humeur's picture
Training in progress, step 4000
3a55d77
raw
history blame
20.4 kB
{
"best_metric": 19.186699618063358,
"best_model_checkpoint": "./whisper-small-sv-en/checkpoint-3000",
"epoch": 5.174644243208279,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 4.2000000000000006e-07,
"loss": 4.8208,
"step": 25
},
{
"epoch": 0.06,
"learning_rate": 9.200000000000001e-07,
"loss": 3.8662,
"step": 50
},
{
"epoch": 0.1,
"learning_rate": 1.42e-06,
"loss": 2.7859,
"step": 75
},
{
"epoch": 0.13,
"learning_rate": 1.9200000000000003e-06,
"loss": 1.7098,
"step": 100
},
{
"epoch": 0.16,
"learning_rate": 2.42e-06,
"loss": 1.197,
"step": 125
},
{
"epoch": 0.19,
"learning_rate": 2.92e-06,
"loss": 1.0401,
"step": 150
},
{
"epoch": 0.23,
"learning_rate": 3.4200000000000007e-06,
"loss": 0.9443,
"step": 175
},
{
"epoch": 0.26,
"learning_rate": 3.920000000000001e-06,
"loss": 0.8663,
"step": 200
},
{
"epoch": 0.29,
"learning_rate": 4.42e-06,
"loss": 0.7198,
"step": 225
},
{
"epoch": 0.32,
"learning_rate": 4.92e-06,
"loss": 0.6317,
"step": 250
},
{
"epoch": 0.36,
"learning_rate": 5.420000000000001e-06,
"loss": 0.509,
"step": 275
},
{
"epoch": 0.39,
"learning_rate": 5.92e-06,
"loss": 0.3145,
"step": 300
},
{
"epoch": 0.42,
"learning_rate": 6.42e-06,
"loss": 0.3392,
"step": 325
},
{
"epoch": 0.45,
"learning_rate": 6.92e-06,
"loss": 0.2963,
"step": 350
},
{
"epoch": 0.49,
"learning_rate": 7.420000000000001e-06,
"loss": 0.3166,
"step": 375
},
{
"epoch": 0.52,
"learning_rate": 7.92e-06,
"loss": 0.3222,
"step": 400
},
{
"epoch": 0.55,
"learning_rate": 8.42e-06,
"loss": 0.3137,
"step": 425
},
{
"epoch": 0.58,
"learning_rate": 8.920000000000001e-06,
"loss": 0.305,
"step": 450
},
{
"epoch": 0.61,
"learning_rate": 9.42e-06,
"loss": 0.2771,
"step": 475
},
{
"epoch": 0.65,
"learning_rate": 9.920000000000002e-06,
"loss": 0.3114,
"step": 500
},
{
"epoch": 0.68,
"learning_rate": 9.940000000000001e-06,
"loss": 0.2985,
"step": 525
},
{
"epoch": 0.71,
"learning_rate": 9.86857142857143e-06,
"loss": 0.2943,
"step": 550
},
{
"epoch": 0.74,
"learning_rate": 9.797142857142858e-06,
"loss": 0.3044,
"step": 575
},
{
"epoch": 0.78,
"learning_rate": 9.725714285714287e-06,
"loss": 0.3105,
"step": 600
},
{
"epoch": 0.81,
"learning_rate": 9.654285714285716e-06,
"loss": 0.2937,
"step": 625
},
{
"epoch": 0.84,
"learning_rate": 9.582857142857143e-06,
"loss": 0.3074,
"step": 650
},
{
"epoch": 0.87,
"learning_rate": 9.511428571428572e-06,
"loss": 0.274,
"step": 675
},
{
"epoch": 0.91,
"learning_rate": 9.440000000000001e-06,
"loss": 0.2884,
"step": 700
},
{
"epoch": 0.94,
"learning_rate": 9.368571428571428e-06,
"loss": 0.292,
"step": 725
},
{
"epoch": 0.97,
"learning_rate": 9.297142857142857e-06,
"loss": 0.286,
"step": 750
},
{
"epoch": 1.0,
"learning_rate": 9.225714285714286e-06,
"loss": 0.2801,
"step": 775
},
{
"epoch": 1.03,
"learning_rate": 9.154285714285715e-06,
"loss": 0.1413,
"step": 800
},
{
"epoch": 1.07,
"learning_rate": 9.082857142857143e-06,
"loss": 0.1485,
"step": 825
},
{
"epoch": 1.1,
"learning_rate": 9.011428571428572e-06,
"loss": 0.1515,
"step": 850
},
{
"epoch": 1.13,
"learning_rate": 8.94e-06,
"loss": 0.1314,
"step": 875
},
{
"epoch": 1.16,
"learning_rate": 8.86857142857143e-06,
"loss": 0.1528,
"step": 900
},
{
"epoch": 1.2,
"learning_rate": 8.797142857142857e-06,
"loss": 0.1579,
"step": 925
},
{
"epoch": 1.23,
"learning_rate": 8.725714285714286e-06,
"loss": 0.1537,
"step": 950
},
{
"epoch": 1.26,
"learning_rate": 8.654285714285715e-06,
"loss": 0.1498,
"step": 975
},
{
"epoch": 1.29,
"learning_rate": 8.582857142857144e-06,
"loss": 0.1398,
"step": 1000
},
{
"epoch": 1.29,
"eval_loss": 0.2894088923931122,
"eval_runtime": 2356.7832,
"eval_samples_per_second": 1.273,
"eval_steps_per_second": 0.159,
"eval_wer": 21.154796674904517,
"step": 1000
},
{
"epoch": 1.33,
"learning_rate": 8.511428571428571e-06,
"loss": 0.1565,
"step": 1025
},
{
"epoch": 1.36,
"learning_rate": 8.44e-06,
"loss": 0.1416,
"step": 1050
},
{
"epoch": 1.39,
"learning_rate": 8.36857142857143e-06,
"loss": 0.1397,
"step": 1075
},
{
"epoch": 1.42,
"learning_rate": 8.297142857142859e-06,
"loss": 0.1514,
"step": 1100
},
{
"epoch": 1.46,
"learning_rate": 8.225714285714288e-06,
"loss": 0.1418,
"step": 1125
},
{
"epoch": 1.49,
"learning_rate": 8.154285714285715e-06,
"loss": 0.1448,
"step": 1150
},
{
"epoch": 1.52,
"learning_rate": 8.082857142857144e-06,
"loss": 0.141,
"step": 1175
},
{
"epoch": 1.55,
"learning_rate": 8.011428571428573e-06,
"loss": 0.1558,
"step": 1200
},
{
"epoch": 1.58,
"learning_rate": 7.94e-06,
"loss": 0.1413,
"step": 1225
},
{
"epoch": 1.62,
"learning_rate": 7.86857142857143e-06,
"loss": 0.1337,
"step": 1250
},
{
"epoch": 1.65,
"learning_rate": 7.797142857142858e-06,
"loss": 0.1434,
"step": 1275
},
{
"epoch": 1.68,
"learning_rate": 7.725714285714286e-06,
"loss": 0.1314,
"step": 1300
},
{
"epoch": 1.71,
"learning_rate": 7.654285714285715e-06,
"loss": 0.1465,
"step": 1325
},
{
"epoch": 1.75,
"learning_rate": 7.5828571428571444e-06,
"loss": 0.1267,
"step": 1350
},
{
"epoch": 1.78,
"learning_rate": 7.511428571428572e-06,
"loss": 0.1541,
"step": 1375
},
{
"epoch": 1.81,
"learning_rate": 7.440000000000001e-06,
"loss": 0.14,
"step": 1400
},
{
"epoch": 1.84,
"learning_rate": 7.36857142857143e-06,
"loss": 0.1309,
"step": 1425
},
{
"epoch": 1.88,
"learning_rate": 7.297142857142858e-06,
"loss": 0.1262,
"step": 1450
},
{
"epoch": 1.91,
"learning_rate": 7.225714285714286e-06,
"loss": 0.1466,
"step": 1475
},
{
"epoch": 1.94,
"learning_rate": 7.154285714285715e-06,
"loss": 0.1336,
"step": 1500
},
{
"epoch": 1.97,
"learning_rate": 7.082857142857143e-06,
"loss": 0.145,
"step": 1525
},
{
"epoch": 2.01,
"learning_rate": 7.011428571428572e-06,
"loss": 0.1124,
"step": 1550
},
{
"epoch": 2.04,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.0503,
"step": 1575
},
{
"epoch": 2.07,
"learning_rate": 6.868571428571429e-06,
"loss": 0.0518,
"step": 1600
},
{
"epoch": 2.1,
"learning_rate": 6.797142857142858e-06,
"loss": 0.0553,
"step": 1625
},
{
"epoch": 2.13,
"learning_rate": 6.725714285714287e-06,
"loss": 0.0553,
"step": 1650
},
{
"epoch": 2.17,
"learning_rate": 6.654285714285716e-06,
"loss": 0.0578,
"step": 1675
},
{
"epoch": 2.2,
"learning_rate": 6.582857142857143e-06,
"loss": 0.0576,
"step": 1700
},
{
"epoch": 2.23,
"learning_rate": 6.511428571428572e-06,
"loss": 0.0509,
"step": 1725
},
{
"epoch": 2.26,
"learning_rate": 6.440000000000001e-06,
"loss": 0.0525,
"step": 1750
},
{
"epoch": 2.3,
"learning_rate": 6.368571428571429e-06,
"loss": 0.0548,
"step": 1775
},
{
"epoch": 2.33,
"learning_rate": 6.297142857142857e-06,
"loss": 0.0479,
"step": 1800
},
{
"epoch": 2.36,
"learning_rate": 6.225714285714286e-06,
"loss": 0.0538,
"step": 1825
},
{
"epoch": 2.39,
"learning_rate": 6.1542857142857145e-06,
"loss": 0.0528,
"step": 1850
},
{
"epoch": 2.43,
"learning_rate": 6.0828571428571435e-06,
"loss": 0.0532,
"step": 1875
},
{
"epoch": 2.46,
"learning_rate": 6.011428571428572e-06,
"loss": 0.0492,
"step": 1900
},
{
"epoch": 2.49,
"learning_rate": 5.94e-06,
"loss": 0.0527,
"step": 1925
},
{
"epoch": 2.52,
"learning_rate": 5.868571428571429e-06,
"loss": 0.0463,
"step": 1950
},
{
"epoch": 2.55,
"learning_rate": 5.797142857142858e-06,
"loss": 0.051,
"step": 1975
},
{
"epoch": 2.59,
"learning_rate": 5.725714285714287e-06,
"loss": 0.0476,
"step": 2000
},
{
"epoch": 2.59,
"eval_loss": 0.28524845838546753,
"eval_runtime": 2654.5637,
"eval_samples_per_second": 1.13,
"eval_steps_per_second": 0.141,
"eval_wer": 19.829251853516062,
"step": 2000
},
{
"epoch": 2.62,
"learning_rate": 5.654285714285714e-06,
"loss": 0.0456,
"step": 2025
},
{
"epoch": 2.65,
"learning_rate": 5.582857142857143e-06,
"loss": 0.0525,
"step": 2050
},
{
"epoch": 2.68,
"learning_rate": 5.511428571428572e-06,
"loss": 0.0506,
"step": 2075
},
{
"epoch": 2.72,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.0557,
"step": 2100
},
{
"epoch": 2.75,
"learning_rate": 5.368571428571429e-06,
"loss": 0.057,
"step": 2125
},
{
"epoch": 2.78,
"learning_rate": 5.297142857142858e-06,
"loss": 0.0572,
"step": 2150
},
{
"epoch": 2.81,
"learning_rate": 5.225714285714286e-06,
"loss": 0.0445,
"step": 2175
},
{
"epoch": 2.85,
"learning_rate": 5.154285714285715e-06,
"loss": 0.0552,
"step": 2200
},
{
"epoch": 2.88,
"learning_rate": 5.082857142857144e-06,
"loss": 0.0511,
"step": 2225
},
{
"epoch": 2.91,
"learning_rate": 5.011428571428571e-06,
"loss": 0.0455,
"step": 2250
},
{
"epoch": 2.94,
"learning_rate": 4.94e-06,
"loss": 0.0559,
"step": 2275
},
{
"epoch": 2.98,
"learning_rate": 4.868571428571429e-06,
"loss": 0.0493,
"step": 2300
},
{
"epoch": 3.01,
"learning_rate": 4.797142857142857e-06,
"loss": 0.05,
"step": 2325
},
{
"epoch": 3.04,
"learning_rate": 4.725714285714286e-06,
"loss": 0.0214,
"step": 2350
},
{
"epoch": 3.07,
"learning_rate": 4.6542857142857145e-06,
"loss": 0.0244,
"step": 2375
},
{
"epoch": 3.1,
"learning_rate": 4.5828571428571435e-06,
"loss": 0.0209,
"step": 2400
},
{
"epoch": 3.14,
"learning_rate": 4.511428571428572e-06,
"loss": 0.0224,
"step": 2425
},
{
"epoch": 3.17,
"learning_rate": 4.440000000000001e-06,
"loss": 0.0253,
"step": 2450
},
{
"epoch": 3.2,
"learning_rate": 4.368571428571429e-06,
"loss": 0.0228,
"step": 2475
},
{
"epoch": 3.23,
"learning_rate": 4.297142857142858e-06,
"loss": 0.0232,
"step": 2500
},
{
"epoch": 3.27,
"learning_rate": 4.225714285714286e-06,
"loss": 0.02,
"step": 2525
},
{
"epoch": 3.3,
"learning_rate": 4.154285714285714e-06,
"loss": 0.0214,
"step": 2550
},
{
"epoch": 3.33,
"learning_rate": 4.082857142857143e-06,
"loss": 0.0201,
"step": 2575
},
{
"epoch": 3.36,
"learning_rate": 4.011428571428571e-06,
"loss": 0.0178,
"step": 2600
},
{
"epoch": 3.4,
"learning_rate": 3.94e-06,
"loss": 0.0221,
"step": 2625
},
{
"epoch": 3.43,
"learning_rate": 3.8685714285714286e-06,
"loss": 0.0199,
"step": 2650
},
{
"epoch": 3.46,
"learning_rate": 3.7971428571428576e-06,
"loss": 0.0182,
"step": 2675
},
{
"epoch": 3.49,
"learning_rate": 3.7257142857142857e-06,
"loss": 0.0197,
"step": 2700
},
{
"epoch": 3.53,
"learning_rate": 3.6542857142857148e-06,
"loss": 0.0213,
"step": 2725
},
{
"epoch": 3.56,
"learning_rate": 3.582857142857143e-06,
"loss": 0.0212,
"step": 2750
},
{
"epoch": 3.59,
"learning_rate": 3.511428571428572e-06,
"loss": 0.0195,
"step": 2775
},
{
"epoch": 3.62,
"learning_rate": 3.44e-06,
"loss": 0.019,
"step": 2800
},
{
"epoch": 3.65,
"learning_rate": 3.3685714285714287e-06,
"loss": 0.019,
"step": 2825
},
{
"epoch": 3.69,
"learning_rate": 3.2971428571428577e-06,
"loss": 0.0221,
"step": 2850
},
{
"epoch": 3.72,
"learning_rate": 3.225714285714286e-06,
"loss": 0.0207,
"step": 2875
},
{
"epoch": 3.75,
"learning_rate": 3.154285714285715e-06,
"loss": 0.0228,
"step": 2900
},
{
"epoch": 3.78,
"learning_rate": 3.082857142857143e-06,
"loss": 0.0201,
"step": 2925
},
{
"epoch": 3.82,
"learning_rate": 3.0114285714285716e-06,
"loss": 0.0199,
"step": 2950
},
{
"epoch": 3.85,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.0198,
"step": 2975
},
{
"epoch": 3.88,
"learning_rate": 2.868571428571429e-06,
"loss": 0.019,
"step": 3000
},
{
"epoch": 3.88,
"eval_loss": 0.2968597114086151,
"eval_runtime": 2272.5975,
"eval_samples_per_second": 1.32,
"eval_steps_per_second": 0.165,
"eval_wer": 19.186699618063358,
"step": 3000
},
{
"epoch": 3.91,
"learning_rate": 2.797142857142857e-06,
"loss": 0.0224,
"step": 3025
},
{
"epoch": 3.95,
"learning_rate": 2.725714285714286e-06,
"loss": 0.02,
"step": 3050
},
{
"epoch": 3.98,
"learning_rate": 2.654285714285714e-06,
"loss": 0.0233,
"step": 3075
},
{
"epoch": 4.01,
"learning_rate": 2.582857142857143e-06,
"loss": 0.0176,
"step": 3100
},
{
"epoch": 4.04,
"learning_rate": 2.5114285714285718e-06,
"loss": 0.0106,
"step": 3125
},
{
"epoch": 4.08,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.0083,
"step": 3150
},
{
"epoch": 4.11,
"learning_rate": 2.3685714285714285e-06,
"loss": 0.008,
"step": 3175
},
{
"epoch": 4.14,
"learning_rate": 2.297142857142857e-06,
"loss": 0.0083,
"step": 3200
},
{
"epoch": 4.17,
"learning_rate": 2.2257142857142857e-06,
"loss": 0.0067,
"step": 3225
},
{
"epoch": 4.2,
"learning_rate": 2.1542857142857147e-06,
"loss": 0.0076,
"step": 3250
},
{
"epoch": 4.24,
"learning_rate": 2.0828571428571433e-06,
"loss": 0.0103,
"step": 3275
},
{
"epoch": 4.27,
"learning_rate": 2.0114285714285715e-06,
"loss": 0.008,
"step": 3300
},
{
"epoch": 4.3,
"learning_rate": 1.94e-06,
"loss": 0.0088,
"step": 3325
},
{
"epoch": 4.33,
"learning_rate": 1.8685714285714289e-06,
"loss": 0.0082,
"step": 3350
},
{
"epoch": 4.37,
"learning_rate": 1.7971428571428572e-06,
"loss": 0.0085,
"step": 3375
},
{
"epoch": 4.4,
"learning_rate": 1.7257142857142858e-06,
"loss": 0.0082,
"step": 3400
},
{
"epoch": 4.43,
"learning_rate": 1.6542857142857144e-06,
"loss": 0.0082,
"step": 3425
},
{
"epoch": 4.46,
"learning_rate": 1.582857142857143e-06,
"loss": 0.0079,
"step": 3450
},
{
"epoch": 4.5,
"learning_rate": 1.5114285714285714e-06,
"loss": 0.0081,
"step": 3475
},
{
"epoch": 4.53,
"learning_rate": 1.44e-06,
"loss": 0.0087,
"step": 3500
},
{
"epoch": 4.56,
"learning_rate": 1.3685714285714286e-06,
"loss": 0.0063,
"step": 3525
},
{
"epoch": 4.59,
"learning_rate": 1.2971428571428574e-06,
"loss": 0.0073,
"step": 3550
},
{
"epoch": 4.62,
"learning_rate": 1.2257142857142857e-06,
"loss": 0.0079,
"step": 3575
},
{
"epoch": 4.66,
"learning_rate": 1.1542857142857143e-06,
"loss": 0.0076,
"step": 3600
},
{
"epoch": 4.69,
"learning_rate": 1.082857142857143e-06,
"loss": 0.0093,
"step": 3625
},
{
"epoch": 4.72,
"learning_rate": 1.0114285714285715e-06,
"loss": 0.0073,
"step": 3650
},
{
"epoch": 4.75,
"learning_rate": 9.400000000000001e-07,
"loss": 0.0066,
"step": 3675
},
{
"epoch": 4.79,
"learning_rate": 8.685714285714286e-07,
"loss": 0.0083,
"step": 3700
},
{
"epoch": 4.82,
"learning_rate": 7.971428571428572e-07,
"loss": 0.0065,
"step": 3725
},
{
"epoch": 4.85,
"learning_rate": 7.257142857142857e-07,
"loss": 0.008,
"step": 3750
},
{
"epoch": 4.88,
"learning_rate": 6.542857142857144e-07,
"loss": 0.0076,
"step": 3775
},
{
"epoch": 4.92,
"learning_rate": 5.82857142857143e-07,
"loss": 0.0069,
"step": 3800
},
{
"epoch": 4.95,
"learning_rate": 5.114285714285714e-07,
"loss": 0.0069,
"step": 3825
},
{
"epoch": 4.98,
"learning_rate": 4.4e-07,
"loss": 0.008,
"step": 3850
},
{
"epoch": 5.01,
"learning_rate": 3.685714285714286e-07,
"loss": 0.0054,
"step": 3875
},
{
"epoch": 5.05,
"learning_rate": 2.9714285714285715e-07,
"loss": 0.0043,
"step": 3900
},
{
"epoch": 5.08,
"learning_rate": 2.2571428571428574e-07,
"loss": 0.0043,
"step": 3925
},
{
"epoch": 5.11,
"learning_rate": 1.542857142857143e-07,
"loss": 0.0041,
"step": 3950
},
{
"epoch": 5.14,
"learning_rate": 8.285714285714285e-08,
"loss": 0.0043,
"step": 3975
},
{
"epoch": 5.17,
"learning_rate": 1.142857142857143e-08,
"loss": 0.0046,
"step": 4000
},
{
"epoch": 5.17,
"eval_loss": 0.3207452893257141,
"eval_runtime": 2511.1438,
"eval_samples_per_second": 1.195,
"eval_steps_per_second": 0.149,
"eval_wer": 19.200179734891034,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 6,
"total_flos": 1.84579222044672e+19,
"trial_name": null,
"trial_params": null
}