diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,5430 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 20.0,
+  "eval_steps": 500,
+  "global_step": 900,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.4074074074074075e-06,
+      "loss": 6.3594,
+      "step": 1
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.4814814814814815e-05,
+      "loss": 6.3594,
+      "step": 2
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 6.3008,
+      "step": 3
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 6.0938,
+      "step": 4
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 5.7578,
+      "step": 5
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 5.4258,
+      "step": 6
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.185185185185185e-05,
+      "loss": 5.5195,
+      "step": 7
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.925925925925926e-05,
+      "loss": 5.457,
+      "step": 8
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 5.4727,
+      "step": 9
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 7.407407407407407e-05,
+      "loss": 5.4062,
+      "step": 10
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 8.148148148148148e-05,
+      "loss": 5.3789,
+      "step": 11
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 8.888888888888889e-05,
+      "loss": 5.3086,
+      "step": 12
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.62962962962963e-05,
+      "loss": 5.375,
+      "step": 13
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001037037037037037,
+      "loss": 5.1719,
+      "step": 14
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00011111111111111112,
+      "loss": 5.1719,
+      "step": 15
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00011851851851851852,
+      "loss": 5.1836,
+      "step": 16
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00012592592592592592,
+      "loss": 5.2617,
+      "step": 17
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00013333333333333334,
+      "loss": 5.0781,
+      "step": 18
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00014074074074074076,
+      "loss": 5.0859,
+      "step": 19
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00014814814814814815,
+      "loss": 5.1914,
+      "step": 20
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00015555555555555556,
+      "loss": 5.2227,
+      "step": 21
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00016296296296296295,
+      "loss": 5.1094,
+      "step": 22
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00017037037037037037,
+      "loss": 5.0156,
+      "step": 23
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00017777777777777779,
+      "loss": 5.1289,
+      "step": 24
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001851851851851852,
+      "loss": 5.0703,
+      "step": 25
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001925925925925926,
+      "loss": 4.9531,
+      "step": 26
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0002,
+      "loss": 4.8281,
+      "step": 27
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.00019999935249847796,
+      "loss": 4.7852,
+      "step": 28
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.00019999741000229694,
+      "loss": 4.7969,
+      "step": 29
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019999417253661235,
+      "loss": 4.8047,
+      "step": 30
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.00019998964014334946,
+      "loss": 4.6992,
+      "step": 31
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.00019998381288120295,
+      "loss": 4.7617,
+      "step": 32
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.00019997669082563597,
+      "loss": 4.8281,
+      "step": 33
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.0001999682740688794,
+      "loss": 4.625,
+      "step": 34
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001999585627199305,
+      "loss": 4.5703,
+      "step": 35
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019994755690455152,
+      "loss": 4.6289,
+      "step": 36
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.00019993525676526805,
+      "loss": 4.582,
+      "step": 37
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.00019992166246136738,
+      "loss": 4.5508,
+      "step": 38
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.00019990677416889608,
+      "loss": 4.5039,
+      "step": 39
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.000199890592080658,
+      "loss": 4.4609,
+      "step": 40
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.00019987311640621173,
+      "loss": 4.4102,
+      "step": 41
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.0001998543473718677,
+      "loss": 4.3477,
+      "step": 42
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.00019983428522068556,
+      "loss": 4.2051,
+      "step": 43
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001998129302124707,
+      "loss": 4.25,
+      "step": 44
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00019979028262377118,
+      "loss": 4.0957,
+      "step": 45
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.00019976634274787393,
+      "loss": 4.0957,
+      "step": 46
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.00019974111089480112,
+      "loss": 4.084,
+      "step": 47
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00019971458739130598,
+      "loss": 4.0977,
+      "step": 48
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.00019968677258086866,
+      "loss": 4.0059,
+      "step": 49
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00019965766682369186,
+      "loss": 4.0938,
+      "step": 50
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.000199627270496696,
+      "loss": 4.0527,
+      "step": 51
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.00019959558399351444,
+      "loss": 4.1016,
+      "step": 52
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.00019956260772448833,
+      "loss": 4.0449,
+      "step": 53
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.0001995283421166614,
+      "loss": 3.9648,
+      "step": 54
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.0001994927876137743,
+      "loss": 3.9746,
+      "step": 55
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.00019945594467625895,
+      "loss": 3.8984,
+      "step": 56
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.00019941781378123244,
+      "loss": 3.873,
+      "step": 57
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.00019937839542249108,
+      "loss": 3.7598,
+      "step": 58
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001993376901105038,
+      "loss": 3.8691,
+      "step": 59
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00019929569837240564,
+      "loss": 3.8906,
+      "step": 60
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.0001992524207519909,
+      "loss": 3.7227,
+      "step": 61
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.00019920785780970604,
+      "loss": 3.7578,
+      "step": 62
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.00019916201012264254,
+      "loss": 3.7109,
+      "step": 63
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.00019911487828452932,
+      "loss": 3.6836,
+      "step": 64
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.00019906646290572514,
+      "loss": 3.6445,
+      "step": 65
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00019901676461321068,
+      "loss": 3.7051,
+      "step": 66
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.00019896578405058027,
+      "loss": 3.6348,
+      "step": 67
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.00019891352187803376,
+      "loss": 3.582,
+      "step": 68
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.00019885997877236788,
+      "loss": 3.668,
+      "step": 69
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001988051554269675,
+      "loss": 3.6387,
+      "step": 70
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.0001987490525517966,
+      "loss": 3.4805,
+      "step": 71
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00019869167087338907,
+      "loss": 3.4434,
+      "step": 72
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.00019863301113483948,
+      "loss": 3.4824,
+      "step": 73
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.00019857307409579318,
+      "loss": 3.4902,
+      "step": 74
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00019851186053243666,
+      "loss": 3.3926,
+      "step": 75
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001984493712374874,
+      "loss": 3.4238,
+      "step": 76
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.00019838560702018374,
+      "loss": 3.375,
+      "step": 77
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.00019832056870627417,
+      "loss": 3.5664,
+      "step": 78
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.00019825425713800688,
+      "loss": 3.3262,
+      "step": 79
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.00019818667317411865,
+      "loss": 3.5371,
+      "step": 80
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.0001981178176898239,
+      "loss": 3.5586,
+      "step": 81
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.00019804769157680328,
+      "loss": 3.4258,
+      "step": 82
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.000197976295743192,
+      "loss": 3.334,
+      "step": 83
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.00019790363111356837,
+      "loss": 3.3652,
+      "step": 84
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.0001978296986289415,
+      "loss": 3.5098,
+      "step": 85
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.00019775449924673934,
+      "loss": 3.4902,
+      "step": 86
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.00019767803394079615,
+      "loss": 3.3887,
+      "step": 87
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 0.00019760030370134004,
+      "loss": 3.4141,
+      "step": 88
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 0.0001975213095349799,
+      "loss": 3.2051,
+      "step": 89
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.00019744105246469263,
+      "loss": 2.9727,
+      "step": 90
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 0.0001973595335298097,
+      "loss": 3.0605,
+      "step": 91
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 0.0001972767537860038,
+      "loss": 3.1074,
+      "step": 92
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 0.0001971927143052752,
+      "loss": 2.9453,
+      "step": 93
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 0.00019710741617593767,
+      "loss": 2.9258,
+      "step": 94
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 0.00019702086050260456,
+      "loss": 3.0449,
+      "step": 95
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 0.00019693304840617457,
+      "loss": 2.9863,
+      "step": 96
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 0.00019684398102381694,
+      "loss": 2.8984,
+      "step": 97
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 0.000196753659508957,
+      "loss": 2.8652,
+      "step": 98
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 0.00019666208503126112,
+      "loss": 3.0098,
+      "step": 99
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 0.0001965692587766216,
+      "loss": 2.7969,
+      "step": 100
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 0.00019647518194714124,
+      "loss": 2.873,
+      "step": 101
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 0.00019637985576111778,
+      "loss": 3.0078,
+      "step": 102
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 0.00019628328145302837,
+      "loss": 2.8848,
+      "step": 103
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 0.00019618546027351316,
+      "loss": 2.8301,
+      "step": 104
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 0.0001960863934893594,
+      "loss": 3.1016,
+      "step": 105
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 0.00019598608238348493,
+      "loss": 2.7852,
+      "step": 106
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 0.0001958845282549217,
+      "loss": 3.0176,
+      "step": 107
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 0.00019578173241879872,
+      "loss": 2.8535,
+      "step": 108
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 0.0001956776962063252,
+      "loss": 2.9648,
+      "step": 109
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 0.00019557242096477327,
+      "loss": 2.9141,
+      "step": 110
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 0.00019546590805746052,
+      "loss": 2.9336,
+      "step": 111
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 0.00019535815886373233,
+      "loss": 2.9141,
+      "step": 112
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 0.00019524917477894404,
+      "loss": 2.8945,
+      "step": 113
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 0.00019513895721444286,
+      "loss": 2.7754,
+      "step": 114
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 0.00019502750759754962,
+      "loss": 2.793,
+      "step": 115
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 0.0001949148273715402,
+      "loss": 2.8418,
+      "step": 116
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 0.00019480091799562704,
+      "loss": 2.707,
+      "step": 117
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 0.00019468578094493998,
+      "loss": 2.7285,
+      "step": 118
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 0.00019456941771050734,
+      "loss": 2.9492,
+      "step": 119
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 0.00019445182979923654,
+      "loss": 2.7891,
+      "step": 120
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 0.00019433301873389463,
+      "loss": 2.793,
+      "step": 121
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 0.00019421298605308847,
+      "loss": 2.7969,
+      "step": 122
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 0.000194091733311245,
+      "loss": 2.873,
+      "step": 123
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 0.00019396926207859084,
+      "loss": 2.8613,
+      "step": 124
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 0.00019384557394113228,
+      "loss": 2.8047,
+      "step": 125
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 0.00019372067050063438,
+      "loss": 2.9355,
+      "step": 126
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 0.00019359455337460054,
+      "loss": 2.8828,
+      "step": 127
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 0.00019346722419625136,
+      "loss": 2.8652,
+      "step": 128
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 0.0001933386846145036,
+      "loss": 2.9199,
+      "step": 129
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 0.00019320893629394873,
+      "loss": 2.9102,
+      "step": 130
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 0.00019307798091483145,
+      "loss": 2.7422,
+      "step": 131
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 0.00019294582017302797,
+      "loss": 2.8496,
+      "step": 132
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 0.0001928124557800238,
+      "loss": 2.8379,
+      "step": 133
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 0.000192677889462892,
+      "loss": 2.7012,
+      "step": 134
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.00019254212296427044,
+      "loss": 2.6133,
+      "step": 135
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 0.00019240515804233937,
+      "loss": 2.498,
+      "step": 136
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 0.00019226699647079874,
+      "loss": 2.4883,
+      "step": 137
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 0.0001921276400388451,
+      "loss": 2.2422,
+      "step": 138
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 0.0001919870905511485,
+      "loss": 2.4473,
+      "step": 139
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 0.00019184534982782904,
+      "loss": 2.541,
+      "step": 140
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 0.00019170241970443343,
+      "loss": 2.416,
+      "step": 141
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 0.00019155830203191113,
+      "loss": 2.4648,
+      "step": 142
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 0.00019141299867659036,
+      "loss": 2.5273,
+      "step": 143
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 0.00019126651152015403,
+      "loss": 2.4902,
+      "step": 144
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 0.00019111884245961522,
+      "loss": 2.4023,
+      "step": 145
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 0.00019096999340729282,
+      "loss": 2.4238,
+      "step": 146
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 0.00019081996629078657,
+      "loss": 2.4004,
+      "step": 147
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 0.00019066876305295216,
+      "loss": 2.3242,
+      "step": 148
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 0.00019051638565187616,
+      "loss": 2.4141,
+      "step": 149
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 0.00019036283606085053,
+      "loss": 2.5547,
+      "step": 150
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 0.0001902081162683472,
+      "loss": 2.4004,
+      "step": 151
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 0.00019005222827799212,
+      "loss": 2.3369,
+      "step": 152
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 0.00018989517410853955,
+      "loss": 2.4609,
+      "step": 153
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 0.00018973695579384576,
+      "loss": 2.332,
+      "step": 154
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 0.00018957757538284273,
+      "loss": 2.2852,
+      "step": 155
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 0.00018941703493951164,
+      "loss": 2.3906,
+      "step": 156
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 0.00018925533654285615,
+      "loss": 2.3008,
+      "step": 157
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 0.00018909248228687536,
+      "loss": 2.3008,
+      "step": 158
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 0.00018892847428053693,
+      "loss": 2.3271,
+      "step": 159
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 0.00018876331464774945,
+      "loss": 2.3242,
+      "step": 160
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 0.00018859700552733516,
+      "loss": 2.3652,
+      "step": 161
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 0.00018842954907300236,
+      "loss": 2.3555,
+      "step": 162
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 0.00018826094745331715,
+      "loss": 2.3613,
+      "step": 163
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 0.00018809120285167565,
+      "loss": 2.3691,
+      "step": 164
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 0.00018792031746627563,
+      "loss": 2.3066,
+      "step": 165
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 0.000187748293510088,
+      "loss": 2.3613,
+      "step": 166
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 0.0001875751332108283,
+      "loss": 2.3457,
+      "step": 167
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 0.0001874008388109276,
+      "loss": 2.3652,
+      "step": 168
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 0.0001872254125675037,
+      "loss": 2.3506,
+      "step": 169
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 0.0001870488567523318,
+      "loss": 2.3672,
+      "step": 170
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 0.00018687117365181512,
+      "loss": 2.2666,
+      "step": 171
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 0.00018669236556695515,
+      "loss": 2.5156,
+      "step": 172
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.00018651243481332213,
+      "loss": 2.3926,
+      "step": 173
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 0.00018633138372102468,
+      "loss": 2.4102,
+      "step": 174
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 0.00018614921463468002,
+      "loss": 2.4082,
+      "step": 175
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 0.00018596592991338334,
+      "loss": 2.2695,
+      "step": 176
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 0.00018578153193067745,
+      "loss": 2.418,
+      "step": 177
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 0.0001855960230745217,
+      "loss": 2.3223,
+      "step": 178
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 0.0001854094057472615,
+      "loss": 2.3398,
+      "step": 179
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 0.00018522168236559695,
+      "loss": 2.2422,
+      "step": 180
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 0.00018503285536055147,
+      "loss": 2.0654,
+      "step": 181
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 0.0001848429271774406,
+      "loss": 2.0186,
+      "step": 182
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 0.00018465190027584005,
+      "loss": 2.0049,
+      "step": 183
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 0.00018445977712955403,
+      "loss": 2.0869,
+      "step": 184
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 0.0001842665602265831,
+      "loss": 2.0547,
+      "step": 185
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 0.00018407225206909208,
+      "loss": 2.0264,
+      "step": 186
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 0.0001838768551733775,
+      "loss": 1.8994,
+      "step": 187
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 0.00018368037206983515,
+      "loss": 2.0342,
+      "step": 188
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 0.00018348280530292713,
+      "loss": 2.0576,
+      "step": 189
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 0.00018328415743114912,
+      "loss": 2.0615,
+      "step": 190
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 0.00018308443102699713,
+      "loss": 2.0518,
+      "step": 191
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 0.00018288362867693414,
+      "loss": 2.0537,
+      "step": 192
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 0.0001826817529813567,
+      "loss": 1.9531,
+      "step": 193
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 0.00018247880655456125,
+      "loss": 2.0049,
+      "step": 194
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 0.00018227479202471015,
+      "loss": 2.0693,
+      "step": 195
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 0.00018206971203379773,
+      "loss": 2.0088,
+      "step": 196
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 0.00018186356923761626,
+      "loss": 1.9189,
+      "step": 197
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 0.0001816563663057211,
+      "loss": 2.0371,
+      "step": 198
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 0.00018144810592139656,
+      "loss": 2.1172,
+      "step": 199
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 0.00018123879078162097,
+      "loss": 1.9873,
+      "step": 200
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 0.00018102842359703176,
+      "loss": 2.0508,
+      "step": 201
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 0.0001808170070918904,
+      "loss": 2.0654,
+      "step": 202
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 0.00018060454400404695,
+      "loss": 2.0928,
+      "step": 203
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 0.000180391037084905,
+      "loss": 2.0166,
+      "step": 204
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 0.0001801764890993856,
+      "loss": 2.0518,
+      "step": 205
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 0.0001799609028258917,
+      "loss": 1.9688,
+      "step": 206
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 0.00017974428105627208,
+      "loss": 1.9512,
+      "step": 207
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 0.0001795266265957853,
+      "loss": 2.0537,
+      "step": 208
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 0.0001793079422630632,
+      "loss": 2.0195,
+      "step": 209
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 0.00017908823089007457,
+      "loss": 2.0605,
+      "step": 210
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 0.00017886749532208837,
+      "loss": 2.1709,
+      "step": 211
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 0.00017864573841763694,
+      "loss": 1.9629,
+      "step": 212
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 0.00017842296304847893,
+      "loss": 2.126,
+      "step": 213
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 0.00017819917209956215,
+      "loss": 2.0469,
+      "step": 214
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 0.00017797436846898619,
+      "loss": 2.0107,
+      "step": 215
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 0.00017774855506796496,
+      "loss": 2.0693,
+      "step": 216
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 0.0001775217348207888,
+      "loss": 2.0762,
+      "step": 217
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 0.00017729391066478688,
+      "loss": 2.0322,
+      "step": 218
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 0.00017706508555028893,
+      "loss": 2.0098,
+      "step": 219
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 0.00017683526244058716,
+      "loss": 2.0449,
+      "step": 220
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 0.0001766044443118978,
+      "loss": 2.0615,
+      "step": 221
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 0.0001763726341533227,
+      "loss": 1.9961,
+      "step": 222
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 0.00017613983496681044,
+      "loss": 2.0596,
+      "step": 223
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 0.00017590604976711754,
+      "loss": 2.0498,
+      "step": 224
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 0.00017567128158176953,
+      "loss": 2.0449,
+      "step": 225
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 0.00017543553345102152,
+      "loss": 1.7109,
+      "step": 226
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 0.00017519880842781894,
+      "loss": 1.6914,
+      "step": 227
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 0.0001749611095777581,
+      "loss": 1.8711,
+      "step": 228
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 0.00017472243997904626,
+      "loss": 1.6826,
+      "step": 229
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 0.00017448280272246212,
+      "loss": 1.8135,
+      "step": 230
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 0.00017424220091131535,
+      "loss": 1.7578,
+      "step": 231
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 0.00017400063766140678,
+      "loss": 1.7686,
+      "step": 232
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 0.0001737581161009878,
+      "loss": 1.7383,
+      "step": 233
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 0.00017351463937072004,
+      "loss": 1.6729,
+      "step": 234
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 0.00017327021062363458,
+      "loss": 1.6133,
+      "step": 235
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 0.00017302483302509108,
+      "loss": 1.7119,
+      "step": 236
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 0.00017277850975273696,
+      "loss": 1.6914,
+      "step": 237
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 0.00017253124399646606,
+      "loss": 1.7627,
+      "step": 238
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 0.00017228303895837748,
+      "loss": 1.7412,
+      "step": 239
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 0.000172033897852734,
+      "loss": 1.7412,
+      "step": 240
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 0.00017178382390592057,
+      "loss": 1.8037,
+      "step": 241
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 0.0001715328203564023,
+      "loss": 1.6709,
+      "step": 242
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 0.00017128089045468294,
+      "loss": 1.6562,
+      "step": 243
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 0.00017102803746326227,
+      "loss": 1.6533,
+      "step": 244
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 0.00017077426465659433,
+      "loss": 1.752,
+      "step": 245
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 0.0001705195753210446,
+      "loss": 1.7041,
+      "step": 246
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 0.00017026397275484773,
+      "loss": 1.7969,
+      "step": 247
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 0.0001700074602680648,
+      "loss": 1.7539,
+      "step": 248
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 0.0001697500411825403,
+      "loss": 1.79,
+      "step": 249
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 0.00016949171883185918,
+      "loss": 1.8535,
+      "step": 250
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 0.0001692324965613038,
+      "loss": 1.8213,
+      "step": 251
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 0.00016897237772781044,
+      "loss": 1.8232,
+      "step": 252
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 0.00016871136569992587,
+      "loss": 1.8584,
+      "step": 253
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 0.00016844946385776384,
+      "loss": 1.8145,
+      "step": 254
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 0.0001681866755929612,
+      "loss": 1.7588,
+      "step": 255
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 0.00016792300430863396,
+      "loss": 1.8535,
+      "step": 256
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 0.0001676584534193332,
+      "loss": 1.8047,
+      "step": 257
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 0.00016739302635100108,
+      "loss": 1.7617,
+      "step": 258
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 0.00016712672654092622,
+      "loss": 1.8027,
+      "step": 259
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 0.0001668595574376992,
+      "loss": 1.8447,
+      "step": 260
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 0.00016659152250116812,
+      "loss": 1.8174,
+      "step": 261
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 0.0001663226252023935,
+      "loss": 1.707,
+      "step": 262
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 0.00016605286902360357,
+      "loss": 1.8584,
+      "step": 263
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 0.00016578225745814907,
+      "loss": 1.7969,
+      "step": 264
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 0.000165510794010458,
+      "loss": 1.7988,
+      "step": 265
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 0.00016523848219599023,
+      "loss": 1.793,
+      "step": 266
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 0.00016496532554119214,
+      "loss": 1.8994,
+      "step": 267
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 0.0001646913275834506,
+      "loss": 1.7725,
+      "step": 268
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 0.00016441649187104763,
+      "loss": 1.8682,
+      "step": 269
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 0.000164140821963114,
+      "loss": 1.7598,
+      "step": 270
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 0.00016386432142958342,
+      "loss": 1.4678,
+      "step": 271
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 0.00016358699385114625,
+      "loss": 1.4043,
+      "step": 272
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 0.000163308842819203,
+      "loss": 1.5,
+      "step": 273
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 0.00016302987193581806,
+      "loss": 1.5537,
+      "step": 274
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 0.00016275008481367287,
+      "loss": 1.4404,
+      "step": 275
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 0.00016246948507601914,
+      "loss": 1.5049,
+      "step": 276
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 0.00016218807635663202,
+      "loss": 1.4834,
+      "step": 277
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 0.00016190586229976304,
+      "loss": 1.5586,
+      "step": 278
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 0.00016162284656009274,
+      "loss": 1.501,
+      "step": 279
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 0.00016133903280268362,
+      "loss": 1.4824,
+      "step": 280
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 0.0001610544247029325,
+      "loss": 1.5479,
+      "step": 281
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 0.0001607690259465229,
+      "loss": 1.4863,
+      "step": 282
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 0.0001604828402293774,
+      "loss": 1.3975,
+      "step": 283
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 0.00016019587125760978,
+      "loss": 1.4707,
+      "step": 284
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 0.00015990812274747692,
+      "loss": 1.5244,
+      "step": 285
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 0.00015961959842533083,
+      "loss": 1.4512,
+      "step": 286
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 0.00015933030202757022,
+      "loss": 1.4209,
+      "step": 287
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 0.00015904023730059228,
+      "loss": 1.4961,
+      "step": 288
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 0.00015874940800074402,
+      "loss": 1.4639,
+      "step": 289
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 0.00015845781789427377,
+      "loss": 1.5713,
+      "step": 290
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 0.00015816547075728226,
+      "loss": 1.4365,
+      "step": 291
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 0.00015787237037567385,
+      "loss": 1.5361,
+      "step": 292
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 0.0001575785205451073,
+      "loss": 1.4453,
+      "step": 293
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 0.000157283925070947,
+      "loss": 1.5576,
+      "step": 294
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 0.0001569885877682132,
+      "loss": 1.4775,
+      "step": 295
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 0.000156692512461533,
+      "loss": 1.542,
+      "step": 296
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 0.00015639570298509064,
+      "loss": 1.5146,
+      "step": 297
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 0.00015609816318257788,
+      "loss": 1.582,
+      "step": 298
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 0.00015579989690714423,
+      "loss": 1.4658,
+      "step": 299
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 0.000155500908021347,
+      "loss": 1.6279,
+      "step": 300
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 0.00015520120039710139,
+      "loss": 1.5908,
+      "step": 301
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 0.0001549007779156302,
+      "loss": 1.4941,
+      "step": 302
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 0.00015459964446741382,
+      "loss": 1.5684,
+      "step": 303
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 0.0001542978039521395,
+      "loss": 1.6045,
+      "step": 304
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 0.00015399526027865107,
+      "loss": 1.5527,
+      "step": 305
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 0.0001536920173648984,
+      "loss": 1.5781,
+      "step": 306
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 0.00015338807913788636,
+      "loss": 1.5352,
+      "step": 307
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 0.0001530834495336243,
+      "loss": 1.6094,
+      "step": 308
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 0.00015277813249707487,
+      "loss": 1.6562,
+      "step": 309
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 0.000152472131982103,
+      "loss": 1.6064,
+      "step": 310
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 0.0001521654519514246,
+      "loss": 1.5303,
+      "step": 311
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 0.0001518580963765555,
+      "loss": 1.5273,
+      "step": 312
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 0.00015155006923775965,
+      "loss": 1.6074,
+      "step": 313
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 0.00015124137452399795,
+      "loss": 1.6553,
+      "step": 314
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 0.00015093201623287631,
+      "loss": 1.5859,
+      "step": 315
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 0.00015062199837059405,
+      "loss": 1.3838,
+      "step": 316
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 0.00015031132495189187,
+      "loss": 1.2764,
+      "step": 317
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 1.2939,
+      "step": 318
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 0.00014968802754658614,
+      "loss": 1.4043,
+      "step": 319
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 0.0001493754116317029,
+      "loss": 1.3164,
+      "step": 320
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 0.00014906215630373606,
+      "loss": 1.3623,
+      "step": 321
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 0.00014874826561935155,
+      "loss": 1.2451,
+      "step": 322
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 0.00014843374364344333,
+      "loss": 1.3223,
+      "step": 323
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 0.00014811859444908052,
+      "loss": 1.3125,
+      "step": 324
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 0.0001478028221174548,
+      "loss": 1.3672,
+      "step": 325
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 0.00014748643073782752,
+      "loss": 1.2632,
+      "step": 326
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 0.00014716942440747664,
+      "loss": 1.2617,
+      "step": 327
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 0.00014685180723164376,
+      "loss": 1.3564,
+      "step": 328
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 0.000146533583323481,
+      "loss": 1.3525,
+      "step": 329
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 0.0001462147568039977,
+      "loss": 1.3721,
+      "step": 330
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 0.00014589533180200693,
+      "loss": 1.2017,
+      "step": 331
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 0.00014557531245407225,
+      "loss": 1.3311,
+      "step": 332
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 0.00014525470290445392,
+      "loss": 1.3828,
+      "step": 333
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 0.00014493350730505533,
+      "loss": 1.2324,
+      "step": 334
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 0.0001446117298153693,
+      "loss": 1.3467,
+      "step": 335
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 0.00014428937460242417,
+      "loss": 1.3291,
+      "step": 336
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 0.00014396644584072972,
+      "loss": 1.29,
+      "step": 337
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 0.00014364294771222324,
+      "loss": 1.2803,
+      "step": 338
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 0.00014331888440621533,
+      "loss": 1.4258,
+      "step": 339
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 0.00014299426011933568,
+      "loss": 1.2871,
+      "step": 340
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 0.0001426690790554787,
+      "loss": 1.4482,
+      "step": 341
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 0.00014234334542574906,
+      "loss": 1.3438,
+      "step": 342
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 0.00014201706344840712,
+      "loss": 1.3965,
+      "step": 343
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 0.00014169023734881452,
+      "loss": 1.3115,
+      "step": 344
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 0.00014136287135937915,
+      "loss": 1.3428,
+      "step": 345
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 0.00014103496971950053,
+      "loss": 1.291,
+      "step": 346
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 0.0001407065366755149,
+      "loss": 1.4287,
+      "step": 347
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 0.00014037757648064018,
+      "loss": 1.3296,
+      "step": 348
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 0.00014004809339492088,
+      "loss": 1.3203,
+      "step": 349
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 0.00013971809168517298,
+      "loss": 1.3623,
+      "step": 350
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 0.00013938757562492873,
+      "loss": 1.3008,
+      "step": 351
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 0.0001390565494943811,
+      "loss": 1.4365,
+      "step": 352
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 0.00013872501758032863,
+      "loss": 1.2852,
+      "step": 353
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 0.00013839298417611963,
+      "loss": 1.4785,
+      "step": 354
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 0.00013806045358159683,
+      "loss": 1.2998,
+      "step": 355
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 0.00013772743010304154,
+      "loss": 1.373,
+      "step": 356
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 0.00013739391805311793,
+      "loss": 1.416,
+      "step": 357
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 0.00013705992175081728,
+      "loss": 1.5264,
+      "step": 358
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 0.00013672544552140176,
+      "loss": 1.4072,
+      "step": 359
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 0.00013639049369634876,
+      "loss": 1.2705,
+      "step": 360
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 0.00013605507061329464,
+      "loss": 1.1445,
+      "step": 361
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 0.0001357191806159785,
+      "loss": 1.1514,
+      "step": 362
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 0.0001353828280541861,
+      "loss": 1.0825,
+      "step": 363
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 0.00013504601728369327,
+      "loss": 1.1479,
+      "step": 364
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 0.00013470875266620978,
+      "loss": 1.0366,
+      "step": 365
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 0.00013437103856932264,
+      "loss": 1.1152,
+      "step": 366
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 0.00013403287936643977,
+      "loss": 1.1719,
+      "step": 367
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 0.00013369427943673312,
+      "loss": 1.1543,
+      "step": 368
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 0.00013335524316508208,
+      "loss": 1.0898,
+      "step": 369
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 0.00013301577494201664,
+      "loss": 1.2358,
+      "step": 370
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 0.0001326758791636607,
+      "loss": 1.0977,
+      "step": 371
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 0.00013233556023167485,
+      "loss": 1.1689,
+      "step": 372
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 0.0001319948225531997,
+      "loss": 1.1787,
+      "step": 373
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 0.0001316536705407985,
+      "loss": 1.2651,
+      "step": 374
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 0.00013131210861240026,
+      "loss": 1.1973,
+      "step": 375
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 0.0001309701411912423,
+      "loss": 1.1792,
+      "step": 376
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 0.00013062777270581312,
+      "loss": 1.208,
+      "step": 377
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 0.00013028500758979506,
+      "loss": 1.1719,
+      "step": 378
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 0.00012994185028200684,
+      "loss": 1.1865,
+      "step": 379
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 0.00012959830522634596,
+      "loss": 1.1963,
+      "step": 380
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 0.00012925437687173142,
+      "loss": 1.1533,
+      "step": 381
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 0.00012891006967204584,
+      "loss": 1.1924,
+      "step": 382
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 0.00012856538808607795,
+      "loss": 1.208,
+      "step": 383
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 0.00012822033657746478,
+      "loss": 1.1924,
+      "step": 384
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 0.0001278749196146339,
+      "loss": 1.1641,
+      "step": 385
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 0.0001275291416707454,
+      "loss": 1.228,
+      "step": 386
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 0.0001271830072236343,
+      "loss": 1.207,
+      "step": 387
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 0.00012683652075575218,
+      "loss": 1.1543,
+      "step": 388
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 0.00012648968675410928,
+      "loss": 1.23,
+      "step": 389
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 0.00012614250971021657,
+      "loss": 1.2354,
+      "step": 390
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 0.0001257949941200273,
+      "loss": 1.207,
+      "step": 391
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 0.00012544714448387893,
+      "loss": 1.2148,
+      "step": 392
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 0.00012509896530643488,
+      "loss": 1.2241,
+      "step": 393
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 0.00012475046109662605,
+      "loss": 1.1982,
+      "step": 394
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 0.0001244016363675926,
+      "loss": 1.1904,
+      "step": 395
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 0.00012405249563662537,
+      "loss": 1.2168,
+      "step": 396
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 0.00012370304342510747,
+      "loss": 1.1934,
+      "step": 397
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 0.00012335328425845565,
+      "loss": 1.2246,
+      "step": 398
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 0.00012300322266606178,
+      "loss": 1.2129,
+      "step": 399
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 0.00012265286318123415,
+      "loss": 1.3145,
+      "step": 400
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 0.00012230221034113874,
+      "loss": 1.2236,
+      "step": 401
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 0.00012195126868674051,
+      "loss": 1.1782,
+      "step": 402
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 0.00012160004276274453,
+      "loss": 1.1338,
+      "step": 403
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 0.00012124853711753727,
+      "loss": 1.2559,
+      "step": 404
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 0.00012089675630312754,
+      "loss": 1.1895,
+      "step": 405
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 0.00012054470487508751,
+      "loss": 0.9648,
+      "step": 406
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 0.00012019238739249397,
+      "loss": 1.082,
+      "step": 407
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 0.000119839808417869,
+      "loss": 1.0996,
+      "step": 408
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 0.00011948697251712109,
+      "loss": 1.1094,
+      "step": 409
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 0.00011913388425948584,
+      "loss": 1.0088,
+      "step": 410
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 0.00011878054821746703,
+      "loss": 1.0249,
+      "step": 411
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 0.00011842696896677708,
+      "loss": 0.9722,
+      "step": 412
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 0.00011807315108627806,
+      "loss": 0.9707,
+      "step": 413
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 0.0001177190991579223,
+      "loss": 1.0815,
+      "step": 414
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 0.00011736481776669306,
+      "loss": 1.0107,
+      "step": 415
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 0.0001170103115005451,
+      "loss": 1.0142,
+      "step": 416
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 0.00011665558495034546,
+      "loss": 1.0698,
+      "step": 417
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 0.00011630064270981367,
+      "loss": 0.9937,
+      "step": 418
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 0.00011594548937546257,
+      "loss": 0.9829,
+      "step": 419
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 0.00011559012954653865,
+      "loss": 1.0083,
+      "step": 420
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 0.00011523456782496253,
+      "loss": 1.0469,
+      "step": 421
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 0.00011487880881526934,
+      "loss": 1.1035,
+      "step": 422
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 0.00011452285712454904,
+      "loss": 1.0723,
+      "step": 423
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 0.00011416671736238689,
+      "loss": 1.0435,
+      "step": 424
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 0.00011381039414080365,
+      "loss": 1.1118,
+      "step": 425
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 0.00011345389207419588,
+      "loss": 1.0879,
+      "step": 426
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 0.00011309721577927619,
+      "loss": 1.0337,
+      "step": 427
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 0.00011274036987501348,
+      "loss": 0.9663,
+      "step": 428
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 0.00011238335898257304,
+      "loss": 0.979,
+      "step": 429
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 0.0001120261877252568,
+      "loss": 1.0454,
+      "step": 430
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 0.00011166886072844342,
+      "loss": 1.0103,
+      "step": 431
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 0.00011131138261952845,
+      "loss": 1.1553,
+      "step": 432
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 0.00011095375802786419,
+      "loss": 1.0342,
+      "step": 433
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 0.00011059599158470002,
+      "loss": 1.0591,
+      "step": 434
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 0.00011023808792312227,
+      "loss": 1.0347,
+      "step": 435
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 0.00010988005167799427,
+      "loss": 1.0762,
+      "step": 436
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 0.00010952188748589625,
+      "loss": 1.1216,
+      "step": 437
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 0.0001091635999850655,
+      "loss": 1.1182,
+      "step": 438
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 0.00010880519381533592,
+      "loss": 1.0396,
+      "step": 439
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 0.00010844667361807842,
+      "loss": 0.9829,
+      "step": 440
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 0.00010808804403614043,
+      "loss": 1.0903,
+      "step": 441
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 0.00010772930971378596,
+      "loss": 1.0249,
+      "step": 442
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 0.00010737047529663545,
+      "loss": 1.0493,
+      "step": 443
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 0.00010701154543160541,
+      "loss": 1.1426,
+      "step": 444
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 0.00010665252476684864,
+      "loss": 1.0591,
+      "step": 445
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 0.0001062934179516936,
+      "loss": 1.0083,
+      "step": 446
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 0.00010593422963658452,
+      "loss": 1.0356,
+      "step": 447
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 0.00010557496447302102,
+      "loss": 1.1387,
+      "step": 448
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 0.00010521562711349788,
+      "loss": 1.0908,
+      "step": 449
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 0.00010485622221144484,
+      "loss": 1.1177,
+      "step": 450
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 0.00010449675442116634,
+      "loss": 0.9346,
+      "step": 451
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 0.00010413722839778117,
+      "loss": 0.96,
+      "step": 452
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 0.00010377764879716234,
+      "loss": 0.9424,
+      "step": 453
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 0.00010341802027587659,
+      "loss": 0.9385,
+      "step": 454
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 0.00010305834749112421,
+      "loss": 0.8706,
+      "step": 455
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 0.00010269863510067872,
+      "loss": 0.8843,
+      "step": 456
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 0.00010233888776282649,
+      "loss": 0.8813,
+      "step": 457
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 0.00010197911013630659,
+      "loss": 0.9258,
+      "step": 458
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 0.00010161930688025017,
+      "loss": 0.9229,
+      "step": 459
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 0.00010125948265412033,
+      "loss": 0.8975,
+      "step": 460
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 0.0001008996421176518,
+      "loss": 0.8848,
+      "step": 461
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 0.00010053978993079045,
+      "loss": 0.9873,
+      "step": 462
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 0.00010017993075363305,
+      "loss": 0.9463,
+      "step": 463
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 9.982006924636697e-05,
+      "loss": 0.9058,
+      "step": 464
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 9.946021006920959e-05,
+      "loss": 1.0679,
+      "step": 465
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 9.910035788234822e-05,
+      "loss": 0.9067,
+      "step": 466
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 9.874051734587968e-05,
+      "loss": 0.9453,
+      "step": 467
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 9.838069311974986e-05,
+      "loss": 0.9468,
+      "step": 468
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 9.802088986369342e-05,
+      "loss": 0.9111,
+      "step": 469
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 9.766111223717352e-05,
+      "loss": 1.0132,
+      "step": 470
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 9.730136489932133e-05,
+      "loss": 0.9922,
+      "step": 471
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 9.694165250887584e-05,
+      "loss": 0.9736,
+      "step": 472
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 9.658197972412345e-05,
+      "loss": 0.9551,
+      "step": 473
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 9.622235120283769e-05,
+      "loss": 0.9575,
+      "step": 474
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 9.586277160221884e-05,
+      "loss": 0.9448,
+      "step": 475
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 9.550324557883373e-05,
+      "loss": 0.9775,
+      "step": 476
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 9.514377778855521e-05,
+      "loss": 1.0044,
+      "step": 477
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 9.478437288650213e-05,
+      "loss": 0.9067,
+      "step": 478
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 9.442503552697899e-05,
+      "loss": 0.9453,
+      "step": 479
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 9.406577036341548e-05,
+      "loss": 1.0103,
+      "step": 480
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 9.37065820483064e-05,
+      "loss": 1.0801,
+      "step": 481
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 9.334747523315137e-05,
+      "loss": 0.957,
+      "step": 482
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 9.298845456839459e-05,
+      "loss": 0.9092,
+      "step": 483
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 9.262952470336458e-05,
+      "loss": 0.9854,
+      "step": 484
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 9.227069028621406e-05,
+      "loss": 0.9238,
+      "step": 485
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 9.19119559638596e-05,
+      "loss": 0.9814,
+      "step": 486
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 9.15533263819216e-05,
+      "loss": 1.0024,
+      "step": 487
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 9.119480618466409e-05,
+      "loss": 0.9043,
+      "step": 488
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 9.083640001493454e-05,
+      "loss": 0.98,
+      "step": 489
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 9.047811251410376e-05,
+      "loss": 1.0547,
+      "step": 490
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 9.011994832200577e-05,
+      "loss": 0.9736,
+      "step": 491
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 8.976191207687775e-05,
+      "loss": 1.0371,
+      "step": 492
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 8.94040084153e-05,
+      "loss": 0.9771,
+      "step": 493
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 8.904624197213585e-05,
+      "loss": 0.9429,
+      "step": 494
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 8.868861738047158e-05,
+      "loss": 0.9692,
+      "step": 495
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 8.83311392715566e-05,
+      "loss": 0.8188,
+      "step": 496
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 8.797381227474324e-05,
+      "loss": 0.8862,
+      "step": 497
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 8.7616641017427e-05,
+      "loss": 0.8545,
+      "step": 498
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 8.725963012498657e-05,
+      "loss": 0.8179,
+      "step": 499
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 8.690278422072384e-05,
+      "loss": 0.9062,
+      "step": 500
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 8.654610792580415e-05,
+      "loss": 0.811,
+      "step": 501
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 8.61896058591964e-05,
+      "loss": 0.7476,
+      "step": 502
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 8.583328263761316e-05,
+      "loss": 0.835,
+      "step": 503
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 8.5477142875451e-05,
+      "loss": 0.8276,
+      "step": 504
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 8.512119118473067e-05,
+      "loss": 0.8247,
+      "step": 505
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 8.476543217503748e-05,
+      "loss": 0.8823,
+      "step": 506
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 8.440987045346134e-05,
+      "loss": 0.8638,
+      "step": 507
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 8.405451062453744e-05,
+      "loss": 0.8462,
+      "step": 508
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 8.369935729018634e-05,
+      "loss": 0.8252,
+      "step": 509
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 8.334441504965455e-05,
+      "loss": 0.7397,
+      "step": 510
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 8.29896884994549e-05,
+      "loss": 0.8184,
+      "step": 511
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 8.263518223330697e-05,
+      "loss": 0.9321,
+      "step": 512
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 8.228090084207774e-05,
+      "loss": 0.8481,
+      "step": 513
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 8.192684891372198e-05,
+      "loss": 0.8745,
+      "step": 514
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 8.157303103322296e-05,
+      "loss": 0.8618,
+      "step": 515
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 8.1219451782533e-05,
+      "loss": 0.9106,
+      "step": 516
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 8.086611574051417e-05,
+      "loss": 0.8784,
+      "step": 517
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 8.051302748287895e-05,
+      "loss": 0.8667,
+      "step": 518
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 8.016019158213101e-05,
+      "loss": 0.8765,
+      "step": 519
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 7.980761260750607e-05,
+      "loss": 0.9463,
+      "step": 520
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 7.945529512491251e-05,
+      "loss": 0.9443,
+      "step": 521
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 7.91032436968725e-05,
+      "loss": 0.9102,
+      "step": 522
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 7.875146288246275e-05,
+      "loss": 0.9639,
+      "step": 523
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 7.839995723725548e-05,
+      "loss": 0.8325,
+      "step": 524
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 7.804873131325954e-05,
+      "loss": 0.9619,
+      "step": 525
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 7.76977896588613e-05,
+      "loss": 0.8569,
+      "step": 526
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 7.734713681876589e-05,
+      "loss": 0.8643,
+      "step": 527
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 7.699677733393826e-05,
+      "loss": 0.8789,
+      "step": 528
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 7.66467157415444e-05,
+      "loss": 0.8901,
+      "step": 529
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 7.629695657489257e-05,
+      "loss": 0.9199,
+      "step": 530
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 7.594750436337467e-05,
+      "loss": 0.9204,
+      "step": 531
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 7.55983636324074e-05,
+      "loss": 0.9624,
+      "step": 532
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 7.524953890337395e-05,
+      "loss": 0.9517,
+      "step": 533
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 7.490103469356513e-05,
+      "loss": 0.8242,
+      "step": 534
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 7.455285551612105e-05,
+      "loss": 0.979,
+      "step": 535
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 7.42050058799727e-05,
+      "loss": 0.9175,
+      "step": 536
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 7.385749028978346e-05,
+      "loss": 0.936,
+      "step": 537
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 7.351031324589074e-05,
+      "loss": 0.833,
+      "step": 538
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 7.316347924424787e-05,
+      "loss": 0.9663,
+      "step": 539
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 7.281699277636572e-05,
+      "loss": 0.8643,
+      "step": 540
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 7.24708583292546e-05,
+      "loss": 0.7593,
+      "step": 541
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 7.212508038536613e-05,
+      "loss": 0.7949,
+      "step": 542
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 7.177966342253524e-05,
+      "loss": 0.7622,
+      "step": 543
+    },
+    {
+      "epoch": 12.09,
+      "learning_rate": 7.143461191392206e-05,
+      "loss": 0.7656,
+      "step": 544
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 7.108993032795418e-05,
+      "loss": 0.749,
+      "step": 545
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 7.07456231282686e-05,
+      "loss": 0.7344,
+      "step": 546
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 7.040169477365403e-05,
+      "loss": 0.7153,
+      "step": 547
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 7.005814971799318e-05,
+      "loss": 0.7266,
+      "step": 548
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 6.971499241020495e-05,
+      "loss": 0.7949,
+      "step": 549
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 6.93722272941869e-05,
+      "loss": 0.7397,
+      "step": 550
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 6.902985880875773e-05,
+      "loss": 0.7422,
+      "step": 551
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 6.868789138759976e-05,
+      "loss": 0.7798,
+      "step": 552
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 6.83463294592015e-05,
+      "loss": 0.8408,
+      "step": 553
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 6.800517744680032e-05,
+      "loss": 0.8447,
+      "step": 554
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 6.766443976832517e-05,
+      "loss": 0.8174,
+      "step": 555
+    },
+    {
+      "epoch": 12.36,
+      "learning_rate": 6.732412083633936e-05,
+      "loss": 0.7407,
+      "step": 556
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 6.698422505798338e-05,
+      "loss": 0.7998,
+      "step": 557
+    },
+    {
+      "epoch": 12.4,
+      "learning_rate": 6.664475683491796e-05,
+      "loss": 0.8213,
+      "step": 558
+    },
+    {
+      "epoch": 12.42,
+      "learning_rate": 6.630572056326687e-05,
+      "loss": 0.7622,
+      "step": 559
+    },
+    {
+      "epoch": 12.44,
+      "learning_rate": 6.59671206335602e-05,
+      "loss": 0.8169,
+      "step": 560
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 6.562896143067734e-05,
+      "loss": 0.7935,
+      "step": 561
+    },
+    {
+      "epoch": 12.49,
+      "learning_rate": 6.529124733379024e-05,
+      "loss": 0.7744,
+      "step": 562
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 6.495398271630675e-05,
+      "loss": 0.835,
+      "step": 563
+    },
+    {
+      "epoch": 12.53,
+      "learning_rate": 6.461717194581393e-05,
+      "loss": 0.8018,
+      "step": 564
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 6.428081938402149e-05,
+      "loss": 0.7261,
+      "step": 565
+    },
+    {
+      "epoch": 12.58,
+      "learning_rate": 6.394492938670538e-05,
+      "loss": 0.8813,
+      "step": 566
+    },
+    {
+      "epoch": 12.6,
+      "learning_rate": 6.360950630365126e-05,
+      "loss": 0.8125,
+      "step": 567
+    },
+    {
+      "epoch": 12.62,
+      "learning_rate": 6.327455447859827e-05,
+      "loss": 0.8042,
+      "step": 568
+    },
+    {
+      "epoch": 12.64,
+      "learning_rate": 6.294007824918276e-05,
+      "loss": 0.7339,
+      "step": 569
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 6.260608194688206e-05,
+      "loss": 0.8389,
+      "step": 570
+    },
+    {
+      "epoch": 12.69,
+      "learning_rate": 6.227256989695848e-05,
+      "loss": 0.8213,
+      "step": 571
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 6.193954641840318e-05,
+      "loss": 0.7446,
+      "step": 572
+    },
+    {
+      "epoch": 12.73,
+      "learning_rate": 6.160701582388038e-05,
+      "loss": 0.8755,
+      "step": 573
+    },
+    {
+      "epoch": 12.76,
+      "learning_rate": 6.12749824196714e-05,
+      "loss": 0.7798,
+      "step": 574
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 6.0943450505618917e-05,
+      "loss": 0.8774,
+      "step": 575
+    },
+    {
+      "epoch": 12.8,
+      "learning_rate": 6.061242437507131e-05,
+      "loss": 0.7964,
+      "step": 576
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 6.028190831482703e-05,
+      "loss": 0.793,
+      "step": 577
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 5.995190660507915e-05,
+      "loss": 0.8257,
+      "step": 578
+    },
+    {
+      "epoch": 12.87,
+      "learning_rate": 5.962242351935985e-05,
+      "loss": 0.8042,
+      "step": 579
+    },
+    {
+      "epoch": 12.89,
+      "learning_rate": 5.929346332448511e-05,
+      "loss": 0.832,
+      "step": 580
+    },
+    {
+      "epoch": 12.91,
+      "learning_rate": 5.89650302804995e-05,
+      "loss": 0.8218,
+      "step": 581
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 5.863712864062089e-05,
+      "loss": 0.8047,
+      "step": 582
+    },
+    {
+      "epoch": 12.96,
+      "learning_rate": 5.8309762651185484e-05,
+      "loss": 0.8052,
+      "step": 583
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 5.7982936551592906e-05,
+      "loss": 0.791,
+      "step": 584
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 5.765665457425102e-05,
+      "loss": 0.8032,
+      "step": 585
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 5.733092094452135e-05,
+      "loss": 0.7715,
+      "step": 586
+    },
+    {
+      "epoch": 13.04,
+      "learning_rate": 5.700573988066433e-05,
+      "loss": 0.7065,
+      "step": 587
+    },
+    {
+      "epoch": 13.07,
+      "learning_rate": 5.668111559378471e-05,
+      "loss": 0.7202,
+      "step": 588
+    },
+    {
+      "epoch": 13.09,
+      "learning_rate": 5.6357052287776765e-05,
+      "loss": 0.7153,
+      "step": 589
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 5.6033554159270294e-05,
+      "loss": 0.7456,
+      "step": 590
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 5.571062539757581e-05,
+      "loss": 0.7344,
+      "step": 591
+    },
+    {
+      "epoch": 13.16,
+      "learning_rate": 5.538827018463069e-05,
+      "loss": 0.6982,
+      "step": 592
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 5.50664926949447e-05,
+      "loss": 0.7637,
+      "step": 593
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 5.474529709554612e-05,
+      "loss": 0.7227,
+      "step": 594
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 5.4424687545927776e-05,
+      "loss": 0.6914,
+      "step": 595
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 5.410466819799306e-05,
+      "loss": 0.7603,
+      "step": 596
+    },
+    {
+      "epoch": 13.27,
+      "learning_rate": 5.378524319600231e-05,
+      "loss": 0.707,
+      "step": 597
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 5.346641667651897e-05,
+      "loss": 0.7505,
+      "step": 598
+    },
+    {
+      "epoch": 13.31,
+      "learning_rate": 5.314819276835625e-05,
+      "loss": 0.7998,
+      "step": 599
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 5.283057559252341e-05,
+      "loss": 0.7847,
+      "step": 600
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 5.25135692621725e-05,
+      "loss": 0.7686,
+      "step": 601
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 5.219717788254521e-05,
+      "loss": 0.687,
+      "step": 602
+    },
+    {
+      "epoch": 13.4,
+      "learning_rate": 5.1881405550919493e-05,
+      "loss": 0.7051,
+      "step": 603
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 5.156625635655672e-05,
+      "loss": 0.7412,
+      "step": 604
+    },
+    {
+      "epoch": 13.44,
+      "learning_rate": 5.12517343806485e-05,
+      "loss": 0.729,
+      "step": 605
+    },
+    {
+      "epoch": 13.47,
+      "learning_rate": 5.0937843696263966e-05,
+      "loss": 0.7505,
+      "step": 606
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 5.062458836829711e-05,
+      "loss": 0.7002,
+      "step": 607
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 5.03119724534139e-05,
+      "loss": 0.7573,
+      "step": 608
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 5.000000000000002e-05,
+      "loss": 0.7241,
+      "step": 609
+    },
+    {
+      "epoch": 13.56,
+      "learning_rate": 4.96886750481082e-05,
+      "loss": 0.6753,
+      "step": 610
+    },
+    {
+      "epoch": 13.58,
+      "learning_rate": 4.9378001629406e-05,
+      "loss": 0.7456,
+      "step": 611
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 4.9067983767123736e-05,
+      "loss": 0.7471,
+      "step": 612
+    },
+    {
+      "epoch": 13.62,
+      "learning_rate": 4.875862547600207e-05,
+      "loss": 0.6768,
+      "step": 613
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 4.8449930762240355e-05,
+      "loss": 0.7339,
+      "step": 614
+    },
+    {
+      "epoch": 13.67,
+      "learning_rate": 4.814190362344454e-05,
+      "loss": 0.7017,
+      "step": 615
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 4.783454804857539e-05,
+      "loss": 0.8296,
+      "step": 616
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 4.752786801789703e-05,
+      "loss": 0.7427,
+      "step": 617
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 4.722186750292511e-05,
+      "loss": 0.7837,
+      "step": 618
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 4.6916550466375684e-05,
+      "loss": 0.7305,
+      "step": 619
+    },
+    {
+      "epoch": 13.78,
+      "learning_rate": 4.661192086211366e-05,
+      "loss": 0.7734,
+      "step": 620
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 4.630798263510162e-05,
+      "loss": 0.7188,
+      "step": 621
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 4.600473972134894e-05,
+      "loss": 0.7505,
+      "step": 622
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 4.570219604786051e-05,
+      "loss": 0.6426,
+      "step": 623
+    },
+    {
+      "epoch": 13.87,
+      "learning_rate": 4.540035553258619e-05,
+      "loss": 0.7739,
+      "step": 624
+    },
+    {
+      "epoch": 13.89,
+      "learning_rate": 4.5099222084369805e-05,
+      "loss": 0.7842,
+      "step": 625
+    },
+    {
+      "epoch": 13.91,
+      "learning_rate": 4.479879960289863e-05,
+      "loss": 0.7319,
+      "step": 626
+    },
+    {
+      "epoch": 13.93,
+      "learning_rate": 4.449909197865303e-05,
+      "loss": 0.7788,
+      "step": 627
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 4.420010309285577e-05,
+      "loss": 0.7427,
+      "step": 628
+    },
+    {
+      "epoch": 13.98,
+      "learning_rate": 4.3901836817422124e-05,
+      "loss": 0.7358,
+      "step": 629
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 4.360429701490934e-05,
+      "loss": 0.7964,
+      "step": 630
+    },
+    {
+      "epoch": 14.02,
+      "learning_rate": 4.3307487538467006e-05,
+      "loss": 0.7236,
+      "step": 631
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 4.301141223178684e-05,
+      "loss": 0.689,
+      "step": 632
+    },
+    {
+      "epoch": 14.07,
+      "learning_rate": 4.271607492905303e-05,
+      "loss": 0.6389,
+      "step": 633
+    },
+    {
+      "epoch": 14.09,
+      "learning_rate": 4.242147945489272e-05,
+      "loss": 0.6812,
+      "step": 634
+    },
+    {
+      "epoch": 14.11,
+      "learning_rate": 4.212762962432619e-05,
+      "loss": 0.6782,
+      "step": 635
+    },
+    {
+      "epoch": 14.13,
+      "learning_rate": 4.183452924271776e-05,
+      "loss": 0.6899,
+      "step": 636
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 4.154218210572627e-05,
+      "loss": 0.7407,
+      "step": 637
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 4.125059199925599e-05,
+      "loss": 0.6567,
+      "step": 638
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 4.0959762699407766e-05,
+      "loss": 0.6587,
+      "step": 639
+    },
+    {
+      "epoch": 14.22,
+      "learning_rate": 4.06696979724298e-05,
+      "loss": 0.7441,
+      "step": 640
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 4.038040157466918e-05,
+      "loss": 0.707,
+      "step": 641
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 4.009187725252309e-05,
+      "loss": 0.7041,
+      "step": 642
+    },
+    {
+      "epoch": 14.29,
+      "learning_rate": 3.980412874239021e-05,
+      "loss": 0.6216,
+      "step": 643
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 3.95171597706226e-05,
+      "loss": 0.6748,
+      "step": 644
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 3.9230974053477086e-05,
+      "loss": 0.6362,
+      "step": 645
+    },
+    {
+      "epoch": 14.36,
+      "learning_rate": 3.8945575297067506e-05,
+      "loss": 0.6045,
+      "step": 646
+    },
+    {
+      "epoch": 14.38,
+      "learning_rate": 3.866096719731639e-05,
+      "loss": 0.7173,
+      "step": 647
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 3.8377153439907266e-05,
+      "loss": 0.6299,
+      "step": 648
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 3.809413770023701e-05,
+      "loss": 0.6499,
+      "step": 649
+    },
+    {
+      "epoch": 14.44,
+      "learning_rate": 3.7811923643367974e-05,
+      "loss": 0.6914,
+      "step": 650
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 3.7530514923980884e-05,
+      "loss": 0.7236,
+      "step": 651
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 3.724991518632717e-05,
+      "loss": 0.7598,
+      "step": 652
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 3.697012806418194e-05,
+      "loss": 0.6279,
+      "step": 653
+    },
+    {
+      "epoch": 14.53,
+      "learning_rate": 3.669115718079702e-05,
+      "loss": 0.7622,
+      "step": 654
+    },
+    {
+      "epoch": 14.56,
+      "learning_rate": 3.641300614885378e-05,
+      "loss": 0.6772,
+      "step": 655
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 3.61356785704166e-05,
+      "loss": 0.7568,
+      "step": 656
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 3.585917803688603e-05,
+      "loss": 0.7065,
+      "step": 657
+    },
+    {
+      "epoch": 14.62,
+      "learning_rate": 3.558350812895238e-05,
+      "loss": 0.6743,
+      "step": 658
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 3.530867241654942e-05,
+      "loss": 0.6709,
+      "step": 659
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 3.503467445880789e-05,
+      "loss": 0.71,
+      "step": 660
+    },
+    {
+      "epoch": 14.69,
+      "learning_rate": 3.476151780400979e-05,
+      "loss": 0.6836,
+      "step": 661
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 3.448920598954203e-05,
+      "loss": 0.6973,
+      "step": 662
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 3.421774254185096e-05,
+      "loss": 0.7256,
+      "step": 663
+    },
+    {
+      "epoch": 14.76,
+      "learning_rate": 3.394713097639647e-05,
+      "loss": 0.7334,
+      "step": 664
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 3.367737479760652e-05,
+      "loss": 0.6675,
+      "step": 665
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 3.340847749883191e-05,
+      "loss": 0.7363,
+      "step": 666
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 3.31404425623008e-05,
+      "loss": 0.7139,
+      "step": 667
+    },
+    {
+      "epoch": 14.84,
+      "learning_rate": 3.287327345907381e-05,
+      "loss": 0.6792,
+      "step": 668
+    },
+    {
+      "epoch": 14.87,
+      "learning_rate": 3.2606973648998915e-05,
+      "loss": 0.6704,
+      "step": 669
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 3.2341546580666796e-05,
+      "loss": 0.7329,
+      "step": 670
+    },
+    {
+      "epoch": 14.91,
+      "learning_rate": 3.207699569136608e-05,
+      "loss": 0.709,
+      "step": 671
+    },
+    {
+      "epoch": 14.93,
+      "learning_rate": 3.1813324407038825e-05,
+      "loss": 0.6294,
+      "step": 672
+    },
+    {
+      "epoch": 14.96,
+      "learning_rate": 3.1550536142236145e-05,
+      "loss": 0.7065,
+      "step": 673
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 3.128863430007414e-05,
+      "loss": 0.7358,
+      "step": 674
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 3.102762227218957e-05,
+      "loss": 0.7622,
+      "step": 675
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 3.0767503438696213e-05,
+      "loss": 0.6484,
+      "step": 676
+    },
+    {
+      "epoch": 15.04,
+      "learning_rate": 3.0508281168140806e-05,
+      "loss": 0.6318,
+      "step": 677
+    },
+    {
+      "epoch": 15.07,
+      "learning_rate": 3.0249958817459722e-05,
+      "loss": 0.6851,
+      "step": 678
+    },
+    {
+      "epoch": 15.09,
+      "learning_rate": 2.999253973193522e-05,
+      "loss": 0.6123,
+      "step": 679
+    },
+    {
+      "epoch": 15.11,
+      "learning_rate": 2.9736027245152275e-05,
+      "loss": 0.603,
+      "step": 680
+    },
+    {
+      "epoch": 15.13,
+      "learning_rate": 2.9480424678955443e-05,
+      "loss": 0.6152,
+      "step": 681
+    },
+    {
+      "epoch": 15.16,
+      "learning_rate": 2.9225735343405693e-05,
+      "loss": 0.6743,
+      "step": 682
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 2.897196253673773e-05,
+      "loss": 0.6562,
+      "step": 683
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 2.8719109545317103e-05,
+      "loss": 0.6621,
+      "step": 684
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 2.8467179643597697e-05,
+      "loss": 0.6233,
+      "step": 685
+    },
+    {
+      "epoch": 15.24,
+      "learning_rate": 2.8216176094079482e-05,
+      "loss": 0.6812,
+      "step": 686
+    },
+    {
+      "epoch": 15.27,
+      "learning_rate": 2.7966102147265994e-05,
+      "loss": 0.687,
+      "step": 687
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 2.7716961041622534e-05,
+      "loss": 0.6611,
+      "step": 688
+    },
+    {
+      "epoch": 15.31,
+      "learning_rate": 2.746875600353398e-05,
+      "loss": 0.6675,
+      "step": 689
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 2.722149024726307e-05,
+      "loss": 0.6821,
+      "step": 690
+    },
+    {
+      "epoch": 15.36,
+      "learning_rate": 2.697516697490896e-05,
+      "loss": 0.6733,
+      "step": 691
+    },
+    {
+      "epoch": 15.38,
+      "learning_rate": 2.6729789376365456e-05,
+      "loss": 0.6606,
+      "step": 692
+    },
+    {
+      "epoch": 15.4,
+      "learning_rate": 2.6485360629279987e-05,
+      "loss": 0.6973,
+      "step": 693
+    },
+    {
+      "epoch": 15.42,
+      "learning_rate": 2.624188389901221e-05,
+      "loss": 0.6748,
+      "step": 694
+    },
+    {
+      "epoch": 15.44,
+      "learning_rate": 2.599936233859326e-05,
+      "loss": 0.7324,
+      "step": 695
+    },
+    {
+      "epoch": 15.47,
+      "learning_rate": 2.5757799088684654e-05,
+      "loss": 0.6538,
+      "step": 696
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 2.5517197277537886e-05,
+      "loss": 0.6245,
+      "step": 697
+    },
+    {
+      "epoch": 15.51,
+      "learning_rate": 2.527756002095373e-05,
+      "loss": 0.5972,
+      "step": 698
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 2.5038890422241958e-05,
+      "loss": 0.6587,
+      "step": 699
+    },
+    {
+      "epoch": 15.56,
+      "learning_rate": 2.480119157218108e-05,
+      "loss": 0.6523,
+      "step": 700
+    },
+    {
+      "epoch": 15.58,
+      "learning_rate": 2.4564466548978525e-05,
+      "loss": 0.5957,
+      "step": 701
+    },
+    {
+      "epoch": 15.6,
+      "learning_rate": 2.432871841823047e-05,
+      "loss": 0.625,
+      "step": 702
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 2.4093950232882456e-05,
+      "loss": 0.6392,
+      "step": 703
+    },
+    {
+      "epoch": 15.64,
+      "learning_rate": 2.3860165033189587e-05,
+      "loss": 0.6982,
+      "step": 704
+    },
+    {
+      "epoch": 15.67,
+      "learning_rate": 2.3627365846677306e-05,
+      "loss": 0.6118,
+      "step": 705
+    },
+    {
+      "epoch": 15.69,
+      "learning_rate": 2.339555568810221e-05,
+      "loss": 0.6895,
+      "step": 706
+    },
+    {
+      "epoch": 15.71,
+      "learning_rate": 2.3164737559412854e-05,
+      "loss": 0.6025,
+      "step": 707
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 2.2934914449711087e-05,
+      "loss": 0.6487,
+      "step": 708
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 2.2706089335213122e-05,
+      "loss": 0.6846,
+      "step": 709
+    },
+    {
+      "epoch": 15.78,
+      "learning_rate": 2.247826517921121e-05,
+      "loss": 0.6377,
+      "step": 710
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 2.2251444932035094e-05,
+      "loss": 0.6797,
+      "step": 711
+    },
+    {
+      "epoch": 15.82,
+      "learning_rate": 2.2025631531013824e-05,
+      "loss": 0.7083,
+      "step": 712
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 2.1800827900437894e-05,
+      "loss": 0.7168,
+      "step": 713
+    },
+    {
+      "epoch": 15.87,
+      "learning_rate": 2.157703695152109e-05,
+      "loss": 0.6914,
+      "step": 714
+    },
+    {
+      "epoch": 15.89,
+      "learning_rate": 2.135426158236309e-05,
+      "loss": 0.6123,
+      "step": 715
+    },
+    {
+      "epoch": 15.91,
+      "learning_rate": 2.1132504677911658e-05,
+      "loss": 0.6733,
+      "step": 716
+    },
+    {
+      "epoch": 15.93,
+      "learning_rate": 2.091176910992545e-05,
+      "loss": 0.645,
+      "step": 717
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 2.069205773693683e-05,
+      "loss": 0.7017,
+      "step": 718
+    },
+    {
+      "epoch": 15.98,
+      "learning_rate": 2.0473373404214723e-05,
+      "loss": 0.6704,
+      "step": 719
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.025571894372794e-05,
+      "loss": 0.6509,
+      "step": 720
+    },
+    {
+      "epoch": 16.02,
+      "learning_rate": 2.003909717410831e-05,
+      "loss": 0.6914,
+      "step": 721
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 1.9823510900614417e-05,
+      "loss": 0.6296,
+      "step": 722
+    },
+    {
+      "epoch": 16.07,
+      "learning_rate": 1.9608962915094996e-05,
+      "loss": 0.6333,
+      "step": 723
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 1.9395455995953036e-05,
+      "loss": 0.6514,
+      "step": 724
+    },
+    {
+      "epoch": 16.11,
+      "learning_rate": 1.9182992908109644e-05,
+      "loss": 0.6104,
+      "step": 725
+    },
+    {
+      "epoch": 16.13,
+      "learning_rate": 1.897157640296825e-05,
+      "loss": 0.6318,
+      "step": 726
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 1.8761209218379016e-05,
+      "loss": 0.606,
+      "step": 727
+    },
+    {
+      "epoch": 16.18,
+      "learning_rate": 1.855189407860344e-05,
+      "loss": 0.6172,
+      "step": 728
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 1.8343633694278895e-05,
+      "loss": 0.6108,
+      "step": 729
+    },
+    {
+      "epoch": 16.22,
+      "learning_rate": 1.813643076238375e-05,
+      "loss": 0.6021,
+      "step": 730
+    },
+    {
+      "epoch": 16.24,
+      "learning_rate": 1.7930287966202265e-05,
+      "loss": 0.5974,
+      "step": 731
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 1.772520797528988e-05,
+      "loss": 0.6309,
+      "step": 732
+    },
+    {
+      "epoch": 16.29,
+      "learning_rate": 1.752119344543879e-05,
+      "loss": 0.623,
+      "step": 733
+    },
+    {
+      "epoch": 16.31,
+      "learning_rate": 1.731824701864331e-05,
+      "loss": 0.668,
+      "step": 734
+    },
+    {
+      "epoch": 16.33,
+      "learning_rate": 1.7116371323065883e-05,
+      "loss": 0.5984,
+      "step": 735
+    },
+    {
+      "epoch": 16.36,
+      "learning_rate": 1.6915568973002905e-05,
+      "loss": 0.6904,
+      "step": 736
+    },
+    {
+      "epoch": 16.38,
+      "learning_rate": 1.6715842568850893e-05,
+      "loss": 0.627,
+      "step": 737
+    },
+    {
+      "epoch": 16.4,
+      "learning_rate": 1.65171946970729e-05,
+      "loss": 0.5972,
+      "step": 738
+    },
+    {
+      "epoch": 16.42,
+      "learning_rate": 1.631962793016487e-05,
+      "loss": 0.623,
+      "step": 739
+    },
+    {
+      "epoch": 16.44,
+      "learning_rate": 1.6123144826622504e-05,
+      "loss": 0.6499,
+      "step": 740
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 1.592774793090792e-05,
+      "loss": 0.6597,
+      "step": 741
+    },
+    {
+      "epoch": 16.49,
+      "learning_rate": 1.5733439773416915e-05,
+      "loss": 0.6045,
+      "step": 742
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 1.554022287044602e-05,
+      "loss": 0.6943,
+      "step": 743
+    },
+    {
+      "epoch": 16.53,
+      "learning_rate": 1.534809972415998e-05,
+      "loss": 0.5886,
+      "step": 744
+    },
+    {
+      "epoch": 16.56,
+      "learning_rate": 1.5157072822559437e-05,
+      "loss": 0.6047,
+      "step": 745
+    },
+    {
+      "epoch": 16.58,
+      "learning_rate": 1.4967144639448538e-05,
+      "loss": 0.6401,
+      "step": 746
+    },
+    {
+      "epoch": 16.6,
+      "learning_rate": 1.4778317634403083e-05,
+      "loss": 0.6108,
+      "step": 747
+    },
+    {
+      "epoch": 16.62,
+      "learning_rate": 1.4590594252738522e-05,
+      "loss": 0.6514,
+      "step": 748
+    },
+    {
+      "epoch": 16.64,
+      "learning_rate": 1.4403976925478312e-05,
+      "loss": 0.6289,
+      "step": 749
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 1.4218468069322578e-05,
+      "loss": 0.6323,
+      "step": 750
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 1.4034070086616647e-05,
+      "loss": 0.6553,
+      "step": 751
+    },
+    {
+      "epoch": 16.71,
+      "learning_rate": 1.3850785365319984e-05,
+      "loss": 0.6082,
+      "step": 752
+    },
+    {
+      "epoch": 16.73,
+      "learning_rate": 1.3668616278975343e-05,
+      "loss": 0.6484,
+      "step": 753
+    },
+    {
+      "epoch": 16.76,
+      "learning_rate": 1.3487565186677897e-05,
+      "loss": 0.6587,
+      "step": 754
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 1.3307634433044846e-05,
+      "loss": 0.6204,
+      "step": 755
+    },
+    {
+      "epoch": 16.8,
+      "learning_rate": 1.3128826348184887e-05,
+      "loss": 0.6255,
+      "step": 756
+    },
+    {
+      "epoch": 16.82,
+      "learning_rate": 1.2951143247668197e-05,
+      "loss": 0.5898,
+      "step": 757
+    },
+    {
+      "epoch": 16.84,
+      "learning_rate": 1.2774587432496321e-05,
+      "loss": 0.624,
+      "step": 758
+    },
+    {
+      "epoch": 16.87,
+      "learning_rate": 1.2599161189072427e-05,
+      "loss": 0.6282,
+      "step": 759
+    },
+    {
+      "epoch": 16.89,
+      "learning_rate": 1.2424866789171729e-05,
+      "loss": 0.6191,
+      "step": 760
+    },
+    {
+      "epoch": 16.91,
+      "learning_rate": 1.2251706489911984e-05,
+      "loss": 0.6267,
+      "step": 761
+    },
+    {
+      "epoch": 16.93,
+      "learning_rate": 1.2079682533724379e-05,
+      "loss": 0.6519,
+      "step": 762
+    },
+    {
+      "epoch": 16.96,
+      "learning_rate": 1.1908797148324358e-05,
+      "loss": 0.6577,
+      "step": 763
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 1.173905254668285e-05,
+      "loss": 0.6426,
+      "step": 764
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 1.1570450926997655e-05,
+      "loss": 0.6863,
+      "step": 765
+    },
+    {
+      "epoch": 17.02,
+      "learning_rate": 1.140299447266483e-05,
+      "loss": 0.6089,
+      "step": 766
+    },
+    {
+      "epoch": 17.04,
+      "learning_rate": 1.1236685352250597e-05,
+      "loss": 0.6086,
+      "step": 767
+    },
+    {
+      "epoch": 17.07,
+      "learning_rate": 1.1071525719463095e-05,
+      "loss": 0.6118,
+      "step": 768
+    },
+    {
+      "epoch": 17.09,
+      "learning_rate": 1.0907517713124638e-05,
+      "loss": 0.5825,
+      "step": 769
+    },
+    {
+      "epoch": 17.11,
+      "learning_rate": 1.0744663457143878e-05,
+      "loss": 0.6562,
+      "step": 770
+    },
+    {
+      "epoch": 17.13,
+      "learning_rate": 1.0582965060488359e-05,
+      "loss": 0.5364,
+      "step": 771
+    },
+    {
+      "epoch": 17.16,
+      "learning_rate": 1.042242461715729e-05,
+      "loss": 0.5894,
+      "step": 772
+    },
+    {
+      "epoch": 17.18,
+      "learning_rate": 1.026304420615426e-05,
+      "loss": 0.5815,
+      "step": 773
+    },
+    {
+      "epoch": 17.2,
+      "learning_rate": 1.010482589146048e-05,
+      "loss": 0.606,
+      "step": 774
+    },
+    {
+      "epoch": 17.22,
+      "learning_rate": 9.947771722007915e-06,
+      "loss": 0.5908,
+      "step": 775
+    },
+    {
+      "epoch": 17.24,
+      "learning_rate": 9.791883731652828e-06,
+      "loss": 0.5938,
+      "step": 776
+    },
+    {
+      "epoch": 17.27,
+      "learning_rate": 9.637163939149485e-06,
+      "loss": 0.594,
+      "step": 777
+    },
+    {
+      "epoch": 17.29,
+      "learning_rate": 9.48361434812386e-06,
+      "loss": 0.6211,
+      "step": 778
+    },
+    {
+      "epoch": 17.31,
+      "learning_rate": 9.33123694704784e-06,
+      "loss": 0.6226,
+      "step": 779
+    },
+    {
+      "epoch": 17.33,
+      "learning_rate": 9.180033709213454e-06,
+      "loss": 0.573,
+      "step": 780
+    },
+    {
+      "epoch": 17.36,
+      "learning_rate": 9.030006592707174e-06,
+      "loss": 0.5986,
+      "step": 781
+    },
+    {
+      "epoch": 17.38,
+      "learning_rate": 8.881157540384777e-06,
+      "loss": 0.6294,
+      "step": 782
+    },
+    {
+      "epoch": 17.4,
+      "learning_rate": 8.733488479845997e-06,
+      "loss": 0.627,
+      "step": 783
+    },
+    {
+      "epoch": 17.42,
+      "learning_rate": 8.587001323409638e-06,
+      "loss": 0.5654,
+      "step": 784
+    },
+    {
+      "epoch": 17.44,
+      "learning_rate": 8.441697968088891e-06,
+      "loss": 0.5955,
+      "step": 785
+    },
+    {
+      "epoch": 17.47,
+      "learning_rate": 8.297580295566575e-06,
+      "loss": 0.6597,
+      "step": 786
+    },
+    {
+      "epoch": 17.49,
+      "learning_rate": 8.154650172170975e-06,
+      "loss": 0.6575,
+      "step": 787
+    },
+    {
+      "epoch": 17.51,
+      "learning_rate": 8.012909448851514e-06,
+      "loss": 0.6304,
+      "step": 788
+    },
+    {
+      "epoch": 17.53,
+      "learning_rate": 7.872359961154906e-06,
+      "loss": 0.6387,
+      "step": 789
+    },
+    {
+      "epoch": 17.56,
+      "learning_rate": 7.733003529201278e-06,
+      "loss": 0.6089,
+      "step": 790
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 7.594841957660637e-06,
+      "loss": 0.6738,
+      "step": 791
+    },
+    {
+      "epoch": 17.6,
+      "learning_rate": 7.457877035729588e-06,
+      "loss": 0.6523,
+      "step": 792
+    },
+    {
+      "epoch": 17.62,
+      "learning_rate": 7.322110537108007e-06,
+      "loss": 0.627,
+      "step": 793
+    },
+    {
+      "epoch": 17.64,
+      "learning_rate": 7.187544219976205e-06,
+      "loss": 0.6855,
+      "step": 794
+    },
+    {
+      "epoch": 17.67,
+      "learning_rate": 7.054179826972074e-06,
+      "loss": 0.5837,
+      "step": 795
+    },
+    {
+      "epoch": 17.69,
+      "learning_rate": 6.9220190851685516e-06,
+      "loss": 0.6162,
+      "step": 796
+    },
+    {
+      "epoch": 17.71,
+      "learning_rate": 6.7910637060512924e-06,
+      "loss": 0.6091,
+      "step": 797
+    },
+    {
+      "epoch": 17.73,
+      "learning_rate": 6.661315385496425e-06,
+      "loss": 0.5933,
+      "step": 798
+    },
+    {
+      "epoch": 17.76,
+      "learning_rate": 6.5327758037486585e-06,
+      "loss": 0.5962,
+      "step": 799
+    },
+    {
+      "epoch": 17.78,
+      "learning_rate": 6.405446625399481e-06,
+      "loss": 0.6714,
+      "step": 800
+    },
+    {
+      "epoch": 17.8,
+      "learning_rate": 6.2793294993656494e-06,
+      "loss": 0.5962,
+      "step": 801
+    },
+    {
+      "epoch": 17.82,
+      "learning_rate": 6.1544260588677575e-06,
+      "loss": 0.6211,
+      "step": 802
+    },
+    {
+      "epoch": 17.84,
+      "learning_rate": 6.030737921409169e-06,
+      "loss": 0.6191,
+      "step": 803
+    },
+    {
+      "epoch": 17.87,
+      "learning_rate": 5.908266688755049e-06,
+      "loss": 0.6064,
+      "step": 804
+    },
+    {
+      "epoch": 17.89,
+      "learning_rate": 5.787013946911546e-06,
+      "loss": 0.6123,
+      "step": 805
+    },
+    {
+      "epoch": 17.91,
+      "learning_rate": 5.666981266105398e-06,
+      "loss": 0.6287,
+      "step": 806
+    },
+    {
+      "epoch": 17.93,
+      "learning_rate": 5.54817020076347e-06,
+      "loss": 0.6648,
+      "step": 807
+    },
+    {
+      "epoch": 17.96,
+      "learning_rate": 5.430582289492659e-06,
+      "loss": 0.6086,
+      "step": 808
+    },
+    {
+      "epoch": 17.98,
+      "learning_rate": 5.314219055060022e-06,
+      "loss": 0.6147,
+      "step": 809
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 5.199082004372957e-06,
+      "loss": 0.6768,
+      "step": 810
+    },
+    {
+      "epoch": 18.02,
+      "learning_rate": 5.085172628459778e-06,
+      "loss": 0.6069,
+      "step": 811
+    },
+    {
+      "epoch": 18.04,
+      "learning_rate": 4.972492402450402e-06,
+      "loss": 0.6294,
+      "step": 812
+    },
+    {
+      "epoch": 18.07,
+      "learning_rate": 4.861042785557146e-06,
+      "loss": 0.5703,
+      "step": 813
+    },
+    {
+      "epoch": 18.09,
+      "learning_rate": 4.750825221055965e-06,
+      "loss": 0.5938,
+      "step": 814
+    },
+    {
+      "epoch": 18.11,
+      "learning_rate": 4.641841136267666e-06,
+      "loss": 0.6108,
+      "step": 815
+    },
+    {
+      "epoch": 18.13,
+      "learning_rate": 4.534091942539475e-06,
+      "loss": 0.605,
+      "step": 816
+    },
+    {
+      "epoch": 18.16,
+      "learning_rate": 4.427579035226725e-06,
+      "loss": 0.5735,
+      "step": 817
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 4.322303793674798e-06,
+      "loss": 0.603,
+      "step": 818
+    },
+    {
+      "epoch": 18.2,
+      "learning_rate": 4.2182675812012965e-06,
+      "loss": 0.572,
+      "step": 819
+    },
+    {
+      "epoch": 18.22,
+      "learning_rate": 4.115471745078314e-06,
+      "loss": 0.6404,
+      "step": 820
+    },
+    {
+      "epoch": 18.24,
+      "learning_rate": 4.0139176165150835e-06,
+      "loss": 0.5867,
+      "step": 821
+    },
+    {
+      "epoch": 18.27,
+      "learning_rate": 3.913606510640644e-06,
+      "loss": 0.5813,
+      "step": 822
+    },
+    {
+      "epoch": 18.29,
+      "learning_rate": 3.8145397264868656e-06,
+      "loss": 0.645,
+      "step": 823
+    },
+    {
+      "epoch": 18.31,
+      "learning_rate": 3.7167185469716426e-06,
+      "loss": 0.6445,
+      "step": 824
+    },
+    {
+      "epoch": 18.33,
+      "learning_rate": 3.620144238882206e-06,
+      "loss": 0.5747,
+      "step": 825
+    },
+    {
+      "epoch": 18.36,
+      "learning_rate": 3.5248180528588024e-06,
+      "loss": 0.5996,
+      "step": 826
+    },
+    {
+      "epoch": 18.38,
+      "learning_rate": 3.4307412233784308e-06,
+      "loss": 0.5754,
+      "step": 827
+    },
+    {
+      "epoch": 18.4,
+      "learning_rate": 3.3379149687388867e-06,
+      "loss": 0.6387,
+      "step": 828
+    },
+    {
+      "epoch": 18.42,
+      "learning_rate": 3.2463404910430206e-06,
+      "loss": 0.5884,
+      "step": 829
+    },
+    {
+      "epoch": 18.44,
+      "learning_rate": 3.1560189761830728e-06,
+      "loss": 0.6265,
+      "step": 830
+    },
+    {
+      "epoch": 18.47,
+      "learning_rate": 3.06695159382544e-06,
+      "loss": 0.572,
+      "step": 831
+    },
+    {
+      "epoch": 18.49,
+      "learning_rate": 2.9791394973954225e-06,
+      "loss": 0.583,
+      "step": 832
+    },
+    {
+      "epoch": 18.51,
+      "learning_rate": 2.892583824062334e-06,
+      "loss": 0.5796,
+      "step": 833
+    },
+    {
+      "epoch": 18.53,
+      "learning_rate": 2.8072856947248037e-06,
+      "loss": 0.623,
+      "step": 834
+    },
+    {
+      "epoch": 18.56,
+      "learning_rate": 2.723246213996178e-06,
+      "loss": 0.5913,
+      "step": 835
+    },
+    {
+      "epoch": 18.58,
+      "learning_rate": 2.640466470190317e-06,
+      "loss": 0.5737,
+      "step": 836
+    },
+    {
+      "epoch": 18.6,
+      "learning_rate": 2.5589475353073988e-06,
+      "loss": 0.6025,
+      "step": 837
+    },
+    {
+      "epoch": 18.62,
+      "learning_rate": 2.478690465020117e-06,
+      "loss": 0.6367,
+      "step": 838
+    },
+    {
+      "epoch": 18.64,
+      "learning_rate": 2.3996962986599814e-06,
+      "loss": 0.5872,
+      "step": 839
+    },
+    {
+      "epoch": 18.67,
+      "learning_rate": 2.3219660592038285e-06,
+      "loss": 0.5713,
+      "step": 840
+    },
+    {
+      "epoch": 18.69,
+      "learning_rate": 2.2455007532606655e-06,
+      "loss": 0.6096,
+      "step": 841
+    },
+    {
+      "epoch": 18.71,
+      "learning_rate": 2.170301371058503e-06,
+      "loss": 0.6313,
+      "step": 842
+    },
+    {
+      "epoch": 18.73,
+      "learning_rate": 2.0963688864316323e-06,
+      "loss": 0.6387,
+      "step": 843
+    },
+    {
+      "epoch": 18.76,
+      "learning_rate": 2.0237042568080012e-06,
+      "loss": 0.6155,
+      "step": 844
+    },
+    {
+      "epoch": 18.78,
+      "learning_rate": 1.9523084231967358e-06,
+      "loss": 0.6047,
+      "step": 845
+    },
+    {
+      "epoch": 18.8,
+      "learning_rate": 1.882182310176095e-06,
+      "loss": 0.5659,
+      "step": 846
+    },
+    {
+      "epoch": 18.82,
+      "learning_rate": 1.8133268258813563e-06,
+      "loss": 0.6431,
+      "step": 847
+    },
+    {
+      "epoch": 18.84,
+      "learning_rate": 1.745742861993138e-06,
+      "loss": 0.5503,
+      "step": 848
+    },
+    {
+      "epoch": 18.87,
+      "learning_rate": 1.6794312937258417e-06,
+      "loss": 0.647,
+      "step": 849
+    },
+    {
+      "epoch": 18.89,
+      "learning_rate": 1.6143929798162704e-06,
+      "loss": 0.5608,
+      "step": 850
+    },
+    {
+      "epoch": 18.91,
+      "learning_rate": 1.550628762512596e-06,
+      "loss": 0.5947,
+      "step": 851
+    },
+    {
+      "epoch": 18.93,
+      "learning_rate": 1.488139467563354e-06,
+      "loss": 0.624,
+      "step": 852
+    },
+    {
+      "epoch": 18.96,
+      "learning_rate": 1.4269259042068327e-06,
+      "loss": 0.5898,
+      "step": 853
+    },
+    {
+      "epoch": 18.98,
+      "learning_rate": 1.3669888651605345e-06,
+      "loss": 0.6182,
+      "step": 854
+    },
+    {
+      "epoch": 19.0,
+      "learning_rate": 1.30832912661093e-06,
+      "loss": 0.5952,
+      "step": 855
+    },
+    {
+      "epoch": 19.02,
+      "learning_rate": 1.2509474482034433e-06,
+      "loss": 0.6436,
+      "step": 856
+    },
+    {
+      "epoch": 19.04,
+      "learning_rate": 1.1948445730325163e-06,
+      "loss": 0.6177,
+      "step": 857
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 1.1400212276321376e-06,
+      "loss": 0.5146,
+      "step": 858
+    },
+    {
+      "epoch": 19.09,
+      "learning_rate": 1.0864781219662611e-06,
+      "loss": 0.6079,
+      "step": 859
+    },
+    {
+      "epoch": 19.11,
+      "learning_rate": 1.034215949419748e-06,
+      "loss": 0.625,
+      "step": 860
+    },
+    {
+      "epoch": 19.13,
+      "learning_rate": 9.832353867893386e-07,
+      "loss": 0.5764,
+      "step": 861
+    },
+    {
+      "epoch": 19.16,
+      "learning_rate": 9.335370942748389e-07,
+      "loss": 0.6313,
+      "step": 862
+    },
+    {
+      "epoch": 19.18,
+      "learning_rate": 8.851217154706826e-07,
+      "loss": 0.5654,
+      "step": 863
+    },
+    {
+      "epoch": 19.2,
+      "learning_rate": 8.379898773574924e-07,
+      "loss": 0.6265,
+      "step": 864
+    },
+    {
+      "epoch": 19.22,
+      "learning_rate": 7.921421902939874e-07,
+      "loss": 0.6167,
+      "step": 865
+    },
+    {
+      "epoch": 19.24,
+      "learning_rate": 7.475792480091226e-07,
+      "loss": 0.5459,
+      "step": 866
+    },
+    {
+      "epoch": 19.27,
+      "learning_rate": 7.043016275943615e-07,
+      "loss": 0.5786,
+      "step": 867
+    },
+    {
+      "epoch": 19.29,
+      "learning_rate": 6.623098894962044e-07,
+      "loss": 0.5356,
+      "step": 868
+    },
+    {
+      "epoch": 19.31,
+      "learning_rate": 6.216045775089275e-07,
+      "loss": 0.5835,
+      "step": 869
+    },
+    {
+      "epoch": 19.33,
+      "learning_rate": 5.821862187675775e-07,
+      "loss": 0.6123,
+      "step": 870
+    },
+    {
+      "epoch": 19.36,
+      "learning_rate": 5.440553237410772e-07,
+      "loss": 0.6187,
+      "step": 871
+    },
+    {
+      "epoch": 19.38,
+      "learning_rate": 5.072123862256972e-07,
+      "loss": 0.5815,
+      "step": 872
+    },
+    {
+      "epoch": 19.4,
+      "learning_rate": 4.7165788333860536e-07,
+      "loss": 0.6494,
+      "step": 873
+    },
+    {
+      "epoch": 19.42,
+      "learning_rate": 4.373922755116722e-07,
+      "loss": 0.5793,
+      "step": 874
+    },
+    {
+      "epoch": 19.44,
+      "learning_rate": 4.044160064855751e-07,
+      "loss": 0.5752,
+      "step": 875
+    },
+    {
+      "epoch": 19.47,
+      "learning_rate": 3.727295033040035e-07,
+      "loss": 0.6265,
+      "step": 876
+    },
+    {
+      "epoch": 19.49,
+      "learning_rate": 3.4233317630814053e-07,
+      "loss": 0.6616,
+      "step": 877
+    },
+    {
+      "epoch": 19.51,
+      "learning_rate": 3.132274191313345e-07,
+      "loss": 0.5957,
+      "step": 878
+    },
+    {
+      "epoch": 19.53,
+      "learning_rate": 2.854126086940356e-07,
+      "loss": 0.624,
+      "step": 879
+    },
+    {
+      "epoch": 19.56,
+      "learning_rate": 2.588891051988895e-07,
+      "loss": 0.5566,
+      "step": 880
+    },
+    {
+      "epoch": 19.58,
+      "learning_rate": 2.3365725212607381e-07,
+      "loss": 0.5947,
+      "step": 881
+    },
+    {
+      "epoch": 19.6,
+      "learning_rate": 2.0971737622883515e-07,
+      "loss": 0.5977,
+      "step": 882
+    },
+    {
+      "epoch": 19.62,
+      "learning_rate": 1.8706978752931482e-07,
+      "loss": 0.6406,
+      "step": 883
+    },
+    {
+      "epoch": 19.64,
+      "learning_rate": 1.6571477931446312e-07,
+      "loss": 0.624,
+      "step": 884
+    },
+    {
+      "epoch": 19.67,
+      "learning_rate": 1.4565262813230894e-07,
+      "loss": 0.5942,
+      "step": 885
+    },
+    {
+      "epoch": 19.69,
+      "learning_rate": 1.2688359378829618e-07,
+      "loss": 0.5977,
+      "step": 886
+    },
+    {
+      "epoch": 19.71,
+      "learning_rate": 1.0940791934198613e-07,
+      "loss": 0.6108,
+      "step": 887
+    },
+    {
+      "epoch": 19.73,
+      "learning_rate": 9.32258311039269e-08,
+      "loss": 0.5933,
+      "step": 888
+    },
+    {
+      "epoch": 19.76,
+      "learning_rate": 7.833753863263349e-08,
+      "loss": 0.6123,
+      "step": 889
+    },
+    {
+      "epoch": 19.78,
+      "learning_rate": 6.474323473194543e-08,
+      "loss": 0.5894,
+      "step": 890
+    },
+    {
+      "epoch": 19.8,
+      "learning_rate": 5.2443095448506674e-08,
+      "loss": 0.5991,
+      "step": 891
+    },
+    {
+      "epoch": 19.82,
+      "learning_rate": 4.143728006951175e-08,
+      "loss": 0.6077,
+      "step": 892
+    },
+    {
+      "epoch": 19.84,
+      "learning_rate": 3.1725931120596405e-08,
+      "loss": 0.6455,
+      "step": 893
+    },
+    {
+      "epoch": 19.87,
+      "learning_rate": 2.3309174364027907e-08,
+      "loss": 0.5962,
+      "step": 894
+    },
+    {
+      "epoch": 19.89,
+      "learning_rate": 1.6187118797061917e-08,
+      "loss": 0.6064,
+      "step": 895
+    },
+    {
+      "epoch": 19.91,
+      "learning_rate": 1.0359856650532519e-08,
+      "loss": 0.5718,
+      "step": 896
+    },
+    {
+      "epoch": 19.93,
+      "learning_rate": 5.827463387653165e-09,
+      "loss": 0.5679,
+      "step": 897
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 2.589997703072999e-09,
+      "loss": 0.5688,
+      "step": 898
+    },
+    {
+      "epoch": 19.98,
+      "learning_rate": 6.475015220552827e-10,
+      "loss": 0.6479,
+      "step": 899
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 0.0,
+      "loss": 0.6077,
+      "step": 900
+    },
+    {
+      "epoch": 20.0,
+      "step": 900,
+      "total_flos": 1.5961750176771604e+18,
+      "train_loss": 1.5027020941840277,
+      "train_runtime": 6381.0295,
+      "train_samples_per_second": 71.835,
+      "train_steps_per_second": 0.141
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 900,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "total_flos": 1.5961750176771604e+18,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}