diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,7560 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9996017522899243,
+  "eval_steps": 500,
+  "global_step": 1255,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 7.894736842105263e-07,
+      "loss": 1.788,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.5789473684210526e-06,
+      "loss": 1.6468,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.368421052631579e-06,
+      "loss": 1.6049,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.157894736842105e-06,
+      "loss": 1.635,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.9473684210526315e-06,
+      "loss": 1.6297,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.736842105263158e-06,
+      "loss": 1.576,
+      "step": 6
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.526315789473684e-06,
+      "loss": 1.5663,
+      "step": 7
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.31578947368421e-06,
+      "loss": 1.5252,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.105263157894737e-06,
+      "loss": 1.6506,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.894736842105263e-06,
+      "loss": 1.6023,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.68421052631579e-06,
+      "loss": 1.422,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.473684210526315e-06,
+      "loss": 1.5656,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.0263157894736843e-05,
+      "loss": 1.5206,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.1052631578947368e-05,
+      "loss": 1.4842,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.1842105263157895e-05,
+      "loss": 1.4483,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.263157894736842e-05,
+      "loss": 1.4382,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.3421052631578948e-05,
+      "loss": 1.492,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.4210526315789473e-05,
+      "loss": 1.3404,
+      "step": 18
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5e-05,
+      "loss": 1.4589,
+      "step": 19
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5789473684210526e-05,
+      "loss": 1.4015,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.6578947368421053e-05,
+      "loss": 1.2486,
+      "step": 21
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.736842105263158e-05,
+      "loss": 1.3298,
+      "step": 22
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.8157894736842107e-05,
+      "loss": 1.3353,
+      "step": 23
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.894736842105263e-05,
+      "loss": 1.3825,
+      "step": 24
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9736842105263158e-05,
+      "loss": 1.3048,
+      "step": 25
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.0526315789473685e-05,
+      "loss": 1.3904,
+      "step": 26
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.1315789473684212e-05,
+      "loss": 1.304,
+      "step": 27
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.2105263157894736e-05,
+      "loss": 1.2266,
+      "step": 28
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.2894736842105263e-05,
+      "loss": 1.2776,
+      "step": 29
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.368421052631579e-05,
+      "loss": 1.2053,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.4473684210526318e-05,
+      "loss": 1.2372,
+      "step": 31
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.526315789473684e-05,
+      "loss": 1.2616,
+      "step": 32
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.605263157894737e-05,
+      "loss": 1.2076,
+      "step": 33
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.6842105263157896e-05,
+      "loss": 1.2036,
+      "step": 34
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.7631578947368423e-05,
+      "loss": 1.2481,
+      "step": 35
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.8421052631578946e-05,
+      "loss": 1.2431,
+      "step": 36
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.9210526315789474e-05,
+      "loss": 1.1937,
+      "step": 37
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3e-05,
+      "loss": 1.2575,
+      "step": 38
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.9999950021915023e-05,
+      "loss": 1.2368,
+      "step": 39
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.9999800087993127e-05,
+      "loss": 1.2664,
+      "step": 40
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.999955019923344e-05,
+      "loss": 1.1121,
+      "step": 41
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.999920035730115e-05,
+      "loss": 1.1825,
+      "step": 42
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.9998750564527518e-05,
+      "loss": 1.17,
+      "step": 43
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.999820082390985e-05,
+      "loss": 1.175,
+      "step": 44
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9997551139111472e-05,
+      "loss": 1.2812,
+      "step": 45
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9996801514461723e-05,
+      "loss": 1.2321,
+      "step": 46
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9995951954955903e-05,
+      "loss": 1.1229,
+      "step": 47
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9995002466255275e-05,
+      "loss": 1.1741,
+      "step": 48
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.999395305468697e-05,
+      "loss": 1.1601,
+      "step": 49
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9992803727244006e-05,
+      "loss": 1.1887,
+      "step": 50
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9991554491585217e-05,
+      "loss": 1.1192,
+      "step": 51
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9990205356035175e-05,
+      "loss": 1.0924,
+      "step": 52
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9988756329584185e-05,
+      "loss": 1.2883,
+      "step": 53
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9987207421888187e-05,
+      "loss": 1.1356,
+      "step": 54
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9985558643268707e-05,
+      "loss": 1.1007,
+      "step": 55
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.9983810004712783e-05,
+      "loss": 1.0989,
+      "step": 56
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9981961517872897e-05,
+      "loss": 1.093,
+      "step": 57
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.998001319506689e-05,
+      "loss": 1.1173,
+      "step": 58
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9977965049277894e-05,
+      "loss": 1.1725,
+      "step": 59
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9975817094154226e-05,
+      "loss": 1.1395,
+      "step": 60
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9973569344009308e-05,
+      "loss": 1.1729,
+      "step": 61
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9971221813821574e-05,
+      "loss": 1.1524,
+      "step": 62
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9968774519234374e-05,
+      "loss": 1.1547,
+      "step": 63
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9966227476555842e-05,
+      "loss": 1.0897,
+      "step": 64
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.996358070275883e-05,
+      "loss": 1.1039,
+      "step": 65
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9960834215480753e-05,
+      "loss": 1.1224,
+      "step": 66
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.995798803302351e-05,
+      "loss": 1.1798,
+      "step": 67
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.995504217435333e-05,
+      "loss": 1.1363,
+      "step": 68
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9951996659100666e-05,
+      "loss": 1.0771,
+      "step": 69
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9948851507560048e-05,
+      "loss": 1.1228,
+      "step": 70
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.994560674068997e-05,
+      "loss": 1.1815,
+      "step": 71
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9942262380112722e-05,
+      "loss": 1.1178,
+      "step": 72
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9938818448114274e-05,
+      "loss": 1.1174,
+      "step": 73
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9935274967644117e-05,
+      "loss": 1.1426,
+      "step": 74
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.993163196231508e-05,
+      "loss": 1.0569,
+      "step": 75
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9927889456403238e-05,
+      "loss": 1.1451,
+      "step": 76
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.992404747484769e-05,
+      "loss": 1.0541,
+      "step": 77
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.992010604325042e-05,
+      "loss": 1.1749,
+      "step": 78
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9916065187876113e-05,
+      "loss": 1.0359,
+      "step": 79
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9911924935652014e-05,
+      "loss": 1.0685,
+      "step": 80
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.9907685314167692e-05,
+      "loss": 1.0912,
+      "step": 81
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.990334635167491e-05,
+      "loss": 1.0654,
+      "step": 82
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9898908077087406e-05,
+      "loss": 1.1598,
+      "step": 83
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9894370519980707e-05,
+      "loss": 1.1121,
+      "step": 84
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9889733710591928e-05,
+      "loss": 1.126,
+      "step": 85
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9884997679819598e-05,
+      "loss": 1.1072,
+      "step": 86
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9880162459223403e-05,
+      "loss": 1.0508,
+      "step": 87
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.987522808102403e-05,
+      "loss": 1.076,
+      "step": 88
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.98701945781029e-05,
+      "loss": 1.1565,
+      "step": 89
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9865061984002007e-05,
+      "loss": 1.1658,
+      "step": 90
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9859830332923637e-05,
+      "loss": 1.0659,
+      "step": 91
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9854499659730182e-05,
+      "loss": 1.0217,
+      "step": 92
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9849069999943886e-05,
+      "loss": 1.0959,
+      "step": 93
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.9843541389746617e-05,
+      "loss": 1.041,
+      "step": 94
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.983791386597962e-05,
+      "loss": 1.0572,
+      "step": 95
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9832187466143274e-05,
+      "loss": 1.0609,
+      "step": 96
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.982636222839685e-05,
+      "loss": 1.0127,
+      "step": 97
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9820438191558243e-05,
+      "loss": 1.0477,
+      "step": 98
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9814415395103724e-05,
+      "loss": 1.0464,
+      "step": 99
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9808293879167664e-05,
+      "loss": 1.07,
+      "step": 100
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9802073684542284e-05,
+      "loss": 1.0456,
+      "step": 101
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9795754852677383e-05,
+      "loss": 1.0769,
+      "step": 102
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9789337425680027e-05,
+      "loss": 1.0964,
+      "step": 103
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9782821446314328e-05,
+      "loss": 1.1342,
+      "step": 104
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9776206958001095e-05,
+      "loss": 1.009,
+      "step": 105
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.9769494004817592e-05,
+      "loss": 1.0591,
+      "step": 106
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.976268263149723e-05,
+      "loss": 1.0852,
+      "step": 107
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.9755772883429263e-05,
+      "loss": 1.0595,
+      "step": 108
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.974876480665848e-05,
+      "loss": 1.0609,
+      "step": 109
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.974165844788492e-05,
+      "loss": 1.0142,
+      "step": 110
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.9734453854463536e-05,
+      "loss": 1.0692,
+      "step": 111
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.972715107440391e-05,
+      "loss": 1.0298,
+      "step": 112
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.97197501563699e-05,
+      "loss": 1.0419,
+      "step": 113
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.9712251149679334e-05,
+      "loss": 1.0061,
+      "step": 114
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.9704654104303676e-05,
+      "loss": 1.0436,
+      "step": 115
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.969695907086769e-05,
+      "loss": 1.1109,
+      "step": 116
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.968916610064913e-05,
+      "loss": 1.0953,
+      "step": 117
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.968127524557835e-05,
+      "loss": 1.1115,
+      "step": 118
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.967328655823799e-05,
+      "loss": 1.018,
+      "step": 119
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.966520009186263e-05,
+      "loss": 1.0061,
+      "step": 120
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.965701590033841e-05,
+      "loss": 1.0244,
+      "step": 121
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9648734038202703e-05,
+      "loss": 1.0281,
+      "step": 122
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9640354560643714e-05,
+      "loss": 1.0535,
+      "step": 123
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9631877523500145e-05,
+      "loss": 1.0232,
+      "step": 124
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.962330298326081e-05,
+      "loss": 1.0496,
+      "step": 125
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.961463099706425e-05,
+      "loss": 1.1165,
+      "step": 126
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9605861622698372e-05,
+      "loss": 0.9616,
+      "step": 127
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9596994918600043e-05,
+      "loss": 0.8639,
+      "step": 128
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9588030943854723e-05,
+      "loss": 1.0163,
+      "step": 129
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9578969758196038e-05,
+      "loss": 0.9984,
+      "step": 130
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9569811422005427e-05,
+      "loss": 1.0669,
+      "step": 131
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.95605559963117e-05,
+      "loss": 1.0264,
+      "step": 132
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.955120354279065e-05,
+      "loss": 1.0708,
+      "step": 133
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.954175412376464e-05,
+      "loss": 1.1603,
+      "step": 134
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9532207802202184e-05,
+      "loss": 0.9794,
+      "step": 135
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9522564641717533e-05,
+      "loss": 1.0434,
+      "step": 136
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9512824706570245e-05,
+      "loss": 1.0086,
+      "step": 137
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9502988061664763e-05,
+      "loss": 1.0718,
+      "step": 138
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9493054772549977e-05,
+      "loss": 1.0703,
+      "step": 139
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9483024905418784e-05,
+      "loss": 1.1021,
+      "step": 140
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9472898527107666e-05,
+      "loss": 1.0574,
+      "step": 141
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9462675705096216e-05,
+      "loss": 1.0371,
+      "step": 142
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9452356507506718e-05,
+      "loss": 1.0102,
+      "step": 143
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9441941003103653e-05,
+      "loss": 1.0869,
+      "step": 144
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.94314292612933e-05,
+      "loss": 1.0326,
+      "step": 145
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9420821352123213e-05,
+      "loss": 0.9899,
+      "step": 146
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.94101173462818e-05,
+      "loss": 0.9533,
+      "step": 147
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9399317315097806e-05,
+      "loss": 1.0127,
+      "step": 148
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.93884213305399e-05,
+      "loss": 1.0719,
+      "step": 149
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9377429465216126e-05,
+      "loss": 1.0817,
+      "step": 150
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9366341792373473e-05,
+      "loss": 0.9923,
+      "step": 151
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9355158385897367e-05,
+      "loss": 1.0294,
+      "step": 152
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9343879320311165e-05,
+      "loss": 1.0223,
+      "step": 153
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9332504670775682e-05,
+      "loss": 1.0778,
+      "step": 154
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.932103451308869e-05,
+      "loss": 0.9678,
+      "step": 155
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.9309468923684374e-05,
+      "loss": 1.0409,
+      "step": 156
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9297807979632877e-05,
+      "loss": 1.0972,
+      "step": 157
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.928605175863975e-05,
+      "loss": 1.0165,
+      "step": 158
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.927420033904545e-05,
+      "loss": 1.0683,
+      "step": 159
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.926225379982481e-05,
+      "loss": 1.0364,
+      "step": 160
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9250212220586522e-05,
+      "loss": 1.0199,
+      "step": 161
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9238075681572588e-05,
+      "loss": 1.0646,
+      "step": 162
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9225844263657807e-05,
+      "loss": 1.0281,
+      "step": 163
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9213518048349227e-05,
+      "loss": 1.0829,
+      "step": 164
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9201097117785595e-05,
+      "loss": 1.0494,
+      "step": 165
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.918858155473683e-05,
+      "loss": 0.9888,
+      "step": 166
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9175971442603437e-05,
+      "loss": 1.0469,
+      "step": 167
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.916326686541599e-05,
+      "loss": 1.0234,
+      "step": 168
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9150467907834552e-05,
+      "loss": 1.0097,
+      "step": 169
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9137574655148098e-05,
+      "loss": 1.0587,
+      "step": 170
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.912458719327398e-05,
+      "loss": 1.0411,
+      "step": 171
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9111505608757322e-05,
+      "loss": 0.9451,
+      "step": 172
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.909832998877046e-05,
+      "loss": 1.0578,
+      "step": 173
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9085060421112375e-05,
+      "loss": 1.0808,
+      "step": 174
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9071696994208067e-05,
+      "loss": 0.9804,
+      "step": 175
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9058239797108e-05,
+      "loss": 1.0908,
+      "step": 176
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9044688919487506e-05,
+      "loss": 1.0445,
+      "step": 177
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.903104445164617e-05,
+      "loss": 1.0187,
+      "step": 178
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.901730648450724e-05,
+      "loss": 0.9725,
+      "step": 179
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9003475109617023e-05,
+      "loss": 1.0418,
+      "step": 180
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.8989550419144267e-05,
+      "loss": 1.0936,
+      "step": 181
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.897553250587956e-05,
+      "loss": 1.0275,
+      "step": 182
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8961421463234693e-05,
+      "loss": 0.9522,
+      "step": 183
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.894721738524205e-05,
+      "loss": 1.0556,
+      "step": 184
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8932920366553982e-05,
+      "loss": 1.0753,
+      "step": 185
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.891853050244217e-05,
+      "loss": 1.0513,
+      "step": 186
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8904047888796996e-05,
+      "loss": 1.0521,
+      "step": 187
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8889472622126894e-05,
+      "loss": 1.0781,
+      "step": 188
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8874804799557724e-05,
+      "loss": 1.0589,
+      "step": 189
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.886004451883211e-05,
+      "loss": 1.0243,
+      "step": 190
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8845191878308797e-05,
+      "loss": 1.0204,
+      "step": 191
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8830246976961977e-05,
+      "loss": 1.143,
+      "step": 192
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8815209914380672e-05,
+      "loss": 1.0046,
+      "step": 193
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.8800080790768022e-05,
+      "loss": 0.9957,
+      "step": 194
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8784859706940637e-05,
+      "loss": 1.008,
+      "step": 195
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.876954676432794e-05,
+      "loss": 0.9919,
+      "step": 196
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.875414206497147e-05,
+      "loss": 0.9517,
+      "step": 197
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.873864571152421e-05,
+      "loss": 0.9965,
+      "step": 198
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8723057807249902e-05,
+      "loss": 0.9456,
+      "step": 199
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8707378456022358e-05,
+      "loss": 0.9865,
+      "step": 200
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.869160776232477e-05,
+      "loss": 1.0567,
+      "step": 201
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8675745831249024e-05,
+      "loss": 1.0489,
+      "step": 202
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8659792768494966e-05,
+      "loss": 1.0587,
+      "step": 203
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8643748680369745e-05,
+      "loss": 1.1367,
+      "step": 204
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.862761367378705e-05,
+      "loss": 1.0152,
+      "step": 205
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8611387856266463e-05,
+      "loss": 1.082,
+      "step": 206
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.8595071335932685e-05,
+      "loss": 0.978,
+      "step": 207
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8578664221514826e-05,
+      "loss": 1.014,
+      "step": 208
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.856216662234573e-05,
+      "loss": 1.0423,
+      "step": 209
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8545578648361173e-05,
+      "loss": 1.0006,
+      "step": 210
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8528900410099178e-05,
+      "loss": 1.0373,
+      "step": 211
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8512132018699267e-05,
+      "loss": 1.1298,
+      "step": 212
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.849527358590172e-05,
+      "loss": 1.1089,
+      "step": 213
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8478325224046828e-05,
+      "loss": 0.976,
+      "step": 214
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.846128704607415e-05,
+      "loss": 1.0484,
+      "step": 215
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.844415916552175e-05,
+      "loss": 1.0419,
+      "step": 216
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.842694169652545e-05,
+      "loss": 1.0755,
+      "step": 217
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.840963475381807e-05,
+      "loss": 1.1313,
+      "step": 218
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.8392238452728654e-05,
+      "loss": 0.9376,
+      "step": 219
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8374752909181713e-05,
+      "loss": 0.9967,
+      "step": 220
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8357178239696443e-05,
+      "loss": 1.0122,
+      "step": 221
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8339514561385953e-05,
+      "loss": 0.9879,
+      "step": 222
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.832176199195649e-05,
+      "loss": 0.9948,
+      "step": 223
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8303920649706637e-05,
+      "loss": 1.0257,
+      "step": 224
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8285990653526552e-05,
+      "loss": 1.0573,
+      "step": 225
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8267972122897143e-05,
+      "loss": 1.0697,
+      "step": 226
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8249865177889302e-05,
+      "loss": 1.0125,
+      "step": 227
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.823166993916308e-05,
+      "loss": 0.9986,
+      "step": 228
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8213386527966913e-05,
+      "loss": 0.9547,
+      "step": 229
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.819501506613678e-05,
+      "loss": 1.013,
+      "step": 230
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8176555676095408e-05,
+      "loss": 1.0598,
+      "step": 231
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.8158008480851465e-05,
+      "loss": 0.9885,
+      "step": 232
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.813937360399872e-05,
+      "loss": 0.9745,
+      "step": 233
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.8120651169715237e-05,
+      "loss": 1.0145,
+      "step": 234
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.8101841302762533e-05,
+      "loss": 1.0939,
+      "step": 235
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.8082944128484763e-05,
+      "loss": 1.0058,
+      "step": 236
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.806395977280787e-05,
+      "loss": 1.1187,
+      "step": 237
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.8044888362238753e-05,
+      "loss": 0.9783,
+      "step": 238
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.8025730023864425e-05,
+      "loss": 0.9609,
+      "step": 239
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.8006484885351152e-05,
+      "loss": 1.002,
+      "step": 240
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.7987153074943632e-05,
+      "loss": 1.0225,
+      "step": 241
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.7967734721464114e-05,
+      "loss": 1.0135,
+      "step": 242
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.794822995431154e-05,
+      "loss": 1.0561,
+      "step": 243
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.7928638903460703e-05,
+      "loss": 1.0292,
+      "step": 244
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7908961699461363e-05,
+      "loss": 0.9945,
+      "step": 245
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7889198473437385e-05,
+      "loss": 1.0148,
+      "step": 246
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7869349357085856e-05,
+      "loss": 1.0075,
+      "step": 247
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7849414482676222e-05,
+      "loss": 0.933,
+      "step": 248
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.78293939830494e-05,
+      "loss": 1.0077,
+      "step": 249
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7809287991616876e-05,
+      "loss": 0.9821,
+      "step": 250
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.778909664235986e-05,
+      "loss": 1.0797,
+      "step": 251
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.776882006982833e-05,
+      "loss": 0.9748,
+      "step": 252
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7748458409140202e-05,
+      "loss": 1.0633,
+      "step": 253
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7728011795980377e-05,
+      "loss": 0.9683,
+      "step": 254
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7707480366599865e-05,
+      "loss": 1.0112,
+      "step": 255
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.768686425781487e-05,
+      "loss": 0.9832,
+      "step": 256
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.7666163607005875e-05,
+      "loss": 0.978,
+      "step": 257
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7645378552116734e-05,
+      "loss": 0.9563,
+      "step": 258
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7624509231653747e-05,
+      "loss": 1.0818,
+      "step": 259
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7603555784684734e-05,
+      "loss": 0.9814,
+      "step": 260
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7582518350838114e-05,
+      "loss": 1.0487,
+      "step": 261
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7561397070301975e-05,
+      "loss": 0.9749,
+      "step": 262
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.754019208382314e-05,
+      "loss": 0.9604,
+      "step": 263
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.751890353270622e-05,
+      "loss": 1.0739,
+      "step": 264
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.749753155881269e-05,
+      "loss": 1.0931,
+      "step": 265
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7476076304559926e-05,
+      "loss": 1.0174,
+      "step": 266
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.745453791292026e-05,
+      "loss": 1.1114,
+      "step": 267
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.7432916527420037e-05,
+      "loss": 1.0449,
+      "step": 268
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.741121229213865e-05,
+      "loss": 1.0161,
+      "step": 269
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.738942535170758e-05,
+      "loss": 0.9802,
+      "step": 270
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7367555851309433e-05,
+      "loss": 1.0514,
+      "step": 271
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7345603936676978e-05,
+      "loss": 1.0708,
+      "step": 272
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7323569754092168e-05,
+      "loss": 1.069,
+      "step": 273
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7301453450385173e-05,
+      "loss": 1.0149,
+      "step": 274
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.727925517293339e-05,
+      "loss": 0.9965,
+      "step": 275
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.725697506966047e-05,
+      "loss": 1.017,
+      "step": 276
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7234613289035333e-05,
+      "loss": 1.027,
+      "step": 277
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7212169980071186e-05,
+      "loss": 0.9985,
+      "step": 278
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.718964529232449e-05,
+      "loss": 1.0769,
+      "step": 279
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.716703937589403e-05,
+      "loss": 0.9525,
+      "step": 280
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.714435238141986e-05,
+      "loss": 1.087,
+      "step": 281
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7121584460082304e-05,
+      "loss": 1.0002,
+      "step": 282
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.709873576360099e-05,
+      "loss": 1.1103,
+      "step": 283
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.7075806444233787e-05,
+      "loss": 0.9968,
+      "step": 284
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.7052796654775833e-05,
+      "loss": 0.9958,
+      "step": 285
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.702970654855849e-05,
+      "loss": 0.9353,
+      "step": 286
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.7006536279448327e-05,
+      "loss": 0.9545,
+      "step": 287
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.6983286001846096e-05,
+      "loss": 0.9062,
+      "step": 288
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.6959955870685715e-05,
+      "loss": 0.9569,
+      "step": 289
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.6936546041433228e-05,
+      "loss": 0.9843,
+      "step": 290
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.6913056670085746e-05,
+      "loss": 0.9891,
+      "step": 291
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.688948791317045e-05,
+      "loss": 0.9494,
+      "step": 292
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.686583992774352e-05,
+      "loss": 1.0039,
+      "step": 293
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.684211287138909e-05,
+      "loss": 1.0229,
+      "step": 294
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.6818306902218204e-05,
+      "loss": 0.9418,
+      "step": 295
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6794422178867767e-05,
+      "loss": 0.9146,
+      "step": 296
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6770458860499466e-05,
+      "loss": 0.9599,
+      "step": 297
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6746417106798747e-05,
+      "loss": 0.9696,
+      "step": 298
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.672229707797371e-05,
+      "loss": 0.9337,
+      "step": 299
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6698098934754074e-05,
+      "loss": 1.0708,
+      "step": 300
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6673822838390086e-05,
+      "loss": 0.9915,
+      "step": 301
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6649468950651452e-05,
+      "loss": 0.9919,
+      "step": 302
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6625037433826262e-05,
+      "loss": 0.9606,
+      "step": 303
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6600528450719906e-05,
+      "loss": 0.8912,
+      "step": 304
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6575942164653985e-05,
+      "loss": 0.9752,
+      "step": 305
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6551278739465242e-05,
+      "loss": 0.85,
+      "step": 306
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.6526538339504434e-05,
+      "loss": 0.9676,
+      "step": 307
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.650172112963528e-05,
+      "loss": 0.9529,
+      "step": 308
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6476827275233315e-05,
+      "loss": 1.0061,
+      "step": 309
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6451856942184844e-05,
+      "loss": 1.0193,
+      "step": 310
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.642681029688578e-05,
+      "loss": 0.9645,
+      "step": 311
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6401687506240577e-05,
+      "loss": 0.9248,
+      "step": 312
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6376488737661096e-05,
+      "loss": 0.9074,
+      "step": 313
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6351214159065497e-05,
+      "loss": 1.0399,
+      "step": 314
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6325863938877116e-05,
+      "loss": 0.9657,
+      "step": 315
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6300438246023353e-05,
+      "loss": 1.0181,
+      "step": 316
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6274937249934526e-05,
+      "loss": 0.918,
+      "step": 317
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6249361120542766e-05,
+      "loss": 0.9334,
+      "step": 318
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6223710028280867e-05,
+      "loss": 1.0437,
+      "step": 319
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.6197984144081163e-05,
+      "loss": 0.9442,
+      "step": 320
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.617218363937437e-05,
+      "loss": 0.9717,
+      "step": 321
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.6146308686088464e-05,
+      "loss": 1.1193,
+      "step": 322
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.6120359456647536e-05,
+      "loss": 1.0027,
+      "step": 323
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.6094336123970614e-05,
+      "loss": 0.9899,
+      "step": 324
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.606823886147055e-05,
+      "loss": 0.9836,
+      "step": 325
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.6042067843052837e-05,
+      "loss": 0.9187,
+      "step": 326
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.601582324311445e-05,
+      "loss": 0.9582,
+      "step": 327
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.598950523654272e-05,
+      "loss": 0.9912,
+      "step": 328
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.59631139987141e-05,
+      "loss": 0.9662,
+      "step": 329
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.593664970549308e-05,
+      "loss": 0.9338,
+      "step": 330
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.5910112533230944e-05,
+      "loss": 0.9308,
+      "step": 331
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.5883502658764634e-05,
+      "loss": 1.0139,
+      "step": 332
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5856820259415557e-05,
+      "loss": 1.0075,
+      "step": 333
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.583006551298841e-05,
+      "loss": 0.9364,
+      "step": 334
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.580323859776999e-05,
+      "loss": 1.0592,
+      "step": 335
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5776339692528018e-05,
+      "loss": 1.0622,
+      "step": 336
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5749368976509923e-05,
+      "loss": 0.9643,
+      "step": 337
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.572232662944167e-05,
+      "loss": 0.9455,
+      "step": 338
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5695212831526556e-05,
+      "loss": 0.9171,
+      "step": 339
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5668027763444005e-05,
+      "loss": 1.0297,
+      "step": 340
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5640771606348377e-05,
+      "loss": 0.9981,
+      "step": 341
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5613444541867734e-05,
+      "loss": 0.9892,
+      "step": 342
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5586046752102663e-05,
+      "loss": 0.9553,
+      "step": 343
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.555857841962504e-05,
+      "loss": 1.0124,
+      "step": 344
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.5531039727476808e-05,
+      "loss": 1.0656,
+      "step": 345
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5503430859168794e-05,
+      "loss": 0.9775,
+      "step": 346
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5475751998679438e-05,
+      "loss": 0.9301,
+      "step": 347
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5448003330453603e-05,
+      "loss": 0.9599,
+      "step": 348
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5420185039401323e-05,
+      "loss": 1.035,
+      "step": 349
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.53922973108966e-05,
+      "loss": 0.9714,
+      "step": 350
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.536434033077612e-05,
+      "loss": 0.9336,
+      "step": 351
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5336314285338063e-05,
+      "loss": 0.891,
+      "step": 352
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5308219361340848e-05,
+      "loss": 1.021,
+      "step": 353
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.528005574600187e-05,
+      "loss": 0.9881,
+      "step": 354
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5251823626996272e-05,
+      "loss": 0.9069,
+      "step": 355
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5223523192455687e-05,
+      "loss": 1.0172,
+      "step": 356
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.5195154630966982e-05,
+      "loss": 0.9946,
+      "step": 357
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.5166718131571003e-05,
+      "loss": 1.0056,
+      "step": 358
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.513821388376133e-05,
+      "loss": 1.0377,
+      "step": 359
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.510964207748299e-05,
+      "loss": 0.8984,
+      "step": 360
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.5081002903131202e-05,
+      "loss": 0.9681,
+      "step": 361
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.5052296551550115e-05,
+      "loss": 1.0663,
+      "step": 362
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.502352321403152e-05,
+      "loss": 1.0292,
+      "step": 363
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.499468308231359e-05,
+      "loss": 1.0103,
+      "step": 364
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.4965776348579613e-05,
+      "loss": 1.0016,
+      "step": 365
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.493680320545667e-05,
+      "loss": 1.0202,
+      "step": 366
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.4907763846014388e-05,
+      "loss": 1.0551,
+      "step": 367
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.4878658463763652e-05,
+      "loss": 0.9298,
+      "step": 368
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.484948725265529e-05,
+      "loss": 0.9412,
+      "step": 369
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.4820250407078803e-05,
+      "loss": 1.0358,
+      "step": 370
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4790948121861076e-05,
+      "loss": 0.967,
+      "step": 371
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4761580592265043e-05,
+      "loss": 0.9725,
+      "step": 372
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.473214801398843e-05,
+      "loss": 0.9665,
+      "step": 373
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.470265058316242e-05,
+      "loss": 0.9622,
+      "step": 374
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4673088496350363e-05,
+      "loss": 0.9336,
+      "step": 375
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4643461950546454e-05,
+      "loss": 1.0701,
+      "step": 376
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4613771143174432e-05,
+      "loss": 0.9188,
+      "step": 377
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4584016272086255e-05,
+      "loss": 0.9702,
+      "step": 378
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.455419753556079e-05,
+      "loss": 0.9674,
+      "step": 379
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4524315132302477e-05,
+      "loss": 0.9572,
+      "step": 380
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4494369261440025e-05,
+      "loss": 1.0073,
+      "step": 381
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.4464360122525074e-05,
+      "loss": 0.9793,
+      "step": 382
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4434287915530857e-05,
+      "loss": 1.0432,
+      "step": 383
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.440415284085089e-05,
+      "loss": 1.0014,
+      "step": 384
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.437395509929761e-05,
+      "loss": 0.8925,
+      "step": 385
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.434369489210106e-05,
+      "loss": 0.9744,
+      "step": 386
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4313372420907528e-05,
+      "loss": 1.0367,
+      "step": 387
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4282987887778226e-05,
+      "loss": 0.956,
+      "step": 388
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4252541495187923e-05,
+      "loss": 0.968,
+      "step": 389
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4222033446023602e-05,
+      "loss": 0.9396,
+      "step": 390
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.419146394358312e-05,
+      "loss": 0.9804,
+      "step": 391
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4160833191573832e-05,
+      "loss": 0.8895,
+      "step": 392
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4130141394111244e-05,
+      "loss": 0.9808,
+      "step": 393
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.409938875571766e-05,
+      "loss": 0.9737,
+      "step": 394
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.4068575481320814e-05,
+      "loss": 1.0355,
+      "step": 395
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.4037701776252494e-05,
+      "loss": 1.008,
+      "step": 396
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.4006767846247194e-05,
+      "loss": 0.977,
+      "step": 397
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3975773897440716e-05,
+      "loss": 1.0697,
+      "step": 398
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3944720136368824e-05,
+      "loss": 1.0171,
+      "step": 399
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.391360676996586e-05,
+      "loss": 0.8689,
+      "step": 400
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3882434005563343e-05,
+      "loss": 0.9595,
+      "step": 401
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3851202050888623e-05,
+      "loss": 1.054,
+      "step": 402
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3819911114063465e-05,
+      "loss": 0.9709,
+      "step": 403
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3788561403602683e-05,
+      "loss": 0.9585,
+      "step": 404
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.375715312841275e-05,
+      "loss": 1.0443,
+      "step": 405
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3725686497790388e-05,
+      "loss": 0.9903,
+      "step": 406
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3694161721421184e-05,
+      "loss": 0.9089,
+      "step": 407
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.3662579009378205e-05,
+      "loss": 1.0071,
+      "step": 408
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.363093857212058e-05,
+      "loss": 0.9785,
+      "step": 409
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3599240620492097e-05,
+      "loss": 1.034,
+      "step": 410
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3567485365719817e-05,
+      "loss": 1.003,
+      "step": 411
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.353567301941265e-05,
+      "loss": 0.9635,
+      "step": 412
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.350380379355995e-05,
+      "loss": 0.8601,
+      "step": 413
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3471877900530093e-05,
+      "loss": 0.9926,
+      "step": 414
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3439895553069087e-05,
+      "loss": 0.9806,
+      "step": 415
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.340785696429913e-05,
+      "loss": 0.9878,
+      "step": 416
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.337576234771719e-05,
+      "loss": 0.9716,
+      "step": 417
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3343611917193596e-05,
+      "loss": 0.983,
+      "step": 418
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3311405886970617e-05,
+      "loss": 0.9522,
+      "step": 419
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.3279144471661006e-05,
+      "loss": 0.9404,
+      "step": 420
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.32468278862466e-05,
+      "loss": 0.9858,
+      "step": 421
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.3214456346076882e-05,
+      "loss": 0.9126,
+      "step": 422
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.3182030066867513e-05,
+      "loss": 0.976,
+      "step": 423
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.314954926469895e-05,
+      "loss": 0.8809,
+      "step": 424
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.3117014156014965e-05,
+      "loss": 0.9875,
+      "step": 425
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.3084424957621207e-05,
+      "loss": 0.9189,
+      "step": 426
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.3051781886683782e-05,
+      "loss": 0.9091,
+      "step": 427
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.301908516072777e-05,
+      "loss": 0.9039,
+      "step": 428
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.298633499763582e-05,
+      "loss": 0.8917,
+      "step": 429
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.2953531615646634e-05,
+      "loss": 1.008,
+      "step": 430
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.2920675233353593e-05,
+      "loss": 0.9837,
+      "step": 431
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.288776606970323e-05,
+      "loss": 1.0046,
+      "step": 432
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.2854804343993813e-05,
+      "loss": 0.9512,
+      "step": 433
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.2821790275873855e-05,
+      "loss": 0.9456,
+      "step": 434
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.2788724085340694e-05,
+      "loss": 0.9403,
+      "step": 435
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.2755605992738965e-05,
+      "loss": 0.9991,
+      "step": 436
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.272243621875919e-05,
+      "loss": 0.9382,
+      "step": 437
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.268921498443627e-05,
+      "loss": 0.9075,
+      "step": 438
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.265594251114803e-05,
+      "loss": 0.9842,
+      "step": 439
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.2622619020613734e-05,
+      "loss": 0.9466,
+      "step": 440
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.258924473489262e-05,
+      "loss": 0.9371,
+      "step": 441
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.2555819876382392e-05,
+      "loss": 0.9836,
+      "step": 442
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.252234466781779e-05,
+      "loss": 1.0178,
+      "step": 443
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.2488819332269046e-05,
+      "loss": 0.9388,
+      "step": 444
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.245524409314044e-05,
+      "loss": 0.9447,
+      "step": 445
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2421619174168786e-05,
+      "loss": 0.9853,
+      "step": 446
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2387944799421964e-05,
+      "loss": 0.912,
+      "step": 447
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2354221193297412e-05,
+      "loss": 0.9338,
+      "step": 448
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2320448580520626e-05,
+      "loss": 1.0299,
+      "step": 449
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2286627186143674e-05,
+      "loss": 0.9733,
+      "step": 450
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.225275723554369e-05,
+      "loss": 0.9888,
+      "step": 451
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2218838954421386e-05,
+      "loss": 0.9935,
+      "step": 452
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2184872568799516e-05,
+      "loss": 1.0319,
+      "step": 453
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.215085830502141e-05,
+      "loss": 0.9224,
+      "step": 454
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.211679638974943e-05,
+      "loss": 1.0248,
+      "step": 455
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.208268704996348e-05,
+      "loss": 0.9948,
+      "step": 456
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.20485305129595e-05,
+      "loss": 0.9898,
+      "step": 457
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.2014327006347927e-05,
+      "loss": 0.958,
+      "step": 458
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.198007675805219e-05,
+      "loss": 0.9185,
+      "step": 459
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.194577999630721e-05,
+      "loss": 0.9818,
+      "step": 460
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1911436949657843e-05,
+      "loss": 0.8751,
+      "step": 461
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1877047846957378e-05,
+      "loss": 0.8988,
+      "step": 462
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1842612917366022e-05,
+      "loss": 0.9746,
+      "step": 463
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1808132390349357e-05,
+      "loss": 0.9333,
+      "step": 464
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.17736064956768e-05,
+      "loss": 0.9397,
+      "step": 465
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1739035463420106e-05,
+      "loss": 1.0369,
+      "step": 466
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.17044195239518e-05,
+      "loss": 0.9371,
+      "step": 467
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1669758907943668e-05,
+      "loss": 0.9578,
+      "step": 468
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.1635053846365204e-05,
+      "loss": 0.9704,
+      "step": 469
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.160030457048208e-05,
+      "loss": 0.9761,
+      "step": 470
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.156551131185459e-05,
+      "loss": 0.9526,
+      "step": 471
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1530674302336135e-05,
+      "loss": 0.9107,
+      "step": 472
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1495793774071643e-05,
+      "loss": 0.9796,
+      "step": 473
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1460869959496058e-05,
+      "loss": 0.9176,
+      "step": 474
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1425903091332753e-05,
+      "loss": 0.9488,
+      "step": 475
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.139089340259201e-05,
+      "loss": 0.9712,
+      "step": 476
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1355841126569462e-05,
+      "loss": 0.9139,
+      "step": 477
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1320746496844522e-05,
+      "loss": 0.9608,
+      "step": 478
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1285609747278846e-05,
+      "loss": 0.9615,
+      "step": 479
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.125043111201475e-05,
+      "loss": 0.9872,
+      "step": 480
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1215210825473693e-05,
+      "loss": 0.9621,
+      "step": 481
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1179949122354656e-05,
+      "loss": 0.9473,
+      "step": 482
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.1144646237632637e-05,
+      "loss": 0.9414,
+      "step": 483
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.110930240655704e-05,
+      "loss": 0.9806,
+      "step": 484
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.107391786465013e-05,
+      "loss": 0.9645,
+      "step": 485
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.103849284770546e-05,
+      "loss": 0.9555,
+      "step": 486
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.1003027591786303e-05,
+      "loss": 0.9221,
+      "step": 487
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.096752233322406e-05,
+      "loss": 0.8472,
+      "step": 488
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.0931977308616718e-05,
+      "loss": 0.9792,
+      "step": 489
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.089639275482724e-05,
+      "loss": 0.872,
+      "step": 490
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.0860768908982004e-05,
+      "loss": 0.9301,
+      "step": 491
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.0825106008469223e-05,
+      "loss": 0.9127,
+      "step": 492
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.0789404290937364e-05,
+      "loss": 0.9266,
+      "step": 493
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.0753663994293555e-05,
+      "loss": 0.9425,
+      "step": 494
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.0717885356702002e-05,
+      "loss": 1.0128,
+      "step": 495
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.068206861658242e-05,
+      "loss": 0.9979,
+      "step": 496
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0646214012608415e-05,
+      "loss": 1.0044,
+      "step": 497
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.061032178370591e-05,
+      "loss": 1.0387,
+      "step": 498
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.057439216905156e-05,
+      "loss": 0.9365,
+      "step": 499
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0538425408071138e-05,
+      "loss": 0.9162,
+      "step": 500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0502421740437953e-05,
+      "loss": 0.9848,
+      "step": 501
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0466381406071263e-05,
+      "loss": 0.9849,
+      "step": 502
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0430304645134646e-05,
+      "loss": 0.9119,
+      "step": 503
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0394191698034425e-05,
+      "loss": 0.8846,
+      "step": 504
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0358042805418065e-05,
+      "loss": 0.9792,
+      "step": 505
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0321858208172548e-05,
+      "loss": 0.971,
+      "step": 506
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0285638147422796e-05,
+      "loss": 0.8285,
+      "step": 507
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.0249382864530043e-05,
+      "loss": 0.9235,
+      "step": 508
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.021309260109024e-05,
+      "loss": 0.9606,
+      "step": 509
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.0176767598932433e-05,
+      "loss": 0.9684,
+      "step": 510
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.0140408100117163e-05,
+      "loss": 0.97,
+      "step": 511
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.0104014346934847e-05,
+      "loss": 0.8283,
+      "step": 512
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.0067586581904163e-05,
+      "loss": 0.9822,
+      "step": 513
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.003112504777044e-05,
+      "loss": 0.9781,
+      "step": 514
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9994629987504016e-05,
+      "loss": 0.8996,
+      "step": 515
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.995810164429867e-05,
+      "loss": 0.8587,
+      "step": 516
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9921540261569952e-05,
+      "loss": 0.8702,
+      "step": 517
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9884946082953578e-05,
+      "loss": 0.9922,
+      "step": 518
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9848319352303808e-05,
+      "loss": 0.9938,
+      "step": 519
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9811660313691824e-05,
+      "loss": 0.9392,
+      "step": 520
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9774969211404104e-05,
+      "loss": 0.9438,
+      "step": 521
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.973824628994078e-05,
+      "loss": 0.9544,
+      "step": 522
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9701491794014028e-05,
+      "loss": 1.0099,
+      "step": 523
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9664705968546422e-05,
+      "loss": 1.0053,
+      "step": 524
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.962788905866931e-05,
+      "loss": 0.8912,
+      "step": 525
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9591041309721177e-05,
+      "loss": 0.8648,
+      "step": 526
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9554162967246014e-05,
+      "loss": 0.9566,
+      "step": 527
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9517254276991677e-05,
+      "loss": 0.929,
+      "step": 528
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9480315484908264e-05,
+      "loss": 0.9746,
+      "step": 529
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.944334683714644e-05,
+      "loss": 0.9771,
+      "step": 530
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9406348580055842e-05,
+      "loss": 0.9009,
+      "step": 531
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.936932096018341e-05,
+      "loss": 0.9065,
+      "step": 532
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.933226422427174e-05,
+      "loss": 0.8756,
+      "step": 533
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9295178619257464e-05,
+      "loss": 0.9544,
+      "step": 534
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.925806439226959e-05,
+      "loss": 0.8725,
+      "step": 535
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.922092179062784e-05,
+      "loss": 0.8904,
+      "step": 536
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.918375106184103e-05,
+      "loss": 0.9524,
+      "step": 537
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.914655245360541e-05,
+      "loss": 0.994,
+      "step": 538
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9109326213803003e-05,
+      "loss": 0.9115,
+      "step": 539
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9072072590499963e-05,
+      "loss": 0.9057,
+      "step": 540
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.903479183194493e-05,
+      "loss": 0.9804,
+      "step": 541
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.899748418656736e-05,
+      "loss": 0.9641,
+      "step": 542
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.896014990297587e-05,
+      "loss": 0.9731,
+      "step": 543
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8922789229956595e-05,
+      "loss": 0.7873,
+      "step": 544
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8885402416471524e-05,
+      "loss": 0.9611,
+      "step": 545
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.884798971165683e-05,
+      "loss": 0.8223,
+      "step": 546
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8810551364821236e-05,
+      "loss": 0.9157,
+      "step": 547
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8773087625444316e-05,
+      "loss": 0.9715,
+      "step": 548
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.873559874317486e-05,
+      "loss": 0.9427,
+      "step": 549
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8698084967829223e-05,
+      "loss": 0.9615,
+      "step": 550
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8660546549389606e-05,
+      "loss": 0.9208,
+      "step": 551
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8622983738002458e-05,
+      "loss": 0.9692,
+      "step": 552
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.858539678397676e-05,
+      "loss": 0.8826,
+      "step": 553
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.854778593778237e-05,
+      "loss": 0.9508,
+      "step": 554
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.851015145004837e-05,
+      "loss": 0.887,
+      "step": 555
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.847249357156138e-05,
+      "loss": 0.9147,
+      "step": 556
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.843481255326388e-05,
+      "loss": 0.9296,
+      "step": 557
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8397108646252555e-05,
+      "loss": 0.9269,
+      "step": 558
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8359382101776623e-05,
+      "loss": 0.9047,
+      "step": 559
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.832163317123613e-05,
+      "loss": 0.9727,
+      "step": 560
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8283862106180315e-05,
+      "loss": 0.8893,
+      "step": 561
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.824606915830592e-05,
+      "loss": 0.9507,
+      "step": 562
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.820825457945549e-05,
+      "loss": 0.9289,
+      "step": 563
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8170418621615725e-05,
+      "loss": 1.0256,
+      "step": 564
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.813256153691579e-05,
+      "loss": 0.9699,
+      "step": 565
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.809468357762563e-05,
+      "loss": 0.8613,
+      "step": 566
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.805678499615429e-05,
+      "loss": 0.8757,
+      "step": 567
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8018866045048246e-05,
+      "loss": 0.9408,
+      "step": 568
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7980926976989703e-05,
+      "loss": 0.9468,
+      "step": 569
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.794296804479492e-05,
+      "loss": 0.9434,
+      "step": 570
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.790498950141254e-05,
+      "loss": 0.9781,
+      "step": 571
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7866991599921868e-05,
+      "loss": 0.9103,
+      "step": 572
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.782897459353122e-05,
+      "loss": 0.9667,
+      "step": 573
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7790938735576225e-05,
+      "loss": 0.983,
+      "step": 574
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.775288427951812e-05,
+      "loss": 0.9491,
+      "step": 575
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7714811478942083e-05,
+      "loss": 0.9076,
+      "step": 576
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.767672058755555e-05,
+      "loss": 0.858,
+      "step": 577
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7638611859186475e-05,
+      "loss": 0.9354,
+      "step": 578
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7600485547781707e-05,
+      "loss": 0.9156,
+      "step": 579
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.756234190740525e-05,
+      "loss": 0.9813,
+      "step": 580
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7524181192236576e-05,
+      "loss": 0.9721,
+      "step": 581
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7486003656568953e-05,
+      "loss": 0.9717,
+      "step": 582
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7447809554807733e-05,
+      "loss": 0.9478,
+      "step": 583
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.740959914146865e-05,
+      "loss": 0.8862,
+      "step": 584
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.737137267117615e-05,
+      "loss": 0.8416,
+      "step": 585
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7333130398661665e-05,
+      "loss": 1.009,
+      "step": 586
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7294872578761932e-05,
+      "loss": 0.9451,
+      "step": 587
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.72565994664173e-05,
+      "loss": 0.9073,
+      "step": 588
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7218311316670017e-05,
+      "loss": 0.883,
+      "step": 589
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7180008384662537e-05,
+      "loss": 0.9379,
+      "step": 590
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7141690925635814e-05,
+      "loss": 0.9396,
+      "step": 591
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.710335919492761e-05,
+      "loss": 0.8925,
+      "step": 592
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7065013447970798e-05,
+      "loss": 0.9299,
+      "step": 593
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.702665394029164e-05,
+      "loss": 0.9151,
+      "step": 594
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.6988280927508108e-05,
+      "loss": 0.8699,
+      "step": 595
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.694989466532815e-05,
+      "loss": 0.8902,
+      "step": 596
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.691149540954802e-05,
+      "loss": 1.0281,
+      "step": 597
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6873083416050562e-05,
+      "loss": 0.8233,
+      "step": 598
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.683465894080348e-05,
+      "loss": 0.9091,
+      "step": 599
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6796222239857668e-05,
+      "loss": 0.8978,
+      "step": 600
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6757773569345498e-05,
+      "loss": 0.9266,
+      "step": 601
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6719313185479077e-05,
+      "loss": 0.9816,
+      "step": 602
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6680841344548595e-05,
+      "loss": 0.9314,
+      "step": 603
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.664235830292057e-05,
+      "loss": 1.0253,
+      "step": 604
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6603864317036172e-05,
+      "loss": 1.033,
+      "step": 605
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6565359643409483e-05,
+      "loss": 0.9375,
+      "step": 606
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6526844538625835e-05,
+      "loss": 0.9096,
+      "step": 607
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.6488319259340033e-05,
+      "loss": 0.9376,
+      "step": 608
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6449784062274705e-05,
+      "loss": 1.0182,
+      "step": 609
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6411239204218576e-05,
+      "loss": 0.939,
+      "step": 610
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6372684942024727e-05,
+      "loss": 0.9372,
+      "step": 611
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6334121532608917e-05,
+      "loss": 0.9067,
+      "step": 612
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6295549232947868e-05,
+      "loss": 1.0143,
+      "step": 613
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.625696830007753e-05,
+      "loss": 0.8966,
+      "step": 614
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6218378991091388e-05,
+      "loss": 0.9856,
+      "step": 615
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6179781563138747e-05,
+      "loss": 0.9286,
+      "step": 616
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.614117627342301e-05,
+      "loss": 0.914,
+      "step": 617
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.610256337919997e-05,
+      "loss": 0.9879,
+      "step": 618
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6063943137776097e-05,
+      "loss": 0.9074,
+      "step": 619
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6025315806506817e-05,
+      "loss": 0.9416,
+      "step": 620
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.59866816427948e-05,
+      "loss": 0.9037,
+      "step": 621
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5948040904088256e-05,
+      "loss": 0.9799,
+      "step": 622
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5909393847879192e-05,
+      "loss": 0.917,
+      "step": 623
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5870740731701727e-05,
+      "loss": 0.9934,
+      "step": 624
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5832081813130364e-05,
+      "loss": 0.941,
+      "step": 625
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.579341734977825e-05,
+      "loss": 0.9941,
+      "step": 626
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5754747599295507e-05,
+      "loss": 0.8864,
+      "step": 627
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5716072819367476e-05,
+      "loss": 0.9664,
+      "step": 628
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.567739326771302e-05,
+      "loss": 0.9597,
+      "step": 629
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5638709202082787e-05,
+      "loss": 0.9146,
+      "step": 630
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5600020880257524e-05,
+      "loss": 0.8144,
+      "step": 631
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.5561328560046316e-05,
+      "loss": 0.9143,
+      "step": 632
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.552263249928491e-05,
+      "loss": 0.8747,
+      "step": 633
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.548393295583398e-05,
+      "loss": 0.9267,
+      "step": 634
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5445230187577396e-05,
+      "loss": 0.9406,
+      "step": 635
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.540652445242053e-05,
+      "loss": 0.9565,
+      "step": 636
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5367816008288508e-05,
+      "loss": 0.9112,
+      "step": 637
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5329105113124522e-05,
+      "loss": 0.9987,
+      "step": 638
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5290392024888098e-05,
+      "loss": 0.9238,
+      "step": 639
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5251677001553367e-05,
+      "loss": 0.9682,
+      "step": 640
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5212960301107357e-05,
+      "loss": 0.897,
+      "step": 641
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5174242181548281e-05,
+      "loss": 0.9401,
+      "step": 642
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5135522900883798e-05,
+      "loss": 0.9359,
+      "step": 643
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.5096802717129303e-05,
+      "loss": 0.9004,
+      "step": 644
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.505808188830622e-05,
+      "loss": 0.9929,
+      "step": 645
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.501936067244027e-05,
+      "loss": 0.8523,
+      "step": 646
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4980639327559733e-05,
+      "loss": 0.9562,
+      "step": 647
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4941918111693777e-05,
+      "loss": 0.9317,
+      "step": 648
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4903197282870696e-05,
+      "loss": 0.9596,
+      "step": 649
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.486447709911621e-05,
+      "loss": 0.99,
+      "step": 650
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4825757818451721e-05,
+      "loss": 0.9651,
+      "step": 651
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4787039698892646e-05,
+      "loss": 0.9244,
+      "step": 652
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4748322998446637e-05,
+      "loss": 0.9054,
+      "step": 653
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4709607975111903e-05,
+      "loss": 0.9202,
+      "step": 654
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4670894886875475e-05,
+      "loss": 0.967,
+      "step": 655
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.46321839917115e-05,
+      "loss": 0.97,
+      "step": 656
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4593475547579477e-05,
+      "loss": 0.9899,
+      "step": 657
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.4554769812422605e-05,
+      "loss": 0.9486,
+      "step": 658
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.451606704416602e-05,
+      "loss": 0.9047,
+      "step": 659
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.447736750071509e-05,
+      "loss": 0.9155,
+      "step": 660
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4438671439953683e-05,
+      "loss": 1.0275,
+      "step": 661
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4399979119742483e-05,
+      "loss": 0.9246,
+      "step": 662
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4361290797917215e-05,
+      "loss": 0.9826,
+      "step": 663
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4322606732286983e-05,
+      "loss": 1.0408,
+      "step": 664
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4283927180632525e-05,
+      "loss": 0.9537,
+      "step": 665
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4245252400704495e-05,
+      "loss": 0.9365,
+      "step": 666
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4206582650221753e-05,
+      "loss": 1.0367,
+      "step": 667
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4167918186869645e-05,
+      "loss": 0.9446,
+      "step": 668
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4129259268298275e-05,
+      "loss": 1.0397,
+      "step": 669
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.409060615212081e-05,
+      "loss": 0.8751,
+      "step": 670
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.4051959095911746e-05,
+      "loss": 0.9155,
+      "step": 671
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.40133183572052e-05,
+      "loss": 0.9271,
+      "step": 672
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3974684193493182e-05,
+      "loss": 0.8246,
+      "step": 673
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3936056862223907e-05,
+      "loss": 0.9546,
+      "step": 674
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3897436620800034e-05,
+      "loss": 0.9329,
+      "step": 675
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3858823726576993e-05,
+      "loss": 0.9219,
+      "step": 676
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3820218436861255e-05,
+      "loss": 0.9045,
+      "step": 677
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3781621008908613e-05,
+      "loss": 0.9724,
+      "step": 678
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3743031699922473e-05,
+      "loss": 0.9322,
+      "step": 679
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3704450767052136e-05,
+      "loss": 0.8907,
+      "step": 680
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3665878467391083e-05,
+      "loss": 0.9023,
+      "step": 681
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3627315057975277e-05,
+      "loss": 0.9785,
+      "step": 682
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3588760795781426e-05,
+      "loss": 0.8949,
+      "step": 683
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.3550215937725292e-05,
+      "loss": 0.9757,
+      "step": 684
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3511680740659968e-05,
+      "loss": 0.8672,
+      "step": 685
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3473155461374173e-05,
+      "loss": 0.8976,
+      "step": 686
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3434640356590516e-05,
+      "loss": 0.9655,
+      "step": 687
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3396135682963832e-05,
+      "loss": 0.9477,
+      "step": 688
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3357641697079431e-05,
+      "loss": 1.0196,
+      "step": 689
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3319158655451408e-05,
+      "loss": 0.991,
+      "step": 690
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3280686814520922e-05,
+      "loss": 0.8988,
+      "step": 691
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.324222643065451e-05,
+      "loss": 0.9223,
+      "step": 692
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3203777760142333e-05,
+      "loss": 0.9274,
+      "step": 693
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3165341059196523e-05,
+      "loss": 0.9313,
+      "step": 694
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3126916583949442e-05,
+      "loss": 0.9886,
+      "step": 695
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.3088504590451976e-05,
+      "loss": 0.9194,
+      "step": 696
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.3050105334671848e-05,
+      "loss": 0.8953,
+      "step": 697
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.3011719072491897e-05,
+      "loss": 1.044,
+      "step": 698
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.297334605970836e-05,
+      "loss": 0.8883,
+      "step": 699
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2934986552029202e-05,
+      "loss": 0.8691,
+      "step": 700
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.289664080507239e-05,
+      "loss": 0.9946,
+      "step": 701
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.285830907436419e-05,
+      "loss": 0.8919,
+      "step": 702
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2819991615337467e-05,
+      "loss": 0.8646,
+      "step": 703
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2781688683329985e-05,
+      "loss": 0.9263,
+      "step": 704
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2743400533582702e-05,
+      "loss": 0.997,
+      "step": 705
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2705127421238069e-05,
+      "loss": 0.9437,
+      "step": 706
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.266686960133834e-05,
+      "loss": 0.898,
+      "step": 707
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2628627328823854e-05,
+      "loss": 0.949,
+      "step": 708
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.2590400858531347e-05,
+      "loss": 0.8241,
+      "step": 709
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2552190445192273e-05,
+      "loss": 0.953,
+      "step": 710
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.251399634343105e-05,
+      "loss": 0.8051,
+      "step": 711
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2475818807763427e-05,
+      "loss": 0.912,
+      "step": 712
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2437658092594755e-05,
+      "loss": 0.8697,
+      "step": 713
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2399514452218294e-05,
+      "loss": 0.9747,
+      "step": 714
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2361388140813526e-05,
+      "loss": 1.0016,
+      "step": 715
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2323279412444457e-05,
+      "loss": 0.859,
+      "step": 716
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2285188521057917e-05,
+      "loss": 0.94,
+      "step": 717
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2247115720481884e-05,
+      "loss": 1.0023,
+      "step": 718
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2209061264423779e-05,
+      "loss": 0.9381,
+      "step": 719
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.217102540646878e-05,
+      "loss": 0.956,
+      "step": 720
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.2133008400078131e-05,
+      "loss": 0.9008,
+      "step": 721
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.2095010498587465e-05,
+      "loss": 0.9539,
+      "step": 722
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.205703195520508e-05,
+      "loss": 0.8502,
+      "step": 723
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.2019073023010302e-05,
+      "loss": 0.9065,
+      "step": 724
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1981133954951758e-05,
+      "loss": 0.8938,
+      "step": 725
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1943215003845711e-05,
+      "loss": 0.9786,
+      "step": 726
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1905316422374371e-05,
+      "loss": 0.9735,
+      "step": 727
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1867438463084213e-05,
+      "loss": 0.8954,
+      "step": 728
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1829581378384277e-05,
+      "loss": 0.935,
+      "step": 729
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1791745420544512e-05,
+      "loss": 0.9075,
+      "step": 730
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1753930841694081e-05,
+      "loss": 0.9107,
+      "step": 731
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1716137893819682e-05,
+      "loss": 1.0386,
+      "step": 732
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1678366828763871e-05,
+      "loss": 1.0021,
+      "step": 733
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.1640617898223385e-05,
+      "loss": 0.9871,
+      "step": 734
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1602891353747446e-05,
+      "loss": 0.8783,
+      "step": 735
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1565187446736124e-05,
+      "loss": 0.913,
+      "step": 736
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1527506428438623e-05,
+      "loss": 0.8356,
+      "step": 737
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1489848549951627e-05,
+      "loss": 0.8997,
+      "step": 738
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1452214062217633e-05,
+      "loss": 0.9397,
+      "step": 739
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1414603216023245e-05,
+      "loss": 0.8982,
+      "step": 740
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1377016261997545e-05,
+      "loss": 0.9882,
+      "step": 741
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1339453450610396e-05,
+      "loss": 0.9695,
+      "step": 742
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1301915032170783e-05,
+      "loss": 1.0146,
+      "step": 743
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.126440125682514e-05,
+      "loss": 0.9857,
+      "step": 744
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1226912374555693e-05,
+      "loss": 0.9251,
+      "step": 745
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1189448635178768e-05,
+      "loss": 0.8778,
+      "step": 746
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.1152010288343168e-05,
+      "loss": 0.9689,
+      "step": 747
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.1114597583528479e-05,
+      "loss": 0.9139,
+      "step": 748
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.1077210770043406e-05,
+      "loss": 0.9087,
+      "step": 749
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.1039850097024132e-05,
+      "loss": 0.9861,
+      "step": 750
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.1002515813432644e-05,
+      "loss": 0.886,
+      "step": 751
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0965208168055072e-05,
+      "loss": 0.9431,
+      "step": 752
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0927927409500038e-05,
+      "loss": 0.8983,
+      "step": 753
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0890673786197001e-05,
+      "loss": 0.9813,
+      "step": 754
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0853447546394593e-05,
+      "loss": 0.9632,
+      "step": 755
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0816248938158972e-05,
+      "loss": 0.9013,
+      "step": 756
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0779078209372167e-05,
+      "loss": 0.8823,
+      "step": 757
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0741935607730416e-05,
+      "loss": 0.9423,
+      "step": 758
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.0704821380742537e-05,
+      "loss": 0.937,
+      "step": 759
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0667735775728262e-05,
+      "loss": 0.9248,
+      "step": 760
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0630679039816594e-05,
+      "loss": 1.0116,
+      "step": 761
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0593651419944159e-05,
+      "loss": 0.8833,
+      "step": 762
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0556653162853563e-05,
+      "loss": 0.8218,
+      "step": 763
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0519684515091743e-05,
+      "loss": 0.9722,
+      "step": 764
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0482745723008325e-05,
+      "loss": 0.9138,
+      "step": 765
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.044583703275399e-05,
+      "loss": 0.9653,
+      "step": 766
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0408958690278828e-05,
+      "loss": 0.9293,
+      "step": 767
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0372110941330692e-05,
+      "loss": 0.9596,
+      "step": 768
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0335294031453584e-05,
+      "loss": 0.9235,
+      "step": 769
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0298508205985977e-05,
+      "loss": 0.9266,
+      "step": 770
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0261753710059221e-05,
+      "loss": 0.9246,
+      "step": 771
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.0225030788595898e-05,
+      "loss": 0.943,
+      "step": 772
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.0188339686308177e-05,
+      "loss": 0.9238,
+      "step": 773
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.0151680647696193e-05,
+      "loss": 0.9783,
+      "step": 774
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.0115053917046428e-05,
+      "loss": 0.933,
+      "step": 775
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.0078459738430052e-05,
+      "loss": 0.9175,
+      "step": 776
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.0041898355701331e-05,
+      "loss": 0.9058,
+      "step": 777
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.0005370012495984e-05,
+      "loss": 0.8996,
+      "step": 778
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.968874952229566e-06,
+      "loss": 0.9221,
+      "step": 779
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.932413418095836e-06,
+      "loss": 0.9027,
+      "step": 780
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.895985653065157e-06,
+      "loss": 0.974,
+      "step": 781
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.859591899882841e-06,
+      "loss": 0.9232,
+      "step": 782
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.823232401067572e-06,
+      "loss": 0.9178,
+      "step": 783
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.786907398909764e-06,
+      "loss": 0.8668,
+      "step": 784
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.75061713546996e-06,
+      "loss": 0.9458,
+      "step": 785
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.714361852577208e-06,
+      "loss": 0.9042,
+      "step": 786
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.678141791827458e-06,
+      "loss": 0.9393,
+      "step": 787
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.641957194581941e-06,
+      "loss": 0.9414,
+      "step": 788
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.605808301965578e-06,
+      "loss": 0.9388,
+      "step": 789
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.569695354865357e-06,
+      "loss": 0.9882,
+      "step": 790
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.53361859392874e-06,
+      "loss": 0.898,
+      "step": 791
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.497578259562046e-06,
+      "loss": 0.9496,
+      "step": 792
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.461574591928872e-06,
+      "loss": 0.9255,
+      "step": 793
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.425607830948446e-06,
+      "loss": 0.9281,
+      "step": 794
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.389678216294094e-06,
+      "loss": 0.8992,
+      "step": 795
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.35378598739159e-06,
+      "loss": 0.9071,
+      "step": 796
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.31793138341758e-06,
+      "loss": 0.9346,
+      "step": 797
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.282114643297998e-06,
+      "loss": 0.9196,
+      "step": 798
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.24633600570645e-06,
+      "loss": 0.8975,
+      "step": 799
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.21059570906264e-06,
+      "loss": 0.8432,
+      "step": 800
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.17489399153078e-06,
+      "loss": 0.9419,
+      "step": 801
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.139231091018e-06,
+      "loss": 0.8897,
+      "step": 802
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.103607245172763e-06,
+      "loss": 0.9583,
+      "step": 803
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.06802269138328e-06,
+      "loss": 0.969,
+      "step": 804
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.032477666775941e-06,
+      "loss": 0.9107,
+      "step": 805
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.996972408213699e-06,
+      "loss": 0.9028,
+      "step": 806
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.96150715229454e-06,
+      "loss": 1.0098,
+      "step": 807
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.926082135349871e-06,
+      "loss": 0.9059,
+      "step": 808
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 8.890697593442961e-06,
+      "loss": 0.8659,
+      "step": 809
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.855353762367364e-06,
+      "loss": 1.0066,
+      "step": 810
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.820050877645347e-06,
+      "loss": 0.8749,
+      "step": 811
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.784789174526312e-06,
+      "loss": 0.9071,
+      "step": 812
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.74956888798525e-06,
+      "loss": 0.8426,
+      "step": 813
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.714390252721159e-06,
+      "loss": 0.9893,
+      "step": 814
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.679253503155477e-06,
+      "loss": 0.9664,
+      "step": 815
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.644158873430537e-06,
+      "loss": 0.9614,
+      "step": 816
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.609106597407994e-06,
+      "loss": 0.8595,
+      "step": 817
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.574096908667253e-06,
+      "loss": 0.8594,
+      "step": 818
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.539130040503948e-06,
+      "loss": 0.912,
+      "step": 819
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.504206225928356e-06,
+      "loss": 0.8807,
+      "step": 820
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.469325697663866e-06,
+      "loss": 0.844,
+      "step": 821
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 8.434488688145408e-06,
+      "loss": 0.8883,
+      "step": 822
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.399695429517926e-06,
+      "loss": 0.9652,
+      "step": 823
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.364946153634801e-06,
+      "loss": 0.9337,
+      "step": 824
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.330241092056338e-06,
+      "loss": 0.9217,
+      "step": 825
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.295580476048204e-06,
+      "loss": 0.9112,
+      "step": 826
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.2609645365799e-06,
+      "loss": 0.9701,
+      "step": 827
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.226393504323202e-06,
+      "loss": 0.8521,
+      "step": 828
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.191867609650646e-06,
+      "loss": 0.9262,
+      "step": 829
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.157387082633977e-06,
+      "loss": 0.8975,
+      "step": 830
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.12295215304262e-06,
+      "loss": 0.9621,
+      "step": 831
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.08856305034216e-06,
+      "loss": 0.8529,
+      "step": 832
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.054220003692789e-06,
+      "loss": 0.8913,
+      "step": 833
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 8.019923241947804e-06,
+      "loss": 0.9164,
+      "step": 834
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.985672993652078e-06,
+      "loss": 1.0326,
+      "step": 835
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.951469487040501e-06,
+      "loss": 0.927,
+      "step": 836
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.91731295003652e-06,
+      "loss": 0.9178,
+      "step": 837
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.883203610250573e-06,
+      "loss": 0.9256,
+      "step": 838
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.849141694978591e-06,
+      "loss": 0.9089,
+      "step": 839
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.81512743120048e-06,
+      "loss": 1.0522,
+      "step": 840
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.781161045578619e-06,
+      "loss": 0.9271,
+      "step": 841
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.747242764456311e-06,
+      "loss": 0.9554,
+      "step": 842
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.71337281385633e-06,
+      "loss": 0.8547,
+      "step": 843
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.679551419479377e-06,
+      "loss": 0.9227,
+      "step": 844
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.645778806702589e-06,
+      "loss": 0.8863,
+      "step": 845
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.612055200578033e-06,
+      "loss": 0.8566,
+      "step": 846
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 7.57838082583122e-06,
+      "loss": 0.8583,
+      "step": 847
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.544755906859568e-06,
+      "loss": 0.8625,
+      "step": 848
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.511180667730958e-06,
+      "loss": 0.9582,
+      "step": 849
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.477655332182213e-06,
+      "loss": 0.834,
+      "step": 850
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.444180123617608e-06,
+      "loss": 1.007,
+      "step": 851
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.4107552651073835e-06,
+      "loss": 0.8334,
+      "step": 852
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.377380979386264e-06,
+      "loss": 0.9075,
+      "step": 853
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.344057488851968e-06,
+      "loss": 0.8303,
+      "step": 854
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.310785015563728e-06,
+      "loss": 0.8706,
+      "step": 855
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.277563781240807e-06,
+      "loss": 0.8678,
+      "step": 856
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.244394007261031e-06,
+      "loss": 0.8915,
+      "step": 857
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.2112759146593054e-06,
+      "loss": 0.9032,
+      "step": 858
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.178209724126146e-06,
+      "loss": 0.927,
+      "step": 859
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 7.145195656006193e-06,
+      "loss": 0.9392,
+      "step": 860
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 7.112233930296772e-06,
+      "loss": 0.8926,
+      "step": 861
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 7.0793247666464085e-06,
+      "loss": 0.8567,
+      "step": 862
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 7.046468384353364e-06,
+      "loss": 0.9115,
+      "step": 863
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 7.013665002364182e-06,
+      "loss": 0.9485,
+      "step": 864
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.980914839272231e-06,
+      "loss": 0.9463,
+      "step": 865
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.948218113316221e-06,
+      "loss": 0.8793,
+      "step": 866
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.915575042378794e-06,
+      "loss": 0.9208,
+      "step": 867
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.882985843985036e-06,
+      "loss": 0.9435,
+      "step": 868
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.850450735301048e-06,
+      "loss": 0.9469,
+      "step": 869
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.817969933132484e-06,
+      "loss": 0.8745,
+      "step": 870
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.785543653923126e-06,
+      "loss": 0.894,
+      "step": 871
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 6.7531721137534e-06,
+      "loss": 0.8571,
+      "step": 872
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.720855528338997e-06,
+      "loss": 0.9925,
+      "step": 873
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.688594113029388e-06,
+      "loss": 0.9521,
+      "step": 874
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.656388082806407e-06,
+      "loss": 0.8844,
+      "step": 875
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.624237652282815e-06,
+      "loss": 0.8511,
+      "step": 876
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.5921430357008735e-06,
+      "loss": 0.884,
+      "step": 877
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.560104446930913e-06,
+      "loss": 0.8456,
+      "step": 878
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.528122099469905e-06,
+      "loss": 0.9186,
+      "step": 879
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.496196206440051e-06,
+      "loss": 0.9124,
+      "step": 880
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.464326980587349e-06,
+      "loss": 0.9302,
+      "step": 881
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.432514634280181e-06,
+      "loss": 0.978,
+      "step": 882
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.400759379507906e-06,
+      "loss": 0.9276,
+      "step": 883
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.369061427879425e-06,
+      "loss": 0.9798,
+      "step": 884
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 6.337420990621795e-06,
+      "loss": 0.8926,
+      "step": 885
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.305838278578816e-06,
+      "loss": 0.9032,
+      "step": 886
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.274313502209613e-06,
+      "loss": 0.9097,
+      "step": 887
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.242846871587247e-06,
+      "loss": 0.9917,
+      "step": 888
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.211438596397318e-06,
+      "loss": 0.9064,
+      "step": 889
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.180088885936539e-06,
+      "loss": 0.8861,
+      "step": 890
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.14879794911138e-06,
+      "loss": 0.8284,
+      "step": 891
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.117565994436658e-06,
+      "loss": 0.9067,
+      "step": 892
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.086393230034141e-06,
+      "loss": 0.9236,
+      "step": 893
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.055279863631174e-06,
+      "loss": 0.8879,
+      "step": 894
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 6.024226102559288e-06,
+      "loss": 0.8689,
+      "step": 895
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 5.993232153752811e-06,
+      "loss": 0.9199,
+      "step": 896
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 5.962298223747508e-06,
+      "loss": 0.9556,
+      "step": 897
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.931424518679189e-06,
+      "loss": 0.8988,
+      "step": 898
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.900611244282341e-06,
+      "loss": 0.9441,
+      "step": 899
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.869858605888759e-06,
+      "loss": 0.9155,
+      "step": 900
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.839166808426172e-06,
+      "loss": 0.9004,
+      "step": 901
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.808536056416881e-06,
+      "loss": 0.9643,
+      "step": 902
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.777966553976396e-06,
+      "loss": 0.9183,
+      "step": 903
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.7474585048120774e-06,
+      "loss": 0.9752,
+      "step": 904
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.717012112221774e-06,
+      "loss": 0.9233,
+      "step": 905
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.686627579092471e-06,
+      "loss": 0.8949,
+      "step": 906
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.6563051078989454e-06,
+      "loss": 0.9216,
+      "step": 907
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.626044900702393e-06,
+      "loss": 1.0398,
+      "step": 908
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.595847159149111e-06,
+      "loss": 0.8968,
+      "step": 909
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 5.565712084469142e-06,
+      "loss": 0.8461,
+      "step": 910
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.535639877474924e-06,
+      "loss": 0.8884,
+      "step": 911
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.505630738559977e-06,
+      "loss": 0.9692,
+      "step": 912
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.475684867697525e-06,
+      "loss": 0.9418,
+      "step": 913
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.445802464439213e-06,
+      "loss": 0.9365,
+      "step": 914
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.415983727913746e-06,
+      "loss": 0.894,
+      "step": 915
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.386228856825568e-06,
+      "loss": 0.8533,
+      "step": 916
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.356538049453548e-06,
+      "loss": 0.8698,
+      "step": 917
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.3269115036496435e-06,
+      "loss": 0.9611,
+      "step": 918
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.297349416837585e-06,
+      "loss": 0.9793,
+      "step": 919
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.267851986011576e-06,
+      "loss": 0.9228,
+      "step": 920
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.2384194077349615e-06,
+      "loss": 0.9306,
+      "step": 921
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 5.209051878138929e-06,
+      "loss": 0.8999,
+      "step": 922
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.179749592921197e-06,
+      "loss": 0.9408,
+      "step": 923
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.150512747344714e-06,
+      "loss": 0.9673,
+      "step": 924
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.12134153623635e-06,
+      "loss": 0.9741,
+      "step": 925
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.09223615398561e-06,
+      "loss": 0.8946,
+      "step": 926
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.06319679454333e-06,
+      "loss": 0.8995,
+      "step": 927
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.034223651420385e-06,
+      "loss": 0.817,
+      "step": 928
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 5.005316917686404e-06,
+      "loss": 0.9278,
+      "step": 929
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.9764767859684844e-06,
+      "loss": 0.9351,
+      "step": 930
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.94770344844989e-06,
+      "loss": 0.9465,
+      "step": 931
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.9189970968687995e-06,
+      "loss": 0.9035,
+      "step": 932
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.8903579225170105e-06,
+      "loss": 0.8824,
+      "step": 933
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.861786116238669e-06,
+      "loss": 0.9253,
+      "step": 934
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.833281868428995e-06,
+      "loss": 0.8165,
+      "step": 935
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.804845369033025e-06,
+      "loss": 0.9724,
+      "step": 936
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.776476807544318e-06,
+      "loss": 0.9704,
+      "step": 937
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.748176373003731e-06,
+      "loss": 0.9293,
+      "step": 938
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.71994425399813e-06,
+      "loss": 0.9149,
+      "step": 939
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.6917806386591525e-06,
+      "loss": 0.9284,
+      "step": 940
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.663685714661936e-06,
+      "loss": 0.8985,
+      "step": 941
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.635659669223888e-06,
+      "loss": 0.7858,
+      "step": 942
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.607702689103408e-06,
+      "loss": 0.8616,
+      "step": 943
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.5798149605986785e-06,
+      "loss": 0.9155,
+      "step": 944
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.551996669546402e-06,
+      "loss": 0.8642,
+      "step": 945
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.5242480013205645e-06,
+      "loss": 0.8928,
+      "step": 946
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.496569140831208e-06,
+      "loss": 0.8577,
+      "step": 947
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.468960272523196e-06,
+      "loss": 0.9069,
+      "step": 948
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.441421580374968e-06,
+      "loss": 0.9197,
+      "step": 949
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.413953247897341e-06,
+      "loss": 0.9451,
+      "step": 950
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.386555458132265e-06,
+      "loss": 0.8569,
+      "step": 951
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.359228393651622e-06,
+      "loss": 0.8836,
+      "step": 952
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.331972236555992e-06,
+      "loss": 0.8856,
+      "step": 953
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.304787168473448e-06,
+      "loss": 0.9374,
+      "step": 954
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.277673370558333e-06,
+      "loss": 0.8898,
+      "step": 955
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.250631023490079e-06,
+      "loss": 0.8957,
+      "step": 956
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.223660307471981e-06,
+      "loss": 0.9062,
+      "step": 957
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.1967614022300065e-06,
+      "loss": 0.9192,
+      "step": 958
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.169934487011588e-06,
+      "loss": 0.905,
+      "step": 959
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.143179740584446e-06,
+      "loss": 0.9047,
+      "step": 960
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.116497341235368e-06,
+      "loss": 0.9244,
+      "step": 961
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.089887466769056e-06,
+      "loss": 0.9878,
+      "step": 962
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.06335029450692e-06,
+      "loss": 0.8381,
+      "step": 963
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.036886001285898e-06,
+      "loss": 0.9131,
+      "step": 964
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.010494763457284e-06,
+      "loss": 0.8756,
+      "step": 965
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.984176756885551e-06,
+      "loss": 0.8406,
+      "step": 966
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.95793215694717e-06,
+      "loss": 0.9298,
+      "step": 967
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.931761138529454e-06,
+      "loss": 0.882,
+      "step": 968
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.905663876029388e-06,
+      "loss": 0.9165,
+      "step": 969
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.8796405433524685e-06,
+      "loss": 0.9273,
+      "step": 970
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.853691313911536e-06,
+      "loss": 0.9951,
+      "step": 971
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.827816360625638e-06,
+      "loss": 0.981,
+      "step": 972
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.8020158559188438e-06,
+      "loss": 0.9606,
+      "step": 973
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.776289971719135e-06,
+      "loss": 0.9386,
+      "step": 974
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.7506388794572375e-06,
+      "loss": 0.9781,
+      "step": 975
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.7250627500654764e-06,
+      "loss": 0.8723,
+      "step": 976
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.6995617539766495e-06,
+      "loss": 0.9369,
+      "step": 977
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.6741360611228837e-06,
+      "loss": 0.8287,
+      "step": 978
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.6487858409345048e-06,
+      "loss": 0.9418,
+      "step": 979
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.623511262338904e-06,
+      "loss": 0.9153,
+      "step": 980
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.5983124937594225e-06,
+      "loss": 0.9466,
+      "step": 981
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.5731897031142202e-06,
+      "loss": 0.9911,
+      "step": 982
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.5481430578151563e-06,
+      "loss": 1.0334,
+      "step": 983
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.523172724766688e-06,
+      "loss": 0.8694,
+      "step": 984
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.4982788703647267e-06,
+      "loss": 0.8907,
+      "step": 985
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.4734616604955678e-06,
+      "loss": 0.9765,
+      "step": 986
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.448721260534761e-06,
+      "loss": 0.8967,
+      "step": 987
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.424057835346015e-06,
+      "loss": 0.8573,
+      "step": 988
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.399471549280097e-06,
+      "loss": 0.9149,
+      "step": 989
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.374962566173743e-06,
+      "loss": 0.7952,
+      "step": 990
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.3505310493485513e-06,
+      "loss": 0.8786,
+      "step": 991
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.326177161609917e-06,
+      "loss": 1.0054,
+      "step": 992
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.3019010652459284e-06,
+      "loss": 0.9059,
+      "step": 993
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.2777029220262913e-06,
+      "loss": 0.9266,
+      "step": 994
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.253582893201256e-06,
+      "loss": 0.9261,
+      "step": 995
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.22954113950054e-06,
+      "loss": 0.8389,
+      "step": 996
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.205577821132242e-06,
+      "loss": 0.859,
+      "step": 997
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.1816930977818e-06,
+      "loss": 0.9442,
+      "step": 998
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.1578871286109134e-06,
+      "loss": 0.9438,
+      "step": 999
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.134160072256481e-06,
+      "loss": 0.8908,
+      "step": 1000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.110512086829551e-06,
+      "loss": 0.971,
+      "step": 1001
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.0869433299142556e-06,
+      "loss": 0.9553,
+      "step": 1002
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.0634539585667765e-06,
+      "loss": 0.948,
+      "step": 1003
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.0400441293142844e-06,
+      "loss": 0.93,
+      "step": 1004
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.016713998153905e-06,
+      "loss": 0.9653,
+      "step": 1005
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.9934637205516763e-06,
+      "loss": 0.8828,
+      "step": 1006
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.9702934514415094e-06,
+      "loss": 0.9352,
+      "step": 1007
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.9472033452241674e-06,
+      "loss": 0.8574,
+      "step": 1008
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.9241935557662124e-06,
+      "loss": 0.9766,
+      "step": 1009
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.9012642363990134e-06,
+      "loss": 0.8619,
+      "step": 1010
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.8784155399176967e-06,
+      "loss": 0.8757,
+      "step": 1011
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.8556476185801434e-06,
+      "loss": 0.904,
+      "step": 1012
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.8329606241059656e-06,
+      "loss": 0.8383,
+      "step": 1013
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.8103547076755087e-06,
+      "loss": 0.9893,
+      "step": 1014
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.7878300199288194e-06,
+      "loss": 0.8978,
+      "step": 1015
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.7653867109646656e-06,
+      "loss": 0.8993,
+      "step": 1016
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.743024930339532e-06,
+      "loss": 0.8902,
+      "step": 1017
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.7207448270666134e-06,
+      "loss": 0.8769,
+      "step": 1018
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.698546549614829e-06,
+      "loss": 0.9586,
+      "step": 1019
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.676430245907836e-06,
+      "loss": 0.9119,
+      "step": 1020
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.654396063323026e-06,
+      "loss": 0.9516,
+      "step": 1021
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.6324441486905708e-06,
+      "loss": 0.8741,
+      "step": 1022
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.6105746482924233e-06,
+      "loss": 0.8818,
+      "step": 1023
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.588787707861353e-06,
+      "loss": 0.8651,
+      "step": 1024
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.5670834725799636e-06,
+      "loss": 0.8875,
+      "step": 1025
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.545462087079741e-06,
+      "loss": 0.9192,
+      "step": 1026
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.5239236954400745e-06,
+      "loss": 0.9061,
+      "step": 1027
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.5024684411873094e-06,
+      "loss": 0.9128,
+      "step": 1028
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.4810964672937797e-06,
+      "loss": 0.9011,
+      "step": 1029
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.459807916176862e-06,
+      "loss": 0.9884,
+      "step": 1030
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.438602929698026e-06,
+      "loss": 0.8248,
+      "step": 1031
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.4174816491618913e-06,
+      "loss": 0.9054,
+      "step": 1032
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.396444215315271e-06,
+      "loss": 0.9364,
+      "step": 1033
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.375490768346255e-06,
+      "loss": 0.887,
+      "step": 1034
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.3546214478832655e-06,
+      "loss": 0.8645,
+      "step": 1035
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.3338363929941248e-06,
+      "loss": 0.91,
+      "step": 1036
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.3131357421851307e-06,
+      "loss": 0.8663,
+      "step": 1037
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.292519633400138e-06,
+      "loss": 0.935,
+      "step": 1038
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.271988204019627e-06,
+      "loss": 0.8133,
+      "step": 1039
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.2515415908598014e-06,
+      "loss": 0.9104,
+      "step": 1040
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.2311799301716713e-06,
+      "loss": 0.8804,
+      "step": 1041
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.210903357640143e-06,
+      "loss": 0.9427,
+      "step": 1042
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1907120083831227e-06,
+      "loss": 0.8962,
+      "step": 1043
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.170606016950606e-06,
+      "loss": 0.8068,
+      "step": 1044
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1505855173237805e-06,
+      "loss": 0.9421,
+      "step": 1045
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.130650642914148e-06,
+      "loss": 0.8753,
+      "step": 1046
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.110801526562619e-06,
+      "loss": 0.9257,
+      "step": 1047
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.0910383005386384e-06,
+      "loss": 0.9131,
+      "step": 1048
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0713610965392984e-06,
+      "loss": 0.9031,
+      "step": 1049
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0517700456884612e-06,
+      "loss": 0.8097,
+      "step": 1050
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.032265278535889e-06,
+      "loss": 0.9228,
+      "step": 1051
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0128469250563674e-06,
+      "loss": 0.8525,
+      "step": 1052
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.993515114648848e-06,
+      "loss": 0.9544,
+      "step": 1053
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.974269976135578e-06,
+      "loss": 1.0175,
+      "step": 1054
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9551116377612466e-06,
+      "loss": 0.885,
+      "step": 1055
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.936040227192132e-06,
+      "loss": 1.0087,
+      "step": 1056
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9170558715152382e-06,
+      "loss": 0.9186,
+      "step": 1057
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.8981586972374687e-06,
+      "loss": 0.8634,
+      "step": 1058
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.8793488302847655e-06,
+      "loss": 0.878,
+      "step": 1059
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.8606263960012804e-06,
+      "loss": 0.8705,
+      "step": 1060
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.8419915191485342e-06,
+      "loss": 0.9127,
+      "step": 1061
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.8234443239045934e-06,
+      "loss": 1.0295,
+      "step": 1062
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.8049849338632229e-06,
+      "loss": 0.9081,
+      "step": 1063
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.786613472033088e-06,
+      "loss": 0.9098,
+      "step": 1064
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.7683300608369186e-06,
+      "loss": 0.8893,
+      "step": 1065
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.7501348221107012e-06,
+      "loss": 0.9547,
+      "step": 1066
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.7320278771028585e-06,
+      "loss": 0.8818,
+      "step": 1067
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.7140093464734512e-06,
+      "loss": 0.9251,
+      "step": 1068
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.6960793502933641e-06,
+      "loss": 0.9055,
+      "step": 1069
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.6782380080435122e-06,
+      "loss": 0.9564,
+      "step": 1070
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.6604854386140484e-06,
+      "loss": 1.0138,
+      "step": 1071
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.64282176030356e-06,
+      "loss": 0.9017,
+      "step": 1072
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.6252470908182887e-06,
+      "loss": 0.9158,
+      "step": 1073
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.6077615472713468e-06,
+      "loss": 0.8742,
+      "step": 1074
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.5903652461819323e-06,
+      "loss": 0.8461,
+      "step": 1075
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.5730583034745499e-06,
+      "loss": 0.9001,
+      "step": 1076
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.555840834478251e-06,
+      "loss": 0.9565,
+      "step": 1077
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.5387129539258488e-06,
+      "loss": 0.8527,
+      "step": 1078
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.521674775953169e-06,
+      "loss": 0.9138,
+      "step": 1079
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.5047264140982809e-06,
+      "loss": 0.9305,
+      "step": 1080
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4878679813007335e-06,
+      "loss": 0.9682,
+      "step": 1081
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4710995899008239e-06,
+      "loss": 0.8728,
+      "step": 1082
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4544213516388293e-06,
+      "loss": 0.9173,
+      "step": 1083
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4378333776542708e-06,
+      "loss": 0.8543,
+      "step": 1084
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4213357784851733e-06,
+      "loss": 0.9278,
+      "step": 1085
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4049286640673208e-06,
+      "loss": 0.9384,
+      "step": 1086
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.388612143733538e-06,
+      "loss": 0.9573,
+      "step": 1087
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.3723863262129504e-06,
+      "loss": 0.9424,
+      "step": 1088
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.3562513196302607e-06,
+      "loss": 0.8589,
+      "step": 1089
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.3402072315050334e-06,
+      "loss": 0.8682,
+      "step": 1090
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.3242541687509802e-06,
+      "loss": 0.9345,
+      "step": 1091
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.3083922376752305e-06,
+      "loss": 0.9033,
+      "step": 1092
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.292621543977645e-06,
+      "loss": 0.9366,
+      "step": 1093
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2769421927501025e-06,
+      "loss": 0.9592,
+      "step": 1094
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.261354288475793e-06,
+      "loss": 0.8785,
+      "step": 1095
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.245857935028531e-06,
+      "loss": 0.9186,
+      "step": 1096
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2304532356720611e-06,
+      "loss": 0.9535,
+      "step": 1097
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.215140293059364e-06,
+      "loss": 0.8886,
+      "step": 1098
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1999192092319823e-06,
+      "loss": 0.8991,
+      "step": 1099
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1847900856193283e-06,
+      "loss": 0.8352,
+      "step": 1100
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.169753023038021e-06,
+      "loss": 0.9402,
+      "step": 1101
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.154808121691206e-06,
+      "loss": 0.9105,
+      "step": 1102
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.139955481167892e-06,
+      "loss": 0.984,
+      "step": 1103
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.1251952004422778e-06,
+      "loss": 0.8567,
+      "step": 1104
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.110527377873108e-06,
+      "loss": 0.8323,
+      "step": 1105
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.0959521112030063e-06,
+      "loss": 0.8885,
+      "step": 1106
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.0814694975578304e-06,
+      "loss": 0.8541,
+      "step": 1107
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.0670796334460182e-06,
+      "loss": 0.8145,
+      "step": 1108
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.0527826147579517e-06,
+      "loss": 0.9298,
+      "step": 1109
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.0385785367653084e-06,
+      "loss": 0.9454,
+      "step": 1110
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.0244674941204391e-06,
+      "loss": 0.9275,
+      "step": 1111
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.0104495808557307e-06,
+      "loss": 0.9141,
+      "step": 1112
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.965248903829783e-07,
+      "loss": 0.9528,
+      "step": 1113
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.826935154927602e-07,
+      "loss": 0.9063,
+      "step": 1114
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.689555483538326e-07,
+      "loss": 0.8957,
+      "step": 1115
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.55311080512496e-07,
+      "loss": 0.857,
+      "step": 1116
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.417602028920008e-07,
+      "loss": 0.9378,
+      "step": 1117
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.28303005791935e-07,
+      "loss": 0.9682,
+      "step": 1118
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.149395788876259e-07,
+      "loss": 0.8814,
+      "step": 1119
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.016700112295385e-07,
+      "loss": 0.9531,
+      "step": 1120
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.884943912426802e-07,
+      "loss": 0.9501,
+      "step": 1121
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.754128067260225e-07,
+      "loss": 0.8622,
+      "step": 1122
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.624253448519015e-07,
+      "loss": 0.9006,
+      "step": 1123
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.495320921654481e-07,
+      "loss": 0.9238,
+      "step": 1124
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.367331345840057e-07,
+      "loss": 0.8481,
+      "step": 1125
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.2402855739656e-07,
+      "loss": 0.875,
+      "step": 1126
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.114184452631717e-07,
+      "loss": 0.8492,
+      "step": 1127
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.989028822144035e-07,
+      "loss": 0.899,
+      "step": 1128
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.864819516507749e-07,
+      "loss": 0.8797,
+      "step": 1129
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.741557363421936e-07,
+      "loss": 0.9493,
+      "step": 1130
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.619243184274139e-07,
+      "loss": 0.9869,
+      "step": 1131
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.497877794134783e-07,
+      "loss": 0.9622,
+      "step": 1132
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.377462001751906e-07,
+      "loss": 0.8955,
+      "step": 1133
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.257996609545503e-07,
+      "loss": 0.8615,
+      "step": 1134
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.139482413602521e-07,
+      "loss": 0.9611,
+      "step": 1135
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 7.021920203671262e-07,
+      "loss": 0.939,
+      "step": 1136
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.90531076315628e-07,
+      "loss": 1.0209,
+      "step": 1137
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.789654869113133e-07,
+      "loss": 0.8476,
+      "step": 1138
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.674953292243168e-07,
+      "loss": 0.8891,
+      "step": 1139
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.561206796888364e-07,
+      "loss": 0.9066,
+      "step": 1140
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.448416141026364e-07,
+      "loss": 0.9207,
+      "step": 1141
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.336582076265279e-07,
+      "loss": 0.9437,
+      "step": 1142
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.225705347838767e-07,
+      "loss": 0.9325,
+      "step": 1143
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.115786694601039e-07,
+      "loss": 0.9032,
+      "step": 1144
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 6.006826849021929e-07,
+      "loss": 0.8681,
+      "step": 1145
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.898826537182034e-07,
+      "loss": 0.8696,
+      "step": 1146
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.791786478767846e-07,
+      "loss": 0.8958,
+      "step": 1147
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.685707387066996e-07,
+      "loss": 0.9463,
+      "step": 1148
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.580589968963451e-07,
+      "loss": 0.9417,
+      "step": 1149
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.476434924932855e-07,
+      "loss": 0.9069,
+      "step": 1150
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.373242949037832e-07,
+      "loss": 0.9018,
+      "step": 1151
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.27101472892334e-07,
+      "loss": 0.9047,
+      "step": 1152
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.169750945812141e-07,
+      "loss": 1.0009,
+      "step": 1153
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 5.069452274500235e-07,
+      "loss": 0.9415,
+      "step": 1154
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.970119383352373e-07,
+      "loss": 0.8321,
+      "step": 1155
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.871752934297547e-07,
+      "loss": 0.9402,
+      "step": 1156
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.774353582824687e-07,
+      "loss": 0.9061,
+      "step": 1157
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.677921977978178e-07,
+      "loss": 0.9791,
+      "step": 1158
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.5824587623536307e-07,
+      "loss": 0.8981,
+      "step": 1159
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.487964572093517e-07,
+      "loss": 0.829,
+      "step": 1160
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.3944400368830076e-07,
+      "loss": 0.9181,
+      "step": 1161
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.301885779945741e-07,
+      "loss": 0.9249,
+      "step": 1162
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.21030241803963e-07,
+      "loss": 0.92,
+      "step": 1163
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.1196905614528105e-07,
+      "loss": 0.9102,
+      "step": 1164
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.030050813999564e-07,
+      "loss": 0.8954,
+      "step": 1165
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.9413837730162885e-07,
+      "loss": 0.8923,
+      "step": 1166
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.8536900293574985e-07,
+      "loss": 0.9308,
+      "step": 1167
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.766970167391914e-07,
+      "loss": 0.9934,
+      "step": 1168
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.6812247649985307e-07,
+      "loss": 0.9034,
+      "step": 1169
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.5964543935628534e-07,
+      "loss": 0.9447,
+      "step": 1170
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.512659617972969e-07,
+      "loss": 0.9493,
+      "step": 1171
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.4298409966158626e-07,
+      "loss": 0.9032,
+      "step": 1172
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.3479990813737095e-07,
+      "loss": 0.8204,
+      "step": 1173
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.2671344176200877e-07,
+      "loss": 0.9172,
+      "step": 1174
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.187247544216537e-07,
+      "loss": 0.9594,
+      "step": 1175
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.108338993508708e-07,
+      "loss": 0.8955,
+      "step": 1176
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.030409291323083e-07,
+      "loss": 0.9261,
+      "step": 1177
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.953458956963279e-07,
+      "loss": 0.9782,
+      "step": 1178
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.8774885032066823e-07,
+      "loss": 0.9273,
+      "step": 1179
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.802498436301004e-07,
+      "loss": 0.9445,
+      "step": 1180
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.728489255960898e-07,
+      "loss": 0.9797,
+      "step": 1181
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.655461455364644e-07,
+      "loss": 0.9347,
+      "step": 1182
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.583415521150839e-07,
+      "loss": 0.8988,
+      "step": 1183
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.51235193341523e-07,
+      "loss": 0.9675,
+      "step": 1184
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.4422711657073994e-07,
+      "loss": 0.9532,
+      "step": 1185
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.3731736850276852e-07,
+      "loss": 0.8979,
+      "step": 1186
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.3050599518240843e-07,
+      "loss": 0.9359,
+      "step": 1187
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.2379304199890872e-07,
+      "loss": 0.9676,
+      "step": 1188
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.171785536856763e-07,
+      "loss": 0.8714,
+      "step": 1189
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.1066257431997305e-07,
+      "loss": 0.9554,
+      "step": 1190
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.042451473226209e-07,
+      "loss": 0.8396,
+      "step": 1191
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9792631545771545e-07,
+      "loss": 0.9382,
+      "step": 1192
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9170612083233951e-07,
+      "loss": 0.9333,
+      "step": 1193
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.8558460489628005e-07,
+      "loss": 0.8191,
+      "step": 1194
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.795618084417583e-07,
+      "loss": 0.9112,
+      "step": 1195
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.7363777160315009e-07,
+      "loss": 0.8743,
+      "step": 1196
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.6781253385672602e-07,
+      "loss": 0.9295,
+      "step": 1197
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.6208613402037998e-07,
+      "loss": 0.8404,
+      "step": 1198
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.564586102533827e-07,
+      "loss": 0.8898,
+      "step": 1199
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.5093000005611367e-07,
+      "loss": 0.8972,
+      "step": 1200
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4550034026981797e-07,
+      "loss": 0.9398,
+      "step": 1201
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4016966707636304e-07,
+      "loss": 0.8681,
+      "step": 1202
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.3493801599799406e-07,
+      "loss": 0.8424,
+      "step": 1203
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.2980542189709898e-07,
+      "loss": 0.9181,
+      "step": 1204
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.2477191897597383e-07,
+      "loss": 0.9816,
+      "step": 1205
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1983754077659614e-07,
+      "loss": 0.8971,
+      "step": 1206
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1500232018040347e-07,
+      "loss": 0.8958,
+      "step": 1207
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1026628940807026e-07,
+      "loss": 0.8427,
+      "step": 1208
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0562948001929473e-07,
+      "loss": 0.8661,
+      "step": 1209
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0109192291259394e-07,
+      "loss": 0.9822,
+      "step": 1210
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.665364832508905e-08,
+      "loss": 0.8565,
+      "step": 1211
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.231468583230707e-08,
+      "loss": 0.8447,
+      "step": 1212
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 8.807506434798773e-08,
+      "loss": 0.9506,
+      "step": 1213
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 8.393481212388531e-08,
+      "loss": 0.9366,
+      "step": 1214
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 7.989395674958378e-08,
+      "loss": 0.8939,
+      "step": 1215
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 7.595252515231021e-08,
+      "loss": 0.8881,
+      "step": 1216
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 7.211054359676172e-08,
+      "loss": 0.8559,
+      "step": 1217
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.836803768491884e-08,
+      "loss": 0.8862,
+      "step": 1218
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.472503235588734e-08,
+      "loss": 0.9038,
+      "step": 1219
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.118155188572338e-08,
+      "loss": 1.012,
+      "step": 1220
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.7737619887276996e-08,
+      "loss": 0.7791,
+      "step": 1221
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.439325931003047e-08,
+      "loss": 0.8757,
+      "step": 1222
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.11484924399519e-08,
+      "loss": 0.9185,
+      "step": 1223
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.8003340899335226e-08,
+      "loss": 0.9487,
+      "step": 1224
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.4957825646668725e-08,
+      "loss": 0.9287,
+      "step": 1225
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.2011966976488455e-08,
+      "loss": 0.9495,
+      "step": 1226
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.916578451924668e-08,
+      "loss": 0.8305,
+      "step": 1227
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.641929724117199e-08,
+      "loss": 0.9476,
+      "step": 1228
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.377252344415771e-08,
+      "loss": 0.9445,
+      "step": 1229
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.12254807656287e-08,
+      "loss": 0.8579,
+      "step": 1230
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.877818617842476e-08,
+      "loss": 0.8287,
+      "step": 1231
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.6430655990694052e-08,
+      "loss": 0.9856,
+      "step": 1232
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.4182905845776538e-08,
+      "loss": 0.9435,
+      "step": 1233
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.2034950722107372e-08,
+      "loss": 1.0177,
+      "step": 1234
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.9986804933110337e-08,
+      "loss": 0.9644,
+      "step": 1235
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.8038482127106238e-08,
+      "loss": 0.8956,
+      "step": 1236
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.6189995287217983e-08,
+      "loss": 0.8339,
+      "step": 1237
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.4441356731293986e-08,
+      "loss": 0.8086,
+      "step": 1238
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.2792578111813225e-08,
+      "loss": 0.9421,
+      "step": 1239
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.124367041581531e-08,
+      "loss": 0.9646,
+      "step": 1240
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.794643964825544e-09,
+      "loss": 0.8867,
+      "step": 1241
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 8.445508414784975e-09,
+      "loss": 0.9142,
+      "step": 1242
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.196272755992106e-09,
+      "loss": 0.8577,
+      "step": 1243
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.046945313031294e-09,
+      "loss": 0.976,
+      "step": 1244
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.997533744727778e-09,
+      "loss": 0.9328,
+      "step": 1245
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.048045044094395e-09,
+      "loss": 0.8511,
+      "step": 1246
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.198485538278284e-09,
+      "loss": 0.9721,
+      "step": 1247
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.448860888527582e-09,
+      "loss": 0.9169,
+      "step": 1248
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.7991760901514553e-09,
+      "loss": 0.8859,
+      "step": 1249
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2494354724817969e-09,
+      "loss": 0.9526,
+      "step": 1250
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.99642698851577e-10,
+      "loss": 0.795,
+      "step": 1251
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.498007665615367e-10,
+      "loss": 0.9409,
+      "step": 1252
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.9991200687352606e-10,
+      "loss": 0.9111,
+      "step": 1253
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.997808497886336e-11,
+      "loss": 0.996,
+      "step": 1254
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0,
+      "loss": 0.893,
+      "step": 1255
+    },
+    {
+      "epoch": 1.0,
+      "step": 1255,
+      "total_flos": 107943621021696.0,
+      "train_loss": 0.9700585531048566,
+      "train_runtime": 1232.7914,
+      "train_samples_per_second": 48.874,
+      "train_steps_per_second": 1.018
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 1255,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 24000,
+  "total_flos": 107943621021696.0,
+  "train_batch_size": 6,
+  "trial_name": null,
+  "trial_params": null
+}