chitanda's picture
Add files using upload-large-folder tool
ecd4bfe verified
raw
history blame
47.1 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.7692307692307693,
"eval_steps": 500,
"global_step": 400,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 6.25e-07,
"loss": 3.7473,
"step": 1
},
{
"epoch": 0.0,
"learning_rate": 1.25e-06,
"loss": 0.0,
"step": 2
},
{
"epoch": 0.01,
"learning_rate": 1.8750000000000003e-06,
"loss": 0.0,
"step": 3
},
{
"epoch": 0.01,
"learning_rate": 2.5e-06,
"loss": 0.0,
"step": 4
},
{
"epoch": 0.01,
"learning_rate": 3.125e-06,
"loss": 0.0,
"step": 5
},
{
"epoch": 0.01,
"learning_rate": 3.7500000000000005e-06,
"loss": 0.0,
"step": 6
},
{
"epoch": 0.01,
"learning_rate": 4.3750000000000005e-06,
"loss": 0.0,
"step": 7
},
{
"epoch": 0.02,
"learning_rate": 5e-06,
"loss": 0.0,
"step": 8
},
{
"epoch": 0.02,
"learning_rate": 5.625e-06,
"loss": 0.0,
"step": 9
},
{
"epoch": 0.02,
"learning_rate": 6.25e-06,
"loss": 0.0,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 6.875e-06,
"loss": 0.0,
"step": 11
},
{
"epoch": 0.02,
"learning_rate": 7.500000000000001e-06,
"loss": 0.0,
"step": 12
},
{
"epoch": 0.03,
"learning_rate": 8.125000000000001e-06,
"loss": 0.0,
"step": 13
},
{
"epoch": 0.03,
"learning_rate": 8.750000000000001e-06,
"loss": 0.0,
"step": 14
},
{
"epoch": 0.03,
"learning_rate": 9.375000000000001e-06,
"loss": 0.0,
"step": 15
},
{
"epoch": 0.03,
"learning_rate": 1e-05,
"loss": 0.0,
"step": 16
},
{
"epoch": 0.03,
"learning_rate": 1.0625e-05,
"loss": 0.0,
"step": 17
},
{
"epoch": 0.03,
"learning_rate": 1.125e-05,
"loss": 0.0,
"step": 18
},
{
"epoch": 0.04,
"learning_rate": 1.1875e-05,
"loss": 0.0,
"step": 19
},
{
"epoch": 0.04,
"learning_rate": 1.25e-05,
"loss": 0.0,
"step": 20
},
{
"epoch": 0.04,
"learning_rate": 1.3125e-05,
"loss": 0.0,
"step": 21
},
{
"epoch": 0.04,
"learning_rate": 1.375e-05,
"loss": 0.0,
"step": 22
},
{
"epoch": 0.04,
"learning_rate": 1.4375e-05,
"loss": 0.0,
"step": 23
},
{
"epoch": 0.05,
"learning_rate": 1.5000000000000002e-05,
"loss": 0.0,
"step": 24
},
{
"epoch": 0.05,
"learning_rate": 1.5625e-05,
"loss": 0.0,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 1.6250000000000002e-05,
"loss": 0.0,
"step": 26
},
{
"epoch": 0.05,
"learning_rate": 1.6875e-05,
"loss": 0.0,
"step": 27
},
{
"epoch": 0.05,
"learning_rate": 1.7500000000000002e-05,
"loss": 0.0,
"step": 28
},
{
"epoch": 0.06,
"learning_rate": 1.8125e-05,
"loss": 0.0,
"step": 29
},
{
"epoch": 0.06,
"learning_rate": 1.8750000000000002e-05,
"loss": 0.0,
"step": 30
},
{
"epoch": 0.06,
"learning_rate": 1.9375e-05,
"loss": 0.0,
"step": 31
},
{
"epoch": 0.06,
"learning_rate": 2e-05,
"loss": 0.0,
"step": 32
},
{
"epoch": 0.06,
"learning_rate": 1.9999951432210905e-05,
"loss": 0.0,
"step": 33
},
{
"epoch": 0.07,
"learning_rate": 1.9999805729315383e-05,
"loss": 0.0,
"step": 34
},
{
"epoch": 0.07,
"learning_rate": 1.999956289272873e-05,
"loss": 0.0,
"step": 35
},
{
"epoch": 0.07,
"learning_rate": 1.999922292480975e-05,
"loss": 0.0,
"step": 36
},
{
"epoch": 0.07,
"learning_rate": 1.9998785828860744e-05,
"loss": 0.0,
"step": 37
},
{
"epoch": 0.07,
"learning_rate": 1.9998251609127465e-05,
"loss": 0.0,
"step": 38
},
{
"epoch": 0.07,
"learning_rate": 1.999762027079909e-05,
"loss": 0.0,
"step": 39
},
{
"epoch": 0.08,
"learning_rate": 1.9996891820008165e-05,
"loss": 0.0,
"step": 40
},
{
"epoch": 0.08,
"learning_rate": 1.9996066263830533e-05,
"loss": 0.0,
"step": 41
},
{
"epoch": 0.08,
"learning_rate": 1.9995143610285275e-05,
"loss": 0.0,
"step": 42
},
{
"epoch": 0.08,
"learning_rate": 1.9994123868334655e-05,
"loss": 0.0,
"step": 43
},
{
"epoch": 0.08,
"learning_rate": 1.9993007047883988e-05,
"loss": 0.0,
"step": 44
},
{
"epoch": 0.09,
"learning_rate": 1.999179315978157e-05,
"loss": 0.0,
"step": 45
},
{
"epoch": 0.09,
"learning_rate": 1.999048221581858e-05,
"loss": 0.0,
"step": 46
},
{
"epoch": 0.09,
"learning_rate": 1.9989074228728942e-05,
"loss": 0.0,
"step": 47
},
{
"epoch": 0.09,
"learning_rate": 1.9987569212189224e-05,
"loss": 0.0,
"step": 48
},
{
"epoch": 0.09,
"learning_rate": 1.9985967180818493e-05,
"loss": 0.0,
"step": 49
},
{
"epoch": 0.1,
"learning_rate": 1.998426815017817e-05,
"loss": 0.0,
"step": 50
},
{
"epoch": 0.1,
"learning_rate": 1.998247213677188e-05,
"loss": 0.0,
"step": 51
},
{
"epoch": 0.1,
"learning_rate": 1.9980579158045322e-05,
"loss": 0.0,
"step": 52
},
{
"epoch": 0.1,
"learning_rate": 1.9978589232386036e-05,
"loss": 0.0,
"step": 53
},
{
"epoch": 0.1,
"learning_rate": 1.997650237912329e-05,
"loss": 0.0,
"step": 54
},
{
"epoch": 0.11,
"learning_rate": 1.997431861852785e-05,
"loss": 0.0,
"step": 55
},
{
"epoch": 0.11,
"learning_rate": 1.9972037971811802e-05,
"loss": 0.0,
"step": 56
},
{
"epoch": 0.11,
"learning_rate": 1.996966046112834e-05,
"loss": 0.0,
"step": 57
},
{
"epoch": 0.11,
"learning_rate": 1.996718610957155e-05,
"loss": 0.0,
"step": 58
},
{
"epoch": 0.11,
"learning_rate": 1.9964614941176194e-05,
"loss": 0.0,
"step": 59
},
{
"epoch": 0.12,
"learning_rate": 1.9961946980917457e-05,
"loss": 0.0,
"step": 60
},
{
"epoch": 0.12,
"learning_rate": 1.995918225471073e-05,
"loss": 0.0,
"step": 61
},
{
"epoch": 0.12,
"learning_rate": 1.9956320789411338e-05,
"loss": 0.0,
"step": 62
},
{
"epoch": 0.12,
"learning_rate": 1.9953362612814294e-05,
"loss": 0.0,
"step": 63
},
{
"epoch": 0.12,
"learning_rate": 1.9950307753654016e-05,
"loss": 0.0,
"step": 64
},
{
"epoch": 0.12,
"learning_rate": 1.994715624160405e-05,
"loss": 0.0,
"step": 65
},
{
"epoch": 0.13,
"learning_rate": 1.99439081072768e-05,
"loss": 0.0,
"step": 66
},
{
"epoch": 0.13,
"learning_rate": 1.9940563382223196e-05,
"loss": 0.0,
"step": 67
},
{
"epoch": 0.13,
"learning_rate": 1.9937122098932428e-05,
"loss": 0.0,
"step": 68
},
{
"epoch": 0.13,
"learning_rate": 1.9933584290831593e-05,
"loss": 0.0,
"step": 69
},
{
"epoch": 0.13,
"learning_rate": 1.9929949992285397e-05,
"loss": 0.0,
"step": 70
},
{
"epoch": 0.14,
"learning_rate": 1.992621923859581e-05,
"loss": 0.0,
"step": 71
},
{
"epoch": 0.14,
"learning_rate": 1.9922392066001724e-05,
"loss": 0.0,
"step": 72
},
{
"epoch": 0.14,
"learning_rate": 1.99184685116786e-05,
"loss": 0.0,
"step": 73
},
{
"epoch": 0.14,
"learning_rate": 1.9914448613738107e-05,
"loss": 0.0,
"step": 74
},
{
"epoch": 0.14,
"learning_rate": 1.991033241122776e-05,
"loss": 0.0,
"step": 75
},
{
"epoch": 0.15,
"learning_rate": 1.9906119944130527e-05,
"loss": 0.0,
"step": 76
},
{
"epoch": 0.15,
"learning_rate": 1.9901811253364458e-05,
"loss": 0.0,
"step": 77
},
{
"epoch": 0.15,
"learning_rate": 1.9897406380782262e-05,
"loss": 0.0,
"step": 78
},
{
"epoch": 0.15,
"learning_rate": 1.989290536917093e-05,
"loss": 0.0,
"step": 79
},
{
"epoch": 0.15,
"learning_rate": 1.9888308262251286e-05,
"loss": 0.0,
"step": 80
},
{
"epoch": 0.16,
"learning_rate": 1.988361510467761e-05,
"loss": 0.0,
"step": 81
},
{
"epoch": 0.16,
"learning_rate": 1.9878825942037147e-05,
"loss": 0.0,
"step": 82
},
{
"epoch": 0.16,
"learning_rate": 1.9873940820849714e-05,
"loss": 0.0,
"step": 83
},
{
"epoch": 0.16,
"learning_rate": 1.9868959788567213e-05,
"loss": 0.0,
"step": 84
},
{
"epoch": 0.16,
"learning_rate": 1.9863882893573188e-05,
"loss": 0.0,
"step": 85
},
{
"epoch": 0.17,
"learning_rate": 1.985871018518236e-05,
"loss": 0.0,
"step": 86
},
{
"epoch": 0.17,
"learning_rate": 1.9853441713640123e-05,
"loss": 0.0,
"step": 87
},
{
"epoch": 0.17,
"learning_rate": 1.9848077530122083e-05,
"loss": 0.0,
"step": 88
},
{
"epoch": 0.17,
"learning_rate": 1.9842617686733546e-05,
"loss": 0.0,
"step": 89
},
{
"epoch": 0.17,
"learning_rate": 1.9837062236509013e-05,
"loss": 0.0,
"step": 90
},
{
"epoch": 0.17,
"learning_rate": 1.983141123341168e-05,
"loss": 0.0,
"step": 91
},
{
"epoch": 0.18,
"learning_rate": 1.9825664732332886e-05,
"loss": 0.0,
"step": 92
},
{
"epoch": 0.18,
"learning_rate": 1.9819822789091597e-05,
"loss": 0.0,
"step": 93
},
{
"epoch": 0.18,
"learning_rate": 1.981388546043388e-05,
"loss": 0.0,
"step": 94
},
{
"epoch": 0.18,
"learning_rate": 1.9807852804032306e-05,
"loss": 0.0,
"step": 95
},
{
"epoch": 0.18,
"learning_rate": 1.9801724878485438e-05,
"loss": 0.0,
"step": 96
},
{
"epoch": 0.19,
"learning_rate": 1.979550174331724e-05,
"loss": 0.0,
"step": 97
},
{
"epoch": 0.19,
"learning_rate": 1.9789183458976485e-05,
"loss": 0.0,
"step": 98
},
{
"epoch": 0.19,
"learning_rate": 1.97827700868362e-05,
"loss": 0.0,
"step": 99
},
{
"epoch": 0.19,
"learning_rate": 1.977626168919305e-05,
"loss": 0.0,
"step": 100
},
{
"epoch": 0.19,
"learning_rate": 1.9769658329266718e-05,
"loss": 0.0,
"step": 101
},
{
"epoch": 0.2,
"learning_rate": 1.9762960071199334e-05,
"loss": 0.0,
"step": 102
},
{
"epoch": 0.2,
"learning_rate": 1.9756166980054812e-05,
"loss": 0.0,
"step": 103
},
{
"epoch": 0.2,
"learning_rate": 1.9749279121818235e-05,
"loss": 0.0,
"step": 104
},
{
"epoch": 0.2,
"learning_rate": 1.9742296563395218e-05,
"loss": 0.0,
"step": 105
},
{
"epoch": 0.2,
"learning_rate": 1.9735219372611232e-05,
"loss": 0.0,
"step": 106
},
{
"epoch": 0.21,
"learning_rate": 1.9728047618210995e-05,
"loss": 0.0,
"step": 107
},
{
"epoch": 0.21,
"learning_rate": 1.9720781369857747e-05,
"loss": 0.0,
"step": 108
},
{
"epoch": 0.21,
"learning_rate": 1.9713420698132614e-05,
"loss": 0.0,
"step": 109
},
{
"epoch": 0.21,
"learning_rate": 1.970596567453391e-05,
"loss": 0.0,
"step": 110
},
{
"epoch": 0.21,
"learning_rate": 1.9698416371476434e-05,
"loss": 0.0,
"step": 111
},
{
"epoch": 0.22,
"learning_rate": 1.969077286229078e-05,
"loss": 0.0,
"step": 112
},
{
"epoch": 0.22,
"learning_rate": 1.9683035221222617e-05,
"loss": 0.0,
"step": 113
},
{
"epoch": 0.22,
"learning_rate": 1.9675203523431964e-05,
"loss": 0.0,
"step": 114
},
{
"epoch": 0.22,
"learning_rate": 1.9667277844992476e-05,
"loss": 0.0,
"step": 115
},
{
"epoch": 0.22,
"learning_rate": 1.9659258262890683e-05,
"loss": 0.0,
"step": 116
},
{
"epoch": 0.23,
"learning_rate": 1.9651144855025265e-05,
"loss": 0.0,
"step": 117
},
{
"epoch": 0.23,
"learning_rate": 1.964293770020628e-05,
"loss": 0.0,
"step": 118
},
{
"epoch": 0.23,
"learning_rate": 1.9634636878154393e-05,
"loss": 0.0,
"step": 119
},
{
"epoch": 0.23,
"learning_rate": 1.962624246950012e-05,
"loss": 0.0,
"step": 120
},
{
"epoch": 0.23,
"learning_rate": 1.9617754555783045e-05,
"loss": 0.0,
"step": 121
},
{
"epoch": 0.23,
"learning_rate": 1.9609173219450998e-05,
"loss": 0.0,
"step": 122
},
{
"epoch": 0.24,
"learning_rate": 1.960049854385929e-05,
"loss": 0.0,
"step": 123
},
{
"epoch": 0.24,
"learning_rate": 1.9591730613269878e-05,
"loss": 0.0,
"step": 124
},
{
"epoch": 0.24,
"learning_rate": 1.9582869512850576e-05,
"loss": 0.0,
"step": 125
},
{
"epoch": 0.24,
"learning_rate": 1.957391532867418e-05,
"loss": 0.0,
"step": 126
},
{
"epoch": 0.24,
"learning_rate": 1.956486814771769e-05,
"loss": 0.0,
"step": 127
},
{
"epoch": 0.25,
"learning_rate": 1.955572805786141e-05,
"loss": 0.0,
"step": 128
},
{
"epoch": 0.25,
"learning_rate": 1.9546495147888134e-05,
"loss": 0.0,
"step": 129
},
{
"epoch": 0.25,
"learning_rate": 1.953716950748227e-05,
"loss": 0.0,
"step": 130
},
{
"epoch": 0.25,
"learning_rate": 1.9527751227228964e-05,
"loss": 0.0,
"step": 131
},
{
"epoch": 0.25,
"learning_rate": 1.9518240398613226e-05,
"loss": 0.0,
"step": 132
},
{
"epoch": 0.26,
"learning_rate": 1.9508637114019037e-05,
"loss": 0.0,
"step": 133
},
{
"epoch": 0.26,
"learning_rate": 1.9498941466728462e-05,
"loss": 0.0,
"step": 134
},
{
"epoch": 0.26,
"learning_rate": 1.9489153550920726e-05,
"loss": 0.0,
"step": 135
},
{
"epoch": 0.26,
"learning_rate": 1.947927346167132e-05,
"loss": 0.0,
"step": 136
},
{
"epoch": 0.26,
"learning_rate": 1.946930129495106e-05,
"loss": 0.0,
"step": 137
},
{
"epoch": 0.27,
"learning_rate": 1.945923714762516e-05,
"loss": 0.0,
"step": 138
},
{
"epoch": 0.27,
"learning_rate": 1.9449081117452304e-05,
"loss": 0.0,
"step": 139
},
{
"epoch": 0.27,
"learning_rate": 1.9438833303083677e-05,
"loss": 0.0,
"step": 140
},
{
"epoch": 0.27,
"learning_rate": 1.9428493804062013e-05,
"loss": 0.0,
"step": 141
},
{
"epoch": 0.27,
"learning_rate": 1.9418062720820636e-05,
"loss": 0.0,
"step": 142
},
{
"epoch": 0.28,
"learning_rate": 1.9407540154682473e-05,
"loss": 0.0,
"step": 143
},
{
"epoch": 0.28,
"learning_rate": 1.9396926207859085e-05,
"loss": 0.0,
"step": 144
},
{
"epoch": 0.28,
"learning_rate": 1.9386220983449652e-05,
"loss": 0.0,
"step": 145
},
{
"epoch": 0.28,
"learning_rate": 1.9375424585439994e-05,
"loss": 0.0,
"step": 146
},
{
"epoch": 0.28,
"learning_rate": 1.9364537118701542e-05,
"loss": 0.0,
"step": 147
},
{
"epoch": 0.28,
"learning_rate": 1.935355868899034e-05,
"loss": 0.0,
"step": 148
},
{
"epoch": 0.29,
"learning_rate": 1.9342489402945997e-05,
"loss": 0.0,
"step": 149
},
{
"epoch": 0.29,
"learning_rate": 1.9331329368090664e-05,
"loss": 0.0,
"step": 150
},
{
"epoch": 0.29,
"learning_rate": 1.932007869282799e-05,
"loss": 0.0,
"step": 151
},
{
"epoch": 0.29,
"learning_rate": 1.9308737486442045e-05,
"loss": 0.0,
"step": 152
},
{
"epoch": 0.29,
"learning_rate": 1.9297305859096305e-05,
"loss": 0.0,
"step": 153
},
{
"epoch": 0.3,
"learning_rate": 1.9285783921832537e-05,
"loss": 0.0,
"step": 154
},
{
"epoch": 0.3,
"learning_rate": 1.927417178656975e-05,
"loss": 0.0,
"step": 155
},
{
"epoch": 0.3,
"learning_rate": 1.926246956610309e-05,
"loss": 0.0,
"step": 156
},
{
"epoch": 0.3,
"learning_rate": 1.9250677374102752e-05,
"loss": 0.0,
"step": 157
},
{
"epoch": 0.3,
"learning_rate": 1.9238795325112867e-05,
"loss": 0.0,
"step": 158
},
{
"epoch": 0.31,
"learning_rate": 1.9226823534550418e-05,
"loss": 0.0,
"step": 159
},
{
"epoch": 0.31,
"learning_rate": 1.921476211870408e-05,
"loss": 0.0,
"step": 160
},
{
"epoch": 0.31,
"learning_rate": 1.9202611194733107e-05,
"loss": 0.0,
"step": 161
},
{
"epoch": 0.31,
"learning_rate": 1.9190370880666206e-05,
"loss": 0.0,
"step": 162
},
{
"epoch": 0.31,
"learning_rate": 1.9178041295400383e-05,
"loss": 0.0,
"step": 163
},
{
"epoch": 0.32,
"learning_rate": 1.9165622558699763e-05,
"loss": 0.0,
"step": 164
},
{
"epoch": 0.32,
"learning_rate": 1.9153114791194475e-05,
"loss": 0.0,
"step": 165
},
{
"epoch": 0.32,
"learning_rate": 1.9140518114379433e-05,
"loss": 0.0,
"step": 166
},
{
"epoch": 0.32,
"learning_rate": 1.912783265061319e-05,
"loss": 0.0,
"step": 167
},
{
"epoch": 0.32,
"learning_rate": 1.9115058523116734e-05,
"loss": 0.0,
"step": 168
},
{
"epoch": 0.33,
"learning_rate": 1.9102195855972287e-05,
"loss": 0.0,
"step": 169
},
{
"epoch": 0.33,
"learning_rate": 1.908924477412211e-05,
"loss": 0.0,
"step": 170
},
{
"epoch": 0.33,
"learning_rate": 1.9076205403367287e-05,
"loss": 0.0,
"step": 171
},
{
"epoch": 0.33,
"learning_rate": 1.9063077870366504e-05,
"loss": 0.0,
"step": 172
},
{
"epoch": 0.33,
"learning_rate": 1.90498623026348e-05,
"loss": 0.0,
"step": 173
},
{
"epoch": 0.33,
"learning_rate": 1.903655882854237e-05,
"loss": 0.0,
"step": 174
},
{
"epoch": 0.34,
"learning_rate": 1.9023167577313267e-05,
"loss": 0.0,
"step": 175
},
{
"epoch": 0.34,
"learning_rate": 1.900968867902419e-05,
"loss": 0.0,
"step": 176
},
{
"epoch": 0.34,
"learning_rate": 1.8996122264603202e-05,
"loss": 0.0,
"step": 177
},
{
"epoch": 0.34,
"learning_rate": 1.898246846582844e-05,
"loss": 0.0,
"step": 178
},
{
"epoch": 0.34,
"learning_rate": 1.8968727415326885e-05,
"loss": 0.0,
"step": 179
},
{
"epoch": 0.35,
"learning_rate": 1.895489924657301e-05,
"loss": 0.0,
"step": 180
},
{
"epoch": 0.35,
"learning_rate": 1.894098409388754e-05,
"loss": 0.0,
"step": 181
},
{
"epoch": 0.35,
"learning_rate": 1.8926982092436117e-05,
"loss": 0.0,
"step": 182
},
{
"epoch": 0.35,
"learning_rate": 1.8912893378227984e-05,
"loss": 0.0,
"step": 183
},
{
"epoch": 0.35,
"learning_rate": 1.8898718088114688e-05,
"loss": 0.0,
"step": 184
},
{
"epoch": 0.36,
"learning_rate": 1.8884456359788725e-05,
"loss": 0.0,
"step": 185
},
{
"epoch": 0.36,
"learning_rate": 1.887010833178222e-05,
"loss": 0.0,
"step": 186
},
{
"epoch": 0.36,
"learning_rate": 1.8855674143465567e-05,
"loss": 0.0,
"step": 187
},
{
"epoch": 0.36,
"learning_rate": 1.8841153935046098e-05,
"loss": 0.0,
"step": 188
},
{
"epoch": 0.36,
"learning_rate": 1.8826547847566692e-05,
"loss": 0.0,
"step": 189
},
{
"epoch": 0.37,
"learning_rate": 1.8811856022904423e-05,
"loss": 0.0,
"step": 190
},
{
"epoch": 0.37,
"learning_rate": 1.8797078603769184e-05,
"loss": 0.0,
"step": 191
},
{
"epoch": 0.37,
"learning_rate": 1.8782215733702286e-05,
"loss": 0.0,
"step": 192
},
{
"epoch": 0.37,
"learning_rate": 1.876726755707508e-05,
"loss": 0.0,
"step": 193
},
{
"epoch": 0.37,
"learning_rate": 1.8752234219087538e-05,
"loss": 0.0,
"step": 194
},
{
"epoch": 0.38,
"learning_rate": 1.8737115865766865e-05,
"loss": 0.0,
"step": 195
},
{
"epoch": 0.38,
"learning_rate": 1.8721912643966055e-05,
"loss": 0.0,
"step": 196
},
{
"epoch": 0.38,
"learning_rate": 1.8706624701362485e-05,
"loss": 0.0,
"step": 197
},
{
"epoch": 0.38,
"learning_rate": 1.8691252186456465e-05,
"loss": 0.0,
"step": 198
},
{
"epoch": 0.38,
"learning_rate": 1.8675795248569816e-05,
"loss": 0.0,
"step": 199
},
{
"epoch": 0.38,
"learning_rate": 1.866025403784439e-05,
"loss": 0.0,
"step": 200
},
{
"epoch": 0.39,
"learning_rate": 1.8644628705240636e-05,
"loss": 0.0,
"step": 201
},
{
"epoch": 0.39,
"learning_rate": 1.862891940253613e-05,
"loss": 0.0,
"step": 202
},
{
"epoch": 0.39,
"learning_rate": 1.8613126282324092e-05,
"loss": 0.0,
"step": 203
},
{
"epoch": 0.39,
"learning_rate": 1.8597249498011906e-05,
"loss": 0.0,
"step": 204
},
{
"epoch": 0.39,
"learning_rate": 1.858128920381963e-05,
"loss": 0.0,
"step": 205
},
{
"epoch": 0.4,
"learning_rate": 1.8565245554778516e-05,
"loss": 0.0,
"step": 206
},
{
"epoch": 0.4,
"learning_rate": 1.854911870672947e-05,
"loss": 0.0,
"step": 207
},
{
"epoch": 0.4,
"learning_rate": 1.8532908816321557e-05,
"loss": 0.0,
"step": 208
},
{
"epoch": 0.4,
"learning_rate": 1.8516616041010495e-05,
"loss": 0.0,
"step": 209
},
{
"epoch": 0.4,
"learning_rate": 1.8500240539057093e-05,
"loss": 0.0,
"step": 210
},
{
"epoch": 0.41,
"learning_rate": 1.848378246952574e-05,
"loss": 0.0,
"step": 211
},
{
"epoch": 0.41,
"learning_rate": 1.8467241992282842e-05,
"loss": 0.0,
"step": 212
},
{
"epoch": 0.41,
"learning_rate": 1.8450619267995283e-05,
"loss": 0.0,
"step": 213
},
{
"epoch": 0.41,
"learning_rate": 1.843391445812886e-05,
"loss": 0.0,
"step": 214
},
{
"epoch": 0.41,
"learning_rate": 1.84171277249467e-05,
"loss": 0.0,
"step": 215
},
{
"epoch": 0.42,
"learning_rate": 1.8400259231507716e-05,
"loss": 0.0,
"step": 216
},
{
"epoch": 0.42,
"learning_rate": 1.8383309141664992e-05,
"loss": 0.0,
"step": 217
},
{
"epoch": 0.42,
"learning_rate": 1.83662776200642e-05,
"loss": 0.0,
"step": 218
},
{
"epoch": 0.42,
"learning_rate": 1.8349164832142015e-05,
"loss": 0.0,
"step": 219
},
{
"epoch": 0.42,
"learning_rate": 1.833197094412449e-05,
"loss": 0.0,
"step": 220
},
{
"epoch": 0.42,
"learning_rate": 1.8314696123025456e-05,
"loss": 0.0,
"step": 221
},
{
"epoch": 0.43,
"learning_rate": 1.8297340536644877e-05,
"loss": 0.0,
"step": 222
},
{
"epoch": 0.43,
"learning_rate": 1.827990435356725e-05,
"loss": 0.0,
"step": 223
},
{
"epoch": 0.43,
"learning_rate": 1.826238774315995e-05,
"loss": 0.0,
"step": 224
},
{
"epoch": 0.43,
"learning_rate": 1.8244790875571582e-05,
"loss": 0.0,
"step": 225
},
{
"epoch": 0.43,
"learning_rate": 1.8227113921730336e-05,
"loss": 0.0,
"step": 226
},
{
"epoch": 0.44,
"learning_rate": 1.8209357053342325e-05,
"loss": 0.0,
"step": 227
},
{
"epoch": 0.44,
"learning_rate": 1.819152044288992e-05,
"loss": 0.0,
"step": 228
},
{
"epoch": 0.44,
"learning_rate": 1.8173604263630066e-05,
"loss": 0.0,
"step": 229
},
{
"epoch": 0.44,
"learning_rate": 1.8155608689592604e-05,
"loss": 0.0,
"step": 230
},
{
"epoch": 0.44,
"learning_rate": 1.8137533895578585e-05,
"loss": 0.0,
"step": 231
},
{
"epoch": 0.45,
"learning_rate": 1.811938005715857e-05,
"loss": 0.0,
"step": 232
},
{
"epoch": 0.45,
"learning_rate": 1.8101147350670905e-05,
"loss": 0.0,
"step": 233
},
{
"epoch": 0.45,
"learning_rate": 1.8082835953220055e-05,
"loss": 0.0,
"step": 234
},
{
"epoch": 0.45,
"learning_rate": 1.806444604267483e-05,
"loss": 0.0,
"step": 235
},
{
"epoch": 0.45,
"learning_rate": 1.8045977797666685e-05,
"loss": 0.0,
"step": 236
},
{
"epoch": 0.46,
"learning_rate": 1.8027431397587993e-05,
"loss": 0.0,
"step": 237
},
{
"epoch": 0.46,
"learning_rate": 1.8008807022590283e-05,
"loss": 0.0,
"step": 238
},
{
"epoch": 0.46,
"learning_rate": 1.7990104853582494e-05,
"loss": 0.0,
"step": 239
},
{
"epoch": 0.46,
"learning_rate": 1.7971325072229227e-05,
"loss": 0.0,
"step": 240
},
{
"epoch": 0.46,
"learning_rate": 1.7952467860948975e-05,
"loss": 0.0,
"step": 241
},
{
"epoch": 0.47,
"learning_rate": 1.7933533402912354e-05,
"loss": 0.0,
"step": 242
},
{
"epoch": 0.47,
"learning_rate": 1.791452188204031e-05,
"loss": 0.0,
"step": 243
},
{
"epoch": 0.47,
"learning_rate": 1.7895433483002356e-05,
"loss": 0.0,
"step": 244
},
{
"epoch": 0.47,
"learning_rate": 1.7876268391214756e-05,
"loss": 0.0,
"step": 245
},
{
"epoch": 0.47,
"learning_rate": 1.785702679283874e-05,
"loss": 0.0,
"step": 246
},
{
"epoch": 0.47,
"learning_rate": 1.7837708874778683e-05,
"loss": 0.0,
"step": 247
},
{
"epoch": 0.48,
"learning_rate": 1.78183148246803e-05,
"loss": 0.0,
"step": 248
},
{
"epoch": 0.48,
"learning_rate": 1.7798844830928818e-05,
"loss": 0.0,
"step": 249
},
{
"epoch": 0.48,
"learning_rate": 1.777929908264715e-05,
"loss": 0.0,
"step": 250
},
{
"epoch": 0.48,
"learning_rate": 1.775967776969405e-05,
"loss": 0.0,
"step": 251
},
{
"epoch": 0.48,
"learning_rate": 1.7739981082662275e-05,
"loss": 0.0,
"step": 252
},
{
"epoch": 0.49,
"learning_rate": 1.772020921287674e-05,
"loss": 0.0,
"step": 253
},
{
"epoch": 0.49,
"learning_rate": 1.7700362352392632e-05,
"loss": 0.0,
"step": 254
},
{
"epoch": 0.49,
"learning_rate": 1.7680440693993586e-05,
"loss": 0.0,
"step": 255
},
{
"epoch": 0.49,
"learning_rate": 1.766044443118978e-05,
"loss": 0.0,
"step": 256
},
{
"epoch": 0.49,
"learning_rate": 1.7640373758216075e-05,
"loss": 0.0,
"step": 257
},
{
"epoch": 0.5,
"learning_rate": 1.762022887003011e-05,
"loss": 0.0,
"step": 258
},
{
"epoch": 0.5,
"learning_rate": 1.7600009962310417e-05,
"loss": 0.0,
"step": 259
},
{
"epoch": 0.5,
"learning_rate": 1.757971723145453e-05,
"loss": 0.0,
"step": 260
},
{
"epoch": 0.5,
"learning_rate": 1.7559350874577066e-05,
"loss": 0.0,
"step": 261
},
{
"epoch": 0.5,
"learning_rate": 1.75389110895078e-05,
"loss": 0.0,
"step": 262
},
{
"epoch": 0.51,
"learning_rate": 1.7518398074789776e-05,
"loss": 0.0,
"step": 263
},
{
"epoch": 0.51,
"learning_rate": 1.7497812029677344e-05,
"loss": 0.0,
"step": 264
},
{
"epoch": 0.51,
"learning_rate": 1.7477153154134244e-05,
"loss": 0.0,
"step": 265
},
{
"epoch": 0.51,
"learning_rate": 1.7456421648831658e-05,
"loss": 0.0,
"step": 266
},
{
"epoch": 0.51,
"learning_rate": 1.743561771514626e-05,
"loss": 0.0,
"step": 267
},
{
"epoch": 0.52,
"learning_rate": 1.741474155515827e-05,
"loss": 0.0,
"step": 268
},
{
"epoch": 0.52,
"learning_rate": 1.739379337164946e-05,
"loss": 0.0,
"step": 269
},
{
"epoch": 0.52,
"learning_rate": 1.737277336810124e-05,
"loss": 0.0,
"step": 270
},
{
"epoch": 0.52,
"learning_rate": 1.7351681748692622e-05,
"loss": 0.0,
"step": 271
},
{
"epoch": 0.52,
"learning_rate": 1.7330518718298263e-05,
"loss": 0.0,
"step": 272
},
{
"epoch": 0.53,
"learning_rate": 1.7309284482486494e-05,
"loss": 0.0,
"step": 273
},
{
"epoch": 0.53,
"learning_rate": 1.7287979247517285e-05,
"loss": 0.0,
"step": 274
},
{
"epoch": 0.53,
"learning_rate": 1.7266603220340273e-05,
"loss": 0.0,
"step": 275
},
{
"epoch": 0.53,
"learning_rate": 1.7245156608592727e-05,
"loss": 0.0,
"step": 276
},
{
"epoch": 0.53,
"learning_rate": 1.7223639620597556e-05,
"loss": 0.0,
"step": 277
},
{
"epoch": 0.53,
"learning_rate": 1.7202052465361268e-05,
"loss": 0.0,
"step": 278
},
{
"epoch": 0.54,
"learning_rate": 1.718039535257194e-05,
"loss": 0.0,
"step": 279
},
{
"epoch": 0.54,
"learning_rate": 1.7158668492597186e-05,
"loss": 0.0,
"step": 280
},
{
"epoch": 0.54,
"learning_rate": 1.7136872096482123e-05,
"loss": 0.0,
"step": 281
},
{
"epoch": 0.54,
"learning_rate": 1.7115006375947304e-05,
"loss": 0.0,
"step": 282
},
{
"epoch": 0.54,
"learning_rate": 1.7093071543386667e-05,
"loss": 0.0,
"step": 283
},
{
"epoch": 0.55,
"learning_rate": 1.7071067811865477e-05,
"loss": 0.0,
"step": 284
},
{
"epoch": 0.55,
"learning_rate": 1.7048995395118253e-05,
"loss": 0.0,
"step": 285
},
{
"epoch": 0.55,
"learning_rate": 1.7026854507546694e-05,
"loss": 0.0,
"step": 286
},
{
"epoch": 0.55,
"learning_rate": 1.7004645364217584e-05,
"loss": 0.0,
"step": 287
},
{
"epoch": 0.55,
"learning_rate": 1.698236818086073e-05,
"loss": 0.0,
"step": 288
},
{
"epoch": 0.56,
"learning_rate": 1.6960023173866834e-05,
"loss": 0.0,
"step": 289
},
{
"epoch": 0.56,
"learning_rate": 1.693761056028542e-05,
"loss": 0.0,
"step": 290
},
{
"epoch": 0.56,
"learning_rate": 1.6915130557822698e-05,
"loss": 0.0,
"step": 291
},
{
"epoch": 0.56,
"learning_rate": 1.689258338483947e-05,
"loss": 0.0,
"step": 292
},
{
"epoch": 0.56,
"learning_rate": 1.686996926034902e-05,
"loss": 0.0,
"step": 293
},
{
"epoch": 0.57,
"learning_rate": 1.6847288404014937e-05,
"loss": 0.0,
"step": 294
},
{
"epoch": 0.57,
"learning_rate": 1.682454103614904e-05,
"loss": 0.0,
"step": 295
},
{
"epoch": 0.57,
"learning_rate": 1.6801727377709195e-05,
"loss": 0.0,
"step": 296
},
{
"epoch": 0.57,
"learning_rate": 1.67788476502972e-05,
"loss": 0.0,
"step": 297
},
{
"epoch": 0.57,
"learning_rate": 1.6755902076156606e-05,
"loss": 0.0,
"step": 298
},
{
"epoch": 0.57,
"learning_rate": 1.6732890878170573e-05,
"loss": 0.0,
"step": 299
},
{
"epoch": 0.58,
"learning_rate": 1.67098142798597e-05,
"loss": 0.0,
"step": 300
},
{
"epoch": 0.58,
"learning_rate": 1.668667250537987e-05,
"loss": 0.0,
"step": 301
},
{
"epoch": 0.58,
"learning_rate": 1.6663465779520042e-05,
"loss": 0.0,
"step": 302
},
{
"epoch": 0.58,
"learning_rate": 1.6640194327700087e-05,
"loss": 0.0,
"step": 303
},
{
"epoch": 0.58,
"learning_rate": 1.6616858375968596e-05,
"loss": 0.0,
"step": 304
},
{
"epoch": 0.59,
"learning_rate": 1.659345815100069e-05,
"loss": 0.0,
"step": 305
},
{
"epoch": 0.59,
"learning_rate": 1.6569993880095807e-05,
"loss": 0.0,
"step": 306
},
{
"epoch": 0.59,
"learning_rate": 1.6546465791175498e-05,
"loss": 0.0,
"step": 307
},
{
"epoch": 0.59,
"learning_rate": 1.6522874112781213e-05,
"loss": 0.0,
"step": 308
},
{
"epoch": 0.59,
"learning_rate": 1.6499219074072087e-05,
"loss": 0.0,
"step": 309
},
{
"epoch": 0.6,
"learning_rate": 1.6475500904822707e-05,
"loss": 0.0,
"step": 310
},
{
"epoch": 0.6,
"learning_rate": 1.645171983542088e-05,
"loss": 0.0,
"step": 311
},
{
"epoch": 0.6,
"learning_rate": 1.6427876096865394e-05,
"loss": 0.0,
"step": 312
},
{
"epoch": 0.6,
"learning_rate": 1.640396992076379e-05,
"loss": 0.0,
"step": 313
},
{
"epoch": 0.6,
"learning_rate": 1.6380001539330088e-05,
"loss": 0.0,
"step": 314
},
{
"epoch": 0.61,
"learning_rate": 1.6355971185382547e-05,
"loss": 0.0,
"step": 315
},
{
"epoch": 0.61,
"learning_rate": 1.6331879092341402e-05,
"loss": 0.0,
"step": 316
},
{
"epoch": 0.61,
"learning_rate": 1.6307725494226586e-05,
"loss": 0.0,
"step": 317
},
{
"epoch": 0.61,
"learning_rate": 1.6283510625655474e-05,
"loss": 0.0,
"step": 318
},
{
"epoch": 0.61,
"learning_rate": 1.6259234721840595e-05,
"loss": 0.0,
"step": 319
},
{
"epoch": 0.62,
"learning_rate": 1.6234898018587336e-05,
"loss": 0.0,
"step": 320
},
{
"epoch": 0.62,
"learning_rate": 1.6210500752291682e-05,
"loss": 0.0,
"step": 321
},
{
"epoch": 0.62,
"learning_rate": 1.6186043159937884e-05,
"loss": 0.0,
"step": 322
},
{
"epoch": 0.62,
"learning_rate": 1.616152547909618e-05,
"loss": 0.0,
"step": 323
},
{
"epoch": 0.62,
"learning_rate": 1.6136947947920477e-05,
"loss": 0.0,
"step": 324
},
{
"epoch": 0.62,
"learning_rate": 1.611231080514605e-05,
"loss": 0.0,
"step": 325
},
{
"epoch": 0.63,
"learning_rate": 1.608761429008721e-05,
"loss": 0.0,
"step": 326
},
{
"epoch": 0.63,
"learning_rate": 1.606285864263498e-05,
"loss": 0.0,
"step": 327
},
{
"epoch": 0.63,
"learning_rate": 1.6038044103254775e-05,
"loss": 0.0,
"step": 328
},
{
"epoch": 0.63,
"learning_rate": 1.601317091298406e-05,
"loss": 0.0,
"step": 329
},
{
"epoch": 0.63,
"learning_rate": 1.5988239313430004e-05,
"loss": 0.0,
"step": 330
},
{
"epoch": 0.64,
"learning_rate": 1.5963249546767144e-05,
"loss": 0.0,
"step": 331
},
{
"epoch": 0.64,
"learning_rate": 1.5938201855735017e-05,
"loss": 0.0,
"step": 332
},
{
"epoch": 0.64,
"learning_rate": 1.5913096483635827e-05,
"loss": 0.0,
"step": 333
},
{
"epoch": 0.64,
"learning_rate": 1.5887933674332048e-05,
"loss": 0.0,
"step": 334
},
{
"epoch": 0.64,
"learning_rate": 1.5862713672244092e-05,
"loss": 0.0,
"step": 335
},
{
"epoch": 0.65,
"learning_rate": 1.5837436722347902e-05,
"loss": 0.0,
"step": 336
},
{
"epoch": 0.65,
"learning_rate": 1.5812103070172592e-05,
"loss": 0.0,
"step": 337
},
{
"epoch": 0.65,
"learning_rate": 1.578671296179806e-05,
"loss": 0.0,
"step": 338
},
{
"epoch": 0.65,
"learning_rate": 1.5761266643852587e-05,
"loss": 0.0,
"step": 339
},
{
"epoch": 0.65,
"learning_rate": 1.573576436351046e-05,
"loss": 0.0,
"step": 340
},
{
"epoch": 0.66,
"learning_rate": 1.5710206368489555e-05,
"loss": 0.0,
"step": 341
},
{
"epoch": 0.66,
"learning_rate": 1.5684592907048925e-05,
"loss": 0.0,
"step": 342
},
{
"epoch": 0.66,
"learning_rate": 1.5658924227986415e-05,
"loss": 0.0,
"step": 343
},
{
"epoch": 0.66,
"learning_rate": 1.563320058063622e-05,
"loss": 0.0,
"step": 344
},
{
"epoch": 0.66,
"learning_rate": 1.560742221486648e-05,
"loss": 0.0,
"step": 345
},
{
"epoch": 0.67,
"learning_rate": 1.5581589381076843e-05,
"loss": 0.0,
"step": 346
},
{
"epoch": 0.67,
"learning_rate": 1.5555702330196024e-05,
"loss": 0.0,
"step": 347
},
{
"epoch": 0.67,
"learning_rate": 1.5529761313679396e-05,
"loss": 0.0,
"step": 348
},
{
"epoch": 0.67,
"learning_rate": 1.5503766583506522e-05,
"loss": 0.0,
"step": 349
},
{
"epoch": 0.67,
"learning_rate": 1.5477718392178716e-05,
"loss": 0.0,
"step": 350
},
{
"epoch": 0.68,
"learning_rate": 1.545161699271659e-05,
"loss": 0.0,
"step": 351
},
{
"epoch": 0.68,
"learning_rate": 1.5425462638657597e-05,
"loss": 0.0,
"step": 352
},
{
"epoch": 0.68,
"learning_rate": 1.5399255584053568e-05,
"loss": 0.0,
"step": 353
},
{
"epoch": 0.68,
"learning_rate": 1.5372996083468242e-05,
"loss": 0.0,
"step": 354
},
{
"epoch": 0.68,
"learning_rate": 1.5346684391974792e-05,
"loss": 0.0,
"step": 355
},
{
"epoch": 0.68,
"learning_rate": 1.5320320765153367e-05,
"loss": 0.0,
"step": 356
},
{
"epoch": 0.69,
"learning_rate": 1.529390545908857e-05,
"loss": 0.0,
"step": 357
},
{
"epoch": 0.69,
"learning_rate": 1.526743873036701e-05,
"loss": 0.0,
"step": 358
},
{
"epoch": 0.69,
"learning_rate": 1.5240920836074777e-05,
"loss": 0.0,
"step": 359
},
{
"epoch": 0.69,
"learning_rate": 1.5214352033794981e-05,
"loss": 0.0,
"step": 360
},
{
"epoch": 0.69,
"learning_rate": 1.5187732581605217e-05,
"loss": 0.0,
"step": 361
},
{
"epoch": 0.7,
"learning_rate": 1.5161062738075068e-05,
"loss": 0.0,
"step": 362
},
{
"epoch": 0.7,
"learning_rate": 1.5134342762263606e-05,
"loss": 0.0,
"step": 363
},
{
"epoch": 0.7,
"learning_rate": 1.5107572913716859e-05,
"loss": 0.0,
"step": 364
},
{
"epoch": 0.7,
"learning_rate": 1.5080753452465296e-05,
"loss": 0.0,
"step": 365
},
{
"epoch": 0.7,
"learning_rate": 1.505388463902131e-05,
"loss": 0.0,
"step": 366
},
{
"epoch": 0.71,
"learning_rate": 1.502696673437667e-05,
"loss": 0.0,
"step": 367
},
{
"epoch": 0.71,
"learning_rate": 1.5000000000000002e-05,
"loss": 0.0,
"step": 368
},
{
"epoch": 0.71,
"learning_rate": 1.4972984697834238e-05,
"loss": 0.0,
"step": 369
},
{
"epoch": 0.71,
"learning_rate": 1.4945921090294076e-05,
"loss": 0.0,
"step": 370
},
{
"epoch": 0.71,
"learning_rate": 1.4918809440263435e-05,
"loss": 0.0,
"step": 371
},
{
"epoch": 0.72,
"learning_rate": 1.4891650011092896e-05,
"loss": 0.0,
"step": 372
},
{
"epoch": 0.72,
"learning_rate": 1.486444306659714e-05,
"loss": 0.0,
"step": 373
},
{
"epoch": 0.72,
"learning_rate": 1.4837188871052399e-05,
"loss": 0.0,
"step": 374
},
{
"epoch": 0.72,
"learning_rate": 1.4809887689193878e-05,
"loss": 0.0,
"step": 375
},
{
"epoch": 0.72,
"learning_rate": 1.4782539786213184e-05,
"loss": 0.0,
"step": 376
},
{
"epoch": 0.72,
"learning_rate": 1.4755145427755755e-05,
"loss": 0.0,
"step": 377
},
{
"epoch": 0.73,
"learning_rate": 1.4727704879918272e-05,
"loss": 0.0,
"step": 378
},
{
"epoch": 0.73,
"learning_rate": 1.4700218409246087e-05,
"loss": 0.0,
"step": 379
},
{
"epoch": 0.73,
"learning_rate": 1.4672686282730622e-05,
"loss": 0.0,
"step": 380
},
{
"epoch": 0.73,
"learning_rate": 1.4645108767806778e-05,
"loss": 0.0,
"step": 381
},
{
"epoch": 0.73,
"learning_rate": 1.4617486132350343e-05,
"loss": 0.0,
"step": 382
},
{
"epoch": 0.74,
"learning_rate": 1.4589818644675378e-05,
"loss": 0.0,
"step": 383
},
{
"epoch": 0.74,
"learning_rate": 1.4562106573531632e-05,
"loss": 0.0,
"step": 384
},
{
"epoch": 0.74,
"learning_rate": 1.4534350188101905e-05,
"loss": 0.0,
"step": 385
},
{
"epoch": 0.74,
"learning_rate": 1.4506549757999456e-05,
"loss": 0.0,
"step": 386
},
{
"epoch": 0.74,
"learning_rate": 1.4478705553265363e-05,
"loss": 0.0,
"step": 387
},
{
"epoch": 0.75,
"learning_rate": 1.4450817844365924e-05,
"loss": 0.0,
"step": 388
},
{
"epoch": 0.75,
"learning_rate": 1.4422886902190014e-05,
"loss": 0.0,
"step": 389
},
{
"epoch": 0.75,
"learning_rate": 1.4394912998046451e-05,
"loss": 0.0,
"step": 390
},
{
"epoch": 0.75,
"learning_rate": 1.436689640366137e-05,
"loss": 0.0,
"step": 391
},
{
"epoch": 0.75,
"learning_rate": 1.4338837391175582e-05,
"loss": 0.0,
"step": 392
},
{
"epoch": 0.76,
"learning_rate": 1.4310736233141926e-05,
"loss": 0.0,
"step": 393
},
{
"epoch": 0.76,
"learning_rate": 1.4282593202522627e-05,
"loss": 0.0,
"step": 394
},
{
"epoch": 0.76,
"learning_rate": 1.4254408572686642e-05,
"loss": 0.0,
"step": 395
},
{
"epoch": 0.76,
"learning_rate": 1.4226182617406996e-05,
"loss": 0.0,
"step": 396
},
{
"epoch": 0.76,
"learning_rate": 1.4197915610858143e-05,
"loss": 0.0,
"step": 397
},
{
"epoch": 0.77,
"learning_rate": 1.4169607827613284e-05,
"loss": 0.0,
"step": 398
},
{
"epoch": 0.77,
"learning_rate": 1.4141259542641706e-05,
"loss": 0.0,
"step": 399
},
{
"epoch": 0.77,
"learning_rate": 1.4112871031306118e-05,
"loss": 0.0,
"step": 400
}
],
"logging_steps": 1.0,
"max_steps": 1040,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 100,
"total_flos": 1.995362282724393e+17,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}