tmnam20's picture
Training in progress, step 111000, checkpoint
9a02b62
{
"best_metric": 0.9822403788566589,
"best_model_checkpoint": "/kaggle/output/checkpoint-111000",
"epoch": 4.522490221642764,
"eval_steps": 1000,
"global_step": 111000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 2.7777777777777777e-11,
"loss": 1.1078,
"step": 1
},
{
"epoch": 0.04,
"learning_rate": 2.7750000000000004e-08,
"loss": 1.128,
"step": 1000
},
{
"epoch": 0.04,
"eval_accuracy": 0.330938123752495,
"eval_loss": 1.1083118915557861,
"eval_runtime": 12.3979,
"eval_samples_per_second": 404.1,
"eval_steps_per_second": 50.573,
"step": 1000
},
{
"epoch": 0.08,
"learning_rate": 5.5527777777777784e-08,
"loss": 1.1164,
"step": 2000
},
{
"epoch": 0.08,
"eval_accuracy": 0.3323353293413174,
"eval_loss": 1.102672815322876,
"eval_runtime": 12.2861,
"eval_samples_per_second": 407.778,
"eval_steps_per_second": 51.033,
"step": 2000
},
{
"epoch": 0.12,
"learning_rate": 8.327777777777778e-08,
"loss": 1.1158,
"step": 3000
},
{
"epoch": 0.12,
"eval_accuracy": 0.34211576846307384,
"eval_loss": 1.1024774312973022,
"eval_runtime": 12.3746,
"eval_samples_per_second": 404.863,
"eval_steps_per_second": 50.668,
"step": 3000
},
{
"epoch": 0.16,
"learning_rate": 1.1105555555555557e-07,
"loss": 1.1115,
"step": 4000
},
{
"epoch": 0.16,
"eval_accuracy": 0.34890219560878244,
"eval_loss": 1.1012382507324219,
"eval_runtime": 12.5405,
"eval_samples_per_second": 399.505,
"eval_steps_per_second": 49.998,
"step": 4000
},
{
"epoch": 0.2,
"learning_rate": 1.3880555555555558e-07,
"loss": 1.1141,
"step": 5000
},
{
"epoch": 0.2,
"eval_accuracy": 0.3534930139720559,
"eval_loss": 1.1002192497253418,
"eval_runtime": 12.3001,
"eval_samples_per_second": 407.314,
"eval_steps_per_second": 50.975,
"step": 5000
},
{
"epoch": 0.24,
"learning_rate": 1.6658333333333335e-07,
"loss": 1.112,
"step": 6000
},
{
"epoch": 0.24,
"eval_accuracy": 0.3524950099800399,
"eval_loss": 1.100056767463684,
"eval_runtime": 12.4062,
"eval_samples_per_second": 403.831,
"eval_steps_per_second": 50.539,
"step": 6000
},
{
"epoch": 0.29,
"learning_rate": 1.9433333333333334e-07,
"loss": 1.1115,
"step": 7000
},
{
"epoch": 0.29,
"eval_accuracy": 0.34331337325349304,
"eval_loss": 1.098986029624939,
"eval_runtime": 12.4456,
"eval_samples_per_second": 402.551,
"eval_steps_per_second": 50.379,
"step": 7000
},
{
"epoch": 0.33,
"learning_rate": 2.2211111111111114e-07,
"loss": 1.109,
"step": 8000
},
{
"epoch": 0.33,
"eval_accuracy": 0.3347305389221557,
"eval_loss": 1.098678469657898,
"eval_runtime": 12.5978,
"eval_samples_per_second": 397.688,
"eval_steps_per_second": 49.771,
"step": 8000
},
{
"epoch": 0.37,
"learning_rate": 2.4986111111111113e-07,
"loss": 1.1117,
"step": 9000
},
{
"epoch": 0.37,
"eval_accuracy": 0.34311377245508984,
"eval_loss": 1.0984419584274292,
"eval_runtime": 12.313,
"eval_samples_per_second": 406.886,
"eval_steps_per_second": 50.922,
"step": 9000
},
{
"epoch": 0.41,
"learning_rate": 2.776388888888889e-07,
"loss": 1.1081,
"step": 10000
},
{
"epoch": 0.41,
"eval_accuracy": 0.35429141716566864,
"eval_loss": 1.0968743562698364,
"eval_runtime": 12.4088,
"eval_samples_per_second": 403.744,
"eval_steps_per_second": 50.528,
"step": 10000
},
{
"epoch": 0.45,
"learning_rate": 3.053888888888889e-07,
"loss": 1.107,
"step": 11000
},
{
"epoch": 0.45,
"eval_accuracy": 0.3722554890219561,
"eval_loss": 1.0950665473937988,
"eval_runtime": 12.5775,
"eval_samples_per_second": 398.33,
"eval_steps_per_second": 49.851,
"step": 11000
},
{
"epoch": 0.49,
"learning_rate": 3.331666666666667e-07,
"loss": 1.1057,
"step": 12000
},
{
"epoch": 0.49,
"eval_accuracy": 0.37544910179640717,
"eval_loss": 1.0940738916397095,
"eval_runtime": 12.5186,
"eval_samples_per_second": 400.203,
"eval_steps_per_second": 50.085,
"step": 12000
},
{
"epoch": 0.53,
"learning_rate": 3.609166666666667e-07,
"loss": 1.1069,
"step": 13000
},
{
"epoch": 0.53,
"eval_accuracy": 0.3401197604790419,
"eval_loss": 1.0937162637710571,
"eval_runtime": 12.2594,
"eval_samples_per_second": 408.666,
"eval_steps_per_second": 51.144,
"step": 13000
},
{
"epoch": 0.57,
"learning_rate": 3.886944444444445e-07,
"loss": 1.1046,
"step": 14000
},
{
"epoch": 0.57,
"eval_accuracy": 0.37604790419161677,
"eval_loss": 1.0926916599273682,
"eval_runtime": 12.256,
"eval_samples_per_second": 408.78,
"eval_steps_per_second": 51.159,
"step": 14000
},
{
"epoch": 0.61,
"learning_rate": 4.164444444444445e-07,
"loss": 1.1046,
"step": 15000
},
{
"epoch": 0.61,
"eval_accuracy": 0.39201596806387223,
"eval_loss": 1.0912976264953613,
"eval_runtime": 12.5019,
"eval_samples_per_second": 400.74,
"eval_steps_per_second": 50.152,
"step": 15000
},
{
"epoch": 0.65,
"learning_rate": 4.442222222222223e-07,
"loss": 1.1052,
"step": 16000
},
{
"epoch": 0.65,
"eval_accuracy": 0.393812375249501,
"eval_loss": 1.0901482105255127,
"eval_runtime": 12.3407,
"eval_samples_per_second": 405.973,
"eval_steps_per_second": 50.807,
"step": 16000
},
{
"epoch": 0.69,
"learning_rate": 4.7197222222222224e-07,
"loss": 1.1011,
"step": 17000
},
{
"epoch": 0.69,
"eval_accuracy": 0.3932135728542914,
"eval_loss": 1.0901678800582886,
"eval_runtime": 12.3644,
"eval_samples_per_second": 405.197,
"eval_steps_per_second": 50.71,
"step": 17000
},
{
"epoch": 0.73,
"learning_rate": 4.997500000000001e-07,
"loss": 1.1,
"step": 18000
},
{
"epoch": 0.73,
"eval_accuracy": 0.38143712574850297,
"eval_loss": 1.0897998809814453,
"eval_runtime": 12.3516,
"eval_samples_per_second": 405.615,
"eval_steps_per_second": 50.763,
"step": 18000
},
{
"epoch": 0.77,
"learning_rate": 5.275277777777778e-07,
"loss": 1.1007,
"step": 19000
},
{
"epoch": 0.77,
"eval_accuracy": 0.40678642714570856,
"eval_loss": 1.0876121520996094,
"eval_runtime": 12.6608,
"eval_samples_per_second": 395.708,
"eval_steps_per_second": 49.523,
"step": 19000
},
{
"epoch": 0.81,
"learning_rate": 5.552777777777778e-07,
"loss": 1.1013,
"step": 20000
},
{
"epoch": 0.81,
"eval_accuracy": 0.4053892215568862,
"eval_loss": 1.0876350402832031,
"eval_runtime": 12.3243,
"eval_samples_per_second": 406.513,
"eval_steps_per_second": 50.875,
"step": 20000
},
{
"epoch": 0.86,
"learning_rate": 5.830555555555556e-07,
"loss": 1.0987,
"step": 21000
},
{
"epoch": 0.86,
"eval_accuracy": 0.40858283433133735,
"eval_loss": 1.085940957069397,
"eval_runtime": 12.3217,
"eval_samples_per_second": 406.598,
"eval_steps_per_second": 50.886,
"step": 21000
},
{
"epoch": 0.9,
"learning_rate": 6.108055555555556e-07,
"loss": 1.0986,
"step": 22000
},
{
"epoch": 0.9,
"eval_accuracy": 0.3972055888223553,
"eval_loss": 1.085978388786316,
"eval_runtime": 12.4285,
"eval_samples_per_second": 403.107,
"eval_steps_per_second": 50.449,
"step": 22000
},
{
"epoch": 0.94,
"learning_rate": 6.385833333333334e-07,
"loss": 1.0944,
"step": 23000
},
{
"epoch": 0.94,
"eval_accuracy": 0.41097804391217563,
"eval_loss": 1.0837496519088745,
"eval_runtime": 12.7175,
"eval_samples_per_second": 393.946,
"eval_steps_per_second": 49.302,
"step": 23000
},
{
"epoch": 0.98,
"learning_rate": 6.663611111111112e-07,
"loss": 1.095,
"step": 24000
},
{
"epoch": 0.98,
"eval_accuracy": 0.4121756487025948,
"eval_loss": 1.0820704698562622,
"eval_runtime": 12.3323,
"eval_samples_per_second": 406.251,
"eval_steps_per_second": 50.842,
"step": 24000
},
{
"epoch": 1.02,
"learning_rate": 6.941111111111112e-07,
"loss": 1.0915,
"step": 25000
},
{
"epoch": 1.02,
"eval_accuracy": 0.4177644710578842,
"eval_loss": 1.0762931108474731,
"eval_runtime": 12.3887,
"eval_samples_per_second": 404.399,
"eval_steps_per_second": 50.61,
"step": 25000
},
{
"epoch": 1.06,
"learning_rate": 7.218888888888889e-07,
"loss": 1.0889,
"step": 26000
},
{
"epoch": 1.06,
"eval_accuracy": 0.43253493013972055,
"eval_loss": 1.0695326328277588,
"eval_runtime": 12.5611,
"eval_samples_per_second": 398.85,
"eval_steps_per_second": 49.916,
"step": 26000
},
{
"epoch": 1.1,
"learning_rate": 7.49638888888889e-07,
"loss": 1.0827,
"step": 27000
},
{
"epoch": 1.1,
"eval_accuracy": 0.4343313373253493,
"eval_loss": 1.0662548542022705,
"eval_runtime": 12.641,
"eval_samples_per_second": 396.328,
"eval_steps_per_second": 49.6,
"step": 27000
},
{
"epoch": 1.14,
"learning_rate": 7.774166666666668e-07,
"loss": 1.0765,
"step": 28000
},
{
"epoch": 1.14,
"eval_accuracy": 0.4315369261477046,
"eval_loss": 1.0609794855117798,
"eval_runtime": 12.4637,
"eval_samples_per_second": 401.967,
"eval_steps_per_second": 50.306,
"step": 28000
},
{
"epoch": 1.18,
"learning_rate": 8.051666666666667e-07,
"loss": 1.0736,
"step": 29000
},
{
"epoch": 1.18,
"eval_accuracy": 0.4377245508982036,
"eval_loss": 1.0587859153747559,
"eval_runtime": 12.3801,
"eval_samples_per_second": 404.683,
"eval_steps_per_second": 50.646,
"step": 29000
},
{
"epoch": 1.22,
"learning_rate": 8.329444444444445e-07,
"loss": 1.0726,
"step": 30000
},
{
"epoch": 1.22,
"eval_accuracy": 0.4249500998003992,
"eval_loss": 1.0559338331222534,
"eval_runtime": 12.5945,
"eval_samples_per_second": 397.792,
"eval_steps_per_second": 49.784,
"step": 30000
},
{
"epoch": 1.26,
"learning_rate": 8.606944444444445e-07,
"loss": 1.0751,
"step": 31000
},
{
"epoch": 1.26,
"eval_accuracy": 0.4427145708582834,
"eval_loss": 1.0576916933059692,
"eval_runtime": 12.6324,
"eval_samples_per_second": 396.599,
"eval_steps_per_second": 49.634,
"step": 31000
},
{
"epoch": 1.3,
"learning_rate": 8.884722222222224e-07,
"loss": 1.0708,
"step": 32000
},
{
"epoch": 1.3,
"eval_accuracy": 0.42634730538922155,
"eval_loss": 1.0571542978286743,
"eval_runtime": 12.4696,
"eval_samples_per_second": 401.776,
"eval_steps_per_second": 50.282,
"step": 32000
},
{
"epoch": 1.34,
"learning_rate": 9.162222222222223e-07,
"loss": 1.0711,
"step": 33000
},
{
"epoch": 1.34,
"eval_accuracy": 0.4419161676646707,
"eval_loss": 1.0544849634170532,
"eval_runtime": 12.415,
"eval_samples_per_second": 403.544,
"eval_steps_per_second": 50.503,
"step": 33000
},
{
"epoch": 1.39,
"learning_rate": 9.440000000000001e-07,
"loss": 1.0666,
"step": 34000
},
{
"epoch": 1.39,
"eval_accuracy": 0.43333333333333335,
"eval_loss": 1.05018150806427,
"eval_runtime": 12.6226,
"eval_samples_per_second": 396.906,
"eval_steps_per_second": 49.673,
"step": 34000
},
{
"epoch": 1.43,
"learning_rate": 9.7175e-07,
"loss": 1.0698,
"step": 35000
},
{
"epoch": 1.43,
"eval_accuracy": 0.4377245508982036,
"eval_loss": 1.0488406419754028,
"eval_runtime": 12.6246,
"eval_samples_per_second": 396.844,
"eval_steps_per_second": 49.665,
"step": 35000
},
{
"epoch": 1.47,
"learning_rate": 9.995277777777778e-07,
"loss": 1.067,
"step": 36000
},
{
"epoch": 1.47,
"eval_accuracy": 0.4407185628742515,
"eval_loss": 1.047359824180603,
"eval_runtime": 12.4039,
"eval_samples_per_second": 403.906,
"eval_steps_per_second": 50.549,
"step": 36000
},
{
"epoch": 1.51,
"learning_rate": 1.027277777777778e-06,
"loss": 1.0636,
"step": 37000
},
{
"epoch": 1.51,
"eval_accuracy": 0.44451097804391215,
"eval_loss": 1.0455204248428345,
"eval_runtime": 12.3836,
"eval_samples_per_second": 404.569,
"eval_steps_per_second": 50.632,
"step": 37000
},
{
"epoch": 1.55,
"learning_rate": 1.0550555555555557e-06,
"loss": 1.0658,
"step": 38000
},
{
"epoch": 1.55,
"eval_accuracy": 0.4407185628742515,
"eval_loss": 1.0461794137954712,
"eval_runtime": 12.5546,
"eval_samples_per_second": 399.057,
"eval_steps_per_second": 49.942,
"step": 38000
},
{
"epoch": 1.59,
"learning_rate": 1.0828333333333334e-06,
"loss": 1.0635,
"step": 39000
},
{
"epoch": 1.59,
"eval_accuracy": 0.4367265469061876,
"eval_loss": 1.0461270809173584,
"eval_runtime": 12.5509,
"eval_samples_per_second": 399.173,
"eval_steps_per_second": 49.956,
"step": 39000
},
{
"epoch": 1.63,
"learning_rate": 1.1105833333333335e-06,
"loss": 1.0596,
"step": 40000
},
{
"epoch": 1.63,
"eval_accuracy": 0.449500998003992,
"eval_loss": 1.0418727397918701,
"eval_runtime": 12.4044,
"eval_samples_per_second": 403.889,
"eval_steps_per_second": 50.547,
"step": 40000
},
{
"epoch": 1.67,
"learning_rate": 1.1383611111111113e-06,
"loss": 1.0595,
"step": 41000
},
{
"epoch": 1.67,
"eval_accuracy": 0.4409181636726547,
"eval_loss": 1.0417557954788208,
"eval_runtime": 12.3642,
"eval_samples_per_second": 405.202,
"eval_steps_per_second": 50.711,
"step": 41000
},
{
"epoch": 1.71,
"learning_rate": 1.1661111111111111e-06,
"loss": 1.054,
"step": 42000
},
{
"epoch": 1.71,
"eval_accuracy": 0.4357285429141717,
"eval_loss": 1.0383535623550415,
"eval_runtime": 12.3634,
"eval_samples_per_second": 405.23,
"eval_steps_per_second": 50.714,
"step": 42000
},
{
"epoch": 1.75,
"learning_rate": 1.193888888888889e-06,
"loss": 1.0543,
"step": 43000
},
{
"epoch": 1.75,
"eval_accuracy": 0.43812375249500995,
"eval_loss": 1.0463169813156128,
"eval_runtime": 12.6398,
"eval_samples_per_second": 396.366,
"eval_steps_per_second": 49.605,
"step": 43000
},
{
"epoch": 1.79,
"learning_rate": 1.221638888888889e-06,
"loss": 1.0579,
"step": 44000
},
{
"epoch": 1.79,
"eval_accuracy": 0.4415169660678643,
"eval_loss": 1.0388957262039185,
"eval_runtime": 12.4542,
"eval_samples_per_second": 402.274,
"eval_steps_per_second": 50.344,
"step": 44000
},
{
"epoch": 1.83,
"learning_rate": 1.2494166666666668e-06,
"loss": 1.0561,
"step": 45000
},
{
"epoch": 1.83,
"eval_accuracy": 0.46187624750499,
"eval_loss": 1.031552791595459,
"eval_runtime": 12.4082,
"eval_samples_per_second": 403.764,
"eval_steps_per_second": 50.531,
"step": 45000
},
{
"epoch": 1.87,
"learning_rate": 1.2771666666666668e-06,
"loss": 1.0516,
"step": 46000
},
{
"epoch": 1.87,
"eval_accuracy": 0.4538922155688623,
"eval_loss": 1.0312999486923218,
"eval_runtime": 12.3695,
"eval_samples_per_second": 405.029,
"eval_steps_per_second": 50.689,
"step": 46000
},
{
"epoch": 1.91,
"learning_rate": 1.3049444444444446e-06,
"loss": 1.0471,
"step": 47000
},
{
"epoch": 1.91,
"eval_accuracy": 0.46387225548902195,
"eval_loss": 1.027535319328308,
"eval_runtime": 12.5846,
"eval_samples_per_second": 398.107,
"eval_steps_per_second": 49.823,
"step": 47000
},
{
"epoch": 1.96,
"learning_rate": 1.3326944444444447e-06,
"loss": 1.054,
"step": 48000
},
{
"epoch": 1.96,
"eval_accuracy": 0.4568862275449102,
"eval_loss": 1.0308438539505005,
"eval_runtime": 12.5646,
"eval_samples_per_second": 398.74,
"eval_steps_per_second": 49.902,
"step": 48000
},
{
"epoch": 2.0,
"learning_rate": 1.3604722222222224e-06,
"loss": 1.0445,
"step": 49000
},
{
"epoch": 2.0,
"eval_accuracy": 0.47045908183632734,
"eval_loss": 1.0230755805969238,
"eval_runtime": 12.4305,
"eval_samples_per_second": 403.041,
"eval_steps_per_second": 50.44,
"step": 49000
},
{
"epoch": 2.04,
"learning_rate": 1.3882222222222223e-06,
"loss": 1.0479,
"step": 50000
},
{
"epoch": 2.04,
"eval_accuracy": 0.47305389221556887,
"eval_loss": 1.0201445817947388,
"eval_runtime": 12.3672,
"eval_samples_per_second": 405.103,
"eval_steps_per_second": 50.698,
"step": 50000
},
{
"epoch": 2.08,
"learning_rate": 1.416e-06,
"loss": 1.0482,
"step": 51000
},
{
"epoch": 2.08,
"eval_accuracy": 0.4678642714570858,
"eval_loss": 1.0167714357376099,
"eval_runtime": 12.5982,
"eval_samples_per_second": 397.677,
"eval_steps_per_second": 49.769,
"step": 51000
},
{
"epoch": 2.12,
"learning_rate": 1.4437500000000002e-06,
"loss": 1.0406,
"step": 52000
},
{
"epoch": 2.12,
"eval_accuracy": 0.4722554890219561,
"eval_loss": 1.0175116062164307,
"eval_runtime": 12.6041,
"eval_samples_per_second": 397.49,
"eval_steps_per_second": 49.746,
"step": 52000
},
{
"epoch": 2.16,
"learning_rate": 1.471527777777778e-06,
"loss": 1.0457,
"step": 53000
},
{
"epoch": 2.16,
"eval_accuracy": 0.4748502994011976,
"eval_loss": 1.010857343673706,
"eval_runtime": 12.3728,
"eval_samples_per_second": 404.921,
"eval_steps_per_second": 50.676,
"step": 53000
},
{
"epoch": 2.2,
"learning_rate": 1.4993055555555557e-06,
"loss": 1.0399,
"step": 54000
},
{
"epoch": 2.2,
"eval_accuracy": 0.48522954091816367,
"eval_loss": 1.0106219053268433,
"eval_runtime": 12.3952,
"eval_samples_per_second": 404.188,
"eval_steps_per_second": 50.584,
"step": 54000
},
{
"epoch": 2.24,
"learning_rate": 1.5270555555555558e-06,
"loss": 1.0412,
"step": 55000
},
{
"epoch": 2.24,
"eval_accuracy": 0.4750499001996008,
"eval_loss": 1.0111658573150635,
"eval_runtime": 12.4858,
"eval_samples_per_second": 401.257,
"eval_steps_per_second": 50.217,
"step": 55000
},
{
"epoch": 2.28,
"learning_rate": 1.5548333333333335e-06,
"loss": 1.0347,
"step": 56000
},
{
"epoch": 2.28,
"eval_accuracy": 0.47924151696606787,
"eval_loss": 1.0121456384658813,
"eval_runtime": 12.6206,
"eval_samples_per_second": 396.97,
"eval_steps_per_second": 49.681,
"step": 56000
},
{
"epoch": 2.32,
"learning_rate": 1.5825833333333334e-06,
"loss": 1.0324,
"step": 57000
},
{
"epoch": 2.32,
"eval_accuracy": 0.4738522954091816,
"eval_loss": 1.0100611448287964,
"eval_runtime": 12.3748,
"eval_samples_per_second": 404.857,
"eval_steps_per_second": 50.668,
"step": 57000
},
{
"epoch": 2.36,
"learning_rate": 1.6103611111111112e-06,
"loss": 1.0346,
"step": 58000
},
{
"epoch": 2.36,
"eval_accuracy": 0.4694610778443114,
"eval_loss": 1.0124626159667969,
"eval_runtime": 12.4716,
"eval_samples_per_second": 401.711,
"eval_steps_per_second": 50.274,
"step": 58000
},
{
"epoch": 2.4,
"learning_rate": 1.638111111111111e-06,
"loss": 1.0309,
"step": 59000
},
{
"epoch": 2.4,
"eval_accuracy": 0.48982035928143713,
"eval_loss": 1.0022122859954834,
"eval_runtime": 12.4734,
"eval_samples_per_second": 401.655,
"eval_steps_per_second": 50.267,
"step": 59000
},
{
"epoch": 2.44,
"learning_rate": 1.665888888888889e-06,
"loss": 1.0351,
"step": 60000
},
{
"epoch": 2.44,
"eval_accuracy": 0.4802395209580838,
"eval_loss": 1.004167914390564,
"eval_runtime": 12.6287,
"eval_samples_per_second": 396.714,
"eval_steps_per_second": 49.649,
"step": 60000
},
{
"epoch": 2.49,
"learning_rate": 1.693638888888889e-06,
"loss": 1.0312,
"step": 61000
},
{
"epoch": 2.49,
"eval_accuracy": 0.48323353293413174,
"eval_loss": 1.0065865516662598,
"eval_runtime": 12.4143,
"eval_samples_per_second": 403.567,
"eval_steps_per_second": 50.506,
"step": 61000
},
{
"epoch": 2.53,
"learning_rate": 1.7214166666666666e-06,
"loss": 1.0329,
"step": 62000
},
{
"epoch": 2.53,
"eval_accuracy": 0.47924151696606787,
"eval_loss": 1.0041630268096924,
"eval_runtime": 12.4326,
"eval_samples_per_second": 402.974,
"eval_steps_per_second": 50.432,
"step": 62000
},
{
"epoch": 2.57,
"learning_rate": 1.7491666666666667e-06,
"loss": 1.0355,
"step": 63000
},
{
"epoch": 2.57,
"eval_accuracy": 0.49261477045908186,
"eval_loss": 1.0007619857788086,
"eval_runtime": 12.4051,
"eval_samples_per_second": 403.866,
"eval_steps_per_second": 50.544,
"step": 63000
},
{
"epoch": 2.61,
"learning_rate": 1.7769444444444447e-06,
"loss": 1.0297,
"step": 64000
},
{
"epoch": 2.61,
"eval_accuracy": 0.4942115768463074,
"eval_loss": 0.9986936450004578,
"eval_runtime": 12.5727,
"eval_samples_per_second": 398.482,
"eval_steps_per_second": 49.87,
"step": 64000
},
{
"epoch": 2.65,
"learning_rate": 1.8046944444444446e-06,
"loss": 1.0277,
"step": 65000
},
{
"epoch": 2.65,
"eval_accuracy": 0.49121756487025947,
"eval_loss": 1.0008972883224487,
"eval_runtime": 12.4544,
"eval_samples_per_second": 402.267,
"eval_steps_per_second": 50.344,
"step": 65000
},
{
"epoch": 2.69,
"learning_rate": 1.8324722222222223e-06,
"loss": 1.0341,
"step": 66000
},
{
"epoch": 2.69,
"eval_accuracy": 0.499001996007984,
"eval_loss": 0.9953575134277344,
"eval_runtime": 12.4973,
"eval_samples_per_second": 400.886,
"eval_steps_per_second": 50.171,
"step": 66000
},
{
"epoch": 2.73,
"learning_rate": 1.8602222222222222e-06,
"loss": 1.0279,
"step": 67000
},
{
"epoch": 2.73,
"eval_accuracy": 0.493812375249501,
"eval_loss": 0.998134195804596,
"eval_runtime": 12.6589,
"eval_samples_per_second": 395.769,
"eval_steps_per_second": 49.53,
"step": 67000
},
{
"epoch": 2.77,
"learning_rate": 1.8880000000000002e-06,
"loss": 1.03,
"step": 68000
},
{
"epoch": 2.77,
"eval_accuracy": 0.5051896207584831,
"eval_loss": 0.9947481751441956,
"eval_runtime": 12.7939,
"eval_samples_per_second": 391.594,
"eval_steps_per_second": 49.008,
"step": 68000
},
{
"epoch": 2.81,
"learning_rate": 1.9157500000000003e-06,
"loss": 1.0283,
"step": 69000
},
{
"epoch": 2.81,
"eval_accuracy": 0.5001996007984032,
"eval_loss": 0.997833251953125,
"eval_runtime": 12.8129,
"eval_samples_per_second": 391.012,
"eval_steps_per_second": 48.935,
"step": 69000
},
{
"epoch": 2.85,
"learning_rate": 1.943527777777778e-06,
"loss": 1.0288,
"step": 70000
},
{
"epoch": 2.85,
"eval_accuracy": 0.49181636726546907,
"eval_loss": 1.00169038772583,
"eval_runtime": 12.5936,
"eval_samples_per_second": 397.822,
"eval_steps_per_second": 49.787,
"step": 70000
},
{
"epoch": 2.89,
"learning_rate": 1.9712777777777777e-06,
"loss": 1.0238,
"step": 71000
},
{
"epoch": 2.89,
"eval_accuracy": 0.4944111776447106,
"eval_loss": 1.001316785812378,
"eval_runtime": 12.485,
"eval_samples_per_second": 401.282,
"eval_steps_per_second": 50.22,
"step": 71000
},
{
"epoch": 2.93,
"learning_rate": 1.9990555555555557e-06,
"loss": 1.0268,
"step": 72000
},
{
"epoch": 2.93,
"eval_accuracy": 0.49241516966067866,
"eval_loss": 1.0005468130111694,
"eval_runtime": 12.501,
"eval_samples_per_second": 400.768,
"eval_steps_per_second": 50.156,
"step": 72000
},
{
"epoch": 2.97,
"learning_rate": 2.026805555555556e-06,
"loss": 1.0266,
"step": 73000
},
{
"epoch": 2.97,
"eval_accuracy": 0.4940119760479042,
"eval_loss": 0.9998964071273804,
"eval_runtime": 12.6227,
"eval_samples_per_second": 396.904,
"eval_steps_per_second": 49.672,
"step": 73000
},
{
"epoch": 3.01,
"learning_rate": 2.0545833333333335e-06,
"loss": 1.0222,
"step": 74000
},
{
"epoch": 3.01,
"eval_accuracy": 0.501996007984032,
"eval_loss": 0.9927905201911926,
"eval_runtime": 12.7924,
"eval_samples_per_second": 391.64,
"eval_steps_per_second": 49.014,
"step": 74000
},
{
"epoch": 3.06,
"learning_rate": 2.0823333333333334e-06,
"loss": 1.0198,
"step": 75000
},
{
"epoch": 3.06,
"eval_accuracy": 0.5025948103792415,
"eval_loss": 0.9925487041473389,
"eval_runtime": 12.6297,
"eval_samples_per_second": 396.684,
"eval_steps_per_second": 49.645,
"step": 75000
},
{
"epoch": 3.1,
"learning_rate": 2.1101111111111113e-06,
"loss": 1.0203,
"step": 76000
},
{
"epoch": 3.1,
"eval_accuracy": 0.5027944111776447,
"eval_loss": 0.9909006953239441,
"eval_runtime": 12.4265,
"eval_samples_per_second": 403.17,
"eval_steps_per_second": 50.457,
"step": 76000
},
{
"epoch": 3.14,
"learning_rate": 2.137861111111111e-06,
"loss": 1.022,
"step": 77000
},
{
"epoch": 3.14,
"eval_accuracy": 0.48303393213572854,
"eval_loss": 1.010209083557129,
"eval_runtime": 12.3999,
"eval_samples_per_second": 404.037,
"eval_steps_per_second": 50.565,
"step": 77000
},
{
"epoch": 3.18,
"learning_rate": 2.1656388888888888e-06,
"loss": 1.0133,
"step": 78000
},
{
"epoch": 3.18,
"eval_accuracy": 0.5075848303393213,
"eval_loss": 0.9907119274139404,
"eval_runtime": 12.4429,
"eval_samples_per_second": 402.64,
"eval_steps_per_second": 50.39,
"step": 78000
},
{
"epoch": 3.22,
"learning_rate": 2.193388888888889e-06,
"loss": 1.0207,
"step": 79000
},
{
"epoch": 3.22,
"eval_accuracy": 0.4944111776447106,
"eval_loss": 1.0015305280685425,
"eval_runtime": 12.7143,
"eval_samples_per_second": 394.046,
"eval_steps_per_second": 49.315,
"step": 79000
},
{
"epoch": 3.26,
"learning_rate": 2.221166666666667e-06,
"loss": 1.0172,
"step": 80000
},
{
"epoch": 3.26,
"eval_accuracy": 0.49540918163672654,
"eval_loss": 0.9991635680198669,
"eval_runtime": 12.7122,
"eval_samples_per_second": 394.109,
"eval_steps_per_second": 49.323,
"step": 80000
},
{
"epoch": 3.3,
"learning_rate": 2.248916666666667e-06,
"loss": 1.0185,
"step": 81000
},
{
"epoch": 3.3,
"eval_accuracy": 0.49001996007984033,
"eval_loss": 1.0026898384094238,
"eval_runtime": 12.6101,
"eval_samples_per_second": 397.299,
"eval_steps_per_second": 49.722,
"step": 81000
},
{
"epoch": 3.34,
"learning_rate": 2.2766944444444444e-06,
"loss": 1.0287,
"step": 82000
},
{
"epoch": 3.34,
"eval_accuracy": 0.4998003992015968,
"eval_loss": 0.9921432733535767,
"eval_runtime": 12.4639,
"eval_samples_per_second": 401.961,
"eval_steps_per_second": 50.305,
"step": 82000
},
{
"epoch": 3.38,
"learning_rate": 2.3044444444444443e-06,
"loss": 1.0142,
"step": 83000
},
{
"epoch": 3.38,
"eval_accuracy": 0.5017964071856288,
"eval_loss": 0.9938461184501648,
"eval_runtime": 12.4125,
"eval_samples_per_second": 403.625,
"eval_steps_per_second": 50.514,
"step": 83000
},
{
"epoch": 3.42,
"learning_rate": 2.3322222222222223e-06,
"loss": 1.0186,
"step": 84000
},
{
"epoch": 3.42,
"eval_accuracy": 0.5089820359281437,
"eval_loss": 0.9863815307617188,
"eval_runtime": 12.6556,
"eval_samples_per_second": 395.872,
"eval_steps_per_second": 49.543,
"step": 84000
},
{
"epoch": 3.46,
"learning_rate": 2.3599722222222226e-06,
"loss": 1.0151,
"step": 85000
},
{
"epoch": 3.46,
"eval_accuracy": 0.5041916167664671,
"eval_loss": 0.9891662001609802,
"eval_runtime": 12.6529,
"eval_samples_per_second": 395.958,
"eval_steps_per_second": 49.554,
"step": 85000
},
{
"epoch": 3.5,
"learning_rate": 2.38775e-06,
"loss": 1.0268,
"step": 86000
},
{
"epoch": 3.5,
"eval_accuracy": 0.5043912175648703,
"eval_loss": 0.9857804179191589,
"eval_runtime": 12.4197,
"eval_samples_per_second": 403.391,
"eval_steps_per_second": 50.484,
"step": 86000
},
{
"epoch": 3.54,
"learning_rate": 2.4155e-06,
"loss": 1.0152,
"step": 87000
},
{
"epoch": 3.54,
"eval_accuracy": 0.5075848303393213,
"eval_loss": 0.9841725826263428,
"eval_runtime": 12.4152,
"eval_samples_per_second": 403.538,
"eval_steps_per_second": 50.503,
"step": 87000
},
{
"epoch": 3.59,
"learning_rate": 2.443277777777778e-06,
"loss": 1.0201,
"step": 88000
},
{
"epoch": 3.59,
"eval_accuracy": 0.5065868263473053,
"eval_loss": 0.9841102957725525,
"eval_runtime": 12.4716,
"eval_samples_per_second": 401.711,
"eval_steps_per_second": 50.274,
"step": 88000
},
{
"epoch": 3.63,
"learning_rate": 2.471027777777778e-06,
"loss": 1.0113,
"step": 89000
},
{
"epoch": 3.63,
"eval_accuracy": 0.5079840319361277,
"eval_loss": 0.9851917028427124,
"eval_runtime": 12.6634,
"eval_samples_per_second": 395.627,
"eval_steps_per_second": 49.513,
"step": 89000
},
{
"epoch": 3.67,
"learning_rate": 2.498805555555556e-06,
"loss": 1.0195,
"step": 90000
},
{
"epoch": 3.67,
"eval_accuracy": 0.5039920159680639,
"eval_loss": 0.9855921864509583,
"eval_runtime": 12.6226,
"eval_samples_per_second": 396.908,
"eval_steps_per_second": 49.673,
"step": 90000
},
{
"epoch": 3.71,
"learning_rate": 2.5265555555555557e-06,
"loss": 1.0202,
"step": 91000
},
{
"epoch": 3.71,
"eval_accuracy": 0.5001996007984032,
"eval_loss": 0.9933446645736694,
"eval_runtime": 12.4993,
"eval_samples_per_second": 400.821,
"eval_steps_per_second": 50.163,
"step": 91000
},
{
"epoch": 3.75,
"learning_rate": 2.5543333333333337e-06,
"loss": 1.0183,
"step": 92000
},
{
"epoch": 3.75,
"eval_accuracy": 0.5,
"eval_loss": 0.9928790926933289,
"eval_runtime": 12.4539,
"eval_samples_per_second": 402.285,
"eval_steps_per_second": 50.346,
"step": 92000
},
{
"epoch": 3.79,
"learning_rate": 2.5820833333333335e-06,
"loss": 1.02,
"step": 93000
},
{
"epoch": 3.79,
"eval_accuracy": 0.5011976047904192,
"eval_loss": 0.9867545962333679,
"eval_runtime": 12.4054,
"eval_samples_per_second": 403.855,
"eval_steps_per_second": 50.542,
"step": 93000
},
{
"epoch": 3.83,
"learning_rate": 2.609861111111111e-06,
"loss": 1.0144,
"step": 94000
},
{
"epoch": 3.83,
"eval_accuracy": 0.5027944111776447,
"eval_loss": 0.9887515902519226,
"eval_runtime": 12.8101,
"eval_samples_per_second": 391.097,
"eval_steps_per_second": 48.946,
"step": 94000
},
{
"epoch": 3.87,
"learning_rate": 2.6376111111111114e-06,
"loss": 1.0166,
"step": 95000
},
{
"epoch": 3.87,
"eval_accuracy": 0.5077844311377245,
"eval_loss": 0.9852330684661865,
"eval_runtime": 12.6394,
"eval_samples_per_second": 396.379,
"eval_steps_per_second": 49.607,
"step": 95000
},
{
"epoch": 3.91,
"learning_rate": 2.6653888888888894e-06,
"loss": 1.0173,
"step": 96000
},
{
"epoch": 3.91,
"eval_accuracy": 0.5053892215568863,
"eval_loss": 0.9848083257675171,
"eval_runtime": 12.5421,
"eval_samples_per_second": 399.456,
"eval_steps_per_second": 49.992,
"step": 96000
},
{
"epoch": 3.95,
"learning_rate": 2.6931388888888892e-06,
"loss": 1.0159,
"step": 97000
},
{
"epoch": 3.95,
"eval_accuracy": 0.5051896207584831,
"eval_loss": 0.9930784702301025,
"eval_runtime": 12.6498,
"eval_samples_per_second": 396.054,
"eval_steps_per_second": 49.566,
"step": 97000
},
{
"epoch": 3.99,
"learning_rate": 2.7209166666666668e-06,
"loss": 1.0126,
"step": 98000
},
{
"epoch": 3.99,
"eval_accuracy": 0.5037924151696607,
"eval_loss": 0.9865526556968689,
"eval_runtime": 12.4969,
"eval_samples_per_second": 400.899,
"eval_steps_per_second": 50.172,
"step": 98000
},
{
"epoch": 4.03,
"learning_rate": 2.7486666666666666e-06,
"loss": 1.0129,
"step": 99000
},
{
"epoch": 4.03,
"eval_accuracy": 0.5073852295409181,
"eval_loss": 0.9890230298042297,
"eval_runtime": 12.7759,
"eval_samples_per_second": 392.143,
"eval_steps_per_second": 49.077,
"step": 99000
},
{
"epoch": 4.07,
"learning_rate": 2.7764444444444446e-06,
"loss": 1.0145,
"step": 100000
},
{
"epoch": 4.07,
"eval_accuracy": 0.506187624750499,
"eval_loss": 0.9834555387496948,
"eval_runtime": 12.6734,
"eval_samples_per_second": 395.316,
"eval_steps_per_second": 49.474,
"step": 100000
},
{
"epoch": 4.12,
"learning_rate": 2.804194444444445e-06,
"loss": 1.0164,
"step": 101000
},
{
"epoch": 4.12,
"eval_accuracy": 0.5081836327345309,
"eval_loss": 0.9881391525268555,
"eval_runtime": 12.4161,
"eval_samples_per_second": 403.509,
"eval_steps_per_second": 50.499,
"step": 101000
},
{
"epoch": 4.16,
"learning_rate": 2.8319722222222225e-06,
"loss": 1.0131,
"step": 102000
},
{
"epoch": 4.16,
"eval_accuracy": 0.5007984031936128,
"eval_loss": 0.9866752624511719,
"eval_runtime": 12.6211,
"eval_samples_per_second": 396.953,
"eval_steps_per_second": 49.679,
"step": 102000
},
{
"epoch": 4.2,
"learning_rate": 2.8597222222222223e-06,
"loss": 1.0172,
"step": 103000
},
{
"epoch": 4.2,
"eval_accuracy": 0.5117764471057884,
"eval_loss": 0.9857889413833618,
"eval_runtime": 12.4246,
"eval_samples_per_second": 403.233,
"eval_steps_per_second": 50.464,
"step": 103000
},
{
"epoch": 4.24,
"learning_rate": 2.8875000000000003e-06,
"loss": 1.0118,
"step": 104000
},
{
"epoch": 4.24,
"eval_accuracy": 0.501996007984032,
"eval_loss": 0.9917896389961243,
"eval_runtime": 12.7707,
"eval_samples_per_second": 392.304,
"eval_steps_per_second": 49.097,
"step": 104000
},
{
"epoch": 4.28,
"learning_rate": 2.91525e-06,
"loss": 1.0082,
"step": 105000
},
{
"epoch": 4.28,
"eval_accuracy": 0.4940119760479042,
"eval_loss": 1.0039118528366089,
"eval_runtime": 12.7128,
"eval_samples_per_second": 394.092,
"eval_steps_per_second": 49.32,
"step": 105000
},
{
"epoch": 4.32,
"learning_rate": 2.9430277777777777e-06,
"loss": 1.006,
"step": 106000
},
{
"epoch": 4.32,
"eval_accuracy": 0.5023952095808383,
"eval_loss": 0.9883149862289429,
"eval_runtime": 12.7217,
"eval_samples_per_second": 393.814,
"eval_steps_per_second": 49.286,
"step": 106000
},
{
"epoch": 4.36,
"learning_rate": 2.970777777777778e-06,
"loss": 1.008,
"step": 107000
},
{
"epoch": 4.36,
"eval_accuracy": 0.5023952095808383,
"eval_loss": 0.9926295876502991,
"eval_runtime": 12.4187,
"eval_samples_per_second": 403.424,
"eval_steps_per_second": 50.488,
"step": 107000
},
{
"epoch": 4.4,
"learning_rate": 2.998555555555556e-06,
"loss": 1.0078,
"step": 108000
},
{
"epoch": 4.4,
"eval_accuracy": 0.5069860279441117,
"eval_loss": 0.9883684515953064,
"eval_runtime": 12.4562,
"eval_samples_per_second": 402.21,
"eval_steps_per_second": 50.336,
"step": 108000
},
{
"epoch": 4.44,
"learning_rate": 3.026305555555556e-06,
"loss": 1.0052,
"step": 109000
},
{
"epoch": 4.44,
"eval_accuracy": 0.5083832335329341,
"eval_loss": 0.9893190264701843,
"eval_runtime": 12.4193,
"eval_samples_per_second": 403.404,
"eval_steps_per_second": 50.486,
"step": 109000
},
{
"epoch": 4.48,
"learning_rate": 3.0540833333333334e-06,
"loss": 1.0102,
"step": 110000
},
{
"epoch": 4.48,
"eval_accuracy": 0.5059880239520959,
"eval_loss": 0.9876956939697266,
"eval_runtime": 12.7204,
"eval_samples_per_second": 393.854,
"eval_steps_per_second": 49.291,
"step": 110000
},
{
"epoch": 4.52,
"learning_rate": 3.0818333333333333e-06,
"loss": 1.004,
"step": 111000
},
{
"epoch": 4.52,
"eval_accuracy": 0.5099800399201597,
"eval_loss": 0.9822403788566589,
"eval_runtime": 12.7671,
"eval_samples_per_second": 392.416,
"eval_steps_per_second": 49.111,
"step": 111000
}
],
"logging_steps": 1000,
"max_steps": 10000000,
"num_train_epochs": 408,
"save_steps": 1000,
"total_flos": 3.867663332907418e+16,
"trial_name": null,
"trial_params": null
}