diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,4686 @@
+{
+  "best_metric": 0.011437707580626011,
+  "best_model_checkpoint": "/home/paperspace/Data/models/asianpaints/llm3br256/checkpoint-400",
+  "epoch": 4.9655172413793105,
+  "eval_steps": 5,
+  "global_step": 540,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.009195402298850575,
+      "grad_norm": 0.13789793848991394,
+      "learning_rate": 1.8518518518518519e-06,
+      "loss": 0.0591,
+      "step": 1
+    },
+    {
+      "epoch": 0.01839080459770115,
+      "grad_norm": 0.13098689913749695,
+      "learning_rate": 3.7037037037037037e-06,
+      "loss": 0.0572,
+      "step": 2
+    },
+    {
+      "epoch": 0.027586206896551724,
+      "grad_norm": 0.12751829624176025,
+      "learning_rate": 5.555555555555556e-06,
+      "loss": 0.0586,
+      "step": 3
+    },
+    {
+      "epoch": 0.0367816091954023,
+      "grad_norm": 0.12321432679891586,
+      "learning_rate": 7.4074074074074075e-06,
+      "loss": 0.0521,
+      "step": 4
+    },
+    {
+      "epoch": 0.04597701149425287,
+      "grad_norm": 0.12428242713212967,
+      "learning_rate": 9.259259259259259e-06,
+      "loss": 0.0567,
+      "step": 5
+    },
+    {
+      "epoch": 0.04597701149425287,
+      "eval_loss": 0.0583593025803566,
+      "eval_runtime": 10.2736,
+      "eval_samples_per_second": 4.867,
+      "eval_steps_per_second": 1.265,
+      "step": 5
+    },
+    {
+      "epoch": 0.05517241379310345,
+      "grad_norm": 0.10534916073083878,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 0.0485,
+      "step": 6
+    },
+    {
+      "epoch": 0.06436781609195402,
+      "grad_norm": 0.08470993489027023,
+      "learning_rate": 1.2962962962962962e-05,
+      "loss": 0.0439,
+      "step": 7
+    },
+    {
+      "epoch": 0.0735632183908046,
+      "grad_norm": 0.07256518304347992,
+      "learning_rate": 1.4814814814814815e-05,
+      "loss": 0.0402,
+      "step": 8
+    },
+    {
+      "epoch": 0.08275862068965517,
+      "grad_norm": 0.060682956129312515,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.0374,
+      "step": 9
+    },
+    {
+      "epoch": 0.09195402298850575,
+      "grad_norm": 0.05973079800605774,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.0378,
+      "step": 10
+    },
+    {
+      "epoch": 0.09195402298850575,
+      "eval_loss": 0.0383976586163044,
+      "eval_runtime": 7.5937,
+      "eval_samples_per_second": 6.584,
+      "eval_steps_per_second": 1.712,
+      "step": 10
+    },
+    {
+      "epoch": 0.10114942528735632,
+      "grad_norm": 0.06681109964847565,
+      "learning_rate": 2.037037037037037e-05,
+      "loss": 0.0335,
+      "step": 11
+    },
+    {
+      "epoch": 0.1103448275862069,
+      "grad_norm": 0.07075705379247665,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.0314,
+      "step": 12
+    },
+    {
+      "epoch": 0.11954022988505747,
+      "grad_norm": 0.05252094194293022,
+      "learning_rate": 2.4074074074074074e-05,
+      "loss": 0.0347,
+      "step": 13
+    },
+    {
+      "epoch": 0.12873563218390804,
+      "grad_norm": 0.04202576354146004,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.03,
+      "step": 14
+    },
+    {
+      "epoch": 0.13793103448275862,
+      "grad_norm": 0.03977281600236893,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.0301,
+      "step": 15
+    },
+    {
+      "epoch": 0.13793103448275862,
+      "eval_loss": 0.0318371020257473,
+      "eval_runtime": 7.612,
+      "eval_samples_per_second": 6.569,
+      "eval_steps_per_second": 1.708,
+      "step": 15
+    },
+    {
+      "epoch": 0.1471264367816092,
+      "grad_norm": 0.04173818975687027,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.0286,
+      "step": 16
+    },
+    {
+      "epoch": 0.15632183908045977,
+      "grad_norm": 0.039005085825920105,
+      "learning_rate": 3.148148148148148e-05,
+      "loss": 0.0281,
+      "step": 17
+    },
+    {
+      "epoch": 0.16551724137931034,
+      "grad_norm": 0.03674433380365372,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.0248,
+      "step": 18
+    },
+    {
+      "epoch": 0.17471264367816092,
+      "grad_norm": 0.035750504583120346,
+      "learning_rate": 3.518518518518519e-05,
+      "loss": 0.0256,
+      "step": 19
+    },
+    {
+      "epoch": 0.1839080459770115,
+      "grad_norm": 0.027469012886285782,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.0248,
+      "step": 20
+    },
+    {
+      "epoch": 0.1839080459770115,
+      "eval_loss": 0.028074176982045174,
+      "eval_runtime": 7.6024,
+      "eval_samples_per_second": 6.577,
+      "eval_steps_per_second": 1.71,
+      "step": 20
+    },
+    {
+      "epoch": 0.19310344827586207,
+      "grad_norm": 0.026447350159287453,
+      "learning_rate": 3.888888888888889e-05,
+      "loss": 0.0231,
+      "step": 21
+    },
+    {
+      "epoch": 0.20229885057471264,
+      "grad_norm": 0.031306445598602295,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.0236,
+      "step": 22
+    },
+    {
+      "epoch": 0.21149425287356322,
+      "grad_norm": 0.029817912727594376,
+      "learning_rate": 4.259259259259259e-05,
+      "loss": 0.0227,
+      "step": 23
+    },
+    {
+      "epoch": 0.2206896551724138,
+      "grad_norm": 0.03197651728987694,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.0277,
+      "step": 24
+    },
+    {
+      "epoch": 0.22988505747126436,
+      "grad_norm": 0.03044748678803444,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 0.0241,
+      "step": 25
+    },
+    {
+      "epoch": 0.22988505747126436,
+      "eval_loss": 0.025574278086423874,
+      "eval_runtime": 7.6017,
+      "eval_samples_per_second": 6.577,
+      "eval_steps_per_second": 1.71,
+      "step": 25
+    },
+    {
+      "epoch": 0.23908045977011494,
+      "grad_norm": 0.02689642831683159,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.0253,
+      "step": 26
+    },
+    {
+      "epoch": 0.2482758620689655,
+      "grad_norm": 0.0260293111205101,
+      "learning_rate": 5e-05,
+      "loss": 0.0231,
+      "step": 27
+    },
+    {
+      "epoch": 0.2574712643678161,
+      "grad_norm": 0.025041967630386353,
+      "learning_rate": 5.185185185185185e-05,
+      "loss": 0.0244,
+      "step": 28
+    },
+    {
+      "epoch": 0.26666666666666666,
+      "grad_norm": 0.025733083486557007,
+      "learning_rate": 5.370370370370371e-05,
+      "loss": 0.0219,
+      "step": 29
+    },
+    {
+      "epoch": 0.27586206896551724,
+      "grad_norm": 0.023718759417533875,
+      "learning_rate": 5.555555555555556e-05,
+      "loss": 0.021,
+      "step": 30
+    },
+    {
+      "epoch": 0.27586206896551724,
+      "eval_loss": 0.0234465803951025,
+      "eval_runtime": 7.6024,
+      "eval_samples_per_second": 6.577,
+      "eval_steps_per_second": 1.71,
+      "step": 30
+    },
+    {
+      "epoch": 0.2850574712643678,
+      "grad_norm": 0.02254614047706127,
+      "learning_rate": 5.740740740740741e-05,
+      "loss": 0.0213,
+      "step": 31
+    },
+    {
+      "epoch": 0.2942528735632184,
+      "grad_norm": 0.024035446345806122,
+      "learning_rate": 5.925925925925926e-05,
+      "loss": 0.0215,
+      "step": 32
+    },
+    {
+      "epoch": 0.30344827586206896,
+      "grad_norm": 0.02866807021200657,
+      "learning_rate": 6.111111111111112e-05,
+      "loss": 0.0211,
+      "step": 33
+    },
+    {
+      "epoch": 0.31264367816091954,
+      "grad_norm": 0.02149030566215515,
+      "learning_rate": 6.296296296296296e-05,
+      "loss": 0.0196,
+      "step": 34
+    },
+    {
+      "epoch": 0.3218390804597701,
+      "grad_norm": 0.021923156455159187,
+      "learning_rate": 6.481481481481482e-05,
+      "loss": 0.0213,
+      "step": 35
+    },
+    {
+      "epoch": 0.3218390804597701,
+      "eval_loss": 0.022549211978912354,
+      "eval_runtime": 7.607,
+      "eval_samples_per_second": 6.573,
+      "eval_steps_per_second": 1.709,
+      "step": 35
+    },
+    {
+      "epoch": 0.3310344827586207,
+      "grad_norm": 0.022337742149829865,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 0.0191,
+      "step": 36
+    },
+    {
+      "epoch": 0.34022988505747126,
+      "grad_norm": 0.025172578170895576,
+      "learning_rate": 6.851851851851852e-05,
+      "loss": 0.0194,
+      "step": 37
+    },
+    {
+      "epoch": 0.34942528735632183,
+      "grad_norm": 0.01893443800508976,
+      "learning_rate": 7.037037037037038e-05,
+      "loss": 0.0209,
+      "step": 38
+    },
+    {
+      "epoch": 0.3586206896551724,
+      "grad_norm": 0.022978560999035835,
+      "learning_rate": 7.222222222222222e-05,
+      "loss": 0.0195,
+      "step": 39
+    },
+    {
+      "epoch": 0.367816091954023,
+      "grad_norm": 0.021215129643678665,
+      "learning_rate": 7.407407407407407e-05,
+      "loss": 0.0211,
+      "step": 40
+    },
+    {
+      "epoch": 0.367816091954023,
+      "eval_loss": 0.0213603638112545,
+      "eval_runtime": 7.5795,
+      "eval_samples_per_second": 6.597,
+      "eval_steps_per_second": 1.715,
+      "step": 40
+    },
+    {
+      "epoch": 0.37701149425287356,
+      "grad_norm": 0.022082313895225525,
+      "learning_rate": 7.592592592592593e-05,
+      "loss": 0.0199,
+      "step": 41
+    },
+    {
+      "epoch": 0.38620689655172413,
+      "grad_norm": 0.020405782386660576,
+      "learning_rate": 7.777777777777778e-05,
+      "loss": 0.0188,
+      "step": 42
+    },
+    {
+      "epoch": 0.3954022988505747,
+      "grad_norm": 0.020502161234617233,
+      "learning_rate": 7.962962962962964e-05,
+      "loss": 0.0203,
+      "step": 43
+    },
+    {
+      "epoch": 0.4045977011494253,
+      "grad_norm": 0.020312348380684853,
+      "learning_rate": 8.148148148148148e-05,
+      "loss": 0.0184,
+      "step": 44
+    },
+    {
+      "epoch": 0.41379310344827586,
+      "grad_norm": 0.01888488046824932,
+      "learning_rate": 8.333333333333334e-05,
+      "loss": 0.0185,
+      "step": 45
+    },
+    {
+      "epoch": 0.41379310344827586,
+      "eval_loss": 0.020027954131364822,
+      "eval_runtime": 7.5799,
+      "eval_samples_per_second": 6.596,
+      "eval_steps_per_second": 1.715,
+      "step": 45
+    },
+    {
+      "epoch": 0.42298850574712643,
+      "grad_norm": 0.016626283526420593,
+      "learning_rate": 8.518518518518518e-05,
+      "loss": 0.0175,
+      "step": 46
+    },
+    {
+      "epoch": 0.432183908045977,
+      "grad_norm": 0.01808527484536171,
+      "learning_rate": 8.703703703703704e-05,
+      "loss": 0.017,
+      "step": 47
+    },
+    {
+      "epoch": 0.4413793103448276,
+      "grad_norm": 0.02142912708222866,
+      "learning_rate": 8.888888888888889e-05,
+      "loss": 0.018,
+      "step": 48
+    },
+    {
+      "epoch": 0.45057471264367815,
+      "grad_norm": 0.017742585390806198,
+      "learning_rate": 9.074074074074075e-05,
+      "loss": 0.0165,
+      "step": 49
+    },
+    {
+      "epoch": 0.45977011494252873,
+      "grad_norm": 0.022063937038183212,
+      "learning_rate": 9.25925925925926e-05,
+      "loss": 0.0162,
+      "step": 50
+    },
+    {
+      "epoch": 0.45977011494252873,
+      "eval_loss": 0.019592924043536186,
+      "eval_runtime": 7.5839,
+      "eval_samples_per_second": 6.593,
+      "eval_steps_per_second": 1.714,
+      "step": 50
+    },
+    {
+      "epoch": 0.4689655172413793,
+      "grad_norm": 0.018968017771840096,
+      "learning_rate": 9.444444444444444e-05,
+      "loss": 0.0183,
+      "step": 51
+    },
+    {
+      "epoch": 0.4781609195402299,
+      "grad_norm": 0.024733267724514008,
+      "learning_rate": 9.62962962962963e-05,
+      "loss": 0.0189,
+      "step": 52
+    },
+    {
+      "epoch": 0.48735632183908045,
+      "grad_norm": 0.021883539855480194,
+      "learning_rate": 9.814814814814815e-05,
+      "loss": 0.0188,
+      "step": 53
+    },
+    {
+      "epoch": 0.496551724137931,
+      "grad_norm": 0.017187397927045822,
+      "learning_rate": 0.0001,
+      "loss": 0.0167,
+      "step": 54
+    },
+    {
+      "epoch": 0.5057471264367817,
+      "grad_norm": 0.018074609339237213,
+      "learning_rate": 9.99989553622803e-05,
+      "loss": 0.0177,
+      "step": 55
+    },
+    {
+      "epoch": 0.5057471264367817,
+      "eval_loss": 0.0189231988042593,
+      "eval_runtime": 7.6176,
+      "eval_samples_per_second": 6.564,
+      "eval_steps_per_second": 1.707,
+      "step": 55
+    },
+    {
+      "epoch": 0.5149425287356322,
+      "grad_norm": 0.017995629459619522,
+      "learning_rate": 9.999582149277187e-05,
+      "loss": 0.0162,
+      "step": 56
+    },
+    {
+      "epoch": 0.5241379310344828,
+      "grad_norm": 0.018664268776774406,
+      "learning_rate": 9.999059852242507e-05,
+      "loss": 0.0153,
+      "step": 57
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 0.017265547066926956,
+      "learning_rate": 9.998328666948438e-05,
+      "loss": 0.0189,
+      "step": 58
+    },
+    {
+      "epoch": 0.542528735632184,
+      "grad_norm": 0.01721198670566082,
+      "learning_rate": 9.997388623947928e-05,
+      "loss": 0.0158,
+      "step": 59
+    },
+    {
+      "epoch": 0.5517241379310345,
+      "grad_norm": 0.0194217711687088,
+      "learning_rate": 9.996239762521151e-05,
+      "loss": 0.0168,
+      "step": 60
+    },
+    {
+      "epoch": 0.5517241379310345,
+      "eval_loss": 0.018419023603200912,
+      "eval_runtime": 7.5901,
+      "eval_samples_per_second": 6.588,
+      "eval_steps_per_second": 1.713,
+      "step": 60
+    },
+    {
+      "epoch": 0.5609195402298851,
+      "grad_norm": 0.019492702558636665,
+      "learning_rate": 9.994882130673868e-05,
+      "loss": 0.0186,
+      "step": 61
+    },
+    {
+      "epoch": 0.5701149425287356,
+      "grad_norm": 0.01645253598690033,
+      "learning_rate": 9.993315785135416e-05,
+      "loss": 0.0175,
+      "step": 62
+    },
+    {
+      "epoch": 0.5793103448275863,
+      "grad_norm": 0.018438637256622314,
+      "learning_rate": 9.991540791356342e-05,
+      "loss": 0.0148,
+      "step": 63
+    },
+    {
+      "epoch": 0.5885057471264368,
+      "grad_norm": 0.016131950542330742,
+      "learning_rate": 9.989557223505661e-05,
+      "loss": 0.0163,
+      "step": 64
+    },
+    {
+      "epoch": 0.5977011494252874,
+      "grad_norm": 0.018596597015857697,
+      "learning_rate": 9.987365164467767e-05,
+      "loss": 0.017,
+      "step": 65
+    },
+    {
+      "epoch": 0.5977011494252874,
+      "eval_loss": 0.01821504347026348,
+      "eval_runtime": 7.5805,
+      "eval_samples_per_second": 6.596,
+      "eval_steps_per_second": 1.715,
+      "step": 65
+    },
+    {
+      "epoch": 0.6068965517241379,
+      "grad_norm": 0.017436878755688667,
+      "learning_rate": 9.98496470583896e-05,
+      "loss": 0.016,
+      "step": 66
+    },
+    {
+      "epoch": 0.6160919540229886,
+      "grad_norm": 0.017340335994958878,
+      "learning_rate": 9.982355947923629e-05,
+      "loss": 0.017,
+      "step": 67
+    },
+    {
+      "epoch": 0.6252873563218391,
+      "grad_norm": 0.017135461792349815,
+      "learning_rate": 9.979538999730047e-05,
+      "loss": 0.0149,
+      "step": 68
+    },
+    {
+      "epoch": 0.6344827586206897,
+      "grad_norm": 0.018239280208945274,
+      "learning_rate": 9.976513978965829e-05,
+      "loss": 0.0165,
+      "step": 69
+    },
+    {
+      "epoch": 0.6436781609195402,
+      "grad_norm": 0.016211749985814095,
+      "learning_rate": 9.973281012033007e-05,
+      "loss": 0.0143,
+      "step": 70
+    },
+    {
+      "epoch": 0.6436781609195402,
+      "eval_loss": 0.017700908705592155,
+      "eval_runtime": 7.5856,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 70
+    },
+    {
+      "epoch": 0.6528735632183909,
+      "grad_norm": 0.016638092696666718,
+      "learning_rate": 9.969840234022749e-05,
+      "loss": 0.0154,
+      "step": 71
+    },
+    {
+      "epoch": 0.6620689655172414,
+      "grad_norm": 0.020088642835617065,
+      "learning_rate": 9.966191788709716e-05,
+      "loss": 0.0163,
+      "step": 72
+    },
+    {
+      "epoch": 0.671264367816092,
+      "grad_norm": 0.01989225298166275,
+      "learning_rate": 9.962335828546048e-05,
+      "loss": 0.0155,
+      "step": 73
+    },
+    {
+      "epoch": 0.6804597701149425,
+      "grad_norm": 0.01787324622273445,
+      "learning_rate": 9.958272514655006e-05,
+      "loss": 0.0145,
+      "step": 74
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 0.018312491476535797,
+      "learning_rate": 9.954002016824227e-05,
+      "loss": 0.0143,
+      "step": 75
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "eval_loss": 0.017578113824129105,
+      "eval_runtime": 7.6044,
+      "eval_samples_per_second": 6.575,
+      "eval_steps_per_second": 1.71,
+      "step": 75
+    },
+    {
+      "epoch": 0.6988505747126437,
+      "grad_norm": 0.019170604646205902,
+      "learning_rate": 9.949524513498636e-05,
+      "loss": 0.0168,
+      "step": 76
+    },
+    {
+      "epoch": 0.7080459770114943,
+      "grad_norm": 0.022306382656097412,
+      "learning_rate": 9.944840191772987e-05,
+      "loss": 0.0159,
+      "step": 77
+    },
+    {
+      "epoch": 0.7172413793103448,
+      "grad_norm": 0.01719992235302925,
+      "learning_rate": 9.939949247384046e-05,
+      "loss": 0.0167,
+      "step": 78
+    },
+    {
+      "epoch": 0.7264367816091954,
+      "grad_norm": 0.020334061235189438,
+      "learning_rate": 9.934851884702414e-05,
+      "loss": 0.0159,
+      "step": 79
+    },
+    {
+      "epoch": 0.735632183908046,
+      "grad_norm": 0.016988877207040787,
+      "learning_rate": 9.929548316723982e-05,
+      "loss": 0.0155,
+      "step": 80
+    },
+    {
+      "epoch": 0.735632183908046,
+      "eval_loss": 0.01759597845375538,
+      "eval_runtime": 7.5797,
+      "eval_samples_per_second": 6.597,
+      "eval_steps_per_second": 1.715,
+      "step": 80
+    },
+    {
+      "epoch": 0.7448275862068966,
+      "grad_norm": 0.01999749429523945,
+      "learning_rate": 9.924038765061042e-05,
+      "loss": 0.0156,
+      "step": 81
+    },
+    {
+      "epoch": 0.7540229885057471,
+      "grad_norm": 0.021031200885772705,
+      "learning_rate": 9.918323459933005e-05,
+      "loss": 0.017,
+      "step": 82
+    },
+    {
+      "epoch": 0.7632183908045977,
+      "grad_norm": 0.015833275392651558,
+      "learning_rate": 9.912402640156811e-05,
+      "loss": 0.0151,
+      "step": 83
+    },
+    {
+      "epoch": 0.7724137931034483,
+      "grad_norm": 0.02128308266401291,
+      "learning_rate": 9.906276553136923e-05,
+      "loss": 0.0174,
+      "step": 84
+    },
+    {
+      "epoch": 0.7816091954022989,
+      "grad_norm": 0.01704588159918785,
+      "learning_rate": 9.899945454855006e-05,
+      "loss": 0.0162,
+      "step": 85
+    },
+    {
+      "epoch": 0.7816091954022989,
+      "eval_loss": 0.016866466030478477,
+      "eval_runtime": 7.5833,
+      "eval_samples_per_second": 6.593,
+      "eval_steps_per_second": 1.714,
+      "step": 85
+    },
+    {
+      "epoch": 0.7908045977011494,
+      "grad_norm": 0.01643301546573639,
+      "learning_rate": 9.893409609859222e-05,
+      "loss": 0.0152,
+      "step": 86
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.017949821427464485,
+      "learning_rate": 9.88666929125318e-05,
+      "loss": 0.0156,
+      "step": 87
+    },
+    {
+      "epoch": 0.8091954022988506,
+      "grad_norm": 0.018931446596980095,
+      "learning_rate": 9.879724780684519e-05,
+      "loss": 0.0162,
+      "step": 88
+    },
+    {
+      "epoch": 0.8183908045977012,
+      "grad_norm": 0.012615452520549297,
+      "learning_rate": 9.872576368333151e-05,
+      "loss": 0.0138,
+      "step": 89
+    },
+    {
+      "epoch": 0.8275862068965517,
+      "grad_norm": 0.016196150332689285,
+      "learning_rate": 9.865224352899119e-05,
+      "loss": 0.0164,
+      "step": 90
+    },
+    {
+      "epoch": 0.8275862068965517,
+      "eval_loss": 0.01638174243271351,
+      "eval_runtime": 7.5993,
+      "eval_samples_per_second": 6.58,
+      "eval_steps_per_second": 1.711,
+      "step": 90
+    },
+    {
+      "epoch": 0.8367816091954023,
+      "grad_norm": 0.016620157286524773,
+      "learning_rate": 9.857669041590134e-05,
+      "loss": 0.0158,
+      "step": 91
+    },
+    {
+      "epoch": 0.8459770114942529,
+      "grad_norm": 0.014475880190730095,
+      "learning_rate": 9.849910750108717e-05,
+      "loss": 0.0149,
+      "step": 92
+    },
+    {
+      "epoch": 0.8551724137931035,
+      "grad_norm": 0.016380243003368378,
+      "learning_rate": 9.84194980263903e-05,
+      "loss": 0.0147,
+      "step": 93
+    },
+    {
+      "epoch": 0.864367816091954,
+      "grad_norm": 0.015107197687029839,
+      "learning_rate": 9.83378653183331e-05,
+      "loss": 0.0148,
+      "step": 94
+    },
+    {
+      "epoch": 0.8735632183908046,
+      "grad_norm": 0.01710473746061325,
+      "learning_rate": 9.825421278797983e-05,
+      "loss": 0.0154,
+      "step": 95
+    },
+    {
+      "epoch": 0.8735632183908046,
+      "eval_loss": 0.0161591824144125,
+      "eval_runtime": 7.6104,
+      "eval_samples_per_second": 6.57,
+      "eval_steps_per_second": 1.708,
+      "step": 95
+    },
+    {
+      "epoch": 0.8827586206896552,
+      "grad_norm": 0.015920141711831093,
+      "learning_rate": 9.816854393079403e-05,
+      "loss": 0.0159,
+      "step": 96
+    },
+    {
+      "epoch": 0.8919540229885058,
+      "grad_norm": 0.015056249685585499,
+      "learning_rate": 9.808086232649246e-05,
+      "loss": 0.0154,
+      "step": 97
+    },
+    {
+      "epoch": 0.9011494252873563,
+      "grad_norm": 0.016236839815974236,
+      "learning_rate": 9.799117163889559e-05,
+      "loss": 0.0152,
+      "step": 98
+    },
+    {
+      "epoch": 0.9103448275862069,
+      "grad_norm": 0.014755873009562492,
+      "learning_rate": 9.789947561577445e-05,
+      "loss": 0.0135,
+      "step": 99
+    },
+    {
+      "epoch": 0.9195402298850575,
+      "grad_norm": 0.015870464965701103,
+      "learning_rate": 9.780577808869398e-05,
+      "loss": 0.0164,
+      "step": 100
+    },
+    {
+      "epoch": 0.9195402298850575,
+      "eval_loss": 0.015854647383093834,
+      "eval_runtime": 7.5871,
+      "eval_samples_per_second": 6.59,
+      "eval_steps_per_second": 1.713,
+      "step": 100
+    },
+    {
+      "epoch": 0.9287356321839081,
+      "grad_norm": 0.015317806974053383,
+      "learning_rate": 9.771008297285307e-05,
+      "loss": 0.0144,
+      "step": 101
+    },
+    {
+      "epoch": 0.9379310344827586,
+      "grad_norm": 0.015459987334907055,
+      "learning_rate": 9.761239426692077e-05,
+      "loss": 0.0136,
+      "step": 102
+    },
+    {
+      "epoch": 0.9471264367816092,
+      "grad_norm": 0.01605418138206005,
+      "learning_rate": 9.751271605286941e-05,
+      "loss": 0.0133,
+      "step": 103
+    },
+    {
+      "epoch": 0.9563218390804598,
+      "grad_norm": 0.01653791405260563,
+      "learning_rate": 9.741105249580383e-05,
+      "loss": 0.0142,
+      "step": 104
+    },
+    {
+      "epoch": 0.9655172413793104,
+      "grad_norm": 0.017847595736384392,
+      "learning_rate": 9.730740784378753e-05,
+      "loss": 0.0156,
+      "step": 105
+    },
+    {
+      "epoch": 0.9655172413793104,
+      "eval_loss": 0.01596776954829693,
+      "eval_runtime": 7.5993,
+      "eval_samples_per_second": 6.58,
+      "eval_steps_per_second": 1.711,
+      "step": 105
+    },
+    {
+      "epoch": 0.9747126436781609,
+      "grad_norm": 0.01591021567583084,
+      "learning_rate": 9.7201786427665e-05,
+      "loss": 0.0142,
+      "step": 106
+    },
+    {
+      "epoch": 0.9839080459770115,
+      "grad_norm": 0.019531626254320145,
+      "learning_rate": 9.709419266088086e-05,
+      "loss": 0.0149,
+      "step": 107
+    },
+    {
+      "epoch": 0.993103448275862,
+      "grad_norm": 0.017293043434619904,
+      "learning_rate": 9.698463103929542e-05,
+      "loss": 0.0145,
+      "step": 108
+    },
+    {
+      "epoch": 1.0022988505747126,
+      "grad_norm": 0.01688973978161812,
+      "learning_rate": 9.687310614099675e-05,
+      "loss": 0.0165,
+      "step": 109
+    },
+    {
+      "epoch": 1.0114942528735633,
+      "grad_norm": 0.021913839504122734,
+      "learning_rate": 9.67596226261095e-05,
+      "loss": 0.0145,
+      "step": 110
+    },
+    {
+      "epoch": 1.0114942528735633,
+      "eval_loss": 0.01590101048350334,
+      "eval_runtime": 7.5856,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 110
+    },
+    {
+      "epoch": 1.0206896551724138,
+      "grad_norm": 0.018519939854741096,
+      "learning_rate": 9.664418523660004e-05,
+      "loss": 0.0131,
+      "step": 111
+    },
+    {
+      "epoch": 1.0298850574712644,
+      "grad_norm": 0.020266039296984673,
+      "learning_rate": 9.652679879607843e-05,
+      "loss": 0.0129,
+      "step": 112
+    },
+    {
+      "epoch": 1.0390804597701149,
+      "grad_norm": 0.021503202617168427,
+      "learning_rate": 9.640746820959684e-05,
+      "loss": 0.0147,
+      "step": 113
+    },
+    {
+      "epoch": 1.0482758620689656,
+      "grad_norm": 0.01816105656325817,
+      "learning_rate": 9.628619846344454e-05,
+      "loss": 0.0131,
+      "step": 114
+    },
+    {
+      "epoch": 1.0574712643678161,
+      "grad_norm": 0.02214297465980053,
+      "learning_rate": 9.616299462493952e-05,
+      "loss": 0.0133,
+      "step": 115
+    },
+    {
+      "epoch": 1.0574712643678161,
+      "eval_loss": 0.015581479296088219,
+      "eval_runtime": 7.5911,
+      "eval_samples_per_second": 6.587,
+      "eval_steps_per_second": 1.713,
+      "step": 115
+    },
+    {
+      "epoch": 1.0666666666666667,
+      "grad_norm": 0.022218503057956696,
+      "learning_rate": 9.603786184221693e-05,
+      "loss": 0.0154,
+      "step": 116
+    },
+    {
+      "epoch": 1.0758620689655172,
+      "grad_norm": 0.019172191619873047,
+      "learning_rate": 9.591080534401371e-05,
+      "loss": 0.0136,
+      "step": 117
+    },
+    {
+      "epoch": 1.085057471264368,
+      "grad_norm": 0.021756643429398537,
+      "learning_rate": 9.57818304394503e-05,
+      "loss": 0.0142,
+      "step": 118
+    },
+    {
+      "epoch": 1.0942528735632184,
+      "grad_norm": 0.018046511337161064,
+      "learning_rate": 9.565094251780871e-05,
+      "loss": 0.0144,
+      "step": 119
+    },
+    {
+      "epoch": 1.103448275862069,
+      "grad_norm": 0.02018323354423046,
+      "learning_rate": 9.551814704830734e-05,
+      "loss": 0.0126,
+      "step": 120
+    },
+    {
+      "epoch": 1.103448275862069,
+      "eval_loss": 0.01549120806157589,
+      "eval_runtime": 7.5874,
+      "eval_samples_per_second": 6.59,
+      "eval_steps_per_second": 1.713,
+      "step": 120
+    },
+    {
+      "epoch": 1.1126436781609195,
+      "grad_norm": 0.02136109583079815,
+      "learning_rate": 9.538344957987244e-05,
+      "loss": 0.0144,
+      "step": 121
+    },
+    {
+      "epoch": 1.1218390804597702,
+      "grad_norm": 0.0185832716524601,
+      "learning_rate": 9.524685574090627e-05,
+      "loss": 0.0149,
+      "step": 122
+    },
+    {
+      "epoch": 1.1310344827586207,
+      "grad_norm": 0.01844456046819687,
+      "learning_rate": 9.51083712390519e-05,
+      "loss": 0.0137,
+      "step": 123
+    },
+    {
+      "epoch": 1.1402298850574712,
+      "grad_norm": 0.018193133175373077,
+      "learning_rate": 9.496800186095466e-05,
+      "loss": 0.0131,
+      "step": 124
+    },
+    {
+      "epoch": 1.1494252873563218,
+      "grad_norm": 0.01977015659213066,
+      "learning_rate": 9.482575347202047e-05,
+      "loss": 0.0145,
+      "step": 125
+    },
+    {
+      "epoch": 1.1494252873563218,
+      "eval_loss": 0.015383351594209671,
+      "eval_runtime": 7.6046,
+      "eval_samples_per_second": 6.575,
+      "eval_steps_per_second": 1.709,
+      "step": 125
+    },
+    {
+      "epoch": 1.1586206896551725,
+      "grad_norm": 0.016328683122992516,
+      "learning_rate": 9.468163201617062e-05,
+      "loss": 0.0133,
+      "step": 126
+    },
+    {
+      "epoch": 1.167816091954023,
+      "grad_norm": 0.0158162210136652,
+      "learning_rate": 9.453564351559348e-05,
+      "loss": 0.0145,
+      "step": 127
+    },
+    {
+      "epoch": 1.1770114942528735,
+      "grad_norm": 0.01766177825629711,
+      "learning_rate": 9.438779407049281e-05,
+      "loss": 0.0133,
+      "step": 128
+    },
+    {
+      "epoch": 1.186206896551724,
+      "grad_norm": 0.015718543902039528,
+      "learning_rate": 9.423808985883289e-05,
+      "loss": 0.013,
+      "step": 129
+    },
+    {
+      "epoch": 1.1954022988505748,
+      "grad_norm": 0.01557285524904728,
+      "learning_rate": 9.40865371360804e-05,
+      "loss": 0.0125,
+      "step": 130
+    },
+    {
+      "epoch": 1.1954022988505748,
+      "eval_loss": 0.014952810481190681,
+      "eval_runtime": 7.5808,
+      "eval_samples_per_second": 6.596,
+      "eval_steps_per_second": 1.715,
+      "step": 130
+    },
+    {
+      "epoch": 1.2045977011494253,
+      "grad_norm": 0.014815889298915863,
+      "learning_rate": 9.393314223494296e-05,
+      "loss": 0.0111,
+      "step": 131
+    },
+    {
+      "epoch": 1.2137931034482758,
+      "grad_norm": 0.01469665952026844,
+      "learning_rate": 9.377791156510455e-05,
+      "loss": 0.0132,
+      "step": 132
+    },
+    {
+      "epoch": 1.2229885057471264,
+      "grad_norm": 0.014182367362082005,
+      "learning_rate": 9.362085161295769e-05,
+      "loss": 0.0119,
+      "step": 133
+    },
+    {
+      "epoch": 1.232183908045977,
+      "grad_norm": 0.016701558604836464,
+      "learning_rate": 9.346196894133239e-05,
+      "loss": 0.0119,
+      "step": 134
+    },
+    {
+      "epoch": 1.2413793103448276,
+      "grad_norm": 0.017275458201766014,
+      "learning_rate": 9.330127018922194e-05,
+      "loss": 0.0122,
+      "step": 135
+    },
+    {
+      "epoch": 1.2413793103448276,
+      "eval_loss": 0.014751402661204338,
+      "eval_runtime": 7.5882,
+      "eval_samples_per_second": 6.589,
+      "eval_steps_per_second": 1.713,
+      "step": 135
+    },
+    {
+      "epoch": 1.2505747126436781,
+      "grad_norm": 0.014996697194874287,
+      "learning_rate": 9.313876207150543e-05,
+      "loss": 0.0128,
+      "step": 136
+    },
+    {
+      "epoch": 1.2597701149425287,
+      "grad_norm": 0.01727742701768875,
+      "learning_rate": 9.297445137866727e-05,
+      "loss": 0.0119,
+      "step": 137
+    },
+    {
+      "epoch": 1.2689655172413792,
+      "grad_norm": 0.017734119668602943,
+      "learning_rate": 9.280834497651334e-05,
+      "loss": 0.0133,
+      "step": 138
+    },
+    {
+      "epoch": 1.27816091954023,
+      "grad_norm": 0.017755163833498955,
+      "learning_rate": 9.264044980588416e-05,
+      "loss": 0.0129,
+      "step": 139
+    },
+    {
+      "epoch": 1.2873563218390804,
+      "grad_norm": 0.015634773299098015,
+      "learning_rate": 9.247077288236488e-05,
+      "loss": 0.0127,
+      "step": 140
+    },
+    {
+      "epoch": 1.2873563218390804,
+      "eval_loss": 0.014655789360404015,
+      "eval_runtime": 7.5803,
+      "eval_samples_per_second": 6.596,
+      "eval_steps_per_second": 1.715,
+      "step": 140
+    },
+    {
+      "epoch": 1.296551724137931,
+      "grad_norm": 0.014848074875772,
+      "learning_rate": 9.229932129599205e-05,
+      "loss": 0.0126,
+      "step": 141
+    },
+    {
+      "epoch": 1.3057471264367817,
+      "grad_norm": 0.015292002819478512,
+      "learning_rate": 9.212610221095748e-05,
+      "loss": 0.0126,
+      "step": 142
+    },
+    {
+      "epoch": 1.3149425287356322,
+      "grad_norm": 0.01488974317908287,
+      "learning_rate": 9.195112286530873e-05,
+      "loss": 0.0113,
+      "step": 143
+    },
+    {
+      "epoch": 1.3241379310344827,
+      "grad_norm": 0.015857763588428497,
+      "learning_rate": 9.177439057064683e-05,
+      "loss": 0.0118,
+      "step": 144
+    },
+    {
+      "epoch": 1.3333333333333333,
+      "grad_norm": 0.01526150293648243,
+      "learning_rate": 9.159591271182058e-05,
+      "loss": 0.0139,
+      "step": 145
+    },
+    {
+      "epoch": 1.3333333333333333,
+      "eval_loss": 0.014428853057324886,
+      "eval_runtime": 7.5934,
+      "eval_samples_per_second": 6.585,
+      "eval_steps_per_second": 1.712,
+      "step": 145
+    },
+    {
+      "epoch": 1.3425287356321838,
+      "grad_norm": 0.015166752971708775,
+      "learning_rate": 9.141569674661817e-05,
+      "loss": 0.0122,
+      "step": 146
+    },
+    {
+      "epoch": 1.3517241379310345,
+      "grad_norm": 0.014061689376831055,
+      "learning_rate": 9.123375020545535e-05,
+      "loss": 0.0123,
+      "step": 147
+    },
+    {
+      "epoch": 1.360919540229885,
+      "grad_norm": 0.015425407327711582,
+      "learning_rate": 9.105008069106093e-05,
+      "loss": 0.0117,
+      "step": 148
+    },
+    {
+      "epoch": 1.3701149425287356,
+      "grad_norm": 0.018539879471063614,
+      "learning_rate": 9.086469587815904e-05,
+      "loss": 0.0123,
+      "step": 149
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 0.014866286888718605,
+      "learning_rate": 9.067760351314838e-05,
+      "loss": 0.0122,
+      "step": 150
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "eval_loss": 0.014377254992723465,
+      "eval_runtime": 7.6093,
+      "eval_samples_per_second": 6.571,
+      "eval_steps_per_second": 1.708,
+      "step": 150
+    },
+    {
+      "epoch": 1.3885057471264368,
+      "grad_norm": 0.014935806393623352,
+      "learning_rate": 9.048881141377863e-05,
+      "loss": 0.0126,
+      "step": 151
+    },
+    {
+      "epoch": 1.3977011494252873,
+      "grad_norm": 0.017417676746845245,
+      "learning_rate": 9.029832746882371e-05,
+      "loss": 0.013,
+      "step": 152
+    },
+    {
+      "epoch": 1.4068965517241379,
+      "grad_norm": 0.015856370329856873,
+      "learning_rate": 9.01061596377522e-05,
+      "loss": 0.0114,
+      "step": 153
+    },
+    {
+      "epoch": 1.4160919540229884,
+      "grad_norm": 0.015301518142223358,
+      "learning_rate": 8.991231595039465e-05,
+      "loss": 0.0136,
+      "step": 154
+    },
+    {
+      "epoch": 1.4252873563218391,
+      "grad_norm": 0.017993515357375145,
+      "learning_rate": 8.97168045066082e-05,
+      "loss": 0.0138,
+      "step": 155
+    },
+    {
+      "epoch": 1.4252873563218391,
+      "eval_loss": 0.013892178423702717,
+      "eval_runtime": 7.5856,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 155
+    },
+    {
+      "epoch": 1.4344827586206896,
+      "grad_norm": 0.014389860443770885,
+      "learning_rate": 8.951963347593797e-05,
+      "loss": 0.0118,
+      "step": 156
+    },
+    {
+      "epoch": 1.4436781609195402,
+      "grad_norm": 0.013852115720510483,
+      "learning_rate": 8.932081109727582e-05,
+      "loss": 0.011,
+      "step": 157
+    },
+    {
+      "epoch": 1.452873563218391,
+      "grad_norm": 0.01491858996450901,
+      "learning_rate": 8.912034567851599e-05,
+      "loss": 0.0115,
+      "step": 158
+    },
+    {
+      "epoch": 1.4620689655172414,
+      "grad_norm": 0.014816755428910255,
+      "learning_rate": 8.891824559620801e-05,
+      "loss": 0.0115,
+      "step": 159
+    },
+    {
+      "epoch": 1.471264367816092,
+      "grad_norm": 0.017679313197731972,
+      "learning_rate": 8.871451929520663e-05,
+      "loss": 0.0143,
+      "step": 160
+    },
+    {
+      "epoch": 1.471264367816092,
+      "eval_loss": 0.013947720639407635,
+      "eval_runtime": 7.588,
+      "eval_samples_per_second": 6.589,
+      "eval_steps_per_second": 1.713,
+      "step": 160
+    },
+    {
+      "epoch": 1.4804597701149425,
+      "grad_norm": 0.015570229850709438,
+      "learning_rate": 8.850917528831899e-05,
+      "loss": 0.0131,
+      "step": 161
+    },
+    {
+      "epoch": 1.489655172413793,
+      "grad_norm": 0.014714114367961884,
+      "learning_rate": 8.83022221559489e-05,
+      "loss": 0.0109,
+      "step": 162
+    },
+    {
+      "epoch": 1.4988505747126437,
+      "grad_norm": 0.015797005966305733,
+      "learning_rate": 8.809366854573831e-05,
+      "loss": 0.0117,
+      "step": 163
+    },
+    {
+      "epoch": 1.5080459770114942,
+      "grad_norm": 0.01803259737789631,
+      "learning_rate": 8.78835231722059e-05,
+      "loss": 0.013,
+      "step": 164
+    },
+    {
+      "epoch": 1.5172413793103448,
+      "grad_norm": 0.015303310938179493,
+      "learning_rate": 8.767179481638303e-05,
+      "loss": 0.0124,
+      "step": 165
+    },
+    {
+      "epoch": 1.5172413793103448,
+      "eval_loss": 0.013795728795230389,
+      "eval_runtime": 7.5973,
+      "eval_samples_per_second": 6.581,
+      "eval_steps_per_second": 1.711,
+      "step": 165
+    },
+    {
+      "epoch": 1.5264367816091955,
+      "grad_norm": 0.01512802392244339,
+      "learning_rate": 8.745849232544681e-05,
+      "loss": 0.0113,
+      "step": 166
+    },
+    {
+      "epoch": 1.535632183908046,
+      "grad_norm": 0.01602269895374775,
+      "learning_rate": 8.724362461235029e-05,
+      "loss": 0.0137,
+      "step": 167
+    },
+    {
+      "epoch": 1.5448275862068965,
+      "grad_norm": 0.015303592197597027,
+      "learning_rate": 8.702720065545024e-05,
+      "loss": 0.0117,
+      "step": 168
+    },
+    {
+      "epoch": 1.5540229885057473,
+      "grad_norm": 0.014865975826978683,
+      "learning_rate": 8.680922949813178e-05,
+      "loss": 0.0122,
+      "step": 169
+    },
+    {
+      "epoch": 1.5632183908045976,
+      "grad_norm": 0.01662212237715721,
+      "learning_rate": 8.658972024843062e-05,
+      "loss": 0.0124,
+      "step": 170
+    },
+    {
+      "epoch": 1.5632183908045976,
+      "eval_loss": 0.013545939698815346,
+      "eval_runtime": 7.6043,
+      "eval_samples_per_second": 6.575,
+      "eval_steps_per_second": 1.71,
+      "step": 170
+    },
+    {
+      "epoch": 1.5724137931034483,
+      "grad_norm": 0.016805054619908333,
+      "learning_rate": 8.636868207865244e-05,
+      "loss": 0.0126,
+      "step": 171
+    },
+    {
+      "epoch": 1.5816091954022988,
+      "grad_norm": 0.015283631160855293,
+      "learning_rate": 8.614612422498964e-05,
+      "loss": 0.0129,
+      "step": 172
+    },
+    {
+      "epoch": 1.5908045977011493,
+      "grad_norm": 0.014891255646944046,
+      "learning_rate": 8.592205598713539e-05,
+      "loss": 0.0133,
+      "step": 173
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 0.01469021663069725,
+      "learning_rate": 8.569648672789497e-05,
+      "loss": 0.0113,
+      "step": 174
+    },
+    {
+      "epoch": 1.6091954022988506,
+      "grad_norm": 0.02000586874783039,
+      "learning_rate": 8.546942587279465e-05,
+      "loss": 0.0138,
+      "step": 175
+    },
+    {
+      "epoch": 1.6091954022988506,
+      "eval_loss": 0.01319777499884367,
+      "eval_runtime": 7.6152,
+      "eval_samples_per_second": 6.566,
+      "eval_steps_per_second": 1.707,
+      "step": 175
+    },
+    {
+      "epoch": 1.6183908045977011,
+      "grad_norm": 0.018022779375314713,
+      "learning_rate": 8.524088290968781e-05,
+      "loss": 0.0127,
+      "step": 176
+    },
+    {
+      "epoch": 1.6275862068965519,
+      "grad_norm": 0.01829381100833416,
+      "learning_rate": 8.501086738835843e-05,
+      "loss": 0.0131,
+      "step": 177
+    },
+    {
+      "epoch": 1.6367816091954022,
+      "grad_norm": 0.016279663890600204,
+      "learning_rate": 8.47793889201221e-05,
+      "loss": 0.0118,
+      "step": 178
+    },
+    {
+      "epoch": 1.645977011494253,
+      "grad_norm": 0.018548633903265,
+      "learning_rate": 8.45464571774244e-05,
+      "loss": 0.0149,
+      "step": 179
+    },
+    {
+      "epoch": 1.6551724137931034,
+      "grad_norm": 0.016604119911789894,
+      "learning_rate": 8.43120818934367e-05,
+      "loss": 0.0112,
+      "step": 180
+    },
+    {
+      "epoch": 1.6551724137931034,
+      "eval_loss": 0.013563334941864014,
+      "eval_runtime": 7.6136,
+      "eval_samples_per_second": 6.567,
+      "eval_steps_per_second": 1.707,
+      "step": 180
+    },
+    {
+      "epoch": 1.664367816091954,
+      "grad_norm": 0.016890155151486397,
+      "learning_rate": 8.407627286164948e-05,
+      "loss": 0.0121,
+      "step": 181
+    },
+    {
+      "epoch": 1.6735632183908047,
+      "grad_norm": 0.014763073064386845,
+      "learning_rate": 8.383903993546311e-05,
+      "loss": 0.0118,
+      "step": 182
+    },
+    {
+      "epoch": 1.6827586206896552,
+      "grad_norm": 0.016909824684262276,
+      "learning_rate": 8.360039302777612e-05,
+      "loss": 0.0123,
+      "step": 183
+    },
+    {
+      "epoch": 1.6919540229885057,
+      "grad_norm": 0.015340789221227169,
+      "learning_rate": 8.336034211057098e-05,
+      "loss": 0.0124,
+      "step": 184
+    },
+    {
+      "epoch": 1.7011494252873565,
+      "grad_norm": 0.01562805287539959,
+      "learning_rate": 8.31188972144974e-05,
+      "loss": 0.0102,
+      "step": 185
+    },
+    {
+      "epoch": 1.7011494252873565,
+      "eval_loss": 0.013533576391637325,
+      "eval_runtime": 7.5972,
+      "eval_samples_per_second": 6.581,
+      "eval_steps_per_second": 1.711,
+      "step": 185
+    },
+    {
+      "epoch": 1.7103448275862068,
+      "grad_norm": 0.0161032285541296,
+      "learning_rate": 8.28760684284532e-05,
+      "loss": 0.0104,
+      "step": 186
+    },
+    {
+      "epoch": 1.7195402298850575,
+      "grad_norm": 0.017128756269812584,
+      "learning_rate": 8.263186589916273e-05,
+      "loss": 0.0129,
+      "step": 187
+    },
+    {
+      "epoch": 1.728735632183908,
+      "grad_norm": 0.018114803358912468,
+      "learning_rate": 8.238629983075294e-05,
+      "loss": 0.0124,
+      "step": 188
+    },
+    {
+      "epoch": 1.7379310344827585,
+      "grad_norm": 0.019191117957234383,
+      "learning_rate": 8.213938048432697e-05,
+      "loss": 0.0135,
+      "step": 189
+    },
+    {
+      "epoch": 1.7471264367816093,
+      "grad_norm": 0.01860605925321579,
+      "learning_rate": 8.18911181775353e-05,
+      "loss": 0.0135,
+      "step": 190
+    },
+    {
+      "epoch": 1.7471264367816093,
+      "eval_loss": 0.013264240697026253,
+      "eval_runtime": 7.5732,
+      "eval_samples_per_second": 6.602,
+      "eval_steps_per_second": 1.717,
+      "step": 190
+    },
+    {
+      "epoch": 1.7563218390804598,
+      "grad_norm": 0.016007546335458755,
+      "learning_rate": 8.164152328414476e-05,
+      "loss": 0.0118,
+      "step": 191
+    },
+    {
+      "epoch": 1.7655172413793103,
+      "grad_norm": 0.017371613532304764,
+      "learning_rate": 8.139060623360493e-05,
+      "loss": 0.0126,
+      "step": 192
+    },
+    {
+      "epoch": 1.774712643678161,
+      "grad_norm": 0.017206458374857903,
+      "learning_rate": 8.113837751061246e-05,
+      "loss": 0.0103,
+      "step": 193
+    },
+    {
+      "epoch": 1.7839080459770114,
+      "grad_norm": 0.016995050013065338,
+      "learning_rate": 8.088484765467286e-05,
+      "loss": 0.0111,
+      "step": 194
+    },
+    {
+      "epoch": 1.793103448275862,
+      "grad_norm": 0.013478915207087994,
+      "learning_rate": 8.063002725966015e-05,
+      "loss": 0.01,
+      "step": 195
+    },
+    {
+      "epoch": 1.793103448275862,
+      "eval_loss": 0.013534443452954292,
+      "eval_runtime": 7.5956,
+      "eval_samples_per_second": 6.583,
+      "eval_steps_per_second": 1.712,
+      "step": 195
+    },
+    {
+      "epoch": 1.8022988505747126,
+      "grad_norm": 0.021648980677127838,
+      "learning_rate": 8.037392697337418e-05,
+      "loss": 0.0143,
+      "step": 196
+    },
+    {
+      "epoch": 1.8114942528735631,
+      "grad_norm": 0.015355809591710567,
+      "learning_rate": 8.011655749709575e-05,
+      "loss": 0.011,
+      "step": 197
+    },
+    {
+      "epoch": 1.8206896551724139,
+      "grad_norm": 0.014543996192514896,
+      "learning_rate": 7.985792958513931e-05,
+      "loss": 0.0101,
+      "step": 198
+    },
+    {
+      "epoch": 1.8298850574712644,
+      "grad_norm": 0.01613461971282959,
+      "learning_rate": 7.95980540444038e-05,
+      "loss": 0.0124,
+      "step": 199
+    },
+    {
+      "epoch": 1.839080459770115,
+      "grad_norm": 0.017040664330124855,
+      "learning_rate": 7.93369417339209e-05,
+      "loss": 0.0115,
+      "step": 200
+    },
+    {
+      "epoch": 1.839080459770115,
+      "eval_loss": 0.01306515745818615,
+      "eval_runtime": 7.5974,
+      "eval_samples_per_second": 6.581,
+      "eval_steps_per_second": 1.711,
+      "step": 200
+    },
+    {
+      "epoch": 1.8482758620689657,
+      "grad_norm": 0.01773899607360363,
+      "learning_rate": 7.907460356440133e-05,
+      "loss": 0.0133,
+      "step": 201
+    },
+    {
+      "epoch": 1.857471264367816,
+      "grad_norm": 0.012051481753587723,
+      "learning_rate": 7.881105049777901e-05,
+      "loss": 0.0093,
+      "step": 202
+    },
+    {
+      "epoch": 1.8666666666666667,
+      "grad_norm": 0.017261862754821777,
+      "learning_rate": 7.854629354675291e-05,
+      "loss": 0.0125,
+      "step": 203
+    },
+    {
+      "epoch": 1.8758620689655172,
+      "grad_norm": 0.01634056121110916,
+      "learning_rate": 7.828034377432693e-05,
+      "loss": 0.0111,
+      "step": 204
+    },
+    {
+      "epoch": 1.8850574712643677,
+      "grad_norm": 0.016423719003796577,
+      "learning_rate": 7.801321229334764e-05,
+      "loss": 0.0113,
+      "step": 205
+    },
+    {
+      "epoch": 1.8850574712643677,
+      "eval_loss": 0.012708608992397785,
+      "eval_runtime": 7.5908,
+      "eval_samples_per_second": 6.587,
+      "eval_steps_per_second": 1.713,
+      "step": 205
+    },
+    {
+      "epoch": 1.8942528735632185,
+      "grad_norm": 0.015081810764968395,
+      "learning_rate": 7.774491026603985e-05,
+      "loss": 0.013,
+      "step": 206
+    },
+    {
+      "epoch": 1.903448275862069,
+      "grad_norm": 0.0127123286947608,
+      "learning_rate": 7.74754489035403e-05,
+      "loss": 0.0106,
+      "step": 207
+    },
+    {
+      "epoch": 1.9126436781609195,
+      "grad_norm": 0.016098791733384132,
+      "learning_rate": 7.720483946542914e-05,
+      "loss": 0.0112,
+      "step": 208
+    },
+    {
+      "epoch": 1.9218390804597703,
+      "grad_norm": 0.01502301823347807,
+      "learning_rate": 7.69330932592594e-05,
+      "loss": 0.0119,
+      "step": 209
+    },
+    {
+      "epoch": 1.9310344827586206,
+      "grad_norm": 0.016128163784742355,
+      "learning_rate": 7.666022164008457e-05,
+      "loss": 0.0107,
+      "step": 210
+    },
+    {
+      "epoch": 1.9310344827586206,
+      "eval_loss": 0.012785980477929115,
+      "eval_runtime": 7.5896,
+      "eval_samples_per_second": 6.588,
+      "eval_steps_per_second": 1.713,
+      "step": 210
+    },
+    {
+      "epoch": 1.9402298850574713,
+      "grad_norm": 0.014872719533741474,
+      "learning_rate": 7.63862360099841e-05,
+      "loss": 0.0106,
+      "step": 211
+    },
+    {
+      "epoch": 1.9494252873563218,
+      "grad_norm": 0.016305750235915184,
+      "learning_rate": 7.611114781758692e-05,
+      "loss": 0.0107,
+      "step": 212
+    },
+    {
+      "epoch": 1.9586206896551723,
+      "grad_norm": 0.014701612293720245,
+      "learning_rate": 7.583496855759316e-05,
+      "loss": 0.0124,
+      "step": 213
+    },
+    {
+      "epoch": 1.967816091954023,
+      "grad_norm": 0.0179608054459095,
+      "learning_rate": 7.555770977029367e-05,
+      "loss": 0.0115,
+      "step": 214
+    },
+    {
+      "epoch": 1.9770114942528736,
+      "grad_norm": 0.016583839431405067,
+      "learning_rate": 7.527938304108795e-05,
+      "loss": 0.0122,
+      "step": 215
+    },
+    {
+      "epoch": 1.9770114942528736,
+      "eval_loss": 0.01276768371462822,
+      "eval_runtime": 7.6299,
+      "eval_samples_per_second": 6.553,
+      "eval_steps_per_second": 1.704,
+      "step": 215
+    },
+    {
+      "epoch": 1.986206896551724,
+      "grad_norm": 0.014455633237957954,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.0109,
+      "step": 216
+    },
+    {
+      "epoch": 1.9954022988505749,
+      "grad_norm": 0.014771837741136551,
+      "learning_rate": 7.471957232119234e-05,
+      "loss": 0.0114,
+      "step": 217
+    },
+    {
+      "epoch": 2.004597701149425,
+      "grad_norm": 0.019322361797094345,
+      "learning_rate": 7.443811172247821e-05,
+      "loss": 0.0123,
+      "step": 218
+    },
+    {
+      "epoch": 2.013793103448276,
+      "grad_norm": 0.011849606409668922,
+      "learning_rate": 7.415562996483192e-05,
+      "loss": 0.0081,
+      "step": 219
+    },
+    {
+      "epoch": 2.0229885057471266,
+      "grad_norm": 0.01762513443827629,
+      "learning_rate": 7.387213885189746e-05,
+      "loss": 0.0099,
+      "step": 220
+    },
+    {
+      "epoch": 2.0229885057471266,
+      "eval_loss": 0.012783367186784744,
+      "eval_runtime": 7.6029,
+      "eval_samples_per_second": 6.576,
+      "eval_steps_per_second": 1.71,
+      "step": 220
+    },
+    {
+      "epoch": 2.032183908045977,
+      "grad_norm": 0.0132211335003376,
+      "learning_rate": 7.358765022949519e-05,
+      "loss": 0.0088,
+      "step": 221
+    },
+    {
+      "epoch": 2.0413793103448277,
+      "grad_norm": 0.01868424378335476,
+      "learning_rate": 7.330217598512695e-05,
+      "loss": 0.0112,
+      "step": 222
+    },
+    {
+      "epoch": 2.050574712643678,
+      "grad_norm": 0.016267532482743263,
+      "learning_rate": 7.30157280474793e-05,
+      "loss": 0.0092,
+      "step": 223
+    },
+    {
+      "epoch": 2.0597701149425287,
+      "grad_norm": 0.014060255140066147,
+      "learning_rate": 7.272831838592503e-05,
+      "loss": 0.0091,
+      "step": 224
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "grad_norm": 0.017770465463399887,
+      "learning_rate": 7.243995901002312e-05,
+      "loss": 0.0121,
+      "step": 225
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "eval_loss": 0.012876266613602638,
+      "eval_runtime": 7.5829,
+      "eval_samples_per_second": 6.594,
+      "eval_steps_per_second": 1.714,
+      "step": 225
+    },
+    {
+      "epoch": 2.0781609195402297,
+      "grad_norm": 0.01353057473897934,
+      "learning_rate": 7.215066196901676e-05,
+      "loss": 0.0095,
+      "step": 226
+    },
+    {
+      "epoch": 2.0873563218390805,
+      "grad_norm": 0.01702524535357952,
+      "learning_rate": 7.186043935133005e-05,
+      "loss": 0.0095,
+      "step": 227
+    },
+    {
+      "epoch": 2.0965517241379312,
+      "grad_norm": 0.019207211211323738,
+      "learning_rate": 7.156930328406268e-05,
+      "loss": 0.0099,
+      "step": 228
+    },
+    {
+      "epoch": 2.1057471264367815,
+      "grad_norm": 0.017192739993333817,
+      "learning_rate": 7.127726593248337e-05,
+      "loss": 0.0108,
+      "step": 229
+    },
+    {
+      "epoch": 2.1149425287356323,
+      "grad_norm": 0.01726214587688446,
+      "learning_rate": 7.098433949952146e-05,
+      "loss": 0.0103,
+      "step": 230
+    },
+    {
+      "epoch": 2.1149425287356323,
+      "eval_loss": 0.012764949351549149,
+      "eval_runtime": 7.5742,
+      "eval_samples_per_second": 6.601,
+      "eval_steps_per_second": 1.716,
+      "step": 230
+    },
+    {
+      "epoch": 2.1241379310344826,
+      "grad_norm": 0.016892077401280403,
+      "learning_rate": 7.069053622525696e-05,
+      "loss": 0.0095,
+      "step": 231
+    },
+    {
+      "epoch": 2.1333333333333333,
+      "grad_norm": 0.01737057976424694,
+      "learning_rate": 7.039586838640919e-05,
+      "loss": 0.0101,
+      "step": 232
+    },
+    {
+      "epoch": 2.142528735632184,
+      "grad_norm": 0.017956653609871864,
+      "learning_rate": 7.01003482958237e-05,
+      "loss": 0.009,
+      "step": 233
+    },
+    {
+      "epoch": 2.1517241379310343,
+      "grad_norm": 0.018064096570014954,
+      "learning_rate": 6.980398830195785e-05,
+      "loss": 0.0115,
+      "step": 234
+    },
+    {
+      "epoch": 2.160919540229885,
+      "grad_norm": 0.017505567520856857,
+      "learning_rate": 6.950680078836474e-05,
+      "loss": 0.01,
+      "step": 235
+    },
+    {
+      "epoch": 2.160919540229885,
+      "eval_loss": 0.012695999816060066,
+      "eval_runtime": 7.6048,
+      "eval_samples_per_second": 6.575,
+      "eval_steps_per_second": 1.709,
+      "step": 235
+    },
+    {
+      "epoch": 2.170114942528736,
+      "grad_norm": 0.01884928159415722,
+      "learning_rate": 6.920879817317589e-05,
+      "loss": 0.0108,
+      "step": 236
+    },
+    {
+      "epoch": 2.179310344827586,
+      "grad_norm": 0.01585654355585575,
+      "learning_rate": 6.890999290858214e-05,
+      "loss": 0.0089,
+      "step": 237
+    },
+    {
+      "epoch": 2.188505747126437,
+      "grad_norm": 0.015242693945765495,
+      "learning_rate": 6.861039748031351e-05,
+      "loss": 0.0094,
+      "step": 238
+    },
+    {
+      "epoch": 2.197701149425287,
+      "grad_norm": 0.01854494772851467,
+      "learning_rate": 6.83100244071174e-05,
+      "loss": 0.0099,
+      "step": 239
+    },
+    {
+      "epoch": 2.206896551724138,
+      "grad_norm": 0.016470277681946754,
+      "learning_rate": 6.800888624023553e-05,
+      "loss": 0.0089,
+      "step": 240
+    },
+    {
+      "epoch": 2.206896551724138,
+      "eval_loss": 0.01267631258815527,
+      "eval_runtime": 7.5818,
+      "eval_samples_per_second": 6.595,
+      "eval_steps_per_second": 1.715,
+      "step": 240
+    },
+    {
+      "epoch": 2.2160919540229886,
+      "grad_norm": 0.017559630796313286,
+      "learning_rate": 6.770699556287939e-05,
+      "loss": 0.0105,
+      "step": 241
+    },
+    {
+      "epoch": 2.225287356321839,
+      "grad_norm": 0.016117800027132034,
+      "learning_rate": 6.740436498970452e-05,
+      "loss": 0.0099,
+      "step": 242
+    },
+    {
+      "epoch": 2.2344827586206897,
+      "grad_norm": 0.01739688403904438,
+      "learning_rate": 6.710100716628344e-05,
+      "loss": 0.0099,
+      "step": 243
+    },
+    {
+      "epoch": 2.2436781609195404,
+      "grad_norm": 0.01701669581234455,
+      "learning_rate": 6.679693476857711e-05,
+      "loss": 0.0085,
+      "step": 244
+    },
+    {
+      "epoch": 2.2528735632183907,
+      "grad_norm": 0.017626378685235977,
+      "learning_rate": 6.649216050240539e-05,
+      "loss": 0.0089,
+      "step": 245
+    },
+    {
+      "epoch": 2.2528735632183907,
+      "eval_loss": 0.012675927020609379,
+      "eval_runtime": 7.5971,
+      "eval_samples_per_second": 6.581,
+      "eval_steps_per_second": 1.711,
+      "step": 245
+    },
+    {
+      "epoch": 2.2620689655172415,
+      "grad_norm": 0.0238157007843256,
+      "learning_rate": 6.618669710291606e-05,
+      "loss": 0.0092,
+      "step": 246
+    },
+    {
+      "epoch": 2.2712643678160918,
+      "grad_norm": 0.017967745661735535,
+      "learning_rate": 6.588055733405266e-05,
+      "loss": 0.0102,
+      "step": 247
+    },
+    {
+      "epoch": 2.2804597701149425,
+      "grad_norm": 0.015404468402266502,
+      "learning_rate": 6.557375398802123e-05,
+      "loss": 0.0095,
+      "step": 248
+    },
+    {
+      "epoch": 2.2896551724137932,
+      "grad_norm": 0.015446755103766918,
+      "learning_rate": 6.526629988475567e-05,
+      "loss": 0.0092,
+      "step": 249
+    },
+    {
+      "epoch": 2.2988505747126435,
+      "grad_norm": 0.016809450462460518,
+      "learning_rate": 6.495820787138209e-05,
+      "loss": 0.0105,
+      "step": 250
+    },
+    {
+      "epoch": 2.2988505747126435,
+      "eval_loss": 0.012465451844036579,
+      "eval_runtime": 7.6011,
+      "eval_samples_per_second": 6.578,
+      "eval_steps_per_second": 1.71,
+      "step": 250
+    },
+    {
+      "epoch": 2.3080459770114943,
+      "grad_norm": 0.015970921143889427,
+      "learning_rate": 6.464949082168204e-05,
+      "loss": 0.0093,
+      "step": 251
+    },
+    {
+      "epoch": 2.317241379310345,
+      "grad_norm": 0.017121300101280212,
+      "learning_rate": 6.434016163555452e-05,
+      "loss": 0.0102,
+      "step": 252
+    },
+    {
+      "epoch": 2.3264367816091953,
+      "grad_norm": 0.01726561225950718,
+      "learning_rate": 6.403023323847695e-05,
+      "loss": 0.0093,
+      "step": 253
+    },
+    {
+      "epoch": 2.335632183908046,
+      "grad_norm": 0.014097567647695541,
+      "learning_rate": 6.371971858096508e-05,
+      "loss": 0.0096,
+      "step": 254
+    },
+    {
+      "epoch": 2.344827586206897,
+      "grad_norm": 0.015593893826007843,
+      "learning_rate": 6.340863063803188e-05,
+      "loss": 0.0093,
+      "step": 255
+    },
+    {
+      "epoch": 2.344827586206897,
+      "eval_loss": 0.012436559423804283,
+      "eval_runtime": 7.5957,
+      "eval_samples_per_second": 6.583,
+      "eval_steps_per_second": 1.712,
+      "step": 255
+    },
+    {
+      "epoch": 2.354022988505747,
+      "grad_norm": 0.01714129000902176,
+      "learning_rate": 6.30969824086453e-05,
+      "loss": 0.0107,
+      "step": 256
+    },
+    {
+      "epoch": 2.363218390804598,
+      "grad_norm": 0.018002189695835114,
+      "learning_rate": 6.27847869151852e-05,
+      "loss": 0.0103,
+      "step": 257
+    },
+    {
+      "epoch": 2.372413793103448,
+      "grad_norm": 0.015969395637512207,
+      "learning_rate": 6.247205720289907e-05,
+      "loss": 0.0099,
+      "step": 258
+    },
+    {
+      "epoch": 2.381609195402299,
+      "grad_norm": 0.020000826567411423,
+      "learning_rate": 6.215880633935708e-05,
+      "loss": 0.0105,
+      "step": 259
+    },
+    {
+      "epoch": 2.3908045977011496,
+      "grad_norm": 0.0170020442456007,
+      "learning_rate": 6.184504741390596e-05,
+      "loss": 0.0097,
+      "step": 260
+    },
+    {
+      "epoch": 2.3908045977011496,
+      "eval_loss": 0.01255771517753601,
+      "eval_runtime": 7.5967,
+      "eval_samples_per_second": 6.582,
+      "eval_steps_per_second": 1.711,
+      "step": 260
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 0.018086500465869904,
+      "learning_rate": 6.153079353712201e-05,
+      "loss": 0.0116,
+      "step": 261
+    },
+    {
+      "epoch": 2.4091954022988507,
+      "grad_norm": 0.017995236441493034,
+      "learning_rate": 6.121605784026339e-05,
+      "loss": 0.0098,
+      "step": 262
+    },
+    {
+      "epoch": 2.418390804597701,
+      "grad_norm": 0.016703004017472267,
+      "learning_rate": 6.09008534747213e-05,
+      "loss": 0.0085,
+      "step": 263
+    },
+    {
+      "epoch": 2.4275862068965517,
+      "grad_norm": 0.01667882315814495,
+      "learning_rate": 6.058519361147055e-05,
+      "loss": 0.0096,
+      "step": 264
+    },
+    {
+      "epoch": 2.4367816091954024,
+      "grad_norm": 0.014494238421320915,
+      "learning_rate": 6.02690914405191e-05,
+      "loss": 0.0091,
+      "step": 265
+    },
+    {
+      "epoch": 2.4367816091954024,
+      "eval_loss": 0.0126058179885149,
+      "eval_runtime": 7.5992,
+      "eval_samples_per_second": 6.58,
+      "eval_steps_per_second": 1.711,
+      "step": 265
+    },
+    {
+      "epoch": 2.4459770114942527,
+      "grad_norm": 0.018936868757009506,
+      "learning_rate": 5.995256017035703e-05,
+      "loss": 0.0102,
+      "step": 266
+    },
+    {
+      "epoch": 2.4551724137931035,
+      "grad_norm": 0.016619672998785973,
+      "learning_rate": 5.963561302740449e-05,
+      "loss": 0.0091,
+      "step": 267
+    },
+    {
+      "epoch": 2.464367816091954,
+      "grad_norm": 0.01624908670783043,
+      "learning_rate": 5.9318263255459116e-05,
+      "loss": 0.0097,
+      "step": 268
+    },
+    {
+      "epoch": 2.4735632183908045,
+      "grad_norm": 0.015271931886672974,
+      "learning_rate": 5.900052411514257e-05,
+      "loss": 0.0095,
+      "step": 269
+    },
+    {
+      "epoch": 2.4827586206896552,
+      "grad_norm": 0.01610173098742962,
+      "learning_rate": 5.868240888334653e-05,
+      "loss": 0.0095,
+      "step": 270
+    },
+    {
+      "epoch": 2.4827586206896552,
+      "eval_loss": 0.01236729696393013,
+      "eval_runtime": 7.5958,
+      "eval_samples_per_second": 6.583,
+      "eval_steps_per_second": 1.711,
+      "step": 270
+    },
+    {
+      "epoch": 2.491954022988506,
+      "grad_norm": 0.013876480981707573,
+      "learning_rate": 5.836393085267776e-05,
+      "loss": 0.0091,
+      "step": 271
+    },
+    {
+      "epoch": 2.5011494252873563,
+      "grad_norm": 0.014515344053506851,
+      "learning_rate": 5.804510333090287e-05,
+      "loss": 0.0087,
+      "step": 272
+    },
+    {
+      "epoch": 2.510344827586207,
+      "grad_norm": 0.015616660006344318,
+      "learning_rate": 5.772593964039203e-05,
+      "loss": 0.0088,
+      "step": 273
+    },
+    {
+      "epoch": 2.5195402298850573,
+      "grad_norm": 0.016088545322418213,
+      "learning_rate": 5.740645311756245e-05,
+      "loss": 0.0095,
+      "step": 274
+    },
+    {
+      "epoch": 2.528735632183908,
+      "grad_norm": 0.014027421362698078,
+      "learning_rate": 5.708665711232103e-05,
+      "loss": 0.0094,
+      "step": 275
+    },
+    {
+      "epoch": 2.528735632183908,
+      "eval_loss": 0.012253012508153915,
+      "eval_runtime": 7.5717,
+      "eval_samples_per_second": 6.604,
+      "eval_steps_per_second": 1.717,
+      "step": 275
+    },
+    {
+      "epoch": 2.5379310344827584,
+      "grad_norm": 0.015265434980392456,
+      "learning_rate": 5.6766564987506566e-05,
+      "loss": 0.0102,
+      "step": 276
+    },
+    {
+      "epoch": 2.547126436781609,
+      "grad_norm": 0.01524539664387703,
+      "learning_rate": 5.644619011833133e-05,
+      "loss": 0.0089,
+      "step": 277
+    },
+    {
+      "epoch": 2.55632183908046,
+      "grad_norm": 0.01515972800552845,
+      "learning_rate": 5.6125545891822274e-05,
+      "loss": 0.009,
+      "step": 278
+    },
+    {
+      "epoch": 2.56551724137931,
+      "grad_norm": 0.017095813527703285,
+      "learning_rate": 5.5804645706261514e-05,
+      "loss": 0.0103,
+      "step": 279
+    },
+    {
+      "epoch": 2.574712643678161,
+      "grad_norm": 0.014950582757592201,
+      "learning_rate": 5.548350297062659e-05,
+      "loss": 0.0092,
+      "step": 280
+    },
+    {
+      "epoch": 2.574712643678161,
+      "eval_loss": 0.011931383982300758,
+      "eval_runtime": 7.6218,
+      "eval_samples_per_second": 6.56,
+      "eval_steps_per_second": 1.706,
+      "step": 280
+    },
+    {
+      "epoch": 2.5839080459770116,
+      "grad_norm": 0.015228749252855778,
+      "learning_rate": 5.516213110403009e-05,
+      "loss": 0.0093,
+      "step": 281
+    },
+    {
+      "epoch": 2.593103448275862,
+      "grad_norm": 0.020853353664278984,
+      "learning_rate": 5.484054353515896e-05,
+      "loss": 0.0116,
+      "step": 282
+    },
+    {
+      "epoch": 2.6022988505747127,
+      "grad_norm": 0.013651091605424881,
+      "learning_rate": 5.451875370171341e-05,
+      "loss": 0.0076,
+      "step": 283
+    },
+    {
+      "epoch": 2.6114942528735634,
+      "grad_norm": 0.015777967870235443,
+      "learning_rate": 5.419677504984534e-05,
+      "loss": 0.0088,
+      "step": 284
+    },
+    {
+      "epoch": 2.6206896551724137,
+      "grad_norm": 0.014214356429874897,
+      "learning_rate": 5.387462103359655e-05,
+      "loss": 0.0084,
+      "step": 285
+    },
+    {
+      "epoch": 2.6206896551724137,
+      "eval_loss": 0.012128188274800777,
+      "eval_runtime": 7.6207,
+      "eval_samples_per_second": 6.561,
+      "eval_steps_per_second": 1.706,
+      "step": 285
+    },
+    {
+      "epoch": 2.6298850574712644,
+      "grad_norm": 0.02143423818051815,
+      "learning_rate": 5.355230511433651e-05,
+      "loss": 0.0109,
+      "step": 286
+    },
+    {
+      "epoch": 2.639080459770115,
+      "grad_norm": 0.016306888312101364,
+      "learning_rate": 5.32298407601999e-05,
+      "loss": 0.009,
+      "step": 287
+    },
+    {
+      "epoch": 2.6482758620689655,
+      "grad_norm": 0.015416650101542473,
+      "learning_rate": 5.290724144552379e-05,
+      "loss": 0.0091,
+      "step": 288
+    },
+    {
+      "epoch": 2.657471264367816,
+      "grad_norm": 0.018417850136756897,
+      "learning_rate": 5.258452065028473e-05,
+      "loss": 0.0096,
+      "step": 289
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "grad_norm": 0.016621043905615807,
+      "learning_rate": 5.226169185953532e-05,
+      "loss": 0.0098,
+      "step": 290
+    },
+    {
+      "epoch": 2.6666666666666665,
+      "eval_loss": 0.01195099763572216,
+      "eval_runtime": 7.5972,
+      "eval_samples_per_second": 6.581,
+      "eval_steps_per_second": 1.711,
+      "step": 290
+    },
+    {
+      "epoch": 2.6758620689655173,
+      "grad_norm": 0.016165705397725105,
+      "learning_rate": 5.193876856284085e-05,
+      "loss": 0.0088,
+      "step": 291
+    },
+    {
+      "epoch": 2.6850574712643676,
+      "grad_norm": 0.015073074027895927,
+      "learning_rate": 5.1615764253715536e-05,
+      "loss": 0.0089,
+      "step": 292
+    },
+    {
+      "epoch": 2.6942528735632183,
+      "grad_norm": 0.019986115396022797,
+      "learning_rate": 5.129269242905882e-05,
+      "loss": 0.0104,
+      "step": 293
+    },
+    {
+      "epoch": 2.703448275862069,
+      "grad_norm": 0.01431217696517706,
+      "learning_rate": 5.096956658859122e-05,
+      "loss": 0.0087,
+      "step": 294
+    },
+    {
+      "epoch": 2.7126436781609193,
+      "grad_norm": 0.0174038615077734,
+      "learning_rate": 5.064640023429043e-05,
+      "loss": 0.0097,
+      "step": 295
+    },
+    {
+      "epoch": 2.7126436781609193,
+      "eval_loss": 0.012155573815107346,
+      "eval_runtime": 7.6146,
+      "eval_samples_per_second": 6.566,
+      "eval_steps_per_second": 1.707,
+      "step": 295
+    },
+    {
+      "epoch": 2.72183908045977,
+      "grad_norm": 0.01756209135055542,
+      "learning_rate": 5.0323206869826966e-05,
+      "loss": 0.0095,
+      "step": 296
+    },
+    {
+      "epoch": 2.731034482758621,
+      "grad_norm": 0.018789948895573616,
+      "learning_rate": 5e-05,
+      "loss": 0.0094,
+      "step": 297
+    },
+    {
+      "epoch": 2.740229885057471,
+      "grad_norm": 0.01564960367977619,
+      "learning_rate": 4.967679313017303e-05,
+      "loss": 0.0079,
+      "step": 298
+    },
+    {
+      "epoch": 2.749425287356322,
+      "grad_norm": 0.015516403131186962,
+      "learning_rate": 4.9353599765709584e-05,
+      "loss": 0.0088,
+      "step": 299
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "grad_norm": 0.017091600224375725,
+      "learning_rate": 4.903043341140879e-05,
+      "loss": 0.0093,
+      "step": 300
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "eval_loss": 0.012067421339452267,
+      "eval_runtime": 7.6195,
+      "eval_samples_per_second": 6.562,
+      "eval_steps_per_second": 1.706,
+      "step": 300
+    },
+    {
+      "epoch": 2.767816091954023,
+      "grad_norm": 0.020157935097813606,
+      "learning_rate": 4.870730757094121e-05,
+      "loss": 0.0099,
+      "step": 301
+    },
+    {
+      "epoch": 2.7770114942528736,
+      "grad_norm": 0.016472700983285904,
+      "learning_rate": 4.8384235746284476e-05,
+      "loss": 0.0077,
+      "step": 302
+    },
+    {
+      "epoch": 2.7862068965517244,
+      "grad_norm": 0.016428852453827858,
+      "learning_rate": 4.806123143715916e-05,
+      "loss": 0.0084,
+      "step": 303
+    },
+    {
+      "epoch": 2.7954022988505747,
+      "grad_norm": 0.017473004758358,
+      "learning_rate": 4.7738308140464685e-05,
+      "loss": 0.0089,
+      "step": 304
+    },
+    {
+      "epoch": 2.8045977011494254,
+      "grad_norm": 0.015140421688556671,
+      "learning_rate": 4.7415479349715275e-05,
+      "loss": 0.0096,
+      "step": 305
+    },
+    {
+      "epoch": 2.8045977011494254,
+      "eval_loss": 0.011938001029193401,
+      "eval_runtime": 7.5857,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 305
+    },
+    {
+      "epoch": 2.8137931034482757,
+      "grad_norm": 0.01662864163517952,
+      "learning_rate": 4.709275855447621e-05,
+      "loss": 0.0091,
+      "step": 306
+    },
+    {
+      "epoch": 2.8229885057471265,
+      "grad_norm": 0.019283581525087357,
+      "learning_rate": 4.677015923980011e-05,
+      "loss": 0.0101,
+      "step": 307
+    },
+    {
+      "epoch": 2.8321839080459768,
+      "grad_norm": 0.016029009595513344,
+      "learning_rate": 4.6447694885663514e-05,
+      "loss": 0.009,
+      "step": 308
+    },
+    {
+      "epoch": 2.8413793103448275,
+      "grad_norm": 0.020449141040444374,
+      "learning_rate": 4.612537896640346e-05,
+      "loss": 0.0089,
+      "step": 309
+    },
+    {
+      "epoch": 2.8505747126436782,
+      "grad_norm": 0.017856908962130547,
+      "learning_rate": 4.5803224950154656e-05,
+      "loss": 0.0097,
+      "step": 310
+    },
+    {
+      "epoch": 2.8505747126436782,
+      "eval_loss": 0.01165664754807949,
+      "eval_runtime": 7.5917,
+      "eval_samples_per_second": 6.586,
+      "eval_steps_per_second": 1.712,
+      "step": 310
+    },
+    {
+      "epoch": 2.8597701149425285,
+      "grad_norm": 0.014575740322470665,
+      "learning_rate": 4.54812462982866e-05,
+      "loss": 0.0082,
+      "step": 311
+    },
+    {
+      "epoch": 2.8689655172413793,
+      "grad_norm": 0.017482686787843704,
+      "learning_rate": 4.515945646484105e-05,
+      "loss": 0.0105,
+      "step": 312
+    },
+    {
+      "epoch": 2.87816091954023,
+      "grad_norm": 0.014415116980671883,
+      "learning_rate": 4.4837868895969936e-05,
+      "loss": 0.0085,
+      "step": 313
+    },
+    {
+      "epoch": 2.8873563218390803,
+      "grad_norm": 0.01374480128288269,
+      "learning_rate": 4.451649702937342e-05,
+      "loss": 0.008,
+      "step": 314
+    },
+    {
+      "epoch": 2.896551724137931,
+      "grad_norm": 0.015432983636856079,
+      "learning_rate": 4.4195354293738484e-05,
+      "loss": 0.0101,
+      "step": 315
+    },
+    {
+      "epoch": 2.896551724137931,
+      "eval_loss": 0.011835181154310703,
+      "eval_runtime": 7.5858,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 315
+    },
+    {
+      "epoch": 2.905747126436782,
+      "grad_norm": 0.01638283021748066,
+      "learning_rate": 4.387445410817774e-05,
+      "loss": 0.0091,
+      "step": 316
+    },
+    {
+      "epoch": 2.914942528735632,
+      "grad_norm": 0.016213780269026756,
+      "learning_rate": 4.355380988166867e-05,
+      "loss": 0.0085,
+      "step": 317
+    },
+    {
+      "epoch": 2.924137931034483,
+      "grad_norm": 0.012681382708251476,
+      "learning_rate": 4.323343501249346e-05,
+      "loss": 0.0089,
+      "step": 318
+    },
+    {
+      "epoch": 2.9333333333333336,
+      "grad_norm": 0.01257992722094059,
+      "learning_rate": 4.2913342887678985e-05,
+      "loss": 0.0088,
+      "step": 319
+    },
+    {
+      "epoch": 2.942528735632184,
+      "grad_norm": 0.014623278751969337,
+      "learning_rate": 4.259354688243757e-05,
+      "loss": 0.0088,
+      "step": 320
+    },
+    {
+      "epoch": 2.942528735632184,
+      "eval_loss": 0.011810057796537876,
+      "eval_runtime": 7.5985,
+      "eval_samples_per_second": 6.58,
+      "eval_steps_per_second": 1.711,
+      "step": 320
+    },
+    {
+      "epoch": 2.9517241379310346,
+      "grad_norm": 0.016209494322538376,
+      "learning_rate": 4.227406035960798e-05,
+      "loss": 0.0093,
+      "step": 321
+    },
+    {
+      "epoch": 2.960919540229885,
+      "grad_norm": 0.013485006988048553,
+      "learning_rate": 4.195489666909713e-05,
+      "loss": 0.0081,
+      "step": 322
+    },
+    {
+      "epoch": 2.9701149425287356,
+      "grad_norm": 0.012734117917716503,
+      "learning_rate": 4.1636069147322246e-05,
+      "loss": 0.0072,
+      "step": 323
+    },
+    {
+      "epoch": 2.979310344827586,
+      "grad_norm": 0.02069373056292534,
+      "learning_rate": 4.131759111665349e-05,
+      "loss": 0.0094,
+      "step": 324
+    },
+    {
+      "epoch": 2.9885057471264367,
+      "grad_norm": 0.018109718337655067,
+      "learning_rate": 4.099947588485744e-05,
+      "loss": 0.0096,
+      "step": 325
+    },
+    {
+      "epoch": 2.9885057471264367,
+      "eval_loss": 0.011837408877909184,
+      "eval_runtime": 7.6092,
+      "eval_samples_per_second": 6.571,
+      "eval_steps_per_second": 1.708,
+      "step": 325
+    },
+    {
+      "epoch": 2.9977011494252874,
+      "grad_norm": 0.02106163091957569,
+      "learning_rate": 4.06817367445409e-05,
+      "loss": 0.0093,
+      "step": 326
+    },
+    {
+      "epoch": 3.0068965517241377,
+      "grad_norm": 0.013615073636174202,
+      "learning_rate": 4.036438697259551e-05,
+      "loss": 0.0072,
+      "step": 327
+    },
+    {
+      "epoch": 3.0160919540229885,
+      "grad_norm": 0.01178102008998394,
+      "learning_rate": 4.004743982964298e-05,
+      "loss": 0.0069,
+      "step": 328
+    },
+    {
+      "epoch": 3.025287356321839,
+      "grad_norm": 0.01662798412144184,
+      "learning_rate": 3.97309085594809e-05,
+      "loss": 0.0095,
+      "step": 329
+    },
+    {
+      "epoch": 3.0344827586206895,
+      "grad_norm": 0.016467638313770294,
+      "learning_rate": 3.941480638852948e-05,
+      "loss": 0.0078,
+      "step": 330
+    },
+    {
+      "epoch": 3.0344827586206895,
+      "eval_loss": 0.011860949918627739,
+      "eval_runtime": 7.5804,
+      "eval_samples_per_second": 6.596,
+      "eval_steps_per_second": 1.715,
+      "step": 330
+    },
+    {
+      "epoch": 3.0436781609195402,
+      "grad_norm": 0.01591646671295166,
+      "learning_rate": 3.909914652527871e-05,
+      "loss": 0.0073,
+      "step": 331
+    },
+    {
+      "epoch": 3.052873563218391,
+      "grad_norm": 0.01757364347577095,
+      "learning_rate": 3.878394215973663e-05,
+      "loss": 0.0074,
+      "step": 332
+    },
+    {
+      "epoch": 3.0620689655172413,
+      "grad_norm": 0.017451900988817215,
+      "learning_rate": 3.846920646287799e-05,
+      "loss": 0.0077,
+      "step": 333
+    },
+    {
+      "epoch": 3.071264367816092,
+      "grad_norm": 0.01796400360763073,
+      "learning_rate": 3.815495258609404e-05,
+      "loss": 0.0077,
+      "step": 334
+    },
+    {
+      "epoch": 3.0804597701149423,
+      "grad_norm": 0.01757225953042507,
+      "learning_rate": 3.784119366064293e-05,
+      "loss": 0.0064,
+      "step": 335
+    },
+    {
+      "epoch": 3.0804597701149423,
+      "eval_loss": 0.01191374659538269,
+      "eval_runtime": 7.5955,
+      "eval_samples_per_second": 6.583,
+      "eval_steps_per_second": 1.712,
+      "step": 335
+    },
+    {
+      "epoch": 3.089655172413793,
+      "grad_norm": 0.01832747273147106,
+      "learning_rate": 3.752794279710094e-05,
+      "loss": 0.007,
+      "step": 336
+    },
+    {
+      "epoch": 3.098850574712644,
+      "grad_norm": 0.018249882385134697,
+      "learning_rate": 3.721521308481482e-05,
+      "loss": 0.0083,
+      "step": 337
+    },
+    {
+      "epoch": 3.108045977011494,
+      "grad_norm": 0.017133980989456177,
+      "learning_rate": 3.6903017591354706e-05,
+      "loss": 0.0062,
+      "step": 338
+    },
+    {
+      "epoch": 3.117241379310345,
+      "grad_norm": 0.018890004605054855,
+      "learning_rate": 3.6591369361968124e-05,
+      "loss": 0.0068,
+      "step": 339
+    },
+    {
+      "epoch": 3.1264367816091956,
+      "grad_norm": 0.015423454344272614,
+      "learning_rate": 3.628028141903493e-05,
+      "loss": 0.0073,
+      "step": 340
+    },
+    {
+      "epoch": 3.1264367816091956,
+      "eval_loss": 0.01213847566395998,
+      "eval_runtime": 7.5868,
+      "eval_samples_per_second": 6.59,
+      "eval_steps_per_second": 1.713,
+      "step": 340
+    },
+    {
+      "epoch": 3.135632183908046,
+      "grad_norm": 0.015354766510426998,
+      "learning_rate": 3.596976676152306e-05,
+      "loss": 0.0074,
+      "step": 341
+    },
+    {
+      "epoch": 3.1448275862068966,
+      "grad_norm": 0.018978066742420197,
+      "learning_rate": 3.5659838364445505e-05,
+      "loss": 0.0071,
+      "step": 342
+    },
+    {
+      "epoch": 3.154022988505747,
+      "grad_norm": 0.021498536691069603,
+      "learning_rate": 3.535050917831797e-05,
+      "loss": 0.0072,
+      "step": 343
+    },
+    {
+      "epoch": 3.1632183908045977,
+      "grad_norm": 0.01662536896765232,
+      "learning_rate": 3.5041792128617927e-05,
+      "loss": 0.0063,
+      "step": 344
+    },
+    {
+      "epoch": 3.1724137931034484,
+      "grad_norm": 0.016221007332205772,
+      "learning_rate": 3.473370011524435e-05,
+      "loss": 0.0066,
+      "step": 345
+    },
+    {
+      "epoch": 3.1724137931034484,
+      "eval_loss": 0.012059099040925503,
+      "eval_runtime": 7.5892,
+      "eval_samples_per_second": 6.588,
+      "eval_steps_per_second": 1.713,
+      "step": 345
+    },
+    {
+      "epoch": 3.1816091954022987,
+      "grad_norm": 0.021475255489349365,
+      "learning_rate": 3.442624601197877e-05,
+      "loss": 0.0076,
+      "step": 346
+    },
+    {
+      "epoch": 3.1908045977011494,
+      "grad_norm": 0.015069660730659962,
+      "learning_rate": 3.4119442665947344e-05,
+      "loss": 0.0064,
+      "step": 347
+    },
+    {
+      "epoch": 3.2,
+      "grad_norm": 0.01803274266421795,
+      "learning_rate": 3.381330289708396e-05,
+      "loss": 0.0082,
+      "step": 348
+    },
+    {
+      "epoch": 3.2091954022988505,
+      "grad_norm": 0.01855016127228737,
+      "learning_rate": 3.350783949759462e-05,
+      "loss": 0.0072,
+      "step": 349
+    },
+    {
+      "epoch": 3.218390804597701,
+      "grad_norm": 0.017640260979533195,
+      "learning_rate": 3.3203065231422904e-05,
+      "loss": 0.0067,
+      "step": 350
+    },
+    {
+      "epoch": 3.218390804597701,
+      "eval_loss": 0.011746257543563843,
+      "eval_runtime": 7.5783,
+      "eval_samples_per_second": 6.598,
+      "eval_steps_per_second": 1.715,
+      "step": 350
+    },
+    {
+      "epoch": 3.227586206896552,
+      "grad_norm": 0.018065195530653,
+      "learning_rate": 3.289899283371657e-05,
+      "loss": 0.0082,
+      "step": 351
+    },
+    {
+      "epoch": 3.2367816091954023,
+      "grad_norm": 0.01586788520216942,
+      "learning_rate": 3.2595635010295475e-05,
+      "loss": 0.007,
+      "step": 352
+    },
+    {
+      "epoch": 3.245977011494253,
+      "grad_norm": 0.01758764311671257,
+      "learning_rate": 3.2293004437120624e-05,
+      "loss": 0.0073,
+      "step": 353
+    },
+    {
+      "epoch": 3.2551724137931033,
+      "grad_norm": 0.018066026270389557,
+      "learning_rate": 3.199111375976449e-05,
+      "loss": 0.0075,
+      "step": 354
+    },
+    {
+      "epoch": 3.264367816091954,
+      "grad_norm": 0.02090166136622429,
+      "learning_rate": 3.1689975592882603e-05,
+      "loss": 0.007,
+      "step": 355
+    },
+    {
+      "epoch": 3.264367816091954,
+      "eval_loss": 0.011750459671020508,
+      "eval_runtime": 7.5842,
+      "eval_samples_per_second": 6.593,
+      "eval_steps_per_second": 1.714,
+      "step": 355
+    },
+    {
+      "epoch": 3.2735632183908048,
+      "grad_norm": 0.0185533594340086,
+      "learning_rate": 3.1389602519686515e-05,
+      "loss": 0.0078,
+      "step": 356
+    },
+    {
+      "epoch": 3.282758620689655,
+      "grad_norm": 0.018230192363262177,
+      "learning_rate": 3.109000709141788e-05,
+      "loss": 0.007,
+      "step": 357
+    },
+    {
+      "epoch": 3.291954022988506,
+      "grad_norm": 0.016010766848921776,
+      "learning_rate": 3.079120182682412e-05,
+      "loss": 0.0061,
+      "step": 358
+    },
+    {
+      "epoch": 3.301149425287356,
+      "grad_norm": 0.019986698403954506,
+      "learning_rate": 3.049319921163526e-05,
+      "loss": 0.0085,
+      "step": 359
+    },
+    {
+      "epoch": 3.310344827586207,
+      "grad_norm": 0.01619451493024826,
+      "learning_rate": 3.019601169804216e-05,
+      "loss": 0.0072,
+      "step": 360
+    },
+    {
+      "epoch": 3.310344827586207,
+      "eval_loss": 0.011605262756347656,
+      "eval_runtime": 7.5872,
+      "eval_samples_per_second": 6.59,
+      "eval_steps_per_second": 1.713,
+      "step": 360
+    },
+    {
+      "epoch": 3.3195402298850576,
+      "grad_norm": 0.017471255734562874,
+      "learning_rate": 2.9899651704176325e-05,
+      "loss": 0.0075,
+      "step": 361
+    },
+    {
+      "epoch": 3.328735632183908,
+      "grad_norm": 0.015944065526127815,
+      "learning_rate": 2.9604131613590824e-05,
+      "loss": 0.0059,
+      "step": 362
+    },
+    {
+      "epoch": 3.3379310344827586,
+      "grad_norm": 0.02053840085864067,
+      "learning_rate": 2.9309463774743046e-05,
+      "loss": 0.0068,
+      "step": 363
+    },
+    {
+      "epoch": 3.3471264367816094,
+      "grad_norm": 0.022107843309640884,
+      "learning_rate": 2.901566050047855e-05,
+      "loss": 0.0076,
+      "step": 364
+    },
+    {
+      "epoch": 3.3563218390804597,
+      "grad_norm": 0.022166235372424126,
+      "learning_rate": 2.872273406751664e-05,
+      "loss": 0.0074,
+      "step": 365
+    },
+    {
+      "epoch": 3.3563218390804597,
+      "eval_loss": 0.011744478717446327,
+      "eval_runtime": 7.5789,
+      "eval_samples_per_second": 6.597,
+      "eval_steps_per_second": 1.715,
+      "step": 365
+    },
+    {
+      "epoch": 3.3655172413793104,
+      "grad_norm": 0.018682653084397316,
+      "learning_rate": 2.8430696715937337e-05,
+      "loss": 0.007,
+      "step": 366
+    },
+    {
+      "epoch": 3.374712643678161,
+      "grad_norm": 0.023426858708262444,
+      "learning_rate": 2.8139560648669962e-05,
+      "loss": 0.0062,
+      "step": 367
+    },
+    {
+      "epoch": 3.3839080459770114,
+      "grad_norm": 0.019427569583058357,
+      "learning_rate": 2.7849338030983257e-05,
+      "loss": 0.0065,
+      "step": 368
+    },
+    {
+      "epoch": 3.393103448275862,
+      "grad_norm": 0.015961861237883568,
+      "learning_rate": 2.7560040989976892e-05,
+      "loss": 0.0062,
+      "step": 369
+    },
+    {
+      "epoch": 3.4022988505747125,
+      "grad_norm": 0.018800493329763412,
+      "learning_rate": 2.7271681614074973e-05,
+      "loss": 0.0067,
+      "step": 370
+    },
+    {
+      "epoch": 3.4022988505747125,
+      "eval_loss": 0.011655151844024658,
+      "eval_runtime": 7.5992,
+      "eval_samples_per_second": 6.58,
+      "eval_steps_per_second": 1.711,
+      "step": 370
+    },
+    {
+      "epoch": 3.4114942528735632,
+      "grad_norm": 0.02263784408569336,
+      "learning_rate": 2.6984271952520722e-05,
+      "loss": 0.0074,
+      "step": 371
+    },
+    {
+      "epoch": 3.420689655172414,
+      "grad_norm": 0.01691795513033867,
+      "learning_rate": 2.6697824014873075e-05,
+      "loss": 0.0069,
+      "step": 372
+    },
+    {
+      "epoch": 3.4298850574712643,
+      "grad_norm": 0.016610443592071533,
+      "learning_rate": 2.641234977050484e-05,
+      "loss": 0.007,
+      "step": 373
+    },
+    {
+      "epoch": 3.439080459770115,
+      "grad_norm": 0.013923544436693192,
+      "learning_rate": 2.612786114810255e-05,
+      "loss": 0.006,
+      "step": 374
+    },
+    {
+      "epoch": 3.4482758620689653,
+      "grad_norm": 0.016392454504966736,
+      "learning_rate": 2.5844370035168073e-05,
+      "loss": 0.0072,
+      "step": 375
+    },
+    {
+      "epoch": 3.4482758620689653,
+      "eval_loss": 0.01167826447635889,
+      "eval_runtime": 7.5952,
+      "eval_samples_per_second": 6.583,
+      "eval_steps_per_second": 1.712,
+      "step": 375
+    },
+    {
+      "epoch": 3.457471264367816,
+      "grad_norm": 0.017350174486637115,
+      "learning_rate": 2.5561888277521794e-05,
+      "loss": 0.0073,
+      "step": 376
+    },
+    {
+      "epoch": 3.466666666666667,
+      "grad_norm": 0.016711924225091934,
+      "learning_rate": 2.528042767880766e-05,
+      "loss": 0.0063,
+      "step": 377
+    },
+    {
+      "epoch": 3.475862068965517,
+      "grad_norm": 0.019175494089722633,
+      "learning_rate": 2.500000000000001e-05,
+      "loss": 0.0073,
+      "step": 378
+    },
+    {
+      "epoch": 3.485057471264368,
+      "grad_norm": 0.018161507323384285,
+      "learning_rate": 2.4720616958912053e-05,
+      "loss": 0.0077,
+      "step": 379
+    },
+    {
+      "epoch": 3.4942528735632186,
+      "grad_norm": 0.017984988167881966,
+      "learning_rate": 2.4442290229706344e-05,
+      "loss": 0.0069,
+      "step": 380
+    },
+    {
+      "epoch": 3.4942528735632186,
+      "eval_loss": 0.011730669066309929,
+      "eval_runtime": 7.5885,
+      "eval_samples_per_second": 6.589,
+      "eval_steps_per_second": 1.713,
+      "step": 380
+    },
+    {
+      "epoch": 3.503448275862069,
+      "grad_norm": 0.016771584749221802,
+      "learning_rate": 2.4165031442406855e-05,
+      "loss": 0.0066,
+      "step": 381
+    },
+    {
+      "epoch": 3.5126436781609196,
+      "grad_norm": 0.016384167596697807,
+      "learning_rate": 2.3888852182413085e-05,
+      "loss": 0.0068,
+      "step": 382
+    },
+    {
+      "epoch": 3.5218390804597703,
+      "grad_norm": 0.016403399407863617,
+      "learning_rate": 2.361376399001592e-05,
+      "loss": 0.0069,
+      "step": 383
+    },
+    {
+      "epoch": 3.5310344827586206,
+      "grad_norm": 0.017848780378699303,
+      "learning_rate": 2.333977835991545e-05,
+      "loss": 0.0077,
+      "step": 384
+    },
+    {
+      "epoch": 3.5402298850574714,
+      "grad_norm": 0.02174249291419983,
+      "learning_rate": 2.3066906740740623e-05,
+      "loss": 0.0076,
+      "step": 385
+    },
+    {
+      "epoch": 3.5402298850574714,
+      "eval_loss": 0.011591249145567417,
+      "eval_runtime": 7.5857,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 385
+    },
+    {
+      "epoch": 3.5494252873563217,
+      "grad_norm": 0.018354855477809906,
+      "learning_rate": 2.2795160534570864e-05,
+      "loss": 0.0073,
+      "step": 386
+    },
+    {
+      "epoch": 3.5586206896551724,
+      "grad_norm": 0.01478376891463995,
+      "learning_rate": 2.25245510964597e-05,
+      "loss": 0.0067,
+      "step": 387
+    },
+    {
+      "epoch": 3.5678160919540227,
+      "grad_norm": 0.020804837346076965,
+      "learning_rate": 2.225508973396016e-05,
+      "loss": 0.0066,
+      "step": 388
+    },
+    {
+      "epoch": 3.5770114942528735,
+      "grad_norm": 0.017460566014051437,
+      "learning_rate": 2.198678770665238e-05,
+      "loss": 0.0062,
+      "step": 389
+    },
+    {
+      "epoch": 3.586206896551724,
+      "grad_norm": 0.01834852248430252,
+      "learning_rate": 2.171965622567308e-05,
+      "loss": 0.0068,
+      "step": 390
+    },
+    {
+      "epoch": 3.586206896551724,
+      "eval_loss": 0.011386090889573097,
+      "eval_runtime": 7.5843,
+      "eval_samples_per_second": 6.593,
+      "eval_steps_per_second": 1.714,
+      "step": 390
+    },
+    {
+      "epoch": 3.5954022988505745,
+      "grad_norm": 0.015800220891833305,
+      "learning_rate": 2.1453706453247087e-05,
+      "loss": 0.0064,
+      "step": 391
+    },
+    {
+      "epoch": 3.6045977011494252,
+      "grad_norm": 0.017407681792974472,
+      "learning_rate": 2.1188949502220983e-05,
+      "loss": 0.0073,
+      "step": 392
+    },
+    {
+      "epoch": 3.613793103448276,
+      "grad_norm": 0.01731077767908573,
+      "learning_rate": 2.0925396435598664e-05,
+      "loss": 0.0063,
+      "step": 393
+    },
+    {
+      "epoch": 3.6229885057471263,
+      "grad_norm": 0.01906469836831093,
+      "learning_rate": 2.066305826607911e-05,
+      "loss": 0.0077,
+      "step": 394
+    },
+    {
+      "epoch": 3.632183908045977,
+      "grad_norm": 0.017172574996948242,
+      "learning_rate": 2.0401945955596206e-05,
+      "loss": 0.0074,
+      "step": 395
+    },
+    {
+      "epoch": 3.632183908045977,
+      "eval_loss": 0.011461833491921425,
+      "eval_runtime": 7.6124,
+      "eval_samples_per_second": 6.568,
+      "eval_steps_per_second": 1.708,
+      "step": 395
+    },
+    {
+      "epoch": 3.6413793103448278,
+      "grad_norm": 0.019811954349279404,
+      "learning_rate": 2.0142070414860704e-05,
+      "loss": 0.0077,
+      "step": 396
+    },
+    {
+      "epoch": 3.650574712643678,
+      "grad_norm": 0.01807416044175625,
+      "learning_rate": 1.9883442502904283e-05,
+      "loss": 0.0075,
+      "step": 397
+    },
+    {
+      "epoch": 3.659770114942529,
+      "grad_norm": 0.021201424300670624,
+      "learning_rate": 1.9626073026625818e-05,
+      "loss": 0.0077,
+      "step": 398
+    },
+    {
+      "epoch": 3.6689655172413795,
+      "grad_norm": 0.01894940622150898,
+      "learning_rate": 1.936997274033986e-05,
+      "loss": 0.0062,
+      "step": 399
+    },
+    {
+      "epoch": 3.67816091954023,
+      "grad_norm": 0.016950421035289764,
+      "learning_rate": 1.9115152345327152e-05,
+      "loss": 0.0065,
+      "step": 400
+    },
+    {
+      "epoch": 3.67816091954023,
+      "eval_loss": 0.011437707580626011,
+      "eval_runtime": 7.6082,
+      "eval_samples_per_second": 6.572,
+      "eval_steps_per_second": 1.709,
+      "step": 400
+    },
+    {
+      "epoch": 3.6873563218390806,
+      "grad_norm": 0.01757856458425522,
+      "learning_rate": 1.8861622489387555e-05,
+      "loss": 0.0062,
+      "step": 401
+    },
+    {
+      "epoch": 3.696551724137931,
+      "grad_norm": 0.020201781764626503,
+      "learning_rate": 1.8609393766395085e-05,
+      "loss": 0.0079,
+      "step": 402
+    },
+    {
+      "epoch": 3.7057471264367816,
+      "grad_norm": 0.015953106805682182,
+      "learning_rate": 1.835847671585526e-05,
+      "loss": 0.0068,
+      "step": 403
+    },
+    {
+      "epoch": 3.714942528735632,
+      "grad_norm": 0.018131552264094353,
+      "learning_rate": 1.8108881822464696e-05,
+      "loss": 0.0066,
+      "step": 404
+    },
+    {
+      "epoch": 3.7241379310344827,
+      "grad_norm": 0.018745753914117813,
+      "learning_rate": 1.7860619515673033e-05,
+      "loss": 0.007,
+      "step": 405
+    },
+    {
+      "epoch": 3.7241379310344827,
+      "eval_loss": 0.011230139061808586,
+      "eval_runtime": 7.5892,
+      "eval_samples_per_second": 6.588,
+      "eval_steps_per_second": 1.713,
+      "step": 405
+    },
+    {
+      "epoch": 3.7333333333333334,
+      "grad_norm": 0.019100049510598183,
+      "learning_rate": 1.7613700169247056e-05,
+      "loss": 0.0078,
+      "step": 406
+    },
+    {
+      "epoch": 3.7425287356321837,
+      "grad_norm": 0.018397876992821693,
+      "learning_rate": 1.7368134100837287e-05,
+      "loss": 0.0077,
+      "step": 407
+    },
+    {
+      "epoch": 3.7517241379310344,
+      "grad_norm": 0.01846805214881897,
+      "learning_rate": 1.7123931571546827e-05,
+      "loss": 0.0073,
+      "step": 408
+    },
+    {
+      "epoch": 3.760919540229885,
+      "grad_norm": 0.019604753702878952,
+      "learning_rate": 1.6881102785502616e-05,
+      "loss": 0.0081,
+      "step": 409
+    },
+    {
+      "epoch": 3.7701149425287355,
+      "grad_norm": 0.019953811541199684,
+      "learning_rate": 1.6639657889429018e-05,
+      "loss": 0.0064,
+      "step": 410
+    },
+    {
+      "epoch": 3.7701149425287355,
+      "eval_loss": 0.01116408035159111,
+      "eval_runtime": 7.6015,
+      "eval_samples_per_second": 6.578,
+      "eval_steps_per_second": 1.71,
+      "step": 410
+    },
+    {
+      "epoch": 3.779310344827586,
+      "grad_norm": 0.018632952123880386,
+      "learning_rate": 1.639960697222388e-05,
+      "loss": 0.0063,
+      "step": 411
+    },
+    {
+      "epoch": 3.788505747126437,
+      "grad_norm": 0.018017446622252464,
+      "learning_rate": 1.6160960064536908e-05,
+      "loss": 0.0074,
+      "step": 412
+    },
+    {
+      "epoch": 3.7977011494252872,
+      "grad_norm": 0.016660140827298164,
+      "learning_rate": 1.592372713835055e-05,
+      "loss": 0.0068,
+      "step": 413
+    },
+    {
+      "epoch": 3.806896551724138,
+      "grad_norm": 0.016572022810578346,
+      "learning_rate": 1.5687918106563326e-05,
+      "loss": 0.007,
+      "step": 414
+    },
+    {
+      "epoch": 3.8160919540229887,
+      "grad_norm": 0.017351387068629265,
+      "learning_rate": 1.545354282257562e-05,
+      "loss": 0.0073,
+      "step": 415
+    },
+    {
+      "epoch": 3.8160919540229887,
+      "eval_loss": 0.011113017797470093,
+      "eval_runtime": 7.5996,
+      "eval_samples_per_second": 6.579,
+      "eval_steps_per_second": 1.711,
+      "step": 415
+    },
+    {
+      "epoch": 3.825287356321839,
+      "grad_norm": 0.017648274078965187,
+      "learning_rate": 1.52206110798779e-05,
+      "loss": 0.0058,
+      "step": 416
+    },
+    {
+      "epoch": 3.8344827586206898,
+      "grad_norm": 0.019093746319413185,
+      "learning_rate": 1.4989132611641576e-05,
+      "loss": 0.0071,
+      "step": 417
+    },
+    {
+      "epoch": 3.84367816091954,
+      "grad_norm": 0.0174701064825058,
+      "learning_rate": 1.4759117090312197e-05,
+      "loss": 0.0061,
+      "step": 418
+    },
+    {
+      "epoch": 3.852873563218391,
+      "grad_norm": 0.01850154623389244,
+      "learning_rate": 1.453057412720536e-05,
+      "loss": 0.0067,
+      "step": 419
+    },
+    {
+      "epoch": 3.862068965517241,
+      "grad_norm": 0.01711280085146427,
+      "learning_rate": 1.4303513272105057e-05,
+      "loss": 0.0065,
+      "step": 420
+    },
+    {
+      "epoch": 3.862068965517241,
+      "eval_loss": 0.011265527456998825,
+      "eval_runtime": 7.5976,
+      "eval_samples_per_second": 6.581,
+      "eval_steps_per_second": 1.711,
+      "step": 420
+    },
+    {
+      "epoch": 3.871264367816092,
+      "grad_norm": 0.019471535459160805,
+      "learning_rate": 1.4077944012864636e-05,
+      "loss": 0.0072,
+      "step": 421
+    },
+    {
+      "epoch": 3.8804597701149426,
+      "grad_norm": 0.020510006695985794,
+      "learning_rate": 1.3853875775010355e-05,
+      "loss": 0.0078,
+      "step": 422
+    },
+    {
+      "epoch": 3.889655172413793,
+      "grad_norm": 0.021030457690358162,
+      "learning_rate": 1.3631317921347563e-05,
+      "loss": 0.007,
+      "step": 423
+    },
+    {
+      "epoch": 3.8988505747126436,
+      "grad_norm": 0.01627511717379093,
+      "learning_rate": 1.3410279751569399e-05,
+      "loss": 0.0066,
+      "step": 424
+    },
+    {
+      "epoch": 3.9080459770114944,
+      "grad_norm": 0.018043728545308113,
+      "learning_rate": 1.3190770501868243e-05,
+      "loss": 0.0069,
+      "step": 425
+    },
+    {
+      "epoch": 3.9080459770114944,
+      "eval_loss": 0.011119170114398003,
+      "eval_runtime": 7.5908,
+      "eval_samples_per_second": 6.587,
+      "eval_steps_per_second": 1.713,
+      "step": 425
+    },
+    {
+      "epoch": 3.9172413793103447,
+      "grad_norm": 0.01714397221803665,
+      "learning_rate": 1.297279934454978e-05,
+      "loss": 0.007,
+      "step": 426
+    },
+    {
+      "epoch": 3.9264367816091954,
+      "grad_norm": 0.017846615985035896,
+      "learning_rate": 1.2756375387649716e-05,
+      "loss": 0.007,
+      "step": 427
+    },
+    {
+      "epoch": 3.935632183908046,
+      "grad_norm": 0.01910008303821087,
+      "learning_rate": 1.25415076745532e-05,
+      "loss": 0.0069,
+      "step": 428
+    },
+    {
+      "epoch": 3.9448275862068964,
+      "grad_norm": 0.015933765098452568,
+      "learning_rate": 1.2328205183616965e-05,
+      "loss": 0.0064,
+      "step": 429
+    },
+    {
+      "epoch": 3.954022988505747,
+      "grad_norm": 0.018724463880062103,
+      "learning_rate": 1.2116476827794104e-05,
+      "loss": 0.0065,
+      "step": 430
+    },
+    {
+      "epoch": 3.954022988505747,
+      "eval_loss": 0.01107843779027462,
+      "eval_runtime": 7.5853,
+      "eval_samples_per_second": 6.592,
+      "eval_steps_per_second": 1.714,
+      "step": 430
+    },
+    {
+      "epoch": 3.963218390804598,
+      "grad_norm": 0.018081670626997948,
+      "learning_rate": 1.1906331454261704e-05,
+      "loss": 0.006,
+      "step": 431
+    },
+    {
+      "epoch": 3.972413793103448,
+      "grad_norm": 0.023170167580246925,
+      "learning_rate": 1.1697777844051105e-05,
+      "loss": 0.0071,
+      "step": 432
+    },
+    {
+      "epoch": 3.981609195402299,
+      "grad_norm": 0.017531078308820724,
+      "learning_rate": 1.1490824711681025e-05,
+      "loss": 0.0068,
+      "step": 433
+    },
+    {
+      "epoch": 3.9908045977011493,
+      "grad_norm": 0.018324997276067734,
+      "learning_rate": 1.1285480704793377e-05,
+      "loss": 0.0068,
+      "step": 434
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.020008638501167297,
+      "learning_rate": 1.1081754403791999e-05,
+      "loss": 0.0076,
+      "step": 435
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.0111029502004385,
+      "eval_runtime": 7.5917,
+      "eval_samples_per_second": 6.586,
+      "eval_steps_per_second": 1.712,
+      "step": 435
+    },
+    {
+      "epoch": 4.00919540229885,
+      "grad_norm": 0.014917205087840557,
+      "learning_rate": 1.0879654321484012e-05,
+      "loss": 0.0051,
+      "step": 436
+    },
+    {
+      "epoch": 4.0183908045977015,
+      "grad_norm": 0.014704613946378231,
+      "learning_rate": 1.0679188902724191e-05,
+      "loss": 0.0051,
+      "step": 437
+    },
+    {
+      "epoch": 4.027586206896552,
+      "grad_norm": 0.01385421585291624,
+      "learning_rate": 1.0480366524062042e-05,
+      "loss": 0.0052,
+      "step": 438
+    },
+    {
+      "epoch": 4.036781609195402,
+      "grad_norm": 0.017096884548664093,
+      "learning_rate": 1.0283195493391823e-05,
+      "loss": 0.006,
+      "step": 439
+    },
+    {
+      "epoch": 4.045977011494253,
+      "grad_norm": 0.020865080878138542,
+      "learning_rate": 1.008768404960535e-05,
+      "loss": 0.0047,
+      "step": 440
+    },
+    {
+      "epoch": 4.045977011494253,
+      "eval_loss": 0.011515479534864426,
+      "eval_runtime": 7.5882,
+      "eval_samples_per_second": 6.589,
+      "eval_steps_per_second": 1.713,
+      "step": 440
+    },
+    {
+      "epoch": 4.055172413793104,
+      "grad_norm": 0.016257256269454956,
+      "learning_rate": 9.893840362247809e-06,
+      "loss": 0.0051,
+      "step": 441
+    },
+    {
+      "epoch": 4.064367816091954,
+      "grad_norm": 0.01711389049887657,
+      "learning_rate": 9.701672531176286e-06,
+      "loss": 0.0056,
+      "step": 442
+    },
+    {
+      "epoch": 4.073563218390804,
+      "grad_norm": 0.018976813182234764,
+      "learning_rate": 9.511188586221376e-06,
+      "loss": 0.0056,
+      "step": 443
+    },
+    {
+      "epoch": 4.082758620689655,
+      "grad_norm": 0.016858583316206932,
+      "learning_rate": 9.322396486851626e-06,
+      "loss": 0.0051,
+      "step": 444
+    },
+    {
+      "epoch": 4.091954022988506,
+      "grad_norm": 0.016142934560775757,
+      "learning_rate": 9.135304121840976e-06,
+      "loss": 0.0053,
+      "step": 445
+    },
+    {
+      "epoch": 4.091954022988506,
+      "eval_loss": 0.01191615965217352,
+      "eval_runtime": 7.5902,
+      "eval_samples_per_second": 6.587,
+      "eval_steps_per_second": 1.713,
+      "step": 445
+    },
+    {
+      "epoch": 4.101149425287356,
+      "grad_norm": 0.024709809571504593,
+      "learning_rate": 8.949919308939082e-06,
+      "loss": 0.0058,
+      "step": 446
+    },
+    {
+      "epoch": 4.110344827586207,
+      "grad_norm": 0.015532772988080978,
+      "learning_rate": 8.766249794544662e-06,
+      "loss": 0.0052,
+      "step": 447
+    },
+    {
+      "epoch": 4.119540229885057,
+      "grad_norm": 0.015922829508781433,
+      "learning_rate": 8.584303253381847e-06,
+      "loss": 0.0057,
+      "step": 448
+    },
+    {
+      "epoch": 4.128735632183908,
+      "grad_norm": 0.024093857035040855,
+      "learning_rate": 8.404087288179424e-06,
+      "loss": 0.0052,
+      "step": 449
+    },
+    {
+      "epoch": 4.137931034482759,
+      "grad_norm": 0.018626809120178223,
+      "learning_rate": 8.225609429353187e-06,
+      "loss": 0.0053,
+      "step": 450
+    },
+    {
+      "epoch": 4.137931034482759,
+      "eval_loss": 0.011974362656474113,
+      "eval_runtime": 7.6303,
+      "eval_samples_per_second": 6.553,
+      "eval_steps_per_second": 1.704,
+      "step": 450
+    },
+    {
+      "epoch": 4.147126436781609,
+      "grad_norm": 0.02421395108103752,
+      "learning_rate": 8.048877134691268e-06,
+      "loss": 0.0067,
+      "step": 451
+    },
+    {
+      "epoch": 4.1563218390804595,
+      "grad_norm": 0.019452223554253578,
+      "learning_rate": 7.873897789042523e-06,
+      "loss": 0.0055,
+      "step": 452
+    },
+    {
+      "epoch": 4.165517241379311,
+      "grad_norm": 0.01944682002067566,
+      "learning_rate": 7.700678704007947e-06,
+      "loss": 0.0051,
+      "step": 453
+    },
+    {
+      "epoch": 4.174712643678161,
+      "grad_norm": 0.0154283307492733,
+      "learning_rate": 7.529227117635135e-06,
+      "loss": 0.0053,
+      "step": 454
+    },
+    {
+      "epoch": 4.183908045977011,
+      "grad_norm": 0.02008649706840515,
+      "learning_rate": 7.35955019411585e-06,
+      "loss": 0.0055,
+      "step": 455
+    },
+    {
+      "epoch": 4.183908045977011,
+      "eval_loss": 0.01187676191329956,
+      "eval_runtime": 7.6617,
+      "eval_samples_per_second": 6.526,
+      "eval_steps_per_second": 1.697,
+      "step": 455
+    },
+    {
+      "epoch": 4.1931034482758625,
+      "grad_norm": 0.01813184656202793,
+      "learning_rate": 7.191655023486682e-06,
+      "loss": 0.0054,
+      "step": 456
+    },
+    {
+      "epoch": 4.202298850574713,
+      "grad_norm": 0.019368577748537064,
+      "learning_rate": 7.02554862133275e-06,
+      "loss": 0.0067,
+      "step": 457
+    },
+    {
+      "epoch": 4.211494252873563,
+      "grad_norm": 0.02604977786540985,
+      "learning_rate": 6.861237928494579e-06,
+      "loss": 0.0044,
+      "step": 458
+    },
+    {
+      "epoch": 4.220689655172414,
+      "grad_norm": 0.016761422157287598,
+      "learning_rate": 6.698729810778065e-06,
+      "loss": 0.0053,
+      "step": 459
+    },
+    {
+      "epoch": 4.2298850574712645,
+      "grad_norm": 0.015887631103396416,
+      "learning_rate": 6.53803105866761e-06,
+      "loss": 0.0053,
+      "step": 460
+    },
+    {
+      "epoch": 4.2298850574712645,
+      "eval_loss": 0.011668752878904343,
+      "eval_runtime": 7.5746,
+      "eval_samples_per_second": 6.601,
+      "eval_steps_per_second": 1.716,
+      "step": 460
+    },
+    {
+      "epoch": 4.239080459770115,
+      "grad_norm": 0.020124757662415504,
+      "learning_rate": 6.379148387042316e-06,
+      "loss": 0.0045,
+      "step": 461
+    },
+    {
+      "epoch": 4.248275862068965,
+      "grad_norm": 0.015533823519945145,
+      "learning_rate": 6.222088434895462e-06,
+      "loss": 0.0045,
+      "step": 462
+    },
+    {
+      "epoch": 4.257471264367816,
+      "grad_norm": 0.01983262039721012,
+      "learning_rate": 6.066857765057055e-06,
+      "loss": 0.0057,
+      "step": 463
+    },
+    {
+      "epoch": 4.266666666666667,
+      "grad_norm": 0.01795038767158985,
+      "learning_rate": 5.9134628639196e-06,
+      "loss": 0.0058,
+      "step": 464
+    },
+    {
+      "epoch": 4.275862068965517,
+      "grad_norm": 0.021382272243499756,
+      "learning_rate": 5.7619101411671095e-06,
+      "loss": 0.0053,
+      "step": 465
+    },
+    {
+      "epoch": 4.275862068965517,
+      "eval_loss": 0.011690051294863224,
+      "eval_runtime": 7.6002,
+      "eval_samples_per_second": 6.579,
+      "eval_steps_per_second": 1.71,
+      "step": 465
+    },
+    {
+      "epoch": 4.285057471264368,
+      "grad_norm": 0.019680311903357506,
+      "learning_rate": 5.6122059295072085e-06,
+      "loss": 0.0055,
+      "step": 466
+    },
+    {
+      "epoch": 4.294252873563218,
+      "grad_norm": 0.018579309806227684,
+      "learning_rate": 5.464356484406535e-06,
+      "loss": 0.0047,
+      "step": 467
+    },
+    {
+      "epoch": 4.303448275862069,
+      "grad_norm": 0.020488621667027473,
+      "learning_rate": 5.318367983829392e-06,
+      "loss": 0.0061,
+      "step": 468
+    },
+    {
+      "epoch": 4.31264367816092,
+      "grad_norm": 0.018219981342554092,
+      "learning_rate": 5.174246527979531e-06,
+      "loss": 0.005,
+      "step": 469
+    },
+    {
+      "epoch": 4.32183908045977,
+      "grad_norm": 0.01655331440269947,
+      "learning_rate": 5.031998139045352e-06,
+      "loss": 0.0053,
+      "step": 470
+    },
+    {
+      "epoch": 4.32183908045977,
+      "eval_loss": 0.011675745248794556,
+      "eval_runtime": 7.6014,
+      "eval_samples_per_second": 6.578,
+      "eval_steps_per_second": 1.71,
+      "step": 470
+    },
+    {
+      "epoch": 4.3310344827586205,
+      "grad_norm": 0.01707630045711994,
+      "learning_rate": 4.891628760948114e-06,
+      "loss": 0.0047,
+      "step": 471
+    },
+    {
+      "epoch": 4.340229885057472,
+      "grad_norm": 0.015081142075359821,
+      "learning_rate": 4.7531442590937335e-06,
+      "loss": 0.0054,
+      "step": 472
+    },
+    {
+      "epoch": 4.349425287356322,
+      "grad_norm": 0.021828968077898026,
+      "learning_rate": 4.616550420127563e-06,
+      "loss": 0.0052,
+      "step": 473
+    },
+    {
+      "epoch": 4.358620689655172,
+      "grad_norm": 0.01851847767829895,
+      "learning_rate": 4.4818529516926726e-06,
+      "loss": 0.006,
+      "step": 474
+    },
+    {
+      "epoch": 4.3678160919540225,
+      "grad_norm": 0.01676092855632305,
+      "learning_rate": 4.349057482191299e-06,
+      "loss": 0.0058,
+      "step": 475
+    },
+    {
+      "epoch": 4.3678160919540225,
+      "eval_loss": 0.011636043898761272,
+      "eval_runtime": 7.5917,
+      "eval_samples_per_second": 6.586,
+      "eval_steps_per_second": 1.712,
+      "step": 475
+    },
+    {
+      "epoch": 4.377011494252874,
+      "grad_norm": 0.018017876893281937,
+      "learning_rate": 4.218169560549706e-06,
+      "loss": 0.0055,
+      "step": 476
+    },
+    {
+      "epoch": 4.386206896551724,
+      "grad_norm": 0.018705466762185097,
+      "learning_rate": 4.089194655986306e-06,
+      "loss": 0.005,
+      "step": 477
+    },
+    {
+      "epoch": 4.395402298850574,
+      "grad_norm": 0.016500435769557953,
+      "learning_rate": 3.962138157783085e-06,
+      "loss": 0.0044,
+      "step": 478
+    },
+    {
+      "epoch": 4.4045977011494255,
+      "grad_norm": 0.016678597778081894,
+      "learning_rate": 3.837005375060482e-06,
+      "loss": 0.0049,
+      "step": 479
+    },
+    {
+      "epoch": 4.413793103448276,
+      "grad_norm": 0.01764843426644802,
+      "learning_rate": 3.7138015365554833e-06,
+      "loss": 0.0053,
+      "step": 480
+    },
+    {
+      "epoch": 4.413793103448276,
+      "eval_loss": 0.011646818369626999,
+      "eval_runtime": 7.5865,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 480
+    },
+    {
+      "epoch": 4.422988505747126,
+      "grad_norm": 0.018464326858520508,
+      "learning_rate": 3.5925317904031587e-06,
+      "loss": 0.0046,
+      "step": 481
+    },
+    {
+      "epoch": 4.432183908045977,
+      "grad_norm": 0.016741527244448662,
+      "learning_rate": 3.4732012039215776e-06,
+      "loss": 0.0051,
+      "step": 482
+    },
+    {
+      "epoch": 4.441379310344828,
+      "grad_norm": 0.015406622551381588,
+      "learning_rate": 3.3558147633999728e-06,
+      "loss": 0.0051,
+      "step": 483
+    },
+    {
+      "epoch": 4.450574712643678,
+      "grad_norm": 0.018325865268707275,
+      "learning_rate": 3.2403773738905187e-06,
+      "loss": 0.0049,
+      "step": 484
+    },
+    {
+      "epoch": 4.459770114942529,
+      "grad_norm": 0.01707405038177967,
+      "learning_rate": 3.126893859003249e-06,
+      "loss": 0.0053,
+      "step": 485
+    },
+    {
+      "epoch": 4.459770114942529,
+      "eval_loss": 0.011750674806535244,
+      "eval_runtime": 7.5766,
+      "eval_samples_per_second": 6.599,
+      "eval_steps_per_second": 1.716,
+      "step": 485
+    },
+    {
+      "epoch": 4.468965517241379,
+      "grad_norm": 0.01917904056608677,
+      "learning_rate": 3.0153689607045845e-06,
+      "loss": 0.0051,
+      "step": 486
+    },
+    {
+      "epoch": 4.47816091954023,
+      "grad_norm": 0.02045316807925701,
+      "learning_rate": 2.9058073391191375e-06,
+      "loss": 0.0054,
+      "step": 487
+    },
+    {
+      "epoch": 4.487356321839081,
+      "grad_norm": 0.017735499888658524,
+      "learning_rate": 2.798213572335001e-06,
+      "loss": 0.0061,
+      "step": 488
+    },
+    {
+      "epoch": 4.496551724137931,
+      "grad_norm": 0.01793784834444523,
+      "learning_rate": 2.692592156212487e-06,
+      "loss": 0.0063,
+      "step": 489
+    },
+    {
+      "epoch": 4.505747126436781,
+      "grad_norm": 0.01671445369720459,
+      "learning_rate": 2.5889475041961765e-06,
+      "loss": 0.0051,
+      "step": 490
+    },
+    {
+      "epoch": 4.505747126436781,
+      "eval_loss": 0.011716877110302448,
+      "eval_runtime": 7.5908,
+      "eval_samples_per_second": 6.587,
+      "eval_steps_per_second": 1.713,
+      "step": 490
+    },
+    {
+      "epoch": 4.514942528735633,
+      "grad_norm": 0.01827097497880459,
+      "learning_rate": 2.4872839471306084e-06,
+      "loss": 0.0052,
+      "step": 491
+    },
+    {
+      "epoch": 4.524137931034483,
+      "grad_norm": 0.018106255680322647,
+      "learning_rate": 2.3876057330792346e-06,
+      "loss": 0.0061,
+      "step": 492
+    },
+    {
+      "epoch": 4.533333333333333,
+      "grad_norm": 0.021528033539652824,
+      "learning_rate": 2.2899170271469428e-06,
+      "loss": 0.0057,
+      "step": 493
+    },
+    {
+      "epoch": 4.5425287356321835,
+      "grad_norm": 0.018092922866344452,
+      "learning_rate": 2.1942219113060212e-06,
+      "loss": 0.0045,
+      "step": 494
+    },
+    {
+      "epoch": 4.551724137931035,
+      "grad_norm": 0.018459059298038483,
+      "learning_rate": 2.100524384225555e-06,
+      "loss": 0.0053,
+      "step": 495
+    },
+    {
+      "epoch": 4.551724137931035,
+      "eval_loss": 0.011710132472217083,
+      "eval_runtime": 7.5789,
+      "eval_samples_per_second": 6.597,
+      "eval_steps_per_second": 1.715,
+      "step": 495
+    },
+    {
+      "epoch": 4.560919540229885,
+      "grad_norm": 0.018115947023034096,
+      "learning_rate": 2.0088283611044036e-06,
+      "loss": 0.005,
+      "step": 496
+    },
+    {
+      "epoch": 4.570114942528735,
+      "grad_norm": 0.021283112466335297,
+      "learning_rate": 1.9191376735075427e-06,
+      "loss": 0.0051,
+      "step": 497
+    },
+    {
+      "epoch": 4.5793103448275865,
+      "grad_norm": 0.019610431045293808,
+      "learning_rate": 1.8314560692059835e-06,
+      "loss": 0.0051,
+      "step": 498
+    },
+    {
+      "epoch": 4.588505747126437,
+      "grad_norm": 0.02000650204718113,
+      "learning_rate": 1.7457872120201779e-06,
+      "loss": 0.0049,
+      "step": 499
+    },
+    {
+      "epoch": 4.597701149425287,
+      "grad_norm": 0.019521350041031837,
+      "learning_rate": 1.6621346816668992e-06,
+      "loss": 0.0059,
+      "step": 500
+    },
+    {
+      "epoch": 4.597701149425287,
+      "eval_loss": 0.01169579103589058,
+      "eval_runtime": 7.5788,
+      "eval_samples_per_second": 6.597,
+      "eval_steps_per_second": 1.715,
+      "step": 500
+    },
+    {
+      "epoch": 4.606896551724138,
+      "grad_norm": 0.019491534680128098,
+      "learning_rate": 1.5805019736097104e-06,
+      "loss": 0.0058,
+      "step": 501
+    },
+    {
+      "epoch": 4.6160919540229886,
+      "grad_norm": 0.01981920190155506,
+      "learning_rate": 1.5008924989128258e-06,
+      "loss": 0.0058,
+      "step": 502
+    },
+    {
+      "epoch": 4.625287356321839,
+      "grad_norm": 0.01777651347219944,
+      "learning_rate": 1.4233095840986753e-06,
+      "loss": 0.0055,
+      "step": 503
+    },
+    {
+      "epoch": 4.63448275862069,
+      "grad_norm": 0.019680196419358253,
+      "learning_rate": 1.3477564710088098e-06,
+      "loss": 0.0049,
+      "step": 504
+    },
+    {
+      "epoch": 4.64367816091954,
+      "grad_norm": 0.01789335533976555,
+      "learning_rate": 1.2742363166685034e-06,
+      "loss": 0.0055,
+      "step": 505
+    },
+    {
+      "epoch": 4.64367816091954,
+      "eval_loss": 0.011671243235468864,
+      "eval_runtime": 7.6173,
+      "eval_samples_per_second": 6.564,
+      "eval_steps_per_second": 1.707,
+      "step": 505
+    },
+    {
+      "epoch": 4.652873563218391,
+      "grad_norm": 0.01689436286687851,
+      "learning_rate": 1.2027521931548214e-06,
+      "loss": 0.0045,
+      "step": 506
+    },
+    {
+      "epoch": 4.662068965517241,
+      "grad_norm": 0.016485173255205154,
+      "learning_rate": 1.1333070874682216e-06,
+      "loss": 0.0055,
+      "step": 507
+    },
+    {
+      "epoch": 4.671264367816092,
+      "grad_norm": 0.02044127695262432,
+      "learning_rate": 1.0659039014077944e-06,
+      "loss": 0.0046,
+      "step": 508
+    },
+    {
+      "epoch": 4.680459770114942,
+      "grad_norm": 0.019011957570910454,
+      "learning_rate": 1.0005454514499414e-06,
+      "loss": 0.0055,
+      "step": 509
+    },
+    {
+      "epoch": 4.689655172413794,
+      "grad_norm": 0.016083214432001114,
+      "learning_rate": 9.372344686307655e-07,
+      "loss": 0.0054,
+      "step": 510
+    },
+    {
+      "epoch": 4.689655172413794,
+      "eval_loss": 0.011643964797258377,
+      "eval_runtime": 7.5863,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 510
+    },
+    {
+      "epoch": 4.698850574712644,
+      "grad_norm": 0.027418499812483788,
+      "learning_rate": 8.759735984318895e-07,
+      "loss": 0.0057,
+      "step": 511
+    },
+    {
+      "epoch": 4.708045977011494,
+      "grad_norm": 0.016228538006544113,
+      "learning_rate": 8.167654006699443e-07,
+      "loss": 0.0054,
+      "step": 512
+    },
+    {
+      "epoch": 4.7172413793103445,
+      "grad_norm": 0.02453775331377983,
+      "learning_rate": 7.596123493895991e-07,
+      "loss": 0.0059,
+      "step": 513
+    },
+    {
+      "epoch": 4.726436781609196,
+      "grad_norm": 0.019689669832587242,
+      "learning_rate": 7.04516832760177e-07,
+      "loss": 0.0062,
+      "step": 514
+    },
+    {
+      "epoch": 4.735632183908046,
+      "grad_norm": 0.017698440700769424,
+      "learning_rate": 6.514811529758747e-07,
+      "loss": 0.0055,
+      "step": 515
+    },
+    {
+      "epoch": 4.735632183908046,
+      "eval_loss": 0.011669199913740158,
+      "eval_runtime": 7.5858,
+      "eval_samples_per_second": 6.591,
+      "eval_steps_per_second": 1.714,
+      "step": 515
+    },
+    {
+      "epoch": 4.744827586206896,
+      "grad_norm": 0.0169072262942791,
+      "learning_rate": 6.005075261595494e-07,
+      "loss": 0.0053,
+      "step": 516
+    },
+    {
+      "epoch": 4.7540229885057474,
+      "grad_norm": 0.02021600864827633,
+      "learning_rate": 5.515980822701439e-07,
+      "loss": 0.0055,
+      "step": 517
+    },
+    {
+      "epoch": 4.763218390804598,
+      "grad_norm": 0.021086223423480988,
+      "learning_rate": 5.047548650136513e-07,
+      "loss": 0.0053,
+      "step": 518
+    },
+    {
+      "epoch": 4.772413793103448,
+      "grad_norm": 0.02129381150007248,
+      "learning_rate": 4.5997983175773417e-07,
+      "loss": 0.0054,
+      "step": 519
+    },
+    {
+      "epoch": 4.781609195402299,
+      "grad_norm": 0.018404850736260414,
+      "learning_rate": 4.1727485344994486e-07,
+      "loss": 0.0056,
+      "step": 520
+    },
+    {
+      "epoch": 4.781609195402299,
+      "eval_loss": 0.01162272784858942,
+      "eval_runtime": 7.644,
+      "eval_samples_per_second": 6.541,
+      "eval_steps_per_second": 1.701,
+      "step": 520
+    },
+    {
+      "epoch": 4.7908045977011495,
+      "grad_norm": 0.01828380487859249,
+      "learning_rate": 3.766417145395218e-07,
+      "loss": 0.0049,
+      "step": 521
+    },
+    {
+      "epoch": 4.8,
+      "grad_norm": 0.01730126328766346,
+      "learning_rate": 3.380821129028489e-07,
+      "loss": 0.0055,
+      "step": 522
+    },
+    {
+      "epoch": 4.809195402298851,
+      "grad_norm": 0.017449945211410522,
+      "learning_rate": 3.0159765977250673e-07,
+      "loss": 0.0057,
+      "step": 523
+    },
+    {
+      "epoch": 4.818390804597701,
+      "grad_norm": 0.01770183816552162,
+      "learning_rate": 2.671898796699268e-07,
+      "loss": 0.005,
+      "step": 524
+    },
+    {
+      "epoch": 4.827586206896552,
+      "grad_norm": 0.016674352809786797,
+      "learning_rate": 2.3486021034170857e-07,
+      "loss": 0.0048,
+      "step": 525
+    },
+    {
+      "epoch": 4.827586206896552,
+      "eval_loss": 0.011633777059614658,
+      "eval_runtime": 7.591,
+      "eval_samples_per_second": 6.587,
+      "eval_steps_per_second": 1.713,
+      "step": 525
+    },
+    {
+      "epoch": 4.836781609195402,
+      "grad_norm": 0.016963541507720947,
+      "learning_rate": 2.0461000269953456e-07,
+      "loss": 0.0061,
+      "step": 526
+    },
+    {
+      "epoch": 4.845977011494253,
+      "grad_norm": 0.018301161006093025,
+      "learning_rate": 1.7644052076371542e-07,
+      "loss": 0.0054,
+      "step": 527
+    },
+    {
+      "epoch": 4.855172413793103,
+      "grad_norm": 0.018041977658867836,
+      "learning_rate": 1.503529416103988e-07,
+      "loss": 0.0048,
+      "step": 528
+    },
+    {
+      "epoch": 4.864367816091954,
+      "grad_norm": 0.016023052856326103,
+      "learning_rate": 1.2634835532233657e-07,
+      "loss": 0.0053,
+      "step": 529
+    },
+    {
+      "epoch": 4.873563218390805,
+      "grad_norm": 0.01638129912316799,
+      "learning_rate": 1.044277649433989e-07,
+      "loss": 0.0049,
+      "step": 530
+    },
+    {
+      "epoch": 4.873563218390805,
+      "eval_loss": 0.01163394283503294,
+      "eval_runtime": 7.5775,
+      "eval_samples_per_second": 6.598,
+      "eval_steps_per_second": 1.716,
+      "step": 530
+    },
+    {
+      "epoch": 4.882758620689655,
+      "grad_norm": 0.02009713463485241,
+      "learning_rate": 8.459208643659122e-08,
+      "loss": 0.0047,
+      "step": 531
+    },
+    {
+      "epoch": 4.8919540229885055,
+      "grad_norm": 0.018126314505934715,
+      "learning_rate": 6.684214864584038e-08,
+      "loss": 0.0048,
+      "step": 532
+    },
+    {
+      "epoch": 4.901149425287357,
+      "grad_norm": 0.014890230260789394,
+      "learning_rate": 5.11786932613223e-08,
+      "loss": 0.005,
+      "step": 533
+    },
+    {
+      "epoch": 4.910344827586207,
+      "grad_norm": 0.016795996576547623,
+      "learning_rate": 3.760237478849793e-08,
+      "loss": 0.0044,
+      "step": 534
+    },
+    {
+      "epoch": 4.919540229885057,
+      "grad_norm": 0.016731204465031624,
+      "learning_rate": 2.6113760520735108e-08,
+      "loss": 0.0043,
+      "step": 535
+    },
+    {
+      "epoch": 4.919540229885057,
+      "eval_loss": 0.011621917597949505,
+      "eval_runtime": 7.6095,
+      "eval_samples_per_second": 6.571,
+      "eval_steps_per_second": 1.708,
+      "step": 535
+    },
+    {
+      "epoch": 4.928735632183908,
+      "grad_norm": 0.02216600440442562,
+      "learning_rate": 1.6713330515627513e-08,
+      "loss": 0.0054,
+      "step": 536
+    },
+    {
+      "epoch": 4.937931034482759,
+      "grad_norm": 0.026815466582775116,
+      "learning_rate": 9.401477574932926e-09,
+      "loss": 0.0048,
+      "step": 537
+    },
+    {
+      "epoch": 4.947126436781609,
+      "grad_norm": 0.01739557273685932,
+      "learning_rate": 4.178507228136397e-09,
+      "loss": 0.0051,
+      "step": 538
+    },
+    {
+      "epoch": 4.956321839080459,
+      "grad_norm": 0.017503153532743454,
+      "learning_rate": 1.0446377197104173e-09,
+      "loss": 0.0048,
+      "step": 539
+    },
+    {
+      "epoch": 4.9655172413793105,
+      "grad_norm": 0.015828318893909454,
+      "learning_rate": 0.0,
+      "loss": 0.0046,
+      "step": 540
+    },
+    {
+      "epoch": 4.9655172413793105,
+      "eval_loss": 0.011626788415014744,
+      "eval_runtime": 7.5888,
+      "eval_samples_per_second": 6.589,
+      "eval_steps_per_second": 1.713,
+      "step": 540
+    },
+    {
+      "epoch": 4.9655172413793105,
+      "step": 540,
+      "total_flos": 1.0432767520256164e+18,
+      "train_loss": 0.011179022066709067,
+      "train_runtime": 8360.2497,
+      "train_samples_per_second": 2.081,
+      "train_steps_per_second": 0.065
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 540,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.0432767520256164e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}