diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,8186 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.943189596167009,
+  "global_step": 1075,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0,
+      "loss": 1.8413,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.637054741273227e-06,
+      "loss": 1.6001,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.934520379430466e-06,
+      "loss": 1.5977,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.1274109482546454e-05,
+      "loss": 1.521,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.3088835776180316e-05,
+      "loss": 1.5742,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.512109398841858,
+      "eval_runtime": 1.433,
+      "eval_samples_per_second": 41.87,
+      "eval_steps_per_second": 10.468,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.457157512070369e-05,
+      "loss": 1.4844,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5825213373821546e-05,
+      "loss": 1.4834,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.691116422381968e-05,
+      "loss": 1.4336,
+      "step": 8
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.786904075886093e-05,
+      "loss": 1.4746,
+      "step": 9
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8725890517453548e-05,
+      "loss": 1.4351,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "eval_loss": 1.3938151597976685,
+      "eval_runtime": 1.4306,
+      "eval_samples_per_second": 41.941,
+      "eval_steps_per_second": 10.485,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.950100540794989e-05,
+      "loss": 1.4609,
+      "step": 11
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.0208629861976917e-05,
+      "loss": 1.3857,
+      "step": 12
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.085958125794301e-05,
+      "loss": 1.4692,
+      "step": 13
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.1462268115094774e-05,
+      "loss": 1.3394,
+      "step": 14
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.2023356155610784e-05,
+      "loss": 1.4116,
+      "step": 15
+    },
+    {
+      "epoch": 0.04,
+      "eval_loss": 1.3675130605697632,
+      "eval_runtime": 1.4323,
+      "eval_samples_per_second": 41.889,
+      "eval_steps_per_second": 10.472,
+      "step": 15
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.254821896509291e-05,
+      "loss": 1.4302,
+      "step": 16
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.3041251789048363e-05,
+      "loss": 1.3269,
+      "step": 17
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.3506095500134153e-05,
+      "loss": 1.3823,
+      "step": 18
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.3945799930242155e-05,
+      "loss": 1.2886,
+      "step": 19
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.436294525872677e-05,
+      "loss": 1.3508,
+      "step": 20
+    },
+    {
+      "epoch": 0.05,
+      "eval_loss": 1.3350911140441895,
+      "eval_runtime": 1.4275,
+      "eval_samples_per_second": 42.031,
+      "eval_steps_per_second": 10.508,
+      "step": 20
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.4759733753252014e-05,
+      "loss": 1.3735,
+      "step": 21
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.513806014922312e-05,
+      "loss": 1.2954,
+      "step": 22
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.5499566371834378e-05,
+      "loss": 1.4253,
+      "step": 23
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.5845684603250148e-05,
+      "loss": 1.3831,
+      "step": 24
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.6177671552360633e-05,
+      "loss": 1.4072,
+      "step": 25
+    },
+    {
+      "epoch": 0.07,
+      "eval_loss": 1.326757788658142,
+      "eval_runtime": 1.4313,
+      "eval_samples_per_second": 41.919,
+      "eval_steps_per_second": 10.48,
+      "step": 25
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.6496635999216235e-05,
+      "loss": 1.3521,
+      "step": 26
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 2.6803561138291392e-05,
+      "loss": 1.3958,
+      "step": 27
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.7099322856368e-05,
+      "loss": 1.2773,
+      "step": 28
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.7384704801599113e-05,
+      "loss": 1.313,
+      "step": 29
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.766041089688401e-05,
+      "loss": 1.3447,
+      "step": 30
+    },
+    {
+      "epoch": 0.08,
+      "eval_loss": 1.3093750476837158,
+      "eval_runtime": 1.4386,
+      "eval_samples_per_second": 41.707,
+      "eval_steps_per_second": 10.427,
+      "step": 30
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 2.7927075800664663e-05,
+      "loss": 1.3066,
+      "step": 31
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.8185273706366137e-05,
+      "loss": 1.3921,
+      "step": 32
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.8435525787380355e-05,
+      "loss": 1.2788,
+      "step": 33
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 2.867830653032159e-05,
+      "loss": 1.3257,
+      "step": 34
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.8914049150001866e-05,
+      "loss": 1.3452,
+      "step": 35
+    },
+    {
+      "epoch": 0.1,
+      "eval_loss": 1.2986328601837158,
+      "eval_runtime": 1.4314,
+      "eval_samples_per_second": 41.918,
+      "eval_steps_per_second": 10.48,
+      "step": 35
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.914315024140738e-05,
+      "loss": 1.2688,
+      "step": 36
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.9365973794160442e-05,
+      "loss": 1.3618,
+      "step": 37
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 2.958285467151538e-05,
+      "loss": 1.436,
+      "step": 38
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.979410163737347e-05,
+      "loss": 1.3599,
+      "step": 39
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9999999999999997e-05,
+      "loss": 1.2979,
+      "step": 40
+    },
+    {
+      "epoch": 0.11,
+      "eval_loss": 1.2783854007720947,
+      "eval_runtime": 1.4346,
+      "eval_samples_per_second": 41.824,
+      "eval_steps_per_second": 10.456,
+      "step": 40
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 3e-05,
+      "loss": 1.3557,
+      "step": 41
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 2.9978873239436622e-05,
+      "loss": 1.3674,
+      "step": 42
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.995774647887324e-05,
+      "loss": 1.312,
+      "step": 43
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.993661971830986e-05,
+      "loss": 1.3704,
+      "step": 44
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 2.991549295774648e-05,
+      "loss": 1.3169,
+      "step": 45
+    },
+    {
+      "epoch": 0.12,
+      "eval_loss": 1.2796224355697632,
+      "eval_runtime": 1.4418,
+      "eval_samples_per_second": 41.614,
+      "eval_steps_per_second": 10.404,
+      "step": 45
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.98943661971831e-05,
+      "loss": 1.2617,
+      "step": 46
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9873239436619718e-05,
+      "loss": 1.2371,
+      "step": 47
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.985211267605634e-05,
+      "loss": 1.2886,
+      "step": 48
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 2.9830985915492957e-05,
+      "loss": 1.3472,
+      "step": 49
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.980985915492958e-05,
+      "loss": 1.2905,
+      "step": 50
+    },
+    {
+      "epoch": 0.14,
+      "eval_loss": 1.2742838859558105,
+      "eval_runtime": 1.4302,
+      "eval_samples_per_second": 41.952,
+      "eval_steps_per_second": 10.488,
+      "step": 50
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.97887323943662e-05,
+      "loss": 1.3145,
+      "step": 51
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 2.9767605633802818e-05,
+      "loss": 1.4238,
+      "step": 52
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.974647887323944e-05,
+      "loss": 1.311,
+      "step": 53
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.9725352112676057e-05,
+      "loss": 1.2935,
+      "step": 54
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.9704225352112678e-05,
+      "loss": 1.311,
+      "step": 55
+    },
+    {
+      "epoch": 0.15,
+      "eval_loss": 1.2705078125,
+      "eval_runtime": 1.4308,
+      "eval_samples_per_second": 41.933,
+      "eval_steps_per_second": 10.483,
+      "step": 55
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 2.9683098591549296e-05,
+      "loss": 1.3501,
+      "step": 56
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.9661971830985917e-05,
+      "loss": 1.335,
+      "step": 57
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.9640845070422535e-05,
+      "loss": 1.3174,
+      "step": 58
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.9619718309859156e-05,
+      "loss": 1.2109,
+      "step": 59
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.9598591549295778e-05,
+      "loss": 1.2927,
+      "step": 60
+    },
+    {
+      "epoch": 0.16,
+      "eval_loss": 1.265234351158142,
+      "eval_runtime": 1.4332,
+      "eval_samples_per_second": 41.865,
+      "eval_steps_per_second": 10.466,
+      "step": 60
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.9577464788732395e-05,
+      "loss": 1.3411,
+      "step": 61
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.9556338028169017e-05,
+      "loss": 1.3955,
+      "step": 62
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.9535211267605635e-05,
+      "loss": 1.3262,
+      "step": 63
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.9514084507042256e-05,
+      "loss": 1.271,
+      "step": 64
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.9492957746478874e-05,
+      "loss": 1.356,
+      "step": 65
+    },
+    {
+      "epoch": 0.18,
+      "eval_loss": 1.2729166746139526,
+      "eval_runtime": 1.4466,
+      "eval_samples_per_second": 41.478,
+      "eval_steps_per_second": 10.369,
+      "step": 65
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.9471830985915495e-05,
+      "loss": 1.2725,
+      "step": 66
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.9450704225352113e-05,
+      "loss": 1.3506,
+      "step": 67
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.9429577464788734e-05,
+      "loss": 1.2678,
+      "step": 68
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.9408450704225355e-05,
+      "loss": 1.3501,
+      "step": 69
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.9387323943661973e-05,
+      "loss": 1.3433,
+      "step": 70
+    },
+    {
+      "epoch": 0.19,
+      "eval_loss": 1.2571614980697632,
+      "eval_runtime": 1.4285,
+      "eval_samples_per_second": 42.003,
+      "eval_steps_per_second": 10.501,
+      "step": 70
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.9366197183098594e-05,
+      "loss": 1.2864,
+      "step": 71
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.9345070422535212e-05,
+      "loss": 1.2654,
+      "step": 72
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.9323943661971834e-05,
+      "loss": 1.1904,
+      "step": 73
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.930281690140845e-05,
+      "loss": 1.3525,
+      "step": 74
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.9281690140845073e-05,
+      "loss": 1.3159,
+      "step": 75
+    },
+    {
+      "epoch": 0.21,
+      "eval_loss": 1.2537109851837158,
+      "eval_runtime": 1.4354,
+      "eval_samples_per_second": 41.801,
+      "eval_steps_per_second": 10.45,
+      "step": 75
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.926056338028169e-05,
+      "loss": 1.2974,
+      "step": 76
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.9239436619718312e-05,
+      "loss": 1.3271,
+      "step": 77
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.9218309859154933e-05,
+      "loss": 1.2312,
+      "step": 78
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.919718309859155e-05,
+      "loss": 1.3052,
+      "step": 79
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.9176056338028172e-05,
+      "loss": 1.291,
+      "step": 80
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 1.2563151121139526,
+      "eval_runtime": 1.434,
+      "eval_samples_per_second": 41.841,
+      "eval_steps_per_second": 10.46,
+      "step": 80
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.915492957746479e-05,
+      "loss": 1.3936,
+      "step": 81
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.913380281690141e-05,
+      "loss": 1.3757,
+      "step": 82
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.911267605633803e-05,
+      "loss": 1.2722,
+      "step": 83
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.909154929577465e-05,
+      "loss": 1.2693,
+      "step": 84
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.9070422535211268e-05,
+      "loss": 1.2332,
+      "step": 85
+    },
+    {
+      "epoch": 0.23,
+      "eval_loss": 1.2557942867279053,
+      "eval_runtime": 1.43,
+      "eval_samples_per_second": 41.959,
+      "eval_steps_per_second": 10.49,
+      "step": 85
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.904929577464789e-05,
+      "loss": 1.3708,
+      "step": 86
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.9028169014084507e-05,
+      "loss": 1.374,
+      "step": 87
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.900704225352113e-05,
+      "loss": 1.282,
+      "step": 88
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.898591549295775e-05,
+      "loss": 1.2688,
+      "step": 89
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.8964788732394368e-05,
+      "loss": 1.2637,
+      "step": 90
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.2471354007720947,
+      "eval_runtime": 1.4419,
+      "eval_samples_per_second": 41.611,
+      "eval_steps_per_second": 10.403,
+      "step": 90
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.894366197183099e-05,
+      "loss": 1.3179,
+      "step": 91
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.8922535211267607e-05,
+      "loss": 1.3,
+      "step": 92
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 2.8901408450704228e-05,
+      "loss": 1.2358,
+      "step": 93
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.8880281690140846e-05,
+      "loss": 1.2954,
+      "step": 94
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.8859154929577467e-05,
+      "loss": 1.3037,
+      "step": 95
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 1.2426432371139526,
+      "eval_runtime": 1.4301,
+      "eval_samples_per_second": 41.954,
+      "eval_steps_per_second": 10.489,
+      "step": 95
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 2.8838028169014085e-05,
+      "loss": 1.2231,
+      "step": 96
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.8816901408450706e-05,
+      "loss": 1.3064,
+      "step": 97
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.8795774647887328e-05,
+      "loss": 1.2656,
+      "step": 98
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.8774647887323946e-05,
+      "loss": 1.2505,
+      "step": 99
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 2.8753521126760567e-05,
+      "loss": 1.2241,
+      "step": 100
+    },
+    {
+      "epoch": 0.27,
+      "eval_loss": 1.2454427480697632,
+      "eval_runtime": 1.4351,
+      "eval_samples_per_second": 41.81,
+      "eval_steps_per_second": 10.452,
+      "step": 100
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.873239436619718e-05,
+      "loss": 1.3816,
+      "step": 101
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.8711267605633803e-05,
+      "loss": 1.2681,
+      "step": 102
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.869014084507042e-05,
+      "loss": 1.2505,
+      "step": 103
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 2.866901408450704e-05,
+      "loss": 1.3008,
+      "step": 104
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.864788732394366e-05,
+      "loss": 1.3772,
+      "step": 105
+    },
+    {
+      "epoch": 0.29,
+      "eval_loss": 1.2433593273162842,
+      "eval_runtime": 1.4392,
+      "eval_samples_per_second": 41.689,
+      "eval_steps_per_second": 10.422,
+      "step": 105
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.862676056338028e-05,
+      "loss": 1.2874,
+      "step": 106
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 2.8605633802816902e-05,
+      "loss": 1.2388,
+      "step": 107
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.858450704225352e-05,
+      "loss": 1.3691,
+      "step": 108
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.856338028169014e-05,
+      "loss": 1.3364,
+      "step": 109
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.854225352112676e-05,
+      "loss": 1.1995,
+      "step": 110
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 1.2469401359558105,
+      "eval_runtime": 1.4361,
+      "eval_samples_per_second": 41.78,
+      "eval_steps_per_second": 10.445,
+      "step": 110
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2.852112676056338e-05,
+      "loss": 1.2988,
+      "step": 111
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.8499999999999998e-05,
+      "loss": 1.3396,
+      "step": 112
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.847887323943662e-05,
+      "loss": 1.354,
+      "step": 113
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.8457746478873237e-05,
+      "loss": 1.1848,
+      "step": 114
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 2.843661971830986e-05,
+      "loss": 1.313,
+      "step": 115
+    },
+    {
+      "epoch": 0.31,
+      "eval_loss": 1.2402994632720947,
+      "eval_runtime": 1.4302,
+      "eval_samples_per_second": 41.952,
+      "eval_steps_per_second": 10.488,
+      "step": 115
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.841549295774648e-05,
+      "loss": 1.2319,
+      "step": 116
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.8394366197183098e-05,
+      "loss": 1.3093,
+      "step": 117
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 2.837323943661972e-05,
+      "loss": 1.3391,
+      "step": 118
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.8352112676056337e-05,
+      "loss": 1.2075,
+      "step": 119
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.8330985915492958e-05,
+      "loss": 1.1575,
+      "step": 120
+    },
+    {
+      "epoch": 0.33,
+      "eval_loss": 1.234765648841858,
+      "eval_runtime": 1.4323,
+      "eval_samples_per_second": 41.891,
+      "eval_steps_per_second": 10.473,
+      "step": 120
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.8309859154929576e-05,
+      "loss": 1.2388,
+      "step": 121
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.8288732394366197e-05,
+      "loss": 1.1997,
+      "step": 122
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.8267605633802815e-05,
+      "loss": 1.3142,
+      "step": 123
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.8246478873239436e-05,
+      "loss": 1.3218,
+      "step": 124
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.8225352112676058e-05,
+      "loss": 1.2168,
+      "step": 125
+    },
+    {
+      "epoch": 0.34,
+      "eval_loss": 1.2285807132720947,
+      "eval_runtime": 1.4322,
+      "eval_samples_per_second": 41.894,
+      "eval_steps_per_second": 10.473,
+      "step": 125
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.8204225352112675e-05,
+      "loss": 1.197,
+      "step": 126
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8183098591549297e-05,
+      "loss": 1.3164,
+      "step": 127
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8161971830985915e-05,
+      "loss": 1.3022,
+      "step": 128
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8140845070422536e-05,
+      "loss": 1.2588,
+      "step": 129
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.8119718309859154e-05,
+      "loss": 1.2776,
+      "step": 130
+    },
+    {
+      "epoch": 0.36,
+      "eval_loss": 1.2305989265441895,
+      "eval_runtime": 1.4359,
+      "eval_samples_per_second": 41.786,
+      "eval_steps_per_second": 10.447,
+      "step": 130
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.8098591549295775e-05,
+      "loss": 1.2236,
+      "step": 131
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.8077464788732393e-05,
+      "loss": 1.3384,
+      "step": 132
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.8056338028169014e-05,
+      "loss": 1.1338,
+      "step": 133
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.8035211267605635e-05,
+      "loss": 1.2407,
+      "step": 134
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.8014084507042253e-05,
+      "loss": 1.2847,
+      "step": 135
+    },
+    {
+      "epoch": 0.37,
+      "eval_loss": 1.2239583730697632,
+      "eval_runtime": 1.4377,
+      "eval_samples_per_second": 41.735,
+      "eval_steps_per_second": 10.434,
+      "step": 135
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.7992957746478874e-05,
+      "loss": 1.2094,
+      "step": 136
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.7971830985915492e-05,
+      "loss": 1.2795,
+      "step": 137
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.7950704225352114e-05,
+      "loss": 1.2324,
+      "step": 138
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.792957746478873e-05,
+      "loss": 1.2852,
+      "step": 139
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.7908450704225353e-05,
+      "loss": 1.25,
+      "step": 140
+    },
+    {
+      "epoch": 0.38,
+      "eval_loss": 1.2257161140441895,
+      "eval_runtime": 1.4271,
+      "eval_samples_per_second": 42.043,
+      "eval_steps_per_second": 10.511,
+      "step": 140
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.788732394366197e-05,
+      "loss": 1.1743,
+      "step": 141
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.7866197183098592e-05,
+      "loss": 1.3108,
+      "step": 142
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.7845070422535213e-05,
+      "loss": 1.27,
+      "step": 143
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.782394366197183e-05,
+      "loss": 1.3096,
+      "step": 144
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.7802816901408452e-05,
+      "loss": 1.1907,
+      "step": 145
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 1.230078101158142,
+      "eval_runtime": 1.4371,
+      "eval_samples_per_second": 41.751,
+      "eval_steps_per_second": 10.438,
+      "step": 145
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.778169014084507e-05,
+      "loss": 1.1763,
+      "step": 146
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.776056338028169e-05,
+      "loss": 1.2397,
+      "step": 147
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.773943661971831e-05,
+      "loss": 1.2581,
+      "step": 148
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.771830985915493e-05,
+      "loss": 1.2305,
+      "step": 149
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.769718309859155e-05,
+      "loss": 1.314,
+      "step": 150
+    },
+    {
+      "epoch": 0.41,
+      "eval_loss": 1.2252603769302368,
+      "eval_runtime": 1.4346,
+      "eval_samples_per_second": 41.825,
+      "eval_steps_per_second": 10.456,
+      "step": 150
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.767605633802817e-05,
+      "loss": 1.2285,
+      "step": 151
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.765492957746479e-05,
+      "loss": 1.2856,
+      "step": 152
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.763380281690141e-05,
+      "loss": 1.2625,
+      "step": 153
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.761267605633803e-05,
+      "loss": 1.2461,
+      "step": 154
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.7591549295774648e-05,
+      "loss": 1.1582,
+      "step": 155
+    },
+    {
+      "epoch": 0.42,
+      "eval_loss": 1.2229167222976685,
+      "eval_runtime": 1.4263,
+      "eval_samples_per_second": 42.067,
+      "eval_steps_per_second": 10.517,
+      "step": 155
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.757042253521127e-05,
+      "loss": 1.0647,
+      "step": 156
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.7549295774647887e-05,
+      "loss": 1.2742,
+      "step": 157
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.7528169014084508e-05,
+      "loss": 1.1929,
+      "step": 158
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.7507042253521126e-05,
+      "loss": 1.2329,
+      "step": 159
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.7485915492957747e-05,
+      "loss": 1.2146,
+      "step": 160
+    },
+    {
+      "epoch": 0.44,
+      "eval_loss": 1.2184244394302368,
+      "eval_runtime": 1.4291,
+      "eval_samples_per_second": 41.984,
+      "eval_steps_per_second": 10.496,
+      "step": 160
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.746478873239437e-05,
+      "loss": 1.3047,
+      "step": 161
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 2.7443661971830986e-05,
+      "loss": 1.2275,
+      "step": 162
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.7422535211267608e-05,
+      "loss": 1.3066,
+      "step": 163
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.7401408450704226e-05,
+      "loss": 1.248,
+      "step": 164
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.7380281690140847e-05,
+      "loss": 1.2434,
+      "step": 165
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 1.2188801765441895,
+      "eval_runtime": 1.4359,
+      "eval_samples_per_second": 41.787,
+      "eval_steps_per_second": 10.447,
+      "step": 165
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 2.7359154929577465e-05,
+      "loss": 1.2832,
+      "step": 166
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.7338028169014086e-05,
+      "loss": 1.2917,
+      "step": 167
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.7316901408450704e-05,
+      "loss": 1.1401,
+      "step": 168
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.7295774647887325e-05,
+      "loss": 1.3042,
+      "step": 169
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.7274647887323943e-05,
+      "loss": 1.228,
+      "step": 170
+    },
+    {
+      "epoch": 0.47,
+      "eval_loss": 1.217187523841858,
+      "eval_runtime": 1.4356,
+      "eval_samples_per_second": 41.793,
+      "eval_steps_per_second": 10.448,
+      "step": 170
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.7253521126760564e-05,
+      "loss": 1.3154,
+      "step": 171
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.7232394366197185e-05,
+      "loss": 1.2849,
+      "step": 172
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.7211267605633803e-05,
+      "loss": 1.2803,
+      "step": 173
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.7190140845070425e-05,
+      "loss": 1.2161,
+      "step": 174
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.7169014084507042e-05,
+      "loss": 1.2734,
+      "step": 175
+    },
+    {
+      "epoch": 0.48,
+      "eval_loss": 1.2192057371139526,
+      "eval_runtime": 1.4309,
+      "eval_samples_per_second": 41.933,
+      "eval_steps_per_second": 10.483,
+      "step": 175
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.7147887323943664e-05,
+      "loss": 1.2029,
+      "step": 176
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.712676056338028e-05,
+      "loss": 1.2407,
+      "step": 177
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.7105633802816903e-05,
+      "loss": 1.146,
+      "step": 178
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.708450704225352e-05,
+      "loss": 1.1736,
+      "step": 179
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.7063380281690142e-05,
+      "loss": 1.1899,
+      "step": 180
+    },
+    {
+      "epoch": 0.49,
+      "eval_loss": 1.2156250476837158,
+      "eval_runtime": 1.4401,
+      "eval_samples_per_second": 41.663,
+      "eval_steps_per_second": 10.416,
+      "step": 180
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.7042253521126763e-05,
+      "loss": 1.2048,
+      "step": 181
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.702112676056338e-05,
+      "loss": 1.2163,
+      "step": 182
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.7000000000000002e-05,
+      "loss": 1.2424,
+      "step": 183
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.697887323943662e-05,
+      "loss": 1.2559,
+      "step": 184
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.695774647887324e-05,
+      "loss": 1.2351,
+      "step": 185
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 1.2085286378860474,
+      "eval_runtime": 1.4288,
+      "eval_samples_per_second": 41.992,
+      "eval_steps_per_second": 10.498,
+      "step": 185
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.693661971830986e-05,
+      "loss": 1.2964,
+      "step": 186
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.691549295774648e-05,
+      "loss": 1.2847,
+      "step": 187
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.68943661971831e-05,
+      "loss": 1.2422,
+      "step": 188
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.687323943661972e-05,
+      "loss": 1.22,
+      "step": 189
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.685211267605634e-05,
+      "loss": 1.1626,
+      "step": 190
+    },
+    {
+      "epoch": 0.52,
+      "eval_loss": 1.2111979722976685,
+      "eval_runtime": 1.4352,
+      "eval_samples_per_second": 41.806,
+      "eval_steps_per_second": 10.451,
+      "step": 190
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.683098591549296e-05,
+      "loss": 1.2283,
+      "step": 191
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.680985915492958e-05,
+      "loss": 1.2148,
+      "step": 192
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.6788732394366198e-05,
+      "loss": 1.1924,
+      "step": 193
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.676760563380282e-05,
+      "loss": 1.2214,
+      "step": 194
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.6746478873239437e-05,
+      "loss": 1.3418,
+      "step": 195
+    },
+    {
+      "epoch": 0.53,
+      "eval_loss": 1.2102214097976685,
+      "eval_runtime": 1.4354,
+      "eval_samples_per_second": 41.8,
+      "eval_steps_per_second": 10.45,
+      "step": 195
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.672535211267606e-05,
+      "loss": 1.1953,
+      "step": 196
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.6704225352112676e-05,
+      "loss": 1.2544,
+      "step": 197
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.6683098591549298e-05,
+      "loss": 1.2188,
+      "step": 198
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.666197183098592e-05,
+      "loss": 1.3069,
+      "step": 199
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6640845070422537e-05,
+      "loss": 1.166,
+      "step": 200
+    },
+    {
+      "epoch": 0.55,
+      "eval_loss": 1.2101562023162842,
+      "eval_runtime": 1.4326,
+      "eval_samples_per_second": 41.881,
+      "eval_steps_per_second": 10.47,
+      "step": 200
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6619718309859158e-05,
+      "loss": 1.1699,
+      "step": 201
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6598591549295776e-05,
+      "loss": 1.1731,
+      "step": 202
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6577464788732397e-05,
+      "loss": 1.3354,
+      "step": 203
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6556338028169015e-05,
+      "loss": 1.2603,
+      "step": 204
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6535211267605636e-05,
+      "loss": 1.1487,
+      "step": 205
+    },
+    {
+      "epoch": 0.56,
+      "eval_loss": 1.2061848640441895,
+      "eval_runtime": 1.4445,
+      "eval_samples_per_second": 41.537,
+      "eval_steps_per_second": 10.384,
+      "step": 205
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6514084507042254e-05,
+      "loss": 1.1214,
+      "step": 206
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.6492957746478875e-05,
+      "loss": 1.3149,
+      "step": 207
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.6471830985915497e-05,
+      "loss": 1.2737,
+      "step": 208
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.6450704225352114e-05,
+      "loss": 1.2397,
+      "step": 209
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.6429577464788736e-05,
+      "loss": 1.2202,
+      "step": 210
+    },
+    {
+      "epoch": 0.57,
+      "eval_loss": 1.2047525644302368,
+      "eval_runtime": 1.4324,
+      "eval_samples_per_second": 41.888,
+      "eval_steps_per_second": 10.472,
+      "step": 210
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.6408450704225354e-05,
+      "loss": 1.2939,
+      "step": 211
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.6387323943661975e-05,
+      "loss": 1.1943,
+      "step": 212
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.6366197183098593e-05,
+      "loss": 1.2336,
+      "step": 213
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.6345070422535214e-05,
+      "loss": 1.2444,
+      "step": 214
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.6323943661971832e-05,
+      "loss": 1.1924,
+      "step": 215
+    },
+    {
+      "epoch": 0.59,
+      "eval_loss": 1.2043620347976685,
+      "eval_runtime": 1.4355,
+      "eval_samples_per_second": 41.799,
+      "eval_steps_per_second": 10.45,
+      "step": 215
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.6302816901408453e-05,
+      "loss": 1.2258,
+      "step": 216
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.6281690140845074e-05,
+      "loss": 1.229,
+      "step": 217
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.6260563380281692e-05,
+      "loss": 1.1682,
+      "step": 218
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.623943661971831e-05,
+      "loss": 1.2185,
+      "step": 219
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.6218309859154928e-05,
+      "loss": 1.3457,
+      "step": 220
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 1.2009114027023315,
+      "eval_runtime": 1.4288,
+      "eval_samples_per_second": 41.994,
+      "eval_steps_per_second": 10.498,
+      "step": 220
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.619718309859155e-05,
+      "loss": 1.2004,
+      "step": 221
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.6176056338028167e-05,
+      "loss": 1.3208,
+      "step": 222
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.6154929577464788e-05,
+      "loss": 1.2263,
+      "step": 223
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.6133802816901406e-05,
+      "loss": 1.2146,
+      "step": 224
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.6112676056338027e-05,
+      "loss": 1.166,
+      "step": 225
+    },
+    {
+      "epoch": 0.62,
+      "eval_loss": 1.2024089097976685,
+      "eval_runtime": 1.432,
+      "eval_samples_per_second": 41.9,
+      "eval_steps_per_second": 10.475,
+      "step": 225
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.609154929577465e-05,
+      "loss": 1.2329,
+      "step": 226
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.6070422535211267e-05,
+      "loss": 1.2402,
+      "step": 227
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.6049295774647888e-05,
+      "loss": 1.2241,
+      "step": 228
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.6028169014084506e-05,
+      "loss": 1.2109,
+      "step": 229
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.6007042253521127e-05,
+      "loss": 1.2432,
+      "step": 230
+    },
+    {
+      "epoch": 0.63,
+      "eval_loss": 1.2039062976837158,
+      "eval_runtime": 1.4346,
+      "eval_samples_per_second": 41.823,
+      "eval_steps_per_second": 10.456,
+      "step": 230
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.5985915492957745e-05,
+      "loss": 1.3118,
+      "step": 231
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.5964788732394366e-05,
+      "loss": 1.2065,
+      "step": 232
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.5943661971830984e-05,
+      "loss": 1.2214,
+      "step": 233
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.5922535211267605e-05,
+      "loss": 1.2268,
+      "step": 234
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.5901408450704226e-05,
+      "loss": 1.2969,
+      "step": 235
+    },
+    {
+      "epoch": 0.64,
+      "eval_loss": 1.1999348402023315,
+      "eval_runtime": 1.4461,
+      "eval_samples_per_second": 41.49,
+      "eval_steps_per_second": 10.372,
+      "step": 235
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.5880281690140844e-05,
+      "loss": 1.1392,
+      "step": 236
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.5859154929577466e-05,
+      "loss": 1.2727,
+      "step": 237
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.5838028169014083e-05,
+      "loss": 1.2529,
+      "step": 238
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 2.5816901408450705e-05,
+      "loss": 1.1973,
+      "step": 239
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.5795774647887323e-05,
+      "loss": 1.2119,
+      "step": 240
+    },
+    {
+      "epoch": 0.66,
+      "eval_loss": 1.2022135257720947,
+      "eval_runtime": 1.4478,
+      "eval_samples_per_second": 41.442,
+      "eval_steps_per_second": 10.361,
+      "step": 240
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.5774647887323944e-05,
+      "loss": 1.2729,
+      "step": 241
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.575352112676056e-05,
+      "loss": 1.1356,
+      "step": 242
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.5732394366197183e-05,
+      "loss": 1.3494,
+      "step": 243
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.5711267605633804e-05,
+      "loss": 1.2461,
+      "step": 244
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.5690140845070422e-05,
+      "loss": 1.1863,
+      "step": 245
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 1.2085286378860474,
+      "eval_runtime": 1.4304,
+      "eval_samples_per_second": 41.946,
+      "eval_steps_per_second": 10.486,
+      "step": 245
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.5669014084507043e-05,
+      "loss": 1.155,
+      "step": 246
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.564788732394366e-05,
+      "loss": 1.1946,
+      "step": 247
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.5626760563380282e-05,
+      "loss": 1.1743,
+      "step": 248
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.56056338028169e-05,
+      "loss": 1.1975,
+      "step": 249
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.558450704225352e-05,
+      "loss": 1.1992,
+      "step": 250
+    },
+    {
+      "epoch": 0.68,
+      "eval_loss": 1.2000651359558105,
+      "eval_runtime": 1.4386,
+      "eval_samples_per_second": 41.707,
+      "eval_steps_per_second": 10.427,
+      "step": 250
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.556338028169014e-05,
+      "loss": 1.2861,
+      "step": 251
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.554225352112676e-05,
+      "loss": 1.3091,
+      "step": 252
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.552112676056338e-05,
+      "loss": 1.207,
+      "step": 253
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.55e-05,
+      "loss": 1.2073,
+      "step": 254
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.547887323943662e-05,
+      "loss": 1.1787,
+      "step": 255
+    },
+    {
+      "epoch": 0.7,
+      "eval_loss": 1.194726586341858,
+      "eval_runtime": 1.4385,
+      "eval_samples_per_second": 41.71,
+      "eval_steps_per_second": 10.427,
+      "step": 255
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.545774647887324e-05,
+      "loss": 1.1829,
+      "step": 256
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.543661971830986e-05,
+      "loss": 1.2554,
+      "step": 257
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.5415492957746478e-05,
+      "loss": 1.2048,
+      "step": 258
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.53943661971831e-05,
+      "loss": 1.2253,
+      "step": 259
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.5373239436619717e-05,
+      "loss": 1.2515,
+      "step": 260
+    },
+    {
+      "epoch": 0.71,
+      "eval_loss": 1.1985677480697632,
+      "eval_runtime": 1.4434,
+      "eval_samples_per_second": 41.568,
+      "eval_steps_per_second": 10.392,
+      "step": 260
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.535211267605634e-05,
+      "loss": 1.0757,
+      "step": 261
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.5330985915492956e-05,
+      "loss": 1.2317,
+      "step": 262
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.5309859154929578e-05,
+      "loss": 1.2808,
+      "step": 263
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.52887323943662e-05,
+      "loss": 1.1843,
+      "step": 264
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.5267605633802817e-05,
+      "loss": 1.3093,
+      "step": 265
+    },
+    {
+      "epoch": 0.73,
+      "eval_loss": 1.1974608898162842,
+      "eval_runtime": 1.4268,
+      "eval_samples_per_second": 42.051,
+      "eval_steps_per_second": 10.513,
+      "step": 265
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.5246478873239438e-05,
+      "loss": 1.2212,
+      "step": 266
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.5225352112676056e-05,
+      "loss": 1.2544,
+      "step": 267
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.5204225352112677e-05,
+      "loss": 1.2085,
+      "step": 268
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5183098591549295e-05,
+      "loss": 1.1968,
+      "step": 269
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5161971830985916e-05,
+      "loss": 1.2449,
+      "step": 270
+    },
+    {
+      "epoch": 0.74,
+      "eval_loss": 1.191992163658142,
+      "eval_runtime": 1.4269,
+      "eval_samples_per_second": 42.049,
+      "eval_steps_per_second": 10.512,
+      "step": 270
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5140845070422534e-05,
+      "loss": 1.2354,
+      "step": 271
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5119718309859155e-05,
+      "loss": 1.3428,
+      "step": 272
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5098591549295777e-05,
+      "loss": 1.2842,
+      "step": 273
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5077464788732394e-05,
+      "loss": 1.2488,
+      "step": 274
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5056338028169016e-05,
+      "loss": 1.2515,
+      "step": 275
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 1.190039038658142,
+      "eval_runtime": 1.4323,
+      "eval_samples_per_second": 41.892,
+      "eval_steps_per_second": 10.473,
+      "step": 275
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.5035211267605634e-05,
+      "loss": 1.3669,
+      "step": 276
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.5014084507042255e-05,
+      "loss": 1.1516,
+      "step": 277
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4992957746478873e-05,
+      "loss": 1.2585,
+      "step": 278
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4971830985915494e-05,
+      "loss": 1.2407,
+      "step": 279
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4950704225352112e-05,
+      "loss": 1.1912,
+      "step": 280
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 1.1897135972976685,
+      "eval_runtime": 1.4284,
+      "eval_samples_per_second": 42.006,
+      "eval_steps_per_second": 10.501,
+      "step": 280
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4929577464788733e-05,
+      "loss": 1.2305,
+      "step": 281
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4908450704225354e-05,
+      "loss": 1.2048,
+      "step": 282
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4887323943661972e-05,
+      "loss": 1.3276,
+      "step": 283
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4866197183098593e-05,
+      "loss": 1.2456,
+      "step": 284
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.484507042253521e-05,
+      "loss": 1.0989,
+      "step": 285
+    },
+    {
+      "epoch": 0.78,
+      "eval_loss": 1.1903645992279053,
+      "eval_runtime": 1.4346,
+      "eval_samples_per_second": 41.823,
+      "eval_steps_per_second": 10.456,
+      "step": 285
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4823943661971833e-05,
+      "loss": 1.2542,
+      "step": 286
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.480281690140845e-05,
+      "loss": 1.198,
+      "step": 287
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.4781690140845072e-05,
+      "loss": 1.1479,
+      "step": 288
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.476056338028169e-05,
+      "loss": 1.2915,
+      "step": 289
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.473943661971831e-05,
+      "loss": 1.1667,
+      "step": 290
+    },
+    {
+      "epoch": 0.79,
+      "eval_loss": 1.1932942867279053,
+      "eval_runtime": 1.434,
+      "eval_samples_per_second": 41.842,
+      "eval_steps_per_second": 10.461,
+      "step": 290
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.4718309859154932e-05,
+      "loss": 1.1479,
+      "step": 291
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.469718309859155e-05,
+      "loss": 1.3481,
+      "step": 292
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.467605633802817e-05,
+      "loss": 1.2959,
+      "step": 293
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.465492957746479e-05,
+      "loss": 1.2476,
+      "step": 294
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.463380281690141e-05,
+      "loss": 1.2771,
+      "step": 295
+    },
+    {
+      "epoch": 0.81,
+      "eval_loss": 1.1938802003860474,
+      "eval_runtime": 1.4332,
+      "eval_samples_per_second": 41.864,
+      "eval_steps_per_second": 10.466,
+      "step": 295
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.4612676056338028e-05,
+      "loss": 1.2058,
+      "step": 296
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.459154929577465e-05,
+      "loss": 1.1758,
+      "step": 297
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.4570422535211267e-05,
+      "loss": 1.166,
+      "step": 298
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.454929577464789e-05,
+      "loss": 1.2974,
+      "step": 299
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.452816901408451e-05,
+      "loss": 1.2161,
+      "step": 300
+    },
+    {
+      "epoch": 0.82,
+      "eval_loss": 1.1921875476837158,
+      "eval_runtime": 1.4317,
+      "eval_samples_per_second": 41.907,
+      "eval_steps_per_second": 10.477,
+      "step": 300
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.4507042253521128e-05,
+      "loss": 1.3,
+      "step": 301
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.448591549295775e-05,
+      "loss": 1.2354,
+      "step": 302
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.4464788732394367e-05,
+      "loss": 1.2419,
+      "step": 303
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.4443661971830988e-05,
+      "loss": 1.2397,
+      "step": 304
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.4422535211267606e-05,
+      "loss": 1.1924,
+      "step": 305
+    },
+    {
+      "epoch": 0.84,
+      "eval_loss": 1.1932291984558105,
+      "eval_runtime": 1.4334,
+      "eval_samples_per_second": 41.859,
+      "eval_steps_per_second": 10.465,
+      "step": 305
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.4401408450704227e-05,
+      "loss": 1.1882,
+      "step": 306
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.4380281690140845e-05,
+      "loss": 1.3137,
+      "step": 307
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.4359154929577466e-05,
+      "loss": 1.2275,
+      "step": 308
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.4338028169014088e-05,
+      "loss": 1.2288,
+      "step": 309
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.4316901408450705e-05,
+      "loss": 1.1711,
+      "step": 310
+    },
+    {
+      "epoch": 0.85,
+      "eval_loss": 1.1925780773162842,
+      "eval_runtime": 1.4464,
+      "eval_samples_per_second": 41.484,
+      "eval_steps_per_second": 10.371,
+      "step": 310
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.4295774647887327e-05,
+      "loss": 1.292,
+      "step": 311
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.4274647887323945e-05,
+      "loss": 1.2173,
+      "step": 312
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 2.4253521126760566e-05,
+      "loss": 1.218,
+      "step": 313
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 2.4232394366197184e-05,
+      "loss": 1.1931,
+      "step": 314
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 2.4211267605633805e-05,
+      "loss": 1.2764,
+      "step": 315
+    },
+    {
+      "epoch": 0.86,
+      "eval_loss": 1.1875,
+      "eval_runtime": 1.4513,
+      "eval_samples_per_second": 41.344,
+      "eval_steps_per_second": 10.336,
+      "step": 315
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.4190140845070423e-05,
+      "loss": 1.1992,
+      "step": 316
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.4169014084507044e-05,
+      "loss": 1.1848,
+      "step": 317
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.4147887323943665e-05,
+      "loss": 1.1208,
+      "step": 318
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.4126760563380283e-05,
+      "loss": 1.3013,
+      "step": 319
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4105633802816904e-05,
+      "loss": 1.177,
+      "step": 320
+    },
+    {
+      "epoch": 0.88,
+      "eval_loss": 1.1837239265441895,
+      "eval_runtime": 1.4348,
+      "eval_samples_per_second": 41.816,
+      "eval_steps_per_second": 10.454,
+      "step": 320
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4084507042253522e-05,
+      "loss": 1.1162,
+      "step": 321
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4063380281690144e-05,
+      "loss": 1.2046,
+      "step": 322
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.404225352112676e-05,
+      "loss": 1.1917,
+      "step": 323
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.4021126760563383e-05,
+      "loss": 1.1958,
+      "step": 324
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.4e-05,
+      "loss": 1.3025,
+      "step": 325
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 1.180078148841858,
+      "eval_runtime": 1.4319,
+      "eval_samples_per_second": 41.903,
+      "eval_steps_per_second": 10.476,
+      "step": 325
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.3978873239436622e-05,
+      "loss": 1.2603,
+      "step": 326
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.3957746478873243e-05,
+      "loss": 1.2471,
+      "step": 327
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.393661971830986e-05,
+      "loss": 1.1423,
+      "step": 328
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.3915492957746482e-05,
+      "loss": 1.2944,
+      "step": 329
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 2.38943661971831e-05,
+      "loss": 1.1836,
+      "step": 330
+    },
+    {
+      "epoch": 0.9,
+      "eval_loss": 1.1818358898162842,
+      "eval_runtime": 1.4368,
+      "eval_samples_per_second": 41.759,
+      "eval_steps_per_second": 10.44,
+      "step": 330
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 2.387323943661972e-05,
+      "loss": 1.2336,
+      "step": 331
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 2.385211267605634e-05,
+      "loss": 1.0251,
+      "step": 332
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 2.383098591549296e-05,
+      "loss": 1.1006,
+      "step": 333
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 2.380985915492958e-05,
+      "loss": 1.2842,
+      "step": 334
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.37887323943662e-05,
+      "loss": 1.1885,
+      "step": 335
+    },
+    {
+      "epoch": 0.92,
+      "eval_loss": 1.1787760257720947,
+      "eval_runtime": 1.4404,
+      "eval_samples_per_second": 41.655,
+      "eval_steps_per_second": 10.414,
+      "step": 335
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.3767605633802817e-05,
+      "loss": 1.1519,
+      "step": 336
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.3746478873239435e-05,
+      "loss": 1.2004,
+      "step": 337
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3725352112676057e-05,
+      "loss": 1.2603,
+      "step": 338
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3704225352112674e-05,
+      "loss": 1.1917,
+      "step": 339
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3683098591549296e-05,
+      "loss": 1.1919,
+      "step": 340
+    },
+    {
+      "epoch": 0.93,
+      "eval_loss": 1.1794922351837158,
+      "eval_runtime": 1.4286,
+      "eval_samples_per_second": 41.999,
+      "eval_steps_per_second": 10.5,
+      "step": 340
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3661971830985914e-05,
+      "loss": 1.1816,
+      "step": 341
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.3640845070422535e-05,
+      "loss": 1.292,
+      "step": 342
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.3619718309859153e-05,
+      "loss": 1.0908,
+      "step": 343
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.3598591549295774e-05,
+      "loss": 1.2385,
+      "step": 344
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.3577464788732392e-05,
+      "loss": 1.2476,
+      "step": 345
+    },
+    {
+      "epoch": 0.94,
+      "eval_loss": 1.177734375,
+      "eval_runtime": 1.4399,
+      "eval_samples_per_second": 41.669,
+      "eval_steps_per_second": 10.417,
+      "step": 345
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.3556338028169013e-05,
+      "loss": 1.2288,
+      "step": 346
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.3535211267605634e-05,
+      "loss": 1.2603,
+      "step": 347
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.3514084507042252e-05,
+      "loss": 1.1631,
+      "step": 348
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.3492957746478874e-05,
+      "loss": 1.2427,
+      "step": 349
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.347183098591549e-05,
+      "loss": 1.2163,
+      "step": 350
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 1.1764322519302368,
+      "eval_runtime": 1.4393,
+      "eval_samples_per_second": 41.688,
+      "eval_steps_per_second": 10.422,
+      "step": 350
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.3450704225352113e-05,
+      "loss": 1.1926,
+      "step": 351
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.342957746478873e-05,
+      "loss": 1.2625,
+      "step": 352
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.3408450704225352e-05,
+      "loss": 1.1846,
+      "step": 353
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.338732394366197e-05,
+      "loss": 1.2764,
+      "step": 354
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.336619718309859e-05,
+      "loss": 1.1787,
+      "step": 355
+    },
+    {
+      "epoch": 0.97,
+      "eval_loss": 1.1746094226837158,
+      "eval_runtime": 1.4348,
+      "eval_samples_per_second": 41.818,
+      "eval_steps_per_second": 10.454,
+      "step": 355
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.3345070422535212e-05,
+      "loss": 1.1868,
+      "step": 356
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.332394366197183e-05,
+      "loss": 1.0908,
+      "step": 357
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.330281690140845e-05,
+      "loss": 1.1816,
+      "step": 358
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.328169014084507e-05,
+      "loss": 1.3042,
+      "step": 359
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.326056338028169e-05,
+      "loss": 1.2336,
+      "step": 360
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 1.1708333492279053,
+      "eval_runtime": 1.4308,
+      "eval_samples_per_second": 41.936,
+      "eval_steps_per_second": 10.484,
+      "step": 360
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.3239436619718308e-05,
+      "loss": 1.1492,
+      "step": 361
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.321830985915493e-05,
+      "loss": 1.1057,
+      "step": 362
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.3197183098591547e-05,
+      "loss": 1.1528,
+      "step": 363
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.317605633802817e-05,
+      "loss": 1.1187,
+      "step": 364
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.315492957746479e-05,
+      "loss": 1.2441,
+      "step": 365
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.1701823472976685,
+      "eval_runtime": 1.4491,
+      "eval_samples_per_second": 41.405,
+      "eval_steps_per_second": 10.351,
+      "step": 365
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.3133802816901408e-05,
+      "loss": 0.9822,
+      "step": 366
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.311267605633803e-05,
+      "loss": 0.9307,
+      "step": 367
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 2.3091549295774647e-05,
+      "loss": 0.8552,
+      "step": 368
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 2.3070422535211268e-05,
+      "loss": 0.8684,
+      "step": 369
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 2.3049295774647886e-05,
+      "loss": 0.8901,
+      "step": 370
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 1.2126953601837158,
+      "eval_runtime": 1.4444,
+      "eval_samples_per_second": 41.539,
+      "eval_steps_per_second": 10.385,
+      "step": 370
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 2.3028169014084507e-05,
+      "loss": 0.7668,
+      "step": 371
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 2.3007042253521125e-05,
+      "loss": 0.8608,
+      "step": 372
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 2.2985915492957746e-05,
+      "loss": 0.8435,
+      "step": 373
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 2.2964788732394368e-05,
+      "loss": 0.8599,
+      "step": 374
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 2.2943661971830986e-05,
+      "loss": 0.821,
+      "step": 375
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 1.213281273841858,
+      "eval_runtime": 1.4333,
+      "eval_samples_per_second": 41.861,
+      "eval_steps_per_second": 10.465,
+      "step": 375
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 2.2922535211267607e-05,
+      "loss": 0.8513,
+      "step": 376
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 2.2901408450704225e-05,
+      "loss": 0.8655,
+      "step": 377
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 2.2880281690140846e-05,
+      "loss": 0.8665,
+      "step": 378
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 2.2859154929577464e-05,
+      "loss": 0.9539,
+      "step": 379
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 2.2838028169014085e-05,
+      "loss": 0.8265,
+      "step": 380
+    },
+    {
+      "epoch": 1.04,
+      "eval_loss": 1.2233723402023315,
+      "eval_runtime": 1.448,
+      "eval_samples_per_second": 41.436,
+      "eval_steps_per_second": 10.359,
+      "step": 380
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 2.2816901408450703e-05,
+      "loss": 0.8921,
+      "step": 381
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2.2795774647887324e-05,
+      "loss": 0.9325,
+      "step": 382
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2.2774647887323945e-05,
+      "loss": 0.8176,
+      "step": 383
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2.2753521126760563e-05,
+      "loss": 0.887,
+      "step": 384
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2.2732394366197185e-05,
+      "loss": 0.8979,
+      "step": 385
+    },
+    {
+      "epoch": 1.05,
+      "eval_loss": 1.2147786617279053,
+      "eval_runtime": 1.4364,
+      "eval_samples_per_second": 41.772,
+      "eval_steps_per_second": 10.443,
+      "step": 385
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 2.2711267605633802e-05,
+      "loss": 0.8177,
+      "step": 386
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 2.2690140845070424e-05,
+      "loss": 0.8152,
+      "step": 387
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 2.266901408450704e-05,
+      "loss": 0.8688,
+      "step": 388
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 2.2647887323943663e-05,
+      "loss": 0.8811,
+      "step": 389
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 2.262676056338028e-05,
+      "loss": 0.905,
+      "step": 390
+    },
+    {
+      "epoch": 1.07,
+      "eval_loss": 1.2212239503860474,
+      "eval_runtime": 1.4335,
+      "eval_samples_per_second": 41.855,
+      "eval_steps_per_second": 10.464,
+      "step": 390
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 2.2605633802816902e-05,
+      "loss": 0.8643,
+      "step": 391
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 2.2584507042253523e-05,
+      "loss": 0.9551,
+      "step": 392
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.256338028169014e-05,
+      "loss": 0.802,
+      "step": 393
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.2542253521126762e-05,
+      "loss": 0.8674,
+      "step": 394
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.252112676056338e-05,
+      "loss": 0.9437,
+      "step": 395
+    },
+    {
+      "epoch": 1.08,
+      "eval_loss": 1.210546851158142,
+      "eval_runtime": 1.4355,
+      "eval_samples_per_second": 41.796,
+      "eval_steps_per_second": 10.449,
+      "step": 395
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 2.25e-05,
+      "loss": 0.8477,
+      "step": 396
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 2.247887323943662e-05,
+      "loss": 0.876,
+      "step": 397
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 2.245774647887324e-05,
+      "loss": 0.8899,
+      "step": 398
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 2.243661971830986e-05,
+      "loss": 0.8422,
+      "step": 399
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 2.241549295774648e-05,
+      "loss": 0.8831,
+      "step": 400
+    },
+    {
+      "epoch": 1.1,
+      "eval_loss": 1.2247395515441895,
+      "eval_runtime": 1.4367,
+      "eval_samples_per_second": 41.763,
+      "eval_steps_per_second": 10.441,
+      "step": 400
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 2.23943661971831e-05,
+      "loss": 0.917,
+      "step": 401
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 2.237323943661972e-05,
+      "loss": 0.8765,
+      "step": 402
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 2.235211267605634e-05,
+      "loss": 0.8333,
+      "step": 403
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 2.2330985915492958e-05,
+      "loss": 0.8477,
+      "step": 404
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 2.230985915492958e-05,
+      "loss": 0.8745,
+      "step": 405
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 1.2182291746139526,
+      "eval_runtime": 1.4515,
+      "eval_samples_per_second": 41.336,
+      "eval_steps_per_second": 10.334,
+      "step": 405
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 2.2288732394366197e-05,
+      "loss": 0.8936,
+      "step": 406
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 2.2267605633802818e-05,
+      "loss": 0.8827,
+      "step": 407
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 2.2246478873239436e-05,
+      "loss": 0.7937,
+      "step": 408
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 2.2225352112676057e-05,
+      "loss": 0.8853,
+      "step": 409
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 2.220422535211268e-05,
+      "loss": 0.9104,
+      "step": 410
+    },
+    {
+      "epoch": 1.12,
+      "eval_loss": 1.2125650644302368,
+      "eval_runtime": 1.4494,
+      "eval_samples_per_second": 41.397,
+      "eval_steps_per_second": 10.349,
+      "step": 410
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 2.2183098591549297e-05,
+      "loss": 0.8668,
+      "step": 411
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 2.2161971830985918e-05,
+      "loss": 0.9046,
+      "step": 412
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 2.2140845070422536e-05,
+      "loss": 0.9089,
+      "step": 413
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 2.2119718309859157e-05,
+      "loss": 0.9058,
+      "step": 414
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 2.2098591549295775e-05,
+      "loss": 0.8799,
+      "step": 415
+    },
+    {
+      "epoch": 1.14,
+      "eval_loss": 1.2117838859558105,
+      "eval_runtime": 1.4538,
+      "eval_samples_per_second": 41.271,
+      "eval_steps_per_second": 10.318,
+      "step": 415
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 2.2077464788732396e-05,
+      "loss": 0.8544,
+      "step": 416
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 2.2056338028169014e-05,
+      "loss": 0.802,
+      "step": 417
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 2.2035211267605635e-05,
+      "loss": 0.8296,
+      "step": 418
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 2.2014084507042253e-05,
+      "loss": 0.8459,
+      "step": 419
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 2.1992957746478874e-05,
+      "loss": 0.9197,
+      "step": 420
+    },
+    {
+      "epoch": 1.15,
+      "eval_loss": 1.2139322757720947,
+      "eval_runtime": 1.4304,
+      "eval_samples_per_second": 41.947,
+      "eval_steps_per_second": 10.487,
+      "step": 420
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 2.1971830985915496e-05,
+      "loss": 0.8813,
+      "step": 421
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 2.1950704225352113e-05,
+      "loss": 0.8796,
+      "step": 422
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 2.1929577464788735e-05,
+      "loss": 0.8816,
+      "step": 423
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 2.1908450704225353e-05,
+      "loss": 0.9091,
+      "step": 424
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 2.1887323943661974e-05,
+      "loss": 0.8579,
+      "step": 425
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 1.2127604484558105,
+      "eval_runtime": 1.4334,
+      "eval_samples_per_second": 41.86,
+      "eval_steps_per_second": 10.465,
+      "step": 425
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 2.186619718309859e-05,
+      "loss": 0.8621,
+      "step": 426
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 2.1845070422535213e-05,
+      "loss": 0.8894,
+      "step": 427
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 2.182394366197183e-05,
+      "loss": 0.8411,
+      "step": 428
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 2.1802816901408452e-05,
+      "loss": 0.8411,
+      "step": 429
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 2.1781690140845073e-05,
+      "loss": 0.9041,
+      "step": 430
+    },
+    {
+      "epoch": 1.18,
+      "eval_loss": 1.2069661617279053,
+      "eval_runtime": 1.436,
+      "eval_samples_per_second": 41.781,
+      "eval_steps_per_second": 10.445,
+      "step": 430
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 2.176056338028169e-05,
+      "loss": 0.9076,
+      "step": 431
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 2.1739436619718312e-05,
+      "loss": 0.8462,
+      "step": 432
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 2.171830985915493e-05,
+      "loss": 0.8042,
+      "step": 433
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 2.169718309859155e-05,
+      "loss": 0.9231,
+      "step": 434
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 2.167605633802817e-05,
+      "loss": 0.8839,
+      "step": 435
+    },
+    {
+      "epoch": 1.19,
+      "eval_loss": 1.2218749523162842,
+      "eval_runtime": 1.4295,
+      "eval_samples_per_second": 41.973,
+      "eval_steps_per_second": 10.493,
+      "step": 435
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 2.165492957746479e-05,
+      "loss": 0.8591,
+      "step": 436
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.163380281690141e-05,
+      "loss": 0.8113,
+      "step": 437
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.161267605633803e-05,
+      "loss": 0.8362,
+      "step": 438
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.159154929577465e-05,
+      "loss": 0.8313,
+      "step": 439
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.157042253521127e-05,
+      "loss": 0.8557,
+      "step": 440
+    },
+    {
+      "epoch": 1.2,
+      "eval_loss": 1.2255208492279053,
+      "eval_runtime": 1.427,
+      "eval_samples_per_second": 42.047,
+      "eval_steps_per_second": 10.512,
+      "step": 440
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.154929577464789e-05,
+      "loss": 0.9014,
+      "step": 441
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.1528169014084508e-05,
+      "loss": 0.8671,
+      "step": 442
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.150704225352113e-05,
+      "loss": 0.8746,
+      "step": 443
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.1485915492957747e-05,
+      "loss": 0.856,
+      "step": 444
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.146478873239437e-05,
+      "loss": 0.9116,
+      "step": 445
+    },
+    {
+      "epoch": 1.22,
+      "eval_loss": 1.2089192867279053,
+      "eval_runtime": 1.436,
+      "eval_samples_per_second": 41.782,
+      "eval_steps_per_second": 10.445,
+      "step": 445
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.1443661971830986e-05,
+      "loss": 0.9019,
+      "step": 446
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.1422535211267608e-05,
+      "loss": 0.8657,
+      "step": 447
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.140140845070423e-05,
+      "loss": 0.8328,
+      "step": 448
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.1380281690140847e-05,
+      "loss": 0.8241,
+      "step": 449
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.1359154929577468e-05,
+      "loss": 0.8831,
+      "step": 450
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 1.2138020992279053,
+      "eval_runtime": 1.4351,
+      "eval_samples_per_second": 41.809,
+      "eval_steps_per_second": 10.452,
+      "step": 450
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.1338028169014086e-05,
+      "loss": 0.8911,
+      "step": 451
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.1316901408450707e-05,
+      "loss": 0.92,
+      "step": 452
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.1295774647887325e-05,
+      "loss": 0.7701,
+      "step": 453
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.1274647887323946e-05,
+      "loss": 0.9236,
+      "step": 454
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.1253521126760564e-05,
+      "loss": 0.8821,
+      "step": 455
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 1.204492211341858,
+      "eval_runtime": 1.4398,
+      "eval_samples_per_second": 41.672,
+      "eval_steps_per_second": 10.418,
+      "step": 455
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.1232394366197182e-05,
+      "loss": 0.8435,
+      "step": 456
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.1211267605633803e-05,
+      "loss": 0.8264,
+      "step": 457
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.119014084507042e-05,
+      "loss": 0.8826,
+      "step": 458
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.1169014084507042e-05,
+      "loss": 0.8324,
+      "step": 459
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.114788732394366e-05,
+      "loss": 0.8179,
+      "step": 460
+    },
+    {
+      "epoch": 1.26,
+      "eval_loss": 1.2137370109558105,
+      "eval_runtime": 1.4314,
+      "eval_samples_per_second": 41.917,
+      "eval_steps_per_second": 10.479,
+      "step": 460
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.112676056338028e-05,
+      "loss": 0.8853,
+      "step": 461
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.11056338028169e-05,
+      "loss": 0.8638,
+      "step": 462
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.108450704225352e-05,
+      "loss": 0.7971,
+      "step": 463
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.106338028169014e-05,
+      "loss": 0.9268,
+      "step": 464
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.104225352112676e-05,
+      "loss": 0.946,
+      "step": 465
+    },
+    {
+      "epoch": 1.27,
+      "eval_loss": 1.2068359851837158,
+      "eval_runtime": 1.44,
+      "eval_samples_per_second": 41.665,
+      "eval_steps_per_second": 10.416,
+      "step": 465
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.102112676056338e-05,
+      "loss": 0.8165,
+      "step": 466
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.1e-05,
+      "loss": 0.8152,
+      "step": 467
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.097887323943662e-05,
+      "loss": 0.8623,
+      "step": 468
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.0957746478873238e-05,
+      "loss": 0.8386,
+      "step": 469
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.093661971830986e-05,
+      "loss": 0.8969,
+      "step": 470
+    },
+    {
+      "epoch": 1.29,
+      "eval_loss": 1.2259114980697632,
+      "eval_runtime": 1.4412,
+      "eval_samples_per_second": 41.633,
+      "eval_steps_per_second": 10.408,
+      "step": 470
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.0915492957746477e-05,
+      "loss": 0.8374,
+      "step": 471
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.08943661971831e-05,
+      "loss": 0.8318,
+      "step": 472
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.0873239436619716e-05,
+      "loss": 0.8282,
+      "step": 473
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.0852112676056337e-05,
+      "loss": 0.9204,
+      "step": 474
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.083098591549296e-05,
+      "loss": 0.8723,
+      "step": 475
+    },
+    {
+      "epoch": 1.3,
+      "eval_loss": 1.2182291746139526,
+      "eval_runtime": 1.4468,
+      "eval_samples_per_second": 41.471,
+      "eval_steps_per_second": 10.368,
+      "step": 475
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.0809859154929577e-05,
+      "loss": 0.8938,
+      "step": 476
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.0788732394366198e-05,
+      "loss": 0.7826,
+      "step": 477
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.0767605633802816e-05,
+      "loss": 0.8999,
+      "step": 478
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.0746478873239437e-05,
+      "loss": 0.8669,
+      "step": 479
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.0725352112676055e-05,
+      "loss": 0.8767,
+      "step": 480
+    },
+    {
+      "epoch": 1.31,
+      "eval_loss": 1.2123697996139526,
+      "eval_runtime": 1.4312,
+      "eval_samples_per_second": 41.923,
+      "eval_steps_per_second": 10.481,
+      "step": 480
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.0704225352112676e-05,
+      "loss": 0.8484,
+      "step": 481
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.0683098591549294e-05,
+      "loss": 0.9216,
+      "step": 482
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.0661971830985915e-05,
+      "loss": 0.8774,
+      "step": 483
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.0640845070422536e-05,
+      "loss": 0.8782,
+      "step": 484
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.0619718309859154e-05,
+      "loss": 0.9365,
+      "step": 485
+    },
+    {
+      "epoch": 1.33,
+      "eval_loss": 1.2236979007720947,
+      "eval_runtime": 1.4395,
+      "eval_samples_per_second": 41.681,
+      "eval_steps_per_second": 10.42,
+      "step": 485
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.0598591549295776e-05,
+      "loss": 0.8958,
+      "step": 486
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.0577464788732393e-05,
+      "loss": 0.9382,
+      "step": 487
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.0556338028169015e-05,
+      "loss": 0.8831,
+      "step": 488
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.0535211267605633e-05,
+      "loss": 0.9192,
+      "step": 489
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.0514084507042254e-05,
+      "loss": 0.8523,
+      "step": 490
+    },
+    {
+      "epoch": 1.34,
+      "eval_loss": 1.2074869871139526,
+      "eval_runtime": 1.4432,
+      "eval_samples_per_second": 41.574,
+      "eval_steps_per_second": 10.393,
+      "step": 490
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.0492957746478872e-05,
+      "loss": 0.8975,
+      "step": 491
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.0471830985915493e-05,
+      "loss": 0.8699,
+      "step": 492
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.0450704225352114e-05,
+      "loss": 0.8765,
+      "step": 493
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.0429577464788732e-05,
+      "loss": 0.8689,
+      "step": 494
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.0408450704225353e-05,
+      "loss": 0.9319,
+      "step": 495
+    },
+    {
+      "epoch": 1.36,
+      "eval_loss": 1.2223957777023315,
+      "eval_runtime": 1.4367,
+      "eval_samples_per_second": 41.763,
+      "eval_steps_per_second": 10.441,
+      "step": 495
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.038732394366197e-05,
+      "loss": 0.8204,
+      "step": 496
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.0366197183098592e-05,
+      "loss": 0.892,
+      "step": 497
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.034507042253521e-05,
+      "loss": 0.8547,
+      "step": 498
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.032394366197183e-05,
+      "loss": 0.803,
+      "step": 499
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.030281690140845e-05,
+      "loss": 0.8696,
+      "step": 500
+    },
+    {
+      "epoch": 1.37,
+      "eval_loss": 1.2244791984558105,
+      "eval_runtime": 1.4387,
+      "eval_samples_per_second": 41.705,
+      "eval_steps_per_second": 10.426,
+      "step": 500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.028169014084507e-05,
+      "loss": 0.7708,
+      "step": 501
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.026056338028169e-05,
+      "loss": 0.8496,
+      "step": 502
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.023943661971831e-05,
+      "loss": 0.8406,
+      "step": 503
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.021830985915493e-05,
+      "loss": 0.9111,
+      "step": 504
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.019718309859155e-05,
+      "loss": 0.8455,
+      "step": 505
+    },
+    {
+      "epoch": 1.38,
+      "eval_loss": 1.2137370109558105,
+      "eval_runtime": 1.4462,
+      "eval_samples_per_second": 41.489,
+      "eval_steps_per_second": 10.372,
+      "step": 505
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.017605633802817e-05,
+      "loss": 0.9314,
+      "step": 506
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.0154929577464788e-05,
+      "loss": 0.8757,
+      "step": 507
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.013380281690141e-05,
+      "loss": 1.0178,
+      "step": 508
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.0112676056338027e-05,
+      "loss": 0.8362,
+      "step": 509
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.009154929577465e-05,
+      "loss": 0.9626,
+      "step": 510
+    },
+    {
+      "epoch": 1.4,
+      "eval_loss": 1.21484375,
+      "eval_runtime": 1.4386,
+      "eval_samples_per_second": 41.706,
+      "eval_steps_per_second": 10.426,
+      "step": 510
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.0070422535211266e-05,
+      "loss": 0.9087,
+      "step": 511
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.0049295774647888e-05,
+      "loss": 0.8828,
+      "step": 512
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.002816901408451e-05,
+      "loss": 0.9663,
+      "step": 513
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.0007042253521127e-05,
+      "loss": 0.9438,
+      "step": 514
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.9985915492957748e-05,
+      "loss": 0.8258,
+      "step": 515
+    },
+    {
+      "epoch": 1.41,
+      "eval_loss": 1.217187523841858,
+      "eval_runtime": 1.4394,
+      "eval_samples_per_second": 41.684,
+      "eval_steps_per_second": 10.421,
+      "step": 515
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.9964788732394366e-05,
+      "loss": 0.8794,
+      "step": 516
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.9943661971830987e-05,
+      "loss": 0.8335,
+      "step": 517
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.9922535211267605e-05,
+      "loss": 0.8965,
+      "step": 518
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.9901408450704226e-05,
+      "loss": 0.8911,
+      "step": 519
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.9880281690140844e-05,
+      "loss": 0.9067,
+      "step": 520
+    },
+    {
+      "epoch": 1.42,
+      "eval_loss": 1.2129557132720947,
+      "eval_runtime": 1.4398,
+      "eval_samples_per_second": 41.672,
+      "eval_steps_per_second": 10.418,
+      "step": 520
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9859154929577465e-05,
+      "loss": 0.9497,
+      "step": 521
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9838028169014087e-05,
+      "loss": 0.9263,
+      "step": 522
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9816901408450705e-05,
+      "loss": 0.9077,
+      "step": 523
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9795774647887326e-05,
+      "loss": 0.8921,
+      "step": 524
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.9774647887323944e-05,
+      "loss": 0.8765,
+      "step": 525
+    },
+    {
+      "epoch": 1.44,
+      "eval_loss": 1.2154947519302368,
+      "eval_runtime": 1.4306,
+      "eval_samples_per_second": 41.942,
+      "eval_steps_per_second": 10.485,
+      "step": 525
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.9753521126760565e-05,
+      "loss": 0.8879,
+      "step": 526
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.9732394366197183e-05,
+      "loss": 0.8462,
+      "step": 527
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9711267605633804e-05,
+      "loss": 0.9807,
+      "step": 528
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9690140845070422e-05,
+      "loss": 0.8496,
+      "step": 529
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9669014084507043e-05,
+      "loss": 0.8538,
+      "step": 530
+    },
+    {
+      "epoch": 1.45,
+      "eval_loss": 1.2236979007720947,
+      "eval_runtime": 1.4302,
+      "eval_samples_per_second": 41.954,
+      "eval_steps_per_second": 10.488,
+      "step": 530
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9647887323943664e-05,
+      "loss": 0.936,
+      "step": 531
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9626760563380282e-05,
+      "loss": 0.9097,
+      "step": 532
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9605633802816904e-05,
+      "loss": 0.8235,
+      "step": 533
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.958450704225352e-05,
+      "loss": 0.9517,
+      "step": 534
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9563380281690143e-05,
+      "loss": 0.9333,
+      "step": 535
+    },
+    {
+      "epoch": 1.46,
+      "eval_loss": 1.2077473402023315,
+      "eval_runtime": 1.4338,
+      "eval_samples_per_second": 41.848,
+      "eval_steps_per_second": 10.462,
+      "step": 535
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.954225352112676e-05,
+      "loss": 0.8149,
+      "step": 536
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.9521126760563382e-05,
+      "loss": 0.8643,
+      "step": 537
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.95e-05,
+      "loss": 0.917,
+      "step": 538
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.947887323943662e-05,
+      "loss": 0.9011,
+      "step": 539
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.9457746478873242e-05,
+      "loss": 0.8516,
+      "step": 540
+    },
+    {
+      "epoch": 1.48,
+      "eval_loss": 1.2184895277023315,
+      "eval_runtime": 1.4391,
+      "eval_samples_per_second": 41.692,
+      "eval_steps_per_second": 10.423,
+      "step": 540
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.943661971830986e-05,
+      "loss": 0.8347,
+      "step": 541
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.941549295774648e-05,
+      "loss": 0.8069,
+      "step": 542
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.93943661971831e-05,
+      "loss": 0.9397,
+      "step": 543
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.937323943661972e-05,
+      "loss": 0.8274,
+      "step": 544
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.9352112676056338e-05,
+      "loss": 0.8267,
+      "step": 545
+    },
+    {
+      "epoch": 1.49,
+      "eval_loss": 1.219140648841858,
+      "eval_runtime": 1.4449,
+      "eval_samples_per_second": 41.527,
+      "eval_steps_per_second": 10.382,
+      "step": 545
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.933098591549296e-05,
+      "loss": 0.9163,
+      "step": 546
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.9309859154929577e-05,
+      "loss": 0.8577,
+      "step": 547
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.92887323943662e-05,
+      "loss": 0.822,
+      "step": 548
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.926760563380282e-05,
+      "loss": 0.9133,
+      "step": 549
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.9246478873239438e-05,
+      "loss": 0.896,
+      "step": 550
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 1.2112630605697632,
+      "eval_runtime": 1.4339,
+      "eval_samples_per_second": 41.844,
+      "eval_steps_per_second": 10.461,
+      "step": 550
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.922535211267606e-05,
+      "loss": 0.9255,
+      "step": 551
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.9204225352112677e-05,
+      "loss": 0.8713,
+      "step": 552
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.9183098591549298e-05,
+      "loss": 0.8396,
+      "step": 553
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.9161971830985916e-05,
+      "loss": 0.9153,
+      "step": 554
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.9140845070422537e-05,
+      "loss": 0.8562,
+      "step": 555
+    },
+    {
+      "epoch": 1.52,
+      "eval_loss": 1.2151693105697632,
+      "eval_runtime": 1.4301,
+      "eval_samples_per_second": 41.954,
+      "eval_steps_per_second": 10.489,
+      "step": 555
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.9119718309859155e-05,
+      "loss": 0.8232,
+      "step": 556
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.9098591549295776e-05,
+      "loss": 0.9653,
+      "step": 557
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.9077464788732398e-05,
+      "loss": 0.8181,
+      "step": 558
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.9056338028169016e-05,
+      "loss": 0.9258,
+      "step": 559
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.9035211267605637e-05,
+      "loss": 0.8547,
+      "step": 560
+    },
+    {
+      "epoch": 1.53,
+      "eval_loss": 1.221289038658142,
+      "eval_runtime": 1.4411,
+      "eval_samples_per_second": 41.634,
+      "eval_steps_per_second": 10.408,
+      "step": 560
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.9014084507042255e-05,
+      "loss": 0.8914,
+      "step": 561
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.8992957746478876e-05,
+      "loss": 0.8821,
+      "step": 562
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.8971830985915494e-05,
+      "loss": 0.8115,
+      "step": 563
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.8950704225352115e-05,
+      "loss": 0.9021,
+      "step": 564
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.8929577464788733e-05,
+      "loss": 0.8772,
+      "step": 565
+    },
+    {
+      "epoch": 1.55,
+      "eval_loss": 1.2206380367279053,
+      "eval_runtime": 1.4377,
+      "eval_samples_per_second": 41.733,
+      "eval_steps_per_second": 10.433,
+      "step": 565
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.8908450704225354e-05,
+      "loss": 0.8928,
+      "step": 566
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.8887323943661975e-05,
+      "loss": 0.8528,
+      "step": 567
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.8866197183098593e-05,
+      "loss": 0.814,
+      "step": 568
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.8845070422535215e-05,
+      "loss": 0.8362,
+      "step": 569
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.8823943661971832e-05,
+      "loss": 0.8918,
+      "step": 570
+    },
+    {
+      "epoch": 1.56,
+      "eval_loss": 1.2133463621139526,
+      "eval_runtime": 1.4294,
+      "eval_samples_per_second": 41.976,
+      "eval_steps_per_second": 10.494,
+      "step": 570
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.8802816901408454e-05,
+      "loss": 0.864,
+      "step": 571
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.878169014084507e-05,
+      "loss": 0.8894,
+      "step": 572
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.8760563380281693e-05,
+      "loss": 0.8896,
+      "step": 573
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.8739436619718307e-05,
+      "loss": 0.9551,
+      "step": 574
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.871830985915493e-05,
+      "loss": 0.8472,
+      "step": 575
+    },
+    {
+      "epoch": 1.57,
+      "eval_loss": 1.2088541984558105,
+      "eval_runtime": 1.4337,
+      "eval_samples_per_second": 41.851,
+      "eval_steps_per_second": 10.463,
+      "step": 575
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.869718309859155e-05,
+      "loss": 0.8984,
+      "step": 576
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.8676056338028168e-05,
+      "loss": 0.9644,
+      "step": 577
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.865492957746479e-05,
+      "loss": 0.9417,
+      "step": 578
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.8633802816901407e-05,
+      "loss": 0.842,
+      "step": 579
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.8612676056338028e-05,
+      "loss": 0.8782,
+      "step": 580
+    },
+    {
+      "epoch": 1.59,
+      "eval_loss": 1.2000651359558105,
+      "eval_runtime": 1.4492,
+      "eval_samples_per_second": 41.403,
+      "eval_steps_per_second": 10.351,
+      "step": 580
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.8591549295774646e-05,
+      "loss": 0.8918,
+      "step": 581
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.8570422535211267e-05,
+      "loss": 0.8528,
+      "step": 582
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.8549295774647885e-05,
+      "loss": 0.8352,
+      "step": 583
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.8528169014084506e-05,
+      "loss": 0.9197,
+      "step": 584
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.8507042253521128e-05,
+      "loss": 0.8721,
+      "step": 585
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 1.211328148841858,
+      "eval_runtime": 1.4496,
+      "eval_samples_per_second": 41.392,
+      "eval_steps_per_second": 10.348,
+      "step": 585
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.8485915492957745e-05,
+      "loss": 0.8601,
+      "step": 586
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.8464788732394367e-05,
+      "loss": 0.8799,
+      "step": 587
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.8443661971830985e-05,
+      "loss": 0.9102,
+      "step": 588
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.8422535211267606e-05,
+      "loss": 0.7679,
+      "step": 589
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8401408450704224e-05,
+      "loss": 0.8848,
+      "step": 590
+    },
+    {
+      "epoch": 1.62,
+      "eval_loss": 1.2179036140441895,
+      "eval_runtime": 1.4338,
+      "eval_samples_per_second": 41.848,
+      "eval_steps_per_second": 10.462,
+      "step": 590
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8380281690140845e-05,
+      "loss": 0.8168,
+      "step": 591
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8359154929577463e-05,
+      "loss": 0.9133,
+      "step": 592
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8338028169014084e-05,
+      "loss": 0.8425,
+      "step": 593
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.8316901408450702e-05,
+      "loss": 0.8164,
+      "step": 594
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.8295774647887323e-05,
+      "loss": 0.8441,
+      "step": 595
+    },
+    {
+      "epoch": 1.63,
+      "eval_loss": 1.2106770277023315,
+      "eval_runtime": 1.4332,
+      "eval_samples_per_second": 41.865,
+      "eval_steps_per_second": 10.466,
+      "step": 595
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.8274647887323944e-05,
+      "loss": 0.8716,
+      "step": 596
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.8253521126760562e-05,
+      "loss": 0.9121,
+      "step": 597
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.8232394366197184e-05,
+      "loss": 0.8623,
+      "step": 598
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.82112676056338e-05,
+      "loss": 0.9595,
+      "step": 599
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.8190140845070423e-05,
+      "loss": 0.9221,
+      "step": 600
+    },
+    {
+      "epoch": 1.64,
+      "eval_loss": 1.2062499523162842,
+      "eval_runtime": 1.4468,
+      "eval_samples_per_second": 41.472,
+      "eval_steps_per_second": 10.368,
+      "step": 600
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.816901408450704e-05,
+      "loss": 0.79,
+      "step": 601
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.8147887323943662e-05,
+      "loss": 0.8618,
+      "step": 602
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.812676056338028e-05,
+      "loss": 0.9082,
+      "step": 603
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.81056338028169e-05,
+      "loss": 0.832,
+      "step": 604
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.8084507042253522e-05,
+      "loss": 0.9185,
+      "step": 605
+    },
+    {
+      "epoch": 1.66,
+      "eval_loss": 1.204687476158142,
+      "eval_runtime": 1.4366,
+      "eval_samples_per_second": 41.767,
+      "eval_steps_per_second": 10.442,
+      "step": 605
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.806338028169014e-05,
+      "loss": 0.9214,
+      "step": 606
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.804225352112676e-05,
+      "loss": 0.9387,
+      "step": 607
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.802112676056338e-05,
+      "loss": 0.9233,
+      "step": 608
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.8e-05,
+      "loss": 0.887,
+      "step": 609
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.797887323943662e-05,
+      "loss": 0.8347,
+      "step": 610
+    },
+    {
+      "epoch": 1.67,
+      "eval_loss": 1.2030599117279053,
+      "eval_runtime": 1.4455,
+      "eval_samples_per_second": 41.507,
+      "eval_steps_per_second": 10.377,
+      "step": 610
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.795774647887324e-05,
+      "loss": 0.8621,
+      "step": 611
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.7936619718309857e-05,
+      "loss": 0.8155,
+      "step": 612
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.791549295774648e-05,
+      "loss": 0.8303,
+      "step": 613
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.78943661971831e-05,
+      "loss": 0.8142,
+      "step": 614
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.7873239436619718e-05,
+      "loss": 0.8111,
+      "step": 615
+    },
+    {
+      "epoch": 1.68,
+      "eval_loss": 1.2149739265441895,
+      "eval_runtime": 1.4409,
+      "eval_samples_per_second": 41.641,
+      "eval_steps_per_second": 10.41,
+      "step": 615
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.785211267605634e-05,
+      "loss": 1.0157,
+      "step": 616
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.7830985915492957e-05,
+      "loss": 0.9094,
+      "step": 617
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.7809859154929578e-05,
+      "loss": 0.8875,
+      "step": 618
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.7788732394366196e-05,
+      "loss": 0.9014,
+      "step": 619
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.7767605633802817e-05,
+      "loss": 0.8735,
+      "step": 620
+    },
+    {
+      "epoch": 1.7,
+      "eval_loss": 1.1962239742279053,
+      "eval_runtime": 1.4303,
+      "eval_samples_per_second": 41.95,
+      "eval_steps_per_second": 10.488,
+      "step": 620
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.7746478873239435e-05,
+      "loss": 0.9038,
+      "step": 621
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.7725352112676056e-05,
+      "loss": 0.8218,
+      "step": 622
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.7704225352112678e-05,
+      "loss": 0.8892,
+      "step": 623
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.7683098591549296e-05,
+      "loss": 0.9238,
+      "step": 624
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.7661971830985917e-05,
+      "loss": 0.8682,
+      "step": 625
+    },
+    {
+      "epoch": 1.71,
+      "eval_loss": 1.208984375,
+      "eval_runtime": 1.431,
+      "eval_samples_per_second": 41.928,
+      "eval_steps_per_second": 10.482,
+      "step": 625
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.7640845070422535e-05,
+      "loss": 0.8586,
+      "step": 626
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.7619718309859156e-05,
+      "loss": 0.885,
+      "step": 627
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.7598591549295774e-05,
+      "loss": 0.8191,
+      "step": 628
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.7577464788732395e-05,
+      "loss": 0.8259,
+      "step": 629
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.7556338028169013e-05,
+      "loss": 0.7971,
+      "step": 630
+    },
+    {
+      "epoch": 1.72,
+      "eval_loss": 1.2097004652023315,
+      "eval_runtime": 1.4333,
+      "eval_samples_per_second": 41.86,
+      "eval_steps_per_second": 10.465,
+      "step": 630
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.7535211267605634e-05,
+      "loss": 0.9509,
+      "step": 631
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.7514084507042255e-05,
+      "loss": 0.9155,
+      "step": 632
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.7492957746478873e-05,
+      "loss": 0.8767,
+      "step": 633
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.7471830985915495e-05,
+      "loss": 0.8826,
+      "step": 634
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.7450704225352112e-05,
+      "loss": 0.9143,
+      "step": 635
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 1.1940103769302368,
+      "eval_runtime": 1.4387,
+      "eval_samples_per_second": 41.704,
+      "eval_steps_per_second": 10.426,
+      "step": 635
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.7429577464788734e-05,
+      "loss": 0.8936,
+      "step": 636
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.740845070422535e-05,
+      "loss": 0.8992,
+      "step": 637
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.7387323943661973e-05,
+      "loss": 0.9531,
+      "step": 638
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.736619718309859e-05,
+      "loss": 0.8372,
+      "step": 639
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.7345070422535212e-05,
+      "loss": 0.8633,
+      "step": 640
+    },
+    {
+      "epoch": 1.75,
+      "eval_loss": 1.2049479484558105,
+      "eval_runtime": 1.4278,
+      "eval_samples_per_second": 42.022,
+      "eval_steps_per_second": 10.506,
+      "step": 640
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.7323943661971833e-05,
+      "loss": 0.8406,
+      "step": 641
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.730281690140845e-05,
+      "loss": 0.9092,
+      "step": 642
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.7281690140845072e-05,
+      "loss": 0.865,
+      "step": 643
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.726056338028169e-05,
+      "loss": 0.8376,
+      "step": 644
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.723943661971831e-05,
+      "loss": 0.8643,
+      "step": 645
+    },
+    {
+      "epoch": 1.77,
+      "eval_loss": 1.201171875,
+      "eval_runtime": 1.4457,
+      "eval_samples_per_second": 41.503,
+      "eval_steps_per_second": 10.376,
+      "step": 645
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.721830985915493e-05,
+      "loss": 0.8903,
+      "step": 646
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.719718309859155e-05,
+      "loss": 0.9695,
+      "step": 647
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.717605633802817e-05,
+      "loss": 0.8241,
+      "step": 648
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.715492957746479e-05,
+      "loss": 0.8745,
+      "step": 649
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.713380281690141e-05,
+      "loss": 0.9146,
+      "step": 650
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 1.1956380605697632,
+      "eval_runtime": 1.4491,
+      "eval_samples_per_second": 41.405,
+      "eval_steps_per_second": 10.351,
+      "step": 650
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.711267605633803e-05,
+      "loss": 0.8906,
+      "step": 651
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.709154929577465e-05,
+      "loss": 0.9258,
+      "step": 652
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.7070422535211268e-05,
+      "loss": 0.8684,
+      "step": 653
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.704929577464789e-05,
+      "loss": 0.8652,
+      "step": 654
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.7028169014084507e-05,
+      "loss": 0.829,
+      "step": 655
+    },
+    {
+      "epoch": 1.79,
+      "eval_loss": 1.2042317390441895,
+      "eval_runtime": 1.4504,
+      "eval_samples_per_second": 41.367,
+      "eval_steps_per_second": 10.342,
+      "step": 655
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.700704225352113e-05,
+      "loss": 0.9016,
+      "step": 656
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.6985915492957746e-05,
+      "loss": 0.9138,
+      "step": 657
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.6964788732394367e-05,
+      "loss": 0.9636,
+      "step": 658
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.694366197183099e-05,
+      "loss": 0.8638,
+      "step": 659
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.6922535211267607e-05,
+      "loss": 0.9082,
+      "step": 660
+    },
+    {
+      "epoch": 1.81,
+      "eval_loss": 1.2052733898162842,
+      "eval_runtime": 1.4291,
+      "eval_samples_per_second": 41.985,
+      "eval_steps_per_second": 10.496,
+      "step": 660
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.6901408450704228e-05,
+      "loss": 0.8015,
+      "step": 661
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.6880281690140846e-05,
+      "loss": 0.8662,
+      "step": 662
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.6859154929577467e-05,
+      "loss": 0.8728,
+      "step": 663
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.6838028169014085e-05,
+      "loss": 0.9229,
+      "step": 664
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.6816901408450706e-05,
+      "loss": 0.8278,
+      "step": 665
+    },
+    {
+      "epoch": 1.82,
+      "eval_loss": 1.2009114027023315,
+      "eval_runtime": 1.4393,
+      "eval_samples_per_second": 41.688,
+      "eval_steps_per_second": 10.422,
+      "step": 665
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.6795774647887324e-05,
+      "loss": 0.8169,
+      "step": 666
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.6774647887323945e-05,
+      "loss": 0.863,
+      "step": 667
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.6753521126760563e-05,
+      "loss": 0.9324,
+      "step": 668
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.6732394366197184e-05,
+      "loss": 0.9136,
+      "step": 669
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.6711267605633806e-05,
+      "loss": 0.9482,
+      "step": 670
+    },
+    {
+      "epoch": 1.83,
+      "eval_loss": 1.2078125476837158,
+      "eval_runtime": 1.4325,
+      "eval_samples_per_second": 41.884,
+      "eval_steps_per_second": 10.471,
+      "step": 670
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.6690140845070424e-05,
+      "loss": 0.7996,
+      "step": 671
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.6669014084507045e-05,
+      "loss": 0.8726,
+      "step": 672
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.6647887323943663e-05,
+      "loss": 0.9144,
+      "step": 673
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6626760563380284e-05,
+      "loss": 0.855,
+      "step": 674
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6605633802816902e-05,
+      "loss": 0.8313,
+      "step": 675
+    },
+    {
+      "epoch": 1.85,
+      "eval_loss": 1.1989582777023315,
+      "eval_runtime": 1.431,
+      "eval_samples_per_second": 41.927,
+      "eval_steps_per_second": 10.482,
+      "step": 675
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6584507042253523e-05,
+      "loss": 0.8845,
+      "step": 676
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.656338028169014e-05,
+      "loss": 0.9253,
+      "step": 677
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6542253521126762e-05,
+      "loss": 0.8608,
+      "step": 678
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6521126760563383e-05,
+      "loss": 0.9197,
+      "step": 679
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.65e-05,
+      "loss": 0.8721,
+      "step": 680
+    },
+    {
+      "epoch": 1.86,
+      "eval_loss": 1.2052083015441895,
+      "eval_runtime": 1.4396,
+      "eval_samples_per_second": 41.678,
+      "eval_steps_per_second": 10.419,
+      "step": 680
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6478873239436623e-05,
+      "loss": 0.8586,
+      "step": 681
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.645774647887324e-05,
+      "loss": 0.9263,
+      "step": 682
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.643661971830986e-05,
+      "loss": 0.8904,
+      "step": 683
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.641549295774648e-05,
+      "loss": 0.9122,
+      "step": 684
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.63943661971831e-05,
+      "loss": 0.8408,
+      "step": 685
+    },
+    {
+      "epoch": 1.88,
+      "eval_loss": 1.2055989503860474,
+      "eval_runtime": 1.4421,
+      "eval_samples_per_second": 41.605,
+      "eval_steps_per_second": 10.401,
+      "step": 685
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.637323943661972e-05,
+      "loss": 0.8971,
+      "step": 686
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.635211267605634e-05,
+      "loss": 0.8931,
+      "step": 687
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.633098591549296e-05,
+      "loss": 0.8677,
+      "step": 688
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.630985915492958e-05,
+      "loss": 0.9255,
+      "step": 689
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.62887323943662e-05,
+      "loss": 0.8643,
+      "step": 690
+    },
+    {
+      "epoch": 1.89,
+      "eval_loss": 1.200781226158142,
+      "eval_runtime": 1.4405,
+      "eval_samples_per_second": 41.651,
+      "eval_steps_per_second": 10.413,
+      "step": 690
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.6267605633802818e-05,
+      "loss": 0.9697,
+      "step": 691
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.6246478873239436e-05,
+      "loss": 0.8462,
+      "step": 692
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.6225352112676054e-05,
+      "loss": 0.8672,
+      "step": 693
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.6204225352112675e-05,
+      "loss": 0.8135,
+      "step": 694
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.6183098591549293e-05,
+      "loss": 0.8759,
+      "step": 695
+    },
+    {
+      "epoch": 1.9,
+      "eval_loss": 1.2013020515441895,
+      "eval_runtime": 1.4298,
+      "eval_samples_per_second": 41.965,
+      "eval_steps_per_second": 10.491,
+      "step": 695
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.6161971830985914e-05,
+      "loss": 0.8938,
+      "step": 696
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.6140845070422536e-05,
+      "loss": 0.8403,
+      "step": 697
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.6119718309859153e-05,
+      "loss": 0.8376,
+      "step": 698
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.6098591549295775e-05,
+      "loss": 0.8423,
+      "step": 699
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.6077464788732393e-05,
+      "loss": 0.8359,
+      "step": 700
+    },
+    {
+      "epoch": 1.92,
+      "eval_loss": 1.2037760019302368,
+      "eval_runtime": 1.4313,
+      "eval_samples_per_second": 41.92,
+      "eval_steps_per_second": 10.48,
+      "step": 700
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.6056338028169014e-05,
+      "loss": 0.9102,
+      "step": 701
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.603521126760563e-05,
+      "loss": 0.8345,
+      "step": 702
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.6014084507042253e-05,
+      "loss": 0.9136,
+      "step": 703
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.599295774647887e-05,
+      "loss": 0.824,
+      "step": 704
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.5971830985915492e-05,
+      "loss": 0.8311,
+      "step": 705
+    },
+    {
+      "epoch": 1.93,
+      "eval_loss": 1.2041015625,
+      "eval_runtime": 1.4474,
+      "eval_samples_per_second": 41.454,
+      "eval_steps_per_second": 10.363,
+      "step": 705
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.5950704225352113e-05,
+      "loss": 0.8589,
+      "step": 706
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.592957746478873e-05,
+      "loss": 0.9204,
+      "step": 707
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.5908450704225352e-05,
+      "loss": 0.8625,
+      "step": 708
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.588732394366197e-05,
+      "loss": 0.9077,
+      "step": 709
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.586619718309859e-05,
+      "loss": 0.8379,
+      "step": 710
+    },
+    {
+      "epoch": 1.94,
+      "eval_loss": 1.194726586341858,
+      "eval_runtime": 1.4327,
+      "eval_samples_per_second": 41.879,
+      "eval_steps_per_second": 10.47,
+      "step": 710
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.584507042253521e-05,
+      "loss": 0.8887,
+      "step": 711
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.582394366197183e-05,
+      "loss": 0.8384,
+      "step": 712
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.580281690140845e-05,
+      "loss": 0.8286,
+      "step": 713
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.578169014084507e-05,
+      "loss": 0.9221,
+      "step": 714
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.576056338028169e-05,
+      "loss": 0.9094,
+      "step": 715
+    },
+    {
+      "epoch": 1.96,
+      "eval_loss": 1.1968098878860474,
+      "eval_runtime": 1.4313,
+      "eval_samples_per_second": 41.921,
+      "eval_steps_per_second": 10.48,
+      "step": 715
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.573943661971831e-05,
+      "loss": 0.8813,
+      "step": 716
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.571830985915493e-05,
+      "loss": 0.8218,
+      "step": 717
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.5697183098591548e-05,
+      "loss": 0.8811,
+      "step": 718
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.567605633802817e-05,
+      "loss": 0.9333,
+      "step": 719
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.5654929577464787e-05,
+      "loss": 0.8589,
+      "step": 720
+    },
+    {
+      "epoch": 1.97,
+      "eval_loss": 1.1916015148162842,
+      "eval_runtime": 1.437,
+      "eval_samples_per_second": 41.753,
+      "eval_steps_per_second": 10.438,
+      "step": 720
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.563380281690141e-05,
+      "loss": 0.8844,
+      "step": 721
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.5612676056338026e-05,
+      "loss": 0.8267,
+      "step": 722
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.5591549295774648e-05,
+      "loss": 0.8459,
+      "step": 723
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.557042253521127e-05,
+      "loss": 0.8961,
+      "step": 724
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.5549295774647887e-05,
+      "loss": 0.8535,
+      "step": 725
+    },
+    {
+      "epoch": 1.98,
+      "eval_loss": 1.1966146230697632,
+      "eval_runtime": 1.4269,
+      "eval_samples_per_second": 42.048,
+      "eval_steps_per_second": 10.512,
+      "step": 725
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.5528169014084508e-05,
+      "loss": 0.8774,
+      "step": 726
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.5507042253521126e-05,
+      "loss": 0.8381,
+      "step": 727
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.5485915492957747e-05,
+      "loss": 0.873,
+      "step": 728
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.5464788732394365e-05,
+      "loss": 0.8857,
+      "step": 729
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.5443661971830986e-05,
+      "loss": 0.9001,
+      "step": 730
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 1.191796898841858,
+      "eval_runtime": 1.4322,
+      "eval_samples_per_second": 41.893,
+      "eval_steps_per_second": 10.473,
+      "step": 730
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.5422535211267604e-05,
+      "loss": 0.7343,
+      "step": 731
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.5401408450704225e-05,
+      "loss": 0.5074,
+      "step": 732
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.5380281690140847e-05,
+      "loss": 0.5366,
+      "step": 733
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.5359154929577464e-05,
+      "loss": 0.6183,
+      "step": 734
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.5338028169014086e-05,
+      "loss": 0.5245,
+      "step": 735
+    },
+    {
+      "epoch": 2.01,
+      "eval_loss": 1.4343098402023315,
+      "eval_runtime": 1.4356,
+      "eval_samples_per_second": 41.795,
+      "eval_steps_per_second": 10.449,
+      "step": 735
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.5316901408450704e-05,
+      "loss": 0.5635,
+      "step": 736
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.5295774647887325e-05,
+      "loss": 0.4658,
+      "step": 737
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.5274647887323943e-05,
+      "loss": 0.5979,
+      "step": 738
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.5253521126760564e-05,
+      "loss": 0.5694,
+      "step": 739
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.5232394366197183e-05,
+      "loss": 0.5286,
+      "step": 740
+    },
+    {
+      "epoch": 2.03,
+      "eval_loss": 1.2919921875,
+      "eval_runtime": 1.4508,
+      "eval_samples_per_second": 41.357,
+      "eval_steps_per_second": 10.339,
+      "step": 740
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.5211267605633803e-05,
+      "loss": 0.5486,
+      "step": 741
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.5190140845070423e-05,
+      "loss": 0.5698,
+      "step": 742
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.5169014084507042e-05,
+      "loss": 0.4885,
+      "step": 743
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.5147887323943662e-05,
+      "loss": 0.5508,
+      "step": 744
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.5126760563380281e-05,
+      "loss": 0.5443,
+      "step": 745
+    },
+    {
+      "epoch": 2.04,
+      "eval_loss": 1.3705729246139526,
+      "eval_runtime": 1.4336,
+      "eval_samples_per_second": 41.854,
+      "eval_steps_per_second": 10.463,
+      "step": 745
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.5105633802816901e-05,
+      "loss": 0.5132,
+      "step": 746
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.5084507042253522e-05,
+      "loss": 0.5052,
+      "step": 747
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.5063380281690142e-05,
+      "loss": 0.537,
+      "step": 748
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.5042253521126761e-05,
+      "loss": 0.5079,
+      "step": 749
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.502112676056338e-05,
+      "loss": 0.5879,
+      "step": 750
+    },
+    {
+      "epoch": 2.05,
+      "eval_loss": 1.3911458253860474,
+      "eval_runtime": 1.4379,
+      "eval_samples_per_second": 41.727,
+      "eval_steps_per_second": 10.432,
+      "step": 750
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.5e-05,
+      "loss": 0.5283,
+      "step": 751
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.497887323943662e-05,
+      "loss": 0.5541,
+      "step": 752
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.495774647887324e-05,
+      "loss": 0.5309,
+      "step": 753
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.4936619718309859e-05,
+      "loss": 0.5353,
+      "step": 754
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.4915492957746479e-05,
+      "loss": 0.5544,
+      "step": 755
+    },
+    {
+      "epoch": 2.07,
+      "eval_loss": 1.3301432132720947,
+      "eval_runtime": 1.4319,
+      "eval_samples_per_second": 41.901,
+      "eval_steps_per_second": 10.475,
+      "step": 755
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.48943661971831e-05,
+      "loss": 0.4685,
+      "step": 756
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.487323943661972e-05,
+      "loss": 0.5537,
+      "step": 757
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.4852112676056339e-05,
+      "loss": 0.5389,
+      "step": 758
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.4830985915492959e-05,
+      "loss": 0.56,
+      "step": 759
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.4809859154929578e-05,
+      "loss": 0.4534,
+      "step": 760
+    },
+    {
+      "epoch": 2.08,
+      "eval_loss": 1.4022786617279053,
+      "eval_runtime": 1.4527,
+      "eval_samples_per_second": 41.302,
+      "eval_steps_per_second": 10.325,
+      "step": 760
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.4788732394366198e-05,
+      "loss": 0.6017,
+      "step": 761
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.4767605633802817e-05,
+      "loss": 0.5728,
+      "step": 762
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.4746478873239437e-05,
+      "loss": 0.5605,
+      "step": 763
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.4725352112676056e-05,
+      "loss": 0.4794,
+      "step": 764
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.4704225352112678e-05,
+      "loss": 0.4332,
+      "step": 765
+    },
+    {
+      "epoch": 2.09,
+      "eval_loss": 1.3668619394302368,
+      "eval_runtime": 1.4358,
+      "eval_samples_per_second": 41.788,
+      "eval_steps_per_second": 10.447,
+      "step": 765
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.4683098591549297e-05,
+      "loss": 0.5717,
+      "step": 766
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.4661971830985917e-05,
+      "loss": 0.5641,
+      "step": 767
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.4640845070422536e-05,
+      "loss": 0.5409,
+      "step": 768
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.4619718309859156e-05,
+      "loss": 0.5383,
+      "step": 769
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.4598591549295775e-05,
+      "loss": 0.6,
+      "step": 770
+    },
+    {
+      "epoch": 2.11,
+      "eval_loss": 1.377343773841858,
+      "eval_runtime": 1.4361,
+      "eval_samples_per_second": 41.781,
+      "eval_steps_per_second": 10.445,
+      "step": 770
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.4577464788732395e-05,
+      "loss": 0.5009,
+      "step": 771
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.4556338028169015e-05,
+      "loss": 0.4661,
+      "step": 772
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.4535211267605634e-05,
+      "loss": 0.4633,
+      "step": 773
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.4514084507042254e-05,
+      "loss": 0.4951,
+      "step": 774
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.4492957746478875e-05,
+      "loss": 0.4669,
+      "step": 775
+    },
+    {
+      "epoch": 2.12,
+      "eval_loss": 1.4046224355697632,
+      "eval_runtime": 1.4407,
+      "eval_samples_per_second": 41.647,
+      "eval_steps_per_second": 10.412,
+      "step": 775
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.4471830985915495e-05,
+      "loss": 0.5579,
+      "step": 776
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.4450704225352114e-05,
+      "loss": 0.5409,
+      "step": 777
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.4429577464788734e-05,
+      "loss": 0.5773,
+      "step": 778
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.4408450704225353e-05,
+      "loss": 0.5281,
+      "step": 779
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.4387323943661973e-05,
+      "loss": 0.5716,
+      "step": 780
+    },
+    {
+      "epoch": 2.14,
+      "eval_loss": 1.3707681894302368,
+      "eval_runtime": 1.4435,
+      "eval_samples_per_second": 41.567,
+      "eval_steps_per_second": 10.392,
+      "step": 780
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.436619718309859e-05,
+      "loss": 0.4774,
+      "step": 781
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.434507042253521e-05,
+      "loss": 0.5137,
+      "step": 782
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.432394366197183e-05,
+      "loss": 0.53,
+      "step": 783
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.4302816901408451e-05,
+      "loss": 0.4743,
+      "step": 784
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.428169014084507e-05,
+      "loss": 0.4795,
+      "step": 785
+    },
+    {
+      "epoch": 2.15,
+      "eval_loss": 1.3768880367279053,
+      "eval_runtime": 1.4358,
+      "eval_samples_per_second": 41.789,
+      "eval_steps_per_second": 10.447,
+      "step": 785
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.426056338028169e-05,
+      "loss": 0.5816,
+      "step": 786
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.423943661971831e-05,
+      "loss": 0.5929,
+      "step": 787
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.421830985915493e-05,
+      "loss": 0.5054,
+      "step": 788
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.4197183098591549e-05,
+      "loss": 0.5183,
+      "step": 789
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.4176056338028168e-05,
+      "loss": 0.5703,
+      "step": 790
+    },
+    {
+      "epoch": 2.16,
+      "eval_loss": 1.391210913658142,
+      "eval_runtime": 1.4338,
+      "eval_samples_per_second": 41.847,
+      "eval_steps_per_second": 10.462,
+      "step": 790
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.4154929577464788e-05,
+      "loss": 0.4634,
+      "step": 791
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.4133802816901408e-05,
+      "loss": 0.4516,
+      "step": 792
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.4112676056338029e-05,
+      "loss": 0.5274,
+      "step": 793
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.4091549295774648e-05,
+      "loss": 0.4973,
+      "step": 794
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.4070422535211268e-05,
+      "loss": 0.5012,
+      "step": 795
+    },
+    {
+      "epoch": 2.18,
+      "eval_loss": 1.3770833015441895,
+      "eval_runtime": 1.4334,
+      "eval_samples_per_second": 41.857,
+      "eval_steps_per_second": 10.464,
+      "step": 795
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.4049295774647887e-05,
+      "loss": 0.5264,
+      "step": 796
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.4028169014084507e-05,
+      "loss": 0.5173,
+      "step": 797
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.4007042253521127e-05,
+      "loss": 0.5386,
+      "step": 798
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.3985915492957746e-05,
+      "loss": 0.4919,
+      "step": 799
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.3964788732394366e-05,
+      "loss": 0.4817,
+      "step": 800
+    },
+    {
+      "epoch": 2.19,
+      "eval_loss": 1.3916666507720947,
+      "eval_runtime": 1.446,
+      "eval_samples_per_second": 41.495,
+      "eval_steps_per_second": 10.374,
+      "step": 800
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.3943661971830985e-05,
+      "loss": 0.5125,
+      "step": 801
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.3922535211267607e-05,
+      "loss": 0.5457,
+      "step": 802
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.3901408450704226e-05,
+      "loss": 0.5438,
+      "step": 803
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.3880281690140846e-05,
+      "loss": 0.5513,
+      "step": 804
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.3859154929577465e-05,
+      "loss": 0.5916,
+      "step": 805
+    },
+    {
+      "epoch": 2.2,
+      "eval_loss": 1.3783854246139526,
+      "eval_runtime": 1.4332,
+      "eval_samples_per_second": 41.863,
+      "eval_steps_per_second": 10.466,
+      "step": 805
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.3838028169014085e-05,
+      "loss": 0.5011,
+      "step": 806
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.3816901408450704e-05,
+      "loss": 0.5632,
+      "step": 807
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.3795774647887324e-05,
+      "loss": 0.4752,
+      "step": 808
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.3774647887323943e-05,
+      "loss": 0.5488,
+      "step": 809
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.3753521126760563e-05,
+      "loss": 0.5339,
+      "step": 810
+    },
+    {
+      "epoch": 2.22,
+      "eval_loss": 1.3803385496139526,
+      "eval_runtime": 1.4366,
+      "eval_samples_per_second": 41.764,
+      "eval_steps_per_second": 10.441,
+      "step": 810
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.3732394366197184e-05,
+      "loss": 0.4406,
+      "step": 811
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.3711267605633804e-05,
+      "loss": 0.5819,
+      "step": 812
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.3690140845070423e-05,
+      "loss": 0.5516,
+      "step": 813
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.3669014084507043e-05,
+      "loss": 0.5632,
+      "step": 814
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.3647887323943663e-05,
+      "loss": 0.4542,
+      "step": 815
+    },
+    {
+      "epoch": 2.23,
+      "eval_loss": 1.3854166269302368,
+      "eval_runtime": 1.4315,
+      "eval_samples_per_second": 41.914,
+      "eval_steps_per_second": 10.479,
+      "step": 815
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.3626760563380282e-05,
+      "loss": 0.4274,
+      "step": 816
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.3605633802816902e-05,
+      "loss": 0.5807,
+      "step": 817
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.3584507042253521e-05,
+      "loss": 0.484,
+      "step": 818
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.356338028169014e-05,
+      "loss": 0.5043,
+      "step": 819
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.354225352112676e-05,
+      "loss": 0.5442,
+      "step": 820
+    },
+    {
+      "epoch": 2.25,
+      "eval_loss": 1.3823567628860474,
+      "eval_runtime": 1.4363,
+      "eval_samples_per_second": 41.773,
+      "eval_steps_per_second": 10.443,
+      "step": 820
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.3521126760563382e-05,
+      "loss": 0.4875,
+      "step": 821
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.3500000000000001e-05,
+      "loss": 0.4908,
+      "step": 822
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.347887323943662e-05,
+      "loss": 0.4437,
+      "step": 823
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.345774647887324e-05,
+      "loss": 0.5343,
+      "step": 824
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.343661971830986e-05,
+      "loss": 0.528,
+      "step": 825
+    },
+    {
+      "epoch": 2.26,
+      "eval_loss": 1.4011067152023315,
+      "eval_runtime": 1.4304,
+      "eval_samples_per_second": 41.946,
+      "eval_steps_per_second": 10.486,
+      "step": 825
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.341549295774648e-05,
+      "loss": 0.5262,
+      "step": 826
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.3394366197183099e-05,
+      "loss": 0.5643,
+      "step": 827
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.3373239436619719e-05,
+      "loss": 0.4787,
+      "step": 828
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.3352112676056338e-05,
+      "loss": 0.5173,
+      "step": 829
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.333098591549296e-05,
+      "loss": 0.5061,
+      "step": 830
+    },
+    {
+      "epoch": 2.27,
+      "eval_loss": 1.389062523841858,
+      "eval_runtime": 1.4293,
+      "eval_samples_per_second": 41.978,
+      "eval_steps_per_second": 10.494,
+      "step": 830
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.3309859154929579e-05,
+      "loss": 0.5092,
+      "step": 831
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.3288732394366199e-05,
+      "loss": 0.5569,
+      "step": 832
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.3267605633802818e-05,
+      "loss": 0.5636,
+      "step": 833
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.3246478873239438e-05,
+      "loss": 0.5132,
+      "step": 834
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.3225352112676057e-05,
+      "loss": 0.4971,
+      "step": 835
+    },
+    {
+      "epoch": 2.29,
+      "eval_loss": 1.37890625,
+      "eval_runtime": 1.439,
+      "eval_samples_per_second": 41.697,
+      "eval_steps_per_second": 10.424,
+      "step": 835
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.3204225352112677e-05,
+      "loss": 0.6008,
+      "step": 836
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.3183098591549296e-05,
+      "loss": 0.4877,
+      "step": 837
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.3161971830985916e-05,
+      "loss": 0.5432,
+      "step": 838
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.3140845070422537e-05,
+      "loss": 0.537,
+      "step": 839
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.3119718309859155e-05,
+      "loss": 0.496,
+      "step": 840
+    },
+    {
+      "epoch": 2.3,
+      "eval_loss": 1.4156900644302368,
+      "eval_runtime": 1.4359,
+      "eval_samples_per_second": 41.784,
+      "eval_steps_per_second": 10.446,
+      "step": 840
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.3098591549295775e-05,
+      "loss": 0.4435,
+      "step": 841
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.3077464788732394e-05,
+      "loss": 0.5853,
+      "step": 842
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.3056338028169014e-05,
+      "loss": 0.536,
+      "step": 843
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.3035211267605633e-05,
+      "loss": 0.5402,
+      "step": 844
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.3014084507042253e-05,
+      "loss": 0.4872,
+      "step": 845
+    },
+    {
+      "epoch": 2.31,
+      "eval_loss": 1.3957682847976685,
+      "eval_runtime": 1.4366,
+      "eval_samples_per_second": 41.766,
+      "eval_steps_per_second": 10.441,
+      "step": 845
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.2992957746478872e-05,
+      "loss": 0.5928,
+      "step": 846
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.2971830985915492e-05,
+      "loss": 0.5397,
+      "step": 847
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.2950704225352113e-05,
+      "loss": 0.496,
+      "step": 848
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.2929577464788733e-05,
+      "loss": 0.4941,
+      "step": 849
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.2908450704225352e-05,
+      "loss": 0.4929,
+      "step": 850
+    },
+    {
+      "epoch": 2.33,
+      "eval_loss": 1.3940755128860474,
+      "eval_runtime": 1.4501,
+      "eval_samples_per_second": 41.377,
+      "eval_steps_per_second": 10.344,
+      "step": 850
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.2887323943661972e-05,
+      "loss": 0.5119,
+      "step": 851
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.2866197183098591e-05,
+      "loss": 0.5687,
+      "step": 852
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.2845070422535211e-05,
+      "loss": 0.5491,
+      "step": 853
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.282394366197183e-05,
+      "loss": 0.5395,
+      "step": 854
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.280281690140845e-05,
+      "loss": 0.5294,
+      "step": 855
+    },
+    {
+      "epoch": 2.34,
+      "eval_loss": 1.3971353769302368,
+      "eval_runtime": 1.4329,
+      "eval_samples_per_second": 41.872,
+      "eval_steps_per_second": 10.468,
+      "step": 855
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.278169014084507e-05,
+      "loss": 0.588,
+      "step": 856
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.276056338028169e-05,
+      "loss": 0.4789,
+      "step": 857
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.273943661971831e-05,
+      "loss": 0.4979,
+      "step": 858
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.271830985915493e-05,
+      "loss": 0.4799,
+      "step": 859
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.269718309859155e-05,
+      "loss": 0.4684,
+      "step": 860
+    },
+    {
+      "epoch": 2.35,
+      "eval_loss": 1.3858072757720947,
+      "eval_runtime": 1.433,
+      "eval_samples_per_second": 41.87,
+      "eval_steps_per_second": 10.467,
+      "step": 860
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.267605633802817e-05,
+      "loss": 0.4322,
+      "step": 861
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.2654929577464789e-05,
+      "loss": 0.5402,
+      "step": 862
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.2633802816901408e-05,
+      "loss": 0.568,
+      "step": 863
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.2612676056338028e-05,
+      "loss": 0.5497,
+      "step": 864
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.2591549295774647e-05,
+      "loss": 0.5945,
+      "step": 865
+    },
+    {
+      "epoch": 2.37,
+      "eval_loss": 1.3839843273162842,
+      "eval_runtime": 1.4291,
+      "eval_samples_per_second": 41.985,
+      "eval_steps_per_second": 10.496,
+      "step": 865
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.2570422535211267e-05,
+      "loss": 0.4505,
+      "step": 866
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.2549295774647888e-05,
+      "loss": 0.5322,
+      "step": 867
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.2528169014084508e-05,
+      "loss": 0.4523,
+      "step": 868
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.2507042253521127e-05,
+      "loss": 0.5189,
+      "step": 869
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.2485915492957747e-05,
+      "loss": 0.454,
+      "step": 870
+    },
+    {
+      "epoch": 2.38,
+      "eval_loss": 1.3895833492279053,
+      "eval_runtime": 1.437,
+      "eval_samples_per_second": 41.754,
+      "eval_steps_per_second": 10.438,
+      "step": 870
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.2464788732394367e-05,
+      "loss": 0.5118,
+      "step": 871
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.2443661971830986e-05,
+      "loss": 0.4862,
+      "step": 872
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.2422535211267606e-05,
+      "loss": 0.5452,
+      "step": 873
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.2401408450704225e-05,
+      "loss": 0.6139,
+      "step": 874
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.2380281690140845e-05,
+      "loss": 0.5459,
+      "step": 875
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 1.3820312023162842,
+      "eval_runtime": 1.4337,
+      "eval_samples_per_second": 41.849,
+      "eval_steps_per_second": 10.462,
+      "step": 875
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.2359154929577466e-05,
+      "loss": 0.533,
+      "step": 876
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.2338028169014086e-05,
+      "loss": 0.5431,
+      "step": 877
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.2316901408450705e-05,
+      "loss": 0.4568,
+      "step": 878
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.2295774647887325e-05,
+      "loss": 0.5712,
+      "step": 879
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.2274647887323944e-05,
+      "loss": 0.6194,
+      "step": 880
+    },
+    {
+      "epoch": 2.41,
+      "eval_loss": 1.3489583730697632,
+      "eval_runtime": 1.4459,
+      "eval_samples_per_second": 41.498,
+      "eval_steps_per_second": 10.374,
+      "step": 880
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.2253521126760564e-05,
+      "loss": 0.4916,
+      "step": 881
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.2232394366197183e-05,
+      "loss": 0.5111,
+      "step": 882
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.2211267605633803e-05,
+      "loss": 0.4934,
+      "step": 883
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.2190140845070423e-05,
+      "loss": 0.5272,
+      "step": 884
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.2169014084507044e-05,
+      "loss": 0.5668,
+      "step": 885
+    },
+    {
+      "epoch": 2.42,
+      "eval_loss": 1.3908202648162842,
+      "eval_runtime": 1.4342,
+      "eval_samples_per_second": 41.834,
+      "eval_steps_per_second": 10.459,
+      "step": 885
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.2147887323943663e-05,
+      "loss": 0.4552,
+      "step": 886
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.2126760563380283e-05,
+      "loss": 0.4707,
+      "step": 887
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.2105633802816902e-05,
+      "loss": 0.4704,
+      "step": 888
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.2084507042253522e-05,
+      "loss": 0.4577,
+      "step": 889
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.2063380281690142e-05,
+      "loss": 0.493,
+      "step": 890
+    },
+    {
+      "epoch": 2.44,
+      "eval_loss": 1.4181640148162842,
+      "eval_runtime": 1.4371,
+      "eval_samples_per_second": 41.752,
+      "eval_steps_per_second": 10.438,
+      "step": 890
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.2042253521126761e-05,
+      "loss": 0.524,
+      "step": 891
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.202112676056338e-05,
+      "loss": 0.4525,
+      "step": 892
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.2e-05,
+      "loss": 0.557,
+      "step": 893
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.1978873239436622e-05,
+      "loss": 0.5846,
+      "step": 894
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.1957746478873241e-05,
+      "loss": 0.5938,
+      "step": 895
+    },
+    {
+      "epoch": 2.45,
+      "eval_loss": 1.3720052242279053,
+      "eval_runtime": 1.4488,
+      "eval_samples_per_second": 41.415,
+      "eval_steps_per_second": 10.354,
+      "step": 895
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.193661971830986e-05,
+      "loss": 0.5188,
+      "step": 896
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.191549295774648e-05,
+      "loss": 0.4122,
+      "step": 897
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.18943661971831e-05,
+      "loss": 0.5168,
+      "step": 898
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.1873239436619718e-05,
+      "loss": 0.5632,
+      "step": 899
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.1852112676056337e-05,
+      "loss": 0.4961,
+      "step": 900
+    },
+    {
+      "epoch": 2.46,
+      "eval_loss": 1.3768229484558105,
+      "eval_runtime": 1.4331,
+      "eval_samples_per_second": 41.868,
+      "eval_steps_per_second": 10.467,
+      "step": 900
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.1830985915492957e-05,
+      "loss": 0.4666,
+      "step": 901
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.1809859154929576e-05,
+      "loss": 0.5946,
+      "step": 902
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.1788732394366196e-05,
+      "loss": 0.5327,
+      "step": 903
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.1767605633802817e-05,
+      "loss": 0.5181,
+      "step": 904
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.1746478873239437e-05,
+      "loss": 0.6107,
+      "step": 905
+    },
+    {
+      "epoch": 2.48,
+      "eval_loss": 1.4018880128860474,
+      "eval_runtime": 1.4327,
+      "eval_samples_per_second": 41.88,
+      "eval_steps_per_second": 10.47,
+      "step": 905
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.1725352112676056e-05,
+      "loss": 0.6012,
+      "step": 906
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.1704225352112676e-05,
+      "loss": 0.5709,
+      "step": 907
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.1683098591549295e-05,
+      "loss": 0.5638,
+      "step": 908
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.1661971830985915e-05,
+      "loss": 0.6099,
+      "step": 909
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.1640845070422535e-05,
+      "loss": 0.5499,
+      "step": 910
+    },
+    {
+      "epoch": 2.49,
+      "eval_loss": 1.3791015148162842,
+      "eval_runtime": 1.4434,
+      "eval_samples_per_second": 41.57,
+      "eval_steps_per_second": 10.392,
+      "step": 910
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.1619718309859154e-05,
+      "loss": 0.562,
+      "step": 911
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.1598591549295774e-05,
+      "loss": 0.4686,
+      "step": 912
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.1577464788732395e-05,
+      "loss": 0.4535,
+      "step": 913
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.1556338028169015e-05,
+      "loss": 0.555,
+      "step": 914
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.1535211267605634e-05,
+      "loss": 0.5012,
+      "step": 915
+    },
+    {
+      "epoch": 2.51,
+      "eval_loss": 1.3942056894302368,
+      "eval_runtime": 1.441,
+      "eval_samples_per_second": 41.639,
+      "eval_steps_per_second": 10.41,
+      "step": 915
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.1514084507042254e-05,
+      "loss": 0.6003,
+      "step": 916
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.1492957746478873e-05,
+      "loss": 0.4349,
+      "step": 917
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.1471830985915493e-05,
+      "loss": 0.5399,
+      "step": 918
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.1450704225352112e-05,
+      "loss": 0.5651,
+      "step": 919
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.1429577464788732e-05,
+      "loss": 0.5103,
+      "step": 920
+    },
+    {
+      "epoch": 2.52,
+      "eval_loss": 1.3930989503860474,
+      "eval_runtime": 1.4306,
+      "eval_samples_per_second": 41.942,
+      "eval_steps_per_second": 10.485,
+      "step": 920
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.1408450704225351e-05,
+      "loss": 0.458,
+      "step": 921
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.1387323943661973e-05,
+      "loss": 0.5641,
+      "step": 922
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.1366197183098592e-05,
+      "loss": 0.5621,
+      "step": 923
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.1345070422535212e-05,
+      "loss": 0.5449,
+      "step": 924
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.1323943661971831e-05,
+      "loss": 0.4758,
+      "step": 925
+    },
+    {
+      "epoch": 2.53,
+      "eval_loss": 1.4047526121139526,
+      "eval_runtime": 1.4295,
+      "eval_samples_per_second": 41.974,
+      "eval_steps_per_second": 10.493,
+      "step": 925
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.1302816901408451e-05,
+      "loss": 0.5479,
+      "step": 926
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.128169014084507e-05,
+      "loss": 0.5792,
+      "step": 927
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.126056338028169e-05,
+      "loss": 0.5737,
+      "step": 928
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.123943661971831e-05,
+      "loss": 0.5726,
+      "step": 929
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.121830985915493e-05,
+      "loss": 0.5357,
+      "step": 930
+    },
+    {
+      "epoch": 2.55,
+      "eval_loss": 1.3789713382720947,
+      "eval_runtime": 1.4464,
+      "eval_samples_per_second": 41.483,
+      "eval_steps_per_second": 10.371,
+      "step": 930
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.119718309859155e-05,
+      "loss": 0.5409,
+      "step": 931
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.117605633802817e-05,
+      "loss": 0.5325,
+      "step": 932
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.115492957746479e-05,
+      "loss": 0.4935,
+      "step": 933
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.1133802816901409e-05,
+      "loss": 0.5735,
+      "step": 934
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.1112676056338029e-05,
+      "loss": 0.5063,
+      "step": 935
+    },
+    {
+      "epoch": 2.56,
+      "eval_loss": 1.3848307132720947,
+      "eval_runtime": 1.4481,
+      "eval_samples_per_second": 41.433,
+      "eval_steps_per_second": 10.358,
+      "step": 935
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.1091549295774648e-05,
+      "loss": 0.4995,
+      "step": 936
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.1070422535211268e-05,
+      "loss": 0.4915,
+      "step": 937
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.1049295774647887e-05,
+      "loss": 0.4227,
+      "step": 938
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.1028169014084507e-05,
+      "loss": 0.5187,
+      "step": 939
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.1007042253521127e-05,
+      "loss": 0.524,
+      "step": 940
+    },
+    {
+      "epoch": 2.57,
+      "eval_loss": 1.4117838144302368,
+      "eval_runtime": 1.4342,
+      "eval_samples_per_second": 41.836,
+      "eval_steps_per_second": 10.459,
+      "step": 940
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0985915492957748e-05,
+      "loss": 0.54,
+      "step": 941
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0964788732394367e-05,
+      "loss": 0.4993,
+      "step": 942
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0943661971830987e-05,
+      "loss": 0.4892,
+      "step": 943
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.0922535211267606e-05,
+      "loss": 0.4188,
+      "step": 944
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.0901408450704226e-05,
+      "loss": 0.5554,
+      "step": 945
+    },
+    {
+      "epoch": 2.59,
+      "eval_loss": 1.3720052242279053,
+      "eval_runtime": 1.4273,
+      "eval_samples_per_second": 42.037,
+      "eval_steps_per_second": 10.509,
+      "step": 945
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.0880281690140846e-05,
+      "loss": 0.5872,
+      "step": 946
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.0859154929577465e-05,
+      "loss": 0.5447,
+      "step": 947
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.0838028169014085e-05,
+      "loss": 0.4872,
+      "step": 948
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.0816901408450704e-05,
+      "loss": 0.5611,
+      "step": 949
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.0795774647887326e-05,
+      "loss": 0.4644,
+      "step": 950
+    },
+    {
+      "epoch": 2.6,
+      "eval_loss": 1.3852864503860474,
+      "eval_runtime": 1.4321,
+      "eval_samples_per_second": 41.896,
+      "eval_steps_per_second": 10.474,
+      "step": 950
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.0774647887323945e-05,
+      "loss": 0.5711,
+      "step": 951
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.0753521126760565e-05,
+      "loss": 0.517,
+      "step": 952
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.0732394366197184e-05,
+      "loss": 0.4472,
+      "step": 953
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.0711267605633804e-05,
+      "loss": 0.4996,
+      "step": 954
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.0690140845070423e-05,
+      "loss": 0.4794,
+      "step": 955
+    },
+    {
+      "epoch": 2.61,
+      "eval_loss": 1.4037760496139526,
+      "eval_runtime": 1.4383,
+      "eval_samples_per_second": 41.716,
+      "eval_steps_per_second": 10.429,
+      "step": 955
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.0669014084507043e-05,
+      "loss": 0.4844,
+      "step": 956
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.0647887323943662e-05,
+      "loss": 0.5343,
+      "step": 957
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.0626760563380282e-05,
+      "loss": 0.5375,
+      "step": 958
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.0605633802816902e-05,
+      "loss": 0.5455,
+      "step": 959
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.0584507042253521e-05,
+      "loss": 0.5909,
+      "step": 960
+    },
+    {
+      "epoch": 2.63,
+      "eval_loss": 1.3699870109558105,
+      "eval_runtime": 1.453,
+      "eval_samples_per_second": 41.295,
+      "eval_steps_per_second": 10.324,
+      "step": 960
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.056338028169014e-05,
+      "loss": 0.4382,
+      "step": 961
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.054225352112676e-05,
+      "loss": 0.5161,
+      "step": 962
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.052112676056338e-05,
+      "loss": 0.5566,
+      "step": 963
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.05e-05,
+      "loss": 0.5203,
+      "step": 964
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.0478873239436619e-05,
+      "loss": 0.5033,
+      "step": 965
+    },
+    {
+      "epoch": 2.64,
+      "eval_loss": 1.3866536617279053,
+      "eval_runtime": 1.4385,
+      "eval_samples_per_second": 41.71,
+      "eval_steps_per_second": 10.427,
+      "step": 965
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.0457746478873239e-05,
+      "loss": 0.5171,
+      "step": 966
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.0436619718309858e-05,
+      "loss": 0.5975,
+      "step": 967
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.041549295774648e-05,
+      "loss": 0.4913,
+      "step": 968
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.0394366197183099e-05,
+      "loss": 0.4594,
+      "step": 969
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.0373239436619718e-05,
+      "loss": 0.4857,
+      "step": 970
+    },
+    {
+      "epoch": 2.66,
+      "eval_loss": 1.4055989980697632,
+      "eval_runtime": 1.4471,
+      "eval_samples_per_second": 41.463,
+      "eval_steps_per_second": 10.366,
+      "step": 970
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.0352112676056338e-05,
+      "loss": 0.5778,
+      "step": 971
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.0330985915492958e-05,
+      "loss": 0.4927,
+      "step": 972
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.0309859154929577e-05,
+      "loss": 0.6121,
+      "step": 973
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.0288732394366197e-05,
+      "loss": 0.5162,
+      "step": 974
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.0267605633802816e-05,
+      "loss": 0.5024,
+      "step": 975
+    },
+    {
+      "epoch": 2.67,
+      "eval_loss": 1.3922525644302368,
+      "eval_runtime": 1.4516,
+      "eval_samples_per_second": 41.332,
+      "eval_steps_per_second": 10.333,
+      "step": 975
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.0246478873239436e-05,
+      "loss": 0.5154,
+      "step": 976
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.0225352112676057e-05,
+      "loss": 0.5673,
+      "step": 977
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.0204225352112677e-05,
+      "loss": 0.6453,
+      "step": 978
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.0183098591549296e-05,
+      "loss": 0.4724,
+      "step": 979
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.0161971830985916e-05,
+      "loss": 0.4335,
+      "step": 980
+    },
+    {
+      "epoch": 2.68,
+      "eval_loss": 1.3710286617279053,
+      "eval_runtime": 1.4519,
+      "eval_samples_per_second": 41.325,
+      "eval_steps_per_second": 10.331,
+      "step": 980
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.0140845070422535e-05,
+      "loss": 0.5408,
+      "step": 981
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.0119718309859155e-05,
+      "loss": 0.4648,
+      "step": 982
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.0098591549295774e-05,
+      "loss": 0.5084,
+      "step": 983
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.0077464788732394e-05,
+      "loss": 0.7192,
+      "step": 984
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.0056338028169014e-05,
+      "loss": 0.4689,
+      "step": 985
+    },
+    {
+      "epoch": 2.7,
+      "eval_loss": 1.3856770992279053,
+      "eval_runtime": 1.4332,
+      "eval_samples_per_second": 41.864,
+      "eval_steps_per_second": 10.466,
+      "step": 985
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.0035211267605633e-05,
+      "loss": 0.5226,
+      "step": 986
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.0014084507042254e-05,
+      "loss": 0.5311,
+      "step": 987
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.992957746478874e-06,
+      "loss": 0.4846,
+      "step": 988
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.971830985915494e-06,
+      "loss": 0.5513,
+      "step": 989
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.950704225352113e-06,
+      "loss": 0.4435,
+      "step": 990
+    },
+    {
+      "epoch": 2.71,
+      "eval_loss": 1.3846354484558105,
+      "eval_runtime": 1.4397,
+      "eval_samples_per_second": 41.676,
+      "eval_steps_per_second": 10.419,
+      "step": 990
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.929577464788733e-06,
+      "loss": 0.5568,
+      "step": 991
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.908450704225352e-06,
+      "loss": 0.5648,
+      "step": 992
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.887323943661972e-06,
+      "loss": 0.4664,
+      "step": 993
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.866197183098591e-06,
+      "loss": 0.5535,
+      "step": 994
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.845070422535211e-06,
+      "loss": 0.5517,
+      "step": 995
+    },
+    {
+      "epoch": 2.72,
+      "eval_loss": 1.3796875476837158,
+      "eval_runtime": 1.4328,
+      "eval_samples_per_second": 41.877,
+      "eval_steps_per_second": 10.469,
+      "step": 995
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.823943661971832e-06,
+      "loss": 0.5344,
+      "step": 996
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.802816901408452e-06,
+      "loss": 0.5807,
+      "step": 997
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.781690140845071e-06,
+      "loss": 0.5213,
+      "step": 998
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.760563380281691e-06,
+      "loss": 0.6564,
+      "step": 999
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.73943661971831e-06,
+      "loss": 0.4532,
+      "step": 1000
+    },
+    {
+      "epoch": 2.74,
+      "eval_loss": 1.3842447996139526,
+      "eval_runtime": 1.4323,
+      "eval_samples_per_second": 41.89,
+      "eval_steps_per_second": 10.472,
+      "step": 1000
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.71830985915493e-06,
+      "loss": 0.472,
+      "step": 1001
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.69718309859155e-06,
+      "loss": 0.5457,
+      "step": 1002
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.676056338028169e-06,
+      "loss": 0.5155,
+      "step": 1003
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.654929577464789e-06,
+      "loss": 0.5092,
+      "step": 1004
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.63380281690141e-06,
+      "loss": 0.5408,
+      "step": 1005
+    },
+    {
+      "epoch": 2.75,
+      "eval_loss": 1.3984375,
+      "eval_runtime": 1.4322,
+      "eval_samples_per_second": 41.894,
+      "eval_steps_per_second": 10.474,
+      "step": 1005
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.61267605633803e-06,
+      "loss": 0.4234,
+      "step": 1006
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 9.591549295774649e-06,
+      "loss": 0.5258,
+      "step": 1007
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 9.570422535211269e-06,
+      "loss": 0.4932,
+      "step": 1008
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 9.549295774647888e-06,
+      "loss": 0.5014,
+      "step": 1009
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.528169014084508e-06,
+      "loss": 0.468,
+      "step": 1010
+    },
+    {
+      "epoch": 2.77,
+      "eval_loss": 1.3862630128860474,
+      "eval_runtime": 1.4422,
+      "eval_samples_per_second": 41.604,
+      "eval_steps_per_second": 10.401,
+      "step": 1010
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.507042253521127e-06,
+      "loss": 0.5214,
+      "step": 1011
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.485915492957747e-06,
+      "loss": 0.442,
+      "step": 1012
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.464788732394366e-06,
+      "loss": 0.5084,
+      "step": 1013
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 9.443661971830988e-06,
+      "loss": 0.4957,
+      "step": 1014
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 9.422535211267607e-06,
+      "loss": 0.4692,
+      "step": 1015
+    },
+    {
+      "epoch": 2.78,
+      "eval_loss": 1.4053385257720947,
+      "eval_runtime": 1.4325,
+      "eval_samples_per_second": 41.885,
+      "eval_steps_per_second": 10.471,
+      "step": 1015
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 9.401408450704227e-06,
+      "loss": 0.5197,
+      "step": 1016
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 9.380281690140846e-06,
+      "loss": 0.5587,
+      "step": 1017
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 9.359154929577464e-06,
+      "loss": 0.5634,
+      "step": 1018
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 9.338028169014084e-06,
+      "loss": 0.6197,
+      "step": 1019
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 9.316901408450703e-06,
+      "loss": 0.523,
+      "step": 1020
+    },
+    {
+      "epoch": 2.79,
+      "eval_loss": 1.3915364742279053,
+      "eval_runtime": 1.436,
+      "eval_samples_per_second": 41.782,
+      "eval_steps_per_second": 10.446,
+      "step": 1020
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.295774647887323e-06,
+      "loss": 0.5531,
+      "step": 1021
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.274647887323943e-06,
+      "loss": 0.6052,
+      "step": 1022
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.253521126760564e-06,
+      "loss": 0.5475,
+      "step": 1023
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.232394366197183e-06,
+      "loss": 0.5607,
+      "step": 1024
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 9.211267605633803e-06,
+      "loss": 0.5332,
+      "step": 1025
+    },
+    {
+      "epoch": 2.81,
+      "eval_loss": 1.374414086341858,
+      "eval_runtime": 1.4534,
+      "eval_samples_per_second": 41.284,
+      "eval_steps_per_second": 10.321,
+      "step": 1025
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 9.190140845070422e-06,
+      "loss": 0.4565,
+      "step": 1026
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 9.169014084507042e-06,
+      "loss": 0.4909,
+      "step": 1027
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 9.147887323943662e-06,
+      "loss": 0.5252,
+      "step": 1028
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.126760563380281e-06,
+      "loss": 0.4759,
+      "step": 1029
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.1056338028169e-06,
+      "loss": 0.4613,
+      "step": 1030
+    },
+    {
+      "epoch": 2.82,
+      "eval_loss": 1.4142577648162842,
+      "eval_runtime": 1.4333,
+      "eval_samples_per_second": 41.862,
+      "eval_steps_per_second": 10.466,
+      "step": 1030
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.08450704225352e-06,
+      "loss": 0.5212,
+      "step": 1031
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9.06338028169014e-06,
+      "loss": 0.5292,
+      "step": 1032
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9.042253521126761e-06,
+      "loss": 0.5511,
+      "step": 1033
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9.02112676056338e-06,
+      "loss": 0.5376,
+      "step": 1034
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9e-06,
+      "loss": 0.5605,
+      "step": 1035
+    },
+    {
+      "epoch": 2.83,
+      "eval_loss": 1.3954427242279053,
+      "eval_runtime": 1.4305,
+      "eval_samples_per_second": 41.943,
+      "eval_steps_per_second": 10.486,
+      "step": 1035
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.97887323943662e-06,
+      "loss": 0.4914,
+      "step": 1036
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.95774647887324e-06,
+      "loss": 0.5815,
+      "step": 1037
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.936619718309859e-06,
+      "loss": 0.5569,
+      "step": 1038
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 8.915492957746478e-06,
+      "loss": 0.5591,
+      "step": 1039
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.894366197183098e-06,
+      "loss": 0.4716,
+      "step": 1040
+    },
+    {
+      "epoch": 2.85,
+      "eval_loss": 1.3823567628860474,
+      "eval_runtime": 1.4343,
+      "eval_samples_per_second": 41.832,
+      "eval_steps_per_second": 10.458,
+      "step": 1040
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.873239436619718e-06,
+      "loss": 0.5032,
+      "step": 1041
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 8.852112676056339e-06,
+      "loss": 0.4578,
+      "step": 1042
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.830985915492958e-06,
+      "loss": 0.4716,
+      "step": 1043
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.809859154929578e-06,
+      "loss": 0.5526,
+      "step": 1044
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.788732394366198e-06,
+      "loss": 0.4974,
+      "step": 1045
+    },
+    {
+      "epoch": 2.86,
+      "eval_loss": 1.4059895277023315,
+      "eval_runtime": 1.4359,
+      "eval_samples_per_second": 41.785,
+      "eval_steps_per_second": 10.446,
+      "step": 1045
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 8.767605633802817e-06,
+      "loss": 0.4811,
+      "step": 1046
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.746478873239437e-06,
+      "loss": 0.5292,
+      "step": 1047
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.725352112676056e-06,
+      "loss": 0.433,
+      "step": 1048
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.704225352112676e-06,
+      "loss": 0.5885,
+      "step": 1049
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.683098591549295e-06,
+      "loss": 0.4904,
+      "step": 1050
+    },
+    {
+      "epoch": 2.87,
+      "eval_loss": 1.37890625,
+      "eval_runtime": 1.4331,
+      "eval_samples_per_second": 41.868,
+      "eval_steps_per_second": 10.467,
+      "step": 1050
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.661971830985917e-06,
+      "loss": 0.5411,
+      "step": 1051
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.640845070422536e-06,
+      "loss": 0.5583,
+      "step": 1052
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.619718309859156e-06,
+      "loss": 0.5664,
+      "step": 1053
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.598591549295775e-06,
+      "loss": 0.5432,
+      "step": 1054
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.577464788732395e-06,
+      "loss": 0.5286,
+      "step": 1055
+    },
+    {
+      "epoch": 2.89,
+      "eval_loss": 1.3819010257720947,
+      "eval_runtime": 1.441,
+      "eval_samples_per_second": 41.637,
+      "eval_steps_per_second": 10.409,
+      "step": 1055
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.556338028169014e-06,
+      "loss": 0.4473,
+      "step": 1056
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 8.535211267605634e-06,
+      "loss": 0.552,
+      "step": 1057
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 8.514084507042254e-06,
+      "loss": 0.5642,
+      "step": 1058
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 8.492957746478873e-06,
+      "loss": 0.4626,
+      "step": 1059
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 8.471830985915494e-06,
+      "loss": 0.4858,
+      "step": 1060
+    },
+    {
+      "epoch": 2.9,
+      "eval_loss": 1.397851586341858,
+      "eval_runtime": 1.4487,
+      "eval_samples_per_second": 41.417,
+      "eval_steps_per_second": 10.354,
+      "step": 1060
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 8.450704225352114e-06,
+      "loss": 0.531,
+      "step": 1061
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 8.429577464788734e-06,
+      "loss": 0.4705,
+      "step": 1062
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 8.408450704225353e-06,
+      "loss": 0.4227,
+      "step": 1063
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 8.387323943661973e-06,
+      "loss": 0.4742,
+      "step": 1064
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.366197183098592e-06,
+      "loss": 0.5513,
+      "step": 1065
+    },
+    {
+      "epoch": 2.92,
+      "eval_loss": 1.4163411855697632,
+      "eval_runtime": 1.4316,
+      "eval_samples_per_second": 41.91,
+      "eval_steps_per_second": 10.478,
+      "step": 1065
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.345070422535212e-06,
+      "loss": 0.4987,
+      "step": 1066
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.323943661971831e-06,
+      "loss": 0.5222,
+      "step": 1067
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.302816901408451e-06,
+      "loss": 0.5615,
+      "step": 1068
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 8.28169014084507e-06,
+      "loss": 0.5894,
+      "step": 1069
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 8.260563380281692e-06,
+      "loss": 0.4795,
+      "step": 1070
+    },
+    {
+      "epoch": 2.93,
+      "eval_loss": 1.3757160902023315,
+      "eval_runtime": 1.4327,
+      "eval_samples_per_second": 41.878,
+      "eval_steps_per_second": 10.469,
+      "step": 1070
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 8.239436619718311e-06,
+      "loss": 0.4339,
+      "step": 1071
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 8.21830985915493e-06,
+      "loss": 0.5669,
+      "step": 1072
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 8.19718309859155e-06,
+      "loss": 0.533,
+      "step": 1073
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 8.17605633802817e-06,
+      "loss": 0.4661,
+      "step": 1074
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 8.15492957746479e-06,
+      "loss": 0.504,
+      "step": 1075
+    },
+    {
+      "epoch": 2.94,
+      "eval_loss": 1.390234351158142,
+      "eval_runtime": 1.4378,
+      "eval_samples_per_second": 41.73,
+      "eval_steps_per_second": 10.432,
+      "step": 1075
+    }
+  ],
+  "max_steps": 1460,
+  "num_train_epochs": 4,
+  "total_flos": 4.48654593325269e+17,
+  "trial_name": null,
+  "trial_params": null
+}