fsicoli's picture
f7c7b967f2bccbc5b3ac50e22c6321a1541f3aa5257597a541ecc235d0ec2b01
b50e049 verified
raw
history blame
19.9 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.032258064516129,
"eval_steps": 1000,
"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": 1.1499999999999999e-08,
"loss": 0.7822,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 2.4e-08,
"loss": 1.3149,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 3.6499999999999996e-08,
"loss": 0.7809,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 4.9e-08,
"loss": 1.2916,
"step": 100
},
{
"epoch": 0.03,
"learning_rate": 6.15e-08,
"loss": 0.7645,
"step": 125
},
{
"epoch": 0.04,
"learning_rate": 7.399999999999999e-08,
"loss": 1.3379,
"step": 150
},
{
"epoch": 0.05,
"learning_rate": 8.649999999999999e-08,
"loss": 0.7722,
"step": 175
},
{
"epoch": 0.05,
"learning_rate": 9.9e-08,
"loss": 1.281,
"step": 200
},
{
"epoch": 0.06,
"learning_rate": 1.115e-07,
"loss": 0.6998,
"step": 225
},
{
"epoch": 0.06,
"learning_rate": 1.24e-07,
"loss": 1.1532,
"step": 250
},
{
"epoch": 0.07,
"learning_rate": 1.365e-07,
"loss": 0.583,
"step": 275
},
{
"epoch": 0.08,
"learning_rate": 1.49e-07,
"loss": 0.6693,
"step": 300
},
{
"epoch": 0.08,
"learning_rate": 1.615e-07,
"loss": 0.3395,
"step": 325
},
{
"epoch": 0.09,
"learning_rate": 1.7399999999999997e-07,
"loss": 0.3597,
"step": 350
},
{
"epoch": 0.1,
"learning_rate": 1.8649999999999998e-07,
"loss": 0.22,
"step": 375
},
{
"epoch": 0.1,
"learning_rate": 1.99e-07,
"loss": 0.3449,
"step": 400
},
{
"epoch": 0.11,
"learning_rate": 2.1149999999999998e-07,
"loss": 0.2123,
"step": 425
},
{
"epoch": 0.12,
"learning_rate": 2.24e-07,
"loss": 0.3362,
"step": 450
},
{
"epoch": 0.12,
"learning_rate": 2.3649999999999998e-07,
"loss": 0.175,
"step": 475
},
{
"epoch": 0.13,
"learning_rate": 2.4899999999999997e-07,
"loss": 0.3051,
"step": 500
},
{
"epoch": 0.14,
"learning_rate": 2.615e-07,
"loss": 0.176,
"step": 525
},
{
"epoch": 0.14,
"learning_rate": 2.74e-07,
"loss": 0.2517,
"step": 550
},
{
"epoch": 0.15,
"learning_rate": 2.865e-07,
"loss": 0.1863,
"step": 575
},
{
"epoch": 0.15,
"learning_rate": 2.9899999999999996e-07,
"loss": 0.2614,
"step": 600
},
{
"epoch": 0.16,
"learning_rate": 3.115e-07,
"loss": 0.1942,
"step": 625
},
{
"epoch": 0.17,
"learning_rate": 3.24e-07,
"loss": 0.2901,
"step": 650
},
{
"epoch": 0.17,
"learning_rate": 3.3650000000000003e-07,
"loss": 0.1587,
"step": 675
},
{
"epoch": 0.18,
"learning_rate": 3.4899999999999996e-07,
"loss": 0.2128,
"step": 700
},
{
"epoch": 0.19,
"learning_rate": 3.6149999999999995e-07,
"loss": 0.1647,
"step": 725
},
{
"epoch": 0.19,
"learning_rate": 3.74e-07,
"loss": 0.2473,
"step": 750
},
{
"epoch": 0.2,
"learning_rate": 3.8649999999999997e-07,
"loss": 0.1415,
"step": 775
},
{
"epoch": 0.21,
"learning_rate": 3.99e-07,
"loss": 0.2367,
"step": 800
},
{
"epoch": 0.21,
"learning_rate": 4.1149999999999995e-07,
"loss": 0.1498,
"step": 825
},
{
"epoch": 0.22,
"learning_rate": 4.24e-07,
"loss": 0.2161,
"step": 850
},
{
"epoch": 0.23,
"learning_rate": 4.3649999999999997e-07,
"loss": 0.144,
"step": 875
},
{
"epoch": 0.23,
"learning_rate": 4.49e-07,
"loss": 0.2316,
"step": 900
},
{
"epoch": 0.24,
"learning_rate": 4.615e-07,
"loss": 0.1449,
"step": 925
},
{
"epoch": 0.25,
"learning_rate": 4.7399999999999993e-07,
"loss": 0.1513,
"step": 950
},
{
"epoch": 0.25,
"learning_rate": 4.864999999999999e-07,
"loss": 0.1855,
"step": 975
},
{
"epoch": 0.26,
"learning_rate": 4.99e-07,
"loss": 0.199,
"step": 1000
},
{
"epoch": 0.26,
"eval_loss": 0.15628743171691895,
"eval_runtime": 184706.1006,
"eval_samples_per_second": 0.051,
"eval_steps_per_second": 0.006,
"eval_wer": 0.1124272786037492,
"step": 1000
},
{
"epoch": 0.26,
"learning_rate": 5.114999999999999e-07,
"loss": 0.1301,
"step": 1025
},
{
"epoch": 0.27,
"learning_rate": 5.24e-07,
"loss": 0.2193,
"step": 1050
},
{
"epoch": 0.28,
"learning_rate": 5.365e-07,
"loss": 0.1552,
"step": 1075
},
{
"epoch": 0.28,
"learning_rate": 5.490000000000001e-07,
"loss": 0.1865,
"step": 1100
},
{
"epoch": 0.29,
"learning_rate": 5.614999999999999e-07,
"loss": 0.1618,
"step": 1125
},
{
"epoch": 0.3,
"learning_rate": 5.739999999999999e-07,
"loss": 0.2259,
"step": 1150
},
{
"epoch": 0.3,
"learning_rate": 5.865e-07,
"loss": 0.1418,
"step": 1175
},
{
"epoch": 0.31,
"learning_rate": 5.989999999999999e-07,
"loss": 0.1916,
"step": 1200
},
{
"epoch": 0.32,
"learning_rate": 6.115e-07,
"loss": 0.1295,
"step": 1225
},
{
"epoch": 0.32,
"learning_rate": 6.24e-07,
"loss": 0.2112,
"step": 1250
},
{
"epoch": 0.33,
"learning_rate": 6.365e-07,
"loss": 0.1231,
"step": 1275
},
{
"epoch": 0.34,
"learning_rate": 6.49e-07,
"loss": 0.1914,
"step": 1300
},
{
"epoch": 0.34,
"learning_rate": 6.614999999999999e-07,
"loss": 0.1485,
"step": 1325
},
{
"epoch": 0.35,
"learning_rate": 6.74e-07,
"loss": 0.1958,
"step": 1350
},
{
"epoch": 0.35,
"learning_rate": 6.865e-07,
"loss": 0.1452,
"step": 1375
},
{
"epoch": 0.36,
"learning_rate": 6.989999999999999e-07,
"loss": 0.1624,
"step": 1400
},
{
"epoch": 0.37,
"learning_rate": 7.115e-07,
"loss": 0.1518,
"step": 1425
},
{
"epoch": 0.37,
"learning_rate": 7.24e-07,
"loss": 0.1935,
"step": 1450
},
{
"epoch": 0.38,
"learning_rate": 7.365e-07,
"loss": 0.1138,
"step": 1475
},
{
"epoch": 0.39,
"learning_rate": 7.489999999999999e-07,
"loss": 0.16,
"step": 1500
},
{
"epoch": 0.39,
"learning_rate": 7.614999999999999e-07,
"loss": 0.1279,
"step": 1525
},
{
"epoch": 0.4,
"learning_rate": 7.74e-07,
"loss": 0.1862,
"step": 1550
},
{
"epoch": 0.41,
"learning_rate": 7.864999999999999e-07,
"loss": 0.1537,
"step": 1575
},
{
"epoch": 0.41,
"learning_rate": 7.99e-07,
"loss": 0.1915,
"step": 1600
},
{
"epoch": 0.42,
"learning_rate": 8.115e-07,
"loss": 0.1303,
"step": 1625
},
{
"epoch": 0.43,
"learning_rate": 8.24e-07,
"loss": 0.1967,
"step": 1650
},
{
"epoch": 0.43,
"learning_rate": 8.365e-07,
"loss": 0.1418,
"step": 1675
},
{
"epoch": 0.44,
"learning_rate": 8.489999999999999e-07,
"loss": 0.194,
"step": 1700
},
{
"epoch": 0.45,
"learning_rate": 8.615e-07,
"loss": 0.1377,
"step": 1725
},
{
"epoch": 0.45,
"learning_rate": 8.739999999999999e-07,
"loss": 0.1534,
"step": 1750
},
{
"epoch": 0.46,
"learning_rate": 8.864999999999999e-07,
"loss": 0.1416,
"step": 1775
},
{
"epoch": 0.46,
"learning_rate": 8.99e-07,
"loss": 0.1869,
"step": 1800
},
{
"epoch": 0.47,
"learning_rate": 9.115e-07,
"loss": 0.1507,
"step": 1825
},
{
"epoch": 0.48,
"learning_rate": 9.24e-07,
"loss": 0.1571,
"step": 1850
},
{
"epoch": 0.48,
"learning_rate": 9.365e-07,
"loss": 0.1348,
"step": 1875
},
{
"epoch": 0.49,
"learning_rate": 9.489999999999999e-07,
"loss": 0.1981,
"step": 1900
},
{
"epoch": 0.5,
"learning_rate": 9.615e-07,
"loss": 0.1324,
"step": 1925
},
{
"epoch": 0.5,
"learning_rate": 9.74e-07,
"loss": 0.1712,
"step": 1950
},
{
"epoch": 0.51,
"learning_rate": 9.865e-07,
"loss": 0.1169,
"step": 1975
},
{
"epoch": 0.52,
"learning_rate": 9.989999999999999e-07,
"loss": 0.1654,
"step": 2000
},
{
"epoch": 0.52,
"eval_loss": 0.1500108540058136,
"eval_runtime": 188234.7285,
"eval_samples_per_second": 0.05,
"eval_steps_per_second": 0.006,
"eval_wer": 0.10521978021978022,
"step": 2000
},
{
"epoch": 0.52,
"learning_rate": 9.923333333333332e-07,
"loss": 0.1202,
"step": 2025
},
{
"epoch": 0.53,
"learning_rate": 9.84e-07,
"loss": 0.186,
"step": 2050
},
{
"epoch": 0.54,
"learning_rate": 9.756666666666666e-07,
"loss": 0.1106,
"step": 2075
},
{
"epoch": 0.54,
"learning_rate": 9.673333333333332e-07,
"loss": 0.1926,
"step": 2100
},
{
"epoch": 0.55,
"learning_rate": 9.589999999999998e-07,
"loss": 0.1178,
"step": 2125
},
{
"epoch": 0.55,
"learning_rate": 9.506666666666667e-07,
"loss": 0.1744,
"step": 2150
},
{
"epoch": 0.56,
"learning_rate": 9.423333333333333e-07,
"loss": 0.1263,
"step": 2175
},
{
"epoch": 0.57,
"learning_rate": 9.34e-07,
"loss": 0.1818,
"step": 2200
},
{
"epoch": 0.57,
"learning_rate": 9.256666666666666e-07,
"loss": 0.1466,
"step": 2225
},
{
"epoch": 0.58,
"learning_rate": 9.173333333333333e-07,
"loss": 0.1357,
"step": 2250
},
{
"epoch": 0.59,
"learning_rate": 9.09e-07,
"loss": 0.148,
"step": 2275
},
{
"epoch": 0.59,
"learning_rate": 9.006666666666666e-07,
"loss": 0.1684,
"step": 2300
},
{
"epoch": 0.6,
"learning_rate": 8.923333333333333e-07,
"loss": 0.1368,
"step": 2325
},
{
"epoch": 0.61,
"learning_rate": 8.839999999999999e-07,
"loss": 0.188,
"step": 2350
},
{
"epoch": 0.61,
"learning_rate": 8.756666666666666e-07,
"loss": 0.1299,
"step": 2375
},
{
"epoch": 0.62,
"learning_rate": 8.673333333333332e-07,
"loss": 0.1461,
"step": 2400
},
{
"epoch": 0.63,
"learning_rate": 8.59e-07,
"loss": 0.1569,
"step": 2425
},
{
"epoch": 0.63,
"learning_rate": 8.506666666666667e-07,
"loss": 0.1527,
"step": 2450
},
{
"epoch": 0.64,
"learning_rate": 8.423333333333334e-07,
"loss": 0.1041,
"step": 2475
},
{
"epoch": 0.65,
"learning_rate": 8.34e-07,
"loss": 0.157,
"step": 2500
},
{
"epoch": 0.65,
"learning_rate": 8.256666666666666e-07,
"loss": 0.1364,
"step": 2525
},
{
"epoch": 0.66,
"learning_rate": 8.173333333333333e-07,
"loss": 0.1685,
"step": 2550
},
{
"epoch": 0.66,
"learning_rate": 8.09e-07,
"loss": 0.1344,
"step": 2575
},
{
"epoch": 0.67,
"learning_rate": 8.006666666666666e-07,
"loss": 0.1589,
"step": 2600
},
{
"epoch": 0.68,
"learning_rate": 7.923333333333333e-07,
"loss": 0.1334,
"step": 2625
},
{
"epoch": 0.68,
"learning_rate": 7.84e-07,
"loss": 0.1297,
"step": 2650
},
{
"epoch": 0.69,
"learning_rate": 7.756666666666665e-07,
"loss": 0.1326,
"step": 2675
},
{
"epoch": 0.7,
"learning_rate": 7.673333333333332e-07,
"loss": 0.179,
"step": 2700
},
{
"epoch": 0.7,
"learning_rate": 7.59e-07,
"loss": 0.1242,
"step": 2725
},
{
"epoch": 0.71,
"learning_rate": 7.506666666666667e-07,
"loss": 0.1714,
"step": 2750
},
{
"epoch": 0.72,
"learning_rate": 7.423333333333333e-07,
"loss": 0.1503,
"step": 2775
},
{
"epoch": 0.72,
"learning_rate": 7.34e-07,
"loss": 0.131,
"step": 2800
},
{
"epoch": 0.73,
"learning_rate": 7.256666666666667e-07,
"loss": 0.118,
"step": 2825
},
{
"epoch": 0.74,
"learning_rate": 7.173333333333333e-07,
"loss": 0.1729,
"step": 2850
},
{
"epoch": 0.74,
"learning_rate": 7.089999999999999e-07,
"loss": 0.1193,
"step": 2875
},
{
"epoch": 0.75,
"learning_rate": 7.006666666666666e-07,
"loss": 0.1723,
"step": 2900
},
{
"epoch": 0.75,
"learning_rate": 6.923333333333333e-07,
"loss": 0.1393,
"step": 2925
},
{
"epoch": 0.76,
"learning_rate": 6.84e-07,
"loss": 0.1259,
"step": 2950
},
{
"epoch": 0.77,
"learning_rate": 6.756666666666666e-07,
"loss": 0.1228,
"step": 2975
},
{
"epoch": 0.77,
"learning_rate": 6.673333333333334e-07,
"loss": 0.1794,
"step": 3000
},
{
"epoch": 0.77,
"eval_loss": 0.13788852095603943,
"eval_runtime": 198193.9708,
"eval_samples_per_second": 0.047,
"eval_steps_per_second": 0.006,
"eval_wer": 0.09970911441499677,
"step": 3000
},
{
"epoch": 0.78,
"learning_rate": 6.59e-07,
"loss": 0.1334,
"step": 3025
},
{
"epoch": 0.79,
"learning_rate": 6.506666666666666e-07,
"loss": 0.1584,
"step": 3050
},
{
"epoch": 0.79,
"learning_rate": 6.423333333333333e-07,
"loss": 0.1331,
"step": 3075
},
{
"epoch": 0.8,
"learning_rate": 6.346666666666666e-07,
"loss": 0.1226,
"step": 3100
},
{
"epoch": 0.81,
"learning_rate": 6.263333333333332e-07,
"loss": 0.135,
"step": 3125
},
{
"epoch": 0.81,
"learning_rate": 6.18e-07,
"loss": 0.1538,
"step": 3150
},
{
"epoch": 0.82,
"learning_rate": 6.096666666666667e-07,
"loss": 0.1405,
"step": 3175
},
{
"epoch": 0.83,
"learning_rate": 6.013333333333334e-07,
"loss": 0.1534,
"step": 3200
},
{
"epoch": 0.83,
"learning_rate": 5.93e-07,
"loss": 0.1628,
"step": 3225
},
{
"epoch": 0.84,
"learning_rate": 5.846666666666667e-07,
"loss": 0.2002,
"step": 3250
},
{
"epoch": 0.85,
"learning_rate": 5.763333333333333e-07,
"loss": 0.1155,
"step": 3275
},
{
"epoch": 0.85,
"learning_rate": 5.679999999999999e-07,
"loss": 0.179,
"step": 3300
},
{
"epoch": 0.86,
"learning_rate": 5.596666666666666e-07,
"loss": 0.1471,
"step": 3325
},
{
"epoch": 0.86,
"learning_rate": 5.513333333333333e-07,
"loss": 0.1386,
"step": 3350
},
{
"epoch": 0.87,
"learning_rate": 5.43e-07,
"loss": 0.1185,
"step": 3375
},
{
"epoch": 0.88,
"learning_rate": 5.346666666666666e-07,
"loss": 0.1418,
"step": 3400
},
{
"epoch": 0.88,
"learning_rate": 5.263333333333333e-07,
"loss": 0.1015,
"step": 3425
},
{
"epoch": 0.89,
"learning_rate": 5.18e-07,
"loss": 0.1525,
"step": 3450
},
{
"epoch": 0.9,
"learning_rate": 5.096666666666667e-07,
"loss": 0.1212,
"step": 3475
},
{
"epoch": 0.9,
"learning_rate": 5.013333333333333e-07,
"loss": 0.1623,
"step": 3500
},
{
"epoch": 0.91,
"learning_rate": 4.93e-07,
"loss": 0.1178,
"step": 3525
},
{
"epoch": 0.92,
"learning_rate": 4.846666666666667e-07,
"loss": 0.1618,
"step": 3550
},
{
"epoch": 0.92,
"learning_rate": 4.763333333333333e-07,
"loss": 0.1187,
"step": 3575
},
{
"epoch": 0.93,
"learning_rate": 4.68e-07,
"loss": 0.1382,
"step": 3600
},
{
"epoch": 0.94,
"learning_rate": 4.5966666666666667e-07,
"loss": 0.1211,
"step": 3625
},
{
"epoch": 0.94,
"learning_rate": 4.5133333333333327e-07,
"loss": 0.1582,
"step": 3650
},
{
"epoch": 0.95,
"learning_rate": 4.43e-07,
"loss": 0.1201,
"step": 3675
},
{
"epoch": 0.95,
"learning_rate": 4.3466666666666664e-07,
"loss": 0.1571,
"step": 3700
},
{
"epoch": 0.96,
"learning_rate": 4.263333333333333e-07,
"loss": 0.1247,
"step": 3725
},
{
"epoch": 0.97,
"learning_rate": 4.1799999999999996e-07,
"loss": 0.1648,
"step": 3750
},
{
"epoch": 0.97,
"learning_rate": 4.0966666666666667e-07,
"loss": 0.1313,
"step": 3775
},
{
"epoch": 0.98,
"learning_rate": 4.0133333333333333e-07,
"loss": 0.1528,
"step": 3800
},
{
"epoch": 0.99,
"learning_rate": 3.93e-07,
"loss": 0.1252,
"step": 3825
},
{
"epoch": 0.99,
"learning_rate": 3.8466666666666664e-07,
"loss": 0.1674,
"step": 3850
},
{
"epoch": 1.0,
"learning_rate": 3.7633333333333335e-07,
"loss": 0.1192,
"step": 3875
},
{
"epoch": 1.01,
"learning_rate": 3.6799999999999996e-07,
"loss": 0.1054,
"step": 3900
},
{
"epoch": 1.01,
"learning_rate": 3.5966666666666667e-07,
"loss": 0.1353,
"step": 3925
},
{
"epoch": 1.02,
"learning_rate": 3.5133333333333333e-07,
"loss": 0.1004,
"step": 3950
},
{
"epoch": 1.03,
"learning_rate": 3.43e-07,
"loss": 0.1382,
"step": 3975
},
{
"epoch": 1.03,
"learning_rate": 3.3466666666666665e-07,
"loss": 0.0821,
"step": 4000
},
{
"epoch": 1.03,
"eval_loss": 0.13208560645580292,
"eval_runtime": 194476.4747,
"eval_samples_per_second": 0.048,
"eval_steps_per_second": 0.006,
"eval_wer": 1.000743374272786,
"step": 4000
}
],
"logging_steps": 25,
"max_steps": 5000,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 1000,
"total_flos": 1.0871315081330688e+20,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}