{
  "best_metric": 0.11646050214767456,
  "best_model_checkpoint": "./vit-indian-food/checkpoint-1500",
  "epoch": 7.853403141361256,
  "eval_steps": 50,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "grad_norm": 7.558130264282227,
      "learning_rate": 0.00019947643979057592,
      "loss": 2.4369,
      "step": 5
    },
    {
      "epoch": 0.05,
      "grad_norm": 6.5493950843811035,
      "learning_rate": 0.00019895287958115185,
      "loss": 1.6214,
      "step": 10
    },
    {
      "epoch": 0.08,
      "grad_norm": 6.205695629119873,
      "learning_rate": 0.00019842931937172776,
      "loss": 1.197,
      "step": 15
    },
    {
      "epoch": 0.1,
      "grad_norm": 3.8880319595336914,
      "learning_rate": 0.00019790575916230367,
      "loss": 0.7391,
      "step": 20
    },
    {
      "epoch": 0.13,
      "grad_norm": 6.315922260284424,
      "learning_rate": 0.0001973821989528796,
      "loss": 0.6288,
      "step": 25
    },
    {
      "epoch": 0.16,
      "grad_norm": 4.008142471313477,
      "learning_rate": 0.0001968586387434555,
      "loss": 0.3582,
      "step": 30
    },
    {
      "epoch": 0.18,
      "grad_norm": 4.005300998687744,
      "learning_rate": 0.00019633507853403142,
      "loss": 0.3882,
      "step": 35
    },
    {
      "epoch": 0.21,
      "grad_norm": 3.9506969451904297,
      "learning_rate": 0.00019581151832460733,
      "loss": 0.39,
      "step": 40
    },
    {
      "epoch": 0.24,
      "grad_norm": 4.074863910675049,
      "learning_rate": 0.00019528795811518326,
      "loss": 0.3413,
      "step": 45
    },
    {
      "epoch": 0.26,
      "grad_norm": 4.617203712463379,
      "learning_rate": 0.00019476439790575917,
      "loss": 0.358,
      "step": 50
    },
    {
      "epoch": 0.26,
      "eval_accuracy": 0.9081364829396326,
      "eval_loss": 0.32936015725135803,
      "eval_precision": 0.9220515144455581,
      "eval_recall": 0.9081364829396326,
      "eval_runtime": 17.5763,
      "eval_samples_per_second": 43.354,
      "eval_steps_per_second": 2.731,
      "step": 50
    },
    {
      "epoch": 0.29,
      "grad_norm": 4.685304641723633,
      "learning_rate": 0.0001942408376963351,
      "loss": 0.3669,
      "step": 55
    },
    {
      "epoch": 0.31,
      "grad_norm": 3.119244337081909,
      "learning_rate": 0.000193717277486911,
      "loss": 0.1722,
      "step": 60
    },
    {
      "epoch": 0.34,
      "grad_norm": 6.17639684677124,
      "learning_rate": 0.00019319371727748692,
      "loss": 0.405,
      "step": 65
    },
    {
      "epoch": 0.37,
      "grad_norm": 4.263852119445801,
      "learning_rate": 0.00019267015706806283,
      "loss": 0.5013,
      "step": 70
    },
    {
      "epoch": 0.39,
      "grad_norm": 2.03020977973938,
      "learning_rate": 0.00019214659685863877,
      "loss": 0.2251,
      "step": 75
    },
    {
      "epoch": 0.42,
      "grad_norm": 5.0216193199157715,
      "learning_rate": 0.00019162303664921465,
      "loss": 0.3329,
      "step": 80
    },
    {
      "epoch": 0.45,
      "grad_norm": 7.692380428314209,
      "learning_rate": 0.00019109947643979058,
      "loss": 0.2191,
      "step": 85
    },
    {
      "epoch": 0.47,
      "grad_norm": 5.645913600921631,
      "learning_rate": 0.0001905759162303665,
      "loss": 0.3437,
      "step": 90
    },
    {
      "epoch": 0.5,
      "grad_norm": 5.066718578338623,
      "learning_rate": 0.00019005235602094243,
      "loss": 0.4461,
      "step": 95
    },
    {
      "epoch": 0.52,
      "grad_norm": 4.659124851226807,
      "learning_rate": 0.00018952879581151833,
      "loss": 0.5051,
      "step": 100
    },
    {
      "epoch": 0.52,
      "eval_accuracy": 0.9291338582677166,
      "eval_loss": 0.22470374405384064,
      "eval_precision": 0.9372539303361345,
      "eval_recall": 0.9291338582677166,
      "eval_runtime": 17.4785,
      "eval_samples_per_second": 43.597,
      "eval_steps_per_second": 2.746,
      "step": 100
    },
    {
      "epoch": 0.55,
      "grad_norm": 4.5200324058532715,
      "learning_rate": 0.00018900523560209424,
      "loss": 0.2615,
      "step": 105
    },
    {
      "epoch": 0.58,
      "grad_norm": 1.2165738344192505,
      "learning_rate": 0.00018848167539267018,
      "loss": 0.2415,
      "step": 110
    },
    {
      "epoch": 0.6,
      "grad_norm": 6.801146030426025,
      "learning_rate": 0.00018795811518324608,
      "loss": 0.3234,
      "step": 115
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8039401173591614,
      "learning_rate": 0.00018743455497382202,
      "loss": 0.1962,
      "step": 120
    },
    {
      "epoch": 0.65,
      "grad_norm": 4.639632701873779,
      "learning_rate": 0.0001869109947643979,
      "loss": 0.2755,
      "step": 125
    },
    {
      "epoch": 0.68,
      "grad_norm": 3.4670636653900146,
      "learning_rate": 0.00018638743455497384,
      "loss": 0.3463,
      "step": 130
    },
    {
      "epoch": 0.71,
      "grad_norm": 6.07156229019165,
      "learning_rate": 0.00018586387434554974,
      "loss": 0.3047,
      "step": 135
    },
    {
      "epoch": 0.73,
      "grad_norm": 6.347087860107422,
      "learning_rate": 0.00018534031413612568,
      "loss": 0.2364,
      "step": 140
    },
    {
      "epoch": 0.76,
      "grad_norm": 5.963706970214844,
      "learning_rate": 0.00018481675392670156,
      "loss": 0.4148,
      "step": 145
    },
    {
      "epoch": 0.79,
      "grad_norm": 4.037764072418213,
      "learning_rate": 0.0001842931937172775,
      "loss": 0.1361,
      "step": 150
    },
    {
      "epoch": 0.79,
      "eval_accuracy": 0.9409448818897638,
      "eval_loss": 0.21655863523483276,
      "eval_precision": 0.9454930944052194,
      "eval_recall": 0.9409448818897638,
      "eval_runtime": 17.4607,
      "eval_samples_per_second": 43.641,
      "eval_steps_per_second": 2.749,
      "step": 150
    },
    {
      "epoch": 0.81,
      "grad_norm": 4.717739582061768,
      "learning_rate": 0.0001837696335078534,
      "loss": 0.1183,
      "step": 155
    },
    {
      "epoch": 0.84,
      "grad_norm": 3.9427084922790527,
      "learning_rate": 0.00018324607329842934,
      "loss": 0.1138,
      "step": 160
    },
    {
      "epoch": 0.86,
      "grad_norm": 4.915616989135742,
      "learning_rate": 0.00018272251308900525,
      "loss": 0.2988,
      "step": 165
    },
    {
      "epoch": 0.89,
      "grad_norm": 6.3770432472229,
      "learning_rate": 0.00018219895287958115,
      "loss": 0.1744,
      "step": 170
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.7859281301498413,
      "learning_rate": 0.00018167539267015706,
      "loss": 0.2486,
      "step": 175
    },
    {
      "epoch": 0.94,
      "grad_norm": 3.75736141204834,
      "learning_rate": 0.000181151832460733,
      "loss": 0.2637,
      "step": 180
    },
    {
      "epoch": 0.97,
      "grad_norm": 6.771958351135254,
      "learning_rate": 0.0001806282722513089,
      "loss": 0.2305,
      "step": 185
    },
    {
      "epoch": 0.99,
      "grad_norm": 4.951119899749756,
      "learning_rate": 0.0001801047120418848,
      "loss": 0.1108,
      "step": 190
    },
    {
      "epoch": 1.02,
      "grad_norm": 3.362152576446533,
      "learning_rate": 0.00017958115183246075,
      "loss": 0.133,
      "step": 195
    },
    {
      "epoch": 1.05,
      "grad_norm": 0.8346299529075623,
      "learning_rate": 0.00017905759162303666,
      "loss": 0.0611,
      "step": 200
    },
    {
      "epoch": 1.05,
      "eval_accuracy": 0.9488188976377953,
      "eval_loss": 0.18689152598381042,
      "eval_precision": 0.9533647964651322,
      "eval_recall": 0.9488188976377953,
      "eval_runtime": 17.6111,
      "eval_samples_per_second": 43.268,
      "eval_steps_per_second": 2.726,
      "step": 200
    },
    {
      "epoch": 1.07,
      "grad_norm": 0.5746695399284363,
      "learning_rate": 0.0001785340314136126,
      "loss": 0.0241,
      "step": 205
    },
    {
      "epoch": 1.1,
      "grad_norm": 2.0743372440338135,
      "learning_rate": 0.0001780104712041885,
      "loss": 0.0662,
      "step": 210
    },
    {
      "epoch": 1.13,
      "grad_norm": 1.6814557313919067,
      "learning_rate": 0.0001774869109947644,
      "loss": 0.0635,
      "step": 215
    },
    {
      "epoch": 1.15,
      "grad_norm": 2.118739128112793,
      "learning_rate": 0.00017696335078534032,
      "loss": 0.1024,
      "step": 220
    },
    {
      "epoch": 1.18,
      "grad_norm": 3.4306182861328125,
      "learning_rate": 0.00017643979057591625,
      "loss": 0.0201,
      "step": 225
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.28945377469062805,
      "learning_rate": 0.00017591623036649216,
      "loss": 0.1151,
      "step": 230
    },
    {
      "epoch": 1.23,
      "grad_norm": 4.6208648681640625,
      "learning_rate": 0.00017539267015706807,
      "loss": 0.2324,
      "step": 235
    },
    {
      "epoch": 1.26,
      "grad_norm": 1.6085395812988281,
      "learning_rate": 0.00017486910994764398,
      "loss": 0.0429,
      "step": 240
    },
    {
      "epoch": 1.28,
      "grad_norm": 5.860565185546875,
      "learning_rate": 0.0001743455497382199,
      "loss": 0.0567,
      "step": 245
    },
    {
      "epoch": 1.31,
      "grad_norm": 5.421276092529297,
      "learning_rate": 0.00017382198952879582,
      "loss": 0.1037,
      "step": 250
    },
    {
      "epoch": 1.31,
      "eval_accuracy": 0.968503937007874,
      "eval_loss": 0.11793948709964752,
      "eval_precision": 0.9693355319509627,
      "eval_recall": 0.968503937007874,
      "eval_runtime": 17.5701,
      "eval_samples_per_second": 43.369,
      "eval_steps_per_second": 2.732,
      "step": 250
    },
    {
      "epoch": 1.34,
      "grad_norm": 3.9477956295013428,
      "learning_rate": 0.00017329842931937175,
      "loss": 0.0285,
      "step": 255
    },
    {
      "epoch": 1.36,
      "grad_norm": 7.515362739562988,
      "learning_rate": 0.00017277486910994763,
      "loss": 0.2195,
      "step": 260
    },
    {
      "epoch": 1.39,
      "grad_norm": 0.1676739752292633,
      "learning_rate": 0.00017225130890052357,
      "loss": 0.0172,
      "step": 265
    },
    {
      "epoch": 1.41,
      "grad_norm": 8.235359191894531,
      "learning_rate": 0.00017172774869109948,
      "loss": 0.1756,
      "step": 270
    },
    {
      "epoch": 1.44,
      "grad_norm": 7.9198832511901855,
      "learning_rate": 0.0001712041884816754,
      "loss": 0.0663,
      "step": 275
    },
    {
      "epoch": 1.47,
      "grad_norm": 6.388627529144287,
      "learning_rate": 0.00017068062827225132,
      "loss": 0.1467,
      "step": 280
    },
    {
      "epoch": 1.49,
      "grad_norm": 0.09213338047266006,
      "learning_rate": 0.00017015706806282723,
      "loss": 0.0684,
      "step": 285
    },
    {
      "epoch": 1.52,
      "grad_norm": 1.0691012144088745,
      "learning_rate": 0.00016963350785340316,
      "loss": 0.0954,
      "step": 290
    },
    {
      "epoch": 1.54,
      "grad_norm": 6.120029449462891,
      "learning_rate": 0.00016910994764397907,
      "loss": 0.0587,
      "step": 295
    },
    {
      "epoch": 1.57,
      "grad_norm": 0.0928138718008995,
      "learning_rate": 0.00016858638743455498,
      "loss": 0.0294,
      "step": 300
    },
    {
      "epoch": 1.57,
      "eval_accuracy": 0.963254593175853,
      "eval_loss": 0.11820019036531448,
      "eval_precision": 0.9645155573207865,
      "eval_recall": 0.963254593175853,
      "eval_runtime": 17.5014,
      "eval_samples_per_second": 43.539,
      "eval_steps_per_second": 2.743,
      "step": 300
    },
    {
      "epoch": 1.6,
      "grad_norm": 0.3131069242954254,
      "learning_rate": 0.0001680628272251309,
      "loss": 0.1184,
      "step": 305
    },
    {
      "epoch": 1.62,
      "grad_norm": 0.417850524187088,
      "learning_rate": 0.00016753926701570682,
      "loss": 0.0053,
      "step": 310
    },
    {
      "epoch": 1.65,
      "grad_norm": 5.371811389923096,
      "learning_rate": 0.00016701570680628273,
      "loss": 0.1099,
      "step": 315
    },
    {
      "epoch": 1.68,
      "grad_norm": 4.09129524230957,
      "learning_rate": 0.00016649214659685867,
      "loss": 0.2501,
      "step": 320
    },
    {
      "epoch": 1.7,
      "grad_norm": 0.6515682935714722,
      "learning_rate": 0.00016596858638743455,
      "loss": 0.0528,
      "step": 325
    },
    {
      "epoch": 1.73,
      "grad_norm": 0.5656059384346008,
      "learning_rate": 0.00016544502617801048,
      "loss": 0.1378,
      "step": 330
    },
    {
      "epoch": 1.75,
      "grad_norm": 0.10378353297710419,
      "learning_rate": 0.0001649214659685864,
      "loss": 0.0329,
      "step": 335
    },
    {
      "epoch": 1.78,
      "grad_norm": 2.387305736541748,
      "learning_rate": 0.00016439790575916233,
      "loss": 0.0353,
      "step": 340
    },
    {
      "epoch": 1.81,
      "grad_norm": 0.05628238245844841,
      "learning_rate": 0.0001638743455497382,
      "loss": 0.0138,
      "step": 345
    },
    {
      "epoch": 1.83,
      "grad_norm": 0.04053734615445137,
      "learning_rate": 0.00016335078534031414,
      "loss": 0.0082,
      "step": 350
    },
    {
      "epoch": 1.83,
      "eval_accuracy": 0.963254593175853,
      "eval_loss": 0.11838709563016891,
      "eval_precision": 0.9654537871036863,
      "eval_recall": 0.963254593175853,
      "eval_runtime": 17.4034,
      "eval_samples_per_second": 43.785,
      "eval_steps_per_second": 2.758,
      "step": 350
    },
    {
      "epoch": 1.86,
      "grad_norm": 2.287546396255493,
      "learning_rate": 0.00016282722513089005,
      "loss": 0.02,
      "step": 355
    },
    {
      "epoch": 1.88,
      "grad_norm": 4.561453342437744,
      "learning_rate": 0.00016230366492146599,
      "loss": 0.203,
      "step": 360
    },
    {
      "epoch": 1.91,
      "grad_norm": 4.0477294921875,
      "learning_rate": 0.0001617801047120419,
      "loss": 0.0646,
      "step": 365
    },
    {
      "epoch": 1.94,
      "grad_norm": 1.0380969047546387,
      "learning_rate": 0.0001612565445026178,
      "loss": 0.0848,
      "step": 370
    },
    {
      "epoch": 1.96,
      "grad_norm": 4.385131359100342,
      "learning_rate": 0.0001607329842931937,
      "loss": 0.166,
      "step": 375
    },
    {
      "epoch": 1.99,
      "grad_norm": 0.036621347069740295,
      "learning_rate": 0.00016020942408376964,
      "loss": 0.0871,
      "step": 380
    },
    {
      "epoch": 2.02,
      "grad_norm": 0.5929269194602966,
      "learning_rate": 0.00015968586387434555,
      "loss": 0.1271,
      "step": 385
    },
    {
      "epoch": 2.04,
      "grad_norm": 0.26594266295433044,
      "learning_rate": 0.00015916230366492146,
      "loss": 0.0114,
      "step": 390
    },
    {
      "epoch": 2.07,
      "grad_norm": 0.020196596160531044,
      "learning_rate": 0.0001586387434554974,
      "loss": 0.0081,
      "step": 395
    },
    {
      "epoch": 2.09,
      "grad_norm": 0.6509531140327454,
      "learning_rate": 0.0001581151832460733,
      "loss": 0.0206,
      "step": 400
    },
    {
      "epoch": 2.09,
      "eval_accuracy": 0.963254593175853,
      "eval_loss": 0.130891352891922,
      "eval_precision": 0.964741369994443,
      "eval_recall": 0.963254593175853,
      "eval_runtime": 17.4584,
      "eval_samples_per_second": 43.647,
      "eval_steps_per_second": 2.749,
      "step": 400
    },
    {
      "epoch": 2.12,
      "grad_norm": 0.04883525148034096,
      "learning_rate": 0.00015759162303664924,
      "loss": 0.047,
      "step": 405
    },
    {
      "epoch": 2.15,
      "grad_norm": 1.4584039449691772,
      "learning_rate": 0.00015706806282722515,
      "loss": 0.0393,
      "step": 410
    },
    {
      "epoch": 2.17,
      "grad_norm": 0.3986053466796875,
      "learning_rate": 0.00015654450261780105,
      "loss": 0.0606,
      "step": 415
    },
    {
      "epoch": 2.2,
      "grad_norm": 0.11985136568546295,
      "learning_rate": 0.00015602094240837696,
      "loss": 0.0379,
      "step": 420
    },
    {
      "epoch": 2.23,
      "grad_norm": 0.07443628460168839,
      "learning_rate": 0.0001554973821989529,
      "loss": 0.0157,
      "step": 425
    },
    {
      "epoch": 2.25,
      "grad_norm": 0.03219376876950264,
      "learning_rate": 0.0001549738219895288,
      "loss": 0.01,
      "step": 430
    },
    {
      "epoch": 2.28,
      "grad_norm": 0.17020490765571594,
      "learning_rate": 0.00015445026178010471,
      "loss": 0.0093,
      "step": 435
    },
    {
      "epoch": 2.3,
      "grad_norm": 2.6877171993255615,
      "learning_rate": 0.00015392670157068062,
      "loss": 0.021,
      "step": 440
    },
    {
      "epoch": 2.33,
      "grad_norm": 0.010094034485518932,
      "learning_rate": 0.00015340314136125656,
      "loss": 0.0083,
      "step": 445
    },
    {
      "epoch": 2.36,
      "grad_norm": 0.0077184755355119705,
      "learning_rate": 0.00015287958115183247,
      "loss": 0.0246,
      "step": 450
    },
    {
      "epoch": 2.36,
      "eval_accuracy": 0.9553805774278216,
      "eval_loss": 0.17534801363945007,
      "eval_precision": 0.9586171915743325,
      "eval_recall": 0.9553805774278216,
      "eval_runtime": 17.505,
      "eval_samples_per_second": 43.53,
      "eval_steps_per_second": 2.742,
      "step": 450
    },
    {
      "epoch": 2.38,
      "grad_norm": 5.202419757843018,
      "learning_rate": 0.0001523560209424084,
      "loss": 0.032,
      "step": 455
    },
    {
      "epoch": 2.41,
      "grad_norm": 0.018663793802261353,
      "learning_rate": 0.00015183246073298428,
      "loss": 0.0221,
      "step": 460
    },
    {
      "epoch": 2.43,
      "grad_norm": 1.0761280059814453,
      "learning_rate": 0.00015130890052356022,
      "loss": 0.0055,
      "step": 465
    },
    {
      "epoch": 2.46,
      "grad_norm": 0.8688761591911316,
      "learning_rate": 0.00015078534031413612,
      "loss": 0.0852,
      "step": 470
    },
    {
      "epoch": 2.49,
      "grad_norm": 0.06378360092639923,
      "learning_rate": 0.00015026178010471206,
      "loss": 0.0031,
      "step": 475
    },
    {
      "epoch": 2.51,
      "grad_norm": 7.236721992492676,
      "learning_rate": 0.00014973821989528797,
      "loss": 0.0106,
      "step": 480
    },
    {
      "epoch": 2.54,
      "grad_norm": 0.013875219039618969,
      "learning_rate": 0.00014921465968586388,
      "loss": 0.0032,
      "step": 485
    },
    {
      "epoch": 2.57,
      "grad_norm": 0.031233886256814003,
      "learning_rate": 0.0001486910994764398,
      "loss": 0.0047,
      "step": 490
    },
    {
      "epoch": 2.59,
      "grad_norm": 0.2528940439224243,
      "learning_rate": 0.00014816753926701572,
      "loss": 0.0081,
      "step": 495
    },
    {
      "epoch": 2.62,
      "grad_norm": 0.02138712629675865,
      "learning_rate": 0.00014764397905759163,
      "loss": 0.0161,
      "step": 500
    },
    {
      "epoch": 2.62,
      "eval_accuracy": 0.9514435695538058,
      "eval_loss": 0.16896109282970428,
      "eval_precision": 0.9537490891546931,
      "eval_recall": 0.9514435695538058,
      "eval_runtime": 17.5851,
      "eval_samples_per_second": 43.332,
      "eval_steps_per_second": 2.73,
      "step": 500
    },
    {
      "epoch": 2.64,
      "grad_norm": 5.179097652435303,
      "learning_rate": 0.00014712041884816754,
      "loss": 0.0271,
      "step": 505
    },
    {
      "epoch": 2.67,
      "grad_norm": 0.06704265624284744,
      "learning_rate": 0.00014659685863874347,
      "loss": 0.0072,
      "step": 510
    },
    {
      "epoch": 2.7,
      "grad_norm": 0.15809708833694458,
      "learning_rate": 0.00014607329842931938,
      "loss": 0.0027,
      "step": 515
    },
    {
      "epoch": 2.72,
      "grad_norm": 6.423127174377441,
      "learning_rate": 0.00014554973821989531,
      "loss": 0.0238,
      "step": 520
    },
    {
      "epoch": 2.75,
      "grad_norm": 0.010630112141370773,
      "learning_rate": 0.0001450261780104712,
      "loss": 0.0271,
      "step": 525
    },
    {
      "epoch": 2.77,
      "grad_norm": 2.8193652629852295,
      "learning_rate": 0.00014450261780104713,
      "loss": 0.1816,
      "step": 530
    },
    {
      "epoch": 2.8,
      "grad_norm": 0.004105111118406057,
      "learning_rate": 0.00014397905759162304,
      "loss": 0.1842,
      "step": 535
    },
    {
      "epoch": 2.83,
      "grad_norm": 0.20598988234996796,
      "learning_rate": 0.00014345549738219897,
      "loss": 0.0302,
      "step": 540
    },
    {
      "epoch": 2.85,
      "grad_norm": 0.7738168835639954,
      "learning_rate": 0.00014293193717277485,
      "loss": 0.0256,
      "step": 545
    },
    {
      "epoch": 2.88,
      "grad_norm": 0.07728663086891174,
      "learning_rate": 0.0001424083769633508,
      "loss": 0.0341,
      "step": 550
    },
    {
      "epoch": 2.88,
      "eval_accuracy": 0.9488188976377953,
      "eval_loss": 0.18862025439739227,
      "eval_precision": 0.9543073149026239,
      "eval_recall": 0.9488188976377953,
      "eval_runtime": 17.3957,
      "eval_samples_per_second": 43.804,
      "eval_steps_per_second": 2.759,
      "step": 550
    },
    {
      "epoch": 2.91,
      "grad_norm": 12.730446815490723,
      "learning_rate": 0.0001418848167539267,
      "loss": 0.0664,
      "step": 555
    },
    {
      "epoch": 2.93,
      "grad_norm": 8.560800552368164,
      "learning_rate": 0.00014136125654450263,
      "loss": 0.0518,
      "step": 560
    },
    {
      "epoch": 2.96,
      "grad_norm": 1.7465717792510986,
      "learning_rate": 0.00014083769633507854,
      "loss": 0.0236,
      "step": 565
    },
    {
      "epoch": 2.98,
      "grad_norm": 5.386238098144531,
      "learning_rate": 0.00014031413612565445,
      "loss": 0.0449,
      "step": 570
    },
    {
      "epoch": 3.01,
      "grad_norm": 0.008860424160957336,
      "learning_rate": 0.00013979057591623038,
      "loss": 0.0168,
      "step": 575
    },
    {
      "epoch": 3.04,
      "grad_norm": 0.0057926299050450325,
      "learning_rate": 0.0001392670157068063,
      "loss": 0.0191,
      "step": 580
    },
    {
      "epoch": 3.06,
      "grad_norm": 0.5266512036323547,
      "learning_rate": 0.00013874345549738223,
      "loss": 0.0023,
      "step": 585
    },
    {
      "epoch": 3.09,
      "grad_norm": 0.03807899355888367,
      "learning_rate": 0.0001382198952879581,
      "loss": 0.0259,
      "step": 590
    },
    {
      "epoch": 3.12,
      "grad_norm": 0.41590607166290283,
      "learning_rate": 0.00013769633507853404,
      "loss": 0.0224,
      "step": 595
    },
    {
      "epoch": 3.14,
      "grad_norm": 0.003112237202003598,
      "learning_rate": 0.00013717277486910995,
      "loss": 0.0022,
      "step": 600
    },
    {
      "epoch": 3.14,
      "eval_accuracy": 0.963254593175853,
      "eval_loss": 0.1624463051557541,
      "eval_precision": 0.9649964224541913,
      "eval_recall": 0.963254593175853,
      "eval_runtime": 17.4244,
      "eval_samples_per_second": 43.732,
      "eval_steps_per_second": 2.755,
      "step": 600
    },
    {
      "epoch": 3.17,
      "grad_norm": 0.013240986503660679,
      "learning_rate": 0.00013664921465968589,
      "loss": 0.0096,
      "step": 605
    },
    {
      "epoch": 3.19,
      "grad_norm": 0.13229414820671082,
      "learning_rate": 0.0001361256544502618,
      "loss": 0.0016,
      "step": 610
    },
    {
      "epoch": 3.22,
      "grad_norm": 0.2047564685344696,
      "learning_rate": 0.0001356020942408377,
      "loss": 0.0029,
      "step": 615
    },
    {
      "epoch": 3.25,
      "grad_norm": 5.075094699859619,
      "learning_rate": 0.0001350785340314136,
      "loss": 0.1541,
      "step": 620
    },
    {
      "epoch": 3.27,
      "grad_norm": 0.11219122260808945,
      "learning_rate": 0.00013455497382198955,
      "loss": 0.0016,
      "step": 625
    },
    {
      "epoch": 3.3,
      "grad_norm": 0.0553029403090477,
      "learning_rate": 0.00013403141361256545,
      "loss": 0.0016,
      "step": 630
    },
    {
      "epoch": 3.32,
      "grad_norm": 0.1759098470211029,
      "learning_rate": 0.00013350785340314136,
      "loss": 0.0616,
      "step": 635
    },
    {
      "epoch": 3.35,
      "grad_norm": 0.29212847352027893,
      "learning_rate": 0.00013298429319371727,
      "loss": 0.0074,
      "step": 640
    },
    {
      "epoch": 3.38,
      "grad_norm": 0.02130221202969551,
      "learning_rate": 0.0001324607329842932,
      "loss": 0.0013,
      "step": 645
    },
    {
      "epoch": 3.4,
      "grad_norm": 1.0668525695800781,
      "learning_rate": 0.0001319371727748691,
      "loss": 0.0044,
      "step": 650
    },
    {
      "epoch": 3.4,
      "eval_accuracy": 0.937007874015748,
      "eval_loss": 0.22285035252571106,
      "eval_precision": 0.9410975298976415,
      "eval_recall": 0.937007874015748,
      "eval_runtime": 17.5444,
      "eval_samples_per_second": 43.433,
      "eval_steps_per_second": 2.736,
      "step": 650
    },
    {
      "epoch": 3.43,
      "grad_norm": 0.27807894349098206,
      "learning_rate": 0.00013141361256544505,
      "loss": 0.0669,
      "step": 655
    },
    {
      "epoch": 3.46,
      "grad_norm": 0.17472220957279205,
      "learning_rate": 0.00013089005235602096,
      "loss": 0.0066,
      "step": 660
    },
    {
      "epoch": 3.48,
      "grad_norm": 0.07417164742946625,
      "learning_rate": 0.00013036649214659686,
      "loss": 0.0045,
      "step": 665
    },
    {
      "epoch": 3.51,
      "grad_norm": 0.007742606569081545,
      "learning_rate": 0.00012984293193717277,
      "loss": 0.0008,
      "step": 670
    },
    {
      "epoch": 3.53,
      "grad_norm": 0.00367438024841249,
      "learning_rate": 0.0001293193717277487,
      "loss": 0.003,
      "step": 675
    },
    {
      "epoch": 3.56,
      "grad_norm": 0.0222078375518322,
      "learning_rate": 0.00012879581151832462,
      "loss": 0.0017,
      "step": 680
    },
    {
      "epoch": 3.59,
      "grad_norm": 6.7436299324035645,
      "learning_rate": 0.00012827225130890052,
      "loss": 0.0179,
      "step": 685
    },
    {
      "epoch": 3.61,
      "grad_norm": 0.0018994753481820226,
      "learning_rate": 0.00012774869109947646,
      "loss": 0.0008,
      "step": 690
    },
    {
      "epoch": 3.64,
      "grad_norm": 0.005384970456361771,
      "learning_rate": 0.00012722513089005237,
      "loss": 0.0018,
      "step": 695
    },
    {
      "epoch": 3.66,
      "grad_norm": 0.0077237836085259914,
      "learning_rate": 0.00012670157068062827,
      "loss": 0.0012,
      "step": 700
    },
    {
      "epoch": 3.66,
      "eval_accuracy": 0.9501312335958005,
      "eval_loss": 0.17090748250484467,
      "eval_precision": 0.9525764956521452,
      "eval_recall": 0.9501312335958005,
      "eval_runtime": 17.5175,
      "eval_samples_per_second": 43.499,
      "eval_steps_per_second": 2.74,
      "step": 700
    },
    {
      "epoch": 3.69,
      "grad_norm": 0.016971301287412643,
      "learning_rate": 0.00012617801047120418,
      "loss": 0.0672,
      "step": 705
    },
    {
      "epoch": 3.72,
      "grad_norm": 0.37131184339523315,
      "learning_rate": 0.00012565445026178012,
      "loss": 0.0032,
      "step": 710
    },
    {
      "epoch": 3.74,
      "grad_norm": 0.020785262808203697,
      "learning_rate": 0.00012513089005235603,
      "loss": 0.0036,
      "step": 715
    },
    {
      "epoch": 3.77,
      "grad_norm": 0.003909711726009846,
      "learning_rate": 0.00012460732984293196,
      "loss": 0.0005,
      "step": 720
    },
    {
      "epoch": 3.8,
      "grad_norm": 0.32289111614227295,
      "learning_rate": 0.00012408376963350784,
      "loss": 0.0067,
      "step": 725
    },
    {
      "epoch": 3.82,
      "grad_norm": 0.011649747379124165,
      "learning_rate": 0.00012356020942408378,
      "loss": 0.0004,
      "step": 730
    },
    {
      "epoch": 3.85,
      "grad_norm": 0.009505286812782288,
      "learning_rate": 0.00012303664921465968,
      "loss": 0.0004,
      "step": 735
    },
    {
      "epoch": 3.87,
      "grad_norm": 0.016009971499443054,
      "learning_rate": 0.00012251308900523562,
      "loss": 0.0132,
      "step": 740
    },
    {
      "epoch": 3.9,
      "grad_norm": 0.07612348347902298,
      "learning_rate": 0.00012198952879581151,
      "loss": 0.0012,
      "step": 745
    },
    {
      "epoch": 3.93,
      "grad_norm": 0.05622846260666847,
      "learning_rate": 0.00012146596858638744,
      "loss": 0.0058,
      "step": 750
    },
    {
      "epoch": 3.93,
      "eval_accuracy": 0.958005249343832,
      "eval_loss": 0.17302347719669342,
      "eval_precision": 0.9602924010464745,
      "eval_recall": 0.958005249343832,
      "eval_runtime": 17.4895,
      "eval_samples_per_second": 43.569,
      "eval_steps_per_second": 2.744,
      "step": 750
    },
    {
      "epoch": 3.95,
      "grad_norm": 0.050630178302526474,
      "learning_rate": 0.00012094240837696336,
      "loss": 0.0008,
      "step": 755
    },
    {
      "epoch": 3.98,
      "grad_norm": 0.02956685610115528,
      "learning_rate": 0.00012041884816753928,
      "loss": 0.0017,
      "step": 760
    },
    {
      "epoch": 4.01,
      "grad_norm": 0.038056932389736176,
      "learning_rate": 0.0001198952879581152,
      "loss": 0.001,
      "step": 765
    },
    {
      "epoch": 4.03,
      "grad_norm": 0.006303295027464628,
      "learning_rate": 0.0001193717277486911,
      "loss": 0.0003,
      "step": 770
    },
    {
      "epoch": 4.06,
      "grad_norm": 0.019060682505369186,
      "learning_rate": 0.00011884816753926702,
      "loss": 0.0005,
      "step": 775
    },
    {
      "epoch": 4.08,
      "grad_norm": 0.9220014810562134,
      "learning_rate": 0.00011832460732984294,
      "loss": 0.0014,
      "step": 780
    },
    {
      "epoch": 4.11,
      "grad_norm": 0.0184442438185215,
      "learning_rate": 0.00011780104712041886,
      "loss": 0.0003,
      "step": 785
    },
    {
      "epoch": 4.14,
      "grad_norm": 0.003219619393348694,
      "learning_rate": 0.00011727748691099475,
      "loss": 0.0003,
      "step": 790
    },
    {
      "epoch": 4.16,
      "grad_norm": 0.006713965907692909,
      "learning_rate": 0.00011675392670157068,
      "loss": 0.0044,
      "step": 795
    },
    {
      "epoch": 4.19,
      "grad_norm": 0.0026614165399223566,
      "learning_rate": 0.0001162303664921466,
      "loss": 0.0003,
      "step": 800
    },
    {
      "epoch": 4.19,
      "eval_accuracy": 0.9488188976377953,
      "eval_loss": 0.20851939916610718,
      "eval_precision": 0.9513993632135537,
      "eval_recall": 0.9488188976377953,
      "eval_runtime": 17.4973,
      "eval_samples_per_second": 43.55,
      "eval_steps_per_second": 2.743,
      "step": 800
    },
    {
      "epoch": 4.21,
      "grad_norm": 0.2586454749107361,
      "learning_rate": 0.00011570680628272252,
      "loss": 0.001,
      "step": 805
    },
    {
      "epoch": 4.24,
      "grad_norm": 0.005604247096925974,
      "learning_rate": 0.00011518324607329844,
      "loss": 0.0027,
      "step": 810
    },
    {
      "epoch": 4.27,
      "grad_norm": 0.00321913487277925,
      "learning_rate": 0.00011465968586387435,
      "loss": 0.0002,
      "step": 815
    },
    {
      "epoch": 4.29,
      "grad_norm": 0.00729359732940793,
      "learning_rate": 0.00011413612565445027,
      "loss": 0.0068,
      "step": 820
    },
    {
      "epoch": 4.32,
      "grad_norm": 2.0007290840148926,
      "learning_rate": 0.00011361256544502619,
      "loss": 0.002,
      "step": 825
    },
    {
      "epoch": 4.35,
      "grad_norm": 0.0031692993361502886,
      "learning_rate": 0.00011308900523560211,
      "loss": 0.0005,
      "step": 830
    },
    {
      "epoch": 4.37,
      "grad_norm": 0.007637929171323776,
      "learning_rate": 0.00011256544502617801,
      "loss": 0.0002,
      "step": 835
    },
    {
      "epoch": 4.4,
      "grad_norm": 0.003380405716598034,
      "learning_rate": 0.00011204188481675393,
      "loss": 0.0002,
      "step": 840
    },
    {
      "epoch": 4.42,
      "grad_norm": 0.0019728606566786766,
      "learning_rate": 0.00011151832460732985,
      "loss": 0.0005,
      "step": 845
    },
    {
      "epoch": 4.45,
      "grad_norm": 0.0047022090293467045,
      "learning_rate": 0.00011099476439790577,
      "loss": 0.0003,
      "step": 850
    },
    {
      "epoch": 4.45,
      "eval_accuracy": 0.9698162729658792,
      "eval_loss": 0.1316651999950409,
      "eval_precision": 0.9708177710817824,
      "eval_recall": 0.9698162729658792,
      "eval_runtime": 17.4736,
      "eval_samples_per_second": 43.609,
      "eval_steps_per_second": 2.747,
      "step": 850
    },
    {
      "epoch": 4.48,
      "grad_norm": 0.003034041728824377,
      "learning_rate": 0.0001104712041884817,
      "loss": 0.0002,
      "step": 855
    },
    {
      "epoch": 4.5,
      "grad_norm": 0.0816386267542839,
      "learning_rate": 0.00010994764397905759,
      "loss": 0.0005,
      "step": 860
    },
    {
      "epoch": 4.53,
      "grad_norm": 0.010150614194571972,
      "learning_rate": 0.00010942408376963351,
      "loss": 0.0004,
      "step": 865
    },
    {
      "epoch": 4.55,
      "grad_norm": 0.003652422921732068,
      "learning_rate": 0.00010890052356020943,
      "loss": 0.0002,
      "step": 870
    },
    {
      "epoch": 4.58,
      "grad_norm": 0.0027291346341371536,
      "learning_rate": 0.00010837696335078535,
      "loss": 0.0002,
      "step": 875
    },
    {
      "epoch": 4.61,
      "grad_norm": 0.0021885058376938105,
      "learning_rate": 0.00010785340314136125,
      "loss": 0.0002,
      "step": 880
    },
    {
      "epoch": 4.63,
      "grad_norm": 0.0019014202989637852,
      "learning_rate": 0.00010732984293193717,
      "loss": 0.0002,
      "step": 885
    },
    {
      "epoch": 4.66,
      "grad_norm": 0.9188545942306519,
      "learning_rate": 0.00010680628272251309,
      "loss": 0.0032,
      "step": 890
    },
    {
      "epoch": 4.69,
      "grad_norm": 0.0015793447382748127,
      "learning_rate": 0.00010628272251308901,
      "loss": 0.0002,
      "step": 895
    },
    {
      "epoch": 4.71,
      "grad_norm": 0.0016003657365217805,
      "learning_rate": 0.00010575916230366492,
      "loss": 0.0002,
      "step": 900
    },
    {
      "epoch": 4.71,
      "eval_accuracy": 0.9698162729658792,
      "eval_loss": 0.12266764789819717,
      "eval_precision": 0.9709897987678096,
      "eval_recall": 0.9698162729658792,
      "eval_runtime": 17.6013,
      "eval_samples_per_second": 43.292,
      "eval_steps_per_second": 2.727,
      "step": 900
    },
    {
      "epoch": 4.74,
      "grad_norm": 0.001760563114657998,
      "learning_rate": 0.00010523560209424084,
      "loss": 0.0002,
      "step": 905
    },
    {
      "epoch": 4.76,
      "grad_norm": 0.14921420812606812,
      "learning_rate": 0.00010471204188481676,
      "loss": 0.0007,
      "step": 910
    },
    {
      "epoch": 4.79,
      "grad_norm": 0.008730238303542137,
      "learning_rate": 0.00010418848167539269,
      "loss": 0.0005,
      "step": 915
    },
    {
      "epoch": 4.82,
      "grad_norm": 0.0021068351343274117,
      "learning_rate": 0.0001036649214659686,
      "loss": 0.0004,
      "step": 920
    },
    {
      "epoch": 4.84,
      "grad_norm": 0.0018881463911384344,
      "learning_rate": 0.0001031413612565445,
      "loss": 0.0002,
      "step": 925
    },
    {
      "epoch": 4.87,
      "grad_norm": 0.010380023159086704,
      "learning_rate": 0.00010261780104712042,
      "loss": 0.0002,
      "step": 930
    },
    {
      "epoch": 4.9,
      "grad_norm": 0.005413604434579611,
      "learning_rate": 0.00010209424083769635,
      "loss": 0.0435,
      "step": 935
    },
    {
      "epoch": 4.92,
      "grad_norm": 0.013811178505420685,
      "learning_rate": 0.00010157068062827227,
      "loss": 0.0002,
      "step": 940
    },
    {
      "epoch": 4.95,
      "grad_norm": 0.0024414085783064365,
      "learning_rate": 0.00010104712041884816,
      "loss": 0.0014,
      "step": 945
    },
    {
      "epoch": 4.97,
      "grad_norm": 0.00615642499178648,
      "learning_rate": 0.00010052356020942408,
      "loss": 0.0004,
      "step": 950
    },
    {
      "epoch": 4.97,
      "eval_accuracy": 0.9711286089238845,
      "eval_loss": 0.12034053355455399,
      "eval_precision": 0.9732646926949845,
      "eval_recall": 0.9711286089238845,
      "eval_runtime": 17.6401,
      "eval_samples_per_second": 43.197,
      "eval_steps_per_second": 2.721,
      "step": 950
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.5034066438674927,
      "learning_rate": 0.0001,
      "loss": 0.0035,
      "step": 955
    },
    {
      "epoch": 5.03,
      "grad_norm": 0.0023136490490287542,
      "learning_rate": 9.947643979057593e-05,
      "loss": 0.0002,
      "step": 960
    },
    {
      "epoch": 5.05,
      "grad_norm": 0.025553325191140175,
      "learning_rate": 9.895287958115183e-05,
      "loss": 0.0003,
      "step": 965
    },
    {
      "epoch": 5.08,
      "grad_norm": 0.0021953664254397154,
      "learning_rate": 9.842931937172776e-05,
      "loss": 0.0002,
      "step": 970
    },
    {
      "epoch": 5.1,
      "grad_norm": 0.0016761135775595903,
      "learning_rate": 9.790575916230366e-05,
      "loss": 0.0002,
      "step": 975
    },
    {
      "epoch": 5.13,
      "grad_norm": 0.001983917085453868,
      "learning_rate": 9.738219895287959e-05,
      "loss": 0.0002,
      "step": 980
    },
    {
      "epoch": 5.16,
      "grad_norm": 0.008785477839410305,
      "learning_rate": 9.68586387434555e-05,
      "loss": 0.0003,
      "step": 985
    },
    {
      "epoch": 5.18,
      "grad_norm": 0.00397130474448204,
      "learning_rate": 9.633507853403142e-05,
      "loss": 0.0002,
      "step": 990
    },
    {
      "epoch": 5.21,
      "grad_norm": 0.0015918458811938763,
      "learning_rate": 9.581151832460732e-05,
      "loss": 0.0003,
      "step": 995
    },
    {
      "epoch": 5.24,
      "grad_norm": 0.0018075347179546952,
      "learning_rate": 9.528795811518324e-05,
      "loss": 0.0002,
      "step": 1000
    },
    {
      "epoch": 5.24,
      "eval_accuracy": 0.958005249343832,
      "eval_loss": 0.19232842326164246,
      "eval_precision": 0.960493858579392,
      "eval_recall": 0.958005249343832,
      "eval_runtime": 17.4659,
      "eval_samples_per_second": 43.628,
      "eval_steps_per_second": 2.748,
      "step": 1000
    },
    {
      "epoch": 5.26,
      "grad_norm": 0.0014983582077547908,
      "learning_rate": 9.476439790575917e-05,
      "loss": 0.0003,
      "step": 1005
    },
    {
      "epoch": 5.29,
      "grad_norm": 0.004939761478453875,
      "learning_rate": 9.424083769633509e-05,
      "loss": 0.0025,
      "step": 1010
    },
    {
      "epoch": 5.31,
      "grad_norm": 0.003176321741193533,
      "learning_rate": 9.371727748691101e-05,
      "loss": 0.0002,
      "step": 1015
    },
    {
      "epoch": 5.34,
      "grad_norm": 0.008645892143249512,
      "learning_rate": 9.319371727748692e-05,
      "loss": 0.0002,
      "step": 1020
    },
    {
      "epoch": 5.37,
      "grad_norm": 0.009872050024569035,
      "learning_rate": 9.267015706806284e-05,
      "loss": 0.0002,
      "step": 1025
    },
    {
      "epoch": 5.39,
      "grad_norm": 0.0015197212342172861,
      "learning_rate": 9.214659685863875e-05,
      "loss": 0.0001,
      "step": 1030
    },
    {
      "epoch": 5.42,
      "grad_norm": 0.0035685019101947546,
      "learning_rate": 9.162303664921467e-05,
      "loss": 0.0002,
      "step": 1035
    },
    {
      "epoch": 5.45,
      "grad_norm": 0.0015960617456585169,
      "learning_rate": 9.109947643979058e-05,
      "loss": 0.0002,
      "step": 1040
    },
    {
      "epoch": 5.47,
      "grad_norm": 0.0019863061606884003,
      "learning_rate": 9.05759162303665e-05,
      "loss": 0.0036,
      "step": 1045
    },
    {
      "epoch": 5.5,
      "grad_norm": 0.0011158788111060858,
      "learning_rate": 9.00523560209424e-05,
      "loss": 0.0001,
      "step": 1050
    },
    {
      "epoch": 5.5,
      "eval_accuracy": 0.9711286089238845,
      "eval_loss": 0.11267632246017456,
      "eval_precision": 0.9717079331418645,
      "eval_recall": 0.9711286089238845,
      "eval_runtime": 17.5861,
      "eval_samples_per_second": 43.33,
      "eval_steps_per_second": 2.729,
      "step": 1050
    },
    {
      "epoch": 5.52,
      "grad_norm": 0.003642949042841792,
      "learning_rate": 8.952879581151833e-05,
      "loss": 0.0002,
      "step": 1055
    },
    {
      "epoch": 5.55,
      "grad_norm": 2.638040781021118,
      "learning_rate": 8.900523560209425e-05,
      "loss": 0.0021,
      "step": 1060
    },
    {
      "epoch": 5.58,
      "grad_norm": 5.575833320617676,
      "learning_rate": 8.848167539267016e-05,
      "loss": 0.0068,
      "step": 1065
    },
    {
      "epoch": 5.6,
      "grad_norm": 0.0017755662556737661,
      "learning_rate": 8.795811518324608e-05,
      "loss": 0.0001,
      "step": 1070
    },
    {
      "epoch": 5.63,
      "grad_norm": 0.0034077195450663567,
      "learning_rate": 8.743455497382199e-05,
      "loss": 0.0003,
      "step": 1075
    },
    {
      "epoch": 5.65,
      "grad_norm": 0.003383865812793374,
      "learning_rate": 8.691099476439791e-05,
      "loss": 0.0004,
      "step": 1080
    },
    {
      "epoch": 5.68,
      "grad_norm": 0.0037927927915006876,
      "learning_rate": 8.638743455497382e-05,
      "loss": 0.0025,
      "step": 1085
    },
    {
      "epoch": 5.71,
      "grad_norm": 0.008559320122003555,
      "learning_rate": 8.586387434554974e-05,
      "loss": 0.0002,
      "step": 1090
    },
    {
      "epoch": 5.73,
      "grad_norm": 0.0041591702029109,
      "learning_rate": 8.534031413612566e-05,
      "loss": 0.0002,
      "step": 1095
    },
    {
      "epoch": 5.76,
      "grad_norm": 0.0011583847226575017,
      "learning_rate": 8.481675392670158e-05,
      "loss": 0.0003,
      "step": 1100
    },
    {
      "epoch": 5.76,
      "eval_accuracy": 0.9711286089238845,
      "eval_loss": 0.13462768495082855,
      "eval_precision": 0.9722333613195212,
      "eval_recall": 0.9711286089238845,
      "eval_runtime": 17.6983,
      "eval_samples_per_second": 43.055,
      "eval_steps_per_second": 2.712,
      "step": 1100
    },
    {
      "epoch": 5.79,
      "grad_norm": 0.0030487922485917807,
      "learning_rate": 8.429319371727749e-05,
      "loss": 0.0002,
      "step": 1105
    },
    {
      "epoch": 5.81,
      "grad_norm": 0.003955108113586903,
      "learning_rate": 8.376963350785341e-05,
      "loss": 0.0002,
      "step": 1110
    },
    {
      "epoch": 5.84,
      "grad_norm": 0.002146199345588684,
      "learning_rate": 8.324607329842933e-05,
      "loss": 0.0001,
      "step": 1115
    },
    {
      "epoch": 5.86,
      "grad_norm": 0.0010180504759773612,
      "learning_rate": 8.272251308900524e-05,
      "loss": 0.0002,
      "step": 1120
    },
    {
      "epoch": 5.89,
      "grad_norm": 0.0008741291239857674,
      "learning_rate": 8.219895287958116e-05,
      "loss": 0.0001,
      "step": 1125
    },
    {
      "epoch": 5.92,
      "grad_norm": 0.0017273599514737725,
      "learning_rate": 8.167539267015707e-05,
      "loss": 0.0071,
      "step": 1130
    },
    {
      "epoch": 5.94,
      "grad_norm": 0.003460003063082695,
      "learning_rate": 8.115183246073299e-05,
      "loss": 0.0006,
      "step": 1135
    },
    {
      "epoch": 5.97,
      "grad_norm": 0.0010187524603679776,
      "learning_rate": 8.06282722513089e-05,
      "loss": 0.0001,
      "step": 1140
    },
    {
      "epoch": 5.99,
      "grad_norm": 0.0035252266097813845,
      "learning_rate": 8.010471204188482e-05,
      "loss": 0.0009,
      "step": 1145
    },
    {
      "epoch": 6.02,
      "grad_norm": 0.0013756396947428584,
      "learning_rate": 7.958115183246073e-05,
      "loss": 0.0001,
      "step": 1150
    },
    {
      "epoch": 6.02,
      "eval_accuracy": 0.9711286089238845,
      "eval_loss": 0.11907853186130524,
      "eval_precision": 0.9719758650287736,
      "eval_recall": 0.9711286089238845,
      "eval_runtime": 18.1879,
      "eval_samples_per_second": 41.896,
      "eval_steps_per_second": 2.639,
      "step": 1150
    },
    {
      "epoch": 6.05,
      "grad_norm": 0.0016478670295327902,
      "learning_rate": 7.905759162303665e-05,
      "loss": 0.0001,
      "step": 1155
    },
    {
      "epoch": 6.07,
      "grad_norm": 0.0019091927679255605,
      "learning_rate": 7.853403141361257e-05,
      "loss": 0.0001,
      "step": 1160
    },
    {
      "epoch": 6.1,
      "grad_norm": 0.00418177992105484,
      "learning_rate": 7.801047120418848e-05,
      "loss": 0.0001,
      "step": 1165
    },
    {
      "epoch": 6.13,
      "grad_norm": 0.0010454648872837424,
      "learning_rate": 7.74869109947644e-05,
      "loss": 0.0001,
      "step": 1170
    },
    {
      "epoch": 6.15,
      "grad_norm": 0.003557993331924081,
      "learning_rate": 7.696335078534031e-05,
      "loss": 0.0001,
      "step": 1175
    },
    {
      "epoch": 6.18,
      "grad_norm": 0.00358203100040555,
      "learning_rate": 7.643979057591623e-05,
      "loss": 0.0002,
      "step": 1180
    },
    {
      "epoch": 6.2,
      "grad_norm": 0.0017727952217683196,
      "learning_rate": 7.591623036649214e-05,
      "loss": 0.0002,
      "step": 1185
    },
    {
      "epoch": 6.23,
      "grad_norm": 0.056761130690574646,
      "learning_rate": 7.539267015706806e-05,
      "loss": 0.0003,
      "step": 1190
    },
    {
      "epoch": 6.26,
      "grad_norm": 0.0010181496618315578,
      "learning_rate": 7.486910994764398e-05,
      "loss": 0.0003,
      "step": 1195
    },
    {
      "epoch": 6.28,
      "grad_norm": 0.003052978776395321,
      "learning_rate": 7.43455497382199e-05,
      "loss": 0.0001,
      "step": 1200
    },
    {
      "epoch": 6.28,
      "eval_accuracy": 0.968503937007874,
      "eval_loss": 0.1217927485704422,
      "eval_precision": 0.9692342594033277,
      "eval_recall": 0.968503937007874,
      "eval_runtime": 17.611,
      "eval_samples_per_second": 43.268,
      "eval_steps_per_second": 2.726,
      "step": 1200
    },
    {
      "epoch": 6.31,
      "grad_norm": 0.0031509266700595617,
      "learning_rate": 7.382198952879581e-05,
      "loss": 0.0002,
      "step": 1205
    },
    {
      "epoch": 6.34,
      "grad_norm": 0.0011931936023756862,
      "learning_rate": 7.329842931937174e-05,
      "loss": 0.0002,
      "step": 1210
    },
    {
      "epoch": 6.36,
      "grad_norm": 0.0014170074136927724,
      "learning_rate": 7.277486910994766e-05,
      "loss": 0.0002,
      "step": 1215
    },
    {
      "epoch": 6.39,
      "grad_norm": 0.0008681151666678488,
      "learning_rate": 7.225130890052356e-05,
      "loss": 0.0002,
      "step": 1220
    },
    {
      "epoch": 6.41,
      "grad_norm": 0.0015088602667674422,
      "learning_rate": 7.172774869109949e-05,
      "loss": 0.0002,
      "step": 1225
    },
    {
      "epoch": 6.44,
      "grad_norm": 0.0011560139246284962,
      "learning_rate": 7.12041884816754e-05,
      "loss": 0.0002,
      "step": 1230
    },
    {
      "epoch": 6.47,
      "grad_norm": 0.0059287999756634235,
      "learning_rate": 7.068062827225132e-05,
      "loss": 0.0002,
      "step": 1235
    },
    {
      "epoch": 6.49,
      "grad_norm": 0.0017274218844249845,
      "learning_rate": 7.015706806282722e-05,
      "loss": 0.0002,
      "step": 1240
    },
    {
      "epoch": 6.52,
      "grad_norm": 0.000886244117282331,
      "learning_rate": 6.963350785340315e-05,
      "loss": 0.0001,
      "step": 1245
    },
    {
      "epoch": 6.54,
      "grad_norm": 0.0028559649363160133,
      "learning_rate": 6.910994764397905e-05,
      "loss": 0.0002,
      "step": 1250
    },
    {
      "epoch": 6.54,
      "eval_accuracy": 0.968503937007874,
      "eval_loss": 0.11844318360090256,
      "eval_precision": 0.9692177012971396,
      "eval_recall": 0.968503937007874,
      "eval_runtime": 17.5917,
      "eval_samples_per_second": 43.316,
      "eval_steps_per_second": 2.729,
      "step": 1250
    },
    {
      "epoch": 6.57,
      "grad_norm": 0.0011471702018752694,
      "learning_rate": 6.858638743455498e-05,
      "loss": 0.0001,
      "step": 1255
    },
    {
      "epoch": 6.6,
      "grad_norm": 0.0015491463709622622,
      "learning_rate": 6.80628272251309e-05,
      "loss": 0.0001,
      "step": 1260
    },
    {
      "epoch": 6.62,
      "grad_norm": 0.0011632316745817661,
      "learning_rate": 6.75392670157068e-05,
      "loss": 0.0001,
      "step": 1265
    },
    {
      "epoch": 6.65,
      "grad_norm": 0.006089572329074144,
      "learning_rate": 6.701570680628273e-05,
      "loss": 0.0001,
      "step": 1270
    },
    {
      "epoch": 6.68,
      "grad_norm": 0.0025600052904337645,
      "learning_rate": 6.649214659685863e-05,
      "loss": 0.0001,
      "step": 1275
    },
    {
      "epoch": 6.7,
      "grad_norm": 0.001310704043135047,
      "learning_rate": 6.596858638743456e-05,
      "loss": 0.0001,
      "step": 1280
    },
    {
      "epoch": 6.73,
      "grad_norm": 0.0050185080617666245,
      "learning_rate": 6.544502617801048e-05,
      "loss": 0.0001,
      "step": 1285
    },
    {
      "epoch": 6.75,
      "grad_norm": 0.0030103351455181837,
      "learning_rate": 6.492146596858639e-05,
      "loss": 0.0001,
      "step": 1290
    },
    {
      "epoch": 6.78,
      "grad_norm": 0.0009425992611795664,
      "learning_rate": 6.439790575916231e-05,
      "loss": 0.0001,
      "step": 1295
    },
    {
      "epoch": 6.81,
      "grad_norm": 0.0016329910140484571,
      "learning_rate": 6.387434554973823e-05,
      "loss": 0.0002,
      "step": 1300
    },
    {
      "epoch": 6.81,
      "eval_accuracy": 0.9698162729658792,
      "eval_loss": 0.11775030940771103,
      "eval_precision": 0.9703878401175062,
      "eval_recall": 0.9698162729658792,
      "eval_runtime": 17.4725,
      "eval_samples_per_second": 43.611,
      "eval_steps_per_second": 2.747,
      "step": 1300
    },
    {
      "epoch": 6.83,
      "grad_norm": 0.0049902000464499,
      "learning_rate": 6.335078534031414e-05,
      "loss": 0.0002,
      "step": 1305
    },
    {
      "epoch": 6.86,
      "grad_norm": 0.0012894049286842346,
      "learning_rate": 6.282722513089006e-05,
      "loss": 0.0001,
      "step": 1310
    },
    {
      "epoch": 6.88,
      "grad_norm": 0.0013533816672861576,
      "learning_rate": 6.230366492146598e-05,
      "loss": 0.0001,
      "step": 1315
    },
    {
      "epoch": 6.91,
      "grad_norm": 0.0015272346790879965,
      "learning_rate": 6.178010471204189e-05,
      "loss": 0.0001,
      "step": 1320
    },
    {
      "epoch": 6.94,
      "grad_norm": 0.0021596469450742006,
      "learning_rate": 6.125654450261781e-05,
      "loss": 0.0001,
      "step": 1325
    },
    {
      "epoch": 6.96,
      "grad_norm": 0.0010469523258507252,
      "learning_rate": 6.073298429319372e-05,
      "loss": 0.0001,
      "step": 1330
    },
    {
      "epoch": 6.99,
      "grad_norm": 0.004506214987486601,
      "learning_rate": 6.020942408376964e-05,
      "loss": 0.0001,
      "step": 1335
    },
    {
      "epoch": 7.02,
      "grad_norm": 0.002483614254742861,
      "learning_rate": 5.968586387434555e-05,
      "loss": 0.0001,
      "step": 1340
    },
    {
      "epoch": 7.04,
      "grad_norm": 0.0011110632913187146,
      "learning_rate": 5.916230366492147e-05,
      "loss": 0.0001,
      "step": 1345
    },
    {
      "epoch": 7.07,
      "grad_norm": 0.0009813542710617185,
      "learning_rate": 5.863874345549738e-05,
      "loss": 0.0001,
      "step": 1350
    },
    {
      "epoch": 7.07,
      "eval_accuracy": 0.9698162729658792,
      "eval_loss": 0.1171058714389801,
      "eval_precision": 0.9703878401175062,
      "eval_recall": 0.9698162729658792,
      "eval_runtime": 17.4755,
      "eval_samples_per_second": 43.604,
      "eval_steps_per_second": 2.747,
      "step": 1350
    },
    {
      "epoch": 7.09,
      "grad_norm": 0.0015625334344804287,
      "learning_rate": 5.81151832460733e-05,
      "loss": 0.0001,
      "step": 1355
    },
    {
      "epoch": 7.12,
      "grad_norm": 0.0011336584575474262,
      "learning_rate": 5.759162303664922e-05,
      "loss": 0.0001,
      "step": 1360
    },
    {
      "epoch": 7.15,
      "grad_norm": 0.002448306418955326,
      "learning_rate": 5.7068062827225135e-05,
      "loss": 0.0001,
      "step": 1365
    },
    {
      "epoch": 7.17,
      "grad_norm": 0.003220533486455679,
      "learning_rate": 5.654450261780106e-05,
      "loss": 0.0001,
      "step": 1370
    },
    {
      "epoch": 7.2,
      "grad_norm": 0.0035826002713292837,
      "learning_rate": 5.6020942408376965e-05,
      "loss": 0.0001,
      "step": 1375
    },
    {
      "epoch": 7.23,
      "grad_norm": 0.0011589195346459746,
      "learning_rate": 5.5497382198952887e-05,
      "loss": 0.0001,
      "step": 1380
    },
    {
      "epoch": 7.25,
      "grad_norm": 0.00346089294180274,
      "learning_rate": 5.4973821989528795e-05,
      "loss": 0.0001,
      "step": 1385
    },
    {
      "epoch": 7.28,
      "grad_norm": 0.0009584302315488458,
      "learning_rate": 5.4450261780104716e-05,
      "loss": 0.0001,
      "step": 1390
    },
    {
      "epoch": 7.3,
      "grad_norm": 0.0017620971193537116,
      "learning_rate": 5.3926701570680624e-05,
      "loss": 0.0001,
      "step": 1395
    },
    {
      "epoch": 7.33,
      "grad_norm": 0.0009925129124894738,
      "learning_rate": 5.3403141361256546e-05,
      "loss": 0.0001,
      "step": 1400
    },
    {
      "epoch": 7.33,
      "eval_accuracy": 0.9698162729658792,
      "eval_loss": 0.11688791215419769,
      "eval_precision": 0.9704864266182309,
      "eval_recall": 0.9698162729658792,
      "eval_runtime": 17.5351,
      "eval_samples_per_second": 43.456,
      "eval_steps_per_second": 2.737,
      "step": 1400
    },
    {
      "epoch": 7.36,
      "grad_norm": 0.0010989775182679296,
      "learning_rate": 5.287958115183246e-05,
      "loss": 0.0001,
      "step": 1405
    },
    {
      "epoch": 7.38,
      "grad_norm": 0.001141382148489356,
      "learning_rate": 5.235602094240838e-05,
      "loss": 0.0001,
      "step": 1410
    },
    {
      "epoch": 7.41,
      "grad_norm": 0.0018131214892491698,
      "learning_rate": 5.18324607329843e-05,
      "loss": 0.0001,
      "step": 1415
    },
    {
      "epoch": 7.43,
      "grad_norm": 0.0016311927465721965,
      "learning_rate": 5.130890052356021e-05,
      "loss": 0.0002,
      "step": 1420
    },
    {
      "epoch": 7.46,
      "grad_norm": 0.0019679146353155375,
      "learning_rate": 5.0785340314136134e-05,
      "loss": 0.0001,
      "step": 1425
    },
    {
      "epoch": 7.49,
      "grad_norm": 0.0030663602519780397,
      "learning_rate": 5.026178010471204e-05,
      "loss": 0.0001,
      "step": 1430
    },
    {
      "epoch": 7.51,
      "grad_norm": 0.0022476576268672943,
      "learning_rate": 4.973821989528796e-05,
      "loss": 0.0001,
      "step": 1435
    },
    {
      "epoch": 7.54,
      "grad_norm": 0.001257985015399754,
      "learning_rate": 4.921465968586388e-05,
      "loss": 0.0001,
      "step": 1440
    },
    {
      "epoch": 7.57,
      "grad_norm": 0.002487305086106062,
      "learning_rate": 4.869109947643979e-05,
      "loss": 0.0001,
      "step": 1445
    },
    {
      "epoch": 7.59,
      "grad_norm": 0.0023026170674711466,
      "learning_rate": 4.816753926701571e-05,
      "loss": 0.0001,
      "step": 1450
    },
    {
      "epoch": 7.59,
      "eval_accuracy": 0.9711286089238845,
      "eval_loss": 0.11671730875968933,
      "eval_precision": 0.9718133010833329,
      "eval_recall": 0.9711286089238845,
      "eval_runtime": 17.5661,
      "eval_samples_per_second": 43.379,
      "eval_steps_per_second": 2.733,
      "step": 1450
    },
    {
      "epoch": 7.62,
      "grad_norm": 0.0014070291072130203,
      "learning_rate": 4.764397905759162e-05,
      "loss": 0.0001,
      "step": 1455
    },
    {
      "epoch": 7.64,
      "grad_norm": 0.0011672518448904157,
      "learning_rate": 4.7120418848167544e-05,
      "loss": 0.0001,
      "step": 1460
    },
    {
      "epoch": 7.67,
      "grad_norm": 0.0015011669602245092,
      "learning_rate": 4.659685863874346e-05,
      "loss": 0.0001,
      "step": 1465
    },
    {
      "epoch": 7.7,
      "grad_norm": 0.001993841025978327,
      "learning_rate": 4.6073298429319374e-05,
      "loss": 0.0001,
      "step": 1470
    },
    {
      "epoch": 7.72,
      "grad_norm": 0.0008604762842878699,
      "learning_rate": 4.554973821989529e-05,
      "loss": 0.0001,
      "step": 1475
    },
    {
      "epoch": 7.75,
      "grad_norm": 0.002782499184831977,
      "learning_rate": 4.50261780104712e-05,
      "loss": 0.0001,
      "step": 1480
    },
    {
      "epoch": 7.77,
      "grad_norm": 0.002074967371299863,
      "learning_rate": 4.4502617801047125e-05,
      "loss": 0.0001,
      "step": 1485
    },
    {
      "epoch": 7.8,
      "grad_norm": 0.0031987845432013273,
      "learning_rate": 4.397905759162304e-05,
      "loss": 0.0001,
      "step": 1490
    },
    {
      "epoch": 7.83,
      "grad_norm": 0.0007084137760102749,
      "learning_rate": 4.3455497382198955e-05,
      "loss": 0.0001,
      "step": 1495
    },
    {
      "epoch": 7.85,
      "grad_norm": 0.0019736553076654673,
      "learning_rate": 4.293193717277487e-05,
      "loss": 0.0001,
      "step": 1500
    },
    {
      "epoch": 7.85,
      "eval_accuracy": 0.9711286089238845,
      "eval_loss": 0.11646050214767456,
      "eval_precision": 0.9718133010833329,
      "eval_recall": 0.9711286089238845,
      "eval_runtime": 17.4763,
      "eval_samples_per_second": 43.602,
      "eval_steps_per_second": 2.747,
      "step": 1500
    }
  ],
  "logging_steps": 5,
  "max_steps": 1910,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1.8551418818712515e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}