lab2_id2223 / checkpoint-3000 /trainer_state.json
humeur's picture
Trainer step-3000
37e4c66
raw
history blame
15.4 kB
{
"best_metric": 19.299033924960682,
"best_model_checkpoint": "./lab2_id2223/checkpoint-3000",
"epoch": 3.8809831824062098,
"global_step": 3000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 4.800000000000001e-07,
"loss": 0.2225,
"step": 25
},
{
"epoch": 0.06,
"learning_rate": 9.800000000000001e-07,
"loss": 0.2081,
"step": 50
},
{
"epoch": 0.1,
"learning_rate": 1.48e-06,
"loss": 0.1992,
"step": 75
},
{
"epoch": 0.13,
"learning_rate": 1.98e-06,
"loss": 0.2014,
"step": 100
},
{
"epoch": 0.16,
"learning_rate": 2.4800000000000004e-06,
"loss": 0.1914,
"step": 125
},
{
"epoch": 0.19,
"learning_rate": 2.9800000000000003e-06,
"loss": 0.1978,
"step": 150
},
{
"epoch": 0.23,
"learning_rate": 3.48e-06,
"loss": 0.1988,
"step": 175
},
{
"epoch": 0.26,
"learning_rate": 3.980000000000001e-06,
"loss": 0.2153,
"step": 200
},
{
"epoch": 0.29,
"learning_rate": 4.48e-06,
"loss": 0.1907,
"step": 225
},
{
"epoch": 0.32,
"learning_rate": 4.980000000000001e-06,
"loss": 0.1976,
"step": 250
},
{
"epoch": 0.36,
"learning_rate": 5.480000000000001e-06,
"loss": 0.2256,
"step": 275
},
{
"epoch": 0.39,
"learning_rate": 5.98e-06,
"loss": 0.1805,
"step": 300
},
{
"epoch": 0.42,
"learning_rate": 6.480000000000001e-06,
"loss": 0.2144,
"step": 325
},
{
"epoch": 0.45,
"learning_rate": 6.98e-06,
"loss": 0.195,
"step": 350
},
{
"epoch": 0.49,
"learning_rate": 7.48e-06,
"loss": 0.2053,
"step": 375
},
{
"epoch": 0.52,
"learning_rate": 7.980000000000002e-06,
"loss": 0.2157,
"step": 400
},
{
"epoch": 0.55,
"learning_rate": 8.48e-06,
"loss": 0.2062,
"step": 425
},
{
"epoch": 0.58,
"learning_rate": 8.96e-06,
"loss": 0.1979,
"step": 450
},
{
"epoch": 0.61,
"learning_rate": 9.460000000000001e-06,
"loss": 0.1868,
"step": 475
},
{
"epoch": 0.65,
"learning_rate": 9.960000000000001e-06,
"loss": 0.2211,
"step": 500
},
{
"epoch": 0.68,
"learning_rate": 9.934285714285715e-06,
"loss": 0.2049,
"step": 525
},
{
"epoch": 0.71,
"learning_rate": 9.862857142857144e-06,
"loss": 0.1989,
"step": 550
},
{
"epoch": 0.74,
"learning_rate": 9.791428571428571e-06,
"loss": 0.2127,
"step": 575
},
{
"epoch": 0.78,
"learning_rate": 9.72e-06,
"loss": 0.2261,
"step": 600
},
{
"epoch": 0.81,
"learning_rate": 9.648571428571429e-06,
"loss": 0.2081,
"step": 625
},
{
"epoch": 0.84,
"learning_rate": 9.577142857142858e-06,
"loss": 0.2281,
"step": 650
},
{
"epoch": 0.87,
"learning_rate": 9.505714285714287e-06,
"loss": 0.204,
"step": 675
},
{
"epoch": 0.91,
"learning_rate": 9.434285714285714e-06,
"loss": 0.2053,
"step": 700
},
{
"epoch": 0.94,
"learning_rate": 9.362857142857143e-06,
"loss": 0.2208,
"step": 725
},
{
"epoch": 0.97,
"learning_rate": 9.291428571428572e-06,
"loss": 0.2076,
"step": 750
},
{
"epoch": 1.0,
"learning_rate": 9.220000000000002e-06,
"loss": 0.1941,
"step": 775
},
{
"epoch": 1.03,
"learning_rate": 9.148571428571429e-06,
"loss": 0.0949,
"step": 800
},
{
"epoch": 1.07,
"learning_rate": 9.077142857142858e-06,
"loss": 0.1011,
"step": 825
},
{
"epoch": 1.1,
"learning_rate": 9.005714285714287e-06,
"loss": 0.1078,
"step": 850
},
{
"epoch": 1.13,
"learning_rate": 8.934285714285716e-06,
"loss": 0.0956,
"step": 875
},
{
"epoch": 1.16,
"learning_rate": 8.862857142857143e-06,
"loss": 0.1093,
"step": 900
},
{
"epoch": 1.2,
"learning_rate": 8.791428571428572e-06,
"loss": 0.1133,
"step": 925
},
{
"epoch": 1.23,
"learning_rate": 8.720000000000001e-06,
"loss": 0.1128,
"step": 950
},
{
"epoch": 1.26,
"learning_rate": 8.64857142857143e-06,
"loss": 0.1068,
"step": 975
},
{
"epoch": 1.29,
"learning_rate": 8.577142857142858e-06,
"loss": 0.1015,
"step": 1000
},
{
"epoch": 1.29,
"eval_loss": 0.2880099415779114,
"eval_runtime": 2234.5268,
"eval_samples_per_second": 1.343,
"eval_steps_per_second": 0.168,
"eval_wer": 20.41339024938216,
"step": 1000
},
{
"epoch": 1.33,
"learning_rate": 8.505714285714287e-06,
"loss": 0.1154,
"step": 1025
},
{
"epoch": 1.36,
"learning_rate": 8.434285714285716e-06,
"loss": 0.1035,
"step": 1050
},
{
"epoch": 1.39,
"learning_rate": 8.362857142857143e-06,
"loss": 0.103,
"step": 1075
},
{
"epoch": 1.42,
"learning_rate": 8.291428571428572e-06,
"loss": 0.1089,
"step": 1100
},
{
"epoch": 1.46,
"learning_rate": 8.220000000000001e-06,
"loss": 0.0986,
"step": 1125
},
{
"epoch": 1.49,
"learning_rate": 8.148571428571428e-06,
"loss": 0.0999,
"step": 1150
},
{
"epoch": 1.52,
"learning_rate": 8.077142857142857e-06,
"loss": 0.1075,
"step": 1175
},
{
"epoch": 1.55,
"learning_rate": 8.005714285714286e-06,
"loss": 0.1155,
"step": 1200
},
{
"epoch": 1.58,
"learning_rate": 7.934285714285715e-06,
"loss": 0.0978,
"step": 1225
},
{
"epoch": 1.62,
"learning_rate": 7.862857142857143e-06,
"loss": 0.1035,
"step": 1250
},
{
"epoch": 1.65,
"learning_rate": 7.791428571428572e-06,
"loss": 0.1085,
"step": 1275
},
{
"epoch": 1.68,
"learning_rate": 7.72e-06,
"loss": 0.1008,
"step": 1300
},
{
"epoch": 1.71,
"learning_rate": 7.64857142857143e-06,
"loss": 0.1083,
"step": 1325
},
{
"epoch": 1.75,
"learning_rate": 7.577142857142857e-06,
"loss": 0.0918,
"step": 1350
},
{
"epoch": 1.78,
"learning_rate": 7.505714285714286e-06,
"loss": 0.1124,
"step": 1375
},
{
"epoch": 1.81,
"learning_rate": 7.434285714285715e-06,
"loss": 0.0997,
"step": 1400
},
{
"epoch": 1.84,
"learning_rate": 7.362857142857144e-06,
"loss": 0.0954,
"step": 1425
},
{
"epoch": 1.88,
"learning_rate": 7.291428571428571e-06,
"loss": 0.0946,
"step": 1450
},
{
"epoch": 1.91,
"learning_rate": 7.22e-06,
"loss": 0.1,
"step": 1475
},
{
"epoch": 1.94,
"learning_rate": 7.148571428571429e-06,
"loss": 0.0961,
"step": 1500
},
{
"epoch": 1.97,
"learning_rate": 7.077142857142858e-06,
"loss": 0.1074,
"step": 1525
},
{
"epoch": 2.01,
"learning_rate": 7.0057142857142865e-06,
"loss": 0.0821,
"step": 1550
},
{
"epoch": 2.04,
"learning_rate": 6.934285714285715e-06,
"loss": 0.0392,
"step": 1575
},
{
"epoch": 2.07,
"learning_rate": 6.862857142857144e-06,
"loss": 0.0371,
"step": 1600
},
{
"epoch": 2.1,
"learning_rate": 6.791428571428572e-06,
"loss": 0.0441,
"step": 1625
},
{
"epoch": 2.13,
"learning_rate": 6.720000000000001e-06,
"loss": 0.0416,
"step": 1650
},
{
"epoch": 2.17,
"learning_rate": 6.648571428571429e-06,
"loss": 0.0434,
"step": 1675
},
{
"epoch": 2.2,
"learning_rate": 6.577142857142857e-06,
"loss": 0.0457,
"step": 1700
},
{
"epoch": 2.23,
"learning_rate": 6.505714285714286e-06,
"loss": 0.0376,
"step": 1725
},
{
"epoch": 2.26,
"learning_rate": 6.434285714285715e-06,
"loss": 0.039,
"step": 1750
},
{
"epoch": 2.3,
"learning_rate": 6.3628571428571426e-06,
"loss": 0.0405,
"step": 1775
},
{
"epoch": 2.33,
"learning_rate": 6.2914285714285716e-06,
"loss": 0.0352,
"step": 1800
},
{
"epoch": 2.36,
"learning_rate": 6.220000000000001e-06,
"loss": 0.0397,
"step": 1825
},
{
"epoch": 2.39,
"learning_rate": 6.14857142857143e-06,
"loss": 0.0396,
"step": 1850
},
{
"epoch": 2.43,
"learning_rate": 6.077142857142858e-06,
"loss": 0.0388,
"step": 1875
},
{
"epoch": 2.46,
"learning_rate": 6.005714285714286e-06,
"loss": 0.038,
"step": 1900
},
{
"epoch": 2.49,
"learning_rate": 5.934285714285715e-06,
"loss": 0.0391,
"step": 1925
},
{
"epoch": 2.52,
"learning_rate": 5.862857142857143e-06,
"loss": 0.0358,
"step": 1950
},
{
"epoch": 2.55,
"learning_rate": 5.791428571428572e-06,
"loss": 0.0399,
"step": 1975
},
{
"epoch": 2.59,
"learning_rate": 5.72e-06,
"loss": 0.0387,
"step": 2000
},
{
"epoch": 2.59,
"eval_loss": 0.29592692852020264,
"eval_runtime": 2465.3193,
"eval_samples_per_second": 1.217,
"eval_steps_per_second": 0.152,
"eval_wer": 19.680970568411592,
"step": 2000
},
{
"epoch": 2.62,
"learning_rate": 5.6485714285714285e-06,
"loss": 0.0349,
"step": 2025
},
{
"epoch": 2.65,
"learning_rate": 5.5771428571428575e-06,
"loss": 0.0395,
"step": 2050
},
{
"epoch": 2.68,
"learning_rate": 5.5057142857142865e-06,
"loss": 0.0403,
"step": 2075
},
{
"epoch": 2.72,
"learning_rate": 5.4342857142857155e-06,
"loss": 0.043,
"step": 2100
},
{
"epoch": 2.75,
"learning_rate": 5.362857142857143e-06,
"loss": 0.0443,
"step": 2125
},
{
"epoch": 2.78,
"learning_rate": 5.291428571428572e-06,
"loss": 0.0421,
"step": 2150
},
{
"epoch": 2.81,
"learning_rate": 5.220000000000001e-06,
"loss": 0.0354,
"step": 2175
},
{
"epoch": 2.85,
"learning_rate": 5.14857142857143e-06,
"loss": 0.0393,
"step": 2200
},
{
"epoch": 2.88,
"learning_rate": 5.077142857142857e-06,
"loss": 0.0371,
"step": 2225
},
{
"epoch": 2.91,
"learning_rate": 5.005714285714286e-06,
"loss": 0.0354,
"step": 2250
},
{
"epoch": 2.94,
"learning_rate": 4.934285714285715e-06,
"loss": 0.043,
"step": 2275
},
{
"epoch": 2.98,
"learning_rate": 4.862857142857143e-06,
"loss": 0.0356,
"step": 2300
},
{
"epoch": 3.01,
"learning_rate": 4.7914285714285715e-06,
"loss": 0.038,
"step": 2325
},
{
"epoch": 3.04,
"learning_rate": 4.7200000000000005e-06,
"loss": 0.0162,
"step": 2350
},
{
"epoch": 3.07,
"learning_rate": 4.648571428571429e-06,
"loss": 0.0175,
"step": 2375
},
{
"epoch": 3.1,
"learning_rate": 4.577142857142858e-06,
"loss": 0.0143,
"step": 2400
},
{
"epoch": 3.14,
"learning_rate": 4.505714285714286e-06,
"loss": 0.0165,
"step": 2425
},
{
"epoch": 3.17,
"learning_rate": 4.434285714285715e-06,
"loss": 0.0176,
"step": 2450
},
{
"epoch": 3.2,
"learning_rate": 4.362857142857143e-06,
"loss": 0.0148,
"step": 2475
},
{
"epoch": 3.23,
"learning_rate": 4.291428571428572e-06,
"loss": 0.0157,
"step": 2500
},
{
"epoch": 3.27,
"learning_rate": 4.22e-06,
"loss": 0.0147,
"step": 2525
},
{
"epoch": 3.3,
"learning_rate": 4.151428571428572e-06,
"loss": 0.016,
"step": 2550
},
{
"epoch": 3.33,
"learning_rate": 4.08e-06,
"loss": 0.0153,
"step": 2575
},
{
"epoch": 3.36,
"learning_rate": 4.008571428571429e-06,
"loss": 0.0135,
"step": 2600
},
{
"epoch": 3.4,
"learning_rate": 3.937142857142858e-06,
"loss": 0.0194,
"step": 2625
},
{
"epoch": 3.43,
"learning_rate": 3.865714285714286e-06,
"loss": 0.0155,
"step": 2650
},
{
"epoch": 3.46,
"learning_rate": 3.7942857142857147e-06,
"loss": 0.0127,
"step": 2675
},
{
"epoch": 3.49,
"learning_rate": 3.722857142857143e-06,
"loss": 0.0148,
"step": 2700
},
{
"epoch": 3.53,
"learning_rate": 3.651428571428572e-06,
"loss": 0.0167,
"step": 2725
},
{
"epoch": 3.56,
"learning_rate": 3.58e-06,
"loss": 0.0181,
"step": 2750
},
{
"epoch": 3.59,
"learning_rate": 3.508571428571429e-06,
"loss": 0.0155,
"step": 2775
},
{
"epoch": 3.62,
"learning_rate": 3.437142857142857e-06,
"loss": 0.0157,
"step": 2800
},
{
"epoch": 3.65,
"learning_rate": 3.3657142857142862e-06,
"loss": 0.0142,
"step": 2825
},
{
"epoch": 3.69,
"learning_rate": 3.2942857142857144e-06,
"loss": 0.0156,
"step": 2850
},
{
"epoch": 3.72,
"learning_rate": 3.222857142857143e-06,
"loss": 0.0156,
"step": 2875
},
{
"epoch": 3.75,
"learning_rate": 3.151428571428572e-06,
"loss": 0.0189,
"step": 2900
},
{
"epoch": 3.78,
"learning_rate": 3.08e-06,
"loss": 0.0133,
"step": 2925
},
{
"epoch": 3.82,
"learning_rate": 3.008571428571429e-06,
"loss": 0.0151,
"step": 2950
},
{
"epoch": 3.85,
"learning_rate": 2.9371428571428573e-06,
"loss": 0.0176,
"step": 2975
},
{
"epoch": 3.88,
"learning_rate": 2.865714285714286e-06,
"loss": 0.0126,
"step": 3000
},
{
"epoch": 3.88,
"eval_loss": 0.3103422224521637,
"eval_runtime": 2600.9697,
"eval_samples_per_second": 1.153,
"eval_steps_per_second": 0.144,
"eval_wer": 19.299033924960682,
"step": 3000
}
],
"max_steps": 4000,
"num_train_epochs": 6,
"total_flos": 1.384517316575232e+19,
"trial_name": null,
"trial_params": null
}