diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,16891 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "global_step": 2811,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.3529411764705883e-07,
+      "loss": 1.3447,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.7058823529411767e-07,
+      "loss": 1.4492,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 7.058823529411766e-07,
+      "loss": 1.4688,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.411764705882353e-07,
+      "loss": 1.4482,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.1764705882352942e-06,
+      "loss": 1.4541,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.4117647058823531e-06,
+      "loss": 1.4482,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.6470588235294118e-06,
+      "loss": 1.4248,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.8823529411764707e-06,
+      "loss": 1.3877,
+      "step": 8
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.1176470588235296e-06,
+      "loss": 1.3916,
+      "step": 9
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.3529411764705885e-06,
+      "loss": 1.3477,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.5882352941176473e-06,
+      "loss": 1.2939,
+      "step": 11
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.8235294117647062e-06,
+      "loss": 0.2609,
+      "step": 12
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.058823529411765e-06,
+      "loss": 1.334,
+      "step": 13
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.2941176470588236e-06,
+      "loss": 0.2573,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.529411764705883e-06,
+      "loss": 0.312,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.7647058823529414e-06,
+      "loss": 1.3384,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.374,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.235294117647059e-06,
+      "loss": 1.3809,
+      "step": 18
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.4705882352941184e-06,
+      "loss": 1.3652,
+      "step": 19
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.705882352941177e-06,
+      "loss": 1.3613,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.941176470588236e-06,
+      "loss": 1.2832,
+      "step": 21
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.176470588235295e-06,
+      "loss": 1.2832,
+      "step": 22
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.411764705882353e-06,
+      "loss": 0.2936,
+      "step": 23
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.6470588235294125e-06,
+      "loss": 1.2686,
+      "step": 24
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.882352941176471e-06,
+      "loss": 0.3242,
+      "step": 25
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.11764705882353e-06,
+      "loss": 1.2607,
+      "step": 26
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.352941176470589e-06,
+      "loss": 1.2383,
+      "step": 27
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.588235294117647e-06,
+      "loss": 1.2314,
+      "step": 28
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.8235294117647065e-06,
+      "loss": 1.2324,
+      "step": 29
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.058823529411766e-06,
+      "loss": 1.1982,
+      "step": 30
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.294117647058823e-06,
+      "loss": 0.2825,
+      "step": 31
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.529411764705883e-06,
+      "loss": 1.2158,
+      "step": 32
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.764705882352941e-06,
+      "loss": 1.1836,
+      "step": 33
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.207,
+      "step": 34
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.23529411764706e-06,
+      "loss": 1.1885,
+      "step": 35
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.470588235294118e-06,
+      "loss": 1.1172,
+      "step": 36
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.705882352941177e-06,
+      "loss": 1.1875,
+      "step": 37
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.941176470588237e-06,
+      "loss": 1.1689,
+      "step": 38
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.176470588235294e-06,
+      "loss": 1.165,
+      "step": 39
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.411764705882354e-06,
+      "loss": 1.1924,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.647058823529412e-06,
+      "loss": 1.1875,
+      "step": 41
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.882352941176472e-06,
+      "loss": 1.1621,
+      "step": 42
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.011764705882353e-05,
+      "loss": 1.0859,
+      "step": 43
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.035294117647059e-05,
+      "loss": 1.1592,
+      "step": 44
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0588235294117648e-05,
+      "loss": 1.1318,
+      "step": 45
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0823529411764706e-05,
+      "loss": 1.1162,
+      "step": 46
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1058823529411766e-05,
+      "loss": 1.1377,
+      "step": 47
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1294117647058825e-05,
+      "loss": 1.1309,
+      "step": 48
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1529411764705882e-05,
+      "loss": 1.124,
+      "step": 49
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1764705882352942e-05,
+      "loss": 1.1592,
+      "step": 50
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2e-05,
+      "loss": 1.1504,
+      "step": 51
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.223529411764706e-05,
+      "loss": 1.1221,
+      "step": 52
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2470588235294119e-05,
+      "loss": 1.1074,
+      "step": 53
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2705882352941177e-05,
+      "loss": 1.085,
+      "step": 54
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2941176470588238e-05,
+      "loss": 1.0996,
+      "step": 55
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3176470588235294e-05,
+      "loss": 1.1201,
+      "step": 56
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3411764705882353e-05,
+      "loss": 1.127,
+      "step": 57
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3647058823529413e-05,
+      "loss": 1.1172,
+      "step": 58
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3882352941176471e-05,
+      "loss": 1.1016,
+      "step": 59
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4117647058823532e-05,
+      "loss": 0.2997,
+      "step": 60
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.435294117647059e-05,
+      "loss": 1.0308,
+      "step": 61
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4588235294117647e-05,
+      "loss": 1.0908,
+      "step": 62
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4823529411764707e-05,
+      "loss": 1.082,
+      "step": 63
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5058823529411765e-05,
+      "loss": 1.126,
+      "step": 64
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5294117647058822e-05,
+      "loss": 1.0723,
+      "step": 65
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5529411764705882e-05,
+      "loss": 0.2864,
+      "step": 66
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5764705882352943e-05,
+      "loss": 1.1123,
+      "step": 67
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 0.2864,
+      "step": 68
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.623529411764706e-05,
+      "loss": 1.126,
+      "step": 69
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.647058823529412e-05,
+      "loss": 1.0479,
+      "step": 70
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.670588235294118e-05,
+      "loss": 1.085,
+      "step": 71
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6941176470588237e-05,
+      "loss": 1.062,
+      "step": 72
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7176470588235293e-05,
+      "loss": 0.2759,
+      "step": 73
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7411764705882353e-05,
+      "loss": 1.0889,
+      "step": 74
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7647058823529414e-05,
+      "loss": 1.0732,
+      "step": 75
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7882352941176474e-05,
+      "loss": 1.0996,
+      "step": 76
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.811764705882353e-05,
+      "loss": 1.0615,
+      "step": 77
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8352941176470587e-05,
+      "loss": 0.3083,
+      "step": 78
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8588235294117647e-05,
+      "loss": 1.0752,
+      "step": 79
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8823529411764708e-05,
+      "loss": 1.0967,
+      "step": 80
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9058823529411764e-05,
+      "loss": 1.0889,
+      "step": 81
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9294117647058825e-05,
+      "loss": 1.0703,
+      "step": 82
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9529411764705885e-05,
+      "loss": 1.0601,
+      "step": 83
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9764705882352945e-05,
+      "loss": 1.0859,
+      "step": 84
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2e-05,
+      "loss": 1.0596,
+      "step": 85
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999993359236735e-05,
+      "loss": 1.0859,
+      "step": 86
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999997343695575e-05,
+      "loss": 1.0776,
+      "step": 87
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999994023318351e-05,
+      "loss": 1.0527,
+      "step": 88
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999893747964108e-05,
+      "loss": 1.0977,
+      "step": 89
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999833981359296e-05,
+      "loss": 1.0557,
+      "step": 90
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999760933448443e-05,
+      "loss": 1.0312,
+      "step": 91
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999674604328568e-05,
+      "loss": 1.0811,
+      "step": 92
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999574994114336e-05,
+      "loss": 1.0581,
+      "step": 93
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999462102938037e-05,
+      "loss": 0.3038,
+      "step": 94
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999335930949612e-05,
+      "loss": 1.085,
+      "step": 95
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999196478316634e-05,
+      "loss": 1.082,
+      "step": 96
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999043745224324e-05,
+      "loss": 1.0557,
+      "step": 97
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9998877731875524e-05,
+      "loss": 1.042,
+      "step": 98
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999869843849074e-05,
+      "loss": 1.0483,
+      "step": 99
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9998505865308085e-05,
+      "loss": 1.0508,
+      "step": 100
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9998300012583333e-05,
+      "loss": 1.0347,
+      "step": 101
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9998080880589888e-05,
+      "loss": 1.0596,
+      "step": 102
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999784846961879e-05,
+      "loss": 1.0371,
+      "step": 103
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999760277997872e-05,
+      "loss": 1.0293,
+      "step": 104
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9997343811995985e-05,
+      "loss": 1.0278,
+      "step": 105
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9997071566014537e-05,
+      "loss": 1.0469,
+      "step": 106
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999678604239596e-05,
+      "loss": 1.0479,
+      "step": 107
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9996487241519473e-05,
+      "loss": 1.0439,
+      "step": 108
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999617516378193e-05,
+      "loss": 1.0566,
+      "step": 109
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9995849809597814e-05,
+      "loss": 1.0469,
+      "step": 110
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9995511179399253e-05,
+      "loss": 0.998,
+      "step": 111
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999515927363599e-05,
+      "loss": 1.0088,
+      "step": 112
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9994794092775418e-05,
+      "loss": 1.0312,
+      "step": 113
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9994415637302545e-05,
+      "loss": 1.0459,
+      "step": 114
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9994023907720027e-05,
+      "loss": 1.0273,
+      "step": 115
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9993618904548135e-05,
+      "loss": 0.9883,
+      "step": 116
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999320062832477e-05,
+      "loss": 0.2969,
+      "step": 117
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999276907960548e-05,
+      "loss": 1.084,
+      "step": 118
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9992324258963414e-05,
+      "loss": 1.0464,
+      "step": 119
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999186616698937e-05,
+      "loss": 1.041,
+      "step": 120
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999139480429176e-05,
+      "loss": 1.022,
+      "step": 121
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9990910171496626e-05,
+      "loss": 1.0454,
+      "step": 122
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9990412269247637e-05,
+      "loss": 1.0586,
+      "step": 123
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9989901098206083e-05,
+      "loss": 0.2816,
+      "step": 124
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9989376659050878e-05,
+      "loss": 1.0098,
+      "step": 125
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9988838952478552e-05,
+      "loss": 0.3008,
+      "step": 126
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9988287979203264e-05,
+      "loss": 1.0454,
+      "step": 127
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9987723739956793e-05,
+      "loss": 1.084,
+      "step": 128
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9987146235488532e-05,
+      "loss": 1.0186,
+      "step": 129
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9986555466565493e-05,
+      "loss": 1.0615,
+      "step": 130
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9985951433972313e-05,
+      "loss": 0.9497,
+      "step": 131
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.998533413851124e-05,
+      "loss": 1.0239,
+      "step": 132
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.998470358100213e-05,
+      "loss": 1.043,
+      "step": 133
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9984059762282464e-05,
+      "loss": 1.041,
+      "step": 134
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9983402683207334e-05,
+      "loss": 1.0415,
+      "step": 135
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9982732344649434e-05,
+      "loss": 1.0791,
+      "step": 136
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9982048747499082e-05,
+      "loss": 0.2792,
+      "step": 137
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9981351892664195e-05,
+      "loss": 1.0254,
+      "step": 138
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.998064178107031e-05,
+      "loss": 0.2675,
+      "step": 139
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.997991841366055e-05,
+      "loss": 1.0396,
+      "step": 140
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.997918179139567e-05,
+      "loss": 1.0576,
+      "step": 141
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9978431915254017e-05,
+      "loss": 0.3071,
+      "step": 142
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9977668786231536e-05,
+      "loss": 1.0122,
+      "step": 143
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9976892405341776e-05,
+      "loss": 1.0381,
+      "step": 144
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9976102773615894e-05,
+      "loss": 1.0083,
+      "step": 145
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9975299892102638e-05,
+      "loss": 1.0039,
+      "step": 146
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.997448376186836e-05,
+      "loss": 1.0161,
+      "step": 147
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.997365438399701e-05,
+      "loss": 1.0225,
+      "step": 148
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9972811759590117e-05,
+      "loss": 1.0303,
+      "step": 149
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9971955889766825e-05,
+      "loss": 1.0283,
+      "step": 150
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9971086775663856e-05,
+      "loss": 1.0273,
+      "step": 151
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9970204418435524e-05,
+      "loss": 0.9976,
+      "step": 152
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.996930881925374e-05,
+      "loss": 1.0273,
+      "step": 153
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.996839997930799e-05,
+      "loss": 1.0278,
+      "step": 154
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.996747789980536e-05,
+      "loss": 0.2766,
+      "step": 155
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.99665425819705e-05,
+      "loss": 1.0088,
+      "step": 156
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9965594027045668e-05,
+      "loss": 1.0288,
+      "step": 157
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9964632236290682e-05,
+      "loss": 1.0,
+      "step": 158
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9963657210982947e-05,
+      "loss": 1.0073,
+      "step": 159
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.996266895241745e-05,
+      "loss": 1.0244,
+      "step": 160
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9961667461906743e-05,
+      "loss": 1.0142,
+      "step": 161
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9960652740780967e-05,
+      "loss": 1.0312,
+      "step": 162
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995962479038782e-05,
+      "loss": 0.3104,
+      "step": 163
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9958583612092577e-05,
+      "loss": 1.0386,
+      "step": 164
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9957529207278082e-05,
+      "loss": 1.0327,
+      "step": 165
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995646157734475e-05,
+      "loss": 1.0254,
+      "step": 166
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995538072371055e-05,
+      "loss": 1.0078,
+      "step": 167
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9954286647811026e-05,
+      "loss": 0.3153,
+      "step": 168
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9953179351099276e-05,
+      "loss": 0.3143,
+      "step": 169
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9952058835045957e-05,
+      "loss": 0.3008,
+      "step": 170
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9950925101139292e-05,
+      "loss": 0.999,
+      "step": 171
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.994977815088504e-05,
+      "loss": 1.021,
+      "step": 172
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.994861798580654e-05,
+      "loss": 1.0205,
+      "step": 173
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9947444607444662e-05,
+      "loss": 0.958,
+      "step": 174
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.994625801735783e-05,
+      "loss": 1.0083,
+      "step": 175
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9945058217122017e-05,
+      "loss": 1.0244,
+      "step": 176
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9943845208330742e-05,
+      "loss": 1.0078,
+      "step": 177
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9942618992595072e-05,
+      "loss": 1.0151,
+      "step": 178
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9941379571543597e-05,
+      "loss": 0.3308,
+      "step": 179
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9940126946822468e-05,
+      "loss": 1.0156,
+      "step": 180
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9938861120095353e-05,
+      "loss": 0.3149,
+      "step": 181
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.993758209304347e-05,
+      "loss": 0.9941,
+      "step": 182
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9936289867365557e-05,
+      "loss": 0.938,
+      "step": 183
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9934984444777893e-05,
+      "loss": 0.2964,
+      "step": 184
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9933665827014272e-05,
+      "loss": 1.0283,
+      "step": 185
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9932334015826023e-05,
+      "loss": 1.0122,
+      "step": 186
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9930989012981992e-05,
+      "loss": 1.0137,
+      "step": 187
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9929630820268554e-05,
+      "loss": 0.9912,
+      "step": 188
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.992825943948959e-05,
+      "loss": 1.0156,
+      "step": 189
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9926874872466507e-05,
+      "loss": 1.0103,
+      "step": 190
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9925477121038218e-05,
+      "loss": 0.9888,
+      "step": 191
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9924066187061156e-05,
+      "loss": 1.0361,
+      "step": 192
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.992264207240925e-05,
+      "loss": 1.0146,
+      "step": 193
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9921204778973947e-05,
+      "loss": 0.9863,
+      "step": 194
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.991975430866419e-05,
+      "loss": 1.0073,
+      "step": 195
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9918290663406417e-05,
+      "loss": 0.9521,
+      "step": 196
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9916813845144587e-05,
+      "loss": 1.0112,
+      "step": 197
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9915323855840133e-05,
+      "loss": 1.0122,
+      "step": 198
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9913820697471988e-05,
+      "loss": 0.9731,
+      "step": 199
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9912304372036572e-05,
+      "loss": 0.9902,
+      "step": 200
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9910774881547803e-05,
+      "loss": 0.9893,
+      "step": 201
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9909232228037075e-05,
+      "loss": 1.04,
+      "step": 202
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.990767641355327e-05,
+      "loss": 1.0034,
+      "step": 203
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9906107440162743e-05,
+      "loss": 1.0122,
+      "step": 204
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9904525309949332e-05,
+      "loss": 0.9946,
+      "step": 205
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9902930025014347e-05,
+      "loss": 0.9692,
+      "step": 206
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9901321587476573e-05,
+      "loss": 1.0166,
+      "step": 207
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.989969999947226e-05,
+      "loss": 1.0039,
+      "step": 208
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9898065263155118e-05,
+      "loss": 0.9619,
+      "step": 209
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9896417380696334e-05,
+      "loss": 1.0029,
+      "step": 210
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.989475635428454e-05,
+      "loss": 0.9673,
+      "step": 211
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.989308218612584e-05,
+      "loss": 0.9854,
+      "step": 212
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9891394878443783e-05,
+      "loss": 1.0093,
+      "step": 213
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.988969443347937e-05,
+      "loss": 1.0269,
+      "step": 214
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.988798085349105e-05,
+      "loss": 0.9761,
+      "step": 215
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.988625414075472e-05,
+      "loss": 0.998,
+      "step": 216
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9884514297563722e-05,
+      "loss": 0.9839,
+      "step": 217
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9882761326228826e-05,
+      "loss": 0.3733,
+      "step": 218
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9880995229078253e-05,
+      "loss": 0.3003,
+      "step": 219
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9879216008457642e-05,
+      "loss": 0.2975,
+      "step": 220
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9877423666730075e-05,
+      "loss": 1.0239,
+      "step": 221
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9875618206276056e-05,
+      "loss": 1.0166,
+      "step": 222
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9873799629493507e-05,
+      "loss": 1.0142,
+      "step": 223
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.987196793879778e-05,
+      "loss": 0.9951,
+      "step": 224
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9870123136621638e-05,
+      "loss": 0.9873,
+      "step": 225
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9868265225415263e-05,
+      "loss": 1.0034,
+      "step": 226
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.986639420764624e-05,
+      "loss": 1.0083,
+      "step": 227
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.986451008579957e-05,
+      "loss": 1.0259,
+      "step": 228
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9862612862377652e-05,
+      "loss": 0.324,
+      "step": 229
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9860702539900288e-05,
+      "loss": 0.9893,
+      "step": 230
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.985877912090468e-05,
+      "loss": 1.0229,
+      "step": 231
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.985684260794542e-05,
+      "loss": 0.978,
+      "step": 232
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9854893003594492e-05,
+      "loss": 1.0444,
+      "step": 233
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9852930310441275e-05,
+      "loss": 1.0039,
+      "step": 234
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9850954531092515e-05,
+      "loss": 1.0142,
+      "step": 235
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.984896566817236e-05,
+      "loss": 1.0117,
+      "step": 236
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.984696372432231e-05,
+      "loss": 0.9902,
+      "step": 237
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9844948702201265e-05,
+      "loss": 0.9717,
+      "step": 238
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9842920604485474e-05,
+      "loss": 1.0127,
+      "step": 239
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9840879433868562e-05,
+      "loss": 1.0137,
+      "step": 240
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9838825193061518e-05,
+      "loss": 1.0112,
+      "step": 241
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9836757884792682e-05,
+      "loss": 1.0034,
+      "step": 242
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.983467751180776e-05,
+      "loss": 1.0234,
+      "step": 243
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9832584076869805e-05,
+      "loss": 0.3257,
+      "step": 244
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9830477582759213e-05,
+      "loss": 0.9922,
+      "step": 245
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9828358032273737e-05,
+      "loss": 0.3086,
+      "step": 246
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9826225428228455e-05,
+      "loss": 1.0083,
+      "step": 247
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.98240797734558e-05,
+      "loss": 0.9995,
+      "step": 248
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9821921070805522e-05,
+      "loss": 0.9971,
+      "step": 249
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.981974932314471e-05,
+      "loss": 1.0005,
+      "step": 250
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9817564533357775e-05,
+      "loss": 1.0215,
+      "step": 251
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9815366704346456e-05,
+      "loss": 0.9746,
+      "step": 252
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.98131558390298e-05,
+      "loss": 0.9917,
+      "step": 253
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9810931940344175e-05,
+      "loss": 0.9834,
+      "step": 254
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.980869501124326e-05,
+      "loss": 0.9888,
+      "step": 255
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9806445054698037e-05,
+      "loss": 1.0024,
+      "step": 256
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9804182073696793e-05,
+      "loss": 1.0117,
+      "step": 257
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.980190607124511e-05,
+      "loss": 0.9902,
+      "step": 258
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.979961705036587e-05,
+      "loss": 1.0215,
+      "step": 259
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9797315014099238e-05,
+      "loss": 1.0,
+      "step": 260
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.979499996550267e-05,
+      "loss": 0.363,
+      "step": 261
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9792671907650913e-05,
+      "loss": 1.0127,
+      "step": 262
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9790330843635967e-05,
+      "loss": 0.9839,
+      "step": 263
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9787976776567133e-05,
+      "loss": 1.0107,
+      "step": 264
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9785609709570973e-05,
+      "loss": 1.0073,
+      "step": 265
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9783229645791305e-05,
+      "loss": 0.96,
+      "step": 266
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9780836588389225e-05,
+      "loss": 0.999,
+      "step": 267
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9778430540543077e-05,
+      "loss": 1.0376,
+      "step": 268
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9776011505448455e-05,
+      "loss": 0.9194,
+      "step": 269
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9773579486318213e-05,
+      "loss": 0.9536,
+      "step": 270
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9771134486382436e-05,
+      "loss": 0.9795,
+      "step": 271
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9768676508888467e-05,
+      "loss": 0.9648,
+      "step": 272
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.976620555710087e-05,
+      "loss": 0.9937,
+      "step": 273
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9763721634301442e-05,
+      "loss": 0.9658,
+      "step": 274
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.976122474378922e-05,
+      "loss": 0.3225,
+      "step": 275
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.975871488888045e-05,
+      "loss": 0.9971,
+      "step": 276
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9756192072908605e-05,
+      "loss": 1.0181,
+      "step": 277
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.975365629922437e-05,
+      "loss": 0.9927,
+      "step": 278
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.975110757119564e-05,
+      "loss": 1.001,
+      "step": 279
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.974854589220751e-05,
+      "loss": 0.9609,
+      "step": 280
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9745971265662286e-05,
+      "loss": 0.979,
+      "step": 281
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9743383694979462e-05,
+      "loss": 0.9146,
+      "step": 282
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9740783183595726e-05,
+      "loss": 0.9971,
+      "step": 283
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.973816973496496e-05,
+      "loss": 0.9722,
+      "step": 284
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.973554335255822e-05,
+      "loss": 1.0059,
+      "step": 285
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.973290403986374e-05,
+      "loss": 0.9966,
+      "step": 286
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.973025180038693e-05,
+      "loss": 0.9883,
+      "step": 287
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9727586637650373e-05,
+      "loss": 0.9946,
+      "step": 288
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.972490855519381e-05,
+      "loss": 0.9751,
+      "step": 289
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9722217556574144e-05,
+      "loss": 0.9888,
+      "step": 290
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9719513645365426e-05,
+      "loss": 0.9614,
+      "step": 291
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9716796825158873e-05,
+      "loss": 0.9785,
+      "step": 292
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.971406709956283e-05,
+      "loss": 0.9829,
+      "step": 293
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.971132447220279e-05,
+      "loss": 0.306,
+      "step": 294
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.970856894672139e-05,
+      "loss": 1.0122,
+      "step": 295
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.970580052677837e-05,
+      "loss": 0.9736,
+      "step": 296
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9703019216050627e-05,
+      "loss": 0.293,
+      "step": 297
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9700225018232167e-05,
+      "loss": 0.9946,
+      "step": 298
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9697417937034106e-05,
+      "loss": 1.0054,
+      "step": 299
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.969459797618468e-05,
+      "loss": 0.9531,
+      "step": 300
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9691765139429227e-05,
+      "loss": 1.001,
+      "step": 301
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9688919430530183e-05,
+      "loss": 0.9829,
+      "step": 302
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9686060853267088e-05,
+      "loss": 0.9956,
+      "step": 303
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.968318941143657e-05,
+      "loss": 0.9512,
+      "step": 304
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9680305108852335e-05,
+      "loss": 0.9766,
+      "step": 305
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9677407949345184e-05,
+      "loss": 0.9834,
+      "step": 306
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9674497936762984e-05,
+      "loss": 0.3025,
+      "step": 307
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9671575074970678e-05,
+      "loss": 0.9561,
+      "step": 308
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.966863936785027e-05,
+      "loss": 0.9829,
+      "step": 309
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9665690819300827e-05,
+      "loss": 0.9756,
+      "step": 310
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9662729433238477e-05,
+      "loss": 0.9072,
+      "step": 311
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9659755213596388e-05,
+      "loss": 0.9565,
+      "step": 312
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.965676816432478e-05,
+      "loss": 1.0015,
+      "step": 313
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9653768289390904e-05,
+      "loss": 0.9316,
+      "step": 314
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.965075559277906e-05,
+      "loss": 0.9702,
+      "step": 315
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.964773007849057e-05,
+      "loss": 0.9858,
+      "step": 316
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.964469175054377e-05,
+      "loss": 1.0059,
+      "step": 317
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9641640612974028e-05,
+      "loss": 1.0034,
+      "step": 318
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.963857666983372e-05,
+      "loss": 0.9746,
+      "step": 319
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.963549992519223e-05,
+      "loss": 0.3408,
+      "step": 320
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9632410383135946e-05,
+      "loss": 0.9517,
+      "step": 321
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9629308047768254e-05,
+      "loss": 0.9639,
+      "step": 322
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9626192923209524e-05,
+      "loss": 0.9829,
+      "step": 323
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.962306501359712e-05,
+      "loss": 0.9629,
+      "step": 324
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.961992432308538e-05,
+      "loss": 0.3079,
+      "step": 325
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.961677085584562e-05,
+      "loss": 0.9634,
+      "step": 326
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9613604616066137e-05,
+      "loss": 0.9639,
+      "step": 327
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9610425607952165e-05,
+      "loss": 0.9976,
+      "step": 328
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.960723383572592e-05,
+      "loss": 0.9839,
+      "step": 329
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9604029303626563e-05,
+      "loss": 1.0044,
+      "step": 330
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9600812015910203e-05,
+      "loss": 0.9883,
+      "step": 331
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9597581976849886e-05,
+      "loss": 0.958,
+      "step": 332
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9594339190735594e-05,
+      "loss": 0.9561,
+      "step": 333
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9591083661874252e-05,
+      "loss": 0.9634,
+      "step": 334
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.958781539458969e-05,
+      "loss": 0.9429,
+      "step": 335
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9584534393222667e-05,
+      "loss": 0.9746,
+      "step": 336
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.958124066213086e-05,
+      "loss": 0.9697,
+      "step": 337
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.957793420568884e-05,
+      "loss": 0.9551,
+      "step": 338
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.957461502828809e-05,
+      "loss": 0.9624,
+      "step": 339
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9571283134336983e-05,
+      "loss": 0.2908,
+      "step": 340
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9567938528260778e-05,
+      "loss": 1.0,
+      "step": 341
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.956458121450163e-05,
+      "loss": 0.9668,
+      "step": 342
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9561211197518564e-05,
+      "loss": 0.9521,
+      "step": 343
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.955782848178747e-05,
+      "loss": 1.0073,
+      "step": 344
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9554433071801117e-05,
+      "loss": 0.9414,
+      "step": 345
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9551024972069127e-05,
+      "loss": 0.9355,
+      "step": 346
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9547604187117974e-05,
+      "loss": 0.9712,
+      "step": 347
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9544170721490982e-05,
+      "loss": 0.936,
+      "step": 348
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9540724579748323e-05,
+      "loss": 0.9717,
+      "step": 349
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9537265766466993e-05,
+      "loss": 0.9722,
+      "step": 350
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9533794286240828e-05,
+      "loss": 0.9609,
+      "step": 351
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9530310143680483e-05,
+      "loss": 0.9453,
+      "step": 352
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.952681334341343e-05,
+      "loss": 0.9424,
+      "step": 353
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9523303890083954e-05,
+      "loss": 0.9453,
+      "step": 354
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9519781788353148e-05,
+      "loss": 0.9688,
+      "step": 355
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9516247042898895e-05,
+      "loss": 0.9199,
+      "step": 356
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9512699658415882e-05,
+      "loss": 0.9639,
+      "step": 357
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.950913963961557e-05,
+      "loss": 0.9414,
+      "step": 358
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9505566991226214e-05,
+      "loss": 0.9648,
+      "step": 359
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9501981717992836e-05,
+      "loss": 0.9688,
+      "step": 360
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9498383824677223e-05,
+      "loss": 0.937,
+      "step": 361
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9494773316057928e-05,
+      "loss": 0.9976,
+      "step": 362
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9491150196930258e-05,
+      "loss": 0.3168,
+      "step": 363
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.948751447210627e-05,
+      "loss": 0.9995,
+      "step": 364
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9483866146414756e-05,
+      "loss": 0.3029,
+      "step": 365
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9480205224701253e-05,
+      "loss": 0.9585,
+      "step": 366
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9476531711828027e-05,
+      "loss": 0.9556,
+      "step": 367
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.947284561267406e-05,
+      "loss": 0.938,
+      "step": 368
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.946914693213505e-05,
+      "loss": 0.9995,
+      "step": 369
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.946543567512342e-05,
+      "loss": 0.9561,
+      "step": 370
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.946171184656828e-05,
+      "loss": 0.9492,
+      "step": 371
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9457975451415442e-05,
+      "loss": 0.9609,
+      "step": 372
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.945422649462741e-05,
+      "loss": 0.9565,
+      "step": 373
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.945046498118337e-05,
+      "loss": 0.9473,
+      "step": 374
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.944669091607919e-05,
+      "loss": 0.9707,
+      "step": 375
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.94429043043274e-05,
+      "loss": 0.9775,
+      "step": 376
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.94391051509572e-05,
+      "loss": 0.9595,
+      "step": 377
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9435293461014447e-05,
+      "loss": 0.9761,
+      "step": 378
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9431469239561646e-05,
+      "loss": 0.9648,
+      "step": 379
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9427632491677944e-05,
+      "loss": 0.3652,
+      "step": 380
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9423783222459135e-05,
+      "loss": 0.9805,
+      "step": 381
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.941992143701763e-05,
+      "loss": 0.9922,
+      "step": 382
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.941604714048247e-05,
+      "loss": 0.9722,
+      "step": 383
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9412160337999317e-05,
+      "loss": 0.9585,
+      "step": 384
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.940826103473043e-05,
+      "loss": 0.9502,
+      "step": 385
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9404349235854688e-05,
+      "loss": 0.9844,
+      "step": 386
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9400424946567552e-05,
+      "loss": 0.9858,
+      "step": 387
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9396488172081076e-05,
+      "loss": 0.9678,
+      "step": 388
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.93925389176239e-05,
+      "loss": 0.3152,
+      "step": 389
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9388577188441236e-05,
+      "loss": 0.3254,
+      "step": 390
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9384602989794868e-05,
+      "loss": 0.9902,
+      "step": 391
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9380616326963133e-05,
+      "loss": 0.9321,
+      "step": 392
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.937661720524093e-05,
+      "loss": 0.9424,
+      "step": 393
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9372605629939706e-05,
+      "loss": 0.957,
+      "step": 394
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9368581606387442e-05,
+      "loss": 0.3423,
+      "step": 395
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9364545139928657e-05,
+      "loss": 0.9844,
+      "step": 396
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9360496235924396e-05,
+      "loss": 0.9482,
+      "step": 397
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9356434899752216e-05,
+      "loss": 0.9434,
+      "step": 398
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.93523611368062e-05,
+      "loss": 0.3315,
+      "step": 399
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.934827495249692e-05,
+      "loss": 0.3235,
+      "step": 400
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9344176352251456e-05,
+      "loss": 1.019,
+      "step": 401
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9340065341513373e-05,
+      "loss": 0.9194,
+      "step": 402
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.933594192574272e-05,
+      "loss": 0.959,
+      "step": 403
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9331806110416027e-05,
+      "loss": 0.9985,
+      "step": 404
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9327657901026284e-05,
+      "loss": 0.9941,
+      "step": 405
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9323497303082948e-05,
+      "loss": 0.9824,
+      "step": 406
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9319324322111928e-05,
+      "loss": 0.3033,
+      "step": 407
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.931513896365558e-05,
+      "loss": 0.9521,
+      "step": 408
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9310941233272698e-05,
+      "loss": 0.9443,
+      "step": 409
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9306731136538507e-05,
+      "loss": 0.9536,
+      "step": 410
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9302508679044662e-05,
+      "loss": 0.9961,
+      "step": 411
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.929827386639923e-05,
+      "loss": 0.9507,
+      "step": 412
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.929402670422669e-05,
+      "loss": 0.959,
+      "step": 413
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9289767198167918e-05,
+      "loss": 0.9229,
+      "step": 414
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9285495353880187e-05,
+      "loss": 0.958,
+      "step": 415
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9281211177037168e-05,
+      "loss": 0.9766,
+      "step": 416
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.927691467332889e-05,
+      "loss": 0.9829,
+      "step": 417
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9272605848461773e-05,
+      "loss": 0.9365,
+      "step": 418
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.926828470815859e-05,
+      "loss": 0.9697,
+      "step": 419
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.926395125815848e-05,
+      "loss": 0.9341,
+      "step": 420
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9259605504216922e-05,
+      "loss": 0.9551,
+      "step": 421
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9255247452105747e-05,
+      "loss": 0.9644,
+      "step": 422
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.92508771076131e-05,
+      "loss": 0.958,
+      "step": 423
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.924649447654348e-05,
+      "loss": 0.9556,
+      "step": 424
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9242099564717683e-05,
+      "loss": 0.9517,
+      "step": 425
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9237692377972822e-05,
+      "loss": 0.9512,
+      "step": 426
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9233272922162318e-05,
+      "loss": 0.9482,
+      "step": 427
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.922884120315588e-05,
+      "loss": 0.2826,
+      "step": 428
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.922439722683951e-05,
+      "loss": 0.9443,
+      "step": 429
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.921994099911548e-05,
+      "loss": 0.9575,
+      "step": 430
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.921547252590235e-05,
+      "loss": 0.9429,
+      "step": 431
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.921099181313493e-05,
+      "loss": 0.3293,
+      "step": 432
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.920649886676429e-05,
+      "loss": 0.9028,
+      "step": 433
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.920199369275775e-05,
+      "loss": 0.9404,
+      "step": 434
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9197476297098868e-05,
+      "loss": 0.9302,
+      "step": 435
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9192946685787435e-05,
+      "loss": 0.9824,
+      "step": 436
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9188404864839465e-05,
+      "loss": 0.9595,
+      "step": 437
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.918385084028719e-05,
+      "loss": 0.3547,
+      "step": 438
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.917928461817906e-05,
+      "loss": 0.9868,
+      "step": 439
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9174706204579695e-05,
+      "loss": 0.9839,
+      "step": 440
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9170115605569945e-05,
+      "loss": 0.2928,
+      "step": 441
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9165512827246818e-05,
+      "loss": 0.9551,
+      "step": 442
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9160897875723505e-05,
+      "loss": 0.9902,
+      "step": 443
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.915627075712937e-05,
+      "loss": 0.9282,
+      "step": 444
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9151631477609932e-05,
+      "loss": 0.9619,
+      "step": 445
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9146980043326864e-05,
+      "loss": 0.3416,
+      "step": 446
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9142316460457974e-05,
+      "loss": 0.9648,
+      "step": 447
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.913764073519722e-05,
+      "loss": 0.9624,
+      "step": 448
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9132952873754675e-05,
+      "loss": 0.9512,
+      "step": 449
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9128252882356535e-05,
+      "loss": 0.9731,
+      "step": 450
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9123540767245107e-05,
+      "loss": 0.9619,
+      "step": 451
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.91188165346788e-05,
+      "loss": 0.9429,
+      "step": 452
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.911408019093211e-05,
+      "loss": 0.9209,
+      "step": 453
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9109331742295634e-05,
+      "loss": 0.958,
+      "step": 454
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.910457119507603e-05,
+      "loss": 0.9438,
+      "step": 455
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9099798555596033e-05,
+      "loss": 0.9536,
+      "step": 456
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9095013830194432e-05,
+      "loss": 0.9595,
+      "step": 457
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9090217025226083e-05,
+      "loss": 0.9858,
+      "step": 458
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.908540814706187e-05,
+      "loss": 0.9575,
+      "step": 459
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9080587202088715e-05,
+      "loss": 0.9927,
+      "step": 460
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9075754196709574e-05,
+      "loss": 0.9595,
+      "step": 461
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.907090913734341e-05,
+      "loss": 0.3218,
+      "step": 462
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9066052030425206e-05,
+      "loss": 0.9795,
+      "step": 463
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9061182882405938e-05,
+      "loss": 0.9888,
+      "step": 464
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9056301699752578e-05,
+      "loss": 0.9438,
+      "step": 465
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.905140848894809e-05,
+      "loss": 0.3088,
+      "step": 466
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9046503256491395e-05,
+      "loss": 0.9521,
+      "step": 467
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9041586008897392e-05,
+      "loss": 0.9331,
+      "step": 468
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.903665675269694e-05,
+      "loss": 0.9512,
+      "step": 469
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.903171549443684e-05,
+      "loss": 0.9697,
+      "step": 470
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9026762240679843e-05,
+      "loss": 0.333,
+      "step": 471
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.902179699800462e-05,
+      "loss": 0.9424,
+      "step": 472
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9016819773005774e-05,
+      "loss": 0.9526,
+      "step": 473
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9011830572293818e-05,
+      "loss": 0.9097,
+      "step": 474
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9006829402495174e-05,
+      "loss": 0.936,
+      "step": 475
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.900181627025216e-05,
+      "loss": 0.3311,
+      "step": 476
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8996791182222977e-05,
+      "loss": 0.3125,
+      "step": 477
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.899175414508171e-05,
+      "loss": 0.9902,
+      "step": 478
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8986705165518318e-05,
+      "loss": 0.9673,
+      "step": 479
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.898164425023861e-05,
+      "loss": 0.9058,
+      "step": 480
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8976571405964258e-05,
+      "loss": 0.9653,
+      "step": 481
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8971486639432776e-05,
+      "loss": 0.9946,
+      "step": 482
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8966389957397503e-05,
+      "loss": 0.9614,
+      "step": 483
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8961281366627612e-05,
+      "loss": 0.3185,
+      "step": 484
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8956160873908097e-05,
+      "loss": 0.9863,
+      "step": 485
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8951028486039752e-05,
+      "loss": 0.9712,
+      "step": 486
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8945884209839172e-05,
+      "loss": 0.9409,
+      "step": 487
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.894072805213874e-05,
+      "loss": 0.9585,
+      "step": 488
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8935560019786618e-05,
+      "loss": 0.9731,
+      "step": 489
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.893038011964674e-05,
+      "loss": 0.9819,
+      "step": 490
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8925188358598815e-05,
+      "loss": 0.9253,
+      "step": 491
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8919984743538286e-05,
+      "loss": 0.9814,
+      "step": 492
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8914769281376345e-05,
+      "loss": 0.9487,
+      "step": 493
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.890954197903993e-05,
+      "loss": 0.9668,
+      "step": 494
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8904302843471692e-05,
+      "loss": 0.9248,
+      "step": 495
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8899051881630004e-05,
+      "loss": 0.9512,
+      "step": 496
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8893789100488945e-05,
+      "loss": 0.9697,
+      "step": 497
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8888514507038288e-05,
+      "loss": 0.9375,
+      "step": 498
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.888322810828351e-05,
+      "loss": 0.9497,
+      "step": 499
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.887792991124574e-05,
+      "loss": 0.9761,
+      "step": 500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8872619922961802e-05,
+      "loss": 0.3135,
+      "step": 501
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8867298150484167e-05,
+      "loss": 0.937,
+      "step": 502
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8861964600880963e-05,
+      "loss": 0.9746,
+      "step": 503
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.885661928123596e-05,
+      "loss": 0.3066,
+      "step": 504
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8851262198648555e-05,
+      "loss": 0.2814,
+      "step": 505
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8845893360233774e-05,
+      "loss": 0.9629,
+      "step": 506
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.884051277312225e-05,
+      "loss": 0.9707,
+      "step": 507
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.883512044446023e-05,
+      "loss": 0.9771,
+      "step": 508
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8829716381409545e-05,
+      "loss": 0.9185,
+      "step": 509
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.882430059114762e-05,
+      "loss": 0.9111,
+      "step": 510
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8818873080867445e-05,
+      "loss": 0.9385,
+      "step": 511
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8813433857777586e-05,
+      "loss": 0.9492,
+      "step": 512
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8807982929102164e-05,
+      "loss": 0.9224,
+      "step": 513
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.880252030208084e-05,
+      "loss": 0.3162,
+      "step": 514
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.879704598396882e-05,
+      "loss": 0.9248,
+      "step": 515
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.879155998203683e-05,
+      "loss": 0.96,
+      "step": 516
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.878606230357112e-05,
+      "loss": 0.958,
+      "step": 517
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8780552955873445e-05,
+      "loss": 0.3354,
+      "step": 518
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8775031946261065e-05,
+      "loss": 0.9663,
+      "step": 519
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8769499282066716e-05,
+      "loss": 0.9463,
+      "step": 520
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8763954970638628e-05,
+      "loss": 0.9614,
+      "step": 521
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8758399019340487e-05,
+      "loss": 0.9243,
+      "step": 522
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.875283143555145e-05,
+      "loss": 0.9517,
+      "step": 523
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8747252226666113e-05,
+      "loss": 0.8979,
+      "step": 524
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.874166140009452e-05,
+      "loss": 0.9531,
+      "step": 525
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8736058963262138e-05,
+      "loss": 0.9302,
+      "step": 526
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8730444923609865e-05,
+      "loss": 0.9453,
+      "step": 527
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8724819288593997e-05,
+      "loss": 0.9497,
+      "step": 528
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8719182065686242e-05,
+      "loss": 0.3474,
+      "step": 529
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.871353326237369e-05,
+      "loss": 0.9717,
+      "step": 530
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8707872886158806e-05,
+      "loss": 0.9478,
+      "step": 531
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.870220094455945e-05,
+      "loss": 0.9468,
+      "step": 532
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8696517445108807e-05,
+      "loss": 0.9399,
+      "step": 533
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.869082239535545e-05,
+      "loss": 0.9629,
+      "step": 534
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.868511580286326e-05,
+      "loss": 0.3027,
+      "step": 535
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8679397675211475e-05,
+      "loss": 0.3225,
+      "step": 536
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8673668019994632e-05,
+      "loss": 0.9009,
+      "step": 537
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8667926844822593e-05,
+      "loss": 0.9507,
+      "step": 538
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8662174157320515e-05,
+      "loss": 0.9971,
+      "step": 539
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8656409965128843e-05,
+      "loss": 0.3037,
+      "step": 540
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8650634275903304e-05,
+      "loss": 0.9292,
+      "step": 541
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8644847097314897e-05,
+      "loss": 0.3042,
+      "step": 542
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8639048437049875e-05,
+      "loss": 0.9702,
+      "step": 543
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.863323830280975e-05,
+      "loss": 0.8633,
+      "step": 544
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.862741670231126e-05,
+      "loss": 0.9497,
+      "step": 545
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8621583643286387e-05,
+      "loss": 0.9365,
+      "step": 546
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8615739133482315e-05,
+      "loss": 0.9385,
+      "step": 547
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.860988318066145e-05,
+      "loss": 0.9478,
+      "step": 548
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8604015792601395e-05,
+      "loss": 0.9199,
+      "step": 549
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.859813697709493e-05,
+      "loss": 0.9331,
+      "step": 550
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8592246741950027e-05,
+      "loss": 0.9351,
+      "step": 551
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8586345094989807e-05,
+      "loss": 0.9473,
+      "step": 552
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8580432044052567e-05,
+      "loss": 0.96,
+      "step": 553
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.857450759699174e-05,
+      "loss": 0.9053,
+      "step": 554
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8568571761675893e-05,
+      "loss": 0.939,
+      "step": 555
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8562624545988727e-05,
+      "loss": 0.938,
+      "step": 556
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.855666595782904e-05,
+      "loss": 0.9653,
+      "step": 557
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.855069600511076e-05,
+      "loss": 0.2932,
+      "step": 558
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.854471469576289e-05,
+      "loss": 0.9253,
+      "step": 559
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8538722037729523e-05,
+      "loss": 0.9653,
+      "step": 560
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.853271803896982e-05,
+      "loss": 0.9497,
+      "step": 561
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8526702707458013e-05,
+      "loss": 0.9561,
+      "step": 562
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8520676051183377e-05,
+      "loss": 0.9292,
+      "step": 563
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8514638078150236e-05,
+      "loss": 0.9336,
+      "step": 564
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8508588796377936e-05,
+      "loss": 0.9575,
+      "step": 565
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.850252821390084e-05,
+      "loss": 0.9346,
+      "step": 566
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8496456338768345e-05,
+      "loss": 0.917,
+      "step": 567
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8490373179044813e-05,
+      "loss": 0.9219,
+      "step": 568
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.848427874280961e-05,
+      "loss": 0.9233,
+      "step": 569
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8478173038157084e-05,
+      "loss": 0.9619,
+      "step": 570
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.847205607319654e-05,
+      "loss": 0.9375,
+      "step": 571
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8465927856052238e-05,
+      "loss": 0.9385,
+      "step": 572
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8459788394863388e-05,
+      "loss": 0.9185,
+      "step": 573
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8453637697784136e-05,
+      "loss": 0.9229,
+      "step": 574
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8447475772983542e-05,
+      "loss": 0.9229,
+      "step": 575
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.844130262864558e-05,
+      "loss": 0.3158,
+      "step": 576
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8435118272969135e-05,
+      "loss": 0.3179,
+      "step": 577
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.842892271416797e-05,
+      "loss": 0.9258,
+      "step": 578
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8422715960470737e-05,
+      "loss": 0.9448,
+      "step": 579
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.841649802012095e-05,
+      "loss": 0.9331,
+      "step": 580
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8410268901376983e-05,
+      "loss": 0.9321,
+      "step": 581
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8404028612512057e-05,
+      "loss": 0.9438,
+      "step": 582
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.839777716181423e-05,
+      "loss": 0.9556,
+      "step": 583
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8391514557586377e-05,
+      "loss": 0.9629,
+      "step": 584
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8385240808146197e-05,
+      "loss": 0.936,
+      "step": 585
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.837895592182619e-05,
+      "loss": 0.3201,
+      "step": 586
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.837265990697364e-05,
+      "loss": 0.9312,
+      "step": 587
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.836635277195061e-05,
+      "loss": 0.9458,
+      "step": 588
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8360034525133953e-05,
+      "loss": 0.9497,
+      "step": 589
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8353705174915252e-05,
+      "loss": 0.3167,
+      "step": 590
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8347364729700857e-05,
+      "loss": 0.9478,
+      "step": 591
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8341013197911842e-05,
+      "loss": 0.9385,
+      "step": 592
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.833465058798402e-05,
+      "loss": 0.9487,
+      "step": 593
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.83282769083679e-05,
+      "loss": 0.9517,
+      "step": 594
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8321892167528707e-05,
+      "loss": 0.2986,
+      "step": 595
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.831549637394634e-05,
+      "loss": 0.9873,
+      "step": 596
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8309089536115406e-05,
+      "loss": 0.3058,
+      "step": 597
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8302671662545154e-05,
+      "loss": 0.9639,
+      "step": 598
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.82962427617595e-05,
+      "loss": 0.9741,
+      "step": 599
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8289802842297004e-05,
+      "loss": 0.918,
+      "step": 600
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8283351912710867e-05,
+      "loss": 0.9419,
+      "step": 601
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.827688998156891e-05,
+      "loss": 0.9619,
+      "step": 602
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8270417057453554e-05,
+      "loss": 0.937,
+      "step": 603
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.826393314896184e-05,
+      "loss": 0.9478,
+      "step": 604
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8257438264705382e-05,
+      "loss": 0.9404,
+      "step": 605
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8250932413310383e-05,
+      "loss": 0.9307,
+      "step": 606
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8244415603417603e-05,
+      "loss": 0.9253,
+      "step": 607
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8237887843682364e-05,
+      "loss": 0.9385,
+      "step": 608
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8231349142774525e-05,
+      "loss": 0.9155,
+      "step": 609
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8224799509378478e-05,
+      "loss": 0.9487,
+      "step": 610
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8218238952193136e-05,
+      "loss": 0.877,
+      "step": 611
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8211667479931924e-05,
+      "loss": 0.9341,
+      "step": 612
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8205085101322754e-05,
+      "loss": 0.9331,
+      "step": 613
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8198491825108035e-05,
+      "loss": 0.3113,
+      "step": 614
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8191887660044646e-05,
+      "loss": 0.9727,
+      "step": 615
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8185272614903917e-05,
+      "loss": 0.9072,
+      "step": 616
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.817864669847165e-05,
+      "loss": 0.9648,
+      "step": 617
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.817200991954806e-05,
+      "loss": 0.9229,
+      "step": 618
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8165362286947817e-05,
+      "loss": 0.917,
+      "step": 619
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8158703809499983e-05,
+      "loss": 0.9185,
+      "step": 620
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.815203449604803e-05,
+      "loss": 0.9683,
+      "step": 621
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8145354355449836e-05,
+      "loss": 0.8916,
+      "step": 622
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8138663396577633e-05,
+      "loss": 0.3062,
+      "step": 623
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8131961628318048e-05,
+      "loss": 0.9336,
+      "step": 624
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8125249059572042e-05,
+      "loss": 0.9185,
+      "step": 625
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8118525699254942e-05,
+      "loss": 0.9043,
+      "step": 626
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8111791556296386e-05,
+      "loss": 0.9331,
+      "step": 627
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8105046639640355e-05,
+      "loss": 0.3308,
+      "step": 628
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8098290958245116e-05,
+      "loss": 0.9272,
+      "step": 629
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.809152452108325e-05,
+      "loss": 0.9404,
+      "step": 630
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8084747337141622e-05,
+      "loss": 0.9233,
+      "step": 631
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8077959415421366e-05,
+      "loss": 0.9224,
+      "step": 632
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8071160764937875e-05,
+      "loss": 0.9429,
+      "step": 633
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8064351394720794e-05,
+      "loss": 0.9351,
+      "step": 634
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.805753131381401e-05,
+      "loss": 0.9316,
+      "step": 635
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8050700531275632e-05,
+      "loss": 0.9209,
+      "step": 636
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8043859056177976e-05,
+      "loss": 0.9429,
+      "step": 637
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.803700689760757e-05,
+      "loss": 0.9224,
+      "step": 638
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8030144064665127e-05,
+      "loss": 0.9497,
+      "step": 639
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8023270566465537e-05,
+      "loss": 0.9468,
+      "step": 640
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.801638641213785e-05,
+      "loss": 0.8979,
+      "step": 641
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.800949161082528e-05,
+      "loss": 0.9385,
+      "step": 642
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.800258617168517e-05,
+      "loss": 0.9487,
+      "step": 643
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7995670103889003e-05,
+      "loss": 0.9658,
+      "step": 644
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.798874341662237e-05,
+      "loss": 0.9268,
+      "step": 645
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7981806119084966e-05,
+      "loss": 0.9375,
+      "step": 646
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7974858220490586e-05,
+      "loss": 0.3413,
+      "step": 647
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7967899730067098e-05,
+      "loss": 0.915,
+      "step": 648
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.796093065705644e-05,
+      "loss": 0.2961,
+      "step": 649
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7953951010714605e-05,
+      "loss": 0.9121,
+      "step": 650
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7946960800311623e-05,
+      "loss": 0.877,
+      "step": 651
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7939960035131574e-05,
+      "loss": 0.9238,
+      "step": 652
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.793294872447253e-05,
+      "loss": 0.9185,
+      "step": 653
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7925926877646587e-05,
+      "loss": 0.8896,
+      "step": 654
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.791889450397983e-05,
+      "loss": 0.916,
+      "step": 655
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.791185161281232e-05,
+      "loss": 0.9434,
+      "step": 656
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.79047982134981e-05,
+      "loss": 0.9365,
+      "step": 657
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7897734315405152e-05,
+      "loss": 0.9248,
+      "step": 658
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7890659927915418e-05,
+      "loss": 0.9365,
+      "step": 659
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7883575060424763e-05,
+      "loss": 0.3225,
+      "step": 660
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7876479722342972e-05,
+      "loss": 0.9106,
+      "step": 661
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7869373923093733e-05,
+      "loss": 0.9165,
+      "step": 662
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.786225767211464e-05,
+      "loss": 0.9316,
+      "step": 663
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7855130978857152e-05,
+      "loss": 0.9365,
+      "step": 664
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7847993852786612e-05,
+      "loss": 0.9526,
+      "step": 665
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.784084630338221e-05,
+      "loss": 0.9443,
+      "step": 666
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7833688340136982e-05,
+      "loss": 0.9292,
+      "step": 667
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7826519972557797e-05,
+      "loss": 0.3101,
+      "step": 668
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7819341210165347e-05,
+      "loss": 0.9146,
+      "step": 669
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7812152062494115e-05,
+      "loss": 0.937,
+      "step": 670
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7804952539092393e-05,
+      "loss": 0.8999,
+      "step": 671
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.779774264952224e-05,
+      "loss": 0.9097,
+      "step": 672
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.77905224033595e-05,
+      "loss": 0.9614,
+      "step": 673
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7783291810193755e-05,
+      "loss": 0.9385,
+      "step": 674
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7776050879628338e-05,
+      "loss": 0.3066,
+      "step": 675
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.776879962128031e-05,
+      "loss": 0.9238,
+      "step": 676
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.776153804478045e-05,
+      "loss": 0.917,
+      "step": 677
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.775426615977324e-05,
+      "loss": 0.3237,
+      "step": 678
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7746983975916852e-05,
+      "loss": 0.9429,
+      "step": 679
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.773969150288314e-05,
+      "loss": 0.8984,
+      "step": 680
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7732388750357618e-05,
+      "loss": 0.9204,
+      "step": 681
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7725075728039457e-05,
+      "loss": 0.9302,
+      "step": 682
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7717752445641473e-05,
+      "loss": 0.9536,
+      "step": 683
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7710418912890096e-05,
+      "loss": 0.2993,
+      "step": 684
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.770307513952538e-05,
+      "loss": 0.959,
+      "step": 685
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7695721135300975e-05,
+      "loss": 0.9482,
+      "step": 686
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7688356909984125e-05,
+      "loss": 0.9214,
+      "step": 687
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.768098247335564e-05,
+      "loss": 0.9048,
+      "step": 688
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7673597835209904e-05,
+      "loss": 0.9048,
+      "step": 689
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.766620300535484e-05,
+      "loss": 0.9453,
+      "step": 690
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7658797993611907e-05,
+      "loss": 0.9185,
+      "step": 691
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7651382809816094e-05,
+      "loss": 0.9058,
+      "step": 692
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7643957463815904e-05,
+      "loss": 0.9136,
+      "step": 693
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7636521965473324e-05,
+      "loss": 0.9214,
+      "step": 694
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7629076324663827e-05,
+      "loss": 0.9224,
+      "step": 695
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7621620551276366e-05,
+      "loss": 0.8896,
+      "step": 696
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7614154655213343e-05,
+      "loss": 0.9546,
+      "step": 697
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7606678646390617e-05,
+      "loss": 0.9497,
+      "step": 698
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7599192534737453e-05,
+      "loss": 0.9194,
+      "step": 699
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7591696330196562e-05,
+      "loss": 0.9199,
+      "step": 700
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7584190042724047e-05,
+      "loss": 0.939,
+      "step": 701
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7576673682289396e-05,
+      "loss": 0.9199,
+      "step": 702
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.756914725887549e-05,
+      "loss": 0.9062,
+      "step": 703
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.756161078247857e-05,
+      "loss": 0.314,
+      "step": 704
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.755406426310822e-05,
+      "loss": 0.9346,
+      "step": 705
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7546507710787374e-05,
+      "loss": 0.9585,
+      "step": 706
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7538941135552286e-05,
+      "loss": 0.917,
+      "step": 707
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7531364547452523e-05,
+      "loss": 0.3107,
+      "step": 708
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.752377795655095e-05,
+      "loss": 0.3273,
+      "step": 709
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7516181372923722e-05,
+      "loss": 0.9419,
+      "step": 710
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7508574806660256e-05,
+      "loss": 0.9336,
+      "step": 711
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7500958267863235e-05,
+      "loss": 0.8735,
+      "step": 712
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7493331766648588e-05,
+      "loss": 0.8677,
+      "step": 713
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.748569531314547e-05,
+      "loss": 0.8906,
+      "step": 714
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7478048917496255e-05,
+      "loss": 0.9038,
+      "step": 715
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.747039258985653e-05,
+      "loss": 0.9424,
+      "step": 716
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.746272634039506e-05,
+      "loss": 0.9385,
+      "step": 717
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.74550501792938e-05,
+      "loss": 0.9014,
+      "step": 718
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.744736411674786e-05,
+      "loss": 0.9072,
+      "step": 719
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7439668162965504e-05,
+      "loss": 0.9326,
+      "step": 720
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.743196232816814e-05,
+      "loss": 0.9409,
+      "step": 721
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7424246622590284e-05,
+      "loss": 0.9004,
+      "step": 722
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7416521056479577e-05,
+      "loss": 0.9736,
+      "step": 723
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7408785640096744e-05,
+      "loss": 0.9429,
+      "step": 724
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.740104038371561e-05,
+      "loss": 0.9136,
+      "step": 725
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7393285297623048e-05,
+      "loss": 0.333,
+      "step": 726
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7385520392118998e-05,
+      "loss": 0.321,
+      "step": 727
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.737774567751644e-05,
+      "loss": 0.9688,
+      "step": 728
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7369961164141383e-05,
+      "loss": 0.915,
+      "step": 729
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.736216686233285e-05,
+      "loss": 0.9219,
+      "step": 730
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7354362782442864e-05,
+      "loss": 0.9043,
+      "step": 731
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.734654893483643e-05,
+      "loss": 0.9141,
+      "step": 732
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7338725329891532e-05,
+      "loss": 0.9531,
+      "step": 733
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7330891977999114e-05,
+      "loss": 0.3259,
+      "step": 734
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7323048889563062e-05,
+      "loss": 0.9482,
+      "step": 735
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.73151960750002e-05,
+      "loss": 0.9316,
+      "step": 736
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.730733354474025e-05,
+      "loss": 0.9253,
+      "step": 737
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7299461309225867e-05,
+      "loss": 0.9009,
+      "step": 738
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7291579378912574e-05,
+      "loss": 0.96,
+      "step": 739
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.728368776426878e-05,
+      "loss": 0.9165,
+      "step": 740
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7275786475775757e-05,
+      "loss": 0.9229,
+      "step": 741
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7267875523927614e-05,
+      "loss": 0.9277,
+      "step": 742
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.725995491923131e-05,
+      "loss": 0.9321,
+      "step": 743
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7252024672206617e-05,
+      "loss": 0.937,
+      "step": 744
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7244084793386108e-05,
+      "loss": 0.9243,
+      "step": 745
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7236135293315158e-05,
+      "loss": 0.8804,
+      "step": 746
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7228176182551917e-05,
+      "loss": 0.9551,
+      "step": 747
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7220207471667296e-05,
+      "loss": 0.3008,
+      "step": 748
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7212229171244966e-05,
+      "loss": 0.915,
+      "step": 749
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7204241291881322e-05,
+      "loss": 0.3313,
+      "step": 750
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7196243844185488e-05,
+      "loss": 0.9233,
+      "step": 751
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7188236838779297e-05,
+      "loss": 0.916,
+      "step": 752
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.718022028629727e-05,
+      "loss": 0.3118,
+      "step": 753
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.717219419738662e-05,
+      "loss": 0.96,
+      "step": 754
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7164158582707215e-05,
+      "loss": 0.9551,
+      "step": 755
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7156113452931572e-05,
+      "loss": 0.9351,
+      "step": 756
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.714805881874486e-05,
+      "loss": 0.916,
+      "step": 757
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7139994690844865e-05,
+      "loss": 0.3274,
+      "step": 758
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7131921079941965e-05,
+      "loss": 0.2922,
+      "step": 759
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7123837996759165e-05,
+      "loss": 0.9292,
+      "step": 760
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7115745452032023e-05,
+      "loss": 0.9277,
+      "step": 761
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7107643456508674e-05,
+      "loss": 0.9292,
+      "step": 762
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.709953202094981e-05,
+      "loss": 0.8779,
+      "step": 763
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.709141115612865e-05,
+      "loss": 0.9419,
+      "step": 764
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.708328087283095e-05,
+      "loss": 0.9199,
+      "step": 765
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7075141181854956e-05,
+      "loss": 0.9346,
+      "step": 766
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.706699209401143e-05,
+      "loss": 0.2859,
+      "step": 767
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7058833620123602e-05,
+      "loss": 0.9321,
+      "step": 768
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.705066577102717e-05,
+      "loss": 0.9199,
+      "step": 769
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.704248855757028e-05,
+      "loss": 0.8862,
+      "step": 770
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.703430199061353e-05,
+      "loss": 0.9277,
+      "step": 771
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7026106081029922e-05,
+      "loss": 0.9375,
+      "step": 772
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7017900839704877e-05,
+      "loss": 0.8979,
+      "step": 773
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.700968627753621e-05,
+      "loss": 0.9385,
+      "step": 774
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.700146240543411e-05,
+      "loss": 0.9297,
+      "step": 775
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.699322923432114e-05,
+      "loss": 0.9087,
+      "step": 776
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6984986775132202e-05,
+      "loss": 0.9014,
+      "step": 777
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6976735038814548e-05,
+      "loss": 0.9375,
+      "step": 778
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6968474036327733e-05,
+      "loss": 0.9033,
+      "step": 779
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6960203778643638e-05,
+      "loss": 0.9019,
+      "step": 780
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6951924276746425e-05,
+      "loss": 0.895,
+      "step": 781
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6943635541632532e-05,
+      "loss": 0.9146,
+      "step": 782
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6935337584310674e-05,
+      "loss": 0.9175,
+      "step": 783
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6927030415801797e-05,
+      "loss": 0.9438,
+      "step": 784
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.691871404713909e-05,
+      "loss": 0.9487,
+      "step": 785
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6910388489367963e-05,
+      "loss": 0.8945,
+      "step": 786
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6902053753546026e-05,
+      "loss": 0.9282,
+      "step": 787
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6893709850743082e-05,
+      "loss": 0.3271,
+      "step": 788
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6885356792041107e-05,
+      "loss": 0.9077,
+      "step": 789
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6876994588534234e-05,
+      "loss": 0.9443,
+      "step": 790
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.686862325132875e-05,
+      "loss": 0.9448,
+      "step": 791
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.686024279154307e-05,
+      "loss": 0.9272,
+      "step": 792
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.685185322030772e-05,
+      "loss": 0.3267,
+      "step": 793
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6843454548765335e-05,
+      "loss": 0.9355,
+      "step": 794
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.683504678807063e-05,
+      "loss": 0.9385,
+      "step": 795
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6826629949390396e-05,
+      "loss": 0.9478,
+      "step": 796
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.681820404390348e-05,
+      "loss": 0.9258,
+      "step": 797
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.680976908280077e-05,
+      "loss": 0.8984,
+      "step": 798
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.680132507728518e-05,
+      "loss": 0.9316,
+      "step": 799
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.679287203857164e-05,
+      "loss": 0.9048,
+      "step": 800
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.678440997788708e-05,
+      "loss": 0.9282,
+      "step": 801
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.67759389064704e-05,
+      "loss": 0.9087,
+      "step": 802
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6767458835572487e-05,
+      "loss": 0.9097,
+      "step": 803
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6758969776456158e-05,
+      "loss": 0.9062,
+      "step": 804
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6750471740396187e-05,
+      "loss": 0.3,
+      "step": 805
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6741964738679257e-05,
+      "loss": 0.9326,
+      "step": 806
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.673344878260397e-05,
+      "loss": 0.9312,
+      "step": 807
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.672492388348082e-05,
+      "loss": 0.8843,
+      "step": 808
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.671639005263216e-05,
+      "loss": 0.9121,
+      "step": 809
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6707847301392237e-05,
+      "loss": 0.9341,
+      "step": 810
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6699295641107116e-05,
+      "loss": 0.9102,
+      "step": 811
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.669073508313471e-05,
+      "loss": 0.9302,
+      "step": 812
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6682165638844753e-05,
+      "loss": 0.3071,
+      "step": 813
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6673587319618765e-05,
+      "loss": 0.9312,
+      "step": 814
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6665000136850076e-05,
+      "loss": 0.9316,
+      "step": 815
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6656404101943766e-05,
+      "loss": 0.8984,
+      "step": 816
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6647799226316684e-05,
+      "loss": 0.9019,
+      "step": 817
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.663918552139742e-05,
+      "loss": 0.9414,
+      "step": 818
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6630562998626287e-05,
+      "loss": 0.3345,
+      "step": 819
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.662193166945531e-05,
+      "loss": 0.9102,
+      "step": 820
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.661329154534822e-05,
+      "loss": 0.9087,
+      "step": 821
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6604642637780405e-05,
+      "loss": 0.936,
+      "step": 822
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6595984958238952e-05,
+      "loss": 0.9136,
+      "step": 823
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6587318518222572e-05,
+      "loss": 0.9551,
+      "step": 824
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.657864332924162e-05,
+      "loss": 0.9717,
+      "step": 825
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6569959402818075e-05,
+      "loss": 0.9185,
+      "step": 826
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6561266750485517e-05,
+      "loss": 0.9414,
+      "step": 827
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.655256538378911e-05,
+      "loss": 0.938,
+      "step": 828
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6543855314285598e-05,
+      "loss": 0.9263,
+      "step": 829
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.653513655354329e-05,
+      "loss": 0.3113,
+      "step": 830
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6526409113142022e-05,
+      "loss": 0.9077,
+      "step": 831
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6517673004673178e-05,
+      "loss": 0.9131,
+      "step": 832
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6508928239739632e-05,
+      "loss": 0.9194,
+      "step": 833
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6500174829955774e-05,
+      "loss": 0.9004,
+      "step": 834
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6491412786947468e-05,
+      "loss": 0.9253,
+      "step": 835
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.648264212235204e-05,
+      "loss": 0.8813,
+      "step": 836
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.647386284781828e-05,
+      "loss": 0.9326,
+      "step": 837
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6465074975006393e-05,
+      "loss": 0.9272,
+      "step": 838
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6456278515588023e-05,
+      "loss": 0.9058,
+      "step": 839
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6447473481246215e-05,
+      "loss": 0.9531,
+      "step": 840
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.643865988367539e-05,
+      "loss": 0.9448,
+      "step": 841
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6429837734581356e-05,
+      "loss": 0.8379,
+      "step": 842
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6421007045681273e-05,
+      "loss": 0.9087,
+      "step": 843
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6412167828703637e-05,
+      "loss": 0.9365,
+      "step": 844
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.640332009538829e-05,
+      "loss": 0.9448,
+      "step": 845
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6394463857486366e-05,
+      "loss": 0.9106,
+      "step": 846
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.63855991267603e-05,
+      "loss": 0.9248,
+      "step": 847
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6376725914983808e-05,
+      "loss": 0.9453,
+      "step": 848
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.636784423394187e-05,
+      "loss": 0.9199,
+      "step": 849
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6358954095430718e-05,
+      "loss": 0.9092,
+      "step": 850
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6350055511257806e-05,
+      "loss": 0.8984,
+      "step": 851
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.634114849324182e-05,
+      "loss": 0.9282,
+      "step": 852
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6332233053212632e-05,
+      "loss": 0.9111,
+      "step": 853
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6323309203011315e-05,
+      "loss": 0.3684,
+      "step": 854
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6314376954490097e-05,
+      "loss": 0.9502,
+      "step": 855
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.630543631951237e-05,
+      "loss": 0.8999,
+      "step": 856
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6296487309952666e-05,
+      "loss": 0.9258,
+      "step": 857
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6287529937696632e-05,
+      "loss": 0.3223,
+      "step": 858
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6278564214641027e-05,
+      "loss": 0.9272,
+      "step": 859
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6269590152693698e-05,
+      "loss": 0.8765,
+      "step": 860
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.626060776377357e-05,
+      "loss": 0.9033,
+      "step": 861
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6251617059810628e-05,
+      "loss": 0.8931,
+      "step": 862
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.62426180527459e-05,
+      "loss": 0.8779,
+      "step": 863
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6233610754531438e-05,
+      "loss": 0.9351,
+      "step": 864
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.622459517713031e-05,
+      "loss": 0.936,
+      "step": 865
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6215571332516578e-05,
+      "loss": 0.9072,
+      "step": 866
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.620653923267529e-05,
+      "loss": 0.9155,
+      "step": 867
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.619749888960245e-05,
+      "loss": 0.3438,
+      "step": 868
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6188450315305012e-05,
+      "loss": 0.8833,
+      "step": 869
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.617939352180087e-05,
+      "loss": 0.8979,
+      "step": 870
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.617032852111882e-05,
+      "loss": 0.8984,
+      "step": 871
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6161255325298576e-05,
+      "loss": 0.9067,
+      "step": 872
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6152173946390715e-05,
+      "loss": 0.3101,
+      "step": 873
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6143084396456706e-05,
+      "loss": 0.8906,
+      "step": 874
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6133986687568854e-05,
+      "loss": 0.9243,
+      "step": 875
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6124880831810306e-05,
+      "loss": 0.9365,
+      "step": 876
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6115766841275027e-05,
+      "loss": 0.9282,
+      "step": 877
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.610664472806779e-05,
+      "loss": 0.9121,
+      "step": 878
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.609751450430415e-05,
+      "loss": 0.9668,
+      "step": 879
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.608837618211044e-05,
+      "loss": 0.9053,
+      "step": 880
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.607922977362375e-05,
+      "loss": 0.9365,
+      "step": 881
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6070075290991903e-05,
+      "loss": 0.9146,
+      "step": 882
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6060912746373453e-05,
+      "loss": 0.9326,
+      "step": 883
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6051742151937655e-05,
+      "loss": 0.8984,
+      "step": 884
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6042563519864452e-05,
+      "loss": 0.9019,
+      "step": 885
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.6033376862344486e-05,
+      "loss": 0.897,
+      "step": 886
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.6024182191579024e-05,
+      "loss": 0.9238,
+      "step": 887
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.6014979519779997e-05,
+      "loss": 0.9253,
+      "step": 888
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.6005768859169965e-05,
+      "loss": 0.874,
+      "step": 889
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.599655022198208e-05,
+      "loss": 0.9072,
+      "step": 890
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5987323620460106e-05,
+      "loss": 0.915,
+      "step": 891
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.597808906685838e-05,
+      "loss": 0.9092,
+      "step": 892
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5968846573441794e-05,
+      "loss": 0.9087,
+      "step": 893
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5959596152485793e-05,
+      "loss": 0.9302,
+      "step": 894
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5950337816276347e-05,
+      "loss": 0.8965,
+      "step": 895
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.594107157710994e-05,
+      "loss": 0.9258,
+      "step": 896
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5931797447293553e-05,
+      "loss": 0.9272,
+      "step": 897
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5922515439144645e-05,
+      "loss": 0.9009,
+      "step": 898
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5913225564991142e-05,
+      "loss": 0.3391,
+      "step": 899
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5903927837171413e-05,
+      "loss": 0.8911,
+      "step": 900
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5894622268034258e-05,
+      "loss": 0.917,
+      "step": 901
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.58853088699389e-05,
+      "loss": 0.325,
+      "step": 902
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5875987655254947e-05,
+      "loss": 0.8926,
+      "step": 903
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5866658636362396e-05,
+      "loss": 0.9072,
+      "step": 904
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.585732182565161e-05,
+      "loss": 0.9453,
+      "step": 905
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5847977235523293e-05,
+      "loss": 0.3188,
+      "step": 906
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5838624878388498e-05,
+      "loss": 0.8716,
+      "step": 907
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5829264766668578e-05,
+      "loss": 0.9189,
+      "step": 908
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5819896912795186e-05,
+      "loss": 0.9438,
+      "step": 909
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5810521329210266e-05,
+      "loss": 0.3347,
+      "step": 910
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5801138028366026e-05,
+      "loss": 0.9346,
+      "step": 911
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5791747022724917e-05,
+      "loss": 0.9019,
+      "step": 912
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.578234832475963e-05,
+      "loss": 0.9014,
+      "step": 913
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5772941946953072e-05,
+      "loss": 0.876,
+      "step": 914
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.576352790179835e-05,
+      "loss": 0.9097,
+      "step": 915
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.575410620179875e-05,
+      "loss": 0.9111,
+      "step": 916
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.574467685946773e-05,
+      "loss": 0.9214,
+      "step": 917
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.573523988732889e-05,
+      "loss": 0.9033,
+      "step": 918
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.572579529791598e-05,
+      "loss": 0.9092,
+      "step": 919
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5716343103772847e-05,
+      "loss": 0.8955,
+      "step": 920
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5706883317453455e-05,
+      "loss": 0.897,
+      "step": 921
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.569741595152184e-05,
+      "loss": 0.9136,
+      "step": 922
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.568794101855211e-05,
+      "loss": 0.9121,
+      "step": 923
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5678458531128422e-05,
+      "loss": 0.9058,
+      "step": 924
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5668968501844966e-05,
+      "loss": 0.3064,
+      "step": 925
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5659470943305956e-05,
+      "loss": 0.9307,
+      "step": 926
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.564996586812559e-05,
+      "loss": 0.8672,
+      "step": 927
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5640453288928063e-05,
+      "loss": 0.9028,
+      "step": 928
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5630933218347536e-05,
+      "loss": 0.9321,
+      "step": 929
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.562140566902811e-05,
+      "loss": 0.9287,
+      "step": 930
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5611870653623826e-05,
+      "loss": 0.8999,
+      "step": 931
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5602328184798643e-05,
+      "loss": 0.8999,
+      "step": 932
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5592778275226413e-05,
+      "loss": 0.9248,
+      "step": 933
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.558322093759088e-05,
+      "loss": 0.8794,
+      "step": 934
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5573656184585643e-05,
+      "loss": 0.9209,
+      "step": 935
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.556408402891415e-05,
+      "loss": 0.9194,
+      "step": 936
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.555450448328969e-05,
+      "loss": 0.3494,
+      "step": 937
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5544917560435366e-05,
+      "loss": 0.9233,
+      "step": 938
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5535323273084062e-05,
+      "loss": 0.9204,
+      "step": 939
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.552572163397847e-05,
+      "loss": 0.8999,
+      "step": 940
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.551611265587102e-05,
+      "loss": 0.897,
+      "step": 941
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5506496351523915e-05,
+      "loss": 0.915,
+      "step": 942
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5496872733709067e-05,
+      "loss": 0.9048,
+      "step": 943
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.548724181520811e-05,
+      "loss": 0.8867,
+      "step": 944
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.547760360881238e-05,
+      "loss": 0.9106,
+      "step": 945
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5467958127322872e-05,
+      "loss": 0.8687,
+      "step": 946
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5458305383550275e-05,
+      "loss": 0.9141,
+      "step": 947
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5448645390314892e-05,
+      "loss": 0.9072,
+      "step": 948
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5438978160446684e-05,
+      "loss": 0.8232,
+      "step": 949
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5429303706785192e-05,
+      "loss": 0.9194,
+      "step": 950
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5419622042179575e-05,
+      "loss": 0.9253,
+      "step": 951
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.540993317948856e-05,
+      "loss": 0.8896,
+      "step": 952
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5400237131580443e-05,
+      "loss": 0.8926,
+      "step": 953
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5390533911333046e-05,
+      "loss": 0.9023,
+      "step": 954
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5380823531633727e-05,
+      "loss": 0.9082,
+      "step": 955
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5371106005379363e-05,
+      "loss": 0.8975,
+      "step": 956
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.53613813454763e-05,
+      "loss": 0.8682,
+      "step": 957
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5351649564840375e-05,
+      "loss": 0.8838,
+      "step": 958
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.534191067639688e-05,
+      "loss": 0.8784,
+      "step": 959
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5332164693080545e-05,
+      "loss": 0.9106,
+      "step": 960
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5322411627835526e-05,
+      "loss": 0.9067,
+      "step": 961
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5312651493615375e-05,
+      "loss": 0.3069,
+      "step": 962
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5302884303383046e-05,
+      "loss": 0.3168,
+      "step": 963
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.529311007011086e-05,
+      "loss": 0.9263,
+      "step": 964
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5283328806780488e-05,
+      "loss": 0.915,
+      "step": 965
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.527354052638294e-05,
+      "loss": 0.9009,
+      "step": 966
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.526374524191855e-05,
+      "loss": 0.9185,
+      "step": 967
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5253942966396947e-05,
+      "loss": 0.9233,
+      "step": 968
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.524413371283705e-05,
+      "loss": 0.3433,
+      "step": 969
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5234317494267048e-05,
+      "loss": 0.9009,
+      "step": 970
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5224494323724374e-05,
+      "loss": 0.8853,
+      "step": 971
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5214664214255703e-05,
+      "loss": 0.9072,
+      "step": 972
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5204827178916919e-05,
+      "loss": 0.9067,
+      "step": 973
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5194983230773104e-05,
+      "loss": 0.9072,
+      "step": 974
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5185132382898524e-05,
+      "loss": 0.9141,
+      "step": 975
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5175274648376612e-05,
+      "loss": 0.9121,
+      "step": 976
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5165410040299942e-05,
+      "loss": 0.8994,
+      "step": 977
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.515553857177022e-05,
+      "loss": 0.9272,
+      "step": 978
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5145660255898262e-05,
+      "loss": 0.8999,
+      "step": 979
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5135775105803983e-05,
+      "loss": 0.8911,
+      "step": 980
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5125883134616363e-05,
+      "loss": 0.3323,
+      "step": 981
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5115984355473462e-05,
+      "loss": 0.915,
+      "step": 982
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5106078781522363e-05,
+      "loss": 0.8838,
+      "step": 983
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5096166425919176e-05,
+      "loss": 0.9062,
+      "step": 984
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5086247301829028e-05,
+      "loss": 0.9609,
+      "step": 985
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5076321422426033e-05,
+      "loss": 0.8979,
+      "step": 986
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5066388800893266e-05,
+      "loss": 0.9165,
+      "step": 987
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5056449450422773e-05,
+      "loss": 0.9028,
+      "step": 988
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5046503384215521e-05,
+      "loss": 0.9185,
+      "step": 989
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.503655061548141e-05,
+      "loss": 0.3342,
+      "step": 990
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.502659115743923e-05,
+      "loss": 0.9087,
+      "step": 991
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5016625023316669e-05,
+      "loss": 0.9087,
+      "step": 992
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5006652226350272e-05,
+      "loss": 0.918,
+      "step": 993
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.4996672779785431e-05,
+      "loss": 0.8823,
+      "step": 994
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.4986686696876381e-05,
+      "loss": 0.2981,
+      "step": 995
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.4976693990886161e-05,
+      "loss": 0.8647,
+      "step": 996
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.4966694675086611e-05,
+      "loss": 0.9082,
+      "step": 997
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4956688762758347e-05,
+      "loss": 0.9136,
+      "step": 998
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4946676267190751e-05,
+      "loss": 0.8608,
+      "step": 999
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4936657201681943e-05,
+      "loss": 0.8926,
+      "step": 1000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4926631579538775e-05,
+      "loss": 0.3054,
+      "step": 1001
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4916599414076797e-05,
+      "loss": 0.9141,
+      "step": 1002
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4906560718620265e-05,
+      "loss": 0.9102,
+      "step": 1003
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4896515506502092e-05,
+      "loss": 0.9136,
+      "step": 1004
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4886463791063854e-05,
+      "loss": 0.3042,
+      "step": 1005
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4876405585655768e-05,
+      "loss": 0.9165,
+      "step": 1006
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.486634090363666e-05,
+      "loss": 0.9058,
+      "step": 1007
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4856269758373972e-05,
+      "loss": 0.9141,
+      "step": 1008
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4846192163243713e-05,
+      "loss": 0.9214,
+      "step": 1009
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4836108131630476e-05,
+      "loss": 0.9272,
+      "step": 1010
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.482601767692739e-05,
+      "loss": 0.9229,
+      "step": 1011
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4815920812536117e-05,
+      "loss": 0.9121,
+      "step": 1012
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4805817551866839e-05,
+      "loss": 0.8896,
+      "step": 1013
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4795707908338228e-05,
+      "loss": 0.915,
+      "step": 1014
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.478559189537743e-05,
+      "loss": 0.3567,
+      "step": 1015
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4775469526420058e-05,
+      "loss": 0.9209,
+      "step": 1016
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4765340814910163e-05,
+      "loss": 0.8892,
+      "step": 1017
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4755205774300218e-05,
+      "loss": 0.3171,
+      "step": 1018
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4745064418051107e-05,
+      "loss": 0.3381,
+      "step": 1019
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4734916759632097e-05,
+      "loss": 0.876,
+      "step": 1020
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4724762812520825e-05,
+      "loss": 0.896,
+      "step": 1021
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4714602590203288e-05,
+      "loss": 0.3103,
+      "step": 1022
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4704436106173813e-05,
+      "loss": 0.9458,
+      "step": 1023
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4694263373935036e-05,
+      "loss": 0.9209,
+      "step": 1024
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4684084406997903e-05,
+      "loss": 0.8945,
+      "step": 1025
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4673899218881636e-05,
+      "loss": 0.8887,
+      "step": 1026
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4663707823113717e-05,
+      "loss": 0.877,
+      "step": 1027
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4653510233229877e-05,
+      "loss": 0.8804,
+      "step": 1028
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4643306462774071e-05,
+      "loss": 0.3257,
+      "step": 1029
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.463309652529847e-05,
+      "loss": 0.8989,
+      "step": 1030
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.462288043436342e-05,
+      "loss": 0.3047,
+      "step": 1031
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4612658203537455e-05,
+      "loss": 0.8735,
+      "step": 1032
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4602429846397254e-05,
+      "loss": 0.9106,
+      "step": 1033
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4592195376527643e-05,
+      "loss": 0.9082,
+      "step": 1034
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4581954807521555e-05,
+      "loss": 0.9312,
+      "step": 1035
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.457170815298003e-05,
+      "loss": 0.894,
+      "step": 1036
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4561455426512192e-05,
+      "loss": 0.9214,
+      "step": 1037
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4551196641735223e-05,
+      "loss": 0.9048,
+      "step": 1038
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4540931812274359e-05,
+      "loss": 0.8755,
+      "step": 1039
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4530660951762859e-05,
+      "loss": 0.9302,
+      "step": 1040
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4520384073841991e-05,
+      "loss": 0.8901,
+      "step": 1041
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.451010119216102e-05,
+      "loss": 0.8193,
+      "step": 1042
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4499812320377182e-05,
+      "loss": 0.8789,
+      "step": 1043
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4489517472155673e-05,
+      "loss": 0.8438,
+      "step": 1044
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4479216661169618e-05,
+      "loss": 0.8921,
+      "step": 1045
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4468909901100068e-05,
+      "loss": 0.9004,
+      "step": 1046
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4458597205635973e-05,
+      "loss": 0.8848,
+      "step": 1047
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.444827858847417e-05,
+      "loss": 0.9072,
+      "step": 1048
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4437954063319352e-05,
+      "loss": 0.894,
+      "step": 1049
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4427623643884072e-05,
+      "loss": 0.9043,
+      "step": 1050
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4417287343888698e-05,
+      "loss": 0.8804,
+      "step": 1051
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4406945177061418e-05,
+      "loss": 0.8901,
+      "step": 1052
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4396597157138205e-05,
+      "loss": 0.8833,
+      "step": 1053
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4386243297862813e-05,
+      "loss": 0.9009,
+      "step": 1054
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4375883612986744e-05,
+      "loss": 0.9116,
+      "step": 1055
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4365518116269242e-05,
+      "loss": 0.9082,
+      "step": 1056
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.435514682147727e-05,
+      "loss": 0.2952,
+      "step": 1057
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.434476974238549e-05,
+      "loss": 0.9033,
+      "step": 1058
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4334386892776246e-05,
+      "loss": 0.8867,
+      "step": 1059
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4323998286439551e-05,
+      "loss": 0.9146,
+      "step": 1060
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4313603937173058e-05,
+      "loss": 0.9346,
+      "step": 1061
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4303203858782047e-05,
+      "loss": 0.9116,
+      "step": 1062
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4292798065079413e-05,
+      "loss": 0.8809,
+      "step": 1063
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4282386569885636e-05,
+      "loss": 0.9087,
+      "step": 1064
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4271969387028773e-05,
+      "loss": 0.9062,
+      "step": 1065
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.426154653034443e-05,
+      "loss": 0.9141,
+      "step": 1066
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4251118013675758e-05,
+      "loss": 0.9014,
+      "step": 1067
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4240683850873416e-05,
+      "loss": 0.3062,
+      "step": 1068
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.423024405579556e-05,
+      "loss": 0.9062,
+      "step": 1069
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4219798642307839e-05,
+      "loss": 0.8486,
+      "step": 1070
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4209347624283352e-05,
+      "loss": 0.9048,
+      "step": 1071
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4198891015602648e-05,
+      "loss": 0.8325,
+      "step": 1072
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4188428830153698e-05,
+      "loss": 0.9204,
+      "step": 1073
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4177961081831884e-05,
+      "loss": 0.8774,
+      "step": 1074
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4167487784539973e-05,
+      "loss": 0.9028,
+      "step": 1075
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4157008952188102e-05,
+      "loss": 0.314,
+      "step": 1076
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4146524598693758e-05,
+      "loss": 0.3347,
+      "step": 1077
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4136034737981768e-05,
+      "loss": 0.8979,
+      "step": 1078
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4125539383984264e-05,
+      "loss": 0.8892,
+      "step": 1079
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4115038550640678e-05,
+      "loss": 0.8931,
+      "step": 1080
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4104532251897722e-05,
+      "loss": 0.9077,
+      "step": 1081
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4094020501709365e-05,
+      "loss": 0.9136,
+      "step": 1082
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4083503314036813e-05,
+      "loss": 0.9082,
+      "step": 1083
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.40729807028485e-05,
+      "loss": 0.9248,
+      "step": 1084
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4062452682120056e-05,
+      "loss": 0.8809,
+      "step": 1085
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4051919265834305e-05,
+      "loss": 0.9072,
+      "step": 1086
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4041380467981225e-05,
+      "loss": 0.3154,
+      "step": 1087
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4030836302557956e-05,
+      "loss": 0.894,
+      "step": 1088
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4020286783568753e-05,
+      "loss": 0.9272,
+      "step": 1089
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.400973192502499e-05,
+      "loss": 0.9067,
+      "step": 1090
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3999171740945132e-05,
+      "loss": 0.8818,
+      "step": 1091
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3988606245354713e-05,
+      "loss": 0.8784,
+      "step": 1092
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3978035452286325e-05,
+      "loss": 0.8979,
+      "step": 1093
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3967459375779595e-05,
+      "loss": 0.897,
+      "step": 1094
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3956878029881167e-05,
+      "loss": 0.9175,
+      "step": 1095
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3946291428644683e-05,
+      "loss": 0.9146,
+      "step": 1096
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3935699586130767e-05,
+      "loss": 0.8955,
+      "step": 1097
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3925102516407001e-05,
+      "loss": 0.8931,
+      "step": 1098
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3914500233547909e-05,
+      "loss": 0.8784,
+      "step": 1099
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3903892751634949e-05,
+      "loss": 0.9146,
+      "step": 1100
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3893280084756468e-05,
+      "loss": 0.8823,
+      "step": 1101
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3882662247007714e-05,
+      "loss": 0.8828,
+      "step": 1102
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3872039252490796e-05,
+      "loss": 0.8887,
+      "step": 1103
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3861411115314663e-05,
+      "loss": 0.9062,
+      "step": 1104
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3850777849595114e-05,
+      "loss": 0.8828,
+      "step": 1105
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3840139469454745e-05,
+      "loss": 0.8926,
+      "step": 1106
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.382949598902295e-05,
+      "loss": 0.3464,
+      "step": 1107
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3818847422435894e-05,
+      "loss": 0.9028,
+      "step": 1108
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.38081937838365e-05,
+      "loss": 0.917,
+      "step": 1109
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3797535087374427e-05,
+      "loss": 0.8823,
+      "step": 1110
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3786871347206053e-05,
+      "loss": 0.8975,
+      "step": 1111
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3776202577494447e-05,
+      "loss": 0.9155,
+      "step": 1112
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3765528792409368e-05,
+      "loss": 0.8999,
+      "step": 1113
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3754850006127232e-05,
+      "loss": 0.3591,
+      "step": 1114
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3744166232831093e-05,
+      "loss": 0.938,
+      "step": 1115
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3733477486710638e-05,
+      "loss": 0.8945,
+      "step": 1116
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3722783781962155e-05,
+      "loss": 0.9028,
+      "step": 1117
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3712085132788509e-05,
+      "loss": 0.8662,
+      "step": 1118
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3701381553399147e-05,
+      "loss": 0.9004,
+      "step": 1119
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.369067305801005e-05,
+      "loss": 0.8672,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3679959660843736e-05,
+      "loss": 0.9053,
+      "step": 1121
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3669241376129236e-05,
+      "loss": 0.894,
+      "step": 1122
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3658518218102064e-05,
+      "loss": 0.9233,
+      "step": 1123
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3647790201004213e-05,
+      "loss": 0.2965,
+      "step": 1124
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3637057339084125e-05,
+      "loss": 0.8828,
+      "step": 1125
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3626319646596682e-05,
+      "loss": 0.8701,
+      "step": 1126
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.361557713780318e-05,
+      "loss": 0.9038,
+      "step": 1127
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3604829826971304e-05,
+      "loss": 0.8906,
+      "step": 1128
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3594077728375129e-05,
+      "loss": 0.8989,
+      "step": 1129
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3583320856295084e-05,
+      "loss": 0.9048,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3572559225017932e-05,
+      "loss": 0.9023,
+      "step": 1131
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3561792848836764e-05,
+      "loss": 0.9062,
+      "step": 1132
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3551021742050974e-05,
+      "loss": 0.8774,
+      "step": 1133
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3540245918966235e-05,
+      "loss": 0.9009,
+      "step": 1134
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3529465393894485e-05,
+      "loss": 0.3213,
+      "step": 1135
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3518680181153903e-05,
+      "loss": 0.9204,
+      "step": 1136
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3507890295068902e-05,
+      "loss": 0.3174,
+      "step": 1137
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.34970957499701e-05,
+      "loss": 0.8965,
+      "step": 1138
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3486296560194292e-05,
+      "loss": 0.9072,
+      "step": 1139
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.347549274008446e-05,
+      "loss": 0.8955,
+      "step": 1140
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3464684303989723e-05,
+      "loss": 0.3147,
+      "step": 1141
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3453871266265331e-05,
+      "loss": 0.877,
+      "step": 1142
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3443053641272656e-05,
+      "loss": 0.8975,
+      "step": 1143
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3432231443379147e-05,
+      "loss": 0.8975,
+      "step": 1144
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.342140468695834e-05,
+      "loss": 0.894,
+      "step": 1145
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3410573386389818e-05,
+      "loss": 0.9199,
+      "step": 1146
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3399737556059203e-05,
+      "loss": 0.8857,
+      "step": 1147
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3388897210358129e-05,
+      "loss": 0.8896,
+      "step": 1148
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3378052363684238e-05,
+      "loss": 0.9097,
+      "step": 1149
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3367203030441129e-05,
+      "loss": 0.874,
+      "step": 1150
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3356349225038384e-05,
+      "loss": 0.8901,
+      "step": 1151
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3345490961891508e-05,
+      "loss": 0.8799,
+      "step": 1152
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3334628255421932e-05,
+      "loss": 0.897,
+      "step": 1153
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3323761120056996e-05,
+      "loss": 0.8413,
+      "step": 1154
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3312889570229901e-05,
+      "loss": 0.9204,
+      "step": 1155
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3302013620379734e-05,
+      "loss": 0.8994,
+      "step": 1156
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3291133284951418e-05,
+      "loss": 0.9155,
+      "step": 1157
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.328024857839569e-05,
+      "loss": 0.9312,
+      "step": 1158
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3269359515169113e-05,
+      "loss": 0.9053,
+      "step": 1159
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.325846610973402e-05,
+      "loss": 0.8853,
+      "step": 1160
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3247568376558515e-05,
+      "loss": 0.916,
+      "step": 1161
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3236666330116456e-05,
+      "loss": 0.855,
+      "step": 1162
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3225759984887416e-05,
+      "loss": 0.8848,
+      "step": 1163
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3214849355356696e-05,
+      "loss": 0.8838,
+      "step": 1164
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3203934456015275e-05,
+      "loss": 0.9028,
+      "step": 1165
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.31930153013598e-05,
+      "loss": 0.8857,
+      "step": 1166
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3182091905892581e-05,
+      "loss": 0.8916,
+      "step": 1167
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3171164284121556e-05,
+      "loss": 0.856,
+      "step": 1168
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3160232450560268e-05,
+      "loss": 0.8906,
+      "step": 1169
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3149296419727868e-05,
+      "loss": 0.8623,
+      "step": 1170
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3138356206149069e-05,
+      "loss": 0.9146,
+      "step": 1171
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3127411824354146e-05,
+      "loss": 0.8594,
+      "step": 1172
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3116463288878914e-05,
+      "loss": 0.8789,
+      "step": 1173
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.310551061426469e-05,
+      "loss": 0.8555,
+      "step": 1174
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3094553815058304e-05,
+      "loss": 0.8936,
+      "step": 1175
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3083592905812058e-05,
+      "loss": 0.877,
+      "step": 1176
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.307262790108371e-05,
+      "loss": 0.9146,
+      "step": 1177
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3061658815436463e-05,
+      "loss": 0.8672,
+      "step": 1178
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3050685663438931e-05,
+      "loss": 0.8867,
+      "step": 1179
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3039708459665146e-05,
+      "loss": 0.9399,
+      "step": 1180
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3028727218694503e-05,
+      "loss": 0.9053,
+      "step": 1181
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3017741955111762e-05,
+      "loss": 0.832,
+      "step": 1182
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.300675268350704e-05,
+      "loss": 0.8896,
+      "step": 1183
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2995759418475763e-05,
+      "loss": 0.9185,
+      "step": 1184
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2984762174618664e-05,
+      "loss": 0.8828,
+      "step": 1185
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2973760966541764e-05,
+      "loss": 0.897,
+      "step": 1186
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2962755808856341e-05,
+      "loss": 0.877,
+      "step": 1187
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.295174671617893e-05,
+      "loss": 0.8813,
+      "step": 1188
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2940733703131287e-05,
+      "loss": 0.8486,
+      "step": 1189
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2929716784340369e-05,
+      "loss": 0.8916,
+      "step": 1190
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.291869597443833e-05,
+      "loss": 0.8779,
+      "step": 1191
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2907671288062493e-05,
+      "loss": 0.8657,
+      "step": 1192
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2896642739855311e-05,
+      "loss": 0.9106,
+      "step": 1193
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2885610344464394e-05,
+      "loss": 0.8701,
+      "step": 1194
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2874574116542439e-05,
+      "loss": 0.9185,
+      "step": 1195
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2863534070747239e-05,
+      "loss": 0.3547,
+      "step": 1196
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2852490221741669e-05,
+      "loss": 0.8848,
+      "step": 1197
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2841442584193637e-05,
+      "loss": 0.8857,
+      "step": 1198
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.28303911727761e-05,
+      "loss": 0.9082,
+      "step": 1199
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2819336002167018e-05,
+      "loss": 0.9097,
+      "step": 1200
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2808277087049338e-05,
+      "loss": 0.8896,
+      "step": 1201
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2797214442110997e-05,
+      "loss": 0.3032,
+      "step": 1202
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.278614808204487e-05,
+      "loss": 0.3232,
+      "step": 1203
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2775078021548775e-05,
+      "loss": 0.9272,
+      "step": 1204
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2764004275325443e-05,
+      "loss": 0.9053,
+      "step": 1205
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2752926858082493e-05,
+      "loss": 0.9048,
+      "step": 1206
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2741845784532436e-05,
+      "loss": 0.8652,
+      "step": 1207
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2730761069392625e-05,
+      "loss": 0.9155,
+      "step": 1208
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2719672727385249e-05,
+      "loss": 0.8745,
+      "step": 1209
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2708580773237325e-05,
+      "loss": 0.8589,
+      "step": 1210
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2697485221680663e-05,
+      "loss": 0.8848,
+      "step": 1211
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2686386087451838e-05,
+      "loss": 0.8496,
+      "step": 1212
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2675283385292212e-05,
+      "loss": 0.8623,
+      "step": 1213
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2664177129947853e-05,
+      "loss": 0.874,
+      "step": 1214
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2653067336169566e-05,
+      "loss": 0.9243,
+      "step": 1215
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2641954018712863e-05,
+      "loss": 0.3171,
+      "step": 1216
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2630837192337913e-05,
+      "loss": 0.9038,
+      "step": 1217
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2619716871809568e-05,
+      "loss": 0.8789,
+      "step": 1218
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2608593071897311e-05,
+      "loss": 0.894,
+      "step": 1219
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2597465807375238e-05,
+      "loss": 0.8818,
+      "step": 1220
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2586335093022064e-05,
+      "loss": 0.8525,
+      "step": 1221
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2575200943621071e-05,
+      "loss": 0.344,
+      "step": 1222
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2564063373960109e-05,
+      "loss": 0.9199,
+      "step": 1223
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2552922398831577e-05,
+      "loss": 0.8857,
+      "step": 1224
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2541778033032383e-05,
+      "loss": 0.8906,
+      "step": 1225
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2530630291363952e-05,
+      "loss": 0.875,
+      "step": 1226
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2519479188632184e-05,
+      "loss": 0.9146,
+      "step": 1227
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2508324739647444e-05,
+      "loss": 0.9028,
+      "step": 1228
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2497166959224546e-05,
+      "loss": 0.8916,
+      "step": 1229
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2486005862182727e-05,
+      "loss": 0.8721,
+      "step": 1230
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.247484146334562e-05,
+      "loss": 0.8711,
+      "step": 1231
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2463673777541262e-05,
+      "loss": 0.3464,
+      "step": 1232
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2452502819602035e-05,
+      "loss": 0.9033,
+      "step": 1233
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.244132860436468e-05,
+      "loss": 0.876,
+      "step": 1234
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2430151146670261e-05,
+      "loss": 0.8975,
+      "step": 1235
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2418970461364147e-05,
+      "loss": 0.895,
+      "step": 1236
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2407786563296e-05,
+      "loss": 0.8779,
+      "step": 1237
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2396599467319737e-05,
+      "loss": 0.9033,
+      "step": 1238
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2385409188293528e-05,
+      "loss": 0.8877,
+      "step": 1239
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2374215741079783e-05,
+      "loss": 0.8984,
+      "step": 1240
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2363019140545096e-05,
+      "loss": 0.8896,
+      "step": 1241
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2351819401560266e-05,
+      "loss": 0.353,
+      "step": 1242
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.234061653900026e-05,
+      "loss": 0.9199,
+      "step": 1243
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2329410567744184e-05,
+      "loss": 0.9087,
+      "step": 1244
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2318201502675285e-05,
+      "loss": 0.8984,
+      "step": 1245
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2306989358680905e-05,
+      "loss": 0.855,
+      "step": 1246
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2295774150652486e-05,
+      "loss": 0.8706,
+      "step": 1247
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2284555893485541e-05,
+      "loss": 0.325,
+      "step": 1248
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2273334602079621e-05,
+      "loss": 0.9062,
+      "step": 1249
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2262110291338313e-05,
+      "loss": 0.9009,
+      "step": 1250
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2250882976169228e-05,
+      "loss": 0.8936,
+      "step": 1251
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2239652671483935e-05,
+      "loss": 0.8618,
+      "step": 1252
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2228419392198008e-05,
+      "loss": 0.8838,
+      "step": 1253
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2217183153230954e-05,
+      "loss": 0.8945,
+      "step": 1254
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2205943969506207e-05,
+      "loss": 0.8877,
+      "step": 1255
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2194701855951128e-05,
+      "loss": 0.8706,
+      "step": 1256
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2183456827496951e-05,
+      "loss": 0.8975,
+      "step": 1257
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2172208899078794e-05,
+      "loss": 0.9023,
+      "step": 1258
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2160958085635628e-05,
+      "loss": 0.8887,
+      "step": 1259
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2149704402110243e-05,
+      "loss": 0.8555,
+      "step": 1260
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.213844786344925e-05,
+      "loss": 0.8901,
+      "step": 1261
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2127188484603057e-05,
+      "loss": 0.3303,
+      "step": 1262
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.211592628052583e-05,
+      "loss": 0.8833,
+      "step": 1263
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2104661266175502e-05,
+      "loss": 0.8726,
+      "step": 1264
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2093393456513724e-05,
+      "loss": 0.8652,
+      "step": 1265
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2082122866505871e-05,
+      "loss": 0.8901,
+      "step": 1266
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2070849511121014e-05,
+      "loss": 0.8994,
+      "step": 1267
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2059573405331879e-05,
+      "loss": 0.8882,
+      "step": 1268
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2048294564114859e-05,
+      "loss": 0.9146,
+      "step": 1269
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2037013002449984e-05,
+      "loss": 0.8794,
+      "step": 1270
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2025728735320878e-05,
+      "loss": 0.8906,
+      "step": 1271
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2014441777714783e-05,
+      "loss": 0.8975,
+      "step": 1272
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2003152144622493e-05,
+      "loss": 0.3406,
+      "step": 1273
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1991859851038362e-05,
+      "loss": 0.9224,
+      "step": 1274
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.198056491196029e-05,
+      "loss": 0.8921,
+      "step": 1275
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.196926734238967e-05,
+      "loss": 0.8818,
+      "step": 1276
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1957967157331404e-05,
+      "loss": 0.895,
+      "step": 1277
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1946664371793865e-05,
+      "loss": 0.3389,
+      "step": 1278
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1935359000788873e-05,
+      "loss": 0.8921,
+      "step": 1279
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.192405105933169e-05,
+      "loss": 0.9438,
+      "step": 1280
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1912740562440988e-05,
+      "loss": 0.8643,
+      "step": 1281
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1901427525138826e-05,
+      "loss": 0.9048,
+      "step": 1282
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.189011196245066e-05,
+      "loss": 0.8906,
+      "step": 1283
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1878793889405269e-05,
+      "loss": 0.8462,
+      "step": 1284
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1867473321034786e-05,
+      "loss": 0.8457,
+      "step": 1285
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1856150272374662e-05,
+      "loss": 0.8735,
+      "step": 1286
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1844824758463626e-05,
+      "loss": 0.8838,
+      "step": 1287
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.183349679434369e-05,
+      "loss": 0.303,
+      "step": 1288
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1822166395060124e-05,
+      "loss": 0.8672,
+      "step": 1289
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1810833575661422e-05,
+      "loss": 0.3358,
+      "step": 1290
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1799498351199303e-05,
+      "loss": 0.8726,
+      "step": 1291
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1788160736728675e-05,
+      "loss": 0.8677,
+      "step": 1292
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1776820747307615e-05,
+      "loss": 0.8989,
+      "step": 1293
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.176547839799737e-05,
+      "loss": 0.8643,
+      "step": 1294
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1754133703862302e-05,
+      "loss": 0.8867,
+      "step": 1295
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1742786679969902e-05,
+      "loss": 0.8818,
+      "step": 1296
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1731437341390748e-05,
+      "loss": 0.8467,
+      "step": 1297
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1720085703198491e-05,
+      "loss": 0.9058,
+      "step": 1298
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.170873178046985e-05,
+      "loss": 0.8813,
+      "step": 1299
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1697375588284556e-05,
+      "loss": 0.8721,
+      "step": 1300
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1686017141725367e-05,
+      "loss": 0.9014,
+      "step": 1301
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1674656455878043e-05,
+      "loss": 0.335,
+      "step": 1302
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1663293545831302e-05,
+      "loss": 0.3535,
+      "step": 1303
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1651928426676822e-05,
+      "loss": 0.8823,
+      "step": 1304
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1640561113509222e-05,
+      "loss": 0.9082,
+      "step": 1305
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1629191621426024e-05,
+      "loss": 0.8877,
+      "step": 1306
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.161781996552765e-05,
+      "loss": 0.9072,
+      "step": 1307
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1606446160917395e-05,
+      "loss": 0.8833,
+      "step": 1308
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1595070222701408e-05,
+      "loss": 0.8701,
+      "step": 1309
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1583692165988676e-05,
+      "loss": 0.8862,
+      "step": 1310
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1572312005890986e-05,
+      "loss": 0.895,
+      "step": 1311
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1560929757522934e-05,
+      "loss": 0.8901,
+      "step": 1312
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1549545436001888e-05,
+      "loss": 0.876,
+      "step": 1313
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1538159056447955e-05,
+      "loss": 0.9111,
+      "step": 1314
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.152677063398399e-05,
+      "loss": 0.8828,
+      "step": 1315
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1515380183735556e-05,
+      "loss": 0.8726,
+      "step": 1316
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1503987720830908e-05,
+      "loss": 0.8975,
+      "step": 1317
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1492593260400979e-05,
+      "loss": 0.8892,
+      "step": 1318
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1481196817579352e-05,
+      "loss": 0.2927,
+      "step": 1319
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1469798407502238e-05,
+      "loss": 0.8843,
+      "step": 1320
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1458398045308471e-05,
+      "loss": 0.8774,
+      "step": 1321
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1446995746139468e-05,
+      "loss": 0.8721,
+      "step": 1322
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1435591525139228e-05,
+      "loss": 0.8696,
+      "step": 1323
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1424185397454292e-05,
+      "loss": 0.8584,
+      "step": 1324
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1412777378233742e-05,
+      "loss": 0.3362,
+      "step": 1325
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1401367482629169e-05,
+      "loss": 0.3108,
+      "step": 1326
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1389955725794655e-05,
+      "loss": 0.877,
+      "step": 1327
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1378542122886752e-05,
+      "loss": 0.9321,
+      "step": 1328
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1367126689064473e-05,
+      "loss": 0.9009,
+      "step": 1329
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1355709439489255e-05,
+      "loss": 0.9028,
+      "step": 1330
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1344290389324949e-05,
+      "loss": 0.9092,
+      "step": 1331
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.133286955373779e-05,
+      "loss": 0.8691,
+      "step": 1332
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.13214469478964e-05,
+      "loss": 0.8721,
+      "step": 1333
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1310022586971739e-05,
+      "loss": 0.8843,
+      "step": 1334
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.12985964861371e-05,
+      "loss": 0.8838,
+      "step": 1335
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1287168660568093e-05,
+      "loss": 0.8833,
+      "step": 1336
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1275739125442618e-05,
+      "loss": 0.8799,
+      "step": 1337
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1264307895940834e-05,
+      "loss": 0.9053,
+      "step": 1338
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1252874987245163e-05,
+      "loss": 0.8804,
+      "step": 1339
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1241440414540255e-05,
+      "loss": 0.8623,
+      "step": 1340
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1230004193012964e-05,
+      "loss": 0.877,
+      "step": 1341
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1218566337852344e-05,
+      "loss": 0.8901,
+      "step": 1342
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1207126864249604e-05,
+      "loss": 0.8892,
+      "step": 1343
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1195685787398117e-05,
+      "loss": 0.3159,
+      "step": 1344
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1184243122493381e-05,
+      "loss": 0.897,
+      "step": 1345
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1172798884732998e-05,
+      "loss": 0.8511,
+      "step": 1346
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1161353089316664e-05,
+      "loss": 0.3484,
+      "step": 1347
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1149905751446143e-05,
+      "loss": 0.8618,
+      "step": 1348
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1138456886325247e-05,
+      "loss": 0.8589,
+      "step": 1349
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1127006509159814e-05,
+      "loss": 0.8843,
+      "step": 1350
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1115554635157698e-05,
+      "loss": 0.8657,
+      "step": 1351
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.110410127952873e-05,
+      "loss": 0.3108,
+      "step": 1352
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1092646457484721e-05,
+      "loss": 0.9019,
+      "step": 1353
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1081190184239418e-05,
+      "loss": 0.8774,
+      "step": 1354
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1069732475008504e-05,
+      "loss": 0.8613,
+      "step": 1355
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1058273345009565e-05,
+      "loss": 0.8755,
+      "step": 1356
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1046812809462073e-05,
+      "loss": 0.9048,
+      "step": 1357
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1035350883587372e-05,
+      "loss": 0.8423,
+      "step": 1358
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1023887582608645e-05,
+      "loss": 0.876,
+      "step": 1359
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1012422921750911e-05,
+      "loss": 0.8789,
+      "step": 1360
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1000956916240985e-05,
+      "loss": 0.897,
+      "step": 1361
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.0989489581307475e-05,
+      "loss": 0.3064,
+      "step": 1362
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.097802093218075e-05,
+      "loss": 0.896,
+      "step": 1363
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0966550984092932e-05,
+      "loss": 0.8813,
+      "step": 1364
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0955079752277859e-05,
+      "loss": 0.8818,
+      "step": 1365
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0943607251971077e-05,
+      "loss": 0.3281,
+      "step": 1366
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0932133498409821e-05,
+      "loss": 0.8833,
+      "step": 1367
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0920658506832984e-05,
+      "loss": 0.8755,
+      "step": 1368
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0909182292481108e-05,
+      "loss": 0.8794,
+      "step": 1369
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.089770487059636e-05,
+      "loss": 0.875,
+      "step": 1370
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0886226256422501e-05,
+      "loss": 0.8774,
+      "step": 1371
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0874746465204893e-05,
+      "loss": 0.8706,
+      "step": 1372
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0863265512190444e-05,
+      "loss": 0.8696,
+      "step": 1373
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0851783412627615e-05,
+      "loss": 0.9033,
+      "step": 1374
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0840300181766383e-05,
+      "loss": 0.8901,
+      "step": 1375
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0828815834858238e-05,
+      "loss": 0.8408,
+      "step": 1376
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0817330387156142e-05,
+      "loss": 0.8779,
+      "step": 1377
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0805843853914524e-05,
+      "loss": 0.8657,
+      "step": 1378
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.079435625038925e-05,
+      "loss": 0.9102,
+      "step": 1379
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0782867591837619e-05,
+      "loss": 0.8696,
+      "step": 1380
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0771377893518314e-05,
+      "loss": 0.8682,
+      "step": 1381
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0759887170691414e-05,
+      "loss": 0.8682,
+      "step": 1382
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0748395438618353e-05,
+      "loss": 0.3286,
+      "step": 1383
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0736902712561902e-05,
+      "loss": 0.8516,
+      "step": 1384
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0725409007786161e-05,
+      "loss": 0.8618,
+      "step": 1385
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0713914339556519e-05,
+      "loss": 0.8789,
+      "step": 1386
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0702418723139654e-05,
+      "loss": 0.8716,
+      "step": 1387
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0690922173803495e-05,
+      "loss": 0.8608,
+      "step": 1388
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0679424706817221e-05,
+      "loss": 0.8662,
+      "step": 1389
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0667926337451217e-05,
+      "loss": 0.299,
+      "step": 1390
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0656427080977077e-05,
+      "loss": 0.3157,
+      "step": 1391
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0644926952667568e-05,
+      "loss": 0.8765,
+      "step": 1392
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0633425967796614e-05,
+      "loss": 0.3079,
+      "step": 1393
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0621924141639279e-05,
+      "loss": 0.304,
+      "step": 1394
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0610421489471748e-05,
+      "loss": 0.8638,
+      "step": 1395
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0598918026571293e-05,
+      "loss": 0.8994,
+      "step": 1396
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0587413768216273e-05,
+      "loss": 0.8335,
+      "step": 1397
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0575908729686097e-05,
+      "loss": 0.8945,
+      "step": 1398
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0564402926261216e-05,
+      "loss": 0.8691,
+      "step": 1399
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0552896373223089e-05,
+      "loss": 0.8828,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0541389085854177e-05,
+      "loss": 0.8579,
+      "step": 1401
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0529881079437913e-05,
+      "loss": 0.8838,
+      "step": 1402
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.051837236925869e-05,
+      "loss": 0.8843,
+      "step": 1403
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0506862970601826e-05,
+      "loss": 0.8511,
+      "step": 1404
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0495352898753563e-05,
+      "loss": 0.3455,
+      "step": 1405
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0483842169001035e-05,
+      "loss": 0.8667,
+      "step": 1406
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0472330796632244e-05,
+      "loss": 0.8823,
+      "step": 1407
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0460818796936053e-05,
+      "loss": 0.8164,
+      "step": 1408
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0449306185202155e-05,
+      "loss": 0.8755,
+      "step": 1409
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0437792976721056e-05,
+      "loss": 0.8594,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.042627918678405e-05,
+      "loss": 0.8809,
+      "step": 1411
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0414764830683213e-05,
+      "loss": 0.8838,
+      "step": 1412
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0403249923711365e-05,
+      "loss": 0.8599,
+      "step": 1413
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0391734481162059e-05,
+      "loss": 0.8589,
+      "step": 1414
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0380218518329564e-05,
+      "loss": 0.9014,
+      "step": 1415
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0368702050508834e-05,
+      "loss": 0.873,
+      "step": 1416
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0357185092995499e-05,
+      "loss": 0.8608,
+      "step": 1417
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0345667661085833e-05,
+      "loss": 0.812,
+      "step": 1418
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0334149770076747e-05,
+      "loss": 0.8853,
+      "step": 1419
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0322631435265757e-05,
+      "loss": 0.8652,
+      "step": 1420
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0311112671950969e-05,
+      "loss": 0.8765,
+      "step": 1421
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0299593495431061e-05,
+      "loss": 0.8726,
+      "step": 1422
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0288073921005258e-05,
+      "loss": 0.8711,
+      "step": 1423
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.027655396397331e-05,
+      "loss": 0.3193,
+      "step": 1424
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0265033639635483e-05,
+      "loss": 0.8765,
+      "step": 1425
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0253512963292524e-05,
+      "loss": 0.3357,
+      "step": 1426
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0241991950245648e-05,
+      "loss": 0.8804,
+      "step": 1427
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0230470615796523e-05,
+      "loss": 0.3245,
+      "step": 1428
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0218948975247238e-05,
+      "loss": 0.3213,
+      "step": 1429
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0207427043900292e-05,
+      "loss": 0.8521,
+      "step": 1430
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0195904837058563e-05,
+      "loss": 0.8916,
+      "step": 1431
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0184382370025307e-05,
+      "loss": 0.896,
+      "step": 1432
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0172859658104117e-05,
+      "loss": 0.8901,
+      "step": 1433
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0161336716598915e-05,
+      "loss": 0.876,
+      "step": 1434
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0149813560813924e-05,
+      "loss": 0.3372,
+      "step": 1435
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0138290206053653e-05,
+      "loss": 0.8794,
+      "step": 1436
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0126766667622878e-05,
+      "loss": 0.2992,
+      "step": 1437
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0115242960826619e-05,
+      "loss": 0.8892,
+      "step": 1438
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0103719100970115e-05,
+      "loss": 0.8872,
+      "step": 1439
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.009219510335881e-05,
+      "loss": 0.8672,
+      "step": 1440
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0080670983298335e-05,
+      "loss": 0.8599,
+      "step": 1441
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0069146756094481e-05,
+      "loss": 0.3188,
+      "step": 1442
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0057622437053178e-05,
+      "loss": 0.853,
+      "step": 1443
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.004609804148048e-05,
+      "loss": 0.8647,
+      "step": 1444
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.003457358468255e-05,
+      "loss": 0.8623,
+      "step": 1445
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.002304908196562e-05,
+      "loss": 0.8735,
+      "step": 1446
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.001152454863599e-05,
+      "loss": 0.8804,
+      "step": 1447
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1e-05,
+      "loss": 0.8936,
+      "step": 1448
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.988475451364013e-06,
+      "loss": 0.9038,
+      "step": 1449
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.976950918034386e-06,
+      "loss": 0.8745,
+      "step": 1450
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.965426415317451e-06,
+      "loss": 0.8911,
+      "step": 1451
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.953901958519521e-06,
+      "loss": 0.8813,
+      "step": 1452
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.942377562946825e-06,
+      "loss": 0.8662,
+      "step": 1453
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.930853243905522e-06,
+      "loss": 0.3402,
+      "step": 1454
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.919329016701668e-06,
+      "loss": 0.8555,
+      "step": 1455
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.907804896641195e-06,
+      "loss": 0.8789,
+      "step": 1456
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.896280899029887e-06,
+      "loss": 0.8813,
+      "step": 1457
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.884757039173385e-06,
+      "loss": 0.8696,
+      "step": 1458
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.873233332377125e-06,
+      "loss": 0.8921,
+      "step": 1459
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.86170979394635e-06,
+      "loss": 0.8662,
+      "step": 1460
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.850186439186083e-06,
+      "loss": 0.8647,
+      "step": 1461
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.838663283401085e-06,
+      "loss": 0.8931,
+      "step": 1462
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.827140341895885e-06,
+      "loss": 0.875,
+      "step": 1463
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.815617629974696e-06,
+      "loss": 0.8555,
+      "step": 1464
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.804095162941439e-06,
+      "loss": 0.9141,
+      "step": 1465
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.792572956099715e-06,
+      "loss": 0.8604,
+      "step": 1466
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.781051024752762e-06,
+      "loss": 0.8794,
+      "step": 1467
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.769529384203479e-06,
+      "loss": 0.8589,
+      "step": 1468
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.758008049754353e-06,
+      "loss": 0.895,
+      "step": 1469
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.746487036707479e-06,
+      "loss": 0.3403,
+      "step": 1470
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.734966360364519e-06,
+      "loss": 0.8579,
+      "step": 1471
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.723446036026695e-06,
+      "loss": 0.8501,
+      "step": 1472
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.711926078994744e-06,
+      "loss": 0.916,
+      "step": 1473
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.70040650456894e-06,
+      "loss": 0.8735,
+      "step": 1474
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.688887328049034e-06,
+      "loss": 0.8574,
+      "step": 1475
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.677368564734246e-06,
+      "loss": 0.853,
+      "step": 1476
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.665850229923258e-06,
+      "loss": 0.8428,
+      "step": 1477
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.654332338914166e-06,
+      "loss": 0.8765,
+      "step": 1478
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.642814907004505e-06,
+      "loss": 0.8496,
+      "step": 1479
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.631297949491168e-06,
+      "loss": 0.8613,
+      "step": 1480
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.619781481670437e-06,
+      "loss": 0.8594,
+      "step": 1481
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.608265518837946e-06,
+      "loss": 0.9229,
+      "step": 1482
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.596750076288642e-06,
+      "loss": 0.8823,
+      "step": 1483
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.58523516931679e-06,
+      "loss": 0.8721,
+      "step": 1484
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.573720813215954e-06,
+      "loss": 0.8857,
+      "step": 1485
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.562207023278948e-06,
+      "loss": 0.8726,
+      "step": 1486
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.550693814797847e-06,
+      "loss": 0.8765,
+      "step": 1487
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.539181203063952e-06,
+      "loss": 0.8706,
+      "step": 1488
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.527669203367756e-06,
+      "loss": 0.3413,
+      "step": 1489
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.516157830998967e-06,
+      "loss": 0.8721,
+      "step": 1490
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.504647101246438e-06,
+      "loss": 0.8857,
+      "step": 1491
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.493137029398176e-06,
+      "loss": 0.8599,
+      "step": 1492
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.481627630741315e-06,
+      "loss": 0.874,
+      "step": 1493
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.470118920562087e-06,
+      "loss": 0.8604,
+      "step": 1494
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.458610914145826e-06,
+      "loss": 0.8623,
+      "step": 1495
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.447103626776913e-06,
+      "loss": 0.8506,
+      "step": 1496
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.435597073738787e-06,
+      "loss": 0.8833,
+      "step": 1497
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.424091270313904e-06,
+      "loss": 0.8208,
+      "step": 1498
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.41258623178373e-06,
+      "loss": 0.3242,
+      "step": 1499
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.40108197342871e-06,
+      "loss": 0.8877,
+      "step": 1500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.389578510528256e-06,
+      "loss": 0.3186,
+      "step": 1501
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.378075858360723e-06,
+      "loss": 0.8799,
+      "step": 1502
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.36657403220339e-06,
+      "loss": 0.8638,
+      "step": 1503
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.355073047332437e-06,
+      "loss": 0.8818,
+      "step": 1504
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.343572919022924e-06,
+      "loss": 0.8789,
+      "step": 1505
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.332073662548785e-06,
+      "loss": 0.8926,
+      "step": 1506
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.320575293182782e-06,
+      "loss": 0.8408,
+      "step": 1507
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.309077826196507e-06,
+      "loss": 0.8657,
+      "step": 1508
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.297581276860353e-06,
+      "loss": 0.8638,
+      "step": 1509
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.286085660443486e-06,
+      "loss": 0.8315,
+      "step": 1510
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.274590992213844e-06,
+      "loss": 0.8867,
+      "step": 1511
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.2630972874381e-06,
+      "loss": 0.8516,
+      "step": 1512
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.25160456138165e-06,
+      "loss": 0.8945,
+      "step": 1513
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.240112829308589e-06,
+      "loss": 0.8892,
+      "step": 1514
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.228622106481691e-06,
+      "loss": 0.876,
+      "step": 1515
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.217132408162383e-06,
+      "loss": 0.8564,
+      "step": 1516
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.205643749610751e-06,
+      "loss": 0.8691,
+      "step": 1517
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.19415614608548e-06,
+      "loss": 0.8711,
+      "step": 1518
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.182669612843861e-06,
+      "loss": 0.8569,
+      "step": 1519
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.171184165141766e-06,
+      "loss": 0.874,
+      "step": 1520
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.15969981823362e-06,
+      "loss": 0.8813,
+      "step": 1521
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.148216587372389e-06,
+      "loss": 0.322,
+      "step": 1522
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.136734487809559e-06,
+      "loss": 0.8809,
+      "step": 1523
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.12525353479511e-06,
+      "loss": 0.8501,
+      "step": 1524
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.113773743577502e-06,
+      "loss": 0.8711,
+      "step": 1525
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.102295129403646e-06,
+      "loss": 0.3259,
+      "step": 1526
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.090817707518893e-06,
+      "loss": 0.8359,
+      "step": 1527
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.079341493167019e-06,
+      "loss": 0.8721,
+      "step": 1528
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.067866501590182e-06,
+      "loss": 0.8833,
+      "step": 1529
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.056392748028926e-06,
+      "loss": 0.8789,
+      "step": 1530
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.044920247722146e-06,
+      "loss": 0.8623,
+      "step": 1531
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.033449015907068e-06,
+      "loss": 0.3152,
+      "step": 1532
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.021979067819252e-06,
+      "loss": 0.8657,
+      "step": 1533
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.010510418692528e-06,
+      "loss": 0.9019,
+      "step": 1534
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.999043083759016e-06,
+      "loss": 0.876,
+      "step": 1535
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.987577078249094e-06,
+      "loss": 0.8789,
+      "step": 1536
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.976112417391358e-06,
+      "loss": 0.8721,
+      "step": 1537
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.964649116412632e-06,
+      "loss": 0.8789,
+      "step": 1538
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.953187190537929e-06,
+      "loss": 0.8394,
+      "step": 1539
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.941726654990438e-06,
+      "loss": 0.3157,
+      "step": 1540
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.9302675249915e-06,
+      "loss": 0.8545,
+      "step": 1541
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.918809815760585e-06,
+      "loss": 0.8677,
+      "step": 1542
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.90735354251528e-06,
+      "loss": 0.877,
+      "step": 1543
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.895898720471272e-06,
+      "loss": 0.855,
+      "step": 1544
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.884445364842304e-06,
+      "loss": 0.8491,
+      "step": 1545
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.872993490840188e-06,
+      "loss": 0.8887,
+      "step": 1546
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.861543113674758e-06,
+      "loss": 0.8975,
+      "step": 1547
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.850094248553862e-06,
+      "loss": 0.8745,
+      "step": 1548
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.838646910683338e-06,
+      "loss": 0.8604,
+      "step": 1549
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.827201115267005e-06,
+      "loss": 0.3096,
+      "step": 1550
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.815756877506622e-06,
+      "loss": 0.9077,
+      "step": 1551
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.804314212601886e-06,
+      "loss": 0.8472,
+      "step": 1552
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.7928731357504e-06,
+      "loss": 0.8735,
+      "step": 1553
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.78143366214766e-06,
+      "loss": 0.8833,
+      "step": 1554
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.769995806987037e-06,
+      "loss": 0.8652,
+      "step": 1555
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.758559585459746e-06,
+      "loss": 0.9058,
+      "step": 1556
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.747125012754839e-06,
+      "loss": 0.8872,
+      "step": 1557
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.735692104059171e-06,
+      "loss": 0.3145,
+      "step": 1558
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.724260874557384e-06,
+      "loss": 0.9097,
+      "step": 1559
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.712831339431908e-06,
+      "loss": 0.3381,
+      "step": 1560
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.701403513862901e-06,
+      "loss": 0.8896,
+      "step": 1561
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.689977413028266e-06,
+      "loss": 0.8774,
+      "step": 1562
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.678553052103605e-06,
+      "loss": 0.3163,
+      "step": 1563
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.667130446262214e-06,
+      "loss": 0.8838,
+      "step": 1564
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.655709610675056e-06,
+      "loss": 0.8945,
+      "step": 1565
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.644290560510748e-06,
+      "loss": 0.8691,
+      "step": 1566
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.632873310935528e-06,
+      "loss": 0.8638,
+      "step": 1567
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.621457877113252e-06,
+      "loss": 0.8682,
+      "step": 1568
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.610044274205352e-06,
+      "loss": 0.8672,
+      "step": 1569
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.598632517370835e-06,
+      "loss": 0.8525,
+      "step": 1570
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.58722262176626e-06,
+      "loss": 0.8765,
+      "step": 1571
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.57581460254571e-06,
+      "loss": 0.8774,
+      "step": 1572
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.564408474860774e-06,
+      "loss": 0.874,
+      "step": 1573
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.553004253860535e-06,
+      "loss": 0.8921,
+      "step": 1574
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.541601954691534e-06,
+      "loss": 0.8511,
+      "step": 1575
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.530201592497764e-06,
+      "loss": 0.8101,
+      "step": 1576
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.518803182420651e-06,
+      "loss": 0.8901,
+      "step": 1577
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.507406739599023e-06,
+      "loss": 0.814,
+      "step": 1578
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.496012279169097e-06,
+      "loss": 0.8789,
+      "step": 1579
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.48461981626445e-06,
+      "loss": 0.8618,
+      "step": 1580
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.473229366016014e-06,
+      "loss": 0.895,
+      "step": 1581
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.461840943552049e-06,
+      "loss": 0.3264,
+      "step": 1582
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.450454563998117e-06,
+      "loss": 0.8506,
+      "step": 1583
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.439070242477067e-06,
+      "loss": 0.8394,
+      "step": 1584
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.427687994109017e-06,
+      "loss": 0.8657,
+      "step": 1585
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.416307834011329e-06,
+      "loss": 0.8633,
+      "step": 1586
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.404929777298592e-06,
+      "loss": 0.8823,
+      "step": 1587
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.393553839082608e-06,
+      "loss": 0.8584,
+      "step": 1588
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.382180034472353e-06,
+      "loss": 0.8545,
+      "step": 1589
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.370808378573983e-06,
+      "loss": 0.3069,
+      "step": 1590
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.359438886490783e-06,
+      "loss": 0.8555,
+      "step": 1591
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.34807157332318e-06,
+      "loss": 0.8755,
+      "step": 1592
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.336706454168701e-06,
+      "loss": 0.894,
+      "step": 1593
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.325343544121958e-06,
+      "loss": 0.8857,
+      "step": 1594
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.313982858274634e-06,
+      "loss": 0.895,
+      "step": 1595
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.30262441171545e-06,
+      "loss": 0.8506,
+      "step": 1596
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.291268219530153e-06,
+      "loss": 0.897,
+      "step": 1597
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.279914296801509e-06,
+      "loss": 0.8662,
+      "step": 1598
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.268562658609254e-06,
+      "loss": 0.8555,
+      "step": 1599
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.2572133200301e-06,
+      "loss": 0.8374,
+      "step": 1600
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.245866296137701e-06,
+      "loss": 0.8174,
+      "step": 1601
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.234521602002634e-06,
+      "loss": 0.8726,
+      "step": 1602
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.223179252692385e-06,
+      "loss": 0.8589,
+      "step": 1603
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.211839263271328e-06,
+      "loss": 0.8442,
+      "step": 1604
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.200501648800698e-06,
+      "loss": 0.8608,
+      "step": 1605
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.189166424338582e-06,
+      "loss": 0.8877,
+      "step": 1606
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.17783360493988e-06,
+      "loss": 0.8877,
+      "step": 1607
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.166503205656312e-06,
+      "loss": 0.8716,
+      "step": 1608
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.155175241536377e-06,
+      "loss": 0.8853,
+      "step": 1609
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.143849727625341e-06,
+      "loss": 0.8511,
+      "step": 1610
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.132526678965215e-06,
+      "loss": 0.8716,
+      "step": 1611
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.121206110594736e-06,
+      "loss": 0.8345,
+      "step": 1612
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.109888037549346e-06,
+      "loss": 0.8711,
+      "step": 1613
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.098572474861174e-06,
+      "loss": 0.8467,
+      "step": 1614
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.087259437559017e-06,
+      "loss": 0.8994,
+      "step": 1615
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.075948940668313e-06,
+      "loss": 0.8711,
+      "step": 1616
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.06464099921113e-06,
+      "loss": 0.8374,
+      "step": 1617
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.05333562820614e-06,
+      "loss": 0.8428,
+      "step": 1618
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.042032842668598e-06,
+      "loss": 0.8374,
+      "step": 1619
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.030732657610332e-06,
+      "loss": 0.8457,
+      "step": 1620
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.019435088039714e-06,
+      "loss": 0.8594,
+      "step": 1621
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.008140148961642e-06,
+      "loss": 0.8823,
+      "step": 1622
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.996847855377514e-06,
+      "loss": 0.3025,
+      "step": 1623
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.985558222285222e-06,
+      "loss": 0.8691,
+      "step": 1624
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.974271264679122e-06,
+      "loss": 0.8667,
+      "step": 1625
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.962986997550019e-06,
+      "loss": 0.8325,
+      "step": 1626
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.951705435885143e-06,
+      "loss": 0.8755,
+      "step": 1627
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.940426594668124e-06,
+      "loss": 0.8833,
+      "step": 1628
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.929150488878991e-06,
+      "loss": 0.8628,
+      "step": 1629
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.917877133494129e-06,
+      "loss": 0.8584,
+      "step": 1630
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.906606543486278e-06,
+      "loss": 0.8335,
+      "step": 1631
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.895338733824503e-06,
+      "loss": 0.9028,
+      "step": 1632
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.884073719474174e-06,
+      "loss": 0.8286,
+      "step": 1633
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.872811515396948e-06,
+      "loss": 0.8691,
+      "step": 1634
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.86155213655075e-06,
+      "loss": 0.8447,
+      "step": 1635
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.85029559788976e-06,
+      "loss": 0.3008,
+      "step": 1636
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.839041914364375e-06,
+      "loss": 0.8838,
+      "step": 1637
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.827791100921208e-06,
+      "loss": 0.322,
+      "step": 1638
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.816543172503052e-06,
+      "loss": 0.8604,
+      "step": 1639
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.805298144048877e-06,
+      "loss": 0.8848,
+      "step": 1640
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.794056030493793e-06,
+      "loss": 0.8662,
+      "step": 1641
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.782816846769047e-06,
+      "loss": 0.3115,
+      "step": 1642
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.771580607801994e-06,
+      "loss": 0.8735,
+      "step": 1643
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.760347328516066e-06,
+      "loss": 0.8794,
+      "step": 1644
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.749117023830779e-06,
+      "loss": 0.8511,
+      "step": 1645
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.737889708661685e-06,
+      "loss": 0.8706,
+      "step": 1646
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.72666539792038e-06,
+      "loss": 0.8896,
+      "step": 1647
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.715444106514462e-06,
+      "loss": 0.8818,
+      "step": 1648
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.704225849347517e-06,
+      "loss": 0.874,
+      "step": 1649
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.693010641319098e-06,
+      "loss": 0.8613,
+      "step": 1650
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.681798497324717e-06,
+      "loss": 0.8623,
+      "step": 1651
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.670589432255816e-06,
+      "loss": 0.8633,
+      "step": 1652
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.659383460999742e-06,
+      "loss": 0.9028,
+      "step": 1653
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.648180598439737e-06,
+      "loss": 0.8521,
+      "step": 1654
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.63698085945491e-06,
+      "loss": 0.8657,
+      "step": 1655
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.6257842589202236e-06,
+      "loss": 0.8804,
+      "step": 1656
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.614590811706473e-06,
+      "loss": 0.8643,
+      "step": 1657
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.603400532680266e-06,
+      "loss": 0.8618,
+      "step": 1658
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.592213436704004e-06,
+      "loss": 0.3346,
+      "step": 1659
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.581029538635856e-06,
+      "loss": 0.8525,
+      "step": 1660
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.569848853329742e-06,
+      "loss": 0.8555,
+      "step": 1661
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.558671395635322e-06,
+      "loss": 0.8701,
+      "step": 1662
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.547497180397968e-06,
+      "loss": 0.8755,
+      "step": 1663
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.536326222458741e-06,
+      "loss": 0.8965,
+      "step": 1664
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.525158536654382e-06,
+      "loss": 0.8901,
+      "step": 1665
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.513994137817277e-06,
+      "loss": 0.8545,
+      "step": 1666
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.502833040775457e-06,
+      "loss": 0.856,
+      "step": 1667
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.4916752603525565e-06,
+      "loss": 0.332,
+      "step": 1668
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.480520811367817e-06,
+      "loss": 0.874,
+      "step": 1669
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.469369708636052e-06,
+      "loss": 0.8838,
+      "step": 1670
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.45822196696762e-06,
+      "loss": 0.8438,
+      "step": 1671
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.447077601168426e-06,
+      "loss": 0.8364,
+      "step": 1672
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.435936626039891e-06,
+      "loss": 0.8613,
+      "step": 1673
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.424799056378932e-06,
+      "loss": 0.8638,
+      "step": 1674
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.41366490697794e-06,
+      "loss": 0.8936,
+      "step": 1675
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.402534192624767e-06,
+      "loss": 0.8687,
+      "step": 1676
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.391406928102695e-06,
+      "loss": 0.8442,
+      "step": 1677
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.380283128190433e-06,
+      "loss": 0.8569,
+      "step": 1678
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.369162807662087e-06,
+      "loss": 0.8711,
+      "step": 1679
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.358045981287141e-06,
+      "loss": 0.8511,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.3469326638304365e-06,
+      "loss": 0.8462,
+      "step": 1681
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.335822870052153e-06,
+      "loss": 0.8687,
+      "step": 1682
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.324716614707794e-06,
+      "loss": 0.8354,
+      "step": 1683
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.313613912548161e-06,
+      "loss": 0.8667,
+      "step": 1684
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.302514778319341e-06,
+      "loss": 0.8428,
+      "step": 1685
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.291419226762677e-06,
+      "loss": 0.8486,
+      "step": 1686
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.280327272614753e-06,
+      "loss": 0.8604,
+      "step": 1687
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.26923893060738e-06,
+      "loss": 0.8521,
+      "step": 1688
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.2581542154675654e-06,
+      "loss": 0.8506,
+      "step": 1689
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.247073141917506e-06,
+      "loss": 0.873,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.23599572467456e-06,
+      "loss": 0.3445,
+      "step": 1691
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.224921978451229e-06,
+      "loss": 0.8555,
+      "step": 1692
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.2138519179551335e-06,
+      "loss": 0.8701,
+      "step": 1693
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.202785557889007e-06,
+      "loss": 0.853,
+      "step": 1694
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.1917229129506626e-06,
+      "loss": 0.834,
+      "step": 1695
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.180663997832984e-06,
+      "loss": 0.8804,
+      "step": 1696
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.169608827223902e-06,
+      "loss": 0.8486,
+      "step": 1697
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.158557415806364e-06,
+      "loss": 0.8701,
+      "step": 1698
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.147509778258334e-06,
+      "loss": 0.8872,
+      "step": 1699
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.13646592925276e-06,
+      "loss": 0.8887,
+      "step": 1700
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.125425883457564e-06,
+      "loss": 0.8682,
+      "step": 1701
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.114389655535607e-06,
+      "loss": 0.8779,
+      "step": 1702
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.10335726014469e-06,
+      "loss": 0.8696,
+      "step": 1703
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.0923287119375125e-06,
+      "loss": 0.8521,
+      "step": 1704
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.081304025561668e-06,
+      "loss": 0.8394,
+      "step": 1705
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.070283215659632e-06,
+      "loss": 0.8452,
+      "step": 1706
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.059266296868715e-06,
+      "loss": 0.8506,
+      "step": 1707
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.048253283821074e-06,
+      "loss": 0.853,
+      "step": 1708
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.037244191143662e-06,
+      "loss": 0.833,
+      "step": 1709
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.026239033458242e-06,
+      "loss": 0.374,
+      "step": 1710
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.015237825381338e-06,
+      "loss": 0.8467,
+      "step": 1711
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.0042405815242375e-06,
+      "loss": 0.8701,
+      "step": 1712
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.993247316492962e-06,
+      "loss": 0.855,
+      "step": 1713
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.982258044888239e-06,
+      "loss": 0.8506,
+      "step": 1714
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.971272781305503e-06,
+      "loss": 0.8877,
+      "step": 1715
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.960291540334856e-06,
+      "loss": 0.8623,
+      "step": 1716
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.949314336561068e-06,
+      "loss": 0.835,
+      "step": 1717
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.93834118456354e-06,
+      "loss": 0.8467,
+      "step": 1718
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.927372098916294e-06,
+      "loss": 0.8486,
+      "step": 1719
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.9164070941879465e-06,
+      "loss": 0.8867,
+      "step": 1720
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.9054461849417e-06,
+      "loss": 0.856,
+      "step": 1721
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.894489385735312e-06,
+      "loss": 0.8535,
+      "step": 1722
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.88353671112109e-06,
+      "loss": 0.8564,
+      "step": 1723
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.872588175645857e-06,
+      "loss": 0.8438,
+      "step": 1724
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.861643793850934e-06,
+      "loss": 0.8799,
+      "step": 1725
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.850703580272136e-06,
+      "loss": 0.854,
+      "step": 1726
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.839767549439733e-06,
+      "loss": 0.8584,
+      "step": 1727
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.828835715878446e-06,
+      "loss": 0.8491,
+      "step": 1728
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.81790809410742e-06,
+      "loss": 0.3262,
+      "step": 1729
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.806984698640202e-06,
+      "loss": 0.8403,
+      "step": 1730
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.796065543984729e-06,
+      "loss": 0.8696,
+      "step": 1731
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.7851506446433045e-06,
+      "loss": 0.8452,
+      "step": 1732
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.774240015112583e-06,
+      "loss": 0.3364,
+      "step": 1733
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.763333669883547e-06,
+      "loss": 0.8174,
+      "step": 1734
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.752431623441488e-06,
+      "loss": 0.8691,
+      "step": 1735
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.741533890265983e-06,
+      "loss": 0.853,
+      "step": 1736
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.73064048483089e-06,
+      "loss": 0.8384,
+      "step": 1737
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.719751421604309e-06,
+      "loss": 0.8594,
+      "step": 1738
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.708866715048586e-06,
+      "loss": 0.8496,
+      "step": 1739
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.6979863796202686e-06,
+      "loss": 0.835,
+      "step": 1740
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.687110429770103e-06,
+      "loss": 0.8589,
+      "step": 1741
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.67623887994301e-06,
+      "loss": 0.2932,
+      "step": 1742
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.6653717445780675e-06,
+      "loss": 0.8633,
+      "step": 1743
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.6545090381084934e-06,
+      "loss": 0.8711,
+      "step": 1744
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.6436507749616195e-06,
+      "loss": 0.3335,
+      "step": 1745
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.632796969558874e-06,
+      "loss": 0.8223,
+      "step": 1746
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.6219476363157685e-06,
+      "loss": 0.8184,
+      "step": 1747
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.611102789641873e-06,
+      "loss": 0.877,
+      "step": 1748
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.600262443940799e-06,
+      "loss": 0.8569,
+      "step": 1749
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.589426613610184e-06,
+      "loss": 0.876,
+      "step": 1750
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.578595313041665e-06,
+      "loss": 0.8428,
+      "step": 1751
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.567768556620857e-06,
+      "loss": 0.8721,
+      "step": 1752
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.556946358727349e-06,
+      "loss": 0.8569,
+      "step": 1753
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.546128733734669e-06,
+      "loss": 0.811,
+      "step": 1754
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.535315696010278e-06,
+      "loss": 0.8696,
+      "step": 1755
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.524507259915542e-06,
+      "loss": 0.8706,
+      "step": 1756
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.51370343980571e-06,
+      "loss": 0.8467,
+      "step": 1757
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.502904250029905e-06,
+      "loss": 0.9087,
+      "step": 1758
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.492109704931101e-06,
+      "loss": 0.8599,
+      "step": 1759
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.481319818846099e-06,
+      "loss": 0.832,
+      "step": 1760
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.470534606105519e-06,
+      "loss": 0.8447,
+      "step": 1761
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.459754081033769e-06,
+      "loss": 0.8389,
+      "step": 1762
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.448978257949028e-06,
+      "loss": 0.8647,
+      "step": 1763
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.438207151163238e-06,
+      "loss": 0.8525,
+      "step": 1764
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.42744077498207e-06,
+      "loss": 0.8467,
+      "step": 1765
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.416679143704918e-06,
+      "loss": 0.8574,
+      "step": 1766
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.405922271624874e-06,
+      "loss": 0.8652,
+      "step": 1767
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.395170173028699e-06,
+      "loss": 0.3452,
+      "step": 1768
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.384422862196824e-06,
+      "loss": 0.8496,
+      "step": 1769
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.373680353403318e-06,
+      "loss": 0.8447,
+      "step": 1770
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.362942660915875e-06,
+      "loss": 0.8672,
+      "step": 1771
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.35220979899579e-06,
+      "loss": 0.8584,
+      "step": 1772
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.341481781897939e-06,
+      "loss": 0.8486,
+      "step": 1773
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.330758623870768e-06,
+      "loss": 0.8662,
+      "step": 1774
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.320040339156267e-06,
+      "loss": 0.8184,
+      "step": 1775
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.309326941989953e-06,
+      "loss": 0.8335,
+      "step": 1776
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.298618446600856e-06,
+      "loss": 0.8765,
+      "step": 1777
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.287914867211494e-06,
+      "loss": 0.8159,
+      "step": 1778
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.27721621803785e-06,
+      "loss": 0.8447,
+      "step": 1779
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.266522513289363e-06,
+      "loss": 0.8594,
+      "step": 1780
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.255833767168907e-06,
+      "loss": 0.8574,
+      "step": 1781
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.24514999387277e-06,
+      "loss": 0.8638,
+      "step": 1782
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.234471207590636e-06,
+      "loss": 0.8916,
+      "step": 1783
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.2237974225055565e-06,
+      "loss": 0.8813,
+      "step": 1784
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.213128652793952e-06,
+      "loss": 0.8184,
+      "step": 1785
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.202464912625576e-06,
+      "loss": 0.2932,
+      "step": 1786
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.1918062161635005e-06,
+      "loss": 0.8486,
+      "step": 1787
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.181152577564109e-06,
+      "loss": 0.8828,
+      "step": 1788
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.170504010977053e-06,
+      "loss": 0.3525,
+      "step": 1789
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.159860530545258e-06,
+      "loss": 0.8486,
+      "step": 1790
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.149222150404889e-06,
+      "loss": 0.8296,
+      "step": 1791
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.138588884685337e-06,
+      "loss": 0.8438,
+      "step": 1792
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.127960747509207e-06,
+      "loss": 0.8452,
+      "step": 1793
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.117337752992288e-06,
+      "loss": 0.8599,
+      "step": 1794
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.106719915243533e-06,
+      "loss": 0.855,
+      "step": 1795
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.0961072483650526e-06,
+      "loss": 0.8418,
+      "step": 1796
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.08549976645209e-06,
+      "loss": 0.3049,
+      "step": 1797
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.0748974835930005e-06,
+      "loss": 0.8716,
+      "step": 1798
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.064300413869237e-06,
+      "loss": 0.8794,
+      "step": 1799
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.053708571355321e-06,
+      "loss": 0.8574,
+      "step": 1800
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.043121970118837e-06,
+      "loss": 0.8481,
+      "step": 1801
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.032540624220408e-06,
+      "loss": 0.8486,
+      "step": 1802
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.0219645477136764e-06,
+      "loss": 0.3186,
+      "step": 1803
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.01139375464529e-06,
+      "loss": 0.8433,
+      "step": 1804
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.000828259054872e-06,
+      "loss": 0.8521,
+      "step": 1805
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.990268074975012e-06,
+      "loss": 0.8101,
+      "step": 1806
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.97971321643125e-06,
+      "loss": 0.2965,
+      "step": 1807
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.9691636974420465e-06,
+      "loss": 0.855,
+      "step": 1808
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.958619532018775e-06,
+      "loss": 0.8545,
+      "step": 1809
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.9480807341657e-06,
+      "loss": 0.2793,
+      "step": 1810
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.937547317879946e-06,
+      "loss": 0.8843,
+      "step": 1811
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.927019297151504e-06,
+      "loss": 0.8623,
+      "step": 1812
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.916496685963191e-06,
+      "loss": 0.8447,
+      "step": 1813
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.905979498290637e-06,
+      "loss": 0.8135,
+      "step": 1814
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.89546774810228e-06,
+      "loss": 0.853,
+      "step": 1815
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.8849614493593256e-06,
+      "loss": 0.8384,
+      "step": 1816
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.87446061601574e-06,
+      "loss": 0.8403,
+      "step": 1817
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.8639652620182354e-06,
+      "loss": 0.8286,
+      "step": 1818
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.853475401306241e-06,
+      "loss": 0.2705,
+      "step": 1819
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.842991047811899e-06,
+      "loss": 0.8823,
+      "step": 1820
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.83251221546003e-06,
+      "loss": 0.875,
+      "step": 1821
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.82203891816812e-06,
+      "loss": 0.8345,
+      "step": 1822
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.811571169846304e-06,
+      "loss": 0.8716,
+      "step": 1823
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.801108984397355e-06,
+      "loss": 0.8237,
+      "step": 1824
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.790652375716653e-06,
+      "loss": 0.8188,
+      "step": 1825
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.780201357692163e-06,
+      "loss": 0.7905,
+      "step": 1826
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.769755944204443e-06,
+      "loss": 0.8691,
+      "step": 1827
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.75931614912659e-06,
+      "loss": 0.3455,
+      "step": 1828
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.748881986324245e-06,
+      "loss": 0.8545,
+      "step": 1829
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.7384534696555695e-06,
+      "loss": 0.8745,
+      "step": 1830
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.728030612971231e-06,
+      "loss": 0.8491,
+      "step": 1831
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.717613430114366e-06,
+      "loss": 0.8569,
+      "step": 1832
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.7072019349205925e-06,
+      "loss": 0.8784,
+      "step": 1833
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.69679614121796e-06,
+      "loss": 0.8774,
+      "step": 1834
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.686396062826946e-06,
+      "loss": 0.8477,
+      "step": 1835
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.6760017135604505e-06,
+      "loss": 0.8193,
+      "step": 1836
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.665613107223755e-06,
+      "loss": 0.8564,
+      "step": 1837
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.655230257614514e-06,
+      "loss": 0.8877,
+      "step": 1838
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.644853178522734e-06,
+      "loss": 0.8828,
+      "step": 1839
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.634481883730762e-06,
+      "loss": 0.8589,
+      "step": 1840
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.624116387013259e-06,
+      "loss": 0.8574,
+      "step": 1841
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.613756702137188e-06,
+      "loss": 0.8496,
+      "step": 1842
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.603402842861797e-06,
+      "loss": 0.8677,
+      "step": 1843
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.5930548229385876e-06,
+      "loss": 0.8657,
+      "step": 1844
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.5827126561113045e-06,
+      "loss": 0.8281,
+      "step": 1845
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.57237635611593e-06,
+      "loss": 0.8408,
+      "step": 1846
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.562045936680649e-06,
+      "loss": 0.8735,
+      "step": 1847
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.5517214115258325e-06,
+      "loss": 0.8169,
+      "step": 1848
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.54140279436403e-06,
+      "loss": 0.8745,
+      "step": 1849
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.531090098899938e-06,
+      "loss": 0.856,
+      "step": 1850
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.520783338830386e-06,
+      "loss": 0.8467,
+      "step": 1851
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.51048252784433e-06,
+      "loss": 0.8384,
+      "step": 1852
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.500187679622819e-06,
+      "loss": 0.8516,
+      "step": 1853
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.48989880783898e-06,
+      "loss": 0.854,
+      "step": 1854
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.479615926158013e-06,
+      "loss": 0.3126,
+      "step": 1855
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.469339048237148e-06,
+      "loss": 0.8496,
+      "step": 1856
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.459068187725644e-06,
+      "loss": 0.8657,
+      "step": 1857
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.448803358264776e-06,
+      "loss": 0.8408,
+      "step": 1858
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.438544573487811e-06,
+      "loss": 0.8389,
+      "step": 1859
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.428291847019973e-06,
+      "loss": 0.3077,
+      "step": 1860
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.4180451924784475e-06,
+      "loss": 0.8472,
+      "step": 1861
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.407804623472357e-06,
+      "loss": 0.8428,
+      "step": 1862
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.397570153602747e-06,
+      "loss": 0.8745,
+      "step": 1863
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.3873417964625465e-06,
+      "loss": 0.8735,
+      "step": 1864
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.377119565636584e-06,
+      "loss": 0.8354,
+      "step": 1865
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.366903474701537e-06,
+      "loss": 0.8589,
+      "step": 1866
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.35669353722593e-06,
+      "loss": 0.8179,
+      "step": 1867
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.346489766770124e-06,
+      "loss": 0.8433,
+      "step": 1868
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.336292176886287e-06,
+      "loss": 0.8521,
+      "step": 1869
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.326100781118366e-06,
+      "loss": 0.8633,
+      "step": 1870
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.3159155930021e-06,
+      "loss": 0.8638,
+      "step": 1871
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.305736626064969e-06,
+      "loss": 0.8472,
+      "step": 1872
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.295563893826191e-06,
+      "loss": 0.7876,
+      "step": 1873
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.285397409796711e-06,
+      "loss": 0.7993,
+      "step": 1874
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.275237187479176e-06,
+      "loss": 0.3403,
+      "step": 1875
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.265083240367909e-06,
+      "loss": 0.854,
+      "step": 1876
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.254935581948897e-06,
+      "loss": 0.8545,
+      "step": 1877
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.244794225699782e-06,
+      "loss": 0.8486,
+      "step": 1878
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.23465918508984e-06,
+      "loss": 0.8311,
+      "step": 1879
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.224530473579943e-06,
+      "loss": 0.8589,
+      "step": 1880
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.214408104622573e-06,
+      "loss": 0.8594,
+      "step": 1881
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.204292091661778e-06,
+      "loss": 0.8613,
+      "step": 1882
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.194182448133163e-06,
+      "loss": 0.8511,
+      "step": 1883
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.184079187463884e-06,
+      "loss": 0.2994,
+      "step": 1884
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.173982323072615e-06,
+      "loss": 0.8862,
+      "step": 1885
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.163891868369526e-06,
+      "loss": 0.3151,
+      "step": 1886
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.153807836756288e-06,
+      "loss": 0.8574,
+      "step": 1887
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.143730241626034e-06,
+      "loss": 0.8599,
+      "step": 1888
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.133659096363341e-06,
+      "loss": 0.8589,
+      "step": 1889
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.123594414344233e-06,
+      "loss": 0.8599,
+      "step": 1890
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.113536208936147e-06,
+      "loss": 0.8384,
+      "step": 1891
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.103484493497913e-06,
+      "loss": 0.855,
+      "step": 1892
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.093439281379738e-06,
+      "loss": 0.8354,
+      "step": 1893
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.083400585923206e-06,
+      "loss": 0.8325,
+      "step": 1894
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.073368420461229e-06,
+      "loss": 0.8345,
+      "step": 1895
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.063342798318057e-06,
+      "loss": 0.8843,
+      "step": 1896
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.053323732809252e-06,
+      "loss": 0.8164,
+      "step": 1897
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.043311237241657e-06,
+      "loss": 0.8267,
+      "step": 1898
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.033305324913392e-06,
+      "loss": 0.8735,
+      "step": 1899
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.0233060091138395e-06,
+      "loss": 0.8408,
+      "step": 1900
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.0133133031236215e-06,
+      "loss": 0.8633,
+      "step": 1901
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.003327220214569e-06,
+      "loss": 0.3289,
+      "step": 1902
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.993347773649732e-06,
+      "loss": 0.8345,
+      "step": 1903
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.983374976683335e-06,
+      "loss": 0.8408,
+      "step": 1904
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.973408842560772e-06,
+      "loss": 0.8447,
+      "step": 1905
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.963449384518593e-06,
+      "loss": 0.856,
+      "step": 1906
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.953496615784482e-06,
+      "loss": 0.8398,
+      "step": 1907
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.943550549577233e-06,
+      "loss": 0.8457,
+      "step": 1908
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.933611199106736e-06,
+      "loss": 0.855,
+      "step": 1909
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.923678577573973e-06,
+      "loss": 0.8467,
+      "step": 1910
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.913752698170972e-06,
+      "loss": 0.8535,
+      "step": 1911
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.903833574080825e-06,
+      "loss": 0.8452,
+      "step": 1912
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.893921218477642e-06,
+      "loss": 0.8315,
+      "step": 1913
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.884015644526543e-06,
+      "loss": 0.304,
+      "step": 1914
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.874116865383638e-06,
+      "loss": 0.8687,
+      "step": 1915
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.86422489419602e-06,
+      "loss": 0.811,
+      "step": 1916
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.85433974410174e-06,
+      "loss": 0.8247,
+      "step": 1917
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.844461428229782e-06,
+      "loss": 0.8438,
+      "step": 1918
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.834589959700061e-06,
+      "loss": 0.3276,
+      "step": 1919
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8247253516233924e-06,
+      "loss": 0.8364,
+      "step": 1920
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.814867617101479e-06,
+      "loss": 0.3193,
+      "step": 1921
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.805016769226898e-06,
+      "loss": 0.8643,
+      "step": 1922
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.795172821083084e-06,
+      "loss": 0.8223,
+      "step": 1923
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.785335785744301e-06,
+      "loss": 0.8569,
+      "step": 1924
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.7755056762756255e-06,
+      "loss": 0.332,
+      "step": 1925
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.765682505732956e-06,
+      "loss": 0.3081,
+      "step": 1926
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.755866287162952e-06,
+      "loss": 0.8467,
+      "step": 1927
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.746057033603054e-06,
+      "loss": 0.8545,
+      "step": 1928
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.736254758081454e-06,
+      "loss": 0.8291,
+      "step": 1929
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.726459473617065e-06,
+      "loss": 0.3162,
+      "step": 1930
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.7166711932195155e-06,
+      "loss": 0.874,
+      "step": 1931
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.706889929889145e-06,
+      "loss": 0.8623,
+      "step": 1932
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.697115696616955e-06,
+      "loss": 0.8667,
+      "step": 1933
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.687348506384626e-06,
+      "loss": 0.8408,
+      "step": 1934
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.677588372164479e-06,
+      "loss": 0.8643,
+      "step": 1935
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.667835306919459e-06,
+      "loss": 0.8408,
+      "step": 1936
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.658089323603123e-06,
+      "loss": 0.8384,
+      "step": 1937
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.648350435159627e-06,
+      "loss": 0.8594,
+      "step": 1938
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.638618654523705e-06,
+      "loss": 0.8418,
+      "step": 1939
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.628893994620645e-06,
+      "loss": 0.8325,
+      "step": 1940
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.619176468366274e-06,
+      "loss": 0.8579,
+      "step": 1941
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.609466088666959e-06,
+      "loss": 0.8413,
+      "step": 1942
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.599762868419561e-06,
+      "loss": 0.8379,
+      "step": 1943
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.590066820511438e-06,
+      "loss": 0.8525,
+      "step": 1944
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.580377957820427e-06,
+      "loss": 0.876,
+      "step": 1945
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.570696293214814e-06,
+      "loss": 0.8569,
+      "step": 1946
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.561021839553323e-06,
+      "loss": 0.8809,
+      "step": 1947
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.55135460968511e-06,
+      "loss": 0.8198,
+      "step": 1948
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.541694616449729e-06,
+      "loss": 0.854,
+      "step": 1949
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.532041872677128e-06,
+      "loss": 0.8721,
+      "step": 1950
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.5223963911876265e-06,
+      "loss": 0.8608,
+      "step": 1951
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.512758184791893e-06,
+      "loss": 0.8521,
+      "step": 1952
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.503127266290935e-06,
+      "loss": 0.8262,
+      "step": 1953
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.493503648476085e-06,
+      "loss": 0.8335,
+      "step": 1954
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.48388734412898e-06,
+      "loss": 0.8359,
+      "step": 1955
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.474278366021533e-06,
+      "loss": 0.8569,
+      "step": 1956
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.464676726915939e-06,
+      "loss": 0.8149,
+      "step": 1957
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.455082439564642e-06,
+      "loss": 0.8267,
+      "step": 1958
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.445495516710312e-06,
+      "loss": 0.8657,
+      "step": 1959
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.43591597108585e-06,
+      "loss": 0.8574,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.426343815414361e-06,
+      "loss": 0.8232,
+      "step": 1961
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.416779062409125e-06,
+      "loss": 0.8481,
+      "step": 1962
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.407221724773587e-06,
+      "loss": 0.8643,
+      "step": 1963
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.397671815201362e-06,
+      "loss": 0.8857,
+      "step": 1964
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.388129346376177e-06,
+      "loss": 0.8418,
+      "step": 1965
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3785943309718935e-06,
+      "loss": 0.8159,
+      "step": 1966
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.369066781652469e-06,
+      "loss": 0.8071,
+      "step": 1967
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.359546711071943e-06,
+      "loss": 0.8535,
+      "step": 1968
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.350034131874414e-06,
+      "loss": 0.8506,
+      "step": 1969
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3405290566940475e-06,
+      "loss": 0.8403,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.331031498155035e-06,
+      "loss": 0.8291,
+      "step": 1971
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.321541468871578e-06,
+      "loss": 0.8481,
+      "step": 1972
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.312058981447893e-06,
+      "loss": 0.311,
+      "step": 1973
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.302584048478164e-06,
+      "loss": 0.3103,
+      "step": 1974
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.293116682546546e-06,
+      "loss": 0.8672,
+      "step": 1975
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.283656896227151e-06,
+      "loss": 0.8359,
+      "step": 1976
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.2742047020840214e-06,
+      "loss": 0.8145,
+      "step": 1977
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.264760112671111e-06,
+      "loss": 0.855,
+      "step": 1978
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.255323140532272e-06,
+      "loss": 0.8315,
+      "step": 1979
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.2458937982012525e-06,
+      "loss": 0.8379,
+      "step": 1980
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.236472098201651e-06,
+      "loss": 0.3,
+      "step": 1981
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.227058053046926e-06,
+      "loss": 0.8335,
+      "step": 1982
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.217651675240371e-06,
+      "loss": 0.8452,
+      "step": 1983
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.208252977275087e-06,
+      "loss": 0.8174,
+      "step": 1984
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.198861971633977e-06,
+      "loss": 0.8252,
+      "step": 1985
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.189478670789737e-06,
+      "loss": 0.8198,
+      "step": 1986
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.180103087204817e-06,
+      "loss": 0.8633,
+      "step": 1987
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.170735233331424e-06,
+      "loss": 0.873,
+      "step": 1988
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.161375121611504e-06,
+      "loss": 0.2961,
+      "step": 1989
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.152022764476709e-06,
+      "loss": 0.856,
+      "step": 1990
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.142678174348395e-06,
+      "loss": 0.8423,
+      "step": 1991
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.133341363637607e-06,
+      "loss": 0.8662,
+      "step": 1992
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.1240123447450575e-06,
+      "loss": 0.8267,
+      "step": 1993
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.1146911300611045e-06,
+      "loss": 0.8428,
+      "step": 1994
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.105377731965743e-06,
+      "loss": 0.8237,
+      "step": 1995
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.096072162828591e-06,
+      "loss": 0.8608,
+      "step": 1996
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.08677443500886e-06,
+      "loss": 0.8247,
+      "step": 1997
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.077484560855354e-06,
+      "loss": 0.832,
+      "step": 1998
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.0682025527064486e-06,
+      "loss": 0.8491,
+      "step": 1999
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.058928422890064e-06,
+      "loss": 0.8652,
+      "step": 2000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.049662183723655e-06,
+      "loss": 0.8491,
+      "step": 2001
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.040403847514211e-06,
+      "loss": 0.2705,
+      "step": 2002
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.031153426558209e-06,
+      "loss": 0.8218,
+      "step": 2003
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.0219109331416215e-06,
+      "loss": 0.8418,
+      "step": 2004
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.012676379539896e-06,
+      "loss": 0.8447,
+      "step": 2005
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.0034497780179245e-06,
+      "loss": 0.8447,
+      "step": 2006
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.9942311408300395e-06,
+      "loss": 0.8271,
+      "step": 2007
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.985020480220002e-06,
+      "loss": 0.8652,
+      "step": 2008
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.975817808420978e-06,
+      "loss": 0.3035,
+      "step": 2009
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.9666231376555195e-06,
+      "loss": 0.8086,
+      "step": 2010
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.957436480135547e-06,
+      "loss": 0.8584,
+      "step": 2011
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.948257848062351e-06,
+      "loss": 0.8247,
+      "step": 2012
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.93908725362655e-06,
+      "loss": 0.8579,
+      "step": 2013
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.929924709008096e-06,
+      "loss": 0.8452,
+      "step": 2014
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.920770226376251e-06,
+      "loss": 0.8481,
+      "step": 2015
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.911623817889563e-06,
+      "loss": 0.2876,
+      "step": 2016
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.902485495695853e-06,
+      "loss": 0.8408,
+      "step": 2017
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.8933552719322164e-06,
+      "loss": 0.3396,
+      "step": 2018
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.884233158724976e-06,
+      "loss": 0.8306,
+      "step": 2019
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.875119168189696e-06,
+      "loss": 0.8589,
+      "step": 2020
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.866013312431148e-06,
+      "loss": 0.8447,
+      "step": 2021
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.856915603543297e-06,
+      "loss": 0.8335,
+      "step": 2022
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.847826053609286e-06,
+      "loss": 0.8433,
+      "step": 2023
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.83874467470143e-06,
+      "loss": 0.8335,
+      "step": 2024
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.82967147888118e-06,
+      "loss": 0.835,
+      "step": 2025
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.820606478199134e-06,
+      "loss": 0.8823,
+      "step": 2026
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.8115496846949885e-06,
+      "loss": 0.8438,
+      "step": 2027
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.802501110397553e-06,
+      "loss": 0.8784,
+      "step": 2028
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7934607673247116e-06,
+      "loss": 0.8418,
+      "step": 2029
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.784428667483421e-06,
+      "loss": 0.8613,
+      "step": 2030
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.775404822869694e-06,
+      "loss": 0.8574,
+      "step": 2031
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.766389245468568e-06,
+      "loss": 0.3101,
+      "step": 2032
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.757381947254104e-06,
+      "loss": 0.855,
+      "step": 2033
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7483829401893766e-06,
+      "loss": 0.312,
+      "step": 2034
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.739392236226432e-06,
+      "loss": 0.853,
+      "step": 2035
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7304098473063034e-06,
+      "loss": 0.8213,
+      "step": 2036
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7214357853589765e-06,
+      "loss": 0.8608,
+      "step": 2037
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7124700623033725e-06,
+      "loss": 0.8794,
+      "step": 2038
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7035126900473363e-06,
+      "loss": 0.8452,
+      "step": 2039
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6945636804876328e-06,
+      "loss": 0.8828,
+      "step": 2040
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6856230455099053e-06,
+      "loss": 0.8291,
+      "step": 2041
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.67669079698869e-06,
+      "loss": 0.8394,
+      "step": 2042
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.667766946787369e-06,
+      "loss": 0.8672,
+      "step": 2043
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.658851506758184e-06,
+      "loss": 0.8491,
+      "step": 2044
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.649944488742194e-06,
+      "loss": 0.8589,
+      "step": 2045
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6410459045692824e-06,
+      "loss": 0.8423,
+      "step": 2046
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.632155766058131e-06,
+      "loss": 0.8311,
+      "step": 2047
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6232740850161953e-06,
+      "loss": 0.3455,
+      "step": 2048
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.614400873239703e-06,
+      "loss": 0.8394,
+      "step": 2049
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6055361425136383e-06,
+      "loss": 0.8535,
+      "step": 2050
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.596679904611715e-06,
+      "loss": 0.8271,
+      "step": 2051
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5878321712963617e-06,
+      "loss": 0.8848,
+      "step": 2052
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5789929543187317e-06,
+      "loss": 0.8291,
+      "step": 2053
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5701622654186485e-06,
+      "loss": 0.7563,
+      "step": 2054
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5613401163246118e-06,
+      "loss": 0.8457,
+      "step": 2055
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5525265187537896e-06,
+      "loss": 0.8589,
+      "step": 2056
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.543721484411976e-06,
+      "loss": 0.8579,
+      "step": 2057
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.534925024993606e-06,
+      "loss": 0.8389,
+      "step": 2058
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5261371521817247e-06,
+      "loss": 0.8584,
+      "step": 2059
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5173578776479624e-06,
+      "loss": 0.8433,
+      "step": 2060
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5085872130525345e-06,
+      "loss": 0.8574,
+      "step": 2061
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.4998251700442244e-06,
+      "loss": 0.8345,
+      "step": 2062
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.491071760260368e-06,
+      "loss": 0.8721,
+      "step": 2063
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.482326995326826e-06,
+      "loss": 0.8179,
+      "step": 2064
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.473590886857977e-06,
+      "loss": 0.8457,
+      "step": 2065
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.464863446456713e-06,
+      "loss": 0.8335,
+      "step": 2066
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.4561446857144054e-06,
+      "loss": 0.8311,
+      "step": 2067
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.447434616210891e-06,
+      "loss": 0.8325,
+      "step": 2068
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.4387332495144866e-06,
+      "loss": 0.8149,
+      "step": 2069
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.4300405971819274e-06,
+      "loss": 0.2987,
+      "step": 2070
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.42135667075838e-06,
+      "loss": 0.8232,
+      "step": 2071
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.41268148177743e-06,
+      "loss": 0.8467,
+      "step": 2072
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.4040150417610483e-06,
+      "loss": 0.8262,
+      "step": 2073
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3953573622195923e-06,
+      "loss": 0.8169,
+      "step": 2074
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3867084546517847e-06,
+      "loss": 0.8408,
+      "step": 2075
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3780683305446926e-06,
+      "loss": 0.8545,
+      "step": 2076
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3694370013737153e-06,
+      "loss": 0.8252,
+      "step": 2077
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.360814478602584e-06,
+      "loss": 0.8149,
+      "step": 2078
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.352200773683317e-06,
+      "loss": 0.8394,
+      "step": 2079
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3435958980562364e-06,
+      "loss": 0.8789,
+      "step": 2080
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3349998631499247e-06,
+      "loss": 0.8442,
+      "step": 2081
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3264126803812346e-06,
+      "loss": 0.8345,
+      "step": 2082
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.317834361155252e-06,
+      "loss": 0.8301,
+      "step": 2083
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.309264916865288e-06,
+      "loss": 0.8271,
+      "step": 2084
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3007043588928866e-06,
+      "loss": 0.8491,
+      "step": 2085
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.292152698607768e-06,
+      "loss": 0.8203,
+      "step": 2086
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2836099473678384e-06,
+      "loss": 0.8154,
+      "step": 2087
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2750761165191846e-06,
+      "loss": 0.8667,
+      "step": 2088
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.266551217396029e-06,
+      "loss": 0.8496,
+      "step": 2089
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.258035261320741e-06,
+      "loss": 0.8311,
+      "step": 2090
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2495282596038156e-06,
+      "loss": 0.8252,
+      "step": 2091
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.241030223543845e-06,
+      "loss": 0.8169,
+      "step": 2092
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2325411644275164e-06,
+      "loss": 0.8369,
+      "step": 2093
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2240610935296e-06,
+      "loss": 0.8311,
+      "step": 2094
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.215590022112921e-06,
+      "loss": 0.8633,
+      "step": 2095
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.2071279614283614e-06,
+      "loss": 0.7969,
+      "step": 2096
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1986749227148215e-06,
+      "loss": 0.833,
+      "step": 2097
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.190230917199234e-06,
+      "loss": 0.8462,
+      "step": 2098
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.181795956096522e-06,
+      "loss": 0.8555,
+      "step": 2099
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1733700506096032e-06,
+      "loss": 0.8506,
+      "step": 2100
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1649532119293713e-06,
+      "loss": 0.2937,
+      "step": 2101
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1565454512346684e-06,
+      "loss": 0.8755,
+      "step": 2102
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1481467796922804e-06,
+      "loss": 0.8408,
+      "step": 2103
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.139757208456933e-06,
+      "loss": 0.8516,
+      "step": 2104
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.131376748671253e-06,
+      "loss": 0.8506,
+      "step": 2105
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.123005411465766e-06,
+      "loss": 0.8384,
+      "step": 2106
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1146432079588963e-06,
+      "loss": 0.8188,
+      "step": 2107
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1062901492569207e-06,
+      "loss": 0.8447,
+      "step": 2108
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0979462464539744e-06,
+      "loss": 0.8677,
+      "step": 2109
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.089611510632039e-06,
+      "loss": 0.8569,
+      "step": 2110
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0812859528609106e-06,
+      "loss": 0.8525,
+      "step": 2111
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0729695841982064e-06,
+      "loss": 0.8252,
+      "step": 2112
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.064662415689328e-06,
+      "loss": 0.8125,
+      "step": 2113
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.056364458367469e-06,
+      "loss": 0.8267,
+      "step": 2114
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0480757232535773e-06,
+      "loss": 0.8184,
+      "step": 2115
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0397962213563614e-06,
+      "loss": 0.8726,
+      "step": 2116
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.031525963672267e-06,
+      "loss": 0.2885,
+      "step": 2117
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0232649611854568e-06,
+      "loss": 0.7778,
+      "step": 2118
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0150132248677976e-06,
+      "loss": 0.8335,
+      "step": 2119
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0067707656788638e-06,
+      "loss": 0.3062,
+      "step": 2120
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.9985375945658934e-06,
+      "loss": 0.8574,
+      "step": 2121
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.990313722463791e-06,
+      "loss": 0.8115,
+      "step": 2122
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9820991602951255e-06,
+      "loss": 0.3176,
+      "step": 2123
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.973893918970082e-06,
+      "loss": 0.8306,
+      "step": 2124
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.965698009386473e-06,
+      "loss": 0.8481,
+      "step": 2125
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9575114424297213e-06,
+      "loss": 0.8418,
+      "step": 2126
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9493342289728334e-06,
+      "loss": 0.8267,
+      "step": 2127
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9411663798764013e-06,
+      "loss": 0.8418,
+      "step": 2128
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9330079059885708e-06,
+      "loss": 0.8535,
+      "step": 2129
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9248588181450456e-06,
+      "loss": 0.8525,
+      "step": 2130
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.916719127169053e-06,
+      "loss": 0.8408,
+      "step": 2131
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9085888438713515e-06,
+      "loss": 0.8384,
+      "step": 2132
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9004679790501922e-06,
+      "loss": 0.853,
+      "step": 2133
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8923565434913283e-06,
+      "loss": 0.8462,
+      "step": 2134
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8842545479679796e-06,
+      "loss": 0.8149,
+      "step": 2135
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.876162003240839e-06,
+      "loss": 0.8188,
+      "step": 2136
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8680789200580373e-06,
+      "loss": 0.8252,
+      "step": 2137
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8600053091551373e-06,
+      "loss": 0.8472,
+      "step": 2138
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.851941181255139e-06,
+      "loss": 0.3127,
+      "step": 2139
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8438865470684284e-06,
+      "loss": 0.8354,
+      "step": 2140
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.835841417292788e-06,
+      "loss": 0.8569,
+      "step": 2141
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8278058026133835e-06,
+      "loss": 0.8525,
+      "step": 2142
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8197797137027338e-06,
+      "loss": 0.8682,
+      "step": 2143
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8117631612207084e-06,
+      "loss": 0.8262,
+      "step": 2144
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8037561558145154e-06,
+      "loss": 0.2977,
+      "step": 2145
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.7957587081186834e-06,
+      "loss": 0.834,
+      "step": 2146
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.7877708287550366e-06,
+      "loss": 0.8574,
+      "step": 2147
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.7797925283327053e-06,
+      "loss": 0.8652,
+      "step": 2148
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.771823817448085e-06,
+      "loss": 0.8301,
+      "step": 2149
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.7638647066848456e-06,
+      "loss": 0.8389,
+      "step": 2150
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.755915206613895e-06,
+      "loss": 0.8608,
+      "step": 2151
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7479753277933884e-06,
+      "loss": 0.8267,
+      "step": 2152
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.740045080768694e-06,
+      "loss": 0.8369,
+      "step": 2153
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7321244760723852e-06,
+      "loss": 0.8354,
+      "step": 2154
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.724213524224246e-06,
+      "loss": 0.8154,
+      "step": 2155
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.716312235731222e-06,
+      "loss": 0.8252,
+      "step": 2156
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7084206210874277e-06,
+      "loss": 0.8638,
+      "step": 2157
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.700538690774137e-06,
+      "loss": 0.8301,
+      "step": 2158
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6926664552597537e-06,
+      "loss": 0.8389,
+      "step": 2159
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.684803924999807e-06,
+      "loss": 0.8242,
+      "step": 2160
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6769511104369384e-06,
+      "loss": 0.8823,
+      "step": 2161
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.669108022000888e-06,
+      "loss": 0.8535,
+      "step": 2162
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.661274670108469e-06,
+      "loss": 0.8394,
+      "step": 2163
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.653451065163574e-06,
+      "loss": 0.8599,
+      "step": 2164
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.645637217557139e-06,
+      "loss": 0.3252,
+      "step": 2165
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6378331376671517e-06,
+      "loss": 0.3086,
+      "step": 2166
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.630038835858617e-06,
+      "loss": 0.8345,
+      "step": 2167
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6222543224835618e-06,
+      "loss": 0.8091,
+      "step": 2168
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6144796078810065e-06,
+      "loss": 0.8267,
+      "step": 2169
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6067147023769556e-06,
+      "loss": 0.7991,
+      "step": 2170
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.598959616284391e-06,
+      "loss": 0.8433,
+      "step": 2171
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5912143599032556e-06,
+      "loss": 0.8516,
+      "step": 2172
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5834789435204245e-06,
+      "loss": 0.8623,
+      "step": 2173
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.575753377409719e-06,
+      "loss": 0.8374,
+      "step": 2174
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5680376718318657e-06,
+      "loss": 0.8569,
+      "step": 2175
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.560331837034494e-06,
+      "loss": 0.854,
+      "step": 2176
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5526358832521424e-06,
+      "loss": 0.3154,
+      "step": 2177
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.544949820706204e-06,
+      "loss": 0.8501,
+      "step": 2178
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.5372736596049417e-06,
+      "loss": 0.8418,
+      "step": 2179
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.529607410143474e-06,
+      "loss": 0.8296,
+      "step": 2180
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.521951082503746e-06,
+      "loss": 0.8413,
+      "step": 2181
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.514304686854534e-06,
+      "loss": 0.8677,
+      "step": 2182
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.5066682333514136e-06,
+      "loss": 0.8345,
+      "step": 2183
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4990417321367678e-06,
+      "loss": 0.8584,
+      "step": 2184
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.491425193339748e-06,
+      "loss": 0.8599,
+      "step": 2185
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4838186270762823e-06,
+      "loss": 0.8379,
+      "step": 2186
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4762220434490504e-06,
+      "loss": 0.8481,
+      "step": 2187
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.46863545254748e-06,
+      "loss": 0.8325,
+      "step": 2188
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.461058864447716e-06,
+      "loss": 0.8218,
+      "step": 2189
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.45349228921263e-06,
+      "loss": 0.8442,
+      "step": 2190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.445935736891785e-06,
+      "loss": 0.8257,
+      "step": 2191
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.438389217521431e-06,
+      "loss": 0.2865,
+      "step": 2192
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.43085274112451e-06,
+      "loss": 0.8247,
+      "step": 2193
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.423326317710606e-06,
+      "loss": 0.8496,
+      "step": 2194
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4158099572759564e-06,
+      "loss": 0.8306,
+      "step": 2195
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.40830366980344e-06,
+      "loss": 0.2992,
+      "step": 2196
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.40080746526255e-06,
+      "loss": 0.3,
+      "step": 2197
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.393321353609388e-06,
+      "loss": 0.8223,
+      "step": 2198
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.385845344786656e-06,
+      "loss": 0.876,
+      "step": 2199
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3783794487236367e-06,
+      "loss": 0.8184,
+      "step": 2200
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3709236753361777e-06,
+      "loss": 0.8472,
+      "step": 2201
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3634780345266805e-06,
+      "loss": 0.3197,
+      "step": 2202
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3560425361840976e-06,
+      "loss": 0.8311,
+      "step": 2203
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3486171901839063e-06,
+      "loss": 0.8501,
+      "step": 2204
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3412020063880957e-06,
+      "loss": 0.8052,
+      "step": 2205
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3337969946451667e-06,
+      "loss": 0.8374,
+      "step": 2206
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3264021647901014e-06,
+      "loss": 0.8369,
+      "step": 2207
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.319017526644359e-06,
+      "loss": 0.8721,
+      "step": 2208
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.311643090015877e-06,
+      "loss": 0.8511,
+      "step": 2209
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3042788646990267e-06,
+      "loss": 0.8047,
+      "step": 2210
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.296924860474621e-06,
+      "loss": 0.8418,
+      "step": 2211
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2895810871099057e-06,
+      "loss": 0.8188,
+      "step": 2212
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.282247554358531e-06,
+      "loss": 0.8721,
+      "step": 2213
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2749242719605435e-06,
+      "loss": 0.2896,
+      "step": 2214
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.267611249642383e-06,
+      "loss": 0.8359,
+      "step": 2215
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2603084971168643e-06,
+      "loss": 0.8423,
+      "step": 2216
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.25301602408315e-06,
+      "loss": 0.8052,
+      "step": 2217
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2457338402267637e-06,
+      "loss": 0.8647,
+      "step": 2218
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2384619552195518e-06,
+      "loss": 0.8325,
+      "step": 2219
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.231200378719693e-06,
+      "loss": 0.8452,
+      "step": 2220
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2239491203716644e-06,
+      "loss": 0.8379,
+      "step": 2221
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.216708189806248e-06,
+      "loss": 0.835,
+      "step": 2222
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2094775966405045e-06,
+      "loss": 0.8511,
+      "step": 2223
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.202257350477761e-06,
+      "loss": 0.8071,
+      "step": 2224
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.19504746090761e-06,
+      "loss": 0.8447,
+      "step": 2225
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.187847937505887e-06,
+      "loss": 0.8218,
+      "step": 2226
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1806587898346553e-06,
+      "loss": 0.8091,
+      "step": 2227
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.173480027442203e-06,
+      "loss": 0.855,
+      "step": 2228
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1663116598630207e-06,
+      "loss": 0.853,
+      "step": 2229
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.159153696617793e-06,
+      "loss": 0.8438,
+      "step": 2230
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1520061472133903e-06,
+      "loss": 0.8335,
+      "step": 2231
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1448690211428514e-06,
+      "loss": 0.8555,
+      "step": 2232
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1377423278853627e-06,
+      "loss": 0.8452,
+      "step": 2233
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.130626076906269e-06,
+      "loss": 0.3401,
+      "step": 2234
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1235202776570297e-06,
+      "loss": 0.8271,
+      "step": 2235
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.116424939575239e-06,
+      "loss": 0.8438,
+      "step": 2236
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1093400720845813e-06,
+      "loss": 0.3022,
+      "step": 2237
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1022656845948486e-06,
+      "loss": 0.793,
+      "step": 2238
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0952017865019036e-06,
+      "loss": 0.8428,
+      "step": 2239
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0881483871876807e-06,
+      "loss": 0.8237,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.081105496020173e-06,
+      "loss": 0.3278,
+      "step": 2241
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.074073122353416e-06,
+      "loss": 0.8271,
+      "step": 2242
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.067051275527472e-06,
+      "loss": 0.8511,
+      "step": 2243
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0600399648684297e-06,
+      "loss": 0.833,
+      "step": 2244
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0530391996883782e-06,
+      "loss": 0.8247,
+      "step": 2245
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0460489892854e-06,
+      "loss": 0.8384,
+      "step": 2246
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0390693429435626e-06,
+      "loss": 0.8096,
+      "step": 2247
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.032100269932905e-06,
+      "loss": 0.8115,
+      "step": 2248
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0251417795094166e-06,
+      "loss": 0.8242,
+      "step": 2249
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.018193880915037e-06,
+      "loss": 0.8452,
+      "step": 2250
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0112565833776364e-06,
+      "loss": 0.856,
+      "step": 2251
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0043298961110003e-06,
+      "loss": 0.8296,
+      "step": 2252
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.99741382831483e-06,
+      "loss": 0.853,
+      "step": 2253
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9905083891747235e-06,
+      "loss": 0.3037,
+      "step": 2254
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.983613587862153e-06,
+      "loss": 0.854,
+      "step": 2255
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9767294335344668e-06,
+      "loss": 0.8252,
+      "step": 2256
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9698559353348735e-06,
+      "loss": 0.8027,
+      "step": 2257
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.962993102392431e-06,
+      "loss": 0.2931,
+      "step": 2258
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9561409438220245e-06,
+      "loss": 0.8496,
+      "step": 2259
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9492994687243715e-06,
+      "loss": 0.8413,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9424686861859933e-06,
+      "loss": 0.8252,
+      "step": 2261
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9356486052792076e-06,
+      "loss": 0.8184,
+      "step": 2262
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9288392350621275e-06,
+      "loss": 0.8237,
+      "step": 2263
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9220405845786372e-06,
+      "loss": 0.3109,
+      "step": 2264
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.915252662858378e-06,
+      "loss": 0.8188,
+      "step": 2265
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9084754789167516e-06,
+      "loss": 0.8438,
+      "step": 2266
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.901709041754889e-06,
+      "loss": 0.8428,
+      "step": 2267
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8949533603596505e-06,
+      "loss": 0.8442,
+      "step": 2268
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8882084437036142e-06,
+      "loss": 0.8345,
+      "step": 2269
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8814743007450608e-06,
+      "loss": 0.8369,
+      "step": 2270
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8747509404279595e-06,
+      "loss": 0.8462,
+      "step": 2271
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.868038371681955e-06,
+      "loss": 0.8174,
+      "step": 2272
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8613366034223668e-06,
+      "loss": 0.8228,
+      "step": 2273
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8546456445501682e-06,
+      "loss": 0.7979,
+      "step": 2274
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8479655039519683e-06,
+      "loss": 0.8281,
+      "step": 2275
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8412961905000205e-06,
+      "loss": 0.8516,
+      "step": 2276
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8346377130521864e-06,
+      "loss": 0.8472,
+      "step": 2277
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.827990080451939e-06,
+      "loss": 0.7822,
+      "step": 2278
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8213533015283524e-06,
+      "loss": 0.8403,
+      "step": 2279
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8147273850960845e-06,
+      "loss": 0.3223,
+      "step": 2280
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8081123399553569e-06,
+      "loss": 0.8188,
+      "step": 2281
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8015081748919672e-06,
+      "loss": 0.8271,
+      "step": 2282
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.794914898677249e-06,
+      "loss": 0.2708,
+      "step": 2283
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7883325200680802e-06,
+      "loss": 0.3103,
+      "step": 2284
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7817610478068659e-06,
+      "loss": 0.3262,
+      "step": 2285
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.775200490621526e-06,
+      "loss": 0.311,
+      "step": 2286
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7686508572254802e-06,
+      "loss": 0.8677,
+      "step": 2287
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.762112156317638e-06,
+      "loss": 0.8359,
+      "step": 2288
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7555843965823992e-06,
+      "loss": 0.297,
+      "step": 2289
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.749067586689619e-06,
+      "loss": 0.8438,
+      "step": 2290
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7425617352946178e-06,
+      "loss": 0.874,
+      "step": 2291
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7360668510381628e-06,
+      "loss": 0.8394,
+      "step": 2292
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7295829425464494e-06,
+      "loss": 0.8574,
+      "step": 2293
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7231100184310955e-06,
+      "loss": 0.29,
+      "step": 2294
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7166480872891333e-06,
+      "loss": 0.8223,
+      "step": 2295
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7101971577029985e-06,
+      "loss": 0.8618,
+      "step": 2296
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7037572382405031e-06,
+      "loss": 0.2986,
+      "step": 2297
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.69732833745485e-06,
+      "loss": 0.8062,
+      "step": 2298
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6909104638845986e-06,
+      "loss": 0.8477,
+      "step": 2299
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6845036260536595e-06,
+      "loss": 0.7935,
+      "step": 2300
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6781078324712973e-06,
+      "loss": 0.8208,
+      "step": 2301
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6717230916321026e-06,
+      "loss": 0.8408,
+      "step": 2302
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6653494120159842e-06,
+      "loss": 0.8389,
+      "step": 2303
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6589868020881582e-06,
+      "loss": 0.855,
+      "step": 2304
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6526352702991478e-06,
+      "loss": 0.3049,
+      "step": 2305
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.646294825084751e-06,
+      "loss": 0.8467,
+      "step": 2306
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6399654748660498e-06,
+      "loss": 0.8301,
+      "step": 2307
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6336472280493908e-06,
+      "loss": 0.854,
+      "step": 2308
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6273400930263672e-06,
+      "loss": 0.8721,
+      "step": 2309
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6210440781738146e-06,
+      "loss": 0.8223,
+      "step": 2310
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.614759191853803e-06,
+      "loss": 0.8452,
+      "step": 2311
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6084854424136253e-06,
+      "loss": 0.8433,
+      "step": 2312
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6022228381857729e-06,
+      "loss": 0.8491,
+      "step": 2313
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5959713874879445e-06,
+      "loss": 0.8198,
+      "step": 2314
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5897310986230196e-06,
+      "loss": 0.8721,
+      "step": 2315
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.583501979879052e-06,
+      "loss": 0.3275,
+      "step": 2316
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5772840395292632e-06,
+      "loss": 0.8408,
+      "step": 2317
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.57107728583203e-06,
+      "loss": 0.8184,
+      "step": 2318
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5648817270308648e-06,
+      "loss": 0.8677,
+      "step": 2319
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5586973713544207e-06,
+      "loss": 0.8252,
+      "step": 2320
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.552524227016462e-06,
+      "loss": 0.8179,
+      "step": 2321
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5463623022158659e-06,
+      "loss": 0.8237,
+      "step": 2322
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5402116051366111e-06,
+      "loss": 0.8403,
+      "step": 2323
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5340721439477635e-06,
+      "loss": 0.8364,
+      "step": 2324
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5279439268034634e-06,
+      "loss": 0.8452,
+      "step": 2325
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.521826961842917e-06,
+      "loss": 0.8589,
+      "step": 2326
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.515721257190389e-06,
+      "loss": 0.8281,
+      "step": 2327
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5096268209551902e-06,
+      "loss": 0.8564,
+      "step": 2328
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5035436612316567e-06,
+      "loss": 0.8223,
+      "step": 2329
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4974717860991583e-06,
+      "loss": 0.8555,
+      "step": 2330
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4914112036220696e-06,
+      "loss": 0.2887,
+      "step": 2331
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4853619218497672e-06,
+      "loss": 0.813,
+      "step": 2332
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4793239488166222e-06,
+      "loss": 0.856,
+      "step": 2333
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4732972925419887e-06,
+      "loss": 0.8438,
+      "step": 2334
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4672819610301802e-06,
+      "loss": 0.8345,
+      "step": 2335
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.461277962270481e-06,
+      "loss": 0.8311,
+      "step": 2336
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.455285304237114e-06,
+      "loss": 0.8667,
+      "step": 2337
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4493039948892418e-06,
+      "loss": 0.3086,
+      "step": 2338
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4433340421709597e-06,
+      "loss": 0.8643,
+      "step": 2339
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4373754540112783e-06,
+      "loss": 0.8257,
+      "step": 2340
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4314282383241097e-06,
+      "loss": 0.3242,
+      "step": 2341
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4254924030082618e-06,
+      "loss": 0.8213,
+      "step": 2342
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.419567955947434e-06,
+      "loss": 0.7876,
+      "step": 2343
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4136549050101933e-06,
+      "loss": 0.8604,
+      "step": 2344
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4077532580499753e-06,
+      "loss": 0.8403,
+      "step": 2345
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.40186302290507e-06,
+      "loss": 0.8481,
+      "step": 2346
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3959842073986085e-06,
+      "loss": 0.7876,
+      "step": 2347
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3901168193385506e-06,
+      "loss": 0.8115,
+      "step": 2348
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.384260866517686e-06,
+      "loss": 0.8335,
+      "step": 2349
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3784163567136177e-06,
+      "loss": 0.8188,
+      "step": 2350
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.372583297688741e-06,
+      "loss": 0.833,
+      "step": 2351
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3667616971902542e-06,
+      "loss": 0.8447,
+      "step": 2352
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3609515629501279e-06,
+      "loss": 0.8091,
+      "step": 2353
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3551529026851063e-06,
+      "loss": 0.856,
+      "step": 2354
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3493657240966974e-06,
+      "loss": 0.3005,
+      "step": 2355
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3435900348711594e-06,
+      "loss": 0.7788,
+      "step": 2356
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3378258426794888e-06,
+      "loss": 0.8491,
+      "step": 2357
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.332073155177408e-06,
+      "loss": 0.323,
+      "step": 2358
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3263319800053698e-06,
+      "loss": 0.8447,
+      "step": 2359
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.320602324788527e-06,
+      "loss": 0.8462,
+      "step": 2360
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3148841971367387e-06,
+      "loss": 0.3262,
+      "step": 2361
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3091776046445537e-06,
+      "loss": 0.856,
+      "step": 2362
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3034825548911944e-06,
+      "loss": 0.8452,
+      "step": 2363
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2977990554405551e-06,
+      "loss": 0.3098,
+      "step": 2364
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2921271138411927e-06,
+      "loss": 0.8579,
+      "step": 2365
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.286466737626314e-06,
+      "loss": 0.8359,
+      "step": 2366
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2808179343137583e-06,
+      "loss": 0.8364,
+      "step": 2367
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2751807114060033e-06,
+      "loss": 0.8242,
+      "step": 2368
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2695550763901376e-06,
+      "loss": 0.8086,
+      "step": 2369
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2639410367378623e-06,
+      "loss": 0.8252,
+      "step": 2370
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.258338599905482e-06,
+      "loss": 0.8477,
+      "step": 2371
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.25274777333389e-06,
+      "loss": 0.835,
+      "step": 2372
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2471685644485543e-06,
+      "loss": 0.8281,
+      "step": 2373
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.241600980659514e-06,
+      "loss": 0.8154,
+      "step": 2374
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2360450293613757e-06,
+      "loss": 0.3516,
+      "step": 2375
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2305007179332851e-06,
+      "loss": 0.8633,
+      "step": 2376
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2249680537389375e-06,
+      "loss": 0.8311,
+      "step": 2377
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2194470441265572e-06,
+      "loss": 0.8145,
+      "step": 2378
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.213937696428885e-06,
+      "loss": 0.8252,
+      "step": 2379
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.208440017963175e-06,
+      "loss": 0.8306,
+      "step": 2380
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2029540160311859e-06,
+      "loss": 0.8472,
+      "step": 2381
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1974796979191628e-06,
+      "loss": 0.8477,
+      "step": 2382
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1920170708978374e-06,
+      "loss": 0.8623,
+      "step": 2383
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.186566142222414e-06,
+      "loss": 0.8228,
+      "step": 2384
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.181126919132557e-06,
+      "loss": 0.8394,
+      "step": 2385
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1756994088523832e-06,
+      "loss": 0.8384,
+      "step": 2386
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1702836185904543e-06,
+      "loss": 0.8472,
+      "step": 2387
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1648795555397719e-06,
+      "loss": 0.8579,
+      "step": 2388
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1594872268777513e-06,
+      "loss": 0.8325,
+      "step": 2389
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1541066397662283e-06,
+      "loss": 0.7847,
+      "step": 2390
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1487378013514483e-06,
+      "loss": 0.8164,
+      "step": 2391
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1433807187640422e-06,
+      "loss": 0.8467,
+      "step": 2392
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1380353991190373e-06,
+      "loss": 0.8306,
+      "step": 2393
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1327018495158348e-06,
+      "loss": 0.8281,
+      "step": 2394
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1273800770382027e-06,
+      "loss": 0.8184,
+      "step": 2395
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1220700887542635e-06,
+      "loss": 0.8359,
+      "step": 2396
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1167718917164961e-06,
+      "loss": 0.7686,
+      "step": 2397
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1114854929617114e-06,
+      "loss": 0.8257,
+      "step": 2398
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1062108995110566e-06,
+      "loss": 0.853,
+      "step": 2399
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1009481183699977e-06,
+      "loss": 0.8389,
+      "step": 2400
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.0956971565283114e-06,
+      "loss": 0.8511,
+      "step": 2401
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.0904580209600724e-06,
+      "loss": 0.8364,
+      "step": 2402
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.0852307186236554e-06,
+      "loss": 0.8398,
+      "step": 2403
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0800152564617184e-06,
+      "loss": 0.8394,
+      "step": 2404
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.074811641401189e-06,
+      "loss": 0.8271,
+      "step": 2405
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0696198803532598e-06,
+      "loss": 0.7842,
+      "step": 2406
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0644399802133876e-06,
+      "loss": 0.8608,
+      "step": 2407
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0592719478612635e-06,
+      "loss": 0.8057,
+      "step": 2408
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.054115790160829e-06,
+      "loss": 0.8428,
+      "step": 2409
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0489715139602475e-06,
+      "loss": 0.8184,
+      "step": 2410
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0438391260919034e-06,
+      "loss": 0.8413,
+      "step": 2411
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0387186333723875e-06,
+      "loss": 0.832,
+      "step": 2412
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0336100426025008e-06,
+      "loss": 0.8433,
+      "step": 2413
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0285133605672283e-06,
+      "loss": 0.3145,
+      "step": 2414
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0234285940357424e-06,
+      "loss": 0.8354,
+      "step": 2415
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0183557497613915e-06,
+      "loss": 0.8354,
+      "step": 2416
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0132948344816863e-06,
+      "loss": 0.8394,
+      "step": 2417
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0082458549182915e-06,
+      "loss": 0.8447,
+      "step": 2418
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.003208817777025e-06,
+      "loss": 0.8247,
+      "step": 2419
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.981837297478437e-07,
+      "loss": 0.8477,
+      "step": 2420
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.931705975048279e-07,
+      "loss": 0.8579,
+      "step": 2421
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.881694277061848e-07,
+      "loss": 0.2976,
+      "step": 2422
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.831802269942304e-07,
+      "loss": 0.8281,
+      "step": 2423
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.782030019953836e-07,
+      "loss": 0.832,
+      "step": 2424
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.73237759320159e-07,
+      "loss": 0.832,
+      "step": 2425
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.682845055631618e-07,
+      "loss": 0.8452,
+      "step": 2426
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.633432473030635e-07,
+      "loss": 0.8721,
+      "step": 2427
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.584139911026102e-07,
+      "loss": 0.8433,
+      "step": 2428
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.534967435086095e-07,
+      "loss": 0.8174,
+      "step": 2429
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.485915110519139e-07,
+      "loss": 0.316,
+      "step": 2430
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.436983002474209e-07,
+      "loss": 0.8022,
+      "step": 2431
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.388171175940641e-07,
+      "loss": 0.8789,
+      "step": 2432
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.339479695747988e-07,
+      "loss": 0.8047,
+      "step": 2433
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.290908626565931e-07,
+      "loss": 0.8511,
+      "step": 2434
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.242458032904311e-07,
+      "loss": 0.8418,
+      "step": 2435
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.194127979112865e-07,
+      "loss": 0.8462,
+      "step": 2436
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.145918529381314e-07,
+      "loss": 0.8169,
+      "step": 2437
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.097829747739185e-07,
+      "loss": 0.8184,
+      "step": 2438
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.049861698055696e-07,
+      "loss": 0.8213,
+      "step": 2439
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.00201444403972e-07,
+      "loss": 0.8652,
+      "step": 2440
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.954288049239734e-07,
+      "loss": 0.8467,
+      "step": 2441
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.90668257704369e-07,
+      "loss": 0.8154,
+      "step": 2442
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.859198090678923e-07,
+      "loss": 0.8467,
+      "step": 2443
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.811834653212037e-07,
+      "loss": 0.8457,
+      "step": 2444
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.764592327548948e-07,
+      "loss": 0.8228,
+      "step": 2445
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.717471176434655e-07,
+      "loss": 0.8472,
+      "step": 2446
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.670471262453251e-07,
+      "loss": 0.8203,
+      "step": 2447
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.623592648027812e-07,
+      "loss": 0.8594,
+      "step": 2448
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.57683539542028e-07,
+      "loss": 0.8477,
+      "step": 2449
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.5301995667314e-07,
+      "loss": 0.835,
+      "step": 2450
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.483685223900706e-07,
+      "loss": 0.3289,
+      "step": 2451
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.437292428706312e-07,
+      "loss": 0.8271,
+      "step": 2452
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.391021242764962e-07,
+      "loss": 0.8174,
+      "step": 2453
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.344871727531856e-07,
+      "loss": 0.8252,
+      "step": 2454
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.298843944300583e-07,
+      "loss": 0.8369,
+      "step": 2455
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.252937954203055e-07,
+      "loss": 0.835,
+      "step": 2456
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.207153818209446e-07,
+      "loss": 0.8408,
+      "step": 2457
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.161491597128091e-07,
+      "loss": 0.8257,
+      "step": 2458
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.115951351605378e-07,
+      "loss": 0.8379,
+      "step": 2459
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.07053314212568e-07,
+      "loss": 0.8389,
+      "step": 2460
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.025237029011368e-07,
+      "loss": 0.7964,
+      "step": 2461
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.980063072422528e-07,
+      "loss": 0.8545,
+      "step": 2462
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.935011332357113e-07,
+      "loss": 0.8179,
+      "step": 2463
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.890081868650723e-07,
+      "loss": 0.8369,
+      "step": 2464
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.845274740976527e-07,
+      "loss": 0.3173,
+      "step": 2465
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.80059000884521e-07,
+      "loss": 0.832,
+      "step": 2466
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.756027731604943e-07,
+      "loss": 0.8418,
+      "step": 2467
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.711587968441225e-07,
+      "loss": 0.8491,
+      "step": 2468
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.667270778376834e-07,
+      "loss": 0.8408,
+      "step": 2469
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.623076220271797e-07,
+      "loss": 0.8262,
+      "step": 2470
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.579004352823205e-07,
+      "loss": 0.8594,
+      "step": 2471
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.535055234565225e-07,
+      "loss": 0.8618,
+      "step": 2472
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.491228923868999e-07,
+      "loss": 0.8242,
+      "step": 2473
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.447525478942585e-07,
+      "loss": 0.8428,
+      "step": 2474
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.4039449578308e-07,
+      "loss": 0.8354,
+      "step": 2475
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.360487418415218e-07,
+      "loss": 0.3071,
+      "step": 2476
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.317152918414116e-07,
+      "loss": 0.8408,
+      "step": 2477
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.273941515382299e-07,
+      "loss": 0.8638,
+      "step": 2478
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.230853266711124e-07,
+      "loss": 0.8223,
+      "step": 2479
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.187888229628371e-07,
+      "loss": 0.8682,
+      "step": 2480
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.145046461198146e-07,
+      "loss": 0.853,
+      "step": 2481
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.102328018320859e-07,
+      "loss": 0.8516,
+      "step": 2482
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.059732957733145e-07,
+      "loss": 0.8745,
+      "step": 2483
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.017261336007719e-07,
+      "loss": 0.302,
+      "step": 2484
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.97491320955338e-07,
+      "loss": 0.8545,
+      "step": 2485
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.932688634614948e-07,
+      "loss": 0.8076,
+      "step": 2486
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.890587667273064e-07,
+      "loss": 0.8198,
+      "step": 2487
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.84861036344423e-07,
+      "loss": 0.8184,
+      "step": 2488
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.806756778880752e-07,
+      "loss": 0.814,
+      "step": 2489
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.76502696917053e-07,
+      "loss": 0.8447,
+      "step": 2490
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.723420989737184e-07,
+      "loss": 0.8164,
+      "step": 2491
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.681938895839746e-07,
+      "loss": 0.8389,
+      "step": 2492
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.64058074257281e-07,
+      "loss": 0.8242,
+      "step": 2493
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.599346584866284e-07,
+      "loss": 0.8218,
+      "step": 2494
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.558236477485458e-07,
+      "loss": 0.8125,
+      "step": 2495
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.517250475030812e-07,
+      "loss": 0.8369,
+      "step": 2496
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.476388631938024e-07,
+      "loss": 0.8408,
+      "step": 2497
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.435651002477838e-07,
+      "loss": 0.8047,
+      "step": 2498
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.395037640756074e-07,
+      "loss": 0.3184,
+      "step": 2499
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.354548600713439e-07,
+      "loss": 0.8472,
+      "step": 2500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.314183936125584e-07,
+      "loss": 0.8457,
+      "step": 2501
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.273943700602958e-07,
+      "loss": 0.8418,
+      "step": 2502
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.233827947590709e-07,
+      "loss": 0.3394,
+      "step": 2503
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.193836730368697e-07,
+      "loss": 0.8164,
+      "step": 2504
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.153970102051354e-07,
+      "loss": 0.8574,
+      "step": 2505
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.11422811558765e-07,
+      "loss": 0.834,
+      "step": 2506
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.074610823761029e-07,
+      "loss": 0.8052,
+      "step": 2507
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.035118279189267e-07,
+      "loss": 0.8418,
+      "step": 2508
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.995750534324518e-07,
+      "loss": 0.8164,
+      "step": 2509
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.956507641453146e-07,
+      "loss": 0.8301,
+      "step": 2510
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.917389652695693e-07,
+      "loss": 0.8306,
+      "step": 2511
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.878396620006865e-07,
+      "loss": 0.8276,
+      "step": 2512
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.839528595175314e-07,
+      "loss": 0.8462,
+      "step": 2513
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.800785629823724e-07,
+      "loss": 0.3243,
+      "step": 2514
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.762167775408678e-07,
+      "loss": 0.854,
+      "step": 2515
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.723675083220581e-07,
+      "loss": 0.8169,
+      "step": 2516
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.685307604383561e-07,
+      "loss": 0.8418,
+      "step": 2517
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.647065389855544e-07,
+      "loss": 0.8281,
+      "step": 2518
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.608948490428023e-07,
+      "loss": 0.3018,
+      "step": 2519
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.570956956726014e-07,
+      "loss": 0.8164,
+      "step": 2520
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.533090839208133e-07,
+      "loss": 0.8359,
+      "step": 2521
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.495350188166304e-07,
+      "loss": 0.835,
+      "step": 2522
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.45773505372591e-07,
+      "loss": 0.8135,
+      "step": 2523
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.420245485845598e-07,
+      "loss": 0.8257,
+      "step": 2524
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.382881534317231e-07,
+      "loss": 0.8647,
+      "step": 2525
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.345643248765809e-07,
+      "loss": 0.8218,
+      "step": 2526
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.308530678649504e-07,
+      "loss": 0.8394,
+      "step": 2527
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.271543873259433e-07,
+      "loss": 0.8457,
+      "step": 2528
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.234682881719766e-07,
+      "loss": 0.3374,
+      "step": 2529
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.197947752987475e-07,
+      "loss": 0.8232,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.161338535852467e-07,
+      "loss": 0.8618,
+      "step": 2531
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.124855278937358e-07,
+      "loss": 0.8628,
+      "step": 2532
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.08849803069743e-07,
+      "loss": 0.855,
+      "step": 2533
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.052266839420738e-07,
+      "loss": 0.8408,
+      "step": 2534
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.016161753227799e-07,
+      "loss": 0.8721,
+      "step": 2535
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.980182820071666e-07,
+      "loss": 0.8252,
+      "step": 2536
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.944330087737881e-07,
+      "loss": 0.8164,
+      "step": 2537
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.90860360384432e-07,
+      "loss": 0.2892,
+      "step": 2538
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.873003415841215e-07,
+      "loss": 0.854,
+      "step": 2539
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.837529571011068e-07,
+      "loss": 0.8662,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.802182116468556e-07,
+      "loss": 0.8257,
+      "step": 2541
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.766961099160461e-07,
+      "loss": 0.8281,
+      "step": 2542
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.731866565865717e-07,
+      "loss": 0.8408,
+      "step": 2543
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.696898563195185e-07,
+      "loss": 0.7983,
+      "step": 2544
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.6620571375917356e-07,
+      "loss": 0.8071,
+      "step": 2545
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.6273423353300873e-07,
+      "loss": 0.8574,
+      "step": 2546
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.5927542025168025e-07,
+      "loss": 0.8579,
+      "step": 2547
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.5582927850902016e-07,
+      "loss": 0.3223,
+      "step": 2548
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.523958128820283e-07,
+      "loss": 0.8223,
+      "step": 2549
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.4897502793087576e-07,
+      "loss": 0.8037,
+      "step": 2550
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.4556692819888504e-07,
+      "loss": 0.3086,
+      "step": 2551
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.421715182125308e-07,
+      "loss": 0.8105,
+      "step": 2552
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.3878880248143904e-07,
+      "loss": 0.8472,
+      "step": 2553
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.354187854983716e-07,
+      "loss": 0.811,
+      "step": 2554
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.3206147173922133e-07,
+      "loss": 0.8228,
+      "step": 2555
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.287168656630203e-07,
+      "loss": 0.8442,
+      "step": 2556
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.25384971711913e-07,
+      "loss": 0.8101,
+      "step": 2557
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.220657943111617e-07,
+      "loss": 0.8262,
+      "step": 2558
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.187593378691435e-07,
+      "loss": 0.8755,
+      "step": 2559
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1546560677733436e-07,
+      "loss": 0.8335,
+      "step": 2560
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1218460541031404e-07,
+      "loss": 0.8652,
+      "step": 2561
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0891633812575214e-07,
+      "loss": 0.834,
+      "step": 2562
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0566080926440765e-07,
+      "loss": 0.2931,
+      "step": 2563
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.024180231501185e-07,
+      "loss": 0.8394,
+      "step": 2564
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.991879840897994e-07,
+      "loss": 0.8545,
+      "step": 2565
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.959706963734389e-07,
+      "loss": 0.3104,
+      "step": 2566
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.9276616427408207e-07,
+      "loss": 0.8325,
+      "step": 2567
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.895743920478379e-07,
+      "loss": 0.8301,
+      "step": 2568
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.8639538393386854e-07,
+      "loss": 0.8213,
+      "step": 2569
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.8322914415438006e-07,
+      "loss": 0.8125,
+      "step": 2570
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.8007567691462187e-07,
+      "loss": 0.8364,
+      "step": 2571
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.769349864028826e-07,
+      "loss": 0.7969,
+      "step": 2572
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.738070767904778e-07,
+      "loss": 0.8379,
+      "step": 2573
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.706919522317476e-07,
+      "loss": 0.8218,
+      "step": 2574
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.675896168640536e-07,
+      "loss": 0.8208,
+      "step": 2575
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.645000748077709e-07,
+      "loss": 0.8428,
+      "step": 2576
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.6142333016628286e-07,
+      "loss": 0.812,
+      "step": 2577
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.5835938702597605e-07,
+      "loss": 0.8311,
+      "step": 2578
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.553082494562354e-07,
+      "loss": 0.835,
+      "step": 2579
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.5226992150943494e-07,
+      "loss": 0.8145,
+      "step": 2580
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.492444072209411e-07,
+      "loss": 0.853,
+      "step": 2581
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.462317106090962e-07,
+      "loss": 0.8105,
+      "step": 2582
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.43231835675224e-07,
+      "loss": 0.2976,
+      "step": 2583
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4024478640361404e-07,
+      "loss": 0.8149,
+      "step": 2584
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.37270566761525e-07,
+      "loss": 0.8438,
+      "step": 2585
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.3430918069917384e-07,
+      "loss": 0.8521,
+      "step": 2586
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.313606321497309e-07,
+      "loss": 0.2815,
+      "step": 2587
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.2842492502932475e-07,
+      "loss": 0.8281,
+      "step": 2588
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.255020632370176e-07,
+      "loss": 0.8403,
+      "step": 2589
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.2259205065481857e-07,
+      "loss": 0.8291,
+      "step": 2590
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1969489114766715e-07,
+      "loss": 0.8555,
+      "step": 2591
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1681058856343314e-07,
+      "loss": 0.8291,
+      "step": 2592
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1393914673291335e-07,
+      "loss": 0.8223,
+      "step": 2593
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1108056946981826e-07,
+      "loss": 0.812,
+      "step": 2594
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.082348605707752e-07,
+      "loss": 0.8247,
+      "step": 2595
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.0540202381532104e-07,
+      "loss": 0.8389,
+      "step": 2596
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.0258206296589487e-07,
+      "loss": 0.8379,
+      "step": 2597
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.9977498176783416e-07,
+      "loss": 0.8369,
+      "step": 2598
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.9698078394937325e-07,
+      "loss": 0.8213,
+      "step": 2599
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.9419947322163133e-07,
+      "loss": 0.8345,
+      "step": 2600
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.914310532786158e-07,
+      "loss": 0.8169,
+      "step": 2601
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.886755277972109e-07,
+      "loss": 0.833,
+      "step": 2602
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.859329004371703e-07,
+      "loss": 0.8184,
+      "step": 2603
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8320317484112793e-07,
+      "loss": 0.8286,
+      "step": 2604
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8048635463457485e-07,
+      "loss": 0.334,
+      "step": 2605
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.777824434258591e-07,
+      "loss": 0.811,
+      "step": 2606
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.750914448061925e-07,
+      "loss": 0.7783,
+      "step": 2607
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.7241336234962943e-07,
+      "loss": 0.8325,
+      "step": 2608
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.697481996130713e-07,
+      "loss": 0.855,
+      "step": 2609
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.6709596013626326e-07,
+      "loss": 0.833,
+      "step": 2610
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.644566474417831e-07,
+      "loss": 0.3235,
+      "step": 2611
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.618302650350413e-07,
+      "loss": 0.8086,
+      "step": 2612
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.592168164042741e-07,
+      "loss": 0.8403,
+      "step": 2613
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.5661630502054057e-07,
+      "loss": 0.814,
+      "step": 2614
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.5402873433771793e-07,
+      "loss": 0.8398,
+      "step": 2615
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.5145410779249256e-07,
+      "loss": 0.812,
+      "step": 2616
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.488924288043648e-07,
+      "loss": 0.8198,
+      "step": 2617
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.463437007756331e-07,
+      "loss": 0.834,
+      "step": 2618
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.4380792709139513e-07,
+      "loss": 0.2955,
+      "step": 2619
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.4128511111955135e-07,
+      "loss": 0.8457,
+      "step": 2620
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.387752562107826e-07,
+      "loss": 0.8433,
+      "step": 2621
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3627836569855789e-07,
+      "loss": 0.8418,
+      "step": 2622
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3379444289913344e-07,
+      "loss": 0.8657,
+      "step": 2623
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3132349111153584e-07,
+      "loss": 0.8306,
+      "step": 2624
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.2886551361756326e-07,
+      "loss": 0.3546,
+      "step": 2625
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.2642051368178986e-07,
+      "loss": 0.8076,
+      "step": 2626
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.2398849455154693e-07,
+      "loss": 0.8262,
+      "step": 2627
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.215694594569251e-07,
+      "loss": 0.8379,
+      "step": 2628
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1916341161077547e-07,
+      "loss": 0.8486,
+      "step": 2629
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.167703542086952e-07,
+      "loss": 0.8169,
+      "step": 2630
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.143902904290296e-07,
+      "loss": 0.8599,
+      "step": 2631
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1202322343286786e-07,
+      "loss": 0.3279,
+      "step": 2632
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.0966915636403518e-07,
+      "loss": 0.8296,
+      "step": 2633
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.073280923490928e-07,
+      "loss": 0.8506,
+      "step": 2634
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.050000344973302e-07,
+      "loss": 0.813,
+      "step": 2635
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.026849859007629e-07,
+      "loss": 0.8276,
+      "step": 2636
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.0038294963413251e-07,
+      "loss": 0.8555,
+      "step": 2637
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9809392875488998e-07,
+      "loss": 0.8208,
+      "step": 2638
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9581792630320784e-07,
+      "loss": 0.8647,
+      "step": 2639
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9355494530196362e-07,
+      "loss": 0.8535,
+      "step": 2640
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9130498875673975e-07,
+      "loss": 0.8252,
+      "step": 2641
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8906805965582586e-07,
+      "loss": 0.8354,
+      "step": 2642
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8684416097020318e-07,
+      "loss": 0.8096,
+      "step": 2643
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8463329565354682e-07,
+      "loss": 0.8369,
+      "step": 2644
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.824354666422268e-07,
+      "loss": 0.8413,
+      "step": 2645
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8025067685529362e-07,
+      "loss": 0.8403,
+      "step": 2646
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7807892919448178e-07,
+      "loss": 0.8145,
+      "step": 2647
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7592022654420282e-07,
+      "loss": 0.855,
+      "step": 2648
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7377457177154554e-07,
+      "loss": 0.8467,
+      "step": 2649
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7164196772626596e-07,
+      "loss": 0.7988,
+      "step": 2650
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.6952241724078723e-07,
+      "loss": 0.8491,
+      "step": 2651
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.6741592313019638e-07,
+      "loss": 0.8257,
+      "step": 2652
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.6532248819223995e-07,
+      "loss": 0.8052,
+      "step": 2653
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.6324211520731824e-07,
+      "loss": 0.8506,
+      "step": 2654
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.6117480693848442e-07,
+      "loss": 0.3062,
+      "step": 2655
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.591205661314399e-07,
+      "loss": 0.8188,
+      "step": 2656
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5707939551452778e-07,
+      "loss": 0.8408,
+      "step": 2657
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.550512977987373e-07,
+      "loss": 0.8599,
+      "step": 2658
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5303627567769043e-07,
+      "loss": 0.8623,
+      "step": 2659
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5103433182764416e-07,
+      "loss": 0.8262,
+      "step": 2660
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4904546890748606e-07,
+      "loss": 0.8345,
+      "step": 2661
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4706968955872868e-07,
+      "loss": 0.3093,
+      "step": 2662
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4510699640550852e-07,
+      "loss": 0.856,
+      "step": 2663
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4315739205458257e-07,
+      "loss": 0.8403,
+      "step": 2664
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.412208790953229e-07,
+      "loss": 0.8135,
+      "step": 2665
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3929746009971434e-07,
+      "loss": 0.8657,
+      "step": 2666
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3738713762235124e-07,
+      "loss": 0.303,
+      "step": 2667
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.354899142004329e-07,
+      "loss": 0.8164,
+      "step": 2668
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3360579235376148e-07,
+      "loss": 0.8159,
+      "step": 2669
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.317347745847386e-07,
+      "loss": 0.79,
+      "step": 2670
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2987686337836202e-07,
+      "loss": 0.8223,
+      "step": 2671
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2803206120222123e-07,
+      "loss": 0.8564,
+      "step": 2672
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2620037050649404e-07,
+      "loss": 0.7759,
+      "step": 2673
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2438179372394553e-07,
+      "loss": 0.835,
+      "step": 2674
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.225763332699259e-07,
+      "loss": 0.8154,
+      "step": 2675
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.207839915423592e-07,
+      "loss": 0.834,
+      "step": 2676
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.190047709217501e-07,
+      "loss": 0.8389,
+      "step": 2677
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1723867377117503e-07,
+      "loss": 0.853,
+      "step": 2678
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1548570243627988e-07,
+      "loss": 0.8418,
+      "step": 2679
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1374585924528004e-07,
+      "loss": 0.8467,
+      "step": 2680
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1201914650895152e-07,
+      "loss": 0.8164,
+      "step": 2681
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1030556652063095e-07,
+      "loss": 0.8501,
+      "step": 2682
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.0860512155621783e-07,
+      "loss": 0.2771,
+      "step": 2683
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.0691781387416e-07,
+      "loss": 0.8403,
+      "step": 2684
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0524364571546042e-07,
+      "loss": 0.8032,
+      "step": 2685
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0358261930366931e-07,
+      "loss": 0.8213,
+      "step": 2686
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.019347368448842e-07,
+      "loss": 0.8496,
+      "step": 2687
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0030000052774436e-07,
+      "loss": 0.8018,
+      "step": 2688
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.867841252342747e-08,
+      "loss": 0.8413,
+      "step": 2689
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.706997498565295e-08,
+      "loss": 0.8345,
+      "step": 2690
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.547469005066979e-08,
+      "loss": 0.3064,
+      "step": 2691
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.389255983725864e-08,
+      "loss": 0.833,
+      "step": 2692
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.232358644673311e-08,
+      "loss": 0.8418,
+      "step": 2693
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.07677719629263e-08,
+      "loss": 0.8677,
+      "step": 2694
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.922511845219972e-08,
+      "loss": 0.854,
+      "step": 2695
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.769562796343001e-08,
+      "loss": 0.8335,
+      "step": 2696
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.617930252801665e-08,
+      "loss": 0.8374,
+      "step": 2697
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.46761441598698e-08,
+      "loss": 0.8286,
+      "step": 2698
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.31861548554147e-08,
+      "loss": 0.8403,
+      "step": 2699
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.170933659358282e-08,
+      "loss": 0.8457,
+      "step": 2700
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.024569133581517e-08,
+      "loss": 0.811,
+      "step": 2701
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.879522102605564e-08,
+      "loss": 0.8472,
+      "step": 2702
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.7357927590751e-08,
+      "loss": 0.8438,
+      "step": 2703
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.593381293884538e-08,
+      "loss": 0.8247,
+      "step": 2704
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.452287896178134e-08,
+      "loss": 0.8096,
+      "step": 2705
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.312512753349322e-08,
+      "loss": 0.8384,
+      "step": 2706
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.174056051041045e-08,
+      "loss": 0.856,
+      "step": 2707
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.036917973144764e-08,
+      "loss": 0.8125,
+      "step": 2708
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.901098701800779e-08,
+      "loss": 0.321,
+      "step": 2709
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.766598417398018e-08,
+      "loss": 0.8467,
+      "step": 2710
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.633417298573142e-08,
+      "loss": 0.8457,
+      "step": 2711
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.50155552221099e-08,
+      "loss": 0.8071,
+      "step": 2712
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.371013263444469e-08,
+      "loss": 0.8472,
+      "step": 2713
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.241790695653227e-08,
+      "loss": 0.3169,
+      "step": 2714
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.113887990464862e-08,
+      "loss": 0.8374,
+      "step": 2715
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.9873053177536e-08,
+      "loss": 0.832,
+      "step": 2716
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.862042845640403e-08,
+      "loss": 0.8481,
+      "step": 2717
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.73810074049308e-08,
+      "loss": 0.8291,
+      "step": 2718
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.61547916692573e-08,
+      "loss": 0.8359,
+      "step": 2719
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.494178287798524e-08,
+      "loss": 0.8301,
+      "step": 2720
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.3741982642173675e-08,
+      "loss": 0.8271,
+      "step": 2721
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.255539255534126e-08,
+      "loss": 0.7317,
+      "step": 2722
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.1382014193461783e-08,
+      "loss": 0.8574,
+      "step": 2723
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.022184911495864e-08,
+      "loss": 0.8394,
+      "step": 2724
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.9074898860711485e-08,
+      "loss": 0.8525,
+      "step": 2725
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.7941164954042886e-08,
+      "loss": 0.8271,
+      "step": 2726
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.6820648900725016e-08,
+      "loss": 0.8428,
+      "step": 2727
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.57133521889741e-08,
+      "loss": 0.3162,
+      "step": 2728
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.4619276289450394e-08,
+      "loss": 0.8423,
+      "step": 2729
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.3538422655251544e-08,
+      "loss": 0.8252,
+      "step": 2730
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.247079272191812e-08,
+      "loss": 0.8281,
+      "step": 2731
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.141638790742475e-08,
+      "loss": 0.8057,
+      "step": 2732
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.037520961218233e-08,
+      "loss": 0.8364,
+      "step": 2733
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.934725921903471e-08,
+      "loss": 0.8081,
+      "step": 2734
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.833253809325643e-08,
+      "loss": 0.8301,
+      "step": 2735
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.7331047582551684e-08,
+      "loss": 0.8398,
+      "step": 2736
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.634278901705424e-08,
+      "loss": 0.3025,
+      "step": 2737
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.5367763709320825e-08,
+      "loss": 0.8462,
+      "step": 2738
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.4405972954334454e-08,
+      "loss": 0.8169,
+      "step": 2739
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.3457418029499974e-08,
+      "loss": 0.7891,
+      "step": 2740
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.252210019464408e-08,
+      "loss": 0.8066,
+      "step": 2741
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.160002069201085e-08,
+      "loss": 0.8403,
+      "step": 2742
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.069118074626176e-08,
+      "loss": 0.8325,
+      "step": 2743
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.9795581564475706e-08,
+      "loss": 0.8257,
+      "step": 2744
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.8913224336145628e-08,
+      "loss": 0.8442,
+      "step": 2745
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.8044110233176323e-08,
+      "loss": 0.835,
+      "step": 2746
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.7188240409883325e-08,
+      "loss": 0.8169,
+      "step": 2747
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.6345616002994013e-08,
+      "loss": 0.8252,
+      "step": 2748
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.5516238131640945e-08,
+      "loss": 0.8384,
+      "step": 2749
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.4700107897364102e-08,
+      "loss": 0.8398,
+      "step": 2750
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.389722638410974e-08,
+      "loss": 0.8413,
+      "step": 2751
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.3107594658228205e-08,
+      "loss": 0.8203,
+      "step": 2752
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.2331213768468363e-08,
+      "loss": 0.8296,
+      "step": 2753
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.1568084745985373e-08,
+      "loss": 0.8096,
+      "step": 2754
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.0818208604328482e-08,
+      "loss": 0.2881,
+      "step": 2755
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.008158633944879e-08,
+      "loss": 0.8398,
+      "step": 2756
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.9358218929693695e-08,
+      "loss": 0.835,
+      "step": 2757
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.86481073358058e-08,
+      "loss": 0.8228,
+      "step": 2758
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.7951252500920668e-08,
+      "loss": 0.8218,
+      "step": 2759
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.7267655350566848e-08,
+      "loss": 0.814,
+      "step": 2760
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.659731679266807e-08,
+      "loss": 0.2866,
+      "step": 2761
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.59402377175355e-08,
+      "loss": 0.8101,
+      "step": 2762
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.5296418997869932e-08,
+      "loss": 0.8203,
+      "step": 2763
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.4665861488761813e-08,
+      "loss": 0.8218,
+      "step": 2764
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.4048566027685673e-08,
+      "loss": 0.814,
+      "step": 2765
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3444533434506801e-08,
+      "loss": 0.811,
+      "step": 2766
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.2853764511471245e-08,
+      "loss": 0.3325,
+      "step": 2767
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.2276260043210253e-08,
+      "loss": 0.8306,
+      "step": 2768
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.1712020796738056e-08,
+      "loss": 0.7998,
+      "step": 2769
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.1161047521451862e-08,
+      "loss": 0.7983,
+      "step": 2770
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0623340949125204e-08,
+      "loss": 0.8403,
+      "step": 2771
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0098901793917926e-08,
+      "loss": 0.8281,
+      "step": 2772
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.587730752362855e-09,
+      "loss": 0.2878,
+      "step": 2773
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.089828503374697e-09,
+      "loss": 0.8354,
+      "step": 2774
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 8.605195708242254e-09,
+      "loss": 0.8677,
+      "step": 2775
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 8.133833010632864e-09,
+      "loss": 0.8257,
+      "step": 2776
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.67574103658797e-09,
+      "loss": 0.8213,
+      "step": 2777
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.230920394524221e-09,
+      "loss": 0.8271,
+      "step": 2778
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.799371675230149e-09,
+      "loss": 0.8276,
+      "step": 2779
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.38109545186838e-09,
+      "loss": 0.3162,
+      "step": 2780
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.976092279974533e-09,
+      "loss": 0.3257,
+      "step": 2781
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.584362697453882e-09,
+      "loss": 0.8525,
+      "step": 2782
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.205907224583584e-09,
+      "loss": 0.8057,
+      "step": 2783
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.840726364010451e-09,
+      "loss": 0.8159,
+      "step": 2784
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.488820600749844e-09,
+      "loss": 0.7939,
+      "step": 2785
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.150190402185672e-09,
+      "loss": 0.8286,
+      "step": 2786
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.824836218072614e-09,
+      "loss": 0.8657,
+      "step": 2787
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.5127584805283443e-09,
+      "loss": 0.8223,
+      "step": 2788
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.2139576040413067e-09,
+      "loss": 0.8423,
+      "step": 2789
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.9284339854651624e-09,
+      "loss": 0.2948,
+      "step": 2790
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.656188004016569e-09,
+      "loss": 0.8232,
+      "step": 2791
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.397220021281843e-09,
+      "loss": 0.8481,
+      "step": 2792
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.1515303812091883e-09,
+      "loss": 0.3164,
+      "step": 2793
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.919119410113135e-09,
+      "loss": 0.8037,
+      "step": 2794
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.6999874166678809e-09,
+      "loss": 0.8101,
+      "step": 2795
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.4941346919172817e-09,
+      "loss": 0.8623,
+      "step": 2796
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.301561509263749e-09,
+      "loss": 0.8296,
+      "step": 2797
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.1222681244749123e-09,
+      "loss": 0.8286,
+      "step": 2798
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 9.562547756780672e-10,
+      "loss": 0.8413,
+      "step": 2799
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 8.035216833657266e-10,
+      "loss": 0.8555,
+      "step": 2800
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 6.640690503889601e-10,
+      "loss": 0.8408,
+      "step": 2801
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.378970619651646e-10,
+      "loss": 0.8291,
+      "step": 2802
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.2500588566696254e-10,
+      "loss": 0.8521,
+      "step": 2803
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.2539567143330396e-10,
+      "loss": 0.8633,
+      "step": 2804
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.39066551560585e-10,
+      "loss": 0.3044,
+      "step": 2805
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.6601864070708852e-10,
+      "loss": 0.8525,
+      "step": 2806
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0625203589187394e-10,
+      "loss": 0.345,
+      "step": 2807
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.976681649366712e-11,
+      "loss": 0.8062,
+      "step": 2808
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.656304425308065e-11,
+      "loss": 0.8374,
+      "step": 2809
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 6.64076326817309e-12,
+      "loss": 0.8149,
+      "step": 2810
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0,
+      "loss": 0.3982,
+      "step": 2811
+    },
+    {
+      "epoch": 1.0,
+      "step": 2811,
+      "total_flos": 714224002678784.0,
+      "train_loss": 0.8310348418462069,
+      "train_runtime": 10495.8902,
+      "train_samples_per_second": 34.278,
+      "train_steps_per_second": 0.268
+    }
+  ],
+  "max_steps": 2811,
+  "num_train_epochs": 1,
+  "total_flos": 714224002678784.0,
+  "trial_name": null,
+  "trial_params": null
+}