diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,5781 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 20,
+  "global_step": 900,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 5e-06,
+      "loss": 1.0235,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1e-05,
+      "loss": 1.0135,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.5e-05,
+      "loss": 1.0064,
+      "step": 3
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2e-05,
+      "loss": 1.0729,
+      "step": 4
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.5e-05,
+      "loss": 1.0523,
+      "step": 5
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3e-05,
+      "loss": 0.9615,
+      "step": 6
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3.5e-05,
+      "loss": 1.0731,
+      "step": 7
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4e-05,
+      "loss": 1.0445,
+      "step": 8
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.5e-05,
+      "loss": 1.0198,
+      "step": 9
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5e-05,
+      "loss": 0.9868,
+      "step": 10
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.500000000000001e-05,
+      "loss": 1.0009,
+      "step": 11
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 6e-05,
+      "loss": 1.0435,
+      "step": 12
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 6.500000000000001e-05,
+      "loss": 0.9577,
+      "step": 13
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7e-05,
+      "loss": 0.9346,
+      "step": 14
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 1.0129,
+      "step": 15
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8e-05,
+      "loss": 1.0224,
+      "step": 16
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 8.5e-05,
+      "loss": 0.9113,
+      "step": 17
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9e-05,
+      "loss": 0.8728,
+      "step": 18
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.5e-05,
+      "loss": 0.9496,
+      "step": 19
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001,
+      "loss": 0.9233,
+      "step": 20
+    },
+    {
+      "epoch": 0.09,
+      "eval_loss": 0.9231593012809753,
+      "eval_runtime": 15.9254,
+      "eval_samples_per_second": 5.086,
+      "eval_steps_per_second": 1.319,
+      "step": 20
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001,
+      "loss": 0.8718,
+      "step": 21
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001,
+      "loss": 0.9282,
+      "step": 22
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001,
+      "loss": 0.9457,
+      "step": 23
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001,
+      "loss": 0.9121,
+      "step": 24
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001,
+      "loss": 0.9357,
+      "step": 25
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001,
+      "loss": 0.9325,
+      "step": 26
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001,
+      "loss": 0.8547,
+      "step": 27
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001,
+      "loss": 0.8566,
+      "step": 28
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001,
+      "loss": 0.9179,
+      "step": 29
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001,
+      "loss": 0.8754,
+      "step": 30
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001,
+      "loss": 0.9256,
+      "step": 31
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001,
+      "loss": 0.8349,
+      "step": 32
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001,
+      "loss": 0.9279,
+      "step": 33
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001,
+      "loss": 0.8994,
+      "step": 34
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001,
+      "loss": 0.9025,
+      "step": 35
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001,
+      "loss": 0.8652,
+      "step": 36
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001,
+      "loss": 0.8222,
+      "step": 37
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001,
+      "loss": 0.8097,
+      "step": 38
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001,
+      "loss": 0.8703,
+      "step": 39
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001,
+      "loss": 0.8543,
+      "step": 40
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 0.8435553908348083,
+      "eval_runtime": 16.1057,
+      "eval_samples_per_second": 5.029,
+      "eval_steps_per_second": 1.304,
+      "step": 40
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001,
+      "loss": 0.8032,
+      "step": 41
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001,
+      "loss": 0.8572,
+      "step": 42
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.0001,
+      "loss": 0.8282,
+      "step": 43
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001,
+      "loss": 0.8609,
+      "step": 44
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001,
+      "loss": 0.8895,
+      "step": 45
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001,
+      "loss": 0.8384,
+      "step": 46
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.0001,
+      "loss": 0.7752,
+      "step": 47
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.0001,
+      "loss": 0.8397,
+      "step": 48
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001,
+      "loss": 0.8508,
+      "step": 49
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001,
+      "loss": 0.8815,
+      "step": 50
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001,
+      "loss": 0.8748,
+      "step": 51
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001,
+      "loss": 0.849,
+      "step": 52
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001,
+      "loss": 0.8635,
+      "step": 53
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001,
+      "loss": 0.7904,
+      "step": 54
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001,
+      "loss": 0.8354,
+      "step": 55
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001,
+      "loss": 0.8526,
+      "step": 56
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001,
+      "loss": 0.8978,
+      "step": 57
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001,
+      "loss": 0.8748,
+      "step": 58
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.0001,
+      "loss": 0.8604,
+      "step": 59
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.0001,
+      "loss": 0.8139,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "eval_loss": 0.8141185641288757,
+      "eval_runtime": 16.2813,
+      "eval_samples_per_second": 4.975,
+      "eval_steps_per_second": 1.29,
+      "step": 60
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.0001,
+      "loss": 0.7797,
+      "step": 61
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001,
+      "loss": 0.7715,
+      "step": 62
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001,
+      "loss": 0.8165,
+      "step": 63
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001,
+      "loss": 0.8253,
+      "step": 64
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001,
+      "loss": 0.8235,
+      "step": 65
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001,
+      "loss": 0.779,
+      "step": 66
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001,
+      "loss": 0.8127,
+      "step": 67
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001,
+      "loss": 0.7749,
+      "step": 68
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001,
+      "loss": 0.7512,
+      "step": 69
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001,
+      "loss": 0.7847,
+      "step": 70
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001,
+      "loss": 0.83,
+      "step": 71
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001,
+      "loss": 0.7835,
+      "step": 72
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001,
+      "loss": 0.8435,
+      "step": 73
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001,
+      "loss": 0.8983,
+      "step": 74
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.0001,
+      "loss": 0.7755,
+      "step": 75
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001,
+      "loss": 0.8256,
+      "step": 76
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001,
+      "loss": 0.8179,
+      "step": 77
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001,
+      "loss": 0.8223,
+      "step": 78
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.0001,
+      "loss": 0.7997,
+      "step": 79
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001,
+      "loss": 0.8398,
+      "step": 80
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 0.7990346550941467,
+      "eval_runtime": 16.4981,
+      "eval_samples_per_second": 4.91,
+      "eval_steps_per_second": 1.273,
+      "step": 80
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001,
+      "loss": 0.7612,
+      "step": 81
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001,
+      "loss": 0.8065,
+      "step": 82
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001,
+      "loss": 0.7967,
+      "step": 83
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001,
+      "loss": 0.8038,
+      "step": 84
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001,
+      "loss": 0.8354,
+      "step": 85
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001,
+      "loss": 0.8315,
+      "step": 86
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001,
+      "loss": 0.8988,
+      "step": 87
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.0001,
+      "loss": 0.729,
+      "step": 88
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0001,
+      "loss": 0.8303,
+      "step": 89
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0001,
+      "loss": 0.8048,
+      "step": 90
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0001,
+      "loss": 0.7968,
+      "step": 91
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.0001,
+      "loss": 0.7863,
+      "step": 92
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.0001,
+      "loss": 0.8174,
+      "step": 93
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001,
+      "loss": 0.7449,
+      "step": 94
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.0001,
+      "loss": 0.8019,
+      "step": 95
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001,
+      "loss": 0.7778,
+      "step": 96
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001,
+      "loss": 0.8435,
+      "step": 97
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001,
+      "loss": 0.8459,
+      "step": 98
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001,
+      "loss": 0.7166,
+      "step": 99
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001,
+      "loss": 0.8702,
+      "step": 100
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 0.7921832203865051,
+      "eval_runtime": 16.9075,
+      "eval_samples_per_second": 4.791,
+      "eval_steps_per_second": 1.242,
+      "step": 100
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7872,
+      "step": 101
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7837,
+      "step": 102
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001,
+      "loss": 0.8252,
+      "step": 103
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.0001,
+      "loss": 0.7955,
+      "step": 104
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7678,
+      "step": 105
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.0001,
+      "loss": 0.8488,
+      "step": 106
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001,
+      "loss": 0.76,
+      "step": 107
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001,
+      "loss": 0.8094,
+      "step": 108
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.0001,
+      "loss": 0.8204,
+      "step": 109
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001,
+      "loss": 0.898,
+      "step": 110
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.0001,
+      "loss": 0.8026,
+      "step": 111
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001,
+      "loss": 0.7875,
+      "step": 112
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001,
+      "loss": 0.8613,
+      "step": 113
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001,
+      "loss": 0.8025,
+      "step": 114
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.0001,
+      "loss": 0.8118,
+      "step": 115
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001,
+      "loss": 0.8256,
+      "step": 116
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001,
+      "loss": 0.7432,
+      "step": 117
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 0.0001,
+      "loss": 0.8224,
+      "step": 118
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001,
+      "loss": 0.6951,
+      "step": 119
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001,
+      "loss": 0.7556,
+      "step": 120
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 0.7822813987731934,
+      "eval_runtime": 17.0173,
+      "eval_samples_per_second": 4.76,
+      "eval_steps_per_second": 1.234,
+      "step": 120
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001,
+      "loss": 0.7782,
+      "step": 121
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 0.0001,
+      "loss": 0.7315,
+      "step": 122
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.0001,
+      "loss": 0.7938,
+      "step": 123
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 0.0001,
+      "loss": 0.7419,
+      "step": 124
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001,
+      "loss": 0.7177,
+      "step": 125
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001,
+      "loss": 0.7828,
+      "step": 126
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.0001,
+      "loss": 0.8034,
+      "step": 127
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001,
+      "loss": 0.7335,
+      "step": 128
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 0.0001,
+      "loss": 0.7807,
+      "step": 129
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001,
+      "loss": 0.7476,
+      "step": 130
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 0.0001,
+      "loss": 0.7756,
+      "step": 131
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.0001,
+      "loss": 0.7472,
+      "step": 132
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 0.0001,
+      "loss": 0.7191,
+      "step": 133
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0001,
+      "loss": 0.7797,
+      "step": 134
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0001,
+      "loss": 0.7729,
+      "step": 135
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 0.0001,
+      "loss": 0.7858,
+      "step": 136
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001,
+      "loss": 0.8139,
+      "step": 137
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 0.0001,
+      "loss": 0.8744,
+      "step": 138
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001,
+      "loss": 0.7423,
+      "step": 139
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.0001,
+      "loss": 0.7628,
+      "step": 140
+    },
+    {
+      "epoch": 0.62,
+      "eval_loss": 0.7766249179840088,
+      "eval_runtime": 16.9687,
+      "eval_samples_per_second": 4.774,
+      "eval_steps_per_second": 1.238,
+      "step": 140
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.0001,
+      "loss": 0.822,
+      "step": 141
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 0.0001,
+      "loss": 0.761,
+      "step": 142
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001,
+      "loss": 0.8807,
+      "step": 143
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001,
+      "loss": 0.8208,
+      "step": 144
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 0.0001,
+      "loss": 0.7718,
+      "step": 145
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001,
+      "loss": 0.7358,
+      "step": 146
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 0.0001,
+      "loss": 0.8005,
+      "step": 147
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001,
+      "loss": 0.7668,
+      "step": 148
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 0.0001,
+      "loss": 0.768,
+      "step": 149
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.0001,
+      "loss": 0.7511,
+      "step": 150
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.0001,
+      "loss": 0.7509,
+      "step": 151
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.0001,
+      "loss": 0.8193,
+      "step": 152
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.0001,
+      "loss": 0.8586,
+      "step": 153
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7708,
+      "step": 154
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.0001,
+      "loss": 0.7745,
+      "step": 155
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 0.0001,
+      "loss": 0.7466,
+      "step": 156
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.0001,
+      "loss": 0.7983,
+      "step": 157
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 0.0001,
+      "loss": 0.8352,
+      "step": 158
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.0001,
+      "loss": 0.7518,
+      "step": 159
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 0.0001,
+      "loss": 0.778,
+      "step": 160
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 0.773330807685852,
+      "eval_runtime": 16.9837,
+      "eval_samples_per_second": 4.769,
+      "eval_steps_per_second": 1.236,
+      "step": 160
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0001,
+      "loss": 0.8105,
+      "step": 161
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0001,
+      "loss": 0.8049,
+      "step": 162
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 0.0001,
+      "loss": 0.7976,
+      "step": 163
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001,
+      "loss": 0.7693,
+      "step": 164
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 0.0001,
+      "loss": 0.7313,
+      "step": 165
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.0001,
+      "loss": 0.7604,
+      "step": 166
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.0001,
+      "loss": 0.733,
+      "step": 167
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.0001,
+      "loss": 0.7193,
+      "step": 168
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 0.0001,
+      "loss": 0.7506,
+      "step": 169
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.0001,
+      "loss": 0.7268,
+      "step": 170
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.0001,
+      "loss": 0.7561,
+      "step": 171
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 0.0001,
+      "loss": 0.7508,
+      "step": 172
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.0001,
+      "loss": 0.7854,
+      "step": 173
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 0.0001,
+      "loss": 0.7944,
+      "step": 174
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001,
+      "loss": 0.7924,
+      "step": 175
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0001,
+      "loss": 0.7462,
+      "step": 176
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001,
+      "loss": 0.7869,
+      "step": 177
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 0.0001,
+      "loss": 0.8291,
+      "step": 178
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.0001,
+      "loss": 0.7804,
+      "step": 179
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.0001,
+      "loss": 0.7769,
+      "step": 180
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 0.7669081687927246,
+      "eval_runtime": 17.2984,
+      "eval_samples_per_second": 4.683,
+      "eval_steps_per_second": 1.214,
+      "step": 180
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.0001,
+      "loss": 0.841,
+      "step": 181
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.0001,
+      "loss": 0.706,
+      "step": 182
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 0.0001,
+      "loss": 0.8322,
+      "step": 183
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.0001,
+      "loss": 0.7529,
+      "step": 184
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 0.0001,
+      "loss": 0.8146,
+      "step": 185
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.0001,
+      "loss": 0.7243,
+      "step": 186
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 0.0001,
+      "loss": 0.7634,
+      "step": 187
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001,
+      "loss": 0.7147,
+      "step": 188
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001,
+      "loss": 0.714,
+      "step": 189
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 0.0001,
+      "loss": 0.7562,
+      "step": 190
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001,
+      "loss": 0.7708,
+      "step": 191
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 0.0001,
+      "loss": 0.7409,
+      "step": 192
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.0001,
+      "loss": 0.7815,
+      "step": 193
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 0.0001,
+      "loss": 0.704,
+      "step": 194
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.0001,
+      "loss": 0.7501,
+      "step": 195
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.0001,
+      "loss": 0.7045,
+      "step": 196
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.0001,
+      "loss": 0.7915,
+      "step": 197
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.0001,
+      "loss": 0.8724,
+      "step": 198
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 0.0001,
+      "loss": 0.7284,
+      "step": 199
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.0001,
+      "loss": 0.7684,
+      "step": 200
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 0.7664858102798462,
+      "eval_runtime": 17.143,
+      "eval_samples_per_second": 4.725,
+      "eval_steps_per_second": 1.225,
+      "step": 200
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.0001,
+      "loss": 0.8412,
+      "step": 201
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.0001,
+      "loss": 0.8131,
+      "step": 202
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 0.0001,
+      "loss": 0.7809,
+      "step": 203
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.0001,
+      "loss": 0.8583,
+      "step": 204
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 0.0001,
+      "loss": 0.7514,
+      "step": 205
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001,
+      "loss": 0.7778,
+      "step": 206
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001,
+      "loss": 0.7876,
+      "step": 207
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 0.0001,
+      "loss": 0.7057,
+      "step": 208
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.0001,
+      "loss": 0.6806,
+      "step": 209
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.0001,
+      "loss": 0.7065,
+      "step": 210
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.0001,
+      "loss": 0.7371,
+      "step": 211
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 0.0001,
+      "loss": 0.7292,
+      "step": 212
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.0001,
+      "loss": 0.7259,
+      "step": 213
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 0.0001,
+      "loss": 0.7409,
+      "step": 214
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.0001,
+      "loss": 0.7547,
+      "step": 215
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.0001,
+      "loss": 0.8032,
+      "step": 216
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 0.0001,
+      "loss": 0.7809,
+      "step": 217
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001,
+      "loss": 0.6974,
+      "step": 218
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 0.0001,
+      "loss": 0.7896,
+      "step": 219
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001,
+      "loss": 0.7194,
+      "step": 220
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 0.7614033818244934,
+      "eval_runtime": 16.9073,
+      "eval_samples_per_second": 4.791,
+      "eval_steps_per_second": 1.242,
+      "step": 220
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 0.0001,
+      "loss": 0.741,
+      "step": 221
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001,
+      "loss": 0.8645,
+      "step": 222
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.0001,
+      "loss": 0.8182,
+      "step": 223
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0001,
+      "loss": 0.7674,
+      "step": 224
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0001,
+      "loss": 0.8312,
+      "step": 225
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0001,
+      "loss": 0.8204,
+      "step": 226
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.0001,
+      "loss": 0.7917,
+      "step": 227
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 0.0001,
+      "loss": 0.7615,
+      "step": 228
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.0001,
+      "loss": 0.7588,
+      "step": 229
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 0.0001,
+      "loss": 0.7964,
+      "step": 230
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.0001,
+      "loss": 0.7347,
+      "step": 231
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 0.0001,
+      "loss": 0.7586,
+      "step": 232
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.0001,
+      "loss": 0.7572,
+      "step": 233
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.0001,
+      "loss": 0.7978,
+      "step": 234
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 0.0001,
+      "loss": 0.7628,
+      "step": 235
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001,
+      "loss": 0.8498,
+      "step": 236
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 0.0001,
+      "loss": 0.7301,
+      "step": 237
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.0001,
+      "loss": 0.7103,
+      "step": 238
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 0.0001,
+      "loss": 0.7977,
+      "step": 239
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.0001,
+      "loss": 0.7463,
+      "step": 240
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 0.7582252621650696,
+      "eval_runtime": 17.2276,
+      "eval_samples_per_second": 4.702,
+      "eval_steps_per_second": 1.219,
+      "step": 240
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.0001,
+      "loss": 0.6327,
+      "step": 241
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.0001,
+      "loss": 0.8081,
+      "step": 242
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.0001,
+      "loss": 0.7364,
+      "step": 243
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 0.0001,
+      "loss": 0.6608,
+      "step": 244
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.0001,
+      "loss": 0.6498,
+      "step": 245
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 0.0001,
+      "loss": 0.6468,
+      "step": 246
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.0001,
+      "loss": 0.7269,
+      "step": 247
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 0.0001,
+      "loss": 0.7557,
+      "step": 248
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.0001,
+      "loss": 0.7812,
+      "step": 249
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.0001,
+      "loss": 0.6935,
+      "step": 250
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001,
+      "loss": 0.7123,
+      "step": 251
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6821,
+      "step": 252
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6922,
+      "step": 253
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001,
+      "loss": 0.7613,
+      "step": 254
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 0.0001,
+      "loss": 0.6746,
+      "step": 255
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.0001,
+      "loss": 0.7271,
+      "step": 256
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 0.0001,
+      "loss": 0.7275,
+      "step": 257
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.0001,
+      "loss": 0.6947,
+      "step": 258
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 0.0001,
+      "loss": 0.8417,
+      "step": 259
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.0001,
+      "loss": 0.7855,
+      "step": 260
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 0.7556370496749878,
+      "eval_runtime": 17.1316,
+      "eval_samples_per_second": 4.728,
+      "eval_steps_per_second": 1.226,
+      "step": 260
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.0001,
+      "loss": 0.6919,
+      "step": 261
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 0.0001,
+      "loss": 0.7388,
+      "step": 262
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.0001,
+      "loss": 0.672,
+      "step": 263
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 0.0001,
+      "loss": 0.7528,
+      "step": 264
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001,
+      "loss": 0.7236,
+      "step": 265
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 0.0001,
+      "loss": 0.6413,
+      "step": 266
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.0001,
+      "loss": 0.718,
+      "step": 267
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 0.0001,
+      "loss": 0.7429,
+      "step": 268
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.0001,
+      "loss": 0.7222,
+      "step": 269
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.0001,
+      "loss": 0.7373,
+      "step": 270
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.0001,
+      "loss": 0.7616,
+      "step": 271
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.0001,
+      "loss": 0.7173,
+      "step": 272
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 0.0001,
+      "loss": 0.6467,
+      "step": 273
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.0001,
+      "loss": 0.728,
+      "step": 274
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 0.0001,
+      "loss": 0.7671,
+      "step": 275
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.0001,
+      "loss": 0.8053,
+      "step": 276
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.0001,
+      "loss": 0.7396,
+      "step": 277
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001,
+      "loss": 0.6888,
+      "step": 278
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001,
+      "loss": 0.7664,
+      "step": 279
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 0.0001,
+      "loss": 0.8129,
+      "step": 280
+    },
+    {
+      "epoch": 1.24,
+      "eval_loss": 0.7567616105079651,
+      "eval_runtime": 17.0165,
+      "eval_samples_per_second": 4.76,
+      "eval_steps_per_second": 1.234,
+      "step": 280
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.0001,
+      "loss": 0.772,
+      "step": 281
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 0.0001,
+      "loss": 0.7414,
+      "step": 282
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.0001,
+      "loss": 0.7488,
+      "step": 283
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 0.0001,
+      "loss": 0.8666,
+      "step": 284
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.0001,
+      "loss": 0.8907,
+      "step": 285
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 0.0001,
+      "loss": 0.8586,
+      "step": 286
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.0001,
+      "loss": 0.7379,
+      "step": 287
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.0001,
+      "loss": 0.6713,
+      "step": 288
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 0.0001,
+      "loss": 0.7425,
+      "step": 289
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0001,
+      "loss": 0.7629,
+      "step": 290
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 0.0001,
+      "loss": 0.73,
+      "step": 291
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001,
+      "loss": 0.7305,
+      "step": 292
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 0.0001,
+      "loss": 0.7231,
+      "step": 293
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001,
+      "loss": 0.7598,
+      "step": 294
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 0.0001,
+      "loss": 0.6481,
+      "step": 295
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.0001,
+      "loss": 0.7588,
+      "step": 296
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.0001,
+      "loss": 0.649,
+      "step": 297
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 0.0001,
+      "loss": 0.7817,
+      "step": 298
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.0001,
+      "loss": 0.7339,
+      "step": 299
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.0001,
+      "loss": 0.8407,
+      "step": 300
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 0.7525928616523743,
+      "eval_runtime": 16.7573,
+      "eval_samples_per_second": 4.834,
+      "eval_steps_per_second": 1.253,
+      "step": 300
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.0001,
+      "loss": 0.768,
+      "step": 301
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.0001,
+      "loss": 0.7663,
+      "step": 302
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.0001,
+      "loss": 0.732,
+      "step": 303
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 0.0001,
+      "loss": 0.7703,
+      "step": 304
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.0001,
+      "loss": 0.737,
+      "step": 305
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.0001,
+      "loss": 0.7285,
+      "step": 306
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.0001,
+      "loss": 0.7678,
+      "step": 307
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001,
+      "loss": 0.6982,
+      "step": 308
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 0.0001,
+      "loss": 0.6761,
+      "step": 309
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001,
+      "loss": 0.7661,
+      "step": 310
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 0.0001,
+      "loss": 0.7322,
+      "step": 311
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001,
+      "loss": 0.6654,
+      "step": 312
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 0.0001,
+      "loss": 0.7539,
+      "step": 313
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.0001,
+      "loss": 0.6856,
+      "step": 314
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.0001,
+      "loss": 0.7662,
+      "step": 315
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 0.0001,
+      "loss": 0.6932,
+      "step": 316
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.0001,
+      "loss": 0.743,
+      "step": 317
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 0.0001,
+      "loss": 0.7537,
+      "step": 318
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.0001,
+      "loss": 0.6932,
+      "step": 319
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 0.0001,
+      "loss": 0.7703,
+      "step": 320
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 0.7526760101318359,
+      "eval_runtime": 16.8167,
+      "eval_samples_per_second": 4.817,
+      "eval_steps_per_second": 1.249,
+      "step": 320
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.0001,
+      "loss": 0.662,
+      "step": 321
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 0.0001,
+      "loss": 0.7589,
+      "step": 322
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001,
+      "loss": 0.6523,
+      "step": 323
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001,
+      "loss": 0.6995,
+      "step": 324
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 0.0001,
+      "loss": 0.692,
+      "step": 325
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7513,
+      "step": 326
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7594,
+      "step": 327
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.0001,
+      "loss": 0.7783,
+      "step": 328
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 0.0001,
+      "loss": 0.7287,
+      "step": 329
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7738,
+      "step": 330
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7922,
+      "step": 331
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.0001,
+      "loss": 0.7355,
+      "step": 332
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.0001,
+      "loss": 0.7133,
+      "step": 333
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 0.0001,
+      "loss": 0.6896,
+      "step": 334
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.0001,
+      "loss": 0.7258,
+      "step": 335
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 0.0001,
+      "loss": 0.7496,
+      "step": 336
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001,
+      "loss": 0.7564,
+      "step": 337
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 0.0001,
+      "loss": 0.7179,
+      "step": 338
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.0001,
+      "loss": 0.7005,
+      "step": 339
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 0.0001,
+      "loss": 0.721,
+      "step": 340
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 0.7456800937652588,
+      "eval_runtime": 16.7294,
+      "eval_samples_per_second": 4.842,
+      "eval_steps_per_second": 1.255,
+      "step": 340
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.0001,
+      "loss": 0.7219,
+      "step": 341
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6403,
+      "step": 342
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6903,
+      "step": 343
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.0001,
+      "loss": 0.6609,
+      "step": 344
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 0.0001,
+      "loss": 0.698,
+      "step": 345
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.0001,
+      "loss": 0.6995,
+      "step": 346
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 0.0001,
+      "loss": 0.6995,
+      "step": 347
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0001,
+      "loss": 0.6153,
+      "step": 348
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 0.0001,
+      "loss": 0.7042,
+      "step": 349
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001,
+      "loss": 0.7039,
+      "step": 350
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001,
+      "loss": 0.8076,
+      "step": 351
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0001,
+      "loss": 0.6823,
+      "step": 352
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.0001,
+      "loss": 0.7132,
+      "step": 353
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 0.0001,
+      "loss": 0.7476,
+      "step": 354
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.0001,
+      "loss": 0.6946,
+      "step": 355
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 0.0001,
+      "loss": 0.6729,
+      "step": 356
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.0001,
+      "loss": 0.7566,
+      "step": 357
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 0.0001,
+      "loss": 0.7116,
+      "step": 358
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.0001,
+      "loss": 0.792,
+      "step": 359
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.0001,
+      "loss": 0.7415,
+      "step": 360
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 0.7429763078689575,
+      "eval_runtime": 16.6571,
+      "eval_samples_per_second": 4.863,
+      "eval_steps_per_second": 1.261,
+      "step": 360
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.0001,
+      "loss": 0.7385,
+      "step": 361
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.0001,
+      "loss": 0.7372,
+      "step": 362
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.0001,
+      "loss": 0.6997,
+      "step": 363
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.0001,
+      "loss": 0.6783,
+      "step": 364
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 0.0001,
+      "loss": 0.6913,
+      "step": 365
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001,
+      "loss": 0.7702,
+      "step": 366
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 0.0001,
+      "loss": 0.7074,
+      "step": 367
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0001,
+      "loss": 0.7402,
+      "step": 368
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0001,
+      "loss": 0.7522,
+      "step": 369
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 0.0001,
+      "loss": 0.6422,
+      "step": 370
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.0001,
+      "loss": 0.6199,
+      "step": 371
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 0.0001,
+      "loss": 0.6999,
+      "step": 372
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.0001,
+      "loss": 0.7034,
+      "step": 373
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 0.0001,
+      "loss": 0.6757,
+      "step": 374
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.0001,
+      "loss": 0.7773,
+      "step": 375
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.0001,
+      "loss": 0.8073,
+      "step": 376
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7425,
+      "step": 377
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7389,
+      "step": 378
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7709,
+      "step": 379
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001,
+      "loss": 0.7646,
+      "step": 380
+    },
+    {
+      "epoch": 1.69,
+      "eval_loss": 0.7447208166122437,
+      "eval_runtime": 16.7015,
+      "eval_samples_per_second": 4.85,
+      "eval_steps_per_second": 1.257,
+      "step": 380
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 0.0001,
+      "loss": 0.719,
+      "step": 381
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001,
+      "loss": 0.7294,
+      "step": 382
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 0.0001,
+      "loss": 0.7995,
+      "step": 383
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001,
+      "loss": 0.7291,
+      "step": 384
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 0.0001,
+      "loss": 0.7088,
+      "step": 385
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001,
+      "loss": 0.7394,
+      "step": 386
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001,
+      "loss": 0.7576,
+      "step": 387
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 0.0001,
+      "loss": 0.7786,
+      "step": 388
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001,
+      "loss": 0.7835,
+      "step": 389
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.0001,
+      "loss": 0.7119,
+      "step": 390
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.0001,
+      "loss": 0.7275,
+      "step": 391
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 0.0001,
+      "loss": 0.745,
+      "step": 392
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.0001,
+      "loss": 0.7427,
+      "step": 393
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 0.0001,
+      "loss": 0.7822,
+      "step": 394
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.0001,
+      "loss": 0.69,
+      "step": 395
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.0001,
+      "loss": 0.72,
+      "step": 396
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 0.0001,
+      "loss": 0.7831,
+      "step": 397
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001,
+      "loss": 0.6883,
+      "step": 398
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 0.0001,
+      "loss": 0.711,
+      "step": 399
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0001,
+      "loss": 0.7606,
+      "step": 400
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 0.7409220337867737,
+      "eval_runtime": 16.8388,
+      "eval_samples_per_second": 4.81,
+      "eval_steps_per_second": 1.247,
+      "step": 400
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.0001,
+      "loss": 0.7286,
+      "step": 401
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.0001,
+      "loss": 0.6967,
+      "step": 402
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 0.0001,
+      "loss": 0.7288,
+      "step": 403
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.0001,
+      "loss": 0.7113,
+      "step": 404
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.0001,
+      "loss": 0.7094,
+      "step": 405
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 0.0001,
+      "loss": 0.6739,
+      "step": 406
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.0001,
+      "loss": 0.835,
+      "step": 407
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 0.0001,
+      "loss": 0.7148,
+      "step": 408
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.0001,
+      "loss": 0.786,
+      "step": 409
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 0.0001,
+      "loss": 0.7136,
+      "step": 410
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.0001,
+      "loss": 0.7714,
+      "step": 411
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 0.0001,
+      "loss": 0.685,
+      "step": 412
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001,
+      "loss": 0.6909,
+      "step": 413
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001,
+      "loss": 0.6932,
+      "step": 414
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 0.0001,
+      "loss": 0.7211,
+      "step": 415
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001,
+      "loss": 0.7558,
+      "step": 416
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 0.0001,
+      "loss": 0.7207,
+      "step": 417
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.0001,
+      "loss": 0.7191,
+      "step": 418
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.0001,
+      "loss": 0.6973,
+      "step": 419
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.0001,
+      "loss": 0.7157,
+      "step": 420
+    },
+    {
+      "epoch": 1.87,
+      "eval_loss": 0.739484429359436,
+      "eval_runtime": 16.7177,
+      "eval_samples_per_second": 4.845,
+      "eval_steps_per_second": 1.256,
+      "step": 420
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.0001,
+      "loss": 0.6833,
+      "step": 421
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.0001,
+      "loss": 0.88,
+      "step": 422
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.0001,
+      "loss": 0.7182,
+      "step": 423
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 0.0001,
+      "loss": 0.6655,
+      "step": 424
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.0001,
+      "loss": 0.7204,
+      "step": 425
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.0001,
+      "loss": 0.7894,
+      "step": 426
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.0001,
+      "loss": 0.776,
+      "step": 427
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 0.0001,
+      "loss": 0.7675,
+      "step": 428
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001,
+      "loss": 0.781,
+      "step": 429
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 0.0001,
+      "loss": 0.7166,
+      "step": 430
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001,
+      "loss": 0.7649,
+      "step": 431
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001,
+      "loss": 0.7263,
+      "step": 432
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 0.0001,
+      "loss": 0.6987,
+      "step": 433
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.0001,
+      "loss": 0.629,
+      "step": 434
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 0.0001,
+      "loss": 0.6829,
+      "step": 435
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.0001,
+      "loss": 0.7115,
+      "step": 436
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 0.0001,
+      "loss": 0.7318,
+      "step": 437
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 0.0001,
+      "loss": 0.7037,
+      "step": 438
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 0.0001,
+      "loss": 0.6787,
+      "step": 439
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 0.0001,
+      "loss": 0.7436,
+      "step": 440
+    },
+    {
+      "epoch": 1.96,
+      "eval_loss": 0.7383676767349243,
+      "eval_runtime": 16.8368,
+      "eval_samples_per_second": 4.811,
+      "eval_steps_per_second": 1.247,
+      "step": 440
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 0.0001,
+      "loss": 0.6661,
+      "step": 441
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 0.0001,
+      "loss": 0.6892,
+      "step": 442
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 0.0001,
+      "loss": 0.7032,
+      "step": 443
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 0.0001,
+      "loss": 0.6904,
+      "step": 444
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 0.0001,
+      "loss": 0.6905,
+      "step": 445
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 0.0001,
+      "loss": 0.6586,
+      "step": 446
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 0.0001,
+      "loss": 0.6711,
+      "step": 447
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 0.0001,
+      "loss": 0.7361,
+      "step": 448
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.0001,
+      "loss": 0.7098,
+      "step": 449
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.0001,
+      "loss": 0.7271,
+      "step": 450
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.0001,
+      "loss": 0.7657,
+      "step": 451
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 0.0001,
+      "loss": 0.6906,
+      "step": 452
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 0.0001,
+      "loss": 0.7584,
+      "step": 453
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 0.0001,
+      "loss": 0.6881,
+      "step": 454
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 0.0001,
+      "loss": 0.8123,
+      "step": 455
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 0.0001,
+      "loss": 0.7077,
+      "step": 456
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 0.0001,
+      "loss": 0.7166,
+      "step": 457
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 0.0001,
+      "loss": 0.6904,
+      "step": 458
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 0.0001,
+      "loss": 0.7855,
+      "step": 459
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 0.0001,
+      "loss": 0.7021,
+      "step": 460
+    },
+    {
+      "epoch": 2.04,
+      "eval_loss": 0.7379825711250305,
+      "eval_runtime": 16.8091,
+      "eval_samples_per_second": 4.819,
+      "eval_steps_per_second": 1.249,
+      "step": 460
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 0.0001,
+      "loss": 0.6584,
+      "step": 461
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 0.0001,
+      "loss": 0.7779,
+      "step": 462
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 0.0001,
+      "loss": 0.777,
+      "step": 463
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 0.0001,
+      "loss": 0.6805,
+      "step": 464
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 0.0001,
+      "loss": 0.7154,
+      "step": 465
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 0.0001,
+      "loss": 0.611,
+      "step": 466
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 0.0001,
+      "loss": 0.6393,
+      "step": 467
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 0.0001,
+      "loss": 0.6244,
+      "step": 468
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 0.0001,
+      "loss": 0.6941,
+      "step": 469
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 0.0001,
+      "loss": 0.6542,
+      "step": 470
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 0.0001,
+      "loss": 0.6351,
+      "step": 471
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 0.0001,
+      "loss": 0.7268,
+      "step": 472
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 0.0001,
+      "loss": 0.7127,
+      "step": 473
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 0.0001,
+      "loss": 0.6683,
+      "step": 474
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 0.0001,
+      "loss": 0.6442,
+      "step": 475
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 0.0001,
+      "loss": 0.7109,
+      "step": 476
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6536,
+      "step": 477
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6947,
+      "step": 478
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 0.0001,
+      "loss": 0.6727,
+      "step": 479
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 0.0001,
+      "loss": 0.6706,
+      "step": 480
+    },
+    {
+      "epoch": 2.13,
+      "eval_loss": 0.7374832034111023,
+      "eval_runtime": 16.816,
+      "eval_samples_per_second": 4.817,
+      "eval_steps_per_second": 1.249,
+      "step": 480
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 0.0001,
+      "loss": 0.642,
+      "step": 481
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 0.0001,
+      "loss": 0.6969,
+      "step": 482
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 0.0001,
+      "loss": 0.6929,
+      "step": 483
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 0.0001,
+      "loss": 0.6866,
+      "step": 484
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 0.0001,
+      "loss": 0.61,
+      "step": 485
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 0.0001,
+      "loss": 0.6935,
+      "step": 486
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 0.0001,
+      "loss": 0.6535,
+      "step": 487
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 0.0001,
+      "loss": 0.6148,
+      "step": 488
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 0.0001,
+      "loss": 0.633,
+      "step": 489
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 0.0001,
+      "loss": 0.6691,
+      "step": 490
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 0.0001,
+      "loss": 0.6587,
+      "step": 491
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 0.0001,
+      "loss": 0.6674,
+      "step": 492
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 0.0001,
+      "loss": 0.6451,
+      "step": 493
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 0.0001,
+      "loss": 0.6007,
+      "step": 494
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 0.0001,
+      "loss": 0.63,
+      "step": 495
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 0.0001,
+      "loss": 0.6647,
+      "step": 496
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 0.0001,
+      "loss": 0.6268,
+      "step": 497
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 0.0001,
+      "loss": 0.6336,
+      "step": 498
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 0.0001,
+      "loss": 0.6086,
+      "step": 499
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 0.0001,
+      "loss": 0.6903,
+      "step": 500
+    },
+    {
+      "epoch": 2.22,
+      "eval_loss": 0.7428303956985474,
+      "eval_runtime": 16.8269,
+      "eval_samples_per_second": 4.814,
+      "eval_steps_per_second": 1.248,
+      "step": 500
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 0.0001,
+      "loss": 0.7179,
+      "step": 501
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 0.0001,
+      "loss": 0.6935,
+      "step": 502
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 0.0001,
+      "loss": 0.7188,
+      "step": 503
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 0.0001,
+      "loss": 0.7431,
+      "step": 504
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 0.0001,
+      "loss": 0.7689,
+      "step": 505
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 0.0001,
+      "loss": 0.7093,
+      "step": 506
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 0.0001,
+      "loss": 0.7496,
+      "step": 507
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 0.0001,
+      "loss": 0.6739,
+      "step": 508
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 0.0001,
+      "loss": 0.6908,
+      "step": 509
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 0.0001,
+      "loss": 0.7479,
+      "step": 510
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 0.0001,
+      "loss": 0.7326,
+      "step": 511
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 0.0001,
+      "loss": 0.6806,
+      "step": 512
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 0.0001,
+      "loss": 0.6873,
+      "step": 513
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 0.0001,
+      "loss": 0.715,
+      "step": 514
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 0.0001,
+      "loss": 0.7345,
+      "step": 515
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 0.0001,
+      "loss": 0.7357,
+      "step": 516
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 0.0001,
+      "loss": 0.6662,
+      "step": 517
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 0.0001,
+      "loss": 0.6833,
+      "step": 518
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 0.0001,
+      "loss": 0.6079,
+      "step": 519
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 0.0001,
+      "loss": 0.7135,
+      "step": 520
+    },
+    {
+      "epoch": 2.31,
+      "eval_loss": 0.734011173248291,
+      "eval_runtime": 16.649,
+      "eval_samples_per_second": 4.865,
+      "eval_steps_per_second": 1.261,
+      "step": 520
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 0.0001,
+      "loss": 0.729,
+      "step": 521
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 0.0001,
+      "loss": 0.6364,
+      "step": 522
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 0.0001,
+      "loss": 0.6216,
+      "step": 523
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 0.0001,
+      "loss": 0.7562,
+      "step": 524
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 0.0001,
+      "loss": 0.6287,
+      "step": 525
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 0.0001,
+      "loss": 0.6435,
+      "step": 526
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 0.0001,
+      "loss": 0.7105,
+      "step": 527
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 0.0001,
+      "loss": 0.6931,
+      "step": 528
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 0.0001,
+      "loss": 0.7115,
+      "step": 529
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 0.0001,
+      "loss": 0.6406,
+      "step": 530
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 0.0001,
+      "loss": 0.6595,
+      "step": 531
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 0.0001,
+      "loss": 0.6851,
+      "step": 532
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 0.0001,
+      "loss": 0.6446,
+      "step": 533
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 0.0001,
+      "loss": 0.6522,
+      "step": 534
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 0.0001,
+      "loss": 0.6917,
+      "step": 535
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 0.0001,
+      "loss": 0.671,
+      "step": 536
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 0.0001,
+      "loss": 0.7605,
+      "step": 537
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 0.0001,
+      "loss": 0.5984,
+      "step": 538
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 0.0001,
+      "loss": 0.6499,
+      "step": 539
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 0.0001,
+      "loss": 0.6795,
+      "step": 540
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 0.734115719795227,
+      "eval_runtime": 17.0499,
+      "eval_samples_per_second": 4.751,
+      "eval_steps_per_second": 1.232,
+      "step": 540
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 0.0001,
+      "loss": 0.7156,
+      "step": 541
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 0.0001,
+      "loss": 0.6773,
+      "step": 542
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 0.0001,
+      "loss": 0.6589,
+      "step": 543
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 0.0001,
+      "loss": 0.7106,
+      "step": 544
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 0.0001,
+      "loss": 0.6798,
+      "step": 545
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 0.0001,
+      "loss": 0.6733,
+      "step": 546
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 0.0001,
+      "loss": 0.722,
+      "step": 547
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 0.0001,
+      "loss": 0.5598,
+      "step": 548
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 0.0001,
+      "loss": 0.6286,
+      "step": 549
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 0.0001,
+      "loss": 0.7219,
+      "step": 550
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7907,
+      "step": 551
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7635,
+      "step": 552
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 0.0001,
+      "loss": 0.7163,
+      "step": 553
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 0.0001,
+      "loss": 0.6866,
+      "step": 554
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7523,
+      "step": 555
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7165,
+      "step": 556
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 0.0001,
+      "loss": 0.6766,
+      "step": 557
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 0.0001,
+      "loss": 0.7717,
+      "step": 558
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 0.0001,
+      "loss": 0.7346,
+      "step": 559
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 0.0001,
+      "loss": 0.705,
+      "step": 560
+    },
+    {
+      "epoch": 2.49,
+      "eval_loss": 0.7371004223823547,
+      "eval_runtime": 16.4728,
+      "eval_samples_per_second": 4.917,
+      "eval_steps_per_second": 1.275,
+      "step": 560
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 0.0001,
+      "loss": 0.6723,
+      "step": 561
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 0.0001,
+      "loss": 0.6548,
+      "step": 562
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 0.0001,
+      "loss": 0.6814,
+      "step": 563
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 0.0001,
+      "loss": 0.7153,
+      "step": 564
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 0.0001,
+      "loss": 0.6452,
+      "step": 565
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6938,
+      "step": 566
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6396,
+      "step": 567
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6502,
+      "step": 568
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 0.0001,
+      "loss": 0.6748,
+      "step": 569
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 0.0001,
+      "loss": 0.6681,
+      "step": 570
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 0.0001,
+      "loss": 0.7478,
+      "step": 571
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 0.0001,
+      "loss": 0.7108,
+      "step": 572
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 0.0001,
+      "loss": 0.743,
+      "step": 573
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 0.0001,
+      "loss": 0.6489,
+      "step": 574
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 0.0001,
+      "loss": 0.6889,
+      "step": 575
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 0.0001,
+      "loss": 0.7132,
+      "step": 576
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 0.0001,
+      "loss": 0.6254,
+      "step": 577
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 0.0001,
+      "loss": 0.7012,
+      "step": 578
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 0.0001,
+      "loss": 0.7193,
+      "step": 579
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 0.0001,
+      "loss": 0.6853,
+      "step": 580
+    },
+    {
+      "epoch": 2.58,
+      "eval_loss": 0.7324036955833435,
+      "eval_runtime": 16.8087,
+      "eval_samples_per_second": 4.819,
+      "eval_steps_per_second": 1.249,
+      "step": 580
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 0.0001,
+      "loss": 0.6266,
+      "step": 581
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 0.0001,
+      "loss": 0.7268,
+      "step": 582
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 0.0001,
+      "loss": 0.6741,
+      "step": 583
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 0.0001,
+      "loss": 0.6661,
+      "step": 584
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 0.0001,
+      "loss": 0.6975,
+      "step": 585
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 0.0001,
+      "loss": 0.684,
+      "step": 586
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 0.0001,
+      "loss": 0.6384,
+      "step": 587
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 0.0001,
+      "loss": 0.6374,
+      "step": 588
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 0.0001,
+      "loss": 0.6731,
+      "step": 589
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 0.0001,
+      "loss": 0.6921,
+      "step": 590
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 0.0001,
+      "loss": 0.6296,
+      "step": 591
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 0.0001,
+      "loss": 0.6464,
+      "step": 592
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 0.0001,
+      "loss": 0.6565,
+      "step": 593
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 0.0001,
+      "loss": 0.6775,
+      "step": 594
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 0.0001,
+      "loss": 0.6897,
+      "step": 595
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 0.0001,
+      "loss": 0.6463,
+      "step": 596
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 0.0001,
+      "loss": 0.7714,
+      "step": 597
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 0.0001,
+      "loss": 0.6419,
+      "step": 598
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 0.0001,
+      "loss": 0.612,
+      "step": 599
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 0.0001,
+      "loss": 0.5813,
+      "step": 600
+    },
+    {
+      "epoch": 2.67,
+      "eval_loss": 0.7384765148162842,
+      "eval_runtime": 17.1978,
+      "eval_samples_per_second": 4.71,
+      "eval_steps_per_second": 1.221,
+      "step": 600
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 0.0001,
+      "loss": 0.7514,
+      "step": 601
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7116,
+      "step": 602
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 0.0001,
+      "loss": 0.6747,
+      "step": 603
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7278,
+      "step": 604
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 0.0001,
+      "loss": 0.7821,
+      "step": 605
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 0.0001,
+      "loss": 0.7186,
+      "step": 606
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 0.0001,
+      "loss": 0.7919,
+      "step": 607
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 0.0001,
+      "loss": 0.793,
+      "step": 608
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 0.0001,
+      "loss": 0.6905,
+      "step": 609
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 0.0001,
+      "loss": 0.7031,
+      "step": 610
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 0.0001,
+      "loss": 0.683,
+      "step": 611
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 0.0001,
+      "loss": 0.6863,
+      "step": 612
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 0.0001,
+      "loss": 0.7223,
+      "step": 613
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 0.0001,
+      "loss": 0.7148,
+      "step": 614
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 0.0001,
+      "loss": 0.7005,
+      "step": 615
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 0.0001,
+      "loss": 0.6737,
+      "step": 616
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 0.0001,
+      "loss": 0.7109,
+      "step": 617
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 0.0001,
+      "loss": 0.6446,
+      "step": 618
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 0.0001,
+      "loss": 0.7636,
+      "step": 619
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 0.0001,
+      "loss": 0.6154,
+      "step": 620
+    },
+    {
+      "epoch": 2.76,
+      "eval_loss": 0.7297637462615967,
+      "eval_runtime": 17.6875,
+      "eval_samples_per_second": 4.58,
+      "eval_steps_per_second": 1.187,
+      "step": 620
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 0.0001,
+      "loss": 0.5922,
+      "step": 621
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 0.0001,
+      "loss": 0.6575,
+      "step": 622
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 0.0001,
+      "loss": 0.7034,
+      "step": 623
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 0.0001,
+      "loss": 0.6845,
+      "step": 624
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 0.0001,
+      "loss": 0.7364,
+      "step": 625
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 0.0001,
+      "loss": 0.6517,
+      "step": 626
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 0.0001,
+      "loss": 0.6223,
+      "step": 627
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 0.0001,
+      "loss": 0.6525,
+      "step": 628
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 0.0001,
+      "loss": 0.665,
+      "step": 629
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 0.0001,
+      "loss": 0.6621,
+      "step": 630
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 0.0001,
+      "loss": 0.7138,
+      "step": 631
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 0.0001,
+      "loss": 0.5946,
+      "step": 632
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 0.0001,
+      "loss": 0.702,
+      "step": 633
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 0.0001,
+      "loss": 0.6865,
+      "step": 634
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 0.0001,
+      "loss": 0.7088,
+      "step": 635
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 0.0001,
+      "loss": 0.7178,
+      "step": 636
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 0.0001,
+      "loss": 0.711,
+      "step": 637
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 0.0001,
+      "loss": 0.7497,
+      "step": 638
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 0.0001,
+      "loss": 0.6651,
+      "step": 639
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 0.0001,
+      "loss": 0.6896,
+      "step": 640
+    },
+    {
+      "epoch": 2.84,
+      "eval_loss": 0.7296957969665527,
+      "eval_runtime": 17.6413,
+      "eval_samples_per_second": 4.591,
+      "eval_steps_per_second": 1.19,
+      "step": 640
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 0.0001,
+      "loss": 0.6705,
+      "step": 641
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 0.0001,
+      "loss": 0.6571,
+      "step": 642
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 0.0001,
+      "loss": 0.6644,
+      "step": 643
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 0.0001,
+      "loss": 0.6748,
+      "step": 644
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 0.0001,
+      "loss": 0.6662,
+      "step": 645
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 0.0001,
+      "loss": 0.6011,
+      "step": 646
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 0.0001,
+      "loss": 0.6546,
+      "step": 647
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 0.0001,
+      "loss": 0.726,
+      "step": 648
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 0.0001,
+      "loss": 0.6394,
+      "step": 649
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 0.0001,
+      "loss": 0.63,
+      "step": 650
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 0.0001,
+      "loss": 0.7665,
+      "step": 651
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 0.0001,
+      "loss": 0.6785,
+      "step": 652
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 0.0001,
+      "loss": 0.6973,
+      "step": 653
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 0.0001,
+      "loss": 0.7283,
+      "step": 654
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 0.0001,
+      "loss": 0.7365,
+      "step": 655
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 0.0001,
+      "loss": 0.6906,
+      "step": 656
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 0.0001,
+      "loss": 0.6758,
+      "step": 657
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 0.0001,
+      "loss": 0.656,
+      "step": 658
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 0.0001,
+      "loss": 0.7011,
+      "step": 659
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 0.0001,
+      "loss": 0.6314,
+      "step": 660
+    },
+    {
+      "epoch": 2.93,
+      "eval_loss": 0.7283563613891602,
+      "eval_runtime": 17.6296,
+      "eval_samples_per_second": 4.595,
+      "eval_steps_per_second": 1.191,
+      "step": 660
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 0.0001,
+      "loss": 0.6825,
+      "step": 661
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 0.0001,
+      "loss": 0.7082,
+      "step": 662
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 0.0001,
+      "loss": 0.6026,
+      "step": 663
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 0.0001,
+      "loss": 0.6308,
+      "step": 664
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 0.0001,
+      "loss": 0.7622,
+      "step": 665
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 0.0001,
+      "loss": 0.6601,
+      "step": 666
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 0.0001,
+      "loss": 0.6863,
+      "step": 667
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 0.0001,
+      "loss": 0.7205,
+      "step": 668
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 0.0001,
+      "loss": 0.622,
+      "step": 669
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 0.0001,
+      "loss": 0.5898,
+      "step": 670
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 0.0001,
+      "loss": 0.6379,
+      "step": 671
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 0.0001,
+      "loss": 0.6464,
+      "step": 672
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 0.0001,
+      "loss": 0.6001,
+      "step": 673
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.0001,
+      "loss": 0.6351,
+      "step": 674
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.0001,
+      "loss": 0.716,
+      "step": 675
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.0001,
+      "loss": 0.7913,
+      "step": 676
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 0.0001,
+      "loss": 0.6763,
+      "step": 677
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 0.0001,
+      "loss": 0.6952,
+      "step": 678
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 0.0001,
+      "loss": 0.6933,
+      "step": 679
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 0.0001,
+      "loss": 0.7001,
+      "step": 680
+    },
+    {
+      "epoch": 3.02,
+      "eval_loss": 0.7300973534584045,
+      "eval_runtime": 17.9948,
+      "eval_samples_per_second": 4.501,
+      "eval_steps_per_second": 1.167,
+      "step": 680
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 0.0001,
+      "loss": 0.6507,
+      "step": 681
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 0.0001,
+      "loss": 0.6635,
+      "step": 682
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 0.0001,
+      "loss": 0.6627,
+      "step": 683
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 0.0001,
+      "loss": 0.6461,
+      "step": 684
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 0.0001,
+      "loss": 0.6493,
+      "step": 685
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 0.0001,
+      "loss": 0.6392,
+      "step": 686
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 0.0001,
+      "loss": 0.6457,
+      "step": 687
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 0.0001,
+      "loss": 0.6702,
+      "step": 688
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 0.0001,
+      "loss": 0.6652,
+      "step": 689
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 0.0001,
+      "loss": 0.6266,
+      "step": 690
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 0.0001,
+      "loss": 0.6648,
+      "step": 691
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 0.0001,
+      "loss": 0.5778,
+      "step": 692
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 0.0001,
+      "loss": 0.6581,
+      "step": 693
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 0.0001,
+      "loss": 0.7089,
+      "step": 694
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 0.0001,
+      "loss": 0.6478,
+      "step": 695
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 0.0001,
+      "loss": 0.5843,
+      "step": 696
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 0.0001,
+      "loss": 0.5995,
+      "step": 697
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 0.0001,
+      "loss": 0.6076,
+      "step": 698
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 0.0001,
+      "loss": 0.7328,
+      "step": 699
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 0.0001,
+      "loss": 0.5725,
+      "step": 700
+    },
+    {
+      "epoch": 3.11,
+      "eval_loss": 0.7330772280693054,
+      "eval_runtime": 18.8269,
+      "eval_samples_per_second": 4.302,
+      "eval_steps_per_second": 1.115,
+      "step": 700
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6693,
+      "step": 701
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6445,
+      "step": 702
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 0.0001,
+      "loss": 0.6478,
+      "step": 703
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 0.0001,
+      "loss": 0.661,
+      "step": 704
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 0.0001,
+      "loss": 0.6351,
+      "step": 705
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 0.0001,
+      "loss": 0.5796,
+      "step": 706
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 0.0001,
+      "loss": 0.607,
+      "step": 707
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 0.0001,
+      "loss": 0.6227,
+      "step": 708
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 0.0001,
+      "loss": 0.6283,
+      "step": 709
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 0.0001,
+      "loss": 0.6654,
+      "step": 710
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 0.0001,
+      "loss": 0.6348,
+      "step": 711
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 0.0001,
+      "loss": 0.5861,
+      "step": 712
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 0.0001,
+      "loss": 0.5794,
+      "step": 713
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 0.0001,
+      "loss": 0.6011,
+      "step": 714
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 0.0001,
+      "loss": 0.5992,
+      "step": 715
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 0.0001,
+      "loss": 0.624,
+      "step": 716
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 0.0001,
+      "loss": 0.636,
+      "step": 717
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 0.0001,
+      "loss": 0.5255,
+      "step": 718
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 0.0001,
+      "loss": 0.6403,
+      "step": 719
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 0.0001,
+      "loss": 0.5832,
+      "step": 720
+    },
+    {
+      "epoch": 3.2,
+      "eval_loss": 0.7388717532157898,
+      "eval_runtime": 18.8118,
+      "eval_samples_per_second": 4.306,
+      "eval_steps_per_second": 1.116,
+      "step": 720
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 0.0001,
+      "loss": 0.5699,
+      "step": 721
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 0.0001,
+      "loss": 0.6878,
+      "step": 722
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 0.0001,
+      "loss": 0.5529,
+      "step": 723
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 0.0001,
+      "loss": 0.5619,
+      "step": 724
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 0.0001,
+      "loss": 0.5764,
+      "step": 725
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 0.0001,
+      "loss": 0.7349,
+      "step": 726
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 0.0001,
+      "loss": 0.7237,
+      "step": 727
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 0.0001,
+      "loss": 0.6804,
+      "step": 728
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 0.0001,
+      "loss": 0.6911,
+      "step": 729
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 0.0001,
+      "loss": 0.6582,
+      "step": 730
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 0.0001,
+      "loss": 0.6994,
+      "step": 731
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 0.0001,
+      "loss": 0.6864,
+      "step": 732
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 0.0001,
+      "loss": 0.7323,
+      "step": 733
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 0.0001,
+      "loss": 0.6921,
+      "step": 734
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 0.0001,
+      "loss": 0.6115,
+      "step": 735
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 0.0001,
+      "loss": 0.6423,
+      "step": 736
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 0.0001,
+      "loss": 0.629,
+      "step": 737
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 0.0001,
+      "loss": 0.6512,
+      "step": 738
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 0.0001,
+      "loss": 0.6384,
+      "step": 739
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 0.0001,
+      "loss": 0.629,
+      "step": 740
+    },
+    {
+      "epoch": 3.29,
+      "eval_loss": 0.7320723533630371,
+      "eval_runtime": 18.8594,
+      "eval_samples_per_second": 4.295,
+      "eval_steps_per_second": 1.114,
+      "step": 740
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 0.0001,
+      "loss": 0.6447,
+      "step": 741
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 0.0001,
+      "loss": 0.5919,
+      "step": 742
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 0.0001,
+      "loss": 0.6678,
+      "step": 743
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 0.0001,
+      "loss": 0.5893,
+      "step": 744
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 0.0001,
+      "loss": 0.593,
+      "step": 745
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 0.0001,
+      "loss": 0.6514,
+      "step": 746
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 0.0001,
+      "loss": 0.6722,
+      "step": 747
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 0.0001,
+      "loss": 0.5961,
+      "step": 748
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 0.0001,
+      "loss": 0.6213,
+      "step": 749
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 0.0001,
+      "loss": 0.6091,
+      "step": 750
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 0.0001,
+      "loss": 0.6327,
+      "step": 751
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 0.0001,
+      "loss": 0.6781,
+      "step": 752
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 0.0001,
+      "loss": 0.609,
+      "step": 753
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 0.0001,
+      "loss": 0.6491,
+      "step": 754
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 0.0001,
+      "loss": 0.6231,
+      "step": 755
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 0.0001,
+      "loss": 0.6353,
+      "step": 756
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 0.0001,
+      "loss": 0.7171,
+      "step": 757
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 0.0001,
+      "loss": 0.702,
+      "step": 758
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 0.0001,
+      "loss": 0.6205,
+      "step": 759
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 0.0001,
+      "loss": 0.6037,
+      "step": 760
+    },
+    {
+      "epoch": 3.38,
+      "eval_loss": 0.735625147819519,
+      "eval_runtime": 18.9293,
+      "eval_samples_per_second": 4.279,
+      "eval_steps_per_second": 1.109,
+      "step": 760
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 0.0001,
+      "loss": 0.6143,
+      "step": 761
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 0.0001,
+      "loss": 0.6124,
+      "step": 762
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 0.0001,
+      "loss": 0.5528,
+      "step": 763
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 0.0001,
+      "loss": 0.5531,
+      "step": 764
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 0.0001,
+      "loss": 0.5917,
+      "step": 765
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 0.0001,
+      "loss": 0.5962,
+      "step": 766
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 0.0001,
+      "loss": 0.6423,
+      "step": 767
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 0.0001,
+      "loss": 0.613,
+      "step": 768
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 0.0001,
+      "loss": 0.6326,
+      "step": 769
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 0.0001,
+      "loss": 0.6229,
+      "step": 770
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 0.0001,
+      "loss": 0.6613,
+      "step": 771
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 0.0001,
+      "loss": 0.6048,
+      "step": 772
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 0.0001,
+      "loss": 0.5452,
+      "step": 773
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 0.0001,
+      "loss": 0.5207,
+      "step": 774
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 0.0001,
+      "loss": 0.5988,
+      "step": 775
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 0.0001,
+      "loss": 0.7064,
+      "step": 776
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 0.0001,
+      "loss": 0.6414,
+      "step": 777
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 0.0001,
+      "loss": 0.7127,
+      "step": 778
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 0.0001,
+      "loss": 0.7062,
+      "step": 779
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7391,
+      "step": 780
+    },
+    {
+      "epoch": 3.47,
+      "eval_loss": 0.7382710576057434,
+      "eval_runtime": 19.0816,
+      "eval_samples_per_second": 4.245,
+      "eval_steps_per_second": 1.101,
+      "step": 780
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 0.0001,
+      "loss": 0.7217,
+      "step": 781
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 0.0001,
+      "loss": 0.6948,
+      "step": 782
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 0.0001,
+      "loss": 0.7053,
+      "step": 783
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 0.0001,
+      "loss": 0.7944,
+      "step": 784
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 0.0001,
+      "loss": 0.683,
+      "step": 785
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 0.0001,
+      "loss": 0.7693,
+      "step": 786
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 0.0001,
+      "loss": 0.6613,
+      "step": 787
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 0.0001,
+      "loss": 0.6312,
+      "step": 788
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 0.0001,
+      "loss": 0.7292,
+      "step": 789
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 0.0001,
+      "loss": 0.6532,
+      "step": 790
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6847,
+      "step": 791
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6814,
+      "step": 792
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 0.0001,
+      "loss": 0.6137,
+      "step": 793
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 0.0001,
+      "loss": 0.6355,
+      "step": 794
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 0.0001,
+      "loss": 0.5924,
+      "step": 795
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 0.0001,
+      "loss": 0.5793,
+      "step": 796
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 0.0001,
+      "loss": 0.5977,
+      "step": 797
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 0.0001,
+      "loss": 0.7161,
+      "step": 798
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 0.0001,
+      "loss": 0.6635,
+      "step": 799
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 0.0001,
+      "loss": 0.6454,
+      "step": 800
+    },
+    {
+      "epoch": 3.56,
+      "eval_loss": 0.7320914268493652,
+      "eval_runtime": 19.2836,
+      "eval_samples_per_second": 4.2,
+      "eval_steps_per_second": 1.089,
+      "step": 800
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 0.0001,
+      "loss": 0.642,
+      "step": 801
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 0.0001,
+      "loss": 0.5794,
+      "step": 802
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 0.0001,
+      "loss": 0.637,
+      "step": 803
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 0.0001,
+      "loss": 0.6212,
+      "step": 804
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 0.0001,
+      "loss": 0.6635,
+      "step": 805
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 0.0001,
+      "loss": 0.6311,
+      "step": 806
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 0.0001,
+      "loss": 0.5901,
+      "step": 807
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 0.0001,
+      "loss": 0.648,
+      "step": 808
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 0.0001,
+      "loss": 0.6348,
+      "step": 809
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 0.0001,
+      "loss": 0.6346,
+      "step": 810
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 0.0001,
+      "loss": 0.6892,
+      "step": 811
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 0.0001,
+      "loss": 0.6337,
+      "step": 812
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 0.0001,
+      "loss": 0.6348,
+      "step": 813
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 0.0001,
+      "loss": 0.6576,
+      "step": 814
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 0.0001,
+      "loss": 0.6202,
+      "step": 815
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 0.0001,
+      "loss": 0.6148,
+      "step": 816
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 0.0001,
+      "loss": 0.668,
+      "step": 817
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 0.0001,
+      "loss": 0.6663,
+      "step": 818
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 0.0001,
+      "loss": 0.6004,
+      "step": 819
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 0.0001,
+      "loss": 0.5944,
+      "step": 820
+    },
+    {
+      "epoch": 3.64,
+      "eval_loss": 0.7350344061851501,
+      "eval_runtime": 19.0874,
+      "eval_samples_per_second": 4.244,
+      "eval_steps_per_second": 1.1,
+      "step": 820
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 0.0001,
+      "loss": 0.5598,
+      "step": 821
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 0.0001,
+      "loss": 0.5836,
+      "step": 822
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 0.0001,
+      "loss": 0.5664,
+      "step": 823
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 0.0001,
+      "loss": 0.7259,
+      "step": 824
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 0.0001,
+      "loss": 0.5605,
+      "step": 825
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 0.0001,
+      "loss": 0.7895,
+      "step": 826
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 0.0001,
+      "loss": 0.6594,
+      "step": 827
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 0.0001,
+      "loss": 0.6306,
+      "step": 828
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 0.0001,
+      "loss": 0.7212,
+      "step": 829
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 0.0001,
+      "loss": 0.6964,
+      "step": 830
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 0.0001,
+      "loss": 0.7085,
+      "step": 831
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 0.0001,
+      "loss": 0.7088,
+      "step": 832
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 0.0001,
+      "loss": 0.7106,
+      "step": 833
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 0.0001,
+      "loss": 0.6695,
+      "step": 834
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 0.0001,
+      "loss": 0.6925,
+      "step": 835
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 0.0001,
+      "loss": 0.6677,
+      "step": 836
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 0.0001,
+      "loss": 0.6597,
+      "step": 837
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 0.0001,
+      "loss": 0.6428,
+      "step": 838
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 0.0001,
+      "loss": 0.6708,
+      "step": 839
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 0.0001,
+      "loss": 0.6382,
+      "step": 840
+    },
+    {
+      "epoch": 3.73,
+      "eval_loss": 0.7277283668518066,
+      "eval_runtime": 19.254,
+      "eval_samples_per_second": 4.207,
+      "eval_steps_per_second": 1.091,
+      "step": 840
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 0.0001,
+      "loss": 0.6431,
+      "step": 841
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 0.0001,
+      "loss": 0.5893,
+      "step": 842
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 0.0001,
+      "loss": 0.5862,
+      "step": 843
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 0.0001,
+      "loss": 0.608,
+      "step": 844
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 0.0001,
+      "loss": 0.5948,
+      "step": 845
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 0.0001,
+      "loss": 0.7078,
+      "step": 846
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 0.0001,
+      "loss": 0.6372,
+      "step": 847
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 0.0001,
+      "loss": 0.6611,
+      "step": 848
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 0.0001,
+      "loss": 0.6026,
+      "step": 849
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 0.0001,
+      "loss": 0.651,
+      "step": 850
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 0.0001,
+      "loss": 0.6261,
+      "step": 851
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 0.0001,
+      "loss": 0.6362,
+      "step": 852
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 0.0001,
+      "loss": 0.5957,
+      "step": 853
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 0.0001,
+      "loss": 0.6254,
+      "step": 854
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 0.0001,
+      "loss": 0.6521,
+      "step": 855
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 0.0001,
+      "loss": 0.6361,
+      "step": 856
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 0.0001,
+      "loss": 0.6474,
+      "step": 857
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 0.0001,
+      "loss": 0.6121,
+      "step": 858
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 0.0001,
+      "loss": 0.5983,
+      "step": 859
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 0.0001,
+      "loss": 0.6431,
+      "step": 860
+    },
+    {
+      "epoch": 3.82,
+      "eval_loss": 0.729424774646759,
+      "eval_runtime": 19.1466,
+      "eval_samples_per_second": 4.231,
+      "eval_steps_per_second": 1.097,
+      "step": 860
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 0.0001,
+      "loss": 0.5838,
+      "step": 861
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 0.0001,
+      "loss": 0.7613,
+      "step": 862
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.0001,
+      "loss": 0.6654,
+      "step": 863
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.0001,
+      "loss": 0.5563,
+      "step": 864
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.0001,
+      "loss": 0.6706,
+      "step": 865
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 0.0001,
+      "loss": 0.6083,
+      "step": 866
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 0.0001,
+      "loss": 0.5665,
+      "step": 867
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 0.0001,
+      "loss": 0.6089,
+      "step": 868
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 0.0001,
+      "loss": 0.5937,
+      "step": 869
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 0.0001,
+      "loss": 0.5958,
+      "step": 870
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 0.0001,
+      "loss": 0.5194,
+      "step": 871
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 0.0001,
+      "loss": 0.5701,
+      "step": 872
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 0.0001,
+      "loss": 0.6191,
+      "step": 873
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 0.0001,
+      "loss": 0.6032,
+      "step": 874
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 0.0001,
+      "loss": 0.5789,
+      "step": 875
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 0.0001,
+      "loss": 0.6796,
+      "step": 876
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 0.0001,
+      "loss": 0.7132,
+      "step": 877
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 0.0001,
+      "loss": 0.7038,
+      "step": 878
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 0.0001,
+      "loss": 0.737,
+      "step": 879
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 0.0001,
+      "loss": 0.6938,
+      "step": 880
+    },
+    {
+      "epoch": 3.91,
+      "eval_loss": 0.7355621457099915,
+      "eval_runtime": 19.0965,
+      "eval_samples_per_second": 4.242,
+      "eval_steps_per_second": 1.1,
+      "step": 880
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 0.0001,
+      "loss": 0.7009,
+      "step": 881
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 0.0001,
+      "loss": 0.6425,
+      "step": 882
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 0.0001,
+      "loss": 0.6306,
+      "step": 883
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 0.0001,
+      "loss": 0.6771,
+      "step": 884
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 0.0001,
+      "loss": 0.6381,
+      "step": 885
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 0.0001,
+      "loss": 0.6741,
+      "step": 886
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 0.0001,
+      "loss": 0.6783,
+      "step": 887
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 0.0001,
+      "loss": 0.6619,
+      "step": 888
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 0.0001,
+      "loss": 0.6493,
+      "step": 889
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 0.0001,
+      "loss": 0.6248,
+      "step": 890
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 0.0001,
+      "loss": 0.5975,
+      "step": 891
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 0.0001,
+      "loss": 0.6597,
+      "step": 892
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 0.0001,
+      "loss": 0.5691,
+      "step": 893
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 0.0001,
+      "loss": 0.6303,
+      "step": 894
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 0.0001,
+      "loss": 0.6042,
+      "step": 895
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 0.0001,
+      "loss": 0.5876,
+      "step": 896
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 0.0001,
+      "loss": 0.6245,
+      "step": 897
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 0.0001,
+      "loss": 0.5628,
+      "step": 898
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 0.0001,
+      "loss": 0.5753,
+      "step": 899
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 0.0001,
+      "loss": 0.5886,
+      "step": 900
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.7348082661628723,
+      "eval_runtime": 18.8837,
+      "eval_samples_per_second": 4.289,
+      "eval_steps_per_second": 1.112,
+      "step": 900
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2250,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 20,
+  "total_flos": 2.1376566399074304e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}