unispeech-large-1500h-cv-timit / trainer_state.json
patrickvonplaten's picture
End of training
2fb2942
raw
history blame
41.8 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 20.0,
"global_step": 2900,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.07,
"learning_rate": 9e-07,
"loss": 8.8086,
"step": 10
},
{
"epoch": 0.14,
"learning_rate": 1.9e-06,
"loss": 8.2467,
"step": 20
},
{
"epoch": 0.21,
"learning_rate": 2.9e-06,
"loss": 7.817,
"step": 30
},
{
"epoch": 0.28,
"learning_rate": 3.9e-06,
"loss": 7.8611,
"step": 40
},
{
"epoch": 0.34,
"learning_rate": 4.800000000000001e-06,
"loss": 7.8146,
"step": 50
},
{
"epoch": 0.41,
"learning_rate": 5.8e-06,
"loss": 7.3444,
"step": 60
},
{
"epoch": 0.48,
"learning_rate": 6.800000000000001e-06,
"loss": 6.6011,
"step": 70
},
{
"epoch": 0.55,
"learning_rate": 7.7e-06,
"loss": 6.7507,
"step": 80
},
{
"epoch": 0.62,
"learning_rate": 8.7e-06,
"loss": 5.8342,
"step": 90
},
{
"epoch": 0.69,
"learning_rate": 9.7e-06,
"loss": 4.64,
"step": 100
},
{
"epoch": 0.69,
"eval_loss": 3.9716715812683105,
"eval_runtime": 76.8676,
"eval_samples_per_second": 21.856,
"eval_steps_per_second": 21.856,
"eval_wer": 0.9980704293294742,
"step": 100
},
{
"epoch": 0.76,
"learning_rate": 1.0700000000000001e-05,
"loss": 3.5055,
"step": 110
},
{
"epoch": 0.83,
"learning_rate": 1.1700000000000001e-05,
"loss": 3.2186,
"step": 120
},
{
"epoch": 0.9,
"learning_rate": 1.27e-05,
"loss": 3.0579,
"step": 130
},
{
"epoch": 0.97,
"learning_rate": 1.3700000000000001e-05,
"loss": 2.9804,
"step": 140
},
{
"epoch": 1.03,
"learning_rate": 1.47e-05,
"loss": 2.9674,
"step": 150
},
{
"epoch": 1.1,
"learning_rate": 1.5700000000000002e-05,
"loss": 2.9162,
"step": 160
},
{
"epoch": 1.17,
"learning_rate": 1.6700000000000003e-05,
"loss": 2.8751,
"step": 170
},
{
"epoch": 1.24,
"learning_rate": 1.77e-05,
"loss": 2.8725,
"step": 180
},
{
"epoch": 1.31,
"learning_rate": 1.87e-05,
"loss": 2.8274,
"step": 190
},
{
"epoch": 1.38,
"learning_rate": 1.97e-05,
"loss": 2.6793,
"step": 200
},
{
"epoch": 1.38,
"eval_loss": 2.62642240524292,
"eval_runtime": 77.0778,
"eval_samples_per_second": 21.796,
"eval_steps_per_second": 21.796,
"eval_wer": 1.0,
"step": 200
},
{
"epoch": 1.45,
"learning_rate": 2.07e-05,
"loss": 2.5617,
"step": 210
},
{
"epoch": 1.52,
"learning_rate": 2.1700000000000002e-05,
"loss": 2.494,
"step": 220
},
{
"epoch": 1.59,
"learning_rate": 2.2700000000000003e-05,
"loss": 2.2241,
"step": 230
},
{
"epoch": 1.66,
"learning_rate": 2.37e-05,
"loss": 1.9545,
"step": 240
},
{
"epoch": 1.72,
"learning_rate": 2.47e-05,
"loss": 1.8201,
"step": 250
},
{
"epoch": 1.79,
"learning_rate": 2.57e-05,
"loss": 1.7222,
"step": 260
},
{
"epoch": 1.86,
"learning_rate": 2.6700000000000002e-05,
"loss": 1.2897,
"step": 270
},
{
"epoch": 1.93,
"learning_rate": 2.7700000000000002e-05,
"loss": 1.3516,
"step": 280
},
{
"epoch": 2.0,
"learning_rate": 2.87e-05,
"loss": 1.456,
"step": 290
},
{
"epoch": 2.07,
"learning_rate": 2.97e-05,
"loss": 1.2221,
"step": 300
},
{
"epoch": 2.07,
"eval_loss": 0.999927818775177,
"eval_runtime": 77.2494,
"eval_samples_per_second": 21.748,
"eval_steps_per_second": 21.748,
"eval_wer": 0.7166976776238716,
"step": 300
},
{
"epoch": 2.14,
"learning_rate": 3.07e-05,
"loss": 0.9479,
"step": 310
},
{
"epoch": 2.21,
"learning_rate": 3.1700000000000005e-05,
"loss": 1.0016,
"step": 320
},
{
"epoch": 2.28,
"learning_rate": 3.27e-05,
"loss": 1.1811,
"step": 330
},
{
"epoch": 2.34,
"learning_rate": 3.3700000000000006e-05,
"loss": 0.8885,
"step": 340
},
{
"epoch": 2.41,
"learning_rate": 3.4699999999999996e-05,
"loss": 0.8771,
"step": 350
},
{
"epoch": 2.48,
"learning_rate": 3.57e-05,
"loss": 0.9365,
"step": 360
},
{
"epoch": 2.55,
"learning_rate": 3.6700000000000004e-05,
"loss": 0.9235,
"step": 370
},
{
"epoch": 2.62,
"learning_rate": 3.77e-05,
"loss": 0.68,
"step": 380
},
{
"epoch": 2.69,
"learning_rate": 3.8700000000000006e-05,
"loss": 0.7075,
"step": 390
},
{
"epoch": 2.76,
"learning_rate": 3.97e-05,
"loss": 0.9009,
"step": 400
},
{
"epoch": 2.76,
"eval_loss": 0.6509363055229187,
"eval_runtime": 76.8309,
"eval_samples_per_second": 21.866,
"eval_steps_per_second": 21.866,
"eval_wer": 0.5569567913996278,
"step": 400
},
{
"epoch": 2.83,
"learning_rate": 4.07e-05,
"loss": 0.6544,
"step": 410
},
{
"epoch": 2.9,
"learning_rate": 4.17e-05,
"loss": 0.5842,
"step": 420
},
{
"epoch": 2.97,
"learning_rate": 4.27e-05,
"loss": 0.6458,
"step": 430
},
{
"epoch": 3.03,
"learning_rate": 4.3700000000000005e-05,
"loss": 0.771,
"step": 440
},
{
"epoch": 3.1,
"learning_rate": 4.47e-05,
"loss": 0.4638,
"step": 450
},
{
"epoch": 3.17,
"learning_rate": 4.5700000000000006e-05,
"loss": 0.457,
"step": 460
},
{
"epoch": 3.24,
"learning_rate": 4.6700000000000003e-05,
"loss": 0.5669,
"step": 470
},
{
"epoch": 3.31,
"learning_rate": 4.77e-05,
"loss": 0.5471,
"step": 480
},
{
"epoch": 3.38,
"learning_rate": 4.87e-05,
"loss": 0.3918,
"step": 490
},
{
"epoch": 3.45,
"learning_rate": 4.97e-05,
"loss": 0.4352,
"step": 500
},
{
"epoch": 3.45,
"eval_loss": 0.4682411551475525,
"eval_runtime": 77.9703,
"eval_samples_per_second": 21.547,
"eval_steps_per_second": 21.547,
"eval_wer": 0.4331886155330439,
"step": 500
},
{
"epoch": 3.52,
"learning_rate": 5.0700000000000006e-05,
"loss": 0.5856,
"step": 510
},
{
"epoch": 3.59,
"learning_rate": 5.17e-05,
"loss": 0.3473,
"step": 520
},
{
"epoch": 3.66,
"learning_rate": 5.270000000000001e-05,
"loss": 0.3429,
"step": 530
},
{
"epoch": 3.72,
"learning_rate": 5.3700000000000004e-05,
"loss": 0.4302,
"step": 540
},
{
"epoch": 3.79,
"learning_rate": 5.470000000000001e-05,
"loss": 0.519,
"step": 550
},
{
"epoch": 3.86,
"learning_rate": 5.5700000000000005e-05,
"loss": 0.3218,
"step": 560
},
{
"epoch": 3.93,
"learning_rate": 5.6699999999999996e-05,
"loss": 0.3232,
"step": 570
},
{
"epoch": 4.0,
"learning_rate": 5.77e-05,
"loss": 0.4794,
"step": 580
},
{
"epoch": 4.07,
"learning_rate": 5.87e-05,
"loss": 0.3111,
"step": 590
},
{
"epoch": 4.14,
"learning_rate": 5.97e-05,
"loss": 0.227,
"step": 600
},
{
"epoch": 4.14,
"eval_loss": 0.3661280572414398,
"eval_runtime": 77.4476,
"eval_samples_per_second": 21.692,
"eval_steps_per_second": 21.692,
"eval_wer": 0.35648818137964305,
"step": 600
},
{
"epoch": 4.21,
"learning_rate": 6.07e-05,
"loss": 0.2526,
"step": 610
},
{
"epoch": 4.28,
"learning_rate": 6.170000000000001e-05,
"loss": 0.4031,
"step": 620
},
{
"epoch": 4.34,
"learning_rate": 6.27e-05,
"loss": 0.224,
"step": 630
},
{
"epoch": 4.41,
"learning_rate": 6.37e-05,
"loss": 0.2333,
"step": 640
},
{
"epoch": 4.48,
"learning_rate": 6.47e-05,
"loss": 0.2988,
"step": 650
},
{
"epoch": 4.55,
"learning_rate": 6.570000000000001e-05,
"loss": 0.3293,
"step": 660
},
{
"epoch": 4.62,
"learning_rate": 6.670000000000001e-05,
"loss": 0.1798,
"step": 670
},
{
"epoch": 4.69,
"learning_rate": 6.77e-05,
"loss": 0.2183,
"step": 680
},
{
"epoch": 4.76,
"learning_rate": 6.87e-05,
"loss": 0.346,
"step": 690
},
{
"epoch": 4.83,
"learning_rate": 6.97e-05,
"loss": 0.2169,
"step": 700
},
{
"epoch": 4.83,
"eval_loss": 0.32439830899238586,
"eval_runtime": 75.9099,
"eval_samples_per_second": 22.131,
"eval_steps_per_second": 22.131,
"eval_wer": 0.32030873130728416,
"step": 700
},
{
"epoch": 4.9,
"learning_rate": 7.07e-05,
"loss": 0.1615,
"step": 710
},
{
"epoch": 4.97,
"learning_rate": 7.17e-05,
"loss": 0.2357,
"step": 720
},
{
"epoch": 5.03,
"learning_rate": 7.27e-05,
"loss": 0.2958,
"step": 730
},
{
"epoch": 5.1,
"learning_rate": 7.37e-05,
"loss": 0.1503,
"step": 740
},
{
"epoch": 5.17,
"learning_rate": 7.47e-05,
"loss": 0.1435,
"step": 750
},
{
"epoch": 5.24,
"learning_rate": 7.570000000000001e-05,
"loss": 0.2361,
"step": 760
},
{
"epoch": 5.31,
"learning_rate": 7.670000000000001e-05,
"loss": 0.2031,
"step": 770
},
{
"epoch": 5.38,
"learning_rate": 7.77e-05,
"loss": 0.134,
"step": 780
},
{
"epoch": 5.45,
"learning_rate": 7.87e-05,
"loss": 0.1531,
"step": 790
},
{
"epoch": 5.52,
"learning_rate": 7.970000000000001e-05,
"loss": 0.2687,
"step": 800
},
{
"epoch": 5.52,
"eval_loss": 0.31372734904289246,
"eval_runtime": 75.8556,
"eval_samples_per_second": 22.147,
"eval_steps_per_second": 22.147,
"eval_wer": 0.2981186685962373,
"step": 800
},
{
"epoch": 5.59,
"learning_rate": 8.070000000000001e-05,
"loss": 0.1379,
"step": 810
},
{
"epoch": 5.66,
"learning_rate": 8.17e-05,
"loss": 0.1228,
"step": 820
},
{
"epoch": 5.72,
"learning_rate": 8.27e-05,
"loss": 0.1692,
"step": 830
},
{
"epoch": 5.79,
"learning_rate": 8.37e-05,
"loss": 0.2325,
"step": 840
},
{
"epoch": 5.86,
"learning_rate": 8.47e-05,
"loss": 0.1253,
"step": 850
},
{
"epoch": 5.93,
"learning_rate": 8.57e-05,
"loss": 0.1254,
"step": 860
},
{
"epoch": 6.0,
"learning_rate": 8.67e-05,
"loss": 0.2237,
"step": 870
},
{
"epoch": 6.07,
"learning_rate": 8.77e-05,
"loss": 0.1176,
"step": 880
},
{
"epoch": 6.14,
"learning_rate": 8.87e-05,
"loss": 0.1008,
"step": 890
},
{
"epoch": 6.21,
"learning_rate": 8.970000000000001e-05,
"loss": 0.127,
"step": 900
},
{
"epoch": 6.21,
"eval_loss": 0.3220064342021942,
"eval_runtime": 76.8075,
"eval_samples_per_second": 21.873,
"eval_steps_per_second": 21.873,
"eval_wer": 0.28275101647026396,
"step": 900
},
{
"epoch": 6.28,
"learning_rate": 9.070000000000001e-05,
"loss": 0.1952,
"step": 910
},
{
"epoch": 6.34,
"learning_rate": 9.17e-05,
"loss": 0.1143,
"step": 920
},
{
"epoch": 6.41,
"learning_rate": 9.27e-05,
"loss": 0.112,
"step": 930
},
{
"epoch": 6.48,
"learning_rate": 9.370000000000001e-05,
"loss": 0.1511,
"step": 940
},
{
"epoch": 6.55,
"learning_rate": 9.47e-05,
"loss": 0.1672,
"step": 950
},
{
"epoch": 6.62,
"learning_rate": 9.57e-05,
"loss": 0.0958,
"step": 960
},
{
"epoch": 6.69,
"learning_rate": 9.67e-05,
"loss": 0.1164,
"step": 970
},
{
"epoch": 6.76,
"learning_rate": 9.77e-05,
"loss": 0.2184,
"step": 980
},
{
"epoch": 6.83,
"learning_rate": 9.87e-05,
"loss": 0.1318,
"step": 990
},
{
"epoch": 6.9,
"learning_rate": 9.970000000000001e-05,
"loss": 0.0922,
"step": 1000
},
{
"epoch": 6.9,
"eval_loss": 0.3075151741504669,
"eval_runtime": 76.116,
"eval_samples_per_second": 22.072,
"eval_steps_per_second": 22.072,
"eval_wer": 0.2708290262559438,
"step": 1000
},
{
"epoch": 6.97,
"learning_rate": 9.963157894736843e-05,
"loss": 0.1205,
"step": 1010
},
{
"epoch": 7.03,
"learning_rate": 9.910526315789475e-05,
"loss": 0.1653,
"step": 1020
},
{
"epoch": 7.1,
"learning_rate": 9.857894736842106e-05,
"loss": 0.0842,
"step": 1030
},
{
"epoch": 7.17,
"learning_rate": 9.805263157894737e-05,
"loss": 0.0826,
"step": 1040
},
{
"epoch": 7.24,
"learning_rate": 9.752631578947369e-05,
"loss": 0.1316,
"step": 1050
},
{
"epoch": 7.31,
"learning_rate": 9.7e-05,
"loss": 0.1306,
"step": 1060
},
{
"epoch": 7.38,
"learning_rate": 9.647368421052631e-05,
"loss": 0.0812,
"step": 1070
},
{
"epoch": 7.45,
"learning_rate": 9.594736842105264e-05,
"loss": 0.0962,
"step": 1080
},
{
"epoch": 7.52,
"learning_rate": 9.542105263157895e-05,
"loss": 0.1571,
"step": 1090
},
{
"epoch": 7.59,
"learning_rate": 9.489473684210527e-05,
"loss": 0.0965,
"step": 1100
},
{
"epoch": 7.59,
"eval_loss": 0.2778594195842743,
"eval_runtime": 76.267,
"eval_samples_per_second": 22.028,
"eval_steps_per_second": 22.028,
"eval_wer": 0.2575976845151954,
"step": 1100
},
{
"epoch": 7.66,
"learning_rate": 9.436842105263158e-05,
"loss": 0.0788,
"step": 1110
},
{
"epoch": 7.72,
"learning_rate": 9.384210526315789e-05,
"loss": 0.1061,
"step": 1120
},
{
"epoch": 7.79,
"learning_rate": 9.331578947368422e-05,
"loss": 0.1347,
"step": 1130
},
{
"epoch": 7.86,
"learning_rate": 9.278947368421053e-05,
"loss": 0.0662,
"step": 1140
},
{
"epoch": 7.93,
"learning_rate": 9.226315789473686e-05,
"loss": 0.0833,
"step": 1150
},
{
"epoch": 8.0,
"learning_rate": 9.173684210526317e-05,
"loss": 0.1915,
"step": 1160
},
{
"epoch": 8.07,
"learning_rate": 9.121052631578948e-05,
"loss": 0.0755,
"step": 1170
},
{
"epoch": 8.14,
"learning_rate": 9.06842105263158e-05,
"loss": 0.0561,
"step": 1180
},
{
"epoch": 8.21,
"learning_rate": 9.015789473684211e-05,
"loss": 0.0825,
"step": 1190
},
{
"epoch": 8.28,
"learning_rate": 8.963157894736842e-05,
"loss": 0.1298,
"step": 1200
},
{
"epoch": 8.28,
"eval_loss": 0.31107842922210693,
"eval_runtime": 76.1554,
"eval_samples_per_second": 22.06,
"eval_steps_per_second": 22.06,
"eval_wer": 0.2480187444007994,
"step": 1200
},
{
"epoch": 8.34,
"learning_rate": 8.910526315789474e-05,
"loss": 0.0738,
"step": 1210
},
{
"epoch": 8.41,
"learning_rate": 8.857894736842106e-05,
"loss": 0.071,
"step": 1220
},
{
"epoch": 8.48,
"learning_rate": 8.805263157894737e-05,
"loss": 0.1156,
"step": 1230
},
{
"epoch": 8.55,
"learning_rate": 8.752631578947369e-05,
"loss": 0.1039,
"step": 1240
},
{
"epoch": 8.62,
"learning_rate": 8.7e-05,
"loss": 0.0685,
"step": 1250
},
{
"epoch": 8.69,
"learning_rate": 8.647368421052631e-05,
"loss": 0.0773,
"step": 1260
},
{
"epoch": 8.76,
"learning_rate": 8.594736842105263e-05,
"loss": 0.1302,
"step": 1270
},
{
"epoch": 8.83,
"learning_rate": 8.542105263157894e-05,
"loss": 0.0742,
"step": 1280
},
{
"epoch": 8.9,
"learning_rate": 8.489473684210527e-05,
"loss": 0.0671,
"step": 1290
},
{
"epoch": 8.97,
"learning_rate": 8.436842105263158e-05,
"loss": 0.0855,
"step": 1300
},
{
"epoch": 8.97,
"eval_loss": 0.30214402079582214,
"eval_runtime": 76.6723,
"eval_samples_per_second": 21.911,
"eval_steps_per_second": 21.911,
"eval_wer": 0.24209220591275585,
"step": 1300
},
{
"epoch": 9.03,
"learning_rate": 8.38421052631579e-05,
"loss": 0.1214,
"step": 1310
},
{
"epoch": 9.1,
"learning_rate": 8.331578947368422e-05,
"loss": 0.0566,
"step": 1320
},
{
"epoch": 9.17,
"learning_rate": 8.278947368421053e-05,
"loss": 0.0625,
"step": 1330
},
{
"epoch": 9.24,
"learning_rate": 8.226315789473684e-05,
"loss": 0.1029,
"step": 1340
},
{
"epoch": 9.31,
"learning_rate": 8.173684210526317e-05,
"loss": 0.0785,
"step": 1350
},
{
"epoch": 9.38,
"learning_rate": 8.121052631578948e-05,
"loss": 0.0498,
"step": 1360
},
{
"epoch": 9.45,
"learning_rate": 8.06842105263158e-05,
"loss": 0.0666,
"step": 1370
},
{
"epoch": 9.52,
"learning_rate": 8.015789473684211e-05,
"loss": 0.1167,
"step": 1380
},
{
"epoch": 9.59,
"learning_rate": 7.963157894736842e-05,
"loss": 0.0529,
"step": 1390
},
{
"epoch": 9.66,
"learning_rate": 7.910526315789474e-05,
"loss": 0.0629,
"step": 1400
},
{
"epoch": 9.66,
"eval_loss": 0.31220361590385437,
"eval_runtime": 76.1235,
"eval_samples_per_second": 22.069,
"eval_steps_per_second": 22.069,
"eval_wer": 0.25105092688305425,
"step": 1400
},
{
"epoch": 9.72,
"learning_rate": 7.857894736842105e-05,
"loss": 0.0754,
"step": 1410
},
{
"epoch": 9.79,
"learning_rate": 7.805263157894738e-05,
"loss": 0.0925,
"step": 1420
},
{
"epoch": 9.86,
"learning_rate": 7.752631578947369e-05,
"loss": 0.0511,
"step": 1430
},
{
"epoch": 9.93,
"learning_rate": 7.7e-05,
"loss": 0.0549,
"step": 1440
},
{
"epoch": 10.0,
"learning_rate": 7.647368421052631e-05,
"loss": 0.1087,
"step": 1450
},
{
"epoch": 10.07,
"learning_rate": 7.594736842105263e-05,
"loss": 0.0578,
"step": 1460
},
{
"epoch": 10.14,
"learning_rate": 7.542105263157895e-05,
"loss": 0.0438,
"step": 1470
},
{
"epoch": 10.21,
"learning_rate": 7.489473684210527e-05,
"loss": 0.0517,
"step": 1480
},
{
"epoch": 10.28,
"learning_rate": 7.43684210526316e-05,
"loss": 0.1108,
"step": 1490
},
{
"epoch": 10.34,
"learning_rate": 7.38421052631579e-05,
"loss": 0.0471,
"step": 1500
},
{
"epoch": 10.34,
"eval_loss": 0.29652971029281616,
"eval_runtime": 76.0614,
"eval_samples_per_second": 22.087,
"eval_steps_per_second": 22.087,
"eval_wer": 0.23678588656880986,
"step": 1500
},
{
"epoch": 10.41,
"learning_rate": 7.331578947368422e-05,
"loss": 0.0457,
"step": 1510
},
{
"epoch": 10.48,
"learning_rate": 7.278947368421053e-05,
"loss": 0.081,
"step": 1520
},
{
"epoch": 10.55,
"learning_rate": 7.226315789473685e-05,
"loss": 0.073,
"step": 1530
},
{
"epoch": 10.62,
"learning_rate": 7.173684210526316e-05,
"loss": 0.0482,
"step": 1540
},
{
"epoch": 10.69,
"learning_rate": 7.121052631578947e-05,
"loss": 0.0601,
"step": 1550
},
{
"epoch": 10.76,
"learning_rate": 7.06842105263158e-05,
"loss": 0.1035,
"step": 1560
},
{
"epoch": 10.83,
"learning_rate": 7.015789473684211e-05,
"loss": 0.0575,
"step": 1570
},
{
"epoch": 10.9,
"learning_rate": 6.963157894736842e-05,
"loss": 0.0438,
"step": 1580
},
{
"epoch": 10.97,
"learning_rate": 6.910526315789474e-05,
"loss": 0.0693,
"step": 1590
},
{
"epoch": 11.03,
"learning_rate": 6.857894736842105e-05,
"loss": 0.0871,
"step": 1600
},
{
"epoch": 11.03,
"eval_loss": 0.3247065544128418,
"eval_runtime": 76.1105,
"eval_samples_per_second": 22.073,
"eval_steps_per_second": 22.073,
"eval_wer": 0.23871545723933568,
"step": 1600
},
{
"epoch": 11.1,
"learning_rate": 6.805263157894736e-05,
"loss": 0.0389,
"step": 1610
},
{
"epoch": 11.17,
"learning_rate": 6.752631578947368e-05,
"loss": 0.0368,
"step": 1620
},
{
"epoch": 11.24,
"learning_rate": 6.7e-05,
"loss": 0.0671,
"step": 1630
},
{
"epoch": 11.31,
"learning_rate": 6.647368421052632e-05,
"loss": 0.0606,
"step": 1640
},
{
"epoch": 11.38,
"learning_rate": 6.594736842105264e-05,
"loss": 0.0471,
"step": 1650
},
{
"epoch": 11.45,
"learning_rate": 6.542105263157895e-05,
"loss": 0.0437,
"step": 1660
},
{
"epoch": 11.52,
"learning_rate": 6.489473684210527e-05,
"loss": 0.0677,
"step": 1670
},
{
"epoch": 11.59,
"learning_rate": 6.436842105263158e-05,
"loss": 0.0409,
"step": 1680
},
{
"epoch": 11.66,
"learning_rate": 6.384210526315791e-05,
"loss": 0.027,
"step": 1690
},
{
"epoch": 11.72,
"learning_rate": 6.331578947368422e-05,
"loss": 0.0503,
"step": 1700
},
{
"epoch": 11.72,
"eval_loss": 0.3359486162662506,
"eval_runtime": 76.9661,
"eval_samples_per_second": 21.828,
"eval_steps_per_second": 21.828,
"eval_wer": 0.2363034939011784,
"step": 1700
},
{
"epoch": 11.79,
"learning_rate": 6.278947368421053e-05,
"loss": 0.0746,
"step": 1710
},
{
"epoch": 11.86,
"learning_rate": 6.226315789473685e-05,
"loss": 0.0421,
"step": 1720
},
{
"epoch": 11.93,
"learning_rate": 6.173684210526316e-05,
"loss": 0.0492,
"step": 1730
},
{
"epoch": 12.0,
"learning_rate": 6.121052631578947e-05,
"loss": 0.093,
"step": 1740
},
{
"epoch": 12.07,
"learning_rate": 6.0684210526315785e-05,
"loss": 0.0436,
"step": 1750
},
{
"epoch": 12.14,
"learning_rate": 6.015789473684211e-05,
"loss": 0.0371,
"step": 1760
},
{
"epoch": 12.21,
"learning_rate": 5.9631578947368425e-05,
"loss": 0.0513,
"step": 1770
},
{
"epoch": 12.28,
"learning_rate": 5.9105263157894744e-05,
"loss": 0.0732,
"step": 1780
},
{
"epoch": 12.34,
"learning_rate": 5.857894736842106e-05,
"loss": 0.0434,
"step": 1790
},
{
"epoch": 12.41,
"learning_rate": 5.805263157894737e-05,
"loss": 0.0402,
"step": 1800
},
{
"epoch": 12.41,
"eval_loss": 0.29758504033088684,
"eval_runtime": 76.2359,
"eval_samples_per_second": 22.037,
"eval_steps_per_second": 22.037,
"eval_wer": 0.23320239818069052,
"step": 1800
},
{
"epoch": 12.48,
"learning_rate": 5.752631578947368e-05,
"loss": 0.0635,
"step": 1810
},
{
"epoch": 12.55,
"learning_rate": 5.6999999999999996e-05,
"loss": 0.0722,
"step": 1820
},
{
"epoch": 12.62,
"learning_rate": 5.647368421052632e-05,
"loss": 0.0365,
"step": 1830
},
{
"epoch": 12.69,
"learning_rate": 5.5947368421052636e-05,
"loss": 0.0464,
"step": 1840
},
{
"epoch": 12.76,
"learning_rate": 5.542105263157895e-05,
"loss": 0.0769,
"step": 1850
},
{
"epoch": 12.83,
"learning_rate": 5.489473684210527e-05,
"loss": 0.0402,
"step": 1860
},
{
"epoch": 12.9,
"learning_rate": 5.436842105263158e-05,
"loss": 0.0364,
"step": 1870
},
{
"epoch": 12.97,
"learning_rate": 5.3842105263157895e-05,
"loss": 0.0487,
"step": 1880
},
{
"epoch": 13.03,
"learning_rate": 5.331578947368421e-05,
"loss": 0.0787,
"step": 1890
},
{
"epoch": 13.1,
"learning_rate": 5.2789473684210534e-05,
"loss": 0.0336,
"step": 1900
},
{
"epoch": 13.1,
"eval_loss": 0.31386885046958923,
"eval_runtime": 75.7907,
"eval_samples_per_second": 22.166,
"eval_steps_per_second": 22.166,
"eval_wer": 0.23209978636896147,
"step": 1900
},
{
"epoch": 13.17,
"learning_rate": 5.226315789473685e-05,
"loss": 0.0357,
"step": 1910
},
{
"epoch": 13.24,
"learning_rate": 5.173684210526316e-05,
"loss": 0.0448,
"step": 1920
},
{
"epoch": 13.31,
"learning_rate": 5.121052631578947e-05,
"loss": 0.0473,
"step": 1930
},
{
"epoch": 13.38,
"learning_rate": 5.068421052631579e-05,
"loss": 0.0352,
"step": 1940
},
{
"epoch": 13.45,
"learning_rate": 5.0157894736842106e-05,
"loss": 0.0411,
"step": 1950
},
{
"epoch": 13.52,
"learning_rate": 4.9631578947368426e-05,
"loss": 0.0777,
"step": 1960
},
{
"epoch": 13.59,
"learning_rate": 4.910526315789474e-05,
"loss": 0.0351,
"step": 1970
},
{
"epoch": 13.66,
"learning_rate": 4.857894736842106e-05,
"loss": 0.0334,
"step": 1980
},
{
"epoch": 13.72,
"learning_rate": 4.805263157894737e-05,
"loss": 0.0536,
"step": 1990
},
{
"epoch": 13.79,
"learning_rate": 4.7526315789473684e-05,
"loss": 0.0634,
"step": 2000
},
{
"epoch": 13.79,
"eval_loss": 0.3188297748565674,
"eval_runtime": 76.164,
"eval_samples_per_second": 22.058,
"eval_steps_per_second": 22.058,
"eval_wer": 0.23092826131899938,
"step": 2000
},
{
"epoch": 13.86,
"learning_rate": 4.7e-05,
"loss": 0.0265,
"step": 2010
},
{
"epoch": 13.93,
"learning_rate": 4.647368421052632e-05,
"loss": 0.0291,
"step": 2020
},
{
"epoch": 14.0,
"learning_rate": 4.594736842105264e-05,
"loss": 0.072,
"step": 2030
},
{
"epoch": 14.07,
"learning_rate": 4.542105263157895e-05,
"loss": 0.04,
"step": 2040
},
{
"epoch": 14.14,
"learning_rate": 4.489473684210527e-05,
"loss": 0.0303,
"step": 2050
},
{
"epoch": 14.21,
"learning_rate": 4.436842105263158e-05,
"loss": 0.0309,
"step": 2060
},
{
"epoch": 14.28,
"learning_rate": 4.3842105263157895e-05,
"loss": 0.0628,
"step": 2070
},
{
"epoch": 14.34,
"learning_rate": 4.3315789473684215e-05,
"loss": 0.0307,
"step": 2080
},
{
"epoch": 14.41,
"learning_rate": 4.278947368421053e-05,
"loss": 0.0316,
"step": 2090
},
{
"epoch": 14.48,
"learning_rate": 4.226315789473684e-05,
"loss": 0.0429,
"step": 2100
},
{
"epoch": 14.48,
"eval_loss": 0.31454598903656006,
"eval_runtime": 76.7727,
"eval_samples_per_second": 21.883,
"eval_steps_per_second": 21.883,
"eval_wer": 0.23354696437185582,
"step": 2100
},
{
"epoch": 14.55,
"learning_rate": 4.1736842105263154e-05,
"loss": 0.051,
"step": 2110
},
{
"epoch": 14.62,
"learning_rate": 4.1210526315789474e-05,
"loss": 0.0269,
"step": 2120
},
{
"epoch": 14.69,
"learning_rate": 4.0684210526315794e-05,
"loss": 0.0379,
"step": 2130
},
{
"epoch": 14.76,
"learning_rate": 4.015789473684211e-05,
"loss": 0.0766,
"step": 2140
},
{
"epoch": 14.83,
"learning_rate": 3.9631578947368426e-05,
"loss": 0.0332,
"step": 2150
},
{
"epoch": 14.9,
"learning_rate": 3.910526315789474e-05,
"loss": 0.0302,
"step": 2160
},
{
"epoch": 14.97,
"learning_rate": 3.857894736842105e-05,
"loss": 0.033,
"step": 2170
},
{
"epoch": 15.03,
"learning_rate": 3.8052631578947365e-05,
"loss": 0.0545,
"step": 2180
},
{
"epoch": 15.1,
"learning_rate": 3.7526315789473685e-05,
"loss": 0.0292,
"step": 2190
},
{
"epoch": 15.17,
"learning_rate": 3.7e-05,
"loss": 0.028,
"step": 2200
},
{
"epoch": 15.17,
"eval_loss": 0.3244403302669525,
"eval_runtime": 76.3321,
"eval_samples_per_second": 22.009,
"eval_steps_per_second": 22.009,
"eval_wer": 0.22424367721039212,
"step": 2200
},
{
"epoch": 15.24,
"learning_rate": 3.647368421052632e-05,
"loss": 0.0499,
"step": 2210
},
{
"epoch": 15.31,
"learning_rate": 3.594736842105264e-05,
"loss": 0.0385,
"step": 2220
},
{
"epoch": 15.38,
"learning_rate": 3.542105263157895e-05,
"loss": 0.0338,
"step": 2230
},
{
"epoch": 15.45,
"learning_rate": 3.4894736842105264e-05,
"loss": 0.0309,
"step": 2240
},
{
"epoch": 15.52,
"learning_rate": 3.436842105263158e-05,
"loss": 0.0508,
"step": 2250
},
{
"epoch": 15.59,
"learning_rate": 3.3842105263157896e-05,
"loss": 0.035,
"step": 2260
},
{
"epoch": 15.66,
"learning_rate": 3.331578947368421e-05,
"loss": 0.0273,
"step": 2270
},
{
"epoch": 15.72,
"learning_rate": 3.278947368421052e-05,
"loss": 0.0462,
"step": 2280
},
{
"epoch": 15.79,
"learning_rate": 3.226315789473684e-05,
"loss": 0.0585,
"step": 2290
},
{
"epoch": 15.86,
"learning_rate": 3.173684210526316e-05,
"loss": 0.0255,
"step": 2300
},
{
"epoch": 15.86,
"eval_loss": 0.291358083486557,
"eval_runtime": 76.4196,
"eval_samples_per_second": 21.984,
"eval_steps_per_second": 21.984,
"eval_wer": 0.21962649024877678,
"step": 2300
},
{
"epoch": 15.93,
"learning_rate": 3.1210526315789475e-05,
"loss": 0.0253,
"step": 2310
},
{
"epoch": 16.0,
"learning_rate": 3.0684210526315795e-05,
"loss": 0.0788,
"step": 2320
},
{
"epoch": 16.07,
"learning_rate": 3.0157894736842108e-05,
"loss": 0.028,
"step": 2330
},
{
"epoch": 16.14,
"learning_rate": 2.963157894736842e-05,
"loss": 0.03,
"step": 2340
},
{
"epoch": 16.21,
"learning_rate": 2.910526315789474e-05,
"loss": 0.0416,
"step": 2350
},
{
"epoch": 16.28,
"learning_rate": 2.8578947368421057e-05,
"loss": 0.0643,
"step": 2360
},
{
"epoch": 16.34,
"learning_rate": 2.805263157894737e-05,
"loss": 0.0295,
"step": 2370
},
{
"epoch": 16.41,
"learning_rate": 2.7526315789473683e-05,
"loss": 0.041,
"step": 2380
},
{
"epoch": 16.48,
"learning_rate": 2.7000000000000002e-05,
"loss": 0.0422,
"step": 2390
},
{
"epoch": 16.55,
"learning_rate": 2.647368421052632e-05,
"loss": 0.0406,
"step": 2400
},
{
"epoch": 16.55,
"eval_loss": 0.32486462593078613,
"eval_runtime": 75.8688,
"eval_samples_per_second": 22.143,
"eval_steps_per_second": 22.143,
"eval_wer": 0.2201777961546413,
"step": 2400
},
{
"epoch": 16.62,
"learning_rate": 2.5947368421052632e-05,
"loss": 0.0229,
"step": 2410
},
{
"epoch": 16.69,
"learning_rate": 2.542105263157895e-05,
"loss": 0.0344,
"step": 2420
},
{
"epoch": 16.76,
"learning_rate": 2.4894736842105264e-05,
"loss": 0.0578,
"step": 2430
},
{
"epoch": 16.83,
"learning_rate": 2.436842105263158e-05,
"loss": 0.0324,
"step": 2440
},
{
"epoch": 16.9,
"learning_rate": 2.3842105263157897e-05,
"loss": 0.0226,
"step": 2450
},
{
"epoch": 16.97,
"learning_rate": 2.331578947368421e-05,
"loss": 0.0294,
"step": 2460
},
{
"epoch": 17.03,
"learning_rate": 2.2789473684210527e-05,
"loss": 0.0543,
"step": 2470
},
{
"epoch": 17.1,
"learning_rate": 2.2263157894736843e-05,
"loss": 0.029,
"step": 2480
},
{
"epoch": 17.17,
"learning_rate": 2.173684210526316e-05,
"loss": 0.0272,
"step": 2490
},
{
"epoch": 17.24,
"learning_rate": 2.1210526315789476e-05,
"loss": 0.0512,
"step": 2500
},
{
"epoch": 17.24,
"eval_loss": 0.303699254989624,
"eval_runtime": 76.7934,
"eval_samples_per_second": 21.877,
"eval_steps_per_second": 21.877,
"eval_wer": 0.21983322996347598,
"step": 2500
},
{
"epoch": 17.31,
"learning_rate": 2.068421052631579e-05,
"loss": 0.0364,
"step": 2510
},
{
"epoch": 17.38,
"learning_rate": 2.0157894736842105e-05,
"loss": 0.0199,
"step": 2520
},
{
"epoch": 17.45,
"learning_rate": 1.963157894736842e-05,
"loss": 0.0345,
"step": 2530
},
{
"epoch": 17.52,
"learning_rate": 1.9105263157894738e-05,
"loss": 0.0505,
"step": 2540
},
{
"epoch": 17.59,
"learning_rate": 1.8578947368421054e-05,
"loss": 0.0304,
"step": 2550
},
{
"epoch": 17.66,
"learning_rate": 1.8052631578947367e-05,
"loss": 0.023,
"step": 2560
},
{
"epoch": 17.72,
"learning_rate": 1.7526315789473683e-05,
"loss": 0.0317,
"step": 2570
},
{
"epoch": 17.79,
"learning_rate": 1.7000000000000003e-05,
"loss": 0.0423,
"step": 2580
},
{
"epoch": 17.86,
"learning_rate": 1.6473684210526316e-05,
"loss": 0.0175,
"step": 2590
},
{
"epoch": 17.93,
"learning_rate": 1.5947368421052633e-05,
"loss": 0.0269,
"step": 2600
},
{
"epoch": 17.93,
"eval_loss": 0.32178086042404175,
"eval_runtime": 76.0013,
"eval_samples_per_second": 22.105,
"eval_steps_per_second": 22.105,
"eval_wer": 0.22424367721039212,
"step": 2600
},
{
"epoch": 18.0,
"learning_rate": 1.5421052631578946e-05,
"loss": 0.0504,
"step": 2610
},
{
"epoch": 18.07,
"learning_rate": 1.4894736842105264e-05,
"loss": 0.0292,
"step": 2620
},
{
"epoch": 18.14,
"learning_rate": 1.4368421052631582e-05,
"loss": 0.0248,
"step": 2630
},
{
"epoch": 18.21,
"learning_rate": 1.3842105263157895e-05,
"loss": 0.03,
"step": 2640
},
{
"epoch": 18.28,
"learning_rate": 1.3315789473684213e-05,
"loss": 0.0461,
"step": 2650
},
{
"epoch": 18.34,
"learning_rate": 1.2789473684210526e-05,
"loss": 0.03,
"step": 2660
},
{
"epoch": 18.41,
"learning_rate": 1.2263157894736844e-05,
"loss": 0.0249,
"step": 2670
},
{
"epoch": 18.48,
"learning_rate": 1.1736842105263158e-05,
"loss": 0.0446,
"step": 2680
},
{
"epoch": 18.55,
"learning_rate": 1.1210526315789475e-05,
"loss": 0.0366,
"step": 2690
},
{
"epoch": 18.62,
"learning_rate": 1.068421052631579e-05,
"loss": 0.0287,
"step": 2700
},
{
"epoch": 18.62,
"eval_loss": 0.3106098771095276,
"eval_runtime": 75.9286,
"eval_samples_per_second": 22.126,
"eval_steps_per_second": 22.126,
"eval_wer": 0.21852387843704776,
"step": 2700
},
{
"epoch": 18.69,
"learning_rate": 1.0157894736842106e-05,
"loss": 0.0283,
"step": 2710
},
{
"epoch": 18.76,
"learning_rate": 9.631578947368422e-06,
"loss": 0.0485,
"step": 2720
},
{
"epoch": 18.83,
"learning_rate": 9.105263157894737e-06,
"loss": 0.0259,
"step": 2730
},
{
"epoch": 18.9,
"learning_rate": 8.578947368421053e-06,
"loss": 0.0226,
"step": 2740
},
{
"epoch": 18.97,
"learning_rate": 8.052631578947368e-06,
"loss": 0.0323,
"step": 2750
},
{
"epoch": 19.03,
"learning_rate": 7.526315789473684e-06,
"loss": 0.0453,
"step": 2760
},
{
"epoch": 19.1,
"learning_rate": 7.000000000000001e-06,
"loss": 0.0215,
"step": 2770
},
{
"epoch": 19.17,
"learning_rate": 6.473684210526316e-06,
"loss": 0.0253,
"step": 2780
},
{
"epoch": 19.24,
"learning_rate": 5.947368421052632e-06,
"loss": 0.0461,
"step": 2790
},
{
"epoch": 19.31,
"learning_rate": 5.421052631578947e-06,
"loss": 0.0319,
"step": 2800
},
{
"epoch": 19.31,
"eval_loss": 0.31244826316833496,
"eval_runtime": 76.1154,
"eval_samples_per_second": 22.072,
"eval_steps_per_second": 22.072,
"eval_wer": 0.22169388739576873,
"step": 2800
},
{
"epoch": 19.38,
"learning_rate": 4.894736842105263e-06,
"loss": 0.0243,
"step": 2810
},
{
"epoch": 19.45,
"learning_rate": 4.368421052631579e-06,
"loss": 0.0362,
"step": 2820
},
{
"epoch": 19.52,
"learning_rate": 3.842105263157895e-06,
"loss": 0.039,
"step": 2830
},
{
"epoch": 19.59,
"learning_rate": 3.315789473684211e-06,
"loss": 0.0239,
"step": 2840
},
{
"epoch": 19.66,
"learning_rate": 2.7894736842105266e-06,
"loss": 0.0208,
"step": 2850
},
{
"epoch": 19.72,
"learning_rate": 2.2631578947368426e-06,
"loss": 0.0345,
"step": 2860
},
{
"epoch": 19.79,
"learning_rate": 1.7368421052631579e-06,
"loss": 0.0403,
"step": 2870
},
{
"epoch": 19.86,
"learning_rate": 1.2105263157894738e-06,
"loss": 0.0227,
"step": 2880
},
{
"epoch": 19.93,
"learning_rate": 6.842105263157896e-07,
"loss": 0.0213,
"step": 2890
},
{
"epoch": 20.0,
"learning_rate": 1.5789473684210527e-07,
"loss": 0.0494,
"step": 2900
},
{
"epoch": 20.0,
"eval_loss": 0.3099028766155243,
"eval_runtime": 77.4031,
"eval_samples_per_second": 21.705,
"eval_steps_per_second": 21.705,
"eval_wer": 0.21955757701054374,
"step": 2900
},
{
"epoch": 20.0,
"step": 2900,
"total_flos": 8.88017577624112e+18,
"train_loss": 0.5385495065306795,
"train_runtime": 7090.1183,
"train_samples_per_second": 13.032,
"train_steps_per_second": 0.409
}
],
"max_steps": 2900,
"num_train_epochs": 20,
"total_flos": 8.88017577624112e+18,
"trial_name": null,
"trial_params": null
}