xlsr300m_cv_7.0_nl_lm / trainer_state.json
Iskaj's picture
Upload lm-boosted decoder
e0290ad
{
"best_metric": 0.4362342655658722,
"best_model_checkpoint": "/local_disk0/X-LSR_CV_60_EP/checkpoint-2400",
"epoch": 59.55555555555556,
"global_step": 26800,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.11,
"learning_rate": 2.9999999999999997e-05,
"loss": 11.3191,
"step": 50
},
{
"epoch": 0.22,
"learning_rate": 5.9999999999999995e-05,
"loss": 5.2444,
"step": 100
},
{
"epoch": 0.33,
"learning_rate": 8.999999999999999e-05,
"loss": 3.4938,
"step": 150
},
{
"epoch": 0.44,
"learning_rate": 0.00011999999999999999,
"loss": 3.0165,
"step": 200
},
{
"epoch": 0.56,
"learning_rate": 0.00015,
"loss": 2.9377,
"step": 250
},
{
"epoch": 0.67,
"learning_rate": 0.00017999999999999998,
"loss": 2.9175,
"step": 300
},
{
"epoch": 0.78,
"learning_rate": 0.00020999999999999998,
"loss": 2.8908,
"step": 350
},
{
"epoch": 0.89,
"learning_rate": 0.00023999999999999998,
"loss": 2.8684,
"step": 400
},
{
"epoch": 0.89,
"eval_loss": 2.944436550140381,
"eval_runtime": 232.2253,
"eval_samples_per_second": 24.58,
"eval_steps_per_second": 3.075,
"eval_wer": 0.9999796528781004,
"step": 400
},
{
"epoch": 1.0,
"learning_rate": 0.00027,
"loss": 2.8525,
"step": 450
},
{
"epoch": 1.11,
"learning_rate": 0.0003,
"loss": 2.6961,
"step": 500
},
{
"epoch": 1.22,
"learning_rate": 0.0002994339622641509,
"loss": 1.6145,
"step": 550
},
{
"epoch": 1.33,
"learning_rate": 0.00029886792452830187,
"loss": 1.0359,
"step": 600
},
{
"epoch": 1.44,
"learning_rate": 0.0002983018867924528,
"loss": 0.8086,
"step": 650
},
{
"epoch": 1.56,
"learning_rate": 0.00029773584905660376,
"loss": 0.7171,
"step": 700
},
{
"epoch": 1.67,
"learning_rate": 0.0002971698113207547,
"loss": 0.6152,
"step": 750
},
{
"epoch": 1.78,
"learning_rate": 0.00029660377358490565,
"loss": 0.5631,
"step": 800
},
{
"epoch": 1.78,
"eval_loss": 0.641059398651123,
"eval_runtime": 232.9944,
"eval_samples_per_second": 24.498,
"eval_steps_per_second": 3.064,
"eval_wer": 0.5468289010519461,
"step": 800
},
{
"epoch": 1.89,
"learning_rate": 0.00029603773584905657,
"loss": 0.521,
"step": 850
},
{
"epoch": 2.0,
"learning_rate": 0.00029547169811320755,
"loss": 0.4953,
"step": 900
},
{
"epoch": 2.11,
"learning_rate": 0.00029490566037735847,
"loss": 0.4331,
"step": 950
},
{
"epoch": 2.22,
"learning_rate": 0.0002943396226415094,
"loss": 0.396,
"step": 1000
},
{
"epoch": 2.33,
"learning_rate": 0.00029377358490566036,
"loss": 0.4066,
"step": 1050
},
{
"epoch": 2.44,
"learning_rate": 0.0002932075471698113,
"loss": 0.3839,
"step": 1100
},
{
"epoch": 2.56,
"learning_rate": 0.00029264150943396225,
"loss": 0.3705,
"step": 1150
},
{
"epoch": 2.67,
"learning_rate": 0.00029207547169811317,
"loss": 0.3707,
"step": 1200
},
{
"epoch": 2.67,
"eval_loss": 0.5500322580337524,
"eval_runtime": 232.2885,
"eval_samples_per_second": 24.573,
"eval_steps_per_second": 3.074,
"eval_wer": 0.4608012696604065,
"step": 1200
},
{
"epoch": 2.78,
"learning_rate": 0.00029150943396226414,
"loss": 0.3813,
"step": 1250
},
{
"epoch": 2.89,
"learning_rate": 0.00029094339622641506,
"loss": 0.3572,
"step": 1300
},
{
"epoch": 3.0,
"learning_rate": 0.00029037735849056604,
"loss": 0.3483,
"step": 1350
},
{
"epoch": 3.11,
"learning_rate": 0.00028981132075471696,
"loss": 0.2772,
"step": 1400
},
{
"epoch": 3.22,
"learning_rate": 0.0002892452830188679,
"loss": 0.2977,
"step": 1450
},
{
"epoch": 3.33,
"learning_rate": 0.00028867924528301885,
"loss": 0.2802,
"step": 1500
},
{
"epoch": 3.44,
"learning_rate": 0.00028811320754716977,
"loss": 0.2913,
"step": 1550
},
{
"epoch": 3.56,
"learning_rate": 0.00028754716981132074,
"loss": 0.2899,
"step": 1600
},
{
"epoch": 3.56,
"eval_loss": 0.5040135383605957,
"eval_runtime": 232.8201,
"eval_samples_per_second": 24.517,
"eval_steps_per_second": 3.067,
"eval_wer": 0.420412232689686,
"step": 1600
},
{
"epoch": 3.67,
"learning_rate": 0.00028698113207547166,
"loss": 0.2802,
"step": 1650
},
{
"epoch": 3.78,
"learning_rate": 0.00028641509433962264,
"loss": 0.2769,
"step": 1700
},
{
"epoch": 3.89,
"learning_rate": 0.00028584905660377356,
"loss": 0.2632,
"step": 1750
},
{
"epoch": 4.0,
"learning_rate": 0.00028528301886792453,
"loss": 0.2784,
"step": 1800
},
{
"epoch": 4.11,
"learning_rate": 0.00028471698113207545,
"loss": 0.2341,
"step": 1850
},
{
"epoch": 4.22,
"learning_rate": 0.00028415094339622637,
"loss": 0.2346,
"step": 1900
},
{
"epoch": 4.33,
"learning_rate": 0.00028358490566037734,
"loss": 0.2301,
"step": 1950
},
{
"epoch": 4.44,
"learning_rate": 0.00028301886792452826,
"loss": 0.2376,
"step": 2000
},
{
"epoch": 4.44,
"eval_loss": 0.4637417495250702,
"eval_runtime": 233.0278,
"eval_samples_per_second": 24.495,
"eval_steps_per_second": 3.064,
"eval_wer": 0.3989663662074999,
"step": 2000
},
{
"epoch": 4.56,
"learning_rate": 0.00028245283018867923,
"loss": 0.2474,
"step": 2050
},
{
"epoch": 4.67,
"learning_rate": 0.00028188679245283015,
"loss": 0.2325,
"step": 2100
},
{
"epoch": 4.78,
"learning_rate": 0.00028132075471698113,
"loss": 0.2477,
"step": 2150
},
{
"epoch": 4.89,
"learning_rate": 0.00028075471698113205,
"loss": 0.2319,
"step": 2200
},
{
"epoch": 5.0,
"learning_rate": 0.000280188679245283,
"loss": 0.2342,
"step": 2250
},
{
"epoch": 5.11,
"learning_rate": 0.00027962264150943394,
"loss": 0.2176,
"step": 2300
},
{
"epoch": 5.22,
"learning_rate": 0.00027905660377358486,
"loss": 0.2015,
"step": 2350
},
{
"epoch": 5.33,
"learning_rate": 0.00027849056603773583,
"loss": 0.2063,
"step": 2400
},
{
"epoch": 5.33,
"eval_loss": 0.4362342655658722,
"eval_runtime": 234.4496,
"eval_samples_per_second": 24.346,
"eval_steps_per_second": 3.045,
"eval_wer": 0.3932488249537103,
"step": 2400
},
{
"epoch": 5.44,
"learning_rate": 0.00027792452830188675,
"loss": 0.2116,
"step": 2450
},
{
"epoch": 5.56,
"learning_rate": 0.0002773584905660377,
"loss": 0.2084,
"step": 2500
},
{
"epoch": 5.67,
"learning_rate": 0.00027679245283018865,
"loss": 0.2122,
"step": 2550
},
{
"epoch": 5.78,
"learning_rate": 0.0002762264150943396,
"loss": 0.2063,
"step": 2600
},
{
"epoch": 5.89,
"learning_rate": 0.00027566037735849054,
"loss": 0.1902,
"step": 2650
},
{
"epoch": 6.0,
"learning_rate": 0.0002750943396226415,
"loss": 0.2067,
"step": 2700
},
{
"epoch": 6.11,
"learning_rate": 0.00027452830188679243,
"loss": 0.1845,
"step": 2750
},
{
"epoch": 6.22,
"learning_rate": 0.0002739622641509434,
"loss": 0.1773,
"step": 2800
},
{
"epoch": 6.22,
"eval_loss": 0.4410901367664337,
"eval_runtime": 235.6532,
"eval_samples_per_second": 24.222,
"eval_steps_per_second": 3.03,
"eval_wer": 0.3876737135532179,
"step": 2800
},
{
"epoch": 6.33,
"learning_rate": 0.0002733962264150943,
"loss": 0.1793,
"step": 2850
},
{
"epoch": 6.44,
"learning_rate": 0.00027283018867924524,
"loss": 0.1764,
"step": 2900
},
{
"epoch": 6.56,
"learning_rate": 0.0002722641509433962,
"loss": 0.1882,
"step": 2950
},
{
"epoch": 6.67,
"learning_rate": 0.00027169811320754714,
"loss": 0.1892,
"step": 3000
},
{
"epoch": 6.78,
"learning_rate": 0.0002711320754716981,
"loss": 0.1941,
"step": 3050
},
{
"epoch": 6.89,
"learning_rate": 0.00027056603773584903,
"loss": 0.1858,
"step": 3100
},
{
"epoch": 7.0,
"learning_rate": 0.00027,
"loss": 0.1817,
"step": 3150
},
{
"epoch": 7.11,
"learning_rate": 0.0002694339622641509,
"loss": 0.1676,
"step": 3200
},
{
"epoch": 7.11,
"eval_loss": 0.479378342628479,
"eval_runtime": 234.1851,
"eval_samples_per_second": 24.374,
"eval_steps_per_second": 3.049,
"eval_wer": 0.38449956253687917,
"step": 3200
},
{
"epoch": 7.22,
"learning_rate": 0.0002688679245283019,
"loss": 0.1728,
"step": 3250
},
{
"epoch": 7.33,
"learning_rate": 0.0002683018867924528,
"loss": 0.1597,
"step": 3300
},
{
"epoch": 7.44,
"learning_rate": 0.00026773584905660374,
"loss": 0.1695,
"step": 3350
},
{
"epoch": 7.56,
"learning_rate": 0.0002671698113207547,
"loss": 0.1662,
"step": 3400
},
{
"epoch": 7.67,
"learning_rate": 0.00026660377358490563,
"loss": 0.1612,
"step": 3450
},
{
"epoch": 7.78,
"learning_rate": 0.0002660377358490566,
"loss": 0.1611,
"step": 3500
},
{
"epoch": 7.89,
"learning_rate": 0.0002654716981132075,
"loss": 0.1669,
"step": 3550
},
{
"epoch": 8.0,
"learning_rate": 0.0002649056603773585,
"loss": 0.1728,
"step": 3600
},
{
"epoch": 8.0,
"eval_loss": 0.4429037272930145,
"eval_runtime": 233.3956,
"eval_samples_per_second": 24.456,
"eval_steps_per_second": 3.059,
"eval_wer": 0.37750015260341424,
"step": 3600
},
{
"epoch": 8.11,
"learning_rate": 0.0002643396226415094,
"loss": 0.1543,
"step": 3650
},
{
"epoch": 8.22,
"learning_rate": 0.0002637735849056604,
"loss": 0.152,
"step": 3700
},
{
"epoch": 8.33,
"learning_rate": 0.0002632075471698113,
"loss": 0.1552,
"step": 3750
},
{
"epoch": 8.44,
"learning_rate": 0.0002626415094339622,
"loss": 0.16,
"step": 3800
},
{
"epoch": 8.56,
"learning_rate": 0.0002620754716981132,
"loss": 0.1645,
"step": 3850
},
{
"epoch": 8.67,
"learning_rate": 0.0002615094339622641,
"loss": 0.158,
"step": 3900
},
{
"epoch": 8.78,
"learning_rate": 0.0002609433962264151,
"loss": 0.1654,
"step": 3950
},
{
"epoch": 8.89,
"learning_rate": 0.000260377358490566,
"loss": 0.1556,
"step": 4000
},
{
"epoch": 8.89,
"eval_loss": 0.4799572825431824,
"eval_runtime": 233.7994,
"eval_samples_per_second": 24.414,
"eval_steps_per_second": 3.054,
"eval_wer": 0.38354324780759763,
"step": 4000
},
{
"epoch": 9.0,
"learning_rate": 0.000259811320754717,
"loss": 0.1627,
"step": 4050
},
{
"epoch": 9.11,
"learning_rate": 0.0002592452830188679,
"loss": 0.1443,
"step": 4100
},
{
"epoch": 9.22,
"learning_rate": 0.0002586792452830189,
"loss": 0.1527,
"step": 4150
},
{
"epoch": 9.33,
"learning_rate": 0.0002581132075471698,
"loss": 0.1516,
"step": 4200
},
{
"epoch": 9.44,
"learning_rate": 0.0002575471698113207,
"loss": 0.1554,
"step": 4250
},
{
"epoch": 9.56,
"learning_rate": 0.0002569811320754717,
"loss": 0.1502,
"step": 4300
},
{
"epoch": 9.67,
"learning_rate": 0.0002564150943396226,
"loss": 0.1496,
"step": 4350
},
{
"epoch": 9.78,
"learning_rate": 0.0002558490566037736,
"loss": 0.1514,
"step": 4400
},
{
"epoch": 9.78,
"eval_loss": 0.46985962986946106,
"eval_runtime": 235.3127,
"eval_samples_per_second": 24.257,
"eval_steps_per_second": 3.034,
"eval_wer": 0.3755264817791523,
"step": 4400
},
{
"epoch": 9.89,
"learning_rate": 0.0002552830188679245,
"loss": 0.1501,
"step": 4450
},
{
"epoch": 10.0,
"learning_rate": 0.0002547169811320755,
"loss": 0.1464,
"step": 4500
},
{
"epoch": 10.11,
"learning_rate": 0.0002541509433962264,
"loss": 0.1354,
"step": 4550
},
{
"epoch": 10.22,
"learning_rate": 0.00025358490566037737,
"loss": 0.1315,
"step": 4600
},
{
"epoch": 10.33,
"learning_rate": 0.0002530188679245283,
"loss": 0.1365,
"step": 4650
},
{
"epoch": 10.44,
"learning_rate": 0.0002524528301886792,
"loss": 0.1399,
"step": 4700
},
{
"epoch": 10.56,
"learning_rate": 0.0002518867924528302,
"loss": 0.1351,
"step": 4750
},
{
"epoch": 10.67,
"learning_rate": 0.0002513207547169811,
"loss": 0.1405,
"step": 4800
},
{
"epoch": 10.67,
"eval_loss": 0.47201868891716003,
"eval_runtime": 234.4469,
"eval_samples_per_second": 24.347,
"eval_steps_per_second": 3.045,
"eval_wer": 0.3793517406962785,
"step": 4800
},
{
"epoch": 10.78,
"learning_rate": 0.0002507547169811321,
"loss": 0.1446,
"step": 4850
},
{
"epoch": 10.89,
"learning_rate": 0.000250188679245283,
"loss": 0.1402,
"step": 4900
},
{
"epoch": 11.0,
"learning_rate": 0.00024962264150943397,
"loss": 0.1417,
"step": 4950
},
{
"epoch": 11.11,
"learning_rate": 0.0002490566037735849,
"loss": 0.1326,
"step": 5000
},
{
"epoch": 11.22,
"learning_rate": 0.00024849056603773586,
"loss": 0.1409,
"step": 5050
},
{
"epoch": 11.33,
"learning_rate": 0.0002479245283018868,
"loss": 0.1279,
"step": 5100
},
{
"epoch": 11.44,
"learning_rate": 0.0002473584905660377,
"loss": 0.1333,
"step": 5150
},
{
"epoch": 11.56,
"learning_rate": 0.0002467924528301887,
"loss": 0.1317,
"step": 5200
},
{
"epoch": 11.56,
"eval_loss": 0.5062007904052734,
"eval_runtime": 233.2872,
"eval_samples_per_second": 24.468,
"eval_steps_per_second": 3.061,
"eval_wer": 0.3787006327954911,
"step": 5200
},
{
"epoch": 11.67,
"learning_rate": 0.0002462264150943396,
"loss": 0.1309,
"step": 5250
},
{
"epoch": 11.78,
"learning_rate": 0.00024566037735849057,
"loss": 0.1344,
"step": 5300
},
{
"epoch": 11.89,
"learning_rate": 0.0002450943396226415,
"loss": 0.1339,
"step": 5350
},
{
"epoch": 12.0,
"learning_rate": 0.00024452830188679246,
"loss": 0.1403,
"step": 5400
},
{
"epoch": 12.11,
"learning_rate": 0.00024396226415094338,
"loss": 0.1208,
"step": 5450
},
{
"epoch": 12.22,
"learning_rate": 0.00024339622641509433,
"loss": 0.1278,
"step": 5500
},
{
"epoch": 12.33,
"learning_rate": 0.00024283018867924527,
"loss": 0.1148,
"step": 5550
},
{
"epoch": 12.44,
"learning_rate": 0.00024226415094339622,
"loss": 0.1204,
"step": 5600
},
{
"epoch": 12.44,
"eval_loss": 0.4777355492115021,
"eval_runtime": 233.4682,
"eval_samples_per_second": 24.449,
"eval_steps_per_second": 3.058,
"eval_wer": 0.36860846033328587,
"step": 5600
},
{
"epoch": 12.56,
"learning_rate": 0.00024169811320754717,
"loss": 0.1237,
"step": 5650
},
{
"epoch": 12.67,
"learning_rate": 0.0002411320754716981,
"loss": 0.124,
"step": 5700
},
{
"epoch": 12.78,
"learning_rate": 0.00024056603773584906,
"loss": 0.1358,
"step": 5750
},
{
"epoch": 12.89,
"learning_rate": 0.00023999999999999998,
"loss": 0.1243,
"step": 5800
},
{
"epoch": 13.0,
"learning_rate": 0.00023943396226415093,
"loss": 0.1255,
"step": 5850
},
{
"epoch": 13.11,
"learning_rate": 0.00023886792452830187,
"loss": 0.1158,
"step": 5900
},
{
"epoch": 13.22,
"learning_rate": 0.00023830188679245282,
"loss": 0.1135,
"step": 5950
},
{
"epoch": 13.33,
"learning_rate": 0.00023773584905660377,
"loss": 0.12,
"step": 6000
},
{
"epoch": 13.33,
"eval_loss": 0.5171375870704651,
"eval_runtime": 233.056,
"eval_samples_per_second": 24.492,
"eval_steps_per_second": 3.064,
"eval_wer": 0.3718233055934238,
"step": 6000
},
{
"epoch": 13.44,
"learning_rate": 0.0002371698113207547,
"loss": 0.1218,
"step": 6050
},
{
"epoch": 13.56,
"learning_rate": 0.00023660377358490566,
"loss": 0.1244,
"step": 6100
},
{
"epoch": 13.67,
"learning_rate": 0.0002360377358490566,
"loss": 0.1238,
"step": 6150
},
{
"epoch": 13.78,
"learning_rate": 0.00023547169811320755,
"loss": 0.1188,
"step": 6200
},
{
"epoch": 13.89,
"learning_rate": 0.0002349056603773585,
"loss": 0.1167,
"step": 6250
},
{
"epoch": 14.0,
"learning_rate": 0.00023433962264150942,
"loss": 0.1212,
"step": 6300
},
{
"epoch": 14.11,
"learning_rate": 0.00023377358490566036,
"loss": 0.1209,
"step": 6350
},
{
"epoch": 14.22,
"learning_rate": 0.0002332075471698113,
"loss": 0.1176,
"step": 6400
},
{
"epoch": 14.22,
"eval_loss": 0.5209046602249146,
"eval_runtime": 232.7218,
"eval_samples_per_second": 24.527,
"eval_steps_per_second": 3.068,
"eval_wer": 0.37357315807679004,
"step": 6400
},
{
"epoch": 14.33,
"learning_rate": 0.00023264150943396226,
"loss": 0.1108,
"step": 6450
},
{
"epoch": 14.44,
"learning_rate": 0.0002320754716981132,
"loss": 0.1201,
"step": 6500
},
{
"epoch": 14.56,
"learning_rate": 0.00023150943396226415,
"loss": 0.1135,
"step": 6550
},
{
"epoch": 14.67,
"learning_rate": 0.0002309433962264151,
"loss": 0.111,
"step": 6600
},
{
"epoch": 14.78,
"learning_rate": 0.00023037735849056604,
"loss": 0.1169,
"step": 6650
},
{
"epoch": 14.89,
"learning_rate": 0.000229811320754717,
"loss": 0.12,
"step": 6700
},
{
"epoch": 15.0,
"learning_rate": 0.0002292452830188679,
"loss": 0.1145,
"step": 6750
},
{
"epoch": 15.11,
"learning_rate": 0.00022867924528301886,
"loss": 0.1102,
"step": 6800
},
{
"epoch": 15.11,
"eval_loss": 0.5242590308189392,
"eval_runtime": 233.3294,
"eval_samples_per_second": 24.463,
"eval_steps_per_second": 3.06,
"eval_wer": 0.37780535943190835,
"step": 6800
},
{
"epoch": 15.22,
"learning_rate": 0.0002281132075471698,
"loss": 0.0994,
"step": 6850
},
{
"epoch": 15.33,
"learning_rate": 0.00022754716981132075,
"loss": 0.1196,
"step": 6900
},
{
"epoch": 15.44,
"learning_rate": 0.0002269811320754717,
"loss": 0.1125,
"step": 6950
},
{
"epoch": 15.56,
"learning_rate": 0.00022641509433962264,
"loss": 0.1122,
"step": 7000
},
{
"epoch": 15.67,
"learning_rate": 0.0002258490566037736,
"loss": 0.1235,
"step": 7050
},
{
"epoch": 15.78,
"learning_rate": 0.00022528301886792453,
"loss": 0.1088,
"step": 7100
},
{
"epoch": 15.89,
"learning_rate": 0.00022471698113207543,
"loss": 0.1089,
"step": 7150
},
{
"epoch": 16.0,
"learning_rate": 0.00022415094339622637,
"loss": 0.1097,
"step": 7200
},
{
"epoch": 16.0,
"eval_loss": 0.4982779622077942,
"eval_runtime": 233.2014,
"eval_samples_per_second": 24.477,
"eval_steps_per_second": 3.062,
"eval_wer": 0.3620159928378131,
"step": 7200
},
{
"epoch": 16.11,
"learning_rate": 0.00022358490566037732,
"loss": 0.1023,
"step": 7250
},
{
"epoch": 16.22,
"learning_rate": 0.00022301886792452827,
"loss": 0.097,
"step": 7300
},
{
"epoch": 16.33,
"learning_rate": 0.0002224528301886792,
"loss": 0.1002,
"step": 7350
},
{
"epoch": 16.44,
"learning_rate": 0.00022188679245283016,
"loss": 0.1051,
"step": 7400
},
{
"epoch": 16.56,
"learning_rate": 0.0002213207547169811,
"loss": 0.1114,
"step": 7450
},
{
"epoch": 16.67,
"learning_rate": 0.00022075471698113205,
"loss": 0.0984,
"step": 7500
},
{
"epoch": 16.78,
"learning_rate": 0.00022018867924528297,
"loss": 0.1121,
"step": 7550
},
{
"epoch": 16.89,
"learning_rate": 0.00021962264150943392,
"loss": 0.1091,
"step": 7600
},
{
"epoch": 16.89,
"eval_loss": 0.4976480007171631,
"eval_runtime": 232.5964,
"eval_samples_per_second": 24.54,
"eval_steps_per_second": 3.07,
"eval_wer": 0.366166805705333,
"step": 7600
},
{
"epoch": 17.0,
"learning_rate": 0.00021905660377358486,
"loss": 0.1061,
"step": 7650
},
{
"epoch": 17.11,
"learning_rate": 0.0002184905660377358,
"loss": 0.0962,
"step": 7700
},
{
"epoch": 17.22,
"learning_rate": 0.00021792452830188676,
"loss": 0.101,
"step": 7750
},
{
"epoch": 17.33,
"learning_rate": 0.0002173584905660377,
"loss": 0.1109,
"step": 7800
},
{
"epoch": 17.44,
"learning_rate": 0.00021679245283018865,
"loss": 0.1021,
"step": 7850
},
{
"epoch": 17.56,
"learning_rate": 0.0002162264150943396,
"loss": 0.1011,
"step": 7900
},
{
"epoch": 17.67,
"learning_rate": 0.00021566037735849054,
"loss": 0.0992,
"step": 7950
},
{
"epoch": 17.78,
"learning_rate": 0.00021509433962264146,
"loss": 0.104,
"step": 8000
},
{
"epoch": 17.78,
"eval_loss": 0.5483397245407104,
"eval_runtime": 232.4366,
"eval_samples_per_second": 24.557,
"eval_steps_per_second": 3.072,
"eval_wer": 0.3651901438541518,
"step": 8000
},
{
"epoch": 17.89,
"learning_rate": 0.0002145283018867924,
"loss": 0.1047,
"step": 8050
},
{
"epoch": 18.0,
"learning_rate": 0.00021396226415094336,
"loss": 0.1026,
"step": 8100
},
{
"epoch": 18.11,
"learning_rate": 0.0002133962264150943,
"loss": 0.0929,
"step": 8150
},
{
"epoch": 18.22,
"learning_rate": 0.00021283018867924525,
"loss": 0.0933,
"step": 8200
},
{
"epoch": 18.33,
"learning_rate": 0.0002122641509433962,
"loss": 0.0971,
"step": 8250
},
{
"epoch": 18.44,
"learning_rate": 0.00021169811320754714,
"loss": 0.097,
"step": 8300
},
{
"epoch": 18.56,
"learning_rate": 0.0002111320754716981,
"loss": 0.097,
"step": 8350
},
{
"epoch": 18.67,
"learning_rate": 0.00021056603773584904,
"loss": 0.1014,
"step": 8400
},
{
"epoch": 18.67,
"eval_loss": 0.5109674334526062,
"eval_runtime": 233.1396,
"eval_samples_per_second": 24.483,
"eval_steps_per_second": 3.063,
"eval_wer": 0.3620159928378131,
"step": 8400
},
{
"epoch": 18.78,
"learning_rate": 0.00020999999999999998,
"loss": 0.0936,
"step": 8450
},
{
"epoch": 18.89,
"learning_rate": 0.0002094339622641509,
"loss": 0.0948,
"step": 8500
},
{
"epoch": 19.0,
"learning_rate": 0.00020886792452830185,
"loss": 0.0928,
"step": 8550
},
{
"epoch": 19.11,
"learning_rate": 0.0002083018867924528,
"loss": 0.0912,
"step": 8600
},
{
"epoch": 19.22,
"learning_rate": 0.00020773584905660374,
"loss": 0.0918,
"step": 8650
},
{
"epoch": 19.33,
"learning_rate": 0.0002071698113207547,
"loss": 0.0902,
"step": 8700
},
{
"epoch": 19.44,
"learning_rate": 0.00020660377358490563,
"loss": 0.0961,
"step": 8750
},
{
"epoch": 19.56,
"learning_rate": 0.00020603773584905658,
"loss": 0.0921,
"step": 8800
},
{
"epoch": 19.56,
"eval_loss": 0.4945477545261383,
"eval_runtime": 232.6384,
"eval_samples_per_second": 24.536,
"eval_steps_per_second": 3.069,
"eval_wer": 0.3609375953771339,
"step": 8800
},
{
"epoch": 19.67,
"learning_rate": 0.00020547169811320753,
"loss": 0.0957,
"step": 8850
},
{
"epoch": 19.78,
"learning_rate": 0.00020491698113207546,
"loss": 0.0958,
"step": 8900
},
{
"epoch": 19.89,
"learning_rate": 0.0002043509433962264,
"loss": 0.0952,
"step": 8950
},
{
"epoch": 20.0,
"learning_rate": 0.00020378490566037735,
"loss": 0.0847,
"step": 9000
},
{
"epoch": 20.11,
"learning_rate": 0.00020321886792452827,
"loss": 0.0921,
"step": 9050
},
{
"epoch": 20.22,
"learning_rate": 0.00020265283018867922,
"loss": 0.0885,
"step": 9100
},
{
"epoch": 20.33,
"learning_rate": 0.00020208679245283017,
"loss": 0.0881,
"step": 9150
},
{
"epoch": 20.44,
"learning_rate": 0.0002015207547169811,
"loss": 0.0943,
"step": 9200
},
{
"epoch": 20.44,
"eval_loss": 0.5395381450653076,
"eval_runtime": 232.8056,
"eval_samples_per_second": 24.518,
"eval_steps_per_second": 3.067,
"eval_wer": 0.36470181292856124,
"step": 9200
},
{
"epoch": 20.56,
"learning_rate": 0.00020095471698113206,
"loss": 0.0888,
"step": 9250
},
{
"epoch": 20.67,
"learning_rate": 0.000200388679245283,
"loss": 0.0894,
"step": 9300
},
{
"epoch": 20.78,
"learning_rate": 0.00019982264150943395,
"loss": 0.0883,
"step": 9350
},
{
"epoch": 20.89,
"learning_rate": 0.0001992566037735849,
"loss": 0.0873,
"step": 9400
},
{
"epoch": 21.0,
"learning_rate": 0.00019869056603773584,
"loss": 0.095,
"step": 9450
},
{
"epoch": 21.11,
"learning_rate": 0.00019812452830188676,
"loss": 0.087,
"step": 9500
},
{
"epoch": 21.22,
"learning_rate": 0.0001975584905660377,
"loss": 0.0795,
"step": 9550
},
{
"epoch": 21.33,
"learning_rate": 0.00019699245283018866,
"loss": 0.0877,
"step": 9600
},
{
"epoch": 21.33,
"eval_loss": 0.516017735004425,
"eval_runtime": 232.5407,
"eval_samples_per_second": 24.546,
"eval_steps_per_second": 3.07,
"eval_wer": 0.37127393330213443,
"step": 9600
},
{
"epoch": 21.44,
"learning_rate": 0.0001964264150943396,
"loss": 0.0902,
"step": 9650
},
{
"epoch": 21.56,
"learning_rate": 0.00019586037735849055,
"loss": 0.0898,
"step": 9700
},
{
"epoch": 21.67,
"learning_rate": 0.0001952943396226415,
"loss": 0.0934,
"step": 9750
},
{
"epoch": 21.78,
"learning_rate": 0.00019472830188679244,
"loss": 0.0845,
"step": 9800
},
{
"epoch": 21.89,
"learning_rate": 0.0001941622641509434,
"loss": 0.0916,
"step": 9850
},
{
"epoch": 22.0,
"learning_rate": 0.00019359622641509434,
"loss": 0.0878,
"step": 9900
},
{
"epoch": 22.11,
"learning_rate": 0.00019303018867924526,
"loss": 0.0864,
"step": 9950
},
{
"epoch": 22.22,
"learning_rate": 0.0001924641509433962,
"loss": 0.0768,
"step": 10000
},
{
"epoch": 22.22,
"eval_loss": 0.5633887052536011,
"eval_runtime": 233.9063,
"eval_samples_per_second": 24.403,
"eval_steps_per_second": 3.053,
"eval_wer": 0.36429487049056913,
"step": 10000
},
{
"epoch": 22.33,
"learning_rate": 0.00019189811320754715,
"loss": 0.0888,
"step": 10050
},
{
"epoch": 22.44,
"learning_rate": 0.0001913320754716981,
"loss": 0.0876,
"step": 10100
},
{
"epoch": 22.56,
"learning_rate": 0.00019076603773584904,
"loss": 0.0835,
"step": 10150
},
{
"epoch": 22.67,
"learning_rate": 0.0001902,
"loss": 0.0766,
"step": 10200
},
{
"epoch": 22.78,
"learning_rate": 0.00018963396226415093,
"loss": 0.0909,
"step": 10250
},
{
"epoch": 22.89,
"learning_rate": 0.00018906792452830188,
"loss": 0.08,
"step": 10300
},
{
"epoch": 23.0,
"learning_rate": 0.00018850188679245283,
"loss": 0.081,
"step": 10350
},
{
"epoch": 23.11,
"learning_rate": 0.00018793584905660375,
"loss": 0.0744,
"step": 10400
},
{
"epoch": 23.11,
"eval_loss": 0.5204855799674988,
"eval_runtime": 235.6787,
"eval_samples_per_second": 24.219,
"eval_steps_per_second": 3.03,
"eval_wer": 0.3642745233686695,
"step": 10400
},
{
"epoch": 23.22,
"learning_rate": 0.0001873698113207547,
"loss": 0.0817,
"step": 10450
},
{
"epoch": 23.33,
"learning_rate": 0.00018680377358490564,
"loss": 0.0791,
"step": 10500
},
{
"epoch": 23.44,
"learning_rate": 0.0001862377358490566,
"loss": 0.0779,
"step": 10550
},
{
"epoch": 23.56,
"learning_rate": 0.00018567169811320753,
"loss": 0.0818,
"step": 10600
},
{
"epoch": 23.67,
"learning_rate": 0.00018510566037735848,
"loss": 0.0801,
"step": 10650
},
{
"epoch": 23.78,
"learning_rate": 0.00018453962264150943,
"loss": 0.0843,
"step": 10700
},
{
"epoch": 23.89,
"learning_rate": 0.00018397358490566037,
"loss": 0.0832,
"step": 10750
},
{
"epoch": 24.0,
"learning_rate": 0.00018340754716981132,
"loss": 0.0852,
"step": 10800
},
{
"epoch": 24.0,
"eval_loss": 0.5427780151367188,
"eval_runtime": 234.1837,
"eval_samples_per_second": 24.374,
"eval_steps_per_second": 3.049,
"eval_wer": 0.36130384357132683,
"step": 10800
},
{
"epoch": 24.11,
"learning_rate": 0.00018284150943396224,
"loss": 0.0779,
"step": 10850
},
{
"epoch": 24.22,
"learning_rate": 0.00018227547169811319,
"loss": 0.0734,
"step": 10900
},
{
"epoch": 24.33,
"learning_rate": 0.00018170943396226413,
"loss": 0.0843,
"step": 10950
},
{
"epoch": 24.44,
"learning_rate": 0.00018114339622641508,
"loss": 0.0777,
"step": 11000
},
{
"epoch": 24.56,
"learning_rate": 0.00018057735849056602,
"loss": 0.0782,
"step": 11050
},
{
"epoch": 24.67,
"learning_rate": 0.00018001132075471697,
"loss": 0.0783,
"step": 11100
},
{
"epoch": 24.78,
"learning_rate": 0.00017944528301886792,
"loss": 0.076,
"step": 11150
},
{
"epoch": 24.89,
"learning_rate": 0.00017887924528301886,
"loss": 0.0732,
"step": 11200
},
{
"epoch": 24.89,
"eval_loss": 0.551848292350769,
"eval_runtime": 232.8874,
"eval_samples_per_second": 24.51,
"eval_steps_per_second": 3.066,
"eval_wer": 0.3628298777137974,
"step": 11200
},
{
"epoch": 25.0,
"learning_rate": 0.0001783245283018868,
"loss": 0.0861,
"step": 11250
},
{
"epoch": 25.11,
"learning_rate": 0.0001777698113207547,
"loss": 0.0812,
"step": 11300
},
{
"epoch": 25.22,
"learning_rate": 0.00017720377358490565,
"loss": 0.0765,
"step": 11350
},
{
"epoch": 25.33,
"learning_rate": 0.00017663773584905657,
"loss": 0.0775,
"step": 11400
},
{
"epoch": 25.44,
"learning_rate": 0.00017607169811320752,
"loss": 0.0762,
"step": 11450
},
{
"epoch": 25.56,
"learning_rate": 0.00017550566037735846,
"loss": 0.0782,
"step": 11500
},
{
"epoch": 25.67,
"learning_rate": 0.0001749396226415094,
"loss": 0.077,
"step": 11550
},
{
"epoch": 25.78,
"learning_rate": 0.00017437358490566036,
"loss": 0.0725,
"step": 11600
},
{
"epoch": 25.78,
"eval_loss": 0.5757771134376526,
"eval_runtime": 233.1827,
"eval_samples_per_second": 24.479,
"eval_steps_per_second": 3.062,
"eval_wer": 0.37282031456650455,
"step": 11600
},
{
"epoch": 25.89,
"learning_rate": 0.0001738075471698113,
"loss": 0.08,
"step": 11650
},
{
"epoch": 26.0,
"learning_rate": 0.00017324150943396225,
"loss": 0.0743,
"step": 11700
},
{
"epoch": 26.11,
"learning_rate": 0.0001726754716981132,
"loss": 0.0722,
"step": 11750
},
{
"epoch": 26.22,
"learning_rate": 0.00017210943396226414,
"loss": 0.0725,
"step": 11800
},
{
"epoch": 26.33,
"learning_rate": 0.00017154339622641506,
"loss": 0.0708,
"step": 11850
},
{
"epoch": 26.44,
"learning_rate": 0.000170977358490566,
"loss": 0.081,
"step": 11900
},
{
"epoch": 26.56,
"learning_rate": 0.00017041132075471695,
"loss": 0.0704,
"step": 11950
},
{
"epoch": 26.67,
"learning_rate": 0.0001698452830188679,
"loss": 0.0691,
"step": 12000
},
{
"epoch": 26.67,
"eval_loss": 0.5725019574165344,
"eval_runtime": 232.5631,
"eval_samples_per_second": 24.544,
"eval_steps_per_second": 3.07,
"eval_wer": 0.3584552465053818,
"step": 12000
},
{
"epoch": 26.78,
"learning_rate": 0.00016927924528301885,
"loss": 0.0766,
"step": 12050
},
{
"epoch": 26.89,
"learning_rate": 0.0001687132075471698,
"loss": 0.0767,
"step": 12100
},
{
"epoch": 27.0,
"learning_rate": 0.00016814716981132074,
"loss": 0.0748,
"step": 12150
},
{
"epoch": 27.11,
"learning_rate": 0.0001675811320754717,
"loss": 0.0645,
"step": 12200
},
{
"epoch": 27.22,
"learning_rate": 0.00016701509433962263,
"loss": 0.0701,
"step": 12250
},
{
"epoch": 27.33,
"learning_rate": 0.00016644905660377358,
"loss": 0.0702,
"step": 12300
},
{
"epoch": 27.44,
"learning_rate": 0.0001658830188679245,
"loss": 0.0711,
"step": 12350
},
{
"epoch": 27.56,
"learning_rate": 0.00016531698113207545,
"loss": 0.0664,
"step": 12400
},
{
"epoch": 27.56,
"eval_loss": 0.5794127583503723,
"eval_runtime": 233.0106,
"eval_samples_per_second": 24.497,
"eval_steps_per_second": 3.064,
"eval_wer": 0.35994058640405313,
"step": 12400
},
{
"epoch": 27.67,
"learning_rate": 0.0001647509433962264,
"loss": 0.0692,
"step": 12450
},
{
"epoch": 27.78,
"learning_rate": 0.00016418490566037734,
"loss": 0.0705,
"step": 12500
},
{
"epoch": 27.89,
"learning_rate": 0.00016361886792452829,
"loss": 0.0705,
"step": 12550
},
{
"epoch": 28.0,
"learning_rate": 0.00016305283018867923,
"loss": 0.0655,
"step": 12600
},
{
"epoch": 28.11,
"learning_rate": 0.00016248679245283018,
"loss": 0.0723,
"step": 12650
},
{
"epoch": 28.22,
"learning_rate": 0.00016192075471698113,
"loss": 0.0716,
"step": 12700
},
{
"epoch": 28.33,
"learning_rate": 0.00016135471698113207,
"loss": 0.0709,
"step": 12750
},
{
"epoch": 28.44,
"learning_rate": 0.000160788679245283,
"loss": 0.0663,
"step": 12800
},
{
"epoch": 28.44,
"eval_loss": 0.5777014493942261,
"eval_runtime": 235.2144,
"eval_samples_per_second": 24.267,
"eval_steps_per_second": 3.036,
"eval_wer": 0.35607463324312777,
"step": 12800
},
{
"epoch": 28.56,
"learning_rate": 0.00016022264150943394,
"loss": 0.0687,
"step": 12850
},
{
"epoch": 28.67,
"learning_rate": 0.00015965660377358488,
"loss": 0.0736,
"step": 12900
},
{
"epoch": 28.78,
"learning_rate": 0.00015909056603773583,
"loss": 0.0675,
"step": 12950
},
{
"epoch": 28.89,
"learning_rate": 0.00015852452830188678,
"loss": 0.0695,
"step": 13000
},
{
"epoch": 29.0,
"learning_rate": 0.00015795849056603772,
"loss": 0.0719,
"step": 13050
},
{
"epoch": 29.11,
"learning_rate": 0.00015739245283018867,
"loss": 0.0675,
"step": 13100
},
{
"epoch": 29.22,
"learning_rate": 0.00015682641509433962,
"loss": 0.0692,
"step": 13150
},
{
"epoch": 29.33,
"learning_rate": 0.00015626037735849056,
"loss": 0.0671,
"step": 13200
},
{
"epoch": 29.33,
"eval_loss": 0.573063850402832,
"eval_runtime": 235.4966,
"eval_samples_per_second": 24.238,
"eval_steps_per_second": 3.032,
"eval_wer": 0.35485380592915133,
"step": 13200
},
{
"epoch": 29.44,
"learning_rate": 0.00015569433962264148,
"loss": 0.0648,
"step": 13250
},
{
"epoch": 29.56,
"learning_rate": 0.00015512830188679243,
"loss": 0.0768,
"step": 13300
},
{
"epoch": 29.67,
"learning_rate": 0.00015456226415094338,
"loss": 0.0627,
"step": 13350
},
{
"epoch": 29.78,
"learning_rate": 0.00015399622641509432,
"loss": 0.0657,
"step": 13400
},
{
"epoch": 29.89,
"learning_rate": 0.00015343018867924527,
"loss": 0.0698,
"step": 13450
},
{
"epoch": 30.0,
"learning_rate": 0.00015286415094339622,
"loss": 0.0683,
"step": 13500
},
{
"epoch": 30.11,
"learning_rate": 0.00015229811320754716,
"loss": 0.0616,
"step": 13550
},
{
"epoch": 30.22,
"learning_rate": 0.0001517320754716981,
"loss": 0.0649,
"step": 13600
},
{
"epoch": 30.22,
"eval_loss": 0.5660180449485779,
"eval_runtime": 233.6416,
"eval_samples_per_second": 24.431,
"eval_steps_per_second": 3.056,
"eval_wer": 0.3600423220135512,
"step": 13600
},
{
"epoch": 30.33,
"learning_rate": 0.00015116603773584905,
"loss": 0.0674,
"step": 13650
},
{
"epoch": 30.44,
"learning_rate": 0.00015059999999999997,
"loss": 0.0645,
"step": 13700
},
{
"epoch": 30.56,
"learning_rate": 0.00015003396226415092,
"loss": 0.0643,
"step": 13750
},
{
"epoch": 30.67,
"learning_rate": 0.00014946792452830187,
"loss": 0.0631,
"step": 13800
},
{
"epoch": 30.78,
"learning_rate": 0.00014890188679245281,
"loss": 0.064,
"step": 13850
},
{
"epoch": 30.89,
"learning_rate": 0.00014833584905660376,
"loss": 0.0666,
"step": 13900
},
{
"epoch": 31.0,
"learning_rate": 0.0001477698113207547,
"loss": 0.0656,
"step": 13950
},
{
"epoch": 31.11,
"learning_rate": 0.00014720377358490565,
"loss": 0.0614,
"step": 14000
},
{
"epoch": 31.11,
"eval_loss": 0.5769771933555603,
"eval_runtime": 234.4224,
"eval_samples_per_second": 24.349,
"eval_steps_per_second": 3.046,
"eval_wer": 0.35595255051173014,
"step": 14000
},
{
"epoch": 31.22,
"learning_rate": 0.0001466377358490566,
"loss": 0.0609,
"step": 14050
},
{
"epoch": 31.33,
"learning_rate": 0.00014607169811320755,
"loss": 0.0615,
"step": 14100
},
{
"epoch": 31.44,
"learning_rate": 0.0001455056603773585,
"loss": 0.0595,
"step": 14150
},
{
"epoch": 31.56,
"learning_rate": 0.0001449396226415094,
"loss": 0.0535,
"step": 14200
},
{
"epoch": 31.67,
"learning_rate": 0.00014437358490566036,
"loss": 0.0662,
"step": 14250
},
{
"epoch": 31.78,
"learning_rate": 0.0001438075471698113,
"loss": 0.0636,
"step": 14300
},
{
"epoch": 31.89,
"learning_rate": 0.00014324150943396225,
"loss": 0.065,
"step": 14350
},
{
"epoch": 32.0,
"learning_rate": 0.0001426754716981132,
"loss": 0.0605,
"step": 14400
},
{
"epoch": 32.0,
"eval_loss": 0.5667794942855835,
"eval_runtime": 236.4511,
"eval_samples_per_second": 24.14,
"eval_steps_per_second": 3.02,
"eval_wer": 0.35633914582782267,
"step": 14400
},
{
"epoch": 32.11,
"learning_rate": 0.00014210943396226414,
"loss": 0.0576,
"step": 14450
},
{
"epoch": 32.22,
"learning_rate": 0.0001415433962264151,
"loss": 0.0576,
"step": 14500
},
{
"epoch": 32.33,
"learning_rate": 0.00014097735849056604,
"loss": 0.0567,
"step": 14550
},
{
"epoch": 32.44,
"learning_rate": 0.00014041132075471698,
"loss": 0.0638,
"step": 14600
},
{
"epoch": 32.56,
"learning_rate": 0.0001398452830188679,
"loss": 0.0563,
"step": 14650
},
{
"epoch": 32.67,
"learning_rate": 0.00013927924528301885,
"loss": 0.0606,
"step": 14700
},
{
"epoch": 32.78,
"learning_rate": 0.0001387132075471698,
"loss": 0.066,
"step": 14750
},
{
"epoch": 32.89,
"learning_rate": 0.00013814716981132074,
"loss": 0.0594,
"step": 14800
},
{
"epoch": 32.89,
"eval_loss": 0.5590910911560059,
"eval_runtime": 236.6548,
"eval_samples_per_second": 24.12,
"eval_steps_per_second": 3.017,
"eval_wer": 0.3464707917065131,
"step": 14800
},
{
"epoch": 33.0,
"learning_rate": 0.0001375811320754717,
"loss": 0.0567,
"step": 14850
},
{
"epoch": 33.11,
"learning_rate": 0.00013701509433962264,
"loss": 0.0593,
"step": 14900
},
{
"epoch": 33.22,
"learning_rate": 0.00013644905660377358,
"loss": 0.0503,
"step": 14950
},
{
"epoch": 33.33,
"learning_rate": 0.00013588301886792453,
"loss": 0.0548,
"step": 15000
},
{
"epoch": 33.44,
"learning_rate": 0.00013531698113207548,
"loss": 0.058,
"step": 15050
},
{
"epoch": 33.56,
"learning_rate": 0.0001347509433962264,
"loss": 0.0593,
"step": 15100
},
{
"epoch": 33.67,
"learning_rate": 0.00013418490566037734,
"loss": 0.0576,
"step": 15150
},
{
"epoch": 33.78,
"learning_rate": 0.0001336188679245283,
"loss": 0.0622,
"step": 15200
},
{
"epoch": 33.78,
"eval_loss": 0.6206709742546082,
"eval_runtime": 237.9375,
"eval_samples_per_second": 23.989,
"eval_steps_per_second": 3.001,
"eval_wer": 0.35127271247482045,
"step": 15200
},
{
"epoch": 33.89,
"learning_rate": 0.00013305283018867923,
"loss": 0.0554,
"step": 15250
},
{
"epoch": 34.0,
"learning_rate": 0.00013248679245283018,
"loss": 0.0536,
"step": 15300
},
{
"epoch": 34.11,
"learning_rate": 0.00013192075471698113,
"loss": 0.0585,
"step": 15350
},
{
"epoch": 34.22,
"learning_rate": 0.00013135471698113207,
"loss": 0.0604,
"step": 15400
},
{
"epoch": 34.33,
"learning_rate": 0.00013078867924528302,
"loss": 0.0586,
"step": 15450
},
{
"epoch": 34.44,
"learning_rate": 0.00013023396226415093,
"loss": 0.0648,
"step": 15500
},
{
"epoch": 34.56,
"learning_rate": 0.00012966792452830187,
"loss": 0.0534,
"step": 15550
},
{
"epoch": 34.67,
"learning_rate": 0.00012910188679245282,
"loss": 0.0597,
"step": 15600
},
{
"epoch": 34.67,
"eval_loss": 0.545280933380127,
"eval_runtime": 234.5113,
"eval_samples_per_second": 24.34,
"eval_steps_per_second": 3.045,
"eval_wer": 0.35078438154922986,
"step": 15600
},
{
"epoch": 34.78,
"learning_rate": 0.00012853584905660377,
"loss": 0.058,
"step": 15650
},
{
"epoch": 34.89,
"learning_rate": 0.0001279698113207547,
"loss": 0.0557,
"step": 15700
},
{
"epoch": 35.0,
"learning_rate": 0.00012740377358490566,
"loss": 0.0557,
"step": 15750
},
{
"epoch": 35.11,
"learning_rate": 0.0001268377358490566,
"loss": 0.0529,
"step": 15800
},
{
"epoch": 35.22,
"learning_rate": 0.00012627169811320753,
"loss": 0.0544,
"step": 15850
},
{
"epoch": 35.33,
"learning_rate": 0.00012570566037735847,
"loss": 0.0538,
"step": 15900
},
{
"epoch": 35.44,
"learning_rate": 0.00012513962264150942,
"loss": 0.0517,
"step": 15950
},
{
"epoch": 35.56,
"learning_rate": 0.00012457358490566037,
"loss": 0.0566,
"step": 16000
},
{
"epoch": 35.56,
"eval_loss": 0.6024277210235596,
"eval_runtime": 233.9908,
"eval_samples_per_second": 24.394,
"eval_steps_per_second": 3.051,
"eval_wer": 0.3531649948114839,
"step": 16000
},
{
"epoch": 35.67,
"learning_rate": 0.0001240075471698113,
"loss": 0.051,
"step": 16050
},
{
"epoch": 35.78,
"learning_rate": 0.00012344150943396226,
"loss": 0.0591,
"step": 16100
},
{
"epoch": 35.89,
"learning_rate": 0.0001228754716981132,
"loss": 0.0569,
"step": 16150
},
{
"epoch": 36.0,
"learning_rate": 0.00012230943396226415,
"loss": 0.0563,
"step": 16200
},
{
"epoch": 36.11,
"learning_rate": 0.00012174339622641508,
"loss": 0.0533,
"step": 16250
},
{
"epoch": 36.22,
"learning_rate": 0.00012117735849056603,
"loss": 0.0566,
"step": 16300
},
{
"epoch": 36.33,
"learning_rate": 0.00012061132075471698,
"loss": 0.0556,
"step": 16350
},
{
"epoch": 36.44,
"learning_rate": 0.00012004528301886791,
"loss": 0.0524,
"step": 16400
},
{
"epoch": 36.44,
"eval_loss": 0.6054043173789978,
"eval_runtime": 232.789,
"eval_samples_per_second": 24.52,
"eval_steps_per_second": 3.067,
"eval_wer": 0.3538161027122714,
"step": 16400
},
{
"epoch": 36.56,
"learning_rate": 0.00011947924528301886,
"loss": 0.0568,
"step": 16450
},
{
"epoch": 36.67,
"learning_rate": 0.0001189132075471698,
"loss": 0.0554,
"step": 16500
},
{
"epoch": 36.78,
"learning_rate": 0.00011834716981132075,
"loss": 0.0499,
"step": 16550
},
{
"epoch": 36.89,
"learning_rate": 0.0001177811320754717,
"loss": 0.0518,
"step": 16600
},
{
"epoch": 37.0,
"learning_rate": 0.00011721509433962263,
"loss": 0.0571,
"step": 16650
},
{
"epoch": 37.11,
"learning_rate": 0.00011664905660377358,
"loss": 0.0542,
"step": 16700
},
{
"epoch": 37.22,
"learning_rate": 0.00011608301886792452,
"loss": 0.0502,
"step": 16750
},
{
"epoch": 37.33,
"learning_rate": 0.00011551698113207547,
"loss": 0.045,
"step": 16800
},
{
"epoch": 37.33,
"eval_loss": 0.6067692041397095,
"eval_runtime": 233.7086,
"eval_samples_per_second": 24.424,
"eval_steps_per_second": 3.055,
"eval_wer": 0.3464097503408143,
"step": 16800
},
{
"epoch": 37.44,
"learning_rate": 0.0001149509433962264,
"loss": 0.0536,
"step": 16850
},
{
"epoch": 37.56,
"learning_rate": 0.00011438490566037735,
"loss": 0.0513,
"step": 16900
},
{
"epoch": 37.67,
"learning_rate": 0.0001138188679245283,
"loss": 0.0513,
"step": 16950
},
{
"epoch": 37.78,
"learning_rate": 0.00011325283018867924,
"loss": 0.0546,
"step": 17000
},
{
"epoch": 37.89,
"learning_rate": 0.00011268679245283019,
"loss": 0.0552,
"step": 17050
},
{
"epoch": 38.0,
"learning_rate": 0.00011212075471698112,
"loss": 0.0511,
"step": 17100
},
{
"epoch": 38.11,
"learning_rate": 0.00011155471698113207,
"loss": 0.0517,
"step": 17150
},
{
"epoch": 38.22,
"learning_rate": 0.00011098867924528301,
"loss": 0.0526,
"step": 17200
},
{
"epoch": 38.22,
"eval_loss": 0.5827322602272034,
"eval_runtime": 234.6281,
"eval_samples_per_second": 24.328,
"eval_steps_per_second": 3.043,
"eval_wer": 0.34946181862575537,
"step": 17200
},
{
"epoch": 38.33,
"learning_rate": 0.00011042264150943396,
"loss": 0.0452,
"step": 17250
},
{
"epoch": 38.44,
"learning_rate": 0.0001098566037735849,
"loss": 0.0471,
"step": 17300
},
{
"epoch": 38.56,
"learning_rate": 0.00010929056603773584,
"loss": 0.049,
"step": 17350
},
{
"epoch": 38.67,
"learning_rate": 0.00010873584905660376,
"loss": 0.0551,
"step": 17400
},
{
"epoch": 38.78,
"learning_rate": 0.0001081698113207547,
"loss": 0.0522,
"step": 17450
},
{
"epoch": 38.89,
"learning_rate": 0.00010760377358490565,
"loss": 0.0504,
"step": 17500
},
{
"epoch": 39.0,
"learning_rate": 0.0001070377358490566,
"loss": 0.0474,
"step": 17550
},
{
"epoch": 39.11,
"learning_rate": 0.00010647169811320753,
"loss": 0.0437,
"step": 17600
},
{
"epoch": 39.11,
"eval_loss": 0.6006141901016235,
"eval_runtime": 234.8995,
"eval_samples_per_second": 24.3,
"eval_steps_per_second": 3.04,
"eval_wer": 0.33963415874824504,
"step": 17600
},
{
"epoch": 39.22,
"learning_rate": 0.00010590566037735848,
"loss": 0.0444,
"step": 17650
},
{
"epoch": 39.33,
"learning_rate": 0.00010533962264150943,
"loss": 0.0496,
"step": 17700
},
{
"epoch": 39.44,
"learning_rate": 0.00010477358490566037,
"loss": 0.0514,
"step": 17750
},
{
"epoch": 39.56,
"learning_rate": 0.0001042075471698113,
"loss": 0.0506,
"step": 17800
},
{
"epoch": 39.67,
"learning_rate": 0.00010364150943396225,
"loss": 0.0494,
"step": 17850
},
{
"epoch": 39.78,
"learning_rate": 0.0001030754716981132,
"loss": 0.046,
"step": 17900
},
{
"epoch": 39.89,
"learning_rate": 0.00010250943396226414,
"loss": 0.0515,
"step": 17950
},
{
"epoch": 40.0,
"learning_rate": 0.00010194339622641509,
"loss": 0.0498,
"step": 18000
},
{
"epoch": 40.0,
"eval_loss": 0.546351969242096,
"eval_runtime": 234.7053,
"eval_samples_per_second": 24.32,
"eval_steps_per_second": 3.042,
"eval_wer": 0.3422996317170936,
"step": 18000
},
{
"epoch": 40.11,
"learning_rate": 0.00010137735849056602,
"loss": 0.0514,
"step": 18050
},
{
"epoch": 40.22,
"learning_rate": 0.00010081132075471697,
"loss": 0.0454,
"step": 18100
},
{
"epoch": 40.33,
"learning_rate": 0.00010024528301886792,
"loss": 0.044,
"step": 18150
},
{
"epoch": 40.44,
"learning_rate": 9.967924528301886e-05,
"loss": 0.047,
"step": 18200
},
{
"epoch": 40.56,
"learning_rate": 9.911320754716981e-05,
"loss": 0.0412,
"step": 18250
},
{
"epoch": 40.67,
"learning_rate": 9.854716981132074e-05,
"loss": 0.0484,
"step": 18300
},
{
"epoch": 40.78,
"learning_rate": 9.798113207547169e-05,
"loss": 0.0413,
"step": 18350
},
{
"epoch": 40.89,
"learning_rate": 9.741509433962264e-05,
"loss": 0.0494,
"step": 18400
},
{
"epoch": 40.89,
"eval_loss": 0.6018606424331665,
"eval_runtime": 234.1001,
"eval_samples_per_second": 24.383,
"eval_steps_per_second": 3.05,
"eval_wer": 0.3394713817730482,
"step": 18400
},
{
"epoch": 41.0,
"learning_rate": 9.684905660377358e-05,
"loss": 0.0462,
"step": 18450
},
{
"epoch": 41.11,
"learning_rate": 9.628301886792452e-05,
"loss": 0.0404,
"step": 18500
},
{
"epoch": 41.22,
"learning_rate": 9.571698113207546e-05,
"loss": 0.0452,
"step": 18550
},
{
"epoch": 41.33,
"learning_rate": 9.515094339622641e-05,
"loss": 0.042,
"step": 18600
},
{
"epoch": 41.44,
"learning_rate": 9.458490566037736e-05,
"loss": 0.0449,
"step": 18650
},
{
"epoch": 41.56,
"learning_rate": 9.40188679245283e-05,
"loss": 0.0431,
"step": 18700
},
{
"epoch": 41.67,
"learning_rate": 9.345283018867923e-05,
"loss": 0.0407,
"step": 18750
},
{
"epoch": 41.78,
"learning_rate": 9.288679245283018e-05,
"loss": 0.0476,
"step": 18800
},
{
"epoch": 41.78,
"eval_loss": 0.5934082865715027,
"eval_runtime": 234.8837,
"eval_samples_per_second": 24.301,
"eval_steps_per_second": 3.04,
"eval_wer": 0.3392068691883533,
"step": 18800
},
{
"epoch": 41.89,
"learning_rate": 9.232075471698113e-05,
"loss": 0.0475,
"step": 18850
},
{
"epoch": 42.0,
"learning_rate": 9.175471698113207e-05,
"loss": 0.0422,
"step": 18900
},
{
"epoch": 42.11,
"learning_rate": 9.118867924528302e-05,
"loss": 0.0424,
"step": 18950
},
{
"epoch": 42.22,
"learning_rate": 9.062264150943395e-05,
"loss": 0.0446,
"step": 19000
},
{
"epoch": 42.33,
"learning_rate": 9.00566037735849e-05,
"loss": 0.0435,
"step": 19050
},
{
"epoch": 42.44,
"learning_rate": 8.949056603773585e-05,
"loss": 0.0423,
"step": 19100
},
{
"epoch": 42.56,
"learning_rate": 8.892452830188679e-05,
"loss": 0.0403,
"step": 19150
},
{
"epoch": 42.67,
"learning_rate": 8.835849056603773e-05,
"loss": 0.0414,
"step": 19200
},
{
"epoch": 42.67,
"eval_loss": 0.6187946796417236,
"eval_runtime": 235.2867,
"eval_samples_per_second": 24.26,
"eval_steps_per_second": 3.035,
"eval_wer": 0.3374366695830875,
"step": 19200
},
{
"epoch": 42.78,
"learning_rate": 8.779245283018867e-05,
"loss": 0.0444,
"step": 19250
},
{
"epoch": 42.89,
"learning_rate": 8.722641509433962e-05,
"loss": 0.0476,
"step": 19300
},
{
"epoch": 43.0,
"learning_rate": 8.666037735849057e-05,
"loss": 0.0426,
"step": 19350
},
{
"epoch": 43.11,
"learning_rate": 8.609433962264151e-05,
"loss": 0.04,
"step": 19400
},
{
"epoch": 43.22,
"learning_rate": 8.552830188679245e-05,
"loss": 0.0452,
"step": 19450
},
{
"epoch": 43.33,
"learning_rate": 8.496226415094339e-05,
"loss": 0.0396,
"step": 19500
},
{
"epoch": 43.44,
"learning_rate": 8.439622641509434e-05,
"loss": 0.0404,
"step": 19550
},
{
"epoch": 43.56,
"learning_rate": 8.383018867924528e-05,
"loss": 0.0382,
"step": 19600
},
{
"epoch": 43.56,
"eval_loss": 0.6085843443870544,
"eval_runtime": 233.4151,
"eval_samples_per_second": 24.454,
"eval_steps_per_second": 3.059,
"eval_wer": 0.33273648442427817,
"step": 19600
},
{
"epoch": 43.67,
"learning_rate": 8.326415094339622e-05,
"loss": 0.0408,
"step": 19650
},
{
"epoch": 43.78,
"learning_rate": 8.269811320754716e-05,
"loss": 0.0443,
"step": 19700
},
{
"epoch": 43.89,
"learning_rate": 8.213207547169811e-05,
"loss": 0.0464,
"step": 19750
},
{
"epoch": 44.0,
"learning_rate": 8.156603773584906e-05,
"loss": 0.0383,
"step": 19800
},
{
"epoch": 44.11,
"learning_rate": 8.1e-05,
"loss": 0.0447,
"step": 19850
},
{
"epoch": 44.22,
"learning_rate": 8.043396226415094e-05,
"loss": 0.0419,
"step": 19900
},
{
"epoch": 44.33,
"learning_rate": 7.986792452830188e-05,
"loss": 0.0389,
"step": 19950
},
{
"epoch": 44.44,
"learning_rate": 7.930188679245283e-05,
"loss": 0.0403,
"step": 20000
},
{
"epoch": 44.44,
"eval_loss": 0.6319227814674377,
"eval_runtime": 233.7986,
"eval_samples_per_second": 24.414,
"eval_steps_per_second": 3.054,
"eval_wer": 0.3353002217836287,
"step": 20000
},
{
"epoch": 44.56,
"learning_rate": 7.873584905660378e-05,
"loss": 0.0397,
"step": 20050
},
{
"epoch": 44.67,
"learning_rate": 7.816981132075472e-05,
"loss": 0.0406,
"step": 20100
},
{
"epoch": 44.78,
"learning_rate": 7.760377358490566e-05,
"loss": 0.039,
"step": 20150
},
{
"epoch": 44.89,
"learning_rate": 7.70377358490566e-05,
"loss": 0.0388,
"step": 20200
},
{
"epoch": 45.0,
"learning_rate": 7.647169811320755e-05,
"loss": 0.0391,
"step": 20250
},
{
"epoch": 45.11,
"learning_rate": 7.59056603773585e-05,
"loss": 0.0343,
"step": 20300
},
{
"epoch": 45.22,
"learning_rate": 7.533962264150943e-05,
"loss": 0.0381,
"step": 20350
},
{
"epoch": 45.33,
"learning_rate": 7.477358490566037e-05,
"loss": 0.0391,
"step": 20400
},
{
"epoch": 45.33,
"eval_loss": 0.6092292666435242,
"eval_runtime": 234.709,
"eval_samples_per_second": 24.319,
"eval_steps_per_second": 3.042,
"eval_wer": 0.337253545485991,
"step": 20400
},
{
"epoch": 45.44,
"learning_rate": 7.420754716981131e-05,
"loss": 0.0404,
"step": 20450
},
{
"epoch": 45.56,
"learning_rate": 7.364150943396225e-05,
"loss": 0.0441,
"step": 20500
},
{
"epoch": 45.67,
"learning_rate": 7.30754716981132e-05,
"loss": 0.0374,
"step": 20550
},
{
"epoch": 45.78,
"learning_rate": 7.250943396226415e-05,
"loss": 0.0379,
"step": 20600
},
{
"epoch": 45.89,
"learning_rate": 7.19433962264151e-05,
"loss": 0.04,
"step": 20650
},
{
"epoch": 46.0,
"learning_rate": 7.137735849056603e-05,
"loss": 0.041,
"step": 20700
},
{
"epoch": 46.11,
"learning_rate": 7.081132075471697e-05,
"loss": 0.0367,
"step": 20750
},
{
"epoch": 46.22,
"learning_rate": 7.024528301886792e-05,
"loss": 0.0364,
"step": 20800
},
{
"epoch": 46.22,
"eval_loss": 0.6104596257209778,
"eval_runtime": 238.3111,
"eval_samples_per_second": 23.952,
"eval_steps_per_second": 2.996,
"eval_wer": 0.33383522900685697,
"step": 20800
},
{
"epoch": 46.33,
"learning_rate": 6.967924528301887e-05,
"loss": 0.0379,
"step": 20850
},
{
"epoch": 46.44,
"learning_rate": 6.91132075471698e-05,
"loss": 0.0392,
"step": 20900
},
{
"epoch": 46.56,
"learning_rate": 6.854716981132075e-05,
"loss": 0.0402,
"step": 20950
},
{
"epoch": 46.67,
"learning_rate": 6.798113207547169e-05,
"loss": 0.042,
"step": 21000
},
{
"epoch": 46.78,
"learning_rate": 6.741509433962264e-05,
"loss": 0.0366,
"step": 21050
},
{
"epoch": 46.89,
"learning_rate": 6.684905660377359e-05,
"loss": 0.0377,
"step": 21100
},
{
"epoch": 47.0,
"learning_rate": 6.628301886792452e-05,
"loss": 0.0345,
"step": 21150
},
{
"epoch": 47.11,
"learning_rate": 6.571698113207546e-05,
"loss": 0.0408,
"step": 21200
},
{
"epoch": 47.11,
"eval_loss": 0.6161568760871887,
"eval_runtime": 235.8491,
"eval_samples_per_second": 24.202,
"eval_steps_per_second": 3.027,
"eval_wer": 0.3335910635440617,
"step": 21200
},
{
"epoch": 47.22,
"learning_rate": 6.515094339622641e-05,
"loss": 0.0377,
"step": 21250
},
{
"epoch": 47.33,
"learning_rate": 6.459622641509433e-05,
"loss": 0.0387,
"step": 21300
},
{
"epoch": 47.44,
"learning_rate": 6.403018867924528e-05,
"loss": 0.0349,
"step": 21350
},
{
"epoch": 47.56,
"learning_rate": 6.346415094339622e-05,
"loss": 0.0398,
"step": 21400
},
{
"epoch": 47.67,
"learning_rate": 6.289811320754717e-05,
"loss": 0.036,
"step": 21450
},
{
"epoch": 47.78,
"learning_rate": 6.233207547169812e-05,
"loss": 0.0359,
"step": 21500
},
{
"epoch": 47.89,
"learning_rate": 6.176603773584905e-05,
"loss": 0.0363,
"step": 21550
},
{
"epoch": 48.0,
"learning_rate": 6.12e-05,
"loss": 0.0347,
"step": 21600
},
{
"epoch": 48.0,
"eval_loss": 0.5747588276863098,
"eval_runtime": 234.422,
"eval_samples_per_second": 24.349,
"eval_steps_per_second": 3.046,
"eval_wer": 0.32937920931084297,
"step": 21600
},
{
"epoch": 48.11,
"learning_rate": 6.0633962264150937e-05,
"loss": 0.0357,
"step": 21650
},
{
"epoch": 48.22,
"learning_rate": 6.0067924528301876e-05,
"loss": 0.036,
"step": 21700
},
{
"epoch": 48.33,
"learning_rate": 5.950188679245282e-05,
"loss": 0.0362,
"step": 21750
},
{
"epoch": 48.44,
"learning_rate": 5.893584905660376e-05,
"loss": 0.0313,
"step": 21800
},
{
"epoch": 48.56,
"learning_rate": 5.836981132075471e-05,
"loss": 0.0365,
"step": 21850
},
{
"epoch": 48.67,
"learning_rate": 5.780377358490565e-05,
"loss": 0.0326,
"step": 21900
},
{
"epoch": 48.78,
"learning_rate": 5.7237735849056595e-05,
"loss": 0.0339,
"step": 21950
},
{
"epoch": 48.89,
"learning_rate": 5.667169811320754e-05,
"loss": 0.0372,
"step": 22000
},
{
"epoch": 48.89,
"eval_loss": 0.5893652439117432,
"eval_runtime": 233.9923,
"eval_samples_per_second": 24.394,
"eval_steps_per_second": 3.051,
"eval_wer": 0.3295826805298391,
"step": 22000
},
{
"epoch": 49.0,
"learning_rate": 5.610566037735848e-05,
"loss": 0.0346,
"step": 22050
},
{
"epoch": 49.11,
"learning_rate": 5.553962264150943e-05,
"loss": 0.0356,
"step": 22100
},
{
"epoch": 49.22,
"learning_rate": 5.497358490566037e-05,
"loss": 0.0371,
"step": 22150
},
{
"epoch": 49.33,
"learning_rate": 5.4407547169811314e-05,
"loss": 0.0342,
"step": 22200
},
{
"epoch": 49.44,
"learning_rate": 5.3841509433962254e-05,
"loss": 0.0323,
"step": 22250
},
{
"epoch": 49.56,
"learning_rate": 5.32754716981132e-05,
"loss": 0.0326,
"step": 22300
},
{
"epoch": 49.67,
"learning_rate": 5.270943396226415e-05,
"loss": 0.0349,
"step": 22350
},
{
"epoch": 49.78,
"learning_rate": 5.214339622641509e-05,
"loss": 0.0378,
"step": 22400
},
{
"epoch": 49.78,
"eval_loss": 0.6031844019889832,
"eval_runtime": 234.1475,
"eval_samples_per_second": 24.378,
"eval_steps_per_second": 3.049,
"eval_wer": 0.33098663194091194,
"step": 22400
},
{
"epoch": 49.89,
"learning_rate": 5.157735849056603e-05,
"loss": 0.03,
"step": 22450
},
{
"epoch": 50.0,
"learning_rate": 5.101132075471697e-05,
"loss": 0.0345,
"step": 22500
},
{
"epoch": 50.11,
"learning_rate": 5.044528301886792e-05,
"loss": 0.0358,
"step": 22550
},
{
"epoch": 50.22,
"learning_rate": 4.987924528301886e-05,
"loss": 0.0345,
"step": 22600
},
{
"epoch": 50.33,
"learning_rate": 4.9313207547169806e-05,
"loss": 0.035,
"step": 22650
},
{
"epoch": 50.44,
"learning_rate": 4.874716981132075e-05,
"loss": 0.0331,
"step": 22700
},
{
"epoch": 50.56,
"learning_rate": 4.818113207547169e-05,
"loss": 0.0339,
"step": 22750
},
{
"epoch": 50.67,
"learning_rate": 4.761509433962264e-05,
"loss": 0.0371,
"step": 22800
},
{
"epoch": 50.67,
"eval_loss": 0.5830812454223633,
"eval_runtime": 238.3507,
"eval_samples_per_second": 23.948,
"eval_steps_per_second": 2.996,
"eval_wer": 0.3274665798522799,
"step": 22800
},
{
"epoch": 50.78,
"learning_rate": 4.704905660377358e-05,
"loss": 0.0345,
"step": 22850
},
{
"epoch": 50.89,
"learning_rate": 4.6483018867924525e-05,
"loss": 0.0325,
"step": 22900
},
{
"epoch": 51.0,
"learning_rate": 4.5916981132075465e-05,
"loss": 0.0387,
"step": 22950
},
{
"epoch": 51.11,
"learning_rate": 4.535094339622641e-05,
"loss": 0.0367,
"step": 23000
},
{
"epoch": 51.22,
"learning_rate": 4.478490566037736e-05,
"loss": 0.0317,
"step": 23050
},
{
"epoch": 51.33,
"learning_rate": 4.42188679245283e-05,
"loss": 0.0349,
"step": 23100
},
{
"epoch": 51.44,
"learning_rate": 4.3652830188679244e-05,
"loss": 0.0322,
"step": 23150
},
{
"epoch": 51.56,
"learning_rate": 4.3086792452830184e-05,
"loss": 0.0323,
"step": 23200
},
{
"epoch": 51.56,
"eval_loss": 0.5856512784957886,
"eval_runtime": 235.7185,
"eval_samples_per_second": 24.215,
"eval_steps_per_second": 3.029,
"eval_wer": 0.3265713064886972,
"step": 23200
},
{
"epoch": 51.67,
"learning_rate": 4.252075471698113e-05,
"loss": 0.0329,
"step": 23250
},
{
"epoch": 51.78,
"learning_rate": 4.195471698113207e-05,
"loss": 0.0348,
"step": 23300
},
{
"epoch": 51.89,
"learning_rate": 4.1388679245283016e-05,
"loss": 0.0324,
"step": 23350
},
{
"epoch": 52.0,
"learning_rate": 4.0822641509433956e-05,
"loss": 0.0311,
"step": 23400
},
{
"epoch": 52.11,
"learning_rate": 4.02566037735849e-05,
"loss": 0.031,
"step": 23450
},
{
"epoch": 52.22,
"learning_rate": 3.969056603773585e-05,
"loss": 0.0329,
"step": 23500
},
{
"epoch": 52.33,
"learning_rate": 3.912452830188679e-05,
"loss": 0.031,
"step": 23550
},
{
"epoch": 52.44,
"learning_rate": 3.8558490566037735e-05,
"loss": 0.0313,
"step": 23600
},
{
"epoch": 52.44,
"eval_loss": 0.591876745223999,
"eval_runtime": 234.2309,
"eval_samples_per_second": 24.369,
"eval_steps_per_second": 3.048,
"eval_wer": 0.3222170224021812,
"step": 23600
},
{
"epoch": 52.56,
"learning_rate": 3.7992452830188675e-05,
"loss": 0.0326,
"step": 23650
},
{
"epoch": 52.67,
"learning_rate": 3.742641509433962e-05,
"loss": 0.0277,
"step": 23700
},
{
"epoch": 52.78,
"learning_rate": 3.686037735849056e-05,
"loss": 0.0316,
"step": 23750
},
{
"epoch": 52.89,
"learning_rate": 3.629433962264151e-05,
"loss": 0.0309,
"step": 23800
},
{
"epoch": 53.0,
"learning_rate": 3.5728301886792454e-05,
"loss": 0.032,
"step": 23850
},
{
"epoch": 53.11,
"learning_rate": 3.5162264150943394e-05,
"loss": 0.0289,
"step": 23900
},
{
"epoch": 53.22,
"learning_rate": 3.459622641509434e-05,
"loss": 0.0284,
"step": 23950
},
{
"epoch": 53.33,
"learning_rate": 3.403018867924528e-05,
"loss": 0.0309,
"step": 24000
},
{
"epoch": 53.33,
"eval_loss": 0.5990718007087708,
"eval_runtime": 234.3903,
"eval_samples_per_second": 24.353,
"eval_steps_per_second": 3.046,
"eval_wer": 0.32537082629662034,
"step": 24000
},
{
"epoch": 53.44,
"learning_rate": 3.346415094339622e-05,
"loss": 0.0284,
"step": 24050
},
{
"epoch": 53.56,
"learning_rate": 3.289811320754717e-05,
"loss": 0.0299,
"step": 24100
},
{
"epoch": 53.67,
"learning_rate": 3.2332075471698106e-05,
"loss": 0.0315,
"step": 24150
},
{
"epoch": 53.78,
"learning_rate": 3.176603773584905e-05,
"loss": 0.0323,
"step": 24200
},
{
"epoch": 53.89,
"learning_rate": 3.119999999999999e-05,
"loss": 0.0319,
"step": 24250
},
{
"epoch": 54.0,
"learning_rate": 3.063396226415094e-05,
"loss": 0.0301,
"step": 24300
},
{
"epoch": 54.11,
"learning_rate": 3.0067924528301882e-05,
"loss": 0.0297,
"step": 24350
},
{
"epoch": 54.22,
"learning_rate": 2.9501886792452825e-05,
"loss": 0.0322,
"step": 24400
},
{
"epoch": 54.22,
"eval_loss": 0.6152312755584717,
"eval_runtime": 234.7575,
"eval_samples_per_second": 24.314,
"eval_steps_per_second": 3.041,
"eval_wer": 0.3252894378090219,
"step": 24400
},
{
"epoch": 54.33,
"learning_rate": 2.893584905660377e-05,
"loss": 0.0283,
"step": 24450
},
{
"epoch": 54.44,
"learning_rate": 2.8369811320754715e-05,
"loss": 0.0322,
"step": 24500
},
{
"epoch": 54.56,
"learning_rate": 2.7803773584905658e-05,
"loss": 0.0297,
"step": 24550
},
{
"epoch": 54.67,
"learning_rate": 2.72377358490566e-05,
"loss": 0.0291,
"step": 24600
},
{
"epoch": 54.78,
"learning_rate": 2.6671698113207544e-05,
"loss": 0.0323,
"step": 24650
},
{
"epoch": 54.89,
"learning_rate": 2.6105660377358488e-05,
"loss": 0.0305,
"step": 24700
},
{
"epoch": 55.0,
"learning_rate": 2.553962264150943e-05,
"loss": 0.0268,
"step": 24750
},
{
"epoch": 55.11,
"learning_rate": 2.4973584905660374e-05,
"loss": 0.0304,
"step": 24800
},
{
"epoch": 55.11,
"eval_loss": 0.603844404220581,
"eval_runtime": 234.1113,
"eval_samples_per_second": 24.382,
"eval_steps_per_second": 3.05,
"eval_wer": 0.3228681303029686,
"step": 24800
},
{
"epoch": 55.22,
"learning_rate": 2.440754716981132e-05,
"loss": 0.0276,
"step": 24850
},
{
"epoch": 55.33,
"learning_rate": 2.3841509433962263e-05,
"loss": 0.0266,
"step": 24900
},
{
"epoch": 55.44,
"learning_rate": 2.3275471698113207e-05,
"loss": 0.0292,
"step": 24950
},
{
"epoch": 55.56,
"learning_rate": 2.270943396226415e-05,
"loss": 0.0316,
"step": 25000
},
{
"epoch": 55.67,
"learning_rate": 2.2143396226415093e-05,
"loss": 0.0288,
"step": 25050
},
{
"epoch": 55.78,
"learning_rate": 2.1577358490566036e-05,
"loss": 0.0293,
"step": 25100
},
{
"epoch": 55.89,
"learning_rate": 2.101132075471698e-05,
"loss": 0.0302,
"step": 25150
},
{
"epoch": 56.0,
"learning_rate": 2.0456603773584902e-05,
"loss": 0.0288,
"step": 25200
},
{
"epoch": 56.0,
"eval_loss": 0.5989021062850952,
"eval_runtime": 235.0378,
"eval_samples_per_second": 24.285,
"eval_steps_per_second": 3.038,
"eval_wer": 0.32227806376788004,
"step": 25200
},
{
"epoch": 56.11,
"learning_rate": 1.9890566037735846e-05,
"loss": 0.0309,
"step": 25250
},
{
"epoch": 56.22,
"learning_rate": 1.9324528301886792e-05,
"loss": 0.0277,
"step": 25300
},
{
"epoch": 56.33,
"learning_rate": 1.8758490566037735e-05,
"loss": 0.031,
"step": 25350
},
{
"epoch": 56.44,
"learning_rate": 1.819245283018868e-05,
"loss": 0.0276,
"step": 25400
},
{
"epoch": 56.56,
"learning_rate": 1.762641509433962e-05,
"loss": 0.0273,
"step": 25450
},
{
"epoch": 56.67,
"learning_rate": 1.7060377358490565e-05,
"loss": 0.028,
"step": 25500
},
{
"epoch": 56.78,
"learning_rate": 1.6494339622641508e-05,
"loss": 0.0287,
"step": 25550
},
{
"epoch": 56.89,
"learning_rate": 1.592830188679245e-05,
"loss": 0.0307,
"step": 25600
},
{
"epoch": 56.89,
"eval_loss": 0.5922682285308838,
"eval_runtime": 236.4109,
"eval_samples_per_second": 24.144,
"eval_steps_per_second": 3.02,
"eval_wer": 0.32016196309032086,
"step": 25600
},
{
"epoch": 57.0,
"learning_rate": 1.5362264150943397e-05,
"loss": 0.027,
"step": 25650
},
{
"epoch": 57.11,
"learning_rate": 1.4796226415094337e-05,
"loss": 0.0277,
"step": 25700
},
{
"epoch": 57.22,
"learning_rate": 1.423018867924528e-05,
"loss": 0.0265,
"step": 25750
},
{
"epoch": 57.33,
"learning_rate": 1.3664150943396225e-05,
"loss": 0.0266,
"step": 25800
},
{
"epoch": 57.44,
"learning_rate": 1.3098113207547168e-05,
"loss": 0.0277,
"step": 25850
},
{
"epoch": 57.56,
"learning_rate": 1.2532075471698111e-05,
"loss": 0.0246,
"step": 25900
},
{
"epoch": 57.67,
"learning_rate": 1.1966037735849054e-05,
"loss": 0.0278,
"step": 25950
},
{
"epoch": 57.78,
"learning_rate": 1.14e-05,
"loss": 0.0258,
"step": 26000
},
{
"epoch": 57.78,
"eval_loss": 0.6004139184951782,
"eval_runtime": 234.8842,
"eval_samples_per_second": 24.301,
"eval_steps_per_second": 3.04,
"eval_wer": 0.31922599548293895,
"step": 26000
},
{
"epoch": 57.89,
"learning_rate": 1.0833962264150942e-05,
"loss": 0.0252,
"step": 26050
},
{
"epoch": 58.0,
"learning_rate": 1.0267924528301886e-05,
"loss": 0.027,
"step": 26100
},
{
"epoch": 58.11,
"learning_rate": 9.701886792452829e-06,
"loss": 0.0277,
"step": 26150
},
{
"epoch": 58.22,
"learning_rate": 9.135849056603773e-06,
"loss": 0.0237,
"step": 26200
},
{
"epoch": 58.33,
"learning_rate": 8.569811320754717e-06,
"loss": 0.0232,
"step": 26250
},
{
"epoch": 58.44,
"learning_rate": 8.00377358490566e-06,
"loss": 0.028,
"step": 26300
},
{
"epoch": 58.56,
"learning_rate": 7.437735849056603e-06,
"loss": 0.0293,
"step": 26350
},
{
"epoch": 58.67,
"learning_rate": 6.871698113207546e-06,
"loss": 0.0261,
"step": 26400
},
{
"epoch": 58.67,
"eval_loss": 0.5958569049835205,
"eval_runtime": 234.8641,
"eval_samples_per_second": 24.303,
"eval_steps_per_second": 3.04,
"eval_wer": 0.3189004415325452,
"step": 26400
},
{
"epoch": 58.78,
"learning_rate": 6.30566037735849e-06,
"loss": 0.0293,
"step": 26450
},
{
"epoch": 58.89,
"learning_rate": 5.739622641509433e-06,
"loss": 0.0278,
"step": 26500
},
{
"epoch": 59.0,
"learning_rate": 5.173584905660377e-06,
"loss": 0.0262,
"step": 26550
},
{
"epoch": 59.11,
"learning_rate": 4.60754716981132e-06,
"loss": 0.025,
"step": 26600
},
{
"epoch": 59.22,
"learning_rate": 4.041509433962263e-06,
"loss": 0.0286,
"step": 26650
},
{
"epoch": 59.33,
"learning_rate": 3.4754716981132073e-06,
"loss": 0.0295,
"step": 26700
},
{
"epoch": 59.44,
"learning_rate": 2.909433962264151e-06,
"loss": 0.0249,
"step": 26750
},
{
"epoch": 59.56,
"learning_rate": 2.343396226415094e-06,
"loss": 0.0277,
"step": 26800
},
{
"epoch": 59.56,
"eval_loss": 0.593723475933075,
"eval_runtime": 234.2801,
"eval_samples_per_second": 24.364,
"eval_steps_per_second": 3.048,
"eval_wer": 0.318188292266059,
"step": 26800
}
],
"max_steps": 27000,
"num_train_epochs": 60,
"total_flos": 1.1861566534187504e+20,
"trial_name": null,
"trial_params": null
}