diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,16717 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 12.0,
+  "global_step": 2760,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.998188405797101e-07,
+      "loss": 0.3554,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.996376811594203e-07,
+      "loss": 0.2575,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.994565217391304e-07,
+      "loss": 0.0431,
+      "step": 3
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.992753623188406e-07,
+      "loss": 0.2741,
+      "step": 4
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.990942028985507e-07,
+      "loss": 0.3406,
+      "step": 5
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.989130434782608e-07,
+      "loss": 0.5719,
+      "step": 6
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.98731884057971e-07,
+      "loss": 0.3005,
+      "step": 7
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.985507246376812e-07,
+      "loss": 0.152,
+      "step": 8
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.983695652173913e-07,
+      "loss": 0.0841,
+      "step": 9
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.981884057971014e-07,
+      "loss": 0.6203,
+      "step": 10
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.980072463768116e-07,
+      "loss": 0.3752,
+      "step": 11
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.978260869565217e-07,
+      "loss": 0.1462,
+      "step": 12
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.976449275362319e-07,
+      "loss": 0.2784,
+      "step": 13
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.974637681159419e-07,
+      "loss": 0.4657,
+      "step": 14
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.972826086956522e-07,
+      "loss": 0.02,
+      "step": 15
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.971014492753623e-07,
+      "loss": 0.4022,
+      "step": 16
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.969202898550724e-07,
+      "loss": 0.4789,
+      "step": 17
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.967391304347827e-07,
+      "loss": 0.0424,
+      "step": 18
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.965579710144927e-07,
+      "loss": 0.6924,
+      "step": 19
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.963768115942029e-07,
+      "loss": 0.0353,
+      "step": 20
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.96195652173913e-07,
+      "loss": 0.426,
+      "step": 21
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.960144927536232e-07,
+      "loss": 0.3873,
+      "step": 22
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.958333333333333e-07,
+      "loss": 0.2104,
+      "step": 23
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.956521739130435e-07,
+      "loss": 0.0368,
+      "step": 24
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.954710144927535e-07,
+      "loss": 0.0834,
+      "step": 25
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.952898550724638e-07,
+      "loss": 0.0539,
+      "step": 26
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.951086956521739e-07,
+      "loss": 0.081,
+      "step": 27
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.94927536231884e-07,
+      "loss": 0.3546,
+      "step": 28
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.947463768115942e-07,
+      "loss": 0.5031,
+      "step": 29
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.945652173913043e-07,
+      "loss": 0.0729,
+      "step": 30
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.943840579710145e-07,
+      "loss": 0.02,
+      "step": 31
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.942028985507246e-07,
+      "loss": 0.011,
+      "step": 32
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.940217391304347e-07,
+      "loss": 0.1114,
+      "step": 33
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.93840579710145e-07,
+      "loss": 0.1102,
+      "step": 34
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.936594202898551e-07,
+      "loss": 0.0762,
+      "step": 35
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.934782608695652e-07,
+      "loss": 0.0979,
+      "step": 36
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.932971014492754e-07,
+      "loss": 0.195,
+      "step": 37
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.931159420289855e-07,
+      "loss": 0.0311,
+      "step": 38
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.929347826086956e-07,
+      "loss": 0.125,
+      "step": 39
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.927536231884058e-07,
+      "loss": 0.2491,
+      "step": 40
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.925724637681159e-07,
+      "loss": 0.2129,
+      "step": 41
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.923913043478261e-07,
+      "loss": 0.2091,
+      "step": 42
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.922101449275362e-07,
+      "loss": 0.0648,
+      "step": 43
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.920289855072463e-07,
+      "loss": 0.0376,
+      "step": 44
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.918478260869566e-07,
+      "loss": 0.0426,
+      "step": 45
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.916666666666666e-07,
+      "loss": 0.157,
+      "step": 46
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.914855072463768e-07,
+      "loss": 0.2103,
+      "step": 47
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.913043478260869e-07,
+      "loss": 0.1734,
+      "step": 48
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.911231884057971e-07,
+      "loss": 0.1613,
+      "step": 49
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.909420289855072e-07,
+      "loss": 0.3365,
+      "step": 50
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.907608695652174e-07,
+      "loss": 0.6747,
+      "step": 51
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.905797101449275e-07,
+      "loss": 0.0683,
+      "step": 52
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.903985507246377e-07,
+      "loss": 0.3365,
+      "step": 53
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.902173913043478e-07,
+      "loss": 0.1388,
+      "step": 54
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.900362318840579e-07,
+      "loss": 0.0145,
+      "step": 55
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.898550724637682e-07,
+      "loss": 0.2425,
+      "step": 56
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.896739130434782e-07,
+      "loss": 0.0449,
+      "step": 57
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.894927536231884e-07,
+      "loss": 0.0357,
+      "step": 58
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.893115942028985e-07,
+      "loss": 0.1447,
+      "step": 59
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.891304347826087e-07,
+      "loss": 0.3243,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.889492753623189e-07,
+      "loss": 0.0161,
+      "step": 61
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.887681159420289e-07,
+      "loss": 0.146,
+      "step": 62
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.885869565217391e-07,
+      "loss": 0.256,
+      "step": 63
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.884057971014493e-07,
+      "loss": 0.1023,
+      "step": 64
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.882246376811594e-07,
+      "loss": 0.0589,
+      "step": 65
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.880434782608695e-07,
+      "loss": 0.1908,
+      "step": 66
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.878623188405797e-07,
+      "loss": 0.0435,
+      "step": 67
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.876811594202898e-07,
+      "loss": 0.1935,
+      "step": 68
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.875e-07,
+      "loss": 0.0147,
+      "step": 69
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.873188405797101e-07,
+      "loss": 0.2193,
+      "step": 70
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.871376811594202e-07,
+      "loss": 0.0467,
+      "step": 71
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.869565217391305e-07,
+      "loss": 0.1861,
+      "step": 72
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.867753623188405e-07,
+      "loss": 0.162,
+      "step": 73
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.865942028985507e-07,
+      "loss": 0.148,
+      "step": 74
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.864130434782608e-07,
+      "loss": 0.6446,
+      "step": 75
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.86231884057971e-07,
+      "loss": 0.5805,
+      "step": 76
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.860507246376812e-07,
+      "loss": 0.7285,
+      "step": 77
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.858695652173913e-07,
+      "loss": 0.507,
+      "step": 78
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.856884057971015e-07,
+      "loss": 0.0135,
+      "step": 79
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.855072463768116e-07,
+      "loss": 0.4134,
+      "step": 80
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.853260869565217e-07,
+      "loss": 0.4095,
+      "step": 81
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.851449275362318e-07,
+      "loss": 0.0491,
+      "step": 82
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.849637681159421e-07,
+      "loss": 0.0047,
+      "step": 83
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.847826086956521e-07,
+      "loss": 0.0098,
+      "step": 84
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.846014492753623e-07,
+      "loss": 0.1208,
+      "step": 85
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.844202898550724e-07,
+      "loss": 0.0254,
+      "step": 86
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.842391304347826e-07,
+      "loss": 0.1558,
+      "step": 87
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.840579710144928e-07,
+      "loss": 0.0122,
+      "step": 88
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.838768115942028e-07,
+      "loss": 0.0179,
+      "step": 89
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.83695652173913e-07,
+      "loss": 0.007,
+      "step": 90
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.835144927536232e-07,
+      "loss": 0.3568,
+      "step": 91
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.833333333333333e-07,
+      "loss": 0.2258,
+      "step": 92
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.831521739130434e-07,
+      "loss": 0.073,
+      "step": 93
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.829710144927537e-07,
+      "loss": 0.0057,
+      "step": 94
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.827898550724638e-07,
+      "loss": 0.1071,
+      "step": 95
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.826086956521739e-07,
+      "loss": 0.0125,
+      "step": 96
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.82427536231884e-07,
+      "loss": 0.1459,
+      "step": 97
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.822463768115942e-07,
+      "loss": 0.3828,
+      "step": 98
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.820652173913044e-07,
+      "loss": 0.1476,
+      "step": 99
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.818840579710144e-07,
+      "loss": 0.1975,
+      "step": 100
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.817028985507246e-07,
+      "loss": 0.0114,
+      "step": 101
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.815217391304348e-07,
+      "loss": 0.1646,
+      "step": 102
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.813405797101449e-07,
+      "loss": 0.1279,
+      "step": 103
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.811594202898551e-07,
+      "loss": 0.108,
+      "step": 104
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.809782608695652e-07,
+      "loss": 0.1946,
+      "step": 105
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.807971014492754e-07,
+      "loss": 0.2246,
+      "step": 106
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.806159420289855e-07,
+      "loss": 0.1337,
+      "step": 107
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.804347826086956e-07,
+      "loss": 0.0041,
+      "step": 108
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.802536231884057e-07,
+      "loss": 0.1564,
+      "step": 109
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.80072463768116e-07,
+      "loss": 0.0116,
+      "step": 110
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.79891304347826e-07,
+      "loss": 0.2226,
+      "step": 111
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.797101449275362e-07,
+      "loss": 0.182,
+      "step": 112
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.795289855072463e-07,
+      "loss": 0.0148,
+      "step": 113
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.793478260869565e-07,
+      "loss": 0.0535,
+      "step": 114
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.791666666666667e-07,
+      "loss": 0.1571,
+      "step": 115
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.789855072463767e-07,
+      "loss": 0.0095,
+      "step": 116
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.78804347826087e-07,
+      "loss": 0.0064,
+      "step": 117
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.786231884057971e-07,
+      "loss": 0.0697,
+      "step": 118
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.784420289855072e-07,
+      "loss": 0.0202,
+      "step": 119
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.782608695652174e-07,
+      "loss": 0.0306,
+      "step": 120
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.780797101449276e-07,
+      "loss": 0.0275,
+      "step": 121
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.778985507246377e-07,
+      "loss": 0.4697,
+      "step": 122
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.777173913043478e-07,
+      "loss": 0.0161,
+      "step": 123
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.775362318840579e-07,
+      "loss": 0.2296,
+      "step": 124
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.773550724637681e-07,
+      "loss": 0.2935,
+      "step": 125
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.771739130434783e-07,
+      "loss": 0.0263,
+      "step": 126
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.769927536231883e-07,
+      "loss": 0.0206,
+      "step": 127
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.7681159420289854e-07,
+      "loss": 0.0218,
+      "step": 128
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.7663043478260867e-07,
+      "loss": 0.0357,
+      "step": 129
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.7644927536231884e-07,
+      "loss": 0.0843,
+      "step": 130
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.7626811594202897e-07,
+      "loss": 0.2699,
+      "step": 131
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.7608695652173915e-07,
+      "loss": 0.0382,
+      "step": 132
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.759057971014492e-07,
+      "loss": 0.0764,
+      "step": 133
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.757246376811594e-07,
+      "loss": 0.0363,
+      "step": 134
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.755434782608695e-07,
+      "loss": 0.0493,
+      "step": 135
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.753623188405797e-07,
+      "loss": 0.221,
+      "step": 136
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.751811594202899e-07,
+      "loss": 0.476,
+      "step": 137
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.7499999999999995e-07,
+      "loss": 0.1724,
+      "step": 138
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.7481884057971013e-07,
+      "loss": 0.1706,
+      "step": 139
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.7463768115942026e-07,
+      "loss": 0.1496,
+      "step": 140
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.7445652173913043e-07,
+      "loss": 0.0468,
+      "step": 141
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.7427536231884056e-07,
+      "loss": 0.0888,
+      "step": 142
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.740942028985507e-07,
+      "loss": 0.0106,
+      "step": 143
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.739130434782608e-07,
+      "loss": 0.0375,
+      "step": 144
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.73731884057971e-07,
+      "loss": 0.009,
+      "step": 145
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.7355072463768117e-07,
+      "loss": 0.0559,
+      "step": 146
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.733695652173913e-07,
+      "loss": 0.2127,
+      "step": 147
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.7318840579710147e-07,
+      "loss": 0.046,
+      "step": 148
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.7300724637681154e-07,
+      "loss": 0.0491,
+      "step": 149
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.728260869565217e-07,
+      "loss": 0.0487,
+      "step": 150
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.7264492753623184e-07,
+      "loss": 0.0076,
+      "step": 151
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.72463768115942e-07,
+      "loss": 0.6671,
+      "step": 152
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.7228260869565215e-07,
+      "loss": 0.0069,
+      "step": 153
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.7210144927536227e-07,
+      "loss": 0.0077,
+      "step": 154
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.7192028985507245e-07,
+      "loss": 0.042,
+      "step": 155
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.717391304347826e-07,
+      "loss": 0.0188,
+      "step": 156
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.7155797101449275e-07,
+      "loss": 0.1749,
+      "step": 157
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.713768115942029e-07,
+      "loss": 0.0055,
+      "step": 158
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.7119565217391306e-07,
+      "loss": 0.0138,
+      "step": 159
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.7101449275362313e-07,
+      "loss": 0.0093,
+      "step": 160
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.708333333333333e-07,
+      "loss": 0.0304,
+      "step": 161
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.706521739130435e-07,
+      "loss": 0.0433,
+      "step": 162
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.704710144927536e-07,
+      "loss": 0.0243,
+      "step": 163
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.702898550724638e-07,
+      "loss": 0.4288,
+      "step": 164
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.7010869565217386e-07,
+      "loss": 0.0291,
+      "step": 165
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.6992753623188404e-07,
+      "loss": 0.0235,
+      "step": 166
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.6974637681159416e-07,
+      "loss": 0.0061,
+      "step": 167
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.6956521739130434e-07,
+      "loss": 0.3967,
+      "step": 168
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.6938405797101447e-07,
+      "loss": 0.0128,
+      "step": 169
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.692028985507246e-07,
+      "loss": 0.2921,
+      "step": 170
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.6902173913043477e-07,
+      "loss": 0.0167,
+      "step": 171
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.688405797101449e-07,
+      "loss": 0.0147,
+      "step": 172
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.686594202898551e-07,
+      "loss": 0.0916,
+      "step": 173
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.684782608695652e-07,
+      "loss": 0.1918,
+      "step": 174
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.682971014492754e-07,
+      "loss": 0.0147,
+      "step": 175
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.6811594202898545e-07,
+      "loss": 0.2914,
+      "step": 176
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.6793478260869563e-07,
+      "loss": 0.1225,
+      "step": 177
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.6775362318840575e-07,
+      "loss": 0.2643,
+      "step": 178
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.6757246376811593e-07,
+      "loss": 0.1541,
+      "step": 179
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.673913043478261e-07,
+      "loss": 0.0103,
+      "step": 180
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.672101449275362e-07,
+      "loss": 0.2637,
+      "step": 181
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.6702898550724636e-07,
+      "loss": 0.0171,
+      "step": 182
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.668478260869565e-07,
+      "loss": 0.01,
+      "step": 183
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.6666666666666666e-07,
+      "loss": 0.3712,
+      "step": 184
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.664855072463768e-07,
+      "loss": 0.0688,
+      "step": 185
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.663043478260869e-07,
+      "loss": 0.0053,
+      "step": 186
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.6612318840579704e-07,
+      "loss": 0.4034,
+      "step": 187
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.659420289855072e-07,
+      "loss": 0.0158,
+      "step": 188
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.657608695652174e-07,
+      "loss": 0.1276,
+      "step": 189
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.655797101449275e-07,
+      "loss": 0.0085,
+      "step": 190
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.653985507246377e-07,
+      "loss": 0.0351,
+      "step": 191
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.6521739130434777e-07,
+      "loss": 0.1687,
+      "step": 192
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.6503623188405795e-07,
+      "loss": 0.065,
+      "step": 193
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.648550724637681e-07,
+      "loss": 0.0263,
+      "step": 194
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.6467391304347825e-07,
+      "loss": 0.1883,
+      "step": 195
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.6449275362318843e-07,
+      "loss": 0.689,
+      "step": 196
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.643115942028985e-07,
+      "loss": 0.018,
+      "step": 197
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.641304347826087e-07,
+      "loss": 0.0402,
+      "step": 198
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.639492753623188e-07,
+      "loss": 0.2798,
+      "step": 199
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.63768115942029e-07,
+      "loss": 0.0137,
+      "step": 200
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.635869565217391e-07,
+      "loss": 0.0081,
+      "step": 201
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.634057971014493e-07,
+      "loss": 0.0034,
+      "step": 202
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.6322463768115936e-07,
+      "loss": 0.007,
+      "step": 203
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.6304347826086954e-07,
+      "loss": 0.1567,
+      "step": 204
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.628623188405797e-07,
+      "loss": 0.0852,
+      "step": 205
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.6268115942028984e-07,
+      "loss": 0.0355,
+      "step": 206
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.625e-07,
+      "loss": 0.0429,
+      "step": 207
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.623188405797101e-07,
+      "loss": 0.0041,
+      "step": 208
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.6213768115942027e-07,
+      "loss": 0.016,
+      "step": 209
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.619565217391304e-07,
+      "loss": 0.4588,
+      "step": 210
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.6177536231884057e-07,
+      "loss": 0.3083,
+      "step": 211
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.615942028985507e-07,
+      "loss": 0.0101,
+      "step": 212
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.614130434782608e-07,
+      "loss": 0.0094,
+      "step": 213
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.61231884057971e-07,
+      "loss": 0.2028,
+      "step": 214
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.610507246376811e-07,
+      "loss": 0.066,
+      "step": 215
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.608695652173913e-07,
+      "loss": 0.3119,
+      "step": 216
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.6068840579710143e-07,
+      "loss": 0.0223,
+      "step": 217
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.605072463768116e-07,
+      "loss": 0.3056,
+      "step": 218
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.603260869565217e-07,
+      "loss": 0.4894,
+      "step": 219
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.6014492753623186e-07,
+      "loss": 0.005,
+      "step": 220
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.59963768115942e-07,
+      "loss": 0.0038,
+      "step": 221
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.5978260869565216e-07,
+      "loss": 0.0286,
+      "step": 222
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.5960144927536234e-07,
+      "loss": 0.0815,
+      "step": 223
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.594202898550724e-07,
+      "loss": 0.009,
+      "step": 224
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.592391304347826e-07,
+      "loss": 0.1529,
+      "step": 225
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.590579710144927e-07,
+      "loss": 0.0336,
+      "step": 226
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.588768115942029e-07,
+      "loss": 0.7381,
+      "step": 227
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.58695652173913e-07,
+      "loss": 0.3864,
+      "step": 228
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.585144927536232e-07,
+      "loss": 0.0311,
+      "step": 229
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.5833333333333327e-07,
+      "loss": 0.0021,
+      "step": 230
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8848039215686274,
+      "eval_combined_score": 0.8997522330166186,
+      "eval_f1": 0.9147005444646098,
+      "eval_loss": 0.40174856781959534,
+      "eval_runtime": 10.1205,
+      "eval_samples_per_second": 40.314,
+      "eval_steps_per_second": 2.569,
+      "step": 230
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.5815217391304345e-07,
+      "loss": 0.3594,
+      "step": 231
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.579710144927536e-07,
+      "loss": 0.0081,
+      "step": 232
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.5778985507246375e-07,
+      "loss": 0.3366,
+      "step": 233
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.5760869565217393e-07,
+      "loss": 0.1448,
+      "step": 234
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.57427536231884e-07,
+      "loss": 0.0229,
+      "step": 235
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.572463768115942e-07,
+      "loss": 0.0116,
+      "step": 236
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.570652173913043e-07,
+      "loss": 0.0857,
+      "step": 237
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.568840579710145e-07,
+      "loss": 0.0344,
+      "step": 238
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.5670289855072466e-07,
+      "loss": 0.0076,
+      "step": 239
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.5652173913043473e-07,
+      "loss": 0.0104,
+      "step": 240
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.563405797101449e-07,
+      "loss": 0.0087,
+      "step": 241
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.5615942028985504e-07,
+      "loss": 0.0127,
+      "step": 242
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.559782608695652e-07,
+      "loss": 0.1234,
+      "step": 243
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.5579710144927534e-07,
+      "loss": 0.0043,
+      "step": 244
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.556159420289855e-07,
+      "loss": 0.2919,
+      "step": 245
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.554347826086956e-07,
+      "loss": 0.2864,
+      "step": 246
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.5525362318840577e-07,
+      "loss": 0.0039,
+      "step": 247
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.5507246376811595e-07,
+      "loss": 0.0076,
+      "step": 248
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.5489130434782607e-07,
+      "loss": 0.0127,
+      "step": 249
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.5471014492753625e-07,
+      "loss": 0.6405,
+      "step": 250
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.545289855072463e-07,
+      "loss": 0.1236,
+      "step": 251
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.543478260869565e-07,
+      "loss": 0.2507,
+      "step": 252
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.541666666666666e-07,
+      "loss": 0.0119,
+      "step": 253
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.539855072463768e-07,
+      "loss": 0.105,
+      "step": 254
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.5380434782608693e-07,
+      "loss": 0.2818,
+      "step": 255
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.536231884057971e-07,
+      "loss": 0.3385,
+      "step": 256
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.5344202898550723e-07,
+      "loss": 0.0178,
+      "step": 257
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.5326086956521736e-07,
+      "loss": 0.1041,
+      "step": 258
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.5307971014492753e-07,
+      "loss": 0.6599,
+      "step": 259
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.5289855072463766e-07,
+      "loss": 0.0071,
+      "step": 260
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.5271739130434784e-07,
+      "loss": 0.0425,
+      "step": 261
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.525362318840579e-07,
+      "loss": 0.5427,
+      "step": 262
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.523550724637681e-07,
+      "loss": 0.0113,
+      "step": 263
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.521739130434782e-07,
+      "loss": 0.0049,
+      "step": 264
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.519927536231884e-07,
+      "loss": 0.1725,
+      "step": 265
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.5181159420289857e-07,
+      "loss": 0.4347,
+      "step": 266
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.5163043478260864e-07,
+      "loss": 0.5354,
+      "step": 267
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.514492753623188e-07,
+      "loss": 0.2619,
+      "step": 268
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.5126811594202894e-07,
+      "loss": 0.8186,
+      "step": 269
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.510869565217391e-07,
+      "loss": 0.3548,
+      "step": 270
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.5090579710144925e-07,
+      "loss": 0.3907,
+      "step": 271
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.507246376811594e-07,
+      "loss": 0.2943,
+      "step": 272
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.5054347826086955e-07,
+      "loss": 0.1195,
+      "step": 273
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.503623188405797e-07,
+      "loss": 0.3433,
+      "step": 274
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.5018115942028985e-07,
+      "loss": 0.008,
+      "step": 275
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.5e-07,
+      "loss": 0.0039,
+      "step": 276
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.4981884057971016e-07,
+      "loss": 0.0892,
+      "step": 277
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.4963768115942023e-07,
+      "loss": 0.0211,
+      "step": 278
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.494565217391304e-07,
+      "loss": 0.0101,
+      "step": 279
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.4927536231884053e-07,
+      "loss": 0.0236,
+      "step": 280
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.490942028985507e-07,
+      "loss": 0.1325,
+      "step": 281
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.489130434782609e-07,
+      "loss": 0.0267,
+      "step": 282
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.48731884057971e-07,
+      "loss": 0.5038,
+      "step": 283
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.4855072463768114e-07,
+      "loss": 0.1644,
+      "step": 284
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.4836956521739127e-07,
+      "loss": 0.0044,
+      "step": 285
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.4818840579710144e-07,
+      "loss": 0.0861,
+      "step": 286
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.4800724637681157e-07,
+      "loss": 0.0193,
+      "step": 287
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.4782608695652175e-07,
+      "loss": 0.0095,
+      "step": 288
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.476449275362318e-07,
+      "loss": 0.0123,
+      "step": 289
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.47463768115942e-07,
+      "loss": 0.0216,
+      "step": 290
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.472826086956522e-07,
+      "loss": 0.0624,
+      "step": 291
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.471014492753623e-07,
+      "loss": 0.5781,
+      "step": 292
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.469202898550725e-07,
+      "loss": 0.0047,
+      "step": 293
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.4673913043478255e-07,
+      "loss": 0.1187,
+      "step": 294
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.4655797101449273e-07,
+      "loss": 0.0111,
+      "step": 295
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.4637681159420285e-07,
+      "loss": 0.0683,
+      "step": 296
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.4619565217391303e-07,
+      "loss": 0.0179,
+      "step": 297
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.460144927536232e-07,
+      "loss": 0.0195,
+      "step": 298
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.4583333333333334e-07,
+      "loss": 0.0101,
+      "step": 299
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.4565217391304346e-07,
+      "loss": 0.5054,
+      "step": 300
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.454710144927536e-07,
+      "loss": 0.0366,
+      "step": 301
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.4528985507246376e-07,
+      "loss": 0.0157,
+      "step": 302
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.451086956521739e-07,
+      "loss": 0.6046,
+      "step": 303
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.4492753623188407e-07,
+      "loss": 0.0162,
+      "step": 304
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.4474637681159414e-07,
+      "loss": 0.1077,
+      "step": 305
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.445652173913043e-07,
+      "loss": 0.011,
+      "step": 306
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.443840579710145e-07,
+      "loss": 0.0273,
+      "step": 307
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.442028985507246e-07,
+      "loss": 0.1484,
+      "step": 308
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.440217391304348e-07,
+      "loss": 0.0085,
+      "step": 309
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.4384057971014487e-07,
+      "loss": 0.3413,
+      "step": 310
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.4365942028985505e-07,
+      "loss": 0.0751,
+      "step": 311
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.434782608695652e-07,
+      "loss": 0.0817,
+      "step": 312
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.4329710144927535e-07,
+      "loss": 0.1593,
+      "step": 313
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.431159420289855e-07,
+      "loss": 0.5342,
+      "step": 314
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.4293478260869566e-07,
+      "loss": 0.105,
+      "step": 315
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.427536231884058e-07,
+      "loss": 0.0065,
+      "step": 316
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.425724637681159e-07,
+      "loss": 0.6542,
+      "step": 317
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.423913043478261e-07,
+      "loss": 0.0304,
+      "step": 318
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.422101449275362e-07,
+      "loss": 0.0147,
+      "step": 319
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.420289855072464e-07,
+      "loss": 0.1346,
+      "step": 320
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.4184782608695646e-07,
+      "loss": 0.0638,
+      "step": 321
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.4166666666666664e-07,
+      "loss": 0.0735,
+      "step": 322
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.4148550724637676e-07,
+      "loss": 0.0789,
+      "step": 323
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.4130434782608694e-07,
+      "loss": 0.0073,
+      "step": 324
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.411231884057971e-07,
+      "loss": 0.0874,
+      "step": 325
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.4094202898550724e-07,
+      "loss": 0.0041,
+      "step": 326
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.4076086956521737e-07,
+      "loss": 0.0061,
+      "step": 327
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.405797101449275e-07,
+      "loss": 0.2491,
+      "step": 328
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.4039855072463767e-07,
+      "loss": 0.0083,
+      "step": 329
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.402173913043478e-07,
+      "loss": 0.1808,
+      "step": 330
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.40036231884058e-07,
+      "loss": 0.2216,
+      "step": 331
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.3985507246376805e-07,
+      "loss": 0.4314,
+      "step": 332
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.3967391304347823e-07,
+      "loss": 0.0834,
+      "step": 333
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.394927536231884e-07,
+      "loss": 0.3467,
+      "step": 334
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.3931159420289853e-07,
+      "loss": 0.3611,
+      "step": 335
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.391304347826087e-07,
+      "loss": 0.5974,
+      "step": 336
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.389492753623188e-07,
+      "loss": 0.0082,
+      "step": 337
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.3876811594202896e-07,
+      "loss": 0.0754,
+      "step": 338
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.385869565217391e-07,
+      "loss": 0.2882,
+      "step": 339
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.3840579710144926e-07,
+      "loss": 0.0039,
+      "step": 340
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.3822463768115944e-07,
+      "loss": 0.0242,
+      "step": 341
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.3804347826086956e-07,
+      "loss": 0.0262,
+      "step": 342
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.378623188405797e-07,
+      "loss": 0.0097,
+      "step": 343
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.376811594202898e-07,
+      "loss": 0.0127,
+      "step": 344
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.375e-07,
+      "loss": 0.0073,
+      "step": 345
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.373188405797101e-07,
+      "loss": 0.0058,
+      "step": 346
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.371376811594203e-07,
+      "loss": 0.0093,
+      "step": 347
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.3695652173913037e-07,
+      "loss": 0.0203,
+      "step": 348
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.3677536231884055e-07,
+      "loss": 0.1264,
+      "step": 349
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.365942028985507e-07,
+      "loss": 0.008,
+      "step": 350
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.3641304347826085e-07,
+      "loss": 0.0298,
+      "step": 351
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.3623188405797103e-07,
+      "loss": 0.0119,
+      "step": 352
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.3605072463768115e-07,
+      "loss": 0.0774,
+      "step": 353
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.358695652173913e-07,
+      "loss": 0.0069,
+      "step": 354
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.356884057971014e-07,
+      "loss": 0.2596,
+      "step": 355
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.355072463768116e-07,
+      "loss": 0.0587,
+      "step": 356
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.353260869565217e-07,
+      "loss": 0.0241,
+      "step": 357
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.351449275362319e-07,
+      "loss": 0.4977,
+      "step": 358
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.34963768115942e-07,
+      "loss": 0.0947,
+      "step": 359
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.3478260869565214e-07,
+      "loss": 0.1855,
+      "step": 360
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.346014492753623e-07,
+      "loss": 0.0117,
+      "step": 361
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.3442028985507244e-07,
+      "loss": 0.0723,
+      "step": 362
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.342391304347826e-07,
+      "loss": 0.0305,
+      "step": 363
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.340579710144927e-07,
+      "loss": 0.0192,
+      "step": 364
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.3387681159420287e-07,
+      "loss": 0.006,
+      "step": 365
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.33695652173913e-07,
+      "loss": 0.277,
+      "step": 366
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.3351449275362317e-07,
+      "loss": 0.0397,
+      "step": 367
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.3333333333333335e-07,
+      "loss": 0.1262,
+      "step": 368
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.331521739130435e-07,
+      "loss": 0.0094,
+      "step": 369
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.329710144927536e-07,
+      "loss": 0.0606,
+      "step": 370
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.327898550724637e-07,
+      "loss": 0.1139,
+      "step": 371
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.326086956521739e-07,
+      "loss": 0.3338,
+      "step": 372
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.3242753623188403e-07,
+      "loss": 0.0055,
+      "step": 373
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.322463768115942e-07,
+      "loss": 0.502,
+      "step": 374
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.320652173913043e-07,
+      "loss": 0.0086,
+      "step": 375
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.3188405797101446e-07,
+      "loss": 0.036,
+      "step": 376
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.3170289855072463e-07,
+      "loss": 0.2287,
+      "step": 377
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.3152173913043476e-07,
+      "loss": 0.0588,
+      "step": 378
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.3134057971014494e-07,
+      "loss": 0.0243,
+      "step": 379
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.3115942028985506e-07,
+      "loss": 0.0615,
+      "step": 380
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.309782608695652e-07,
+      "loss": 0.1099,
+      "step": 381
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.307971014492753e-07,
+      "loss": 0.1008,
+      "step": 382
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.306159420289855e-07,
+      "loss": 0.0211,
+      "step": 383
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.3043478260869567e-07,
+      "loss": 0.0394,
+      "step": 384
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.302536231884058e-07,
+      "loss": 0.0097,
+      "step": 385
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.300724637681159e-07,
+      "loss": 0.0181,
+      "step": 386
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.2989130434782605e-07,
+      "loss": 0.1603,
+      "step": 387
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.297101449275362e-07,
+      "loss": 0.0388,
+      "step": 388
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.2952898550724635e-07,
+      "loss": 0.0046,
+      "step": 389
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.2934782608695653e-07,
+      "loss": 0.0081,
+      "step": 390
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.291666666666666e-07,
+      "loss": 0.0708,
+      "step": 391
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.289855072463768e-07,
+      "loss": 0.0141,
+      "step": 392
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.2880434782608695e-07,
+      "loss": 0.015,
+      "step": 393
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.286231884057971e-07,
+      "loss": 0.1401,
+      "step": 394
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.2844202898550726e-07,
+      "loss": 0.3438,
+      "step": 395
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.282608695652174e-07,
+      "loss": 0.0121,
+      "step": 396
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.280797101449275e-07,
+      "loss": 0.0126,
+      "step": 397
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.2789855072463763e-07,
+      "loss": 0.0384,
+      "step": 398
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.277173913043478e-07,
+      "loss": 0.006,
+      "step": 399
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.2753623188405794e-07,
+      "loss": 0.1013,
+      "step": 400
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.273550724637681e-07,
+      "loss": 0.0272,
+      "step": 401
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.2717391304347824e-07,
+      "loss": 0.0029,
+      "step": 402
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.2699275362318837e-07,
+      "loss": 0.0051,
+      "step": 403
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.2681159420289854e-07,
+      "loss": 0.006,
+      "step": 404
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.2663043478260867e-07,
+      "loss": 0.0167,
+      "step": 405
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.2644927536231885e-07,
+      "loss": 0.0079,
+      "step": 406
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.262681159420289e-07,
+      "loss": 0.0669,
+      "step": 407
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.260869565217391e-07,
+      "loss": 0.1183,
+      "step": 408
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.259057971014493e-07,
+      "loss": 0.0083,
+      "step": 409
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.257246376811594e-07,
+      "loss": 0.0203,
+      "step": 410
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.255434782608696e-07,
+      "loss": 0.1972,
+      "step": 411
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.253623188405797e-07,
+      "loss": 0.4777,
+      "step": 412
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.2518115942028983e-07,
+      "loss": 0.0067,
+      "step": 413
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.2499999999999995e-07,
+      "loss": 0.167,
+      "step": 414
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.2481884057971013e-07,
+      "loss": 0.0295,
+      "step": 415
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.2463768115942026e-07,
+      "loss": 0.561,
+      "step": 416
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.2445652173913044e-07,
+      "loss": 0.053,
+      "step": 417
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.2427536231884056e-07,
+      "loss": 0.0889,
+      "step": 418
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.240942028985507e-07,
+      "loss": 0.0385,
+      "step": 419
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.2391304347826086e-07,
+      "loss": 0.0121,
+      "step": 420
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.23731884057971e-07,
+      "loss": 0.3258,
+      "step": 421
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.2355072463768117e-07,
+      "loss": 0.0138,
+      "step": 422
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.233695652173913e-07,
+      "loss": 0.0151,
+      "step": 423
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.231884057971014e-07,
+      "loss": 0.0191,
+      "step": 424
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.2300724637681154e-07,
+      "loss": 0.0138,
+      "step": 425
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.228260869565217e-07,
+      "loss": 0.0774,
+      "step": 426
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.226449275362319e-07,
+      "loss": 0.0121,
+      "step": 427
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.22463768115942e-07,
+      "loss": 0.0614,
+      "step": 428
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.2228260869565215e-07,
+      "loss": 0.011,
+      "step": 429
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.221014492753623e-07,
+      "loss": 0.2909,
+      "step": 430
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.2192028985507245e-07,
+      "loss": 0.1674,
+      "step": 431
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.217391304347826e-07,
+      "loss": 0.0032,
+      "step": 432
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.2155797101449276e-07,
+      "loss": 0.0404,
+      "step": 433
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.2137681159420283e-07,
+      "loss": 0.127,
+      "step": 434
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.21195652173913e-07,
+      "loss": 0.0688,
+      "step": 435
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.210144927536232e-07,
+      "loss": 0.0089,
+      "step": 436
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.208333333333333e-07,
+      "loss": 0.4372,
+      "step": 437
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.206521739130435e-07,
+      "loss": 0.398,
+      "step": 438
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.204710144927536e-07,
+      "loss": 0.0433,
+      "step": 439
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.2028985507246374e-07,
+      "loss": 0.007,
+      "step": 440
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.2010869565217386e-07,
+      "loss": 0.0709,
+      "step": 441
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.1992753623188404e-07,
+      "loss": 0.019,
+      "step": 442
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.197463768115942e-07,
+      "loss": 0.0094,
+      "step": 443
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.1956521739130434e-07,
+      "loss": 0.0264,
+      "step": 444
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.1938405797101447e-07,
+      "loss": 0.0208,
+      "step": 445
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.192028985507246e-07,
+      "loss": 0.004,
+      "step": 446
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.1902173913043477e-07,
+      "loss": 0.0088,
+      "step": 447
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.188405797101449e-07,
+      "loss": 0.0044,
+      "step": 448
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.186594202898551e-07,
+      "loss": 0.0582,
+      "step": 449
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.184782608695652e-07,
+      "loss": 0.0044,
+      "step": 450
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.1829710144927533e-07,
+      "loss": 0.0841,
+      "step": 451
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.181159420289855e-07,
+      "loss": 0.0272,
+      "step": 452
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.1793478260869563e-07,
+      "loss": 0.0273,
+      "step": 453
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.177536231884058e-07,
+      "loss": 0.0307,
+      "step": 454
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.1757246376811593e-07,
+      "loss": 0.0914,
+      "step": 455
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.1739130434782606e-07,
+      "loss": 0.0265,
+      "step": 456
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.172101449275362e-07,
+      "loss": 0.0094,
+      "step": 457
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.1702898550724636e-07,
+      "loss": 0.0053,
+      "step": 458
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.168478260869565e-07,
+      "loss": 0.1291,
+      "step": 459
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.1666666666666667e-07,
+      "loss": 0.0026,
+      "step": 460
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8872549019607843,
+      "eval_combined_score": 0.902260544505572,
+      "eval_f1": 0.9172661870503597,
+      "eval_loss": 0.410470575094223,
+      "eval_runtime": 10.0164,
+      "eval_samples_per_second": 40.733,
+      "eval_steps_per_second": 2.596,
+      "step": 460
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.164855072463768e-07,
+      "loss": 0.1736,
+      "step": 461
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.163043478260869e-07,
+      "loss": 0.0176,
+      "step": 462
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 4.161231884057971e-07,
+      "loss": 0.0082,
+      "step": 463
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.159420289855072e-07,
+      "loss": 0.2297,
+      "step": 464
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 4.157608695652174e-07,
+      "loss": 0.0379,
+      "step": 465
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.155797101449275e-07,
+      "loss": 0.275,
+      "step": 466
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.1539855072463765e-07,
+      "loss": 0.2237,
+      "step": 467
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.1521739130434777e-07,
+      "loss": 0.1986,
+      "step": 468
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.1503623188405795e-07,
+      "loss": 0.0522,
+      "step": 469
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.1485507246376813e-07,
+      "loss": 0.0369,
+      "step": 470
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.1467391304347825e-07,
+      "loss": 0.0254,
+      "step": 471
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.144927536231884e-07,
+      "loss": 0.007,
+      "step": 472
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.143115942028985e-07,
+      "loss": 0.0691,
+      "step": 473
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.141304347826087e-07,
+      "loss": 0.0107,
+      "step": 474
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.139492753623188e-07,
+      "loss": 0.0289,
+      "step": 475
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.13768115942029e-07,
+      "loss": 0.003,
+      "step": 476
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.1358695652173916e-07,
+      "loss": 0.3286,
+      "step": 477
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.1340579710144924e-07,
+      "loss": 0.0369,
+      "step": 478
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.132246376811594e-07,
+      "loss": 0.0136,
+      "step": 479
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.1304347826086954e-07,
+      "loss": 0.1732,
+      "step": 480
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.128623188405797e-07,
+      "loss": 0.0042,
+      "step": 481
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.1268115942028984e-07,
+      "loss": 0.3188,
+      "step": 482
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.1249999999999997e-07,
+      "loss": 0.1283,
+      "step": 483
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.123188405797101e-07,
+      "loss": 0.0273,
+      "step": 484
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.1213768115942027e-07,
+      "loss": 0.0776,
+      "step": 485
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.1195652173913045e-07,
+      "loss": 0.0167,
+      "step": 486
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.117753623188406e-07,
+      "loss": 0.0036,
+      "step": 487
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.115942028985507e-07,
+      "loss": 0.4265,
+      "step": 488
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.114130434782608e-07,
+      "loss": 0.006,
+      "step": 489
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.11231884057971e-07,
+      "loss": 0.0239,
+      "step": 490
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1105072463768113e-07,
+      "loss": 0.0379,
+      "step": 491
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.108695652173913e-07,
+      "loss": 0.0107,
+      "step": 492
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.1068840579710143e-07,
+      "loss": 0.004,
+      "step": 493
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.1050724637681156e-07,
+      "loss": 0.3443,
+      "step": 494
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 4.1032608695652173e-07,
+      "loss": 0.3406,
+      "step": 495
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.1014492753623186e-07,
+      "loss": 0.149,
+      "step": 496
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 4.0996376811594204e-07,
+      "loss": 0.0726,
+      "step": 497
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.0978260869565216e-07,
+      "loss": 0.7459,
+      "step": 498
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.096014492753623e-07,
+      "loss": 0.1177,
+      "step": 499
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 4.094202898550724e-07,
+      "loss": 0.0064,
+      "step": 500
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.092391304347826e-07,
+      "loss": 0.8413,
+      "step": 501
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 4.090579710144927e-07,
+      "loss": 0.0281,
+      "step": 502
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.088768115942029e-07,
+      "loss": 0.1825,
+      "step": 503
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 4.0869565217391307e-07,
+      "loss": 0.0208,
+      "step": 504
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.0851449275362315e-07,
+      "loss": 0.0036,
+      "step": 505
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.083333333333333e-07,
+      "loss": 0.0582,
+      "step": 506
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 4.0815217391304345e-07,
+      "loss": 0.0038,
+      "step": 507
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.0797101449275363e-07,
+      "loss": 0.3333,
+      "step": 508
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 4.0778985507246375e-07,
+      "loss": 0.022,
+      "step": 509
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.076086956521739e-07,
+      "loss": 0.2766,
+      "step": 510
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 4.07427536231884e-07,
+      "loss": 0.0319,
+      "step": 511
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.072463768115942e-07,
+      "loss": 0.0289,
+      "step": 512
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.0706521739130436e-07,
+      "loss": 0.0045,
+      "step": 513
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 4.068840579710145e-07,
+      "loss": 0.0216,
+      "step": 514
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.067028985507246e-07,
+      "loss": 0.131,
+      "step": 515
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 4.0652173913043473e-07,
+      "loss": 0.0999,
+      "step": 516
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.063405797101449e-07,
+      "loss": 0.0571,
+      "step": 517
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 4.0615942028985504e-07,
+      "loss": 0.2837,
+      "step": 518
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.059782608695652e-07,
+      "loss": 0.0258,
+      "step": 519
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.057971014492754e-07,
+      "loss": 0.0287,
+      "step": 520
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.0561594202898547e-07,
+      "loss": 0.0039,
+      "step": 521
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.0543478260869564e-07,
+      "loss": 0.0056,
+      "step": 522
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.0525362318840577e-07,
+      "loss": 0.0127,
+      "step": 523
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.0507246376811595e-07,
+      "loss": 0.3452,
+      "step": 524
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.0489130434782607e-07,
+      "loss": 0.0119,
+      "step": 525
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.047101449275362e-07,
+      "loss": 0.0519,
+      "step": 526
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.045289855072463e-07,
+      "loss": 0.0125,
+      "step": 527
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.043478260869565e-07,
+      "loss": 0.0481,
+      "step": 528
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.041666666666667e-07,
+      "loss": 0.206,
+      "step": 529
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.039855072463768e-07,
+      "loss": 0.0104,
+      "step": 530
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.0380434782608693e-07,
+      "loss": 0.0273,
+      "step": 531
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.0362318840579705e-07,
+      "loss": 0.0206,
+      "step": 532
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.0344202898550723e-07,
+      "loss": 0.0137,
+      "step": 533
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.0326086956521736e-07,
+      "loss": 0.0048,
+      "step": 534
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.0307971014492754e-07,
+      "loss": 0.0281,
+      "step": 535
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.0289855072463766e-07,
+      "loss": 0.3819,
+      "step": 536
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.027173913043478e-07,
+      "loss": 0.4088,
+      "step": 537
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.0253623188405796e-07,
+      "loss": 0.0639,
+      "step": 538
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.023550724637681e-07,
+      "loss": 0.1967,
+      "step": 539
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.0217391304347827e-07,
+      "loss": 0.0182,
+      "step": 540
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.019927536231884e-07,
+      "loss": 0.1689,
+      "step": 541
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.018115942028985e-07,
+      "loss": 0.0037,
+      "step": 542
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.0163043478260864e-07,
+      "loss": 0.1464,
+      "step": 543
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.014492753623188e-07,
+      "loss": 0.0319,
+      "step": 544
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.0126811594202895e-07,
+      "loss": 0.0252,
+      "step": 545
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.010869565217391e-07,
+      "loss": 0.1536,
+      "step": 546
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.009057971014493e-07,
+      "loss": 0.0085,
+      "step": 547
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.007246376811594e-07,
+      "loss": 0.0491,
+      "step": 548
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.0054347826086955e-07,
+      "loss": 0.019,
+      "step": 549
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.003623188405797e-07,
+      "loss": 0.0111,
+      "step": 550
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.0018115942028986e-07,
+      "loss": 0.003,
+      "step": 551
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4e-07,
+      "loss": 0.0954,
+      "step": 552
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 3.998188405797101e-07,
+      "loss": 0.1872,
+      "step": 553
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.996376811594203e-07,
+      "loss": 0.0169,
+      "step": 554
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.994565217391304e-07,
+      "loss": 0.3957,
+      "step": 555
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.992753623188406e-07,
+      "loss": 0.2421,
+      "step": 556
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.990942028985507e-07,
+      "loss": 0.0149,
+      "step": 557
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.9891304347826084e-07,
+      "loss": 0.2924,
+      "step": 558
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.9873188405797096e-07,
+      "loss": 0.0232,
+      "step": 559
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.9855072463768114e-07,
+      "loss": 0.4088,
+      "step": 560
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.9836956521739127e-07,
+      "loss": 0.004,
+      "step": 561
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.9818840579710145e-07,
+      "loss": 0.0265,
+      "step": 562
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.980072463768116e-07,
+      "loss": 0.0465,
+      "step": 563
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.978260869565217e-07,
+      "loss": 0.047,
+      "step": 564
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.976449275362319e-07,
+      "loss": 0.0202,
+      "step": 565
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.97463768115942e-07,
+      "loss": 0.0059,
+      "step": 566
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.972826086956522e-07,
+      "loss": 0.0292,
+      "step": 567
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.971014492753623e-07,
+      "loss": 0.0059,
+      "step": 568
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.9692028985507243e-07,
+      "loss": 0.0075,
+      "step": 569
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.9673913043478255e-07,
+      "loss": 0.0038,
+      "step": 570
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.9655797101449273e-07,
+      "loss": 0.2557,
+      "step": 571
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.963768115942029e-07,
+      "loss": 0.1168,
+      "step": 572
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.9619565217391303e-07,
+      "loss": 0.2964,
+      "step": 573
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.960144927536232e-07,
+      "loss": 0.2318,
+      "step": 574
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.958333333333333e-07,
+      "loss": 0.0092,
+      "step": 575
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.9565217391304346e-07,
+      "loss": 0.0925,
+      "step": 576
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.954710144927536e-07,
+      "loss": 0.0705,
+      "step": 577
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.9528985507246377e-07,
+      "loss": 0.0644,
+      "step": 578
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.9510869565217394e-07,
+      "loss": 0.0181,
+      "step": 579
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.94927536231884e-07,
+      "loss": 0.027,
+      "step": 580
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.947463768115942e-07,
+      "loss": 0.0882,
+      "step": 581
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.945652173913043e-07,
+      "loss": 0.0056,
+      "step": 582
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.943840579710145e-07,
+      "loss": 0.0098,
+      "step": 583
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.942028985507246e-07,
+      "loss": 0.0076,
+      "step": 584
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.9402173913043475e-07,
+      "loss": 0.023,
+      "step": 585
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.9384057971014487e-07,
+      "loss": 0.3907,
+      "step": 586
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.9365942028985505e-07,
+      "loss": 0.0621,
+      "step": 587
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.9347826086956523e-07,
+      "loss": 0.1628,
+      "step": 588
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.9329710144927535e-07,
+      "loss": 0.1415,
+      "step": 589
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.9311594202898553e-07,
+      "loss": 0.6444,
+      "step": 590
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.929347826086956e-07,
+      "loss": 0.3134,
+      "step": 591
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.927536231884058e-07,
+      "loss": 0.2566,
+      "step": 592
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 3.925724637681159e-07,
+      "loss": 0.0029,
+      "step": 593
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 3.923913043478261e-07,
+      "loss": 0.1318,
+      "step": 594
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 3.922101449275362e-07,
+      "loss": 0.0375,
+      "step": 595
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 3.9202898550724634e-07,
+      "loss": 0.0142,
+      "step": 596
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 3.918478260869565e-07,
+      "loss": 0.3516,
+      "step": 597
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 3.9166666666666664e-07,
+      "loss": 0.0161,
+      "step": 598
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 3.914855072463768e-07,
+      "loss": 0.1442,
+      "step": 599
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 3.9130434782608694e-07,
+      "loss": 0.0187,
+      "step": 600
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 3.911231884057971e-07,
+      "loss": 0.0066,
+      "step": 601
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 3.909420289855072e-07,
+      "loss": 0.6487,
+      "step": 602
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 3.9076086956521737e-07,
+      "loss": 0.0069,
+      "step": 603
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 3.905797101449275e-07,
+      "loss": 0.2932,
+      "step": 604
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 3.903985507246377e-07,
+      "loss": 0.1761,
+      "step": 605
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 3.9021739130434785e-07,
+      "loss": 0.1269,
+      "step": 606
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.900362318840579e-07,
+      "loss": 0.0357,
+      "step": 607
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.898550724637681e-07,
+      "loss": 0.2455,
+      "step": 608
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 3.8967391304347823e-07,
+      "loss": 0.0221,
+      "step": 609
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 3.894927536231884e-07,
+      "loss": 0.1672,
+      "step": 610
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.8931159420289853e-07,
+      "loss": 0.0796,
+      "step": 611
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.8913043478260866e-07,
+      "loss": 0.0281,
+      "step": 612
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.889492753623188e-07,
+      "loss": 0.0045,
+      "step": 613
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.8876811594202896e-07,
+      "loss": 0.2352,
+      "step": 614
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.8858695652173914e-07,
+      "loss": 0.0272,
+      "step": 615
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 3.8840579710144926e-07,
+      "loss": 0.0152,
+      "step": 616
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 3.8822463768115944e-07,
+      "loss": 0.0092,
+      "step": 617
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 3.880434782608695e-07,
+      "loss": 0.0227,
+      "step": 618
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 3.878623188405797e-07,
+      "loss": 0.019,
+      "step": 619
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 3.876811594202898e-07,
+      "loss": 0.0088,
+      "step": 620
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 3.875e-07,
+      "loss": 0.0082,
+      "step": 621
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 3.873188405797102e-07,
+      "loss": 0.2902,
+      "step": 622
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 3.8713768115942025e-07,
+      "loss": 0.3015,
+      "step": 623
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 3.869565217391304e-07,
+      "loss": 0.0843,
+      "step": 624
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 3.8677536231884055e-07,
+      "loss": 0.0164,
+      "step": 625
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 3.8659420289855073e-07,
+      "loss": 0.0333,
+      "step": 626
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 3.8641304347826085e-07,
+      "loss": 0.0215,
+      "step": 627
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 3.8623188405797103e-07,
+      "loss": 0.0072,
+      "step": 628
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 3.860507246376811e-07,
+      "loss": 0.0048,
+      "step": 629
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 3.858695652173913e-07,
+      "loss": 0.0079,
+      "step": 630
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 3.8568840579710146e-07,
+      "loss": 0.1312,
+      "step": 631
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 3.855072463768116e-07,
+      "loss": 0.0076,
+      "step": 632
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 3.8532608695652176e-07,
+      "loss": 0.0266,
+      "step": 633
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 3.8514492753623183e-07,
+      "loss": 0.0084,
+      "step": 634
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 3.84963768115942e-07,
+      "loss": 0.0055,
+      "step": 635
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 3.8478260869565214e-07,
+      "loss": 0.151,
+      "step": 636
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 3.846014492753623e-07,
+      "loss": 0.0036,
+      "step": 637
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 3.8442028985507244e-07,
+      "loss": 0.1113,
+      "step": 638
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 3.8423913043478257e-07,
+      "loss": 0.0087,
+      "step": 639
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 3.8405797101449274e-07,
+      "loss": 0.4738,
+      "step": 640
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 3.8387681159420287e-07,
+      "loss": 0.3524,
+      "step": 641
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 3.8369565217391305e-07,
+      "loss": 0.0123,
+      "step": 642
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 3.8351449275362317e-07,
+      "loss": 0.0049,
+      "step": 643
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 3.8333333333333335e-07,
+      "loss": 0.3096,
+      "step": 644
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 3.831521739130434e-07,
+      "loss": 0.0465,
+      "step": 645
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 3.829710144927536e-07,
+      "loss": 0.0341,
+      "step": 646
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 3.8278985507246373e-07,
+      "loss": 0.0158,
+      "step": 647
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 3.826086956521739e-07,
+      "loss": 0.123,
+      "step": 648
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 3.824275362318841e-07,
+      "loss": 0.004,
+      "step": 649
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.8224637681159416e-07,
+      "loss": 0.336,
+      "step": 650
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.8206521739130433e-07,
+      "loss": 0.0065,
+      "step": 651
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.8188405797101446e-07,
+      "loss": 0.2844,
+      "step": 652
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.8170289855072464e-07,
+      "loss": 0.436,
+      "step": 653
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.8152173913043476e-07,
+      "loss": 0.0825,
+      "step": 654
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.813405797101449e-07,
+      "loss": 0.4895,
+      "step": 655
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.81159420289855e-07,
+      "loss": 0.005,
+      "step": 656
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 3.809782608695652e-07,
+      "loss": 0.0258,
+      "step": 657
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 3.8079710144927537e-07,
+      "loss": 0.0654,
+      "step": 658
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 3.806159420289855e-07,
+      "loss": 0.3691,
+      "step": 659
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 3.8043478260869567e-07,
+      "loss": 0.2798,
+      "step": 660
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 3.8025362318840574e-07,
+      "loss": 0.0259,
+      "step": 661
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 3.800724637681159e-07,
+      "loss": 0.6486,
+      "step": 662
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 3.7989130434782605e-07,
+      "loss": 0.0241,
+      "step": 663
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.797101449275362e-07,
+      "loss": 0.2396,
+      "step": 664
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.795289855072464e-07,
+      "loss": 0.0066,
+      "step": 665
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.793478260869565e-07,
+      "loss": 0.0173,
+      "step": 666
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.7916666666666665e-07,
+      "loss": 0.07,
+      "step": 667
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.789855072463768e-07,
+      "loss": 0.0085,
+      "step": 668
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 3.7880434782608696e-07,
+      "loss": 0.2512,
+      "step": 669
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 3.786231884057971e-07,
+      "loss": 0.0166,
+      "step": 670
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 3.7844202898550726e-07,
+      "loss": 0.005,
+      "step": 671
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 3.7826086956521733e-07,
+      "loss": 0.0065,
+      "step": 672
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.780797101449275e-07,
+      "loss": 0.2184,
+      "step": 673
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.778985507246377e-07,
+      "loss": 0.0245,
+      "step": 674
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.777173913043478e-07,
+      "loss": 0.1576,
+      "step": 675
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 3.77536231884058e-07,
+      "loss": 0.2655,
+      "step": 676
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 3.7735507246376806e-07,
+      "loss": 0.0079,
+      "step": 677
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.7717391304347824e-07,
+      "loss": 0.2925,
+      "step": 678
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.7699275362318837e-07,
+      "loss": 0.0097,
+      "step": 679
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 3.7681159420289855e-07,
+      "loss": 0.0022,
+      "step": 680
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 3.7663043478260867e-07,
+      "loss": 0.0089,
+      "step": 681
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.764492753623188e-07,
+      "loss": 0.0047,
+      "step": 682
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.76268115942029e-07,
+      "loss": 0.025,
+      "step": 683
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.760869565217391e-07,
+      "loss": 0.0196,
+      "step": 684
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.759057971014493e-07,
+      "loss": 0.1376,
+      "step": 685
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.757246376811594e-07,
+      "loss": 0.1593,
+      "step": 686
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 3.755434782608696e-07,
+      "loss": 0.029,
+      "step": 687
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 3.7536231884057965e-07,
+      "loss": 0.0451,
+      "step": 688
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.7518115942028983e-07,
+      "loss": 0.0036,
+      "step": 689
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.75e-07,
+      "loss": 0.0026,
+      "step": 690
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8946078431372549,
+      "eval_combined_score": 0.9091156444815938,
+      "eval_f1": 0.9236234458259325,
+      "eval_loss": 0.3707413673400879,
+      "eval_runtime": 8.7667,
+      "eval_samples_per_second": 46.54,
+      "eval_steps_per_second": 2.966,
+      "step": 690
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.7481884057971013e-07,
+      "loss": 0.0068,
+      "step": 691
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.746376811594203e-07,
+      "loss": 0.2153,
+      "step": 692
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.744565217391304e-07,
+      "loss": 0.0876,
+      "step": 693
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 3.7427536231884056e-07,
+      "loss": 0.0138,
+      "step": 694
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 3.740942028985507e-07,
+      "loss": 0.0061,
+      "step": 695
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.7391304347826087e-07,
+      "loss": 0.0469,
+      "step": 696
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.73731884057971e-07,
+      "loss": 0.056,
+      "step": 697
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.7355072463768117e-07,
+      "loss": 0.0046,
+      "step": 698
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.733695652173913e-07,
+      "loss": 0.0256,
+      "step": 699
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.731884057971014e-07,
+      "loss": 0.0227,
+      "step": 700
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.730072463768116e-07,
+      "loss": 0.0141,
+      "step": 701
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.728260869565217e-07,
+      "loss": 0.4175,
+      "step": 702
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 3.726449275362319e-07,
+      "loss": 0.184,
+      "step": 703
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 3.72463768115942e-07,
+      "loss": 0.2781,
+      "step": 704
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.7228260869565215e-07,
+      "loss": 0.315,
+      "step": 705
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.721014492753623e-07,
+      "loss": 0.0074,
+      "step": 706
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.7192028985507246e-07,
+      "loss": 0.0102,
+      "step": 707
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 3.7173913043478263e-07,
+      "loss": 0.0044,
+      "step": 708
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 3.715579710144927e-07,
+      "loss": 0.0251,
+      "step": 709
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.713768115942029e-07,
+      "loss": 0.1116,
+      "step": 710
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.71195652173913e-07,
+      "loss": 0.0168,
+      "step": 711
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.710144927536232e-07,
+      "loss": 0.0353,
+      "step": 712
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.708333333333333e-07,
+      "loss": 0.1118,
+      "step": 713
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.706521739130435e-07,
+      "loss": 0.1331,
+      "step": 714
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.7047101449275356e-07,
+      "loss": 0.3309,
+      "step": 715
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.7028985507246374e-07,
+      "loss": 0.0168,
+      "step": 716
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.701086956521739e-07,
+      "loss": 0.0042,
+      "step": 717
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.6992753623188404e-07,
+      "loss": 0.0301,
+      "step": 718
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 3.697463768115942e-07,
+      "loss": 0.055,
+      "step": 719
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 3.695652173913043e-07,
+      "loss": 0.2141,
+      "step": 720
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 3.6938405797101447e-07,
+      "loss": 0.0094,
+      "step": 721
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.692028985507246e-07,
+      "loss": 0.03,
+      "step": 722
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.690217391304348e-07,
+      "loss": 0.0063,
+      "step": 723
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 3.6884057971014495e-07,
+      "loss": 0.491,
+      "step": 724
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 3.686594202898551e-07,
+      "loss": 0.1236,
+      "step": 725
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.684782608695652e-07,
+      "loss": 0.1352,
+      "step": 726
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.6829710144927533e-07,
+      "loss": 0.1155,
+      "step": 727
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.681159420289855e-07,
+      "loss": 0.1595,
+      "step": 728
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.6793478260869563e-07,
+      "loss": 0.0052,
+      "step": 729
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.677536231884058e-07,
+      "loss": 0.0625,
+      "step": 730
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.675724637681159e-07,
+      "loss": 0.0127,
+      "step": 731
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.6739130434782606e-07,
+      "loss": 0.0389,
+      "step": 732
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 3.6721014492753624e-07,
+      "loss": 0.1466,
+      "step": 733
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 3.6702898550724636e-07,
+      "loss": 0.0368,
+      "step": 734
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.6684782608695654e-07,
+      "loss": 0.3592,
+      "step": 735
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.666666666666666e-07,
+      "loss": 0.1775,
+      "step": 736
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.664855072463768e-07,
+      "loss": 0.3018,
+      "step": 737
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.663043478260869e-07,
+      "loss": 0.033,
+      "step": 738
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.661231884057971e-07,
+      "loss": 0.4745,
+      "step": 739
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 3.659420289855072e-07,
+      "loss": 0.4027,
+      "step": 740
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 3.657608695652174e-07,
+      "loss": 0.0254,
+      "step": 741
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.655797101449275e-07,
+      "loss": 0.1019,
+      "step": 742
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.6539855072463765e-07,
+      "loss": 0.0159,
+      "step": 743
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.6521739130434783e-07,
+      "loss": 0.0446,
+      "step": 744
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 3.6503623188405795e-07,
+      "loss": 0.3428,
+      "step": 745
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 3.6485507246376813e-07,
+      "loss": 0.009,
+      "step": 746
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.646739130434782e-07,
+      "loss": 0.1195,
+      "step": 747
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.644927536231884e-07,
+      "loss": 0.0096,
+      "step": 748
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 3.643115942028985e-07,
+      "loss": 0.0055,
+      "step": 749
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 3.641304347826087e-07,
+      "loss": 0.0103,
+      "step": 750
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.6394927536231886e-07,
+      "loss": 0.1507,
+      "step": 751
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.6376811594202894e-07,
+      "loss": 0.1586,
+      "step": 752
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.635869565217391e-07,
+      "loss": 0.1951,
+      "step": 753
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 3.6340579710144924e-07,
+      "loss": 1.0777,
+      "step": 754
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 3.632246376811594e-07,
+      "loss": 0.2979,
+      "step": 755
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 3.6304347826086954e-07,
+      "loss": 0.0372,
+      "step": 756
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 3.628623188405797e-07,
+      "loss": 0.0145,
+      "step": 757
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.626811594202898e-07,
+      "loss": 0.0232,
+      "step": 758
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.6249999999999997e-07,
+      "loss": 0.0066,
+      "step": 759
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.6231884057971015e-07,
+      "loss": 0.3453,
+      "step": 760
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 3.621376811594203e-07,
+      "loss": 0.0155,
+      "step": 761
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 3.6195652173913045e-07,
+      "loss": 0.0915,
+      "step": 762
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.617753623188405e-07,
+      "loss": 0.1527,
+      "step": 763
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.615942028985507e-07,
+      "loss": 0.1679,
+      "step": 764
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.6141304347826083e-07,
+      "loss": 0.0143,
+      "step": 765
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.61231884057971e-07,
+      "loss": 0.0355,
+      "step": 766
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.610507246376812e-07,
+      "loss": 0.2229,
+      "step": 767
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.608695652173913e-07,
+      "loss": 0.2891,
+      "step": 768
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.6068840579710143e-07,
+      "loss": 0.0178,
+      "step": 769
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.6050724637681156e-07,
+      "loss": 0.0186,
+      "step": 770
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.6032608695652174e-07,
+      "loss": 0.1925,
+      "step": 771
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.6014492753623186e-07,
+      "loss": 0.1039,
+      "step": 772
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.5996376811594204e-07,
+      "loss": 0.3549,
+      "step": 773
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.597826086956521e-07,
+      "loss": 0.1091,
+      "step": 774
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.596014492753623e-07,
+      "loss": 0.2333,
+      "step": 775
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.5942028985507247e-07,
+      "loss": 0.0266,
+      "step": 776
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.592391304347826e-07,
+      "loss": 0.0091,
+      "step": 777
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.5905797101449277e-07,
+      "loss": 0.0028,
+      "step": 778
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 3.5887681159420284e-07,
+      "loss": 0.0051,
+      "step": 779
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 3.58695652173913e-07,
+      "loss": 0.0145,
+      "step": 780
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.5851449275362315e-07,
+      "loss": 0.2478,
+      "step": 781
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.583333333333333e-07,
+      "loss": 0.1927,
+      "step": 782
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.5815217391304345e-07,
+      "loss": 0.4606,
+      "step": 783
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 3.5797101449275363e-07,
+      "loss": 0.006,
+      "step": 784
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 3.5778985507246375e-07,
+      "loss": 0.113,
+      "step": 785
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 3.576086956521739e-07,
+      "loss": 0.5206,
+      "step": 786
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 3.5742753623188406e-07,
+      "loss": 0.02,
+      "step": 787
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.572463768115942e-07,
+      "loss": 0.057,
+      "step": 788
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.5706521739130436e-07,
+      "loss": 0.0056,
+      "step": 789
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.5688405797101443e-07,
+      "loss": 0.0916,
+      "step": 790
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.567028985507246e-07,
+      "loss": 0.0198,
+      "step": 791
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.5652173913043474e-07,
+      "loss": 0.0056,
+      "step": 792
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 3.563405797101449e-07,
+      "loss": 0.0926,
+      "step": 793
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 3.561594202898551e-07,
+      "loss": 0.0303,
+      "step": 794
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.559782608695652e-07,
+      "loss": 0.1987,
+      "step": 795
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.5579710144927534e-07,
+      "loss": 0.0846,
+      "step": 796
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.5561594202898547e-07,
+      "loss": 0.0309,
+      "step": 797
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.5543478260869565e-07,
+      "loss": 0.1853,
+      "step": 798
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.5525362318840577e-07,
+      "loss": 0.2297,
+      "step": 799
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.5507246376811595e-07,
+      "loss": 0.0152,
+      "step": 800
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.54891304347826e-07,
+      "loss": 0.0042,
+      "step": 801
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 3.547101449275362e-07,
+      "loss": 0.2511,
+      "step": 802
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 3.545289855072464e-07,
+      "loss": 0.0687,
+      "step": 803
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.543478260869565e-07,
+      "loss": 0.0591,
+      "step": 804
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.541666666666667e-07,
+      "loss": 0.1775,
+      "step": 805
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.5398550724637675e-07,
+      "loss": 0.0158,
+      "step": 806
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 3.5380434782608693e-07,
+      "loss": 0.0148,
+      "step": 807
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 3.5362318840579706e-07,
+      "loss": 0.0055,
+      "step": 808
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 3.5344202898550724e-07,
+      "loss": 0.1367,
+      "step": 809
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 3.532608695652174e-07,
+      "loss": 0.0437,
+      "step": 810
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.5307971014492754e-07,
+      "loss": 0.0187,
+      "step": 811
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.5289855072463766e-07,
+      "loss": 0.2786,
+      "step": 812
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.527173913043478e-07,
+      "loss": 0.3367,
+      "step": 813
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 3.5253623188405797e-07,
+      "loss": 0.0131,
+      "step": 814
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 3.523550724637681e-07,
+      "loss": 0.0916,
+      "step": 815
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.5217391304347827e-07,
+      "loss": 0.0132,
+      "step": 816
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.5199275362318834e-07,
+      "loss": 0.5121,
+      "step": 817
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 3.518115942028985e-07,
+      "loss": 0.1656,
+      "step": 818
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 3.516304347826087e-07,
+      "loss": 0.0832,
+      "step": 819
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.514492753623188e-07,
+      "loss": 0.0496,
+      "step": 820
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.51268115942029e-07,
+      "loss": 0.039,
+      "step": 821
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.5108695652173913e-07,
+      "loss": 0.3077,
+      "step": 822
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 3.5090579710144925e-07,
+      "loss": 0.1057,
+      "step": 823
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 3.507246376811594e-07,
+      "loss": 0.0621,
+      "step": 824
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.5054347826086956e-07,
+      "loss": 0.0035,
+      "step": 825
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.503623188405797e-07,
+      "loss": 0.2523,
+      "step": 826
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.5018115942028986e-07,
+      "loss": 0.0281,
+      "step": 827
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.5e-07,
+      "loss": 0.0038,
+      "step": 828
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.498188405797101e-07,
+      "loss": 0.0973,
+      "step": 829
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.496376811594203e-07,
+      "loss": 0.0089,
+      "step": 830
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.494565217391304e-07,
+      "loss": 0.2538,
+      "step": 831
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.492753623188406e-07,
+      "loss": 0.0076,
+      "step": 832
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.4909420289855066e-07,
+      "loss": 0.3548,
+      "step": 833
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.4891304347826084e-07,
+      "loss": 0.2206,
+      "step": 834
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.48731884057971e-07,
+      "loss": 0.183,
+      "step": 835
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.4855072463768114e-07,
+      "loss": 0.3489,
+      "step": 836
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.483695652173913e-07,
+      "loss": 0.0079,
+      "step": 837
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.4818840579710145e-07,
+      "loss": 0.7327,
+      "step": 838
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.4800724637681157e-07,
+      "loss": 0.0571,
+      "step": 839
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.478260869565217e-07,
+      "loss": 0.064,
+      "step": 840
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.476449275362319e-07,
+      "loss": 0.4016,
+      "step": 841
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.47463768115942e-07,
+      "loss": 0.0055,
+      "step": 842
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.472826086956522e-07,
+      "loss": 0.1305,
+      "step": 843
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.471014492753623e-07,
+      "loss": 0.1713,
+      "step": 844
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.4692028985507243e-07,
+      "loss": 0.0746,
+      "step": 845
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.467391304347826e-07,
+      "loss": 0.1359,
+      "step": 846
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.4655797101449273e-07,
+      "loss": 0.0196,
+      "step": 847
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.463768115942029e-07,
+      "loss": 0.1568,
+      "step": 848
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.4619565217391304e-07,
+      "loss": 0.1685,
+      "step": 849
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.4601449275362316e-07,
+      "loss": 0.3564,
+      "step": 850
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.458333333333333e-07,
+      "loss": 0.0156,
+      "step": 851
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.4565217391304346e-07,
+      "loss": 0.3732,
+      "step": 852
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.4547101449275364e-07,
+      "loss": 0.0464,
+      "step": 853
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.4528985507246377e-07,
+      "loss": 0.7348,
+      "step": 854
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.451086956521739e-07,
+      "loss": 0.003,
+      "step": 855
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.44927536231884e-07,
+      "loss": 0.0081,
+      "step": 856
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.447463768115942e-07,
+      "loss": 0.1079,
+      "step": 857
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.445652173913043e-07,
+      "loss": 0.1506,
+      "step": 858
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.443840579710145e-07,
+      "loss": 0.0109,
+      "step": 859
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.4420289855072457e-07,
+      "loss": 0.3453,
+      "step": 860
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.4402173913043475e-07,
+      "loss": 0.0063,
+      "step": 861
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.4384057971014493e-07,
+      "loss": 0.0857,
+      "step": 862
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.4365942028985505e-07,
+      "loss": 0.3483,
+      "step": 863
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.4347826086956523e-07,
+      "loss": 0.0132,
+      "step": 864
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.4329710144927536e-07,
+      "loss": 0.0046,
+      "step": 865
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.431159420289855e-07,
+      "loss": 0.0403,
+      "step": 866
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.429347826086956e-07,
+      "loss": 0.0839,
+      "step": 867
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.427536231884058e-07,
+      "loss": 0.0198,
+      "step": 868
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.4257246376811596e-07,
+      "loss": 0.0171,
+      "step": 869
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.423913043478261e-07,
+      "loss": 0.3275,
+      "step": 870
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.422101449275362e-07,
+      "loss": 0.0447,
+      "step": 871
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.4202898550724634e-07,
+      "loss": 0.0063,
+      "step": 872
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.418478260869565e-07,
+      "loss": 0.2072,
+      "step": 873
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.4166666666666664e-07,
+      "loss": 0.0102,
+      "step": 874
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.414855072463768e-07,
+      "loss": 0.0531,
+      "step": 875
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 3.413043478260869e-07,
+      "loss": 0.0048,
+      "step": 876
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 3.4112318840579707e-07,
+      "loss": 0.3341,
+      "step": 877
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 3.4094202898550725e-07,
+      "loss": 0.2271,
+      "step": 878
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 3.407608695652174e-07,
+      "loss": 0.0174,
+      "step": 879
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.4057971014492755e-07,
+      "loss": 0.0505,
+      "step": 880
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.403985507246377e-07,
+      "loss": 0.3291,
+      "step": 881
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.402173913043478e-07,
+      "loss": 0.0628,
+      "step": 882
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 3.4003623188405793e-07,
+      "loss": 0.1227,
+      "step": 883
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 3.398550724637681e-07,
+      "loss": 0.0431,
+      "step": 884
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 3.3967391304347823e-07,
+      "loss": 0.0618,
+      "step": 885
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 3.394927536231884e-07,
+      "loss": 0.0077,
+      "step": 886
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 3.3931159420289853e-07,
+      "loss": 0.3023,
+      "step": 887
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 3.3913043478260866e-07,
+      "loss": 0.0753,
+      "step": 888
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.3894927536231884e-07,
+      "loss": 0.3343,
+      "step": 889
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.3876811594202896e-07,
+      "loss": 0.0031,
+      "step": 890
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.3858695652173914e-07,
+      "loss": 0.4601,
+      "step": 891
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 3.3840579710144927e-07,
+      "loss": 0.0319,
+      "step": 892
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 3.382246376811594e-07,
+      "loss": 0.0048,
+      "step": 893
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.380434782608695e-07,
+      "loss": 0.0105,
+      "step": 894
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.378623188405797e-07,
+      "loss": 0.0514,
+      "step": 895
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 3.3768115942028987e-07,
+      "loss": 0.0073,
+      "step": 896
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 3.375e-07,
+      "loss": 0.0187,
+      "step": 897
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 3.373188405797101e-07,
+      "loss": 0.2798,
+      "step": 898
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.3713768115942025e-07,
+      "loss": 0.0893,
+      "step": 899
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.369565217391304e-07,
+      "loss": 0.0156,
+      "step": 900
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 3.3677536231884055e-07,
+      "loss": 0.0111,
+      "step": 901
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 3.3659420289855073e-07,
+      "loss": 0.0181,
+      "step": 902
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.364130434782608e-07,
+      "loss": 0.3899,
+      "step": 903
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.36231884057971e-07,
+      "loss": 0.0655,
+      "step": 904
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.3605072463768116e-07,
+      "loss": 0.3357,
+      "step": 905
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 3.358695652173913e-07,
+      "loss": 0.0456,
+      "step": 906
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 3.3568840579710146e-07,
+      "loss": 0.047,
+      "step": 907
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 3.355072463768116e-07,
+      "loss": 0.007,
+      "step": 908
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 3.353260869565217e-07,
+      "loss": 0.0036,
+      "step": 909
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.3514492753623184e-07,
+      "loss": 0.2502,
+      "step": 910
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.34963768115942e-07,
+      "loss": 0.0075,
+      "step": 911
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.347826086956522e-07,
+      "loss": 0.0209,
+      "step": 912
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.346014492753623e-07,
+      "loss": 0.36,
+      "step": 913
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.3442028985507244e-07,
+      "loss": 0.0176,
+      "step": 914
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.3423913043478257e-07,
+      "loss": 0.6118,
+      "step": 915
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.3405797101449275e-07,
+      "loss": 0.3893,
+      "step": 916
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 3.3387681159420287e-07,
+      "loss": 0.1723,
+      "step": 917
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 3.3369565217391305e-07,
+      "loss": 0.1238,
+      "step": 918
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.335144927536232e-07,
+      "loss": 0.0141,
+      "step": 919
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.333333333333333e-07,
+      "loss": 0.0037,
+      "step": 920
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8946078431372549,
+      "eval_combined_score": 0.9087042806350548,
+      "eval_f1": 0.9228007181328547,
+      "eval_loss": 0.3892787992954254,
+      "eval_runtime": 9.1624,
+      "eval_samples_per_second": 44.53,
+      "eval_steps_per_second": 2.838,
+      "step": 920
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 3.331521739130435e-07,
+      "loss": 0.0316,
+      "step": 921
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 3.329710144927536e-07,
+      "loss": 0.1486,
+      "step": 922
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 3.327898550724638e-07,
+      "loss": 0.1529,
+      "step": 923
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 3.326086956521739e-07,
+      "loss": 0.0192,
+      "step": 924
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 3.3242753623188403e-07,
+      "loss": 0.013,
+      "step": 925
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 3.3224637681159416e-07,
+      "loss": 0.3189,
+      "step": 926
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 3.3206521739130434e-07,
+      "loss": 0.0162,
+      "step": 927
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 3.3188405797101446e-07,
+      "loss": 0.066,
+      "step": 928
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 3.3170289855072464e-07,
+      "loss": 0.0261,
+      "step": 929
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 3.3152173913043476e-07,
+      "loss": 0.0134,
+      "step": 930
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 3.313405797101449e-07,
+      "loss": 0.0972,
+      "step": 931
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 3.3115942028985507e-07,
+      "loss": 0.021,
+      "step": 932
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 3.309782608695652e-07,
+      "loss": 0.0923,
+      "step": 933
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 3.3079710144927537e-07,
+      "loss": 0.0852,
+      "step": 934
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 3.306159420289855e-07,
+      "loss": 0.0073,
+      "step": 935
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 3.304347826086956e-07,
+      "loss": 0.2197,
+      "step": 936
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 3.3025362318840575e-07,
+      "loss": 0.0285,
+      "step": 937
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 3.300724637681159e-07,
+      "loss": 0.0108,
+      "step": 938
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 3.298913043478261e-07,
+      "loss": 0.0166,
+      "step": 939
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 3.2971014492753623e-07,
+      "loss": 0.0232,
+      "step": 940
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 3.2952898550724635e-07,
+      "loss": 0.1692,
+      "step": 941
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 3.293478260869565e-07,
+      "loss": 0.014,
+      "step": 942
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 3.2916666666666666e-07,
+      "loss": 0.1665,
+      "step": 943
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 3.289855072463768e-07,
+      "loss": 0.0047,
+      "step": 944
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 3.2880434782608696e-07,
+      "loss": 0.0484,
+      "step": 945
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 3.2862318840579714e-07,
+      "loss": 0.3112,
+      "step": 946
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 3.284420289855072e-07,
+      "loss": 0.0516,
+      "step": 947
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 3.282608695652174e-07,
+      "loss": 0.0068,
+      "step": 948
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 3.280797101449275e-07,
+      "loss": 0.0087,
+      "step": 949
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 3.278985507246377e-07,
+      "loss": 0.6003,
+      "step": 950
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 3.277173913043478e-07,
+      "loss": 0.0105,
+      "step": 951
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 3.2753623188405794e-07,
+      "loss": 0.1512,
+      "step": 952
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 3.2735507246376807e-07,
+      "loss": 0.0855,
+      "step": 953
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 3.2717391304347824e-07,
+      "loss": 0.3914,
+      "step": 954
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 3.269927536231884e-07,
+      "loss": 0.1221,
+      "step": 955
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 3.2681159420289855e-07,
+      "loss": 0.0199,
+      "step": 956
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 3.2663043478260867e-07,
+      "loss": 0.0735,
+      "step": 957
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 3.264492753623188e-07,
+      "loss": 0.1051,
+      "step": 958
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 3.26268115942029e-07,
+      "loss": 0.0388,
+      "step": 959
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 3.260869565217391e-07,
+      "loss": 0.2412,
+      "step": 960
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 3.259057971014493e-07,
+      "loss": 0.0128,
+      "step": 961
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 3.257246376811594e-07,
+      "loss": 0.0126,
+      "step": 962
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 3.2554347826086953e-07,
+      "loss": 0.0799,
+      "step": 963
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 3.253623188405797e-07,
+      "loss": 0.0568,
+      "step": 964
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 3.2518115942028983e-07,
+      "loss": 0.0104,
+      "step": 965
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 3.25e-07,
+      "loss": 0.1761,
+      "step": 966
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 3.2481884057971014e-07,
+      "loss": 0.0236,
+      "step": 967
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 3.2463768115942026e-07,
+      "loss": 0.1802,
+      "step": 968
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 3.244565217391304e-07,
+      "loss": 0.0115,
+      "step": 969
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 3.2427536231884057e-07,
+      "loss": 0.1939,
+      "step": 970
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 3.240942028985507e-07,
+      "loss": 0.0074,
+      "step": 971
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 3.2391304347826087e-07,
+      "loss": 0.0954,
+      "step": 972
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 3.2373188405797105e-07,
+      "loss": 0.0072,
+      "step": 973
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 3.235507246376811e-07,
+      "loss": 0.0267,
+      "step": 974
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 3.233695652173913e-07,
+      "loss": 0.0368,
+      "step": 975
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 3.231884057971014e-07,
+      "loss": 0.0254,
+      "step": 976
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 3.230072463768116e-07,
+      "loss": 0.0995,
+      "step": 977
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 3.228260869565217e-07,
+      "loss": 0.0649,
+      "step": 978
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 3.2264492753623185e-07,
+      "loss": 0.0136,
+      "step": 979
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 3.2246376811594203e-07,
+      "loss": 0.1891,
+      "step": 980
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 3.2228260869565215e-07,
+      "loss": 0.1091,
+      "step": 981
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 3.2210144927536233e-07,
+      "loss": 0.0073,
+      "step": 982
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 3.2192028985507246e-07,
+      "loss": 0.0144,
+      "step": 983
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 3.217391304347826e-07,
+      "loss": 0.0338,
+      "step": 984
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 3.215579710144927e-07,
+      "loss": 0.0183,
+      "step": 985
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 3.213768115942029e-07,
+      "loss": 0.0132,
+      "step": 986
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 3.21195652173913e-07,
+      "loss": 0.0071,
+      "step": 987
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 3.210144927536232e-07,
+      "loss": 0.2068,
+      "step": 988
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 3.2083333333333337e-07,
+      "loss": 0.064,
+      "step": 989
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 3.2065217391304344e-07,
+      "loss": 0.2764,
+      "step": 990
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 3.204710144927536e-07,
+      "loss": 0.0113,
+      "step": 991
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 3.2028985507246374e-07,
+      "loss": 0.2342,
+      "step": 992
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 3.201086956521739e-07,
+      "loss": 0.0163,
+      "step": 993
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 3.1992753623188405e-07,
+      "loss": 0.0909,
+      "step": 994
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 3.1974637681159417e-07,
+      "loss": 0.0106,
+      "step": 995
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 3.195652173913043e-07,
+      "loss": 0.1551,
+      "step": 996
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 3.193840579710145e-07,
+      "loss": 0.2833,
+      "step": 997
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 3.1920289855072465e-07,
+      "loss": 0.0405,
+      "step": 998
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 3.190217391304348e-07,
+      "loss": 0.0541,
+      "step": 999
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 3.188405797101449e-07,
+      "loss": 0.0767,
+      "step": 1000
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 3.1865942028985503e-07,
+      "loss": 0.0112,
+      "step": 1001
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 3.184782608695652e-07,
+      "loss": 0.2512,
+      "step": 1002
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 3.1829710144927533e-07,
+      "loss": 0.2694,
+      "step": 1003
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 3.181159420289855e-07,
+      "loss": 0.4526,
+      "step": 1004
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 3.179347826086957e-07,
+      "loss": 0.0387,
+      "step": 1005
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 3.1775362318840576e-07,
+      "loss": 0.2418,
+      "step": 1006
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 3.1757246376811594e-07,
+      "loss": 0.1541,
+      "step": 1007
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 3.1739130434782606e-07,
+      "loss": 0.0737,
+      "step": 1008
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 3.1721014492753624e-07,
+      "loss": 0.1735,
+      "step": 1009
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 3.1702898550724637e-07,
+      "loss": 0.323,
+      "step": 1010
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 3.168478260869565e-07,
+      "loss": 0.0084,
+      "step": 1011
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 3.166666666666666e-07,
+      "loss": 0.0094,
+      "step": 1012
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 3.164855072463768e-07,
+      "loss": 0.0155,
+      "step": 1013
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 3.1630434782608697e-07,
+      "loss": 0.0096,
+      "step": 1014
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 3.161231884057971e-07,
+      "loss": 0.1935,
+      "step": 1015
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 3.159420289855073e-07,
+      "loss": 0.0086,
+      "step": 1016
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 3.1576086956521735e-07,
+      "loss": 0.0064,
+      "step": 1017
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 3.1557971014492753e-07,
+      "loss": 0.005,
+      "step": 1018
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 3.1539855072463765e-07,
+      "loss": 0.2494,
+      "step": 1019
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 3.1521739130434783e-07,
+      "loss": 0.0831,
+      "step": 1020
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 3.1503623188405796e-07,
+      "loss": 0.1613,
+      "step": 1021
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 3.148550724637681e-07,
+      "loss": 0.0037,
+      "step": 1022
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 3.1467391304347826e-07,
+      "loss": 0.0644,
+      "step": 1023
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 3.144927536231884e-07,
+      "loss": 0.2127,
+      "step": 1024
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 3.1431159420289856e-07,
+      "loss": 0.1774,
+      "step": 1025
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 3.141304347826087e-07,
+      "loss": 0.0194,
+      "step": 1026
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 3.139492753623188e-07,
+      "loss": 0.0199,
+      "step": 1027
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 3.1376811594202894e-07,
+      "loss": 0.0208,
+      "step": 1028
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 3.135869565217391e-07,
+      "loss": 0.0083,
+      "step": 1029
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 3.1340579710144924e-07,
+      "loss": 0.0438,
+      "step": 1030
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 3.132246376811594e-07,
+      "loss": 0.2676,
+      "step": 1031
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 3.130434782608696e-07,
+      "loss": 0.1898,
+      "step": 1032
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 3.1286231884057967e-07,
+      "loss": 0.0542,
+      "step": 1033
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 3.1268115942028985e-07,
+      "loss": 0.0226,
+      "step": 1034
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 3.1249999999999997e-07,
+      "loss": 0.0101,
+      "step": 1035
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 3.1231884057971015e-07,
+      "loss": 0.0191,
+      "step": 1036
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 3.121376811594203e-07,
+      "loss": 0.3766,
+      "step": 1037
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 3.119565217391304e-07,
+      "loss": 0.1336,
+      "step": 1038
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 3.117753623188405e-07,
+      "loss": 0.1327,
+      "step": 1039
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 3.115942028985507e-07,
+      "loss": 0.0069,
+      "step": 1040
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 3.114130434782609e-07,
+      "loss": 0.2647,
+      "step": 1041
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 3.11231884057971e-07,
+      "loss": 0.1891,
+      "step": 1042
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 3.110507246376812e-07,
+      "loss": 0.0113,
+      "step": 1043
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 3.1086956521739126e-07,
+      "loss": 0.3465,
+      "step": 1044
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 3.1068840579710144e-07,
+      "loss": 0.1512,
+      "step": 1045
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 3.1050724637681156e-07,
+      "loss": 0.0035,
+      "step": 1046
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 3.1032608695652174e-07,
+      "loss": 0.0996,
+      "step": 1047
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 3.101449275362319e-07,
+      "loss": 0.6926,
+      "step": 1048
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 3.09963768115942e-07,
+      "loss": 0.4219,
+      "step": 1049
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 3.0978260869565217e-07,
+      "loss": 0.3628,
+      "step": 1050
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 3.096014492753623e-07,
+      "loss": 0.6683,
+      "step": 1051
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 3.0942028985507247e-07,
+      "loss": 0.1532,
+      "step": 1052
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.092391304347826e-07,
+      "loss": 0.0478,
+      "step": 1053
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 3.090579710144927e-07,
+      "loss": 0.0235,
+      "step": 1054
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.0887681159420285e-07,
+      "loss": 0.0145,
+      "step": 1055
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 3.08695652173913e-07,
+      "loss": 0.075,
+      "step": 1056
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.085144927536232e-07,
+      "loss": 0.3058,
+      "step": 1057
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.0833333333333333e-07,
+      "loss": 0.0298,
+      "step": 1058
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 3.081521739130435e-07,
+      "loss": 0.0242,
+      "step": 1059
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 3.079710144927536e-07,
+      "loss": 0.0087,
+      "step": 1060
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 3.0778985507246376e-07,
+      "loss": 0.1052,
+      "step": 1061
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 3.076086956521739e-07,
+      "loss": 0.0172,
+      "step": 1062
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 3.0742753623188406e-07,
+      "loss": 0.015,
+      "step": 1063
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 3.072463768115942e-07,
+      "loss": 0.1115,
+      "step": 1064
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 3.070652173913043e-07,
+      "loss": 0.5237,
+      "step": 1065
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 3.068840579710145e-07,
+      "loss": 0.4865,
+      "step": 1066
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 3.067028985507246e-07,
+      "loss": 0.0475,
+      "step": 1067
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 3.065217391304348e-07,
+      "loss": 0.1189,
+      "step": 1068
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 3.063405797101449e-07,
+      "loss": 0.1729,
+      "step": 1069
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 3.061594202898551e-07,
+      "loss": 0.2113,
+      "step": 1070
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 3.0597826086956517e-07,
+      "loss": 0.1478,
+      "step": 1071
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 3.0579710144927535e-07,
+      "loss": 0.2319,
+      "step": 1072
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 3.0561594202898547e-07,
+      "loss": 0.2426,
+      "step": 1073
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 3.0543478260869565e-07,
+      "loss": 0.3024,
+      "step": 1074
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 3.0525362318840583e-07,
+      "loss": 0.0956,
+      "step": 1075
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 3.050724637681159e-07,
+      "loss": 0.6275,
+      "step": 1076
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 3.048913043478261e-07,
+      "loss": 0.0604,
+      "step": 1077
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 3.047101449275362e-07,
+      "loss": 0.0482,
+      "step": 1078
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 3.045289855072464e-07,
+      "loss": 0.0133,
+      "step": 1079
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 3.043478260869565e-07,
+      "loss": 0.1096,
+      "step": 1080
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 3.0416666666666663e-07,
+      "loss": 0.8055,
+      "step": 1081
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 3.0398550724637676e-07,
+      "loss": 0.376,
+      "step": 1082
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 3.0380434782608693e-07,
+      "loss": 0.0652,
+      "step": 1083
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 3.036231884057971e-07,
+      "loss": 0.4945,
+      "step": 1084
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 3.0344202898550724e-07,
+      "loss": 0.474,
+      "step": 1085
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 3.032608695652174e-07,
+      "loss": 0.0156,
+      "step": 1086
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 3.030797101449275e-07,
+      "loss": 0.0063,
+      "step": 1087
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 3.0289855072463767e-07,
+      "loss": 0.1155,
+      "step": 1088
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 3.027173913043478e-07,
+      "loss": 0.1641,
+      "step": 1089
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 3.0253623188405797e-07,
+      "loss": 0.3304,
+      "step": 1090
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 3.0235507246376815e-07,
+      "loss": 0.2717,
+      "step": 1091
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 3.021739130434782e-07,
+      "loss": 0.1411,
+      "step": 1092
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 3.019927536231884e-07,
+      "loss": 0.4309,
+      "step": 1093
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 3.018115942028985e-07,
+      "loss": 0.4501,
+      "step": 1094
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 3.016304347826087e-07,
+      "loss": 0.2649,
+      "step": 1095
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 3.014492753623188e-07,
+      "loss": 0.0091,
+      "step": 1096
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 3.0126811594202895e-07,
+      "loss": 0.179,
+      "step": 1097
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 3.010869565217391e-07,
+      "loss": 0.1972,
+      "step": 1098
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 3.0090579710144925e-07,
+      "loss": 0.1309,
+      "step": 1099
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 3.0072463768115943e-07,
+      "loss": 0.1497,
+      "step": 1100
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 3.0054347826086956e-07,
+      "loss": 0.0915,
+      "step": 1101
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 3.0036231884057974e-07,
+      "loss": 0.1156,
+      "step": 1102
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 3.001811594202898e-07,
+      "loss": 0.493,
+      "step": 1103
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 3e-07,
+      "loss": 0.0969,
+      "step": 1104
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 2.998188405797101e-07,
+      "loss": 0.0344,
+      "step": 1105
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 2.996376811594203e-07,
+      "loss": 0.008,
+      "step": 1106
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 2.994565217391304e-07,
+      "loss": 0.5317,
+      "step": 1107
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.9927536231884054e-07,
+      "loss": 0.007,
+      "step": 1108
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.990942028985507e-07,
+      "loss": 0.0064,
+      "step": 1109
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 2.9891304347826084e-07,
+      "loss": 0.0166,
+      "step": 1110
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 2.98731884057971e-07,
+      "loss": 0.0182,
+      "step": 1111
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 2.9855072463768115e-07,
+      "loss": 0.4441,
+      "step": 1112
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 2.983695652173913e-07,
+      "loss": 0.0152,
+      "step": 1113
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 2.981884057971014e-07,
+      "loss": 0.0636,
+      "step": 1114
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.980072463768116e-07,
+      "loss": 0.0558,
+      "step": 1115
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.9782608695652175e-07,
+      "loss": 0.379,
+      "step": 1116
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 2.976449275362319e-07,
+      "loss": 0.0099,
+      "step": 1117
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 2.9746376811594206e-07,
+      "loss": 0.063,
+      "step": 1118
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 2.9728260869565213e-07,
+      "loss": 0.0576,
+      "step": 1119
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 2.971014492753623e-07,
+      "loss": 0.1816,
+      "step": 1120
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 2.9692028985507243e-07,
+      "loss": 0.004,
+      "step": 1121
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.967391304347826e-07,
+      "loss": 0.7476,
+      "step": 1122
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.9655797101449274e-07,
+      "loss": 0.1144,
+      "step": 1123
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 2.9637681159420286e-07,
+      "loss": 0.5578,
+      "step": 1124
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 2.9619565217391304e-07,
+      "loss": 0.1301,
+      "step": 1125
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.9601449275362316e-07,
+      "loss": 0.8318,
+      "step": 1126
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.9583333333333334e-07,
+      "loss": 0.0126,
+      "step": 1127
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.9565217391304347e-07,
+      "loss": 0.1437,
+      "step": 1128
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.9547101449275365e-07,
+      "loss": 0.3581,
+      "step": 1129
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.952898550724637e-07,
+      "loss": 0.1987,
+      "step": 1130
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 2.951086956521739e-07,
+      "loss": 0.3525,
+      "step": 1131
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 2.94927536231884e-07,
+      "loss": 0.3636,
+      "step": 1132
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.947463768115942e-07,
+      "loss": 0.3876,
+      "step": 1133
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.945652173913044e-07,
+      "loss": 0.079,
+      "step": 1134
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.9438405797101445e-07,
+      "loss": 0.018,
+      "step": 1135
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.9420289855072463e-07,
+      "loss": 0.008,
+      "step": 1136
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.9402173913043475e-07,
+      "loss": 0.2247,
+      "step": 1137
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 2.9384057971014493e-07,
+      "loss": 0.0117,
+      "step": 1138
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 2.9365942028985506e-07,
+      "loss": 0.3542,
+      "step": 1139
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 2.9347826086956523e-07,
+      "loss": 0.1193,
+      "step": 1140
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 2.932971014492753e-07,
+      "loss": 0.1028,
+      "step": 1141
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.931159420289855e-07,
+      "loss": 0.2848,
+      "step": 1142
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.9293478260869566e-07,
+      "loss": 0.0045,
+      "step": 1143
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.927536231884058e-07,
+      "loss": 0.0935,
+      "step": 1144
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.9257246376811597e-07,
+      "loss": 0.0057,
+      "step": 1145
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.9239130434782604e-07,
+      "loss": 0.0076,
+      "step": 1146
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 2.922101449275362e-07,
+      "loss": 0.3515,
+      "step": 1147
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 2.9202898550724634e-07,
+      "loss": 0.0224,
+      "step": 1148
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.918478260869565e-07,
+      "loss": 0.3875,
+      "step": 1149
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.916666666666667e-07,
+      "loss": 1.324,
+      "step": 1150
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.8897058823529411,
+      "eval_combined_score": 0.9050299323269131,
+      "eval_f1": 0.9203539823008849,
+      "eval_loss": 0.38705751299858093,
+      "eval_runtime": 9.3861,
+      "eval_samples_per_second": 43.469,
+      "eval_steps_per_second": 2.77,
+      "step": 1150
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.9148550724637677e-07,
+      "loss": 0.0408,
+      "step": 1151
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 2.9130434782608695e-07,
+      "loss": 0.0305,
+      "step": 1152
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 2.9112318840579707e-07,
+      "loss": 0.3799,
+      "step": 1153
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 2.9094202898550725e-07,
+      "loss": 0.0044,
+      "step": 1154
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 2.907608695652174e-07,
+      "loss": 0.4001,
+      "step": 1155
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 2.9057971014492755e-07,
+      "loss": 0.1075,
+      "step": 1156
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 2.9039855072463763e-07,
+      "loss": 0.018,
+      "step": 1157
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 2.902173913043478e-07,
+      "loss": 0.2045,
+      "step": 1158
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 2.90036231884058e-07,
+      "loss": 0.0369,
+      "step": 1159
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 2.898550724637681e-07,
+      "loss": 0.224,
+      "step": 1160
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 2.896739130434783e-07,
+      "loss": 0.0558,
+      "step": 1161
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 2.8949275362318836e-07,
+      "loss": 0.3322,
+      "step": 1162
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 2.8931159420289854e-07,
+      "loss": 0.0122,
+      "step": 1163
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 2.8913043478260866e-07,
+      "loss": 0.0062,
+      "step": 1164
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.8894927536231884e-07,
+      "loss": 0.0671,
+      "step": 1165
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.8876811594202897e-07,
+      "loss": 0.0067,
+      "step": 1166
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.8858695652173914e-07,
+      "loss": 0.0211,
+      "step": 1167
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 2.8840579710144927e-07,
+      "loss": 0.01,
+      "step": 1168
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 2.882246376811594e-07,
+      "loss": 0.0666,
+      "step": 1169
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 2.8804347826086957e-07,
+      "loss": 0.5114,
+      "step": 1170
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 2.878623188405797e-07,
+      "loss": 0.0167,
+      "step": 1171
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.876811594202899e-07,
+      "loss": 0.0266,
+      "step": 1172
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.8749999999999995e-07,
+      "loss": 0.151,
+      "step": 1173
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.873188405797101e-07,
+      "loss": 0.1577,
+      "step": 1174
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 2.8713768115942025e-07,
+      "loss": 0.0191,
+      "step": 1175
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 2.8695652173913043e-07,
+      "loss": 0.0098,
+      "step": 1176
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 2.867753623188406e-07,
+      "loss": 0.0061,
+      "step": 1177
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 2.865942028985507e-07,
+      "loss": 0.0339,
+      "step": 1178
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 2.8641304347826086e-07,
+      "loss": 0.0258,
+      "step": 1179
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 2.86231884057971e-07,
+      "loss": 0.2181,
+      "step": 1180
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 2.8605072463768116e-07,
+      "loss": 0.1308,
+      "step": 1181
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 2.858695652173913e-07,
+      "loss": 0.1897,
+      "step": 1182
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 2.8568840579710146e-07,
+      "loss": 0.0992,
+      "step": 1183
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 2.8550724637681154e-07,
+      "loss": 0.1294,
+      "step": 1184
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 2.853260869565217e-07,
+      "loss": 0.0149,
+      "step": 1185
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 2.851449275362319e-07,
+      "loss": 0.1364,
+      "step": 1186
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 2.84963768115942e-07,
+      "loss": 0.0427,
+      "step": 1187
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 2.847826086956522e-07,
+      "loss": 0.1179,
+      "step": 1188
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 2.8460144927536227e-07,
+      "loss": 0.2878,
+      "step": 1189
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 2.8442028985507245e-07,
+      "loss": 0.0133,
+      "step": 1190
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 2.8423913043478257e-07,
+      "loss": 0.0186,
+      "step": 1191
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 2.8405797101449275e-07,
+      "loss": 0.0092,
+      "step": 1192
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 2.8387681159420293e-07,
+      "loss": 0.4775,
+      "step": 1193
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 2.8369565217391305e-07,
+      "loss": 0.4026,
+      "step": 1194
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 2.835144927536232e-07,
+      "loss": 0.1284,
+      "step": 1195
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 2.833333333333333e-07,
+      "loss": 0.1387,
+      "step": 1196
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 2.831521739130435e-07,
+      "loss": 0.3526,
+      "step": 1197
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 2.829710144927536e-07,
+      "loss": 0.008,
+      "step": 1198
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 2.827898550724638e-07,
+      "loss": 0.2757,
+      "step": 1199
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 2.8260869565217386e-07,
+      "loss": 0.2317,
+      "step": 1200
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 2.8242753623188403e-07,
+      "loss": 0.0099,
+      "step": 1201
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 2.822463768115942e-07,
+      "loss": 0.0711,
+      "step": 1202
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 2.8206521739130434e-07,
+      "loss": 0.0075,
+      "step": 1203
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 2.818840579710145e-07,
+      "loss": 0.0336,
+      "step": 1204
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 2.817028985507246e-07,
+      "loss": 0.0429,
+      "step": 1205
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 2.8152173913043477e-07,
+      "loss": 0.2111,
+      "step": 1206
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 2.813405797101449e-07,
+      "loss": 0.106,
+      "step": 1207
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 2.8115942028985507e-07,
+      "loss": 0.0038,
+      "step": 1208
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 2.809782608695652e-07,
+      "loss": 0.3583,
+      "step": 1209
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 2.8079710144927537e-07,
+      "loss": 0.0119,
+      "step": 1210
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.806159420289855e-07,
+      "loss": 0.1366,
+      "step": 1211
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.804347826086956e-07,
+      "loss": 0.0072,
+      "step": 1212
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.802536231884058e-07,
+      "loss": 0.2164,
+      "step": 1213
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 2.8007246376811593e-07,
+      "loss": 0.3016,
+      "step": 1214
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 2.798913043478261e-07,
+      "loss": 0.0134,
+      "step": 1215
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 2.797101449275362e-07,
+      "loss": 0.2477,
+      "step": 1216
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 2.7952898550724636e-07,
+      "loss": 0.6198,
+      "step": 1217
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 2.793478260869565e-07,
+      "loss": 0.0163,
+      "step": 1218
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 2.7916666666666666e-07,
+      "loss": 0.0321,
+      "step": 1219
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 2.7898550724637684e-07,
+      "loss": 0.1752,
+      "step": 1220
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 2.788043478260869e-07,
+      "loss": 0.1418,
+      "step": 1221
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 2.786231884057971e-07,
+      "loss": 0.4174,
+      "step": 1222
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 2.784420289855072e-07,
+      "loss": 0.4245,
+      "step": 1223
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 2.782608695652174e-07,
+      "loss": 0.2839,
+      "step": 1224
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.780797101449275e-07,
+      "loss": 0.0062,
+      "step": 1225
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.778985507246377e-07,
+      "loss": 0.0264,
+      "step": 1226
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.777173913043478e-07,
+      "loss": 0.2683,
+      "step": 1227
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 2.7753623188405794e-07,
+      "loss": 0.01,
+      "step": 1228
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 2.773550724637681e-07,
+      "loss": 0.0092,
+      "step": 1229
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 2.7717391304347825e-07,
+      "loss": 0.0172,
+      "step": 1230
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 2.769927536231884e-07,
+      "loss": 0.3386,
+      "step": 1231
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 2.768115942028985e-07,
+      "loss": 0.2101,
+      "step": 1232
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 2.766304347826087e-07,
+      "loss": 0.5643,
+      "step": 1233
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 2.764492753623188e-07,
+      "loss": 0.052,
+      "step": 1234
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 2.76268115942029e-07,
+      "loss": 0.3403,
+      "step": 1235
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 2.7608695652173916e-07,
+      "loss": 0.072,
+      "step": 1236
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 2.759057971014493e-07,
+      "loss": 0.1012,
+      "step": 1237
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 2.757246376811594e-07,
+      "loss": 0.1471,
+      "step": 1238
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 2.7554347826086953e-07,
+      "loss": 0.0067,
+      "step": 1239
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 2.753623188405797e-07,
+      "loss": 0.9499,
+      "step": 1240
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 2.7518115942028984e-07,
+      "loss": 0.0589,
+      "step": 1241
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 2.75e-07,
+      "loss": 0.0321,
+      "step": 1242
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 2.748188405797101e-07,
+      "loss": 0.0075,
+      "step": 1243
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 2.7463768115942026e-07,
+      "loss": 0.0486,
+      "step": 1244
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 2.7445652173913044e-07,
+      "loss": 0.0634,
+      "step": 1245
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 2.7427536231884057e-07,
+      "loss": 0.22,
+      "step": 1246
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 2.7409420289855075e-07,
+      "loss": 0.2852,
+      "step": 1247
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 2.739130434782608e-07,
+      "loss": 0.006,
+      "step": 1248
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 2.73731884057971e-07,
+      "loss": 0.1862,
+      "step": 1249
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 2.735507246376811e-07,
+      "loss": 0.0375,
+      "step": 1250
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.733695652173913e-07,
+      "loss": 0.0884,
+      "step": 1251
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.731884057971014e-07,
+      "loss": 0.0977,
+      "step": 1252
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 2.730072463768116e-07,
+      "loss": 0.2483,
+      "step": 1253
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 2.7282608695652173e-07,
+      "loss": 0.0089,
+      "step": 1254
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 2.7264492753623185e-07,
+      "loss": 0.0228,
+      "step": 1255
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 2.7246376811594203e-07,
+      "loss": 0.0231,
+      "step": 1256
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 2.7228260869565216e-07,
+      "loss": 0.0146,
+      "step": 1257
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 2.7210144927536233e-07,
+      "loss": 0.0064,
+      "step": 1258
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 2.719202898550724e-07,
+      "loss": 0.0112,
+      "step": 1259
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 2.717391304347826e-07,
+      "loss": 0.0066,
+      "step": 1260
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 2.7155797101449276e-07,
+      "loss": 0.249,
+      "step": 1261
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 2.713768115942029e-07,
+      "loss": 0.0207,
+      "step": 1262
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 2.7119565217391307e-07,
+      "loss": 0.2571,
+      "step": 1263
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.710144927536232e-07,
+      "loss": 0.1773,
+      "step": 1264
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.708333333333333e-07,
+      "loss": 0.5297,
+      "step": 1265
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.7065217391304344e-07,
+      "loss": 0.0497,
+      "step": 1266
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 2.704710144927536e-07,
+      "loss": 0.0479,
+      "step": 1267
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 2.7028985507246375e-07,
+      "loss": 0.6137,
+      "step": 1268
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 2.701086956521739e-07,
+      "loss": 0.0985,
+      "step": 1269
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 2.6992753623188405e-07,
+      "loss": 0.1527,
+      "step": 1270
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 2.697463768115942e-07,
+      "loss": 0.0981,
+      "step": 1271
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 2.6956521739130435e-07,
+      "loss": 0.3516,
+      "step": 1272
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 2.693840579710145e-07,
+      "loss": 0.0336,
+      "step": 1273
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 2.6920289855072466e-07,
+      "loss": 0.0841,
+      "step": 1274
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 2.6902173913043473e-07,
+      "loss": 0.2298,
+      "step": 1275
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 2.688405797101449e-07,
+      "loss": 0.0413,
+      "step": 1276
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 2.6865942028985503e-07,
+      "loss": 0.4771,
+      "step": 1277
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 2.684782608695652e-07,
+      "loss": 0.0454,
+      "step": 1278
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 2.682971014492754e-07,
+      "loss": 0.0812,
+      "step": 1279
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 2.681159420289855e-07,
+      "loss": 0.24,
+      "step": 1280
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 2.6793478260869564e-07,
+      "loss": 0.3943,
+      "step": 1281
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 2.6775362318840576e-07,
+      "loss": 0.0283,
+      "step": 1282
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 2.6757246376811594e-07,
+      "loss": 0.0097,
+      "step": 1283
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 2.6739130434782607e-07,
+      "loss": 0.0155,
+      "step": 1284
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 2.6721014492753624e-07,
+      "loss": 0.2408,
+      "step": 1285
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 2.670289855072463e-07,
+      "loss": 0.0561,
+      "step": 1286
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 2.668478260869565e-07,
+      "loss": 0.5175,
+      "step": 1287
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 2.6666666666666667e-07,
+      "loss": 0.0084,
+      "step": 1288
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 2.664855072463768e-07,
+      "loss": 0.1346,
+      "step": 1289
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 2.66304347826087e-07,
+      "loss": 0.1784,
+      "step": 1290
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 2.661231884057971e-07,
+      "loss": 0.2247,
+      "step": 1291
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 2.659420289855072e-07,
+      "loss": 0.1441,
+      "step": 1292
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 2.6576086956521735e-07,
+      "loss": 0.0112,
+      "step": 1293
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 2.6557971014492753e-07,
+      "loss": 0.2246,
+      "step": 1294
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 2.653985507246377e-07,
+      "loss": 0.3938,
+      "step": 1295
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 2.6521739130434783e-07,
+      "loss": 0.0126,
+      "step": 1296
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 2.6503623188405796e-07,
+      "loss": 0.024,
+      "step": 1297
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 2.648550724637681e-07,
+      "loss": 0.5345,
+      "step": 1298
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 2.6467391304347826e-07,
+      "loss": 0.0473,
+      "step": 1299
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 2.644927536231884e-07,
+      "loss": 0.4863,
+      "step": 1300
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 2.6431159420289856e-07,
+      "loss": 0.1661,
+      "step": 1301
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 2.6413043478260864e-07,
+      "loss": 0.0688,
+      "step": 1302
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 2.639492753623188e-07,
+      "loss": 0.2084,
+      "step": 1303
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 2.63768115942029e-07,
+      "loss": 0.0143,
+      "step": 1304
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 2.635869565217391e-07,
+      "loss": 0.3095,
+      "step": 1305
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 2.634057971014493e-07,
+      "loss": 0.0357,
+      "step": 1306
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 2.632246376811594e-07,
+      "loss": 0.0482,
+      "step": 1307
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 2.6304347826086955e-07,
+      "loss": 0.0764,
+      "step": 1308
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 2.6286231884057967e-07,
+      "loss": 0.1098,
+      "step": 1309
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 2.6268115942028985e-07,
+      "loss": 0.2982,
+      "step": 1310
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 2.625e-07,
+      "loss": 0.1211,
+      "step": 1311
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 2.6231884057971015e-07,
+      "loss": 0.0142,
+      "step": 1312
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 2.621376811594203e-07,
+      "loss": 0.058,
+      "step": 1313
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 2.619565217391304e-07,
+      "loss": 0.2052,
+      "step": 1314
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 2.617753623188406e-07,
+      "loss": 0.0434,
+      "step": 1315
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 2.615942028985507e-07,
+      "loss": 0.3418,
+      "step": 1316
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 2.614130434782609e-07,
+      "loss": 0.2306,
+      "step": 1317
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 2.61231884057971e-07,
+      "loss": 0.2538,
+      "step": 1318
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 2.6105072463768114e-07,
+      "loss": 0.3121,
+      "step": 1319
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 2.6086956521739126e-07,
+      "loss": 0.2583,
+      "step": 1320
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 2.6068840579710144e-07,
+      "loss": 0.3814,
+      "step": 1321
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 2.605072463768116e-07,
+      "loss": 0.0531,
+      "step": 1322
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 2.6032608695652174e-07,
+      "loss": 0.0353,
+      "step": 1323
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 2.6014492753623187e-07,
+      "loss": 0.1002,
+      "step": 1324
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 2.59963768115942e-07,
+      "loss": 0.3467,
+      "step": 1325
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 2.5978260869565217e-07,
+      "loss": 0.0269,
+      "step": 1326
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 2.596014492753623e-07,
+      "loss": 0.2959,
+      "step": 1327
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 2.5942028985507247e-07,
+      "loss": 0.0466,
+      "step": 1328
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 2.5923913043478255e-07,
+      "loss": 0.0306,
+      "step": 1329
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 2.590579710144927e-07,
+      "loss": 0.588,
+      "step": 1330
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 2.588768115942029e-07,
+      "loss": 0.0603,
+      "step": 1331
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 2.5869565217391303e-07,
+      "loss": 0.1414,
+      "step": 1332
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 2.585144927536232e-07,
+      "loss": 0.0259,
+      "step": 1333
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 2.5833333333333333e-07,
+      "loss": 0.1491,
+      "step": 1334
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 2.5815217391304346e-07,
+      "loss": 0.0099,
+      "step": 1335
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 2.579710144927536e-07,
+      "loss": 0.5819,
+      "step": 1336
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 2.5778985507246376e-07,
+      "loss": 0.8519,
+      "step": 1337
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 2.5760869565217394e-07,
+      "loss": 0.169,
+      "step": 1338
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 2.5742753623188406e-07,
+      "loss": 0.1375,
+      "step": 1339
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 2.572463768115942e-07,
+      "loss": 0.0209,
+      "step": 1340
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 2.570652173913043e-07,
+      "loss": 0.4712,
+      "step": 1341
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 2.568840579710145e-07,
+      "loss": 0.0178,
+      "step": 1342
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 2.567028985507246e-07,
+      "loss": 0.0653,
+      "step": 1343
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 2.565217391304348e-07,
+      "loss": 0.0251,
+      "step": 1344
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 2.5634057971014487e-07,
+      "loss": 0.0312,
+      "step": 1345
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 2.5615942028985504e-07,
+      "loss": 0.3878,
+      "step": 1346
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 2.559782608695652e-07,
+      "loss": 0.006,
+      "step": 1347
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 2.5579710144927535e-07,
+      "loss": 0.0607,
+      "step": 1348
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 2.556159420289855e-07,
+      "loss": 0.0189,
+      "step": 1349
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 2.5543478260869565e-07,
+      "loss": 0.0073,
+      "step": 1350
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 2.552536231884058e-07,
+      "loss": 0.0518,
+      "step": 1351
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 2.550724637681159e-07,
+      "loss": 0.1715,
+      "step": 1352
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 2.548913043478261e-07,
+      "loss": 0.122,
+      "step": 1353
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 2.547101449275362e-07,
+      "loss": 0.0323,
+      "step": 1354
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 2.545289855072464e-07,
+      "loss": 0.0141,
+      "step": 1355
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 2.543478260869565e-07,
+      "loss": 0.0274,
+      "step": 1356
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 2.5416666666666663e-07,
+      "loss": 0.0105,
+      "step": 1357
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 2.539855072463768e-07,
+      "loss": 0.005,
+      "step": 1358
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 2.5380434782608694e-07,
+      "loss": 0.0366,
+      "step": 1359
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 2.536231884057971e-07,
+      "loss": 0.2073,
+      "step": 1360
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 2.5344202898550724e-07,
+      "loss": 0.2838,
+      "step": 1361
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 2.5326086956521736e-07,
+      "loss": 0.0619,
+      "step": 1362
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 2.530797101449275e-07,
+      "loss": 0.0209,
+      "step": 1363
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 2.5289855072463767e-07,
+      "loss": 0.1411,
+      "step": 1364
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 2.5271739130434785e-07,
+      "loss": 0.6116,
+      "step": 1365
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 2.5253623188405797e-07,
+      "loss": 0.0954,
+      "step": 1366
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 2.523550724637681e-07,
+      "loss": 0.1679,
+      "step": 1367
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 2.521739130434782e-07,
+      "loss": 0.266,
+      "step": 1368
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 2.519927536231884e-07,
+      "loss": 0.3313,
+      "step": 1369
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 2.518115942028985e-07,
+      "loss": 0.0181,
+      "step": 1370
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 2.516304347826087e-07,
+      "loss": 0.0205,
+      "step": 1371
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 2.5144927536231883e-07,
+      "loss": 0.0253,
+      "step": 1372
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 2.5126811594202895e-07,
+      "loss": 0.2875,
+      "step": 1373
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 2.5108695652173913e-07,
+      "loss": 0.3063,
+      "step": 1374
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 2.5090579710144926e-07,
+      "loss": 0.2231,
+      "step": 1375
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 2.5072463768115944e-07,
+      "loss": 0.0423,
+      "step": 1376
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 2.5054347826086956e-07,
+      "loss": 0.071,
+      "step": 1377
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 2.503623188405797e-07,
+      "loss": 0.0221,
+      "step": 1378
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.501811594202898e-07,
+      "loss": 0.0639,
+      "step": 1379
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.5e-07,
+      "loss": 0.0227,
+      "step": 1380
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.8897058823529411,
+      "eval_combined_score": 0.9048884651551561,
+      "eval_f1": 0.9200710479573712,
+      "eval_loss": 0.3950679898262024,
+      "eval_runtime": 8.775,
+      "eval_samples_per_second": 46.496,
+      "eval_steps_per_second": 2.963,
+      "step": 1380
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.4981884057971017e-07,
+      "loss": 0.2332,
+      "step": 1381
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 2.496376811594203e-07,
+      "loss": 0.0563,
+      "step": 1382
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 2.494565217391304e-07,
+      "loss": 0.3695,
+      "step": 1383
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 2.492753623188406e-07,
+      "loss": 0.2653,
+      "step": 1384
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 2.490942028985507e-07,
+      "loss": 0.0138,
+      "step": 1385
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 2.4891304347826085e-07,
+      "loss": 0.0194,
+      "step": 1386
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 2.4873188405797097e-07,
+      "loss": 0.2302,
+      "step": 1387
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 2.4855072463768115e-07,
+      "loss": 0.1392,
+      "step": 1388
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 2.4836956521739133e-07,
+      "loss": 0.0354,
+      "step": 1389
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 2.4818840579710145e-07,
+      "loss": 0.3695,
+      "step": 1390
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 2.480072463768116e-07,
+      "loss": 0.0086,
+      "step": 1391
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 2.4782608695652176e-07,
+      "loss": 0.0278,
+      "step": 1392
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 2.476449275362319e-07,
+      "loss": 0.2141,
+      "step": 1393
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 2.47463768115942e-07,
+      "loss": 0.0257,
+      "step": 1394
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 2.4728260869565213e-07,
+      "loss": 0.276,
+      "step": 1395
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 2.471014492753623e-07,
+      "loss": 0.0067,
+      "step": 1396
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 2.469202898550725e-07,
+      "loss": 0.3242,
+      "step": 1397
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 2.467391304347826e-07,
+      "loss": 0.0203,
+      "step": 1398
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 2.4655797101449274e-07,
+      "loss": 0.0237,
+      "step": 1399
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 2.463768115942029e-07,
+      "loss": 0.0203,
+      "step": 1400
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 2.4619565217391304e-07,
+      "loss": 0.0065,
+      "step": 1401
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 2.4601449275362317e-07,
+      "loss": 0.3641,
+      "step": 1402
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 2.458333333333333e-07,
+      "loss": 0.0171,
+      "step": 1403
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 2.4565217391304347e-07,
+      "loss": 0.116,
+      "step": 1404
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 2.454710144927536e-07,
+      "loss": 0.3526,
+      "step": 1405
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 2.4528985507246377e-07,
+      "loss": 0.0171,
+      "step": 1406
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 2.451086956521739e-07,
+      "loss": 0.3775,
+      "step": 1407
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 2.449275362318841e-07,
+      "loss": 0.0088,
+      "step": 1408
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 2.447463768115942e-07,
+      "loss": 0.0837,
+      "step": 1409
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 2.445652173913043e-07,
+      "loss": 0.2256,
+      "step": 1410
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 2.4438405797101445e-07,
+      "loss": 0.2489,
+      "step": 1411
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 2.4420289855072463e-07,
+      "loss": 0.012,
+      "step": 1412
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 2.4402173913043475e-07,
+      "loss": 0.0084,
+      "step": 1413
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 2.438405797101449e-07,
+      "loss": 0.0089,
+      "step": 1414
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 2.4365942028985506e-07,
+      "loss": 0.0228,
+      "step": 1415
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 2.4347826086956524e-07,
+      "loss": 0.2783,
+      "step": 1416
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 2.4329710144927536e-07,
+      "loss": 0.1131,
+      "step": 1417
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 2.431159420289855e-07,
+      "loss": 0.0499,
+      "step": 1418
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 2.4293478260869566e-07,
+      "loss": 0.0782,
+      "step": 1419
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 2.427536231884058e-07,
+      "loss": 0.0033,
+      "step": 1420
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 2.425724637681159e-07,
+      "loss": 0.0411,
+      "step": 1421
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 2.4239130434782604e-07,
+      "loss": 0.0114,
+      "step": 1422
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 2.422101449275362e-07,
+      "loss": 0.3854,
+      "step": 1423
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 2.420289855072464e-07,
+      "loss": 0.0349,
+      "step": 1424
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 2.418478260869565e-07,
+      "loss": 0.0128,
+      "step": 1425
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 2.4166666666666665e-07,
+      "loss": 0.1169,
+      "step": 1426
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 2.414855072463768e-07,
+      "loss": 0.1662,
+      "step": 1427
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 2.4130434782608695e-07,
+      "loss": 0.0112,
+      "step": 1428
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 2.411231884057971e-07,
+      "loss": 0.4801,
+      "step": 1429
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 2.409420289855072e-07,
+      "loss": 0.241,
+      "step": 1430
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 2.407608695652174e-07,
+      "loss": 0.2264,
+      "step": 1431
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 2.4057971014492756e-07,
+      "loss": 0.3174,
+      "step": 1432
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 2.403985507246377e-07,
+      "loss": 0.0397,
+      "step": 1433
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 2.402173913043478e-07,
+      "loss": 0.0652,
+      "step": 1434
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 2.40036231884058e-07,
+      "loss": 0.0948,
+      "step": 1435
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 2.398550724637681e-07,
+      "loss": 0.032,
+      "step": 1436
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 2.3967391304347824e-07,
+      "loss": 0.4069,
+      "step": 1437
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 2.3949275362318836e-07,
+      "loss": 0.0509,
+      "step": 1438
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 2.3931159420289854e-07,
+      "loss": 0.0681,
+      "step": 1439
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 2.391304347826087e-07,
+      "loss": 0.0398,
+      "step": 1440
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 2.3894927536231884e-07,
+      "loss": 0.3615,
+      "step": 1441
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 2.3876811594202897e-07,
+      "loss": 0.1264,
+      "step": 1442
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 2.3858695652173915e-07,
+      "loss": 0.6726,
+      "step": 1443
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 2.3840579710144927e-07,
+      "loss": 0.3551,
+      "step": 1444
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 2.3822463768115942e-07,
+      "loss": 0.8274,
+      "step": 1445
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 2.3804347826086957e-07,
+      "loss": 0.0098,
+      "step": 1446
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 2.378623188405797e-07,
+      "loss": 0.3142,
+      "step": 1447
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 2.3768115942028985e-07,
+      "loss": 0.0154,
+      "step": 1448
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 2.3749999999999998e-07,
+      "loss": 0.0211,
+      "step": 1449
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 2.3731884057971013e-07,
+      "loss": 0.206,
+      "step": 1450
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 2.3713768115942028e-07,
+      "loss": 0.1029,
+      "step": 1451
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 2.369565217391304e-07,
+      "loss": 0.0281,
+      "step": 1452
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 2.3677536231884058e-07,
+      "loss": 0.0439,
+      "step": 1453
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 2.3659420289855073e-07,
+      "loss": 0.1739,
+      "step": 1454
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 2.3641304347826086e-07,
+      "loss": 0.0359,
+      "step": 1455
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 2.36231884057971e-07,
+      "loss": 0.0095,
+      "step": 1456
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 2.3605072463768114e-07,
+      "loss": 0.0078,
+      "step": 1457
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 2.358695652173913e-07,
+      "loss": 0.1336,
+      "step": 1458
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 2.3568840579710144e-07,
+      "loss": 0.2985,
+      "step": 1459
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 2.3550724637681156e-07,
+      "loss": 0.5205,
+      "step": 1460
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 2.3532608695652174e-07,
+      "loss": 0.025,
+      "step": 1461
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 2.351449275362319e-07,
+      "loss": 0.121,
+      "step": 1462
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 2.3496376811594202e-07,
+      "loss": 0.1871,
+      "step": 1463
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 2.3478260869565217e-07,
+      "loss": 0.0321,
+      "step": 1464
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 2.346014492753623e-07,
+      "loss": 0.2441,
+      "step": 1465
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 2.3442028985507245e-07,
+      "loss": 0.1726,
+      "step": 1466
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 2.342391304347826e-07,
+      "loss": 0.0756,
+      "step": 1467
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 2.3405797101449273e-07,
+      "loss": 0.1389,
+      "step": 1468
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 2.3387681159420288e-07,
+      "loss": 0.0074,
+      "step": 1469
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 2.3369565217391305e-07,
+      "loss": 0.3997,
+      "step": 1470
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 2.3351449275362318e-07,
+      "loss": 0.2251,
+      "step": 1471
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 2.3333333333333333e-07,
+      "loss": 0.6389,
+      "step": 1472
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 2.3315217391304346e-07,
+      "loss": 0.0115,
+      "step": 1473
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 2.329710144927536e-07,
+      "loss": 0.2778,
+      "step": 1474
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 2.3278985507246376e-07,
+      "loss": 0.0257,
+      "step": 1475
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 2.3260869565217389e-07,
+      "loss": 0.105,
+      "step": 1476
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 2.3242753623188404e-07,
+      "loss": 0.1022,
+      "step": 1477
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 2.3224637681159422e-07,
+      "loss": 0.0145,
+      "step": 1478
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 2.3206521739130434e-07,
+      "loss": 0.1123,
+      "step": 1479
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 2.318840579710145e-07,
+      "loss": 0.0596,
+      "step": 1480
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 2.3170289855072464e-07,
+      "loss": 0.0808,
+      "step": 1481
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 2.3152173913043477e-07,
+      "loss": 0.009,
+      "step": 1482
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 2.3134057971014492e-07,
+      "loss": 0.0105,
+      "step": 1483
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 2.3115942028985505e-07,
+      "loss": 0.1346,
+      "step": 1484
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 2.309782608695652e-07,
+      "loss": 0.0675,
+      "step": 1485
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 2.3079710144927535e-07,
+      "loss": 0.0284,
+      "step": 1486
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 2.306159420289855e-07,
+      "loss": 0.2602,
+      "step": 1487
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 2.3043478260869565e-07,
+      "loss": 0.3282,
+      "step": 1488
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 2.302536231884058e-07,
+      "loss": 0.0101,
+      "step": 1489
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 2.3007246376811593e-07,
+      "loss": 0.0209,
+      "step": 1490
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 2.2989130434782608e-07,
+      "loss": 0.3097,
+      "step": 1491
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 2.297101449275362e-07,
+      "loss": 0.0465,
+      "step": 1492
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 2.2952898550724636e-07,
+      "loss": 0.0227,
+      "step": 1493
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 2.293478260869565e-07,
+      "loss": 0.2238,
+      "step": 1494
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 2.2916666666666663e-07,
+      "loss": 0.0176,
+      "step": 1495
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 2.289855072463768e-07,
+      "loss": 0.0342,
+      "step": 1496
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 2.2880434782608696e-07,
+      "loss": 0.229,
+      "step": 1497
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 2.286231884057971e-07,
+      "loss": 0.0168,
+      "step": 1498
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 2.2844202898550724e-07,
+      "loss": 0.0716,
+      "step": 1499
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 2.2826086956521737e-07,
+      "loss": 0.2004,
+      "step": 1500
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 2.2807971014492752e-07,
+      "loss": 0.0855,
+      "step": 1501
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 2.2789855072463767e-07,
+      "loss": 0.0527,
+      "step": 1502
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 2.277173913043478e-07,
+      "loss": 0.3369,
+      "step": 1503
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 2.2753623188405797e-07,
+      "loss": 0.0281,
+      "step": 1504
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 2.2735507246376812e-07,
+      "loss": 0.0393,
+      "step": 1505
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 2.2717391304347825e-07,
+      "loss": 0.1753,
+      "step": 1506
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 2.269927536231884e-07,
+      "loss": 0.2146,
+      "step": 1507
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 2.2681159420289855e-07,
+      "loss": 0.0111,
+      "step": 1508
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 2.2663043478260868e-07,
+      "loss": 0.1614,
+      "step": 1509
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 2.2644927536231883e-07,
+      "loss": 0.0167,
+      "step": 1510
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 2.2626811594202895e-07,
+      "loss": 0.205,
+      "step": 1511
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 2.260869565217391e-07,
+      "loss": 0.195,
+      "step": 1512
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 2.2590579710144928e-07,
+      "loss": 0.0075,
+      "step": 1513
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 2.257246376811594e-07,
+      "loss": 0.014,
+      "step": 1514
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 2.2554347826086956e-07,
+      "loss": 0.2985,
+      "step": 1515
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 2.253623188405797e-07,
+      "loss": 0.0442,
+      "step": 1516
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 2.2518115942028984e-07,
+      "loss": 0.5812,
+      "step": 1517
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 2.25e-07,
+      "loss": 0.2859,
+      "step": 1518
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 2.2481884057971011e-07,
+      "loss": 0.0091,
+      "step": 1519
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 2.2463768115942027e-07,
+      "loss": 0.5299,
+      "step": 1520
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 2.2445652173913044e-07,
+      "loss": 0.3824,
+      "step": 1521
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 2.2427536231884057e-07,
+      "loss": 0.0076,
+      "step": 1522
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 2.2409420289855072e-07,
+      "loss": 0.2998,
+      "step": 1523
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 2.2391304347826087e-07,
+      "loss": 0.4091,
+      "step": 1524
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 2.23731884057971e-07,
+      "loss": 0.041,
+      "step": 1525
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 2.2355072463768115e-07,
+      "loss": 0.0255,
+      "step": 1526
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 2.2336956521739128e-07,
+      "loss": 0.6094,
+      "step": 1527
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 2.2318840579710143e-07,
+      "loss": 0.0814,
+      "step": 1528
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 2.230072463768116e-07,
+      "loss": 0.0175,
+      "step": 1529
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 2.2282608695652173e-07,
+      "loss": 0.2616,
+      "step": 1530
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 2.2264492753623188e-07,
+      "loss": 0.5975,
+      "step": 1531
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 2.2246376811594203e-07,
+      "loss": 0.3662,
+      "step": 1532
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 2.2228260869565216e-07,
+      "loss": 0.0961,
+      "step": 1533
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 2.221014492753623e-07,
+      "loss": 0.0584,
+      "step": 1534
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 2.2192028985507244e-07,
+      "loss": 0.0796,
+      "step": 1535
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 2.217391304347826e-07,
+      "loss": 0.0062,
+      "step": 1536
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 2.2155797101449274e-07,
+      "loss": 0.175,
+      "step": 1537
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 2.213768115942029e-07,
+      "loss": 0.0177,
+      "step": 1538
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 2.2119565217391304e-07,
+      "loss": 0.0201,
+      "step": 1539
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 2.210144927536232e-07,
+      "loss": 0.0089,
+      "step": 1540
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 2.2083333333333332e-07,
+      "loss": 0.0817,
+      "step": 1541
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 2.2065217391304347e-07,
+      "loss": 0.3409,
+      "step": 1542
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 2.2047101449275362e-07,
+      "loss": 0.0121,
+      "step": 1543
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 2.2028985507246375e-07,
+      "loss": 0.0097,
+      "step": 1544
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 2.201086956521739e-07,
+      "loss": 0.194,
+      "step": 1545
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 2.1992753623188402e-07,
+      "loss": 0.0318,
+      "step": 1546
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 2.197463768115942e-07,
+      "loss": 0.206,
+      "step": 1547
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 2.1956521739130435e-07,
+      "loss": 0.0531,
+      "step": 1548
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 2.1938405797101448e-07,
+      "loss": 0.0263,
+      "step": 1549
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 2.1920289855072463e-07,
+      "loss": 0.351,
+      "step": 1550
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 2.1902173913043478e-07,
+      "loss": 0.5096,
+      "step": 1551
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 2.188405797101449e-07,
+      "loss": 0.0303,
+      "step": 1552
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 2.1865942028985506e-07,
+      "loss": 0.3943,
+      "step": 1553
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 2.1847826086956518e-07,
+      "loss": 0.0754,
+      "step": 1554
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 2.1829710144927536e-07,
+      "loss": 0.4236,
+      "step": 1555
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 2.1811594202898551e-07,
+      "loss": 0.1951,
+      "step": 1556
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 2.1793478260869564e-07,
+      "loss": 0.1875,
+      "step": 1557
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 2.177536231884058e-07,
+      "loss": 0.3761,
+      "step": 1558
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 2.1757246376811594e-07,
+      "loss": 0.0477,
+      "step": 1559
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 2.1739130434782607e-07,
+      "loss": 0.2532,
+      "step": 1560
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 2.1721014492753622e-07,
+      "loss": 0.1149,
+      "step": 1561
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 2.1702898550724634e-07,
+      "loss": 0.0382,
+      "step": 1562
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 2.168478260869565e-07,
+      "loss": 0.0193,
+      "step": 1563
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 2.1666666666666667e-07,
+      "loss": 0.2805,
+      "step": 1564
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 2.164855072463768e-07,
+      "loss": 0.16,
+      "step": 1565
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 2.1630434782608695e-07,
+      "loss": 0.0154,
+      "step": 1566
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 2.161231884057971e-07,
+      "loss": 0.1567,
+      "step": 1567
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 2.1594202898550723e-07,
+      "loss": 0.026,
+      "step": 1568
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 2.1576086956521738e-07,
+      "loss": 0.0866,
+      "step": 1569
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 2.1557971014492753e-07,
+      "loss": 0.3575,
+      "step": 1570
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 2.1539855072463766e-07,
+      "loss": 0.035,
+      "step": 1571
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 2.1521739130434783e-07,
+      "loss": 0.7372,
+      "step": 1572
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 2.1503623188405796e-07,
+      "loss": 0.2017,
+      "step": 1573
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 2.148550724637681e-07,
+      "loss": 0.0352,
+      "step": 1574
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 2.1467391304347826e-07,
+      "loss": 0.2838,
+      "step": 1575
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 2.144927536231884e-07,
+      "loss": 0.0665,
+      "step": 1576
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 2.1431159420289854e-07,
+      "loss": 0.2798,
+      "step": 1577
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 2.141304347826087e-07,
+      "loss": 0.3058,
+      "step": 1578
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 2.1394927536231882e-07,
+      "loss": 0.0417,
+      "step": 1579
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 2.1376811594202897e-07,
+      "loss": 0.2781,
+      "step": 1580
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 2.1358695652173912e-07,
+      "loss": 0.0257,
+      "step": 1581
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 2.1340579710144927e-07,
+      "loss": 0.2425,
+      "step": 1582
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 2.1322463768115942e-07,
+      "loss": 0.0056,
+      "step": 1583
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 2.1304347826086955e-07,
+      "loss": 0.2768,
+      "step": 1584
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 2.128623188405797e-07,
+      "loss": 0.0119,
+      "step": 1585
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 2.1268115942028985e-07,
+      "loss": 0.3056,
+      "step": 1586
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 2.1249999999999998e-07,
+      "loss": 0.1489,
+      "step": 1587
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 2.1231884057971013e-07,
+      "loss": 0.06,
+      "step": 1588
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 2.1213768115942028e-07,
+      "loss": 0.0087,
+      "step": 1589
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 2.1195652173913043e-07,
+      "loss": 0.129,
+      "step": 1590
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 2.1177536231884058e-07,
+      "loss": 0.3147,
+      "step": 1591
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 2.115942028985507e-07,
+      "loss": 0.36,
+      "step": 1592
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 2.1141304347826086e-07,
+      "loss": 0.4874,
+      "step": 1593
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 2.11231884057971e-07,
+      "loss": 0.1013,
+      "step": 1594
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 2.1105072463768114e-07,
+      "loss": 0.0469,
+      "step": 1595
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 2.108695652173913e-07,
+      "loss": 0.2513,
+      "step": 1596
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 2.1068840579710141e-07,
+      "loss": 0.2501,
+      "step": 1597
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 2.105072463768116e-07,
+      "loss": 0.3041,
+      "step": 1598
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 2.1032608695652174e-07,
+      "loss": 0.0429,
+      "step": 1599
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 2.1014492753623187e-07,
+      "loss": 0.366,
+      "step": 1600
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 2.0996376811594202e-07,
+      "loss": 0.1956,
+      "step": 1601
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 2.0978260869565217e-07,
+      "loss": 0.2162,
+      "step": 1602
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 2.096014492753623e-07,
+      "loss": 0.0099,
+      "step": 1603
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 2.0942028985507245e-07,
+      "loss": 0.3227,
+      "step": 1604
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 2.092391304347826e-07,
+      "loss": 0.1571,
+      "step": 1605
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 2.0905797101449275e-07,
+      "loss": 0.1171,
+      "step": 1606
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 2.088768115942029e-07,
+      "loss": 0.1739,
+      "step": 1607
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 2.0869565217391303e-07,
+      "loss": 0.0271,
+      "step": 1608
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 2.0851449275362318e-07,
+      "loss": 0.0162,
+      "step": 1609
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 2.0833333333333333e-07,
+      "loss": 0.0081,
+      "step": 1610
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.8823529411764706,
+      "eval_combined_score": 0.8989229494614748,
+      "eval_f1": 0.915492957746479,
+      "eval_loss": 0.3817954659461975,
+      "eval_runtime": 8.7859,
+      "eval_samples_per_second": 46.438,
+      "eval_steps_per_second": 2.959,
+      "step": 1610
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 2.0815217391304346e-07,
+      "loss": 0.1201,
+      "step": 1611
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 2.079710144927536e-07,
+      "loss": 0.0453,
+      "step": 1612
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 2.0778985507246376e-07,
+      "loss": 0.1335,
+      "step": 1613
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 2.0760869565217389e-07,
+      "loss": 0.4293,
+      "step": 1614
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 2.0742753623188406e-07,
+      "loss": 0.2781,
+      "step": 1615
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 2.072463768115942e-07,
+      "loss": 0.0719,
+      "step": 1616
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 2.0706521739130434e-07,
+      "loss": 0.2414,
+      "step": 1617
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 2.068840579710145e-07,
+      "loss": 0.314,
+      "step": 1618
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 2.0670289855072462e-07,
+      "loss": 0.1571,
+      "step": 1619
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 2.0652173913043477e-07,
+      "loss": 0.4826,
+      "step": 1620
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 2.0634057971014492e-07,
+      "loss": 0.1231,
+      "step": 1621
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 2.0615942028985505e-07,
+      "loss": 0.4466,
+      "step": 1622
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 2.0597826086956522e-07,
+      "loss": 0.1576,
+      "step": 1623
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 2.0579710144927535e-07,
+      "loss": 0.0853,
+      "step": 1624
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 2.056159420289855e-07,
+      "loss": 0.0242,
+      "step": 1625
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 2.0543478260869565e-07,
+      "loss": 0.3374,
+      "step": 1626
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 2.0525362318840578e-07,
+      "loss": 0.0287,
+      "step": 1627
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 2.0507246376811593e-07,
+      "loss": 0.0216,
+      "step": 1628
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 2.0489130434782608e-07,
+      "loss": 0.0545,
+      "step": 1629
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 2.047101449275362e-07,
+      "loss": 0.236,
+      "step": 1630
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 2.0452898550724636e-07,
+      "loss": 0.2441,
+      "step": 1631
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 2.0434782608695654e-07,
+      "loss": 0.2171,
+      "step": 1632
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 2.0416666666666666e-07,
+      "loss": 0.129,
+      "step": 1633
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 2.0398550724637681e-07,
+      "loss": 0.2509,
+      "step": 1634
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 2.0380434782608694e-07,
+      "loss": 0.3161,
+      "step": 1635
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 2.036231884057971e-07,
+      "loss": 0.416,
+      "step": 1636
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 2.0344202898550724e-07,
+      "loss": 0.3959,
+      "step": 1637
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 2.0326086956521737e-07,
+      "loss": 0.0364,
+      "step": 1638
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 2.0307971014492752e-07,
+      "loss": 0.0365,
+      "step": 1639
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 2.028985507246377e-07,
+      "loss": 0.0796,
+      "step": 1640
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 2.0271739130434782e-07,
+      "loss": 0.2673,
+      "step": 1641
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 2.0253623188405797e-07,
+      "loss": 0.3085,
+      "step": 1642
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 2.023550724637681e-07,
+      "loss": 0.004,
+      "step": 1643
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 2.0217391304347825e-07,
+      "loss": 0.7061,
+      "step": 1644
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 2.019927536231884e-07,
+      "loss": 0.1152,
+      "step": 1645
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 2.0181159420289853e-07,
+      "loss": 0.374,
+      "step": 1646
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 2.0163043478260868e-07,
+      "loss": 0.157,
+      "step": 1647
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 2.0144927536231883e-07,
+      "loss": 0.0365,
+      "step": 1648
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 2.0126811594202898e-07,
+      "loss": 0.2403,
+      "step": 1649
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 2.0108695652173913e-07,
+      "loss": 0.1387,
+      "step": 1650
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 2.0090579710144926e-07,
+      "loss": 0.0068,
+      "step": 1651
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 2.007246376811594e-07,
+      "loss": 0.7509,
+      "step": 1652
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 2.0054347826086956e-07,
+      "loss": 0.016,
+      "step": 1653
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 2.003623188405797e-07,
+      "loss": 0.2687,
+      "step": 1654
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 2.0018115942028984e-07,
+      "loss": 0.3449,
+      "step": 1655
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 2e-07,
+      "loss": 0.0188,
+      "step": 1656
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 1.9981884057971014e-07,
+      "loss": 0.019,
+      "step": 1657
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.996376811594203e-07,
+      "loss": 0.1773,
+      "step": 1658
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.9945652173913042e-07,
+      "loss": 0.1772,
+      "step": 1659
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 1.9927536231884057e-07,
+      "loss": 0.1114,
+      "step": 1660
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 1.9909420289855072e-07,
+      "loss": 0.159,
+      "step": 1661
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.9891304347826085e-07,
+      "loss": 0.1015,
+      "step": 1662
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.98731884057971e-07,
+      "loss": 0.0124,
+      "step": 1663
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.9855072463768115e-07,
+      "loss": 0.6877,
+      "step": 1664
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 1.9836956521739128e-07,
+      "loss": 0.15,
+      "step": 1665
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 1.9818840579710145e-07,
+      "loss": 0.0385,
+      "step": 1666
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 1.980072463768116e-07,
+      "loss": 0.3542,
+      "step": 1667
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 1.9782608695652173e-07,
+      "loss": 0.7173,
+      "step": 1668
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 1.9764492753623188e-07,
+      "loss": 0.0275,
+      "step": 1669
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 1.97463768115942e-07,
+      "loss": 0.2517,
+      "step": 1670
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.9728260869565216e-07,
+      "loss": 0.0133,
+      "step": 1671
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.971014492753623e-07,
+      "loss": 0.0135,
+      "step": 1672
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.9692028985507244e-07,
+      "loss": 0.126,
+      "step": 1673
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 1.9673913043478261e-07,
+      "loss": 0.15,
+      "step": 1674
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 1.9655797101449277e-07,
+      "loss": 0.0092,
+      "step": 1675
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 1.963768115942029e-07,
+      "loss": 0.0082,
+      "step": 1676
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 1.9619565217391304e-07,
+      "loss": 0.2186,
+      "step": 1677
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.9601449275362317e-07,
+      "loss": 0.1853,
+      "step": 1678
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.9583333333333332e-07,
+      "loss": 0.1034,
+      "step": 1679
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.9565217391304347e-07,
+      "loss": 0.0337,
+      "step": 1680
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.954710144927536e-07,
+      "loss": 0.1126,
+      "step": 1681
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.9528985507246375e-07,
+      "loss": 0.126,
+      "step": 1682
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 1.9510869565217393e-07,
+      "loss": 0.2321,
+      "step": 1683
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 1.9492753623188405e-07,
+      "loss": 0.026,
+      "step": 1684
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.947463768115942e-07,
+      "loss": 0.1372,
+      "step": 1685
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.9456521739130433e-07,
+      "loss": 0.029,
+      "step": 1686
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.9438405797101448e-07,
+      "loss": 0.0052,
+      "step": 1687
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 1.9420289855072463e-07,
+      "loss": 0.1388,
+      "step": 1688
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 1.9402173913043476e-07,
+      "loss": 0.0429,
+      "step": 1689
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 1.938405797101449e-07,
+      "loss": 0.0189,
+      "step": 1690
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 1.936594202898551e-07,
+      "loss": 0.0235,
+      "step": 1691
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.934782608695652e-07,
+      "loss": 0.068,
+      "step": 1692
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.9329710144927536e-07,
+      "loss": 0.0258,
+      "step": 1693
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 1.9311594202898552e-07,
+      "loss": 0.172,
+      "step": 1694
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 1.9293478260869564e-07,
+      "loss": 0.0881,
+      "step": 1695
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 1.927536231884058e-07,
+      "loss": 0.3024,
+      "step": 1696
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 1.9257246376811592e-07,
+      "loss": 0.1695,
+      "step": 1697
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 1.9239130434782607e-07,
+      "loss": 0.0164,
+      "step": 1698
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.9221014492753622e-07,
+      "loss": 0.1792,
+      "step": 1699
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.9202898550724637e-07,
+      "loss": 0.0599,
+      "step": 1700
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.9184782608695652e-07,
+      "loss": 0.0257,
+      "step": 1701
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.9166666666666668e-07,
+      "loss": 0.0915,
+      "step": 1702
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.914855072463768e-07,
+      "loss": 0.0145,
+      "step": 1703
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 1.9130434782608695e-07,
+      "loss": 0.0313,
+      "step": 1704
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 1.9112318840579708e-07,
+      "loss": 0.008,
+      "step": 1705
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.9094202898550723e-07,
+      "loss": 0.129,
+      "step": 1706
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.9076086956521738e-07,
+      "loss": 0.0258,
+      "step": 1707
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 1.905797101449275e-07,
+      "loss": 0.1444,
+      "step": 1708
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 1.9039855072463768e-07,
+      "loss": 0.3305,
+      "step": 1709
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 1.9021739130434784e-07,
+      "loss": 0.0244,
+      "step": 1710
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 1.9003623188405796e-07,
+      "loss": 0.0104,
+      "step": 1711
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 1.898550724637681e-07,
+      "loss": 0.161,
+      "step": 1712
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.8967391304347824e-07,
+      "loss": 0.0183,
+      "step": 1713
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.894927536231884e-07,
+      "loss": 0.0898,
+      "step": 1714
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.8931159420289854e-07,
+      "loss": 0.0208,
+      "step": 1715
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.8913043478260867e-07,
+      "loss": 0.0705,
+      "step": 1716
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 1.8894927536231884e-07,
+      "loss": 0.2916,
+      "step": 1717
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 1.88768115942029e-07,
+      "loss": 0.5314,
+      "step": 1718
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 1.8858695652173912e-07,
+      "loss": 0.1374,
+      "step": 1719
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.8840579710144927e-07,
+      "loss": 0.0188,
+      "step": 1720
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.882246376811594e-07,
+      "loss": 0.0097,
+      "step": 1721
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 1.8804347826086955e-07,
+      "loss": 0.0178,
+      "step": 1722
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 1.878623188405797e-07,
+      "loss": 0.0672,
+      "step": 1723
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.8768115942028983e-07,
+      "loss": 0.0447,
+      "step": 1724
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.875e-07,
+      "loss": 0.2747,
+      "step": 1725
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.8731884057971016e-07,
+      "loss": 0.3252,
+      "step": 1726
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.8713768115942028e-07,
+      "loss": 0.2576,
+      "step": 1727
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.8695652173913043e-07,
+      "loss": 0.009,
+      "step": 1728
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 1.8677536231884058e-07,
+      "loss": 0.0324,
+      "step": 1729
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 1.865942028985507e-07,
+      "loss": 0.1526,
+      "step": 1730
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.8641304347826086e-07,
+      "loss": 0.0355,
+      "step": 1731
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.86231884057971e-07,
+      "loss": 0.2926,
+      "step": 1732
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.8605072463768114e-07,
+      "loss": 0.0224,
+      "step": 1733
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 1.8586956521739132e-07,
+      "loss": 0.0126,
+      "step": 1734
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 1.8568840579710144e-07,
+      "loss": 0.0253,
+      "step": 1735
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 1.855072463768116e-07,
+      "loss": 0.2458,
+      "step": 1736
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 1.8532608695652175e-07,
+      "loss": 0.0944,
+      "step": 1737
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.8514492753623187e-07,
+      "loss": 0.0593,
+      "step": 1738
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.8496376811594202e-07,
+      "loss": 0.4496,
+      "step": 1739
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.8478260869565215e-07,
+      "loss": 0.0269,
+      "step": 1740
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.846014492753623e-07,
+      "loss": 0.0437,
+      "step": 1741
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.8442028985507248e-07,
+      "loss": 0.0735,
+      "step": 1742
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 1.842391304347826e-07,
+      "loss": 0.0432,
+      "step": 1743
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 1.8405797101449275e-07,
+      "loss": 0.1016,
+      "step": 1744
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 1.838768115942029e-07,
+      "loss": 0.0237,
+      "step": 1745
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 1.8369565217391303e-07,
+      "loss": 0.2967,
+      "step": 1746
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.8351449275362318e-07,
+      "loss": 0.0456,
+      "step": 1747
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.833333333333333e-07,
+      "loss": 0.0169,
+      "step": 1748
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.8315217391304346e-07,
+      "loss": 0.0138,
+      "step": 1749
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 1.829710144927536e-07,
+      "loss": 0.1276,
+      "step": 1750
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 1.8278985507246376e-07,
+      "loss": 0.0164,
+      "step": 1751
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 1.8260869565217391e-07,
+      "loss": 0.0222,
+      "step": 1752
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 1.8242753623188407e-07,
+      "loss": 0.348,
+      "step": 1753
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.822463768115942e-07,
+      "loss": 0.2323,
+      "step": 1754
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.8206521739130434e-07,
+      "loss": 0.0177,
+      "step": 1755
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.8188405797101447e-07,
+      "loss": 0.0484,
+      "step": 1756
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.8170289855072462e-07,
+      "loss": 0.2645,
+      "step": 1757
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.8152173913043477e-07,
+      "loss": 0.2438,
+      "step": 1758
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 1.813405797101449e-07,
+      "loss": 0.0313,
+      "step": 1759
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 1.8115942028985507e-07,
+      "loss": 0.0716,
+      "step": 1760
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 1.8097826086956523e-07,
+      "loss": 0.0537,
+      "step": 1761
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 1.8079710144927535e-07,
+      "loss": 0.1296,
+      "step": 1762
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.806159420289855e-07,
+      "loss": 0.1652,
+      "step": 1763
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.8043478260869565e-07,
+      "loss": 0.1705,
+      "step": 1764
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.8025362318840578e-07,
+      "loss": 0.008,
+      "step": 1765
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.8007246376811593e-07,
+      "loss": 0.4339,
+      "step": 1766
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.7989130434782606e-07,
+      "loss": 0.2435,
+      "step": 1767
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 1.7971014492753623e-07,
+      "loss": 0.2826,
+      "step": 1768
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 1.7952898550724639e-07,
+      "loss": 0.0618,
+      "step": 1769
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.793478260869565e-07,
+      "loss": 0.2177,
+      "step": 1770
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.7916666666666666e-07,
+      "loss": 0.09,
+      "step": 1771
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.7898550724637681e-07,
+      "loss": 0.0841,
+      "step": 1772
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 1.7880434782608694e-07,
+      "loss": 0.4194,
+      "step": 1773
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 1.786231884057971e-07,
+      "loss": 0.0056,
+      "step": 1774
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 1.7844202898550722e-07,
+      "loss": 0.2403,
+      "step": 1775
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 1.7826086956521737e-07,
+      "loss": 0.3229,
+      "step": 1776
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.7807971014492755e-07,
+      "loss": 0.3268,
+      "step": 1777
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.7789855072463767e-07,
+      "loss": 0.0295,
+      "step": 1778
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.7771739130434782e-07,
+      "loss": 0.0343,
+      "step": 1779
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.7753623188405797e-07,
+      "loss": 0.0043,
+      "step": 1780
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.773550724637681e-07,
+      "loss": 0.3764,
+      "step": 1781
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 1.7717391304347825e-07,
+      "loss": 0.0636,
+      "step": 1782
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 1.7699275362318838e-07,
+      "loss": 0.0062,
+      "step": 1783
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.7681159420289853e-07,
+      "loss": 0.1182,
+      "step": 1784
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.766304347826087e-07,
+      "loss": 0.0406,
+      "step": 1785
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 1.7644927536231883e-07,
+      "loss": 0.0827,
+      "step": 1786
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 1.7626811594202898e-07,
+      "loss": 0.0061,
+      "step": 1787
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 1.7608695652173914e-07,
+      "loss": 0.2301,
+      "step": 1788
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 1.7590579710144926e-07,
+      "loss": 0.1834,
+      "step": 1789
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 1.757246376811594e-07,
+      "loss": 0.0238,
+      "step": 1790
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.7554347826086956e-07,
+      "loss": 0.0683,
+      "step": 1791
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.753623188405797e-07,
+      "loss": 0.1156,
+      "step": 1792
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 1.7518115942028984e-07,
+      "loss": 0.4358,
+      "step": 1793
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 1.75e-07,
+      "loss": 0.0759,
+      "step": 1794
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 1.7481884057971014e-07,
+      "loss": 0.0196,
+      "step": 1795
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 1.746376811594203e-07,
+      "loss": 0.374,
+      "step": 1796
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 1.7445652173913042e-07,
+      "loss": 0.0873,
+      "step": 1797
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 1.7427536231884057e-07,
+      "loss": 0.1456,
+      "step": 1798
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 1.7409420289855072e-07,
+      "loss": 0.3769,
+      "step": 1799
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.7391304347826085e-07,
+      "loss": 0.1817,
+      "step": 1800
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.73731884057971e-07,
+      "loss": 0.1244,
+      "step": 1801
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.7355072463768115e-07,
+      "loss": 0.0139,
+      "step": 1802
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 1.733695652173913e-07,
+      "loss": 0.4218,
+      "step": 1803
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 1.7318840579710146e-07,
+      "loss": 0.0107,
+      "step": 1804
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.7300724637681158e-07,
+      "loss": 0.0244,
+      "step": 1805
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.7282608695652173e-07,
+      "loss": 0.1767,
+      "step": 1806
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 1.7264492753623188e-07,
+      "loss": 0.0356,
+      "step": 1807
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 1.72463768115942e-07,
+      "loss": 0.0059,
+      "step": 1808
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.7228260869565216e-07,
+      "loss": 0.0182,
+      "step": 1809
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.7210144927536229e-07,
+      "loss": 0.0066,
+      "step": 1810
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.7192028985507246e-07,
+      "loss": 0.1298,
+      "step": 1811
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 1.7173913043478262e-07,
+      "loss": 0.1198,
+      "step": 1812
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 1.7155797101449274e-07,
+      "loss": 0.0242,
+      "step": 1813
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 1.713768115942029e-07,
+      "loss": 0.3789,
+      "step": 1814
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 1.7119565217391304e-07,
+      "loss": 0.0107,
+      "step": 1815
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 1.7101449275362317e-07,
+      "loss": 0.1819,
+      "step": 1816
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 1.7083333333333332e-07,
+      "loss": 0.2928,
+      "step": 1817
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 1.7065217391304345e-07,
+      "loss": 0.213,
+      "step": 1818
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.7047101449275362e-07,
+      "loss": 0.0278,
+      "step": 1819
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.7028985507246378e-07,
+      "loss": 0.0293,
+      "step": 1820
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 1.701086956521739e-07,
+      "loss": 0.0054,
+      "step": 1821
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 1.6992753623188405e-07,
+      "loss": 0.0341,
+      "step": 1822
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.697463768115942e-07,
+      "loss": 0.0246,
+      "step": 1823
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.6956521739130433e-07,
+      "loss": 0.1213,
+      "step": 1824
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.6938405797101448e-07,
+      "loss": 0.2937,
+      "step": 1825
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.6920289855072463e-07,
+      "loss": 0.0326,
+      "step": 1826
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.6902173913043476e-07,
+      "loss": 0.0891,
+      "step": 1827
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 1.6884057971014494e-07,
+      "loss": 0.0211,
+      "step": 1828
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 1.6865942028985506e-07,
+      "loss": 0.5623,
+      "step": 1829
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 1.684782608695652e-07,
+      "loss": 0.0319,
+      "step": 1830
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 1.6829710144927536e-07,
+      "loss": 0.0041,
+      "step": 1831
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.681159420289855e-07,
+      "loss": 0.1657,
+      "step": 1832
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.6793478260869564e-07,
+      "loss": 0.3594,
+      "step": 1833
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.677536231884058e-07,
+      "loss": 0.06,
+      "step": 1834
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 1.6757246376811592e-07,
+      "loss": 0.2607,
+      "step": 1835
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 1.673913043478261e-07,
+      "loss": 0.0275,
+      "step": 1836
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 1.6721014492753622e-07,
+      "loss": 0.0077,
+      "step": 1837
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 1.6702898550724637e-07,
+      "loss": 0.4303,
+      "step": 1838
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.6684782608695653e-07,
+      "loss": 0.0169,
+      "step": 1839
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.6666666666666665e-07,
+      "loss": 0.0054,
+      "step": 1840
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8872549019607843,
+      "eval_combined_score": 0.9027021840764775,
+      "eval_f1": 0.9181494661921709,
+      "eval_loss": 0.3902093470096588,
+      "eval_runtime": 8.8207,
+      "eval_samples_per_second": 46.255,
+      "eval_steps_per_second": 2.948,
+      "step": 1840
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 1.664855072463768e-07,
+      "loss": 0.0155,
+      "step": 1841
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 1.6630434782608695e-07,
+      "loss": 0.4551,
+      "step": 1842
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 1.6612318840579708e-07,
+      "loss": 0.0121,
+      "step": 1843
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.6594202898550723e-07,
+      "loss": 0.0163,
+      "step": 1844
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.6576086956521738e-07,
+      "loss": 0.1871,
+      "step": 1845
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.6557971014492753e-07,
+      "loss": 0.0057,
+      "step": 1846
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.6539855072463769e-07,
+      "loss": 0.0138,
+      "step": 1847
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.652173913043478e-07,
+      "loss": 0.011,
+      "step": 1848
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.6503623188405796e-07,
+      "loss": 0.0049,
+      "step": 1849
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.6485507246376811e-07,
+      "loss": 0.0778,
+      "step": 1850
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.6467391304347824e-07,
+      "loss": 0.014,
+      "step": 1851
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.644927536231884e-07,
+      "loss": 0.1946,
+      "step": 1852
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.6431159420289857e-07,
+      "loss": 0.0726,
+      "step": 1853
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.641304347826087e-07,
+      "loss": 0.0374,
+      "step": 1854
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.6394927536231885e-07,
+      "loss": 0.1955,
+      "step": 1855
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.6376811594202897e-07,
+      "loss": 0.1685,
+      "step": 1856
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.6358695652173912e-07,
+      "loss": 0.1769,
+      "step": 1857
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.6340579710144927e-07,
+      "loss": 0.3073,
+      "step": 1858
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.632246376811594e-07,
+      "loss": 0.5301,
+      "step": 1859
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.6304347826086955e-07,
+      "loss": 0.0149,
+      "step": 1860
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.628623188405797e-07,
+      "loss": 0.0084,
+      "step": 1861
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.6268115942028985e-07,
+      "loss": 0.2829,
+      "step": 1862
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.625e-07,
+      "loss": 0.0117,
+      "step": 1863
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.6231884057971013e-07,
+      "loss": 0.4676,
+      "step": 1864
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.6213768115942028e-07,
+      "loss": 0.0556,
+      "step": 1865
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.6195652173913043e-07,
+      "loss": 0.191,
+      "step": 1866
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.6177536231884056e-07,
+      "loss": 0.288,
+      "step": 1867
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.615942028985507e-07,
+      "loss": 0.4041,
+      "step": 1868
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.6141304347826086e-07,
+      "loss": 0.0084,
+      "step": 1869
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.6123188405797101e-07,
+      "loss": 0.0446,
+      "step": 1870
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.6105072463768117e-07,
+      "loss": 0.6713,
+      "step": 1871
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.608695652173913e-07,
+      "loss": 0.3212,
+      "step": 1872
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.6068840579710144e-07,
+      "loss": 0.0278,
+      "step": 1873
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.605072463768116e-07,
+      "loss": 0.0461,
+      "step": 1874
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.6032608695652172e-07,
+      "loss": 0.0154,
+      "step": 1875
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.6014492753623187e-07,
+      "loss": 0.0051,
+      "step": 1876
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.5996376811594202e-07,
+      "loss": 0.2654,
+      "step": 1877
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.5978260869565215e-07,
+      "loss": 0.019,
+      "step": 1878
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.5960144927536233e-07,
+      "loss": 0.0255,
+      "step": 1879
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.5942028985507245e-07,
+      "loss": 0.1402,
+      "step": 1880
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.592391304347826e-07,
+      "loss": 0.0519,
+      "step": 1881
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.5905797101449275e-07,
+      "loss": 0.2332,
+      "step": 1882
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.5887681159420288e-07,
+      "loss": 0.3535,
+      "step": 1883
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.5869565217391303e-07,
+      "loss": 0.11,
+      "step": 1884
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.5851449275362318e-07,
+      "loss": 0.0542,
+      "step": 1885
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.583333333333333e-07,
+      "loss": 0.2982,
+      "step": 1886
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.5815217391304349e-07,
+      "loss": 0.1259,
+      "step": 1887
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.5797101449275364e-07,
+      "loss": 0.0458,
+      "step": 1888
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.5778985507246376e-07,
+      "loss": 0.0357,
+      "step": 1889
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.5760869565217392e-07,
+      "loss": 0.1908,
+      "step": 1890
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.5742753623188404e-07,
+      "loss": 0.1795,
+      "step": 1891
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.572463768115942e-07,
+      "loss": 0.4618,
+      "step": 1892
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.5706521739130434e-07,
+      "loss": 0.0164,
+      "step": 1893
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.5688405797101447e-07,
+      "loss": 0.7443,
+      "step": 1894
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.5670289855072462e-07,
+      "loss": 0.257,
+      "step": 1895
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.565217391304348e-07,
+      "loss": 0.0163,
+      "step": 1896
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.5634057971014492e-07,
+      "loss": 0.2919,
+      "step": 1897
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.5615942028985508e-07,
+      "loss": 0.0208,
+      "step": 1898
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.559782608695652e-07,
+      "loss": 0.268,
+      "step": 1899
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.5579710144927535e-07,
+      "loss": 0.3731,
+      "step": 1900
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.556159420289855e-07,
+      "loss": 0.1472,
+      "step": 1901
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.5543478260869563e-07,
+      "loss": 0.0724,
+      "step": 1902
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.5525362318840578e-07,
+      "loss": 0.0172,
+      "step": 1903
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.5507246376811596e-07,
+      "loss": 0.203,
+      "step": 1904
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.5489130434782608e-07,
+      "loss": 0.0162,
+      "step": 1905
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.5471014492753624e-07,
+      "loss": 0.0962,
+      "step": 1906
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.5452898550724636e-07,
+      "loss": 0.2673,
+      "step": 1907
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.543478260869565e-07,
+      "loss": 0.0303,
+      "step": 1908
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.5416666666666666e-07,
+      "loss": 0.3059,
+      "step": 1909
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.539855072463768e-07,
+      "loss": 0.0304,
+      "step": 1910
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.5380434782608694e-07,
+      "loss": 0.046,
+      "step": 1911
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.536231884057971e-07,
+      "loss": 0.0193,
+      "step": 1912
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.5344202898550724e-07,
+      "loss": 0.1047,
+      "step": 1913
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.532608695652174e-07,
+      "loss": 0.5978,
+      "step": 1914
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.5307971014492755e-07,
+      "loss": 0.1131,
+      "step": 1915
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.5289855072463767e-07,
+      "loss": 0.3664,
+      "step": 1916
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.5271739130434782e-07,
+      "loss": 0.0596,
+      "step": 1917
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.5253623188405795e-07,
+      "loss": 0.0164,
+      "step": 1918
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.523550724637681e-07,
+      "loss": 0.0082,
+      "step": 1919
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.5217391304347825e-07,
+      "loss": 0.1418,
+      "step": 1920
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.5199275362318838e-07,
+      "loss": 0.0307,
+      "step": 1921
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.5181159420289856e-07,
+      "loss": 0.116,
+      "step": 1922
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.516304347826087e-07,
+      "loss": 0.0317,
+      "step": 1923
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.5144927536231883e-07,
+      "loss": 0.0291,
+      "step": 1924
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.5126811594202898e-07,
+      "loss": 0.0042,
+      "step": 1925
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.510869565217391e-07,
+      "loss": 0.0373,
+      "step": 1926
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.5090579710144926e-07,
+      "loss": 0.2558,
+      "step": 1927
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.507246376811594e-07,
+      "loss": 0.0158,
+      "step": 1928
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.5054347826086954e-07,
+      "loss": 0.1087,
+      "step": 1929
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.5036231884057972e-07,
+      "loss": 0.3221,
+      "step": 1930
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.5018115942028987e-07,
+      "loss": 0.1207,
+      "step": 1931
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.5e-07,
+      "loss": 0.098,
+      "step": 1932
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.4981884057971014e-07,
+      "loss": 0.0117,
+      "step": 1933
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.4963768115942027e-07,
+      "loss": 0.0255,
+      "step": 1934
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.4945652173913042e-07,
+      "loss": 0.3015,
+      "step": 1935
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.4927536231884057e-07,
+      "loss": 0.0262,
+      "step": 1936
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.490942028985507e-07,
+      "loss": 0.2457,
+      "step": 1937
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.4891304347826088e-07,
+      "loss": 0.1065,
+      "step": 1938
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.4873188405797103e-07,
+      "loss": 0.0916,
+      "step": 1939
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.4855072463768115e-07,
+      "loss": 0.0157,
+      "step": 1940
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 1.483695652173913e-07,
+      "loss": 0.0106,
+      "step": 1941
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 1.4818840579710143e-07,
+      "loss": 0.3427,
+      "step": 1942
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.4800724637681158e-07,
+      "loss": 0.007,
+      "step": 1943
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.4782608695652173e-07,
+      "loss": 0.253,
+      "step": 1944
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.4764492753623186e-07,
+      "loss": 0.1538,
+      "step": 1945
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.47463768115942e-07,
+      "loss": 0.1617,
+      "step": 1946
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.472826086956522e-07,
+      "loss": 0.1743,
+      "step": 1947
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.4710144927536231e-07,
+      "loss": 0.0171,
+      "step": 1948
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.4692028985507247e-07,
+      "loss": 0.5306,
+      "step": 1949
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.4673913043478262e-07,
+      "loss": 0.0546,
+      "step": 1950
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.4655797101449274e-07,
+      "loss": 0.282,
+      "step": 1951
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.463768115942029e-07,
+      "loss": 0.0835,
+      "step": 1952
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.4619565217391302e-07,
+      "loss": 0.0084,
+      "step": 1953
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.4601449275362317e-07,
+      "loss": 0.2495,
+      "step": 1954
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.4583333333333335e-07,
+      "loss": 0.0871,
+      "step": 1955
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.4565217391304347e-07,
+      "loss": 0.197,
+      "step": 1956
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.4547101449275363e-07,
+      "loss": 0.092,
+      "step": 1957
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.4528985507246378e-07,
+      "loss": 0.0105,
+      "step": 1958
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.451086956521739e-07,
+      "loss": 0.0285,
+      "step": 1959
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.4492753623188405e-07,
+      "loss": 0.125,
+      "step": 1960
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.4474637681159418e-07,
+      "loss": 0.2004,
+      "step": 1961
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.4456521739130433e-07,
+      "loss": 0.2828,
+      "step": 1962
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.4438405797101448e-07,
+      "loss": 0.01,
+      "step": 1963
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.4420289855072463e-07,
+      "loss": 0.0214,
+      "step": 1964
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.4402173913043479e-07,
+      "loss": 0.0513,
+      "step": 1965
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.4384057971014494e-07,
+      "loss": 0.1726,
+      "step": 1966
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.4365942028985506e-07,
+      "loss": 0.4055,
+      "step": 1967
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 1.4347826086956521e-07,
+      "loss": 0.0058,
+      "step": 1968
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 1.4329710144927534e-07,
+      "loss": 0.1825,
+      "step": 1969
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.431159420289855e-07,
+      "loss": 0.2149,
+      "step": 1970
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.4293478260869564e-07,
+      "loss": 0.0126,
+      "step": 1971
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.4275362318840577e-07,
+      "loss": 0.11,
+      "step": 1972
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.4257246376811595e-07,
+      "loss": 0.033,
+      "step": 1973
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.423913043478261e-07,
+      "loss": 0.123,
+      "step": 1974
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.4221014492753622e-07,
+      "loss": 0.013,
+      "step": 1975
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.4202898550724637e-07,
+      "loss": 0.0532,
+      "step": 1976
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.4184782608695653e-07,
+      "loss": 0.0632,
+      "step": 1977
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.4166666666666665e-07,
+      "loss": 0.335,
+      "step": 1978
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.414855072463768e-07,
+      "loss": 0.0544,
+      "step": 1979
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 1.4130434782608693e-07,
+      "loss": 0.0481,
+      "step": 1980
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 1.411231884057971e-07,
+      "loss": 0.1653,
+      "step": 1981
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 1.4094202898550726e-07,
+      "loss": 0.2556,
+      "step": 1982
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 1.4076086956521738e-07,
+      "loss": 0.0077,
+      "step": 1983
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 1.4057971014492753e-07,
+      "loss": 0.0508,
+      "step": 1984
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 1.4039855072463769e-07,
+      "loss": 0.0083,
+      "step": 1985
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 1.402173913043478e-07,
+      "loss": 0.3602,
+      "step": 1986
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 1.4003623188405796e-07,
+      "loss": 0.0256,
+      "step": 1987
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 1.398550724637681e-07,
+      "loss": 0.3324,
+      "step": 1988
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 1.3967391304347824e-07,
+      "loss": 0.0146,
+      "step": 1989
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 1.3949275362318842e-07,
+      "loss": 0.3512,
+      "step": 1990
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 1.3931159420289854e-07,
+      "loss": 0.0053,
+      "step": 1991
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 1.391304347826087e-07,
+      "loss": 0.072,
+      "step": 1992
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 1.3894927536231885e-07,
+      "loss": 0.3749,
+      "step": 1993
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 1.3876811594202897e-07,
+      "loss": 0.3107,
+      "step": 1994
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 1.3858695652173912e-07,
+      "loss": 0.0073,
+      "step": 1995
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 1.3840579710144925e-07,
+      "loss": 0.0333,
+      "step": 1996
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 1.382246376811594e-07,
+      "loss": 0.2379,
+      "step": 1997
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 1.3804347826086958e-07,
+      "loss": 0.005,
+      "step": 1998
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 1.378623188405797e-07,
+      "loss": 0.0838,
+      "step": 1999
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 1.3768115942028986e-07,
+      "loss": 0.3811,
+      "step": 2000
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 1.375e-07,
+      "loss": 0.0191,
+      "step": 2001
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 1.3731884057971013e-07,
+      "loss": 0.023,
+      "step": 2002
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 1.3713768115942028e-07,
+      "loss": 0.0115,
+      "step": 2003
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 1.369565217391304e-07,
+      "loss": 0.0143,
+      "step": 2004
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 1.3677536231884056e-07,
+      "loss": 0.0207,
+      "step": 2005
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 1.365942028985507e-07,
+      "loss": 0.2305,
+      "step": 2006
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 1.3641304347826086e-07,
+      "loss": 0.0239,
+      "step": 2007
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 1.3623188405797102e-07,
+      "loss": 0.3981,
+      "step": 2008
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 1.3605072463768117e-07,
+      "loss": 0.1325,
+      "step": 2009
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 1.358695652173913e-07,
+      "loss": 0.0067,
+      "step": 2010
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 1.3568840579710144e-07,
+      "loss": 0.0209,
+      "step": 2011
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 1.355072463768116e-07,
+      "loss": 0.028,
+      "step": 2012
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 1.3532608695652172e-07,
+      "loss": 0.2172,
+      "step": 2013
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 1.3514492753623187e-07,
+      "loss": 0.1055,
+      "step": 2014
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 1.3496376811594202e-07,
+      "loss": 0.1886,
+      "step": 2015
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 1.3478260869565218e-07,
+      "loss": 0.006,
+      "step": 2016
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 1.3460144927536233e-07,
+      "loss": 0.0111,
+      "step": 2017
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 1.3442028985507245e-07,
+      "loss": 0.3443,
+      "step": 2018
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 1.342391304347826e-07,
+      "loss": 0.19,
+      "step": 2019
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 1.3405797101449276e-07,
+      "loss": 0.0094,
+      "step": 2020
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 1.3387681159420288e-07,
+      "loss": 0.5131,
+      "step": 2021
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 1.3369565217391303e-07,
+      "loss": 0.0234,
+      "step": 2022
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 1.3351449275362316e-07,
+      "loss": 0.0179,
+      "step": 2023
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 1.3333333333333334e-07,
+      "loss": 0.1598,
+      "step": 2024
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 1.331521739130435e-07,
+      "loss": 0.177,
+      "step": 2025
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 1.329710144927536e-07,
+      "loss": 0.4147,
+      "step": 2026
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 1.3278985507246376e-07,
+      "loss": 0.1755,
+      "step": 2027
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 1.3260869565217392e-07,
+      "loss": 0.0185,
+      "step": 2028
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 1.3242753623188404e-07,
+      "loss": 0.0299,
+      "step": 2029
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 1.322463768115942e-07,
+      "loss": 0.109,
+      "step": 2030
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 1.3206521739130432e-07,
+      "loss": 0.1516,
+      "step": 2031
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 1.318840579710145e-07,
+      "loss": 0.0535,
+      "step": 2032
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 1.3170289855072465e-07,
+      "loss": 0.4152,
+      "step": 2033
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 1.3152173913043477e-07,
+      "loss": 0.0086,
+      "step": 2034
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 1.3134057971014492e-07,
+      "loss": 0.1951,
+      "step": 2035
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 1.3115942028985508e-07,
+      "loss": 0.0858,
+      "step": 2036
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 1.309782608695652e-07,
+      "loss": 0.1257,
+      "step": 2037
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 1.3079710144927535e-07,
+      "loss": 0.1249,
+      "step": 2038
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 1.306159420289855e-07,
+      "loss": 0.0062,
+      "step": 2039
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 1.3043478260869563e-07,
+      "loss": 0.4516,
+      "step": 2040
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 1.302536231884058e-07,
+      "loss": 0.0135,
+      "step": 2041
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 1.3007246376811593e-07,
+      "loss": 0.0096,
+      "step": 2042
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 1.2989130434782609e-07,
+      "loss": 0.0061,
+      "step": 2043
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 1.2971014492753624e-07,
+      "loss": 0.057,
+      "step": 2044
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 1.2952898550724636e-07,
+      "loss": 0.3508,
+      "step": 2045
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 1.2934782608695651e-07,
+      "loss": 0.0051,
+      "step": 2046
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 1.2916666666666667e-07,
+      "loss": 0.3586,
+      "step": 2047
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 1.289855072463768e-07,
+      "loss": 0.0895,
+      "step": 2048
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 1.2880434782608697e-07,
+      "loss": 0.0051,
+      "step": 2049
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 1.286231884057971e-07,
+      "loss": 0.0095,
+      "step": 2050
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 1.2844202898550725e-07,
+      "loss": 0.0096,
+      "step": 2051
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 1.282608695652174e-07,
+      "loss": 0.5975,
+      "step": 2052
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 1.2807971014492752e-07,
+      "loss": 0.1532,
+      "step": 2053
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 1.2789855072463767e-07,
+      "loss": 0.3372,
+      "step": 2054
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 1.2771739130434783e-07,
+      "loss": 0.1776,
+      "step": 2055
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 1.2753623188405795e-07,
+      "loss": 0.2098,
+      "step": 2056
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 1.273550724637681e-07,
+      "loss": 0.1367,
+      "step": 2057
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 1.2717391304347825e-07,
+      "loss": 0.0148,
+      "step": 2058
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 1.269927536231884e-07,
+      "loss": 0.2808,
+      "step": 2059
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 1.2681159420289856e-07,
+      "loss": 0.7086,
+      "step": 2060
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 1.2663043478260868e-07,
+      "loss": 0.0071,
+      "step": 2061
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 1.2644927536231883e-07,
+      "loss": 0.025,
+      "step": 2062
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 1.2626811594202899e-07,
+      "loss": 0.4704,
+      "step": 2063
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 1.260869565217391e-07,
+      "loss": 0.1014,
+      "step": 2064
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 1.2590579710144926e-07,
+      "loss": 0.0153,
+      "step": 2065
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 1.2572463768115941e-07,
+      "loss": 0.1526,
+      "step": 2066
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 1.2554347826086957e-07,
+      "loss": 0.1688,
+      "step": 2067
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 1.2536231884057972e-07,
+      "loss": 0.0256,
+      "step": 2068
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 1.2518115942028984e-07,
+      "loss": 0.0203,
+      "step": 2069
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 1.25e-07,
+      "loss": 0.0383,
+      "step": 2070
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.8921568627450981,
+      "eval_combined_score": 0.907346037006352,
+      "eval_f1": 0.9225352112676057,
+      "eval_loss": 0.36586612462997437,
+      "eval_runtime": 9.084,
+      "eval_samples_per_second": 44.914,
+      "eval_steps_per_second": 2.862,
+      "step": 2070
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 1.2481884057971015e-07,
+      "loss": 0.3214,
+      "step": 2071
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 1.246376811594203e-07,
+      "loss": 0.0204,
+      "step": 2072
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 1.2445652173913042e-07,
+      "loss": 0.3614,
+      "step": 2073
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 1.2427536231884057e-07,
+      "loss": 0.0507,
+      "step": 2074
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 1.2409420289855073e-07,
+      "loss": 0.1034,
+      "step": 2075
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 1.2391304347826088e-07,
+      "loss": 0.1199,
+      "step": 2076
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 1.23731884057971e-07,
+      "loss": 0.0244,
+      "step": 2077
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 1.2355072463768115e-07,
+      "loss": 0.9242,
+      "step": 2078
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 1.233695652173913e-07,
+      "loss": 0.2319,
+      "step": 2079
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 1.2318840579710146e-07,
+      "loss": 0.3737,
+      "step": 2080
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 1.2300724637681158e-07,
+      "loss": 0.0125,
+      "step": 2081
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 1.2282608695652173e-07,
+      "loss": 0.2396,
+      "step": 2082
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 1.2264492753623189e-07,
+      "loss": 0.3656,
+      "step": 2083
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 1.2246376811594204e-07,
+      "loss": 0.3728,
+      "step": 2084
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 1.2228260869565216e-07,
+      "loss": 0.0079,
+      "step": 2085
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 1.2210144927536231e-07,
+      "loss": 0.2723,
+      "step": 2086
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 1.2192028985507244e-07,
+      "loss": 0.0116,
+      "step": 2087
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 1.2173913043478262e-07,
+      "loss": 0.0059,
+      "step": 2088
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 1.2155797101449274e-07,
+      "loss": 0.0425,
+      "step": 2089
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 1.213768115942029e-07,
+      "loss": 0.0344,
+      "step": 2090
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 1.2119565217391302e-07,
+      "loss": 0.1583,
+      "step": 2091
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 1.210144927536232e-07,
+      "loss": 0.0105,
+      "step": 2092
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 1.2083333333333332e-07,
+      "loss": 0.3955,
+      "step": 2093
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 1.2065217391304348e-07,
+      "loss": 0.0291,
+      "step": 2094
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 1.204710144927536e-07,
+      "loss": 0.0141,
+      "step": 2095
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 1.2028985507246378e-07,
+      "loss": 0.0116,
+      "step": 2096
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 1.201086956521739e-07,
+      "loss": 0.0061,
+      "step": 2097
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 1.1992753623188406e-07,
+      "loss": 0.1289,
+      "step": 2098
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 1.1974637681159418e-07,
+      "loss": 0.0218,
+      "step": 2099
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 1.1956521739130436e-07,
+      "loss": 0.7285,
+      "step": 2100
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 1.1938405797101448e-07,
+      "loss": 0.2038,
+      "step": 2101
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 1.1920289855072464e-07,
+      "loss": 0.0648,
+      "step": 2102
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 1.1902173913043479e-07,
+      "loss": 0.1997,
+      "step": 2103
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 1.1884057971014493e-07,
+      "loss": 0.1405,
+      "step": 2104
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 1.1865942028985506e-07,
+      "loss": 0.0288,
+      "step": 2105
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 1.184782608695652e-07,
+      "loss": 0.2476,
+      "step": 2106
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 1.1829710144927537e-07,
+      "loss": 0.0148,
+      "step": 2107
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 1.181159420289855e-07,
+      "loss": 0.0263,
+      "step": 2108
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 1.1793478260869564e-07,
+      "loss": 0.0261,
+      "step": 2109
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 1.1775362318840578e-07,
+      "loss": 0.0635,
+      "step": 2110
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 1.1757246376811595e-07,
+      "loss": 0.4629,
+      "step": 2111
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 1.1739130434782609e-07,
+      "loss": 0.2669,
+      "step": 2112
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 1.1721014492753622e-07,
+      "loss": 0.0268,
+      "step": 2113
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 1.1702898550724636e-07,
+      "loss": 0.2451,
+      "step": 2114
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 1.1684782608695653e-07,
+      "loss": 0.0057,
+      "step": 2115
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 1.1666666666666667e-07,
+      "loss": 0.0324,
+      "step": 2116
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 1.164855072463768e-07,
+      "loss": 0.1054,
+      "step": 2117
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 1.1630434782608694e-07,
+      "loss": 0.0234,
+      "step": 2118
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 1.1612318840579711e-07,
+      "loss": 0.0617,
+      "step": 2119
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 1.1594202898550725e-07,
+      "loss": 0.1154,
+      "step": 2120
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 1.1576086956521738e-07,
+      "loss": 0.2234,
+      "step": 2121
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 1.1557971014492752e-07,
+      "loss": 0.0216,
+      "step": 2122
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 1.1539855072463767e-07,
+      "loss": 0.4074,
+      "step": 2123
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 1.1521739130434783e-07,
+      "loss": 0.058,
+      "step": 2124
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 1.1503623188405796e-07,
+      "loss": 0.0123,
+      "step": 2125
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 1.148550724637681e-07,
+      "loss": 0.2287,
+      "step": 2126
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 1.1467391304347825e-07,
+      "loss": 0.0287,
+      "step": 2127
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 1.144927536231884e-07,
+      "loss": 0.0523,
+      "step": 2128
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 1.1431159420289854e-07,
+      "loss": 0.0269,
+      "step": 2129
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 1.1413043478260868e-07,
+      "loss": 0.2575,
+      "step": 2130
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 1.1394927536231883e-07,
+      "loss": 0.2621,
+      "step": 2131
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 1.1376811594202899e-07,
+      "loss": 0.028,
+      "step": 2132
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 1.1358695652173912e-07,
+      "loss": 0.0501,
+      "step": 2133
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 1.1340579710144928e-07,
+      "loss": 0.3011,
+      "step": 2134
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 1.1322463768115941e-07,
+      "loss": 0.2111,
+      "step": 2135
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 1.1304347826086955e-07,
+      "loss": 0.1955,
+      "step": 2136
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 1.128623188405797e-07,
+      "loss": 0.1926,
+      "step": 2137
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 1.1268115942028986e-07,
+      "loss": 0.0247,
+      "step": 2138
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 1.125e-07,
+      "loss": 0.332,
+      "step": 2139
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 1.1231884057971013e-07,
+      "loss": 0.0704,
+      "step": 2140
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 1.1213768115942028e-07,
+      "loss": 0.0591,
+      "step": 2141
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 1.1195652173913044e-07,
+      "loss": 0.3785,
+      "step": 2142
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 1.1177536231884058e-07,
+      "loss": 0.0086,
+      "step": 2143
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 1.1159420289855071e-07,
+      "loss": 0.3973,
+      "step": 2144
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 1.1141304347826087e-07,
+      "loss": 0.0767,
+      "step": 2145
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 1.1123188405797102e-07,
+      "loss": 0.1916,
+      "step": 2146
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 1.1105072463768116e-07,
+      "loss": 0.1409,
+      "step": 2147
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 1.108695652173913e-07,
+      "loss": 0.0678,
+      "step": 2148
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 1.1068840579710145e-07,
+      "loss": 0.698,
+      "step": 2149
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 1.105072463768116e-07,
+      "loss": 0.157,
+      "step": 2150
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 1.1032608695652174e-07,
+      "loss": 0.0551,
+      "step": 2151
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 1.1014492753623187e-07,
+      "loss": 0.4498,
+      "step": 2152
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 1.0996376811594201e-07,
+      "loss": 0.0601,
+      "step": 2153
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 1.0978260869565218e-07,
+      "loss": 0.3492,
+      "step": 2154
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 1.0960144927536232e-07,
+      "loss": 0.0115,
+      "step": 2155
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 1.0942028985507245e-07,
+      "loss": 0.0044,
+      "step": 2156
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 1.0923913043478259e-07,
+      "loss": 0.0106,
+      "step": 2157
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 1.0905797101449276e-07,
+      "loss": 0.0192,
+      "step": 2158
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.088768115942029e-07,
+      "loss": 0.1005,
+      "step": 2159
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.0869565217391303e-07,
+      "loss": 0.0374,
+      "step": 2160
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.0851449275362317e-07,
+      "loss": 0.2407,
+      "step": 2161
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.0833333333333334e-07,
+      "loss": 0.0525,
+      "step": 2162
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.0815217391304348e-07,
+      "loss": 0.0046,
+      "step": 2163
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.0797101449275361e-07,
+      "loss": 0.023,
+      "step": 2164
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.0778985507246377e-07,
+      "loss": 0.0131,
+      "step": 2165
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.0760869565217392e-07,
+      "loss": 0.3866,
+      "step": 2166
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.0742753623188406e-07,
+      "loss": 0.0586,
+      "step": 2167
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.072463768115942e-07,
+      "loss": 0.0276,
+      "step": 2168
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.0706521739130435e-07,
+      "loss": 0.1011,
+      "step": 2169
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.0688405797101448e-07,
+      "loss": 0.0866,
+      "step": 2170
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.0670289855072464e-07,
+      "loss": 0.0486,
+      "step": 2171
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.0652173913043477e-07,
+      "loss": 0.0125,
+      "step": 2172
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.0634057971014493e-07,
+      "loss": 0.1627,
+      "step": 2173
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.0615942028985506e-07,
+      "loss": 0.3001,
+      "step": 2174
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.0597826086956522e-07,
+      "loss": 0.0403,
+      "step": 2175
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.0579710144927535e-07,
+      "loss": 0.0079,
+      "step": 2176
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.056159420289855e-07,
+      "loss": 0.395,
+      "step": 2177
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.0543478260869564e-07,
+      "loss": 0.0078,
+      "step": 2178
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.052536231884058e-07,
+      "loss": 0.0064,
+      "step": 2179
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.0507246376811593e-07,
+      "loss": 0.1995,
+      "step": 2180
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.0489130434782609e-07,
+      "loss": 0.4201,
+      "step": 2181
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.0471014492753622e-07,
+      "loss": 0.0638,
+      "step": 2182
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.0452898550724638e-07,
+      "loss": 0.264,
+      "step": 2183
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.0434782608695651e-07,
+      "loss": 0.1746,
+      "step": 2184
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.0416666666666667e-07,
+      "loss": 0.2041,
+      "step": 2185
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.039855072463768e-07,
+      "loss": 0.0555,
+      "step": 2186
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.0380434782608694e-07,
+      "loss": 0.236,
+      "step": 2187
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.036231884057971e-07,
+      "loss": 0.0171,
+      "step": 2188
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.0344202898550725e-07,
+      "loss": 0.1046,
+      "step": 2189
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.0326086956521738e-07,
+      "loss": 0.0167,
+      "step": 2190
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.0307971014492752e-07,
+      "loss": 0.0394,
+      "step": 2191
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.0289855072463767e-07,
+      "loss": 0.3342,
+      "step": 2192
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.0271739130434783e-07,
+      "loss": 0.0229,
+      "step": 2193
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.0253623188405797e-07,
+      "loss": 0.1353,
+      "step": 2194
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.023550724637681e-07,
+      "loss": 0.0239,
+      "step": 2195
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0217391304347827e-07,
+      "loss": 0.0346,
+      "step": 2196
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0199275362318841e-07,
+      "loss": 0.3768,
+      "step": 2197
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0181159420289855e-07,
+      "loss": 0.3121,
+      "step": 2198
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0163043478260868e-07,
+      "loss": 0.1975,
+      "step": 2199
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 1.0144927536231885e-07,
+      "loss": 0.0411,
+      "step": 2200
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 1.0126811594202899e-07,
+      "loss": 0.1544,
+      "step": 2201
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 1.0108695652173913e-07,
+      "loss": 0.0092,
+      "step": 2202
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 1.0090579710144926e-07,
+      "loss": 0.1473,
+      "step": 2203
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 1.0072463768115942e-07,
+      "loss": 0.0231,
+      "step": 2204
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 1.0054347826086957e-07,
+      "loss": 0.2838,
+      "step": 2205
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 1.003623188405797e-07,
+      "loss": 0.3561,
+      "step": 2206
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 1.0018115942028984e-07,
+      "loss": 0.4274,
+      "step": 2207
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 1e-07,
+      "loss": 0.007,
+      "step": 2208
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 9.981884057971015e-08,
+      "loss": 0.0122,
+      "step": 2209
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 9.963768115942029e-08,
+      "loss": 0.0065,
+      "step": 2210
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 9.945652173913042e-08,
+      "loss": 0.0309,
+      "step": 2211
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 9.927536231884058e-08,
+      "loss": 0.0086,
+      "step": 2212
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 9.909420289855073e-08,
+      "loss": 0.0437,
+      "step": 2213
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 9.891304347826087e-08,
+      "loss": 0.014,
+      "step": 2214
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 9.8731884057971e-08,
+      "loss": 0.0192,
+      "step": 2215
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 9.855072463768116e-08,
+      "loss": 0.3965,
+      "step": 2216
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 9.836956521739131e-08,
+      "loss": 0.044,
+      "step": 2217
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 9.818840579710145e-08,
+      "loss": 0.2313,
+      "step": 2218
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 9.800724637681158e-08,
+      "loss": 0.011,
+      "step": 2219
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 9.782608695652174e-08,
+      "loss": 0.4078,
+      "step": 2220
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 9.764492753623187e-08,
+      "loss": 0.0047,
+      "step": 2221
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 9.746376811594203e-08,
+      "loss": 0.0472,
+      "step": 2222
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 9.728260869565216e-08,
+      "loss": 0.2484,
+      "step": 2223
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 9.710144927536232e-08,
+      "loss": 0.0047,
+      "step": 2224
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 9.692028985507245e-08,
+      "loss": 0.0784,
+      "step": 2225
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 9.67391304347826e-08,
+      "loss": 0.094,
+      "step": 2226
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 9.655797101449276e-08,
+      "loss": 0.0471,
+      "step": 2227
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 9.63768115942029e-08,
+      "loss": 0.0153,
+      "step": 2228
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 9.619565217391303e-08,
+      "loss": 0.2941,
+      "step": 2229
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 9.601449275362319e-08,
+      "loss": 0.0608,
+      "step": 2230
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 9.583333333333334e-08,
+      "loss": 0.0814,
+      "step": 2231
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 9.565217391304348e-08,
+      "loss": 0.2729,
+      "step": 2232
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 9.547101449275361e-08,
+      "loss": 0.317,
+      "step": 2233
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 9.528985507246375e-08,
+      "loss": 0.1331,
+      "step": 2234
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 9.510869565217392e-08,
+      "loss": 0.1587,
+      "step": 2235
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 9.492753623188406e-08,
+      "loss": 0.0096,
+      "step": 2236
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 9.47463768115942e-08,
+      "loss": 0.0418,
+      "step": 2237
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 9.456521739130433e-08,
+      "loss": 0.0492,
+      "step": 2238
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 9.43840579710145e-08,
+      "loss": 0.0735,
+      "step": 2239
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 9.420289855072464e-08,
+      "loss": 0.1205,
+      "step": 2240
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 9.402173913043477e-08,
+      "loss": 0.0956,
+      "step": 2241
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 9.384057971014491e-08,
+      "loss": 0.0394,
+      "step": 2242
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 9.365942028985508e-08,
+      "loss": 0.0405,
+      "step": 2243
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 9.347826086956522e-08,
+      "loss": 0.0206,
+      "step": 2244
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 9.329710144927536e-08,
+      "loss": 0.0411,
+      "step": 2245
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 9.31159420289855e-08,
+      "loss": 0.0165,
+      "step": 2246
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 9.293478260869566e-08,
+      "loss": 0.2973,
+      "step": 2247
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 9.27536231884058e-08,
+      "loss": 0.1723,
+      "step": 2248
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 9.257246376811594e-08,
+      "loss": 0.0129,
+      "step": 2249
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 9.239130434782607e-08,
+      "loss": 0.0295,
+      "step": 2250
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 9.221014492753624e-08,
+      "loss": 0.2568,
+      "step": 2251
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 9.202898550724638e-08,
+      "loss": 0.0057,
+      "step": 2252
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 9.184782608695652e-08,
+      "loss": 0.0224,
+      "step": 2253
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 9.166666666666665e-08,
+      "loss": 0.4097,
+      "step": 2254
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 9.14855072463768e-08,
+      "loss": 0.0745,
+      "step": 2255
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 9.130434782608696e-08,
+      "loss": 0.0078,
+      "step": 2256
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 9.11231884057971e-08,
+      "loss": 0.4228,
+      "step": 2257
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 9.094202898550723e-08,
+      "loss": 0.3581,
+      "step": 2258
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 9.076086956521739e-08,
+      "loss": 0.0222,
+      "step": 2259
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 9.057971014492754e-08,
+      "loss": 0.0069,
+      "step": 2260
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 9.039855072463768e-08,
+      "loss": 0.0085,
+      "step": 2261
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 9.021739130434783e-08,
+      "loss": 0.2762,
+      "step": 2262
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 9.003623188405797e-08,
+      "loss": 0.1959,
+      "step": 2263
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 8.985507246376812e-08,
+      "loss": 0.0096,
+      "step": 2264
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 8.967391304347826e-08,
+      "loss": 0.2439,
+      "step": 2265
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 8.949275362318841e-08,
+      "loss": 0.2657,
+      "step": 2266
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 8.931159420289855e-08,
+      "loss": 0.0478,
+      "step": 2267
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 8.913043478260868e-08,
+      "loss": 0.2622,
+      "step": 2268
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 8.894927536231884e-08,
+      "loss": 0.0312,
+      "step": 2269
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 8.876811594202899e-08,
+      "loss": 0.2144,
+      "step": 2270
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 8.858695652173913e-08,
+      "loss": 0.0153,
+      "step": 2271
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 8.840579710144926e-08,
+      "loss": 0.1296,
+      "step": 2272
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 8.822463768115942e-08,
+      "loss": 0.5336,
+      "step": 2273
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 8.804347826086957e-08,
+      "loss": 0.0507,
+      "step": 2274
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 8.78623188405797e-08,
+      "loss": 0.0117,
+      "step": 2275
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 8.768115942028984e-08,
+      "loss": 0.0108,
+      "step": 2276
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 8.75e-08,
+      "loss": 0.0237,
+      "step": 2277
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 8.731884057971015e-08,
+      "loss": 0.0304,
+      "step": 2278
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 8.713768115942029e-08,
+      "loss": 0.0456,
+      "step": 2279
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 8.695652173913042e-08,
+      "loss": 0.3048,
+      "step": 2280
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 8.677536231884058e-08,
+      "loss": 0.1712,
+      "step": 2281
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 8.659420289855073e-08,
+      "loss": 0.2436,
+      "step": 2282
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 8.641304347826087e-08,
+      "loss": 0.0211,
+      "step": 2283
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 8.6231884057971e-08,
+      "loss": 0.2316,
+      "step": 2284
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 8.605072463768114e-08,
+      "loss": 0.0709,
+      "step": 2285
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 8.586956521739131e-08,
+      "loss": 0.0241,
+      "step": 2286
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 8.568840579710145e-08,
+      "loss": 0.1683,
+      "step": 2287
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 8.550724637681158e-08,
+      "loss": 0.1667,
+      "step": 2288
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 8.532608695652172e-08,
+      "loss": 0.4022,
+      "step": 2289
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 8.514492753623189e-08,
+      "loss": 0.2171,
+      "step": 2290
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 8.496376811594203e-08,
+      "loss": 0.1875,
+      "step": 2291
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 8.478260869565216e-08,
+      "loss": 0.0079,
+      "step": 2292
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 8.460144927536232e-08,
+      "loss": 0.0162,
+      "step": 2293
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 8.442028985507247e-08,
+      "loss": 0.1112,
+      "step": 2294
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 8.42391304347826e-08,
+      "loss": 0.0184,
+      "step": 2295
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 8.405797101449275e-08,
+      "loss": 0.4017,
+      "step": 2296
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 8.38768115942029e-08,
+      "loss": 0.4073,
+      "step": 2297
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 8.369565217391305e-08,
+      "loss": 0.0112,
+      "step": 2298
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 8.351449275362319e-08,
+      "loss": 0.2845,
+      "step": 2299
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 8.333333333333333e-08,
+      "loss": 0.3861,
+      "step": 2300
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.8651960784313726,
+      "eval_combined_score": 0.8840971573814711,
+      "eval_f1": 0.9029982363315697,
+      "eval_loss": 0.4259922504425049,
+      "eval_runtime": 8.8168,
+      "eval_samples_per_second": 46.275,
+      "eval_steps_per_second": 2.949,
+      "step": 2300
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 8.315217391304348e-08,
+      "loss": 0.0119,
+      "step": 2301
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 8.297101449275362e-08,
+      "loss": 0.0121,
+      "step": 2302
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 8.278985507246377e-08,
+      "loss": 0.1428,
+      "step": 2303
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 8.26086956521739e-08,
+      "loss": 0.4148,
+      "step": 2304
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 8.242753623188406e-08,
+      "loss": 0.0195,
+      "step": 2305
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 8.22463768115942e-08,
+      "loss": 0.2681,
+      "step": 2306
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 8.206521739130435e-08,
+      "loss": 0.0538,
+      "step": 2307
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 8.188405797101449e-08,
+      "loss": 0.2419,
+      "step": 2308
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 8.170289855072464e-08,
+      "loss": 0.0322,
+      "step": 2309
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 8.152173913043478e-08,
+      "loss": 0.0053,
+      "step": 2310
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 8.134057971014493e-08,
+      "loss": 0.0786,
+      "step": 2311
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 8.115942028985507e-08,
+      "loss": 0.0525,
+      "step": 2312
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 8.097826086956522e-08,
+      "loss": 0.1565,
+      "step": 2313
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 8.079710144927536e-08,
+      "loss": 0.0074,
+      "step": 2314
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 8.061594202898551e-08,
+      "loss": 0.0161,
+      "step": 2315
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 8.043478260869565e-08,
+      "loss": 0.0072,
+      "step": 2316
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 8.02536231884058e-08,
+      "loss": 0.01,
+      "step": 2317
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 8.007246376811594e-08,
+      "loss": 0.0728,
+      "step": 2318
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 7.989130434782607e-08,
+      "loss": 0.0932,
+      "step": 2319
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 7.971014492753623e-08,
+      "loss": 0.0268,
+      "step": 2320
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 7.952898550724638e-08,
+      "loss": 0.118,
+      "step": 2321
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 7.934782608695652e-08,
+      "loss": 0.137,
+      "step": 2322
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 7.916666666666665e-08,
+      "loss": 0.041,
+      "step": 2323
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 7.898550724637682e-08,
+      "loss": 0.0762,
+      "step": 2324
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 7.880434782608696e-08,
+      "loss": 0.2416,
+      "step": 2325
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 7.86231884057971e-08,
+      "loss": 0.1186,
+      "step": 2326
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 7.844202898550723e-08,
+      "loss": 0.126,
+      "step": 2327
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 7.82608695652174e-08,
+      "loss": 0.0185,
+      "step": 2328
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 7.807971014492754e-08,
+      "loss": 0.0171,
+      "step": 2329
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 7.789855072463768e-08,
+      "loss": 0.0269,
+      "step": 2330
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 7.771739130434781e-08,
+      "loss": 0.0186,
+      "step": 2331
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 7.753623188405798e-08,
+      "loss": 0.2938,
+      "step": 2332
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 7.735507246376812e-08,
+      "loss": 0.0208,
+      "step": 2333
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 7.717391304347826e-08,
+      "loss": 0.0142,
+      "step": 2334
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 7.69927536231884e-08,
+      "loss": 0.3289,
+      "step": 2335
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 7.681159420289855e-08,
+      "loss": 0.1486,
+      "step": 2336
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 7.66304347826087e-08,
+      "loss": 0.0224,
+      "step": 2337
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 7.644927536231884e-08,
+      "loss": 0.1559,
+      "step": 2338
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 7.626811594202897e-08,
+      "loss": 0.1076,
+      "step": 2339
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 7.608695652173913e-08,
+      "loss": 0.3062,
+      "step": 2340
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 7.590579710144928e-08,
+      "loss": 0.8031,
+      "step": 2341
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 7.572463768115942e-08,
+      "loss": 0.0155,
+      "step": 2342
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 7.554347826086955e-08,
+      "loss": 0.3764,
+      "step": 2343
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 7.53623188405797e-08,
+      "loss": 0.1399,
+      "step": 2344
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 7.518115942028986e-08,
+      "loss": 0.1432,
+      "step": 2345
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 7.5e-08,
+      "loss": 0.012,
+      "step": 2346
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 7.481884057971014e-08,
+      "loss": 0.0218,
+      "step": 2347
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 7.463768115942029e-08,
+      "loss": 0.006,
+      "step": 2348
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 7.445652173913044e-08,
+      "loss": 0.0559,
+      "step": 2349
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 7.427536231884058e-08,
+      "loss": 0.0142,
+      "step": 2350
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 7.409420289855072e-08,
+      "loss": 0.0481,
+      "step": 2351
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 7.391304347826087e-08,
+      "loss": 0.0907,
+      "step": 2352
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 7.3731884057971e-08,
+      "loss": 0.0044,
+      "step": 2353
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 7.355072463768116e-08,
+      "loss": 0.0248,
+      "step": 2354
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 7.336956521739131e-08,
+      "loss": 0.0123,
+      "step": 2355
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 7.318840579710145e-08,
+      "loss": 0.215,
+      "step": 2356
+    },
+    {
+      "epoch": 10.25,
+      "learning_rate": 7.300724637681159e-08,
+      "loss": 0.0648,
+      "step": 2357
+    },
+    {
+      "epoch": 10.25,
+      "learning_rate": 7.282608695652174e-08,
+      "loss": 0.0184,
+      "step": 2358
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 7.264492753623189e-08,
+      "loss": 0.0069,
+      "step": 2359
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 7.246376811594203e-08,
+      "loss": 0.1791,
+      "step": 2360
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 7.228260869565217e-08,
+      "loss": 0.0329,
+      "step": 2361
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 7.210144927536232e-08,
+      "loss": 0.0139,
+      "step": 2362
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 7.192028985507247e-08,
+      "loss": 0.2377,
+      "step": 2363
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 7.173913043478261e-08,
+      "loss": 0.0075,
+      "step": 2364
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 7.155797101449275e-08,
+      "loss": 0.1068,
+      "step": 2365
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 7.137681159420288e-08,
+      "loss": 0.183,
+      "step": 2366
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 7.119565217391305e-08,
+      "loss": 0.2207,
+      "step": 2367
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 7.101449275362319e-08,
+      "loss": 0.2975,
+      "step": 2368
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 7.083333333333333e-08,
+      "loss": 0.0067,
+      "step": 2369
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 7.065217391304346e-08,
+      "loss": 0.3651,
+      "step": 2370
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 7.047101449275363e-08,
+      "loss": 0.0085,
+      "step": 2371
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 7.028985507246377e-08,
+      "loss": 0.3227,
+      "step": 2372
+    },
+    {
+      "epoch": 10.32,
+      "learning_rate": 7.01086956521739e-08,
+      "loss": 0.3214,
+      "step": 2373
+    },
+    {
+      "epoch": 10.32,
+      "learning_rate": 6.992753623188404e-08,
+      "loss": 0.1884,
+      "step": 2374
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 6.974637681159421e-08,
+      "loss": 0.369,
+      "step": 2375
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 6.956521739130435e-08,
+      "loss": 0.0728,
+      "step": 2376
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 6.938405797101449e-08,
+      "loss": 0.2053,
+      "step": 2377
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 6.920289855072462e-08,
+      "loss": 0.102,
+      "step": 2378
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 6.902173913043479e-08,
+      "loss": 0.1538,
+      "step": 2379
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 6.884057971014493e-08,
+      "loss": 0.1706,
+      "step": 2380
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 6.865942028985507e-08,
+      "loss": 0.0258,
+      "step": 2381
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 6.84782608695652e-08,
+      "loss": 0.0666,
+      "step": 2382
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 6.829710144927536e-08,
+      "loss": 0.008,
+      "step": 2383
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 6.811594202898551e-08,
+      "loss": 0.2863,
+      "step": 2384
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 6.793478260869565e-08,
+      "loss": 0.3288,
+      "step": 2385
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 6.77536231884058e-08,
+      "loss": 0.0799,
+      "step": 2386
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 6.757246376811594e-08,
+      "loss": 0.06,
+      "step": 2387
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 6.739130434782609e-08,
+      "loss": 0.4472,
+      "step": 2388
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 6.721014492753623e-08,
+      "loss": 0.0166,
+      "step": 2389
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 6.702898550724638e-08,
+      "loss": 0.011,
+      "step": 2390
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 6.684782608695652e-08,
+      "loss": 0.0333,
+      "step": 2391
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 6.666666666666667e-08,
+      "loss": 0.3874,
+      "step": 2392
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 6.64855072463768e-08,
+      "loss": 0.3683,
+      "step": 2393
+    },
+    {
+      "epoch": 10.41,
+      "learning_rate": 6.630434782608696e-08,
+      "loss": 0.0323,
+      "step": 2394
+    },
+    {
+      "epoch": 10.41,
+      "learning_rate": 6.61231884057971e-08,
+      "loss": 0.4114,
+      "step": 2395
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 6.594202898550725e-08,
+      "loss": 0.0429,
+      "step": 2396
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 6.576086956521739e-08,
+      "loss": 0.0084,
+      "step": 2397
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 6.557971014492754e-08,
+      "loss": 0.3101,
+      "step": 2398
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 6.539855072463768e-08,
+      "loss": 0.0117,
+      "step": 2399
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 6.521739130434782e-08,
+      "loss": 0.0746,
+      "step": 2400
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 6.503623188405797e-08,
+      "loss": 0.6855,
+      "step": 2401
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 6.485507246376812e-08,
+      "loss": 0.3405,
+      "step": 2402
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 6.467391304347826e-08,
+      "loss": 0.0699,
+      "step": 2403
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 6.44927536231884e-08,
+      "loss": 0.0365,
+      "step": 2404
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 6.431159420289855e-08,
+      "loss": 0.0539,
+      "step": 2405
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 6.41304347826087e-08,
+      "loss": 0.0491,
+      "step": 2406
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 6.394927536231884e-08,
+      "loss": 0.3278,
+      "step": 2407
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 6.376811594202898e-08,
+      "loss": 0.0546,
+      "step": 2408
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 6.358695652173913e-08,
+      "loss": 0.0045,
+      "step": 2409
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 6.340579710144928e-08,
+      "loss": 0.0631,
+      "step": 2410
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 6.322463768115942e-08,
+      "loss": 0.0216,
+      "step": 2411
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 6.304347826086956e-08,
+      "loss": 0.559,
+      "step": 2412
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 6.286231884057971e-08,
+      "loss": 0.1843,
+      "step": 2413
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 6.268115942028986e-08,
+      "loss": 0.0377,
+      "step": 2414
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 6.25e-08,
+      "loss": 0.0172,
+      "step": 2415
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 6.231884057971015e-08,
+      "loss": 0.2519,
+      "step": 2416
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 6.213768115942029e-08,
+      "loss": 0.0244,
+      "step": 2417
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 6.195652173913044e-08,
+      "loss": 0.0698,
+      "step": 2418
+    },
+    {
+      "epoch": 10.52,
+      "learning_rate": 6.177536231884058e-08,
+      "loss": 0.1862,
+      "step": 2419
+    },
+    {
+      "epoch": 10.52,
+      "learning_rate": 6.159420289855073e-08,
+      "loss": 0.0135,
+      "step": 2420
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 6.141304347826087e-08,
+      "loss": 0.0798,
+      "step": 2421
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 6.123188405797102e-08,
+      "loss": 0.128,
+      "step": 2422
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 6.105072463768116e-08,
+      "loss": 0.0559,
+      "step": 2423
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 6.086956521739131e-08,
+      "loss": 0.0075,
+      "step": 2424
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 6.068840579710145e-08,
+      "loss": 0.0882,
+      "step": 2425
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 6.05072463768116e-08,
+      "loss": 0.2089,
+      "step": 2426
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 6.032608695652174e-08,
+      "loss": 0.3118,
+      "step": 2427
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 6.014492753623189e-08,
+      "loss": 0.1534,
+      "step": 2428
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 5.996376811594203e-08,
+      "loss": 0.0147,
+      "step": 2429
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 5.978260869565218e-08,
+      "loss": 0.3376,
+      "step": 2430
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 5.960144927536232e-08,
+      "loss": 0.1212,
+      "step": 2431
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 5.942028985507246e-08,
+      "loss": 0.0215,
+      "step": 2432
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 5.92391304347826e-08,
+      "loss": 0.06,
+      "step": 2433
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 5.905797101449275e-08,
+      "loss": 0.0574,
+      "step": 2434
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 5.887681159420289e-08,
+      "loss": 0.0254,
+      "step": 2435
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 5.869565217391304e-08,
+      "loss": 0.0069,
+      "step": 2436
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 5.851449275362318e-08,
+      "loss": 0.2238,
+      "step": 2437
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 5.833333333333333e-08,
+      "loss": 0.0095,
+      "step": 2438
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 5.815217391304347e-08,
+      "loss": 0.0593,
+      "step": 2439
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 5.797101449275362e-08,
+      "loss": 0.0451,
+      "step": 2440
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 5.778985507246376e-08,
+      "loss": 0.4327,
+      "step": 2441
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 5.760869565217391e-08,
+      "loss": 0.0088,
+      "step": 2442
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 5.742753623188405e-08,
+      "loss": 0.0433,
+      "step": 2443
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 5.72463768115942e-08,
+      "loss": 0.7125,
+      "step": 2444
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 5.706521739130434e-08,
+      "loss": 0.0431,
+      "step": 2445
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 5.688405797101449e-08,
+      "loss": 0.0259,
+      "step": 2446
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 5.670289855072464e-08,
+      "loss": 0.0264,
+      "step": 2447
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 5.6521739130434777e-08,
+      "loss": 0.0296,
+      "step": 2448
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 5.634057971014493e-08,
+      "loss": 0.1006,
+      "step": 2449
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 5.6159420289855067e-08,
+      "loss": 0.2048,
+      "step": 2450
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 5.597826086956522e-08,
+      "loss": 0.0095,
+      "step": 2451
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 5.579710144927536e-08,
+      "loss": 0.1389,
+      "step": 2452
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 5.561594202898551e-08,
+      "loss": 0.4487,
+      "step": 2453
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 5.543478260869565e-08,
+      "loss": 0.1058,
+      "step": 2454
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 5.52536231884058e-08,
+      "loss": 0.0538,
+      "step": 2455
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 5.507246376811594e-08,
+      "loss": 0.0538,
+      "step": 2456
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 5.489130434782609e-08,
+      "loss": 0.0228,
+      "step": 2457
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 5.471014492753623e-08,
+      "loss": 0.5987,
+      "step": 2458
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 5.452898550724638e-08,
+      "loss": 0.5354,
+      "step": 2459
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 5.434782608695652e-08,
+      "loss": 0.2397,
+      "step": 2460
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 5.416666666666667e-08,
+      "loss": 0.106,
+      "step": 2461
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 5.398550724637681e-08,
+      "loss": 0.0836,
+      "step": 2462
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 5.380434782608696e-08,
+      "loss": 0.3414,
+      "step": 2463
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 5.36231884057971e-08,
+      "loss": 0.026,
+      "step": 2464
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 5.344202898550724e-08,
+      "loss": 0.2771,
+      "step": 2465
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 5.326086956521739e-08,
+      "loss": 0.3864,
+      "step": 2466
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 5.307971014492753e-08,
+      "loss": 0.0153,
+      "step": 2467
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 5.289855072463768e-08,
+      "loss": 0.4104,
+      "step": 2468
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 5.271739130434782e-08,
+      "loss": 0.0136,
+      "step": 2469
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 5.253623188405797e-08,
+      "loss": 0.3096,
+      "step": 2470
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 5.235507246376811e-08,
+      "loss": 0.0155,
+      "step": 2471
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 5.217391304347826e-08,
+      "loss": 0.0157,
+      "step": 2472
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 5.19927536231884e-08,
+      "loss": 0.4385,
+      "step": 2473
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 5.181159420289855e-08,
+      "loss": 0.0087,
+      "step": 2474
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 5.163043478260869e-08,
+      "loss": 0.0877,
+      "step": 2475
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 5.144927536231884e-08,
+      "loss": 0.0354,
+      "step": 2476
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 5.126811594202898e-08,
+      "loss": 0.0306,
+      "step": 2477
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 5.1086956521739134e-08,
+      "loss": 0.3061,
+      "step": 2478
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 5.090579710144927e-08,
+      "loss": 0.1214,
+      "step": 2479
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 5.0724637681159424e-08,
+      "loss": 0.0326,
+      "step": 2480
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 5.054347826086956e-08,
+      "loss": 0.089,
+      "step": 2481
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 5.036231884057971e-08,
+      "loss": 0.0037,
+      "step": 2482
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 5.018115942028985e-08,
+      "loss": 0.0633,
+      "step": 2483
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 5e-08,
+      "loss": 0.0206,
+      "step": 2484
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 4.981884057971014e-08,
+      "loss": 0.2289,
+      "step": 2485
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 4.963768115942029e-08,
+      "loss": 0.206,
+      "step": 2486
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 4.945652173913043e-08,
+      "loss": 0.0108,
+      "step": 2487
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 4.927536231884058e-08,
+      "loss": 0.0145,
+      "step": 2488
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 4.909420289855072e-08,
+      "loss": 0.1829,
+      "step": 2489
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.891304347826087e-08,
+      "loss": 0.0243,
+      "step": 2490
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.873188405797101e-08,
+      "loss": 0.0103,
+      "step": 2491
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.855072463768116e-08,
+      "loss": 0.0382,
+      "step": 2492
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 4.83695652173913e-08,
+      "loss": 0.0103,
+      "step": 2493
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 4.818840579710145e-08,
+      "loss": 0.0167,
+      "step": 2494
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 4.800724637681159e-08,
+      "loss": 0.2958,
+      "step": 2495
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 4.782608695652174e-08,
+      "loss": 0.1134,
+      "step": 2496
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 4.7644927536231877e-08,
+      "loss": 0.0215,
+      "step": 2497
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 4.746376811594203e-08,
+      "loss": 0.4622,
+      "step": 2498
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 4.7282608695652167e-08,
+      "loss": 0.0328,
+      "step": 2499
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 4.710144927536232e-08,
+      "loss": 0.0272,
+      "step": 2500
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 4.6920289855072457e-08,
+      "loss": 0.1542,
+      "step": 2501
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 4.673913043478261e-08,
+      "loss": 0.2742,
+      "step": 2502
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 4.655797101449275e-08,
+      "loss": 0.0052,
+      "step": 2503
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 4.63768115942029e-08,
+      "loss": 0.3738,
+      "step": 2504
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 4.619565217391304e-08,
+      "loss": 0.0211,
+      "step": 2505
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 4.601449275362319e-08,
+      "loss": 0.0925,
+      "step": 2506
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 4.583333333333333e-08,
+      "loss": 0.32,
+      "step": 2507
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 4.565217391304348e-08,
+      "loss": 0.0258,
+      "step": 2508
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 4.547101449275362e-08,
+      "loss": 0.3299,
+      "step": 2509
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 4.528985507246377e-08,
+      "loss": 0.023,
+      "step": 2510
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 4.5108695652173914e-08,
+      "loss": 0.023,
+      "step": 2511
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 4.492753623188406e-08,
+      "loss": 0.0092,
+      "step": 2512
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 4.4746376811594204e-08,
+      "loss": 0.0088,
+      "step": 2513
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 4.456521739130434e-08,
+      "loss": 0.4817,
+      "step": 2514
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 4.4384057971014494e-08,
+      "loss": 0.0522,
+      "step": 2515
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.420289855072463e-08,
+      "loss": 0.0198,
+      "step": 2516
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.4021739130434784e-08,
+      "loss": 0.0265,
+      "step": 2517
+    },
+    {
+      "epoch": 10.95,
+      "learning_rate": 4.384057971014492e-08,
+      "loss": 0.0387,
+      "step": 2518
+    },
+    {
+      "epoch": 10.95,
+      "learning_rate": 4.3659420289855074e-08,
+      "loss": 0.0252,
+      "step": 2519
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 4.347826086956521e-08,
+      "loss": 0.1275,
+      "step": 2520
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 4.3297101449275364e-08,
+      "loss": 0.1267,
+      "step": 2521
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 4.31159420289855e-08,
+      "loss": 0.0712,
+      "step": 2522
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 4.2934782608695654e-08,
+      "loss": 0.0608,
+      "step": 2523
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 4.275362318840579e-08,
+      "loss": 0.2419,
+      "step": 2524
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 4.2572463768115944e-08,
+      "loss": 0.0114,
+      "step": 2525
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 4.239130434782608e-08,
+      "loss": 0.3378,
+      "step": 2526
+    },
+    {
+      "epoch": 10.99,
+      "learning_rate": 4.2210144927536234e-08,
+      "loss": 0.4217,
+      "step": 2527
+    },
+    {
+      "epoch": 10.99,
+      "learning_rate": 4.202898550724637e-08,
+      "loss": 0.0631,
+      "step": 2528
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 4.1847826086956524e-08,
+      "loss": 0.2973,
+      "step": 2529
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 4.166666666666666e-08,
+      "loss": 0.0028,
+      "step": 2530
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.8946078431372549,
+      "eval_combined_score": 0.9089795008912656,
+      "eval_f1": 0.9233511586452763,
+      "eval_loss": 0.3619478940963745,
+      "eval_runtime": 9.2314,
+      "eval_samples_per_second": 44.197,
+      "eval_steps_per_second": 2.816,
+      "step": 2530
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 4.148550724637681e-08,
+      "loss": 0.3846,
+      "step": 2531
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 4.130434782608695e-08,
+      "loss": 0.4157,
+      "step": 2532
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 4.11231884057971e-08,
+      "loss": 0.1946,
+      "step": 2533
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 4.094202898550724e-08,
+      "loss": 0.0185,
+      "step": 2534
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 4.076086956521739e-08,
+      "loss": 0.0681,
+      "step": 2535
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.057971014492753e-08,
+      "loss": 0.0182,
+      "step": 2536
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.039855072463768e-08,
+      "loss": 0.7905,
+      "step": 2537
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.021739130434782e-08,
+      "loss": 0.0232,
+      "step": 2538
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 4.003623188405797e-08,
+      "loss": 0.0081,
+      "step": 2539
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 3.985507246376811e-08,
+      "loss": 0.0254,
+      "step": 2540
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 3.967391304347826e-08,
+      "loss": 0.2483,
+      "step": 2541
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 3.949275362318841e-08,
+      "loss": 0.4101,
+      "step": 2542
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.931159420289855e-08,
+      "loss": 0.2726,
+      "step": 2543
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 3.91304347826087e-08,
+      "loss": 0.1715,
+      "step": 2544
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 3.894927536231884e-08,
+      "loss": 0.4379,
+      "step": 2545
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 3.876811594202899e-08,
+      "loss": 0.051,
+      "step": 2546
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 3.858695652173913e-08,
+      "loss": 0.6221,
+      "step": 2547
+    },
+    {
+      "epoch": 11.08,
+      "learning_rate": 3.840579710144927e-08,
+      "loss": 0.133,
+      "step": 2548
+    },
+    {
+      "epoch": 11.08,
+      "learning_rate": 3.822463768115942e-08,
+      "loss": 0.5378,
+      "step": 2549
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 3.804347826086956e-08,
+      "loss": 0.0079,
+      "step": 2550
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 3.786231884057971e-08,
+      "loss": 0.1594,
+      "step": 2551
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.768115942028985e-08,
+      "loss": 0.1192,
+      "step": 2552
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.75e-08,
+      "loss": 0.3616,
+      "step": 2553
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 3.7318840579710143e-08,
+      "loss": 0.0848,
+      "step": 2554
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 3.713768115942029e-08,
+      "loss": 0.0314,
+      "step": 2555
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 3.6956521739130433e-08,
+      "loss": 0.0948,
+      "step": 2556
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 3.677536231884058e-08,
+      "loss": 0.0731,
+      "step": 2557
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 3.6594202898550723e-08,
+      "loss": 0.0977,
+      "step": 2558
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 3.641304347826087e-08,
+      "loss": 0.2225,
+      "step": 2559
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 3.6231884057971014e-08,
+      "loss": 0.0157,
+      "step": 2560
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 3.605072463768116e-08,
+      "loss": 0.4609,
+      "step": 2561
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.5869565217391304e-08,
+      "loss": 0.1208,
+      "step": 2562
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 3.568840579710144e-08,
+      "loss": 0.1116,
+      "step": 2563
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 3.5507246376811594e-08,
+      "loss": 0.0559,
+      "step": 2564
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 3.532608695652173e-08,
+      "loss": 0.2852,
+      "step": 2565
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 3.5144927536231884e-08,
+      "loss": 0.161,
+      "step": 2566
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 3.496376811594202e-08,
+      "loss": 0.0207,
+      "step": 2567
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 3.4782608695652174e-08,
+      "loss": 0.0276,
+      "step": 2568
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 3.460144927536231e-08,
+      "loss": 0.0321,
+      "step": 2569
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 3.4420289855072464e-08,
+      "loss": 0.0288,
+      "step": 2570
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.42391304347826e-08,
+      "loss": 0.4073,
+      "step": 2571
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 3.4057971014492754e-08,
+      "loss": 0.055,
+      "step": 2572
+    },
+    {
+      "epoch": 11.19,
+      "learning_rate": 3.38768115942029e-08,
+      "loss": 0.0377,
+      "step": 2573
+    },
+    {
+      "epoch": 11.19,
+      "learning_rate": 3.3695652173913044e-08,
+      "loss": 0.0138,
+      "step": 2574
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 3.351449275362319e-08,
+      "loss": 0.1369,
+      "step": 2575
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 3.3333333333333334e-08,
+      "loss": 0.1962,
+      "step": 2576
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 3.315217391304348e-08,
+      "loss": 0.0454,
+      "step": 2577
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 3.2971014492753624e-08,
+      "loss": 0.0619,
+      "step": 2578
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 3.278985507246377e-08,
+      "loss": 0.0172,
+      "step": 2579
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.260869565217391e-08,
+      "loss": 0.0172,
+      "step": 2580
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 3.242753623188406e-08,
+      "loss": 0.0943,
+      "step": 2581
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 3.22463768115942e-08,
+      "loss": 0.055,
+      "step": 2582
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 3.206521739130435e-08,
+      "loss": 0.4246,
+      "step": 2583
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 3.188405797101449e-08,
+      "loss": 0.0369,
+      "step": 2584
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 3.170289855072464e-08,
+      "loss": 0.0328,
+      "step": 2585
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 3.152173913043478e-08,
+      "loss": 0.1434,
+      "step": 2586
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.134057971014493e-08,
+      "loss": 0.0249,
+      "step": 2587
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 3.1159420289855074e-08,
+      "loss": 0.0231,
+      "step": 2588
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 3.097826086956522e-08,
+      "loss": 0.3619,
+      "step": 2589
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 3.0797101449275364e-08,
+      "loss": 0.1027,
+      "step": 2590
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 3.061594202898551e-08,
+      "loss": 0.0799,
+      "step": 2591
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 3.0434782608695655e-08,
+      "loss": 0.2879,
+      "step": 2592
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 3.02536231884058e-08,
+      "loss": 0.3086,
+      "step": 2593
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 3.0072463768115945e-08,
+      "loss": 0.2721,
+      "step": 2594
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 2.989130434782609e-08,
+      "loss": 0.0846,
+      "step": 2595
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 2.971014492753623e-08,
+      "loss": 0.0081,
+      "step": 2596
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 2.9528985507246376e-08,
+      "loss": 0.3924,
+      "step": 2597
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 2.934782608695652e-08,
+      "loss": 0.169,
+      "step": 2598
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 2.9166666666666666e-08,
+      "loss": 0.1229,
+      "step": 2599
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 2.898550724637681e-08,
+      "loss": 0.0179,
+      "step": 2600
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 2.8804347826086957e-08,
+      "loss": 0.0452,
+      "step": 2601
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 2.86231884057971e-08,
+      "loss": 0.0372,
+      "step": 2602
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 2.8442028985507247e-08,
+      "loss": 0.5749,
+      "step": 2603
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 2.8260869565217388e-08,
+      "loss": 0.2381,
+      "step": 2604
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 2.8079710144927533e-08,
+      "loss": 0.3063,
+      "step": 2605
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 2.789855072463768e-08,
+      "loss": 0.2578,
+      "step": 2606
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 2.7717391304347823e-08,
+      "loss": 0.0384,
+      "step": 2607
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 2.753623188405797e-08,
+      "loss": 0.0062,
+      "step": 2608
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 2.7355072463768113e-08,
+      "loss": 0.1248,
+      "step": 2609
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 2.717391304347826e-08,
+      "loss": 0.1497,
+      "step": 2610
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 2.6992753623188404e-08,
+      "loss": 0.2666,
+      "step": 2611
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 2.681159420289855e-08,
+      "loss": 0.1911,
+      "step": 2612
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 2.6630434782608694e-08,
+      "loss": 0.0324,
+      "step": 2613
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 2.644927536231884e-08,
+      "loss": 0.0305,
+      "step": 2614
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 2.6268115942028984e-08,
+      "loss": 0.5515,
+      "step": 2615
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 2.608695652173913e-08,
+      "loss": 0.1066,
+      "step": 2616
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 2.5905797101449274e-08,
+      "loss": 0.1782,
+      "step": 2617
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 2.572463768115942e-08,
+      "loss": 0.0548,
+      "step": 2618
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 2.5543478260869567e-08,
+      "loss": 0.0417,
+      "step": 2619
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 2.5362318840579712e-08,
+      "loss": 0.0469,
+      "step": 2620
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 2.5181159420289854e-08,
+      "loss": 0.3549,
+      "step": 2621
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 2.5e-08,
+      "loss": 0.1697,
+      "step": 2622
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 2.4818840579710144e-08,
+      "loss": 0.0286,
+      "step": 2623
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 2.463768115942029e-08,
+      "loss": 0.0962,
+      "step": 2624
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 2.4456521739130434e-08,
+      "loss": 0.147,
+      "step": 2625
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 2.427536231884058e-08,
+      "loss": 0.0052,
+      "step": 2626
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 2.4094202898550724e-08,
+      "loss": 0.3852,
+      "step": 2627
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 2.391304347826087e-08,
+      "loss": 0.0402,
+      "step": 2628
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 2.3731884057971014e-08,
+      "loss": 0.0517,
+      "step": 2629
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 2.355072463768116e-08,
+      "loss": 0.5935,
+      "step": 2630
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 2.3369565217391304e-08,
+      "loss": 0.2443,
+      "step": 2631
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 2.318840579710145e-08,
+      "loss": 0.0168,
+      "step": 2632
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 2.3007246376811594e-08,
+      "loss": 0.1257,
+      "step": 2633
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 2.282608695652174e-08,
+      "loss": 0.0063,
+      "step": 2634
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 2.2644927536231884e-08,
+      "loss": 0.2767,
+      "step": 2635
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 2.246376811594203e-08,
+      "loss": 0.0226,
+      "step": 2636
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 2.228260869565217e-08,
+      "loss": 0.0211,
+      "step": 2637
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 2.2101449275362316e-08,
+      "loss": 0.0378,
+      "step": 2638
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 2.192028985507246e-08,
+      "loss": 0.0446,
+      "step": 2639
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 2.1739130434782606e-08,
+      "loss": 0.0068,
+      "step": 2640
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 2.155797101449275e-08,
+      "loss": 0.1999,
+      "step": 2641
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 2.1376811594202896e-08,
+      "loss": 0.0274,
+      "step": 2642
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 2.119565217391304e-08,
+      "loss": 0.0936,
+      "step": 2643
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 2.1014492753623186e-08,
+      "loss": 0.1155,
+      "step": 2644
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 2.083333333333333e-08,
+      "loss": 0.2681,
+      "step": 2645
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 2.0652173913043476e-08,
+      "loss": 0.0373,
+      "step": 2646
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 2.047101449275362e-08,
+      "loss": 0.0306,
+      "step": 2647
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 2.0289855072463766e-08,
+      "loss": 0.0871,
+      "step": 2648
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 2.010869565217391e-08,
+      "loss": 0.0852,
+      "step": 2649
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 1.9927536231884056e-08,
+      "loss": 0.0655,
+      "step": 2650
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 1.9746376811594205e-08,
+      "loss": 0.4805,
+      "step": 2651
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 1.956521739130435e-08,
+      "loss": 0.2606,
+      "step": 2652
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 1.9384057971014495e-08,
+      "loss": 0.1044,
+      "step": 2653
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 1.9202898550724637e-08,
+      "loss": 0.0138,
+      "step": 2654
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 1.902173913043478e-08,
+      "loss": 0.0452,
+      "step": 2655
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 1.8840579710144927e-08,
+      "loss": 0.027,
+      "step": 2656
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 1.8659420289855072e-08,
+      "loss": 0.0506,
+      "step": 2657
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 1.8478260869565217e-08,
+      "loss": 0.2928,
+      "step": 2658
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 1.8297101449275362e-08,
+      "loss": 0.2717,
+      "step": 2659
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 1.8115942028985507e-08,
+      "loss": 0.0428,
+      "step": 2660
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 1.7934782608695652e-08,
+      "loss": 1.1035,
+      "step": 2661
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 1.7753623188405797e-08,
+      "loss": 0.0044,
+      "step": 2662
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 1.7572463768115942e-08,
+      "loss": 0.1961,
+      "step": 2663
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 1.7391304347826087e-08,
+      "loss": 0.0437,
+      "step": 2664
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 1.7210144927536232e-08,
+      "loss": 0.0719,
+      "step": 2665
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 1.7028985507246377e-08,
+      "loss": 0.0159,
+      "step": 2666
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 1.6847826086956522e-08,
+      "loss": 0.2896,
+      "step": 2667
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 1.6666666666666667e-08,
+      "loss": 0.0094,
+      "step": 2668
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 1.6485507246376812e-08,
+      "loss": 0.1072,
+      "step": 2669
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 1.6304347826086954e-08,
+      "loss": 0.0594,
+      "step": 2670
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 1.61231884057971e-08,
+      "loss": 0.4341,
+      "step": 2671
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 1.5942028985507244e-08,
+      "loss": 0.4571,
+      "step": 2672
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 1.576086956521739e-08,
+      "loss": 0.0505,
+      "step": 2673
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 1.5579710144927537e-08,
+      "loss": 0.0274,
+      "step": 2674
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 1.5398550724637682e-08,
+      "loss": 0.305,
+      "step": 2675
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 1.5217391304347827e-08,
+      "loss": 0.0329,
+      "step": 2676
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 1.5036231884057972e-08,
+      "loss": 0.3676,
+      "step": 2677
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 1.4855072463768116e-08,
+      "loss": 0.0459,
+      "step": 2678
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 1.467391304347826e-08,
+      "loss": 0.1596,
+      "step": 2679
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 1.4492753623188406e-08,
+      "loss": 0.1405,
+      "step": 2680
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 1.431159420289855e-08,
+      "loss": 0.1973,
+      "step": 2681
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 1.4130434782608694e-08,
+      "loss": 0.0227,
+      "step": 2682
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 1.394927536231884e-08,
+      "loss": 0.0964,
+      "step": 2683
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 1.3768115942028984e-08,
+      "loss": 0.0219,
+      "step": 2684
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 1.358695652173913e-08,
+      "loss": 0.0252,
+      "step": 2685
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 1.3405797101449274e-08,
+      "loss": 0.0095,
+      "step": 2686
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 1.322463768115942e-08,
+      "loss": 0.0168,
+      "step": 2687
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 1.3043478260869564e-08,
+      "loss": 0.0138,
+      "step": 2688
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 1.286231884057971e-08,
+      "loss": 0.2176,
+      "step": 2689
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 1.2681159420289856e-08,
+      "loss": 0.2204,
+      "step": 2690
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 1.25e-08,
+      "loss": 0.3693,
+      "step": 2691
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 1.2318840579710144e-08,
+      "loss": 0.2058,
+      "step": 2692
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 1.213768115942029e-08,
+      "loss": 0.2971,
+      "step": 2693
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 1.1956521739130435e-08,
+      "loss": 0.0917,
+      "step": 2694
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 1.177536231884058e-08,
+      "loss": 0.0071,
+      "step": 2695
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 1.1594202898550725e-08,
+      "loss": 0.0188,
+      "step": 2696
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 1.141304347826087e-08,
+      "loss": 0.357,
+      "step": 2697
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 1.1231884057971015e-08,
+      "loss": 0.0161,
+      "step": 2698
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 1.1050724637681158e-08,
+      "loss": 0.044,
+      "step": 2699
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 1.0869565217391303e-08,
+      "loss": 0.0415,
+      "step": 2700
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 1.0688405797101448e-08,
+      "loss": 0.0349,
+      "step": 2701
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 1.0507246376811593e-08,
+      "loss": 0.007,
+      "step": 2702
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 1.0326086956521738e-08,
+      "loss": 0.0335,
+      "step": 2703
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 1.0144927536231883e-08,
+      "loss": 0.0112,
+      "step": 2704
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 9.963768115942028e-09,
+      "loss": 0.0918,
+      "step": 2705
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 9.782608695652175e-09,
+      "loss": 0.2982,
+      "step": 2706
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 9.601449275362318e-09,
+      "loss": 0.0176,
+      "step": 2707
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 9.420289855072463e-09,
+      "loss": 0.1529,
+      "step": 2708
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 9.239130434782608e-09,
+      "loss": 0.0134,
+      "step": 2709
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 9.057971014492753e-09,
+      "loss": 0.1326,
+      "step": 2710
+    },
+    {
+      "epoch": 11.79,
+      "learning_rate": 8.876811594202898e-09,
+      "loss": 0.0298,
+      "step": 2711
+    },
+    {
+      "epoch": 11.79,
+      "learning_rate": 8.695652173913043e-09,
+      "loss": 0.6665,
+      "step": 2712
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 8.514492753623188e-09,
+      "loss": 0.5776,
+      "step": 2713
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 8.333333333333334e-09,
+      "loss": 0.0381,
+      "step": 2714
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 8.152173913043477e-09,
+      "loss": 0.5108,
+      "step": 2715
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 7.971014492753622e-09,
+      "loss": 0.1376,
+      "step": 2716
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 7.789855072463769e-09,
+      "loss": 0.0466,
+      "step": 2717
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 7.608695652173914e-09,
+      "loss": 0.0082,
+      "step": 2718
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 7.427536231884058e-09,
+      "loss": 0.1375,
+      "step": 2719
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 7.246376811594203e-09,
+      "loss": 0.1397,
+      "step": 2720
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 7.065217391304347e-09,
+      "loss": 0.0884,
+      "step": 2721
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 6.884057971014492e-09,
+      "loss": 0.163,
+      "step": 2722
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 6.702898550724637e-09,
+      "loss": 0.0373,
+      "step": 2723
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 6.521739130434782e-09,
+      "loss": 0.328,
+      "step": 2724
+    },
+    {
+      "epoch": 11.85,
+      "learning_rate": 6.340579710144928e-09,
+      "loss": 0.1126,
+      "step": 2725
+    },
+    {
+      "epoch": 11.85,
+      "learning_rate": 6.159420289855072e-09,
+      "loss": 0.0126,
+      "step": 2726
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 5.978260869565217e-09,
+      "loss": 0.1461,
+      "step": 2727
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 5.797101449275362e-09,
+      "loss": 0.0996,
+      "step": 2728
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 5.615942028985507e-09,
+      "loss": 0.0131,
+      "step": 2729
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 5.4347826086956515e-09,
+      "loss": 0.0143,
+      "step": 2730
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 5.2536231884057966e-09,
+      "loss": 0.0967,
+      "step": 2731
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 5.072463768115942e-09,
+      "loss": 0.0073,
+      "step": 2732
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 4.8913043478260875e-09,
+      "loss": 0.1054,
+      "step": 2733
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 4.710144927536232e-09,
+      "loss": 0.3318,
+      "step": 2734
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 4.528985507246377e-09,
+      "loss": 0.0649,
+      "step": 2735
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 4.347826086956522e-09,
+      "loss": 0.0598,
+      "step": 2736
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 4.166666666666667e-09,
+      "loss": 0.5363,
+      "step": 2737
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 3.985507246376811e-09,
+      "loss": 0.0162,
+      "step": 2738
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 3.804347826086957e-09,
+      "loss": 0.0296,
+      "step": 2739
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 3.6231884057971014e-09,
+      "loss": 0.0161,
+      "step": 2740
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.442028985507246e-09,
+      "loss": 0.1584,
+      "step": 2741
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 3.260869565217391e-09,
+      "loss": 0.1024,
+      "step": 2742
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 3.079710144927536e-09,
+      "loss": 0.0183,
+      "step": 2743
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 2.898550724637681e-09,
+      "loss": 0.0965,
+      "step": 2744
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 2.7173913043478258e-09,
+      "loss": 0.135,
+      "step": 2745
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 2.536231884057971e-09,
+      "loss": 0.0493,
+      "step": 2746
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 2.355072463768116e-09,
+      "loss": 0.317,
+      "step": 2747
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 2.173913043478261e-09,
+      "loss": 0.0109,
+      "step": 2748
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 1.9927536231884055e-09,
+      "loss": 0.0046,
+      "step": 2749
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 1.8115942028985507e-09,
+      "loss": 0.0104,
+      "step": 2750
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 1.6304347826086955e-09,
+      "loss": 0.0144,
+      "step": 2751
+    },
+    {
+      "epoch": 11.97,
+      "learning_rate": 1.4492753623188406e-09,
+      "loss": 0.0211,
+      "step": 2752
+    },
+    {
+      "epoch": 11.97,
+      "learning_rate": 1.2681159420289854e-09,
+      "loss": 0.0136,
+      "step": 2753
+    },
+    {
+      "epoch": 11.97,
+      "learning_rate": 1.0869565217391304e-09,
+      "loss": 0.0308,
+      "step": 2754
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 9.057971014492754e-10,
+      "loss": 0.054,
+      "step": 2755
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 7.246376811594203e-10,
+      "loss": 0.0839,
+      "step": 2756
+    },
+    {
+      "epoch": 11.99,
+      "learning_rate": 5.434782608695652e-10,
+      "loss": 0.0107,
+      "step": 2757
+    },
+    {
+      "epoch": 11.99,
+      "learning_rate": 3.6231884057971014e-10,
+      "loss": 0.3865,
+      "step": 2758
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 1.8115942028985507e-10,
+      "loss": 0.1425,
+      "step": 2759
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 0.0,
+      "loss": 0.0957,
+      "step": 2760
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.8970588235294118,
+      "eval_combined_score": 0.9114269382664727,
+      "eval_f1": 0.9257950530035336,
+      "eval_loss": 0.38710883259773254,
+      "eval_runtime": 9.2231,
+      "eval_samples_per_second": 44.237,
+      "eval_steps_per_second": 2.819,
+      "step": 2760
+    },
+    {
+      "epoch": 12.0,
+      "step": 2760,
+      "total_flos": 4213608833064960.0,
+      "train_loss": 0.13981886300810045,
+      "train_runtime": 2408.1272,
+      "train_samples_per_second": 18.278,
+      "train_steps_per_second": 1.146
+    }
+  ],
+  "max_steps": 2760,
+  "num_train_epochs": 12,
+  "total_flos": 4213608833064960.0,
+  "trial_name": null,
+  "trial_params": null
+}