2789 lines
47 KiB
JSON
2789 lines
47 KiB
JSON
{
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 2.9890820865345735,
|
|
"global_step": 462,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.01,
|
|
"learning_rate": 0,
|
|
"loss": 10.8463,
|
|
"step": 1
|
|
},
|
|
{
|
|
"epoch": 0.01,
|
|
"learning_rate": 0,
|
|
"loss": 11.3118,
|
|
"step": 2
|
|
},
|
|
{
|
|
"epoch": 0.02,
|
|
"learning_rate": 0,
|
|
"loss": 11.0038,
|
|
"step": 3
|
|
},
|
|
{
|
|
"epoch": 0.03,
|
|
"learning_rate": 0,
|
|
"loss": 11.1465,
|
|
"step": 4
|
|
},
|
|
{
|
|
"epoch": 0.03,
|
|
"learning_rate": 0,
|
|
"loss": 11.3677,
|
|
"step": 5
|
|
},
|
|
{
|
|
"epoch": 0.04,
|
|
"learning_rate": 0,
|
|
"loss": 11.3668,
|
|
"step": 6
|
|
},
|
|
{
|
|
"epoch": 0.05,
|
|
"learning_rate": 0.0,
|
|
"loss": 10.8335,
|
|
"step": 7
|
|
},
|
|
{
|
|
"epoch": 0.05,
|
|
"learning_rate": 0.0,
|
|
"loss": 11.1042,
|
|
"step": 8
|
|
},
|
|
{
|
|
"epoch": 0.06,
|
|
"learning_rate": 2e-05,
|
|
"loss": 11.0963,
|
|
"step": 9
|
|
},
|
|
{
|
|
"epoch": 0.06,
|
|
"learning_rate": 2e-05,
|
|
"loss": 11.3511,
|
|
"step": 10
|
|
},
|
|
{
|
|
"epoch": 0.07,
|
|
"learning_rate": 2e-05,
|
|
"loss": 9.441,
|
|
"step": 11
|
|
},
|
|
{
|
|
"epoch": 0.08,
|
|
"learning_rate": 2e-05,
|
|
"loss": 9.1196,
|
|
"step": 12
|
|
},
|
|
{
|
|
"epoch": 0.08,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.403,
|
|
"step": 13
|
|
},
|
|
{
|
|
"epoch": 0.09,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.6426,
|
|
"step": 14
|
|
},
|
|
{
|
|
"epoch": 0.1,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.7776,
|
|
"step": 15
|
|
},
|
|
{
|
|
"epoch": 0.1,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.9274,
|
|
"step": 16
|
|
},
|
|
{
|
|
"epoch": 0.11,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.2598,
|
|
"step": 17
|
|
},
|
|
{
|
|
"epoch": 0.12,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.2732,
|
|
"step": 18
|
|
},
|
|
{
|
|
"epoch": 0.12,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.2438,
|
|
"step": 19
|
|
},
|
|
{
|
|
"epoch": 0.13,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.0727,
|
|
"step": 20
|
|
},
|
|
{
|
|
"epoch": 0.14,
|
|
"learning_rate": 2e-05,
|
|
"loss": 8.1909,
|
|
"step": 21
|
|
},
|
|
{
|
|
"epoch": 0.14,
|
|
"learning_rate": 2e-05,
|
|
"loss": 7.7707,
|
|
"step": 22
|
|
},
|
|
{
|
|
"epoch": 0.15,
|
|
"learning_rate": 2e-05,
|
|
"loss": 7.4861,
|
|
"step": 23
|
|
},
|
|
{
|
|
"epoch": 0.16,
|
|
"learning_rate": 2e-05,
|
|
"loss": 7.4279,
|
|
"step": 24
|
|
},
|
|
{
|
|
"epoch": 0.16,
|
|
"learning_rate": 2e-05,
|
|
"loss": 7.3696,
|
|
"step": 25
|
|
},
|
|
{
|
|
"epoch": 0.17,
|
|
"learning_rate": 2e-05,
|
|
"loss": 7.243,
|
|
"step": 26
|
|
},
|
|
{
|
|
"epoch": 0.17,
|
|
"learning_rate": 2e-05,
|
|
"loss": 6.9497,
|
|
"step": 27
|
|
},
|
|
{
|
|
"epoch": 0.18,
|
|
"learning_rate": 2e-05,
|
|
"loss": 6.9924,
|
|
"step": 28
|
|
},
|
|
{
|
|
"epoch": 0.19,
|
|
"learning_rate": 2e-05,
|
|
"loss": 6.5594,
|
|
"step": 29
|
|
},
|
|
{
|
|
"epoch": 0.19,
|
|
"learning_rate": 2e-05,
|
|
"loss": 6.7199,
|
|
"step": 30
|
|
},
|
|
{
|
|
"epoch": 0.2,
|
|
"learning_rate": 2e-05,
|
|
"loss": 6.2285,
|
|
"step": 31
|
|
},
|
|
{
|
|
"epoch": 0.21,
|
|
"learning_rate": 2e-05,
|
|
"loss": 6.0328,
|
|
"step": 32
|
|
},
|
|
{
|
|
"epoch": 0.21,
|
|
"learning_rate": 2e-05,
|
|
"loss": 5.821,
|
|
"step": 33
|
|
},
|
|
{
|
|
"epoch": 0.22,
|
|
"learning_rate": 2e-05,
|
|
"loss": 5.7585,
|
|
"step": 34
|
|
},
|
|
{
|
|
"epoch": 0.23,
|
|
"learning_rate": 2e-05,
|
|
"loss": 5.6209,
|
|
"step": 35
|
|
},
|
|
{
|
|
"epoch": 0.23,
|
|
"learning_rate": 2e-05,
|
|
"loss": 5.3515,
|
|
"step": 36
|
|
},
|
|
{
|
|
"epoch": 0.24,
|
|
"learning_rate": 2e-05,
|
|
"loss": 5.2036,
|
|
"step": 37
|
|
},
|
|
{
|
|
"epoch": 0.25,
|
|
"learning_rate": 2e-05,
|
|
"loss": 5.0071,
|
|
"step": 38
|
|
},
|
|
{
|
|
"epoch": 0.25,
|
|
"learning_rate": 2e-05,
|
|
"loss": 4.8688,
|
|
"step": 39
|
|
},
|
|
{
|
|
"epoch": 0.26,
|
|
"learning_rate": 2e-05,
|
|
"loss": 4.6462,
|
|
"step": 40
|
|
},
|
|
{
|
|
"epoch": 0.27,
|
|
"learning_rate": 2e-05,
|
|
"loss": 4.536,
|
|
"step": 41
|
|
},
|
|
{
|
|
"epoch": 0.27,
|
|
"learning_rate": 2e-05,
|
|
"loss": 4.0728,
|
|
"step": 42
|
|
},
|
|
{
|
|
"epoch": 0.28,
|
|
"learning_rate": 2e-05,
|
|
"loss": 3.8805,
|
|
"step": 43
|
|
},
|
|
{
|
|
"epoch": 0.28,
|
|
"learning_rate": 2e-05,
|
|
"loss": 3.9002,
|
|
"step": 44
|
|
},
|
|
{
|
|
"epoch": 0.29,
|
|
"learning_rate": 2e-05,
|
|
"loss": 3.3943,
|
|
"step": 45
|
|
},
|
|
{
|
|
"epoch": 0.3,
|
|
"learning_rate": 2e-05,
|
|
"loss": 3.3925,
|
|
"step": 46
|
|
},
|
|
{
|
|
"epoch": 0.3,
|
|
"learning_rate": 2e-05,
|
|
"loss": 3.0984,
|
|
"step": 47
|
|
},
|
|
{
|
|
"epoch": 0.31,
|
|
"learning_rate": 2e-05,
|
|
"loss": 2.8044,
|
|
"step": 48
|
|
},
|
|
{
|
|
"epoch": 0.32,
|
|
"learning_rate": 2e-05,
|
|
"loss": 2.6679,
|
|
"step": 49
|
|
},
|
|
{
|
|
"epoch": 0.32,
|
|
"learning_rate": 2e-05,
|
|
"loss": 2.4378,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.33,
|
|
"learning_rate": 2e-05,
|
|
"loss": 2.1403,
|
|
"step": 51
|
|
},
|
|
{
|
|
"epoch": 0.34,
|
|
"learning_rate": 2e-05,
|
|
"loss": 2.0139,
|
|
"step": 52
|
|
},
|
|
{
|
|
"epoch": 0.34,
|
|
"learning_rate": 2e-05,
|
|
"loss": 1.8027,
|
|
"step": 53
|
|
},
|
|
{
|
|
"epoch": 0.35,
|
|
"learning_rate": 2e-05,
|
|
"loss": 1.5626,
|
|
"step": 54
|
|
},
|
|
{
|
|
"epoch": 0.36,
|
|
"learning_rate": 2e-05,
|
|
"loss": 1.3481,
|
|
"step": 55
|
|
},
|
|
{
|
|
"epoch": 0.36,
|
|
"learning_rate": 2e-05,
|
|
"loss": 1.17,
|
|
"step": 56
|
|
},
|
|
{
|
|
"epoch": 0.37,
|
|
"learning_rate": 2e-05,
|
|
"loss": 1.0311,
|
|
"step": 57
|
|
},
|
|
{
|
|
"epoch": 0.38,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.8989,
|
|
"step": 58
|
|
},
|
|
{
|
|
"epoch": 0.38,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.7375,
|
|
"step": 59
|
|
},
|
|
{
|
|
"epoch": 0.39,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.6133,
|
|
"step": 60
|
|
},
|
|
{
|
|
"epoch": 0.39,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.4984,
|
|
"step": 61
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.4197,
|
|
"step": 62
|
|
},
|
|
{
|
|
"epoch": 0.41,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.349,
|
|
"step": 63
|
|
},
|
|
{
|
|
"epoch": 0.41,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.2899,
|
|
"step": 64
|
|
},
|
|
{
|
|
"epoch": 0.42,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.2383,
|
|
"step": 65
|
|
},
|
|
{
|
|
"epoch": 0.43,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.2003,
|
|
"step": 66
|
|
},
|
|
{
|
|
"epoch": 0.43,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.1653,
|
|
"step": 67
|
|
},
|
|
{
|
|
"epoch": 0.44,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.1448,
|
|
"step": 68
|
|
},
|
|
{
|
|
"epoch": 0.45,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.118,
|
|
"step": 69
|
|
},
|
|
{
|
|
"epoch": 0.45,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.1192,
|
|
"step": 70
|
|
},
|
|
{
|
|
"epoch": 0.46,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.1045,
|
|
"step": 71
|
|
},
|
|
{
|
|
"epoch": 0.47,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.1026,
|
|
"step": 72
|
|
},
|
|
{
|
|
"epoch": 0.47,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0894,
|
|
"step": 73
|
|
},
|
|
{
|
|
"epoch": 0.48,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0847,
|
|
"step": 74
|
|
},
|
|
{
|
|
"epoch": 0.49,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0838,
|
|
"step": 75
|
|
},
|
|
{
|
|
"epoch": 0.49,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0812,
|
|
"step": 76
|
|
},
|
|
{
|
|
"epoch": 0.5,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0852,
|
|
"step": 77
|
|
},
|
|
{
|
|
"epoch": 0.5,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0834,
|
|
"step": 78
|
|
},
|
|
{
|
|
"epoch": 0.51,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0805,
|
|
"step": 79
|
|
},
|
|
{
|
|
"epoch": 0.52,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0743,
|
|
"step": 80
|
|
},
|
|
{
|
|
"epoch": 0.52,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0747,
|
|
"step": 81
|
|
},
|
|
{
|
|
"epoch": 0.53,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0738,
|
|
"step": 82
|
|
},
|
|
{
|
|
"epoch": 0.54,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0845,
|
|
"step": 83
|
|
},
|
|
{
|
|
"epoch": 0.54,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0687,
|
|
"step": 84
|
|
},
|
|
{
|
|
"epoch": 0.55,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0515,
|
|
"step": 85
|
|
},
|
|
{
|
|
"epoch": 0.56,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0695,
|
|
"step": 86
|
|
},
|
|
{
|
|
"epoch": 0.56,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0615,
|
|
"step": 87
|
|
},
|
|
{
|
|
"epoch": 0.57,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0677,
|
|
"step": 88
|
|
},
|
|
{
|
|
"epoch": 0.58,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0692,
|
|
"step": 89
|
|
},
|
|
{
|
|
"epoch": 0.58,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0564,
|
|
"step": 90
|
|
},
|
|
{
|
|
"epoch": 0.59,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.058,
|
|
"step": 91
|
|
},
|
|
{
|
|
"epoch": 0.6,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0637,
|
|
"step": 92
|
|
},
|
|
{
|
|
"epoch": 0.6,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0711,
|
|
"step": 93
|
|
},
|
|
{
|
|
"epoch": 0.61,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0525,
|
|
"step": 94
|
|
},
|
|
{
|
|
"epoch": 0.61,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0518,
|
|
"step": 95
|
|
},
|
|
{
|
|
"epoch": 0.62,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0603,
|
|
"step": 96
|
|
},
|
|
{
|
|
"epoch": 0.63,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0509,
|
|
"step": 97
|
|
},
|
|
{
|
|
"epoch": 0.63,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0599,
|
|
"step": 98
|
|
},
|
|
{
|
|
"epoch": 0.64,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0639,
|
|
"step": 99
|
|
},
|
|
{
|
|
"epoch": 0.65,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0611,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.65,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0462,
|
|
"step": 101
|
|
},
|
|
{
|
|
"epoch": 0.66,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.051,
|
|
"step": 102
|
|
},
|
|
{
|
|
"epoch": 0.67,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0621,
|
|
"step": 103
|
|
},
|
|
{
|
|
"epoch": 0.67,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0597,
|
|
"step": 104
|
|
},
|
|
{
|
|
"epoch": 0.68,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0574,
|
|
"step": 105
|
|
},
|
|
{
|
|
"epoch": 0.69,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0573,
|
|
"step": 106
|
|
},
|
|
{
|
|
"epoch": 0.69,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0575,
|
|
"step": 107
|
|
},
|
|
{
|
|
"epoch": 0.7,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0563,
|
|
"step": 108
|
|
},
|
|
{
|
|
"epoch": 0.71,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0454,
|
|
"step": 109
|
|
},
|
|
{
|
|
"epoch": 0.71,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0442,
|
|
"step": 110
|
|
},
|
|
{
|
|
"epoch": 0.72,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0602,
|
|
"step": 111
|
|
},
|
|
{
|
|
"epoch": 0.72,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0578,
|
|
"step": 112
|
|
},
|
|
{
|
|
"epoch": 0.73,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0524,
|
|
"step": 113
|
|
},
|
|
{
|
|
"epoch": 0.74,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0567,
|
|
"step": 114
|
|
},
|
|
{
|
|
"epoch": 0.74,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0476,
|
|
"step": 115
|
|
},
|
|
{
|
|
"epoch": 0.75,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0576,
|
|
"step": 116
|
|
},
|
|
{
|
|
"epoch": 0.76,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0481,
|
|
"step": 117
|
|
},
|
|
{
|
|
"epoch": 0.76,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0496,
|
|
"step": 118
|
|
},
|
|
{
|
|
"epoch": 0.77,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0485,
|
|
"step": 119
|
|
},
|
|
{
|
|
"epoch": 0.78,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0538,
|
|
"step": 120
|
|
},
|
|
{
|
|
"epoch": 0.78,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0508,
|
|
"step": 121
|
|
},
|
|
{
|
|
"epoch": 0.79,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0444,
|
|
"step": 122
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0494,
|
|
"step": 123
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0505,
|
|
"step": 124
|
|
},
|
|
{
|
|
"epoch": 0.81,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.045,
|
|
"step": 125
|
|
},
|
|
{
|
|
"epoch": 0.82,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0492,
|
|
"step": 126
|
|
},
|
|
{
|
|
"epoch": 0.82,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0608,
|
|
"step": 127
|
|
},
|
|
{
|
|
"epoch": 0.83,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0549,
|
|
"step": 128
|
|
},
|
|
{
|
|
"epoch": 0.83,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0526,
|
|
"step": 129
|
|
},
|
|
{
|
|
"epoch": 0.84,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0525,
|
|
"step": 130
|
|
},
|
|
{
|
|
"epoch": 0.85,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0455,
|
|
"step": 131
|
|
},
|
|
{
|
|
"epoch": 0.85,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0528,
|
|
"step": 132
|
|
},
|
|
{
|
|
"epoch": 0.86,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0613,
|
|
"step": 133
|
|
},
|
|
{
|
|
"epoch": 0.87,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0612,
|
|
"step": 134
|
|
},
|
|
{
|
|
"epoch": 0.87,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0449,
|
|
"step": 135
|
|
},
|
|
{
|
|
"epoch": 0.88,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0572,
|
|
"step": 136
|
|
},
|
|
{
|
|
"epoch": 0.89,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0513,
|
|
"step": 137
|
|
},
|
|
{
|
|
"epoch": 0.89,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0637,
|
|
"step": 138
|
|
},
|
|
{
|
|
"epoch": 0.9,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0539,
|
|
"step": 139
|
|
},
|
|
{
|
|
"epoch": 0.91,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0575,
|
|
"step": 140
|
|
},
|
|
{
|
|
"epoch": 0.91,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.056,
|
|
"step": 141
|
|
},
|
|
{
|
|
"epoch": 0.92,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0574,
|
|
"step": 142
|
|
},
|
|
{
|
|
"epoch": 0.93,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0588,
|
|
"step": 143
|
|
},
|
|
{
|
|
"epoch": 0.93,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0495,
|
|
"step": 144
|
|
},
|
|
{
|
|
"epoch": 0.94,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0494,
|
|
"step": 145
|
|
},
|
|
{
|
|
"epoch": 0.94,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0497,
|
|
"step": 146
|
|
},
|
|
{
|
|
"epoch": 0.95,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0583,
|
|
"step": 147
|
|
},
|
|
{
|
|
"epoch": 0.96,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0469,
|
|
"step": 148
|
|
},
|
|
{
|
|
"epoch": 0.96,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0572,
|
|
"step": 149
|
|
},
|
|
{
|
|
"epoch": 0.97,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0605,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.98,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0509,
|
|
"step": 151
|
|
},
|
|
{
|
|
"epoch": 0.98,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0498,
|
|
"step": 152
|
|
},
|
|
{
|
|
"epoch": 0.99,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0519,
|
|
"step": 153
|
|
},
|
|
{
|
|
"epoch": 1.0,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0564,
|
|
"step": 154
|
|
},
|
|
{
|
|
"epoch": 1.0,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0535,
|
|
"step": 155
|
|
},
|
|
{
|
|
"epoch": 1.01,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0367,
|
|
"step": 156
|
|
},
|
|
{
|
|
"epoch": 1.02,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0502,
|
|
"step": 157
|
|
},
|
|
{
|
|
"epoch": 1.02,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0503,
|
|
"step": 158
|
|
},
|
|
{
|
|
"epoch": 1.03,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0424,
|
|
"step": 159
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.049,
|
|
"step": 160
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0421,
|
|
"step": 161
|
|
},
|
|
{
|
|
"epoch": 1.05,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0429,
|
|
"step": 162
|
|
},
|
|
{
|
|
"epoch": 1.05,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0517,
|
|
"step": 163
|
|
},
|
|
{
|
|
"epoch": 1.06,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0521,
|
|
"step": 164
|
|
},
|
|
{
|
|
"epoch": 1.07,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0446,
|
|
"step": 165
|
|
},
|
|
{
|
|
"epoch": 1.07,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0411,
|
|
"step": 166
|
|
},
|
|
{
|
|
"epoch": 1.08,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0334,
|
|
"step": 167
|
|
},
|
|
{
|
|
"epoch": 1.09,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0381,
|
|
"step": 168
|
|
},
|
|
{
|
|
"epoch": 1.09,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0511,
|
|
"step": 169
|
|
},
|
|
{
|
|
"epoch": 1.1,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0424,
|
|
"step": 170
|
|
},
|
|
{
|
|
"epoch": 1.11,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0434,
|
|
"step": 171
|
|
},
|
|
{
|
|
"epoch": 1.11,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0452,
|
|
"step": 172
|
|
},
|
|
{
|
|
"epoch": 1.12,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0433,
|
|
"step": 173
|
|
},
|
|
{
|
|
"epoch": 1.13,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0424,
|
|
"step": 174
|
|
},
|
|
{
|
|
"epoch": 1.13,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0476,
|
|
"step": 175
|
|
},
|
|
{
|
|
"epoch": 1.14,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0532,
|
|
"step": 176
|
|
},
|
|
{
|
|
"epoch": 1.15,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0447,
|
|
"step": 177
|
|
},
|
|
{
|
|
"epoch": 1.15,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0449,
|
|
"step": 178
|
|
},
|
|
{
|
|
"epoch": 1.16,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0421,
|
|
"step": 179
|
|
},
|
|
{
|
|
"epoch": 1.16,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0439,
|
|
"step": 180
|
|
},
|
|
{
|
|
"epoch": 1.17,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0416,
|
|
"step": 181
|
|
},
|
|
{
|
|
"epoch": 1.18,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0422,
|
|
"step": 182
|
|
},
|
|
{
|
|
"epoch": 1.18,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0448,
|
|
"step": 183
|
|
},
|
|
{
|
|
"epoch": 1.19,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0469,
|
|
"step": 184
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0471,
|
|
"step": 185
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0436,
|
|
"step": 186
|
|
},
|
|
{
|
|
"epoch": 1.21,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0447,
|
|
"step": 187
|
|
},
|
|
{
|
|
"epoch": 1.22,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0512,
|
|
"step": 188
|
|
},
|
|
{
|
|
"epoch": 1.22,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0463,
|
|
"step": 189
|
|
},
|
|
{
|
|
"epoch": 1.23,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0434,
|
|
"step": 190
|
|
},
|
|
{
|
|
"epoch": 1.24,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0486,
|
|
"step": 191
|
|
},
|
|
{
|
|
"epoch": 1.24,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0387,
|
|
"step": 192
|
|
},
|
|
{
|
|
"epoch": 1.25,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0379,
|
|
"step": 193
|
|
},
|
|
{
|
|
"epoch": 1.26,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0389,
|
|
"step": 194
|
|
},
|
|
{
|
|
"epoch": 1.26,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0408,
|
|
"step": 195
|
|
},
|
|
{
|
|
"epoch": 1.27,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0449,
|
|
"step": 196
|
|
},
|
|
{
|
|
"epoch": 1.27,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0433,
|
|
"step": 197
|
|
},
|
|
{
|
|
"epoch": 1.28,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0487,
|
|
"step": 198
|
|
},
|
|
{
|
|
"epoch": 1.29,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0314,
|
|
"step": 199
|
|
},
|
|
{
|
|
"epoch": 1.29,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0352,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 1.3,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0428,
|
|
"step": 201
|
|
},
|
|
{
|
|
"epoch": 1.31,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0397,
|
|
"step": 202
|
|
},
|
|
{
|
|
"epoch": 1.31,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0341,
|
|
"step": 203
|
|
},
|
|
{
|
|
"epoch": 1.32,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0373,
|
|
"step": 204
|
|
},
|
|
{
|
|
"epoch": 1.33,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0402,
|
|
"step": 205
|
|
},
|
|
{
|
|
"epoch": 1.33,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0408,
|
|
"step": 206
|
|
},
|
|
{
|
|
"epoch": 1.34,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.043,
|
|
"step": 207
|
|
},
|
|
{
|
|
"epoch": 1.35,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0405,
|
|
"step": 208
|
|
},
|
|
{
|
|
"epoch": 1.35,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0444,
|
|
"step": 209
|
|
},
|
|
{
|
|
"epoch": 1.36,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0412,
|
|
"step": 210
|
|
},
|
|
{
|
|
"epoch": 1.37,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0368,
|
|
"step": 211
|
|
},
|
|
{
|
|
"epoch": 1.37,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0524,
|
|
"step": 212
|
|
},
|
|
{
|
|
"epoch": 1.38,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0439,
|
|
"step": 213
|
|
},
|
|
{
|
|
"epoch": 1.38,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0444,
|
|
"step": 214
|
|
},
|
|
{
|
|
"epoch": 1.39,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0415,
|
|
"step": 215
|
|
},
|
|
{
|
|
"epoch": 1.4,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.041,
|
|
"step": 216
|
|
},
|
|
{
|
|
"epoch": 1.4,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0416,
|
|
"step": 217
|
|
},
|
|
{
|
|
"epoch": 1.41,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0458,
|
|
"step": 218
|
|
},
|
|
{
|
|
"epoch": 1.42,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0399,
|
|
"step": 219
|
|
},
|
|
{
|
|
"epoch": 1.42,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0468,
|
|
"step": 220
|
|
},
|
|
{
|
|
"epoch": 1.43,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0516,
|
|
"step": 221
|
|
},
|
|
{
|
|
"epoch": 1.44,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0434,
|
|
"step": 222
|
|
},
|
|
{
|
|
"epoch": 1.44,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0492,
|
|
"step": 223
|
|
},
|
|
{
|
|
"epoch": 1.45,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0432,
|
|
"step": 224
|
|
},
|
|
{
|
|
"epoch": 1.46,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0451,
|
|
"step": 225
|
|
},
|
|
{
|
|
"epoch": 1.46,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0486,
|
|
"step": 226
|
|
},
|
|
{
|
|
"epoch": 1.47,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0359,
|
|
"step": 227
|
|
},
|
|
{
|
|
"epoch": 1.48,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0423,
|
|
"step": 228
|
|
},
|
|
{
|
|
"epoch": 1.48,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0395,
|
|
"step": 229
|
|
},
|
|
{
|
|
"epoch": 1.49,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0513,
|
|
"step": 230
|
|
},
|
|
{
|
|
"epoch": 1.49,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0381,
|
|
"step": 231
|
|
},
|
|
{
|
|
"epoch": 1.5,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0384,
|
|
"step": 232
|
|
},
|
|
{
|
|
"epoch": 1.51,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0471,
|
|
"step": 233
|
|
},
|
|
{
|
|
"epoch": 1.51,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.041,
|
|
"step": 234
|
|
},
|
|
{
|
|
"epoch": 1.52,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0479,
|
|
"step": 235
|
|
},
|
|
{
|
|
"epoch": 1.53,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.047,
|
|
"step": 236
|
|
},
|
|
{
|
|
"epoch": 1.53,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0441,
|
|
"step": 237
|
|
},
|
|
{
|
|
"epoch": 1.54,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0505,
|
|
"step": 238
|
|
},
|
|
{
|
|
"epoch": 1.55,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0419,
|
|
"step": 239
|
|
},
|
|
{
|
|
"epoch": 1.55,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0409,
|
|
"step": 240
|
|
},
|
|
{
|
|
"epoch": 1.56,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0444,
|
|
"step": 241
|
|
},
|
|
{
|
|
"epoch": 1.57,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0392,
|
|
"step": 242
|
|
},
|
|
{
|
|
"epoch": 1.57,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.037,
|
|
"step": 243
|
|
},
|
|
{
|
|
"epoch": 1.58,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0444,
|
|
"step": 244
|
|
},
|
|
{
|
|
"epoch": 1.59,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0361,
|
|
"step": 245
|
|
},
|
|
{
|
|
"epoch": 1.59,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0468,
|
|
"step": 246
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.052,
|
|
"step": 247
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0453,
|
|
"step": 248
|
|
},
|
|
{
|
|
"epoch": 1.61,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0423,
|
|
"step": 249
|
|
},
|
|
{
|
|
"epoch": 1.62,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0385,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 1.62,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0462,
|
|
"step": 251
|
|
},
|
|
{
|
|
"epoch": 1.63,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0362,
|
|
"step": 252
|
|
},
|
|
{
|
|
"epoch": 1.64,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0388,
|
|
"step": 253
|
|
},
|
|
{
|
|
"epoch": 1.64,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0433,
|
|
"step": 254
|
|
},
|
|
{
|
|
"epoch": 1.65,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.051,
|
|
"step": 255
|
|
},
|
|
{
|
|
"epoch": 1.66,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0362,
|
|
"step": 256
|
|
},
|
|
{
|
|
"epoch": 1.66,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0444,
|
|
"step": 257
|
|
},
|
|
{
|
|
"epoch": 1.67,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0413,
|
|
"step": 258
|
|
},
|
|
{
|
|
"epoch": 1.68,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0331,
|
|
"step": 259
|
|
},
|
|
{
|
|
"epoch": 1.68,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0425,
|
|
"step": 260
|
|
},
|
|
{
|
|
"epoch": 1.69,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0508,
|
|
"step": 261
|
|
},
|
|
{
|
|
"epoch": 1.7,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0445,
|
|
"step": 262
|
|
},
|
|
{
|
|
"epoch": 1.7,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0366,
|
|
"step": 263
|
|
},
|
|
{
|
|
"epoch": 1.71,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0401,
|
|
"step": 264
|
|
},
|
|
{
|
|
"epoch": 1.71,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.037,
|
|
"step": 265
|
|
},
|
|
{
|
|
"epoch": 1.72,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0464,
|
|
"step": 266
|
|
},
|
|
{
|
|
"epoch": 1.73,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0405,
|
|
"step": 267
|
|
},
|
|
{
|
|
"epoch": 1.73,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0455,
|
|
"step": 268
|
|
},
|
|
{
|
|
"epoch": 1.74,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0385,
|
|
"step": 269
|
|
},
|
|
{
|
|
"epoch": 1.75,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0459,
|
|
"step": 270
|
|
},
|
|
{
|
|
"epoch": 1.75,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0464,
|
|
"step": 271
|
|
},
|
|
{
|
|
"epoch": 1.76,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0381,
|
|
"step": 272
|
|
},
|
|
{
|
|
"epoch": 1.77,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0356,
|
|
"step": 273
|
|
},
|
|
{
|
|
"epoch": 1.77,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0454,
|
|
"step": 274
|
|
},
|
|
{
|
|
"epoch": 1.78,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0398,
|
|
"step": 275
|
|
},
|
|
{
|
|
"epoch": 1.79,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0393,
|
|
"step": 276
|
|
},
|
|
{
|
|
"epoch": 1.79,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0399,
|
|
"step": 277
|
|
},
|
|
{
|
|
"epoch": 1.8,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.045,
|
|
"step": 278
|
|
},
|
|
{
|
|
"epoch": 1.81,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0448,
|
|
"step": 279
|
|
},
|
|
{
|
|
"epoch": 1.81,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0441,
|
|
"step": 280
|
|
},
|
|
{
|
|
"epoch": 1.82,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.042,
|
|
"step": 281
|
|
},
|
|
{
|
|
"epoch": 1.82,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0387,
|
|
"step": 282
|
|
},
|
|
{
|
|
"epoch": 1.83,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0383,
|
|
"step": 283
|
|
},
|
|
{
|
|
"epoch": 1.84,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0377,
|
|
"step": 284
|
|
},
|
|
{
|
|
"epoch": 1.84,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0433,
|
|
"step": 285
|
|
},
|
|
{
|
|
"epoch": 1.85,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0424,
|
|
"step": 286
|
|
},
|
|
{
|
|
"epoch": 1.86,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0459,
|
|
"step": 287
|
|
},
|
|
{
|
|
"epoch": 1.86,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0402,
|
|
"step": 288
|
|
},
|
|
{
|
|
"epoch": 1.87,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0389,
|
|
"step": 289
|
|
},
|
|
{
|
|
"epoch": 1.88,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0394,
|
|
"step": 290
|
|
},
|
|
{
|
|
"epoch": 1.88,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0412,
|
|
"step": 291
|
|
},
|
|
{
|
|
"epoch": 1.89,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0493,
|
|
"step": 292
|
|
},
|
|
{
|
|
"epoch": 1.9,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0468,
|
|
"step": 293
|
|
},
|
|
{
|
|
"epoch": 1.9,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0407,
|
|
"step": 294
|
|
},
|
|
{
|
|
"epoch": 1.91,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0376,
|
|
"step": 295
|
|
},
|
|
{
|
|
"epoch": 1.92,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.042,
|
|
"step": 296
|
|
},
|
|
{
|
|
"epoch": 1.92,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0367,
|
|
"step": 297
|
|
},
|
|
{
|
|
"epoch": 1.93,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0365,
|
|
"step": 298
|
|
},
|
|
{
|
|
"epoch": 1.93,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0449,
|
|
"step": 299
|
|
},
|
|
{
|
|
"epoch": 1.94,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0452,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 1.95,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0427,
|
|
"step": 301
|
|
},
|
|
{
|
|
"epoch": 1.95,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0453,
|
|
"step": 302
|
|
},
|
|
{
|
|
"epoch": 1.96,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0351,
|
|
"step": 303
|
|
},
|
|
{
|
|
"epoch": 1.97,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.036,
|
|
"step": 304
|
|
},
|
|
{
|
|
"epoch": 1.97,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0443,
|
|
"step": 305
|
|
},
|
|
{
|
|
"epoch": 1.98,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0369,
|
|
"step": 306
|
|
},
|
|
{
|
|
"epoch": 1.99,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0376,
|
|
"step": 307
|
|
},
|
|
{
|
|
"epoch": 1.99,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.048,
|
|
"step": 308
|
|
},
|
|
{
|
|
"epoch": 2.0,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0366,
|
|
"step": 309
|
|
},
|
|
{
|
|
"epoch": 2.01,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0337,
|
|
"step": 310
|
|
},
|
|
{
|
|
"epoch": 2.01,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0278,
|
|
"step": 311
|
|
},
|
|
{
|
|
"epoch": 2.02,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0279,
|
|
"step": 312
|
|
},
|
|
{
|
|
"epoch": 2.03,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0307,
|
|
"step": 313
|
|
},
|
|
{
|
|
"epoch": 2.03,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0303,
|
|
"step": 314
|
|
},
|
|
{
|
|
"epoch": 2.04,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0286,
|
|
"step": 315
|
|
},
|
|
{
|
|
"epoch": 2.04,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0233,
|
|
"step": 316
|
|
},
|
|
{
|
|
"epoch": 2.05,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0285,
|
|
"step": 317
|
|
},
|
|
{
|
|
"epoch": 2.06,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0258,
|
|
"step": 318
|
|
},
|
|
{
|
|
"epoch": 2.06,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0262,
|
|
"step": 319
|
|
},
|
|
{
|
|
"epoch": 2.07,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0268,
|
|
"step": 320
|
|
},
|
|
{
|
|
"epoch": 2.08,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0273,
|
|
"step": 321
|
|
},
|
|
{
|
|
"epoch": 2.08,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0258,
|
|
"step": 322
|
|
},
|
|
{
|
|
"epoch": 2.09,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0261,
|
|
"step": 323
|
|
},
|
|
{
|
|
"epoch": 2.1,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0269,
|
|
"step": 324
|
|
},
|
|
{
|
|
"epoch": 2.1,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0268,
|
|
"step": 325
|
|
},
|
|
{
|
|
"epoch": 2.11,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0244,
|
|
"step": 326
|
|
},
|
|
{
|
|
"epoch": 2.12,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.026,
|
|
"step": 327
|
|
},
|
|
{
|
|
"epoch": 2.12,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.023,
|
|
"step": 328
|
|
},
|
|
{
|
|
"epoch": 2.13,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0259,
|
|
"step": 329
|
|
},
|
|
{
|
|
"epoch": 2.14,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0288,
|
|
"step": 330
|
|
},
|
|
{
|
|
"epoch": 2.14,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0264,
|
|
"step": 331
|
|
},
|
|
{
|
|
"epoch": 2.15,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0243,
|
|
"step": 332
|
|
},
|
|
{
|
|
"epoch": 2.15,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0319,
|
|
"step": 333
|
|
},
|
|
{
|
|
"epoch": 2.16,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0265,
|
|
"step": 334
|
|
},
|
|
{
|
|
"epoch": 2.17,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0241,
|
|
"step": 335
|
|
},
|
|
{
|
|
"epoch": 2.17,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.028,
|
|
"step": 336
|
|
},
|
|
{
|
|
"epoch": 2.18,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0318,
|
|
"step": 337
|
|
},
|
|
{
|
|
"epoch": 2.19,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0307,
|
|
"step": 338
|
|
},
|
|
{
|
|
"epoch": 2.19,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.026,
|
|
"step": 339
|
|
},
|
|
{
|
|
"epoch": 2.2,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0276,
|
|
"step": 340
|
|
},
|
|
{
|
|
"epoch": 2.21,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.023,
|
|
"step": 341
|
|
},
|
|
{
|
|
"epoch": 2.21,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.028,
|
|
"step": 342
|
|
},
|
|
{
|
|
"epoch": 2.22,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0299,
|
|
"step": 343
|
|
},
|
|
{
|
|
"epoch": 2.23,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0239,
|
|
"step": 344
|
|
},
|
|
{
|
|
"epoch": 2.23,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0269,
|
|
"step": 345
|
|
},
|
|
{
|
|
"epoch": 2.24,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0284,
|
|
"step": 346
|
|
},
|
|
{
|
|
"epoch": 2.25,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0243,
|
|
"step": 347
|
|
},
|
|
{
|
|
"epoch": 2.25,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0237,
|
|
"step": 348
|
|
},
|
|
{
|
|
"epoch": 2.26,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0185,
|
|
"step": 349
|
|
},
|
|
{
|
|
"epoch": 2.26,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0235,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 2.27,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0199,
|
|
"step": 351
|
|
},
|
|
{
|
|
"epoch": 2.28,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0263,
|
|
"step": 352
|
|
},
|
|
{
|
|
"epoch": 2.28,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0271,
|
|
"step": 353
|
|
},
|
|
{
|
|
"epoch": 2.29,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0282,
|
|
"step": 354
|
|
},
|
|
{
|
|
"epoch": 2.3,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0248,
|
|
"step": 355
|
|
},
|
|
{
|
|
"epoch": 2.3,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0281,
|
|
"step": 356
|
|
},
|
|
{
|
|
"epoch": 2.31,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0247,
|
|
"step": 357
|
|
},
|
|
{
|
|
"epoch": 2.32,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.024,
|
|
"step": 358
|
|
},
|
|
{
|
|
"epoch": 2.32,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0311,
|
|
"step": 359
|
|
},
|
|
{
|
|
"epoch": 2.33,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0338,
|
|
"step": 360
|
|
},
|
|
{
|
|
"epoch": 2.34,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.026,
|
|
"step": 361
|
|
},
|
|
{
|
|
"epoch": 2.34,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0255,
|
|
"step": 362
|
|
},
|
|
{
|
|
"epoch": 2.35,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0273,
|
|
"step": 363
|
|
},
|
|
{
|
|
"epoch": 2.36,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0302,
|
|
"step": 364
|
|
},
|
|
{
|
|
"epoch": 2.36,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0248,
|
|
"step": 365
|
|
},
|
|
{
|
|
"epoch": 2.37,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0266,
|
|
"step": 366
|
|
},
|
|
{
|
|
"epoch": 2.37,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0288,
|
|
"step": 367
|
|
},
|
|
{
|
|
"epoch": 2.38,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0344,
|
|
"step": 368
|
|
},
|
|
{
|
|
"epoch": 2.39,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0244,
|
|
"step": 369
|
|
},
|
|
{
|
|
"epoch": 2.39,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0238,
|
|
"step": 370
|
|
},
|
|
{
|
|
"epoch": 2.4,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0279,
|
|
"step": 371
|
|
},
|
|
{
|
|
"epoch": 2.41,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0299,
|
|
"step": 372
|
|
},
|
|
{
|
|
"epoch": 2.41,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0211,
|
|
"step": 373
|
|
},
|
|
{
|
|
"epoch": 2.42,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0257,
|
|
"step": 374
|
|
},
|
|
{
|
|
"epoch": 2.43,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0342,
|
|
"step": 375
|
|
},
|
|
{
|
|
"epoch": 2.43,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0236,
|
|
"step": 376
|
|
},
|
|
{
|
|
"epoch": 2.44,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.035,
|
|
"step": 377
|
|
},
|
|
{
|
|
"epoch": 2.45,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0316,
|
|
"step": 378
|
|
},
|
|
{
|
|
"epoch": 2.45,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0254,
|
|
"step": 379
|
|
},
|
|
{
|
|
"epoch": 2.46,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0281,
|
|
"step": 380
|
|
},
|
|
{
|
|
"epoch": 2.47,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0246,
|
|
"step": 381
|
|
},
|
|
{
|
|
"epoch": 2.47,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0299,
|
|
"step": 382
|
|
},
|
|
{
|
|
"epoch": 2.48,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0269,
|
|
"step": 383
|
|
},
|
|
{
|
|
"epoch": 2.48,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0285,
|
|
"step": 384
|
|
},
|
|
{
|
|
"epoch": 2.49,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0346,
|
|
"step": 385
|
|
},
|
|
{
|
|
"epoch": 2.5,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.024,
|
|
"step": 386
|
|
},
|
|
{
|
|
"epoch": 2.5,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0353,
|
|
"step": 387
|
|
},
|
|
{
|
|
"epoch": 2.51,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0303,
|
|
"step": 388
|
|
},
|
|
{
|
|
"epoch": 2.52,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.03,
|
|
"step": 389
|
|
},
|
|
{
|
|
"epoch": 2.52,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0277,
|
|
"step": 390
|
|
},
|
|
{
|
|
"epoch": 2.53,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0232,
|
|
"step": 391
|
|
},
|
|
{
|
|
"epoch": 2.54,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0314,
|
|
"step": 392
|
|
},
|
|
{
|
|
"epoch": 2.54,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0259,
|
|
"step": 393
|
|
},
|
|
{
|
|
"epoch": 2.55,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0304,
|
|
"step": 394
|
|
},
|
|
{
|
|
"epoch": 2.56,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0274,
|
|
"step": 395
|
|
},
|
|
{
|
|
"epoch": 2.56,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0303,
|
|
"step": 396
|
|
},
|
|
{
|
|
"epoch": 2.57,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0279,
|
|
"step": 397
|
|
},
|
|
{
|
|
"epoch": 2.58,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0281,
|
|
"step": 398
|
|
},
|
|
{
|
|
"epoch": 2.58,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0285,
|
|
"step": 399
|
|
},
|
|
{
|
|
"epoch": 2.59,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0355,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 2.59,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0279,
|
|
"step": 401
|
|
},
|
|
{
|
|
"epoch": 2.6,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0312,
|
|
"step": 402
|
|
},
|
|
{
|
|
"epoch": 2.61,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0258,
|
|
"step": 403
|
|
},
|
|
{
|
|
"epoch": 2.61,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0304,
|
|
"step": 404
|
|
},
|
|
{
|
|
"epoch": 2.62,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0271,
|
|
"step": 405
|
|
},
|
|
{
|
|
"epoch": 2.63,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0271,
|
|
"step": 406
|
|
},
|
|
{
|
|
"epoch": 2.63,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0248,
|
|
"step": 407
|
|
},
|
|
{
|
|
"epoch": 2.64,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0261,
|
|
"step": 408
|
|
},
|
|
{
|
|
"epoch": 2.65,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0284,
|
|
"step": 409
|
|
},
|
|
{
|
|
"epoch": 2.65,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0288,
|
|
"step": 410
|
|
},
|
|
{
|
|
"epoch": 2.66,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0296,
|
|
"step": 411
|
|
},
|
|
{
|
|
"epoch": 2.67,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.028,
|
|
"step": 412
|
|
},
|
|
{
|
|
"epoch": 2.67,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0232,
|
|
"step": 413
|
|
},
|
|
{
|
|
"epoch": 2.68,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0281,
|
|
"step": 414
|
|
},
|
|
{
|
|
"epoch": 2.68,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0337,
|
|
"step": 415
|
|
},
|
|
{
|
|
"epoch": 2.69,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0228,
|
|
"step": 416
|
|
},
|
|
{
|
|
"epoch": 2.7,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0319,
|
|
"step": 417
|
|
},
|
|
{
|
|
"epoch": 2.7,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0321,
|
|
"step": 418
|
|
},
|
|
{
|
|
"epoch": 2.71,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0312,
|
|
"step": 419
|
|
},
|
|
{
|
|
"epoch": 2.72,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0276,
|
|
"step": 420
|
|
},
|
|
{
|
|
"epoch": 2.72,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0241,
|
|
"step": 421
|
|
},
|
|
{
|
|
"epoch": 2.73,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0302,
|
|
"step": 422
|
|
},
|
|
{
|
|
"epoch": 2.74,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0271,
|
|
"step": 423
|
|
},
|
|
{
|
|
"epoch": 2.74,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0319,
|
|
"step": 424
|
|
},
|
|
{
|
|
"epoch": 2.75,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0269,
|
|
"step": 425
|
|
},
|
|
{
|
|
"epoch": 2.76,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0247,
|
|
"step": 426
|
|
},
|
|
{
|
|
"epoch": 2.76,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0241,
|
|
"step": 427
|
|
},
|
|
{
|
|
"epoch": 2.77,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0245,
|
|
"step": 428
|
|
},
|
|
{
|
|
"epoch": 2.78,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0262,
|
|
"step": 429
|
|
},
|
|
{
|
|
"epoch": 2.78,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0307,
|
|
"step": 430
|
|
},
|
|
{
|
|
"epoch": 2.79,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0293,
|
|
"step": 431
|
|
},
|
|
{
|
|
"epoch": 2.79,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0235,
|
|
"step": 432
|
|
},
|
|
{
|
|
"epoch": 2.8,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0257,
|
|
"step": 433
|
|
},
|
|
{
|
|
"epoch": 2.81,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0301,
|
|
"step": 434
|
|
},
|
|
{
|
|
"epoch": 2.81,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0266,
|
|
"step": 435
|
|
},
|
|
{
|
|
"epoch": 2.82,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.024,
|
|
"step": 436
|
|
},
|
|
{
|
|
"epoch": 2.83,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0279,
|
|
"step": 437
|
|
},
|
|
{
|
|
"epoch": 2.83,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0285,
|
|
"step": 438
|
|
},
|
|
{
|
|
"epoch": 2.84,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0264,
|
|
"step": 439
|
|
},
|
|
{
|
|
"epoch": 2.85,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0307,
|
|
"step": 440
|
|
},
|
|
{
|
|
"epoch": 2.85,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0273,
|
|
"step": 441
|
|
},
|
|
{
|
|
"epoch": 2.86,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0306,
|
|
"step": 442
|
|
},
|
|
{
|
|
"epoch": 2.87,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0273,
|
|
"step": 443
|
|
},
|
|
{
|
|
"epoch": 2.87,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0241,
|
|
"step": 444
|
|
},
|
|
{
|
|
"epoch": 2.88,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0333,
|
|
"step": 445
|
|
},
|
|
{
|
|
"epoch": 2.89,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0251,
|
|
"step": 446
|
|
},
|
|
{
|
|
"epoch": 2.89,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.026,
|
|
"step": 447
|
|
},
|
|
{
|
|
"epoch": 2.9,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0309,
|
|
"step": 448
|
|
},
|
|
{
|
|
"epoch": 2.9,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0317,
|
|
"step": 449
|
|
},
|
|
{
|
|
"epoch": 2.91,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0298,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 2.92,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0303,
|
|
"step": 451
|
|
},
|
|
{
|
|
"epoch": 2.92,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0317,
|
|
"step": 452
|
|
},
|
|
{
|
|
"epoch": 2.93,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0271,
|
|
"step": 453
|
|
},
|
|
{
|
|
"epoch": 2.94,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0301,
|
|
"step": 454
|
|
},
|
|
{
|
|
"epoch": 2.94,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0272,
|
|
"step": 455
|
|
},
|
|
{
|
|
"epoch": 2.95,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0302,
|
|
"step": 456
|
|
},
|
|
{
|
|
"epoch": 2.96,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.031,
|
|
"step": 457
|
|
},
|
|
{
|
|
"epoch": 2.96,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0293,
|
|
"step": 458
|
|
},
|
|
{
|
|
"epoch": 2.97,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.023,
|
|
"step": 459
|
|
},
|
|
{
|
|
"epoch": 2.98,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0263,
|
|
"step": 460
|
|
},
|
|
{
|
|
"epoch": 2.98,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0279,
|
|
"step": 461
|
|
},
|
|
{
|
|
"epoch": 2.99,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.0272,
|
|
"step": 462
|
|
}
|
|
],
|
|
"max_steps": 462,
|
|
"num_train_epochs": 3,
|
|
"total_flos": 68006471270400.0,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|