whisper-tiny-it / trainer_state.json
luigisaetta's picture
End of training
ece5ecd
raw
history blame
15.1 kB
{
"best_metric": 0.2660653426588741,
"best_model_checkpoint": "./whisper-tiny-it/checkpoint-3000",
"epoch": 5.025125628140704,
"global_step": 3000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 4.2000000000000006e-07,
"loss": 1.6014,
"step": 25
},
{
"epoch": 0.08,
"learning_rate": 9.200000000000001e-07,
"loss": 1.4799,
"step": 50
},
{
"epoch": 0.13,
"learning_rate": 1.42e-06,
"loss": 1.2794,
"step": 75
},
{
"epoch": 0.17,
"learning_rate": 1.9200000000000003e-06,
"loss": 1.0656,
"step": 100
},
{
"epoch": 0.21,
"learning_rate": 2.42e-06,
"loss": 0.8907,
"step": 125
},
{
"epoch": 0.25,
"learning_rate": 2.92e-06,
"loss": 0.8231,
"step": 150
},
{
"epoch": 0.29,
"learning_rate": 3.4200000000000007e-06,
"loss": 0.7702,
"step": 175
},
{
"epoch": 0.34,
"learning_rate": 3.920000000000001e-06,
"loss": 0.7254,
"step": 200
},
{
"epoch": 0.38,
"learning_rate": 4.42e-06,
"loss": 0.6913,
"step": 225
},
{
"epoch": 0.42,
"learning_rate": 4.92e-06,
"loss": 0.673,
"step": 250
},
{
"epoch": 0.46,
"learning_rate": 5.420000000000001e-06,
"loss": 0.6347,
"step": 275
},
{
"epoch": 0.5,
"learning_rate": 5.92e-06,
"loss": 0.6315,
"step": 300
},
{
"epoch": 0.54,
"learning_rate": 6.42e-06,
"loss": 0.611,
"step": 325
},
{
"epoch": 0.59,
"learning_rate": 6.92e-06,
"loss": 0.5992,
"step": 350
},
{
"epoch": 0.63,
"learning_rate": 7.420000000000001e-06,
"loss": 0.5815,
"step": 375
},
{
"epoch": 0.67,
"learning_rate": 7.92e-06,
"loss": 0.5757,
"step": 400
},
{
"epoch": 0.71,
"learning_rate": 8.42e-06,
"loss": 0.5605,
"step": 425
},
{
"epoch": 0.75,
"learning_rate": 8.920000000000001e-06,
"loss": 0.541,
"step": 450
},
{
"epoch": 0.8,
"learning_rate": 9.42e-06,
"loss": 0.5292,
"step": 475
},
{
"epoch": 0.84,
"learning_rate": 9.920000000000002e-06,
"loss": 0.5198,
"step": 500
},
{
"epoch": 0.88,
"learning_rate": 9.916000000000001e-06,
"loss": 0.5194,
"step": 525
},
{
"epoch": 0.92,
"learning_rate": 9.816e-06,
"loss": 0.5057,
"step": 550
},
{
"epoch": 0.96,
"learning_rate": 9.716000000000002e-06,
"loss": 0.4999,
"step": 575
},
{
"epoch": 1.01,
"learning_rate": 9.616e-06,
"loss": 0.4956,
"step": 600
},
{
"epoch": 1.05,
"learning_rate": 9.516e-06,
"loss": 0.4549,
"step": 625
},
{
"epoch": 1.09,
"learning_rate": 9.416000000000001e-06,
"loss": 0.448,
"step": 650
},
{
"epoch": 1.13,
"learning_rate": 9.316e-06,
"loss": 0.4505,
"step": 675
},
{
"epoch": 1.17,
"learning_rate": 9.216000000000001e-06,
"loss": 0.4407,
"step": 700
},
{
"epoch": 1.21,
"learning_rate": 9.116e-06,
"loss": 0.4432,
"step": 725
},
{
"epoch": 1.26,
"learning_rate": 9.016e-06,
"loss": 0.4384,
"step": 750
},
{
"epoch": 1.3,
"learning_rate": 8.916e-06,
"loss": 0.4335,
"step": 775
},
{
"epoch": 1.34,
"learning_rate": 8.816000000000002e-06,
"loss": 0.4328,
"step": 800
},
{
"epoch": 1.38,
"learning_rate": 8.716000000000001e-06,
"loss": 0.423,
"step": 825
},
{
"epoch": 1.42,
"learning_rate": 8.616000000000002e-06,
"loss": 0.4225,
"step": 850
},
{
"epoch": 1.47,
"learning_rate": 8.516000000000001e-06,
"loss": 0.4171,
"step": 875
},
{
"epoch": 1.51,
"learning_rate": 8.416e-06,
"loss": 0.4132,
"step": 900
},
{
"epoch": 1.55,
"learning_rate": 8.316000000000001e-06,
"loss": 0.408,
"step": 925
},
{
"epoch": 1.59,
"learning_rate": 8.216e-06,
"loss": 0.406,
"step": 950
},
{
"epoch": 1.63,
"learning_rate": 8.116e-06,
"loss": 0.4044,
"step": 975
},
{
"epoch": 1.68,
"learning_rate": 8.016e-06,
"loss": 0.3997,
"step": 1000
},
{
"epoch": 1.68,
"eval_loss": 0.4650561511516571,
"eval_runtime": 819.8227,
"eval_samples_per_second": 18.293,
"eval_steps_per_second": 0.144,
"eval_wer": 0.30382830785234255,
"step": 1000
},
{
"epoch": 1.72,
"learning_rate": 7.916e-06,
"loss": 0.402,
"step": 1025
},
{
"epoch": 1.76,
"learning_rate": 7.816000000000001e-06,
"loss": 0.3956,
"step": 1050
},
{
"epoch": 1.8,
"learning_rate": 7.716e-06,
"loss": 0.3886,
"step": 1075
},
{
"epoch": 1.84,
"learning_rate": 7.616000000000001e-06,
"loss": 0.3929,
"step": 1100
},
{
"epoch": 1.88,
"learning_rate": 7.516000000000001e-06,
"loss": 0.3879,
"step": 1125
},
{
"epoch": 1.93,
"learning_rate": 7.416000000000001e-06,
"loss": 0.3882,
"step": 1150
},
{
"epoch": 1.97,
"learning_rate": 7.316000000000001e-06,
"loss": 0.3843,
"step": 1175
},
{
"epoch": 2.01,
"learning_rate": 7.216000000000001e-06,
"loss": 0.3716,
"step": 1200
},
{
"epoch": 2.05,
"learning_rate": 7.116000000000001e-06,
"loss": 0.3502,
"step": 1225
},
{
"epoch": 2.09,
"learning_rate": 7.016e-06,
"loss": 0.3528,
"step": 1250
},
{
"epoch": 2.14,
"learning_rate": 6.916e-06,
"loss": 0.3514,
"step": 1275
},
{
"epoch": 2.18,
"learning_rate": 6.8160000000000005e-06,
"loss": 0.3501,
"step": 1300
},
{
"epoch": 2.22,
"learning_rate": 6.716000000000001e-06,
"loss": 0.3558,
"step": 1325
},
{
"epoch": 2.26,
"learning_rate": 6.616e-06,
"loss": 0.3459,
"step": 1350
},
{
"epoch": 2.3,
"learning_rate": 6.516e-06,
"loss": 0.3466,
"step": 1375
},
{
"epoch": 2.35,
"learning_rate": 6.416e-06,
"loss": 0.3447,
"step": 1400
},
{
"epoch": 2.39,
"learning_rate": 6.316000000000001e-06,
"loss": 0.3413,
"step": 1425
},
{
"epoch": 2.43,
"learning_rate": 6.216000000000001e-06,
"loss": 0.3418,
"step": 1450
},
{
"epoch": 2.47,
"learning_rate": 6.116000000000001e-06,
"loss": 0.3407,
"step": 1475
},
{
"epoch": 2.51,
"learning_rate": 6.0160000000000005e-06,
"loss": 0.3356,
"step": 1500
},
{
"epoch": 2.55,
"learning_rate": 5.916000000000001e-06,
"loss": 0.3379,
"step": 1525
},
{
"epoch": 2.6,
"learning_rate": 5.816000000000001e-06,
"loss": 0.3358,
"step": 1550
},
{
"epoch": 2.64,
"learning_rate": 5.716000000000001e-06,
"loss": 0.3408,
"step": 1575
},
{
"epoch": 2.68,
"learning_rate": 5.616e-06,
"loss": 0.3377,
"step": 1600
},
{
"epoch": 2.72,
"learning_rate": 5.516e-06,
"loss": 0.3263,
"step": 1625
},
{
"epoch": 2.76,
"learning_rate": 5.416e-06,
"loss": 0.3305,
"step": 1650
},
{
"epoch": 2.81,
"learning_rate": 5.3160000000000004e-06,
"loss": 0.3284,
"step": 1675
},
{
"epoch": 2.85,
"learning_rate": 5.216e-06,
"loss": 0.3347,
"step": 1700
},
{
"epoch": 2.89,
"learning_rate": 5.116000000000001e-06,
"loss": 0.3375,
"step": 1725
},
{
"epoch": 2.93,
"learning_rate": 5.016000000000001e-06,
"loss": 0.3279,
"step": 1750
},
{
"epoch": 2.97,
"learning_rate": 4.916e-06,
"loss": 0.3262,
"step": 1775
},
{
"epoch": 3.02,
"learning_rate": 4.816e-06,
"loss": 0.3264,
"step": 1800
},
{
"epoch": 3.06,
"learning_rate": 4.716e-06,
"loss": 0.3137,
"step": 1825
},
{
"epoch": 3.1,
"learning_rate": 4.616e-06,
"loss": 0.3087,
"step": 1850
},
{
"epoch": 3.14,
"learning_rate": 4.5160000000000005e-06,
"loss": 0.3107,
"step": 1875
},
{
"epoch": 3.18,
"learning_rate": 4.416000000000001e-06,
"loss": 0.3086,
"step": 1900
},
{
"epoch": 3.22,
"learning_rate": 4.316e-06,
"loss": 0.2998,
"step": 1925
},
{
"epoch": 3.27,
"learning_rate": 4.216e-06,
"loss": 0.3079,
"step": 1950
},
{
"epoch": 3.31,
"learning_rate": 4.116000000000001e-06,
"loss": 0.3013,
"step": 1975
},
{
"epoch": 3.35,
"learning_rate": 4.016e-06,
"loss": 0.3026,
"step": 2000
},
{
"epoch": 3.35,
"eval_loss": 0.4085945188999176,
"eval_runtime": 821.5774,
"eval_samples_per_second": 18.254,
"eval_steps_per_second": 0.144,
"eval_wer": 0.274339250763061,
"step": 2000
},
{
"epoch": 3.39,
"learning_rate": 3.916e-06,
"loss": 0.3047,
"step": 2025
},
{
"epoch": 3.43,
"learning_rate": 3.816e-06,
"loss": 0.3017,
"step": 2050
},
{
"epoch": 3.48,
"learning_rate": 3.716e-06,
"loss": 0.3049,
"step": 2075
},
{
"epoch": 3.52,
"learning_rate": 3.616e-06,
"loss": 0.31,
"step": 2100
},
{
"epoch": 3.56,
"learning_rate": 3.5160000000000007e-06,
"loss": 0.3018,
"step": 2125
},
{
"epoch": 3.6,
"learning_rate": 3.4160000000000004e-06,
"loss": 0.3013,
"step": 2150
},
{
"epoch": 3.64,
"learning_rate": 3.3160000000000005e-06,
"loss": 0.304,
"step": 2175
},
{
"epoch": 3.69,
"learning_rate": 3.216e-06,
"loss": 0.3021,
"step": 2200
},
{
"epoch": 3.73,
"learning_rate": 3.1160000000000003e-06,
"loss": 0.298,
"step": 2225
},
{
"epoch": 3.77,
"learning_rate": 3.016e-06,
"loss": 0.298,
"step": 2250
},
{
"epoch": 3.81,
"learning_rate": 2.9160000000000005e-06,
"loss": 0.2978,
"step": 2275
},
{
"epoch": 3.85,
"learning_rate": 2.8160000000000002e-06,
"loss": 0.3026,
"step": 2300
},
{
"epoch": 3.89,
"learning_rate": 2.7160000000000003e-06,
"loss": 0.2967,
"step": 2325
},
{
"epoch": 3.94,
"learning_rate": 2.616e-06,
"loss": 0.2939,
"step": 2350
},
{
"epoch": 3.98,
"learning_rate": 2.516e-06,
"loss": 0.2975,
"step": 2375
},
{
"epoch": 4.02,
"learning_rate": 2.4160000000000002e-06,
"loss": 0.2932,
"step": 2400
},
{
"epoch": 4.06,
"learning_rate": 2.3160000000000004e-06,
"loss": 0.2867,
"step": 2425
},
{
"epoch": 4.1,
"learning_rate": 2.216e-06,
"loss": 0.2868,
"step": 2450
},
{
"epoch": 4.15,
"learning_rate": 2.116e-06,
"loss": 0.2851,
"step": 2475
},
{
"epoch": 4.19,
"learning_rate": 2.0160000000000003e-06,
"loss": 0.2788,
"step": 2500
},
{
"epoch": 4.23,
"learning_rate": 1.916e-06,
"loss": 0.2804,
"step": 2525
},
{
"epoch": 4.27,
"learning_rate": 1.8160000000000003e-06,
"loss": 0.2863,
"step": 2550
},
{
"epoch": 4.31,
"learning_rate": 1.7160000000000002e-06,
"loss": 0.2794,
"step": 2575
},
{
"epoch": 4.36,
"learning_rate": 1.616e-06,
"loss": 0.2845,
"step": 2600
},
{
"epoch": 4.4,
"learning_rate": 1.5160000000000002e-06,
"loss": 0.2835,
"step": 2625
},
{
"epoch": 4.44,
"learning_rate": 1.416e-06,
"loss": 0.2814,
"step": 2650
},
{
"epoch": 4.48,
"learning_rate": 1.316e-06,
"loss": 0.2867,
"step": 2675
},
{
"epoch": 4.52,
"learning_rate": 1.216e-06,
"loss": 0.2803,
"step": 2700
},
{
"epoch": 4.56,
"learning_rate": 1.1160000000000002e-06,
"loss": 0.2845,
"step": 2725
},
{
"epoch": 4.61,
"learning_rate": 1.016e-06,
"loss": 0.2815,
"step": 2750
},
{
"epoch": 4.65,
"learning_rate": 9.160000000000001e-07,
"loss": 0.2837,
"step": 2775
},
{
"epoch": 4.69,
"learning_rate": 8.160000000000001e-07,
"loss": 0.2884,
"step": 2800
},
{
"epoch": 4.73,
"learning_rate": 7.16e-07,
"loss": 0.2853,
"step": 2825
},
{
"epoch": 4.77,
"learning_rate": 6.160000000000001e-07,
"loss": 0.284,
"step": 2850
},
{
"epoch": 4.82,
"learning_rate": 5.16e-07,
"loss": 0.2822,
"step": 2875
},
{
"epoch": 4.86,
"learning_rate": 4.16e-07,
"loss": 0.2876,
"step": 2900
},
{
"epoch": 4.9,
"learning_rate": 3.160000000000001e-07,
"loss": 0.2804,
"step": 2925
},
{
"epoch": 4.94,
"learning_rate": 2.1600000000000003e-07,
"loss": 0.2828,
"step": 2950
},
{
"epoch": 4.98,
"learning_rate": 1.16e-07,
"loss": 0.2833,
"step": 2975
},
{
"epoch": 5.03,
"learning_rate": 1.6e-08,
"loss": 0.2874,
"step": 3000
},
{
"epoch": 5.03,
"eval_loss": 0.3958355188369751,
"eval_runtime": 842.4001,
"eval_samples_per_second": 17.803,
"eval_steps_per_second": 0.14,
"eval_wer": 0.2660653426588741,
"step": 3000
},
{
"epoch": 5.03,
"step": 3000,
"total_flos": 1.8880222255172813e+19,
"train_loss": 0.4164268921216329,
"train_runtime": 23892.2245,
"train_samples_per_second": 32.144,
"train_steps_per_second": 0.126
}
],
"max_steps": 3000,
"num_train_epochs": 6,
"total_flos": 1.8880222255172813e+19,
"trial_name": null,
"trial_params": null
}