whisper-small-it-3 / checkpoint-3000 /trainer_state.json
GIanlucaRub's picture
Copied from Google Drive
077030d
{
"best_metric": 22.199112590127566,
"best_model_checkpoint": "./whisper-small-it-3/checkpoint-2000",
"epoch": 2.8625954198473282,
"global_step": 3000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 4.2000000000000006e-07,
"loss": 3.6819,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 9.200000000000001e-07,
"loss": 2.9971,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 1.42e-06,
"loss": 2.1984,
"step": 75
},
{
"epoch": 0.1,
"learning_rate": 1.9200000000000003e-06,
"loss": 1.4257,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 2.42e-06,
"loss": 1.0526,
"step": 125
},
{
"epoch": 0.14,
"learning_rate": 2.92e-06,
"loss": 0.9106,
"step": 150
},
{
"epoch": 0.17,
"learning_rate": 3.4200000000000007e-06,
"loss": 0.8346,
"step": 175
},
{
"epoch": 0.19,
"learning_rate": 3.920000000000001e-06,
"loss": 0.7408,
"step": 200
},
{
"epoch": 0.21,
"learning_rate": 4.42e-06,
"loss": 0.6563,
"step": 225
},
{
"epoch": 0.24,
"learning_rate": 4.92e-06,
"loss": 0.6005,
"step": 250
},
{
"epoch": 0.26,
"learning_rate": 5.420000000000001e-06,
"loss": 0.473,
"step": 275
},
{
"epoch": 0.29,
"learning_rate": 5.92e-06,
"loss": 0.3355,
"step": 300
},
{
"epoch": 0.31,
"learning_rate": 6.42e-06,
"loss": 0.3183,
"step": 325
},
{
"epoch": 0.33,
"learning_rate": 6.92e-06,
"loss": 0.2788,
"step": 350
},
{
"epoch": 0.36,
"learning_rate": 7.420000000000001e-06,
"loss": 0.304,
"step": 375
},
{
"epoch": 0.38,
"learning_rate": 7.92e-06,
"loss": 0.278,
"step": 400
},
{
"epoch": 0.41,
"learning_rate": 8.42e-06,
"loss": 0.2932,
"step": 425
},
{
"epoch": 0.43,
"learning_rate": 8.920000000000001e-06,
"loss": 0.2813,
"step": 450
},
{
"epoch": 0.45,
"learning_rate": 9.42e-06,
"loss": 0.2675,
"step": 475
},
{
"epoch": 0.48,
"learning_rate": 9.920000000000002e-06,
"loss": 0.2845,
"step": 500
},
{
"epoch": 0.5,
"learning_rate": 9.940000000000001e-06,
"loss": 0.2827,
"step": 525
},
{
"epoch": 0.52,
"learning_rate": 9.86857142857143e-06,
"loss": 0.2772,
"step": 550
},
{
"epoch": 0.55,
"learning_rate": 9.797142857142858e-06,
"loss": 0.2928,
"step": 575
},
{
"epoch": 0.57,
"learning_rate": 9.725714285714287e-06,
"loss": 0.2784,
"step": 600
},
{
"epoch": 0.6,
"learning_rate": 9.654285714285716e-06,
"loss": 0.2913,
"step": 625
},
{
"epoch": 0.62,
"learning_rate": 9.582857142857143e-06,
"loss": 0.2732,
"step": 650
},
{
"epoch": 0.64,
"learning_rate": 9.511428571428572e-06,
"loss": 0.2557,
"step": 675
},
{
"epoch": 0.67,
"learning_rate": 9.440000000000001e-06,
"loss": 0.2736,
"step": 700
},
{
"epoch": 0.69,
"learning_rate": 9.368571428571428e-06,
"loss": 0.2707,
"step": 725
},
{
"epoch": 0.72,
"learning_rate": 9.297142857142857e-06,
"loss": 0.2478,
"step": 750
},
{
"epoch": 0.74,
"learning_rate": 9.225714285714286e-06,
"loss": 0.2828,
"step": 775
},
{
"epoch": 0.76,
"learning_rate": 9.154285714285715e-06,
"loss": 0.2504,
"step": 800
},
{
"epoch": 0.79,
"learning_rate": 9.082857142857143e-06,
"loss": 0.2538,
"step": 825
},
{
"epoch": 0.81,
"learning_rate": 9.011428571428572e-06,
"loss": 0.2703,
"step": 850
},
{
"epoch": 0.83,
"learning_rate": 8.94e-06,
"loss": 0.2658,
"step": 875
},
{
"epoch": 0.86,
"learning_rate": 8.86857142857143e-06,
"loss": 0.2688,
"step": 900
},
{
"epoch": 0.88,
"learning_rate": 8.797142857142857e-06,
"loss": 0.2834,
"step": 925
},
{
"epoch": 0.91,
"learning_rate": 8.725714285714286e-06,
"loss": 0.2305,
"step": 950
},
{
"epoch": 0.93,
"learning_rate": 8.654285714285715e-06,
"loss": 0.2449,
"step": 975
},
{
"epoch": 0.95,
"learning_rate": 8.582857142857144e-06,
"loss": 0.2545,
"step": 1000
},
{
"epoch": 0.95,
"eval_loss": 0.38721346855163574,
"eval_runtime": 628.7523,
"eval_samples_per_second": 2.386,
"eval_steps_per_second": 0.299,
"eval_wer": 24.889073765945646,
"step": 1000
},
{
"epoch": 0.98,
"learning_rate": 8.511428571428571e-06,
"loss": 0.2866,
"step": 1025
},
{
"epoch": 1.0,
"learning_rate": 8.44e-06,
"loss": 0.251,
"step": 1050
},
{
"epoch": 1.03,
"learning_rate": 8.36857142857143e-06,
"loss": 0.1246,
"step": 1075
},
{
"epoch": 1.05,
"learning_rate": 8.297142857142859e-06,
"loss": 0.1265,
"step": 1100
},
{
"epoch": 1.07,
"learning_rate": 8.225714285714288e-06,
"loss": 0.1327,
"step": 1125
},
{
"epoch": 1.1,
"learning_rate": 8.154285714285715e-06,
"loss": 0.1472,
"step": 1150
},
{
"epoch": 1.12,
"learning_rate": 8.082857142857144e-06,
"loss": 0.1328,
"step": 1175
},
{
"epoch": 1.15,
"learning_rate": 8.011428571428573e-06,
"loss": 0.126,
"step": 1200
},
{
"epoch": 1.17,
"learning_rate": 7.94e-06,
"loss": 0.1357,
"step": 1225
},
{
"epoch": 1.19,
"learning_rate": 7.86857142857143e-06,
"loss": 0.1226,
"step": 1250
},
{
"epoch": 1.22,
"learning_rate": 7.797142857142858e-06,
"loss": 0.1328,
"step": 1275
},
{
"epoch": 1.24,
"learning_rate": 7.725714285714286e-06,
"loss": 0.1353,
"step": 1300
},
{
"epoch": 1.26,
"learning_rate": 7.654285714285715e-06,
"loss": 0.1259,
"step": 1325
},
{
"epoch": 1.29,
"learning_rate": 7.5828571428571444e-06,
"loss": 0.1398,
"step": 1350
},
{
"epoch": 1.31,
"learning_rate": 7.511428571428572e-06,
"loss": 0.1253,
"step": 1375
},
{
"epoch": 1.34,
"learning_rate": 7.440000000000001e-06,
"loss": 0.1499,
"step": 1400
},
{
"epoch": 1.36,
"learning_rate": 7.36857142857143e-06,
"loss": 0.1342,
"step": 1425
},
{
"epoch": 1.38,
"learning_rate": 7.297142857142858e-06,
"loss": 0.1073,
"step": 1450
},
{
"epoch": 1.41,
"learning_rate": 7.225714285714286e-06,
"loss": 0.1412,
"step": 1475
},
{
"epoch": 1.43,
"learning_rate": 7.154285714285715e-06,
"loss": 0.1383,
"step": 1500
},
{
"epoch": 1.46,
"learning_rate": 7.082857142857143e-06,
"loss": 0.1272,
"step": 1525
},
{
"epoch": 1.48,
"learning_rate": 7.011428571428572e-06,
"loss": 0.1402,
"step": 1550
},
{
"epoch": 1.5,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.1302,
"step": 1575
},
{
"epoch": 1.53,
"learning_rate": 6.868571428571429e-06,
"loss": 0.1243,
"step": 1600
},
{
"epoch": 1.55,
"learning_rate": 6.797142857142858e-06,
"loss": 0.1345,
"step": 1625
},
{
"epoch": 1.57,
"learning_rate": 6.725714285714287e-06,
"loss": 0.1202,
"step": 1650
},
{
"epoch": 1.6,
"learning_rate": 6.654285714285716e-06,
"loss": 0.1355,
"step": 1675
},
{
"epoch": 1.62,
"learning_rate": 6.582857142857143e-06,
"loss": 0.1215,
"step": 1700
},
{
"epoch": 1.65,
"learning_rate": 6.511428571428572e-06,
"loss": 0.1279,
"step": 1725
},
{
"epoch": 1.67,
"learning_rate": 6.440000000000001e-06,
"loss": 0.1383,
"step": 1750
},
{
"epoch": 1.69,
"learning_rate": 6.368571428571429e-06,
"loss": 0.1261,
"step": 1775
},
{
"epoch": 1.72,
"learning_rate": 6.297142857142857e-06,
"loss": 0.1309,
"step": 1800
},
{
"epoch": 1.74,
"learning_rate": 6.225714285714286e-06,
"loss": 0.1265,
"step": 1825
},
{
"epoch": 1.77,
"learning_rate": 6.1542857142857145e-06,
"loss": 0.1067,
"step": 1850
},
{
"epoch": 1.79,
"learning_rate": 6.0828571428571435e-06,
"loss": 0.1268,
"step": 1875
},
{
"epoch": 1.81,
"learning_rate": 6.011428571428572e-06,
"loss": 0.1352,
"step": 1900
},
{
"epoch": 1.84,
"learning_rate": 5.94e-06,
"loss": 0.1261,
"step": 1925
},
{
"epoch": 1.86,
"learning_rate": 5.868571428571429e-06,
"loss": 0.1254,
"step": 1950
},
{
"epoch": 1.88,
"learning_rate": 5.797142857142858e-06,
"loss": 0.1263,
"step": 1975
},
{
"epoch": 1.91,
"learning_rate": 5.725714285714287e-06,
"loss": 0.129,
"step": 2000
},
{
"epoch": 1.91,
"eval_loss": 0.3681599199771881,
"eval_runtime": 629.0815,
"eval_samples_per_second": 2.384,
"eval_steps_per_second": 0.299,
"eval_wer": 22.199112590127566,
"step": 2000
},
{
"epoch": 1.93,
"learning_rate": 5.654285714285714e-06,
"loss": 0.1253,
"step": 2025
},
{
"epoch": 1.96,
"learning_rate": 5.582857142857143e-06,
"loss": 0.1169,
"step": 2050
},
{
"epoch": 1.98,
"learning_rate": 5.511428571428572e-06,
"loss": 0.1344,
"step": 2075
},
{
"epoch": 2.0,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.1149,
"step": 2100
},
{
"epoch": 2.03,
"learning_rate": 5.368571428571429e-06,
"loss": 0.0464,
"step": 2125
},
{
"epoch": 2.05,
"learning_rate": 5.297142857142858e-06,
"loss": 0.0519,
"step": 2150
},
{
"epoch": 2.08,
"learning_rate": 5.225714285714286e-06,
"loss": 0.0583,
"step": 2175
},
{
"epoch": 2.1,
"learning_rate": 5.154285714285715e-06,
"loss": 0.0562,
"step": 2200
},
{
"epoch": 2.12,
"learning_rate": 5.082857142857144e-06,
"loss": 0.0574,
"step": 2225
},
{
"epoch": 2.15,
"learning_rate": 5.011428571428571e-06,
"loss": 0.0523,
"step": 2250
},
{
"epoch": 2.17,
"learning_rate": 4.94e-06,
"loss": 0.0585,
"step": 2275
},
{
"epoch": 2.19,
"learning_rate": 4.868571428571429e-06,
"loss": 0.0585,
"step": 2300
},
{
"epoch": 2.22,
"learning_rate": 4.797142857142857e-06,
"loss": 0.0598,
"step": 2325
},
{
"epoch": 2.24,
"learning_rate": 4.725714285714286e-06,
"loss": 0.063,
"step": 2350
},
{
"epoch": 2.27,
"learning_rate": 4.6542857142857145e-06,
"loss": 0.0516,
"step": 2375
},
{
"epoch": 2.29,
"learning_rate": 4.5828571428571435e-06,
"loss": 0.0514,
"step": 2400
},
{
"epoch": 2.31,
"learning_rate": 4.511428571428572e-06,
"loss": 0.0495,
"step": 2425
},
{
"epoch": 2.34,
"learning_rate": 4.440000000000001e-06,
"loss": 0.0486,
"step": 2450
},
{
"epoch": 2.36,
"learning_rate": 4.368571428571429e-06,
"loss": 0.0496,
"step": 2475
},
{
"epoch": 2.39,
"learning_rate": 4.297142857142858e-06,
"loss": 0.0562,
"step": 2500
},
{
"epoch": 2.41,
"learning_rate": 4.225714285714286e-06,
"loss": 0.0559,
"step": 2525
},
{
"epoch": 2.43,
"learning_rate": 4.154285714285714e-06,
"loss": 0.0536,
"step": 2550
},
{
"epoch": 2.46,
"learning_rate": 4.082857142857143e-06,
"loss": 0.0549,
"step": 2575
},
{
"epoch": 2.48,
"learning_rate": 4.011428571428571e-06,
"loss": 0.0616,
"step": 2600
},
{
"epoch": 2.5,
"learning_rate": 3.94e-06,
"loss": 0.0551,
"step": 2625
},
{
"epoch": 2.53,
"learning_rate": 3.8685714285714286e-06,
"loss": 0.0506,
"step": 2650
},
{
"epoch": 2.55,
"learning_rate": 3.7971428571428576e-06,
"loss": 0.0584,
"step": 2675
},
{
"epoch": 2.58,
"learning_rate": 3.7257142857142857e-06,
"loss": 0.0589,
"step": 2700
},
{
"epoch": 2.6,
"learning_rate": 3.6542857142857148e-06,
"loss": 0.047,
"step": 2725
},
{
"epoch": 2.62,
"learning_rate": 3.582857142857143e-06,
"loss": 0.0517,
"step": 2750
},
{
"epoch": 2.65,
"learning_rate": 3.511428571428572e-06,
"loss": 0.0543,
"step": 2775
},
{
"epoch": 2.67,
"learning_rate": 3.44e-06,
"loss": 0.0466,
"step": 2800
},
{
"epoch": 2.7,
"learning_rate": 3.3685714285714287e-06,
"loss": 0.0492,
"step": 2825
},
{
"epoch": 2.72,
"learning_rate": 3.2971428571428577e-06,
"loss": 0.0589,
"step": 2850
},
{
"epoch": 2.74,
"learning_rate": 3.225714285714286e-06,
"loss": 0.0531,
"step": 2875
},
{
"epoch": 2.77,
"learning_rate": 3.154285714285715e-06,
"loss": 0.0539,
"step": 2900
},
{
"epoch": 2.79,
"learning_rate": 3.082857142857143e-06,
"loss": 0.0614,
"step": 2925
},
{
"epoch": 2.81,
"learning_rate": 3.0114285714285716e-06,
"loss": 0.056,
"step": 2950
},
{
"epoch": 2.84,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.0601,
"step": 2975
},
{
"epoch": 2.86,
"learning_rate": 2.868571428571429e-06,
"loss": 0.0534,
"step": 3000
},
{
"epoch": 2.86,
"eval_loss": 0.3771337866783142,
"eval_runtime": 606.2144,
"eval_samples_per_second": 2.474,
"eval_steps_per_second": 0.31,
"eval_wer": 22.469495285635052,
"step": 3000
}
],
"max_steps": 4000,
"num_train_epochs": 4,
"total_flos": 1.384805901975552e+19,
"trial_name": null,
"trial_params": null
}