2042 lines
49 KiB
JSON
2042 lines
49 KiB
JSON
|
{
|
||
|
|
"best_metric": 1.0,
|
||
|
|
"best_model_checkpoint": "../data/wav2vec2-large-xlsr-53-thai/checkpoint-1000",
|
||
|
|
"epoch": 99.77827050997783,
|
||
|
|
"global_step": 135000,
|
||
|
|
"is_hyper_param_search": false,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"epoch": 0.74,
|
||
|
|
"learning_rate": 9.970000000000001e-05,
|
||
|
|
"loss": 7.1725,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.74,
|
||
|
|
"eval_loss": 3.039940595626831,
|
||
|
|
"eval_runtime": 185.7874,
|
||
|
|
"eval_samples_per_second": 16.293,
|
||
|
|
"eval_steps_per_second": 0.511,
|
||
|
|
"eval_wer": 1.0,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.48,
|
||
|
|
"learning_rate": 9.925763216679077e-05,
|
||
|
|
"loss": 1.0788,
|
||
|
|
"step": 2000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.48,
|
||
|
|
"eval_loss": 0.6130112409591675,
|
||
|
|
"eval_runtime": 176.9194,
|
||
|
|
"eval_samples_per_second": 17.109,
|
||
|
|
"eval_steps_per_second": 0.537,
|
||
|
|
"eval_wer": 0.6820256330103157,
|
||
|
|
"step": 2000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.22,
|
||
|
|
"learning_rate": 9.851303052866717e-05,
|
||
|
|
"loss": 0.6339,
|
||
|
|
"step": 3000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.22,
|
||
|
|
"eval_loss": 0.49624550342559814,
|
||
|
|
"eval_runtime": 175.6715,
|
||
|
|
"eval_samples_per_second": 17.231,
|
||
|
|
"eval_steps_per_second": 0.541,
|
||
|
|
"eval_wer": 0.6168802750859643,
|
||
|
|
"step": 3000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.96,
|
||
|
|
"learning_rate": 9.776842889054357e-05,
|
||
|
|
"loss": 0.5295,
|
||
|
|
"step": 4000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.96,
|
||
|
|
"eval_loss": 0.4389719069004059,
|
||
|
|
"eval_runtime": 176.6072,
|
||
|
|
"eval_samples_per_second": 17.14,
|
||
|
|
"eval_steps_per_second": 0.538,
|
||
|
|
"eval_wer": 0.577117849327915,
|
||
|
|
"step": 4000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.7,
|
||
|
|
"learning_rate": 9.702382725241996e-05,
|
||
|
|
"loss": 0.4616,
|
||
|
|
"step": 5000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.7,
|
||
|
|
"eval_loss": 0.4208119809627533,
|
||
|
|
"eval_runtime": 196.7481,
|
||
|
|
"eval_samples_per_second": 15.385,
|
||
|
|
"eval_steps_per_second": 0.483,
|
||
|
|
"eval_wer": 0.5618005626758362,
|
||
|
|
"step": 5000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.43,
|
||
|
|
"learning_rate": 9.627922561429635e-05,
|
||
|
|
"loss": 0.4206,
|
||
|
|
"step": 6000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.43,
|
||
|
|
"eval_loss": 0.4009915888309479,
|
||
|
|
"eval_runtime": 180.4803,
|
||
|
|
"eval_samples_per_second": 16.772,
|
||
|
|
"eval_steps_per_second": 0.526,
|
||
|
|
"eval_wer": 0.5411691153485464,
|
||
|
|
"step": 6000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.17,
|
||
|
|
"learning_rate": 9.553462397617276e-05,
|
||
|
|
"loss": 0.3908,
|
||
|
|
"step": 7000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.17,
|
||
|
|
"eval_loss": 0.399889200925827,
|
||
|
|
"eval_runtime": 189.7255,
|
||
|
|
"eval_samples_per_second": 15.955,
|
||
|
|
"eval_steps_per_second": 0.501,
|
||
|
|
"eval_wer": 0.5384807752422632,
|
||
|
|
"step": 7000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.91,
|
||
|
|
"learning_rate": 9.479002233804914e-05,
|
||
|
|
"loss": 0.3659,
|
||
|
|
"step": 8000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.91,
|
||
|
|
"eval_loss": 0.3719017505645752,
|
||
|
|
"eval_runtime": 184.5994,
|
||
|
|
"eval_samples_per_second": 16.398,
|
||
|
|
"eval_steps_per_second": 0.515,
|
||
|
|
"eval_wer": 0.5141606752110034,
|
||
|
|
"step": 8000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 6.65,
|
||
|
|
"learning_rate": 9.404542069992555e-05,
|
||
|
|
"loss": 0.338,
|
||
|
|
"step": 9000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 6.65,
|
||
|
|
"eval_loss": 0.35533592104911804,
|
||
|
|
"eval_runtime": 188.5325,
|
||
|
|
"eval_samples_per_second": 16.056,
|
||
|
|
"eval_steps_per_second": 0.504,
|
||
|
|
"eval_wer": 0.5101594248202563,
|
||
|
|
"step": 9000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 7.39,
|
||
|
|
"learning_rate": 9.330156366344007e-05,
|
||
|
|
"loss": 0.3222,
|
||
|
|
"step": 10000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 7.39,
|
||
|
|
"eval_loss": 0.40722423791885376,
|
||
|
|
"eval_runtime": 195.2723,
|
||
|
|
"eval_samples_per_second": 15.501,
|
||
|
|
"eval_steps_per_second": 0.487,
|
||
|
|
"eval_wer": 0.49309159112222567,
|
||
|
|
"step": 10000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 8.13,
|
||
|
|
"learning_rate": 9.255696202531646e-05,
|
||
|
|
"loss": 0.31,
|
||
|
|
"step": 11000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 8.13,
|
||
|
|
"eval_loss": 0.37148234248161316,
|
||
|
|
"eval_runtime": 182.6314,
|
||
|
|
"eval_samples_per_second": 16.574,
|
||
|
|
"eval_steps_per_second": 0.52,
|
||
|
|
"eval_wer": 0.48715223507346045,
|
||
|
|
"step": 11000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 8.87,
|
||
|
|
"learning_rate": 9.181310498883098e-05,
|
||
|
|
"loss": 0.2901,
|
||
|
|
"step": 12000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 8.87,
|
||
|
|
"eval_loss": 0.365583598613739,
|
||
|
|
"eval_runtime": 175.3968,
|
||
|
|
"eval_samples_per_second": 17.258,
|
||
|
|
"eval_steps_per_second": 0.542,
|
||
|
|
"eval_wer": 0.4882775867458581,
|
||
|
|
"step": 12000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 9.61,
|
||
|
|
"learning_rate": 9.106850335070738e-05,
|
||
|
|
"loss": 0.276,
|
||
|
|
"step": 13000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 9.61,
|
||
|
|
"eval_loss": 0.37835174798965454,
|
||
|
|
"eval_runtime": 175.9962,
|
||
|
|
"eval_samples_per_second": 17.199,
|
||
|
|
"eval_steps_per_second": 0.54,
|
||
|
|
"eval_wer": 0.4897155361050328,
|
||
|
|
"step": 13000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 10.35,
|
||
|
|
"learning_rate": 9.032539091586003e-05,
|
||
|
|
"loss": 0.2628,
|
||
|
|
"step": 14000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 10.35,
|
||
|
|
"eval_loss": 0.4127887189388275,
|
||
|
|
"eval_runtime": 180.0512,
|
||
|
|
"eval_samples_per_second": 16.812,
|
||
|
|
"eval_steps_per_second": 0.528,
|
||
|
|
"eval_wer": 0.48871522350734603,
|
||
|
|
"step": 14000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 11.09,
|
||
|
|
"learning_rate": 8.958078927773641e-05,
|
||
|
|
"loss": 0.256,
|
||
|
|
"step": 15000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 11.09,
|
||
|
|
"eval_loss": 0.3795180022716522,
|
||
|
|
"eval_runtime": 174.1363,
|
||
|
|
"eval_samples_per_second": 17.383,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.46939668646452015,
|
||
|
|
"step": 15000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 11.83,
|
||
|
|
"learning_rate": 8.88361876396128e-05,
|
||
|
|
"loss": 0.2455,
|
||
|
|
"step": 16000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 11.83,
|
||
|
|
"eval_loss": 0.3748489320278168,
|
||
|
|
"eval_runtime": 174.8629,
|
||
|
|
"eval_samples_per_second": 17.311,
|
||
|
|
"eval_steps_per_second": 0.543,
|
||
|
|
"eval_wer": 0.46970928415129726,
|
||
|
|
"step": 16000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 12.56,
|
||
|
|
"learning_rate": 8.809233060312732e-05,
|
||
|
|
"loss": 0.2377,
|
||
|
|
"step": 17000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 12.56,
|
||
|
|
"eval_loss": 0.4371410012245178,
|
||
|
|
"eval_runtime": 173.1913,
|
||
|
|
"eval_samples_per_second": 17.478,
|
||
|
|
"eval_steps_per_second": 0.549,
|
||
|
|
"eval_wer": 0.4650203188496405,
|
||
|
|
"step": 17000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 13.3,
|
||
|
|
"learning_rate": 8.734772896500373e-05,
|
||
|
|
"loss": 0.2279,
|
||
|
|
"step": 18000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 13.3,
|
||
|
|
"eval_loss": 0.38019847869873047,
|
||
|
|
"eval_runtime": 177.0444,
|
||
|
|
"eval_samples_per_second": 17.097,
|
||
|
|
"eval_steps_per_second": 0.537,
|
||
|
|
"eval_wer": 0.45339168490153176,
|
||
|
|
"step": 18000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 14.04,
|
||
|
|
"learning_rate": 8.660387192851825e-05,
|
||
|
|
"loss": 0.2247,
|
||
|
|
"step": 19000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 14.04,
|
||
|
|
"eval_loss": 0.3687661588191986,
|
||
|
|
"eval_runtime": 197.8618,
|
||
|
|
"eval_samples_per_second": 15.299,
|
||
|
|
"eval_steps_per_second": 0.48,
|
||
|
|
"eval_wer": 0.4567677399187246,
|
||
|
|
"step": 19000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 14.78,
|
||
|
|
"learning_rate": 8.585927029039465e-05,
|
||
|
|
"loss": 0.2139,
|
||
|
|
"step": 20000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 14.78,
|
||
|
|
"eval_loss": 0.3616667091846466,
|
||
|
|
"eval_runtime": 177.8968,
|
||
|
|
"eval_samples_per_second": 17.015,
|
||
|
|
"eval_steps_per_second": 0.534,
|
||
|
|
"eval_wer": 0.446639574867146,
|
||
|
|
"step": 20000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 15.52,
|
||
|
|
"learning_rate": 8.511541325390916e-05,
|
||
|
|
"loss": 0.2072,
|
||
|
|
"step": 21000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 15.52,
|
||
|
|
"eval_loss": 0.4171736538410187,
|
||
|
|
"eval_runtime": 171.8647,
|
||
|
|
"eval_samples_per_second": 17.613,
|
||
|
|
"eval_steps_per_second": 0.553,
|
||
|
|
"eval_wer": 0.45883088465145355,
|
||
|
|
"step": 21000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 16.26,
|
||
|
|
"learning_rate": 8.437081161578556e-05,
|
||
|
|
"loss": 0.1999,
|
||
|
|
"step": 22000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 16.26,
|
||
|
|
"eval_loss": 0.7755644917488098,
|
||
|
|
"eval_runtime": 171.1132,
|
||
|
|
"eval_samples_per_second": 17.69,
|
||
|
|
"eval_steps_per_second": 0.555,
|
||
|
|
"eval_wer": 0.45126602063144733,
|
||
|
|
"step": 22000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 17.0,
|
||
|
|
"learning_rate": 8.362695457930007e-05,
|
||
|
|
"loss": 0.1958,
|
||
|
|
"step": 23000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 17.0,
|
||
|
|
"eval_loss": 0.41745948791503906,
|
||
|
|
"eval_runtime": 172.9131,
|
||
|
|
"eval_samples_per_second": 17.506,
|
||
|
|
"eval_steps_per_second": 0.549,
|
||
|
|
"eval_wer": 0.44382619568615195,
|
||
|
|
"step": 23000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 17.74,
|
||
|
|
"learning_rate": 8.288235294117647e-05,
|
||
|
|
"loss": 0.1866,
|
||
|
|
"step": 24000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 17.74,
|
||
|
|
"eval_loss": 0.4865190386772156,
|
||
|
|
"eval_runtime": 173.9742,
|
||
|
|
"eval_samples_per_second": 17.399,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.45401688027508597,
|
||
|
|
"step": 24000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 18.48,
|
||
|
|
"learning_rate": 8.213924050632912e-05,
|
||
|
|
"loss": 0.1854,
|
||
|
|
"step": 25000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 18.48,
|
||
|
|
"eval_loss": 0.5420244336128235,
|
||
|
|
"eval_runtime": 175.6968,
|
||
|
|
"eval_samples_per_second": 17.229,
|
||
|
|
"eval_steps_per_second": 0.541,
|
||
|
|
"eval_wer": 0.4475148483901219,
|
||
|
|
"step": 25000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 19.22,
|
||
|
|
"learning_rate": 8.139463886820552e-05,
|
||
|
|
"loss": 0.1777,
|
||
|
|
"step": 26000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 19.22,
|
||
|
|
"eval_loss": 0.4428083300590515,
|
||
|
|
"eval_runtime": 174.8946,
|
||
|
|
"eval_samples_per_second": 17.308,
|
||
|
|
"eval_steps_per_second": 0.543,
|
||
|
|
"eval_wer": 0.4501406689590497,
|
||
|
|
"step": 26000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 19.96,
|
||
|
|
"learning_rate": 8.065003723008191e-05,
|
||
|
|
"loss": 0.1764,
|
||
|
|
"step": 27000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 19.96,
|
||
|
|
"eval_loss": 0.44021618366241455,
|
||
|
|
"eval_runtime": 174.1095,
|
||
|
|
"eval_samples_per_second": 17.386,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.44307596123788684,
|
||
|
|
"step": 27000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 20.69,
|
||
|
|
"learning_rate": 7.990543559195831e-05,
|
||
|
|
"loss": 0.1698,
|
||
|
|
"step": 28000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 20.69,
|
||
|
|
"eval_loss": 0.48001521825790405,
|
||
|
|
"eval_runtime": 173.9598,
|
||
|
|
"eval_samples_per_second": 17.401,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.4456392622694592,
|
||
|
|
"step": 28000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 21.43,
|
||
|
|
"learning_rate": 7.91608339538347e-05,
|
||
|
|
"loss": 0.1625,
|
||
|
|
"step": 29000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 21.43,
|
||
|
|
"eval_loss": 0.4736146628856659,
|
||
|
|
"eval_runtime": 173.7886,
|
||
|
|
"eval_samples_per_second": 17.418,
|
||
|
|
"eval_steps_per_second": 0.547,
|
||
|
|
"eval_wer": 0.44676461394185685,
|
||
|
|
"step": 29000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 22.17,
|
||
|
|
"learning_rate": 7.841697691734922e-05,
|
||
|
|
"loss": 0.1611,
|
||
|
|
"step": 30000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 22.17,
|
||
|
|
"eval_loss": 0.4669973850250244,
|
||
|
|
"eval_runtime": 174.8827,
|
||
|
|
"eval_samples_per_second": 17.309,
|
||
|
|
"eval_steps_per_second": 0.543,
|
||
|
|
"eval_wer": 0.43813691778680836,
|
||
|
|
"step": 30000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 22.91,
|
||
|
|
"learning_rate": 7.767237527922562e-05,
|
||
|
|
"loss": 0.1557,
|
||
|
|
"step": 31000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 22.91,
|
||
|
|
"eval_loss": 0.49789130687713623,
|
||
|
|
"eval_runtime": 176.2993,
|
||
|
|
"eval_samples_per_second": 17.17,
|
||
|
|
"eval_steps_per_second": 0.539,
|
||
|
|
"eval_wer": 0.4390121913097843,
|
||
|
|
"step": 31000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 23.65,
|
||
|
|
"learning_rate": 7.692851824274014e-05,
|
||
|
|
"loss": 0.1531,
|
||
|
|
"step": 32000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 23.65,
|
||
|
|
"eval_loss": 0.49486255645751953,
|
||
|
|
"eval_runtime": 178.461,
|
||
|
|
"eval_samples_per_second": 16.962,
|
||
|
|
"eval_steps_per_second": 0.532,
|
||
|
|
"eval_wer": 0.4311347296030009,
|
||
|
|
"step": 32000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 24.39,
|
||
|
|
"learning_rate": 7.618391660461653e-05,
|
||
|
|
"loss": 0.1485,
|
||
|
|
"step": 33000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 24.39,
|
||
|
|
"eval_loss": 0.5052071809768677,
|
||
|
|
"eval_runtime": 197.1526,
|
||
|
|
"eval_samples_per_second": 15.354,
|
||
|
|
"eval_steps_per_second": 0.482,
|
||
|
|
"eval_wer": 0.43432322600812756,
|
||
|
|
"step": 33000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 25.13,
|
||
|
|
"learning_rate": 7.543931496649293e-05,
|
||
|
|
"loss": 0.149,
|
||
|
|
"step": 34000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 25.13,
|
||
|
|
"eval_loss": 0.4628446400165558,
|
||
|
|
"eval_runtime": 191.152,
|
||
|
|
"eval_samples_per_second": 15.836,
|
||
|
|
"eval_steps_per_second": 0.497,
|
||
|
|
"eval_wer": 0.42638324476398876,
|
||
|
|
"step": 34000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 25.87,
|
||
|
|
"learning_rate": 7.469545793000745e-05,
|
||
|
|
"loss": 0.1439,
|
||
|
|
"step": 35000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 25.87,
|
||
|
|
"eval_loss": 0.4972988963127136,
|
||
|
|
"eval_runtime": 199.2887,
|
||
|
|
"eval_samples_per_second": 15.189,
|
||
|
|
"eval_steps_per_second": 0.477,
|
||
|
|
"eval_wer": 0.4331353547983745,
|
||
|
|
"step": 35000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 26.61,
|
||
|
|
"learning_rate": 7.395160089352196e-05,
|
||
|
|
"loss": 0.14,
|
||
|
|
"step": 36000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 26.61,
|
||
|
|
"eval_loss": 0.5113205313682556,
|
||
|
|
"eval_runtime": 177.8533,
|
||
|
|
"eval_samples_per_second": 17.02,
|
||
|
|
"eval_steps_per_second": 0.534,
|
||
|
|
"eval_wer": 0.4366364488902782,
|
||
|
|
"step": 36000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 27.35,
|
||
|
|
"learning_rate": 7.320699925539837e-05,
|
||
|
|
"loss": 0.1373,
|
||
|
|
"step": 37000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 27.35,
|
||
|
|
"eval_loss": 0.5227766036987305,
|
||
|
|
"eval_runtime": 197.1721,
|
||
|
|
"eval_samples_per_second": 15.352,
|
||
|
|
"eval_steps_per_second": 0.482,
|
||
|
|
"eval_wer": 0.4340731478587058,
|
||
|
|
"step": 37000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 28.09,
|
||
|
|
"learning_rate": 7.246239761727476e-05,
|
||
|
|
"loss": 0.134,
|
||
|
|
"step": 38000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 28.09,
|
||
|
|
"eval_loss": 0.5289518237113953,
|
||
|
|
"eval_runtime": 186.9942,
|
||
|
|
"eval_samples_per_second": 16.188,
|
||
|
|
"eval_steps_per_second": 0.508,
|
||
|
|
"eval_wer": 0.42744607689903097,
|
||
|
|
"step": 38000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 28.82,
|
||
|
|
"learning_rate": 7.171779597915115e-05,
|
||
|
|
"loss": 0.1309,
|
||
|
|
"step": 39000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 28.82,
|
||
|
|
"eval_loss": 0.5079456567764282,
|
||
|
|
"eval_runtime": 177.9839,
|
||
|
|
"eval_samples_per_second": 17.007,
|
||
|
|
"eval_steps_per_second": 0.534,
|
||
|
|
"eval_wer": 0.4363238512035011,
|
||
|
|
"step": 39000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 29.56,
|
||
|
|
"learning_rate": 7.097319434102755e-05,
|
||
|
|
"loss": 0.129,
|
||
|
|
"step": 40000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 29.56,
|
||
|
|
"eval_loss": 0.5366144776344299,
|
||
|
|
"eval_runtime": 174.4769,
|
||
|
|
"eval_samples_per_second": 17.349,
|
||
|
|
"eval_steps_per_second": 0.544,
|
||
|
|
"eval_wer": 0.4248202563301032,
|
||
|
|
"step": 40000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 30.3,
|
||
|
|
"learning_rate": 7.02300819061802e-05,
|
||
|
|
"loss": 0.126,
|
||
|
|
"step": 41000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 30.3,
|
||
|
|
"eval_loss": 0.5051801204681396,
|
||
|
|
"eval_runtime": 174.1717,
|
||
|
|
"eval_samples_per_second": 17.379,
|
||
|
|
"eval_steps_per_second": 0.545,
|
||
|
|
"eval_wer": 0.42275711159737417,
|
||
|
|
"step": 41000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 31.04,
|
||
|
|
"learning_rate": 6.94854802680566e-05,
|
||
|
|
"loss": 0.1255,
|
||
|
|
"step": 42000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 31.04,
|
||
|
|
"eval_loss": 0.5653140544891357,
|
||
|
|
"eval_runtime": 175.3338,
|
||
|
|
"eval_samples_per_second": 17.264,
|
||
|
|
"eval_steps_per_second": 0.542,
|
||
|
|
"eval_wer": 0.43063457330415755,
|
||
|
|
"step": 42000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 31.78,
|
||
|
|
"learning_rate": 6.874087862993299e-05,
|
||
|
|
"loss": 0.122,
|
||
|
|
"step": 43000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 31.78,
|
||
|
|
"eval_loss": 0.5453576445579529,
|
||
|
|
"eval_runtime": 176.4408,
|
||
|
|
"eval_samples_per_second": 17.156,
|
||
|
|
"eval_steps_per_second": 0.538,
|
||
|
|
"eval_wer": 0.42800875273522976,
|
||
|
|
"step": 43000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 32.52,
|
||
|
|
"learning_rate": 6.799627699180938e-05,
|
||
|
|
"loss": 0.1191,
|
||
|
|
"step": 44000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 32.52,
|
||
|
|
"eval_loss": 0.6267871260643005,
|
||
|
|
"eval_runtime": 172.551,
|
||
|
|
"eval_samples_per_second": 17.543,
|
||
|
|
"eval_steps_per_second": 0.551,
|
||
|
|
"eval_wer": 0.42400750234448265,
|
||
|
|
"step": 44000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 33.26,
|
||
|
|
"learning_rate": 6.725167535368579e-05,
|
||
|
|
"loss": 0.1151,
|
||
|
|
"step": 45000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 33.26,
|
||
|
|
"eval_loss": 0.5381397008895874,
|
||
|
|
"eval_runtime": 174.0505,
|
||
|
|
"eval_samples_per_second": 17.392,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.4221944357611754,
|
||
|
|
"step": 45000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 34.0,
|
||
|
|
"learning_rate": 6.650707371556217e-05,
|
||
|
|
"loss": 0.115,
|
||
|
|
"step": 46000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 34.0,
|
||
|
|
"eval_loss": 0.5648242235183716,
|
||
|
|
"eval_runtime": 174.2088,
|
||
|
|
"eval_samples_per_second": 17.376,
|
||
|
|
"eval_steps_per_second": 0.545,
|
||
|
|
"eval_wer": 0.4198812128790247,
|
||
|
|
"step": 46000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 34.74,
|
||
|
|
"learning_rate": 6.576247207743858e-05,
|
||
|
|
"loss": 0.1124,
|
||
|
|
"step": 47000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 34.74,
|
||
|
|
"eval_loss": 0.613673985004425,
|
||
|
|
"eval_runtime": 193.9522,
|
||
|
|
"eval_samples_per_second": 15.607,
|
||
|
|
"eval_steps_per_second": 0.49,
|
||
|
|
"eval_wer": 0.4295092216317599,
|
||
|
|
"step": 47000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 35.48,
|
||
|
|
"learning_rate": 6.50186150409531e-05,
|
||
|
|
"loss": 0.1101,
|
||
|
|
"step": 48000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 35.48,
|
||
|
|
"eval_loss": 0.5470528602600098,
|
||
|
|
"eval_runtime": 193.2438,
|
||
|
|
"eval_samples_per_second": 15.664,
|
||
|
|
"eval_steps_per_second": 0.492,
|
||
|
|
"eval_wer": 0.42206939668646454,
|
||
|
|
"step": 48000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 36.22,
|
||
|
|
"learning_rate": 6.427475800446761e-05,
|
||
|
|
"loss": 0.1061,
|
||
|
|
"step": 49000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 36.22,
|
||
|
|
"eval_loss": 0.5942440032958984,
|
||
|
|
"eval_runtime": 200.9573,
|
||
|
|
"eval_samples_per_second": 15.063,
|
||
|
|
"eval_steps_per_second": 0.473,
|
||
|
|
"eval_wer": 0.4258205689277899,
|
||
|
|
"step": 49000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 36.95,
|
||
|
|
"learning_rate": 6.353015636634401e-05,
|
||
|
|
"loss": 0.1065,
|
||
|
|
"step": 50000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 36.95,
|
||
|
|
"eval_loss": 0.5583751201629639,
|
||
|
|
"eval_runtime": 178.0909,
|
||
|
|
"eval_samples_per_second": 16.997,
|
||
|
|
"eval_steps_per_second": 0.533,
|
||
|
|
"eval_wer": 0.4255704907783682,
|
||
|
|
"step": 50000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 37.69,
|
||
|
|
"learning_rate": 6.27855547282204e-05,
|
||
|
|
"loss": 0.1043,
|
||
|
|
"step": 51000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 37.69,
|
||
|
|
"eval_loss": 0.7426488399505615,
|
||
|
|
"eval_runtime": 188.9691,
|
||
|
|
"eval_samples_per_second": 16.018,
|
||
|
|
"eval_steps_per_second": 0.503,
|
||
|
|
"eval_wer": 0.4210065645514223,
|
||
|
|
"step": 51000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 38.43,
|
||
|
|
"learning_rate": 6.204169769173492e-05,
|
||
|
|
"loss": 0.103,
|
||
|
|
"step": 52000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 38.43,
|
||
|
|
"eval_loss": 0.7028630375862122,
|
||
|
|
"eval_runtime": 182.3453,
|
||
|
|
"eval_samples_per_second": 16.6,
|
||
|
|
"eval_steps_per_second": 0.521,
|
||
|
|
"eval_wer": 0.42338230697092843,
|
||
|
|
"step": 52000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 39.17,
|
||
|
|
"learning_rate": 6.129709605361132e-05,
|
||
|
|
"loss": 0.1016,
|
||
|
|
"step": 53000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 39.17,
|
||
|
|
"eval_loss": 0.644480288028717,
|
||
|
|
"eval_runtime": 172.8589,
|
||
|
|
"eval_samples_per_second": 17.511,
|
||
|
|
"eval_steps_per_second": 0.55,
|
||
|
|
"eval_wer": 0.42319474835886217,
|
||
|
|
"step": 53000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 39.91,
|
||
|
|
"learning_rate": 6.055323901712584e-05,
|
||
|
|
"loss": 0.0977,
|
||
|
|
"step": 54000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 39.91,
|
||
|
|
"eval_loss": 0.6561093926429749,
|
||
|
|
"eval_runtime": 174.7926,
|
||
|
|
"eval_samples_per_second": 17.318,
|
||
|
|
"eval_steps_per_second": 0.544,
|
||
|
|
"eval_wer": 0.41394185683025947,
|
||
|
|
"step": 54000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 40.65,
|
||
|
|
"learning_rate": 5.980863737900224e-05,
|
||
|
|
"loss": 0.097,
|
||
|
|
"step": 55000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 40.65,
|
||
|
|
"eval_loss": 0.6229739785194397,
|
||
|
|
"eval_runtime": 176.8753,
|
||
|
|
"eval_samples_per_second": 17.114,
|
||
|
|
"eval_steps_per_second": 0.537,
|
||
|
|
"eval_wer": 0.41450453266645826,
|
||
|
|
"step": 55000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 41.39,
|
||
|
|
"learning_rate": 5.9064780342516755e-05,
|
||
|
|
"loss": 0.0963,
|
||
|
|
"step": 56000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 41.39,
|
||
|
|
"eval_loss": 0.5873845219612122,
|
||
|
|
"eval_runtime": 175.9014,
|
||
|
|
"eval_samples_per_second": 17.209,
|
||
|
|
"eval_steps_per_second": 0.54,
|
||
|
|
"eval_wer": 0.41713035323538605,
|
||
|
|
"step": 56000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 42.13,
|
||
|
|
"learning_rate": 5.832017870439316e-05,
|
||
|
|
"loss": 0.0943,
|
||
|
|
"step": 57000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 42.13,
|
||
|
|
"eval_loss": 0.5930323600769043,
|
||
|
|
"eval_runtime": 174.9181,
|
||
|
|
"eval_samples_per_second": 17.305,
|
||
|
|
"eval_steps_per_second": 0.543,
|
||
|
|
"eval_wer": 0.418005626758362,
|
||
|
|
"step": 57000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 42.87,
|
||
|
|
"learning_rate": 5.757557706626955e-05,
|
||
|
|
"loss": 0.093,
|
||
|
|
"step": 58000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 42.87,
|
||
|
|
"eval_loss": 0.6292551755905151,
|
||
|
|
"eval_runtime": 175.0843,
|
||
|
|
"eval_samples_per_second": 17.289,
|
||
|
|
"eval_steps_per_second": 0.543,
|
||
|
|
"eval_wer": 0.4175054704595186,
|
||
|
|
"step": 58000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 43.61,
|
||
|
|
"learning_rate": 5.683097542814595e-05,
|
||
|
|
"loss": 0.0896,
|
||
|
|
"step": 59000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 43.61,
|
||
|
|
"eval_loss": 0.6171667575836182,
|
||
|
|
"eval_runtime": 175.3688,
|
||
|
|
"eval_samples_per_second": 17.261,
|
||
|
|
"eval_steps_per_second": 0.542,
|
||
|
|
"eval_wer": 0.41319162238199436,
|
||
|
|
"step": 59000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 44.35,
|
||
|
|
"learning_rate": 5.608637379002234e-05,
|
||
|
|
"loss": 0.0911,
|
||
|
|
"step": 60000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 44.35,
|
||
|
|
"eval_loss": 0.6143432259559631,
|
||
|
|
"eval_runtime": 178.3103,
|
||
|
|
"eval_samples_per_second": 16.976,
|
||
|
|
"eval_steps_per_second": 0.533,
|
||
|
|
"eval_wer": 0.4172553923100969,
|
||
|
|
"step": 60000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 45.08,
|
||
|
|
"learning_rate": 5.534326135517498e-05,
|
||
|
|
"loss": 0.0885,
|
||
|
|
"step": 61000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 45.08,
|
||
|
|
"eval_loss": 0.6145235300064087,
|
||
|
|
"eval_runtime": 191.9563,
|
||
|
|
"eval_samples_per_second": 15.769,
|
||
|
|
"eval_steps_per_second": 0.495,
|
||
|
|
"eval_wer": 0.4133166614567052,
|
||
|
|
"step": 61000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 45.82,
|
||
|
|
"learning_rate": 5.459865971705138e-05,
|
||
|
|
"loss": 0.0872,
|
||
|
|
"step": 62000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 45.82,
|
||
|
|
"eval_loss": 0.8499487638473511,
|
||
|
|
"eval_runtime": 177.455,
|
||
|
|
"eval_samples_per_second": 17.058,
|
||
|
|
"eval_steps_per_second": 0.535,
|
||
|
|
"eval_wer": 0.41187871209753046,
|
||
|
|
"step": 62000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 46.56,
|
||
|
|
"learning_rate": 5.385405807892777e-05,
|
||
|
|
"loss": 0.0883,
|
||
|
|
"step": 63000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 46.56,
|
||
|
|
"eval_loss": 0.7073235511779785,
|
||
|
|
"eval_runtime": 199.205,
|
||
|
|
"eval_samples_per_second": 15.195,
|
||
|
|
"eval_steps_per_second": 0.477,
|
||
|
|
"eval_wer": 0.4190684588934042,
|
||
|
|
"step": 63000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 47.3,
|
||
|
|
"learning_rate": 5.3109456440804175e-05,
|
||
|
|
"loss": 0.0844,
|
||
|
|
"step": 64000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 47.3,
|
||
|
|
"eval_loss": 0.6707965135574341,
|
||
|
|
"eval_runtime": 203.0283,
|
||
|
|
"eval_samples_per_second": 14.909,
|
||
|
|
"eval_steps_per_second": 0.468,
|
||
|
|
"eval_wer": 0.4145670522038137,
|
||
|
|
"step": 64000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 48.04,
|
||
|
|
"learning_rate": 5.2365599404318685e-05,
|
||
|
|
"loss": 0.0851,
|
||
|
|
"step": 65000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 48.04,
|
||
|
|
"eval_loss": 0.6518837213516235,
|
||
|
|
"eval_runtime": 184.8621,
|
||
|
|
"eval_samples_per_second": 16.374,
|
||
|
|
"eval_steps_per_second": 0.514,
|
||
|
|
"eval_wer": 0.4101281650515786,
|
||
|
|
"step": 65000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 48.78,
|
||
|
|
"learning_rate": 5.162099776619509e-05,
|
||
|
|
"loss": 0.0832,
|
||
|
|
"step": 66000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 48.78,
|
||
|
|
"eval_loss": 0.6594052910804749,
|
||
|
|
"eval_runtime": 196.4902,
|
||
|
|
"eval_samples_per_second": 15.405,
|
||
|
|
"eval_steps_per_second": 0.483,
|
||
|
|
"eval_wer": 0.41387933729290405,
|
||
|
|
"step": 66000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 49.52,
|
||
|
|
"learning_rate": 5.0877140729709605e-05,
|
||
|
|
"loss": 0.0822,
|
||
|
|
"step": 67000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 49.52,
|
||
|
|
"eval_loss": 0.653843104839325,
|
||
|
|
"eval_runtime": 186.7472,
|
||
|
|
"eval_samples_per_second": 16.209,
|
||
|
|
"eval_steps_per_second": 0.509,
|
||
|
|
"eval_wer": 0.415942482025633,
|
||
|
|
"step": 67000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 50.26,
|
||
|
|
"learning_rate": 5.013253909158601e-05,
|
||
|
|
"loss": 0.0833,
|
||
|
|
"step": 68000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 50.26,
|
||
|
|
"eval_loss": 0.6541770100593567,
|
||
|
|
"eval_runtime": 175.2691,
|
||
|
|
"eval_samples_per_second": 17.271,
|
||
|
|
"eval_steps_per_second": 0.542,
|
||
|
|
"eval_wer": 0.41844326351984995,
|
||
|
|
"step": 68000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 51.0,
|
||
|
|
"learning_rate": 4.9387937453462405e-05,
|
||
|
|
"loss": 0.079,
|
||
|
|
"step": 69000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 51.0,
|
||
|
|
"eval_loss": 0.6602481007575989,
|
||
|
|
"eval_runtime": 176.6295,
|
||
|
|
"eval_samples_per_second": 17.138,
|
||
|
|
"eval_steps_per_second": 0.538,
|
||
|
|
"eval_wer": 0.41212879024695215,
|
||
|
|
"step": 69000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 51.74,
|
||
|
|
"learning_rate": 4.8644080416976914e-05,
|
||
|
|
"loss": 0.0768,
|
||
|
|
"step": 70000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 51.74,
|
||
|
|
"eval_loss": 0.6583812832832336,
|
||
|
|
"eval_runtime": 175.3489,
|
||
|
|
"eval_samples_per_second": 17.263,
|
||
|
|
"eval_steps_per_second": 0.542,
|
||
|
|
"eval_wer": 0.41556736480150047,
|
||
|
|
"step": 70000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 52.48,
|
||
|
|
"learning_rate": 4.790022338049144e-05,
|
||
|
|
"loss": 0.0775,
|
||
|
|
"step": 71000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 52.48,
|
||
|
|
"eval_loss": 0.6354698538780212,
|
||
|
|
"eval_runtime": 173.883,
|
||
|
|
"eval_samples_per_second": 17.408,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.4156924038762113,
|
||
|
|
"step": 71000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 53.22,
|
||
|
|
"learning_rate": 4.7155621742367834e-05,
|
||
|
|
"loss": 0.0762,
|
||
|
|
"step": 72000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 53.22,
|
||
|
|
"eval_loss": 0.6745539307594299,
|
||
|
|
"eval_runtime": 174.1736,
|
||
|
|
"eval_samples_per_second": 17.379,
|
||
|
|
"eval_steps_per_second": 0.545,
|
||
|
|
"eval_wer": 0.40675211003438577,
|
||
|
|
"step": 72000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 53.95,
|
||
|
|
"learning_rate": 4.641176470588236e-05,
|
||
|
|
"loss": 0.0761,
|
||
|
|
"step": 73000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 53.95,
|
||
|
|
"eval_loss": 0.6509293913841248,
|
||
|
|
"eval_runtime": 176.7298,
|
||
|
|
"eval_samples_per_second": 17.128,
|
||
|
|
"eval_steps_per_second": 0.538,
|
||
|
|
"eval_wer": 0.4095029696780244,
|
||
|
|
"step": 73000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 54.69,
|
||
|
|
"learning_rate": 4.566716306775875e-05,
|
||
|
|
"loss": 0.0742,
|
||
|
|
"step": 74000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 54.69,
|
||
|
|
"eval_loss": 0.6646940112113953,
|
||
|
|
"eval_runtime": 177.5457,
|
||
|
|
"eval_samples_per_second": 17.049,
|
||
|
|
"eval_steps_per_second": 0.535,
|
||
|
|
"eval_wer": 0.41312910284463894,
|
||
|
|
"step": 74000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 55.43,
|
||
|
|
"learning_rate": 4.492330603127327e-05,
|
||
|
|
"loss": 0.0738,
|
||
|
|
"step": 75000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 55.43,
|
||
|
|
"eval_loss": 0.6746073365211487,
|
||
|
|
"eval_runtime": 182.277,
|
||
|
|
"eval_samples_per_second": 16.607,
|
||
|
|
"eval_steps_per_second": 0.521,
|
||
|
|
"eval_wer": 0.4097530478274461,
|
||
|
|
"step": 75000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 56.17,
|
||
|
|
"learning_rate": 4.417870439314967e-05,
|
||
|
|
"loss": 0.072,
|
||
|
|
"step": 76000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 56.17,
|
||
|
|
"eval_loss": 0.6681593060493469,
|
||
|
|
"eval_runtime": 189.5798,
|
||
|
|
"eval_samples_per_second": 15.967,
|
||
|
|
"eval_steps_per_second": 0.501,
|
||
|
|
"eval_wer": 0.41081587996248825,
|
||
|
|
"step": 76000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 56.91,
|
||
|
|
"learning_rate": 4.3434102755026064e-05,
|
||
|
|
"loss": 0.0715,
|
||
|
|
"step": 77000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 56.91,
|
||
|
|
"eval_loss": 0.6641326546669006,
|
||
|
|
"eval_runtime": 199.1459,
|
||
|
|
"eval_samples_per_second": 15.2,
|
||
|
|
"eval_steps_per_second": 0.477,
|
||
|
|
"eval_wer": 0.4114410753360425,
|
||
|
|
"step": 77000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 57.65,
|
||
|
|
"learning_rate": 4.2690245718540587e-05,
|
||
|
|
"loss": 0.0717,
|
||
|
|
"step": 78000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 57.65,
|
||
|
|
"eval_loss": 0.6959729790687561,
|
||
|
|
"eval_runtime": 198.8176,
|
||
|
|
"eval_samples_per_second": 15.225,
|
||
|
|
"eval_steps_per_second": 0.478,
|
||
|
|
"eval_wer": 0.41106595811191,
|
||
|
|
"step": 78000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 58.39,
|
||
|
|
"learning_rate": 4.194564408041698e-05,
|
||
|
|
"loss": 0.0711,
|
||
|
|
"step": 79000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 58.39,
|
||
|
|
"eval_loss": 0.663927435874939,
|
||
|
|
"eval_runtime": 201.2001,
|
||
|
|
"eval_samples_per_second": 15.045,
|
||
|
|
"eval_steps_per_second": 0.472,
|
||
|
|
"eval_wer": 0.4081900593935605,
|
||
|
|
"step": 79000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 59.13,
|
||
|
|
"learning_rate": 4.1201042442293374e-05,
|
||
|
|
"loss": 0.0706,
|
||
|
|
"step": 80000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 59.13,
|
||
|
|
"eval_loss": 0.6455733180046082,
|
||
|
|
"eval_runtime": 180.6704,
|
||
|
|
"eval_samples_per_second": 16.754,
|
||
|
|
"eval_steps_per_second": 0.526,
|
||
|
|
"eval_wer": 0.41156611441075336,
|
||
|
|
"step": 80000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 59.87,
|
||
|
|
"learning_rate": 4.045644080416977e-05,
|
||
|
|
"loss": 0.0685,
|
||
|
|
"step": 81000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 59.87,
|
||
|
|
"eval_loss": 0.6929445862770081,
|
||
|
|
"eval_runtime": 176.1838,
|
||
|
|
"eval_samples_per_second": 17.181,
|
||
|
|
"eval_steps_per_second": 0.539,
|
||
|
|
"eval_wer": 0.4169427946233198,
|
||
|
|
"step": 81000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 60.61,
|
||
|
|
"learning_rate": 3.971258376768429e-05,
|
||
|
|
"loss": 0.0666,
|
||
|
|
"step": 82000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 60.61,
|
||
|
|
"eval_loss": 0.6526830792427063,
|
||
|
|
"eval_runtime": 175.1162,
|
||
|
|
"eval_samples_per_second": 17.286,
|
||
|
|
"eval_steps_per_second": 0.542,
|
||
|
|
"eval_wer": 0.40987808690215694,
|
||
|
|
"step": 82000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 61.35,
|
||
|
|
"learning_rate": 3.896798212956069e-05,
|
||
|
|
"loss": 0.0677,
|
||
|
|
"step": 83000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 61.35,
|
||
|
|
"eval_loss": 0.708766758441925,
|
||
|
|
"eval_runtime": 174.527,
|
||
|
|
"eval_samples_per_second": 17.344,
|
||
|
|
"eval_steps_per_second": 0.544,
|
||
|
|
"eval_wer": 0.4113785557986871,
|
||
|
|
"step": 83000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 62.08,
|
||
|
|
"learning_rate": 3.822338049143708e-05,
|
||
|
|
"loss": 0.0667,
|
||
|
|
"step": 84000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 62.08,
|
||
|
|
"eval_loss": 0.6928330063819885,
|
||
|
|
"eval_runtime": 178.7497,
|
||
|
|
"eval_samples_per_second": 16.934,
|
||
|
|
"eval_steps_per_second": 0.531,
|
||
|
|
"eval_wer": 0.4085026570803376,
|
||
|
|
"step": 84000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 62.82,
|
||
|
|
"learning_rate": 3.74795234549516e-05,
|
||
|
|
"loss": 0.065,
|
||
|
|
"step": 85000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 62.82,
|
||
|
|
"eval_loss": 0.7296735048294067,
|
||
|
|
"eval_runtime": 174.5663,
|
||
|
|
"eval_samples_per_second": 17.34,
|
||
|
|
"eval_steps_per_second": 0.544,
|
||
|
|
"eval_wer": 0.40962800875273525,
|
||
|
|
"step": 85000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 63.56,
|
||
|
|
"learning_rate": 3.6734921816828e-05,
|
||
|
|
"loss": 0.0655,
|
||
|
|
"step": 86000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 63.56,
|
||
|
|
"eval_loss": 0.7482656836509705,
|
||
|
|
"eval_runtime": 173.9987,
|
||
|
|
"eval_samples_per_second": 17.397,
|
||
|
|
"eval_steps_per_second": 0.546,
|
||
|
|
"eval_wer": 0.4106908408877774,
|
||
|
|
"step": 86000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 64.3,
|
||
|
|
"learning_rate": 3.59903201787044e-05,
|
||
|
|
"loss": 0.0636,
|
||
|
|
"step": 87000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 64.3,
|
||
|
|
"eval_loss": 0.7502027750015259,
|
||
|
|
"eval_runtime": 172.4262,
|
||
|
|
"eval_samples_per_second": 17.555,
|
||
|
|
"eval_steps_per_second": 0.551,
|
||
|
|
"eval_wer": 0.41875586120662706,
|
||
|
|
"step": 87000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 65.04,
|
||
|
|
"learning_rate": 3.5245718540580794e-05,
|
||
|
|
"loss": 0.0645,
|
||
|
|
"step": 88000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 65.04,
|
||
|
|
"eval_loss": 0.7395268678665161,
|
||
|
|
"eval_runtime": 172.5532,
|
||
|
|
"eval_samples_per_second": 17.542,
|
||
|
|
"eval_steps_per_second": 0.551,
|
||
|
|
"eval_wer": 0.40793998124413877,
|
||
|
|
"step": 88000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 65.78,
|
||
|
|
"learning_rate": 3.450186150409531e-05,
|
||
|
|
"loss": 0.0628,
|
||
|
|
"step": 89000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 65.78,
|
||
|
|
"eval_loss": 0.6981734037399292,
|
||
|
|
"eval_runtime": 186.2785,
|
||
|
|
"eval_samples_per_second": 16.25,
|
||
|
|
"eval_steps_per_second": 0.51,
|
||
|
|
"eval_wer": 0.4075023444826508,
|
||
|
|
"step": 89000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 66.52,
|
||
|
|
"learning_rate": 3.375725986597171e-05,
|
||
|
|
"loss": 0.0618,
|
||
|
|
"step": 90000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 66.52,
|
||
|
|
"eval_loss": 0.7291818261146545,
|
||
|
|
"eval_runtime": 195.0277,
|
||
|
|
"eval_samples_per_second": 15.521,
|
||
|
|
"eval_steps_per_second": 0.487,
|
||
|
|
"eval_wer": 0.4113160362613317,
|
||
|
|
"step": 90000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 67.26,
|
||
|
|
"learning_rate": 3.301340282948623e-05,
|
||
|
|
"loss": 0.0612,
|
||
|
|
"step": 91000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 67.26,
|
||
|
|
"eval_loss": 0.7428854703903198,
|
||
|
|
"eval_runtime": 201.9566,
|
||
|
|
"eval_samples_per_second": 14.988,
|
||
|
|
"eval_steps_per_second": 0.47,
|
||
|
|
"eval_wer": 0.40994060643951236,
|
||
|
|
"step": 91000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 68.0,
|
||
|
|
"learning_rate": 3.2269545793000746e-05,
|
||
|
|
"loss": 0.0613,
|
||
|
|
"step": 92000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 68.0,
|
||
|
|
"eval_loss": 0.7314553260803223,
|
||
|
|
"eval_runtime": 181.5248,
|
||
|
|
"eval_samples_per_second": 16.675,
|
||
|
|
"eval_steps_per_second": 0.523,
|
||
|
|
"eval_wer": 0.41150359487339794,
|
||
|
|
"step": 92000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 68.74,
|
||
|
|
"learning_rate": 3.152494415487714e-05,
|
||
|
|
"loss": 0.0607,
|
||
|
|
"step": 93000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 68.74,
|
||
|
|
"eval_loss": 0.7307026386260986,
|
||
|
|
"eval_runtime": 263.2133,
|
||
|
|
"eval_samples_per_second": 11.5,
|
||
|
|
"eval_steps_per_second": 0.361,
|
||
|
|
"eval_wer": 0.40725226633322914,
|
||
|
|
"step": 93000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 69.48,
|
||
|
|
"learning_rate": 3.078034251675354e-05,
|
||
|
|
"loss": 0.0597,
|
||
|
|
"step": 94000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 69.48,
|
||
|
|
"eval_loss": 0.6700393557548523,
|
||
|
|
"eval_runtime": 261.1836,
|
||
|
|
"eval_samples_per_second": 11.59,
|
||
|
|
"eval_steps_per_second": 0.364,
|
||
|
|
"eval_wer": 0.40206314473272897,
|
||
|
|
"step": 94000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 70.21,
|
||
|
|
"learning_rate": 3.0035740878629936e-05,
|
||
|
|
"loss": 0.0586,
|
||
|
|
"step": 95000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 70.21,
|
||
|
|
"eval_loss": 0.7129278779029846,
|
||
|
|
"eval_runtime": 174.8869,
|
||
|
|
"eval_samples_per_second": 17.308,
|
||
|
|
"eval_steps_per_second": 0.543,
|
||
|
|
"eval_wer": 0.40737730540794,
|
||
|
|
"step": 95000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 70.95,
|
||
|
|
"learning_rate": 2.9291883842144452e-05,
|
||
|
|
"loss": 0.0585,
|
||
|
|
"step": 96000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 70.95,
|
||
|
|
"eval_loss": 0.7538037896156311,
|
||
|
|
"eval_runtime": 175.7698,
|
||
|
|
"eval_samples_per_second": 17.221,
|
||
|
|
"eval_steps_per_second": 0.54,
|
||
|
|
"eval_wer": 0.4142544545170366,
|
||
|
|
"step": 96000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 71.69,
|
||
|
|
"learning_rate": 2.8547282204020846e-05,
|
||
|
|
"loss": 0.0573,
|
||
|
|
"step": 97000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 71.69,
|
||
|
|
"eval_loss": 0.703994631767273,
|
||
|
|
"eval_runtime": 176.9961,
|
||
|
|
"eval_samples_per_second": 17.102,
|
||
|
|
"eval_steps_per_second": 0.537,
|
||
|
|
"eval_wer": 0.40487652391372303,
|
||
|
|
"step": 97000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 72.43,
|
||
|
|
"learning_rate": 2.780342516753537e-05,
|
||
|
|
"loss": 0.0575,
|
||
|
|
"step": 98000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 72.43,
|
||
|
|
"eval_loss": 0.7418561577796936,
|
||
|
|
"eval_runtime": 177.6888,
|
||
|
|
"eval_samples_per_second": 17.035,
|
||
|
|
"eval_steps_per_second": 0.535,
|
||
|
|
"eval_wer": 0.4070647077211629,
|
||
|
|
"step": 98000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 73.17,
|
||
|
|
"learning_rate": 2.7058823529411766e-05,
|
||
|
|
"loss": 0.0567,
|
||
|
|
"step": 99000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 73.17,
|
||
|
|
"eval_loss": 0.7433472871780396,
|
||
|
|
"eval_runtime": 175.8142,
|
||
|
|
"eval_samples_per_second": 17.217,
|
||
|
|
"eval_steps_per_second": 0.54,
|
||
|
|
"eval_wer": 0.4093154110659581,
|
||
|
|
"step": 99000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 73.91,
|
||
|
|
"learning_rate": 2.631422189128816e-05,
|
||
|
|
"loss": 0.054,
|
||
|
|
"step": 100000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 73.91,
|
||
|
|
"eval_loss": 0.6921048760414124,
|
||
|
|
"eval_runtime": 180.3464,
|
||
|
|
"eval_samples_per_second": 16.784,
|
||
|
|
"eval_steps_per_second": 0.527,
|
||
|
|
"eval_wer": 0.40256330103157234,
|
||
|
|
"step": 100000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 74.65,
|
||
|
|
"learning_rate": 2.5570364854802682e-05,
|
||
|
|
"loss": 0.0571,
|
||
|
|
"step": 101000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 74.65,
|
||
|
|
"eval_loss": 0.7526727318763733,
|
||
|
|
"eval_runtime": 176.5656,
|
||
|
|
"eval_samples_per_second": 17.144,
|
||
|
|
"eval_steps_per_second": 0.538,
|
||
|
|
"eval_wer": 0.4104407627383557,
|
||
|
|
"step": 101000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 75.39,
|
||
|
|
"learning_rate": 2.48265078183172e-05,
|
||
|
|
"loss": 0.0532,
|
||
|
|
"step": 102000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 75.39,
|
||
|
|
"eval_loss": 0.717917263507843,
|
||
|
|
"eval_runtime": 180.7508,
|
||
|
|
"eval_samples_per_second": 16.747,
|
||
|
|
"eval_steps_per_second": 0.526,
|
||
|
|
"eval_wer": 0.40693966864645204,
|
||
|
|
"step": 102000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 76.13,
|
||
|
|
"learning_rate": 2.40819061801936e-05,
|
||
|
|
"loss": 0.0549,
|
||
|
|
"step": 103000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 76.13,
|
||
|
|
"eval_loss": 0.7699146866798401,
|
||
|
|
"eval_runtime": 199.8758,
|
||
|
|
"eval_samples_per_second": 15.144,
|
||
|
|
"eval_steps_per_second": 0.475,
|
||
|
|
"eval_wer": 0.4080650203188496,
|
||
|
|
"step": 103000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 76.87,
|
||
|
|
"learning_rate": 2.3338049143708118e-05,
|
||
|
|
"loss": 0.0533,
|
||
|
|
"step": 104000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 76.87,
|
||
|
|
"eval_loss": 0.7196021676063538,
|
||
|
|
"eval_runtime": 186.8201,
|
||
|
|
"eval_samples_per_second": 16.203,
|
||
|
|
"eval_steps_per_second": 0.509,
|
||
|
|
"eval_wer": 0.4058143169740544,
|
||
|
|
"step": 104000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 77.61,
|
||
|
|
"learning_rate": 2.2593447505584515e-05,
|
||
|
|
"loss": 0.0538,
|
||
|
|
"step": 105000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 77.61,
|
||
|
|
"eval_loss": 0.7532729506492615,
|
||
|
|
"eval_runtime": 186.7527,
|
||
|
|
"eval_samples_per_second": 16.209,
|
||
|
|
"eval_steps_per_second": 0.509,
|
||
|
|
"eval_wer": 0.4068146295717412,
|
||
|
|
"step": 105000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 78.34,
|
||
|
|
"learning_rate": 2.1848845867460908e-05,
|
||
|
|
"loss": 0.0533,
|
||
|
|
"step": 106000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 78.34,
|
||
|
|
"eval_loss": 0.7514676451683044,
|
||
|
|
"eval_runtime": 203.9001,
|
||
|
|
"eval_samples_per_second": 14.846,
|
||
|
|
"eval_steps_per_second": 0.466,
|
||
|
|
"eval_wer": 0.40762738355736167,
|
||
|
|
"step": 106000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 79.08,
|
||
|
|
"learning_rate": 2.1104244229337305e-05,
|
||
|
|
"loss": 0.0517,
|
||
|
|
"step": 107000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 79.08,
|
||
|
|
"eval_loss": 0.766864001750946,
|
||
|
|
"eval_runtime": 183.9434,
|
||
|
|
"eval_samples_per_second": 16.456,
|
||
|
|
"eval_steps_per_second": 0.516,
|
||
|
|
"eval_wer": 0.4061894341981869,
|
||
|
|
"step": 107000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 79.82,
|
||
|
|
"learning_rate": 2.0359642591213702e-05,
|
||
|
|
"loss": 0.0521,
|
||
|
|
"step": 108000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 79.82,
|
||
|
|
"eval_loss": 0.7399023771286011,
|
||
|
|
"eval_runtime": 175.691,
|
||
|
|
"eval_samples_per_second": 17.229,
|
||
|
|
"eval_steps_per_second": 0.541,
|
||
|
|
"eval_wer": 0.4015004688965302,
|
||
|
|
"step": 108000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 80.56,
|
||
|
|
"learning_rate": 1.961578555472822e-05,
|
||
|
|
"loss": 0.0525,
|
||
|
|
"step": 109000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 80.56,
|
||
|
|
"eval_loss": 0.7517969608306885,
|
||
|
|
"eval_runtime": 185.511,
|
||
|
|
"eval_samples_per_second": 16.317,
|
||
|
|
"eval_steps_per_second": 0.512,
|
||
|
|
"eval_wer": 0.4044388871522351,
|
||
|
|
"step": 109000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 81.3,
|
||
|
|
"learning_rate": 1.887118391660462e-05,
|
||
|
|
"loss": 0.0505,
|
||
|
|
"step": 110000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 81.3,
|
||
|
|
"eval_loss": 0.7839346528053284,
|
||
|
|
"eval_runtime": 181.6126,
|
||
|
|
"eval_samples_per_second": 16.667,
|
||
|
|
"eval_steps_per_second": 0.523,
|
||
|
|
"eval_wer": 0.4077524226320725,
|
||
|
|
"step": 110000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 82.04,
|
||
|
|
"learning_rate": 1.8126582278481012e-05,
|
||
|
|
"loss": 0.0508,
|
||
|
|
"step": 111000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 82.04,
|
||
|
|
"eval_loss": 0.7692478895187378,
|
||
|
|
"eval_runtime": 182.2532,
|
||
|
|
"eval_samples_per_second": 16.609,
|
||
|
|
"eval_steps_per_second": 0.521,
|
||
|
|
"eval_wer": 0.4070647077211629,
|
||
|
|
"step": 111000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 82.78,
|
||
|
|
"learning_rate": 1.7383469843633657e-05,
|
||
|
|
"loss": 0.0496,
|
||
|
|
"step": 112000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 82.78,
|
||
|
|
"eval_loss": 0.7821455597877502,
|
||
|
|
"eval_runtime": 227.3787,
|
||
|
|
"eval_samples_per_second": 13.313,
|
||
|
|
"eval_steps_per_second": 0.418,
|
||
|
|
"eval_wer": 0.40987808690215694,
|
||
|
|
"step": 112000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 83.52,
|
||
|
|
"learning_rate": 1.663886820551005e-05,
|
||
|
|
"loss": 0.0493,
|
||
|
|
"step": 113000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 83.52,
|
||
|
|
"eval_loss": 0.7528874278068542,
|
||
|
|
"eval_runtime": 217.8709,
|
||
|
|
"eval_samples_per_second": 13.894,
|
||
|
|
"eval_steps_per_second": 0.436,
|
||
|
|
"eval_wer": 0.4055642388246327,
|
||
|
|
"step": 113000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 84.26,
|
||
|
|
"learning_rate": 1.5894266567386448e-05,
|
||
|
|
"loss": 0.0498,
|
||
|
|
"step": 114000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 84.26,
|
||
|
|
"eval_loss": 0.7717328071594238,
|
||
|
|
"eval_runtime": 216.4941,
|
||
|
|
"eval_samples_per_second": 13.982,
|
||
|
|
"eval_steps_per_second": 0.439,
|
||
|
|
"eval_wer": 0.40493904345107845,
|
||
|
|
"step": 114000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 85.0,
|
||
|
|
"learning_rate": 1.5149664929262845e-05,
|
||
|
|
"loss": 0.0499,
|
||
|
|
"step": 115000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 85.0,
|
||
|
|
"eval_loss": 0.757793128490448,
|
||
|
|
"eval_runtime": 213.4299,
|
||
|
|
"eval_samples_per_second": 14.183,
|
||
|
|
"eval_steps_per_second": 0.445,
|
||
|
|
"eval_wer": 0.40406376992810256,
|
||
|
|
"step": 115000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 85.74,
|
||
|
|
"learning_rate": 1.4405807892777364e-05,
|
||
|
|
"loss": 0.0485,
|
||
|
|
"step": 116000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 85.74,
|
||
|
|
"eval_loss": 0.7553907632827759,
|
||
|
|
"eval_runtime": 181.6761,
|
||
|
|
"eval_samples_per_second": 16.662,
|
||
|
|
"eval_steps_per_second": 0.523,
|
||
|
|
"eval_wer": 0.4044388871522351,
|
||
|
|
"step": 116000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 86.47,
|
||
|
|
"learning_rate": 1.3661206254653761e-05,
|
||
|
|
"loss": 0.0479,
|
||
|
|
"step": 117000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 86.47,
|
||
|
|
"eval_loss": 0.8072551488876343,
|
||
|
|
"eval_runtime": 188.7645,
|
||
|
|
"eval_samples_per_second": 16.036,
|
||
|
|
"eval_steps_per_second": 0.503,
|
||
|
|
"eval_wer": 0.40625195373554235,
|
||
|
|
"step": 117000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 87.21,
|
||
|
|
"learning_rate": 1.2917349218168282e-05,
|
||
|
|
"loss": 0.0462,
|
||
|
|
"step": 118000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 87.21,
|
||
|
|
"eval_loss": 0.7592765092849731,
|
||
|
|
"eval_runtime": 187.3254,
|
||
|
|
"eval_samples_per_second": 16.159,
|
||
|
|
"eval_steps_per_second": 0.507,
|
||
|
|
"eval_wer": 0.4025007814942169,
|
||
|
|
"step": 118000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 87.95,
|
||
|
|
"learning_rate": 1.2172747580044677e-05,
|
||
|
|
"loss": 0.0475,
|
||
|
|
"step": 119000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 87.95,
|
||
|
|
"eval_loss": 0.7812408208847046,
|
||
|
|
"eval_runtime": 227.3028,
|
||
|
|
"eval_samples_per_second": 13.317,
|
||
|
|
"eval_steps_per_second": 0.418,
|
||
|
|
"eval_wer": 0.40493904345107845,
|
||
|
|
"step": 119000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 88.69,
|
||
|
|
"learning_rate": 1.1428890543559197e-05,
|
||
|
|
"loss": 0.0455,
|
||
|
|
"step": 120000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 88.69,
|
||
|
|
"eval_loss": 0.779983401298523,
|
||
|
|
"eval_runtime": 223.0101,
|
||
|
|
"eval_samples_per_second": 13.573,
|
||
|
|
"eval_steps_per_second": 0.426,
|
||
|
|
"eval_wer": 0.40487652391372303,
|
||
|
|
"step": 120000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 89.43,
|
||
|
|
"learning_rate": 1.0684288905435592e-05,
|
||
|
|
"loss": 0.0451,
|
||
|
|
"step": 121000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 89.43,
|
||
|
|
"eval_loss": 0.8092254400253296,
|
||
|
|
"eval_runtime": 251.7762,
|
||
|
|
"eval_samples_per_second": 12.023,
|
||
|
|
"eval_steps_per_second": 0.377,
|
||
|
|
"eval_wer": 0.40737730540794,
|
||
|
|
"step": 121000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 90.17,
|
||
|
|
"learning_rate": 9.939687267311989e-06,
|
||
|
|
"loss": 0.0455,
|
||
|
|
"step": 122000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 90.17,
|
||
|
|
"eval_loss": 0.7754979133605957,
|
||
|
|
"eval_runtime": 271.2522,
|
||
|
|
"eval_samples_per_second": 11.159,
|
||
|
|
"eval_steps_per_second": 0.35,
|
||
|
|
"eval_wer": 0.4074398249452954,
|
||
|
|
"step": 122000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 90.91,
|
||
|
|
"learning_rate": 9.195085629188386e-06,
|
||
|
|
"loss": 0.0476,
|
||
|
|
"step": 123000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 90.91,
|
||
|
|
"eval_loss": 0.7846540212631226,
|
||
|
|
"eval_runtime": 244.3416,
|
||
|
|
"eval_samples_per_second": 12.388,
|
||
|
|
"eval_steps_per_second": 0.389,
|
||
|
|
"eval_wer": 0.40593935604876524,
|
||
|
|
"step": 123000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 91.65,
|
||
|
|
"learning_rate": 8.451228592702905e-06,
|
||
|
|
"loss": 0.0461,
|
||
|
|
"step": 124000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 91.65,
|
||
|
|
"eval_loss": 0.7784237861633301,
|
||
|
|
"eval_runtime": 267.8746,
|
||
|
|
"eval_samples_per_second": 11.3,
|
||
|
|
"eval_steps_per_second": 0.355,
|
||
|
|
"eval_wer": 0.4034385745545483,
|
||
|
|
"step": 124000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 92.39,
|
||
|
|
"learning_rate": 7.7066269545793e-06,
|
||
|
|
"loss": 0.0454,
|
||
|
|
"step": 125000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 92.39,
|
||
|
|
"eval_loss": 0.7823372483253479,
|
||
|
|
"eval_runtime": 207.7458,
|
||
|
|
"eval_samples_per_second": 14.571,
|
||
|
|
"eval_steps_per_second": 0.457,
|
||
|
|
"eval_wer": 0.403251015942482,
|
||
|
|
"step": 125000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 93.13,
|
||
|
|
"learning_rate": 6.9620253164556965e-06,
|
||
|
|
"loss": 0.0453,
|
||
|
|
"step": 126000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 93.13,
|
||
|
|
"eval_loss": 0.7685350179672241,
|
||
|
|
"eval_runtime": 180.1025,
|
||
|
|
"eval_samples_per_second": 16.807,
|
||
|
|
"eval_steps_per_second": 0.527,
|
||
|
|
"eval_wer": 0.4015004688965302,
|
||
|
|
"step": 126000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 93.87,
|
||
|
|
"learning_rate": 6.2174236783320925e-06,
|
||
|
|
"loss": 0.0451,
|
||
|
|
"step": 127000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 93.87,
|
||
|
|
"eval_loss": 0.7680614590644836,
|
||
|
|
"eval_runtime": 177.9917,
|
||
|
|
"eval_samples_per_second": 17.006,
|
||
|
|
"eval_steps_per_second": 0.534,
|
||
|
|
"eval_wer": 0.4,
|
||
|
|
"step": 127000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 94.6,
|
||
|
|
"learning_rate": 5.473566641846612e-06,
|
||
|
|
"loss": 0.0428,
|
||
|
|
"step": 128000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 94.6,
|
||
|
|
"eval_loss": 0.7936143279075623,
|
||
|
|
"eval_runtime": 180.9951,
|
||
|
|
"eval_samples_per_second": 16.724,
|
||
|
|
"eval_steps_per_second": 0.525,
|
||
|
|
"eval_wer": 0.4034385745545483,
|
||
|
|
"step": 128000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 95.34,
|
||
|
|
"learning_rate": 4.728965003723009e-06,
|
||
|
|
"loss": 0.0445,
|
||
|
|
"step": 129000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 95.34,
|
||
|
|
"eval_loss": 0.7666231989860535,
|
||
|
|
"eval_runtime": 203.0139,
|
||
|
|
"eval_samples_per_second": 14.91,
|
||
|
|
"eval_steps_per_second": 0.468,
|
||
|
|
"eval_wer": 0.39968740231322286,
|
||
|
|
"step": 129000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 96.08,
|
||
|
|
"learning_rate": 3.984363365599404e-06,
|
||
|
|
"loss": 0.0441,
|
||
|
|
"step": 130000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 96.08,
|
||
|
|
"eval_loss": 0.7681093811988831,
|
||
|
|
"eval_runtime": 206.6264,
|
||
|
|
"eval_samples_per_second": 14.65,
|
||
|
|
"eval_steps_per_second": 0.46,
|
||
|
|
"eval_wer": 0.4008752735229759,
|
||
|
|
"step": 130000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 96.82,
|
||
|
|
"learning_rate": 3.2412509307520474e-06,
|
||
|
|
"loss": 0.043,
|
||
|
|
"step": 131000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 96.82,
|
||
|
|
"eval_loss": 0.7704082131385803,
|
||
|
|
"eval_runtime": 196.2406,
|
||
|
|
"eval_samples_per_second": 15.425,
|
||
|
|
"eval_steps_per_second": 0.484,
|
||
|
|
"eval_wer": 0.4008752735229759,
|
||
|
|
"step": 131000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 97.56,
|
||
|
|
"learning_rate": 2.496649292628444e-06,
|
||
|
|
"loss": 0.043,
|
||
|
|
"step": 132000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 97.56,
|
||
|
|
"eval_loss": 0.7688604593276978,
|
||
|
|
"eval_runtime": 208.2948,
|
||
|
|
"eval_samples_per_second": 14.532,
|
||
|
|
"eval_steps_per_second": 0.456,
|
||
|
|
"eval_wer": 0.3981244138793373,
|
||
|
|
"step": 132000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 98.3,
|
||
|
|
"learning_rate": 1.7520476545048399e-06,
|
||
|
|
"loss": 0.0429,
|
||
|
|
"step": 133000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 98.3,
|
||
|
|
"eval_loss": 0.7736847996711731,
|
||
|
|
"eval_runtime": 207.1539,
|
||
|
|
"eval_samples_per_second": 14.612,
|
||
|
|
"eval_steps_per_second": 0.459,
|
||
|
|
"eval_wer": 0.3992497655517349,
|
||
|
|
"step": 133000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 99.04,
|
||
|
|
"learning_rate": 1.0081906180193596e-06,
|
||
|
|
"loss": 0.0434,
|
||
|
|
"step": 134000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 99.04,
|
||
|
|
"eval_loss": 0.7795138955116272,
|
||
|
|
"eval_runtime": 196.4909,
|
||
|
|
"eval_samples_per_second": 15.405,
|
||
|
|
"eval_steps_per_second": 0.483,
|
||
|
|
"eval_wer": 0.4012503907471085,
|
||
|
|
"step": 134000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 99.78,
|
||
|
|
"learning_rate": 2.635889798957558e-07,
|
||
|
|
"loss": 0.0433,
|
||
|
|
"step": 135000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 99.78,
|
||
|
|
"eval_loss": 0.7773757576942444,
|
||
|
|
"eval_runtime": 198.9794,
|
||
|
|
"eval_samples_per_second": 15.213,
|
||
|
|
"eval_steps_per_second": 0.477,
|
||
|
|
"eval_wer": 0.4008127539856205,
|
||
|
|
"step": 135000
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"max_steps": 135300,
|
||
|
|
"num_train_epochs": 100,
|
||
|
|
"total_flos": 1.184566523613751e+21,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|