{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6666666666666666,
  "eval_steps": 100,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 20.25166668892518,
      "learning_rate": 8.333333333333334e-06,
      "loss": 11.0527,
      "step": 25
    },
    {
      "epoch": 0.02,
      "grad_norm": 9.032903124299999,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 8.6558,
      "step": 50
    },
    {
      "epoch": 0.03,
      "grad_norm": 5.7572068330082,
      "learning_rate": 2.5e-05,
      "loss": 8.501,
      "step": 75
    },
    {
      "epoch": 0.03,
      "grad_norm": 3.7554051243303186,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 8.4932,
      "step": 100
    },
    {
      "epoch": 0.03,
      "eval_loss": 8.403280258178711,
      "eval_runtime": 2.444,
      "eval_samples_per_second": 2.864,
      "eval_steps_per_second": 0.409,
      "step": 100
    },
    {
      "epoch": 0.04,
      "grad_norm": 5.05967019179598,
      "learning_rate": 4.166666666666667e-05,
      "loss": 8.3719,
      "step": 125
    },
    {
      "epoch": 0.05,
      "grad_norm": 4.847800073002136,
      "learning_rate": 5e-05,
      "loss": 8.1324,
      "step": 150
    },
    {
      "epoch": 0.06,
      "grad_norm": 4.83727692533756,
      "learning_rate": 4.999050767562379e-05,
      "loss": 7.725,
      "step": 175
    },
    {
      "epoch": 0.07,
      "grad_norm": 7.288277420776578,
      "learning_rate": 4.996203791083291e-05,
      "loss": 6.8915,
      "step": 200
    },
    {
      "epoch": 0.07,
      "eval_loss": 7.34761381149292,
      "eval_runtime": 1.9955,
      "eval_samples_per_second": 3.508,
      "eval_steps_per_second": 0.501,
      "step": 200
    },
    {
      "epoch": 0.07,
      "grad_norm": 7.159902843784087,
      "learning_rate": 4.991461232516675e-05,
      "loss": 5.7354,
      "step": 225
    },
    {
      "epoch": 0.08,
      "grad_norm": 11.491480752541028,
      "learning_rate": 4.984826693294874e-05,
      "loss": 4.5382,
      "step": 250
    },
    {
      "epoch": 0.09,
      "grad_norm": 8.64216534708584,
      "learning_rate": 4.976305211593758e-05,
      "loss": 3.4957,
      "step": 275
    },
    {
      "epoch": 0.1,
      "grad_norm": 9.463504654920019,
      "learning_rate": 4.965903258506806e-05,
      "loss": 2.57,
      "step": 300
    },
    {
      "epoch": 0.1,
      "eval_loss": 6.328631401062012,
      "eval_runtime": 2.0084,
      "eval_samples_per_second": 3.485,
      "eval_steps_per_second": 0.498,
      "step": 300
    },
    {
      "epoch": 0.11,
      "grad_norm": 21.508882454052266,
      "learning_rate": 4.953628733131045e-05,
      "loss": 1.8991,
      "step": 325
    },
    {
      "epoch": 0.12,
      "grad_norm": 5.804271629648969,
      "learning_rate": 4.9394909565685894e-05,
      "loss": 1.3398,
      "step": 350
    },
    {
      "epoch": 0.12,
      "grad_norm": 9.385609144407736,
      "learning_rate": 4.923500664848326e-05,
      "loss": 1.0789,
      "step": 375
    },
    {
      "epoch": 0.13,
      "grad_norm": 5.7800077293904915,
      "learning_rate": 4.905670000773126e-05,
      "loss": 0.8359,
      "step": 400
    },
    {
      "epoch": 0.13,
      "eval_loss": 6.389462471008301,
      "eval_runtime": 1.9225,
      "eval_samples_per_second": 3.641,
      "eval_steps_per_second": 0.52,
      "step": 400
    },
    {
      "epoch": 0.14,
      "grad_norm": 6.188182955826137,
      "learning_rate": 4.886012504698769e-05,
      "loss": 0.6113,
      "step": 425
    },
    {
      "epoch": 0.15,
      "grad_norm": 5.433461636544655,
      "learning_rate": 4.864543104251587e-05,
      "loss": 0.5115,
      "step": 450
    },
    {
      "epoch": 0.16,
      "grad_norm": 5.263910737398804,
      "learning_rate": 4.841278102992637e-05,
      "loss": 0.4475,
      "step": 475
    },
    {
      "epoch": 0.17,
      "grad_norm": 4.489583987078794,
      "learning_rate": 4.8162351680370044e-05,
      "loss": 0.3845,
      "step": 500
    },
    {
      "epoch": 0.17,
      "eval_loss": 6.672789573669434,
      "eval_runtime": 2.0659,
      "eval_samples_per_second": 3.388,
      "eval_steps_per_second": 0.484,
      "step": 500
    },
    {
      "epoch": 0.17,
      "grad_norm": 5.196381164846382,
      "learning_rate": 4.789433316637644e-05,
      "loss": 0.3733,
      "step": 525
    },
    {
      "epoch": 0.18,
      "grad_norm": 3.8044414274380562,
      "learning_rate": 4.760892901743944e-05,
      "loss": 0.3566,
      "step": 550
    },
    {
      "epoch": 0.19,
      "grad_norm": 3.4787421022056098,
      "learning_rate": 4.730635596545985e-05,
      "loss": 0.3141,
      "step": 575
    },
    {
      "epoch": 0.2,
      "grad_norm": 9.827489853335932,
      "learning_rate": 4.698684378016222e-05,
      "loss": 0.3107,
      "step": 600
    },
    {
      "epoch": 0.2,
      "eval_loss": 6.502092361450195,
      "eval_runtime": 2.426,
      "eval_samples_per_second": 2.885,
      "eval_steps_per_second": 0.412,
      "step": 600
    },
    {
      "epoch": 0.21,
      "grad_norm": 3.569090915517714,
      "learning_rate": 4.665063509461097e-05,
      "loss": 0.3029,
      "step": 625
    },
    {
      "epoch": 0.22,
      "grad_norm": 8.808604180091518,
      "learning_rate": 4.629798522095818e-05,
      "loss": 0.2851,
      "step": 650
    },
    {
      "epoch": 0.23,
      "grad_norm": 6.532098366400892,
      "learning_rate": 4.592916195656322e-05,
      "loss": 0.2543,
      "step": 675
    },
    {
      "epoch": 0.23,
      "grad_norm": 3.86761849222122,
      "learning_rate": 4.554444538063113e-05,
      "loss": 0.2296,
      "step": 700
    },
    {
      "epoch": 0.23,
      "eval_loss": 6.935348987579346,
      "eval_runtime": 2.2326,
      "eval_samples_per_second": 3.135,
      "eval_steps_per_second": 0.448,
      "step": 700
    },
    {
      "epoch": 0.24,
      "grad_norm": 3.229812499737018,
      "learning_rate": 4.514412764152446e-05,
      "loss": 0.202,
      "step": 725
    },
    {
      "epoch": 0.25,
      "grad_norm": 4.16567522590051,
      "learning_rate": 4.4728512734909844e-05,
      "loss": 0.1716,
      "step": 750
    },
    {
      "epoch": 0.26,
      "grad_norm": 2.949438330771571,
      "learning_rate": 4.4297916272908024e-05,
      "loss": 0.1658,
      "step": 775
    },
    {
      "epoch": 0.27,
      "grad_norm": 3.866403478843968,
      "learning_rate": 4.385266524442241e-05,
      "loss": 0.164,
      "step": 800
    },
    {
      "epoch": 0.27,
      "eval_loss": 6.8642449378967285,
      "eval_runtime": 2.0035,
      "eval_samples_per_second": 3.494,
      "eval_steps_per_second": 0.499,
      "step": 800
    },
    {
      "epoch": 0.28,
      "grad_norm": 2.5637077564702793,
      "learning_rate": 4.3393097766828293e-05,
      "loss": 0.1646,
      "step": 825
    },
    {
      "epoch": 0.28,
      "grad_norm": 3.2208975889798532,
      "learning_rate": 4.2919562829211283e-05,
      "loss": 0.1649,
      "step": 850
    },
    {
      "epoch": 0.29,
      "grad_norm": 4.840386620025996,
      "learning_rate": 4.2432420027349886e-05,
      "loss": 0.1565,
      "step": 875
    },
    {
      "epoch": 0.3,
      "grad_norm": 2.732903459358712,
      "learning_rate": 4.193203929064353e-05,
      "loss": 0.149,
      "step": 900
    },
    {
      "epoch": 0.3,
      "eval_loss": 6.954349517822266,
      "eval_runtime": 1.9483,
      "eval_samples_per_second": 3.593,
      "eval_steps_per_second": 0.513,
      "step": 900
    },
    {
      "epoch": 0.31,
      "grad_norm": 5.731747492996942,
      "learning_rate": 4.141880060119336e-05,
      "loss": 0.1414,
      "step": 925
    },
    {
      "epoch": 0.32,
      "grad_norm": 2.6801789029945224,
      "learning_rate": 4.089309370524921e-05,
      "loss": 0.1349,
      "step": 950
    },
    {
      "epoch": 0.33,
      "grad_norm": 3.545251177735513,
      "learning_rate": 4.03553178172417e-05,
      "loss": 0.1349,
      "step": 975
    },
    {
      "epoch": 0.33,
      "grad_norm": 2.805948855969587,
      "learning_rate": 3.9805881316624506e-05,
      "loss": 0.1288,
      "step": 1000
    },
    {
      "epoch": 0.33,
      "eval_loss": 7.151072025299072,
      "eval_runtime": 2.0089,
      "eval_samples_per_second": 3.484,
      "eval_steps_per_second": 0.498,
      "step": 1000
    },
    {
      "epoch": 0.34,
      "grad_norm": 3.359524866205609,
      "learning_rate": 3.9245201437756654e-05,
      "loss": 0.133,
      "step": 1025
    },
    {
      "epoch": 0.35,
      "grad_norm": 2.4513945515951034,
      "learning_rate": 3.867370395306068e-05,
      "loss": 0.1283,
      "step": 1050
    },
    {
      "epoch": 0.36,
      "grad_norm": 2.7251550008679124,
      "learning_rate": 3.8091822849696954e-05,
      "loss": 0.1296,
      "step": 1075
    },
    {
      "epoch": 0.37,
      "grad_norm": 2.3408176490646735,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.1174,
      "step": 1100
    },
    {
      "epoch": 0.37,
      "eval_loss": 7.04253625869751,
      "eval_runtime": 2.1128,
      "eval_samples_per_second": 3.313,
      "eval_steps_per_second": 0.473,
      "step": 1100
    },
    {
      "epoch": 0.38,
      "grad_norm": 2.345049499419834,
      "learning_rate": 3.689868482592684e-05,
      "loss": 0.1222,
      "step": 1125
    },
    {
      "epoch": 0.38,
      "grad_norm": 2.252502312200347,
      "learning_rate": 3.628833395777224e-05,
      "loss": 0.1123,
      "step": 1150
    },
    {
      "epoch": 0.39,
      "grad_norm": 2.255479925896646,
      "learning_rate": 3.5669410887410095e-05,
      "loss": 0.1109,
      "step": 1175
    },
    {
      "epoch": 0.4,
      "grad_norm": 5.433572539291945,
      "learning_rate": 3.504238561632424e-05,
      "loss": 0.1055,
      "step": 1200
    },
    {
      "epoch": 0.4,
      "eval_loss": 6.962005138397217,
      "eval_runtime": 2.3652,
      "eval_samples_per_second": 2.96,
      "eval_steps_per_second": 0.423,
      "step": 1200
    },
    {
      "epoch": 0.41,
      "grad_norm": 1.894777201625373,
      "learning_rate": 3.4407734298695895e-05,
      "loss": 0.0971,
      "step": 1225
    },
    {
      "epoch": 0.42,
      "grad_norm": 13.586848448823995,
      "learning_rate": 3.376593887981887e-05,
      "loss": 0.097,
      "step": 1250
    },
    {
      "epoch": 0.42,
      "grad_norm": 1.910777490633546,
      "learning_rate": 3.311748673011709e-05,
      "loss": 0.095,
      "step": 1275
    },
    {
      "epoch": 0.43,
      "grad_norm": 2.213597934122813,
      "learning_rate": 3.246287027504237e-05,
      "loss": 0.0923,
      "step": 1300
    },
    {
      "epoch": 0.43,
      "eval_loss": 7.195333003997803,
      "eval_runtime": 1.8349,
      "eval_samples_per_second": 3.815,
      "eval_steps_per_second": 0.545,
      "step": 1300
    },
    {
      "epoch": 0.44,
      "grad_norm": 1.8556555959500387,
      "learning_rate": 3.180258662113338e-05,
      "loss": 0.0899,
      "step": 1325
    },
    {
      "epoch": 0.45,
      "grad_norm": 1.8171299789610367,
      "learning_rate": 3.1137137178519985e-05,
      "loss": 0.088,
      "step": 1350
    },
    {
      "epoch": 0.46,
      "grad_norm": 1.6125472661304423,
      "learning_rate": 3.04670272801594e-05,
      "loss": 0.0848,
      "step": 1375
    },
    {
      "epoch": 0.47,
      "grad_norm": 1.6146107012694884,
      "learning_rate": 2.9792765798093465e-05,
      "loss": 0.0823,
      "step": 1400
    },
    {
      "epoch": 0.47,
      "eval_loss": 7.175464630126953,
      "eval_runtime": 2.305,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 0.434,
      "step": 1400
    },
    {
      "epoch": 0.47,
      "grad_norm": 1.6339734948096556,
      "learning_rate": 2.9114864757018352e-05,
      "loss": 0.0775,
      "step": 1425
    },
    {
      "epoch": 0.48,
      "grad_norm": 1.3803334577313728,
      "learning_rate": 2.8433838945460205e-05,
      "loss": 0.0778,
      "step": 1450
    },
    {
      "epoch": 0.49,
      "grad_norm": 4.1834888729216155,
      "learning_rate": 2.7750205524851986e-05,
      "loss": 0.0724,
      "step": 1475
    },
    {
      "epoch": 0.5,
      "grad_norm": 1.3957828771921776,
      "learning_rate": 2.7064483636808313e-05,
      "loss": 0.0743,
      "step": 1500
    },
    {
      "epoch": 0.5,
      "eval_loss": 7.146019458770752,
      "eval_runtime": 2.0516,
      "eval_samples_per_second": 3.412,
      "eval_steps_per_second": 0.487,
      "step": 1500
    },
    {
      "epoch": 0.51,
      "grad_norm": 2.3774136895238387,
      "learning_rate": 2.6377194008896637e-05,
      "loss": 0.0721,
      "step": 1525
    },
    {
      "epoch": 0.52,
      "grad_norm": 1.4294952617652887,
      "learning_rate": 2.5688858559204053e-05,
      "loss": 0.07,
      "step": 1550
    },
    {
      "epoch": 0.53,
      "grad_norm": 1.7503317291445033,
      "learning_rate": 2.5e-05,
      "loss": 0.0689,
      "step": 1575
    },
    {
      "epoch": 0.53,
      "grad_norm": 1.8141634588926774,
      "learning_rate": 2.4311141440795953e-05,
      "loss": 0.0692,
      "step": 1600
    },
    {
      "epoch": 0.53,
      "eval_loss": 7.258533477783203,
      "eval_runtime": 2.0233,
      "eval_samples_per_second": 3.46,
      "eval_steps_per_second": 0.494,
      "step": 1600
    },
    {
      "epoch": 0.54,
      "grad_norm": 1.5420686364317522,
      "learning_rate": 2.3622805991103362e-05,
      "loss": 0.0659,
      "step": 1625
    },
    {
      "epoch": 0.55,
      "grad_norm": 1.3888120359776974,
      "learning_rate": 2.2935516363191693e-05,
      "loss": 0.0634,
      "step": 1650
    },
    {
      "epoch": 0.56,
      "grad_norm": 1.2546701670441247,
      "learning_rate": 2.224979447514802e-05,
      "loss": 0.0645,
      "step": 1675
    },
    {
      "epoch": 0.57,
      "grad_norm": 1.2106681885887598,
      "learning_rate": 2.1566161054539798e-05,
      "loss": 0.0607,
      "step": 1700
    },
    {
      "epoch": 0.57,
      "eval_loss": 7.4217352867126465,
      "eval_runtime": 2.0821,
      "eval_samples_per_second": 3.362,
      "eval_steps_per_second": 0.48,
      "step": 1700
    },
    {
      "epoch": 0.57,
      "grad_norm": 1.8579148754464851,
      "learning_rate": 2.088513524298165e-05,
      "loss": 0.0628,
      "step": 1725
    },
    {
      "epoch": 0.58,
      "grad_norm": 2.182376129737232,
      "learning_rate": 2.0207234201906547e-05,
      "loss": 0.0599,
      "step": 1750
    },
    {
      "epoch": 0.59,
      "grad_norm": 1.322832200578898,
      "learning_rate": 1.9532972719840607e-05,
      "loss": 0.0572,
      "step": 1775
    },
    {
      "epoch": 0.6,
      "grad_norm": 2.8440718747760045,
      "learning_rate": 1.8862862821480025e-05,
      "loss": 0.0566,
      "step": 1800
    },
    {
      "epoch": 0.6,
      "eval_loss": 7.199245452880859,
      "eval_runtime": 2.2957,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.436,
      "step": 1800
    },
    {
      "epoch": 0.61,
      "grad_norm": 1.0329563875586547,
      "learning_rate": 1.819741337886662e-05,
      "loss": 0.0524,
      "step": 1825
    },
    {
      "epoch": 0.62,
      "grad_norm": 1.4496258780971825,
      "learning_rate": 1.7537129724957642e-05,
      "loss": 0.0559,
      "step": 1850
    },
    {
      "epoch": 0.62,
      "grad_norm": 1.419911688184789,
      "learning_rate": 1.6882513269882917e-05,
      "loss": 0.0541,
      "step": 1875
    },
    {
      "epoch": 0.63,
      "grad_norm": 1.1492711217382043,
      "learning_rate": 1.6234061120181142e-05,
      "loss": 0.0522,
      "step": 1900
    },
    {
      "epoch": 0.63,
      "eval_loss": 7.336871147155762,
      "eval_runtime": 1.9969,
      "eval_samples_per_second": 3.505,
      "eval_steps_per_second": 0.501,
      "step": 1900
    },
    {
      "epoch": 0.64,
      "grad_norm": 1.1314206062794803,
      "learning_rate": 1.5592265701304114e-05,
      "loss": 0.0515,
      "step": 1925
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.992667486842559,
      "learning_rate": 1.495761438367577e-05,
      "loss": 0.0529,
      "step": 1950
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.9083563595052039,
      "learning_rate": 1.433058911258991e-05,
      "loss": 0.0481,
      "step": 1975
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.9921561622348036,
      "learning_rate": 1.3711666042227772e-05,
      "loss": 0.0504,
      "step": 2000
    },
    {
      "epoch": 0.67,
      "eval_loss": 7.345071792602539,
      "eval_runtime": 2.216,
      "eval_samples_per_second": 3.159,
      "eval_steps_per_second": 0.451,
      "step": 2000
    }
  ],
  "logging_steps": 25,
  "max_steps": 3000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 1000,
  "total_flos": 104676741611520.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}