_rate": 0.00040429917867241145,
      "loss": 3.0341,
      "step": 48550
    },
    {
      "epoch": 13.078596037898363,
      "grad_norm": 0.36502817273139954,
      "learning_rate": 0.00040409721287195364,
      "loss": 3.0445,
      "step": 48600
    },
    {
      "epoch": 13.092054263565892,
      "grad_norm": 0.3744087219238281,
      "learning_rate": 0.0004038952470714959,
      "loss": 3.0327,
      "step": 48650
    },
    {
      "epoch": 13.10551248923342,
      "grad_norm": 0.3876210153102875,
      "learning_rate": 0.0004036932812710381,
      "loss": 3.0489,
      "step": 48700
    },
    {
      "epoch": 13.118970714900948,
      "grad_norm": 0.36698827147483826,
      "learning_rate": 0.0004034913154705803,
      "loss": 3.0473,
      "step": 48750
    },
    {
      "epoch": 13.132428940568476,
      "grad_norm": 0.3697962462902069,
      "learning_rate": 0.0004032893496701225,
      "loss": 3.0524,
      "step": 48800
    },
    {
      "epoch": 13.145887166236003,
      "grad_norm": 0.35449784994125366,
      "learning_rate": 0.0004030873838696647,
      "loss": 3.0573,
      "step": 48850
    },
    {
      "epoch": 13.159345391903532,
      "grad_norm": 0.36308717727661133,
      "learning_rate": 0.00040288541806920693,
      "loss": 3.0631,
      "step": 48900
    },
    {
      "epoch": 13.17280361757106,
      "grad_norm": 0.36822566390037537,
      "learning_rate": 0.0004026834522687491,
      "loss": 3.0673,
      "step": 48950
    },
    {
      "epoch": 13.186261843238588,
      "grad_norm": 0.39083588123321533,
      "learning_rate": 0.0004024814864682913,
      "loss": 3.0567,
      "step": 49000
    },
    {
      "epoch": 13.186261843238588,
      "eval_accuracy": 0.39197395256825956,
      "eval_loss": 3.307344913482666,
      "eval_runtime": 83.1676,
      "eval_samples_per_second": 216.575,
      "eval_steps_per_second": 13.539,
      "step": 49000
    },
    {
      "epoch": 13.199720068906116,
      "grad_norm": 0.36873242259025574,
      "learning_rate": 0.0004022795206678335,
      "loss": 3.0442,
      "step": 49050
    },
    {
      "epoch": 13.213178294573643,
      "grad_norm": 0.356784850358963,
      "learning_rate": 0.00040207755486737577,
      "loss": 3.0689,
      "step": 49100
    },
    {
      "epoch": 13.226636520241172,
      "grad_norm": 0.3866429924964905,
      "learning_rate": 0.00040187558906691796,
      "loss": 3.057,
      "step": 49150
    },
    {
      "epoch": 13.2400947459087,
      "grad_norm": 0.3816594183444977,
      "learning_rate": 0.00040167362326646016,
      "loss": 3.0636,
      "step": 49200
    },
    {
      "epoch": 13.253552971576227,
      "grad_norm": 0.39061596989631653,
      "learning_rate": 0.00040147165746600235,
      "loss": 3.0786,
      "step": 49250
    },
    {
      "epoch": 13.267011197243756,
      "grad_norm": 0.3700222373008728,
      "learning_rate": 0.00040126969166554455,
      "loss": 3.0775,
      "step": 49300
    },
    {
      "epoch": 13.280469422911283,
      "grad_norm": 0.3660484552383423,
      "learning_rate": 0.00040106772586508686,
      "loss": 3.0839,
      "step": 49350
    },
    {
      "epoch": 13.293927648578812,
      "grad_norm": 0.3573848307132721,
      "learning_rate": 0.00040086576006462905,
      "loss": 3.0697,
      "step": 49400
    },
    {
      "epoch": 13.30738587424634,
      "grad_norm": 0.36954015493392944,
      "learning_rate": 0.00040066379426417125,
      "loss": 3.0801,
      "step": 49450
    },
    {
      "epoch": 13.320844099913867,
      "grad_norm": 0.38545018434524536,
      "learning_rate": 0.00040046182846371344,
      "loss": 3.0714,
      "step": 49500
    },
    {
      "epoch": 13.334302325581396,
      "grad_norm": 0.3988477289676666,
      "learning_rate": 0.0004002598626632557,
      "loss": 3.0719,
      "step": 49550
    },
    {
      "epoch": 13.347760551248923,
      "grad_norm": 0.36968204379081726,
      "learning_rate": 0.0004000578968627979,
      "loss": 3.079,
      "step": 49600
    },
    {
      "epoch": 13.361218776916452,
      "grad_norm": 0.39189064502716064,
      "learning_rate": 0.0003998559310623401,
      "loss": 3.083,
      "step": 49650
    },
    {
      "epoch": 13.37467700258398,
      "grad_norm": 0.3649806082248688,
      "learning_rate": 0.0003996539652618823,
      "loss": 3.0832,
      "step": 49700
    },
    {
      "epoch": 13.388135228251507,
      "grad_norm": 0.408478707075119,
      "learning_rate": 0.0003994519994614245,
      "loss": 3.0807,
      "step": 49750
    },
    {
      "epoch": 13.401593453919036,
      "grad_norm": 0.3362438678741455,
      "learning_rate": 0.00039925003366096673,
      "loss": 3.088,
      "step": 49800
    },
    {
      "epoch": 13.415051679586563,
      "grad_norm": 0.3814699351787567,
      "learning_rate": 0.0003990480678605089,
      "loss": 3.0846,
      "step": 49850
    },
    {
      "epoch": 13.428509905254092,
      "grad_norm": 0.4107760787010193,
      "learning_rate": 0.0003988461020600511,
      "loss": 3.0907,
      "step": 49900
    },
    {
      "epoch": 13.44196813092162,
      "grad_norm": 0.3375151753425598,
      "learning_rate": 0.0003986441362595933,
      "loss": 3.078,
      "step": 49950
    },
    {
      "epoch": 13.455426356589147,
      "grad_norm": 0.3659836947917938,
      "learning_rate": 0.0003984421704591355,
      "loss": 3.0894,
      "step": 50000
    },
    {
      "epoch": 13.455426356589147,
      "eval_accuracy": 0.3926573412897599,
      "eval_loss": 3.30159592628479,
      "eval_runtime": 83.2283,
      "eval_samples_per_second": 216.417,
      "eval_steps_per_second": 13.529,
      "step": 50000
    },
    {
      "epoch": 13.468884582256676,
      "grad_norm": 0.3724363148212433,
      "learning_rate": 0.00039824020465867776,
      "loss": 3.0855,
      "step": 50050
    },
    {
      "epoch": 13.482342807924203,
      "grad_norm": 0.3457666039466858,
      "learning_rate": 0.00039803823885821996,
      "loss": 3.0845,
      "step": 50100
    },
    {
      "epoch": 13.49580103359173,
      "grad_norm": 0.3506271541118622,
      "learning_rate": 0.00039783627305776216,
      "loss": 3.0944,
      "step": 50150
    },
    {
      "epoch": 13.50925925925926,
      "grad_norm": 0.3566408157348633,
      "learning_rate": 0.00039763430725730435,
      "loss": 3.1032,
      "step": 50200
    },
    {
      "epoch": 13.522717484926787,
      "grad_norm": 0.3639317750930786,
      "learning_rate": 0.00039743234145684666,
      "loss": 3.0951,
      "step": 50250
    },
    {
      "epoch": 13.536175710594316,
      "grad_norm": 0.39604026079177856,
      "learning_rate": 0.00039723037565638885,
      "loss": 3.0997,
      "step": 50300
    },
    {
      "epoch": 13.549633936261843,
      "grad_norm": 0.368399053812027,
      "learning_rate": 0.00039702840985593105,
      "loss": 3.0991,
      "step": 50350
    },
    {
      "epoch": 13.56309216192937,
      "grad_norm": 0.34584465622901917,
      "learning_rate": 0.00039682644405547324,
      "loss": 3.0976,
      "step": 50400
    },
    {
      "epoch": 13.5765503875969,
      "grad_norm": 0.36537596583366394,
      "learning_rate": 0.00039662447825501544,
      "loss": 3.1096,
      "step": 50450
    },
    {
      "epoch": 13.590008613264427,
      "grad_norm": 0.37870803475379944,
      "learning_rate": 0.0003964225124545577,
      "loss": 3.0884,
      "step": 50500
    },
    {
      "epoch": 13.603466838931956,
      "grad_norm": 0.3623582124710083,
      "learning_rate": 0.0003962205466540999,
      "loss": 3.1062,
      "step": 50550
    },
    {
      "epoch": 13.616925064599483,
      "grad_norm": 0.38529542088508606,
      "learning_rate": 0.0003960185808536421,
      "loss": 3.1028,
      "step": 50600
    },
    {
      "epoch": 13.63038329026701,
      "grad_norm": 0.3744858205318451,
      "learning_rate": 0.0003958166150531843,
      "loss": 3.097,
      "step": 50650
    },
    {
      "epoch": 13.64384151593454,
      "grad_norm": 0.3665783405303955,
      "learning_rate": 0.0003956146492527265,
      "loss": 3.0995,
      "step": 50700
    },
    {
      "epoch": 13.657299741602067,
      "grad_norm": 0.42016690969467163,
      "learning_rate": 0.0003954126834522687,
      "loss": 3.0945,
      "step": 50750
    },
    {
      "epoch": 13.670757967269594,
      "grad_norm": 0.39203500747680664,
      "learning_rate": 0.0003952107176518109,
      "loss": 3.0959,
      "step": 50800
    },
    {
      "epoch": 13.684216192937123,
      "grad_norm": 0.36824801564216614,
      "learning_rate": 0.0003950087518513531,
      "loss": 3.1006,
      "step": 50850
    },
    {
      "epoch": 13.69767441860465,
      "grad_norm": 0.3983285129070282,
      "learning_rate": 0.0003948067860508953,
      "loss": 3.101,
      "step": 50900
    },
    {
      "epoch": 13.71113264427218,
      "grad_norm": 0.37094029784202576,
      "learning_rate": 0.0003946048202504375,
      "loss": 3.0981,
      "step": 50950
    },
    {
      "epoch": 13.724590869939707,
      "grad_norm": 0.37548765540122986,
      "learning_rate": 0.00039440285444997976,
      "loss": 3.101,
      "step": 51000
    },
    {
      "epoch": 13.724590869939707,
      "eval_accuracy": 0.39297991380393066,
      "eval_loss": 3.2925641536712646,
      "eval_runtime": 83.3334,
      "eval_samples_per_second": 216.144,
      "eval_steps_per_second": 13.512,
      "step": 51000
    },
    {
      "epoch": 13.738049095607234,
      "grad_norm": 0.38235723972320557,
      "learning_rate": 0.00039420088864952196,
      "loss": 3.108,
      "step": 51050
    },
    {
      "epoch": 13.751507321274763,
      "grad_norm": 0.3521125912666321,
      "learning_rate": 0.00039399892284906415,
      "loss": 3.1136,
      "step": 51100
    },
    {
      "epoch": 13.76496554694229,
      "grad_norm": 0.3683144748210907,
      "learning_rate": 0.00039379695704860646,
      "loss": 3.1035,
      "step": 51150
    },
    {
      "epoch": 13.77842377260982,
      "grad_norm": 0.36110466718673706,
      "learning_rate": 0.00039359499124814865,
      "loss": 3.1044,
      "step": 51200
    },
    {
      "epoch": 13.791881998277347,
      "grad_norm": 0.3849784731864929,
      "learning_rate": 0.00039339302544769085,
      "loss": 3.098,
      "step": 51250
    },
    {
      "epoch": 13.805340223944874,
      "grad_norm": 0.37069833278656006,
      "learning_rate": 0.00039319105964723305,
      "loss": 3.1065,
      "step": 51300
    },
    {
      "epoch": 13.818798449612403,
      "grad_norm": 0.35221436619758606,
      "learning_rate": 0.00039298909384677524,
      "loss": 3.1062,
      "step": 51350
    },
    {
      "epoch": 13.83225667527993,
      "grad_norm": 0.3762305974960327,
      "learning_rate": 0.0003927871280463175,
      "loss": 3.1065,
      "step": 51400
    },
    {
      "epoch": 13.84571490094746,
      "grad_norm": 0.3478483259677887,
      "learning_rate": 0.0003925851622458597,
      "loss": 3.1079,
      "step": 51450
    },
    {
      "epoch": 13.859173126614987,
      "grad_norm": 0.37982606887817383,
      "learning_rate": 0.0003923831964454019,
      "loss": 3.1113,
      "step": 51500
    },
    {
      "epoch": 13.872631352282514,
      "grad_norm": 0.3709942400455475,
      "learning_rate": 0.0003921812306449441,
      "loss": 3.1142,
      "step": 51550
    },
    {
      "epoch": 13.886089577950044,
      "grad_norm": 0.4197673201560974,
      "learning_rate": 0.0003919792648444863,
      "loss": 3.1027,
      "step": 51600
    },
    {
      "epoch": 13.89954780361757,
      "grad_norm": 0.37212076783180237,
      "learning_rate": 0.0003917772990440285,
      "loss": 3.1148,
      "step": 51650
    },
    {
      "epoch": 13.9130060292851,
      "grad_norm": 0.3714587092399597,
      "learning_rate": 0.0003915753332435707,
      "loss": 3.112,
      "step": 51700
    },
    {
      "epoch": 13.926464254952627,
      "grad_norm": 0.366169273853302,
      "learning_rate": 0.0003913733674431129,
      "loss": 3.1141,
      "step": 51750
    },
    {
      "epoch": 13.939922480620154,
      "grad_norm": 0.3851865530014038,
      "learning_rate": 0.0003911714016426551,
      "loss": 3.1172,
      "step": 51800
    },
    {
      "epoch": 13.953380706287684,
      "grad_norm": 0.3594266474246979,
      "learning_rate": 0.0003909694358421973,
      "loss": 3.1064,
      "step": 51850
    },
    {
      "epoch": 13.96683893195521,
      "grad_norm": 0.35453295707702637,
      "learning_rate": 0.00039076747004173956,
      "loss": 3.1108,
      "step": 51900
    },
    {
      "epoch": 13.98029715762274,
      "grad_norm": 0.35283157229423523,
      "learning_rate": 0.00039056550424128176,
      "loss": 3.1167,
      "step": 51950
    },
    {
      "epoch": 13.993755383290267,
      "grad_norm": 0.36269649863243103,
      "learning_rate": 0.00039036353844082395,
      "loss": 3.1171,
      "step": 52000
    },
    {
      "epoch": 13.993755383290267,
      "eval_accuracy": 0.393478711517827,
      "eval_loss": 3.287684917449951,
      "eval_runtime": 83.2052,
      "eval_samples_per_second": 216.477,
      "eval_steps_per_second": 13.533,
      "step": 52000
    },
    {
      "epoch": 14.006998277347115,
      "grad_norm": 0.37233373522758484,
      "learning_rate": 0.0003901615726403662,
      "loss": 3.0692,
      "step": 52050
    },
    {
      "epoch": 14.020456503014643,
      "grad_norm": 0.36130645871162415,
      "learning_rate": 0.00038995960683990845,
      "loss": 3.0228,
      "step": 52100
    },
    {
      "epoch": 14.03391472868217,
      "grad_norm": 0.3740144670009613,
      "learning_rate": 0.00038975764103945065,
      "loss": 3.0162,
      "step": 52150
    },
    {
      "epoch": 14.047372954349699,
      "grad_norm": 0.35414737462997437,
      "learning_rate": 0.00038955567523899285,
      "loss": 3.0209,
      "step": 52200
    },
    {
      "epoch": 14.060831180017226,
      "grad_norm": 0.3854324519634247,
      "learning_rate": 0.00038935370943853504,
      "loss": 3.0184,
      "step": 52250
    },
    {
      "epoch": 14.074289405684755,
      "grad_norm": 0.38975897431373596,
      "learning_rate": 0.00038915174363807724,
      "loss": 3.0218,
      "step": 52300
    },
    {
      "epoch": 14.087747631352283,
      "grad_norm": 0.3861389756202698,
      "learning_rate": 0.0003889497778376195,
      "loss": 3.0356,
      "step": 52350
    },
    {
      "epoch": 14.10120585701981,
      "grad_norm": 0.3889559507369995,
      "learning_rate": 0.0003887478120371617,
      "loss": 3.0364,
      "step": 52400
    },
    {
      "epoch": 14.114664082687339,
      "grad_norm": 0.3829285800457001,
      "learning_rate": 0.0003885458462367039,
      "loss": 3.0403,
      "step": 52450
    },
    {
      "epoch": 14.128122308354866,
      "grad_norm": 0.37157002091407776,
      "learning_rate": 0.0003883438804362461,
      "loss": 3.039,
      "step": 52500
    },
    {
      "epoch": 14.141580534022394,
      "grad_norm": 0.37536025047302246,
      "learning_rate": 0.0003881419146357883,
      "loss": 3.0317,
      "step": 52550
    },
    {
      "epoch": 14.155038759689923,
      "grad_norm": 0.36771246790885925,
      "learning_rate": 0.0003879399488353305,
      "loss": 3.0499,
      "step": 52600
    },
    {
      "epoch": 14.16849698535745,
      "grad_norm": 0.40229716897010803,
      "learning_rate": 0.0003877379830348727,
      "loss": 3.0463,
      "step": 52650
    },
    {
      "epoch": 14.18195521102498,
      "grad_norm": 0.38133320212364197,
      "learning_rate": 0.0003875360172344149,
      "loss": 3.0487,
      "step": 52700
    },
    {
      "epoch": 14.195413436692506,
      "grad_norm": 0.3584715723991394,
      "learning_rate": 0.0003873340514339571,
      "loss": 3.0386,
      "step": 52750
    },
    {
      "epoch": 14.208871662360034,
      "grad_norm": 0.38643693923950195,
      "learning_rate": 0.0003871320856334993,
      "loss": 3.0538,
      "step": 52800
    },
    {
      "epoch": 14.222329888027563,
      "grad_norm": 0.3699548542499542,
      "learning_rate": 0.00038693011983304156,
      "loss": 3.0646,
      "step": 52850
    },
    {
      "epoch": 14.23578811369509,
      "grad_norm": 0.37531349062919617,
      "learning_rate": 0.0003867281540325838,
      "loss": 3.0559,
      "step": 52900
    },
    {
      "epoch": 14.24924633936262,
      "grad_norm": 0.3837898373603821,
      "learning_rate": 0.000386526188232126,
      "loss": 3.0553,
      "step": 52950
    },
    {
      "epoch": 14.262704565030146,
      "grad_norm": 0.38276827335357666,
      "learning_rate": 0.00038632422243166825,
      "loss": 3.0578,
      "step": 53000
    },
    {
      "epoch": 14.262704565030146,
      "eval_accuracy": 0.39303771393109094,
      "eval_loss": 3.3016796112060547,
      "eval_runtime": 83.3444,
      "eval_samples_per_second": 216.115,
      "eval_steps_per_second": 13.51,
      "step": 53000
    },
    {
      "epoch": 14.276162790697674,
      "grad_norm": 0.40769684314727783,
      "learning_rate": 0.00038612225663121045,
      "loss": 3.0476,
      "step": 53050
    },
    {
      "epoch": 14.289621016365203,
      "grad_norm": 0.3898642063140869,
      "learning_rate": 0.00038592029083075265,
      "loss": 3.0569,
      "step": 53100
    },
    {
      "epoch": 14.30307924203273,
      "grad_norm": 0.36371198296546936,
      "learning_rate": 0.00038571832503029484,
      "loss": 3.0503,
      "step": 53150
    },
    {
      "epoch": 14.31653746770026,
      "grad_norm": 0.396706759929657,
      "learning_rate": 0.00038551635922983704,
      "loss": 3.0647,
      "step": 53200
    },
    {
      "epoch": 14.329995693367787,
      "grad_norm": 0.40205034613609314,
      "learning_rate": 0.0003853143934293793,
      "loss": 3.0602,
      "step": 53250
    },
    {
      "epoch": 14.343453919035314,
      "grad_norm": 0.3814631402492523,
      "learning_rate": 0.0003851124276289215,
      "loss": 3.0651,
      "step": 53300
    },
    {
      "epoch": 14.356912144702843,
      "grad_norm": 0.3645440340042114,
      "learning_rate": 0.0003849104618284637,
      "loss": 3.065,
      "step": 53350
    },
    {
      "epoch": 14.37037037037037,
      "grad_norm": 0.3672763705253601,
      "learning_rate": 0.0003847084960280059,
      "loss": 3.0706,
      "step": 53400
    },
    {
      "epoch": 14.383828596037898,
      "grad_norm": 0.3941022455692291,
      "learning_rate": 0.0003845065302275481,
      "loss": 3.0728,
      "step": 53450
    },
    {
      "epoch": 14.397286821705427,
      "grad_norm": 0.38894933462142944,
      "learning_rate": 0.0003843045644270903,
      "loss": 3.0597,
      "step": 53500
    },
    {
      "epoch": 14.410745047372954,
      "grad_norm": 0.363395094871521,
      "learning_rate": 0.0003841025986266325,
      "loss": 3.0795,
      "step": 53550
    },
    {
      "epoch": 14.424203273040483,
      "grad_norm": 0.40244677662849426,
      "learning_rate": 0.0003839006328261747,
      "loss": 3.0619,
      "step": 53600
    },
    {
      "epoch": 14.43766149870801,
      "grad_norm": 0.3906124532222748,
      "learning_rate": 0.0003836986670257169,
      "loss": 3.0721,
      "step": 53650
    },
    {
      "epoch": 14.451119724375538,
      "grad_norm": 0.3765961229801178,
      "learning_rate": 0.0003834967012252591,
      "loss": 3.0793,
      "step": 53700
    },
    {
      "epoch": 14.464577950043067,
      "grad_norm": 0.3905157148838043,
      "learning_rate": 0.00038329473542480136,
      "loss": 3.0669,
      "step": 53750
    },
    {
      "epoch": 14.478036175710594,
      "grad_norm": 0.38133829832077026,
      "learning_rate": 0.0003830927696243436,
      "loss": 3.0859,
      "step": 53800
    },
    {
      "epoch": 14.491494401378123,
      "grad_norm": 0.39759406447410583,
      "learning_rate": 0.0003828908038238858,
      "loss": 3.074,
      "step": 53850
    },
    {
      "epoch": 14.50495262704565,
      "grad_norm": 0.38885143399238586,
      "learning_rate": 0.000382688838023428,
      "loss": 3.0819,
      "step": 53900
    },
    {
      "epoch": 14.518410852713178,
      "grad_norm": 0.3945043683052063,
      "learning_rate": 0.00038248687222297025,
      "loss": 3.076,
      "step": 53950
    },
    {
      "epoch": 14.531869078380707,
      "grad_norm": 0.3805738091468811,
      "learning_rate": 0.00038228490642251245,
      "loss": 3.0831,
      "step": 54000
    },
    {
      "epoch": 14.531869078380707,
      "eval_accuracy": 0.392992299545465,
      "eval_loss": 3.2958991527557373,
      "eval_runtime": 83.3491,
      "eval_samples_per_second": 216.103,
      "eval_steps_per_second": 13.509,
      "step": 54000
    },
    {
      "epoch": 14.545327304048234,
      "grad_norm": 0.39582914113998413,
      "learning_rate": 0.00038208294062205464,
      "loss": 3.082,
      "step": 54050
    },
    {
      "epoch": 14.558785529715763,
      "grad_norm": 0.3954962491989136,
      "learning_rate": 0.00038188097482159684,
      "loss": 3.0661,
      "step": 54100
    },
    {
      "epoch": 14.57224375538329,
      "grad_norm": 0.3633665144443512,
      "learning_rate": 0.00038167900902113904,
      "loss": 3.0756,
      "step": 54150
    },
    {
      "epoch": 14.585701981050818,
      "grad_norm": 0.38512715697288513,
      "learning_rate": 0.0003814770432206813,
      "loss": 3.095,
      "step": 54200
    },
    {
      "epoch": 14.599160206718347,
      "grad_norm": 0.35448163747787476,
      "learning_rate": 0.0003812750774202235,
      "loss": 3.0822,
      "step": 54250
    },
    {
      "epoch": 14.612618432385874,
      "grad_norm": 0.3850742280483246,
      "learning_rate": 0.0003810731116197657,
      "loss": 3.0823,
      "step": 54300
    },
    {
      "epoch": 14.626076658053403,
      "grad_norm": 0.3849395215511322,
      "learning_rate": 0.0003808711458193079,
      "loss": 3.0836,
      "step": 54350
    },
    {
      "epoch": 14.63953488372093,
      "grad_norm": 0.3653835356235504,
      "learning_rate": 0.00038066918001885007,
      "loss": 3.0884,
      "step": 54400
    },
    {
      "epoch": 14.652993109388458,
      "grad_norm": 0.37795838713645935,
      "learning_rate": 0.0003804672142183923,
      "loss": 3.0862,
      "step": 54450
    },
    {
      "epoch": 14.666451335055987,
      "grad_norm": 0.38842540979385376,
      "learning_rate": 0.0003802652484179345,
      "loss": 3.0867,
      "step": 54500
    },
    {
      "epoch": 14.679909560723514,
      "grad_norm": 0.38113725185394287,
      "learning_rate": 0.0003800632826174767,
      "loss": 3.0887,
      "step": 54550
    },
    {
      "epoch": 14.693367786391041,
      "grad_norm": 0.4157524108886719,
      "learning_rate": 0.0003798613168170189,
      "loss": 3.0878,
      "step": 54600
    },
    {
      "epoch": 14.70682601205857,
      "grad_norm": 0.3839399218559265,
      "learning_rate": 0.0003796593510165611,
      "loss": 3.0831,
      "step": 54650
    },
    {
      "epoch": 14.720284237726098,
      "grad_norm": 0.37682539224624634,
      "learning_rate": 0.0003794573852161034,
      "loss": 3.0831,
      "step": 54700
    },
    {
      "epoch": 14.733742463393627,
      "grad_norm": 0.3848724961280823,
      "learning_rate": 0.0003792554194156456,
      "loss": 3.0933,
      "step": 54750
    },
    {
      "epoch": 14.747200689061154,
      "grad_norm": 0.40955615043640137,
      "learning_rate": 0.0003790534536151878,
      "loss": 3.0961,
      "step": 54800
    },
    {
      "epoch": 14.760658914728682,
      "grad_norm": 0.4166955053806305,
      "learning_rate": 0.00037885148781473005,
      "loss": 3.0891,
      "step": 54850
    },
    {
      "epoch": 14.77411714039621,
      "grad_norm": 0.34451085329055786,
      "learning_rate": 0.00037864952201427225,
      "loss": 3.0798,
      "step": 54900
    },
    {
      "epoch": 14.787575366063738,
      "grad_norm": 0.3508240282535553,
      "learning_rate": 0.00037844755621381444,
      "loss": 3.1008,
      "step": 54950
    },
    {
      "epoch": 14.801033591731267,
      "grad_norm": 0.3818090558052063,
      "learning_rate": 0.00037824559041335664,
      "loss": 3.0911,
      "step": 55000
    },
    {
      "epoch": 14.801033591731267,
      "eval_accuracy": 0.393826924689911,
      "eval_loss": 3.28652024269104,
      "eval_runtime": 83.4285,
      "eval_samples_per_second": 215.898,
      "eval_steps_per_second": 13.497,
      "step": 55000
    },
    {
      "epoch": 14.814491817398794,
      "grad_norm": 0.376193106174469,
      "learning_rate": 0.00037804362461289884,
      "loss": 3.0854,
      "step": 55050
    },
    {
      "epoch": 14.827950043066322,
      "grad_norm": 0.36976513266563416,
      "learning_rate": 0.0003778416588124411,
      "loss": 3.0889,
      "step": 55100
    },
    {
      "epoch": 14.84140826873385,
      "grad_norm": 0.3586288392543793,
      "learning_rate": 0.0003776396930119833,
      "loss": 3.0969,
      "step": 55150
    },
    {
      "epoch": 14.854866494401378,
      "grad_norm": 0.37360259890556335,
      "learning_rate": 0.0003774377272115255,
      "loss": 3.0902,
      "step": 55200
    },
    {
      "epoch": 14.868324720068905,
      "grad_norm": 0.397998571395874,
      "learning_rate": 0.0003772357614110677,
      "loss": 3.0929,
      "step": 55250
    },
    {
      "epoch": 14.881782945736434,
      "grad_norm": 0.36560338735580444,
      "learning_rate": 0.00037703379561060987,
      "loss": 3.0913,
      "step": 55300
    },
    {
      "epoch": 14.895241171403962,
      "grad_norm": 0.39389437437057495,
      "learning_rate": 0.0003768318298101521,
      "loss": 3.0949,
      "step": 55350
    },
    {
      "epoch": 14.90869939707149,
      "grad_norm": 0.3835701644420624,
      "learning_rate": 0.0003766298640096943,
      "loss": 3.1048,
      "step": 55400
    },
    {
      "epoch": 14.922157622739018,
      "grad_norm": 0.3771551549434662,
      "learning_rate": 0.0003764278982092365,
      "loss": 3.1091,
      "step": 55450
    },
    {
      "epoch": 14.935615848406545,
      "grad_norm": 0.38621798157691956,
      "learning_rate": 0.0003762259324087787,
      "loss": 3.0897,
      "step": 55500
    },
    {
      "epoch": 14.949074074074074,
      "grad_norm": 0.3835623264312744,
      "learning_rate": 0.0003760239666083209,
      "loss": 3.1042,
      "step": 55550
    },
    {
      "epoch": 14.962532299741602,
      "grad_norm": 0.3699501156806946,
      "learning_rate": 0.0003758220008078632,
      "loss": 3.0936,
      "step": 55600
    },
    {
      "epoch": 14.97599052540913,
      "grad_norm": 0.3850233256816864,
      "learning_rate": 0.0003756200350074054,
      "loss": 3.0955,
      "step": 55650
    },
    {
      "epoch": 14.989448751076658,
      "grad_norm": 0.41616466641426086,
      "learning_rate": 0.0003754180692069476,
      "loss": 3.1047,
      "step": 55700
    },
    {
      "epoch": 15.002691645133506,
      "grad_norm": 0.3711931109428406,
      "learning_rate": 0.0003752161034064898,
      "loss": 3.0735,
      "step": 55750
    },
    {
      "epoch": 15.016149870801033,
      "grad_norm": 0.34310173988342285,
      "learning_rate": 0.00037501413760603205,
      "loss": 3.0098,
      "step": 55800
    },
    {
      "epoch": 15.02960809646856,
      "grad_norm": 0.39193013310432434,
      "learning_rate": 0.00037481217180557425,
      "loss": 3.011,
      "step": 55850
    },
    {
      "epoch": 15.04306632213609,
      "grad_norm": 0.37037384510040283,
      "learning_rate": 0.00037461020600511644,
      "loss": 3.0014,
      "step": 55900
    },
    {
      "epoch": 15.056524547803617,
      "grad_norm": 0.36951473355293274,
      "learning_rate": 0.00037440824020465864,
      "loss": 3.0097,
      "step": 55950
    },
    {
      "epoch": 15.069982773471146,
      "grad_norm": 0.3650955259799957,
      "learning_rate": 0.00037420627440420083,
      "loss": 3.0138,
      "step": 56000
    },
    {
      "epoch": 15.069982773471146,
      "eval_accuracy": 0.3936073493948153,
      "eval_loss": 3.3001368045806885,
      "eval_runtime": 83.2587,
      "eval_samples_per_second": 216.338,
      "eval_steps_per_second": 13.524,
      "step": 56000
    },
    {
      "epoch": 15.083440999138674,
      "grad_norm": 0.38153523206710815,
      "learning_rate": 0.0003740043086037431,
      "loss": 3.0167,
      "step": 56050
    },
    {
      "epoch": 15.0968992248062,
      "grad_norm": 0.38686785101890564,
      "learning_rate": 0.0003738023428032853,
      "loss": 3.0214,
      "step": 56100
    },
    {
      "epoch": 15.11035745047373,
      "grad_norm": 0.3664006292819977,
      "learning_rate": 0.0003736003770028275,
      "loss": 3.0116,
      "step": 56150
    },
    {
      "epoch": 15.123815676141257,
      "grad_norm": 0.400959849357605,
      "learning_rate": 0.00037339841120236967,
      "loss": 3.0311,
      "step": 56200
    },
    {
      "epoch": 15.137273901808786,
      "grad_norm": 0.36764273047447205,
      "learning_rate": 0.00037319644540191187,
      "loss": 3.0123,
      "step": 56250
    },
    {
      "epoch": 15.150732127476314,
      "grad_norm": 0.3967522084712982,
      "learning_rate": 0.0003729944796014541,
      "loss": 3.0298,
      "step": 56300
    },
    {
      "epoch": 15.164190353143841,
      "grad_norm": 0.40533536672592163,
      "learning_rate": 0.0003727925138009963,
      "loss": 3.0188,
      "step": 56350
    },
    {
      "epoch": 15.17764857881137,
      "grad_norm": 0.37758809328079224,
      "learning_rate": 0.0003725905480005385,
      "loss": 3.03,
      "step": 56400
    },
    {
      "epoch": 15.191106804478897,
      "grad_norm": 0.40048253536224365,
      "learning_rate": 0.0003723885822000807,
      "loss": 3.024,
      "step": 56450
    },
    {
      "epoch": 15.204565030146426,
      "grad_norm": 0.390164315700531,
      "learning_rate": 0.000372186616399623,
      "loss": 3.0491,
      "step": 56500
    },
    {
      "epoch": 15.218023255813954,
      "grad_norm": 0.4010895788669586,
      "learning_rate": 0.0003719846505991652,
      "loss": 3.0408,
      "step": 56550
    },
    {
      "epoch": 15.231481481481481,
      "grad_norm": 0.37081724405288696,
      "learning_rate": 0.0003717826847987074,
      "loss": 3.0425,
      "step": 56600
    },
    {
      "epoch": 15.24493970714901,
      "grad_norm": 0.3892953395843506,
      "learning_rate": 0.0003715807189982496,
      "loss": 3.0366,
      "step": 56650
    },
    {
      "epoch": 15.258397932816537,
      "grad_norm": 0.39661431312561035,
      "learning_rate": 0.00037137875319779185,
      "loss": 3.0406,
      "step": 56700
    },
    {
      "epoch": 15.271856158484065,
      "grad_norm": 0.390493780374527,
      "learning_rate": 0.00037117678739733405,
      "loss": 3.0368,
      "step": 56750
    },
    {
      "epoch": 15.285314384151594,
      "grad_norm": 0.4005759060382843,
      "learning_rate": 0.00037097482159687624,
      "loss": 3.0354,
      "step": 56800
    },
    {
      "epoch": 15.298772609819121,
      "grad_norm": 0.3855637013912201,
      "learning_rate": 0.00037077285579641844,
      "loss": 3.0453,
      "step": 56850
    },
    {
      "epoch": 15.31223083548665,
      "grad_norm": 0.3907672166824341,
      "learning_rate": 0.00037057088999596063,
      "loss": 3.055,
      "step": 56900
    },
    {
      "epoch": 15.325689061154177,
      "grad_norm": 0.3794354796409607,
      "learning_rate": 0.0003703689241955029,
      "loss": 3.05,
      "step": 56950
    },
    {
      "epoch": 15.339147286821705,
      "grad_norm": 0.39498579502105713,
      "learning_rate": 0.0003701669583950451,
      "loss": 3.05,
      "step": 57000
    },
    {
      "epoch": 15.339147286821705,
      "eval_accuracy": 0.3933595259172728,
      "eval_loss": 3.2969160079956055,
      "eval_runtime": 83.4361,
      "eval_samples_per_second": 215.878,
      "eval_steps_per_second": 13.495,
      "step": 57000
    },
    {
      "epoch": 15.352605512489234,
      "grad_norm": 0.38030511140823364,
      "learning_rate": 0.0003699649925945873,
      "loss": 3.0488,
      "step": 57050
    },
    {
      "epoch": 15.366063738156761,
      "grad_norm": 0.4138135015964508,
      "learning_rate": 0.0003697630267941295,
      "loss": 3.0577,
      "step": 57100
    },
    {
      "epoch": 15.37952196382429,
      "grad_norm": 0.4302813708782196,
      "learning_rate": 0.00036956106099367167,
      "loss": 3.0501,
      "step": 57150
    },
    {
      "epoch": 15.392980189491817,
      "grad_norm": 0.3731640875339508,
      "learning_rate": 0.0003693590951932139,
      "loss": 3.0541,
      "step": 57200
    },
    {
      "epoch": 15.406438415159345,
      "grad_norm": 0.3728298842906952,
      "learning_rate": 0.0003691571293927561,
      "loss": 3.0713,
      "step": 57250
    },
    {
      "epoch": 15.419896640826874,
      "grad_norm": 0.3899688124656677,
      "learning_rate": 0.0003689551635922983,
      "loss": 3.0526,
      "step": 57300
    },
    {
      "epoch": 15.433354866494401,
      "grad_norm": 0.3961775302886963,
      "learning_rate": 0.00036875319779184056,
      "loss": 3.0579,
      "step": 57350
    },
    {
      "epoch": 15.44681309216193,
      "grad_norm": 0.3863848149776459,
      "learning_rate": 0.0003685512319913828,
      "loss": 3.0651,
      "step": 57400
    },
    {
      "epoch": 15.460271317829458,
      "grad_norm": 0.3688678741455078,
      "learning_rate": 0.000368349266190925,
      "loss": 3.07,
      "step": 57450
    },
    {
      "epoch": 15.473729543496985,
      "grad_norm": 0.3774935007095337,
      "learning_rate": 0.0003681473003904672,
      "loss": 3.0658,
      "step": 57500
    },
    {
      "epoch": 15.487187769164514,
      "grad_norm": 0.39579784870147705,
      "learning_rate": 0.0003679453345900094,
      "loss": 3.045,
      "step": 57550
    },
    {
      "epoch": 15.500645994832041,
      "grad_norm": 0.3768142759799957,
      "learning_rate": 0.0003677433687895516,
      "loss": 3.0567,
      "step": 57600
    },
    {
      "epoch": 15.514104220499568,
      "grad_norm": 0.36016687750816345,
      "learning_rate": 0.00036754140298909385,
      "loss": 3.0532,
      "step": 57650
    },
    {
      "epoch": 15.527562446167098,
      "grad_norm": 0.3630627989768982,
      "learning_rate": 0.00036733943718863604,
      "loss": 3.064,
      "step": 57700
    },
    {
      "epoch": 15.541020671834625,
      "grad_norm": 0.4007406532764435,
      "learning_rate": 0.00036713747138817824,
      "loss": 3.0607,
      "step": 57750
    },
    {
      "epoch": 15.554478897502154,
      "grad_norm": 0.3902122378349304,
      "learning_rate": 0.00036693550558772044,
      "loss": 3.068,
      "step": 57800
    },
    {
      "epoch": 15.567937123169681,
      "grad_norm": 0.39091888070106506,
      "learning_rate": 0.00036673353978726263,
      "loss": 3.0537,
      "step": 57850
    },
    {
      "epoch": 15.581395348837209,
      "grad_norm": 0.3762982189655304,
      "learning_rate": 0.0003665315739868049,
      "loss": 3.074,
      "step": 57900
    },
    {
      "epoch": 15.594853574504738,
      "grad_norm": 0.3759395182132721,
      "learning_rate": 0.0003663296081863471,
      "loss": 3.0703,
      "step": 57950
    },
    {
      "epoch": 15.608311800172265,
      "grad_norm": 0.4088388979434967,
      "learning_rate": 0.0003661276423858893,
      "loss": 3.0701,
      "step": 58000
    },
    {
      "epoch": 15.608311800172265,
      "eval_accuracy": 0.39394132982882035,
      "eval_loss": 3.2907497882843018,
      "eval_runtime": 83.3883,
      "eval_samples_per_second": 216.002,
      "eval_steps_per_second": 13.503,
      "step": 58000
    },
    {
      "epoch": 15.621770025839794,
      "grad_norm": 0.4267113506793976,
      "learning_rate": 0.00036592567658543147,
      "loss": 3.079,
      "step": 58050
    },
    {
      "epoch": 15.635228251507321,
      "grad_norm": 0.3806926906108856,
      "learning_rate": 0.00036572371078497367,
      "loss": 3.0623,
      "step": 58100
    },
    {
      "epoch": 15.648686477174849,
      "grad_norm": 0.38087034225463867,
      "learning_rate": 0.0003655217449845159,
      "loss": 3.0702,
      "step": 58150
    },
    {
      "epoch": 15.662144702842378,
      "grad_norm": 0.3637595474720001,
      "learning_rate": 0.0003653197791840581,
      "loss": 3.0801,
      "step": 58200
    },
    {
      "epoch": 15.675602928509905,
      "grad_norm": 0.37766537070274353,
      "learning_rate": 0.00036511781338360036,
      "loss": 3.0653,
      "step": 58250
    },
    {
      "epoch": 15.689061154177434,
      "grad_norm": 0.37575387954711914,
      "learning_rate": 0.00036491584758314256,
      "loss": 3.0805,
      "step": 58300
    },
    {
      "epoch": 15.702519379844961,
      "grad_norm": 0.38047921657562256,
      "learning_rate": 0.0003647138817826848,
      "loss": 3.0729,
      "step": 58350
    },
    {
      "epoch": 15.715977605512489,
      "grad_norm": 0.39415618777275085,
      "learning_rate": 0.000364511915982227,
      "loss": 3.0791,
      "step": 58400
    },
    {
      "epoch": 15.729435831180018,
      "grad_norm": 0.40556368231773376,
      "learning_rate": 0.0003643099501817692,
      "loss": 3.0657,
      "step": 58450
    },
    {
      "epoch": 15.742894056847545,
      "grad_norm": 0.41106295585632324,
      "learning_rate": 0.0003641079843813114,
      "loss": 3.0894,
      "step": 58500
    },
    {
      "epoch": 15.756352282515074,
      "grad_norm": 0.3911437392234802,
      "learning_rate": 0.0003639060185808536,
      "loss": 3.0875,
      "step": 58550
    },
    {
      "epoch": 15.769810508182601,
      "grad_norm": 0.3743394613265991,
      "learning_rate": 0.00036370405278039584,
      "loss": 3.0757,
      "step": 58600
    },
    {
      "epoch": 15.783268733850129,
      "grad_norm": 0.3689190149307251,
      "learning_rate": 0.00036350208697993804,
      "loss": 3.0773,
      "step": 58650
    },
    {
      "epoch": 15.796726959517658,
      "grad_norm": 0.4005644619464874,
      "learning_rate": 0.00036330012117948024,
      "loss": 3.0809,
      "step": 58700
    },
    {
      "epoch": 15.810185185185185,
      "grad_norm": 0.3765542507171631,
      "learning_rate": 0.00036309815537902243,
      "loss": 3.0807,
      "step": 58750
    },
    {
      "epoch": 15.823643410852712,
      "grad_norm": 0.3823806643486023,
      "learning_rate": 0.0003628961895785647,
      "loss": 3.0784,
      "step": 58800
    },
    {
      "epoch": 15.837101636520241,
      "grad_norm": 0.3556201457977295,
      "learning_rate": 0.0003626942237781069,
      "loss": 3.0851,
      "step": 58850
    },
    {
      "epoch": 15.850559862187769,
      "grad_norm": 0.3880035877227783,
      "learning_rate": 0.0003624922579776491,
      "loss": 3.0749,
      "step": 58900
    },
    {
      "epoch": 15.864018087855298,
      "grad_norm": 0.3721064627170563,
      "learning_rate": 0.00036229029217719127,
      "loss": 3.0836,
      "step": 58950
    },
    {
      "epoch": 15.877476313522825,
      "grad_norm": 0.3672582805156708,
      "learning_rate": 0.00036208832637673347,
      "loss": 3.0819,
      "step": 59000
    },
    {
      "epoch": 15.877476313522825,
      "eval_accuracy": 0.3944164425282036,
      "eval_loss": 3.284586191177368,
      "eval_runtime": 86.2284,
      "eval_samples_per_second": 208.887,
      "eval_steps_per_second": 13.058,
      "step": 59000
    },
    {
      "epoch": 15.890934539190352,
      "grad_norm": 0.3808155953884125,
      "learning_rate": 0.0003618863605762757,
      "loss": 3.078,
      "step": 59050
    },
    {
      "epoch": 15.904392764857882,
      "grad_norm": 0.38692107796669006,
      "learning_rate": 0.0003616843947758179,
      "loss": 3.0884,
      "step": 59100
    },
    {
      "epoch": 15.917850990525409,
      "grad_norm": 0.4037911593914032,
      "learning_rate": 0.00036148242897536016,
      "loss": 3.0807,
      "step": 59150
    },
    {
      "epoch": 15.931309216192938,
      "grad_norm": 0.4004386365413666,
      "learning_rate": 0.00036128046317490236,
      "loss": 3.0834,
      "step": 59200
    },
    {
      "epoch": 15.944767441860465,
      "grad_norm": 0.37851184606552124,
      "learning_rate": 0.0003610784973744446,
      "loss": 3.0883,
      "step": 59250
    },
    {
      "epoch": 15.958225667527993,
      "grad_norm": 0.38049548864364624,
      "learning_rate": 0.0003608765315739868,
      "loss": 3.0786,
      "step": 59300
    },
    {
      "epoch": 15.971683893195522,
      "grad_norm": 0.3860936760902405,
      "learning_rate": 0.000360674565773529,
      "loss": 3.0925,
      "step": 59350
    },
    {
      "epoch": 15.985142118863049,
      "grad_norm": 0.3885321617126465,
      "learning_rate": 0.0003604725999730712,
      "loss": 3.0794,
      "step": 59400
    },
    {
      "epoch": 15.998600344530576,
      "grad_norm": 0.3719334900379181,
      "learning_rate": 0.0003602706341726134,
      "loss": 3.0893,
      "step": 59450
    },
    {
      "epoch": 16.011843238587424,
      "grad_norm": 0.4153191149234772,
      "learning_rate": 0.00036006866837215564,
      "loss": 2.9898,
      "step": 59500
    },
    {
      "epoch": 16.02530146425495,
      "grad_norm": 0.37304985523223877,
      "learning_rate": 0.00035986670257169784,
      "loss": 2.9825,
      "step": 59550
    },
    {
      "epoch": 16.03875968992248,
      "grad_norm": 0.39161500334739685,
      "learning_rate": 0.00035966473677124004,
      "loss": 2.9904,
      "step": 59600
    },
    {
      "epoch": 16.05221791559001,
      "grad_norm": 0.36925414204597473,
      "learning_rate": 0.00035946277097078223,
      "loss": 3.0016,
      "step": 59650
    },
    {
      "epoch": 16.065676141257537,
      "grad_norm": 0.41601887345314026,
      "learning_rate": 0.00035926080517032443,
      "loss": 3.0069,
      "step": 59700
    },
    {
      "epoch": 16.079134366925064,
      "grad_norm": 0.40431052446365356,
      "learning_rate": 0.0003590588393698667,
      "loss": 3.002,
      "step": 59750
    },
    {
      "epoch": 16.09259259259259,
      "grad_norm": 0.40011581778526306,
      "learning_rate": 0.0003588568735694089,
      "loss": 3.0053,
      "step": 59800
    },
    {
      "epoch": 16.10605081826012,
      "grad_norm": 0.38849717378616333,
      "learning_rate": 0.00035865490776895107,
      "loss": 3.014,
      "step": 59850
    },
    {
      "epoch": 16.11950904392765,
      "grad_norm": 0.39223939180374146,
      "learning_rate": 0.00035845294196849327,
      "loss": 3.0027,
      "step": 59900
    },
    {
      "epoch": 16.132967269595177,
      "grad_norm": 0.399728000164032,
      "learning_rate": 0.00035825097616803546,
      "loss": 3.0086,
      "step": 59950
    },
    {
      "epoch": 16.146425495262704,
      "grad_norm": 0.3876127302646637,
      "learning_rate": 0.0003580490103675777,
      "loss": 3.0113,
      "step": 60000
    },
    {
      "epoch": 16.146425495262704,
      "eval_accuracy": 0.3938716871944036,
      "eval_loss": 3.299686908721924,
      "eval_runtime": 83.7969,
      "eval_samples_per_second": 214.948,
      "eval_steps_per_second": 13.437,
      "step": 60000
    },
    {
      "epoch": 16.15988372093023,
      "grad_norm": 0.38624656200408936,
      "learning_rate": 0.00035784704456711996,
      "loss": 3.0209,
      "step": 60050
    },
    {
      "epoch": 16.17334194659776,
      "grad_norm": 0.39339882135391235,
      "learning_rate": 0.00035764507876666216,
      "loss": 3.0209,
      "step": 60100
    },
    {
      "epoch": 16.18680017226529,
      "grad_norm": 0.3788904845714569,
      "learning_rate": 0.00035744311296620436,
      "loss": 3.0243,
      "step": 60150
    },
    {
      "epoch": 16.200258397932817,
      "grad_norm": 0.38240116834640503,
      "learning_rate": 0.0003572411471657466,
      "loss": 3.0134,
      "step": 60200
    },
    {
      "epoch": 16.213716623600344,
      "grad_norm": 0.41567012667655945,
      "learning_rate": 0.0003570391813652888,
      "loss": 3.028,
      "step": 60250
    },
    {
      "epoch": 16.227174849267872,
      "grad_norm": 0.37342318892478943,
      "learning_rate": 0.000356837215564831,
      "loss": 3.0323,
      "step": 60300
    },
    {
      "epoch": 16.2406330749354,
      "grad_norm": 0.40529847145080566,
      "learning_rate": 0.0003566352497643732,
      "loss": 3.0224,
      "step": 60350
    },
    {
      "epoch": 16.25409130060293,
      "grad_norm": 0.3941630423069,
      "learning_rate": 0.0003564332839639154,
      "loss": 3.0321,
      "step": 60400
    },
    {
      "epoch": 16.267549526270457,
      "grad_norm": 0.3903161287307739,
      "learning_rate": 0.00035623131816345764,
      "loss": 3.0398,
      "step": 60450
    },
    {
      "epoch": 16.281007751937985,
      "grad_norm": 0.44505786895751953,
      "learning_rate": 0.00035602935236299984,
      "loss": 3.037,
      "step": 60500
    },
    {
      "epoch": 16.294465977605512,
      "grad_norm": 0.39874106645584106,
      "learning_rate": 0.00035582738656254203,
      "loss": 3.0353,
      "step": 60550
    },
    {
      "epoch": 16.30792420327304,
      "grad_norm": 0.3974197804927826,
      "learning_rate": 0.00035562542076208423,
      "loss": 3.0441,
      "step": 60600
    },
    {
      "epoch": 16.32138242894057,
      "grad_norm": 0.40742236375808716,
      "learning_rate": 0.0003554234549616264,
      "loss": 3.0304,
      "step": 60650
    },
    {
      "epoch": 16.334840654608097,
      "grad_norm": 0.41979461908340454,
      "learning_rate": 0.0003552214891611687,
      "loss": 3.0415,
      "step": 60700
    },
    {
      "epoch": 16.348298880275625,
      "grad_norm": 0.3914986848831177,
      "learning_rate": 0.00035501952336071087,
      "loss": 3.0405,
      "step": 60750
    },
    {
      "epoch": 16.361757105943152,
      "grad_norm": 0.38728559017181396,
      "learning_rate": 0.00035481755756025307,
      "loss": 3.0378,
      "step": 60800
    },
    {
      "epoch": 16.37521533161068,
      "grad_norm": 0.41403523087501526,
      "learning_rate": 0.00035461559175979526,
      "loss": 3.0435,
      "step": 60850
    },
    {
      "epoch": 16.38867355727821,
      "grad_norm": 0.4332049787044525,
      "learning_rate": 0.0003544136259593375,
      "loss": 3.048,
      "step": 60900
    },
    {
      "epoch": 16.402131782945737,
      "grad_norm": 0.3733983635902405,
      "learning_rate": 0.00035421166015887977,
      "loss": 3.0381,
      "step": 60950
    },
    {
      "epoch": 16.415590008613265,
      "grad_norm": 0.3889237642288208,
      "learning_rate": 0.00035400969435842196,
      "loss": 3.051,
      "step": 61000
    },
    {
      "epoch": 16.415590008613265,
      "eval_accuracy": 0.3940545398523185,
      "eval_loss": 3.29353928565979,
      "eval_runtime": 83.9248,
      "eval_samples_per_second": 214.621,
      "eval_steps_per_second": 13.417,
      "step": 61000
    },
    {
      "epoch": 16.429048234280792,
      "grad_norm": 0.39081403613090515,
      "learning_rate": 0.00035380772855796416,
      "loss": 3.0478,
      "step": 61050
    },
    {
      "epoch": 16.44250645994832,
      "grad_norm": 0.3659651577472687,
      "learning_rate": 0.0003536057627575064,
      "loss": 3.0439,
      "step": 61100
    },
    {
      "epoch": 16.45596468561585,
      "grad_norm": 0.4061603844165802,
      "learning_rate": 0.0003534037969570486,
      "loss": 3.0454,
      "step": 61150
    },
    {
      "epoch": 16.469422911283377,
      "grad_norm": 0.38700902462005615,
      "learning_rate": 0.0003532018311565908,
      "loss": 3.0432,
      "step": 61200
    },
    {
      "epoch": 16.482881136950905,
      "grad_norm": 0.38399988412857056,
      "learning_rate": 0.000352999865356133,
      "loss": 3.0488,
      "step": 61250
    },
    {
      "epoch": 16.496339362618432,
      "grad_norm": 0.4063047170639038,
      "learning_rate": 0.0003527978995556752,
      "loss": 3.0418,
      "step": 61300
    },
    {
      "epoch": 16.50979758828596,
      "grad_norm": 0.3732617199420929,
      "learning_rate": 0.00035259593375521744,
      "loss": 3.0538,
      "step": 61350
    },
    {
      "epoch": 16.52325581395349,
      "grad_norm": 0.38415974378585815,
      "learning_rate": 0.00035239396795475964,
      "loss": 3.0442,
      "step": 61400
    },
    {
      "epoch": 16.536714039621017,
      "grad_norm": 0.4004860818386078,
      "learning_rate": 0.00035219200215430183,
      "loss": 3.0538,
      "step": 61450
    },
    {
      "epoch": 16.550172265288545,
      "grad_norm": 0.37568730115890503,
      "learning_rate": 0.00035199003635384403,
      "loss": 3.0554,
      "step": 61500
    },
    {
      "epoch": 16.563630490956072,
      "grad_norm": 0.3945615887641907,
      "learning_rate": 0.0003517880705533862,
      "loss": 3.0596,
      "step": 61550
    },
    {
      "epoch": 16.5770887166236,
      "grad_norm": 0.36448490619659424,
      "learning_rate": 0.0003515861047529285,
      "loss": 3.0586,
      "step": 61600
    },
    {
      "epoch": 16.590546942291127,
      "grad_norm": 0.39387741684913635,
      "learning_rate": 0.0003513841389524707,
      "loss": 3.0564,
      "step": 61650
    },
    {
      "epoch": 16.604005167958658,
      "grad_norm": 0.4100702404975891,
      "learning_rate": 0.00035118217315201287,
      "loss": 3.0571,
      "step": 61700
    },
    {
      "epoch": 16.617463393626185,
      "grad_norm": 0.3951115012168884,
      "learning_rate": 0.00035098020735155507,
      "loss": 3.0549,
      "step": 61750
    },
    {
      "epoch": 16.630921619293712,
      "grad_norm": 0.3702566921710968,
      "learning_rate": 0.00035077824155109737,
      "loss": 3.0636,
      "step": 61800
    },
    {
      "epoch": 16.64437984496124,
      "grad_norm": 0.40223056077957153,
      "learning_rate": 0.00035057627575063957,
      "loss": 3.0665,
      "step": 61850
    },
    {
      "epoch": 16.657838070628767,
      "grad_norm": 0.3815414309501648,
      "learning_rate": 0.00035037430995018176,
      "loss": 3.0504,
      "step": 61900
    },
    {
      "epoch": 16.671296296296298,
      "grad_norm": 0.38790836930274963,
      "learning_rate": 0.00035017234414972396,
      "loss": 3.0658,
      "step": 61950
    },
    {
      "epoch": 16.684754521963825,
      "grad_norm": 0.4092879891395569,
      "learning_rate": 0.00034997037834926615,
      "loss": 3.0578,
      "step": 62000
    },
    {
      "epoch": 16.684754521963825,
      "eval_accuracy": 0.3945778917555724,
      "eval_loss": 3.2841830253601074,
      "eval_runtime": 84.1034,
      "eval_samples_per_second": 214.165,
      "eval_steps_per_second": 13.388,
      "step": 62000
    },
    {
      "epoch": 16.698212747631352,
      "grad_norm": 0.4112163782119751,
      "learning_rate": 0.0003497684125488084,
      "loss": 3.0558,
      "step": 62050
    },
    {
      "epoch": 16.71167097329888,
      "grad_norm": 0.3894757032394409,
      "learning_rate": 0.0003495664467483506,
      "loss": 3.0564,
      "step": 62100
    },
    {
      "epoch": 16.725129198966407,
      "grad_norm": 0.392168790102005,
      "learning_rate": 0.0003493644809478928,
      "loss": 3.0548,
      "step": 62150
    },
    {
      "epoch": 16.738587424633938,
      "grad_norm": 0.38225874304771423,
      "learning_rate": 0.000349162515147435,
      "loss": 3.0542,
      "step": 62200
    },
    {
      "epoch": 16.752045650301465,
      "grad_norm": 0.36075863242149353,
      "learning_rate": 0.0003489605493469772,
      "loss": 3.0565,
      "step": 62250
    },
    {
      "epoch": 16.765503875968992,
      "grad_norm": 0.3718566596508026,
      "learning_rate": 0.00034875858354651944,
      "loss": 3.0648,
      "step": 62300
    },
    {
      "epoch": 16.77896210163652,
      "grad_norm": 0.41544222831726074,
      "learning_rate": 0.00034855661774606164,
      "loss": 3.0557,
      "step": 62350
    },
    {
      "epoch": 16.792420327304047,
      "grad_norm": 0.40402117371559143,
      "learning_rate": 0.00034835465194560383,
      "loss": 3.067,
      "step": 62400
    },
    {
      "epoch": 16.805878552971578,
      "grad_norm": 0.42445436120033264,
      "learning_rate": 0.00034815268614514603,
      "loss": 3.0731,
      "step": 62450
    },
    {
      "epoch": 16.819336778639105,
      "grad_norm": 0.3735466003417969,
      "learning_rate": 0.0003479507203446882,
      "loss": 3.0772,
      "step": 62500
    },
    {
      "epoch": 16.832795004306632,
      "grad_norm": 0.38858041167259216,
      "learning_rate": 0.0003477487545442305,
      "loss": 3.0736,
      "step": 62550
    },
    {
      "epoch": 16.84625322997416,
      "grad_norm": 0.3718245029449463,
      "learning_rate": 0.00034754678874377267,
      "loss": 3.0677,
      "step": 62600
    },
    {
      "epoch": 16.859711455641687,
      "grad_norm": 0.3866599202156067,
      "learning_rate": 0.00034734482294331487,
      "loss": 3.0696,
      "step": 62650
    },
    {
      "epoch": 16.873169681309218,
      "grad_norm": 0.3618182837963104,
      "learning_rate": 0.00034714285714285717,
      "loss": 3.0662,
      "step": 62700
    },
    {
      "epoch": 16.886627906976745,
      "grad_norm": 0.3741232454776764,
      "learning_rate": 0.00034694089134239937,
      "loss": 3.0677,
      "step": 62750
    },
    {
      "epoch": 16.900086132644272,
      "grad_norm": 0.4058481454849243,
      "learning_rate": 0.00034673892554194156,
      "loss": 3.0745,
      "step": 62800
    },
    {
      "epoch": 16.9135443583118,
      "grad_norm": 0.37404704093933105,
      "learning_rate": 0.00034653695974148376,
      "loss": 3.0665,
      "step": 62850
    },
    {
      "epoch": 16.927002583979327,
      "grad_norm": 0.3875318169593811,
      "learning_rate": 0.00034633499394102595,
      "loss": 3.0633,
      "step": 62900
    },
    {
      "epoch": 16.940460809646858,
      "grad_norm": 0.4056463837623596,
      "learning_rate": 0.0003461330281405682,
      "loss": 3.0624,
      "step": 62950
    },
    {
      "epoch": 16.953919035314385,
      "grad_norm": 0.3895954489707947,
      "learning_rate": 0.0003459310623401104,
      "loss": 3.065,
      "step": 63000
    },
    {
      "epoch": 16.953919035314385,
      "eval_accuracy": 0.3950468115841885,
      "eval_loss": 3.2801597118377686,
      "eval_runtime": 84.1836,
      "eval_samples_per_second": 213.961,
      "eval_steps_per_second": 13.376,
      "step": 63000
    },
    {
      "epoch": 16.967377260981912,
      "grad_norm": 0.4403444528579712,
      "learning_rate": 0.0003457290965396526,
      "loss": 3.0758,
      "step": 63050
    },
    {
      "epoch": 16.98083548664944,
      "grad_norm": 0.3798871636390686,
      "learning_rate": 0.0003455271307391948,
      "loss": 3.0797,
      "step": 63100
    },
    {
      "epoch": 16.994293712316967,
      "grad_norm": 0.39273741841316223,
      "learning_rate": 0.000345325164938737,
      "loss": 3.0703,
      "step": 63150
    },
    {
      "epoch": 17.007536606373815,
      "grad_norm": 0.43220987915992737,
      "learning_rate": 0.00034512319913827924,
      "loss": 3.0157,
      "step": 63200
    },
    {
      "epoch": 17.020994832041342,
      "grad_norm": 0.39727628231048584,
      "learning_rate": 0.00034492123333782144,
      "loss": 2.9755,
      "step": 63250
    },
    {
      "epoch": 17.034453057708873,
      "grad_norm": 0.3941941261291504,
      "learning_rate": 0.00034471926753736363,
      "loss": 2.9771,
      "step": 63300
    },
    {
      "epoch": 17.0479112833764,
      "grad_norm": 0.3985773026943207,
      "learning_rate": 0.00034451730173690583,
      "loss": 2.9746,
      "step": 63350
    },
    {
      "epoch": 17.061369509043928,
      "grad_norm": 0.4089372754096985,
      "learning_rate": 0.000344315335936448,
      "loss": 2.9802,
      "step": 63400
    },
    {
      "epoch": 17.074827734711455,
      "grad_norm": 0.37851646542549133,
      "learning_rate": 0.0003441133701359903,
      "loss": 2.9907,
      "step": 63450
    },
    {
      "epoch": 17.088285960378983,
      "grad_norm": 0.39602717757225037,
      "learning_rate": 0.00034391140433553247,
      "loss": 2.9954,
      "step": 63500
    },
    {
      "epoch": 17.101744186046513,
      "grad_norm": 0.37470999360084534,
      "learning_rate": 0.00034370943853507467,
      "loss": 2.9986,
      "step": 63550
    },
    {
      "epoch": 17.11520241171404,
      "grad_norm": 0.37743017077445984,
      "learning_rate": 0.0003435074727346169,
      "loss": 2.9932,
      "step": 63600
    },
    {
      "epoch": 17.128660637381568,
      "grad_norm": 0.40411996841430664,
      "learning_rate": 0.00034330550693415917,
      "loss": 2.9936,
      "step": 63650
    },
    {
      "epoch": 17.142118863049095,
      "grad_norm": 0.3877135217189789,
      "learning_rate": 0.00034310354113370136,
      "loss": 3.0072,
      "step": 63700
    },
    {
      "epoch": 17.155577088716623,
      "grad_norm": 0.4010166823863983,
      "learning_rate": 0.00034290157533324356,
      "loss": 3.0141,
      "step": 63750
    },
    {
      "epoch": 17.16903531438415,
      "grad_norm": 0.3890071213245392,
      "learning_rate": 0.00034269960953278576,
      "loss": 2.999,
      "step": 63800
    },
    {
      "epoch": 17.18249354005168,
      "grad_norm": 0.3823590576648712,
      "learning_rate": 0.00034249764373232795,
      "loss": 3.0136,
      "step": 63850
    },
    {
      "epoch": 17.195951765719208,
      "grad_norm": 0.39487892389297485,
      "learning_rate": 0.0003422956779318702,
      "loss": 3.0047,
      "step": 63900
    },
    {
      "epoch": 17.209409991386735,
      "grad_norm": 0.43239912390708923,
      "learning_rate": 0.0003420937121314124,
      "loss": 3.0127,
      "step": 63950
    },
    {
      "epoch": 17.222868217054263,
      "grad_norm": 0.39804336428642273,
      "learning_rate": 0.0003418917463309546,
      "loss": 3.0174,
      "step": 64000
    },
    {
      "epoch": 17.222868217054263,
      "eval_accuracy": 0.39433289309627456,
      "eval_loss": 3.2937676906585693,
      "eval_runtime": 84.2941,
      "eval_samples_per_second": 213.681,
      "eval_steps_per_second": 13.358,
      "step": 64000
    },
    {
      "epoch": 17.23632644272179,
      "grad_norm": 0.4015428423881531,
      "learning_rate": 0.0003416897805304968,
      "loss": 3.012,
      "step": 64050
    },
    {
      "epoch": 17.24978466838932,
      "grad_norm": 0.39993274211883545,
      "learning_rate": 0.000341487814730039,
      "loss": 3.0209,
      "step": 64100
    },
    {
      "epoch": 17.263242894056848,
      "grad_norm": 0.43030300736427307,
      "learning_rate": 0.00034128584892958124,
      "loss": 3.012,
      "step": 64150
    },
    {
      "epoch": 17.276701119724375,
      "grad_norm": 0.39758017659187317,
      "learning_rate": 0.00034108388312912343,
      "loss": 3.0147,
      "step": 64200
    },
    {
      "epoch": 17.290159345391903,
      "grad_norm": 0.4015481770038605,
      "learning_rate": 0.00034088191732866563,
      "loss": 3.0259,
      "step": 64250
    },
    {
      "epoch": 17.30361757105943,
      "grad_norm": 0.44094589352607727,
      "learning_rate": 0.0003406799515282078,
      "loss": 3.0314,
      "step": 64300
    },
    {
      "epoch": 17.31707579672696,
      "grad_norm": 0.4031984508037567,
      "learning_rate": 0.00034047798572775,
      "loss": 3.0234,
      "step": 64350
    },
    {
      "epoch": 17.330534022394488,
      "grad_norm": 0.4179072678089142,
      "learning_rate": 0.00034027601992729227,
      "loss": 3.0175,
      "step": 64400
    },
    {
      "epoch": 17.343992248062015,
      "grad_norm": 0.39360618591308594,
      "learning_rate": 0.00034007405412683447,
      "loss": 3.0304,
      "step": 64450
    },
    {
      "epoch": 17.357450473729543,
      "grad_norm": 0.3782019019126892,
      "learning_rate": 0.0003398720883263767,
      "loss": 3.0321,
      "step": 64500
    },
    {
      "epoch": 17.37090869939707,
      "grad_norm": 0.38941699266433716,
      "learning_rate": 0.00033967012252591897,
      "loss": 3.039,
      "step": 64550
    },
    {
      "epoch": 17.3843669250646,
      "grad_norm": 0.3850061297416687,
      "learning_rate": 0.00033946815672546116,
      "loss": 3.0367,
      "step": 64600
    },
    {
      "epoch": 17.397825150732128,
      "grad_norm": 0.39331889152526855,
      "learning_rate": 0.00033926619092500336,
      "loss": 3.0347,
      "step": 64650
    },
    {
      "epoch": 17.411283376399656,
      "grad_norm": 0.39823418855667114,
      "learning_rate": 0.00033906422512454556,
      "loss": 3.0254,
      "step": 64700
    },
    {
      "epoch": 17.424741602067183,
      "grad_norm": 0.4005449712276459,
      "learning_rate": 0.00033886225932408775,
      "loss": 3.021,
      "step": 64750
    },
    {
      "epoch": 17.43819982773471,
      "grad_norm": 0.3958333432674408,
      "learning_rate": 0.00033866029352363,
      "loss": 3.04,
      "step": 64800
    },
    {
      "epoch": 17.45165805340224,
      "grad_norm": 0.41515442728996277,
      "learning_rate": 0.0003384583277231722,
      "loss": 3.0348,
      "step": 64850
    },
    {
      "epoch": 17.46511627906977,
      "grad_norm": 0.39790430665016174,
      "learning_rate": 0.0003382563619227144,
      "loss": 3.0355,
      "step": 64900
    },
    {
      "epoch": 17.478574504737296,
      "grad_norm": 0.39022842049598694,
      "learning_rate": 0.0003380543961222566,
      "loss": 3.0234,
      "step": 64950
    },
    {
      "epoch": 17.492032730404823,
      "grad_norm": 0.3833022713661194,
      "learning_rate": 0.0003378524303217988,
      "loss": 3.0362,
      "step": 65000
    },
    {
      "epoch": 17.492032730404823,
      "eval_accuracy": 0.39422153006932104,
      "eval_loss": 3.2917582988739014,
      "eval_runtime": 84.1874,
      "eval_samples_per_second": 213.951,
      "eval_steps_per_second": 13.375,
      "step": 65000
    },
    {
      "epoch": 17.50549095607235,
      "grad_norm": 0.38307058811187744,
      "learning_rate": 0.00033765046452134104,
      "loss": 3.0399,
      "step": 65050
    },
    {
      "epoch": 17.51894918173988,
      "grad_norm": 0.38385099172592163,
      "learning_rate": 0.00033744849872088323,
      "loss": 3.0317,
      "step": 65100
    },
    {
      "epoch": 17.53240740740741,
      "grad_norm": 0.4235006272792816,
      "learning_rate": 0.00033724653292042543,
      "loss": 3.0362,
      "step": 65150
    },
    {
      "epoch": 17.545865633074936,
      "grad_norm": 0.3770826458930969,
      "learning_rate": 0.0003370445671199676,
      "loss": 3.0256,
      "step": 65200
    },
    {
      "epoch": 17.559323858742463,
      "grad_norm": 0.4297742545604706,
      "learning_rate": 0.0003368426013195098,
      "loss": 3.0339,
      "step": 65250
    },
    {
      "epoch": 17.57278208440999,
      "grad_norm": 0.4212315082550049,
      "learning_rate": 0.00033664063551905207,
      "loss": 3.0387,
      "step": 65300
    },
    {
      "epoch": 17.58624031007752,
      "grad_norm": 0.3919636011123657,
      "learning_rate": 0.00033643866971859427,
      "loss": 3.0402,
      "step": 65350
    },
    {
      "epoch": 17.59969853574505,
      "grad_norm": 0.427120566368103,
      "learning_rate": 0.0003362367039181365,
      "loss": 3.0405,
      "step": 65400
    },
    {
      "epoch": 17.613156761412576,
      "grad_norm": 0.3810177743434906,
      "learning_rate": 0.0003360347381176787,
      "loss": 3.0462,
      "step": 65450
    },
    {
      "epoch": 17.626614987080103,
      "grad_norm": 0.41916796565055847,
      "learning_rate": 0.00033583277231722097,
      "loss": 3.0445,
      "step": 65500
    },
    {
      "epoch": 17.64007321274763,
      "grad_norm": 0.3881797790527344,
      "learning_rate": 0.00033563080651676316,
      "loss": 3.0509,
      "step": 65550
    },
    {
      "epoch": 17.653531438415158,
      "grad_norm": 0.397521436214447,
      "learning_rate": 0.00033542884071630536,
      "loss": 3.0488,
      "step": 65600
    },
    {
      "epoch": 17.66698966408269,
      "grad_norm": 0.40210968255996704,
      "learning_rate": 0.00033522687491584755,
      "loss": 3.0561,
      "step": 65650
    },
    {
      "epoch": 17.680447889750216,
      "grad_norm": 0.4065394103527069,
      "learning_rate": 0.00033502490911538975,
      "loss": 3.038,
      "step": 65700
    },
    {
      "epoch": 17.693906115417743,
      "grad_norm": 0.3900739550590515,
      "learning_rate": 0.000334822943314932,
      "loss": 3.0569,
      "step": 65750
    },
    {
      "epoch": 17.70736434108527,
      "grad_norm": 0.3864552974700928,
      "learning_rate": 0.0003346209775144742,
      "loss": 3.0566,
      "step": 65800
    },
    {
      "epoch": 17.720822566752798,
      "grad_norm": 0.38607335090637207,
      "learning_rate": 0.0003344190117140164,
      "loss": 3.0517,
      "step": 65850
    },
    {
      "epoch": 17.73428079242033,
      "grad_norm": 0.40046197175979614,
      "learning_rate": 0.0003342170459135586,
      "loss": 3.0499,
      "step": 65900
    },
    {
      "epoch": 17.747739018087856,
      "grad_norm": 0.4442780613899231,
      "learning_rate": 0.0003340150801131008,
      "loss": 3.0564,
      "step": 65950
    },
    {
      "epoch": 17.761197243755383,
      "grad_norm": 0.3828612267971039,
      "learning_rate": 0.00033381311431264303,
      "loss": 3.0606,
      "step": 66000
    },
    {
      "epoch": 17.761197243755383,
      "eval_accuracy": 0.3952208638468027,
      "eval_loss": 3.2804038524627686,
      "eval_runtime": 84.3724,
      "eval_samples_per_second": 213.482,
      "eval_steps_per_second": 13.346,
      "step": 66000
    },
    {
      "epoch": 17.77465546942291,
      "grad_norm": 0.39253178238868713,
      "learning_rate": 0.00033361114851218523,
      "loss": 3.0497,
      "step": 66050
    },
    {
      "epoch": 17.788113695090438,
      "grad_norm": 0.4305365979671478,
      "learning_rate": 0.0003334091827117274,
      "loss": 3.0593,
      "step": 66100
    },
    {
      "epoch": 17.80157192075797,
      "grad_norm": 0.39081278443336487,
      "learning_rate": 0.0003332072169112696,
      "loss": 3.0569,
      "step": 66150
    },
    {
      "epoch": 17.815030146425496,
      "grad_norm": 0.39869746565818787,
      "learning_rate": 0.0003330052511108118,
      "loss": 3.0546,
      "step": 66200
    },
    {
      "epoch": 17.828488372093023,
      "grad_norm": 0.4080706536769867,
      "learning_rate": 0.0003328032853103541,
      "loss": 3.0462,
      "step": 66250
    },
    {
      "epoch": 17.84194659776055,
      "grad_norm": 0.3890393376350403,
      "learning_rate": 0.0003326013195098963,
      "loss": 3.055,
      "step": 66300
    },
    {
      "epoch": 17.855404823428078,
      "grad_norm": 0.38458430767059326,
      "learning_rate": 0.0003323993537094385,
      "loss": 3.0524,
      "step": 66350
    },
    {
      "epoch": 17.86886304909561,
      "grad_norm": 0.4325410723686218,
      "learning_rate": 0.00033219738790898077,
      "loss": 3.0612,
      "step": 66400
    },
    {
      "epoch": 17.882321274763136,
      "grad_norm": 0.4040071666240692,
      "learning_rate": 0.00033199542210852296,
      "loss": 3.0582,
      "step": 66450
    },
    {
      "epoch": 17.895779500430663,
      "grad_norm": 0.391096830368042,
      "learning_rate": 0.00033179345630806516,
      "loss": 3.0613,
      "step": 66500
    },
    {
      "epoch": 17.90923772609819,
      "grad_norm": 0.38018929958343506,
      "learning_rate": 0.00033159149050760735,
      "loss": 3.0525,
      "step": 66550
    },
    {
      "epoch": 17.922695951765718,
      "grad_norm": 0.3955528736114502,
      "learning_rate": 0.00033138952470714955,
      "loss": 3.0504,
      "step": 66600
    },
    {
      "epoch": 17.93615417743325,
      "grad_norm": 0.3869403302669525,
      "learning_rate": 0.0003311875589066918,
      "loss": 3.062,
      "step": 66650
    },
    {
      "epoch": 17.949612403100776,
      "grad_norm": 0.3936130702495575,
      "learning_rate": 0.000330985593106234,
      "loss": 3.0632,
      "step": 66700
    },
    {
      "epoch": 17.963070628768303,
      "grad_norm": 0.35933396220207214,
      "learning_rate": 0.0003307836273057762,
      "loss": 3.0597,
      "step": 66750
    },
    {
      "epoch": 17.97652885443583,
      "grad_norm": 0.37747249007225037,
      "learning_rate": 0.0003305816615053184,
      "loss": 3.0455,
      "step": 66800
    },
    {
      "epoch": 17.989987080103358,
      "grad_norm": 0.37566208839416504,
      "learning_rate": 0.0003303796957048606,
      "loss": 3.0704,
      "step": 66850
    },
    {
      "epoch": 18.003229974160206,
      "grad_norm": 0.3803272545337677,
      "learning_rate": 0.00033017772990440284,
      "loss": 3.0406,
      "step": 66900
    },
    {
      "epoch": 18.016688199827733,
      "grad_norm": 0.37502166628837585,
      "learning_rate": 0.00032997576410394503,
      "loss": 2.9641,
      "step": 66950
    },
    {
      "epoch": 18.030146425495264,
      "grad_norm": 0.4064948558807373,
      "learning_rate": 0.00032977379830348723,
      "loss": 2.9624,
      "step": 67000
    },
    {
      "epoch": 18.030146425495264,
      "eval_accuracy": 0.3943790680098895,
      "eval_loss": 3.2948076725006104,
      "eval_runtime": 84.2303,
      "eval_samples_per_second": 213.842,
      "eval_steps_per_second": 13.368,
      "step": 67000
    },
    {
      "epoch": 18.04360465116279,
      "grad_norm": 0.42101815342903137,
      "learning_rate": 0.0003295718325030294,
      "loss": 2.9775,
      "step": 67050
    },
    {
      "epoch": 18.05706287683032,
      "grad_norm": 0.393537312746048,
      "learning_rate": 0.0003293698667025716,
      "loss": 2.9704,
      "step": 67100
    },
    {
      "epoch": 18.070521102497846,
      "grad_norm": 0.412462443113327,
      "learning_rate": 0.0003291679009021139,
      "loss": 2.9863,
      "step": 67150
    },
    {
      "epoch": 18.083979328165373,
      "grad_norm": 0.39957910776138306,
      "learning_rate": 0.0003289659351016561,
      "loss": 2.9814,
      "step": 67200
    },
    {
      "epoch": 18.097437553832904,
      "grad_norm": 0.4040756821632385,
      "learning_rate": 0.0003287639693011983,
      "loss": 2.9795,
      "step": 67250
    },
    {
      "epoch": 18.11089577950043,
      "grad_norm": 0.3962806165218353,
      "learning_rate": 0.0003285620035007405,
      "loss": 2.9904,
      "step": 67300
    },
    {
      "epoch": 18.12435400516796,
      "grad_norm": 0.41427522897720337,
      "learning_rate": 0.00032836003770028276,
      "loss": 2.9795,
      "step": 67350
    },
    {
      "epoch": 18.137812230835486,
      "grad_norm": 0.42045581340789795,
      "learning_rate": 0.00032815807189982496,
      "loss": 2.9898,
      "step": 67400
    },
    {
      "epoch": 18.151270456503013,
      "grad_norm": 0.4016278088092804,
      "learning_rate": 0.00032795610609936715,
      "loss": 2.9928,
      "step": 67450
    },
    {
      "epoch": 18.164728682170544,
      "grad_norm": 0.3914085030555725,
      "learning_rate": 0.00032775414029890935,
      "loss": 2.9785,
      "step": 67500
    },
    {
      "epoch": 18.17818690783807,
      "grad_norm": 0.40669935941696167,
      "learning_rate": 0.00032755217449845155,
      "loss": 2.985,
      "step": 67550
    },
    {
      "epoch": 18.1916451335056,
      "grad_norm": 0.4045898914337158,
      "learning_rate": 0.0003273502086979938,
      "loss": 3.0007,
      "step": 67600
    },
    {
      "epoch": 18.205103359173126,
      "grad_norm": 0.4130690097808838,
      "learning_rate": 0.000327148242897536,
      "loss": 3.0049,
      "step": 67650
    },
    {
      "epoch": 18.218561584840653,
      "grad_norm": 0.39090612530708313,
      "learning_rate": 0.0003269462770970782,
      "loss": 3.0011,
      "step": 67700
    },
    {
      "epoch": 18.232019810508184,
      "grad_norm": 0.40354859828948975,
      "learning_rate": 0.0003267443112966204,
      "loss": 2.9969,
      "step": 67750
    },
    {
      "epoch": 18.24547803617571,
      "grad_norm": 0.3883422911167145,
      "learning_rate": 0.0003265423454961626,
      "loss": 3.0096,
      "step": 67800
    },
    {
      "epoch": 18.25893626184324,
      "grad_norm": 0.395225465297699,
      "learning_rate": 0.00032634037969570483,
      "loss": 3.0052,
      "step": 67850
    },
    {
      "epoch": 18.272394487510766,
      "grad_norm": 0.39701804518699646,
      "learning_rate": 0.00032613841389524703,
      "loss": 3.0102,
      "step": 67900
    },
    {
      "epoch": 18.285852713178294,
      "grad_norm": 0.41126975417137146,
      "learning_rate": 0.0003259364480947892,
      "loss": 3.0168,
      "step": 67950
    },
    {
      "epoch": 18.29931093884582,
      "grad_norm": 0.41036221385002136,
      "learning_rate": 0.0003257344822943314,
      "loss": 3.0074,
      "step": 68000
    },
    {
      "epoch": 18.29931093884582,
      "eval_accuracy": 0.39468697319855905,
      "eval_loss": 3.2947328090667725,
      "eval_runtime": 84.7379,
      "eval_samples_per_second": 212.561,
      "eval_steps_per_second": 13.288,
      "step": 68000
    },
    {
      "epoch": 18.31276916451335,
      "grad_norm": 0.42312413454055786,
      "learning_rate": 0.0003255325164938737,
      "loss": 3.0128,
      "step": 68050
    },
    {
      "epoch": 18.32622739018088,
      "grad_norm": 0.41836053133010864,
      "learning_rate": 0.0003253305506934159,
      "loss": 3.0062,
      "step": 68100
    },
    {
      "epoch": 18.339685615848406,
      "grad_norm": 0.38210856914520264,
      "learning_rate": 0.0003251285848929581,
      "loss": 3.0138,
      "step": 68150
    },
    {
      "epoch": 18.353143841515934,
      "grad_norm": 0.41910403966903687,
      "learning_rate": 0.0003249266190925003,
      "loss": 3.0095,
      "step": 68200
    },
    {
      "epoch": 18.36660206718346,
      "grad_norm": 0.3842230439186096,
      "learning_rate": 0.0003247246532920425,
      "loss": 3.0149,
      "step": 68250
    },
    {
      "epoch": 18.38006029285099,
      "grad_norm": 0.42878076434135437,
      "learning_rate": 0.00032452268749158476,
      "loss": 3.021,
      "step": 68300
    },
    {
      "epoch": 18.39351851851852,
      "grad_norm": 0.4169536530971527,
      "learning_rate": 0.00032432072169112696,
      "loss": 3.0203,
      "step": 68350
    },
    {
      "epoch": 18.406976744186046,
      "grad_norm": 0.410892128944397,
      "learning_rate": 0.00032411875589066915,
      "loss": 3.0285,
      "step": 68400
    },
    {
      "epoch": 18.420434969853574,
      "grad_norm": 0.3849635422229767,
      "learning_rate": 0.00032391679009021135,
      "loss": 3.0302,
      "step": 68450
    },
    {
      "epoch": 18.4338931955211,
      "grad_norm": 0.43162012100219727,
      "learning_rate": 0.0003237148242897536,
      "loss": 3.0232,
      "step": 68500
    },
    {
      "epoch": 18.447351421188632,
      "grad_norm": 0.4144338071346283,
      "learning_rate": 0.0003235128584892958,
      "loss": 3.0151,
      "step": 68550
    },
    {
      "epoch": 18.46080964685616,
      "grad_norm": 0.37732717394828796,
      "learning_rate": 0.000323310892688838,
      "loss": 3.0245,
      "step": 68600
    },
    {
      "epoch": 18.474267872523686,
      "grad_norm": 0.43414679169654846,
      "learning_rate": 0.0003231089268883802,
      "loss": 3.0285,
      "step": 68650
    },
    {
      "epoch": 18.487726098191214,
      "grad_norm": 0.38566941022872925,
      "learning_rate": 0.0003229069610879224,
      "loss": 3.0251,
      "step": 68700
    },
    {
      "epoch": 18.50118432385874,
      "grad_norm": 0.4028451442718506,
      "learning_rate": 0.00032270499528746463,
      "loss": 3.0339,
      "step": 68750
    },
    {
      "epoch": 18.514642549526272,
      "grad_norm": 0.40648922324180603,
      "learning_rate": 0.00032250302948700683,
      "loss": 3.0373,
      "step": 68800
    },
    {
      "epoch": 18.5281007751938,
      "grad_norm": 0.40535131096839905,
      "learning_rate": 0.000322301063686549,
      "loss": 3.0345,
      "step": 68850
    },
    {
      "epoch": 18.541559000861326,
      "grad_norm": 0.4239271283149719,
      "learning_rate": 0.0003220990978860912,
      "loss": 3.0323,
      "step": 68900
    },
    {
      "epoch": 18.555017226528854,
      "grad_norm": 0.38849544525146484,
      "learning_rate": 0.0003218971320856335,
      "loss": 3.0314,
      "step": 68950
    },
    {
      "epoch": 18.56847545219638,
      "grad_norm": 0.39953166246414185,
      "learning_rate": 0.0003216951662851757,
      "loss": 3.0378,
      "step": 69000
    },
    {
      "epoch": 18.56847545219638,
      "eval_accuracy": 0.3947981189318015,
      "eval_loss": 3.286586284637451,
      "eval_runtime": 84.2338,
      "eval_samples_per_second": 213.833,
      "eval_steps_per_second": 13.368,
      "step": 69000
    },
    {
      "epoch": 18.581933677863912,
      "grad_norm": 0.45566827058792114,
      "learning_rate": 0.0003214932004847179,
      "loss": 3.0289,
      "step": 69050
    },
    {
      "epoch": 18.59539190353144,
      "grad_norm": 0.4569692015647888,
      "learning_rate": 0.0003212912346842601,
      "loss": 3.0434,
      "step": 69100
    },
    {
      "epoch": 18.608850129198967,
      "grad_norm": 0.4200350046157837,
      "learning_rate": 0.0003210892688838023,
      "loss": 3.036,
      "step": 69150
    },
    {
      "epoch": 18.622308354866494,
      "grad_norm": 0.3737678825855255,
      "learning_rate": 0.00032088730308334456,
      "loss": 3.0303,
      "step": 69200
    },
    {
      "epoch": 18.63576658053402,
      "grad_norm": 0.3913853168487549,
      "learning_rate": 0.00032068533728288676,
      "loss": 3.0362,
      "step": 69250
    },
    {
      "epoch": 18.649224806201552,
      "grad_norm": 0.399463951587677,
      "learning_rate": 0.00032048337148242895,
      "loss": 3.0295,
      "step": 69300
    },
    {
      "epoch": 18.66268303186908,
      "grad_norm": 0.3811693489551544,
      "learning_rate": 0.00032028140568197115,
      "loss": 3.0338,
      "step": 69350
    },
    {
      "epoch": 18.676141257536607,
      "grad_norm": 0.43641650676727295,
      "learning_rate": 0.00032007943988151334,
      "loss": 3.0281,
      "step": 69400
    },
    {
      "epoch": 18.689599483204134,
      "grad_norm": 0.4464429020881653,
      "learning_rate": 0.0003198774740810556,
      "loss": 3.0292,
      "step": 69450
    },
    {
      "epoch": 18.70305770887166,
      "grad_norm": 0.40374037623405457,
      "learning_rate": 0.0003196755082805978,
      "loss": 3.0286,
      "step": 69500
    },
    {
      "epoch": 18.716515934539192,
      "grad_norm": 0.3943737745285034,
      "learning_rate": 0.00031947354248014,
      "loss": 3.036,
      "step": 69550
    },
    {
      "epoch": 18.72997416020672,
      "grad_norm": 0.38929787278175354,
      "learning_rate": 0.0003192715766796822,
      "loss": 3.0363,
      "step": 69600
    },
    {
      "epoch": 18.743432385874247,
      "grad_norm": 0.4019330143928528,
      "learning_rate": 0.0003190696108792244,
      "loss": 3.0236,
      "step": 69650
    },
    {
      "epoch": 18.756890611541774,
      "grad_norm": 0.3955172300338745,
      "learning_rate": 0.00031886764507876663,
      "loss": 3.0431,
      "step": 69700
    },
    {
      "epoch": 18.7703488372093,
      "grad_norm": 0.41721200942993164,
      "learning_rate": 0.0003186656792783088,
      "loss": 3.0335,
      "step": 69750
    },
    {
      "epoch": 18.783807062876832,
      "grad_norm": 0.40068718791007996,
      "learning_rate": 0.000318463713477851,
      "loss": 3.0402,
      "step": 69800
    },
    {
      "epoch": 18.79726528854436,
      "grad_norm": 0.40066292881965637,
      "learning_rate": 0.00031826174767739327,
      "loss": 3.0358,
      "step": 69850
    },
    {
      "epoch": 18.810723514211887,
      "grad_norm": 0.39857134222984314,
      "learning_rate": 0.0003180597818769355,
      "loss": 3.0422,
      "step": 69900
    },
    {
      "epoch": 18.824181739879414,
      "grad_norm": 0.4016108214855194,
      "learning_rate": 0.0003178578160764777,
      "loss": 3.0368,
      "step": 69950
    },
    {
      "epoch": 18.83763996554694,
      "grad_norm": 0.3944569230079651,
      "learning_rate": 0.0003176558502760199,
      "loss": 3.0442,
      "step": 70000
    },
    {
      "epoch": 18.83763996554694,
      "eval_accuracy": 0.3956260079712025,
      "eval_loss": 3.279024124145508,
      "eval_runtime": 84.4934,
      "eval_samples_per_second": 213.177,
      "eval_steps_per_second": 13.326,
      "step": 70000
    },
    {
      "epoch": 18.85109819121447,
      "grad_norm": 0.3991747796535492,
      "learning_rate": 0.0003174538844755621,
      "loss": 3.0441,
      "step": 70050
    },
    {
      "epoch": 18.864556416882,
      "grad_norm": 0.4364408850669861,
      "learning_rate": 0.0003172519186751043,
      "loss": 3.0552,
      "step": 70100
    },
    {
      "epoch": 18.878014642549527,
      "grad_norm": 0.4032851457595825,
      "learning_rate": 0.00031704995287464656,
      "loss": 3.0501,
      "step": 70150
    },
    {
      "epoch": 18.891472868217054,
      "grad_norm": 0.41202491521835327,
      "learning_rate": 0.00031684798707418875,
      "loss": 3.0385,
      "step": 70200
    },
    {
      "epoch": 18.90493109388458,
      "grad_norm": 0.385976642370224,
      "learning_rate": 0.00031664602127373095,
      "loss": 3.045,
      "step": 70250
    },
    {
      "epoch": 18.91838931955211,
      "grad_norm": 0.3908829689025879,
      "learning_rate": 0.00031644405547327315,
      "loss": 3.049,
      "step": 70300
    },
    {
      "epoch": 18.93184754521964,
      "grad_norm": 0.4331992566585541,
      "learning_rate": 0.00031624208967281534,
      "loss": 3.0528,
      "step": 70350
    },
    {
      "epoch": 18.945305770887167,
      "grad_norm": 0.3863152861595154,
      "learning_rate": 0.0003160401238723576,
      "loss": 3.0488,
      "step": 70400
    },
    {
      "epoch": 18.958763996554694,
      "grad_norm": 0.3831482529640198,
      "learning_rate": 0.0003158381580718998,
      "loss": 3.0486,
      "step": 70450
    },
    {
      "epoch": 18.97222222222222,
      "grad_norm": 0.43002840876579285,
      "learning_rate": 0.000315636192271442,
      "loss": 3.0364,
      "step": 70500
    },
    {
      "epoch": 18.98568044788975,
      "grad_norm": 0.4217604100704193,
      "learning_rate": 0.0003154342264709842,
      "loss": 3.0445,
      "step": 70550
    },
    {
      "epoch": 18.99913867355728,
      "grad_norm": 0.3787181079387665,
      "learning_rate": 0.00031523226067052643,
      "loss": 3.0433,
      "step": 70600
    },
    {
      "epoch": 19.012381567614124,
      "grad_norm": 0.3988605737686157,
      "learning_rate": 0.0003150302948700686,
      "loss": 2.9634,
      "step": 70650
    },
    {
      "epoch": 19.025839793281655,
      "grad_norm": 0.4285053610801697,
      "learning_rate": 0.0003148283290696108,
      "loss": 2.9314,
      "step": 70700
    },
    {
      "epoch": 19.039298018949182,
      "grad_norm": 0.38082054257392883,
      "learning_rate": 0.0003146263632691531,
      "loss": 2.9725,
      "step": 70750
    },
    {
      "epoch": 19.05275624461671,
      "grad_norm": 0.4119919538497925,
      "learning_rate": 0.0003144243974686953,
      "loss": 2.9737,
      "step": 70800
    },
    {
      "epoch": 19.066214470284237,
      "grad_norm": 0.39995309710502625,
      "learning_rate": 0.0003142224316682375,
      "loss": 2.9707,
      "step": 70850
    },
    {
      "epoch": 19.079672695951764,
      "grad_norm": 0.4222920835018158,
      "learning_rate": 0.0003140204658677797,
      "loss": 2.9676,
      "step": 70900
    },
    {
      "epoch": 19.093130921619295,
      "grad_norm": 0.408886581659317,
      "learning_rate": 0.0003138185000673219,
      "loss": 2.9693,
      "step": 70950
    },
    {
      "epoch": 19.106589147286822,
      "grad_norm": 0.404677152633667,
      "learning_rate": 0.0003136165342668641,
      "loss": 2.9767,
      "step": 71000
    },
    {
      "epoch": 19.106589147286822,
      "eval_accuracy": 0.3946647005931684,
      "eval_loss": 3.2947773933410645,
      "eval_runtime": 84.5205,
      "eval_samples_per_second": 213.108,
      "eval_steps_per_second": 13.322,
      "step": 71000
    },
    {
      "epoch": 19.12004737295435,
      "grad_norm": 0.4102245271205902,
      "learning_rate": 0.00031341456846640636,
      "loss": 2.9824,
      "step": 71050
    },
    {
      "epoch": 19.133505598621877,
      "grad_norm": 0.4353906810283661,
      "learning_rate": 0.00031321260266594855,
      "loss": 2.9786,
      "step": 71100
    },
    {
      "epoch": 19.146963824289404,
      "grad_norm": 0.4163901209831238,
      "learning_rate": 0.00031301063686549075,
      "loss": 2.9813,
      "step": 71150
    },
    {
      "epoch": 19.160422049956935,
      "grad_norm": 0.4259953200817108,
      "learning_rate": 0.00031280867106503295,
      "loss": 2.9928,
      "step": 71200
    },
    {
      "epoch": 19.173880275624462,
      "grad_norm": 0.4333082437515259,
      "learning_rate": 0.00031260670526457514,
      "loss": 2.9873,
      "step": 71250
    },
    {
      "epoch": 19.18733850129199,
      "grad_norm": 0.4248490631580353,
      "learning_rate": 0.0003124047394641174,
      "loss": 2.9902,
      "step": 71300
    },
    {
      "epoch": 19.200796726959517,
      "grad_norm": 0.3874186873435974,
      "learning_rate": 0.0003122027736636596,
      "loss": 2.9904,
      "step": 71350
    },
    {
      "epoch": 19.214254952627044,
      "grad_norm": 0.4307815730571747,
      "learning_rate": 0.0003120008078632018,
      "loss": 3.0034,
      "step": 71400
    },
    {
      "epoch": 19.227713178294575,
      "grad_norm": 0.40104547142982483,
      "learning_rate": 0.000311798842062744,
      "loss": 2.9947,
      "step": 71450
    },
    {
      "epoch": 19.241171403962102,
      "grad_norm": 0.41799429059028625,
      "learning_rate": 0.0003115968762622862,
      "loss": 2.9995,
      "step": 71500
    },
    {
      "epoch": 19.25462962962963,
      "grad_norm": 0.409606009721756,
      "learning_rate": 0.00031139491046182843,
      "loss": 2.9842,
      "step": 71550
    },
    {
      "epoch": 19.268087855297157,
      "grad_norm": 0.4212263524532318,
      "learning_rate": 0.0003111929446613707,
      "loss": 2.9857,
      "step": 71600
    },
    {
      "epoch": 19.281546080964684,
      "grad_norm": 0.4294273555278778,
      "learning_rate": 0.0003109909788609129,
      "loss": 2.9913,
      "step": 71650
    },
    {
      "epoch": 19.295004306632215,
      "grad_norm": 0.42206907272338867,
      "learning_rate": 0.00031078901306045507,
      "loss": 3.0016,
      "step": 71700
    },
    {
      "epoch": 19.308462532299743,
      "grad_norm": 0.40087857842445374,
      "learning_rate": 0.0003105870472599973,
      "loss": 2.9957,
      "step": 71750
    },
    {
      "epoch": 19.32192075796727,
      "grad_norm": 0.41836878657341003,
      "learning_rate": 0.0003103850814595395,
      "loss": 2.9951,
      "step": 71800
    },
    {
      "epoch": 19.335378983634797,
      "grad_norm": 0.39714816212654114,
      "learning_rate": 0.0003101831156590817,
      "loss": 3.0037,
      "step": 71850
    },
    {
      "epoch": 19.348837209302324,
      "grad_norm": 0.395064115524292,
      "learning_rate": 0.0003099811498586239,
      "loss": 3.0026,
      "step": 71900
    },
    {
      "epoch": 19.362295434969855,
      "grad_norm": 0.44047603011131287,
      "learning_rate": 0.0003097791840581661,
      "loss": 3.0041,
      "step": 71950
    },
    {
      "epoch": 19.375753660637383,
      "grad_norm": 0.41249942779541016,
      "learning_rate": 0.00030957721825770836,
      "loss": 2.9956,
      "step": 72000
    },
    {
      "epoch": 19.375753660637383,
      "eval_accuracy": 0.3946752393381581,
      "eval_loss": 3.2918050289154053,
      "eval_runtime": 84.3014,
      "eval_samples_per_second": 213.662,
      "eval_steps_per_second": 13.357,
      "step": 72000
    },
    {
      "epoch": 19.38921188630491,
      "grad_norm": 0.41538962721824646,
      "learning_rate": 0.00030937525245725055,
      "loss": 3.0063,
      "step": 72050
    },
    {
      "epoch": 19.402670111972437,
      "grad_norm": 0.46369507908821106,
      "learning_rate": 0.00030917328665679275,
      "loss": 2.9987,
      "step": 72100
    },
    {
      "epoch": 19.416128337639964,
      "grad_norm": 0.43894675374031067,
      "learning_rate": 0.00030897132085633494,
      "loss": 3.0218,
      "step": 72150
    },
    {
      "epoch": 19.429586563307495,
      "grad_norm": 0.40366244316101074,
      "learning_rate": 0.00030876935505587714,
      "loss": 2.9982,
      "step": 72200
    },
    {
      "epoch": 19.443044788975023,
      "grad_norm": 0.44417864084243774,
      "learning_rate": 0.0003085673892554194,
      "loss": 3.0062,
      "step": 72250
    },
    {
      "epoch": 19.45650301464255,
      "grad_norm": 0.40489599108695984,
      "learning_rate": 0.0003083654234549616,
      "loss": 3.0149,
      "step": 72300
    },
    {
      "epoch": 19.469961240310077,
      "grad_norm": 0.38643038272857666,
      "learning_rate": 0.0003081634576545038,
      "loss": 3.0139,
      "step": 72350
    },
    {
      "epoch": 19.483419465977605,
      "grad_norm": 0.40237957239151,
      "learning_rate": 0.000307961491854046,
      "loss": 3.0093,
      "step": 72400
    },
    {
      "epoch": 19.496877691645132,
      "grad_norm": 0.43354129791259766,
      "learning_rate": 0.00030775952605358823,
      "loss": 3.0041,
      "step": 72450
    },
    {
      "epoch": 19.510335917312663,
      "grad_norm": 0.4100993275642395,
      "learning_rate": 0.0003075575602531305,
      "loss": 3.0127,
      "step": 72500
    },
    {
      "epoch": 19.52379414298019,
      "grad_norm": 0.39991486072540283,
      "learning_rate": 0.0003073555944526727,
      "loss": 3.0125,
      "step": 72550
    },
    {
      "epoch": 19.537252368647717,
      "grad_norm": 0.4252054989337921,
      "learning_rate": 0.00030715362865221487,
      "loss": 3.0321,
      "step": 72600
    },
    {
      "epoch": 19.550710594315245,
      "grad_norm": 0.4369717836380005,
      "learning_rate": 0.0003069516628517571,
      "loss": 3.0176,
      "step": 72650
    },
    {
      "epoch": 19.564168819982772,
      "grad_norm": 0.43274322152137756,
      "learning_rate": 0.0003067496970512993,
      "loss": 3.0243,
      "step": 72700
    },
    {
      "epoch": 19.577627045650303,
      "grad_norm": 0.4220770001411438,
      "learning_rate": 0.0003065477312508415,
      "loss": 3.0169,
      "step": 72750
    },
    {
      "epoch": 19.59108527131783,
      "grad_norm": 0.4007902443408966,
      "learning_rate": 0.0003063457654503837,
      "loss": 3.0172,
      "step": 72800
    },
    {
      "epoch": 19.604543496985357,
      "grad_norm": 0.4364423453807831,
      "learning_rate": 0.0003061437996499259,
      "loss": 3.0243,
      "step": 72850
    },
    {
      "epoch": 19.618001722652885,
      "grad_norm": 0.415723979473114,
      "learning_rate": 0.00030594183384946816,
      "loss": 3.004,
      "step": 72900
    },
    {
      "epoch": 19.631459948320412,
      "grad_norm": 0.4381922483444214,
      "learning_rate": 0.00030573986804901035,
      "loss": 3.0225,
      "step": 72950
    },
    {
      "epoch": 19.644918173987943,
      "grad_norm": 0.4394854009151459,
      "learning_rate": 0.00030553790224855255,
      "loss": 3.0241,
      "step": 73000
    },
    {
      "epoch": 19.644918173987943,
      "eval_accuracy": 0.39559471767679993,
      "eval_loss": 3.2832705974578857,
      "eval_runtime": 84.3768,
      "eval_samples_per_second": 213.471,
      "eval_steps_per_second": 13.345,
      "step": 73000
    },
    {
      "epoch": 19.65837639965547,
      "grad_norm": 0.38926881551742554,
      "learning_rate": 0.00030533593644809474,
      "loss": 3.0254,
      "step": 73050
    },
    {
      "epoch": 19.671834625322997,
      "grad_norm": 0.41295066475868225,
      "learning_rate": 0.00030513397064763694,
      "loss": 3.0199,
      "step": 73100
    },
    {
      "epoch": 19.685292850990525,
      "grad_norm": 0.3994574546813965,
      "learning_rate": 0.0003049320048471792,
      "loss": 3.0203,
      "step": 73150
    },
    {
      "epoch": 19.698751076658052,
      "grad_norm": 0.39267680048942566,
      "learning_rate": 0.0003047300390467214,
      "loss": 3.0137,
      "step": 73200
    },
    {
      "epoch": 19.712209302325583,
      "grad_norm": 0.4108363687992096,
      "learning_rate": 0.0003045280732462636,
      "loss": 3.0187,
      "step": 73250
    },
    {
      "epoch": 19.72566752799311,
      "grad_norm": 0.39049604535102844,
      "learning_rate": 0.0003043261074458058,
      "loss": 3.0274,
      "step": 73300
    },
    {
      "epoch": 19.739125753660637,
      "grad_norm": 0.425126314163208,
      "learning_rate": 0.000304124141645348,
      "loss": 3.0236,
      "step": 73350
    },
    {
      "epoch": 19.752583979328165,
      "grad_norm": 0.3977421224117279,
      "learning_rate": 0.0003039221758448903,
      "loss": 3.0182,
      "step": 73400
    },
    {
      "epoch": 19.766042204995692,
      "grad_norm": 0.41254788637161255,
      "learning_rate": 0.0003037202100444325,
      "loss": 3.0432,
      "step": 73450
    },
    {
      "epoch": 19.779500430663223,
      "grad_norm": 0.44152772426605225,
      "learning_rate": 0.00030351824424397467,
      "loss": 3.0239,
      "step": 73500
    },
    {
      "epoch": 19.79295865633075,
      "grad_norm": 0.4028533399105072,
      "learning_rate": 0.00030331627844351687,
      "loss": 3.0262,
      "step": 73550
    },
    {
      "epoch": 19.806416881998278,
      "grad_norm": 0.4171806871891022,
      "learning_rate": 0.0003031143126430591,
      "loss": 3.0284,
      "step": 73600
    },
    {
      "epoch": 19.819875107665805,
      "grad_norm": 0.4112701117992401,
      "learning_rate": 0.0003029123468426013,
      "loss": 3.0257,
      "step": 73650
    },
    {
      "epoch": 19.833333333333332,
      "grad_norm": 0.3806881904602051,
      "learning_rate": 0.0003027103810421435,
      "loss": 3.0291,
      "step": 73700
    },
    {
      "epoch": 19.846791559000863,
      "grad_norm": 0.40808114409446716,
      "learning_rate": 0.0003025084152416857,
      "loss": 3.0287,
      "step": 73750
    },
    {
      "epoch": 19.86024978466839,
      "grad_norm": 0.3911757469177246,
      "learning_rate": 0.0003023064494412279,
      "loss": 3.0346,
      "step": 73800
    },
    {
      "epoch": 19.873708010335918,
      "grad_norm": 0.44232165813446045,
      "learning_rate": 0.00030210448364077015,
      "loss": 3.0384,
      "step": 73850
    },
    {
      "epoch": 19.887166236003445,
      "grad_norm": 0.43855389952659607,
      "learning_rate": 0.00030190251784031235,
      "loss": 3.0471,
      "step": 73900
    },
    {
      "epoch": 19.900624461670972,
      "grad_norm": 0.4235590696334839,
      "learning_rate": 0.00030170055203985454,
      "loss": 3.0415,
      "step": 73950
    },
    {
      "epoch": 19.9140826873385,
      "grad_norm": 0.4630587100982666,
      "learning_rate": 0.00030149858623939674,
      "loss": 3.036,
      "step": 74000
    },
    {
      "epoch": 19.9140826873385,
      "eval_accuracy": 0.3960048595565557,
      "eval_loss": 3.278202533721924,
      "eval_runtime": 84.2989,
      "eval_samples_per_second": 213.668,
      "eval_steps_per_second": 13.357,
      "step": 74000
    },
    {
      "epoch": 19.92754091300603,
      "grad_norm": 0.42282310128211975,
      "learning_rate": 0.00030129662043893894,
      "loss": 3.0379,
      "step": 74050
    },
    {
      "epoch": 19.940999138673558,
      "grad_norm": 0.401750385761261,
      "learning_rate": 0.0003010946546384812,
      "loss": 3.0293,
      "step": 74100
    },
    {
      "epoch": 19.954457364341085,
      "grad_norm": 0.38794606924057007,
      "learning_rate": 0.0003008926888380234,
      "loss": 3.0469,
      "step": 74150
    },
    {
      "epoch": 19.967915590008612,
      "grad_norm": 0.4405243694782257,
      "learning_rate": 0.0003006907230375656,
      "loss": 3.0427,
      "step": 74200
    },
    {
      "epoch": 19.98137381567614,
      "grad_norm": 0.391490638256073,
      "learning_rate": 0.0003004887572371078,
      "loss": 3.0396,
      "step": 74250
    },
    {
      "epoch": 19.99483204134367,
      "grad_norm": 0.4083889126777649,
      "learning_rate": 0.0003002867914366501,
      "loss": 3.0334,
      "step": 74300
    },
    {
      "epoch": 20.00807493540052,
      "grad_norm": 0.3883116543292999,
      "learning_rate": 0.0003000848256361923,
      "loss": 2.9895,
      "step": 74350
    },
    {
      "epoch": 20.021533161068046,
      "grad_norm": 0.4304565489292145,
      "learning_rate": 0.00029988285983573447,
      "loss": 2.9404,
      "step": 74400
    },
    {
      "epoch": 20.034991386735573,
      "grad_norm": 0.3939877450466156,
      "learning_rate": 0.00029968089403527667,
      "loss": 2.9541,
      "step": 74450
    },
    {
      "epoch": 20.0484496124031,
      "grad_norm": 0.43768543004989624,
      "learning_rate": 0.0002994789282348189,
      "loss": 2.9494,
      "step": 74500
    },
    {
      "epoch": 20.061907838070628,
      "grad_norm": 0.4543677866458893,
      "learning_rate": 0.0002992769624343611,
      "loss": 2.9591,
      "step": 74550
    },
    {
      "epoch": 20.075366063738155,
      "grad_norm": 0.41045457124710083,
      "learning_rate": 0.0002990749966339033,
      "loss": 2.9561,
      "step": 74600
    },
    {
      "epoch": 20.088824289405686,
      "grad_norm": 0.42093926668167114,
      "learning_rate": 0.0002988730308334455,
      "loss": 2.9635,
      "step": 74650
    },
    {
      "epoch": 20.102282515073213,
      "grad_norm": 0.40229475498199463,
      "learning_rate": 0.0002986710650329877,
      "loss": 2.9509,
      "step": 74700
    },
    {
      "epoch": 20.11574074074074,
      "grad_norm": 0.43167799711227417,
      "learning_rate": 0.00029846909923252995,
      "loss": 2.9681,
      "step": 74750
    },
    {
      "epoch": 20.129198966408268,
      "grad_norm": 0.42425838112831116,
      "learning_rate": 0.00029826713343207215,
      "loss": 2.9664,
      "step": 74800
    },
    {
      "epoch": 20.142657192075795,
      "grad_norm": 0.4355631172657013,
      "learning_rate": 0.00029806516763161435,
      "loss": 2.9735,
      "step": 74850
    },
    {
      "epoch": 20.156115417743326,
      "grad_norm": 0.39577028155326843,
      "learning_rate": 0.00029786320183115654,
      "loss": 2.971,
      "step": 74900
    },
    {
      "epoch": 20.169573643410853,
      "grad_norm": 0.39700061082839966,
      "learning_rate": 0.0002976612360306988,
      "loss": 2.9752,
      "step": 74950
    },
    {
      "epoch": 20.18303186907838,
      "grad_norm": 0.4195544719696045,
      "learning_rate": 0.000297459270230241,
      "loss": 2.9768,
      "step": 75000
    },
    {
      "epoch": 20.18303186907838,
      "eval_accuracy": 0.39521814767541363,
      "eval_loss": 3.2936341762542725,
      "eval_runtime": 84.2068,
      "eval_samples_per_second": 213.902,
      "eval_steps_per_second": 13.372,
      "step": 75000
    },
    {
      "epoch": 20.196490094745908,
      "grad_norm": 0.4240413308143616,
      "learning_rate": 0.0002972573044297832,
      "loss": 2.9758,
      "step": 75050
    },
    {
      "epoch": 20.209948320413435,
      "grad_norm": 0.40852588415145874,
      "learning_rate": 0.00029705533862932543,
      "loss": 2.9619,
      "step": 75100
    },
    {
      "epoch": 20.223406546080966,
      "grad_norm": 0.40699484944343567,
      "learning_rate": 0.00029685337282886763,
      "loss": 2.9874,
      "step": 75150
    },
    {
      "epoch": 20.236864771748493,
      "grad_norm": 0.4012264311313629,
      "learning_rate": 0.0002966514070284098,
      "loss": 2.9782,
      "step": 75200
    },
    {
      "epoch": 20.25032299741602,
      "grad_norm": 0.3994031846523285,
      "learning_rate": 0.000296449441227952,
      "loss": 2.9756,
      "step": 75250
    },
    {
      "epoch": 20.263781223083548,
      "grad_norm": 0.42914170026779175,
      "learning_rate": 0.0002962474754274942,
      "loss": 2.9801,
      "step": 75300
    },
    {
      "epoch": 20.277239448751075,
      "grad_norm": 0.4271126389503479,
      "learning_rate": 0.00029604550962703647,
      "loss": 2.9807,
      "step": 75350
    },
    {
      "epoch": 20.290697674418606,
      "grad_norm": 0.45824435353279114,
      "learning_rate": 0.00029584354382657867,
      "loss": 2.9919,
      "step": 75400
    },
    {
      "epoch": 20.304155900086133,
      "grad_norm": 0.40884071588516235,
      "learning_rate": 0.0002956415780261209,
      "loss": 2.9802,
      "step": 75450
    },
    {
      "epoch": 20.31761412575366,
      "grad_norm": 0.43035104870796204,
      "learning_rate": 0.0002954396122256631,
      "loss": 2.9967,
      "step": 75500
    },
    {
      "epoch": 20.331072351421188,
      "grad_norm": 0.4559783637523651,
      "learning_rate": 0.0002952376464252053,
      "loss": 2.9901,
      "step": 75550
    },
    {
      "epoch": 20.344530577088715,
      "grad_norm": 0.4269450902938843,
      "learning_rate": 0.0002950356806247475,
      "loss": 2.9905,
      "step": 75600
    },
    {
      "epoch": 20.357988802756246,
      "grad_norm": 0.44144493341445923,
      "learning_rate": 0.0002948337148242897,
      "loss": 2.9889,
      "step": 75650
    },
    {
      "epoch": 20.371447028423773,
      "grad_norm": 0.4244430661201477,
      "learning_rate": 0.00029463174902383195,
      "loss": 2.9969,
      "step": 75700
    },
    {
      "epoch": 20.3849052540913,
      "grad_norm": 0.45661237835884094,
      "learning_rate": 0.00029442978322337415,
      "loss": 2.9909,
      "step": 75750
    },
    {
      "epoch": 20.398363479758828,
      "grad_norm": 0.40380531549453735,
      "learning_rate": 0.00029422781742291634,
      "loss": 2.9922,
      "step": 75800
    },
    {
      "epoch": 20.411821705426355,
      "grad_norm": 0.4603881239891052,
      "learning_rate": 0.0002940258516224586,
      "loss": 3.0011,
      "step": 75850
    },
    {
      "epoch": 20.425279931093886,
      "grad_norm": 0.431744247674942,
      "learning_rate": 0.0002938238858220008,
      "loss": 3.0096,
      "step": 75900
    },
    {
      "epoch": 20.438738156761413,
      "grad_norm": 0.4146721363067627,
      "learning_rate": 0.000293621920021543,
      "loss": 2.9948,
      "step": 75950
    },
    {
      "epoch": 20.45219638242894,
      "grad_norm": 0.42351874709129333,
      "learning_rate": 0.0002934199542210852,
      "loss": 3.007,
      "step": 76000
    },
    {
      "epoch": 20.45219638242894,
      "eval_accuracy": 0.3951562189677419,
      "eval_loss": 3.2891974449157715,
      "eval_runtime": 84.1816,
      "eval_samples_per_second": 213.966,
      "eval_steps_per_second": 13.376,
      "step": 76000
    },
    {
      "epoch": 20.465654608096468,
      "grad_norm": 0.4256731867790222,
      "learning_rate": 0.00029321798842062743,
      "loss": 2.9968,
      "step": 76050
    },
    {
      "epoch": 20.479112833763995,
      "grad_norm": 0.3949142098426819,
      "learning_rate": 0.00029301602262016963,
      "loss": 3.0021,
      "step": 76100
    },
    {
      "epoch": 20.492571059431526,
      "grad_norm": 0.4100065529346466,
      "learning_rate": 0.0002928140568197118,
      "loss": 3.0108,
      "step": 76150
    },
    {
      "epoch": 20.506029285099054,
      "grad_norm": 0.4375024735927582,
      "learning_rate": 0.000292612091019254,
      "loss": 3.0207,
      "step": 76200
    },
    {
      "epoch": 20.51948751076658,
      "grad_norm": 0.41435155272483826,
      "learning_rate": 0.0002924101252187962,
      "loss": 2.9932,
      "step": 76250
    },
    {
      "epoch": 20.532945736434108,
      "grad_norm": 0.41171911358833313,
      "learning_rate": 0.00029220815941833847,
      "loss": 2.993,
      "step": 76300
    },
    {
      "epoch": 20.546403962101635,
      "grad_norm": 0.43453913927078247,
      "learning_rate": 0.0002920061936178807,
      "loss": 3.0179,
      "step": 76350
    },
    {
      "epoch": 20.559862187769163,
      "grad_norm": 0.41289013624191284,
      "learning_rate": 0.0002918042278174229,
      "loss": 3.0076,
      "step": 76400
    },
    {
      "epoch": 20.573320413436694,
      "grad_norm": 0.4423086941242218,
      "learning_rate": 0.0002916022620169651,
      "loss": 2.9962,
      "step": 76450
    },
    {
      "epoch": 20.58677863910422,
      "grad_norm": 0.402458131313324,
      "learning_rate": 0.0002914002962165073,
      "loss": 3.0043,
      "step": 76500
    },
    {
      "epoch": 20.600236864771748,
      "grad_norm": 0.43701931834220886,
      "learning_rate": 0.0002911983304160495,
      "loss": 3.0093,
      "step": 76550
    },
    {
      "epoch": 20.613695090439276,
      "grad_norm": 0.44222718477249146,
      "learning_rate": 0.00029099636461559175,
      "loss": 3.0216,
      "step": 76600
    },
    {
      "epoch": 20.627153316106803,
      "grad_norm": 0.45639166235923767,
      "learning_rate": 0.00029079439881513395,
      "loss": 3.0018,
      "step": 76650
    },
    {
      "epoch": 20.640611541774334,
      "grad_norm": 0.4106077253818512,
      "learning_rate": 0.00029059243301467614,
      "loss": 3.0116,
      "step": 76700
    },
    {
      "epoch": 20.65406976744186,
      "grad_norm": 0.41484782099723816,
      "learning_rate": 0.0002903904672142184,
      "loss": 3.0124,
      "step": 76750
    },
    {
      "epoch": 20.66752799310939,
      "grad_norm": 0.45335787534713745,
      "learning_rate": 0.0002901885014137606,
      "loss": 3.0042,
      "step": 76800
    },
    {
      "epoch": 20.680986218776916,
      "grad_norm": 0.44338420033454895,
      "learning_rate": 0.0002899865356133028,
      "loss": 3.0165,
      "step": 76850
    },
    {
      "epoch": 20.694444444444443,
      "grad_norm": 0.4381047785282135,
      "learning_rate": 0.000289784569812845,
      "loss": 3.0096,
      "step": 76900
    },
    {
      "epoch": 20.707902670111974,
      "grad_norm": 0.4340892434120178,
      "learning_rate": 0.00028958260401238723,
      "loss": 3.0114,
      "step": 76950
    },
    {
      "epoch": 20.7213608957795,
      "grad_norm": 0.4346958100795746,
      "learning_rate": 0.00028938063821192943,
      "loss": 3.0079,
      "step": 77000
    },
    {
      "epoch": 20.7213608957795,
      "eval_accuracy": 0.39584275844805356,
      "eval_loss": 3.2841553688049316,
      "eval_runtime": 84.2355,
      "eval_samples_per_second": 213.829,
      "eval_steps_per_second": 13.367,
      "step": 77000
    },
    {
      "epoch": 20.73481912144703,
      "grad_norm": 0.42295241355895996,
      "learning_rate": 0.0002891786724114716,
      "loss": 3.0145,
      "step": 77050
    },
    {
      "epoch": 20.748277347114556,
      "grad_norm": 0.4046490490436554,
      "learning_rate": 0.0002889767066110138,
      "loss": 3.0169,
      "step": 77100
    },
    {
      "epoch": 20.761735572782083,
      "grad_norm": 0.4284420907497406,
      "learning_rate": 0.000288774740810556,
      "loss": 3.005,
      "step": 77150
    },
    {
      "epoch": 20.775193798449614,
      "grad_norm": 0.4255708158016205,
      "learning_rate": 0.00028857277501009827,
      "loss": 3.0204,
      "step": 77200
    },
    {
      "epoch": 20.78865202411714,
      "grad_norm": 0.4019745886325836,
      "learning_rate": 0.00028837080920964046,
      "loss": 3.012,
      "step": 77250
    },
    {
      "epoch": 20.80211024978467,
      "grad_norm": 0.44036272168159485,
      "learning_rate": 0.0002881688434091827,
      "loss": 3.013,
      "step": 77300
    },
    {
      "epoch": 20.815568475452196,
      "grad_norm": 0.4328933656215668,
      "learning_rate": 0.0002879668776087249,
      "loss": 3.0236,
      "step": 77350
    },
    {
      "epoch": 20.829026701119723,
      "grad_norm": 0.3818512260913849,
      "learning_rate": 0.0002877649118082671,
      "loss": 3.0239,
      "step": 77400
    },
    {
      "epoch": 20.842484926787254,
      "grad_norm": 0.40535131096839905,
      "learning_rate": 0.0002875629460078093,
      "loss": 3.0219,
      "step": 77450
    },
    {
      "epoch": 20.85594315245478,
      "grad_norm": 0.47238993644714355,
      "learning_rate": 0.0002873609802073515,
      "loss": 3.0331,
      "step": 77500
    },
    {
      "epoch": 20.86940137812231,
      "grad_norm": 0.42848581075668335,
      "learning_rate": 0.00028715901440689375,
      "loss": 3.0194,
      "step": 77550
    },
    {
      "epoch": 20.882859603789836,
      "grad_norm": 0.44346335530281067,
      "learning_rate": 0.00028695704860643594,
      "loss": 3.0282,
      "step": 77600
    },
    {
      "epoch": 20.896317829457363,
      "grad_norm": 0.42101702094078064,
      "learning_rate": 0.0002867550828059782,
      "loss": 3.0228,
      "step": 77650
    },
    {
      "epoch": 20.909776055124894,
      "grad_norm": 0.4506657123565674,
      "learning_rate": 0.0002865531170055204,
      "loss": 3.0229,
      "step": 77700
    },
    {
      "epoch": 20.92323428079242,
      "grad_norm": 0.4302241802215576,
      "learning_rate": 0.0002863511512050626,
      "loss": 3.026,
      "step": 77750
    },
    {
      "epoch": 20.93669250645995,
      "grad_norm": 0.4115939140319824,
      "learning_rate": 0.0002861491854046048,
      "loss": 3.0291,
      "step": 77800
    },
    {
      "epoch": 20.950150732127476,
      "grad_norm": 0.4474116861820221,
      "learning_rate": 0.000285947219604147,
      "loss": 3.0353,
      "step": 77850
    },
    {
      "epoch": 20.963608957795003,
      "grad_norm": 0.4128603935241699,
      "learning_rate": 0.00028574525380368923,
      "loss": 3.0248,
      "step": 77900
    },
    {
      "epoch": 20.977067183462534,
      "grad_norm": 0.41937094926834106,
      "learning_rate": 0.0002855432880032314,
      "loss": 3.0263,
      "step": 77950
    },
    {
      "epoch": 20.99052540913006,
      "grad_norm": 0.43259957432746887,
      "learning_rate": 0.0002853413222027736,
      "loss": 3.0312,
      "step": 78000
    },
    {
      "epoch": 20.99052540913006,
      "eval_accuracy": 0.3964165224922893,
      "eval_loss": 3.276137113571167,
      "eval_runtime": 84.1807,
      "eval_samples_per_second": 213.968,
      "eval_steps_per_second": 13.376,
      "step": 78000
    },
    {
      "epoch": 21.00376830318691,
      "grad_norm": 0.48113709688186646,
      "learning_rate": 0.00028513935640231587,
      "loss": 2.9942,
      "step": 78050
    },
    {
      "epoch": 21.017226528854437,
      "grad_norm": 0.41671714186668396,
      "learning_rate": 0.00028493739060185807,
      "loss": 2.9406,
      "step": 78100
    },
    {
      "epoch": 21.030684754521964,
      "grad_norm": 0.4281383454799652,
      "learning_rate": 0.00028473542480140026,
      "loss": 2.9371,
      "step": 78150
    },
    {
      "epoch": 21.04414298018949,
      "grad_norm": 0.44374173879623413,
      "learning_rate": 0.0002845334590009425,
      "loss": 2.9454,
      "step": 78200
    },
    {
      "epoch": 21.05760120585702,
      "grad_norm": 0.4133867919445038,
      "learning_rate": 0.0002843314932004847,
      "loss": 2.9385,
      "step": 78250
    },
    {
      "epoch": 21.07105943152455,
      "grad_norm": 0.4318714737892151,
      "learning_rate": 0.0002841295274000269,
      "loss": 2.9437,
      "step": 78300
    },
    {
      "epoch": 21.084517657192077,
      "grad_norm": 0.42526036500930786,
      "learning_rate": 0.0002839275615995691,
      "loss": 2.9492,
      "step": 78350
    },
    {
      "epoch": 21.097975882859604,
      "grad_norm": 0.4428749084472656,
      "learning_rate": 0.0002837255957991113,
      "loss": 2.9529,
      "step": 78400
    },
    {
      "epoch": 21.11143410852713,
      "grad_norm": 0.4705010950565338,
      "learning_rate": 0.00028352362999865355,
      "loss": 2.9707,
      "step": 78450
    },
    {
      "epoch": 21.12489233419466,
      "grad_norm": 0.42135167121887207,
      "learning_rate": 0.00028332166419819574,
      "loss": 2.9605,
      "step": 78500
    },
    {
      "epoch": 21.13835055986219,
      "grad_norm": 0.46383029222488403,
      "learning_rate": 0.000283119698397738,
      "loss": 2.9611,
      "step": 78550
    },
    {
      "epoch": 21.151808785529717,
      "grad_norm": 0.43224141001701355,
      "learning_rate": 0.0002829177325972802,
      "loss": 2.9578,
      "step": 78600
    },
    {
      "epoch": 21.165267011197244,
      "grad_norm": 0.4280335009098053,
      "learning_rate": 0.0002827157667968224,
      "loss": 2.9644,
      "step": 78650
    },
    {
      "epoch": 21.17872523686477,
      "grad_norm": 0.42558160424232483,
      "learning_rate": 0.0002825138009963646,
      "loss": 2.9595,
      "step": 78700
    },
    {
      "epoch": 21.1921834625323,
      "grad_norm": 0.43694210052490234,
      "learning_rate": 0.0002823118351959068,
      "loss": 2.9512,
      "step": 78750
    },
    {
      "epoch": 21.205641688199826,
      "grad_norm": 0.4240587651729584,
      "learning_rate": 0.00028210986939544903,
      "loss": 2.9711,
      "step": 78800
    },
    {
      "epoch": 21.219099913867357,
      "grad_norm": 0.4569544196128845,
      "learning_rate": 0.0002819079035949912,
      "loss": 2.9705,
      "step": 78850
    },
    {
      "epoch": 21.232558139534884,
      "grad_norm": 0.41455137729644775,
      "learning_rate": 0.0002817059377945334,
      "loss": 2.9773,
      "step": 78900
    },
    {
      "epoch": 21.24601636520241,
      "grad_norm": 0.4157683253288269,
      "learning_rate": 0.00028150397199407567,
      "loss": 2.9584,
      "step": 78950
    },
    {
      "epoch": 21.25947459086994,
      "grad_norm": 0.44556525349617004,
      "learning_rate": 0.00028130200619361787,
      "loss": 2.9761,
      "step": 79000
    },
    {
      "epoch": 21.25947459086994,
      "eval_accuracy": 0.3955257269235165,
      "eval_loss": 3.2893779277801514,
      "eval_runtime": 84.4114,
      "eval_samples_per_second": 213.383,
      "eval_steps_per_second": 13.339,
      "step": 79000
    },
    {
      "epoch": 21.272932816537466,
      "grad_norm": 0.43476971983909607,
      "learning_rate": 0.00028110004039316006,
      "loss": 2.9738,
      "step": 79050
    },
    {
      "epoch": 21.286391042204997,
      "grad_norm": 0.4178507328033447,
      "learning_rate": 0.00028089807459270226,
      "loss": 2.9821,
      "step": 79100
    },
    {
      "epoch": 21.299849267872524,
      "grad_norm": 0.4632306694984436,
      "learning_rate": 0.0002806961087922445,
      "loss": 2.9797,
      "step": 79150
    },
    {
      "epoch": 21.31330749354005,
      "grad_norm": 0.439938485622406,
      "learning_rate": 0.0002804941429917867,
      "loss": 2.974,
      "step": 79200
    },
    {
      "epoch": 21.32676571920758,
      "grad_norm": 0.4647805988788605,
      "learning_rate": 0.0002802921771913289,
      "loss": 2.9843,
      "step": 79250
    },
    {
      "epoch": 21.340223944875106,
      "grad_norm": 0.4232933819293976,
      "learning_rate": 0.0002800902113908711,
      "loss": 2.9878,
      "step": 79300
    },
    {
      "epoch": 21.353682170542637,
      "grad_norm": 0.42116060853004456,
      "learning_rate": 0.0002798882455904133,
      "loss": 2.9746,
      "step": 79350
    },
    {
      "epoch": 21.367140396210164,
      "grad_norm": 0.41664060950279236,
      "learning_rate": 0.00027968627978995555,
      "loss": 2.9838,
      "step": 79400
    },
    {
      "epoch": 21.38059862187769,
      "grad_norm": 0.4296671748161316,
      "learning_rate": 0.00027948431398949774,
      "loss": 2.9871,
      "step": 79450
    },
    {
      "epoch": 21.39405684754522,
      "grad_norm": 0.44407474994659424,
      "learning_rate": 0.00027928234818904,
      "loss": 2.9862,
      "step": 79500
    },
    {
      "epoch": 21.407515073212746,
      "grad_norm": 0.45838916301727295,
      "learning_rate": 0.0002790803823885822,
      "loss": 2.9866,
      "step": 79550
    },
    {
      "epoch": 21.420973298880277,
      "grad_norm": 0.4345085322856903,
      "learning_rate": 0.0002788784165881244,
      "loss": 2.9845,
      "step": 79600
    },
    {
      "epoch": 21.434431524547804,
      "grad_norm": 0.46148353815078735,
      "learning_rate": 0.0002786764507876666,
      "loss": 2.9853,
      "step": 79650
    },
    {
      "epoch": 21.44788975021533,
      "grad_norm": 0.4526107907295227,
      "learning_rate": 0.0002784744849872088,
      "loss": 2.9898,
      "step": 79700
    },
    {
      "epoch": 21.46134797588286,
      "grad_norm": 0.4054795205593109,
      "learning_rate": 0.000278272519186751,
      "loss": 2.9847,
      "step": 79750
    },
    {
      "epoch": 21.474806201550386,
      "grad_norm": 0.41127878427505493,
      "learning_rate": 0.0002780705533862932,
      "loss": 2.9959,
      "step": 79800
    },
    {
      "epoch": 21.488264427217917,
      "grad_norm": 0.4218996465206146,
      "learning_rate": 0.0002778685875858355,
      "loss": 2.99,
      "step": 79850
    },
    {
      "epoch": 21.501722652885444,
      "grad_norm": 0.4325758218765259,
      "learning_rate": 0.00027766662178537767,
      "loss": 2.9942,
      "step": 79900
    },
    {
      "epoch": 21.51518087855297,
      "grad_norm": 0.4231953024864197,
      "learning_rate": 0.00027746465598491987,
      "loss": 2.9873,
      "step": 79950
    },
    {
      "epoch": 21.5286391042205,
      "grad_norm": 0.4397009015083313,
      "learning_rate": 0.00027726269018446206,
      "loss": 2.9983,
      "step": 80000
    },
    {
      "epoch": 21.5286391042205,
      "eval_accuracy": 0.3956675110700281,
      "eval_loss": 3.2855608463287354,
      "eval_runtime": 84.4641,
      "eval_samples_per_second": 213.25,
      "eval_steps_per_second": 13.331,
      "step": 80000
    },
    {
      "epoch": 21.542097329888026,
      "grad_norm": 0.4709068238735199,
      "learning_rate": 0.00027706072438400426,
      "loss": 2.9979,
      "step": 80050
    },
    {
      "epoch": 21.555555555555557,
      "grad_norm": 0.4287954866886139,
      "learning_rate": 0.0002768587585835465,
      "loss": 2.9954,
      "step": 80100
    },
    {
      "epoch": 21.569013781223084,
      "grad_norm": 0.41935524344444275,
      "learning_rate": 0.0002766567927830887,
      "loss": 3.0003,
      "step": 80150
    },
    {
      "epoch": 21.58247200689061,
      "grad_norm": 0.4271881580352783,
      "learning_rate": 0.0002764548269826309,
      "loss": 3.001,
      "step": 80200
    },
    {
      "epoch": 21.59593023255814,
      "grad_norm": 0.43717139959335327,
      "learning_rate": 0.0002762528611821731,
      "loss": 2.986,
      "step": 80250
    },
    {
      "epoch": 21.609388458225666,
      "grad_norm": 0.41988781094551086,
      "learning_rate": 0.00027605089538171535,
      "loss": 2.9871,
      "step": 80300
    },
    {
      "epoch": 21.622846683893197,
      "grad_norm": 0.4415890574455261,
      "learning_rate": 0.00027584892958125754,
      "loss": 3.0056,
      "step": 80350
    },
    {
      "epoch": 21.636304909560724,
      "grad_norm": 0.42887449264526367,
      "learning_rate": 0.0002756469637807998,
      "loss": 2.9965,
      "step": 80400
    },
    {
      "epoch": 21.649763135228252,
      "grad_norm": 0.41498321294784546,
      "learning_rate": 0.000275444997980342,
      "loss": 2.9968,
      "step": 80450
    },
    {
      "epoch": 21.66322136089578,
      "grad_norm": 0.41250497102737427,
      "learning_rate": 0.0002752430321798842,
      "loss": 3.0041,
      "step": 80500
    },
    {
      "epoch": 21.676679586563306,
      "grad_norm": 0.4571186304092407,
      "learning_rate": 0.0002750410663794264,
      "loss": 2.9934,
      "step": 80550
    },
    {
      "epoch": 21.690137812230837,
      "grad_norm": 0.4223412275314331,
      "learning_rate": 0.0002748391005789686,
      "loss": 3.0143,
      "step": 80600
    },
    {
      "epoch": 21.703596037898365,
      "grad_norm": 0.4207080602645874,
      "learning_rate": 0.00027463713477851083,
      "loss": 3.0069,
      "step": 80650
    },
    {
      "epoch": 21.717054263565892,
      "grad_norm": 0.44499847292900085,
      "learning_rate": 0.000274435168978053,
      "loss": 3.0036,
      "step": 80700
    },
    {
      "epoch": 21.73051248923342,
      "grad_norm": 0.3987369239330292,
      "learning_rate": 0.0002742332031775953,
      "loss": 2.9944,
      "step": 80750
    },
    {
      "epoch": 21.743970714900946,
      "grad_norm": 0.3991621732711792,
      "learning_rate": 0.00027403123737713747,
      "loss": 2.999,
      "step": 80800
    },
    {
      "epoch": 21.757428940568474,
      "grad_norm": 0.4163661599159241,
      "learning_rate": 0.00027382927157667967,
      "loss": 3.0073,
      "step": 80850
    },
    {
      "epoch": 21.770887166236005,
      "grad_norm": 0.42498746514320374,
      "learning_rate": 0.00027362730577622186,
      "loss": 3.0072,
      "step": 80900
    },
    {
      "epoch": 21.784345391903532,
      "grad_norm": 0.4446844756603241,
      "learning_rate": 0.00027342533997576406,
      "loss": 3.0069,
      "step": 80950
    },
    {
      "epoch": 21.79780361757106,
      "grad_norm": 0.4386047124862671,
      "learning_rate": 0.0002732233741753063,
      "loss": 3.0083,
      "step": 81000
    },
    {
      "epoch": 21.79780361757106,
      "eval_accuracy": 0.39614794746533405,
      "eval_loss": 3.2783379554748535,
      "eval_runtime": 84.3997,
      "eval_samples_per_second": 213.413,
      "eval_steps_per_second": 13.341,
      "step": 81000
    },
    {
      "epoch": 21.811261843238587,
      "grad_norm": 0.43316155672073364,
      "learning_rate": 0.0002730214083748485,
      "loss": 3.0009,
      "step": 81050
    },
    {
      "epoch": 21.824720068906114,
      "grad_norm": 0.41254112124443054,
      "learning_rate": 0.0002728194425743907,
      "loss": 3.0195,
      "step": 81100
    },
    {
      "epoch": 21.838178294573645,
      "grad_norm": 0.4280899465084076,
      "learning_rate": 0.0002726174767739329,
      "loss": 3.0058,
      "step": 81150
    },
    {
      "epoch": 21.851636520241172,
      "grad_norm": 0.425401508808136,
      "learning_rate": 0.00027241551097347515,
      "loss": 3.0115,
      "step": 81200
    },
    {
      "epoch": 21.8650947459087,
      "grad_norm": 0.40731748938560486,
      "learning_rate": 0.00027221354517301734,
      "loss": 3.0066,
      "step": 81250
    },
    {
      "epoch": 21.878552971576227,
      "grad_norm": 0.42605629563331604,
      "learning_rate": 0.00027201157937255954,
      "loss": 3.0125,
      "step": 81300
    },
    {
      "epoch": 21.892011197243754,
      "grad_norm": 0.4229583740234375,
      "learning_rate": 0.0002718096135721018,
      "loss": 3.0118,
      "step": 81350
    },
    {
      "epoch": 21.905469422911285,
      "grad_norm": 0.4372251629829407,
      "learning_rate": 0.000271607647771644,
      "loss": 3.0138,
      "step": 81400
    },
    {
      "epoch": 21.918927648578812,
      "grad_norm": 0.4544908106327057,
      "learning_rate": 0.0002714056819711862,
      "loss": 3.022,
      "step": 81450
    },
    {
      "epoch": 21.93238587424634,
      "grad_norm": 0.4251236319541931,
      "learning_rate": 0.0002712037161707284,
      "loss": 3.0073,
      "step": 81500
    },
    {
      "epoch": 21.945844099913867,
      "grad_norm": 0.41079625487327576,
      "learning_rate": 0.0002710017503702706,
      "loss": 3.0237,
      "step": 81550
    },
    {
      "epoch": 21.959302325581394,
      "grad_norm": 0.4202929139137268,
      "learning_rate": 0.0002707997845698128,
      "loss": 3.0082,
      "step": 81600
    },
    {
      "epoch": 21.972760551248925,
      "grad_norm": 0.4280118942260742,
      "learning_rate": 0.000270597818769355,
      "loss": 3.0193,
      "step": 81650
    },
    {
      "epoch": 21.986218776916452,
      "grad_norm": 0.4100026786327362,
      "learning_rate": 0.00027039585296889727,
      "loss": 3.0194,
      "step": 81700
    },
    {
      "epoch": 21.99967700258398,
      "grad_norm": 0.4353160560131073,
      "learning_rate": 0.00027019388716843947,
      "loss": 3.0147,
      "step": 81750
    },
    {
      "epoch": 22.012919896640827,
      "grad_norm": 0.4785066545009613,
      "learning_rate": 0.00026999192136798166,
      "loss": 2.929,
      "step": 81800
    },
    {
      "epoch": 22.026378122308355,
      "grad_norm": 0.43430209159851074,
      "learning_rate": 0.00026978995556752386,
      "loss": 2.9383,
      "step": 81850
    },
    {
      "epoch": 22.039836347975882,
      "grad_norm": 0.45329535007476807,
      "learning_rate": 0.00026958798976706606,
      "loss": 2.9409,
      "step": 81900
    },
    {
      "epoch": 22.05329457364341,
      "grad_norm": 0.4317338168621063,
      "learning_rate": 0.0002693860239666083,
      "loss": 2.9353,
      "step": 81950
    },
    {
      "epoch": 22.06675279931094,
      "grad_norm": 0.44393497705459595,
      "learning_rate": 0.0002691840581661505,
      "loss": 2.9479,
      "step": 82000
    },
    {
      "epoch": 22.06675279931094,
      "eval_accuracy": 0.3956484978703043,
      "eval_loss": 3.291426420211792,
      "eval_runtime": 84.2418,
      "eval_samples_per_second": 213.813,
      "eval_steps_per_second": 13.366,
      "step": 82000
    },
    {
      "epoch": 22.080211024978468,
      "grad_norm": 0.4554722309112549,
      "learning_rate": 0.0002689820923656927,
      "loss": 2.9453,
      "step": 82050
    },
    {
      "epoch": 22.093669250645995,
      "grad_norm": 0.4372183382511139,
      "learning_rate": 0.00026878012656523495,
      "loss": 2.9461,
      "step": 82100
    },
    {
      "epoch": 22.107127476313522,
      "grad_norm": 0.41931185126304626,
      "learning_rate": 0.00026857816076477714,
      "loss": 2.9408,
      "step": 82150
    },
    {
      "epoch": 22.12058570198105,
      "grad_norm": 0.41515403985977173,
      "learning_rate": 0.00026837619496431934,
      "loss": 2.9544,
      "step": 82200
    },
    {
      "epoch": 22.13404392764858,
      "grad_norm": 0.4686168432235718,
      "learning_rate": 0.0002681742291638616,
      "loss": 2.9531,
      "step": 82250
    },
    {
      "epoch": 22.147502153316108,
      "grad_norm": 0.43255189061164856,
      "learning_rate": 0.0002679722633634038,
      "loss": 2.958,
      "step": 82300
    },
    {
      "epoch": 22.160960378983635,
      "grad_norm": 0.44745513796806335,
      "learning_rate": 0.000267770297562946,
      "loss": 2.9437,
      "step": 82350
    },
    {
      "epoch": 22.174418604651162,
      "grad_norm": 0.4283008277416229,
      "learning_rate": 0.0002675683317624882,
      "loss": 2.9502,
      "step": 82400
    },
    {
      "epoch": 22.18787683031869,
      "grad_norm": 0.42748215794563293,
      "learning_rate": 0.0002673663659620304,
      "loss": 2.9589,
      "step": 82450
    },
    {
      "epoch": 22.20133505598622,
      "grad_norm": 0.4140898585319519,
      "learning_rate": 0.0002671644001615726,
      "loss": 2.9573,
      "step": 82500
    },
    {
      "epoch": 22.214793281653748,
      "grad_norm": 0.4312494099140167,
      "learning_rate": 0.0002669624343611148,
      "loss": 2.954,
      "step": 82550
    },
    {
      "epoch": 22.228251507321275,
      "grad_norm": 0.484730064868927,
      "learning_rate": 0.00026676046856065707,
      "loss": 2.9519,
      "step": 82600
    },
    {
      "epoch": 22.241709732988802,
      "grad_norm": 0.4209139943122864,
      "learning_rate": 0.00026655850276019927,
      "loss": 2.9513,
      "step": 82650
    },
    {
      "epoch": 22.25516795865633,
      "grad_norm": 0.41069546341896057,
      "learning_rate": 0.00026635653695974146,
      "loss": 2.9638,
      "step": 82700
    },
    {
      "epoch": 22.26862618432386,
      "grad_norm": 0.4399559199810028,
      "learning_rate": 0.00026615457115928366,
      "loss": 2.9603,
      "step": 82750
    },
    {
      "epoch": 22.282084409991388,
      "grad_norm": 0.42374998331069946,
      "learning_rate": 0.00026595260535882586,
      "loss": 2.9526,
      "step": 82800
    },
    {
      "epoch": 22.295542635658915,
      "grad_norm": 0.43667072057724,
      "learning_rate": 0.0002657506395583681,
      "loss": 2.9648,
      "step": 82850
    },
    {
      "epoch": 22.309000861326442,
      "grad_norm": 0.42516595125198364,
      "learning_rate": 0.0002655486737579103,
      "loss": 2.9667,
      "step": 82900
    },
    {
      "epoch": 22.32245908699397,
      "grad_norm": 0.4428614377975464,
      "learning_rate": 0.00026534670795745255,
      "loss": 2.9751,
      "step": 82950
    },
    {
      "epoch": 22.3359173126615,
      "grad_norm": 0.43633002042770386,
      "learning_rate": 0.00026514474215699475,
      "loss": 2.9673,
      "step": 83000
    },
    {
      "epoch": 22.3359173126615,
      "eval_accuracy": 0.39550421484611475,
      "eval_loss": 3.2900450229644775,
      "eval_runtime": 84.3594,
      "eval_samples_per_second": 213.515,
      "eval_steps_per_second": 13.348,
      "step": 83000
    },
    {
      "epoch": 22.349375538329028,
      "grad_norm": 0.4040275514125824,
      "learning_rate": 0.00026494277635653694,
      "loss": 2.9692,
      "step": 83050
    },
    {
      "epoch": 22.362833763996555,
      "grad_norm": 0.44113674759864807,
      "learning_rate": 0.00026474081055607914,
      "loss": 2.979,
      "step": 83100
    },
    {
      "epoch": 22.376291989664082,
      "grad_norm": 0.43358901143074036,
      "learning_rate": 0.00026453884475562134,
      "loss": 2.9684,
      "step": 83150
    },
    {
      "epoch": 22.38975021533161,
      "grad_norm": 0.49278175830841064,
      "learning_rate": 0.0002643368789551636,
      "loss": 2.9718,
      "step": 83200
    },
    {
      "epoch": 22.403208440999137,
      "grad_norm": 0.42028820514678955,
      "learning_rate": 0.0002641349131547058,
      "loss": 2.9787,
      "step": 83250
    },
    {
      "epoch": 22.416666666666668,
      "grad_norm": 0.43706896901130676,
      "learning_rate": 0.000263932947354248,
      "loss": 2.9718,
      "step": 83300
    },
    {
      "epoch": 22.430124892334195,
      "grad_norm": 0.4647141098976135,
      "learning_rate": 0.0002637309815537902,
      "loss": 2.9748,
      "step": 83350
    },
    {
      "epoch": 22.443583118001722,
      "grad_norm": 0.42630627751350403,
      "learning_rate": 0.0002635290157533324,
      "loss": 2.9797,
      "step": 83400
    },
    {
      "epoch": 22.45704134366925,
      "grad_norm": 0.44382381439208984,
      "learning_rate": 0.0002633270499528746,
      "loss": 2.9914,
      "step": 83450
    },
    {
      "epoch": 22.470499569336777,
      "grad_norm": 0.4510803818702698,
      "learning_rate": 0.0002631250841524168,
      "loss": 2.9876,
      "step": 83500
    },
    {
      "epoch": 22.483957795004308,
      "grad_norm": 0.42208534479141235,
      "learning_rate": 0.00026292311835195907,
      "loss": 2.9881,
      "step": 83550
    },
    {
      "epoch": 22.497416020671835,
      "grad_norm": 0.4549991190433502,
      "learning_rate": 0.00026272115255150126,
      "loss": 2.9813,
      "step": 83600
    },
    {
      "epoch": 22.510874246339363,
      "grad_norm": 0.44071683287620544,
      "learning_rate": 0.00026251918675104346,
      "loss": 2.9696,
      "step": 83650
    },
    {
      "epoch": 22.52433247200689,
      "grad_norm": 0.44255688786506653,
      "learning_rate": 0.00026231722095058566,
      "loss": 2.9737,
      "step": 83700
    },
    {
      "epoch": 22.537790697674417,
      "grad_norm": 0.4315544664859772,
      "learning_rate": 0.00026211525515012785,
      "loss": 2.9929,
      "step": 83750
    },
    {
      "epoch": 22.551248923341948,
      "grad_norm": 0.42129266262054443,
      "learning_rate": 0.0002619132893496701,
      "loss": 2.9962,
      "step": 83800
    },
    {
      "epoch": 22.564707149009475,
      "grad_norm": 0.4324382543563843,
      "learning_rate": 0.0002617113235492123,
      "loss": 2.9864,
      "step": 83850
    },
    {
      "epoch": 22.578165374677003,
      "grad_norm": 0.43194422125816345,
      "learning_rate": 0.00026150935774875455,
      "loss": 2.9977,
      "step": 83900
    },
    {
      "epoch": 22.59162360034453,
      "grad_norm": 0.4550255239009857,
      "learning_rate": 0.00026130739194829675,
      "loss": 2.9809,
      "step": 83950
    },
    {
      "epoch": 22.605081826012057,
      "grad_norm": 0.4199272692203522,
      "learning_rate": 0.00026110542614783894,
      "loss": 2.9974,
      "step": 84000
    },
    {
      "epoch": 22.605081826012057,
      "eval_accuracy": 0.3962559424397651,
      "eval_loss": 3.2822072505950928,
      "eval_runtime": 84.2178,
      "eval_samples_per_second": 213.874,
      "eval_steps_per_second": 13.37,
      "step": 84000
    },
    {
      "epoch": 22.618540051679588,
      "grad_norm": 0.43133702874183655,
      "learning_rate": 0.00026090346034738114,
      "loss": 2.9937,
      "step": 84050
    },
    {
      "epoch": 22.631998277347115,
      "grad_norm": 0.43230393528938293,
      "learning_rate": 0.0002607014945469234,
      "loss": 2.9925,
      "step": 84100
    },
    {
      "epoch": 22.645456503014643,
      "grad_norm": 0.4332883954048157,
      "learning_rate": 0.0002604995287464656,
      "loss": 2.9918,
      "step": 84150
    },
    {
      "epoch": 22.65891472868217,
      "grad_norm": 0.4412862956523895,
      "learning_rate": 0.0002602975629460078,
      "loss": 2.9791,
      "step": 84200
    },
    {
      "epoch": 22.672372954349697,
      "grad_norm": 0.44808104634284973,
      "learning_rate": 0.00026009559714555,
      "loss": 2.9951,
      "step": 84250
    },
    {
      "epoch": 22.685831180017228,
      "grad_norm": 0.45203959941864014,
      "learning_rate": 0.0002598936313450922,
      "loss": 2.9839,
      "step": 84300
    },
    {
      "epoch": 22.699289405684755,
      "grad_norm": 0.45069366693496704,
      "learning_rate": 0.0002596916655446344,
      "loss": 2.9882,
      "step": 84350
    },
    {
      "epoch": 22.712747631352283,
      "grad_norm": 0.4243113696575165,
      "learning_rate": 0.0002594896997441766,
      "loss": 2.9876,
      "step": 84400
    },
    {
      "epoch": 22.72620585701981,
      "grad_norm": 0.46229085326194763,
      "learning_rate": 0.00025928773394371887,
      "loss": 2.9938,
      "step": 84450
    },
    {
      "epoch": 22.739664082687337,
      "grad_norm": 0.45884108543395996,
      "learning_rate": 0.00025908576814326107,
      "loss": 2.9913,
      "step": 84500
    },
    {
      "epoch": 22.753122308354868,
      "grad_norm": 0.4346989691257477,
      "learning_rate": 0.00025888380234280326,
      "loss": 2.9906,
      "step": 84550
    },
    {
      "epoch": 22.766580534022395,
      "grad_norm": 0.4462386965751648,
      "learning_rate": 0.00025868183654234546,
      "loss": 3.0047,
      "step": 84600
    },
    {
      "epoch": 22.780038759689923,
      "grad_norm": 0.45179876685142517,
      "learning_rate": 0.00025847987074188765,
      "loss": 3.012,
      "step": 84650
    },
    {
      "epoch": 22.79349698535745,
      "grad_norm": 0.41962164640426636,
      "learning_rate": 0.0002582779049414299,
      "loss": 2.996,
      "step": 84700
    },
    {
      "epoch": 22.806955211024977,
      "grad_norm": 0.44738271832466125,
      "learning_rate": 0.0002580759391409721,
      "loss": 2.9975,
      "step": 84750
    },
    {
      "epoch": 22.820413436692505,
      "grad_norm": 0.4586033523082733,
      "learning_rate": 0.00025787397334051435,
      "loss": 3.0053,
      "step": 84800
    },
    {
      "epoch": 22.833871662360036,
      "grad_norm": 0.44295305013656616,
      "learning_rate": 0.00025767200754005655,
      "loss": 2.9959,
      "step": 84850
    },
    {
      "epoch": 22.847329888027563,
      "grad_norm": 0.4162992238998413,
      "learning_rate": 0.00025747004173959874,
      "loss": 2.9962,
      "step": 84900
    },
    {
      "epoch": 22.86078811369509,
      "grad_norm": 0.4316057860851288,
      "learning_rate": 0.00025726807593914094,
      "loss": 2.9985,
      "step": 84950
    },
    {
      "epoch": 22.874246339362617,
      "grad_norm": 0.4657049775123596,
      "learning_rate": 0.00025706611013868313,
      "loss": 2.9881,
      "step": 85000
    },
    {
      "epoch": 22.874246339362617,
      "eval_accuracy": 0.39644976843009205,
      "eval_loss": 3.277489423751831,
      "eval_runtime": 84.2054,
      "eval_samples_per_second": 213.905,
      "eval_steps_per_second": 13.372,
      "step": 85000
    },
    {
      "epoch": 22.887704565030145,
      "grad_norm": 0.4318774938583374,
      "learning_rate": 0.0002568641443382254,
      "loss": 2.981,
      "step": 85050
    },
    {
      "epoch": 22.901162790697676,
      "grad_norm": 0.42645877599716187,
      "learning_rate": 0.0002566621785377676,
      "loss": 3.0126,
      "step": 85100
    },
    {
      "epoch": 22.914621016365203,
      "grad_norm": 0.4816548526287079,
      "learning_rate": 0.0002564602127373098,
      "loss": 3.0005,
      "step": 85150
    },
    {
      "epoch": 22.92807924203273,
      "grad_norm": 0.4206794500350952,
      "learning_rate": 0.00025625824693685203,
      "loss": 2.9963,
      "step": 85200
    },
    {
      "epoch": 22.941537467700257,
      "grad_norm": 0.4292474091053009,
      "learning_rate": 0.0002560562811363942,
      "loss": 2.9969,
      "step": 85250
    },
    {
      "epoch": 22.954995693367785,
      "grad_norm": 0.441545307636261,
      "learning_rate": 0.0002558543153359364,
      "loss": 3.0034,
      "step": 85300
    },
    {
      "epoch": 22.968453919035316,
      "grad_norm": 0.4273262023925781,
      "learning_rate": 0.0002556523495354786,
      "loss": 2.9899,
      "step": 85350
    },
    {
      "epoch": 22.981912144702843,
      "grad_norm": 0.4588988721370697,
      "learning_rate": 0.00025545038373502087,
      "loss": 3.0084,
      "step": 85400
    },
    {
      "epoch": 22.99537037037037,
      "grad_norm": 0.4454367458820343,
      "learning_rate": 0.00025524841793456306,
      "loss": 3.002,
      "step": 85450
    },
    {
      "epoch": 23.00861326442722,
      "grad_norm": 0.4609130918979645,
      "learning_rate": 0.00025504645213410526,
      "loss": 2.9471,
      "step": 85500
    },
    {
      "epoch": 23.022071490094746,
      "grad_norm": 0.4232344925403595,
      "learning_rate": 0.00025484448633364745,
      "loss": 2.9226,
      "step": 85550
    },
    {
      "epoch": 23.035529715762273,
      "grad_norm": 0.4284253716468811,
      "learning_rate": 0.00025464252053318965,
      "loss": 2.9235,
      "step": 85600
    },
    {
      "epoch": 23.0489879414298,
      "grad_norm": 0.47828778624534607,
      "learning_rate": 0.0002544405547327319,
      "loss": 2.938,
      "step": 85650
    },
    {
      "epoch": 23.06244616709733,
      "grad_norm": 0.45083558559417725,
      "learning_rate": 0.0002542385889322741,
      "loss": 2.9261,
      "step": 85700
    },
    {
      "epoch": 23.07590439276486,
      "grad_norm": 0.4257071316242218,
      "learning_rate": 0.00025403662313181635,
      "loss": 2.9283,
      "step": 85750
    },
    {
      "epoch": 23.089362618432386,
      "grad_norm": 0.46276742219924927,
      "learning_rate": 0.00025383465733135854,
      "loss": 2.9286,
      "step": 85800
    },
    {
      "epoch": 23.102820844099913,
      "grad_norm": 0.4386633634567261,
      "learning_rate": 0.00025363269153090074,
      "loss": 2.9142,
      "step": 85850
    },
    {
      "epoch": 23.11627906976744,
      "grad_norm": 0.45438438653945923,
      "learning_rate": 0.00025343072573044294,
      "loss": 2.939,
      "step": 85900
    },
    {
      "epoch": 23.12973729543497,
      "grad_norm": 0.48279038071632385,
      "learning_rate": 0.00025322875992998513,
      "loss": 2.9429,
      "step": 85950
    },
    {
      "epoch": 23.1431955211025,
      "grad_norm": 0.4280843436717987,
      "learning_rate": 0.0002530267941295274,
      "loss": 2.9401,
      "step": 86000
    },
    {
      "epoch": 23.1431955211025,
      "eval_accuracy": 0.3959046871557253,
      "eval_loss": 3.289358615875244,
      "eval_runtime": 84.7092,
      "eval_samples_per_second": 212.633,
      "eval_steps_per_second": 13.293,
      "step": 86000
    },
    {
      "epoch": 23.156653746770026,
      "grad_norm": 0.44154366850852966,
      "learning_rate": 0.0002528248283290696,
      "loss": 2.947,
      "step": 86050
    },
    {
      "epoch": 23.170111972437553,
      "grad_norm": 0.4362318813800812,
      "learning_rate": 0.00025262286252861183,
      "loss": 2.9437,
      "step": 86100
    },
    {
      "epoch": 23.18357019810508,
      "grad_norm": 0.4498545825481415,
      "learning_rate": 0.000252420896728154,
      "loss": 2.9482,
      "step": 86150
    },
    {
      "epoch": 23.19702842377261,
      "grad_norm": 0.43635016679763794,
      "learning_rate": 0.0002522189309276962,
      "loss": 2.9472,
      "step": 86200
    },
    {
      "epoch": 23.21048664944014,
      "grad_norm": 0.4467051923274994,
      "learning_rate": 0.0002520169651272384,
      "loss": 2.9403,
      "step": 86250
    },
    {
      "epoch": 23.223944875107666,
      "grad_norm": 0.4375402629375458,
      "learning_rate": 0.00025181499932678067,
      "loss": 2.9361,
      "step": 86300
    },
    {
      "epoch": 23.237403100775193,
      "grad_norm": 0.438663512468338,
      "learning_rate": 0.00025161303352632286,
      "loss": 2.951,
      "step": 86350
    },
    {
      "epoch": 23.25086132644272,
      "grad_norm": 0.46189165115356445,
      "learning_rate": 0.00025141106772586506,
      "loss": 2.9466,
      "step": 86400
    },
    {
      "epoch": 23.26431955211025,
      "grad_norm": 0.48259103298187256,
      "learning_rate": 0.00025120910192540726,
      "loss": 2.9528,
      "step": 86450
    },
    {
      "epoch": 23.27777777777778,
      "grad_norm": 0.44444501399993896,
      "learning_rate": 0.00025100713612494945,
      "loss": 2.951,
      "step": 86500
    },
    {
      "epoch": 23.291236003445306,
      "grad_norm": 0.45306235551834106,
      "learning_rate": 0.0002508051703244917,
      "loss": 2.9499,
      "step": 86550
    },
    {
      "epoch": 23.304694229112833,
      "grad_norm": 0.43860191106796265,
      "learning_rate": 0.0002506032045240339,
      "loss": 2.9596,
      "step": 86600
    },
    {
      "epoch": 23.31815245478036,
      "grad_norm": 0.42625096440315247,
      "learning_rate": 0.00025040123872357615,
      "loss": 2.9567,
      "step": 86650
    },
    {
      "epoch": 23.33161068044789,
      "grad_norm": 0.4213877320289612,
      "learning_rate": 0.00025019927292311834,
      "loss": 2.9555,
      "step": 86700
    },
    {
      "epoch": 23.34506890611542,
      "grad_norm": 0.47949302196502686,
      "learning_rate": 0.00024999730712266054,
      "loss": 2.9491,
      "step": 86750
    },
    {
      "epoch": 23.358527131782946,
      "grad_norm": 0.4475306570529938,
      "learning_rate": 0.00024979534132220274,
      "loss": 2.9665,
      "step": 86800
    },
    {
      "epoch": 23.371985357450473,
      "grad_norm": 0.48562973737716675,
      "learning_rate": 0.00024959337552174493,
      "loss": 2.9681,
      "step": 86850
    },
    {
      "epoch": 23.385443583118,
      "grad_norm": 0.4420519769191742,
      "learning_rate": 0.0002493914097212872,
      "loss": 2.9748,
      "step": 86900
    },
    {
      "epoch": 23.39890180878553,
      "grad_norm": 0.4443926513195038,
      "learning_rate": 0.0002491894439208294,
      "loss": 2.9666,
      "step": 86950
    },
    {
      "epoch": 23.41236003445306,
      "grad_norm": 0.45205479860305786,
      "learning_rate": 0.00024898747812037163,
      "loss": 2.9705,
      "step": 87000
    },
    {
      "epoch": 23.41236003445306,
      "eval_accuracy": 0.39622302244252905,
      "eval_loss": 3.2870099544525146,
      "eval_runtime": 84.4021,
      "eval_samples_per_second": 213.407,
      "eval_steps_per_second": 13.341,
      "step": 87000
    },
    {
      "epoch": 23.425818260120586,
      "grad_norm": 0.438326358795166,
      "learning_rate": 0.0002487855123199138,
      "loss": 2.9766,
      "step": 87050
    },
    {
      "epoch": 23.439276485788113,
      "grad_norm": 0.45632097125053406,
      "learning_rate": 0.000248583546519456,
      "loss": 2.9607,
      "step": 87100
    },
    {
      "epoch": 23.45273471145564,
      "grad_norm": 0.45612555742263794,
      "learning_rate": 0.0002483815807189982,
      "loss": 2.9737,
      "step": 87150
    },
    {
      "epoch": 23.466192937123168,
      "grad_norm": 0.4510277807712555,
      "learning_rate": 0.0002481796149185404,
      "loss": 2.9566,
      "step": 87200
    },
    {
      "epoch": 23.4796511627907,
      "grad_norm": 0.4482479393482208,
      "learning_rate": 0.00024797764911808266,
      "loss": 2.9681,
      "step": 87250
    },
    {
      "epoch": 23.493109388458226,
      "grad_norm": 0.4364250600337982,
      "learning_rate": 0.00024777568331762486,
      "loss": 2.9681,
      "step": 87300
    },
    {
      "epoch": 23.506567614125753,
      "grad_norm": 0.4434226453304291,
      "learning_rate": 0.00024757371751716706,
      "loss": 2.9757,
      "step": 87350
    },
    {
      "epoch": 23.52002583979328,
      "grad_norm": 0.4489511251449585,
      "learning_rate": 0.0002473717517167093,
      "loss": 2.9684,
      "step": 87400
    },
    {
      "epoch": 23.533484065460808,
      "grad_norm": 0.4310791492462158,
      "learning_rate": 0.0002471697859162515,
      "loss": 2.9736,
      "step": 87450
    },
    {
      "epoch": 23.54694229112834,
      "grad_norm": 0.43977832794189453,
      "learning_rate": 0.0002469678201157937,
      "loss": 2.9804,
      "step": 87500
    },
    {
      "epoch": 23.560400516795866,
      "grad_norm": 0.4800080955028534,
      "learning_rate": 0.0002467658543153359,
      "loss": 2.9741,
      "step": 87550
    },
    {
      "epoch": 23.573858742463393,
      "grad_norm": 0.43800172209739685,
      "learning_rate": 0.00024656388851487815,
      "loss": 2.9741,
      "step": 87600
    },
    {
      "epoch": 23.58731696813092,
      "grad_norm": 0.44847339391708374,
      "learning_rate": 0.00024636192271442034,
      "loss": 2.9877,
      "step": 87650
    },
    {
      "epoch": 23.600775193798448,
      "grad_norm": 0.4332613945007324,
      "learning_rate": 0.00024615995691396254,
      "loss": 2.9724,
      "step": 87700
    },
    {
      "epoch": 23.61423341946598,
      "grad_norm": 0.45360738039016724,
      "learning_rate": 0.00024595799111350473,
      "loss": 2.9813,
      "step": 87750
    },
    {
      "epoch": 23.627691645133506,
      "grad_norm": 0.46419477462768555,
      "learning_rate": 0.00024575602531304693,
      "loss": 2.9826,
      "step": 87800
    },
    {
      "epoch": 23.641149870801033,
      "grad_norm": 0.43647441267967224,
      "learning_rate": 0.0002455540595125892,
      "loss": 2.9771,
      "step": 87850
    },
    {
      "epoch": 23.65460809646856,
      "grad_norm": 0.47725701332092285,
      "learning_rate": 0.00024535209371213143,
      "loss": 2.9701,
      "step": 87900
    },
    {
      "epoch": 23.668066322136088,
      "grad_norm": 0.4467654526233673,
      "learning_rate": 0.0002451501279116736,
      "loss": 2.9804,
      "step": 87950
    },
    {
      "epoch": 23.68152454780362,
      "grad_norm": 0.4790620505809784,
      "learning_rate": 0.0002449481621112158,
      "loss": 2.9741,
      "step": 88000
    },
    {
      "epoch": 23.68152454780362,
      "eval_accuracy": 0.39642869094011257,
      "eval_loss": 3.280144453048706,
      "eval_runtime": 84.3647,
      "eval_samples_per_second": 213.502,
      "eval_steps_per_second": 13.347,
      "step": 88000
    },
    {
      "epoch": 23.694982773471146,
      "grad_norm": 0.4683910310268402,
      "learning_rate": 0.000244746196310758,
      "loss": 2.9842,
      "step": 88050
    },
    {
      "epoch": 23.708440999138674,
      "grad_norm": 0.4699527323246002,
      "learning_rate": 0.0002445442305103002,
      "loss": 2.9769,
      "step": 88100
    },
    {
      "epoch": 23.7218992248062,
      "grad_norm": 0.441974014043808,
      "learning_rate": 0.00024434226470984246,
      "loss": 2.9869,
      "step": 88150
    },
    {
      "epoch": 23.735357450473728,
      "grad_norm": 0.4502110481262207,
      "learning_rate": 0.00024414029890938463,
      "loss": 2.9824,
      "step": 88200
    },
    {
      "epoch": 23.74881567614126,
      "grad_norm": 0.4554987847805023,
      "learning_rate": 0.00024393833310892686,
      "loss": 2.99,
      "step": 88250
    },
    {
      "epoch": 23.762273901808786,
      "grad_norm": 0.43465709686279297,
      "learning_rate": 0.0002437363673084691,
      "loss": 2.9847,
      "step": 88300
    },
    {
      "epoch": 23.775732127476314,
      "grad_norm": 0.4699002504348755,
      "learning_rate": 0.0002435344015080113,
      "loss": 2.9943,
      "step": 88350
    },
    {
      "epoch": 23.78919035314384,
      "grad_norm": 0.4456266760826111,
      "learning_rate": 0.0002433324357075535,
      "loss": 2.9872,
      "step": 88400
    },
    {
      "epoch": 23.802648578811368,
      "grad_norm": 0.46145564317703247,
      "learning_rate": 0.00024313046990709572,
      "loss": 2.9856,
      "step": 88450
    },
    {
      "epoch": 23.8161068044789,
      "grad_norm": 0.4430699050426483,
      "learning_rate": 0.00024292850410663792,
      "loss": 2.9957,
      "step": 88500
    },
    {
      "epoch": 23.829565030146426,
      "grad_norm": 0.46495842933654785,
      "learning_rate": 0.00024272653830618014,
      "loss": 2.9767,
      "step": 88550
    },
    {
      "epoch": 23.843023255813954,
      "grad_norm": 0.43364599347114563,
      "learning_rate": 0.00024252457250572234,
      "loss": 2.9894,
      "step": 88600
    },
    {
      "epoch": 23.85648148148148,
      "grad_norm": 0.4535362124443054,
      "learning_rate": 0.00024232260670526453,
      "loss": 2.9934,
      "step": 88650
    },
    {
      "epoch": 23.86993970714901,
      "grad_norm": 0.4435168504714966,
      "learning_rate": 0.00024212064090480676,
      "loss": 2.9941,
      "step": 88700
    },
    {
      "epoch": 23.88339793281654,
      "grad_norm": 0.4610646665096283,
      "learning_rate": 0.00024191867510434898,
      "loss": 2.9944,
      "step": 88750
    },
    {
      "epoch": 23.896856158484066,
      "grad_norm": 0.466998815536499,
      "learning_rate": 0.0002417167093038912,
      "loss": 2.9985,
      "step": 88800
    },
    {
      "epoch": 23.910314384151594,
      "grad_norm": 0.4404110908508301,
      "learning_rate": 0.0002415147435034334,
      "loss": 2.9946,
      "step": 88850
    },
    {
      "epoch": 23.92377260981912,
      "grad_norm": 0.43099141120910645,
      "learning_rate": 0.00024131277770297562,
      "loss": 2.9921,
      "step": 88900
    },
    {
      "epoch": 23.93723083548665,
      "grad_norm": 0.4604959785938263,
      "learning_rate": 0.00024111081190251782,
      "loss": 2.9974,
      "step": 88950
    },
    {
      "epoch": 23.95068906115418,
      "grad_norm": 0.4126609265804291,
      "learning_rate": 0.00024090884610206002,
      "loss": 2.9865,
      "step": 89000
    },
    {
      "epoch": 23.95068906115418,
      "eval_accuracy": 0.39700060798780373,
      "eval_loss": 3.2750401496887207,
      "eval_runtime": 84.5116,
      "eval_samples_per_second": 213.13,
      "eval_steps_per_second": 13.324,
      "step": 89000
    },
    {
      "epoch": 23.964147286821706,
      "grad_norm": 0.42181965708732605,
      "learning_rate": 0.00024070688030160224,
      "loss": 3.0037,
      "step": 89050
    },
    {
      "epoch": 23.977605512489234,
      "grad_norm": 0.452213853597641,
      "learning_rate": 0.00024050491450114443,
      "loss": 2.9859,
      "step": 89100
    },
    {
      "epoch": 23.99106373815676,
      "grad_norm": 0.4670252501964569,
      "learning_rate": 0.00024030294870068666,
      "loss": 2.9958,
      "step": 89150
    },
    {
      "epoch": 24.00430663221361,
      "grad_norm": 0.4971795976161957,
      "learning_rate": 0.00024010098290022888,
      "loss": 2.9672,
      "step": 89200
    },
    {
      "epoch": 24.017764857881136,
      "grad_norm": 0.4482610821723938,
      "learning_rate": 0.0002398990170997711,
      "loss": 2.9152,
      "step": 89250
    },
    {
      "epoch": 24.031223083548664,
      "grad_norm": 0.4758496582508087,
      "learning_rate": 0.0002396970512993133,
      "loss": 2.9161,
      "step": 89300
    },
    {
      "epoch": 24.044681309216195,
      "grad_norm": 0.4543116092681885,
      "learning_rate": 0.00023949508549885552,
      "loss": 2.9114,
      "step": 89350
    },
    {
      "epoch": 24.058139534883722,
      "grad_norm": 0.4571596384048462,
      "learning_rate": 0.00023929311969839772,
      "loss": 2.9164,
      "step": 89400
    },
    {
      "epoch": 24.07159776055125,
      "grad_norm": 0.4987284243106842,
      "learning_rate": 0.00023909115389793992,
      "loss": 2.9166,
      "step": 89450
    },
    {
      "epoch": 24.085055986218777,
      "grad_norm": 0.4575844705104828,
      "learning_rate": 0.00023888918809748214,
      "loss": 2.9232,
      "step": 89500
    },
    {
      "epoch": 24.098514211886304,
      "grad_norm": 0.4510669708251953,
      "learning_rate": 0.00023868722229702433,
      "loss": 2.9222,
      "step": 89550
    },
    {
      "epoch": 24.11197243755383,
      "grad_norm": 0.45965123176574707,
      "learning_rate": 0.00023848525649656656,
      "loss": 2.9339,
      "step": 89600
    },
    {
      "epoch": 24.125430663221362,
      "grad_norm": 0.4233417212963104,
      "learning_rate": 0.00023828329069610878,
      "loss": 2.9215,
      "step": 89650
    },
    {
      "epoch": 24.13888888888889,
      "grad_norm": 0.44202446937561035,
      "learning_rate": 0.000238081324895651,
      "loss": 2.9268,
      "step": 89700
    },
    {
      "epoch": 24.152347114556417,
      "grad_norm": 0.45001307129859924,
      "learning_rate": 0.0002378793590951932,
      "loss": 2.9262,
      "step": 89750
    },
    {
      "epoch": 24.165805340223944,
      "grad_norm": 0.47428709268569946,
      "learning_rate": 0.0002376773932947354,
      "loss": 2.9411,
      "step": 89800
    },
    {
      "epoch": 24.17926356589147,
      "grad_norm": 0.4623621106147766,
      "learning_rate": 0.00023747542749427762,
      "loss": 2.9332,
      "step": 89850
    },
    {
      "epoch": 24.192721791559002,
      "grad_norm": 0.439605712890625,
      "learning_rate": 0.00023727346169381982,
      "loss": 2.9391,
      "step": 89900
    },
    {
      "epoch": 24.20618001722653,
      "grad_norm": 0.4671004116535187,
      "learning_rate": 0.00023707149589336204,
      "loss": 2.9257,
      "step": 89950
    },
    {
      "epoch": 24.219638242894057,
      "grad_norm": 0.45063653588294983,
      "learning_rate": 0.00023686953009290424,
      "loss": 2.9469,
      "step": 90000
    },
    {
      "epoch": 24.219638242894057,
      "eval_accuracy": 0.3961077481287752,
      "eval_loss": 3.2907376289367676,
      "eval_runtime": 84.5507,
      "eval_samples_per_second": 213.032,
      "eval_steps_per_second": 13.317,
      "step": 90000
    },
    {
      "epoch": 24.233096468561584,
      "grad_norm": 0.4493043124675751,
      "learning_rate": 0.00023666756429244643,
      "loss": 2.9382,
      "step": 90050
    },
    {
      "epoch": 24.24655469422911,
      "grad_norm": 0.4637029767036438,
      "learning_rate": 0.00023646559849198868,
      "loss": 2.9453,
      "step": 90100
    },
    {
      "epoch": 24.260012919896642,
      "grad_norm": 0.4731544852256775,
      "learning_rate": 0.00023626363269153088,
      "loss": 2.9403,
      "step": 90150
    },
    {
      "epoch": 24.27347114556417,
      "grad_norm": 0.4753323793411255,
      "learning_rate": 0.0002360616668910731,
      "loss": 2.9524,
      "step": 90200
    },
    {
      "epoch": 24.286929371231697,
      "grad_norm": 0.43400460481643677,
      "learning_rate": 0.0002358597010906153,
      "loss": 2.9453,
      "step": 90250
    },
    {
      "epoch": 24.300387596899224,
      "grad_norm": 0.45669588446617126,
      "learning_rate": 0.00023565773529015752,
      "loss": 2.9584,
      "step": 90300
    },
    {
      "epoch": 24.31384582256675,
      "grad_norm": 0.4351573884487152,
      "learning_rate": 0.00023545576948969972,
      "loss": 2.9432,
      "step": 90350
    },
    {
      "epoch": 24.327304048234282,
      "grad_norm": 0.4844313859939575,
      "learning_rate": 0.00023525380368924194,
      "loss": 2.9543,
      "step": 90400
    },
    {
      "epoch": 24.34076227390181,
      "grad_norm": 0.4621292054653168,
      "learning_rate": 0.00023505183788878414,
      "loss": 2.9483,
      "step": 90450
    },
    {
      "epoch": 24.354220499569337,
      "grad_norm": 0.47124025225639343,
      "learning_rate": 0.00023484987208832633,
      "loss": 2.9648,
      "step": 90500
    },
    {
      "epoch": 24.367678725236864,
      "grad_norm": 0.43759214878082275,
      "learning_rate": 0.00023464790628786858,
      "loss": 2.9608,
      "step": 90550
    },
    {
      "epoch": 24.38113695090439,
      "grad_norm": 0.46638819575309753,
      "learning_rate": 0.00023444594048741078,
      "loss": 2.9561,
      "step": 90600
    },
    {
      "epoch": 24.394595176571922,
      "grad_norm": 0.47136223316192627,
      "learning_rate": 0.000234243974686953,
      "loss": 2.9534,
      "step": 90650
    },
    {
      "epoch": 24.40805340223945,
      "grad_norm": 0.45332977175712585,
      "learning_rate": 0.0002340420088864952,
      "loss": 2.9429,
      "step": 90700
    },
    {
      "epoch": 24.421511627906977,
      "grad_norm": 0.4295179545879364,
      "learning_rate": 0.00023384004308603742,
      "loss": 2.9604,
      "step": 90750
    },
    {
      "epoch": 24.434969853574504,
      "grad_norm": 0.44807741045951843,
      "learning_rate": 0.00023363807728557962,
      "loss": 2.9575,
      "step": 90800
    },
    {
      "epoch": 24.44842807924203,
      "grad_norm": 0.4464702010154724,
      "learning_rate": 0.0002334361114851218,
      "loss": 2.9665,
      "step": 90850
    },
    {
      "epoch": 24.461886304909562,
      "grad_norm": 0.42546746134757996,
      "learning_rate": 0.00023323414568466404,
      "loss": 2.9541,
      "step": 90900
    },
    {
      "epoch": 24.47534453057709,
      "grad_norm": 0.46260973811149597,
      "learning_rate": 0.00023303217988420623,
      "loss": 2.9679,
      "step": 90950
    },
    {
      "epoch": 24.488802756244617,
      "grad_norm": 0.4452535808086395,
      "learning_rate": 0.00023283021408374848,
      "loss": 2.9622,
      "step": 91000
    },
    {
      "epoch": 24.488802756244617,
      "eval_accuracy": 0.39678146728013025,
      "eval_loss": 3.2826082706451416,
      "eval_runtime": 84.4752,
      "eval_samples_per_second": 213.222,
      "eval_steps_per_second": 13.329,
      "step": 91000
    },
    {
      "epoch": 24.502260981912144,
      "grad_norm": 0.44478604197502136,
      "learning_rate": 0.00023262824828329068,
      "loss": 2.9602,
      "step": 91050
    },
    {
      "epoch": 24.51571920757967,
      "grad_norm": 0.4625186026096344,
      "learning_rate": 0.0002324262824828329,
      "loss": 2.9719,
      "step": 91100
    },
    {
      "epoch": 24.529177433247202,
      "grad_norm": 0.45984506607055664,
      "learning_rate": 0.0002322243166823751,
      "loss": 2.9671,
      "step": 91150
    },
    {
      "epoch": 24.54263565891473,
      "grad_norm": 0.42750629782676697,
      "learning_rate": 0.0002320223508819173,
      "loss": 2.9661,
      "step": 91200
    },
    {
      "epoch": 24.556093884582257,
      "grad_norm": 0.45963552594184875,
      "learning_rate": 0.00023182038508145952,
      "loss": 2.9601,
      "step": 91250
    },
    {
      "epoch": 24.569552110249784,
      "grad_norm": 0.45868661999702454,
      "learning_rate": 0.0002316184192810017,
      "loss": 2.9586,
      "step": 91300
    },
    {
      "epoch": 24.58301033591731,
      "grad_norm": 0.44878077507019043,
      "learning_rate": 0.00023141645348054394,
      "loss": 2.967,
      "step": 91350
    },
    {
      "epoch": 24.596468561584842,
      "grad_norm": 0.4787624478340149,
      "learning_rate": 0.00023121448768008613,
      "loss": 2.9715,
      "step": 91400
    },
    {
      "epoch": 24.60992678725237,
      "grad_norm": 0.4520249366760254,
      "learning_rate": 0.00023101252187962838,
      "loss": 2.9671,
      "step": 91450
    },
    {
      "epoch": 24.623385012919897,
      "grad_norm": 0.48157331347465515,
      "learning_rate": 0.00023081055607917058,
      "loss": 2.9613,
      "step": 91500
    },
    {
      "epoch": 24.636843238587424,
      "grad_norm": 0.4596235454082489,
      "learning_rate": 0.0002306085902787128,
      "loss": 2.9667,
      "step": 91550
    },
    {
      "epoch": 24.65030146425495,
      "grad_norm": 0.44594740867614746,
      "learning_rate": 0.000230406624478255,
      "loss": 2.971,
      "step": 91600
    },
    {
      "epoch": 24.66375968992248,
      "grad_norm": 0.45610886812210083,
      "learning_rate": 0.0002302046586777972,
      "loss": 2.9681,
      "step": 91650
    },
    {
      "epoch": 24.67721791559001,
      "grad_norm": 0.46916455030441284,
      "learning_rate": 0.00023000269287733942,
      "loss": 2.9711,
      "step": 91700
    },
    {
      "epoch": 24.690676141257537,
      "grad_norm": 0.4882563054561615,
      "learning_rate": 0.00022980072707688161,
      "loss": 2.9659,
      "step": 91750
    },
    {
      "epoch": 24.704134366925064,
      "grad_norm": 0.4524535834789276,
      "learning_rate": 0.00022959876127642384,
      "loss": 2.9676,
      "step": 91800
    },
    {
      "epoch": 24.71759259259259,
      "grad_norm": 0.4534264802932739,
      "learning_rate": 0.00022939679547596603,
      "loss": 2.9763,
      "step": 91850
    },
    {
      "epoch": 24.73105081826012,
      "grad_norm": 0.4541570842266083,
      "learning_rate": 0.00022919482967550828,
      "loss": 2.9837,
      "step": 91900
    },
    {
      "epoch": 24.74450904392765,
      "grad_norm": 0.4537384808063507,
      "learning_rate": 0.00022899286387505048,
      "loss": 2.9694,
      "step": 91950
    },
    {
      "epoch": 24.757967269595177,
      "grad_norm": 0.43848171830177307,
      "learning_rate": 0.00022879089807459268,
      "loss": 2.9704,
      "step": 92000
    },
    {
      "epoch": 24.757967269595177,
      "eval_accuracy": 0.3968452429843466,
      "eval_loss": 3.279933452606201,
      "eval_runtime": 84.3195,
      "eval_samples_per_second": 213.616,
      "eval_steps_per_second": 13.354,
      "step": 92000
    },
    {
      "epoch": 24.771425495262704,
      "grad_norm": 0.45227792859077454,
      "learning_rate": 0.0002285889322741349,
      "loss": 2.9856,
      "step": 92050
    },
    {
      "epoch": 24.78488372093023,
      "grad_norm": 0.44843050837516785,
      "learning_rate": 0.0002283869664736771,
      "loss": 2.9728,
      "step": 92100
    },
    {
      "epoch": 24.79834194659776,
      "grad_norm": 0.455285906791687,
      "learning_rate": 0.00022818500067321932,
      "loss": 2.9708,
      "step": 92150
    },
    {
      "epoch": 24.81180017226529,
      "grad_norm": 0.45710763335227966,
      "learning_rate": 0.00022798303487276151,
      "loss": 2.9744,
      "step": 92200
    },
    {
      "epoch": 24.825258397932817,
      "grad_norm": 0.4339000880718231,
      "learning_rate": 0.0002277810690723037,
      "loss": 2.976,
      "step": 92250
    },
    {
      "epoch": 24.838716623600344,
      "grad_norm": 0.4364817142486572,
      "learning_rate": 0.00022757910327184596,
      "loss": 2.9744,
      "step": 92300
    },
    {
      "epoch": 24.852174849267872,
      "grad_norm": 0.4568532705307007,
      "learning_rate": 0.00022737713747138818,
      "loss": 2.9904,
      "step": 92350
    },
    {
      "epoch": 24.8656330749354,
      "grad_norm": 0.4795514643192291,
      "learning_rate": 0.00022717517167093038,
      "loss": 2.9797,
      "step": 92400
    },
    {
      "epoch": 24.87909130060293,
      "grad_norm": 0.47505083680152893,
      "learning_rate": 0.00022697320587047258,
      "loss": 2.9805,
      "step": 92450
    },
    {
      "epoch": 24.892549526270457,
      "grad_norm": 0.4372386038303375,
      "learning_rate": 0.0002267712400700148,
      "loss": 2.9872,
      "step": 92500
    },
    {
      "epoch": 24.906007751937985,
      "grad_norm": 0.4770386517047882,
      "learning_rate": 0.000226569274269557,
      "loss": 2.9903,
      "step": 92550
    },
    {
      "epoch": 24.919465977605512,
      "grad_norm": 0.45319411158561707,
      "learning_rate": 0.00022636730846909922,
      "loss": 2.9664,
      "step": 92600
    },
    {
      "epoch": 24.93292420327304,
      "grad_norm": 0.4697992503643036,
      "learning_rate": 0.00022616534266864141,
      "loss": 2.9843,
      "step": 92650
    },
    {
      "epoch": 24.94638242894057,
      "grad_norm": 0.4561481773853302,
      "learning_rate": 0.0002259633768681836,
      "loss": 2.9747,
      "step": 92700
    },
    {
      "epoch": 24.959840654608097,
      "grad_norm": 0.4588293135166168,
      "learning_rate": 0.00022576141106772586,
      "loss": 2.969,
      "step": 92750
    },
    {
      "epoch": 24.973298880275625,
      "grad_norm": 0.48158708214759827,
      "learning_rate": 0.00022555944526726806,
      "loss": 2.985,
      "step": 92800
    },
    {
      "epoch": 24.986757105943152,
      "grad_norm": 0.44852331280708313,
      "learning_rate": 0.00022535747946681028,
      "loss": 2.9814,
      "step": 92850
    },
    {
      "epoch": 25.0,
      "grad_norm": 0.9995496869087219,
      "learning_rate": 0.00022515551366635248,
      "loss": 2.9722,
      "step": 92900
    },
    {
      "epoch": 25.013458225667527,
      "grad_norm": 0.4474904239177704,
      "learning_rate": 0.0002249535478658947,
      "loss": 2.898,
      "step": 92950
    },
    {
      "epoch": 25.026916451335055,
      "grad_norm": 0.4777752161026001,
      "learning_rate": 0.0002247515820654369,
      "loss": 2.9062,
      "step": 93000
    },
    {
      "epoch": 25.026916451335055,
      "eval_accuracy": 0.39641043826837774,
      "eval_loss": 3.286386013031006,
      "eval_runtime": 84.1972,
      "eval_samples_per_second": 213.926,
      "eval_steps_per_second": 13.373,
      "step": 93000
    },
    {
      "epoch": 25.040374677002585,
      "grad_norm": 0.46315711736679077,
      "learning_rate": 0.0002245496162649791,
      "loss": 2.9057,
      "step": 93050
    },
    {
      "epoch": 25.053832902670113,
      "grad_norm": 0.46571436524391174,
      "learning_rate": 0.00022434765046452131,
      "loss": 2.9087,
      "step": 93100
    },
    {
      "epoch": 25.06729112833764,
      "grad_norm": 0.4476296007633209,
      "learning_rate": 0.0002241456846640635,
      "loss": 2.9127,
      "step": 93150
    },
    {
      "epoch": 25.080749354005167,
      "grad_norm": 0.47912272810935974,
      "learning_rate": 0.00022394371886360576,
      "loss": 2.9086,
      "step": 93200
    },
    {
      "epoch": 25.094207579672695,
      "grad_norm": 0.5244853496551514,
      "learning_rate": 0.00022374175306314796,
      "loss": 2.9135,
      "step": 93250
    },
    {
      "epoch": 25.107665805340226,
      "grad_norm": 0.47766560316085815,
      "learning_rate": 0.00022353978726269018,
      "loss": 2.9176,
      "step": 93300
    },
    {
      "epoch": 25.121124031007753,
      "grad_norm": 0.4383608400821686,
      "learning_rate": 0.00022333782146223238,
      "loss": 2.9296,
      "step": 93350
    },
    {
      "epoch": 25.13458225667528,
      "grad_norm": 0.45898735523223877,
      "learning_rate": 0.0002231358556617746,
      "loss": 2.9278,
      "step": 93400
    },
    {
      "epoch": 25.148040482342807,
      "grad_norm": 0.47586843371391296,
      "learning_rate": 0.0002229338898613168,
      "loss": 2.9204,
      "step": 93450
    },
    {
      "epoch": 25.161498708010335,
      "grad_norm": 0.47148630023002625,
      "learning_rate": 0.000222731924060859,
      "loss": 2.9385,
      "step": 93500
    },
    {
      "epoch": 25.174956933677866,
      "grad_norm": 0.44435662031173706,
      "learning_rate": 0.00022252995826040122,
      "loss": 2.9306,
      "step": 93550
    },
    {
      "epoch": 25.188415159345393,
      "grad_norm": 0.4764225482940674,
      "learning_rate": 0.0002223279924599434,
      "loss": 2.9267,
      "step": 93600
    },
    {
      "epoch": 25.20187338501292,
      "grad_norm": 0.46969074010849,
      "learning_rate": 0.00022212602665948566,
      "loss": 2.9237,
      "step": 93650
    },
    {
      "epoch": 25.215331610680447,
      "grad_norm": 0.44603314995765686,
      "learning_rate": 0.00022192406085902786,
      "loss": 2.9317,
      "step": 93700
    },
    {
      "epoch": 25.228789836347975,
      "grad_norm": 0.4528902769088745,
      "learning_rate": 0.00022172209505857008,
      "loss": 2.9429,
      "step": 93750
    },
    {
      "epoch": 25.242248062015506,
      "grad_norm": 0.5264995694160461,
      "learning_rate": 0.00022152012925811228,
      "loss": 2.9356,
      "step": 93800
    },
    {
      "epoch": 25.255706287683033,
      "grad_norm": 0.45114192366600037,
      "learning_rate": 0.00022131816345765447,
      "loss": 2.9461,
      "step": 93850
    },
    {
      "epoch": 25.26916451335056,
      "grad_norm": 0.5063636302947998,
      "learning_rate": 0.0002211161976571967,
      "loss": 2.9366,
      "step": 93900
    },
    {
      "epoch": 25.282622739018088,
      "grad_norm": 0.44546231627464294,
      "learning_rate": 0.0002209142318567389,
      "loss": 2.9423,
      "step": 93950
    },
    {
      "epoch": 25.296080964685615,
      "grad_norm": 0.4745674431324005,
      "learning_rate": 0.00022071226605628112,
      "loss": 2.9439,
      "step": 94000
    },
    {
      "epoch": 25.296080964685615,
      "eval_accuracy": 0.3964877948295396,
      "eval_loss": 3.287885904312134,
      "eval_runtime": 83.308,
      "eval_samples_per_second": 216.21,
      "eval_steps_per_second": 13.516,
      "step": 94000
    },
    {
      "epoch": 25.309539190353142,
      "grad_norm": 0.4587896764278412,
      "learning_rate": 0.0002205103002558233,
      "loss": 2.9311,
      "step": 94050
    },
    {
      "epoch": 25.322997416020673,
      "grad_norm": 0.4565125107765198,
      "learning_rate": 0.00022030833445536556,
      "loss": 2.9197,
      "step": 94100
    },
    {
      "epoch": 25.3364556416882,
      "grad_norm": 0.44744712114334106,
      "learning_rate": 0.00022010636865490776,
      "loss": 2.9223,
      "step": 94150
    },
    {
      "epoch": 25.349913867355728,
      "grad_norm": 0.460087388753891,
      "learning_rate": 0.00021990440285444998,
      "loss": 2.9415,
      "step": 94200
    },
    {
      "epoch": 25.363372093023255,
      "grad_norm": 0.49062275886535645,
      "learning_rate": 0.00021970243705399218,
      "loss": 2.9499,
      "step": 94250
    },
    {
      "epoch": 25.376830318690782,
      "grad_norm": 0.47762060165405273,
      "learning_rate": 0.00021950047125353437,
      "loss": 2.9395,
      "step": 94300
    },
    {
      "epoch": 25.390288544358313,
      "grad_norm": 0.46556368470191956,
      "learning_rate": 0.0002192985054530766,
      "loss": 2.9315,
      "step": 94350
    },
    {
      "epoch": 25.40374677002584,
      "grad_norm": 0.47144410014152527,
      "learning_rate": 0.0002190965396526188,
      "loss": 2.9388,
      "step": 94400
    },
    {
      "epoch": 25.417204995693368,
      "grad_norm": 0.4625619351863861,
      "learning_rate": 0.00021889457385216102,
      "loss": 2.9488,
      "step": 94450
    },
    {
      "epoch": 25.430663221360895,
      "grad_norm": 0.4657180905342102,
      "learning_rate": 0.0002186926080517032,
      "loss": 2.9426,
      "step": 94500
    },
    {
      "epoch": 25.444121447028422,
      "grad_norm": 0.4904530942440033,
      "learning_rate": 0.00021849064225124546,
      "loss": 2.9506,
      "step": 94550
    },
    {
      "epoch": 25.457579672695953,
      "grad_norm": 0.46393194794654846,
      "learning_rate": 0.00021828867645078766,
      "loss": 2.9448,
      "step": 94600
    },
    {
      "epoch": 25.47103789836348,
      "grad_norm": 0.4495938718318939,
      "learning_rate": 0.00021808671065032985,
      "loss": 2.948,
      "step": 94650
    },
    {
      "epoch": 25.484496124031008,
      "grad_norm": 0.517048180103302,
      "learning_rate": 0.00021788474484987208,
      "loss": 2.9419,
      "step": 94700
    },
    {
      "epoch": 25.497954349698535,
      "grad_norm": 0.49732452630996704,
      "learning_rate": 0.00021768277904941427,
      "loss": 2.9534,
      "step": 94750
    },
    {
      "epoch": 25.511412575366062,
      "grad_norm": 0.46624764800071716,
      "learning_rate": 0.0002174808132489565,
      "loss": 2.9585,
      "step": 94800
    },
    {
      "epoch": 25.524870801033593,
      "grad_norm": 0.5083665251731873,
      "learning_rate": 0.0002172788474484987,
      "loss": 2.959,
      "step": 94850
    },
    {
      "epoch": 25.53832902670112,
      "grad_norm": 0.4506458640098572,
      "learning_rate": 0.0002170768816480409,
      "loss": 2.9549,
      "step": 94900
    },
    {
      "epoch": 25.551787252368648,
      "grad_norm": 0.45113179087638855,
      "learning_rate": 0.0002168749158475831,
      "loss": 2.9551,
      "step": 94950
    },
    {
      "epoch": 25.565245478036175,
      "grad_norm": 0.4966818392276764,
      "learning_rate": 0.00021667295004712534,
      "loss": 2.9474,
      "step": 95000
    },
    {
      "epoch": 25.565245478036175,
      "eval_accuracy": 0.39686925393942635,
      "eval_loss": 3.282986640930176,
      "eval_runtime": 83.2523,
      "eval_samples_per_second": 216.354,
      "eval_steps_per_second": 13.525,
      "step": 95000
    },
    {
      "epoch": 25.578703703703702,
      "grad_norm": 0.4647633731365204,
      "learning_rate": 0.00021647098424666756,
      "loss": 2.9585,
      "step": 95050
    },
    {
      "epoch": 25.592161929371233,
      "grad_norm": 0.4701825678348541,
      "learning_rate": 0.00021626901844620976,
      "loss": 2.9567,
      "step": 95100
    },
    {
      "epoch": 25.60562015503876,
      "grad_norm": 0.45861950516700745,
      "learning_rate": 0.00021606705264575198,
      "loss": 2.9636,
      "step": 95150
    },
    {
      "epoch": 25.619078380706288,
      "grad_norm": 0.4622119069099426,
      "learning_rate": 0.00021586508684529417,
      "loss": 2.9653,
      "step": 95200
    },
    {
      "epoch": 25.632536606373815,
      "grad_norm": 0.4949166476726532,
      "learning_rate": 0.0002156631210448364,
      "loss": 2.9618,
      "step": 95250
    },
    {
      "epoch": 25.645994832041342,
      "grad_norm": 0.4535043239593506,
      "learning_rate": 0.0002154611552443786,
      "loss": 2.9644,
      "step": 95300
    },
    {
      "epoch": 25.659453057708873,
      "grad_norm": 0.4799681603908539,
      "learning_rate": 0.0002152591894439208,
      "loss": 2.9582,
      "step": 95350
    },
    {
      "epoch": 25.6729112833764,
      "grad_norm": 0.46847987174987793,
      "learning_rate": 0.000215057223643463,
      "loss": 2.9495,
      "step": 95400
    },
    {
      "epoch": 25.686369509043928,
      "grad_norm": 0.4924308657646179,
      "learning_rate": 0.00021485525784300524,
      "loss": 2.9512,
      "step": 95450
    },
    {
      "epoch": 25.699827734711455,
      "grad_norm": 0.4683797061443329,
      "learning_rate": 0.00021465329204254746,
      "loss": 2.9673,
      "step": 95500
    },
    {
      "epoch": 25.713285960378983,
      "grad_norm": 0.48432156443595886,
      "learning_rate": 0.00021445132624208966,
      "loss": 2.9674,
      "step": 95550
    },
    {
      "epoch": 25.72674418604651,
      "grad_norm": 0.46764570474624634,
      "learning_rate": 0.00021424936044163188,
      "loss": 2.9675,
      "step": 95600
    },
    {
      "epoch": 25.74020241171404,
      "grad_norm": 0.4551568031311035,
      "learning_rate": 0.00021404739464117407,
      "loss": 2.9624,
      "step": 95650
    },
    {
      "epoch": 25.753660637381568,
      "grad_norm": 0.48788902163505554,
      "learning_rate": 0.00021384542884071627,
      "loss": 2.9667,
      "step": 95700
    },
    {
      "epoch": 25.767118863049095,
      "grad_norm": 0.46860629320144653,
      "learning_rate": 0.0002136434630402585,
      "loss": 2.9644,
      "step": 95750
    },
    {
      "epoch": 25.780577088716623,
      "grad_norm": 0.4676167368888855,
      "learning_rate": 0.0002134414972398007,
      "loss": 2.9705,
      "step": 95800
    },
    {
      "epoch": 25.794035314384153,
      "grad_norm": 0.46130162477493286,
      "learning_rate": 0.0002132395314393429,
      "loss": 2.962,
      "step": 95850
    },
    {
      "epoch": 25.80749354005168,
      "grad_norm": 0.4687904119491577,
      "learning_rate": 0.00021303756563888514,
      "loss": 2.9669,
      "step": 95900
    },
    {
      "epoch": 25.820951765719208,
      "grad_norm": 0.43597960472106934,
      "learning_rate": 0.00021283559983842736,
      "loss": 2.976,
      "step": 95950
    },
    {
      "epoch": 25.834409991386735,
      "grad_norm": 0.4919257164001465,
      "learning_rate": 0.00021263363403796956,
      "loss": 2.966,
      "step": 96000
    },
    {
      "epoch": 25.834409991386735,
      "eval_accuracy": 0.39693509393389836,
      "eval_loss": 3.2789909839630127,
      "eval_runtime": 83.7019,
      "eval_samples_per_second": 215.192,
      "eval_steps_per_second": 13.453,
      "step": 96000
    },
    {
      "epoch": 25.847868217054263,
      "grad_norm": 0.4576070308685303,
      "learning_rate": 0.00021243166823751175,
      "loss": 2.9521,
      "step": 96050
    },
    {
      "epoch": 25.86132644272179,
      "grad_norm": 0.45357754826545715,
      "learning_rate": 0.00021222970243705398,
      "loss": 2.9583,
      "step": 96100
    },
    {
      "epoch": 25.87478466838932,
      "grad_norm": 0.47627171874046326,
      "learning_rate": 0.00021202773663659617,
      "loss": 2.9624,
      "step": 96150
    },
    {
      "epoch": 25.888242894056848,
      "grad_norm": 0.46957269310951233,
      "learning_rate": 0.0002118257708361384,
      "loss": 2.9553,
      "step": 96200
    },
    {
      "epoch": 25.901701119724375,
      "grad_norm": 0.4511432945728302,
      "learning_rate": 0.0002116238050356806,
      "loss": 2.967,
      "step": 96250
    },
    {
      "epoch": 25.915159345391903,
      "grad_norm": 0.4503208100795746,
      "learning_rate": 0.00021142183923522281,
      "loss": 2.9781,
      "step": 96300
    },
    {
      "epoch": 25.92861757105943,
      "grad_norm": 0.4600832760334015,
      "learning_rate": 0.00021121987343476504,
      "loss": 2.9741,
      "step": 96350
    },
    {
      "epoch": 25.94207579672696,
      "grad_norm": 0.4865885376930237,
      "learning_rate": 0.00021101790763430726,
      "loss": 2.9742,
      "step": 96400
    },
    {
      "epoch": 25.955534022394488,
      "grad_norm": 0.44817084074020386,
      "learning_rate": 0.00021081594183384946,
      "loss": 2.9739,
      "step": 96450
    },
    {
      "epoch": 25.968992248062015,
      "grad_norm": 0.4550294280052185,
      "learning_rate": 0.00021061397603339165,
      "loss": 2.9716,
      "step": 96500
    },
    {
      "epoch": 25.982450473729543,
      "grad_norm": 0.47818243503570557,
      "learning_rate": 0.00021041201023293388,
      "loss": 2.9627,
      "step": 96550
    },
    {
      "epoch": 25.99590869939707,
      "grad_norm": 0.44311001896858215,
      "learning_rate": 0.00021021004443247607,
      "loss": 2.9881,
      "step": 96600
    },
    {
      "epoch": 26.009151593453918,
      "grad_norm": 0.46623387932777405,
      "learning_rate": 0.0002100080786320183,
      "loss": 2.9238,
      "step": 96650
    },
    {
      "epoch": 26.022609819121445,
      "grad_norm": 0.4871658384799957,
      "learning_rate": 0.0002098061128315605,
      "loss": 2.8875,
      "step": 96700
    },
    {
      "epoch": 26.036068044788976,
      "grad_norm": 0.47384247183799744,
      "learning_rate": 0.00020960414703110274,
      "loss": 2.8927,
      "step": 96750
    },
    {
      "epoch": 26.049526270456504,
      "grad_norm": 0.49783265590667725,
      "learning_rate": 0.00020940218123064494,
      "loss": 2.8936,
      "step": 96800
    },
    {
      "epoch": 26.06298449612403,
      "grad_norm": 0.46611055731773376,
      "learning_rate": 0.00020920021543018713,
      "loss": 2.8996,
      "step": 96850
    },
    {
      "epoch": 26.076442721791558,
      "grad_norm": 0.4554579257965088,
      "learning_rate": 0.00020899824962972936,
      "loss": 2.908,
      "step": 96900
    },
    {
      "epoch": 26.089900947459086,
      "grad_norm": 0.4305267333984375,
      "learning_rate": 0.00020879628382927155,
      "loss": 2.918,
      "step": 96950
    },
    {
      "epoch": 26.103359173126616,
      "grad_norm": 0.4762580096721649,
      "learning_rate": 0.00020859431802881378,
      "loss": 2.9114,
      "step": 97000
    },
    {
      "epoch": 26.103359173126616,
      "eval_accuracy": 0.3964316244052128,
      "eval_loss": 3.2921440601348877,
      "eval_runtime": 83.6181,
      "eval_samples_per_second": 215.408,
      "eval_steps_per_second": 13.466,
      "step": 97000
    },
    {
      "epoch": 26.116817398794144,
      "grad_norm": 0.4823869466781616,
      "learning_rate": 0.00020839235222835597,
      "loss": 2.9166,
      "step": 97050
    },
    {
      "epoch": 26.13027562446167,
      "grad_norm": 0.5001690983772278,
      "learning_rate": 0.00020819038642789817,
      "loss": 2.9095,
      "step": 97100
    },
    {
      "epoch": 26.1437338501292,
      "grad_norm": 0.4815123975276947,
      "learning_rate": 0.0002079884206274404,
      "loss": 2.9056,
      "step": 97150
    },
    {
      "epoch": 26.157192075796726,
      "grad_norm": 0.472492516040802,
      "learning_rate": 0.00020778645482698264,
      "loss": 2.9101,
      "step": 97200
    },
    {
      "epoch": 26.170650301464256,
      "grad_norm": 0.4798564612865448,
      "learning_rate": 0.00020758448902652484,
      "loss": 2.9224,
      "step": 97250
    },
    {
      "epoch": 26.184108527131784,
      "grad_norm": 0.4941078722476959,
      "learning_rate": 0.00020738252322606703,
      "loss": 2.9259,
      "step": 97300
    },
    {
      "epoch": 26.19756675279931,
      "grad_norm": 0.47859832644462585,
      "learning_rate": 0.00020718055742560926,
      "loss": 2.912,
      "step": 97350
    },
    {
      "epoch": 26.21102497846684,
      "grad_norm": 0.4728969931602478,
      "learning_rate": 0.00020697859162515145,
      "loss": 2.921,
      "step": 97400
    },
    {
      "epoch": 26.224483204134366,
      "grad_norm": 0.4580983817577362,
      "learning_rate": 0.00020677662582469368,
      "loss": 2.9153,
      "step": 97450
    },
    {
      "epoch": 26.237941429801896,
      "grad_norm": 0.44771474599838257,
      "learning_rate": 0.00020657466002423587,
      "loss": 2.9296,
      "step": 97500
    },
    {
      "epoch": 26.251399655469424,
      "grad_norm": 0.4706652760505676,
      "learning_rate": 0.00020637269422377807,
      "loss": 2.9202,
      "step": 97550
    },
    {
      "epoch": 26.26485788113695,
      "grad_norm": 0.47193393111228943,
      "learning_rate": 0.0002061707284233203,
      "loss": 2.9311,
      "step": 97600
    },
    {
      "epoch": 26.27831610680448,
      "grad_norm": 0.46989378333091736,
      "learning_rate": 0.00020596876262286251,
      "loss": 2.9232,
      "step": 97650
    },
    {
      "epoch": 26.291774332472006,
      "grad_norm": 0.48723989725112915,
      "learning_rate": 0.00020576679682240474,
      "loss": 2.9268,
      "step": 97700
    },
    {
      "epoch": 26.305232558139537,
      "grad_norm": 0.47716909646987915,
      "learning_rate": 0.00020556483102194693,
      "loss": 2.9206,
      "step": 97750
    },
    {
      "epoch": 26.318690783807064,
      "grad_norm": 0.4953400492668152,
      "learning_rate": 0.00020536286522148916,
      "loss": 2.9408,
      "step": 97800
    },
    {
      "epoch": 26.33214900947459,
      "grad_norm": 0.45694100856781006,
      "learning_rate": 0.00020516089942103135,
      "loss": 2.9265,
      "step": 97850
    },
    {
      "epoch": 26.34560723514212,
      "grad_norm": 0.4777909815311432,
      "learning_rate": 0.00020495893362057355,
      "loss": 2.9366,
      "step": 97900
    },
    {
      "epoch": 26.359065460809646,
      "grad_norm": 0.48248589038848877,
      "learning_rate": 0.00020475696782011577,
      "loss": 2.9211,
      "step": 97950
    },
    {
      "epoch": 26.372523686477177,
      "grad_norm": 0.47510600090026855,
      "learning_rate": 0.00020455500201965797,
      "loss": 2.9351,
      "step": 98000
    },
    {
      "epoch": 26.372523686477177,
      "eval_accuracy": 0.39670932576803547,
      "eval_loss": 3.2853517532348633,
      "eval_runtime": 84.0898,
      "eval_samples_per_second": 214.199,
      "eval_steps_per_second": 13.39,
      "step": 98000
    },
    {
      "epoch": 26.385981912144704,
      "grad_norm": 0.4783083200454712,
      "learning_rate": 0.0002043530362192002,
      "loss": 2.9364,
      "step": 98050
    },
    {
      "epoch": 26.39944013781223,
      "grad_norm": 0.5013649463653564,
      "learning_rate": 0.00020415107041874242,
      "loss": 2.9384,
      "step": 98100
    },
    {
      "epoch": 26.41289836347976,
      "grad_norm": 0.45403429865837097,
      "learning_rate": 0.00020394910461828464,
      "loss": 2.9344,
      "step": 98150
    },
    {
      "epoch": 26.426356589147286,
      "grad_norm": 0.4686357080936432,
      "learning_rate": 0.00020374713881782683,
      "loss": 2.9323,
      "step": 98200
    },
    {
      "epoch": 26.439814814814813,
      "grad_norm": 0.46750739216804504,
      "learning_rate": 0.00020354517301736906,
      "loss": 2.9383,
      "step": 98250
    },
    {
      "epoch": 26.453273040482344,
      "grad_norm": 0.4663945734500885,
      "learning_rate": 0.00020334320721691125,
      "loss": 2.9419,
      "step": 98300
    },
    {
      "epoch": 26.46673126614987,
      "grad_norm": 0.46350347995758057,
      "learning_rate": 0.00020314124141645345,
      "loss": 2.9406,
      "step": 98350
    },
    {
      "epoch": 26.4801894918174,
      "grad_norm": 0.5216088891029358,
      "learning_rate": 0.00020293927561599567,
      "loss": 2.947,
      "step": 98400
    },
    {
      "epoch": 26.493647717484926,
      "grad_norm": 0.47216132283210754,
      "learning_rate": 0.00020273730981553787,
      "loss": 2.9383,
      "step": 98450
    },
    {
      "epoch": 26.507105943152453,
      "grad_norm": 0.44430720806121826,
      "learning_rate": 0.0002025353440150801,
      "loss": 2.938,
      "step": 98500
    },
    {
      "epoch": 26.520564168819984,
      "grad_norm": 0.4635902941226959,
      "learning_rate": 0.00020233337821462232,
      "loss": 2.9437,
      "step": 98550
    },
    {
      "epoch": 26.53402239448751,
      "grad_norm": 0.5018345713615417,
      "learning_rate": 0.00020213141241416454,
      "loss": 2.9515,
      "step": 98600
    },
    {
      "epoch": 26.54748062015504,
      "grad_norm": 0.4643080532550812,
      "learning_rate": 0.00020192944661370674,
      "loss": 2.9381,
      "step": 98650
    },
    {
      "epoch": 26.560938845822566,
      "grad_norm": 0.45210981369018555,
      "learning_rate": 0.00020172748081324893,
      "loss": 2.9402,
      "step": 98700
    },
    {
      "epoch": 26.574397071490093,
      "grad_norm": 0.4725451171398163,
      "learning_rate": 0.00020152551501279115,
      "loss": 2.937,
      "step": 98750
    },
    {
      "epoch": 26.587855297157624,
      "grad_norm": 0.47395259141921997,
      "learning_rate": 0.00020132354921233335,
      "loss": 2.9421,
      "step": 98800
    },
    {
      "epoch": 26.60131352282515,
      "grad_norm": 0.4714031517505646,
      "learning_rate": 0.00020112158341187557,
      "loss": 2.9508,
      "step": 98850
    },
    {
      "epoch": 26.61477174849268,
      "grad_norm": 0.49274519085884094,
      "learning_rate": 0.00020091961761141777,
      "loss": 2.949,
      "step": 98900
    },
    {
      "epoch": 26.628229974160206,
      "grad_norm": 0.47620511054992676,
      "learning_rate": 0.00020071765181095997,
      "loss": 2.946,
      "step": 98950
    },
    {
      "epoch": 26.641688199827733,
      "grad_norm": 0.47118687629699707,
      "learning_rate": 0.00020051568601050222,
      "loss": 2.955,
      "step": 99000
    },
    {
      "epoch": 26.641688199827733,
      "eval_accuracy": 0.3971610793934724,
      "eval_loss": 3.2821288108825684,
      "eval_runtime": 83.9214,
      "eval_samples_per_second": 214.629,
      "eval_steps_per_second": 13.417,
      "step": 99000
    },
    {
      "epoch": 26.655146425495264,
      "grad_norm": 0.4601459205150604,
      "learning_rate": 0.00020031372021004444,
      "loss": 2.942,
      "step": 99050
    },
    {
      "epoch": 26.66860465116279,
      "grad_norm": 0.48537662625312805,
      "learning_rate": 0.00020011175440958664,
      "loss": 2.9549,
      "step": 99100
    },
    {
      "epoch": 26.68206287683032,
      "grad_norm": 0.47839945554733276,
      "learning_rate": 0.00019990978860912883,
      "loss": 2.9483,
      "step": 99150
    },
    {
      "epoch": 26.695521102497846,
      "grad_norm": 0.5007160305976868,
      "learning_rate": 0.00019970782280867105,
      "loss": 2.9443,
      "step": 99200
    },
    {
      "epoch": 26.708979328165373,
      "grad_norm": 0.45769360661506653,
      "learning_rate": 0.00019950585700821325,
      "loss": 2.9489,
      "step": 99250
    },
    {
      "epoch": 26.722437553832904,
      "grad_norm": 0.4725143313407898,
      "learning_rate": 0.00019930389120775547,
      "loss": 2.953,
      "step": 99300
    },
    {
      "epoch": 26.73589577950043,
      "grad_norm": 0.4660271108150482,
      "learning_rate": 0.00019910192540729767,
      "loss": 2.9595,
      "step": 99350
    },
    {
      "epoch": 26.74935400516796,
      "grad_norm": 0.5152368545532227,
      "learning_rate": 0.00019889995960683987,
      "loss": 2.9463,
      "step": 99400
    },
    {
      "epoch": 26.762812230835486,
      "grad_norm": 0.442242830991745,
      "learning_rate": 0.00019869799380638212,
      "loss": 2.96,
      "step": 99450
    },
    {
      "epoch": 26.776270456503013,
      "grad_norm": 0.49033448100090027,
      "learning_rate": 0.0001984960280059243,
      "loss": 2.9534,
      "step": 99500
    },
    {
      "epoch": 26.789728682170544,
      "grad_norm": 0.4927358627319336,
      "learning_rate": 0.00019829406220546654,
      "loss": 2.9348,
      "step": 99550
    },
    {
      "epoch": 26.80318690783807,
      "grad_norm": 0.476701557636261,
      "learning_rate": 0.00019809209640500873,
      "loss": 2.9437,
      "step": 99600
    },
    {
      "epoch": 26.8166451335056,
      "grad_norm": 0.5028373003005981,
      "learning_rate": 0.00019789013060455096,
      "loss": 2.9598,
      "step": 99650
    },
    {
      "epoch": 26.830103359173126,
      "grad_norm": 0.4803355634212494,
      "learning_rate": 0.00019768816480409315,
      "loss": 2.9621,
      "step": 99700
    },
    {
      "epoch": 26.843561584840653,
      "grad_norm": 0.4716373682022095,
      "learning_rate": 0.00019748619900363535,
      "loss": 2.9636,
      "step": 99750
    },
    {
      "epoch": 26.857019810508184,
      "grad_norm": 0.4548377990722656,
      "learning_rate": 0.00019728423320317757,
      "loss": 2.9674,
      "step": 99800
    },
    {
      "epoch": 26.87047803617571,
      "grad_norm": 0.4843604564666748,
      "learning_rate": 0.00019708226740271977,
      "loss": 2.9526,
      "step": 99850
    },
    {
      "epoch": 26.88393626184324,
      "grad_norm": 0.4458632469177246,
      "learning_rate": 0.00019688030160226202,
      "loss": 2.9544,
      "step": 99900
    },
    {
      "epoch": 26.897394487510766,
      "grad_norm": 0.4735550582408905,
      "learning_rate": 0.0001966783358018042,
      "loss": 2.9628,
      "step": 99950
    },
    {
      "epoch": 26.910852713178294,
      "grad_norm": 0.4771996736526489,
      "learning_rate": 0.00019647637000134644,
      "loss": 2.9637,
      "step": 100000
    },
    {
      "epoch": 26.910852713178294,
      "eval_accuracy": 0.3976530323554682,
      "eval_loss": 3.2714831829071045,
      "eval_runtime": 83.8651,
      "eval_samples_per_second": 214.773,
      "eval_steps_per_second": 13.426,
      "step": 100000
    },
    {
      "epoch": 26.92431093884582,
      "grad_norm": 0.5054681897163391,
      "learning_rate": 0.00019627440420088863,
      "loss": 2.9601,
      "step": 100050
    },
    {
      "epoch": 26.93776916451335,
      "grad_norm": 0.4774108827114105,
      "learning_rate": 0.00019607243840043086,
      "loss": 2.9651,
      "step": 100100
    },
    {
      "epoch": 26.95122739018088,
      "grad_norm": 0.4822835624217987,
      "learning_rate": 0.00019587047259997305,
      "loss": 2.9602,
      "step": 100150
    },
    {
      "epoch": 26.964685615848406,
      "grad_norm": 0.47141721844673157,
      "learning_rate": 0.00019566850679951525,
      "loss": 2.9669,
      "step": 100200
    },
    {
      "epoch": 26.978143841515934,
      "grad_norm": 0.49402523040771484,
      "learning_rate": 0.00019546654099905747,
      "loss": 2.9719,
      "step": 100250
    },
    {
      "epoch": 26.99160206718346,
      "grad_norm": 0.4692891836166382,
      "learning_rate": 0.00019526457519859967,
      "loss": 2.9655,
      "step": 100300
    },
    {
      "epoch": 27.00484496124031,
      "grad_norm": 0.4445691704750061,
      "learning_rate": 0.00019506260939814192,
      "loss": 2.938,
      "step": 100350
    },
    {
      "epoch": 27.01830318690784,
      "grad_norm": 0.4529949426651001,
      "learning_rate": 0.0001948606435976841,
      "loss": 2.8878,
      "step": 100400
    },
    {
      "epoch": 27.031761412575367,
      "grad_norm": 0.4826950132846832,
      "learning_rate": 0.00019465867779722634,
      "loss": 2.8936,
      "step": 100450
    },
    {
      "epoch": 27.045219638242894,
      "grad_norm": 0.4961959719657898,
      "learning_rate": 0.00019445671199676853,
      "loss": 2.887,
      "step": 100500
    },
    {
      "epoch": 27.05867786391042,
      "grad_norm": 0.47039660811424255,
      "learning_rate": 0.00019425474619631073,
      "loss": 2.8874,
      "step": 100550
    },
    {
      "epoch": 27.07213608957795,
      "grad_norm": 0.49506014585494995,
      "learning_rate": 0.00019405278039585295,
      "loss": 2.9017,
      "step": 100600
    },
    {
      "epoch": 27.085594315245476,
      "grad_norm": 0.4895104169845581,
      "learning_rate": 0.00019385081459539515,
      "loss": 2.9064,
      "step": 100650
    },
    {
      "epoch": 27.099052540913007,
      "grad_norm": 0.48083260655403137,
      "learning_rate": 0.00019364884879493737,
      "loss": 2.8897,
      "step": 100700
    },
    {
      "epoch": 27.112510766580534,
      "grad_norm": 0.4734748601913452,
      "learning_rate": 0.00019344688299447957,
      "loss": 2.8981,
      "step": 100750
    },
    {
      "epoch": 27.125968992248062,
      "grad_norm": 0.49400684237480164,
      "learning_rate": 0.00019324491719402182,
      "loss": 2.9027,
      "step": 100800
    },
    {
      "epoch": 27.13942721791559,
      "grad_norm": 0.49420422315597534,
      "learning_rate": 0.00019304295139356401,
      "loss": 2.8874,
      "step": 100850
    },
    {
      "epoch": 27.152885443583116,
      "grad_norm": 0.4779294729232788,
      "learning_rate": 0.0001928409855931062,
      "loss": 2.9169,
      "step": 100900
    },
    {
      "epoch": 27.166343669250647,
      "grad_norm": 0.4682933986186981,
      "learning_rate": 0.00019263901979264843,
      "loss": 2.8994,
      "step": 100950
    },
    {
      "epoch": 27.179801894918175,
      "grad_norm": 0.4925904870033264,
      "learning_rate": 0.00019243705399219063,
      "loss": 2.9047,
      "step": 101000
    },
    {
      "epoch": 27.179801894918175,
      "eval_accuracy": 0.39650202756761854,
      "eval_loss": 3.291107654571533,
      "eval_runtime": 83.7959,
      "eval_samples_per_second": 214.951,
      "eval_steps_per_second": 13.437,
      "step": 101000
    },
    {
      "epoch": 27.193260120585702,
      "grad_norm": 0.4778890013694763,
      "learning_rate": 0.00019223508819173285,
      "loss": 2.9085,
      "step": 101050
    },
    {
      "epoch": 27.20671834625323,
      "grad_norm": 0.4523789584636688,
      "learning_rate": 0.00019203312239127505,
      "loss": 2.9155,
      "step": 101100
    },
    {
      "epoch": 27.220176571920756,
      "grad_norm": 0.4899305999279022,
      "learning_rate": 0.00019183115659081727,
      "loss": 2.9123,
      "step": 101150
    },
    {
      "epoch": 27.233634797588287,
      "grad_norm": 0.47693610191345215,
      "learning_rate": 0.00019162919079035947,
      "loss": 2.9102,
      "step": 101200
    },
    {
      "epoch": 27.247093023255815,
      "grad_norm": 0.4802713394165039,
      "learning_rate": 0.00019142722498990172,
      "loss": 2.9122,
      "step": 101250
    },
    {
      "epoch": 27.260551248923342,
      "grad_norm": 0.4723227918148041,
      "learning_rate": 0.00019122525918944391,
      "loss": 2.9112,
      "step": 101300
    },
    {
      "epoch": 27.27400947459087,
      "grad_norm": 0.47506293654441833,
      "learning_rate": 0.0001910232933889861,
      "loss": 2.9032,
      "step": 101350
    },
    {
      "epoch": 27.287467700258397,
      "grad_norm": 0.4856734573841095,
      "learning_rate": 0.00019082132758852833,
      "loss": 2.9086,
      "step": 101400
    },
    {
      "epoch": 27.300925925925927,
      "grad_norm": 0.48981887102127075,
      "learning_rate": 0.00019061936178807053,
      "loss": 2.9159,
      "step": 101450
    },
    {
      "epoch": 27.314384151593455,
      "grad_norm": 0.48034849762916565,
      "learning_rate": 0.00019041739598761275,
      "loss": 2.9197,
      "step": 101500
    },
    {
      "epoch": 27.327842377260982,
      "grad_norm": 0.4707762897014618,
      "learning_rate": 0.00019021543018715495,
      "loss": 2.9102,
      "step": 101550
    },
    {
      "epoch": 27.34130060292851,
      "grad_norm": 0.4944721758365631,
      "learning_rate": 0.00019001346438669714,
      "loss": 2.9205,
      "step": 101600
    },
    {
      "epoch": 27.354758828596037,
      "grad_norm": 0.45902949571609497,
      "learning_rate": 0.0001898114985862394,
      "loss": 2.9231,
      "step": 101650
    },
    {
      "epoch": 27.368217054263567,
      "grad_norm": 0.47691816091537476,
      "learning_rate": 0.0001896095327857816,
      "loss": 2.9278,
      "step": 101700
    },
    {
      "epoch": 27.381675279931095,
      "grad_norm": 0.5126941204071045,
      "learning_rate": 0.00018940756698532381,
      "loss": 2.9298,
      "step": 101750
    },
    {
      "epoch": 27.395133505598622,
      "grad_norm": 0.4797493517398834,
      "learning_rate": 0.000189205601184866,
      "loss": 2.9261,
      "step": 101800
    },
    {
      "epoch": 27.40859173126615,
      "grad_norm": 0.47926071286201477,
      "learning_rate": 0.00018900363538440823,
      "loss": 2.9394,
      "step": 101850
    },
    {
      "epoch": 27.422049956933677,
      "grad_norm": 0.4621308743953705,
      "learning_rate": 0.00018880166958395043,
      "loss": 2.9311,
      "step": 101900
    },
    {
      "epoch": 27.435508182601207,
      "grad_norm": 0.497580349445343,
      "learning_rate": 0.00018859970378349263,
      "loss": 2.9275,
      "step": 101950
    },
    {
      "epoch": 27.448966408268735,
      "grad_norm": 0.47857317328453064,
      "learning_rate": 0.00018839773798303485,
      "loss": 2.9261,
      "step": 102000
    },
    {
      "epoch": 27.448966408268735,
      "eval_accuracy": 0.39700788732712655,
      "eval_loss": 3.2868685722351074,
      "eval_runtime": 83.3893,
      "eval_samples_per_second": 215.999,
      "eval_steps_per_second": 13.503,
      "step": 102000
    },
    {
      "epoch": 27.462424633936262,
      "grad_norm": 0.49953895807266235,
      "learning_rate": 0.00018819577218257705,
      "loss": 2.94,
      "step": 102050
    },
    {
      "epoch": 27.47588285960379,
      "grad_norm": 0.5003396272659302,
      "learning_rate": 0.0001879938063821193,
      "loss": 2.9426,
      "step": 102100
    },
    {
      "epoch": 27.489341085271317,
      "grad_norm": 0.49409356713294983,
      "learning_rate": 0.0001877918405816615,
      "loss": 2.9304,
      "step": 102150
    },
    {
      "epoch": 27.502799310938848,
      "grad_norm": 0.48523804545402527,
      "learning_rate": 0.00018758987478120371,
      "loss": 2.9272,
      "step": 102200
    },
    {
      "epoch": 27.516257536606375,
      "grad_norm": 0.5118574500083923,
      "learning_rate": 0.0001873879089807459,
      "loss": 2.9363,
      "step": 102250
    },
    {
      "epoch": 27.529715762273902,
      "grad_norm": 0.4614475667476654,
      "learning_rate": 0.00018718594318028813,
      "loss": 2.9246,
      "step": 102300
    },
    {
      "epoch": 27.54317398794143,
      "grad_norm": 0.5088652968406677,
      "learning_rate": 0.00018698397737983033,
      "loss": 2.928,
      "step": 102350
    },
    {
      "epoch": 27.556632213608957,
      "grad_norm": 0.4668629467487335,
      "learning_rate": 0.00018678201157937253,
      "loss": 2.9364,
      "step": 102400
    },
    {
      "epoch": 27.570090439276484,
      "grad_norm": 0.45806413888931274,
      "learning_rate": 0.00018658004577891475,
      "loss": 2.9297,
      "step": 102450
    },
    {
      "epoch": 27.583548664944015,
      "grad_norm": 0.47820380330085754,
      "learning_rate": 0.00018637807997845695,
      "loss": 2.9393,
      "step": 102500
    },
    {
      "epoch": 27.597006890611542,
      "grad_norm": 0.48876985907554626,
      "learning_rate": 0.0001861761141779992,
      "loss": 2.9367,
      "step": 102550
    },
    {
      "epoch": 27.61046511627907,
      "grad_norm": 0.4883100092411041,
      "learning_rate": 0.0001859741483775414,
      "loss": 2.9347,
      "step": 102600
    },
    {
      "epoch": 27.623923341946597,
      "grad_norm": 0.4798702895641327,
      "learning_rate": 0.00018577218257708362,
      "loss": 2.9481,
      "step": 102650
    },
    {
      "epoch": 27.637381567614124,
      "grad_norm": 0.5159857273101807,
      "learning_rate": 0.0001855702167766258,
      "loss": 2.9428,
      "step": 102700
    },
    {
      "epoch": 27.650839793281655,
      "grad_norm": 0.47942033410072327,
      "learning_rate": 0.000185368250976168,
      "loss": 2.94,
      "step": 102750
    },
    {
      "epoch": 27.664298018949182,
      "grad_norm": 0.4900752305984497,
      "learning_rate": 0.00018516628517571023,
      "loss": 2.9366,
      "step": 102800
    },
    {
      "epoch": 27.67775624461671,
      "grad_norm": 0.48997023701667786,
      "learning_rate": 0.00018496431937525243,
      "loss": 2.9462,
      "step": 102850
    },
    {
      "epoch": 27.691214470284237,
      "grad_norm": 0.4761708080768585,
      "learning_rate": 0.00018476235357479465,
      "loss": 2.9438,
      "step": 102900
    },
    {
      "epoch": 27.704672695951764,
      "grad_norm": 0.4778705835342407,
      "learning_rate": 0.00018456038777433685,
      "loss": 2.9411,
      "step": 102950
    },
    {
      "epoch": 27.718130921619295,
      "grad_norm": 0.47539815306663513,
      "learning_rate": 0.0001843584219738791,
      "loss": 2.9289,
      "step": 103000
    },
    {
      "epoch": 27.718130921619295,
      "eval_accuracy": 0.39738760808732426,
      "eval_loss": 3.2796902656555176,
      "eval_runtime": 83.5612,
      "eval_samples_per_second": 215.555,
      "eval_steps_per_second": 13.475,
      "step": 103000
    },
    {
      "epoch": 27.731589147286822,
      "grad_norm": 0.4928041696548462,
      "learning_rate": 0.0001841564561734213,
      "loss": 2.9338,
      "step": 103050
    },
    {
      "epoch": 27.74504737295435,
      "grad_norm": 0.45715710520744324,
      "learning_rate": 0.00018395449037296352,
      "loss": 2.9516,
      "step": 103100
    },
    {
      "epoch": 27.758505598621877,
      "grad_norm": 0.513388991355896,
      "learning_rate": 0.0001837525245725057,
      "loss": 2.9555,
      "step": 103150
    },
    {
      "epoch": 27.771963824289404,
      "grad_norm": 0.4790501296520233,
      "learning_rate": 0.0001835505587720479,
      "loss": 2.9415,
      "step": 103200
    },
    {
      "epoch": 27.785422049956935,
      "grad_norm": 0.4711688160896301,
      "learning_rate": 0.00018334859297159013,
      "loss": 2.9406,
      "step": 103250
    },
    {
      "epoch": 27.798880275624462,
      "grad_norm": 0.4755244255065918,
      "learning_rate": 0.00018314662717113233,
      "loss": 2.9452,
      "step": 103300
    },
    {
      "epoch": 27.81233850129199,
      "grad_norm": 0.48446419835090637,
      "learning_rate": 0.00018294466137067455,
      "loss": 2.9535,
      "step": 103350
    },
    {
      "epoch": 27.825796726959517,
      "grad_norm": 0.5106527805328369,
      "learning_rate": 0.00018274269557021675,
      "loss": 2.9491,
      "step": 103400
    },
    {
      "epoch": 27.839254952627044,
      "grad_norm": 0.4658513069152832,
      "learning_rate": 0.000182540729769759,
      "loss": 2.9409,
      "step": 103450
    },
    {
      "epoch": 27.852713178294575,
      "grad_norm": 0.4776923358440399,
      "learning_rate": 0.0001823387639693012,
      "loss": 2.9458,
      "step": 103500
    },
    {
      "epoch": 27.866171403962102,
      "grad_norm": 0.4554797410964966,
      "learning_rate": 0.0001821367981688434,
      "loss": 2.9528,
      "step": 103550
    },
    {
      "epoch": 27.87962962962963,
      "grad_norm": 0.4616670310497284,
      "learning_rate": 0.0001819348323683856,
      "loss": 2.9565,
      "step": 103600
    },
    {
      "epoch": 27.893087855297157,
      "grad_norm": 0.4706868827342987,
      "learning_rate": 0.0001817328665679278,
      "loss": 2.9533,
      "step": 103650
    },
    {
      "epoch": 27.906546080964684,
      "grad_norm": 0.47426125407218933,
      "learning_rate": 0.00018153090076747003,
      "loss": 2.9484,
      "step": 103700
    },
    {
      "epoch": 27.920004306632215,
      "grad_norm": 0.5088492631912231,
      "learning_rate": 0.00018132893496701223,
      "loss": 2.9629,
      "step": 103750
    },
    {
      "epoch": 27.933462532299743,
      "grad_norm": 0.4598265290260315,
      "learning_rate": 0.00018112696916655442,
      "loss": 2.9575,
      "step": 103800
    },
    {
      "epoch": 27.94692075796727,
      "grad_norm": 0.47561395168304443,
      "learning_rate": 0.00018092500336609665,
      "loss": 2.9553,
      "step": 103850
    },
    {
      "epoch": 27.960378983634797,
      "grad_norm": 0.4709970951080322,
      "learning_rate": 0.0001807230375656389,
      "loss": 2.9453,
      "step": 103900
    },
    {
      "epoch": 27.973837209302324,
      "grad_norm": 0.50580233335495,
      "learning_rate": 0.0001805210717651811,
      "loss": 2.9417,
      "step": 103950
    },
    {
      "epoch": 27.987295434969855,
      "grad_norm": 0.48298802971839905,
      "learning_rate": 0.0001803191059647233,
      "loss": 2.9416,
      "step": 104000
    },
    {
      "epoch": 27.987295434969855,
      "eval_accuracy": 0.3978392530659056,
      "eval_loss": 3.272890090942383,
      "eval_runtime": 83.4118,
      "eval_samples_per_second": 215.941,
      "eval_steps_per_second": 13.499,
      "step": 104000
    },
    {
      "epoch": 28.0005383290267,
      "grad_norm": 0.5029809474945068,
      "learning_rate": 0.0001801171401642655,
      "loss": 2.9482,
      "step": 104050
    },
    {
      "epoch": 28.01399655469423,
      "grad_norm": 0.4830522835254669,
      "learning_rate": 0.0001799151743638077,
      "loss": 2.888,
      "step": 104100
    },
    {
      "epoch": 28.027454780361758,
      "grad_norm": 0.46248242259025574,
      "learning_rate": 0.00017971320856334993,
      "loss": 2.8743,
      "step": 104150
    },
    {
      "epoch": 28.040913006029285,
      "grad_norm": 0.4835436940193176,
      "learning_rate": 0.00017951124276289213,
      "loss": 2.8806,
      "step": 104200
    },
    {
      "epoch": 28.054371231696813,
      "grad_norm": 0.46218568086624146,
      "learning_rate": 0.00017930927696243432,
      "loss": 2.8808,
      "step": 104250
    },
    {
      "epoch": 28.06782945736434,
      "grad_norm": 0.46646058559417725,
      "learning_rate": 0.00017910731116197655,
      "loss": 2.8763,
      "step": 104300
    },
    {
      "epoch": 28.08128768303187,
      "grad_norm": 0.5135084390640259,
      "learning_rate": 0.00017890534536151877,
      "loss": 2.8779,
      "step": 104350
    },
    {
      "epoch": 28.094745908699398,
      "grad_norm": 0.5373383164405823,
      "learning_rate": 0.000178703379561061,
      "loss": 2.9005,
      "step": 104400
    },
    {
      "epoch": 28.108204134366925,
      "grad_norm": 0.4764528274536133,
      "learning_rate": 0.0001785014137606032,
      "loss": 2.9056,
      "step": 104450
    },
    {
      "epoch": 28.121662360034453,
      "grad_norm": 0.5029706954956055,
      "learning_rate": 0.0001782994479601454,
      "loss": 2.8854,
      "step": 104500
    },
    {
      "epoch": 28.13512058570198,
      "grad_norm": 0.4830996096134186,
      "learning_rate": 0.0001780974821596876,
      "loss": 2.8918,
      "step": 104550
    },
    {
      "epoch": 28.14857881136951,
      "grad_norm": 0.4923493266105652,
      "learning_rate": 0.0001778955163592298,
      "loss": 2.8969,
      "step": 104600
    },
    {
      "epoch": 28.162037037037038,
      "grad_norm": 0.5185808539390564,
      "learning_rate": 0.00017769355055877203,
      "loss": 2.8965,
      "step": 104650
    },
    {
      "epoch": 28.175495262704565,
      "grad_norm": 0.49304184317588806,
      "learning_rate": 0.00017749158475831422,
      "loss": 2.8943,
      "step": 104700
    },
    {
      "epoch": 28.188953488372093,
      "grad_norm": 0.48696431517601013,
      "learning_rate": 0.00017728961895785645,
      "loss": 2.8983,
      "step": 104750
    },
    {
      "epoch": 28.20241171403962,
      "grad_norm": 0.47151580452919006,
      "learning_rate": 0.00017708765315739867,
      "loss": 2.9009,
      "step": 104800
    },
    {
      "epoch": 28.215869939707147,
      "grad_norm": 0.512098491191864,
      "learning_rate": 0.0001768856873569409,
      "loss": 2.9057,
      "step": 104850
    },
    {
      "epoch": 28.229328165374678,
      "grad_norm": 0.48440688848495483,
      "learning_rate": 0.0001766837215564831,
      "loss": 2.9098,
      "step": 104900
    },
    {
      "epoch": 28.242786391042205,
      "grad_norm": 0.4709826409816742,
      "learning_rate": 0.0001764817557560253,
      "loss": 2.9088,
      "step": 104950
    },
    {
      "epoch": 28.256244616709733,
      "grad_norm": 0.4686014950275421,
      "learning_rate": 0.0001762797899555675,
      "loss": 2.9046,
      "step": 105000
    },
    {
      "epoch": 28.256244616709733,
      "eval_accuracy": 0.39703537498158437,
      "eval_loss": 3.2879161834716797,
      "eval_runtime": 84.0151,
      "eval_samples_per_second": 214.39,
      "eval_steps_per_second": 13.402,
      "step": 105000
    },
    {
      "epoch": 28.26970284237726,
      "grad_norm": 0.47056883573532104,
      "learning_rate": 0.0001760778241551097,
      "loss": 2.9219,
      "step": 105050
    },
    {
      "epoch": 28.283161068044787,
      "grad_norm": 0.4826996326446533,
      "learning_rate": 0.00017587585835465193,
      "loss": 2.9077,
      "step": 105100
    },
    {
      "epoch": 28.296619293712318,
      "grad_norm": 0.47803038358688354,
      "learning_rate": 0.00017567389255419412,
      "loss": 2.9056,
      "step": 105150
    },
    {
      "epoch": 28.310077519379846,
      "grad_norm": 0.4889366030693054,
      "learning_rate": 0.00017547192675373635,
      "loss": 2.9104,
      "step": 105200
    },
    {
      "epoch": 28.323535745047373,
      "grad_norm": 0.4798263907432556,
      "learning_rate": 0.00017526996095327857,
      "loss": 2.9023,
      "step": 105250
    },
    {
      "epoch": 28.3369939707149,
      "grad_norm": 0.47523853182792664,
      "learning_rate": 0.0001750679951528208,
      "loss": 2.9126,
      "step": 105300
    },
    {
      "epoch": 28.350452196382427,
      "grad_norm": 0.47547391057014465,
      "learning_rate": 0.000174866029352363,
      "loss": 2.9019,
      "step": 105350
    },
    {
      "epoch": 28.36391042204996,
      "grad_norm": 0.507918655872345,
      "learning_rate": 0.0001746640635519052,
      "loss": 2.9088,
      "step": 105400
    },
    {
      "epoch": 28.377368647717486,
      "grad_norm": 0.46947479248046875,
      "learning_rate": 0.0001744620977514474,
      "loss": 2.9047,
      "step": 105450
    },
    {
      "epoch": 28.390826873385013,
      "grad_norm": 0.50822514295578,
      "learning_rate": 0.0001742601319509896,
      "loss": 2.9046,
      "step": 105500
    },
    {
      "epoch": 28.40428509905254,
      "grad_norm": 0.48622313141822815,
      "learning_rate": 0.00017405816615053183,
      "loss": 2.9211,
      "step": 105550
    },
    {
      "epoch": 28.417743324720067,
      "grad_norm": 0.49788394570350647,
      "learning_rate": 0.00017385620035007403,
      "loss": 2.9059,
      "step": 105600
    },
    {
      "epoch": 28.4312015503876,
      "grad_norm": 0.5572716593742371,
      "learning_rate": 0.00017365423454961622,
      "loss": 2.9158,
      "step": 105650
    },
    {
      "epoch": 28.444659776055126,
      "grad_norm": 0.5081602931022644,
      "learning_rate": 0.00017345226874915847,
      "loss": 2.9197,
      "step": 105700
    },
    {
      "epoch": 28.458118001722653,
      "grad_norm": 0.5340498685836792,
      "learning_rate": 0.00017325030294870067,
      "loss": 2.9138,
      "step": 105750
    },
    {
      "epoch": 28.47157622739018,
      "grad_norm": 0.4911704957485199,
      "learning_rate": 0.0001730483371482429,
      "loss": 2.9194,
      "step": 105800
    },
    {
      "epoch": 28.485034453057708,
      "grad_norm": 0.5058697462081909,
      "learning_rate": 0.0001728463713477851,
      "loss": 2.926,
      "step": 105850
    },
    {
      "epoch": 28.49849267872524,
      "grad_norm": 0.5085639357566833,
      "learning_rate": 0.0001726444055473273,
      "loss": 2.9267,
      "step": 105900
    },
    {
      "epoch": 28.511950904392766,
      "grad_norm": 0.5275683403015137,
      "learning_rate": 0.0001724424397468695,
      "loss": 2.9251,
      "step": 105950
    },
    {
      "epoch": 28.525409130060293,
      "grad_norm": 0.5193195343017578,
      "learning_rate": 0.00017224047394641173,
      "loss": 2.9343,
      "step": 106000
    },
    {
      "epoch": 28.525409130060293,
      "eval_accuracy": 0.3973411072331427,
      "eval_loss": 3.2808279991149902,
      "eval_runtime": 83.9538,
      "eval_samples_per_second": 214.547,
      "eval_steps_per_second": 13.412,
      "step": 106000
    },
    {
      "epoch": 28.53886735572782,
      "grad_norm": 0.49010545015335083,
      "learning_rate": 0.00017203850814595393,
      "loss": 2.9254,
      "step": 106050
    },
    {
      "epoch": 28.552325581395348,
      "grad_norm": 0.4868125021457672,
      "learning_rate": 0.00017183654234549612,
      "loss": 2.9375,
      "step": 106100
    },
    {
      "epoch": 28.56578380706288,
      "grad_norm": 0.46857622265815735,
      "learning_rate": 0.00017163457654503837,
      "loss": 2.9227,
      "step": 106150
    },
    {
      "epoch": 28.579242032730406,
      "grad_norm": 0.5180408358573914,
      "learning_rate": 0.00017143261074458057,
      "loss": 2.9294,
      "step": 106200
    },
    {
      "epoch": 28.592700258397933,
      "grad_norm": 0.47722306847572327,
      "learning_rate": 0.0001712306449441228,
      "loss": 2.9235,
      "step": 106250
    },
    {
      "epoch": 28.60615848406546,
      "grad_norm": 0.48222780227661133,
      "learning_rate": 0.000171028679143665,
      "loss": 2.9311,
      "step": 106300
    },
    {
      "epoch": 28.619616709732988,
      "grad_norm": 0.4867691397666931,
      "learning_rate": 0.0001708267133432072,
      "loss": 2.9299,
      "step": 106350
    },
    {
      "epoch": 28.63307493540052,
      "grad_norm": 0.47849902510643005,
      "learning_rate": 0.0001706247475427494,
      "loss": 2.9198,
      "step": 106400
    },
    {
      "epoch": 28.646533161068046,
      "grad_norm": 0.5070969462394714,
      "learning_rate": 0.0001704227817422916,
      "loss": 2.9292,
      "step": 106450
    },
    {
      "epoch": 28.659991386735573,
      "grad_norm": 0.4817515015602112,
      "learning_rate": 0.00017022081594183383,
      "loss": 2.9275,
      "step": 106500
    },
    {
      "epoch": 28.6734496124031,
      "grad_norm": 0.5026460886001587,
      "learning_rate": 0.00017001885014137605,
      "loss": 2.9389,
      "step": 106550
    },
    {
      "epoch": 28.686907838070628,
      "grad_norm": 0.4838036000728607,
      "learning_rate": 0.00016981688434091827,
      "loss": 2.9249,
      "step": 106600
    },
    {
      "epoch": 28.70036606373816,
      "grad_norm": 0.5039952397346497,
      "learning_rate": 0.00016961491854046047,
      "loss": 2.9318,
      "step": 106650
    },
    {
      "epoch": 28.713824289405686,
      "grad_norm": 0.5017082691192627,
      "learning_rate": 0.0001694129527400027,
      "loss": 2.9341,
      "step": 106700
    },
    {
      "epoch": 28.727282515073213,
      "grad_norm": 0.4967496693134308,
      "learning_rate": 0.0001692109869395449,
      "loss": 2.9396,
      "step": 106750
    },
    {
      "epoch": 28.74074074074074,
      "grad_norm": 0.5065209269523621,
      "learning_rate": 0.00016900902113908708,
      "loss": 2.9364,
      "step": 106800
    },
    {
      "epoch": 28.754198966408268,
      "grad_norm": 0.4952838718891144,
      "learning_rate": 0.0001688070553386293,
      "loss": 2.932,
      "step": 106850
    },
    {
      "epoch": 28.767657192075795,
      "grad_norm": 0.46681851148605347,
      "learning_rate": 0.0001686050895381715,
      "loss": 2.9359,
      "step": 106900
    },
    {
      "epoch": 28.781115417743326,
      "grad_norm": 0.5006422996520996,
      "learning_rate": 0.00016840312373771373,
      "loss": 2.94,
      "step": 106950
    },
    {
      "epoch": 28.794573643410853,
      "grad_norm": 0.48693737387657166,
      "learning_rate": 0.00016820115793725595,
      "loss": 2.9354,
      "step": 107000
    },
    {
      "epoch": 28.794573643410853,
      "eval_accuracy": 0.3976146800154539,
      "eval_loss": 3.279836893081665,
      "eval_runtime": 83.847,
      "eval_samples_per_second": 214.82,
      "eval_steps_per_second": 13.429,
      "step": 107000
    },
    {
      "epoch": 28.80803186907838,
      "grad_norm": 0.4770016074180603,
      "learning_rate": 0.00016799919213679817,
      "loss": 2.9343,
      "step": 107050
    },
    {
      "epoch": 28.821490094745908,
      "grad_norm": 0.5120300650596619,
      "learning_rate": 0.00016779722633634037,
      "loss": 2.9285,
      "step": 107100
    },
    {
      "epoch": 28.834948320413435,
      "grad_norm": 0.5253459215164185,
      "learning_rate": 0.0001675952605358826,
      "loss": 2.9444,
      "step": 107150
    },
    {
      "epoch": 28.848406546080966,
      "grad_norm": 0.5212303996086121,
      "learning_rate": 0.0001673932947354248,
      "loss": 2.9396,
      "step": 107200
    },
    {
      "epoch": 28.861864771748493,
      "grad_norm": 0.46815282106399536,
      "learning_rate": 0.00016719132893496698,
      "loss": 2.9358,
      "step": 107250
    },
    {
      "epoch": 28.87532299741602,
      "grad_norm": 0.4672527313232422,
      "learning_rate": 0.0001669893631345092,
      "loss": 2.9428,
      "step": 107300
    },
    {
      "epoch": 28.888781223083548,
      "grad_norm": 0.5072224736213684,
      "learning_rate": 0.0001667873973340514,
      "loss": 2.929,
      "step": 107350
    },
    {
      "epoch": 28.902239448751075,
      "grad_norm": 0.47793617844581604,
      "learning_rate": 0.00016658543153359363,
      "loss": 2.9454,
      "step": 107400
    },
    {
      "epoch": 28.915697674418606,
      "grad_norm": 0.4852517545223236,
      "learning_rate": 0.00016638346573313585,
      "loss": 2.9434,
      "step": 107450
    },
    {
      "epoch": 28.929155900086133,
      "grad_norm": 0.5230569243431091,
      "learning_rate": 0.00016618149993267807,
      "loss": 2.936,
      "step": 107500
    },
    {
      "epoch": 28.94261412575366,
      "grad_norm": 0.47650372982025146,
      "learning_rate": 0.00016597953413222027,
      "loss": 2.936,
      "step": 107550
    },
    {
      "epoch": 28.956072351421188,
      "grad_norm": 0.5041089653968811,
      "learning_rate": 0.00016577756833176247,
      "loss": 2.9287,
      "step": 107600
    },
    {
      "epoch": 28.969530577088715,
      "grad_norm": 0.48725366592407227,
      "learning_rate": 0.0001655756025313047,
      "loss": 2.9442,
      "step": 107650
    },
    {
      "epoch": 28.982988802756246,
      "grad_norm": 0.5112107396125793,
      "learning_rate": 0.00016537363673084688,
      "loss": 2.9407,
      "step": 107700
    },
    {
      "epoch": 28.996447028423773,
      "grad_norm": 0.48162296414375305,
      "learning_rate": 0.0001651716709303891,
      "loss": 2.9424,
      "step": 107750
    },
    {
      "epoch": 29.00968992248062,
      "grad_norm": 0.4751878082752228,
      "learning_rate": 0.0001649697051299313,
      "loss": 2.8944,
      "step": 107800
    },
    {
      "epoch": 29.02314814814815,
      "grad_norm": 0.49943700432777405,
      "learning_rate": 0.0001647677393294735,
      "loss": 2.8711,
      "step": 107850
    },
    {
      "epoch": 29.036606373815676,
      "grad_norm": 0.48914316296577454,
      "learning_rate": 0.00016456577352901575,
      "loss": 2.8756,
      "step": 107900
    },
    {
      "epoch": 29.050064599483203,
      "grad_norm": 0.49434104561805725,
      "learning_rate": 0.00016436380772855797,
      "loss": 2.8756,
      "step": 107950
    },
    {
      "epoch": 29.06352282515073,
      "grad_norm": 0.48990321159362793,
      "learning_rate": 0.00016416184192810017,
      "loss": 2.871,
      "step": 108000
    },
    {
      "epoch": 29.06352282515073,
      "eval_accuracy": 0.3971213146443358,
      "eval_loss": 3.2863667011260986,
      "eval_runtime": 84.2038,
      "eval_samples_per_second": 213.91,
      "eval_steps_per_second": 13.372,
      "step": 108000
    },
    {
      "epoch": 29.07698105081826,
      "grad_norm": 0.49509531259536743,
      "learning_rate": 0.00016395987612764237,
      "loss": 2.878,
      "step": 108050
    },
    {
      "epoch": 29.09043927648579,
      "grad_norm": 0.5226240158081055,
      "learning_rate": 0.0001637579103271846,
      "loss": 2.8785,
      "step": 108100
    },
    {
      "epoch": 29.103897502153316,
      "grad_norm": 0.5094099640846252,
      "learning_rate": 0.00016355594452672679,
      "loss": 2.8687,
      "step": 108150
    },
    {
      "epoch": 29.117355727820843,
      "grad_norm": 0.4777970314025879,
      "learning_rate": 0.000163353978726269,
      "loss": 2.8918,
      "step": 108200
    },
    {
      "epoch": 29.13081395348837,
      "grad_norm": 0.5035146474838257,
      "learning_rate": 0.0001631520129258112,
      "loss": 2.8844,
      "step": 108250
    },
    {
      "epoch": 29.1442721791559,
      "grad_norm": 0.49367451667785645,
      "learning_rate": 0.0001629500471253534,
      "loss": 2.8811,
      "step": 108300
    },
    {
      "epoch": 29.15773040482343,
      "grad_norm": 0.4895753860473633,
      "learning_rate": 0.00016274808132489565,
      "loss": 2.8944,
      "step": 108350
    },
    {
      "epoch": 29.171188630490956,
      "grad_norm": 0.49232521653175354,
      "learning_rate": 0.00016254611552443785,
      "loss": 2.8815,
      "step": 108400
    },
    {
      "epoch": 29.184646856158484,
      "grad_norm": 0.5064100027084351,
      "learning_rate": 0.00016234414972398007,
      "loss": 2.8921,
      "step": 108450
    },
    {
      "epoch": 29.19810508182601,
      "grad_norm": 0.4918973743915558,
      "learning_rate": 0.00016214218392352227,
      "loss": 2.8809,
      "step": 108500
    },
    {
      "epoch": 29.21156330749354,
      "grad_norm": 0.4989228844642639,
      "learning_rate": 0.0001619402181230645,
      "loss": 2.884,
      "step": 108550
    },
    {
      "epoch": 29.22502153316107,
      "grad_norm": 0.4891134798526764,
      "learning_rate": 0.00016173825232260669,
      "loss": 2.9,
      "step": 108600
    },
    {
      "epoch": 29.238479758828596,
      "grad_norm": 0.4895051419734955,
      "learning_rate": 0.00016153628652214888,
      "loss": 2.8805,
      "step": 108650
    },
    {
      "epoch": 29.251937984496124,
      "grad_norm": 0.5228111743927002,
      "learning_rate": 0.0001613343207216911,
      "loss": 2.8874,
      "step": 108700
    },
    {
      "epoch": 29.26539621016365,
      "grad_norm": 0.48809224367141724,
      "learning_rate": 0.0001611323549212333,
      "loss": 2.8869,
      "step": 108750
    },
    {
      "epoch": 29.27885443583118,
      "grad_norm": 0.4924210011959076,
      "learning_rate": 0.00016093038912077555,
      "loss": 2.9054,
      "step": 108800
    },
    {
      "epoch": 29.29231266149871,
      "grad_norm": 0.5170987844467163,
      "learning_rate": 0.00016072842332031775,
      "loss": 2.9026,
      "step": 108850
    },
    {
      "epoch": 29.305770887166236,
      "grad_norm": 0.4849761426448822,
      "learning_rate": 0.00016052645751985997,
      "loss": 2.8882,
      "step": 108900
    },
    {
      "epoch": 29.319229112833764,
      "grad_norm": 0.4915884733200073,
      "learning_rate": 0.00016032449171940217,
      "loss": 2.9057,
      "step": 108950
    },
    {
      "epoch": 29.33268733850129,
      "grad_norm": 0.48439425230026245,
      "learning_rate": 0.0001601225259189444,
      "loss": 2.9031,
      "step": 109000
    },
    {
      "epoch": 29.33268733850129,
      "eval_accuracy": 0.3973558832054994,
      "eval_loss": 3.2866954803466797,
      "eval_runtime": 84.0378,
      "eval_samples_per_second": 214.332,
      "eval_steps_per_second": 13.399,
      "step": 109000
    },
    {
      "epoch": 29.34614556416882,
      "grad_norm": 0.48902153968811035,
      "learning_rate": 0.00015992056011848659,
      "loss": 2.8975,
      "step": 109050
    },
    {
      "epoch": 29.35960378983635,
      "grad_norm": 0.5074155926704407,
      "learning_rate": 0.00015971859431802878,
      "loss": 2.8957,
      "step": 109100
    },
    {
      "epoch": 29.373062015503876,
      "grad_norm": 0.4753362834453583,
      "learning_rate": 0.000159516628517571,
      "loss": 2.9135,
      "step": 109150
    },
    {
      "epoch": 29.386520241171404,
      "grad_norm": 0.49768611788749695,
      "learning_rate": 0.0001593146627171132,
      "loss": 2.9044,
      "step": 109200
    },
    {
      "epoch": 29.39997846683893,
      "grad_norm": 0.48798155784606934,
      "learning_rate": 0.00015911269691665545,
      "loss": 2.9125,
      "step": 109250
    },
    {
      "epoch": 29.41343669250646,
      "grad_norm": 0.4994659125804901,
      "learning_rate": 0.00015891073111619765,
      "loss": 2.9075,
      "step": 109300
    },
    {
      "epoch": 29.42689491817399,
      "grad_norm": 0.5110604763031006,
      "learning_rate": 0.00015870876531573987,
      "loss": 2.9191,
      "step": 109350
    },
    {
      "epoch": 29.440353143841516,
      "grad_norm": 0.49628448486328125,
      "learning_rate": 0.00015850679951528207,
      "loss": 2.9099,
      "step": 109400
    },
    {
      "epoch": 29.453811369509044,
      "grad_norm": 0.5112541317939758,
      "learning_rate": 0.00015830483371482426,
      "loss": 2.9164,
      "step": 109450
    },
    {
      "epoch": 29.46726959517657,
      "grad_norm": 0.5015026926994324,
      "learning_rate": 0.00015810286791436649,
      "loss": 2.9162,
      "step": 109500
    },
    {
      "epoch": 29.4807278208441,
      "grad_norm": 0.5245726704597473,
      "learning_rate": 0.00015790090211390868,
      "loss": 2.8955,
      "step": 109550
    },
    {
      "epoch": 29.49418604651163,
      "grad_norm": 0.5183436274528503,
      "learning_rate": 0.0001576989363134509,
      "loss": 2.9129,
      "step": 109600
    },
    {
      "epoch": 29.507644272179157,
      "grad_norm": 0.5848057270050049,
      "learning_rate": 0.0001574969705129931,
      "loss": 2.9056,
      "step": 109650
    },
    {
      "epoch": 29.521102497846684,
      "grad_norm": 0.5125300884246826,
      "learning_rate": 0.00015729500471253535,
      "loss": 2.909,
      "step": 109700
    },
    {
      "epoch": 29.53456072351421,
      "grad_norm": 0.5223360657691956,
      "learning_rate": 0.00015709303891207755,
      "loss": 2.9271,
      "step": 109750
    },
    {
      "epoch": 29.54801894918174,
      "grad_norm": 0.4989023506641388,
      "learning_rate": 0.00015689107311161977,
      "loss": 2.9126,
      "step": 109800
    },
    {
      "epoch": 29.56147717484927,
      "grad_norm": 0.4915999472141266,
      "learning_rate": 0.00015668910731116197,
      "loss": 2.9169,
      "step": 109850
    },
    {
      "epoch": 29.574935400516797,
      "grad_norm": 0.5052797794342041,
      "learning_rate": 0.00015648714151070416,
      "loss": 2.9242,
      "step": 109900
    },
    {
      "epoch": 29.588393626184324,
      "grad_norm": 0.5059757232666016,
      "learning_rate": 0.0001562851757102464,
      "loss": 2.9155,
      "step": 109950
    },
    {
      "epoch": 29.60185185185185,
      "grad_norm": 0.49516811966896057,
      "learning_rate": 0.00015608320990978858,
      "loss": 2.9175,
      "step": 110000
    },
    {
      "epoch": 29.60185185185185,
      "eval_accuracy": 0.39757187315436154,
      "eval_loss": 3.280747652053833,
      "eval_runtime": 83.9577,
      "eval_samples_per_second": 214.537,
      "eval_steps_per_second": 13.412,
      "step": 110000
    },
    {
      "epoch": 29.61531007751938,
      "grad_norm": 0.49822351336479187,
      "learning_rate": 0.0001558812441093308,
      "loss": 2.9191,
      "step": 110050
    },
    {
      "epoch": 29.62876830318691,
      "grad_norm": 0.518943727016449,
      "learning_rate": 0.000155679278308873,
      "loss": 2.9205,
      "step": 110100
    },
    {
      "epoch": 29.642226528854437,
      "grad_norm": 0.5562917590141296,
      "learning_rate": 0.00015547731250841525,
      "loss": 2.9141,
      "step": 110150
    },
    {
      "epoch": 29.655684754521964,
      "grad_norm": 0.5198628902435303,
      "learning_rate": 0.00015527534670795745,
      "loss": 2.9207,
      "step": 110200
    },
    {
      "epoch": 29.66914298018949,
      "grad_norm": 0.4802551865577698,
      "learning_rate": 0.00015507338090749964,
      "loss": 2.9129,
      "step": 110250
    },
    {
      "epoch": 29.68260120585702,
      "grad_norm": 0.5034878849983215,
      "learning_rate": 0.00015487141510704187,
      "loss": 2.9178,
      "step": 110300
    },
    {
      "epoch": 29.69605943152455,
      "grad_norm": 0.49084168672561646,
      "learning_rate": 0.00015466944930658406,
      "loss": 2.9193,
      "step": 110350
    },
    {
      "epoch": 29.709517657192077,
      "grad_norm": 0.490837961435318,
      "learning_rate": 0.0001544674835061263,
      "loss": 2.9108,
      "step": 110400
    },
    {
      "epoch": 29.722975882859604,
      "grad_norm": 0.5120250582695007,
      "learning_rate": 0.00015426551770566848,
      "loss": 2.912,
      "step": 110450
    },
    {
      "epoch": 29.73643410852713,
      "grad_norm": 0.5083431601524353,
      "learning_rate": 0.00015406355190521068,
      "loss": 2.9311,
      "step": 110500
    },
    {
      "epoch": 29.74989233419466,
      "grad_norm": 0.5029120445251465,
      "learning_rate": 0.0001538615861047529,
      "loss": 2.9308,
      "step": 110550
    },
    {
      "epoch": 29.76335055986219,
      "grad_norm": 0.4910016655921936,
      "learning_rate": 0.00015365962030429513,
      "loss": 2.9271,
      "step": 110600
    },
    {
      "epoch": 29.776808785529717,
      "grad_norm": 0.5051091313362122,
      "learning_rate": 0.00015345765450383735,
      "loss": 2.9172,
      "step": 110650
    },
    {
      "epoch": 29.790267011197244,
      "grad_norm": 0.5030775666236877,
      "learning_rate": 0.00015325568870337955,
      "loss": 2.9283,
      "step": 110700
    },
    {
      "epoch": 29.80372523686477,
      "grad_norm": 0.4919271171092987,
      "learning_rate": 0.00015305372290292177,
      "loss": 2.9327,
      "step": 110750
    },
    {
      "epoch": 29.8171834625323,
      "grad_norm": 0.5042194128036499,
      "learning_rate": 0.00015285175710246396,
      "loss": 2.9341,
      "step": 110800
    },
    {
      "epoch": 29.830641688199826,
      "grad_norm": 0.477828711271286,
      "learning_rate": 0.0001526497913020062,
      "loss": 2.9274,
      "step": 110850
    },
    {
      "epoch": 29.844099913867357,
      "grad_norm": 0.5096775889396667,
      "learning_rate": 0.00015244782550154838,
      "loss": 2.9346,
      "step": 110900
    },
    {
      "epoch": 29.857558139534884,
      "grad_norm": 0.49817365407943726,
      "learning_rate": 0.00015224585970109058,
      "loss": 2.9296,
      "step": 110950
    },
    {
      "epoch": 29.87101636520241,
      "grad_norm": 0.4907272756099701,
      "learning_rate": 0.00015204389390063283,
      "loss": 2.9304,
      "step": 111000
    },
    {
      "epoch": 29.87101636520241,
      "eval_accuracy": 0.3980782761481474,
      "eval_loss": 3.2766919136047363,
      "eval_runtime": 84.1125,
      "eval_samples_per_second": 214.142,
      "eval_steps_per_second": 13.387,
      "step": 111000
    },
    {
      "epoch": 29.88447459086994,
      "grad_norm": 0.5253500938415527,
      "learning_rate": 0.00015184192810017503,
      "loss": 2.9289,
      "step": 111050
    },
    {
      "epoch": 29.897932816537466,
      "grad_norm": 0.4913543462753296,
      "learning_rate": 0.00015163996229971725,
      "loss": 2.9185,
      "step": 111100
    },
    {
      "epoch": 29.911391042204997,
      "grad_norm": 0.492617130279541,
      "learning_rate": 0.00015143799649925945,
      "loss": 2.9429,
      "step": 111150
    },
    {
      "epoch": 29.924849267872524,
      "grad_norm": 0.5264730453491211,
      "learning_rate": 0.00015123603069880167,
      "loss": 2.9273,
      "step": 111200
    },
    {
      "epoch": 29.93830749354005,
      "grad_norm": 0.5183423757553101,
      "learning_rate": 0.00015103406489834386,
      "loss": 2.9241,
      "step": 111250
    },
    {
      "epoch": 29.95176571920758,
      "grad_norm": 0.5091369152069092,
      "learning_rate": 0.00015083209909788606,
      "loss": 2.9362,
      "step": 111300
    },
    {
      "epoch": 29.965223944875106,
      "grad_norm": 0.5015203952789307,
      "learning_rate": 0.00015063013329742828,
      "loss": 2.9342,
      "step": 111350
    },
    {
      "epoch": 29.978682170542637,
      "grad_norm": 0.4941323697566986,
      "learning_rate": 0.00015042816749697048,
      "loss": 2.9284,
      "step": 111400
    },
    {
      "epoch": 29.992140396210164,
      "grad_norm": 0.5067252516746521,
      "learning_rate": 0.00015022620169651273,
      "loss": 2.9318,
      "step": 111450
    },
    {
      "epoch": 30.005383290267012,
      "grad_norm": 0.48924145102500916,
      "learning_rate": 0.00015002423589605493,
      "loss": 2.9068,
      "step": 111500
    },
    {
      "epoch": 30.01884151593454,
      "grad_norm": 0.5250173807144165,
      "learning_rate": 0.00014982227009559712,
      "loss": 2.8651,
      "step": 111550
    },
    {
      "epoch": 30.032299741602067,
      "grad_norm": 0.49640926718711853,
      "learning_rate": 0.00014962030429513935,
      "loss": 2.8643,
      "step": 111600
    },
    {
      "epoch": 30.045757967269594,
      "grad_norm": 0.5105226039886475,
      "learning_rate": 0.00014941833849468154,
      "loss": 2.8743,
      "step": 111650
    },
    {
      "epoch": 30.05921619293712,
      "grad_norm": 0.48008471727371216,
      "learning_rate": 0.00014921637269422377,
      "loss": 2.8642,
      "step": 111700
    },
    {
      "epoch": 30.072674418604652,
      "grad_norm": 0.500213086605072,
      "learning_rate": 0.00014901440689376596,
      "loss": 2.862,
      "step": 111750
    },
    {
      "epoch": 30.08613264427218,
      "grad_norm": 0.506698489189148,
      "learning_rate": 0.00014881244109330818,
      "loss": 2.878,
      "step": 111800
    },
    {
      "epoch": 30.099590869939707,
      "grad_norm": 0.49218326807022095,
      "learning_rate": 0.0001486104752928504,
      "loss": 2.8731,
      "step": 111850
    },
    {
      "epoch": 30.113049095607234,
      "grad_norm": 0.5353227257728577,
      "learning_rate": 0.0001484085094923926,
      "loss": 2.8818,
      "step": 111900
    },
    {
      "epoch": 30.12650732127476,
      "grad_norm": 0.523637056350708,
      "learning_rate": 0.00014820654369193483,
      "loss": 2.8805,
      "step": 111950
    },
    {
      "epoch": 30.139965546942292,
      "grad_norm": 0.5045579671859741,
      "learning_rate": 0.00014800457789147702,
      "loss": 2.8797,
      "step": 112000
    },
    {
      "epoch": 30.139965546942292,
      "eval_accuracy": 0.39747952332713177,
      "eval_loss": 3.288581609725952,
      "eval_runtime": 83.7476,
      "eval_samples_per_second": 215.075,
      "eval_steps_per_second": 13.445,
      "step": 112000
    },
    {
      "epoch": 30.15342377260982,
      "grad_norm": 0.5382460355758667,
      "learning_rate": 0.00014780261209101925,
      "loss": 2.8683,
      "step": 112050
    },
    {
      "epoch": 30.166881998277347,
      "grad_norm": 0.49853479862213135,
      "learning_rate": 0.00014760064629056144,
      "loss": 2.8768,
      "step": 112100
    },
    {
      "epoch": 30.180340223944874,
      "grad_norm": 0.5337037444114685,
      "learning_rate": 0.00014739868049010367,
      "loss": 2.8751,
      "step": 112150
    },
    {
      "epoch": 30.1937984496124,
      "grad_norm": 0.486288458108902,
      "learning_rate": 0.00014719671468964586,
      "loss": 2.8913,
      "step": 112200
    },
    {
      "epoch": 30.207256675279933,
      "grad_norm": 0.4975809156894684,
      "learning_rate": 0.00014699474888918808,
      "loss": 2.8825,
      "step": 112250
    },
    {
      "epoch": 30.22071490094746,
      "grad_norm": 0.5215346813201904,
      "learning_rate": 0.0001467927830887303,
      "loss": 2.8828,
      "step": 112300
    },
    {
      "epoch": 30.234173126614987,
      "grad_norm": 0.5139694213867188,
      "learning_rate": 0.0001465908172882725,
      "loss": 2.872,
      "step": 112350
    },
    {
      "epoch": 30.247631352282514,
      "grad_norm": 0.5134860873222351,
      "learning_rate": 0.0001463888514878147,
      "loss": 2.8777,
      "step": 112400
    },
    {
      "epoch": 30.26108957795004,
      "grad_norm": 0.508992075920105,
      "learning_rate": 0.00014618688568735692,
      "loss": 2.8766,
      "step": 112450
    },
    {
      "epoch": 30.274547803617573,
      "grad_norm": 0.5374428629875183,
      "learning_rate": 0.00014598491988689915,
      "loss": 2.8917,
      "step": 112500
    },
    {
      "epoch": 30.2880060292851,
      "grad_norm": 0.503511905670166,
      "learning_rate": 0.00014578295408644134,
      "loss": 2.8927,
      "step": 112550
    },
    {
      "epoch": 30.301464254952627,
      "grad_norm": 0.4890826940536499,
      "learning_rate": 0.00014558098828598357,
      "loss": 2.9034,
      "step": 112600
    },
    {
      "epoch": 30.314922480620154,
      "grad_norm": 0.5056515336036682,
      "learning_rate": 0.00014537902248552576,
      "loss": 2.8881,
      "step": 112650
    },
    {
      "epoch": 30.328380706287682,
      "grad_norm": 0.520115315914154,
      "learning_rate": 0.00014517705668506799,
      "loss": 2.8921,
      "step": 112700
    },
    {
      "epoch": 30.341838931955213,
      "grad_norm": 0.4846559762954712,
      "learning_rate": 0.0001449750908846102,
      "loss": 2.8974,
      "step": 112750
    },
    {
      "epoch": 30.35529715762274,
      "grad_norm": 0.5147880911827087,
      "learning_rate": 0.0001447731250841524,
      "loss": 2.8928,
      "step": 112800
    },
    {
      "epoch": 30.368755383290267,
      "grad_norm": 0.5425703525543213,
      "learning_rate": 0.0001445711592836946,
      "loss": 2.9002,
      "step": 112850
    },
    {
      "epoch": 30.382213608957795,
      "grad_norm": 0.5446016192436218,
      "learning_rate": 0.00014436919348323682,
      "loss": 2.8988,
      "step": 112900
    },
    {
      "epoch": 30.395671834625322,
      "grad_norm": 0.48981142044067383,
      "learning_rate": 0.00014416722768277905,
      "loss": 2.8945,
      "step": 112950
    },
    {
      "epoch": 30.409130060292853,
      "grad_norm": 0.5191630721092224,
      "learning_rate": 0.00014396526188232124,
      "loss": 2.9046,
      "step": 113000
    },
    {
      "epoch": 30.409130060292853,
      "eval_accuracy": 0.39749147448124383,
      "eval_loss": 3.285137891769409,
      "eval_runtime": 83.9831,
      "eval_samples_per_second": 214.472,
      "eval_steps_per_second": 13.407,
      "step": 113000
    },
    {
      "epoch": 30.42258828596038,
      "grad_norm": 0.4975392520427704,
      "learning_rate": 0.00014376329608186347,
      "loss": 2.9018,
      "step": 113050
    },
    {
      "epoch": 30.436046511627907,
      "grad_norm": 0.49719271063804626,
      "learning_rate": 0.00014356133028140566,
      "loss": 2.8995,
      "step": 113100
    },
    {
      "epoch": 30.449504737295435,
      "grad_norm": 0.49405014514923096,
      "learning_rate": 0.00014335936448094789,
      "loss": 2.8976,
      "step": 113150
    },
    {
      "epoch": 30.462962962962962,
      "grad_norm": 0.5243561267852783,
      "learning_rate": 0.00014315739868049008,
      "loss": 2.8968,
      "step": 113200
    },
    {
      "epoch": 30.47642118863049,
      "grad_norm": 0.49951159954071045,
      "learning_rate": 0.0001429554328800323,
      "loss": 2.9026,
      "step": 113250
    },
    {
      "epoch": 30.48987941429802,
      "grad_norm": 0.5445297956466675,
      "learning_rate": 0.0001427534670795745,
      "loss": 2.9019,
      "step": 113300
    },
    {
      "epoch": 30.503337639965547,
      "grad_norm": 0.49440625309944153,
      "learning_rate": 0.00014255150127911672,
      "loss": 2.9085,
      "step": 113350
    },
    {
      "epoch": 30.516795865633075,
      "grad_norm": 0.5150471329689026,
      "learning_rate": 0.00014234953547865895,
      "loss": 2.9015,
      "step": 113400
    },
    {
      "epoch": 30.530254091300602,
      "grad_norm": 0.5206648707389832,
      "learning_rate": 0.00014214756967820114,
      "loss": 2.9063,
      "step": 113450
    },
    {
      "epoch": 30.54371231696813,
      "grad_norm": 0.5529144406318665,
      "learning_rate": 0.00014194560387774334,
      "loss": 2.8991,
      "step": 113500
    },
    {
      "epoch": 30.55717054263566,
      "grad_norm": 0.5388721227645874,
      "learning_rate": 0.00014174363807728556,
      "loss": 2.8996,
      "step": 113550
    },
    {
      "epoch": 30.570628768303187,
      "grad_norm": 0.4762554168701172,
      "learning_rate": 0.00014154167227682779,
      "loss": 2.9152,
      "step": 113600
    },
    {
      "epoch": 30.584086993970715,
      "grad_norm": 0.5123204588890076,
      "learning_rate": 0.00014133970647636998,
      "loss": 2.9115,
      "step": 113650
    },
    {
      "epoch": 30.597545219638242,
      "grad_norm": 0.5319657921791077,
      "learning_rate": 0.0001411377406759122,
      "loss": 2.9054,
      "step": 113700
    },
    {
      "epoch": 30.61100344530577,
      "grad_norm": 0.5207548141479492,
      "learning_rate": 0.0001409357748754544,
      "loss": 2.9148,
      "step": 113750
    },
    {
      "epoch": 30.6244616709733,
      "grad_norm": 0.512873649597168,
      "learning_rate": 0.00014073380907499662,
      "loss": 2.9115,
      "step": 113800
    },
    {
      "epoch": 30.637919896640827,
      "grad_norm": 0.5313157439231873,
      "learning_rate": 0.00014053184327453885,
      "loss": 2.8988,
      "step": 113850
    },
    {
      "epoch": 30.651378122308355,
      "grad_norm": 0.5074067115783691,
      "learning_rate": 0.00014032987747408104,
      "loss": 2.9006,
      "step": 113900
    },
    {
      "epoch": 30.664836347975882,
      "grad_norm": 0.49639374017715454,
      "learning_rate": 0.00014012791167362324,
      "loss": 2.9073,
      "step": 113950
    },
    {
      "epoch": 30.67829457364341,
      "grad_norm": 0.5084487199783325,
      "learning_rate": 0.00013992594587316546,
      "loss": 2.9068,
      "step": 114000
    },
    {
      "epoch": 30.67829457364341,
      "eval_accuracy": 0.39820126438864634,
      "eval_loss": 3.2775118350982666,
      "eval_runtime": 84.2536,
      "eval_samples_per_second": 213.783,
      "eval_steps_per_second": 13.364,
      "step": 114000
    },
    {
      "epoch": 30.69175279931094,
      "grad_norm": 0.5242618918418884,
      "learning_rate": 0.0001397239800727077,
      "loss": 2.9196,
      "step": 114050
    },
    {
      "epoch": 30.705211024978468,
      "grad_norm": 0.5300239324569702,
      "learning_rate": 0.00013952201427224988,
      "loss": 2.9037,
      "step": 114100
    },
    {
      "epoch": 30.718669250645995,
      "grad_norm": 0.48562678694725037,
      "learning_rate": 0.0001393200484717921,
      "loss": 2.9132,
      "step": 114150
    },
    {
      "epoch": 30.732127476313522,
      "grad_norm": 0.538164496421814,
      "learning_rate": 0.0001391180826713343,
      "loss": 2.919,
      "step": 114200
    },
    {
      "epoch": 30.74558570198105,
      "grad_norm": 0.518423855304718,
      "learning_rate": 0.00013891611687087653,
      "loss": 2.9107,
      "step": 114250
    },
    {
      "epoch": 30.75904392764858,
      "grad_norm": 0.5549152493476868,
      "learning_rate": 0.00013871415107041872,
      "loss": 2.9134,
      "step": 114300
    },
    {
      "epoch": 30.772502153316108,
      "grad_norm": 0.4988192021846771,
      "learning_rate": 0.00013851218526996094,
      "loss": 2.916,
      "step": 114350
    },
    {
      "epoch": 30.785960378983635,
      "grad_norm": 0.5223867893218994,
      "learning_rate": 0.00013831021946950314,
      "loss": 2.9046,
      "step": 114400
    },
    {
      "epoch": 30.799418604651162,
      "grad_norm": 0.5547376275062561,
      "learning_rate": 0.00013810825366904536,
      "loss": 2.9202,
      "step": 114450
    },
    {
      "epoch": 30.81287683031869,
      "grad_norm": 0.48123112320899963,
      "learning_rate": 0.0001379062878685876,
      "loss": 2.9069,
      "step": 114500
    },
    {
      "epoch": 30.82633505598622,
      "grad_norm": 0.5459958910942078,
      "learning_rate": 0.00013770432206812978,
      "loss": 2.9168,
      "step": 114550
    },
    {
      "epoch": 30.839793281653748,
      "grad_norm": 0.5234883427619934,
      "learning_rate": 0.00013750235626767198,
      "loss": 2.9114,
      "step": 114600
    },
    {
      "epoch": 30.853251507321275,
      "grad_norm": 0.5141161680221558,
      "learning_rate": 0.0001373003904672142,
      "loss": 2.9134,
      "step": 114650
    },
    {
      "epoch": 30.866709732988802,
      "grad_norm": 0.5167535543441772,
      "learning_rate": 0.00013709842466675643,
      "loss": 2.9253,
      "step": 114700
    },
    {
      "epoch": 30.88016795865633,
      "grad_norm": 0.535083532333374,
      "learning_rate": 0.00013689645886629862,
      "loss": 2.9163,
      "step": 114750
    },
    {
      "epoch": 30.89362618432386,
      "grad_norm": 0.5264659523963928,
      "learning_rate": 0.00013669449306584084,
      "loss": 2.915,
      "step": 114800
    },
    {
      "epoch": 30.907084409991388,
      "grad_norm": 0.5105305314064026,
      "learning_rate": 0.00013649252726538304,
      "loss": 2.9142,
      "step": 114850
    },
    {
      "epoch": 30.920542635658915,
      "grad_norm": 0.5078901052474976,
      "learning_rate": 0.00013629056146492526,
      "loss": 2.912,
      "step": 114900
    },
    {
      "epoch": 30.934000861326442,
      "grad_norm": 0.5027287602424622,
      "learning_rate": 0.0001360885956644675,
      "loss": 2.9141,
      "step": 114950
    },
    {
      "epoch": 30.94745908699397,
      "grad_norm": 0.5367311835289001,
      "learning_rate": 0.00013588662986400968,
      "loss": 2.92,
      "step": 115000
    },
    {
      "epoch": 30.94745908699397,
      "eval_accuracy": 0.39823385844531567,
      "eval_loss": 3.2751386165618896,
      "eval_runtime": 84.2856,
      "eval_samples_per_second": 213.702,
      "eval_steps_per_second": 13.359,
      "step": 115000
    },
    {
      "epoch": 30.9609173126615,
      "grad_norm": 0.515602171421051,
      "learning_rate": 0.00013568466406355188,
      "loss": 2.9224,
      "step": 115050
    },
    {
      "epoch": 30.974375538329028,
      "grad_norm": 0.5385026931762695,
      "learning_rate": 0.0001354826982630941,
      "loss": 2.9246,
      "step": 115100
    },
    {
      "epoch": 30.987833763996555,
      "grad_norm": 0.5618683695793152,
      "learning_rate": 0.00013528073246263633,
      "loss": 2.9148,
      "step": 115150
    },
    {
      "epoch": 31.001076658053403,
      "grad_norm": 0.5228216648101807,
      "learning_rate": 0.00013507876666217852,
      "loss": 2.9174,
      "step": 115200
    },
    {
      "epoch": 31.01453488372093,
      "grad_norm": 0.5337773561477661,
      "learning_rate": 0.00013487680086172075,
      "loss": 2.8479,
      "step": 115250
    },
    {
      "epoch": 31.027993109388458,
      "grad_norm": 0.4986388385295868,
      "learning_rate": 0.00013467483506126294,
      "loss": 2.8559,
      "step": 115300
    },
    {
      "epoch": 31.041451335055985,
      "grad_norm": 0.5042169094085693,
      "learning_rate": 0.00013447286926080516,
      "loss": 2.8558,
      "step": 115350
    },
    {
      "epoch": 31.054909560723516,
      "grad_norm": 0.5418523550033569,
      "learning_rate": 0.00013427090346034736,
      "loss": 2.8677,
      "step": 115400
    },
    {
      "epoch": 31.068367786391043,
      "grad_norm": 0.5390510559082031,
      "learning_rate": 0.00013406893765988958,
      "loss": 2.8706,
      "step": 115450
    },
    {
      "epoch": 31.08182601205857,
      "grad_norm": 0.5468896627426147,
      "learning_rate": 0.00013386697185943178,
      "loss": 2.8598,
      "step": 115500
    },
    {
      "epoch": 31.095284237726098,
      "grad_norm": 0.5302108526229858,
      "learning_rate": 0.000133665006058974,
      "loss": 2.8575,
      "step": 115550
    },
    {
      "epoch": 31.108742463393625,
      "grad_norm": 0.5282973647117615,
      "learning_rate": 0.00013346304025851623,
      "loss": 2.857,
      "step": 115600
    },
    {
      "epoch": 31.122200689061152,
      "grad_norm": 0.5058830976486206,
      "learning_rate": 0.00013326107445805842,
      "loss": 2.8662,
      "step": 115650
    },
    {
      "epoch": 31.135658914728683,
      "grad_norm": 0.5025017857551575,
      "learning_rate": 0.00013305910865760065,
      "loss": 2.8768,
      "step": 115700
    },
    {
      "epoch": 31.14911714039621,
      "grad_norm": 0.5279959440231323,
      "learning_rate": 0.00013285714285714284,
      "loss": 2.8686,
      "step": 115750
    },
    {
      "epoch": 31.162575366063738,
      "grad_norm": 0.5327432751655579,
      "learning_rate": 0.00013265517705668506,
      "loss": 2.8695,
      "step": 115800
    },
    {
      "epoch": 31.176033591731265,
      "grad_norm": 0.509773313999176,
      "learning_rate": 0.00013245321125622726,
      "loss": 2.868,
      "step": 115850
    },
    {
      "epoch": 31.189491817398793,
      "grad_norm": 0.5246791243553162,
      "learning_rate": 0.00013225124545576948,
      "loss": 2.8735,
      "step": 115900
    },
    {
      "epoch": 31.202950043066323,
      "grad_norm": 0.5291630625724792,
      "learning_rate": 0.00013204927965531168,
      "loss": 2.8675,
      "step": 115950
    },
    {
      "epoch": 31.21640826873385,
      "grad_norm": 0.5185742378234863,
      "learning_rate": 0.0001318473138548539,
      "loss": 2.8819,
      "step": 116000
    },
    {
      "epoch": 31.21640826873385,
      "eval_accuracy": 0.39740749046189255,
      "eval_loss": 3.2877187728881836,
      "eval_runtime": 84.0977,
      "eval_samples_per_second": 214.179,
      "eval_steps_per_second": 13.389,
      "step": 116000
    },
    {
      "epoch": 31.229866494401378,
      "grad_norm": 0.502913773059845,
      "learning_rate": 0.00013164534805439613,
      "loss": 2.8652,
      "step": 116050
    },
    {
      "epoch": 31.243324720068905,
      "grad_norm": 0.5236504077911377,
      "learning_rate": 0.00013144338225393832,
      "loss": 2.8694,
      "step": 116100
    },
    {
      "epoch": 31.256782945736433,
      "grad_norm": 0.5153883099555969,
      "learning_rate": 0.00013124141645348052,
      "loss": 2.8824,
      "step": 116150
    },
    {
      "epoch": 31.270241171403963,
      "grad_norm": 0.5336629152297974,
      "learning_rate": 0.00013103945065302274,
      "loss": 2.8785,
      "step": 116200
    },
    {
      "epoch": 31.28369939707149,
      "grad_norm": 0.509537935256958,
      "learning_rate": 0.00013083748485256497,
      "loss": 2.873,
      "step": 116250
    },
    {
      "epoch": 31.297157622739018,
      "grad_norm": 0.49638795852661133,
      "learning_rate": 0.00013063551905210716,
      "loss": 2.8774,
      "step": 116300
    },
    {
      "epoch": 31.310615848406545,
      "grad_norm": 0.5222117304801941,
      "learning_rate": 0.00013043355325164938,
      "loss": 2.8789,
      "step": 116350
    },
    {
      "epoch": 31.324074074074073,
      "grad_norm": 0.5545364022254944,
      "learning_rate": 0.00013023158745119158,
      "loss": 2.8803,
      "step": 116400
    },
    {
      "epoch": 31.337532299741603,
      "grad_norm": 0.4968852400779724,
      "learning_rate": 0.00013002962165073378,
      "loss": 2.8819,
      "step": 116450
    },
    {
      "epoch": 31.35099052540913,
      "grad_norm": 0.5044577121734619,
      "learning_rate": 0.000129827655850276,
      "loss": 2.8825,
      "step": 116500
    },
    {
      "epoch": 31.364448751076658,
      "grad_norm": 0.5312719941139221,
      "learning_rate": 0.00012962569004981822,
      "loss": 2.8717,
      "step": 116550
    },
    {
      "epoch": 31.377906976744185,
      "grad_norm": 0.5037994980812073,
      "learning_rate": 0.00012942372424936042,
      "loss": 2.8886,
      "step": 116600
    },
    {
      "epoch": 31.391365202411713,
      "grad_norm": 0.48807597160339355,
      "learning_rate": 0.00012922175844890264,
      "loss": 2.887,
      "step": 116650
    },
    {
      "epoch": 31.404823428079244,
      "grad_norm": 0.5140440464019775,
      "learning_rate": 0.00012901979264844487,
      "loss": 2.8877,
      "step": 116700
    },
    {
      "epoch": 31.41828165374677,
      "grad_norm": 0.5027121901512146,
      "learning_rate": 0.00012881782684798706,
      "loss": 2.8817,
      "step": 116750
    },
    {
      "epoch": 31.431739879414298,
      "grad_norm": 0.553657054901123,
      "learning_rate": 0.00012861586104752928,
      "loss": 2.8978,
      "step": 116800
    },
    {
      "epoch": 31.445198105081825,
      "grad_norm": 0.506435751914978,
      "learning_rate": 0.00012841389524707148,
      "loss": 2.8916,
      "step": 116850
    },
    {
      "epoch": 31.458656330749353,
      "grad_norm": 0.562160313129425,
      "learning_rate": 0.0001282119294466137,
      "loss": 2.8851,
      "step": 116900
    },
    {
      "epoch": 31.472114556416884,
      "grad_norm": 0.5217667818069458,
      "learning_rate": 0.0001280099636461559,
      "loss": 2.8827,
      "step": 116950
    },
    {
      "epoch": 31.48557278208441,
      "grad_norm": 0.5382517576217651,
      "learning_rate": 0.00012780799784569812,
      "loss": 2.9002,
      "step": 117000
    },
    {
      "epoch": 31.48557278208441,
      "eval_accuracy": 0.3978187188102039,
      "eval_loss": 3.2845301628112793,
      "eval_runtime": 84.3616,
      "eval_samples_per_second": 213.51,
      "eval_steps_per_second": 13.347,
      "step": 117000
    },
    {
      "epoch": 31.499031007751938,
      "grad_norm": 0.5561841130256653,
      "learning_rate": 0.00012760603204524032,
      "loss": 2.8919,
      "step": 117050
    },
    {
      "epoch": 31.512489233419466,
      "grad_norm": 0.5291392207145691,
      "learning_rate": 0.00012740406624478254,
      "loss": 2.8964,
      "step": 117100
    },
    {
      "epoch": 31.525947459086993,
      "grad_norm": 0.4918546974658966,
      "learning_rate": 0.00012720210044432477,
      "loss": 2.8958,
      "step": 117150
    },
    {
      "epoch": 31.539405684754524,
      "grad_norm": 0.49700331687927246,
      "learning_rate": 0.00012700013464386696,
      "loss": 2.8983,
      "step": 117200
    },
    {
      "epoch": 31.55286391042205,
      "grad_norm": 0.5119349956512451,
      "learning_rate": 0.00012679816884340916,
      "loss": 2.8947,
      "step": 117250
    },
    {
      "epoch": 31.56632213608958,
      "grad_norm": 0.5518103241920471,
      "learning_rate": 0.00012659620304295138,
      "loss": 2.9039,
      "step": 117300
    },
    {
      "epoch": 31.579780361757106,
      "grad_norm": 0.543112576007843,
      "learning_rate": 0.0001263942372424936,
      "loss": 2.903,
      "step": 117350
    },
    {
      "epoch": 31.593238587424633,
      "grad_norm": 0.49518176913261414,
      "learning_rate": 0.0001261922714420358,
      "loss": 2.8945,
      "step": 117400
    },
    {
      "epoch": 31.606696813092164,
      "grad_norm": 0.5011724233627319,
      "learning_rate": 0.00012599030564157802,
      "loss": 2.9032,
      "step": 117450
    },
    {
      "epoch": 31.62015503875969,
      "grad_norm": 0.5476980209350586,
      "learning_rate": 0.00012578833984112022,
      "loss": 2.9047,
      "step": 117500
    },
    {
      "epoch": 31.63361326442722,
      "grad_norm": 0.5187918543815613,
      "learning_rate": 0.00012558637404066244,
      "loss": 2.8962,
      "step": 117550
    },
    {
      "epoch": 31.647071490094746,
      "grad_norm": 0.5203251838684082,
      "learning_rate": 0.00012538440824020467,
      "loss": 2.8932,
      "step": 117600
    },
    {
      "epoch": 31.660529715762273,
      "grad_norm": 0.49319368600845337,
      "learning_rate": 0.00012518244243974686,
      "loss": 2.8982,
      "step": 117650
    },
    {
      "epoch": 31.6739879414298,
      "grad_norm": 0.5315812230110168,
      "learning_rate": 0.00012498047663928906,
      "loss": 2.9009,
      "step": 117700
    },
    {
      "epoch": 31.68744616709733,
      "grad_norm": 0.5264846682548523,
      "learning_rate": 0.00012477851083883128,
      "loss": 2.9001,
      "step": 117750
    },
    {
      "epoch": 31.70090439276486,
      "grad_norm": 0.5249003767967224,
      "learning_rate": 0.0001245765450383735,
      "loss": 2.9071,
      "step": 117800
    },
    {
      "epoch": 31.714362618432386,
      "grad_norm": 0.5264038443565369,
      "learning_rate": 0.0001243745792379157,
      "loss": 2.8998,
      "step": 117850
    },
    {
      "epoch": 31.727820844099913,
      "grad_norm": 0.5106028318405151,
      "learning_rate": 0.00012417261343745792,
      "loss": 2.9045,
      "step": 117900
    },
    {
      "epoch": 31.74127906976744,
      "grad_norm": 0.5077876448631287,
      "learning_rate": 0.00012397064763700012,
      "loss": 2.9034,
      "step": 117950
    },
    {
      "epoch": 31.75473729543497,
      "grad_norm": 0.5172459483146667,
      "learning_rate": 0.00012376868183654232,
      "loss": 2.8976,
      "step": 118000
    },
    {
      "epoch": 31.75473729543497,
      "eval_accuracy": 0.3982812284743417,
      "eval_loss": 3.278949737548828,
      "eval_runtime": 84.2828,
      "eval_samples_per_second": 213.709,
      "eval_steps_per_second": 13.36,
      "step": 118000
    },
    {
      "epoch": 31.7681955211025,
      "grad_norm": 0.533544659614563,
      "learning_rate": 0.00012356671603608454,
      "loss": 2.9008,
      "step": 118050
    },
    {
      "epoch": 31.781653746770026,
      "grad_norm": 0.5425562262535095,
      "learning_rate": 0.00012336475023562676,
      "loss": 2.9001,
      "step": 118100
    },
    {
      "epoch": 31.795111972437553,
      "grad_norm": 0.5205543041229248,
      "learning_rate": 0.00012316278443516896,
      "loss": 2.9033,
      "step": 118150
    },
    {
      "epoch": 31.80857019810508,
      "grad_norm": 0.5152506232261658,
      "learning_rate": 0.00012296081863471118,
      "loss": 2.8986,
      "step": 118200
    },
    {
      "epoch": 31.82202842377261,
      "grad_norm": 0.5451449751853943,
      "learning_rate": 0.0001227588528342534,
      "loss": 2.893,
      "step": 118250
    },
    {
      "epoch": 31.83548664944014,
      "grad_norm": 0.5216442942619324,
      "learning_rate": 0.0001225568870337956,
      "loss": 2.9036,
      "step": 118300
    },
    {
      "epoch": 31.848944875107666,
      "grad_norm": 0.5051542520523071,
      "learning_rate": 0.0001223549212333378,
      "loss": 2.9151,
      "step": 118350
    },
    {
      "epoch": 31.862403100775193,
      "grad_norm": 0.5095641016960144,
      "learning_rate": 0.00012215295543288002,
      "loss": 2.911,
      "step": 118400
    },
    {
      "epoch": 31.87586132644272,
      "grad_norm": 0.5289960503578186,
      "learning_rate": 0.00012195098963242223,
      "loss": 2.9136,
      "step": 118450
    },
    {
      "epoch": 31.88931955211025,
      "grad_norm": 0.5164697766304016,
      "learning_rate": 0.00012174902383196445,
      "loss": 2.9122,
      "step": 118500
    },
    {
      "epoch": 31.90277777777778,
      "grad_norm": 0.5209395289421082,
      "learning_rate": 0.00012154705803150665,
      "loss": 2.9047,
      "step": 118550
    },
    {
      "epoch": 31.916236003445306,
      "grad_norm": 0.5136258602142334,
      "learning_rate": 0.00012134509223104886,
      "loss": 2.908,
      "step": 118600
    },
    {
      "epoch": 31.929694229112833,
      "grad_norm": 0.5150439143180847,
      "learning_rate": 0.00012114312643059107,
      "loss": 2.9137,
      "step": 118650
    },
    {
      "epoch": 31.94315245478036,
      "grad_norm": 0.5110583901405334,
      "learning_rate": 0.00012094116063013329,
      "loss": 2.9014,
      "step": 118700
    },
    {
      "epoch": 31.95661068044789,
      "grad_norm": 0.4987826645374298,
      "learning_rate": 0.0001207391948296755,
      "loss": 2.9139,
      "step": 118750
    },
    {
      "epoch": 31.97006890611542,
      "grad_norm": 0.5320709347724915,
      "learning_rate": 0.00012053722902921771,
      "loss": 2.9127,
      "step": 118800
    },
    {
      "epoch": 31.983527131782946,
      "grad_norm": 0.5164916515350342,
      "learning_rate": 0.00012033526322875992,
      "loss": 2.9032,
      "step": 118850
    },
    {
      "epoch": 31.996985357450473,
      "grad_norm": 0.526587724685669,
      "learning_rate": 0.00012013329742830212,
      "loss": 2.9163,
      "step": 118900
    },
    {
      "epoch": 32.01022825150732,
      "grad_norm": 0.5279465317726135,
      "learning_rate": 0.00011993133162784434,
      "loss": 2.8669,
      "step": 118950
    },
    {
      "epoch": 32.02368647717485,
      "grad_norm": 0.5259037613868713,
      "learning_rate": 0.00011972936582738655,
      "loss": 2.8533,
      "step": 119000
    },
    {
      "epoch": 32.02368647717485,
      "eval_accuracy": 0.397786016106679,
      "eval_loss": 3.285374641418457,
      "eval_runtime": 84.1514,
      "eval_samples_per_second": 214.043,
      "eval_steps_per_second": 13.381,
      "step": 119000
    },
    {
      "epoch": 32.037144702842376,
      "grad_norm": 0.5307081937789917,
      "learning_rate": 0.00011952740002692876,
      "loss": 2.8442,
      "step": 119050
    },
    {
      "epoch": 32.0506029285099,
      "grad_norm": 0.5095431208610535,
      "learning_rate": 0.00011932543422647097,
      "loss": 2.8613,
      "step": 119100
    },
    {
      "epoch": 32.06406115417743,
      "grad_norm": 0.52222740650177,
      "learning_rate": 0.00011912346842601319,
      "loss": 2.8608,
      "step": 119150
    },
    {
      "epoch": 32.07751937984496,
      "grad_norm": 0.5147879123687744,
      "learning_rate": 0.0001189215026255554,
      "loss": 2.8635,
      "step": 119200
    },
    {
      "epoch": 32.09097760551249,
      "grad_norm": 0.506778359413147,
      "learning_rate": 0.00011871953682509761,
      "loss": 2.85,
      "step": 119250
    },
    {
      "epoch": 32.10443583118002,
      "grad_norm": 0.509945273399353,
      "learning_rate": 0.00011851757102463981,
      "loss": 2.8524,
      "step": 119300
    },
    {
      "epoch": 32.11789405684755,
      "grad_norm": 0.539358913898468,
      "learning_rate": 0.00011831560522418203,
      "loss": 2.8589,
      "step": 119350
    },
    {
      "epoch": 32.131352282515074,
      "grad_norm": 0.5196101665496826,
      "learning_rate": 0.00011811363942372424,
      "loss": 2.8532,
      "step": 119400
    },
    {
      "epoch": 32.1448105081826,
      "grad_norm": 0.5178312063217163,
      "learning_rate": 0.00011791167362326645,
      "loss": 2.8566,
      "step": 119450
    },
    {
      "epoch": 32.15826873385013,
      "grad_norm": 0.5040403604507446,
      "learning_rate": 0.00011770970782280866,
      "loss": 2.8624,
      "step": 119500
    },
    {
      "epoch": 32.171726959517656,
      "grad_norm": 0.522476315498352,
      "learning_rate": 0.00011750774202235087,
      "loss": 2.8606,
      "step": 119550
    },
    {
      "epoch": 32.18518518518518,
      "grad_norm": 0.5482622981071472,
      "learning_rate": 0.00011730577622189309,
      "loss": 2.8635,
      "step": 119600
    },
    {
      "epoch": 32.19864341085271,
      "grad_norm": 0.5259464979171753,
      "learning_rate": 0.00011710381042143529,
      "loss": 2.8532,
      "step": 119650
    },
    {
      "epoch": 32.21210163652024,
      "grad_norm": 0.554554283618927,
      "learning_rate": 0.0001169018446209775,
      "loss": 2.8685,
      "step": 119700
    },
    {
      "epoch": 32.22555986218777,
      "grad_norm": 0.5307154655456543,
      "learning_rate": 0.00011669987882051971,
      "loss": 2.8654,
      "step": 119750
    },
    {
      "epoch": 32.2390180878553,
      "grad_norm": 0.5246207118034363,
      "learning_rate": 0.00011649791302006193,
      "loss": 2.8725,
      "step": 119800
    },
    {
      "epoch": 32.25247631352283,
      "grad_norm": 0.497340589761734,
      "learning_rate": 0.00011629594721960414,
      "loss": 2.8747,
      "step": 119850
    },
    {
      "epoch": 32.265934539190354,
      "grad_norm": 0.5233364701271057,
      "learning_rate": 0.00011609398141914635,
      "loss": 2.8645,
      "step": 119900
    },
    {
      "epoch": 32.27939276485788,
      "grad_norm": 0.5302901864051819,
      "learning_rate": 0.00011589201561868856,
      "loss": 2.8607,
      "step": 119950
    },
    {
      "epoch": 32.29285099052541,
      "grad_norm": 0.4998948574066162,
      "learning_rate": 0.00011569004981823076,
      "loss": 2.8772,
      "step": 120000
    },
    {
      "epoch": 32.29285099052541,
      "eval_accuracy": 0.3977070298426837,
      "eval_loss": 3.2868707180023193,
      "eval_runtime": 84.1065,
      "eval_samples_per_second": 214.157,
      "eval_steps_per_second": 13.388,
      "step": 120000
    },
    {
      "epoch": 32.306309216192936,
      "grad_norm": 0.5445365905761719,
      "learning_rate": 0.00011548808401777298,
      "loss": 2.8473,
      "step": 120050
    },
    {
      "epoch": 32.31976744186046,
      "grad_norm": 0.5282999873161316,
      "learning_rate": 0.00011528611821731519,
      "loss": 2.8531,
      "step": 120100
    },
    {
      "epoch": 32.33322566752799,
      "grad_norm": 0.5160717368125916,
      "learning_rate": 0.0001150841524168574,
      "loss": 2.85,
      "step": 120150
    },
    {
      "epoch": 32.34668389319552,
      "grad_norm": 0.5283183455467224,
      "learning_rate": 0.00011488218661639961,
      "loss": 2.8473,
      "step": 120200
    },
    {
      "epoch": 32.36014211886305,
      "grad_norm": 0.5090058445930481,
      "learning_rate": 0.00011468022081594183,
      "loss": 2.8662,
      "step": 120250
    },
    {
      "epoch": 32.37360034453058,
      "grad_norm": 0.4993782639503479,
      "learning_rate": 0.00011447825501548404,
      "loss": 2.8537,
      "step": 120300
    },
    {
      "epoch": 32.38705857019811,
      "grad_norm": 0.5348803997039795,
      "learning_rate": 0.00011427628921502625,
      "loss": 2.8453,
      "step": 120350
    },
    {
      "epoch": 32.400516795865634,
      "grad_norm": 0.5482392311096191,
      "learning_rate": 0.00011407432341456845,
      "loss": 2.8524,
      "step": 120400
    },
    {
      "epoch": 32.41397502153316,
      "grad_norm": 0.5049187541007996,
      "learning_rate": 0.00011387235761411066,
      "loss": 2.8509,
      "step": 120450
    },
    {
      "epoch": 32.42743324720069,
      "grad_norm": 0.544791579246521,
      "learning_rate": 0.00011367039181365288,
      "loss": 2.8615,
      "step": 120500
    },
    {
      "epoch": 32.440891472868216,
      "grad_norm": 0.5117326378822327,
      "learning_rate": 0.00011346842601319509,
      "loss": 2.8648,
      "step": 120550
    },
    {
      "epoch": 32.454349698535744,
      "grad_norm": 0.5414500832557678,
      "learning_rate": 0.0001132664602127373,
      "loss": 2.8611,
      "step": 120600
    },
    {
      "epoch": 32.46780792420327,
      "grad_norm": 0.5438842177391052,
      "learning_rate": 0.00011306449441227951,
      "loss": 2.8621,
      "step": 120650
    },
    {
      "epoch": 32.4812661498708,
      "grad_norm": 0.5444427728652954,
      "learning_rate": 0.00011286252861182173,
      "loss": 2.8641,
      "step": 120700
    },
    {
      "epoch": 32.494724375538325,
      "grad_norm": 0.5395984053611755,
      "learning_rate": 0.00011266056281136394,
      "loss": 2.86,
      "step": 120750
    },
    {
      "epoch": 32.50818260120586,
      "grad_norm": 0.5177767872810364,
      "learning_rate": 0.00011245859701090614,
      "loss": 2.8679,
      "step": 120800
    },
    {
      "epoch": 32.52164082687339,
      "grad_norm": 0.5380182266235352,
      "learning_rate": 0.00011225663121044835,
      "loss": 2.8627,
      "step": 120850
    },
    {
      "epoch": 32.535099052540914,
      "grad_norm": 0.5435357093811035,
      "learning_rate": 0.00011205466540999056,
      "loss": 2.8756,
      "step": 120900
    },
    {
      "epoch": 32.54855727820844,
      "grad_norm": 0.5082008838653564,
      "learning_rate": 0.00011185269960953278,
      "loss": 2.869,
      "step": 120950
    },
    {
      "epoch": 32.56201550387597,
      "grad_norm": 0.5337185263633728,
      "learning_rate": 0.00011165073380907499,
      "loss": 2.8754,
      "step": 121000
    },
    {
      "epoch": 32.56201550387597,
      "eval_accuracy": 0.39746159659596364,
      "eval_loss": 3.2868685722351074,
      "eval_runtime": 186.5501,
      "eval_samples_per_second": 96.553,
      "eval_steps_per_second": 6.036,
      "step": 121000
    },
    {
      "epoch": 32.575473729543496,
      "grad_norm": 0.5145751237869263,
      "learning_rate": 0.0001114487680086172,
      "loss": 2.8751,
      "step": 121050
    },
    {
      "epoch": 32.588931955211024,
      "grad_norm": 0.5249476432800293,
      "learning_rate": 0.0001112468022081594,
      "loss": 2.8711,
      "step": 121100
    },
    {
      "epoch": 32.60239018087855,
      "grad_norm": 0.5283523201942444,
      "learning_rate": 0.00011104483640770163,
      "loss": 2.8686,
      "step": 121150
    },
    {
      "epoch": 32.61584840654608,
      "grad_norm": 0.5112684965133667,
      "learning_rate": 0.00011084287060724383,
      "loss": 2.8617,
      "step": 121200
    },
    {
      "epoch": 32.629306632213606,
      "grad_norm": 0.5741111636161804,
      "learning_rate": 0.00011064090480678604,
      "loss": 2.8731,
      "step": 121250
    },
    {
      "epoch": 32.64276485788114,
      "grad_norm": 0.5316374897956848,
      "learning_rate": 0.00011043893900632825,
      "loss": 2.875,
      "step": 121300
    },
    {
      "epoch": 32.65622308354867,
      "grad_norm": 0.5303090214729309,
      "learning_rate": 0.00011023697320587046,
      "loss": 2.8673,
      "step": 121350
    },
    {
      "epoch": 32.669681309216195,
      "grad_norm": 0.5476655960083008,
      "learning_rate": 0.00011003500740541268,
      "loss": 2.8739,
      "step": 121400
    },
    {
      "epoch": 32.68313953488372,
      "grad_norm": 0.5116368532180786,
      "learning_rate": 0.00010983304160495489,
      "loss": 2.865,
      "step": 121450
    },
    {
      "epoch": 32.69659776055125,
      "grad_norm": 0.5238583087921143,
      "learning_rate": 0.00010963107580449709,
      "loss": 2.8644,
      "step": 121500
    },
    {
      "epoch": 32.71005598621878,
      "grad_norm": 0.5408053994178772,
      "learning_rate": 0.0001094291100040393,
      "loss": 2.8771,
      "step": 121550
    },
    {
      "epoch": 32.723514211886304,
      "grad_norm": 0.535257875919342,
      "learning_rate": 0.00010922714420358152,
      "loss": 2.8728,
      "step": 121600
    },
    {
      "epoch": 32.73697243755383,
      "grad_norm": 0.533491313457489,
      "learning_rate": 0.00010902517840312373,
      "loss": 2.8837,
      "step": 121650
    },
    {
      "epoch": 32.75043066322136,
      "grad_norm": 0.5391260981559753,
      "learning_rate": 0.00010882321260266594,
      "loss": 2.8816,
      "step": 121700
    },
    {
      "epoch": 32.763888888888886,
      "grad_norm": 0.5323964357376099,
      "learning_rate": 0.00010862124680220815,
      "loss": 2.8742,
      "step": 121750
    },
    {
      "epoch": 32.77734711455642,
      "grad_norm": 0.5182379484176636,
      "learning_rate": 0.00010841928100175037,
      "loss": 2.8804,
      "step": 121800
    },
    {
      "epoch": 32.79080534022395,
      "grad_norm": 0.536156177520752,
      "learning_rate": 0.00010821731520129258,
      "loss": 2.8807,
      "step": 121850
    },
    {
      "epoch": 32.804263565891475,
      "grad_norm": 0.5116010308265686,
      "learning_rate": 0.00010801534940083478,
      "loss": 2.8769,
      "step": 121900
    },
    {
      "epoch": 32.817721791559,
      "grad_norm": 0.5143781304359436,
      "learning_rate": 0.00010781338360037699,
      "loss": 2.8861,
      "step": 121950
    },
    {
      "epoch": 32.83118001722653,
      "grad_norm": 0.5126392245292664,
      "learning_rate": 0.0001076114177999192,
      "loss": 2.889,
      "step": 122000
    },
    {
      "epoch": 32.83118001722653,
      "eval_accuracy": 0.39802036737413155,
      "eval_loss": 3.2823173999786377,
      "eval_runtime": 186.3391,
      "eval_samples_per_second": 96.663,
      "eval_steps_per_second": 6.043,
      "step": 122000
    },
    {
      "epoch": 32.84463824289406,
      "grad_norm": 0.5267850756645203,
      "learning_rate": 0.00010740945199946142,
      "loss": 2.8867,
      "step": 122050
    },
    {
      "epoch": 32.858096468561584,
      "grad_norm": 0.5378038287162781,
      "learning_rate": 0.00010720748619900363,
      "loss": 2.8817,
      "step": 122100
    },
    {
      "epoch": 32.87155469422911,
      "grad_norm": 0.559219479560852,
      "learning_rate": 0.00010700552039854584,
      "loss": 2.884,
      "step": 122150
    },
    {
      "epoch": 32.88501291989664,
      "grad_norm": 0.5389184951782227,
      "learning_rate": 0.00010680355459808805,
      "loss": 2.8876,
      "step": 122200
    },
    {
      "epoch": 32.898471145564166,
      "grad_norm": 0.5781436562538147,
      "learning_rate": 0.00010660158879763027,
      "loss": 2.8824,
      "step": 122250
    },
    {
      "epoch": 32.9119293712317,
      "grad_norm": 0.5194249749183655,
      "learning_rate": 0.00010639962299717247,
      "loss": 2.8726,
      "step": 122300
    },
    {
      "epoch": 32.92538759689923,
      "grad_norm": 0.5154020190238953,
      "learning_rate": 0.00010619765719671468,
      "loss": 2.8828,
      "step": 122350
    },
    {
      "epoch": 32.938845822566755,
      "grad_norm": 0.5082391500473022,
      "learning_rate": 0.00010599569139625689,
      "loss": 2.8871,
      "step": 122400
    },
    {
      "epoch": 32.95230404823428,
      "grad_norm": 0.5316715240478516,
      "learning_rate": 0.0001057937255957991,
      "loss": 2.894,
      "step": 122450
    },
    {
      "epoch": 32.96576227390181,
      "grad_norm": 0.5585044026374817,
      "learning_rate": 0.00010559175979534132,
      "loss": 2.8915,
      "step": 122500
    },
    {
      "epoch": 32.97922049956934,
      "grad_norm": 0.5465138554573059,
      "learning_rate": 0.00010538979399488353,
      "loss": 2.8973,
      "step": 122550
    },
    {
      "epoch": 32.992678725236864,
      "grad_norm": 0.5389026999473572,
      "learning_rate": 0.00010518782819442573,
      "loss": 2.8928,
      "step": 122600
    },
    {
      "epoch": 33.006190783807064,
      "grad_norm": 0.5233954191207886,
      "learning_rate": 0.00010498586239396794,
      "loss": 2.9347,
      "step": 122650
    },
    {
      "epoch": 33.01964900947459,
      "grad_norm": 0.5135116577148438,
      "learning_rate": 0.00010478389659351016,
      "loss": 2.8385,
      "step": 122700
    },
    {
      "epoch": 33.03310723514212,
      "grad_norm": 0.5033681988716125,
      "learning_rate": 0.00010458193079305237,
      "loss": 2.8538,
      "step": 122750
    },
    {
      "epoch": 33.046565460809646,
      "grad_norm": 0.5341190695762634,
      "learning_rate": 0.00010437996499259458,
      "loss": 2.8325,
      "step": 122800
    },
    {
      "epoch": 33.06002368647717,
      "grad_norm": 0.546575129032135,
      "learning_rate": 0.00010417799919213679,
      "loss": 2.8402,
      "step": 122850
    },
    {
      "epoch": 33.0734819121447,
      "grad_norm": 0.5185466408729553,
      "learning_rate": 0.000103976033391679,
      "loss": 2.8568,
      "step": 122900
    },
    {
      "epoch": 33.08694013781223,
      "grad_norm": 0.5232506394386292,
      "learning_rate": 0.00010377406759122122,
      "loss": 2.8471,
      "step": 122950
    },
    {
      "epoch": 33.10039836347976,
      "grad_norm": 0.4987465739250183,
      "learning_rate": 0.00010357210179076342,
      "loss": 2.8458,
      "step": 123000
    },
    {
      "epoch": 33.10039836347976,
      "eval_accuracy": 0.3978827118081314,
      "eval_loss": 3.2873218059539795,
      "eval_runtime": 184.8755,
      "eval_samples_per_second": 97.428,
      "eval_steps_per_second": 6.091,
      "step": 123000
    },
    {
      "epoch": 33.11385658914729,
      "grad_norm": 0.5143473744392395,
      "learning_rate": 0.00010337013599030563,
      "loss": 2.8395,
      "step": 123050
    },
    {
      "epoch": 33.12731481481482,
      "grad_norm": 0.5328736305236816,
      "learning_rate": 0.00010316817018984784,
      "loss": 2.8439,
      "step": 123100
    },
    {
      "epoch": 33.140773040482344,
      "grad_norm": 0.5229068398475647,
      "learning_rate": 0.00010296620438939006,
      "loss": 2.8693,
      "step": 123150
    },
    {
      "epoch": 33.15423126614987,
      "grad_norm": 0.5527721047401428,
      "learning_rate": 0.00010276423858893227,
      "loss": 2.852,
      "step": 123200
    },
    {
      "epoch": 33.1676894918174,
      "grad_norm": 0.5370103716850281,
      "learning_rate": 0.00010256227278847448,
      "loss": 2.868,
      "step": 123250
    },
    {
      "epoch": 33.181147717484926,
      "grad_norm": 0.525336503982544,
      "learning_rate": 0.00010236030698801669,
      "loss": 2.8642,
      "step": 123300
    },
    {
      "epoch": 33.19460594315245,
      "grad_norm": 0.530510663986206,
      "learning_rate": 0.00010215834118755888,
      "loss": 2.854,
      "step": 123350
    },
    {
      "epoch": 33.20806416881998,
      "grad_norm": 0.5082398653030396,
      "learning_rate": 0.00010195637538710111,
      "loss": 2.8547,
      "step": 123400
    },
    {
      "epoch": 33.22152239448751,
      "grad_norm": 0.5090308785438538,
      "learning_rate": 0.00010175440958664332,
      "loss": 2.8586,
      "step": 123450
    },
    {
      "epoch": 33.23498062015504,
      "grad_norm": 0.5598142147064209,
      "learning_rate": 0.00010155244378618553,
      "loss": 2.8538,
      "step": 123500
    },
    {
      "epoch": 33.24843884582257,
      "grad_norm": 0.5287654399871826,
      "learning_rate": 0.00010135047798572774,
      "loss": 2.8616,
      "step": 123550
    },
    {
      "epoch": 33.2618970714901,
      "grad_norm": 0.5196240544319153,
      "learning_rate": 0.00010114851218526996,
      "loss": 2.8648,
      "step": 123600
    },
    {
      "epoch": 33.275355297157624,
      "grad_norm": 0.517903208732605,
      "learning_rate": 0.00010094654638481217,
      "loss": 2.8583,
      "step": 123650
    },
    {
      "epoch": 33.28881352282515,
      "grad_norm": 0.5230700969696045,
      "learning_rate": 0.00010074458058435438,
      "loss": 2.8642,
      "step": 123700
    },
    {
      "epoch": 33.30227174849268,
      "grad_norm": 0.5415684580802917,
      "learning_rate": 0.00010054261478389658,
      "loss": 2.8586,
      "step": 123750
    },
    {
      "epoch": 33.315729974160206,
      "grad_norm": 0.52119380235672,
      "learning_rate": 0.0001003406489834388,
      "loss": 2.8692,
      "step": 123800
    },
    {
      "epoch": 33.32918819982773,
      "grad_norm": 0.5184343457221985,
      "learning_rate": 0.00010013868318298101,
      "loss": 2.8756,
      "step": 123850
    },
    {
      "epoch": 33.34264642549526,
      "grad_norm": 0.5174550414085388,
      "learning_rate": 9.993671738252322e-05,
      "loss": 2.8626,
      "step": 123900
    },
    {
      "epoch": 33.35610465116279,
      "grad_norm": 0.5095716118812561,
      "learning_rate": 9.973475158206543e-05,
      "loss": 2.859,
      "step": 123950
    },
    {
      "epoch": 33.369562876830315,
      "grad_norm": 0.5363544821739197,
      "learning_rate": 9.953278578160764e-05,
      "loss": 2.8752,
      "step": 124000
    },
    {
      "epoch": 33.369562876830315,
      "eval_accuracy": 0.3979835360900952,
      "eval_loss": 3.2857799530029297,
      "eval_runtime": 185.7823,
      "eval_samples_per_second": 96.952,
      "eval_steps_per_second": 6.061,
      "step": 124000
    },
    {
      "epoch": 33.38302110249785,
      "grad_norm": 0.5235912799835205,
      "learning_rate": 9.933081998114986e-05,
      "loss": 2.8715,
      "step": 124050
    },
    {
      "epoch": 33.39647932816538,
      "grad_norm": 0.5178714990615845,
      "learning_rate": 9.912885418069207e-05,
      "loss": 2.8689,
      "step": 124100
    },
    {
      "epoch": 33.409937553832904,
      "grad_norm": 0.5316716432571411,
      "learning_rate": 9.892688838023427e-05,
      "loss": 2.8585,
      "step": 124150
    },
    {
      "epoch": 33.42339577950043,
      "grad_norm": 0.5281285047531128,
      "learning_rate": 9.872492257977648e-05,
      "loss": 2.8723,
      "step": 124200
    },
    {
      "epoch": 33.43685400516796,
      "grad_norm": 0.5472725033760071,
      "learning_rate": 9.85229567793187e-05,
      "loss": 2.87,
      "step": 124250
    },
    {
      "epoch": 33.450312230835486,
      "grad_norm": 0.49894216656684875,
      "learning_rate": 9.832099097886091e-05,
      "loss": 2.8793,
      "step": 124300
    },
    {
      "epoch": 33.46377045650301,
      "grad_norm": 0.572883129119873,
      "learning_rate": 9.811902517840312e-05,
      "loss": 2.8711,
      "step": 124350
    },
    {
      "epoch": 33.47722868217054,
      "grad_norm": 0.5105341076850891,
      "learning_rate": 9.791705937794533e-05,
      "loss": 2.8587,
      "step": 124400
    },
    {
      "epoch": 33.49068690783807,
      "grad_norm": 0.5502719879150391,
      "learning_rate": 9.771509357748752e-05,
      "loss": 2.8666,
      "step": 124450
    },
    {
      "epoch": 33.504145133505595,
      "grad_norm": 0.5212557911872864,
      "learning_rate": 9.751312777702975e-05,
      "loss": 2.8761,
      "step": 124500
    },
    {
      "epoch": 33.51760335917313,
      "grad_norm": 0.5381674766540527,
      "learning_rate": 9.731116197657196e-05,
      "loss": 2.8736,
      "step": 124550
    },
    {
      "epoch": 33.53106158484066,
      "grad_norm": 0.544306755065918,
      "learning_rate": 9.710919617611417e-05,
      "loss": 2.8795,
      "step": 124600
    },
    {
      "epoch": 33.544519810508184,
      "grad_norm": 0.5742482542991638,
      "learning_rate": 9.690723037565638e-05,
      "loss": 2.8771,
      "step": 124650
    },
    {
      "epoch": 33.55797803617571,
      "grad_norm": 0.5406849980354309,
      "learning_rate": 9.67052645751986e-05,
      "loss": 2.878,
      "step": 124700
    },
    {
      "epoch": 33.57143626184324,
      "grad_norm": 0.5390380024909973,
      "learning_rate": 9.650329877474081e-05,
      "loss": 2.879,
      "step": 124750
    },
    {
      "epoch": 33.584894487510766,
      "grad_norm": 0.511185884475708,
      "learning_rate": 9.630133297428302e-05,
      "loss": 2.8822,
      "step": 124800
    },
    {
      "epoch": 33.59835271317829,
      "grad_norm": 0.5297926664352417,
      "learning_rate": 9.609936717382521e-05,
      "loss": 2.8849,
      "step": 124850
    },
    {
      "epoch": 33.61181093884582,
      "grad_norm": 0.5385516285896301,
      "learning_rate": 9.589740137336742e-05,
      "loss": 2.8806,
      "step": 124900
    },
    {
      "epoch": 33.62526916451335,
      "grad_norm": 0.5299723148345947,
      "learning_rate": 9.569543557290965e-05,
      "loss": 2.8761,
      "step": 124950
    },
    {
      "epoch": 33.638727390180875,
      "grad_norm": 0.5565329194068909,
      "learning_rate": 9.549346977245186e-05,
      "loss": 2.8878,
      "step": 125000
    },
    {
      "epoch": 33.638727390180875,
      "eval_accuracy": 0.39817518914331085,
      "eval_loss": 3.282923698425293,
      "eval_runtime": 185.7187,
      "eval_samples_per_second": 96.985,
      "eval_steps_per_second": 6.063,
      "step": 125000
    },
    {
      "epoch": 33.65218561584841,
      "grad_norm": 0.5189403891563416,
      "learning_rate": 9.529150397199407e-05,
      "loss": 2.8695,
      "step": 125050
    },
    {
      "epoch": 33.66564384151594,
      "grad_norm": 0.5315220952033997,
      "learning_rate": 9.508953817153628e-05,
      "loss": 2.884,
      "step": 125100
    },
    {
      "epoch": 33.679102067183464,
      "grad_norm": 0.5463157892227173,
      "learning_rate": 9.48875723710785e-05,
      "loss": 2.8829,
      "step": 125150
    },
    {
      "epoch": 33.69256029285099,
      "grad_norm": 0.5383089184761047,
      "learning_rate": 9.468560657062071e-05,
      "loss": 2.8782,
      "step": 125200
    },
    {
      "epoch": 33.70601851851852,
      "grad_norm": 0.5510347485542297,
      "learning_rate": 9.44836407701629e-05,
      "loss": 2.8732,
      "step": 125250
    },
    {
      "epoch": 33.719476744186046,
      "grad_norm": 0.534978985786438,
      "learning_rate": 9.428167496970512e-05,
      "loss": 2.8869,
      "step": 125300
    },
    {
      "epoch": 33.732934969853574,
      "grad_norm": 0.5539801716804504,
      "learning_rate": 9.407970916924732e-05,
      "loss": 2.8891,
      "step": 125350
    },
    {
      "epoch": 33.7463931955211,
      "grad_norm": 0.520168125629425,
      "learning_rate": 9.387774336878955e-05,
      "loss": 2.8779,
      "step": 125400
    },
    {
      "epoch": 33.75985142118863,
      "grad_norm": 0.5352253317832947,
      "learning_rate": 9.367577756833176e-05,
      "loss": 2.8777,
      "step": 125450
    },
    {
      "epoch": 33.773309646856156,
      "grad_norm": 0.5400208234786987,
      "learning_rate": 9.347381176787397e-05,
      "loss": 2.8739,
      "step": 125500
    },
    {
      "epoch": 33.78676787252369,
      "grad_norm": 0.5117079615592957,
      "learning_rate": 9.327184596741616e-05,
      "loss": 2.8682,
      "step": 125550
    },
    {
      "epoch": 33.80022609819122,
      "grad_norm": 0.5133141875267029,
      "learning_rate": 9.30698801669584e-05,
      "loss": 2.8912,
      "step": 125600
    },
    {
      "epoch": 33.813684323858745,
      "grad_norm": 0.5294487476348877,
      "learning_rate": 9.28679143665006e-05,
      "loss": 2.8867,
      "step": 125650
    },
    {
      "epoch": 33.82714254952627,
      "grad_norm": 0.5136508345603943,
      "learning_rate": 9.26659485660428e-05,
      "loss": 2.8757,
      "step": 125700
    },
    {
      "epoch": 33.8406007751938,
      "grad_norm": 0.5041595697402954,
      "learning_rate": 9.246398276558502e-05,
      "loss": 2.8892,
      "step": 125750
    },
    {
      "epoch": 33.85405900086133,
      "grad_norm": 0.5109428763389587,
      "learning_rate": 9.226201696512723e-05,
      "loss": 2.8785,
      "step": 125800
    },
    {
      "epoch": 33.867517226528854,
      "grad_norm": 0.5545689463615417,
      "learning_rate": 9.206005116466945e-05,
      "loss": 2.879,
      "step": 125850
    },
    {
      "epoch": 33.88097545219638,
      "grad_norm": 0.566541314125061,
      "learning_rate": 9.185808536421166e-05,
      "loss": 2.8762,
      "step": 125900
    },
    {
      "epoch": 33.89443367786391,
      "grad_norm": 0.5466289520263672,
      "learning_rate": 9.165611956375385e-05,
      "loss": 2.8979,
      "step": 125950
    },
    {
      "epoch": 33.907891903531436,
      "grad_norm": 0.545253574848175,
      "learning_rate": 9.145415376329606e-05,
      "loss": 2.8909,
      "step": 126000
    },
    {
      "epoch": 33.907891903531436,
      "eval_accuracy": 0.39849808759804833,
      "eval_loss": 3.2789785861968994,
      "eval_runtime": 185.5746,
      "eval_samples_per_second": 97.061,
      "eval_steps_per_second": 6.068,
      "step": 126000
    },
    {
      "epoch": 33.92135012919896,
      "grad_norm": 0.5617538690567017,
      "learning_rate": 9.125218796283829e-05,
      "loss": 2.8737,
      "step": 126050
    },
    {
      "epoch": 33.9348083548665,
      "grad_norm": 0.5632882714271545,
      "learning_rate": 9.10502221623805e-05,
      "loss": 2.8973,
      "step": 126100
    },
    {
      "epoch": 33.948266580534025,
      "grad_norm": 0.5469334721565247,
      "learning_rate": 9.08482563619227e-05,
      "loss": 2.8853,
      "step": 126150
    },
    {
      "epoch": 33.96172480620155,
      "grad_norm": 0.5518969893455505,
      "learning_rate": 9.064629056146492e-05,
      "loss": 2.8821,
      "step": 126200
    },
    {
      "epoch": 33.97518303186908,
      "grad_norm": 0.5617625713348389,
      "learning_rate": 9.044432476100714e-05,
      "loss": 2.8948,
      "step": 126250
    },
    {
      "epoch": 33.98864125753661,
      "grad_norm": 0.5358351469039917,
      "learning_rate": 9.024235896054935e-05,
      "loss": 2.8871,
      "step": 126300
    },
    {
      "epoch": 34.00188415159345,
      "grad_norm": 0.5151727199554443,
      "learning_rate": 9.004039316009154e-05,
      "loss": 2.8792,
      "step": 126350
    },
    {
      "epoch": 34.01534237726098,
      "grad_norm": 0.5237066745758057,
      "learning_rate": 8.983842735963375e-05,
      "loss": 2.8371,
      "step": 126400
    },
    {
      "epoch": 34.02880060292851,
      "grad_norm": 0.5409257411956787,
      "learning_rate": 8.963646155917596e-05,
      "loss": 2.8324,
      "step": 126450
    },
    {
      "epoch": 34.04225882859604,
      "grad_norm": 0.5065373182296753,
      "learning_rate": 8.943449575871819e-05,
      "loss": 2.8374,
      "step": 126500
    },
    {
      "epoch": 34.05571705426357,
      "grad_norm": 0.528031051158905,
      "learning_rate": 8.92325299582604e-05,
      "loss": 2.822,
      "step": 126550
    },
    {
      "epoch": 34.069175279931095,
      "grad_norm": 0.5282293558120728,
      "learning_rate": 8.90305641578026e-05,
      "loss": 2.8444,
      "step": 126600
    },
    {
      "epoch": 34.08263350559862,
      "grad_norm": 0.5434133410453796,
      "learning_rate": 8.882859835734482e-05,
      "loss": 2.8394,
      "step": 126650
    },
    {
      "epoch": 34.09609173126615,
      "grad_norm": 0.5360873937606812,
      "learning_rate": 8.862663255688704e-05,
      "loss": 2.8363,
      "step": 126700
    },
    {
      "epoch": 34.10954995693368,
      "grad_norm": 0.5426977276802063,
      "learning_rate": 8.842466675642924e-05,
      "loss": 2.8571,
      "step": 126750
    },
    {
      "epoch": 34.123008182601204,
      "grad_norm": 0.5531218647956848,
      "learning_rate": 8.822270095597145e-05,
      "loss": 2.8467,
      "step": 126800
    },
    {
      "epoch": 34.13646640826873,
      "grad_norm": 0.5302387475967407,
      "learning_rate": 8.802073515551365e-05,
      "loss": 2.8528,
      "step": 126850
    },
    {
      "epoch": 34.14992463393626,
      "grad_norm": 0.5007009506225586,
      "learning_rate": 8.781876935505586e-05,
      "loss": 2.8388,
      "step": 126900
    },
    {
      "epoch": 34.16338285960379,
      "grad_norm": 0.5326977968215942,
      "learning_rate": 8.761680355459809e-05,
      "loss": 2.836,
      "step": 126950
    },
    {
      "epoch": 34.17684108527132,
      "grad_norm": 0.5608571767807007,
      "learning_rate": 8.74148377541403e-05,
      "loss": 2.8487,
      "step": 127000
    },
    {
      "epoch": 34.17684108527132,
      "eval_accuracy": 0.3979312769525687,
      "eval_loss": 3.288566827774048,
      "eval_runtime": 185.4209,
      "eval_samples_per_second": 97.141,
      "eval_steps_per_second": 6.073,
      "step": 127000
    },
    {
      "epoch": 34.19029931093885,
      "grad_norm": 0.5379499793052673,
      "learning_rate": 8.721287195368251e-05,
      "loss": 2.8465,
      "step": 127050
    },
    {
      "epoch": 34.203757536606375,
      "grad_norm": 0.5316014289855957,
      "learning_rate": 8.70109061532247e-05,
      "loss": 2.8394,
      "step": 127100
    },
    {
      "epoch": 34.2172157622739,
      "grad_norm": 0.5188553929328918,
      "learning_rate": 8.680894035276693e-05,
      "loss": 2.848,
      "step": 127150
    },
    {
      "epoch": 34.23067398794143,
      "grad_norm": 0.5482759475708008,
      "learning_rate": 8.660697455230914e-05,
      "loss": 2.8465,
      "step": 127200
    },
    {
      "epoch": 34.24413221360896,
      "grad_norm": 0.5545268654823303,
      "learning_rate": 8.640500875185135e-05,
      "loss": 2.8657,
      "step": 127250
    },
    {
      "epoch": 34.257590439276484,
      "grad_norm": 0.5227739214897156,
      "learning_rate": 8.620304295139356e-05,
      "loss": 2.8452,
      "step": 127300
    },
    {
      "epoch": 34.27104866494401,
      "grad_norm": 0.5258796215057373,
      "learning_rate": 8.600107715093576e-05,
      "loss": 2.8596,
      "step": 127350
    },
    {
      "epoch": 34.28450689061154,
      "grad_norm": 0.5534136295318604,
      "learning_rate": 8.579911135047799e-05,
      "loss": 2.8433,
      "step": 127400
    },
    {
      "epoch": 34.29796511627907,
      "grad_norm": 0.5434840321540833,
      "learning_rate": 8.559714555002018e-05,
      "loss": 2.8582,
      "step": 127450
    },
    {
      "epoch": 34.3114233419466,
      "grad_norm": 0.5373007655143738,
      "learning_rate": 8.53951797495624e-05,
      "loss": 2.862,
      "step": 127500
    },
    {
      "epoch": 34.32488156761413,
      "grad_norm": 0.51643967628479,
      "learning_rate": 8.51932139491046e-05,
      "loss": 2.8568,
      "step": 127550
    },
    {
      "epoch": 34.338339793281655,
      "grad_norm": 0.5431195497512817,
      "learning_rate": 8.499124814864683e-05,
      "loss": 2.8548,
      "step": 127600
    },
    {
      "epoch": 34.35179801894918,
      "grad_norm": 0.5500590801239014,
      "learning_rate": 8.478928234818904e-05,
      "loss": 2.8533,
      "step": 127650
    },
    {
      "epoch": 34.36525624461671,
      "grad_norm": 0.5351163148880005,
      "learning_rate": 8.458731654773125e-05,
      "loss": 2.8513,
      "step": 127700
    },
    {
      "epoch": 34.37871447028424,
      "grad_norm": 0.5352818369865417,
      "learning_rate": 8.438535074727346e-05,
      "loss": 2.8694,
      "step": 127750
    },
    {
      "epoch": 34.392172695951764,
      "grad_norm": 0.5493792295455933,
      "learning_rate": 8.418338494681565e-05,
      "loss": 2.8605,
      "step": 127800
    },
    {
      "epoch": 34.40563092161929,
      "grad_norm": 0.5486342310905457,
      "learning_rate": 8.398141914635787e-05,
      "loss": 2.855,
      "step": 127850
    },
    {
      "epoch": 34.41908914728682,
      "grad_norm": 0.5304096937179565,
      "learning_rate": 8.377945334590008e-05,
      "loss": 2.8568,
      "step": 127900
    },
    {
      "epoch": 34.43254737295435,
      "grad_norm": 0.5243949890136719,
      "learning_rate": 8.35774875454423e-05,
      "loss": 2.8686,
      "step": 127950
    },
    {
      "epoch": 34.44600559862188,
      "grad_norm": 0.5273032188415527,
      "learning_rate": 8.33755217449845e-05,
      "loss": 2.8584,
      "step": 128000
    },
    {
      "epoch": 34.44600559862188,
      "eval_accuracy": 0.3982789468903749,
      "eval_loss": 3.2838759422302246,
      "eval_runtime": 185.3112,
      "eval_samples_per_second": 97.199,
      "eval_steps_per_second": 6.076,
      "step": 128000
    },
    {
      "epoch": 34.45946382428941,
      "grad_norm": 0.5518338084220886,
      "learning_rate": 8.317355594452673e-05,
      "loss": 2.861,
      "step": 128050
    },
    {
      "epoch": 34.472922049956935,
      "grad_norm": 0.5366113185882568,
      "learning_rate": 8.297159014406894e-05,
      "loss": 2.8608,
      "step": 128100
    },
    {
      "epoch": 34.48638027562446,
      "grad_norm": 0.5235980153083801,
      "learning_rate": 8.276962434361115e-05,
      "loss": 2.8558,
      "step": 128150
    },
    {
      "epoch": 34.49983850129199,
      "grad_norm": 0.557013213634491,
      "learning_rate": 8.256765854315334e-05,
      "loss": 2.8723,
      "step": 128200
    },
    {
      "epoch": 34.51329672695952,
      "grad_norm": 0.5463767647743225,
      "learning_rate": 8.236569274269555e-05,
      "loss": 2.862,
      "step": 128250
    },
    {
      "epoch": 34.526754952627044,
      "grad_norm": 0.5397362112998962,
      "learning_rate": 8.216372694223778e-05,
      "loss": 2.8606,
      "step": 128300
    },
    {
      "epoch": 34.54021317829457,
      "grad_norm": 0.530078113079071,
      "learning_rate": 8.196176114177998e-05,
      "loss": 2.8632,
      "step": 128350
    },
    {
      "epoch": 34.5536714039621,
      "grad_norm": 0.5380610227584839,
      "learning_rate": 8.17597953413222e-05,
      "loss": 2.8611,
      "step": 128400
    },
    {
      "epoch": 34.567129629629626,
      "grad_norm": 0.5572210550308228,
      "learning_rate": 8.15578295408644e-05,
      "loss": 2.8626,
      "step": 128450
    },
    {
      "epoch": 34.58058785529716,
      "grad_norm": 0.5364845395088196,
      "learning_rate": 8.135586374040663e-05,
      "loss": 2.8647,
      "step": 128500
    },
    {
      "epoch": 34.59404608096469,
      "grad_norm": 0.5398139953613281,
      "learning_rate": 8.115389793994884e-05,
      "loss": 2.8655,
      "step": 128550
    },
    {
      "epoch": 34.607504306632215,
      "grad_norm": 0.5546800494194031,
      "learning_rate": 8.095193213949103e-05,
      "loss": 2.8695,
      "step": 128600
    },
    {
      "epoch": 34.62096253229974,
      "grad_norm": 0.5232498645782471,
      "learning_rate": 8.074996633903324e-05,
      "loss": 2.8647,
      "step": 128650
    },
    {
      "epoch": 34.63442075796727,
      "grad_norm": 0.5261814594268799,
      "learning_rate": 8.054800053857547e-05,
      "loss": 2.8669,
      "step": 128700
    },
    {
      "epoch": 34.6478789836348,
      "grad_norm": 0.522723376750946,
      "learning_rate": 8.034603473811768e-05,
      "loss": 2.867,
      "step": 128750
    },
    {
      "epoch": 34.661337209302324,
      "grad_norm": 0.5273510813713074,
      "learning_rate": 8.014406893765989e-05,
      "loss": 2.8667,
      "step": 128800
    },
    {
      "epoch": 34.67479543496985,
      "grad_norm": 0.5781503915786743,
      "learning_rate": 7.99421031372021e-05,
      "loss": 2.8696,
      "step": 128850
    },
    {
      "epoch": 34.68825366063738,
      "grad_norm": 0.5295001268386841,
      "learning_rate": 7.974013733674429e-05,
      "loss": 2.8723,
      "step": 128900
    },
    {
      "epoch": 34.701711886304906,
      "grad_norm": 0.5671932101249695,
      "learning_rate": 7.953817153628653e-05,
      "loss": 2.8732,
      "step": 128950
    },
    {
      "epoch": 34.71517011197244,
      "grad_norm": 0.5496615171432495,
      "learning_rate": 7.933620573582872e-05,
      "loss": 2.8711,
      "step": 129000
    },
    {
      "epoch": 34.71517011197244,
      "eval_accuracy": 0.3987073414418655,
      "eval_loss": 3.2790117263793945,
      "eval_runtime": 185.2229,
      "eval_samples_per_second": 97.245,
      "eval_steps_per_second": 6.079,
      "step": 129000
    },
    {
      "epoch": 34.72862833763997,
      "grad_norm": 0.534195065498352,
      "learning_rate": 7.913423993537093e-05,
      "loss": 2.8648,
      "step": 129050
    },
    {
      "epoch": 34.742086563307495,
      "grad_norm": 0.5406900644302368,
      "learning_rate": 7.893227413491314e-05,
      "loss": 2.8747,
      "step": 129100
    },
    {
      "epoch": 34.75554478897502,
      "grad_norm": 0.536567747592926,
      "learning_rate": 7.873030833445537e-05,
      "loss": 2.8731,
      "step": 129150
    },
    {
      "epoch": 34.76900301464255,
      "grad_norm": 0.5142216682434082,
      "learning_rate": 7.852834253399758e-05,
      "loss": 2.8666,
      "step": 129200
    },
    {
      "epoch": 34.78246124031008,
      "grad_norm": 0.540355384349823,
      "learning_rate": 7.832637673353979e-05,
      "loss": 2.8639,
      "step": 129250
    },
    {
      "epoch": 34.795919465977605,
      "grad_norm": 0.5532994866371155,
      "learning_rate": 7.812441093308198e-05,
      "loss": 2.8721,
      "step": 129300
    },
    {
      "epoch": 34.80937769164513,
      "grad_norm": 0.5396466255187988,
      "learning_rate": 7.792244513262419e-05,
      "loss": 2.8635,
      "step": 129350
    },
    {
      "epoch": 34.82283591731266,
      "grad_norm": 0.5356307625770569,
      "learning_rate": 7.772047933216641e-05,
      "loss": 2.8618,
      "step": 129400
    },
    {
      "epoch": 34.83629414298019,
      "grad_norm": 0.5366500616073608,
      "learning_rate": 7.751851353170862e-05,
      "loss": 2.8724,
      "step": 129450
    },
    {
      "epoch": 34.84975236864772,
      "grad_norm": 0.5566040873527527,
      "learning_rate": 7.731654773125083e-05,
      "loss": 2.8804,
      "step": 129500
    },
    {
      "epoch": 34.86321059431525,
      "grad_norm": 0.5199340581893921,
      "learning_rate": 7.711458193079304e-05,
      "loss": 2.868,
      "step": 129550
    },
    {
      "epoch": 34.876668819982775,
      "grad_norm": 0.5459930300712585,
      "learning_rate": 7.691261613033527e-05,
      "loss": 2.8722,
      "step": 129600
    },
    {
      "epoch": 34.8901270456503,
      "grad_norm": 0.5455026626586914,
      "learning_rate": 7.671065032987748e-05,
      "loss": 2.8759,
      "step": 129650
    },
    {
      "epoch": 34.90358527131783,
      "grad_norm": 0.5485450029373169,
      "learning_rate": 7.650868452941967e-05,
      "loss": 2.88,
      "step": 129700
    },
    {
      "epoch": 34.91704349698536,
      "grad_norm": 0.5579107403755188,
      "learning_rate": 7.630671872896188e-05,
      "loss": 2.8716,
      "step": 129750
    },
    {
      "epoch": 34.930501722652885,
      "grad_norm": 0.524882972240448,
      "learning_rate": 7.610475292850409e-05,
      "loss": 2.8661,
      "step": 129800
    },
    {
      "epoch": 34.94395994832041,
      "grad_norm": 0.5295476317405701,
      "learning_rate": 7.590278712804632e-05,
      "loss": 2.8774,
      "step": 129850
    },
    {
      "epoch": 34.95741817398794,
      "grad_norm": 0.5301741361618042,
      "learning_rate": 7.570082132758852e-05,
      "loss": 2.8824,
      "step": 129900
    },
    {
      "epoch": 34.97087639965547,
      "grad_norm": 0.5374778509140015,
      "learning_rate": 7.549885552713073e-05,
      "loss": 2.8632,
      "step": 129950
    },
    {
      "epoch": 34.984334625323,
      "grad_norm": 0.5610101819038391,
      "learning_rate": 7.529688972667294e-05,
      "loss": 2.8672,
      "step": 130000
    },
    {
      "epoch": 34.984334625323,
      "eval_accuracy": 0.39880936083924046,
      "eval_loss": 3.2768795490264893,
      "eval_runtime": 185.2906,
      "eval_samples_per_second": 97.209,
      "eval_steps_per_second": 6.077,
      "step": 130000
    },
    {
      "epoch": 34.99779285099053,
      "grad_norm": 0.5390916466712952,
      "learning_rate": 7.509492392621517e-05,
      "loss": 2.8797,
      "step": 130050
    },
    {
      "epoch": 35.01103574504737,
      "grad_norm": 0.519458532333374,
      "learning_rate": 7.489295812575736e-05,
      "loss": 2.843,
      "step": 130100
    },
    {
      "epoch": 35.0244939707149,
      "grad_norm": 0.5529184937477112,
      "learning_rate": 7.469099232529957e-05,
      "loss": 2.8206,
      "step": 130150
    },
    {
      "epoch": 35.03795219638243,
      "grad_norm": 0.5352154970169067,
      "learning_rate": 7.448902652484178e-05,
      "loss": 2.8268,
      "step": 130200
    },
    {
      "epoch": 35.051410422049955,
      "grad_norm": 0.550682008266449,
      "learning_rate": 7.428706072438399e-05,
      "loss": 2.8395,
      "step": 130250
    },
    {
      "epoch": 35.06486864771748,
      "grad_norm": 0.5407971739768982,
      "learning_rate": 7.40850949239262e-05,
      "loss": 2.8207,
      "step": 130300
    },
    {
      "epoch": 35.078326873385016,
      "grad_norm": 0.5525979399681091,
      "learning_rate": 7.388312912346843e-05,
      "loss": 2.8337,
      "step": 130350
    },
    {
      "epoch": 35.091785099052544,
      "grad_norm": 0.5697442889213562,
      "learning_rate": 7.368116332301062e-05,
      "loss": 2.8242,
      "step": 130400
    },
    {
      "epoch": 35.10524332472007,
      "grad_norm": 0.5499669909477234,
      "learning_rate": 7.347919752255284e-05,
      "loss": 2.8311,
      "step": 130450
    },
    {
      "epoch": 35.1187015503876,
      "grad_norm": 0.5380462408065796,
      "learning_rate": 7.327723172209505e-05,
      "loss": 2.8203,
      "step": 130500
    },
    {
      "epoch": 35.132159776055126,
      "grad_norm": 0.5371805429458618,
      "learning_rate": 7.307526592163726e-05,
      "loss": 2.84,
      "step": 130550
    },
    {
      "epoch": 35.14561800172265,
      "grad_norm": 0.5366227030754089,
      "learning_rate": 7.287330012117947e-05,
      "loss": 2.8302,
      "step": 130600
    },
    {
      "epoch": 35.15907622739018,
      "grad_norm": 0.5230523347854614,
      "learning_rate": 7.267133432072168e-05,
      "loss": 2.8365,
      "step": 130650
    },
    {
      "epoch": 35.17253445305771,
      "grad_norm": 0.5528269410133362,
      "learning_rate": 7.246936852026389e-05,
      "loss": 2.8478,
      "step": 130700
    },
    {
      "epoch": 35.185992678725235,
      "grad_norm": 0.5290082097053528,
      "learning_rate": 7.22674027198061e-05,
      "loss": 2.8474,
      "step": 130750
    },
    {
      "epoch": 35.19945090439276,
      "grad_norm": 0.5147569179534912,
      "learning_rate": 7.206543691934831e-05,
      "loss": 2.8417,
      "step": 130800
    },
    {
      "epoch": 35.21290913006029,
      "grad_norm": 0.5535443425178528,
      "learning_rate": 7.186347111889052e-05,
      "loss": 2.8383,
      "step": 130850
    },
    {
      "epoch": 35.226367355727824,
      "grad_norm": 0.5383947491645813,
      "learning_rate": 7.166150531843274e-05,
      "loss": 2.8467,
      "step": 130900
    },
    {
      "epoch": 35.23982558139535,
      "grad_norm": 0.5462708473205566,
      "learning_rate": 7.145953951797495e-05,
      "loss": 2.8381,
      "step": 130950
    },
    {
      "epoch": 35.25328380706288,
      "grad_norm": 0.5454403758049011,
      "learning_rate": 7.125757371751716e-05,
      "loss": 2.8497,
      "step": 131000
    },
    {
      "epoch": 35.25328380706288,
      "eval_accuracy": 0.39848255109770264,
      "eval_loss": 3.2834692001342773,
      "eval_runtime": 185.1989,
      "eval_samples_per_second": 97.258,
      "eval_steps_per_second": 6.08,
      "step": 131000
    },
    {
      "epoch": 35.266742032730406,
      "grad_norm": 0.5249785780906677,
      "learning_rate": 7.105560791705937e-05,
      "loss": 2.849,
      "step": 131050
    },
    {
      "epoch": 35.28020025839793,
      "grad_norm": 0.5564918518066406,
      "learning_rate": 7.085364211660158e-05,
      "loss": 2.8459,
      "step": 131100
    },
    {
      "epoch": 35.29365848406546,
      "grad_norm": 0.5393915176391602,
      "learning_rate": 7.065167631614379e-05,
      "loss": 2.8503,
      "step": 131150
    },
    {
      "epoch": 35.30711670973299,
      "grad_norm": 0.5674166083335876,
      "learning_rate": 7.0449710515686e-05,
      "loss": 2.8461,
      "step": 131200
    },
    {
      "epoch": 35.320574935400515,
      "grad_norm": 0.5445375442504883,
      "learning_rate": 7.024774471522821e-05,
      "loss": 2.8415,
      "step": 131250
    },
    {
      "epoch": 35.33403316106804,
      "grad_norm": 0.5473485589027405,
      "learning_rate": 7.004577891477042e-05,
      "loss": 2.8508,
      "step": 131300
    },
    {
      "epoch": 35.34749138673557,
      "grad_norm": 0.5318377614021301,
      "learning_rate": 6.984381311431263e-05,
      "loss": 2.8518,
      "step": 131350
    },
    {
      "epoch": 35.360949612403104,
      "grad_norm": 0.542001485824585,
      "learning_rate": 6.964184731385484e-05,
      "loss": 2.8553,
      "step": 131400
    },
    {
      "epoch": 35.37440783807063,
      "grad_norm": 0.5563853979110718,
      "learning_rate": 6.943988151339706e-05,
      "loss": 2.8542,
      "step": 131450
    },
    {
      "epoch": 35.38786606373816,
      "grad_norm": 0.5513464212417603,
      "learning_rate": 6.923791571293927e-05,
      "loss": 2.8616,
      "step": 131500
    },
    {
      "epoch": 35.401324289405686,
      "grad_norm": 0.5498881936073303,
      "learning_rate": 6.903594991248148e-05,
      "loss": 2.8538,
      "step": 131550
    },
    {
      "epoch": 35.41478251507321,
      "grad_norm": 0.5461167097091675,
      "learning_rate": 6.88339841120237e-05,
      "loss": 2.8503,
      "step": 131600
    },
    {
      "epoch": 35.42824074074074,
      "grad_norm": 0.5541663765907288,
      "learning_rate": 6.86320183115659e-05,
      "loss": 2.8513,
      "step": 131650
    },
    {
      "epoch": 35.44169896640827,
      "grad_norm": 0.5767017602920532,
      "learning_rate": 6.843005251110811e-05,
      "loss": 2.8524,
      "step": 131700
    },
    {
      "epoch": 35.455157192075795,
      "grad_norm": 0.5335713028907776,
      "learning_rate": 6.822808671065032e-05,
      "loss": 2.8446,
      "step": 131750
    },
    {
      "epoch": 35.46861541774332,
      "grad_norm": 0.5331796407699585,
      "learning_rate": 6.802612091019253e-05,
      "loss": 2.8525,
      "step": 131800
    },
    {
      "epoch": 35.48207364341085,
      "grad_norm": 0.5452083349227905,
      "learning_rate": 6.782415510973474e-05,
      "loss": 2.8561,
      "step": 131850
    },
    {
      "epoch": 35.495531869078384,
      "grad_norm": 0.538600504398346,
      "learning_rate": 6.762218930927696e-05,
      "loss": 2.8545,
      "step": 131900
    },
    {
      "epoch": 35.50899009474591,
      "grad_norm": 0.5454013347625732,
      "learning_rate": 6.742022350881916e-05,
      "loss": 2.8518,
      "step": 131950
    },
    {
      "epoch": 35.52244832041344,
      "grad_norm": 0.5553362965583801,
      "learning_rate": 6.721825770836138e-05,
      "loss": 2.8554,
      "step": 132000
    },
    {
      "epoch": 35.52244832041344,
      "eval_accuracy": 0.3986885455358528,
      "eval_loss": 3.2840888500213623,
      "eval_runtime": 185.2215,
      "eval_samples_per_second": 97.246,
      "eval_steps_per_second": 6.079,
      "step": 132000
    },
    {
      "epoch": 35.535906546080966,
      "grad_norm": 0.5484082102775574,
      "learning_rate": 6.70162919079036e-05,
      "loss": 2.856,
      "step": 132050
    },
    {
      "epoch": 35.54936477174849,
      "grad_norm": 0.5512910485267639,
      "learning_rate": 6.68143261074458e-05,
      "loss": 2.8579,
      "step": 132100
    },
    {
      "epoch": 35.56282299741602,
      "grad_norm": 0.5726379156112671,
      "learning_rate": 6.661236030698801e-05,
      "loss": 2.8596,
      "step": 132150
    },
    {
      "epoch": 35.57628122308355,
      "grad_norm": 0.5693790316581726,
      "learning_rate": 6.641039450653022e-05,
      "loss": 2.854,
      "step": 132200
    },
    {
      "epoch": 35.589739448751075,
      "grad_norm": 0.5748031139373779,
      "learning_rate": 6.620842870607243e-05,
      "loss": 2.8482,
      "step": 132250
    },
    {
      "epoch": 35.6031976744186,
      "grad_norm": 0.5548589825630188,
      "learning_rate": 6.600646290561464e-05,
      "loss": 2.8342,
      "step": 132300
    },
    {
      "epoch": 35.61665590008613,
      "grad_norm": 0.521531343460083,
      "learning_rate": 6.580449710515685e-05,
      "loss": 2.8522,
      "step": 132350
    },
    {
      "epoch": 35.630114125753664,
      "grad_norm": 0.5538045763969421,
      "learning_rate": 6.560253130469906e-05,
      "loss": 2.8566,
      "step": 132400
    },
    {
      "epoch": 35.64357235142119,
      "grad_norm": 0.5404136180877686,
      "learning_rate": 6.540056550424128e-05,
      "loss": 2.8622,
      "step": 132450
    },
    {
      "epoch": 35.65703057708872,
      "grad_norm": 0.5683730244636536,
      "learning_rate": 6.519859970378348e-05,
      "loss": 2.8548,
      "step": 132500
    },
    {
      "epoch": 35.670488802756246,
      "grad_norm": 0.5365909337997437,
      "learning_rate": 6.49966339033257e-05,
      "loss": 2.8428,
      "step": 132550
    },
    {
      "epoch": 35.68394702842377,
      "grad_norm": 0.5329694151878357,
      "learning_rate": 6.479466810286791e-05,
      "loss": 2.8479,
      "step": 132600
    },
    {
      "epoch": 35.6974052540913,
      "grad_norm": 0.5340180397033691,
      "learning_rate": 6.459270230241012e-05,
      "loss": 2.8607,
      "step": 132650
    },
    {
      "epoch": 35.71086347975883,
      "grad_norm": 0.5344703793525696,
      "learning_rate": 6.439073650195233e-05,
      "loss": 2.8592,
      "step": 132700
    },
    {
      "epoch": 35.724321705426355,
      "grad_norm": 0.572593629360199,
      "learning_rate": 6.418877070149454e-05,
      "loss": 2.8562,
      "step": 132750
    },
    {
      "epoch": 35.73777993109388,
      "grad_norm": 0.5305546522140503,
      "learning_rate": 6.398680490103675e-05,
      "loss": 2.8578,
      "step": 132800
    },
    {
      "epoch": 35.75123815676141,
      "grad_norm": 0.5552554726600647,
      "learning_rate": 6.378483910057896e-05,
      "loss": 2.848,
      "step": 132850
    },
    {
      "epoch": 35.76469638242894,
      "grad_norm": 0.5427165627479553,
      "learning_rate": 6.358287330012117e-05,
      "loss": 2.857,
      "step": 132900
    },
    {
      "epoch": 35.77815460809647,
      "grad_norm": 0.55644291639328,
      "learning_rate": 6.338090749966338e-05,
      "loss": 2.8574,
      "step": 132950
    },
    {
      "epoch": 35.791612833764,
      "grad_norm": 0.5329557061195374,
      "learning_rate": 6.31789416992056e-05,
      "loss": 2.8521,
      "step": 133000
    },
    {
      "epoch": 35.791612833764,
      "eval_accuracy": 0.3988379349622539,
      "eval_loss": 3.2778732776641846,
      "eval_runtime": 185.1178,
      "eval_samples_per_second": 97.3,
      "eval_steps_per_second": 6.083,
      "step": 133000
    },
    {
      "epoch": 35.805071059431526,
      "grad_norm": 0.5707139372825623,
      "learning_rate": 6.29769758987478e-05,
      "loss": 2.8582,
      "step": 133050
    },
    {
      "epoch": 35.81852928509905,
      "grad_norm": 0.5158944725990295,
      "learning_rate": 6.277501009829002e-05,
      "loss": 2.851,
      "step": 133100
    },
    {
      "epoch": 35.83198751076658,
      "grad_norm": 0.5252084732055664,
      "learning_rate": 6.257304429783223e-05,
      "loss": 2.8645,
      "step": 133150
    },
    {
      "epoch": 35.84544573643411,
      "grad_norm": 0.5452003479003906,
      "learning_rate": 6.237107849737443e-05,
      "loss": 2.855,
      "step": 133200
    },
    {
      "epoch": 35.858903962101635,
      "grad_norm": 0.5610634684562683,
      "learning_rate": 6.216911269691665e-05,
      "loss": 2.8595,
      "step": 133250
    },
    {
      "epoch": 35.87236218776916,
      "grad_norm": 0.571576714515686,
      "learning_rate": 6.196714689645886e-05,
      "loss": 2.8608,
      "step": 133300
    },
    {
      "epoch": 35.88582041343669,
      "grad_norm": 0.5424999594688416,
      "learning_rate": 6.176518109600107e-05,
      "loss": 2.8542,
      "step": 133350
    },
    {
      "epoch": 35.89927863910422,
      "grad_norm": 0.5375037789344788,
      "learning_rate": 6.156321529554328e-05,
      "loss": 2.8602,
      "step": 133400
    },
    {
      "epoch": 35.91273686477175,
      "grad_norm": 0.53477543592453,
      "learning_rate": 6.136124949508549e-05,
      "loss": 2.8473,
      "step": 133450
    },
    {
      "epoch": 35.92619509043928,
      "grad_norm": 0.5400429368019104,
      "learning_rate": 6.11592836946277e-05,
      "loss": 2.867,
      "step": 133500
    },
    {
      "epoch": 35.939653316106806,
      "grad_norm": 0.5509105920791626,
      "learning_rate": 6.095731789416992e-05,
      "loss": 2.8671,
      "step": 133550
    },
    {
      "epoch": 35.953111541774334,
      "grad_norm": 0.5655390024185181,
      "learning_rate": 6.075535209371213e-05,
      "loss": 2.867,
      "step": 133600
    },
    {
      "epoch": 35.96656976744186,
      "grad_norm": 0.5542401671409607,
      "learning_rate": 6.055338629325433e-05,
      "loss": 2.8567,
      "step": 133650
    },
    {
      "epoch": 35.98002799310939,
      "grad_norm": 0.5564590692520142,
      "learning_rate": 6.0351420492796546e-05,
      "loss": 2.8599,
      "step": 133700
    },
    {
      "epoch": 35.993486218776916,
      "grad_norm": 0.5613747835159302,
      "learning_rate": 6.0149454692338756e-05,
      "loss": 2.8563,
      "step": 133750
    },
    {
      "epoch": 36.00672911283377,
      "grad_norm": 0.5513243079185486,
      "learning_rate": 5.994748889188097e-05,
      "loss": 2.8507,
      "step": 133800
    },
    {
      "epoch": 36.020187338501295,
      "grad_norm": 0.5331861972808838,
      "learning_rate": 5.9745523091423175e-05,
      "loss": 2.8216,
      "step": 133850
    },
    {
      "epoch": 36.03364556416882,
      "grad_norm": 0.555422842502594,
      "learning_rate": 5.954355729096539e-05,
      "loss": 2.8205,
      "step": 133900
    },
    {
      "epoch": 36.04710378983635,
      "grad_norm": 0.5661362409591675,
      "learning_rate": 5.93415914905076e-05,
      "loss": 2.824,
      "step": 133950
    },
    {
      "epoch": 36.060562015503876,
      "grad_norm": 0.5626558661460876,
      "learning_rate": 5.913962569004982e-05,
      "loss": 2.8176,
      "step": 134000
    },
    {
      "epoch": 36.060562015503876,
      "eval_accuracy": 0.3983750993575494,
      "eval_loss": 3.2855114936828613,
      "eval_runtime": 185.0832,
      "eval_samples_per_second": 97.318,
      "eval_steps_per_second": 6.084,
      "step": 134000
    },
    {
      "epoch": 36.074020241171404,
      "grad_norm": 0.535280168056488,
      "learning_rate": 5.893765988959202e-05,
      "loss": 2.8101,
      "step": 134050
    },
    {
      "epoch": 36.08747846683893,
      "grad_norm": 0.5265828371047974,
      "learning_rate": 5.873569408913424e-05,
      "loss": 2.8213,
      "step": 134100
    },
    {
      "epoch": 36.10093669250646,
      "grad_norm": 0.5543688535690308,
      "learning_rate": 5.8533728288676447e-05,
      "loss": 2.8345,
      "step": 134150
    },
    {
      "epoch": 36.114394918173986,
      "grad_norm": 0.5701776742935181,
      "learning_rate": 5.8331762488218656e-05,
      "loss": 2.8308,
      "step": 134200
    },
    {
      "epoch": 36.12785314384151,
      "grad_norm": 0.5481976866722107,
      "learning_rate": 5.8129796687760866e-05,
      "loss": 2.8317,
      "step": 134250
    },
    {
      "epoch": 36.14131136950905,
      "grad_norm": 0.5497786998748779,
      "learning_rate": 5.7927830887303076e-05,
      "loss": 2.8286,
      "step": 134300
    },
    {
      "epoch": 36.154769595176575,
      "grad_norm": 0.5546992421150208,
      "learning_rate": 5.772586508684529e-05,
      "loss": 2.8228,
      "step": 134350
    },
    {
      "epoch": 36.1682278208441,
      "grad_norm": 0.5465831756591797,
      "learning_rate": 5.7523899286387495e-05,
      "loss": 2.8262,
      "step": 134400
    },
    {
      "epoch": 36.18168604651163,
      "grad_norm": 0.5486997365951538,
      "learning_rate": 5.732193348592971e-05,
      "loss": 2.8317,
      "step": 134450
    },
    {
      "epoch": 36.19514427217916,
      "grad_norm": 0.5851626992225647,
      "learning_rate": 5.711996768547192e-05,
      "loss": 2.8307,
      "step": 134500
    },
    {
      "epoch": 36.208602497846684,
      "grad_norm": 0.5671719312667847,
      "learning_rate": 5.691800188501414e-05,
      "loss": 2.824,
      "step": 134550
    },
    {
      "epoch": 36.22206072351421,
      "grad_norm": 0.5485758781433105,
      "learning_rate": 5.671603608455634e-05,
      "loss": 2.8278,
      "step": 134600
    },
    {
      "epoch": 36.23551894918174,
      "grad_norm": 0.5623011589050293,
      "learning_rate": 5.651407028409855e-05,
      "loss": 2.8345,
      "step": 134650
    },
    {
      "epoch": 36.248977174849266,
      "grad_norm": 0.5373241901397705,
      "learning_rate": 5.6312104483640766e-05,
      "loss": 2.8249,
      "step": 134700
    },
    {
      "epoch": 36.26243540051679,
      "grad_norm": 0.5760046243667603,
      "learning_rate": 5.6110138683182976e-05,
      "loss": 2.8385,
      "step": 134750
    },
    {
      "epoch": 36.27589362618432,
      "grad_norm": 0.5395543575286865,
      "learning_rate": 5.5908172882725186e-05,
      "loss": 2.8368,
      "step": 134800
    },
    {
      "epoch": 36.289351851851855,
      "grad_norm": 0.5136194825172424,
      "learning_rate": 5.5706207082267395e-05,
      "loss": 2.8355,
      "step": 134850
    },
    {
      "epoch": 36.30281007751938,
      "grad_norm": 0.5381130576133728,
      "learning_rate": 5.550424128180961e-05,
      "loss": 2.8456,
      "step": 134900
    },
    {
      "epoch": 36.31626830318691,
      "grad_norm": 0.5677443146705627,
      "learning_rate": 5.530227548135182e-05,
      "loss": 2.8347,
      "step": 134950
    },
    {
      "epoch": 36.32972652885444,
      "grad_norm": 0.5214958786964417,
      "learning_rate": 5.510030968089403e-05,
      "loss": 2.8358,
      "step": 135000
    },
    {
      "epoch": 36.32972652885444,
      "eval_accuracy": 0.3986302021744147,
      "eval_loss": 3.283909320831299,
      "eval_runtime": 185.0405,
      "eval_samples_per_second": 97.341,
      "eval_steps_per_second": 6.085,
      "step": 135000
    },
    {
      "epoch": 36.343184754521964,
      "grad_norm": 0.5448713898658752,
      "learning_rate": 5.489834388043624e-05,
      "loss": 2.8389,
      "step": 135050
    },
    {
      "epoch": 36.35664298018949,
      "grad_norm": 0.5596902966499329,
      "learning_rate": 5.469637807997846e-05,
      "loss": 2.8404,
      "step": 135100
    },
    {
      "epoch": 36.37010120585702,
      "grad_norm": 0.5640928149223328,
      "learning_rate": 5.449441227952067e-05,
      "loss": 2.8364,
      "step": 135150
    },
    {
      "epoch": 36.383559431524546,
      "grad_norm": 0.5564897656440735,
      "learning_rate": 5.429244647906287e-05,
      "loss": 2.8239,
      "step": 135200
    },
    {
      "epoch": 36.39701765719207,
      "grad_norm": 0.5490913391113281,
      "learning_rate": 5.4090480678605086e-05,
      "loss": 2.8481,
      "step": 135250
    },
    {
      "epoch": 36.4104758828596,
      "grad_norm": 0.5441024303436279,
      "learning_rate": 5.3888514878147296e-05,
      "loss": 2.8347,
      "step": 135300
    },
    {
      "epoch": 36.423934108527135,
      "grad_norm": 0.555292010307312,
      "learning_rate": 5.3686549077689505e-05,
      "loss": 2.84,
      "step": 135350
    },
    {
      "epoch": 36.43739233419466,
      "grad_norm": 0.5425558090209961,
      "learning_rate": 5.3484583277231715e-05,
      "loss": 2.8416,
      "step": 135400
    },
    {
      "epoch": 36.45085055986219,
      "grad_norm": 0.5508972406387329,
      "learning_rate": 5.328261747677393e-05,
      "loss": 2.8283,
      "step": 135450
    },
    {
      "epoch": 36.46430878552972,
      "grad_norm": 0.5342379808425903,
      "learning_rate": 5.308065167631614e-05,
      "loss": 2.8472,
      "step": 135500
    },
    {
      "epoch": 36.477767011197244,
      "grad_norm": 0.538860023021698,
      "learning_rate": 5.287868587585835e-05,
      "loss": 2.8417,
      "step": 135550
    },
    {
      "epoch": 36.49122523686477,
      "grad_norm": 0.5174806118011475,
      "learning_rate": 5.267672007540056e-05,
      "loss": 2.836,
      "step": 135600
    },
    {
      "epoch": 36.5046834625323,
      "grad_norm": 0.538652241230011,
      "learning_rate": 5.247475427494277e-05,
      "loss": 2.8371,
      "step": 135650
    },
    {
      "epoch": 36.518141688199826,
      "grad_norm": 0.5652296543121338,
      "learning_rate": 5.2272788474484986e-05,
      "loss": 2.8288,
      "step": 135700
    },
    {
      "epoch": 36.53159991386735,
      "grad_norm": 0.5380600690841675,
      "learning_rate": 5.207082267402719e-05,
      "loss": 2.8364,
      "step": 135750
    },
    {
      "epoch": 36.54505813953488,
      "grad_norm": 0.5351638793945312,
      "learning_rate": 5.1868856873569406e-05,
      "loss": 2.8392,
      "step": 135800
    },
    {
      "epoch": 36.558516365202415,
      "grad_norm": 0.5255042314529419,
      "learning_rate": 5.1666891073111615e-05,
      "loss": 2.8522,
      "step": 135850
    },
    {
      "epoch": 36.57197459086994,
      "grad_norm": 0.5579770803451538,
      "learning_rate": 5.146492527265383e-05,
      "loss": 2.8542,
      "step": 135900
    },
    {
      "epoch": 36.58543281653747,
      "grad_norm": 0.5614697933197021,
      "learning_rate": 5.1262959472196035e-05,
      "loss": 2.8524,
      "step": 135950
    },
    {
      "epoch": 36.598891042205,
      "grad_norm": 0.5513339042663574,
      "learning_rate": 5.106099367173825e-05,
      "loss": 2.8456,
      "step": 136000
    },
    {
      "epoch": 36.598891042205,
      "eval_accuracy": 0.39873395992147875,
      "eval_loss": 3.2819995880126953,
      "eval_runtime": 185.0751,
      "eval_samples_per_second": 97.323,
      "eval_steps_per_second": 6.084,
      "step": 136000
    },
    {
      "epoch": 36.612349267872524,
      "grad_norm": 0.5263423919677734,
      "learning_rate": 5.085902787128046e-05,
      "loss": 2.8455,
      "step": 136050
    },
    {
      "epoch": 36.62580749354005,
      "grad_norm": 0.5566456913948059,
      "learning_rate": 5.0657062070822664e-05,
      "loss": 2.8521,
      "step": 136100
    },
    {
      "epoch": 36.63926571920758,
      "grad_norm": 0.560788631439209,
      "learning_rate": 5.045509627036488e-05,
      "loss": 2.845,
      "step": 136150
    },
    {
      "epoch": 36.652723944875106,
      "grad_norm": 0.5380584597587585,
      "learning_rate": 5.025313046990709e-05,
      "loss": 2.8351,
      "step": 136200
    },
    {
      "epoch": 36.66618217054263,
      "grad_norm": 0.5398831963539124,
      "learning_rate": 5.0051164669449306e-05,
      "loss": 2.8393,
      "step": 136250
    },
    {
      "epoch": 36.67964039621016,
      "grad_norm": 0.5925846099853516,
      "learning_rate": 4.984919886899151e-05,
      "loss": 2.856,
      "step": 136300
    },
    {
      "epoch": 36.693098621877695,
      "grad_norm": 0.5680198669433594,
      "learning_rate": 4.9647233068533725e-05,
      "loss": 2.8538,
      "step": 136350
    },
    {
      "epoch": 36.70655684754522,
      "grad_norm": 0.5498775243759155,
      "learning_rate": 4.9445267268075935e-05,
      "loss": 2.8447,
      "step": 136400
    },
    {
      "epoch": 36.72001507321275,
      "grad_norm": 0.6519020199775696,
      "learning_rate": 4.924330146761815e-05,
      "loss": 2.8363,
      "step": 136450
    },
    {
      "epoch": 36.73347329888028,
      "grad_norm": 0.542931079864502,
      "learning_rate": 4.9041335667160354e-05,
      "loss": 2.8475,
      "step": 136500
    },
    {
      "epoch": 36.746931524547804,
      "grad_norm": 0.5331995487213135,
      "learning_rate": 4.883936986670257e-05,
      "loss": 2.8461,
      "step": 136550
    },
    {
      "epoch": 36.76038975021533,
      "grad_norm": 0.5860674381256104,
      "learning_rate": 4.863740406624478e-05,
      "loss": 2.8563,
      "step": 136600
    },
    {
      "epoch": 36.77384797588286,
      "grad_norm": 0.5560420751571655,
      "learning_rate": 4.843543826578698e-05,
      "loss": 2.8406,
      "step": 136650
    },
    {
      "epoch": 36.787306201550386,
      "grad_norm": 0.5390902161598206,
      "learning_rate": 4.82334724653292e-05,
      "loss": 2.8624,
      "step": 136700
    },
    {
      "epoch": 36.80076442721791,
      "grad_norm": 0.5597389936447144,
      "learning_rate": 4.803150666487141e-05,
      "loss": 2.8441,
      "step": 136750
    },
    {
      "epoch": 36.81422265288544,
      "grad_norm": 0.5445606708526611,
      "learning_rate": 4.7829540864413626e-05,
      "loss": 2.8501,
      "step": 136800
    },
    {
      "epoch": 36.82768087855297,
      "grad_norm": 0.5534197688102722,
      "learning_rate": 4.762757506395583e-05,
      "loss": 2.8494,
      "step": 136850
    },
    {
      "epoch": 36.8411391042205,
      "grad_norm": 0.535542368888855,
      "learning_rate": 4.7425609263498045e-05,
      "loss": 2.8489,
      "step": 136900
    },
    {
      "epoch": 36.85459732988803,
      "grad_norm": 0.5582454204559326,
      "learning_rate": 4.7223643463040255e-05,
      "loss": 2.8509,
      "step": 136950
    },
    {
      "epoch": 36.86805555555556,
      "grad_norm": 0.5511658787727356,
      "learning_rate": 4.702167766258247e-05,
      "loss": 2.8367,
      "step": 137000
    },
    {
      "epoch": 36.86805555555556,
      "eval_accuracy": 0.39909684041906396,
      "eval_loss": 3.276746988296509,
      "eval_runtime": 185.1452,
      "eval_samples_per_second": 97.286,
      "eval_steps_per_second": 6.082,
      "step": 137000
    },
    {
      "epoch": 36.881513781223084,
      "grad_norm": 0.5359370112419128,
      "learning_rate": 4.6819711862124674e-05,
      "loss": 2.8548,
      "step": 137050
    },
    {
      "epoch": 36.89497200689061,
      "grad_norm": 0.5714049935340881,
      "learning_rate": 4.6617746061666884e-05,
      "loss": 2.8506,
      "step": 137100
    },
    {
      "epoch": 36.90843023255814,
      "grad_norm": 0.5278427600860596,
      "learning_rate": 4.64157802612091e-05,
      "loss": 2.8537,
      "step": 137150
    },
    {
      "epoch": 36.921888458225666,
      "grad_norm": 0.5692651867866516,
      "learning_rate": 4.62138144607513e-05,
      "loss": 2.849,
      "step": 137200
    },
    {
      "epoch": 36.935346683893194,
      "grad_norm": 0.5612398386001587,
      "learning_rate": 4.601184866029352e-05,
      "loss": 2.8476,
      "step": 137250
    },
    {
      "epoch": 36.94880490956072,
      "grad_norm": 0.5397350788116455,
      "learning_rate": 4.580988285983573e-05,
      "loss": 2.8502,
      "step": 137300
    },
    {
      "epoch": 36.96226313522825,
      "grad_norm": 0.5517504215240479,
      "learning_rate": 4.5607917059377945e-05,
      "loss": 2.8528,
      "step": 137350
    },
    {
      "epoch": 36.97572136089578,
      "grad_norm": 0.5510405898094177,
      "learning_rate": 4.540595125892015e-05,
      "loss": 2.8463,
      "step": 137400
    },
    {
      "epoch": 36.98917958656331,
      "grad_norm": 0.5432448983192444,
      "learning_rate": 4.5203985458462365e-05,
      "loss": 2.8421,
      "step": 137450
    },
    {
      "epoch": 37.002422480620154,
      "grad_norm": 0.5367938876152039,
      "learning_rate": 4.5002019658004574e-05,
      "loss": 2.8457,
      "step": 137500
    },
    {
      "epoch": 37.01588070628768,
      "grad_norm": 0.5568806529045105,
      "learning_rate": 4.480005385754679e-05,
      "loss": 2.8216,
      "step": 137550
    },
    {
      "epoch": 37.02933893195521,
      "grad_norm": 0.5297728776931763,
      "learning_rate": 4.4598088057088994e-05,
      "loss": 2.8143,
      "step": 137600
    },
    {
      "epoch": 37.042797157622736,
      "grad_norm": 0.5489912629127502,
      "learning_rate": 4.43961222566312e-05,
      "loss": 2.806,
      "step": 137650
    },
    {
      "epoch": 37.056255383290264,
      "grad_norm": 0.5584233403205872,
      "learning_rate": 4.419415645617342e-05,
      "loss": 2.8124,
      "step": 137700
    },
    {
      "epoch": 37.0697136089578,
      "grad_norm": 0.5610176920890808,
      "learning_rate": 4.399219065571562e-05,
      "loss": 2.8285,
      "step": 137750
    },
    {
      "epoch": 37.083171834625325,
      "grad_norm": 0.5475199818611145,
      "learning_rate": 4.379022485525784e-05,
      "loss": 2.8177,
      "step": 137800
    },
    {
      "epoch": 37.09663006029285,
      "grad_norm": 0.5631940364837646,
      "learning_rate": 4.358825905480005e-05,
      "loss": 2.8203,
      "step": 137850
    },
    {
      "epoch": 37.11008828596038,
      "grad_norm": 0.5738129615783691,
      "learning_rate": 4.3386293254342265e-05,
      "loss": 2.8178,
      "step": 137900
    },
    {
      "epoch": 37.12354651162791,
      "grad_norm": 0.5770145058631897,
      "learning_rate": 4.318432745388447e-05,
      "loss": 2.8254,
      "step": 137950
    },
    {
      "epoch": 37.137004737295435,
      "grad_norm": 0.5465659499168396,
      "learning_rate": 4.2982361653426684e-05,
      "loss": 2.8204,
      "step": 138000
    },
    {
      "epoch": 37.137004737295435,
      "eval_accuracy": 0.3987451505476019,
      "eval_loss": 3.2841856479644775,
      "eval_runtime": 185.0894,
      "eval_samples_per_second": 97.315,
      "eval_steps_per_second": 6.084,
      "step": 138000
    },
    {
      "epoch": 37.15046296296296,
      "grad_norm": 0.5489895343780518,
      "learning_rate": 4.2780395852968894e-05,
      "loss": 2.8239,
      "step": 138050
    },
    {
      "epoch": 37.16392118863049,
      "grad_norm": 0.5280149579048157,
      "learning_rate": 4.2578430052511104e-05,
      "loss": 2.8289,
      "step": 138100
    },
    {
      "epoch": 37.17737941429802,
      "grad_norm": 0.5438522100448608,
      "learning_rate": 4.237646425205331e-05,
      "loss": 2.814,
      "step": 138150
    },
    {
      "epoch": 37.190837639965544,
      "grad_norm": 0.5496814846992493,
      "learning_rate": 4.217449845159552e-05,
      "loss": 2.8292,
      "step": 138200
    },
    {
      "epoch": 37.20429586563308,
      "grad_norm": 0.5470216274261475,
      "learning_rate": 4.197253265113774e-05,
      "loss": 2.8256,
      "step": 138250
    },
    {
      "epoch": 37.217754091300606,
      "grad_norm": 0.5606210231781006,
      "learning_rate": 4.177056685067994e-05,
      "loss": 2.8306,
      "step": 138300
    },
    {
      "epoch": 37.23121231696813,
      "grad_norm": 0.5440297722816467,
      "learning_rate": 4.156860105022216e-05,
      "loss": 2.8278,
      "step": 138350
    },
    {
      "epoch": 37.24467054263566,
      "grad_norm": 0.5358032584190369,
      "learning_rate": 4.136663524976437e-05,
      "loss": 2.8254,
      "step": 138400
    },
    {
      "epoch": 37.25812876830319,
      "grad_norm": 0.5529847145080566,
      "learning_rate": 4.1164669449306585e-05,
      "loss": 2.8224,
      "step": 138450
    },
    {
      "epoch": 37.271586993970715,
      "grad_norm": 0.5472120046615601,
      "learning_rate": 4.096270364884879e-05,
      "loss": 2.818,
      "step": 138500
    },
    {
      "epoch": 37.28504521963824,
      "grad_norm": 0.5636007785797119,
      "learning_rate": 4.0760737848391004e-05,
      "loss": 2.8179,
      "step": 138550
    },
    {
      "epoch": 37.29850344530577,
      "grad_norm": 0.5373379588127136,
      "learning_rate": 4.0558772047933214e-05,
      "loss": 2.8263,
      "step": 138600
    },
    {
      "epoch": 37.3119616709733,
      "grad_norm": 0.5434703826904297,
      "learning_rate": 4.0356806247475423e-05,
      "loss": 2.8245,
      "step": 138650
    },
    {
      "epoch": 37.325419896640824,
      "grad_norm": 0.5582107305526733,
      "learning_rate": 4.015484044701763e-05,
      "loss": 2.8128,
      "step": 138700
    },
    {
      "epoch": 37.33887812230836,
      "grad_norm": 0.5430098176002502,
      "learning_rate": 3.995287464655984e-05,
      "loss": 2.8321,
      "step": 138750
    },
    {
      "epoch": 37.352336347975886,
      "grad_norm": 0.5941494107246399,
      "learning_rate": 3.975090884610206e-05,
      "loss": 2.8342,
      "step": 138800
    },
    {
      "epoch": 37.36579457364341,
      "grad_norm": 0.561592698097229,
      "learning_rate": 3.954894304564427e-05,
      "loss": 2.835,
      "step": 138850
    },
    {
      "epoch": 37.37925279931094,
      "grad_norm": 0.5493949055671692,
      "learning_rate": 3.934697724518648e-05,
      "loss": 2.8335,
      "step": 138900
    },
    {
      "epoch": 37.39271102497847,
      "grad_norm": 0.5652241706848145,
      "learning_rate": 3.914501144472869e-05,
      "loss": 2.8281,
      "step": 138950
    },
    {
      "epoch": 37.406169250645995,
      "grad_norm": 0.5507028102874756,
      "learning_rate": 3.8943045644270905e-05,
      "loss": 2.8306,
      "step": 139000
    },
    {
      "epoch": 37.406169250645995,
      "eval_accuracy": 0.398903014428737,
      "eval_loss": 3.2826850414276123,
      "eval_runtime": 185.0422,
      "eval_samples_per_second": 97.34,
      "eval_steps_per_second": 6.085,
      "step": 139000
    },
    {
      "epoch": 37.41962747631352,
      "grad_norm": 0.5337747931480408,
      "learning_rate": 3.8741079843813114e-05,
      "loss": 2.8402,
      "step": 139050
    },
    {
      "epoch": 37.43308570198105,
      "grad_norm": 0.5691428780555725,
      "learning_rate": 3.853911404335532e-05,
      "loss": 2.8324,
      "step": 139100
    },
    {
      "epoch": 37.44654392764858,
      "grad_norm": 0.5328547954559326,
      "learning_rate": 3.8337148242897533e-05,
      "loss": 2.8208,
      "step": 139150
    },
    {
      "epoch": 37.460002153316104,
      "grad_norm": 0.5412877202033997,
      "learning_rate": 3.813518244243974e-05,
      "loss": 2.8336,
      "step": 139200
    },
    {
      "epoch": 37.47346037898363,
      "grad_norm": 0.5513226985931396,
      "learning_rate": 3.793321664198195e-05,
      "loss": 2.8282,
      "step": 139250
    },
    {
      "epoch": 37.486918604651166,
      "grad_norm": 0.571736752986908,
      "learning_rate": 3.773125084152416e-05,
      "loss": 2.8317,
      "step": 139300
    },
    {
      "epoch": 37.50037683031869,
      "grad_norm": 0.5565307140350342,
      "learning_rate": 3.752928504106638e-05,
      "loss": 2.8289,
      "step": 139350
    },
    {
      "epoch": 37.51383505598622,
      "grad_norm": 0.5516583323478699,
      "learning_rate": 3.732731924060859e-05,
      "loss": 2.8312,
      "step": 139400
    },
    {
      "epoch": 37.52729328165375,
      "grad_norm": 0.5443484783172607,
      "learning_rate": 3.71253534401508e-05,
      "loss": 2.8316,
      "step": 139450
    },
    {
      "epoch": 37.540751507321275,
      "grad_norm": 0.542640209197998,
      "learning_rate": 3.692338763969301e-05,
      "loss": 2.834,
      "step": 139500
    },
    {
      "epoch": 37.5542097329888,
      "grad_norm": 0.5565000772476196,
      "learning_rate": 3.672142183923522e-05,
      "loss": 2.8275,
      "step": 139550
    },
    {
      "epoch": 37.56766795865633,
      "grad_norm": 0.5553215146064758,
      "learning_rate": 3.6519456038777434e-05,
      "loss": 2.8363,
      "step": 139600
    },
    {
      "epoch": 37.58112618432386,
      "grad_norm": 0.536067008972168,
      "learning_rate": 3.6317490238319643e-05,
      "loss": 2.8296,
      "step": 139650
    },
    {
      "epoch": 37.594584409991384,
      "grad_norm": 0.5395115613937378,
      "learning_rate": 3.611552443786185e-05,
      "loss": 2.8354,
      "step": 139700
    },
    {
      "epoch": 37.60804263565891,
      "grad_norm": 0.5381755232810974,
      "learning_rate": 3.591355863740406e-05,
      "loss": 2.8319,
      "step": 139750
    },
    {
      "epoch": 37.621500861326446,
      "grad_norm": 0.5646410584449768,
      "learning_rate": 3.571159283694627e-05,
      "loss": 2.8335,
      "step": 139800
    },
    {
      "epoch": 37.63495908699397,
      "grad_norm": 0.5553077459335327,
      "learning_rate": 3.550962703648848e-05,
      "loss": 2.8444,
      "step": 139850
    },
    {
      "epoch": 37.6484173126615,
      "grad_norm": 0.5285354852676392,
      "learning_rate": 3.530766123603069e-05,
      "loss": 2.826,
      "step": 139900
    },
    {
      "epoch": 37.66187553832903,
      "grad_norm": 0.5623334050178528,
      "learning_rate": 3.510569543557291e-05,
      "loss": 2.8264,
      "step": 139950
    },
    {
      "epoch": 37.675333763996555,
      "grad_norm": 0.5664812326431274,
      "learning_rate": 3.490372963511512e-05,
      "loss": 2.8403,
      "step": 140000
    },
    {
      "epoch": 37.675333763996555,
      "eval_accuracy": 0.3991455142103568,
      "eval_loss": 3.2796993255615234,
      "eval_runtime": 185.0168,
      "eval_samples_per_second": 97.353,
      "eval_steps_per_second": 6.086,
      "step": 140000
    },
    {
      "epoch": 37.675333763996555,
      "step": 140000,
      "total_flos": 2.925872922230784e+18,
      "train_loss": 0.40768963841029576,
      "train_runtime": 37932.5985,
      "train_samples_per_second": 313.415,
      "train_steps_per_second": 3.919
    }
  ],
  "logging_steps": 50,
  "max_steps": 148640,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 10000,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 20,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 20
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.925872922230784e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}