xls-r-uyghur-cv8 / trainer_state.json
lucio's picture
End of training
a658482
raw
history blame
32.5 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 99.99602649006623,
"global_step": 18800,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.53,
"learning_rate": 2.0000000000000003e-06,
"loss": 15.5103,
"step": 100
},
{
"epoch": 1.06,
"learning_rate": 3.980000000000001e-06,
"loss": 10.25,
"step": 200
},
{
"epoch": 1.59,
"learning_rate": 5.98e-06,
"loss": 5.588,
"step": 300
},
{
"epoch": 2.13,
"learning_rate": 7.980000000000002e-06,
"loss": 4.6387,
"step": 400
},
{
"epoch": 2.66,
"learning_rate": 9.980000000000001e-06,
"loss": 4.1169,
"step": 500
},
{
"epoch": 2.66,
"eval_loss": 4.01455020904541,
"eval_runtime": 162.7143,
"eval_samples_per_second": 16.864,
"eval_steps_per_second": 2.108,
"eval_wer": 1.0,
"step": 500
},
{
"epoch": 3.19,
"learning_rate": 1.198e-05,
"loss": 3.796,
"step": 600
},
{
"epoch": 3.72,
"learning_rate": 1.398e-05,
"loss": 3.4906,
"step": 700
},
{
"epoch": 4.25,
"learning_rate": 1.5980000000000003e-05,
"loss": 3.3596,
"step": 800
},
{
"epoch": 4.78,
"learning_rate": 1.798e-05,
"loss": 3.2609,
"step": 900
},
{
"epoch": 5.32,
"learning_rate": 1.9980000000000002e-05,
"loss": 3.2512,
"step": 1000
},
{
"epoch": 5.32,
"eval_loss": 3.234210252761841,
"eval_runtime": 160.5159,
"eval_samples_per_second": 17.095,
"eval_steps_per_second": 2.137,
"eval_wer": 1.0,
"step": 1000
},
{
"epoch": 5.85,
"learning_rate": 2.1980000000000003e-05,
"loss": 3.1927,
"step": 1100
},
{
"epoch": 6.38,
"learning_rate": 2.3980000000000004e-05,
"loss": 3.1722,
"step": 1200
},
{
"epoch": 6.91,
"learning_rate": 2.5980000000000002e-05,
"loss": 3.1008,
"step": 1300
},
{
"epoch": 7.45,
"learning_rate": 2.7980000000000003e-05,
"loss": 2.9719,
"step": 1400
},
{
"epoch": 7.97,
"learning_rate": 2.9980000000000004e-05,
"loss": 2.5435,
"step": 1500
},
{
"epoch": 7.97,
"eval_loss": 1.8155322074890137,
"eval_runtime": 146.5621,
"eval_samples_per_second": 18.722,
"eval_steps_per_second": 2.34,
"eval_wer": 1.0286354695576598,
"step": 1500
},
{
"epoch": 8.51,
"learning_rate": 3.198e-05,
"loss": 2.1032,
"step": 1600
},
{
"epoch": 9.04,
"learning_rate": 3.398e-05,
"loss": 1.8413,
"step": 1700
},
{
"epoch": 9.57,
"learning_rate": 3.5980000000000004e-05,
"loss": 1.7079,
"step": 1800
},
{
"epoch": 10.11,
"learning_rate": 3.7980000000000006e-05,
"loss": 1.6319,
"step": 1900
},
{
"epoch": 10.64,
"learning_rate": 3.998000000000001e-05,
"loss": 1.5575,
"step": 2000
},
{
"epoch": 10.64,
"eval_loss": 0.6345986127853394,
"eval_runtime": 149.5836,
"eval_samples_per_second": 18.344,
"eval_steps_per_second": 2.293,
"eval_wer": 0.705777411829285,
"step": 2000
},
{
"epoch": 11.17,
"learning_rate": 3.976428571428572e-05,
"loss": 1.5137,
"step": 2100
},
{
"epoch": 11.7,
"learning_rate": 3.952857142857143e-05,
"loss": 1.468,
"step": 2200
},
{
"epoch": 12.23,
"learning_rate": 3.9290476190476196e-05,
"loss": 1.4546,
"step": 2300
},
{
"epoch": 12.76,
"learning_rate": 3.905238095238096e-05,
"loss": 1.4071,
"step": 2400
},
{
"epoch": 13.3,
"learning_rate": 3.881428571428572e-05,
"loss": 1.3979,
"step": 2500
},
{
"epoch": 13.3,
"eval_loss": 0.48850002884864807,
"eval_runtime": 146.6274,
"eval_samples_per_second": 18.714,
"eval_steps_per_second": 2.339,
"eval_wer": 0.6320387905402315,
"step": 2500
},
{
"epoch": 13.83,
"learning_rate": 3.857857142857143e-05,
"loss": 1.3532,
"step": 2600
},
{
"epoch": 14.36,
"learning_rate": 3.834047619047619e-05,
"loss": 1.3544,
"step": 2700
},
{
"epoch": 14.89,
"learning_rate": 3.810238095238096e-05,
"loss": 1.335,
"step": 2800
},
{
"epoch": 15.42,
"learning_rate": 3.7864285714285715e-05,
"loss": 1.3402,
"step": 2900
},
{
"epoch": 15.95,
"learning_rate": 3.762619047619048e-05,
"loss": 1.2874,
"step": 3000
},
{
"epoch": 15.95,
"eval_loss": 0.4270566701889038,
"eval_runtime": 150.2121,
"eval_samples_per_second": 18.268,
"eval_steps_per_second": 2.283,
"eval_wer": 0.6088468048122226,
"step": 3000
},
{
"epoch": 16.49,
"learning_rate": 3.7388095238095244e-05,
"loss": 1.2897,
"step": 3100
},
{
"epoch": 17.02,
"learning_rate": 3.715e-05,
"loss": 1.2965,
"step": 3200
},
{
"epoch": 17.55,
"learning_rate": 3.6911904761904766e-05,
"loss": 1.2752,
"step": 3300
},
{
"epoch": 18.08,
"learning_rate": 3.6673809523809524e-05,
"loss": 1.2672,
"step": 3400
},
{
"epoch": 18.61,
"learning_rate": 3.643571428571429e-05,
"loss": 1.2383,
"step": 3500
},
{
"epoch": 18.61,
"eval_loss": 0.3889118432998657,
"eval_runtime": 149.102,
"eval_samples_per_second": 18.404,
"eval_steps_per_second": 2.3,
"eval_wer": 0.586889895247244,
"step": 3500
},
{
"epoch": 19.15,
"learning_rate": 3.619761904761905e-05,
"loss": 1.2479,
"step": 3600
},
{
"epoch": 19.68,
"learning_rate": 3.595952380952381e-05,
"loss": 1.2281,
"step": 3700
},
{
"epoch": 20.21,
"learning_rate": 3.5721428571428575e-05,
"loss": 1.2223,
"step": 3800
},
{
"epoch": 20.74,
"learning_rate": 3.548333333333333e-05,
"loss": 1.2006,
"step": 3900
},
{
"epoch": 21.28,
"learning_rate": 3.52452380952381e-05,
"loss": 1.2054,
"step": 4000
},
{
"epoch": 21.28,
"eval_loss": 0.3609465956687927,
"eval_runtime": 157.7853,
"eval_samples_per_second": 17.391,
"eval_steps_per_second": 2.174,
"eval_wer": 0.5792964640226889,
"step": 4000
},
{
"epoch": 21.81,
"learning_rate": 3.500714285714286e-05,
"loss": 1.1953,
"step": 4100
},
{
"epoch": 22.34,
"learning_rate": 3.476904761904762e-05,
"loss": 1.2022,
"step": 4200
},
{
"epoch": 22.87,
"learning_rate": 3.4530952380952384e-05,
"loss": 1.1841,
"step": 4300
},
{
"epoch": 23.4,
"learning_rate": 3.429285714285715e-05,
"loss": 1.1941,
"step": 4400
},
{
"epoch": 23.93,
"learning_rate": 3.4054761904761906e-05,
"loss": 1.1866,
"step": 4500
},
{
"epoch": 23.93,
"eval_loss": 0.3450409471988678,
"eval_runtime": 147.8034,
"eval_samples_per_second": 18.565,
"eval_steps_per_second": 2.321,
"eval_wer": 0.5513471478889347,
"step": 4500
},
{
"epoch": 24.47,
"learning_rate": 3.381666666666667e-05,
"loss": 1.1688,
"step": 4600
},
{
"epoch": 25.0,
"learning_rate": 3.357857142857143e-05,
"loss": 1.1772,
"step": 4700
},
{
"epoch": 25.53,
"learning_rate": 3.3342857142857146e-05,
"loss": 1.1641,
"step": 4800
},
{
"epoch": 26.06,
"learning_rate": 3.3104761904761904e-05,
"loss": 1.1567,
"step": 4900
},
{
"epoch": 26.59,
"learning_rate": 3.286666666666667e-05,
"loss": 1.1332,
"step": 5000
},
{
"epoch": 26.59,
"eval_loss": 0.32143130898475647,
"eval_runtime": 151.6356,
"eval_samples_per_second": 18.096,
"eval_steps_per_second": 2.262,
"eval_wer": 0.5378985407803851,
"step": 5000
},
{
"epoch": 27.13,
"learning_rate": 3.262857142857143e-05,
"loss": 1.1436,
"step": 5100
},
{
"epoch": 27.66,
"learning_rate": 3.23904761904762e-05,
"loss": 1.1234,
"step": 5200
},
{
"epoch": 28.19,
"learning_rate": 3.215476190476191e-05,
"loss": 1.1273,
"step": 5300
},
{
"epoch": 28.72,
"learning_rate": 3.191666666666667e-05,
"loss": 1.1237,
"step": 5400
},
{
"epoch": 29.25,
"learning_rate": 3.167857142857143e-05,
"loss": 1.135,
"step": 5500
},
{
"epoch": 29.25,
"eval_loss": 0.3122180998325348,
"eval_runtime": 157.9044,
"eval_samples_per_second": 17.378,
"eval_steps_per_second": 2.172,
"eval_wer": 0.538401719957916,
"step": 5500
},
{
"epoch": 29.78,
"learning_rate": 3.1440476190476194e-05,
"loss": 1.1127,
"step": 5600
},
{
"epoch": 30.32,
"learning_rate": 3.120238095238095e-05,
"loss": 1.1288,
"step": 5700
},
{
"epoch": 30.85,
"learning_rate": 3.096428571428572e-05,
"loss": 1.11,
"step": 5800
},
{
"epoch": 31.38,
"learning_rate": 3.072619047619048e-05,
"loss": 1.1098,
"step": 5900
},
{
"epoch": 31.91,
"learning_rate": 3.048809523809524e-05,
"loss": 1.0992,
"step": 6000
},
{
"epoch": 31.91,
"eval_loss": 0.29478520154953003,
"eval_runtime": 149.1488,
"eval_samples_per_second": 18.398,
"eval_steps_per_second": 2.3,
"eval_wer": 0.5078450208133205,
"step": 6000
},
{
"epoch": 32.45,
"learning_rate": 3.025e-05,
"loss": 1.1043,
"step": 6100
},
{
"epoch": 32.97,
"learning_rate": 3.0011904761904765e-05,
"loss": 1.0963,
"step": 6200
},
{
"epoch": 33.51,
"learning_rate": 2.9773809523809526e-05,
"loss": 1.1059,
"step": 6300
},
{
"epoch": 34.04,
"learning_rate": 2.953571428571429e-05,
"loss": 1.095,
"step": 6400
},
{
"epoch": 34.57,
"learning_rate": 2.9297619047619048e-05,
"loss": 1.0707,
"step": 6500
},
{
"epoch": 34.57,
"eval_loss": 0.2927539348602295,
"eval_runtime": 153.3467,
"eval_samples_per_second": 17.894,
"eval_steps_per_second": 2.237,
"eval_wer": 0.5127853254654408,
"step": 6500
},
{
"epoch": 35.11,
"learning_rate": 2.9059523809523812e-05,
"loss": 1.0887,
"step": 6600
},
{
"epoch": 35.64,
"learning_rate": 2.8821428571428574e-05,
"loss": 1.0763,
"step": 6700
},
{
"epoch": 36.17,
"learning_rate": 2.8583333333333335e-05,
"loss": 1.0805,
"step": 6800
},
{
"epoch": 36.7,
"learning_rate": 2.83452380952381e-05,
"loss": 1.0675,
"step": 6900
},
{
"epoch": 37.23,
"learning_rate": 2.8107142857142857e-05,
"loss": 1.0754,
"step": 7000
},
{
"epoch": 37.23,
"eval_loss": 0.28569138050079346,
"eval_runtime": 156.3448,
"eval_samples_per_second": 17.551,
"eval_steps_per_second": 2.194,
"eval_wer": 0.5016696399981703,
"step": 7000
},
{
"epoch": 37.76,
"learning_rate": 2.786904761904762e-05,
"loss": 1.0695,
"step": 7100
},
{
"epoch": 38.3,
"learning_rate": 2.7630952380952383e-05,
"loss": 1.0747,
"step": 7200
},
{
"epoch": 38.83,
"learning_rate": 2.7392857142857147e-05,
"loss": 1.0551,
"step": 7300
},
{
"epoch": 39.36,
"learning_rate": 2.715714285714286e-05,
"loss": 1.0511,
"step": 7400
},
{
"epoch": 39.89,
"learning_rate": 2.6919047619047622e-05,
"loss": 1.0461,
"step": 7500
},
{
"epoch": 39.89,
"eval_loss": 0.27913743257522583,
"eval_runtime": 146.6834,
"eval_samples_per_second": 18.707,
"eval_steps_per_second": 2.338,
"eval_wer": 0.509949224646631,
"step": 7500
},
{
"epoch": 40.42,
"learning_rate": 2.6680952380952387e-05,
"loss": 1.0547,
"step": 7600
},
{
"epoch": 40.95,
"learning_rate": 2.6442857142857144e-05,
"loss": 1.0443,
"step": 7700
},
{
"epoch": 41.49,
"learning_rate": 2.6204761904761905e-05,
"loss": 1.0582,
"step": 7800
},
{
"epoch": 42.02,
"learning_rate": 2.596666666666667e-05,
"loss": 1.0588,
"step": 7900
},
{
"epoch": 42.55,
"learning_rate": 2.572857142857143e-05,
"loss": 1.0328,
"step": 8000
},
{
"epoch": 42.55,
"eval_loss": 0.2728850841522217,
"eval_runtime": 150.8954,
"eval_samples_per_second": 18.185,
"eval_steps_per_second": 2.273,
"eval_wer": 0.511961941356754,
"step": 8000
},
{
"epoch": 43.08,
"learning_rate": 2.5490476190476196e-05,
"loss": 1.0464,
"step": 8100
},
{
"epoch": 43.61,
"learning_rate": 2.5252380952380953e-05,
"loss": 1.0289,
"step": 8200
},
{
"epoch": 44.15,
"learning_rate": 2.5014285714285714e-05,
"loss": 1.026,
"step": 8300
},
{
"epoch": 44.68,
"learning_rate": 2.477619047619048e-05,
"loss": 1.0288,
"step": 8400
},
{
"epoch": 45.21,
"learning_rate": 2.453809523809524e-05,
"loss": 1.0201,
"step": 8500
},
{
"epoch": 45.21,
"eval_loss": 0.2654191255569458,
"eval_runtime": 150.9056,
"eval_samples_per_second": 18.184,
"eval_steps_per_second": 2.273,
"eval_wer": 0.47198206852385527,
"step": 8500
},
{
"epoch": 45.74,
"learning_rate": 2.4300000000000005e-05,
"loss": 1.0167,
"step": 8600
},
{
"epoch": 46.28,
"learning_rate": 2.4061904761904762e-05,
"loss": 1.0183,
"step": 8700
},
{
"epoch": 46.81,
"learning_rate": 2.3823809523809523e-05,
"loss": 1.0093,
"step": 8800
},
{
"epoch": 47.34,
"learning_rate": 2.3585714285714288e-05,
"loss": 1.0236,
"step": 8900
},
{
"epoch": 47.87,
"learning_rate": 2.334761904761905e-05,
"loss": 1.0035,
"step": 9000
},
{
"epoch": 47.87,
"eval_loss": 0.26225191354751587,
"eval_runtime": 146.6476,
"eval_samples_per_second": 18.712,
"eval_steps_per_second": 2.339,
"eval_wer": 0.4658981748318924,
"step": 9000
},
{
"epoch": 48.4,
"learning_rate": 2.3109523809523813e-05,
"loss": 1.0059,
"step": 9100
},
{
"epoch": 48.93,
"learning_rate": 2.287142857142857e-05,
"loss": 1.0031,
"step": 9200
},
{
"epoch": 49.47,
"learning_rate": 2.2633333333333332e-05,
"loss": 1.0092,
"step": 9300
},
{
"epoch": 50.0,
"learning_rate": 2.2395238095238097e-05,
"loss": 0.9908,
"step": 9400
},
{
"epoch": 50.53,
"learning_rate": 2.215952380952381e-05,
"loss": 1.0069,
"step": 9500
},
{
"epoch": 50.53,
"eval_loss": 0.25685444474220276,
"eval_runtime": 152.7548,
"eval_samples_per_second": 17.963,
"eval_steps_per_second": 2.245,
"eval_wer": 0.45931110196239877,
"step": 9500
},
{
"epoch": 51.06,
"learning_rate": 2.1921428571428572e-05,
"loss": 1.0056,
"step": 9600
},
{
"epoch": 51.59,
"learning_rate": 2.1683333333333336e-05,
"loss": 0.9943,
"step": 9700
},
{
"epoch": 52.13,
"learning_rate": 2.1445238095238097e-05,
"loss": 0.9985,
"step": 9800
},
{
"epoch": 52.66,
"learning_rate": 2.120714285714286e-05,
"loss": 0.9858,
"step": 9900
},
{
"epoch": 53.19,
"learning_rate": 2.096904761904762e-05,
"loss": 0.9998,
"step": 10000
},
{
"epoch": 53.19,
"eval_loss": 0.251886785030365,
"eval_runtime": 152.7789,
"eval_samples_per_second": 17.961,
"eval_steps_per_second": 2.245,
"eval_wer": 0.44051049814738574,
"step": 10000
},
{
"epoch": 53.72,
"learning_rate": 2.0730952380952384e-05,
"loss": 0.9829,
"step": 10100
},
{
"epoch": 54.25,
"learning_rate": 2.0492857142857145e-05,
"loss": 0.9944,
"step": 10200
},
{
"epoch": 54.78,
"learning_rate": 2.025476190476191e-05,
"loss": 0.9777,
"step": 10300
},
{
"epoch": 55.32,
"learning_rate": 2.0019047619047624e-05,
"loss": 0.9828,
"step": 10400
},
{
"epoch": 55.85,
"learning_rate": 1.978095238095238e-05,
"loss": 0.9762,
"step": 10500
},
{
"epoch": 55.85,
"eval_loss": 0.2504919767379761,
"eval_runtime": 153.619,
"eval_samples_per_second": 17.862,
"eval_steps_per_second": 2.233,
"eval_wer": 0.45876217922327434,
"step": 10500
},
{
"epoch": 56.38,
"learning_rate": 1.9542857142857143e-05,
"loss": 0.9843,
"step": 10600
},
{
"epoch": 56.91,
"learning_rate": 1.9304761904761907e-05,
"loss": 0.9686,
"step": 10700
},
{
"epoch": 57.45,
"learning_rate": 1.9066666666666668e-05,
"loss": 0.9776,
"step": 10800
},
{
"epoch": 57.97,
"learning_rate": 1.882857142857143e-05,
"loss": 0.9743,
"step": 10900
},
{
"epoch": 58.51,
"learning_rate": 1.859047619047619e-05,
"loss": 0.9755,
"step": 11000
},
{
"epoch": 58.51,
"eval_loss": 0.2478867620229721,
"eval_runtime": 157.921,
"eval_samples_per_second": 17.376,
"eval_steps_per_second": 2.172,
"eval_wer": 0.4563835140204016,
"step": 11000
},
{
"epoch": 59.04,
"learning_rate": 1.8352380952380955e-05,
"loss": 0.9598,
"step": 11100
},
{
"epoch": 59.57,
"learning_rate": 1.8114285714285716e-05,
"loss": 0.9621,
"step": 11200
},
{
"epoch": 60.11,
"learning_rate": 1.7876190476190477e-05,
"loss": 0.9711,
"step": 11300
},
{
"epoch": 60.64,
"learning_rate": 1.7638095238095238e-05,
"loss": 0.9584,
"step": 11400
},
{
"epoch": 61.17,
"learning_rate": 1.7400000000000003e-05,
"loss": 0.9624,
"step": 11500
},
{
"epoch": 61.17,
"eval_loss": 0.24601028859615326,
"eval_runtime": 151.7998,
"eval_samples_per_second": 18.076,
"eval_steps_per_second": 2.26,
"eval_wer": 0.42976076117286494,
"step": 11500
},
{
"epoch": 61.7,
"learning_rate": 1.7161904761904764e-05,
"loss": 0.9489,
"step": 11600
},
{
"epoch": 62.23,
"learning_rate": 1.6923809523809525e-05,
"loss": 0.9563,
"step": 11700
},
{
"epoch": 62.76,
"learning_rate": 1.6685714285714286e-05,
"loss": 0.9483,
"step": 11800
},
{
"epoch": 63.3,
"learning_rate": 1.644761904761905e-05,
"loss": 0.9703,
"step": 11900
},
{
"epoch": 63.83,
"learning_rate": 1.6209523809523812e-05,
"loss": 0.9494,
"step": 12000
},
{
"epoch": 63.83,
"eval_loss": 0.2402362823486328,
"eval_runtime": 160.0311,
"eval_samples_per_second": 17.147,
"eval_steps_per_second": 2.143,
"eval_wer": 0.41823338365125107,
"step": 12000
},
{
"epoch": 64.36,
"learning_rate": 1.5971428571428573e-05,
"loss": 0.9528,
"step": 12100
},
{
"epoch": 64.89,
"learning_rate": 1.5733333333333334e-05,
"loss": 0.9422,
"step": 12200
},
{
"epoch": 65.42,
"learning_rate": 1.5497619047619048e-05,
"loss": 0.9481,
"step": 12300
},
{
"epoch": 65.95,
"learning_rate": 1.5259523809523812e-05,
"loss": 0.9464,
"step": 12400
},
{
"epoch": 66.49,
"learning_rate": 1.5021428571428574e-05,
"loss": 0.948,
"step": 12500
},
{
"epoch": 66.49,
"eval_loss": 0.2412397861480713,
"eval_runtime": 153.0299,
"eval_samples_per_second": 17.931,
"eval_steps_per_second": 2.241,
"eval_wer": 0.4211609715932482,
"step": 12500
},
{
"epoch": 67.02,
"learning_rate": 1.4783333333333335e-05,
"loss": 0.9524,
"step": 12600
},
{
"epoch": 67.55,
"learning_rate": 1.4545238095238097e-05,
"loss": 0.9389,
"step": 12700
},
{
"epoch": 68.08,
"learning_rate": 1.4307142857142857e-05,
"loss": 0.936,
"step": 12800
},
{
"epoch": 68.61,
"learning_rate": 1.406904761904762e-05,
"loss": 0.9284,
"step": 12900
},
{
"epoch": 69.15,
"learning_rate": 1.3830952380952383e-05,
"loss": 0.9312,
"step": 13000
},
{
"epoch": 69.15,
"eval_loss": 0.2352364957332611,
"eval_runtime": 155.983,
"eval_samples_per_second": 17.592,
"eval_steps_per_second": 2.199,
"eval_wer": 0.39700837107177167,
"step": 13000
},
{
"epoch": 69.68,
"learning_rate": 1.3592857142857144e-05,
"loss": 0.9291,
"step": 13100
},
{
"epoch": 70.21,
"learning_rate": 1.3354761904761906e-05,
"loss": 0.9319,
"step": 13200
},
{
"epoch": 70.74,
"learning_rate": 1.311666666666667e-05,
"loss": 0.9168,
"step": 13300
},
{
"epoch": 71.28,
"learning_rate": 1.2878571428571429e-05,
"loss": 0.9359,
"step": 13400
},
{
"epoch": 71.81,
"learning_rate": 1.2640476190476192e-05,
"loss": 0.9172,
"step": 13500
},
{
"epoch": 71.81,
"eval_loss": 0.23573005199432373,
"eval_runtime": 152.9039,
"eval_samples_per_second": 17.946,
"eval_steps_per_second": 2.243,
"eval_wer": 0.3926169891587759,
"step": 13500
},
{
"epoch": 72.34,
"learning_rate": 1.2402380952380953e-05,
"loss": 0.938,
"step": 13600
},
{
"epoch": 72.87,
"learning_rate": 1.2164285714285715e-05,
"loss": 0.9146,
"step": 13700
},
{
"epoch": 73.4,
"learning_rate": 1.1926190476190478e-05,
"loss": 0.9253,
"step": 13800
},
{
"epoch": 73.93,
"learning_rate": 1.1688095238095238e-05,
"loss": 0.9192,
"step": 13900
},
{
"epoch": 74.47,
"learning_rate": 1.145e-05,
"loss": 0.9101,
"step": 14000
},
{
"epoch": 74.47,
"eval_loss": 0.23052847385406494,
"eval_runtime": 144.6085,
"eval_samples_per_second": 18.975,
"eval_steps_per_second": 2.372,
"eval_wer": 0.39046704176387176,
"step": 14000
},
{
"epoch": 75.0,
"learning_rate": 1.1211904761904763e-05,
"loss": 0.9204,
"step": 14100
},
{
"epoch": 75.53,
"learning_rate": 1.0973809523809524e-05,
"loss": 0.9244,
"step": 14200
},
{
"epoch": 76.06,
"learning_rate": 1.0735714285714287e-05,
"loss": 0.9251,
"step": 14300
},
{
"epoch": 76.59,
"learning_rate": 1.0497619047619048e-05,
"loss": 0.9149,
"step": 14400
},
{
"epoch": 77.13,
"learning_rate": 1.0259523809523811e-05,
"loss": 0.9177,
"step": 14500
},
{
"epoch": 77.13,
"eval_loss": 0.23065772652626038,
"eval_runtime": 154.0486,
"eval_samples_per_second": 17.813,
"eval_steps_per_second": 2.227,
"eval_wer": 0.3837884817711907,
"step": 14500
},
{
"epoch": 77.66,
"learning_rate": 1.0021428571428572e-05,
"loss": 0.902,
"step": 14600
},
{
"epoch": 78.19,
"learning_rate": 9.783333333333335e-06,
"loss": 0.9127,
"step": 14700
},
{
"epoch": 78.72,
"learning_rate": 9.545238095238096e-06,
"loss": 0.9078,
"step": 14800
},
{
"epoch": 79.25,
"learning_rate": 9.307142857142857e-06,
"loss": 0.9066,
"step": 14900
},
{
"epoch": 79.78,
"learning_rate": 9.06904761904762e-06,
"loss": 0.9083,
"step": 15000
},
{
"epoch": 79.78,
"eval_loss": 0.2312641590833664,
"eval_runtime": 159.1357,
"eval_samples_per_second": 17.243,
"eval_steps_per_second": 2.155,
"eval_wer": 0.3799917661589131,
"step": 15000
},
{
"epoch": 80.32,
"learning_rate": 8.830952380952381e-06,
"loss": 0.9123,
"step": 15100
},
{
"epoch": 80.85,
"learning_rate": 8.592857142857144e-06,
"loss": 0.902,
"step": 15200
},
{
"epoch": 81.38,
"learning_rate": 8.354761904761905e-06,
"loss": 0.9114,
"step": 15300
},
{
"epoch": 81.91,
"learning_rate": 8.116666666666666e-06,
"loss": 0.8936,
"step": 15400
},
{
"epoch": 82.45,
"learning_rate": 7.878571428571429e-06,
"loss": 0.9068,
"step": 15500
},
{
"epoch": 82.45,
"eval_loss": 0.22753386199474335,
"eval_runtime": 149.5268,
"eval_samples_per_second": 18.351,
"eval_steps_per_second": 2.294,
"eval_wer": 0.3742280773981062,
"step": 15500
},
{
"epoch": 82.97,
"learning_rate": 7.640476190476192e-06,
"loss": 0.8974,
"step": 15600
},
{
"epoch": 83.51,
"learning_rate": 7.402380952380953e-06,
"loss": 0.9025,
"step": 15700
},
{
"epoch": 84.04,
"learning_rate": 7.164285714285715e-06,
"loss": 0.9008,
"step": 15800
},
{
"epoch": 84.57,
"learning_rate": 6.926190476190476e-06,
"loss": 0.8975,
"step": 15900
},
{
"epoch": 85.11,
"learning_rate": 6.688095238095239e-06,
"loss": 0.9087,
"step": 16000
},
{
"epoch": 85.11,
"eval_loss": 0.22831058502197266,
"eval_runtime": 151.3152,
"eval_samples_per_second": 18.134,
"eval_steps_per_second": 2.267,
"eval_wer": 0.3746855130140433,
"step": 16000
},
{
"epoch": 85.64,
"learning_rate": 6.450000000000001e-06,
"loss": 0.8852,
"step": 16100
},
{
"epoch": 86.17,
"learning_rate": 6.211904761904762e-06,
"loss": 0.8967,
"step": 16200
},
{
"epoch": 86.7,
"learning_rate": 5.973809523809524e-06,
"loss": 0.8862,
"step": 16300
},
{
"epoch": 87.23,
"learning_rate": 5.735714285714287e-06,
"loss": 0.9042,
"step": 16400
},
{
"epoch": 87.76,
"learning_rate": 5.497619047619048e-06,
"loss": 0.8838,
"step": 16500
},
{
"epoch": 87.76,
"eval_loss": 0.2285744994878769,
"eval_runtime": 157.9121,
"eval_samples_per_second": 17.377,
"eval_steps_per_second": 2.172,
"eval_wer": 0.37770458807922785,
"step": 16500
},
{
"epoch": 88.3,
"learning_rate": 5.25952380952381e-06,
"loss": 0.8952,
"step": 16600
},
{
"epoch": 88.83,
"learning_rate": 5.021428571428572e-06,
"loss": 0.8893,
"step": 16700
},
{
"epoch": 89.36,
"learning_rate": 4.783333333333334e-06,
"loss": 0.8933,
"step": 16800
},
{
"epoch": 89.89,
"learning_rate": 4.545238095238095e-06,
"loss": 0.8796,
"step": 16900
},
{
"epoch": 90.42,
"learning_rate": 4.307142857142858e-06,
"loss": 0.8868,
"step": 17000
},
{
"epoch": 90.42,
"eval_loss": 0.22693119943141937,
"eval_runtime": 145.6646,
"eval_samples_per_second": 18.838,
"eval_steps_per_second": 2.355,
"eval_wer": 0.37216961712638946,
"step": 17000
},
{
"epoch": 90.95,
"learning_rate": 4.069047619047619e-06,
"loss": 0.8852,
"step": 17100
},
{
"epoch": 91.49,
"learning_rate": 3.830952380952382e-06,
"loss": 0.8895,
"step": 17200
},
{
"epoch": 92.02,
"learning_rate": 3.592857142857143e-06,
"loss": 0.8858,
"step": 17300
},
{
"epoch": 92.55,
"learning_rate": 3.3547619047619052e-06,
"loss": 0.8852,
"step": 17400
},
{
"epoch": 93.08,
"learning_rate": 3.1166666666666668e-06,
"loss": 0.8895,
"step": 17500
},
{
"epoch": 93.08,
"eval_loss": 0.22457998991012573,
"eval_runtime": 151.9978,
"eval_samples_per_second": 18.053,
"eval_steps_per_second": 2.257,
"eval_wer": 0.37143772014089016,
"step": 17500
},
{
"epoch": 93.61,
"learning_rate": 2.8785714285714287e-06,
"loss": 0.8835,
"step": 17600
},
{
"epoch": 94.15,
"learning_rate": 2.6404761904761907e-06,
"loss": 0.8846,
"step": 17700
},
{
"epoch": 94.68,
"learning_rate": 2.4023809523809527e-06,
"loss": 0.8826,
"step": 17800
},
{
"epoch": 95.21,
"learning_rate": 2.1642857142857146e-06,
"loss": 0.8916,
"step": 17900
},
{
"epoch": 95.74,
"learning_rate": 1.926190476190476e-06,
"loss": 0.8926,
"step": 18000
},
{
"epoch": 95.74,
"eval_loss": 0.2240794152021408,
"eval_runtime": 144.8784,
"eval_samples_per_second": 18.94,
"eval_steps_per_second": 2.368,
"eval_wer": 0.37052284890901604,
"step": 18000
},
{
"epoch": 96.28,
"learning_rate": 1.6880952380952381e-06,
"loss": 0.8885,
"step": 18100
},
{
"epoch": 96.81,
"learning_rate": 1.45e-06,
"loss": 0.8728,
"step": 18200
},
{
"epoch": 97.34,
"learning_rate": 1.211904761904762e-06,
"loss": 0.887,
"step": 18300
},
{
"epoch": 97.87,
"learning_rate": 9.73809523809524e-07,
"loss": 0.8809,
"step": 18400
},
{
"epoch": 98.4,
"learning_rate": 7.404761904761905e-07,
"loss": 0.8856,
"step": 18500
},
{
"epoch": 98.4,
"eval_loss": 0.22416575253009796,
"eval_runtime": 184.721,
"eval_samples_per_second": 14.855,
"eval_steps_per_second": 1.857,
"eval_wer": 0.3693335163075797,
"step": 18500
},
{
"epoch": 98.93,
"learning_rate": 5.023809523809524e-07,
"loss": 0.8738,
"step": 18600
},
{
"epoch": 99.47,
"learning_rate": 2.6428571428571433e-07,
"loss": 0.8831,
"step": 18700
},
{
"epoch": 100.0,
"learning_rate": 2.619047619047619e-08,
"loss": 0.877,
"step": 18800
},
{
"epoch": 100.0,
"step": 18800,
"total_flos": 1.0790071156798875e+20,
"train_loss": 1.3463122907597969,
"train_runtime": 52469.8271,
"train_samples_per_second": 11.5,
"train_steps_per_second": 0.358
}
],
"max_steps": 18800,
"num_train_epochs": 100,
"total_flos": 1.0790071156798875e+20,
"trial_name": null,
"trial_params": null
}