whisper-tiny-it-8 / checkpoint-4000 /trainer_state.json
FCameCode's picture
Upload 48 files
c2ca7f9
{
"best_metric": 56.905158069883534,
"best_model_checkpoint": "./whisper-tiny-it-8/checkpoint-3000",
"epoch": 3.816793893129771,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 4.0000000000000003e-07,
"loss": 8.8054,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 9.000000000000001e-07,
"loss": 7.9613,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 1.4000000000000001e-06,
"loss": 6.7905,
"step": 75
},
{
"epoch": 0.1,
"learning_rate": 1.9000000000000002e-06,
"loss": 5.8035,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 2.4000000000000003e-06,
"loss": 5.1,
"step": 125
},
{
"epoch": 0.14,
"learning_rate": 2.9e-06,
"loss": 4.7572,
"step": 150
},
{
"epoch": 0.17,
"learning_rate": 3.4000000000000005e-06,
"loss": 4.6419,
"step": 175
},
{
"epoch": 0.19,
"learning_rate": 3.900000000000001e-06,
"loss": 4.4402,
"step": 200
},
{
"epoch": 0.21,
"learning_rate": 4.4e-06,
"loss": 4.3455,
"step": 225
},
{
"epoch": 0.24,
"learning_rate": 4.9000000000000005e-06,
"loss": 4.3137,
"step": 250
},
{
"epoch": 0.26,
"learning_rate": 5.400000000000001e-06,
"loss": 4.2972,
"step": 275
},
{
"epoch": 0.29,
"learning_rate": 5.9e-06,
"loss": 4.2256,
"step": 300
},
{
"epoch": 0.31,
"learning_rate": 6.4000000000000006e-06,
"loss": 4.1941,
"step": 325
},
{
"epoch": 0.33,
"learning_rate": 6.9e-06,
"loss": 4.0644,
"step": 350
},
{
"epoch": 0.36,
"learning_rate": 7.4e-06,
"loss": 3.8409,
"step": 375
},
{
"epoch": 0.38,
"learning_rate": 7.9e-06,
"loss": 2.9889,
"step": 400
},
{
"epoch": 0.41,
"learning_rate": 8.400000000000001e-06,
"loss": 2.3795,
"step": 425
},
{
"epoch": 0.43,
"learning_rate": 8.900000000000001e-06,
"loss": 1.9844,
"step": 450
},
{
"epoch": 0.45,
"learning_rate": 9.4e-06,
"loss": 1.6867,
"step": 475
},
{
"epoch": 0.48,
"learning_rate": 9.9e-06,
"loss": 1.6208,
"step": 500
},
{
"epoch": 0.5,
"learning_rate": 9.942857142857145e-06,
"loss": 1.4984,
"step": 525
},
{
"epoch": 0.52,
"learning_rate": 9.871428571428572e-06,
"loss": 1.3849,
"step": 550
},
{
"epoch": 0.55,
"learning_rate": 9.800000000000001e-06,
"loss": 1.4037,
"step": 575
},
{
"epoch": 0.57,
"learning_rate": 9.72857142857143e-06,
"loss": 1.3242,
"step": 600
},
{
"epoch": 0.6,
"learning_rate": 9.657142857142859e-06,
"loss": 1.2818,
"step": 625
},
{
"epoch": 0.62,
"learning_rate": 9.585714285714286e-06,
"loss": 1.2702,
"step": 650
},
{
"epoch": 0.64,
"learning_rate": 9.514285714285715e-06,
"loss": 1.1514,
"step": 675
},
{
"epoch": 0.67,
"learning_rate": 9.442857142857144e-06,
"loss": 1.17,
"step": 700
},
{
"epoch": 0.69,
"learning_rate": 9.371428571428572e-06,
"loss": 1.1697,
"step": 725
},
{
"epoch": 0.72,
"learning_rate": 9.3e-06,
"loss": 1.1135,
"step": 750
},
{
"epoch": 0.74,
"learning_rate": 9.22857142857143e-06,
"loss": 1.149,
"step": 775
},
{
"epoch": 0.76,
"learning_rate": 9.157142857142857e-06,
"loss": 1.0708,
"step": 800
},
{
"epoch": 0.79,
"learning_rate": 9.085714285714286e-06,
"loss": 1.0837,
"step": 825
},
{
"epoch": 0.81,
"learning_rate": 9.014285714285715e-06,
"loss": 1.0944,
"step": 850
},
{
"epoch": 0.83,
"learning_rate": 8.942857142857142e-06,
"loss": 1.0532,
"step": 875
},
{
"epoch": 0.86,
"learning_rate": 8.871428571428571e-06,
"loss": 1.0309,
"step": 900
},
{
"epoch": 0.88,
"learning_rate": 8.8e-06,
"loss": 1.0917,
"step": 925
},
{
"epoch": 0.91,
"learning_rate": 8.72857142857143e-06,
"loss": 1.0036,
"step": 950
},
{
"epoch": 0.93,
"learning_rate": 8.657142857142858e-06,
"loss": 1.0038,
"step": 975
},
{
"epoch": 0.95,
"learning_rate": 8.585714285714286e-06,
"loss": 1.0065,
"step": 1000
},
{
"epoch": 0.95,
"eval_loss": 1.2800816297531128,
"eval_runtime": 266.6418,
"eval_samples_per_second": 5.626,
"eval_steps_per_second": 0.705,
"eval_wer": 69.238768718802,
"step": 1000
},
{
"epoch": 0.98,
"learning_rate": 8.514285714285715e-06,
"loss": 1.0273,
"step": 1025
},
{
"epoch": 1.0,
"learning_rate": 8.442857142857144e-06,
"loss": 0.9565,
"step": 1050
},
{
"epoch": 1.03,
"learning_rate": 8.371428571428573e-06,
"loss": 0.8123,
"step": 1075
},
{
"epoch": 1.05,
"learning_rate": 8.3e-06,
"loss": 0.815,
"step": 1100
},
{
"epoch": 1.07,
"learning_rate": 8.22857142857143e-06,
"loss": 0.8377,
"step": 1125
},
{
"epoch": 1.1,
"learning_rate": 8.157142857142858e-06,
"loss": 0.8618,
"step": 1150
},
{
"epoch": 1.12,
"learning_rate": 8.085714285714287e-06,
"loss": 0.8205,
"step": 1175
},
{
"epoch": 1.15,
"learning_rate": 8.014285714285715e-06,
"loss": 0.7786,
"step": 1200
},
{
"epoch": 1.17,
"learning_rate": 7.942857142857144e-06,
"loss": 0.8262,
"step": 1225
},
{
"epoch": 1.19,
"learning_rate": 7.871428571428573e-06,
"loss": 0.7959,
"step": 1250
},
{
"epoch": 1.22,
"learning_rate": 7.800000000000002e-06,
"loss": 0.8014,
"step": 1275
},
{
"epoch": 1.24,
"learning_rate": 7.72857142857143e-06,
"loss": 0.7926,
"step": 1300
},
{
"epoch": 1.26,
"learning_rate": 7.657142857142858e-06,
"loss": 0.7425,
"step": 1325
},
{
"epoch": 1.29,
"learning_rate": 7.585714285714286e-06,
"loss": 0.791,
"step": 1350
},
{
"epoch": 1.31,
"learning_rate": 7.514285714285715e-06,
"loss": 0.7548,
"step": 1375
},
{
"epoch": 1.34,
"learning_rate": 7.442857142857144e-06,
"loss": 0.811,
"step": 1400
},
{
"epoch": 1.36,
"learning_rate": 7.371428571428571e-06,
"loss": 0.7539,
"step": 1425
},
{
"epoch": 1.38,
"learning_rate": 7.3e-06,
"loss": 0.7304,
"step": 1450
},
{
"epoch": 1.41,
"learning_rate": 7.2285714285714294e-06,
"loss": 0.7771,
"step": 1475
},
{
"epoch": 1.43,
"learning_rate": 7.1571428571428584e-06,
"loss": 0.784,
"step": 1500
},
{
"epoch": 1.46,
"learning_rate": 7.085714285714286e-06,
"loss": 0.7612,
"step": 1525
},
{
"epoch": 1.48,
"learning_rate": 7.014285714285715e-06,
"loss": 0.7805,
"step": 1550
},
{
"epoch": 1.5,
"learning_rate": 6.942857142857144e-06,
"loss": 0.7835,
"step": 1575
},
{
"epoch": 1.53,
"learning_rate": 6.871428571428572e-06,
"loss": 0.7226,
"step": 1600
},
{
"epoch": 1.55,
"learning_rate": 6.800000000000001e-06,
"loss": 0.7638,
"step": 1625
},
{
"epoch": 1.57,
"learning_rate": 6.728571428571429e-06,
"loss": 0.7078,
"step": 1650
},
{
"epoch": 1.6,
"learning_rate": 6.657142857142857e-06,
"loss": 0.7265,
"step": 1675
},
{
"epoch": 1.62,
"learning_rate": 6.585714285714286e-06,
"loss": 0.7209,
"step": 1700
},
{
"epoch": 1.65,
"learning_rate": 6.514285714285715e-06,
"loss": 0.7052,
"step": 1725
},
{
"epoch": 1.67,
"learning_rate": 6.442857142857143e-06,
"loss": 0.7082,
"step": 1750
},
{
"epoch": 1.69,
"learning_rate": 6.371428571428572e-06,
"loss": 0.7255,
"step": 1775
},
{
"epoch": 1.72,
"learning_rate": 6.300000000000001e-06,
"loss": 0.7278,
"step": 1800
},
{
"epoch": 1.74,
"learning_rate": 6.22857142857143e-06,
"loss": 0.7112,
"step": 1825
},
{
"epoch": 1.77,
"learning_rate": 6.157142857142858e-06,
"loss": 0.6747,
"step": 1850
},
{
"epoch": 1.79,
"learning_rate": 6.085714285714286e-06,
"loss": 0.6794,
"step": 1875
},
{
"epoch": 1.81,
"learning_rate": 6.014285714285715e-06,
"loss": 0.7097,
"step": 1900
},
{
"epoch": 1.84,
"learning_rate": 5.942857142857143e-06,
"loss": 0.6871,
"step": 1925
},
{
"epoch": 1.86,
"learning_rate": 5.871428571428572e-06,
"loss": 0.6971,
"step": 1950
},
{
"epoch": 1.88,
"learning_rate": 5.8e-06,
"loss": 0.6698,
"step": 1975
},
{
"epoch": 1.91,
"learning_rate": 5.7285714285714285e-06,
"loss": 0.6735,
"step": 2000
},
{
"epoch": 1.91,
"eval_loss": 1.0736733675003052,
"eval_runtime": 263.868,
"eval_samples_per_second": 5.685,
"eval_steps_per_second": 0.712,
"eval_wer": 59.1444814198558,
"step": 2000
},
{
"epoch": 1.93,
"learning_rate": 5.6571428571428576e-06,
"loss": 0.6878,
"step": 2025
},
{
"epoch": 1.96,
"learning_rate": 5.5857142857142866e-06,
"loss": 0.657,
"step": 2050
},
{
"epoch": 1.98,
"learning_rate": 5.514285714285714e-06,
"loss": 0.6893,
"step": 2075
},
{
"epoch": 2.0,
"learning_rate": 5.442857142857143e-06,
"loss": 0.6431,
"step": 2100
},
{
"epoch": 2.03,
"learning_rate": 5.371428571428572e-06,
"loss": 0.5387,
"step": 2125
},
{
"epoch": 2.05,
"learning_rate": 5.300000000000001e-06,
"loss": 0.5448,
"step": 2150
},
{
"epoch": 2.08,
"learning_rate": 5.22857142857143e-06,
"loss": 0.5627,
"step": 2175
},
{
"epoch": 2.1,
"learning_rate": 5.157142857142857e-06,
"loss": 0.5601,
"step": 2200
},
{
"epoch": 2.12,
"learning_rate": 5.085714285714286e-06,
"loss": 0.5506,
"step": 2225
},
{
"epoch": 2.15,
"learning_rate": 5.014285714285715e-06,
"loss": 0.5368,
"step": 2250
},
{
"epoch": 2.17,
"learning_rate": 4.9428571428571435e-06,
"loss": 0.5878,
"step": 2275
},
{
"epoch": 2.19,
"learning_rate": 4.871428571428572e-06,
"loss": 0.5591,
"step": 2300
},
{
"epoch": 2.22,
"learning_rate": 4.800000000000001e-06,
"loss": 0.5557,
"step": 2325
},
{
"epoch": 2.24,
"learning_rate": 4.728571428571429e-06,
"loss": 0.5722,
"step": 2350
},
{
"epoch": 2.27,
"learning_rate": 4.657142857142857e-06,
"loss": 0.5888,
"step": 2375
},
{
"epoch": 2.29,
"learning_rate": 4.585714285714286e-06,
"loss": 0.5553,
"step": 2400
},
{
"epoch": 2.31,
"learning_rate": 4.514285714285714e-06,
"loss": 0.5231,
"step": 2425
},
{
"epoch": 2.34,
"learning_rate": 4.442857142857143e-06,
"loss": 0.5489,
"step": 2450
},
{
"epoch": 2.36,
"learning_rate": 4.371428571428572e-06,
"loss": 0.539,
"step": 2475
},
{
"epoch": 2.39,
"learning_rate": 4.3e-06,
"loss": 0.5743,
"step": 2500
},
{
"epoch": 2.41,
"learning_rate": 4.228571428571429e-06,
"loss": 0.5524,
"step": 2525
},
{
"epoch": 2.43,
"learning_rate": 4.1571428571428575e-06,
"loss": 0.5245,
"step": 2550
},
{
"epoch": 2.46,
"learning_rate": 4.0857142857142865e-06,
"loss": 0.548,
"step": 2575
},
{
"epoch": 2.48,
"learning_rate": 4.014285714285715e-06,
"loss": 0.5343,
"step": 2600
},
{
"epoch": 2.5,
"learning_rate": 3.942857142857143e-06,
"loss": 0.5425,
"step": 2625
},
{
"epoch": 2.53,
"learning_rate": 3.871428571428572e-06,
"loss": 0.5518,
"step": 2650
},
{
"epoch": 2.55,
"learning_rate": 3.8000000000000005e-06,
"loss": 0.5514,
"step": 2675
},
{
"epoch": 2.58,
"learning_rate": 3.7285714285714286e-06,
"loss": 0.5478,
"step": 2700
},
{
"epoch": 2.6,
"learning_rate": 3.6571428571428576e-06,
"loss": 0.5416,
"step": 2725
},
{
"epoch": 2.62,
"learning_rate": 3.5857142857142862e-06,
"loss": 0.5508,
"step": 2750
},
{
"epoch": 2.65,
"learning_rate": 3.5142857142857144e-06,
"loss": 0.5565,
"step": 2775
},
{
"epoch": 2.67,
"learning_rate": 3.4428571428571434e-06,
"loss": 0.5403,
"step": 2800
},
{
"epoch": 2.7,
"learning_rate": 3.3714285714285716e-06,
"loss": 0.5276,
"step": 2825
},
{
"epoch": 2.72,
"learning_rate": 3.3000000000000006e-06,
"loss": 0.566,
"step": 2850
},
{
"epoch": 2.74,
"learning_rate": 3.2285714285714288e-06,
"loss": 0.5128,
"step": 2875
},
{
"epoch": 2.77,
"learning_rate": 3.1571428571428573e-06,
"loss": 0.5455,
"step": 2900
},
{
"epoch": 2.79,
"learning_rate": 3.085714285714286e-06,
"loss": 0.5543,
"step": 2925
},
{
"epoch": 2.81,
"learning_rate": 3.0142857142857145e-06,
"loss": 0.5532,
"step": 2950
},
{
"epoch": 2.84,
"learning_rate": 2.9428571428571427e-06,
"loss": 0.5395,
"step": 2975
},
{
"epoch": 2.86,
"learning_rate": 2.8714285714285717e-06,
"loss": 0.5258,
"step": 3000
},
{
"epoch": 2.86,
"eval_loss": 1.0115019083023071,
"eval_runtime": 269.7805,
"eval_samples_per_second": 5.56,
"eval_steps_per_second": 0.697,
"eval_wer": 56.905158069883534,
"step": 3000
},
{
"epoch": 2.89,
"learning_rate": 2.8000000000000003e-06,
"loss": 0.5565,
"step": 3025
},
{
"epoch": 2.91,
"learning_rate": 2.728571428571429e-06,
"loss": 0.5461,
"step": 3050
},
{
"epoch": 2.93,
"learning_rate": 2.6571428571428575e-06,
"loss": 0.5455,
"step": 3075
},
{
"epoch": 2.96,
"learning_rate": 2.5857142857142856e-06,
"loss": 0.5553,
"step": 3100
},
{
"epoch": 2.98,
"learning_rate": 2.5142857142857147e-06,
"loss": 0.5281,
"step": 3125
},
{
"epoch": 3.01,
"learning_rate": 2.442857142857143e-06,
"loss": 0.516,
"step": 3150
},
{
"epoch": 3.03,
"learning_rate": 2.371428571428572e-06,
"loss": 0.4755,
"step": 3175
},
{
"epoch": 3.05,
"learning_rate": 2.3000000000000004e-06,
"loss": 0.4388,
"step": 3200
},
{
"epoch": 3.08,
"learning_rate": 2.228571428571429e-06,
"loss": 0.444,
"step": 3225
},
{
"epoch": 3.1,
"learning_rate": 2.157142857142857e-06,
"loss": 0.4704,
"step": 3250
},
{
"epoch": 3.12,
"learning_rate": 2.0857142857142858e-06,
"loss": 0.4611,
"step": 3275
},
{
"epoch": 3.15,
"learning_rate": 2.0142857142857144e-06,
"loss": 0.4756,
"step": 3300
},
{
"epoch": 3.17,
"learning_rate": 1.942857142857143e-06,
"loss": 0.4623,
"step": 3325
},
{
"epoch": 3.2,
"learning_rate": 1.8714285714285715e-06,
"loss": 0.4465,
"step": 3350
},
{
"epoch": 3.22,
"learning_rate": 1.8000000000000001e-06,
"loss": 0.4818,
"step": 3375
},
{
"epoch": 3.24,
"learning_rate": 1.7285714285714287e-06,
"loss": 0.4584,
"step": 3400
},
{
"epoch": 3.27,
"learning_rate": 1.657142857142857e-06,
"loss": 0.4421,
"step": 3425
},
{
"epoch": 3.29,
"learning_rate": 1.5857142857142857e-06,
"loss": 0.4468,
"step": 3450
},
{
"epoch": 3.32,
"learning_rate": 1.5142857142857145e-06,
"loss": 0.4412,
"step": 3475
},
{
"epoch": 3.34,
"learning_rate": 1.442857142857143e-06,
"loss": 0.4693,
"step": 3500
},
{
"epoch": 3.36,
"learning_rate": 1.3714285714285717e-06,
"loss": 0.4527,
"step": 3525
},
{
"epoch": 3.39,
"learning_rate": 1.3e-06,
"loss": 0.4624,
"step": 3550
},
{
"epoch": 3.41,
"learning_rate": 1.2285714285714286e-06,
"loss": 0.4745,
"step": 3575
},
{
"epoch": 3.44,
"learning_rate": 1.1571428571428572e-06,
"loss": 0.4406,
"step": 3600
},
{
"epoch": 3.46,
"learning_rate": 1.0857142857142858e-06,
"loss": 0.476,
"step": 3625
},
{
"epoch": 3.48,
"learning_rate": 1.0142857142857144e-06,
"loss": 0.4736,
"step": 3650
},
{
"epoch": 3.51,
"learning_rate": 9.42857142857143e-07,
"loss": 0.4618,
"step": 3675
},
{
"epoch": 3.53,
"learning_rate": 8.714285714285716e-07,
"loss": 0.4581,
"step": 3700
},
{
"epoch": 3.55,
"learning_rate": 8.000000000000001e-07,
"loss": 0.4767,
"step": 3725
},
{
"epoch": 3.58,
"learning_rate": 7.285714285714287e-07,
"loss": 0.4436,
"step": 3750
},
{
"epoch": 3.6,
"learning_rate": 6.571428571428571e-07,
"loss": 0.4595,
"step": 3775
},
{
"epoch": 3.63,
"learning_rate": 5.857142857142857e-07,
"loss": 0.461,
"step": 3800
},
{
"epoch": 3.65,
"learning_rate": 5.142857142857143e-07,
"loss": 0.4506,
"step": 3825
},
{
"epoch": 3.67,
"learning_rate": 4.4285714285714286e-07,
"loss": 0.4434,
"step": 3850
},
{
"epoch": 3.7,
"learning_rate": 3.7142857142857145e-07,
"loss": 0.4722,
"step": 3875
},
{
"epoch": 3.72,
"learning_rate": 3.0000000000000004e-07,
"loss": 0.4963,
"step": 3900
},
{
"epoch": 3.75,
"learning_rate": 2.285714285714286e-07,
"loss": 0.4678,
"step": 3925
},
{
"epoch": 3.77,
"learning_rate": 1.5714285714285717e-07,
"loss": 0.4692,
"step": 3950
},
{
"epoch": 3.79,
"learning_rate": 8.571428571428573e-08,
"loss": 0.4507,
"step": 3975
},
{
"epoch": 3.82,
"learning_rate": 1.4285714285714288e-08,
"loss": 0.4449,
"step": 4000
},
{
"epoch": 3.82,
"eval_loss": 0.999731183052063,
"eval_runtime": 272.3393,
"eval_samples_per_second": 5.508,
"eval_steps_per_second": 0.69,
"eval_wer": 58.36106489184692,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 4,
"total_flos": 1.57509165109248e+18,
"trial_name": null,
"trial_params": null
}