sanchit-gandhi's picture
Add training and evaluation results
4909500
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 7.334963325183374,
"global_step": 3000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.06,
"learning_rate": 4.6000000000000004e-07,
"loss": 0.8917,
"step": 25
},
{
"epoch": 0.12,
"learning_rate": 9.400000000000001e-07,
"loss": 0.8196,
"step": 50
},
{
"epoch": 0.18,
"learning_rate": 1.44e-06,
"loss": 0.6476,
"step": 75
},
{
"epoch": 0.24,
"learning_rate": 1.94e-06,
"loss": 0.4701,
"step": 100
},
{
"epoch": 0.31,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.5104,
"step": 125
},
{
"epoch": 0.37,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.3547,
"step": 150
},
{
"epoch": 0.43,
"learning_rate": 3.44e-06,
"loss": 0.4427,
"step": 175
},
{
"epoch": 0.49,
"learning_rate": 3.94e-06,
"loss": 0.3209,
"step": 200
},
{
"epoch": 0.55,
"learning_rate": 4.440000000000001e-06,
"loss": 0.3846,
"step": 225
},
{
"epoch": 0.61,
"learning_rate": 4.94e-06,
"loss": 0.252,
"step": 250
},
{
"epoch": 0.67,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.3344,
"step": 275
},
{
"epoch": 0.73,
"learning_rate": 5.94e-06,
"loss": 0.2421,
"step": 300
},
{
"epoch": 0.79,
"learning_rate": 6.440000000000001e-06,
"loss": 0.3204,
"step": 325
},
{
"epoch": 0.86,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.2263,
"step": 350
},
{
"epoch": 0.92,
"learning_rate": 7.440000000000001e-06,
"loss": 0.2931,
"step": 375
},
{
"epoch": 0.98,
"learning_rate": 7.94e-06,
"loss": 0.1939,
"step": 400
},
{
"epoch": 1.04,
"learning_rate": 8.44e-06,
"loss": 0.2428,
"step": 425
},
{
"epoch": 1.1,
"learning_rate": 8.94e-06,
"loss": 0.1915,
"step": 450
},
{
"epoch": 1.16,
"learning_rate": 9.440000000000001e-06,
"loss": 0.1955,
"step": 475
},
{
"epoch": 1.22,
"learning_rate": 9.940000000000001e-06,
"loss": 0.1694,
"step": 500
},
{
"epoch": 1.28,
"learning_rate": 9.951111111111111e-06,
"loss": 0.1946,
"step": 525
},
{
"epoch": 1.34,
"learning_rate": 9.895555555555557e-06,
"loss": 0.1883,
"step": 550
},
{
"epoch": 1.41,
"learning_rate": 9.84e-06,
"loss": 0.1889,
"step": 575
},
{
"epoch": 1.47,
"learning_rate": 9.784444444444445e-06,
"loss": 0.1598,
"step": 600
},
{
"epoch": 1.53,
"learning_rate": 9.72888888888889e-06,
"loss": 0.1764,
"step": 625
},
{
"epoch": 1.59,
"learning_rate": 9.673333333333334e-06,
"loss": 0.1652,
"step": 650
},
{
"epoch": 1.65,
"learning_rate": 9.617777777777778e-06,
"loss": 0.1718,
"step": 675
},
{
"epoch": 1.71,
"learning_rate": 9.562222222222223e-06,
"loss": 0.157,
"step": 700
},
{
"epoch": 1.77,
"learning_rate": 9.506666666666667e-06,
"loss": 0.1685,
"step": 725
},
{
"epoch": 1.83,
"learning_rate": 9.451111111111112e-06,
"loss": 0.1597,
"step": 750
},
{
"epoch": 1.89,
"learning_rate": 9.395555555555556e-06,
"loss": 0.1641,
"step": 775
},
{
"epoch": 1.96,
"learning_rate": 9.340000000000002e-06,
"loss": 0.1446,
"step": 800
},
{
"epoch": 2.02,
"learning_rate": 9.284444444444444e-06,
"loss": 0.1322,
"step": 825
},
{
"epoch": 2.08,
"learning_rate": 9.22888888888889e-06,
"loss": 0.0919,
"step": 850
},
{
"epoch": 2.14,
"learning_rate": 9.173333333333334e-06,
"loss": 0.0799,
"step": 875
},
{
"epoch": 2.2,
"learning_rate": 9.117777777777778e-06,
"loss": 0.0925,
"step": 900
},
{
"epoch": 2.26,
"learning_rate": 9.062222222222224e-06,
"loss": 0.0725,
"step": 925
},
{
"epoch": 2.32,
"learning_rate": 9.006666666666666e-06,
"loss": 0.1034,
"step": 950
},
{
"epoch": 2.38,
"learning_rate": 8.951111111111112e-06,
"loss": 0.082,
"step": 975
},
{
"epoch": 2.44,
"learning_rate": 8.895555555555556e-06,
"loss": 0.1011,
"step": 1000
},
{
"epoch": 2.44,
"eval_loss": 0.30747494101524353,
"eval_runtime": 799.5663,
"eval_samples_per_second": 3.619,
"eval_steps_per_second": 0.226,
"eval_wer": 0.3463133835604842,
"step": 1000
},
{
"epoch": 2.51,
"learning_rate": 8.84e-06,
"loss": 0.0762,
"step": 1025
},
{
"epoch": 2.57,
"learning_rate": 8.784444444444446e-06,
"loss": 0.1048,
"step": 1050
},
{
"epoch": 2.63,
"learning_rate": 8.72888888888889e-06,
"loss": 0.0804,
"step": 1075
},
{
"epoch": 2.69,
"learning_rate": 8.673333333333334e-06,
"loss": 0.1036,
"step": 1100
},
{
"epoch": 2.75,
"learning_rate": 8.617777777777778e-06,
"loss": 0.0789,
"step": 1125
},
{
"epoch": 2.81,
"learning_rate": 8.562222222222224e-06,
"loss": 0.0976,
"step": 1150
},
{
"epoch": 2.87,
"learning_rate": 8.506666666666668e-06,
"loss": 0.0822,
"step": 1175
},
{
"epoch": 2.93,
"learning_rate": 8.451111111111112e-06,
"loss": 0.0981,
"step": 1200
},
{
"epoch": 3.0,
"learning_rate": 8.395555555555557e-06,
"loss": 0.0591,
"step": 1225
},
{
"epoch": 3.06,
"learning_rate": 8.34e-06,
"loss": 0.0594,
"step": 1250
},
{
"epoch": 3.12,
"learning_rate": 8.284444444444446e-06,
"loss": 0.0348,
"step": 1275
},
{
"epoch": 3.18,
"learning_rate": 8.22888888888889e-06,
"loss": 0.062,
"step": 1300
},
{
"epoch": 3.24,
"learning_rate": 8.173333333333334e-06,
"loss": 0.035,
"step": 1325
},
{
"epoch": 3.3,
"learning_rate": 8.11777777777778e-06,
"loss": 0.0612,
"step": 1350
},
{
"epoch": 3.36,
"learning_rate": 8.062222222222222e-06,
"loss": 0.0328,
"step": 1375
},
{
"epoch": 3.42,
"learning_rate": 8.006666666666667e-06,
"loss": 0.0643,
"step": 1400
},
{
"epoch": 3.48,
"learning_rate": 7.951111111111111e-06,
"loss": 0.0306,
"step": 1425
},
{
"epoch": 3.55,
"learning_rate": 7.895555555555557e-06,
"loss": 0.0601,
"step": 1450
},
{
"epoch": 3.61,
"learning_rate": 7.840000000000001e-06,
"loss": 0.0304,
"step": 1475
},
{
"epoch": 3.67,
"learning_rate": 7.784444444444445e-06,
"loss": 0.0582,
"step": 1500
},
{
"epoch": 3.73,
"learning_rate": 7.72888888888889e-06,
"loss": 0.0341,
"step": 1525
},
{
"epoch": 3.79,
"learning_rate": 7.673333333333333e-06,
"loss": 0.0625,
"step": 1550
},
{
"epoch": 3.85,
"learning_rate": 7.617777777777778e-06,
"loss": 0.0365,
"step": 1575
},
{
"epoch": 3.91,
"learning_rate": 7.562222222222223e-06,
"loss": 0.0622,
"step": 1600
},
{
"epoch": 3.97,
"learning_rate": 7.506666666666668e-06,
"loss": 0.032,
"step": 1625
},
{
"epoch": 4.03,
"learning_rate": 7.451111111111111e-06,
"loss": 0.041,
"step": 1650
},
{
"epoch": 4.1,
"learning_rate": 7.395555555555556e-06,
"loss": 0.0217,
"step": 1675
},
{
"epoch": 4.16,
"learning_rate": 7.340000000000001e-06,
"loss": 0.0267,
"step": 1700
},
{
"epoch": 4.22,
"learning_rate": 7.284444444444445e-06,
"loss": 0.0241,
"step": 1725
},
{
"epoch": 4.28,
"learning_rate": 7.22888888888889e-06,
"loss": 0.0292,
"step": 1750
},
{
"epoch": 4.34,
"learning_rate": 7.173333333333335e-06,
"loss": 0.0204,
"step": 1775
},
{
"epoch": 4.4,
"learning_rate": 7.117777777777778e-06,
"loss": 0.0234,
"step": 1800
},
{
"epoch": 4.46,
"learning_rate": 7.062222222222223e-06,
"loss": 0.0211,
"step": 1825
},
{
"epoch": 4.52,
"learning_rate": 7.006666666666667e-06,
"loss": 0.031,
"step": 1850
},
{
"epoch": 4.58,
"learning_rate": 6.951111111111112e-06,
"loss": 0.0196,
"step": 1875
},
{
"epoch": 4.65,
"learning_rate": 6.8955555555555565e-06,
"loss": 0.0306,
"step": 1900
},
{
"epoch": 4.71,
"learning_rate": 6.8400000000000014e-06,
"loss": 0.0212,
"step": 1925
},
{
"epoch": 4.77,
"learning_rate": 6.784444444444445e-06,
"loss": 0.0271,
"step": 1950
},
{
"epoch": 4.83,
"learning_rate": 6.7288888888888895e-06,
"loss": 0.0224,
"step": 1975
},
{
"epoch": 4.89,
"learning_rate": 6.6733333333333335e-06,
"loss": 0.0264,
"step": 2000
},
{
"epoch": 4.89,
"eval_loss": 0.355797678232193,
"eval_runtime": 796.3076,
"eval_samples_per_second": 3.634,
"eval_steps_per_second": 0.227,
"eval_wer": 0.3312875645475324,
"step": 2000
},
{
"epoch": 4.95,
"learning_rate": 6.617777777777778e-06,
"loss": 0.0229,
"step": 2025
},
{
"epoch": 5.01,
"learning_rate": 6.562222222222223e-06,
"loss": 0.0245,
"step": 2050
},
{
"epoch": 5.07,
"learning_rate": 6.5066666666666665e-06,
"loss": 0.0149,
"step": 2075
},
{
"epoch": 5.13,
"learning_rate": 6.451111111111111e-06,
"loss": 0.0114,
"step": 2100
},
{
"epoch": 5.2,
"learning_rate": 6.395555555555556e-06,
"loss": 0.0126,
"step": 2125
},
{
"epoch": 5.26,
"learning_rate": 6.34e-06,
"loss": 0.0093,
"step": 2150
},
{
"epoch": 5.32,
"learning_rate": 6.284444444444445e-06,
"loss": 0.014,
"step": 2175
},
{
"epoch": 5.38,
"learning_rate": 6.22888888888889e-06,
"loss": 0.0099,
"step": 2200
},
{
"epoch": 5.44,
"learning_rate": 6.173333333333333e-06,
"loss": 0.0121,
"step": 2225
},
{
"epoch": 5.5,
"learning_rate": 6.117777777777778e-06,
"loss": 0.008,
"step": 2250
},
{
"epoch": 5.56,
"learning_rate": 6.062222222222223e-06,
"loss": 0.0156,
"step": 2275
},
{
"epoch": 5.62,
"learning_rate": 6.006666666666667e-06,
"loss": 0.0132,
"step": 2300
},
{
"epoch": 5.68,
"learning_rate": 5.951111111111112e-06,
"loss": 0.0132,
"step": 2325
},
{
"epoch": 5.75,
"learning_rate": 5.895555555555557e-06,
"loss": 0.0093,
"step": 2350
},
{
"epoch": 5.81,
"learning_rate": 5.84e-06,
"loss": 0.0161,
"step": 2375
},
{
"epoch": 5.87,
"learning_rate": 5.784444444444445e-06,
"loss": 0.0114,
"step": 2400
},
{
"epoch": 5.93,
"learning_rate": 5.72888888888889e-06,
"loss": 0.016,
"step": 2425
},
{
"epoch": 5.99,
"learning_rate": 5.673333333333334e-06,
"loss": 0.0084,
"step": 2450
},
{
"epoch": 6.05,
"learning_rate": 5.617777777777779e-06,
"loss": 0.0114,
"step": 2475
},
{
"epoch": 6.11,
"learning_rate": 5.562222222222222e-06,
"loss": 0.0044,
"step": 2500
},
{
"epoch": 6.17,
"learning_rate": 5.506666666666667e-06,
"loss": 0.007,
"step": 2525
},
{
"epoch": 6.23,
"learning_rate": 5.451111111111112e-06,
"loss": 0.0035,
"step": 2550
},
{
"epoch": 6.3,
"learning_rate": 5.3955555555555565e-06,
"loss": 0.009,
"step": 2575
},
{
"epoch": 6.36,
"learning_rate": 5.3400000000000005e-06,
"loss": 0.0038,
"step": 2600
},
{
"epoch": 6.42,
"learning_rate": 5.2844444444444454e-06,
"loss": 0.008,
"step": 2625
},
{
"epoch": 6.48,
"learning_rate": 5.228888888888889e-06,
"loss": 0.0035,
"step": 2650
},
{
"epoch": 6.54,
"learning_rate": 5.1733333333333335e-06,
"loss": 0.0079,
"step": 2675
},
{
"epoch": 6.6,
"learning_rate": 5.117777777777778e-06,
"loss": 0.0041,
"step": 2700
},
{
"epoch": 6.66,
"learning_rate": 5.062222222222222e-06,
"loss": 0.0077,
"step": 2725
},
{
"epoch": 6.72,
"learning_rate": 5.006666666666667e-06,
"loss": 0.004,
"step": 2750
},
{
"epoch": 6.78,
"learning_rate": 4.951111111111111e-06,
"loss": 0.0091,
"step": 2775
},
{
"epoch": 6.85,
"learning_rate": 4.895555555555556e-06,
"loss": 0.0041,
"step": 2800
},
{
"epoch": 6.91,
"learning_rate": 4.84e-06,
"loss": 0.0072,
"step": 2825
},
{
"epoch": 6.97,
"learning_rate": 4.784444444444445e-06,
"loss": 0.0042,
"step": 2850
},
{
"epoch": 7.03,
"learning_rate": 4.728888888888889e-06,
"loss": 0.0053,
"step": 2875
},
{
"epoch": 7.09,
"learning_rate": 4.673333333333333e-06,
"loss": 0.0033,
"step": 2900
},
{
"epoch": 7.15,
"learning_rate": 4.617777777777778e-06,
"loss": 0.0022,
"step": 2925
},
{
"epoch": 7.21,
"learning_rate": 4.562222222222222e-06,
"loss": 0.003,
"step": 2950
},
{
"epoch": 7.27,
"learning_rate": 4.506666666666667e-06,
"loss": 0.003,
"step": 2975
},
{
"epoch": 7.33,
"learning_rate": 4.451111111111112e-06,
"loss": 0.0025,
"step": 3000
},
{
"epoch": 7.33,
"eval_loss": 0.42144426703453064,
"eval_runtime": 791.1382,
"eval_samples_per_second": 3.658,
"eval_steps_per_second": 0.229,
"eval_wer": 0.3259121307034623,
"step": 3000
}
],
"max_steps": 5000,
"num_train_epochs": 13,
"total_flos": 1.384401882415104e+19,
"trial_name": null,
"trial_params": null
}