server-4000 / trainer_state.json
lekiet258's picture
push checkpoint-4000
aea2f44
{
"best_metric": 143.92346010911677,
"best_model_checkpoint": "./whisper-small-vn/checkpoint-1000",
"epoch": 30.075187969924812,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.19,
"learning_rate": 4.0000000000000003e-07,
"loss": 4.7195,
"step": 25
},
{
"epoch": 0.38,
"learning_rate": 9.000000000000001e-07,
"loss": 3.5121,
"step": 50
},
{
"epoch": 0.56,
"learning_rate": 1.4000000000000001e-06,
"loss": 2.3064,
"step": 75
},
{
"epoch": 0.75,
"learning_rate": 1.9000000000000002e-06,
"loss": 1.8246,
"step": 100
},
{
"epoch": 0.94,
"learning_rate": 2.4000000000000003e-06,
"loss": 1.7025,
"step": 125
},
{
"epoch": 1.13,
"learning_rate": 2.9e-06,
"loss": 1.5639,
"step": 150
},
{
"epoch": 1.32,
"learning_rate": 3.4000000000000005e-06,
"loss": 1.3348,
"step": 175
},
{
"epoch": 1.5,
"learning_rate": 3.900000000000001e-06,
"loss": 1.2405,
"step": 200
},
{
"epoch": 1.69,
"learning_rate": 4.4e-06,
"loss": 1.2094,
"step": 225
},
{
"epoch": 1.88,
"learning_rate": 4.9000000000000005e-06,
"loss": 1.1991,
"step": 250
},
{
"epoch": 2.07,
"learning_rate": 5.400000000000001e-06,
"loss": 1.0683,
"step": 275
},
{
"epoch": 2.26,
"learning_rate": 5.9e-06,
"loss": 0.9159,
"step": 300
},
{
"epoch": 2.44,
"learning_rate": 6.4000000000000006e-06,
"loss": 0.9682,
"step": 325
},
{
"epoch": 2.63,
"learning_rate": 6.9e-06,
"loss": 0.887,
"step": 350
},
{
"epoch": 2.82,
"learning_rate": 7.4e-06,
"loss": 0.8728,
"step": 375
},
{
"epoch": 3.01,
"learning_rate": 7.9e-06,
"loss": 0.8998,
"step": 400
},
{
"epoch": 3.2,
"learning_rate": 8.400000000000001e-06,
"loss": 0.6549,
"step": 425
},
{
"epoch": 3.38,
"learning_rate": 8.900000000000001e-06,
"loss": 0.6609,
"step": 450
},
{
"epoch": 3.57,
"learning_rate": 9.4e-06,
"loss": 0.567,
"step": 475
},
{
"epoch": 3.76,
"learning_rate": 9.9e-06,
"loss": 0.645,
"step": 500
},
{
"epoch": 3.95,
"learning_rate": 9.942857142857145e-06,
"loss": 0.6003,
"step": 525
},
{
"epoch": 4.14,
"learning_rate": 9.871428571428572e-06,
"loss": 0.4924,
"step": 550
},
{
"epoch": 4.32,
"learning_rate": 9.800000000000001e-06,
"loss": 0.4309,
"step": 575
},
{
"epoch": 4.51,
"learning_rate": 9.72857142857143e-06,
"loss": 0.3875,
"step": 600
},
{
"epoch": 4.7,
"learning_rate": 9.657142857142859e-06,
"loss": 0.3996,
"step": 625
},
{
"epoch": 4.89,
"learning_rate": 9.585714285714286e-06,
"loss": 0.4067,
"step": 650
},
{
"epoch": 5.08,
"learning_rate": 9.514285714285715e-06,
"loss": 0.3232,
"step": 675
},
{
"epoch": 5.26,
"learning_rate": 9.442857142857144e-06,
"loss": 0.2838,
"step": 700
},
{
"epoch": 5.45,
"learning_rate": 9.371428571428572e-06,
"loss": 0.2525,
"step": 725
},
{
"epoch": 5.64,
"learning_rate": 9.3e-06,
"loss": 0.2575,
"step": 750
},
{
"epoch": 5.83,
"learning_rate": 9.22857142857143e-06,
"loss": 0.2609,
"step": 775
},
{
"epoch": 6.02,
"learning_rate": 9.157142857142857e-06,
"loss": 0.2553,
"step": 800
},
{
"epoch": 6.2,
"learning_rate": 9.085714285714286e-06,
"loss": 0.1759,
"step": 825
},
{
"epoch": 6.39,
"learning_rate": 9.014285714285715e-06,
"loss": 0.1816,
"step": 850
},
{
"epoch": 6.58,
"learning_rate": 8.942857142857142e-06,
"loss": 0.1731,
"step": 875
},
{
"epoch": 6.77,
"learning_rate": 8.871428571428571e-06,
"loss": 0.183,
"step": 900
},
{
"epoch": 6.95,
"learning_rate": 8.8e-06,
"loss": 0.1946,
"step": 925
},
{
"epoch": 7.14,
"learning_rate": 8.72857142857143e-06,
"loss": 0.1591,
"step": 950
},
{
"epoch": 7.33,
"learning_rate": 8.657142857142858e-06,
"loss": 0.1947,
"step": 975
},
{
"epoch": 7.52,
"learning_rate": 8.585714285714286e-06,
"loss": 0.1948,
"step": 1000
},
{
"epoch": 7.52,
"eval_loss": 1.0915887355804443,
"eval_runtime": 131.3357,
"eval_samples_per_second": 2.01,
"eval_steps_per_second": 0.251,
"eval_wer": 143.92346010911677,
"step": 1000
},
{
"epoch": 7.71,
"learning_rate": 8.514285714285715e-06,
"loss": 0.1567,
"step": 1025
},
{
"epoch": 7.89,
"learning_rate": 8.442857142857144e-06,
"loss": 0.1648,
"step": 1050
},
{
"epoch": 8.08,
"learning_rate": 8.371428571428573e-06,
"loss": 0.1552,
"step": 1075
},
{
"epoch": 8.27,
"learning_rate": 8.3e-06,
"loss": 0.1317,
"step": 1100
},
{
"epoch": 8.46,
"learning_rate": 8.22857142857143e-06,
"loss": 0.1327,
"step": 1125
},
{
"epoch": 8.65,
"learning_rate": 8.157142857142858e-06,
"loss": 0.1272,
"step": 1150
},
{
"epoch": 8.83,
"learning_rate": 8.085714285714287e-06,
"loss": 0.131,
"step": 1175
},
{
"epoch": 9.02,
"learning_rate": 8.014285714285715e-06,
"loss": 0.0967,
"step": 1200
},
{
"epoch": 9.21,
"learning_rate": 7.942857142857144e-06,
"loss": 0.0836,
"step": 1225
},
{
"epoch": 9.4,
"learning_rate": 7.871428571428573e-06,
"loss": 0.0855,
"step": 1250
},
{
"epoch": 9.59,
"learning_rate": 7.800000000000002e-06,
"loss": 0.0852,
"step": 1275
},
{
"epoch": 9.77,
"learning_rate": 7.72857142857143e-06,
"loss": 0.086,
"step": 1300
},
{
"epoch": 9.96,
"learning_rate": 7.657142857142858e-06,
"loss": 0.0859,
"step": 1325
},
{
"epoch": 10.15,
"learning_rate": 7.585714285714286e-06,
"loss": 0.075,
"step": 1350
},
{
"epoch": 10.34,
"learning_rate": 7.514285714285715e-06,
"loss": 0.0712,
"step": 1375
},
{
"epoch": 10.53,
"learning_rate": 7.442857142857144e-06,
"loss": 0.069,
"step": 1400
},
{
"epoch": 10.71,
"learning_rate": 7.371428571428571e-06,
"loss": 0.0699,
"step": 1425
},
{
"epoch": 10.9,
"learning_rate": 7.3e-06,
"loss": 0.0694,
"step": 1450
},
{
"epoch": 11.09,
"learning_rate": 7.2285714285714294e-06,
"loss": 0.0635,
"step": 1475
},
{
"epoch": 11.28,
"learning_rate": 7.1571428571428584e-06,
"loss": 0.0611,
"step": 1500
},
{
"epoch": 11.47,
"learning_rate": 7.085714285714286e-06,
"loss": 0.0628,
"step": 1525
},
{
"epoch": 11.65,
"learning_rate": 7.014285714285715e-06,
"loss": 0.0546,
"step": 1550
},
{
"epoch": 11.84,
"learning_rate": 6.942857142857144e-06,
"loss": 0.0622,
"step": 1575
},
{
"epoch": 12.03,
"learning_rate": 6.871428571428572e-06,
"loss": 0.0592,
"step": 1600
},
{
"epoch": 12.22,
"learning_rate": 6.800000000000001e-06,
"loss": 0.0507,
"step": 1625
},
{
"epoch": 12.41,
"learning_rate": 6.728571428571429e-06,
"loss": 0.0591,
"step": 1650
},
{
"epoch": 12.59,
"learning_rate": 6.657142857142857e-06,
"loss": 0.0506,
"step": 1675
},
{
"epoch": 12.78,
"learning_rate": 6.585714285714286e-06,
"loss": 0.0519,
"step": 1700
},
{
"epoch": 12.97,
"learning_rate": 6.514285714285715e-06,
"loss": 0.0512,
"step": 1725
},
{
"epoch": 13.16,
"learning_rate": 6.442857142857143e-06,
"loss": 0.0491,
"step": 1750
},
{
"epoch": 13.35,
"learning_rate": 6.371428571428572e-06,
"loss": 0.0465,
"step": 1775
},
{
"epoch": 13.53,
"learning_rate": 6.300000000000001e-06,
"loss": 0.0439,
"step": 1800
},
{
"epoch": 13.72,
"learning_rate": 6.22857142857143e-06,
"loss": 0.0389,
"step": 1825
},
{
"epoch": 13.91,
"learning_rate": 6.157142857142858e-06,
"loss": 0.0412,
"step": 1850
},
{
"epoch": 14.1,
"learning_rate": 6.085714285714286e-06,
"loss": 0.0376,
"step": 1875
},
{
"epoch": 14.29,
"learning_rate": 6.014285714285715e-06,
"loss": 0.0333,
"step": 1900
},
{
"epoch": 14.47,
"learning_rate": 5.942857142857143e-06,
"loss": 0.0345,
"step": 1925
},
{
"epoch": 14.66,
"learning_rate": 5.871428571428572e-06,
"loss": 0.0329,
"step": 1950
},
{
"epoch": 14.85,
"learning_rate": 5.8e-06,
"loss": 0.0273,
"step": 1975
},
{
"epoch": 15.04,
"learning_rate": 5.7285714285714285e-06,
"loss": 0.0248,
"step": 2000
},
{
"epoch": 15.04,
"eval_loss": 1.180077075958252,
"eval_runtime": 132.654,
"eval_samples_per_second": 1.99,
"eval_steps_per_second": 0.249,
"eval_wer": 156.08444690440422,
"step": 2000
},
{
"epoch": 15.23,
"learning_rate": 5.6571428571428576e-06,
"loss": 0.0201,
"step": 2025
},
{
"epoch": 15.41,
"learning_rate": 5.5857142857142866e-06,
"loss": 0.0245,
"step": 2050
},
{
"epoch": 15.6,
"learning_rate": 5.514285714285714e-06,
"loss": 0.0204,
"step": 2075
},
{
"epoch": 15.79,
"learning_rate": 5.442857142857143e-06,
"loss": 0.0242,
"step": 2100
},
{
"epoch": 15.98,
"learning_rate": 5.371428571428572e-06,
"loss": 0.0188,
"step": 2125
},
{
"epoch": 16.17,
"learning_rate": 5.300000000000001e-06,
"loss": 0.0121,
"step": 2150
},
{
"epoch": 16.35,
"learning_rate": 5.22857142857143e-06,
"loss": 0.0105,
"step": 2175
},
{
"epoch": 16.54,
"learning_rate": 5.157142857142857e-06,
"loss": 0.0126,
"step": 2200
},
{
"epoch": 16.73,
"learning_rate": 5.085714285714286e-06,
"loss": 0.011,
"step": 2225
},
{
"epoch": 16.92,
"learning_rate": 5.014285714285715e-06,
"loss": 0.0134,
"step": 2250
},
{
"epoch": 17.11,
"learning_rate": 4.9428571428571435e-06,
"loss": 0.0079,
"step": 2275
},
{
"epoch": 17.29,
"learning_rate": 4.871428571428572e-06,
"loss": 0.0069,
"step": 2300
},
{
"epoch": 17.48,
"learning_rate": 4.800000000000001e-06,
"loss": 0.0049,
"step": 2325
},
{
"epoch": 17.67,
"learning_rate": 4.728571428571429e-06,
"loss": 0.0076,
"step": 2350
},
{
"epoch": 17.86,
"learning_rate": 4.657142857142857e-06,
"loss": 0.0056,
"step": 2375
},
{
"epoch": 18.05,
"learning_rate": 4.585714285714286e-06,
"loss": 0.0053,
"step": 2400
},
{
"epoch": 18.23,
"learning_rate": 4.514285714285714e-06,
"loss": 0.0037,
"step": 2425
},
{
"epoch": 18.42,
"learning_rate": 4.442857142857143e-06,
"loss": 0.0026,
"step": 2450
},
{
"epoch": 18.61,
"learning_rate": 4.371428571428572e-06,
"loss": 0.0034,
"step": 2475
},
{
"epoch": 18.8,
"learning_rate": 4.3e-06,
"loss": 0.0038,
"step": 2500
},
{
"epoch": 18.98,
"learning_rate": 4.228571428571429e-06,
"loss": 0.0035,
"step": 2525
},
{
"epoch": 19.17,
"learning_rate": 4.1571428571428575e-06,
"loss": 0.0038,
"step": 2550
},
{
"epoch": 19.36,
"learning_rate": 4.0857142857142865e-06,
"loss": 0.0021,
"step": 2575
},
{
"epoch": 19.55,
"learning_rate": 4.014285714285715e-06,
"loss": 0.0019,
"step": 2600
},
{
"epoch": 19.74,
"learning_rate": 3.942857142857143e-06,
"loss": 0.0032,
"step": 2625
},
{
"epoch": 19.92,
"learning_rate": 3.871428571428572e-06,
"loss": 0.0017,
"step": 2650
},
{
"epoch": 20.11,
"learning_rate": 3.8000000000000005e-06,
"loss": 0.0016,
"step": 2675
},
{
"epoch": 20.3,
"learning_rate": 3.7285714285714286e-06,
"loss": 0.0018,
"step": 2700
},
{
"epoch": 20.49,
"learning_rate": 3.6571428571428576e-06,
"loss": 0.0024,
"step": 2725
},
{
"epoch": 20.68,
"learning_rate": 3.5857142857142862e-06,
"loss": 0.0017,
"step": 2750
},
{
"epoch": 20.86,
"learning_rate": 3.5142857142857144e-06,
"loss": 0.0028,
"step": 2775
},
{
"epoch": 21.05,
"learning_rate": 3.4428571428571434e-06,
"loss": 0.0015,
"step": 2800
},
{
"epoch": 21.24,
"learning_rate": 3.3714285714285716e-06,
"loss": 0.0012,
"step": 2825
},
{
"epoch": 21.43,
"learning_rate": 3.3000000000000006e-06,
"loss": 0.0017,
"step": 2850
},
{
"epoch": 21.62,
"learning_rate": 3.2285714285714288e-06,
"loss": 0.0024,
"step": 2875
},
{
"epoch": 21.8,
"learning_rate": 3.1571428571428573e-06,
"loss": 0.0016,
"step": 2900
},
{
"epoch": 21.99,
"learning_rate": 3.085714285714286e-06,
"loss": 0.0014,
"step": 2925
},
{
"epoch": 22.18,
"learning_rate": 3.0142857142857145e-06,
"loss": 0.0011,
"step": 2950
},
{
"epoch": 22.37,
"learning_rate": 2.9428571428571427e-06,
"loss": 0.0012,
"step": 2975
},
{
"epoch": 22.56,
"learning_rate": 2.8714285714285717e-06,
"loss": 0.0014,
"step": 3000
},
{
"epoch": 22.56,
"eval_loss": 1.2573596239089966,
"eval_runtime": 130.1823,
"eval_samples_per_second": 2.028,
"eval_steps_per_second": 0.253,
"eval_wer": 163.200759073298,
"step": 3000
},
{
"epoch": 22.74,
"learning_rate": 2.8000000000000003e-06,
"loss": 0.0013,
"step": 3025
},
{
"epoch": 22.93,
"learning_rate": 2.728571428571429e-06,
"loss": 0.0013,
"step": 3050
},
{
"epoch": 23.12,
"learning_rate": 2.6571428571428575e-06,
"loss": 0.0015,
"step": 3075
},
{
"epoch": 23.31,
"learning_rate": 2.5857142857142856e-06,
"loss": 0.0019,
"step": 3100
},
{
"epoch": 23.5,
"learning_rate": 2.5142857142857147e-06,
"loss": 0.0009,
"step": 3125
},
{
"epoch": 23.68,
"learning_rate": 2.442857142857143e-06,
"loss": 0.0013,
"step": 3150
},
{
"epoch": 23.87,
"learning_rate": 2.371428571428572e-06,
"loss": 0.001,
"step": 3175
},
{
"epoch": 24.06,
"learning_rate": 2.3000000000000004e-06,
"loss": 0.0013,
"step": 3200
},
{
"epoch": 24.25,
"learning_rate": 2.228571428571429e-06,
"loss": 0.0009,
"step": 3225
},
{
"epoch": 24.44,
"learning_rate": 2.157142857142857e-06,
"loss": 0.0011,
"step": 3250
},
{
"epoch": 24.62,
"learning_rate": 2.0857142857142858e-06,
"loss": 0.0009,
"step": 3275
},
{
"epoch": 24.81,
"learning_rate": 2.0142857142857144e-06,
"loss": 0.001,
"step": 3300
},
{
"epoch": 25.0,
"learning_rate": 1.942857142857143e-06,
"loss": 0.001,
"step": 3325
},
{
"epoch": 25.19,
"learning_rate": 1.8714285714285715e-06,
"loss": 0.0011,
"step": 3350
},
{
"epoch": 25.38,
"learning_rate": 1.8000000000000001e-06,
"loss": 0.0008,
"step": 3375
},
{
"epoch": 25.56,
"learning_rate": 1.7285714285714287e-06,
"loss": 0.0008,
"step": 3400
},
{
"epoch": 25.75,
"learning_rate": 1.657142857142857e-06,
"loss": 0.0013,
"step": 3425
},
{
"epoch": 25.94,
"learning_rate": 1.5857142857142857e-06,
"loss": 0.0009,
"step": 3450
},
{
"epoch": 26.13,
"learning_rate": 1.5142857142857145e-06,
"loss": 0.0008,
"step": 3475
},
{
"epoch": 26.32,
"learning_rate": 1.442857142857143e-06,
"loss": 0.0009,
"step": 3500
},
{
"epoch": 26.5,
"learning_rate": 1.3714285714285717e-06,
"loss": 0.0008,
"step": 3525
},
{
"epoch": 26.69,
"learning_rate": 1.3e-06,
"loss": 0.0011,
"step": 3550
},
{
"epoch": 26.88,
"learning_rate": 1.2285714285714286e-06,
"loss": 0.0008,
"step": 3575
},
{
"epoch": 27.07,
"learning_rate": 1.1571428571428572e-06,
"loss": 0.0011,
"step": 3600
},
{
"epoch": 27.26,
"learning_rate": 1.0857142857142858e-06,
"loss": 0.0013,
"step": 3625
},
{
"epoch": 27.44,
"learning_rate": 1.0142857142857144e-06,
"loss": 0.0008,
"step": 3650
},
{
"epoch": 27.63,
"learning_rate": 9.42857142857143e-07,
"loss": 0.0008,
"step": 3675
},
{
"epoch": 27.82,
"learning_rate": 8.714285714285716e-07,
"loss": 0.0008,
"step": 3700
},
{
"epoch": 28.01,
"learning_rate": 8.000000000000001e-07,
"loss": 0.0007,
"step": 3725
},
{
"epoch": 28.2,
"learning_rate": 7.285714285714287e-07,
"loss": 0.0008,
"step": 3750
},
{
"epoch": 28.38,
"learning_rate": 6.571428571428571e-07,
"loss": 0.001,
"step": 3775
},
{
"epoch": 28.57,
"learning_rate": 5.857142857142857e-07,
"loss": 0.0009,
"step": 3800
},
{
"epoch": 28.76,
"learning_rate": 5.142857142857143e-07,
"loss": 0.0007,
"step": 3825
},
{
"epoch": 28.95,
"learning_rate": 4.4285714285714286e-07,
"loss": 0.0008,
"step": 3850
},
{
"epoch": 29.14,
"learning_rate": 3.7142857142857145e-07,
"loss": 0.0007,
"step": 3875
},
{
"epoch": 29.32,
"learning_rate": 3.0000000000000004e-07,
"loss": 0.0007,
"step": 3900
},
{
"epoch": 29.51,
"learning_rate": 2.285714285714286e-07,
"loss": 0.0007,
"step": 3925
},
{
"epoch": 29.7,
"learning_rate": 1.5714285714285717e-07,
"loss": 0.001,
"step": 3950
},
{
"epoch": 29.89,
"learning_rate": 8.571428571428573e-08,
"loss": 0.0009,
"step": 3975
},
{
"epoch": 30.08,
"learning_rate": 1.4285714285714288e-08,
"loss": 0.0008,
"step": 4000
},
{
"epoch": 30.08,
"eval_loss": 1.2812294960021973,
"eval_runtime": 131.9154,
"eval_samples_per_second": 2.001,
"eval_steps_per_second": 0.25,
"eval_wer": 164.2682058986321,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 31,
"total_flos": 9.1741298761728e+18,
"trial_name": null,
"trial_params": null
}