whisper-tiny-it-7 / checkpoint-2000 /trainer_state.json
GIanlucaRub's picture
Copied from Google Drive
8f173c0
raw
history blame
10.2 kB
{
"best_metric": 160.23294509151413,
"best_model_checkpoint": "./whisper-tiny-it-7/checkpoint-2000",
"epoch": 1.9083969465648853,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 4e-08,
"loss": 8.8525,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 9e-08,
"loss": 8.857,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 1.4e-07,
"loss": 8.6107,
"step": 75
},
{
"epoch": 0.1,
"learning_rate": 1.8999999999999998e-07,
"loss": 8.1659,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 2.4e-07,
"loss": 7.7359,
"step": 125
},
{
"epoch": 0.14,
"learning_rate": 2.9e-07,
"loss": 7.2816,
"step": 150
},
{
"epoch": 0.17,
"learning_rate": 3.4000000000000003e-07,
"loss": 6.9829,
"step": 175
},
{
"epoch": 0.19,
"learning_rate": 3.8999999999999997e-07,
"loss": 6.5257,
"step": 200
},
{
"epoch": 0.21,
"learning_rate": 4.3999999999999997e-07,
"loss": 6.1594,
"step": 225
},
{
"epoch": 0.24,
"learning_rate": 4.9e-07,
"loss": 5.9187,
"step": 250
},
{
"epoch": 0.26,
"learning_rate": 5.4e-07,
"loss": 5.7123,
"step": 275
},
{
"epoch": 0.29,
"learning_rate": 5.9e-07,
"loss": 5.4899,
"step": 300
},
{
"epoch": 0.31,
"learning_rate": 6.4e-07,
"loss": 5.1635,
"step": 325
},
{
"epoch": 0.33,
"learning_rate": 6.9e-07,
"loss": 4.9477,
"step": 350
},
{
"epoch": 0.36,
"learning_rate": 7.4e-07,
"loss": 4.8882,
"step": 375
},
{
"epoch": 0.38,
"learning_rate": 7.9e-07,
"loss": 4.7571,
"step": 400
},
{
"epoch": 0.41,
"learning_rate": 8.399999999999999e-07,
"loss": 4.7402,
"step": 425
},
{
"epoch": 0.43,
"learning_rate": 8.9e-07,
"loss": 4.6719,
"step": 450
},
{
"epoch": 0.45,
"learning_rate": 9.399999999999999e-07,
"loss": 4.5627,
"step": 475
},
{
"epoch": 0.48,
"learning_rate": 9.9e-07,
"loss": 4.6078,
"step": 500
},
{
"epoch": 0.5,
"learning_rate": 9.942857142857144e-07,
"loss": 4.5315,
"step": 525
},
{
"epoch": 0.52,
"learning_rate": 9.871428571428572e-07,
"loss": 4.4441,
"step": 550
},
{
"epoch": 0.55,
"learning_rate": 9.8e-07,
"loss": 4.4877,
"step": 575
},
{
"epoch": 0.57,
"learning_rate": 9.72857142857143e-07,
"loss": 4.4972,
"step": 600
},
{
"epoch": 0.6,
"learning_rate": 9.657142857142857e-07,
"loss": 4.4206,
"step": 625
},
{
"epoch": 0.62,
"learning_rate": 9.585714285714285e-07,
"loss": 4.4121,
"step": 650
},
{
"epoch": 0.64,
"learning_rate": 9.514285714285713e-07,
"loss": 4.3059,
"step": 675
},
{
"epoch": 0.67,
"learning_rate": 9.442857142857142e-07,
"loss": 4.3611,
"step": 700
},
{
"epoch": 0.69,
"learning_rate": 9.371428571428571e-07,
"loss": 4.343,
"step": 725
},
{
"epoch": 0.72,
"learning_rate": 9.3e-07,
"loss": 4.3366,
"step": 750
},
{
"epoch": 0.74,
"learning_rate": 9.228571428571428e-07,
"loss": 4.3313,
"step": 775
},
{
"epoch": 0.76,
"learning_rate": 9.157142857142856e-07,
"loss": 4.317,
"step": 800
},
{
"epoch": 0.79,
"learning_rate": 9.085714285714285e-07,
"loss": 4.2746,
"step": 825
},
{
"epoch": 0.81,
"learning_rate": 9.014285714285714e-07,
"loss": 4.2817,
"step": 850
},
{
"epoch": 0.83,
"learning_rate": 8.942857142857142e-07,
"loss": 4.2529,
"step": 875
},
{
"epoch": 0.86,
"learning_rate": 8.871428571428571e-07,
"loss": 4.248,
"step": 900
},
{
"epoch": 0.88,
"learning_rate": 8.799999999999999e-07,
"loss": 4.2731,
"step": 925
},
{
"epoch": 0.91,
"learning_rate": 8.728571428571428e-07,
"loss": 4.2709,
"step": 950
},
{
"epoch": 0.93,
"learning_rate": 8.657142857142857e-07,
"loss": 4.2254,
"step": 975
},
{
"epoch": 0.95,
"learning_rate": 8.585714285714285e-07,
"loss": 4.204,
"step": 1000
},
{
"epoch": 0.95,
"eval_loss": 4.17011833190918,
"eval_runtime": 1604.0964,
"eval_samples_per_second": 0.935,
"eval_steps_per_second": 0.117,
"eval_wer": 644.20410427066,
"step": 1000
},
{
"epoch": 0.98,
"learning_rate": 8.514285714285714e-07,
"loss": 4.2357,
"step": 1025
},
{
"epoch": 1.0,
"learning_rate": 8.442857142857142e-07,
"loss": 4.1855,
"step": 1050
},
{
"epoch": 1.03,
"learning_rate": 8.371428571428571e-07,
"loss": 4.1227,
"step": 1075
},
{
"epoch": 1.05,
"learning_rate": 8.299999999999999e-07,
"loss": 4.1665,
"step": 1100
},
{
"epoch": 1.07,
"learning_rate": 8.228571428571428e-07,
"loss": 4.1772,
"step": 1125
},
{
"epoch": 1.1,
"learning_rate": 8.157142857142856e-07,
"loss": 4.1451,
"step": 1150
},
{
"epoch": 1.12,
"learning_rate": 8.085714285714285e-07,
"loss": 4.0833,
"step": 1175
},
{
"epoch": 1.15,
"learning_rate": 8.014285714285713e-07,
"loss": 4.1437,
"step": 1200
},
{
"epoch": 1.17,
"learning_rate": 7.942857142857142e-07,
"loss": 4.1286,
"step": 1225
},
{
"epoch": 1.19,
"learning_rate": 7.871428571428571e-07,
"loss": 4.1125,
"step": 1250
},
{
"epoch": 1.22,
"learning_rate": 7.799999999999999e-07,
"loss": 4.0927,
"step": 1275
},
{
"epoch": 1.24,
"learning_rate": 7.728571428571428e-07,
"loss": 4.1013,
"step": 1300
},
{
"epoch": 1.26,
"learning_rate": 7.657142857142856e-07,
"loss": 4.0453,
"step": 1325
},
{
"epoch": 1.29,
"learning_rate": 7.585714285714285e-07,
"loss": 4.1317,
"step": 1350
},
{
"epoch": 1.31,
"learning_rate": 7.514285714285715e-07,
"loss": 4.0638,
"step": 1375
},
{
"epoch": 1.34,
"learning_rate": 7.442857142857143e-07,
"loss": 4.1043,
"step": 1400
},
{
"epoch": 1.36,
"learning_rate": 7.371428571428571e-07,
"loss": 4.0761,
"step": 1425
},
{
"epoch": 1.38,
"learning_rate": 7.3e-07,
"loss": 3.9902,
"step": 1450
},
{
"epoch": 1.41,
"learning_rate": 7.228571428571429e-07,
"loss": 4.0019,
"step": 1475
},
{
"epoch": 1.43,
"learning_rate": 7.157142857142858e-07,
"loss": 4.0683,
"step": 1500
},
{
"epoch": 1.46,
"learning_rate": 7.085714285714285e-07,
"loss": 3.9523,
"step": 1525
},
{
"epoch": 1.48,
"learning_rate": 7.014285714285714e-07,
"loss": 3.9421,
"step": 1550
},
{
"epoch": 1.5,
"learning_rate": 6.942857142857143e-07,
"loss": 3.7698,
"step": 1575
},
{
"epoch": 1.53,
"learning_rate": 6.871428571428572e-07,
"loss": 3.7565,
"step": 1600
},
{
"epoch": 1.55,
"learning_rate": 6.800000000000001e-07,
"loss": 3.6593,
"step": 1625
},
{
"epoch": 1.57,
"learning_rate": 6.728571428571428e-07,
"loss": 3.4785,
"step": 1650
},
{
"epoch": 1.6,
"learning_rate": 6.657142857142857e-07,
"loss": 3.4676,
"step": 1675
},
{
"epoch": 1.62,
"learning_rate": 6.585714285714286e-07,
"loss": 3.3216,
"step": 1700
},
{
"epoch": 1.65,
"learning_rate": 6.514285714285715e-07,
"loss": 3.2336,
"step": 1725
},
{
"epoch": 1.67,
"learning_rate": 6.442857142857142e-07,
"loss": 3.1206,
"step": 1750
},
{
"epoch": 1.69,
"learning_rate": 6.371428571428571e-07,
"loss": 3.0662,
"step": 1775
},
{
"epoch": 1.72,
"learning_rate": 6.3e-07,
"loss": 2.972,
"step": 1800
},
{
"epoch": 1.74,
"learning_rate": 6.228571428571429e-07,
"loss": 2.8847,
"step": 1825
},
{
"epoch": 1.77,
"learning_rate": 6.157142857142858e-07,
"loss": 2.8173,
"step": 1850
},
{
"epoch": 1.79,
"learning_rate": 6.085714285714285e-07,
"loss": 2.7043,
"step": 1875
},
{
"epoch": 1.81,
"learning_rate": 6.014285714285714e-07,
"loss": 2.7055,
"step": 1900
},
{
"epoch": 1.84,
"learning_rate": 5.942857142857143e-07,
"loss": 2.7277,
"step": 1925
},
{
"epoch": 1.86,
"learning_rate": 5.871428571428572e-07,
"loss": 2.6155,
"step": 1950
},
{
"epoch": 1.88,
"learning_rate": 5.8e-07,
"loss": 2.533,
"step": 1975
},
{
"epoch": 1.91,
"learning_rate": 5.728571428571428e-07,
"loss": 2.4631,
"step": 2000
},
{
"epoch": 1.91,
"eval_loss": 2.7998619079589844,
"eval_runtime": 865.3409,
"eval_samples_per_second": 1.733,
"eval_steps_per_second": 0.217,
"eval_wer": 160.23294509151413,
"step": 2000
}
],
"max_steps": 4000,
"num_train_epochs": 4,
"total_flos": 7.8763199164416e+17,
"trial_name": null,
"trial_params": null
}