{
  "best_global_step": 225,
  "best_metric": 0.36194586753845215,
  "best_model_checkpoint": "/content/output/checkpoint-225",
  "epoch": 13.0,
  "eval_steps": 500,
  "global_step": 975,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.13333333333333333,
      "grad_norm": 1.7940553426742554,
      "learning_rate": 5.917159763313609e-06,
      "loss": 1.3972,
      "mean_token_accuracy": 0.7618847399950027,
      "num_tokens": 31152.0,
      "step": 10
    },
    {
      "epoch": 0.26666666666666666,
      "grad_norm": 1.0511088371276855,
      "learning_rate": 1.1834319526627219e-05,
      "loss": 1.2288,
      "mean_token_accuracy": 0.7673537939786911,
      "num_tokens": 61773.0,
      "step": 20
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8096975088119507,
      "learning_rate": 1.7751479289940828e-05,
      "loss": 0.9255,
      "mean_token_accuracy": 0.7948055118322372,
      "num_tokens": 94758.0,
      "step": 30
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 0.6213255524635315,
      "learning_rate": 2.3668639053254438e-05,
      "loss": 0.7366,
      "mean_token_accuracy": 0.8212289035320282,
      "num_tokens": 125915.0,
      "step": 40
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 0.42887863516807556,
      "learning_rate": 2.958579881656805e-05,
      "loss": 0.5556,
      "mean_token_accuracy": 0.8463607966899872,
      "num_tokens": 159364.0,
      "step": 50
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.5108550786972046,
      "learning_rate": 3.5502958579881656e-05,
      "loss": 0.4676,
      "mean_token_accuracy": 0.8614094287157059,
      "num_tokens": 190374.0,
      "step": 60
    },
    {
      "epoch": 0.9333333333333333,
      "grad_norm": 0.4310983121395111,
      "learning_rate": 4.142011834319527e-05,
      "loss": 0.4214,
      "mean_token_accuracy": 0.8710317760705948,
      "num_tokens": 222156.0,
      "step": 70
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.4043419659137726,
      "eval_mean_token_accuracy": 0.8734077858924866,
      "eval_num_tokens": 238011.0,
      "eval_runtime": 50.8925,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 75
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 0.35767048597335815,
      "learning_rate": 4.7337278106508875e-05,
      "loss": 0.3901,
      "mean_token_accuracy": 0.8787548273801804,
      "num_tokens": 255825.0,
      "step": 80
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.3684692680835724,
      "learning_rate": 5.3254437869822495e-05,
      "loss": 0.3956,
      "mean_token_accuracy": 0.8787799149751663,
      "num_tokens": 287370.0,
      "step": 90
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 0.35148611664772034,
      "learning_rate": 5.91715976331361e-05,
      "loss": 0.3822,
      "mean_token_accuracy": 0.8814697057008744,
      "num_tokens": 315883.0,
      "step": 100
    },
    {
      "epoch": 1.4666666666666668,
      "grad_norm": 0.39074671268463135,
      "learning_rate": 6.50887573964497e-05,
      "loss": 0.3703,
      "mean_token_accuracy": 0.8836676925420761,
      "num_tokens": 347324.0,
      "step": 110
    },
    {
      "epoch": 1.6,
      "grad_norm": 0.38130292296409607,
      "learning_rate": 7.100591715976331e-05,
      "loss": 0.3762,
      "mean_token_accuracy": 0.8851484537124634,
      "num_tokens": 379374.0,
      "step": 120
    },
    {
      "epoch": 1.7333333333333334,
      "grad_norm": 0.4091486632823944,
      "learning_rate": 7.692307692307693e-05,
      "loss": 0.3446,
      "mean_token_accuracy": 0.8894414573907852,
      "num_tokens": 411843.0,
      "step": 130
    },
    {
      "epoch": 1.8666666666666667,
      "grad_norm": 0.361017644405365,
      "learning_rate": 8.284023668639054e-05,
      "loss": 0.3494,
      "mean_token_accuracy": 0.8915452778339386,
      "num_tokens": 444017.0,
      "step": 140
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.3487047851085663,
      "learning_rate": 8.875739644970414e-05,
      "loss": 0.3674,
      "mean_token_accuracy": 0.8846240967512131,
      "num_tokens": 476022.0,
      "step": 150
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.3620273470878601,
      "eval_mean_token_accuracy": 0.8875497984886169,
      "eval_num_tokens": 476022.0,
      "eval_runtime": 50.8961,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 150
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 0.480814665555954,
      "learning_rate": 9.467455621301775e-05,
      "loss": 0.2974,
      "mean_token_accuracy": 0.9050114572048187,
      "num_tokens": 508445.0,
      "step": 160
    },
    {
      "epoch": 2.2666666666666666,
      "grad_norm": 0.40775951743125916,
      "learning_rate": 9.989539748953976e-05,
      "loss": 0.2896,
      "mean_token_accuracy": 0.9084265738725662,
      "num_tokens": 541530.0,
      "step": 170
    },
    {
      "epoch": 2.4,
      "grad_norm": 0.5174453854560852,
      "learning_rate": 9.884937238493725e-05,
      "loss": 0.29,
      "mean_token_accuracy": 0.9073476493358612,
      "num_tokens": 572212.0,
      "step": 180
    },
    {
      "epoch": 2.533333333333333,
      "grad_norm": 0.5889921188354492,
      "learning_rate": 9.780334728033474e-05,
      "loss": 0.2828,
      "mean_token_accuracy": 0.909446981549263,
      "num_tokens": 603269.0,
      "step": 190
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 0.5635319352149963,
      "learning_rate": 9.675732217573223e-05,
      "loss": 0.2744,
      "mean_token_accuracy": 0.9101487964391708,
      "num_tokens": 635792.0,
      "step": 200
    },
    {
      "epoch": 2.8,
      "grad_norm": 0.609230637550354,
      "learning_rate": 9.571129707112972e-05,
      "loss": 0.2863,
      "mean_token_accuracy": 0.9104649156332016,
      "num_tokens": 666470.0,
      "step": 210
    },
    {
      "epoch": 2.9333333333333336,
      "grad_norm": 0.5209991931915283,
      "learning_rate": 9.46652719665272e-05,
      "loss": 0.3027,
      "mean_token_accuracy": 0.9050683736801147,
      "num_tokens": 699119.0,
      "step": 220
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.36194586753845215,
      "eval_mean_token_accuracy": 0.8886684679985046,
      "eval_num_tokens": 714033.0,
      "eval_runtime": 50.8962,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 225
    },
    {
      "epoch": 3.066666666666667,
      "grad_norm": 0.5214265584945679,
      "learning_rate": 9.361924686192469e-05,
      "loss": 0.2371,
      "mean_token_accuracy": 0.9248212277889252,
      "num_tokens": 729520.0,
      "step": 230
    },
    {
      "epoch": 3.2,
      "grad_norm": 0.7459234595298767,
      "learning_rate": 9.257322175732218e-05,
      "loss": 0.1874,
      "mean_token_accuracy": 0.939895498752594,
      "num_tokens": 759510.0,
      "step": 240
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 0.614969789981842,
      "learning_rate": 9.152719665271967e-05,
      "loss": 0.1754,
      "mean_token_accuracy": 0.943790751695633,
      "num_tokens": 791875.0,
      "step": 250
    },
    {
      "epoch": 3.466666666666667,
      "grad_norm": 0.8585807085037231,
      "learning_rate": 9.048117154811716e-05,
      "loss": 0.1794,
      "mean_token_accuracy": 0.9401535421609879,
      "num_tokens": 821467.0,
      "step": 260
    },
    {
      "epoch": 3.6,
      "grad_norm": 0.6798614263534546,
      "learning_rate": 8.943514644351465e-05,
      "loss": 0.1754,
      "mean_token_accuracy": 0.9439484149217605,
      "num_tokens": 854000.0,
      "step": 270
    },
    {
      "epoch": 3.7333333333333334,
      "grad_norm": 0.6594717502593994,
      "learning_rate": 8.838912133891214e-05,
      "loss": 0.179,
      "mean_token_accuracy": 0.9433883309364319,
      "num_tokens": 885296.0,
      "step": 280
    },
    {
      "epoch": 3.8666666666666667,
      "grad_norm": 0.6909582614898682,
      "learning_rate": 8.734309623430963e-05,
      "loss": 0.1855,
      "mean_token_accuracy": 0.9395292073488235,
      "num_tokens": 919387.0,
      "step": 290
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.7755655646324158,
      "learning_rate": 8.629707112970712e-05,
      "loss": 0.1851,
      "mean_token_accuracy": 0.9409447878599166,
      "num_tokens": 952044.0,
      "step": 300
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.39551886916160583,
      "eval_mean_token_accuracy": 0.887341833114624,
      "eval_num_tokens": 952044.0,
      "eval_runtime": 50.8937,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 300
    },
    {
      "epoch": 4.133333333333334,
      "grad_norm": 0.6951614022254944,
      "learning_rate": 8.525104602510461e-05,
      "loss": 0.0942,
      "mean_token_accuracy": 0.9724623709917068,
      "num_tokens": 984335.0,
      "step": 310
    },
    {
      "epoch": 4.266666666666667,
      "grad_norm": 0.6803505420684814,
      "learning_rate": 8.42050209205021e-05,
      "loss": 0.1073,
      "mean_token_accuracy": 0.9653927534818649,
      "num_tokens": 1014854.0,
      "step": 320
    },
    {
      "epoch": 4.4,
      "grad_norm": 0.7284913063049316,
      "learning_rate": 8.315899581589958e-05,
      "loss": 0.1054,
      "mean_token_accuracy": 0.9662943929433823,
      "num_tokens": 1047986.0,
      "step": 330
    },
    {
      "epoch": 4.533333333333333,
      "grad_norm": 0.6803147792816162,
      "learning_rate": 8.211297071129707e-05,
      "loss": 0.1065,
      "mean_token_accuracy": 0.9682819366455078,
      "num_tokens": 1079663.0,
      "step": 340
    },
    {
      "epoch": 4.666666666666667,
      "grad_norm": 0.6864665150642395,
      "learning_rate": 8.106694560669456e-05,
      "loss": 0.1085,
      "mean_token_accuracy": 0.9646553307771683,
      "num_tokens": 1109495.0,
      "step": 350
    },
    {
      "epoch": 4.8,
      "grad_norm": 0.7594568729400635,
      "learning_rate": 8.002092050209205e-05,
      "loss": 0.1079,
      "mean_token_accuracy": 0.9674631506204605,
      "num_tokens": 1141598.0,
      "step": 360
    },
    {
      "epoch": 4.933333333333334,
      "grad_norm": 0.6360507607460022,
      "learning_rate": 7.897489539748954e-05,
      "loss": 0.118,
      "mean_token_accuracy": 0.9628556370735168,
      "num_tokens": 1172834.0,
      "step": 370
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.45760437846183777,
      "eval_mean_token_accuracy": 0.8874933505058289,
      "eval_num_tokens": 1190055.0,
      "eval_runtime": 50.9015,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 375
    },
    {
      "epoch": 5.066666666666666,
      "grad_norm": 0.4860494136810303,
      "learning_rate": 7.792887029288704e-05,
      "loss": 0.0845,
      "mean_token_accuracy": 0.9751860290765763,
      "num_tokens": 1206554.0,
      "step": 380
    },
    {
      "epoch": 5.2,
      "grad_norm": 0.6970198750495911,
      "learning_rate": 7.688284518828453e-05,
      "loss": 0.065,
      "mean_token_accuracy": 0.9810720026493073,
      "num_tokens": 1237792.0,
      "step": 390
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 0.8283060193061829,
      "learning_rate": 7.583682008368202e-05,
      "loss": 0.0663,
      "mean_token_accuracy": 0.9797625124454499,
      "num_tokens": 1270203.0,
      "step": 400
    },
    {
      "epoch": 5.466666666666667,
      "grad_norm": 0.8080848455429077,
      "learning_rate": 7.479079497907951e-05,
      "loss": 0.0712,
      "mean_token_accuracy": 0.9780115723609925,
      "num_tokens": 1301414.0,
      "step": 410
    },
    {
      "epoch": 5.6,
      "grad_norm": 0.7461147904396057,
      "learning_rate": 7.3744769874477e-05,
      "loss": 0.0678,
      "mean_token_accuracy": 0.9800621330738067,
      "num_tokens": 1332421.0,
      "step": 420
    },
    {
      "epoch": 5.733333333333333,
      "grad_norm": 0.5786488056182861,
      "learning_rate": 7.269874476987449e-05,
      "loss": 0.067,
      "mean_token_accuracy": 0.980746528506279,
      "num_tokens": 1365542.0,
      "step": 430
    },
    {
      "epoch": 5.866666666666667,
      "grad_norm": 0.6561855673789978,
      "learning_rate": 7.165271966527197e-05,
      "loss": 0.0702,
      "mean_token_accuracy": 0.9795888513326645,
      "num_tokens": 1396981.0,
      "step": 440
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.8029749989509583,
      "learning_rate": 7.060669456066946e-05,
      "loss": 0.0721,
      "mean_token_accuracy": 0.9787283718585968,
      "num_tokens": 1428066.0,
      "step": 450
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.5095443725585938,
      "eval_mean_token_accuracy": 0.8847648048400879,
      "eval_num_tokens": 1428066.0,
      "eval_runtime": 50.9006,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 450
    },
    {
      "epoch": 6.133333333333334,
      "grad_norm": 0.5984286665916443,
      "learning_rate": 6.956066945606695e-05,
      "loss": 0.0439,
      "mean_token_accuracy": 0.9883647084236145,
      "num_tokens": 1459612.0,
      "step": 460
    },
    {
      "epoch": 6.266666666666667,
      "grad_norm": 0.7371001839637756,
      "learning_rate": 6.851464435146444e-05,
      "loss": 0.0497,
      "mean_token_accuracy": 0.9854195445775986,
      "num_tokens": 1490815.0,
      "step": 470
    },
    {
      "epoch": 6.4,
      "grad_norm": 0.5508085489273071,
      "learning_rate": 6.746861924686193e-05,
      "loss": 0.0477,
      "mean_token_accuracy": 0.9866609632968902,
      "num_tokens": 1520943.0,
      "step": 480
    },
    {
      "epoch": 6.533333333333333,
      "grad_norm": 0.7761865854263306,
      "learning_rate": 6.642259414225942e-05,
      "loss": 0.0449,
      "mean_token_accuracy": 0.9868600249290467,
      "num_tokens": 1552633.0,
      "step": 490
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 0.6642009615898132,
      "learning_rate": 6.537656903765691e-05,
      "loss": 0.0472,
      "mean_token_accuracy": 0.9871610105037689,
      "num_tokens": 1587798.0,
      "step": 500
    },
    {
      "epoch": 6.8,
      "grad_norm": 0.6801927089691162,
      "learning_rate": 6.43305439330544e-05,
      "loss": 0.0527,
      "mean_token_accuracy": 0.9849318206310272,
      "num_tokens": 1617597.0,
      "step": 510
    },
    {
      "epoch": 6.933333333333334,
      "grad_norm": 0.6705682873725891,
      "learning_rate": 6.32845188284519e-05,
      "loss": 0.0505,
      "mean_token_accuracy": 0.9851646542549133,
      "num_tokens": 1648212.0,
      "step": 520
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.5524249076843262,
      "eval_mean_token_accuracy": 0.8823326063156128,
      "eval_num_tokens": 1666077.0,
      "eval_runtime": 50.8934,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 525
    },
    {
      "epoch": 7.066666666666666,
      "grad_norm": 1.0225543975830078,
      "learning_rate": 6.223849372384938e-05,
      "loss": 0.0441,
      "mean_token_accuracy": 0.9876838475465775,
      "num_tokens": 1682282.0,
      "step": 530
    },
    {
      "epoch": 7.2,
      "grad_norm": 0.7063732743263245,
      "learning_rate": 6.119246861924686e-05,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9909861594438553,
      "num_tokens": 1714966.0,
      "step": 540
    },
    {
      "epoch": 7.333333333333333,
      "grad_norm": 0.4836845099925995,
      "learning_rate": 6.014644351464436e-05,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9898746073246002,
      "num_tokens": 1745662.0,
      "step": 550
    },
    {
      "epoch": 7.466666666666667,
      "grad_norm": 0.5448402166366577,
      "learning_rate": 5.910041841004185e-05,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9898856014013291,
      "num_tokens": 1778368.0,
      "step": 560
    },
    {
      "epoch": 7.6,
      "grad_norm": 0.4760727882385254,
      "learning_rate": 5.8054393305439334e-05,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9897664040327072,
      "num_tokens": 1810571.0,
      "step": 570
    },
    {
      "epoch": 7.733333333333333,
      "grad_norm": 0.43467867374420166,
      "learning_rate": 5.7008368200836825e-05,
      "loss": 0.0388,
      "mean_token_accuracy": 0.9894861280918121,
      "num_tokens": 1842310.0,
      "step": 580
    },
    {
      "epoch": 7.866666666666667,
      "grad_norm": 0.8253856897354126,
      "learning_rate": 5.5962343096234316e-05,
      "loss": 0.0398,
      "mean_token_accuracy": 0.9881662368774414,
      "num_tokens": 1873435.0,
      "step": 590
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.5036665797233582,
      "learning_rate": 5.4916317991631806e-05,
      "loss": 0.0421,
      "mean_token_accuracy": 0.9883299976587295,
      "num_tokens": 1904088.0,
      "step": 600
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.5785043239593506,
      "eval_mean_token_accuracy": 0.8845841264724732,
      "eval_num_tokens": 1904088.0,
      "eval_runtime": 50.8908,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 600
    },
    {
      "epoch": 8.133333333333333,
      "grad_norm": 0.24711328744888306,
      "learning_rate": 5.38702928870293e-05,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9923271119594574,
      "num_tokens": 1936241.0,
      "step": 610
    },
    {
      "epoch": 8.266666666666667,
      "grad_norm": 0.44961780309677124,
      "learning_rate": 5.282426778242678e-05,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9920144468545914,
      "num_tokens": 1969247.0,
      "step": 620
    },
    {
      "epoch": 8.4,
      "grad_norm": 0.3607077896595001,
      "learning_rate": 5.177824267782427e-05,
      "loss": 0.032,
      "mean_token_accuracy": 0.9911372780799865,
      "num_tokens": 2001136.0,
      "step": 630
    },
    {
      "epoch": 8.533333333333333,
      "grad_norm": 0.34154054522514343,
      "learning_rate": 5.073221757322176e-05,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9917692422866822,
      "num_tokens": 2033446.0,
      "step": 640
    },
    {
      "epoch": 8.666666666666666,
      "grad_norm": 0.5010355710983276,
      "learning_rate": 4.968619246861925e-05,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9908981770277023,
      "num_tokens": 2064800.0,
      "step": 650
    },
    {
      "epoch": 8.8,
      "grad_norm": 0.47791624069213867,
      "learning_rate": 4.864016736401674e-05,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9913896471261978,
      "num_tokens": 2095573.0,
      "step": 660
    },
    {
      "epoch": 8.933333333333334,
      "grad_norm": 0.45093876123428345,
      "learning_rate": 4.759414225941423e-05,
      "loss": 0.0371,
      "mean_token_accuracy": 0.9898908942937851,
      "num_tokens": 2126001.0,
      "step": 670
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.6148654222488403,
      "eval_mean_token_accuracy": 0.8842980313301086,
      "eval_num_tokens": 2142099.0,
      "eval_runtime": 50.8928,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 675
    },
    {
      "epoch": 9.066666666666666,
      "grad_norm": 0.5800787806510925,
      "learning_rate": 4.654811715481171e-05,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9921304136514664,
      "num_tokens": 2157400.0,
      "step": 680
    },
    {
      "epoch": 9.2,
      "grad_norm": 0.6982473731040955,
      "learning_rate": 4.5502092050209203e-05,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9932474166154861,
      "num_tokens": 2189379.0,
      "step": 690
    },
    {
      "epoch": 9.333333333333334,
      "grad_norm": 0.3595369756221771,
      "learning_rate": 4.4456066945606694e-05,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9921673446893692,
      "num_tokens": 2219752.0,
      "step": 700
    },
    {
      "epoch": 9.466666666666667,
      "grad_norm": 0.27210283279418945,
      "learning_rate": 4.3410041841004185e-05,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9919114917516708,
      "num_tokens": 2251801.0,
      "step": 710
    },
    {
      "epoch": 9.6,
      "grad_norm": 0.45378750562667847,
      "learning_rate": 4.2364016736401676e-05,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9923428118228912,
      "num_tokens": 2283166.0,
      "step": 720
    },
    {
      "epoch": 9.733333333333333,
      "grad_norm": 0.21831561625003815,
      "learning_rate": 4.131799163179916e-05,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9930205404758453,
      "num_tokens": 2315753.0,
      "step": 730
    },
    {
      "epoch": 9.866666666666667,
      "grad_norm": 0.3201010525226593,
      "learning_rate": 4.027196652719665e-05,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9924349457025528,
      "num_tokens": 2348020.0,
      "step": 740
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.25126612186431885,
      "learning_rate": 3.922594142259414e-05,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9927875697612762,
      "num_tokens": 2380110.0,
      "step": 750
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.6573454141616821,
      "eval_mean_token_accuracy": 0.8831826686859131,
      "eval_num_tokens": 2380110.0,
      "eval_runtime": 50.8906,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 750
    },
    {
      "epoch": 10.133333333333333,
      "grad_norm": 0.5593218803405762,
      "learning_rate": 3.817991631799163e-05,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9936788022518158,
      "num_tokens": 2412729.0,
      "step": 760
    },
    {
      "epoch": 10.266666666666667,
      "grad_norm": 0.3219089210033417,
      "learning_rate": 3.713389121338912e-05,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9932976424694061,
      "num_tokens": 2443952.0,
      "step": 770
    },
    {
      "epoch": 10.4,
      "grad_norm": 0.14705878496170044,
      "learning_rate": 3.6087866108786614e-05,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9936914265155792,
      "num_tokens": 2475185.0,
      "step": 780
    },
    {
      "epoch": 10.533333333333333,
      "grad_norm": 0.2050597369670868,
      "learning_rate": 3.50418410041841e-05,
      "loss": 0.024,
      "mean_token_accuracy": 0.9928701817989349,
      "num_tokens": 2506064.0,
      "step": 790
    },
    {
      "epoch": 10.666666666666666,
      "grad_norm": 0.30800071358680725,
      "learning_rate": 3.399581589958159e-05,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9936599344015121,
      "num_tokens": 2537568.0,
      "step": 800
    },
    {
      "epoch": 10.8,
      "grad_norm": 0.6075056195259094,
      "learning_rate": 3.294979079497908e-05,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9924876004457474,
      "num_tokens": 2568278.0,
      "step": 810
    },
    {
      "epoch": 10.933333333333334,
      "grad_norm": 0.46507900953292847,
      "learning_rate": 3.190376569037657e-05,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9935177505016327,
      "num_tokens": 2601290.0,
      "step": 820
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.7078420519828796,
      "eval_mean_token_accuracy": 0.8819138598442078,
      "eval_num_tokens": 2618121.0,
      "eval_runtime": 50.8936,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 825
    },
    {
      "epoch": 11.066666666666666,
      "grad_norm": 0.13493098318576813,
      "learning_rate": 3.085774058577406e-05,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9937657356262207,
      "num_tokens": 2631945.0,
      "step": 830
    },
    {
      "epoch": 11.2,
      "grad_norm": 0.27274149656295776,
      "learning_rate": 2.981171548117155e-05,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9939479857683182,
      "num_tokens": 2664278.0,
      "step": 840
    },
    {
      "epoch": 11.333333333333334,
      "grad_norm": 0.12694329023361206,
      "learning_rate": 2.8765690376569036e-05,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9940280497074128,
      "num_tokens": 2697487.0,
      "step": 850
    },
    {
      "epoch": 11.466666666666667,
      "grad_norm": 0.30919721722602844,
      "learning_rate": 2.7719665271966527e-05,
      "loss": 0.0208,
      "mean_token_accuracy": 0.993863531947136,
      "num_tokens": 2729287.0,
      "step": 860
    },
    {
      "epoch": 11.6,
      "grad_norm": 0.346965491771698,
      "learning_rate": 2.6673640167364018e-05,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9938155382871627,
      "num_tokens": 2760283.0,
      "step": 870
    },
    {
      "epoch": 11.733333333333333,
      "grad_norm": 0.22179089486598969,
      "learning_rate": 2.5627615062761505e-05,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9939759701490403,
      "num_tokens": 2793986.0,
      "step": 880
    },
    {
      "epoch": 11.866666666666667,
      "grad_norm": 0.16201026737689972,
      "learning_rate": 2.4581589958158996e-05,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9931929171085357,
      "num_tokens": 2824816.0,
      "step": 890
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.20347487926483154,
      "learning_rate": 2.3535564853556487e-05,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9934813290834427,
      "num_tokens": 2856132.0,
      "step": 900
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.7223861813545227,
      "eval_mean_token_accuracy": 0.8824266290664673,
      "eval_num_tokens": 2856132.0,
      "eval_runtime": 50.8977,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 900
    },
    {
      "epoch": 12.133333333333333,
      "grad_norm": 0.10506761819124222,
      "learning_rate": 2.2489539748953974e-05,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9947699904441833,
      "num_tokens": 2889294.0,
      "step": 910
    },
    {
      "epoch": 12.266666666666667,
      "grad_norm": 0.22959034144878387,
      "learning_rate": 2.1443514644351465e-05,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9941918909549713,
      "num_tokens": 2920564.0,
      "step": 920
    },
    {
      "epoch": 12.4,
      "grad_norm": 0.14024658501148224,
      "learning_rate": 2.0397489539748953e-05,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9944606810808182,
      "num_tokens": 2952215.0,
      "step": 930
    },
    {
      "epoch": 12.533333333333333,
      "grad_norm": 0.1049821749329567,
      "learning_rate": 1.9351464435146444e-05,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9942614287137985,
      "num_tokens": 2984367.0,
      "step": 940
    },
    {
      "epoch": 12.666666666666666,
      "grad_norm": 0.14182651042938232,
      "learning_rate": 1.8305439330543934e-05,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9938604891300201,
      "num_tokens": 3015120.0,
      "step": 950
    },
    {
      "epoch": 12.8,
      "grad_norm": 0.13577786087989807,
      "learning_rate": 1.7259414225941422e-05,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9941955178976059,
      "num_tokens": 3046751.0,
      "step": 960
    },
    {
      "epoch": 12.933333333333334,
      "grad_norm": 0.18529069423675537,
      "learning_rate": 1.6213389121338913e-05,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9938661843538285,
      "num_tokens": 3078212.0,
      "step": 970
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.7834702134132385,
      "eval_mean_token_accuracy": 0.8812357115745545,
      "eval_num_tokens": 3094143.0,
      "eval_runtime": 50.9004,
      "eval_samples_per_second": 1.965,
      "eval_steps_per_second": 0.491,
      "step": 975
    }
  ],
  "logging_steps": 10,
  "max_steps": 1125,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.854305945570427e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}