whisper-large-ur / trainer_state.json
testphase73's picture
Upload 10 files
4d66165
{
"best_metric": 75.2080188861798,
"best_model_checkpoint": "./whisper-large-ur/checkpoint-4000",
"epoch": 1.0764262648008611,
"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": 3.8e-07,
"loss": 4.3261,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 8.8e-07,
"loss": 3.7304,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 1.3800000000000001e-06,
"loss": 2.49,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 1.8800000000000002e-06,
"loss": 2.0671,
"step": 100
},
{
"epoch": 0.03,
"learning_rate": 2.38e-06,
"loss": 1.6215,
"step": 125
},
{
"epoch": 0.04,
"learning_rate": 2.88e-06,
"loss": 1.3597,
"step": 150
},
{
"epoch": 0.05,
"learning_rate": 3.3800000000000007e-06,
"loss": 1.3107,
"step": 175
},
{
"epoch": 0.05,
"learning_rate": 3.88e-06,
"loss": 1.2027,
"step": 200
},
{
"epoch": 0.06,
"learning_rate": 4.38e-06,
"loss": 1.1403,
"step": 225
},
{
"epoch": 0.07,
"learning_rate": 4.880000000000001e-06,
"loss": 1.2392,
"step": 250
},
{
"epoch": 0.07,
"learning_rate": 5.380000000000001e-06,
"loss": 1.1253,
"step": 275
},
{
"epoch": 0.08,
"learning_rate": 5.8800000000000005e-06,
"loss": 1.003,
"step": 300
},
{
"epoch": 0.09,
"learning_rate": 6.380000000000001e-06,
"loss": 1.2717,
"step": 325
},
{
"epoch": 0.09,
"learning_rate": 6.88e-06,
"loss": 1.0027,
"step": 350
},
{
"epoch": 0.1,
"learning_rate": 7.3800000000000005e-06,
"loss": 0.9987,
"step": 375
},
{
"epoch": 0.11,
"learning_rate": 7.88e-06,
"loss": 1.0718,
"step": 400
},
{
"epoch": 0.11,
"learning_rate": 8.380000000000001e-06,
"loss": 1.1424,
"step": 425
},
{
"epoch": 0.12,
"learning_rate": 8.880000000000001e-06,
"loss": 1.0544,
"step": 450
},
{
"epoch": 0.13,
"learning_rate": 9.38e-06,
"loss": 1.0535,
"step": 475
},
{
"epoch": 0.13,
"learning_rate": 9.88e-06,
"loss": 1.0831,
"step": 500
},
{
"epoch": 0.14,
"learning_rate": 9.945714285714286e-06,
"loss": 0.9913,
"step": 525
},
{
"epoch": 0.15,
"learning_rate": 9.874285714285715e-06,
"loss": 1.0124,
"step": 550
},
{
"epoch": 0.15,
"learning_rate": 9.802857142857144e-06,
"loss": 1.0512,
"step": 575
},
{
"epoch": 0.16,
"learning_rate": 9.731428571428573e-06,
"loss": 0.9447,
"step": 600
},
{
"epoch": 0.17,
"learning_rate": 9.66e-06,
"loss": 0.9097,
"step": 625
},
{
"epoch": 0.17,
"learning_rate": 9.58857142857143e-06,
"loss": 1.0378,
"step": 650
},
{
"epoch": 0.18,
"learning_rate": 9.517142857142859e-06,
"loss": 0.9416,
"step": 675
},
{
"epoch": 0.19,
"learning_rate": 9.445714285714288e-06,
"loss": 0.9182,
"step": 700
},
{
"epoch": 0.2,
"learning_rate": 9.374285714285715e-06,
"loss": 0.973,
"step": 725
},
{
"epoch": 0.2,
"learning_rate": 9.302857142857144e-06,
"loss": 0.9704,
"step": 750
},
{
"epoch": 0.21,
"learning_rate": 9.231428571428573e-06,
"loss": 0.8628,
"step": 775
},
{
"epoch": 0.22,
"learning_rate": 9.16e-06,
"loss": 0.9047,
"step": 800
},
{
"epoch": 0.22,
"learning_rate": 9.08857142857143e-06,
"loss": 0.9525,
"step": 825
},
{
"epoch": 0.23,
"learning_rate": 9.017142857142858e-06,
"loss": 0.9475,
"step": 850
},
{
"epoch": 0.24,
"learning_rate": 8.945714285714286e-06,
"loss": 0.845,
"step": 875
},
{
"epoch": 0.24,
"learning_rate": 8.874285714285715e-06,
"loss": 1.0463,
"step": 900
},
{
"epoch": 0.25,
"learning_rate": 8.802857142857144e-06,
"loss": 0.8325,
"step": 925
},
{
"epoch": 0.26,
"learning_rate": 8.731428571428571e-06,
"loss": 0.7611,
"step": 950
},
{
"epoch": 0.26,
"learning_rate": 8.66e-06,
"loss": 0.8264,
"step": 975
},
{
"epoch": 0.27,
"learning_rate": 8.588571428571429e-06,
"loss": 0.8408,
"step": 1000
},
{
"epoch": 0.27,
"eval_loss": 0.9657530188560486,
"eval_runtime": 867.2046,
"eval_samples_per_second": 3.808,
"eval_steps_per_second": 0.476,
"eval_wer": 168.88424474631373,
"step": 1000
},
{
"epoch": 0.28,
"learning_rate": 8.517142857142858e-06,
"loss": 0.8489,
"step": 1025
},
{
"epoch": 0.28,
"learning_rate": 8.445714285714285e-06,
"loss": 0.8329,
"step": 1050
},
{
"epoch": 0.29,
"learning_rate": 8.374285714285714e-06,
"loss": 0.8605,
"step": 1075
},
{
"epoch": 0.3,
"learning_rate": 8.302857142857143e-06,
"loss": 0.8589,
"step": 1100
},
{
"epoch": 0.3,
"learning_rate": 8.231428571428572e-06,
"loss": 0.9028,
"step": 1125
},
{
"epoch": 0.31,
"learning_rate": 8.16e-06,
"loss": 0.9279,
"step": 1150
},
{
"epoch": 0.32,
"learning_rate": 8.088571428571429e-06,
"loss": 0.8998,
"step": 1175
},
{
"epoch": 0.32,
"learning_rate": 8.017142857142858e-06,
"loss": 0.7627,
"step": 1200
},
{
"epoch": 0.33,
"learning_rate": 7.945714285714287e-06,
"loss": 0.7937,
"step": 1225
},
{
"epoch": 0.34,
"learning_rate": 7.874285714285716e-06,
"loss": 0.8665,
"step": 1250
},
{
"epoch": 0.34,
"learning_rate": 7.802857142857143e-06,
"loss": 0.9183,
"step": 1275
},
{
"epoch": 0.35,
"learning_rate": 7.731428571428572e-06,
"loss": 0.9306,
"step": 1300
},
{
"epoch": 0.36,
"learning_rate": 7.660000000000001e-06,
"loss": 0.8631,
"step": 1325
},
{
"epoch": 0.36,
"learning_rate": 7.588571428571429e-06,
"loss": 0.7727,
"step": 1350
},
{
"epoch": 0.37,
"learning_rate": 7.5171428571428575e-06,
"loss": 0.8144,
"step": 1375
},
{
"epoch": 0.38,
"learning_rate": 7.445714285714286e-06,
"loss": 0.8637,
"step": 1400
},
{
"epoch": 0.38,
"learning_rate": 7.374285714285715e-06,
"loss": 0.7421,
"step": 1425
},
{
"epoch": 0.39,
"learning_rate": 7.302857142857144e-06,
"loss": 0.855,
"step": 1450
},
{
"epoch": 0.4,
"learning_rate": 7.231428571428573e-06,
"loss": 0.8864,
"step": 1475
},
{
"epoch": 0.4,
"learning_rate": 7.16e-06,
"loss": 0.8096,
"step": 1500
},
{
"epoch": 0.41,
"learning_rate": 7.088571428571429e-06,
"loss": 0.7856,
"step": 1525
},
{
"epoch": 0.42,
"learning_rate": 7.017142857142858e-06,
"loss": 0.858,
"step": 1550
},
{
"epoch": 0.42,
"learning_rate": 6.945714285714287e-06,
"loss": 0.8314,
"step": 1575
},
{
"epoch": 0.43,
"learning_rate": 6.874285714285714e-06,
"loss": 0.8743,
"step": 1600
},
{
"epoch": 0.44,
"learning_rate": 6.8028571428571434e-06,
"loss": 0.835,
"step": 1625
},
{
"epoch": 0.44,
"learning_rate": 6.7314285714285724e-06,
"loss": 0.7496,
"step": 1650
},
{
"epoch": 0.45,
"learning_rate": 6.660000000000001e-06,
"loss": 0.8331,
"step": 1675
},
{
"epoch": 0.46,
"learning_rate": 6.588571428571429e-06,
"loss": 0.7762,
"step": 1700
},
{
"epoch": 0.46,
"learning_rate": 6.517142857142858e-06,
"loss": 0.8316,
"step": 1725
},
{
"epoch": 0.47,
"learning_rate": 6.445714285714286e-06,
"loss": 0.8765,
"step": 1750
},
{
"epoch": 0.48,
"learning_rate": 6.374285714285715e-06,
"loss": 0.7997,
"step": 1775
},
{
"epoch": 0.48,
"learning_rate": 6.302857142857144e-06,
"loss": 0.9058,
"step": 1800
},
{
"epoch": 0.49,
"learning_rate": 6.231428571428571e-06,
"loss": 0.76,
"step": 1825
},
{
"epoch": 0.5,
"learning_rate": 6.16e-06,
"loss": 0.8158,
"step": 1850
},
{
"epoch": 0.5,
"learning_rate": 6.088571428571429e-06,
"loss": 0.8641,
"step": 1875
},
{
"epoch": 0.51,
"learning_rate": 6.017142857142858e-06,
"loss": 0.8444,
"step": 1900
},
{
"epoch": 0.52,
"learning_rate": 5.945714285714286e-06,
"loss": 0.6553,
"step": 1925
},
{
"epoch": 0.52,
"learning_rate": 5.874285714285715e-06,
"loss": 0.7227,
"step": 1950
},
{
"epoch": 0.53,
"learning_rate": 5.802857142857144e-06,
"loss": 0.6912,
"step": 1975
},
{
"epoch": 0.54,
"learning_rate": 5.731428571428572e-06,
"loss": 0.8024,
"step": 2000
},
{
"epoch": 0.54,
"eval_loss": 0.8614732027053833,
"eval_runtime": 751.5965,
"eval_samples_per_second": 4.393,
"eval_steps_per_second": 0.549,
"eval_wer": 79.90634312473392,
"step": 2000
},
{
"epoch": 0.54,
"learning_rate": 5.66e-06,
"loss": 0.6802,
"step": 2025
},
{
"epoch": 0.55,
"learning_rate": 5.588571428571429e-06,
"loss": 0.8505,
"step": 2050
},
{
"epoch": 0.56,
"learning_rate": 5.517142857142857e-06,
"loss": 0.7384,
"step": 2075
},
{
"epoch": 0.57,
"learning_rate": 5.445714285714286e-06,
"loss": 0.7666,
"step": 2100
},
{
"epoch": 0.57,
"learning_rate": 5.374285714285715e-06,
"loss": 0.7384,
"step": 2125
},
{
"epoch": 0.58,
"learning_rate": 5.3028571428571425e-06,
"loss": 0.6857,
"step": 2150
},
{
"epoch": 0.59,
"learning_rate": 5.2314285714285716e-06,
"loss": 0.845,
"step": 2175
},
{
"epoch": 0.59,
"learning_rate": 5.1600000000000006e-06,
"loss": 0.7361,
"step": 2200
},
{
"epoch": 0.6,
"learning_rate": 5.08857142857143e-06,
"loss": 0.7968,
"step": 2225
},
{
"epoch": 0.61,
"learning_rate": 5.017142857142857e-06,
"loss": 0.7221,
"step": 2250
},
{
"epoch": 0.61,
"learning_rate": 4.945714285714286e-06,
"loss": 0.7486,
"step": 2275
},
{
"epoch": 0.62,
"learning_rate": 4.874285714285715e-06,
"loss": 0.8008,
"step": 2300
},
{
"epoch": 0.63,
"learning_rate": 4.802857142857143e-06,
"loss": 0.7099,
"step": 2325
},
{
"epoch": 0.63,
"learning_rate": 4.731428571428572e-06,
"loss": 0.7624,
"step": 2350
},
{
"epoch": 0.64,
"learning_rate": 4.66e-06,
"loss": 0.7452,
"step": 2375
},
{
"epoch": 0.65,
"learning_rate": 4.588571428571429e-06,
"loss": 0.7053,
"step": 2400
},
{
"epoch": 0.65,
"learning_rate": 4.5171428571428575e-06,
"loss": 0.7442,
"step": 2425
},
{
"epoch": 0.66,
"learning_rate": 4.445714285714286e-06,
"loss": 0.8031,
"step": 2450
},
{
"epoch": 0.67,
"learning_rate": 4.374285714285715e-06,
"loss": 0.7198,
"step": 2475
},
{
"epoch": 0.67,
"learning_rate": 4.302857142857143e-06,
"loss": 0.7593,
"step": 2500
},
{
"epoch": 0.68,
"learning_rate": 4.231428571428572e-06,
"loss": 0.7465,
"step": 2525
},
{
"epoch": 0.69,
"learning_rate": 4.16e-06,
"loss": 0.6923,
"step": 2550
},
{
"epoch": 0.69,
"learning_rate": 4.088571428571429e-06,
"loss": 0.7078,
"step": 2575
},
{
"epoch": 0.7,
"learning_rate": 4.017142857142857e-06,
"loss": 0.7868,
"step": 2600
},
{
"epoch": 0.71,
"learning_rate": 3.945714285714286e-06,
"loss": 0.7817,
"step": 2625
},
{
"epoch": 0.71,
"learning_rate": 3.874285714285715e-06,
"loss": 0.8695,
"step": 2650
},
{
"epoch": 0.72,
"learning_rate": 3.802857142857143e-06,
"loss": 0.6811,
"step": 2675
},
{
"epoch": 0.73,
"learning_rate": 3.731428571428572e-06,
"loss": 0.773,
"step": 2700
},
{
"epoch": 0.73,
"learning_rate": 3.66e-06,
"loss": 0.8696,
"step": 2725
},
{
"epoch": 0.74,
"learning_rate": 3.588571428571429e-06,
"loss": 0.6929,
"step": 2750
},
{
"epoch": 0.75,
"learning_rate": 3.5171428571428573e-06,
"loss": 0.6453,
"step": 2775
},
{
"epoch": 0.75,
"learning_rate": 3.4457142857142863e-06,
"loss": 0.7387,
"step": 2800
},
{
"epoch": 0.76,
"learning_rate": 3.3742857142857145e-06,
"loss": 0.6669,
"step": 2825
},
{
"epoch": 0.77,
"learning_rate": 3.302857142857143e-06,
"loss": 0.6659,
"step": 2850
},
{
"epoch": 0.77,
"learning_rate": 3.2314285714285716e-06,
"loss": 0.6333,
"step": 2875
},
{
"epoch": 0.78,
"learning_rate": 3.1600000000000002e-06,
"loss": 0.6892,
"step": 2900
},
{
"epoch": 0.79,
"learning_rate": 3.0885714285714284e-06,
"loss": 0.7451,
"step": 2925
},
{
"epoch": 0.79,
"learning_rate": 3.0171428571428574e-06,
"loss": 0.7031,
"step": 2950
},
{
"epoch": 0.8,
"learning_rate": 2.945714285714286e-06,
"loss": 0.7443,
"step": 2975
},
{
"epoch": 0.81,
"learning_rate": 2.8742857142857146e-06,
"loss": 0.735,
"step": 3000
},
{
"epoch": 0.81,
"eval_loss": 0.8074357509613037,
"eval_runtime": 759.2924,
"eval_samples_per_second": 4.349,
"eval_steps_per_second": 0.544,
"eval_wer": 84.75560199698131,
"step": 3000
},
{
"epoch": 0.81,
"learning_rate": 2.802857142857143e-06,
"loss": 0.7539,
"step": 3025
},
{
"epoch": 0.82,
"learning_rate": 2.7314285714285714e-06,
"loss": 0.7019,
"step": 3050
},
{
"epoch": 0.83,
"learning_rate": 2.6600000000000004e-06,
"loss": 0.6256,
"step": 3075
},
{
"epoch": 0.83,
"learning_rate": 2.5885714285714285e-06,
"loss": 0.7012,
"step": 3100
},
{
"epoch": 0.84,
"learning_rate": 2.5171428571428575e-06,
"loss": 0.5999,
"step": 3125
},
{
"epoch": 0.85,
"learning_rate": 2.445714285714286e-06,
"loss": 0.7535,
"step": 3150
},
{
"epoch": 0.85,
"learning_rate": 2.3742857142857147e-06,
"loss": 0.6353,
"step": 3175
},
{
"epoch": 0.86,
"learning_rate": 2.302857142857143e-06,
"loss": 0.6801,
"step": 3200
},
{
"epoch": 0.87,
"learning_rate": 2.2314285714285715e-06,
"loss": 0.6153,
"step": 3225
},
{
"epoch": 0.87,
"learning_rate": 2.16e-06,
"loss": 0.71,
"step": 3250
},
{
"epoch": 0.88,
"learning_rate": 2.0885714285714287e-06,
"loss": 0.6666,
"step": 3275
},
{
"epoch": 0.89,
"learning_rate": 2.0171428571428573e-06,
"loss": 0.6849,
"step": 3300
},
{
"epoch": 0.89,
"learning_rate": 1.945714285714286e-06,
"loss": 0.6465,
"step": 3325
},
{
"epoch": 0.9,
"learning_rate": 1.8742857142857142e-06,
"loss": 0.7247,
"step": 3350
},
{
"epoch": 0.91,
"learning_rate": 1.8028571428571432e-06,
"loss": 0.5913,
"step": 3375
},
{
"epoch": 0.91,
"learning_rate": 1.7314285714285716e-06,
"loss": 0.6817,
"step": 3400
},
{
"epoch": 0.92,
"learning_rate": 1.6600000000000002e-06,
"loss": 0.7188,
"step": 3425
},
{
"epoch": 0.93,
"learning_rate": 1.5885714285714288e-06,
"loss": 0.6657,
"step": 3450
},
{
"epoch": 0.94,
"learning_rate": 1.5171428571428574e-06,
"loss": 0.604,
"step": 3475
},
{
"epoch": 0.94,
"learning_rate": 1.4457142857142858e-06,
"loss": 0.6521,
"step": 3500
},
{
"epoch": 0.95,
"learning_rate": 1.3742857142857143e-06,
"loss": 0.6956,
"step": 3525
},
{
"epoch": 0.96,
"learning_rate": 1.302857142857143e-06,
"loss": 0.6587,
"step": 3550
},
{
"epoch": 0.96,
"learning_rate": 1.2314285714285715e-06,
"loss": 0.7181,
"step": 3575
},
{
"epoch": 0.97,
"learning_rate": 1.1600000000000001e-06,
"loss": 0.7944,
"step": 3600
},
{
"epoch": 0.98,
"learning_rate": 1.0885714285714287e-06,
"loss": 0.7407,
"step": 3625
},
{
"epoch": 0.98,
"learning_rate": 1.0171428571428573e-06,
"loss": 0.6756,
"step": 3650
},
{
"epoch": 0.99,
"learning_rate": 9.457142857142858e-07,
"loss": 0.696,
"step": 3675
},
{
"epoch": 1.0,
"learning_rate": 8.742857142857144e-07,
"loss": 0.6427,
"step": 3700
},
{
"epoch": 1.0,
"learning_rate": 8.028571428571429e-07,
"loss": 0.6364,
"step": 3725
},
{
"epoch": 1.01,
"learning_rate": 7.314285714285715e-07,
"loss": 0.6172,
"step": 3750
},
{
"epoch": 1.02,
"learning_rate": 6.6e-07,
"loss": 0.5513,
"step": 3775
},
{
"epoch": 1.02,
"learning_rate": 5.885714285714286e-07,
"loss": 0.5374,
"step": 3800
},
{
"epoch": 1.03,
"learning_rate": 5.171428571428572e-07,
"loss": 0.5309,
"step": 3825
},
{
"epoch": 1.04,
"learning_rate": 4.457142857142858e-07,
"loss": 0.564,
"step": 3850
},
{
"epoch": 1.04,
"learning_rate": 3.7428571428571434e-07,
"loss": 0.5684,
"step": 3875
},
{
"epoch": 1.05,
"learning_rate": 3.028571428571429e-07,
"loss": 0.6113,
"step": 3900
},
{
"epoch": 1.06,
"learning_rate": 2.3142857142857144e-07,
"loss": 0.5005,
"step": 3925
},
{
"epoch": 1.06,
"learning_rate": 1.6e-07,
"loss": 0.5409,
"step": 3950
},
{
"epoch": 1.07,
"learning_rate": 8.857142857142858e-08,
"loss": 0.6808,
"step": 3975
},
{
"epoch": 1.08,
"learning_rate": 1.7142857142857143e-08,
"loss": 0.692,
"step": 4000
},
{
"epoch": 1.08,
"eval_loss": 0.7873697876930237,
"eval_runtime": 749.5073,
"eval_samples_per_second": 4.406,
"eval_steps_per_second": 0.551,
"eval_wer": 75.2080188861798,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 2,
"total_flos": 1.9695108096e+17,
"trial_name": null,
"trial_params": null
}