{
  "best_metric": 0.4087750315666199,
  "best_model_checkpoint": "deberta-v3-large-lr55-e6-bs4-acc6-dsv3/checkpoint-213",
  "epoch": 12.0,
  "eval_steps": 500,
  "global_step": 512,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 1.8460181951522827,
      "learning_rate": 5.496726190476191e-06,
      "loss": 0.6818,
      "step": 1
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.9176546335220337,
      "learning_rate": 5.493452380952381e-06,
      "loss": 0.7102,
      "step": 2
    },
    {
      "epoch": 0.07,
      "grad_norm": 2.076033353805542,
      "learning_rate": 5.490178571428571e-06,
      "loss": 0.712,
      "step": 3
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.9934174418449402,
      "learning_rate": 5.486904761904762e-06,
      "loss": 0.6877,
      "step": 4
    },
    {
      "epoch": 0.12,
      "grad_norm": 3.206833600997925,
      "learning_rate": 5.483630952380952e-06,
      "loss": 0.7138,
      "step": 5
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.930253803730011,
      "learning_rate": 5.480357142857143e-06,
      "loss": 0.675,
      "step": 6
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.047216773033142,
      "learning_rate": 5.477083333333333e-06,
      "loss": 0.7025,
      "step": 7
    },
    {
      "epoch": 0.19,
      "grad_norm": 2.5932674407958984,
      "learning_rate": 5.473809523809524e-06,
      "loss": 0.6736,
      "step": 8
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.1506260633468628,
      "learning_rate": 5.4705357142857145e-06,
      "loss": 0.6687,
      "step": 9
    },
    {
      "epoch": 0.23,
      "grad_norm": 2.9675679206848145,
      "learning_rate": 5.4672619047619045e-06,
      "loss": 0.7021,
      "step": 10
    },
    {
      "epoch": 0.26,
      "grad_norm": 3.620849132537842,
      "learning_rate": 5.463988095238095e-06,
      "loss": 0.682,
      "step": 11
    },
    {
      "epoch": 0.28,
      "grad_norm": 2.2752580642700195,
      "learning_rate": 5.460714285714286e-06,
      "loss": 0.6633,
      "step": 12
    },
    {
      "epoch": 0.3,
      "grad_norm": 3.6715457439422607,
      "learning_rate": 5.457440476190476e-06,
      "loss": 0.6926,
      "step": 13
    },
    {
      "epoch": 0.33,
      "grad_norm": 3.4275119304656982,
      "learning_rate": 5.454166666666667e-06,
      "loss": 0.6945,
      "step": 14
    },
    {
      "epoch": 0.35,
      "grad_norm": 2.6921539306640625,
      "learning_rate": 5.450892857142857e-06,
      "loss": 0.6876,
      "step": 15
    },
    {
      "epoch": 0.38,
      "grad_norm": 3.6308276653289795,
      "learning_rate": 5.4476190476190476e-06,
      "loss": 0.7249,
      "step": 16
    },
    {
      "epoch": 0.4,
      "grad_norm": 1.0595824718475342,
      "learning_rate": 5.444345238095238e-06,
      "loss": 0.6804,
      "step": 17
    },
    {
      "epoch": 0.42,
      "grad_norm": 2.645174026489258,
      "learning_rate": 5.441071428571428e-06,
      "loss": 0.6768,
      "step": 18
    },
    {
      "epoch": 0.45,
      "grad_norm": 1.631472110748291,
      "learning_rate": 5.437797619047619e-06,
      "loss": 0.6812,
      "step": 19
    },
    {
      "epoch": 0.47,
      "grad_norm": 3.3360867500305176,
      "learning_rate": 5.43452380952381e-06,
      "loss": 0.6817,
      "step": 20
    },
    {
      "epoch": 0.49,
      "grad_norm": 1.531082034111023,
      "learning_rate": 5.43125e-06,
      "loss": 0.6624,
      "step": 21
    },
    {
      "epoch": 0.52,
      "grad_norm": 2.1303884983062744,
      "learning_rate": 5.427976190476191e-06,
      "loss": 0.6926,
      "step": 22
    },
    {
      "epoch": 0.54,
      "grad_norm": 2.1816420555114746,
      "learning_rate": 5.4247023809523815e-06,
      "loss": 0.7173,
      "step": 23
    },
    {
      "epoch": 0.56,
      "grad_norm": 3.382896900177002,
      "learning_rate": 5.4214285714285714e-06,
      "loss": 0.7172,
      "step": 24
    },
    {
      "epoch": 0.59,
      "grad_norm": 4.177649021148682,
      "learning_rate": 5.418154761904762e-06,
      "loss": 0.6932,
      "step": 25
    },
    {
      "epoch": 0.61,
      "grad_norm": 6.432918548583984,
      "learning_rate": 5.414880952380952e-06,
      "loss": 0.6679,
      "step": 26
    },
    {
      "epoch": 0.63,
      "grad_norm": 1.7472172975540161,
      "learning_rate": 5.411607142857142e-06,
      "loss": 0.6704,
      "step": 27
    },
    {
      "epoch": 0.66,
      "grad_norm": 2.1172163486480713,
      "learning_rate": 5.408333333333333e-06,
      "loss": 0.6637,
      "step": 28
    },
    {
      "epoch": 0.68,
      "grad_norm": 1.6855010986328125,
      "learning_rate": 5.405059523809524e-06,
      "loss": 0.656,
      "step": 29
    },
    {
      "epoch": 0.7,
      "grad_norm": 2.4389238357543945,
      "learning_rate": 5.401785714285714e-06,
      "loss": 0.6562,
      "step": 30
    },
    {
      "epoch": 0.73,
      "grad_norm": 3.0723414421081543,
      "learning_rate": 5.3985119047619045e-06,
      "loss": 0.6812,
      "step": 31
    },
    {
      "epoch": 0.75,
      "grad_norm": 2.1830461025238037,
      "learning_rate": 5.3952380952380944e-06,
      "loss": 0.6448,
      "step": 32
    },
    {
      "epoch": 0.77,
      "grad_norm": 3.109797716140747,
      "learning_rate": 5.391964285714285e-06,
      "loss": 0.6841,
      "step": 33
    },
    {
      "epoch": 0.8,
      "grad_norm": 2.5417990684509277,
      "learning_rate": 5.388690476190476e-06,
      "loss": 0.6428,
      "step": 34
    },
    {
      "epoch": 0.82,
      "grad_norm": 2.0981853008270264,
      "learning_rate": 5.385416666666666e-06,
      "loss": 0.6493,
      "step": 35
    },
    {
      "epoch": 0.84,
      "grad_norm": 1.927240014076233,
      "learning_rate": 5.382142857142857e-06,
      "loss": 0.6496,
      "step": 36
    },
    {
      "epoch": 0.87,
      "grad_norm": 4.362425327301025,
      "learning_rate": 5.378869047619048e-06,
      "loss": 0.6666,
      "step": 37
    },
    {
      "epoch": 0.89,
      "grad_norm": 6.46484899520874,
      "learning_rate": 5.3755952380952376e-06,
      "loss": 0.7947,
      "step": 38
    },
    {
      "epoch": 0.91,
      "grad_norm": 3.2664661407470703,
      "learning_rate": 5.372321428571428e-06,
      "loss": 0.608,
      "step": 39
    },
    {
      "epoch": 0.94,
      "grad_norm": 2.2718191146850586,
      "learning_rate": 5.369047619047619e-06,
      "loss": 0.6247,
      "step": 40
    },
    {
      "epoch": 0.96,
      "grad_norm": 5.372130870819092,
      "learning_rate": 5.365773809523809e-06,
      "loss": 0.7885,
      "step": 41
    },
    {
      "epoch": 0.98,
      "grad_norm": 1.7550398111343384,
      "learning_rate": 5.3625e-06,
      "loss": 0.6425,
      "step": 42
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.6681818181818182,
      "eval_loss": 0.6295949816703796,
      "eval_runtime": 1.9777,
      "eval_samples_per_second": 111.242,
      "eval_steps_per_second": 27.81,
      "step": 42
    },
    {
      "epoch": 1.01,
      "grad_norm": 3.6129467487335205,
      "learning_rate": 5.35922619047619e-06,
      "loss": 0.7519,
      "step": 43
    },
    {
      "epoch": 1.03,
      "grad_norm": 4.161266326904297,
      "learning_rate": 5.355952380952381e-06,
      "loss": 0.6534,
      "step": 44
    },
    {
      "epoch": 1.05,
      "grad_norm": 2.0768189430236816,
      "learning_rate": 5.3526785714285715e-06,
      "loss": 0.6805,
      "step": 45
    },
    {
      "epoch": 1.08,
      "grad_norm": 2.847532272338867,
      "learning_rate": 5.349404761904761e-06,
      "loss": 0.6691,
      "step": 46
    },
    {
      "epoch": 1.1,
      "grad_norm": 2.2965738773345947,
      "learning_rate": 5.346130952380952e-06,
      "loss": 0.6695,
      "step": 47
    },
    {
      "epoch": 1.12,
      "grad_norm": 2.5308315753936768,
      "learning_rate": 5.342857142857143e-06,
      "loss": 0.6215,
      "step": 48
    },
    {
      "epoch": 1.15,
      "grad_norm": 2.581547975540161,
      "learning_rate": 5.339583333333333e-06,
      "loss": 0.5902,
      "step": 49
    },
    {
      "epoch": 1.17,
      "grad_norm": 3.935713768005371,
      "learning_rate": 5.336309523809524e-06,
      "loss": 0.5963,
      "step": 50
    },
    {
      "epoch": 1.2,
      "grad_norm": 3.4248504638671875,
      "learning_rate": 5.333035714285714e-06,
      "loss": 0.6649,
      "step": 51
    },
    {
      "epoch": 1.22,
      "grad_norm": 7.237908840179443,
      "learning_rate": 5.3297619047619045e-06,
      "loss": 0.7107,
      "step": 52
    },
    {
      "epoch": 1.24,
      "grad_norm": 2.9264256954193115,
      "learning_rate": 5.326488095238095e-06,
      "loss": 0.6806,
      "step": 53
    },
    {
      "epoch": 1.27,
      "grad_norm": 2.699674606323242,
      "learning_rate": 5.323214285714285e-06,
      "loss": 0.6276,
      "step": 54
    },
    {
      "epoch": 1.29,
      "grad_norm": 3.520171642303467,
      "learning_rate": 5.319940476190476e-06,
      "loss": 0.6278,
      "step": 55
    },
    {
      "epoch": 1.31,
      "grad_norm": 2.0494229793548584,
      "learning_rate": 5.316666666666667e-06,
      "loss": 0.5886,
      "step": 56
    },
    {
      "epoch": 1.34,
      "grad_norm": 2.4809341430664062,
      "learning_rate": 5.313392857142857e-06,
      "loss": 0.6001,
      "step": 57
    },
    {
      "epoch": 1.36,
      "grad_norm": 5.9523820877075195,
      "learning_rate": 5.310119047619048e-06,
      "loss": 0.6717,
      "step": 58
    },
    {
      "epoch": 1.38,
      "grad_norm": 4.1546502113342285,
      "learning_rate": 5.306845238095238e-06,
      "loss": 0.6483,
      "step": 59
    },
    {
      "epoch": 1.41,
      "grad_norm": 3.3143973350524902,
      "learning_rate": 5.303571428571428e-06,
      "loss": 0.5775,
      "step": 60
    },
    {
      "epoch": 1.43,
      "grad_norm": 7.167575359344482,
      "learning_rate": 5.300297619047619e-06,
      "loss": 0.6182,
      "step": 61
    },
    {
      "epoch": 1.45,
      "grad_norm": 1.888744592666626,
      "learning_rate": 5.297023809523809e-06,
      "loss": 0.5694,
      "step": 62
    },
    {
      "epoch": 1.48,
      "grad_norm": 4.127984046936035,
      "learning_rate": 5.29375e-06,
      "loss": 0.5242,
      "step": 63
    },
    {
      "epoch": 1.5,
      "grad_norm": 3.9158403873443604,
      "learning_rate": 5.290476190476191e-06,
      "loss": 0.5515,
      "step": 64
    },
    {
      "epoch": 1.52,
      "grad_norm": 3.5147557258605957,
      "learning_rate": 5.287202380952381e-06,
      "loss": 0.5724,
      "step": 65
    },
    {
      "epoch": 1.55,
      "grad_norm": 4.157176494598389,
      "learning_rate": 5.2839285714285715e-06,
      "loss": 0.6203,
      "step": 66
    },
    {
      "epoch": 1.57,
      "grad_norm": 5.852452754974365,
      "learning_rate": 5.280654761904762e-06,
      "loss": 0.6141,
      "step": 67
    },
    {
      "epoch": 1.59,
      "grad_norm": 4.920362949371338,
      "learning_rate": 5.277380952380952e-06,
      "loss": 0.6591,
      "step": 68
    },
    {
      "epoch": 1.62,
      "grad_norm": 13.502717018127441,
      "learning_rate": 5.274107142857143e-06,
      "loss": 0.5012,
      "step": 69
    },
    {
      "epoch": 1.64,
      "grad_norm": 3.996752977371216,
      "learning_rate": 5.270833333333333e-06,
      "loss": 0.5382,
      "step": 70
    },
    {
      "epoch": 1.66,
      "grad_norm": 6.804556369781494,
      "learning_rate": 5.267559523809524e-06,
      "loss": 0.6235,
      "step": 71
    },
    {
      "epoch": 1.69,
      "grad_norm": 3.19149112701416,
      "learning_rate": 5.264285714285715e-06,
      "loss": 0.6962,
      "step": 72
    },
    {
      "epoch": 1.71,
      "grad_norm": 5.843012809753418,
      "learning_rate": 5.2610119047619045e-06,
      "loss": 0.619,
      "step": 73
    },
    {
      "epoch": 1.73,
      "grad_norm": 8.353483200073242,
      "learning_rate": 5.257738095238095e-06,
      "loss": 0.5508,
      "step": 74
    },
    {
      "epoch": 1.76,
      "grad_norm": 4.157430648803711,
      "learning_rate": 5.254464285714286e-06,
      "loss": 0.5476,
      "step": 75
    },
    {
      "epoch": 1.78,
      "grad_norm": 7.451327800750732,
      "learning_rate": 5.251190476190476e-06,
      "loss": 0.563,
      "step": 76
    },
    {
      "epoch": 1.8,
      "grad_norm": 2.8933181762695312,
      "learning_rate": 5.247916666666667e-06,
      "loss": 0.537,
      "step": 77
    },
    {
      "epoch": 1.83,
      "grad_norm": 5.481462478637695,
      "learning_rate": 5.244642857142858e-06,
      "loss": 0.4794,
      "step": 78
    },
    {
      "epoch": 1.85,
      "grad_norm": 3.3026862144470215,
      "learning_rate": 5.241369047619047e-06,
      "loss": 0.5493,
      "step": 79
    },
    {
      "epoch": 1.88,
      "grad_norm": 4.270678520202637,
      "learning_rate": 5.238095238095238e-06,
      "loss": 0.5098,
      "step": 80
    },
    {
      "epoch": 1.9,
      "grad_norm": 6.741820812225342,
      "learning_rate": 5.234821428571428e-06,
      "loss": 0.587,
      "step": 81
    },
    {
      "epoch": 1.92,
      "grad_norm": 12.857108116149902,
      "learning_rate": 5.231547619047618e-06,
      "loss": 0.6178,
      "step": 82
    },
    {
      "epoch": 1.95,
      "grad_norm": 14.573826789855957,
      "learning_rate": 5.228273809523809e-06,
      "loss": 0.5698,
      "step": 83
    },
    {
      "epoch": 1.97,
      "grad_norm": 16.281530380249023,
      "learning_rate": 5.225e-06,
      "loss": 0.5929,
      "step": 84
    },
    {
      "epoch": 1.99,
      "grad_norm": 8.965470314025879,
      "learning_rate": 5.22172619047619e-06,
      "loss": 0.583,
      "step": 85
    },
    {
      "epoch": 1.99,
      "eval_accuracy": 0.7,
      "eval_loss": 0.5743765830993652,
      "eval_runtime": 1.9707,
      "eval_samples_per_second": 111.636,
      "eval_steps_per_second": 27.909,
      "step": 85
    },
    {
      "epoch": 2.02,
      "grad_norm": 7.075886249542236,
      "learning_rate": 5.218452380952381e-06,
      "loss": 0.5969,
      "step": 86
    },
    {
      "epoch": 2.04,
      "grad_norm": 5.647829055786133,
      "learning_rate": 5.2151785714285715e-06,
      "loss": 0.5581,
      "step": 87
    },
    {
      "epoch": 2.06,
      "grad_norm": 12.890501976013184,
      "learning_rate": 5.2119047619047615e-06,
      "loss": 0.5742,
      "step": 88
    },
    {
      "epoch": 2.09,
      "grad_norm": 6.206398963928223,
      "learning_rate": 5.208630952380952e-06,
      "loss": 0.4763,
      "step": 89
    },
    {
      "epoch": 2.11,
      "grad_norm": 15.82968521118164,
      "learning_rate": 5.205357142857142e-06,
      "loss": 0.5015,
      "step": 90
    },
    {
      "epoch": 2.13,
      "grad_norm": 2.781341314315796,
      "learning_rate": 5.202083333333333e-06,
      "loss": 0.6048,
      "step": 91
    },
    {
      "epoch": 2.16,
      "grad_norm": 31.130008697509766,
      "learning_rate": 5.198809523809524e-06,
      "loss": 0.6021,
      "step": 92
    },
    {
      "epoch": 2.18,
      "grad_norm": 41.88652038574219,
      "learning_rate": 5.195535714285714e-06,
      "loss": 0.5294,
      "step": 93
    },
    {
      "epoch": 2.2,
      "grad_norm": 13.205155372619629,
      "learning_rate": 5.1922619047619046e-06,
      "loss": 0.5571,
      "step": 94
    },
    {
      "epoch": 2.23,
      "grad_norm": 4.5637054443359375,
      "learning_rate": 5.188988095238095e-06,
      "loss": 0.415,
      "step": 95
    },
    {
      "epoch": 2.25,
      "grad_norm": 8.613974571228027,
      "learning_rate": 5.185714285714285e-06,
      "loss": 0.665,
      "step": 96
    },
    {
      "epoch": 2.27,
      "grad_norm": 16.234405517578125,
      "learning_rate": 5.182440476190476e-06,
      "loss": 0.5181,
      "step": 97
    },
    {
      "epoch": 2.3,
      "grad_norm": 6.523887634277344,
      "learning_rate": 5.179166666666666e-06,
      "loss": 0.4848,
      "step": 98
    },
    {
      "epoch": 2.32,
      "grad_norm": 9.602444648742676,
      "learning_rate": 5.175892857142857e-06,
      "loss": 0.4226,
      "step": 99
    },
    {
      "epoch": 2.34,
      "grad_norm": 5.141808032989502,
      "learning_rate": 5.172619047619048e-06,
      "loss": 0.4066,
      "step": 100
    },
    {
      "epoch": 2.37,
      "grad_norm": 14.732641220092773,
      "learning_rate": 5.169345238095238e-06,
      "loss": 0.4309,
      "step": 101
    },
    {
      "epoch": 2.39,
      "grad_norm": 13.681814193725586,
      "learning_rate": 5.166071428571428e-06,
      "loss": 0.5336,
      "step": 102
    },
    {
      "epoch": 2.41,
      "grad_norm": 9.389954566955566,
      "learning_rate": 5.162797619047619e-06,
      "loss": 0.4979,
      "step": 103
    },
    {
      "epoch": 2.44,
      "grad_norm": 3.370234727859497,
      "learning_rate": 5.159523809523809e-06,
      "loss": 0.5064,
      "step": 104
    },
    {
      "epoch": 2.46,
      "grad_norm": 6.832080841064453,
      "learning_rate": 5.15625e-06,
      "loss": 0.5425,
      "step": 105
    },
    {
      "epoch": 2.48,
      "grad_norm": 5.864181041717529,
      "learning_rate": 5.152976190476191e-06,
      "loss": 0.4536,
      "step": 106
    },
    {
      "epoch": 2.51,
      "grad_norm": 4.189244747161865,
      "learning_rate": 5.149702380952381e-06,
      "loss": 0.5783,
      "step": 107
    },
    {
      "epoch": 2.53,
      "grad_norm": 13.515569686889648,
      "learning_rate": 5.1464285714285715e-06,
      "loss": 0.6054,
      "step": 108
    },
    {
      "epoch": 2.55,
      "grad_norm": 8.88909912109375,
      "learning_rate": 5.1431547619047615e-06,
      "loss": 0.4679,
      "step": 109
    },
    {
      "epoch": 2.58,
      "grad_norm": 12.870144844055176,
      "learning_rate": 5.139880952380952e-06,
      "loss": 0.4515,
      "step": 110
    },
    {
      "epoch": 2.6,
      "grad_norm": 7.6494927406311035,
      "learning_rate": 5.136607142857143e-06,
      "loss": 0.4118,
      "step": 111
    },
    {
      "epoch": 2.62,
      "grad_norm": 5.379680633544922,
      "learning_rate": 5.133333333333333e-06,
      "loss": 0.4404,
      "step": 112
    },
    {
      "epoch": 2.65,
      "grad_norm": 12.46003246307373,
      "learning_rate": 5.130059523809524e-06,
      "loss": 0.5413,
      "step": 113
    },
    {
      "epoch": 2.67,
      "grad_norm": 16.497268676757812,
      "learning_rate": 5.126785714285715e-06,
      "loss": 0.4424,
      "step": 114
    },
    {
      "epoch": 2.7,
      "grad_norm": 27.879886627197266,
      "learning_rate": 5.123511904761905e-06,
      "loss": 0.6332,
      "step": 115
    },
    {
      "epoch": 2.72,
      "grad_norm": 24.27963638305664,
      "learning_rate": 5.120238095238095e-06,
      "loss": 0.6507,
      "step": 116
    },
    {
      "epoch": 2.74,
      "grad_norm": 9.079182624816895,
      "learning_rate": 5.116964285714285e-06,
      "loss": 0.4821,
      "step": 117
    },
    {
      "epoch": 2.77,
      "grad_norm": 8.210986137390137,
      "learning_rate": 5.113690476190476e-06,
      "loss": 0.4236,
      "step": 118
    },
    {
      "epoch": 2.79,
      "grad_norm": 9.538209915161133,
      "learning_rate": 5.110416666666667e-06,
      "loss": 0.5458,
      "step": 119
    },
    {
      "epoch": 2.81,
      "grad_norm": 11.242250442504883,
      "learning_rate": 5.107142857142857e-06,
      "loss": 0.3717,
      "step": 120
    },
    {
      "epoch": 2.84,
      "grad_norm": 7.99606466293335,
      "learning_rate": 5.103869047619048e-06,
      "loss": 0.463,
      "step": 121
    },
    {
      "epoch": 2.86,
      "grad_norm": 8.157175064086914,
      "learning_rate": 5.1005952380952385e-06,
      "loss": 0.4463,
      "step": 122
    },
    {
      "epoch": 2.88,
      "grad_norm": 6.33786678314209,
      "learning_rate": 5.0973214285714284e-06,
      "loss": 0.4483,
      "step": 123
    },
    {
      "epoch": 2.91,
      "grad_norm": 12.092632293701172,
      "learning_rate": 5.094047619047619e-06,
      "loss": 0.4729,
      "step": 124
    },
    {
      "epoch": 2.93,
      "grad_norm": 6.666680335998535,
      "learning_rate": 5.09077380952381e-06,
      "loss": 0.4628,
      "step": 125
    },
    {
      "epoch": 2.95,
      "grad_norm": 13.772384643554688,
      "learning_rate": 5.0875e-06,
      "loss": 0.2814,
      "step": 126
    },
    {
      "epoch": 2.98,
      "grad_norm": 28.22318458557129,
      "learning_rate": 5.084226190476191e-06,
      "loss": 0.6727,
      "step": 127
    },
    {
      "epoch": 3.0,
      "grad_norm": 25.690156936645508,
      "learning_rate": 5.080952380952381e-06,
      "loss": 0.4848,
      "step": 128
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7227272727272728,
      "eval_loss": 0.6313565373420715,
      "eval_runtime": 1.9701,
      "eval_samples_per_second": 111.669,
      "eval_steps_per_second": 27.917,
      "step": 128
    },
    {
      "epoch": 3.02,
      "grad_norm": 20.357358932495117,
      "learning_rate": 5.0776785714285715e-06,
      "loss": 0.6101,
      "step": 129
    },
    {
      "epoch": 3.05,
      "grad_norm": 24.226192474365234,
      "learning_rate": 5.074404761904762e-06,
      "loss": 0.4901,
      "step": 130
    },
    {
      "epoch": 3.07,
      "grad_norm": 21.727924346923828,
      "learning_rate": 5.071130952380952e-06,
      "loss": 0.3983,
      "step": 131
    },
    {
      "epoch": 3.09,
      "grad_norm": 8.311666488647461,
      "learning_rate": 5.067857142857142e-06,
      "loss": 0.4098,
      "step": 132
    },
    {
      "epoch": 3.12,
      "grad_norm": 8.897052764892578,
      "learning_rate": 5.064583333333333e-06,
      "loss": 0.4665,
      "step": 133
    },
    {
      "epoch": 3.14,
      "grad_norm": 24.099576950073242,
      "learning_rate": 5.061309523809523e-06,
      "loss": 0.4198,
      "step": 134
    },
    {
      "epoch": 3.16,
      "grad_norm": 10.50734806060791,
      "learning_rate": 5.058035714285714e-06,
      "loss": 0.3888,
      "step": 135
    },
    {
      "epoch": 3.19,
      "grad_norm": 22.82213592529297,
      "learning_rate": 5.054761904761905e-06,
      "loss": 0.4525,
      "step": 136
    },
    {
      "epoch": 3.21,
      "grad_norm": 37.29859161376953,
      "learning_rate": 5.0514880952380946e-06,
      "loss": 0.486,
      "step": 137
    },
    {
      "epoch": 3.23,
      "grad_norm": 14.720842361450195,
      "learning_rate": 5.048214285714285e-06,
      "loss": 0.4094,
      "step": 138
    },
    {
      "epoch": 3.26,
      "grad_norm": 13.443496704101562,
      "learning_rate": 5.044940476190476e-06,
      "loss": 0.3444,
      "step": 139
    },
    {
      "epoch": 3.28,
      "grad_norm": 10.844743728637695,
      "learning_rate": 5.041666666666666e-06,
      "loss": 0.3291,
      "step": 140
    },
    {
      "epoch": 3.3,
      "grad_norm": 10.810824394226074,
      "learning_rate": 5.038392857142857e-06,
      "loss": 0.4249,
      "step": 141
    },
    {
      "epoch": 3.33,
      "grad_norm": 5.066943168640137,
      "learning_rate": 5.035119047619048e-06,
      "loss": 0.3591,
      "step": 142
    },
    {
      "epoch": 3.35,
      "grad_norm": 75.52027130126953,
      "learning_rate": 5.031845238095238e-06,
      "loss": 0.4469,
      "step": 143
    },
    {
      "epoch": 3.38,
      "grad_norm": 41.847652435302734,
      "learning_rate": 5.0285714285714285e-06,
      "loss": 0.4158,
      "step": 144
    },
    {
      "epoch": 3.4,
      "grad_norm": 17.77975082397461,
      "learning_rate": 5.025297619047618e-06,
      "loss": 0.2459,
      "step": 145
    },
    {
      "epoch": 3.42,
      "grad_norm": 41.41980743408203,
      "learning_rate": 5.022023809523809e-06,
      "loss": 0.5155,
      "step": 146
    },
    {
      "epoch": 3.45,
      "grad_norm": 42.26784896850586,
      "learning_rate": 5.01875e-06,
      "loss": 0.3646,
      "step": 147
    },
    {
      "epoch": 3.47,
      "grad_norm": 15.376642227172852,
      "learning_rate": 5.01547619047619e-06,
      "loss": 0.411,
      "step": 148
    },
    {
      "epoch": 3.49,
      "grad_norm": 17.662681579589844,
      "learning_rate": 5.012202380952381e-06,
      "loss": 0.3776,
      "step": 149
    },
    {
      "epoch": 3.52,
      "grad_norm": 12.808351516723633,
      "learning_rate": 5.0089285714285716e-06,
      "loss": 0.3035,
      "step": 150
    },
    {
      "epoch": 3.54,
      "grad_norm": 30.4917049407959,
      "learning_rate": 5.0056547619047615e-06,
      "loss": 0.5275,
      "step": 151
    },
    {
      "epoch": 3.56,
      "grad_norm": 9.420913696289062,
      "learning_rate": 5.002380952380952e-06,
      "loss": 0.4193,
      "step": 152
    },
    {
      "epoch": 3.59,
      "grad_norm": 20.13055419921875,
      "learning_rate": 4.999107142857142e-06,
      "loss": 0.5197,
      "step": 153
    },
    {
      "epoch": 3.61,
      "grad_norm": 3.9424262046813965,
      "learning_rate": 4.995833333333333e-06,
      "loss": 0.3119,
      "step": 154
    },
    {
      "epoch": 3.63,
      "grad_norm": 4.024766445159912,
      "learning_rate": 4.992559523809524e-06,
      "loss": 0.3072,
      "step": 155
    },
    {
      "epoch": 3.66,
      "grad_norm": 16.67523956298828,
      "learning_rate": 4.989285714285714e-06,
      "loss": 0.4243,
      "step": 156
    },
    {
      "epoch": 3.68,
      "grad_norm": 14.94045352935791,
      "learning_rate": 4.986011904761905e-06,
      "loss": 0.4539,
      "step": 157
    },
    {
      "epoch": 3.7,
      "grad_norm": 14.900248527526855,
      "learning_rate": 4.9827380952380954e-06,
      "loss": 0.3834,
      "step": 158
    },
    {
      "epoch": 3.73,
      "grad_norm": 59.21234893798828,
      "learning_rate": 4.979464285714285e-06,
      "loss": 0.6183,
      "step": 159
    },
    {
      "epoch": 3.75,
      "grad_norm": 9.785917282104492,
      "learning_rate": 4.976190476190476e-06,
      "loss": 0.6158,
      "step": 160
    },
    {
      "epoch": 3.77,
      "grad_norm": 47.627864837646484,
      "learning_rate": 4.972916666666667e-06,
      "loss": 0.7627,
      "step": 161
    },
    {
      "epoch": 3.8,
      "grad_norm": 33.516265869140625,
      "learning_rate": 4.969642857142857e-06,
      "loss": 0.6755,
      "step": 162
    },
    {
      "epoch": 3.82,
      "grad_norm": 31.58948516845703,
      "learning_rate": 4.966369047619048e-06,
      "loss": 0.3291,
      "step": 163
    },
    {
      "epoch": 3.84,
      "grad_norm": 6.718880653381348,
      "learning_rate": 4.963095238095238e-06,
      "loss": 0.3233,
      "step": 164
    },
    {
      "epoch": 3.87,
      "grad_norm": 17.761388778686523,
      "learning_rate": 4.9598214285714285e-06,
      "loss": 0.332,
      "step": 165
    },
    {
      "epoch": 3.89,
      "grad_norm": 24.450733184814453,
      "learning_rate": 4.956547619047619e-06,
      "loss": 0.4093,
      "step": 166
    },
    {
      "epoch": 3.91,
      "grad_norm": 5.664928913116455,
      "learning_rate": 4.953273809523809e-06,
      "loss": 0.458,
      "step": 167
    },
    {
      "epoch": 3.94,
      "grad_norm": 8.065629005432129,
      "learning_rate": 4.95e-06,
      "loss": 0.503,
      "step": 168
    },
    {
      "epoch": 3.96,
      "grad_norm": 8.75709056854248,
      "learning_rate": 4.946726190476191e-06,
      "loss": 0.4183,
      "step": 169
    },
    {
      "epoch": 3.98,
      "grad_norm": 29.46023941040039,
      "learning_rate": 4.943452380952381e-06,
      "loss": 0.4187,
      "step": 170
    },
    {
      "epoch": 3.98,
      "eval_accuracy": 0.7727272727272727,
      "eval_loss": 0.48904669284820557,
      "eval_runtime": 1.961,
      "eval_samples_per_second": 112.185,
      "eval_steps_per_second": 28.046,
      "step": 170
    },
    {
      "epoch": 4.01,
      "grad_norm": 2.7368314266204834,
      "learning_rate": 4.940178571428572e-06,
      "loss": 0.4246,
      "step": 171
    },
    {
      "epoch": 4.03,
      "grad_norm": 4.692036151885986,
      "learning_rate": 4.9369047619047615e-06,
      "loss": 0.303,
      "step": 172
    },
    {
      "epoch": 4.05,
      "grad_norm": 21.547483444213867,
      "learning_rate": 4.933630952380952e-06,
      "loss": 0.3867,
      "step": 173
    },
    {
      "epoch": 4.08,
      "grad_norm": 8.194944381713867,
      "learning_rate": 4.930357142857143e-06,
      "loss": 0.1695,
      "step": 174
    },
    {
      "epoch": 4.1,
      "grad_norm": 3.5077579021453857,
      "learning_rate": 4.927083333333333e-06,
      "loss": 0.3142,
      "step": 175
    },
    {
      "epoch": 4.12,
      "grad_norm": 33.11520004272461,
      "learning_rate": 4.923809523809524e-06,
      "loss": 0.2809,
      "step": 176
    },
    {
      "epoch": 4.15,
      "grad_norm": 22.623079299926758,
      "learning_rate": 4.920535714285715e-06,
      "loss": 0.5169,
      "step": 177
    },
    {
      "epoch": 4.17,
      "grad_norm": 50.30414581298828,
      "learning_rate": 4.917261904761905e-06,
      "loss": 0.5075,
      "step": 178
    },
    {
      "epoch": 4.2,
      "grad_norm": 22.926681518554688,
      "learning_rate": 4.9139880952380955e-06,
      "loss": 0.3153,
      "step": 179
    },
    {
      "epoch": 4.22,
      "grad_norm": 4.504119396209717,
      "learning_rate": 4.910714285714286e-06,
      "loss": 0.3218,
      "step": 180
    },
    {
      "epoch": 4.24,
      "grad_norm": 26.036670684814453,
      "learning_rate": 4.907440476190476e-06,
      "loss": 0.4977,
      "step": 181
    },
    {
      "epoch": 4.27,
      "grad_norm": 3.1559319496154785,
      "learning_rate": 4.904166666666667e-06,
      "loss": 0.2435,
      "step": 182
    },
    {
      "epoch": 4.29,
      "grad_norm": 10.12259292602539,
      "learning_rate": 4.900892857142857e-06,
      "loss": 0.2531,
      "step": 183
    },
    {
      "epoch": 4.31,
      "grad_norm": 14.153977394104004,
      "learning_rate": 4.897619047619047e-06,
      "loss": 0.2939,
      "step": 184
    },
    {
      "epoch": 4.34,
      "grad_norm": 27.15022850036621,
      "learning_rate": 4.894345238095238e-06,
      "loss": 0.3503,
      "step": 185
    },
    {
      "epoch": 4.36,
      "grad_norm": 14.092424392700195,
      "learning_rate": 4.8910714285714285e-06,
      "loss": 0.2999,
      "step": 186
    },
    {
      "epoch": 4.38,
      "grad_norm": 21.638959884643555,
      "learning_rate": 4.8877976190476185e-06,
      "loss": 0.2351,
      "step": 187
    },
    {
      "epoch": 4.41,
      "grad_norm": 10.01887321472168,
      "learning_rate": 4.884523809523809e-06,
      "loss": 0.313,
      "step": 188
    },
    {
      "epoch": 4.43,
      "grad_norm": 13.962733268737793,
      "learning_rate": 4.881249999999999e-06,
      "loss": 0.3914,
      "step": 189
    },
    {
      "epoch": 4.45,
      "grad_norm": 12.105456352233887,
      "learning_rate": 4.87797619047619e-06,
      "loss": 0.3625,
      "step": 190
    },
    {
      "epoch": 4.48,
      "grad_norm": 11.165689468383789,
      "learning_rate": 4.874702380952381e-06,
      "loss": 0.425,
      "step": 191
    },
    {
      "epoch": 4.5,
      "grad_norm": 13.687613487243652,
      "learning_rate": 4.871428571428571e-06,
      "loss": 0.3409,
      "step": 192
    },
    {
      "epoch": 4.52,
      "grad_norm": 16.423450469970703,
      "learning_rate": 4.8681547619047616e-06,
      "loss": 0.4593,
      "step": 193
    },
    {
      "epoch": 4.55,
      "grad_norm": 10.233711242675781,
      "learning_rate": 4.864880952380952e-06,
      "loss": 0.2118,
      "step": 194
    },
    {
      "epoch": 4.57,
      "grad_norm": 9.759559631347656,
      "learning_rate": 4.861607142857142e-06,
      "loss": 0.3881,
      "step": 195
    },
    {
      "epoch": 4.59,
      "grad_norm": 16.481748580932617,
      "learning_rate": 4.858333333333333e-06,
      "loss": 0.3053,
      "step": 196
    },
    {
      "epoch": 4.62,
      "grad_norm": 17.1108455657959,
      "learning_rate": 4.855059523809524e-06,
      "loss": 0.4328,
      "step": 197
    },
    {
      "epoch": 4.64,
      "grad_norm": 18.17218589782715,
      "learning_rate": 4.851785714285714e-06,
      "loss": 0.3355,
      "step": 198
    },
    {
      "epoch": 4.66,
      "grad_norm": 10.150218963623047,
      "learning_rate": 4.848511904761905e-06,
      "loss": 0.244,
      "step": 199
    },
    {
      "epoch": 4.69,
      "grad_norm": 8.062572479248047,
      "learning_rate": 4.845238095238095e-06,
      "loss": 0.2057,
      "step": 200
    },
    {
      "epoch": 4.71,
      "grad_norm": 11.194631576538086,
      "learning_rate": 4.8419642857142854e-06,
      "loss": 0.2364,
      "step": 201
    },
    {
      "epoch": 4.73,
      "grad_norm": 14.064760208129883,
      "learning_rate": 4.838690476190476e-06,
      "loss": 0.2463,
      "step": 202
    },
    {
      "epoch": 4.76,
      "grad_norm": 18.526512145996094,
      "learning_rate": 4.835416666666666e-06,
      "loss": 0.2456,
      "step": 203
    },
    {
      "epoch": 4.78,
      "grad_norm": 17.577804565429688,
      "learning_rate": 4.832142857142857e-06,
      "loss": 0.2601,
      "step": 204
    },
    {
      "epoch": 4.8,
      "grad_norm": 11.843798637390137,
      "learning_rate": 4.828869047619048e-06,
      "loss": 0.1637,
      "step": 205
    },
    {
      "epoch": 4.83,
      "grad_norm": 12.153199195861816,
      "learning_rate": 4.825595238095238e-06,
      "loss": 0.1723,
      "step": 206
    },
    {
      "epoch": 4.85,
      "grad_norm": 10.96668529510498,
      "learning_rate": 4.8223214285714285e-06,
      "loss": 0.3315,
      "step": 207
    },
    {
      "epoch": 4.88,
      "grad_norm": 8.555688858032227,
      "learning_rate": 4.8190476190476185e-06,
      "loss": 0.2321,
      "step": 208
    },
    {
      "epoch": 4.9,
      "grad_norm": 12.00627326965332,
      "learning_rate": 4.815773809523809e-06,
      "loss": 0.2672,
      "step": 209
    },
    {
      "epoch": 4.92,
      "grad_norm": 22.342628479003906,
      "learning_rate": 4.8125e-06,
      "loss": 0.4607,
      "step": 210
    },
    {
      "epoch": 4.95,
      "grad_norm": 25.33961296081543,
      "learning_rate": 4.80922619047619e-06,
      "loss": 0.2068,
      "step": 211
    },
    {
      "epoch": 4.97,
      "grad_norm": 22.976959228515625,
      "learning_rate": 4.805952380952381e-06,
      "loss": 0.3447,
      "step": 212
    },
    {
      "epoch": 4.99,
      "grad_norm": 19.871631622314453,
      "learning_rate": 4.802678571428572e-06,
      "loss": 0.1563,
      "step": 213
    },
    {
      "epoch": 4.99,
      "eval_accuracy": 0.8545454545454545,
      "eval_loss": 0.4087750315666199,
      "eval_runtime": 1.9518,
      "eval_samples_per_second": 112.716,
      "eval_steps_per_second": 28.179,
      "step": 213
    },
    {
      "epoch": 5.02,
      "grad_norm": 6.280614852905273,
      "learning_rate": 4.799404761904762e-06,
      "loss": 0.1367,
      "step": 214
    },
    {
      "epoch": 5.04,
      "grad_norm": 18.553333282470703,
      "learning_rate": 4.796130952380952e-06,
      "loss": 0.3876,
      "step": 215
    },
    {
      "epoch": 5.06,
      "grad_norm": 10.598409652709961,
      "learning_rate": 4.792857142857143e-06,
      "loss": 0.2598,
      "step": 216
    },
    {
      "epoch": 5.09,
      "grad_norm": 28.292373657226562,
      "learning_rate": 4.789583333333333e-06,
      "loss": 0.1943,
      "step": 217
    },
    {
      "epoch": 5.11,
      "grad_norm": 29.251375198364258,
      "learning_rate": 4.786309523809524e-06,
      "loss": 0.3162,
      "step": 218
    },
    {
      "epoch": 5.13,
      "grad_norm": 54.16463088989258,
      "learning_rate": 4.783035714285714e-06,
      "loss": 0.3793,
      "step": 219
    },
    {
      "epoch": 5.16,
      "grad_norm": 13.832413673400879,
      "learning_rate": 4.779761904761905e-06,
      "loss": 0.134,
      "step": 220
    },
    {
      "epoch": 5.18,
      "grad_norm": 17.206539154052734,
      "learning_rate": 4.7764880952380955e-06,
      "loss": 0.3387,
      "step": 221
    },
    {
      "epoch": 5.2,
      "grad_norm": 22.453044891357422,
      "learning_rate": 4.7732142857142855e-06,
      "loss": 0.2228,
      "step": 222
    },
    {
      "epoch": 5.23,
      "grad_norm": 34.940757751464844,
      "learning_rate": 4.769940476190476e-06,
      "loss": 0.2947,
      "step": 223
    },
    {
      "epoch": 5.25,
      "grad_norm": 10.684017181396484,
      "learning_rate": 4.766666666666667e-06,
      "loss": 0.303,
      "step": 224
    },
    {
      "epoch": 5.27,
      "grad_norm": 12.175122261047363,
      "learning_rate": 4.763392857142857e-06,
      "loss": 0.2265,
      "step": 225
    },
    {
      "epoch": 5.3,
      "grad_norm": 7.9341139793396,
      "learning_rate": 4.760119047619048e-06,
      "loss": 0.2761,
      "step": 226
    },
    {
      "epoch": 5.32,
      "grad_norm": 28.212831497192383,
      "learning_rate": 4.756845238095238e-06,
      "loss": 0.3097,
      "step": 227
    },
    {
      "epoch": 5.34,
      "grad_norm": 31.275775909423828,
      "learning_rate": 4.7535714285714286e-06,
      "loss": 0.3684,
      "step": 228
    },
    {
      "epoch": 5.37,
      "grad_norm": 23.49386978149414,
      "learning_rate": 4.750297619047619e-06,
      "loss": 0.221,
      "step": 229
    },
    {
      "epoch": 5.39,
      "grad_norm": 25.68425941467285,
      "learning_rate": 4.747023809523809e-06,
      "loss": 0.6258,
      "step": 230
    },
    {
      "epoch": 5.41,
      "grad_norm": 25.499523162841797,
      "learning_rate": 4.74375e-06,
      "loss": 0.3263,
      "step": 231
    },
    {
      "epoch": 5.44,
      "grad_norm": 17.70143699645996,
      "learning_rate": 4.740476190476191e-06,
      "loss": 0.2089,
      "step": 232
    },
    {
      "epoch": 5.46,
      "grad_norm": 14.780723571777344,
      "learning_rate": 4.737202380952381e-06,
      "loss": 0.1825,
      "step": 233
    },
    {
      "epoch": 5.48,
      "grad_norm": 5.02520227432251,
      "learning_rate": 4.733928571428572e-06,
      "loss": 0.2236,
      "step": 234
    },
    {
      "epoch": 5.51,
      "grad_norm": 26.148418426513672,
      "learning_rate": 4.7306547619047625e-06,
      "loss": 0.2121,
      "step": 235
    },
    {
      "epoch": 5.53,
      "grad_norm": 36.72833251953125,
      "learning_rate": 4.727380952380952e-06,
      "loss": 0.4922,
      "step": 236
    },
    {
      "epoch": 5.55,
      "grad_norm": 16.5570125579834,
      "learning_rate": 4.724107142857142e-06,
      "loss": 0.25,
      "step": 237
    },
    {
      "epoch": 5.58,
      "grad_norm": 36.25120544433594,
      "learning_rate": 4.720833333333333e-06,
      "loss": 0.4475,
      "step": 238
    },
    {
      "epoch": 5.6,
      "grad_norm": 13.410614013671875,
      "learning_rate": 4.717559523809523e-06,
      "loss": 0.2065,
      "step": 239
    },
    {
      "epoch": 5.62,
      "grad_norm": 18.893632888793945,
      "learning_rate": 4.714285714285714e-06,
      "loss": 0.2649,
      "step": 240
    },
    {
      "epoch": 5.65,
      "grad_norm": 9.750955581665039,
      "learning_rate": 4.711011904761905e-06,
      "loss": 0.2054,
      "step": 241
    },
    {
      "epoch": 5.67,
      "grad_norm": 9.189886093139648,
      "learning_rate": 4.707738095238095e-06,
      "loss": 0.1694,
      "step": 242
    },
    {
      "epoch": 5.7,
      "grad_norm": 6.946423530578613,
      "learning_rate": 4.7044642857142855e-06,
      "loss": 0.186,
      "step": 243
    },
    {
      "epoch": 5.72,
      "grad_norm": 12.687536239624023,
      "learning_rate": 4.701190476190476e-06,
      "loss": 0.1277,
      "step": 244
    },
    {
      "epoch": 5.74,
      "grad_norm": 30.949155807495117,
      "learning_rate": 4.697916666666666e-06,
      "loss": 0.2875,
      "step": 245
    },
    {
      "epoch": 5.77,
      "grad_norm": 11.202219009399414,
      "learning_rate": 4.694642857142857e-06,
      "loss": 0.2124,
      "step": 246
    },
    {
      "epoch": 5.79,
      "grad_norm": 2.138655185699463,
      "learning_rate": 4.691369047619047e-06,
      "loss": 0.067,
      "step": 247
    },
    {
      "epoch": 5.81,
      "grad_norm": 10.137479782104492,
      "learning_rate": 4.688095238095238e-06,
      "loss": 0.1081,
      "step": 248
    },
    {
      "epoch": 5.84,
      "grad_norm": 10.702754020690918,
      "learning_rate": 4.684821428571429e-06,
      "loss": 0.1297,
      "step": 249
    },
    {
      "epoch": 5.86,
      "grad_norm": 11.548544883728027,
      "learning_rate": 4.6815476190476185e-06,
      "loss": 0.1005,
      "step": 250
    },
    {
      "epoch": 5.88,
      "grad_norm": 22.75307273864746,
      "learning_rate": 4.678273809523809e-06,
      "loss": 0.1593,
      "step": 251
    },
    {
      "epoch": 5.91,
      "grad_norm": 9.360298156738281,
      "learning_rate": 4.675e-06,
      "loss": 0.0642,
      "step": 252
    },
    {
      "epoch": 5.93,
      "grad_norm": 19.112125396728516,
      "learning_rate": 4.67172619047619e-06,
      "loss": 0.2293,
      "step": 253
    },
    {
      "epoch": 5.95,
      "grad_norm": 28.329204559326172,
      "learning_rate": 4.668452380952381e-06,
      "loss": 0.2197,
      "step": 254
    },
    {
      "epoch": 5.98,
      "grad_norm": 18.537385940551758,
      "learning_rate": 4.665178571428571e-06,
      "loss": 0.3401,
      "step": 255
    },
    {
      "epoch": 6.0,
      "grad_norm": 38.293357849121094,
      "learning_rate": 4.661904761904762e-06,
      "loss": 0.3536,
      "step": 256
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8454545454545455,
      "eval_loss": 0.488827645778656,
      "eval_runtime": 1.962,
      "eval_samples_per_second": 112.13,
      "eval_steps_per_second": 28.033,
      "step": 256
    },
    {
      "epoch": 6.02,
      "grad_norm": 4.5169901847839355,
      "learning_rate": 4.6586309523809524e-06,
      "loss": 0.0597,
      "step": 257
    },
    {
      "epoch": 6.05,
      "grad_norm": 8.103499412536621,
      "learning_rate": 4.655357142857142e-06,
      "loss": 0.0922,
      "step": 258
    },
    {
      "epoch": 6.07,
      "grad_norm": 24.916545867919922,
      "learning_rate": 4.652083333333333e-06,
      "loss": 0.1184,
      "step": 259
    },
    {
      "epoch": 6.09,
      "grad_norm": 20.035037994384766,
      "learning_rate": 4.648809523809524e-06,
      "loss": 0.2025,
      "step": 260
    },
    {
      "epoch": 6.12,
      "grad_norm": 6.929393291473389,
      "learning_rate": 4.645535714285714e-06,
      "loss": 0.2023,
      "step": 261
    },
    {
      "epoch": 6.14,
      "grad_norm": 15.74714183807373,
      "learning_rate": 4.642261904761905e-06,
      "loss": 0.2039,
      "step": 262
    },
    {
      "epoch": 6.16,
      "grad_norm": 10.057015419006348,
      "learning_rate": 4.6389880952380955e-06,
      "loss": 0.0961,
      "step": 263
    },
    {
      "epoch": 6.19,
      "grad_norm": 1.1386276483535767,
      "learning_rate": 4.6357142857142855e-06,
      "loss": 0.0461,
      "step": 264
    },
    {
      "epoch": 6.21,
      "grad_norm": 3.349055767059326,
      "learning_rate": 4.632440476190476e-06,
      "loss": 0.181,
      "step": 265
    },
    {
      "epoch": 6.23,
      "grad_norm": 47.79439163208008,
      "learning_rate": 4.629166666666666e-06,
      "loss": 0.1307,
      "step": 266
    },
    {
      "epoch": 6.26,
      "grad_norm": 8.228836059570312,
      "learning_rate": 4.625892857142857e-06,
      "loss": 0.2784,
      "step": 267
    },
    {
      "epoch": 6.28,
      "grad_norm": 9.270495414733887,
      "learning_rate": 4.622619047619048e-06,
      "loss": 0.1872,
      "step": 268
    },
    {
      "epoch": 6.3,
      "grad_norm": 0.9865421652793884,
      "learning_rate": 4.619345238095238e-06,
      "loss": 0.0396,
      "step": 269
    },
    {
      "epoch": 6.33,
      "grad_norm": 46.00654602050781,
      "learning_rate": 4.616071428571429e-06,
      "loss": 0.2206,
      "step": 270
    },
    {
      "epoch": 6.35,
      "grad_norm": 35.620689392089844,
      "learning_rate": 4.612797619047619e-06,
      "loss": 0.0977,
      "step": 271
    },
    {
      "epoch": 6.38,
      "grad_norm": 16.84827995300293,
      "learning_rate": 4.609523809523809e-06,
      "loss": 0.0652,
      "step": 272
    },
    {
      "epoch": 6.4,
      "grad_norm": 11.778346061706543,
      "learning_rate": 4.60625e-06,
      "loss": 0.0619,
      "step": 273
    },
    {
      "epoch": 6.42,
      "grad_norm": 12.670008659362793,
      "learning_rate": 4.60297619047619e-06,
      "loss": 0.1638,
      "step": 274
    },
    {
      "epoch": 6.45,
      "grad_norm": 3.050675392150879,
      "learning_rate": 4.599702380952381e-06,
      "loss": 0.1787,
      "step": 275
    },
    {
      "epoch": 6.47,
      "grad_norm": 9.351134300231934,
      "learning_rate": 4.596428571428572e-06,
      "loss": 0.2906,
      "step": 276
    },
    {
      "epoch": 6.49,
      "grad_norm": 18.061622619628906,
      "learning_rate": 4.593154761904762e-06,
      "loss": 0.1441,
      "step": 277
    },
    {
      "epoch": 6.52,
      "grad_norm": 5.385902404785156,
      "learning_rate": 4.5898809523809525e-06,
      "loss": 0.0326,
      "step": 278
    },
    {
      "epoch": 6.54,
      "grad_norm": 1.4941680431365967,
      "learning_rate": 4.586607142857143e-06,
      "loss": 0.0304,
      "step": 279
    },
    {
      "epoch": 6.56,
      "grad_norm": 8.055174827575684,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.3047,
      "step": 280
    },
    {
      "epoch": 6.59,
      "grad_norm": 11.676218032836914,
      "learning_rate": 4.580059523809524e-06,
      "loss": 0.4685,
      "step": 281
    },
    {
      "epoch": 6.61,
      "grad_norm": 3.277010917663574,
      "learning_rate": 4.576785714285715e-06,
      "loss": 0.0433,
      "step": 282
    },
    {
      "epoch": 6.63,
      "grad_norm": 14.3773832321167,
      "learning_rate": 4.573511904761905e-06,
      "loss": 0.3162,
      "step": 283
    },
    {
      "epoch": 6.66,
      "grad_norm": 5.165549278259277,
      "learning_rate": 4.5702380952380956e-06,
      "loss": 0.1855,
      "step": 284
    },
    {
      "epoch": 6.68,
      "grad_norm": 23.437685012817383,
      "learning_rate": 4.5669642857142855e-06,
      "loss": 0.12,
      "step": 285
    },
    {
      "epoch": 6.7,
      "grad_norm": 3.0389316082000732,
      "learning_rate": 4.563690476190476e-06,
      "loss": 0.0439,
      "step": 286
    },
    {
      "epoch": 6.73,
      "grad_norm": 35.046241760253906,
      "learning_rate": 4.560416666666667e-06,
      "loss": 0.1244,
      "step": 287
    },
    {
      "epoch": 6.75,
      "grad_norm": 28.730819702148438,
      "learning_rate": 4.557142857142857e-06,
      "loss": 0.2398,
      "step": 288
    },
    {
      "epoch": 6.77,
      "grad_norm": 19.567346572875977,
      "learning_rate": 4.553869047619047e-06,
      "loss": 0.0952,
      "step": 289
    },
    {
      "epoch": 6.8,
      "grad_norm": 17.299070358276367,
      "learning_rate": 4.550595238095238e-06,
      "loss": 0.2421,
      "step": 290
    },
    {
      "epoch": 6.82,
      "grad_norm": 4.0813422203063965,
      "learning_rate": 4.547321428571428e-06,
      "loss": 0.1717,
      "step": 291
    },
    {
      "epoch": 6.84,
      "grad_norm": 15.247352600097656,
      "learning_rate": 4.544047619047619e-06,
      "loss": 0.1287,
      "step": 292
    },
    {
      "epoch": 6.87,
      "grad_norm": 22.969430923461914,
      "learning_rate": 4.540773809523809e-06,
      "loss": 0.0944,
      "step": 293
    },
    {
      "epoch": 6.89,
      "grad_norm": 14.135355949401855,
      "learning_rate": 4.537499999999999e-06,
      "loss": 0.0501,
      "step": 294
    },
    {
      "epoch": 6.91,
      "grad_norm": 6.1480607986450195,
      "learning_rate": 4.53422619047619e-06,
      "loss": 0.0491,
      "step": 295
    },
    {
      "epoch": 6.94,
      "grad_norm": 5.824121952056885,
      "learning_rate": 4.530952380952381e-06,
      "loss": 0.4394,
      "step": 296
    },
    {
      "epoch": 6.96,
      "grad_norm": 3.691617727279663,
      "learning_rate": 4.527678571428571e-06,
      "loss": 0.0313,
      "step": 297
    },
    {
      "epoch": 6.98,
      "grad_norm": 1.4261701107025146,
      "learning_rate": 4.524404761904762e-06,
      "loss": 0.1709,
      "step": 298
    },
    {
      "epoch": 6.98,
      "eval_accuracy": 0.85,
      "eval_loss": 0.5375058054924011,
      "eval_runtime": 1.9645,
      "eval_samples_per_second": 111.989,
      "eval_steps_per_second": 27.997,
      "step": 298
    },
    {
      "epoch": 7.01,
      "grad_norm": 37.57842254638672,
      "learning_rate": 4.5211309523809525e-06,
      "loss": 0.299,
      "step": 299
    },
    {
      "epoch": 7.03,
      "grad_norm": 19.698734283447266,
      "learning_rate": 4.5178571428571424e-06,
      "loss": 0.3081,
      "step": 300
    },
    {
      "epoch": 7.05,
      "grad_norm": 0.6386685371398926,
      "learning_rate": 4.514583333333333e-06,
      "loss": 0.024,
      "step": 301
    },
    {
      "epoch": 7.08,
      "grad_norm": 1.1291331052780151,
      "learning_rate": 4.511309523809523e-06,
      "loss": 0.0343,
      "step": 302
    },
    {
      "epoch": 7.1,
      "grad_norm": 3.5354042053222656,
      "learning_rate": 4.508035714285714e-06,
      "loss": 0.1509,
      "step": 303
    },
    {
      "epoch": 7.12,
      "grad_norm": 0.7393876314163208,
      "learning_rate": 4.504761904761905e-06,
      "loss": 0.0237,
      "step": 304
    },
    {
      "epoch": 7.15,
      "grad_norm": 2.4936347007751465,
      "learning_rate": 4.501488095238095e-06,
      "loss": 0.0299,
      "step": 305
    },
    {
      "epoch": 7.17,
      "grad_norm": 28.873003005981445,
      "learning_rate": 4.4982142857142855e-06,
      "loss": 0.0815,
      "step": 306
    },
    {
      "epoch": 7.2,
      "grad_norm": 53.65907669067383,
      "learning_rate": 4.494940476190476e-06,
      "loss": 0.1271,
      "step": 307
    },
    {
      "epoch": 7.22,
      "grad_norm": 0.4335859715938568,
      "learning_rate": 4.491666666666666e-06,
      "loss": 0.0183,
      "step": 308
    },
    {
      "epoch": 7.24,
      "grad_norm": 2.026947259902954,
      "learning_rate": 4.488392857142857e-06,
      "loss": 0.1534,
      "step": 309
    },
    {
      "epoch": 7.27,
      "grad_norm": 0.4929918348789215,
      "learning_rate": 4.485119047619047e-06,
      "loss": 0.0192,
      "step": 310
    },
    {
      "epoch": 7.29,
      "grad_norm": 24.899736404418945,
      "learning_rate": 4.481845238095238e-06,
      "loss": 0.0524,
      "step": 311
    },
    {
      "epoch": 7.31,
      "grad_norm": 0.6321877837181091,
      "learning_rate": 4.478571428571429e-06,
      "loss": 0.0205,
      "step": 312
    },
    {
      "epoch": 7.34,
      "grad_norm": 1.5749492645263672,
      "learning_rate": 4.475297619047619e-06,
      "loss": 0.1671,
      "step": 313
    },
    {
      "epoch": 7.36,
      "grad_norm": 0.4469403624534607,
      "learning_rate": 4.472023809523809e-06,
      "loss": 0.0154,
      "step": 314
    },
    {
      "epoch": 7.38,
      "grad_norm": 10.251415252685547,
      "learning_rate": 4.46875e-06,
      "loss": 0.1924,
      "step": 315
    },
    {
      "epoch": 7.41,
      "grad_norm": 4.918264865875244,
      "learning_rate": 4.46547619047619e-06,
      "loss": 0.499,
      "step": 316
    },
    {
      "epoch": 7.43,
      "grad_norm": 0.3266950845718384,
      "learning_rate": 4.462202380952381e-06,
      "loss": 0.0131,
      "step": 317
    },
    {
      "epoch": 7.45,
      "grad_norm": 11.629830360412598,
      "learning_rate": 4.458928571428572e-06,
      "loss": 0.3543,
      "step": 318
    },
    {
      "epoch": 7.48,
      "grad_norm": 35.164634704589844,
      "learning_rate": 4.455654761904762e-06,
      "loss": 0.1511,
      "step": 319
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.5766786932945251,
      "learning_rate": 4.4523809523809525e-06,
      "loss": 0.0165,
      "step": 320
    },
    {
      "epoch": 7.52,
      "grad_norm": 30.383398056030273,
      "learning_rate": 4.4491071428571425e-06,
      "loss": 0.152,
      "step": 321
    },
    {
      "epoch": 7.55,
      "grad_norm": 26.80906867980957,
      "learning_rate": 4.445833333333333e-06,
      "loss": 0.2207,
      "step": 322
    },
    {
      "epoch": 7.57,
      "grad_norm": 0.6242391467094421,
      "learning_rate": 4.442559523809524e-06,
      "loss": 0.0129,
      "step": 323
    },
    {
      "epoch": 7.59,
      "grad_norm": 8.999675750732422,
      "learning_rate": 4.439285714285714e-06,
      "loss": 0.1625,
      "step": 324
    },
    {
      "epoch": 7.62,
      "grad_norm": 0.30194684863090515,
      "learning_rate": 4.436011904761905e-06,
      "loss": 0.0107,
      "step": 325
    },
    {
      "epoch": 7.64,
      "grad_norm": 11.341878890991211,
      "learning_rate": 4.432738095238096e-06,
      "loss": 0.1591,
      "step": 326
    },
    {
      "epoch": 7.66,
      "grad_norm": 1.6119003295898438,
      "learning_rate": 4.4294642857142856e-06,
      "loss": 0.0128,
      "step": 327
    },
    {
      "epoch": 7.69,
      "grad_norm": 58.44689178466797,
      "learning_rate": 4.426190476190476e-06,
      "loss": 0.1724,
      "step": 328
    },
    {
      "epoch": 7.71,
      "grad_norm": 1.108935832977295,
      "learning_rate": 4.422916666666666e-06,
      "loss": 0.0137,
      "step": 329
    },
    {
      "epoch": 7.73,
      "grad_norm": 0.3343135118484497,
      "learning_rate": 4.419642857142857e-06,
      "loss": 0.0115,
      "step": 330
    },
    {
      "epoch": 7.76,
      "grad_norm": 2.9117400646209717,
      "learning_rate": 4.416369047619048e-06,
      "loss": 0.0137,
      "step": 331
    },
    {
      "epoch": 7.78,
      "grad_norm": 0.3950476348400116,
      "learning_rate": 4.413095238095238e-06,
      "loss": 0.01,
      "step": 332
    },
    {
      "epoch": 7.8,
      "grad_norm": 0.46693718433380127,
      "learning_rate": 4.409821428571429e-06,
      "loss": 0.0109,
      "step": 333
    },
    {
      "epoch": 7.83,
      "grad_norm": 0.3062843680381775,
      "learning_rate": 4.4065476190476195e-06,
      "loss": 0.0082,
      "step": 334
    },
    {
      "epoch": 7.85,
      "grad_norm": 66.0848388671875,
      "learning_rate": 4.4032738095238094e-06,
      "loss": 0.0962,
      "step": 335
    },
    {
      "epoch": 7.88,
      "grad_norm": 0.5519910454750061,
      "learning_rate": 4.4e-06,
      "loss": 0.0084,
      "step": 336
    },
    {
      "epoch": 7.9,
      "grad_norm": 0.20682856440544128,
      "learning_rate": 4.396726190476191e-06,
      "loss": 0.0077,
      "step": 337
    },
    {
      "epoch": 7.92,
      "grad_norm": 23.140066146850586,
      "learning_rate": 4.393452380952381e-06,
      "loss": 0.0632,
      "step": 338
    },
    {
      "epoch": 7.95,
      "grad_norm": 1.7450965642929077,
      "learning_rate": 4.390178571428572e-06,
      "loss": 0.0097,
      "step": 339
    },
    {
      "epoch": 7.97,
      "grad_norm": 14.685700416564941,
      "learning_rate": 4.386904761904762e-06,
      "loss": 0.1616,
      "step": 340
    },
    {
      "epoch": 7.99,
      "grad_norm": 0.1476498693227768,
      "learning_rate": 4.3836309523809525e-06,
      "loss": 0.0067,
      "step": 341
    },
    {
      "epoch": 7.99,
      "eval_accuracy": 0.8863636363636364,
      "eval_loss": 0.5613145232200623,
      "eval_runtime": 1.9469,
      "eval_samples_per_second": 113.002,
      "eval_steps_per_second": 28.251,
      "step": 341
    },
    {
      "epoch": 8.02,
      "grad_norm": 0.2792944312095642,
      "learning_rate": 4.3803571428571425e-06,
      "loss": 0.0069,
      "step": 342
    },
    {
      "epoch": 8.04,
      "grad_norm": 0.5173578262329102,
      "learning_rate": 4.377083333333333e-06,
      "loss": 0.0073,
      "step": 343
    },
    {
      "epoch": 8.06,
      "grad_norm": 2.705310344696045,
      "learning_rate": 4.373809523809523e-06,
      "loss": 0.0116,
      "step": 344
    },
    {
      "epoch": 8.09,
      "grad_norm": 9.17689037322998,
      "learning_rate": 4.370535714285714e-06,
      "loss": 0.1767,
      "step": 345
    },
    {
      "epoch": 8.11,
      "grad_norm": 3.7868285179138184,
      "learning_rate": 4.367261904761904e-06,
      "loss": 0.0087,
      "step": 346
    },
    {
      "epoch": 8.13,
      "grad_norm": 0.3689323663711548,
      "learning_rate": 4.363988095238095e-06,
      "loss": 0.0061,
      "step": 347
    },
    {
      "epoch": 8.16,
      "grad_norm": 0.20344328880310059,
      "learning_rate": 4.360714285714286e-06,
      "loss": 0.0057,
      "step": 348
    },
    {
      "epoch": 8.18,
      "grad_norm": 37.69071578979492,
      "learning_rate": 4.3574404761904755e-06,
      "loss": 0.2596,
      "step": 349
    },
    {
      "epoch": 8.2,
      "grad_norm": 0.16801539063453674,
      "learning_rate": 4.354166666666666e-06,
      "loss": 0.0051,
      "step": 350
    },
    {
      "epoch": 8.23,
      "grad_norm": 5.060468673706055,
      "learning_rate": 4.350892857142857e-06,
      "loss": 0.2096,
      "step": 351
    },
    {
      "epoch": 8.25,
      "grad_norm": 1.643596887588501,
      "learning_rate": 4.347619047619047e-06,
      "loss": 0.0076,
      "step": 352
    },
    {
      "epoch": 8.27,
      "grad_norm": 0.7824279069900513,
      "learning_rate": 4.344345238095238e-06,
      "loss": 0.0059,
      "step": 353
    },
    {
      "epoch": 8.3,
      "grad_norm": 2.7976343631744385,
      "learning_rate": 4.341071428571429e-06,
      "loss": 0.0067,
      "step": 354
    },
    {
      "epoch": 8.32,
      "grad_norm": 13.010490417480469,
      "learning_rate": 4.337797619047619e-06,
      "loss": 0.3109,
      "step": 355
    },
    {
      "epoch": 8.34,
      "grad_norm": 0.10449977964162827,
      "learning_rate": 4.3345238095238094e-06,
      "loss": 0.0039,
      "step": 356
    },
    {
      "epoch": 8.37,
      "grad_norm": 0.4954240024089813,
      "learning_rate": 4.331249999999999e-06,
      "loss": 0.0064,
      "step": 357
    },
    {
      "epoch": 8.39,
      "grad_norm": 1.7399401664733887,
      "learning_rate": 4.32797619047619e-06,
      "loss": 0.0056,
      "step": 358
    },
    {
      "epoch": 8.41,
      "grad_norm": 7.1369147300720215,
      "learning_rate": 4.324702380952381e-06,
      "loss": 0.2385,
      "step": 359
    },
    {
      "epoch": 8.44,
      "grad_norm": 23.376176834106445,
      "learning_rate": 4.321428571428571e-06,
      "loss": 0.3003,
      "step": 360
    },
    {
      "epoch": 8.46,
      "grad_norm": 0.34425580501556396,
      "learning_rate": 4.318154761904762e-06,
      "loss": 0.0044,
      "step": 361
    },
    {
      "epoch": 8.48,
      "grad_norm": 0.21413090825080872,
      "learning_rate": 4.3148809523809526e-06,
      "loss": 0.004,
      "step": 362
    },
    {
      "epoch": 8.51,
      "grad_norm": 3.548433303833008,
      "learning_rate": 4.3116071428571425e-06,
      "loss": 0.123,
      "step": 363
    },
    {
      "epoch": 8.53,
      "grad_norm": 1.6423828601837158,
      "learning_rate": 4.308333333333333e-06,
      "loss": 0.2496,
      "step": 364
    },
    {
      "epoch": 8.55,
      "grad_norm": 2.283780813217163,
      "learning_rate": 4.305059523809523e-06,
      "loss": 0.2397,
      "step": 365
    },
    {
      "epoch": 8.58,
      "grad_norm": 0.362973153591156,
      "learning_rate": 4.301785714285714e-06,
      "loss": 0.0064,
      "step": 366
    },
    {
      "epoch": 8.6,
      "grad_norm": 19.265830993652344,
      "learning_rate": 4.298511904761905e-06,
      "loss": 0.0201,
      "step": 367
    },
    {
      "epoch": 8.62,
      "grad_norm": 0.20540741086006165,
      "learning_rate": 4.295238095238095e-06,
      "loss": 0.0047,
      "step": 368
    },
    {
      "epoch": 8.65,
      "grad_norm": 3.41279935836792,
      "learning_rate": 4.291964285714286e-06,
      "loss": 0.0088,
      "step": 369
    },
    {
      "epoch": 8.67,
      "grad_norm": 10.215970039367676,
      "learning_rate": 4.288690476190476e-06,
      "loss": 0.0162,
      "step": 370
    },
    {
      "epoch": 8.7,
      "grad_norm": 0.7728336453437805,
      "learning_rate": 4.285416666666666e-06,
      "loss": 0.008,
      "step": 371
    },
    {
      "epoch": 8.72,
      "grad_norm": 2.603362798690796,
      "learning_rate": 4.282142857142857e-06,
      "loss": 0.0122,
      "step": 372
    },
    {
      "epoch": 8.74,
      "grad_norm": 58.749488830566406,
      "learning_rate": 4.278869047619048e-06,
      "loss": 0.304,
      "step": 373
    },
    {
      "epoch": 8.77,
      "grad_norm": 8.352728843688965,
      "learning_rate": 4.275595238095238e-06,
      "loss": 0.1469,
      "step": 374
    },
    {
      "epoch": 8.79,
      "grad_norm": 0.2930574417114258,
      "learning_rate": 4.272321428571429e-06,
      "loss": 0.0067,
      "step": 375
    },
    {
      "epoch": 8.81,
      "grad_norm": 0.6395189166069031,
      "learning_rate": 4.269047619047619e-06,
      "loss": 0.0114,
      "step": 376
    },
    {
      "epoch": 8.84,
      "grad_norm": 21.4405517578125,
      "learning_rate": 4.2657738095238095e-06,
      "loss": 0.0245,
      "step": 377
    },
    {
      "epoch": 8.86,
      "grad_norm": 20.98697280883789,
      "learning_rate": 4.2625e-06,
      "loss": 0.2003,
      "step": 378
    },
    {
      "epoch": 8.88,
      "grad_norm": 0.16704310476779938,
      "learning_rate": 4.25922619047619e-06,
      "loss": 0.0045,
      "step": 379
    },
    {
      "epoch": 8.91,
      "grad_norm": 0.30824193358421326,
      "learning_rate": 4.255952380952381e-06,
      "loss": 0.0063,
      "step": 380
    },
    {
      "epoch": 8.93,
      "grad_norm": 38.134681701660156,
      "learning_rate": 4.252678571428572e-06,
      "loss": 0.0395,
      "step": 381
    },
    {
      "epoch": 8.95,
      "grad_norm": 0.32065123319625854,
      "learning_rate": 4.249404761904762e-06,
      "loss": 0.0055,
      "step": 382
    },
    {
      "epoch": 8.98,
      "grad_norm": 2.818316698074341,
      "learning_rate": 4.2461309523809526e-06,
      "loss": 0.0077,
      "step": 383
    },
    {
      "epoch": 9.0,
      "grad_norm": 8.902009010314941,
      "learning_rate": 4.2428571428571425e-06,
      "loss": 0.0103,
      "step": 384
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8772727272727273,
      "eval_loss": 0.6385995149612427,
      "eval_runtime": 1.9711,
      "eval_samples_per_second": 111.612,
      "eval_steps_per_second": 27.903,
      "step": 384
    },
    {
      "epoch": 9.02,
      "grad_norm": 50.5201301574707,
      "learning_rate": 4.239583333333333e-06,
      "loss": 0.1206,
      "step": 385
    },
    {
      "epoch": 9.05,
      "grad_norm": 0.10049311071634293,
      "learning_rate": 4.236309523809524e-06,
      "loss": 0.0038,
      "step": 386
    },
    {
      "epoch": 9.07,
      "grad_norm": 0.11850426346063614,
      "learning_rate": 4.233035714285714e-06,
      "loss": 0.0038,
      "step": 387
    },
    {
      "epoch": 9.09,
      "grad_norm": 0.10592693090438843,
      "learning_rate": 4.229761904761905e-06,
      "loss": 0.0037,
      "step": 388
    },
    {
      "epoch": 9.12,
      "grad_norm": 0.4136406481266022,
      "learning_rate": 4.226488095238096e-06,
      "loss": 0.0049,
      "step": 389
    },
    {
      "epoch": 9.14,
      "grad_norm": 0.09526017308235168,
      "learning_rate": 4.223214285714286e-06,
      "loss": 0.0034,
      "step": 390
    },
    {
      "epoch": 9.16,
      "grad_norm": 0.08932331204414368,
      "learning_rate": 4.2199404761904764e-06,
      "loss": 0.0034,
      "step": 391
    },
    {
      "epoch": 9.19,
      "grad_norm": 12.363450050354004,
      "learning_rate": 4.216666666666667e-06,
      "loss": 0.1276,
      "step": 392
    },
    {
      "epoch": 9.21,
      "grad_norm": 0.19276930391788483,
      "learning_rate": 4.213392857142857e-06,
      "loss": 0.0036,
      "step": 393
    },
    {
      "epoch": 9.23,
      "grad_norm": 0.10427949577569962,
      "learning_rate": 4.210119047619047e-06,
      "loss": 0.0041,
      "step": 394
    },
    {
      "epoch": 9.26,
      "grad_norm": 0.07943836599588394,
      "learning_rate": 4.206845238095238e-06,
      "loss": 0.0029,
      "step": 395
    },
    {
      "epoch": 9.28,
      "grad_norm": 0.18391607701778412,
      "learning_rate": 4.203571428571428e-06,
      "loss": 0.0036,
      "step": 396
    },
    {
      "epoch": 9.3,
      "grad_norm": 0.3423359990119934,
      "learning_rate": 4.200297619047619e-06,
      "loss": 0.0037,
      "step": 397
    },
    {
      "epoch": 9.33,
      "grad_norm": 41.31951904296875,
      "learning_rate": 4.1970238095238095e-06,
      "loss": 0.215,
      "step": 398
    },
    {
      "epoch": 9.35,
      "grad_norm": 0.1459679901599884,
      "learning_rate": 4.1937499999999994e-06,
      "loss": 0.0035,
      "step": 399
    },
    {
      "epoch": 9.38,
      "grad_norm": 0.0786140188574791,
      "learning_rate": 4.19047619047619e-06,
      "loss": 0.0027,
      "step": 400
    },
    {
      "epoch": 9.4,
      "grad_norm": 1.5436954498291016,
      "learning_rate": 4.18720238095238e-06,
      "loss": 0.086,
      "step": 401
    },
    {
      "epoch": 9.42,
      "grad_norm": 1.1575084924697876,
      "learning_rate": 4.183928571428571e-06,
      "loss": 0.0048,
      "step": 402
    },
    {
      "epoch": 9.45,
      "grad_norm": 0.07273612916469574,
      "learning_rate": 4.180654761904762e-06,
      "loss": 0.0027,
      "step": 403
    },
    {
      "epoch": 9.47,
      "grad_norm": 2.0359814167022705,
      "learning_rate": 4.177380952380952e-06,
      "loss": 0.2268,
      "step": 404
    },
    {
      "epoch": 9.49,
      "grad_norm": 0.06620334088802338,
      "learning_rate": 4.1741071428571426e-06,
      "loss": 0.0024,
      "step": 405
    },
    {
      "epoch": 9.52,
      "grad_norm": 0.9721193313598633,
      "learning_rate": 4.170833333333333e-06,
      "loss": 0.0038,
      "step": 406
    },
    {
      "epoch": 9.54,
      "grad_norm": 1.0129694938659668,
      "learning_rate": 4.167559523809523e-06,
      "loss": 0.0047,
      "step": 407
    },
    {
      "epoch": 9.56,
      "grad_norm": 2.4980669021606445,
      "learning_rate": 4.164285714285714e-06,
      "loss": 0.2284,
      "step": 408
    },
    {
      "epoch": 9.59,
      "grad_norm": 0.1482830047607422,
      "learning_rate": 4.161011904761905e-06,
      "loss": 0.0033,
      "step": 409
    },
    {
      "epoch": 9.61,
      "grad_norm": 96.48265075683594,
      "learning_rate": 4.157738095238095e-06,
      "loss": 0.0618,
      "step": 410
    },
    {
      "epoch": 9.63,
      "grad_norm": 0.1294354349374771,
      "learning_rate": 4.154464285714286e-06,
      "loss": 0.0031,
      "step": 411
    },
    {
      "epoch": 9.66,
      "grad_norm": 55.35332107543945,
      "learning_rate": 4.151190476190476e-06,
      "loss": 0.1618,
      "step": 412
    },
    {
      "epoch": 9.68,
      "grad_norm": 3.1412291526794434,
      "learning_rate": 4.147916666666666e-06,
      "loss": 0.1818,
      "step": 413
    },
    {
      "epoch": 9.7,
      "grad_norm": 2.8581583499908447,
      "learning_rate": 4.144642857142857e-06,
      "loss": 0.0895,
      "step": 414
    },
    {
      "epoch": 9.73,
      "grad_norm": 0.1566578447818756,
      "learning_rate": 4.141369047619047e-06,
      "loss": 0.004,
      "step": 415
    },
    {
      "epoch": 9.75,
      "grad_norm": 0.20300696790218353,
      "learning_rate": 4.138095238095238e-06,
      "loss": 0.0045,
      "step": 416
    },
    {
      "epoch": 9.77,
      "grad_norm": 28.00294303894043,
      "learning_rate": 4.134821428571429e-06,
      "loss": 0.0955,
      "step": 417
    },
    {
      "epoch": 9.8,
      "grad_norm": 4.9537553787231445,
      "learning_rate": 4.131547619047619e-06,
      "loss": 0.0108,
      "step": 418
    },
    {
      "epoch": 9.82,
      "grad_norm": 1.2664889097213745,
      "learning_rate": 4.1282738095238095e-06,
      "loss": 0.0144,
      "step": 419
    },
    {
      "epoch": 9.84,
      "grad_norm": 21.156566619873047,
      "learning_rate": 4.125e-06,
      "loss": 0.0984,
      "step": 420
    },
    {
      "epoch": 9.87,
      "grad_norm": 14.108423233032227,
      "learning_rate": 4.12172619047619e-06,
      "loss": 0.0181,
      "step": 421
    },
    {
      "epoch": 9.89,
      "grad_norm": 0.3327002227306366,
      "learning_rate": 4.118452380952381e-06,
      "loss": 0.0059,
      "step": 422
    },
    {
      "epoch": 9.91,
      "grad_norm": 0.26034343242645264,
      "learning_rate": 4.115178571428571e-06,
      "loss": 0.0047,
      "step": 423
    },
    {
      "epoch": 9.94,
      "grad_norm": 0.19648778438568115,
      "learning_rate": 4.111904761904762e-06,
      "loss": 0.0043,
      "step": 424
    },
    {
      "epoch": 9.96,
      "grad_norm": 95.51915740966797,
      "learning_rate": 4.108630952380953e-06,
      "loss": 0.0498,
      "step": 425
    },
    {
      "epoch": 9.98,
      "grad_norm": 0.3881193995475769,
      "learning_rate": 4.1053571428571426e-06,
      "loss": 0.0037,
      "step": 426
    },
    {
      "epoch": 9.98,
      "eval_accuracy": 0.8636363636363636,
      "eval_loss": 0.6795494556427002,
      "eval_runtime": 1.9632,
      "eval_samples_per_second": 112.063,
      "eval_steps_per_second": 28.016,
      "step": 426
    },
    {
      "epoch": 10.01,
      "grad_norm": 117.20142364501953,
      "learning_rate": 4.102083333333333e-06,
      "loss": 0.0613,
      "step": 427
    },
    {
      "epoch": 10.03,
      "grad_norm": 0.3005952537059784,
      "learning_rate": 4.098809523809524e-06,
      "loss": 0.0036,
      "step": 428
    },
    {
      "epoch": 10.05,
      "grad_norm": 0.22291749715805054,
      "learning_rate": 4.095535714285714e-06,
      "loss": 0.004,
      "step": 429
    },
    {
      "epoch": 10.08,
      "grad_norm": 0.16808579862117767,
      "learning_rate": 4.092261904761905e-06,
      "loss": 0.0035,
      "step": 430
    },
    {
      "epoch": 10.1,
      "grad_norm": 0.153959259390831,
      "learning_rate": 4.088988095238095e-06,
      "loss": 0.0036,
      "step": 431
    },
    {
      "epoch": 10.12,
      "grad_norm": 0.32211458683013916,
      "learning_rate": 4.085714285714286e-06,
      "loss": 0.0038,
      "step": 432
    },
    {
      "epoch": 10.15,
      "grad_norm": 1.331497311592102,
      "learning_rate": 4.0824404761904765e-06,
      "loss": 0.0044,
      "step": 433
    },
    {
      "epoch": 10.17,
      "grad_norm": 0.26392656564712524,
      "learning_rate": 4.0791666666666664e-06,
      "loss": 0.0032,
      "step": 434
    },
    {
      "epoch": 10.2,
      "grad_norm": 0.24153968691825867,
      "learning_rate": 4.075892857142857e-06,
      "loss": 0.0029,
      "step": 435
    },
    {
      "epoch": 10.22,
      "grad_norm": 52.36206817626953,
      "learning_rate": 4.072619047619048e-06,
      "loss": 0.0645,
      "step": 436
    },
    {
      "epoch": 10.24,
      "grad_norm": 0.12544922530651093,
      "learning_rate": 4.069345238095238e-06,
      "loss": 0.0025,
      "step": 437
    },
    {
      "epoch": 10.27,
      "grad_norm": 0.14890314638614655,
      "learning_rate": 4.066071428571429e-06,
      "loss": 0.0024,
      "step": 438
    },
    {
      "epoch": 10.29,
      "grad_norm": 0.050663262605667114,
      "learning_rate": 4.06279761904762e-06,
      "loss": 0.002,
      "step": 439
    },
    {
      "epoch": 10.31,
      "grad_norm": 0.07523107528686523,
      "learning_rate": 4.0595238095238095e-06,
      "loss": 0.0023,
      "step": 440
    },
    {
      "epoch": 10.34,
      "grad_norm": 0.04829227551817894,
      "learning_rate": 4.05625e-06,
      "loss": 0.0017,
      "step": 441
    },
    {
      "epoch": 10.36,
      "grad_norm": 0.0691770538687706,
      "learning_rate": 4.05297619047619e-06,
      "loss": 0.0023,
      "step": 442
    },
    {
      "epoch": 10.38,
      "grad_norm": 0.06425254791975021,
      "learning_rate": 4.049702380952381e-06,
      "loss": 0.002,
      "step": 443
    },
    {
      "epoch": 10.41,
      "grad_norm": 0.061768341809511185,
      "learning_rate": 4.046428571428572e-06,
      "loss": 0.0021,
      "step": 444
    },
    {
      "epoch": 10.43,
      "grad_norm": 2.5151658058166504,
      "learning_rate": 4.043154761904762e-06,
      "loss": 0.1136,
      "step": 445
    },
    {
      "epoch": 10.45,
      "grad_norm": 0.042002830654382706,
      "learning_rate": 4.039880952380953e-06,
      "loss": 0.0016,
      "step": 446
    },
    {
      "epoch": 10.48,
      "grad_norm": 0.064276322722435,
      "learning_rate": 4.036607142857143e-06,
      "loss": 0.0022,
      "step": 447
    },
    {
      "epoch": 10.5,
      "grad_norm": 0.06543342024087906,
      "learning_rate": 4.0333333333333326e-06,
      "loss": 0.0018,
      "step": 448
    },
    {
      "epoch": 10.52,
      "grad_norm": 0.04980894550681114,
      "learning_rate": 4.030059523809523e-06,
      "loss": 0.0017,
      "step": 449
    },
    {
      "epoch": 10.55,
      "grad_norm": 0.11392680555582047,
      "learning_rate": 4.026785714285714e-06,
      "loss": 0.0024,
      "step": 450
    },
    {
      "epoch": 10.57,
      "grad_norm": 0.15568284690380096,
      "learning_rate": 4.023511904761904e-06,
      "loss": 0.0021,
      "step": 451
    },
    {
      "epoch": 10.59,
      "grad_norm": 0.05342809110879898,
      "learning_rate": 4.020238095238095e-06,
      "loss": 0.0018,
      "step": 452
    },
    {
      "epoch": 10.62,
      "grad_norm": 9.609216690063477,
      "learning_rate": 4.016964285714286e-06,
      "loss": 0.1925,
      "step": 453
    },
    {
      "epoch": 10.64,
      "grad_norm": 18.55946922302246,
      "learning_rate": 4.013690476190476e-06,
      "loss": 0.0191,
      "step": 454
    },
    {
      "epoch": 10.66,
      "grad_norm": 1.1658024787902832,
      "learning_rate": 4.0104166666666665e-06,
      "loss": 0.0035,
      "step": 455
    },
    {
      "epoch": 10.69,
      "grad_norm": 0.062038108706474304,
      "learning_rate": 4.007142857142857e-06,
      "loss": 0.0017,
      "step": 456
    },
    {
      "epoch": 10.71,
      "grad_norm": 0.1960655152797699,
      "learning_rate": 4.003869047619047e-06,
      "loss": 0.0018,
      "step": 457
    },
    {
      "epoch": 10.73,
      "grad_norm": 0.9941339492797852,
      "learning_rate": 4.000595238095238e-06,
      "loss": 0.0802,
      "step": 458
    },
    {
      "epoch": 10.76,
      "grad_norm": 0.2770668864250183,
      "learning_rate": 3.997321428571428e-06,
      "loss": 0.002,
      "step": 459
    },
    {
      "epoch": 10.78,
      "grad_norm": 0.07205989211797714,
      "learning_rate": 3.994047619047619e-06,
      "loss": 0.0018,
      "step": 460
    },
    {
      "epoch": 10.8,
      "grad_norm": 0.17621174454689026,
      "learning_rate": 3.9907738095238096e-06,
      "loss": 0.0026,
      "step": 461
    },
    {
      "epoch": 10.83,
      "grad_norm": 0.6454585790634155,
      "learning_rate": 3.9874999999999995e-06,
      "loss": 0.0023,
      "step": 462
    },
    {
      "epoch": 10.85,
      "grad_norm": 0.9011671543121338,
      "learning_rate": 3.98422619047619e-06,
      "loss": 0.0775,
      "step": 463
    },
    {
      "epoch": 10.88,
      "grad_norm": 0.19559772312641144,
      "learning_rate": 3.980952380952381e-06,
      "loss": 0.0018,
      "step": 464
    },
    {
      "epoch": 10.9,
      "grad_norm": 0.06330468505620956,
      "learning_rate": 3.977678571428571e-06,
      "loss": 0.0016,
      "step": 465
    },
    {
      "epoch": 10.92,
      "grad_norm": 0.05253860726952553,
      "learning_rate": 3.974404761904762e-06,
      "loss": 0.0015,
      "step": 466
    },
    {
      "epoch": 10.95,
      "grad_norm": 0.16123701632022858,
      "learning_rate": 3.971130952380952e-06,
      "loss": 0.0019,
      "step": 467
    },
    {
      "epoch": 10.97,
      "grad_norm": 2.395909070968628,
      "learning_rate": 3.967857142857143e-06,
      "loss": 0.2863,
      "step": 468
    },
    {
      "epoch": 10.99,
      "grad_norm": 2.7706680297851562,
      "learning_rate": 3.964583333333333e-06,
      "loss": 0.2554,
      "step": 469
    },
    {
      "epoch": 10.99,
      "eval_accuracy": 0.8590909090909091,
      "eval_loss": 0.861154317855835,
      "eval_runtime": 1.9879,
      "eval_samples_per_second": 110.668,
      "eval_steps_per_second": 27.667,
      "step": 469
    },
    {
      "epoch": 11.02,
      "grad_norm": 0.11244534701108932,
      "learning_rate": 3.961309523809523e-06,
      "loss": 0.0019,
      "step": 470
    },
    {
      "epoch": 11.04,
      "grad_norm": 0.27306148409843445,
      "learning_rate": 3.958035714285714e-06,
      "loss": 0.0025,
      "step": 471
    },
    {
      "epoch": 11.06,
      "grad_norm": 0.06686495989561081,
      "learning_rate": 3.954761904761905e-06,
      "loss": 0.0017,
      "step": 472
    },
    {
      "epoch": 11.09,
      "grad_norm": 0.07426581531763077,
      "learning_rate": 3.951488095238095e-06,
      "loss": 0.002,
      "step": 473
    },
    {
      "epoch": 11.11,
      "grad_norm": 0.06276604533195496,
      "learning_rate": 3.948214285714286e-06,
      "loss": 0.0017,
      "step": 474
    },
    {
      "epoch": 11.13,
      "grad_norm": 114.83184051513672,
      "learning_rate": 3.9449404761904765e-06,
      "loss": 0.0446,
      "step": 475
    },
    {
      "epoch": 11.16,
      "grad_norm": 0.07929594069719315,
      "learning_rate": 3.9416666666666665e-06,
      "loss": 0.0019,
      "step": 476
    },
    {
      "epoch": 11.18,
      "grad_norm": 3.3116915225982666,
      "learning_rate": 3.938392857142857e-06,
      "loss": 0.2205,
      "step": 477
    },
    {
      "epoch": 11.2,
      "grad_norm": 0.2510410249233246,
      "learning_rate": 3.935119047619047e-06,
      "loss": 0.0028,
      "step": 478
    },
    {
      "epoch": 11.23,
      "grad_norm": 0.11125629395246506,
      "learning_rate": 3.931845238095238e-06,
      "loss": 0.0021,
      "step": 479
    },
    {
      "epoch": 11.25,
      "grad_norm": 0.33865758776664734,
      "learning_rate": 3.928571428571429e-06,
      "loss": 0.0036,
      "step": 480
    },
    {
      "epoch": 11.27,
      "grad_norm": 1.4806885719299316,
      "learning_rate": 3.925297619047619e-06,
      "loss": 0.0062,
      "step": 481
    },
    {
      "epoch": 11.3,
      "grad_norm": 0.4448486268520355,
      "learning_rate": 3.92202380952381e-06,
      "loss": 0.0026,
      "step": 482
    },
    {
      "epoch": 11.32,
      "grad_norm": 3.580939769744873,
      "learning_rate": 3.91875e-06,
      "loss": 0.1848,
      "step": 483
    },
    {
      "epoch": 11.34,
      "grad_norm": 0.07441940158605576,
      "learning_rate": 3.91547619047619e-06,
      "loss": 0.0019,
      "step": 484
    },
    {
      "epoch": 11.37,
      "grad_norm": 0.05166240781545639,
      "learning_rate": 3.912202380952381e-06,
      "loss": 0.0017,
      "step": 485
    },
    {
      "epoch": 11.39,
      "grad_norm": 0.08746666461229324,
      "learning_rate": 3.908928571428571e-06,
      "loss": 0.0021,
      "step": 486
    },
    {
      "epoch": 11.41,
      "grad_norm": 0.2311132699251175,
      "learning_rate": 3.905654761904762e-06,
      "loss": 0.0029,
      "step": 487
    },
    {
      "epoch": 11.44,
      "grad_norm": 0.10750468820333481,
      "learning_rate": 3.902380952380953e-06,
      "loss": 0.002,
      "step": 488
    },
    {
      "epoch": 11.46,
      "grad_norm": 0.29891398549079895,
      "learning_rate": 3.899107142857143e-06,
      "loss": 0.003,
      "step": 489
    },
    {
      "epoch": 11.48,
      "grad_norm": 0.13320636749267578,
      "learning_rate": 3.8958333333333334e-06,
      "loss": 0.0025,
      "step": 490
    },
    {
      "epoch": 11.51,
      "grad_norm": 0.11678657680749893,
      "learning_rate": 3.892559523809524e-06,
      "loss": 0.0028,
      "step": 491
    },
    {
      "epoch": 11.53,
      "grad_norm": 0.15591733157634735,
      "learning_rate": 3.889285714285714e-06,
      "loss": 0.0021,
      "step": 492
    },
    {
      "epoch": 11.55,
      "grad_norm": 35.02848434448242,
      "learning_rate": 3.886011904761905e-06,
      "loss": 0.0162,
      "step": 493
    },
    {
      "epoch": 11.58,
      "grad_norm": 14.46584415435791,
      "learning_rate": 3.882738095238096e-06,
      "loss": 0.0833,
      "step": 494
    },
    {
      "epoch": 11.6,
      "grad_norm": 0.19305245578289032,
      "learning_rate": 3.879464285714286e-06,
      "loss": 0.002,
      "step": 495
    },
    {
      "epoch": 11.62,
      "grad_norm": 0.814598023891449,
      "learning_rate": 3.8761904761904765e-06,
      "loss": 0.0697,
      "step": 496
    },
    {
      "epoch": 11.65,
      "grad_norm": 0.03925960883498192,
      "learning_rate": 3.8729166666666665e-06,
      "loss": 0.0015,
      "step": 497
    },
    {
      "epoch": 11.67,
      "grad_norm": 0.07029083371162415,
      "learning_rate": 3.869642857142857e-06,
      "loss": 0.0017,
      "step": 498
    },
    {
      "epoch": 11.7,
      "grad_norm": 0.048684656620025635,
      "learning_rate": 3.866369047619047e-06,
      "loss": 0.0016,
      "step": 499
    },
    {
      "epoch": 11.72,
      "grad_norm": 0.07016553729772568,
      "learning_rate": 3.863095238095238e-06,
      "loss": 0.002,
      "step": 500
    },
    {
      "epoch": 11.74,
      "grad_norm": 0.10889595001935959,
      "learning_rate": 3.859821428571428e-06,
      "loss": 0.0018,
      "step": 501
    },
    {
      "epoch": 11.77,
      "grad_norm": 0.04481940343976021,
      "learning_rate": 3.856547619047619e-06,
      "loss": 0.0015,
      "step": 502
    },
    {
      "epoch": 11.79,
      "grad_norm": 0.05054694786667824,
      "learning_rate": 3.853273809523809e-06,
      "loss": 0.0016,
      "step": 503
    },
    {
      "epoch": 11.81,
      "grad_norm": 0.07786347717046738,
      "learning_rate": 3.8499999999999996e-06,
      "loss": 0.0015,
      "step": 504
    },
    {
      "epoch": 11.84,
      "grad_norm": 0.10747917741537094,
      "learning_rate": 3.84672619047619e-06,
      "loss": 0.002,
      "step": 505
    },
    {
      "epoch": 11.86,
      "grad_norm": 3.685096263885498,
      "learning_rate": 3.84345238095238e-06,
      "loss": 0.0047,
      "step": 506
    },
    {
      "epoch": 11.88,
      "grad_norm": 0.8624734878540039,
      "learning_rate": 3.840178571428571e-06,
      "loss": 0.0707,
      "step": 507
    },
    {
      "epoch": 11.91,
      "grad_norm": 0.03694400563836098,
      "learning_rate": 3.836904761904762e-06,
      "loss": 0.0014,
      "step": 508
    },
    {
      "epoch": 11.93,
      "grad_norm": 0.08262762427330017,
      "learning_rate": 3.833630952380952e-06,
      "loss": 0.0015,
      "step": 509
    },
    {
      "epoch": 11.95,
      "grad_norm": 0.06174321845173836,
      "learning_rate": 3.830357142857143e-06,
      "loss": 0.0014,
      "step": 510
    },
    {
      "epoch": 11.98,
      "grad_norm": 0.04213571548461914,
      "learning_rate": 3.8270833333333335e-06,
      "loss": 0.0013,
      "step": 511
    },
    {
      "epoch": 12.0,
      "grad_norm": 4.956207275390625,
      "learning_rate": 3.823809523809523e-06,
      "loss": 0.2121,
      "step": 512
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8909090909090909,
      "eval_loss": 0.7056229114532471,
      "eval_runtime": 1.9577,
      "eval_samples_per_second": 112.375,
      "eval_steps_per_second": 28.094,
      "step": 512
    }
  ],
  "logging_steps": 1,
  "max_steps": 1680,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 500,
  "total_flos": 5202308973417984.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}