wav2vec2-large-xlsr-persian-v3 / trainer_state.json
m3hrdadfi's picture
Hello Wav2Vec
02535ad
raw
history blame
78.3 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 30.0,
"global_step": 71280,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.13,
"learning_rate": 0.0002472,
"loss": 7.355,
"step": 300
},
{
"epoch": 0.13,
"eval_loss": 3.007211446762085,
"eval_runtime": 459.0118,
"eval_samples_per_second": 11.357,
"eval_wer": 1.0,
"step": 300
},
{
"epoch": 0.25,
"learning_rate": 0.0004114179146651596,
"loss": 1.4963,
"step": 600
},
{
"epoch": 0.25,
"eval_loss": 0.6261181235313416,
"eval_runtime": 447.0386,
"eval_samples_per_second": 11.661,
"eval_wer": 0.5529420917507326,
"step": 600
},
{
"epoch": 0.38,
"learning_rate": 0.0004096716586606386,
"loss": 0.5139,
"step": 900
},
{
"epoch": 0.38,
"eval_loss": 0.47113755345344543,
"eval_runtime": 456.1265,
"eval_samples_per_second": 11.429,
"eval_wer": 0.445372267316719,
"step": 900
},
{
"epoch": 0.51,
"learning_rate": 0.00040792540265611755,
"loss": 0.4189,
"step": 1200
},
{
"epoch": 0.51,
"eval_loss": 0.4321776032447815,
"eval_runtime": 431.3516,
"eval_samples_per_second": 12.085,
"eval_wer": 0.4174684266486865,
"step": 1200
},
{
"epoch": 0.63,
"learning_rate": 0.00040617914665159647,
"loss": 0.3815,
"step": 1500
},
{
"epoch": 0.63,
"eval_loss": 0.3827853798866272,
"eval_runtime": 459.5456,
"eval_samples_per_second": 11.344,
"eval_wer": 0.3828998210627318,
"step": 1500
},
{
"epoch": 0.76,
"learning_rate": 0.00040443289064707544,
"loss": 0.3504,
"step": 1800
},
{
"epoch": 0.76,
"eval_loss": 0.3574122190475464,
"eval_runtime": 457.446,
"eval_samples_per_second": 11.396,
"eval_wer": 0.3792951427608205,
"step": 1800
},
{
"epoch": 0.88,
"learning_rate": 0.00040268663464255436,
"loss": 0.3385,
"step": 2100
},
{
"epoch": 0.88,
"eval_loss": 0.3647058308124542,
"eval_runtime": 432.4987,
"eval_samples_per_second": 12.053,
"eval_wer": 0.3572002800757242,
"step": 2100
},
{
"epoch": 1.01,
"learning_rate": 0.00040094037863803333,
"loss": 0.3205,
"step": 2400
},
{
"epoch": 1.01,
"eval_loss": 0.3391036093235016,
"eval_runtime": 460.5432,
"eval_samples_per_second": 11.319,
"eval_wer": 0.34612691579575217,
"step": 2400
},
{
"epoch": 1.14,
"learning_rate": 0.0003991941226335123,
"loss": 0.2697,
"step": 2700
},
{
"epoch": 1.14,
"eval_loss": 0.34452542662620544,
"eval_runtime": 469.56,
"eval_samples_per_second": 11.102,
"eval_wer": 0.3170042270688001,
"step": 2700
},
{
"epoch": 1.26,
"learning_rate": 0.0003974478666289912,
"loss": 0.2562,
"step": 3000
},
{
"epoch": 1.26,
"eval_loss": 0.32179856300354004,
"eval_runtime": 433.2725,
"eval_samples_per_second": 12.032,
"eval_wer": 0.31285495708098854,
"step": 3000
},
{
"epoch": 1.39,
"learning_rate": 0.0003957016106244702,
"loss": 0.2514,
"step": 3300
},
{
"epoch": 1.39,
"eval_loss": 0.3245706558227539,
"eval_runtime": 453.9529,
"eval_samples_per_second": 11.484,
"eval_wer": 0.30668291797411895,
"step": 3300
},
{
"epoch": 1.52,
"learning_rate": 0.0003939553546199491,
"loss": 0.2545,
"step": 3600
},
{
"epoch": 1.52,
"eval_loss": 0.32298558950424194,
"eval_runtime": 460.4237,
"eval_samples_per_second": 11.322,
"eval_wer": 0.30904281527968674,
"step": 3600
},
{
"epoch": 1.64,
"learning_rate": 0.0003922090986154281,
"loss": 0.2457,
"step": 3900
},
{
"epoch": 1.64,
"eval_loss": 0.31425100564956665,
"eval_runtime": 459.4801,
"eval_samples_per_second": 11.345,
"eval_wer": 0.2933274552008506,
"step": 3900
},
{
"epoch": 1.77,
"learning_rate": 0.000390462842610907,
"loss": 0.2458,
"step": 4200
},
{
"epoch": 1.77,
"eval_loss": 0.312289834022522,
"eval_runtime": 428.9645,
"eval_samples_per_second": 12.153,
"eval_wer": 0.28100930992453516,
"step": 4200
},
{
"epoch": 1.89,
"learning_rate": 0.00038871658660638597,
"loss": 0.2353,
"step": 4500
},
{
"epoch": 1.89,
"eval_loss": 0.29052308201789856,
"eval_runtime": 458.3214,
"eval_samples_per_second": 11.374,
"eval_wer": 0.27597832006431366,
"step": 4500
},
{
"epoch": 2.02,
"learning_rate": 0.00038697033060186494,
"loss": 0.2316,
"step": 4800
},
{
"epoch": 2.02,
"eval_loss": 0.3132184147834778,
"eval_runtime": 450.6205,
"eval_samples_per_second": 11.568,
"eval_wer": 0.26967661627032496,
"step": 4800
},
{
"epoch": 2.15,
"learning_rate": 0.00038522407459734386,
"loss": 0.1995,
"step": 5100
},
{
"epoch": 2.15,
"eval_loss": 0.3037310242652893,
"eval_runtime": 424.704,
"eval_samples_per_second": 12.274,
"eval_wer": 0.2598480329866964,
"step": 5100
},
{
"epoch": 2.27,
"learning_rate": 0.00038347781859282283,
"loss": 0.1934,
"step": 5400
},
{
"epoch": 2.27,
"eval_loss": 0.2947574853897095,
"eval_runtime": 450.7971,
"eval_samples_per_second": 11.564,
"eval_wer": 0.25870698374004825,
"step": 5400
},
{
"epoch": 2.4,
"learning_rate": 0.00038173156258830174,
"loss": 0.1957,
"step": 5700
},
{
"epoch": 2.4,
"eval_loss": 0.30417683720588684,
"eval_runtime": 454.0214,
"eval_samples_per_second": 11.482,
"eval_wer": 0.2659682062187184,
"step": 5700
},
{
"epoch": 2.53,
"learning_rate": 0.0003799853065837807,
"loss": 0.198,
"step": 6000
},
{
"epoch": 2.53,
"eval_loss": 0.2877831757068634,
"eval_runtime": 423.4049,
"eval_samples_per_second": 12.312,
"eval_wer": 0.25683981224553304,
"step": 6000
},
{
"epoch": 2.65,
"learning_rate": 0.00037823905057925963,
"loss": 0.2005,
"step": 6300
},
{
"epoch": 2.65,
"eval_loss": 0.2968423664569855,
"eval_runtime": 449.924,
"eval_samples_per_second": 11.586,
"eval_wer": 0.25951090480018674,
"step": 6300
},
{
"epoch": 2.78,
"learning_rate": 0.00037649279457473866,
"loss": 0.1994,
"step": 6600
},
{
"epoch": 2.78,
"eval_loss": 0.2824803590774536,
"eval_runtime": 452.2888,
"eval_samples_per_second": 11.526,
"eval_wer": 0.25056404138896815,
"step": 6600
},
{
"epoch": 2.9,
"learning_rate": 0.0003747465385702176,
"loss": 0.1869,
"step": 6900
},
{
"epoch": 2.9,
"eval_loss": 0.2733488082885742,
"eval_runtime": 446.0548,
"eval_samples_per_second": 11.687,
"eval_wer": 0.24574051502813724,
"step": 6900
},
{
"epoch": 3.03,
"learning_rate": 0.0003730002825656965,
"loss": 0.1916,
"step": 7200
},
{
"epoch": 3.03,
"eval_loss": 0.2809160351753235,
"eval_runtime": 423.4141,
"eval_samples_per_second": 12.312,
"eval_wer": 0.24529965509193227,
"step": 7200
},
{
"epoch": 3.16,
"learning_rate": 0.00037125402656117547,
"loss": 0.1675,
"step": 7500
},
{
"epoch": 3.16,
"eval_loss": 0.28469306230545044,
"eval_runtime": 443.2959,
"eval_samples_per_second": 11.76,
"eval_wer": 0.25012318145276313,
"step": 7500
},
{
"epoch": 3.28,
"learning_rate": 0.0003695077705566544,
"loss": 0.162,
"step": 7800
},
{
"epoch": 3.28,
"eval_loss": 0.2753625810146332,
"eval_runtime": 450.2372,
"eval_samples_per_second": 11.578,
"eval_wer": 0.23324083918985503,
"step": 7800
},
{
"epoch": 3.41,
"learning_rate": 0.00036776151455213336,
"loss": 0.1595,
"step": 8100
},
{
"epoch": 3.41,
"eval_loss": 0.2853989899158478,
"eval_runtime": 425.7608,
"eval_samples_per_second": 12.244,
"eval_wer": 0.23324083918985503,
"step": 8100
},
{
"epoch": 3.54,
"learning_rate": 0.0003660152585476123,
"loss": 0.1658,
"step": 8400
},
{
"epoch": 3.54,
"eval_loss": 0.28348785638809204,
"eval_runtime": 450.477,
"eval_samples_per_second": 11.572,
"eval_wer": 0.23103653950883016,
"step": 8400
},
{
"epoch": 3.66,
"learning_rate": 0.00036426900254309124,
"loss": 0.1594,
"step": 8700
},
{
"epoch": 3.66,
"eval_loss": 0.27261418104171753,
"eval_runtime": 459.5467,
"eval_samples_per_second": 11.344,
"eval_wer": 0.23287777806592153,
"step": 8700
},
{
"epoch": 3.79,
"learning_rate": 0.0003625227465385702,
"loss": 0.1646,
"step": 9000
},
{
"epoch": 3.79,
"eval_loss": 0.2741488218307495,
"eval_runtime": 421.349,
"eval_samples_per_second": 12.372,
"eval_wer": 0.2319182593812401,
"step": 9000
},
{
"epoch": 3.91,
"learning_rate": 0.00036077649053404913,
"loss": 0.1691,
"step": 9300
},
{
"epoch": 3.91,
"eval_loss": 0.3101365268230438,
"eval_runtime": 444.0677,
"eval_samples_per_second": 11.739,
"eval_wer": 0.2433546847851456,
"step": 9300
},
{
"epoch": 4.04,
"learning_rate": 0.0003590302345295281,
"loss": 0.1729,
"step": 9600
},
{
"epoch": 4.04,
"eval_loss": 0.2734999358654022,
"eval_runtime": 449.6252,
"eval_samples_per_second": 11.594,
"eval_wer": 0.21807007079691917,
"step": 9600
},
{
"epoch": 4.17,
"learning_rate": 0.000357283978525007,
"loss": 0.1416,
"step": 9900
},
{
"epoch": 4.17,
"eval_loss": 0.27035850286483765,
"eval_runtime": 421.6286,
"eval_samples_per_second": 12.364,
"eval_wer": 0.21532117942999404,
"step": 9900
},
{
"epoch": 4.29,
"learning_rate": 0.000355537722520486,
"loss": 0.1334,
"step": 10200
},
{
"epoch": 4.29,
"eval_loss": 0.2867918908596039,
"eval_runtime": 447.7813,
"eval_samples_per_second": 11.642,
"eval_wer": 0.2140763984336506,
"step": 10200
},
{
"epoch": 4.42,
"learning_rate": 0.00035379146651596497,
"loss": 0.1382,
"step": 10500
},
{
"epoch": 4.42,
"eval_loss": 0.29414185881614685,
"eval_runtime": 447.6999,
"eval_samples_per_second": 11.644,
"eval_wer": 0.22981769144991054,
"step": 10500
},
{
"epoch": 4.55,
"learning_rate": 0.0003520452105114439,
"loss": 0.1563,
"step": 10800
},
{
"epoch": 4.55,
"eval_loss": 0.2874026298522949,
"eval_runtime": 448.1528,
"eval_samples_per_second": 11.632,
"eval_wer": 0.2264982754596613,
"step": 10800
},
{
"epoch": 4.67,
"learning_rate": 0.00035029895450692285,
"loss": 0.1522,
"step": 11100
},
{
"epoch": 4.67,
"eval_loss": 0.27748048305511475,
"eval_runtime": 422.3545,
"eval_samples_per_second": 12.343,
"eval_wer": 0.21596950286558958,
"step": 11100
},
{
"epoch": 4.8,
"learning_rate": 0.00034855269850240177,
"loss": 0.152,
"step": 11400
},
{
"epoch": 4.8,
"eval_loss": 0.2689838707447052,
"eval_runtime": 450.681,
"eval_samples_per_second": 11.567,
"eval_wer": 0.221156090350354,
"step": 11400
},
{
"epoch": 4.92,
"learning_rate": 0.00034680644249788074,
"loss": 0.1537,
"step": 11700
},
{
"epoch": 4.92,
"eval_loss": 0.26583364605903625,
"eval_runtime": 439.0624,
"eval_samples_per_second": 11.873,
"eval_wer": 0.2123388916262545,
"step": 11700
},
{
"epoch": 5.05,
"learning_rate": 0.00034506018649335966,
"loss": 0.1394,
"step": 12000
},
{
"epoch": 5.05,
"eval_loss": 0.28286588191986084,
"eval_runtime": 420.652,
"eval_samples_per_second": 12.393,
"eval_wer": 0.20710043826664246,
"step": 12000
},
{
"epoch": 5.18,
"learning_rate": 0.0003433139304888387,
"loss": 0.1318,
"step": 12300
},
{
"epoch": 5.18,
"eval_loss": 0.2701680362224579,
"eval_runtime": 452.9348,
"eval_samples_per_second": 11.509,
"eval_wer": 0.20406628458805529,
"step": 12300
},
{
"epoch": 5.3,
"learning_rate": 0.0003415676744843176,
"loss": 0.1211,
"step": 12600
},
{
"epoch": 5.3,
"eval_loss": 0.27643057703971863,
"eval_runtime": 455.5802,
"eval_samples_per_second": 11.443,
"eval_wer": 0.19257799330930214,
"step": 12600
},
{
"epoch": 5.43,
"learning_rate": 0.0003398214184797965,
"loss": 0.1278,
"step": 12900
},
{
"epoch": 5.43,
"eval_loss": 0.2747470438480377,
"eval_runtime": 420.85,
"eval_samples_per_second": 12.387,
"eval_wer": 0.20979746375871994,
"step": 12900
},
{
"epoch": 5.56,
"learning_rate": 0.0003380751624752755,
"loss": 0.1334,
"step": 13200
},
{
"epoch": 5.56,
"eval_loss": 0.2526901960372925,
"eval_runtime": 447.6704,
"eval_samples_per_second": 11.645,
"eval_wer": 0.19836103835481445,
"step": 13200
},
{
"epoch": 5.68,
"learning_rate": 0.0003363289064707544,
"loss": 0.1262,
"step": 13500
},
{
"epoch": 5.68,
"eval_loss": 0.24875697493553162,
"eval_runtime": 453.8545,
"eval_samples_per_second": 11.486,
"eval_wer": 0.2052591997095511,
"step": 13500
},
{
"epoch": 5.81,
"learning_rate": 0.0003345826504662334,
"loss": 0.1366,
"step": 13800
},
{
"epoch": 5.81,
"eval_loss": 0.2449602484703064,
"eval_runtime": 450.7796,
"eval_samples_per_second": 11.564,
"eval_wer": 0.19501568942714143,
"step": 13800
},
{
"epoch": 5.93,
"learning_rate": 0.00033283639446171235,
"loss": 0.1348,
"step": 14100
},
{
"epoch": 5.93,
"eval_loss": 0.28218716382980347,
"eval_runtime": 422.2899,
"eval_samples_per_second": 12.345,
"eval_wer": 0.22273799953320714,
"step": 14100
},
{
"epoch": 6.06,
"learning_rate": 0.00033109013845719127,
"loss": 0.1325,
"step": 14400
},
{
"epoch": 6.06,
"eval_loss": 0.2777564525604248,
"eval_runtime": 447.1411,
"eval_samples_per_second": 11.659,
"eval_wer": 0.20780062757708564,
"step": 14400
},
{
"epoch": 6.19,
"learning_rate": 0.00032934388245267024,
"loss": 0.1261,
"step": 14700
},
{
"epoch": 6.19,
"eval_loss": 0.28267186880111694,
"eval_runtime": 452.8704,
"eval_samples_per_second": 11.511,
"eval_wer": 0.2153730453048417,
"step": 14700
},
{
"epoch": 6.31,
"learning_rate": 0.00032759762644814916,
"loss": 0.1268,
"step": 15000
},
{
"epoch": 6.31,
"eval_loss": 0.2655777335166931,
"eval_runtime": 421.8213,
"eval_samples_per_second": 12.358,
"eval_wer": 0.20074686859780608,
"step": 15000
},
{
"epoch": 6.44,
"learning_rate": 0.00032585137044362813,
"loss": 0.1146,
"step": 15300
},
{
"epoch": 6.44,
"eval_loss": 0.26334822177886963,
"eval_runtime": 453.2863,
"eval_samples_per_second": 11.5,
"eval_wer": 0.19060709006509166,
"step": 15300
},
{
"epoch": 6.57,
"learning_rate": 0.00032410511443910705,
"loss": 0.1196,
"step": 15600
},
{
"epoch": 6.57,
"eval_loss": 0.2629387676715851,
"eval_runtime": 453.8742,
"eval_samples_per_second": 11.486,
"eval_wer": 0.1959752081118228,
"step": 15600
},
{
"epoch": 6.69,
"learning_rate": 0.0003223588584345861,
"loss": 0.1162,
"step": 15900
},
{
"epoch": 6.69,
"eval_loss": 0.2686946392059326,
"eval_runtime": 423.221,
"eval_samples_per_second": 12.317,
"eval_wer": 0.19566401286273696,
"step": 15900
},
{
"epoch": 6.82,
"learning_rate": 0.000320612602430065,
"loss": 0.1178,
"step": 16200
},
{
"epoch": 6.82,
"eval_loss": 0.26208436489105225,
"eval_runtime": 455.8818,
"eval_samples_per_second": 11.435,
"eval_wer": 0.19716812323331864,
"step": 16200
},
{
"epoch": 6.94,
"learning_rate": 0.0003188663464255439,
"loss": 0.1172,
"step": 16500
},
{
"epoch": 6.94,
"eval_loss": 0.2683062255382538,
"eval_runtime": 453.6948,
"eval_samples_per_second": 11.49,
"eval_wer": 0.18845465625891444,
"step": 16500
},
{
"epoch": 7.07,
"learning_rate": 0.0003171200904210229,
"loss": 0.1137,
"step": 16800
},
{
"epoch": 7.07,
"eval_loss": 0.246523916721344,
"eval_runtime": 446.7642,
"eval_samples_per_second": 11.668,
"eval_wer": 0.18466844739503643,
"step": 16800
},
{
"epoch": 7.2,
"learning_rate": 0.0003153738344165018,
"loss": 0.1015,
"step": 17100
},
{
"epoch": 7.2,
"eval_loss": 0.261563241481781,
"eval_runtime": 423.2985,
"eval_samples_per_second": 12.315,
"eval_wer": 0.18783226576074272,
"step": 17100
},
{
"epoch": 7.32,
"learning_rate": 0.00031362757841198077,
"loss": 0.1048,
"step": 17400
},
{
"epoch": 7.32,
"eval_loss": 0.26347172260284424,
"eval_runtime": 455.9994,
"eval_samples_per_second": 11.432,
"eval_wer": 0.18096003734342989,
"step": 17400
},
{
"epoch": 7.45,
"learning_rate": 0.0003118813224074597,
"loss": 0.1085,
"step": 17700
},
{
"epoch": 7.45,
"eval_loss": 0.25925150513648987,
"eval_runtime": 451.158,
"eval_samples_per_second": 11.555,
"eval_wer": 0.19657166567257073,
"step": 17700
},
{
"epoch": 7.58,
"learning_rate": 0.0003101350664029387,
"loss": 0.1165,
"step": 18000
},
{
"epoch": 7.58,
"eval_loss": 0.25639158487319946,
"eval_runtime": 422.5184,
"eval_samples_per_second": 12.338,
"eval_wer": 0.1952490858639558,
"step": 18000
},
{
"epoch": 7.7,
"learning_rate": 0.00030838881039841763,
"loss": 0.1132,
"step": 18300
},
{
"epoch": 7.7,
"eval_loss": 0.2572135627269745,
"eval_runtime": 450.0834,
"eval_samples_per_second": 11.582,
"eval_wer": 0.19268172505899744,
"step": 18300
},
{
"epoch": 7.83,
"learning_rate": 0.00030664255439389655,
"loss": 0.1105,
"step": 18600
},
{
"epoch": 7.83,
"eval_loss": 0.26829710602760315,
"eval_runtime": 447.2579,
"eval_samples_per_second": 11.655,
"eval_wer": 0.18083037265631077,
"step": 18600
},
{
"epoch": 7.95,
"learning_rate": 0.0003048962983893755,
"loss": 0.1114,
"step": 18900
},
{
"epoch": 7.95,
"eval_loss": 0.2628280818462372,
"eval_runtime": 420.5057,
"eval_samples_per_second": 12.397,
"eval_wer": 0.18710614351287572,
"step": 18900
},
{
"epoch": 8.08,
"learning_rate": 0.00030315004238485444,
"loss": 0.1008,
"step": 19200
},
{
"epoch": 8.08,
"eval_loss": 0.2631722390651703,
"eval_runtime": 454.7744,
"eval_samples_per_second": 11.463,
"eval_wer": 0.1872358081999948,
"step": 19200
},
{
"epoch": 8.21,
"learning_rate": 0.0003014037863803334,
"loss": 0.0916,
"step": 19500
},
{
"epoch": 8.21,
"eval_loss": 0.2808758616447449,
"eval_runtime": 447.7804,
"eval_samples_per_second": 11.642,
"eval_wer": 0.1823863489017401,
"step": 19500
},
{
"epoch": 8.33,
"learning_rate": 0.0002996575303758124,
"loss": 0.0991,
"step": 19800
},
{
"epoch": 8.33,
"eval_loss": 0.26910626888275146,
"eval_runtime": 418.5271,
"eval_samples_per_second": 12.456,
"eval_wer": 0.18023391509556289,
"step": 19800
},
{
"epoch": 8.46,
"learning_rate": 0.0002979112743712913,
"loss": 0.0969,
"step": 20100
},
{
"epoch": 8.46,
"eval_loss": 0.24965764582157135,
"eval_runtime": 443.9577,
"eval_samples_per_second": 11.742,
"eval_wer": 0.1918259381240113,
"step": 20100
},
{
"epoch": 8.59,
"learning_rate": 0.00029616501836677027,
"loss": 0.1008,
"step": 20400
},
{
"epoch": 8.59,
"eval_loss": 0.2633902132511139,
"eval_runtime": 446.6554,
"eval_samples_per_second": 11.671,
"eval_wer": 0.17963745753481497,
"step": 20400
},
{
"epoch": 8.71,
"learning_rate": 0.0002944187623622492,
"loss": 0.1003,
"step": 20700
},
{
"epoch": 8.71,
"eval_loss": 0.26895150542259216,
"eval_runtime": 447.0422,
"eval_samples_per_second": 11.661,
"eval_wer": 0.17255776561811156,
"step": 20700
},
{
"epoch": 8.84,
"learning_rate": 0.00029267250635772816,
"loss": 0.1059,
"step": 21000
},
{
"epoch": 8.84,
"eval_loss": 0.25897544622421265,
"eval_runtime": 418.2635,
"eval_samples_per_second": 12.463,
"eval_wer": 0.18676901532636603,
"step": 21000
},
{
"epoch": 8.96,
"learning_rate": 0.0002909262503532071,
"loss": 0.1011,
"step": 21300
},
{
"epoch": 8.96,
"eval_loss": 0.27191728353500366,
"eval_runtime": 439.9733,
"eval_samples_per_second": 11.848,
"eval_wer": 0.17569565104639404,
"step": 21300
},
{
"epoch": 9.09,
"learning_rate": 0.0002891799943486861,
"loss": 0.086,
"step": 21600
},
{
"epoch": 9.09,
"eval_loss": 0.24117550253868103,
"eval_runtime": 451.9133,
"eval_samples_per_second": 11.535,
"eval_wer": 0.16464821970384586,
"step": 21600
},
{
"epoch": 9.22,
"learning_rate": 0.000287433738344165,
"loss": 0.0896,
"step": 21900
},
{
"epoch": 9.22,
"eval_loss": 0.2518727481365204,
"eval_runtime": 420.864,
"eval_samples_per_second": 12.386,
"eval_wer": 0.1694198801898291,
"step": 21900
},
{
"epoch": 9.34,
"learning_rate": 0.00028568748233964394,
"loss": 0.0853,
"step": 22200
},
{
"epoch": 9.34,
"eval_loss": 0.25565269589424133,
"eval_runtime": 442.9887,
"eval_samples_per_second": 11.768,
"eval_wer": 0.17079432587329166,
"step": 22200
},
{
"epoch": 9.47,
"learning_rate": 0.0002839412263351229,
"loss": 0.0792,
"step": 22500
},
{
"epoch": 9.47,
"eval_loss": 0.24455079436302185,
"eval_runtime": 447.726,
"eval_samples_per_second": 11.643,
"eval_wer": 0.16685251938487072,
"step": 22500
},
{
"epoch": 9.6,
"learning_rate": 0.0002821949703306018,
"loss": 0.0864,
"step": 22800
},
{
"epoch": 9.6,
"eval_loss": 0.2556290626525879,
"eval_runtime": 421.6345,
"eval_samples_per_second": 12.364,
"eval_wer": 0.16667098882290396,
"step": 22800
},
{
"epoch": 9.72,
"learning_rate": 0.0002804487143260808,
"loss": 0.092,
"step": 23100
},
{
"epoch": 9.72,
"eval_loss": 0.2547804117202759,
"eval_runtime": 444.7489,
"eval_samples_per_second": 11.721,
"eval_wer": 0.17476206529913643,
"step": 23100
},
{
"epoch": 9.85,
"learning_rate": 0.00027870245832155977,
"loss": 0.0915,
"step": 23400
},
{
"epoch": 9.85,
"eval_loss": 0.23804427683353424,
"eval_runtime": 440.9531,
"eval_samples_per_second": 11.822,
"eval_wer": 0.16879748969165737,
"step": 23400
},
{
"epoch": 9.97,
"learning_rate": 0.00027695620231703874,
"loss": 0.0926,
"step": 23700
},
{
"epoch": 9.97,
"eval_loss": 0.2537280321121216,
"eval_runtime": 446.3024,
"eval_samples_per_second": 11.68,
"eval_wer": 0.1642332927050647,
"step": 23700
},
{
"epoch": 10.1,
"learning_rate": 0.00027520994631251766,
"loss": 0.0819,
"step": 24000
},
{
"epoch": 10.1,
"eval_loss": 0.24399155378341675,
"eval_runtime": 420.6608,
"eval_samples_per_second": 12.392,
"eval_wer": 0.16405176214309794,
"step": 24000
},
{
"epoch": 10.23,
"learning_rate": 0.0002734636903079966,
"loss": 0.08,
"step": 24300
},
{
"epoch": 10.23,
"eval_loss": 0.24013200402259827,
"eval_runtime": 448.905,
"eval_samples_per_second": 11.613,
"eval_wer": 0.15940976634423382,
"step": 24300
},
{
"epoch": 10.35,
"learning_rate": 0.00027171743430347555,
"loss": 0.0775,
"step": 24600
},
{
"epoch": 10.35,
"eval_loss": 0.2440200299024582,
"eval_runtime": 449.5925,
"eval_samples_per_second": 11.595,
"eval_wer": 0.1610435414019346,
"step": 24600
},
{
"epoch": 10.48,
"learning_rate": 0.00026997117829895447,
"loss": 0.0772,
"step": 24900
},
{
"epoch": 10.48,
"eval_loss": 0.27414917945861816,
"eval_runtime": 419.8275,
"eval_samples_per_second": 12.417,
"eval_wer": 0.166230128886699,
"step": 24900
},
{
"epoch": 10.61,
"learning_rate": 0.00026822492229443344,
"loss": 0.0808,
"step": 25200
},
{
"epoch": 10.61,
"eval_loss": 0.2687024176120758,
"eval_runtime": 443.5669,
"eval_samples_per_second": 11.752,
"eval_wer": 0.1630663105209927,
"step": 25200
},
{
"epoch": 10.73,
"learning_rate": 0.0002664786662899124,
"loss": 0.0801,
"step": 25500
},
{
"epoch": 10.73,
"eval_loss": 0.25520530343055725,
"eval_runtime": 450.4913,
"eval_samples_per_second": 11.572,
"eval_wer": 0.1645704208915744,
"step": 25500
},
{
"epoch": 10.86,
"learning_rate": 0.0002647324102853913,
"loss": 0.0852,
"step": 25800
},
{
"epoch": 10.86,
"eval_loss": 0.24987785518169403,
"eval_runtime": 423.7767,
"eval_samples_per_second": 12.301,
"eval_wer": 0.16527061020201758,
"step": 25800
},
{
"epoch": 10.98,
"learning_rate": 0.0002629861542808703,
"loss": 0.0822,
"step": 26100
},
{
"epoch": 10.98,
"eval_loss": 0.24487553536891937,
"eval_runtime": 453.0579,
"eval_samples_per_second": 11.506,
"eval_wer": 0.1553901610435414,
"step": 26100
},
{
"epoch": 11.11,
"learning_rate": 0.0002612398982763492,
"loss": 0.0693,
"step": 26400
},
{
"epoch": 11.11,
"eval_loss": 0.2533997893333435,
"eval_runtime": 453.0359,
"eval_samples_per_second": 11.507,
"eval_wer": 0.14846606675138094,
"step": 26400
},
{
"epoch": 11.24,
"learning_rate": 0.0002594936422718282,
"loss": 0.0694,
"step": 26700
},
{
"epoch": 11.24,
"eval_loss": 0.2577211260795593,
"eval_runtime": 448.0425,
"eval_samples_per_second": 11.635,
"eval_wer": 0.15048883587043904,
"step": 26700
},
{
"epoch": 11.36,
"learning_rate": 0.0002577473862673071,
"loss": 0.0718,
"step": 27000
},
{
"epoch": 11.36,
"eval_loss": 0.24945984780788422,
"eval_runtime": 422.4019,
"eval_samples_per_second": 12.341,
"eval_wer": 0.15331552604963564,
"step": 27000
},
{
"epoch": 11.49,
"learning_rate": 0.00025600113026278613,
"loss": 0.0712,
"step": 27300
},
{
"epoch": 11.49,
"eval_loss": 0.25240784883499146,
"eval_runtime": 445.1796,
"eval_samples_per_second": 11.71,
"eval_wer": 0.15487150229506497,
"step": 27300
},
{
"epoch": 11.62,
"learning_rate": 0.00025425487425826505,
"loss": 0.0724,
"step": 27600
},
{
"epoch": 11.62,
"eval_loss": 0.2555846869945526,
"eval_runtime": 442.017,
"eval_samples_per_second": 11.794,
"eval_wer": 0.15507896579445554,
"step": 27600
},
{
"epoch": 11.74,
"learning_rate": 0.00025250861825374397,
"loss": 0.0733,
"step": 27900
},
{
"epoch": 11.74,
"eval_loss": 0.2552262246608734,
"eval_runtime": 422.3705,
"eval_samples_per_second": 12.342,
"eval_wer": 0.15331552604963564,
"step": 27900
},
{
"epoch": 11.87,
"learning_rate": 0.00025076236224922294,
"loss": 0.0739,
"step": 28200
},
{
"epoch": 11.87,
"eval_loss": 0.25745663046836853,
"eval_runtime": 446.6268,
"eval_samples_per_second": 11.672,
"eval_wer": 0.15181141567905396,
"step": 28200
},
{
"epoch": 11.99,
"learning_rate": 0.00024901610624470185,
"loss": 0.0756,
"step": 28500
},
{
"epoch": 11.99,
"eval_loss": 0.25365495681762695,
"eval_runtime": 445.0479,
"eval_samples_per_second": 11.713,
"eval_wer": 0.15199294624102072,
"step": 28500
},
{
"epoch": 12.12,
"learning_rate": 0.0002472698502401808,
"loss": 0.0623,
"step": 28800
},
{
"epoch": 12.12,
"eval_loss": 0.25725752115249634,
"eval_runtime": 422.5366,
"eval_samples_per_second": 12.337,
"eval_wer": 0.15450844117113147,
"step": 28800
},
{
"epoch": 12.25,
"learning_rate": 0.0002455235942356598,
"loss": 0.0642,
"step": 29100
},
{
"epoch": 12.25,
"eval_loss": 0.2661702334880829,
"eval_runtime": 452.9004,
"eval_samples_per_second": 11.51,
"eval_wer": 0.1533673919244833,
"step": 29100
},
{
"epoch": 12.37,
"learning_rate": 0.00024377733823113874,
"loss": 0.0656,
"step": 29400
},
{
"epoch": 12.37,
"eval_loss": 0.27368617057800293,
"eval_runtime": 444.5703,
"eval_samples_per_second": 11.726,
"eval_wer": 0.15121495811830607,
"step": 29400
},
{
"epoch": 12.5,
"learning_rate": 0.0002420310822266177,
"loss": 0.0713,
"step": 29700
},
{
"epoch": 12.5,
"eval_loss": 0.27105453610420227,
"eval_runtime": 422.4149,
"eval_samples_per_second": 12.341,
"eval_wer": 0.1492699878115194,
"step": 29700
},
{
"epoch": 12.63,
"learning_rate": 0.00024028482622209663,
"loss": 0.0667,
"step": 30000
},
{
"epoch": 12.63,
"eval_loss": 0.24904416501522064,
"eval_runtime": 450.1403,
"eval_samples_per_second": 11.581,
"eval_wer": 0.150203573558777,
"step": 30000
},
{
"epoch": 12.75,
"learning_rate": 0.00023853857021757558,
"loss": 0.0629,
"step": 30300
},
{
"epoch": 12.75,
"eval_loss": 0.26200103759765625,
"eval_runtime": 443.2234,
"eval_samples_per_second": 11.762,
"eval_wer": 0.1519670133035969,
"step": 30300
},
{
"epoch": 12.88,
"learning_rate": 0.00023679231421305452,
"loss": 0.0657,
"step": 30600
},
{
"epoch": 12.88,
"eval_loss": 0.2610265016555786,
"eval_runtime": 438.2352,
"eval_samples_per_second": 11.895,
"eval_wer": 0.1481030056274474,
"step": 30600
},
{
"epoch": 13.01,
"learning_rate": 0.0002350460582085335,
"loss": 0.0712,
"step": 30900
},
{
"epoch": 13.01,
"eval_loss": 0.2637769281864166,
"eval_runtime": 423.4731,
"eval_samples_per_second": 12.31,
"eval_wer": 0.1468841575685278,
"step": 30900
},
{
"epoch": 13.13,
"learning_rate": 0.00023329980220401244,
"loss": 0.0624,
"step": 31200
},
{
"epoch": 13.13,
"eval_loss": 0.25238868594169617,
"eval_runtime": 449.8496,
"eval_samples_per_second": 11.588,
"eval_wer": 0.14999611005938643,
"step": 31200
},
{
"epoch": 13.26,
"learning_rate": 0.00023155354619949138,
"loss": 0.0592,
"step": 31500
},
{
"epoch": 13.26,
"eval_loss": 0.25298765301704407,
"eval_runtime": 451.2139,
"eval_samples_per_second": 11.553,
"eval_wer": 0.1419309665205778,
"step": 31500
},
{
"epoch": 13.38,
"learning_rate": 0.00022980729019497033,
"loss": 0.0634,
"step": 31800
},
{
"epoch": 13.38,
"eval_loss": 0.2578509449958801,
"eval_runtime": 421.4953,
"eval_samples_per_second": 12.368,
"eval_wer": 0.15080003111952492,
"step": 31800
},
{
"epoch": 13.51,
"learning_rate": 0.00022806103419044927,
"loss": 0.0575,
"step": 32100
},
{
"epoch": 13.51,
"eval_loss": 0.24078154563903809,
"eval_runtime": 445.6242,
"eval_samples_per_second": 11.698,
"eval_wer": 0.1440574673893312,
"step": 32100
},
{
"epoch": 13.64,
"learning_rate": 0.00022631477818592821,
"loss": 0.0571,
"step": 32400
},
{
"epoch": 13.64,
"eval_loss": 0.24118457734584808,
"eval_runtime": 441.1228,
"eval_samples_per_second": 11.818,
"eval_wer": 0.13801509296958067,
"step": 32400
},
{
"epoch": 13.76,
"learning_rate": 0.00022456852218140716,
"loss": 0.0583,
"step": 32700
},
{
"epoch": 13.76,
"eval_loss": 0.2555965483188629,
"eval_runtime": 420.9676,
"eval_samples_per_second": 12.383,
"eval_wer": 0.1451207178237079,
"step": 32700
},
{
"epoch": 13.89,
"learning_rate": 0.00022282226617688613,
"loss": 0.0564,
"step": 33000
},
{
"epoch": 13.89,
"eval_loss": 0.25013622641563416,
"eval_runtime": 445.1032,
"eval_samples_per_second": 11.712,
"eval_wer": 0.1420606312076969,
"step": 33000
},
{
"epoch": 14.02,
"learning_rate": 0.00022107601017236508,
"loss": 0.0628,
"step": 33300
},
{
"epoch": 14.02,
"eval_loss": 0.2687903344631195,
"eval_runtime": 448.3298,
"eval_samples_per_second": 11.628,
"eval_wer": 0.14968491481030055,
"step": 33300
},
{
"epoch": 14.14,
"learning_rate": 0.00021932975416784402,
"loss": 0.0565,
"step": 33600
},
{
"epoch": 14.14,
"eval_loss": 0.2639634609222412,
"eval_runtime": 444.628,
"eval_samples_per_second": 11.724,
"eval_wer": 0.14154197245922046,
"step": 33600
},
{
"epoch": 14.27,
"learning_rate": 0.00021758349816332296,
"loss": 0.0539,
"step": 33900
},
{
"epoch": 14.27,
"eval_loss": 0.25850602984428406,
"eval_runtime": 422.0625,
"eval_samples_per_second": 12.351,
"eval_wer": 0.14177536889603484,
"step": 33900
},
{
"epoch": 14.39,
"learning_rate": 0.0002158372421588019,
"loss": 0.0515,
"step": 34200
},
{
"epoch": 14.39,
"eval_loss": 0.2456846386194229,
"eval_runtime": 439.8595,
"eval_samples_per_second": 11.852,
"eval_wer": 0.13741863540883276,
"step": 34200
},
{
"epoch": 14.52,
"learning_rate": 0.00021409098615428085,
"loss": 0.0535,
"step": 34500
},
{
"epoch": 14.52,
"eval_loss": 0.2575737535953522,
"eval_runtime": 450.7209,
"eval_samples_per_second": 11.566,
"eval_wer": 0.13684811078550868,
"step": 34500
},
{
"epoch": 14.65,
"learning_rate": 0.00021234473014975982,
"loss": 0.0528,
"step": 34800
},
{
"epoch": 14.65,
"eval_loss": 0.24703530967235565,
"eval_runtime": 422.3035,
"eval_samples_per_second": 12.344,
"eval_wer": 0.13889681284199062,
"step": 34800
},
{
"epoch": 14.77,
"learning_rate": 0.00021059847414523877,
"loss": 0.0556,
"step": 35100
},
{
"epoch": 14.77,
"eval_loss": 0.2394973188638687,
"eval_runtime": 452.3855,
"eval_samples_per_second": 11.523,
"eval_wer": 0.13962293508985762,
"step": 35100
},
{
"epoch": 14.9,
"learning_rate": 0.00020885221814071771,
"loss": 0.0562,
"step": 35400
},
{
"epoch": 14.9,
"eval_loss": 0.2461041510105133,
"eval_runtime": 443.6609,
"eval_samples_per_second": 11.75,
"eval_wer": 0.13835222115609036,
"step": 35400
},
{
"epoch": 15.03,
"learning_rate": 0.00020710596213619666,
"loss": 0.055,
"step": 35700
},
{
"epoch": 15.03,
"eval_loss": 0.261261522769928,
"eval_runtime": 421.1886,
"eval_samples_per_second": 12.377,
"eval_wer": 0.14125671014755842,
"step": 35700
},
{
"epoch": 15.15,
"learning_rate": 0.0002053597061316756,
"loss": 0.0504,
"step": 36000
},
{
"epoch": 15.15,
"eval_loss": 0.2512105405330658,
"eval_runtime": 449.4571,
"eval_samples_per_second": 11.598,
"eval_wer": 0.14104924664816784,
"step": 36000
},
{
"epoch": 15.28,
"learning_rate": 0.00020361345012715457,
"loss": 0.0492,
"step": 36300
},
{
"epoch": 15.28,
"eval_loss": 0.2652667164802551,
"eval_runtime": 447.4649,
"eval_samples_per_second": 11.65,
"eval_wer": 0.1455875106973367,
"step": 36300
},
{
"epoch": 15.4,
"learning_rate": 0.0002018671941226335,
"loss": 0.0495,
"step": 36600
},
{
"epoch": 15.4,
"eval_loss": 0.25162389874458313,
"eval_runtime": 420.3357,
"eval_samples_per_second": 12.402,
"eval_wer": 0.1370037084100516,
"step": 36600
},
{
"epoch": 15.53,
"learning_rate": 0.00020012093811811244,
"loss": 0.0549,
"step": 36900
},
{
"epoch": 15.53,
"eval_loss": 0.2618192136287689,
"eval_runtime": 445.4016,
"eval_samples_per_second": 11.704,
"eval_wer": 0.14042685614999612,
"step": 36900
},
{
"epoch": 15.66,
"learning_rate": 0.0001983746821135914,
"loss": 0.0497,
"step": 37200
},
{
"epoch": 15.66,
"eval_loss": 0.254682332277298,
"eval_runtime": 449.7454,
"eval_samples_per_second": 11.591,
"eval_wer": 0.13679624491066103,
"step": 37200
},
{
"epoch": 15.78,
"learning_rate": 0.00019662842610907035,
"loss": 0.0517,
"step": 37500
},
{
"epoch": 15.78,
"eval_loss": 0.2562556862831116,
"eval_runtime": 452.5481,
"eval_samples_per_second": 11.519,
"eval_wer": 0.13721117190944218,
"step": 37500
},
{
"epoch": 15.91,
"learning_rate": 0.0001948821701045493,
"loss": 0.0541,
"step": 37800
},
{
"epoch": 15.91,
"eval_loss": 0.25422725081443787,
"eval_runtime": 421.3988,
"eval_samples_per_second": 12.371,
"eval_wer": 0.13549959803946993,
"step": 37800
},
{
"epoch": 16.04,
"learning_rate": 0.00019313591410002827,
"loss": 0.0494,
"step": 38100
},
{
"epoch": 16.04,
"eval_loss": 0.2502134144306183,
"eval_runtime": 457.8032,
"eval_samples_per_second": 11.387,
"eval_wer": 0.13267290786027333,
"step": 38100
},
{
"epoch": 16.16,
"learning_rate": 0.00019138965809550719,
"loss": 0.0447,
"step": 38400
},
{
"epoch": 16.16,
"eval_loss": 0.2713734805583954,
"eval_runtime": 453.8402,
"eval_samples_per_second": 11.486,
"eval_wer": 0.13220611498664453,
"step": 38400
},
{
"epoch": 16.29,
"learning_rate": 0.00018964340209098613,
"loss": 0.0441,
"step": 38700
},
{
"epoch": 16.29,
"eval_loss": 0.25368762016296387,
"eval_runtime": 421.2775,
"eval_samples_per_second": 12.374,
"eval_wer": 0.13114286455226784,
"step": 38700
},
{
"epoch": 16.41,
"learning_rate": 0.0001878971460864651,
"loss": 0.0457,
"step": 39000
},
{
"epoch": 16.41,
"eval_loss": 0.24593111872673035,
"eval_runtime": 450.5495,
"eval_samples_per_second": 11.57,
"eval_wer": 0.13430668291797412,
"step": 39000
},
{
"epoch": 16.54,
"learning_rate": 0.00018615089008194405,
"loss": 0.0484,
"step": 39300
},
{
"epoch": 16.54,
"eval_loss": 0.24097564816474915,
"eval_runtime": 447.1986,
"eval_samples_per_second": 11.657,
"eval_wer": 0.1310909986774202,
"step": 39300
},
{
"epoch": 16.67,
"learning_rate": 0.000184404634077423,
"loss": 0.0435,
"step": 39600
},
{
"epoch": 16.67,
"eval_loss": 0.24710460007190704,
"eval_runtime": 420.3947,
"eval_samples_per_second": 12.4,
"eval_wer": 0.13150592567620134,
"step": 39600
},
{
"epoch": 16.79,
"learning_rate": 0.00018265837807290196,
"loss": 0.045,
"step": 39900
},
{
"epoch": 16.79,
"eval_loss": 0.24996981024742126,
"eval_runtime": 445.0069,
"eval_samples_per_second": 11.714,
"eval_wer": 0.12870516843442856,
"step": 39900
},
{
"epoch": 16.92,
"learning_rate": 0.00018091212206838088,
"loss": 0.0476,
"step": 40200
},
{
"epoch": 16.92,
"eval_loss": 0.24158866703510284,
"eval_runtime": 450.2986,
"eval_samples_per_second": 11.577,
"eval_wer": 0.12753818625035657,
"step": 40200
},
{
"epoch": 17.05,
"learning_rate": 0.00017916586606385983,
"loss": 0.0421,
"step": 40500
},
{
"epoch": 17.05,
"eval_loss": 0.25376275181770325,
"eval_runtime": 439.9451,
"eval_samples_per_second": 11.849,
"eval_wer": 0.12727885687611837,
"step": 40500
},
{
"epoch": 17.17,
"learning_rate": 0.0001774196100593388,
"loss": 0.043,
"step": 40800
},
{
"epoch": 17.17,
"eval_loss": 0.2612881064414978,
"eval_runtime": 420.2448,
"eval_samples_per_second": 12.405,
"eval_wer": 0.13638131791187988,
"step": 40800
},
{
"epoch": 17.3,
"learning_rate": 0.00017567335405481774,
"loss": 0.0421,
"step": 41100
},
{
"epoch": 17.3,
"eval_loss": 0.25012627243995667,
"eval_runtime": 442.3555,
"eval_samples_per_second": 11.785,
"eval_wer": 0.129949949430772,
"step": 41100
},
{
"epoch": 17.42,
"learning_rate": 0.00017392709805029669,
"loss": 0.0455,
"step": 41400
},
{
"epoch": 17.42,
"eval_loss": 0.27013105154037476,
"eval_runtime": 439.1226,
"eval_samples_per_second": 11.871,
"eval_wer": 0.13282850548481626,
"step": 41400
},
{
"epoch": 17.55,
"learning_rate": 0.00017218084204577566,
"loss": 0.0447,
"step": 41700
},
{
"epoch": 17.55,
"eval_loss": 0.25483304262161255,
"eval_runtime": 422.116,
"eval_samples_per_second": 12.35,
"eval_wer": 0.13098726692772492,
"step": 41700
},
{
"epoch": 17.68,
"learning_rate": 0.0001704345860412546,
"loss": 0.0422,
"step": 42000
},
{
"epoch": 17.68,
"eval_loss": 0.25715696811676025,
"eval_runtime": 455.1849,
"eval_samples_per_second": 11.452,
"eval_wer": 0.12795311324913772,
"step": 42000
},
{
"epoch": 17.8,
"learning_rate": 0.00016868833003673352,
"loss": 0.0421,
"step": 42300
},
{
"epoch": 17.8,
"eval_loss": 0.25144657492637634,
"eval_runtime": 438.538,
"eval_samples_per_second": 11.887,
"eval_wer": 0.13508467104068878,
"step": 42300
},
{
"epoch": 17.93,
"learning_rate": 0.00016694207403221246,
"loss": 0.0421,
"step": 42600
},
{
"epoch": 17.93,
"eval_loss": 0.25014957785606384,
"eval_runtime": 419.3857,
"eval_samples_per_second": 12.43,
"eval_wer": 0.12906822955836209,
"step": 42600
},
{
"epoch": 18.06,
"learning_rate": 0.00016519581802769144,
"loss": 0.0412,
"step": 42900
},
{
"epoch": 18.06,
"eval_loss": 0.25013405084609985,
"eval_runtime": 447.7509,
"eval_samples_per_second": 11.643,
"eval_wer": 0.12891263193381913,
"step": 42900
},
{
"epoch": 18.18,
"learning_rate": 0.00016344956202317038,
"loss": 0.0365,
"step": 43200
},
{
"epoch": 18.18,
"eval_loss": 0.24244625866413116,
"eval_runtime": 452.5128,
"eval_samples_per_second": 11.52,
"eval_wer": 0.13249137729830657,
"step": 43200
},
{
"epoch": 18.31,
"learning_rate": 0.00016170330601864932,
"loss": 0.037,
"step": 43500
},
{
"epoch": 18.31,
"eval_loss": 0.24644562602043152,
"eval_runtime": 447.1204,
"eval_samples_per_second": 11.659,
"eval_wer": 0.12963875418168616,
"step": 43500
},
{
"epoch": 18.43,
"learning_rate": 0.0001599570500141283,
"loss": 0.0398,
"step": 43800
},
{
"epoch": 18.43,
"eval_loss": 0.24118588864803314,
"eval_runtime": 420.9169,
"eval_samples_per_second": 12.385,
"eval_wer": 0.12551541713129846,
"step": 43800
},
{
"epoch": 18.56,
"learning_rate": 0.00015821079400960721,
"loss": 0.0398,
"step": 44100
},
{
"epoch": 18.56,
"eval_loss": 0.2401241511106491,
"eval_runtime": 443.2738,
"eval_samples_per_second": 11.76,
"eval_wer": 0.12642306994113223,
"step": 44100
},
{
"epoch": 18.69,
"learning_rate": 0.00015646453800508616,
"loss": 0.0404,
"step": 44400
},
{
"epoch": 18.69,
"eval_loss": 0.24249568581581116,
"eval_runtime": 454.0524,
"eval_samples_per_second": 11.481,
"eval_wer": 0.12416690438525972,
"step": 44400
},
{
"epoch": 18.81,
"learning_rate": 0.00015471828200056513,
"loss": 0.0385,
"step": 44700
},
{
"epoch": 18.81,
"eval_loss": 0.2565571069717407,
"eval_runtime": 421.1105,
"eval_samples_per_second": 12.379,
"eval_wer": 0.13007961411789112,
"step": 44700
},
{
"epoch": 18.94,
"learning_rate": 0.00015297202599604407,
"loss": 0.0409,
"step": 45000
},
{
"epoch": 18.94,
"eval_loss": 0.24367791414260864,
"eval_runtime": 440.1807,
"eval_samples_per_second": 11.843,
"eval_wer": 0.12823837556079978,
"step": 45000
},
{
"epoch": 19.07,
"learning_rate": 0.00015122576999152302,
"loss": 0.0386,
"step": 45300
},
{
"epoch": 19.07,
"eval_loss": 0.2512024939060211,
"eval_runtime": 448.55,
"eval_samples_per_second": 11.622,
"eval_wer": 0.1251004901325173,
"step": 45300
},
{
"epoch": 19.19,
"learning_rate": 0.000149479513987002,
"loss": 0.0382,
"step": 45600
},
{
"epoch": 19.19,
"eval_loss": 0.26347675919532776,
"eval_runtime": 420.744,
"eval_samples_per_second": 12.39,
"eval_wer": 0.1246077643214647,
"step": 45600
},
{
"epoch": 19.32,
"learning_rate": 0.00014773325798248094,
"loss": 0.0356,
"step": 45900
},
{
"epoch": 19.32,
"eval_loss": 0.24348150193691254,
"eval_runtime": 444.991,
"eval_samples_per_second": 11.715,
"eval_wer": 0.11929151214958118,
"step": 45900
},
{
"epoch": 19.44,
"learning_rate": 0.00014598700197795985,
"loss": 0.0347,
"step": 46200
},
{
"epoch": 19.44,
"eval_loss": 0.24703486263751984,
"eval_runtime": 450.7542,
"eval_samples_per_second": 11.565,
"eval_wer": 0.11866912165140946,
"step": 46200
},
{
"epoch": 19.57,
"learning_rate": 0.00014424074597343882,
"loss": 0.0351,
"step": 46500
},
{
"epoch": 19.57,
"eval_loss": 0.25269144773483276,
"eval_runtime": 420.9044,
"eval_samples_per_second": 12.385,
"eval_wer": 0.12250719639013512,
"step": 46500
},
{
"epoch": 19.7,
"learning_rate": 0.00014249448996891777,
"loss": 0.0387,
"step": 46800
},
{
"epoch": 19.7,
"eval_loss": 0.2696446180343628,
"eval_runtime": 447.6183,
"eval_samples_per_second": 11.646,
"eval_wer": 0.12899043074609062,
"step": 46800
},
{
"epoch": 19.82,
"learning_rate": 0.0001407482339643967,
"loss": 0.0381,
"step": 47100
},
{
"epoch": 19.82,
"eval_loss": 0.24737927317619324,
"eval_runtime": 454.8982,
"eval_samples_per_second": 11.46,
"eval_wer": 0.123544513887088,
"step": 47100
},
{
"epoch": 19.95,
"learning_rate": 0.00013900197795987568,
"loss": 0.0353,
"step": 47400
},
{
"epoch": 19.95,
"eval_loss": 0.24795196950435638,
"eval_runtime": 453.6312,
"eval_samples_per_second": 11.492,
"eval_wer": 0.12318145276315448,
"step": 47400
},
{
"epoch": 20.08,
"learning_rate": 0.00013725572195535463,
"loss": 0.0339,
"step": 47700
},
{
"epoch": 20.08,
"eval_loss": 0.25215083360671997,
"eval_runtime": 420.9875,
"eval_samples_per_second": 12.383,
"eval_wer": 0.12250719639013512,
"step": 47700
},
{
"epoch": 20.2,
"learning_rate": 0.00013550946595083355,
"loss": 0.0329,
"step": 48000
},
{
"epoch": 20.2,
"eval_loss": 0.25695204734802246,
"eval_runtime": 443.8662,
"eval_samples_per_second": 11.745,
"eval_wer": 0.11952490858639558,
"step": 48000
},
{
"epoch": 20.33,
"learning_rate": 0.00013376320994631252,
"loss": 0.0316,
"step": 48300
},
{
"epoch": 20.33,
"eval_loss": 0.25657811760902405,
"eval_runtime": 443.4461,
"eval_samples_per_second": 11.756,
"eval_wer": 0.11999170146002437,
"step": 48300
},
{
"epoch": 20.45,
"learning_rate": 0.00013201695394179146,
"loss": 0.0339,
"step": 48600
},
{
"epoch": 20.45,
"eval_loss": 0.2606305181980133,
"eval_runtime": 420.9039,
"eval_samples_per_second": 12.385,
"eval_wer": 0.11809859702808537,
"step": 48600
},
{
"epoch": 20.58,
"learning_rate": 0.0001302706979372704,
"loss": 0.0324,
"step": 48900
},
{
"epoch": 20.58,
"eval_loss": 0.2572144865989685,
"eval_runtime": 456.9807,
"eval_samples_per_second": 11.407,
"eval_wer": 0.11488291278753145,
"step": 48900
},
{
"epoch": 20.71,
"learning_rate": 0.00012852444193274935,
"loss": 0.0306,
"step": 49200
},
{
"epoch": 20.71,
"eval_loss": 0.25449544191360474,
"eval_runtime": 449.6003,
"eval_samples_per_second": 11.595,
"eval_wer": 0.11783926765384715,
"step": 49200
},
{
"epoch": 20.83,
"learning_rate": 0.00012677818592822832,
"loss": 0.0339,
"step": 49500
},
{
"epoch": 20.83,
"eval_loss": 0.24496378004550934,
"eval_runtime": 420.0292,
"eval_samples_per_second": 12.411,
"eval_wer": 0.1147013822255647,
"step": 49500
},
{
"epoch": 20.96,
"learning_rate": 0.00012503192992370724,
"loss": 0.0344,
"step": 49800
},
{
"epoch": 20.96,
"eval_loss": 0.23759284615516663,
"eval_runtime": 450.494,
"eval_samples_per_second": 11.572,
"eval_wer": 0.12004356733487202,
"step": 49800
},
{
"epoch": 21.09,
"learning_rate": 0.00012328567391918619,
"loss": 0.0302,
"step": 50100
},
{
"epoch": 21.09,
"eval_loss": 0.2617592215538025,
"eval_runtime": 450.3195,
"eval_samples_per_second": 11.576,
"eval_wer": 0.11584243147221286,
"step": 50100
},
{
"epoch": 21.21,
"learning_rate": 0.00012153941791466516,
"loss": 0.0312,
"step": 50400
},
{
"epoch": 21.21,
"eval_loss": 0.25956034660339355,
"eval_runtime": 440.9223,
"eval_samples_per_second": 11.823,
"eval_wer": 0.11462358341329322,
"step": 50400
},
{
"epoch": 21.34,
"learning_rate": 0.0001197931619101441,
"loss": 0.0296,
"step": 50700
},
{
"epoch": 21.34,
"eval_loss": 0.25406599044799805,
"eval_runtime": 422.3783,
"eval_samples_per_second": 12.342,
"eval_wer": 0.11480511397525998,
"step": 50700
},
{
"epoch": 21.46,
"learning_rate": 0.00011804690590562305,
"loss": 0.0288,
"step": 51000
},
{
"epoch": 21.46,
"eval_loss": 0.25333333015441895,
"eval_runtime": 451.8621,
"eval_samples_per_second": 11.537,
"eval_wer": 0.11270454604393039,
"step": 51000
},
{
"epoch": 21.59,
"learning_rate": 0.000116300649901102,
"loss": 0.0305,
"step": 51300
},
{
"epoch": 21.59,
"eval_loss": 0.2583317756652832,
"eval_runtime": 448.0342,
"eval_samples_per_second": 11.635,
"eval_wer": 0.11542750447343171,
"step": 51300
},
{
"epoch": 21.72,
"learning_rate": 0.00011455439389658095,
"loss": 0.0298,
"step": 51600
},
{
"epoch": 21.72,
"eval_loss": 0.23815418779850006,
"eval_runtime": 423.0616,
"eval_samples_per_second": 12.322,
"eval_wer": 0.11524597391146495,
"step": 51600
},
{
"epoch": 21.84,
"learning_rate": 0.0001128081378920599,
"loss": 0.0291,
"step": 51900
},
{
"epoch": 21.84,
"eval_loss": 0.2578916549682617,
"eval_runtime": 448.7674,
"eval_samples_per_second": 11.616,
"eval_wer": 0.1138974611654262,
"step": 51900
},
{
"epoch": 21.97,
"learning_rate": 0.00011106188188753885,
"loss": 0.0295,
"step": 52200
},
{
"epoch": 21.97,
"eval_loss": 0.26773908734321594,
"eval_runtime": 447.9798,
"eval_samples_per_second": 11.637,
"eval_wer": 0.1136640647286118,
"step": 52200
},
{
"epoch": 22.1,
"learning_rate": 0.0001093156258830178,
"loss": 0.0279,
"step": 52500
},
{
"epoch": 22.1,
"eval_loss": 0.24904872477054596,
"eval_runtime": 420.4369,
"eval_samples_per_second": 12.399,
"eval_wer": 0.1128342107310495,
"step": 52500
},
{
"epoch": 22.22,
"learning_rate": 0.00010756936987849674,
"loss": 0.0283,
"step": 52800
},
{
"epoch": 22.22,
"eval_loss": 0.2628434896469116,
"eval_runtime": 449.8543,
"eval_samples_per_second": 11.588,
"eval_wer": 0.11433832110163118,
"step": 52800
},
{
"epoch": 22.35,
"learning_rate": 0.0001058231138739757,
"loss": 0.0277,
"step": 53100
},
{
"epoch": 22.35,
"eval_loss": 0.26064106822013855,
"eval_runtime": 447.314,
"eval_samples_per_second": 11.654,
"eval_wer": 0.11280827779362568,
"step": 53100
},
{
"epoch": 22.47,
"learning_rate": 0.00010407685786945464,
"loss": 0.0263,
"step": 53400
},
{
"epoch": 22.47,
"eval_loss": 0.264713853597641,
"eval_runtime": 446.9042,
"eval_samples_per_second": 11.665,
"eval_wer": 0.11210808848318249,
"step": 53400
},
{
"epoch": 22.6,
"learning_rate": 0.0001023306018649336,
"loss": 0.0278,
"step": 53700
},
{
"epoch": 22.6,
"eval_loss": 0.2553078830242157,
"eval_runtime": 421.9865,
"eval_samples_per_second": 12.353,
"eval_wer": 0.1103446487383626,
"step": 53700
},
{
"epoch": 22.73,
"learning_rate": 0.00010058434586041255,
"loss": 0.028,
"step": 54000
},
{
"epoch": 22.73,
"eval_loss": 0.25996631383895874,
"eval_runtime": 448.115,
"eval_samples_per_second": 11.633,
"eval_wer": 0.11226368610772543,
"step": 54000
},
{
"epoch": 22.85,
"learning_rate": 9.883808985589149e-05,
"loss": 0.0289,
"step": 54300
},
{
"epoch": 22.85,
"eval_loss": 0.2588217258453369,
"eval_runtime": 457.0011,
"eval_samples_per_second": 11.407,
"eval_wer": 0.11221182023287778,
"step": 54300
},
{
"epoch": 22.98,
"learning_rate": 9.709183385137043e-05,
"loss": 0.0288,
"step": 54600
},
{
"epoch": 22.98,
"eval_loss": 0.2563655972480774,
"eval_runtime": 421.5067,
"eval_samples_per_second": 12.368,
"eval_wer": 0.1130676071678639,
"step": 54600
},
{
"epoch": 23.11,
"learning_rate": 9.534557784684939e-05,
"loss": 0.0259,
"step": 54900
},
{
"epoch": 23.11,
"eval_loss": 0.25361210107803345,
"eval_runtime": 454.0477,
"eval_samples_per_second": 11.481,
"eval_wer": 0.10935919711625736,
"step": 54900
},
{
"epoch": 23.23,
"learning_rate": 9.359932184232834e-05,
"loss": 0.0278,
"step": 55200
},
{
"epoch": 23.23,
"eval_loss": 0.26548078656196594,
"eval_runtime": 453.6451,
"eval_samples_per_second": 11.491,
"eval_wer": 0.10974819117761468,
"step": 55200
},
{
"epoch": 23.36,
"learning_rate": 9.185306583780728e-05,
"loss": 0.0254,
"step": 55500
},
{
"epoch": 23.36,
"eval_loss": 0.27383095026016235,
"eval_runtime": 421.4102,
"eval_samples_per_second": 12.37,
"eval_wer": 0.11286014366847333,
"step": 55500
},
{
"epoch": 23.48,
"learning_rate": 9.010680983328624e-05,
"loss": 0.0276,
"step": 55800
},
{
"epoch": 23.48,
"eval_loss": 0.2656802833080292,
"eval_runtime": 449.064,
"eval_samples_per_second": 11.609,
"eval_wer": 0.1109411062991105,
"step": 55800
},
{
"epoch": 23.61,
"learning_rate": 8.836055382876518e-05,
"loss": 0.026,
"step": 56100
},
{
"epoch": 23.61,
"eval_loss": 0.27263152599334717,
"eval_runtime": 447.0215,
"eval_samples_per_second": 11.662,
"eval_wer": 0.11044838048805788,
"step": 56100
},
{
"epoch": 23.74,
"learning_rate": 8.661429782424413e-05,
"loss": 0.0258,
"step": 56400
},
{
"epoch": 23.74,
"eval_loss": 0.26607978343963623,
"eval_runtime": 422.7932,
"eval_samples_per_second": 12.33,
"eval_wer": 0.10775135499598039,
"step": 56400
},
{
"epoch": 23.86,
"learning_rate": 8.486804181972309e-05,
"loss": 0.0252,
"step": 56700
},
{
"epoch": 23.86,
"eval_loss": 0.26961830258369446,
"eval_runtime": 442.9007,
"eval_samples_per_second": 11.77,
"eval_wer": 0.10891833718005238,
"step": 56700
},
{
"epoch": 23.99,
"learning_rate": 8.312178581520203e-05,
"loss": 0.0255,
"step": 57000
},
{
"epoch": 23.99,
"eval_loss": 0.25439324975013733,
"eval_runtime": 447.4952,
"eval_samples_per_second": 11.649,
"eval_wer": 0.10808848318249008,
"step": 57000
},
{
"epoch": 24.12,
"learning_rate": 8.137552981068098e-05,
"loss": 0.0235,
"step": 57300
},
{
"epoch": 24.12,
"eval_loss": 0.26512473821640015,
"eval_runtime": 447.2758,
"eval_samples_per_second": 11.655,
"eval_wer": 0.1096444594279194,
"step": 57300
},
{
"epoch": 24.24,
"learning_rate": 7.962927380615993e-05,
"loss": 0.0237,
"step": 57600
},
{
"epoch": 24.24,
"eval_loss": 0.2661215364933014,
"eval_runtime": 422.3577,
"eval_samples_per_second": 12.343,
"eval_wer": 0.10912580067944296,
"step": 57600
},
{
"epoch": 24.37,
"learning_rate": 7.788301780163888e-05,
"loss": 0.0259,
"step": 57900
},
{
"epoch": 24.37,
"eval_loss": 0.2603491246700287,
"eval_runtime": 447.2071,
"eval_samples_per_second": 11.657,
"eval_wer": 0.1081144161199139,
"step": 57900
},
{
"epoch": 24.49,
"learning_rate": 7.613676179711782e-05,
"loss": 0.022,
"step": 58200
},
{
"epoch": 24.49,
"eval_loss": 0.25694188475608826,
"eval_runtime": 447.9008,
"eval_samples_per_second": 11.639,
"eval_wer": 0.10653250693706076,
"step": 58200
},
{
"epoch": 24.62,
"learning_rate": 7.439050579259678e-05,
"loss": 0.0239,
"step": 58500
},
{
"epoch": 24.62,
"eval_loss": 0.2633558511734009,
"eval_runtime": 422.3281,
"eval_samples_per_second": 12.343,
"eval_wer": 0.10868494074323799,
"step": 58500
},
{
"epoch": 24.75,
"learning_rate": 7.264424978807573e-05,
"loss": 0.0226,
"step": 58800
},
{
"epoch": 24.75,
"eval_loss": 0.2715343236923218,
"eval_runtime": 451.351,
"eval_samples_per_second": 11.55,
"eval_wer": 0.10580638468919375,
"step": 58800
},
{
"epoch": 24.87,
"learning_rate": 7.089799378355467e-05,
"loss": 0.0236,
"step": 59100
},
{
"epoch": 24.87,
"eval_loss": 0.25477010011672974,
"eval_runtime": 442.3516,
"eval_samples_per_second": 11.785,
"eval_wer": 0.10487279894193616,
"step": 59100
},
{
"epoch": 25.0,
"learning_rate": 6.915173777903363e-05,
"loss": 0.0221,
"step": 59400
},
{
"epoch": 25.0,
"eval_loss": 0.26137542724609375,
"eval_runtime": 423.9191,
"eval_samples_per_second": 12.297,
"eval_wer": 0.10531365887814113,
"step": 59400
},
{
"epoch": 25.13,
"learning_rate": 6.740548177451257e-05,
"loss": 0.0228,
"step": 59700
},
{
"epoch": 25.13,
"eval_loss": 0.2630428969860077,
"eval_runtime": 457.649,
"eval_samples_per_second": 11.391,
"eval_wer": 0.10508026244132673,
"step": 59700
},
{
"epoch": 25.25,
"learning_rate": 6.565922576999152e-05,
"loss": 0.0211,
"step": 60000
},
{
"epoch": 25.25,
"eval_loss": 0.26376327872276306,
"eval_runtime": 438.7038,
"eval_samples_per_second": 11.883,
"eval_wer": 0.10510619537875054,
"step": 60000
},
{
"epoch": 25.38,
"learning_rate": 6.391296976547048e-05,
"loss": 0.0216,
"step": 60300
},
{
"epoch": 25.38,
"eval_loss": 0.27177637815475464,
"eval_runtime": 453.0435,
"eval_samples_per_second": 11.507,
"eval_wer": 0.10443193900573118,
"step": 60300
},
{
"epoch": 25.51,
"learning_rate": 6.216671376094942e-05,
"loss": 0.0241,
"step": 60600
},
{
"epoch": 25.51,
"eval_loss": 0.2619257867336273,
"eval_runtime": 422.4781,
"eval_samples_per_second": 12.339,
"eval_wer": 0.10329088975908302,
"step": 60600
},
{
"epoch": 25.63,
"learning_rate": 6.042045775642837e-05,
"loss": 0.0211,
"step": 60900
},
{
"epoch": 25.63,
"eval_loss": 0.26482462882995605,
"eval_runtime": 439.7256,
"eval_samples_per_second": 11.855,
"eval_wer": 0.10217577344985866,
"step": 60900
},
{
"epoch": 25.76,
"learning_rate": 5.867420175190732e-05,
"loss": 0.0216,
"step": 61200
},
{
"epoch": 25.76,
"eval_loss": 0.2719128131866455,
"eval_runtime": 444.9118,
"eval_samples_per_second": 11.717,
"eval_wer": 0.10227950519955395,
"step": 61200
},
{
"epoch": 25.88,
"learning_rate": 5.692794574738626e-05,
"loss": 0.0209,
"step": 61500
},
{
"epoch": 25.88,
"eval_loss": 0.2721671164035797,
"eval_runtime": 422.0845,
"eval_samples_per_second": 12.351,
"eval_wer": 0.10134591945229636,
"step": 61500
},
{
"epoch": 26.01,
"learning_rate": 5.518168974286522e-05,
"loss": 0.0212,
"step": 61800
},
{
"epoch": 26.01,
"eval_loss": 0.26508814096450806,
"eval_runtime": 445.5664,
"eval_samples_per_second": 11.7,
"eval_wer": 0.10051606545473406,
"step": 61800
},
{
"epoch": 26.14,
"learning_rate": 5.3435433738344156e-05,
"loss": 0.0204,
"step": 62100
},
{
"epoch": 26.14,
"eval_loss": 0.26492705941200256,
"eval_runtime": 452.3889,
"eval_samples_per_second": 11.523,
"eval_wer": 0.10220170638728249,
"step": 62100
},
{
"epoch": 26.26,
"learning_rate": 5.168917773382311e-05,
"loss": 0.0202,
"step": 62400
},
{
"epoch": 26.26,
"eval_loss": 0.278484046459198,
"eval_runtime": 422.0533,
"eval_samples_per_second": 12.352,
"eval_wer": 0.10313529213454008,
"step": 62400
},
{
"epoch": 26.39,
"learning_rate": 4.994292172930206e-05,
"loss": 0.0213,
"step": 62700
},
{
"epoch": 26.39,
"eval_loss": 0.272367924451828,
"eval_runtime": 442.0178,
"eval_samples_per_second": 11.794,
"eval_wer": 0.10033453489276731,
"step": 62700
},
{
"epoch": 26.52,
"learning_rate": 4.819666572478101e-05,
"loss": 0.0215,
"step": 63000
},
{
"epoch": 26.52,
"eval_loss": 0.2681904137134552,
"eval_runtime": 443.7076,
"eval_samples_per_second": 11.749,
"eval_wer": 0.10157931588911076,
"step": 63000
},
{
"epoch": 26.64,
"learning_rate": 4.645040972025996e-05,
"loss": 0.0188,
"step": 63300
},
{
"epoch": 26.64,
"eval_loss": 0.2568943500518799,
"eval_runtime": 423.3689,
"eval_samples_per_second": 12.313,
"eval_wer": 0.10023080314307202,
"step": 63300
},
{
"epoch": 26.77,
"learning_rate": 4.4704153715738906e-05,
"loss": 0.0192,
"step": 63600
},
{
"epoch": 26.77,
"eval_loss": 0.26485371589660645,
"eval_runtime": 442.7329,
"eval_samples_per_second": 11.775,
"eval_wer": 0.10111252301548196,
"step": 63600
},
{
"epoch": 26.89,
"learning_rate": 4.295789771121786e-05,
"loss": 0.0209,
"step": 63900
},
{
"epoch": 26.89,
"eval_loss": 0.25886571407318115,
"eval_runtime": 451.6321,
"eval_samples_per_second": 11.543,
"eval_wer": 0.099504680895205,
"step": 63900
},
{
"epoch": 27.02,
"learning_rate": 4.121164170669681e-05,
"loss": 0.0194,
"step": 64200
},
{
"epoch": 27.02,
"eval_loss": 0.2557186484336853,
"eval_runtime": 442.5519,
"eval_samples_per_second": 11.779,
"eval_wer": 0.09893415627188092,
"step": 64200
},
{
"epoch": 27.15,
"learning_rate": 3.946538570217575e-05,
"loss": 0.0196,
"step": 64500
},
{
"epoch": 27.15,
"eval_loss": 0.25841090083122253,
"eval_runtime": 424.1061,
"eval_samples_per_second": 12.292,
"eval_wer": 0.09989367495656233,
"step": 64500
},
{
"epoch": 27.27,
"learning_rate": 3.77191296976547e-05,
"loss": 0.019,
"step": 64800
},
{
"epoch": 27.27,
"eval_loss": 0.2668149173259735,
"eval_runtime": 448.6143,
"eval_samples_per_second": 11.62,
"eval_wer": 0.09877855864733798,
"step": 64800
},
{
"epoch": 27.4,
"learning_rate": 3.5972873693133656e-05,
"loss": 0.0181,
"step": 65100
},
{
"epoch": 27.4,
"eval_loss": 0.2628437876701355,
"eval_runtime": 450.825,
"eval_samples_per_second": 11.563,
"eval_wer": 0.09768937527553746,
"step": 65100
},
{
"epoch": 27.53,
"learning_rate": 3.42266176886126e-05,
"loss": 0.0175,
"step": 65400
},
{
"epoch": 27.53,
"eval_loss": 0.26442092657089233,
"eval_runtime": 420.8378,
"eval_samples_per_second": 12.387,
"eval_wer": 0.0998677420191385,
"step": 65400
},
{
"epoch": 27.65,
"learning_rate": 3.248036168409155e-05,
"loss": 0.0201,
"step": 65700
},
{
"epoch": 27.65,
"eval_loss": 0.2726499140262604,
"eval_runtime": 445.7776,
"eval_samples_per_second": 11.694,
"eval_wer": 0.09924535152096678,
"step": 65700
},
{
"epoch": 27.78,
"learning_rate": 3.07341056795705e-05,
"loss": 0.0182,
"step": 66000
},
{
"epoch": 27.78,
"eval_loss": 0.265592485666275,
"eval_runtime": 453.8658,
"eval_samples_per_second": 11.486,
"eval_wer": 0.09704105183994191,
"step": 66000
},
{
"epoch": 27.9,
"learning_rate": 2.8987849675049448e-05,
"loss": 0.0205,
"step": 66300
},
{
"epoch": 27.9,
"eval_loss": 0.2697185277938843,
"eval_runtime": 421.7967,
"eval_samples_per_second": 12.359,
"eval_wer": 0.09831176577370919,
"step": 66300
},
{
"epoch": 28.03,
"learning_rate": 2.7241593670528396e-05,
"loss": 0.0197,
"step": 66600
},
{
"epoch": 28.03,
"eval_loss": 0.26293227076530457,
"eval_runtime": 454.9345,
"eval_samples_per_second": 11.459,
"eval_wer": 0.09675578952827987,
"step": 66600
},
{
"epoch": 28.16,
"learning_rate": 2.5495337666007347e-05,
"loss": 0.018,
"step": 66900
},
{
"epoch": 28.16,
"eval_loss": 0.2722020149230957,
"eval_runtime": 451.9089,
"eval_samples_per_second": 11.536,
"eval_wer": 0.09683358834055134,
"step": 66900
},
{
"epoch": 28.28,
"learning_rate": 2.3749081661486295e-05,
"loss": 0.0172,
"step": 67200
},
{
"epoch": 28.28,
"eval_loss": 0.2774170935153961,
"eval_runtime": 451.1229,
"eval_samples_per_second": 11.556,
"eval_wer": 0.09685952127797516,
"step": 67200
},
{
"epoch": 28.41,
"learning_rate": 2.2002825656965243e-05,
"loss": 0.0158,
"step": 67500
},
{
"epoch": 28.41,
"eval_loss": 0.27579203248023987,
"eval_runtime": 425.5661,
"eval_samples_per_second": 12.25,
"eval_wer": 0.09735224708902777,
"step": 67500
},
{
"epoch": 28.54,
"learning_rate": 2.025656965244419e-05,
"loss": 0.0189,
"step": 67800
},
{
"epoch": 28.54,
"eval_loss": 0.2757831811904907,
"eval_runtime": 453.6044,
"eval_samples_per_second": 11.492,
"eval_wer": 0.09717071652706102,
"step": 67800
},
{
"epoch": 28.66,
"learning_rate": 1.8510313647923142e-05,
"loss": 0.0188,
"step": 68100
},
{
"epoch": 28.66,
"eval_loss": 0.281482994556427,
"eval_runtime": 443.0118,
"eval_samples_per_second": 11.767,
"eval_wer": 0.09660019190373693,
"step": 68100
},
{
"epoch": 28.79,
"learning_rate": 1.676405764340209e-05,
"loss": 0.0167,
"step": 68400
},
{
"epoch": 28.79,
"eval_loss": 0.2745157480239868,
"eval_runtime": 422.5171,
"eval_samples_per_second": 12.338,
"eval_wer": 0.09587406965586992,
"step": 68400
},
{
"epoch": 28.91,
"learning_rate": 1.501780163888104e-05,
"loss": 0.0149,
"step": 68700
},
{
"epoch": 28.91,
"eval_loss": 0.27913355827331543,
"eval_runtime": 451.1316,
"eval_samples_per_second": 11.555,
"eval_wer": 0.09660019190373693,
"step": 68700
},
{
"epoch": 29.04,
"learning_rate": 1.3271545634359989e-05,
"loss": 0.0168,
"step": 69000
},
{
"epoch": 29.04,
"eval_loss": 0.27442270517349243,
"eval_runtime": 453.3058,
"eval_samples_per_second": 11.5,
"eval_wer": 0.0956147402816317,
"step": 69000
},
{
"epoch": 29.17,
"learning_rate": 1.1525289629838937e-05,
"loss": 0.0163,
"step": 69300
},
{
"epoch": 29.17,
"eval_loss": 0.27673137187957764,
"eval_runtime": 423.4053,
"eval_samples_per_second": 12.312,
"eval_wer": 0.09512201447057908,
"step": 69300
},
{
"epoch": 29.29,
"learning_rate": 9.779033625317885e-06,
"loss": 0.0177,
"step": 69600
},
{
"epoch": 29.29,
"eval_loss": 0.2782154083251953,
"eval_runtime": 450.2277,
"eval_samples_per_second": 11.579,
"eval_wer": 0.09512201447057908,
"step": 69600
},
{
"epoch": 29.42,
"learning_rate": 8.032777620796835e-06,
"loss": 0.0166,
"step": 69900
},
{
"epoch": 29.42,
"eval_loss": 0.2790413200855255,
"eval_runtime": 446.9628,
"eval_samples_per_second": 11.663,
"eval_wer": 0.09553694146936023,
"step": 69900
},
{
"epoch": 29.55,
"learning_rate": 6.286521616275784e-06,
"loss": 0.0152,
"step": 70200
},
{
"epoch": 29.55,
"eval_loss": 0.2786215543746948,
"eval_runtime": 447.9811,
"eval_samples_per_second": 11.637,
"eval_wer": 0.0953813438448173,
"step": 70200
},
{
"epoch": 29.67,
"learning_rate": 4.540265611754733e-06,
"loss": 0.0157,
"step": 70500
},
{
"epoch": 29.67,
"eval_loss": 0.27765703201293945,
"eval_runtime": 423.4541,
"eval_samples_per_second": 12.311,
"eval_wer": 0.09517388034542673,
"step": 70500
},
{
"epoch": 29.8,
"learning_rate": 2.7940096072336816e-06,
"loss": 0.0175,
"step": 70800
},
{
"epoch": 29.8,
"eval_loss": 0.2788052558898926,
"eval_runtime": 450.7945,
"eval_samples_per_second": 11.564,
"eval_wer": 0.09486268509634087,
"step": 70800
},
{
"epoch": 29.92,
"learning_rate": 1.0477536027126306e-06,
"loss": 0.0166,
"step": 71100
},
{
"epoch": 29.92,
"eval_loss": 0.2789298892021179,
"eval_runtime": 455.4523,
"eval_samples_per_second": 11.446,
"eval_wer": 0.09540727678224112,
"step": 71100
},
{
"epoch": 30.0,
"step": 71280,
"total_flos": 4.535964846237055e+20,
"train_runtime": 470143.8106,
"train_samples_per_second": 0.152
}
],
"max_steps": 71280,
"num_train_epochs": 30,
"total_flos": 4.535964846237055e+20,
"trial_name": null,
"trial_params": null
}