instructscore_data2text / trainer_state.json
Wendaxu's picture
Update from wendaxu
d4eae12
Raw
History Blame Contribute Delete
46.6 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.988717091516924,
"global_step": 447,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 0,
"loss": 9.6951,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 0,
"loss": 10.3828,
"step": 2
},
{
"epoch": 0.02,
"learning_rate": 0,
"loss": 10.3819,
"step": 3
},
{
"epoch": 0.03,
"learning_rate": 0,
"loss": 9.8255,
"step": 4
},
{
"epoch": 0.03,
"learning_rate": 0,
"loss": 9.7955,
"step": 5
},
{
"epoch": 0.04,
"learning_rate": 0,
"loss": 10.444,
"step": 6
},
{
"epoch": 0.05,
"learning_rate": 0.0,
"loss": 9.5294,
"step": 7
},
{
"epoch": 0.05,
"learning_rate": 0.0,
"loss": 10.4795,
"step": 8
},
{
"epoch": 0.06,
"learning_rate": 0.0,
"loss": 9.4406,
"step": 9
},
{
"epoch": 0.07,
"learning_rate": 2e-05,
"loss": 10.2562,
"step": 10
},
{
"epoch": 0.07,
"learning_rate": 2e-05,
"loss": 10.7465,
"step": 11
},
{
"epoch": 0.08,
"learning_rate": 2e-05,
"loss": 8.6074,
"step": 12
},
{
"epoch": 0.09,
"learning_rate": 2e-05,
"loss": 8.1801,
"step": 13
},
{
"epoch": 0.09,
"learning_rate": 2e-05,
"loss": 7.7766,
"step": 14
},
{
"epoch": 0.1,
"learning_rate": 2e-05,
"loss": 8.0473,
"step": 15
},
{
"epoch": 0.11,
"learning_rate": 2e-05,
"loss": 7.4571,
"step": 16
},
{
"epoch": 0.11,
"learning_rate": 2e-05,
"loss": 8.0565,
"step": 17
},
{
"epoch": 0.12,
"learning_rate": 2e-05,
"loss": 7.7186,
"step": 18
},
{
"epoch": 0.13,
"learning_rate": 2e-05,
"loss": 7.4039,
"step": 19
},
{
"epoch": 0.13,
"learning_rate": 2e-05,
"loss": 7.3152,
"step": 20
},
{
"epoch": 0.14,
"learning_rate": 2e-05,
"loss": 7.8044,
"step": 21
},
{
"epoch": 0.15,
"learning_rate": 2e-05,
"loss": 7.644,
"step": 22
},
{
"epoch": 0.15,
"learning_rate": 2e-05,
"loss": 7.4755,
"step": 23
},
{
"epoch": 0.16,
"learning_rate": 2e-05,
"loss": 7.0238,
"step": 24
},
{
"epoch": 0.17,
"learning_rate": 2e-05,
"loss": 7.0813,
"step": 25
},
{
"epoch": 0.17,
"learning_rate": 2e-05,
"loss": 7.1864,
"step": 26
},
{
"epoch": 0.18,
"learning_rate": 2e-05,
"loss": 6.6825,
"step": 27
},
{
"epoch": 0.19,
"learning_rate": 2e-05,
"loss": 6.412,
"step": 28
},
{
"epoch": 0.19,
"learning_rate": 2e-05,
"loss": 6.2732,
"step": 29
},
{
"epoch": 0.2,
"learning_rate": 2e-05,
"loss": 6.15,
"step": 30
},
{
"epoch": 0.21,
"learning_rate": 2e-05,
"loss": 5.9155,
"step": 31
},
{
"epoch": 0.21,
"learning_rate": 2e-05,
"loss": 5.9589,
"step": 32
},
{
"epoch": 0.22,
"learning_rate": 2e-05,
"loss": 5.8398,
"step": 33
},
{
"epoch": 0.23,
"learning_rate": 2e-05,
"loss": 5.5758,
"step": 34
},
{
"epoch": 0.23,
"learning_rate": 2e-05,
"loss": 5.4107,
"step": 35
},
{
"epoch": 0.24,
"learning_rate": 2e-05,
"loss": 5.079,
"step": 36
},
{
"epoch": 0.25,
"learning_rate": 2e-05,
"loss": 5.2523,
"step": 37
},
{
"epoch": 0.25,
"learning_rate": 2e-05,
"loss": 4.8414,
"step": 38
},
{
"epoch": 0.26,
"learning_rate": 2e-05,
"loss": 4.5972,
"step": 39
},
{
"epoch": 0.27,
"learning_rate": 2e-05,
"loss": 4.4326,
"step": 40
},
{
"epoch": 0.27,
"learning_rate": 2e-05,
"loss": 4.1486,
"step": 41
},
{
"epoch": 0.28,
"learning_rate": 2e-05,
"loss": 4.1405,
"step": 42
},
{
"epoch": 0.29,
"learning_rate": 2e-05,
"loss": 4.0155,
"step": 43
},
{
"epoch": 0.29,
"learning_rate": 2e-05,
"loss": 3.6778,
"step": 44
},
{
"epoch": 0.3,
"learning_rate": 2e-05,
"loss": 3.578,
"step": 45
},
{
"epoch": 0.31,
"learning_rate": 2e-05,
"loss": 3.3746,
"step": 46
},
{
"epoch": 0.31,
"learning_rate": 2e-05,
"loss": 2.8532,
"step": 47
},
{
"epoch": 0.32,
"learning_rate": 2e-05,
"loss": 2.9177,
"step": 48
},
{
"epoch": 0.33,
"learning_rate": 2e-05,
"loss": 2.7433,
"step": 49
},
{
"epoch": 0.33,
"learning_rate": 2e-05,
"loss": 2.5093,
"step": 50
},
{
"epoch": 0.34,
"learning_rate": 2e-05,
"loss": 2.3509,
"step": 51
},
{
"epoch": 0.35,
"learning_rate": 2e-05,
"loss": 2.1956,
"step": 52
},
{
"epoch": 0.35,
"learning_rate": 2e-05,
"loss": 2.1043,
"step": 53
},
{
"epoch": 0.36,
"learning_rate": 2e-05,
"loss": 1.9602,
"step": 54
},
{
"epoch": 0.37,
"learning_rate": 2e-05,
"loss": 1.5427,
"step": 55
},
{
"epoch": 0.37,
"learning_rate": 2e-05,
"loss": 1.5504,
"step": 56
},
{
"epoch": 0.38,
"learning_rate": 2e-05,
"loss": 1.3438,
"step": 57
},
{
"epoch": 0.39,
"learning_rate": 2e-05,
"loss": 1.1685,
"step": 58
},
{
"epoch": 0.39,
"learning_rate": 2e-05,
"loss": 0.9803,
"step": 59
},
{
"epoch": 0.4,
"learning_rate": 2e-05,
"loss": 0.8676,
"step": 60
},
{
"epoch": 0.41,
"learning_rate": 2e-05,
"loss": 0.7606,
"step": 61
},
{
"epoch": 0.41,
"learning_rate": 2e-05,
"loss": 0.625,
"step": 62
},
{
"epoch": 0.42,
"learning_rate": 2e-05,
"loss": 0.529,
"step": 63
},
{
"epoch": 0.43,
"learning_rate": 2e-05,
"loss": 0.4348,
"step": 64
},
{
"epoch": 0.43,
"learning_rate": 2e-05,
"loss": 0.3692,
"step": 65
},
{
"epoch": 0.44,
"learning_rate": 2e-05,
"loss": 0.313,
"step": 66
},
{
"epoch": 0.45,
"learning_rate": 2e-05,
"loss": 0.2629,
"step": 67
},
{
"epoch": 0.45,
"learning_rate": 2e-05,
"loss": 0.2398,
"step": 68
},
{
"epoch": 0.46,
"learning_rate": 2e-05,
"loss": 0.1941,
"step": 69
},
{
"epoch": 0.47,
"learning_rate": 2e-05,
"loss": 0.1701,
"step": 70
},
{
"epoch": 0.47,
"learning_rate": 2e-05,
"loss": 0.1606,
"step": 71
},
{
"epoch": 0.48,
"learning_rate": 2e-05,
"loss": 0.1388,
"step": 72
},
{
"epoch": 0.49,
"learning_rate": 2e-05,
"loss": 0.1279,
"step": 73
},
{
"epoch": 0.49,
"learning_rate": 2e-05,
"loss": 0.1251,
"step": 74
},
{
"epoch": 0.5,
"learning_rate": 2e-05,
"loss": 0.1289,
"step": 75
},
{
"epoch": 0.51,
"learning_rate": 2e-05,
"loss": 0.1032,
"step": 76
},
{
"epoch": 0.51,
"learning_rate": 2e-05,
"loss": 0.1145,
"step": 77
},
{
"epoch": 0.52,
"learning_rate": 2e-05,
"loss": 0.0967,
"step": 78
},
{
"epoch": 0.53,
"learning_rate": 2e-05,
"loss": 0.1181,
"step": 79
},
{
"epoch": 0.53,
"learning_rate": 2e-05,
"loss": 0.0967,
"step": 80
},
{
"epoch": 0.54,
"learning_rate": 2e-05,
"loss": 0.0921,
"step": 81
},
{
"epoch": 0.55,
"learning_rate": 2e-05,
"loss": 0.1068,
"step": 82
},
{
"epoch": 0.55,
"learning_rate": 2e-05,
"loss": 0.1203,
"step": 83
},
{
"epoch": 0.56,
"learning_rate": 2e-05,
"loss": 0.0996,
"step": 84
},
{
"epoch": 0.57,
"learning_rate": 2e-05,
"loss": 0.1081,
"step": 85
},
{
"epoch": 0.58,
"learning_rate": 2e-05,
"loss": 0.1009,
"step": 86
},
{
"epoch": 0.58,
"learning_rate": 2e-05,
"loss": 0.0929,
"step": 87
},
{
"epoch": 0.59,
"learning_rate": 2e-05,
"loss": 0.1018,
"step": 88
},
{
"epoch": 0.6,
"learning_rate": 2e-05,
"loss": 0.0947,
"step": 89
},
{
"epoch": 0.6,
"learning_rate": 2e-05,
"loss": 0.0817,
"step": 90
},
{
"epoch": 0.61,
"learning_rate": 2e-05,
"loss": 0.0704,
"step": 91
},
{
"epoch": 0.62,
"learning_rate": 2e-05,
"loss": 0.087,
"step": 92
},
{
"epoch": 0.62,
"learning_rate": 2e-05,
"loss": 0.0751,
"step": 93
},
{
"epoch": 0.63,
"learning_rate": 2e-05,
"loss": 0.0883,
"step": 94
},
{
"epoch": 0.64,
"learning_rate": 2e-05,
"loss": 0.086,
"step": 95
},
{
"epoch": 0.64,
"learning_rate": 2e-05,
"loss": 0.0931,
"step": 96
},
{
"epoch": 0.65,
"learning_rate": 2e-05,
"loss": 0.0747,
"step": 97
},
{
"epoch": 0.66,
"learning_rate": 2e-05,
"loss": 0.0882,
"step": 98
},
{
"epoch": 0.66,
"learning_rate": 2e-05,
"loss": 0.0822,
"step": 99
},
{
"epoch": 0.67,
"learning_rate": 2e-05,
"loss": 0.0774,
"step": 100
},
{
"epoch": 0.68,
"learning_rate": 2e-05,
"loss": 0.0694,
"step": 101
},
{
"epoch": 0.68,
"learning_rate": 2e-05,
"loss": 0.0751,
"step": 102
},
{
"epoch": 0.69,
"learning_rate": 2e-05,
"loss": 0.0826,
"step": 103
},
{
"epoch": 0.7,
"learning_rate": 2e-05,
"loss": 0.0829,
"step": 104
},
{
"epoch": 0.7,
"learning_rate": 2e-05,
"loss": 0.0698,
"step": 105
},
{
"epoch": 0.71,
"learning_rate": 2e-05,
"loss": 0.079,
"step": 106
},
{
"epoch": 0.72,
"learning_rate": 2e-05,
"loss": 0.0878,
"step": 107
},
{
"epoch": 0.72,
"learning_rate": 2e-05,
"loss": 0.0636,
"step": 108
},
{
"epoch": 0.73,
"learning_rate": 2e-05,
"loss": 0.0666,
"step": 109
},
{
"epoch": 0.74,
"learning_rate": 2e-05,
"loss": 0.0682,
"step": 110
},
{
"epoch": 0.74,
"learning_rate": 2e-05,
"loss": 0.0856,
"step": 111
},
{
"epoch": 0.75,
"learning_rate": 2e-05,
"loss": 0.0849,
"step": 112
},
{
"epoch": 0.76,
"learning_rate": 2e-05,
"loss": 0.0705,
"step": 113
},
{
"epoch": 0.76,
"learning_rate": 2e-05,
"loss": 0.0701,
"step": 114
},
{
"epoch": 0.77,
"learning_rate": 2e-05,
"loss": 0.0722,
"step": 115
},
{
"epoch": 0.78,
"learning_rate": 2e-05,
"loss": 0.0739,
"step": 116
},
{
"epoch": 0.78,
"learning_rate": 2e-05,
"loss": 0.0717,
"step": 117
},
{
"epoch": 0.79,
"learning_rate": 2e-05,
"loss": 0.0765,
"step": 118
},
{
"epoch": 0.8,
"learning_rate": 2e-05,
"loss": 0.0872,
"step": 119
},
{
"epoch": 0.8,
"learning_rate": 2e-05,
"loss": 0.0734,
"step": 120
},
{
"epoch": 0.81,
"learning_rate": 2e-05,
"loss": 0.07,
"step": 121
},
{
"epoch": 0.82,
"learning_rate": 2e-05,
"loss": 0.0756,
"step": 122
},
{
"epoch": 0.82,
"learning_rate": 2e-05,
"loss": 0.0812,
"step": 123
},
{
"epoch": 0.83,
"learning_rate": 2e-05,
"loss": 0.0807,
"step": 124
},
{
"epoch": 0.84,
"learning_rate": 2e-05,
"loss": 0.0753,
"step": 125
},
{
"epoch": 0.84,
"learning_rate": 2e-05,
"loss": 0.0704,
"step": 126
},
{
"epoch": 0.85,
"learning_rate": 2e-05,
"loss": 0.0858,
"step": 127
},
{
"epoch": 0.86,
"learning_rate": 2e-05,
"loss": 0.0709,
"step": 128
},
{
"epoch": 0.86,
"learning_rate": 2e-05,
"loss": 0.0797,
"step": 129
},
{
"epoch": 0.87,
"learning_rate": 2e-05,
"loss": 0.077,
"step": 130
},
{
"epoch": 0.88,
"learning_rate": 2e-05,
"loss": 0.0774,
"step": 131
},
{
"epoch": 0.88,
"learning_rate": 2e-05,
"loss": 0.0858,
"step": 132
},
{
"epoch": 0.89,
"learning_rate": 2e-05,
"loss": 0.0826,
"step": 133
},
{
"epoch": 0.9,
"learning_rate": 2e-05,
"loss": 0.0685,
"step": 134
},
{
"epoch": 0.9,
"learning_rate": 2e-05,
"loss": 0.067,
"step": 135
},
{
"epoch": 0.91,
"learning_rate": 2e-05,
"loss": 0.0759,
"step": 136
},
{
"epoch": 0.92,
"learning_rate": 2e-05,
"loss": 0.0644,
"step": 137
},
{
"epoch": 0.92,
"learning_rate": 2e-05,
"loss": 0.0619,
"step": 138
},
{
"epoch": 0.93,
"learning_rate": 2e-05,
"loss": 0.0802,
"step": 139
},
{
"epoch": 0.94,
"learning_rate": 2e-05,
"loss": 0.0755,
"step": 140
},
{
"epoch": 0.94,
"learning_rate": 2e-05,
"loss": 0.0793,
"step": 141
},
{
"epoch": 0.95,
"learning_rate": 2e-05,
"loss": 0.0731,
"step": 142
},
{
"epoch": 0.96,
"learning_rate": 2e-05,
"loss": 0.0845,
"step": 143
},
{
"epoch": 0.96,
"learning_rate": 2e-05,
"loss": 0.0745,
"step": 144
},
{
"epoch": 0.97,
"learning_rate": 2e-05,
"loss": 0.0861,
"step": 145
},
{
"epoch": 0.98,
"learning_rate": 2e-05,
"loss": 0.0689,
"step": 146
},
{
"epoch": 0.98,
"learning_rate": 2e-05,
"loss": 0.0712,
"step": 147
},
{
"epoch": 0.99,
"learning_rate": 2e-05,
"loss": 0.0561,
"step": 148
},
{
"epoch": 1.0,
"learning_rate": 2e-05,
"loss": 0.0733,
"step": 149
},
{
"epoch": 1.0,
"learning_rate": 2e-05,
"loss": 0.0613,
"step": 150
},
{
"epoch": 1.01,
"learning_rate": 2e-05,
"loss": 0.0544,
"step": 151
},
{
"epoch": 1.02,
"learning_rate": 2e-05,
"loss": 0.0743,
"step": 152
},
{
"epoch": 1.02,
"learning_rate": 2e-05,
"loss": 0.0525,
"step": 153
},
{
"epoch": 1.03,
"learning_rate": 2e-05,
"loss": 0.0596,
"step": 154
},
{
"epoch": 1.04,
"learning_rate": 2e-05,
"loss": 0.0611,
"step": 155
},
{
"epoch": 1.04,
"learning_rate": 2e-05,
"loss": 0.0468,
"step": 156
},
{
"epoch": 1.05,
"learning_rate": 2e-05,
"loss": 0.0575,
"step": 157
},
{
"epoch": 1.06,
"learning_rate": 2e-05,
"loss": 0.055,
"step": 158
},
{
"epoch": 1.06,
"learning_rate": 2e-05,
"loss": 0.0465,
"step": 159
},
{
"epoch": 1.07,
"learning_rate": 2e-05,
"loss": 0.0628,
"step": 160
},
{
"epoch": 1.08,
"learning_rate": 2e-05,
"loss": 0.0542,
"step": 161
},
{
"epoch": 1.08,
"learning_rate": 2e-05,
"loss": 0.0611,
"step": 162
},
{
"epoch": 1.09,
"learning_rate": 2e-05,
"loss": 0.0747,
"step": 163
},
{
"epoch": 1.1,
"learning_rate": 2e-05,
"loss": 0.0605,
"step": 164
},
{
"epoch": 1.1,
"learning_rate": 2e-05,
"loss": 0.052,
"step": 165
},
{
"epoch": 1.11,
"learning_rate": 2e-05,
"loss": 0.0573,
"step": 166
},
{
"epoch": 1.12,
"learning_rate": 2e-05,
"loss": 0.0597,
"step": 167
},
{
"epoch": 1.12,
"learning_rate": 2e-05,
"loss": 0.0703,
"step": 168
},
{
"epoch": 1.13,
"learning_rate": 2e-05,
"loss": 0.0501,
"step": 169
},
{
"epoch": 1.14,
"learning_rate": 2e-05,
"loss": 0.0651,
"step": 170
},
{
"epoch": 1.14,
"learning_rate": 2e-05,
"loss": 0.0651,
"step": 171
},
{
"epoch": 1.15,
"learning_rate": 2e-05,
"loss": 0.0536,
"step": 172
},
{
"epoch": 1.16,
"learning_rate": 2e-05,
"loss": 0.0605,
"step": 173
},
{
"epoch": 1.16,
"learning_rate": 2e-05,
"loss": 0.0625,
"step": 174
},
{
"epoch": 1.17,
"learning_rate": 2e-05,
"loss": 0.0662,
"step": 175
},
{
"epoch": 1.18,
"learning_rate": 2e-05,
"loss": 0.0583,
"step": 176
},
{
"epoch": 1.18,
"learning_rate": 2e-05,
"loss": 0.0499,
"step": 177
},
{
"epoch": 1.19,
"learning_rate": 2e-05,
"loss": 0.0641,
"step": 178
},
{
"epoch": 1.2,
"learning_rate": 2e-05,
"loss": 0.0522,
"step": 179
},
{
"epoch": 1.2,
"learning_rate": 2e-05,
"loss": 0.058,
"step": 180
},
{
"epoch": 1.21,
"learning_rate": 2e-05,
"loss": 0.0496,
"step": 181
},
{
"epoch": 1.22,
"learning_rate": 2e-05,
"loss": 0.0626,
"step": 182
},
{
"epoch": 1.22,
"learning_rate": 2e-05,
"loss": 0.0546,
"step": 183
},
{
"epoch": 1.23,
"learning_rate": 2e-05,
"loss": 0.0557,
"step": 184
},
{
"epoch": 1.24,
"learning_rate": 2e-05,
"loss": 0.0541,
"step": 185
},
{
"epoch": 1.24,
"learning_rate": 2e-05,
"loss": 0.0654,
"step": 186
},
{
"epoch": 1.25,
"learning_rate": 2e-05,
"loss": 0.0705,
"step": 187
},
{
"epoch": 1.26,
"learning_rate": 2e-05,
"loss": 0.057,
"step": 188
},
{
"epoch": 1.26,
"learning_rate": 2e-05,
"loss": 0.0492,
"step": 189
},
{
"epoch": 1.27,
"learning_rate": 2e-05,
"loss": 0.0442,
"step": 190
},
{
"epoch": 1.28,
"learning_rate": 2e-05,
"loss": 0.0628,
"step": 191
},
{
"epoch": 1.28,
"learning_rate": 2e-05,
"loss": 0.0578,
"step": 192
},
{
"epoch": 1.29,
"learning_rate": 2e-05,
"loss": 0.0628,
"step": 193
},
{
"epoch": 1.3,
"learning_rate": 2e-05,
"loss": 0.0571,
"step": 194
},
{
"epoch": 1.3,
"learning_rate": 2e-05,
"loss": 0.0566,
"step": 195
},
{
"epoch": 1.31,
"learning_rate": 2e-05,
"loss": 0.0638,
"step": 196
},
{
"epoch": 1.32,
"learning_rate": 2e-05,
"loss": 0.0663,
"step": 197
},
{
"epoch": 1.32,
"learning_rate": 2e-05,
"loss": 0.0641,
"step": 198
},
{
"epoch": 1.33,
"learning_rate": 2e-05,
"loss": 0.0548,
"step": 199
},
{
"epoch": 1.34,
"learning_rate": 2e-05,
"loss": 0.0554,
"step": 200
},
{
"epoch": 1.34,
"learning_rate": 2e-05,
"loss": 0.0746,
"step": 201
},
{
"epoch": 1.35,
"learning_rate": 2e-05,
"loss": 0.0473,
"step": 202
},
{
"epoch": 1.36,
"learning_rate": 2e-05,
"loss": 0.0598,
"step": 203
},
{
"epoch": 1.36,
"learning_rate": 2e-05,
"loss": 0.0532,
"step": 204
},
{
"epoch": 1.37,
"learning_rate": 2e-05,
"loss": 0.0584,
"step": 205
},
{
"epoch": 1.38,
"learning_rate": 2e-05,
"loss": 0.0565,
"step": 206
},
{
"epoch": 1.38,
"learning_rate": 2e-05,
"loss": 0.0637,
"step": 207
},
{
"epoch": 1.39,
"learning_rate": 2e-05,
"loss": 0.0655,
"step": 208
},
{
"epoch": 1.4,
"learning_rate": 2e-05,
"loss": 0.063,
"step": 209
},
{
"epoch": 1.4,
"learning_rate": 2e-05,
"loss": 0.0696,
"step": 210
},
{
"epoch": 1.41,
"learning_rate": 2e-05,
"loss": 0.047,
"step": 211
},
{
"epoch": 1.42,
"learning_rate": 2e-05,
"loss": 0.0549,
"step": 212
},
{
"epoch": 1.42,
"learning_rate": 2e-05,
"loss": 0.0647,
"step": 213
},
{
"epoch": 1.43,
"learning_rate": 2e-05,
"loss": 0.0809,
"step": 214
},
{
"epoch": 1.44,
"learning_rate": 2e-05,
"loss": 0.045,
"step": 215
},
{
"epoch": 1.44,
"learning_rate": 2e-05,
"loss": 0.0542,
"step": 216
},
{
"epoch": 1.45,
"learning_rate": 2e-05,
"loss": 0.0672,
"step": 217
},
{
"epoch": 1.46,
"learning_rate": 2e-05,
"loss": 0.068,
"step": 218
},
{
"epoch": 1.46,
"learning_rate": 2e-05,
"loss": 0.0624,
"step": 219
},
{
"epoch": 1.47,
"learning_rate": 2e-05,
"loss": 0.0601,
"step": 220
},
{
"epoch": 1.48,
"learning_rate": 2e-05,
"loss": 0.0627,
"step": 221
},
{
"epoch": 1.48,
"learning_rate": 2e-05,
"loss": 0.0458,
"step": 222
},
{
"epoch": 1.49,
"learning_rate": 2e-05,
"loss": 0.061,
"step": 223
},
{
"epoch": 1.5,
"learning_rate": 2e-05,
"loss": 0.0551,
"step": 224
},
{
"epoch": 1.5,
"learning_rate": 2e-05,
"loss": 0.0585,
"step": 225
},
{
"epoch": 1.51,
"learning_rate": 2e-05,
"loss": 0.0593,
"step": 226
},
{
"epoch": 1.52,
"learning_rate": 2e-05,
"loss": 0.0498,
"step": 227
},
{
"epoch": 1.52,
"learning_rate": 2e-05,
"loss": 0.0551,
"step": 228
},
{
"epoch": 1.53,
"learning_rate": 2e-05,
"loss": 0.0558,
"step": 229
},
{
"epoch": 1.54,
"learning_rate": 2e-05,
"loss": 0.069,
"step": 230
},
{
"epoch": 1.54,
"learning_rate": 2e-05,
"loss": 0.0669,
"step": 231
},
{
"epoch": 1.55,
"learning_rate": 2e-05,
"loss": 0.0536,
"step": 232
},
{
"epoch": 1.56,
"learning_rate": 2e-05,
"loss": 0.0554,
"step": 233
},
{
"epoch": 1.56,
"learning_rate": 2e-05,
"loss": 0.0592,
"step": 234
},
{
"epoch": 1.57,
"learning_rate": 2e-05,
"loss": 0.0594,
"step": 235
},
{
"epoch": 1.58,
"learning_rate": 2e-05,
"loss": 0.0559,
"step": 236
},
{
"epoch": 1.58,
"learning_rate": 2e-05,
"loss": 0.0464,
"step": 237
},
{
"epoch": 1.59,
"learning_rate": 2e-05,
"loss": 0.0538,
"step": 238
},
{
"epoch": 1.6,
"learning_rate": 2e-05,
"loss": 0.0581,
"step": 239
},
{
"epoch": 1.6,
"learning_rate": 2e-05,
"loss": 0.0493,
"step": 240
},
{
"epoch": 1.61,
"learning_rate": 2e-05,
"loss": 0.0549,
"step": 241
},
{
"epoch": 1.62,
"learning_rate": 2e-05,
"loss": 0.0465,
"step": 242
},
{
"epoch": 1.62,
"learning_rate": 2e-05,
"loss": 0.0528,
"step": 243
},
{
"epoch": 1.63,
"learning_rate": 2e-05,
"loss": 0.059,
"step": 244
},
{
"epoch": 1.64,
"learning_rate": 2e-05,
"loss": 0.0444,
"step": 245
},
{
"epoch": 1.64,
"learning_rate": 2e-05,
"loss": 0.0559,
"step": 246
},
{
"epoch": 1.65,
"learning_rate": 2e-05,
"loss": 0.0744,
"step": 247
},
{
"epoch": 1.66,
"learning_rate": 2e-05,
"loss": 0.0627,
"step": 248
},
{
"epoch": 1.66,
"learning_rate": 2e-05,
"loss": 0.054,
"step": 249
},
{
"epoch": 1.67,
"learning_rate": 2e-05,
"loss": 0.0618,
"step": 250
},
{
"epoch": 1.68,
"learning_rate": 2e-05,
"loss": 0.0656,
"step": 251
},
{
"epoch": 1.68,
"learning_rate": 2e-05,
"loss": 0.0525,
"step": 252
},
{
"epoch": 1.69,
"learning_rate": 2e-05,
"loss": 0.0568,
"step": 253
},
{
"epoch": 1.7,
"learning_rate": 2e-05,
"loss": 0.0539,
"step": 254
},
{
"epoch": 1.7,
"learning_rate": 2e-05,
"loss": 0.0504,
"step": 255
},
{
"epoch": 1.71,
"learning_rate": 2e-05,
"loss": 0.0465,
"step": 256
},
{
"epoch": 1.72,
"learning_rate": 2e-05,
"loss": 0.0657,
"step": 257
},
{
"epoch": 1.73,
"learning_rate": 2e-05,
"loss": 0.065,
"step": 258
},
{
"epoch": 1.73,
"learning_rate": 2e-05,
"loss": 0.0509,
"step": 259
},
{
"epoch": 1.74,
"learning_rate": 2e-05,
"loss": 0.0607,
"step": 260
},
{
"epoch": 1.75,
"learning_rate": 2e-05,
"loss": 0.0631,
"step": 261
},
{
"epoch": 1.75,
"learning_rate": 2e-05,
"loss": 0.0624,
"step": 262
},
{
"epoch": 1.76,
"learning_rate": 2e-05,
"loss": 0.0531,
"step": 263
},
{
"epoch": 1.77,
"learning_rate": 2e-05,
"loss": 0.0642,
"step": 264
},
{
"epoch": 1.77,
"learning_rate": 2e-05,
"loss": 0.062,
"step": 265
},
{
"epoch": 1.78,
"learning_rate": 2e-05,
"loss": 0.0557,
"step": 266
},
{
"epoch": 1.79,
"learning_rate": 2e-05,
"loss": 0.0639,
"step": 267
},
{
"epoch": 1.79,
"learning_rate": 2e-05,
"loss": 0.0536,
"step": 268
},
{
"epoch": 1.8,
"learning_rate": 2e-05,
"loss": 0.0519,
"step": 269
},
{
"epoch": 1.81,
"learning_rate": 2e-05,
"loss": 0.0598,
"step": 270
},
{
"epoch": 1.81,
"learning_rate": 2e-05,
"loss": 0.0571,
"step": 271
},
{
"epoch": 1.82,
"learning_rate": 2e-05,
"loss": 0.0587,
"step": 272
},
{
"epoch": 1.83,
"learning_rate": 2e-05,
"loss": 0.0604,
"step": 273
},
{
"epoch": 1.83,
"learning_rate": 2e-05,
"loss": 0.056,
"step": 274
},
{
"epoch": 1.84,
"learning_rate": 2e-05,
"loss": 0.0597,
"step": 275
},
{
"epoch": 1.85,
"learning_rate": 2e-05,
"loss": 0.0531,
"step": 276
},
{
"epoch": 1.85,
"learning_rate": 2e-05,
"loss": 0.0645,
"step": 277
},
{
"epoch": 1.86,
"learning_rate": 2e-05,
"loss": 0.0573,
"step": 278
},
{
"epoch": 1.87,
"learning_rate": 2e-05,
"loss": 0.0567,
"step": 279
},
{
"epoch": 1.87,
"learning_rate": 2e-05,
"loss": 0.0584,
"step": 280
},
{
"epoch": 1.88,
"learning_rate": 2e-05,
"loss": 0.0433,
"step": 281
},
{
"epoch": 1.89,
"learning_rate": 2e-05,
"loss": 0.0571,
"step": 282
},
{
"epoch": 1.89,
"learning_rate": 2e-05,
"loss": 0.0592,
"step": 283
},
{
"epoch": 1.9,
"learning_rate": 2e-05,
"loss": 0.0514,
"step": 284
},
{
"epoch": 1.91,
"learning_rate": 2e-05,
"loss": 0.0565,
"step": 285
},
{
"epoch": 1.91,
"learning_rate": 2e-05,
"loss": 0.0564,
"step": 286
},
{
"epoch": 1.92,
"learning_rate": 2e-05,
"loss": 0.0581,
"step": 287
},
{
"epoch": 1.93,
"learning_rate": 2e-05,
"loss": 0.0607,
"step": 288
},
{
"epoch": 1.93,
"learning_rate": 2e-05,
"loss": 0.0572,
"step": 289
},
{
"epoch": 1.94,
"learning_rate": 2e-05,
"loss": 0.0526,
"step": 290
},
{
"epoch": 1.95,
"learning_rate": 2e-05,
"loss": 0.0649,
"step": 291
},
{
"epoch": 1.95,
"learning_rate": 2e-05,
"loss": 0.05,
"step": 292
},
{
"epoch": 1.96,
"learning_rate": 2e-05,
"loss": 0.0655,
"step": 293
},
{
"epoch": 1.97,
"learning_rate": 2e-05,
"loss": 0.0575,
"step": 294
},
{
"epoch": 1.97,
"learning_rate": 2e-05,
"loss": 0.0581,
"step": 295
},
{
"epoch": 1.98,
"learning_rate": 2e-05,
"loss": 0.0574,
"step": 296
},
{
"epoch": 1.99,
"learning_rate": 2e-05,
"loss": 0.0592,
"step": 297
},
{
"epoch": 1.99,
"learning_rate": 2e-05,
"loss": 0.0559,
"step": 298
},
{
"epoch": 2.0,
"learning_rate": 2e-05,
"loss": 0.0577,
"step": 299
},
{
"epoch": 2.01,
"learning_rate": 2e-05,
"loss": 0.0402,
"step": 300
},
{
"epoch": 2.01,
"learning_rate": 2e-05,
"loss": 0.0358,
"step": 301
},
{
"epoch": 2.02,
"learning_rate": 2e-05,
"loss": 0.0363,
"step": 302
},
{
"epoch": 2.03,
"learning_rate": 2e-05,
"loss": 0.0515,
"step": 303
},
{
"epoch": 2.03,
"learning_rate": 2e-05,
"loss": 0.0422,
"step": 304
},
{
"epoch": 2.04,
"learning_rate": 2e-05,
"loss": 0.0376,
"step": 305
},
{
"epoch": 2.05,
"learning_rate": 2e-05,
"loss": 0.0331,
"step": 306
},
{
"epoch": 2.05,
"learning_rate": 2e-05,
"loss": 0.0348,
"step": 307
},
{
"epoch": 2.06,
"learning_rate": 2e-05,
"loss": 0.0389,
"step": 308
},
{
"epoch": 2.07,
"learning_rate": 2e-05,
"loss": 0.0415,
"step": 309
},
{
"epoch": 2.07,
"learning_rate": 2e-05,
"loss": 0.0402,
"step": 310
},
{
"epoch": 2.08,
"learning_rate": 2e-05,
"loss": 0.0431,
"step": 311
},
{
"epoch": 2.09,
"learning_rate": 2e-05,
"loss": 0.0365,
"step": 312
},
{
"epoch": 2.09,
"learning_rate": 2e-05,
"loss": 0.0402,
"step": 313
},
{
"epoch": 2.1,
"learning_rate": 2e-05,
"loss": 0.0388,
"step": 314
},
{
"epoch": 2.11,
"learning_rate": 2e-05,
"loss": 0.0393,
"step": 315
},
{
"epoch": 2.11,
"learning_rate": 2e-05,
"loss": 0.0368,
"step": 316
},
{
"epoch": 2.12,
"learning_rate": 2e-05,
"loss": 0.0427,
"step": 317
},
{
"epoch": 2.13,
"learning_rate": 2e-05,
"loss": 0.0411,
"step": 318
},
{
"epoch": 2.13,
"learning_rate": 2e-05,
"loss": 0.0315,
"step": 319
},
{
"epoch": 2.14,
"learning_rate": 2e-05,
"loss": 0.0322,
"step": 320
},
{
"epoch": 2.15,
"learning_rate": 2e-05,
"loss": 0.0413,
"step": 321
},
{
"epoch": 2.15,
"learning_rate": 2e-05,
"loss": 0.0392,
"step": 322
},
{
"epoch": 2.16,
"learning_rate": 2e-05,
"loss": 0.0385,
"step": 323
},
{
"epoch": 2.17,
"learning_rate": 2e-05,
"loss": 0.0448,
"step": 324
},
{
"epoch": 2.17,
"learning_rate": 2e-05,
"loss": 0.044,
"step": 325
},
{
"epoch": 2.18,
"learning_rate": 2e-05,
"loss": 0.0371,
"step": 326
},
{
"epoch": 2.19,
"learning_rate": 2e-05,
"loss": 0.0455,
"step": 327
},
{
"epoch": 2.19,
"learning_rate": 2e-05,
"loss": 0.0379,
"step": 328
},
{
"epoch": 2.2,
"learning_rate": 2e-05,
"loss": 0.0405,
"step": 329
},
{
"epoch": 2.21,
"learning_rate": 2e-05,
"loss": 0.0338,
"step": 330
},
{
"epoch": 2.21,
"learning_rate": 2e-05,
"loss": 0.0378,
"step": 331
},
{
"epoch": 2.22,
"learning_rate": 2e-05,
"loss": 0.0377,
"step": 332
},
{
"epoch": 2.23,
"learning_rate": 2e-05,
"loss": 0.0422,
"step": 333
},
{
"epoch": 2.23,
"learning_rate": 2e-05,
"loss": 0.0425,
"step": 334
},
{
"epoch": 2.24,
"learning_rate": 2e-05,
"loss": 0.0372,
"step": 335
},
{
"epoch": 2.25,
"learning_rate": 2e-05,
"loss": 0.0391,
"step": 336
},
{
"epoch": 2.25,
"learning_rate": 2e-05,
"loss": 0.0387,
"step": 337
},
{
"epoch": 2.26,
"learning_rate": 2e-05,
"loss": 0.0455,
"step": 338
},
{
"epoch": 2.27,
"learning_rate": 2e-05,
"loss": 0.0333,
"step": 339
},
{
"epoch": 2.27,
"learning_rate": 2e-05,
"loss": 0.0381,
"step": 340
},
{
"epoch": 2.28,
"learning_rate": 2e-05,
"loss": 0.043,
"step": 341
},
{
"epoch": 2.29,
"learning_rate": 2e-05,
"loss": 0.0378,
"step": 342
},
{
"epoch": 2.29,
"learning_rate": 2e-05,
"loss": 0.0384,
"step": 343
},
{
"epoch": 2.3,
"learning_rate": 2e-05,
"loss": 0.037,
"step": 344
},
{
"epoch": 2.31,
"learning_rate": 2e-05,
"loss": 0.039,
"step": 345
},
{
"epoch": 2.31,
"learning_rate": 2e-05,
"loss": 0.0354,
"step": 346
},
{
"epoch": 2.32,
"learning_rate": 2e-05,
"loss": 0.0416,
"step": 347
},
{
"epoch": 2.33,
"learning_rate": 2e-05,
"loss": 0.0411,
"step": 348
},
{
"epoch": 2.33,
"learning_rate": 2e-05,
"loss": 0.0414,
"step": 349
},
{
"epoch": 2.34,
"learning_rate": 2e-05,
"loss": 0.0358,
"step": 350
},
{
"epoch": 2.35,
"learning_rate": 2e-05,
"loss": 0.0378,
"step": 351
},
{
"epoch": 2.35,
"learning_rate": 2e-05,
"loss": 0.0524,
"step": 352
},
{
"epoch": 2.36,
"learning_rate": 2e-05,
"loss": 0.039,
"step": 353
},
{
"epoch": 2.37,
"learning_rate": 2e-05,
"loss": 0.0415,
"step": 354
},
{
"epoch": 2.37,
"learning_rate": 2e-05,
"loss": 0.0381,
"step": 355
},
{
"epoch": 2.38,
"learning_rate": 2e-05,
"loss": 0.0401,
"step": 356
},
{
"epoch": 2.39,
"learning_rate": 2e-05,
"loss": 0.0368,
"step": 357
},
{
"epoch": 2.39,
"learning_rate": 2e-05,
"loss": 0.0412,
"step": 358
},
{
"epoch": 2.4,
"learning_rate": 2e-05,
"loss": 0.0405,
"step": 359
},
{
"epoch": 2.41,
"learning_rate": 2e-05,
"loss": 0.0382,
"step": 360
},
{
"epoch": 2.41,
"learning_rate": 2e-05,
"loss": 0.0488,
"step": 361
},
{
"epoch": 2.42,
"learning_rate": 2e-05,
"loss": 0.0457,
"step": 362
},
{
"epoch": 2.43,
"learning_rate": 2e-05,
"loss": 0.0381,
"step": 363
},
{
"epoch": 2.43,
"learning_rate": 2e-05,
"loss": 0.0479,
"step": 364
},
{
"epoch": 2.44,
"learning_rate": 2e-05,
"loss": 0.0374,
"step": 365
},
{
"epoch": 2.45,
"learning_rate": 2e-05,
"loss": 0.0486,
"step": 366
},
{
"epoch": 2.45,
"learning_rate": 2e-05,
"loss": 0.0453,
"step": 367
},
{
"epoch": 2.46,
"learning_rate": 2e-05,
"loss": 0.0456,
"step": 368
},
{
"epoch": 2.47,
"learning_rate": 2e-05,
"loss": 0.0366,
"step": 369
},
{
"epoch": 2.47,
"learning_rate": 2e-05,
"loss": 0.0342,
"step": 370
},
{
"epoch": 2.48,
"learning_rate": 2e-05,
"loss": 0.0338,
"step": 371
},
{
"epoch": 2.49,
"learning_rate": 2e-05,
"loss": 0.0337,
"step": 372
},
{
"epoch": 2.49,
"learning_rate": 2e-05,
"loss": 0.0367,
"step": 373
},
{
"epoch": 2.5,
"learning_rate": 2e-05,
"loss": 0.038,
"step": 374
},
{
"epoch": 2.51,
"learning_rate": 2e-05,
"loss": 0.0419,
"step": 375
},
{
"epoch": 2.51,
"learning_rate": 2e-05,
"loss": 0.0322,
"step": 376
},
{
"epoch": 2.52,
"learning_rate": 2e-05,
"loss": 0.0325,
"step": 377
},
{
"epoch": 2.53,
"learning_rate": 2e-05,
"loss": 0.0365,
"step": 378
},
{
"epoch": 2.53,
"learning_rate": 2e-05,
"loss": 0.0384,
"step": 379
},
{
"epoch": 2.54,
"learning_rate": 2e-05,
"loss": 0.0443,
"step": 380
},
{
"epoch": 2.55,
"learning_rate": 2e-05,
"loss": 0.0371,
"step": 381
},
{
"epoch": 2.55,
"learning_rate": 2e-05,
"loss": 0.0499,
"step": 382
},
{
"epoch": 2.56,
"learning_rate": 2e-05,
"loss": 0.0464,
"step": 383
},
{
"epoch": 2.57,
"learning_rate": 2e-05,
"loss": 0.0427,
"step": 384
},
{
"epoch": 2.57,
"learning_rate": 2e-05,
"loss": 0.0337,
"step": 385
},
{
"epoch": 2.58,
"learning_rate": 2e-05,
"loss": 0.0396,
"step": 386
},
{
"epoch": 2.59,
"learning_rate": 2e-05,
"loss": 0.0442,
"step": 387
},
{
"epoch": 2.59,
"learning_rate": 2e-05,
"loss": 0.0391,
"step": 388
},
{
"epoch": 2.6,
"learning_rate": 2e-05,
"loss": 0.0457,
"step": 389
},
{
"epoch": 2.61,
"learning_rate": 2e-05,
"loss": 0.0393,
"step": 390
},
{
"epoch": 2.61,
"learning_rate": 2e-05,
"loss": 0.0328,
"step": 391
},
{
"epoch": 2.62,
"learning_rate": 2e-05,
"loss": 0.038,
"step": 392
},
{
"epoch": 2.63,
"learning_rate": 2e-05,
"loss": 0.0416,
"step": 393
},
{
"epoch": 2.63,
"learning_rate": 2e-05,
"loss": 0.0375,
"step": 394
},
{
"epoch": 2.64,
"learning_rate": 2e-05,
"loss": 0.0416,
"step": 395
},
{
"epoch": 2.65,
"learning_rate": 2e-05,
"loss": 0.0441,
"step": 396
},
{
"epoch": 2.65,
"learning_rate": 2e-05,
"loss": 0.0454,
"step": 397
},
{
"epoch": 2.66,
"learning_rate": 2e-05,
"loss": 0.0374,
"step": 398
},
{
"epoch": 2.67,
"learning_rate": 2e-05,
"loss": 0.0332,
"step": 399
},
{
"epoch": 2.67,
"learning_rate": 2e-05,
"loss": 0.0383,
"step": 400
},
{
"epoch": 2.68,
"learning_rate": 2e-05,
"loss": 0.0474,
"step": 401
},
{
"epoch": 2.69,
"learning_rate": 2e-05,
"loss": 0.0511,
"step": 402
},
{
"epoch": 2.69,
"learning_rate": 2e-05,
"loss": 0.0369,
"step": 403
},
{
"epoch": 2.7,
"learning_rate": 2e-05,
"loss": 0.0382,
"step": 404
},
{
"epoch": 2.71,
"learning_rate": 2e-05,
"loss": 0.0368,
"step": 405
},
{
"epoch": 2.71,
"learning_rate": 2e-05,
"loss": 0.0521,
"step": 406
},
{
"epoch": 2.72,
"learning_rate": 2e-05,
"loss": 0.0392,
"step": 407
},
{
"epoch": 2.73,
"learning_rate": 2e-05,
"loss": 0.0523,
"step": 408
},
{
"epoch": 2.73,
"learning_rate": 2e-05,
"loss": 0.0336,
"step": 409
},
{
"epoch": 2.74,
"learning_rate": 2e-05,
"loss": 0.0365,
"step": 410
},
{
"epoch": 2.75,
"learning_rate": 2e-05,
"loss": 0.0386,
"step": 411
},
{
"epoch": 2.75,
"learning_rate": 2e-05,
"loss": 0.047,
"step": 412
},
{
"epoch": 2.76,
"learning_rate": 2e-05,
"loss": 0.0438,
"step": 413
},
{
"epoch": 2.77,
"learning_rate": 2e-05,
"loss": 0.0409,
"step": 414
},
{
"epoch": 2.77,
"learning_rate": 2e-05,
"loss": 0.0494,
"step": 415
},
{
"epoch": 2.78,
"learning_rate": 2e-05,
"loss": 0.0496,
"step": 416
},
{
"epoch": 2.79,
"learning_rate": 2e-05,
"loss": 0.036,
"step": 417
},
{
"epoch": 2.79,
"learning_rate": 2e-05,
"loss": 0.0483,
"step": 418
},
{
"epoch": 2.8,
"learning_rate": 2e-05,
"loss": 0.0439,
"step": 419
},
{
"epoch": 2.81,
"learning_rate": 2e-05,
"loss": 0.0368,
"step": 420
},
{
"epoch": 2.81,
"learning_rate": 2e-05,
"loss": 0.0376,
"step": 421
},
{
"epoch": 2.82,
"learning_rate": 2e-05,
"loss": 0.0457,
"step": 422
},
{
"epoch": 2.83,
"learning_rate": 2e-05,
"loss": 0.0379,
"step": 423
},
{
"epoch": 2.83,
"learning_rate": 2e-05,
"loss": 0.0365,
"step": 424
},
{
"epoch": 2.84,
"learning_rate": 2e-05,
"loss": 0.0328,
"step": 425
},
{
"epoch": 2.85,
"learning_rate": 2e-05,
"loss": 0.0415,
"step": 426
},
{
"epoch": 2.85,
"learning_rate": 2e-05,
"loss": 0.0413,
"step": 427
},
{
"epoch": 2.86,
"learning_rate": 2e-05,
"loss": 0.0376,
"step": 428
},
{
"epoch": 2.87,
"learning_rate": 2e-05,
"loss": 0.0321,
"step": 429
},
{
"epoch": 2.88,
"learning_rate": 2e-05,
"loss": 0.0397,
"step": 430
},
{
"epoch": 2.88,
"learning_rate": 2e-05,
"loss": 0.0386,
"step": 431
},
{
"epoch": 2.89,
"learning_rate": 2e-05,
"loss": 0.0444,
"step": 432
},
{
"epoch": 2.9,
"learning_rate": 2e-05,
"loss": 0.0333,
"step": 433
},
{
"epoch": 2.9,
"learning_rate": 2e-05,
"loss": 0.0478,
"step": 434
},
{
"epoch": 2.91,
"learning_rate": 2e-05,
"loss": 0.0361,
"step": 435
},
{
"epoch": 2.92,
"learning_rate": 2e-05,
"loss": 0.0317,
"step": 436
},
{
"epoch": 2.92,
"learning_rate": 2e-05,
"loss": 0.0534,
"step": 437
},
{
"epoch": 2.93,
"learning_rate": 2e-05,
"loss": 0.0408,
"step": 438
},
{
"epoch": 2.94,
"learning_rate": 2e-05,
"loss": 0.0447,
"step": 439
},
{
"epoch": 2.94,
"learning_rate": 2e-05,
"loss": 0.0426,
"step": 440
},
{
"epoch": 2.95,
"learning_rate": 2e-05,
"loss": 0.0427,
"step": 441
},
{
"epoch": 2.96,
"learning_rate": 2e-05,
"loss": 0.0402,
"step": 442
},
{
"epoch": 2.96,
"learning_rate": 2e-05,
"loss": 0.0413,
"step": 443
},
{
"epoch": 2.97,
"learning_rate": 2e-05,
"loss": 0.0451,
"step": 444
},
{
"epoch": 2.98,
"learning_rate": 2e-05,
"loss": 0.0356,
"step": 445
},
{
"epoch": 2.98,
"learning_rate": 2e-05,
"loss": 0.0368,
"step": 446
},
{
"epoch": 2.99,
"learning_rate": 2e-05,
"loss": 0.0445,
"step": 447
}
],
"max_steps": 447,
"num_train_epochs": 3,
"total_flos": 65798170214400.0,
"trial_name": null,
"trial_params": null
}