{
  "best_metric": 3.1981074810028076,
  "best_model_checkpoint": "checkpoints/it5-base/checkpoint-27001",
  "epoch": 9.751173708920188,
  "eval_steps": 2077,
  "global_step": 27001,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14987360057782592,
      "eval_g2l_cer": 100.4364,
      "eval_g2l_gen_len": 5.3498,
      "eval_g2l_rouge1": 23.9754,
      "eval_g2l_rouge2": 12.4646,
      "eval_g2l_rougeL": 23.7793,
      "eval_g2l_rougeLsum": 23.7651,
      "eval_l2ex_cer": 108.3181,
      "eval_l2ex_gen_len": 40.6639,
      "eval_l2ex_rouge1": 25.3285,
      "eval_l2ex_rouge2": 11.5065,
      "eval_l2ex_rougeL": 22.7839,
      "eval_l2ex_rougeLsum": 22.7439,
      "eval_l2g_cer": 97.287,
      "eval_l2g_gen_len": 23.9843,
      "eval_l2g_rouge1": 25.9551,
      "eval_l2g_rouge2": 11.8024,
      "eval_l2g_rougeL": 24.1441,
      "eval_l2g_rougeLsum": 24.1774,
      "eval_loss": 3.9933629035949707,
      "eval_runtime": 160.4997,
      "eval_samples_per_second": 61.813,
      "eval_steps_per_second": 0.972,
      "step": 415
    },
    {
      "epoch": 0.15023474178403756,
      "grad_norm": 14.013622283935547,
      "learning_rate": 0.00014026974951830443,
      "loss": 5.5983,
      "step": 416
    },
    {
      "epoch": 0.3004694835680751,
      "grad_norm": 47.43400192260742,
      "learning_rate": 0.00028053949903660886,
      "loss": 4.0514,
      "step": 832
    },
    {
      "epoch": 0.4507042253521127,
      "grad_norm": 87.22700500488281,
      "learning_rate": 0.0004208092485549133,
      "loss": 3.8797,
      "step": 1248
    },
    {
      "epoch": 0.6009389671361502,
      "grad_norm": 117.74351501464844,
      "learning_rate": 0.0005610789980732177,
      "loss": 3.7982,
      "step": 1664
    },
    {
      "epoch": 0.7500902853015529,
      "eval_g2l_cer": 52.9873,
      "eval_g2l_gen_len": 3.4046,
      "eval_g2l_rouge1": 36.7676,
      "eval_g2l_rouge2": 27.4065,
      "eval_g2l_rougeL": 36.653,
      "eval_g2l_rougeLsum": 36.6757,
      "eval_l2ex_cer": 82.3042,
      "eval_l2ex_gen_len": 18.221,
      "eval_l2ex_rouge1": 29.9987,
      "eval_l2ex_rouge2": 15.0081,
      "eval_l2ex_rougeL": 27.336,
      "eval_l2ex_rougeLsum": 27.3153,
      "eval_l2g_cer": 72.0808,
      "eval_l2g_gen_len": 10.2902,
      "eval_l2g_rouge1": 32.6068,
      "eval_l2g_rouge2": 20.974,
      "eval_l2g_rougeL": 31.1844,
      "eval_l2g_rougeLsum": 31.2287,
      "eval_loss": 3.5869500637054443,
      "eval_runtime": 137.2692,
      "eval_samples_per_second": 72.274,
      "eval_steps_per_second": 1.136,
      "step": 2077
    },
    {
      "epoch": 0.7511737089201878,
      "grad_norm": 132.14553833007812,
      "learning_rate": 0.0006999999822586204,
      "loss": 3.7522,
      "step": 2080
    },
    {
      "epoch": 0.9014084507042254,
      "grad_norm": 163.82469177246094,
      "learning_rate": 0.0006998044195197278,
      "loss": 3.7004,
      "step": 2496
    },
    {
      "epoch": 1.051643192488263,
      "grad_norm": 134.45631408691406,
      "learning_rate": 0.0006992253249390411,
      "loss": 3.6596,
      "step": 2912
    },
    {
      "epoch": 1.2018779342723005,
      "grad_norm": 110.51844787597656,
      "learning_rate": 0.0006982633339003477,
      "loss": 3.5916,
      "step": 3328
    },
    {
      "epoch": 1.352112676056338,
      "grad_norm": 119.85750579833984,
      "learning_rate": 0.000696919501902232,
      "loss": 3.5667,
      "step": 3744
    },
    {
      "epoch": 1.500180570603106,
      "eval_g2l_cer": 53.0855,
      "eval_g2l_gen_len": 4.9383,
      "eval_g2l_rouge1": 41.8458,
      "eval_g2l_rouge2": 30.8627,
      "eval_g2l_rougeL": 41.5679,
      "eval_g2l_rougeLsum": 41.5758,
      "eval_l2ex_cer": 85.4973,
      "eval_l2ex_gen_len": 30.0565,
      "eval_l2ex_rouge1": 30.6637,
      "eval_l2ex_rouge2": 14.7084,
      "eval_l2ex_rougeL": 27.5063,
      "eval_l2ex_rougeLsum": 27.5683,
      "eval_l2g_cer": 79.9636,
      "eval_l2g_gen_len": 23.344,
      "eval_l2g_rouge1": 38.4413,
      "eval_l2g_rouge2": 25.3548,
      "eval_l2g_rougeL": 36.2653,
      "eval_l2g_rougeLsum": 36.3346,
      "eval_loss": 3.4180662631988525,
      "eval_runtime": 157.8624,
      "eval_samples_per_second": 62.846,
      "eval_steps_per_second": 0.988,
      "step": 4154
    },
    {
      "epoch": 1.5023474178403755,
      "grad_norm": 130.5960235595703,
      "learning_rate": 0.000695195303399979,
      "loss": 3.5481,
      "step": 4160
    },
    {
      "epoch": 1.652582159624413,
      "grad_norm": 128.95892333984375,
      "learning_rate": 0.0006930926301877994,
      "loss": 3.5237,
      "step": 4576
    },
    {
      "epoch": 1.8028169014084507,
      "grad_norm": 129.52719116210938,
      "learning_rate": 0.000690613789323149,
      "loss": 3.5171,
      "step": 4992
    },
    {
      "epoch": 1.9530516431924883,
      "grad_norm": 122.73914337158203,
      "learning_rate": 0.0006877615005954206,
      "loss": 3.5051,
      "step": 5408
    },
    {
      "epoch": 2.103286384976526,
      "grad_norm": 100.69551849365234,
      "learning_rate": 0.0006845388935417848,
      "loss": 3.4086,
      "step": 5824
    },
    {
      "epoch": 2.2502708559046587,
      "eval_g2l_cer": 50.7779,
      "eval_g2l_gen_len": 4.8103,
      "eval_g2l_rouge1": 42.8195,
      "eval_g2l_rouge2": 31.8093,
      "eval_g2l_rougeL": 42.5575,
      "eval_g2l_rougeLsum": 42.5624,
      "eval_l2ex_cer": 83.3686,
      "eval_l2ex_gen_len": 29.3011,
      "eval_l2ex_rouge1": 31.5473,
      "eval_l2ex_rouge2": 15.1028,
      "eval_l2ex_rougeL": 28.0691,
      "eval_l2ex_rougeLsum": 28.1825,
      "eval_l2g_cer": 77.4412,
      "eval_l2g_gen_len": 21.7602,
      "eval_l2g_rouge1": 39.093,
      "eval_l2g_rouge2": 25.9346,
      "eval_l2g_rougeL": 37.0244,
      "eval_l2g_rougeLsum": 37.0876,
      "eval_loss": 3.3524162769317627,
      "eval_runtime": 157.4639,
      "eval_samples_per_second": 63.005,
      "eval_steps_per_second": 0.991,
      "step": 6231
    },
    {
      "epoch": 2.2535211267605635,
      "grad_norm": 101.63910675048828,
      "learning_rate": 0.0006809495040134555,
      "loss": 3.3874,
      "step": 6240
    },
    {
      "epoch": 2.403755868544601,
      "grad_norm": 99.6649398803711,
      "learning_rate": 0.000676997270296147,
      "loss": 3.3843,
      "step": 6656
    },
    {
      "epoch": 2.5539906103286385,
      "grad_norm": 102.08024597167969,
      "learning_rate": 0.0006726865287889783,
      "loss": 3.379,
      "step": 7072
    },
    {
      "epoch": 2.704225352112676,
      "grad_norm": 93.55581665039062,
      "learning_rate": 0.0006680220092465666,
      "loss": 3.3681,
      "step": 7488
    },
    {
      "epoch": 2.8544600938967135,
      "grad_norm": 95.19955444335938,
      "learning_rate": 0.0006630088295895313,
      "loss": 3.3544,
      "step": 7904
    },
    {
      "epoch": 3.0003611412062114,
      "eval_g2l_cer": 47.3806,
      "eval_g2l_gen_len": 3.8524,
      "eval_g2l_rouge1": 44.4831,
      "eval_g2l_rouge2": 34.1168,
      "eval_g2l_rougeL": 44.3015,
      "eval_g2l_rougeLsum": 44.3284,
      "eval_l2ex_cer": 85.008,
      "eval_l2ex_gen_len": 20.6814,
      "eval_l2ex_rouge1": 31.6888,
      "eval_l2ex_rouge2": 16.0577,
      "eval_l2ex_rougeL": 28.2507,
      "eval_l2ex_rougeLsum": 28.3238,
      "eval_l2g_cer": 75.025,
      "eval_l2g_gen_len": 15.1432,
      "eval_l2g_rouge1": 38.7981,
      "eval_l2g_rouge2": 26.3158,
      "eval_l2g_rougeL": 36.9181,
      "eval_l2g_rougeLsum": 36.9696,
      "eval_loss": 3.283109188079834,
      "eval_runtime": 151.9761,
      "eval_samples_per_second": 65.28,
      "eval_steps_per_second": 1.026,
      "step": 8308
    },
    {
      "epoch": 3.004694835680751,
      "grad_norm": 93.19324493408203,
      "learning_rate": 0.0006576524902890998,
      "loss": 3.3457,
      "step": 8320
    },
    {
      "epoch": 3.1549295774647885,
      "grad_norm": 104.09616088867188,
      "learning_rate": 0.0006519588683319802,
      "loss": 3.2525,
      "step": 8736
    },
    {
      "epoch": 3.3051643192488265,
      "grad_norm": 104.19088745117188,
      "learning_rate": 0.0006459342107721181,
      "loss": 3.2478,
      "step": 9152
    },
    {
      "epoch": 3.455399061032864,
      "grad_norm": 83.33271789550781,
      "learning_rate": 0.0006395851278764164,
      "loss": 3.2538,
      "step": 9568
    },
    {
      "epoch": 3.6056338028169015,
      "grad_norm": 97.30130767822266,
      "learning_rate": 0.0006329185858719365,
      "loss": 3.2493,
      "step": 9984
    },
    {
      "epoch": 3.7504514265077646,
      "eval_g2l_cer": 48.347,
      "eval_g2l_gen_len": 4.6986,
      "eval_g2l_rouge1": 45.126,
      "eval_g2l_rouge2": 34.0899,
      "eval_g2l_rougeL": 44.8883,
      "eval_g2l_rougeLsum": 44.8974,
      "eval_l2ex_cer": 83.8663,
      "eval_l2ex_gen_len": 27.5262,
      "eval_l2ex_rouge1": 32.7987,
      "eval_l2ex_rouge2": 15.8604,
      "eval_l2ex_rougeL": 28.7942,
      "eval_l2ex_rougeLsum": 28.8651,
      "eval_l2g_cer": 76.6028,
      "eval_l2g_gen_len": 21.6109,
      "eval_l2g_rouge1": 39.8458,
      "eval_l2g_rouge2": 26.7789,
      "eval_l2g_rougeL": 37.6794,
      "eval_l2g_rougeLsum": 37.7404,
      "eval_loss": 3.271592855453491,
      "eval_runtime": 157.264,
      "eval_samples_per_second": 63.085,
      "eval_steps_per_second": 0.992,
      "step": 10385
    },
    {
      "epoch": 3.755868544600939,
      "grad_norm": 87.18195343017578,
      "learning_rate": 0.0006259418993025392,
      "loss": 3.2387,
      "step": 10400
    },
    {
      "epoch": 3.9061032863849765,
      "grad_norm": 77.52426147460938,
      "learning_rate": 0.0006186627230033526,
      "loss": 3.2361,
      "step": 10816
    },
    {
      "epoch": 4.056338028169014,
      "grad_norm": 101.6712646484375,
      "learning_rate": 0.0006110890437018696,
      "loss": 3.1968,
      "step": 11232
    },
    {
      "epoch": 4.206572769953052,
      "grad_norm": 99.43510437011719,
      "learning_rate": 0.0006032291712548952,
      "loss": 3.1304,
      "step": 11648
    },
    {
      "epoch": 4.356807511737089,
      "grad_norm": 57.04045867919922,
      "learning_rate": 0.0005950917295309531,
      "loss": 3.1457,
      "step": 12064
    },
    {
      "epoch": 4.500541711809317,
      "eval_g2l_cer": 47.8444,
      "eval_g2l_gen_len": 4.6609,
      "eval_g2l_rouge1": 45.1043,
      "eval_g2l_rouge2": 34.1441,
      "eval_g2l_rougeL": 44.9302,
      "eval_g2l_rougeLsum": 44.9463,
      "eval_l2ex_cer": 82.1447,
      "eval_l2ex_gen_len": 26.408,
      "eval_l2ex_rouge1": 33.2278,
      "eval_l2ex_rouge2": 16.1616,
      "eval_l2ex_rougeL": 29.2879,
      "eval_l2ex_rougeLsum": 29.3229,
      "eval_l2g_cer": 76.1324,
      "eval_l2g_gen_len": 20.8059,
      "eval_l2g_rouge1": 40.2785,
      "eval_l2g_rouge2": 27.06,
      "eval_l2g_rougeL": 38.0678,
      "eval_l2g_rougeLsum": 38.1592,
      "eval_loss": 3.266465902328491,
      "eval_runtime": 157.9782,
      "eval_samples_per_second": 62.8,
      "eval_steps_per_second": 0.987,
      "step": 12462
    },
    {
      "epoch": 4.507042253521127,
      "grad_norm": 67.55484771728516,
      "learning_rate": 0.0005866856469481595,
      "loss": 3.145,
      "step": 12480
    },
    {
      "epoch": 4.657276995305164,
      "grad_norm": 76.89288330078125,
      "learning_rate": 0.0005780201466779439,
      "loss": 3.1236,
      "step": 12896
    },
    {
      "epoch": 4.807511737089202,
      "grad_norm": 87.32029724121094,
      "learning_rate": 0.0005691047365253672,
      "loss": 3.1392,
      "step": 13312
    },
    {
      "epoch": 4.957746478873239,
      "grad_norm": 109.48633575439453,
      "learning_rate": 0.0005599491984971357,
      "loss": 3.1438,
      "step": 13728
    },
    {
      "epoch": 5.107981220657277,
      "grad_norm": 66.15205383300781,
      "learning_rate": 0.0005505635780687648,
      "loss": 3.0644,
      "step": 14144
    },
    {
      "epoch": 5.250631997110871,
      "eval_g2l_cer": 46.2679,
      "eval_g2l_gen_len": 3.7853,
      "eval_g2l_rouge1": 45.7284,
      "eval_g2l_rouge2": 35.1094,
      "eval_g2l_rougeL": 45.5565,
      "eval_g2l_rougeLsum": 45.5772,
      "eval_l2ex_cer": 81.8173,
      "eval_l2ex_gen_len": 18.9137,
      "eval_l2ex_rouge1": 33.0179,
      "eval_l2ex_rouge2": 16.8664,
      "eval_l2ex_rougeL": 29.3623,
      "eval_l2ex_rougeLsum": 29.4514,
      "eval_l2g_cer": 70.134,
      "eval_l2g_gen_len": 14.2539,
      "eval_l2g_rouge1": 40.477,
      "eval_l2g_rouge2": 27.4563,
      "eval_l2g_rougeL": 38.398,
      "eval_l2g_rougeLsum": 38.4554,
      "eval_loss": 3.223316192626953,
      "eval_runtime": 147.2815,
      "eval_samples_per_second": 67.361,
      "eval_steps_per_second": 1.059,
      "step": 14539
    },
    {
      "epoch": 5.258215962441315,
      "grad_norm": 120.2764663696289,
      "learning_rate": 0.0005409581731626619,
      "loss": 3.0359,
      "step": 14560
    },
    {
      "epoch": 5.408450704225352,
      "grad_norm": 66.7710189819336,
      "learning_rate": 0.0005311435228492235,
      "loss": 3.0459,
      "step": 14976
    },
    {
      "epoch": 5.55868544600939,
      "grad_norm": 103.15011596679688,
      "learning_rate": 0.0005211303957833462,
      "loss": 3.0499,
      "step": 15392
    },
    {
      "epoch": 5.708920187793427,
      "grad_norm": 91.70545196533203,
      "learning_rate": 0.0005109297783890348,
      "loss": 3.0455,
      "step": 15808
    },
    {
      "epoch": 5.859154929577465,
      "grad_norm": 62.21952819824219,
      "learning_rate": 0.0005005528628050757,
      "loss": 3.0468,
      "step": 16224
    },
    {
      "epoch": 6.000722282412423,
      "eval_g2l_cer": 46.2085,
      "eval_g2l_gen_len": 4.1369,
      "eval_g2l_rouge1": 46.8001,
      "eval_g2l_rouge2": 35.9998,
      "eval_g2l_rougeL": 46.6066,
      "eval_g2l_rougeLsum": 46.6501,
      "eval_l2ex_cer": 79.8794,
      "eval_l2ex_gen_len": 18.741,
      "eval_l2ex_rouge1": 33.8832,
      "eval_l2ex_rouge2": 17.2827,
      "eval_l2ex_rougeL": 29.9003,
      "eval_l2ex_rougeLsum": 29.9671,
      "eval_l2g_cer": 70.1127,
      "eval_l2g_gen_len": 14.8234,
      "eval_l2g_rouge1": 40.9429,
      "eval_l2g_rouge2": 27.7443,
      "eval_l2g_rougeL": 38.7561,
      "eval_l2g_rougeLsum": 38.812,
      "eval_loss": 3.198634624481201,
      "eval_runtime": 147.3585,
      "eval_samples_per_second": 67.326,
      "eval_steps_per_second": 1.059,
      "step": 16616
    },
    {
      "epoch": 6.009389671361502,
      "grad_norm": 60.71475601196289,
      "learning_rate": 0.0004900110346049988,
      "loss": 3.0478,
      "step": 16640
    },
    {
      "epoch": 6.15962441314554,
      "grad_norm": 74.91094207763672,
      "learning_rate": 0.0004793158603048017,
      "loss": 2.9501,
      "step": 17056
    },
    {
      "epoch": 6.309859154929577,
      "grad_norm": 98.47882843017578,
      "learning_rate": 0.00046847907467214455,
      "loss": 2.9655,
      "step": 17472
    },
    {
      "epoch": 6.460093896713615,
      "grad_norm": 56.426727294921875,
      "learning_rate": 0.00045751256785093856,
      "loss": 2.9753,
      "step": 17888
    },
    {
      "epoch": 6.610328638497653,
      "grad_norm": 47.2915153503418,
      "learning_rate": 0.00044642837231545463,
      "loss": 2.9699,
      "step": 18304
    },
    {
      "epoch": 6.750812567713976,
      "eval_g2l_cer": 46.2359,
      "eval_g2l_gen_len": 4.4804,
      "eval_g2l_rouge1": 47.0405,
      "eval_g2l_rouge2": 35.9417,
      "eval_g2l_rougeL": 46.8609,
      "eval_g2l_rougeLsum": 46.8948,
      "eval_l2ex_cer": 81.7484,
      "eval_l2ex_gen_len": 23.0452,
      "eval_l2ex_rouge1": 33.3772,
      "eval_l2ex_rouge2": 16.3629,
      "eval_l2ex_rougeL": 29.2446,
      "eval_l2ex_rougeLsum": 29.3549,
      "eval_l2g_cer": 74.5769,
      "eval_l2g_gen_len": 19.6127,
      "eval_l2g_rouge1": 41.2652,
      "eval_l2g_rouge2": 27.8695,
      "eval_l2g_rougeL": 38.8879,
      "eval_l2g_rougeLsum": 38.9752,
      "eval_loss": 3.2092766761779785,
      "eval_runtime": 156.9963,
      "eval_samples_per_second": 63.193,
      "eval_steps_per_second": 0.994,
      "step": 18693
    },
    {
      "epoch": 6.76056338028169,
      "grad_norm": 40.62468338012695,
      "learning_rate": 0.0004352386496682677,
      "loss": 2.9756,
      "step": 18720
    },
    {
      "epoch": 6.910798122065728,
      "grad_norm": 44.248905181884766,
      "learning_rate": 0.00042395567729652095,
      "loss": 2.9631,
      "step": 19136
    },
    {
      "epoch": 7.061032863849765,
      "grad_norm": 71.54096984863281,
      "learning_rate": 0.00041259183490115084,
      "loss": 2.9346,
      "step": 19552
    },
    {
      "epoch": 7.211267605633803,
      "grad_norm": 63.520111083984375,
      "learning_rate": 0.0004011595909138531,
      "loss": 2.8878,
      "step": 19968
    },
    {
      "epoch": 7.36150234741784,
      "grad_norm": 52.090240478515625,
      "learning_rate": 0.0003896714888166938,
      "loss": 2.8877,
      "step": 20384
    },
    {
      "epoch": 7.500902853015529,
      "eval_g2l_cer": 45.7698,
      "eval_g2l_gen_len": 4.4545,
      "eval_g2l_rouge1": 47.583,
      "eval_g2l_rouge2": 36.2797,
      "eval_g2l_rougeL": 47.433,
      "eval_g2l_rougeLsum": 47.4196,
      "eval_l2ex_cer": 81.5805,
      "eval_l2ex_gen_len": 23.9979,
      "eval_l2ex_rouge1": 33.9624,
      "eval_l2ex_rouge2": 16.6357,
      "eval_l2ex_rougeL": 29.8228,
      "eval_l2ex_rougeLsum": 29.9123,
      "eval_l2g_cer": 73.7736,
      "eval_l2g_gen_len": 19.3481,
      "eval_l2g_rouge1": 41.2698,
      "eval_l2g_rouge2": 27.863,
      "eval_l2g_rougeL": 38.869,
      "eval_l2g_rougeLsum": 38.9562,
      "eval_loss": 3.214210033416748,
      "eval_runtime": 157.1314,
      "eval_samples_per_second": 63.138,
      "eval_steps_per_second": 0.993,
      "step": 20770
    },
    {
      "epoch": 7.511737089201878,
      "grad_norm": 44.08253479003906,
      "learning_rate": 0.0003781401333793737,
      "loss": 2.888,
      "step": 20800
    },
    {
      "epoch": 7.661971830985916,
      "grad_norm": 50.90373611450195,
      "learning_rate": 0.00036657817682924954,
      "loss": 2.9017,
      "step": 21216
    },
    {
      "epoch": 7.812206572769953,
      "grad_norm": 70.20822143554688,
      "learning_rate": 0.0003549983049692832,
      "loss": 2.8951,
      "step": 21632
    },
    {
      "epoch": 7.962441314553991,
      "grad_norm": 24.146835327148438,
      "learning_rate": 0.00034341322325915155,
      "loss": 2.9011,
      "step": 22048
    },
    {
      "epoch": 8.112676056338028,
      "grad_norm": 43.61560821533203,
      "learning_rate": 0.0003318356428747905,
      "loss": 2.8542,
      "step": 22464
    },
    {
      "epoch": 8.250993138317082,
      "eval_g2l_cer": 45.6053,
      "eval_g2l_gen_len": 3.9989,
      "eval_g2l_rouge1": 47.3053,
      "eval_g2l_rouge2": 36.3191,
      "eval_g2l_rougeL": 47.2008,
      "eval_g2l_rougeLsum": 47.2596,
      "eval_l2ex_cer": 81.0598,
      "eval_l2ex_gen_len": 20.2055,
      "eval_l2ex_rouge1": 33.9028,
      "eval_l2ex_rouge2": 16.9085,
      "eval_l2ex_rougeL": 29.7011,
      "eval_l2ex_rougeLsum": 29.7791,
      "eval_l2g_cer": 71.3827,
      "eval_l2g_gen_len": 15.4007,
      "eval_l2g_rouge1": 41.2614,
      "eval_l2g_rouge2": 28.0203,
      "eval_l2g_rougeL": 39.0253,
      "eval_l2g_rougeLsum": 39.0954,
      "eval_loss": 3.200995922088623,
      "eval_runtime": 148.5568,
      "eval_samples_per_second": 66.783,
      "eval_steps_per_second": 1.05,
      "step": 22847
    },
    {
      "epoch": 8.262910798122066,
      "grad_norm": 34.18248748779297,
      "learning_rate": 0.0003202782667616653,
      "loss": 2.8359,
      "step": 22880
    },
    {
      "epoch": 8.413145539906104,
      "grad_norm": 46.180694580078125,
      "learning_rate": 0.00030875377569707275,
      "loss": 2.8378,
      "step": 23296
    },
    {
      "epoch": 8.56338028169014,
      "grad_norm": 42.6771354675293,
      "learning_rate": 0.000297274814376765,
      "loss": 2.8327,
      "step": 23712
    },
    {
      "epoch": 8.713615023474178,
      "grad_norm": 19.82225227355957,
      "learning_rate": 0.0002858539775411638,
      "loss": 2.843,
      "step": 24128
    },
    {
      "epoch": 8.863849765258216,
      "grad_norm": 19.00598907470703,
      "learning_rate": 0.00027450379615638425,
      "loss": 2.8231,
      "step": 24544
    },
    {
      "epoch": 9.001083423618635,
      "eval_g2l_cer": 45.4637,
      "eval_g2l_gen_len": 4.0383,
      "eval_g2l_rouge1": 47.3272,
      "eval_g2l_rouge2": 36.2943,
      "eval_g2l_rougeL": 47.1452,
      "eval_g2l_rougeLsum": 47.2213,
      "eval_l2ex_cer": 80.295,
      "eval_l2ex_gen_len": 19.4255,
      "eval_l2ex_rouge1": 34.2944,
      "eval_l2ex_rouge2": 17.3206,
      "eval_l2ex_rougeL": 30.0912,
      "eval_l2ex_rougeLsum": 30.1909,
      "eval_l2g_cer": 69.8778,
      "eval_l2g_gen_len": 14.8225,
      "eval_l2g_rouge1": 41.2308,
      "eval_l2g_rouge2": 27.9775,
      "eval_l2g_rougeL": 39.0866,
      "eval_l2g_rougeLsum": 39.1425,
      "eval_loss": 3.2024292945861816,
      "eval_runtime": 146.3445,
      "eval_samples_per_second": 67.792,
      "eval_steps_per_second": 1.066,
      "step": 24924
    },
    {
      "epoch": 9.014084507042254,
      "grad_norm": 22.52915382385254,
      "learning_rate": 0.00026323672366523274,
      "loss": 2.8413,
      "step": 24960
    },
    {
      "epoch": 9.164319248826292,
      "grad_norm": 30.310894012451172,
      "learning_rate": 0.0002520651223232633,
      "loss": 2.7777,
      "step": 25376
    },
    {
      "epoch": 9.314553990610328,
      "grad_norm": 31.759746551513672,
      "learning_rate": 0.000241001249634885,
      "loss": 2.7963,
      "step": 25792
    },
    {
      "epoch": 9.464788732394366,
      "grad_norm": 13.962241172790527,
      "learning_rate": 0.00023005724490440243,
      "loss": 2.7968,
      "step": 26208
    },
    {
      "epoch": 9.615023474178404,
      "grad_norm": 20.10390853881836,
      "learning_rate": 0.0002192451159167455,
      "loss": 2.7902,
      "step": 26624
    },
    {
      "epoch": 9.751173708920188,
      "eval_g2l_cer": 45.3083,
      "eval_g2l_gen_len": 4.3145,
      "eval_g2l_rouge1": 47.8559,
      "eval_g2l_rouge2": 36.8346,
      "eval_g2l_rougeL": 47.6856,
      "eval_g2l_rougeLsum": 47.712,
      "eval_l2ex_cer": 80.353,
      "eval_l2ex_gen_len": 21.4861,
      "eval_l2ex_rouge1": 34.306,
      "eval_l2ex_rouge2": 16.8625,
      "eval_l2ex_rougeL": 29.7995,
      "eval_l2ex_rougeLsum": 29.9518,
      "eval_l2g_cer": 71.8262,
      "eval_l2g_gen_len": 17.528,
      "eval_l2g_rouge1": 41.8619,
      "eval_l2g_rouge2": 28.3549,
      "eval_l2g_rougeL": 39.5168,
      "eval_l2g_rougeLsum": 39.5953,
      "eval_loss": 3.1981074810028076,
      "eval_runtime": 153.7484,
      "eval_samples_per_second": 64.527,
      "eval_steps_per_second": 1.015,
      "step": 27001
    }
  ],
  "logging_steps": 416,
  "max_steps": 41535,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 2077,
  "total_flos": 2.9580064377012224e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}