whisper-largev2-mls-dutch / trainer_state.json
sgangireddy's picture
End of training
fa5a4c4
{
"best_metric": 10.591602311347534,
"best_model_checkpoint": "./checkpoint-4000",
"epoch": 1.0,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 4.4e-07,
"loss": 0.9971,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 9.400000000000001e-07,
"loss": 0.8605,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 1.44e-06,
"loss": 0.4011,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 1.94e-06,
"loss": 0.3205,
"step": 100
},
{
"epoch": 0.03,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.4209,
"step": 125
},
{
"epoch": 0.04,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.4295,
"step": 150
},
{
"epoch": 0.04,
"learning_rate": 3.44e-06,
"loss": 0.2947,
"step": 175
},
{
"epoch": 0.05,
"learning_rate": 3.94e-06,
"loss": 0.3056,
"step": 200
},
{
"epoch": 0.06,
"learning_rate": 4.440000000000001e-06,
"loss": 0.3778,
"step": 225
},
{
"epoch": 0.06,
"learning_rate": 4.94e-06,
"loss": 0.2793,
"step": 250
},
{
"epoch": 0.07,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.2478,
"step": 275
},
{
"epoch": 0.07,
"learning_rate": 5.94e-06,
"loss": 0.4359,
"step": 300
},
{
"epoch": 0.08,
"learning_rate": 6.440000000000001e-06,
"loss": 0.3597,
"step": 325
},
{
"epoch": 0.09,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.2937,
"step": 350
},
{
"epoch": 0.09,
"learning_rate": 7.440000000000001e-06,
"loss": 0.1855,
"step": 375
},
{
"epoch": 0.1,
"learning_rate": 7.94e-06,
"loss": 0.1392,
"step": 400
},
{
"epoch": 0.11,
"learning_rate": 8.44e-06,
"loss": 0.2429,
"step": 425
},
{
"epoch": 0.11,
"learning_rate": 8.94e-06,
"loss": 0.2948,
"step": 450
},
{
"epoch": 0.12,
"learning_rate": 9.440000000000001e-06,
"loss": 0.3468,
"step": 475
},
{
"epoch": 0.12,
"learning_rate": 9.940000000000001e-06,
"loss": 0.2761,
"step": 500
},
{
"epoch": 0.13,
"learning_rate": 9.937142857142858e-06,
"loss": 0.258,
"step": 525
},
{
"epoch": 0.14,
"learning_rate": 9.865714285714285e-06,
"loss": 0.1773,
"step": 550
},
{
"epoch": 0.14,
"learning_rate": 9.794285714285714e-06,
"loss": 0.1902,
"step": 575
},
{
"epoch": 0.15,
"learning_rate": 9.722857142857143e-06,
"loss": 0.1806,
"step": 600
},
{
"epoch": 0.16,
"learning_rate": 9.651428571428572e-06,
"loss": 0.2258,
"step": 625
},
{
"epoch": 0.16,
"learning_rate": 9.58e-06,
"loss": 0.2519,
"step": 650
},
{
"epoch": 0.17,
"learning_rate": 9.508571428571429e-06,
"loss": 0.1608,
"step": 675
},
{
"epoch": 0.17,
"learning_rate": 9.437142857142858e-06,
"loss": 0.1346,
"step": 700
},
{
"epoch": 0.18,
"learning_rate": 9.365714285714287e-06,
"loss": 0.2203,
"step": 725
},
{
"epoch": 0.19,
"learning_rate": 9.294285714285714e-06,
"loss": 0.2389,
"step": 750
},
{
"epoch": 0.19,
"learning_rate": 9.222857142857143e-06,
"loss": 0.2412,
"step": 775
},
{
"epoch": 0.2,
"learning_rate": 9.151428571428572e-06,
"loss": 0.1783,
"step": 800
},
{
"epoch": 0.21,
"learning_rate": 9.080000000000001e-06,
"loss": 0.1925,
"step": 825
},
{
"epoch": 0.21,
"learning_rate": 9.00857142857143e-06,
"loss": 0.2095,
"step": 850
},
{
"epoch": 0.22,
"learning_rate": 8.937142857142857e-06,
"loss": 0.1664,
"step": 875
},
{
"epoch": 0.23,
"learning_rate": 8.865714285714287e-06,
"loss": 0.134,
"step": 900
},
{
"epoch": 0.23,
"learning_rate": 8.794285714285716e-06,
"loss": 0.1974,
"step": 925
},
{
"epoch": 0.24,
"learning_rate": 8.722857142857145e-06,
"loss": 0.1902,
"step": 950
},
{
"epoch": 0.24,
"learning_rate": 8.651428571428572e-06,
"loss": 0.1498,
"step": 975
},
{
"epoch": 0.25,
"learning_rate": 8.580000000000001e-06,
"loss": 0.2515,
"step": 1000
},
{
"epoch": 0.25,
"eval_loss": 0.2578780949115753,
"eval_runtime": 10032.042,
"eval_samples_per_second": 0.307,
"eval_steps_per_second": 0.019,
"eval_wer": 12.977609792896416,
"step": 1000
},
{
"epoch": 0.26,
"learning_rate": 8.50857142857143e-06,
"loss": 0.2639,
"step": 1025
},
{
"epoch": 0.26,
"learning_rate": 8.437142857142859e-06,
"loss": 0.166,
"step": 1050
},
{
"epoch": 0.27,
"learning_rate": 8.365714285714286e-06,
"loss": 0.1287,
"step": 1075
},
{
"epoch": 0.28,
"learning_rate": 8.294285714285715e-06,
"loss": 0.2459,
"step": 1100
},
{
"epoch": 0.28,
"learning_rate": 8.222857142857144e-06,
"loss": 0.1961,
"step": 1125
},
{
"epoch": 0.29,
"learning_rate": 8.151428571428572e-06,
"loss": 0.1865,
"step": 1150
},
{
"epoch": 0.29,
"learning_rate": 8.08e-06,
"loss": 0.1367,
"step": 1175
},
{
"epoch": 0.3,
"learning_rate": 8.00857142857143e-06,
"loss": 0.1546,
"step": 1200
},
{
"epoch": 0.31,
"learning_rate": 7.94e-06,
"loss": 0.145,
"step": 1225
},
{
"epoch": 0.31,
"learning_rate": 7.86857142857143e-06,
"loss": 0.2315,
"step": 1250
},
{
"epoch": 0.32,
"learning_rate": 7.797142857142858e-06,
"loss": 0.1752,
"step": 1275
},
{
"epoch": 0.33,
"learning_rate": 7.725714285714286e-06,
"loss": 0.1931,
"step": 1300
},
{
"epoch": 0.33,
"learning_rate": 7.654285714285715e-06,
"loss": 0.1847,
"step": 1325
},
{
"epoch": 0.34,
"learning_rate": 7.5828571428571444e-06,
"loss": 0.2739,
"step": 1350
},
{
"epoch": 0.34,
"learning_rate": 7.511428571428572e-06,
"loss": 0.2192,
"step": 1375
},
{
"epoch": 0.35,
"learning_rate": 7.440000000000001e-06,
"loss": 0.2253,
"step": 1400
},
{
"epoch": 0.36,
"learning_rate": 7.36857142857143e-06,
"loss": 0.2854,
"step": 1425
},
{
"epoch": 0.36,
"learning_rate": 7.297142857142858e-06,
"loss": 0.2926,
"step": 1450
},
{
"epoch": 0.37,
"learning_rate": 7.225714285714286e-06,
"loss": 0.2575,
"step": 1475
},
{
"epoch": 0.38,
"learning_rate": 7.154285714285715e-06,
"loss": 0.1938,
"step": 1500
},
{
"epoch": 0.38,
"learning_rate": 7.082857142857143e-06,
"loss": 0.2697,
"step": 1525
},
{
"epoch": 0.39,
"learning_rate": 7.011428571428572e-06,
"loss": 0.2491,
"step": 1550
},
{
"epoch": 0.39,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.1908,
"step": 1575
},
{
"epoch": 0.4,
"learning_rate": 6.868571428571429e-06,
"loss": 0.1716,
"step": 1600
},
{
"epoch": 0.41,
"learning_rate": 6.797142857142858e-06,
"loss": 0.1867,
"step": 1625
},
{
"epoch": 0.41,
"learning_rate": 6.725714285714287e-06,
"loss": 0.2076,
"step": 1650
},
{
"epoch": 0.42,
"learning_rate": 6.654285714285716e-06,
"loss": 0.2138,
"step": 1675
},
{
"epoch": 0.42,
"learning_rate": 6.582857142857143e-06,
"loss": 0.2347,
"step": 1700
},
{
"epoch": 0.43,
"learning_rate": 6.511428571428572e-06,
"loss": 0.2331,
"step": 1725
},
{
"epoch": 0.44,
"learning_rate": 6.440000000000001e-06,
"loss": 0.1837,
"step": 1750
},
{
"epoch": 0.44,
"learning_rate": 6.368571428571429e-06,
"loss": 0.2191,
"step": 1775
},
{
"epoch": 0.45,
"learning_rate": 6.297142857142857e-06,
"loss": 0.105,
"step": 1800
},
{
"epoch": 0.46,
"learning_rate": 6.225714285714286e-06,
"loss": 0.1324,
"step": 1825
},
{
"epoch": 0.46,
"learning_rate": 6.1542857142857145e-06,
"loss": 0.1817,
"step": 1850
},
{
"epoch": 0.47,
"learning_rate": 6.0828571428571435e-06,
"loss": 0.1742,
"step": 1875
},
{
"epoch": 0.47,
"learning_rate": 6.011428571428572e-06,
"loss": 0.1981,
"step": 1900
},
{
"epoch": 0.48,
"learning_rate": 5.94e-06,
"loss": 0.1463,
"step": 1925
},
{
"epoch": 0.49,
"learning_rate": 5.868571428571429e-06,
"loss": 0.149,
"step": 1950
},
{
"epoch": 0.49,
"learning_rate": 5.797142857142858e-06,
"loss": 0.1474,
"step": 1975
},
{
"epoch": 0.5,
"learning_rate": 5.725714285714287e-06,
"loss": 0.24,
"step": 2000
},
{
"epoch": 0.5,
"eval_loss": 0.23606961965560913,
"eval_runtime": 10036.0205,
"eval_samples_per_second": 0.306,
"eval_steps_per_second": 0.019,
"eval_wer": 11.241766018836083,
"step": 2000
},
{
"epoch": 0.51,
"learning_rate": 5.654285714285714e-06,
"loss": 0.1309,
"step": 2025
},
{
"epoch": 0.51,
"learning_rate": 5.582857142857143e-06,
"loss": 0.1073,
"step": 2050
},
{
"epoch": 0.52,
"learning_rate": 5.511428571428572e-06,
"loss": 0.195,
"step": 2075
},
{
"epoch": 0.53,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.1532,
"step": 2100
},
{
"epoch": 0.53,
"learning_rate": 5.368571428571429e-06,
"loss": 0.1457,
"step": 2125
},
{
"epoch": 0.54,
"learning_rate": 5.297142857142858e-06,
"loss": 0.1536,
"step": 2150
},
{
"epoch": 0.54,
"learning_rate": 5.225714285714286e-06,
"loss": 0.1708,
"step": 2175
},
{
"epoch": 0.55,
"learning_rate": 5.154285714285715e-06,
"loss": 0.1813,
"step": 2200
},
{
"epoch": 0.56,
"learning_rate": 5.082857142857144e-06,
"loss": 0.1274,
"step": 2225
},
{
"epoch": 0.56,
"learning_rate": 5.011428571428571e-06,
"loss": 0.1812,
"step": 2250
},
{
"epoch": 0.57,
"learning_rate": 4.94e-06,
"loss": 0.2039,
"step": 2275
},
{
"epoch": 0.57,
"learning_rate": 4.868571428571429e-06,
"loss": 0.122,
"step": 2300
},
{
"epoch": 0.58,
"learning_rate": 4.797142857142857e-06,
"loss": 0.1524,
"step": 2325
},
{
"epoch": 0.59,
"learning_rate": 4.725714285714286e-06,
"loss": 0.1521,
"step": 2350
},
{
"epoch": 0.59,
"learning_rate": 4.6542857142857145e-06,
"loss": 0.1639,
"step": 2375
},
{
"epoch": 0.6,
"learning_rate": 4.5828571428571435e-06,
"loss": 0.151,
"step": 2400
},
{
"epoch": 0.61,
"learning_rate": 4.511428571428572e-06,
"loss": 0.1837,
"step": 2425
},
{
"epoch": 0.61,
"learning_rate": 4.440000000000001e-06,
"loss": 0.1931,
"step": 2450
},
{
"epoch": 0.62,
"learning_rate": 4.368571428571429e-06,
"loss": 0.2184,
"step": 2475
},
{
"epoch": 0.62,
"learning_rate": 4.297142857142858e-06,
"loss": 0.2267,
"step": 2500
},
{
"epoch": 0.63,
"learning_rate": 4.225714285714286e-06,
"loss": 0.1389,
"step": 2525
},
{
"epoch": 0.64,
"learning_rate": 4.154285714285714e-06,
"loss": 0.1403,
"step": 2550
},
{
"epoch": 0.64,
"learning_rate": 4.082857142857143e-06,
"loss": 0.1452,
"step": 2575
},
{
"epoch": 0.65,
"learning_rate": 4.011428571428571e-06,
"loss": 0.2264,
"step": 2600
},
{
"epoch": 0.66,
"learning_rate": 3.94e-06,
"loss": 0.1998,
"step": 2625
},
{
"epoch": 0.66,
"learning_rate": 3.8685714285714286e-06,
"loss": 0.1258,
"step": 2650
},
{
"epoch": 0.67,
"learning_rate": 3.7971428571428576e-06,
"loss": 0.175,
"step": 2675
},
{
"epoch": 0.68,
"learning_rate": 3.7257142857142857e-06,
"loss": 0.1974,
"step": 2700
},
{
"epoch": 0.68,
"learning_rate": 3.6542857142857148e-06,
"loss": 0.1376,
"step": 2725
},
{
"epoch": 0.69,
"learning_rate": 3.582857142857143e-06,
"loss": 0.1473,
"step": 2750
},
{
"epoch": 0.69,
"learning_rate": 3.511428571428572e-06,
"loss": 0.1731,
"step": 2775
},
{
"epoch": 0.7,
"learning_rate": 3.44e-06,
"loss": 0.2041,
"step": 2800
},
{
"epoch": 0.71,
"learning_rate": 3.3685714285714287e-06,
"loss": 0.2068,
"step": 2825
},
{
"epoch": 0.71,
"learning_rate": 3.2971428571428577e-06,
"loss": 0.1555,
"step": 2850
},
{
"epoch": 0.72,
"learning_rate": 3.225714285714286e-06,
"loss": 0.1401,
"step": 2875
},
{
"epoch": 0.72,
"learning_rate": 3.154285714285715e-06,
"loss": 0.3019,
"step": 2900
},
{
"epoch": 0.73,
"learning_rate": 3.082857142857143e-06,
"loss": 0.1506,
"step": 2925
},
{
"epoch": 0.74,
"learning_rate": 3.0114285714285716e-06,
"loss": 0.2685,
"step": 2950
},
{
"epoch": 0.74,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.1849,
"step": 2975
},
{
"epoch": 0.75,
"learning_rate": 2.868571428571429e-06,
"loss": 0.1308,
"step": 3000
},
{
"epoch": 0.75,
"eval_loss": 0.23346394300460815,
"eval_runtime": 9895.3506,
"eval_samples_per_second": 0.311,
"eval_steps_per_second": 0.02,
"eval_wer": 10.750254699299285,
"step": 3000
},
{
"epoch": 0.76,
"learning_rate": 2.797142857142857e-06,
"loss": 0.1293,
"step": 3025
},
{
"epoch": 0.76,
"learning_rate": 2.725714285714286e-06,
"loss": 0.1276,
"step": 3050
},
{
"epoch": 0.77,
"learning_rate": 2.654285714285714e-06,
"loss": 0.2258,
"step": 3075
},
{
"epoch": 0.78,
"learning_rate": 2.582857142857143e-06,
"loss": 0.1976,
"step": 3100
},
{
"epoch": 0.78,
"learning_rate": 2.5114285714285718e-06,
"loss": 0.1396,
"step": 3125
},
{
"epoch": 0.79,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.1494,
"step": 3150
},
{
"epoch": 0.79,
"learning_rate": 2.3685714285714285e-06,
"loss": 0.156,
"step": 3175
},
{
"epoch": 0.8,
"learning_rate": 2.297142857142857e-06,
"loss": 0.2117,
"step": 3200
},
{
"epoch": 0.81,
"learning_rate": 2.2257142857142857e-06,
"loss": 0.1898,
"step": 3225
},
{
"epoch": 0.81,
"learning_rate": 2.1542857142857147e-06,
"loss": 0.1662,
"step": 3250
},
{
"epoch": 0.82,
"learning_rate": 2.0828571428571433e-06,
"loss": 0.1369,
"step": 3275
},
{
"epoch": 0.82,
"learning_rate": 2.0114285714285715e-06,
"loss": 0.2262,
"step": 3300
},
{
"epoch": 0.83,
"learning_rate": 1.94e-06,
"loss": 0.1123,
"step": 3325
},
{
"epoch": 0.84,
"learning_rate": 1.8685714285714289e-06,
"loss": 0.1412,
"step": 3350
},
{
"epoch": 0.84,
"learning_rate": 1.7971428571428572e-06,
"loss": 0.1642,
"step": 3375
},
{
"epoch": 0.85,
"learning_rate": 1.7257142857142858e-06,
"loss": 0.132,
"step": 3400
},
{
"epoch": 0.86,
"learning_rate": 1.6542857142857144e-06,
"loss": 0.1329,
"step": 3425
},
{
"epoch": 0.86,
"learning_rate": 1.582857142857143e-06,
"loss": 0.1607,
"step": 3450
},
{
"epoch": 0.87,
"learning_rate": 1.5114285714285714e-06,
"loss": 0.1693,
"step": 3475
},
{
"epoch": 0.88,
"learning_rate": 1.44e-06,
"loss": 0.1509,
"step": 3500
},
{
"epoch": 0.88,
"learning_rate": 1.3685714285714286e-06,
"loss": 0.1087,
"step": 3525
},
{
"epoch": 0.89,
"learning_rate": 1.3e-06,
"loss": 0.1381,
"step": 3550
},
{
"epoch": 0.89,
"learning_rate": 1.2285714285714286e-06,
"loss": 0.1128,
"step": 3575
},
{
"epoch": 0.9,
"learning_rate": 1.1571428571428572e-06,
"loss": 0.1646,
"step": 3600
},
{
"epoch": 0.91,
"learning_rate": 1.0857142857142858e-06,
"loss": 0.1514,
"step": 3625
},
{
"epoch": 0.91,
"learning_rate": 1.0142857142857144e-06,
"loss": 0.1552,
"step": 3650
},
{
"epoch": 0.92,
"learning_rate": 9.42857142857143e-07,
"loss": 0.1176,
"step": 3675
},
{
"epoch": 0.93,
"learning_rate": 8.714285714285716e-07,
"loss": 0.1626,
"step": 3700
},
{
"epoch": 0.93,
"learning_rate": 8.000000000000001e-07,
"loss": 0.1459,
"step": 3725
},
{
"epoch": 0.94,
"learning_rate": 7.285714285714287e-07,
"loss": 0.1592,
"step": 3750
},
{
"epoch": 0.94,
"learning_rate": 6.571428571428571e-07,
"loss": 0.1145,
"step": 3775
},
{
"epoch": 0.95,
"learning_rate": 5.857142857142857e-07,
"loss": 0.1451,
"step": 3800
},
{
"epoch": 0.96,
"learning_rate": 5.142857142857143e-07,
"loss": 0.1405,
"step": 3825
},
{
"epoch": 0.96,
"learning_rate": 4.4285714285714286e-07,
"loss": 0.1576,
"step": 3850
},
{
"epoch": 0.97,
"learning_rate": 3.7142857142857145e-07,
"loss": 0.1442,
"step": 3875
},
{
"epoch": 0.97,
"learning_rate": 3.0000000000000004e-07,
"loss": 0.1491,
"step": 3900
},
{
"epoch": 0.98,
"learning_rate": 2.285714285714286e-07,
"loss": 0.1323,
"step": 3925
},
{
"epoch": 0.99,
"learning_rate": 1.5714285714285717e-07,
"loss": 0.1043,
"step": 3950
},
{
"epoch": 0.99,
"learning_rate": 8.571428571428573e-08,
"loss": 0.165,
"step": 3975
},
{
"epoch": 1.0,
"learning_rate": 1.4285714285714288e-08,
"loss": 0.1072,
"step": 4000
},
{
"epoch": 1.0,
"eval_loss": 0.2031349390745163,
"eval_runtime": 9972.0158,
"eval_samples_per_second": 0.308,
"eval_steps_per_second": 0.019,
"eval_wer": 10.591602311347534,
"step": 4000
},
{
"epoch": 1.0,
"step": 4000,
"total_flos": 5.435317813248e+20,
"train_loss": 0.20192023056745528,
"train_runtime": 114914.5801,
"train_samples_per_second": 2.228,
"train_steps_per_second": 0.035
}
],
"max_steps": 4000,
"num_train_epochs": 9223372036854775807,
"total_flos": 5.435317813248e+20,
"trial_name": null,
"trial_params": null
}