diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,7417 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 1233,
+  "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": 0.9651,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 2.2917,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 2.2298,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9e-07,
+      "loss": 1.0824,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 0.8161,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.5e-06,
+      "loss": 0.8268,
+      "step": 6
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.8e-06,
+      "loss": 0.8638,
+      "step": 7
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.1e-06,
+      "loss": 0.8703,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 0.8445,
+      "step": 9
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.7e-06,
+      "loss": 0.909,
+      "step": 10
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3e-06,
+      "loss": 0.9049,
+      "step": 11
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.3e-06,
+      "loss": 0.9104,
+      "step": 12
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.6e-06,
+      "loss": 0.8589,
+      "step": 13
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.9e-06,
+      "loss": 0.8139,
+      "step": 14
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.2e-06,
+      "loss": 0.8234,
+      "step": 15
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.5e-06,
+      "loss": 0.7649,
+      "step": 16
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 0.73,
+      "step": 17
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.1e-06,
+      "loss": 0.6922,
+      "step": 18
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.4e-06,
+      "loss": 0.682,
+      "step": 19
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.7e-06,
+      "loss": 0.6846,
+      "step": 20
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-06,
+      "loss": 0.6739,
+      "step": 21
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.997545008183306e-06,
+      "loss": 0.652,
+      "step": 22
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.995090016366612e-06,
+      "loss": 0.6735,
+      "step": 23
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.992635024549918e-06,
+      "loss": 0.6616,
+      "step": 24
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.990180032733224e-06,
+      "loss": 0.6716,
+      "step": 25
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.987725040916531e-06,
+      "loss": 0.6712,
+      "step": 26
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.985270049099837e-06,
+      "loss": 0.6694,
+      "step": 27
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.982815057283142e-06,
+      "loss": 0.6473,
+      "step": 28
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.980360065466448e-06,
+      "loss": 0.6608,
+      "step": 29
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.977905073649755e-06,
+      "loss": 0.6639,
+      "step": 30
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.975450081833061e-06,
+      "loss": 0.6458,
+      "step": 31
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.972995090016366e-06,
+      "loss": 0.6379,
+      "step": 32
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.970540098199673e-06,
+      "loss": 0.6426,
+      "step": 33
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.968085106382979e-06,
+      "loss": 0.6317,
+      "step": 34
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.965630114566285e-06,
+      "loss": 0.6406,
+      "step": 35
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.963175122749591e-06,
+      "loss": 0.621,
+      "step": 36
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.960720130932897e-06,
+      "loss": 0.637,
+      "step": 37
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.958265139116203e-06,
+      "loss": 0.6385,
+      "step": 38
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.955810147299509e-06,
+      "loss": 0.6053,
+      "step": 39
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.953355155482815e-06,
+      "loss": 0.618,
+      "step": 40
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.950900163666122e-06,
+      "loss": 0.6357,
+      "step": 41
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.948445171849428e-06,
+      "loss": 0.6145,
+      "step": 42
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.9459901800327334e-06,
+      "loss": 0.6196,
+      "step": 43
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.943535188216039e-06,
+      "loss": 0.6251,
+      "step": 44
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.941080196399345e-06,
+      "loss": 0.6202,
+      "step": 45
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.938625204582651e-06,
+      "loss": 0.6321,
+      "step": 46
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.9361702127659575e-06,
+      "loss": 0.6242,
+      "step": 47
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.933715220949264e-06,
+      "loss": 0.6225,
+      "step": 48
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.93126022913257e-06,
+      "loss": 0.5993,
+      "step": 49
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.928805237315876e-06,
+      "loss": 0.6005,
+      "step": 50
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.9263502454991815e-06,
+      "loss": 0.6354,
+      "step": 51
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.923895253682488e-06,
+      "loss": 0.5904,
+      "step": 52
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.921440261865794e-06,
+      "loss": 0.5923,
+      "step": 53
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.9189852700491e-06,
+      "loss": 0.6157,
+      "step": 54
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.916530278232406e-06,
+      "loss": 0.6257,
+      "step": 55
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.914075286415712e-06,
+      "loss": 0.6063,
+      "step": 56
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.911620294599018e-06,
+      "loss": 0.5841,
+      "step": 57
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.9091653027823245e-06,
+      "loss": 0.6036,
+      "step": 58
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.90671031096563e-06,
+      "loss": 0.6141,
+      "step": 59
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.904255319148936e-06,
+      "loss": 0.5933,
+      "step": 60
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.901800327332242e-06,
+      "loss": 0.5851,
+      "step": 61
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.8993453355155485e-06,
+      "loss": 0.5844,
+      "step": 62
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.896890343698854e-06,
+      "loss": 0.5862,
+      "step": 63
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.894435351882161e-06,
+      "loss": 0.5881,
+      "step": 64
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.891980360065467e-06,
+      "loss": 0.6077,
+      "step": 65
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.8895253682487725e-06,
+      "loss": 0.6136,
+      "step": 66
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.887070376432078e-06,
+      "loss": 0.6074,
+      "step": 67
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.884615384615384e-06,
+      "loss": 0.5779,
+      "step": 68
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.882160392798691e-06,
+      "loss": 0.5741,
+      "step": 69
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.879705400981997e-06,
+      "loss": 0.6041,
+      "step": 70
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.877250409165303e-06,
+      "loss": 0.5849,
+      "step": 71
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.874795417348609e-06,
+      "loss": 0.5931,
+      "step": 72
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.872340425531915e-06,
+      "loss": 0.5982,
+      "step": 73
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8698854337152205e-06,
+      "loss": 0.5762,
+      "step": 74
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.867430441898527e-06,
+      "loss": 0.5959,
+      "step": 75
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.864975450081833e-06,
+      "loss": 0.5673,
+      "step": 76
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.86252045826514e-06,
+      "loss": 0.564,
+      "step": 77
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.860065466448445e-06,
+      "loss": 0.613,
+      "step": 78
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.857610474631751e-06,
+      "loss": 0.5768,
+      "step": 79
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.855155482815058e-06,
+      "loss": 0.6206,
+      "step": 80
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.852700490998364e-06,
+      "loss": 0.5924,
+      "step": 81
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.850245499181669e-06,
+      "loss": 0.5674,
+      "step": 82
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.847790507364975e-06,
+      "loss": 0.5724,
+      "step": 83
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.845335515548282e-06,
+      "loss": 0.5991,
+      "step": 84
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.842880523731588e-06,
+      "loss": 0.5868,
+      "step": 85
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.840425531914894e-06,
+      "loss": 0.5989,
+      "step": 86
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8379705400982e-06,
+      "loss": 0.5929,
+      "step": 87
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.835515548281506e-06,
+      "loss": 0.5985,
+      "step": 88
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.833060556464812e-06,
+      "loss": 0.59,
+      "step": 89
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.830605564648117e-06,
+      "loss": 0.5988,
+      "step": 90
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.828150572831424e-06,
+      "loss": 0.5672,
+      "step": 91
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.825695581014731e-06,
+      "loss": 0.5502,
+      "step": 92
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.8232405891980365e-06,
+      "loss": 0.6034,
+      "step": 93
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.820785597381342e-06,
+      "loss": 0.5815,
+      "step": 94
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.818330605564648e-06,
+      "loss": 0.5778,
+      "step": 95
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.815875613747954e-06,
+      "loss": 0.5781,
+      "step": 96
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.8134206219312605e-06,
+      "loss": 0.5862,
+      "step": 97
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.810965630114566e-06,
+      "loss": 0.5738,
+      "step": 98
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.808510638297873e-06,
+      "loss": 0.5988,
+      "step": 99
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.806055646481179e-06,
+      "loss": 0.5718,
+      "step": 100
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.8036006546644845e-06,
+      "loss": 0.5829,
+      "step": 101
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.80114566284779e-06,
+      "loss": 0.575,
+      "step": 102
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.798690671031097e-06,
+      "loss": 0.5919,
+      "step": 103
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.796235679214403e-06,
+      "loss": 0.5684,
+      "step": 104
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7937806873977085e-06,
+      "loss": 0.6091,
+      "step": 105
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.791325695581015e-06,
+      "loss": 0.5685,
+      "step": 106
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.788870703764321e-06,
+      "loss": 0.5639,
+      "step": 107
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.7864157119476275e-06,
+      "loss": 0.5582,
+      "step": 108
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.783960720130933e-06,
+      "loss": 0.5761,
+      "step": 109
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.781505728314239e-06,
+      "loss": 0.5602,
+      "step": 110
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.779050736497545e-06,
+      "loss": 0.5658,
+      "step": 111
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.776595744680851e-06,
+      "loss": 0.5732,
+      "step": 112
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.774140752864157e-06,
+      "loss": 0.5905,
+      "step": 113
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.771685761047464e-06,
+      "loss": 0.5513,
+      "step": 114
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.76923076923077e-06,
+      "loss": 0.5384,
+      "step": 115
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.7667757774140756e-06,
+      "loss": 0.5668,
+      "step": 116
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.764320785597381e-06,
+      "loss": 0.5795,
+      "step": 117
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.761865793780687e-06,
+      "loss": 0.5606,
+      "step": 118
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.759410801963993e-06,
+      "loss": 0.5608,
+      "step": 119
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.7569558101472996e-06,
+      "loss": 0.5698,
+      "step": 120
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.754500818330606e-06,
+      "loss": 0.5475,
+      "step": 121
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.752045826513912e-06,
+      "loss": 0.5664,
+      "step": 122
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.749590834697218e-06,
+      "loss": 0.5838,
+      "step": 123
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.7471358428805236e-06,
+      "loss": 0.578,
+      "step": 124
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.74468085106383e-06,
+      "loss": 0.577,
+      "step": 125
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.742225859247136e-06,
+      "loss": 0.5604,
+      "step": 126
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.739770867430442e-06,
+      "loss": 0.5618,
+      "step": 127
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.737315875613748e-06,
+      "loss": 0.5867,
+      "step": 128
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.734860883797054e-06,
+      "loss": 0.5658,
+      "step": 129
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.73240589198036e-06,
+      "loss": 0.5658,
+      "step": 130
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.729950900163667e-06,
+      "loss": 0.5521,
+      "step": 131
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.727495908346972e-06,
+      "loss": 0.5738,
+      "step": 132
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.725040916530278e-06,
+      "loss": 0.5711,
+      "step": 133
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.722585924713584e-06,
+      "loss": 0.5536,
+      "step": 134
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.720130932896891e-06,
+      "loss": 0.5654,
+      "step": 135
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.717675941080197e-06,
+      "loss": 0.5605,
+      "step": 136
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.715220949263503e-06,
+      "loss": 0.5864,
+      "step": 137
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.712765957446809e-06,
+      "loss": 0.5707,
+      "step": 138
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.710310965630115e-06,
+      "loss": 0.5898,
+      "step": 139
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.7078559738134204e-06,
+      "loss": 0.5735,
+      "step": 140
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.705400981996726e-06,
+      "loss": 0.5676,
+      "step": 141
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.702945990180033e-06,
+      "loss": 0.5822,
+      "step": 142
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.7004909983633395e-06,
+      "loss": 0.5753,
+      "step": 143
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.698036006546645e-06,
+      "loss": 0.5471,
+      "step": 144
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.695581014729951e-06,
+      "loss": 0.5861,
+      "step": 145
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.693126022913257e-06,
+      "loss": 0.5882,
+      "step": 146
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.690671031096563e-06,
+      "loss": 0.5597,
+      "step": 147
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.688216039279869e-06,
+      "loss": 0.5541,
+      "step": 148
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.685761047463175e-06,
+      "loss": 0.5781,
+      "step": 149
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.683306055646482e-06,
+      "loss": 0.5832,
+      "step": 150
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.6808510638297875e-06,
+      "loss": 0.5487,
+      "step": 151
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.678396072013093e-06,
+      "loss": 0.5618,
+      "step": 152
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.6759410801964e-06,
+      "loss": 0.5679,
+      "step": 153
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.673486088379706e-06,
+      "loss": 0.554,
+      "step": 154
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.6710310965630115e-06,
+      "loss": 0.5876,
+      "step": 155
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.668576104746317e-06,
+      "loss": 0.5913,
+      "step": 156
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.666121112929624e-06,
+      "loss": 0.5878,
+      "step": 157
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.66366612111293e-06,
+      "loss": 0.5505,
+      "step": 158
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.661211129296236e-06,
+      "loss": 0.5587,
+      "step": 159
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.658756137479542e-06,
+      "loss": 0.5398,
+      "step": 160
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.656301145662848e-06,
+      "loss": 0.5393,
+      "step": 161
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.653846153846154e-06,
+      "loss": 0.5726,
+      "step": 162
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.6513911620294595e-06,
+      "loss": 0.5633,
+      "step": 163
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.648936170212766e-06,
+      "loss": 0.5444,
+      "step": 164
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.646481178396072e-06,
+      "loss": 0.5496,
+      "step": 165
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.644026186579379e-06,
+      "loss": 0.541,
+      "step": 166
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.641571194762684e-06,
+      "loss": 0.5607,
+      "step": 167
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.63911620294599e-06,
+      "loss": 0.5677,
+      "step": 168
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.636661211129296e-06,
+      "loss": 0.5688,
+      "step": 169
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.634206219312603e-06,
+      "loss": 0.5315,
+      "step": 170
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.631751227495908e-06,
+      "loss": 0.5709,
+      "step": 171
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.629296235679214e-06,
+      "loss": 0.5422,
+      "step": 172
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.626841243862521e-06,
+      "loss": 0.5578,
+      "step": 173
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.624386252045827e-06,
+      "loss": 0.5568,
+      "step": 174
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.621931260229132e-06,
+      "loss": 0.5786,
+      "step": 175
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.619476268412439e-06,
+      "loss": 0.5334,
+      "step": 176
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.617021276595745e-06,
+      "loss": 0.5645,
+      "step": 177
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.614566284779051e-06,
+      "loss": 0.5489,
+      "step": 178
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.612111292962357e-06,
+      "loss": 0.5396,
+      "step": 179
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.609656301145663e-06,
+      "loss": 0.5439,
+      "step": 180
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.60720130932897e-06,
+      "loss": 0.5769,
+      "step": 181
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.6047463175122754e-06,
+      "loss": 0.5809,
+      "step": 182
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.602291325695581e-06,
+      "loss": 0.5371,
+      "step": 183
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.599836333878887e-06,
+      "loss": 0.5573,
+      "step": 184
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.597381342062193e-06,
+      "loss": 0.5749,
+      "step": 185
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.5949263502454995e-06,
+      "loss": 0.5388,
+      "step": 186
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.592471358428805e-06,
+      "loss": 0.558,
+      "step": 187
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.590016366612112e-06,
+      "loss": 0.5346,
+      "step": 188
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.587561374795418e-06,
+      "loss": 0.5427,
+      "step": 189
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.5851063829787235e-06,
+      "loss": 0.5533,
+      "step": 190
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.582651391162029e-06,
+      "loss": 0.5543,
+      "step": 191
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.580196399345336e-06,
+      "loss": 0.5493,
+      "step": 192
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.577741407528642e-06,
+      "loss": 0.5689,
+      "step": 193
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 5.5752864157119475e-06,
+      "loss": 0.5413,
+      "step": 194
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 5.572831423895254e-06,
+      "loss": 0.5459,
+      "step": 195
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 5.57037643207856e-06,
+      "loss": 0.5533,
+      "step": 196
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 5.567921440261866e-06,
+      "loss": 0.5611,
+      "step": 197
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 5.565466448445172e-06,
+      "loss": 0.5631,
+      "step": 198
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 5.563011456628478e-06,
+      "loss": 0.547,
+      "step": 199
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 5.560556464811784e-06,
+      "loss": 0.5386,
+      "step": 200
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 5.55810147299509e-06,
+      "loss": 0.588,
+      "step": 201
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 5.555646481178396e-06,
+      "loss": 0.5555,
+      "step": 202
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 5.553191489361702e-06,
+      "loss": 0.5193,
+      "step": 203
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 5.550736497545009e-06,
+      "loss": 0.5508,
+      "step": 204
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 5.5482815057283145e-06,
+      "loss": 0.5614,
+      "step": 205
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 5.54582651391162e-06,
+      "loss": 0.5202,
+      "step": 206
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 5.543371522094926e-06,
+      "loss": 0.5749,
+      "step": 207
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 5.540916530278232e-06,
+      "loss": 0.5663,
+      "step": 208
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 5.5384615384615385e-06,
+      "loss": 0.5395,
+      "step": 209
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 5.536006546644845e-06,
+      "loss": 0.556,
+      "step": 210
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 5.533551554828151e-06,
+      "loss": 0.5307,
+      "step": 211
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 5.531096563011457e-06,
+      "loss": 0.5445,
+      "step": 212
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.5286415711947625e-06,
+      "loss": 0.5326,
+      "step": 213
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.526186579378068e-06,
+      "loss": 0.538,
+      "step": 214
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.523731587561375e-06,
+      "loss": 0.542,
+      "step": 215
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.521276595744681e-06,
+      "loss": 0.5442,
+      "step": 216
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.518821603927987e-06,
+      "loss": 0.5553,
+      "step": 217
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.516366612111293e-06,
+      "loss": 0.5521,
+      "step": 218
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.513911620294599e-06,
+      "loss": 0.5651,
+      "step": 219
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.511456628477906e-06,
+      "loss": 0.5359,
+      "step": 220
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.509001636661211e-06,
+      "loss": 0.5524,
+      "step": 221
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.506546644844517e-06,
+      "loss": 0.5501,
+      "step": 222
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.504091653027823e-06,
+      "loss": 0.5436,
+      "step": 223
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.50163666121113e-06,
+      "loss": 0.5559,
+      "step": 224
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 5.499181669394435e-06,
+      "loss": 0.5617,
+      "step": 225
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 5.496726677577742e-06,
+      "loss": 0.5474,
+      "step": 226
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 5.494271685761048e-06,
+      "loss": 0.562,
+      "step": 227
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 5.491816693944354e-06,
+      "loss": 0.5356,
+      "step": 228
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 5.489361702127659e-06,
+      "loss": 0.5636,
+      "step": 229
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 5.486906710310965e-06,
+      "loss": 0.5457,
+      "step": 230
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 5.484451718494272e-06,
+      "loss": 0.5679,
+      "step": 231
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 5.4819967266775785e-06,
+      "loss": 0.5514,
+      "step": 232
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 5.479541734860884e-06,
+      "loss": 0.5368,
+      "step": 233
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 5.47708674304419e-06,
+      "loss": 0.5689,
+      "step": 234
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 5.474631751227496e-06,
+      "loss": 0.5405,
+      "step": 235
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 5.472176759410802e-06,
+      "loss": 0.5427,
+      "step": 236
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 5.469721767594108e-06,
+      "loss": 0.545,
+      "step": 237
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 5.467266775777414e-06,
+      "loss": 0.5633,
+      "step": 238
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 5.464811783960721e-06,
+      "loss": 0.5411,
+      "step": 239
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 5.4623567921440265e-06,
+      "loss": 0.573,
+      "step": 240
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 5.459901800327332e-06,
+      "loss": 0.5367,
+      "step": 241
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 5.457446808510638e-06,
+      "loss": 0.5432,
+      "step": 242
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 5.454991816693945e-06,
+      "loss": 0.5386,
+      "step": 243
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 5.4525368248772505e-06,
+      "loss": 0.5651,
+      "step": 244
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 5.450081833060556e-06,
+      "loss": 0.5575,
+      "step": 245
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 5.447626841243863e-06,
+      "loss": 0.5658,
+      "step": 246
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 5.445171849427169e-06,
+      "loss": 0.5386,
+      "step": 247
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 5.4427168576104745e-06,
+      "loss": 0.5279,
+      "step": 248
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 5.440261865793781e-06,
+      "loss": 0.5329,
+      "step": 249
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.437806873977087e-06,
+      "loss": 0.5412,
+      "step": 250
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.435351882160393e-06,
+      "loss": 0.5449,
+      "step": 251
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.4328968903436985e-06,
+      "loss": 0.5591,
+      "step": 252
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.430441898527005e-06,
+      "loss": 0.5413,
+      "step": 253
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.427986906710312e-06,
+      "loss": 0.5558,
+      "step": 254
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.4255319148936176e-06,
+      "loss": 0.5292,
+      "step": 255
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.423076923076923e-06,
+      "loss": 0.5607,
+      "step": 256
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.420621931260229e-06,
+      "loss": 0.5072,
+      "step": 257
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.418166939443535e-06,
+      "loss": 0.5522,
+      "step": 258
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.415711947626841e-06,
+      "loss": 0.5415,
+      "step": 259
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.413256955810147e-06,
+      "loss": 0.5206,
+      "step": 260
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.410801963993454e-06,
+      "loss": 0.5297,
+      "step": 261
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 5.40834697217676e-06,
+      "loss": 0.561,
+      "step": 262
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 5.4058919803600656e-06,
+      "loss": 0.5589,
+      "step": 263
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 5.403436988543371e-06,
+      "loss": 0.5492,
+      "step": 264
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 5.400981996726678e-06,
+      "loss": 0.5312,
+      "step": 265
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 5.398527004909984e-06,
+      "loss": 0.533,
+      "step": 266
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 5.3960720130932896e-06,
+      "loss": 0.5509,
+      "step": 267
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 5.393617021276596e-06,
+      "loss": 0.5538,
+      "step": 268
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 5.391162029459902e-06,
+      "loss": 0.5341,
+      "step": 269
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 5.388707037643208e-06,
+      "loss": 0.5557,
+      "step": 270
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 5.386252045826514e-06,
+      "loss": 0.536,
+      "step": 271
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 5.38379705400982e-06,
+      "loss": 0.5492,
+      "step": 272
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 5.381342062193126e-06,
+      "loss": 0.5423,
+      "step": 273
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 5.378887070376432e-06,
+      "loss": 0.5604,
+      "step": 274
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 5.3764320785597384e-06,
+      "loss": 0.5443,
+      "step": 275
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 5.373977086743044e-06,
+      "loss": 0.5733,
+      "step": 276
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 5.371522094926351e-06,
+      "loss": 0.5364,
+      "step": 277
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 5.369067103109657e-06,
+      "loss": 0.5141,
+      "step": 278
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 5.3666121112929624e-06,
+      "loss": 0.5671,
+      "step": 279
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 5.364157119476268e-06,
+      "loss": 0.5574,
+      "step": 280
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 5.361702127659574e-06,
+      "loss": 0.5418,
+      "step": 281
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 5.359247135842881e-06,
+      "loss": 0.5271,
+      "step": 282
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 5.356792144026187e-06,
+      "loss": 0.5658,
+      "step": 283
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 5.354337152209493e-06,
+      "loss": 0.5563,
+      "step": 284
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 5.351882160392799e-06,
+      "loss": 0.5562,
+      "step": 285
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 5.349427168576105e-06,
+      "loss": 0.528,
+      "step": 286
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.3469721767594104e-06,
+      "loss": 0.5568,
+      "step": 287
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.344517184942717e-06,
+      "loss": 0.5576,
+      "step": 288
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.342062193126023e-06,
+      "loss": 0.5643,
+      "step": 289
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.3396072013093295e-06,
+      "loss": 0.5368,
+      "step": 290
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.337152209492635e-06,
+      "loss": 0.5454,
+      "step": 291
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.334697217675941e-06,
+      "loss": 0.5223,
+      "step": 292
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.332242225859248e-06,
+      "loss": 0.5643,
+      "step": 293
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.3297872340425535e-06,
+      "loss": 0.5287,
+      "step": 294
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.327332242225859e-06,
+      "loss": 0.5387,
+      "step": 295
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.324877250409165e-06,
+      "loss": 0.5328,
+      "step": 296
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.322422258592472e-06,
+      "loss": 0.5735,
+      "step": 297
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.3199672667757775e-06,
+      "loss": 0.5153,
+      "step": 298
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 5.317512274959084e-06,
+      "loss": 0.5499,
+      "step": 299
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 5.31505728314239e-06,
+      "loss": 0.5788,
+      "step": 300
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 5.312602291325696e-06,
+      "loss": 0.5391,
+      "step": 301
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 5.3101472995090015e-06,
+      "loss": 0.5614,
+      "step": 302
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 5.307692307692307e-06,
+      "loss": 0.556,
+      "step": 303
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 5.305237315875614e-06,
+      "loss": 0.5509,
+      "step": 304
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 5.302782324058921e-06,
+      "loss": 0.5336,
+      "step": 305
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.300327332242226e-06,
+      "loss": 0.5253,
+      "step": 306
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.297872340425532e-06,
+      "loss": 0.543,
+      "step": 307
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.295417348608838e-06,
+      "loss": 0.5548,
+      "step": 308
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.292962356792144e-06,
+      "loss": 0.5407,
+      "step": 309
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.29050736497545e-06,
+      "loss": 0.5459,
+      "step": 310
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.288052373158756e-06,
+      "loss": 0.534,
+      "step": 311
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.285597381342063e-06,
+      "loss": 0.5355,
+      "step": 312
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.283142389525369e-06,
+      "loss": 0.549,
+      "step": 313
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.280687397708674e-06,
+      "loss": 0.572,
+      "step": 314
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.27823240589198e-06,
+      "loss": 0.5226,
+      "step": 315
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.275777414075287e-06,
+      "loss": 0.5585,
+      "step": 316
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 5.273322422258593e-06,
+      "loss": 0.5644,
+      "step": 317
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.270867430441898e-06,
+      "loss": 0.5524,
+      "step": 318
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.268412438625205e-06,
+      "loss": 0.5451,
+      "step": 319
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.265957446808511e-06,
+      "loss": 0.5393,
+      "step": 320
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.2635024549918174e-06,
+      "loss": 0.5423,
+      "step": 321
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.261047463175123e-06,
+      "loss": 0.5167,
+      "step": 322
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 5.258592471358429e-06,
+      "loss": 0.5365,
+      "step": 323
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.256137479541735e-06,
+      "loss": 0.5466,
+      "step": 324
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.253682487725041e-06,
+      "loss": 0.5298,
+      "step": 325
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.251227495908347e-06,
+      "loss": 0.5506,
+      "step": 326
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.248772504091653e-06,
+      "loss": 0.5753,
+      "step": 327
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.24631751227496e-06,
+      "loss": 0.5476,
+      "step": 328
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 5.2438625204582655e-06,
+      "loss": 0.5214,
+      "step": 329
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.241407528641571e-06,
+      "loss": 0.5368,
+      "step": 330
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.238952536824877e-06,
+      "loss": 0.5514,
+      "step": 331
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.236497545008183e-06,
+      "loss": 0.5402,
+      "step": 332
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.2340425531914895e-06,
+      "loss": 0.535,
+      "step": 333
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.231587561374795e-06,
+      "loss": 0.5197,
+      "step": 334
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 5.229132569558102e-06,
+      "loss": 0.5149,
+      "step": 335
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.226677577741408e-06,
+      "loss": 0.534,
+      "step": 336
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.2242225859247135e-06,
+      "loss": 0.5233,
+      "step": 337
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.22176759410802e-06,
+      "loss": 0.5343,
+      "step": 338
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.219312602291326e-06,
+      "loss": 0.5319,
+      "step": 339
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.216857610474632e-06,
+      "loss": 0.5182,
+      "step": 340
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.2144026186579375e-06,
+      "loss": 0.5384,
+      "step": 341
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 5.211947626841244e-06,
+      "loss": 0.5059,
+      "step": 342
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.20949263502455e-06,
+      "loss": 0.5541,
+      "step": 343
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.2070376432078565e-06,
+      "loss": 0.5406,
+      "step": 344
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.204582651391162e-06,
+      "loss": 0.5367,
+      "step": 345
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.202127659574468e-06,
+      "loss": 0.571,
+      "step": 346
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.199672667757774e-06,
+      "loss": 0.5293,
+      "step": 347
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 5.19721767594108e-06,
+      "loss": 0.5337,
+      "step": 348
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.194762684124386e-06,
+      "loss": 0.5235,
+      "step": 349
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.192307692307693e-06,
+      "loss": 0.4971,
+      "step": 350
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.189852700490999e-06,
+      "loss": 0.5441,
+      "step": 351
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.1873977086743045e-06,
+      "loss": 0.5261,
+      "step": 352
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.18494271685761e-06,
+      "loss": 0.5458,
+      "step": 353
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 5.182487725040916e-06,
+      "loss": 0.5236,
+      "step": 354
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.180032733224223e-06,
+      "loss": 0.531,
+      "step": 355
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.1775777414075285e-06,
+      "loss": 0.5343,
+      "step": 356
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.175122749590835e-06,
+      "loss": 0.5355,
+      "step": 357
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.172667757774141e-06,
+      "loss": 0.5175,
+      "step": 358
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.170212765957447e-06,
+      "loss": 0.5335,
+      "step": 359
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 5.1677577741407525e-06,
+      "loss": 0.5316,
+      "step": 360
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.165302782324059e-06,
+      "loss": 0.5343,
+      "step": 361
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.162847790507365e-06,
+      "loss": 0.5682,
+      "step": 362
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.160392798690671e-06,
+      "loss": 0.506,
+      "step": 363
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.157937806873977e-06,
+      "loss": 0.524,
+      "step": 364
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.155482815057283e-06,
+      "loss": 0.5064,
+      "step": 365
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 5.15302782324059e-06,
+      "loss": 0.5285,
+      "step": 366
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 5.150572831423896e-06,
+      "loss": 0.5329,
+      "step": 367
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 5.148117839607201e-06,
+      "loss": 0.5581,
+      "step": 368
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 5.145662847790507e-06,
+      "loss": 0.5532,
+      "step": 369
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 5.143207855973813e-06,
+      "loss": 0.5144,
+      "step": 370
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 5.14075286415712e-06,
+      "loss": 0.526,
+      "step": 371
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 5.138297872340426e-06,
+      "loss": 0.5149,
+      "step": 372
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.135842880523732e-06,
+      "loss": 0.5523,
+      "step": 373
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.133387888707038e-06,
+      "loss": 0.5585,
+      "step": 374
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.130932896890344e-06,
+      "loss": 0.5672,
+      "step": 375
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.128477905073649e-06,
+      "loss": 0.5405,
+      "step": 376
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.126022913256956e-06,
+      "loss": 0.5295,
+      "step": 377
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.123567921440262e-06,
+      "loss": 0.5354,
+      "step": 378
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 5.1211129296235685e-06,
+      "loss": 0.5524,
+      "step": 379
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 5.118657937806874e-06,
+      "loss": 0.5487,
+      "step": 380
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 5.11620294599018e-06,
+      "loss": 0.5221,
+      "step": 381
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 5.113747954173486e-06,
+      "loss": 0.5171,
+      "step": 382
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 5.1112929623567925e-06,
+      "loss": 0.5293,
+      "step": 383
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 5.108837970540098e-06,
+      "loss": 0.5438,
+      "step": 384
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 5.106382978723404e-06,
+      "loss": 0.5163,
+      "step": 385
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 5.103927986906711e-06,
+      "loss": 0.5412,
+      "step": 386
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 5.1014729950900165e-06,
+      "loss": 0.5123,
+      "step": 387
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 5.099018003273322e-06,
+      "loss": 0.515,
+      "step": 388
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 5.096563011456629e-06,
+      "loss": 0.5334,
+      "step": 389
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 5.094108019639935e-06,
+      "loss": 0.5158,
+      "step": 390
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 5.0916530278232405e-06,
+      "loss": 0.5438,
+      "step": 391
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.089198036006546e-06,
+      "loss": 0.5035,
+      "step": 392
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.086743044189853e-06,
+      "loss": 0.539,
+      "step": 393
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.0842880523731596e-06,
+      "loss": 0.5575,
+      "step": 394
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.081833060556465e-06,
+      "loss": 0.5455,
+      "step": 395
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.079378068739771e-06,
+      "loss": 0.5344,
+      "step": 396
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.076923076923077e-06,
+      "loss": 0.5555,
+      "step": 397
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.074468085106383e-06,
+      "loss": 0.5507,
+      "step": 398
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.0720130932896885e-06,
+      "loss": 0.5379,
+      "step": 399
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.069558101472995e-06,
+      "loss": 0.5179,
+      "step": 400
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.067103109656302e-06,
+      "loss": 0.5151,
+      "step": 401
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.0646481178396076e-06,
+      "loss": 0.5558,
+      "step": 402
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.062193126022913e-06,
+      "loss": 0.5233,
+      "step": 403
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.059738134206219e-06,
+      "loss": 0.5496,
+      "step": 404
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.057283142389526e-06,
+      "loss": 0.5206,
+      "step": 405
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.0548281505728316e-06,
+      "loss": 0.554,
+      "step": 406
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.052373158756137e-06,
+      "loss": 0.5454,
+      "step": 407
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.049918166939444e-06,
+      "loss": 0.5042,
+      "step": 408
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.04746317512275e-06,
+      "loss": 0.507,
+      "step": 409
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.0450081833060556e-06,
+      "loss": 0.5644,
+      "step": 410
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.042553191489362e-06,
+      "loss": 0.5171,
+      "step": 411
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.040098199672668e-06,
+      "loss": 0.5391,
+      "step": 412
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.037643207855974e-06,
+      "loss": 0.5104,
+      "step": 413
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.03518821603928e-06,
+      "loss": 0.5586,
+      "step": 414
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.032733224222586e-06,
+      "loss": 0.5443,
+      "step": 415
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.030278232405892e-06,
+      "loss": 0.5137,
+      "step": 416
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.027823240589199e-06,
+      "loss": 0.5422,
+      "step": 417
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.0253682487725044e-06,
+      "loss": 0.5508,
+      "step": 418
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.02291325695581e-06,
+      "loss": 0.5394,
+      "step": 419
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.020458265139116e-06,
+      "loss": 0.5297,
+      "step": 420
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.018003273322422e-06,
+      "loss": 0.518,
+      "step": 421
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.0155482815057284e-06,
+      "loss": 0.541,
+      "step": 422
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.013093289689035e-06,
+      "loss": 0.5295,
+      "step": 423
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.010638297872341e-06,
+      "loss": 0.5435,
+      "step": 424
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.008183306055647e-06,
+      "loss": 0.527,
+      "step": 425
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.0057283142389524e-06,
+      "loss": 0.5103,
+      "step": 426
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.003273322422258e-06,
+      "loss": 0.5458,
+      "step": 427
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 5.000818330605565e-06,
+      "loss": 0.5466,
+      "step": 428
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.998363338788871e-06,
+      "loss": 0.5515,
+      "step": 429
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.995908346972177e-06,
+      "loss": 0.5232,
+      "step": 430
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.993453355155483e-06,
+      "loss": 0.5667,
+      "step": 431
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.990998363338789e-06,
+      "loss": 0.5447,
+      "step": 432
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.988543371522095e-06,
+      "loss": 0.5178,
+      "step": 433
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.986088379705401e-06,
+      "loss": 0.5325,
+      "step": 434
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.983633387888707e-06,
+      "loss": 0.5345,
+      "step": 435
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.981178396072013e-06,
+      "loss": 0.5374,
+      "step": 436
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.9787234042553195e-06,
+      "loss": 0.5215,
+      "step": 437
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.976268412438625e-06,
+      "loss": 0.5386,
+      "step": 438
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.973813420621932e-06,
+      "loss": 0.5032,
+      "step": 439
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.971358428805238e-06,
+      "loss": 0.5657,
+      "step": 440
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.9689034369885435e-06,
+      "loss": 0.5227,
+      "step": 441
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.966448445171849e-06,
+      "loss": 0.5081,
+      "step": 442
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.963993453355155e-06,
+      "loss": 0.5491,
+      "step": 443
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.961538461538462e-06,
+      "loss": 0.5257,
+      "step": 444
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.959083469721768e-06,
+      "loss": 0.5189,
+      "step": 445
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.956628477905074e-06,
+      "loss": 0.5322,
+      "step": 446
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.95417348608838e-06,
+      "loss": 0.526,
+      "step": 447
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.951718494271686e-06,
+      "loss": 0.5018,
+      "step": 448
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.9492635024549915e-06,
+      "loss": 0.5379,
+      "step": 449
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.946808510638298e-06,
+      "loss": 0.5503,
+      "step": 450
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.944353518821604e-06,
+      "loss": 0.5398,
+      "step": 451
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.941898527004911e-06,
+      "loss": 0.5196,
+      "step": 452
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.939443535188216e-06,
+      "loss": 0.5457,
+      "step": 453
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.936988543371522e-06,
+      "loss": 0.5653,
+      "step": 454
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.934533551554828e-06,
+      "loss": 0.5331,
+      "step": 455
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.932078559738135e-06,
+      "loss": 0.5187,
+      "step": 456
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.92962356792144e-06,
+      "loss": 0.5152,
+      "step": 457
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.927168576104746e-06,
+      "loss": 0.5036,
+      "step": 458
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.924713584288053e-06,
+      "loss": 0.546,
+      "step": 459
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.922258592471359e-06,
+      "loss": 0.5226,
+      "step": 460
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.919803600654664e-06,
+      "loss": 0.5133,
+      "step": 461
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.917348608837971e-06,
+      "loss": 0.5284,
+      "step": 462
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.914893617021277e-06,
+      "loss": 0.5196,
+      "step": 463
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.912438625204583e-06,
+      "loss": 0.544,
+      "step": 464
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.909983633387888e-06,
+      "loss": 0.5435,
+      "step": 465
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.907528641571195e-06,
+      "loss": 0.5203,
+      "step": 466
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.905073649754501e-06,
+      "loss": 0.5299,
+      "step": 467
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.9026186579378075e-06,
+      "loss": 0.5406,
+      "step": 468
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.900163666121113e-06,
+      "loss": 0.5277,
+      "step": 469
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.897708674304419e-06,
+      "loss": 0.5443,
+      "step": 470
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.895253682487725e-06,
+      "loss": 0.5418,
+      "step": 471
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.892798690671031e-06,
+      "loss": 0.5338,
+      "step": 472
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.890343698854337e-06,
+      "loss": 0.5307,
+      "step": 473
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.887888707037643e-06,
+      "loss": 0.5298,
+      "step": 474
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.88543371522095e-06,
+      "loss": 0.5049,
+      "step": 475
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.8829787234042555e-06,
+      "loss": 0.549,
+      "step": 476
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.880523731587561e-06,
+      "loss": 0.5234,
+      "step": 477
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.878068739770868e-06,
+      "loss": 0.5503,
+      "step": 478
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.875613747954174e-06,
+      "loss": 0.5152,
+      "step": 479
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.8731587561374795e-06,
+      "loss": 0.5162,
+      "step": 480
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.870703764320786e-06,
+      "loss": 0.5392,
+      "step": 481
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.868248772504092e-06,
+      "loss": 0.529,
+      "step": 482
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.865793780687398e-06,
+      "loss": 0.5478,
+      "step": 483
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.863338788870704e-06,
+      "loss": 0.5636,
+      "step": 484
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.86088379705401e-06,
+      "loss": 0.5583,
+      "step": 485
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.858428805237316e-06,
+      "loss": 0.5379,
+      "step": 486
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.855973813420622e-06,
+      "loss": 0.5331,
+      "step": 487
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.853518821603928e-06,
+      "loss": 0.543,
+      "step": 488
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.851063829787234e-06,
+      "loss": 0.5504,
+      "step": 489
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.848608837970541e-06,
+      "loss": 0.5,
+      "step": 490
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8461538461538465e-06,
+      "loss": 0.5087,
+      "step": 491
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.843698854337152e-06,
+      "loss": 0.5249,
+      "step": 492
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.841243862520458e-06,
+      "loss": 0.5309,
+      "step": 493
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.838788870703764e-06,
+      "loss": 0.5268,
+      "step": 494
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.8363338788870705e-06,
+      "loss": 0.4992,
+      "step": 495
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.833878887070376e-06,
+      "loss": 0.5412,
+      "step": 496
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.831423895253683e-06,
+      "loss": 0.5414,
+      "step": 497
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.828968903436989e-06,
+      "loss": 0.5157,
+      "step": 498
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.8265139116202945e-06,
+      "loss": 0.5292,
+      "step": 499
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.8240589198036e-06,
+      "loss": 0.5027,
+      "step": 500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.821603927986907e-06,
+      "loss": 0.5457,
+      "step": 501
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.819148936170213e-06,
+      "loss": 0.5036,
+      "step": 502
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.8166939443535186e-06,
+      "loss": 0.5345,
+      "step": 503
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.814238952536825e-06,
+      "loss": 0.5071,
+      "step": 504
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.811783960720131e-06,
+      "loss": 0.5273,
+      "step": 505
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.809328968903438e-06,
+      "loss": 0.5629,
+      "step": 506
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.806873977086743e-06,
+      "loss": 0.5279,
+      "step": 507
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.804418985270049e-06,
+      "loss": 0.5173,
+      "step": 508
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.801963993453355e-06,
+      "loss": 0.5361,
+      "step": 509
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.799509001636661e-06,
+      "loss": 0.5425,
+      "step": 510
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.797054009819967e-06,
+      "loss": 0.53,
+      "step": 511
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.794599018003274e-06,
+      "loss": 0.5246,
+      "step": 512
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.79214402618658e-06,
+      "loss": 0.5475,
+      "step": 513
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.789689034369886e-06,
+      "loss": 0.5221,
+      "step": 514
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.787234042553191e-06,
+      "loss": 0.5413,
+      "step": 515
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.784779050736497e-06,
+      "loss": 0.5413,
+      "step": 516
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.782324058919803e-06,
+      "loss": 0.5086,
+      "step": 517
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.77986906710311e-06,
+      "loss": 0.5358,
+      "step": 518
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.777414075286416e-06,
+      "loss": 0.5471,
+      "step": 519
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.774959083469722e-06,
+      "loss": 0.5227,
+      "step": 520
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.772504091653028e-06,
+      "loss": 0.5308,
+      "step": 521
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.770049099836334e-06,
+      "loss": 0.5245,
+      "step": 522
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.76759410801964e-06,
+      "loss": 0.5209,
+      "step": 523
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.765139116202946e-06,
+      "loss": 0.5418,
+      "step": 524
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.762684124386252e-06,
+      "loss": 0.5703,
+      "step": 525
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.7602291325695585e-06,
+      "loss": 0.5505,
+      "step": 526
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.757774140752864e-06,
+      "loss": 0.5309,
+      "step": 527
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.75531914893617e-06,
+      "loss": 0.5381,
+      "step": 528
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.752864157119477e-06,
+      "loss": 0.5226,
+      "step": 529
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.7504091653027825e-06,
+      "loss": 0.5191,
+      "step": 530
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.747954173486088e-06,
+      "loss": 0.5088,
+      "step": 531
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.745499181669394e-06,
+      "loss": 0.5367,
+      "step": 532
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.743044189852701e-06,
+      "loss": 0.5507,
+      "step": 533
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.740589198036007e-06,
+      "loss": 0.5172,
+      "step": 534
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.738134206219313e-06,
+      "loss": 0.5306,
+      "step": 535
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.735679214402619e-06,
+      "loss": 0.5469,
+      "step": 536
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.733224222585925e-06,
+      "loss": 0.5436,
+      "step": 537
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.7307692307692305e-06,
+      "loss": 0.5203,
+      "step": 538
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.728314238952536e-06,
+      "loss": 0.5298,
+      "step": 539
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.725859247135843e-06,
+      "loss": 0.5298,
+      "step": 540
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.7234042553191496e-06,
+      "loss": 0.5306,
+      "step": 541
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.720949263502455e-06,
+      "loss": 0.5381,
+      "step": 542
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.718494271685761e-06,
+      "loss": 0.5065,
+      "step": 543
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.716039279869067e-06,
+      "loss": 0.537,
+      "step": 544
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.713584288052373e-06,
+      "loss": 0.5263,
+      "step": 545
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.711129296235679e-06,
+      "loss": 0.5162,
+      "step": 546
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.708674304418985e-06,
+      "loss": 0.5279,
+      "step": 547
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.706219312602292e-06,
+      "loss": 0.5499,
+      "step": 548
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.7037643207855976e-06,
+      "loss": 0.5327,
+      "step": 549
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.701309328968903e-06,
+      "loss": 0.539,
+      "step": 550
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.69885433715221e-06,
+      "loss": 0.5195,
+      "step": 551
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.696399345335516e-06,
+      "loss": 0.5086,
+      "step": 552
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.6939443535188216e-06,
+      "loss": 0.5323,
+      "step": 553
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.691489361702127e-06,
+      "loss": 0.5523,
+      "step": 554
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.689034369885434e-06,
+      "loss": 0.5496,
+      "step": 555
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.68657937806874e-06,
+      "loss": 0.5495,
+      "step": 556
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.6841243862520464e-06,
+      "loss": 0.509,
+      "step": 557
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.681669394435352e-06,
+      "loss": 0.5425,
+      "step": 558
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.679214402618658e-06,
+      "loss": 0.5529,
+      "step": 559
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.676759410801964e-06,
+      "loss": 0.5287,
+      "step": 560
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.67430441898527e-06,
+      "loss": 0.5379,
+      "step": 561
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.671849427168576e-06,
+      "loss": 0.5424,
+      "step": 562
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.669394435351883e-06,
+      "loss": 0.5241,
+      "step": 563
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.666939443535189e-06,
+      "loss": 0.5415,
+      "step": 564
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.6644844517184944e-06,
+      "loss": 0.5141,
+      "step": 565
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.6620294599018e-06,
+      "loss": 0.5195,
+      "step": 566
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.659574468085106e-06,
+      "loss": 0.5435,
+      "step": 567
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.657119476268413e-06,
+      "loss": 0.5499,
+      "step": 568
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.6546644844517184e-06,
+      "loss": 0.525,
+      "step": 569
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.652209492635025e-06,
+      "loss": 0.5337,
+      "step": 570
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.649754500818331e-06,
+      "loss": 0.5328,
+      "step": 571
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.647299509001637e-06,
+      "loss": 0.5154,
+      "step": 572
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.6448445171849424e-06,
+      "loss": 0.5677,
+      "step": 573
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.642389525368249e-06,
+      "loss": 0.5159,
+      "step": 574
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.639934533551555e-06,
+      "loss": 0.4756,
+      "step": 575
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.637479541734861e-06,
+      "loss": 0.5248,
+      "step": 576
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.635024549918167e-06,
+      "loss": 0.5271,
+      "step": 577
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.632569558101473e-06,
+      "loss": 0.5566,
+      "step": 578
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.63011456628478e-06,
+      "loss": 0.5054,
+      "step": 579
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.6276595744680855e-06,
+      "loss": 0.513,
+      "step": 580
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.625204582651391e-06,
+      "loss": 0.5271,
+      "step": 581
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.622749590834697e-06,
+      "loss": 0.5176,
+      "step": 582
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.620294599018003e-06,
+      "loss": 0.5316,
+      "step": 583
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.6178396072013095e-06,
+      "loss": 0.5632,
+      "step": 584
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.615384615384616e-06,
+      "loss": 0.5506,
+      "step": 585
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.612929623567922e-06,
+      "loss": 0.5366,
+      "step": 586
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.610474631751228e-06,
+      "loss": 0.5212,
+      "step": 587
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.6080196399345335e-06,
+      "loss": 0.5271,
+      "step": 588
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.605564648117839e-06,
+      "loss": 0.5044,
+      "step": 589
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.603109656301146e-06,
+      "loss": 0.5338,
+      "step": 590
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.600654664484452e-06,
+      "loss": 0.5251,
+      "step": 591
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.598199672667758e-06,
+      "loss": 0.5257,
+      "step": 592
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.595744680851064e-06,
+      "loss": 0.5489,
+      "step": 593
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.59328968903437e-06,
+      "loss": 0.5391,
+      "step": 594
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.590834697217676e-06,
+      "loss": 0.5396,
+      "step": 595
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.588379705400982e-06,
+      "loss": 0.5366,
+      "step": 596
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.585924713584288e-06,
+      "loss": 0.5063,
+      "step": 597
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.583469721767594e-06,
+      "loss": 0.536,
+      "step": 598
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.581014729950901e-06,
+      "loss": 0.5119,
+      "step": 599
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.578559738134206e-06,
+      "loss": 0.5352,
+      "step": 600
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.576104746317512e-06,
+      "loss": 0.5428,
+      "step": 601
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.573649754500819e-06,
+      "loss": 0.5234,
+      "step": 602
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.571194762684125e-06,
+      "loss": 0.5373,
+      "step": 603
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.56873977086743e-06,
+      "loss": 0.5354,
+      "step": 604
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.566284779050736e-06,
+      "loss": 0.5197,
+      "step": 605
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.563829787234043e-06,
+      "loss": 0.5083,
+      "step": 606
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.5613747954173494e-06,
+      "loss": 0.536,
+      "step": 607
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.558919803600655e-06,
+      "loss": 0.5158,
+      "step": 608
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.556464811783961e-06,
+      "loss": 0.4954,
+      "step": 609
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.554009819967267e-06,
+      "loss": 0.5122,
+      "step": 610
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.551554828150573e-06,
+      "loss": 0.5215,
+      "step": 611
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.549099836333878e-06,
+      "loss": 0.5443,
+      "step": 612
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.546644844517185e-06,
+      "loss": 0.5205,
+      "step": 613
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.544189852700492e-06,
+      "loss": 0.5271,
+      "step": 614
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.5417348608837975e-06,
+      "loss": 0.5283,
+      "step": 615
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.539279869067103e-06,
+      "loss": 0.5336,
+      "step": 616
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.536824877250409e-06,
+      "loss": 0.5328,
+      "step": 617
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.534369885433716e-06,
+      "loss": 0.5358,
+      "step": 618
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.5319148936170215e-06,
+      "loss": 0.5323,
+      "step": 619
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.529459901800327e-06,
+      "loss": 0.5215,
+      "step": 620
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.527004909983634e-06,
+      "loss": 0.5298,
+      "step": 621
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.52454991816694e-06,
+      "loss": 0.495,
+      "step": 622
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.5220949263502455e-06,
+      "loss": 0.5403,
+      "step": 623
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.519639934533552e-06,
+      "loss": 0.502,
+      "step": 624
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.517184942716858e-06,
+      "loss": 0.5147,
+      "step": 625
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.514729950900164e-06,
+      "loss": 0.5315,
+      "step": 626
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.5122749590834695e-06,
+      "loss": 0.5227,
+      "step": 627
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.509819967266776e-06,
+      "loss": 0.5241,
+      "step": 628
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.507364975450082e-06,
+      "loss": 0.5318,
+      "step": 629
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.5049099836333885e-06,
+      "loss": 0.4985,
+      "step": 630
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.502454991816694e-06,
+      "loss": 0.503,
+      "step": 631
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.5e-06,
+      "loss": 0.515,
+      "step": 632
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.497545008183306e-06,
+      "loss": 0.5374,
+      "step": 633
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.495090016366612e-06,
+      "loss": 0.5016,
+      "step": 634
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.492635024549918e-06,
+      "loss": 0.5018,
+      "step": 635
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.490180032733224e-06,
+      "loss": 0.4957,
+      "step": 636
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.487725040916531e-06,
+      "loss": 0.5123,
+      "step": 637
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.4852700490998365e-06,
+      "loss": 0.5044,
+      "step": 638
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.482815057283142e-06,
+      "loss": 0.5132,
+      "step": 639
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.480360065466448e-06,
+      "loss": 0.5124,
+      "step": 640
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.477905073649755e-06,
+      "loss": 0.5087,
+      "step": 641
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.4754500818330605e-06,
+      "loss": 0.5339,
+      "step": 642
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.472995090016366e-06,
+      "loss": 0.5179,
+      "step": 643
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.470540098199673e-06,
+      "loss": 0.5241,
+      "step": 644
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.468085106382979e-06,
+      "loss": 0.5215,
+      "step": 645
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.4656301145662846e-06,
+      "loss": 0.5265,
+      "step": 646
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.463175122749591e-06,
+      "loss": 0.5232,
+      "step": 647
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.460720130932897e-06,
+      "loss": 0.5178,
+      "step": 648
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.458265139116203e-06,
+      "loss": 0.5283,
+      "step": 649
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.4558101472995086e-06,
+      "loss": 0.5108,
+      "step": 650
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.453355155482815e-06,
+      "loss": 0.5147,
+      "step": 651
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.450900163666122e-06,
+      "loss": 0.519,
+      "step": 652
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.448445171849428e-06,
+      "loss": 0.4891,
+      "step": 653
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.445990180032733e-06,
+      "loss": 0.5392,
+      "step": 654
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.443535188216039e-06,
+      "loss": 0.5105,
+      "step": 655
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.441080196399345e-06,
+      "loss": 0.4892,
+      "step": 656
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.438625204582651e-06,
+      "loss": 0.5255,
+      "step": 657
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.436170212765957e-06,
+      "loss": 0.5046,
+      "step": 658
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.433715220949264e-06,
+      "loss": 0.5253,
+      "step": 659
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.43126022913257e-06,
+      "loss": 0.4971,
+      "step": 660
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.428805237315876e-06,
+      "loss": 0.5251,
+      "step": 661
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.426350245499181e-06,
+      "loss": 0.5303,
+      "step": 662
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.423895253682488e-06,
+      "loss": 0.5038,
+      "step": 663
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.421440261865794e-06,
+      "loss": 0.5415,
+      "step": 664
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.4189852700491e-06,
+      "loss": 0.4947,
+      "step": 665
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.416530278232406e-06,
+      "loss": 0.4983,
+      "step": 666
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.414075286415712e-06,
+      "loss": 0.5343,
+      "step": 667
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.411620294599018e-06,
+      "loss": 0.5187,
+      "step": 668
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.4091653027823245e-06,
+      "loss": 0.4729,
+      "step": 669
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.40671031096563e-06,
+      "loss": 0.5064,
+      "step": 670
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.404255319148936e-06,
+      "loss": 0.5197,
+      "step": 671
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.401800327332242e-06,
+      "loss": 0.5191,
+      "step": 672
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.3993453355155485e-06,
+      "loss": 0.5069,
+      "step": 673
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.396890343698854e-06,
+      "loss": 0.487,
+      "step": 674
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.394435351882161e-06,
+      "loss": 0.5261,
+      "step": 675
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.391980360065467e-06,
+      "loss": 0.5175,
+      "step": 676
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.3895253682487725e-06,
+      "loss": 0.4791,
+      "step": 677
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.387070376432078e-06,
+      "loss": 0.4863,
+      "step": 678
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.384615384615384e-06,
+      "loss": 0.5026,
+      "step": 679
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.382160392798691e-06,
+      "loss": 0.5055,
+      "step": 680
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.379705400981997e-06,
+      "loss": 0.5073,
+      "step": 681
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.377250409165303e-06,
+      "loss": 0.5205,
+      "step": 682
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.374795417348609e-06,
+      "loss": 0.5159,
+      "step": 683
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.372340425531915e-06,
+      "loss": 0.5221,
+      "step": 684
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.3698854337152205e-06,
+      "loss": 0.4749,
+      "step": 685
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.367430441898527e-06,
+      "loss": 0.5104,
+      "step": 686
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.364975450081833e-06,
+      "loss": 0.4992,
+      "step": 687
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.3625204582651396e-06,
+      "loss": 0.5046,
+      "step": 688
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.360065466448445e-06,
+      "loss": 0.494,
+      "step": 689
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.357610474631751e-06,
+      "loss": 0.5043,
+      "step": 690
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.355155482815058e-06,
+      "loss": 0.503,
+      "step": 691
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.3527004909983636e-06,
+      "loss": 0.5129,
+      "step": 692
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.350245499181669e-06,
+      "loss": 0.4685,
+      "step": 693
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.347790507364975e-06,
+      "loss": 0.5161,
+      "step": 694
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.345335515548282e-06,
+      "loss": 0.4987,
+      "step": 695
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.342880523731588e-06,
+      "loss": 0.5232,
+      "step": 696
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.340425531914894e-06,
+      "loss": 0.5055,
+      "step": 697
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.3379705400982e-06,
+      "loss": 0.4985,
+      "step": 698
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.335515548281506e-06,
+      "loss": 0.4897,
+      "step": 699
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.333060556464812e-06,
+      "loss": 0.4947,
+      "step": 700
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.330605564648117e-06,
+      "loss": 0.5075,
+      "step": 701
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.328150572831424e-06,
+      "loss": 0.5095,
+      "step": 702
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.325695581014731e-06,
+      "loss": 0.5118,
+      "step": 703
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.3232405891980364e-06,
+      "loss": 0.5124,
+      "step": 704
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.320785597381342e-06,
+      "loss": 0.512,
+      "step": 705
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.318330605564648e-06,
+      "loss": 0.5238,
+      "step": 706
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.315875613747954e-06,
+      "loss": 0.5003,
+      "step": 707
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.3134206219312604e-06,
+      "loss": 0.4673,
+      "step": 708
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.310965630114566e-06,
+      "loss": 0.4942,
+      "step": 709
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.308510638297873e-06,
+      "loss": 0.5196,
+      "step": 710
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.306055646481179e-06,
+      "loss": 0.488,
+      "step": 711
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.3036006546644844e-06,
+      "loss": 0.4947,
+      "step": 712
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.30114566284779e-06,
+      "loss": 0.4909,
+      "step": 713
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.298690671031097e-06,
+      "loss": 0.5184,
+      "step": 714
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.296235679214403e-06,
+      "loss": 0.4897,
+      "step": 715
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.2937806873977084e-06,
+      "loss": 0.5105,
+      "step": 716
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.291325695581015e-06,
+      "loss": 0.4994,
+      "step": 717
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.288870703764321e-06,
+      "loss": 0.5066,
+      "step": 718
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.2864157119476275e-06,
+      "loss": 0.4996,
+      "step": 719
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.283960720130933e-06,
+      "loss": 0.5005,
+      "step": 720
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.281505728314239e-06,
+      "loss": 0.5126,
+      "step": 721
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.279050736497545e-06,
+      "loss": 0.5089,
+      "step": 722
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.276595744680851e-06,
+      "loss": 0.4857,
+      "step": 723
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.274140752864157e-06,
+      "loss": 0.4708,
+      "step": 724
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.271685761047464e-06,
+      "loss": 0.4939,
+      "step": 725
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.26923076923077e-06,
+      "loss": 0.4865,
+      "step": 726
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.2667757774140755e-06,
+      "loss": 0.4834,
+      "step": 727
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.264320785597381e-06,
+      "loss": 0.4813,
+      "step": 728
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.261865793780687e-06,
+      "loss": 0.5078,
+      "step": 729
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.259410801963993e-06,
+      "loss": 0.5134,
+      "step": 730
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.2569558101472995e-06,
+      "loss": 0.4494,
+      "step": 731
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.254500818330606e-06,
+      "loss": 0.4842,
+      "step": 732
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.252045826513912e-06,
+      "loss": 0.4894,
+      "step": 733
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.249590834697218e-06,
+      "loss": 0.49,
+      "step": 734
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.2471358428805235e-06,
+      "loss": 0.4882,
+      "step": 735
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.24468085106383e-06,
+      "loss": 0.4804,
+      "step": 736
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.242225859247136e-06,
+      "loss": 0.4799,
+      "step": 737
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.239770867430442e-06,
+      "loss": 0.464,
+      "step": 738
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.237315875613748e-06,
+      "loss": 0.5139,
+      "step": 739
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.234860883797054e-06,
+      "loss": 0.4891,
+      "step": 740
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.23240589198036e-06,
+      "loss": 0.5039,
+      "step": 741
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.229950900163667e-06,
+      "loss": 0.4829,
+      "step": 742
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.227495908346972e-06,
+      "loss": 0.4715,
+      "step": 743
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.225040916530278e-06,
+      "loss": 0.4905,
+      "step": 744
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.222585924713584e-06,
+      "loss": 0.4933,
+      "step": 745
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.220130932896891e-06,
+      "loss": 0.4914,
+      "step": 746
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.217675941080197e-06,
+      "loss": 0.4835,
+      "step": 747
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.215220949263503e-06,
+      "loss": 0.475,
+      "step": 748
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.212765957446809e-06,
+      "loss": 0.4893,
+      "step": 749
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.210310965630115e-06,
+      "loss": 0.4848,
+      "step": 750
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.20785597381342e-06,
+      "loss": 0.4958,
+      "step": 751
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.205400981996726e-06,
+      "loss": 0.4787,
+      "step": 752
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.202945990180033e-06,
+      "loss": 0.4937,
+      "step": 753
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.2004909983633395e-06,
+      "loss": 0.5131,
+      "step": 754
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.198036006546645e-06,
+      "loss": 0.5042,
+      "step": 755
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.195581014729951e-06,
+      "loss": 0.5046,
+      "step": 756
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.193126022913257e-06,
+      "loss": 0.4927,
+      "step": 757
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.190671031096563e-06,
+      "loss": 0.4885,
+      "step": 758
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.188216039279869e-06,
+      "loss": 0.5303,
+      "step": 759
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.185761047463175e-06,
+      "loss": 0.4858,
+      "step": 760
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.183306055646482e-06,
+      "loss": 0.4787,
+      "step": 761
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.1808510638297875e-06,
+      "loss": 0.523,
+      "step": 762
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.178396072013093e-06,
+      "loss": 0.5041,
+      "step": 763
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.1759410801964e-06,
+      "loss": 0.4618,
+      "step": 764
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.173486088379706e-06,
+      "loss": 0.489,
+      "step": 765
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.1710310965630115e-06,
+      "loss": 0.4974,
+      "step": 766
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.168576104746317e-06,
+      "loss": 0.5085,
+      "step": 767
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.166121112929624e-06,
+      "loss": 0.4758,
+      "step": 768
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.16366612111293e-06,
+      "loss": 0.4871,
+      "step": 769
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.161211129296236e-06,
+      "loss": 0.4909,
+      "step": 770
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.158756137479542e-06,
+      "loss": 0.4828,
+      "step": 771
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.156301145662848e-06,
+      "loss": 0.5117,
+      "step": 772
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.153846153846154e-06,
+      "loss": 0.5197,
+      "step": 773
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.1513911620294595e-06,
+      "loss": 0.4858,
+      "step": 774
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.148936170212766e-06,
+      "loss": 0.4775,
+      "step": 775
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.146481178396072e-06,
+      "loss": 0.498,
+      "step": 776
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.1440261865793785e-06,
+      "loss": 0.4619,
+      "step": 777
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.141571194762684e-06,
+      "loss": 0.4781,
+      "step": 778
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.13911620294599e-06,
+      "loss": 0.4792,
+      "step": 779
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.136661211129296e-06,
+      "loss": 0.4822,
+      "step": 780
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.1342062193126025e-06,
+      "loss": 0.4739,
+      "step": 781
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.131751227495908e-06,
+      "loss": 0.4694,
+      "step": 782
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.129296235679214e-06,
+      "loss": 0.4714,
+      "step": 783
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.126841243862521e-06,
+      "loss": 0.4891,
+      "step": 784
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.1243862520458266e-06,
+      "loss": 0.4946,
+      "step": 785
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.121931260229132e-06,
+      "loss": 0.4605,
+      "step": 786
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.119476268412439e-06,
+      "loss": 0.4881,
+      "step": 787
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.117021276595745e-06,
+      "loss": 0.4907,
+      "step": 788
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.1145662847790506e-06,
+      "loss": 0.4631,
+      "step": 789
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.112111292962357e-06,
+      "loss": 0.4774,
+      "step": 790
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.109656301145663e-06,
+      "loss": 0.5062,
+      "step": 791
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.10720130932897e-06,
+      "loss": 0.49,
+      "step": 792
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.104746317512275e-06,
+      "loss": 0.4727,
+      "step": 793
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.102291325695581e-06,
+      "loss": 0.4739,
+      "step": 794
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.099836333878887e-06,
+      "loss": 0.4704,
+      "step": 795
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.097381342062193e-06,
+      "loss": 0.4823,
+      "step": 796
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.094926350245499e-06,
+      "loss": 0.4814,
+      "step": 797
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.092471358428805e-06,
+      "loss": 0.5081,
+      "step": 798
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.090016366612112e-06,
+      "loss": 0.4603,
+      "step": 799
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.087561374795418e-06,
+      "loss": 0.4965,
+      "step": 800
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.085106382978723e-06,
+      "loss": 0.4865,
+      "step": 801
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.082651391162029e-06,
+      "loss": 0.4723,
+      "step": 802
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.080196399345336e-06,
+      "loss": 0.4722,
+      "step": 803
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.077741407528642e-06,
+      "loss": 0.4899,
+      "step": 804
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.075286415711947e-06,
+      "loss": 0.4547,
+      "step": 805
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.072831423895254e-06,
+      "loss": 0.4619,
+      "step": 806
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.07037643207856e-06,
+      "loss": 0.4964,
+      "step": 807
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.067921440261866e-06,
+      "loss": 0.4626,
+      "step": 808
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.065466448445172e-06,
+      "loss": 0.5171,
+      "step": 809
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.063011456628478e-06,
+      "loss": 0.4766,
+      "step": 810
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.060556464811784e-06,
+      "loss": 0.4615,
+      "step": 811
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.05810147299509e-06,
+      "loss": 0.478,
+      "step": 812
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.055646481178396e-06,
+      "loss": 0.4923,
+      "step": 813
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.053191489361702e-06,
+      "loss": 0.4669,
+      "step": 814
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.050736497545009e-06,
+      "loss": 0.4813,
+      "step": 815
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.0482815057283145e-06,
+      "loss": 0.4771,
+      "step": 816
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.04582651391162e-06,
+      "loss": 0.4916,
+      "step": 817
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.043371522094926e-06,
+      "loss": 0.4784,
+      "step": 818
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.040916530278232e-06,
+      "loss": 0.4787,
+      "step": 819
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.0384615384615385e-06,
+      "loss": 0.4587,
+      "step": 820
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.036006546644845e-06,
+      "loss": 0.4872,
+      "step": 821
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.033551554828151e-06,
+      "loss": 0.4591,
+      "step": 822
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.031096563011457e-06,
+      "loss": 0.5044,
+      "step": 823
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.0286415711947625e-06,
+      "loss": 0.4928,
+      "step": 824
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.026186579378068e-06,
+      "loss": 0.4699,
+      "step": 825
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.023731587561375e-06,
+      "loss": 0.481,
+      "step": 826
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.021276595744681e-06,
+      "loss": 0.4668,
+      "step": 827
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.018821603927987e-06,
+      "loss": 0.4665,
+      "step": 828
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.016366612111293e-06,
+      "loss": 0.4695,
+      "step": 829
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.013911620294599e-06,
+      "loss": 0.4822,
+      "step": 830
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.0114566284779056e-06,
+      "loss": 0.4565,
+      "step": 831
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.009001636661211e-06,
+      "loss": 0.4616,
+      "step": 832
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.006546644844517e-06,
+      "loss": 0.4772,
+      "step": 833
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.004091653027823e-06,
+      "loss": 0.4776,
+      "step": 834
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.0016366612111296e-06,
+      "loss": 0.507,
+      "step": 835
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.999181669394435e-06,
+      "loss": 0.4621,
+      "step": 836
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.996726677577742e-06,
+      "loss": 0.4892,
+      "step": 837
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.994271685761048e-06,
+      "loss": 0.4734,
+      "step": 838
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.991816693944354e-06,
+      "loss": 0.4802,
+      "step": 839
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.989361702127659e-06,
+      "loss": 0.4722,
+      "step": 840
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.986906710310965e-06,
+      "loss": 0.4869,
+      "step": 841
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.984451718494272e-06,
+      "loss": 0.4953,
+      "step": 842
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.9819967266775784e-06,
+      "loss": 0.4789,
+      "step": 843
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.979541734860884e-06,
+      "loss": 0.4688,
+      "step": 844
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.97708674304419e-06,
+      "loss": 0.483,
+      "step": 845
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.974631751227496e-06,
+      "loss": 0.477,
+      "step": 846
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.972176759410802e-06,
+      "loss": 0.5073,
+      "step": 847
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.969721767594108e-06,
+      "loss": 0.4845,
+      "step": 848
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.967266775777414e-06,
+      "loss": 0.4733,
+      "step": 849
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.964811783960721e-06,
+      "loss": 0.4904,
+      "step": 850
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.9623567921440264e-06,
+      "loss": 0.4852,
+      "step": 851
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.959901800327332e-06,
+      "loss": 0.4707,
+      "step": 852
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.957446808510638e-06,
+      "loss": 0.4748,
+      "step": 853
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.954991816693945e-06,
+      "loss": 0.4992,
+      "step": 854
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.9525368248772504e-06,
+      "loss": 0.4727,
+      "step": 855
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.950081833060556e-06,
+      "loss": 0.4778,
+      "step": 856
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.947626841243863e-06,
+      "loss": 0.4884,
+      "step": 857
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.945171849427169e-06,
+      "loss": 0.4703,
+      "step": 858
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.9427168576104745e-06,
+      "loss": 0.4632,
+      "step": 859
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.940261865793781e-06,
+      "loss": 0.4992,
+      "step": 860
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.937806873977087e-06,
+      "loss": 0.4824,
+      "step": 861
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.935351882160393e-06,
+      "loss": 0.4753,
+      "step": 862
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.9328968903436985e-06,
+      "loss": 0.5077,
+      "step": 863
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.930441898527005e-06,
+      "loss": 0.4418,
+      "step": 864
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.927986906710312e-06,
+      "loss": 0.5016,
+      "step": 865
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.9255319148936175e-06,
+      "loss": 0.4688,
+      "step": 866
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.923076923076923e-06,
+      "loss": 0.4628,
+      "step": 867
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.920621931260229e-06,
+      "loss": 0.4857,
+      "step": 868
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.918166939443535e-06,
+      "loss": 0.4734,
+      "step": 869
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.915711947626841e-06,
+      "loss": 0.4881,
+      "step": 870
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.913256955810147e-06,
+      "loss": 0.4726,
+      "step": 871
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.910801963993454e-06,
+      "loss": 0.4906,
+      "step": 872
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.90834697217676e-06,
+      "loss": 0.452,
+      "step": 873
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.9058919803600655e-06,
+      "loss": 0.4803,
+      "step": 874
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.903436988543371e-06,
+      "loss": 0.4707,
+      "step": 875
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.900981996726678e-06,
+      "loss": 0.4448,
+      "step": 876
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.898527004909984e-06,
+      "loss": 0.4597,
+      "step": 877
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.8960720130932895e-06,
+      "loss": 0.4912,
+      "step": 878
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.893617021276596e-06,
+      "loss": 0.5108,
+      "step": 879
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.891162029459902e-06,
+      "loss": 0.4792,
+      "step": 880
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.888707037643208e-06,
+      "loss": 0.4777,
+      "step": 881
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.886252045826514e-06,
+      "loss": 0.4539,
+      "step": 882
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.88379705400982e-06,
+      "loss": 0.4867,
+      "step": 883
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.881342062193126e-06,
+      "loss": 0.5004,
+      "step": 884
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.878887070376432e-06,
+      "loss": 0.4615,
+      "step": 885
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.876432078559738e-06,
+      "loss": 0.4953,
+      "step": 886
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.873977086743044e-06,
+      "loss": 0.4836,
+      "step": 887
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.871522094926351e-06,
+      "loss": 0.4576,
+      "step": 888
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.869067103109657e-06,
+      "loss": 0.5017,
+      "step": 889
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.866612111292962e-06,
+      "loss": 0.4725,
+      "step": 890
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.864157119476268e-06,
+      "loss": 0.5028,
+      "step": 891
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.861702127659574e-06,
+      "loss": 0.4948,
+      "step": 892
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.859247135842881e-06,
+      "loss": 0.4699,
+      "step": 893
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.856792144026187e-06,
+      "loss": 0.4721,
+      "step": 894
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.854337152209493e-06,
+      "loss": 0.4692,
+      "step": 895
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.851882160392799e-06,
+      "loss": 0.5033,
+      "step": 896
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.849427168576105e-06,
+      "loss": 0.5047,
+      "step": 897
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.84697217675941e-06,
+      "loss": 0.4465,
+      "step": 898
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.844517184942717e-06,
+      "loss": 0.4784,
+      "step": 899
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.842062193126023e-06,
+      "loss": 0.507,
+      "step": 900
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.8396072013093295e-06,
+      "loss": 0.4865,
+      "step": 901
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.837152209492635e-06,
+      "loss": 0.471,
+      "step": 902
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.834697217675941e-06,
+      "loss": 0.5058,
+      "step": 903
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.832242225859248e-06,
+      "loss": 0.4874,
+      "step": 904
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.8297872340425535e-06,
+      "loss": 0.4912,
+      "step": 905
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.827332242225859e-06,
+      "loss": 0.4708,
+      "step": 906
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.824877250409165e-06,
+      "loss": 0.486,
+      "step": 907
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.822422258592472e-06,
+      "loss": 0.4864,
+      "step": 908
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.8199672667757775e-06,
+      "loss": 0.4617,
+      "step": 909
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.817512274959084e-06,
+      "loss": 0.4907,
+      "step": 910
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.81505728314239e-06,
+      "loss": 0.4726,
+      "step": 911
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.8126022913256957e-06,
+      "loss": 0.4796,
+      "step": 912
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.8101472995090015e-06,
+      "loss": 0.4908,
+      "step": 913
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.8076923076923077e-06,
+      "loss": 0.4625,
+      "step": 914
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.8052373158756135e-06,
+      "loss": 0.471,
+      "step": 915
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.80278232405892e-06,
+      "loss": 0.5052,
+      "step": 916
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.800327332242226e-06,
+      "loss": 0.4936,
+      "step": 917
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.797872340425532e-06,
+      "loss": 0.4953,
+      "step": 918
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.795417348608838e-06,
+      "loss": 0.4877,
+      "step": 919
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.7929623567921437e-06,
+      "loss": 0.48,
+      "step": 920
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.7905073649754503e-06,
+      "loss": 0.4813,
+      "step": 921
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.7880523731587565e-06,
+      "loss": 0.4748,
+      "step": 922
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.7855973813420623e-06,
+      "loss": 0.4733,
+      "step": 923
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.783142389525368e-06,
+      "loss": 0.4799,
+      "step": 924
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.7806873977086743e-06,
+      "loss": 0.49,
+      "step": 925
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.77823240589198e-06,
+      "loss": 0.4921,
+      "step": 926
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.7757774140752868e-06,
+      "loss": 0.4776,
+      "step": 927
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.7733224222585926e-06,
+      "loss": 0.4509,
+      "step": 928
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.7708674304418988e-06,
+      "loss": 0.482,
+      "step": 929
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.7684124386252046e-06,
+      "loss": 0.4974,
+      "step": 930
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.7659574468085103e-06,
+      "loss": 0.4859,
+      "step": 931
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.763502454991817e-06,
+      "loss": 0.4673,
+      "step": 932
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.761047463175123e-06,
+      "loss": 0.5155,
+      "step": 933
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.758592471358429e-06,
+      "loss": 0.4862,
+      "step": 934
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.7561374795417348e-06,
+      "loss": 0.4948,
+      "step": 935
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.753682487725041e-06,
+      "loss": 0.4822,
+      "step": 936
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.7512274959083468e-06,
+      "loss": 0.4866,
+      "step": 937
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.7487725040916534e-06,
+      "loss": 0.457,
+      "step": 938
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.746317512274959e-06,
+      "loss": 0.4671,
+      "step": 939
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.7438625204582654e-06,
+      "loss": 0.4786,
+      "step": 940
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.741407528641571e-06,
+      "loss": 0.4881,
+      "step": 941
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.738952536824877e-06,
+      "loss": 0.4773,
+      "step": 942
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.736497545008183e-06,
+      "loss": 0.498,
+      "step": 943
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.73404255319149e-06,
+      "loss": 0.5026,
+      "step": 944
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.7315875613747956e-06,
+      "loss": 0.4534,
+      "step": 945
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.7291325695581014e-06,
+      "loss": 0.4746,
+      "step": 946
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.7266775777414076e-06,
+      "loss": 0.4843,
+      "step": 947
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.7242225859247134e-06,
+      "loss": 0.4983,
+      "step": 948
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.72176759410802e-06,
+      "loss": 0.4604,
+      "step": 949
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.719312602291326e-06,
+      "loss": 0.4853,
+      "step": 950
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.716857610474632e-06,
+      "loss": 0.4561,
+      "step": 951
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.714402618657938e-06,
+      "loss": 0.4613,
+      "step": 952
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.7119476268412436e-06,
+      "loss": 0.4574,
+      "step": 953
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.70949263502455e-06,
+      "loss": 0.4965,
+      "step": 954
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.7070376432078565e-06,
+      "loss": 0.4717,
+      "step": 955
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.7045826513911623e-06,
+      "loss": 0.4588,
+      "step": 956
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.702127659574468e-06,
+      "loss": 0.4653,
+      "step": 957
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.6996726677577743e-06,
+      "loss": 0.4654,
+      "step": 958
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.69721767594108e-06,
+      "loss": 0.4679,
+      "step": 959
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.6947626841243867e-06,
+      "loss": 0.4905,
+      "step": 960
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.6923076923076925e-06,
+      "loss": 0.486,
+      "step": 961
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.6898527004909987e-06,
+      "loss": 0.4831,
+      "step": 962
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.6873977086743045e-06,
+      "loss": 0.5004,
+      "step": 963
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.6849427168576103e-06,
+      "loss": 0.4644,
+      "step": 964
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.6824877250409165e-06,
+      "loss": 0.4739,
+      "step": 965
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.680032733224223e-06,
+      "loss": 0.4479,
+      "step": 966
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.677577741407529e-06,
+      "loss": 0.4712,
+      "step": 967
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.6751227495908347e-06,
+      "loss": 0.4741,
+      "step": 968
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.672667757774141e-06,
+      "loss": 0.479,
+      "step": 969
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.6702127659574467e-06,
+      "loss": 0.477,
+      "step": 970
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.6677577741407525e-06,
+      "loss": 0.4732,
+      "step": 971
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.665302782324059e-06,
+      "loss": 0.4798,
+      "step": 972
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.6628477905073654e-06,
+      "loss": 0.4721,
+      "step": 973
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.660392798690671e-06,
+      "loss": 0.4597,
+      "step": 974
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.657937806873977e-06,
+      "loss": 0.4787,
+      "step": 975
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.655482815057283e-06,
+      "loss": 0.4738,
+      "step": 976
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.65302782324059e-06,
+      "loss": 0.4693,
+      "step": 977
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.6505728314238956e-06,
+      "loss": 0.5194,
+      "step": 978
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.6481178396072014e-06,
+      "loss": 0.4583,
+      "step": 979
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.6456628477905076e-06,
+      "loss": 0.45,
+      "step": 980
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.6432078559738134e-06,
+      "loss": 0.4693,
+      "step": 981
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.640752864157119e-06,
+      "loss": 0.4641,
+      "step": 982
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.638297872340426e-06,
+      "loss": 0.4614,
+      "step": 983
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.635842880523732e-06,
+      "loss": 0.4821,
+      "step": 984
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.633387888707038e-06,
+      "loss": 0.5112,
+      "step": 985
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.6309328968903436e-06,
+      "loss": 0.4715,
+      "step": 986
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.62847790507365e-06,
+      "loss": 0.4826,
+      "step": 987
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.6260229132569564e-06,
+      "loss": 0.4499,
+      "step": 988
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.6235679214402622e-06,
+      "loss": 0.4621,
+      "step": 989
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.621112929623568e-06,
+      "loss": 0.5173,
+      "step": 990
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.6186579378068742e-06,
+      "loss": 0.5014,
+      "step": 991
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.61620294599018e-06,
+      "loss": 0.5002,
+      "step": 992
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.613747954173486e-06,
+      "loss": 0.4628,
+      "step": 993
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.6112929623567924e-06,
+      "loss": 0.4677,
+      "step": 994
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.6088379705400987e-06,
+      "loss": 0.4951,
+      "step": 995
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.6063829787234044e-06,
+      "loss": 0.4825,
+      "step": 996
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.6039279869067102e-06,
+      "loss": 0.4794,
+      "step": 997
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.6014729950900164e-06,
+      "loss": 0.4623,
+      "step": 998
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.5990180032733222e-06,
+      "loss": 0.4798,
+      "step": 999
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.596563011456629e-06,
+      "loss": 0.4828,
+      "step": 1000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.5941080196399347e-06,
+      "loss": 0.4579,
+      "step": 1001
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.591653027823241e-06,
+      "loss": 0.4847,
+      "step": 1002
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.5891980360065467e-06,
+      "loss": 0.4666,
+      "step": 1003
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.5867430441898525e-06,
+      "loss": 0.4535,
+      "step": 1004
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.584288052373159e-06,
+      "loss": 0.4721,
+      "step": 1005
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.581833060556465e-06,
+      "loss": 0.4598,
+      "step": 1006
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.579378068739771e-06,
+      "loss": 0.4629,
+      "step": 1007
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.576923076923077e-06,
+      "loss": 0.4869,
+      "step": 1008
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.574468085106383e-06,
+      "loss": 0.4506,
+      "step": 1009
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.572013093289689e-06,
+      "loss": 0.506,
+      "step": 1010
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.5695581014729955e-06,
+      "loss": 0.5079,
+      "step": 1011
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.5671031096563013e-06,
+      "loss": 0.4715,
+      "step": 1012
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.564648117839607e-06,
+      "loss": 0.4887,
+      "step": 1013
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.5621931260229133e-06,
+      "loss": 0.4997,
+      "step": 1014
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.559738134206219e-06,
+      "loss": 0.4916,
+      "step": 1015
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.5572831423895257e-06,
+      "loss": 0.4692,
+      "step": 1016
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.5548281505728315e-06,
+      "loss": 0.4635,
+      "step": 1017
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.5523731587561377e-06,
+      "loss": 0.4728,
+      "step": 1018
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.5499181669394435e-06,
+      "loss": 0.4884,
+      "step": 1019
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.5474631751227497e-06,
+      "loss": 0.4942,
+      "step": 1020
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.5450081833060555e-06,
+      "loss": 0.4652,
+      "step": 1021
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.542553191489362e-06,
+      "loss": 0.4977,
+      "step": 1022
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.540098199672668e-06,
+      "loss": 0.4922,
+      "step": 1023
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.5376432078559737e-06,
+      "loss": 0.4664,
+      "step": 1024
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.53518821603928e-06,
+      "loss": 0.4542,
+      "step": 1025
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.5327332242225858e-06,
+      "loss": 0.4866,
+      "step": 1026
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.530278232405892e-06,
+      "loss": 0.4751,
+      "step": 1027
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.527823240589198e-06,
+      "loss": 0.4706,
+      "step": 1028
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.5253682487725044e-06,
+      "loss": 0.478,
+      "step": 1029
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.52291325695581e-06,
+      "loss": 0.4845,
+      "step": 1030
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.520458265139116e-06,
+      "loss": 0.4943,
+      "step": 1031
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.518003273322422e-06,
+      "loss": 0.4653,
+      "step": 1032
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.515548281505729e-06,
+      "loss": 0.4557,
+      "step": 1033
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.5130932896890346e-06,
+      "loss": 0.5015,
+      "step": 1034
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.5106382978723404e-06,
+      "loss": 0.4975,
+      "step": 1035
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.5081833060556466e-06,
+      "loss": 0.4811,
+      "step": 1036
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.5057283142389524e-06,
+      "loss": 0.4786,
+      "step": 1037
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.503273322422258e-06,
+      "loss": 0.4518,
+      "step": 1038
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.500818330605565e-06,
+      "loss": 0.4999,
+      "step": 1039
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.498363338788871e-06,
+      "loss": 0.4833,
+      "step": 1040
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.495908346972177e-06,
+      "loss": 0.4901,
+      "step": 1041
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.4934533551554826e-06,
+      "loss": 0.4566,
+      "step": 1042
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.490998363338789e-06,
+      "loss": 0.4637,
+      "step": 1043
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.4885433715220946e-06,
+      "loss": 0.5169,
+      "step": 1044
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.4860883797054013e-06,
+      "loss": 0.4851,
+      "step": 1045
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.483633387888707e-06,
+      "loss": 0.4709,
+      "step": 1046
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.4811783960720133e-06,
+      "loss": 0.483,
+      "step": 1047
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.478723404255319e-06,
+      "loss": 0.5135,
+      "step": 1048
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.476268412438625e-06,
+      "loss": 0.4904,
+      "step": 1049
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.4738134206219315e-06,
+      "loss": 0.4525,
+      "step": 1050
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.4713584288052377e-06,
+      "loss": 0.4775,
+      "step": 1051
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.4689034369885435e-06,
+      "loss": 0.4951,
+      "step": 1052
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.4664484451718493e-06,
+      "loss": 0.4614,
+      "step": 1053
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.4639934533551555e-06,
+      "loss": 0.4946,
+      "step": 1054
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.4615384615384613e-06,
+      "loss": 0.4478,
+      "step": 1055
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.459083469721768e-06,
+      "loss": 0.486,
+      "step": 1056
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.4566284779050737e-06,
+      "loss": 0.5032,
+      "step": 1057
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.45417348608838e-06,
+      "loss": 0.4604,
+      "step": 1058
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.4517184942716857e-06,
+      "loss": 0.4811,
+      "step": 1059
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.4492635024549915e-06,
+      "loss": 0.4769,
+      "step": 1060
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.446808510638298e-06,
+      "loss": 0.4715,
+      "step": 1061
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.4443535188216043e-06,
+      "loss": 0.4665,
+      "step": 1062
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.44189852700491e-06,
+      "loss": 0.4848,
+      "step": 1063
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.439443535188216e-06,
+      "loss": 0.4609,
+      "step": 1064
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.436988543371522e-06,
+      "loss": 0.457,
+      "step": 1065
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.434533551554828e-06,
+      "loss": 0.4935,
+      "step": 1066
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.4320785597381346e-06,
+      "loss": 0.4967,
+      "step": 1067
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.4296235679214403e-06,
+      "loss": 0.4838,
+      "step": 1068
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.4271685761047466e-06,
+      "loss": 0.4853,
+      "step": 1069
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.4247135842880523e-06,
+      "loss": 0.4901,
+      "step": 1070
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.422258592471358e-06,
+      "loss": 0.4915,
+      "step": 1071
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.4198036006546643e-06,
+      "loss": 0.489,
+      "step": 1072
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.417348608837971e-06,
+      "loss": 0.4559,
+      "step": 1073
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.4148936170212768e-06,
+      "loss": 0.4575,
+      "step": 1074
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.4124386252045826e-06,
+      "loss": 0.4699,
+      "step": 1075
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.4099836333878888e-06,
+      "loss": 0.4779,
+      "step": 1076
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.4075286415711946e-06,
+      "loss": 0.4761,
+      "step": 1077
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.405073649754501e-06,
+      "loss": 0.4689,
+      "step": 1078
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.402618657937807e-06,
+      "loss": 0.4779,
+      "step": 1079
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.400163666121113e-06,
+      "loss": 0.484,
+      "step": 1080
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.397708674304419e-06,
+      "loss": 0.4972,
+      "step": 1081
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.3952536824877248e-06,
+      "loss": 0.473,
+      "step": 1082
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.392798690671031e-06,
+      "loss": 0.4696,
+      "step": 1083
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.3903436988543376e-06,
+      "loss": 0.4906,
+      "step": 1084
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.3878887070376434e-06,
+      "loss": 0.4836,
+      "step": 1085
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.385433715220949e-06,
+      "loss": 0.4949,
+      "step": 1086
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.3829787234042554e-06,
+      "loss": 0.4631,
+      "step": 1087
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.3805237315875612e-06,
+      "loss": 0.5138,
+      "step": 1088
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.378068739770868e-06,
+      "loss": 0.4701,
+      "step": 1089
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.3756137479541736e-06,
+      "loss": 0.4711,
+      "step": 1090
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.37315875613748e-06,
+      "loss": 0.4593,
+      "step": 1091
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.3707037643207856e-06,
+      "loss": 0.4855,
+      "step": 1092
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.3682487725040914e-06,
+      "loss": 0.4854,
+      "step": 1093
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.3657937806873976e-06,
+      "loss": 0.4766,
+      "step": 1094
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.3633387888707043e-06,
+      "loss": 0.4826,
+      "step": 1095
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.36088379705401e-06,
+      "loss": 0.4576,
+      "step": 1096
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.358428805237316e-06,
+      "loss": 0.4857,
+      "step": 1097
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.355973813420622e-06,
+      "loss": 0.4711,
+      "step": 1098
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.353518821603928e-06,
+      "loss": 0.4867,
+      "step": 1099
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.3510638297872336e-06,
+      "loss": 0.5131,
+      "step": 1100
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.3486088379705403e-06,
+      "loss": 0.5161,
+      "step": 1101
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.3461538461538465e-06,
+      "loss": 0.4848,
+      "step": 1102
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.3436988543371523e-06,
+      "loss": 0.4951,
+      "step": 1103
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.341243862520458e-06,
+      "loss": 0.482,
+      "step": 1104
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.3387888707037643e-06,
+      "loss": 0.4987,
+      "step": 1105
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.336333878887071e-06,
+      "loss": 0.4757,
+      "step": 1106
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.3338788870703767e-06,
+      "loss": 0.4677,
+      "step": 1107
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.3314238952536825e-06,
+      "loss": 0.4537,
+      "step": 1108
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.3289689034369887e-06,
+      "loss": 0.4744,
+      "step": 1109
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.3265139116202945e-06,
+      "loss": 0.4847,
+      "step": 1110
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.3240589198036003e-06,
+      "loss": 0.46,
+      "step": 1111
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.321603927986907e-06,
+      "loss": 0.4773,
+      "step": 1112
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.319148936170213e-06,
+      "loss": 0.496,
+      "step": 1113
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.316693944353519e-06,
+      "loss": 0.4579,
+      "step": 1114
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.3142389525368247e-06,
+      "loss": 0.4891,
+      "step": 1115
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.311783960720131e-06,
+      "loss": 0.4553,
+      "step": 1116
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.3093289689034376e-06,
+      "loss": 0.4938,
+      "step": 1117
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.3068739770867434e-06,
+      "loss": 0.4687,
+      "step": 1118
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.304418985270049e-06,
+      "loss": 0.4761,
+      "step": 1119
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.3019639934533554e-06,
+      "loss": 0.4686,
+      "step": 1120
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.299509001636661e-06,
+      "loss": 0.4566,
+      "step": 1121
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.297054009819967e-06,
+      "loss": 0.5076,
+      "step": 1122
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.2945990180032736e-06,
+      "loss": 0.4804,
+      "step": 1123
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.29214402618658e-06,
+      "loss": 0.4887,
+      "step": 1124
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.2896890343698856e-06,
+      "loss": 0.4785,
+      "step": 1125
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.2872340425531914e-06,
+      "loss": 0.4816,
+      "step": 1126
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.2847790507364976e-06,
+      "loss": 0.4761,
+      "step": 1127
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.2823240589198034e-06,
+      "loss": 0.4784,
+      "step": 1128
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.27986906710311e-06,
+      "loss": 0.4786,
+      "step": 1129
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.277414075286416e-06,
+      "loss": 0.4985,
+      "step": 1130
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.274959083469722e-06,
+      "loss": 0.4825,
+      "step": 1131
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.272504091653028e-06,
+      "loss": 0.4888,
+      "step": 1132
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.2700490998363336e-06,
+      "loss": 0.4789,
+      "step": 1133
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.2675941080196402e-06,
+      "loss": 0.4661,
+      "step": 1134
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.2651391162029464e-06,
+      "loss": 0.4977,
+      "step": 1135
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.2626841243862522e-06,
+      "loss": 0.4815,
+      "step": 1136
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.260229132569558e-06,
+      "loss": 0.486,
+      "step": 1137
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.2577741407528642e-06,
+      "loss": 0.4665,
+      "step": 1138
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.25531914893617e-06,
+      "loss": 0.4904,
+      "step": 1139
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.2528641571194767e-06,
+      "loss": 0.4781,
+      "step": 1140
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.2504091653027825e-06,
+      "loss": 0.4828,
+      "step": 1141
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.2479541734860887e-06,
+      "loss": 0.5269,
+      "step": 1142
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.2454991816693945e-06,
+      "loss": 0.504,
+      "step": 1143
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.2430441898527002e-06,
+      "loss": 0.4972,
+      "step": 1144
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.240589198036007e-06,
+      "loss": 0.4622,
+      "step": 1145
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.238134206219313e-06,
+      "loss": 0.472,
+      "step": 1146
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.235679214402619e-06,
+      "loss": 0.4874,
+      "step": 1147
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.2332242225859247e-06,
+      "loss": 0.4581,
+      "step": 1148
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.230769230769231e-06,
+      "loss": 0.5167,
+      "step": 1149
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.2283142389525367e-06,
+      "loss": 0.4702,
+      "step": 1150
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.2258592471358433e-06,
+      "loss": 0.4856,
+      "step": 1151
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.223404255319149e-06,
+      "loss": 0.4787,
+      "step": 1152
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.2209492635024553e-06,
+      "loss": 0.5037,
+      "step": 1153
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.218494271685761e-06,
+      "loss": 0.4918,
+      "step": 1154
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.216039279869067e-06,
+      "loss": 0.4646,
+      "step": 1155
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.213584288052373e-06,
+      "loss": 0.4964,
+      "step": 1156
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.2111292962356793e-06,
+      "loss": 0.4849,
+      "step": 1157
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.2086743044189855e-06,
+      "loss": 0.4752,
+      "step": 1158
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.2062193126022913e-06,
+      "loss": 0.4676,
+      "step": 1159
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.2037643207855975e-06,
+      "loss": 0.4749,
+      "step": 1160
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.2013093289689033e-06,
+      "loss": 0.4921,
+      "step": 1161
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.19885433715221e-06,
+      "loss": 0.4719,
+      "step": 1162
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.1963993453355157e-06,
+      "loss": 0.4831,
+      "step": 1163
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.1939443535188215e-06,
+      "loss": 0.49,
+      "step": 1164
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.1914893617021277e-06,
+      "loss": 0.4851,
+      "step": 1165
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.1890343698854335e-06,
+      "loss": 0.4883,
+      "step": 1166
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.1865793780687398e-06,
+      "loss": 0.4855,
+      "step": 1167
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.184124386252046e-06,
+      "loss": 0.4803,
+      "step": 1168
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.181669394435352e-06,
+      "loss": 0.4651,
+      "step": 1169
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.179214402618658e-06,
+      "loss": 0.488,
+      "step": 1170
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.176759410801964e-06,
+      "loss": 0.5117,
+      "step": 1171
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.17430441898527e-06,
+      "loss": 0.4931,
+      "step": 1172
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.1718494271685766e-06,
+      "loss": 0.4887,
+      "step": 1173
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.1693944353518824e-06,
+      "loss": 0.4811,
+      "step": 1174
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.166939443535188e-06,
+      "loss": 0.4821,
+      "step": 1175
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.1644844517184944e-06,
+      "loss": 0.4951,
+      "step": 1176
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.1620294599018e-06,
+      "loss": 0.4753,
+      "step": 1177
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.1595744680851064e-06,
+      "loss": 0.5211,
+      "step": 1178
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.1571194762684126e-06,
+      "loss": 0.4757,
+      "step": 1179
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.154664484451719e-06,
+      "loss": 0.4949,
+      "step": 1180
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.1522094926350246e-06,
+      "loss": 0.4676,
+      "step": 1181
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.1497545008183304e-06,
+      "loss": 0.4792,
+      "step": 1182
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.1472995090016366e-06,
+      "loss": 0.4758,
+      "step": 1183
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.1448445171849424e-06,
+      "loss": 0.5046,
+      "step": 1184
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.142389525368249e-06,
+      "loss": 0.4883,
+      "step": 1185
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.139934533551555e-06,
+      "loss": 0.4809,
+      "step": 1186
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.137479541734861e-06,
+      "loss": 0.4946,
+      "step": 1187
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.135024549918167e-06,
+      "loss": 0.4764,
+      "step": 1188
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.1325695581014726e-06,
+      "loss": 0.482,
+      "step": 1189
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.1301145662847793e-06,
+      "loss": 0.5175,
+      "step": 1190
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.1276595744680855e-06,
+      "loss": 0.4407,
+      "step": 1191
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.1252045826513913e-06,
+      "loss": 0.4654,
+      "step": 1192
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.122749590834697e-06,
+      "loss": 0.4698,
+      "step": 1193
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.1202945990180033e-06,
+      "loss": 0.4924,
+      "step": 1194
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.117839607201309e-06,
+      "loss": 0.4757,
+      "step": 1195
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.1153846153846157e-06,
+      "loss": 0.4904,
+      "step": 1196
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.1129296235679215e-06,
+      "loss": 0.4539,
+      "step": 1197
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.1104746317512277e-06,
+      "loss": 0.4809,
+      "step": 1198
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.1080196399345335e-06,
+      "loss": 0.4801,
+      "step": 1199
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.1055646481178393e-06,
+      "loss": 0.5018,
+      "step": 1200
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.103109656301146e-06,
+      "loss": 0.5093,
+      "step": 1201
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.100654664484452e-06,
+      "loss": 0.4945,
+      "step": 1202
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.098199672667758e-06,
+      "loss": 0.4923,
+      "step": 1203
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.0957446808510637e-06,
+      "loss": 0.4712,
+      "step": 1204
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.09328968903437e-06,
+      "loss": 0.4656,
+      "step": 1205
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.0908346972176757e-06,
+      "loss": 0.4831,
+      "step": 1206
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.0883797054009823e-06,
+      "loss": 0.4776,
+      "step": 1207
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.085924713584288e-06,
+      "loss": 0.4848,
+      "step": 1208
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.0834697217675943e-06,
+      "loss": 0.4738,
+      "step": 1209
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.0810147299509e-06,
+      "loss": 0.4893,
+      "step": 1210
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 3.078559738134206e-06,
+      "loss": 0.5125,
+      "step": 1211
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.076104746317512e-06,
+      "loss": 0.4971,
+      "step": 1212
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.0736497545008188e-06,
+      "loss": 0.4862,
+      "step": 1213
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.0711947626841246e-06,
+      "loss": 0.4676,
+      "step": 1214
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.0687397708674303e-06,
+      "loss": 0.4873,
+      "step": 1215
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.0662847790507366e-06,
+      "loss": 0.4666,
+      "step": 1216
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 3.0638297872340424e-06,
+      "loss": 0.4968,
+      "step": 1217
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.061374795417349e-06,
+      "loss": 0.5053,
+      "step": 1218
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.0589198036006548e-06,
+      "loss": 0.4938,
+      "step": 1219
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.056464811783961e-06,
+      "loss": 0.492,
+      "step": 1220
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.0540098199672668e-06,
+      "loss": 0.4789,
+      "step": 1221
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.0515548281505726e-06,
+      "loss": 0.469,
+      "step": 1222
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.0490998363338788e-06,
+      "loss": 0.4865,
+      "step": 1223
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.0466448445171854e-06,
+      "loss": 0.4652,
+      "step": 1224
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.044189852700491e-06,
+      "loss": 0.462,
+      "step": 1225
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.041734860883797e-06,
+      "loss": 0.476,
+      "step": 1226
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.039279869067103e-06,
+      "loss": 0.4763,
+      "step": 1227
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.036824877250409e-06,
+      "loss": 0.4861,
+      "step": 1228
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 3.0343698854337156e-06,
+      "loss": 0.4782,
+      "step": 1229
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.0319148936170214e-06,
+      "loss": 0.4809,
+      "step": 1230
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.0294599018003276e-06,
+      "loss": 0.4884,
+      "step": 1231
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.0270049099836334e-06,
+      "loss": 0.4758,
+      "step": 1232
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 3.0245499181669392e-06,
+      "loss": 0.5012,
+      "step": 1233
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2464,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 2.068501915027007e+19,
+  "trial_name": null,
+  "trial_params": null
+}