wav2vec2-xls-r-300m-npsc-bokmaal / trainer_state.json
versae's picture
End of training
91f970e
raw
history blame
40.1 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 14.999677731227845,
"global_step": 23265,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.06,
"learning_rate": 1e-05,
"loss": 0.0958,
"step": 100
},
{
"epoch": 0.13,
"learning_rate": 1.9900000000000003e-05,
"loss": 0.0914,
"step": 200
},
{
"epoch": 0.19,
"learning_rate": 2.9900000000000002e-05,
"loss": 0.0893,
"step": 300
},
{
"epoch": 0.26,
"learning_rate": 3.99e-05,
"loss": 0.0925,
"step": 400
},
{
"epoch": 0.32,
"learning_rate": 4.99e-05,
"loss": 0.0969,
"step": 500
},
{
"epoch": 0.32,
"eval_loss": 0.17728671431541443,
"eval_runtime": 380.6151,
"eval_samples_per_second": 14.285,
"eval_steps_per_second": 0.893,
"eval_wer": 0.10537617539490327,
"step": 500
},
{
"epoch": 0.39,
"learning_rate": 4.978256094882495e-05,
"loss": 0.0982,
"step": 600
},
{
"epoch": 0.45,
"learning_rate": 4.956292554359763e-05,
"loss": 0.0982,
"step": 700
},
{
"epoch": 0.52,
"learning_rate": 4.9343290138370304e-05,
"loss": 0.0939,
"step": 800
},
{
"epoch": 0.58,
"learning_rate": 4.9125851087195255e-05,
"loss": 0.0957,
"step": 900
},
{
"epoch": 0.64,
"learning_rate": 4.8906215681967934e-05,
"loss": 0.0929,
"step": 1000
},
{
"epoch": 0.64,
"eval_loss": 0.16721801459789276,
"eval_runtime": 381.6558,
"eval_samples_per_second": 14.246,
"eval_steps_per_second": 0.891,
"eval_wer": 0.1061409787521939,
"step": 1000
},
{
"epoch": 0.71,
"learning_rate": 4.8688776630792884e-05,
"loss": 0.0938,
"step": 1100
},
{
"epoch": 0.77,
"learning_rate": 4.8469141225565564e-05,
"loss": 0.088,
"step": 1200
},
{
"epoch": 0.84,
"learning_rate": 4.8251702174390514e-05,
"loss": 0.1015,
"step": 1300
},
{
"epoch": 0.9,
"learning_rate": 4.803206676916319e-05,
"loss": 0.0963,
"step": 1400
},
{
"epoch": 0.97,
"learning_rate": 4.7812431363935866e-05,
"loss": 0.1018,
"step": 1500
},
{
"epoch": 0.97,
"eval_loss": 0.17695783078670502,
"eval_runtime": 379.2394,
"eval_samples_per_second": 14.337,
"eval_steps_per_second": 0.897,
"eval_wer": 0.10669987351329091,
"step": 1500
},
{
"epoch": 1.03,
"learning_rate": 4.7592795958708545e-05,
"loss": 0.0899,
"step": 1600
},
{
"epoch": 1.1,
"learning_rate": 4.7373160553481224e-05,
"loss": 0.0898,
"step": 1700
},
{
"epoch": 1.16,
"learning_rate": 4.7153525148253903e-05,
"loss": 0.0884,
"step": 1800
},
{
"epoch": 1.22,
"learning_rate": 4.6933889743026576e-05,
"loss": 0.0914,
"step": 1900
},
{
"epoch": 1.29,
"learning_rate": 4.6714254337799255e-05,
"loss": 0.0871,
"step": 2000
},
{
"epoch": 1.29,
"eval_loss": 0.18323425948619843,
"eval_runtime": 384.6744,
"eval_samples_per_second": 14.134,
"eval_steps_per_second": 0.884,
"eval_wer": 0.1086805181052487,
"step": 2000
},
{
"epoch": 1.35,
"learning_rate": 4.6494618932571934e-05,
"loss": 0.0918,
"step": 2100
},
{
"epoch": 1.42,
"learning_rate": 4.6274983527344614e-05,
"loss": 0.0909,
"step": 2200
},
{
"epoch": 1.48,
"learning_rate": 4.6057544476169564e-05,
"loss": 0.0988,
"step": 2300
},
{
"epoch": 1.55,
"learning_rate": 4.5837909070942236e-05,
"loss": 0.0967,
"step": 2400
},
{
"epoch": 1.61,
"learning_rate": 4.5618273665714916e-05,
"loss": 0.0908,
"step": 2500
},
{
"epoch": 1.61,
"eval_loss": 0.1829613745212555,
"eval_runtime": 380.5109,
"eval_samples_per_second": 14.289,
"eval_steps_per_second": 0.894,
"eval_wer": 0.11013168344985146,
"step": 2500
},
{
"epoch": 1.68,
"learning_rate": 4.539863826048759e-05,
"loss": 0.0901,
"step": 2600
},
{
"epoch": 1.74,
"learning_rate": 4.5179002855260274e-05,
"loss": 0.0934,
"step": 2700
},
{
"epoch": 1.81,
"learning_rate": 4.495936745003295e-05,
"loss": 0.0958,
"step": 2800
},
{
"epoch": 1.87,
"learning_rate": 4.4739732044805626e-05,
"loss": 0.0909,
"step": 2900
},
{
"epoch": 1.93,
"learning_rate": 4.45200966395783e-05,
"loss": 0.0975,
"step": 3000
},
{
"epoch": 1.93,
"eval_loss": 0.1847657710313797,
"eval_runtime": 376.8362,
"eval_samples_per_second": 14.428,
"eval_steps_per_second": 0.902,
"eval_wer": 0.11004343690862561,
"step": 3000
},
{
"epoch": 2.0,
"learning_rate": 4.4300461234350984e-05,
"loss": 0.0977,
"step": 3100
},
{
"epoch": 2.06,
"learning_rate": 4.408082582912366e-05,
"loss": 0.0949,
"step": 3200
},
{
"epoch": 2.13,
"learning_rate": 4.3861190423896336e-05,
"loss": 0.0916,
"step": 3300
},
{
"epoch": 2.19,
"learning_rate": 4.364155501866901e-05,
"loss": 0.0963,
"step": 3400
},
{
"epoch": 2.26,
"learning_rate": 4.342191961344169e-05,
"loss": 0.0936,
"step": 3500
},
{
"epoch": 2.26,
"eval_loss": 0.18531930446624756,
"eval_runtime": 380.7297,
"eval_samples_per_second": 14.28,
"eval_steps_per_second": 0.893,
"eval_wer": 0.1112690833145401,
"step": 3500
},
{
"epoch": 2.32,
"learning_rate": 4.320228420821437e-05,
"loss": 0.0955,
"step": 3600
},
{
"epoch": 2.39,
"learning_rate": 4.298264880298704e-05,
"loss": 0.093,
"step": 3700
},
{
"epoch": 2.45,
"learning_rate": 4.276301339775972e-05,
"loss": 0.0924,
"step": 3800
},
{
"epoch": 2.51,
"learning_rate": 4.254557434658467e-05,
"loss": 0.1065,
"step": 3900
},
{
"epoch": 2.58,
"learning_rate": 4.232593894135735e-05,
"loss": 0.1025,
"step": 4000
},
{
"epoch": 2.58,
"eval_loss": 0.19584240019321442,
"eval_runtime": 374.8969,
"eval_samples_per_second": 14.503,
"eval_steps_per_second": 0.907,
"eval_wer": 0.11489699667604694,
"step": 4000
},
{
"epoch": 2.64,
"learning_rate": 4.210630353613003e-05,
"loss": 0.101,
"step": 4100
},
{
"epoch": 2.71,
"learning_rate": 4.18866681309027e-05,
"loss": 0.1038,
"step": 4200
},
{
"epoch": 2.77,
"learning_rate": 4.166703272567538e-05,
"loss": 0.1043,
"step": 4300
},
{
"epoch": 2.84,
"learning_rate": 4.144739732044806e-05,
"loss": 0.1057,
"step": 4400
},
{
"epoch": 2.9,
"learning_rate": 4.122776191522074e-05,
"loss": 0.0989,
"step": 4500
},
{
"epoch": 2.9,
"eval_loss": 0.17756135761737823,
"eval_runtime": 376.0235,
"eval_samples_per_second": 14.459,
"eval_steps_per_second": 0.904,
"eval_wer": 0.11230843146675556,
"step": 4500
},
{
"epoch": 2.97,
"learning_rate": 4.100812650999341e-05,
"loss": 0.0944,
"step": 4600
},
{
"epoch": 3.03,
"learning_rate": 4.078849110476609e-05,
"loss": 0.0911,
"step": 4700
},
{
"epoch": 3.09,
"learning_rate": 4.056885569953877e-05,
"loss": 0.0903,
"step": 4800
},
{
"epoch": 3.16,
"learning_rate": 4.034922029431145e-05,
"loss": 0.093,
"step": 4900
},
{
"epoch": 3.22,
"learning_rate": 4.012958488908412e-05,
"loss": 0.0946,
"step": 5000
},
{
"epoch": 3.22,
"eval_loss": 0.18248289823532104,
"eval_runtime": 377.2331,
"eval_samples_per_second": 14.413,
"eval_steps_per_second": 0.901,
"eval_wer": 0.10971986625746419,
"step": 5000
},
{
"epoch": 3.29,
"learning_rate": 3.99099494838568e-05,
"loss": 0.0885,
"step": 5100
},
{
"epoch": 3.35,
"learning_rate": 3.969031407862948e-05,
"loss": 0.0951,
"step": 5200
},
{
"epoch": 3.42,
"learning_rate": 3.947067867340216e-05,
"loss": 0.0829,
"step": 5300
},
{
"epoch": 3.48,
"learning_rate": 3.925104326817483e-05,
"loss": 0.0864,
"step": 5400
},
{
"epoch": 3.55,
"learning_rate": 3.903140786294751e-05,
"loss": 0.0859,
"step": 5500
},
{
"epoch": 3.55,
"eval_loss": 0.1863715648651123,
"eval_runtime": 375.9814,
"eval_samples_per_second": 14.461,
"eval_steps_per_second": 0.904,
"eval_wer": 0.10721954758939865,
"step": 5500
},
{
"epoch": 3.61,
"learning_rate": 3.881177245772018e-05,
"loss": 0.0835,
"step": 5600
},
{
"epoch": 3.67,
"learning_rate": 3.859213705249287e-05,
"loss": 0.0818,
"step": 5700
},
{
"epoch": 3.74,
"learning_rate": 3.837250164726554e-05,
"loss": 0.0829,
"step": 5800
},
{
"epoch": 3.8,
"learning_rate": 3.815286624203822e-05,
"loss": 0.0811,
"step": 5900
},
{
"epoch": 3.87,
"learning_rate": 3.793323083681089e-05,
"loss": 0.0867,
"step": 6000
},
{
"epoch": 3.87,
"eval_loss": 0.18856672942638397,
"eval_runtime": 378.3989,
"eval_samples_per_second": 14.368,
"eval_steps_per_second": 0.899,
"eval_wer": 0.10813142851539902,
"step": 6000
},
{
"epoch": 3.93,
"learning_rate": 3.771359543158357e-05,
"loss": 0.084,
"step": 6100
},
{
"epoch": 4.0,
"learning_rate": 3.749396002635625e-05,
"loss": 0.0865,
"step": 6200
},
{
"epoch": 4.06,
"learning_rate": 3.7274324621128924e-05,
"loss": 0.0816,
"step": 6300
},
{
"epoch": 4.13,
"learning_rate": 3.70546892159016e-05,
"loss": 0.0808,
"step": 6400
},
{
"epoch": 4.19,
"learning_rate": 3.683505381067428e-05,
"loss": 0.0783,
"step": 6500
},
{
"epoch": 4.19,
"eval_loss": 0.18829604983329773,
"eval_runtime": 376.7422,
"eval_samples_per_second": 14.432,
"eval_steps_per_second": 0.902,
"eval_wer": 0.10633708217714022,
"step": 6500
},
{
"epoch": 4.26,
"learning_rate": 3.661541840544696e-05,
"loss": 0.0808,
"step": 6600
},
{
"epoch": 4.32,
"learning_rate": 3.6395783000219634e-05,
"loss": 0.0824,
"step": 6700
},
{
"epoch": 4.38,
"learning_rate": 3.617614759499231e-05,
"loss": 0.084,
"step": 6800
},
{
"epoch": 4.45,
"learning_rate": 3.595651218976499e-05,
"loss": 0.0831,
"step": 6900
},
{
"epoch": 4.51,
"learning_rate": 3.573687678453767e-05,
"loss": 0.0804,
"step": 7000
},
{
"epoch": 4.51,
"eval_loss": 0.18314477801322937,
"eval_runtime": 381.141,
"eval_samples_per_second": 14.265,
"eval_steps_per_second": 0.892,
"eval_wer": 0.10627825114965633,
"step": 7000
},
{
"epoch": 4.58,
"learning_rate": 3.551943773336262e-05,
"loss": 0.0816,
"step": 7100
},
{
"epoch": 4.64,
"learning_rate": 3.5299802328135294e-05,
"loss": 0.0814,
"step": 7200
},
{
"epoch": 4.71,
"learning_rate": 3.5080166922907974e-05,
"loss": 0.0773,
"step": 7300
},
{
"epoch": 4.77,
"learning_rate": 3.486053151768065e-05,
"loss": 0.086,
"step": 7400
},
{
"epoch": 4.84,
"learning_rate": 3.464089611245333e-05,
"loss": 0.0797,
"step": 7500
},
{
"epoch": 4.84,
"eval_loss": 0.18838059902191162,
"eval_runtime": 376.2869,
"eval_samples_per_second": 14.449,
"eval_steps_per_second": 0.904,
"eval_wer": 0.1058272132722798,
"step": 7500
},
{
"epoch": 4.9,
"learning_rate": 3.4421260707226005e-05,
"loss": 0.0841,
"step": 7600
},
{
"epoch": 4.96,
"learning_rate": 3.4201625301998684e-05,
"loss": 0.0817,
"step": 7700
},
{
"epoch": 5.03,
"learning_rate": 3.398198989677136e-05,
"loss": 0.0808,
"step": 7800
},
{
"epoch": 5.09,
"learning_rate": 3.376235449154404e-05,
"loss": 0.0758,
"step": 7900
},
{
"epoch": 5.16,
"learning_rate": 3.3542719086316715e-05,
"loss": 0.0705,
"step": 8000
},
{
"epoch": 5.16,
"eval_loss": 0.1801859736442566,
"eval_runtime": 378.8319,
"eval_samples_per_second": 14.352,
"eval_steps_per_second": 0.897,
"eval_wer": 0.10566052536107542,
"step": 8000
},
{
"epoch": 5.22,
"learning_rate": 3.3323083681089394e-05,
"loss": 0.0804,
"step": 8100
},
{
"epoch": 5.29,
"learning_rate": 3.310344827586207e-05,
"loss": 0.0753,
"step": 8200
},
{
"epoch": 5.35,
"learning_rate": 3.288381287063475e-05,
"loss": 0.072,
"step": 8300
},
{
"epoch": 5.42,
"learning_rate": 3.2664177465407425e-05,
"loss": 0.078,
"step": 8400
},
{
"epoch": 5.48,
"learning_rate": 3.2444542060180104e-05,
"loss": 0.0795,
"step": 8500
},
{
"epoch": 5.48,
"eval_loss": 0.18536211550235748,
"eval_runtime": 379.793,
"eval_samples_per_second": 14.316,
"eval_steps_per_second": 0.895,
"eval_wer": 0.10382695833782737,
"step": 8500
},
{
"epoch": 5.54,
"learning_rate": 3.222490665495278e-05,
"loss": 0.0748,
"step": 8600
},
{
"epoch": 5.61,
"learning_rate": 3.2005271249725456e-05,
"loss": 0.0731,
"step": 8700
},
{
"epoch": 5.67,
"learning_rate": 3.1785635844498135e-05,
"loss": 0.0766,
"step": 8800
},
{
"epoch": 5.74,
"learning_rate": 3.156600043927081e-05,
"loss": 0.072,
"step": 8900
},
{
"epoch": 5.8,
"learning_rate": 3.134636503404349e-05,
"loss": 0.0711,
"step": 9000
},
{
"epoch": 5.8,
"eval_loss": 0.17664773762226105,
"eval_runtime": 376.7989,
"eval_samples_per_second": 14.429,
"eval_steps_per_second": 0.902,
"eval_wer": 0.10324845323423573,
"step": 9000
},
{
"epoch": 5.87,
"learning_rate": 3.1126729628816166e-05,
"loss": 0.0951,
"step": 9100
},
{
"epoch": 5.93,
"learning_rate": 3.0907094223588845e-05,
"loss": 0.095,
"step": 9200
},
{
"epoch": 6.0,
"learning_rate": 3.068745881836152e-05,
"loss": 0.0945,
"step": 9300
},
{
"epoch": 6.06,
"learning_rate": 3.04678234131342e-05,
"loss": 0.0892,
"step": 9400
},
{
"epoch": 6.13,
"learning_rate": 3.0248188007906873e-05,
"loss": 0.0973,
"step": 9500
},
{
"epoch": 6.13,
"eval_loss": 0.16625218093395233,
"eval_runtime": 378.912,
"eval_samples_per_second": 14.349,
"eval_steps_per_second": 0.897,
"eval_wer": 0.10142469138223499,
"step": 9500
},
{
"epoch": 6.19,
"learning_rate": 3.0028552602679556e-05,
"loss": 0.0911,
"step": 9600
},
{
"epoch": 6.25,
"learning_rate": 2.9808917197452228e-05,
"loss": 0.0897,
"step": 9700
},
{
"epoch": 6.32,
"learning_rate": 2.958928179222491e-05,
"loss": 0.085,
"step": 9800
},
{
"epoch": 6.38,
"learning_rate": 2.9369646386997583e-05,
"loss": 0.0897,
"step": 9900
},
{
"epoch": 6.45,
"learning_rate": 2.9150010981770266e-05,
"loss": 0.087,
"step": 10000
},
{
"epoch": 6.45,
"eval_loss": 0.1664039045572281,
"eval_runtime": 378.5078,
"eval_samples_per_second": 14.364,
"eval_steps_per_second": 0.898,
"eval_wer": 0.10139527586849305,
"step": 10000
},
{
"epoch": 6.51,
"learning_rate": 2.893037557654294e-05,
"loss": 0.091,
"step": 10100
},
{
"epoch": 6.58,
"learning_rate": 2.871074017131562e-05,
"loss": 0.0877,
"step": 10200
},
{
"epoch": 6.64,
"learning_rate": 2.8491104766088293e-05,
"loss": 0.0865,
"step": 10300
},
{
"epoch": 6.71,
"learning_rate": 2.8271469360860976e-05,
"loss": 0.0904,
"step": 10400
},
{
"epoch": 6.77,
"learning_rate": 2.805183395563365e-05,
"loss": 0.0962,
"step": 10500
},
{
"epoch": 6.77,
"eval_loss": 0.16310061514377594,
"eval_runtime": 377.2788,
"eval_samples_per_second": 14.411,
"eval_steps_per_second": 0.901,
"eval_wer": 0.10092462764862188,
"step": 10500
},
{
"epoch": 6.83,
"learning_rate": 2.7832198550406324e-05,
"loss": 0.0946,
"step": 10600
},
{
"epoch": 6.9,
"learning_rate": 2.7612563145179004e-05,
"loss": 0.0894,
"step": 10700
},
{
"epoch": 6.96,
"learning_rate": 2.739292773995168e-05,
"loss": 0.0908,
"step": 10800
},
{
"epoch": 7.03,
"learning_rate": 2.717329233472436e-05,
"loss": 0.0864,
"step": 10900
},
{
"epoch": 7.09,
"learning_rate": 2.6953656929497035e-05,
"loss": 0.0857,
"step": 11000
},
{
"epoch": 7.09,
"eval_loss": 0.16591086983680725,
"eval_runtime": 375.5053,
"eval_samples_per_second": 14.479,
"eval_steps_per_second": 0.905,
"eval_wer": 0.10021865531881514,
"step": 11000
},
{
"epoch": 7.16,
"learning_rate": 2.6734021524269714e-05,
"loss": 0.0941,
"step": 11100
},
{
"epoch": 7.22,
"learning_rate": 2.651438611904239e-05,
"loss": 0.0853,
"step": 11200
},
{
"epoch": 7.29,
"learning_rate": 2.629475071381507e-05,
"loss": 0.0835,
"step": 11300
},
{
"epoch": 7.35,
"learning_rate": 2.6075115308587745e-05,
"loss": 0.0834,
"step": 11400
},
{
"epoch": 7.41,
"learning_rate": 2.5855479903360424e-05,
"loss": 0.0882,
"step": 11500
},
{
"epoch": 7.41,
"eval_loss": 0.16680683195590973,
"eval_runtime": 375.61,
"eval_samples_per_second": 14.475,
"eval_steps_per_second": 0.905,
"eval_wer": 0.10069910870993362,
"step": 11500
},
{
"epoch": 7.48,
"learning_rate": 2.56358444981331e-05,
"loss": 0.0817,
"step": 11600
},
{
"epoch": 7.54,
"learning_rate": 2.541620909290578e-05,
"loss": 0.0797,
"step": 11700
},
{
"epoch": 7.61,
"learning_rate": 2.5196573687678455e-05,
"loss": 0.0844,
"step": 11800
},
{
"epoch": 7.67,
"learning_rate": 2.497693828245113e-05,
"loss": 0.0812,
"step": 11900
},
{
"epoch": 7.74,
"learning_rate": 2.475730287722381e-05,
"loss": 0.0784,
"step": 12000
},
{
"epoch": 7.74,
"eval_loss": 0.16882041096687317,
"eval_runtime": 382.3729,
"eval_samples_per_second": 14.219,
"eval_steps_per_second": 0.889,
"eval_wer": 0.09955190367399766,
"step": 12000
},
{
"epoch": 7.8,
"learning_rate": 2.4537667471996486e-05,
"loss": 0.0828,
"step": 12100
},
{
"epoch": 7.87,
"learning_rate": 2.4318032066769165e-05,
"loss": 0.0841,
"step": 12200
},
{
"epoch": 7.93,
"learning_rate": 2.409839666154184e-05,
"loss": 0.0836,
"step": 12300
},
{
"epoch": 7.99,
"learning_rate": 2.387876125631452e-05,
"loss": 0.0763,
"step": 12400
},
{
"epoch": 8.06,
"learning_rate": 2.3659125851087196e-05,
"loss": 0.0838,
"step": 12500
},
{
"epoch": 8.06,
"eval_loss": 0.16750921308994293,
"eval_runtime": 379.814,
"eval_samples_per_second": 14.315,
"eval_steps_per_second": 0.895,
"eval_wer": 0.09836547795307245,
"step": 12500
},
{
"epoch": 8.12,
"learning_rate": 2.3439490445859875e-05,
"loss": 0.0792,
"step": 12600
},
{
"epoch": 8.19,
"learning_rate": 2.321985504063255e-05,
"loss": 0.079,
"step": 12700
},
{
"epoch": 8.25,
"learning_rate": 2.300021963540523e-05,
"loss": 0.0833,
"step": 12800
},
{
"epoch": 8.32,
"learning_rate": 2.2780584230177906e-05,
"loss": 0.0757,
"step": 12900
},
{
"epoch": 8.38,
"learning_rate": 2.2560948824950586e-05,
"loss": 0.0863,
"step": 13000
},
{
"epoch": 8.38,
"eval_loss": 0.16389821469783783,
"eval_runtime": 379.0093,
"eval_samples_per_second": 14.345,
"eval_steps_per_second": 0.897,
"eval_wer": 0.09786541421945934,
"step": 13000
},
{
"epoch": 8.45,
"learning_rate": 2.234131341972326e-05,
"loss": 0.0811,
"step": 13100
},
{
"epoch": 8.51,
"learning_rate": 2.2121678014495937e-05,
"loss": 0.0775,
"step": 13200
},
{
"epoch": 8.57,
"learning_rate": 2.1902042609268617e-05,
"loss": 0.0774,
"step": 13300
},
{
"epoch": 8.64,
"learning_rate": 2.1682407204041292e-05,
"loss": 0.0791,
"step": 13400
},
{
"epoch": 8.7,
"learning_rate": 2.1462771798813968e-05,
"loss": 0.0763,
"step": 13500
},
{
"epoch": 8.7,
"eval_loss": 0.16381791234016418,
"eval_runtime": 379.2251,
"eval_samples_per_second": 14.337,
"eval_steps_per_second": 0.897,
"eval_wer": 0.0980222969594164,
"step": 13500
},
{
"epoch": 8.77,
"learning_rate": 2.1243136393586648e-05,
"loss": 0.0814,
"step": 13600
},
{
"epoch": 8.83,
"learning_rate": 2.1023500988359323e-05,
"loss": 0.0852,
"step": 13700
},
{
"epoch": 8.9,
"learning_rate": 2.0803865583132003e-05,
"loss": 0.0839,
"step": 13800
},
{
"epoch": 8.96,
"learning_rate": 2.058423017790468e-05,
"loss": 0.0803,
"step": 13900
},
{
"epoch": 9.03,
"learning_rate": 2.0364594772677358e-05,
"loss": 0.0822,
"step": 14000
},
{
"epoch": 9.03,
"eval_loss": 0.17092539370059967,
"eval_runtime": 375.1283,
"eval_samples_per_second": 14.494,
"eval_steps_per_second": 0.906,
"eval_wer": 0.09723788325963113,
"step": 14000
},
{
"epoch": 9.09,
"learning_rate": 2.0144959367450034e-05,
"loss": 0.0793,
"step": 14100
},
{
"epoch": 9.16,
"learning_rate": 1.992532396222271e-05,
"loss": 0.0821,
"step": 14200
},
{
"epoch": 9.22,
"learning_rate": 1.970568855699539e-05,
"loss": 0.081,
"step": 14300
},
{
"epoch": 9.28,
"learning_rate": 1.9486053151768065e-05,
"loss": 0.0763,
"step": 14400
},
{
"epoch": 9.35,
"learning_rate": 1.9266417746540744e-05,
"loss": 0.0769,
"step": 14500
},
{
"epoch": 9.35,
"eval_loss": 0.16998609900474548,
"eval_runtime": 376.06,
"eval_samples_per_second": 14.458,
"eval_steps_per_second": 0.904,
"eval_wer": 0.09645346955984586,
"step": 14500
},
{
"epoch": 9.41,
"learning_rate": 1.904678234131342e-05,
"loss": 0.0781,
"step": 14600
},
{
"epoch": 9.48,
"learning_rate": 1.88271469360861e-05,
"loss": 0.0771,
"step": 14700
},
{
"epoch": 9.54,
"learning_rate": 1.8607511530858775e-05,
"loss": 0.0806,
"step": 14800
},
{
"epoch": 9.61,
"learning_rate": 1.8387876125631454e-05,
"loss": 0.074,
"step": 14900
},
{
"epoch": 9.67,
"learning_rate": 1.816824072040413e-05,
"loss": 0.0838,
"step": 15000
},
{
"epoch": 9.67,
"eval_loss": 0.17028485238552094,
"eval_runtime": 378.355,
"eval_samples_per_second": 14.37,
"eval_steps_per_second": 0.899,
"eval_wer": 0.09736535048584624,
"step": 15000
},
{
"epoch": 9.74,
"learning_rate": 1.794860531517681e-05,
"loss": 0.0737,
"step": 15100
},
{
"epoch": 9.8,
"learning_rate": 1.7728969909949485e-05,
"loss": 0.0697,
"step": 15200
},
{
"epoch": 9.86,
"learning_rate": 1.750933450472216e-05,
"loss": 0.0789,
"step": 15300
},
{
"epoch": 9.93,
"learning_rate": 1.7289699099494837e-05,
"loss": 0.0827,
"step": 15400
},
{
"epoch": 9.99,
"learning_rate": 1.7070063694267516e-05,
"loss": 0.0799,
"step": 15500
},
{
"epoch": 9.99,
"eval_loss": 0.16668923199176788,
"eval_runtime": 383.8462,
"eval_samples_per_second": 14.165,
"eval_steps_per_second": 0.886,
"eval_wer": 0.09569847137380254,
"step": 15500
},
{
"epoch": 10.06,
"learning_rate": 1.6850428289040192e-05,
"loss": 0.0752,
"step": 15600
},
{
"epoch": 10.12,
"learning_rate": 1.663079288381287e-05,
"loss": 0.0689,
"step": 15700
},
{
"epoch": 10.19,
"learning_rate": 1.6411157478585547e-05,
"loss": 0.0701,
"step": 15800
},
{
"epoch": 10.25,
"learning_rate": 1.6191522073358226e-05,
"loss": 0.0718,
"step": 15900
},
{
"epoch": 10.32,
"learning_rate": 1.5971886668130902e-05,
"loss": 0.0712,
"step": 16000
},
{
"epoch": 10.32,
"eval_loss": 0.17544205486774445,
"eval_runtime": 378.572,
"eval_samples_per_second": 14.362,
"eval_steps_per_second": 0.898,
"eval_wer": 0.09600243168246933,
"step": 16000
},
{
"epoch": 10.38,
"learning_rate": 1.575225126290358e-05,
"loss": 0.076,
"step": 16100
},
{
"epoch": 10.44,
"learning_rate": 1.5532615857676257e-05,
"loss": 0.0792,
"step": 16200
},
{
"epoch": 10.51,
"learning_rate": 1.5312980452448936e-05,
"loss": 0.0764,
"step": 16300
},
{
"epoch": 10.57,
"learning_rate": 1.5093345047221614e-05,
"loss": 0.0747,
"step": 16400
},
{
"epoch": 10.64,
"learning_rate": 1.4873709641994291e-05,
"loss": 0.0737,
"step": 16500
},
{
"epoch": 10.64,
"eval_loss": 0.17247016727924347,
"eval_runtime": 379.1086,
"eval_samples_per_second": 14.342,
"eval_steps_per_second": 0.897,
"eval_wer": 0.09679665055350192,
"step": 16500
},
{
"epoch": 10.7,
"learning_rate": 1.4654074236766967e-05,
"loss": 0.0704,
"step": 16600
},
{
"epoch": 10.77,
"learning_rate": 1.4434438831539645e-05,
"loss": 0.0746,
"step": 16700
},
{
"epoch": 10.83,
"learning_rate": 1.4214803426312322e-05,
"loss": 0.0747,
"step": 16800
},
{
"epoch": 10.9,
"learning_rate": 1.3995168021085e-05,
"loss": 0.0746,
"step": 16900
},
{
"epoch": 10.96,
"learning_rate": 1.3775532615857678e-05,
"loss": 0.0851,
"step": 17000
},
{
"epoch": 10.96,
"eval_loss": 0.17331218719482422,
"eval_runtime": 384.1505,
"eval_samples_per_second": 14.153,
"eval_steps_per_second": 0.885,
"eval_wer": 0.09575730240128644,
"step": 17000
},
{
"epoch": 11.03,
"learning_rate": 1.3555897210630355e-05,
"loss": 0.0776,
"step": 17100
},
{
"epoch": 11.09,
"learning_rate": 1.3336261805403031e-05,
"loss": 0.0712,
"step": 17200
},
{
"epoch": 11.15,
"learning_rate": 1.3116626400175708e-05,
"loss": 0.0798,
"step": 17300
},
{
"epoch": 11.22,
"learning_rate": 1.2896990994948386e-05,
"loss": 0.0752,
"step": 17400
},
{
"epoch": 11.28,
"learning_rate": 1.2677355589721064e-05,
"loss": 0.076,
"step": 17500
},
{
"epoch": 11.28,
"eval_loss": 0.16816101968288422,
"eval_runtime": 375.8451,
"eval_samples_per_second": 14.466,
"eval_steps_per_second": 0.905,
"eval_wer": 0.09541412140763038,
"step": 17500
},
{
"epoch": 11.35,
"learning_rate": 1.2457720184493741e-05,
"loss": 0.0785,
"step": 17600
},
{
"epoch": 11.41,
"learning_rate": 1.2238084779266417e-05,
"loss": 0.0787,
"step": 17700
},
{
"epoch": 11.48,
"learning_rate": 1.2018449374039095e-05,
"loss": 0.073,
"step": 17800
},
{
"epoch": 11.54,
"learning_rate": 1.1798813968811772e-05,
"loss": 0.0686,
"step": 17900
},
{
"epoch": 11.61,
"learning_rate": 1.157917856358445e-05,
"loss": 0.0712,
"step": 18000
},
{
"epoch": 11.61,
"eval_loss": 0.17131488025188446,
"eval_runtime": 375.69,
"eval_samples_per_second": 14.472,
"eval_steps_per_second": 0.905,
"eval_wer": 0.09434535774167296,
"step": 18000
},
{
"epoch": 11.67,
"learning_rate": 1.1359543158357127e-05,
"loss": 0.0737,
"step": 18100
},
{
"epoch": 11.73,
"learning_rate": 1.1139907753129805e-05,
"loss": 0.0781,
"step": 18200
},
{
"epoch": 11.8,
"learning_rate": 1.0920272347902482e-05,
"loss": 0.084,
"step": 18300
},
{
"epoch": 11.86,
"learning_rate": 1.0702833296727433e-05,
"loss": 0.0841,
"step": 18400
},
{
"epoch": 11.93,
"learning_rate": 1.048319789150011e-05,
"loss": 0.0745,
"step": 18500
},
{
"epoch": 11.93,
"eval_loss": 0.16615267097949982,
"eval_runtime": 379.2398,
"eval_samples_per_second": 14.337,
"eval_steps_per_second": 0.897,
"eval_wer": 0.09507094041397433,
"step": 18500
},
{
"epoch": 11.99,
"learning_rate": 1.0263562486272788e-05,
"loss": 0.0881,
"step": 18600
},
{
"epoch": 12.06,
"learning_rate": 1.0043927081045465e-05,
"loss": 0.0801,
"step": 18700
},
{
"epoch": 12.12,
"learning_rate": 9.824291675818143e-06,
"loss": 0.0863,
"step": 18800
},
{
"epoch": 12.19,
"learning_rate": 9.60465627059082e-06,
"loss": 0.0786,
"step": 18900
},
{
"epoch": 12.25,
"learning_rate": 9.385020865363498e-06,
"loss": 0.0864,
"step": 19000
},
{
"epoch": 12.25,
"eval_loss": 0.16919353604316711,
"eval_runtime": 377.3159,
"eval_samples_per_second": 14.41,
"eval_steps_per_second": 0.901,
"eval_wer": 0.09465912322158707,
"step": 19000
},
{
"epoch": 12.31,
"learning_rate": 9.165385460136176e-06,
"loss": 0.0973,
"step": 19100
},
{
"epoch": 12.38,
"learning_rate": 8.945750054908853e-06,
"loss": 0.0999,
"step": 19200
},
{
"epoch": 12.44,
"learning_rate": 8.726114649681529e-06,
"loss": 0.0922,
"step": 19300
},
{
"epoch": 12.51,
"learning_rate": 8.506479244454206e-06,
"loss": 0.0918,
"step": 19400
},
{
"epoch": 12.57,
"learning_rate": 8.286843839226884e-06,
"loss": 0.0937,
"step": 19500
},
{
"epoch": 12.57,
"eval_loss": 0.16241294145584106,
"eval_runtime": 377.2732,
"eval_samples_per_second": 14.411,
"eval_steps_per_second": 0.901,
"eval_wer": 0.09427672154294174,
"step": 19500
},
{
"epoch": 12.64,
"learning_rate": 8.067208433999562e-06,
"loss": 0.0916,
"step": 19600
},
{
"epoch": 12.7,
"learning_rate": 7.847573028772239e-06,
"loss": 0.0941,
"step": 19700
},
{
"epoch": 12.77,
"learning_rate": 7.627937623544916e-06,
"loss": 0.0919,
"step": 19800
},
{
"epoch": 12.83,
"learning_rate": 7.408302218317593e-06,
"loss": 0.0943,
"step": 19900
},
{
"epoch": 12.89,
"learning_rate": 7.188666813090271e-06,
"loss": 0.0915,
"step": 20000
},
{
"epoch": 12.89,
"eval_loss": 0.1677888035774231,
"eval_runtime": 382.7478,
"eval_samples_per_second": 14.205,
"eval_steps_per_second": 0.888,
"eval_wer": 0.09419828017296322,
"step": 20000
},
{
"epoch": 12.96,
"learning_rate": 6.9690314078629485e-06,
"loss": 0.0882,
"step": 20100
},
{
"epoch": 13.02,
"learning_rate": 6.749396002635624e-06,
"loss": 0.0974,
"step": 20200
},
{
"epoch": 13.09,
"learning_rate": 6.529760597408302e-06,
"loss": 0.0955,
"step": 20300
},
{
"epoch": 13.15,
"learning_rate": 6.3101251921809794e-06,
"loss": 0.0992,
"step": 20400
},
{
"epoch": 13.22,
"learning_rate": 6.092686141005931e-06,
"loss": 0.0926,
"step": 20500
},
{
"epoch": 13.22,
"eval_loss": 0.16411229968070984,
"eval_runtime": 381.1998,
"eval_samples_per_second": 14.263,
"eval_steps_per_second": 0.892,
"eval_wer": 0.09449243531038269,
"step": 20500
},
{
"epoch": 13.28,
"learning_rate": 5.873050735778607e-06,
"loss": 0.0939,
"step": 20600
},
{
"epoch": 13.35,
"learning_rate": 5.655611684603558e-06,
"loss": 0.0974,
"step": 20700
},
{
"epoch": 13.41,
"learning_rate": 5.435976279376235e-06,
"loss": 0.0957,
"step": 20800
},
{
"epoch": 13.48,
"learning_rate": 5.216340874148913e-06,
"loss": 0.0909,
"step": 20900
},
{
"epoch": 13.54,
"learning_rate": 4.99670546892159e-06,
"loss": 0.0912,
"step": 21000
},
{
"epoch": 13.54,
"eval_loss": 0.1665380299091339,
"eval_runtime": 376.5271,
"eval_samples_per_second": 14.44,
"eval_steps_per_second": 0.903,
"eval_wer": 0.09374724229558669,
"step": 21000
},
{
"epoch": 13.6,
"learning_rate": 4.777070063694268e-06,
"loss": 0.09,
"step": 21100
},
{
"epoch": 13.67,
"learning_rate": 4.557434658466945e-06,
"loss": 0.0938,
"step": 21200
},
{
"epoch": 13.73,
"learning_rate": 4.337799253239622e-06,
"loss": 0.0867,
"step": 21300
},
{
"epoch": 13.8,
"learning_rate": 4.1181638480123e-06,
"loss": 0.0844,
"step": 21400
},
{
"epoch": 13.86,
"learning_rate": 3.898528442784977e-06,
"loss": 0.0917,
"step": 21500
},
{
"epoch": 13.86,
"eval_loss": 0.16477684676647186,
"eval_runtime": 375.5992,
"eval_samples_per_second": 14.476,
"eval_steps_per_second": 0.905,
"eval_wer": 0.09364919058311354,
"step": 21500
},
{
"epoch": 13.93,
"learning_rate": 3.678893037557654e-06,
"loss": 0.0936,
"step": 21600
},
{
"epoch": 13.99,
"learning_rate": 3.4592576323303315e-06,
"loss": 0.0902,
"step": 21700
},
{
"epoch": 14.06,
"learning_rate": 3.239622227103009e-06,
"loss": 0.0994,
"step": 21800
},
{
"epoch": 14.12,
"learning_rate": 3.0199868218756866e-06,
"loss": 0.0948,
"step": 21900
},
{
"epoch": 14.18,
"learning_rate": 2.8003514166483638e-06,
"loss": 0.094,
"step": 22000
},
{
"epoch": 14.18,
"eval_loss": 0.16351106762886047,
"eval_runtime": 381.719,
"eval_samples_per_second": 14.243,
"eval_steps_per_second": 0.891,
"eval_wer": 0.09346289232941453,
"step": 22000
},
{
"epoch": 14.25,
"learning_rate": 2.5807160114210413e-06,
"loss": 0.0943,
"step": 22100
},
{
"epoch": 14.31,
"learning_rate": 2.3610806061937184e-06,
"loss": 0.0941,
"step": 22200
},
{
"epoch": 14.38,
"learning_rate": 2.141445200966396e-06,
"loss": 0.0868,
"step": 22300
},
{
"epoch": 14.44,
"learning_rate": 1.9218097957390735e-06,
"loss": 0.0896,
"step": 22400
},
{
"epoch": 14.51,
"learning_rate": 1.7021743905117505e-06,
"loss": 0.0864,
"step": 22500
},
{
"epoch": 14.51,
"eval_loss": 0.16783463954925537,
"eval_runtime": 377.482,
"eval_samples_per_second": 14.403,
"eval_steps_per_second": 0.901,
"eval_wer": 0.09338445095943601,
"step": 22500
},
{
"epoch": 14.57,
"learning_rate": 1.482538985284428e-06,
"loss": 0.0935,
"step": 22600
},
{
"epoch": 14.64,
"learning_rate": 1.2629035800571053e-06,
"loss": 0.0913,
"step": 22700
},
{
"epoch": 14.7,
"learning_rate": 1.0432681748297827e-06,
"loss": 0.093,
"step": 22800
},
{
"epoch": 14.76,
"learning_rate": 8.236327696024599e-07,
"loss": 0.0871,
"step": 22900
},
{
"epoch": 14.83,
"learning_rate": 6.039973643751374e-07,
"loss": 0.0899,
"step": 23000
},
{
"epoch": 14.83,
"eval_loss": 0.16628585755825043,
"eval_runtime": 378.1045,
"eval_samples_per_second": 14.38,
"eval_steps_per_second": 0.899,
"eval_wer": 0.09316873719199506,
"step": 23000
},
{
"epoch": 14.89,
"learning_rate": 3.8436195914781466e-07,
"loss": 0.099,
"step": 23100
},
{
"epoch": 14.96,
"learning_rate": 1.64726553920492e-07,
"loss": 0.0964,
"step": 23200
},
{
"epoch": 15.0,
"step": 23265,
"total_flos": 1.618046661664003e+20,
"train_loss": 0.08625626374438786,
"train_runtime": 75520.7008,
"train_samples_per_second": 9.861,
"train_steps_per_second": 0.308
}
],
"max_steps": 23265,
"num_train_epochs": 15,
"total_flos": 1.618046661664003e+20,
"trial_name": null,
"trial_params": null
}