diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,11608 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 1930,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0,
+      "loss": 2.0553,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0,
+      "loss": 2.0436,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.4482758620689656e-07,
+      "loss": 2.0303,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.896551724137931e-07,
+      "loss": 2.0072,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.0344827586206898e-06,
+      "loss": 2.0294,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3793103448275862e-06,
+      "loss": 1.9984,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.724137931034483e-06,
+      "loss": 1.8882,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.0689655172413796e-06,
+      "loss": 1.8551,
+      "step": 8
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.4137931034482762e-06,
+      "loss": 1.6339,
+      "step": 9
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.7586206896551725e-06,
+      "loss": 1.5546,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3.103448275862069e-06,
+      "loss": 1.3977,
+      "step": 11
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3.448275862068966e-06,
+      "loss": 1.3685,
+      "step": 12
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3.793103448275862e-06,
+      "loss": 1.3312,
+      "step": 13
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.137931034482759e-06,
+      "loss": 1.2566,
+      "step": 14
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.482758620689656e-06,
+      "loss": 1.2518,
+      "step": 15
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.8275862068965525e-06,
+      "loss": 1.1782,
+      "step": 16
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.172413793103449e-06,
+      "loss": 1.1722,
+      "step": 17
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.517241379310345e-06,
+      "loss": 1.1436,
+      "step": 18
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.862068965517242e-06,
+      "loss": 1.104,
+      "step": 19
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 6.206896551724138e-06,
+      "loss": 1.1195,
+      "step": 20
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 6.551724137931035e-06,
+      "loss": 1.1206,
+      "step": 21
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 6.896551724137932e-06,
+      "loss": 1.1008,
+      "step": 22
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7.241379310344828e-06,
+      "loss": 1.1017,
+      "step": 23
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7.586206896551724e-06,
+      "loss": 1.0606,
+      "step": 24
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7.93103448275862e-06,
+      "loss": 1.0637,
+      "step": 25
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8.275862068965518e-06,
+      "loss": 1.0486,
+      "step": 26
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8.620689655172414e-06,
+      "loss": 1.0206,
+      "step": 27
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8.965517241379312e-06,
+      "loss": 1.0629,
+      "step": 28
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.310344827586207e-06,
+      "loss": 1.0151,
+      "step": 29
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.655172413793105e-06,
+      "loss": 1.0542,
+      "step": 30
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1e-05,
+      "loss": 1.0278,
+      "step": 31
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.0344827586206898e-05,
+      "loss": 1.0034,
+      "step": 32
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.0689655172413792e-05,
+      "loss": 1.022,
+      "step": 33
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.103448275862069e-05,
+      "loss": 0.9959,
+      "step": 34
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.1379310344827587e-05,
+      "loss": 0.9914,
+      "step": 35
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.1724137931034483e-05,
+      "loss": 0.9968,
+      "step": 36
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.206896551724138e-05,
+      "loss": 0.9993,
+      "step": 37
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.2413793103448277e-05,
+      "loss": 0.9492,
+      "step": 38
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.2758620689655174e-05,
+      "loss": 0.9835,
+      "step": 39
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.310344827586207e-05,
+      "loss": 0.9752,
+      "step": 40
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.3448275862068967e-05,
+      "loss": 0.9728,
+      "step": 41
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.3793103448275863e-05,
+      "loss": 0.9501,
+      "step": 42
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.4137931034482759e-05,
+      "loss": 0.9659,
+      "step": 43
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.4482758620689657e-05,
+      "loss": 0.9434,
+      "step": 44
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.4827586206896554e-05,
+      "loss": 0.9694,
+      "step": 45
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.5172413793103448e-05,
+      "loss": 0.972,
+      "step": 46
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.5517241379310346e-05,
+      "loss": 0.9327,
+      "step": 47
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.5517241379310346e-05,
+      "loss": 0.9561,
+      "step": 48
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.586206896551724e-05,
+      "loss": 0.9435,
+      "step": 49
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.6206896551724137e-05,
+      "loss": 0.9296,
+      "step": 50
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.6551724137931037e-05,
+      "loss": 0.9451,
+      "step": 51
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.6896551724137932e-05,
+      "loss": 0.9641,
+      "step": 52
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7241379310344828e-05,
+      "loss": 0.9669,
+      "step": 53
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7586206896551724e-05,
+      "loss": 0.9518,
+      "step": 54
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7931034482758623e-05,
+      "loss": 0.9636,
+      "step": 55
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.827586206896552e-05,
+      "loss": 0.9497,
+      "step": 56
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.8620689655172415e-05,
+      "loss": 0.953,
+      "step": 57
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.896551724137931e-05,
+      "loss": 0.8998,
+      "step": 58
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.931034482758621e-05,
+      "loss": 0.9348,
+      "step": 59
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9655172413793106e-05,
+      "loss": 0.9483,
+      "step": 60
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2e-05,
+      "loss": 0.9244,
+      "step": 61
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9999985918206904e-05,
+      "loss": 0.9411,
+      "step": 62
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9999943672867272e-05,
+      "loss": 0.9348,
+      "step": 63
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9999873264100076e-05,
+      "loss": 0.9843,
+      "step": 64
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.999977469210362e-05,
+      "loss": 0.9584,
+      "step": 65
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.999964795715552e-05,
+      "loss": 0.9105,
+      "step": 66
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.99994930596127e-05,
+      "loss": 0.9399,
+      "step": 67
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9999309999911408e-05,
+      "loss": 0.9276,
+      "step": 68
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.999909877856721e-05,
+      "loss": 0.9247,
+      "step": 69
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9998859396174982e-05,
+      "loss": 0.9513,
+      "step": 70
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.99985918534089e-05,
+      "loss": 0.9201,
+      "step": 71
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9998296151022475e-05,
+      "loss": 0.9269,
+      "step": 72
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9997972289848505e-05,
+      "loss": 0.9476,
+      "step": 73
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.999762027079909e-05,
+      "loss": 0.9219,
+      "step": 74
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.999724009486566e-05,
+      "loss": 0.933,
+      "step": 75
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.999683176311891e-05,
+      "loss": 0.9104,
+      "step": 76
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9996395276708856e-05,
+      "loss": 0.9006,
+      "step": 77
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9995930636864803e-05,
+      "loss": 0.9301,
+      "step": 78
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9995437844895337e-05,
+      "loss": 0.9305,
+      "step": 79
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.999491690218834e-05,
+      "loss": 0.9422,
+      "step": 80
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9994367810210974e-05,
+      "loss": 0.9219,
+      "step": 81
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9993790570509678e-05,
+      "loss": 0.9148,
+      "step": 82
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9993185184710167e-05,
+      "loss": 0.92,
+      "step": 83
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9992551654517424e-05,
+      "loss": 0.9294,
+      "step": 84
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9991889981715696e-05,
+      "loss": 0.9173,
+      "step": 85
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9991200168168494e-05,
+      "loss": 0.9208,
+      "step": 86
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.999048221581858e-05,
+      "loss": 0.8996,
+      "step": 87
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.998973612668796e-05,
+      "loss": 0.934,
+      "step": 88
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.998973612668796e-05,
+      "loss": 0.9286,
+      "step": 89
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9988961902877895e-05,
+      "loss": 0.9379,
+      "step": 90
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9988159546568874e-05,
+      "loss": 0.9277,
+      "step": 91
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9987329060020616e-05,
+      "loss": 0.907,
+      "step": 92
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.998647044557208e-05,
+      "loss": 0.9131,
+      "step": 93
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9985583705641418e-05,
+      "loss": 0.944,
+      "step": 94
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.998466884272602e-05,
+      "loss": 0.915,
+      "step": 95
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9983725859402462e-05,
+      "loss": 0.9274,
+      "step": 96
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9982754758326526e-05,
+      "loss": 0.947,
+      "step": 97
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9981755542233175e-05,
+      "loss": 0.9406,
+      "step": 98
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9980728213936568e-05,
+      "loss": 0.9252,
+      "step": 99
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9979672776330024e-05,
+      "loss": 0.9316,
+      "step": 100
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9978589232386036e-05,
+      "loss": 0.9276,
+      "step": 101
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9977477585156252e-05,
+      "loss": 0.9409,
+      "step": 102
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.997633783777147e-05,
+      "loss": 0.9077,
+      "step": 103
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9975169993441627e-05,
+      "loss": 0.9188,
+      "step": 104
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.997397405545579e-05,
+      "loss": 0.934,
+      "step": 105
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9972750027182154e-05,
+      "loss": 0.919,
+      "step": 106
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9971497912068014e-05,
+      "loss": 0.9022,
+      "step": 107
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.997021771363978e-05,
+      "loss": 0.9291,
+      "step": 108
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.996890943550295e-05,
+      "loss": 0.9149,
+      "step": 109
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9967573081342103e-05,
+      "loss": 0.8811,
+      "step": 110
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9966208654920888e-05,
+      "loss": 0.92,
+      "step": 111
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.996481616008203e-05,
+      "loss": 0.9312,
+      "step": 112
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9963395600747285e-05,
+      "loss": 0.8981,
+      "step": 113
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9961946980917457e-05,
+      "loss": 0.9078,
+      "step": 114
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9960470304672382e-05,
+      "loss": 0.9117,
+      "step": 115
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.995896557617091e-05,
+      "loss": 0.8712,
+      "step": 116
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9957432799650897e-05,
+      "loss": 0.8915,
+      "step": 117
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9955871979429188e-05,
+      "loss": 0.8866,
+      "step": 118
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9954283119901613e-05,
+      "loss": 0.9162,
+      "step": 119
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9952666225542976e-05,
+      "loss": 0.8821,
+      "step": 120
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9951021300907025e-05,
+      "loss": 0.9329,
+      "step": 121
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9949348350626456e-05,
+      "loss": 0.8896,
+      "step": 122
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9947647379412905e-05,
+      "loss": 0.9119,
+      "step": 123
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9945918392056912e-05,
+      "loss": 0.9108,
+      "step": 124
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9944161393427923e-05,
+      "loss": 0.8932,
+      "step": 125
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9942376388474282e-05,
+      "loss": 0.9083,
+      "step": 126
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9940563382223196e-05,
+      "loss": 0.8875,
+      "step": 127
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9938722379780747e-05,
+      "loss": 0.9179,
+      "step": 128
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.993685338633186e-05,
+      "loss": 0.9,
+      "step": 129
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9934956407140285e-05,
+      "loss": 0.9103,
+      "step": 130
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9933031447548593e-05,
+      "loss": 0.8908,
+      "step": 131
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9931078512978172e-05,
+      "loss": 0.8838,
+      "step": 132
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9929097608929174e-05,
+      "loss": 0.9242,
+      "step": 133
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.992708874098054e-05,
+      "loss": 0.9018,
+      "step": 134
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.9925051914789968e-05,
+      "loss": 0.9146,
+      "step": 135
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.992298713609388e-05,
+      "loss": 0.8907,
+      "step": 136
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.9920894410707446e-05,
+      "loss": 0.9053,
+      "step": 137
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.991877374452452e-05,
+      "loss": 0.9045,
+      "step": 138
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.991662514351766e-05,
+      "loss": 0.8906,
+      "step": 139
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9914448613738107e-05,
+      "loss": 0.9165,
+      "step": 140
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9912244161315737e-05,
+      "loss": 0.8885,
+      "step": 141
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9910011792459086e-05,
+      "loss": 0.8807,
+      "step": 142
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.99077515134553e-05,
+      "loss": 0.9059,
+      "step": 143
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9905463330670143e-05,
+      "loss": 0.8864,
+      "step": 144
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9903147250547953e-05,
+      "loss": 0.8909,
+      "step": 145
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9900803279611643e-05,
+      "loss": 0.892,
+      "step": 146
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9898431424462674e-05,
+      "loss": 0.8774,
+      "step": 147
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9896031691781043e-05,
+      "loss": 0.8874,
+      "step": 148
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9893604088325258e-05,
+      "loss": 0.8798,
+      "step": 149
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.989114862093232e-05,
+      "loss": 0.9215,
+      "step": 150
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9888665296517702e-05,
+      "loss": 0.8854,
+      "step": 151
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9886154122075344e-05,
+      "loss": 0.8736,
+      "step": 152
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.988361510467761e-05,
+      "loss": 0.9276,
+      "step": 153
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.9881048251475283e-05,
+      "loss": 0.8588,
+      "step": 154
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.987845356969754e-05,
+      "loss": 0.8925,
+      "step": 155
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.9875831066651937e-05,
+      "loss": 0.8953,
+      "step": 156
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.987318074972439e-05,
+      "loss": 0.8999,
+      "step": 157
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9870502626379127e-05,
+      "loss": 0.8927,
+      "step": 158
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9867796704158722e-05,
+      "loss": 0.902,
+      "step": 159
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.986506299068401e-05,
+      "loss": 0.8989,
+      "step": 160
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.986230149365411e-05,
+      "loss": 0.8677,
+      "step": 161
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.985951222084639e-05,
+      "loss": 0.899,
+      "step": 162
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.985669518011644e-05,
+      "loss": 0.8915,
+      "step": 163
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.985385037939806e-05,
+      "loss": 0.92,
+      "step": 164
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9850977826703225e-05,
+      "loss": 0.8846,
+      "step": 165
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9848077530122083e-05,
+      "loss": 0.894,
+      "step": 166
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9845149497822898e-05,
+      "loss": 0.8885,
+      "step": 167
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9842193738052064e-05,
+      "loss": 0.8712,
+      "step": 168
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9839210259134062e-05,
+      "loss": 0.8876,
+      "step": 169
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.983619906947144e-05,
+      "loss": 0.8679,
+      "step": 170
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.983316017754478e-05,
+      "loss": 0.913,
+      "step": 171
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9830093591912694e-05,
+      "loss": 0.8847,
+      "step": 172
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9826999321211793e-05,
+      "loss": 0.8768,
+      "step": 173
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9823877374156647e-05,
+      "loss": 0.886,
+      "step": 174
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9820727759539786e-05,
+      "loss": 0.8922,
+      "step": 175
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.9817550486231645e-05,
+      "loss": 0.877,
+      "step": 176
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.981434556318057e-05,
+      "loss": 0.8864,
+      "step": 177
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.981111299941277e-05,
+      "loss": 0.9434,
+      "step": 178
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.9807852804032306e-05,
+      "loss": 0.8846,
+      "step": 179
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.980456498622106e-05,
+      "loss": 0.9049,
+      "step": 180
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.9801249555238703e-05,
+      "loss": 0.8849,
+      "step": 181
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.979790652042268e-05,
+      "loss": 0.883,
+      "step": 182
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.9794535891188172e-05,
+      "loss": 0.8646,
+      "step": 183
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.9791137677028084e-05,
+      "loss": 0.886,
+      "step": 184
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.9787711887513e-05,
+      "loss": 0.8763,
+      "step": 185
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.978425853229118e-05,
+      "loss": 0.8674,
+      "step": 186
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.9780777621088504e-05,
+      "loss": 0.88,
+      "step": 187
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.977726916370847e-05,
+      "loss": 0.8939,
+      "step": 188
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.977373317003215e-05,
+      "loss": 0.9059,
+      "step": 189
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.977016965001817e-05,
+      "loss": 0.8754,
+      "step": 190
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.9766578613702683e-05,
+      "loss": 0.8939,
+      "step": 191
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.9762960071199334e-05,
+      "loss": 0.8969,
+      "step": 192
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.9759314032699237e-05,
+      "loss": 0.8917,
+      "step": 193
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.975564050847094e-05,
+      "loss": 0.8937,
+      "step": 194
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9751939508860406e-05,
+      "loss": 0.8886,
+      "step": 195
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9748211044290982e-05,
+      "loss": 0.8927,
+      "step": 196
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.974445512526336e-05,
+      "loss": 0.8777,
+      "step": 197
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9740671762355548e-05,
+      "loss": 0.882,
+      "step": 198
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.973686096622286e-05,
+      "loss": 0.8826,
+      "step": 199
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.973302274759786e-05,
+      "loss": 0.8808,
+      "step": 200
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.9729157117290353e-05,
+      "loss": 0.8751,
+      "step": 201
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.9725264086187334e-05,
+      "loss": 0.8594,
+      "step": 202
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.972134366525298e-05,
+      "loss": 0.876,
+      "step": 203
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.97173958655286e-05,
+      "loss": 0.8657,
+      "step": 204
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.9713420698132614e-05,
+      "loss": 0.8762,
+      "step": 205
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.9709418174260523e-05,
+      "loss": 0.9096,
+      "step": 206
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.970538830518486e-05,
+      "loss": 0.9033,
+      "step": 207
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.970133110225519e-05,
+      "loss": 0.8995,
+      "step": 208
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.969724657689805e-05,
+      "loss": 0.8884,
+      "step": 209
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.9693134740616924e-05,
+      "loss": 0.8732,
+      "step": 210
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.9688995604992226e-05,
+      "loss": 0.8898,
+      "step": 211
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.9684829181681236e-05,
+      "loss": 0.8729,
+      "step": 212
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.9680635482418103e-05,
+      "loss": 0.8793,
+      "step": 213
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.9676414519013782e-05,
+      "loss": 0.889,
+      "step": 214
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.967216630335603e-05,
+      "loss": 0.8688,
+      "step": 215
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.9667890847409336e-05,
+      "loss": 0.8745,
+      "step": 216
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.966358816321492e-05,
+      "loss": 0.8801,
+      "step": 217
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 0.8765,
+      "step": 218
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.965490115863118e-05,
+      "loss": 0.8839,
+      "step": 219
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.965051686270758e-05,
+      "loss": 0.8983,
+      "step": 220
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.9646105387467627e-05,
+      "loss": 0.8891,
+      "step": 221
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.9641666745335626e-05,
+      "loss": 0.8915,
+      "step": 222
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.9637200948812378e-05,
+      "loss": 0.8745,
+      "step": 223
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.9632708010475166e-05,
+      "loss": 0.8455,
+      "step": 224
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.962818794297772e-05,
+      "loss": 0.8721,
+      "step": 225
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.962364075905017e-05,
+      "loss": 0.8926,
+      "step": 226
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9619066471499017e-05,
+      "loss": 0.8885,
+      "step": 227
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9614465093207093e-05,
+      "loss": 0.8455,
+      "step": 228
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.960983663713353e-05,
+      "loss": 0.887,
+      "step": 229
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.9605181116313725e-05,
+      "loss": 0.8527,
+      "step": 230
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.960049854385929e-05,
+      "loss": 0.8961,
+      "step": 231
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.9595788932958026e-05,
+      "loss": 0.8965,
+      "step": 232
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.959105229687389e-05,
+      "loss": 0.8987,
+      "step": 233
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9586288648946947e-05,
+      "loss": 0.8958,
+      "step": 234
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9581498002593337e-05,
+      "loss": 0.8967,
+      "step": 235
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.957668037130524e-05,
+      "loss": 0.9013,
+      "step": 236
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9571835768650833e-05,
+      "loss": 0.9098,
+      "step": 237
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9566964208274254e-05,
+      "loss": 0.8728,
+      "step": 238
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.956206570389556e-05,
+      "loss": 0.8951,
+      "step": 239
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9557140269310702e-05,
+      "loss": 0.8865,
+      "step": 240
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9552187918391472e-05,
+      "loss": 0.8833,
+      "step": 241
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.954720866508546e-05,
+      "loss": 0.8746,
+      "step": 242
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9542202523416027e-05,
+      "loss": 0.8688,
+      "step": 243
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.953716950748227e-05,
+      "loss": 0.8625,
+      "step": 244
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9532109631458963e-05,
+      "loss": 0.8755,
+      "step": 245
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9527022909596537e-05,
+      "loss": 0.8789,
+      "step": 246
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9521909356221015e-05,
+      "loss": 0.8594,
+      "step": 247
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9516768985734008e-05,
+      "loss": 0.8786,
+      "step": 248
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9511601812612633e-05,
+      "loss": 0.8745,
+      "step": 249
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.950640785140951e-05,
+      "loss": 0.8721,
+      "step": 250
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9501187116752694e-05,
+      "loss": 0.8825,
+      "step": 251
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.949593962334564e-05,
+      "loss": 0.842,
+      "step": 252
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9490665385967182e-05,
+      "loss": 0.8804,
+      "step": 253
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9485364419471454e-05,
+      "loss": 0.8682,
+      "step": 254
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9480036738787888e-05,
+      "loss": 0.8765,
+      "step": 255
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9474682358921137e-05,
+      "loss": 0.8618,
+      "step": 256
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.946930129495106e-05,
+      "loss": 0.8606,
+      "step": 257
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.9463893562032655e-05,
+      "loss": 0.8755,
+      "step": 258
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.9458459175396042e-05,
+      "loss": 0.8426,
+      "step": 259
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.9452998150346403e-05,
+      "loss": 0.8597,
+      "step": 260
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.9447510502263945e-05,
+      "loss": 0.8776,
+      "step": 261
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.9441996246603848e-05,
+      "loss": 0.8914,
+      "step": 262
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.9436455398896234e-05,
+      "loss": 0.8802,
+      "step": 263
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.9430887974746124e-05,
+      "loss": 0.8843,
+      "step": 264
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.942529398983337e-05,
+      "loss": 0.8849,
+      "step": 265
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.9419673459912652e-05,
+      "loss": 0.8784,
+      "step": 266
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.9414026400813387e-05,
+      "loss": 0.8761,
+      "step": 267
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.940835282843972e-05,
+      "loss": 0.8666,
+      "step": 268
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.9402652758770476e-05,
+      "loss": 0.8553,
+      "step": 269
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.9396926207859085e-05,
+      "loss": 0.8642,
+      "step": 270
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.9391173191833573e-05,
+      "loss": 0.8804,
+      "step": 271
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.9385393726896492e-05,
+      "loss": 0.8802,
+      "step": 272
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9379587829324888e-05,
+      "loss": 0.8822,
+      "step": 273
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9373755515470253e-05,
+      "loss": 0.8743,
+      "step": 274
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9367896801758478e-05,
+      "loss": 0.8748,
+      "step": 275
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9362011704689794e-05,
+      "loss": 0.8944,
+      "step": 276
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9362011704689794e-05,
+      "loss": 0.8997,
+      "step": 277
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.935610024083875e-05,
+      "loss": 0.8804,
+      "step": 278
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9350162426854152e-05,
+      "loss": 0.876,
+      "step": 279
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.9344198279459e-05,
+      "loss": 0.8588,
+      "step": 280
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.9338207815450487e-05,
+      "loss": 0.8618,
+      "step": 281
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.93321910516999e-05,
+      "loss": 0.8716,
+      "step": 282
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.9326148005152606e-05,
+      "loss": 0.8745,
+      "step": 283
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.932007869282799e-05,
+      "loss": 0.8883,
+      "step": 284
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.9313983131819406e-05,
+      "loss": 0.8724,
+      "step": 285
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.9307861339294153e-05,
+      "loss": 0.8666,
+      "step": 286
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.9301713332493386e-05,
+      "loss": 0.8604,
+      "step": 287
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.9295539128732096e-05,
+      "loss": 0.8541,
+      "step": 288
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.928933874539906e-05,
+      "loss": 0.881,
+      "step": 289
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.9283112199956775e-05,
+      "loss": 0.8851,
+      "step": 290
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.927685950994143e-05,
+      "loss": 0.8564,
+      "step": 291
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.927058069296284e-05,
+      "loss": 0.9008,
+      "step": 292
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.926427576670441e-05,
+      "loss": 0.8835,
+      "step": 293
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.9257944748923074e-05,
+      "loss": 0.9124,
+      "step": 294
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.925158765744924e-05,
+      "loss": 0.8536,
+      "step": 295
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.924520451018676e-05,
+      "loss": 0.8748,
+      "step": 296
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.9238795325112867e-05,
+      "loss": 0.8803,
+      "step": 297
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.923236012027813e-05,
+      "loss": 0.863,
+      "step": 298
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.922589891380639e-05,
+      "loss": 0.8643,
+      "step": 299
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.9219411723894717e-05,
+      "loss": 0.8839,
+      "step": 300
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.921289856881337e-05,
+      "loss": 0.8795,
+      "step": 301
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.9206359466905727e-05,
+      "loss": 0.8703,
+      "step": 302
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.9199794436588244e-05,
+      "loss": 0.8782,
+      "step": 303
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.91932034963504e-05,
+      "loss": 0.8773,
+      "step": 304
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.918658666475465e-05,
+      "loss": 0.8596,
+      "step": 305
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.917994396043636e-05,
+      "loss": 0.8708,
+      "step": 306
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9173275402103772e-05,
+      "loss": 0.8847,
+      "step": 307
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9166581008537935e-05,
+      "loss": 0.8763,
+      "step": 308
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.915986079859266e-05,
+      "loss": 0.8736,
+      "step": 309
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9153114791194475e-05,
+      "loss": 0.8596,
+      "step": 310
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9146343005342546e-05,
+      "loss": 0.8635,
+      "step": 311
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9139545460108658e-05,
+      "loss": 0.8665,
+      "step": 312
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.913272217463714e-05,
+      "loss": 0.897,
+      "step": 313
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.91258731681448e-05,
+      "loss": 0.871,
+      "step": 314
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.91189984599209e-05,
+      "loss": 0.8537,
+      "step": 315
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.9112098069327093e-05,
+      "loss": 0.877,
+      "step": 316
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.910517201579734e-05,
+      "loss": 0.8816,
+      "step": 317
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.9098220318837897e-05,
+      "loss": 0.8746,
+      "step": 318
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.909124299802724e-05,
+      "loss": 0.8585,
+      "step": 319
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.9084240073015997e-05,
+      "loss": 0.8547,
+      "step": 320
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.9077211563526928e-05,
+      "loss": 0.8852,
+      "step": 321
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.9070157489354825e-05,
+      "loss": 0.8696,
+      "step": 322
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9063077870366504e-05,
+      "loss": 0.8709,
+      "step": 323
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9055972726500696e-05,
+      "loss": 0.859,
+      "step": 324
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.904884207776804e-05,
+      "loss": 0.8859,
+      "step": 325
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9041685944251005e-05,
+      "loss": 0.8775,
+      "step": 326
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9034504346103825e-05,
+      "loss": 0.8753,
+      "step": 327
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9027297303552456e-05,
+      "loss": 0.8546,
+      "step": 328
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9020064836894514e-05,
+      "loss": 0.857,
+      "step": 329
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9012806966499218e-05,
+      "loss": 0.8743,
+      "step": 330
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9005523712807335e-05,
+      "loss": 0.842,
+      "step": 331
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.8998215096331124e-05,
+      "loss": 0.8875,
+      "step": 332
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.899088113765426e-05,
+      "loss": 0.8508,
+      "step": 333
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.898352185743181e-05,
+      "loss": 0.8339,
+      "step": 334
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8976137276390145e-05,
+      "loss": 0.8566,
+      "step": 335
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8968727415326885e-05,
+      "loss": 0.8727,
+      "step": 336
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8961292295110865e-05,
+      "loss": 0.8806,
+      "step": 337
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8953831936682047e-05,
+      "loss": 0.8947,
+      "step": 338
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8946346361051474e-05,
+      "loss": 0.8868,
+      "step": 339
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8938835589301216e-05,
+      "loss": 0.8925,
+      "step": 340
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8931299642584297e-05,
+      "loss": 0.8691,
+      "step": 341
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8923738542124644e-05,
+      "loss": 0.864,
+      "step": 342
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.891615230921703e-05,
+      "loss": 0.8549,
+      "step": 343
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8908540965227006e-05,
+      "loss": 0.8612,
+      "step": 344
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8900904531590845e-05,
+      "loss": 0.863,
+      "step": 345
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8893243029815487e-05,
+      "loss": 0.8651,
+      "step": 346
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8885556481478468e-05,
+      "loss": 0.8671,
+      "step": 347
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8877844908227857e-05,
+      "loss": 0.8714,
+      "step": 348
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.887010833178222e-05,
+      "loss": 0.843,
+      "step": 349
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.8862346773930524e-05,
+      "loss": 0.9056,
+      "step": 350
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.8854560256532098e-05,
+      "loss": 0.8651,
+      "step": 351
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.8846748801516576e-05,
+      "loss": 0.8575,
+      "step": 352
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.883891243088381e-05,
+      "loss": 0.8608,
+      "step": 353
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.8831051166703832e-05,
+      "loss": 0.8511,
+      "step": 354
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.8823165031116782e-05,
+      "loss": 0.8515,
+      "step": 355
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.881525404633284e-05,
+      "loss": 0.8765,
+      "step": 356
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.8807318234632187e-05,
+      "loss": 0.855,
+      "step": 357
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.8799357618364908e-05,
+      "loss": 0.8834,
+      "step": 358
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.879137221995095e-05,
+      "loss": 0.8745,
+      "step": 359
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.8783362061880063e-05,
+      "loss": 0.8506,
+      "step": 360
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8775327166711724e-05,
+      "loss": 0.8714,
+      "step": 361
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.876726755707508e-05,
+      "loss": 0.8419,
+      "step": 362
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.875918325566888e-05,
+      "loss": 0.8546,
+      "step": 363
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8751074285261418e-05,
+      "loss": 0.859,
+      "step": 364
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.8742940668690464e-05,
+      "loss": 0.8728,
+      "step": 365
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.8734782428863196e-05,
+      "loss": 0.8405,
+      "step": 366
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.8726599588756144e-05,
+      "loss": 0.8669,
+      "step": 367
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.871839217141512e-05,
+      "loss": 0.8515,
+      "step": 368
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.8710160199955158e-05,
+      "loss": 0.8656,
+      "step": 369
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.8701903697560437e-05,
+      "loss": 0.8525,
+      "step": 370
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.869362268748423e-05,
+      "loss": 0.8474,
+      "step": 371
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.8685317193048833e-05,
+      "loss": 0.8627,
+      "step": 372
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.8676987237645494e-05,
+      "loss": 0.8677,
+      "step": 373
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.8668632844734354e-05,
+      "loss": 0.8569,
+      "step": 374
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 0.8488,
+      "step": 375
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.8651850840573315e-05,
+      "loss": 0.8694,
+      "step": 376
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.8643423276587548e-05,
+      "loss": 0.8685,
+      "step": 377
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.863497136962213e-05,
+      "loss": 0.8387,
+      "step": 378
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.862649514348067e-05,
+      "loss": 0.846,
+      "step": 379
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.8617994622035254e-05,
+      "loss": 0.8677,
+      "step": 380
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.86094698292264e-05,
+      "loss": 0.8976,
+      "step": 381
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.8600920789062986e-05,
+      "loss": 0.8618,
+      "step": 382
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.859234752562217e-05,
+      "loss": 0.8509,
+      "step": 383
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.8583750063049338e-05,
+      "loss": 0.8723,
+      "step": 384
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.8575128425558024e-05,
+      "loss": 0.8699,
+      "step": 385
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.856648263742986e-05,
+      "loss": 0.8851,
+      "step": 386
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.8557812723014476e-05,
+      "loss": 0.8318,
+      "step": 387
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.854911870672947e-05,
+      "loss": 0.8209,
+      "step": 388
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.85404006130603e-05,
+      "loss": 0.8235,
+      "step": 389
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.8531658466560248e-05,
+      "loss": 0.8012,
+      "step": 390
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.8522892291850335e-05,
+      "loss": 0.8123,
+      "step": 391
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.8514102113619253e-05,
+      "loss": 0.8032,
+      "step": 392
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.8505287956623298e-05,
+      "loss": 0.8029,
+      "step": 393
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.8496449845686292e-05,
+      "loss": 0.7956,
+      "step": 394
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.848758780569953e-05,
+      "loss": 0.8147,
+      "step": 395
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.8478701861621686e-05,
+      "loss": 0.825,
+      "step": 396
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.8469792038478774e-05,
+      "loss": 0.8312,
+      "step": 397
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.846085836136405e-05,
+      "loss": 0.8092,
+      "step": 398
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.845190085543795e-05,
+      "loss": 0.8087,
+      "step": 399
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.8442919545928023e-05,
+      "loss": 0.8275,
+      "step": 400
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.843391445812886e-05,
+      "loss": 0.8235,
+      "step": 401
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.8424885617402016e-05,
+      "loss": 0.7974,
+      "step": 402
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.8415833049175942e-05,
+      "loss": 0.8043,
+      "step": 403
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.840675677894592e-05,
+      "loss": 0.8124,
+      "step": 404
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.8397656832273982e-05,
+      "loss": 0.8035,
+      "step": 405
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.8388533234788838e-05,
+      "loss": 0.8011,
+      "step": 406
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.8379386012185813e-05,
+      "loss": 0.8064,
+      "step": 407
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.8370215190226768e-05,
+      "loss": 0.7938,
+      "step": 408
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.8361020794740024e-05,
+      "loss": 0.8236,
+      "step": 409
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.8351802851620296e-05,
+      "loss": 0.8387,
+      "step": 410
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.8342561386828613e-05,
+      "loss": 0.8263,
+      "step": 411
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.8333296426392264e-05,
+      "loss": 0.8098,
+      "step": 412
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.8324007996404692e-05,
+      "loss": 0.795,
+      "step": 413
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.8314696123025456e-05,
+      "loss": 0.8033,
+      "step": 414
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.8305360832480118e-05,
+      "loss": 0.7811,
+      "step": 415
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.8296002151060213e-05,
+      "loss": 0.8351,
+      "step": 416
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.828662010512314e-05,
+      "loss": 0.7897,
+      "step": 417
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.827721472109211e-05,
+      "loss": 0.8026,
+      "step": 418
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.8267786025456054e-05,
+      "loss": 0.8083,
+      "step": 419
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.8258334044769558e-05,
+      "loss": 0.8089,
+      "step": 420
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.8248858805652795e-05,
+      "loss": 0.8237,
+      "step": 421
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.823936033479143e-05,
+      "loss": 0.7872,
+      "step": 422
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8229838658936566e-05,
+      "loss": 0.7649,
+      "step": 423
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8220293804904656e-05,
+      "loss": 0.783,
+      "step": 424
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.821072579957744e-05,
+      "loss": 0.8133,
+      "step": 425
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.8201134669901838e-05,
+      "loss": 0.7943,
+      "step": 426
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.819152044288992e-05,
+      "loss": 0.8052,
+      "step": 427
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.8181883145618794e-05,
+      "loss": 0.8219,
+      "step": 428
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.817222280523055e-05,
+      "loss": 0.8032,
+      "step": 429
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.8162539448932162e-05,
+      "loss": 0.8318,
+      "step": 430
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.8152833103995443e-05,
+      "loss": 0.8214,
+      "step": 431
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.8143103797756942e-05,
+      "loss": 0.813,
+      "step": 432
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.8133351557617863e-05,
+      "loss": 0.8008,
+      "step": 433
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.812357641104402e-05,
+      "loss": 0.8174,
+      "step": 434
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.811377838556573e-05,
+      "loss": 0.8159,
+      "step": 435
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.810395750877775e-05,
+      "loss": 0.8136,
+      "step": 436
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.8094113808339184e-05,
+      "loss": 0.8135,
+      "step": 437
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.8084247311973426e-05,
+      "loss": 0.8026,
+      "step": 438
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.807435804746807e-05,
+      "loss": 0.8047,
+      "step": 439
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.806444604267483e-05,
+      "loss": 0.8215,
+      "step": 440
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.805451132550946e-05,
+      "loss": 0.8203,
+      "step": 441
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.80445539239517e-05,
+      "loss": 0.8019,
+      "step": 442
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.8034573866045145e-05,
+      "loss": 0.8087,
+      "step": 443
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.8024571179897236e-05,
+      "loss": 0.8217,
+      "step": 444
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.8014545893679118e-05,
+      "loss": 0.7776,
+      "step": 445
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.8004498035625584e-05,
+      "loss": 0.8448,
+      "step": 446
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.7994427634035016e-05,
+      "loss": 0.8067,
+      "step": 447
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.7984334717269268e-05,
+      "loss": 0.8081,
+      "step": 448
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.7974219313753618e-05,
+      "loss": 0.8023,
+      "step": 449
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.7964081451976673e-05,
+      "loss": 0.7947,
+      "step": 450
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.795392116049028e-05,
+      "loss": 0.8079,
+      "step": 451
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.794373846790947e-05,
+      "loss": 0.8259,
+      "step": 452
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.7933533402912354e-05,
+      "loss": 0.8016,
+      "step": 453
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.7923305994240055e-05,
+      "loss": 0.8137,
+      "step": 454
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.791305627069662e-05,
+      "loss": 0.8068,
+      "step": 455
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.7902784261148952e-05,
+      "loss": 0.8091,
+      "step": 456
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.7892489994526715e-05,
+      "loss": 0.8212,
+      "step": 457
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.7882173499822246e-05,
+      "loss": 0.8102,
+      "step": 458
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.7871834806090502e-05,
+      "loss": 0.798,
+      "step": 459
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.7861473942448953e-05,
+      "loss": 0.8278,
+      "step": 460
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.78510909380775e-05,
+      "loss": 0.8207,
+      "step": 461
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.784068582221841e-05,
+      "loss": 0.8094,
+      "step": 462
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.7830258624176224e-05,
+      "loss": 0.8134,
+      "step": 463
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.7819809373317665e-05,
+      "loss": 0.8055,
+      "step": 464
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.780933809907158e-05,
+      "loss": 0.8071,
+      "step": 465
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.7798844830928818e-05,
+      "loss": 0.8248,
+      "step": 466
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.7788329598442197e-05,
+      "loss": 0.8262,
+      "step": 467
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.7777792431226384e-05,
+      "loss": 0.8256,
+      "step": 468
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.7767233358957813e-05,
+      "loss": 0.7906,
+      "step": 469
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.7756652411374616e-05,
+      "loss": 0.7968,
+      "step": 470
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.7746049618276545e-05,
+      "loss": 0.8322,
+      "step": 471
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.7735425009524867e-05,
+      "loss": 0.8167,
+      "step": 472
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.7724778615042285e-05,
+      "loss": 0.821,
+      "step": 473
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.7714110464812865e-05,
+      "loss": 0.7856,
+      "step": 474
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.770342058888195e-05,
+      "loss": 0.8084,
+      "step": 475
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.7692709017356053e-05,
+      "loss": 0.799,
+      "step": 476
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.7681975780402807e-05,
+      "loss": 0.8066,
+      "step": 477
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.767122090825086e-05,
+      "loss": 0.8333,
+      "step": 478
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.766044443118978e-05,
+      "loss": 0.8276,
+      "step": 479
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.7649646379570003e-05,
+      "loss": 0.8232,
+      "step": 480
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.7638826783802706e-05,
+      "loss": 0.8259,
+      "step": 481
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.762798567435975e-05,
+      "loss": 0.8323,
+      "step": 482
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.761712308177359e-05,
+      "loss": 0.8253,
+      "step": 483
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.7606239036637188e-05,
+      "loss": 0.7976,
+      "step": 484
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.7595333569603908e-05,
+      "loss": 0.8038,
+      "step": 485
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.7584406711387462e-05,
+      "loss": 0.7876,
+      "step": 486
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.7573458492761802e-05,
+      "loss": 0.8145,
+      "step": 487
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.756248894456103e-05,
+      "loss": 0.8185,
+      "step": 488
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.755149809767934e-05,
+      "loss": 0.8489,
+      "step": 489
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.754048598307089e-05,
+      "loss": 0.8151,
+      "step": 490
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.7529452631749743e-05,
+      "loss": 0.8062,
+      "step": 491
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.7518398074789776e-05,
+      "loss": 0.8083,
+      "step": 492
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.7507322343324586e-05,
+      "loss": 0.7873,
+      "step": 493
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.7496225468547407e-05,
+      "loss": 0.821,
+      "step": 494
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.7485107481711014e-05,
+      "loss": 0.8312,
+      "step": 495
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.7473968414127644e-05,
+      "loss": 0.8057,
+      "step": 496
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.746280829716891e-05,
+      "loss": 0.8193,
+      "step": 497
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.7451627162265703e-05,
+      "loss": 0.8375,
+      "step": 498
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.744042504090811e-05,
+      "loss": 0.8209,
+      "step": 499
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.7429201964645314e-05,
+      "loss": 0.8126,
+      "step": 500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.7417957965085532e-05,
+      "loss": 0.7912,
+      "step": 501
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.7406693073895894e-05,
+      "loss": 0.8132,
+      "step": 502
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.7395407322802374e-05,
+      "loss": 0.8276,
+      "step": 503
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.7384100743589698e-05,
+      "loss": 0.8178,
+      "step": 504
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.737277336810124e-05,
+      "loss": 0.8152,
+      "step": 505
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.736142522823896e-05,
+      "loss": 0.8195,
+      "step": 506
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.7350056355963287e-05,
+      "loss": 0.8113,
+      "step": 507
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.733866678329304e-05,
+      "loss": 0.8213,
+      "step": 508
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.7327256542305342e-05,
+      "loss": 0.8265,
+      "step": 509
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.731582566513552e-05,
+      "loss": 0.8363,
+      "step": 510
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.7304374183977032e-05,
+      "loss": 0.8005,
+      "step": 511
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.7292902131081348e-05,
+      "loss": 0.8236,
+      "step": 512
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.7281409538757886e-05,
+      "loss": 0.7989,
+      "step": 513
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.7269896439373904e-05,
+      "loss": 0.8178,
+      "step": 514
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.7258362865354426e-05,
+      "loss": 0.8264,
+      "step": 515
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.7246808849182125e-05,
+      "loss": 0.8018,
+      "step": 516
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.723523442339726e-05,
+      "loss": 0.8046,
+      "step": 517
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.7223639620597556e-05,
+      "loss": 0.8306,
+      "step": 518
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.7212024473438145e-05,
+      "loss": 0.8135,
+      "step": 519
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.7200389014631446e-05,
+      "loss": 0.8225,
+      "step": 520
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.718873327694708e-05,
+      "loss": 0.8248,
+      "step": 521
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.7177057293211786e-05,
+      "loss": 0.7838,
+      "step": 522
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.716536109630932e-05,
+      "loss": 0.8286,
+      "step": 523
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.7153644719180366e-05,
+      "loss": 0.8214,
+      "step": 524
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.7141908194822448e-05,
+      "loss": 0.8296,
+      "step": 525
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.7130151556289823e-05,
+      "loss": 0.8027,
+      "step": 526
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.7118374836693407e-05,
+      "loss": 0.8134,
+      "step": 527
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.710657806920066e-05,
+      "loss": 0.8365,
+      "step": 528
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.7094761287035516e-05,
+      "loss": 0.8129,
+      "step": 529
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.708292452347826e-05,
+      "loss": 0.7981,
+      "step": 530
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 0.8204,
+      "step": 531
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.705919118558991e-05,
+      "loss": 0.7837,
+      "step": 532
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.70472946781004e-05,
+      "loss": 0.8143,
+      "step": 533
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.703537832290178e-05,
+      "loss": 0.8241,
+      "step": 534
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.7023442153554776e-05,
+      "loss": 0.8178,
+      "step": 535
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.7011486203675923e-05,
+      "loss": 0.8191,
+      "step": 536
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.6999510506937465e-05,
+      "loss": 0.7963,
+      "step": 537
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.6987515097067262e-05,
+      "loss": 0.8228,
+      "step": 538
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.6975500007848682e-05,
+      "loss": 0.8265,
+      "step": 539
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.696346527312053e-05,
+      "loss": 0.8206,
+      "step": 540
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.6951410926776934e-05,
+      "loss": 0.8196,
+      "step": 541
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.6939337002767257e-05,
+      "loss": 0.8492,
+      "step": 542
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.6927243535095995e-05,
+      "loss": 0.8179,
+      "step": 543
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.6915130557822698e-05,
+      "loss": 0.8078,
+      "step": 544
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.6902998105061845e-05,
+      "loss": 0.8223,
+      "step": 545
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.6890846210982775e-05,
+      "loss": 0.8325,
+      "step": 546
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.687867490980959e-05,
+      "loss": 0.8006,
+      "step": 547
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.6866484235821025e-05,
+      "loss": 0.8154,
+      "step": 548
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.68542742233504e-05,
+      "loss": 0.8169,
+      "step": 549
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.6842044906785483e-05,
+      "loss": 0.8148,
+      "step": 550
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.6829796320568416e-05,
+      "loss": 0.8226,
+      "step": 551
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.6817528499195614e-05,
+      "loss": 0.8075,
+      "step": 552
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.6805241477217663e-05,
+      "loss": 0.8168,
+      "step": 553
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.6792935289239212e-05,
+      "loss": 0.8215,
+      "step": 554
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.678060996991891e-05,
+      "loss": 0.8015,
+      "step": 555
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.6768265553969272e-05,
+      "loss": 0.7984,
+      "step": 556
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.6755902076156606e-05,
+      "loss": 0.8163,
+      "step": 557
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.674351957130089e-05,
+      "loss": 0.8267,
+      "step": 558
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.67311180742757e-05,
+      "loss": 0.8087,
+      "step": 559
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.6718697620008112e-05,
+      "loss": 0.8116,
+      "step": 560
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.670625824347856e-05,
+      "loss": 0.8076,
+      "step": 561
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.6693799979720804e-05,
+      "loss": 0.8074,
+      "step": 562
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.6681322863821776e-05,
+      "loss": 0.8287,
+      "step": 563
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.666882693092151e-05,
+      "loss": 0.845,
+      "step": 564
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.6656312216213035e-05,
+      "loss": 0.8275,
+      "step": 565
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.6643778754942276e-05,
+      "loss": 0.813,
+      "step": 566
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.6631226582407954e-05,
+      "loss": 0.8271,
+      "step": 567
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.6618655733961487e-05,
+      "loss": 0.8245,
+      "step": 568
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.6606066245006896e-05,
+      "loss": 0.812,
+      "step": 569
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.659345815100069e-05,
+      "loss": 0.8329,
+      "step": 570
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.6580831487451788e-05,
+      "loss": 0.8259,
+      "step": 571
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.6568186289921407e-05,
+      "loss": 0.8126,
+      "step": 572
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.655552259402295e-05,
+      "loss": 0.8316,
+      "step": 573
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.654284043542193e-05,
+      "loss": 0.8107,
+      "step": 574
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.653013984983585e-05,
+      "loss": 0.8083,
+      "step": 575
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.651742087303412e-05,
+      "loss": 0.7931,
+      "step": 576
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.650468354083794e-05,
+      "loss": 0.8292,
+      "step": 577
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.64919278891202e-05,
+      "loss": 0.8317,
+      "step": 578
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.647915395380539e-05,
+      "loss": 0.8217,
+      "step": 579
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.6466361770869495e-05,
+      "loss": 0.8147,
+      "step": 580
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.6453551376339886e-05,
+      "loss": 0.8212,
+      "step": 581
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.6440722806295237e-05,
+      "loss": 0.8094,
+      "step": 582
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.6427876096865394e-05,
+      "loss": 0.8215,
+      "step": 583
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.64150112842313e-05,
+      "loss": 0.8321,
+      "step": 584
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.640212840462488e-05,
+      "loss": 0.8,
+      "step": 585
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.6389227494328946e-05,
+      "loss": 0.8093,
+      "step": 586
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.6376308589677083e-05,
+      "loss": 0.8205,
+      "step": 587
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.6363371727053567e-05,
+      "loss": 0.8179,
+      "step": 588
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.6350416942893234e-05,
+      "loss": 0.8251,
+      "step": 589
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.6337444273681408e-05,
+      "loss": 0.826,
+      "step": 590
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.6324453755953772e-05,
+      "loss": 0.8044,
+      "step": 591
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.6311445426296294e-05,
+      "loss": 0.7917,
+      "step": 592
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.6298419321345078e-05,
+      "loss": 0.8055,
+      "step": 593
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.6285375477786322e-05,
+      "loss": 0.8364,
+      "step": 594
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.6272313932356164e-05,
+      "loss": 0.7997,
+      "step": 595
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.6259234721840595e-05,
+      "loss": 0.8251,
+      "step": 596
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.6246137883075363e-05,
+      "loss": 0.8298,
+      "step": 597
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.6233023452945866e-05,
+      "loss": 0.8503,
+      "step": 598
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.621989146838704e-05,
+      "loss": 0.8218,
+      "step": 599
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.6206741966383268e-05,
+      "loss": 0.8359,
+      "step": 600
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.6193574983968257e-05,
+      "loss": 0.8104,
+      "step": 601
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.6180390558224956e-05,
+      "loss": 0.8077,
+      "step": 602
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.6167188726285433e-05,
+      "loss": 0.7972,
+      "step": 603
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.6153969525330783e-05,
+      "loss": 0.8008,
+      "step": 604
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.614073299259101e-05,
+      "loss": 0.8126,
+      "step": 605
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.6127479165344948e-05,
+      "loss": 0.8069,
+      "step": 606
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.6114208080920125e-05,
+      "loss": 0.8237,
+      "step": 607
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.6100919776692664e-05,
+      "loss": 0.8156,
+      "step": 608
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.608761429008721e-05,
+      "loss": 0.8196,
+      "step": 609
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.607429165857677e-05,
+      "loss": 0.8059,
+      "step": 610
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.6060951919682665e-05,
+      "loss": 0.8287,
+      "step": 611
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.6047595110974376e-05,
+      "loss": 0.7994,
+      "step": 612
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.603422127006947e-05,
+      "loss": 0.8078,
+      "step": 613
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.6020830434633476e-05,
+      "loss": 0.8112,
+      "step": 614
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.600742264237979e-05,
+      "loss": 0.7999,
+      "step": 615
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.5993997931069568e-05,
+      "loss": 0.8102,
+      "step": 616
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.5980556338511602e-05,
+      "loss": 0.824,
+      "step": 617
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.5967097902562242e-05,
+      "loss": 0.8076,
+      "step": 618
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.5953622661125274e-05,
+      "loss": 0.843,
+      "step": 619
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.5940130652151803e-05,
+      "loss": 0.8135,
+      "step": 620
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.592662191364017e-05,
+      "loss": 0.8129,
+      "step": 621
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.5913096483635827e-05,
+      "loss": 0.828,
+      "step": 622
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.5899554400231233e-05,
+      "loss": 0.8227,
+      "step": 623
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.5885995701565748e-05,
+      "loss": 0.8201,
+      "step": 624
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.587242042582554e-05,
+      "loss": 0.8136,
+      "step": 625
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.5858828611243445e-05,
+      "loss": 0.8087,
+      "step": 626
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.584522029609889e-05,
+      "loss": 0.8213,
+      "step": 627
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.5831595518717772e-05,
+      "loss": 0.8055,
+      "step": 628
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.5817954317472347e-05,
+      "loss": 0.8447,
+      "step": 629
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.5804296730781134e-05,
+      "loss": 0.7988,
+      "step": 630
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.579062279710879e-05,
+      "loss": 0.8083,
+      "step": 631
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.577693255496602e-05,
+      "loss": 0.8068,
+      "step": 632
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.5763226042909455e-05,
+      "loss": 0.8229,
+      "step": 633
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.5749503299541547e-05,
+      "loss": 0.8223,
+      "step": 634
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.573576436351046e-05,
+      "loss": 0.8261,
+      "step": 635
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5722009273509972e-05,
+      "loss": 0.8007,
+      "step": 636
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5708238068279343e-05,
+      "loss": 0.7957,
+      "step": 637
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.569445078660323e-05,
+      "loss": 0.7991,
+      "step": 638
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.568064746731156e-05,
+      "loss": 0.8011,
+      "step": 639
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.566682814927943e-05,
+      "loss": 0.8211,
+      "step": 640
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.5652992871426994e-05,
+      "loss": 0.785,
+      "step": 641
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.563914167271936e-05,
+      "loss": 0.8036,
+      "step": 642
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.562527459216647e-05,
+      "loss": 0.8012,
+      "step": 643
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.561139166882299e-05,
+      "loss": 0.8176,
+      "step": 644
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.5597492941788225e-05,
+      "loss": 0.8069,
+      "step": 645
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.5583578450205957e-05,
+      "loss": 0.8098,
+      "step": 646
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.5569648233264395e-05,
+      "loss": 0.8246,
+      "step": 647
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.5555702330196024e-05,
+      "loss": 0.7992,
+      "step": 648
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.5541740780277508e-05,
+      "loss": 0.8192,
+      "step": 649
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.5527763622829577e-05,
+      "loss": 0.8256,
+      "step": 650
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.551377089721692e-05,
+      "loss": 0.7951,
+      "step": 651
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.5499762642848072e-05,
+      "loss": 0.811,
+      "step": 652
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.5485738899175293e-05,
+      "loss": 0.8223,
+      "step": 653
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.547169970569449e-05,
+      "loss": 0.8193,
+      "step": 654
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.5457645101945046e-05,
+      "loss": 0.8221,
+      "step": 655
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.5443575127509786e-05,
+      "loss": 0.8043,
+      "step": 656
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.542948982201479e-05,
+      "loss": 0.8304,
+      "step": 657
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.5415389225129334e-05,
+      "loss": 0.7808,
+      "step": 658
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.5401273376565756e-05,
+      "loss": 0.7949,
+      "step": 659
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.5387142316079345e-05,
+      "loss": 0.7949,
+      "step": 660
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.5372996083468242e-05,
+      "loss": 0.8066,
+      "step": 661
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5358834718573302e-05,
+      "loss": 0.8126,
+      "step": 662
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5344658261278013e-05,
+      "loss": 0.8076,
+      "step": 663
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5330466751508358e-05,
+      "loss": 0.8045,
+      "step": 664
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.5316260229232728e-05,
+      "loss": 0.7985,
+      "step": 665
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.530203873446177e-05,
+      "loss": 0.8073,
+      "step": 666
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.5287802307248325e-05,
+      "loss": 0.8085,
+      "step": 667
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.5273550987687277e-05,
+      "loss": 0.8343,
+      "step": 668
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.525928481591544e-05,
+      "loss": 0.8195,
+      "step": 669
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.5245003832111485e-05,
+      "loss": 0.8047,
+      "step": 670
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.5230708076495777e-05,
+      "loss": 0.8116,
+      "step": 671
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.5216397589330289e-05,
+      "loss": 0.824,
+      "step": 672
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.520207241091849e-05,
+      "loss": 0.8046,
+      "step": 673
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.5187732581605217e-05,
+      "loss": 0.8097,
+      "step": 674
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.5173378141776569e-05,
+      "loss": 0.8132,
+      "step": 675
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.51590091318598e-05,
+      "loss": 0.8298,
+      "step": 676
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.5144625592323197e-05,
+      "loss": 0.8088,
+      "step": 677
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.5130227563675957e-05,
+      "loss": 0.8226,
+      "step": 678
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.5115815086468103e-05,
+      "loss": 0.8068,
+      "step": 679
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.510138820129033e-05,
+      "loss": 0.8295,
+      "step": 680
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.5086946948773924e-05,
+      "loss": 0.8107,
+      "step": 681
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.5072491369590633e-05,
+      "loss": 0.7997,
+      "step": 682
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.5058021504452553e-05,
+      "loss": 0.8285,
+      "step": 683
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.5043537394112008e-05,
+      "loss": 0.8198,
+      "step": 684
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.5029039079361449e-05,
+      "loss": 0.8121,
+      "step": 685
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.5014526601033335e-05,
+      "loss": 0.8196,
+      "step": 686
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.8195,
+      "step": 687
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.4985459317173571e-05,
+      "loss": 0.801,
+      "step": 688
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.4970904593505822e-05,
+      "loss": 0.8167,
+      "step": 689
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.495633586998807e-05,
+      "loss": 0.8088,
+      "step": 690
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.494175318765107e-05,
+      "loss": 0.8198,
+      "step": 691
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.4927156587564886e-05,
+      "loss": 0.8201,
+      "step": 692
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.4912546110838775e-05,
+      "loss": 0.8245,
+      "step": 693
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.489792179862108e-05,
+      "loss": 0.8137,
+      "step": 694
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.4883283692099114e-05,
+      "loss": 0.7726,
+      "step": 695
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.4868631832499026e-05,
+      "loss": 0.8071,
+      "step": 696
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.4853966261085708e-05,
+      "loss": 0.805,
+      "step": 697
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.4839287019162675e-05,
+      "loss": 0.7944,
+      "step": 698
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.4824594148071936e-05,
+      "loss": 0.7955,
+      "step": 699
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.4809887689193878e-05,
+      "loss": 0.8086,
+      "step": 700
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.4795167683947168e-05,
+      "loss": 0.8124,
+      "step": 701
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.4780434173788617e-05,
+      "loss": 0.8353,
+      "step": 702
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.4765687200213079e-05,
+      "loss": 0.812,
+      "step": 703
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.4750926804753315e-05,
+      "loss": 0.7964,
+      "step": 704
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.4736153028979893e-05,
+      "loss": 0.797,
+      "step": 705
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.4721365914501067e-05,
+      "loss": 0.8204,
+      "step": 706
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.4706565502962648e-05,
+      "loss": 0.827,
+      "step": 707
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.4691751836047907e-05,
+      "loss": 0.814,
+      "step": 708
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.4676924955477443e-05,
+      "loss": 0.7938,
+      "step": 709
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.4662084903009066e-05,
+      "loss": 0.805,
+      "step": 710
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.4647231720437687e-05,
+      "loss": 0.8329,
+      "step": 711
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.4632365449595197e-05,
+      "loss": 0.8056,
+      "step": 712
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.4617486132350343e-05,
+      "loss": 0.7969,
+      "step": 713
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.4602593810608619e-05,
+      "loss": 0.8022,
+      "step": 714
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.4587688526312143e-05,
+      "loss": 0.7923,
+      "step": 715
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.4572770321439542e-05,
+      "loss": 0.816,
+      "step": 716
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.4557839238005831e-05,
+      "loss": 0.8073,
+      "step": 717
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.4542895318062295e-05,
+      "loss": 0.8177,
+      "step": 718
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.4527938603696376e-05,
+      "loss": 0.7943,
+      "step": 719
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.4512969137031538e-05,
+      "loss": 0.811,
+      "step": 720
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.449798696022717e-05,
+      "loss": 0.8381,
+      "step": 721
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.4482992115478458e-05,
+      "loss": 0.8073,
+      "step": 722
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.4467984645016259e-05,
+      "loss": 0.8145,
+      "step": 723
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.4452964591106992e-05,
+      "loss": 0.8145,
+      "step": 724
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.4437931996052518e-05,
+      "loss": 0.8092,
+      "step": 725
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.4422886902190014e-05,
+      "loss": 0.8054,
+      "step": 726
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.4407829351891858e-05,
+      "loss": 0.7986,
+      "step": 727
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.4392759387565515e-05,
+      "loss": 0.8167,
+      "step": 728
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.4377677051653404e-05,
+      "loss": 0.8085,
+      "step": 729
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.43625823866328e-05,
+      "loss": 0.814,
+      "step": 730
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.4347475435015686e-05,
+      "loss": 0.8319,
+      "step": 731
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.4332356239348654e-05,
+      "loss": 0.8128,
+      "step": 732
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.4317224842212784e-05,
+      "loss": 0.8112,
+      "step": 733
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.4302081286223522e-05,
+      "loss": 0.8057,
+      "step": 734
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.4286925614030542e-05,
+      "loss": 0.8061,
+      "step": 735
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.4271757868317662e-05,
+      "loss": 0.8231,
+      "step": 736
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.4256578091802686e-05,
+      "loss": 0.7965,
+      "step": 737
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.4241386327237312e-05,
+      "loss": 0.814,
+      "step": 738
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.4226182617406996e-05,
+      "loss": 0.8168,
+      "step": 739
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.4210967005130837e-05,
+      "loss": 0.8155,
+      "step": 740
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.419573953326146e-05,
+      "loss": 0.825,
+      "step": 741
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.4180500244684878e-05,
+      "loss": 0.8111,
+      "step": 742
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.4165249182320401e-05,
+      "loss": 0.8245,
+      "step": 743
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4149986389120485e-05,
+      "loss": 0.8212,
+      "step": 744
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4134711908070631e-05,
+      "loss": 0.8275,
+      "step": 745
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4119425782189256e-05,
+      "loss": 0.8279,
+      "step": 746
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.410412805452757e-05,
+      "loss": 0.8061,
+      "step": 747
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4088818768169462e-05,
+      "loss": 0.8091,
+      "step": 748
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4073497966231376e-05,
+      "loss": 0.8284,
+      "step": 749
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4058165691862176e-05,
+      "loss": 0.8138,
+      "step": 750
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.404282198824305e-05,
+      "loss": 0.8084,
+      "step": 751
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4027466898587375e-05,
+      "loss": 0.8089,
+      "step": 752
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4012100466140579e-05,
+      "loss": 0.8235,
+      "step": 753
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.3996722734180057e-05,
+      "loss": 0.804,
+      "step": 754
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.3981333746015012e-05,
+      "loss": 0.7989,
+      "step": 755
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.3965933544986351e-05,
+      "loss": 0.8038,
+      "step": 756
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.395052217446657e-05,
+      "loss": 0.8121,
+      "step": 757
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.3935099677859609e-05,
+      "loss": 0.8193,
+      "step": 758
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3919666098600753e-05,
+      "loss": 0.7996,
+      "step": 759
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3904221480156492e-05,
+      "loss": 0.7828,
+      "step": 760
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3888765866024414e-05,
+      "loss": 0.822,
+      "step": 761
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3873299299733068e-05,
+      "loss": 0.7771,
+      "step": 762
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3857821824841855e-05,
+      "loss": 0.8235,
+      "step": 763
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3842333484940894e-05,
+      "loss": 0.79,
+      "step": 764
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3826834323650899e-05,
+      "loss": 0.8116,
+      "step": 765
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3811324384623072e-05,
+      "loss": 0.8099,
+      "step": 766
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3795803711538966e-05,
+      "loss": 0.8258,
+      "step": 767
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3780272348110356e-05,
+      "loss": 0.8134,
+      "step": 768
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3764730338079133e-05,
+      "loss": 0.8095,
+      "step": 769
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3749177725217176e-05,
+      "loss": 0.8064,
+      "step": 770
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3733614553326211e-05,
+      "loss": 0.8148,
+      "step": 771
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3718040866237718e-05,
+      "loss": 0.7805,
+      "step": 772
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3702456707812789e-05,
+      "loss": 0.7609,
+      "step": 773
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.368686212194199e-05,
+      "loss": 0.7411,
+      "step": 774
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.3671257152545277e-05,
+      "loss": 0.7382,
+      "step": 775
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.365564184357184e-05,
+      "loss": 0.747,
+      "step": 776
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.3640016238999984e-05,
+      "loss": 0.7219,
+      "step": 777
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.3624380382837017e-05,
+      "loss": 0.7269,
+      "step": 778
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.3608734319119123e-05,
+      "loss": 0.7534,
+      "step": 779
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.3593078091911218e-05,
+      "loss": 0.7318,
+      "step": 780
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.3577411745306858e-05,
+      "loss": 0.7073,
+      "step": 781
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.3561735323428098e-05,
+      "loss": 0.7068,
+      "step": 782
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.3546048870425356e-05,
+      "loss": 0.715,
+      "step": 783
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.3530352430477313e-05,
+      "loss": 0.7323,
+      "step": 784
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.3514646047790776e-05,
+      "loss": 0.7123,
+      "step": 785
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.3498929766600543e-05,
+      "loss": 0.7216,
+      "step": 786
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.34832036311693e-05,
+      "loss": 0.7143,
+      "step": 787
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.3467467685787494e-05,
+      "loss": 0.7136,
+      "step": 788
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.3451721974773179e-05,
+      "loss": 0.7299,
+      "step": 789
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.3435966542471928e-05,
+      "loss": 0.7,
+      "step": 790
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.342020143325669e-05,
+      "loss": 0.7199,
+      "step": 791
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.340442669152766e-05,
+      "loss": 0.7188,
+      "step": 792
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.3388642361712178e-05,
+      "loss": 0.7305,
+      "step": 793
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.3372848488264571e-05,
+      "loss": 0.7068,
+      "step": 794
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.335704511566605e-05,
+      "loss": 0.7286,
+      "step": 795
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.3341232288424581e-05,
+      "loss": 0.7042,
+      "step": 796
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.3325410051074759e-05,
+      "loss": 0.7069,
+      "step": 797
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.3309578448177674e-05,
+      "loss": 0.7136,
+      "step": 798
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.3293737524320798e-05,
+      "loss": 0.715,
+      "step": 799
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.3277887324117857e-05,
+      "loss": 0.7225,
+      "step": 800
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.3262027892208696e-05,
+      "loss": 0.7065,
+      "step": 801
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.3246159273259163e-05,
+      "loss": 0.7093,
+      "step": 802
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.3230281511960981e-05,
+      "loss": 0.7275,
+      "step": 803
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.3214394653031616e-05,
+      "loss": 0.7202,
+      "step": 804
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.3198498741214167e-05,
+      "loss": 0.7532,
+      "step": 805
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.3182593821277219e-05,
+      "loss": 0.7123,
+      "step": 806
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.3166679938014728e-05,
+      "loss": 0.7233,
+      "step": 807
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.3150757136245897e-05,
+      "loss": 0.7397,
+      "step": 808
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.3134825460815048e-05,
+      "loss": 0.7196,
+      "step": 809
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.311888495659149e-05,
+      "loss": 0.7079,
+      "step": 810
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.3102935668469403e-05,
+      "loss": 0.7225,
+      "step": 811
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.3086977641367702e-05,
+      "loss": 0.7278,
+      "step": 812
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.3071010920229909e-05,
+      "loss": 0.7046,
+      "step": 813
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.305503555002404e-05,
+      "loss": 0.7112,
+      "step": 814
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.303905157574247e-05,
+      "loss": 0.7225,
+      "step": 815
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.3023059042401797e-05,
+      "loss": 0.696,
+      "step": 816
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.300705799504273e-05,
+      "loss": 0.7164,
+      "step": 817
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.2991048478729963e-05,
+      "loss": 0.7371,
+      "step": 818
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.297503053855203e-05,
+      "loss": 0.7299,
+      "step": 819
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.2959004219621196e-05,
+      "loss": 0.7285,
+      "step": 820
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.2942969567073325e-05,
+      "loss": 0.733,
+      "step": 821
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.2926926626067745e-05,
+      "loss": 0.7276,
+      "step": 822
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.291087544178713e-05,
+      "loss": 0.7142,
+      "step": 823
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.2894816059437379e-05,
+      "loss": 0.6995,
+      "step": 824
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.2878748524247462e-05,
+      "loss": 0.7294,
+      "step": 825
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.2862672881469327e-05,
+      "loss": 0.723,
+      "step": 826
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.2846589176377748e-05,
+      "loss": 0.7259,
+      "step": 827
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.2830497454270206e-05,
+      "loss": 0.7313,
+      "step": 828
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.2814397760466758e-05,
+      "loss": 0.7295,
+      "step": 829
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.2798290140309924e-05,
+      "loss": 0.7216,
+      "step": 830
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.2782174639164528e-05,
+      "loss": 0.6987,
+      "step": 831
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.2766051302417609e-05,
+      "loss": 0.709,
+      "step": 832
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.2749920175478262e-05,
+      "loss": 0.7125,
+      "step": 833
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.2733781303777524e-05,
+      "loss": 0.7169,
+      "step": 834
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.2717634732768245e-05,
+      "loss": 0.7432,
+      "step": 835
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.2701480507924966e-05,
+      "loss": 0.7224,
+      "step": 836
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.2685318674743769e-05,
+      "loss": 0.7295,
+      "step": 837
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.266914927874218e-05,
+      "loss": 0.725,
+      "step": 838
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.2652972365459008e-05,
+      "loss": 0.7143,
+      "step": 839
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.263678798045425e-05,
+      "loss": 0.7338,
+      "step": 840
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.2620596169308935e-05,
+      "loss": 0.7028,
+      "step": 841
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.260439697762501e-05,
+      "loss": 0.7426,
+      "step": 842
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 0.7219,
+      "step": 843
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.2571976635152921e-05,
+      "loss": 0.7232,
+      "step": 844
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.2555755575672068e-05,
+      "loss": 0.7224,
+      "step": 845
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.2539527318266971e-05,
+      "loss": 0.7327,
+      "step": 846
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.2523291908642219e-05,
+      "loss": 0.7272,
+      "step": 847
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.2507049392522552e-05,
+      "loss": 0.7241,
+      "step": 848
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.2490799815652719e-05,
+      "loss": 0.721,
+      "step": 849
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.2474543223797353e-05,
+      "loss": 0.7423,
+      "step": 850
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.2458279662740853e-05,
+      "loss": 0.7219,
+      "step": 851
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.244200917828723e-05,
+      "loss": 0.7252,
+      "step": 852
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.2425731816260013e-05,
+      "loss": 0.7103,
+      "step": 853
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.2409447622502086e-05,
+      "loss": 0.7216,
+      "step": 854
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.2393156642875579e-05,
+      "loss": 0.7392,
+      "step": 855
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.2376858923261732e-05,
+      "loss": 0.718,
+      "step": 856
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.2360554509560771e-05,
+      "loss": 0.7111,
+      "step": 857
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.2344243447691767e-05,
+      "loss": 0.7306,
+      "step": 858
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.2327925783592522e-05,
+      "loss": 0.7327,
+      "step": 859
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.2311601563219436e-05,
+      "loss": 0.7172,
+      "step": 860
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.2295270832547358e-05,
+      "loss": 0.7199,
+      "step": 861
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.2278933637569486e-05,
+      "loss": 0.7194,
+      "step": 862
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.2262590024297226e-05,
+      "loss": 0.7274,
+      "step": 863
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.2246240038760042e-05,
+      "loss": 0.7153,
+      "step": 864
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.2229883727005365e-05,
+      "loss": 0.7148,
+      "step": 865
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.2213521135098435e-05,
+      "loss": 0.7438,
+      "step": 866
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.2197152309122173e-05,
+      "loss": 0.7393,
+      "step": 867
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.2180777295177066e-05,
+      "loss": 0.7405,
+      "step": 868
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.2164396139381029e-05,
+      "loss": 0.7206,
+      "step": 869
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.214800888786927e-05,
+      "loss": 0.7112,
+      "step": 870
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.2131615586794162e-05,
+      "loss": 0.7183,
+      "step": 871
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.2115216282325123e-05,
+      "loss": 0.7212,
+      "step": 872
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.2098811020648475e-05,
+      "loss": 0.7086,
+      "step": 873
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.2082399847967318e-05,
+      "loss": 0.7401,
+      "step": 874
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.2065982810501404e-05,
+      "loss": 0.7159,
+      "step": 875
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.204955995448699e-05,
+      "loss": 0.728,
+      "step": 876
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.203313132617673e-05,
+      "loss": 0.7275,
+      "step": 877
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.2016696971839539e-05,
+      "loss": 0.7125,
+      "step": 878
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.2000256937760446e-05,
+      "loss": 0.7356,
+      "step": 879
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.1983811270240484e-05,
+      "loss": 0.7225,
+      "step": 880
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.1967360015596554e-05,
+      "loss": 0.7068,
+      "step": 881
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.1950903220161286e-05,
+      "loss": 0.7353,
+      "step": 882
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.1934440930282913e-05,
+      "loss": 0.7484,
+      "step": 883
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.191797319232516e-05,
+      "loss": 0.761,
+      "step": 884
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.1901500052667068e-05,
+      "loss": 0.7294,
+      "step": 885
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.1885021557702914e-05,
+      "loss": 0.7412,
+      "step": 886
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.1868537753842052e-05,
+      "loss": 0.7192,
+      "step": 887
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.1852048687508775e-05,
+      "loss": 0.7286,
+      "step": 888
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.1835554405142216e-05,
+      "loss": 0.7314,
+      "step": 889
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.1819054953196187e-05,
+      "loss": 0.7438,
+      "step": 890
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.180255037813906e-05,
+      "loss": 0.7215,
+      "step": 891
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.1786040726453638e-05,
+      "loss": 0.7293,
+      "step": 892
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.1769526044637023e-05,
+      "loss": 0.7067,
+      "step": 893
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.1753006379200478e-05,
+      "loss": 0.7136,
+      "step": 894
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.1736481776669307e-05,
+      "loss": 0.721,
+      "step": 895
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.1719952283582716e-05,
+      "loss": 0.7355,
+      "step": 896
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.1703417946493682e-05,
+      "loss": 0.7222,
+      "step": 897
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.1686878811968837e-05,
+      "loss": 0.7335,
+      "step": 898
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.167033492658831e-05,
+      "loss": 0.7003,
+      "step": 899
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.1653786336945614e-05,
+      "loss": 0.715,
+      "step": 900
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.163723308964751e-05,
+      "loss": 0.7386,
+      "step": 901
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.1620675231313887e-05,
+      "loss": 0.7255,
+      "step": 902
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.1604112808577603e-05,
+      "loss": 0.7259,
+      "step": 903
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.1587545868084384e-05,
+      "loss": 0.7314,
+      "step": 904
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.1570974456492678e-05,
+      "loss": 0.7495,
+      "step": 905
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.1554398620473516e-05,
+      "loss": 0.717,
+      "step": 906
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.1537818406710406e-05,
+      "loss": 0.731,
+      "step": 907
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.1521233861899168e-05,
+      "loss": 0.7238,
+      "step": 908
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.1504645032747832e-05,
+      "loss": 0.7236,
+      "step": 909
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.1488051965976485e-05,
+      "loss": 0.7238,
+      "step": 910
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.1471454708317163e-05,
+      "loss": 0.7473,
+      "step": 911
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.1454853306513687e-05,
+      "loss": 0.7175,
+      "step": 912
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1438247807321558e-05,
+      "loss": 0.7491,
+      "step": 913
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1421638257507826e-05,
+      "loss": 0.7211,
+      "step": 914
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1405024703850929e-05,
+      "loss": 0.7312,
+      "step": 915
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1388407193140597e-05,
+      "loss": 0.7397,
+      "step": 916
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.1371785772177702e-05,
+      "loss": 0.7143,
+      "step": 917
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.1355160487774119e-05,
+      "loss": 0.7253,
+      "step": 918
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.1338531386752618e-05,
+      "loss": 0.7317,
+      "step": 919
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.1321898515946708e-05,
+      "loss": 0.7133,
+      "step": 920
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.130526192220052e-05,
+      "loss": 0.7346,
+      "step": 921
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.1288621652368663e-05,
+      "loss": 0.7513,
+      "step": 922
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.1271977753316111e-05,
+      "loss": 0.7232,
+      "step": 923
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.1255330271918048e-05,
+      "loss": 0.7214,
+      "step": 924
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.1238679255059753e-05,
+      "loss": 0.7136,
+      "step": 925
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.122202474963646e-05,
+      "loss": 0.7221,
+      "step": 926
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.1205366802553231e-05,
+      "loss": 0.7337,
+      "step": 927
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.1188705460724819e-05,
+      "loss": 0.7342,
+      "step": 928
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.1172040771075536e-05,
+      "loss": 0.7238,
+      "step": 929
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.1155372780539125e-05,
+      "loss": 0.7334,
+      "step": 930
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.1138701536058624e-05,
+      "loss": 0.715,
+      "step": 931
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.1122027084586238e-05,
+      "loss": 0.7269,
+      "step": 932
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.11053494730832e-05,
+      "loss": 0.7331,
+      "step": 933
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.1088668748519646e-05,
+      "loss": 0.724,
+      "step": 934
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.107198495787448e-05,
+      "loss": 0.7333,
+      "step": 935
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.1055298148135236e-05,
+      "loss": 0.716,
+      "step": 936
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.1038608366297956e-05,
+      "loss": 0.7202,
+      "step": 937
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.1021915659367053e-05,
+      "loss": 0.7207,
+      "step": 938
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.1005220074355172e-05,
+      "loss": 0.7377,
+      "step": 939
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.0988521658283072e-05,
+      "loss": 0.7214,
+      "step": 940
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.0971820458179478e-05,
+      "loss": 0.7534,
+      "step": 941
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.0955116521080959e-05,
+      "loss": 0.7368,
+      "step": 942
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.0938409894031793e-05,
+      "loss": 0.7336,
+      "step": 943
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.0921700624083837e-05,
+      "loss": 0.728,
+      "step": 944
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.090498875829638e-05,
+      "loss": 0.7229,
+      "step": 945
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.0888274343736034e-05,
+      "loss": 0.7357,
+      "step": 946
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.0871557427476585e-05,
+      "loss": 0.7188,
+      "step": 947
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.085483805659886e-05,
+      "loss": 0.7369,
+      "step": 948
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.0838116278190607e-05,
+      "loss": 0.7233,
+      "step": 949
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.0821392139346352e-05,
+      "loss": 0.7413,
+      "step": 950
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.0804665687167262e-05,
+      "loss": 0.7177,
+      "step": 951
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.0787936968761029e-05,
+      "loss": 0.7395,
+      "step": 952
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.0771206031241724e-05,
+      "loss": 0.7326,
+      "step": 953
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.0754472921729661e-05,
+      "loss": 0.7441,
+      "step": 954
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.0737737687351284e-05,
+      "loss": 0.7151,
+      "step": 955
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.0721000375239012e-05,
+      "loss": 0.73,
+      "step": 956
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.070426103253112e-05,
+      "loss": 0.7207,
+      "step": 957
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.0687519706371596e-05,
+      "loss": 0.7327,
+      "step": 958
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.0670776443910024e-05,
+      "loss": 0.7199,
+      "step": 959
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.0654031292301432e-05,
+      "loss": 0.7246,
+      "step": 960
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.0637284298706169e-05,
+      "loss": 0.7296,
+      "step": 961
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.0620535510289785e-05,
+      "loss": 0.7402,
+      "step": 962
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.0603784974222862e-05,
+      "loss": 0.7178,
+      "step": 963
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.0587032737680922e-05,
+      "loss": 0.7253,
+      "step": 964
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.0570278847844275e-05,
+      "loss": 0.7242,
+      "step": 965
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.0553523351897879e-05,
+      "loss": 0.7126,
+      "step": 966
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.0536766297031216e-05,
+      "loss": 0.718,
+      "step": 967
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.0520007730438171e-05,
+      "loss": 0.7315,
+      "step": 968
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.0503247699316867e-05,
+      "loss": 0.7462,
+      "step": 969
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.048648625086957e-05,
+      "loss": 0.7362,
+      "step": 970
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.0469723432302528e-05,
+      "loss": 0.7025,
+      "step": 971
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.0452959290825846e-05,
+      "loss": 0.7236,
+      "step": 972
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.0436193873653362e-05,
+      "loss": 0.7287,
+      "step": 973
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.0419427228002504e-05,
+      "loss": 0.7292,
+      "step": 974
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.0402659401094154e-05,
+      "loss": 0.7382,
+      "step": 975
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.0385890440152529e-05,
+      "loss": 0.7129,
+      "step": 976
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.0369120392405038e-05,
+      "loss": 0.7318,
+      "step": 977
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.0352349305082148e-05,
+      "loss": 0.7024,
+      "step": 978
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.0335577225417255e-05,
+      "loss": 0.7262,
+      "step": 979
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.0318804200646553e-05,
+      "loss": 0.7159,
+      "step": 980
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.030203027800889e-05,
+      "loss": 0.7321,
+      "step": 981
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.028525550474565e-05,
+      "loss": 0.7138,
+      "step": 982
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.0268479928100615e-05,
+      "loss": 0.7267,
+      "step": 983
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.0251703595319816e-05,
+      "loss": 0.7122,
+      "step": 984
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.0234926553651423e-05,
+      "loss": 0.7255,
+      "step": 985
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.0218148850345613e-05,
+      "loss": 0.7189,
+      "step": 986
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.0201370532654404e-05,
+      "loss": 0.7204,
+      "step": 987
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.0184591647831562e-05,
+      "loss": 0.7411,
+      "step": 988
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.016781224313244e-05,
+      "loss": 0.726,
+      "step": 989
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.015103236581386e-05,
+      "loss": 0.721,
+      "step": 990
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.0134252063133976e-05,
+      "loss": 0.7081,
+      "step": 991
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.0117471382352138e-05,
+      "loss": 0.709,
+      "step": 992
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.0100690370728756e-05,
+      "loss": 0.7324,
+      "step": 993
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0083909075525183e-05,
+      "loss": 0.7105,
+      "step": 994
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0067127544003564e-05,
+      "loss": 0.7456,
+      "step": 995
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0050345823426704e-05,
+      "loss": 0.7363,
+      "step": 996
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.003356396105795e-05,
+      "loss": 0.7167,
+      "step": 997
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.0016782004161047e-05,
+      "loss": 0.707,
+      "step": 998
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1e-05,
+      "loss": 0.7321,
+      "step": 999
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 9.983217995838955e-06,
+      "loss": 0.715,
+      "step": 1000
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 9.966436038942054e-06,
+      "loss": 0.7366,
+      "step": 1001
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 9.9496541765733e-06,
+      "loss": 0.7196,
+      "step": 1002
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 9.932872455996441e-06,
+      "loss": 0.7393,
+      "step": 1003
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 9.91609092447482e-06,
+      "loss": 0.7211,
+      "step": 1004
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 9.899309629271246e-06,
+      "loss": 0.7045,
+      "step": 1005
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.882528617647866e-06,
+      "loss": 0.7139,
+      "step": 1006
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.865747936866027e-06,
+      "loss": 0.7101,
+      "step": 1007
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.848967634186142e-06,
+      "loss": 0.7208,
+      "step": 1008
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.832187756867562e-06,
+      "loss": 0.7224,
+      "step": 1009
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.815408352168443e-06,
+      "loss": 0.7193,
+      "step": 1010
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.7986294673456e-06,
+      "loss": 0.7237,
+      "step": 1011
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.78185114965439e-06,
+      "loss": 0.75,
+      "step": 1012
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.765073446348579e-06,
+      "loss": 0.727,
+      "step": 1013
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.748296404680189e-06,
+      "loss": 0.7325,
+      "step": 1014
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.73152007189939e-06,
+      "loss": 0.7439,
+      "step": 1015
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.714744495254353e-06,
+      "loss": 0.7163,
+      "step": 1016
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.697969721991114e-06,
+      "loss": 0.7104,
+      "step": 1017
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 9.681195799353452e-06,
+      "loss": 0.7253,
+      "step": 1018
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 9.664422774582748e-06,
+      "loss": 0.7256,
+      "step": 1019
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 9.647650694917857e-06,
+      "loss": 0.7234,
+      "step": 1020
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.630879607594967e-06,
+      "loss": 0.7151,
+      "step": 1021
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.614109559847475e-06,
+      "loss": 0.7297,
+      "step": 1022
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.597340598905851e-06,
+      "loss": 0.7241,
+      "step": 1023
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.580572771997503e-06,
+      "loss": 0.7324,
+      "step": 1024
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 9.563806126346643e-06,
+      "loss": 0.7315,
+      "step": 1025
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 9.547040709174159e-06,
+      "loss": 0.7087,
+      "step": 1026
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 9.530276567697479e-06,
+      "loss": 0.7321,
+      "step": 1027
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 9.513513749130435e-06,
+      "loss": 0.7307,
+      "step": 1028
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.496752300683133e-06,
+      "loss": 0.7202,
+      "step": 1029
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.479992269561832e-06,
+      "loss": 0.7086,
+      "step": 1030
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.463233702968784e-06,
+      "loss": 0.7334,
+      "step": 1031
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.446476648102123e-06,
+      "loss": 0.7352,
+      "step": 1032
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.429721152155728e-06,
+      "loss": 0.7107,
+      "step": 1033
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.412967262319078e-06,
+      "loss": 0.7332,
+      "step": 1034
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.39621502577714e-06,
+      "loss": 0.7321,
+      "step": 1035
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.379464489710217e-06,
+      "loss": 0.7179,
+      "step": 1036
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.362715701293831e-06,
+      "loss": 0.7394,
+      "step": 1037
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.34596870769857e-06,
+      "loss": 0.7356,
+      "step": 1038
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.329223556089976e-06,
+      "loss": 0.7197,
+      "step": 1039
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.312480293628404e-06,
+      "loss": 0.7194,
+      "step": 1040
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.295738967468881e-06,
+      "loss": 0.7316,
+      "step": 1041
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.278999624760988e-06,
+      "loss": 0.7408,
+      "step": 1042
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.262262312648718e-06,
+      "loss": 0.7301,
+      "step": 1043
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.24552707827034e-06,
+      "loss": 0.7205,
+      "step": 1044
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.22879396875828e-06,
+      "loss": 0.7181,
+      "step": 1045
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.212063031238976e-06,
+      "loss": 0.7168,
+      "step": 1046
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.195334312832742e-06,
+      "loss": 0.7395,
+      "step": 1047
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.178607860653652e-06,
+      "loss": 0.7238,
+      "step": 1048
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.161883721809397e-06,
+      "loss": 0.7331,
+      "step": 1049
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.145161943401143e-06,
+      "loss": 0.7341,
+      "step": 1050
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.128442572523418e-06,
+      "loss": 0.7234,
+      "step": 1051
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.11172565626397e-06,
+      "loss": 0.7306,
+      "step": 1052
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.095011241703623e-06,
+      "loss": 0.75,
+      "step": 1053
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.078299375916166e-06,
+      "loss": 0.7291,
+      "step": 1054
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.061590105968208e-06,
+      "loss": 0.7247,
+      "step": 1055
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.044883478919043e-06,
+      "loss": 0.7243,
+      "step": 1056
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.028179541820525e-06,
+      "loss": 0.7139,
+      "step": 1057
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.011478341716933e-06,
+      "loss": 0.7278,
+      "step": 1058
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 8.994779925644832e-06,
+      "loss": 0.7319,
+      "step": 1059
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.97808434063295e-06,
+      "loss": 0.7153,
+      "step": 1060
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.961391633702047e-06,
+      "loss": 0.7247,
+      "step": 1061
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.944701851864767e-06,
+      "loss": 0.7368,
+      "step": 1062
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.928015042125523e-06,
+      "loss": 0.7454,
+      "step": 1063
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.911331251480357e-06,
+      "loss": 0.7042,
+      "step": 1064
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.894650526916802e-06,
+      "loss": 0.7079,
+      "step": 1065
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.877972915413763e-06,
+      "loss": 0.7407,
+      "step": 1066
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.861298463941379e-06,
+      "loss": 0.7348,
+      "step": 1067
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 8.844627219460879e-06,
+      "loss": 0.728,
+      "step": 1068
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 8.827959228924466e-06,
+      "loss": 0.7115,
+      "step": 1069
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 8.811294539275185e-06,
+      "loss": 0.7112,
+      "step": 1070
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 8.79463319744677e-06,
+      "loss": 0.7194,
+      "step": 1071
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 8.777975250363543e-06,
+      "loss": 0.7311,
+      "step": 1072
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 8.761320744940252e-06,
+      "loss": 0.7292,
+      "step": 1073
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 8.744669728081959e-06,
+      "loss": 0.7123,
+      "step": 1074
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 8.728022246683894e-06,
+      "loss": 0.7159,
+      "step": 1075
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 8.711378347631343e-06,
+      "loss": 0.7355,
+      "step": 1076
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 8.694738077799487e-06,
+      "loss": 0.7218,
+      "step": 1077
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 8.678101484053297e-06,
+      "loss": 0.7342,
+      "step": 1078
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 8.661468613247387e-06,
+      "loss": 0.7224,
+      "step": 1079
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 8.644839512225886e-06,
+      "loss": 0.7288,
+      "step": 1080
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 8.628214227822303e-06,
+      "loss": 0.7198,
+      "step": 1081
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 8.611592806859405e-06,
+      "loss": 0.7305,
+      "step": 1082
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 8.594975296149076e-06,
+      "loss": 0.7062,
+      "step": 1083
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 8.578361742492179e-06,
+      "loss": 0.7416,
+      "step": 1084
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 8.561752192678444e-06,
+      "loss": 0.7283,
+      "step": 1085
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 8.54514669348632e-06,
+      "loss": 0.7257,
+      "step": 1086
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 8.528545291682839e-06,
+      "loss": 0.7341,
+      "step": 1087
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 8.511948034023515e-06,
+      "loss": 0.7384,
+      "step": 1088
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 8.49535496725217e-06,
+      "loss": 0.7183,
+      "step": 1089
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 8.478766138100834e-06,
+      "loss": 0.7185,
+      "step": 1090
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.462181593289596e-06,
+      "loss": 0.722,
+      "step": 1091
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.445601379526482e-06,
+      "loss": 0.734,
+      "step": 1092
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.429025543507324e-06,
+      "loss": 0.7262,
+      "step": 1093
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.412454131915616e-06,
+      "loss": 0.7515,
+      "step": 1094
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.395887191422397e-06,
+      "loss": 0.7286,
+      "step": 1095
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.379324768686113e-06,
+      "loss": 0.7156,
+      "step": 1096
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.36276691035249e-06,
+      "loss": 0.7097,
+      "step": 1097
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.346213663054388e-06,
+      "loss": 0.732,
+      "step": 1098
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.32966507341169e-06,
+      "loss": 0.7204,
+      "step": 1099
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.313121188031164e-06,
+      "loss": 0.7143,
+      "step": 1100
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.296582053506318e-06,
+      "loss": 0.7395,
+      "step": 1101
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.280047716417286e-06,
+      "loss": 0.7023,
+      "step": 1102
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.263518223330698e-06,
+      "loss": 0.7278,
+      "step": 1103
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.246993620799526e-06,
+      "loss": 0.7242,
+      "step": 1104
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.23047395536298e-06,
+      "loss": 0.7336,
+      "step": 1105
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.213959273546365e-06,
+      "loss": 0.7118,
+      "step": 1106
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.197449621860944e-06,
+      "loss": 0.7141,
+      "step": 1107
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.180945046803815e-06,
+      "loss": 0.7227,
+      "step": 1108
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.164445594857787e-06,
+      "loss": 0.7197,
+      "step": 1109
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.147951312491227e-06,
+      "loss": 0.7232,
+      "step": 1110
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.131462246157953e-06,
+      "loss": 0.7291,
+      "step": 1111
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.114978442297089e-06,
+      "loss": 0.7301,
+      "step": 1112
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.098499947332935e-06,
+      "loss": 0.7274,
+      "step": 1113
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.082026807674844e-06,
+      "loss": 0.7297,
+      "step": 1114
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.065559069717089e-06,
+      "loss": 0.7106,
+      "step": 1115
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.04909677983872e-06,
+      "loss": 0.7203,
+      "step": 1116
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.032639984403448e-06,
+      "loss": 0.7189,
+      "step": 1117
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 8.016188729759518e-06,
+      "loss": 0.7424,
+      "step": 1118
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 7.999743062239557e-06,
+      "loss": 0.728,
+      "step": 1119
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 7.983303028160464e-06,
+      "loss": 0.7142,
+      "step": 1120
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 7.966868673823272e-06,
+      "loss": 0.7224,
+      "step": 1121
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 7.950440045513014e-06,
+      "loss": 0.7227,
+      "step": 1122
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 7.9340171894986e-06,
+      "loss": 0.7289,
+      "step": 1123
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 7.917600152032684e-06,
+      "loss": 0.7273,
+      "step": 1124
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 7.901188979351527e-06,
+      "loss": 0.7211,
+      "step": 1125
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 7.884783717674878e-06,
+      "loss": 0.7089,
+      "step": 1126
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 7.868384413205842e-06,
+      "loss": 0.7017,
+      "step": 1127
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 7.851991112130734e-06,
+      "loss": 0.7283,
+      "step": 1128
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 7.835603860618973e-06,
+      "loss": 0.7057,
+      "step": 1129
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.819222704822937e-06,
+      "loss": 0.7349,
+      "step": 1130
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.802847690877832e-06,
+      "loss": 0.7255,
+      "step": 1131
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.78647886490157e-06,
+      "loss": 0.7219,
+      "step": 1132
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 7.77011627299464e-06,
+      "loss": 0.7166,
+      "step": 1133
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 7.753759961239965e-06,
+      "loss": 0.7137,
+      "step": 1134
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 7.73740997570278e-06,
+      "loss": 0.7272,
+      "step": 1135
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 7.721066362430517e-06,
+      "loss": 0.7306,
+      "step": 1136
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 7.704729167452647e-06,
+      "loss": 0.7097,
+      "step": 1137
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 7.68839843678057e-06,
+      "loss": 0.7359,
+      "step": 1138
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 7.67207421640748e-06,
+      "loss": 0.73,
+      "step": 1139
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 7.655756552308238e-06,
+      "loss": 0.7101,
+      "step": 1140
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 7.639445490439234e-06,
+      "loss": 0.7413,
+      "step": 1141
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 7.623141076738271e-06,
+      "loss": 0.7241,
+      "step": 1142
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 7.606843357124426e-06,
+      "loss": 0.7161,
+      "step": 1143
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 7.5905523774979175e-06,
+      "loss": 0.7134,
+      "step": 1144
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 7.574268183739989e-06,
+      "loss": 0.7222,
+      "step": 1145
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 7.557990821712769e-06,
+      "loss": 0.7178,
+      "step": 1146
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 7.541720337259151e-06,
+      "loss": 0.7186,
+      "step": 1147
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 7.525456776202648e-06,
+      "loss": 0.7164,
+      "step": 1148
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 7.5092001843472825e-06,
+      "loss": 0.7351,
+      "step": 1149
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 7.492950607477451e-06,
+      "loss": 0.7361,
+      "step": 1150
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 7.476708091357783e-06,
+      "loss": 0.7318,
+      "step": 1151
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 7.460472681733031e-06,
+      "loss": 0.7269,
+      "step": 1152
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 7.444244424327932e-06,
+      "loss": 0.7068,
+      "step": 1153
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 7.42802336484708e-06,
+      "loss": 0.7277,
+      "step": 1154
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 0.709,
+      "step": 1155
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 7.395603022374991e-06,
+      "loss": 0.7448,
+      "step": 1156
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 7.379403830691067e-06,
+      "loss": 0.7159,
+      "step": 1157
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 7.3632120195457515e-06,
+      "loss": 0.6741,
+      "step": 1158
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 7.347027634540993e-06,
+      "loss": 0.6027,
+      "step": 1159
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 7.330850721257825e-06,
+      "loss": 0.5939,
+      "step": 1160
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 7.314681325256232e-06,
+      "loss": 0.597,
+      "step": 1161
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 7.298519492075037e-06,
+      "loss": 0.5818,
+      "step": 1162
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 7.282365267231757e-06,
+      "loss": 0.5857,
+      "step": 1163
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 7.26621869622248e-06,
+      "loss": 0.5734,
+      "step": 1164
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 7.2500798245217426e-06,
+      "loss": 0.5889,
+      "step": 1165
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 7.2339486975823955e-06,
+      "loss": 0.5827,
+      "step": 1166
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 7.217825360835475e-06,
+      "loss": 0.5859,
+      "step": 1167
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 7.201709859690081e-06,
+      "loss": 0.5596,
+      "step": 1168
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 7.185602239533243e-06,
+      "loss": 0.5687,
+      "step": 1169
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 7.169502545729798e-06,
+      "loss": 0.5632,
+      "step": 1170
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 7.153410823622253e-06,
+      "loss": 0.5508,
+      "step": 1171
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 7.137327118530676e-06,
+      "loss": 0.5804,
+      "step": 1172
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 7.12125147575254e-06,
+      "loss": 0.5761,
+      "step": 1173
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 7.1051839405626255e-06,
+      "loss": 0.5696,
+      "step": 1174
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 7.089124558212872e-06,
+      "loss": 0.5797,
+      "step": 1175
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 7.07307337393226e-06,
+      "loss": 0.5685,
+      "step": 1176
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 7.057030432926678e-06,
+      "loss": 0.5453,
+      "step": 1177
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 7.040995780378806e-06,
+      "loss": 0.5601,
+      "step": 1178
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 7.024969461447973e-06,
+      "loss": 0.5354,
+      "step": 1179
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 7.0089515212700374e-06,
+      "loss": 0.5467,
+      "step": 1180
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 6.992942004957271e-06,
+      "loss": 0.5487,
+      "step": 1181
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 6.976940957598206e-06,
+      "loss": 0.5555,
+      "step": 1182
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 6.960948424257532e-06,
+      "loss": 0.5753,
+      "step": 1183
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 6.944964449975962e-06,
+      "loss": 0.5565,
+      "step": 1184
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 6.928989079770094e-06,
+      "loss": 0.56,
+      "step": 1185
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 6.9130223586323005e-06,
+      "loss": 0.5511,
+      "step": 1186
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 6.8970643315306e-06,
+      "loss": 0.5598,
+      "step": 1187
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 6.881115043408512e-06,
+      "loss": 0.5546,
+      "step": 1188
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 6.865174539184956e-06,
+      "loss": 0.5594,
+      "step": 1189
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 6.849242863754109e-06,
+      "loss": 0.5675,
+      "step": 1190
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 6.833320061985278e-06,
+      "loss": 0.5669,
+      "step": 1191
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 6.817406178722785e-06,
+      "loss": 0.5626,
+      "step": 1192
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 6.8015012587858375e-06,
+      "loss": 0.5643,
+      "step": 1193
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 6.785605346968387e-06,
+      "loss": 0.5723,
+      "step": 1194
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 6.7697184880390235e-06,
+      "loss": 0.5642,
+      "step": 1195
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 6.753840726740843e-06,
+      "loss": 0.5657,
+      "step": 1196
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 6.7379721077913095e-06,
+      "loss": 0.5546,
+      "step": 1197
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 6.7221126758821474e-06,
+      "loss": 0.5604,
+      "step": 1198
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 6.706262475679205e-06,
+      "loss": 0.5558,
+      "step": 1199
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 6.690421551822333e-06,
+      "loss": 0.5581,
+      "step": 1200
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 6.674589948925246e-06,
+      "loss": 0.5472,
+      "step": 1201
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 6.6587677115754225e-06,
+      "loss": 0.5521,
+      "step": 1202
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 6.6429548843339554e-06,
+      "loss": 0.5684,
+      "step": 1203
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 6.627151511735433e-06,
+      "loss": 0.5447,
+      "step": 1204
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 6.611357638287823e-06,
+      "loss": 0.5562,
+      "step": 1205
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 6.595573308472338e-06,
+      "loss": 0.5828,
+      "step": 1206
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 6.579798566743314e-06,
+      "loss": 0.556,
+      "step": 1207
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 6.564033457528074e-06,
+      "loss": 0.5579,
+      "step": 1208
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 6.548278025226822e-06,
+      "loss": 0.5596,
+      "step": 1209
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 6.5325323142125085e-06,
+      "loss": 0.5717,
+      "step": 1210
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 6.5167963688306975e-06,
+      "loss": 0.5484,
+      "step": 1211
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 6.501070233399458e-06,
+      "loss": 0.5623,
+      "step": 1212
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 6.485353952209226e-06,
+      "loss": 0.583,
+      "step": 1213
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 6.469647569522687e-06,
+      "loss": 0.5674,
+      "step": 1214
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 6.453951129574644e-06,
+      "loss": 0.5665,
+      "step": 1215
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 6.438264676571901e-06,
+      "loss": 0.5665,
+      "step": 1216
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 6.422588254693141e-06,
+      "loss": 0.5702,
+      "step": 1217
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 6.406921908088783e-06,
+      "loss": 0.5624,
+      "step": 1218
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 6.3912656808808805e-06,
+      "loss": 0.5469,
+      "step": 1219
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 6.375619617162985e-06,
+      "loss": 0.5722,
+      "step": 1220
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 6.3599837610000194e-06,
+      "loss": 0.5553,
+      "step": 1221
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 6.344358156428163e-06,
+      "loss": 0.5562,
+      "step": 1222
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 6.3287428474547256e-06,
+      "loss": 0.554,
+      "step": 1223
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 6.3131378780580134e-06,
+      "loss": 0.5629,
+      "step": 1224
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 6.297543292187216e-06,
+      "loss": 0.5601,
+      "step": 1225
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 6.281959133762283e-06,
+      "loss": 0.5471,
+      "step": 1226
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 6.266385446673791e-06,
+      "loss": 0.5739,
+      "step": 1227
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 6.250822274782828e-06,
+      "loss": 0.558,
+      "step": 1228
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 6.23526966192087e-06,
+      "loss": 0.5633,
+      "step": 1229
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 6.219727651889647e-06,
+      "loss": 0.5568,
+      "step": 1230
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 6.204196288461037e-06,
+      "loss": 0.562,
+      "step": 1231
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 6.18867561537693e-06,
+      "loss": 0.558,
+      "step": 1232
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 6.173165676349103e-06,
+      "loss": 0.5605,
+      "step": 1233
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 6.157666515059111e-06,
+      "loss": 0.5633,
+      "step": 1234
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 6.142178175158149e-06,
+      "loss": 0.5508,
+      "step": 1235
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 6.1267007002669345e-06,
+      "loss": 0.5581,
+      "step": 1236
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 6.111234133975589e-06,
+      "loss": 0.5753,
+      "step": 1237
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 6.095778519843512e-06,
+      "loss": 0.5674,
+      "step": 1238
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 6.080333901399252e-06,
+      "loss": 0.5797,
+      "step": 1239
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 6.064900322140393e-06,
+      "loss": 0.5538,
+      "step": 1240
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 6.049477825533434e-06,
+      "loss": 0.5656,
+      "step": 1241
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 6.0340664550136494e-06,
+      "loss": 0.55,
+      "step": 1242
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 6.018666253984991e-06,
+      "loss": 0.5599,
+      "step": 1243
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 6.003277265819947e-06,
+      "loss": 0.5668,
+      "step": 1244
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 5.987899533859422e-06,
+      "loss": 0.5671,
+      "step": 1245
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 5.97253310141263e-06,
+      "loss": 0.5676,
+      "step": 1246
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 5.957178011756952e-06,
+      "loss": 0.5675,
+      "step": 1247
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 5.941834308137828e-06,
+      "loss": 0.5714,
+      "step": 1248
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 5.92650203376863e-06,
+      "loss": 0.562,
+      "step": 1249
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 5.9111812318305425e-06,
+      "loss": 0.553,
+      "step": 1250
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 5.895871945472434e-06,
+      "loss": 0.5492,
+      "step": 1251
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 5.880574217810749e-06,
+      "loss": 0.5644,
+      "step": 1252
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 5.865288091929375e-06,
+      "loss": 0.5547,
+      "step": 1253
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 5.85001361087952e-06,
+      "loss": 0.5662,
+      "step": 1254
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 5.834750817679606e-06,
+      "loss": 0.5663,
+      "step": 1255
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 5.819499755315128e-06,
+      "loss": 0.5627,
+      "step": 1256
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 5.8042604667385475e-06,
+      "loss": 0.5545,
+      "step": 1257
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 5.789032994869166e-06,
+      "loss": 0.5619,
+      "step": 1258
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 5.773817382593008e-06,
+      "loss": 0.5703,
+      "step": 1259
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 5.75861367276269e-06,
+      "loss": 0.5626,
+      "step": 1260
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 5.74342190819732e-06,
+      "loss": 0.5599,
+      "step": 1261
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 5.7282421316823444e-06,
+      "loss": 0.5565,
+      "step": 1262
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 5.713074385969457e-06,
+      "loss": 0.5601,
+      "step": 1263
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 5.697918713776479e-06,
+      "loss": 0.5611,
+      "step": 1264
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 5.682775157787213e-06,
+      "loss": 0.5555,
+      "step": 1265
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 5.667643760651349e-06,
+      "loss": 0.5656,
+      "step": 1266
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 5.652524564984319e-06,
+      "loss": 0.5496,
+      "step": 1267
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 5.637417613367202e-06,
+      "loss": 0.5697,
+      "step": 1268
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 5.622322948346595e-06,
+      "loss": 0.5679,
+      "step": 1269
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 5.607240612434486e-06,
+      "loss": 0.5795,
+      "step": 1270
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 5.5921706481081405e-06,
+      "loss": 0.5549,
+      "step": 1271
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 5.5771130978099896e-06,
+      "loss": 0.5523,
+      "step": 1272
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 5.562068003947484e-06,
+      "loss": 0.557,
+      "step": 1273
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 5.547035408893009e-06,
+      "loss": 0.5602,
+      "step": 1274
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 5.532015354983742e-06,
+      "loss": 0.5652,
+      "step": 1275
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 5.517007884521544e-06,
+      "loss": 0.5632,
+      "step": 1276
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 5.50201303977283e-06,
+      "loss": 0.5756,
+      "step": 1277
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 5.4870308629684675e-06,
+      "loss": 0.5667,
+      "step": 1278
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 5.47206139630363e-06,
+      "loss": 0.5565,
+      "step": 1279
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 5.457104681937706e-06,
+      "loss": 0.5785,
+      "step": 1280
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 5.442160761994171e-06,
+      "loss": 0.5601,
+      "step": 1281
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 5.4272296785604594e-06,
+      "loss": 0.5586,
+      "step": 1282
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 5.412311473687859e-06,
+      "loss": 0.5675,
+      "step": 1283
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 5.397406189391387e-06,
+      "loss": 0.5574,
+      "step": 1284
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 5.382513867649663e-06,
+      "loss": 0.5714,
+      "step": 1285
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 5.3676345504048075e-06,
+      "loss": 0.5605,
+      "step": 1286
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 5.352768279562315e-06,
+      "loss": 0.5707,
+      "step": 1287
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 5.337915096990935e-06,
+      "loss": 0.5657,
+      "step": 1288
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 5.323075044522559e-06,
+      "loss": 0.5713,
+      "step": 1289
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 5.308248163952097e-06,
+      "loss": 0.5704,
+      "step": 1290
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 5.293434497037355e-06,
+      "loss": 0.5637,
+      "step": 1291
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 5.278634085498937e-06,
+      "loss": 0.5729,
+      "step": 1292
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 5.263846971020108e-06,
+      "loss": 0.5536,
+      "step": 1293
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 5.249073195246685e-06,
+      "loss": 0.5708,
+      "step": 1294
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 5.234312799786921e-06,
+      "loss": 0.563,
+      "step": 1295
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 5.219565826211382e-06,
+      "loss": 0.574,
+      "step": 1296
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 5.204832316052837e-06,
+      "loss": 0.5751,
+      "step": 1297
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 5.190112310806126e-06,
+      "loss": 0.5425,
+      "step": 1298
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 5.175405851928068e-06,
+      "loss": 0.5625,
+      "step": 1299
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 5.160712980837326e-06,
+      "loss": 0.5509,
+      "step": 1300
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 5.146033738914291e-06,
+      "loss": 0.5638,
+      "step": 1301
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 5.131368167500977e-06,
+      "loss": 0.5765,
+      "step": 1302
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 5.116716307900893e-06,
+      "loss": 0.5651,
+      "step": 1303
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 5.1020782013789226e-06,
+      "loss": 0.5674,
+      "step": 1304
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 5.087453889161229e-06,
+      "loss": 0.5629,
+      "step": 1305
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 5.072843412435118e-06,
+      "loss": 0.5638,
+      "step": 1306
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 5.058246812348931e-06,
+      "loss": 0.5477,
+      "step": 1307
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 5.043664130011932e-06,
+      "loss": 0.5542,
+      "step": 1308
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 5.029095406494185e-06,
+      "loss": 0.5643,
+      "step": 1309
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 5.014540682826434e-06,
+      "loss": 0.5484,
+      "step": 1310
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 0.5606,
+      "step": 1311
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.985473398966669e-06,
+      "loss": 0.5738,
+      "step": 1312
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.970960920638552e-06,
+      "loss": 0.5613,
+      "step": 1313
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 4.956462605887994e-06,
+      "loss": 0.5504,
+      "step": 1314
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.9419784955474525e-06,
+      "loss": 0.5566,
+      "step": 1315
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.92750863040937e-06,
+      "loss": 0.5813,
+      "step": 1316
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.913053051226079e-06,
+      "loss": 0.556,
+      "step": 1317
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 4.8986117987096745e-06,
+      "loss": 0.5588,
+      "step": 1318
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.8841849135319015e-06,
+      "loss": 0.5659,
+      "step": 1319
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.869772436324045e-06,
+      "loss": 0.5644,
+      "step": 1320
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.8553744076768074e-06,
+      "loss": 0.5566,
+      "step": 1321
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 4.8409908681402e-06,
+      "loss": 0.5504,
+      "step": 1322
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.826621858223431e-06,
+      "loss": 0.5474,
+      "step": 1323
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.812267418394784e-06,
+      "loss": 0.5533,
+      "step": 1324
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 4.797927589081509e-06,
+      "loss": 0.5545,
+      "step": 1325
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.783602410669712e-06,
+      "loss": 0.5635,
+      "step": 1326
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.769291923504226e-06,
+      "loss": 0.5482,
+      "step": 1327
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.754996167888516e-06,
+      "loss": 0.5599,
+      "step": 1328
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 4.740715184084559e-06,
+      "loss": 0.556,
+      "step": 1329
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.726449012312726e-06,
+      "loss": 0.5645,
+      "step": 1330
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.712197692751673e-06,
+      "loss": 0.5834,
+      "step": 1331
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.697961265538231e-06,
+      "loss": 0.5523,
+      "step": 1332
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 4.683739770767276e-06,
+      "loss": 0.5707,
+      "step": 1333
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.669533248491641e-06,
+      "loss": 0.5432,
+      "step": 1334
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.655341738721989e-06,
+      "loss": 0.5521,
+      "step": 1335
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.641165281426699e-06,
+      "loss": 0.5538,
+      "step": 1336
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 4.627003916531761e-06,
+      "loss": 0.5627,
+      "step": 1337
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.612857683920658e-06,
+      "loss": 0.5653,
+      "step": 1338
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.598726623434249e-06,
+      "loss": 0.5607,
+      "step": 1339
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.584610774870671e-06,
+      "loss": 0.5562,
+      "step": 1340
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 4.570510177985213e-06,
+      "loss": 0.5587,
+      "step": 1341
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.556424872490218e-06,
+      "loss": 0.5453,
+      "step": 1342
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.542354898054953e-06,
+      "loss": 0.5593,
+      "step": 1343
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.528300294305519e-06,
+      "loss": 0.5623,
+      "step": 1344
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 4.5142611008247094e-06,
+      "loss": 0.5522,
+      "step": 1345
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.5002373571519344e-06,
+      "loss": 0.5656,
+      "step": 1346
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.486229102783084e-06,
+      "loss": 0.5394,
+      "step": 1347
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.472236377170426e-06,
+      "loss": 0.5685,
+      "step": 1348
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 4.458259219722495e-06,
+      "loss": 0.5747,
+      "step": 1349
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.444297669803981e-06,
+      "loss": 0.5563,
+      "step": 1350
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.430351766735609e-06,
+      "loss": 0.5605,
+      "step": 1351
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 4.416421549794047e-06,
+      "loss": 0.5686,
+      "step": 1352
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.40250705821178e-06,
+      "loss": 0.5527,
+      "step": 1353
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.388608331177009e-06,
+      "loss": 0.5445,
+      "step": 1354
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.374725407833532e-06,
+      "loss": 0.5532,
+      "step": 1355
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 4.360858327280641e-06,
+      "loss": 0.5659,
+      "step": 1356
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.34700712857301e-06,
+      "loss": 0.5676,
+      "step": 1357
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.333171850720575e-06,
+      "loss": 0.5617,
+      "step": 1358
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.319352532688444e-06,
+      "loss": 0.5593,
+      "step": 1359
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 4.3055492133967725e-06,
+      "loss": 0.556,
+      "step": 1360
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.2917619317206575e-06,
+      "loss": 0.5731,
+      "step": 1361
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.277990726490029e-06,
+      "loss": 0.5587,
+      "step": 1362
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.264235636489542e-06,
+      "loss": 0.5543,
+      "step": 1363
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 4.250496700458457e-06,
+      "loss": 0.5704,
+      "step": 1364
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.236773957090549e-06,
+      "loss": 0.5615,
+      "step": 1365
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.223067445033983e-06,
+      "loss": 0.5511,
+      "step": 1366
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.209377202891212e-06,
+      "loss": 0.5556,
+      "step": 1367
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 4.1957032692188685e-06,
+      "loss": 0.5641,
+      "step": 1368
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.182045682527658e-06,
+      "loss": 0.5443,
+      "step": 1369
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.168404481282233e-06,
+      "loss": 0.5557,
+      "step": 1370
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.154779703901114e-06,
+      "loss": 0.5588,
+      "step": 1371
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 4.141171388756558e-06,
+      "loss": 0.5532,
+      "step": 1372
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.127579574174464e-06,
+      "loss": 0.5671,
+      "step": 1373
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.1140042984342535e-06,
+      "loss": 0.5422,
+      "step": 1374
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.100445599768774e-06,
+      "loss": 0.5494,
+      "step": 1375
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 4.086903516364179e-06,
+      "loss": 0.549,
+      "step": 1376
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.073378086359834e-06,
+      "loss": 0.5479,
+      "step": 1377
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.059869347848201e-06,
+      "loss": 0.5696,
+      "step": 1378
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 4.046377338874731e-06,
+      "loss": 0.5637,
+      "step": 1379
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.03290209743776e-06,
+      "loss": 0.5602,
+      "step": 1380
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.019443661488402e-06,
+      "loss": 0.5587,
+      "step": 1381
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 4.006002068930437e-06,
+      "loss": 0.5713,
+      "step": 1382
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 3.99257735762021e-06,
+      "loss": 0.5368,
+      "step": 1383
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.979169565366524e-06,
+      "loss": 0.5607,
+      "step": 1384
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.9657787299305305e-06,
+      "loss": 0.5605,
+      "step": 1385
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.952404889025626e-06,
+      "loss": 0.5604,
+      "step": 1386
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.939048080317337e-06,
+      "loss": 0.5581,
+      "step": 1387
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.92570834142323e-06,
+      "loss": 0.5423,
+      "step": 1388
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.912385709912794e-06,
+      "loss": 0.5464,
+      "step": 1389
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.899080223307335e-06,
+      "loss": 0.553,
+      "step": 1390
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.885791919079878e-06,
+      "loss": 0.5635,
+      "step": 1391
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.8725208346550535e-06,
+      "loss": 0.5613,
+      "step": 1392
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.8592670074089915e-06,
+      "loss": 0.5566,
+      "step": 1393
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.8460304746692216e-06,
+      "loss": 0.5764,
+      "step": 1394
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.832811273714569e-06,
+      "loss": 0.5479,
+      "step": 1395
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.819609441775045e-06,
+      "loss": 0.5607,
+      "step": 1396
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.8064250160317428e-06,
+      "loss": 0.5594,
+      "step": 1397
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.7932580336167356e-06,
+      "loss": 0.5708,
+      "step": 1398
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.7801085316129615e-06,
+      "loss": 0.554,
+      "step": 1399
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.7669765470541363e-06,
+      "loss": 0.5612,
+      "step": 1400
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.75386211692464e-06,
+      "loss": 0.5555,
+      "step": 1401
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.7407652781594094e-06,
+      "loss": 0.5561,
+      "step": 1402
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.7276860676438398e-06,
+      "loss": 0.5765,
+      "step": 1403
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.714624522213681e-06,
+      "loss": 0.5605,
+      "step": 1404
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.7015806786549245e-06,
+      "loss": 0.5649,
+      "step": 1405
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.688554573703712e-06,
+      "loss": 0.5435,
+      "step": 1406
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.6755462440462288e-06,
+      "loss": 0.5574,
+      "step": 1407
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.662555726318595e-06,
+      "loss": 0.5639,
+      "step": 1408
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.649583057106767e-06,
+      "loss": 0.5461,
+      "step": 1409
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.636628272946434e-06,
+      "loss": 0.5519,
+      "step": 1410
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.6236914103229183e-06,
+      "loss": 0.5609,
+      "step": 1411
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.6107725056710563e-06,
+      "loss": 0.548,
+      "step": 1412
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.5978715953751207e-06,
+      "loss": 0.5621,
+      "step": 1413
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.5849887157687013e-06,
+      "loss": 0.5503,
+      "step": 1414
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.5721239031346067e-06,
+      "loss": 0.5558,
+      "step": 1415
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.5592771937047633e-06,
+      "loss": 0.5568,
+      "step": 1416
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.5464486236601147e-06,
+      "loss": 0.562,
+      "step": 1417
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.533638229130509e-06,
+      "loss": 0.55,
+      "step": 1418
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.5208460461946136e-06,
+      "loss": 0.5585,
+      "step": 1419
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.5080721108798044e-06,
+      "loss": 0.5565,
+      "step": 1420
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.4953164591620627e-06,
+      "loss": 0.5416,
+      "step": 1421
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.4825791269658782e-06,
+      "loss": 0.5705,
+      "step": 1422
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.4698601501641517e-06,
+      "loss": 0.547,
+      "step": 1423
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.457159564578074e-06,
+      "loss": 0.56,
+      "step": 1424
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.4444774059770537e-06,
+      "loss": 0.5699,
+      "step": 1425
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.4318137100785965e-06,
+      "loss": 0.5497,
+      "step": 1426
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.419168512548212e-06,
+      "loss": 0.5601,
+      "step": 1427
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.4065418489993118e-06,
+      "loss": 0.5618,
+      "step": 1428
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.3939337549931106e-06,
+      "loss": 0.5513,
+      "step": 1429
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.3813442660385175e-06,
+      "loss": 0.5665,
+      "step": 1430
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.3687734175920505e-06,
+      "loss": 0.5527,
+      "step": 1431
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.356221245057727e-06,
+      "loss": 0.5596,
+      "step": 1432
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.343687783786966e-06,
+      "loss": 0.5453,
+      "step": 1433
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.3311730690784916e-06,
+      "loss": 0.5418,
+      "step": 1434
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.3186771361782278e-06,
+      "loss": 0.5554,
+      "step": 1435
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.3062000202792e-06,
+      "loss": 0.5556,
+      "step": 1436
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.2937417565214426e-06,
+      "loss": 0.5554,
+      "step": 1437
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.281302379991891e-06,
+      "loss": 0.5536,
+      "step": 1438
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.2688819257242963e-06,
+      "loss": 0.5521,
+      "step": 1439
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.2564804286991137e-06,
+      "loss": 0.5543,
+      "step": 1440
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.2440979238433977e-06,
+      "loss": 0.5642,
+      "step": 1441
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.2317344460307275e-06,
+      "loss": 0.5522,
+      "step": 1442
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.2193900300810908e-06,
+      "loss": 0.5585,
+      "step": 1443
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.207064710760788e-06,
+      "loss": 0.5471,
+      "step": 1444
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.1947585227823395e-06,
+      "loss": 0.5346,
+      "step": 1445
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.1824715008043874e-06,
+      "loss": 0.5464,
+      "step": 1446
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.1702036794315837e-06,
+      "loss": 0.5472,
+      "step": 1447
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.157955093214519e-06,
+      "loss": 0.5579,
+      "step": 1448
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.145725776649602e-06,
+      "loss": 0.567,
+      "step": 1449
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.133515764178976e-06,
+      "loss": 0.5539,
+      "step": 1450
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.1213250901904124e-06,
+      "loss": 0.5695,
+      "step": 1451
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.1091537890172263e-06,
+      "loss": 0.5523,
+      "step": 1452
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.0970018949381598e-06,
+      "loss": 0.5687,
+      "step": 1453
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.0848694421773075e-06,
+      "loss": 0.5519,
+      "step": 1454
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.0727564649040066e-06,
+      "loss": 0.5631,
+      "step": 1455
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.060662997232746e-06,
+      "loss": 0.538,
+      "step": 1456
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.0485890732230674e-06,
+      "loss": 0.5587,
+      "step": 1457
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.036534726879473e-06,
+      "loss": 0.5549,
+      "step": 1458
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.0244999921513207e-06,
+      "loss": 0.5602,
+      "step": 1459
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.012484902932741e-06,
+      "loss": 0.5472,
+      "step": 1460
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.000489493062535e-06,
+      "loss": 0.5455,
+      "step": 1461
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.9885137963240773e-06,
+      "loss": 0.5499,
+      "step": 1462
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.976557846445225e-06,
+      "loss": 0.5572,
+      "step": 1463
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.964621677098225e-06,
+      "loss": 0.5457,
+      "step": 1464
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.952705321899604e-06,
+      "loss": 0.5484,
+      "step": 1465
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.9408088144100934e-06,
+      "loss": 0.547,
+      "step": 1466
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 0.5667,
+      "step": 1467
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.91707547652174e-06,
+      "loss": 0.5577,
+      "step": 1468
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.9052387129644876e-06,
+      "loss": 0.5526,
+      "step": 1469
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.8934219307993393e-06,
+      "loss": 0.5371,
+      "step": 1470
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.8816251633065963e-06,
+      "loss": 0.5438,
+      "step": 1471
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.869848443710178e-06,
+      "loss": 0.5435,
+      "step": 1472
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.8580918051775542e-06,
+      "loss": 0.5519,
+      "step": 1473
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.846355280819636e-06,
+      "loss": 0.5632,
+      "step": 1474
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.834638903690683e-06,
+      "loss": 0.5571,
+      "step": 1475
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.8229427067882165e-06,
+      "loss": 0.553,
+      "step": 1476
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.8112667230529245e-06,
+      "loss": 0.5526,
+      "step": 1477
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.799610985368557e-06,
+      "loss": 0.5562,
+      "step": 1478
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.7879755265618558e-06,
+      "loss": 0.5403,
+      "step": 1479
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.776360379402445e-06,
+      "loss": 0.5466,
+      "step": 1480
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.7647655766027447e-06,
+      "loss": 0.5516,
+      "step": 1481
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.7531911508178764e-06,
+      "loss": 0.5494,
+      "step": 1482
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.7416371346455792e-06,
+      "loss": 0.5624,
+      "step": 1483
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.7301035606260985e-06,
+      "loss": 0.5602,
+      "step": 1484
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.7185904612421177e-06,
+      "loss": 0.5503,
+      "step": 1485
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.7070978689186544e-06,
+      "loss": 0.5443,
+      "step": 1486
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.69562581602297e-06,
+      "loss": 0.5602,
+      "step": 1487
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.6841743348644788e-06,
+      "loss": 0.5648,
+      "step": 1488
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 2.672743457694662e-06,
+      "loss": 0.5524,
+      "step": 1489
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 2.661333216706964e-06,
+      "loss": 0.5613,
+      "step": 1490
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 2.6499436440367165e-06,
+      "loss": 0.5505,
+      "step": 1491
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.6385747717610423e-06,
+      "loss": 0.5464,
+      "step": 1492
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.6272266318987606e-06,
+      "loss": 0.5479,
+      "step": 1493
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.615899256410306e-06,
+      "loss": 0.5594,
+      "step": 1494
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.6045926771976306e-06,
+      "loss": 0.5461,
+      "step": 1495
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 2.593306926104111e-06,
+      "loss": 0.5303,
+      "step": 1496
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 2.5820420349144693e-06,
+      "loss": 0.5653,
+      "step": 1497
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 2.570798035354686e-06,
+      "loss": 0.554,
+      "step": 1498
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 2.559574959091893e-06,
+      "loss": 0.5559,
+      "step": 1499
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 2.548372837734299e-06,
+      "loss": 0.5569,
+      "step": 1500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 2.537191702831091e-06,
+      "loss": 0.5481,
+      "step": 1501
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 2.5260315858723573e-06,
+      "loss": 0.548,
+      "step": 1502
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 2.514892518288988e-06,
+      "loss": 0.5532,
+      "step": 1503
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 2.5037745314525952e-06,
+      "loss": 0.5549,
+      "step": 1504
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 2.492677656675414e-06,
+      "loss": 0.5471,
+      "step": 1505
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 2.4816019252102274e-06,
+      "loss": 0.5432,
+      "step": 1506
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 2.470547368250261e-06,
+      "loss": 0.5626,
+      "step": 1507
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 2.4595140169291155e-06,
+      "loss": 0.5667,
+      "step": 1508
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 2.4485019023206637e-06,
+      "loss": 0.5737,
+      "step": 1509
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 2.43751105543897e-06,
+      "loss": 0.5595,
+      "step": 1510
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 2.4265415072382016e-06,
+      "loss": 0.5695,
+      "step": 1511
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 2.415593288612541e-06,
+      "loss": 0.5405,
+      "step": 1512
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 2.4046664303960942e-06,
+      "loss": 0.5453,
+      "step": 1513
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 2.3937609633628145e-06,
+      "loss": 0.557,
+      "step": 1514
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 2.382876918226409e-06,
+      "loss": 0.5598,
+      "step": 1515
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 2.3720143256402507e-06,
+      "loss": 0.5461,
+      "step": 1516
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 2.3611732161972966e-06,
+      "loss": 0.5566,
+      "step": 1517
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 2.3503536204300017e-06,
+      "loss": 0.554,
+      "step": 1518
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 2.339555568810221e-06,
+      "loss": 0.5629,
+      "step": 1519
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 2.328779091749145e-06,
+      "loss": 0.5521,
+      "step": 1520
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 2.318024219597196e-06,
+      "loss": 0.5625,
+      "step": 1521
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 2.3072909826439517e-06,
+      "loss": 0.5385,
+      "step": 1522
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 2.296579411118055e-06,
+      "loss": 0.5505,
+      "step": 1523
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 2.285889535187139e-06,
+      "loss": 0.5721,
+      "step": 1524
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 2.275221384957719e-06,
+      "loss": 0.5541,
+      "step": 1525
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 2.264574990475136e-06,
+      "loss": 0.5518,
+      "step": 1526
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 2.2539503817234553e-06,
+      "loss": 0.5354,
+      "step": 1527
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 2.2433475886253843e-06,
+      "loss": 0.5467,
+      "step": 1528
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 2.232766641042191e-06,
+      "loss": 0.5505,
+      "step": 1529
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 2.222207568773619e-06,
+      "loss": 0.5497,
+      "step": 1530
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 2.211670401557804e-06,
+      "loss": 0.5464,
+      "step": 1531
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 2.201155169071184e-06,
+      "loss": 0.5475,
+      "step": 1532
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 2.190661900928426e-06,
+      "loss": 0.5468,
+      "step": 1533
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 2.180190626682336e-06,
+      "loss": 0.5467,
+      "step": 1534
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.1697413758237785e-06,
+      "loss": 0.5411,
+      "step": 1535
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.1593141777815906e-06,
+      "loss": 0.5361,
+      "step": 1536
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.148909061922504e-06,
+      "loss": 0.5499,
+      "step": 1537
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.138526057551051e-06,
+      "loss": 0.5605,
+      "step": 1538
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 2.1281651939094996e-06,
+      "loss": 0.5444,
+      "step": 1539
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 2.1178265001777552e-06,
+      "loss": 0.5533,
+      "step": 1540
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 2.107510005473289e-06,
+      "loss": 0.5569,
+      "step": 1541
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 2.0972157388510484e-06,
+      "loss": 0.5489,
+      "step": 1542
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.0869437293033835e-06,
+      "loss": 0.5604,
+      "step": 1543
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.0766940057599505e-06,
+      "loss": 0.4724,
+      "step": 1544
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.0664665970876496e-06,
+      "loss": 0.4131,
+      "step": 1545
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 2.0562615320905333e-06,
+      "loss": 0.4205,
+      "step": 1546
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 2.046078839509721e-06,
+      "loss": 0.4119,
+      "step": 1547
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 2.03591854802333e-06,
+      "loss": 0.3964,
+      "step": 1548
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 2.0257806862463847e-06,
+      "loss": 0.3901,
+      "step": 1549
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 2.0156652827307367e-06,
+      "loss": 0.3777,
+      "step": 1550
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 2.0055723659649907e-06,
+      "loss": 0.3714,
+      "step": 1551
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 1.99550196437442e-06,
+      "loss": 0.3718,
+      "step": 1552
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 1.9854541063208875e-06,
+      "loss": 0.3803,
+      "step": 1553
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.975428820102766e-06,
+      "loss": 0.381,
+      "step": 1554
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.965426133954854e-06,
+      "loss": 0.3828,
+      "step": 1555
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.9554460760483052e-06,
+      "loss": 0.3839,
+      "step": 1556
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.94548867449054e-06,
+      "loss": 0.3727,
+      "step": 1557
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.9355539573251737e-06,
+      "loss": 0.3817,
+      "step": 1558
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.9256419525319316e-06,
+      "loss": 0.3823,
+      "step": 1559
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.9157526880265763e-06,
+      "loss": 0.3715,
+      "step": 1560
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.9058861916608185e-06,
+      "loss": 0.3823,
+      "step": 1561
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.896042491222253e-06,
+      "loss": 0.3669,
+      "step": 1562
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.8862216144342692e-06,
+      "loss": 0.385,
+      "step": 1563
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.8764235889559802e-06,
+      "loss": 0.3585,
+      "step": 1564
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.8666484423821374e-06,
+      "loss": 0.365,
+      "step": 1565
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.8568962022430637e-06,
+      "loss": 0.3656,
+      "step": 1566
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.8471668960045575e-06,
+      "loss": 0.3675,
+      "step": 1567
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.8374605510678377e-06,
+      "loss": 0.384,
+      "step": 1568
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.8277771947694523e-06,
+      "loss": 0.3721,
+      "step": 1569
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.8181168543812067e-06,
+      "loss": 0.3692,
+      "step": 1570
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.808479557110081e-06,
+      "loss": 0.3703,
+      "step": 1571
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.7988653300981663e-06,
+      "loss": 0.3751,
+      "step": 1572
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.789274200422566e-06,
+      "loss": 0.3716,
+      "step": 1573
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.7797061950953442e-06,
+      "loss": 0.3541,
+      "step": 1574
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.7701613410634367e-06,
+      "loss": 0.3827,
+      "step": 1575
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.7606396652085733e-06,
+      "loss": 0.3776,
+      "step": 1576
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.751141194347208e-06,
+      "loss": 0.368,
+      "step": 1577
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.7416659552304449e-06,
+      "loss": 0.3532,
+      "step": 1578
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.7322139745439493e-06,
+      "loss": 0.3649,
+      "step": 1579
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.7227852789078914e-06,
+      "loss": 0.3591,
+      "step": 1580
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.7133798948768598e-06,
+      "loss": 0.3602,
+      "step": 1581
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.7039978489397879e-06,
+      "loss": 0.377,
+      "step": 1582
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.6946391675198838e-06,
+      "loss": 0.3708,
+      "step": 1583
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.6853038769745466e-06,
+      "loss": 0.3815,
+      "step": 1584
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.6759920035953092e-06,
+      "loss": 0.3723,
+      "step": 1585
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.6667035736077385e-06,
+      "loss": 0.3752,
+      "step": 1586
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.6574386131713872e-06,
+      "loss": 0.3618,
+      "step": 1587
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.6481971483797077e-06,
+      "loss": 0.3783,
+      "step": 1588
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.6389792052599796e-06,
+      "loss": 0.3604,
+      "step": 1589
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.6297848097732338e-06,
+      "loss": 0.3728,
+      "step": 1590
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.620613987814189e-06,
+      "loss": 0.3566,
+      "step": 1591
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.6114667652111649e-06,
+      "loss": 0.3655,
+      "step": 1592
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.6023431677260215e-06,
+      "loss": 0.3662,
+      "step": 1593
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.5932432210540827e-06,
+      "loss": 0.3599,
+      "step": 1594
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.584166950824061e-06,
+      "loss": 0.3765,
+      "step": 1595
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.5751143825979875e-06,
+      "loss": 0.3722,
+      "step": 1596
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.566085541871145e-06,
+      "loss": 0.3691,
+      "step": 1597
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.5570804540719808e-06,
+      "loss": 0.3581,
+      "step": 1598
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.5480991445620541e-06,
+      "loss": 0.3584,
+      "step": 1599
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.5391416386359526e-06,
+      "loss": 0.3752,
+      "step": 1600
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.5302079615212274e-06,
+      "loss": 0.3685,
+      "step": 1601
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.5212981383783154e-06,
+      "loss": 0.3612,
+      "step": 1602
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.5124121943004766e-06,
+      "loss": 0.3552,
+      "step": 1603
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.503550154313711e-06,
+      "loss": 0.3631,
+      "step": 1604
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.4947120433767048e-06,
+      "loss": 0.3561,
+      "step": 1605
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.4858978863807472e-06,
+      "loss": 0.3702,
+      "step": 1606
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.4771077081496654e-06,
+      "loss": 0.3569,
+      "step": 1607
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.4683415334397532e-06,
+      "loss": 0.36,
+      "step": 1608
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.4595993869397052e-06,
+      "loss": 0.367,
+      "step": 1609
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.4508812932705364e-06,
+      "loss": 0.3587,
+      "step": 1610
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.4421872769855262e-06,
+      "loss": 0.3698,
+      "step": 1611
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.4335173625701438e-06,
+      "loss": 0.3778,
+      "step": 1612
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.4248715744419773e-06,
+      "loss": 0.3574,
+      "step": 1613
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.4162499369506655e-06,
+      "loss": 0.3645,
+      "step": 1614
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.407652474377832e-06,
+      "loss": 0.3537,
+      "step": 1615
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.3990792109370154e-06,
+      "loss": 0.3556,
+      "step": 1616
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.3905301707735986e-06,
+      "loss": 0.3585,
+      "step": 1617
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.3820053779647457e-06,
+      "loss": 0.3637,
+      "step": 1618
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.373504856519331e-06,
+      "loss": 0.3498,
+      "step": 1619
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.3650286303778715e-06,
+      "loss": 0.3735,
+      "step": 1620
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.3565767234124572e-06,
+      "loss": 0.368,
+      "step": 1621
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.3481491594266905e-06,
+      "loss": 0.3676,
+      "step": 1622
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.339745962155613e-06,
+      "loss": 0.369,
+      "step": 1623
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.3313671552656438e-06,
+      "loss": 0.3587,
+      "step": 1624
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.3230127623545064e-06,
+      "loss": 0.3596,
+      "step": 1625
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.3146828069511708e-06,
+      "loss": 0.3608,
+      "step": 1626
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.3063773125157719e-06,
+      "loss": 0.3728,
+      "step": 1627
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.2980963024395655e-06,
+      "loss": 0.3532,
+      "step": 1628
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.2898398000448441e-06,
+      "loss": 0.3651,
+      "step": 1629
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.2816078285848798e-06,
+      "loss": 0.3606,
+      "step": 1630
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.273400411243857e-06,
+      "loss": 0.3694,
+      "step": 1631
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.2652175711368075e-06,
+      "loss": 0.3732,
+      "step": 1632
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.2570593313095393e-06,
+      "loss": 0.3647,
+      "step": 1633
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.2489257147385836e-06,
+      "loss": 0.3632,
+      "step": 1634
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2408167443311215e-06,
+      "loss": 0.3628,
+      "step": 1635
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2327324429249232e-06,
+      "loss": 0.3674,
+      "step": 1636
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2246728332882785e-06,
+      "loss": 0.3569,
+      "step": 1637
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2166379381199423e-06,
+      "loss": 0.3657,
+      "step": 1638
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.2086277800490554e-06,
+      "loss": 0.3619,
+      "step": 1639
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.2006423816350976e-06,
+      "loss": 0.3903,
+      "step": 1640
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.1926817653678159e-06,
+      "loss": 0.3689,
+      "step": 1641
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.1847459536671591e-06,
+      "loss": 0.3594,
+      "step": 1642
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.1768349688832204e-06,
+      "loss": 0.3675,
+      "step": 1643
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.1689488332961684e-06,
+      "loss": 0.3649,
+      "step": 1644
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.1610875691161916e-06,
+      "loss": 0.3847,
+      "step": 1645
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.1532511984834248e-06,
+      "loss": 0.3642,
+      "step": 1646
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.1454397434679022e-06,
+      "loss": 0.3798,
+      "step": 1647
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.1376532260694794e-06,
+      "loss": 0.3574,
+      "step": 1648
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.129891668217783e-06,
+      "loss": 0.3564,
+      "step": 1649
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.1221550917721434e-06,
+      "loss": 0.357,
+      "step": 1650
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.1144435185215374e-06,
+      "loss": 0.362,
+      "step": 1651
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.1067569701845148e-06,
+      "loss": 0.3636,
+      "step": 1652
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.099095468409156e-06,
+      "loss": 0.3556,
+      "step": 1653
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.0914590347729958e-06,
+      "loss": 0.3557,
+      "step": 1654
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.083847690782972e-06,
+      "loss": 0.3522,
+      "step": 1655
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.0762614578753571e-06,
+      "loss": 0.3594,
+      "step": 1656
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.0687003574157074e-06,
+      "loss": 0.3875,
+      "step": 1657
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.0611644106987863e-06,
+      "loss": 0.349,
+      "step": 1658
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.0536536389485275e-06,
+      "loss": 0.3615,
+      "step": 1659
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.046168063317956e-06,
+      "loss": 0.3562,
+      "step": 1660
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.038707704889138e-06,
+      "loss": 0.362,
+      "step": 1661
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.0312725846731174e-06,
+      "loss": 0.3494,
+      "step": 1662
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.0238627236098619e-06,
+      "loss": 0.3649,
+      "step": 1663
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.0164781425681925e-06,
+      "loss": 0.3696,
+      "step": 1664
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.0091188623457415e-06,
+      "loss": 0.3557,
+      "step": 1665
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 1.0017849036688797e-06,
+      "loss": 0.3616,
+      "step": 1666
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 9.944762871926661e-07,
+      "loss": 0.3603,
+      "step": 1667
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 9.87193033500784e-07,
+      "loss": 0.3618,
+      "step": 1668
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 9.79935163105491e-07,
+      "loss": 0.3721,
+      "step": 1669
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 9.727026964475483e-07,
+      "loss": 0.3647,
+      "step": 1670
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 9.65495653896179e-07,
+      "loss": 0.368,
+      "step": 1671
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 9.583140557489955e-07,
+      "loss": 0.3738,
+      "step": 1672
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 9.51157922231959e-07,
+      "loss": 0.3652,
+      "step": 1673
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 9.440272734993072e-07,
+      "loss": 0.3534,
+      "step": 1674
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 9.369221296335007e-07,
+      "loss": 0.3571,
+      "step": 1675
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 9.298425106451747e-07,
+      "loss": 0.3522,
+      "step": 1676
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 9.227884364730743e-07,
+      "loss": 0.3661,
+      "step": 1677
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 9.157599269840023e-07,
+      "loss": 0.3655,
+      "step": 1678
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 9.08757001972762e-07,
+      "loss": 0.3799,
+      "step": 1679
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 9.017796811621049e-07,
+      "loss": 0.367,
+      "step": 1680
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 8.948279842026631e-07,
+      "loss": 0.3702,
+      "step": 1681
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 8.879019306729108e-07,
+      "loss": 0.3755,
+      "step": 1682
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 8.810015400790994e-07,
+      "loss": 0.3794,
+      "step": 1683
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 8.741268318552021e-07,
+      "loss": 0.3478,
+      "step": 1684
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 8.672778253628622e-07,
+      "loss": 0.3705,
+      "step": 1685
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 8.604545398913422e-07,
+      "loss": 0.3629,
+      "step": 1686
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 8.536569946574546e-07,
+      "loss": 0.3606,
+      "step": 1687
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 8.468852088055291e-07,
+      "loss": 0.3694,
+      "step": 1688
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 8.401392014073406e-07,
+      "loss": 0.3632,
+      "step": 1689
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 8.334189914620672e-07,
+      "loss": 0.3727,
+      "step": 1690
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 8.267245978962291e-07,
+      "loss": 0.3655,
+      "step": 1691
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 8.200560395636414e-07,
+      "loss": 0.3705,
+      "step": 1692
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 8.134133352453522e-07,
+      "loss": 0.3602,
+      "step": 1693
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 8.067965036496017e-07,
+      "loss": 0.3615,
+      "step": 1694
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 8.002055634117578e-07,
+      "loss": 0.3633,
+      "step": 1695
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 7.936405330942754e-07,
+      "loss": 0.3638,
+      "step": 1696
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 7.871014311866321e-07,
+      "loss": 0.3634,
+      "step": 1697
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 7.805882761052852e-07,
+      "loss": 0.3572,
+      "step": 1698
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 7.741010861936138e-07,
+      "loss": 0.36,
+      "step": 1699
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 7.676398797218721e-07,
+      "loss": 0.368,
+      "step": 1700
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 7.612046748871327e-07,
+      "loss": 0.3611,
+      "step": 1701
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 7.547954898132426e-07,
+      "loss": 0.3581,
+      "step": 1702
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 7.48412342550765e-07,
+      "loss": 0.3642,
+      "step": 1703
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 7.420552510769308e-07,
+      "loss": 0.3501,
+      "step": 1704
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 7.357242332955917e-07,
+      "loss": 0.3651,
+      "step": 1705
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 7.294193070371603e-07,
+      "loss": 0.3773,
+      "step": 1706
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 7.231404900585714e-07,
+      "loss": 0.3643,
+      "step": 1707
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 7.168878000432278e-07,
+      "loss": 0.3577,
+      "step": 1708
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 7.106612546009439e-07,
+      "loss": 0.3647,
+      "step": 1709
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 7.044608712679058e-07,
+      "loss": 0.3539,
+      "step": 1710
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 6.98286667506618e-07,
+      "loss": 0.3539,
+      "step": 1711
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 6.9213866070585e-07,
+      "loss": 0.3448,
+      "step": 1712
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 6.860168681805946e-07,
+      "loss": 0.3711,
+      "step": 1713
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 6.799213071720156e-07,
+      "loss": 0.3629,
+      "step": 1714
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 6.738519948473976e-07,
+      "loss": 0.3496,
+      "step": 1715
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 6.678089483001016e-07,
+      "loss": 0.3763,
+      "step": 1716
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 6.617921845495157e-07,
+      "loss": 0.3672,
+      "step": 1717
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 6.558017205410005e-07,
+      "loss": 0.3604,
+      "step": 1718
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 6.498375731458529e-07,
+      "loss": 0.3542,
+      "step": 1719
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 6.438997591612506e-07,
+      "loss": 0.3654,
+      "step": 1720
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 6.379882953102068e-07,
+      "loss": 0.3621,
+      "step": 1721
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 6.321031982415237e-07,
+      "loss": 0.3684,
+      "step": 1722
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 6.262444845297488e-07,
+      "loss": 0.3606,
+      "step": 1723
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 6.204121706751153e-07,
+      "loss": 0.3871,
+      "step": 1724
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 6.146062731035129e-07,
+      "loss": 0.3611,
+      "step": 1725
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 6.088268081664316e-07,
+      "loss": 0.3325,
+      "step": 1726
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 6.030737921409169e-07,
+      "loss": 0.3639,
+      "step": 1727
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 5.973472412295256e-07,
+      "loss": 0.3767,
+      "step": 1728
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 5.916471715602789e-07,
+      "loss": 0.3828,
+      "step": 1729
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 5.859735991866166e-07,
+      "loss": 0.3531,
+      "step": 1730
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 5.803265400873514e-07,
+      "loss": 0.359,
+      "step": 1731
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 5.74706010166629e-07,
+      "loss": 0.3608,
+      "step": 1732
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 5.691120252538784e-07,
+      "loss": 0.3686,
+      "step": 1733
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 5.63544601103767e-07,
+      "loss": 0.3839,
+      "step": 1734
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 5.580037533961546e-07,
+      "loss": 0.3715,
+      "step": 1735
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 5.524894977360574e-07,
+      "loss": 0.3767,
+      "step": 1736
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 5.470018496535967e-07,
+      "loss": 0.3549,
+      "step": 1737
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 5.415408246039588e-07,
+      "loss": 0.3532,
+      "step": 1738
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 5.361064379673465e-07,
+      "loss": 0.3624,
+      "step": 1739
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 5.306987050489442e-07,
+      "loss": 0.3634,
+      "step": 1740
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 5.253176410788641e-07,
+      "loss": 0.3496,
+      "step": 1741
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 5.199632612121131e-07,
+      "loss": 0.3653,
+      "step": 1742
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 5.146355805285452e-07,
+      "loss": 0.3687,
+      "step": 1743
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 5.093346140328193e-07,
+      "loss": 0.3574,
+      "step": 1744
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 5.040603766543594e-07,
+      "loss": 0.3527,
+      "step": 1745
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 4.988128832473105e-07,
+      "loss": 0.361,
+      "step": 1746
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.935921485904926e-07,
+      "loss": 0.3527,
+      "step": 1747
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.883981873873678e-07,
+      "loss": 0.3572,
+      "step": 1748
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.832310142659946e-07,
+      "loss": 0.3609,
+      "step": 1749
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 4.780906437789845e-07,
+      "loss": 0.3499,
+      "step": 1750
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.7297709040346474e-07,
+      "loss": 0.3551,
+      "step": 1751
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.678903685410374e-07,
+      "loss": 0.3636,
+      "step": 1752
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.628304925177318e-07,
+      "loss": 0.3557,
+      "step": 1753
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 4.5779747658397436e-07,
+      "loss": 0.3604,
+      "step": 1754
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.5279133491454406e-07,
+      "loss": 0.349,
+      "step": 1755
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.478120816085307e-07,
+      "loss": 0.3649,
+      "step": 1756
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.428597306892968e-07,
+      "loss": 0.3533,
+      "step": 1757
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 4.3793429610444083e-07,
+      "loss": 0.3694,
+      "step": 1758
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.3303579172574884e-07,
+      "loss": 0.3602,
+      "step": 1759
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.281642313491685e-07,
+      "loss": 0.3645,
+      "step": 1760
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.233196286947605e-07,
+      "loss": 0.3658,
+      "step": 1761
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 4.185019974066629e-07,
+      "loss": 0.3658,
+      "step": 1762
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.137113510530544e-07,
+      "loss": 0.3666,
+      "step": 1763
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.089477031261113e-07,
+      "loss": 0.3649,
+      "step": 1764
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 4.042110670419763e-07,
+      "loss": 0.3552,
+      "step": 1765
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.99501456140714e-07,
+      "loss": 0.3578,
+      "step": 1766
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.9481888368627764e-07,
+      "loss": 0.3724,
+      "step": 1767
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.9016336286647025e-07,
+      "loss": 0.3575,
+      "step": 1768
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.855349067929093e-07,
+      "loss": 0.3679,
+      "step": 1769
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.809335285009863e-07,
+      "loss": 0.3692,
+      "step": 1770
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.76359240949834e-07,
+      "loss": 0.3773,
+      "step": 1771
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.7181205702228384e-07,
+      "loss": 0.3758,
+      "step": 1772
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.6729198952483725e-07,
+      "loss": 0.3508,
+      "step": 1773
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.627990511876267e-07,
+      "loss": 0.366,
+      "step": 1774
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.5833325466437697e-07,
+      "loss": 0.3624,
+      "step": 1775
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.538946125323717e-07,
+      "loss": 0.3483,
+      "step": 1776
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.494831372924212e-07,
+      "loss": 0.3709,
+      "step": 1777
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 3.4509884136882056e-07,
+      "loss": 0.3553,
+      "step": 1778
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 3.4074173710931804e-07,
+      "loss": 0.3592,
+      "step": 1779
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 3.3641183678508326e-07,
+      "loss": 0.3602,
+      "step": 1780
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 3.3210915259066834e-07,
+      "loss": 0.3554,
+      "step": 1781
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 3.278336966439744e-07,
+      "loss": 0.3538,
+      "step": 1782
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 3.235854809862193e-07,
+      "loss": 0.3443,
+      "step": 1783
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 3.1936451758190157e-07,
+      "loss": 0.3655,
+      "step": 1784
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 3.151708183187674e-07,
+      "loss": 0.3625,
+      "step": 1785
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 3.110043950077779e-07,
+      "loss": 0.3547,
+      "step": 1786
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 3.0686525938307563e-07,
+      "loss": 0.3695,
+      "step": 1787
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 3.0275342310195133e-07,
+      "loss": 0.352,
+      "step": 1788
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 2.9866889774481045e-07,
+      "loss": 0.369,
+      "step": 1789
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.9461169481514006e-07,
+      "loss": 0.358,
+      "step": 1790
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.905818257394799e-07,
+      "loss": 0.3579,
+      "step": 1791
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.865793018673857e-07,
+      "loss": 0.354,
+      "step": 1792
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.826041344714003e-07,
+      "loss": 0.3597,
+      "step": 1793
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.786563347470217e-07,
+      "loss": 0.3685,
+      "step": 1794
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.7473591381266705e-07,
+      "loss": 0.3646,
+      "step": 1795
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.7084288270964985e-07,
+      "loss": 0.3732,
+      "step": 1796
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.6697725240214076e-07,
+      "loss": 0.3621,
+      "step": 1797
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.6313903377714225e-07,
+      "loss": 0.3449,
+      "step": 1798
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.593282376444539e-07,
+      "loss": 0.3671,
+      "step": 1799
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.5554487473664404e-07,
+      "loss": 0.359,
+      "step": 1800
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.5178895570901806e-07,
+      "loss": 0.3511,
+      "step": 1801
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.480604911395934e-07,
+      "loss": 0.3585,
+      "step": 1802
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.4435949152906144e-07,
+      "loss": 0.3532,
+      "step": 1803
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.406859673007655e-07,
+      "loss": 0.3566,
+      "step": 1804
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.370399288006664e-07,
+      "loss": 0.3464,
+      "step": 1805
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.334213862973167e-07,
+      "loss": 0.3609,
+      "step": 1806
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.2983034998182997e-07,
+      "loss": 0.3601,
+      "step": 1807
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.2626682996785054e-07,
+      "loss": 0.353,
+      "step": 1808
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.2273083629153148e-07,
+      "loss": 0.365,
+      "step": 1809
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.1922237891149668e-07,
+      "loss": 0.3731,
+      "step": 1810
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.15741467708821e-07,
+      "loss": 0.3605,
+      "step": 1811
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.122881124870002e-07,
+      "loss": 0.3429,
+      "step": 1812
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.088623229719189e-07,
+      "loss": 0.3479,
+      "step": 1813
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.0546410881183032e-07,
+      "loss": 0.3385,
+      "step": 1814
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.0209347957732328e-07,
+      "loss": 0.3592,
+      "step": 1815
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 1.9875044476129868e-07,
+      "loss": 0.3629,
+      "step": 1816
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.9543501377894182e-07,
+      "loss": 0.3535,
+      "step": 1817
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.921471959676957e-07,
+      "loss": 0.3503,
+      "step": 1818
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.8888700058723343e-07,
+      "loss": 0.3679,
+      "step": 1819
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.8565443681943462e-07,
+      "loss": 0.3687,
+      "step": 1820
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.8244951376835907e-07,
+      "loss": 0.3521,
+      "step": 1821
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.7927224046021763e-07,
+      "loss": 0.3639,
+      "step": 1822
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.761226258433524e-07,
+      "loss": 0.3673,
+      "step": 1823
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.730006787882077e-07,
+      "loss": 0.3664,
+      "step": 1824
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.6990640808730695e-07,
+      "loss": 0.3533,
+      "step": 1825
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.668398224552237e-07,
+      "loss": 0.3464,
+      "step": 1826
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.6380093052856482e-07,
+      "loss": 0.3482,
+      "step": 1827
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.6078974086593868e-07,
+      "loss": 0.3808,
+      "step": 1828
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.578062619479359e-07,
+      "loss": 0.3624,
+      "step": 1829
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.5485050217710295e-07,
+      "loss": 0.3811,
+      "step": 1830
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.519224698779198e-07,
+      "loss": 0.3566,
+      "step": 1831
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.490221732967745e-07,
+      "loss": 0.3598,
+      "step": 1832
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.4614962060194303e-07,
+      "loss": 0.3555,
+      "step": 1833
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.4330481988356294e-07,
+      "loss": 0.3502,
+      "step": 1834
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.404877791536141e-07,
+      "loss": 0.3551,
+      "step": 1835
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.3769850634589356e-07,
+      "loss": 0.3701,
+      "step": 1836
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.349370093159952e-07,
+      "loss": 0.3641,
+      "step": 1837
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.322032958412811e-07,
+      "loss": 0.3627,
+      "step": 1838
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.2949737362087156e-07,
+      "loss": 0.3599,
+      "step": 1839
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.2681925027561383e-07,
+      "loss": 0.3605,
+      "step": 1840
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.2416893334806333e-07,
+      "loss": 0.3643,
+      "step": 1841
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.2154643030246273e-07,
+      "loss": 0.3641,
+      "step": 1842
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.1895174852472158e-07,
+      "loss": 0.3577,
+      "step": 1843
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.1638489532239339e-07,
+      "loss": 0.3479,
+      "step": 1844
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.1384587792465873e-07,
+      "loss": 0.3519,
+      "step": 1845
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.1133470348229869e-07,
+      "loss": 0.3455,
+      "step": 1846
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.0885137906768373e-07,
+      "loss": 0.3605,
+      "step": 1847
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.0639591167474483e-07,
+      "loss": 0.3692,
+      "step": 1848
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.0396830821895909e-07,
+      "loss": 0.3542,
+      "step": 1849
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.0156857553732858e-07,
+      "loss": 0.3609,
+      "step": 1850
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 9.919672038835926e-08,
+      "loss": 0.3555,
+      "step": 1851
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 9.68527494520477e-08,
+      "loss": 0.3631,
+      "step": 1852
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 9.453666932985661e-08,
+      "loss": 0.369,
+      "step": 1853
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 9.224848654469932e-08,
+      "loss": 0.3685,
+      "step": 1854
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 8.99882075409153e-08,
+      "loss": 0.3714,
+      "step": 1855
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 8.775583868426363e-08,
+      "loss": 0.3559,
+      "step": 1856
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 8.555138626189619e-08,
+      "loss": 0.3569,
+      "step": 1857
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 8.33748564823389e-08,
+      "loss": 0.3649,
+      "step": 1858
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 8.122625547548279e-08,
+      "loss": 0.358,
+      "step": 1859
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 7.910558929255852e-08,
+      "loss": 0.3606,
+      "step": 1860
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 7.701286390612073e-08,
+      "loss": 0.3542,
+      "step": 1861
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 7.494808521003482e-08,
+      "loss": 0.356,
+      "step": 1862
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 7.291125901946027e-08,
+      "loss": 0.3561,
+      "step": 1863
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 7.090239107082842e-08,
+      "loss": 0.3752,
+      "step": 1864
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 6.892148702183133e-08,
+      "loss": 0.374,
+      "step": 1865
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 6.696855245140743e-08,
+      "loss": 0.3653,
+      "step": 1866
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 6.504359285971818e-08,
+      "loss": 0.3748,
+      "step": 1867
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 6.314661366814245e-08,
+      "loss": 0.3705,
+      "step": 1868
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 6.127762021925221e-08,
+      "loss": 0.363,
+      "step": 1869
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 5.943661777680354e-08,
+      "loss": 0.364,
+      "step": 1870
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 5.7623611525721155e-08,
+      "loss": 0.3513,
+      "step": 1871
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 5.5838606572078404e-08,
+      "loss": 0.3573,
+      "step": 1872
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 5.4081607943090584e-08,
+      "loss": 0.3386,
+      "step": 1873
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 5.235262058709723e-08,
+      "loss": 0.3553,
+      "step": 1874
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 5.0651649373544276e-08,
+      "loss": 0.3513,
+      "step": 1875
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.8978699092977475e-08,
+      "loss": 0.354,
+      "step": 1876
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 4.733377445702569e-08,
+      "loss": 0.3575,
+      "step": 1877
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.571688009838648e-08,
+      "loss": 0.3622,
+      "step": 1878
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.412802057081278e-08,
+      "loss": 0.343,
+      "step": 1879
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.256720034910511e-08,
+      "loss": 0.3523,
+      "step": 1880
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 4.103442382909051e-08,
+      "loss": 0.3667,
+      "step": 1881
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 3.952969532761919e-08,
+      "loss": 0.3577,
+      "step": 1882
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 3.805301908254455e-08,
+      "loss": 0.3564,
+      "step": 1883
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 3.660439925271764e-08,
+      "loss": 0.3514,
+      "step": 1884
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 3.5183839917972695e-08,
+      "loss": 0.3526,
+      "step": 1885
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 3.3791345079111635e-08,
+      "loss": 0.3423,
+      "step": 1886
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 3.242691865790071e-08,
+      "loss": 0.3743,
+      "step": 1887
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 3.109056449705272e-08,
+      "loss": 0.3583,
+      "step": 1888
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 2.9782286360222622e-08,
+      "loss": 0.3553,
+      "step": 1889
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.850208793198861e-08,
+      "loss": 0.3542,
+      "step": 1890
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.7249972817849913e-08,
+      "loss": 0.3656,
+      "step": 1891
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.6025944544211257e-08,
+      "loss": 0.3588,
+      "step": 1892
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.4830006558373975e-08,
+      "loss": 0.3477,
+      "step": 1893
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.366216222853046e-08,
+      "loss": 0.3512,
+      "step": 1894
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.2522414843748618e-08,
+      "loss": 0.3539,
+      "step": 1895
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.1410767613965212e-08,
+      "loss": 0.3713,
+      "step": 1896
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.032722366997697e-08,
+      "loss": 0.3641,
+      "step": 1897
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.927178606343394e-08,
+      "loss": 0.3646,
+      "step": 1898
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.824445776682504e-08,
+      "loss": 0.3475,
+      "step": 1899
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.7245241673476966e-08,
+      "loss": 0.3653,
+      "step": 1900
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.6274140597538624e-08,
+      "loss": 0.3646,
+      "step": 1901
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.5331157273981155e-08,
+      "loss": 0.3619,
+      "step": 1902
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.4416294358582383e-08,
+      "loss": 0.3628,
+      "step": 1903
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.3529554427923475e-08,
+      "loss": 0.3771,
+      "step": 1904
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.267093997938451e-08,
+      "loss": 0.3587,
+      "step": 1905
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.184045343112894e-08,
+      "loss": 0.3533,
+      "step": 1906
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.1038097122106905e-08,
+      "loss": 0.3583,
+      "step": 1907
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.0263873312040818e-08,
+      "loss": 0.3489,
+      "step": 1908
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 9.517784181422018e-09,
+      "loss": 0.3565,
+      "step": 1909
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 8.799831831506345e-09,
+      "loss": 0.3665,
+      "step": 1910
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 8.110018284304132e-09,
+      "loss": 0.3531,
+      "step": 1911
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 7.4483454825768905e-09,
+      "loss": 0.3598,
+      "step": 1912
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 6.814815289833965e-09,
+      "loss": 0.3512,
+      "step": 1913
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 6.209429490323659e-09,
+      "loss": 0.3502,
+      "step": 1914
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 5.632189789027686e-09,
+      "loss": 0.3607,
+      "step": 1915
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 5.083097811662274e-09,
+      "loss": 0.3725,
+      "step": 1916
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.562155104665955e-09,
+      "loss": 0.3575,
+      "step": 1917
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 4.069363135200677e-09,
+      "loss": 0.3665,
+      "step": 1918
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 3.6047232911462506e-09,
+      "loss": 0.3592,
+      "step": 1919
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 3.168236881092579e-09,
+      "loss": 0.3546,
+      "step": 1920
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.7599051343440985e-09,
+      "loss": 0.3517,
+      "step": 1921
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.379729200908676e-09,
+      "loss": 0.3742,
+      "step": 1922
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.0277101514987184e-09,
+      "loss": 0.3532,
+      "step": 1923
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 1.7038489775256239e-09,
+      "loss": 0.3551,
+      "step": 1924
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 1.408146591097559e-09,
+      "loss": 0.3505,
+      "step": 1925
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 1.1406038250205699e-09,
+      "loss": 0.3488,
+      "step": 1926
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 9.012214327897006e-10,
+      "loss": 0.3413,
+      "step": 1927
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 6.900000885923242e-10,
+      "loss": 0.3552,
+      "step": 1928
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 5.069403873025902e-10,
+      "loss": 0.3654,
+      "step": 1929
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 3.5204284448253635e-10,
+      "loss": 0.3138,
+      "step": 1930
+    },
+    {
+      "epoch": 5.0,
+      "step": 1930,
+      "total_flos": 290658866233344.0,
+      "train_loss": 0.6783839724842131,
+      "train_runtime": 84872.0511,
+      "train_samples_per_second": 2.904,
+      "train_steps_per_second": 0.023
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 1930,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 290658866233344.0,
+  "trial_name": null,
+  "trial_params": null
+}