{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.6539440203562341,
  "eval_steps": 200,
  "global_step": 7800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.004240882103477523,
      "grad_norm": 0.4664132297039032,
      "learning_rate": 0.00019,
      "loss": 1.9501,
      "step": 20
    },
    {
      "epoch": 0.008481764206955046,
      "grad_norm": 0.6104708909988403,
      "learning_rate": 0.00019968280467445743,
      "loss": 1.9307,
      "step": 40
    },
    {
      "epoch": 0.01272264631043257,
      "grad_norm": 0.6868094801902771,
      "learning_rate": 0.00019934891485809685,
      "loss": 1.8434,
      "step": 60
    },
    {
      "epoch": 0.016963528413910092,
      "grad_norm": 0.8883188962936401,
      "learning_rate": 0.00019901502504173624,
      "loss": 1.8516,
      "step": 80
    },
    {
      "epoch": 0.021204410517387615,
      "grad_norm": 0.6765041351318359,
      "learning_rate": 0.00019868113522537563,
      "loss": 1.8814,
      "step": 100
    },
    {
      "epoch": 0.02544529262086514,
      "grad_norm": 0.9518934488296509,
      "learning_rate": 0.00019834724540901503,
      "loss": 1.7392,
      "step": 120
    },
    {
      "epoch": 0.029686174724342665,
      "grad_norm": 0.8360865712165833,
      "learning_rate": 0.00019801335559265445,
      "loss": 1.8312,
      "step": 140
    },
    {
      "epoch": 0.033927056827820185,
      "grad_norm": 0.8230759501457214,
      "learning_rate": 0.00019767946577629384,
      "loss": 1.7807,
      "step": 160
    },
    {
      "epoch": 0.03816793893129771,
      "grad_norm": 0.6617977619171143,
      "learning_rate": 0.00019734557595993323,
      "loss": 1.767,
      "step": 180
    },
    {
      "epoch": 0.04240882103477523,
      "grad_norm": 0.742358922958374,
      "learning_rate": 0.00019701168614357262,
      "loss": 1.7217,
      "step": 200
    },
    {
      "epoch": 0.04240882103477523,
      "eval_loss": NaN,
      "eval_runtime": 37.9555,
      "eval_samples_per_second": 26.347,
      "eval_steps_per_second": 3.293,
      "step": 200
    },
    {
      "epoch": 0.04664970313825276,
      "grad_norm": 0.6581922173500061,
      "learning_rate": 0.00019667779632721204,
      "loss": 1.7217,
      "step": 220
    },
    {
      "epoch": 0.05089058524173028,
      "grad_norm": 0.9428125619888306,
      "learning_rate": 0.00019634390651085144,
      "loss": 1.7576,
      "step": 240
    },
    {
      "epoch": 0.0551314673452078,
      "grad_norm": 0.9091566801071167,
      "learning_rate": 0.0001960100166944908,
      "loss": 1.6873,
      "step": 260
    },
    {
      "epoch": 0.05937234944868533,
      "grad_norm": 1.0092300176620483,
      "learning_rate": 0.00019567612687813022,
      "loss": 1.6463,
      "step": 280
    },
    {
      "epoch": 0.06361323155216285,
      "grad_norm": 0.9816062450408936,
      "learning_rate": 0.00019534223706176962,
      "loss": 1.6733,
      "step": 300
    },
    {
      "epoch": 0.06785411365564037,
      "grad_norm": 0.9674614071846008,
      "learning_rate": 0.000195008347245409,
      "loss": 1.6471,
      "step": 320
    },
    {
      "epoch": 0.0720949957591179,
      "grad_norm": 1.0324562788009644,
      "learning_rate": 0.00019467445742904843,
      "loss": 1.6639,
      "step": 340
    },
    {
      "epoch": 0.07633587786259542,
      "grad_norm": 0.9738441705703735,
      "learning_rate": 0.00019434056761268782,
      "loss": 1.6454,
      "step": 360
    },
    {
      "epoch": 0.08057675996607294,
      "grad_norm": 1.0142924785614014,
      "learning_rate": 0.00019400667779632721,
      "loss": 1.5662,
      "step": 380
    },
    {
      "epoch": 0.08481764206955046,
      "grad_norm": 1.3319178819656372,
      "learning_rate": 0.0001936727879799666,
      "loss": 1.6187,
      "step": 400
    },
    {
      "epoch": 0.08481764206955046,
      "eval_loss": NaN,
      "eval_runtime": 37.8512,
      "eval_samples_per_second": 26.419,
      "eval_steps_per_second": 3.302,
      "step": 400
    },
    {
      "epoch": 0.089058524173028,
      "grad_norm": 1.0217434167861938,
      "learning_rate": 0.00019333889816360603,
      "loss": 1.6443,
      "step": 420
    },
    {
      "epoch": 0.09329940627650551,
      "grad_norm": 0.9588495492935181,
      "learning_rate": 0.00019300500834724542,
      "loss": 1.6175,
      "step": 440
    },
    {
      "epoch": 0.09754028837998303,
      "grad_norm": 1.3390898704528809,
      "learning_rate": 0.0001926711185308848,
      "loss": 1.5598,
      "step": 460
    },
    {
      "epoch": 0.10178117048346055,
      "grad_norm": 0.9628949761390686,
      "learning_rate": 0.00019233722871452423,
      "loss": 1.5223,
      "step": 480
    },
    {
      "epoch": 0.10602205258693809,
      "grad_norm": 1.0719170570373535,
      "learning_rate": 0.00019200333889816363,
      "loss": 1.5544,
      "step": 500
    },
    {
      "epoch": 0.1102629346904156,
      "grad_norm": 1.0842593908309937,
      "learning_rate": 0.00019166944908180302,
      "loss": 1.5238,
      "step": 520
    },
    {
      "epoch": 0.11450381679389313,
      "grad_norm": 1.2601176500320435,
      "learning_rate": 0.0001913355592654424,
      "loss": 1.4784,
      "step": 540
    },
    {
      "epoch": 0.11874469889737066,
      "grad_norm": 1.0640774965286255,
      "learning_rate": 0.00019100166944908183,
      "loss": 1.4986,
      "step": 560
    },
    {
      "epoch": 0.12298558100084818,
      "grad_norm": 1.4637383222579956,
      "learning_rate": 0.00019066777963272122,
      "loss": 1.5739,
      "step": 580
    },
    {
      "epoch": 0.1272264631043257,
      "grad_norm": 2.3007261753082275,
      "learning_rate": 0.0001903338898163606,
      "loss": 1.4499,
      "step": 600
    },
    {
      "epoch": 0.1272264631043257,
      "eval_loss": NaN,
      "eval_runtime": 37.8269,
      "eval_samples_per_second": 26.436,
      "eval_steps_per_second": 3.305,
      "step": 600
    },
    {
      "epoch": 0.13146734520780323,
      "grad_norm": 1.212930679321289,
      "learning_rate": 0.00019,
      "loss": 1.4338,
      "step": 620
    },
    {
      "epoch": 0.13570822731128074,
      "grad_norm": 1.016894817352295,
      "learning_rate": 0.0001896661101836394,
      "loss": 1.4397,
      "step": 640
    },
    {
      "epoch": 0.13994910941475827,
      "grad_norm": 1.2239773273468018,
      "learning_rate": 0.0001893322203672788,
      "loss": 1.46,
      "step": 660
    },
    {
      "epoch": 0.1441899915182358,
      "grad_norm": 1.1503028869628906,
      "learning_rate": 0.0001889983305509182,
      "loss": 1.4547,
      "step": 680
    },
    {
      "epoch": 0.1484308736217133,
      "grad_norm": 1.0300050973892212,
      "learning_rate": 0.0001886644407345576,
      "loss": 1.4064,
      "step": 700
    },
    {
      "epoch": 0.15267175572519084,
      "grad_norm": 1.1932989358901978,
      "learning_rate": 0.000188330550918197,
      "loss": 1.3682,
      "step": 720
    },
    {
      "epoch": 0.15691263782866835,
      "grad_norm": 1.3355811834335327,
      "learning_rate": 0.0001879966611018364,
      "loss": 1.3677,
      "step": 740
    },
    {
      "epoch": 0.16115351993214588,
      "grad_norm": 1.5467063188552856,
      "learning_rate": 0.00018766277128547581,
      "loss": 1.4191,
      "step": 760
    },
    {
      "epoch": 0.16539440203562342,
      "grad_norm": 1.5290387868881226,
      "learning_rate": 0.0001873288814691152,
      "loss": 1.3353,
      "step": 780
    },
    {
      "epoch": 0.16963528413910092,
      "grad_norm": 1.5212856531143188,
      "learning_rate": 0.0001869949916527546,
      "loss": 1.3441,
      "step": 800
    },
    {
      "epoch": 0.16963528413910092,
      "eval_loss": NaN,
      "eval_runtime": 37.934,
      "eval_samples_per_second": 26.362,
      "eval_steps_per_second": 3.295,
      "step": 800
    },
    {
      "epoch": 0.17387616624257846,
      "grad_norm": 1.4864256381988525,
      "learning_rate": 0.000186661101836394,
      "loss": 1.2865,
      "step": 820
    },
    {
      "epoch": 0.178117048346056,
      "grad_norm": 1.727289080619812,
      "learning_rate": 0.0001863272120200334,
      "loss": 1.2379,
      "step": 840
    },
    {
      "epoch": 0.1823579304495335,
      "grad_norm": 1.4979146718978882,
      "learning_rate": 0.0001859933222036728,
      "loss": 1.4105,
      "step": 860
    },
    {
      "epoch": 0.18659881255301103,
      "grad_norm": 1.2873855829238892,
      "learning_rate": 0.0001856594323873122,
      "loss": 1.295,
      "step": 880
    },
    {
      "epoch": 0.19083969465648856,
      "grad_norm": 1.845335841178894,
      "learning_rate": 0.0001853255425709516,
      "loss": 1.2669,
      "step": 900
    },
    {
      "epoch": 0.19508057675996607,
      "grad_norm": 1.209466576576233,
      "learning_rate": 0.00018499165275459098,
      "loss": 1.2415,
      "step": 920
    },
    {
      "epoch": 0.1993214588634436,
      "grad_norm": 1.5466270446777344,
      "learning_rate": 0.00018465776293823038,
      "loss": 1.3774,
      "step": 940
    },
    {
      "epoch": 0.2035623409669211,
      "grad_norm": 1.4850807189941406,
      "learning_rate": 0.00018432387312186977,
      "loss": 1.2678,
      "step": 960
    },
    {
      "epoch": 0.20780322307039864,
      "grad_norm": 1.8490267992019653,
      "learning_rate": 0.0001839899833055092,
      "loss": 1.2155,
      "step": 980
    },
    {
      "epoch": 0.21204410517387617,
      "grad_norm": 1.5460015535354614,
      "learning_rate": 0.00018365609348914858,
      "loss": 1.2701,
      "step": 1000
    },
    {
      "epoch": 0.21204410517387617,
      "eval_loss": NaN,
      "eval_runtime": 37.9749,
      "eval_samples_per_second": 26.333,
      "eval_steps_per_second": 3.292,
      "step": 1000
    },
    {
      "epoch": 0.21628498727735368,
      "grad_norm": 1.6261916160583496,
      "learning_rate": 0.00018332220367278798,
      "loss": 1.2321,
      "step": 1020
    },
    {
      "epoch": 0.2205258693808312,
      "grad_norm": 1.7328996658325195,
      "learning_rate": 0.0001829883138564274,
      "loss": 1.1889,
      "step": 1040
    },
    {
      "epoch": 0.22476675148430875,
      "grad_norm": 1.444868564605713,
      "learning_rate": 0.0001826544240400668,
      "loss": 1.2224,
      "step": 1060
    },
    {
      "epoch": 0.22900763358778625,
      "grad_norm": 1.4370837211608887,
      "learning_rate": 0.00018232053422370618,
      "loss": 1.2079,
      "step": 1080
    },
    {
      "epoch": 0.23324851569126379,
      "grad_norm": 1.7325958013534546,
      "learning_rate": 0.00018198664440734557,
      "loss": 1.1496,
      "step": 1100
    },
    {
      "epoch": 0.23748939779474132,
      "grad_norm": 1.6572184562683105,
      "learning_rate": 0.000181652754590985,
      "loss": 1.1805,
      "step": 1120
    },
    {
      "epoch": 0.24173027989821882,
      "grad_norm": 2.1474740505218506,
      "learning_rate": 0.0001813188647746244,
      "loss": 1.1641,
      "step": 1140
    },
    {
      "epoch": 0.24597116200169636,
      "grad_norm": 1.5673317909240723,
      "learning_rate": 0.00018098497495826378,
      "loss": 1.2177,
      "step": 1160
    },
    {
      "epoch": 0.2502120441051739,
      "grad_norm": 1.8548556566238403,
      "learning_rate": 0.00018065108514190317,
      "loss": 1.1209,
      "step": 1180
    },
    {
      "epoch": 0.2544529262086514,
      "grad_norm": 2.2383711338043213,
      "learning_rate": 0.0001803171953255426,
      "loss": 1.126,
      "step": 1200
    },
    {
      "epoch": 0.2544529262086514,
      "eval_loss": NaN,
      "eval_runtime": 37.8124,
      "eval_samples_per_second": 26.446,
      "eval_steps_per_second": 3.306,
      "step": 1200
    },
    {
      "epoch": 0.2586938083121289,
      "grad_norm": 1.7100919485092163,
      "learning_rate": 0.00017998330550918199,
      "loss": 1.1178,
      "step": 1220
    },
    {
      "epoch": 0.26293469041560646,
      "grad_norm": 2.1981756687164307,
      "learning_rate": 0.00017964941569282138,
      "loss": 1.1507,
      "step": 1240
    },
    {
      "epoch": 0.26717557251908397,
      "grad_norm": 1.5975751876831055,
      "learning_rate": 0.00017931552587646077,
      "loss": 1.0739,
      "step": 1260
    },
    {
      "epoch": 0.2714164546225615,
      "grad_norm": 1.7303990125656128,
      "learning_rate": 0.00017898163606010017,
      "loss": 1.082,
      "step": 1280
    },
    {
      "epoch": 0.27565733672603904,
      "grad_norm": 1.694312334060669,
      "learning_rate": 0.00017864774624373956,
      "loss": 1.0806,
      "step": 1300
    },
    {
      "epoch": 0.27989821882951654,
      "grad_norm": 1.8411606550216675,
      "learning_rate": 0.00017831385642737898,
      "loss": 1.063,
      "step": 1320
    },
    {
      "epoch": 0.28413910093299405,
      "grad_norm": 1.8138402700424194,
      "learning_rate": 0.00017797996661101837,
      "loss": 1.0353,
      "step": 1340
    },
    {
      "epoch": 0.2883799830364716,
      "grad_norm": 1.689369559288025,
      "learning_rate": 0.00017764607679465776,
      "loss": 0.9812,
      "step": 1360
    },
    {
      "epoch": 0.2926208651399491,
      "grad_norm": 2.3171262741088867,
      "learning_rate": 0.00017731218697829716,
      "loss": 1.0633,
      "step": 1380
    },
    {
      "epoch": 0.2968617472434266,
      "grad_norm": 2.1252689361572266,
      "learning_rate": 0.00017697829716193658,
      "loss": 1.0333,
      "step": 1400
    },
    {
      "epoch": 0.2968617472434266,
      "eval_loss": NaN,
      "eval_runtime": 37.8223,
      "eval_samples_per_second": 26.439,
      "eval_steps_per_second": 3.305,
      "step": 1400
    },
    {
      "epoch": 0.3011026293469042,
      "grad_norm": 2.131967544555664,
      "learning_rate": 0.00017664440734557597,
      "loss": 1.0446,
      "step": 1420
    },
    {
      "epoch": 0.3053435114503817,
      "grad_norm": 2.162263870239258,
      "learning_rate": 0.00017631051752921536,
      "loss": 1.0348,
      "step": 1440
    },
    {
      "epoch": 0.3095843935538592,
      "grad_norm": 2.072547197341919,
      "learning_rate": 0.00017597662771285478,
      "loss": 1.0648,
      "step": 1460
    },
    {
      "epoch": 0.3138252756573367,
      "grad_norm": 2.038855791091919,
      "learning_rate": 0.00017564273789649418,
      "loss": 1.0076,
      "step": 1480
    },
    {
      "epoch": 0.31806615776081426,
      "grad_norm": 1.827268362045288,
      "learning_rate": 0.00017530884808013357,
      "loss": 1.0256,
      "step": 1500
    },
    {
      "epoch": 0.32230703986429177,
      "grad_norm": 1.92107355594635,
      "learning_rate": 0.00017497495826377296,
      "loss": 1.0388,
      "step": 1520
    },
    {
      "epoch": 0.3265479219677693,
      "grad_norm": 1.9847649335861206,
      "learning_rate": 0.00017464106844741238,
      "loss": 0.9606,
      "step": 1540
    },
    {
      "epoch": 0.33078880407124683,
      "grad_norm": 1.8608839511871338,
      "learning_rate": 0.00017430717863105177,
      "loss": 0.9937,
      "step": 1560
    },
    {
      "epoch": 0.33502968617472434,
      "grad_norm": 2.0207772254943848,
      "learning_rate": 0.00017397328881469117,
      "loss": 0.9853,
      "step": 1580
    },
    {
      "epoch": 0.33927056827820185,
      "grad_norm": 2.1667962074279785,
      "learning_rate": 0.00017363939899833056,
      "loss": 0.9284,
      "step": 1600
    },
    {
      "epoch": 0.33927056827820185,
      "eval_loss": NaN,
      "eval_runtime": 37.8987,
      "eval_samples_per_second": 26.386,
      "eval_steps_per_second": 3.298,
      "step": 1600
    },
    {
      "epoch": 0.3435114503816794,
      "grad_norm": 2.407201051712036,
      "learning_rate": 0.00017330550918196995,
      "loss": 0.9595,
      "step": 1620
    },
    {
      "epoch": 0.3477523324851569,
      "grad_norm": 2.4099292755126953,
      "learning_rate": 0.00017297161936560935,
      "loss": 0.9631,
      "step": 1640
    },
    {
      "epoch": 0.3519932145886344,
      "grad_norm": 3.359679698944092,
      "learning_rate": 0.00017263772954924874,
      "loss": 0.9178,
      "step": 1660
    },
    {
      "epoch": 0.356234096692112,
      "grad_norm": 2.013239860534668,
      "learning_rate": 0.00017230383973288816,
      "loss": 0.9626,
      "step": 1680
    },
    {
      "epoch": 0.3604749787955895,
      "grad_norm": 1.6444381475448608,
      "learning_rate": 0.00017196994991652755,
      "loss": 0.9803,
      "step": 1700
    },
    {
      "epoch": 0.364715860899067,
      "grad_norm": 2.0000648498535156,
      "learning_rate": 0.00017163606010016694,
      "loss": 0.9305,
      "step": 1720
    },
    {
      "epoch": 0.36895674300254455,
      "grad_norm": 1.8234150409698486,
      "learning_rate": 0.00017130217028380636,
      "loss": 0.9241,
      "step": 1740
    },
    {
      "epoch": 0.37319762510602206,
      "grad_norm": 1.971179485321045,
      "learning_rate": 0.00017096828046744576,
      "loss": 0.9001,
      "step": 1760
    },
    {
      "epoch": 0.37743850720949956,
      "grad_norm": 1.9195438623428345,
      "learning_rate": 0.00017063439065108515,
      "loss": 0.887,
      "step": 1780
    },
    {
      "epoch": 0.3816793893129771,
      "grad_norm": 1.79142427444458,
      "learning_rate": 0.00017030050083472454,
      "loss": 0.9172,
      "step": 1800
    },
    {
      "epoch": 0.3816793893129771,
      "eval_loss": NaN,
      "eval_runtime": 37.9666,
      "eval_samples_per_second": 26.339,
      "eval_steps_per_second": 3.292,
      "step": 1800
    },
    {
      "epoch": 0.38592027141645463,
      "grad_norm": 2.0568130016326904,
      "learning_rate": 0.00016996661101836396,
      "loss": 1.0137,
      "step": 1820
    },
    {
      "epoch": 0.39016115351993214,
      "grad_norm": 2.3543286323547363,
      "learning_rate": 0.00016963272120200336,
      "loss": 0.9284,
      "step": 1840
    },
    {
      "epoch": 0.3944020356234097,
      "grad_norm": 2.310962677001953,
      "learning_rate": 0.00016929883138564275,
      "loss": 0.9116,
      "step": 1860
    },
    {
      "epoch": 0.3986429177268872,
      "grad_norm": 2.118211269378662,
      "learning_rate": 0.00016896494156928214,
      "loss": 0.9175,
      "step": 1880
    },
    {
      "epoch": 0.4028837998303647,
      "grad_norm": 1.8690170049667358,
      "learning_rate": 0.00016863105175292156,
      "loss": 0.8685,
      "step": 1900
    },
    {
      "epoch": 0.4071246819338422,
      "grad_norm": 2.0020463466644287,
      "learning_rate": 0.00016829716193656093,
      "loss": 0.8749,
      "step": 1920
    },
    {
      "epoch": 0.4113655640373198,
      "grad_norm": 2.3309547901153564,
      "learning_rate": 0.00016796327212020032,
      "loss": 0.8613,
      "step": 1940
    },
    {
      "epoch": 0.4156064461407973,
      "grad_norm": 2.250720262527466,
      "learning_rate": 0.00016762938230383974,
      "loss": 0.8637,
      "step": 1960
    },
    {
      "epoch": 0.4198473282442748,
      "grad_norm": 2.1057088375091553,
      "learning_rate": 0.00016729549248747913,
      "loss": 0.8546,
      "step": 1980
    },
    {
      "epoch": 0.42408821034775235,
      "grad_norm": 2.093207597732544,
      "learning_rate": 0.00016696160267111853,
      "loss": 0.8296,
      "step": 2000
    },
    {
      "epoch": 0.42408821034775235,
      "eval_loss": NaN,
      "eval_runtime": 37.9941,
      "eval_samples_per_second": 26.32,
      "eval_steps_per_second": 3.29,
      "step": 2000
    },
    {
      "epoch": 0.42832909245122985,
      "grad_norm": 2.342036247253418,
      "learning_rate": 0.00016664440734557597,
      "loss": 0.8409,
      "step": 2020
    },
    {
      "epoch": 0.43256997455470736,
      "grad_norm": 2.3811817169189453,
      "learning_rate": 0.00016631051752921536,
      "loss": 0.8652,
      "step": 2040
    },
    {
      "epoch": 0.4368108566581849,
      "grad_norm": 2.5953927040100098,
      "learning_rate": 0.00016597662771285476,
      "loss": 0.8477,
      "step": 2060
    },
    {
      "epoch": 0.4410517387616624,
      "grad_norm": 2.4263832569122314,
      "learning_rate": 0.00016564273789649415,
      "loss": 0.8568,
      "step": 2080
    },
    {
      "epoch": 0.44529262086513993,
      "grad_norm": 2.08699369430542,
      "learning_rate": 0.00016530884808013357,
      "loss": 0.7701,
      "step": 2100
    },
    {
      "epoch": 0.4495335029686175,
      "grad_norm": 2.477623462677002,
      "learning_rate": 0.00016497495826377296,
      "loss": 0.8086,
      "step": 2120
    },
    {
      "epoch": 0.453774385072095,
      "grad_norm": 3.03116512298584,
      "learning_rate": 0.00016464106844741235,
      "loss": 0.846,
      "step": 2140
    },
    {
      "epoch": 0.4580152671755725,
      "grad_norm": 2.2469892501831055,
      "learning_rate": 0.00016430717863105177,
      "loss": 0.8077,
      "step": 2160
    },
    {
      "epoch": 0.46225614927905007,
      "grad_norm": 2.609234094619751,
      "learning_rate": 0.0001639899833055092,
      "loss": 0.7277,
      "step": 2180
    },
    {
      "epoch": 0.46649703138252757,
      "grad_norm": 3.4459376335144043,
      "learning_rate": 0.00016365609348914859,
      "loss": 0.7989,
      "step": 2200
    },
    {
      "epoch": 0.46649703138252757,
      "eval_loss": NaN,
      "eval_runtime": 37.8877,
      "eval_samples_per_second": 26.394,
      "eval_steps_per_second": 3.299,
      "step": 2200
    },
    {
      "epoch": 0.4707379134860051,
      "grad_norm": 2.198207378387451,
      "learning_rate": 0.00016332220367278798,
      "loss": 0.7732,
      "step": 2220
    },
    {
      "epoch": 0.47497879558948264,
      "grad_norm": 2.7479138374328613,
      "learning_rate": 0.0001629883138564274,
      "loss": 0.7897,
      "step": 2240
    },
    {
      "epoch": 0.47921967769296014,
      "grad_norm": 2.0409023761749268,
      "learning_rate": 0.0001626544240400668,
      "loss": 0.7741,
      "step": 2260
    },
    {
      "epoch": 0.48346055979643765,
      "grad_norm": 2.71423077583313,
      "learning_rate": 0.00016232053422370618,
      "loss": 0.7862,
      "step": 2280
    },
    {
      "epoch": 0.48770144189991516,
      "grad_norm": 2.3507261276245117,
      "learning_rate": 0.00016198664440734558,
      "loss": 0.7562,
      "step": 2300
    },
    {
      "epoch": 0.4919423240033927,
      "grad_norm": 2.036689281463623,
      "learning_rate": 0.000161652754590985,
      "loss": 0.754,
      "step": 2320
    },
    {
      "epoch": 0.4961832061068702,
      "grad_norm": 3.0369062423706055,
      "learning_rate": 0.0001613188647746244,
      "loss": 0.7449,
      "step": 2340
    },
    {
      "epoch": 0.5004240882103478,
      "grad_norm": 2.6256980895996094,
      "learning_rate": 0.00016098497495826378,
      "loss": 0.7403,
      "step": 2360
    },
    {
      "epoch": 0.5046649703138253,
      "grad_norm": 2.3012733459472656,
      "learning_rate": 0.0001606510851419032,
      "loss": 0.7439,
      "step": 2380
    },
    {
      "epoch": 0.5089058524173028,
      "grad_norm": 2.7419307231903076,
      "learning_rate": 0.00016031719532554257,
      "loss": 0.743,
      "step": 2400
    },
    {
      "epoch": 0.5089058524173028,
      "eval_loss": NaN,
      "eval_runtime": 37.8655,
      "eval_samples_per_second": 26.409,
      "eval_steps_per_second": 3.301,
      "step": 2400
    },
    {
      "epoch": 0.5131467345207803,
      "grad_norm": 1.889102816581726,
      "learning_rate": 0.00015998330550918196,
      "loss": 0.738,
      "step": 2420
    },
    {
      "epoch": 0.5173876166242578,
      "grad_norm": 2.076204299926758,
      "learning_rate": 0.00015964941569282138,
      "loss": 0.7599,
      "step": 2440
    },
    {
      "epoch": 0.5216284987277354,
      "grad_norm": 2.6748721599578857,
      "learning_rate": 0.00015931552587646077,
      "loss": 0.7207,
      "step": 2460
    },
    {
      "epoch": 0.5258693808312129,
      "grad_norm": 2.8571300506591797,
      "learning_rate": 0.00015898163606010017,
      "loss": 0.7226,
      "step": 2480
    },
    {
      "epoch": 0.5301102629346904,
      "grad_norm": 2.2073874473571777,
      "learning_rate": 0.00015864774624373956,
      "loss": 0.7187,
      "step": 2500
    },
    {
      "epoch": 0.5343511450381679,
      "grad_norm": 3.2114593982696533,
      "learning_rate": 0.00015831385642737898,
      "loss": 0.6847,
      "step": 2520
    },
    {
      "epoch": 0.5385920271416454,
      "grad_norm": 2.41007661819458,
      "learning_rate": 0.00015797996661101837,
      "loss": 0.6738,
      "step": 2540
    },
    {
      "epoch": 0.542832909245123,
      "grad_norm": 1.9356921911239624,
      "learning_rate": 0.00015764607679465777,
      "loss": 0.7382,
      "step": 2560
    },
    {
      "epoch": 0.5470737913486005,
      "grad_norm": 2.4597182273864746,
      "learning_rate": 0.00015731218697829716,
      "loss": 0.675,
      "step": 2580
    },
    {
      "epoch": 0.5513146734520781,
      "grad_norm": 2.7616782188415527,
      "learning_rate": 0.00015697829716193658,
      "loss": 0.6616,
      "step": 2600
    },
    {
      "epoch": 0.5513146734520781,
      "eval_loss": NaN,
      "eval_runtime": 37.9594,
      "eval_samples_per_second": 26.344,
      "eval_steps_per_second": 3.293,
      "step": 2600
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 2.227827787399292,
      "learning_rate": 0.00015664440734557597,
      "loss": 0.7212,
      "step": 2620
    },
    {
      "epoch": 0.5597964376590331,
      "grad_norm": 2.134922981262207,
      "learning_rate": 0.00015631051752921536,
      "loss": 0.6918,
      "step": 2640
    },
    {
      "epoch": 0.5640373197625106,
      "grad_norm": 2.2440874576568604,
      "learning_rate": 0.00015597662771285478,
      "loss": 0.6634,
      "step": 2660
    },
    {
      "epoch": 0.5682782018659881,
      "grad_norm": 2.5504469871520996,
      "learning_rate": 0.00015564273789649418,
      "loss": 0.677,
      "step": 2680
    },
    {
      "epoch": 0.5725190839694656,
      "grad_norm": 2.585655689239502,
      "learning_rate": 0.00015530884808013357,
      "loss": 0.642,
      "step": 2700
    },
    {
      "epoch": 0.5767599660729432,
      "grad_norm": 3.3703689575195312,
      "learning_rate": 0.00015497495826377296,
      "loss": 0.7237,
      "step": 2720
    },
    {
      "epoch": 0.5810008481764207,
      "grad_norm": 3.0335161685943604,
      "learning_rate": 0.00015464106844741236,
      "loss": 0.6499,
      "step": 2740
    },
    {
      "epoch": 0.5852417302798982,
      "grad_norm": 1.6910688877105713,
      "learning_rate": 0.00015430717863105175,
      "loss": 0.6663,
      "step": 2760
    },
    {
      "epoch": 0.5894826123833757,
      "grad_norm": 2.1536941528320312,
      "learning_rate": 0.00015397328881469114,
      "loss": 0.7011,
      "step": 2780
    },
    {
      "epoch": 0.5937234944868532,
      "grad_norm": 2.1281566619873047,
      "learning_rate": 0.00015363939899833056,
      "loss": 0.6119,
      "step": 2800
    },
    {
      "epoch": 0.5937234944868532,
      "eval_loss": NaN,
      "eval_runtime": 37.7825,
      "eval_samples_per_second": 26.467,
      "eval_steps_per_second": 3.308,
      "step": 2800
    },
    {
      "epoch": 0.5979643765903307,
      "grad_norm": 3.287522315979004,
      "learning_rate": 0.00015330550918196995,
      "loss": 0.6579,
      "step": 2820
    },
    {
      "epoch": 0.6022052586938084,
      "grad_norm": 2.5627098083496094,
      "learning_rate": 0.00015297161936560935,
      "loss": 0.637,
      "step": 2840
    },
    {
      "epoch": 0.6064461407972859,
      "grad_norm": 2.8666200637817383,
      "learning_rate": 0.00015263772954924877,
      "loss": 0.6411,
      "step": 2860
    },
    {
      "epoch": 0.6106870229007634,
      "grad_norm": 3.079150438308716,
      "learning_rate": 0.00015230383973288816,
      "loss": 0.642,
      "step": 2880
    },
    {
      "epoch": 0.6149279050042409,
      "grad_norm": 2.3377933502197266,
      "learning_rate": 0.00015196994991652755,
      "loss": 0.6171,
      "step": 2900
    },
    {
      "epoch": 0.6191687871077184,
      "grad_norm": 3.3924803733825684,
      "learning_rate": 0.00015163606010016695,
      "loss": 0.5948,
      "step": 2920
    },
    {
      "epoch": 0.6234096692111959,
      "grad_norm": 2.9143242835998535,
      "learning_rate": 0.00015130217028380637,
      "loss": 0.6805,
      "step": 2940
    },
    {
      "epoch": 0.6276505513146734,
      "grad_norm": 2.6952388286590576,
      "learning_rate": 0.00015096828046744576,
      "loss": 0.6483,
      "step": 2960
    },
    {
      "epoch": 0.631891433418151,
      "grad_norm": 2.9229159355163574,
      "learning_rate": 0.00015063439065108515,
      "loss": 0.5956,
      "step": 2980
    },
    {
      "epoch": 0.6361323155216285,
      "grad_norm": 2.964066505432129,
      "learning_rate": 0.00015030050083472454,
      "loss": 0.619,
      "step": 3000
    },
    {
      "epoch": 0.6361323155216285,
      "eval_loss": NaN,
      "eval_runtime": 37.9845,
      "eval_samples_per_second": 26.327,
      "eval_steps_per_second": 3.291,
      "step": 3000
    },
    {
      "epoch": 0.640373197625106,
      "grad_norm": 3.002941846847534,
      "learning_rate": 0.00014996661101836396,
      "loss": 0.6143,
      "step": 3020
    },
    {
      "epoch": 0.6446140797285835,
      "grad_norm": 3.1462719440460205,
      "learning_rate": 0.00014963272120200336,
      "loss": 0.5858,
      "step": 3040
    },
    {
      "epoch": 0.648854961832061,
      "grad_norm": 2.8309357166290283,
      "learning_rate": 0.00014929883138564272,
      "loss": 0.6158,
      "step": 3060
    },
    {
      "epoch": 0.6530958439355385,
      "grad_norm": 3.1488876342773438,
      "learning_rate": 0.00014896494156928214,
      "loss": 0.6091,
      "step": 3080
    },
    {
      "epoch": 0.6573367260390162,
      "grad_norm": 3.0378949642181396,
      "learning_rate": 0.00014863105175292154,
      "loss": 0.6128,
      "step": 3100
    },
    {
      "epoch": 0.6615776081424937,
      "grad_norm": 1.9123942852020264,
      "learning_rate": 0.00014829716193656093,
      "loss": 0.5913,
      "step": 3120
    },
    {
      "epoch": 0.6658184902459712,
      "grad_norm": 2.844385862350464,
      "learning_rate": 0.00014796327212020035,
      "loss": 0.5794,
      "step": 3140
    },
    {
      "epoch": 0.6700593723494487,
      "grad_norm": 2.8968617916107178,
      "learning_rate": 0.00014762938230383974,
      "loss": 0.5974,
      "step": 3160
    },
    {
      "epoch": 0.6743002544529262,
      "grad_norm": 2.963021993637085,
      "learning_rate": 0.00014729549248747913,
      "loss": 0.578,
      "step": 3180
    },
    {
      "epoch": 0.6785411365564037,
      "grad_norm": 2.493708610534668,
      "learning_rate": 0.00014696160267111853,
      "loss": 0.6086,
      "step": 3200
    },
    {
      "epoch": 0.6785411365564037,
      "eval_loss": NaN,
      "eval_runtime": 37.8206,
      "eval_samples_per_second": 26.441,
      "eval_steps_per_second": 3.305,
      "step": 3200
    },
    {
      "epoch": 0.6827820186598813,
      "grad_norm": 2.5819506645202637,
      "learning_rate": 0.00014662771285475795,
      "loss": 0.5927,
      "step": 3220
    },
    {
      "epoch": 0.6870229007633588,
      "grad_norm": 2.7373132705688477,
      "learning_rate": 0.00014629382303839734,
      "loss": 0.5703,
      "step": 3240
    },
    {
      "epoch": 0.6912637828668363,
      "grad_norm": 2.321910858154297,
      "learning_rate": 0.00014595993322203673,
      "loss": 0.5988,
      "step": 3260
    },
    {
      "epoch": 0.6955046649703138,
      "grad_norm": 3.2021353244781494,
      "learning_rate": 0.00014562604340567613,
      "loss": 0.5932,
      "step": 3280
    },
    {
      "epoch": 0.6997455470737913,
      "grad_norm": 2.8209946155548096,
      "learning_rate": 0.00014529215358931555,
      "loss": 0.5952,
      "step": 3300
    },
    {
      "epoch": 0.7039864291772688,
      "grad_norm": 2.8019261360168457,
      "learning_rate": 0.00014495826377295494,
      "loss": 0.5452,
      "step": 3320
    },
    {
      "epoch": 0.7082273112807463,
      "grad_norm": 2.4001333713531494,
      "learning_rate": 0.00014462437395659433,
      "loss": 0.5632,
      "step": 3340
    },
    {
      "epoch": 0.712468193384224,
      "grad_norm": 2.9871296882629395,
      "learning_rate": 0.00014429048414023375,
      "loss": 0.5036,
      "step": 3360
    },
    {
      "epoch": 0.7167090754877015,
      "grad_norm": 1.9114025831222534,
      "learning_rate": 0.00014395659432387314,
      "loss": 0.5174,
      "step": 3380
    },
    {
      "epoch": 0.720949957591179,
      "grad_norm": 2.019404172897339,
      "learning_rate": 0.0001436227045075125,
      "loss": 0.5521,
      "step": 3400
    },
    {
      "epoch": 0.720949957591179,
      "eval_loss": NaN,
      "eval_runtime": 37.8434,
      "eval_samples_per_second": 26.425,
      "eval_steps_per_second": 3.303,
      "step": 3400
    },
    {
      "epoch": 0.7251908396946565,
      "grad_norm": 2.5788028240203857,
      "learning_rate": 0.00014328881469115193,
      "loss": 0.5601,
      "step": 3420
    },
    {
      "epoch": 0.729431721798134,
      "grad_norm": 2.6271934509277344,
      "learning_rate": 0.00014295492487479132,
      "loss": 0.5742,
      "step": 3440
    },
    {
      "epoch": 0.7336726039016115,
      "grad_norm": 2.9399425983428955,
      "learning_rate": 0.00014262103505843072,
      "loss": 0.5917,
      "step": 3460
    },
    {
      "epoch": 0.7379134860050891,
      "grad_norm": 2.6985580921173096,
      "learning_rate": 0.0001422871452420701,
      "loss": 0.5729,
      "step": 3480
    },
    {
      "epoch": 0.7421543681085666,
      "grad_norm": 2.4061472415924072,
      "learning_rate": 0.00014195325542570953,
      "loss": 0.5379,
      "step": 3500
    },
    {
      "epoch": 0.7463952502120441,
      "grad_norm": 2.4237020015716553,
      "learning_rate": 0.00014161936560934892,
      "loss": 0.5315,
      "step": 3520
    },
    {
      "epoch": 0.7506361323155216,
      "grad_norm": 2.8194217681884766,
      "learning_rate": 0.00014128547579298831,
      "loss": 0.5216,
      "step": 3540
    },
    {
      "epoch": 0.7548770144189991,
      "grad_norm": 2.7585253715515137,
      "learning_rate": 0.0001409515859766277,
      "loss": 0.5046,
      "step": 3560
    },
    {
      "epoch": 0.7591178965224766,
      "grad_norm": 2.255218505859375,
      "learning_rate": 0.00014061769616026713,
      "loss": 0.5122,
      "step": 3580
    },
    {
      "epoch": 0.7633587786259542,
      "grad_norm": 2.4961671829223633,
      "learning_rate": 0.00014028380634390652,
      "loss": 0.5491,
      "step": 3600
    },
    {
      "epoch": 0.7633587786259542,
      "eval_loss": NaN,
      "eval_runtime": 38.0456,
      "eval_samples_per_second": 26.284,
      "eval_steps_per_second": 3.286,
      "step": 3600
    },
    {
      "epoch": 0.7675996607294318,
      "grad_norm": 2.435004472732544,
      "learning_rate": 0.0001399499165275459,
      "loss": 0.5383,
      "step": 3620
    },
    {
      "epoch": 0.7718405428329093,
      "grad_norm": 2.876791000366211,
      "learning_rate": 0.00013961602671118533,
      "loss": 0.5172,
      "step": 3640
    },
    {
      "epoch": 0.7760814249363868,
      "grad_norm": 2.4962687492370605,
      "learning_rate": 0.00013928213689482473,
      "loss": 0.5853,
      "step": 3660
    },
    {
      "epoch": 0.7803223070398643,
      "grad_norm": 3.0524990558624268,
      "learning_rate": 0.00013894824707846412,
      "loss": 0.5266,
      "step": 3680
    },
    {
      "epoch": 0.7845631891433418,
      "grad_norm": 3.0947062969207764,
      "learning_rate": 0.0001386143572621035,
      "loss": 0.54,
      "step": 3700
    },
    {
      "epoch": 0.7888040712468194,
      "grad_norm": 2.5288310050964355,
      "learning_rate": 0.0001382804674457429,
      "loss": 0.5169,
      "step": 3720
    },
    {
      "epoch": 0.7930449533502969,
      "grad_norm": 2.1172380447387695,
      "learning_rate": 0.0001379465776293823,
      "loss": 0.5054,
      "step": 3740
    },
    {
      "epoch": 0.7972858354537744,
      "grad_norm": 2.9825069904327393,
      "learning_rate": 0.0001376126878130217,
      "loss": 0.4899,
      "step": 3760
    },
    {
      "epoch": 0.8015267175572519,
      "grad_norm": 2.3686017990112305,
      "learning_rate": 0.0001372787979966611,
      "loss": 0.5452,
      "step": 3780
    },
    {
      "epoch": 0.8057675996607294,
      "grad_norm": 2.4265174865722656,
      "learning_rate": 0.0001369449081803005,
      "loss": 0.5123,
      "step": 3800
    },
    {
      "epoch": 0.8057675996607294,
      "eval_loss": NaN,
      "eval_runtime": 38.0721,
      "eval_samples_per_second": 26.266,
      "eval_steps_per_second": 3.283,
      "step": 3800
    },
    {
      "epoch": 0.8100084817642069,
      "grad_norm": 3.1802978515625,
      "learning_rate": 0.0001366110183639399,
      "loss": 0.4792,
      "step": 3820
    },
    {
      "epoch": 0.8142493638676844,
      "grad_norm": 3.5329525470733643,
      "learning_rate": 0.00013627712854757932,
      "loss": 0.4819,
      "step": 3840
    },
    {
      "epoch": 0.818490245971162,
      "grad_norm": 2.5517141819000244,
      "learning_rate": 0.0001359432387312187,
      "loss": 0.4534,
      "step": 3860
    },
    {
      "epoch": 0.8227311280746396,
      "grad_norm": 2.1168999671936035,
      "learning_rate": 0.0001356093489148581,
      "loss": 0.4557,
      "step": 3880
    },
    {
      "epoch": 0.8269720101781171,
      "grad_norm": 2.714913845062256,
      "learning_rate": 0.0001352754590984975,
      "loss": 0.4422,
      "step": 3900
    },
    {
      "epoch": 0.8312128922815946,
      "grad_norm": 2.2053091526031494,
      "learning_rate": 0.00013494156928213691,
      "loss": 0.5149,
      "step": 3920
    },
    {
      "epoch": 0.8354537743850721,
      "grad_norm": 2.389634370803833,
      "learning_rate": 0.0001346076794657763,
      "loss": 0.4453,
      "step": 3940
    },
    {
      "epoch": 0.8396946564885496,
      "grad_norm": 3.0144214630126953,
      "learning_rate": 0.0001342737896494157,
      "loss": 0.5246,
      "step": 3960
    },
    {
      "epoch": 0.8439355385920272,
      "grad_norm": 3.4380931854248047,
      "learning_rate": 0.0001339398998330551,
      "loss": 0.5303,
      "step": 3980
    },
    {
      "epoch": 0.8481764206955047,
      "grad_norm": 3.7437222003936768,
      "learning_rate": 0.0001336060100166945,
      "loss": 0.4912,
      "step": 4000
    },
    {
      "epoch": 0.8481764206955047,
      "eval_loss": NaN,
      "eval_runtime": 38.0261,
      "eval_samples_per_second": 26.298,
      "eval_steps_per_second": 3.287,
      "step": 4000
    },
    {
      "epoch": 0.8524173027989822,
      "grad_norm": 2.9163849353790283,
      "learning_rate": 0.0001332721202003339,
      "loss": 0.5178,
      "step": 4020
    },
    {
      "epoch": 0.8566581849024597,
      "grad_norm": 2.850667953491211,
      "learning_rate": 0.0001329382303839733,
      "loss": 0.5158,
      "step": 4040
    },
    {
      "epoch": 0.8608990670059372,
      "grad_norm": 2.959547281265259,
      "learning_rate": 0.0001326043405676127,
      "loss": 0.4744,
      "step": 4060
    },
    {
      "epoch": 0.8651399491094147,
      "grad_norm": 2.9672493934631348,
      "learning_rate": 0.00013227045075125208,
      "loss": 0.4545,
      "step": 4080
    },
    {
      "epoch": 0.8693808312128923,
      "grad_norm": 2.6477742195129395,
      "learning_rate": 0.00013193656093489148,
      "loss": 0.5046,
      "step": 4100
    },
    {
      "epoch": 0.8736217133163698,
      "grad_norm": 2.9155495166778564,
      "learning_rate": 0.0001316026711185309,
      "loss": 0.4875,
      "step": 4120
    },
    {
      "epoch": 0.8778625954198473,
      "grad_norm": 3.254493474960327,
      "learning_rate": 0.0001312687813021703,
      "loss": 0.5038,
      "step": 4140
    },
    {
      "epoch": 0.8821034775233249,
      "grad_norm": 2.727482557296753,
      "learning_rate": 0.00013093489148580968,
      "loss": 0.4758,
      "step": 4160
    },
    {
      "epoch": 0.8863443596268024,
      "grad_norm": 2.6975488662719727,
      "learning_rate": 0.00013060100166944908,
      "loss": 0.4379,
      "step": 4180
    },
    {
      "epoch": 0.8905852417302799,
      "grad_norm": 2.36287784576416,
      "learning_rate": 0.0001302671118530885,
      "loss": 0.4682,
      "step": 4200
    },
    {
      "epoch": 0.8905852417302799,
      "eval_loss": NaN,
      "eval_runtime": 37.8352,
      "eval_samples_per_second": 26.43,
      "eval_steps_per_second": 3.304,
      "step": 4200
    },
    {
      "epoch": 0.8948261238337574,
      "grad_norm": 2.4240024089813232,
      "learning_rate": 0.0001299332220367279,
      "loss": 0.4486,
      "step": 4220
    },
    {
      "epoch": 0.899067005937235,
      "grad_norm": 2.734173536300659,
      "learning_rate": 0.00012959933222036728,
      "loss": 0.4798,
      "step": 4240
    },
    {
      "epoch": 0.9033078880407125,
      "grad_norm": 2.919610023498535,
      "learning_rate": 0.0001292821368948247,
      "loss": 0.4764,
      "step": 4260
    },
    {
      "epoch": 0.90754877014419,
      "grad_norm": 2.2606446743011475,
      "learning_rate": 0.00012894824707846412,
      "loss": 0.4436,
      "step": 4280
    },
    {
      "epoch": 0.9117896522476675,
      "grad_norm": 3.0298829078674316,
      "learning_rate": 0.0001286143572621035,
      "loss": 0.41,
      "step": 4300
    },
    {
      "epoch": 0.916030534351145,
      "grad_norm": 3.4316773414611816,
      "learning_rate": 0.0001282804674457429,
      "loss": 0.4629,
      "step": 4320
    },
    {
      "epoch": 0.9202714164546225,
      "grad_norm": 3.0519142150878906,
      "learning_rate": 0.00012794657762938233,
      "loss": 0.4743,
      "step": 4340
    },
    {
      "epoch": 0.9245122985581001,
      "grad_norm": 3.2025558948516846,
      "learning_rate": 0.00012761268781302172,
      "loss": 0.3862,
      "step": 4360
    },
    {
      "epoch": 0.9287531806615776,
      "grad_norm": 3.377744436264038,
      "learning_rate": 0.0001272787979966611,
      "loss": 0.4513,
      "step": 4380
    },
    {
      "epoch": 0.9329940627650551,
      "grad_norm": 3.2821669578552246,
      "learning_rate": 0.0001269449081803005,
      "loss": 0.4679,
      "step": 4400
    },
    {
      "epoch": 0.9329940627650551,
      "eval_loss": NaN,
      "eval_runtime": 38.0614,
      "eval_samples_per_second": 26.273,
      "eval_steps_per_second": 3.284,
      "step": 4400
    },
    {
      "epoch": 0.9372349448685326,
      "grad_norm": 2.8143436908721924,
      "learning_rate": 0.00012661101836393992,
      "loss": 0.3916,
      "step": 4420
    },
    {
      "epoch": 0.9414758269720102,
      "grad_norm": 2.3522398471832275,
      "learning_rate": 0.00012627712854757932,
      "loss": 0.4765,
      "step": 4440
    },
    {
      "epoch": 0.9457167090754877,
      "grad_norm": 2.471167802810669,
      "learning_rate": 0.00012594323873121868,
      "loss": 0.4298,
      "step": 4460
    },
    {
      "epoch": 0.9499575911789653,
      "grad_norm": 2.6195333003997803,
      "learning_rate": 0.0001256093489148581,
      "loss": 0.4234,
      "step": 4480
    },
    {
      "epoch": 0.9541984732824428,
      "grad_norm": 2.5602707862854004,
      "learning_rate": 0.0001252754590984975,
      "loss": 0.4108,
      "step": 4500
    },
    {
      "epoch": 0.9584393553859203,
      "grad_norm": 3.301478147506714,
      "learning_rate": 0.0001249415692821369,
      "loss": 0.4493,
      "step": 4520
    },
    {
      "epoch": 0.9626802374893978,
      "grad_norm": 2.927687406539917,
      "learning_rate": 0.0001246076794657763,
      "loss": 0.4196,
      "step": 4540
    },
    {
      "epoch": 0.9669211195928753,
      "grad_norm": 2.7514328956604004,
      "learning_rate": 0.0001242737896494157,
      "loss": 0.459,
      "step": 4560
    },
    {
      "epoch": 0.9711620016963528,
      "grad_norm": 2.1768290996551514,
      "learning_rate": 0.0001239398998330551,
      "loss": 0.4488,
      "step": 4580
    },
    {
      "epoch": 0.9754028837998303,
      "grad_norm": 3.429978132247925,
      "learning_rate": 0.0001236060100166945,
      "loss": 0.4259,
      "step": 4600
    },
    {
      "epoch": 0.9754028837998303,
      "eval_loss": NaN,
      "eval_runtime": 38.103,
      "eval_samples_per_second": 26.245,
      "eval_steps_per_second": 3.281,
      "step": 4600
    },
    {
      "epoch": 0.9796437659033079,
      "grad_norm": 2.6817071437835693,
      "learning_rate": 0.0001232721202003339,
      "loss": 0.4057,
      "step": 4620
    },
    {
      "epoch": 0.9838846480067854,
      "grad_norm": 2.703551769256592,
      "learning_rate": 0.0001229382303839733,
      "loss": 0.4121,
      "step": 4640
    },
    {
      "epoch": 0.9881255301102629,
      "grad_norm": 2.28878116607666,
      "learning_rate": 0.0001226043405676127,
      "loss": 0.4275,
      "step": 4660
    },
    {
      "epoch": 0.9923664122137404,
      "grad_norm": 2.7061429023742676,
      "learning_rate": 0.00012227045075125209,
      "loss": 0.3971,
      "step": 4680
    },
    {
      "epoch": 0.996607294317218,
      "grad_norm": 3.0262553691864014,
      "learning_rate": 0.00012193656093489149,
      "loss": 0.4466,
      "step": 4700
    },
    {
      "epoch": 1.0008481764206956,
      "grad_norm": 2.8571345806121826,
      "learning_rate": 0.0001216026711185309,
      "loss": 0.4158,
      "step": 4720
    },
    {
      "epoch": 1.005089058524173,
      "grad_norm": 2.682584762573242,
      "learning_rate": 0.00012126878130217029,
      "loss": 0.4008,
      "step": 4740
    },
    {
      "epoch": 1.0093299406276506,
      "grad_norm": 2.6532816886901855,
      "learning_rate": 0.0001209348914858097,
      "loss": 0.4246,
      "step": 4760
    },
    {
      "epoch": 1.013570822731128,
      "grad_norm": 3.1518845558166504,
      "learning_rate": 0.0001206010016694491,
      "loss": 0.3965,
      "step": 4780
    },
    {
      "epoch": 1.0178117048346056,
      "grad_norm": 3.080641031265259,
      "learning_rate": 0.00012026711185308848,
      "loss": 0.3847,
      "step": 4800
    },
    {
      "epoch": 1.0178117048346056,
      "eval_loss": NaN,
      "eval_runtime": 37.983,
      "eval_samples_per_second": 26.328,
      "eval_steps_per_second": 3.291,
      "step": 4800
    },
    {
      "epoch": 1.0220525869380832,
      "grad_norm": 2.2755777835845947,
      "learning_rate": 0.00011993322203672788,
      "loss": 0.3943,
      "step": 4820
    },
    {
      "epoch": 1.0262934690415606,
      "grad_norm": 3.423597574234009,
      "learning_rate": 0.00011959933222036728,
      "loss": 0.3909,
      "step": 4840
    },
    {
      "epoch": 1.0305343511450382,
      "grad_norm": 2.2315890789031982,
      "learning_rate": 0.00011926544240400668,
      "loss": 0.4006,
      "step": 4860
    },
    {
      "epoch": 1.0347752332485156,
      "grad_norm": 3.244436740875244,
      "learning_rate": 0.00011893155258764608,
      "loss": 0.4294,
      "step": 4880
    },
    {
      "epoch": 1.0390161153519932,
      "grad_norm": 2.940098285675049,
      "learning_rate": 0.00011859766277128547,
      "loss": 0.4002,
      "step": 4900
    },
    {
      "epoch": 1.0432569974554706,
      "grad_norm": 2.1194727420806885,
      "learning_rate": 0.00011826377295492488,
      "loss": 0.424,
      "step": 4920
    },
    {
      "epoch": 1.0474978795589482,
      "grad_norm": 2.3373231887817383,
      "learning_rate": 0.00011792988313856427,
      "loss": 0.414,
      "step": 4940
    },
    {
      "epoch": 1.0517387616624259,
      "grad_norm": 2.886240243911743,
      "learning_rate": 0.00011759599332220368,
      "loss": 0.4024,
      "step": 4960
    },
    {
      "epoch": 1.0559796437659033,
      "grad_norm": 2.6492719650268555,
      "learning_rate": 0.00011726210350584307,
      "loss": 0.3853,
      "step": 4980
    },
    {
      "epoch": 1.0602205258693809,
      "grad_norm": 2.861229181289673,
      "learning_rate": 0.00011692821368948248,
      "loss": 0.3927,
      "step": 5000
    },
    {
      "epoch": 1.0602205258693809,
      "eval_loss": NaN,
      "eval_runtime": 37.8335,
      "eval_samples_per_second": 26.432,
      "eval_steps_per_second": 3.304,
      "step": 5000
    },
    {
      "epoch": 1.0644614079728583,
      "grad_norm": 2.7827229499816895,
      "learning_rate": 0.00011659432387312189,
      "loss": 0.3816,
      "step": 5020
    },
    {
      "epoch": 1.0687022900763359,
      "grad_norm": 2.5567145347595215,
      "learning_rate": 0.00011626043405676128,
      "loss": 0.3371,
      "step": 5040
    },
    {
      "epoch": 1.0729431721798135,
      "grad_norm": 2.503917694091797,
      "learning_rate": 0.00011592654424040069,
      "loss": 0.3806,
      "step": 5060
    },
    {
      "epoch": 1.077184054283291,
      "grad_norm": 2.6187195777893066,
      "learning_rate": 0.00011559265442404008,
      "loss": 0.3837,
      "step": 5080
    },
    {
      "epoch": 1.0814249363867685,
      "grad_norm": 2.5440361499786377,
      "learning_rate": 0.00011525876460767948,
      "loss": 0.3575,
      "step": 5100
    },
    {
      "epoch": 1.085665818490246,
      "grad_norm": 3.2564001083374023,
      "learning_rate": 0.00011492487479131886,
      "loss": 0.3943,
      "step": 5120
    },
    {
      "epoch": 1.0899067005937235,
      "grad_norm": 2.446729898452759,
      "learning_rate": 0.00011459098497495826,
      "loss": 0.4361,
      "step": 5140
    },
    {
      "epoch": 1.094147582697201,
      "grad_norm": 2.33760666847229,
      "learning_rate": 0.00011425709515859766,
      "loss": 0.3763,
      "step": 5160
    },
    {
      "epoch": 1.0983884648006785,
      "grad_norm": 2.830857515335083,
      "learning_rate": 0.00011392320534223706,
      "loss": 0.4351,
      "step": 5180
    },
    {
      "epoch": 1.1026293469041561,
      "grad_norm": 3.329183578491211,
      "learning_rate": 0.00011358931552587646,
      "loss": 0.3902,
      "step": 5200
    },
    {
      "epoch": 1.1026293469041561,
      "eval_loss": NaN,
      "eval_runtime": 37.8232,
      "eval_samples_per_second": 26.439,
      "eval_steps_per_second": 3.305,
      "step": 5200
    },
    {
      "epoch": 1.1068702290076335,
      "grad_norm": 2.8310630321502686,
      "learning_rate": 0.00011325542570951586,
      "loss": 0.3785,
      "step": 5220
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 3.998307943344116,
      "learning_rate": 0.00011292153589315526,
      "loss": 0.3716,
      "step": 5240
    },
    {
      "epoch": 1.1153519932145886,
      "grad_norm": 3.2324347496032715,
      "learning_rate": 0.00011258764607679465,
      "loss": 0.3527,
      "step": 5260
    },
    {
      "epoch": 1.1195928753180662,
      "grad_norm": 2.707636594772339,
      "learning_rate": 0.00011225375626043406,
      "loss": 0.4084,
      "step": 5280
    },
    {
      "epoch": 1.1238337574215436,
      "grad_norm": 3.636577606201172,
      "learning_rate": 0.00011191986644407347,
      "loss": 0.3585,
      "step": 5300
    },
    {
      "epoch": 1.1280746395250212,
      "grad_norm": 3.26069974899292,
      "learning_rate": 0.00011158597662771286,
      "loss": 0.3408,
      "step": 5320
    },
    {
      "epoch": 1.1323155216284988,
      "grad_norm": 2.04370379447937,
      "learning_rate": 0.00011125208681135227,
      "loss": 0.4031,
      "step": 5340
    },
    {
      "epoch": 1.1365564037319762,
      "grad_norm": 2.825923204421997,
      "learning_rate": 0.00011091819699499166,
      "loss": 0.3902,
      "step": 5360
    },
    {
      "epoch": 1.1407972858354538,
      "grad_norm": 2.3939812183380127,
      "learning_rate": 0.00011058430717863107,
      "loss": 0.3264,
      "step": 5380
    },
    {
      "epoch": 1.1450381679389312,
      "grad_norm": 3.194133758544922,
      "learning_rate": 0.00011025041736227046,
      "loss": 0.4127,
      "step": 5400
    },
    {
      "epoch": 1.1450381679389312,
      "eval_loss": NaN,
      "eval_runtime": 37.8179,
      "eval_samples_per_second": 26.442,
      "eval_steps_per_second": 3.305,
      "step": 5400
    },
    {
      "epoch": 1.1492790500424088,
      "grad_norm": 2.159160614013672,
      "learning_rate": 0.00010991652754590987,
      "loss": 0.3401,
      "step": 5420
    },
    {
      "epoch": 1.1535199321458864,
      "grad_norm": 2.3315744400024414,
      "learning_rate": 0.00010958263772954926,
      "loss": 0.3581,
      "step": 5440
    },
    {
      "epoch": 1.1577608142493638,
      "grad_norm": 2.886758804321289,
      "learning_rate": 0.00010924874791318864,
      "loss": 0.3688,
      "step": 5460
    },
    {
      "epoch": 1.1620016963528415,
      "grad_norm": 2.8764262199401855,
      "learning_rate": 0.00010891485809682804,
      "loss": 0.4109,
      "step": 5480
    },
    {
      "epoch": 1.1662425784563188,
      "grad_norm": 2.9258036613464355,
      "learning_rate": 0.00010858096828046744,
      "loss": 0.3782,
      "step": 5500
    },
    {
      "epoch": 1.1704834605597965,
      "grad_norm": 2.862652063369751,
      "learning_rate": 0.00010824707846410684,
      "loss": 0.3443,
      "step": 5520
    },
    {
      "epoch": 1.174724342663274,
      "grad_norm": 3.548119306564331,
      "learning_rate": 0.00010791318864774625,
      "loss": 0.4126,
      "step": 5540
    },
    {
      "epoch": 1.1789652247667515,
      "grad_norm": 1.9414244890213013,
      "learning_rate": 0.00010757929883138564,
      "loss": 0.3207,
      "step": 5560
    },
    {
      "epoch": 1.183206106870229,
      "grad_norm": 3.13726806640625,
      "learning_rate": 0.00010724540901502505,
      "loss": 0.3563,
      "step": 5580
    },
    {
      "epoch": 1.1874469889737065,
      "grad_norm": 3.2025578022003174,
      "learning_rate": 0.00010691151919866444,
      "loss": 0.3629,
      "step": 5600
    },
    {
      "epoch": 1.1874469889737065,
      "eval_loss": NaN,
      "eval_runtime": 37.9491,
      "eval_samples_per_second": 26.351,
      "eval_steps_per_second": 3.294,
      "step": 5600
    },
    {
      "epoch": 1.191687871077184,
      "grad_norm": 2.1612327098846436,
      "learning_rate": 0.00010657762938230385,
      "loss": 0.3432,
      "step": 5620
    },
    {
      "epoch": 1.1959287531806615,
      "grad_norm": 3.0352044105529785,
      "learning_rate": 0.00010624373956594324,
      "loss": 0.3906,
      "step": 5640
    },
    {
      "epoch": 1.2001696352841391,
      "grad_norm": 2.1630747318267822,
      "learning_rate": 0.00010590984974958265,
      "loss": 0.3683,
      "step": 5660
    },
    {
      "epoch": 1.2044105173876165,
      "grad_norm": 3.0407931804656982,
      "learning_rate": 0.00010557595993322204,
      "loss": 0.3153,
      "step": 5680
    },
    {
      "epoch": 1.2086513994910941,
      "grad_norm": 2.7199888229370117,
      "learning_rate": 0.00010524207011686145,
      "loss": 0.3581,
      "step": 5700
    },
    {
      "epoch": 1.2128922815945717,
      "grad_norm": 2.591113328933716,
      "learning_rate": 0.00010490818030050084,
      "loss": 0.3391,
      "step": 5720
    },
    {
      "epoch": 1.2171331636980491,
      "grad_norm": 2.730194091796875,
      "learning_rate": 0.00010457429048414025,
      "loss": 0.3194,
      "step": 5740
    },
    {
      "epoch": 1.2213740458015268,
      "grad_norm": 2.4739983081817627,
      "learning_rate": 0.00010424040066777965,
      "loss": 0.3184,
      "step": 5760
    },
    {
      "epoch": 1.2256149279050041,
      "grad_norm": 2.8976736068725586,
      "learning_rate": 0.00010390651085141905,
      "loss": 0.3895,
      "step": 5780
    },
    {
      "epoch": 1.2298558100084818,
      "grad_norm": 2.584592580795288,
      "learning_rate": 0.00010357262103505843,
      "loss": 0.3753,
      "step": 5800
    },
    {
      "epoch": 1.2298558100084818,
      "eval_loss": NaN,
      "eval_runtime": 38.1162,
      "eval_samples_per_second": 26.236,
      "eval_steps_per_second": 3.279,
      "step": 5800
    },
    {
      "epoch": 1.2340966921119594,
      "grad_norm": 2.542156219482422,
      "learning_rate": 0.00010323873121869783,
      "loss": 0.3697,
      "step": 5820
    },
    {
      "epoch": 1.2383375742154368,
      "grad_norm": 2.29978346824646,
      "learning_rate": 0.00010290484140233722,
      "loss": 0.3638,
      "step": 5840
    },
    {
      "epoch": 1.2425784563189144,
      "grad_norm": 2.7800185680389404,
      "learning_rate": 0.00010257095158597663,
      "loss": 0.3629,
      "step": 5860
    },
    {
      "epoch": 1.2468193384223918,
      "grad_norm": 2.435959577560425,
      "learning_rate": 0.00010223706176961602,
      "loss": 0.3364,
      "step": 5880
    },
    {
      "epoch": 1.2510602205258694,
      "grad_norm": 3.185875177383423,
      "learning_rate": 0.00010190317195325543,
      "loss": 0.3066,
      "step": 5900
    },
    {
      "epoch": 1.255301102629347,
      "grad_norm": 2.650622844696045,
      "learning_rate": 0.00010156928213689482,
      "loss": 0.3624,
      "step": 5920
    },
    {
      "epoch": 1.2595419847328244,
      "grad_norm": 2.265470504760742,
      "learning_rate": 0.00010123539232053423,
      "loss": 0.3641,
      "step": 5940
    },
    {
      "epoch": 1.263782866836302,
      "grad_norm": 2.2070956230163574,
      "learning_rate": 0.00010090150250417362,
      "loss": 0.3221,
      "step": 5960
    },
    {
      "epoch": 1.2680237489397794,
      "grad_norm": 3.2336478233337402,
      "learning_rate": 0.00010056761268781303,
      "loss": 0.3845,
      "step": 5980
    },
    {
      "epoch": 1.272264631043257,
      "grad_norm": 2.9044620990753174,
      "learning_rate": 0.00010023372287145244,
      "loss": 0.3197,
      "step": 6000
    },
    {
      "epoch": 1.272264631043257,
      "eval_loss": NaN,
      "eval_runtime": 38.0767,
      "eval_samples_per_second": 26.263,
      "eval_steps_per_second": 3.283,
      "step": 6000
    },
    {
      "epoch": 1.2765055131467344,
      "grad_norm": 2.5410068035125732,
      "learning_rate": 9.989983305509183e-05,
      "loss": 0.3648,
      "step": 6020
    },
    {
      "epoch": 1.280746395250212,
      "grad_norm": 3.2223682403564453,
      "learning_rate": 9.956594323873122e-05,
      "loss": 0.3317,
      "step": 6040
    },
    {
      "epoch": 1.2849872773536894,
      "grad_norm": 2.561291456222534,
      "learning_rate": 9.923205342237061e-05,
      "loss": 0.3481,
      "step": 6060
    },
    {
      "epoch": 1.289228159457167,
      "grad_norm": 2.6729624271392822,
      "learning_rate": 9.889816360601002e-05,
      "loss": 0.3733,
      "step": 6080
    },
    {
      "epoch": 1.2934690415606447,
      "grad_norm": 2.9964187145233154,
      "learning_rate": 9.856427378964941e-05,
      "loss": 0.3248,
      "step": 6100
    },
    {
      "epoch": 1.297709923664122,
      "grad_norm": 2.4425148963928223,
      "learning_rate": 9.823038397328882e-05,
      "loss": 0.3631,
      "step": 6120
    },
    {
      "epoch": 1.3019508057675997,
      "grad_norm": 3.486454963684082,
      "learning_rate": 9.789649415692823e-05,
      "loss": 0.3245,
      "step": 6140
    },
    {
      "epoch": 1.306191687871077,
      "grad_norm": 2.1288866996765137,
      "learning_rate": 9.756260434056762e-05,
      "loss": 0.3463,
      "step": 6160
    },
    {
      "epoch": 1.3104325699745547,
      "grad_norm": 3.143205404281616,
      "learning_rate": 9.722871452420703e-05,
      "loss": 0.3357,
      "step": 6180
    },
    {
      "epoch": 1.3146734520780323,
      "grad_norm": 3.161135673522949,
      "learning_rate": 9.68948247078464e-05,
      "loss": 0.3154,
      "step": 6200
    },
    {
      "epoch": 1.3146734520780323,
      "eval_loss": NaN,
      "eval_runtime": 37.8949,
      "eval_samples_per_second": 26.389,
      "eval_steps_per_second": 3.299,
      "step": 6200
    },
    {
      "epoch": 1.3189143341815097,
      "grad_norm": 2.3218791484832764,
      "learning_rate": 9.656093489148581e-05,
      "loss": 0.3233,
      "step": 6220
    },
    {
      "epoch": 1.3231552162849873,
      "grad_norm": 2.413891553878784,
      "learning_rate": 9.62270450751252e-05,
      "loss": 0.3155,
      "step": 6240
    },
    {
      "epoch": 1.3273960983884647,
      "grad_norm": 2.4525930881500244,
      "learning_rate": 9.589315525876461e-05,
      "loss": 0.3283,
      "step": 6260
    },
    {
      "epoch": 1.3316369804919423,
      "grad_norm": 2.1102969646453857,
      "learning_rate": 9.555926544240402e-05,
      "loss": 0.3304,
      "step": 6280
    },
    {
      "epoch": 1.33587786259542,
      "grad_norm": 2.621964454650879,
      "learning_rate": 9.522537562604341e-05,
      "loss": 0.3187,
      "step": 6300
    },
    {
      "epoch": 1.3401187446988974,
      "grad_norm": 2.056610345840454,
      "learning_rate": 9.489148580968282e-05,
      "loss": 0.3352,
      "step": 6320
    },
    {
      "epoch": 1.344359626802375,
      "grad_norm": 2.7109179496765137,
      "learning_rate": 9.455759599332221e-05,
      "loss": 0.2929,
      "step": 6340
    },
    {
      "epoch": 1.3486005089058524,
      "grad_norm": 3.21298885345459,
      "learning_rate": 9.422370617696162e-05,
      "loss": 0.3139,
      "step": 6360
    },
    {
      "epoch": 1.35284139100933,
      "grad_norm": 2.386242389678955,
      "learning_rate": 9.388981636060101e-05,
      "loss": 0.3147,
      "step": 6380
    },
    {
      "epoch": 1.3570822731128076,
      "grad_norm": 2.346001386642456,
      "learning_rate": 9.35559265442404e-05,
      "loss": 0.328,
      "step": 6400
    },
    {
      "epoch": 1.3570822731128076,
      "eval_loss": NaN,
      "eval_runtime": 37.8691,
      "eval_samples_per_second": 26.407,
      "eval_steps_per_second": 3.301,
      "step": 6400
    },
    {
      "epoch": 1.361323155216285,
      "grad_norm": 2.5485541820526123,
      "learning_rate": 9.322203672787981e-05,
      "loss": 0.3027,
      "step": 6420
    },
    {
      "epoch": 1.3655640373197624,
      "grad_norm": 2.338224172592163,
      "learning_rate": 9.28881469115192e-05,
      "loss": 0.2996,
      "step": 6440
    },
    {
      "epoch": 1.36980491942324,
      "grad_norm": 2.043363094329834,
      "learning_rate": 9.255425709515861e-05,
      "loss": 0.3289,
      "step": 6460
    },
    {
      "epoch": 1.3740458015267176,
      "grad_norm": 2.4965274333953857,
      "learning_rate": 9.2220367278798e-05,
      "loss": 0.3214,
      "step": 6480
    },
    {
      "epoch": 1.378286683630195,
      "grad_norm": 2.0820116996765137,
      "learning_rate": 9.18864774624374e-05,
      "loss": 0.2939,
      "step": 6500
    },
    {
      "epoch": 1.3825275657336726,
      "grad_norm": 3.0011324882507324,
      "learning_rate": 9.15525876460768e-05,
      "loss": 0.33,
      "step": 6520
    },
    {
      "epoch": 1.38676844783715,
      "grad_norm": 2.5677077770233154,
      "learning_rate": 9.121869782971619e-05,
      "loss": 0.3258,
      "step": 6540
    },
    {
      "epoch": 1.3910093299406276,
      "grad_norm": 3.5414836406707764,
      "learning_rate": 9.08848080133556e-05,
      "loss": 0.3332,
      "step": 6560
    },
    {
      "epoch": 1.3952502120441053,
      "grad_norm": 2.178295135498047,
      "learning_rate": 9.055091819699499e-05,
      "loss": 0.3056,
      "step": 6580
    },
    {
      "epoch": 1.3994910941475827,
      "grad_norm": 2.7327194213867188,
      "learning_rate": 9.02170283806344e-05,
      "loss": 0.3401,
      "step": 6600
    },
    {
      "epoch": 1.3994910941475827,
      "eval_loss": NaN,
      "eval_runtime": 38.0354,
      "eval_samples_per_second": 26.291,
      "eval_steps_per_second": 3.286,
      "step": 6600
    },
    {
      "epoch": 1.4037319762510603,
      "grad_norm": 2.3425300121307373,
      "learning_rate": 8.988313856427379e-05,
      "loss": 0.3429,
      "step": 6620
    },
    {
      "epoch": 1.4079728583545377,
      "grad_norm": 2.3231027126312256,
      "learning_rate": 8.95492487479132e-05,
      "loss": 0.3125,
      "step": 6640
    },
    {
      "epoch": 1.4122137404580153,
      "grad_norm": 2.060715675354004,
      "learning_rate": 8.921535893155259e-05,
      "loss": 0.3193,
      "step": 6660
    },
    {
      "epoch": 1.416454622561493,
      "grad_norm": 2.686115264892578,
      "learning_rate": 8.8881469115192e-05,
      "loss": 0.3315,
      "step": 6680
    },
    {
      "epoch": 1.4206955046649703,
      "grad_norm": 3.2563514709472656,
      "learning_rate": 8.854757929883139e-05,
      "loss": 0.3034,
      "step": 6700
    },
    {
      "epoch": 1.424936386768448,
      "grad_norm": 2.294642925262451,
      "learning_rate": 8.821368948247078e-05,
      "loss": 0.3124,
      "step": 6720
    },
    {
      "epoch": 1.4291772688719253,
      "grad_norm": 2.9424397945404053,
      "learning_rate": 8.787979966611019e-05,
      "loss": 0.3109,
      "step": 6740
    },
    {
      "epoch": 1.433418150975403,
      "grad_norm": 2.2704017162323,
      "learning_rate": 8.754590984974958e-05,
      "loss": 0.2703,
      "step": 6760
    },
    {
      "epoch": 1.4376590330788805,
      "grad_norm": 3.3748135566711426,
      "learning_rate": 8.721202003338899e-05,
      "loss": 0.3386,
      "step": 6780
    },
    {
      "epoch": 1.441899915182358,
      "grad_norm": 2.500150203704834,
      "learning_rate": 8.687813021702838e-05,
      "loss": 0.3126,
      "step": 6800
    },
    {
      "epoch": 1.441899915182358,
      "eval_loss": NaN,
      "eval_runtime": 38.0024,
      "eval_samples_per_second": 26.314,
      "eval_steps_per_second": 3.289,
      "step": 6800
    },
    {
      "epoch": 1.4461407972858353,
      "grad_norm": 2.9435441493988037,
      "learning_rate": 8.654424040066779e-05,
      "loss": 0.3387,
      "step": 6820
    },
    {
      "epoch": 1.450381679389313,
      "grad_norm": 1.965718388557434,
      "learning_rate": 8.62103505843072e-05,
      "loss": 0.3028,
      "step": 6840
    },
    {
      "epoch": 1.4546225614927906,
      "grad_norm": 2.8321971893310547,
      "learning_rate": 8.587646076794659e-05,
      "loss": 0.3233,
      "step": 6860
    },
    {
      "epoch": 1.458863443596268,
      "grad_norm": 2.4241039752960205,
      "learning_rate": 8.554257095158598e-05,
      "loss": 0.3081,
      "step": 6880
    },
    {
      "epoch": 1.4631043256997456,
      "grad_norm": 2.129350423812866,
      "learning_rate": 8.520868113522537e-05,
      "loss": 0.299,
      "step": 6900
    },
    {
      "epoch": 1.467345207803223,
      "grad_norm": 3.391523838043213,
      "learning_rate": 8.487479131886478e-05,
      "loss": 0.319,
      "step": 6920
    },
    {
      "epoch": 1.4715860899067006,
      "grad_norm": 2.4123945236206055,
      "learning_rate": 8.454090150250417e-05,
      "loss": 0.3261,
      "step": 6940
    },
    {
      "epoch": 1.4758269720101782,
      "grad_norm": 2.5370140075683594,
      "learning_rate": 8.420701168614358e-05,
      "loss": 0.298,
      "step": 6960
    },
    {
      "epoch": 1.4800678541136556,
      "grad_norm": 3.2087557315826416,
      "learning_rate": 8.387312186978298e-05,
      "loss": 0.2916,
      "step": 6980
    },
    {
      "epoch": 1.4843087362171332,
      "grad_norm": 2.307880163192749,
      "learning_rate": 8.353923205342238e-05,
      "loss": 0.2901,
      "step": 7000
    },
    {
      "epoch": 1.4843087362171332,
      "eval_loss": NaN,
      "eval_runtime": 37.9172,
      "eval_samples_per_second": 26.373,
      "eval_steps_per_second": 3.297,
      "step": 7000
    },
    {
      "epoch": 1.4885496183206106,
      "grad_norm": 2.6145153045654297,
      "learning_rate": 8.320534223706178e-05,
      "loss": 0.3129,
      "step": 7020
    },
    {
      "epoch": 1.4927905004240882,
      "grad_norm": 2.5438311100006104,
      "learning_rate": 8.287145242070116e-05,
      "loss": 0.3112,
      "step": 7040
    },
    {
      "epoch": 1.4970313825275658,
      "grad_norm": 2.8195457458496094,
      "learning_rate": 8.253756260434057e-05,
      "loss": 0.3005,
      "step": 7060
    },
    {
      "epoch": 1.5012722646310432,
      "grad_norm": 2.549837112426758,
      "learning_rate": 8.220367278797996e-05,
      "loss": 0.2716,
      "step": 7080
    },
    {
      "epoch": 1.5055131467345206,
      "grad_norm": 2.7534306049346924,
      "learning_rate": 8.186978297161937e-05,
      "loss": 0.2906,
      "step": 7100
    },
    {
      "epoch": 1.5097540288379983,
      "grad_norm": 3.789994239807129,
      "learning_rate": 8.153589315525877e-05,
      "loss": 0.2652,
      "step": 7120
    },
    {
      "epoch": 1.5139949109414759,
      "grad_norm": 2.3409814834594727,
      "learning_rate": 8.120200333889817e-05,
      "loss": 0.2751,
      "step": 7140
    },
    {
      "epoch": 1.5182357930449535,
      "grad_norm": 2.9466300010681152,
      "learning_rate": 8.086811352253757e-05,
      "loss": 0.2342,
      "step": 7160
    },
    {
      "epoch": 1.5224766751484309,
      "grad_norm": 2.7116031646728516,
      "learning_rate": 8.053422370617697e-05,
      "loss": 0.2742,
      "step": 7180
    },
    {
      "epoch": 1.5267175572519083,
      "grad_norm": 2.626523494720459,
      "learning_rate": 8.020033388981636e-05,
      "loss": 0.3174,
      "step": 7200
    },
    {
      "epoch": 1.5267175572519083,
      "eval_loss": NaN,
      "eval_runtime": 37.8367,
      "eval_samples_per_second": 26.429,
      "eval_steps_per_second": 3.304,
      "step": 7200
    },
    {
      "epoch": 1.530958439355386,
      "grad_norm": 3.002830743789673,
      "learning_rate": 7.986644407345575e-05,
      "loss": 0.2723,
      "step": 7220
    },
    {
      "epoch": 1.5351993214588635,
      "grad_norm": 2.650344133377075,
      "learning_rate": 7.953255425709516e-05,
      "loss": 0.2713,
      "step": 7240
    },
    {
      "epoch": 1.5394402035623411,
      "grad_norm": 2.4937496185302734,
      "learning_rate": 7.919866444073457e-05,
      "loss": 0.3015,
      "step": 7260
    },
    {
      "epoch": 1.5436810856658185,
      "grad_norm": 2.3540656566619873,
      "learning_rate": 7.886477462437396e-05,
      "loss": 0.313,
      "step": 7280
    },
    {
      "epoch": 1.547921967769296,
      "grad_norm": 1.8998830318450928,
      "learning_rate": 7.853088480801337e-05,
      "loss": 0.2616,
      "step": 7300
    },
    {
      "epoch": 1.5521628498727735,
      "grad_norm": 2.5842199325561523,
      "learning_rate": 7.819699499165276e-05,
      "loss": 0.3048,
      "step": 7320
    },
    {
      "epoch": 1.5564037319762511,
      "grad_norm": 2.214853525161743,
      "learning_rate": 7.786310517529216e-05,
      "loss": 0.2561,
      "step": 7340
    },
    {
      "epoch": 1.5606446140797285,
      "grad_norm": 2.432750940322876,
      "learning_rate": 7.752921535893156e-05,
      "loss": 0.2826,
      "step": 7360
    },
    {
      "epoch": 1.5648854961832062,
      "grad_norm": 2.189223051071167,
      "learning_rate": 7.719532554257095e-05,
      "loss": 0.2902,
      "step": 7380
    },
    {
      "epoch": 1.5691263782866836,
      "grad_norm": 2.117978572845459,
      "learning_rate": 7.686143572621036e-05,
      "loss": 0.2845,
      "step": 7400
    },
    {
      "epoch": 1.5691263782866836,
      "eval_loss": NaN,
      "eval_runtime": 37.9556,
      "eval_samples_per_second": 26.347,
      "eval_steps_per_second": 3.293,
      "step": 7400
    },
    {
      "epoch": 1.5733672603901612,
      "grad_norm": 3.270785331726074,
      "learning_rate": 7.652754590984975e-05,
      "loss": 0.2926,
      "step": 7420
    },
    {
      "epoch": 1.5776081424936388,
      "grad_norm": 2.4925777912139893,
      "learning_rate": 7.619365609348916e-05,
      "loss": 0.316,
      "step": 7440
    },
    {
      "epoch": 1.5818490245971162,
      "grad_norm": 3.0589041709899902,
      "learning_rate": 7.587646076794657e-05,
      "loss": 0.2708,
      "step": 7460
    },
    {
      "epoch": 1.5860899067005936,
      "grad_norm": 2.6401047706604004,
      "learning_rate": 7.554257095158598e-05,
      "loss": 0.2777,
      "step": 7480
    },
    {
      "epoch": 1.5903307888040712,
      "grad_norm": 2.553291082382202,
      "learning_rate": 7.520868113522537e-05,
      "loss": 0.3136,
      "step": 7500
    },
    {
      "epoch": 1.5945716709075488,
      "grad_norm": 3.430556535720825,
      "learning_rate": 7.487479131886478e-05,
      "loss": 0.2691,
      "step": 7520
    },
    {
      "epoch": 1.5988125530110264,
      "grad_norm": 2.3197848796844482,
      "learning_rate": 7.454090150250419e-05,
      "loss": 0.306,
      "step": 7540
    },
    {
      "epoch": 1.6030534351145038,
      "grad_norm": 2.545103073120117,
      "learning_rate": 7.420701168614358e-05,
      "loss": 0.2906,
      "step": 7560
    },
    {
      "epoch": 1.6072943172179812,
      "grad_norm": 2.177480697631836,
      "learning_rate": 7.387312186978299e-05,
      "loss": 0.2807,
      "step": 7580
    },
    {
      "epoch": 1.6115351993214588,
      "grad_norm": 2.4286954402923584,
      "learning_rate": 7.353923205342236e-05,
      "loss": 0.271,
      "step": 7600
    },
    {
      "epoch": 1.6115351993214588,
      "eval_loss": NaN,
      "eval_runtime": 37.9885,
      "eval_samples_per_second": 26.324,
      "eval_steps_per_second": 3.29,
      "step": 7600
    },
    {
      "epoch": 1.6157760814249365,
      "grad_norm": 2.295604944229126,
      "learning_rate": 7.320534223706177e-05,
      "loss": 0.2663,
      "step": 7620
    },
    {
      "epoch": 1.620016963528414,
      "grad_norm": 2.827371835708618,
      "learning_rate": 7.287145242070116e-05,
      "loss": 0.2769,
      "step": 7640
    },
    {
      "epoch": 1.6242578456318915,
      "grad_norm": 2.231592893600464,
      "learning_rate": 7.253756260434057e-05,
      "loss": 0.239,
      "step": 7660
    },
    {
      "epoch": 1.6284987277353689,
      "grad_norm": 1.9957088232040405,
      "learning_rate": 7.220367278797998e-05,
      "loss": 0.2869,
      "step": 7680
    },
    {
      "epoch": 1.6327396098388465,
      "grad_norm": 2.816500425338745,
      "learning_rate": 7.186978297161937e-05,
      "loss": 0.2765,
      "step": 7700
    },
    {
      "epoch": 1.636980491942324,
      "grad_norm": 2.1216447353363037,
      "learning_rate": 7.153589315525878e-05,
      "loss": 0.2422,
      "step": 7720
    },
    {
      "epoch": 1.6412213740458015,
      "grad_norm": 2.527742385864258,
      "learning_rate": 7.120200333889817e-05,
      "loss": 0.2747,
      "step": 7740
    },
    {
      "epoch": 1.645462256149279,
      "grad_norm": 2.030169725418091,
      "learning_rate": 7.086811352253758e-05,
      "loss": 0.2639,
      "step": 7760
    },
    {
      "epoch": 1.6497031382527565,
      "grad_norm": 2.324866771697998,
      "learning_rate": 7.053422370617695e-05,
      "loss": 0.2387,
      "step": 7780
    },
    {
      "epoch": 1.6539440203562341,
      "grad_norm": 2.7137794494628906,
      "learning_rate": 7.020033388981636e-05,
      "loss": 0.2598,
      "step": 7800
    },
    {
      "epoch": 1.6539440203562341,
      "eval_loss": NaN,
      "eval_runtime": 37.9351,
      "eval_samples_per_second": 26.361,
      "eval_steps_per_second": 3.295,
      "step": 7800
    }
  ],
  "logging_steps": 20,
  "max_steps": 12000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.374158637713326e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}