chitanda's picture
Add files using upload-large-folder tool
ecd4bfe verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.38461538461538464,
"eval_steps": 500,
"global_step": 200,
"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
}
],
"logging_steps": 1.0,
"max_steps": 1040,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 100,
"total_flos": 9.983413310193664e+16,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}