{
  "best_global_step": 6000,
  "best_metric": 0.13940538465976715,
  "best_model_checkpoint": "/data/alamparan/mattext_ckpt/results_2m/2026-02-06/03-23-25/pretrain/checkpoints/robocrys_rep_test-pretrain/checkpoint-6000",
  "epoch": 0.5780903747952597,
  "eval_steps": 50,
  "global_step": 6000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.004817419789960497,
      "grad_norm": 1.3140820264816284,
      "learning_rate": 0.00019998111571442338,
      "loss": 5.9543,
      "step": 50
    },
    {
      "epoch": 0.004817419789960497,
      "eval_loss": 4.040468692779541,
      "eval_runtime": 85.9858,
      "eval_samples_per_second": 221.013,
      "eval_steps_per_second": 4.605,
      "step": 50
    },
    {
      "epoch": 0.009634839579920994,
      "grad_norm": 0.8951072692871094,
      "learning_rate": 0.0001999618460352635,
      "loss": 3.7958,
      "step": 100
    },
    {
      "epoch": 0.009634839579920994,
      "eval_loss": 3.572112798690796,
      "eval_runtime": 87.5351,
      "eval_samples_per_second": 217.102,
      "eval_steps_per_second": 4.524,
      "step": 100
    },
    {
      "epoch": 0.014452259369881492,
      "grad_norm": 1.1552037000656128,
      "learning_rate": 0.00019994257635610368,
      "loss": 3.4955,
      "step": 150
    },
    {
      "epoch": 0.014452259369881492,
      "eval_loss": 3.365239381790161,
      "eval_runtime": 86.7746,
      "eval_samples_per_second": 219.004,
      "eval_steps_per_second": 4.564,
      "step": 150
    },
    {
      "epoch": 0.019269679159841988,
      "grad_norm": 1.1728101968765259,
      "learning_rate": 0.00019992330667694384,
      "loss": 3.3176,
      "step": 200
    },
    {
      "epoch": 0.019269679159841988,
      "eval_loss": 3.2046337127685547,
      "eval_runtime": 86.8432,
      "eval_samples_per_second": 218.831,
      "eval_steps_per_second": 4.56,
      "step": 200
    },
    {
      "epoch": 0.024087098949802484,
      "grad_norm": 0.8054835200309753,
      "learning_rate": 0.000199904036997784,
      "loss": 3.1815,
      "step": 250
    },
    {
      "epoch": 0.024087098949802484,
      "eval_loss": 3.065305709838867,
      "eval_runtime": 85.5782,
      "eval_samples_per_second": 222.066,
      "eval_steps_per_second": 4.627,
      "step": 250
    },
    {
      "epoch": 0.028904518739762984,
      "grad_norm": 1.2382686138153076,
      "learning_rate": 0.00019988476731862417,
      "loss": 3.0462,
      "step": 300
    },
    {
      "epoch": 0.028904518739762984,
      "eval_loss": 2.937901020050049,
      "eval_runtime": 86.036,
      "eval_samples_per_second": 220.884,
      "eval_steps_per_second": 4.603,
      "step": 300
    },
    {
      "epoch": 0.03372193852972348,
      "grad_norm": 1.245282530784607,
      "learning_rate": 0.00019986549763946433,
      "loss": 2.9173,
      "step": 350
    },
    {
      "epoch": 0.03372193852972348,
      "eval_loss": 2.7557334899902344,
      "eval_runtime": 88.8138,
      "eval_samples_per_second": 213.976,
      "eval_steps_per_second": 4.459,
      "step": 350
    },
    {
      "epoch": 0.038539358319683976,
      "grad_norm": 1.2060555219650269,
      "learning_rate": 0.00019984622796030447,
      "loss": 2.7187,
      "step": 400
    },
    {
      "epoch": 0.038539358319683976,
      "eval_loss": 2.480039119720459,
      "eval_runtime": 88.2775,
      "eval_samples_per_second": 215.276,
      "eval_steps_per_second": 4.486,
      "step": 400
    },
    {
      "epoch": 0.04335677810964447,
      "grad_norm": 1.9137617349624634,
      "learning_rate": 0.00019982695828114463,
      "loss": 2.3768,
      "step": 450
    },
    {
      "epoch": 0.04335677810964447,
      "eval_loss": 1.9087786674499512,
      "eval_runtime": 88.071,
      "eval_samples_per_second": 215.78,
      "eval_steps_per_second": 4.496,
      "step": 450
    },
    {
      "epoch": 0.04817419789960497,
      "grad_norm": 1.4742215871810913,
      "learning_rate": 0.00019980768860198477,
      "loss": 1.7234,
      "step": 500
    },
    {
      "epoch": 0.04817419789960497,
      "eval_loss": 1.158734679222107,
      "eval_runtime": 88.5849,
      "eval_samples_per_second": 214.529,
      "eval_steps_per_second": 4.47,
      "step": 500
    },
    {
      "epoch": 0.05299161768956547,
      "grad_norm": 1.0141704082489014,
      "learning_rate": 0.00019978841892282493,
      "loss": 1.1418,
      "step": 550
    },
    {
      "epoch": 0.05299161768956547,
      "eval_loss": 0.8516036868095398,
      "eval_runtime": 87.5002,
      "eval_samples_per_second": 217.188,
      "eval_steps_per_second": 4.526,
      "step": 550
    },
    {
      "epoch": 0.05780903747952597,
      "grad_norm": 0.8414612412452698,
      "learning_rate": 0.0001997691492436651,
      "loss": 0.881,
      "step": 600
    },
    {
      "epoch": 0.05780903747952597,
      "eval_loss": 0.7400951385498047,
      "eval_runtime": 88.6088,
      "eval_samples_per_second": 214.471,
      "eval_steps_per_second": 4.469,
      "step": 600
    },
    {
      "epoch": 0.06262645726948646,
      "grad_norm": 0.694267213344574,
      "learning_rate": 0.00019974987956450526,
      "loss": 0.813,
      "step": 650
    },
    {
      "epoch": 0.06262645726948646,
      "eval_loss": 0.6765585541725159,
      "eval_runtime": 90.0063,
      "eval_samples_per_second": 211.141,
      "eval_steps_per_second": 4.4,
      "step": 650
    },
    {
      "epoch": 0.06744387705944696,
      "grad_norm": 0.6236635446548462,
      "learning_rate": 0.00019973060988534542,
      "loss": 0.7301,
      "step": 700
    },
    {
      "epoch": 0.06744387705944696,
      "eval_loss": 0.6367717981338501,
      "eval_runtime": 90.1272,
      "eval_samples_per_second": 210.857,
      "eval_steps_per_second": 4.394,
      "step": 700
    },
    {
      "epoch": 0.07226129684940746,
      "grad_norm": 0.663361132144928,
      "learning_rate": 0.00019971134020618558,
      "loss": 0.6936,
      "step": 750
    },
    {
      "epoch": 0.07226129684940746,
      "eval_loss": 0.603342592716217,
      "eval_runtime": 88.3855,
      "eval_samples_per_second": 215.013,
      "eval_steps_per_second": 4.48,
      "step": 750
    },
    {
      "epoch": 0.07707871663936795,
      "grad_norm": 0.584972083568573,
      "learning_rate": 0.00019969207052702575,
      "loss": 0.6492,
      "step": 800
    },
    {
      "epoch": 0.07707871663936795,
      "eval_loss": 0.5670896768569946,
      "eval_runtime": 88.8592,
      "eval_samples_per_second": 213.867,
      "eval_steps_per_second": 4.456,
      "step": 800
    },
    {
      "epoch": 0.08189613642932846,
      "grad_norm": 0.6252435445785522,
      "learning_rate": 0.00019967280084786588,
      "loss": 0.6554,
      "step": 850
    },
    {
      "epoch": 0.08189613642932846,
      "eval_loss": 0.5390716791152954,
      "eval_runtime": 88.7469,
      "eval_samples_per_second": 214.137,
      "eval_steps_per_second": 4.462,
      "step": 850
    },
    {
      "epoch": 0.08671355621928895,
      "grad_norm": 0.5381043553352356,
      "learning_rate": 0.00019965353116870605,
      "loss": 0.5847,
      "step": 900
    },
    {
      "epoch": 0.08671355621928895,
      "eval_loss": 0.5227247476577759,
      "eval_runtime": 87.1283,
      "eval_samples_per_second": 218.115,
      "eval_steps_per_second": 4.545,
      "step": 900
    },
    {
      "epoch": 0.09153097600924945,
      "grad_norm": 0.4695926308631897,
      "learning_rate": 0.0001996342614895462,
      "loss": 0.5787,
      "step": 950
    },
    {
      "epoch": 0.09153097600924945,
      "eval_loss": 0.5046476721763611,
      "eval_runtime": 88.0204,
      "eval_samples_per_second": 215.905,
      "eval_steps_per_second": 4.499,
      "step": 950
    },
    {
      "epoch": 0.09634839579920994,
      "grad_norm": 0.49100059270858765,
      "learning_rate": 0.00019961499181038637,
      "loss": 0.5795,
      "step": 1000
    },
    {
      "epoch": 0.09634839579920994,
      "eval_loss": 0.4904717803001404,
      "eval_runtime": 87.871,
      "eval_samples_per_second": 216.272,
      "eval_steps_per_second": 4.507,
      "step": 1000
    },
    {
      "epoch": 0.10116581558917044,
      "grad_norm": 0.48738670349121094,
      "learning_rate": 0.00019959572213122654,
      "loss": 0.5372,
      "step": 1050
    },
    {
      "epoch": 0.10116581558917044,
      "eval_loss": 0.4725087285041809,
      "eval_runtime": 87.3641,
      "eval_samples_per_second": 217.526,
      "eval_steps_per_second": 4.533,
      "step": 1050
    },
    {
      "epoch": 0.10598323537913094,
      "grad_norm": 0.5932963490486145,
      "learning_rate": 0.0001995764524520667,
      "loss": 0.5074,
      "step": 1100
    },
    {
      "epoch": 0.10598323537913094,
      "eval_loss": 0.4346506893634796,
      "eval_runtime": 87.5476,
      "eval_samples_per_second": 217.07,
      "eval_steps_per_second": 4.523,
      "step": 1100
    },
    {
      "epoch": 0.11080065516909143,
      "grad_norm": 0.5904502272605896,
      "learning_rate": 0.00019955718277290684,
      "loss": 0.4764,
      "step": 1150
    },
    {
      "epoch": 0.11080065516909143,
      "eval_loss": 0.4015448987483978,
      "eval_runtime": 88.5379,
      "eval_samples_per_second": 214.642,
      "eval_steps_per_second": 4.473,
      "step": 1150
    },
    {
      "epoch": 0.11561807495905194,
      "grad_norm": 0.46411070227622986,
      "learning_rate": 0.000199537913093747,
      "loss": 0.4346,
      "step": 1200
    },
    {
      "epoch": 0.11561807495905194,
      "eval_loss": 0.38101693987846375,
      "eval_runtime": 87.5237,
      "eval_samples_per_second": 217.13,
      "eval_steps_per_second": 4.524,
      "step": 1200
    },
    {
      "epoch": 0.12043549474901243,
      "grad_norm": 0.49205589294433594,
      "learning_rate": 0.00019951864341458714,
      "loss": 0.4151,
      "step": 1250
    },
    {
      "epoch": 0.12043549474901243,
      "eval_loss": 0.3670279085636139,
      "eval_runtime": 87.1762,
      "eval_samples_per_second": 217.995,
      "eval_steps_per_second": 4.543,
      "step": 1250
    },
    {
      "epoch": 0.12525291453897291,
      "grad_norm": 0.5221843123435974,
      "learning_rate": 0.0001994993737354273,
      "loss": 0.3951,
      "step": 1300
    },
    {
      "epoch": 0.12525291453897291,
      "eval_loss": 0.3488354980945587,
      "eval_runtime": 87.8499,
      "eval_samples_per_second": 216.324,
      "eval_steps_per_second": 4.508,
      "step": 1300
    },
    {
      "epoch": 0.13007033432893342,
      "grad_norm": 0.5032587647438049,
      "learning_rate": 0.00019948010405626746,
      "loss": 0.3593,
      "step": 1350
    },
    {
      "epoch": 0.13007033432893342,
      "eval_loss": 0.33173102140426636,
      "eval_runtime": 86.3749,
      "eval_samples_per_second": 220.018,
      "eval_steps_per_second": 4.585,
      "step": 1350
    },
    {
      "epoch": 0.13488775411889392,
      "grad_norm": 0.5051923394203186,
      "learning_rate": 0.00019946083437710763,
      "loss": 0.3526,
      "step": 1400
    },
    {
      "epoch": 0.13488775411889392,
      "eval_loss": 0.30538395047187805,
      "eval_runtime": 89.1978,
      "eval_samples_per_second": 213.055,
      "eval_steps_per_second": 4.44,
      "step": 1400
    },
    {
      "epoch": 0.13970517390885442,
      "grad_norm": 0.46636977791786194,
      "learning_rate": 0.0001994415646979478,
      "loss": 0.3299,
      "step": 1450
    },
    {
      "epoch": 0.13970517390885442,
      "eval_loss": 0.28566229343414307,
      "eval_runtime": 87.6248,
      "eval_samples_per_second": 216.879,
      "eval_steps_per_second": 4.519,
      "step": 1450
    },
    {
      "epoch": 0.14452259369881493,
      "grad_norm": 0.4614977240562439,
      "learning_rate": 0.00019942229501878795,
      "loss": 0.2968,
      "step": 1500
    },
    {
      "epoch": 0.14452259369881493,
      "eval_loss": 0.2682988941669464,
      "eval_runtime": 88.341,
      "eval_samples_per_second": 215.121,
      "eval_steps_per_second": 4.483,
      "step": 1500
    },
    {
      "epoch": 0.1493400134887754,
      "grad_norm": 0.40138185024261475,
      "learning_rate": 0.00019940302533962812,
      "loss": 0.2822,
      "step": 1550
    },
    {
      "epoch": 0.1493400134887754,
      "eval_loss": 0.2586437165737152,
      "eval_runtime": 87.7499,
      "eval_samples_per_second": 216.57,
      "eval_steps_per_second": 4.513,
      "step": 1550
    },
    {
      "epoch": 0.1541574332787359,
      "grad_norm": 0.4057703912258148,
      "learning_rate": 0.00019938375566046825,
      "loss": 0.271,
      "step": 1600
    },
    {
      "epoch": 0.1541574332787359,
      "eval_loss": 0.25495249032974243,
      "eval_runtime": 87.2931,
      "eval_samples_per_second": 217.703,
      "eval_steps_per_second": 4.536,
      "step": 1600
    },
    {
      "epoch": 0.1589748530686964,
      "grad_norm": 0.32808950543403625,
      "learning_rate": 0.00019936448598130842,
      "loss": 0.2547,
      "step": 1650
    },
    {
      "epoch": 0.1589748530686964,
      "eval_loss": 0.243699312210083,
      "eval_runtime": 86.1589,
      "eval_samples_per_second": 220.569,
      "eval_steps_per_second": 4.596,
      "step": 1650
    },
    {
      "epoch": 0.1637922728586569,
      "grad_norm": 0.40255042910575867,
      "learning_rate": 0.00019934521630214858,
      "loss": 0.2549,
      "step": 1700
    },
    {
      "epoch": 0.1637922728586569,
      "eval_loss": 0.24029429256916046,
      "eval_runtime": 88.882,
      "eval_samples_per_second": 213.812,
      "eval_steps_per_second": 4.455,
      "step": 1700
    },
    {
      "epoch": 0.1686096926486174,
      "grad_norm": 0.4509395956993103,
      "learning_rate": 0.00019932594662298874,
      "loss": 0.2541,
      "step": 1750
    },
    {
      "epoch": 0.1686096926486174,
      "eval_loss": 0.23613183200359344,
      "eval_runtime": 87.9005,
      "eval_samples_per_second": 216.199,
      "eval_steps_per_second": 4.505,
      "step": 1750
    },
    {
      "epoch": 0.1734271124385779,
      "grad_norm": 0.3515077531337738,
      "learning_rate": 0.0001993066769438289,
      "loss": 0.2385,
      "step": 1800
    },
    {
      "epoch": 0.1734271124385779,
      "eval_loss": 0.2291395217180252,
      "eval_runtime": 87.3005,
      "eval_samples_per_second": 217.685,
      "eval_steps_per_second": 4.536,
      "step": 1800
    },
    {
      "epoch": 0.1782445322285384,
      "grad_norm": 0.3758944272994995,
      "learning_rate": 0.00019928740726466907,
      "loss": 0.2406,
      "step": 1850
    },
    {
      "epoch": 0.1782445322285384,
      "eval_loss": 0.22721637785434723,
      "eval_runtime": 87.3928,
      "eval_samples_per_second": 217.455,
      "eval_steps_per_second": 4.531,
      "step": 1850
    },
    {
      "epoch": 0.1830619520184989,
      "grad_norm": 0.33480963110923767,
      "learning_rate": 0.0001992681375855092,
      "loss": 0.2327,
      "step": 1900
    },
    {
      "epoch": 0.1830619520184989,
      "eval_loss": 0.22441919147968292,
      "eval_runtime": 87.7197,
      "eval_samples_per_second": 216.645,
      "eval_steps_per_second": 4.514,
      "step": 1900
    },
    {
      "epoch": 0.1878793718084594,
      "grad_norm": 0.44033002853393555,
      "learning_rate": 0.00019924886790634937,
      "loss": 0.2327,
      "step": 1950
    },
    {
      "epoch": 0.1878793718084594,
      "eval_loss": 0.21973223984241486,
      "eval_runtime": 86.9627,
      "eval_samples_per_second": 218.53,
      "eval_steps_per_second": 4.554,
      "step": 1950
    },
    {
      "epoch": 0.19269679159841988,
      "grad_norm": 0.36066603660583496,
      "learning_rate": 0.0001992295982271895,
      "loss": 0.2204,
      "step": 2000
    },
    {
      "epoch": 0.19269679159841988,
      "eval_loss": 0.21755366027355194,
      "eval_runtime": 85.9832,
      "eval_samples_per_second": 221.02,
      "eval_steps_per_second": 4.606,
      "step": 2000
    },
    {
      "epoch": 0.19751421138838038,
      "grad_norm": 0.3927019536495209,
      "learning_rate": 0.00019921032854802967,
      "loss": 0.2227,
      "step": 2050
    },
    {
      "epoch": 0.19751421138838038,
      "eval_loss": 0.21245667338371277,
      "eval_runtime": 88.7726,
      "eval_samples_per_second": 214.075,
      "eval_steps_per_second": 4.461,
      "step": 2050
    },
    {
      "epoch": 0.20233163117834088,
      "grad_norm": 0.33693307638168335,
      "learning_rate": 0.00019919105886886983,
      "loss": 0.2193,
      "step": 2100
    },
    {
      "epoch": 0.20233163117834088,
      "eval_loss": 0.21231026947498322,
      "eval_runtime": 86.8589,
      "eval_samples_per_second": 218.792,
      "eval_steps_per_second": 4.559,
      "step": 2100
    },
    {
      "epoch": 0.20714905096830138,
      "grad_norm": 0.32883742451667786,
      "learning_rate": 0.00019917178918971,
      "loss": 0.2163,
      "step": 2150
    },
    {
      "epoch": 0.20714905096830138,
      "eval_loss": 0.20846392214298248,
      "eval_runtime": 86.2766,
      "eval_samples_per_second": 220.268,
      "eval_steps_per_second": 4.59,
      "step": 2150
    },
    {
      "epoch": 0.2119664707582619,
      "grad_norm": 0.3935041129589081,
      "learning_rate": 0.00019915251951055016,
      "loss": 0.216,
      "step": 2200
    },
    {
      "epoch": 0.2119664707582619,
      "eval_loss": 0.206617072224617,
      "eval_runtime": 87.55,
      "eval_samples_per_second": 217.065,
      "eval_steps_per_second": 4.523,
      "step": 2200
    },
    {
      "epoch": 0.21678389054822236,
      "grad_norm": 0.41742655634880066,
      "learning_rate": 0.00019913324983139032,
      "loss": 0.2046,
      "step": 2250
    },
    {
      "epoch": 0.21678389054822236,
      "eval_loss": 0.2045588344335556,
      "eval_runtime": 86.8857,
      "eval_samples_per_second": 218.724,
      "eval_steps_per_second": 4.558,
      "step": 2250
    },
    {
      "epoch": 0.22160131033818287,
      "grad_norm": 0.3324650526046753,
      "learning_rate": 0.00019911398015223049,
      "loss": 0.206,
      "step": 2300
    },
    {
      "epoch": 0.22160131033818287,
      "eval_loss": 0.2034013867378235,
      "eval_runtime": 88.0292,
      "eval_samples_per_second": 215.883,
      "eval_steps_per_second": 4.499,
      "step": 2300
    },
    {
      "epoch": 0.22641873012814337,
      "grad_norm": 0.3411875367164612,
      "learning_rate": 0.00019909471047307062,
      "loss": 0.2021,
      "step": 2350
    },
    {
      "epoch": 0.22641873012814337,
      "eval_loss": 0.19746768474578857,
      "eval_runtime": 88.0101,
      "eval_samples_per_second": 215.93,
      "eval_steps_per_second": 4.499,
      "step": 2350
    },
    {
      "epoch": 0.23123614991810387,
      "grad_norm": 0.3764231503009796,
      "learning_rate": 0.00019907544079391079,
      "loss": 0.2028,
      "step": 2400
    },
    {
      "epoch": 0.23123614991810387,
      "eval_loss": 0.19666016101837158,
      "eval_runtime": 87.9384,
      "eval_samples_per_second": 216.106,
      "eval_steps_per_second": 4.503,
      "step": 2400
    },
    {
      "epoch": 0.23605356970806435,
      "grad_norm": 0.377541720867157,
      "learning_rate": 0.00019905617111475095,
      "loss": 0.1995,
      "step": 2450
    },
    {
      "epoch": 0.23605356970806435,
      "eval_loss": 0.19164888560771942,
      "eval_runtime": 88.6219,
      "eval_samples_per_second": 214.439,
      "eval_steps_per_second": 4.468,
      "step": 2450
    },
    {
      "epoch": 0.24087098949802485,
      "grad_norm": 0.38512399792671204,
      "learning_rate": 0.0001990369014355911,
      "loss": 0.1947,
      "step": 2500
    },
    {
      "epoch": 0.24087098949802485,
      "eval_loss": 0.1909172534942627,
      "eval_runtime": 86.7278,
      "eval_samples_per_second": 219.122,
      "eval_steps_per_second": 4.566,
      "step": 2500
    },
    {
      "epoch": 0.24568840928798535,
      "grad_norm": 0.29089486598968506,
      "learning_rate": 0.00019901763175643128,
      "loss": 0.1934,
      "step": 2550
    },
    {
      "epoch": 0.24568840928798535,
      "eval_loss": 0.19215822219848633,
      "eval_runtime": 87.4982,
      "eval_samples_per_second": 217.193,
      "eval_steps_per_second": 4.526,
      "step": 2550
    },
    {
      "epoch": 0.25050582907794583,
      "grad_norm": 0.34340256452560425,
      "learning_rate": 0.00019899836207727144,
      "loss": 0.1893,
      "step": 2600
    },
    {
      "epoch": 0.25050582907794583,
      "eval_loss": 0.18751804530620575,
      "eval_runtime": 89.0572,
      "eval_samples_per_second": 213.391,
      "eval_steps_per_second": 4.447,
      "step": 2600
    },
    {
      "epoch": 0.25532324886790636,
      "grad_norm": 0.31367331743240356,
      "learning_rate": 0.00019897909239811158,
      "loss": 0.1831,
      "step": 2650
    },
    {
      "epoch": 0.25532324886790636,
      "eval_loss": 0.1869846135377884,
      "eval_runtime": 88.3755,
      "eval_samples_per_second": 215.037,
      "eval_steps_per_second": 4.481,
      "step": 2650
    },
    {
      "epoch": 0.26014066865786684,
      "grad_norm": 0.35607218742370605,
      "learning_rate": 0.00019895982271895174,
      "loss": 0.1906,
      "step": 2700
    },
    {
      "epoch": 0.26014066865786684,
      "eval_loss": 0.1840076744556427,
      "eval_runtime": 88.0613,
      "eval_samples_per_second": 215.804,
      "eval_steps_per_second": 4.497,
      "step": 2700
    },
    {
      "epoch": 0.26495808844782737,
      "grad_norm": 0.29454657435417175,
      "learning_rate": 0.00019894055303979188,
      "loss": 0.1878,
      "step": 2750
    },
    {
      "epoch": 0.26495808844782737,
      "eval_loss": 0.1864839792251587,
      "eval_runtime": 87.4663,
      "eval_samples_per_second": 217.272,
      "eval_steps_per_second": 4.527,
      "step": 2750
    },
    {
      "epoch": 0.26977550823778784,
      "grad_norm": 0.33838963508605957,
      "learning_rate": 0.00019892128336063204,
      "loss": 0.1821,
      "step": 2800
    },
    {
      "epoch": 0.26977550823778784,
      "eval_loss": 0.1847812533378601,
      "eval_runtime": 88.0891,
      "eval_samples_per_second": 215.736,
      "eval_steps_per_second": 4.495,
      "step": 2800
    },
    {
      "epoch": 0.2745929280277483,
      "grad_norm": 0.2440441995859146,
      "learning_rate": 0.0001989020136814722,
      "loss": 0.1783,
      "step": 2850
    },
    {
      "epoch": 0.2745929280277483,
      "eval_loss": 0.18052709102630615,
      "eval_runtime": 85.9682,
      "eval_samples_per_second": 221.059,
      "eval_steps_per_second": 4.606,
      "step": 2850
    },
    {
      "epoch": 0.27941034781770885,
      "grad_norm": 0.33384960889816284,
      "learning_rate": 0.00019888274400231237,
      "loss": 0.1796,
      "step": 2900
    },
    {
      "epoch": 0.27941034781770885,
      "eval_loss": 0.17909078299999237,
      "eval_runtime": 87.5176,
      "eval_samples_per_second": 217.145,
      "eval_steps_per_second": 4.525,
      "step": 2900
    },
    {
      "epoch": 0.2842277676076693,
      "grad_norm": 0.29702430963516235,
      "learning_rate": 0.00019886347432315253,
      "loss": 0.1785,
      "step": 2950
    },
    {
      "epoch": 0.2842277676076693,
      "eval_loss": 0.17997683584690094,
      "eval_runtime": 87.9757,
      "eval_samples_per_second": 216.014,
      "eval_steps_per_second": 4.501,
      "step": 2950
    },
    {
      "epoch": 0.28904518739762985,
      "grad_norm": 0.3510948121547699,
      "learning_rate": 0.0001988442046439927,
      "loss": 0.1765,
      "step": 3000
    },
    {
      "epoch": 0.28904518739762985,
      "eval_loss": 0.1761734038591385,
      "eval_runtime": 88.1258,
      "eval_samples_per_second": 215.646,
      "eval_steps_per_second": 4.494,
      "step": 3000
    },
    {
      "epoch": 0.29386260718759033,
      "grad_norm": 0.37029772996902466,
      "learning_rate": 0.00019882493496483286,
      "loss": 0.1737,
      "step": 3050
    },
    {
      "epoch": 0.29386260718759033,
      "eval_loss": 0.17449568212032318,
      "eval_runtime": 88.3713,
      "eval_samples_per_second": 215.047,
      "eval_steps_per_second": 4.481,
      "step": 3050
    },
    {
      "epoch": 0.2986800269775508,
      "grad_norm": 0.2771267592906952,
      "learning_rate": 0.000198805665285673,
      "loss": 0.1761,
      "step": 3100
    },
    {
      "epoch": 0.2986800269775508,
      "eval_loss": 0.1765695959329605,
      "eval_runtime": 87.9476,
      "eval_samples_per_second": 216.083,
      "eval_steps_per_second": 4.503,
      "step": 3100
    },
    {
      "epoch": 0.30349744676751134,
      "grad_norm": 0.338558167219162,
      "learning_rate": 0.00019878639560651316,
      "loss": 0.1766,
      "step": 3150
    },
    {
      "epoch": 0.30349744676751134,
      "eval_loss": 0.17348217964172363,
      "eval_runtime": 86.4764,
      "eval_samples_per_second": 219.759,
      "eval_steps_per_second": 4.579,
      "step": 3150
    },
    {
      "epoch": 0.3083148665574718,
      "grad_norm": 0.36884650588035583,
      "learning_rate": 0.00019876712592735332,
      "loss": 0.1697,
      "step": 3200
    },
    {
      "epoch": 0.3083148665574718,
      "eval_loss": 0.1715058535337448,
      "eval_runtime": 88.0269,
      "eval_samples_per_second": 215.888,
      "eval_steps_per_second": 4.499,
      "step": 3200
    },
    {
      "epoch": 0.31313228634743234,
      "grad_norm": 0.2907465994358063,
      "learning_rate": 0.00019874785624819348,
      "loss": 0.1747,
      "step": 3250
    },
    {
      "epoch": 0.31313228634743234,
      "eval_loss": 0.17374737560749054,
      "eval_runtime": 87.3911,
      "eval_samples_per_second": 217.459,
      "eval_steps_per_second": 4.531,
      "step": 3250
    },
    {
      "epoch": 0.3179497061373928,
      "grad_norm": 0.3033406734466553,
      "learning_rate": 0.00019872858656903365,
      "loss": 0.1706,
      "step": 3300
    },
    {
      "epoch": 0.3179497061373928,
      "eval_loss": 0.17150533199310303,
      "eval_runtime": 88.1112,
      "eval_samples_per_second": 215.682,
      "eval_steps_per_second": 4.494,
      "step": 3300
    },
    {
      "epoch": 0.3227671259273533,
      "grad_norm": 0.3115890324115753,
      "learning_rate": 0.0001987093168898738,
      "loss": 0.1665,
      "step": 3350
    },
    {
      "epoch": 0.3227671259273533,
      "eval_loss": 0.16980254650115967,
      "eval_runtime": 87.4444,
      "eval_samples_per_second": 217.327,
      "eval_steps_per_second": 4.529,
      "step": 3350
    },
    {
      "epoch": 0.3275845457173138,
      "grad_norm": 0.25616976618766785,
      "learning_rate": 0.00019869004721071395,
      "loss": 0.1655,
      "step": 3400
    },
    {
      "epoch": 0.3275845457173138,
      "eval_loss": 0.17015036940574646,
      "eval_runtime": 88.3575,
      "eval_samples_per_second": 215.081,
      "eval_steps_per_second": 4.482,
      "step": 3400
    },
    {
      "epoch": 0.3324019655072743,
      "grad_norm": 0.23346827924251556,
      "learning_rate": 0.0001986707775315541,
      "loss": 0.1673,
      "step": 3450
    },
    {
      "epoch": 0.3324019655072743,
      "eval_loss": 0.1687326282262802,
      "eval_runtime": 90.3569,
      "eval_samples_per_second": 210.321,
      "eval_steps_per_second": 4.383,
      "step": 3450
    },
    {
      "epoch": 0.3372193852972348,
      "grad_norm": 0.3347044289112091,
      "learning_rate": 0.00019865150785239425,
      "loss": 0.163,
      "step": 3500
    },
    {
      "epoch": 0.3372193852972348,
      "eval_loss": 0.17024628818035126,
      "eval_runtime": 88.9183,
      "eval_samples_per_second": 213.724,
      "eval_steps_per_second": 4.454,
      "step": 3500
    },
    {
      "epoch": 0.3420368050871953,
      "grad_norm": 0.34929919242858887,
      "learning_rate": 0.0001986322381732344,
      "loss": 0.1624,
      "step": 3550
    },
    {
      "epoch": 0.3420368050871953,
      "eval_loss": 0.16716831922531128,
      "eval_runtime": 89.8228,
      "eval_samples_per_second": 211.572,
      "eval_steps_per_second": 4.409,
      "step": 3550
    },
    {
      "epoch": 0.3468542248771558,
      "grad_norm": 0.2455097734928131,
      "learning_rate": 0.00019861296849407457,
      "loss": 0.1633,
      "step": 3600
    },
    {
      "epoch": 0.3468542248771558,
      "eval_loss": 0.16651391983032227,
      "eval_runtime": 86.8189,
      "eval_samples_per_second": 218.892,
      "eval_steps_per_second": 4.561,
      "step": 3600
    },
    {
      "epoch": 0.3516716446671163,
      "grad_norm": 0.28978198766708374,
      "learning_rate": 0.00019859369881491474,
      "loss": 0.1646,
      "step": 3650
    },
    {
      "epoch": 0.3516716446671163,
      "eval_loss": 0.16727794706821442,
      "eval_runtime": 87.7801,
      "eval_samples_per_second": 216.496,
      "eval_steps_per_second": 4.511,
      "step": 3650
    },
    {
      "epoch": 0.3564890644570768,
      "grad_norm": 0.292510062456131,
      "learning_rate": 0.0001985744291357549,
      "loss": 0.1625,
      "step": 3700
    },
    {
      "epoch": 0.3564890644570768,
      "eval_loss": 0.16462790966033936,
      "eval_runtime": 87.2466,
      "eval_samples_per_second": 217.819,
      "eval_steps_per_second": 4.539,
      "step": 3700
    },
    {
      "epoch": 0.36130648424703726,
      "grad_norm": 0.26186755299568176,
      "learning_rate": 0.00019855515945659506,
      "loss": 0.1583,
      "step": 3750
    },
    {
      "epoch": 0.36130648424703726,
      "eval_loss": 0.16536560654640198,
      "eval_runtime": 85.6294,
      "eval_samples_per_second": 221.933,
      "eval_steps_per_second": 4.625,
      "step": 3750
    },
    {
      "epoch": 0.3661239040369978,
      "grad_norm": 0.25864389538764954,
      "learning_rate": 0.00019853588977743523,
      "loss": 0.163,
      "step": 3800
    },
    {
      "epoch": 0.3661239040369978,
      "eval_loss": 0.16156961023807526,
      "eval_runtime": 88.8605,
      "eval_samples_per_second": 213.863,
      "eval_steps_per_second": 4.456,
      "step": 3800
    },
    {
      "epoch": 0.37094132382695827,
      "grad_norm": 0.298779159784317,
      "learning_rate": 0.00019851662009827536,
      "loss": 0.1582,
      "step": 3850
    },
    {
      "epoch": 0.37094132382695827,
      "eval_loss": 0.16503678262233734,
      "eval_runtime": 90.0855,
      "eval_samples_per_second": 210.955,
      "eval_steps_per_second": 4.396,
      "step": 3850
    },
    {
      "epoch": 0.3757587436169188,
      "grad_norm": 0.30580171942710876,
      "learning_rate": 0.00019849735041911553,
      "loss": 0.1568,
      "step": 3900
    },
    {
      "epoch": 0.3757587436169188,
      "eval_loss": 0.16093286871910095,
      "eval_runtime": 87.9051,
      "eval_samples_per_second": 216.188,
      "eval_steps_per_second": 4.505,
      "step": 3900
    },
    {
      "epoch": 0.3805761634068793,
      "grad_norm": 0.26903071999549866,
      "learning_rate": 0.0001984780807399557,
      "loss": 0.156,
      "step": 3950
    },
    {
      "epoch": 0.3805761634068793,
      "eval_loss": 0.16180996596813202,
      "eval_runtime": 87.2728,
      "eval_samples_per_second": 217.754,
      "eval_steps_per_second": 4.537,
      "step": 3950
    },
    {
      "epoch": 0.38539358319683975,
      "grad_norm": 0.25209519267082214,
      "learning_rate": 0.00019845881106079585,
      "loss": 0.1559,
      "step": 4000
    },
    {
      "epoch": 0.38539358319683975,
      "eval_loss": 0.15961618721485138,
      "eval_runtime": 89.2611,
      "eval_samples_per_second": 212.903,
      "eval_steps_per_second": 4.436,
      "step": 4000
    },
    {
      "epoch": 0.3902110029868003,
      "grad_norm": 0.29754438996315,
      "learning_rate": 0.00019843954138163602,
      "loss": 0.1546,
      "step": 4050
    },
    {
      "epoch": 0.3902110029868003,
      "eval_loss": 0.16072513163089752,
      "eval_runtime": 87.9158,
      "eval_samples_per_second": 216.161,
      "eval_steps_per_second": 4.504,
      "step": 4050
    },
    {
      "epoch": 0.39502842277676076,
      "grad_norm": 0.259056031703949,
      "learning_rate": 0.00019842027170247618,
      "loss": 0.1576,
      "step": 4100
    },
    {
      "epoch": 0.39502842277676076,
      "eval_loss": 0.16210326552391052,
      "eval_runtime": 87.5661,
      "eval_samples_per_second": 217.025,
      "eval_steps_per_second": 4.522,
      "step": 4100
    },
    {
      "epoch": 0.3998458425667213,
      "grad_norm": 0.2900582253932953,
      "learning_rate": 0.00019840100202331634,
      "loss": 0.1557,
      "step": 4150
    },
    {
      "epoch": 0.3998458425667213,
      "eval_loss": 0.16127543151378632,
      "eval_runtime": 87.2874,
      "eval_samples_per_second": 217.718,
      "eval_steps_per_second": 4.537,
      "step": 4150
    },
    {
      "epoch": 0.40466326235668176,
      "grad_norm": 0.3155025243759155,
      "learning_rate": 0.00019838173234415648,
      "loss": 0.151,
      "step": 4200
    },
    {
      "epoch": 0.40466326235668176,
      "eval_loss": 0.15659989416599274,
      "eval_runtime": 87.5007,
      "eval_samples_per_second": 217.187,
      "eval_steps_per_second": 4.526,
      "step": 4200
    },
    {
      "epoch": 0.40948068214664224,
      "grad_norm": 0.20887872576713562,
      "learning_rate": 0.00019836246266499664,
      "loss": 0.149,
      "step": 4250
    },
    {
      "epoch": 0.40948068214664224,
      "eval_loss": 0.1579655408859253,
      "eval_runtime": 88.1039,
      "eval_samples_per_second": 215.7,
      "eval_steps_per_second": 4.495,
      "step": 4250
    },
    {
      "epoch": 0.41429810193660277,
      "grad_norm": 0.2782476544380188,
      "learning_rate": 0.00019834319298583678,
      "loss": 0.1519,
      "step": 4300
    },
    {
      "epoch": 0.41429810193660277,
      "eval_loss": 0.15905822813510895,
      "eval_runtime": 89.4392,
      "eval_samples_per_second": 212.479,
      "eval_steps_per_second": 4.428,
      "step": 4300
    },
    {
      "epoch": 0.41911552172656324,
      "grad_norm": 0.2640698254108429,
      "learning_rate": 0.00019832392330667694,
      "loss": 0.1499,
      "step": 4350
    },
    {
      "epoch": 0.41911552172656324,
      "eval_loss": 0.15701065957546234,
      "eval_runtime": 88.5724,
      "eval_samples_per_second": 214.559,
      "eval_steps_per_second": 4.471,
      "step": 4350
    },
    {
      "epoch": 0.4239329415165238,
      "grad_norm": 0.2606890797615051,
      "learning_rate": 0.0001983046536275171,
      "loss": 0.1509,
      "step": 4400
    },
    {
      "epoch": 0.4239329415165238,
      "eval_loss": 0.1540563404560089,
      "eval_runtime": 87.9838,
      "eval_samples_per_second": 215.994,
      "eval_steps_per_second": 4.501,
      "step": 4400
    },
    {
      "epoch": 0.42875036130648425,
      "grad_norm": 0.32414618134498596,
      "learning_rate": 0.00019828538394835727,
      "loss": 0.1476,
      "step": 4450
    },
    {
      "epoch": 0.42875036130648425,
      "eval_loss": 0.15688583254814148,
      "eval_runtime": 85.8129,
      "eval_samples_per_second": 221.459,
      "eval_steps_per_second": 4.615,
      "step": 4450
    },
    {
      "epoch": 0.4335677810964447,
      "grad_norm": 0.31708353757858276,
      "learning_rate": 0.00019826611426919743,
      "loss": 0.1477,
      "step": 4500
    },
    {
      "epoch": 0.4335677810964447,
      "eval_loss": 0.15801727771759033,
      "eval_runtime": 89.5976,
      "eval_samples_per_second": 212.104,
      "eval_steps_per_second": 4.42,
      "step": 4500
    },
    {
      "epoch": 0.43838520088640526,
      "grad_norm": 0.23212596774101257,
      "learning_rate": 0.0001982468445900376,
      "loss": 0.146,
      "step": 4550
    },
    {
      "epoch": 0.43838520088640526,
      "eval_loss": 0.15200358629226685,
      "eval_runtime": 87.3248,
      "eval_samples_per_second": 217.624,
      "eval_steps_per_second": 4.535,
      "step": 4550
    },
    {
      "epoch": 0.44320262067636573,
      "grad_norm": 0.2829053997993469,
      "learning_rate": 0.00019822757491087773,
      "loss": 0.1465,
      "step": 4600
    },
    {
      "epoch": 0.44320262067636573,
      "eval_loss": 0.1542133092880249,
      "eval_runtime": 85.3408,
      "eval_samples_per_second": 222.684,
      "eval_steps_per_second": 4.64,
      "step": 4600
    },
    {
      "epoch": 0.4480200404663262,
      "grad_norm": 0.26959696412086487,
      "learning_rate": 0.0001982083052317179,
      "loss": 0.1494,
      "step": 4650
    },
    {
      "epoch": 0.4480200404663262,
      "eval_loss": 0.15412269532680511,
      "eval_runtime": 87.9426,
      "eval_samples_per_second": 216.095,
      "eval_steps_per_second": 4.503,
      "step": 4650
    },
    {
      "epoch": 0.45283746025628674,
      "grad_norm": 0.2577558755874634,
      "learning_rate": 0.00019818903555255806,
      "loss": 0.1465,
      "step": 4700
    },
    {
      "epoch": 0.45283746025628674,
      "eval_loss": 0.15427736937999725,
      "eval_runtime": 88.297,
      "eval_samples_per_second": 215.228,
      "eval_steps_per_second": 4.485,
      "step": 4700
    },
    {
      "epoch": 0.4576548800462472,
      "grad_norm": 0.23728205263614655,
      "learning_rate": 0.00019816976587339822,
      "loss": 0.1425,
      "step": 4750
    },
    {
      "epoch": 0.4576548800462472,
      "eval_loss": 0.15123674273490906,
      "eval_runtime": 87.7725,
      "eval_samples_per_second": 216.514,
      "eval_steps_per_second": 4.512,
      "step": 4750
    },
    {
      "epoch": 0.46247229983620775,
      "grad_norm": 0.2485925406217575,
      "learning_rate": 0.00019815049619423839,
      "loss": 0.1422,
      "step": 4800
    },
    {
      "epoch": 0.46247229983620775,
      "eval_loss": 0.15153329074382782,
      "eval_runtime": 87.8581,
      "eval_samples_per_second": 216.303,
      "eval_steps_per_second": 4.507,
      "step": 4800
    },
    {
      "epoch": 0.4672897196261682,
      "grad_norm": 0.22593127191066742,
      "learning_rate": 0.00019813122651507855,
      "loss": 0.1452,
      "step": 4850
    },
    {
      "epoch": 0.4672897196261682,
      "eval_loss": 0.15019385516643524,
      "eval_runtime": 88.2356,
      "eval_samples_per_second": 215.378,
      "eval_steps_per_second": 4.488,
      "step": 4850
    },
    {
      "epoch": 0.4721071394161287,
      "grad_norm": 0.2340419441461563,
      "learning_rate": 0.0001981119568359187,
      "loss": 0.1409,
      "step": 4900
    },
    {
      "epoch": 0.4721071394161287,
      "eval_loss": 0.14823836088180542,
      "eval_runtime": 87.0907,
      "eval_samples_per_second": 218.209,
      "eval_steps_per_second": 4.547,
      "step": 4900
    },
    {
      "epoch": 0.4769245592060892,
      "grad_norm": 0.24335741996765137,
      "learning_rate": 0.00019809268715675885,
      "loss": 0.1442,
      "step": 4950
    },
    {
      "epoch": 0.4769245592060892,
      "eval_loss": 0.14839066565036774,
      "eval_runtime": 88.1982,
      "eval_samples_per_second": 215.469,
      "eval_steps_per_second": 4.49,
      "step": 4950
    },
    {
      "epoch": 0.4817419789960497,
      "grad_norm": 0.2178025096654892,
      "learning_rate": 0.000198073417477599,
      "loss": 0.1397,
      "step": 5000
    },
    {
      "epoch": 0.4817419789960497,
      "eval_loss": 0.14846083521842957,
      "eval_runtime": 88.7021,
      "eval_samples_per_second": 214.245,
      "eval_steps_per_second": 4.464,
      "step": 5000
    },
    {
      "epoch": 0.48655939878601023,
      "grad_norm": 0.2502117156982422,
      "learning_rate": 0.00019805414779843915,
      "loss": 0.1412,
      "step": 5050
    },
    {
      "epoch": 0.48655939878601023,
      "eval_loss": 0.14794325828552246,
      "eval_runtime": 87.2101,
      "eval_samples_per_second": 217.911,
      "eval_steps_per_second": 4.541,
      "step": 5050
    },
    {
      "epoch": 0.4913768185759707,
      "grad_norm": 0.26042014360427856,
      "learning_rate": 0.0001980348781192793,
      "loss": 0.1407,
      "step": 5100
    },
    {
      "epoch": 0.4913768185759707,
      "eval_loss": 0.14844238758087158,
      "eval_runtime": 87.5921,
      "eval_samples_per_second": 216.96,
      "eval_steps_per_second": 4.521,
      "step": 5100
    },
    {
      "epoch": 0.4961942383659312,
      "grad_norm": 0.27736595273017883,
      "learning_rate": 0.00019801560844011948,
      "loss": 0.1392,
      "step": 5150
    },
    {
      "epoch": 0.4961942383659312,
      "eval_loss": 0.1466793715953827,
      "eval_runtime": 87.8863,
      "eval_samples_per_second": 216.234,
      "eval_steps_per_second": 4.506,
      "step": 5150
    },
    {
      "epoch": 0.5010116581558917,
      "grad_norm": 0.26440733671188354,
      "learning_rate": 0.00019799633876095964,
      "loss": 0.1362,
      "step": 5200
    },
    {
      "epoch": 0.5010116581558917,
      "eval_loss": 0.1452549546957016,
      "eval_runtime": 87.8213,
      "eval_samples_per_second": 216.394,
      "eval_steps_per_second": 4.509,
      "step": 5200
    },
    {
      "epoch": 0.5058290779458522,
      "grad_norm": 0.2524762749671936,
      "learning_rate": 0.0001979770690817998,
      "loss": 0.1411,
      "step": 5250
    },
    {
      "epoch": 0.5058290779458522,
      "eval_loss": 0.14888718724250793,
      "eval_runtime": 87.8648,
      "eval_samples_per_second": 216.287,
      "eval_steps_per_second": 4.507,
      "step": 5250
    },
    {
      "epoch": 0.5106464977358127,
      "grad_norm": 0.2830604612827301,
      "learning_rate": 0.00019795779940263997,
      "loss": 0.1373,
      "step": 5300
    },
    {
      "epoch": 0.5106464977358127,
      "eval_loss": 0.14802096784114838,
      "eval_runtime": 86.7846,
      "eval_samples_per_second": 218.979,
      "eval_steps_per_second": 4.563,
      "step": 5300
    },
    {
      "epoch": 0.5154639175257731,
      "grad_norm": 0.23933689296245575,
      "learning_rate": 0.0001979385297234801,
      "loss": 0.1356,
      "step": 5350
    },
    {
      "epoch": 0.5154639175257731,
      "eval_loss": 0.14858801662921906,
      "eval_runtime": 86.8659,
      "eval_samples_per_second": 218.774,
      "eval_steps_per_second": 4.559,
      "step": 5350
    },
    {
      "epoch": 0.5202813373157337,
      "grad_norm": 0.23482471704483032,
      "learning_rate": 0.00019791926004432027,
      "loss": 0.1375,
      "step": 5400
    },
    {
      "epoch": 0.5202813373157337,
      "eval_loss": 0.14439742267131805,
      "eval_runtime": 86.5083,
      "eval_samples_per_second": 219.678,
      "eval_steps_per_second": 4.578,
      "step": 5400
    },
    {
      "epoch": 0.5250987571056942,
      "grad_norm": 0.2511419653892517,
      "learning_rate": 0.00019789999036516043,
      "loss": 0.1338,
      "step": 5450
    },
    {
      "epoch": 0.5250987571056942,
      "eval_loss": 0.14546434581279755,
      "eval_runtime": 88.2612,
      "eval_samples_per_second": 215.316,
      "eval_steps_per_second": 4.487,
      "step": 5450
    },
    {
      "epoch": 0.5299161768956547,
      "grad_norm": 0.2835318148136139,
      "learning_rate": 0.0001978807206860006,
      "loss": 0.1373,
      "step": 5500
    },
    {
      "epoch": 0.5299161768956547,
      "eval_loss": 0.14437447488307953,
      "eval_runtime": 87.9924,
      "eval_samples_per_second": 215.973,
      "eval_steps_per_second": 4.5,
      "step": 5500
    },
    {
      "epoch": 0.5347335966856152,
      "grad_norm": 0.2739470899105072,
      "learning_rate": 0.00019786145100684076,
      "loss": 0.1377,
      "step": 5550
    },
    {
      "epoch": 0.5347335966856152,
      "eval_loss": 0.14675094187259674,
      "eval_runtime": 88.4052,
      "eval_samples_per_second": 214.965,
      "eval_steps_per_second": 4.479,
      "step": 5550
    },
    {
      "epoch": 0.5395510164755757,
      "grad_norm": 0.23613734543323517,
      "learning_rate": 0.00019784218132768092,
      "loss": 0.1375,
      "step": 5600
    },
    {
      "epoch": 0.5395510164755757,
      "eval_loss": 0.14566753804683685,
      "eval_runtime": 87.4914,
      "eval_samples_per_second": 217.21,
      "eval_steps_per_second": 4.526,
      "step": 5600
    },
    {
      "epoch": 0.5443684362655362,
      "grad_norm": 0.2323722243309021,
      "learning_rate": 0.00019782291164852108,
      "loss": 0.1321,
      "step": 5650
    },
    {
      "epoch": 0.5443684362655362,
      "eval_loss": 0.1452152580022812,
      "eval_runtime": 89.1135,
      "eval_samples_per_second": 213.256,
      "eval_steps_per_second": 4.444,
      "step": 5650
    },
    {
      "epoch": 0.5491858560554966,
      "grad_norm": 0.29475587606430054,
      "learning_rate": 0.00019780364196936122,
      "loss": 0.1335,
      "step": 5700
    },
    {
      "epoch": 0.5491858560554966,
      "eval_loss": 0.142301544547081,
      "eval_runtime": 85.457,
      "eval_samples_per_second": 222.381,
      "eval_steps_per_second": 4.634,
      "step": 5700
    },
    {
      "epoch": 0.5540032758454572,
      "grad_norm": 0.29040685296058655,
      "learning_rate": 0.00019778437229020138,
      "loss": 0.1339,
      "step": 5750
    },
    {
      "epoch": 0.5540032758454572,
      "eval_loss": 0.14422039687633514,
      "eval_runtime": 88.9367,
      "eval_samples_per_second": 213.68,
      "eval_steps_per_second": 4.453,
      "step": 5750
    },
    {
      "epoch": 0.5588206956354177,
      "grad_norm": 0.28566643595695496,
      "learning_rate": 0.00019776510261104152,
      "loss": 0.1334,
      "step": 5800
    },
    {
      "epoch": 0.5588206956354177,
      "eval_loss": 0.14271041750907898,
      "eval_runtime": 87.7795,
      "eval_samples_per_second": 216.497,
      "eval_steps_per_second": 4.511,
      "step": 5800
    },
    {
      "epoch": 0.5636381154253781,
      "grad_norm": 0.2813100218772888,
      "learning_rate": 0.00019774583293188168,
      "loss": 0.1347,
      "step": 5850
    },
    {
      "epoch": 0.5636381154253781,
      "eval_loss": 0.14340683817863464,
      "eval_runtime": 87.6902,
      "eval_samples_per_second": 216.717,
      "eval_steps_per_second": 4.516,
      "step": 5850
    },
    {
      "epoch": 0.5684555352153386,
      "grad_norm": 0.2560277581214905,
      "learning_rate": 0.00019772656325272185,
      "loss": 0.1335,
      "step": 5900
    },
    {
      "epoch": 0.5684555352153386,
      "eval_loss": 0.1437900960445404,
      "eval_runtime": 88.7352,
      "eval_samples_per_second": 214.165,
      "eval_steps_per_second": 4.463,
      "step": 5900
    },
    {
      "epoch": 0.5732729550052992,
      "grad_norm": 0.2642715275287628,
      "learning_rate": 0.000197707293573562,
      "loss": 0.1313,
      "step": 5950
    },
    {
      "epoch": 0.5732729550052992,
      "eval_loss": 0.14153434336185455,
      "eval_runtime": 89.9067,
      "eval_samples_per_second": 211.375,
      "eval_steps_per_second": 4.405,
      "step": 5950
    },
    {
      "epoch": 0.5780903747952597,
      "grad_norm": 0.2073492407798767,
      "learning_rate": 0.00019768802389440217,
      "loss": 0.1323,
      "step": 6000
    },
    {
      "epoch": 0.5780903747952597,
      "eval_loss": 0.13940538465976715,
      "eval_runtime": 89.7245,
      "eval_samples_per_second": 211.804,
      "eval_steps_per_second": 4.414,
      "step": 6000
    }
  ],
  "logging_steps": 50,
  "max_steps": 518950,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 1000,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 10,
        "early_stopping_threshold": 0.001
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.21054265081856e+17,
  "train_batch_size": 192,
  "trial_name": null,
  "trial_params": null
}