wav2vec2-large-xls-r-300m-tr / trainer_state.json
emre's picture
End of training
7128c74
raw
history blame
57.7 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 100.0,
"global_step": 78400,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.64,
"learning_rate": 1.9880000000000003e-05,
"loss": 6.8222,
"step": 500
},
{
"epoch": 0.64,
"eval_loss": 3.502612352371216,
"eval_runtime": 238.4374,
"eval_samples_per_second": 34.974,
"eval_steps_per_second": 1.095,
"eval_wer": 1.0,
"step": 500
},
{
"epoch": 1.28,
"learning_rate": 1.987240051347882e-05,
"loss": 3.2136,
"step": 1000
},
{
"epoch": 1.28,
"eval_loss": 3.0592851638793945,
"eval_runtime": 239.7957,
"eval_samples_per_second": 34.775,
"eval_steps_per_second": 1.088,
"eval_wer": 0.9999770836675298,
"step": 1000
},
{
"epoch": 1.91,
"learning_rate": 1.9744030808729143e-05,
"loss": 2.8882,
"step": 1500
},
{
"epoch": 1.91,
"eval_loss": 2.4670279026031494,
"eval_runtime": 236.3109,
"eval_samples_per_second": 35.288,
"eval_steps_per_second": 1.104,
"eval_wer": 0.9938584228979994,
"step": 1500
},
{
"epoch": 2.55,
"learning_rate": 1.9615661103979463e-05,
"loss": 2.3743,
"step": 2000
},
{
"epoch": 2.55,
"eval_loss": 1.1843628883361816,
"eval_runtime": 239.6417,
"eval_samples_per_second": 34.798,
"eval_steps_per_second": 1.089,
"eval_wer": 0.8656873753924422,
"step": 2000
},
{
"epoch": 3.19,
"learning_rate": 1.9487291399229784e-05,
"loss": 1.9456,
"step": 2500
},
{
"epoch": 3.19,
"eval_loss": 0.8227887153625488,
"eval_runtime": 238.3867,
"eval_samples_per_second": 34.981,
"eval_steps_per_second": 1.095,
"eval_wer": 0.7397162958040195,
"step": 2500
},
{
"epoch": 3.83,
"learning_rate": 1.9358921694480104e-05,
"loss": 1.7781,
"step": 3000
},
{
"epoch": 3.83,
"eval_loss": 0.6826250553131104,
"eval_runtime": 237.1431,
"eval_samples_per_second": 35.164,
"eval_steps_per_second": 1.101,
"eval_wer": 0.6752984852304237,
"step": 3000
},
{
"epoch": 4.46,
"learning_rate": 1.9230551989730427e-05,
"loss": 1.6848,
"step": 3500
},
{
"epoch": 4.46,
"eval_loss": 0.5885177254676819,
"eval_runtime": 235.3547,
"eval_samples_per_second": 35.432,
"eval_steps_per_second": 1.109,
"eval_wer": 0.6139743795402983,
"step": 3500
},
{
"epoch": 5.1,
"learning_rate": 1.9102182284980744e-05,
"loss": 1.6228,
"step": 4000
},
{
"epoch": 5.1,
"eval_loss": 0.527351975440979,
"eval_runtime": 237.0762,
"eval_samples_per_second": 35.174,
"eval_steps_per_second": 1.101,
"eval_wer": 0.5788894745284965,
"step": 4000
},
{
"epoch": 5.74,
"learning_rate": 1.8973812580231068e-05,
"loss": 1.5768,
"step": 4500
},
{
"epoch": 5.74,
"eval_loss": 0.48999691009521484,
"eval_runtime": 235.6055,
"eval_samples_per_second": 35.394,
"eval_steps_per_second": 1.108,
"eval_wer": 0.5518711185461879,
"step": 4500
},
{
"epoch": 6.38,
"learning_rate": 1.8845699614890885e-05,
"loss": 1.5431,
"step": 5000
},
{
"epoch": 6.38,
"eval_loss": 0.45079049468040466,
"eval_runtime": 235.5935,
"eval_samples_per_second": 35.396,
"eval_steps_per_second": 1.108,
"eval_wer": 0.5237986112702523,
"step": 5000
},
{
"epoch": 7.02,
"learning_rate": 1.871732991014121e-05,
"loss": 1.5019,
"step": 5500
},
{
"epoch": 7.02,
"eval_loss": 0.42483240365982056,
"eval_runtime": 235.975,
"eval_samples_per_second": 35.338,
"eval_steps_per_second": 1.106,
"eval_wer": 0.5021426770859592,
"step": 5500
},
{
"epoch": 7.65,
"learning_rate": 1.858921694480103e-05,
"loss": 1.4684,
"step": 6000
},
{
"epoch": 7.65,
"eval_loss": 0.4009439945220947,
"eval_runtime": 236.3695,
"eval_samples_per_second": 35.28,
"eval_steps_per_second": 1.104,
"eval_wer": 0.4827096271512707,
"step": 6000
},
{
"epoch": 8.29,
"learning_rate": 1.846084724005135e-05,
"loss": 1.4635,
"step": 6500
},
{
"epoch": 8.29,
"eval_loss": 0.3830302357673645,
"eval_runtime": 234.8449,
"eval_samples_per_second": 35.509,
"eval_steps_per_second": 1.111,
"eval_wer": 0.4700139789628068,
"step": 6500
},
{
"epoch": 8.93,
"learning_rate": 1.833247753530167e-05,
"loss": 1.4291,
"step": 7000
},
{
"epoch": 8.93,
"eval_loss": 0.3707072138786316,
"eval_runtime": 237.9649,
"eval_samples_per_second": 35.043,
"eval_steps_per_second": 1.097,
"eval_wer": 0.45947246602653713,
"step": 7000
},
{
"epoch": 9.57,
"learning_rate": 1.820410783055199e-05,
"loss": 1.4271,
"step": 7500
},
{
"epoch": 9.57,
"eval_loss": 0.3570108115673065,
"eval_runtime": 236.9986,
"eval_samples_per_second": 35.186,
"eval_steps_per_second": 1.101,
"eval_wer": 0.4514059169970438,
"step": 7500
},
{
"epoch": 10.2,
"learning_rate": 1.807573812580231e-05,
"loss": 1.3938,
"step": 8000
},
{
"epoch": 10.2,
"eval_loss": 0.34792855381965637,
"eval_runtime": 234.9447,
"eval_samples_per_second": 35.493,
"eval_steps_per_second": 1.111,
"eval_wer": 0.4378394481747141,
"step": 8000
},
{
"epoch": 10.84,
"learning_rate": 1.794762516046213e-05,
"loss": 1.3914,
"step": 8500
},
{
"epoch": 10.84,
"eval_loss": 0.3395732045173645,
"eval_runtime": 235.1285,
"eval_samples_per_second": 35.466,
"eval_steps_per_second": 1.11,
"eval_wer": 0.4368082132135573,
"step": 8500
},
{
"epoch": 11.48,
"learning_rate": 1.7819255455712454e-05,
"loss": 1.3767,
"step": 9000
},
{
"epoch": 11.48,
"eval_loss": 0.3253381848335266,
"eval_runtime": 238.3808,
"eval_samples_per_second": 34.982,
"eval_steps_per_second": 1.095,
"eval_wer": 0.42622086761234734,
"step": 9000
},
{
"epoch": 12.12,
"learning_rate": 1.7690885750962775e-05,
"loss": 1.3641,
"step": 9500
},
{
"epoch": 12.12,
"eval_loss": 0.32509171962738037,
"eval_runtime": 234.3424,
"eval_samples_per_second": 35.585,
"eval_steps_per_second": 1.114,
"eval_wer": 0.4178105735958017,
"step": 9500
},
{
"epoch": 12.76,
"learning_rate": 1.7562772785622595e-05,
"loss": 1.355,
"step": 10000
},
{
"epoch": 12.76,
"eval_loss": 0.3138255476951599,
"eval_runtime": 234.3344,
"eval_samples_per_second": 35.586,
"eval_steps_per_second": 1.114,
"eval_wer": 0.4135710520888237,
"step": 10000
},
{
"epoch": 13.39,
"learning_rate": 1.7434403080872915e-05,
"loss": 1.336,
"step": 10500
},
{
"epoch": 13.39,
"eval_loss": 0.31212833523750305,
"eval_runtime": 238.4088,
"eval_samples_per_second": 34.978,
"eval_steps_per_second": 1.095,
"eval_wer": 0.40694823200494995,
"step": 10500
},
{
"epoch": 14.03,
"learning_rate": 1.7306033376123236e-05,
"loss": 1.3292,
"step": 11000
},
{
"epoch": 14.03,
"eval_loss": 0.30410972237586975,
"eval_runtime": 233.9502,
"eval_samples_per_second": 35.644,
"eval_steps_per_second": 1.116,
"eval_wer": 0.4014483122121136,
"step": 11000
},
{
"epoch": 14.67,
"learning_rate": 1.7177663671373556e-05,
"loss": 1.3249,
"step": 11500
},
{
"epoch": 14.67,
"eval_loss": 0.3013543486595154,
"eval_runtime": 234.7846,
"eval_samples_per_second": 35.518,
"eval_steps_per_second": 1.112,
"eval_wer": 0.3931296835254486,
"step": 11500
},
{
"epoch": 15.31,
"learning_rate": 1.704929396662388e-05,
"loss": 1.3156,
"step": 12000
},
{
"epoch": 15.31,
"eval_loss": 0.301371306180954,
"eval_runtime": 237.6831,
"eval_samples_per_second": 35.085,
"eval_steps_per_second": 1.098,
"eval_wer": 0.39294635286568735,
"step": 12000
},
{
"epoch": 15.94,
"learning_rate": 1.69209242618742e-05,
"loss": 1.313,
"step": 12500
},
{
"epoch": 15.94,
"eval_loss": 0.29687899351119995,
"eval_runtime": 234.864,
"eval_samples_per_second": 35.506,
"eval_steps_per_second": 1.111,
"eval_wer": 0.39679629672067285,
"step": 12500
},
{
"epoch": 16.58,
"learning_rate": 1.679255455712452e-05,
"loss": 1.3068,
"step": 13000
},
{
"epoch": 16.58,
"eval_loss": 0.2965061664581299,
"eval_runtime": 239.4104,
"eval_samples_per_second": 34.831,
"eval_steps_per_second": 1.09,
"eval_wer": 0.39663588239338177,
"step": 13000
},
{
"epoch": 17.22,
"learning_rate": 1.666444159178434e-05,
"loss": 1.2785,
"step": 13500
},
{
"epoch": 17.22,
"eval_loss": 0.29434436559677124,
"eval_runtime": 241.1756,
"eval_samples_per_second": 34.576,
"eval_steps_per_second": 1.082,
"eval_wer": 0.38497146916607466,
"step": 13500
},
{
"epoch": 17.86,
"learning_rate": 1.653607188703466e-05,
"loss": 1.2867,
"step": 14000
},
{
"epoch": 17.86,
"eval_loss": 0.291163831949234,
"eval_runtime": 234.1726,
"eval_samples_per_second": 35.61,
"eval_steps_per_second": 1.115,
"eval_wer": 0.37821115108737996,
"step": 14000
},
{
"epoch": 18.49,
"learning_rate": 1.640770218228498e-05,
"loss": 1.2714,
"step": 14500
},
{
"epoch": 18.49,
"eval_loss": 0.281938374042511,
"eval_runtime": 233.7302,
"eval_samples_per_second": 35.678,
"eval_steps_per_second": 1.117,
"eval_wer": 0.3746820358869766,
"step": 14500
},
{
"epoch": 19.13,
"learning_rate": 1.6279332477535305e-05,
"loss": 1.2844,
"step": 15000
},
{
"epoch": 19.13,
"eval_loss": 0.28400421142578125,
"eval_runtime": 234.7428,
"eval_samples_per_second": 35.524,
"eval_steps_per_second": 1.112,
"eval_wer": 0.3740403785778124,
"step": 15000
},
{
"epoch": 19.77,
"learning_rate": 1.6150962772785625e-05,
"loss": 1.2684,
"step": 15500
},
{
"epoch": 19.77,
"eval_loss": 0.2913394272327423,
"eval_runtime": 238.9443,
"eval_samples_per_second": 34.899,
"eval_steps_per_second": 1.092,
"eval_wer": 0.38279441758141025,
"step": 15500
},
{
"epoch": 20.41,
"learning_rate": 1.6022849807445445e-05,
"loss": 1.26,
"step": 16000
},
{
"epoch": 20.41,
"eval_loss": 0.27389487624168396,
"eval_runtime": 235.3865,
"eval_samples_per_second": 35.427,
"eval_steps_per_second": 1.109,
"eval_wer": 0.3674404748264088,
"step": 16000
},
{
"epoch": 21.05,
"learning_rate": 1.5894736842105266e-05,
"loss": 1.2543,
"step": 16500
},
{
"epoch": 21.05,
"eval_loss": 0.2739649713039398,
"eval_runtime": 234.6769,
"eval_samples_per_second": 35.534,
"eval_steps_per_second": 1.112,
"eval_wer": 0.36909045076425967,
"step": 16500
},
{
"epoch": 21.68,
"learning_rate": 1.5766367137355586e-05,
"loss": 1.2532,
"step": 17000
},
{
"epoch": 21.68,
"eval_loss": 0.2709480822086334,
"eval_runtime": 236.2528,
"eval_samples_per_second": 35.297,
"eval_steps_per_second": 1.105,
"eval_wer": 0.37562160551825285,
"step": 17000
},
{
"epoch": 22.32,
"learning_rate": 1.5637997432605906e-05,
"loss": 1.2409,
"step": 17500
},
{
"epoch": 22.32,
"eval_loss": 0.266906201839447,
"eval_runtime": 233.9016,
"eval_samples_per_second": 35.652,
"eval_steps_per_second": 1.116,
"eval_wer": 0.35932809313197517,
"step": 17500
},
{
"epoch": 22.96,
"learning_rate": 1.5509627727856227e-05,
"loss": 1.2404,
"step": 18000
},
{
"epoch": 22.96,
"eval_loss": 0.2673003077507019,
"eval_runtime": 235.0288,
"eval_samples_per_second": 35.481,
"eval_steps_per_second": 1.111,
"eval_wer": 0.35763228452918394,
"step": 18000
},
{
"epoch": 23.6,
"learning_rate": 1.538125802310655e-05,
"loss": 1.2347,
"step": 18500
},
{
"epoch": 23.6,
"eval_loss": 0.2678009271621704,
"eval_runtime": 234.3074,
"eval_samples_per_second": 35.59,
"eval_steps_per_second": 1.114,
"eval_wer": 0.36427802094552786,
"step": 18500
},
{
"epoch": 24.23,
"learning_rate": 1.5252888318356869e-05,
"loss": 1.2351,
"step": 19000
},
{
"epoch": 24.23,
"eval_loss": 0.2715151607990265,
"eval_runtime": 239.5626,
"eval_samples_per_second": 34.809,
"eval_steps_per_second": 1.089,
"eval_wer": 0.36498842725210257,
"step": 19000
},
{
"epoch": 24.87,
"learning_rate": 1.512451861360719e-05,
"loss": 1.2409,
"step": 19500
},
{
"epoch": 24.87,
"eval_loss": 0.2636788487434387,
"eval_runtime": 234.5142,
"eval_samples_per_second": 35.559,
"eval_steps_per_second": 1.113,
"eval_wer": 0.3570822925499003,
"step": 19500
},
{
"epoch": 25.51,
"learning_rate": 1.4996148908857511e-05,
"loss": 1.2152,
"step": 20000
},
{
"epoch": 25.51,
"eval_loss": 0.2785368859767914,
"eval_runtime": 234.4552,
"eval_samples_per_second": 35.568,
"eval_steps_per_second": 1.113,
"eval_wer": 0.36088640373994546,
"step": 20000
},
{
"epoch": 26.15,
"learning_rate": 1.4868035943517332e-05,
"loss": 1.2046,
"step": 20500
},
{
"epoch": 26.15,
"eval_loss": 0.2610293924808502,
"eval_runtime": 233.6859,
"eval_samples_per_second": 35.685,
"eval_steps_per_second": 1.117,
"eval_wer": 0.3508032174530788,
"step": 20500
},
{
"epoch": 26.79,
"learning_rate": 1.4739666238767652e-05,
"loss": 1.2082,
"step": 21000
},
{
"epoch": 26.79,
"eval_loss": 0.26189225912094116,
"eval_runtime": 234.8807,
"eval_samples_per_second": 35.503,
"eval_steps_per_second": 1.111,
"eval_wer": 0.3460595366317575,
"step": 21000
},
{
"epoch": 27.42,
"learning_rate": 1.4611296534017974e-05,
"loss": 1.2109,
"step": 21500
},
{
"epoch": 27.42,
"eval_loss": 0.2596839666366577,
"eval_runtime": 235.2549,
"eval_samples_per_second": 35.447,
"eval_steps_per_second": 1.109,
"eval_wer": 0.3501615601439146,
"step": 21500
},
{
"epoch": 28.06,
"learning_rate": 1.4482926829268294e-05,
"loss": 1.2014,
"step": 22000
},
{
"epoch": 28.06,
"eval_loss": 0.2608143091201782,
"eval_runtime": 234.8521,
"eval_samples_per_second": 35.507,
"eval_steps_per_second": 1.111,
"eval_wer": 0.3468386919357426,
"step": 22000
},
{
"epoch": 28.7,
"learning_rate": 1.4354813863928115e-05,
"loss": 1.1948,
"step": 22500
},
{
"epoch": 28.7,
"eval_loss": 0.2573413550853729,
"eval_runtime": 233.0487,
"eval_samples_per_second": 35.782,
"eval_steps_per_second": 1.12,
"eval_wer": 0.3456699589797649,
"step": 22500
},
{
"epoch": 29.34,
"learning_rate": 1.4226444159178435e-05,
"loss": 1.205,
"step": 23000
},
{
"epoch": 29.34,
"eval_loss": 0.2619355320930481,
"eval_runtime": 236.4546,
"eval_samples_per_second": 35.267,
"eval_steps_per_second": 1.104,
"eval_wer": 0.3463803652863396,
"step": 23000
},
{
"epoch": 29.97,
"learning_rate": 1.4098074454428757e-05,
"loss": 1.2019,
"step": 23500
},
{
"epoch": 29.97,
"eval_loss": 0.2558630406856537,
"eval_runtime": 235.5761,
"eval_samples_per_second": 35.398,
"eval_steps_per_second": 1.108,
"eval_wer": 0.34738868391502625,
"step": 23500
},
{
"epoch": 30.61,
"learning_rate": 1.3969704749679077e-05,
"loss": 1.1917,
"step": 24000
},
{
"epoch": 30.61,
"eval_loss": 0.2601245045661926,
"eval_runtime": 235.3838,
"eval_samples_per_second": 35.427,
"eval_steps_per_second": 1.109,
"eval_wer": 0.34619703462657836,
"step": 24000
},
{
"epoch": 31.25,
"learning_rate": 1.3841591784338897e-05,
"loss": 1.1939,
"step": 24500
},
{
"epoch": 31.25,
"eval_loss": 0.2575133740901947,
"eval_runtime": 235.7288,
"eval_samples_per_second": 35.375,
"eval_steps_per_second": 1.107,
"eval_wer": 0.33874922657377915,
"step": 24500
},
{
"epoch": 31.89,
"learning_rate": 1.3713478818998718e-05,
"loss": 1.1882,
"step": 25000
},
{
"epoch": 31.89,
"eval_loss": 0.2534877359867096,
"eval_runtime": 235.7724,
"eval_samples_per_second": 35.369,
"eval_steps_per_second": 1.107,
"eval_wer": 0.33675550564887596,
"step": 25000
},
{
"epoch": 32.53,
"learning_rate": 1.3585109114249038e-05,
"loss": 1.191,
"step": 25500
},
{
"epoch": 32.53,
"eval_loss": 0.24889491498470306,
"eval_runtime": 236.163,
"eval_samples_per_second": 35.31,
"eval_steps_per_second": 1.105,
"eval_wer": 0.336457593326764,
"step": 25500
},
{
"epoch": 33.16,
"learning_rate": 1.345673940949936e-05,
"loss": 1.1767,
"step": 26000
},
{
"epoch": 33.16,
"eval_loss": 0.2501368224620819,
"eval_runtime": 236.8496,
"eval_samples_per_second": 35.208,
"eval_steps_per_second": 1.102,
"eval_wer": 0.3347159520590325,
"step": 26000
},
{
"epoch": 33.8,
"learning_rate": 1.332836970474968e-05,
"loss": 1.167,
"step": 26500
},
{
"epoch": 33.8,
"eval_loss": 0.25037434697151184,
"eval_runtime": 233.4394,
"eval_samples_per_second": 35.722,
"eval_steps_per_second": 1.118,
"eval_wer": 0.33469303572656234,
"step": 26500
},
{
"epoch": 34.44,
"learning_rate": 1.3200000000000002e-05,
"loss": 1.1678,
"step": 27000
},
{
"epoch": 34.44,
"eval_loss": 0.24800854921340942,
"eval_runtime": 233.3087,
"eval_samples_per_second": 35.742,
"eval_steps_per_second": 1.119,
"eval_wer": 0.33783257327497307,
"step": 27000
},
{
"epoch": 35.08,
"learning_rate": 1.3071630295250321e-05,
"loss": 1.1803,
"step": 27500
},
{
"epoch": 35.08,
"eval_loss": 0.24871966242790222,
"eval_runtime": 235.3613,
"eval_samples_per_second": 35.431,
"eval_steps_per_second": 1.109,
"eval_wer": 0.33448678873433096,
"step": 27500
},
{
"epoch": 35.71,
"learning_rate": 1.2943260590500643e-05,
"loss": 1.167,
"step": 28000
},
{
"epoch": 35.71,
"eval_loss": 0.24418053030967712,
"eval_runtime": 232.8217,
"eval_samples_per_second": 35.817,
"eval_steps_per_second": 1.121,
"eval_wer": 0.3318514105002635,
"step": 28000
},
{
"epoch": 36.35,
"learning_rate": 1.2814890885750963e-05,
"loss": 1.1661,
"step": 28500
},
{
"epoch": 36.35,
"eval_loss": 0.24953289330005646,
"eval_runtime": 237.0041,
"eval_samples_per_second": 35.185,
"eval_steps_per_second": 1.101,
"eval_wer": 0.3333868047757637,
"step": 28500
},
{
"epoch": 36.99,
"learning_rate": 1.2686521181001285e-05,
"loss": 1.164,
"step": 29000
},
{
"epoch": 36.99,
"eval_loss": 0.24715770781040192,
"eval_runtime": 233.2505,
"eval_samples_per_second": 35.751,
"eval_steps_per_second": 1.119,
"eval_wer": 0.32923894859866626,
"step": 29000
},
{
"epoch": 37.63,
"learning_rate": 1.2558151476251605e-05,
"loss": 1.1578,
"step": 29500
},
{
"epoch": 37.63,
"eval_loss": 0.24420544505119324,
"eval_runtime": 239.1685,
"eval_samples_per_second": 34.867,
"eval_steps_per_second": 1.091,
"eval_wer": 0.3242431881201732,
"step": 29500
},
{
"epoch": 38.27,
"learning_rate": 1.2430038510911426e-05,
"loss": 1.1584,
"step": 30000
},
{
"epoch": 38.27,
"eval_loss": 0.243063285946846,
"eval_runtime": 235.3653,
"eval_samples_per_second": 35.43,
"eval_steps_per_second": 1.109,
"eval_wer": 0.3314389165158008,
"step": 30000
},
{
"epoch": 38.9,
"learning_rate": 1.2301668806161746e-05,
"loss": 1.1526,
"step": 30500
},
{
"epoch": 38.9,
"eval_loss": 0.2440960705280304,
"eval_runtime": 238.3272,
"eval_samples_per_second": 34.99,
"eval_steps_per_second": 1.095,
"eval_wer": 0.3347159520590325,
"step": 30500
},
{
"epoch": 39.54,
"learning_rate": 1.2173299101412068e-05,
"loss": 1.1542,
"step": 31000
},
{
"epoch": 39.54,
"eval_loss": 0.2436649352312088,
"eval_runtime": 236.1705,
"eval_samples_per_second": 35.309,
"eval_steps_per_second": 1.105,
"eval_wer": 0.3330430597887114,
"step": 31000
},
{
"epoch": 40.18,
"learning_rate": 1.2044929396662388e-05,
"loss": 1.1508,
"step": 31500
},
{
"epoch": 40.18,
"eval_loss": 0.24334189295768738,
"eval_runtime": 234.2613,
"eval_samples_per_second": 35.597,
"eval_steps_per_second": 1.114,
"eval_wer": 0.3294222792584275,
"step": 31500
},
{
"epoch": 40.82,
"learning_rate": 1.191655969191271e-05,
"loss": 1.1406,
"step": 32000
},
{
"epoch": 40.82,
"eval_loss": 0.24340486526489258,
"eval_runtime": 233.0368,
"eval_samples_per_second": 35.784,
"eval_steps_per_second": 1.12,
"eval_wer": 0.3270618970140019,
"step": 32000
},
{
"epoch": 41.45,
"learning_rate": 1.178870346598203e-05,
"loss": 1.1514,
"step": 32500
},
{
"epoch": 41.45,
"eval_loss": 0.24264076352119446,
"eval_runtime": 238.874,
"eval_samples_per_second": 34.91,
"eval_steps_per_second": 1.093,
"eval_wer": 0.32552650273850176,
"step": 32500
},
{
"epoch": 42.09,
"learning_rate": 1.166033376123235e-05,
"loss": 1.1418,
"step": 33000
},
{
"epoch": 42.09,
"eval_loss": 0.2431638389825821,
"eval_runtime": 238.255,
"eval_samples_per_second": 35.0,
"eval_steps_per_second": 1.095,
"eval_wer": 0.3232807021564269,
"step": 33000
},
{
"epoch": 42.73,
"learning_rate": 1.1531964056482671e-05,
"loss": 1.1365,
"step": 33500
},
{
"epoch": 42.73,
"eval_loss": 0.2436181753873825,
"eval_runtime": 234.1309,
"eval_samples_per_second": 35.617,
"eval_steps_per_second": 1.115,
"eval_wer": 0.32399110846300155,
"step": 33500
},
{
"epoch": 43.37,
"learning_rate": 1.1403594351732992e-05,
"loss": 1.1348,
"step": 34000
},
{
"epoch": 43.37,
"eval_loss": 0.2482648491859436,
"eval_runtime": 238.6079,
"eval_samples_per_second": 34.949,
"eval_steps_per_second": 1.094,
"eval_wer": 0.3257327497307331,
"step": 34000
},
{
"epoch": 44.01,
"learning_rate": 1.1275224646983314e-05,
"loss": 1.1301,
"step": 34500
},
{
"epoch": 44.01,
"eval_loss": 0.24202857911586761,
"eval_runtime": 232.8764,
"eval_samples_per_second": 35.809,
"eval_steps_per_second": 1.121,
"eval_wer": 0.3271077296789422,
"step": 34500
},
{
"epoch": 44.64,
"learning_rate": 1.1146854942233634e-05,
"loss": 1.1268,
"step": 35000
},
{
"epoch": 44.64,
"eval_loss": 0.24715635180473328,
"eval_runtime": 233.1436,
"eval_samples_per_second": 35.768,
"eval_steps_per_second": 1.119,
"eval_wer": 0.3224786305199716,
"step": 35000
},
{
"epoch": 45.28,
"learning_rate": 1.1018485237483956e-05,
"loss": 1.1224,
"step": 35500
},
{
"epoch": 45.28,
"eval_loss": 0.2381853461265564,
"eval_runtime": 238.9288,
"eval_samples_per_second": 34.902,
"eval_steps_per_second": 1.092,
"eval_wer": 0.3204849095950684,
"step": 35500
},
{
"epoch": 45.92,
"learning_rate": 1.0890115532734276e-05,
"loss": 1.1224,
"step": 36000
},
{
"epoch": 45.92,
"eval_loss": 0.23876558244228363,
"eval_runtime": 243.8522,
"eval_samples_per_second": 34.197,
"eval_steps_per_second": 1.07,
"eval_wer": 0.3184224396727548,
"step": 36000
},
{
"epoch": 46.56,
"learning_rate": 1.0762002567394097e-05,
"loss": 1.1198,
"step": 36500
},
{
"epoch": 46.56,
"eval_loss": 0.23819147050380707,
"eval_runtime": 238.5089,
"eval_samples_per_second": 34.963,
"eval_steps_per_second": 1.094,
"eval_wer": 0.32016408094048626,
"step": 36500
},
{
"epoch": 47.19,
"learning_rate": 1.0633632862644417e-05,
"loss": 1.1274,
"step": 37000
},
{
"epoch": 47.19,
"eval_loss": 0.24040183424949646,
"eval_runtime": 237.5269,
"eval_samples_per_second": 35.108,
"eval_steps_per_second": 1.099,
"eval_wer": 0.31720787405183676,
"step": 37000
},
{
"epoch": 47.83,
"learning_rate": 1.0505263157894739e-05,
"loss": 1.1147,
"step": 37500
},
{
"epoch": 47.83,
"eval_loss": 0.23944656550884247,
"eval_runtime": 235.5135,
"eval_samples_per_second": 35.408,
"eval_steps_per_second": 1.108,
"eval_wer": 0.3164287187478516,
"step": 37500
},
{
"epoch": 48.47,
"learning_rate": 1.0377150192554558e-05,
"loss": 1.121,
"step": 38000
},
{
"epoch": 48.47,
"eval_loss": 0.2405969202518463,
"eval_runtime": 233.9426,
"eval_samples_per_second": 35.645,
"eval_steps_per_second": 1.116,
"eval_wer": 0.3202099136054266,
"step": 38000
},
{
"epoch": 49.11,
"learning_rate": 1.024878048780488e-05,
"loss": 1.1109,
"step": 38500
},
{
"epoch": 49.11,
"eval_loss": 0.23839789628982544,
"eval_runtime": 240.6944,
"eval_samples_per_second": 34.646,
"eval_steps_per_second": 1.084,
"eval_wer": 0.31537456745422465,
"step": 38500
},
{
"epoch": 49.74,
"learning_rate": 1.0120410783055198e-05,
"loss": 1.1164,
"step": 39000
},
{
"epoch": 49.74,
"eval_loss": 0.2374895066022873,
"eval_runtime": 237.3857,
"eval_samples_per_second": 35.128,
"eval_steps_per_second": 1.099,
"eval_wer": 0.31686412906478445,
"step": 39000
},
{
"epoch": 50.38,
"learning_rate": 9.992041078305522e-06,
"loss": 1.1105,
"step": 39500
},
{
"epoch": 50.38,
"eval_loss": 0.23865096271038055,
"eval_runtime": 232.9433,
"eval_samples_per_second": 35.798,
"eval_steps_per_second": 1.12,
"eval_wer": 0.31729953938171734,
"step": 39500
},
{
"epoch": 51.02,
"learning_rate": 9.86392811296534e-06,
"loss": 1.1054,
"step": 40000
},
{
"epoch": 51.02,
"eval_loss": 0.23617714643478394,
"eval_runtime": 233.5533,
"eval_samples_per_second": 35.705,
"eval_steps_per_second": 1.118,
"eval_wer": 0.31200586658111235,
"step": 40000
},
{
"epoch": 51.66,
"learning_rate": 9.735558408215663e-06,
"loss": 1.0893,
"step": 40500
},
{
"epoch": 51.66,
"eval_loss": 0.23990453779697418,
"eval_runtime": 234.1326,
"eval_samples_per_second": 35.617,
"eval_steps_per_second": 1.115,
"eval_wer": 0.31299126887732887,
"step": 40500
},
{
"epoch": 52.3,
"learning_rate": 9.607188703465983e-06,
"loss": 1.0913,
"step": 41000
},
{
"epoch": 52.3,
"eval_loss": 0.2356649786233902,
"eval_runtime": 238.2462,
"eval_samples_per_second": 35.002,
"eval_steps_per_second": 1.096,
"eval_wer": 0.3087517473703508,
"step": 41000
},
{
"epoch": 52.93,
"learning_rate": 9.478818998716303e-06,
"loss": 1.1017,
"step": 41500
},
{
"epoch": 52.93,
"eval_loss": 0.23450981080532074,
"eval_runtime": 233.807,
"eval_samples_per_second": 35.666,
"eval_steps_per_second": 1.116,
"eval_wer": 0.3083621697183583,
"step": 41500
},
{
"epoch": 53.57,
"learning_rate": 9.350449293966625e-06,
"loss": 1.0937,
"step": 42000
},
{
"epoch": 53.57,
"eval_loss": 0.23296880722045898,
"eval_runtime": 238.7521,
"eval_samples_per_second": 34.927,
"eval_steps_per_second": 1.093,
"eval_wer": 0.31395375484107524,
"step": 42000
},
{
"epoch": 54.21,
"learning_rate": 9.222079589216945e-06,
"loss": 1.0945,
"step": 42500
},
{
"epoch": 54.21,
"eval_loss": 0.23989614844322205,
"eval_runtime": 233.2848,
"eval_samples_per_second": 35.746,
"eval_steps_per_second": 1.119,
"eval_wer": 0.3106767192978436,
"step": 42500
},
{
"epoch": 54.85,
"learning_rate": 9.093966623876766e-06,
"loss": 1.0933,
"step": 43000
},
{
"epoch": 54.85,
"eval_loss": 0.23831374943256378,
"eval_runtime": 234.0959,
"eval_samples_per_second": 35.622,
"eval_steps_per_second": 1.115,
"eval_wer": 0.31338084652932147,
"step": 43000
},
{
"epoch": 55.48,
"learning_rate": 8.965596919127086e-06,
"loss": 1.0912,
"step": 43500
},
{
"epoch": 55.48,
"eval_loss": 0.2371690571308136,
"eval_runtime": 234.0624,
"eval_samples_per_second": 35.627,
"eval_steps_per_second": 1.115,
"eval_wer": 0.30769759607672387,
"step": 43500
},
{
"epoch": 56.12,
"learning_rate": 8.837227214377408e-06,
"loss": 1.0898,
"step": 44000
},
{
"epoch": 56.12,
"eval_loss": 0.23387998342514038,
"eval_runtime": 233.2957,
"eval_samples_per_second": 35.744,
"eval_steps_per_second": 1.119,
"eval_wer": 0.30829342072094784,
"step": 44000
},
{
"epoch": 56.76,
"learning_rate": 8.708857509627728e-06,
"loss": 1.0903,
"step": 44500
},
{
"epoch": 56.76,
"eval_loss": 0.23673543334007263,
"eval_runtime": 236.1402,
"eval_samples_per_second": 35.314,
"eval_steps_per_second": 1.105,
"eval_wer": 0.3065288631207462,
"step": 44500
},
{
"epoch": 57.4,
"learning_rate": 8.58048780487805e-06,
"loss": 1.0947,
"step": 45000
},
{
"epoch": 57.4,
"eval_loss": 0.23515641689300537,
"eval_runtime": 235.6322,
"eval_samples_per_second": 35.39,
"eval_steps_per_second": 1.108,
"eval_wer": 0.31035589064326147,
"step": 45000
},
{
"epoch": 58.04,
"learning_rate": 8.45211810012837e-06,
"loss": 1.0751,
"step": 45500
},
{
"epoch": 58.04,
"eval_loss": 0.23344124853610992,
"eval_runtime": 236.1866,
"eval_samples_per_second": 35.307,
"eval_steps_per_second": 1.105,
"eval_wer": 0.3084309187157687,
"step": 45500
},
{
"epoch": 58.67,
"learning_rate": 8.32374839537869e-06,
"loss": 1.09,
"step": 46000
},
{
"epoch": 58.67,
"eval_loss": 0.2327876091003418,
"eval_runtime": 233.0944,
"eval_samples_per_second": 35.775,
"eval_steps_per_second": 1.12,
"eval_wer": 0.3100350619886793,
"step": 46000
},
{
"epoch": 59.31,
"learning_rate": 8.195378690629013e-06,
"loss": 1.0876,
"step": 46500
},
{
"epoch": 59.31,
"eval_loss": 0.22762973606586456,
"eval_runtime": 232.6706,
"eval_samples_per_second": 35.84,
"eval_steps_per_second": 1.122,
"eval_wer": 0.3049705525127758,
"step": 46500
},
{
"epoch": 59.95,
"learning_rate": 8.067008985879333e-06,
"loss": 1.076,
"step": 47000
},
{
"epoch": 59.95,
"eval_loss": 0.23087561130523682,
"eval_runtime": 233.5675,
"eval_samples_per_second": 35.703,
"eval_steps_per_second": 1.117,
"eval_wer": 0.304695556523134,
"step": 47000
},
{
"epoch": 60.59,
"learning_rate": 7.938896020539153e-06,
"loss": 1.086,
"step": 47500
},
{
"epoch": 60.59,
"eval_loss": 0.229255810379982,
"eval_runtime": 234.5783,
"eval_samples_per_second": 35.549,
"eval_steps_per_second": 1.113,
"eval_wer": 0.304695556523134,
"step": 47500
},
{
"epoch": 61.22,
"learning_rate": 7.810526315789474e-06,
"loss": 1.082,
"step": 48000
},
{
"epoch": 61.22,
"eval_loss": 0.232817605137825,
"eval_runtime": 239.1878,
"eval_samples_per_second": 34.864,
"eval_steps_per_second": 1.091,
"eval_wer": 0.3027476682631712,
"step": 48000
},
{
"epoch": 61.86,
"learning_rate": 7.682413350449294e-06,
"loss": 1.0714,
"step": 48500
},
{
"epoch": 61.86,
"eval_loss": 0.22896084189414978,
"eval_runtime": 234.6636,
"eval_samples_per_second": 35.536,
"eval_steps_per_second": 1.112,
"eval_wer": 0.3019914292916562,
"step": 48500
},
{
"epoch": 62.5,
"learning_rate": 7.554043645699616e-06,
"loss": 1.0746,
"step": 49000
},
{
"epoch": 62.5,
"eval_loss": 0.2313106656074524,
"eval_runtime": 234.936,
"eval_samples_per_second": 35.495,
"eval_steps_per_second": 1.111,
"eval_wer": 0.30586428947911176,
"step": 49000
},
{
"epoch": 63.14,
"learning_rate": 7.425673940949937e-06,
"loss": 1.076,
"step": 49500
},
{
"epoch": 63.14,
"eval_loss": 0.23419487476348877,
"eval_runtime": 235.1421,
"eval_samples_per_second": 35.464,
"eval_steps_per_second": 1.11,
"eval_wer": 0.3050393015101863,
"step": 49500
},
{
"epoch": 63.78,
"learning_rate": 7.297304236200257e-06,
"loss": 1.0648,
"step": 50000
},
{
"epoch": 63.78,
"eval_loss": 0.22855186462402344,
"eval_runtime": 233.6212,
"eval_samples_per_second": 35.695,
"eval_steps_per_second": 1.117,
"eval_wer": 0.30247267227352936,
"step": 50000
},
{
"epoch": 64.41,
"learning_rate": 7.169191270860078e-06,
"loss": 1.0586,
"step": 50500
},
{
"epoch": 64.41,
"eval_loss": 0.23375432193279266,
"eval_runtime": 233.8132,
"eval_samples_per_second": 35.665,
"eval_steps_per_second": 1.116,
"eval_wer": 0.30435181153608176,
"step": 50500
},
{
"epoch": 65.05,
"learning_rate": 7.040821566110398e-06,
"loss": 1.0753,
"step": 51000
},
{
"epoch": 65.05,
"eval_loss": 0.23079071938991547,
"eval_runtime": 238.4184,
"eval_samples_per_second": 34.976,
"eval_steps_per_second": 1.095,
"eval_wer": 0.30446639319843255,
"step": 51000
},
{
"epoch": 65.69,
"learning_rate": 6.912451861360719e-06,
"loss": 1.0664,
"step": 51500
},
{
"epoch": 65.69,
"eval_loss": 0.22733470797538757,
"eval_runtime": 233.5628,
"eval_samples_per_second": 35.703,
"eval_steps_per_second": 1.117,
"eval_wer": 0.3009372779980292,
"step": 51500
},
{
"epoch": 66.33,
"learning_rate": 6.78408215661104e-06,
"loss": 1.0739,
"step": 52000
},
{
"epoch": 66.33,
"eval_loss": 0.2298172414302826,
"eval_runtime": 234.5885,
"eval_samples_per_second": 35.547,
"eval_steps_per_second": 1.113,
"eval_wer": 0.30272475193070103,
"step": 52000
},
{
"epoch": 66.96,
"learning_rate": 6.655712451861361e-06,
"loss": 1.0695,
"step": 52500
},
{
"epoch": 66.96,
"eval_loss": 0.2247009128332138,
"eval_runtime": 234.2623,
"eval_samples_per_second": 35.597,
"eval_steps_per_second": 1.114,
"eval_wer": 0.29958521438229024,
"step": 52500
},
{
"epoch": 67.6,
"learning_rate": 6.5273427471116825e-06,
"loss": 1.06,
"step": 53000
},
{
"epoch": 67.6,
"eval_loss": 0.22762586176395416,
"eval_runtime": 238.2043,
"eval_samples_per_second": 35.008,
"eval_steps_per_second": 1.096,
"eval_wer": 0.30148726997731284,
"step": 53000
},
{
"epoch": 68.24,
"learning_rate": 6.399229781771502e-06,
"loss": 1.0742,
"step": 53500
},
{
"epoch": 68.24,
"eval_loss": 0.22798208892345428,
"eval_runtime": 234.5841,
"eval_samples_per_second": 35.548,
"eval_steps_per_second": 1.113,
"eval_wer": 0.2974081627976259,
"step": 53500
},
{
"epoch": 68.88,
"learning_rate": 6.270860077021823e-06,
"loss": 1.0618,
"step": 54000
},
{
"epoch": 68.88,
"eval_loss": 0.2291172742843628,
"eval_runtime": 233.7512,
"eval_samples_per_second": 35.675,
"eval_steps_per_second": 1.117,
"eval_wer": 0.2988518917432454,
"step": 54000
},
{
"epoch": 69.52,
"learning_rate": 6.142490372272144e-06,
"loss": 1.062,
"step": 54500
},
{
"epoch": 69.52,
"eval_loss": 0.23021718859672546,
"eval_runtime": 236.2703,
"eval_samples_per_second": 35.294,
"eval_steps_per_second": 1.105,
"eval_wer": 0.2970873341430437,
"step": 54500
},
{
"epoch": 70.15,
"learning_rate": 6.0141206675224654e-06,
"loss": 1.0572,
"step": 55000
},
{
"epoch": 70.15,
"eval_loss": 0.22801047563552856,
"eval_runtime": 232.8765,
"eval_samples_per_second": 35.809,
"eval_steps_per_second": 1.121,
"eval_wer": 0.2989893897380663,
"step": 55000
},
{
"epoch": 70.79,
"learning_rate": 5.885750962772786e-06,
"loss": 1.055,
"step": 55500
},
{
"epoch": 70.79,
"eval_loss": 0.22775860130786896,
"eval_runtime": 234.9515,
"eval_samples_per_second": 35.492,
"eval_steps_per_second": 1.111,
"eval_wer": 0.29834773242890206,
"step": 55500
},
{
"epoch": 71.43,
"learning_rate": 5.757381258023107e-06,
"loss": 1.0553,
"step": 56000
},
{
"epoch": 71.43,
"eval_loss": 0.22817462682724,
"eval_runtime": 233.8475,
"eval_samples_per_second": 35.66,
"eval_steps_per_second": 1.116,
"eval_wer": 0.29905813873547676,
"step": 56000
},
{
"epoch": 72.07,
"learning_rate": 5.629011553273428e-06,
"loss": 1.0509,
"step": 56500
},
{
"epoch": 72.07,
"eval_loss": 0.2261197566986084,
"eval_runtime": 235.4495,
"eval_samples_per_second": 35.417,
"eval_steps_per_second": 1.109,
"eval_wer": 0.29591860118706603,
"step": 56500
},
{
"epoch": 72.7,
"learning_rate": 5.500641848523749e-06,
"loss": 1.0469,
"step": 57000
},
{
"epoch": 72.7,
"eval_loss": 0.22159287333488464,
"eval_runtime": 233.6081,
"eval_samples_per_second": 35.697,
"eval_steps_per_second": 1.117,
"eval_wer": 0.29188532667231937,
"step": 57000
},
{
"epoch": 73.34,
"learning_rate": 5.37227214377407e-06,
"loss": 1.0476,
"step": 57500
},
{
"epoch": 73.34,
"eval_loss": 0.22673606872558594,
"eval_runtime": 233.9171,
"eval_samples_per_second": 35.649,
"eval_steps_per_second": 1.116,
"eval_wer": 0.2988518917432454,
"step": 57500
},
{
"epoch": 73.98,
"learning_rate": 5.24415917843389e-06,
"loss": 1.0494,
"step": 58000
},
{
"epoch": 73.98,
"eval_loss": 0.22603431344032288,
"eval_runtime": 234.549,
"eval_samples_per_second": 35.553,
"eval_steps_per_second": 1.113,
"eval_wer": 0.2960102665169466,
"step": 58000
},
{
"epoch": 74.62,
"learning_rate": 5.115789473684211e-06,
"loss": 1.0517,
"step": 58500
},
{
"epoch": 74.62,
"eval_loss": 0.2297457605600357,
"eval_runtime": 234.3971,
"eval_samples_per_second": 35.576,
"eval_steps_per_second": 1.113,
"eval_wer": 0.29894355707312603,
"step": 58500
},
{
"epoch": 75.26,
"learning_rate": 4.987419768934532e-06,
"loss": 1.0458,
"step": 59000
},
{
"epoch": 75.26,
"eval_loss": 0.2246260643005371,
"eval_runtime": 233.0094,
"eval_samples_per_second": 35.788,
"eval_steps_per_second": 1.12,
"eval_wer": 0.2923436533217224,
"step": 59000
},
{
"epoch": 75.89,
"learning_rate": 4.859306803594352e-06,
"loss": 1.0382,
"step": 59500
},
{
"epoch": 75.89,
"eval_loss": 0.22546544671058655,
"eval_runtime": 233.7397,
"eval_samples_per_second": 35.676,
"eval_steps_per_second": 1.117,
"eval_wer": 0.2921603226619612,
"step": 59500
},
{
"epoch": 76.53,
"learning_rate": 4.730937098844673e-06,
"loss": 1.0462,
"step": 60000
},
{
"epoch": 76.53,
"eval_loss": 0.22583839297294617,
"eval_runtime": 233.483,
"eval_samples_per_second": 35.716,
"eval_steps_per_second": 1.118,
"eval_wer": 0.29539152554025255,
"step": 60000
},
{
"epoch": 77.17,
"learning_rate": 4.602567394094994e-06,
"loss": 1.0375,
"step": 60500
},
{
"epoch": 77.17,
"eval_loss": 0.22505535185337067,
"eval_runtime": 239.1762,
"eval_samples_per_second": 34.866,
"eval_steps_per_second": 1.091,
"eval_wer": 0.2928707289685359,
"step": 60500
},
{
"epoch": 77.81,
"learning_rate": 4.474197689345315e-06,
"loss": 1.0332,
"step": 61000
},
{
"epoch": 77.81,
"eval_loss": 0.22773370146751404,
"eval_runtime": 237.0288,
"eval_samples_per_second": 35.181,
"eval_steps_per_second": 1.101,
"eval_wer": 0.2939936292595733,
"step": 61000
},
{
"epoch": 78.44,
"learning_rate": 4.345827984595636e-06,
"loss": 1.0423,
"step": 61500
},
{
"epoch": 78.44,
"eval_loss": 0.22426965832710266,
"eval_runtime": 233.5981,
"eval_samples_per_second": 35.698,
"eval_steps_per_second": 1.117,
"eval_wer": 0.28961660975777437,
"step": 61500
},
{
"epoch": 79.08,
"learning_rate": 4.217458279845957e-06,
"loss": 1.0379,
"step": 62000
},
{
"epoch": 79.08,
"eval_loss": 0.2274049073457718,
"eval_runtime": 234.6064,
"eval_samples_per_second": 35.545,
"eval_steps_per_second": 1.113,
"eval_wer": 0.29280197997112545,
"step": 62000
},
{
"epoch": 79.72,
"learning_rate": 4.089345314505777e-06,
"loss": 1.0398,
"step": 62500
},
{
"epoch": 79.72,
"eval_loss": 0.22373217344284058,
"eval_runtime": 234.7196,
"eval_samples_per_second": 35.527,
"eval_steps_per_second": 1.112,
"eval_wer": 0.29280197997112545,
"step": 62500
},
{
"epoch": 80.36,
"learning_rate": 3.960975609756098e-06,
"loss": 1.0395,
"step": 63000
},
{
"epoch": 80.36,
"eval_loss": 0.22652971744537354,
"eval_runtime": 233.8582,
"eval_samples_per_second": 35.658,
"eval_steps_per_second": 1.116,
"eval_wer": 0.2956206888649541,
"step": 63000
},
{
"epoch": 80.99,
"learning_rate": 3.832605905006419e-06,
"loss": 1.0397,
"step": 63500
},
{
"epoch": 80.99,
"eval_loss": 0.22396665811538696,
"eval_runtime": 235.3841,
"eval_samples_per_second": 35.427,
"eval_steps_per_second": 1.109,
"eval_wer": 0.29202282466714025,
"step": 63500
},
{
"epoch": 81.63,
"learning_rate": 3.7042362002567396e-06,
"loss": 1.0262,
"step": 64000
},
{
"epoch": 81.63,
"eval_loss": 0.22437626123428345,
"eval_runtime": 239.3117,
"eval_samples_per_second": 34.846,
"eval_steps_per_second": 1.091,
"eval_wer": 0.2933748882828792,
"step": 64000
},
{
"epoch": 82.27,
"learning_rate": 3.5758664955070607e-06,
"loss": 1.0335,
"step": 64500
},
{
"epoch": 82.27,
"eval_loss": 0.22651708126068115,
"eval_runtime": 234.0011,
"eval_samples_per_second": 35.637,
"eval_steps_per_second": 1.115,
"eval_wer": 0.2936269679400509,
"step": 64500
},
{
"epoch": 82.91,
"learning_rate": 3.4474967907573813e-06,
"loss": 1.0385,
"step": 65000
},
{
"epoch": 82.91,
"eval_loss": 0.22380606830120087,
"eval_runtime": 236.2381,
"eval_samples_per_second": 35.299,
"eval_steps_per_second": 1.105,
"eval_wer": 0.29284781263606574,
"step": 65000
},
{
"epoch": 83.55,
"learning_rate": 3.3191270860077024e-06,
"loss": 1.0289,
"step": 65500
},
{
"epoch": 83.55,
"eval_loss": 0.221877783536911,
"eval_runtime": 233.6081,
"eval_samples_per_second": 35.697,
"eval_steps_per_second": 1.117,
"eval_wer": 0.2911520040332745,
"step": 65500
},
{
"epoch": 84.18,
"learning_rate": 3.1910141206675225e-06,
"loss": 1.0372,
"step": 66000
},
{
"epoch": 84.18,
"eval_loss": 0.22357773780822754,
"eval_runtime": 234.3938,
"eval_samples_per_second": 35.577,
"eval_steps_per_second": 1.114,
"eval_wer": 0.2898457730824759,
"step": 66000
},
{
"epoch": 84.82,
"learning_rate": 3.0626444159178436e-06,
"loss": 1.0279,
"step": 66500
},
{
"epoch": 84.82,
"eval_loss": 0.22185130417346954,
"eval_runtime": 235.3542,
"eval_samples_per_second": 35.432,
"eval_steps_per_second": 1.109,
"eval_wer": 0.2902124344019983,
"step": 66500
},
{
"epoch": 85.46,
"learning_rate": 2.9342747111681647e-06,
"loss": 1.0325,
"step": 67000
},
{
"epoch": 85.46,
"eval_loss": 0.22399279475212097,
"eval_runtime": 232.7577,
"eval_samples_per_second": 35.827,
"eval_steps_per_second": 1.121,
"eval_wer": 0.2907853427137521,
"step": 67000
},
{
"epoch": 86.1,
"learning_rate": 2.8059050064184858e-06,
"loss": 1.0202,
"step": 67500
},
{
"epoch": 86.1,
"eval_loss": 0.22061078250408173,
"eval_runtime": 236.8657,
"eval_samples_per_second": 35.206,
"eval_steps_per_second": 1.102,
"eval_wer": 0.28863120746155785,
"step": 67500
},
{
"epoch": 86.73,
"learning_rate": 2.6775353016688064e-06,
"loss": 1.0166,
"step": 68000
},
{
"epoch": 86.73,
"eval_loss": 0.2218720018863678,
"eval_runtime": 233.7184,
"eval_samples_per_second": 35.68,
"eval_steps_per_second": 1.117,
"eval_wer": 0.2885624584641474,
"step": 68000
},
{
"epoch": 87.37,
"learning_rate": 2.5491655969191275e-06,
"loss": 1.0259,
"step": 68500
},
{
"epoch": 87.37,
"eval_loss": 0.2234880030155182,
"eval_runtime": 235.0768,
"eval_samples_per_second": 35.474,
"eval_steps_per_second": 1.11,
"eval_wer": 0.28970827508765495,
"step": 68500
},
{
"epoch": 88.01,
"learning_rate": 2.420795892169448e-06,
"loss": 1.0337,
"step": 69000
},
{
"epoch": 88.01,
"eval_loss": 0.220963254570961,
"eval_runtime": 233.4278,
"eval_samples_per_second": 35.724,
"eval_steps_per_second": 1.118,
"eval_wer": 0.2873249765107592,
"step": 69000
},
{
"epoch": 88.65,
"learning_rate": 2.2926829268292687e-06,
"loss": 1.0264,
"step": 69500
},
{
"epoch": 88.65,
"eval_loss": 0.22157110273838043,
"eval_runtime": 233.5356,
"eval_samples_per_second": 35.708,
"eval_steps_per_second": 1.118,
"eval_wer": 0.2882187134770951,
"step": 69500
},
{
"epoch": 89.29,
"learning_rate": 2.1643132220795894e-06,
"loss": 1.0231,
"step": 70000
},
{
"epoch": 89.29,
"eval_loss": 0.2222645878791809,
"eval_runtime": 238.1657,
"eval_samples_per_second": 35.013,
"eval_steps_per_second": 1.096,
"eval_wer": 0.2899145220798863,
"step": 70000
},
{
"epoch": 89.92,
"learning_rate": 2.03594351732991e-06,
"loss": 1.0281,
"step": 70500
},
{
"epoch": 89.92,
"eval_loss": 0.22140374779701233,
"eval_runtime": 233.9034,
"eval_samples_per_second": 35.651,
"eval_steps_per_second": 1.116,
"eval_wer": 0.28723331118087864,
"step": 70500
},
{
"epoch": 90.56,
"learning_rate": 1.907573812580231e-06,
"loss": 1.0135,
"step": 71000
},
{
"epoch": 90.56,
"eval_loss": 0.22179700434207916,
"eval_runtime": 235.221,
"eval_samples_per_second": 35.452,
"eval_steps_per_second": 1.11,
"eval_wer": 0.28684373352888604,
"step": 71000
},
{
"epoch": 91.2,
"learning_rate": 1.7792041078305522e-06,
"loss": 1.0291,
"step": 71500
},
{
"epoch": 91.2,
"eval_loss": 0.22093360126018524,
"eval_runtime": 236.8894,
"eval_samples_per_second": 35.202,
"eval_steps_per_second": 1.102,
"eval_wer": 0.2862937415496024,
"step": 71500
},
{
"epoch": 91.84,
"learning_rate": 1.6508344030808731e-06,
"loss": 1.0321,
"step": 72000
},
{
"epoch": 91.84,
"eval_loss": 0.219914972782135,
"eval_runtime": 233.825,
"eval_samples_per_second": 35.663,
"eval_steps_per_second": 1.116,
"eval_wer": 0.2875770561679309,
"step": 72000
},
{
"epoch": 92.47,
"learning_rate": 1.5227214377406932e-06,
"loss": 1.028,
"step": 72500
},
{
"epoch": 92.47,
"eval_loss": 0.22144030034542084,
"eval_runtime": 234.5883,
"eval_samples_per_second": 35.547,
"eval_steps_per_second": 1.113,
"eval_wer": 0.28576666590278893,
"step": 72500
},
{
"epoch": 93.11,
"learning_rate": 1.3943517329910141e-06,
"loss": 1.0213,
"step": 73000
},
{
"epoch": 93.11,
"eval_loss": 0.2219470739364624,
"eval_runtime": 234.1897,
"eval_samples_per_second": 35.608,
"eval_steps_per_second": 1.114,
"eval_wer": 0.2874624745055801,
"step": 73000
},
{
"epoch": 93.75,
"learning_rate": 1.2659820282413352e-06,
"loss": 1.0261,
"step": 73500
},
{
"epoch": 93.75,
"eval_loss": 0.22315841913223267,
"eval_runtime": 234.7799,
"eval_samples_per_second": 35.518,
"eval_steps_per_second": 1.112,
"eval_wer": 0.28688956619382633,
"step": 73500
},
{
"epoch": 94.39,
"learning_rate": 1.137612323491656e-06,
"loss": 1.0197,
"step": 74000
},
{
"epoch": 94.39,
"eval_loss": 0.22271272540092468,
"eval_runtime": 233.7941,
"eval_samples_per_second": 35.668,
"eval_steps_per_second": 1.116,
"eval_wer": 0.2866145702041845,
"step": 74000
},
{
"epoch": 95.03,
"learning_rate": 1.009242618741977e-06,
"loss": 1.0298,
"step": 74500
},
{
"epoch": 95.03,
"eval_loss": 0.22284960746765137,
"eval_runtime": 236.8784,
"eval_samples_per_second": 35.204,
"eval_steps_per_second": 1.102,
"eval_wer": 0.28675206819900545,
"step": 74500
},
{
"epoch": 95.66,
"learning_rate": 8.808729139922979e-07,
"loss": 1.0192,
"step": 75000
},
{
"epoch": 95.66,
"eval_loss": 0.22300392389297485,
"eval_runtime": 233.5836,
"eval_samples_per_second": 35.7,
"eval_steps_per_second": 1.117,
"eval_wer": 0.28652290487430393,
"step": 75000
},
{
"epoch": 96.3,
"learning_rate": 7.525032092426188e-07,
"loss": 1.0156,
"step": 75500
},
{
"epoch": 96.3,
"eval_loss": 0.2220136970281601,
"eval_runtime": 231.7898,
"eval_samples_per_second": 35.977,
"eval_steps_per_second": 1.126,
"eval_wer": 0.2869353988587666,
"step": 75500
},
{
"epoch": 96.94,
"learning_rate": 6.241335044929398e-07,
"loss": 1.0075,
"step": 76000
},
{
"epoch": 96.94,
"eval_loss": 0.2222670316696167,
"eval_runtime": 232.6992,
"eval_samples_per_second": 35.836,
"eval_steps_per_second": 1.122,
"eval_wer": 0.2866145702041845,
"step": 76000
},
{
"epoch": 97.58,
"learning_rate": 4.9602053915276e-07,
"loss": 1.0201,
"step": 76500
},
{
"epoch": 97.58,
"eval_loss": 0.2218976616859436,
"eval_runtime": 233.5288,
"eval_samples_per_second": 35.709,
"eval_steps_per_second": 1.118,
"eval_wer": 0.28659165387171437,
"step": 76500
},
{
"epoch": 98.21,
"learning_rate": 3.676508344030809e-07,
"loss": 1.0159,
"step": 77000
},
{
"epoch": 98.21,
"eval_loss": 0.22194130718708038,
"eval_runtime": 233.434,
"eval_samples_per_second": 35.723,
"eval_steps_per_second": 1.118,
"eval_wer": 0.28764580516534133,
"step": 77000
},
{
"epoch": 98.85,
"learning_rate": 2.395378690629012e-07,
"loss": 1.0087,
"step": 77500
},
{
"epoch": 98.85,
"eval_loss": 0.2219318002462387,
"eval_runtime": 235.7197,
"eval_samples_per_second": 35.377,
"eval_steps_per_second": 1.107,
"eval_wer": 0.2873249765107592,
"step": 77500
},
{
"epoch": 99.49,
"learning_rate": 1.111681643132221e-07,
"loss": 1.0159,
"step": 78000
},
{
"epoch": 99.49,
"eval_loss": 0.2223382443189621,
"eval_runtime": 234.8374,
"eval_samples_per_second": 35.51,
"eval_steps_per_second": 1.111,
"eval_wer": 0.2867291518665353,
"step": 78000
},
{
"epoch": 100.0,
"step": 78400,
"total_flos": 1.850628803077507e+20,
"train_loss": 1.2247810986577248,
"train_runtime": 113701.5353,
"train_samples_per_second": 22.038,
"train_steps_per_second": 0.69
}
],
"max_steps": 78400,
"num_train_epochs": 100,
"total_flos": 1.850628803077507e+20,
"trial_name": null,
"trial_params": null
}