diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,8344 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1386,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.1904761904761906e-07,
+      "loss": 1.3946,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.3809523809523811e-07,
+      "loss": 1.4061,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.5714285714285716e-07,
+      "loss": 1.5024,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.7619047619047623e-07,
+      "loss": 1.4133,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.952380952380953e-07,
+      "loss": 1.3062,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.142857142857143e-07,
+      "loss": 1.4196,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 1.4895,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.523809523809525e-07,
+      "loss": 1.3539,
+      "step": 8
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0714285714285714e-06,
+      "loss": 1.3545,
+      "step": 9
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1904761904761906e-06,
+      "loss": 1.3961,
+      "step": 10
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3095238095238096e-06,
+      "loss": 1.3732,
+      "step": 11
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.4285714285714286e-06,
+      "loss": 1.3545,
+      "step": 12
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.5476190476190479e-06,
+      "loss": 1.3962,
+      "step": 13
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6666666666666667e-06,
+      "loss": 1.3635,
+      "step": 14
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7857142857142859e-06,
+      "loss": 1.3969,
+      "step": 15
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.904761904761905e-06,
+      "loss": 1.2621,
+      "step": 16
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.023809523809524e-06,
+      "loss": 1.3088,
+      "step": 17
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.1428571428571427e-06,
+      "loss": 1.3454,
+      "step": 18
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.261904761904762e-06,
+      "loss": 1.3698,
+      "step": 19
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.380952380952381e-06,
+      "loss": 1.2087,
+      "step": 20
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.5e-06,
+      "loss": 1.2755,
+      "step": 21
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.6190476190476192e-06,
+      "loss": 1.318,
+      "step": 22
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.7380952380952387e-06,
+      "loss": 1.2695,
+      "step": 23
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.8571428571428573e-06,
+      "loss": 1.2872,
+      "step": 24
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9761904761904763e-06,
+      "loss": 1.2864,
+      "step": 25
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.0952380952380957e-06,
+      "loss": 1.3348,
+      "step": 26
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.2142857142857147e-06,
+      "loss": 1.3063,
+      "step": 27
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 1.2449,
+      "step": 28
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.4523809523809528e-06,
+      "loss": 1.2698,
+      "step": 29
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.5714285714285718e-06,
+      "loss": 1.2673,
+      "step": 30
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.690476190476191e-06,
+      "loss": 1.2796,
+      "step": 31
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.80952380952381e-06,
+      "loss": 1.3012,
+      "step": 32
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.928571428571429e-06,
+      "loss": 1.2365,
+      "step": 33
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.047619047619048e-06,
+      "loss": 1.2532,
+      "step": 34
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.166666666666667e-06,
+      "loss": 1.3028,
+      "step": 35
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.2857142857142855e-06,
+      "loss": 1.2532,
+      "step": 36
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.404761904761905e-06,
+      "loss": 1.2419,
+      "step": 37
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.523809523809524e-06,
+      "loss": 1.1859,
+      "step": 38
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.642857142857144e-06,
+      "loss": 1.2687,
+      "step": 39
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.761904761904762e-06,
+      "loss": 1.2494,
+      "step": 40
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.880952380952381e-06,
+      "loss": 1.2043,
+      "step": 41
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5e-06,
+      "loss": 1.2295,
+      "step": 42
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.99999317015224e-06,
+      "loss": 1.2781,
+      "step": 43
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999972680646276e-06,
+      "loss": 1.2932,
+      "step": 44
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999938531594062e-06,
+      "loss": 1.23,
+      "step": 45
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999890723182183e-06,
+      "loss": 1.226,
+      "step": 46
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999829255671857e-06,
+      "loss": 1.1824,
+      "step": 47
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999754129398938e-06,
+      "loss": 1.2213,
+      "step": 48
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9996653447739054e-06,
+      "loss": 1.2769,
+      "step": 49
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999562902281866e-06,
+      "loss": 1.2213,
+      "step": 50
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999446802482556e-06,
+      "loss": 1.2102,
+      "step": 51
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9993170460103295e-06,
+      "loss": 1.1609,
+      "step": 52
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999173633574158e-06,
+      "loss": 1.1904,
+      "step": 53
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.999016565957633e-06,
+      "loss": 1.1683,
+      "step": 54
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.99884584401895e-06,
+      "loss": 1.2171,
+      "step": 55
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.998661468690914e-06,
+      "loss": 1.2155,
+      "step": 56
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.998463440980931e-06,
+      "loss": 1.2159,
+      "step": 57
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.998251761970997e-06,
+      "loss": 1.2489,
+      "step": 58
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.998026432817702e-06,
+      "loss": 1.2293,
+      "step": 59
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.997787454752217e-06,
+      "loss": 1.2377,
+      "step": 60
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.99753482908029e-06,
+      "loss": 1.1822,
+      "step": 61
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9972685571822355e-06,
+      "loss": 1.243,
+      "step": 62
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.996988640512931e-06,
+      "loss": 1.235,
+      "step": 63
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9966950806018075e-06,
+      "loss": 1.188,
+      "step": 64
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.996387879052841e-06,
+      "loss": 1.2172,
+      "step": 65
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.996067037544542e-06,
+      "loss": 1.1841,
+      "step": 66
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.99573255782995e-06,
+      "loss": 1.2061,
+      "step": 67
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.995384441736622e-06,
+      "loss": 1.1844,
+      "step": 68
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.995022691166621e-06,
+      "loss": 1.1889,
+      "step": 69
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.994647308096509e-06,
+      "loss": 1.1591,
+      "step": 70
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.994258294577333e-06,
+      "loss": 1.1708,
+      "step": 71
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.993855652734616e-06,
+      "loss": 1.1861,
+      "step": 72
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.993439384768342e-06,
+      "loss": 1.1728,
+      "step": 73
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.993009492952951e-06,
+      "loss": 1.1907,
+      "step": 74
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.992565979637318e-06,
+      "loss": 1.1469,
+      "step": 75
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.992108847244746e-06,
+      "loss": 1.221,
+      "step": 76
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.991638098272951e-06,
+      "loss": 1.1186,
+      "step": 77
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9911537352940485e-06,
+      "loss": 1.1231,
+      "step": 78
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.9906557609545375e-06,
+      "loss": 1.1212,
+      "step": 79
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.99014417797529e-06,
+      "loss": 1.1806,
+      "step": 80
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.989618989151533e-06,
+      "loss": 1.133,
+      "step": 81
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.989080197352834e-06,
+      "loss": 1.1415,
+      "step": 82
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.9885278055230875e-06,
+      "loss": 1.1616,
+      "step": 83
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.987961816680493e-06,
+      "loss": 1.193,
+      "step": 84
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.987382233917545e-06,
+      "loss": 1.089,
+      "step": 85
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9867890604010125e-06,
+      "loss": 1.1837,
+      "step": 86
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9861822993719255e-06,
+      "loss": 1.1375,
+      "step": 87
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.985561954145552e-06,
+      "loss": 1.1079,
+      "step": 88
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.98492802811138e-06,
+      "loss": 1.2162,
+      "step": 89
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.984280524733107e-06,
+      "loss": 1.1609,
+      "step": 90
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.983619447548611e-06,
+      "loss": 1.1058,
+      "step": 91
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.982944800169939e-06,
+      "loss": 1.1134,
+      "step": 92
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.98225658628328e-06,
+      "loss": 1.1145,
+      "step": 93
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.981554809648952e-06,
+      "loss": 1.1627,
+      "step": 94
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.980839474101378e-06,
+      "loss": 1.1349,
+      "step": 95
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.980110583549062e-06,
+      "loss": 1.1472,
+      "step": 96
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.979368141974575e-06,
+      "loss": 1.1521,
+      "step": 97
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.978612153434527e-06,
+      "loss": 1.1047,
+      "step": 98
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.977842622059546e-06,
+      "loss": 1.128,
+      "step": 99
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.97705955205426e-06,
+      "loss": 1.1579,
+      "step": 100
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.976262947697265e-06,
+      "loss": 1.0862,
+      "step": 101
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9754528133411144e-06,
+      "loss": 1.1108,
+      "step": 102
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.9746291534122805e-06,
+      "loss": 1.0984,
+      "step": 103
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.973791972411142e-06,
+      "loss": 1.1147,
+      "step": 104
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.972941274911953e-06,
+      "loss": 1.1634,
+      "step": 105
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.9720770655628216e-06,
+      "loss": 1.2256,
+      "step": 106
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.971199349085683e-06,
+      "loss": 1.1609,
+      "step": 107
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.970308130276273e-06,
+      "loss": 1.0844,
+      "step": 108
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9694034140041015e-06,
+      "loss": 1.1932,
+      "step": 109
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.968485205212429e-06,
+      "loss": 1.1266,
+      "step": 110
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.967553508918236e-06,
+      "loss": 1.15,
+      "step": 111
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.966608330212198e-06,
+      "loss": 1.1118,
+      "step": 112
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.965649674258657e-06,
+      "loss": 1.1662,
+      "step": 113
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.96467754629559e-06,
+      "loss": 1.1065,
+      "step": 114
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.963691951634588e-06,
+      "loss": 1.1719,
+      "step": 115
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.96269289566082e-06,
+      "loss": 1.0891,
+      "step": 116
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.961680383833005e-06,
+      "loss": 1.1007,
+      "step": 117
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9606544216833865e-06,
+      "loss": 1.1109,
+      "step": 118
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.959615014817694e-06,
+      "loss": 1.1269,
+      "step": 119
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.958562168915121e-06,
+      "loss": 1.1305,
+      "step": 120
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.957495889728291e-06,
+      "loss": 1.1418,
+      "step": 121
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9564161830832214e-06,
+      "loss": 1.1291,
+      "step": 122
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.955323054879298e-06,
+      "loss": 1.1304,
+      "step": 123
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.954216511089242e-06,
+      "loss": 1.1179,
+      "step": 124
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.953096557759072e-06,
+      "loss": 1.0964,
+      "step": 125
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9519632010080765e-06,
+      "loss": 1.1108,
+      "step": 126
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.95081644702878e-06,
+      "loss": 1.0967,
+      "step": 127
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.949656302086907e-06,
+      "loss": 1.1283,
+      "step": 128
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.948482772521346e-06,
+      "loss": 1.1233,
+      "step": 129
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.947295864744121e-06,
+      "loss": 1.1071,
+      "step": 130
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.946095585240353e-06,
+      "loss": 1.085,
+      "step": 131
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.944881940568219e-06,
+      "loss": 1.1204,
+      "step": 132
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.94365493735893e-06,
+      "loss": 1.1638,
+      "step": 133
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.942414582316679e-06,
+      "loss": 1.1378,
+      "step": 134
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.941160882218617e-06,
+      "loss": 1.1266,
+      "step": 135
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.939893843914808e-06,
+      "loss": 1.0971,
+      "step": 136
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.938613474328195e-06,
+      "loss": 1.0708,
+      "step": 137
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.937319780454559e-06,
+      "loss": 1.0818,
+      "step": 138
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.93601276936249e-06,
+      "loss": 1.0869,
+      "step": 139
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9346924481933345e-06,
+      "loss": 1.1202,
+      "step": 140
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.933358824161167e-06,
+      "loss": 1.0576,
+      "step": 141
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.932011904552749e-06,
+      "loss": 1.0518,
+      "step": 142
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.930651696727482e-06,
+      "loss": 1.0729,
+      "step": 143
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.929278208117378e-06,
+      "loss": 1.0985,
+      "step": 144
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.92789144622701e-06,
+      "loss": 1.0929,
+      "step": 145
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9264914186334775e-06,
+      "loss": 1.1205,
+      "step": 146
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.925078132986361e-06,
+      "loss": 1.0576,
+      "step": 147
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.923651597007679e-06,
+      "loss": 1.0956,
+      "step": 148
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.922211818491852e-06,
+      "loss": 1.096,
+      "step": 149
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.920758805305654e-06,
+      "loss": 1.0561,
+      "step": 150
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.919292565388172e-06,
+      "loss": 1.0633,
+      "step": 151
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.917813106750763e-06,
+      "loss": 1.113,
+      "step": 152
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9163204374770085e-06,
+      "loss": 1.0652,
+      "step": 153
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.914814565722671e-06,
+      "loss": 1.091,
+      "step": 154
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.913295499715651e-06,
+      "loss": 1.0376,
+      "step": 155
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.91176324775594e-06,
+      "loss": 1.0568,
+      "step": 156
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.910217818215576e-06,
+      "loss": 1.1732,
+      "step": 157
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.908659219538598e-06,
+      "loss": 1.081,
+      "step": 158
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.907087460240999e-06,
+      "loss": 1.0785,
+      "step": 159
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.905502548910681e-06,
+      "loss": 1.087,
+      "step": 160
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9039044942074055e-06,
+      "loss": 1.1409,
+      "step": 161
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.9022933048627496e-06,
+      "loss": 1.1089,
+      "step": 162
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.900668989680055e-06,
+      "loss": 1.0687,
+      "step": 163
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.899031557534383e-06,
+      "loss": 1.0767,
+      "step": 164
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.897381017372462e-06,
+      "loss": 1.1029,
+      "step": 165
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.895717378212644e-06,
+      "loss": 1.0583,
+      "step": 166
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.894040649144849e-06,
+      "loss": 1.0986,
+      "step": 167
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.8923508393305224e-06,
+      "loss": 1.0978,
+      "step": 168
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.8906479580025774e-06,
+      "loss": 1.0686,
+      "step": 169
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.8889320144653525e-06,
+      "loss": 1.0808,
+      "step": 170
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.887203018094552e-06,
+      "loss": 1.1331,
+      "step": 171
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.885460978337201e-06,
+      "loss": 1.1233,
+      "step": 172
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.8837059047115955e-06,
+      "loss": 1.0679,
+      "step": 173
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.881937806807241e-06,
+      "loss": 1.1077,
+      "step": 174
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.880156694284811e-06,
+      "loss": 1.1039,
+      "step": 175
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.8783625768760865e-06,
+      "loss": 1.097,
+      "step": 176
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.876555464383908e-06,
+      "loss": 1.0581,
+      "step": 177
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.8747353666821155e-06,
+      "loss": 1.0559,
+      "step": 178
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.872902293715502e-06,
+      "loss": 1.07,
+      "step": 179
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.871056255499758e-06,
+      "loss": 1.1274,
+      "step": 180
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.869197262121406e-06,
+      "loss": 1.0831,
+      "step": 181
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.867325323737765e-06,
+      "loss": 1.0732,
+      "step": 182
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8654404505768735e-06,
+      "loss": 1.0123,
+      "step": 183
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.863542652937453e-06,
+      "loss": 1.0958,
+      "step": 184
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.861631941188836e-06,
+      "loss": 1.0742,
+      "step": 185
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.859708325770919e-06,
+      "loss": 1.0819,
+      "step": 186
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8577718171941036e-06,
+      "loss": 1.0386,
+      "step": 187
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.855822426039236e-06,
+      "loss": 1.0735,
+      "step": 188
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8538601629575525e-06,
+      "loss": 1.0522,
+      "step": 189
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.851885038670618e-06,
+      "loss": 1.1023,
+      "step": 190
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8498970639702745e-06,
+      "loss": 1.0681,
+      "step": 191
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.84789624971857e-06,
+      "loss": 1.0624,
+      "step": 192
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.845882606847712e-06,
+      "loss": 1.0571,
+      "step": 193
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.843856146359999e-06,
+      "loss": 1.0899,
+      "step": 194
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.841816879327764e-06,
+      "loss": 1.1203,
+      "step": 195
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.839764816893315e-06,
+      "loss": 1.0775,
+      "step": 196
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.83769997026887e-06,
+      "loss": 1.112,
+      "step": 197
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.835622350736499e-06,
+      "loss": 1.0862,
+      "step": 198
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.8335319696480655e-06,
+      "loss": 1.0268,
+      "step": 199
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.831428838425153e-06,
+      "loss": 1.125,
+      "step": 200
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8293129685590164e-06,
+      "loss": 1.1081,
+      "step": 201
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.827184371610511e-06,
+      "loss": 1.1479,
+      "step": 202
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.82504305921003e-06,
+      "loss": 1.0941,
+      "step": 203
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.822889043057446e-06,
+      "loss": 1.0968,
+      "step": 204
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.820722334922039e-06,
+      "loss": 1.0657,
+      "step": 205
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8185429466424375e-06,
+      "loss": 1.0654,
+      "step": 206
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.816350890126556e-06,
+      "loss": 1.0931,
+      "step": 207
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.814146177351523e-06,
+      "loss": 1.117,
+      "step": 208
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.811928820363621e-06,
+      "loss": 1.083,
+      "step": 209
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.809698831278217e-06,
+      "loss": 1.0667,
+      "step": 210
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.807456222279703e-06,
+      "loss": 1.0113,
+      "step": 211
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8052010056214184e-06,
+      "loss": 1.0232,
+      "step": 212
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.802933193625593e-06,
+      "loss": 1.012,
+      "step": 213
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.800652798683277e-06,
+      "loss": 1.0283,
+      "step": 214
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.798359833254269e-06,
+      "loss": 1.0364,
+      "step": 215
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.796054309867053e-06,
+      "loss": 1.0499,
+      "step": 216
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.793736241118729e-06,
+      "loss": 1.0475,
+      "step": 217
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.791405639674941e-06,
+      "loss": 1.0823,
+      "step": 218
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.789062518269812e-06,
+      "loss": 1.0629,
+      "step": 219
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.7867068897058725e-06,
+      "loss": 1.0438,
+      "step": 220
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.784338766853989e-06,
+      "loss": 1.0429,
+      "step": 221
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.781958162653298e-06,
+      "loss": 1.0831,
+      "step": 222
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.779565090111129e-06,
+      "loss": 1.0904,
+      "step": 223
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.77715956230294e-06,
+      "loss": 1.0542,
+      "step": 224
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.774741592372242e-06,
+      "loss": 1.0667,
+      "step": 225
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.7723111935305275e-06,
+      "loss": 1.0663,
+      "step": 226
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.769868379057201e-06,
+      "loss": 1.0283,
+      "step": 227
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.767413162299501e-06,
+      "loss": 1.0474,
+      "step": 228
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.764945556672435e-06,
+      "loss": 1.1026,
+      "step": 229
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.7624655756587e-06,
+      "loss": 1.0322,
+      "step": 230
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.759973232808609e-06,
+      "loss": 1.0415,
+      "step": 231
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.757468541740019e-06,
+      "loss": 1.0144,
+      "step": 232
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.754951516138259e-06,
+      "loss": 1.0599,
+      "step": 233
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.752422169756048e-06,
+      "loss": 1.0688,
+      "step": 234
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.749880516413428e-06,
+      "loss": 1.1168,
+      "step": 235
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.747326569997684e-06,
+      "loss": 1.0417,
+      "step": 236
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.744760344463267e-06,
+      "loss": 1.07,
+      "step": 237
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.742181853831721e-06,
+      "loss": 1.0162,
+      "step": 238
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.739591112191605e-06,
+      "loss": 1.1038,
+      "step": 239
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.736988133698416e-06,
+      "loss": 1.0519,
+      "step": 240
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.73437293257451e-06,
+      "loss": 1.0285,
+      "step": 241
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.731745523109029e-06,
+      "loss": 1.0441,
+      "step": 242
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.729105919657818e-06,
+      "loss": 1.0275,
+      "step": 243
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.7264541366433495e-06,
+      "loss": 1.0335,
+      "step": 244
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.7237901885546405e-06,
+      "loss": 1.0802,
+      "step": 245
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.721114089947181e-06,
+      "loss": 1.0849,
+      "step": 246
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.718425855442847e-06,
+      "loss": 1.0401,
+      "step": 247
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.715725499729826e-06,
+      "loss": 1.0161,
+      "step": 248
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.713013037562531e-06,
+      "loss": 1.0049,
+      "step": 249
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.710288483761524e-06,
+      "loss": 1.0174,
+      "step": 250
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.707551853213439e-06,
+      "loss": 1.0612,
+      "step": 251
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.704803160870888e-06,
+      "loss": 1.0924,
+      "step": 252
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.702042421752393e-06,
+      "loss": 1.0671,
+      "step": 253
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.699269650942296e-06,
+      "loss": 1.0589,
+      "step": 254
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.69648486359068e-06,
+      "loss": 1.0373,
+      "step": 255
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.693688074913282e-06,
+      "loss": 1.0508,
+      "step": 256
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.690879300191416e-06,
+      "loss": 1.0986,
+      "step": 257
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.6880585547718845e-06,
+      "loss": 1.0661,
+      "step": 258
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.685225854066897e-06,
+      "loss": 1.0082,
+      "step": 259
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.682381213553986e-06,
+      "loss": 1.058,
+      "step": 260
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.67952464877592e-06,
+      "loss": 1.0485,
+      "step": 261
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.676656175340621e-06,
+      "loss": 1.0192,
+      "step": 262
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.673775808921078e-06,
+      "loss": 1.0188,
+      "step": 263
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.670883565255264e-06,
+      "loss": 1.0785,
+      "step": 264
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.667979460146045e-06,
+      "loss": 1.0604,
+      "step": 265
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.665063509461098e-06,
+      "loss": 1.0797,
+      "step": 266
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.662135729132821e-06,
+      "loss": 1.095,
+      "step": 267
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.659196135158251e-06,
+      "loss": 1.0508,
+      "step": 268
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.65624474359897e-06,
+      "loss": 1.004,
+      "step": 269
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.653281570581023e-06,
+      "loss": 1.0815,
+      "step": 270
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.6503066322948264e-06,
+      "loss": 1.0403,
+      "step": 271
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.64731994499508e-06,
+      "loss": 1.0166,
+      "step": 272
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.644321525000681e-06,
+      "loss": 1.1433,
+      "step": 273
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.641311388694629e-06,
+      "loss": 0.9864,
+      "step": 274
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.638289552523944e-06,
+      "loss": 1.0397,
+      "step": 275
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.635256032999569e-06,
+      "loss": 1.06,
+      "step": 276
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.6322108466962865e-06,
+      "loss": 1.0367,
+      "step": 277
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.629154010252624e-06,
+      "loss": 1.0056,
+      "step": 278
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.6260855403707625e-06,
+      "loss": 0.9576,
+      "step": 279
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.623005453816447e-06,
+      "loss": 1.0415,
+      "step": 280
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.619913767418898e-06,
+      "loss": 1.0351,
+      "step": 281
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.6168104980707105e-06,
+      "loss": 1.0204,
+      "step": 282
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.6136956627277725e-06,
+      "loss": 1.0511,
+      "step": 283
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.610569278409164e-06,
+      "loss": 1.0148,
+      "step": 284
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.607431362197067e-06,
+      "loss": 1.0709,
+      "step": 285
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.604281931236675e-06,
+      "loss": 1.0407,
+      "step": 286
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.601121002736095e-06,
+      "loss": 1.0465,
+      "step": 287
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.597948593966256e-06,
+      "loss": 0.9542,
+      "step": 288
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.594764722260812e-06,
+      "loss": 0.9913,
+      "step": 289
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.59156940501605e-06,
+      "loss": 1.0946,
+      "step": 290
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.5883626596907945e-06,
+      "loss": 1.0525,
+      "step": 291
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.585144503806312e-06,
+      "loss": 1.0513,
+      "step": 292
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.581914954946215e-06,
+      "loss": 1.112,
+      "step": 293
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.578674030756364e-06,
+      "loss": 1.0894,
+      "step": 294
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.575421748944773e-06,
+      "loss": 1.0049,
+      "step": 295
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.572158127281516e-06,
+      "loss": 0.9974,
+      "step": 296
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.568883183598623e-06,
+      "loss": 1.0169,
+      "step": 297
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.565596935789987e-06,
+      "loss": 1.0418,
+      "step": 298
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.562299401811268e-06,
+      "loss": 1.0091,
+      "step": 299
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.558990599679787e-06,
+      "loss": 1.0447,
+      "step": 300
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.555670547474438e-06,
+      "loss": 1.0131,
+      "step": 301
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.552339263335581e-06,
+      "loss": 1.007,
+      "step": 302
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.548996765464947e-06,
+      "loss": 1.0085,
+      "step": 303
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.5456430721255384e-06,
+      "loss": 1.0338,
+      "step": 304
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.5422782016415255e-06,
+      "loss": 1.0302,
+      "step": 305
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.538902172398151e-06,
+      "loss": 1.0259,
+      "step": 306
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.535515002841628e-06,
+      "loss": 1.0189,
+      "step": 307
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.532116711479039e-06,
+      "loss": 1.0391,
+      "step": 308
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.528707316878233e-06,
+      "loss": 1.0154,
+      "step": 309
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.525286837667726e-06,
+      "loss": 0.995,
+      "step": 310
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.521855292536603e-06,
+      "loss": 1.0459,
+      "step": 311
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.518412700234407e-06,
+      "loss": 1.0192,
+      "step": 312
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.514959079571042e-06,
+      "loss": 1.0181,
+      "step": 313
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.511494449416671e-06,
+      "loss": 0.9982,
+      "step": 314
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.508018828701613e-06,
+      "loss": 1.0017,
+      "step": 315
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.504532236416234e-06,
+      "loss": 1.0567,
+      "step": 316
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.501034691610852e-06,
+      "loss": 1.0419,
+      "step": 317
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.4975262133956235e-06,
+      "loss": 1.0696,
+      "step": 318
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.4940068209404465e-06,
+      "loss": 1.0446,
+      "step": 319
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.490476533474854e-06,
+      "loss": 1.0119,
+      "step": 320
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.486935370287907e-06,
+      "loss": 1.0258,
+      "step": 321
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.4833833507280884e-06,
+      "loss": 0.9649,
+      "step": 322
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.4798204942032005e-06,
+      "loss": 1.001,
+      "step": 323
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.476246820180259e-06,
+      "loss": 1.0449,
+      "step": 324
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.472662348185382e-06,
+      "loss": 0.9918,
+      "step": 325
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.469067097803689e-06,
+      "loss": 0.9844,
+      "step": 326
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.465461088679189e-06,
+      "loss": 1.0193,
+      "step": 327
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.461844340514678e-06,
+      "loss": 0.9866,
+      "step": 328
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.4582168730716264e-06,
+      "loss": 1.0365,
+      "step": 329
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.454578706170075e-06,
+      "loss": 1.0687,
+      "step": 330
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.450929859688524e-06,
+      "loss": 0.9721,
+      "step": 331
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.447270353563828e-06,
+      "loss": 1.0291,
+      "step": 332
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.44360020779108e-06,
+      "loss": 1.048,
+      "step": 333
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.439919442423513e-06,
+      "loss": 0.9766,
+      "step": 334
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.4362280775723775e-06,
+      "loss": 1.0227,
+      "step": 335
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.432526133406843e-06,
+      "loss": 0.9643,
+      "step": 336
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.428813630153881e-06,
+      "loss": 1.0257,
+      "step": 337
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.425090588098158e-06,
+      "loss": 0.9937,
+      "step": 338
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.421357027581921e-06,
+      "loss": 1.0464,
+      "step": 339
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.4176129690048905e-06,
+      "loss": 1.0072,
+      "step": 340
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.413858432824147e-06,
+      "loss": 1.0912,
+      "step": 341
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.410093439554019e-06,
+      "loss": 1.0214,
+      "step": 342
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.406318009765971e-06,
+      "loss": 1.0206,
+      "step": 343
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.40253216408849e-06,
+      "loss": 1.0272,
+      "step": 344
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.398735923206978e-06,
+      "loss": 1.0626,
+      "step": 345
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.394929307863633e-06,
+      "loss": 1.0309,
+      "step": 346
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.391112338857335e-06,
+      "loss": 1.0412,
+      "step": 347
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.38728503704354e-06,
+      "loss": 0.9742,
+      "step": 348
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.383447423334159e-06,
+      "loss": 0.9859,
+      "step": 349
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.379599518697444e-06,
+      "loss": 0.9719,
+      "step": 350
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.375741344157879e-06,
+      "loss": 0.9563,
+      "step": 351
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.371872920796059e-06,
+      "loss": 1.0104,
+      "step": 352
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.367994269748579e-06,
+      "loss": 1.0322,
+      "step": 353
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.364105412207914e-06,
+      "loss": 1.013,
+      "step": 354
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.360206369422311e-06,
+      "loss": 1.0343,
+      "step": 355
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.356297162695662e-06,
+      "loss": 0.9826,
+      "step": 356
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.352377813387398e-06,
+      "loss": 1.0222,
+      "step": 357
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.348448342912365e-06,
+      "loss": 1.0188,
+      "step": 358
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.344508772740714e-06,
+      "loss": 1.0435,
+      "step": 359
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.340559124397774e-06,
+      "loss": 1.0977,
+      "step": 360
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.336599419463943e-06,
+      "loss": 0.9367,
+      "step": 361
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.332629679574566e-06,
+      "loss": 1.0609,
+      "step": 362
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.32864992641982e-06,
+      "loss": 0.999,
+      "step": 363
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.324660181744589e-06,
+      "loss": 1.0526,
+      "step": 364
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.320660467348355e-06,
+      "loss": 1.0102,
+      "step": 365
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.316650805085068e-06,
+      "loss": 1.0429,
+      "step": 366
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.312631216863035e-06,
+      "loss": 1.0345,
+      "step": 367
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.308601724644797e-06,
+      "loss": 1.0044,
+      "step": 368
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.304562350447008e-06,
+      "loss": 0.9693,
+      "step": 369
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.300513116340317e-06,
+      "loss": 1.0062,
+      "step": 370
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.296454044449245e-06,
+      "loss": 1.0395,
+      "step": 371
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.292385156952069e-06,
+      "loss": 1.0303,
+      "step": 372
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.288306476080691e-06,
+      "loss": 0.9308,
+      "step": 373
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.284218024120531e-06,
+      "loss": 1.0221,
+      "step": 374
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2801198234103895e-06,
+      "loss": 1.0183,
+      "step": 375
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.276011896342336e-06,
+      "loss": 1.014,
+      "step": 376
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.271894265361584e-06,
+      "loss": 1.0485,
+      "step": 377
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.267766952966369e-06,
+      "loss": 1.0309,
+      "step": 378
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.263629981707823e-06,
+      "loss": 1.0479,
+      "step": 379
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.259483374189851e-06,
+      "loss": 1.0164,
+      "step": 380
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.255327153069014e-06,
+      "loss": 0.9849,
+      "step": 381
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.251161341054396e-06,
+      "loss": 0.9599,
+      "step": 382
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.246985960907488e-06,
+      "loss": 1.0752,
+      "step": 383
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.242801035442059e-06,
+      "loss": 1.0346,
+      "step": 384
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.2386065875240295e-06,
+      "loss": 0.9934,
+      "step": 385
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.234402640071355e-06,
+      "loss": 1.0038,
+      "step": 386
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.2301892160538895e-06,
+      "loss": 1.0294,
+      "step": 387
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.225966338493272e-06,
+      "loss": 1.0041,
+      "step": 388
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.221734030462787e-06,
+      "loss": 1.0398,
+      "step": 389
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.217492315087255e-06,
+      "loss": 0.9899,
+      "step": 390
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.213241215542888e-06,
+      "loss": 1.024,
+      "step": 391
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.2089807550571786e-06,
+      "loss": 1.033,
+      "step": 392
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.204710956908763e-06,
+      "loss": 1.0132,
+      "step": 393
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.200431844427299e-06,
+      "loss": 1.0346,
+      "step": 394
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.196143440993335e-06,
+      "loss": 1.0303,
+      "step": 395
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.191845770038186e-06,
+      "loss": 1.0072,
+      "step": 396
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.187538855043802e-06,
+      "loss": 1.0371,
+      "step": 397
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.183222719542643e-06,
+      "loss": 1.0141,
+      "step": 398
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.178897387117547e-06,
+      "loss": 0.9848,
+      "step": 399
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.174562881401602e-06,
+      "loss": 1.0418,
+      "step": 400
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.170219226078023e-06,
+      "loss": 1.0123,
+      "step": 401
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1658664448800105e-06,
+      "loss": 1.028,
+      "step": 402
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.161504561590632e-06,
+      "loss": 1.0038,
+      "step": 403
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1571336000426865e-06,
+      "loss": 0.9811,
+      "step": 404
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.152753584118576e-06,
+      "loss": 0.9956,
+      "step": 405
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1483645377501726e-06,
+      "loss": 0.998,
+      "step": 406
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.143966484918692e-06,
+      "loss": 0.9867,
+      "step": 407
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.139559449654561e-06,
+      "loss": 1.0174,
+      "step": 408
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.13514345603728e-06,
+      "loss": 1.0261,
+      "step": 409
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.130718528195303e-06,
+      "loss": 1.0238,
+      "step": 410
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1262846903058966e-06,
+      "loss": 1.0273,
+      "step": 411
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.121841966595009e-06,
+      "loss": 0.9801,
+      "step": 412
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.117390381337144e-06,
+      "loss": 0.969,
+      "step": 413
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.11292995885522e-06,
+      "loss": 1.0162,
+      "step": 414
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.108460723520441e-06,
+      "loss": 1.0157,
+      "step": 415
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.103982699752167e-06,
+      "loss": 1.0397,
+      "step": 416
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.099495912017773e-06,
+      "loss": 1.0227,
+      "step": 417
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.095000384832522e-06,
+      "loss": 1.0193,
+      "step": 418
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.090496142759425e-06,
+      "loss": 1.0011,
+      "step": 419
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.085983210409114e-06,
+      "loss": 1.0794,
+      "step": 420
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0814616124397015e-06,
+      "loss": 1.0219,
+      "step": 421
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.076931373556646e-06,
+      "loss": 0.9527,
+      "step": 422
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.072392518512623e-06,
+      "loss": 1.0097,
+      "step": 423
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.067845072107384e-06,
+      "loss": 0.9824,
+      "step": 424
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.063289059187621e-06,
+      "loss": 0.973,
+      "step": 425
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.058724504646834e-06,
+      "loss": 0.9978,
+      "step": 426
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.054151433425194e-06,
+      "loss": 0.9907,
+      "step": 427
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.049569870509404e-06,
+      "loss": 0.9853,
+      "step": 428
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.044979840932567e-06,
+      "loss": 0.9889,
+      "step": 429
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.040381369774045e-06,
+      "loss": 1.0352,
+      "step": 430
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.035774482159323e-06,
+      "loss": 1.0145,
+      "step": 431
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.031159203259876e-06,
+      "loss": 0.9651,
+      "step": 432
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.026535558293024e-06,
+      "loss": 0.9818,
+      "step": 433
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.021903572521802e-06,
+      "loss": 0.9801,
+      "step": 434
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0172632712548145e-06,
+      "loss": 0.9771,
+      "step": 435
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.012614679846103e-06,
+      "loss": 1.0039,
+      "step": 436
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.007957823695005e-06,
+      "loss": 0.9744,
+      "step": 437
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.003292728246015e-06,
+      "loss": 1.0089,
+      "step": 438
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.998619418988646e-06,
+      "loss": 0.9776,
+      "step": 439
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.993937921457292e-06,
+      "loss": 0.9809,
+      "step": 440
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.989248261231084e-06,
+      "loss": 1.0186,
+      "step": 441
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.984550463933754e-06,
+      "loss": 1.013,
+      "step": 442
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.979844555233496e-06,
+      "loss": 1.0743,
+      "step": 443
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.975130560842821e-06,
+      "loss": 1.0163,
+      "step": 444
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.970408506518419e-06,
+      "loss": 0.9437,
+      "step": 445
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.965678418061023e-06,
+      "loss": 0.9803,
+      "step": 446
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.960940321315257e-06,
+      "loss": 1.0044,
+      "step": 447
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.956194242169506e-06,
+      "loss": 0.9526,
+      "step": 448
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.9514402065557675e-06,
+      "loss": 0.9931,
+      "step": 449
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.946678240449515e-06,
+      "loss": 1.0173,
+      "step": 450
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.941908369869549e-06,
+      "loss": 1.0598,
+      "step": 451
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.937130620877863e-06,
+      "loss": 0.9913,
+      "step": 452
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.9323450195794954e-06,
+      "loss": 0.963,
+      "step": 453
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.927551592122389e-06,
+      "loss": 0.9974,
+      "step": 454
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.922750364697246e-06,
+      "loss": 0.9936,
+      "step": 455
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.9179413635373895e-06,
+      "loss": 1.005,
+      "step": 456
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.913124614918617e-06,
+      "loss": 1.0052,
+      "step": 457
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.908300145159055e-06,
+      "loss": 0.9594,
+      "step": 458
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.9034679806190204e-06,
+      "loss": 1.0006,
+      "step": 459
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.898628147700869e-06,
+      "loss": 0.9919,
+      "step": 460
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.89378067284886e-06,
+      "loss": 0.9984,
+      "step": 461
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.888925582549006e-06,
+      "loss": 0.9885,
+      "step": 462
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.884062903328929e-06,
+      "loss": 0.961,
+      "step": 463
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.879192661757715e-06,
+      "loss": 0.9713,
+      "step": 464
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.874314884445771e-06,
+      "loss": 0.9179,
+      "step": 465
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.869429598044679e-06,
+      "loss": 1.0214,
+      "step": 466
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.864536829247049e-06,
+      "loss": 1.0172,
+      "step": 467
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.859636604786372e-06,
+      "loss": 0.944,
+      "step": 468
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.8547289514368795e-06,
+      "loss": 0.9878,
+      "step": 469
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.849813896013392e-06,
+      "loss": 1.0083,
+      "step": 470
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.8448914653711725e-06,
+      "loss": 0.9519,
+      "step": 471
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.839961686405782e-06,
+      "loss": 0.9831,
+      "step": 472
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.835024586052933e-06,
+      "loss": 0.9964,
+      "step": 473
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.830080191288342e-06,
+      "loss": 0.9891,
+      "step": 474
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.825128529127577e-06,
+      "loss": 1.0317,
+      "step": 475
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.82016962662592e-06,
+      "loss": 0.9975,
+      "step": 476
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.815203510878209e-06,
+      "loss": 0.9853,
+      "step": 477
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.810230209018694e-06,
+      "loss": 0.9227,
+      "step": 478
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.805249748220893e-06,
+      "loss": 0.9984,
+      "step": 479
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.8002621556974367e-06,
+      "loss": 0.9422,
+      "step": 480
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.7952674586999226e-06,
+      "loss": 1.0027,
+      "step": 481
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.790265684518767e-06,
+      "loss": 0.9119,
+      "step": 482
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.785256860483054e-06,
+      "loss": 0.9989,
+      "step": 483
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.7802410139603908e-06,
+      "loss": 1.0217,
+      "step": 484
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.7752181723567484e-06,
+      "loss": 0.9916,
+      "step": 485
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.770188363116324e-06,
+      "loss": 0.9977,
+      "step": 486
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.7651516137213816e-06,
+      "loss": 0.9398,
+      "step": 487
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7601079516921076e-06,
+      "loss": 0.9779,
+      "step": 488
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.755057404586455e-06,
+      "loss": 0.9283,
+      "step": 489
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.9638,
+      "step": 490
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7449357655657854e-06,
+      "loss": 0.935,
+      "step": 491
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7398647289541703e-06,
+      "loss": 0.9566,
+      "step": 492
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.7347869178726804e-06,
+      "loss": 1.003,
+      "step": 493
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.7297023600658586e-06,
+      "loss": 0.9928,
+      "step": 494
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.724611083315109e-06,
+      "loss": 0.9376,
+      "step": 495
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.719513115438548e-06,
+      "loss": 1.0196,
+      "step": 496
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.7144084842908506e-06,
+      "loss": 0.8741,
+      "step": 497
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.7092972177630998e-06,
+      "loss": 0.991,
+      "step": 498
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.7041793437826336e-06,
+      "loss": 0.9298,
+      "step": 499
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.699054890312892e-06,
+      "loss": 0.9917,
+      "step": 500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.6939238853532657e-06,
+      "loss": 0.9415,
+      "step": 501
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6887863569389388e-06,
+      "loss": 0.989,
+      "step": 502
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6836423331407416e-06,
+      "loss": 0.9834,
+      "step": 503
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6784918420649952e-06,
+      "loss": 0.9363,
+      "step": 504
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6733349118533524e-06,
+      "loss": 0.9171,
+      "step": 505
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6681715706826555e-06,
+      "loss": 0.9898,
+      "step": 506
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.663001846764769e-06,
+      "loss": 1.0001,
+      "step": 507
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6578257683464363e-06,
+      "loss": 0.9791,
+      "step": 508
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6526433637091193e-06,
+      "loss": 0.9223,
+      "step": 509
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6474546611688446e-06,
+      "loss": 0.9243,
+      "step": 510
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.642259689076052e-06,
+      "loss": 0.9733,
+      "step": 511
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.6370584758154366e-06,
+      "loss": 1.0013,
+      "step": 512
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.631851049805793e-06,
+      "loss": 0.9615,
+      "step": 513
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.626637439499864e-06,
+      "loss": 0.9997,
+      "step": 514
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.6214176733841792e-06,
+      "loss": 0.9664,
+      "step": 515
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.6161917799789076e-06,
+      "loss": 0.9587,
+      "step": 516
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.610959787837693e-06,
+      "loss": 0.9299,
+      "step": 517
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.6057217255475034e-06,
+      "loss": 0.9615,
+      "step": 518
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.600477621728473e-06,
+      "loss": 0.9934,
+      "step": 519
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.595227505033747e-06,
+      "loss": 0.9956,
+      "step": 520
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.589971404149323e-06,
+      "loss": 1.0085,
+      "step": 521
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5847093477938955e-06,
+      "loss": 0.9524,
+      "step": 522
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5794413647187008e-06,
+      "loss": 0.9432,
+      "step": 523
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5741674837073563e-06,
+      "loss": 0.9506,
+      "step": 524
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5688877335757055e-06,
+      "loss": 0.9459,
+      "step": 525
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5636021431716604e-06,
+      "loss": 0.9213,
+      "step": 526
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5583107413750427e-06,
+      "loss": 1.0094,
+      "step": 527
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.553013557097428e-06,
+      "loss": 0.9227,
+      "step": 528
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.547710619281985e-06,
+      "loss": 0.9418,
+      "step": 529
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.542401956903321e-06,
+      "loss": 0.9495,
+      "step": 530
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.5370875989673198e-06,
+      "loss": 0.9391,
+      "step": 531
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.531767574510987e-06,
+      "loss": 0.9696,
+      "step": 532
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.526441912602288e-06,
+      "loss": 1.0026,
+      "step": 533
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.521110642339991e-06,
+      "loss": 0.9495,
+      "step": 534
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.5157737928535065e-06,
+      "loss": 0.934,
+      "step": 535
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.51043139330273e-06,
+      "loss": 0.9408,
+      "step": 536
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.505083472877884e-06,
+      "loss": 0.9468,
+      "step": 537
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.499730060799352e-06,
+      "loss": 0.9958,
+      "step": 538
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.494371186317528e-06,
+      "loss": 1.018,
+      "step": 539
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4890068787126475e-06,
+      "loss": 0.9772,
+      "step": 540
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.483637167294634e-06,
+      "loss": 0.9492,
+      "step": 541
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4782620814029376e-06,
+      "loss": 0.8934,
+      "step": 542
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.472881650406373e-06,
+      "loss": 0.9354,
+      "step": 543
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4674959037029593e-06,
+      "loss": 0.9369,
+      "step": 544
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4621048707197607e-06,
+      "loss": 0.9564,
+      "step": 545
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4567085809127247e-06,
+      "loss": 0.9633,
+      "step": 546
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.451307063766522e-06,
+      "loss": 0.9656,
+      "step": 547
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4459003487943842e-06,
+      "loss": 0.9444,
+      "step": 548
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4404884655379433e-06,
+      "loss": 0.9879,
+      "step": 549
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4350714435670706e-06,
+      "loss": 0.9441,
+      "step": 550
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.429649312479714e-06,
+      "loss": 1.011,
+      "step": 551
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.424222101901738e-06,
+      "loss": 1.0019,
+      "step": 552
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4187898414867594e-06,
+      "loss": 0.963,
+      "step": 553
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4133525609159883e-06,
+      "loss": 0.9777,
+      "step": 554
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4079102898980633e-06,
+      "loss": 0.9542,
+      "step": 555
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4024630581688895e-06,
+      "loss": 0.9276,
+      "step": 556
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3970108954914782e-06,
+      "loss": 0.9875,
+      "step": 557
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.391553831655783e-06,
+      "loss": 0.9596,
+      "step": 558
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3860918964785354e-06,
+      "loss": 0.9375,
+      "step": 559
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3806251198030843e-06,
+      "loss": 1.0134,
+      "step": 560
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.375153531499231e-06,
+      "loss": 0.9159,
+      "step": 561
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.369677161463068e-06,
+      "loss": 0.9951,
+      "step": 562
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.364196039616815e-06,
+      "loss": 0.9551,
+      "step": 563
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.358710195908653e-06,
+      "loss": 0.9753,
+      "step": 564
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3532196603125646e-06,
+      "loss": 0.9942,
+      "step": 565
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3477244628281667e-06,
+      "loss": 0.9705,
+      "step": 566
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3422246334805504e-06,
+      "loss": 0.9857,
+      "step": 567
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3367202023201128e-06,
+      "loss": 0.9709,
+      "step": 568
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.331211199422395e-06,
+      "loss": 0.9784,
+      "step": 569
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3256976548879183e-06,
+      "loss": 0.9502,
+      "step": 570
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.3201795988420184e-06,
+      "loss": 0.9633,
+      "step": 571
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.3146570614346814e-06,
+      "loss": 0.9547,
+      "step": 572
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.309130072840379e-06,
+      "loss": 0.9232,
+      "step": 573
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.303598663257904e-06,
+      "loss": 0.9421,
+      "step": 574
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.298062862910205e-06,
+      "loss": 0.9837,
+      "step": 575
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.292522702044221e-06,
+      "loss": 0.9768,
+      "step": 576
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.2869782109307147e-06,
+      "loss": 0.9995,
+      "step": 577
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.281429419864112e-06,
+      "loss": 0.9245,
+      "step": 578
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.27587635916233e-06,
+      "loss": 0.9799,
+      "step": 579
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.2703190591666174e-06,
+      "loss": 0.9274,
+      "step": 580
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.264757550241384e-06,
+      "loss": 0.9373,
+      "step": 581
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.259191862774037e-06,
+      "loss": 0.9824,
+      "step": 582
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.2536220271748154e-06,
+      "loss": 0.9795,
+      "step": 583
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.2480480738766222e-06,
+      "loss": 1.0047,
+      "step": 584
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.242470033334859e-06,
+      "loss": 0.9361,
+      "step": 585
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.236887936027261e-06,
+      "loss": 0.9239,
+      "step": 586
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2313018124537273e-06,
+      "loss": 0.9939,
+      "step": 587
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.225711693136156e-06,
+      "loss": 0.9822,
+      "step": 588
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2201176086182796e-06,
+      "loss": 0.9341,
+      "step": 589
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.2145195894654947e-06,
+      "loss": 0.9373,
+      "step": 590
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.2089176662646946e-06,
+      "loss": 0.9862,
+      "step": 591
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.203311869624107e-06,
+      "loss": 0.9104,
+      "step": 592
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.197702230173121e-06,
+      "loss": 0.9938,
+      "step": 593
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.1920887785621233e-06,
+      "loss": 0.9591,
+      "step": 594
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.1864715454623312e-06,
+      "loss": 0.9533,
+      "step": 595
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.180850561565621e-06,
+      "loss": 0.9864,
+      "step": 596
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.175225857584364e-06,
+      "loss": 0.9533,
+      "step": 597
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.169597464251258e-06,
+      "loss": 0.9795,
+      "step": 598
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.1639654123191586e-06,
+      "loss": 0.9674,
+      "step": 599
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.1583297325609117e-06,
+      "loss": 1.0189,
+      "step": 600
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.152690455769186e-06,
+      "loss": 0.9471,
+      "step": 601
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.147047612756302e-06,
+      "loss": 1.0013,
+      "step": 602
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1414012343540667e-06,
+      "loss": 0.9712,
+      "step": 603
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1357513514136044e-06,
+      "loss": 0.8872,
+      "step": 604
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1300979948051874e-06,
+      "loss": 0.9814,
+      "step": 605
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1244411954180677e-06,
+      "loss": 0.9741,
+      "step": 606
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.118780984160307e-06,
+      "loss": 0.917,
+      "step": 607
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.113117391958612e-06,
+      "loss": 0.9762,
+      "step": 608
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.10745044975816e-06,
+      "loss": 0.9604,
+      "step": 609
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.1017801885224332e-06,
+      "loss": 0.9892,
+      "step": 610
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.0961066392330475e-06,
+      "loss": 0.9624,
+      "step": 611
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.0904298328895865e-06,
+      "loss": 0.9717,
+      "step": 612
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.0847498005094277e-06,
+      "loss": 0.9779,
+      "step": 613
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.0790665731275764e-06,
+      "loss": 0.9444,
+      "step": 614
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.073380181796495e-06,
+      "loss": 0.9487,
+      "step": 615
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.0676906575859335e-06,
+      "loss": 0.9953,
+      "step": 616
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0619980315827585e-06,
+      "loss": 0.9632,
+      "step": 617
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.056302334890786e-06,
+      "loss": 0.9855,
+      "step": 618
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0506035986306095e-06,
+      "loss": 1.0145,
+      "step": 619
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0449018539394274e-06,
+      "loss": 0.9352,
+      "step": 620
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.039197131970881e-06,
+      "loss": 0.9987,
+      "step": 621
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0334894638948753e-06,
+      "loss": 0.997,
+      "step": 622
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.027778880897413e-06,
+      "loss": 0.9223,
+      "step": 623
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.022065414180425e-06,
+      "loss": 0.9186,
+      "step": 624
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.016349094961597e-06,
+      "loss": 0.9213,
+      "step": 625
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0106299544742013e-06,
+      "loss": 0.9768,
+      "step": 626
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.0049080239669243e-06,
+      "loss": 0.9867,
+      "step": 627
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.999183334703699e-06,
+      "loss": 1.0262,
+      "step": 628
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.9934559179635282e-06,
+      "loss": 0.9686,
+      "step": 629
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.9877258050403214e-06,
+      "loss": 0.9464,
+      "step": 630
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9819930272427162e-06,
+      "loss": 0.9852,
+      "step": 631
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9762576158939127e-06,
+      "loss": 0.9895,
+      "step": 632
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9705196023315007e-06,
+      "loss": 0.9486,
+      "step": 633
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.964779017907287e-06,
+      "loss": 0.9561,
+      "step": 634
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9590358939871255e-06,
+      "loss": 0.9368,
+      "step": 635
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9532902619507465e-06,
+      "loss": 0.9074,
+      "step": 636
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.947542153191583e-06,
+      "loss": 0.9821,
+      "step": 637
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.941791599116601e-06,
+      "loss": 0.9369,
+      "step": 638
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9360386311461276e-06,
+      "loss": 0.962,
+      "step": 639
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.93028328071368e-06,
+      "loss": 0.9871,
+      "step": 640
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.924525579265791e-06,
+      "loss": 0.9634,
+      "step": 641
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9187655582618413e-06,
+      "loss": 0.9026,
+      "step": 642
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9130032491738837e-06,
+      "loss": 1.0044,
+      "step": 643
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9072386834864723e-06,
+      "loss": 0.9724,
+      "step": 644
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.901471892696493e-06,
+      "loss": 0.9284,
+      "step": 645
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.895702908312987e-06,
+      "loss": 0.9205,
+      "step": 646
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8899317618569843e-06,
+      "loss": 0.9464,
+      "step": 647
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8841584848613254e-06,
+      "loss": 0.9183,
+      "step": 648
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8783831088704923e-06,
+      "loss": 0.9448,
+      "step": 649
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.872605665440436e-06,
+      "loss": 0.9878,
+      "step": 650
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8668261861384045e-06,
+      "loss": 0.8909,
+      "step": 651
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8610447025427685e-06,
+      "loss": 0.9777,
+      "step": 652
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8552612462428497e-06,
+      "loss": 0.9425,
+      "step": 653
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.849475848838749e-06,
+      "loss": 0.9325,
+      "step": 654
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.843688541941174e-06,
+      "loss": 0.9806,
+      "step": 655
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8378993571712638e-06,
+      "loss": 0.9987,
+      "step": 656
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.83210832616042e-06,
+      "loss": 0.9778,
+      "step": 657
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.82631548055013e-06,
+      "loss": 0.9316,
+      "step": 658
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.820520851991796e-06,
+      "loss": 0.9384,
+      "step": 659
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.814724472146564e-06,
+      "loss": 0.9542,
+      "step": 660
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8089263726851474e-06,
+      "loss": 0.9268,
+      "step": 661
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8031265852876537e-06,
+      "loss": 0.9525,
+      "step": 662
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7973251416434176e-06,
+      "loss": 0.9024,
+      "step": 663
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.791522073450819e-06,
+      "loss": 1.0033,
+      "step": 664
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7857174124171166e-06,
+      "loss": 0.9514,
+      "step": 665
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7799111902582697e-06,
+      "loss": 0.9808,
+      "step": 666
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7741034386987707e-06,
+      "loss": 0.9121,
+      "step": 667
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7682941894714664e-06,
+      "loss": 0.9992,
+      "step": 668
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.762483474317387e-06,
+      "loss": 0.9279,
+      "step": 669
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7566713249855715e-06,
+      "loss": 0.9323,
+      "step": 670
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7508577732328975e-06,
+      "loss": 0.97,
+      "step": 671
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7450428508239024e-06,
+      "loss": 0.936,
+      "step": 672
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7392265895306142e-06,
+      "loss": 0.9914,
+      "step": 673
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.733409021132377e-06,
+      "loss": 1.0271,
+      "step": 674
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7275901774156753e-06,
+      "loss": 0.9598,
+      "step": 675
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7217700901739637e-06,
+      "loss": 0.9653,
+      "step": 676
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.715948791207489e-06,
+      "loss": 0.99,
+      "step": 677
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.710126312323119e-06,
+      "loss": 0.9306,
+      "step": 678
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.704302685334171e-06,
+      "loss": 0.9499,
+      "step": 679
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.6984779420602324e-06,
+      "loss": 0.9608,
+      "step": 680
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.6926521143269914e-06,
+      "loss": 0.9216,
+      "step": 681
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.686825233966061e-06,
+      "loss": 0.9423,
+      "step": 682
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6809973328148058e-06,
+      "loss": 0.9326,
+      "step": 683
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6751684427161684e-06,
+      "loss": 1.0065,
+      "step": 684
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.669338595518494e-06,
+      "loss": 0.9795,
+      "step": 685
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.663507823075358e-06,
+      "loss": 0.9598,
+      "step": 686
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6576761572453903e-06,
+      "loss": 1.0292,
+      "step": 687
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.651843629892103e-06,
+      "loss": 0.964,
+      "step": 688
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6460102728837156e-06,
+      "loss": 1.0002,
+      "step": 689
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6401761180929798e-06,
+      "loss": 0.9562,
+      "step": 690
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6343411973970075e-06,
+      "loss": 0.9439,
+      "step": 691
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6285055426770935e-06,
+      "loss": 0.9781,
+      "step": 692
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6226691858185454e-06,
+      "loss": 0.9383,
+      "step": 693
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.616832158710506e-06,
+      "loss": 0.9084,
+      "step": 694
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6109944932457813e-06,
+      "loss": 0.9841,
+      "step": 695
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.6051562213206633e-06,
+      "loss": 0.9822,
+      "step": 696
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.599317374834759e-06,
+      "loss": 0.9477,
+      "step": 697
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.593477985690815e-06,
+      "loss": 0.9092,
+      "step": 698
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.587638085794541e-06,
+      "loss": 1.0064,
+      "step": 699
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.5817977070544408e-06,
+      "loss": 0.9513,
+      "step": 700
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.575956881381631e-06,
+      "loss": 0.9421,
+      "step": 701
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.5701156406896726e-06,
+      "loss": 0.9578,
+      "step": 702
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.564274016894393e-06,
+      "loss": 0.8667,
+      "step": 703
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.5584320419137127e-06,
+      "loss": 0.9441,
+      "step": 704
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5525897476674722e-06,
+      "loss": 0.9641,
+      "step": 705
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5467471660772557e-06,
+      "loss": 0.9316,
+      "step": 706
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5409043290662173e-06,
+      "loss": 0.9771,
+      "step": 707
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.535061268558906e-06,
+      "loss": 0.9355,
+      "step": 708
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5292180164810944e-06,
+      "loss": 0.9627,
+      "step": 709
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5233746047595984e-06,
+      "loss": 0.9704,
+      "step": 710
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5175310653221092e-06,
+      "loss": 0.9779,
+      "step": 711
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.511687430097014e-06,
+      "loss": 0.9617,
+      "step": 712
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5058437310132244e-06,
+      "loss": 0.9194,
+      "step": 713
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5e-06,
+      "loss": 0.9481,
+      "step": 714
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4941562689867755e-06,
+      "loss": 0.9846,
+      "step": 715
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.488312569902987e-06,
+      "loss": 0.9301,
+      "step": 716
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.482468934677891e-06,
+      "loss": 0.9383,
+      "step": 717
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4766253952404024e-06,
+      "loss": 1.0147,
+      "step": 718
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.470781983518906e-06,
+      "loss": 0.9243,
+      "step": 719
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.464938731441095e-06,
+      "loss": 0.9709,
+      "step": 720
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.459095670933783e-06,
+      "loss": 0.9498,
+      "step": 721
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.453252833922745e-06,
+      "loss": 0.9761,
+      "step": 722
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.447410252332528e-06,
+      "loss": 0.9085,
+      "step": 723
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.441567958086288e-06,
+      "loss": 0.923,
+      "step": 724
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.435725983105608e-06,
+      "loss": 0.9624,
+      "step": 725
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.429884359310328e-06,
+      "loss": 0.8979,
+      "step": 726
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.4240431186183695e-06,
+      "loss": 0.9808,
+      "step": 727
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.41820229294556e-06,
+      "loss": 0.9722,
+      "step": 728
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.41236191420546e-06,
+      "loss": 0.9669,
+      "step": 729
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.4065220143091863e-06,
+      "loss": 0.9785,
+      "step": 730
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.400682625165242e-06,
+      "loss": 0.9422,
+      "step": 731
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.394843778679338e-06,
+      "loss": 0.9611,
+      "step": 732
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.3890055067542195e-06,
+      "loss": 1.0296,
+      "step": 733
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.3831678412894947e-06,
+      "loss": 0.9736,
+      "step": 734
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.377330814181455e-06,
+      "loss": 0.97,
+      "step": 735
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.371494457322907e-06,
+      "loss": 0.9341,
+      "step": 736
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3656588026029938e-06,
+      "loss": 0.9229,
+      "step": 737
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3598238819070206e-06,
+      "loss": 1.0024,
+      "step": 738
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3539897271162853e-06,
+      "loss": 0.9758,
+      "step": 739
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.348156370107897e-06,
+      "loss": 0.972,
+      "step": 740
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3423238427546105e-06,
+      "loss": 0.9159,
+      "step": 741
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3364921769246423e-06,
+      "loss": 0.9294,
+      "step": 742
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3306614044815066e-06,
+      "loss": 0.9373,
+      "step": 743
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3248315572838316e-06,
+      "loss": 0.9159,
+      "step": 744
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.319002667185195e-06,
+      "loss": 0.9316,
+      "step": 745
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3131747660339396e-06,
+      "loss": 1.0052,
+      "step": 746
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.30734788567301e-06,
+      "loss": 0.9267,
+      "step": 747
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.301522057939768e-06,
+      "loss": 0.9735,
+      "step": 748
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.2956973146658303e-06,
+      "loss": 0.9429,
+      "step": 749
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.2898736876768816e-06,
+      "loss": 0.9163,
+      "step": 750
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.2840512087925127e-06,
+      "loss": 0.9068,
+      "step": 751
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.278229909826037e-06,
+      "loss": 0.9469,
+      "step": 752
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.272409822584325e-06,
+      "loss": 0.9846,
+      "step": 753
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.2665909788676236e-06,
+      "loss": 0.9236,
+      "step": 754
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.2607734104693866e-06,
+      "loss": 0.9558,
+      "step": 755
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2549571491760985e-06,
+      "loss": 0.8905,
+      "step": 756
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.249142226767104e-06,
+      "loss": 0.9602,
+      "step": 757
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2433286750144293e-06,
+      "loss": 0.941,
+      "step": 758
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.237516525682614e-06,
+      "loss": 0.8972,
+      "step": 759
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.231705810528534e-06,
+      "loss": 0.9617,
+      "step": 760
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.2258965613012293e-06,
+      "loss": 0.9597,
+      "step": 761
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.2200888097417308e-06,
+      "loss": 1.0112,
+      "step": 762
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.2142825875828838e-06,
+      "loss": 0.926,
+      "step": 763
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.2084779265491817e-06,
+      "loss": 0.8959,
+      "step": 764
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.2026748583565824e-06,
+      "loss": 0.9239,
+      "step": 765
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1968734147123467e-06,
+      "loss": 0.9029,
+      "step": 766
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.191073627314854e-06,
+      "loss": 0.9461,
+      "step": 767
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1852755278534373e-06,
+      "loss": 0.9881,
+      "step": 768
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1794791480082046e-06,
+      "loss": 0.9645,
+      "step": 769
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.173684519449872e-06,
+      "loss": 0.9578,
+      "step": 770
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.167891673839581e-06,
+      "loss": 0.9325,
+      "step": 771
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.162100642828737e-06,
+      "loss": 0.9361,
+      "step": 772
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1563114580588267e-06,
+      "loss": 0.893,
+      "step": 773
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.1505241511612522e-06,
+      "loss": 0.9928,
+      "step": 774
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.144738753757151e-06,
+      "loss": 0.9392,
+      "step": 775
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.138955297457233e-06,
+      "loss": 0.9513,
+      "step": 776
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.133173813861596e-06,
+      "loss": 0.9368,
+      "step": 777
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.1273943345595637e-06,
+      "loss": 0.9492,
+      "step": 778
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.1216168911295085e-06,
+      "loss": 0.9814,
+      "step": 779
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.1158415151386746e-06,
+      "loss": 0.9802,
+      "step": 780
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.110068238143016e-06,
+      "loss": 0.9804,
+      "step": 781
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.104297091687013e-06,
+      "loss": 0.9486,
+      "step": 782
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.098528107303508e-06,
+      "loss": 0.9405,
+      "step": 783
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0927613165135285e-06,
+      "loss": 0.9125,
+      "step": 784
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0869967508261175e-06,
+      "loss": 1.0126,
+      "step": 785
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0812344417381595e-06,
+      "loss": 0.9592,
+      "step": 786
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0754744207342097e-06,
+      "loss": 0.9148,
+      "step": 787
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0697167192863205e-06,
+      "loss": 0.9498,
+      "step": 788
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0639613688538733e-06,
+      "loss": 0.9695,
+      "step": 789
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0582084008834003e-06,
+      "loss": 0.9437,
+      "step": 790
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.052457846808419e-06,
+      "loss": 0.9514,
+      "step": 791
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0467097380492547e-06,
+      "loss": 0.9517,
+      "step": 792
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.040964106012876e-06,
+      "loss": 0.9405,
+      "step": 793
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.035220982092714e-06,
+      "loss": 0.9375,
+      "step": 794
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0294803976685006e-06,
+      "loss": 0.9194,
+      "step": 795
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0237423841060877e-06,
+      "loss": 0.9677,
+      "step": 796
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.018006972757285e-06,
+      "loss": 0.8989,
+      "step": 797
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.01227419495968e-06,
+      "loss": 0.899,
+      "step": 798
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.0065440820364718e-06,
+      "loss": 0.9571,
+      "step": 799
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.000816665296302e-06,
+      "loss": 0.9763,
+      "step": 800
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.9950919760330757e-06,
+      "loss": 0.9262,
+      "step": 801
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.9893700455257996e-06,
+      "loss": 0.9256,
+      "step": 802
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.9836509050384035e-06,
+      "loss": 0.9724,
+      "step": 803
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.9779345858195757e-06,
+      "loss": 0.9309,
+      "step": 804
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.972221119102587e-06,
+      "loss": 0.9407,
+      "step": 805
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.9665105361051255e-06,
+      "loss": 0.9503,
+      "step": 806
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.960802868029119e-06,
+      "loss": 0.9203,
+      "step": 807
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9550981460605734e-06,
+      "loss": 0.9184,
+      "step": 808
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.949396401369392e-06,
+      "loss": 0.9031,
+      "step": 809
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9436976651092143e-06,
+      "loss": 0.9571,
+      "step": 810
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.938001968417242e-06,
+      "loss": 0.978,
+      "step": 811
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9323093424140673e-06,
+      "loss": 0.922,
+      "step": 812
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.926619818203506e-06,
+      "loss": 0.9566,
+      "step": 813
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.920933426872425e-06,
+      "loss": 0.9721,
+      "step": 814
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.915250199490573e-06,
+      "loss": 0.9806,
+      "step": 815
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.909570167110415e-06,
+      "loss": 0.9575,
+      "step": 816
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.903893360766953e-06,
+      "loss": 0.9611,
+      "step": 817
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.8982198114775683e-06,
+      "loss": 1.0217,
+      "step": 818
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.8925495502418407e-06,
+      "loss": 0.9288,
+      "step": 819
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.8868826080413876e-06,
+      "loss": 0.9254,
+      "step": 820
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8812190158396931e-06,
+      "loss": 0.987,
+      "step": 821
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8755588045819325e-06,
+      "loss": 0.9327,
+      "step": 822
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.869902005194813e-06,
+      "loss": 0.9929,
+      "step": 823
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8642486485863954e-06,
+      "loss": 0.943,
+      "step": 824
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8585987656459337e-06,
+      "loss": 0.9563,
+      "step": 825
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.852952387243698e-06,
+      "loss": 0.9633,
+      "step": 826
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8473095442308145e-06,
+      "loss": 0.9946,
+      "step": 827
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.841670267439088e-06,
+      "loss": 0.9791,
+      "step": 828
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8360345876808422e-06,
+      "loss": 0.8966,
+      "step": 829
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8304025357487427e-06,
+      "loss": 0.9884,
+      "step": 830
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8247741424156373e-06,
+      "loss": 0.938,
+      "step": 831
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.81914943843438e-06,
+      "loss": 0.9376,
+      "step": 832
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.81352845453767e-06,
+      "loss": 0.9242,
+      "step": 833
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.8079112214378769e-06,
+      "loss": 0.9564,
+      "step": 834
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.80229776982688e-06,
+      "loss": 0.9728,
+      "step": 835
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7966881303758938e-06,
+      "loss": 0.9036,
+      "step": 836
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7910823337353062e-06,
+      "loss": 0.9069,
+      "step": 837
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7854804105345064e-06,
+      "loss": 0.9437,
+      "step": 838
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.779882391381721e-06,
+      "loss": 0.9606,
+      "step": 839
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7742883068638447e-06,
+      "loss": 0.9603,
+      "step": 840
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7686981875462733e-06,
+      "loss": 0.9593,
+      "step": 841
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7631120639727396e-06,
+      "loss": 0.9574,
+      "step": 842
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7575299666651413e-06,
+      "loss": 0.9399,
+      "step": 843
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7519519261233786e-06,
+      "loss": 0.915,
+      "step": 844
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7463779728251844e-06,
+      "loss": 0.9098,
+      "step": 845
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7408081372259633e-06,
+      "loss": 0.9693,
+      "step": 846
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7352424497586163e-06,
+      "loss": 0.8886,
+      "step": 847
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.729680940833383e-06,
+      "loss": 0.9601,
+      "step": 848
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7241236408376702e-06,
+      "loss": 0.8875,
+      "step": 849
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7185705801358892e-06,
+      "loss": 0.9323,
+      "step": 850
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7130217890692857e-06,
+      "loss": 0.9351,
+      "step": 851
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7074772979557802e-06,
+      "loss": 1.0109,
+      "step": 852
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.7019371370897953e-06,
+      "loss": 0.9329,
+      "step": 853
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6964013367420967e-06,
+      "loss": 0.9155,
+      "step": 854
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6908699271596213e-06,
+      "loss": 0.9401,
+      "step": 855
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6853429385653196e-06,
+      "loss": 0.9497,
+      "step": 856
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6798204011579824e-06,
+      "loss": 0.9576,
+      "step": 857
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6743023451120831e-06,
+      "loss": 0.9111,
+      "step": 858
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.668788800577606e-06,
+      "loss": 0.9729,
+      "step": 859
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6632797976798887e-06,
+      "loss": 0.9423,
+      "step": 860
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6577753665194502e-06,
+      "loss": 0.9639,
+      "step": 861
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6522755371718333e-06,
+      "loss": 0.9538,
+      "step": 862
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6467803396874365e-06,
+      "loss": 0.8871,
+      "step": 863
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6412898040913471e-06,
+      "loss": 0.9833,
+      "step": 864
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.635803960383186e-06,
+      "loss": 0.977,
+      "step": 865
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6303228385369318e-06,
+      "loss": 0.9868,
+      "step": 866
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6248464685007692e-06,
+      "loss": 0.9455,
+      "step": 867
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6193748801969164e-06,
+      "loss": 1.0272,
+      "step": 868
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6139081035214654e-06,
+      "loss": 0.9529,
+      "step": 869
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6084461683442176e-06,
+      "loss": 0.9141,
+      "step": 870
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.6029891045085224e-06,
+      "loss": 0.957,
+      "step": 871
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5975369418311113e-06,
+      "loss": 0.9703,
+      "step": 872
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5920897101019384e-06,
+      "loss": 0.9496,
+      "step": 873
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5866474390840126e-06,
+      "loss": 0.9136,
+      "step": 874
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5812101585132416e-06,
+      "loss": 0.9655,
+      "step": 875
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5757778980982627e-06,
+      "loss": 0.9827,
+      "step": 876
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5703506875202868e-06,
+      "loss": 0.9365,
+      "step": 877
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5649285564329296e-06,
+      "loss": 0.8948,
+      "step": 878
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5595115344620575e-06,
+      "loss": 0.9586,
+      "step": 879
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5540996512056164e-06,
+      "loss": 1.0013,
+      "step": 880
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.548692936233479e-06,
+      "loss": 0.9254,
+      "step": 881
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5432914190872757e-06,
+      "loss": 0.9348,
+      "step": 882
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5378951292802397e-06,
+      "loss": 0.9236,
+      "step": 883
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5325040962970417e-06,
+      "loss": 0.9352,
+      "step": 884
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5271183495936273e-06,
+      "loss": 0.9244,
+      "step": 885
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.521737918597063e-06,
+      "loss": 0.968,
+      "step": 886
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5163628327053661e-06,
+      "loss": 0.8781,
+      "step": 887
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5109931212873535e-06,
+      "loss": 0.9887,
+      "step": 888
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5056288136824725e-06,
+      "loss": 0.9156,
+      "step": 889
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.500269939200648e-06,
+      "loss": 0.9018,
+      "step": 890
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.494916527122116e-06,
+      "loss": 0.8993,
+      "step": 891
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4895686066972703e-06,
+      "loss": 0.9816,
+      "step": 892
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4842262071464941e-06,
+      "loss": 0.9634,
+      "step": 893
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.47888935766001e-06,
+      "loss": 0.9772,
+      "step": 894
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4735580873977125e-06,
+      "loss": 0.9437,
+      "step": 895
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4682324254890135e-06,
+      "loss": 0.9355,
+      "step": 896
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4629124010326808e-06,
+      "loss": 0.9336,
+      "step": 897
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4575980430966808e-06,
+      "loss": 0.9904,
+      "step": 898
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4522893807180163e-06,
+      "loss": 0.9514,
+      "step": 899
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.446986442902574e-06,
+      "loss": 0.9456,
+      "step": 900
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4416892586249586e-06,
+      "loss": 0.9479,
+      "step": 901
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4363978568283412e-06,
+      "loss": 0.934,
+      "step": 902
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4311122664242955e-06,
+      "loss": 0.9653,
+      "step": 903
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4258325162926441e-06,
+      "loss": 0.922,
+      "step": 904
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4205586352813e-06,
+      "loss": 0.966,
+      "step": 905
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.415290652206105e-06,
+      "loss": 0.9418,
+      "step": 906
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4100285958506785e-06,
+      "loss": 0.9558,
+      "step": 907
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.404772494966254e-06,
+      "loss": 0.959,
+      "step": 908
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.399522378271527e-06,
+      "loss": 0.9413,
+      "step": 909
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3942782744524974e-06,
+      "loss": 0.93,
+      "step": 910
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.389040212162307e-06,
+      "loss": 0.926,
+      "step": 911
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3838082200210932e-06,
+      "loss": 0.9447,
+      "step": 912
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.378582326615821e-06,
+      "loss": 0.926,
+      "step": 913
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3733625605001365e-06,
+      "loss": 0.9709,
+      "step": 914
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3681489501942077e-06,
+      "loss": 0.9275,
+      "step": 915
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.362941524184564e-06,
+      "loss": 0.9237,
+      "step": 916
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3577403109239485e-06,
+      "loss": 0.9561,
+      "step": 917
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3525453388311554e-06,
+      "loss": 0.9759,
+      "step": 918
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.347356636290882e-06,
+      "loss": 0.9413,
+      "step": 919
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.342174231653564e-06,
+      "loss": 0.919,
+      "step": 920
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3369981532352317e-06,
+      "loss": 0.927,
+      "step": 921
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.331828429317345e-06,
+      "loss": 0.9093,
+      "step": 922
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.326665088146648e-06,
+      "loss": 0.9802,
+      "step": 923
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3215081579350058e-06,
+      "loss": 0.9277,
+      "step": 924
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3163576668592581e-06,
+      "loss": 0.924,
+      "step": 925
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3112136430610623e-06,
+      "loss": 0.9419,
+      "step": 926
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.306076114646735e-06,
+      "loss": 0.8907,
+      "step": 927
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.3009451096871084e-06,
+      "loss": 0.8949,
+      "step": 928
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2958206562173664e-06,
+      "loss": 0.9464,
+      "step": 929
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2907027822369006e-06,
+      "loss": 0.9255,
+      "step": 930
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2855915157091498e-06,
+      "loss": 0.9453,
+      "step": 931
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2804868845614527e-06,
+      "loss": 0.9294,
+      "step": 932
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2753889166848909e-06,
+      "loss": 0.9457,
+      "step": 933
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2702976399341422e-06,
+      "loss": 0.899,
+      "step": 934
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.26521308212732e-06,
+      "loss": 0.9364,
+      "step": 935
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2601352710458314e-06,
+      "loss": 0.9537,
+      "step": 936
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2550642344342155e-06,
+      "loss": 0.9344,
+      "step": 937
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2500000000000007e-06,
+      "loss": 0.9676,
+      "step": 938
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2449425954135452e-06,
+      "loss": 0.9229,
+      "step": 939
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2398920483078939e-06,
+      "loss": 0.9467,
+      "step": 940
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2348483862786188e-06,
+      "loss": 0.9155,
+      "step": 941
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2298116368836772e-06,
+      "loss": 0.9029,
+      "step": 942
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2247818276432522e-06,
+      "loss": 0.9271,
+      "step": 943
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.219758986039611e-06,
+      "loss": 0.9819,
+      "step": 944
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.214743139516946e-06,
+      "loss": 0.9531,
+      "step": 945
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.2097343154812332e-06,
+      "loss": 0.938,
+      "step": 946
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.2047325413000782e-06,
+      "loss": 0.9186,
+      "step": 947
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.1997378443025633e-06,
+      "loss": 0.9543,
+      "step": 948
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.1947502517791073e-06,
+      "loss": 0.9125,
+      "step": 949
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1897697909813058e-06,
+      "loss": 0.9185,
+      "step": 950
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1847964891217923e-06,
+      "loss": 0.9263,
+      "step": 951
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1798303733740801e-06,
+      "loss": 0.8837,
+      "step": 952
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1748714708724232e-06,
+      "loss": 0.8839,
+      "step": 953
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.169919808711659e-06,
+      "loss": 0.9568,
+      "step": 954
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1649754139470679e-06,
+      "loss": 0.9485,
+      "step": 955
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.160038313594219e-06,
+      "loss": 0.9184,
+      "step": 956
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1551085346288296e-06,
+      "loss": 0.9129,
+      "step": 957
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1501861039866095e-06,
+      "loss": 0.9151,
+      "step": 958
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.1452710485631216e-06,
+      "loss": 0.9566,
+      "step": 959
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.140363395213629e-06,
+      "loss": 0.9041,
+      "step": 960
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.1354631707529532e-06,
+      "loss": 0.9707,
+      "step": 961
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.130570401955322e-06,
+      "loss": 0.9317,
+      "step": 962
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.1256851155542297e-06,
+      "loss": 0.9001,
+      "step": 963
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1208073382422866e-06,
+      "loss": 0.8462,
+      "step": 964
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1159370966710723e-06,
+      "loss": 0.9704,
+      "step": 965
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1110744174509952e-06,
+      "loss": 0.9049,
+      "step": 966
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1062193271511408e-06,
+      "loss": 0.9542,
+      "step": 967
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.1013718522991315e-06,
+      "loss": 0.9507,
+      "step": 968
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0965320193809808e-06,
+      "loss": 0.9249,
+      "step": 969
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0916998548409449e-06,
+      "loss": 0.9098,
+      "step": 970
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0868753850813826e-06,
+      "loss": 0.9589,
+      "step": 971
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0820586364626103e-06,
+      "loss": 0.8898,
+      "step": 972
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.077249635302754e-06,
+      "loss": 0.861,
+      "step": 973
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0724484078776121e-06,
+      "loss": 0.9215,
+      "step": 974
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0676549804205048e-06,
+      "loss": 0.9303,
+      "step": 975
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0628693791221373e-06,
+      "loss": 0.9534,
+      "step": 976
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.058091630130451e-06,
+      "loss": 0.9405,
+      "step": 977
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.0533217595504859e-06,
+      "loss": 0.9049,
+      "step": 978
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.0485597934442323e-06,
+      "loss": 0.9695,
+      "step": 979
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.043805757830495e-06,
+      "loss": 0.888,
+      "step": 980
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.0390596786847435e-06,
+      "loss": 0.9354,
+      "step": 981
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0343215819389782e-06,
+      "loss": 0.9306,
+      "step": 982
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0295914934815806e-06,
+      "loss": 0.894,
+      "step": 983
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.02486943915718e-06,
+      "loss": 0.9514,
+      "step": 984
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0201554447665044e-06,
+      "loss": 0.9857,
+      "step": 985
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0154495360662464e-06,
+      "loss": 0.9448,
+      "step": 986
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.0107517387689168e-06,
+      "loss": 0.9353,
+      "step": 987
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.0060620785427083e-06,
+      "loss": 0.935,
+      "step": 988
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.001380581011354e-06,
+      "loss": 0.932,
+      "step": 989
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.967072717539852e-07,
+      "loss": 0.9449,
+      "step": 990
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.920421763049957e-07,
+      "loss": 0.9358,
+      "step": 991
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.873853201538972e-07,
+      "loss": 0.8898,
+      "step": 992
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.82736728745186e-07,
+      "loss": 0.9258,
+      "step": 993
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.780964274781984e-07,
+      "loss": 0.8816,
+      "step": 994
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.734644417069764e-07,
+      "loss": 0.9727,
+      "step": 995
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.688407967401248e-07,
+      "loss": 0.9296,
+      "step": 996
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.642255178406782e-07,
+      "loss": 0.8672,
+      "step": 997
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.596186302259563e-07,
+      "loss": 0.9055,
+      "step": 998
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.550201590674343e-07,
+      "loss": 0.9666,
+      "step": 999
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.504301294905966e-07,
+      "loss": 0.9575,
+      "step": 1000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.458485665748071e-07,
+      "loss": 0.9364,
+      "step": 1001
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.412754953531664e-07,
+      "loss": 0.9803,
+      "step": 1002
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.367109408123803e-07,
+      "loss": 0.922,
+      "step": 1003
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.32154927892617e-07,
+      "loss": 0.9241,
+      "step": 1004
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.276074814873778e-07,
+      "loss": 0.9279,
+      "step": 1005
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.230686264433547e-07,
+      "loss": 0.9172,
+      "step": 1006
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.185383875603004e-07,
+      "loss": 0.936,
+      "step": 1007
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.140167895908867e-07,
+      "loss": 0.9635,
+      "step": 1008
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.095038572405751e-07,
+      "loss": 0.9222,
+      "step": 1009
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 9.04999615167479e-07,
+      "loss": 0.8934,
+      "step": 1010
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 9.005040879822269e-07,
+      "loss": 0.9572,
+      "step": 1011
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.960173002478336e-07,
+      "loss": 0.9636,
+      "step": 1012
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.915392764795592e-07,
+      "loss": 0.9133,
+      "step": 1013
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.870700411447817e-07,
+      "loss": 0.9111,
+      "step": 1014
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.826096186628568e-07,
+      "loss": 0.957,
+      "step": 1015
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.781580334049919e-07,
+      "loss": 0.9085,
+      "step": 1016
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.737153096941045e-07,
+      "loss": 0.9445,
+      "step": 1017
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.692814718046979e-07,
+      "loss": 0.9714,
+      "step": 1018
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.648565439627205e-07,
+      "loss": 0.923,
+      "step": 1019
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.604405503454399e-07,
+      "loss": 0.9227,
+      "step": 1020
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.560335150813081e-07,
+      "loss": 0.9158,
+      "step": 1021
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.516354622498279e-07,
+      "loss": 0.9367,
+      "step": 1022
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.472464158814256e-07,
+      "loss": 0.9377,
+      "step": 1023
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.428663999573142e-07,
+      "loss": 0.9393,
+      "step": 1024
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.384954384093682e-07,
+      "loss": 0.9491,
+      "step": 1025
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.341335551199903e-07,
+      "loss": 0.9279,
+      "step": 1026
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.297807739219777e-07,
+      "loss": 0.9395,
+      "step": 1027
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.254371185983981e-07,
+      "loss": 0.9241,
+      "step": 1028
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.21102612882454e-07,
+      "loss": 0.9359,
+      "step": 1029
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.16777280457357e-07,
+      "loss": 0.9254,
+      "step": 1030
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.124611449561981e-07,
+      "loss": 0.9077,
+      "step": 1031
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.081542299618139e-07,
+      "loss": 0.9492,
+      "step": 1032
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 8.038565590066652e-07,
+      "loss": 0.9483,
+      "step": 1033
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.995681555727011e-07,
+      "loss": 0.9,
+      "step": 1034
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.952890430912374e-07,
+      "loss": 0.902,
+      "step": 1035
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.910192449428216e-07,
+      "loss": 0.9446,
+      "step": 1036
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.867587844571126e-07,
+      "loss": 0.9459,
+      "step": 1037
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.825076849127458e-07,
+      "loss": 0.9479,
+      "step": 1038
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.78265969537213e-07,
+      "loss": 0.9195,
+      "step": 1039
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.740336615067293e-07,
+      "loss": 0.9123,
+      "step": 1040
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.698107839461111e-07,
+      "loss": 0.9716,
+      "step": 1041
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.655973599286459e-07,
+      "loss": 0.9298,
+      "step": 1042
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.613934124759712e-07,
+      "loss": 0.9233,
+      "step": 1043
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.57198964557942e-07,
+      "loss": 0.8915,
+      "step": 1044
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.530140390925125e-07,
+      "loss": 0.9297,
+      "step": 1045
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.488386589456043e-07,
+      "loss": 0.892,
+      "step": 1046
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.446728469309872e-07,
+      "loss": 0.9365,
+      "step": 1047
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.405166258101495e-07,
+      "loss": 0.8845,
+      "step": 1048
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.363700182921784e-07,
+      "loss": 0.8854,
+      "step": 1049
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.322330470336314e-07,
+      "loss": 0.9162,
+      "step": 1050
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.28105734638416e-07,
+      "loss": 0.9666,
+      "step": 1051
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.239881036576652e-07,
+      "loss": 0.9236,
+      "step": 1052
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.198801765896115e-07,
+      "loss": 0.9678,
+      "step": 1053
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.157819758794699e-07,
+      "loss": 0.9487,
+      "step": 1054
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.116935239193085e-07,
+      "loss": 0.9025,
+      "step": 1055
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 7.076148430479321e-07,
+      "loss": 0.9176,
+      "step": 1056
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 7.035459555507549e-07,
+      "loss": 0.9478,
+      "step": 1057
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.994868836596841e-07,
+      "loss": 0.9384,
+      "step": 1058
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.954376495529927e-07,
+      "loss": 0.918,
+      "step": 1059
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.91398275355204e-07,
+      "loss": 0.9073,
+      "step": 1060
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.873687831369655e-07,
+      "loss": 0.9752,
+      "step": 1061
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.833491949149329e-07,
+      "loss": 0.8883,
+      "step": 1062
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.793395326516458e-07,
+      "loss": 0.9172,
+      "step": 1063
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.753398182554116e-07,
+      "loss": 0.9353,
+      "step": 1064
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.713500735801811e-07,
+      "loss": 0.8938,
+      "step": 1065
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.673703204254348e-07,
+      "loss": 0.9346,
+      "step": 1066
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.63400580536058e-07,
+      "loss": 0.9675,
+      "step": 1067
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.594408756022272e-07,
+      "loss": 0.9397,
+      "step": 1068
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.554912272592867e-07,
+      "loss": 0.9519,
+      "step": 1069
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.515516570876351e-07,
+      "loss": 0.9028,
+      "step": 1070
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.476221866126029e-07,
+      "loss": 0.913,
+      "step": 1071
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.437028373043386e-07,
+      "loss": 0.9498,
+      "step": 1072
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.397936305776903e-07,
+      "loss": 0.915,
+      "step": 1073
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.358945877920861e-07,
+      "loss": 0.922,
+      "step": 1074
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.320057302514223e-07,
+      "loss": 0.9657,
+      "step": 1075
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.281270792039418e-07,
+      "loss": 0.9277,
+      "step": 1076
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.242586558421216e-07,
+      "loss": 0.9215,
+      "step": 1077
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.204004813025569e-07,
+      "loss": 0.9901,
+      "step": 1078
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.16552576665842e-07,
+      "loss": 0.9262,
+      "step": 1079
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.127149629564605e-07,
+      "loss": 0.9609,
+      "step": 1080
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.08887661142665e-07,
+      "loss": 0.9429,
+      "step": 1081
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.050706921363672e-07,
+      "loss": 0.935,
+      "step": 1082
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.012640767930217e-07,
+      "loss": 0.9699,
+      "step": 1083
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.974678359115094e-07,
+      "loss": 0.8779,
+      "step": 1084
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.936819902340299e-07,
+      "loss": 0.9135,
+      "step": 1085
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.899065604459814e-07,
+      "loss": 0.9218,
+      "step": 1086
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.861415671758536e-07,
+      "loss": 0.9553,
+      "step": 1087
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.823870309951096e-07,
+      "loss": 0.9289,
+      "step": 1088
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.786429724180801e-07,
+      "loss": 0.9148,
+      "step": 1089
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.749094119018431e-07,
+      "loss": 0.9048,
+      "step": 1090
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.711863698461198e-07,
+      "loss": 0.9105,
+      "step": 1091
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.674738665931575e-07,
+      "loss": 0.9411,
+      "step": 1092
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.637719224276228e-07,
+      "loss": 0.8819,
+      "step": 1093
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.600805575764878e-07,
+      "loss": 0.9297,
+      "step": 1094
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.563997922089193e-07,
+      "loss": 0.949,
+      "step": 1095
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.527296464361725e-07,
+      "loss": 0.9506,
+      "step": 1096
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.490701403114757e-07,
+      "loss": 0.9317,
+      "step": 1097
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.454212938299256e-07,
+      "loss": 0.9344,
+      "step": 1098
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.417831269283741e-07,
+      "loss": 0.9196,
+      "step": 1099
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.381556594853226e-07,
+      "loss": 0.9378,
+      "step": 1100
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.34538911320811e-07,
+      "loss": 0.9185,
+      "step": 1101
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.309329021963116e-07,
+      "loss": 0.8968,
+      "step": 1102
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.273376518146181e-07,
+      "loss": 0.9922,
+      "step": 1103
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.237531798197415e-07,
+      "loss": 0.9227,
+      "step": 1104
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.201795057967995e-07,
+      "loss": 0.9479,
+      "step": 1105
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.166166492719124e-07,
+      "loss": 0.9345,
+      "step": 1106
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.130646297120936e-07,
+      "loss": 0.9065,
+      "step": 1107
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.095234665251464e-07,
+      "loss": 0.9095,
+      "step": 1108
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.059931790595538e-07,
+      "loss": 0.9045,
+      "step": 1109
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.02473786604378e-07,
+      "loss": 0.9432,
+      "step": 1110
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.989653083891493e-07,
+      "loss": 0.9072,
+      "step": 1111
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.954677635837668e-07,
+      "loss": 0.9502,
+      "step": 1112
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.91981171298388e-07,
+      "loss": 0.9157,
+      "step": 1113
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.885055505833291e-07,
+      "loss": 0.8765,
+      "step": 1114
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.850409204289591e-07,
+      "loss": 0.96,
+      "step": 1115
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.815872997655941e-07,
+      "loss": 0.9285,
+      "step": 1116
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.781447074633977e-07,
+      "loss": 0.9059,
+      "step": 1117
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.747131623322737e-07,
+      "loss": 0.9177,
+      "step": 1118
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.712926831217685e-07,
+      "loss": 0.9337,
+      "step": 1119
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.678832885209622e-07,
+      "loss": 0.9281,
+      "step": 1120
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.644849971583729e-07,
+      "loss": 0.9064,
+      "step": 1121
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.610978276018496e-07,
+      "loss": 0.9417,
+      "step": 1122
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.5772179835847584e-07,
+      "loss": 0.9312,
+      "step": 1123
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.543569278744625e-07,
+      "loss": 0.8847,
+      "step": 1124
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.5100323453505344e-07,
+      "loss": 0.9596,
+      "step": 1125
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.476607366644192e-07,
+      "loss": 0.9405,
+      "step": 1126
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.4432945252556285e-07,
+      "loss": 0.9433,
+      "step": 1127
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.410094003202134e-07,
+      "loss": 0.8896,
+      "step": 1128
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.3770059818873335e-07,
+      "loss": 0.9046,
+      "step": 1129
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.344030642100133e-07,
+      "loss": 0.8966,
+      "step": 1130
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.3111681640137763e-07,
+      "loss": 0.9305,
+      "step": 1131
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.278418727184849e-07,
+      "loss": 0.908,
+      "step": 1132
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.2457825105522736e-07,
+      "loss": 0.9385,
+      "step": 1133
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.2132596924363666e-07,
+      "loss": 0.963,
+      "step": 1134
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.180850450537857e-07,
+      "loss": 0.9046,
+      "step": 1135
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.1485549619368765e-07,
+      "loss": 0.9561,
+      "step": 1136
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.116373403092058e-07,
+      "loss": 0.9511,
+      "step": 1137
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.0843059498395063e-07,
+      "loss": 0.9188,
+      "step": 1138
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.052352777391885e-07,
+      "loss": 0.9454,
+      "step": 1139
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.020514060337447e-07,
+      "loss": 0.9239,
+      "step": 1140
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.9887899726390455e-07,
+      "loss": 0.8879,
+      "step": 1141
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.9571806876332527e-07,
+      "loss": 0.9573,
+      "step": 1142
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.925686378029331e-07,
+      "loss": 0.954,
+      "step": 1143
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.894307215908372e-07,
+      "loss": 0.9446,
+      "step": 1144
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.863043372722278e-07,
+      "loss": 0.8918,
+      "step": 1145
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.831895019292897e-07,
+      "loss": 0.942,
+      "step": 1146
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.800862325811025e-07,
+      "loss": 0.9039,
+      "step": 1147
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.7699454618355306e-07,
+      "loss": 0.9342,
+      "step": 1148
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.7391445962923824e-07,
+      "loss": 0.8648,
+      "step": 1149
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.708459897473768e-07,
+      "loss": 0.9602,
+      "step": 1150
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.677891533037134e-07,
+      "loss": 0.9268,
+      "step": 1151
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.647439670004316e-07,
+      "loss": 0.9866,
+      "step": 1152
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.617104474760569e-07,
+      "loss": 0.9394,
+      "step": 1153
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.586886113053717e-07,
+      "loss": 0.9666,
+      "step": 1154
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.5567847499932004e-07,
+      "loss": 0.9307,
+      "step": 1155
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.5268005500492004e-07,
+      "loss": 0.9144,
+      "step": 1156
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.496933677051745e-07,
+      "loss": 0.9273,
+      "step": 1157
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.4671842941897764e-07,
+      "loss": 0.9137,
+      "step": 1158
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.4375525640103103e-07,
+      "loss": 0.9542,
+      "step": 1159
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.408038648417503e-07,
+      "loss": 0.889,
+      "step": 1160
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.3786427086718006e-07,
+      "loss": 0.9567,
+      "step": 1161
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.3493649053890325e-07,
+      "loss": 0.8687,
+      "step": 1162
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.320205398539561e-07,
+      "loss": 0.9059,
+      "step": 1163
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.2911643474473647e-07,
+      "loss": 0.9958,
+      "step": 1164
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.262241910789221e-07,
+      "loss": 0.9087,
+      "step": 1165
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.233438246593795e-07,
+      "loss": 0.9522,
+      "step": 1166
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.2047535122408076e-07,
+      "loss": 0.9409,
+      "step": 1167
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.1761878644601425e-07,
+      "loss": 0.9107,
+      "step": 1168
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.147741459331033e-07,
+      "loss": 0.9359,
+      "step": 1169
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.119414452281158e-07,
+      "loss": 0.9227,
+      "step": 1170
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.0912069980858496e-07,
+      "loss": 0.9746,
+      "step": 1171
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.063119250867186e-07,
+      "loss": 0.9445,
+      "step": 1172
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.035151364093214e-07,
+      "loss": 0.9535,
+      "step": 1173
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.007303490577043e-07,
+      "loss": 0.9401,
+      "step": 1174
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.9795757824760796e-07,
+      "loss": 0.8791,
+      "step": 1175
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.9519683912911267e-07,
+      "loss": 0.9475,
+      "step": 1176
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.924481467865617e-07,
+      "loss": 0.8539,
+      "step": 1177
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.897115162384759e-07,
+      "loss": 0.9217,
+      "step": 1178
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.869869624374699e-07,
+      "loss": 0.9007,
+      "step": 1179
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.8427450027017493e-07,
+      "loss": 0.9116,
+      "step": 1180
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.815741445571529e-07,
+      "loss": 0.9754,
+      "step": 1181
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.788859100528196e-07,
+      "loss": 0.9038,
+      "step": 1182
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.7620981144536014e-07,
+      "loss": 0.9046,
+      "step": 1183
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.7354586335665205e-07,
+      "loss": 0.9065,
+      "step": 1184
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.708940803421825e-07,
+      "loss": 0.9277,
+      "step": 1185
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.6825447689097174e-07,
+      "loss": 0.9043,
+      "step": 1186
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.6562706742549073e-07,
+      "loss": 0.8937,
+      "step": 1187
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.630118663015849e-07,
+      "loss": 0.897,
+      "step": 1188
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.604088878083957e-07,
+      "loss": 0.9135,
+      "step": 1189
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.5781814616827936e-07,
+      "loss": 0.9142,
+      "step": 1190
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.552396555367334e-07,
+      "loss": 0.9192,
+      "step": 1191
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.526734300023162e-07,
+      "loss": 0.9095,
+      "step": 1192
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.501194835865717e-07,
+      "loss": 0.9301,
+      "step": 1193
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.4757783024395244e-07,
+      "loss": 0.8825,
+      "step": 1194
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.450484838617417e-07,
+      "loss": 0.9234,
+      "step": 1195
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.4253145825998134e-07,
+      "loss": 0.8806,
+      "step": 1196
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.400267671913917e-07,
+      "loss": 0.9751,
+      "step": 1197
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.3753442434129998e-07,
+      "loss": 0.8983,
+      "step": 1198
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.350544433275645e-07,
+      "loss": 0.9057,
+      "step": 1199
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.325868377004986e-07,
+      "loss": 0.9366,
+      "step": 1200
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.3013162094279977e-07,
+      "loss": 0.8985,
+      "step": 1201
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.2768880646947268e-07,
+      "loss": 0.956,
+      "step": 1202
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.2525840762775863e-07,
+      "loss": 0.9426,
+      "step": 1203
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.2284043769706026e-07,
+      "loss": 0.95,
+      "step": 1204
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.2043490988887168e-07,
+      "loss": 0.9499,
+      "step": 1205
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1804183734670277e-07,
+      "loss": 0.9425,
+      "step": 1206
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1566123314601118e-07,
+      "loss": 0.9218,
+      "step": 1207
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1329311029412796e-07,
+      "loss": 0.9175,
+      "step": 1208
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.1093748173018846e-07,
+      "loss": 0.9306,
+      "step": 1209
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0859436032505954e-07,
+      "loss": 1.0025,
+      "step": 1210
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0626375888127187e-07,
+      "loss": 0.9397,
+      "step": 1211
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.039456901329473e-07,
+      "loss": 0.9058,
+      "step": 1212
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.0164016674573185e-07,
+      "loss": 0.9113,
+      "step": 1213
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9934720131672357e-07,
+      "loss": 0.9336,
+      "step": 1214
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9706680637440745e-07,
+      "loss": 0.9335,
+      "step": 1215
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9479899437858246e-07,
+      "loss": 0.9406,
+      "step": 1216
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9254377772029847e-07,
+      "loss": 0.929,
+      "step": 1217
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.9030116872178317e-07,
+      "loss": 0.9002,
+      "step": 1218
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8807117963637988e-07,
+      "loss": 0.8925,
+      "step": 1219
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8585382264847795e-07,
+      "loss": 0.979,
+      "step": 1220
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.836491098734447e-07,
+      "loss": 0.9463,
+      "step": 1221
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.8145705335756298e-07,
+      "loss": 0.8732,
+      "step": 1222
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7927766507796208e-07,
+      "loss": 0.9699,
+      "step": 1223
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.771109569425547e-07,
+      "loss": 0.9213,
+      "step": 1224
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7495694078996984e-07,
+      "loss": 0.926,
+      "step": 1225
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7281562838948968e-07,
+      "loss": 0.9389,
+      "step": 1226
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.7068703144098365e-07,
+      "loss": 0.8797,
+      "step": 1227
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6857116157484755e-07,
+      "loss": 0.9517,
+      "step": 1228
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6646803035193532e-07,
+      "loss": 0.9278,
+      "step": 1229
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6437764926350074e-07,
+      "loss": 0.9569,
+      "step": 1230
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6230002973113056e-07,
+      "loss": 0.9627,
+      "step": 1231
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.602351831066862e-07,
+      "loss": 0.9189,
+      "step": 1232
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5818312067223673e-07,
+      "loss": 0.9116,
+      "step": 1233
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.561438536400023e-07,
+      "loss": 0.8899,
+      "step": 1234
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5411739315228909e-07,
+      "loss": 0.8823,
+      "step": 1235
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.5210375028143097e-07,
+      "loss": 0.9304,
+      "step": 1236
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.5010293602972653e-07,
+      "loss": 0.9448,
+      "step": 1237
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.4811496132938196e-07,
+      "loss": 0.9032,
+      "step": 1238
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.4613983704244826e-07,
+      "loss": 0.9009,
+      "step": 1239
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.4417757396076427e-07,
+      "loss": 0.8852,
+      "step": 1240
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.422281828058969e-07,
+      "loss": 0.9707,
+      "step": 1241
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.4029167422908107e-07,
+      "loss": 0.8705,
+      "step": 1242
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.3836805881116495e-07,
+      "loss": 0.8931,
+      "step": 1243
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.364573470625477e-07,
+      "loss": 0.9079,
+      "step": 1244
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.345595494231261e-07,
+      "loss": 0.9481,
+      "step": 1245
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.3267467626223606e-07,
+      "loss": 0.9465,
+      "step": 1246
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.3080273787859367e-07,
+      "loss": 0.9324,
+      "step": 1247
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.2894374450024338e-07,
+      "loss": 0.9284,
+      "step": 1248
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.2709770628449776e-07,
+      "loss": 0.9089,
+      "step": 1249
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.2526463331788503e-07,
+      "loss": 0.9026,
+      "step": 1250
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.234445356160932e-07,
+      "loss": 0.969,
+      "step": 1251
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.2163742312391342e-07,
+      "loss": 0.8986,
+      "step": 1252
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.1984330571518932e-07,
+      "loss": 0.8886,
+      "step": 1253
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.1806219319275918e-07,
+      "loss": 0.8814,
+      "step": 1254
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1629409528840534e-07,
+      "loss": 0.9806,
+      "step": 1255
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1453902166279895e-07,
+      "loss": 0.8926,
+      "step": 1256
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1279698190544918e-07,
+      "loss": 0.9208,
+      "step": 1257
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1106798553464804e-07,
+      "loss": 0.9196,
+      "step": 1258
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0935204199742255e-07,
+      "loss": 0.9761,
+      "step": 1259
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0764916066947795e-07,
+      "loss": 0.9574,
+      "step": 1260
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0595935085515069e-07,
+      "loss": 0.9209,
+      "step": 1261
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.042826217873566e-07,
+      "loss": 0.9717,
+      "step": 1262
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0261898262753811e-07,
+      "loss": 0.9158,
+      "step": 1263
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.0096844246561794e-07,
+      "loss": 0.9216,
+      "step": 1264
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.933101031994547e-08,
+      "loss": 0.8997,
+      "step": 1265
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.770669513725128e-08,
+      "loss": 0.9223,
+      "step": 1266
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.609550579259497e-08,
+      "loss": 0.9327,
+      "step": 1267
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.449745108931985e-08,
+      "loss": 0.9007,
+      "step": 1268
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.291253975900138e-08,
+      "loss": 0.922,
+      "step": 1269
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.134078046140249e-08,
+      "loss": 0.9623,
+      "step": 1270
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.978218178442439e-08,
+      "loss": 0.9201,
+      "step": 1271
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.823675224406052e-08,
+      "loss": 0.95,
+      "step": 1272
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.670450028434946e-08,
+      "loss": 0.9226,
+      "step": 1273
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.518543427732951e-08,
+      "loss": 0.9372,
+      "step": 1274
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.367956252299192e-08,
+      "loss": 0.9366,
+      "step": 1275
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.218689324923729e-08,
+      "loss": 0.9333,
+      "step": 1276
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.070743461182807e-08,
+      "loss": 0.9515,
+      "step": 1277
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.924119469434666e-08,
+      "loss": 0.9177,
+      "step": 1278
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.778818150814854e-08,
+      "loss": 0.9572,
+      "step": 1279
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.634840299232171e-08,
+      "loss": 0.9564,
+      "step": 1280
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.492186701364007e-08,
+      "loss": 0.9186,
+      "step": 1281
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.350858136652262e-08,
+      "loss": 0.9696,
+      "step": 1282
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 7.210855377299014e-08,
+      "loss": 0.9136,
+      "step": 1283
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 7.072179188262252e-08,
+      "loss": 0.9301,
+      "step": 1284
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 6.93483032725184e-08,
+      "loss": 0.9454,
+      "step": 1285
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 6.798809544725171e-08,
+      "loss": 0.8926,
+      "step": 1286
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.664117583883272e-08,
+      "loss": 0.9328,
+      "step": 1287
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.530755180666593e-08,
+      "loss": 0.8939,
+      "step": 1288
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.398723063751083e-08,
+      "loss": 0.9142,
+      "step": 1289
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.268021954544095e-08,
+      "loss": 0.9497,
+      "step": 1290
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.138652567180658e-08,
+      "loss": 0.98,
+      "step": 1291
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 6.010615608519261e-08,
+      "loss": 0.942,
+      "step": 1292
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.883911778138324e-08,
+      "loss": 0.909,
+      "step": 1293
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.758541768332071e-08,
+      "loss": 0.945,
+      "step": 1294
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.634506264107054e-08,
+      "loss": 0.9058,
+      "step": 1295
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.5118059431781e-08,
+      "loss": 0.9496,
+      "step": 1296
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.3904414759648195e-08,
+      "loss": 0.9393,
+      "step": 1297
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.2704135255879085e-08,
+      "loss": 0.9425,
+      "step": 1298
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.151722747865434e-08,
+      "loss": 0.8766,
+      "step": 1299
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.0343697913093904e-08,
+      "loss": 0.9401,
+      "step": 1300
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.918355297122035e-08,
+      "loss": 0.9055,
+      "step": 1301
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.8036798991923925e-08,
+      "loss": 0.9285,
+      "step": 1302
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.690344224092924e-08,
+      "loss": 0.943,
+      "step": 1303
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.57834889107589e-08,
+      "loss": 0.9391,
+      "step": 1304
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.467694512070242e-08,
+      "loss": 0.9177,
+      "step": 1305
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.358381691677932e-08,
+      "loss": 0.9313,
+      "step": 1306
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.25041102717097e-08,
+      "loss": 0.9467,
+      "step": 1307
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.1437831084878974e-08,
+      "loss": 0.9005,
+      "step": 1308
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.038498518230627e-08,
+      "loss": 0.9389,
+      "step": 1309
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.9345578316614396e-08,
+      "loss": 0.9429,
+      "step": 1310
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.831961616699464e-08,
+      "loss": 0.9043,
+      "step": 1311
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.730710433918039e-08,
+      "loss": 0.948,
+      "step": 1312
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.630804836541213e-08,
+      "loss": 0.929,
+      "step": 1313
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.5322453704410286e-08,
+      "loss": 0.9213,
+      "step": 1314
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.4350325741344114e-08,
+      "loss": 0.9501,
+      "step": 1315
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.339166978780256e-08,
+      "loss": 0.9553,
+      "step": 1316
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.2446491081764566e-08,
+      "loss": 0.9134,
+      "step": 1317
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.151479478757186e-08,
+      "loss": 0.9884,
+      "step": 1318
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.059658599589926e-08,
+      "loss": 0.8798,
+      "step": 1319
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.969186972372806e-08,
+      "loss": 0.9179,
+      "step": 1320
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.8800650914317385e-08,
+      "loss": 0.9695,
+      "step": 1321
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7922934437178695e-08,
+      "loss": 0.8738,
+      "step": 1322
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7058725088047466e-08,
+      "loss": 0.8622,
+      "step": 1323
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.6208027588858765e-08,
+      "loss": 0.9552,
+      "step": 1324
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.5370846587720044e-08,
+      "loss": 0.9816,
+      "step": 1325
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.454718665888589e-08,
+      "loss": 0.921,
+      "step": 1326
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.3737052302734432e-08,
+      "loss": 0.8977,
+      "step": 1327
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.294044794574096e-08,
+      "loss": 0.9581,
+      "step": 1328
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.215737794045436e-08,
+      "loss": 0.947,
+      "step": 1329
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.1387846565474047e-08,
+      "loss": 0.949,
+      "step": 1330
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.0631858025425554e-08,
+      "loss": 0.949,
+      "step": 1331
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.9889416450938335e-08,
+      "loss": 0.9196,
+      "step": 1332
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.9160525898622716e-08,
+      "loss": 0.9274,
+      "step": 1333
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.844519035104797e-08,
+      "loss": 0.9248,
+      "step": 1334
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.7743413716720394e-08,
+      "loss": 0.9545,
+      "step": 1335
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.7055199830061653e-08,
+      "loss": 0.9476,
+      "step": 1336
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.6380552451389088e-08,
+      "loss": 0.8985,
+      "step": 1337
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.571947526689349e-08,
+      "loss": 0.9035,
+      "step": 1338
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.507197188862053e-08,
+      "loss": 0.9544,
+      "step": 1339
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.4438045854449357e-08,
+      "loss": 0.9738,
+      "step": 1340
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.3817700628074582e-08,
+      "loss": 0.8631,
+      "step": 1341
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.3210939598987394e-08,
+      "loss": 0.9473,
+      "step": 1342
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.261776608245585e-08,
+      "loss": 0.9364,
+      "step": 1343
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.2038183319507957e-08,
+      "loss": 0.887,
+      "step": 1344
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.1472194476913057e-08,
+      "loss": 0.9236,
+      "step": 1345
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.0919802647165467e-08,
+      "loss": 0.9311,
+      "step": 1346
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.0381010848466988e-08,
+      "loss": 0.9325,
+      "step": 1347
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.85582202471025e-09,
+      "loss": 0.9435,
+      "step": 1348
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.344239045462888e-09,
+      "loss": 0.9315,
+      "step": 1349
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.84626470595229e-09,
+      "loss": 0.9324,
+      "step": 1350
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.361901727049204e-09,
+      "loss": 0.9333,
+      "step": 1351
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.891152755254427e-09,
+      "loss": 0.9213,
+      "step": 1352
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.434020362682703e-09,
+      "loss": 0.9535,
+      "step": 1353
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 6.990507047049677e-09,
+      "loss": 0.8807,
+      "step": 1354
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 6.560615231658296e-09,
+      "loss": 0.8865,
+      "step": 1355
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 6.144347265384931e-09,
+      "loss": 0.9316,
+      "step": 1356
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 5.741705422667443e-09,
+      "loss": 0.9185,
+      "step": 1357
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 5.352691903491303e-09,
+      "loss": 0.9198,
+      "step": 1358
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.977308833379324e-09,
+      "loss": 0.8986,
+      "step": 1359
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.615558263378617e-09,
+      "loss": 0.9193,
+      "step": 1360
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 4.267442170050318e-09,
+      "loss": 0.932,
+      "step": 1361
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.932962455458489e-09,
+      "loss": 0.8555,
+      "step": 1362
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.6121209471595697e-09,
+      "loss": 0.9731,
+      "step": 1363
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.304919398192663e-09,
+      "loss": 0.9081,
+      "step": 1364
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.011359487068988e-09,
+      "loss": 0.9013,
+      "step": 1365
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.7314428177646622e-09,
+      "loss": 0.8808,
+      "step": 1366
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.465170919710158e-09,
+      "loss": 0.9075,
+      "step": 1367
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.2125452477828047e-09,
+      "loss": 0.886,
+      "step": 1368
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.973567182298186e-09,
+      "loss": 0.9549,
+      "step": 1369
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.7482380290034795e-09,
+      "loss": 0.9087,
+      "step": 1370
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.5365590190699608e-09,
+      "loss": 0.9479,
+      "step": 1371
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.3385313090857888e-09,
+      "loss": 0.9602,
+      "step": 1372
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.1541559810504532e-09,
+      "loss": 0.9252,
+      "step": 1373
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 9.834340423678368e-10,
+      "loss": 0.8837,
+      "step": 1374
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 8.263664258420512e-10,
+      "loss": 0.9094,
+      "step": 1375
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 6.829539896716087e-10,
+      "loss": 0.9138,
+      "step": 1376
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 5.531975174444259e-10,
+      "loss": 0.9108,
+      "step": 1377
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.3709771813393864e-10,
+      "loss": 0.9414,
+      "step": 1378
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.346552260954927e-10,
+      "loss": 0.9395,
+      "step": 1379
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.45870601062459e-10,
+      "loss": 0.9607,
+      "step": 1380
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.707443281429022e-10,
+      "loss": 0.9411,
+      "step": 1381
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.0927681781819354e-10,
+      "loss": 0.893,
+      "step": 1382
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 6.14684059385695e-11,
+      "loss": 0.8813,
+      "step": 1383
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.7319353723964657e-11,
+      "loss": 0.8738,
+      "step": 1384
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 6.829847760403408e-12,
+      "loss": 0.9713,
+      "step": 1385
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.0,
+      "loss": 0.9176,
+      "step": 1386
+    },
+    {
+      "epoch": 3.0,
+      "step": 1386,
+      "total_flos": 2.997617413765202e+17,
+      "train_loss": 0.9902922145990066,
+      "train_runtime": 3136.962,
+      "train_samples_per_second": 56.445,
+      "train_steps_per_second": 0.442
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 1386,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "total_flos": 2.997617413765202e+17,
+  "trial_name": null,
+  "trial_params": null
+}