whisper-small-ko / checkpoint-4000 /trainer_state.json
kiru1998's picture
model
715fe46
{
"best_metric": 139.47633434038266,
"best_model_checkpoint": "./whisper-small-ko\\checkpoint-4000",
"epoch": 213.33333333333334,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.33,
"learning_rate": 4.6000000000000004e-07,
"loss": 3.2711,
"step": 25
},
{
"epoch": 2.67,
"learning_rate": 9.600000000000001e-07,
"loss": 2.4436,
"step": 50
},
{
"epoch": 4.0,
"learning_rate": 1.44e-06,
"loss": 1.6348,
"step": 75
},
{
"epoch": 5.33,
"learning_rate": 1.94e-06,
"loss": 0.8506,
"step": 100
},
{
"epoch": 6.67,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.5835,
"step": 125
},
{
"epoch": 8.0,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.4431,
"step": 150
},
{
"epoch": 9.33,
"learning_rate": 3.44e-06,
"loss": 0.3383,
"step": 175
},
{
"epoch": 10.67,
"learning_rate": 3.94e-06,
"loss": 0.2268,
"step": 200
},
{
"epoch": 12.0,
"learning_rate": 4.440000000000001e-06,
"loss": 0.1276,
"step": 225
},
{
"epoch": 13.33,
"learning_rate": 4.94e-06,
"loss": 0.0344,
"step": 250
},
{
"epoch": 14.67,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.0128,
"step": 275
},
{
"epoch": 16.0,
"learning_rate": 5.94e-06,
"loss": 0.0113,
"step": 300
},
{
"epoch": 17.33,
"learning_rate": 6.440000000000001e-06,
"loss": 0.0073,
"step": 325
},
{
"epoch": 18.67,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.0053,
"step": 350
},
{
"epoch": 20.0,
"learning_rate": 7.440000000000001e-06,
"loss": 0.0028,
"step": 375
},
{
"epoch": 21.33,
"learning_rate": 7.94e-06,
"loss": 0.0033,
"step": 400
},
{
"epoch": 22.67,
"learning_rate": 8.44e-06,
"loss": 0.0019,
"step": 425
},
{
"epoch": 24.0,
"learning_rate": 8.94e-06,
"loss": 0.0026,
"step": 450
},
{
"epoch": 25.33,
"learning_rate": 9.440000000000001e-06,
"loss": 0.0011,
"step": 475
},
{
"epoch": 26.67,
"learning_rate": 9.940000000000001e-06,
"loss": 0.0009,
"step": 500
},
{
"epoch": 28.0,
"learning_rate": 9.937142857142858e-06,
"loss": 0.0007,
"step": 525
},
{
"epoch": 29.33,
"learning_rate": 9.865714285714285e-06,
"loss": 0.0007,
"step": 550
},
{
"epoch": 30.67,
"learning_rate": 9.794285714285714e-06,
"loss": 0.0006,
"step": 575
},
{
"epoch": 32.0,
"learning_rate": 9.722857142857143e-06,
"loss": 0.0005,
"step": 600
},
{
"epoch": 33.33,
"learning_rate": 9.651428571428572e-06,
"loss": 0.0005,
"step": 625
},
{
"epoch": 34.67,
"learning_rate": 9.58e-06,
"loss": 0.0004,
"step": 650
},
{
"epoch": 36.0,
"learning_rate": 9.508571428571429e-06,
"loss": 0.0004,
"step": 675
},
{
"epoch": 37.33,
"learning_rate": 9.437142857142858e-06,
"loss": 0.0004,
"step": 700
},
{
"epoch": 38.67,
"learning_rate": 9.365714285714287e-06,
"loss": 0.0004,
"step": 725
},
{
"epoch": 40.0,
"learning_rate": 9.294285714285714e-06,
"loss": 0.0003,
"step": 750
},
{
"epoch": 41.33,
"learning_rate": 9.222857142857143e-06,
"loss": 0.0003,
"step": 775
},
{
"epoch": 42.67,
"learning_rate": 9.151428571428572e-06,
"loss": 0.0003,
"step": 800
},
{
"epoch": 44.0,
"learning_rate": 9.080000000000001e-06,
"loss": 0.0003,
"step": 825
},
{
"epoch": 45.33,
"learning_rate": 9.00857142857143e-06,
"loss": 0.0003,
"step": 850
},
{
"epoch": 46.67,
"learning_rate": 8.937142857142857e-06,
"loss": 0.0003,
"step": 875
},
{
"epoch": 48.0,
"learning_rate": 8.865714285714287e-06,
"loss": 0.0003,
"step": 900
},
{
"epoch": 49.33,
"learning_rate": 8.794285714285716e-06,
"loss": 0.0002,
"step": 925
},
{
"epoch": 50.67,
"learning_rate": 8.722857142857145e-06,
"loss": 0.0002,
"step": 950
},
{
"epoch": 52.0,
"learning_rate": 8.651428571428572e-06,
"loss": 0.0002,
"step": 975
},
{
"epoch": 53.33,
"learning_rate": 8.580000000000001e-06,
"loss": 0.0002,
"step": 1000
},
{
"epoch": 53.33,
"eval_loss": 0.4229048490524292,
"eval_runtime": 65.8019,
"eval_samples_per_second": 1.991,
"eval_steps_per_second": 0.258,
"eval_wer": 168.27794561933536,
"step": 1000
},
{
"epoch": 54.67,
"learning_rate": 8.50857142857143e-06,
"loss": 0.0002,
"step": 1025
},
{
"epoch": 56.0,
"learning_rate": 8.437142857142859e-06,
"loss": 0.0002,
"step": 1050
},
{
"epoch": 57.33,
"learning_rate": 8.365714285714286e-06,
"loss": 0.0002,
"step": 1075
},
{
"epoch": 58.67,
"learning_rate": 8.294285714285715e-06,
"loss": 0.0002,
"step": 1100
},
{
"epoch": 60.0,
"learning_rate": 8.222857142857144e-06,
"loss": 0.0002,
"step": 1125
},
{
"epoch": 61.33,
"learning_rate": 8.151428571428572e-06,
"loss": 0.0002,
"step": 1150
},
{
"epoch": 62.67,
"learning_rate": 8.08e-06,
"loss": 0.0002,
"step": 1175
},
{
"epoch": 64.0,
"learning_rate": 8.00857142857143e-06,
"loss": 0.0002,
"step": 1200
},
{
"epoch": 65.33,
"learning_rate": 7.937142857142857e-06,
"loss": 0.0002,
"step": 1225
},
{
"epoch": 66.67,
"learning_rate": 7.865714285714286e-06,
"loss": 0.0002,
"step": 1250
},
{
"epoch": 68.0,
"learning_rate": 7.794285714285715e-06,
"loss": 0.0001,
"step": 1275
},
{
"epoch": 69.33,
"learning_rate": 7.722857142857142e-06,
"loss": 0.0001,
"step": 1300
},
{
"epoch": 70.67,
"learning_rate": 7.651428571428571e-06,
"loss": 0.0001,
"step": 1325
},
{
"epoch": 72.0,
"learning_rate": 7.58e-06,
"loss": 0.0001,
"step": 1350
},
{
"epoch": 73.33,
"learning_rate": 7.508571428571429e-06,
"loss": 0.0001,
"step": 1375
},
{
"epoch": 74.67,
"learning_rate": 7.4371428571428575e-06,
"loss": 0.0001,
"step": 1400
},
{
"epoch": 76.0,
"learning_rate": 7.365714285714286e-06,
"loss": 0.0001,
"step": 1425
},
{
"epoch": 77.33,
"learning_rate": 7.294285714285715e-06,
"loss": 0.0001,
"step": 1450
},
{
"epoch": 78.67,
"learning_rate": 7.222857142857144e-06,
"loss": 0.0001,
"step": 1475
},
{
"epoch": 80.0,
"learning_rate": 7.151428571428573e-06,
"loss": 0.0001,
"step": 1500
},
{
"epoch": 81.33,
"learning_rate": 7.08e-06,
"loss": 0.0001,
"step": 1525
},
{
"epoch": 82.67,
"learning_rate": 7.008571428571429e-06,
"loss": 0.0001,
"step": 1550
},
{
"epoch": 84.0,
"learning_rate": 6.937142857142858e-06,
"loss": 0.0001,
"step": 1575
},
{
"epoch": 85.33,
"learning_rate": 6.865714285714287e-06,
"loss": 0.0001,
"step": 1600
},
{
"epoch": 86.67,
"learning_rate": 6.794285714285714e-06,
"loss": 0.0001,
"step": 1625
},
{
"epoch": 88.0,
"learning_rate": 6.722857142857143e-06,
"loss": 0.0001,
"step": 1650
},
{
"epoch": 89.33,
"learning_rate": 6.651428571428572e-06,
"loss": 0.0001,
"step": 1675
},
{
"epoch": 90.67,
"learning_rate": 6.5800000000000005e-06,
"loss": 0.0001,
"step": 1700
},
{
"epoch": 92.0,
"learning_rate": 6.5085714285714295e-06,
"loss": 0.0001,
"step": 1725
},
{
"epoch": 93.33,
"learning_rate": 6.437142857142858e-06,
"loss": 0.0001,
"step": 1750
},
{
"epoch": 94.67,
"learning_rate": 6.365714285714286e-06,
"loss": 0.0001,
"step": 1775
},
{
"epoch": 96.0,
"learning_rate": 6.294285714285715e-06,
"loss": 0.0001,
"step": 1800
},
{
"epoch": 97.33,
"learning_rate": 6.222857142857144e-06,
"loss": 0.0001,
"step": 1825
},
{
"epoch": 98.67,
"learning_rate": 6.151428571428571e-06,
"loss": 0.0001,
"step": 1850
},
{
"epoch": 100.0,
"learning_rate": 6.08e-06,
"loss": 0.0001,
"step": 1875
},
{
"epoch": 101.33,
"learning_rate": 6.008571428571429e-06,
"loss": 0.0001,
"step": 1900
},
{
"epoch": 102.67,
"learning_rate": 5.937142857142858e-06,
"loss": 0.0001,
"step": 1925
},
{
"epoch": 104.0,
"learning_rate": 5.865714285714286e-06,
"loss": 0.0001,
"step": 1950
},
{
"epoch": 105.33,
"learning_rate": 5.794285714285715e-06,
"loss": 0.0001,
"step": 1975
},
{
"epoch": 106.67,
"learning_rate": 5.722857142857144e-06,
"loss": 0.0001,
"step": 2000
},
{
"epoch": 106.67,
"eval_loss": 0.4576183259487152,
"eval_runtime": 62.3455,
"eval_samples_per_second": 2.101,
"eval_steps_per_second": 0.273,
"eval_wer": 145.9214501510574,
"step": 2000
},
{
"epoch": 108.0,
"learning_rate": 5.651428571428572e-06,
"loss": 0.0001,
"step": 2025
},
{
"epoch": 109.33,
"learning_rate": 5.580000000000001e-06,
"loss": 0.0001,
"step": 2050
},
{
"epoch": 110.67,
"learning_rate": 5.508571428571429e-06,
"loss": 0.0001,
"step": 2075
},
{
"epoch": 112.0,
"learning_rate": 5.437142857142857e-06,
"loss": 0.0001,
"step": 2100
},
{
"epoch": 113.33,
"learning_rate": 5.365714285714286e-06,
"loss": 0.0001,
"step": 2125
},
{
"epoch": 114.67,
"learning_rate": 5.294285714285715e-06,
"loss": 0.0001,
"step": 2150
},
{
"epoch": 116.0,
"learning_rate": 5.2228571428571425e-06,
"loss": 0.0001,
"step": 2175
},
{
"epoch": 117.33,
"learning_rate": 5.1514285714285715e-06,
"loss": 0.0001,
"step": 2200
},
{
"epoch": 118.67,
"learning_rate": 5.0800000000000005e-06,
"loss": 0.0001,
"step": 2225
},
{
"epoch": 120.0,
"learning_rate": 5.0085714285714295e-06,
"loss": 0.0001,
"step": 2250
},
{
"epoch": 121.33,
"learning_rate": 4.937142857142858e-06,
"loss": 0.0001,
"step": 2275
},
{
"epoch": 122.67,
"learning_rate": 4.865714285714287e-06,
"loss": 0.0001,
"step": 2300
},
{
"epoch": 124.0,
"learning_rate": 4.794285714285715e-06,
"loss": 0.0001,
"step": 2325
},
{
"epoch": 125.33,
"learning_rate": 4.722857142857144e-06,
"loss": 0.0001,
"step": 2350
},
{
"epoch": 126.67,
"learning_rate": 4.651428571428572e-06,
"loss": 0.0001,
"step": 2375
},
{
"epoch": 128.0,
"learning_rate": 4.58e-06,
"loss": 0.0001,
"step": 2400
},
{
"epoch": 129.33,
"learning_rate": 4.508571428571429e-06,
"loss": 0.0001,
"step": 2425
},
{
"epoch": 130.67,
"learning_rate": 4.437142857142857e-06,
"loss": 0.0001,
"step": 2450
},
{
"epoch": 132.0,
"learning_rate": 4.3657142857142855e-06,
"loss": 0.0001,
"step": 2475
},
{
"epoch": 133.33,
"learning_rate": 4.2942857142857146e-06,
"loss": 0.0001,
"step": 2500
},
{
"epoch": 134.67,
"learning_rate": 4.222857142857143e-06,
"loss": 0.0001,
"step": 2525
},
{
"epoch": 136.0,
"learning_rate": 4.151428571428572e-06,
"loss": 0.0001,
"step": 2550
},
{
"epoch": 137.33,
"learning_rate": 4.08e-06,
"loss": 0.0001,
"step": 2575
},
{
"epoch": 138.67,
"learning_rate": 4.008571428571429e-06,
"loss": 0.0001,
"step": 2600
},
{
"epoch": 140.0,
"learning_rate": 3.937142857142858e-06,
"loss": 0.0001,
"step": 2625
},
{
"epoch": 141.33,
"learning_rate": 3.865714285714286e-06,
"loss": 0.0001,
"step": 2650
},
{
"epoch": 142.67,
"learning_rate": 3.7942857142857147e-06,
"loss": 0.0001,
"step": 2675
},
{
"epoch": 144.0,
"learning_rate": 3.722857142857143e-06,
"loss": 0.0001,
"step": 2700
},
{
"epoch": 145.33,
"learning_rate": 3.651428571428572e-06,
"loss": 0.0001,
"step": 2725
},
{
"epoch": 146.67,
"learning_rate": 3.58e-06,
"loss": 0.0001,
"step": 2750
},
{
"epoch": 148.0,
"learning_rate": 3.508571428571429e-06,
"loss": 0.0001,
"step": 2775
},
{
"epoch": 149.33,
"learning_rate": 3.437142857142857e-06,
"loss": 0.0001,
"step": 2800
},
{
"epoch": 150.67,
"learning_rate": 3.3657142857142862e-06,
"loss": 0.0001,
"step": 2825
},
{
"epoch": 152.0,
"learning_rate": 3.2942857142857144e-06,
"loss": 0.0001,
"step": 2850
},
{
"epoch": 153.33,
"learning_rate": 3.222857142857143e-06,
"loss": 0.0001,
"step": 2875
},
{
"epoch": 154.67,
"learning_rate": 3.151428571428572e-06,
"loss": 0.0,
"step": 2900
},
{
"epoch": 156.0,
"learning_rate": 3.08e-06,
"loss": 0.0,
"step": 2925
},
{
"epoch": 157.33,
"learning_rate": 3.008571428571429e-06,
"loss": 0.0,
"step": 2950
},
{
"epoch": 158.67,
"learning_rate": 2.9371428571428573e-06,
"loss": 0.0,
"step": 2975
},
{
"epoch": 160.0,
"learning_rate": 2.865714285714286e-06,
"loss": 0.0,
"step": 3000
},
{
"epoch": 160.0,
"eval_loss": 0.4790264368057251,
"eval_runtime": 63.1579,
"eval_samples_per_second": 2.074,
"eval_steps_per_second": 0.269,
"eval_wer": 147.53272910372607,
"step": 3000
},
{
"epoch": 161.33,
"learning_rate": 2.7942857142857145e-06,
"loss": 0.0,
"step": 3025
},
{
"epoch": 162.67,
"learning_rate": 2.722857142857143e-06,
"loss": 0.0,
"step": 3050
},
{
"epoch": 164.0,
"learning_rate": 2.6514285714285713e-06,
"loss": 0.0,
"step": 3075
},
{
"epoch": 165.33,
"learning_rate": 2.5800000000000003e-06,
"loss": 0.0,
"step": 3100
},
{
"epoch": 166.67,
"learning_rate": 2.5085714285714285e-06,
"loss": 0.0,
"step": 3125
},
{
"epoch": 168.0,
"learning_rate": 2.4371428571428575e-06,
"loss": 0.0,
"step": 3150
},
{
"epoch": 169.33,
"learning_rate": 2.365714285714286e-06,
"loss": 0.0,
"step": 3175
},
{
"epoch": 170.67,
"learning_rate": 2.2942857142857146e-06,
"loss": 0.0,
"step": 3200
},
{
"epoch": 172.0,
"learning_rate": 2.222857142857143e-06,
"loss": 0.0,
"step": 3225
},
{
"epoch": 173.33,
"learning_rate": 2.1514285714285714e-06,
"loss": 0.0,
"step": 3250
},
{
"epoch": 174.67,
"learning_rate": 2.08e-06,
"loss": 0.0,
"step": 3275
},
{
"epoch": 176.0,
"learning_rate": 2.0085714285714286e-06,
"loss": 0.0,
"step": 3300
},
{
"epoch": 177.33,
"learning_rate": 1.9371428571428576e-06,
"loss": 0.0,
"step": 3325
},
{
"epoch": 178.67,
"learning_rate": 1.865714285714286e-06,
"loss": 0.0,
"step": 3350
},
{
"epoch": 180.0,
"learning_rate": 1.7942857142857146e-06,
"loss": 0.0,
"step": 3375
},
{
"epoch": 181.33,
"learning_rate": 1.7228571428571432e-06,
"loss": 0.0,
"step": 3400
},
{
"epoch": 182.67,
"learning_rate": 1.6514285714285715e-06,
"loss": 0.0,
"step": 3425
},
{
"epoch": 184.0,
"learning_rate": 1.5800000000000001e-06,
"loss": 0.0,
"step": 3450
},
{
"epoch": 185.33,
"learning_rate": 1.5085714285714287e-06,
"loss": 0.0,
"step": 3475
},
{
"epoch": 186.67,
"learning_rate": 1.4371428571428573e-06,
"loss": 0.0,
"step": 3500
},
{
"epoch": 188.0,
"learning_rate": 1.3657142857142857e-06,
"loss": 0.0,
"step": 3525
},
{
"epoch": 189.33,
"learning_rate": 1.2942857142857143e-06,
"loss": 0.0,
"step": 3550
},
{
"epoch": 190.67,
"learning_rate": 1.222857142857143e-06,
"loss": 0.0,
"step": 3575
},
{
"epoch": 192.0,
"learning_rate": 1.1514285714285714e-06,
"loss": 0.0,
"step": 3600
},
{
"epoch": 193.33,
"learning_rate": 1.08e-06,
"loss": 0.0,
"step": 3625
},
{
"epoch": 194.67,
"learning_rate": 1.0085714285714286e-06,
"loss": 0.0,
"step": 3650
},
{
"epoch": 196.0,
"learning_rate": 9.371428571428571e-07,
"loss": 0.0,
"step": 3675
},
{
"epoch": 197.33,
"learning_rate": 8.657142857142858e-07,
"loss": 0.0,
"step": 3700
},
{
"epoch": 198.67,
"learning_rate": 7.942857142857144e-07,
"loss": 0.0,
"step": 3725
},
{
"epoch": 200.0,
"learning_rate": 7.228571428571429e-07,
"loss": 0.0,
"step": 3750
},
{
"epoch": 201.33,
"learning_rate": 6.514285714285715e-07,
"loss": 0.0,
"step": 3775
},
{
"epoch": 202.67,
"learning_rate": 5.800000000000001e-07,
"loss": 0.0,
"step": 3800
},
{
"epoch": 204.0,
"learning_rate": 5.085714285714286e-07,
"loss": 0.0,
"step": 3825
},
{
"epoch": 205.33,
"learning_rate": 4.371428571428572e-07,
"loss": 0.0,
"step": 3850
},
{
"epoch": 206.67,
"learning_rate": 3.657142857142858e-07,
"loss": 0.0,
"step": 3875
},
{
"epoch": 208.0,
"learning_rate": 2.942857142857143e-07,
"loss": 0.0,
"step": 3900
},
{
"epoch": 209.33,
"learning_rate": 2.228571428571429e-07,
"loss": 0.0,
"step": 3925
},
{
"epoch": 210.67,
"learning_rate": 1.5142857142857144e-07,
"loss": 0.0,
"step": 3950
},
{
"epoch": 212.0,
"learning_rate": 8e-08,
"loss": 0.0,
"step": 3975
},
{
"epoch": 213.33,
"learning_rate": 8.571428571428572e-09,
"loss": 0.0,
"step": 4000
},
{
"epoch": 213.33,
"eval_loss": 0.48670610785484314,
"eval_runtime": 76.1165,
"eval_samples_per_second": 1.721,
"eval_steps_per_second": 0.223,
"eval_wer": 139.47633434038266,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 223,
"total_flos": 1.828505954967552e+19,
"trial_name": null,
"trial_params": null
}