{
  "best_global_step": 100,
  "best_metric": 0.5112941265106201,
  "best_model_checkpoint": "/root/autodl-tmp/model/lora-textui/stage0_caption-v2/checkpoint-100",
  "epoch": 3.458149779735683,
  "eval_steps": 10,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03524229074889868,
      "grad_norm": 1.2463780641555786,
      "learning_rate": 0.0,
      "loss": 1.9336,
      "step": 1
    },
    {
      "epoch": 0.07048458149779736,
      "grad_norm": 1.3275448083877563,
      "learning_rate": 1e-05,
      "loss": 2.048,
      "step": 2
    },
    {
      "epoch": 0.10572687224669604,
      "grad_norm": 1.2619495391845703,
      "learning_rate": 2e-05,
      "loss": 1.9311,
      "step": 3
    },
    {
      "epoch": 0.14096916299559473,
      "grad_norm": 1.2719473838806152,
      "learning_rate": 3e-05,
      "loss": 1.9814,
      "step": 4
    },
    {
      "epoch": 0.1762114537444934,
      "grad_norm": 1.3939034938812256,
      "learning_rate": 4e-05,
      "loss": 1.9955,
      "step": 5
    },
    {
      "epoch": 0.21145374449339208,
      "grad_norm": 1.3475849628448486,
      "learning_rate": 5e-05,
      "loss": 1.874,
      "step": 6
    },
    {
      "epoch": 0.24669603524229075,
      "grad_norm": 1.2599456310272217,
      "learning_rate": 6e-05,
      "loss": 1.7726,
      "step": 7
    },
    {
      "epoch": 0.28193832599118945,
      "grad_norm": 1.1231255531311035,
      "learning_rate": 7e-05,
      "loss": 1.6056,
      "step": 8
    },
    {
      "epoch": 0.31718061674008813,
      "grad_norm": 0.926201581954956,
      "learning_rate": 8e-05,
      "loss": 1.4883,
      "step": 9
    },
    {
      "epoch": 0.3524229074889868,
      "grad_norm": 0.8478774428367615,
      "learning_rate": 9e-05,
      "loss": 1.3543,
      "step": 10
    },
    {
      "epoch": 0.3524229074889868,
      "eval_loss": 1.2461069822311401,
      "eval_runtime": 25.3282,
      "eval_samples_per_second": 7.975,
      "eval_steps_per_second": 0.513,
      "step": 10
    },
    {
      "epoch": 0.3876651982378855,
      "grad_norm": 0.8932197093963623,
      "learning_rate": 0.0001,
      "loss": 1.253,
      "step": 11
    },
    {
      "epoch": 0.42290748898678415,
      "grad_norm": 0.9408664703369141,
      "learning_rate": 0.00011000000000000002,
      "loss": 1.1504,
      "step": 12
    },
    {
      "epoch": 0.4581497797356828,
      "grad_norm": 0.7685560584068298,
      "learning_rate": 0.00012,
      "loss": 1.1904,
      "step": 13
    },
    {
      "epoch": 0.4933920704845815,
      "grad_norm": 1.0110777616500854,
      "learning_rate": 0.00013000000000000002,
      "loss": 1.065,
      "step": 14
    },
    {
      "epoch": 0.5286343612334802,
      "grad_norm": 1.0698988437652588,
      "learning_rate": 0.00014,
      "loss": 1.0323,
      "step": 15
    },
    {
      "epoch": 0.5638766519823789,
      "grad_norm": 0.9592034220695496,
      "learning_rate": 0.00015000000000000001,
      "loss": 0.967,
      "step": 16
    },
    {
      "epoch": 0.5991189427312775,
      "grad_norm": 0.515692412853241,
      "learning_rate": 0.00016,
      "loss": 0.9886,
      "step": 17
    },
    {
      "epoch": 0.6343612334801763,
      "grad_norm": 0.4879838824272156,
      "learning_rate": 0.00017,
      "loss": 0.9052,
      "step": 18
    },
    {
      "epoch": 0.6696035242290749,
      "grad_norm": 0.5619950294494629,
      "learning_rate": 0.00018,
      "loss": 0.8916,
      "step": 19
    },
    {
      "epoch": 0.7048458149779736,
      "grad_norm": 0.5306786298751831,
      "learning_rate": 0.00019,
      "loss": 0.8508,
      "step": 20
    },
    {
      "epoch": 0.7048458149779736,
      "eval_loss": 0.8256819844245911,
      "eval_runtime": 25.3172,
      "eval_samples_per_second": 7.979,
      "eval_steps_per_second": 0.513,
      "step": 20
    },
    {
      "epoch": 0.7400881057268722,
      "grad_norm": 0.45645713806152344,
      "learning_rate": 0.0002,
      "loss": 0.837,
      "step": 21
    },
    {
      "epoch": 0.775330396475771,
      "grad_norm": 0.3715068995952606,
      "learning_rate": 0.0001999229036240723,
      "loss": 0.8872,
      "step": 22
    },
    {
      "epoch": 0.8105726872246696,
      "grad_norm": 0.3798660337924957,
      "learning_rate": 0.0001996917333733128,
      "loss": 0.8076,
      "step": 23
    },
    {
      "epoch": 0.8458149779735683,
      "grad_norm": 0.4133865237236023,
      "learning_rate": 0.00019930684569549264,
      "loss": 0.8181,
      "step": 24
    },
    {
      "epoch": 0.8810572687224669,
      "grad_norm": 0.3884609043598175,
      "learning_rate": 0.00019876883405951377,
      "loss": 0.8081,
      "step": 25
    },
    {
      "epoch": 0.9162995594713657,
      "grad_norm": 0.3344547748565674,
      "learning_rate": 0.00019807852804032305,
      "loss": 0.7206,
      "step": 26
    },
    {
      "epoch": 0.9515418502202643,
      "grad_norm": 0.2715914845466614,
      "learning_rate": 0.00019723699203976766,
      "loss": 0.7821,
      "step": 27
    },
    {
      "epoch": 0.986784140969163,
      "grad_norm": 0.3191257417201996,
      "learning_rate": 0.00019624552364536473,
      "loss": 0.7356,
      "step": 28
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.43808606266975403,
      "learning_rate": 0.00019510565162951537,
      "loss": 0.7774,
      "step": 29
    },
    {
      "epoch": 1.0352422907488987,
      "grad_norm": 0.32706010341644287,
      "learning_rate": 0.00019381913359224842,
      "loss": 0.7069,
      "step": 30
    },
    {
      "epoch": 1.0352422907488987,
      "eval_loss": 0.6948859691619873,
      "eval_runtime": 25.3163,
      "eval_samples_per_second": 7.979,
      "eval_steps_per_second": 0.514,
      "step": 30
    },
    {
      "epoch": 1.0704845814977975,
      "grad_norm": 0.2936725616455078,
      "learning_rate": 0.0001923879532511287,
      "loss": 0.7228,
      "step": 31
    },
    {
      "epoch": 1.105726872246696,
      "grad_norm": 0.3039261996746063,
      "learning_rate": 0.00019081431738250814,
      "loss": 0.6692,
      "step": 32
    },
    {
      "epoch": 1.1409691629955947,
      "grad_norm": 0.29901978373527527,
      "learning_rate": 0.0001891006524188368,
      "loss": 0.6742,
      "step": 33
    },
    {
      "epoch": 1.1762114537444934,
      "grad_norm": 0.32121536135673523,
      "learning_rate": 0.00018724960070727972,
      "loss": 0.7012,
      "step": 34
    },
    {
      "epoch": 1.2114537444933922,
      "grad_norm": 0.3143290579319,
      "learning_rate": 0.00018526401643540922,
      "loss": 0.6594,
      "step": 35
    },
    {
      "epoch": 1.2466960352422907,
      "grad_norm": 0.3040238320827484,
      "learning_rate": 0.00018314696123025454,
      "loss": 0.6723,
      "step": 36
    },
    {
      "epoch": 1.2819383259911894,
      "grad_norm": 0.2974918782711029,
      "learning_rate": 0.00018090169943749476,
      "loss": 0.6514,
      "step": 37
    },
    {
      "epoch": 1.3171806167400881,
      "grad_norm": 0.2963806390762329,
      "learning_rate": 0.00017853169308807448,
      "loss": 0.6585,
      "step": 38
    },
    {
      "epoch": 1.3524229074889869,
      "grad_norm": 0.2951265871524811,
      "learning_rate": 0.0001760405965600031,
      "loss": 0.6339,
      "step": 39
    },
    {
      "epoch": 1.3876651982378854,
      "grad_norm": 0.3203594386577606,
      "learning_rate": 0.00017343225094356855,
      "loss": 0.655,
      "step": 40
    },
    {
      "epoch": 1.3876651982378854,
      "eval_loss": 0.6199031472206116,
      "eval_runtime": 25.3133,
      "eval_samples_per_second": 7.98,
      "eval_steps_per_second": 0.514,
      "step": 40
    },
    {
      "epoch": 1.422907488986784,
      "grad_norm": 0.29003453254699707,
      "learning_rate": 0.00017071067811865476,
      "loss": 0.6334,
      "step": 41
    },
    {
      "epoch": 1.4581497797356828,
      "grad_norm": 0.29403063654899597,
      "learning_rate": 0.0001678800745532942,
      "loss": 0.644,
      "step": 42
    },
    {
      "epoch": 1.4933920704845816,
      "grad_norm": 0.30763620138168335,
      "learning_rate": 0.00016494480483301836,
      "loss": 0.5925,
      "step": 43
    },
    {
      "epoch": 1.5286343612334803,
      "grad_norm": 0.3372487425804138,
      "learning_rate": 0.00016190939493098344,
      "loss": 0.6486,
      "step": 44
    },
    {
      "epoch": 1.5638766519823788,
      "grad_norm": 0.2898256480693817,
      "learning_rate": 0.00015877852522924732,
      "loss": 0.6039,
      "step": 45
    },
    {
      "epoch": 1.5991189427312775,
      "grad_norm": 0.2907792627811432,
      "learning_rate": 0.00015555702330196023,
      "loss": 0.6244,
      "step": 46
    },
    {
      "epoch": 1.6343612334801763,
      "grad_norm": 0.3023645579814911,
      "learning_rate": 0.0001522498564715949,
      "loss": 0.5672,
      "step": 47
    },
    {
      "epoch": 1.6696035242290748,
      "grad_norm": 0.3130245804786682,
      "learning_rate": 0.00014886212414969553,
      "loss": 0.6356,
      "step": 48
    },
    {
      "epoch": 1.7048458149779737,
      "grad_norm": 0.3192823529243469,
      "learning_rate": 0.00014539904997395468,
      "loss": 0.6105,
      "step": 49
    },
    {
      "epoch": 1.7400881057268722,
      "grad_norm": 0.3187403082847595,
      "learning_rate": 0.0001418659737537428,
      "loss": 0.6023,
      "step": 50
    },
    {
      "epoch": 1.7400881057268722,
      "eval_loss": 0.5753047466278076,
      "eval_runtime": 25.3636,
      "eval_samples_per_second": 7.964,
      "eval_steps_per_second": 0.513,
      "step": 50
    },
    {
      "epoch": 1.775330396475771,
      "grad_norm": 0.3186057507991791,
      "learning_rate": 0.000138268343236509,
      "loss": 0.6216,
      "step": 51
    },
    {
      "epoch": 1.8105726872246697,
      "grad_norm": 0.3059786856174469,
      "learning_rate": 0.0001346117057077493,
      "loss": 0.6124,
      "step": 52
    },
    {
      "epoch": 1.8458149779735682,
      "grad_norm": 0.2928261160850525,
      "learning_rate": 0.00013090169943749476,
      "loss": 0.5445,
      "step": 53
    },
    {
      "epoch": 1.881057268722467,
      "grad_norm": 0.2955411374568939,
      "learning_rate": 0.00012714404498650743,
      "loss": 0.5508,
      "step": 54
    },
    {
      "epoch": 1.9162995594713657,
      "grad_norm": 0.30531296133995056,
      "learning_rate": 0.00012334453638559057,
      "loss": 0.5522,
      "step": 55
    },
    {
      "epoch": 1.9515418502202642,
      "grad_norm": 0.3257448375225067,
      "learning_rate": 0.00011950903220161285,
      "loss": 0.5569,
      "step": 56
    },
    {
      "epoch": 1.9867841409691631,
      "grad_norm": 0.29654088616371155,
      "learning_rate": 0.0001156434465040231,
      "loss": 0.5307,
      "step": 57
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.46664145588874817,
      "learning_rate": 0.00011175373974578378,
      "loss": 0.6131,
      "step": 58
    },
    {
      "epoch": 2.0352422907488985,
      "grad_norm": 0.3071623742580414,
      "learning_rate": 0.0001078459095727845,
      "loss": 0.4965,
      "step": 59
    },
    {
      "epoch": 2.0704845814977975,
      "grad_norm": 0.3435165286064148,
      "learning_rate": 0.00010392598157590688,
      "loss": 0.5086,
      "step": 60
    },
    {
      "epoch": 2.0704845814977975,
      "eval_loss": 0.5449034571647644,
      "eval_runtime": 25.365,
      "eval_samples_per_second": 7.964,
      "eval_steps_per_second": 0.513,
      "step": 60
    },
    {
      "epoch": 2.105726872246696,
      "grad_norm": 0.3053547739982605,
      "learning_rate": 0.0001,
      "loss": 0.5208,
      "step": 61
    },
    {
      "epoch": 2.140969162995595,
      "grad_norm": 0.3197723627090454,
      "learning_rate": 9.607401842409317e-05,
      "loss": 0.5298,
      "step": 62
    },
    {
      "epoch": 2.1762114537444934,
      "grad_norm": 0.31069523096084595,
      "learning_rate": 9.215409042721552e-05,
      "loss": 0.5025,
      "step": 63
    },
    {
      "epoch": 2.211453744493392,
      "grad_norm": 0.3025587797164917,
      "learning_rate": 8.824626025421626e-05,
      "loss": 0.4856,
      "step": 64
    },
    {
      "epoch": 2.246696035242291,
      "grad_norm": 0.32704848051071167,
      "learning_rate": 8.435655349597689e-05,
      "loss": 0.5118,
      "step": 65
    },
    {
      "epoch": 2.2819383259911894,
      "grad_norm": 0.3421045243740082,
      "learning_rate": 8.049096779838719e-05,
      "loss": 0.5256,
      "step": 66
    },
    {
      "epoch": 2.317180616740088,
      "grad_norm": 0.3493560552597046,
      "learning_rate": 7.66554636144095e-05,
      "loss": 0.5396,
      "step": 67
    },
    {
      "epoch": 2.352422907488987,
      "grad_norm": 0.34267452359199524,
      "learning_rate": 7.285595501349258e-05,
      "loss": 0.5633,
      "step": 68
    },
    {
      "epoch": 2.3876651982378854,
      "grad_norm": 0.34627947211265564,
      "learning_rate": 6.909830056250527e-05,
      "loss": 0.5233,
      "step": 69
    },
    {
      "epoch": 2.4229074889867843,
      "grad_norm": 0.3640250265598297,
      "learning_rate": 6.538829429225069e-05,
      "loss": 0.5638,
      "step": 70
    },
    {
      "epoch": 2.4229074889867843,
      "eval_loss": 0.5274443030357361,
      "eval_runtime": 25.3671,
      "eval_samples_per_second": 7.963,
      "eval_steps_per_second": 0.512,
      "step": 70
    },
    {
      "epoch": 2.458149779735683,
      "grad_norm": 0.3336215913295746,
      "learning_rate": 6.173165676349103e-05,
      "loss": 0.5003,
      "step": 71
    },
    {
      "epoch": 2.4933920704845813,
      "grad_norm": 0.3200637102127075,
      "learning_rate": 5.8134026246257225e-05,
      "loss": 0.4974,
      "step": 72
    },
    {
      "epoch": 2.5286343612334803,
      "grad_norm": 0.3232054114341736,
      "learning_rate": 5.4600950026045326e-05,
      "loss": 0.5325,
      "step": 73
    },
    {
      "epoch": 2.563876651982379,
      "grad_norm": 0.327750027179718,
      "learning_rate": 5.113787585030454e-05,
      "loss": 0.4323,
      "step": 74
    },
    {
      "epoch": 2.5991189427312777,
      "grad_norm": 0.325334757566452,
      "learning_rate": 4.7750143528405126e-05,
      "loss": 0.4905,
      "step": 75
    },
    {
      "epoch": 2.6343612334801763,
      "grad_norm": 0.3405950665473938,
      "learning_rate": 4.444297669803981e-05,
      "loss": 0.4649,
      "step": 76
    },
    {
      "epoch": 2.6696035242290748,
      "grad_norm": 0.35329973697662354,
      "learning_rate": 4.12214747707527e-05,
      "loss": 0.5379,
      "step": 77
    },
    {
      "epoch": 2.7048458149779737,
      "grad_norm": 0.34212884306907654,
      "learning_rate": 3.8090605069016595e-05,
      "loss": 0.4885,
      "step": 78
    },
    {
      "epoch": 2.7400881057268722,
      "grad_norm": 0.33281591534614563,
      "learning_rate": 3.5055195166981645e-05,
      "loss": 0.4923,
      "step": 79
    },
    {
      "epoch": 2.7753303964757707,
      "grad_norm": 0.32896408438682556,
      "learning_rate": 3.211992544670582e-05,
      "loss": 0.4747,
      "step": 80
    },
    {
      "epoch": 2.7753303964757707,
      "eval_loss": 0.5167238712310791,
      "eval_runtime": 25.36,
      "eval_samples_per_second": 7.965,
      "eval_steps_per_second": 0.513,
      "step": 80
    },
    {
      "epoch": 2.8105726872246697,
      "grad_norm": 0.34730497002601624,
      "learning_rate": 2.9289321881345254e-05,
      "loss": 0.4887,
      "step": 81
    },
    {
      "epoch": 2.845814977973568,
      "grad_norm": 0.35023897886276245,
      "learning_rate": 2.6567749056431467e-05,
      "loss": 0.4812,
      "step": 82
    },
    {
      "epoch": 2.8810572687224667,
      "grad_norm": 0.34714555740356445,
      "learning_rate": 2.3959403439996907e-05,
      "loss": 0.4947,
      "step": 83
    },
    {
      "epoch": 2.9162995594713657,
      "grad_norm": 0.3485592007637024,
      "learning_rate": 2.146830691192553e-05,
      "loss": 0.5061,
      "step": 84
    },
    {
      "epoch": 2.951541850220264,
      "grad_norm": 0.36132875084877014,
      "learning_rate": 1.9098300562505266e-05,
      "loss": 0.4976,
      "step": 85
    },
    {
      "epoch": 2.986784140969163,
      "grad_norm": 0.3450870215892792,
      "learning_rate": 1.6853038769745467e-05,
      "loss": 0.459,
      "step": 86
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.5119443535804749,
      "learning_rate": 1.4735983564590783e-05,
      "loss": 0.4275,
      "step": 87
    },
    {
      "epoch": 3.0352422907488985,
      "grad_norm": 0.325406551361084,
      "learning_rate": 1.2750399292720283e-05,
      "loss": 0.4549,
      "step": 88
    },
    {
      "epoch": 3.0704845814977975,
      "grad_norm": 0.33169472217559814,
      "learning_rate": 1.0899347581163221e-05,
      "loss": 0.4918,
      "step": 89
    },
    {
      "epoch": 3.105726872246696,
      "grad_norm": 0.3156157433986664,
      "learning_rate": 9.185682617491863e-06,
      "loss": 0.458,
      "step": 90
    },
    {
      "epoch": 3.105726872246696,
      "eval_loss": 0.5118579268455505,
      "eval_runtime": 25.3571,
      "eval_samples_per_second": 7.966,
      "eval_steps_per_second": 0.513,
      "step": 90
    },
    {
      "epoch": 3.140969162995595,
      "grad_norm": 0.34081166982650757,
      "learning_rate": 7.612046748871327e-06,
      "loss": 0.4861,
      "step": 91
    },
    {
      "epoch": 3.1762114537444934,
      "grad_norm": 0.32250353693962097,
      "learning_rate": 6.180866407751595e-06,
      "loss": 0.4881,
      "step": 92
    },
    {
      "epoch": 3.211453744493392,
      "grad_norm": 0.3292059600353241,
      "learning_rate": 4.8943483704846475e-06,
      "loss": 0.4524,
      "step": 93
    },
    {
      "epoch": 3.246696035242291,
      "grad_norm": 0.3330952525138855,
      "learning_rate": 3.7544763546352834e-06,
      "loss": 0.4678,
      "step": 94
    },
    {
      "epoch": 3.2819383259911894,
      "grad_norm": 0.3274575173854828,
      "learning_rate": 2.7630079602323442e-06,
      "loss": 0.5017,
      "step": 95
    },
    {
      "epoch": 3.317180616740088,
      "grad_norm": 0.3240121006965637,
      "learning_rate": 1.921471959676957e-06,
      "loss": 0.4615,
      "step": 96
    },
    {
      "epoch": 3.352422907488987,
      "grad_norm": 0.34071552753448486,
      "learning_rate": 1.231165940486234e-06,
      "loss": 0.4972,
      "step": 97
    },
    {
      "epoch": 3.3876651982378854,
      "grad_norm": 0.3293885588645935,
      "learning_rate": 6.931543045073708e-07,
      "loss": 0.4188,
      "step": 98
    },
    {
      "epoch": 3.4229074889867843,
      "grad_norm": 0.32528501749038696,
      "learning_rate": 3.0826662668720364e-07,
      "loss": 0.4577,
      "step": 99
    },
    {
      "epoch": 3.458149779735683,
      "grad_norm": 0.35076647996902466,
      "learning_rate": 7.709637592770991e-08,
      "loss": 0.4505,
      "step": 100
    },
    {
      "epoch": 3.458149779735683,
      "eval_loss": 0.5112941265106201,
      "eval_runtime": 25.3907,
      "eval_samples_per_second": 7.956,
      "eval_steps_per_second": 0.512,
      "step": 100
    }
  ],
  "logging_steps": 1,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 30,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.237513300119388e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}