xls-r-300m-sv / trainer_state.json
hf-test's picture
End of training
f7ee644
raw
history blame
29.7 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 49.99782451051487,
"global_step": 17200,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.29,
"learning_rate": 3.7125e-06,
"loss": 12.2893,
"step": 100
},
{
"epoch": 0.58,
"learning_rate": 7.4625e-06,
"loss": 6.2066,
"step": 200
},
{
"epoch": 0.87,
"learning_rate": 1.1212499999999998e-05,
"loss": 4.1421,
"step": 300
},
{
"epoch": 1.16,
"learning_rate": 1.49625e-05,
"loss": 3.6807,
"step": 400
},
{
"epoch": 1.45,
"learning_rate": 1.8712499999999997e-05,
"loss": 3.3349,
"step": 500
},
{
"epoch": 1.45,
"eval_loss": 3.2857718467712402,
"eval_runtime": 141.2158,
"eval_samples_per_second": 32.716,
"eval_steps_per_second": 4.093,
"eval_wer": 1.0,
"step": 500
},
{
"epoch": 1.74,
"learning_rate": 2.2462499999999997e-05,
"loss": 3.1383,
"step": 600
},
{
"epoch": 2.03,
"learning_rate": 2.6212499999999997e-05,
"loss": 3.0909,
"step": 700
},
{
"epoch": 2.32,
"learning_rate": 2.99625e-05,
"loss": 3.0188,
"step": 800
},
{
"epoch": 2.61,
"learning_rate": 3.37125e-05,
"loss": 2.9601,
"step": 900
},
{
"epoch": 2.91,
"learning_rate": 3.7462499999999996e-05,
"loss": 2.9298,
"step": 1000
},
{
"epoch": 2.91,
"eval_loss": 2.922468662261963,
"eval_runtime": 142.3546,
"eval_samples_per_second": 32.454,
"eval_steps_per_second": 4.06,
"eval_wer": 1.000030594138163,
"step": 1000
},
{
"epoch": 3.2,
"learning_rate": 4.12125e-05,
"loss": 2.9181,
"step": 1100
},
{
"epoch": 3.49,
"learning_rate": 4.4962499999999995e-05,
"loss": 2.8449,
"step": 1200
},
{
"epoch": 3.78,
"learning_rate": 4.871249999999999e-05,
"loss": 2.7039,
"step": 1300
},
{
"epoch": 4.07,
"learning_rate": 5.2462499999999994e-05,
"loss": 2.3828,
"step": 1400
},
{
"epoch": 4.36,
"learning_rate": 5.62125e-05,
"loss": 2.0839,
"step": 1500
},
{
"epoch": 4.36,
"eval_loss": 1.1546391248703003,
"eval_runtime": 141.9654,
"eval_samples_per_second": 32.543,
"eval_steps_per_second": 4.071,
"eval_wer": 0.829498868016888,
"step": 1500
},
{
"epoch": 4.65,
"learning_rate": 5.9962499999999994e-05,
"loss": 1.9444,
"step": 1600
},
{
"epoch": 4.94,
"learning_rate": 6.37125e-05,
"loss": 1.8594,
"step": 1700
},
{
"epoch": 5.23,
"learning_rate": 6.746249999999999e-05,
"loss": 1.8026,
"step": 1800
},
{
"epoch": 5.52,
"learning_rate": 7.121249999999999e-05,
"loss": 1.7341,
"step": 1900
},
{
"epoch": 5.81,
"learning_rate": 7.49625e-05,
"loss": 1.7093,
"step": 2000
},
{
"epoch": 5.81,
"eval_loss": 0.6826953887939453,
"eval_runtime": 142.0245,
"eval_samples_per_second": 32.53,
"eval_steps_per_second": 4.07,
"eval_wer": 0.5701217646698893,
"step": 2000
},
{
"epoch": 6.1,
"learning_rate": 7.451151315789474e-05,
"loss": 1.6759,
"step": 2100
},
{
"epoch": 6.39,
"learning_rate": 7.402302631578947e-05,
"loss": 1.6272,
"step": 2200
},
{
"epoch": 6.68,
"learning_rate": 7.352960526315788e-05,
"loss": 1.6174,
"step": 2300
},
{
"epoch": 6.97,
"learning_rate": 7.30361842105263e-05,
"loss": 1.6106,
"step": 2400
},
{
"epoch": 7.27,
"learning_rate": 7.254276315789473e-05,
"loss": 1.5855,
"step": 2500
},
{
"epoch": 7.27,
"eval_loss": 0.5596911311149597,
"eval_runtime": 141.6166,
"eval_samples_per_second": 32.623,
"eval_steps_per_second": 4.081,
"eval_wer": 0.49467661995961576,
"step": 2500
},
{
"epoch": 7.56,
"learning_rate": 7.204934210526316e-05,
"loss": 1.5606,
"step": 2600
},
{
"epoch": 7.85,
"learning_rate": 7.155592105263157e-05,
"loss": 1.5163,
"step": 2700
},
{
"epoch": 8.14,
"learning_rate": 7.10625e-05,
"loss": 1.5371,
"step": 2800
},
{
"epoch": 8.43,
"learning_rate": 7.056907894736841e-05,
"loss": 1.5072,
"step": 2900
},
{
"epoch": 8.72,
"learning_rate": 7.007565789473684e-05,
"loss": 1.4831,
"step": 3000
},
{
"epoch": 8.72,
"eval_loss": 0.49225884675979614,
"eval_runtime": 141.5992,
"eval_samples_per_second": 32.627,
"eval_steps_per_second": 4.082,
"eval_wer": 0.45267086826164105,
"step": 3000
},
{
"epoch": 9.01,
"learning_rate": 6.958717105263157e-05,
"loss": 1.5036,
"step": 3100
},
{
"epoch": 9.3,
"learning_rate": 6.909375e-05,
"loss": 1.4641,
"step": 3200
},
{
"epoch": 9.59,
"learning_rate": 6.860032894736842e-05,
"loss": 1.4493,
"step": 3300
},
{
"epoch": 9.88,
"learning_rate": 6.810690789473684e-05,
"loss": 1.4722,
"step": 3400
},
{
"epoch": 10.17,
"learning_rate": 6.761348684210526e-05,
"loss": 1.4416,
"step": 3500
},
{
"epoch": 10.17,
"eval_loss": 0.46698251366615295,
"eval_runtime": 142.3359,
"eval_samples_per_second": 32.458,
"eval_steps_per_second": 4.061,
"eval_wer": 0.4269717922046136,
"step": 3500
},
{
"epoch": 10.46,
"learning_rate": 6.712006578947368e-05,
"loss": 1.4211,
"step": 3600
},
{
"epoch": 10.75,
"learning_rate": 6.66266447368421e-05,
"loss": 1.4086,
"step": 3700
},
{
"epoch": 11.05,
"learning_rate": 6.613322368421052e-05,
"loss": 1.4385,
"step": 3800
},
{
"epoch": 11.34,
"learning_rate": 6.563980263157894e-05,
"loss": 1.3916,
"step": 3900
},
{
"epoch": 11.63,
"learning_rate": 6.514638157894736e-05,
"loss": 1.3848,
"step": 4000
},
{
"epoch": 11.63,
"eval_loss": 0.43410158157348633,
"eval_runtime": 142.891,
"eval_samples_per_second": 32.332,
"eval_steps_per_second": 4.045,
"eval_wer": 0.39799914336413145,
"step": 4000
},
{
"epoch": 11.92,
"learning_rate": 6.465296052631578e-05,
"loss": 1.3851,
"step": 4100
},
{
"epoch": 12.21,
"learning_rate": 6.415953947368421e-05,
"loss": 1.3951,
"step": 4200
},
{
"epoch": 12.5,
"learning_rate": 6.366611842105262e-05,
"loss": 1.377,
"step": 4300
},
{
"epoch": 12.79,
"learning_rate": 6.317269736842105e-05,
"loss": 1.3609,
"step": 4400
},
{
"epoch": 13.08,
"learning_rate": 6.267927631578946e-05,
"loss": 1.3749,
"step": 4500
},
{
"epoch": 13.08,
"eval_loss": 0.42033153772354126,
"eval_runtime": 142.2303,
"eval_samples_per_second": 32.483,
"eval_steps_per_second": 4.064,
"eval_wer": 0.40105855718044425,
"step": 4500
},
{
"epoch": 13.37,
"learning_rate": 6.218585526315789e-05,
"loss": 1.3346,
"step": 4600
},
{
"epoch": 13.66,
"learning_rate": 6.16924342105263e-05,
"loss": 1.3528,
"step": 4700
},
{
"epoch": 13.95,
"learning_rate": 6.119901315789473e-05,
"loss": 1.3348,
"step": 4800
},
{
"epoch": 14.24,
"learning_rate": 6.070559210526316e-05,
"loss": 1.3425,
"step": 4900
},
{
"epoch": 14.53,
"learning_rate": 6.021217105263157e-05,
"loss": 1.3311,
"step": 5000
},
{
"epoch": 14.53,
"eval_loss": 0.4309719204902649,
"eval_runtime": 143.1747,
"eval_samples_per_second": 32.268,
"eval_steps_per_second": 4.037,
"eval_wer": 0.3961023067980175,
"step": 5000
},
{
"epoch": 14.82,
"learning_rate": 5.971875e-05,
"loss": 1.3272,
"step": 5100
},
{
"epoch": 15.12,
"learning_rate": 5.922532894736842e-05,
"loss": 1.3359,
"step": 5200
},
{
"epoch": 15.41,
"learning_rate": 5.873190789473683e-05,
"loss": 1.3071,
"step": 5300
},
{
"epoch": 15.7,
"learning_rate": 5.823848684210526e-05,
"loss": 1.3136,
"step": 5400
},
{
"epoch": 15.99,
"learning_rate": 5.7745065789473685e-05,
"loss": 1.317,
"step": 5500
},
{
"epoch": 15.99,
"eval_loss": 0.3898037374019623,
"eval_runtime": 142.672,
"eval_samples_per_second": 32.382,
"eval_steps_per_second": 4.051,
"eval_wer": 0.43217279569234535,
"step": 5500
},
{
"epoch": 16.28,
"learning_rate": 5.72516447368421e-05,
"loss": 1.2904,
"step": 5600
},
{
"epoch": 16.57,
"learning_rate": 5.675822368421052e-05,
"loss": 1.3134,
"step": 5700
},
{
"epoch": 16.86,
"learning_rate": 5.6264802631578946e-05,
"loss": 1.2875,
"step": 5800
},
{
"epoch": 17.15,
"learning_rate": 5.577138157894736e-05,
"loss": 1.2874,
"step": 5900
},
{
"epoch": 17.44,
"learning_rate": 5.52828947368421e-05,
"loss": 1.2799,
"step": 6000
},
{
"epoch": 17.44,
"eval_loss": 0.3805598020553589,
"eval_runtime": 143.0363,
"eval_samples_per_second": 32.3,
"eval_steps_per_second": 4.041,
"eval_wer": 0.357247751330845,
"step": 6000
},
{
"epoch": 17.73,
"learning_rate": 5.478947368421052e-05,
"loss": 1.2704,
"step": 6100
},
{
"epoch": 18.02,
"learning_rate": 5.429605263157895e-05,
"loss": 1.2974,
"step": 6200
},
{
"epoch": 18.31,
"learning_rate": 5.380263157894736e-05,
"loss": 1.2545,
"step": 6300
},
{
"epoch": 18.6,
"learning_rate": 5.330921052631578e-05,
"loss": 1.2781,
"step": 6400
},
{
"epoch": 18.89,
"learning_rate": 5.281578947368421e-05,
"loss": 1.2771,
"step": 6500
},
{
"epoch": 18.89,
"eval_loss": 0.38284429907798767,
"eval_runtime": 142.0565,
"eval_samples_per_second": 32.522,
"eval_steps_per_second": 4.069,
"eval_wer": 0.34268494156519613,
"step": 6500
},
{
"epoch": 19.19,
"learning_rate": 5.232236842105262e-05,
"loss": 1.2641,
"step": 6600
},
{
"epoch": 19.48,
"learning_rate": 5.182894736842105e-05,
"loss": 1.241,
"step": 6700
},
{
"epoch": 19.77,
"learning_rate": 5.133552631578947e-05,
"loss": 1.2456,
"step": 6800
},
{
"epoch": 20.06,
"learning_rate": 5.084210526315789e-05,
"loss": 1.2597,
"step": 6900
},
{
"epoch": 20.35,
"learning_rate": 5.034868421052631e-05,
"loss": 1.2451,
"step": 7000
},
{
"epoch": 20.35,
"eval_loss": 0.37021246552467346,
"eval_runtime": 141.2064,
"eval_samples_per_second": 32.718,
"eval_steps_per_second": 4.093,
"eval_wer": 0.3358930428929817,
"step": 7000
},
{
"epoch": 20.64,
"learning_rate": 4.9855263157894736e-05,
"loss": 1.2458,
"step": 7100
},
{
"epoch": 20.93,
"learning_rate": 4.936184210526315e-05,
"loss": 1.2285,
"step": 7200
},
{
"epoch": 21.22,
"learning_rate": 4.886842105263158e-05,
"loss": 1.2305,
"step": 7300
},
{
"epoch": 21.51,
"learning_rate": 4.8375e-05,
"loss": 1.2228,
"step": 7400
},
{
"epoch": 21.8,
"learning_rate": 4.7881578947368424e-05,
"loss": 1.2182,
"step": 7500
},
{
"epoch": 21.8,
"eval_loss": 0.3685418367385864,
"eval_runtime": 141.7574,
"eval_samples_per_second": 32.591,
"eval_steps_per_second": 4.077,
"eval_wer": 0.3270207428256746,
"step": 7500
},
{
"epoch": 22.09,
"learning_rate": 4.738815789473684e-05,
"loss": 1.2297,
"step": 7600
},
{
"epoch": 22.38,
"learning_rate": 4.689473684210526e-05,
"loss": 1.2252,
"step": 7700
},
{
"epoch": 22.67,
"learning_rate": 4.6401315789473684e-05,
"loss": 1.2153,
"step": 7800
},
{
"epoch": 22.96,
"learning_rate": 4.59078947368421e-05,
"loss": 1.2132,
"step": 7900
},
{
"epoch": 23.26,
"learning_rate": 4.541940789473684e-05,
"loss": 1.2152,
"step": 8000
},
{
"epoch": 23.26,
"eval_loss": 0.3650066554546356,
"eval_runtime": 141.8703,
"eval_samples_per_second": 32.565,
"eval_steps_per_second": 4.074,
"eval_wer": 0.33078382181973937,
"step": 8000
},
{
"epoch": 23.55,
"learning_rate": 4.492598684210526e-05,
"loss": 1.185,
"step": 8100
},
{
"epoch": 23.84,
"learning_rate": 4.443256578947369e-05,
"loss": 1.2039,
"step": 8200
},
{
"epoch": 24.13,
"learning_rate": 4.39391447368421e-05,
"loss": 1.2078,
"step": 8300
},
{
"epoch": 24.42,
"learning_rate": 4.344572368421052e-05,
"loss": 1.1991,
"step": 8400
},
{
"epoch": 24.71,
"learning_rate": 4.295230263157895e-05,
"loss": 1.1837,
"step": 8500
},
{
"epoch": 24.71,
"eval_loss": 0.35678917169570923,
"eval_runtime": 142.5499,
"eval_samples_per_second": 32.41,
"eval_steps_per_second": 4.055,
"eval_wer": 0.3186991372453038,
"step": 8500
},
{
"epoch": 25.0,
"learning_rate": 4.2463815789473676e-05,
"loss": 1.1982,
"step": 8600
},
{
"epoch": 25.29,
"learning_rate": 4.19703947368421e-05,
"loss": 1.2087,
"step": 8700
},
{
"epoch": 25.58,
"learning_rate": 4.1476973684210523e-05,
"loss": 1.1841,
"step": 8800
},
{
"epoch": 25.87,
"learning_rate": 4.098355263157895e-05,
"loss": 1.1694,
"step": 8900
},
{
"epoch": 26.16,
"learning_rate": 4.0490131578947364e-05,
"loss": 1.1721,
"step": 9000
},
{
"epoch": 26.16,
"eval_loss": 0.3659152686595917,
"eval_runtime": 141.7914,
"eval_samples_per_second": 32.583,
"eval_steps_per_second": 4.076,
"eval_wer": 0.3249403414305819,
"step": 9000
},
{
"epoch": 26.45,
"learning_rate": 3.9996710526315784e-05,
"loss": 1.1585,
"step": 9100
},
{
"epoch": 26.74,
"learning_rate": 3.950328947368421e-05,
"loss": 1.1635,
"step": 9200
},
{
"epoch": 27.03,
"learning_rate": 3.9009868421052624e-05,
"loss": 1.193,
"step": 9300
},
{
"epoch": 27.32,
"learning_rate": 3.851644736842105e-05,
"loss": 1.1751,
"step": 9400
},
{
"epoch": 27.61,
"learning_rate": 3.802302631578947e-05,
"loss": 1.1764,
"step": 9500
},
{
"epoch": 27.61,
"eval_loss": 0.3546680510044098,
"eval_runtime": 142.2113,
"eval_samples_per_second": 32.487,
"eval_steps_per_second": 4.064,
"eval_wer": 0.3145077403169553,
"step": 9500
},
{
"epoch": 27.91,
"learning_rate": 3.752960526315789e-05,
"loss": 1.1551,
"step": 9600
},
{
"epoch": 28.2,
"learning_rate": 3.703618421052631e-05,
"loss": 1.1571,
"step": 9700
},
{
"epoch": 28.49,
"learning_rate": 3.654276315789473e-05,
"loss": 1.1472,
"step": 9800
},
{
"epoch": 28.78,
"learning_rate": 3.604934210526316e-05,
"loss": 1.1515,
"step": 9900
},
{
"epoch": 29.07,
"learning_rate": 3.555592105263158e-05,
"loss": 1.1606,
"step": 10000
},
{
"epoch": 29.07,
"eval_loss": 0.35135164856910706,
"eval_runtime": 141.0593,
"eval_samples_per_second": 32.752,
"eval_steps_per_second": 4.098,
"eval_wer": 0.310377531664933,
"step": 10000
},
{
"epoch": 29.36,
"learning_rate": 3.50625e-05,
"loss": 1.1487,
"step": 10100
},
{
"epoch": 29.65,
"learning_rate": 3.456907894736842e-05,
"loss": 1.1433,
"step": 10200
},
{
"epoch": 29.94,
"learning_rate": 3.407565789473684e-05,
"loss": 1.1556,
"step": 10300
},
{
"epoch": 30.23,
"learning_rate": 3.358223684210526e-05,
"loss": 1.1542,
"step": 10400
},
{
"epoch": 30.52,
"learning_rate": 3.308881578947368e-05,
"loss": 1.1431,
"step": 10500
},
{
"epoch": 30.52,
"eval_loss": 0.3469483554363251,
"eval_runtime": 141.5722,
"eval_samples_per_second": 32.634,
"eval_steps_per_second": 4.083,
"eval_wer": 0.3061555405984213,
"step": 10500
},
{
"epoch": 30.81,
"learning_rate": 3.25953947368421e-05,
"loss": 1.126,
"step": 10600
},
{
"epoch": 31.1,
"learning_rate": 3.2101973684210526e-05,
"loss": 1.1384,
"step": 10700
},
{
"epoch": 31.39,
"learning_rate": 3.1608552631578946e-05,
"loss": 1.1208,
"step": 10800
},
{
"epoch": 31.68,
"learning_rate": 3.1115131578947366e-05,
"loss": 1.1218,
"step": 10900
},
{
"epoch": 31.97,
"learning_rate": 3.0621710526315786e-05,
"loss": 1.1047,
"step": 11000
},
{
"epoch": 31.97,
"eval_loss": 0.3313155770301819,
"eval_runtime": 142.8254,
"eval_samples_per_second": 32.347,
"eval_steps_per_second": 4.047,
"eval_wer": 0.2978951232943768,
"step": 11000
},
{
"epoch": 32.27,
"learning_rate": 3.0128289473684206e-05,
"loss": 1.1228,
"step": 11100
},
{
"epoch": 32.56,
"learning_rate": 2.963486842105263e-05,
"loss": 1.1065,
"step": 11200
},
{
"epoch": 32.85,
"learning_rate": 2.914144736842105e-05,
"loss": 1.1134,
"step": 11300
},
{
"epoch": 33.14,
"learning_rate": 2.864802631578947e-05,
"loss": 1.1138,
"step": 11400
},
{
"epoch": 33.43,
"learning_rate": 2.8154605263157894e-05,
"loss": 1.1315,
"step": 11500
},
{
"epoch": 33.43,
"eval_loss": 0.3297792375087738,
"eval_runtime": 142.5868,
"eval_samples_per_second": 32.401,
"eval_steps_per_second": 4.054,
"eval_wer": 0.29924126537355444,
"step": 11500
},
{
"epoch": 33.72,
"learning_rate": 2.7661184210526314e-05,
"loss": 1.0891,
"step": 11600
},
{
"epoch": 34.01,
"learning_rate": 2.7167763157894734e-05,
"loss": 1.1259,
"step": 11700
},
{
"epoch": 34.3,
"learning_rate": 2.6674342105263157e-05,
"loss": 1.1049,
"step": 11800
},
{
"epoch": 34.59,
"learning_rate": 2.6180921052631577e-05,
"loss": 1.0896,
"step": 11900
},
{
"epoch": 34.88,
"learning_rate": 2.56875e-05,
"loss": 1.1022,
"step": 12000
},
{
"epoch": 34.88,
"eval_loss": 0.3296017050743103,
"eval_runtime": 141.6872,
"eval_samples_per_second": 32.607,
"eval_steps_per_second": 4.079,
"eval_wer": 0.2972526463929511,
"step": 12000
},
{
"epoch": 35.17,
"learning_rate": 2.5194078947368418e-05,
"loss": 1.1032,
"step": 12100
},
{
"epoch": 35.46,
"learning_rate": 2.4700657894736838e-05,
"loss": 1.0896,
"step": 12200
},
{
"epoch": 35.75,
"learning_rate": 2.420723684210526e-05,
"loss": 1.0879,
"step": 12300
},
{
"epoch": 36.05,
"learning_rate": 2.3718749999999997e-05,
"loss": 1.122,
"step": 12400
},
{
"epoch": 36.34,
"learning_rate": 2.322532894736842e-05,
"loss": 1.0935,
"step": 12500
},
{
"epoch": 36.34,
"eval_loss": 0.327805757522583,
"eval_runtime": 141.8254,
"eval_samples_per_second": 32.575,
"eval_steps_per_second": 4.075,
"eval_wer": 0.2926023373921557,
"step": 12500
},
{
"epoch": 36.63,
"learning_rate": 2.273190789473684e-05,
"loss": 1.089,
"step": 12600
},
{
"epoch": 36.92,
"learning_rate": 2.2238486842105264e-05,
"loss": 1.072,
"step": 12700
},
{
"epoch": 37.21,
"learning_rate": 2.174506578947368e-05,
"loss": 1.0993,
"step": 12800
},
{
"epoch": 37.5,
"learning_rate": 2.12516447368421e-05,
"loss": 1.0868,
"step": 12900
},
{
"epoch": 37.79,
"learning_rate": 2.0758223684210525e-05,
"loss": 1.0676,
"step": 13000
},
{
"epoch": 37.79,
"eval_loss": 0.3207636773586273,
"eval_runtime": 141.3445,
"eval_samples_per_second": 32.686,
"eval_steps_per_second": 4.089,
"eval_wer": 0.28675885700299825,
"step": 13000
},
{
"epoch": 38.08,
"learning_rate": 2.0264802631578945e-05,
"loss": 1.098,
"step": 13100
},
{
"epoch": 38.37,
"learning_rate": 1.9771381578947365e-05,
"loss": 1.074,
"step": 13200
},
{
"epoch": 38.66,
"learning_rate": 1.927796052631579e-05,
"loss": 1.0575,
"step": 13300
},
{
"epoch": 38.95,
"learning_rate": 1.878453947368421e-05,
"loss": 1.0775,
"step": 13400
},
{
"epoch": 39.24,
"learning_rate": 1.829111842105263e-05,
"loss": 1.0571,
"step": 13500
},
{
"epoch": 39.24,
"eval_loss": 0.3322257101535797,
"eval_runtime": 141.361,
"eval_samples_per_second": 32.682,
"eval_steps_per_second": 4.089,
"eval_wer": 0.2885333170164596,
"step": 13500
},
{
"epoch": 39.53,
"learning_rate": 1.7797697368421052e-05,
"loss": 1.0639,
"step": 13600
},
{
"epoch": 39.82,
"learning_rate": 1.7304276315789472e-05,
"loss": 1.0656,
"step": 13700
},
{
"epoch": 40.12,
"learning_rate": 1.6810855263157892e-05,
"loss": 1.0896,
"step": 13800
},
{
"epoch": 40.41,
"learning_rate": 1.6317434210526316e-05,
"loss": 1.0577,
"step": 13900
},
{
"epoch": 40.7,
"learning_rate": 1.582894736842105e-05,
"loss": 1.0536,
"step": 14000
},
{
"epoch": 40.7,
"eval_loss": 0.3245023787021637,
"eval_runtime": 141.5405,
"eval_samples_per_second": 32.641,
"eval_steps_per_second": 4.084,
"eval_wer": 0.283118154561586,
"step": 14000
},
{
"epoch": 40.99,
"learning_rate": 1.5335526315789472e-05,
"loss": 1.0625,
"step": 14100
},
{
"epoch": 41.28,
"learning_rate": 1.4842105263157894e-05,
"loss": 1.0675,
"step": 14200
},
{
"epoch": 41.57,
"learning_rate": 1.4348684210526315e-05,
"loss": 1.0562,
"step": 14300
},
{
"epoch": 41.86,
"learning_rate": 1.3855263157894737e-05,
"loss": 1.0601,
"step": 14400
},
{
"epoch": 42.15,
"learning_rate": 1.3361842105263156e-05,
"loss": 1.0525,
"step": 14500
},
{
"epoch": 42.15,
"eval_loss": 0.32854029536247253,
"eval_runtime": 141.2452,
"eval_samples_per_second": 32.709,
"eval_steps_per_second": 4.092,
"eval_wer": 0.2825674600746497,
"step": 14500
},
{
"epoch": 42.44,
"learning_rate": 1.2868421052631577e-05,
"loss": 1.0534,
"step": 14600
},
{
"epoch": 42.73,
"learning_rate": 1.2375e-05,
"loss": 1.0514,
"step": 14700
},
{
"epoch": 43.02,
"learning_rate": 1.188157894736842e-05,
"loss": 1.065,
"step": 14800
},
{
"epoch": 43.31,
"learning_rate": 1.1388157894736841e-05,
"loss": 1.0382,
"step": 14900
},
{
"epoch": 43.6,
"learning_rate": 1.0894736842105263e-05,
"loss": 1.0464,
"step": 15000
},
{
"epoch": 43.6,
"eval_loss": 0.3222814202308655,
"eval_runtime": 140.9953,
"eval_samples_per_second": 32.767,
"eval_steps_per_second": 4.099,
"eval_wer": 0.27963042281098943,
"step": 15000
},
{
"epoch": 43.89,
"learning_rate": 1.0401315789473685e-05,
"loss": 1.0578,
"step": 15100
},
{
"epoch": 44.19,
"learning_rate": 9.907894736842103e-06,
"loss": 1.034,
"step": 15200
},
{
"epoch": 44.48,
"learning_rate": 9.414473684210525e-06,
"loss": 1.0408,
"step": 15300
},
{
"epoch": 44.77,
"learning_rate": 8.921052631578947e-06,
"loss": 1.0396,
"step": 15400
},
{
"epoch": 45.06,
"learning_rate": 8.427631578947367e-06,
"loss": 1.0415,
"step": 15500
},
{
"epoch": 45.06,
"eval_loss": 0.3165771961212158,
"eval_runtime": 141.4647,
"eval_samples_per_second": 32.658,
"eval_steps_per_second": 4.086,
"eval_wer": 0.2773970507250811,
"step": 15500
},
{
"epoch": 45.35,
"learning_rate": 7.934210526315789e-06,
"loss": 1.0306,
"step": 15600
},
{
"epoch": 45.64,
"learning_rate": 7.44078947368421e-06,
"loss": 1.0402,
"step": 15700
},
{
"epoch": 45.93,
"learning_rate": 6.9473684210526315e-06,
"loss": 1.0148,
"step": 15800
},
{
"epoch": 46.22,
"learning_rate": 6.458881578947368e-06,
"loss": 1.0276,
"step": 15900
},
{
"epoch": 46.51,
"learning_rate": 5.965460526315788e-06,
"loss": 1.0356,
"step": 16000
},
{
"epoch": 46.51,
"eval_loss": 0.31774380803108215,
"eval_runtime": 141.9517,
"eval_samples_per_second": 32.546,
"eval_steps_per_second": 4.072,
"eval_wer": 0.27455179587591017,
"step": 16000
},
{
"epoch": 46.8,
"learning_rate": 5.47203947368421e-06,
"loss": 1.0331,
"step": 16100
},
{
"epoch": 47.09,
"learning_rate": 4.978618421052631e-06,
"loss": 1.0442,
"step": 16200
},
{
"epoch": 47.38,
"learning_rate": 4.485197368421052e-06,
"loss": 1.0157,
"step": 16300
},
{
"epoch": 47.67,
"learning_rate": 3.991776315789474e-06,
"loss": 1.0308,
"step": 16400
},
{
"epoch": 47.96,
"learning_rate": 3.498355263157895e-06,
"loss": 1.04,
"step": 16500
},
{
"epoch": 47.96,
"eval_loss": 0.31499630212783813,
"eval_runtime": 141.8201,
"eval_samples_per_second": 32.576,
"eval_steps_per_second": 4.076,
"eval_wer": 0.27354218931652696,
"step": 16500
},
{
"epoch": 48.26,
"learning_rate": 3.0049342105263154e-06,
"loss": 1.0388,
"step": 16600
},
{
"epoch": 48.55,
"learning_rate": 2.5115131578947363e-06,
"loss": 1.0016,
"step": 16700
},
{
"epoch": 48.84,
"learning_rate": 2.0180921052631577e-06,
"loss": 1.0127,
"step": 16800
},
{
"epoch": 49.13,
"learning_rate": 1.524671052631579e-06,
"loss": 1.0343,
"step": 16900
},
{
"epoch": 49.42,
"learning_rate": 1.0312499999999999e-06,
"loss": 1.0209,
"step": 17000
},
{
"epoch": 49.42,
"eval_loss": 0.3174569308757782,
"eval_runtime": 141.608,
"eval_samples_per_second": 32.625,
"eval_steps_per_second": 4.082,
"eval_wer": 0.2730526831059169,
"step": 17000
},
{
"epoch": 49.71,
"learning_rate": 5.378289473684209e-07,
"loss": 1.0189,
"step": 17100
},
{
"epoch": 50.0,
"learning_rate": 4.44078947368421e-08,
"loss": 1.0364,
"step": 17200
},
{
"epoch": 50.0,
"step": 17200,
"total_flos": 5.147257612039286e+19,
"train_loss": 1.449138161858847,
"train_runtime": 28230.9535,
"train_samples_per_second": 19.535,
"train_steps_per_second": 0.609
}
],
"max_steps": 17200,
"num_train_epochs": 50,
"total_flos": 5.147257612039286e+19,
"trial_name": null,
"trial_params": null
}