diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,13288 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 20.0,
+  "eval_steps": 500,
+  "global_step": 2180,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0,
+      "loss": 4.8333,
+      "step": 1
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0,
+      "loss": 2.3065,
+      "step": 2
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0,
+      "loss": 4.4461,
+      "step": 3
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0,
+      "loss": 4.7319,
+      "step": 4
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2e-05,
+      "loss": 5.0436,
+      "step": 5
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2e-05,
+      "loss": 6.2135,
+      "step": 6
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4e-05,
+      "loss": 3.4661,
+      "step": 7
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 6e-05,
+      "loss": 5.5063,
+      "step": 8
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 8e-05,
+      "loss": 2.5579,
+      "step": 9
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001,
+      "loss": 5.7731,
+      "step": 10
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.995402298850575e-05,
+      "loss": 2.9313,
+      "step": 11
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.99080459770115e-05,
+      "loss": 2.2405,
+      "step": 12
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.986206896551724e-05,
+      "loss": 2.2503,
+      "step": 13
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.981609195402299e-05,
+      "loss": 3.7502,
+      "step": 14
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.977011494252874e-05,
+      "loss": 4.8206,
+      "step": 15
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.972413793103448e-05,
+      "loss": 2.1221,
+      "step": 16
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.967816091954023e-05,
+      "loss": 2.681,
+      "step": 17
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.963218390804599e-05,
+      "loss": 2.9936,
+      "step": 18
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.958620689655172e-05,
+      "loss": 1.6324,
+      "step": 19
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.954022988505748e-05,
+      "loss": 2.1403,
+      "step": 20
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.949425287356321e-05,
+      "loss": 1.9128,
+      "step": 21
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.944827586206897e-05,
+      "loss": 3.1573,
+      "step": 22
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.940229885057472e-05,
+      "loss": 2.0411,
+      "step": 23
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.935632183908047e-05,
+      "loss": 2.3495,
+      "step": 24
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.931034482758621e-05,
+      "loss": 1.3057,
+      "step": 25
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.926436781609196e-05,
+      "loss": 2.1814,
+      "step": 26
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.92183908045977e-05,
+      "loss": 1.9983,
+      "step": 27
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.917241379310345e-05,
+      "loss": 2.7244,
+      "step": 28
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.91264367816092e-05,
+      "loss": 3.1512,
+      "step": 29
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.908045977011494e-05,
+      "loss": 3.564,
+      "step": 30
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.903448275862069e-05,
+      "loss": 4.3232,
+      "step": 31
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.898850574712644e-05,
+      "loss": 2.0438,
+      "step": 32
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.894252873563218e-05,
+      "loss": 3.1714,
+      "step": 33
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.889655172413793e-05,
+      "loss": 2.2101,
+      "step": 34
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.885057471264369e-05,
+      "loss": 2.0324,
+      "step": 35
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.880459770114942e-05,
+      "loss": 1.7595,
+      "step": 36
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.875862068965518e-05,
+      "loss": 1.9976,
+      "step": 37
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.871264367816092e-05,
+      "loss": 2.296,
+      "step": 38
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.866666666666668e-05,
+      "loss": 1.2148,
+      "step": 39
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.862068965517242e-05,
+      "loss": 1.5207,
+      "step": 40
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.857471264367817e-05,
+      "loss": 1.5712,
+      "step": 41
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.852873563218391e-05,
+      "loss": 1.5386,
+      "step": 42
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.848275862068966e-05,
+      "loss": 2.1024,
+      "step": 43
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.843678160919541e-05,
+      "loss": 2.0523,
+      "step": 44
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.839080459770115e-05,
+      "loss": 2.8,
+      "step": 45
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.83448275862069e-05,
+      "loss": 2.5403,
+      "step": 46
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.829885057471266e-05,
+      "loss": 1.9564,
+      "step": 47
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.825287356321839e-05,
+      "loss": 1.6888,
+      "step": 48
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.820689655172415e-05,
+      "loss": 2.2113,
+      "step": 49
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.816091954022988e-05,
+      "loss": 2.3537,
+      "step": 50
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.811494252873563e-05,
+      "loss": 1.487,
+      "step": 51
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.806896551724139e-05,
+      "loss": 1.7331,
+      "step": 52
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.802298850574712e-05,
+      "loss": 2.9036,
+      "step": 53
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.797701149425288e-05,
+      "loss": 2.0968,
+      "step": 54
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.793103448275862e-05,
+      "loss": 1.8345,
+      "step": 55
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.788505747126438e-05,
+      "loss": 2.444,
+      "step": 56
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.783908045977012e-05,
+      "loss": 1.623,
+      "step": 57
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.779310344827587e-05,
+      "loss": 2.4327,
+      "step": 58
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.774712643678162e-05,
+      "loss": 2.0198,
+      "step": 59
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.770114942528736e-05,
+      "loss": 2.2694,
+      "step": 60
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 9.765517241379311e-05,
+      "loss": 2.0924,
+      "step": 61
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.760919540229885e-05,
+      "loss": 1.4724,
+      "step": 62
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 9.75632183908046e-05,
+      "loss": 2.0698,
+      "step": 63
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.751724137931035e-05,
+      "loss": 1.2204,
+      "step": 64
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 9.747126436781609e-05,
+      "loss": 1.8652,
+      "step": 65
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.742528735632185e-05,
+      "loss": 1.5699,
+      "step": 66
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.737931034482759e-05,
+      "loss": 1.6397,
+      "step": 67
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.733333333333335e-05,
+      "loss": 1.694,
+      "step": 68
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.728735632183908e-05,
+      "loss": 3.0533,
+      "step": 69
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.724137931034482e-05,
+      "loss": 2.3781,
+      "step": 70
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 9.719540229885058e-05,
+      "loss": 1.3803,
+      "step": 71
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.714942528735632e-05,
+      "loss": 1.744,
+      "step": 72
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 9.710344827586208e-05,
+      "loss": 1.8157,
+      "step": 73
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.705747126436781e-05,
+      "loss": 2.0532,
+      "step": 74
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.701149425287357e-05,
+      "loss": 1.7599,
+      "step": 75
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 9.696551724137932e-05,
+      "loss": 2.1115,
+      "step": 76
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 9.691954022988506e-05,
+      "loss": 1.7123,
+      "step": 77
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 9.687356321839081e-05,
+      "loss": 1.4739,
+      "step": 78
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 9.682758620689656e-05,
+      "loss": 1.5236,
+      "step": 79
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 9.67816091954023e-05,
+      "loss": 1.7303,
+      "step": 80
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 9.673563218390805e-05,
+      "loss": 1.5281,
+      "step": 81
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.66896551724138e-05,
+      "loss": 1.7074,
+      "step": 82
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.664367816091955e-05,
+      "loss": 1.8005,
+      "step": 83
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 9.659770114942529e-05,
+      "loss": 1.7979,
+      "step": 84
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 9.655172413793105e-05,
+      "loss": 1.823,
+      "step": 85
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 9.650574712643678e-05,
+      "loss": 2.5651,
+      "step": 86
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 9.645977011494254e-05,
+      "loss": 1.2539,
+      "step": 87
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 9.641379310344829e-05,
+      "loss": 2.0415,
+      "step": 88
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 9.636781609195403e-05,
+      "loss": 2.8371,
+      "step": 89
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 9.632183908045978e-05,
+      "loss": 1.2387,
+      "step": 90
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 9.627586206896551e-05,
+      "loss": 1.6533,
+      "step": 91
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 9.622988505747127e-05,
+      "loss": 2.0956,
+      "step": 92
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 9.618390804597702e-05,
+      "loss": 1.2122,
+      "step": 93
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.613793103448276e-05,
+      "loss": 1.6876,
+      "step": 94
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.609195402298851e-05,
+      "loss": 2.3298,
+      "step": 95
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 9.604597701149426e-05,
+      "loss": 1.8006,
+      "step": 96
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.6e-05,
+      "loss": 1.7219,
+      "step": 97
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 9.595402298850575e-05,
+      "loss": 1.3732,
+      "step": 98
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 9.59080459770115e-05,
+      "loss": 1.7793,
+      "step": 99
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 9.586206896551725e-05,
+      "loss": 2.0547,
+      "step": 100
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.581609195402299e-05,
+      "loss": 1.3536,
+      "step": 101
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 9.577011494252875e-05,
+      "loss": 2.5,
+      "step": 102
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 9.572413793103448e-05,
+      "loss": 1.2496,
+      "step": 103
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 9.567816091954024e-05,
+      "loss": 2.2696,
+      "step": 104
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.563218390804599e-05,
+      "loss": 1.1918,
+      "step": 105
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.558620689655173e-05,
+      "loss": 1.5118,
+      "step": 106
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.554022988505748e-05,
+      "loss": 1.1344,
+      "step": 107
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.549425287356323e-05,
+      "loss": 1.6437,
+      "step": 108
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 9.544827586206897e-05,
+      "loss": 1.7307,
+      "step": 109
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.4024390243902439,
+      "eval_loss": 1.5326756238937378,
+      "eval_runtime": 13.8933,
+      "eval_samples_per_second": 11.804,
+      "eval_steps_per_second": 1.512,
+      "step": 109
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 9.540229885057472e-05,
+      "loss": 1.5236,
+      "step": 110
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 9.535632183908046e-05,
+      "loss": 0.7238,
+      "step": 111
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.531034482758621e-05,
+      "loss": 1.0976,
+      "step": 112
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.526436781609196e-05,
+      "loss": 0.869,
+      "step": 113
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.52183908045977e-05,
+      "loss": 0.7228,
+      "step": 114
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.517241379310345e-05,
+      "loss": 1.1438,
+      "step": 115
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.51264367816092e-05,
+      "loss": 0.5575,
+      "step": 116
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.508045977011494e-05,
+      "loss": 1.4123,
+      "step": 117
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.503448275862069e-05,
+      "loss": 1.8973,
+      "step": 118
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.498850574712645e-05,
+      "loss": 0.3972,
+      "step": 119
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 9.494252873563218e-05,
+      "loss": 0.7104,
+      "step": 120
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 9.489655172413794e-05,
+      "loss": 0.6647,
+      "step": 121
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 9.485057471264367e-05,
+      "loss": 0.8883,
+      "step": 122
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 9.480459770114943e-05,
+      "loss": 1.2907,
+      "step": 123
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 9.475862068965518e-05,
+      "loss": 1.0642,
+      "step": 124
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 9.471264367816093e-05,
+      "loss": 1.4426,
+      "step": 125
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 9.466666666666667e-05,
+      "loss": 0.8142,
+      "step": 126
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 9.462068965517242e-05,
+      "loss": 0.2699,
+      "step": 127
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 9.457471264367817e-05,
+      "loss": 0.3382,
+      "step": 128
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 9.452873563218391e-05,
+      "loss": 1.3154,
+      "step": 129
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 9.448275862068966e-05,
+      "loss": 0.7058,
+      "step": 130
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 9.44367816091954e-05,
+      "loss": 0.7397,
+      "step": 131
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 9.439080459770115e-05,
+      "loss": 0.9056,
+      "step": 132
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 9.43448275862069e-05,
+      "loss": 1.3027,
+      "step": 133
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 9.429885057471264e-05,
+      "loss": 0.9512,
+      "step": 134
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 9.425287356321839e-05,
+      "loss": 0.7303,
+      "step": 135
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 9.420689655172415e-05,
+      "loss": 1.9017,
+      "step": 136
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 9.416091954022988e-05,
+      "loss": 0.2521,
+      "step": 137
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 9.411494252873564e-05,
+      "loss": 0.9502,
+      "step": 138
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.406896551724137e-05,
+      "loss": 0.7114,
+      "step": 139
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.402298850574713e-05,
+      "loss": 0.4075,
+      "step": 140
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 9.397701149425288e-05,
+      "loss": 1.609,
+      "step": 141
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 9.393103448275863e-05,
+      "loss": 1.7638,
+      "step": 142
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 9.388505747126437e-05,
+      "loss": 0.6199,
+      "step": 143
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 9.383908045977012e-05,
+      "loss": 0.7654,
+      "step": 144
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 9.379310344827587e-05,
+      "loss": 0.3153,
+      "step": 145
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 9.374712643678161e-05,
+      "loss": 0.8762,
+      "step": 146
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 9.370114942528736e-05,
+      "loss": 1.0565,
+      "step": 147
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 9.365517241379312e-05,
+      "loss": 1.3616,
+      "step": 148
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 9.360919540229885e-05,
+      "loss": 0.9263,
+      "step": 149
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 9.35632183908046e-05,
+      "loss": 1.3355,
+      "step": 150
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 9.351724137931034e-05,
+      "loss": 1.0787,
+      "step": 151
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 9.347126436781609e-05,
+      "loss": 1.3131,
+      "step": 152
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 9.342528735632185e-05,
+      "loss": 0.8362,
+      "step": 153
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 9.337931034482758e-05,
+      "loss": 1.7228,
+      "step": 154
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 9.333333333333334e-05,
+      "loss": 1.289,
+      "step": 155
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 9.328735632183908e-05,
+      "loss": 0.7971,
+      "step": 156
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 9.324137931034484e-05,
+      "loss": 1.1699,
+      "step": 157
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 9.319540229885058e-05,
+      "loss": 0.5347,
+      "step": 158
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 9.314942528735633e-05,
+      "loss": 0.8059,
+      "step": 159
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 9.310344827586207e-05,
+      "loss": 0.6804,
+      "step": 160
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 9.305747126436782e-05,
+      "loss": 1.3072,
+      "step": 161
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 9.301149425287357e-05,
+      "loss": 0.4505,
+      "step": 162
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.296551724137931e-05,
+      "loss": 0.5823,
+      "step": 163
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.291954022988506e-05,
+      "loss": 0.6861,
+      "step": 164
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 9.287356321839082e-05,
+      "loss": 1.2158,
+      "step": 165
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 9.282758620689655e-05,
+      "loss": 0.2745,
+      "step": 166
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.278160919540231e-05,
+      "loss": 0.1638,
+      "step": 167
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.273563218390805e-05,
+      "loss": 0.2435,
+      "step": 168
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 9.26896551724138e-05,
+      "loss": 0.9956,
+      "step": 169
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.264367816091954e-05,
+      "loss": 0.5066,
+      "step": 170
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 9.259770114942528e-05,
+      "loss": 0.4705,
+      "step": 171
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.255172413793104e-05,
+      "loss": 0.8163,
+      "step": 172
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.250574712643678e-05,
+      "loss": 0.4386,
+      "step": 173
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 9.245977011494254e-05,
+      "loss": 0.5746,
+      "step": 174
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.241379310344827e-05,
+      "loss": 0.171,
+      "step": 175
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.236781609195403e-05,
+      "loss": 1.1641,
+      "step": 176
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.232183908045978e-05,
+      "loss": 1.142,
+      "step": 177
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.227586206896552e-05,
+      "loss": 1.1509,
+      "step": 178
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 9.222988505747127e-05,
+      "loss": 0.0439,
+      "step": 179
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 9.218390804597701e-05,
+      "loss": 0.4179,
+      "step": 180
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 9.213793103448276e-05,
+      "loss": 0.4088,
+      "step": 181
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 9.209195402298851e-05,
+      "loss": 0.0732,
+      "step": 182
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 9.204597701149425e-05,
+      "loss": 0.0423,
+      "step": 183
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 9.200000000000001e-05,
+      "loss": 0.2141,
+      "step": 184
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 9.195402298850575e-05,
+      "loss": 0.1807,
+      "step": 185
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 9.19080459770115e-05,
+      "loss": 1.7772,
+      "step": 186
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.186206896551724e-05,
+      "loss": 0.9672,
+      "step": 187
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.1816091954023e-05,
+      "loss": 3.4189,
+      "step": 188
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.177011494252875e-05,
+      "loss": 2.2924,
+      "step": 189
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 9.172413793103448e-05,
+      "loss": 0.9232,
+      "step": 190
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 9.167816091954024e-05,
+      "loss": 2.6377,
+      "step": 191
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 9.163218390804597e-05,
+      "loss": 0.1525,
+      "step": 192
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 9.158620689655173e-05,
+      "loss": 0.6139,
+      "step": 193
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 9.154022988505748e-05,
+      "loss": 1.1956,
+      "step": 194
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 9.149425287356322e-05,
+      "loss": 0.2272,
+      "step": 195
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 9.144827586206897e-05,
+      "loss": 0.2746,
+      "step": 196
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 9.140229885057472e-05,
+      "loss": 1.1535,
+      "step": 197
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 9.135632183908046e-05,
+      "loss": 0.2671,
+      "step": 198
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 9.131034482758621e-05,
+      "loss": 0.7807,
+      "step": 199
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 9.126436781609195e-05,
+      "loss": 0.0523,
+      "step": 200
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 9.121839080459771e-05,
+      "loss": 0.3851,
+      "step": 201
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 9.117241379310345e-05,
+      "loss": 0.056,
+      "step": 202
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 9.112643678160921e-05,
+      "loss": 2.6584,
+      "step": 203
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 9.108045977011494e-05,
+      "loss": 0.1002,
+      "step": 204
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 9.10344827586207e-05,
+      "loss": 0.5692,
+      "step": 205
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 9.098850574712645e-05,
+      "loss": 1.2853,
+      "step": 206
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 9.094252873563219e-05,
+      "loss": 2.3472,
+      "step": 207
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.089655172413794e-05,
+      "loss": 0.9564,
+      "step": 208
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 9.085057471264369e-05,
+      "loss": 0.7848,
+      "step": 209
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 9.080459770114943e-05,
+      "loss": 1.0236,
+      "step": 210
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 9.075862068965518e-05,
+      "loss": 1.1465,
+      "step": 211
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 9.071264367816092e-05,
+      "loss": 0.229,
+      "step": 212
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 9.066666666666667e-05,
+      "loss": 0.757,
+      "step": 213
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 9.062068965517242e-05,
+      "loss": 0.1344,
+      "step": 214
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 9.057471264367816e-05,
+      "loss": 0.4208,
+      "step": 215
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 9.052873563218391e-05,
+      "loss": 1.0033,
+      "step": 216
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 9.048275862068966e-05,
+      "loss": 0.1265,
+      "step": 217
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 9.043678160919542e-05,
+      "loss": 1.4519,
+      "step": 218
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.7073170731707317,
+      "eval_loss": 1.234080195426941,
+      "eval_runtime": 13.8793,
+      "eval_samples_per_second": 11.816,
+      "eval_steps_per_second": 1.513,
+      "step": 218
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 9.039080459770115e-05,
+      "loss": 0.0027,
+      "step": 219
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 9.034482758620691e-05,
+      "loss": 0.1063,
+      "step": 220
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 9.029885057471264e-05,
+      "loss": 0.8028,
+      "step": 221
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 9.02528735632184e-05,
+      "loss": 0.3921,
+      "step": 222
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 9.020689655172413e-05,
+      "loss": 0.0123,
+      "step": 223
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 9.01609195402299e-05,
+      "loss": 0.0506,
+      "step": 224
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 9.011494252873564e-05,
+      "loss": 0.0079,
+      "step": 225
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 9.006896551724139e-05,
+      "loss": 0.0484,
+      "step": 226
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 9.002298850574713e-05,
+      "loss": 0.7222,
+      "step": 227
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 8.997701149425288e-05,
+      "loss": 0.0005,
+      "step": 228
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 8.993103448275863e-05,
+      "loss": 0.127,
+      "step": 229
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 8.988505747126437e-05,
+      "loss": 0.097,
+      "step": 230
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 8.983908045977012e-05,
+      "loss": 0.0061,
+      "step": 231
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 8.979310344827586e-05,
+      "loss": 0.0039,
+      "step": 232
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 8.974712643678161e-05,
+      "loss": 0.0005,
+      "step": 233
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 8.970114942528736e-05,
+      "loss": 0.1735,
+      "step": 234
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 8.96551724137931e-05,
+      "loss": 0.2443,
+      "step": 235
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 8.960919540229885e-05,
+      "loss": 0.0007,
+      "step": 236
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 8.956321839080461e-05,
+      "loss": 0.0206,
+      "step": 237
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 8.951724137931034e-05,
+      "loss": 0.0003,
+      "step": 238
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.94712643678161e-05,
+      "loss": 0.0261,
+      "step": 239
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.942528735632183e-05,
+      "loss": 0.0019,
+      "step": 240
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.93793103448276e-05,
+      "loss": 0.0026,
+      "step": 241
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.933333333333334e-05,
+      "loss": 2.2971,
+      "step": 242
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.928735632183909e-05,
+      "loss": 0.0004,
+      "step": 243
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 8.924137931034483e-05,
+      "loss": 0.1492,
+      "step": 244
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 8.919540229885058e-05,
+      "loss": 0.0116,
+      "step": 245
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 8.914942528735633e-05,
+      "loss": 0.0658,
+      "step": 246
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 8.910344827586207e-05,
+      "loss": 0.0107,
+      "step": 247
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 8.905747126436782e-05,
+      "loss": 0.3834,
+      "step": 248
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 8.901149425287358e-05,
+      "loss": 0.0006,
+      "step": 249
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 8.896551724137931e-05,
+      "loss": 0.3983,
+      "step": 250
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 8.891954022988506e-05,
+      "loss": 0.0829,
+      "step": 251
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 8.88735632183908e-05,
+      "loss": 0.2372,
+      "step": 252
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 8.882758620689655e-05,
+      "loss": 0.4324,
+      "step": 253
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 8.878160919540231e-05,
+      "loss": 0.0332,
+      "step": 254
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 8.873563218390804e-05,
+      "loss": 0.0005,
+      "step": 255
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 8.86896551724138e-05,
+      "loss": 0.0023,
+      "step": 256
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 8.864367816091954e-05,
+      "loss": 0.3183,
+      "step": 257
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 8.85977011494253e-05,
+      "loss": 0.5345,
+      "step": 258
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 8.855172413793104e-05,
+      "loss": 0.0129,
+      "step": 259
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 8.850574712643679e-05,
+      "loss": 0.0197,
+      "step": 260
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 8.845977011494253e-05,
+      "loss": 0.0,
+      "step": 261
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 8.841379310344828e-05,
+      "loss": 0.0458,
+      "step": 262
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 8.836781609195403e-05,
+      "loss": 0.0956,
+      "step": 263
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 8.832183908045977e-05,
+      "loss": 0.278,
+      "step": 264
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 8.827586206896552e-05,
+      "loss": 0.1117,
+      "step": 265
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 8.822988505747128e-05,
+      "loss": 0.0,
+      "step": 266
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 8.818390804597701e-05,
+      "loss": 0.1784,
+      "step": 267
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 8.813793103448277e-05,
+      "loss": 0.0005,
+      "step": 268
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 8.80919540229885e-05,
+      "loss": 0.0001,
+      "step": 269
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 8.804597701149425e-05,
+      "loss": 0.0001,
+      "step": 270
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 8.800000000000001e-05,
+      "loss": 0.1491,
+      "step": 271
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 8.795402298850574e-05,
+      "loss": 0.269,
+      "step": 272
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 8.79080459770115e-05,
+      "loss": 0.0,
+      "step": 273
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 8.786206896551724e-05,
+      "loss": 1.8217,
+      "step": 274
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 8.7816091954023e-05,
+      "loss": 1.2318,
+      "step": 275
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 8.777011494252874e-05,
+      "loss": 1.2411,
+      "step": 276
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 8.772413793103449e-05,
+      "loss": 0.0006,
+      "step": 277
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 8.767816091954024e-05,
+      "loss": 0.8842,
+      "step": 278
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 8.763218390804598e-05,
+      "loss": 0.1592,
+      "step": 279
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 8.758620689655173e-05,
+      "loss": 0.467,
+      "step": 280
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 8.754022988505747e-05,
+      "loss": 2.0205,
+      "step": 281
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 8.749425287356322e-05,
+      "loss": 0.0208,
+      "step": 282
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 8.744827586206897e-05,
+      "loss": 0.0001,
+      "step": 283
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 8.740229885057471e-05,
+      "loss": 0.0,
+      "step": 284
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 8.735632183908047e-05,
+      "loss": 0.6909,
+      "step": 285
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 8.73103448275862e-05,
+      "loss": 2.6178,
+      "step": 286
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 8.726436781609197e-05,
+      "loss": 0.0022,
+      "step": 287
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 8.72183908045977e-05,
+      "loss": 0.4241,
+      "step": 288
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 8.717241379310346e-05,
+      "loss": 0.9552,
+      "step": 289
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 8.71264367816092e-05,
+      "loss": 0.0001,
+      "step": 290
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 8.708045977011494e-05,
+      "loss": 0.0,
+      "step": 291
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 8.70344827586207e-05,
+      "loss": 0.0001,
+      "step": 292
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 8.698850574712643e-05,
+      "loss": 0.0004,
+      "step": 293
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 8.694252873563219e-05,
+      "loss": 0.0008,
+      "step": 294
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 8.689655172413794e-05,
+      "loss": 0.0567,
+      "step": 295
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 8.685057471264368e-05,
+      "loss": 0.0014,
+      "step": 296
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 8.680459770114943e-05,
+      "loss": 0.0343,
+      "step": 297
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 8.675862068965518e-05,
+      "loss": 0.0002,
+      "step": 298
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 8.671264367816092e-05,
+      "loss": 0.0005,
+      "step": 299
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.666666666666667e-05,
+      "loss": 0.9325,
+      "step": 300
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.662068965517241e-05,
+      "loss": 0.5408,
+      "step": 301
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 8.657471264367817e-05,
+      "loss": 0.0214,
+      "step": 302
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 8.652873563218391e-05,
+      "loss": 0.0018,
+      "step": 303
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 8.648275862068967e-05,
+      "loss": 0.0003,
+      "step": 304
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 8.64367816091954e-05,
+      "loss": 0.0991,
+      "step": 305
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 8.639080459770116e-05,
+      "loss": 0.0286,
+      "step": 306
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 8.63448275862069e-05,
+      "loss": 0.001,
+      "step": 307
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.629885057471265e-05,
+      "loss": 0.1312,
+      "step": 308
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.62528735632184e-05,
+      "loss": 0.009,
+      "step": 309
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.620689655172413e-05,
+      "loss": 0.0003,
+      "step": 310
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.616091954022989e-05,
+      "loss": 0.0511,
+      "step": 311
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.611494252873564e-05,
+      "loss": 1.1161,
+      "step": 312
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.606896551724138e-05,
+      "loss": 0.0288,
+      "step": 313
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.602298850574713e-05,
+      "loss": 0.0,
+      "step": 314
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.597701149425288e-05,
+      "loss": 0.003,
+      "step": 315
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 8.597701149425288e-05,
+      "loss": 1.0016,
+      "step": 316
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 8.593103448275862e-05,
+      "loss": 0.522,
+      "step": 317
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.588505747126437e-05,
+      "loss": 0.3683,
+      "step": 318
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 8.583908045977012e-05,
+      "loss": 0.1132,
+      "step": 319
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 8.579310344827588e-05,
+      "loss": 0.2868,
+      "step": 320
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 8.574712643678161e-05,
+      "loss": 0.0071,
+      "step": 321
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 8.574712643678161e-05,
+      "loss": 0.7119,
+      "step": 322
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 8.570114942528737e-05,
+      "loss": 0.965,
+      "step": 323
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 8.570114942528737e-05,
+      "loss": 0.6825,
+      "step": 324
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 8.56551724137931e-05,
+      "loss": 0.1487,
+      "step": 325
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 8.560919540229886e-05,
+      "loss": 0.1148,
+      "step": 326
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 8.556321839080461e-05,
+      "loss": 0.0207,
+      "step": 327
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.6280487804878049,
+      "eval_loss": 2.292419910430908,
+      "eval_runtime": 13.8764,
+      "eval_samples_per_second": 11.819,
+      "eval_steps_per_second": 1.513,
+      "step": 327
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 8.551724137931035e-05,
+      "loss": 0.0438,
+      "step": 328
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 8.54712643678161e-05,
+      "loss": 0.0008,
+      "step": 329
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 8.542528735632185e-05,
+      "loss": 0.0096,
+      "step": 330
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 8.537931034482759e-05,
+      "loss": 0.0326,
+      "step": 331
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 8.533333333333334e-05,
+      "loss": 0.0002,
+      "step": 332
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 8.533333333333334e-05,
+      "loss": 0.8887,
+      "step": 333
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 8.528735632183908e-05,
+      "loss": 0.0019,
+      "step": 334
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 8.524137931034483e-05,
+      "loss": 0.2207,
+      "step": 335
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 8.519540229885058e-05,
+      "loss": 0.1587,
+      "step": 336
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 8.514942528735632e-05,
+      "loss": 0.4261,
+      "step": 337
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 8.514942528735632e-05,
+      "loss": 0.7838,
+      "step": 338
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 8.510344827586207e-05,
+      "loss": 0.111,
+      "step": 339
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 8.505747126436782e-05,
+      "loss": 0.122,
+      "step": 340
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 8.501149425287356e-05,
+      "loss": 0.6099,
+      "step": 341
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 8.496551724137931e-05,
+      "loss": 0.5102,
+      "step": 342
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 8.491954022988507e-05,
+      "loss": 1.087,
+      "step": 343
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 8.48735632183908e-05,
+      "loss": 0.9862,
+      "step": 344
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 8.482758620689656e-05,
+      "loss": 0.5406,
+      "step": 345
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 8.47816091954023e-05,
+      "loss": 1.8497,
+      "step": 346
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 8.473563218390805e-05,
+      "loss": 2.7576,
+      "step": 347
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 8.46896551724138e-05,
+      "loss": 4.038,
+      "step": 348
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 8.464367816091955e-05,
+      "loss": 2.2978,
+      "step": 349
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 8.459770114942529e-05,
+      "loss": 2.4796,
+      "step": 350
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 8.455172413793104e-05,
+      "loss": 3.8485,
+      "step": 351
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 8.450574712643679e-05,
+      "loss": 2.7897,
+      "step": 352
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 8.445977011494253e-05,
+      "loss": 3.4341,
+      "step": 353
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 8.441379310344828e-05,
+      "loss": 3.4921,
+      "step": 354
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 8.436781609195402e-05,
+      "loss": 4.8688,
+      "step": 355
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 8.432183908045977e-05,
+      "loss": 4.1337,
+      "step": 356
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 8.427586206896552e-05,
+      "loss": 5.276,
+      "step": 357
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 8.422988505747126e-05,
+      "loss": 3.3783,
+      "step": 358
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 8.418390804597701e-05,
+      "loss": 5.7553,
+      "step": 359
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 8.413793103448277e-05,
+      "loss": 3.6846,
+      "step": 360
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 8.40919540229885e-05,
+      "loss": 5.5257,
+      "step": 361
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 8.404597701149426e-05,
+      "loss": 3.842,
+      "step": 362
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 8.4e-05,
+      "loss": 4.8504,
+      "step": 363
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 8.395402298850576e-05,
+      "loss": 3.0007,
+      "step": 364
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 8.39080459770115e-05,
+      "loss": 5.071,
+      "step": 365
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 8.386206896551725e-05,
+      "loss": 3.5151,
+      "step": 366
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 8.3816091954023e-05,
+      "loss": 3.2012,
+      "step": 367
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 8.377011494252874e-05,
+      "loss": 2.9036,
+      "step": 368
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 8.372413793103449e-05,
+      "loss": 2.6111,
+      "step": 369
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 8.367816091954023e-05,
+      "loss": 2.9521,
+      "step": 370
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 8.363218390804598e-05,
+      "loss": 2.6554,
+      "step": 371
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 8.358620689655174e-05,
+      "loss": 2.1818,
+      "step": 372
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 8.354022988505747e-05,
+      "loss": 2.0701,
+      "step": 373
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 8.349425287356323e-05,
+      "loss": 1.9787,
+      "step": 374
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 8.344827586206896e-05,
+      "loss": 1.8641,
+      "step": 375
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 8.340229885057471e-05,
+      "loss": 2.2446,
+      "step": 376
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 8.335632183908047e-05,
+      "loss": 2.1309,
+      "step": 377
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 8.33103448275862e-05,
+      "loss": 1.6129,
+      "step": 378
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 8.326436781609196e-05,
+      "loss": 2.1021,
+      "step": 379
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 8.32183908045977e-05,
+      "loss": 1.6412,
+      "step": 380
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 8.317241379310346e-05,
+      "loss": 2.2044,
+      "step": 381
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 8.31264367816092e-05,
+      "loss": 2.4867,
+      "step": 382
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 8.308045977011495e-05,
+      "loss": 1.5561,
+      "step": 383
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 8.30344827586207e-05,
+      "loss": 2.0623,
+      "step": 384
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 8.298850574712644e-05,
+      "loss": 1.7909,
+      "step": 385
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 8.294252873563219e-05,
+      "loss": 2.4909,
+      "step": 386
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 8.289655172413793e-05,
+      "loss": 1.6948,
+      "step": 387
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 8.285057471264368e-05,
+      "loss": 1.7915,
+      "step": 388
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 8.280459770114943e-05,
+      "loss": 1.672,
+      "step": 389
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 8.275862068965517e-05,
+      "loss": 1.5677,
+      "step": 390
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 8.271264367816093e-05,
+      "loss": 2.3682,
+      "step": 391
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 8.266666666666667e-05,
+      "loss": 2.375,
+      "step": 392
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 8.262068965517243e-05,
+      "loss": 2.2048,
+      "step": 393
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 8.257471264367816e-05,
+      "loss": 1.5205,
+      "step": 394
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 8.25287356321839e-05,
+      "loss": 2.5317,
+      "step": 395
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 8.248275862068966e-05,
+      "loss": 1.9683,
+      "step": 396
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 8.24367816091954e-05,
+      "loss": 2.1746,
+      "step": 397
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 8.239080459770116e-05,
+      "loss": 1.7266,
+      "step": 398
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 8.234482758620689e-05,
+      "loss": 1.4727,
+      "step": 399
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 8.229885057471265e-05,
+      "loss": 2.223,
+      "step": 400
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 8.22528735632184e-05,
+      "loss": 2.1956,
+      "step": 401
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 8.220689655172414e-05,
+      "loss": 1.4727,
+      "step": 402
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 8.216091954022989e-05,
+      "loss": 2.0462,
+      "step": 403
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 8.211494252873564e-05,
+      "loss": 1.9425,
+      "step": 404
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 8.206896551724138e-05,
+      "loss": 2.0615,
+      "step": 405
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 8.202298850574713e-05,
+      "loss": 1.9185,
+      "step": 406
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 8.197701149425287e-05,
+      "loss": 1.543,
+      "step": 407
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 8.193103448275863e-05,
+      "loss": 1.9578,
+      "step": 408
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 8.188505747126437e-05,
+      "loss": 1.9735,
+      "step": 409
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 8.183908045977013e-05,
+      "loss": 1.7452,
+      "step": 410
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 8.179310344827586e-05,
+      "loss": 1.7625,
+      "step": 411
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 8.174712643678162e-05,
+      "loss": 1.7108,
+      "step": 412
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 8.170114942528737e-05,
+      "loss": 2.1712,
+      "step": 413
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 8.165517241379311e-05,
+      "loss": 1.8252,
+      "step": 414
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 8.160919540229886e-05,
+      "loss": 2.013,
+      "step": 415
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 8.156321839080459e-05,
+      "loss": 2.0345,
+      "step": 416
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 8.151724137931035e-05,
+      "loss": 2.0923,
+      "step": 417
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 8.14712643678161e-05,
+      "loss": 1.729,
+      "step": 418
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 8.142528735632184e-05,
+      "loss": 2.1458,
+      "step": 419
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 8.137931034482759e-05,
+      "loss": 2.2624,
+      "step": 420
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 8.133333333333334e-05,
+      "loss": 1.9985,
+      "step": 421
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 8.128735632183908e-05,
+      "loss": 1.9106,
+      "step": 422
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 8.124137931034483e-05,
+      "loss": 2.5249,
+      "step": 423
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 8.119540229885058e-05,
+      "loss": 2.2891,
+      "step": 424
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 8.114942528735633e-05,
+      "loss": 1.668,
+      "step": 425
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 8.110344827586207e-05,
+      "loss": 1.7,
+      "step": 426
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 8.105747126436783e-05,
+      "loss": 2.1659,
+      "step": 427
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 8.101149425287356e-05,
+      "loss": 1.8076,
+      "step": 428
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 8.096551724137932e-05,
+      "loss": 1.8,
+      "step": 429
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 8.091954022988507e-05,
+      "loss": 2.1989,
+      "step": 430
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 8.087356321839081e-05,
+      "loss": 1.7143,
+      "step": 431
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 8.082758620689656e-05,
+      "loss": 1.9302,
+      "step": 432
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 8.07816091954023e-05,
+      "loss": 1.7413,
+      "step": 433
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 8.073563218390805e-05,
+      "loss": 1.7524,
+      "step": 434
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 8.06896551724138e-05,
+      "loss": 1.9447,
+      "step": 435
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 8.064367816091954e-05,
+      "loss": 1.5647,
+      "step": 436
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.22560975609756098,
+      "eval_loss": 1.9343886375427246,
+      "eval_runtime": 13.7592,
+      "eval_samples_per_second": 11.919,
+      "eval_steps_per_second": 1.526,
+      "step": 436
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 8.059770114942529e-05,
+      "loss": 1.6333,
+      "step": 437
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 8.055172413793104e-05,
+      "loss": 1.6856,
+      "step": 438
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 8.050574712643678e-05,
+      "loss": 2.3695,
+      "step": 439
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 8.045977011494253e-05,
+      "loss": 1.7945,
+      "step": 440
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 8.041379310344828e-05,
+      "loss": 2.1624,
+      "step": 441
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 8.036781609195402e-05,
+      "loss": 1.7616,
+      "step": 442
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 8.032183908045977e-05,
+      "loss": 1.557,
+      "step": 443
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 8.027586206896553e-05,
+      "loss": 1.689,
+      "step": 444
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 8.022988505747126e-05,
+      "loss": 1.4793,
+      "step": 445
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 8.018390804597702e-05,
+      "loss": 1.5692,
+      "step": 446
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 8.013793103448275e-05,
+      "loss": 2.1131,
+      "step": 447
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 8.009195402298851e-05,
+      "loss": 1.9056,
+      "step": 448
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 8.004597701149426e-05,
+      "loss": 1.6618,
+      "step": 449
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 8e-05,
+      "loss": 1.9688,
+      "step": 450
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 7.995402298850575e-05,
+      "loss": 2.0741,
+      "step": 451
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 7.99080459770115e-05,
+      "loss": 1.9928,
+      "step": 452
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 7.986206896551725e-05,
+      "loss": 1.807,
+      "step": 453
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 7.981609195402299e-05,
+      "loss": 1.9935,
+      "step": 454
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 7.977011494252874e-05,
+      "loss": 2.2464,
+      "step": 455
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 7.972413793103448e-05,
+      "loss": 1.6822,
+      "step": 456
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 7.967816091954023e-05,
+      "loss": 1.9629,
+      "step": 457
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 7.963218390804598e-05,
+      "loss": 1.7938,
+      "step": 458
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 7.958620689655172e-05,
+      "loss": 1.98,
+      "step": 459
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 7.954022988505747e-05,
+      "loss": 1.5706,
+      "step": 460
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 7.949425287356323e-05,
+      "loss": 1.9338,
+      "step": 461
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 7.944827586206896e-05,
+      "loss": 1.7342,
+      "step": 462
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 7.940229885057472e-05,
+      "loss": 2.0734,
+      "step": 463
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 7.935632183908046e-05,
+      "loss": 2.0415,
+      "step": 464
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 7.931034482758621e-05,
+      "loss": 1.6935,
+      "step": 465
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 7.926436781609196e-05,
+      "loss": 1.9387,
+      "step": 466
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 7.921839080459771e-05,
+      "loss": 2.0369,
+      "step": 467
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 7.917241379310345e-05,
+      "loss": 1.572,
+      "step": 468
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 7.91264367816092e-05,
+      "loss": 2.2131,
+      "step": 469
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 7.908045977011495e-05,
+      "loss": 2.2298,
+      "step": 470
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 7.903448275862069e-05,
+      "loss": 2.3888,
+      "step": 471
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 7.898850574712644e-05,
+      "loss": 1.9611,
+      "step": 472
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 7.89425287356322e-05,
+      "loss": 1.6646,
+      "step": 473
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 7.889655172413793e-05,
+      "loss": 1.8882,
+      "step": 474
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 7.885057471264368e-05,
+      "loss": 1.4906,
+      "step": 475
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 7.880459770114942e-05,
+      "loss": 2.666,
+      "step": 476
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 7.875862068965517e-05,
+      "loss": 1.8164,
+      "step": 477
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 7.871264367816093e-05,
+      "loss": 1.6232,
+      "step": 478
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 7.866666666666666e-05,
+      "loss": 1.4965,
+      "step": 479
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 7.862068965517242e-05,
+      "loss": 1.7057,
+      "step": 480
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 7.857471264367816e-05,
+      "loss": 2.3079,
+      "step": 481
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 7.852873563218392e-05,
+      "loss": 1.4268,
+      "step": 482
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 7.848275862068966e-05,
+      "loss": 1.832,
+      "step": 483
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 7.843678160919541e-05,
+      "loss": 1.9131,
+      "step": 484
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 7.839080459770115e-05,
+      "loss": 2.2351,
+      "step": 485
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 7.83448275862069e-05,
+      "loss": 1.8963,
+      "step": 486
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 7.829885057471265e-05,
+      "loss": 1.5729,
+      "step": 487
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 7.82528735632184e-05,
+      "loss": 1.666,
+      "step": 488
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 7.820689655172414e-05,
+      "loss": 1.5511,
+      "step": 489
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 7.81609195402299e-05,
+      "loss": 1.5701,
+      "step": 490
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 7.811494252873563e-05,
+      "loss": 1.5034,
+      "step": 491
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 7.806896551724139e-05,
+      "loss": 2.2518,
+      "step": 492
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 7.802298850574713e-05,
+      "loss": 1.6068,
+      "step": 493
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 7.797701149425287e-05,
+      "loss": 1.7186,
+      "step": 494
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 7.793103448275862e-05,
+      "loss": 1.8242,
+      "step": 495
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 7.788505747126436e-05,
+      "loss": 1.7521,
+      "step": 496
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 7.783908045977012e-05,
+      "loss": 2.0605,
+      "step": 497
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 7.779310344827586e-05,
+      "loss": 1.9621,
+      "step": 498
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 7.774712643678162e-05,
+      "loss": 1.597,
+      "step": 499
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 7.770114942528735e-05,
+      "loss": 2.2424,
+      "step": 500
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 7.765517241379311e-05,
+      "loss": 2.0083,
+      "step": 501
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 7.760919540229886e-05,
+      "loss": 2.2173,
+      "step": 502
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 7.75632183908046e-05,
+      "loss": 2.2157,
+      "step": 503
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 7.751724137931035e-05,
+      "loss": 1.9162,
+      "step": 504
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 7.74712643678161e-05,
+      "loss": 1.9079,
+      "step": 505
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 7.742528735632184e-05,
+      "loss": 1.7954,
+      "step": 506
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 7.737931034482759e-05,
+      "loss": 1.8683,
+      "step": 507
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 7.733333333333333e-05,
+      "loss": 2.2458,
+      "step": 508
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 7.72873563218391e-05,
+      "loss": 1.6823,
+      "step": 509
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 7.724137931034483e-05,
+      "loss": 2.757,
+      "step": 510
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 7.719540229885059e-05,
+      "loss": 1.6662,
+      "step": 511
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 7.714942528735632e-05,
+      "loss": 1.8411,
+      "step": 512
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 7.710344827586208e-05,
+      "loss": 1.596,
+      "step": 513
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 7.705747126436783e-05,
+      "loss": 1.6668,
+      "step": 514
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 7.701149425287356e-05,
+      "loss": 2.46,
+      "step": 515
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 7.696551724137932e-05,
+      "loss": 1.681,
+      "step": 516
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 7.691954022988505e-05,
+      "loss": 1.62,
+      "step": 517
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 7.687356321839081e-05,
+      "loss": 1.7124,
+      "step": 518
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 7.682758620689656e-05,
+      "loss": 1.6846,
+      "step": 519
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 7.67816091954023e-05,
+      "loss": 1.769,
+      "step": 520
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 7.673563218390805e-05,
+      "loss": 1.6621,
+      "step": 521
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 7.66896551724138e-05,
+      "loss": 2.542,
+      "step": 522
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 7.664367816091954e-05,
+      "loss": 1.7936,
+      "step": 523
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 7.659770114942529e-05,
+      "loss": 1.8122,
+      "step": 524
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 7.655172413793103e-05,
+      "loss": 1.8184,
+      "step": 525
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 7.65057471264368e-05,
+      "loss": 1.9609,
+      "step": 526
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 7.645977011494253e-05,
+      "loss": 1.5197,
+      "step": 527
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 7.641379310344829e-05,
+      "loss": 1.5723,
+      "step": 528
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 7.636781609195402e-05,
+      "loss": 1.965,
+      "step": 529
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 7.632183908045978e-05,
+      "loss": 1.9922,
+      "step": 530
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 7.627586206896553e-05,
+      "loss": 1.7246,
+      "step": 531
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 7.622988505747127e-05,
+      "loss": 2.3324,
+      "step": 532
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 7.618390804597702e-05,
+      "loss": 1.7375,
+      "step": 533
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 7.613793103448275e-05,
+      "loss": 2.3625,
+      "step": 534
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 7.609195402298851e-05,
+      "loss": 2.2788,
+      "step": 535
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 7.604597701149426e-05,
+      "loss": 1.6123,
+      "step": 536
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 7.6e-05,
+      "loss": 1.6016,
+      "step": 537
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 7.595402298850575e-05,
+      "loss": 1.7178,
+      "step": 538
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 7.59080459770115e-05,
+      "loss": 2.395,
+      "step": 539
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 7.586206896551724e-05,
+      "loss": 1.8619,
+      "step": 540
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 7.581609195402299e-05,
+      "loss": 2.2148,
+      "step": 541
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 7.577011494252874e-05,
+      "loss": 2.3892,
+      "step": 542
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 7.57241379310345e-05,
+      "loss": 2.0655,
+      "step": 543
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 7.567816091954023e-05,
+      "loss": 1.6265,
+      "step": 544
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 7.563218390804599e-05,
+      "loss": 2.2047,
+      "step": 545
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.22560975609756098,
+      "eval_loss": 1.9401201009750366,
+      "eval_runtime": 13.7495,
+      "eval_samples_per_second": 11.928,
+      "eval_steps_per_second": 1.527,
+      "step": 545
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 7.558620689655172e-05,
+      "loss": 1.7788,
+      "step": 546
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 7.554022988505748e-05,
+      "loss": 2.2297,
+      "step": 547
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 7.549425287356321e-05,
+      "loss": 1.7134,
+      "step": 548
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 7.544827586206897e-05,
+      "loss": 1.8003,
+      "step": 549
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 7.540229885057472e-05,
+      "loss": 1.8055,
+      "step": 550
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 7.535632183908047e-05,
+      "loss": 2.0462,
+      "step": 551
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 7.531034482758621e-05,
+      "loss": 1.4995,
+      "step": 552
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 7.526436781609196e-05,
+      "loss": 1.6777,
+      "step": 553
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 7.52183908045977e-05,
+      "loss": 1.6595,
+      "step": 554
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 7.517241379310345e-05,
+      "loss": 1.6724,
+      "step": 555
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 7.51264367816092e-05,
+      "loss": 1.5505,
+      "step": 556
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 7.508045977011494e-05,
+      "loss": 1.7629,
+      "step": 557
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 7.503448275862069e-05,
+      "loss": 1.7228,
+      "step": 558
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 7.498850574712644e-05,
+      "loss": 2.0435,
+      "step": 559
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 7.494252873563218e-05,
+      "loss": 1.9549,
+      "step": 560
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 7.489655172413793e-05,
+      "loss": 1.6808,
+      "step": 561
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 7.485057471264369e-05,
+      "loss": 1.9697,
+      "step": 562
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 7.480459770114942e-05,
+      "loss": 1.9958,
+      "step": 563
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 7.475862068965518e-05,
+      "loss": 1.8628,
+      "step": 564
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 7.471264367816091e-05,
+      "loss": 2.4049,
+      "step": 565
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 7.466666666666667e-05,
+      "loss": 2.0194,
+      "step": 566
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 7.462068965517242e-05,
+      "loss": 1.7708,
+      "step": 567
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 7.457471264367817e-05,
+      "loss": 1.5848,
+      "step": 568
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 7.452873563218391e-05,
+      "loss": 3.0573,
+      "step": 569
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 7.448275862068966e-05,
+      "loss": 1.5132,
+      "step": 570
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 7.44367816091954e-05,
+      "loss": 1.7882,
+      "step": 571
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 7.439080459770115e-05,
+      "loss": 2.3348,
+      "step": 572
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 7.43448275862069e-05,
+      "loss": 1.9481,
+      "step": 573
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 7.429885057471265e-05,
+      "loss": 1.8252,
+      "step": 574
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 7.425287356321839e-05,
+      "loss": 1.811,
+      "step": 575
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 7.420689655172414e-05,
+      "loss": 1.6592,
+      "step": 576
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 7.416091954022988e-05,
+      "loss": 1.756,
+      "step": 577
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 7.411494252873563e-05,
+      "loss": 1.3816,
+      "step": 578
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 7.406896551724139e-05,
+      "loss": 2.3195,
+      "step": 579
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 7.402298850574712e-05,
+      "loss": 1.8683,
+      "step": 580
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 7.397701149425288e-05,
+      "loss": 1.8132,
+      "step": 581
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 7.393103448275862e-05,
+      "loss": 1.7401,
+      "step": 582
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 7.388505747126438e-05,
+      "loss": 1.9826,
+      "step": 583
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 7.383908045977012e-05,
+      "loss": 1.8164,
+      "step": 584
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 7.379310344827587e-05,
+      "loss": 2.3875,
+      "step": 585
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 7.374712643678161e-05,
+      "loss": 2.437,
+      "step": 586
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 7.370114942528736e-05,
+      "loss": 1.6499,
+      "step": 587
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 7.365517241379311e-05,
+      "loss": 1.4867,
+      "step": 588
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 7.360919540229885e-05,
+      "loss": 2.0332,
+      "step": 589
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 7.35632183908046e-05,
+      "loss": 1.7773,
+      "step": 590
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 7.351724137931036e-05,
+      "loss": 2.0991,
+      "step": 591
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 7.347126436781609e-05,
+      "loss": 2.1195,
+      "step": 592
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 7.342528735632185e-05,
+      "loss": 1.6899,
+      "step": 593
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 7.337931034482759e-05,
+      "loss": 1.82,
+      "step": 594
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 7.333333333333333e-05,
+      "loss": 1.8576,
+      "step": 595
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 7.328735632183909e-05,
+      "loss": 1.9465,
+      "step": 596
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 7.324137931034482e-05,
+      "loss": 1.6104,
+      "step": 597
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 7.319540229885058e-05,
+      "loss": 1.5617,
+      "step": 598
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 7.314942528735632e-05,
+      "loss": 1.4771,
+      "step": 599
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 7.310344827586208e-05,
+      "loss": 1.7228,
+      "step": 600
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 7.305747126436781e-05,
+      "loss": 2.2367,
+      "step": 601
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 7.301149425287357e-05,
+      "loss": 1.7345,
+      "step": 602
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 7.296551724137932e-05,
+      "loss": 1.8449,
+      "step": 603
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 7.291954022988506e-05,
+      "loss": 1.7609,
+      "step": 604
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 7.287356321839081e-05,
+      "loss": 1.5413,
+      "step": 605
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 7.282758620689655e-05,
+      "loss": 2.1678,
+      "step": 606
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 7.27816091954023e-05,
+      "loss": 2.9336,
+      "step": 607
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 7.273563218390805e-05,
+      "loss": 1.6719,
+      "step": 608
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 7.26896551724138e-05,
+      "loss": 1.6634,
+      "step": 609
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 7.264367816091955e-05,
+      "loss": 2.616,
+      "step": 610
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 7.259770114942529e-05,
+      "loss": 1.8848,
+      "step": 611
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 7.255172413793105e-05,
+      "loss": 2.2708,
+      "step": 612
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 7.250574712643678e-05,
+      "loss": 1.5667,
+      "step": 613
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 7.245977011494253e-05,
+      "loss": 1.5768,
+      "step": 614
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 7.241379310344828e-05,
+      "loss": 1.4399,
+      "step": 615
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 7.236781609195402e-05,
+      "loss": 2.4998,
+      "step": 616
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 7.232183908045978e-05,
+      "loss": 1.7526,
+      "step": 617
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 7.227586206896551e-05,
+      "loss": 1.9308,
+      "step": 618
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 7.222988505747127e-05,
+      "loss": 1.7547,
+      "step": 619
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 7.218390804597702e-05,
+      "loss": 1.6444,
+      "step": 620
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 7.213793103448276e-05,
+      "loss": 2.5205,
+      "step": 621
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 7.209195402298851e-05,
+      "loss": 2.1214,
+      "step": 622
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 7.204597701149426e-05,
+      "loss": 1.6736,
+      "step": 623
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 7.2e-05,
+      "loss": 2.124,
+      "step": 624
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 7.195402298850575e-05,
+      "loss": 1.6465,
+      "step": 625
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 7.19080459770115e-05,
+      "loss": 1.6675,
+      "step": 626
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 7.186206896551725e-05,
+      "loss": 1.5444,
+      "step": 627
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 7.181609195402299e-05,
+      "loss": 1.7523,
+      "step": 628
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 7.177011494252875e-05,
+      "loss": 1.8066,
+      "step": 629
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 7.172413793103448e-05,
+      "loss": 2.4466,
+      "step": 630
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 7.167816091954024e-05,
+      "loss": 1.9425,
+      "step": 631
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 7.163218390804599e-05,
+      "loss": 1.8034,
+      "step": 632
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 7.158620689655173e-05,
+      "loss": 1.8462,
+      "step": 633
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 7.154022988505748e-05,
+      "loss": 1.8682,
+      "step": 634
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 7.149425287356321e-05,
+      "loss": 1.9159,
+      "step": 635
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 7.144827586206897e-05,
+      "loss": 1.6872,
+      "step": 636
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 7.140229885057472e-05,
+      "loss": 1.7738,
+      "step": 637
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 7.135632183908046e-05,
+      "loss": 1.9723,
+      "step": 638
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 7.131034482758621e-05,
+      "loss": 2.2087,
+      "step": 639
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 7.126436781609196e-05,
+      "loss": 1.8768,
+      "step": 640
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 7.12183908045977e-05,
+      "loss": 1.8112,
+      "step": 641
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 7.117241379310345e-05,
+      "loss": 2.3779,
+      "step": 642
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 7.11264367816092e-05,
+      "loss": 2.0054,
+      "step": 643
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 7.108045977011496e-05,
+      "loss": 1.9657,
+      "step": 644
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 7.103448275862069e-05,
+      "loss": 2.3031,
+      "step": 645
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 7.098850574712645e-05,
+      "loss": 1.9206,
+      "step": 646
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.094252873563218e-05,
+      "loss": 1.7756,
+      "step": 647
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.089655172413794e-05,
+      "loss": 1.8485,
+      "step": 648
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 7.085057471264369e-05,
+      "loss": 1.5212,
+      "step": 649
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 7.080459770114943e-05,
+      "loss": 1.7484,
+      "step": 650
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 7.075862068965518e-05,
+      "loss": 1.7096,
+      "step": 651
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 7.071264367816093e-05,
+      "loss": 1.7694,
+      "step": 652
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 7.066666666666667e-05,
+      "loss": 1.6475,
+      "step": 653
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 7.06206896551724e-05,
+      "loss": 2.016,
+      "step": 654
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.22560975609756098,
+      "eval_loss": 1.8887790441513062,
+      "eval_runtime": 13.7699,
+      "eval_samples_per_second": 11.91,
+      "eval_steps_per_second": 1.525,
+      "step": 654
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 7.057471264367816e-05,
+      "loss": 1.5601,
+      "step": 655
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 7.052873563218391e-05,
+      "loss": 2.0661,
+      "step": 656
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 7.048275862068966e-05,
+      "loss": 1.8234,
+      "step": 657
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 7.04367816091954e-05,
+      "loss": 1.7882,
+      "step": 658
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 7.039080459770115e-05,
+      "loss": 1.9746,
+      "step": 659
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.03448275862069e-05,
+      "loss": 1.8735,
+      "step": 660
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.029885057471264e-05,
+      "loss": 2.1546,
+      "step": 661
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 7.025287356321839e-05,
+      "loss": 1.6602,
+      "step": 662
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 7.020689655172415e-05,
+      "loss": 1.5793,
+      "step": 663
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 7.016091954022988e-05,
+      "loss": 2.124,
+      "step": 664
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 7.011494252873564e-05,
+      "loss": 2.033,
+      "step": 665
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 7.006896551724137e-05,
+      "loss": 2.2248,
+      "step": 666
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 7.002298850574713e-05,
+      "loss": 1.9048,
+      "step": 667
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 6.997701149425288e-05,
+      "loss": 1.9442,
+      "step": 668
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 6.993103448275863e-05,
+      "loss": 1.93,
+      "step": 669
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 6.988505747126437e-05,
+      "loss": 1.9111,
+      "step": 670
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 6.983908045977012e-05,
+      "loss": 1.6963,
+      "step": 671
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 6.979310344827587e-05,
+      "loss": 2.0454,
+      "step": 672
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 6.974712643678161e-05,
+      "loss": 1.8958,
+      "step": 673
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 6.970114942528736e-05,
+      "loss": 1.7178,
+      "step": 674
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 6.96551724137931e-05,
+      "loss": 1.6203,
+      "step": 675
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 6.960919540229885e-05,
+      "loss": 1.707,
+      "step": 676
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 6.95632183908046e-05,
+      "loss": 2.0311,
+      "step": 677
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 6.951724137931034e-05,
+      "loss": 2.6637,
+      "step": 678
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 6.947126436781609e-05,
+      "loss": 1.4525,
+      "step": 679
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 6.942528735632185e-05,
+      "loss": 1.7622,
+      "step": 680
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 6.937931034482758e-05,
+      "loss": 1.8638,
+      "step": 681
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 6.933333333333334e-05,
+      "loss": 1.6286,
+      "step": 682
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 6.928735632183908e-05,
+      "loss": 1.9795,
+      "step": 683
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 6.924137931034484e-05,
+      "loss": 1.7845,
+      "step": 684
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 6.919540229885058e-05,
+      "loss": 1.8701,
+      "step": 685
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 6.914942528735633e-05,
+      "loss": 1.7396,
+      "step": 686
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 6.910344827586207e-05,
+      "loss": 1.5553,
+      "step": 687
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 6.905747126436782e-05,
+      "loss": 1.6608,
+      "step": 688
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 6.901149425287357e-05,
+      "loss": 1.5757,
+      "step": 689
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 6.896551724137931e-05,
+      "loss": 2.6255,
+      "step": 690
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 6.891954022988506e-05,
+      "loss": 1.9033,
+      "step": 691
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 6.887356321839082e-05,
+      "loss": 2.0218,
+      "step": 692
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 6.882758620689655e-05,
+      "loss": 2.0052,
+      "step": 693
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 6.87816091954023e-05,
+      "loss": 1.8732,
+      "step": 694
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 6.873563218390804e-05,
+      "loss": 1.5153,
+      "step": 695
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 6.868965517241379e-05,
+      "loss": 1.7554,
+      "step": 696
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 6.864367816091955e-05,
+      "loss": 2.006,
+      "step": 697
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 6.859770114942528e-05,
+      "loss": 1.8607,
+      "step": 698
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 6.855172413793104e-05,
+      "loss": 1.6597,
+      "step": 699
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 6.850574712643678e-05,
+      "loss": 1.6479,
+      "step": 700
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 6.845977011494254e-05,
+      "loss": 2.8652,
+      "step": 701
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 6.841379310344828e-05,
+      "loss": 2.1702,
+      "step": 702
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 6.836781609195403e-05,
+      "loss": 1.533,
+      "step": 703
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 6.832183908045978e-05,
+      "loss": 1.6243,
+      "step": 704
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 6.827586206896552e-05,
+      "loss": 2.1925,
+      "step": 705
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 6.822988505747127e-05,
+      "loss": 1.89,
+      "step": 706
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 6.818390804597701e-05,
+      "loss": 2.5444,
+      "step": 707
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 6.813793103448276e-05,
+      "loss": 1.6346,
+      "step": 708
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 6.809195402298851e-05,
+      "loss": 1.7669,
+      "step": 709
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 6.804597701149425e-05,
+      "loss": 1.8022,
+      "step": 710
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 6.800000000000001e-05,
+      "loss": 2.0021,
+      "step": 711
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 6.795402298850575e-05,
+      "loss": 1.5726,
+      "step": 712
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 6.79080459770115e-05,
+      "loss": 1.9782,
+      "step": 713
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 6.786206896551724e-05,
+      "loss": 1.6034,
+      "step": 714
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 6.781609195402298e-05,
+      "loss": 1.6787,
+      "step": 715
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 6.777011494252874e-05,
+      "loss": 1.9569,
+      "step": 716
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 6.772413793103448e-05,
+      "loss": 2.4325,
+      "step": 717
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 6.767816091954024e-05,
+      "loss": 1.7354,
+      "step": 718
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 6.763218390804597e-05,
+      "loss": 1.5884,
+      "step": 719
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 6.758620689655173e-05,
+      "loss": 1.5739,
+      "step": 720
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 6.754022988505748e-05,
+      "loss": 1.6934,
+      "step": 721
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 6.749425287356322e-05,
+      "loss": 1.6077,
+      "step": 722
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 6.744827586206897e-05,
+      "loss": 1.7529,
+      "step": 723
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 6.740229885057472e-05,
+      "loss": 1.7295,
+      "step": 724
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 6.735632183908046e-05,
+      "loss": 1.8527,
+      "step": 725
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 6.731034482758621e-05,
+      "loss": 1.625,
+      "step": 726
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 6.726436781609195e-05,
+      "loss": 1.5915,
+      "step": 727
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 6.721839080459771e-05,
+      "loss": 1.8542,
+      "step": 728
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 6.717241379310345e-05,
+      "loss": 1.6406,
+      "step": 729
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 6.71264367816092e-05,
+      "loss": 1.7165,
+      "step": 730
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 6.708045977011494e-05,
+      "loss": 1.7225,
+      "step": 731
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 6.70344827586207e-05,
+      "loss": 1.7093,
+      "step": 732
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 6.698850574712645e-05,
+      "loss": 2.6672,
+      "step": 733
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 6.694252873563218e-05,
+      "loss": 2.3314,
+      "step": 734
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 6.689655172413794e-05,
+      "loss": 2.0335,
+      "step": 735
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 6.685057471264367e-05,
+      "loss": 2.0239,
+      "step": 736
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 6.680459770114943e-05,
+      "loss": 1.7988,
+      "step": 737
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 6.675862068965518e-05,
+      "loss": 1.6471,
+      "step": 738
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 6.671264367816092e-05,
+      "loss": 1.59,
+      "step": 739
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 2.0796,
+      "step": 740
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 6.662068965517242e-05,
+      "loss": 1.6006,
+      "step": 741
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 6.657471264367816e-05,
+      "loss": 2.0755,
+      "step": 742
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 6.652873563218391e-05,
+      "loss": 1.4642,
+      "step": 743
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 6.648275862068966e-05,
+      "loss": 1.4842,
+      "step": 744
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 6.643678160919542e-05,
+      "loss": 1.4954,
+      "step": 745
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 6.639080459770115e-05,
+      "loss": 1.6323,
+      "step": 746
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 6.634482758620691e-05,
+      "loss": 2.0827,
+      "step": 747
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 6.629885057471264e-05,
+      "loss": 1.6825,
+      "step": 748
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 6.62528735632184e-05,
+      "loss": 2.3032,
+      "step": 749
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 6.620689655172415e-05,
+      "loss": 2.332,
+      "step": 750
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 6.616091954022989e-05,
+      "loss": 1.4476,
+      "step": 751
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 6.611494252873564e-05,
+      "loss": 1.7376,
+      "step": 752
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 6.606896551724139e-05,
+      "loss": 2.4022,
+      "step": 753
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 6.602298850574713e-05,
+      "loss": 1.8208,
+      "step": 754
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 6.597701149425288e-05,
+      "loss": 2.1134,
+      "step": 755
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 6.593103448275862e-05,
+      "loss": 1.8843,
+      "step": 756
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 6.588505747126437e-05,
+      "loss": 1.7562,
+      "step": 757
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 6.583908045977012e-05,
+      "loss": 2.5618,
+      "step": 758
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 6.579310344827586e-05,
+      "loss": 1.6574,
+      "step": 759
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 6.574712643678161e-05,
+      "loss": 1.4478,
+      "step": 760
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 6.570114942528736e-05,
+      "loss": 1.756,
+      "step": 761
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 6.56551724137931e-05,
+      "loss": 1.5016,
+      "step": 762
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 6.560919540229885e-05,
+      "loss": 1.625,
+      "step": 763
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.17682926829268292,
+      "eval_loss": 1.9067740440368652,
+      "eval_runtime": 13.7816,
+      "eval_samples_per_second": 11.9,
+      "eval_steps_per_second": 1.524,
+      "step": 763
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 6.556321839080461e-05,
+      "loss": 1.681,
+      "step": 764
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 6.551724137931034e-05,
+      "loss": 2.0745,
+      "step": 765
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 6.54712643678161e-05,
+      "loss": 2.0957,
+      "step": 766
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 6.542528735632183e-05,
+      "loss": 1.736,
+      "step": 767
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 6.53793103448276e-05,
+      "loss": 1.8683,
+      "step": 768
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 6.533333333333334e-05,
+      "loss": 1.7137,
+      "step": 769
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 6.528735632183909e-05,
+      "loss": 1.3867,
+      "step": 770
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 6.524137931034483e-05,
+      "loss": 1.8883,
+      "step": 771
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 6.519540229885058e-05,
+      "loss": 1.863,
+      "step": 772
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 6.514942528735633e-05,
+      "loss": 1.5835,
+      "step": 773
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 6.510344827586207e-05,
+      "loss": 1.3963,
+      "step": 774
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 6.505747126436782e-05,
+      "loss": 1.9046,
+      "step": 775
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 6.501149425287356e-05,
+      "loss": 2.1147,
+      "step": 776
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 6.496551724137931e-05,
+      "loss": 1.6136,
+      "step": 777
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 6.491954022988506e-05,
+      "loss": 1.9569,
+      "step": 778
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 6.48735632183908e-05,
+      "loss": 1.7619,
+      "step": 779
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 6.482758620689655e-05,
+      "loss": 1.9777,
+      "step": 780
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 6.478160919540231e-05,
+      "loss": 1.8252,
+      "step": 781
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 6.473563218390804e-05,
+      "loss": 2.1406,
+      "step": 782
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 6.46896551724138e-05,
+      "loss": 1.6481,
+      "step": 783
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 6.464367816091954e-05,
+      "loss": 1.8084,
+      "step": 784
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 6.45977011494253e-05,
+      "loss": 2.0542,
+      "step": 785
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 6.455172413793104e-05,
+      "loss": 1.9229,
+      "step": 786
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 6.450574712643679e-05,
+      "loss": 1.7456,
+      "step": 787
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 6.445977011494253e-05,
+      "loss": 1.7352,
+      "step": 788
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 6.441379310344828e-05,
+      "loss": 2.5968,
+      "step": 789
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 6.436781609195403e-05,
+      "loss": 1.9098,
+      "step": 790
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 6.432183908045977e-05,
+      "loss": 1.6266,
+      "step": 791
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 6.427586206896552e-05,
+      "loss": 1.8179,
+      "step": 792
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 6.422988505747128e-05,
+      "loss": 1.5225,
+      "step": 793
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 6.418390804597701e-05,
+      "loss": 2.2358,
+      "step": 794
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 6.413793103448276e-05,
+      "loss": 1.6493,
+      "step": 795
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 6.40919540229885e-05,
+      "loss": 2.4172,
+      "step": 796
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 6.404597701149425e-05,
+      "loss": 1.4154,
+      "step": 797
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 6.400000000000001e-05,
+      "loss": 1.7445,
+      "step": 798
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 6.395402298850574e-05,
+      "loss": 1.8722,
+      "step": 799
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 6.39080459770115e-05,
+      "loss": 1.5848,
+      "step": 800
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 6.386206896551724e-05,
+      "loss": 1.4268,
+      "step": 801
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 6.3816091954023e-05,
+      "loss": 1.7007,
+      "step": 802
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 6.377011494252874e-05,
+      "loss": 1.9852,
+      "step": 803
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 6.372413793103449e-05,
+      "loss": 1.6433,
+      "step": 804
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 6.367816091954023e-05,
+      "loss": 1.5815,
+      "step": 805
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 6.363218390804598e-05,
+      "loss": 1.6621,
+      "step": 806
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 6.358620689655173e-05,
+      "loss": 1.6776,
+      "step": 807
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 6.354022988505747e-05,
+      "loss": 1.7482,
+      "step": 808
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 6.349425287356322e-05,
+      "loss": 1.7687,
+      "step": 809
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 6.344827586206897e-05,
+      "loss": 1.5576,
+      "step": 810
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 6.340229885057471e-05,
+      "loss": 2.1086,
+      "step": 811
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 6.335632183908047e-05,
+      "loss": 2.5109,
+      "step": 812
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 6.33103448275862e-05,
+      "loss": 2.0659,
+      "step": 813
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 6.326436781609195e-05,
+      "loss": 1.4474,
+      "step": 814
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 6.32183908045977e-05,
+      "loss": 2.2822,
+      "step": 815
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 6.317241379310344e-05,
+      "loss": 1.7428,
+      "step": 816
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 6.31264367816092e-05,
+      "loss": 1.7007,
+      "step": 817
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 6.308045977011494e-05,
+      "loss": 1.8874,
+      "step": 818
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 6.30344827586207e-05,
+      "loss": 2.1263,
+      "step": 819
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 6.298850574712643e-05,
+      "loss": 1.4705,
+      "step": 820
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 6.294252873563219e-05,
+      "loss": 1.6245,
+      "step": 821
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 6.289655172413794e-05,
+      "loss": 2.3783,
+      "step": 822
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 6.285057471264368e-05,
+      "loss": 1.8276,
+      "step": 823
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 6.280459770114943e-05,
+      "loss": 2.0007,
+      "step": 824
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 6.275862068965517e-05,
+      "loss": 1.6224,
+      "step": 825
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 6.271264367816092e-05,
+      "loss": 2.0405,
+      "step": 826
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 6.266666666666667e-05,
+      "loss": 2.4634,
+      "step": 827
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 6.262068965517241e-05,
+      "loss": 1.7811,
+      "step": 828
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 6.257471264367817e-05,
+      "loss": 1.4993,
+      "step": 829
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 6.25287356321839e-05,
+      "loss": 1.8398,
+      "step": 830
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 6.248275862068967e-05,
+      "loss": 1.6232,
+      "step": 831
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 6.24367816091954e-05,
+      "loss": 2.1147,
+      "step": 832
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 6.239080459770116e-05,
+      "loss": 2.4437,
+      "step": 833
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 6.23448275862069e-05,
+      "loss": 2.0591,
+      "step": 834
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 6.229885057471264e-05,
+      "loss": 1.5824,
+      "step": 835
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 6.22528735632184e-05,
+      "loss": 1.6925,
+      "step": 836
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 6.220689655172413e-05,
+      "loss": 1.8022,
+      "step": 837
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 6.216091954022989e-05,
+      "loss": 1.8923,
+      "step": 838
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 6.211494252873564e-05,
+      "loss": 1.6699,
+      "step": 839
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 6.206896551724138e-05,
+      "loss": 1.6175,
+      "step": 840
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 6.202298850574713e-05,
+      "loss": 1.6865,
+      "step": 841
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 6.197701149425288e-05,
+      "loss": 2.064,
+      "step": 842
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 6.193103448275862e-05,
+      "loss": 1.7609,
+      "step": 843
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 6.188505747126437e-05,
+      "loss": 1.6167,
+      "step": 844
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 6.183908045977011e-05,
+      "loss": 2.0716,
+      "step": 845
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 6.179310344827587e-05,
+      "loss": 1.6818,
+      "step": 846
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 6.174712643678161e-05,
+      "loss": 1.6444,
+      "step": 847
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 6.170114942528737e-05,
+      "loss": 1.6883,
+      "step": 848
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 6.16551724137931e-05,
+      "loss": 1.5643,
+      "step": 849
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 6.160919540229886e-05,
+      "loss": 1.9596,
+      "step": 850
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 6.15632183908046e-05,
+      "loss": 1.8804,
+      "step": 851
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 6.151724137931035e-05,
+      "loss": 2.422,
+      "step": 852
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 6.14712643678161e-05,
+      "loss": 1.6704,
+      "step": 853
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 6.142528735632183e-05,
+      "loss": 2.1497,
+      "step": 854
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 6.137931034482759e-05,
+      "loss": 2.028,
+      "step": 855
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 6.133333333333334e-05,
+      "loss": 1.652,
+      "step": 856
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 6.128735632183908e-05,
+      "loss": 2.2965,
+      "step": 857
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 6.124137931034483e-05,
+      "loss": 2.4193,
+      "step": 858
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 6.119540229885058e-05,
+      "loss": 1.9907,
+      "step": 859
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 6.114942528735632e-05,
+      "loss": 1.6479,
+      "step": 860
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 6.110344827586207e-05,
+      "loss": 1.7487,
+      "step": 861
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 6.105747126436782e-05,
+      "loss": 2.3571,
+      "step": 862
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 6.101149425287357e-05,
+      "loss": 2.0465,
+      "step": 863
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 6.096551724137931e-05,
+      "loss": 1.6825,
+      "step": 864
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 6.091954022988506e-05,
+      "loss": 1.8114,
+      "step": 865
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 6.087356321839081e-05,
+      "loss": 1.8555,
+      "step": 866
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 6.082758620689656e-05,
+      "loss": 1.7729,
+      "step": 867
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 6.07816091954023e-05,
+      "loss": 1.8009,
+      "step": 868
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 6.0735632183908054e-05,
+      "loss": 1.8859,
+      "step": 869
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 6.068965517241379e-05,
+      "loss": 2.0726,
+      "step": 870
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 6.0643678160919546e-05,
+      "loss": 1.6642,
+      "step": 871
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 6.059770114942529e-05,
+      "loss": 2.0002,
+      "step": 872
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.1951219512195122,
+      "eval_loss": 1.8909108638763428,
+      "eval_runtime": 13.8035,
+      "eval_samples_per_second": 11.881,
+      "eval_steps_per_second": 1.521,
+      "step": 872
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 6.0551724137931046e-05,
+      "loss": 1.6178,
+      "step": 873
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 6.0505747126436785e-05,
+      "loss": 1.6294,
+      "step": 874
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 6.0459770114942525e-05,
+      "loss": 1.972,
+      "step": 875
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 6.041379310344828e-05,
+      "loss": 1.7635,
+      "step": 876
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 6.0367816091954024e-05,
+      "loss": 1.756,
+      "step": 877
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 6.032183908045978e-05,
+      "loss": 1.5892,
+      "step": 878
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 6.027586206896552e-05,
+      "loss": 1.6401,
+      "step": 879
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 6.022988505747127e-05,
+      "loss": 1.7856,
+      "step": 880
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 6.018390804597701e-05,
+      "loss": 1.4993,
+      "step": 881
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 6.013793103448276e-05,
+      "loss": 1.5705,
+      "step": 882
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 6.009195402298851e-05,
+      "loss": 2.0047,
+      "step": 883
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 6.004597701149426e-05,
+      "loss": 1.7207,
+      "step": 884
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 6e-05,
+      "loss": 1.9867,
+      "step": 885
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 5.9954022988505755e-05,
+      "loss": 2.202,
+      "step": 886
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 5.9908045977011494e-05,
+      "loss": 1.8605,
+      "step": 887
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 5.986206896551725e-05,
+      "loss": 2.2139,
+      "step": 888
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 5.9816091954022994e-05,
+      "loss": 1.8851,
+      "step": 889
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 5.977011494252874e-05,
+      "loss": 1.6016,
+      "step": 890
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 5.9724137931034486e-05,
+      "loss": 1.5799,
+      "step": 891
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 5.967816091954024e-05,
+      "loss": 1.5456,
+      "step": 892
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 5.963218390804598e-05,
+      "loss": 2.1439,
+      "step": 893
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 5.958620689655172e-05,
+      "loss": 1.8843,
+      "step": 894
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 5.954022988505747e-05,
+      "loss": 1.8838,
+      "step": 895
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 5.949425287356322e-05,
+      "loss": 2.6497,
+      "step": 896
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 5.944827586206897e-05,
+      "loss": 1.9233,
+      "step": 897
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 5.940229885057471e-05,
+      "loss": 1.8294,
+      "step": 898
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 5.9356321839080464e-05,
+      "loss": 1.8224,
+      "step": 899
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 5.93103448275862e-05,
+      "loss": 1.8685,
+      "step": 900
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 5.9264367816091956e-05,
+      "loss": 1.7674,
+      "step": 901
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 5.92183908045977e-05,
+      "loss": 1.6688,
+      "step": 902
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 5.9172413793103456e-05,
+      "loss": 1.4915,
+      "step": 903
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 5.9126436781609195e-05,
+      "loss": 1.7163,
+      "step": 904
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 5.908045977011495e-05,
+      "loss": 1.6834,
+      "step": 905
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 5.903448275862069e-05,
+      "loss": 1.9749,
+      "step": 906
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 5.898850574712644e-05,
+      "loss": 1.8823,
+      "step": 907
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 5.894252873563219e-05,
+      "loss": 2.2928,
+      "step": 908
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 5.889655172413794e-05,
+      "loss": 2.4523,
+      "step": 909
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 5.885057471264368e-05,
+      "loss": 1.71,
+      "step": 910
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 5.880459770114943e-05,
+      "loss": 2.1211,
+      "step": 911
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 5.875862068965517e-05,
+      "loss": 1.7277,
+      "step": 912
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 5.871264367816092e-05,
+      "loss": 1.5869,
+      "step": 913
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 5.866666666666667e-05,
+      "loss": 1.9443,
+      "step": 914
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 5.862068965517241e-05,
+      "loss": 1.6956,
+      "step": 915
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 5.8574712643678165e-05,
+      "loss": 1.6989,
+      "step": 916
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 5.8528735632183904e-05,
+      "loss": 1.62,
+      "step": 917
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 5.848275862068966e-05,
+      "loss": 1.5386,
+      "step": 918
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 5.8436781609195404e-05,
+      "loss": 2.2,
+      "step": 919
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 5.839080459770116e-05,
+      "loss": 1.6717,
+      "step": 920
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 5.8344827586206896e-05,
+      "loss": 1.9118,
+      "step": 921
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 5.829885057471265e-05,
+      "loss": 1.7552,
+      "step": 922
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 5.825287356321839e-05,
+      "loss": 2.0762,
+      "step": 923
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 5.820689655172414e-05,
+      "loss": 1.9935,
+      "step": 924
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 5.816091954022989e-05,
+      "loss": 1.9007,
+      "step": 925
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 5.811494252873564e-05,
+      "loss": 2.12,
+      "step": 926
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 5.806896551724138e-05,
+      "loss": 2.741,
+      "step": 927
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 5.8022988505747134e-05,
+      "loss": 1.7995,
+      "step": 928
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 5.7977011494252874e-05,
+      "loss": 1.7904,
+      "step": 929
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 5.7931034482758627e-05,
+      "loss": 2.2375,
+      "step": 930
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 5.788505747126437e-05,
+      "loss": 1.8127,
+      "step": 931
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 5.7839080459770126e-05,
+      "loss": 1.95,
+      "step": 932
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 5.7793103448275866e-05,
+      "loss": 1.8459,
+      "step": 933
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 5.7747126436781605e-05,
+      "loss": 1.6375,
+      "step": 934
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 5.770114942528736e-05,
+      "loss": 2.4199,
+      "step": 935
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 5.7655172413793105e-05,
+      "loss": 1.7728,
+      "step": 936
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 5.760919540229886e-05,
+      "loss": 1.8201,
+      "step": 937
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 5.75632183908046e-05,
+      "loss": 1.6937,
+      "step": 938
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 5.751724137931035e-05,
+      "loss": 1.6769,
+      "step": 939
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 5.747126436781609e-05,
+      "loss": 1.5275,
+      "step": 940
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 5.742528735632184e-05,
+      "loss": 1.6706,
+      "step": 941
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 5.737931034482759e-05,
+      "loss": 2.318,
+      "step": 942
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 5.7333333333333336e-05,
+      "loss": 1.7674,
+      "step": 943
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 5.728735632183908e-05,
+      "loss": 2.2681,
+      "step": 944
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 5.7241379310344835e-05,
+      "loss": 1.7778,
+      "step": 945
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 5.7195402298850575e-05,
+      "loss": 2.0973,
+      "step": 946
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 5.714942528735633e-05,
+      "loss": 1.4997,
+      "step": 947
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 5.710344827586207e-05,
+      "loss": 2.0493,
+      "step": 948
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 5.705747126436782e-05,
+      "loss": 1.7192,
+      "step": 949
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 5.7011494252873567e-05,
+      "loss": 1.8639,
+      "step": 950
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 5.696551724137932e-05,
+      "loss": 1.7383,
+      "step": 951
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 5.691954022988506e-05,
+      "loss": 1.6157,
+      "step": 952
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 5.68735632183908e-05,
+      "loss": 1.9453,
+      "step": 953
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 5.682758620689655e-05,
+      "loss": 2.1405,
+      "step": 954
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 5.67816091954023e-05,
+      "loss": 2.4316,
+      "step": 955
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 5.673563218390805e-05,
+      "loss": 1.9035,
+      "step": 956
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 5.668965517241379e-05,
+      "loss": 1.5785,
+      "step": 957
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 5.6643678160919544e-05,
+      "loss": 2.3794,
+      "step": 958
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 5.6597701149425283e-05,
+      "loss": 1.9325,
+      "step": 959
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 5.6551724137931037e-05,
+      "loss": 2.1016,
+      "step": 960
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 5.650574712643678e-05,
+      "loss": 1.6899,
+      "step": 961
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 5.6459770114942536e-05,
+      "loss": 2.4627,
+      "step": 962
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 5.6413793103448275e-05,
+      "loss": 1.6683,
+      "step": 963
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 5.636781609195403e-05,
+      "loss": 1.9974,
+      "step": 964
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 5.632183908045977e-05,
+      "loss": 1.5358,
+      "step": 965
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 5.627586206896552e-05,
+      "loss": 1.8221,
+      "step": 966
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 5.622988505747127e-05,
+      "loss": 1.6338,
+      "step": 967
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 5.618390804597702e-05,
+      "loss": 2.0472,
+      "step": 968
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 5.613793103448276e-05,
+      "loss": 1.4399,
+      "step": 969
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 5.609195402298851e-05,
+      "loss": 1.5876,
+      "step": 970
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 5.604597701149425e-05,
+      "loss": 1.8649,
+      "step": 971
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 5.6000000000000006e-05,
+      "loss": 1.7972,
+      "step": 972
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.595402298850575e-05,
+      "loss": 1.9564,
+      "step": 973
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.590804597701149e-05,
+      "loss": 1.8359,
+      "step": 974
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.5862068965517245e-05,
+      "loss": 1.6019,
+      "step": 975
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.5816091954022984e-05,
+      "loss": 1.8045,
+      "step": 976
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.577011494252874e-05,
+      "loss": 1.6011,
+      "step": 977
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.5724137931034484e-05,
+      "loss": 1.6857,
+      "step": 978
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.567816091954024e-05,
+      "loss": 1.7729,
+      "step": 979
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.5632183908045976e-05,
+      "loss": 1.9616,
+      "step": 980
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 5.558620689655173e-05,
+      "loss": 1.7906,
+      "step": 981
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.21951219512195122,
+      "eval_loss": 1.8828065395355225,
+      "eval_runtime": 13.7791,
+      "eval_samples_per_second": 11.902,
+      "eval_steps_per_second": 1.524,
+      "step": 981
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.554022988505747e-05,
+      "loss": 1.5441,
+      "step": 982
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 5.549425287356322e-05,
+      "loss": 2.9479,
+      "step": 983
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 5.544827586206897e-05,
+      "loss": 1.6123,
+      "step": 984
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 5.540229885057472e-05,
+      "loss": 2.0708,
+      "step": 985
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 5.535632183908046e-05,
+      "loss": 1.8866,
+      "step": 986
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 5.5310344827586214e-05,
+      "loss": 1.9072,
+      "step": 987
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 5.5264367816091954e-05,
+      "loss": 1.4451,
+      "step": 988
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 5.521839080459771e-05,
+      "loss": 1.6772,
+      "step": 989
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 5.517241379310345e-05,
+      "loss": 2.1719,
+      "step": 990
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 5.5126436781609206e-05,
+      "loss": 1.8784,
+      "step": 991
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 5.5080459770114946e-05,
+      "loss": 3.1177,
+      "step": 992
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 5.5034482758620685e-05,
+      "loss": 1.7357,
+      "step": 993
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 5.498850574712644e-05,
+      "loss": 1.7998,
+      "step": 994
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 5.4942528735632185e-05,
+      "loss": 1.5747,
+      "step": 995
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 5.489655172413794e-05,
+      "loss": 1.8302,
+      "step": 996
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 5.485057471264368e-05,
+      "loss": 1.5742,
+      "step": 997
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 5.480459770114943e-05,
+      "loss": 1.5708,
+      "step": 998
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 5.475862068965517e-05,
+      "loss": 1.7603,
+      "step": 999
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 5.471264367816092e-05,
+      "loss": 2.3278,
+      "step": 1000
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 5.466666666666666e-05,
+      "loss": 1.6209,
+      "step": 1001
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 5.4620689655172416e-05,
+      "loss": 1.5913,
+      "step": 1002
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 5.457471264367816e-05,
+      "loss": 1.9668,
+      "step": 1003
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 5.4528735632183915e-05,
+      "loss": 1.7516,
+      "step": 1004
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 5.4482758620689655e-05,
+      "loss": 2.2046,
+      "step": 1005
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 5.443678160919541e-05,
+      "loss": 1.6139,
+      "step": 1006
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 5.439080459770115e-05,
+      "loss": 2.0814,
+      "step": 1007
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 5.43448275862069e-05,
+      "loss": 1.7788,
+      "step": 1008
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 5.429885057471265e-05,
+      "loss": 2.279,
+      "step": 1009
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 5.42528735632184e-05,
+      "loss": 1.6279,
+      "step": 1010
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 5.420689655172414e-05,
+      "loss": 1.987,
+      "step": 1011
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 5.416091954022989e-05,
+      "loss": 1.9465,
+      "step": 1012
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 5.411494252873563e-05,
+      "loss": 1.5745,
+      "step": 1013
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 5.406896551724138e-05,
+      "loss": 2.1603,
+      "step": 1014
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 5.402298850574713e-05,
+      "loss": 1.7492,
+      "step": 1015
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 5.397701149425287e-05,
+      "loss": 1.7249,
+      "step": 1016
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 5.3931034482758624e-05,
+      "loss": 1.7231,
+      "step": 1017
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 5.3885057471264364e-05,
+      "loss": 1.5967,
+      "step": 1018
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 5.383908045977012e-05,
+      "loss": 2.3701,
+      "step": 1019
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 5.379310344827586e-05,
+      "loss": 2.0977,
+      "step": 1020
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 5.3747126436781616e-05,
+      "loss": 1.5535,
+      "step": 1021
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 5.3701149425287356e-05,
+      "loss": 2.1187,
+      "step": 1022
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 5.365517241379311e-05,
+      "loss": 2.507,
+      "step": 1023
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 5.360919540229885e-05,
+      "loss": 1.6999,
+      "step": 1024
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 5.35632183908046e-05,
+      "loss": 1.6494,
+      "step": 1025
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 5.351724137931035e-05,
+      "loss": 1.6665,
+      "step": 1026
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 5.34712643678161e-05,
+      "loss": 1.61,
+      "step": 1027
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 5.342528735632184e-05,
+      "loss": 1.5044,
+      "step": 1028
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 5.3379310344827594e-05,
+      "loss": 1.9559,
+      "step": 1029
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 5.333333333333333e-05,
+      "loss": 1.5228,
+      "step": 1030
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 5.3287356321839086e-05,
+      "loss": 1.8773,
+      "step": 1031
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 5.324137931034483e-05,
+      "loss": 1.7751,
+      "step": 1032
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 5.319540229885057e-05,
+      "loss": 1.5228,
+      "step": 1033
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 5.3149425287356325e-05,
+      "loss": 1.9888,
+      "step": 1034
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 5.3103448275862065e-05,
+      "loss": 1.6795,
+      "step": 1035
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 5.305747126436782e-05,
+      "loss": 2.0138,
+      "step": 1036
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 5.3011494252873564e-05,
+      "loss": 1.6971,
+      "step": 1037
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 5.296551724137932e-05,
+      "loss": 1.7777,
+      "step": 1038
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 5.291954022988506e-05,
+      "loss": 2.1331,
+      "step": 1039
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 5.287356321839081e-05,
+      "loss": 1.8372,
+      "step": 1040
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 5.282758620689655e-05,
+      "loss": 1.5228,
+      "step": 1041
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 5.27816091954023e-05,
+      "loss": 1.8153,
+      "step": 1042
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 5.273563218390805e-05,
+      "loss": 1.8493,
+      "step": 1043
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 5.26896551724138e-05,
+      "loss": 2.7891,
+      "step": 1044
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 5.264367816091954e-05,
+      "loss": 1.7777,
+      "step": 1045
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 5.2597701149425295e-05,
+      "loss": 1.8097,
+      "step": 1046
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 5.2551724137931034e-05,
+      "loss": 1.877,
+      "step": 1047
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 5.250574712643679e-05,
+      "loss": 1.7261,
+      "step": 1048
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 5.2459770114942534e-05,
+      "loss": 2.1932,
+      "step": 1049
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 5.241379310344828e-05,
+      "loss": 1.6466,
+      "step": 1050
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 5.2367816091954026e-05,
+      "loss": 1.45,
+      "step": 1051
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 5.232183908045978e-05,
+      "loss": 1.9414,
+      "step": 1052
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 5.227586206896552e-05,
+      "loss": 1.8628,
+      "step": 1053
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 5.222988505747126e-05,
+      "loss": 1.5793,
+      "step": 1054
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 5.218390804597701e-05,
+      "loss": 1.9985,
+      "step": 1055
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 5.213793103448276e-05,
+      "loss": 1.4803,
+      "step": 1056
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 5.209195402298851e-05,
+      "loss": 1.5633,
+      "step": 1057
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 5.204597701149425e-05,
+      "loss": 1.5339,
+      "step": 1058
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 5.2000000000000004e-05,
+      "loss": 1.6622,
+      "step": 1059
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 5.195402298850574e-05,
+      "loss": 1.9722,
+      "step": 1060
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 5.1908045977011496e-05,
+      "loss": 2.5116,
+      "step": 1061
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 5.186206896551724e-05,
+      "loss": 1.6868,
+      "step": 1062
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 5.1816091954022996e-05,
+      "loss": 1.9591,
+      "step": 1063
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 5.1770114942528735e-05,
+      "loss": 1.5259,
+      "step": 1064
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 5.172413793103449e-05,
+      "loss": 1.6895,
+      "step": 1065
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 5.167816091954023e-05,
+      "loss": 1.7995,
+      "step": 1066
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 5.163218390804598e-05,
+      "loss": 1.6273,
+      "step": 1067
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 5.158620689655173e-05,
+      "loss": 1.7168,
+      "step": 1068
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 5.154022988505748e-05,
+      "loss": 1.7664,
+      "step": 1069
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 5.149425287356322e-05,
+      "loss": 1.7959,
+      "step": 1070
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 5.144827586206897e-05,
+      "loss": 1.8727,
+      "step": 1071
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 5.140229885057471e-05,
+      "loss": 1.863,
+      "step": 1072
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 5.135632183908046e-05,
+      "loss": 1.9214,
+      "step": 1073
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 5.131034482758621e-05,
+      "loss": 1.4723,
+      "step": 1074
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 5.126436781609195e-05,
+      "loss": 2.0275,
+      "step": 1075
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 5.1218390804597705e-05,
+      "loss": 1.6831,
+      "step": 1076
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 5.1172413793103444e-05,
+      "loss": 1.9502,
+      "step": 1077
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 5.11264367816092e-05,
+      "loss": 1.6278,
+      "step": 1078
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 5.1080459770114944e-05,
+      "loss": 1.8112,
+      "step": 1079
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 5.10344827586207e-05,
+      "loss": 2.2209,
+      "step": 1080
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 5.0988505747126436e-05,
+      "loss": 1.7811,
+      "step": 1081
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 5.094252873563219e-05,
+      "loss": 1.8164,
+      "step": 1082
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 5.089655172413793e-05,
+      "loss": 1.8353,
+      "step": 1083
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 5.085057471264368e-05,
+      "loss": 1.5037,
+      "step": 1084
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 5.080459770114943e-05,
+      "loss": 2.0915,
+      "step": 1085
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 5.075862068965518e-05,
+      "loss": 2.4845,
+      "step": 1086
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 5.071264367816092e-05,
+      "loss": 1.8055,
+      "step": 1087
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 5.0666666666666674e-05,
+      "loss": 1.7321,
+      "step": 1088
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 5.0620689655172414e-05,
+      "loss": 2.7694,
+      "step": 1089
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 5.057471264367817e-05,
+      "loss": 1.5295,
+      "step": 1090
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.21951219512195122,
+      "eval_loss": 1.8966660499572754,
+      "eval_runtime": 13.7625,
+      "eval_samples_per_second": 11.916,
+      "eval_steps_per_second": 1.526,
+      "step": 1090
+    },
+    {
+      "epoch": 10.01,
+      "learning_rate": 5.052873563218391e-05,
+      "loss": 1.6022,
+      "step": 1091
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 5.0482758620689666e-05,
+      "loss": 1.6204,
+      "step": 1092
+    },
+    {
+      "epoch": 10.03,
+      "learning_rate": 5.0436781609195406e-05,
+      "loss": 1.7482,
+      "step": 1093
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 5.0390804597701145e-05,
+      "loss": 1.6257,
+      "step": 1094
+    },
+    {
+      "epoch": 10.05,
+      "learning_rate": 5.03448275862069e-05,
+      "loss": 1.797,
+      "step": 1095
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 5.0298850574712645e-05,
+      "loss": 1.8939,
+      "step": 1096
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 5.02528735632184e-05,
+      "loss": 2.1351,
+      "step": 1097
+    },
+    {
+      "epoch": 10.07,
+      "learning_rate": 5.020689655172414e-05,
+      "loss": 1.528,
+      "step": 1098
+    },
+    {
+      "epoch": 10.08,
+      "learning_rate": 5.016091954022989e-05,
+      "loss": 1.9282,
+      "step": 1099
+    },
+    {
+      "epoch": 10.09,
+      "learning_rate": 5.011494252873563e-05,
+      "loss": 1.6743,
+      "step": 1100
+    },
+    {
+      "epoch": 10.1,
+      "learning_rate": 5.006896551724138e-05,
+      "loss": 2.1668,
+      "step": 1101
+    },
+    {
+      "epoch": 10.11,
+      "learning_rate": 5.002298850574713e-05,
+      "loss": 1.6883,
+      "step": 1102
+    },
+    {
+      "epoch": 10.12,
+      "learning_rate": 4.9977011494252876e-05,
+      "loss": 1.6938,
+      "step": 1103
+    },
+    {
+      "epoch": 10.13,
+      "learning_rate": 4.993103448275862e-05,
+      "loss": 1.52,
+      "step": 1104
+    },
+    {
+      "epoch": 10.14,
+      "learning_rate": 4.988505747126437e-05,
+      "loss": 2.5396,
+      "step": 1105
+    },
+    {
+      "epoch": 10.15,
+      "learning_rate": 4.9839080459770115e-05,
+      "loss": 1.7773,
+      "step": 1106
+    },
+    {
+      "epoch": 10.16,
+      "learning_rate": 4.979310344827586e-05,
+      "loss": 1.8641,
+      "step": 1107
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 4.974712643678161e-05,
+      "loss": 2.0225,
+      "step": 1108
+    },
+    {
+      "epoch": 10.17,
+      "learning_rate": 4.970114942528736e-05,
+      "loss": 1.6182,
+      "step": 1109
+    },
+    {
+      "epoch": 10.18,
+      "learning_rate": 4.9655172413793107e-05,
+      "loss": 1.611,
+      "step": 1110
+    },
+    {
+      "epoch": 10.19,
+      "learning_rate": 4.960919540229885e-05,
+      "loss": 1.6986,
+      "step": 1111
+    },
+    {
+      "epoch": 10.2,
+      "learning_rate": 4.95632183908046e-05,
+      "loss": 1.5553,
+      "step": 1112
+    },
+    {
+      "epoch": 10.21,
+      "learning_rate": 4.9517241379310346e-05,
+      "loss": 2.1768,
+      "step": 1113
+    },
+    {
+      "epoch": 10.22,
+      "learning_rate": 4.947126436781609e-05,
+      "loss": 2.5448,
+      "step": 1114
+    },
+    {
+      "epoch": 10.23,
+      "learning_rate": 4.9425287356321845e-05,
+      "loss": 1.966,
+      "step": 1115
+    },
+    {
+      "epoch": 10.24,
+      "learning_rate": 4.937931034482759e-05,
+      "loss": 1.5207,
+      "step": 1116
+    },
+    {
+      "epoch": 10.25,
+      "learning_rate": 4.933333333333334e-05,
+      "loss": 1.7253,
+      "step": 1117
+    },
+    {
+      "epoch": 10.26,
+      "learning_rate": 4.9287356321839084e-05,
+      "loss": 2.0926,
+      "step": 1118
+    },
+    {
+      "epoch": 10.27,
+      "learning_rate": 4.924137931034483e-05,
+      "loss": 1.8249,
+      "step": 1119
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 4.9195402298850577e-05,
+      "loss": 1.8302,
+      "step": 1120
+    },
+    {
+      "epoch": 10.28,
+      "learning_rate": 4.914942528735633e-05,
+      "loss": 1.715,
+      "step": 1121
+    },
+    {
+      "epoch": 10.29,
+      "learning_rate": 4.9103448275862076e-05,
+      "loss": 1.7139,
+      "step": 1122
+    },
+    {
+      "epoch": 10.3,
+      "learning_rate": 4.9057471264367816e-05,
+      "loss": 1.7456,
+      "step": 1123
+    },
+    {
+      "epoch": 10.31,
+      "learning_rate": 4.901149425287356e-05,
+      "loss": 1.5951,
+      "step": 1124
+    },
+    {
+      "epoch": 10.32,
+      "learning_rate": 4.896551724137931e-05,
+      "loss": 2.1174,
+      "step": 1125
+    },
+    {
+      "epoch": 10.33,
+      "learning_rate": 4.891954022988506e-05,
+      "loss": 1.6536,
+      "step": 1126
+    },
+    {
+      "epoch": 10.34,
+      "learning_rate": 4.887356321839081e-05,
+      "loss": 1.6761,
+      "step": 1127
+    },
+    {
+      "epoch": 10.35,
+      "learning_rate": 4.8827586206896554e-05,
+      "loss": 1.6738,
+      "step": 1128
+    },
+    {
+      "epoch": 10.36,
+      "learning_rate": 4.87816091954023e-05,
+      "loss": 2.0975,
+      "step": 1129
+    },
+    {
+      "epoch": 10.37,
+      "learning_rate": 4.8735632183908047e-05,
+      "loss": 1.861,
+      "step": 1130
+    },
+    {
+      "epoch": 10.38,
+      "learning_rate": 4.868965517241379e-05,
+      "loss": 1.5345,
+      "step": 1131
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 4.864367816091954e-05,
+      "loss": 1.7257,
+      "step": 1132
+    },
+    {
+      "epoch": 10.39,
+      "learning_rate": 4.859770114942529e-05,
+      "loss": 1.799,
+      "step": 1133
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 4.855172413793104e-05,
+      "loss": 1.5788,
+      "step": 1134
+    },
+    {
+      "epoch": 10.41,
+      "learning_rate": 4.8505747126436785e-05,
+      "loss": 1.9471,
+      "step": 1135
+    },
+    {
+      "epoch": 10.42,
+      "learning_rate": 4.845977011494253e-05,
+      "loss": 1.7962,
+      "step": 1136
+    },
+    {
+      "epoch": 10.43,
+      "learning_rate": 4.841379310344828e-05,
+      "loss": 1.7238,
+      "step": 1137
+    },
+    {
+      "epoch": 10.44,
+      "learning_rate": 4.8367816091954024e-05,
+      "loss": 2.3117,
+      "step": 1138
+    },
+    {
+      "epoch": 10.45,
+      "learning_rate": 4.832183908045978e-05,
+      "loss": 1.7619,
+      "step": 1139
+    },
+    {
+      "epoch": 10.46,
+      "learning_rate": 4.827586206896552e-05,
+      "loss": 2.0288,
+      "step": 1140
+    },
+    {
+      "epoch": 10.47,
+      "learning_rate": 4.822988505747127e-05,
+      "loss": 1.8205,
+      "step": 1141
+    },
+    {
+      "epoch": 10.48,
+      "learning_rate": 4.8183908045977016e-05,
+      "loss": 2.0771,
+      "step": 1142
+    },
+    {
+      "epoch": 10.49,
+      "learning_rate": 4.8137931034482755e-05,
+      "loss": 1.8914,
+      "step": 1143
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 4.809195402298851e-05,
+      "loss": 1.9269,
+      "step": 1144
+    },
+    {
+      "epoch": 10.5,
+      "learning_rate": 4.8045977011494255e-05,
+      "loss": 1.8094,
+      "step": 1145
+    },
+    {
+      "epoch": 10.51,
+      "learning_rate": 4.8e-05,
+      "loss": 1.9557,
+      "step": 1146
+    },
+    {
+      "epoch": 10.52,
+      "learning_rate": 4.795402298850575e-05,
+      "loss": 1.617,
+      "step": 1147
+    },
+    {
+      "epoch": 10.53,
+      "learning_rate": 4.7908045977011494e-05,
+      "loss": 1.4893,
+      "step": 1148
+    },
+    {
+      "epoch": 10.54,
+      "learning_rate": 4.786206896551724e-05,
+      "loss": 1.466,
+      "step": 1149
+    },
+    {
+      "epoch": 10.55,
+      "learning_rate": 4.781609195402299e-05,
+      "loss": 1.8566,
+      "step": 1150
+    },
+    {
+      "epoch": 10.56,
+      "learning_rate": 4.777011494252874e-05,
+      "loss": 2.3088,
+      "step": 1151
+    },
+    {
+      "epoch": 10.57,
+      "learning_rate": 4.7724137931034486e-05,
+      "loss": 1.6722,
+      "step": 1152
+    },
+    {
+      "epoch": 10.58,
+      "learning_rate": 4.767816091954023e-05,
+      "loss": 1.875,
+      "step": 1153
+    },
+    {
+      "epoch": 10.59,
+      "learning_rate": 4.763218390804598e-05,
+      "loss": 2.2614,
+      "step": 1154
+    },
+    {
+      "epoch": 10.6,
+      "learning_rate": 4.7586206896551725e-05,
+      "loss": 1.7057,
+      "step": 1155
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 4.754022988505747e-05,
+      "loss": 2.3504,
+      "step": 1156
+    },
+    {
+      "epoch": 10.61,
+      "learning_rate": 4.7494252873563224e-05,
+      "loss": 2.4084,
+      "step": 1157
+    },
+    {
+      "epoch": 10.62,
+      "learning_rate": 4.744827586206897e-05,
+      "loss": 1.6128,
+      "step": 1158
+    },
+    {
+      "epoch": 10.63,
+      "learning_rate": 4.740229885057472e-05,
+      "loss": 1.7897,
+      "step": 1159
+    },
+    {
+      "epoch": 10.64,
+      "learning_rate": 4.735632183908046e-05,
+      "loss": 2.366,
+      "step": 1160
+    },
+    {
+      "epoch": 10.65,
+      "learning_rate": 4.731034482758621e-05,
+      "loss": 1.688,
+      "step": 1161
+    },
+    {
+      "epoch": 10.66,
+      "learning_rate": 4.7264367816091956e-05,
+      "loss": 1.5409,
+      "step": 1162
+    },
+    {
+      "epoch": 10.67,
+      "learning_rate": 4.72183908045977e-05,
+      "loss": 2.2352,
+      "step": 1163
+    },
+    {
+      "epoch": 10.68,
+      "learning_rate": 4.717241379310345e-05,
+      "loss": 1.6621,
+      "step": 1164
+    },
+    {
+      "epoch": 10.69,
+      "learning_rate": 4.7126436781609195e-05,
+      "loss": 1.6263,
+      "step": 1165
+    },
+    {
+      "epoch": 10.7,
+      "learning_rate": 4.708045977011494e-05,
+      "loss": 1.5474,
+      "step": 1166
+    },
+    {
+      "epoch": 10.71,
+      "learning_rate": 4.703448275862069e-05,
+      "loss": 1.9526,
+      "step": 1167
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 4.698850574712644e-05,
+      "loss": 1.6815,
+      "step": 1168
+    },
+    {
+      "epoch": 10.72,
+      "learning_rate": 4.694252873563219e-05,
+      "loss": 1.6676,
+      "step": 1169
+    },
+    {
+      "epoch": 10.73,
+      "learning_rate": 4.689655172413793e-05,
+      "loss": 1.77,
+      "step": 1170
+    },
+    {
+      "epoch": 10.74,
+      "learning_rate": 4.685057471264368e-05,
+      "loss": 1.494,
+      "step": 1171
+    },
+    {
+      "epoch": 10.75,
+      "learning_rate": 4.6804597701149426e-05,
+      "loss": 1.9129,
+      "step": 1172
+    },
+    {
+      "epoch": 10.76,
+      "learning_rate": 4.675862068965517e-05,
+      "loss": 1.993,
+      "step": 1173
+    },
+    {
+      "epoch": 10.77,
+      "learning_rate": 4.6712643678160925e-05,
+      "loss": 1.6449,
+      "step": 1174
+    },
+    {
+      "epoch": 10.78,
+      "learning_rate": 4.666666666666667e-05,
+      "loss": 2.3529,
+      "step": 1175
+    },
+    {
+      "epoch": 10.79,
+      "learning_rate": 4.662068965517242e-05,
+      "loss": 1.9031,
+      "step": 1176
+    },
+    {
+      "epoch": 10.8,
+      "learning_rate": 4.6574712643678164e-05,
+      "loss": 1.6279,
+      "step": 1177
+    },
+    {
+      "epoch": 10.81,
+      "learning_rate": 4.652873563218391e-05,
+      "loss": 1.8566,
+      "step": 1178
+    },
+    {
+      "epoch": 10.82,
+      "learning_rate": 4.648275862068966e-05,
+      "loss": 2.4531,
+      "step": 1179
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.643678160919541e-05,
+      "loss": 1.8475,
+      "step": 1180
+    },
+    {
+      "epoch": 10.83,
+      "learning_rate": 4.6390804597701156e-05,
+      "loss": 1.5609,
+      "step": 1181
+    },
+    {
+      "epoch": 10.84,
+      "learning_rate": 4.63448275862069e-05,
+      "loss": 2.027,
+      "step": 1182
+    },
+    {
+      "epoch": 10.85,
+      "learning_rate": 4.629885057471264e-05,
+      "loss": 2.1715,
+      "step": 1183
+    },
+    {
+      "epoch": 10.86,
+      "learning_rate": 4.625287356321839e-05,
+      "loss": 1.5091,
+      "step": 1184
+    },
+    {
+      "epoch": 10.87,
+      "learning_rate": 4.6206896551724135e-05,
+      "loss": 2.0099,
+      "step": 1185
+    },
+    {
+      "epoch": 10.88,
+      "learning_rate": 4.616091954022989e-05,
+      "loss": 1.6862,
+      "step": 1186
+    },
+    {
+      "epoch": 10.89,
+      "learning_rate": 4.6114942528735634e-05,
+      "loss": 1.4196,
+      "step": 1187
+    },
+    {
+      "epoch": 10.9,
+      "learning_rate": 4.606896551724138e-05,
+      "loss": 1.6296,
+      "step": 1188
+    },
+    {
+      "epoch": 10.91,
+      "learning_rate": 4.602298850574713e-05,
+      "loss": 1.8882,
+      "step": 1189
+    },
+    {
+      "epoch": 10.92,
+      "learning_rate": 4.597701149425287e-05,
+      "loss": 1.5142,
+      "step": 1190
+    },
+    {
+      "epoch": 10.93,
+      "learning_rate": 4.593103448275862e-05,
+      "loss": 1.9896,
+      "step": 1191
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.588505747126437e-05,
+      "loss": 2.1445,
+      "step": 1192
+    },
+    {
+      "epoch": 10.94,
+      "learning_rate": 4.583908045977012e-05,
+      "loss": 1.9979,
+      "step": 1193
+    },
+    {
+      "epoch": 10.95,
+      "learning_rate": 4.5793103448275865e-05,
+      "loss": 2.0643,
+      "step": 1194
+    },
+    {
+      "epoch": 10.96,
+      "learning_rate": 4.574712643678161e-05,
+      "loss": 1.5688,
+      "step": 1195
+    },
+    {
+      "epoch": 10.97,
+      "learning_rate": 4.570114942528736e-05,
+      "loss": 1.5461,
+      "step": 1196
+    },
+    {
+      "epoch": 10.98,
+      "learning_rate": 4.5655172413793104e-05,
+      "loss": 1.8079,
+      "step": 1197
+    },
+    {
+      "epoch": 10.99,
+      "learning_rate": 4.560919540229886e-05,
+      "loss": 2.1349,
+      "step": 1198
+    },
+    {
+      "epoch": 11.0,
+      "learning_rate": 4.5563218390804604e-05,
+      "loss": 1.7018,
+      "step": 1199
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.23780487804878048,
+      "eval_loss": 1.8844590187072754,
+      "eval_runtime": 13.7761,
+      "eval_samples_per_second": 11.905,
+      "eval_steps_per_second": 1.524,
+      "step": 1199
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 4.551724137931035e-05,
+      "loss": 1.6172,
+      "step": 1200
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 4.5471264367816096e-05,
+      "loss": 1.9045,
+      "step": 1201
+    },
+    {
+      "epoch": 11.03,
+      "learning_rate": 4.542528735632184e-05,
+      "loss": 2.3164,
+      "step": 1202
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 4.537931034482759e-05,
+      "loss": 1.9805,
+      "step": 1203
+    },
+    {
+      "epoch": 11.05,
+      "learning_rate": 4.5333333333333335e-05,
+      "loss": 1.8664,
+      "step": 1204
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 4.528735632183908e-05,
+      "loss": 2.0649,
+      "step": 1205
+    },
+    {
+      "epoch": 11.06,
+      "learning_rate": 4.524137931034483e-05,
+      "loss": 1.8231,
+      "step": 1206
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 4.5195402298850574e-05,
+      "loss": 2.1961,
+      "step": 1207
+    },
+    {
+      "epoch": 11.08,
+      "learning_rate": 4.514942528735632e-05,
+      "loss": 1.5721,
+      "step": 1208
+    },
+    {
+      "epoch": 11.09,
+      "learning_rate": 4.510344827586207e-05,
+      "loss": 1.7613,
+      "step": 1209
+    },
+    {
+      "epoch": 11.1,
+      "learning_rate": 4.505747126436782e-05,
+      "loss": 1.5303,
+      "step": 1210
+    },
+    {
+      "epoch": 11.11,
+      "learning_rate": 4.5011494252873566e-05,
+      "loss": 2.0698,
+      "step": 1211
+    },
+    {
+      "epoch": 11.12,
+      "learning_rate": 4.496551724137931e-05,
+      "loss": 1.6326,
+      "step": 1212
+    },
+    {
+      "epoch": 11.13,
+      "learning_rate": 4.491954022988506e-05,
+      "loss": 1.5959,
+      "step": 1213
+    },
+    {
+      "epoch": 11.14,
+      "learning_rate": 4.4873563218390805e-05,
+      "loss": 1.5667,
+      "step": 1214
+    },
+    {
+      "epoch": 11.15,
+      "learning_rate": 4.482758620689655e-05,
+      "loss": 1.5443,
+      "step": 1215
+    },
+    {
+      "epoch": 11.16,
+      "learning_rate": 4.4781609195402305e-05,
+      "loss": 1.9723,
+      "step": 1216
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 4.473563218390805e-05,
+      "loss": 1.9295,
+      "step": 1217
+    },
+    {
+      "epoch": 11.17,
+      "learning_rate": 4.46896551724138e-05,
+      "loss": 1.7935,
+      "step": 1218
+    },
+    {
+      "epoch": 11.18,
+      "learning_rate": 4.4643678160919544e-05,
+      "loss": 2.0555,
+      "step": 1219
+    },
+    {
+      "epoch": 11.19,
+      "learning_rate": 4.459770114942529e-05,
+      "loss": 1.5651,
+      "step": 1220
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 4.4551724137931036e-05,
+      "loss": 1.4282,
+      "step": 1221
+    },
+    {
+      "epoch": 11.21,
+      "learning_rate": 4.450574712643679e-05,
+      "loss": 2.3602,
+      "step": 1222
+    },
+    {
+      "epoch": 11.22,
+      "learning_rate": 4.445977011494253e-05,
+      "loss": 1.6559,
+      "step": 1223
+    },
+    {
+      "epoch": 11.23,
+      "learning_rate": 4.4413793103448275e-05,
+      "loss": 2.0973,
+      "step": 1224
+    },
+    {
+      "epoch": 11.24,
+      "learning_rate": 4.436781609195402e-05,
+      "loss": 1.9338,
+      "step": 1225
+    },
+    {
+      "epoch": 11.25,
+      "learning_rate": 4.432183908045977e-05,
+      "loss": 2.0715,
+      "step": 1226
+    },
+    {
+      "epoch": 11.26,
+      "learning_rate": 4.427586206896552e-05,
+      "loss": 1.988,
+      "step": 1227
+    },
+    {
+      "epoch": 11.27,
+      "learning_rate": 4.422988505747127e-05,
+      "loss": 1.7847,
+      "step": 1228
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 4.4183908045977014e-05,
+      "loss": 1.5549,
+      "step": 1229
+    },
+    {
+      "epoch": 11.28,
+      "learning_rate": 4.413793103448276e-05,
+      "loss": 2.0698,
+      "step": 1230
+    },
+    {
+      "epoch": 11.29,
+      "learning_rate": 4.4091954022988506e-05,
+      "loss": 1.9976,
+      "step": 1231
+    },
+    {
+      "epoch": 11.3,
+      "learning_rate": 4.404597701149425e-05,
+      "loss": 1.8382,
+      "step": 1232
+    },
+    {
+      "epoch": 11.31,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 1.7964,
+      "step": 1233
+    },
+    {
+      "epoch": 11.32,
+      "learning_rate": 4.395402298850575e-05,
+      "loss": 1.7887,
+      "step": 1234
+    },
+    {
+      "epoch": 11.33,
+      "learning_rate": 4.39080459770115e-05,
+      "loss": 1.6257,
+      "step": 1235
+    },
+    {
+      "epoch": 11.34,
+      "learning_rate": 4.3862068965517245e-05,
+      "loss": 1.8848,
+      "step": 1236
+    },
+    {
+      "epoch": 11.35,
+      "learning_rate": 4.381609195402299e-05,
+      "loss": 1.7344,
+      "step": 1237
+    },
+    {
+      "epoch": 11.36,
+      "learning_rate": 4.377011494252874e-05,
+      "loss": 1.9342,
+      "step": 1238
+    },
+    {
+      "epoch": 11.37,
+      "learning_rate": 4.3724137931034484e-05,
+      "loss": 1.8743,
+      "step": 1239
+    },
+    {
+      "epoch": 11.38,
+      "learning_rate": 4.367816091954024e-05,
+      "loss": 1.6978,
+      "step": 1240
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 4.363218390804598e-05,
+      "loss": 1.6164,
+      "step": 1241
+    },
+    {
+      "epoch": 11.39,
+      "learning_rate": 4.358620689655173e-05,
+      "loss": 1.6506,
+      "step": 1242
+    },
+    {
+      "epoch": 11.4,
+      "learning_rate": 4.354022988505747e-05,
+      "loss": 2.0403,
+      "step": 1243
+    },
+    {
+      "epoch": 11.41,
+      "learning_rate": 4.3494252873563215e-05,
+      "loss": 1.8511,
+      "step": 1244
+    },
+    {
+      "epoch": 11.42,
+      "learning_rate": 4.344827586206897e-05,
+      "loss": 1.5176,
+      "step": 1245
+    },
+    {
+      "epoch": 11.43,
+      "learning_rate": 4.3402298850574715e-05,
+      "loss": 1.6011,
+      "step": 1246
+    },
+    {
+      "epoch": 11.44,
+      "learning_rate": 4.335632183908046e-05,
+      "loss": 1.6582,
+      "step": 1247
+    },
+    {
+      "epoch": 11.45,
+      "learning_rate": 4.331034482758621e-05,
+      "loss": 1.8379,
+      "step": 1248
+    },
+    {
+      "epoch": 11.46,
+      "learning_rate": 4.3264367816091954e-05,
+      "loss": 1.6789,
+      "step": 1249
+    },
+    {
+      "epoch": 11.47,
+      "learning_rate": 4.32183908045977e-05,
+      "loss": 1.9702,
+      "step": 1250
+    },
+    {
+      "epoch": 11.48,
+      "learning_rate": 4.317241379310345e-05,
+      "loss": 1.8424,
+      "step": 1251
+    },
+    {
+      "epoch": 11.49,
+      "learning_rate": 4.31264367816092e-05,
+      "loss": 1.6566,
+      "step": 1252
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 4.3080459770114946e-05,
+      "loss": 2.2305,
+      "step": 1253
+    },
+    {
+      "epoch": 11.5,
+      "learning_rate": 4.303448275862069e-05,
+      "loss": 1.9766,
+      "step": 1254
+    },
+    {
+      "epoch": 11.51,
+      "learning_rate": 4.298850574712644e-05,
+      "loss": 2.3747,
+      "step": 1255
+    },
+    {
+      "epoch": 11.52,
+      "learning_rate": 4.2942528735632185e-05,
+      "loss": 1.5701,
+      "step": 1256
+    },
+    {
+      "epoch": 11.53,
+      "learning_rate": 4.289655172413794e-05,
+      "loss": 2.0265,
+      "step": 1257
+    },
+    {
+      "epoch": 11.54,
+      "learning_rate": 4.2850574712643684e-05,
+      "loss": 1.7035,
+      "step": 1258
+    },
+    {
+      "epoch": 11.55,
+      "learning_rate": 4.280459770114943e-05,
+      "loss": 1.8665,
+      "step": 1259
+    },
+    {
+      "epoch": 11.56,
+      "learning_rate": 4.275862068965518e-05,
+      "loss": 1.7173,
+      "step": 1260
+    },
+    {
+      "epoch": 11.57,
+      "learning_rate": 4.271264367816092e-05,
+      "loss": 1.8296,
+      "step": 1261
+    },
+    {
+      "epoch": 11.58,
+      "learning_rate": 4.266666666666667e-05,
+      "loss": 2.0926,
+      "step": 1262
+    },
+    {
+      "epoch": 11.59,
+      "learning_rate": 4.2620689655172416e-05,
+      "loss": 1.4948,
+      "step": 1263
+    },
+    {
+      "epoch": 11.6,
+      "learning_rate": 4.257471264367816e-05,
+      "loss": 1.8761,
+      "step": 1264
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 4.252873563218391e-05,
+      "loss": 1.6701,
+      "step": 1265
+    },
+    {
+      "epoch": 11.61,
+      "learning_rate": 4.2482758620689655e-05,
+      "loss": 1.7051,
+      "step": 1266
+    },
+    {
+      "epoch": 11.62,
+      "learning_rate": 4.24367816091954e-05,
+      "loss": 1.6675,
+      "step": 1267
+    },
+    {
+      "epoch": 11.63,
+      "learning_rate": 4.239080459770115e-05,
+      "loss": 1.6203,
+      "step": 1268
+    },
+    {
+      "epoch": 11.64,
+      "learning_rate": 4.23448275862069e-05,
+      "loss": 2.0101,
+      "step": 1269
+    },
+    {
+      "epoch": 11.65,
+      "learning_rate": 4.2298850574712647e-05,
+      "loss": 1.6746,
+      "step": 1270
+    },
+    {
+      "epoch": 11.66,
+      "learning_rate": 4.225287356321839e-05,
+      "loss": 2.7695,
+      "step": 1271
+    },
+    {
+      "epoch": 11.67,
+      "learning_rate": 4.220689655172414e-05,
+      "loss": 1.5369,
+      "step": 1272
+    },
+    {
+      "epoch": 11.68,
+      "learning_rate": 4.2160919540229886e-05,
+      "loss": 2.1051,
+      "step": 1273
+    },
+    {
+      "epoch": 11.69,
+      "learning_rate": 4.211494252873563e-05,
+      "loss": 2.0049,
+      "step": 1274
+    },
+    {
+      "epoch": 11.7,
+      "learning_rate": 4.2068965517241385e-05,
+      "loss": 1.8354,
+      "step": 1275
+    },
+    {
+      "epoch": 11.71,
+      "learning_rate": 4.202298850574713e-05,
+      "loss": 1.8571,
+      "step": 1276
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 4.197701149425288e-05,
+      "loss": 2.2498,
+      "step": 1277
+    },
+    {
+      "epoch": 11.72,
+      "learning_rate": 4.1931034482758624e-05,
+      "loss": 1.6582,
+      "step": 1278
+    },
+    {
+      "epoch": 11.73,
+      "learning_rate": 4.188505747126437e-05,
+      "loss": 1.6263,
+      "step": 1279
+    },
+    {
+      "epoch": 11.74,
+      "learning_rate": 4.1839080459770117e-05,
+      "loss": 1.6029,
+      "step": 1280
+    },
+    {
+      "epoch": 11.75,
+      "learning_rate": 4.179310344827587e-05,
+      "loss": 2.0412,
+      "step": 1281
+    },
+    {
+      "epoch": 11.76,
+      "learning_rate": 4.1747126436781616e-05,
+      "loss": 1.5327,
+      "step": 1282
+    },
+    {
+      "epoch": 11.77,
+      "learning_rate": 4.1701149425287356e-05,
+      "loss": 1.8146,
+      "step": 1283
+    },
+    {
+      "epoch": 11.78,
+      "learning_rate": 4.16551724137931e-05,
+      "loss": 1.6007,
+      "step": 1284
+    },
+    {
+      "epoch": 11.79,
+      "learning_rate": 4.160919540229885e-05,
+      "loss": 1.6982,
+      "step": 1285
+    },
+    {
+      "epoch": 11.8,
+      "learning_rate": 4.15632183908046e-05,
+      "loss": 1.6987,
+      "step": 1286
+    },
+    {
+      "epoch": 11.81,
+      "learning_rate": 4.151724137931035e-05,
+      "loss": 2.2549,
+      "step": 1287
+    },
+    {
+      "epoch": 11.82,
+      "learning_rate": 4.1471264367816094e-05,
+      "loss": 1.7121,
+      "step": 1288
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 4.142528735632184e-05,
+      "loss": 2.1203,
+      "step": 1289
+    },
+    {
+      "epoch": 11.83,
+      "learning_rate": 4.1379310344827587e-05,
+      "loss": 2.1497,
+      "step": 1290
+    },
+    {
+      "epoch": 11.84,
+      "learning_rate": 4.133333333333333e-05,
+      "loss": 2.3716,
+      "step": 1291
+    },
+    {
+      "epoch": 11.85,
+      "learning_rate": 4.128735632183908e-05,
+      "loss": 1.4795,
+      "step": 1292
+    },
+    {
+      "epoch": 11.86,
+      "learning_rate": 4.124137931034483e-05,
+      "loss": 1.8084,
+      "step": 1293
+    },
+    {
+      "epoch": 11.87,
+      "learning_rate": 4.119540229885058e-05,
+      "loss": 1.8711,
+      "step": 1294
+    },
+    {
+      "epoch": 11.88,
+      "learning_rate": 4.1149425287356325e-05,
+      "loss": 1.7199,
+      "step": 1295
+    },
+    {
+      "epoch": 11.89,
+      "learning_rate": 4.110344827586207e-05,
+      "loss": 1.4984,
+      "step": 1296
+    },
+    {
+      "epoch": 11.9,
+      "learning_rate": 4.105747126436782e-05,
+      "loss": 1.8105,
+      "step": 1297
+    },
+    {
+      "epoch": 11.91,
+      "learning_rate": 4.1011494252873564e-05,
+      "loss": 1.96,
+      "step": 1298
+    },
+    {
+      "epoch": 11.92,
+      "learning_rate": 4.096551724137932e-05,
+      "loss": 1.7435,
+      "step": 1299
+    },
+    {
+      "epoch": 11.93,
+      "learning_rate": 4.091954022988506e-05,
+      "loss": 1.5999,
+      "step": 1300
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 4.087356321839081e-05,
+      "loss": 1.6416,
+      "step": 1301
+    },
+    {
+      "epoch": 11.94,
+      "learning_rate": 4.0827586206896556e-05,
+      "loss": 1.5633,
+      "step": 1302
+    },
+    {
+      "epoch": 11.95,
+      "learning_rate": 4.0781609195402295e-05,
+      "loss": 1.5324,
+      "step": 1303
+    },
+    {
+      "epoch": 11.96,
+      "learning_rate": 4.073563218390805e-05,
+      "loss": 1.7332,
+      "step": 1304
+    },
+    {
+      "epoch": 11.97,
+      "learning_rate": 4.0689655172413795e-05,
+      "loss": 2.6089,
+      "step": 1305
+    },
+    {
+      "epoch": 11.98,
+      "learning_rate": 4.064367816091954e-05,
+      "loss": 2.0877,
+      "step": 1306
+    },
+    {
+      "epoch": 11.99,
+      "learning_rate": 4.059770114942529e-05,
+      "loss": 1.6229,
+      "step": 1307
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 4.0551724137931034e-05,
+      "loss": 1.8412,
+      "step": 1308
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.2073170731707317,
+      "eval_loss": 1.8808236122131348,
+      "eval_runtime": 13.7567,
+      "eval_samples_per_second": 11.921,
+      "eval_steps_per_second": 1.527,
+      "step": 1308
+    },
+    {
+      "epoch": 12.01,
+      "learning_rate": 4.050574712643678e-05,
+      "loss": 1.6997,
+      "step": 1309
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 4.045977011494253e-05,
+      "loss": 1.8682,
+      "step": 1310
+    },
+    {
+      "epoch": 12.03,
+      "learning_rate": 4.041379310344828e-05,
+      "loss": 2.7541,
+      "step": 1311
+    },
+    {
+      "epoch": 12.04,
+      "learning_rate": 4.0367816091954026e-05,
+      "loss": 1.7106,
+      "step": 1312
+    },
+    {
+      "epoch": 12.05,
+      "learning_rate": 4.032183908045977e-05,
+      "loss": 2.2801,
+      "step": 1313
+    },
+    {
+      "epoch": 12.06,
+      "learning_rate": 4.027586206896552e-05,
+      "loss": 1.6076,
+      "step": 1314
+    },
+    {
+      "epoch": 12.06,
+      "learning_rate": 4.0229885057471265e-05,
+      "loss": 1.9743,
+      "step": 1315
+    },
+    {
+      "epoch": 12.07,
+      "learning_rate": 4.018390804597701e-05,
+      "loss": 1.6813,
+      "step": 1316
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 4.0137931034482764e-05,
+      "loss": 1.4762,
+      "step": 1317
+    },
+    {
+      "epoch": 12.09,
+      "learning_rate": 4.009195402298851e-05,
+      "loss": 1.5628,
+      "step": 1318
+    },
+    {
+      "epoch": 12.1,
+      "learning_rate": 4.004597701149426e-05,
+      "loss": 1.7441,
+      "step": 1319
+    },
+    {
+      "epoch": 12.11,
+      "learning_rate": 4e-05,
+      "loss": 2.4427,
+      "step": 1320
+    },
+    {
+      "epoch": 12.12,
+      "learning_rate": 3.995402298850575e-05,
+      "loss": 1.8439,
+      "step": 1321
+    },
+    {
+      "epoch": 12.13,
+      "learning_rate": 3.9908045977011496e-05,
+      "loss": 1.7676,
+      "step": 1322
+    },
+    {
+      "epoch": 12.14,
+      "learning_rate": 3.986206896551724e-05,
+      "loss": 1.876,
+      "step": 1323
+    },
+    {
+      "epoch": 12.15,
+      "learning_rate": 3.981609195402299e-05,
+      "loss": 1.8944,
+      "step": 1324
+    },
+    {
+      "epoch": 12.16,
+      "learning_rate": 3.9770114942528735e-05,
+      "loss": 1.6557,
+      "step": 1325
+    },
+    {
+      "epoch": 12.17,
+      "learning_rate": 3.972413793103448e-05,
+      "loss": 1.7728,
+      "step": 1326
+    },
+    {
+      "epoch": 12.17,
+      "learning_rate": 3.967816091954023e-05,
+      "loss": 1.9071,
+      "step": 1327
+    },
+    {
+      "epoch": 12.18,
+      "learning_rate": 3.963218390804598e-05,
+      "loss": 2.0858,
+      "step": 1328
+    },
+    {
+      "epoch": 12.19,
+      "learning_rate": 3.958620689655173e-05,
+      "loss": 1.6608,
+      "step": 1329
+    },
+    {
+      "epoch": 12.2,
+      "learning_rate": 3.954022988505747e-05,
+      "loss": 1.8184,
+      "step": 1330
+    },
+    {
+      "epoch": 12.21,
+      "learning_rate": 3.949425287356322e-05,
+      "loss": 1.931,
+      "step": 1331
+    },
+    {
+      "epoch": 12.22,
+      "learning_rate": 3.9448275862068966e-05,
+      "loss": 1.8122,
+      "step": 1332
+    },
+    {
+      "epoch": 12.23,
+      "learning_rate": 3.940229885057471e-05,
+      "loss": 1.6846,
+      "step": 1333
+    },
+    {
+      "epoch": 12.24,
+      "learning_rate": 3.9356321839080465e-05,
+      "loss": 1.9634,
+      "step": 1334
+    },
+    {
+      "epoch": 12.25,
+      "learning_rate": 3.931034482758621e-05,
+      "loss": 1.6266,
+      "step": 1335
+    },
+    {
+      "epoch": 12.26,
+      "learning_rate": 3.926436781609196e-05,
+      "loss": 2.2756,
+      "step": 1336
+    },
+    {
+      "epoch": 12.27,
+      "learning_rate": 3.9218390804597704e-05,
+      "loss": 1.5933,
+      "step": 1337
+    },
+    {
+      "epoch": 12.28,
+      "learning_rate": 3.917241379310345e-05,
+      "loss": 1.8022,
+      "step": 1338
+    },
+    {
+      "epoch": 12.28,
+      "learning_rate": 3.91264367816092e-05,
+      "loss": 1.6755,
+      "step": 1339
+    },
+    {
+      "epoch": 12.29,
+      "learning_rate": 3.908045977011495e-05,
+      "loss": 2.2314,
+      "step": 1340
+    },
+    {
+      "epoch": 12.3,
+      "learning_rate": 3.9034482758620696e-05,
+      "loss": 2.1081,
+      "step": 1341
+    },
+    {
+      "epoch": 12.31,
+      "learning_rate": 3.8988505747126436e-05,
+      "loss": 1.8835,
+      "step": 1342
+    },
+    {
+      "epoch": 12.32,
+      "learning_rate": 3.894252873563218e-05,
+      "loss": 2.1328,
+      "step": 1343
+    },
+    {
+      "epoch": 12.33,
+      "learning_rate": 3.889655172413793e-05,
+      "loss": 2.1217,
+      "step": 1344
+    },
+    {
+      "epoch": 12.34,
+      "learning_rate": 3.8850574712643675e-05,
+      "loss": 1.7944,
+      "step": 1345
+    },
+    {
+      "epoch": 12.35,
+      "learning_rate": 3.880459770114943e-05,
+      "loss": 1.5661,
+      "step": 1346
+    },
+    {
+      "epoch": 12.36,
+      "learning_rate": 3.8758620689655174e-05,
+      "loss": 1.6538,
+      "step": 1347
+    },
+    {
+      "epoch": 12.37,
+      "learning_rate": 3.871264367816092e-05,
+      "loss": 1.9019,
+      "step": 1348
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 3.866666666666667e-05,
+      "loss": 1.826,
+      "step": 1349
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.862068965517241e-05,
+      "loss": 1.7057,
+      "step": 1350
+    },
+    {
+      "epoch": 12.39,
+      "learning_rate": 3.857471264367816e-05,
+      "loss": 1.6004,
+      "step": 1351
+    },
+    {
+      "epoch": 12.4,
+      "learning_rate": 3.852873563218391e-05,
+      "loss": 2.0872,
+      "step": 1352
+    },
+    {
+      "epoch": 12.41,
+      "learning_rate": 3.848275862068966e-05,
+      "loss": 1.8605,
+      "step": 1353
+    },
+    {
+      "epoch": 12.42,
+      "learning_rate": 3.8436781609195405e-05,
+      "loss": 1.8164,
+      "step": 1354
+    },
+    {
+      "epoch": 12.43,
+      "learning_rate": 3.839080459770115e-05,
+      "loss": 1.6991,
+      "step": 1355
+    },
+    {
+      "epoch": 12.44,
+      "learning_rate": 3.83448275862069e-05,
+      "loss": 1.604,
+      "step": 1356
+    },
+    {
+      "epoch": 12.45,
+      "learning_rate": 3.8298850574712644e-05,
+      "loss": 2.5076,
+      "step": 1357
+    },
+    {
+      "epoch": 12.46,
+      "learning_rate": 3.82528735632184e-05,
+      "loss": 1.7762,
+      "step": 1358
+    },
+    {
+      "epoch": 12.47,
+      "learning_rate": 3.8206896551724144e-05,
+      "loss": 2.0272,
+      "step": 1359
+    },
+    {
+      "epoch": 12.48,
+      "learning_rate": 3.816091954022989e-05,
+      "loss": 1.6501,
+      "step": 1360
+    },
+    {
+      "epoch": 12.49,
+      "learning_rate": 3.8114942528735636e-05,
+      "loss": 1.6751,
+      "step": 1361
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 3.8068965517241376e-05,
+      "loss": 2.2643,
+      "step": 1362
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 3.802298850574713e-05,
+      "loss": 2.0036,
+      "step": 1363
+    },
+    {
+      "epoch": 12.51,
+      "learning_rate": 3.7977011494252875e-05,
+      "loss": 1.5226,
+      "step": 1364
+    },
+    {
+      "epoch": 12.52,
+      "learning_rate": 3.793103448275862e-05,
+      "loss": 1.6284,
+      "step": 1365
+    },
+    {
+      "epoch": 12.53,
+      "learning_rate": 3.788505747126437e-05,
+      "loss": 1.5311,
+      "step": 1366
+    },
+    {
+      "epoch": 12.54,
+      "learning_rate": 3.7839080459770114e-05,
+      "loss": 1.7964,
+      "step": 1367
+    },
+    {
+      "epoch": 12.55,
+      "learning_rate": 3.779310344827586e-05,
+      "loss": 2.0977,
+      "step": 1368
+    },
+    {
+      "epoch": 12.56,
+      "learning_rate": 3.774712643678161e-05,
+      "loss": 1.9036,
+      "step": 1369
+    },
+    {
+      "epoch": 12.57,
+      "learning_rate": 3.770114942528736e-05,
+      "loss": 1.742,
+      "step": 1370
+    },
+    {
+      "epoch": 12.58,
+      "learning_rate": 3.7655172413793106e-05,
+      "loss": 2.2972,
+      "step": 1371
+    },
+    {
+      "epoch": 12.59,
+      "learning_rate": 3.760919540229885e-05,
+      "loss": 1.6639,
+      "step": 1372
+    },
+    {
+      "epoch": 12.6,
+      "learning_rate": 3.75632183908046e-05,
+      "loss": 1.5106,
+      "step": 1373
+    },
+    {
+      "epoch": 12.61,
+      "learning_rate": 3.7517241379310345e-05,
+      "loss": 1.6445,
+      "step": 1374
+    },
+    {
+      "epoch": 12.61,
+      "learning_rate": 3.747126436781609e-05,
+      "loss": 1.7145,
+      "step": 1375
+    },
+    {
+      "epoch": 12.62,
+      "learning_rate": 3.7425287356321845e-05,
+      "loss": 1.4827,
+      "step": 1376
+    },
+    {
+      "epoch": 12.63,
+      "learning_rate": 3.737931034482759e-05,
+      "loss": 1.585,
+      "step": 1377
+    },
+    {
+      "epoch": 12.64,
+      "learning_rate": 3.733333333333334e-05,
+      "loss": 2.2886,
+      "step": 1378
+    },
+    {
+      "epoch": 12.65,
+      "learning_rate": 3.7287356321839084e-05,
+      "loss": 1.6084,
+      "step": 1379
+    },
+    {
+      "epoch": 12.66,
+      "learning_rate": 3.724137931034483e-05,
+      "loss": 2.1146,
+      "step": 1380
+    },
+    {
+      "epoch": 12.67,
+      "learning_rate": 3.7195402298850576e-05,
+      "loss": 1.7863,
+      "step": 1381
+    },
+    {
+      "epoch": 12.68,
+      "learning_rate": 3.714942528735632e-05,
+      "loss": 2.0809,
+      "step": 1382
+    },
+    {
+      "epoch": 12.69,
+      "learning_rate": 3.710344827586207e-05,
+      "loss": 2.2632,
+      "step": 1383
+    },
+    {
+      "epoch": 12.7,
+      "learning_rate": 3.7057471264367815e-05,
+      "loss": 1.4661,
+      "step": 1384
+    },
+    {
+      "epoch": 12.71,
+      "learning_rate": 3.701149425287356e-05,
+      "loss": 1.5457,
+      "step": 1385
+    },
+    {
+      "epoch": 12.72,
+      "learning_rate": 3.696551724137931e-05,
+      "loss": 1.5627,
+      "step": 1386
+    },
+    {
+      "epoch": 12.72,
+      "learning_rate": 3.691954022988506e-05,
+      "loss": 2.4417,
+      "step": 1387
+    },
+    {
+      "epoch": 12.73,
+      "learning_rate": 3.687356321839081e-05,
+      "loss": 2.0254,
+      "step": 1388
+    },
+    {
+      "epoch": 12.74,
+      "learning_rate": 3.6827586206896554e-05,
+      "loss": 1.7257,
+      "step": 1389
+    },
+    {
+      "epoch": 12.75,
+      "learning_rate": 3.67816091954023e-05,
+      "loss": 1.9139,
+      "step": 1390
+    },
+    {
+      "epoch": 12.76,
+      "learning_rate": 3.6735632183908046e-05,
+      "loss": 1.6035,
+      "step": 1391
+    },
+    {
+      "epoch": 12.77,
+      "learning_rate": 3.668965517241379e-05,
+      "loss": 2.0908,
+      "step": 1392
+    },
+    {
+      "epoch": 12.78,
+      "learning_rate": 3.6643678160919546e-05,
+      "loss": 1.943,
+      "step": 1393
+    },
+    {
+      "epoch": 12.79,
+      "learning_rate": 3.659770114942529e-05,
+      "loss": 1.6842,
+      "step": 1394
+    },
+    {
+      "epoch": 12.8,
+      "learning_rate": 3.655172413793104e-05,
+      "loss": 1.9285,
+      "step": 1395
+    },
+    {
+      "epoch": 12.81,
+      "learning_rate": 3.6505747126436785e-05,
+      "loss": 1.418,
+      "step": 1396
+    },
+    {
+      "epoch": 12.82,
+      "learning_rate": 3.645977011494253e-05,
+      "loss": 1.9443,
+      "step": 1397
+    },
+    {
+      "epoch": 12.83,
+      "learning_rate": 3.641379310344828e-05,
+      "loss": 2.0679,
+      "step": 1398
+    },
+    {
+      "epoch": 12.83,
+      "learning_rate": 3.6367816091954024e-05,
+      "loss": 1.5817,
+      "step": 1399
+    },
+    {
+      "epoch": 12.84,
+      "learning_rate": 3.632183908045978e-05,
+      "loss": 1.7067,
+      "step": 1400
+    },
+    {
+      "epoch": 12.85,
+      "learning_rate": 3.627586206896552e-05,
+      "loss": 1.9093,
+      "step": 1401
+    },
+    {
+      "epoch": 12.86,
+      "learning_rate": 3.622988505747126e-05,
+      "loss": 1.6146,
+      "step": 1402
+    },
+    {
+      "epoch": 12.87,
+      "learning_rate": 3.618390804597701e-05,
+      "loss": 1.7603,
+      "step": 1403
+    },
+    {
+      "epoch": 12.88,
+      "learning_rate": 3.6137931034482755e-05,
+      "loss": 1.9061,
+      "step": 1404
+    },
+    {
+      "epoch": 12.89,
+      "learning_rate": 3.609195402298851e-05,
+      "loss": 1.6337,
+      "step": 1405
+    },
+    {
+      "epoch": 12.9,
+      "learning_rate": 3.6045977011494255e-05,
+      "loss": 1.7829,
+      "step": 1406
+    },
+    {
+      "epoch": 12.91,
+      "learning_rate": 3.6e-05,
+      "loss": 1.6992,
+      "step": 1407
+    },
+    {
+      "epoch": 12.92,
+      "learning_rate": 3.595402298850575e-05,
+      "loss": 1.6491,
+      "step": 1408
+    },
+    {
+      "epoch": 12.93,
+      "learning_rate": 3.5908045977011494e-05,
+      "loss": 2.2817,
+      "step": 1409
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.586206896551724e-05,
+      "loss": 1.6289,
+      "step": 1410
+    },
+    {
+      "epoch": 12.94,
+      "learning_rate": 3.581609195402299e-05,
+      "loss": 1.4462,
+      "step": 1411
+    },
+    {
+      "epoch": 12.95,
+      "learning_rate": 3.577011494252874e-05,
+      "loss": 1.6452,
+      "step": 1412
+    },
+    {
+      "epoch": 12.96,
+      "learning_rate": 3.5724137931034486e-05,
+      "loss": 2.1898,
+      "step": 1413
+    },
+    {
+      "epoch": 12.97,
+      "learning_rate": 3.567816091954023e-05,
+      "loss": 1.5272,
+      "step": 1414
+    },
+    {
+      "epoch": 12.98,
+      "learning_rate": 3.563218390804598e-05,
+      "loss": 1.6938,
+      "step": 1415
+    },
+    {
+      "epoch": 12.99,
+      "learning_rate": 3.5586206896551725e-05,
+      "loss": 1.8053,
+      "step": 1416
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 3.554022988505748e-05,
+      "loss": 2.4396,
+      "step": 1417
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.20121951219512196,
+      "eval_loss": 1.8816215991973877,
+      "eval_runtime": 13.7857,
+      "eval_samples_per_second": 11.896,
+      "eval_steps_per_second": 1.523,
+      "step": 1417
+    },
+    {
+      "epoch": 13.01,
+      "learning_rate": 3.5494252873563224e-05,
+      "loss": 1.5524,
+      "step": 1418
+    },
+    {
+      "epoch": 13.02,
+      "learning_rate": 3.544827586206897e-05,
+      "loss": 2.1268,
+      "step": 1419
+    },
+    {
+      "epoch": 13.03,
+      "learning_rate": 3.540229885057472e-05,
+      "loss": 1.5194,
+      "step": 1420
+    },
+    {
+      "epoch": 13.04,
+      "learning_rate": 3.535632183908046e-05,
+      "loss": 1.7894,
+      "step": 1421
+    },
+    {
+      "epoch": 13.05,
+      "learning_rate": 3.53103448275862e-05,
+      "loss": 1.6634,
+      "step": 1422
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.5264367816091956e-05,
+      "loss": 1.8446,
+      "step": 1423
+    },
+    {
+      "epoch": 13.06,
+      "learning_rate": 3.52183908045977e-05,
+      "loss": 1.5461,
+      "step": 1424
+    },
+    {
+      "epoch": 13.07,
+      "learning_rate": 3.517241379310345e-05,
+      "loss": 1.7526,
+      "step": 1425
+    },
+    {
+      "epoch": 13.08,
+      "learning_rate": 3.5126436781609195e-05,
+      "loss": 1.721,
+      "step": 1426
+    },
+    {
+      "epoch": 13.09,
+      "learning_rate": 3.508045977011494e-05,
+      "loss": 1.7843,
+      "step": 1427
+    },
+    {
+      "epoch": 13.1,
+      "learning_rate": 3.503448275862069e-05,
+      "loss": 2.3265,
+      "step": 1428
+    },
+    {
+      "epoch": 13.11,
+      "learning_rate": 3.498850574712644e-05,
+      "loss": 1.6883,
+      "step": 1429
+    },
+    {
+      "epoch": 13.12,
+      "learning_rate": 3.4942528735632187e-05,
+      "loss": 2.46,
+      "step": 1430
+    },
+    {
+      "epoch": 13.13,
+      "learning_rate": 3.489655172413793e-05,
+      "loss": 2.0363,
+      "step": 1431
+    },
+    {
+      "epoch": 13.14,
+      "learning_rate": 3.485057471264368e-05,
+      "loss": 1.8346,
+      "step": 1432
+    },
+    {
+      "epoch": 13.15,
+      "learning_rate": 3.4804597701149426e-05,
+      "loss": 1.7017,
+      "step": 1433
+    },
+    {
+      "epoch": 13.16,
+      "learning_rate": 3.475862068965517e-05,
+      "loss": 1.6763,
+      "step": 1434
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 3.4712643678160925e-05,
+      "loss": 1.6313,
+      "step": 1435
+    },
+    {
+      "epoch": 13.17,
+      "learning_rate": 3.466666666666667e-05,
+      "loss": 1.7451,
+      "step": 1436
+    },
+    {
+      "epoch": 13.18,
+      "learning_rate": 3.462068965517242e-05,
+      "loss": 2.2183,
+      "step": 1437
+    },
+    {
+      "epoch": 13.19,
+      "learning_rate": 3.4574712643678164e-05,
+      "loss": 1.6877,
+      "step": 1438
+    },
+    {
+      "epoch": 13.2,
+      "learning_rate": 3.452873563218391e-05,
+      "loss": 1.8016,
+      "step": 1439
+    },
+    {
+      "epoch": 13.21,
+      "learning_rate": 3.4482758620689657e-05,
+      "loss": 1.8276,
+      "step": 1440
+    },
+    {
+      "epoch": 13.22,
+      "learning_rate": 3.443678160919541e-05,
+      "loss": 1.8577,
+      "step": 1441
+    },
+    {
+      "epoch": 13.23,
+      "learning_rate": 3.439080459770115e-05,
+      "loss": 1.5483,
+      "step": 1442
+    },
+    {
+      "epoch": 13.24,
+      "learning_rate": 3.4344827586206896e-05,
+      "loss": 2.0301,
+      "step": 1443
+    },
+    {
+      "epoch": 13.25,
+      "learning_rate": 3.429885057471264e-05,
+      "loss": 2.16,
+      "step": 1444
+    },
+    {
+      "epoch": 13.26,
+      "learning_rate": 3.425287356321839e-05,
+      "loss": 1.7,
+      "step": 1445
+    },
+    {
+      "epoch": 13.27,
+      "learning_rate": 3.420689655172414e-05,
+      "loss": 2.2327,
+      "step": 1446
+    },
+    {
+      "epoch": 13.28,
+      "learning_rate": 3.416091954022989e-05,
+      "loss": 1.6562,
+      "step": 1447
+    },
+    {
+      "epoch": 13.28,
+      "learning_rate": 3.4114942528735634e-05,
+      "loss": 1.6733,
+      "step": 1448
+    },
+    {
+      "epoch": 13.29,
+      "learning_rate": 3.406896551724138e-05,
+      "loss": 1.6128,
+      "step": 1449
+    },
+    {
+      "epoch": 13.3,
+      "learning_rate": 3.4022988505747127e-05,
+      "loss": 1.9048,
+      "step": 1450
+    },
+    {
+      "epoch": 13.31,
+      "learning_rate": 3.397701149425287e-05,
+      "loss": 2.404,
+      "step": 1451
+    },
+    {
+      "epoch": 13.32,
+      "learning_rate": 3.393103448275862e-05,
+      "loss": 1.7293,
+      "step": 1452
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 3.388505747126437e-05,
+      "loss": 1.7227,
+      "step": 1453
+    },
+    {
+      "epoch": 13.34,
+      "learning_rate": 3.383908045977012e-05,
+      "loss": 1.7477,
+      "step": 1454
+    },
+    {
+      "epoch": 13.35,
+      "learning_rate": 3.3793103448275865e-05,
+      "loss": 1.8203,
+      "step": 1455
+    },
+    {
+      "epoch": 13.36,
+      "learning_rate": 3.374712643678161e-05,
+      "loss": 1.5614,
+      "step": 1456
+    },
+    {
+      "epoch": 13.37,
+      "learning_rate": 3.370114942528736e-05,
+      "loss": 1.6593,
+      "step": 1457
+    },
+    {
+      "epoch": 13.38,
+      "learning_rate": 3.3655172413793104e-05,
+      "loss": 2.0024,
+      "step": 1458
+    },
+    {
+      "epoch": 13.39,
+      "learning_rate": 3.360919540229886e-05,
+      "loss": 1.7272,
+      "step": 1459
+    },
+    {
+      "epoch": 13.39,
+      "learning_rate": 3.35632183908046e-05,
+      "loss": 1.5952,
+      "step": 1460
+    },
+    {
+      "epoch": 13.4,
+      "learning_rate": 3.351724137931035e-05,
+      "loss": 2.1549,
+      "step": 1461
+    },
+    {
+      "epoch": 13.41,
+      "learning_rate": 3.347126436781609e-05,
+      "loss": 1.6263,
+      "step": 1462
+    },
+    {
+      "epoch": 13.42,
+      "learning_rate": 3.3425287356321836e-05,
+      "loss": 1.7546,
+      "step": 1463
+    },
+    {
+      "epoch": 13.43,
+      "learning_rate": 3.337931034482759e-05,
+      "loss": 1.6455,
+      "step": 1464
+    },
+    {
+      "epoch": 13.44,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 1.4205,
+      "step": 1465
+    },
+    {
+      "epoch": 13.45,
+      "learning_rate": 3.328735632183908e-05,
+      "loss": 1.4761,
+      "step": 1466
+    },
+    {
+      "epoch": 13.46,
+      "learning_rate": 3.324137931034483e-05,
+      "loss": 1.5075,
+      "step": 1467
+    },
+    {
+      "epoch": 13.47,
+      "learning_rate": 3.3195402298850574e-05,
+      "loss": 1.8545,
+      "step": 1468
+    },
+    {
+      "epoch": 13.48,
+      "learning_rate": 3.314942528735632e-05,
+      "loss": 2.1361,
+      "step": 1469
+    },
+    {
+      "epoch": 13.49,
+      "learning_rate": 3.310344827586207e-05,
+      "loss": 2.0436,
+      "step": 1470
+    },
+    {
+      "epoch": 13.5,
+      "learning_rate": 3.305747126436782e-05,
+      "loss": 1.9092,
+      "step": 1471
+    },
+    {
+      "epoch": 13.5,
+      "learning_rate": 3.3011494252873566e-05,
+      "loss": 2.5329,
+      "step": 1472
+    },
+    {
+      "epoch": 13.51,
+      "learning_rate": 3.296551724137931e-05,
+      "loss": 1.5025,
+      "step": 1473
+    },
+    {
+      "epoch": 13.52,
+      "learning_rate": 3.291954022988506e-05,
+      "loss": 2.0181,
+      "step": 1474
+    },
+    {
+      "epoch": 13.53,
+      "learning_rate": 3.2873563218390805e-05,
+      "loss": 1.4078,
+      "step": 1475
+    },
+    {
+      "epoch": 13.54,
+      "learning_rate": 3.282758620689655e-05,
+      "loss": 1.759,
+      "step": 1476
+    },
+    {
+      "epoch": 13.55,
+      "learning_rate": 3.2781609195402304e-05,
+      "loss": 1.7231,
+      "step": 1477
+    },
+    {
+      "epoch": 13.56,
+      "learning_rate": 3.273563218390805e-05,
+      "loss": 2.1947,
+      "step": 1478
+    },
+    {
+      "epoch": 13.57,
+      "learning_rate": 3.26896551724138e-05,
+      "loss": 2.0677,
+      "step": 1479
+    },
+    {
+      "epoch": 13.58,
+      "learning_rate": 3.264367816091954e-05,
+      "loss": 2.6541,
+      "step": 1480
+    },
+    {
+      "epoch": 13.59,
+      "learning_rate": 3.259770114942529e-05,
+      "loss": 1.814,
+      "step": 1481
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 3.2551724137931036e-05,
+      "loss": 1.6043,
+      "step": 1482
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.250574712643678e-05,
+      "loss": 1.7979,
+      "step": 1483
+    },
+    {
+      "epoch": 13.61,
+      "learning_rate": 3.245977011494253e-05,
+      "loss": 1.6444,
+      "step": 1484
+    },
+    {
+      "epoch": 13.62,
+      "learning_rate": 3.2413793103448275e-05,
+      "loss": 2.0675,
+      "step": 1485
+    },
+    {
+      "epoch": 13.63,
+      "learning_rate": 3.236781609195402e-05,
+      "loss": 1.8717,
+      "step": 1486
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 3.232183908045977e-05,
+      "loss": 1.7233,
+      "step": 1487
+    },
+    {
+      "epoch": 13.65,
+      "learning_rate": 3.227586206896552e-05,
+      "loss": 1.7165,
+      "step": 1488
+    },
+    {
+      "epoch": 13.66,
+      "learning_rate": 3.222988505747127e-05,
+      "loss": 1.8049,
+      "step": 1489
+    },
+    {
+      "epoch": 13.67,
+      "learning_rate": 3.218390804597701e-05,
+      "loss": 2.2933,
+      "step": 1490
+    },
+    {
+      "epoch": 13.68,
+      "learning_rate": 3.213793103448276e-05,
+      "loss": 1.7821,
+      "step": 1491
+    },
+    {
+      "epoch": 13.69,
+      "learning_rate": 3.2091954022988506e-05,
+      "loss": 1.5636,
+      "step": 1492
+    },
+    {
+      "epoch": 13.7,
+      "learning_rate": 3.204597701149425e-05,
+      "loss": 1.715,
+      "step": 1493
+    },
+    {
+      "epoch": 13.71,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 1.5199,
+      "step": 1494
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 3.195402298850575e-05,
+      "loss": 1.9632,
+      "step": 1495
+    },
+    {
+      "epoch": 13.72,
+      "learning_rate": 3.19080459770115e-05,
+      "loss": 2.112,
+      "step": 1496
+    },
+    {
+      "epoch": 13.73,
+      "learning_rate": 3.1862068965517244e-05,
+      "loss": 2.389,
+      "step": 1497
+    },
+    {
+      "epoch": 13.74,
+      "learning_rate": 3.181609195402299e-05,
+      "loss": 1.4465,
+      "step": 1498
+    },
+    {
+      "epoch": 13.75,
+      "learning_rate": 3.177011494252874e-05,
+      "loss": 1.5308,
+      "step": 1499
+    },
+    {
+      "epoch": 13.76,
+      "learning_rate": 3.172413793103448e-05,
+      "loss": 1.521,
+      "step": 1500
+    },
+    {
+      "epoch": 13.77,
+      "learning_rate": 3.1678160919540236e-05,
+      "loss": 1.4801,
+      "step": 1501
+    },
+    {
+      "epoch": 13.78,
+      "learning_rate": 3.1632183908045976e-05,
+      "loss": 1.6025,
+      "step": 1502
+    },
+    {
+      "epoch": 13.79,
+      "learning_rate": 3.158620689655172e-05,
+      "loss": 1.4998,
+      "step": 1503
+    },
+    {
+      "epoch": 13.8,
+      "learning_rate": 3.154022988505747e-05,
+      "loss": 1.5724,
+      "step": 1504
+    },
+    {
+      "epoch": 13.81,
+      "learning_rate": 3.1494252873563215e-05,
+      "loss": 1.6698,
+      "step": 1505
+    },
+    {
+      "epoch": 13.82,
+      "learning_rate": 3.144827586206897e-05,
+      "loss": 2.0186,
+      "step": 1506
+    },
+    {
+      "epoch": 13.83,
+      "learning_rate": 3.1402298850574714e-05,
+      "loss": 2.1341,
+      "step": 1507
+    },
+    {
+      "epoch": 13.83,
+      "learning_rate": 3.135632183908046e-05,
+      "loss": 3.1515,
+      "step": 1508
+    },
+    {
+      "epoch": 13.84,
+      "learning_rate": 3.131034482758621e-05,
+      "loss": 2.0454,
+      "step": 1509
+    },
+    {
+      "epoch": 13.85,
+      "learning_rate": 3.126436781609195e-05,
+      "loss": 2.2539,
+      "step": 1510
+    },
+    {
+      "epoch": 13.86,
+      "learning_rate": 3.12183908045977e-05,
+      "loss": 1.6053,
+      "step": 1511
+    },
+    {
+      "epoch": 13.87,
+      "learning_rate": 3.117241379310345e-05,
+      "loss": 2.0915,
+      "step": 1512
+    },
+    {
+      "epoch": 13.88,
+      "learning_rate": 3.11264367816092e-05,
+      "loss": 1.8654,
+      "step": 1513
+    },
+    {
+      "epoch": 13.89,
+      "learning_rate": 3.1080459770114945e-05,
+      "loss": 1.9774,
+      "step": 1514
+    },
+    {
+      "epoch": 13.9,
+      "learning_rate": 3.103448275862069e-05,
+      "loss": 2.1024,
+      "step": 1515
+    },
+    {
+      "epoch": 13.91,
+      "learning_rate": 3.098850574712644e-05,
+      "loss": 1.5114,
+      "step": 1516
+    },
+    {
+      "epoch": 13.92,
+      "learning_rate": 3.0942528735632184e-05,
+      "loss": 2.2142,
+      "step": 1517
+    },
+    {
+      "epoch": 13.93,
+      "learning_rate": 3.089655172413794e-05,
+      "loss": 1.6768,
+      "step": 1518
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 3.0850574712643684e-05,
+      "loss": 2.1619,
+      "step": 1519
+    },
+    {
+      "epoch": 13.94,
+      "learning_rate": 3.080459770114943e-05,
+      "loss": 1.8649,
+      "step": 1520
+    },
+    {
+      "epoch": 13.95,
+      "learning_rate": 3.0758620689655176e-05,
+      "loss": 1.5361,
+      "step": 1521
+    },
+    {
+      "epoch": 13.96,
+      "learning_rate": 3.0712643678160916e-05,
+      "loss": 1.847,
+      "step": 1522
+    },
+    {
+      "epoch": 13.97,
+      "learning_rate": 3.066666666666667e-05,
+      "loss": 1.8208,
+      "step": 1523
+    },
+    {
+      "epoch": 13.98,
+      "learning_rate": 3.0620689655172415e-05,
+      "loss": 1.9235,
+      "step": 1524
+    },
+    {
+      "epoch": 13.99,
+      "learning_rate": 3.057471264367816e-05,
+      "loss": 1.7407,
+      "step": 1525
+    },
+    {
+      "epoch": 14.0,
+      "learning_rate": 3.052873563218391e-05,
+      "loss": 1.8643,
+      "step": 1526
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.20121951219512196,
+      "eval_loss": 1.8826814889907837,
+      "eval_runtime": 13.751,
+      "eval_samples_per_second": 11.926,
+      "eval_steps_per_second": 1.527,
+      "step": 1526
+    },
+    {
+      "epoch": 14.01,
+      "learning_rate": 3.0482758620689654e-05,
+      "loss": 1.9652,
+      "step": 1527
+    },
+    {
+      "epoch": 14.02,
+      "learning_rate": 3.0436781609195404e-05,
+      "loss": 1.4541,
+      "step": 1528
+    },
+    {
+      "epoch": 14.03,
+      "learning_rate": 3.039080459770115e-05,
+      "loss": 2.0109,
+      "step": 1529
+    },
+    {
+      "epoch": 14.04,
+      "learning_rate": 3.0344827586206897e-05,
+      "loss": 2.3494,
+      "step": 1530
+    },
+    {
+      "epoch": 14.05,
+      "learning_rate": 3.0298850574712646e-05,
+      "loss": 1.7708,
+      "step": 1531
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 3.0252873563218393e-05,
+      "loss": 2.221,
+      "step": 1532
+    },
+    {
+      "epoch": 14.06,
+      "learning_rate": 3.020689655172414e-05,
+      "loss": 1.5771,
+      "step": 1533
+    },
+    {
+      "epoch": 14.07,
+      "learning_rate": 3.016091954022989e-05,
+      "loss": 1.7401,
+      "step": 1534
+    },
+    {
+      "epoch": 14.08,
+      "learning_rate": 3.0114942528735635e-05,
+      "loss": 1.8612,
+      "step": 1535
+    },
+    {
+      "epoch": 14.09,
+      "learning_rate": 3.006896551724138e-05,
+      "loss": 1.8592,
+      "step": 1536
+    },
+    {
+      "epoch": 14.1,
+      "learning_rate": 3.002298850574713e-05,
+      "loss": 1.5413,
+      "step": 1537
+    },
+    {
+      "epoch": 14.11,
+      "learning_rate": 2.9977011494252877e-05,
+      "loss": 1.8957,
+      "step": 1538
+    },
+    {
+      "epoch": 14.12,
+      "learning_rate": 2.9931034482758624e-05,
+      "loss": 1.6045,
+      "step": 1539
+    },
+    {
+      "epoch": 14.13,
+      "learning_rate": 2.988505747126437e-05,
+      "loss": 2.7238,
+      "step": 1540
+    },
+    {
+      "epoch": 14.14,
+      "learning_rate": 2.983908045977012e-05,
+      "loss": 1.7129,
+      "step": 1541
+    },
+    {
+      "epoch": 14.15,
+      "learning_rate": 2.979310344827586e-05,
+      "loss": 2.2707,
+      "step": 1542
+    },
+    {
+      "epoch": 14.16,
+      "learning_rate": 2.974712643678161e-05,
+      "loss": 1.6042,
+      "step": 1543
+    },
+    {
+      "epoch": 14.17,
+      "learning_rate": 2.9701149425287355e-05,
+      "loss": 1.7389,
+      "step": 1544
+    },
+    {
+      "epoch": 14.17,
+      "learning_rate": 2.96551724137931e-05,
+      "loss": 1.5934,
+      "step": 1545
+    },
+    {
+      "epoch": 14.18,
+      "learning_rate": 2.960919540229885e-05,
+      "loss": 1.7236,
+      "step": 1546
+    },
+    {
+      "epoch": 14.19,
+      "learning_rate": 2.9563218390804598e-05,
+      "loss": 1.7969,
+      "step": 1547
+    },
+    {
+      "epoch": 14.2,
+      "learning_rate": 2.9517241379310344e-05,
+      "loss": 1.7384,
+      "step": 1548
+    },
+    {
+      "epoch": 14.21,
+      "learning_rate": 2.9471264367816094e-05,
+      "loss": 1.4736,
+      "step": 1549
+    },
+    {
+      "epoch": 14.22,
+      "learning_rate": 2.942528735632184e-05,
+      "loss": 2.0887,
+      "step": 1550
+    },
+    {
+      "epoch": 14.23,
+      "learning_rate": 2.9379310344827586e-05,
+      "loss": 1.7804,
+      "step": 1551
+    },
+    {
+      "epoch": 14.24,
+      "learning_rate": 2.9333333333333336e-05,
+      "loss": 1.7489,
+      "step": 1552
+    },
+    {
+      "epoch": 14.25,
+      "learning_rate": 2.9287356321839082e-05,
+      "loss": 1.5519,
+      "step": 1553
+    },
+    {
+      "epoch": 14.26,
+      "learning_rate": 2.924137931034483e-05,
+      "loss": 1.6989,
+      "step": 1554
+    },
+    {
+      "epoch": 14.27,
+      "learning_rate": 2.919540229885058e-05,
+      "loss": 2.1844,
+      "step": 1555
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 2.9149425287356325e-05,
+      "loss": 1.9168,
+      "step": 1556
+    },
+    {
+      "epoch": 14.28,
+      "learning_rate": 2.910344827586207e-05,
+      "loss": 2.4704,
+      "step": 1557
+    },
+    {
+      "epoch": 14.29,
+      "learning_rate": 2.905747126436782e-05,
+      "loss": 1.7443,
+      "step": 1558
+    },
+    {
+      "epoch": 14.3,
+      "learning_rate": 2.9011494252873567e-05,
+      "loss": 1.9082,
+      "step": 1559
+    },
+    {
+      "epoch": 14.31,
+      "learning_rate": 2.8965517241379313e-05,
+      "loss": 1.634,
+      "step": 1560
+    },
+    {
+      "epoch": 14.32,
+      "learning_rate": 2.8919540229885063e-05,
+      "loss": 1.7171,
+      "step": 1561
+    },
+    {
+      "epoch": 14.33,
+      "learning_rate": 2.8873563218390803e-05,
+      "loss": 1.5677,
+      "step": 1562
+    },
+    {
+      "epoch": 14.34,
+      "learning_rate": 2.8827586206896552e-05,
+      "loss": 1.5802,
+      "step": 1563
+    },
+    {
+      "epoch": 14.35,
+      "learning_rate": 2.87816091954023e-05,
+      "loss": 1.7687,
+      "step": 1564
+    },
+    {
+      "epoch": 14.36,
+      "learning_rate": 2.8735632183908045e-05,
+      "loss": 1.5824,
+      "step": 1565
+    },
+    {
+      "epoch": 14.37,
+      "learning_rate": 2.8689655172413795e-05,
+      "loss": 2.3631,
+      "step": 1566
+    },
+    {
+      "epoch": 14.38,
+      "learning_rate": 2.864367816091954e-05,
+      "loss": 1.8706,
+      "step": 1567
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 2.8597701149425287e-05,
+      "loss": 1.8057,
+      "step": 1568
+    },
+    {
+      "epoch": 14.39,
+      "learning_rate": 2.8551724137931034e-05,
+      "loss": 1.9193,
+      "step": 1569
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 2.8505747126436783e-05,
+      "loss": 1.606,
+      "step": 1570
+    },
+    {
+      "epoch": 14.41,
+      "learning_rate": 2.845977011494253e-05,
+      "loss": 2.0651,
+      "step": 1571
+    },
+    {
+      "epoch": 14.42,
+      "learning_rate": 2.8413793103448276e-05,
+      "loss": 1.7275,
+      "step": 1572
+    },
+    {
+      "epoch": 14.43,
+      "learning_rate": 2.8367816091954026e-05,
+      "loss": 1.8053,
+      "step": 1573
+    },
+    {
+      "epoch": 14.44,
+      "learning_rate": 2.8321839080459772e-05,
+      "loss": 1.6084,
+      "step": 1574
+    },
+    {
+      "epoch": 14.45,
+      "learning_rate": 2.8275862068965518e-05,
+      "loss": 2.0034,
+      "step": 1575
+    },
+    {
+      "epoch": 14.46,
+      "learning_rate": 2.8229885057471268e-05,
+      "loss": 1.6413,
+      "step": 1576
+    },
+    {
+      "epoch": 14.47,
+      "learning_rate": 2.8183908045977014e-05,
+      "loss": 2.0604,
+      "step": 1577
+    },
+    {
+      "epoch": 14.48,
+      "learning_rate": 2.813793103448276e-05,
+      "loss": 1.7454,
+      "step": 1578
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 2.809195402298851e-05,
+      "loss": 1.7236,
+      "step": 1579
+    },
+    {
+      "epoch": 14.5,
+      "learning_rate": 2.8045977011494257e-05,
+      "loss": 1.4935,
+      "step": 1580
+    },
+    {
+      "epoch": 14.5,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 2.1271,
+      "step": 1581
+    },
+    {
+      "epoch": 14.51,
+      "learning_rate": 2.7954022988505746e-05,
+      "loss": 1.6323,
+      "step": 1582
+    },
+    {
+      "epoch": 14.52,
+      "learning_rate": 2.7908045977011492e-05,
+      "loss": 1.4438,
+      "step": 1583
+    },
+    {
+      "epoch": 14.53,
+      "learning_rate": 2.7862068965517242e-05,
+      "loss": 1.8148,
+      "step": 1584
+    },
+    {
+      "epoch": 14.54,
+      "learning_rate": 2.7816091954022988e-05,
+      "loss": 2.1118,
+      "step": 1585
+    },
+    {
+      "epoch": 14.55,
+      "learning_rate": 2.7770114942528735e-05,
+      "loss": 1.7695,
+      "step": 1586
+    },
+    {
+      "epoch": 14.56,
+      "learning_rate": 2.7724137931034484e-05,
+      "loss": 1.5965,
+      "step": 1587
+    },
+    {
+      "epoch": 14.57,
+      "learning_rate": 2.767816091954023e-05,
+      "loss": 1.9694,
+      "step": 1588
+    },
+    {
+      "epoch": 14.58,
+      "learning_rate": 2.7632183908045977e-05,
+      "loss": 1.592,
+      "step": 1589
+    },
+    {
+      "epoch": 14.59,
+      "learning_rate": 2.7586206896551727e-05,
+      "loss": 1.6727,
+      "step": 1590
+    },
+    {
+      "epoch": 14.6,
+      "learning_rate": 2.7540229885057473e-05,
+      "loss": 1.5648,
+      "step": 1591
+    },
+    {
+      "epoch": 14.61,
+      "learning_rate": 2.749425287356322e-05,
+      "loss": 1.6624,
+      "step": 1592
+    },
+    {
+      "epoch": 14.61,
+      "learning_rate": 2.744827586206897e-05,
+      "loss": 2.1903,
+      "step": 1593
+    },
+    {
+      "epoch": 14.62,
+      "learning_rate": 2.7402298850574715e-05,
+      "loss": 1.6885,
+      "step": 1594
+    },
+    {
+      "epoch": 14.63,
+      "learning_rate": 2.735632183908046e-05,
+      "loss": 2.098,
+      "step": 1595
+    },
+    {
+      "epoch": 14.64,
+      "learning_rate": 2.7310344827586208e-05,
+      "loss": 1.3237,
+      "step": 1596
+    },
+    {
+      "epoch": 14.65,
+      "learning_rate": 2.7264367816091958e-05,
+      "loss": 1.7851,
+      "step": 1597
+    },
+    {
+      "epoch": 14.66,
+      "learning_rate": 2.7218390804597704e-05,
+      "loss": 1.6541,
+      "step": 1598
+    },
+    {
+      "epoch": 14.67,
+      "learning_rate": 2.717241379310345e-05,
+      "loss": 1.473,
+      "step": 1599
+    },
+    {
+      "epoch": 14.68,
+      "learning_rate": 2.71264367816092e-05,
+      "loss": 2.3848,
+      "step": 1600
+    },
+    {
+      "epoch": 14.69,
+      "learning_rate": 2.7080459770114946e-05,
+      "loss": 1.7733,
+      "step": 1601
+    },
+    {
+      "epoch": 14.7,
+      "learning_rate": 2.703448275862069e-05,
+      "loss": 1.372,
+      "step": 1602
+    },
+    {
+      "epoch": 14.71,
+      "learning_rate": 2.6988505747126436e-05,
+      "loss": 1.8623,
+      "step": 1603
+    },
+    {
+      "epoch": 14.72,
+      "learning_rate": 2.6942528735632182e-05,
+      "loss": 1.9893,
+      "step": 1604
+    },
+    {
+      "epoch": 14.72,
+      "learning_rate": 2.689655172413793e-05,
+      "loss": 1.6007,
+      "step": 1605
+    },
+    {
+      "epoch": 14.73,
+      "learning_rate": 2.6850574712643678e-05,
+      "loss": 1.9106,
+      "step": 1606
+    },
+    {
+      "epoch": 14.74,
+      "learning_rate": 2.6804597701149424e-05,
+      "loss": 1.8382,
+      "step": 1607
+    },
+    {
+      "epoch": 14.75,
+      "learning_rate": 2.6758620689655174e-05,
+      "loss": 1.6369,
+      "step": 1608
+    },
+    {
+      "epoch": 14.76,
+      "learning_rate": 2.671264367816092e-05,
+      "loss": 1.8024,
+      "step": 1609
+    },
+    {
+      "epoch": 14.77,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 1.889,
+      "step": 1610
+    },
+    {
+      "epoch": 14.78,
+      "learning_rate": 2.6620689655172416e-05,
+      "loss": 1.5881,
+      "step": 1611
+    },
+    {
+      "epoch": 14.79,
+      "learning_rate": 2.6574712643678163e-05,
+      "loss": 2.0518,
+      "step": 1612
+    },
+    {
+      "epoch": 14.8,
+      "learning_rate": 2.652873563218391e-05,
+      "loss": 1.9819,
+      "step": 1613
+    },
+    {
+      "epoch": 14.81,
+      "learning_rate": 2.648275862068966e-05,
+      "loss": 2.3205,
+      "step": 1614
+    },
+    {
+      "epoch": 14.82,
+      "learning_rate": 2.6436781609195405e-05,
+      "loss": 1.7902,
+      "step": 1615
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 2.639080459770115e-05,
+      "loss": 1.835,
+      "step": 1616
+    },
+    {
+      "epoch": 14.83,
+      "learning_rate": 2.63448275862069e-05,
+      "loss": 1.6623,
+      "step": 1617
+    },
+    {
+      "epoch": 14.84,
+      "learning_rate": 2.6298850574712647e-05,
+      "loss": 1.5251,
+      "step": 1618
+    },
+    {
+      "epoch": 14.85,
+      "learning_rate": 2.6252873563218394e-05,
+      "loss": 2.318,
+      "step": 1619
+    },
+    {
+      "epoch": 14.86,
+      "learning_rate": 2.620689655172414e-05,
+      "loss": 2.1434,
+      "step": 1620
+    },
+    {
+      "epoch": 14.87,
+      "learning_rate": 2.616091954022989e-05,
+      "loss": 1.6634,
+      "step": 1621
+    },
+    {
+      "epoch": 14.88,
+      "learning_rate": 2.611494252873563e-05,
+      "loss": 1.9053,
+      "step": 1622
+    },
+    {
+      "epoch": 14.89,
+      "learning_rate": 2.606896551724138e-05,
+      "loss": 1.6914,
+      "step": 1623
+    },
+    {
+      "epoch": 14.9,
+      "learning_rate": 2.6022988505747125e-05,
+      "loss": 1.9811,
+      "step": 1624
+    },
+    {
+      "epoch": 14.91,
+      "learning_rate": 2.597701149425287e-05,
+      "loss": 2.021,
+      "step": 1625
+    },
+    {
+      "epoch": 14.92,
+      "learning_rate": 2.593103448275862e-05,
+      "loss": 1.9419,
+      "step": 1626
+    },
+    {
+      "epoch": 14.93,
+      "learning_rate": 2.5885057471264368e-05,
+      "loss": 1.5719,
+      "step": 1627
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.5839080459770114e-05,
+      "loss": 1.7918,
+      "step": 1628
+    },
+    {
+      "epoch": 14.94,
+      "learning_rate": 2.5793103448275864e-05,
+      "loss": 1.9688,
+      "step": 1629
+    },
+    {
+      "epoch": 14.95,
+      "learning_rate": 2.574712643678161e-05,
+      "loss": 1.5233,
+      "step": 1630
+    },
+    {
+      "epoch": 14.96,
+      "learning_rate": 2.5701149425287356e-05,
+      "loss": 1.895,
+      "step": 1631
+    },
+    {
+      "epoch": 14.97,
+      "learning_rate": 2.5655172413793106e-05,
+      "loss": 2.812,
+      "step": 1632
+    },
+    {
+      "epoch": 14.98,
+      "learning_rate": 2.5609195402298852e-05,
+      "loss": 1.9437,
+      "step": 1633
+    },
+    {
+      "epoch": 14.99,
+      "learning_rate": 2.55632183908046e-05,
+      "loss": 2.0636,
+      "step": 1634
+    },
+    {
+      "epoch": 15.0,
+      "learning_rate": 2.551724137931035e-05,
+      "loss": 1.7271,
+      "step": 1635
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": 0.22560975609756098,
+      "eval_loss": 1.8843815326690674,
+      "eval_runtime": 13.8081,
+      "eval_samples_per_second": 11.877,
+      "eval_steps_per_second": 1.521,
+      "step": 1635
+    },
+    {
+      "epoch": 15.01,
+      "learning_rate": 2.5471264367816095e-05,
+      "loss": 1.7332,
+      "step": 1636
+    },
+    {
+      "epoch": 15.02,
+      "learning_rate": 2.542528735632184e-05,
+      "loss": 1.7082,
+      "step": 1637
+    },
+    {
+      "epoch": 15.03,
+      "learning_rate": 2.537931034482759e-05,
+      "loss": 1.668,
+      "step": 1638
+    },
+    {
+      "epoch": 15.04,
+      "learning_rate": 2.5333333333333337e-05,
+      "loss": 1.6359,
+      "step": 1639
+    },
+    {
+      "epoch": 15.05,
+      "learning_rate": 2.5287356321839083e-05,
+      "loss": 2.9172,
+      "step": 1640
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 2.5241379310344833e-05,
+      "loss": 1.8018,
+      "step": 1641
+    },
+    {
+      "epoch": 15.06,
+      "learning_rate": 2.5195402298850573e-05,
+      "loss": 1.8132,
+      "step": 1642
+    },
+    {
+      "epoch": 15.07,
+      "learning_rate": 2.5149425287356322e-05,
+      "loss": 1.6418,
+      "step": 1643
+    },
+    {
+      "epoch": 15.08,
+      "learning_rate": 2.510344827586207e-05,
+      "loss": 1.9818,
+      "step": 1644
+    },
+    {
+      "epoch": 15.09,
+      "learning_rate": 2.5057471264367815e-05,
+      "loss": 1.5732,
+      "step": 1645
+    },
+    {
+      "epoch": 15.1,
+      "learning_rate": 2.5011494252873565e-05,
+      "loss": 2.1366,
+      "step": 1646
+    },
+    {
+      "epoch": 15.11,
+      "learning_rate": 2.496551724137931e-05,
+      "loss": 1.7344,
+      "step": 1647
+    },
+    {
+      "epoch": 15.12,
+      "learning_rate": 2.4919540229885057e-05,
+      "loss": 1.8201,
+      "step": 1648
+    },
+    {
+      "epoch": 15.13,
+      "learning_rate": 2.4873563218390804e-05,
+      "loss": 1.6927,
+      "step": 1649
+    },
+    {
+      "epoch": 15.14,
+      "learning_rate": 2.4827586206896553e-05,
+      "loss": 1.6735,
+      "step": 1650
+    },
+    {
+      "epoch": 15.15,
+      "learning_rate": 2.47816091954023e-05,
+      "loss": 2.34,
+      "step": 1651
+    },
+    {
+      "epoch": 15.16,
+      "learning_rate": 2.4735632183908046e-05,
+      "loss": 1.4194,
+      "step": 1652
+    },
+    {
+      "epoch": 15.17,
+      "learning_rate": 2.4689655172413796e-05,
+      "loss": 1.8441,
+      "step": 1653
+    },
+    {
+      "epoch": 15.17,
+      "learning_rate": 2.4643678160919542e-05,
+      "loss": 1.9115,
+      "step": 1654
+    },
+    {
+      "epoch": 15.18,
+      "learning_rate": 2.4597701149425288e-05,
+      "loss": 1.6226,
+      "step": 1655
+    },
+    {
+      "epoch": 15.19,
+      "learning_rate": 2.4551724137931038e-05,
+      "loss": 1.6839,
+      "step": 1656
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 2.450574712643678e-05,
+      "loss": 2.2098,
+      "step": 1657
+    },
+    {
+      "epoch": 15.21,
+      "learning_rate": 2.445977011494253e-05,
+      "loss": 2.1073,
+      "step": 1658
+    },
+    {
+      "epoch": 15.22,
+      "learning_rate": 2.4413793103448277e-05,
+      "loss": 1.2534,
+      "step": 1659
+    },
+    {
+      "epoch": 15.23,
+      "learning_rate": 2.4367816091954023e-05,
+      "loss": 1.6341,
+      "step": 1660
+    },
+    {
+      "epoch": 15.24,
+      "learning_rate": 2.432183908045977e-05,
+      "loss": 1.4996,
+      "step": 1661
+    },
+    {
+      "epoch": 15.25,
+      "learning_rate": 2.427586206896552e-05,
+      "loss": 1.911,
+      "step": 1662
+    },
+    {
+      "epoch": 15.26,
+      "learning_rate": 2.4229885057471266e-05,
+      "loss": 1.7266,
+      "step": 1663
+    },
+    {
+      "epoch": 15.27,
+      "learning_rate": 2.4183908045977012e-05,
+      "loss": 1.8965,
+      "step": 1664
+    },
+    {
+      "epoch": 15.28,
+      "learning_rate": 2.413793103448276e-05,
+      "loss": 1.6419,
+      "step": 1665
+    },
+    {
+      "epoch": 15.28,
+      "learning_rate": 2.4091954022988508e-05,
+      "loss": 1.6712,
+      "step": 1666
+    },
+    {
+      "epoch": 15.29,
+      "learning_rate": 2.4045977011494254e-05,
+      "loss": 1.4689,
+      "step": 1667
+    },
+    {
+      "epoch": 15.3,
+      "learning_rate": 2.4e-05,
+      "loss": 1.8421,
+      "step": 1668
+    },
+    {
+      "epoch": 15.31,
+      "learning_rate": 2.3954022988505747e-05,
+      "loss": 1.633,
+      "step": 1669
+    },
+    {
+      "epoch": 15.32,
+      "learning_rate": 2.3908045977011497e-05,
+      "loss": 1.7638,
+      "step": 1670
+    },
+    {
+      "epoch": 15.33,
+      "learning_rate": 2.3862068965517243e-05,
+      "loss": 1.7814,
+      "step": 1671
+    },
+    {
+      "epoch": 15.34,
+      "learning_rate": 2.381609195402299e-05,
+      "loss": 1.7199,
+      "step": 1672
+    },
+    {
+      "epoch": 15.35,
+      "learning_rate": 2.3770114942528736e-05,
+      "loss": 1.7913,
+      "step": 1673
+    },
+    {
+      "epoch": 15.36,
+      "learning_rate": 2.3724137931034485e-05,
+      "loss": 1.9087,
+      "step": 1674
+    },
+    {
+      "epoch": 15.37,
+      "learning_rate": 2.367816091954023e-05,
+      "loss": 1.5482,
+      "step": 1675
+    },
+    {
+      "epoch": 15.38,
+      "learning_rate": 2.3632183908045978e-05,
+      "loss": 1.8262,
+      "step": 1676
+    },
+    {
+      "epoch": 15.39,
+      "learning_rate": 2.3586206896551724e-05,
+      "loss": 1.9912,
+      "step": 1677
+    },
+    {
+      "epoch": 15.39,
+      "learning_rate": 2.354022988505747e-05,
+      "loss": 1.6935,
+      "step": 1678
+    },
+    {
+      "epoch": 15.4,
+      "learning_rate": 2.349425287356322e-05,
+      "loss": 1.5713,
+      "step": 1679
+    },
+    {
+      "epoch": 15.41,
+      "learning_rate": 2.3448275862068967e-05,
+      "loss": 1.6357,
+      "step": 1680
+    },
+    {
+      "epoch": 15.42,
+      "learning_rate": 2.3402298850574713e-05,
+      "loss": 2.0249,
+      "step": 1681
+    },
+    {
+      "epoch": 15.43,
+      "learning_rate": 2.3356321839080463e-05,
+      "loss": 2.0511,
+      "step": 1682
+    },
+    {
+      "epoch": 15.44,
+      "learning_rate": 2.331034482758621e-05,
+      "loss": 1.7285,
+      "step": 1683
+    },
+    {
+      "epoch": 15.45,
+      "learning_rate": 2.3264367816091955e-05,
+      "loss": 1.62,
+      "step": 1684
+    },
+    {
+      "epoch": 15.46,
+      "learning_rate": 2.3218390804597705e-05,
+      "loss": 1.7941,
+      "step": 1685
+    },
+    {
+      "epoch": 15.47,
+      "learning_rate": 2.317241379310345e-05,
+      "loss": 1.976,
+      "step": 1686
+    },
+    {
+      "epoch": 15.48,
+      "learning_rate": 2.3126436781609194e-05,
+      "loss": 1.4738,
+      "step": 1687
+    },
+    {
+      "epoch": 15.49,
+      "learning_rate": 2.3080459770114944e-05,
+      "loss": 1.7324,
+      "step": 1688
+    },
+    {
+      "epoch": 15.5,
+      "learning_rate": 2.303448275862069e-05,
+      "loss": 1.766,
+      "step": 1689
+    },
+    {
+      "epoch": 15.5,
+      "learning_rate": 2.2988505747126437e-05,
+      "loss": 1.7443,
+      "step": 1690
+    },
+    {
+      "epoch": 15.51,
+      "learning_rate": 2.2942528735632186e-05,
+      "loss": 1.6686,
+      "step": 1691
+    },
+    {
+      "epoch": 15.52,
+      "learning_rate": 2.2896551724137933e-05,
+      "loss": 1.946,
+      "step": 1692
+    },
+    {
+      "epoch": 15.53,
+      "learning_rate": 2.285057471264368e-05,
+      "loss": 1.5653,
+      "step": 1693
+    },
+    {
+      "epoch": 15.54,
+      "learning_rate": 2.280459770114943e-05,
+      "loss": 1.6733,
+      "step": 1694
+    },
+    {
+      "epoch": 15.55,
+      "learning_rate": 2.2758620689655175e-05,
+      "loss": 1.4906,
+      "step": 1695
+    },
+    {
+      "epoch": 15.56,
+      "learning_rate": 2.271264367816092e-05,
+      "loss": 1.7023,
+      "step": 1696
+    },
+    {
+      "epoch": 15.57,
+      "learning_rate": 2.2666666666666668e-05,
+      "loss": 1.7244,
+      "step": 1697
+    },
+    {
+      "epoch": 15.58,
+      "learning_rate": 2.2620689655172414e-05,
+      "loss": 1.5703,
+      "step": 1698
+    },
+    {
+      "epoch": 15.59,
+      "learning_rate": 2.257471264367816e-05,
+      "loss": 1.8245,
+      "step": 1699
+    },
+    {
+      "epoch": 15.6,
+      "learning_rate": 2.252873563218391e-05,
+      "loss": 1.563,
+      "step": 1700
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 2.2482758620689656e-05,
+      "loss": 1.7417,
+      "step": 1701
+    },
+    {
+      "epoch": 15.61,
+      "learning_rate": 2.2436781609195403e-05,
+      "loss": 1.9163,
+      "step": 1702
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 2.2390804597701152e-05,
+      "loss": 1.8657,
+      "step": 1703
+    },
+    {
+      "epoch": 15.63,
+      "learning_rate": 2.23448275862069e-05,
+      "loss": 1.6245,
+      "step": 1704
+    },
+    {
+      "epoch": 15.64,
+      "learning_rate": 2.2298850574712645e-05,
+      "loss": 1.8483,
+      "step": 1705
+    },
+    {
+      "epoch": 15.65,
+      "learning_rate": 2.2252873563218395e-05,
+      "loss": 1.9165,
+      "step": 1706
+    },
+    {
+      "epoch": 15.66,
+      "learning_rate": 2.2206896551724138e-05,
+      "loss": 1.6232,
+      "step": 1707
+    },
+    {
+      "epoch": 15.67,
+      "learning_rate": 2.2160919540229884e-05,
+      "loss": 1.7046,
+      "step": 1708
+    },
+    {
+      "epoch": 15.68,
+      "learning_rate": 2.2114942528735634e-05,
+      "loss": 1.6408,
+      "step": 1709
+    },
+    {
+      "epoch": 15.69,
+      "learning_rate": 2.206896551724138e-05,
+      "loss": 2.2756,
+      "step": 1710
+    },
+    {
+      "epoch": 15.7,
+      "learning_rate": 2.2022988505747126e-05,
+      "loss": 2.1585,
+      "step": 1711
+    },
+    {
+      "epoch": 15.71,
+      "learning_rate": 2.1977011494252876e-05,
+      "loss": 1.5264,
+      "step": 1712
+    },
+    {
+      "epoch": 15.72,
+      "learning_rate": 2.1931034482758622e-05,
+      "loss": 1.7729,
+      "step": 1713
+    },
+    {
+      "epoch": 15.72,
+      "learning_rate": 2.188505747126437e-05,
+      "loss": 2.6007,
+      "step": 1714
+    },
+    {
+      "epoch": 15.73,
+      "learning_rate": 2.183908045977012e-05,
+      "loss": 1.7759,
+      "step": 1715
+    },
+    {
+      "epoch": 15.74,
+      "learning_rate": 2.1793103448275865e-05,
+      "loss": 1.7961,
+      "step": 1716
+    },
+    {
+      "epoch": 15.75,
+      "learning_rate": 2.1747126436781608e-05,
+      "loss": 2.3807,
+      "step": 1717
+    },
+    {
+      "epoch": 15.76,
+      "learning_rate": 2.1701149425287357e-05,
+      "loss": 2.4976,
+      "step": 1718
+    },
+    {
+      "epoch": 15.77,
+      "learning_rate": 2.1655172413793104e-05,
+      "loss": 1.8289,
+      "step": 1719
+    },
+    {
+      "epoch": 15.78,
+      "learning_rate": 2.160919540229885e-05,
+      "loss": 1.8258,
+      "step": 1720
+    },
+    {
+      "epoch": 15.79,
+      "learning_rate": 2.15632183908046e-05,
+      "loss": 1.834,
+      "step": 1721
+    },
+    {
+      "epoch": 15.8,
+      "learning_rate": 2.1517241379310346e-05,
+      "loss": 1.8302,
+      "step": 1722
+    },
+    {
+      "epoch": 15.81,
+      "learning_rate": 2.1471264367816092e-05,
+      "loss": 1.7635,
+      "step": 1723
+    },
+    {
+      "epoch": 15.82,
+      "learning_rate": 2.1425287356321842e-05,
+      "loss": 1.4315,
+      "step": 1724
+    },
+    {
+      "epoch": 15.83,
+      "learning_rate": 2.137931034482759e-05,
+      "loss": 2.0212,
+      "step": 1725
+    },
+    {
+      "epoch": 15.83,
+      "learning_rate": 2.1333333333333335e-05,
+      "loss": 1.7803,
+      "step": 1726
+    },
+    {
+      "epoch": 15.84,
+      "learning_rate": 2.128735632183908e-05,
+      "loss": 1.4383,
+      "step": 1727
+    },
+    {
+      "epoch": 15.85,
+      "learning_rate": 2.1241379310344827e-05,
+      "loss": 2.4782,
+      "step": 1728
+    },
+    {
+      "epoch": 15.86,
+      "learning_rate": 2.1195402298850574e-05,
+      "loss": 2.0928,
+      "step": 1729
+    },
+    {
+      "epoch": 15.87,
+      "learning_rate": 2.1149425287356323e-05,
+      "loss": 1.6455,
+      "step": 1730
+    },
+    {
+      "epoch": 15.88,
+      "learning_rate": 2.110344827586207e-05,
+      "loss": 1.9884,
+      "step": 1731
+    },
+    {
+      "epoch": 15.89,
+      "learning_rate": 2.1057471264367816e-05,
+      "loss": 1.6217,
+      "step": 1732
+    },
+    {
+      "epoch": 15.9,
+      "learning_rate": 2.1011494252873566e-05,
+      "loss": 1.7039,
+      "step": 1733
+    },
+    {
+      "epoch": 15.91,
+      "learning_rate": 2.0965517241379312e-05,
+      "loss": 1.7773,
+      "step": 1734
+    },
+    {
+      "epoch": 15.92,
+      "learning_rate": 2.0919540229885058e-05,
+      "loss": 2.4419,
+      "step": 1735
+    },
+    {
+      "epoch": 15.93,
+      "learning_rate": 2.0873563218390808e-05,
+      "loss": 2.2485,
+      "step": 1736
+    },
+    {
+      "epoch": 15.94,
+      "learning_rate": 2.082758620689655e-05,
+      "loss": 2.0182,
+      "step": 1737
+    },
+    {
+      "epoch": 15.94,
+      "learning_rate": 2.07816091954023e-05,
+      "loss": 1.9006,
+      "step": 1738
+    },
+    {
+      "epoch": 15.95,
+      "learning_rate": 2.0735632183908047e-05,
+      "loss": 2.1702,
+      "step": 1739
+    },
+    {
+      "epoch": 15.96,
+      "learning_rate": 2.0689655172413793e-05,
+      "loss": 2.0101,
+      "step": 1740
+    },
+    {
+      "epoch": 15.97,
+      "learning_rate": 2.064367816091954e-05,
+      "loss": 1.9912,
+      "step": 1741
+    },
+    {
+      "epoch": 15.98,
+      "learning_rate": 2.059770114942529e-05,
+      "loss": 1.7266,
+      "step": 1742
+    },
+    {
+      "epoch": 15.99,
+      "learning_rate": 2.0551724137931036e-05,
+      "loss": 1.7074,
+      "step": 1743
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 2.0505747126436782e-05,
+      "loss": 1.851,
+      "step": 1744
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": 0.21341463414634146,
+      "eval_loss": 1.8719631433486938,
+      "eval_runtime": 13.8073,
+      "eval_samples_per_second": 11.878,
+      "eval_steps_per_second": 1.521,
+      "step": 1744
+    },
+    {
+      "epoch": 16.01,
+      "learning_rate": 2.045977011494253e-05,
+      "loss": 2.4538,
+      "step": 1745
+    },
+    {
+      "epoch": 16.02,
+      "learning_rate": 2.0413793103448278e-05,
+      "loss": 1.5378,
+      "step": 1746
+    },
+    {
+      "epoch": 16.03,
+      "learning_rate": 2.0367816091954024e-05,
+      "loss": 1.6201,
+      "step": 1747
+    },
+    {
+      "epoch": 16.04,
+      "learning_rate": 2.032183908045977e-05,
+      "loss": 1.6374,
+      "step": 1748
+    },
+    {
+      "epoch": 16.05,
+      "learning_rate": 2.0275862068965517e-05,
+      "loss": 1.8158,
+      "step": 1749
+    },
+    {
+      "epoch": 16.06,
+      "learning_rate": 2.0229885057471267e-05,
+      "loss": 1.6787,
+      "step": 1750
+    },
+    {
+      "epoch": 16.06,
+      "learning_rate": 2.0183908045977013e-05,
+      "loss": 1.6675,
+      "step": 1751
+    },
+    {
+      "epoch": 16.07,
+      "learning_rate": 2.013793103448276e-05,
+      "loss": 1.6286,
+      "step": 1752
+    },
+    {
+      "epoch": 16.08,
+      "learning_rate": 2.0091954022988506e-05,
+      "loss": 1.6663,
+      "step": 1753
+    },
+    {
+      "epoch": 16.09,
+      "learning_rate": 2.0045977011494255e-05,
+      "loss": 1.5321,
+      "step": 1754
+    },
+    {
+      "epoch": 16.1,
+      "learning_rate": 2e-05,
+      "loss": 1.7603,
+      "step": 1755
+    },
+    {
+      "epoch": 16.11,
+      "learning_rate": 1.9954022988505748e-05,
+      "loss": 1.6042,
+      "step": 1756
+    },
+    {
+      "epoch": 16.12,
+      "learning_rate": 1.9908045977011494e-05,
+      "loss": 2.1413,
+      "step": 1757
+    },
+    {
+      "epoch": 16.13,
+      "learning_rate": 1.986206896551724e-05,
+      "loss": 1.5905,
+      "step": 1758
+    },
+    {
+      "epoch": 16.14,
+      "learning_rate": 1.981609195402299e-05,
+      "loss": 1.8213,
+      "step": 1759
+    },
+    {
+      "epoch": 16.15,
+      "learning_rate": 1.9770114942528737e-05,
+      "loss": 2.2788,
+      "step": 1760
+    },
+    {
+      "epoch": 16.16,
+      "learning_rate": 1.9724137931034483e-05,
+      "loss": 2.2445,
+      "step": 1761
+    },
+    {
+      "epoch": 16.17,
+      "learning_rate": 1.9678160919540233e-05,
+      "loss": 2.0981,
+      "step": 1762
+    },
+    {
+      "epoch": 16.17,
+      "learning_rate": 1.963218390804598e-05,
+      "loss": 2.1543,
+      "step": 1763
+    },
+    {
+      "epoch": 16.18,
+      "learning_rate": 1.9586206896551725e-05,
+      "loss": 1.7363,
+      "step": 1764
+    },
+    {
+      "epoch": 16.19,
+      "learning_rate": 1.9540229885057475e-05,
+      "loss": 2.0233,
+      "step": 1765
+    },
+    {
+      "epoch": 16.2,
+      "learning_rate": 1.9494252873563218e-05,
+      "loss": 1.675,
+      "step": 1766
+    },
+    {
+      "epoch": 16.21,
+      "learning_rate": 1.9448275862068964e-05,
+      "loss": 1.604,
+      "step": 1767
+    },
+    {
+      "epoch": 16.22,
+      "learning_rate": 1.9402298850574714e-05,
+      "loss": 1.7731,
+      "step": 1768
+    },
+    {
+      "epoch": 16.23,
+      "learning_rate": 1.935632183908046e-05,
+      "loss": 1.6901,
+      "step": 1769
+    },
+    {
+      "epoch": 16.24,
+      "learning_rate": 1.9310344827586207e-05,
+      "loss": 2.116,
+      "step": 1770
+    },
+    {
+      "epoch": 16.25,
+      "learning_rate": 1.9264367816091956e-05,
+      "loss": 1.5864,
+      "step": 1771
+    },
+    {
+      "epoch": 16.26,
+      "learning_rate": 1.9218390804597703e-05,
+      "loss": 1.8122,
+      "step": 1772
+    },
+    {
+      "epoch": 16.27,
+      "learning_rate": 1.917241379310345e-05,
+      "loss": 1.9989,
+      "step": 1773
+    },
+    {
+      "epoch": 16.28,
+      "learning_rate": 1.91264367816092e-05,
+      "loss": 1.7694,
+      "step": 1774
+    },
+    {
+      "epoch": 16.28,
+      "learning_rate": 1.9080459770114945e-05,
+      "loss": 1.8421,
+      "step": 1775
+    },
+    {
+      "epoch": 16.29,
+      "learning_rate": 1.9034482758620688e-05,
+      "loss": 1.8328,
+      "step": 1776
+    },
+    {
+      "epoch": 16.3,
+      "learning_rate": 1.8988505747126438e-05,
+      "loss": 1.6348,
+      "step": 1777
+    },
+    {
+      "epoch": 16.31,
+      "learning_rate": 1.8942528735632184e-05,
+      "loss": 1.9723,
+      "step": 1778
+    },
+    {
+      "epoch": 16.32,
+      "learning_rate": 1.889655172413793e-05,
+      "loss": 1.8636,
+      "step": 1779
+    },
+    {
+      "epoch": 16.33,
+      "learning_rate": 1.885057471264368e-05,
+      "loss": 1.8268,
+      "step": 1780
+    },
+    {
+      "epoch": 16.34,
+      "learning_rate": 1.8804597701149426e-05,
+      "loss": 1.7373,
+      "step": 1781
+    },
+    {
+      "epoch": 16.35,
+      "learning_rate": 1.8758620689655173e-05,
+      "loss": 2.5465,
+      "step": 1782
+    },
+    {
+      "epoch": 16.36,
+      "learning_rate": 1.8712643678160922e-05,
+      "loss": 1.5959,
+      "step": 1783
+    },
+    {
+      "epoch": 16.37,
+      "learning_rate": 1.866666666666667e-05,
+      "loss": 2.1048,
+      "step": 1784
+    },
+    {
+      "epoch": 16.38,
+      "learning_rate": 1.8620689655172415e-05,
+      "loss": 1.9517,
+      "step": 1785
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 1.857471264367816e-05,
+      "loss": 1.7157,
+      "step": 1786
+    },
+    {
+      "epoch": 16.39,
+      "learning_rate": 1.8528735632183908e-05,
+      "loss": 1.5264,
+      "step": 1787
+    },
+    {
+      "epoch": 16.4,
+      "learning_rate": 1.8482758620689654e-05,
+      "loss": 1.7161,
+      "step": 1788
+    },
+    {
+      "epoch": 16.41,
+      "learning_rate": 1.8436781609195404e-05,
+      "loss": 2.3171,
+      "step": 1789
+    },
+    {
+      "epoch": 16.42,
+      "learning_rate": 1.839080459770115e-05,
+      "loss": 1.495,
+      "step": 1790
+    },
+    {
+      "epoch": 16.43,
+      "learning_rate": 1.8344827586206896e-05,
+      "loss": 1.7321,
+      "step": 1791
+    },
+    {
+      "epoch": 16.44,
+      "learning_rate": 1.8298850574712646e-05,
+      "loss": 2.0011,
+      "step": 1792
+    },
+    {
+      "epoch": 16.45,
+      "learning_rate": 1.8252873563218392e-05,
+      "loss": 1.6976,
+      "step": 1793
+    },
+    {
+      "epoch": 16.46,
+      "learning_rate": 1.820689655172414e-05,
+      "loss": 1.5715,
+      "step": 1794
+    },
+    {
+      "epoch": 16.47,
+      "learning_rate": 1.816091954022989e-05,
+      "loss": 1.5409,
+      "step": 1795
+    },
+    {
+      "epoch": 16.48,
+      "learning_rate": 1.811494252873563e-05,
+      "loss": 1.6737,
+      "step": 1796
+    },
+    {
+      "epoch": 16.49,
+      "learning_rate": 1.8068965517241378e-05,
+      "loss": 1.7484,
+      "step": 1797
+    },
+    {
+      "epoch": 16.5,
+      "learning_rate": 1.8022988505747127e-05,
+      "loss": 1.5293,
+      "step": 1798
+    },
+    {
+      "epoch": 16.5,
+      "learning_rate": 1.7977011494252874e-05,
+      "loss": 1.5758,
+      "step": 1799
+    },
+    {
+      "epoch": 16.51,
+      "learning_rate": 1.793103448275862e-05,
+      "loss": 1.6663,
+      "step": 1800
+    },
+    {
+      "epoch": 16.52,
+      "learning_rate": 1.788505747126437e-05,
+      "loss": 1.6991,
+      "step": 1801
+    },
+    {
+      "epoch": 16.53,
+      "learning_rate": 1.7839080459770116e-05,
+      "loss": 1.6533,
+      "step": 1802
+    },
+    {
+      "epoch": 16.54,
+      "learning_rate": 1.7793103448275862e-05,
+      "loss": 1.7358,
+      "step": 1803
+    },
+    {
+      "epoch": 16.55,
+      "learning_rate": 1.7747126436781612e-05,
+      "loss": 2.1868,
+      "step": 1804
+    },
+    {
+      "epoch": 16.56,
+      "learning_rate": 1.770114942528736e-05,
+      "loss": 1.6187,
+      "step": 1805
+    },
+    {
+      "epoch": 16.57,
+      "learning_rate": 1.76551724137931e-05,
+      "loss": 1.8219,
+      "step": 1806
+    },
+    {
+      "epoch": 16.58,
+      "learning_rate": 1.760919540229885e-05,
+      "loss": 1.772,
+      "step": 1807
+    },
+    {
+      "epoch": 16.59,
+      "learning_rate": 1.7563218390804597e-05,
+      "loss": 1.5456,
+      "step": 1808
+    },
+    {
+      "epoch": 16.6,
+      "learning_rate": 1.7517241379310344e-05,
+      "loss": 1.763,
+      "step": 1809
+    },
+    {
+      "epoch": 16.61,
+      "learning_rate": 1.7471264367816093e-05,
+      "loss": 1.4909,
+      "step": 1810
+    },
+    {
+      "epoch": 16.61,
+      "learning_rate": 1.742528735632184e-05,
+      "loss": 1.8175,
+      "step": 1811
+    },
+    {
+      "epoch": 16.62,
+      "learning_rate": 1.7379310344827586e-05,
+      "loss": 1.7096,
+      "step": 1812
+    },
+    {
+      "epoch": 16.63,
+      "learning_rate": 1.7333333333333336e-05,
+      "loss": 2.1813,
+      "step": 1813
+    },
+    {
+      "epoch": 16.64,
+      "learning_rate": 1.7287356321839082e-05,
+      "loss": 2.0566,
+      "step": 1814
+    },
+    {
+      "epoch": 16.65,
+      "learning_rate": 1.7241379310344828e-05,
+      "loss": 1.981,
+      "step": 1815
+    },
+    {
+      "epoch": 16.66,
+      "learning_rate": 1.7195402298850575e-05,
+      "loss": 1.9546,
+      "step": 1816
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 1.714942528735632e-05,
+      "loss": 1.7575,
+      "step": 1817
+    },
+    {
+      "epoch": 16.68,
+      "learning_rate": 1.710344827586207e-05,
+      "loss": 1.7743,
+      "step": 1818
+    },
+    {
+      "epoch": 16.69,
+      "learning_rate": 1.7057471264367817e-05,
+      "loss": 1.9899,
+      "step": 1819
+    },
+    {
+      "epoch": 16.7,
+      "learning_rate": 1.7011494252873563e-05,
+      "loss": 1.4906,
+      "step": 1820
+    },
+    {
+      "epoch": 16.71,
+      "learning_rate": 1.696551724137931e-05,
+      "loss": 2.3783,
+      "step": 1821
+    },
+    {
+      "epoch": 16.72,
+      "learning_rate": 1.691954022988506e-05,
+      "loss": 1.6141,
+      "step": 1822
+    },
+    {
+      "epoch": 16.72,
+      "learning_rate": 1.6873563218390806e-05,
+      "loss": 1.7741,
+      "step": 1823
+    },
+    {
+      "epoch": 16.73,
+      "learning_rate": 1.6827586206896552e-05,
+      "loss": 2.0852,
+      "step": 1824
+    },
+    {
+      "epoch": 16.74,
+      "learning_rate": 1.67816091954023e-05,
+      "loss": 1.6502,
+      "step": 1825
+    },
+    {
+      "epoch": 16.75,
+      "learning_rate": 1.6735632183908045e-05,
+      "loss": 2.1024,
+      "step": 1826
+    },
+    {
+      "epoch": 16.76,
+      "learning_rate": 1.6689655172413794e-05,
+      "loss": 1.7739,
+      "step": 1827
+    },
+    {
+      "epoch": 16.77,
+      "learning_rate": 1.664367816091954e-05,
+      "loss": 1.5786,
+      "step": 1828
+    },
+    {
+      "epoch": 16.78,
+      "learning_rate": 1.6597701149425287e-05,
+      "loss": 2.5854,
+      "step": 1829
+    },
+    {
+      "epoch": 16.79,
+      "learning_rate": 1.6551724137931037e-05,
+      "loss": 1.5301,
+      "step": 1830
+    },
+    {
+      "epoch": 16.8,
+      "learning_rate": 1.6505747126436783e-05,
+      "loss": 1.7145,
+      "step": 1831
+    },
+    {
+      "epoch": 16.81,
+      "learning_rate": 1.645977011494253e-05,
+      "loss": 1.749,
+      "step": 1832
+    },
+    {
+      "epoch": 16.82,
+      "learning_rate": 1.6413793103448276e-05,
+      "loss": 1.7703,
+      "step": 1833
+    },
+    {
+      "epoch": 16.83,
+      "learning_rate": 1.6367816091954025e-05,
+      "loss": 1.363,
+      "step": 1834
+    },
+    {
+      "epoch": 16.83,
+      "learning_rate": 1.632183908045977e-05,
+      "loss": 1.5184,
+      "step": 1835
+    },
+    {
+      "epoch": 16.84,
+      "learning_rate": 1.6275862068965518e-05,
+      "loss": 1.7321,
+      "step": 1836
+    },
+    {
+      "epoch": 16.85,
+      "learning_rate": 1.6229885057471264e-05,
+      "loss": 1.6095,
+      "step": 1837
+    },
+    {
+      "epoch": 16.86,
+      "learning_rate": 1.618390804597701e-05,
+      "loss": 1.7881,
+      "step": 1838
+    },
+    {
+      "epoch": 16.87,
+      "learning_rate": 1.613793103448276e-05,
+      "loss": 2.151,
+      "step": 1839
+    },
+    {
+      "epoch": 16.88,
+      "learning_rate": 1.6091954022988507e-05,
+      "loss": 1.8125,
+      "step": 1840
+    },
+    {
+      "epoch": 16.89,
+      "learning_rate": 1.6045977011494253e-05,
+      "loss": 1.4826,
+      "step": 1841
+    },
+    {
+      "epoch": 16.9,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 2.0067,
+      "step": 1842
+    },
+    {
+      "epoch": 16.91,
+      "learning_rate": 1.595402298850575e-05,
+      "loss": 2.3835,
+      "step": 1843
+    },
+    {
+      "epoch": 16.92,
+      "learning_rate": 1.5908045977011495e-05,
+      "loss": 1.9051,
+      "step": 1844
+    },
+    {
+      "epoch": 16.93,
+      "learning_rate": 1.586206896551724e-05,
+      "loss": 2.9092,
+      "step": 1845
+    },
+    {
+      "epoch": 16.94,
+      "learning_rate": 1.5816091954022988e-05,
+      "loss": 1.6898,
+      "step": 1846
+    },
+    {
+      "epoch": 16.94,
+      "learning_rate": 1.5770114942528734e-05,
+      "loss": 1.5666,
+      "step": 1847
+    },
+    {
+      "epoch": 16.95,
+      "learning_rate": 1.5724137931034484e-05,
+      "loss": 1.4518,
+      "step": 1848
+    },
+    {
+      "epoch": 16.96,
+      "learning_rate": 1.567816091954023e-05,
+      "loss": 1.6247,
+      "step": 1849
+    },
+    {
+      "epoch": 16.97,
+      "learning_rate": 1.5632183908045977e-05,
+      "loss": 1.6055,
+      "step": 1850
+    },
+    {
+      "epoch": 16.98,
+      "learning_rate": 1.5586206896551726e-05,
+      "loss": 1.7147,
+      "step": 1851
+    },
+    {
+      "epoch": 16.99,
+      "learning_rate": 1.5540229885057473e-05,
+      "loss": 1.9076,
+      "step": 1852
+    },
+    {
+      "epoch": 17.0,
+      "learning_rate": 1.549425287356322e-05,
+      "loss": 1.7633,
+      "step": 1853
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": 0.21341463414634146,
+      "eval_loss": 1.8786174058914185,
+      "eval_runtime": 13.7801,
+      "eval_samples_per_second": 11.901,
+      "eval_steps_per_second": 1.524,
+      "step": 1853
+    },
+    {
+      "epoch": 17.01,
+      "learning_rate": 1.544827586206897e-05,
+      "loss": 2.1003,
+      "step": 1854
+    },
+    {
+      "epoch": 17.02,
+      "learning_rate": 1.5402298850574715e-05,
+      "loss": 2.5952,
+      "step": 1855
+    },
+    {
+      "epoch": 17.03,
+      "learning_rate": 1.5356321839080458e-05,
+      "loss": 1.548,
+      "step": 1856
+    },
+    {
+      "epoch": 17.04,
+      "learning_rate": 1.5310344827586208e-05,
+      "loss": 2.0223,
+      "step": 1857
+    },
+    {
+      "epoch": 17.05,
+      "learning_rate": 1.5264367816091954e-05,
+      "loss": 1.7617,
+      "step": 1858
+    },
+    {
+      "epoch": 17.06,
+      "learning_rate": 1.5218390804597702e-05,
+      "loss": 1.6746,
+      "step": 1859
+    },
+    {
+      "epoch": 17.06,
+      "learning_rate": 1.5172413793103448e-05,
+      "loss": 1.9598,
+      "step": 1860
+    },
+    {
+      "epoch": 17.07,
+      "learning_rate": 1.5126436781609196e-05,
+      "loss": 2.2388,
+      "step": 1861
+    },
+    {
+      "epoch": 17.08,
+      "learning_rate": 1.5080459770114944e-05,
+      "loss": 1.7284,
+      "step": 1862
+    },
+    {
+      "epoch": 17.09,
+      "learning_rate": 1.503448275862069e-05,
+      "loss": 1.8538,
+      "step": 1863
+    },
+    {
+      "epoch": 17.1,
+      "learning_rate": 1.4988505747126439e-05,
+      "loss": 1.9424,
+      "step": 1864
+    },
+    {
+      "epoch": 17.11,
+      "learning_rate": 1.4942528735632185e-05,
+      "loss": 1.5018,
+      "step": 1865
+    },
+    {
+      "epoch": 17.12,
+      "learning_rate": 1.489655172413793e-05,
+      "loss": 1.6875,
+      "step": 1866
+    },
+    {
+      "epoch": 17.13,
+      "learning_rate": 1.4850574712643678e-05,
+      "loss": 1.9541,
+      "step": 1867
+    },
+    {
+      "epoch": 17.14,
+      "learning_rate": 1.4804597701149426e-05,
+      "loss": 1.8123,
+      "step": 1868
+    },
+    {
+      "epoch": 17.15,
+      "learning_rate": 1.4758620689655172e-05,
+      "loss": 1.8073,
+      "step": 1869
+    },
+    {
+      "epoch": 17.16,
+      "learning_rate": 1.471264367816092e-05,
+      "loss": 1.841,
+      "step": 1870
+    },
+    {
+      "epoch": 17.17,
+      "learning_rate": 1.4666666666666668e-05,
+      "loss": 2.1097,
+      "step": 1871
+    },
+    {
+      "epoch": 17.17,
+      "learning_rate": 1.4620689655172414e-05,
+      "loss": 1.4635,
+      "step": 1872
+    },
+    {
+      "epoch": 17.18,
+      "learning_rate": 1.4574712643678162e-05,
+      "loss": 1.7179,
+      "step": 1873
+    },
+    {
+      "epoch": 17.19,
+      "learning_rate": 1.452873563218391e-05,
+      "loss": 1.6642,
+      "step": 1874
+    },
+    {
+      "epoch": 17.2,
+      "learning_rate": 1.4482758620689657e-05,
+      "loss": 1.6476,
+      "step": 1875
+    },
+    {
+      "epoch": 17.21,
+      "learning_rate": 1.4436781609195401e-05,
+      "loss": 1.5916,
+      "step": 1876
+    },
+    {
+      "epoch": 17.22,
+      "learning_rate": 1.439080459770115e-05,
+      "loss": 1.527,
+      "step": 1877
+    },
+    {
+      "epoch": 17.23,
+      "learning_rate": 1.4344827586206897e-05,
+      "loss": 1.6084,
+      "step": 1878
+    },
+    {
+      "epoch": 17.24,
+      "learning_rate": 1.4298850574712644e-05,
+      "loss": 1.6357,
+      "step": 1879
+    },
+    {
+      "epoch": 17.25,
+      "learning_rate": 1.4252873563218392e-05,
+      "loss": 1.8091,
+      "step": 1880
+    },
+    {
+      "epoch": 17.26,
+      "learning_rate": 1.4206896551724138e-05,
+      "loss": 1.4702,
+      "step": 1881
+    },
+    {
+      "epoch": 17.27,
+      "learning_rate": 1.4160919540229886e-05,
+      "loss": 2.0314,
+      "step": 1882
+    },
+    {
+      "epoch": 17.28,
+      "learning_rate": 1.4114942528735634e-05,
+      "loss": 1.8582,
+      "step": 1883
+    },
+    {
+      "epoch": 17.28,
+      "learning_rate": 1.406896551724138e-05,
+      "loss": 1.86,
+      "step": 1884
+    },
+    {
+      "epoch": 17.29,
+      "learning_rate": 1.4022988505747128e-05,
+      "loss": 1.8768,
+      "step": 1885
+    },
+    {
+      "epoch": 17.3,
+      "learning_rate": 1.3977011494252873e-05,
+      "loss": 2.2855,
+      "step": 1886
+    },
+    {
+      "epoch": 17.31,
+      "learning_rate": 1.3931034482758621e-05,
+      "loss": 1.5513,
+      "step": 1887
+    },
+    {
+      "epoch": 17.32,
+      "learning_rate": 1.3885057471264367e-05,
+      "loss": 1.91,
+      "step": 1888
+    },
+    {
+      "epoch": 17.33,
+      "learning_rate": 1.3839080459770115e-05,
+      "loss": 1.745,
+      "step": 1889
+    },
+    {
+      "epoch": 17.34,
+      "learning_rate": 1.3793103448275863e-05,
+      "loss": 1.6802,
+      "step": 1890
+    },
+    {
+      "epoch": 17.35,
+      "learning_rate": 1.374712643678161e-05,
+      "loss": 1.5715,
+      "step": 1891
+    },
+    {
+      "epoch": 17.36,
+      "learning_rate": 1.3701149425287358e-05,
+      "loss": 1.5778,
+      "step": 1892
+    },
+    {
+      "epoch": 17.37,
+      "learning_rate": 1.3655172413793104e-05,
+      "loss": 1.8125,
+      "step": 1893
+    },
+    {
+      "epoch": 17.38,
+      "learning_rate": 1.3609195402298852e-05,
+      "loss": 1.6094,
+      "step": 1894
+    },
+    {
+      "epoch": 17.39,
+      "learning_rate": 1.35632183908046e-05,
+      "loss": 1.8483,
+      "step": 1895
+    },
+    {
+      "epoch": 17.39,
+      "learning_rate": 1.3517241379310345e-05,
+      "loss": 1.8512,
+      "step": 1896
+    },
+    {
+      "epoch": 17.4,
+      "learning_rate": 1.3471264367816091e-05,
+      "loss": 1.5658,
+      "step": 1897
+    },
+    {
+      "epoch": 17.41,
+      "learning_rate": 1.3425287356321839e-05,
+      "loss": 1.722,
+      "step": 1898
+    },
+    {
+      "epoch": 17.42,
+      "learning_rate": 1.3379310344827587e-05,
+      "loss": 1.6313,
+      "step": 1899
+    },
+    {
+      "epoch": 17.43,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 1.6457,
+      "step": 1900
+    },
+    {
+      "epoch": 17.44,
+      "learning_rate": 1.3287356321839081e-05,
+      "loss": 1.9386,
+      "step": 1901
+    },
+    {
+      "epoch": 17.45,
+      "learning_rate": 1.324137931034483e-05,
+      "loss": 1.805,
+      "step": 1902
+    },
+    {
+      "epoch": 17.46,
+      "learning_rate": 1.3195402298850576e-05,
+      "loss": 1.5356,
+      "step": 1903
+    },
+    {
+      "epoch": 17.47,
+      "learning_rate": 1.3149425287356324e-05,
+      "loss": 1.8633,
+      "step": 1904
+    },
+    {
+      "epoch": 17.48,
+      "learning_rate": 1.310344827586207e-05,
+      "loss": 1.9469,
+      "step": 1905
+    },
+    {
+      "epoch": 17.49,
+      "learning_rate": 1.3057471264367815e-05,
+      "loss": 1.6068,
+      "step": 1906
+    },
+    {
+      "epoch": 17.5,
+      "learning_rate": 1.3011494252873563e-05,
+      "loss": 1.762,
+      "step": 1907
+    },
+    {
+      "epoch": 17.5,
+      "learning_rate": 1.296551724137931e-05,
+      "loss": 1.5911,
+      "step": 1908
+    },
+    {
+      "epoch": 17.51,
+      "learning_rate": 1.2919540229885057e-05,
+      "loss": 1.6597,
+      "step": 1909
+    },
+    {
+      "epoch": 17.52,
+      "learning_rate": 1.2873563218390805e-05,
+      "loss": 1.5972,
+      "step": 1910
+    },
+    {
+      "epoch": 17.53,
+      "learning_rate": 1.2827586206896553e-05,
+      "loss": 1.6908,
+      "step": 1911
+    },
+    {
+      "epoch": 17.54,
+      "learning_rate": 1.27816091954023e-05,
+      "loss": 1.8957,
+      "step": 1912
+    },
+    {
+      "epoch": 17.55,
+      "learning_rate": 1.2735632183908047e-05,
+      "loss": 2.0771,
+      "step": 1913
+    },
+    {
+      "epoch": 17.56,
+      "learning_rate": 1.2689655172413795e-05,
+      "loss": 1.7178,
+      "step": 1914
+    },
+    {
+      "epoch": 17.57,
+      "learning_rate": 1.2643678160919542e-05,
+      "loss": 1.4637,
+      "step": 1915
+    },
+    {
+      "epoch": 17.58,
+      "learning_rate": 1.2597701149425286e-05,
+      "loss": 2.2137,
+      "step": 1916
+    },
+    {
+      "epoch": 17.59,
+      "learning_rate": 1.2551724137931034e-05,
+      "loss": 1.8975,
+      "step": 1917
+    },
+    {
+      "epoch": 17.6,
+      "learning_rate": 1.2505747126436782e-05,
+      "loss": 1.7241,
+      "step": 1918
+    },
+    {
+      "epoch": 17.61,
+      "learning_rate": 1.2459770114942529e-05,
+      "loss": 1.8488,
+      "step": 1919
+    },
+    {
+      "epoch": 17.61,
+      "learning_rate": 1.2413793103448277e-05,
+      "loss": 2.1903,
+      "step": 1920
+    },
+    {
+      "epoch": 17.62,
+      "learning_rate": 1.2367816091954023e-05,
+      "loss": 1.4601,
+      "step": 1921
+    },
+    {
+      "epoch": 17.63,
+      "learning_rate": 1.2321839080459771e-05,
+      "loss": 1.7508,
+      "step": 1922
+    },
+    {
+      "epoch": 17.64,
+      "learning_rate": 1.2275862068965519e-05,
+      "loss": 2.6745,
+      "step": 1923
+    },
+    {
+      "epoch": 17.65,
+      "learning_rate": 1.2229885057471265e-05,
+      "loss": 1.8154,
+      "step": 1924
+    },
+    {
+      "epoch": 17.66,
+      "learning_rate": 1.2183908045977012e-05,
+      "loss": 1.8402,
+      "step": 1925
+    },
+    {
+      "epoch": 17.67,
+      "learning_rate": 1.213793103448276e-05,
+      "loss": 1.889,
+      "step": 1926
+    },
+    {
+      "epoch": 17.68,
+      "learning_rate": 1.2091954022988506e-05,
+      "loss": 1.8571,
+      "step": 1927
+    },
+    {
+      "epoch": 17.69,
+      "learning_rate": 1.2045977011494254e-05,
+      "loss": 1.6229,
+      "step": 1928
+    },
+    {
+      "epoch": 17.7,
+      "learning_rate": 1.2e-05,
+      "loss": 1.6628,
+      "step": 1929
+    },
+    {
+      "epoch": 17.71,
+      "learning_rate": 1.1954022988505748e-05,
+      "loss": 2.4438,
+      "step": 1930
+    },
+    {
+      "epoch": 17.72,
+      "learning_rate": 1.1908045977011495e-05,
+      "loss": 1.5868,
+      "step": 1931
+    },
+    {
+      "epoch": 17.72,
+      "learning_rate": 1.1862068965517243e-05,
+      "loss": 1.7507,
+      "step": 1932
+    },
+    {
+      "epoch": 17.73,
+      "learning_rate": 1.1816091954022989e-05,
+      "loss": 2.554,
+      "step": 1933
+    },
+    {
+      "epoch": 17.74,
+      "learning_rate": 1.1770114942528735e-05,
+      "loss": 1.6724,
+      "step": 1934
+    },
+    {
+      "epoch": 17.75,
+      "learning_rate": 1.1724137931034483e-05,
+      "loss": 2.1261,
+      "step": 1935
+    },
+    {
+      "epoch": 17.76,
+      "learning_rate": 1.1678160919540231e-05,
+      "loss": 1.6383,
+      "step": 1936
+    },
+    {
+      "epoch": 17.77,
+      "learning_rate": 1.1632183908045978e-05,
+      "loss": 1.6623,
+      "step": 1937
+    },
+    {
+      "epoch": 17.78,
+      "learning_rate": 1.1586206896551726e-05,
+      "loss": 1.5114,
+      "step": 1938
+    },
+    {
+      "epoch": 17.79,
+      "learning_rate": 1.1540229885057472e-05,
+      "loss": 2.1174,
+      "step": 1939
+    },
+    {
+      "epoch": 17.8,
+      "learning_rate": 1.1494252873563218e-05,
+      "loss": 1.5231,
+      "step": 1940
+    },
+    {
+      "epoch": 17.81,
+      "learning_rate": 1.1448275862068966e-05,
+      "loss": 1.881,
+      "step": 1941
+    },
+    {
+      "epoch": 17.82,
+      "learning_rate": 1.1402298850574714e-05,
+      "loss": 2.4178,
+      "step": 1942
+    },
+    {
+      "epoch": 17.83,
+      "learning_rate": 1.135632183908046e-05,
+      "loss": 1.3992,
+      "step": 1943
+    },
+    {
+      "epoch": 17.83,
+      "learning_rate": 1.1310344827586207e-05,
+      "loss": 1.6854,
+      "step": 1944
+    },
+    {
+      "epoch": 17.84,
+      "learning_rate": 1.1264367816091955e-05,
+      "loss": 1.6252,
+      "step": 1945
+    },
+    {
+      "epoch": 17.85,
+      "learning_rate": 1.1218390804597701e-05,
+      "loss": 2.1357,
+      "step": 1946
+    },
+    {
+      "epoch": 17.86,
+      "learning_rate": 1.117241379310345e-05,
+      "loss": 1.748,
+      "step": 1947
+    },
+    {
+      "epoch": 17.87,
+      "learning_rate": 1.1126436781609197e-05,
+      "loss": 1.9517,
+      "step": 1948
+    },
+    {
+      "epoch": 17.88,
+      "learning_rate": 1.1080459770114942e-05,
+      "loss": 1.7772,
+      "step": 1949
+    },
+    {
+      "epoch": 17.89,
+      "learning_rate": 1.103448275862069e-05,
+      "loss": 1.7389,
+      "step": 1950
+    },
+    {
+      "epoch": 17.9,
+      "learning_rate": 1.0988505747126438e-05,
+      "loss": 1.493,
+      "step": 1951
+    },
+    {
+      "epoch": 17.91,
+      "learning_rate": 1.0942528735632184e-05,
+      "loss": 1.543,
+      "step": 1952
+    },
+    {
+      "epoch": 17.92,
+      "learning_rate": 1.0896551724137932e-05,
+      "loss": 1.5745,
+      "step": 1953
+    },
+    {
+      "epoch": 17.93,
+      "learning_rate": 1.0850574712643679e-05,
+      "loss": 1.6556,
+      "step": 1954
+    },
+    {
+      "epoch": 17.94,
+      "learning_rate": 1.0804597701149425e-05,
+      "loss": 2.241,
+      "step": 1955
+    },
+    {
+      "epoch": 17.94,
+      "learning_rate": 1.0758620689655173e-05,
+      "loss": 1.4161,
+      "step": 1956
+    },
+    {
+      "epoch": 17.95,
+      "learning_rate": 1.0712643678160921e-05,
+      "loss": 1.5301,
+      "step": 1957
+    },
+    {
+      "epoch": 17.96,
+      "learning_rate": 1.0666666666666667e-05,
+      "loss": 1.8564,
+      "step": 1958
+    },
+    {
+      "epoch": 17.97,
+      "learning_rate": 1.0620689655172414e-05,
+      "loss": 1.6702,
+      "step": 1959
+    },
+    {
+      "epoch": 17.98,
+      "learning_rate": 1.0574712643678162e-05,
+      "loss": 2.201,
+      "step": 1960
+    },
+    {
+      "epoch": 17.99,
+      "learning_rate": 1.0528735632183908e-05,
+      "loss": 1.8416,
+      "step": 1961
+    },
+    {
+      "epoch": 18.0,
+      "learning_rate": 1.0482758620689656e-05,
+      "loss": 2.6586,
+      "step": 1962
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": 0.25,
+      "eval_loss": 1.8722608089447021,
+      "eval_runtime": 13.8116,
+      "eval_samples_per_second": 11.874,
+      "eval_steps_per_second": 1.52,
+      "step": 1962
+    },
+    {
+      "epoch": 18.01,
+      "learning_rate": 1.0436781609195404e-05,
+      "loss": 1.981,
+      "step": 1963
+    },
+    {
+      "epoch": 18.02,
+      "learning_rate": 1.039080459770115e-05,
+      "loss": 1.4829,
+      "step": 1964
+    },
+    {
+      "epoch": 18.03,
+      "learning_rate": 1.0344827586206897e-05,
+      "loss": 1.799,
+      "step": 1965
+    },
+    {
+      "epoch": 18.04,
+      "learning_rate": 1.0298850574712645e-05,
+      "loss": 1.7013,
+      "step": 1966
+    },
+    {
+      "epoch": 18.05,
+      "learning_rate": 1.0252873563218391e-05,
+      "loss": 1.8411,
+      "step": 1967
+    },
+    {
+      "epoch": 18.06,
+      "learning_rate": 1.0206896551724139e-05,
+      "loss": 2.2943,
+      "step": 1968
+    },
+    {
+      "epoch": 18.06,
+      "learning_rate": 1.0160919540229885e-05,
+      "loss": 1.863,
+      "step": 1969
+    },
+    {
+      "epoch": 18.07,
+      "learning_rate": 1.0114942528735633e-05,
+      "loss": 1.6908,
+      "step": 1970
+    },
+    {
+      "epoch": 18.08,
+      "learning_rate": 1.006896551724138e-05,
+      "loss": 1.7035,
+      "step": 1971
+    },
+    {
+      "epoch": 18.09,
+      "learning_rate": 1.0022988505747128e-05,
+      "loss": 1.8714,
+      "step": 1972
+    },
+    {
+      "epoch": 18.1,
+      "learning_rate": 9.977011494252874e-06,
+      "loss": 1.9688,
+      "step": 1973
+    },
+    {
+      "epoch": 18.11,
+      "learning_rate": 9.93103448275862e-06,
+      "loss": 1.6745,
+      "step": 1974
+    },
+    {
+      "epoch": 18.12,
+      "learning_rate": 9.885057471264368e-06,
+      "loss": 1.7702,
+      "step": 1975
+    },
+    {
+      "epoch": 18.13,
+      "learning_rate": 9.839080459770116e-06,
+      "loss": 1.7617,
+      "step": 1976
+    },
+    {
+      "epoch": 18.14,
+      "learning_rate": 9.793103448275863e-06,
+      "loss": 1.6504,
+      "step": 1977
+    },
+    {
+      "epoch": 18.15,
+      "learning_rate": 9.747126436781609e-06,
+      "loss": 2.1846,
+      "step": 1978
+    },
+    {
+      "epoch": 18.16,
+      "learning_rate": 9.701149425287357e-06,
+      "loss": 1.8273,
+      "step": 1979
+    },
+    {
+      "epoch": 18.17,
+      "learning_rate": 9.655172413793103e-06,
+      "loss": 1.7038,
+      "step": 1980
+    },
+    {
+      "epoch": 18.17,
+      "learning_rate": 9.609195402298851e-06,
+      "loss": 1.7782,
+      "step": 1981
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 9.5632183908046e-06,
+      "loss": 1.473,
+      "step": 1982
+    },
+    {
+      "epoch": 18.19,
+      "learning_rate": 9.517241379310344e-06,
+      "loss": 1.9904,
+      "step": 1983
+    },
+    {
+      "epoch": 18.2,
+      "learning_rate": 9.471264367816092e-06,
+      "loss": 1.7074,
+      "step": 1984
+    },
+    {
+      "epoch": 18.21,
+      "learning_rate": 9.42528735632184e-06,
+      "loss": 1.5938,
+      "step": 1985
+    },
+    {
+      "epoch": 18.22,
+      "learning_rate": 9.379310344827586e-06,
+      "loss": 2.0194,
+      "step": 1986
+    },
+    {
+      "epoch": 18.23,
+      "learning_rate": 9.333333333333334e-06,
+      "loss": 2.173,
+      "step": 1987
+    },
+    {
+      "epoch": 18.24,
+      "learning_rate": 9.28735632183908e-06,
+      "loss": 2.0861,
+      "step": 1988
+    },
+    {
+      "epoch": 18.25,
+      "learning_rate": 9.241379310344827e-06,
+      "loss": 1.7593,
+      "step": 1989
+    },
+    {
+      "epoch": 18.26,
+      "learning_rate": 9.195402298850575e-06,
+      "loss": 1.8021,
+      "step": 1990
+    },
+    {
+      "epoch": 18.27,
+      "learning_rate": 9.149425287356323e-06,
+      "loss": 1.6424,
+      "step": 1991
+    },
+    {
+      "epoch": 18.28,
+      "learning_rate": 9.10344827586207e-06,
+      "loss": 2.0724,
+      "step": 1992
+    },
+    {
+      "epoch": 18.28,
+      "learning_rate": 9.057471264367816e-06,
+      "loss": 1.763,
+      "step": 1993
+    },
+    {
+      "epoch": 18.29,
+      "learning_rate": 9.011494252873564e-06,
+      "loss": 2.1261,
+      "step": 1994
+    },
+    {
+      "epoch": 18.3,
+      "learning_rate": 8.96551724137931e-06,
+      "loss": 1.6613,
+      "step": 1995
+    },
+    {
+      "epoch": 18.31,
+      "learning_rate": 8.919540229885058e-06,
+      "loss": 1.7075,
+      "step": 1996
+    },
+    {
+      "epoch": 18.32,
+      "learning_rate": 8.873563218390806e-06,
+      "loss": 2.3066,
+      "step": 1997
+    },
+    {
+      "epoch": 18.33,
+      "learning_rate": 8.82758620689655e-06,
+      "loss": 2.0594,
+      "step": 1998
+    },
+    {
+      "epoch": 18.34,
+      "learning_rate": 8.781609195402299e-06,
+      "loss": 1.6063,
+      "step": 1999
+    },
+    {
+      "epoch": 18.35,
+      "learning_rate": 8.735632183908047e-06,
+      "loss": 1.3161,
+      "step": 2000
+    },
+    {
+      "epoch": 18.36,
+      "learning_rate": 8.689655172413793e-06,
+      "loss": 1.7985,
+      "step": 2001
+    },
+    {
+      "epoch": 18.37,
+      "learning_rate": 8.643678160919541e-06,
+      "loss": 2.0872,
+      "step": 2002
+    },
+    {
+      "epoch": 18.38,
+      "learning_rate": 8.597701149425287e-06,
+      "loss": 1.5913,
+      "step": 2003
+    },
+    {
+      "epoch": 18.39,
+      "learning_rate": 8.551724137931035e-06,
+      "loss": 1.7886,
+      "step": 2004
+    },
+    {
+      "epoch": 18.39,
+      "learning_rate": 8.505747126436782e-06,
+      "loss": 1.728,
+      "step": 2005
+    },
+    {
+      "epoch": 18.4,
+      "learning_rate": 8.45977011494253e-06,
+      "loss": 1.7227,
+      "step": 2006
+    },
+    {
+      "epoch": 18.41,
+      "learning_rate": 8.413793103448276e-06,
+      "loss": 1.6143,
+      "step": 2007
+    },
+    {
+      "epoch": 18.42,
+      "learning_rate": 8.367816091954022e-06,
+      "loss": 1.6408,
+      "step": 2008
+    },
+    {
+      "epoch": 18.43,
+      "learning_rate": 8.32183908045977e-06,
+      "loss": 1.673,
+      "step": 2009
+    },
+    {
+      "epoch": 18.44,
+      "learning_rate": 8.275862068965518e-06,
+      "loss": 1.5601,
+      "step": 2010
+    },
+    {
+      "epoch": 18.45,
+      "learning_rate": 8.229885057471265e-06,
+      "loss": 1.6007,
+      "step": 2011
+    },
+    {
+      "epoch": 18.46,
+      "learning_rate": 8.183908045977013e-06,
+      "loss": 1.5991,
+      "step": 2012
+    },
+    {
+      "epoch": 18.47,
+      "learning_rate": 8.137931034482759e-06,
+      "loss": 1.4759,
+      "step": 2013
+    },
+    {
+      "epoch": 18.48,
+      "learning_rate": 8.091954022988505e-06,
+      "loss": 1.4889,
+      "step": 2014
+    },
+    {
+      "epoch": 18.49,
+      "learning_rate": 8.045977011494253e-06,
+      "loss": 1.5286,
+      "step": 2015
+    },
+    {
+      "epoch": 18.5,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.5837,
+      "step": 2016
+    },
+    {
+      "epoch": 18.5,
+      "learning_rate": 7.954022988505748e-06,
+      "loss": 1.7769,
+      "step": 2017
+    },
+    {
+      "epoch": 18.51,
+      "learning_rate": 7.908045977011494e-06,
+      "loss": 1.6331,
+      "step": 2018
+    },
+    {
+      "epoch": 18.52,
+      "learning_rate": 7.862068965517242e-06,
+      "loss": 1.7039,
+      "step": 2019
+    },
+    {
+      "epoch": 18.53,
+      "learning_rate": 7.816091954022988e-06,
+      "loss": 1.7415,
+      "step": 2020
+    },
+    {
+      "epoch": 18.54,
+      "learning_rate": 7.770114942528736e-06,
+      "loss": 1.6737,
+      "step": 2021
+    },
+    {
+      "epoch": 18.55,
+      "learning_rate": 7.724137931034484e-06,
+      "loss": 1.828,
+      "step": 2022
+    },
+    {
+      "epoch": 18.56,
+      "learning_rate": 7.678160919540229e-06,
+      "loss": 1.4221,
+      "step": 2023
+    },
+    {
+      "epoch": 18.57,
+      "learning_rate": 7.632183908045977e-06,
+      "loss": 2.0946,
+      "step": 2024
+    },
+    {
+      "epoch": 18.58,
+      "learning_rate": 7.586206896551724e-06,
+      "loss": 1.6146,
+      "step": 2025
+    },
+    {
+      "epoch": 18.59,
+      "learning_rate": 7.540229885057472e-06,
+      "loss": 1.7017,
+      "step": 2026
+    },
+    {
+      "epoch": 18.6,
+      "learning_rate": 7.494252873563219e-06,
+      "loss": 1.4984,
+      "step": 2027
+    },
+    {
+      "epoch": 18.61,
+      "learning_rate": 7.448275862068965e-06,
+      "loss": 1.8439,
+      "step": 2028
+    },
+    {
+      "epoch": 18.61,
+      "learning_rate": 7.402298850574713e-06,
+      "loss": 2.596,
+      "step": 2029
+    },
+    {
+      "epoch": 18.62,
+      "learning_rate": 7.35632183908046e-06,
+      "loss": 1.7111,
+      "step": 2030
+    },
+    {
+      "epoch": 18.63,
+      "learning_rate": 7.310344827586207e-06,
+      "loss": 1.771,
+      "step": 2031
+    },
+    {
+      "epoch": 18.64,
+      "learning_rate": 7.264367816091955e-06,
+      "loss": 2.1641,
+      "step": 2032
+    },
+    {
+      "epoch": 18.65,
+      "learning_rate": 7.218390804597701e-06,
+      "loss": 1.7231,
+      "step": 2033
+    },
+    {
+      "epoch": 18.66,
+      "learning_rate": 7.172413793103449e-06,
+      "loss": 2.0176,
+      "step": 2034
+    },
+    {
+      "epoch": 18.67,
+      "learning_rate": 7.126436781609196e-06,
+      "loss": 1.6999,
+      "step": 2035
+    },
+    {
+      "epoch": 18.68,
+      "learning_rate": 7.080459770114943e-06,
+      "loss": 1.5221,
+      "step": 2036
+    },
+    {
+      "epoch": 18.69,
+      "learning_rate": 7.03448275862069e-06,
+      "loss": 1.8704,
+      "step": 2037
+    },
+    {
+      "epoch": 18.7,
+      "learning_rate": 6.9885057471264365e-06,
+      "loss": 1.659,
+      "step": 2038
+    },
+    {
+      "epoch": 18.71,
+      "learning_rate": 6.942528735632184e-06,
+      "loss": 1.6603,
+      "step": 2039
+    },
+    {
+      "epoch": 18.72,
+      "learning_rate": 6.896551724137932e-06,
+      "loss": 1.8709,
+      "step": 2040
+    },
+    {
+      "epoch": 18.72,
+      "learning_rate": 6.850574712643679e-06,
+      "loss": 1.6874,
+      "step": 2041
+    },
+    {
+      "epoch": 18.73,
+      "learning_rate": 6.804597701149426e-06,
+      "loss": 1.564,
+      "step": 2042
+    },
+    {
+      "epoch": 18.74,
+      "learning_rate": 6.758620689655172e-06,
+      "loss": 1.9025,
+      "step": 2043
+    },
+    {
+      "epoch": 18.75,
+      "learning_rate": 6.7126436781609195e-06,
+      "loss": 1.6248,
+      "step": 2044
+    },
+    {
+      "epoch": 18.76,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 2.3864,
+      "step": 2045
+    },
+    {
+      "epoch": 18.77,
+      "learning_rate": 6.620689655172415e-06,
+      "loss": 1.6193,
+      "step": 2046
+    },
+    {
+      "epoch": 18.78,
+      "learning_rate": 6.574712643678162e-06,
+      "loss": 1.6766,
+      "step": 2047
+    },
+    {
+      "epoch": 18.79,
+      "learning_rate": 6.528735632183907e-06,
+      "loss": 1.6108,
+      "step": 2048
+    },
+    {
+      "epoch": 18.8,
+      "learning_rate": 6.482758620689655e-06,
+      "loss": 1.9281,
+      "step": 2049
+    },
+    {
+      "epoch": 18.81,
+      "learning_rate": 6.4367816091954025e-06,
+      "loss": 2.2736,
+      "step": 2050
+    },
+    {
+      "epoch": 18.82,
+      "learning_rate": 6.39080459770115e-06,
+      "loss": 1.5257,
+      "step": 2051
+    },
+    {
+      "epoch": 18.83,
+      "learning_rate": 6.344827586206898e-06,
+      "loss": 1.6097,
+      "step": 2052
+    },
+    {
+      "epoch": 18.83,
+      "learning_rate": 6.298850574712643e-06,
+      "loss": 2.4777,
+      "step": 2053
+    },
+    {
+      "epoch": 18.84,
+      "learning_rate": 6.252873563218391e-06,
+      "loss": 2.0103,
+      "step": 2054
+    },
+    {
+      "epoch": 18.85,
+      "learning_rate": 6.206896551724138e-06,
+      "loss": 1.8506,
+      "step": 2055
+    },
+    {
+      "epoch": 18.86,
+      "learning_rate": 6.1609195402298855e-06,
+      "loss": 1.633,
+      "step": 2056
+    },
+    {
+      "epoch": 18.87,
+      "learning_rate": 6.114942528735633e-06,
+      "loss": 1.7209,
+      "step": 2057
+    },
+    {
+      "epoch": 18.88,
+      "learning_rate": 6.06896551724138e-06,
+      "loss": 2.1247,
+      "step": 2058
+    },
+    {
+      "epoch": 18.89,
+      "learning_rate": 6.022988505747127e-06,
+      "loss": 1.873,
+      "step": 2059
+    },
+    {
+      "epoch": 18.9,
+      "learning_rate": 5.977011494252874e-06,
+      "loss": 1.9735,
+      "step": 2060
+    },
+    {
+      "epoch": 18.91,
+      "learning_rate": 5.931034482758621e-06,
+      "loss": 1.709,
+      "step": 2061
+    },
+    {
+      "epoch": 18.92,
+      "learning_rate": 5.885057471264368e-06,
+      "loss": 1.6885,
+      "step": 2062
+    },
+    {
+      "epoch": 18.93,
+      "learning_rate": 5.839080459770116e-06,
+      "loss": 2.2174,
+      "step": 2063
+    },
+    {
+      "epoch": 18.94,
+      "learning_rate": 5.793103448275863e-06,
+      "loss": 2.3934,
+      "step": 2064
+    },
+    {
+      "epoch": 18.94,
+      "learning_rate": 5.747126436781609e-06,
+      "loss": 1.5946,
+      "step": 2065
+    },
+    {
+      "epoch": 18.95,
+      "learning_rate": 5.701149425287357e-06,
+      "loss": 2.3485,
+      "step": 2066
+    },
+    {
+      "epoch": 18.96,
+      "learning_rate": 5.6551724137931035e-06,
+      "loss": 1.4891,
+      "step": 2067
+    },
+    {
+      "epoch": 18.97,
+      "learning_rate": 5.609195402298851e-06,
+      "loss": 1.716,
+      "step": 2068
+    },
+    {
+      "epoch": 18.98,
+      "learning_rate": 5.563218390804599e-06,
+      "loss": 1.6224,
+      "step": 2069
+    },
+    {
+      "epoch": 18.99,
+      "learning_rate": 5.517241379310345e-06,
+      "loss": 1.6514,
+      "step": 2070
+    },
+    {
+      "epoch": 19.0,
+      "learning_rate": 5.471264367816092e-06,
+      "loss": 2.0078,
+      "step": 2071
+    },
+    {
+      "epoch": 19.0,
+      "eval_accuracy": 0.24390243902439024,
+      "eval_loss": 1.8769859075546265,
+      "eval_runtime": 13.8611,
+      "eval_samples_per_second": 11.832,
+      "eval_steps_per_second": 1.515,
+      "step": 2071
+    },
+    {
+      "epoch": 19.01,
+      "learning_rate": 5.425287356321839e-06,
+      "loss": 2.113,
+      "step": 2072
+    },
+    {
+      "epoch": 19.02,
+      "learning_rate": 5.3793103448275865e-06,
+      "loss": 2.0407,
+      "step": 2073
+    },
+    {
+      "epoch": 19.03,
+      "learning_rate": 5.333333333333334e-06,
+      "loss": 1.9523,
+      "step": 2074
+    },
+    {
+      "epoch": 19.04,
+      "learning_rate": 5.287356321839081e-06,
+      "loss": 1.5802,
+      "step": 2075
+    },
+    {
+      "epoch": 19.05,
+      "learning_rate": 5.241379310344828e-06,
+      "loss": 1.8424,
+      "step": 2076
+    },
+    {
+      "epoch": 19.06,
+      "learning_rate": 5.195402298850575e-06,
+      "loss": 2.2139,
+      "step": 2077
+    },
+    {
+      "epoch": 19.06,
+      "learning_rate": 5.149425287356322e-06,
+      "loss": 1.7326,
+      "step": 2078
+    },
+    {
+      "epoch": 19.07,
+      "learning_rate": 5.1034482758620695e-06,
+      "loss": 1.4578,
+      "step": 2079
+    },
+    {
+      "epoch": 19.08,
+      "learning_rate": 5.057471264367817e-06,
+      "loss": 1.8058,
+      "step": 2080
+    },
+    {
+      "epoch": 19.09,
+      "learning_rate": 5.011494252873564e-06,
+      "loss": 1.8672,
+      "step": 2081
+    },
+    {
+      "epoch": 19.1,
+      "learning_rate": 4.96551724137931e-06,
+      "loss": 1.6191,
+      "step": 2082
+    },
+    {
+      "epoch": 19.11,
+      "learning_rate": 4.919540229885058e-06,
+      "loss": 1.8558,
+      "step": 2083
+    },
+    {
+      "epoch": 19.12,
+      "learning_rate": 4.8735632183908045e-06,
+      "loss": 1.8053,
+      "step": 2084
+    },
+    {
+      "epoch": 19.13,
+      "learning_rate": 4.827586206896552e-06,
+      "loss": 1.6794,
+      "step": 2085
+    },
+    {
+      "epoch": 19.14,
+      "learning_rate": 4.7816091954023e-06,
+      "loss": 1.5737,
+      "step": 2086
+    },
+    {
+      "epoch": 19.15,
+      "learning_rate": 4.735632183908046e-06,
+      "loss": 1.7139,
+      "step": 2087
+    },
+    {
+      "epoch": 19.16,
+      "learning_rate": 4.689655172413793e-06,
+      "loss": 1.6437,
+      "step": 2088
+    },
+    {
+      "epoch": 19.17,
+      "learning_rate": 4.64367816091954e-06,
+      "loss": 2.0791,
+      "step": 2089
+    },
+    {
+      "epoch": 19.17,
+      "learning_rate": 4.5977011494252875e-06,
+      "loss": 1.6147,
+      "step": 2090
+    },
+    {
+      "epoch": 19.18,
+      "learning_rate": 4.551724137931035e-06,
+      "loss": 1.9152,
+      "step": 2091
+    },
+    {
+      "epoch": 19.19,
+      "learning_rate": 4.505747126436782e-06,
+      "loss": 1.8205,
+      "step": 2092
+    },
+    {
+      "epoch": 19.2,
+      "learning_rate": 4.459770114942529e-06,
+      "loss": 1.7497,
+      "step": 2093
+    },
+    {
+      "epoch": 19.21,
+      "learning_rate": 4.413793103448275e-06,
+      "loss": 1.734,
+      "step": 2094
+    },
+    {
+      "epoch": 19.22,
+      "learning_rate": 4.367816091954023e-06,
+      "loss": 1.6431,
+      "step": 2095
+    },
+    {
+      "epoch": 19.23,
+      "learning_rate": 4.3218390804597705e-06,
+      "loss": 1.6222,
+      "step": 2096
+    },
+    {
+      "epoch": 19.24,
+      "learning_rate": 4.275862068965518e-06,
+      "loss": 2.2694,
+      "step": 2097
+    },
+    {
+      "epoch": 19.25,
+      "learning_rate": 4.229885057471265e-06,
+      "loss": 1.8006,
+      "step": 2098
+    },
+    {
+      "epoch": 19.26,
+      "learning_rate": 4.183908045977011e-06,
+      "loss": 1.9801,
+      "step": 2099
+    },
+    {
+      "epoch": 19.27,
+      "learning_rate": 4.137931034482759e-06,
+      "loss": 1.8374,
+      "step": 2100
+    },
+    {
+      "epoch": 19.28,
+      "learning_rate": 4.091954022988506e-06,
+      "loss": 1.7391,
+      "step": 2101
+    },
+    {
+      "epoch": 19.28,
+      "learning_rate": 4.045977011494253e-06,
+      "loss": 1.2126,
+      "step": 2102
+    },
+    {
+      "epoch": 19.29,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.6657,
+      "step": 2103
+    },
+    {
+      "epoch": 19.3,
+      "learning_rate": 3.954022988505747e-06,
+      "loss": 1.8836,
+      "step": 2104
+    },
+    {
+      "epoch": 19.31,
+      "learning_rate": 3.908045977011494e-06,
+      "loss": 1.5682,
+      "step": 2105
+    },
+    {
+      "epoch": 19.32,
+      "learning_rate": 3.862068965517242e-06,
+      "loss": 1.8257,
+      "step": 2106
+    },
+    {
+      "epoch": 19.33,
+      "learning_rate": 3.8160919540229885e-06,
+      "loss": 1.5636,
+      "step": 2107
+    },
+    {
+      "epoch": 19.34,
+      "learning_rate": 3.770114942528736e-06,
+      "loss": 1.8752,
+      "step": 2108
+    },
+    {
+      "epoch": 19.35,
+      "learning_rate": 3.7241379310344824e-06,
+      "loss": 1.5951,
+      "step": 2109
+    },
+    {
+      "epoch": 19.36,
+      "learning_rate": 3.67816091954023e-06,
+      "loss": 1.8802,
+      "step": 2110
+    },
+    {
+      "epoch": 19.37,
+      "learning_rate": 3.6321839080459776e-06,
+      "loss": 1.6312,
+      "step": 2111
+    },
+    {
+      "epoch": 19.38,
+      "learning_rate": 3.5862068965517243e-06,
+      "loss": 1.5378,
+      "step": 2112
+    },
+    {
+      "epoch": 19.39,
+      "learning_rate": 3.5402298850574715e-06,
+      "loss": 1.7642,
+      "step": 2113
+    },
+    {
+      "epoch": 19.39,
+      "learning_rate": 3.4942528735632182e-06,
+      "loss": 1.6104,
+      "step": 2114
+    },
+    {
+      "epoch": 19.4,
+      "learning_rate": 3.448275862068966e-06,
+      "loss": 2.3239,
+      "step": 2115
+    },
+    {
+      "epoch": 19.41,
+      "learning_rate": 3.402298850574713e-06,
+      "loss": 2.0949,
+      "step": 2116
+    },
+    {
+      "epoch": 19.42,
+      "learning_rate": 3.3563218390804597e-06,
+      "loss": 1.9261,
+      "step": 2117
+    },
+    {
+      "epoch": 19.43,
+      "learning_rate": 3.3103448275862073e-06,
+      "loss": 1.6624,
+      "step": 2118
+    },
+    {
+      "epoch": 19.44,
+      "learning_rate": 3.2643678160919537e-06,
+      "loss": 1.6051,
+      "step": 2119
+    },
+    {
+      "epoch": 19.45,
+      "learning_rate": 3.2183908045977012e-06,
+      "loss": 2.4215,
+      "step": 2120
+    },
+    {
+      "epoch": 19.46,
+      "learning_rate": 3.172413793103449e-06,
+      "loss": 1.5786,
+      "step": 2121
+    },
+    {
+      "epoch": 19.47,
+      "learning_rate": 3.1264367816091956e-06,
+      "loss": 1.6406,
+      "step": 2122
+    },
+    {
+      "epoch": 19.48,
+      "learning_rate": 3.0804597701149427e-06,
+      "loss": 1.8825,
+      "step": 2123
+    },
+    {
+      "epoch": 19.49,
+      "learning_rate": 3.03448275862069e-06,
+      "loss": 1.7821,
+      "step": 2124
+    },
+    {
+      "epoch": 19.5,
+      "learning_rate": 2.988505747126437e-06,
+      "loss": 1.3542,
+      "step": 2125
+    },
+    {
+      "epoch": 19.5,
+      "learning_rate": 2.942528735632184e-06,
+      "loss": 1.5246,
+      "step": 2126
+    },
+    {
+      "epoch": 19.51,
+      "learning_rate": 2.8965517241379314e-06,
+      "loss": 1.9956,
+      "step": 2127
+    },
+    {
+      "epoch": 19.52,
+      "learning_rate": 2.8505747126436786e-06,
+      "loss": 2.1634,
+      "step": 2128
+    },
+    {
+      "epoch": 19.53,
+      "learning_rate": 2.8045977011494253e-06,
+      "loss": 1.8049,
+      "step": 2129
+    },
+    {
+      "epoch": 19.54,
+      "learning_rate": 2.7586206896551725e-06,
+      "loss": 1.555,
+      "step": 2130
+    },
+    {
+      "epoch": 19.55,
+      "learning_rate": 2.7126436781609197e-06,
+      "loss": 1.5581,
+      "step": 2131
+    },
+    {
+      "epoch": 19.56,
+      "learning_rate": 2.666666666666667e-06,
+      "loss": 1.709,
+      "step": 2132
+    },
+    {
+      "epoch": 19.57,
+      "learning_rate": 2.620689655172414e-06,
+      "loss": 1.8364,
+      "step": 2133
+    },
+    {
+      "epoch": 19.58,
+      "learning_rate": 2.574712643678161e-06,
+      "loss": 1.7524,
+      "step": 2134
+    },
+    {
+      "epoch": 19.59,
+      "learning_rate": 2.5287356321839083e-06,
+      "loss": 2.4427,
+      "step": 2135
+    },
+    {
+      "epoch": 19.6,
+      "learning_rate": 2.482758620689655e-06,
+      "loss": 1.5676,
+      "step": 2136
+    },
+    {
+      "epoch": 19.61,
+      "learning_rate": 2.4367816091954022e-06,
+      "loss": 2.2926,
+      "step": 2137
+    },
+    {
+      "epoch": 19.61,
+      "learning_rate": 2.39080459770115e-06,
+      "loss": 2.0679,
+      "step": 2138
+    },
+    {
+      "epoch": 19.62,
+      "learning_rate": 2.3448275862068966e-06,
+      "loss": 1.6076,
+      "step": 2139
+    },
+    {
+      "epoch": 19.63,
+      "learning_rate": 2.2988505747126437e-06,
+      "loss": 1.4717,
+      "step": 2140
+    },
+    {
+      "epoch": 19.64,
+      "learning_rate": 2.252873563218391e-06,
+      "loss": 1.9429,
+      "step": 2141
+    },
+    {
+      "epoch": 19.65,
+      "learning_rate": 2.2068965517241377e-06,
+      "loss": 1.9751,
+      "step": 2142
+    },
+    {
+      "epoch": 19.66,
+      "learning_rate": 2.1609195402298852e-06,
+      "loss": 1.5581,
+      "step": 2143
+    },
+    {
+      "epoch": 19.67,
+      "learning_rate": 2.1149425287356324e-06,
+      "loss": 1.6296,
+      "step": 2144
+    },
+    {
+      "epoch": 19.68,
+      "learning_rate": 2.0689655172413796e-06,
+      "loss": 2.6944,
+      "step": 2145
+    },
+    {
+      "epoch": 19.69,
+      "learning_rate": 2.0229885057471263e-06,
+      "loss": 1.5285,
+      "step": 2146
+    },
+    {
+      "epoch": 19.7,
+      "learning_rate": 1.9770114942528735e-06,
+      "loss": 1.6525,
+      "step": 2147
+    },
+    {
+      "epoch": 19.71,
+      "learning_rate": 1.931034482758621e-06,
+      "loss": 2.1895,
+      "step": 2148
+    },
+    {
+      "epoch": 19.72,
+      "learning_rate": 1.885057471264368e-06,
+      "loss": 2.2881,
+      "step": 2149
+    },
+    {
+      "epoch": 19.72,
+      "learning_rate": 1.839080459770115e-06,
+      "loss": 1.3332,
+      "step": 2150
+    },
+    {
+      "epoch": 19.73,
+      "learning_rate": 1.7931034482758622e-06,
+      "loss": 1.8159,
+      "step": 2151
+    },
+    {
+      "epoch": 19.74,
+      "learning_rate": 1.7471264367816091e-06,
+      "loss": 2.2511,
+      "step": 2152
+    },
+    {
+      "epoch": 19.75,
+      "learning_rate": 1.7011494252873565e-06,
+      "loss": 1.744,
+      "step": 2153
+    },
+    {
+      "epoch": 19.76,
+      "learning_rate": 1.6551724137931037e-06,
+      "loss": 1.7443,
+      "step": 2154
+    },
+    {
+      "epoch": 19.77,
+      "learning_rate": 1.6091954022988506e-06,
+      "loss": 1.4425,
+      "step": 2155
+    },
+    {
+      "epoch": 19.78,
+      "learning_rate": 1.5632183908045978e-06,
+      "loss": 1.8097,
+      "step": 2156
+    },
+    {
+      "epoch": 19.79,
+      "learning_rate": 1.517241379310345e-06,
+      "loss": 1.7868,
+      "step": 2157
+    },
+    {
+      "epoch": 19.8,
+      "learning_rate": 1.471264367816092e-06,
+      "loss": 1.728,
+      "step": 2158
+    },
+    {
+      "epoch": 19.81,
+      "learning_rate": 1.4252873563218393e-06,
+      "loss": 1.5975,
+      "step": 2159
+    },
+    {
+      "epoch": 19.82,
+      "learning_rate": 1.3793103448275862e-06,
+      "loss": 1.7236,
+      "step": 2160
+    },
+    {
+      "epoch": 19.83,
+      "learning_rate": 1.3333333333333334e-06,
+      "loss": 1.8273,
+      "step": 2161
+    },
+    {
+      "epoch": 19.83,
+      "learning_rate": 1.2873563218390806e-06,
+      "loss": 1.5047,
+      "step": 2162
+    },
+    {
+      "epoch": 19.84,
+      "learning_rate": 1.2413793103448275e-06,
+      "loss": 2.2744,
+      "step": 2163
+    },
+    {
+      "epoch": 19.85,
+      "learning_rate": 1.195402298850575e-06,
+      "loss": 2.432,
+      "step": 2164
+    },
+    {
+      "epoch": 19.86,
+      "learning_rate": 1.1494252873563219e-06,
+      "loss": 1.8094,
+      "step": 2165
+    },
+    {
+      "epoch": 19.87,
+      "learning_rate": 1.1034482758620688e-06,
+      "loss": 1.8114,
+      "step": 2166
+    },
+    {
+      "epoch": 19.88,
+      "learning_rate": 1.0574712643678162e-06,
+      "loss": 1.8494,
+      "step": 2167
+    },
+    {
+      "epoch": 19.89,
+      "learning_rate": 1.0114942528735632e-06,
+      "loss": 1.4665,
+      "step": 2168
+    },
+    {
+      "epoch": 19.9,
+      "learning_rate": 9.655172413793105e-07,
+      "loss": 1.5801,
+      "step": 2169
+    },
+    {
+      "epoch": 19.91,
+      "learning_rate": 9.195402298850575e-07,
+      "loss": 1.8141,
+      "step": 2170
+    },
+    {
+      "epoch": 19.92,
+      "learning_rate": 8.735632183908046e-07,
+      "loss": 1.9564,
+      "step": 2171
+    },
+    {
+      "epoch": 19.93,
+      "learning_rate": 8.275862068965518e-07,
+      "loss": 1.7217,
+      "step": 2172
+    },
+    {
+      "epoch": 19.94,
+      "learning_rate": 7.816091954022989e-07,
+      "loss": 1.6777,
+      "step": 2173
+    },
+    {
+      "epoch": 19.94,
+      "learning_rate": 7.35632183908046e-07,
+      "loss": 1.5667,
+      "step": 2174
+    },
+    {
+      "epoch": 19.95,
+      "learning_rate": 6.896551724137931e-07,
+      "loss": 1.6799,
+      "step": 2175
+    },
+    {
+      "epoch": 19.96,
+      "learning_rate": 6.436781609195403e-07,
+      "loss": 1.6877,
+      "step": 2176
+    },
+    {
+      "epoch": 19.97,
+      "learning_rate": 5.977011494252875e-07,
+      "loss": 1.9862,
+      "step": 2177
+    },
+    {
+      "epoch": 19.98,
+      "learning_rate": 5.517241379310344e-07,
+      "loss": 1.658,
+      "step": 2178
+    },
+    {
+      "epoch": 19.99,
+      "learning_rate": 5.057471264367816e-07,
+      "loss": 2.0682,
+      "step": 2179
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 4.5977011494252875e-07,
+      "loss": 1.4072,
+      "step": 2180
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": 0.23780487804878048,
+      "eval_loss": 1.886708378791809,
+      "eval_runtime": 13.8709,
+      "eval_samples_per_second": 11.823,
+      "eval_steps_per_second": 1.514,
+      "step": 2180
+    },
+    {
+      "epoch": 20.0,
+      "step": 2180,
+      "total_flos": 1.2663415269359616e+17,
+      "train_loss": 1.7466916341539105,
+      "train_runtime": 4318.1089,
+      "train_samples_per_second": 3.024,
+      "train_steps_per_second": 0.505
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2180,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "total_flos": 1.2663415269359616e+17,
+  "trial_name": null,
+  "trial_params": null
+}