{
  "best_global_step": 2800,
  "best_metric": 0.9999996161228406,
  "best_model_checkpoint": "/projects/bffw/darora1/llm_ipc/final_models/fj_n3/checkpoint-2800",
  "epoch": 0.9073417721518987,
  "eval_steps": 40,
  "global_step": 2800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0006481012658227848,
      "grad_norm": 52.902732849121094,
      "learning_rate": 2.0000000000000002e-07,
      "loss": 2.8041,
      "step": 2
    },
    {
      "epoch": 0.0012962025316455696,
      "grad_norm": 52.87858200073242,
      "learning_rate": 6.000000000000001e-07,
      "loss": 3.0228,
      "step": 4
    },
    {
      "epoch": 0.0019443037974683545,
      "grad_norm": 59.53159713745117,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 3.0448,
      "step": 6
    },
    {
      "epoch": 0.002592405063291139,
      "grad_norm": 41.063140869140625,
      "learning_rate": 1.4000000000000001e-06,
      "loss": 2.5576,
      "step": 8
    },
    {
      "epoch": 0.0032405063291139242,
      "grad_norm": 46.71913528442383,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 2.5334,
      "step": 10
    },
    {
      "epoch": 0.003888607594936709,
      "grad_norm": 38.23870086669922,
      "learning_rate": 2.2e-06,
      "loss": 2.0494,
      "step": 12
    },
    {
      "epoch": 0.004536708860759494,
      "grad_norm": 26.98814582824707,
      "learning_rate": 2.6e-06,
      "loss": 1.6988,
      "step": 14
    },
    {
      "epoch": 0.005184810126582278,
      "grad_norm": 24.897260665893555,
      "learning_rate": 3e-06,
      "loss": 1.4495,
      "step": 16
    },
    {
      "epoch": 0.005832911392405063,
      "grad_norm": 24.14356231689453,
      "learning_rate": 3.4000000000000005e-06,
      "loss": 1.2273,
      "step": 18
    },
    {
      "epoch": 0.0064810126582278485,
      "grad_norm": 20.519506454467773,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 0.5886,
      "step": 20
    },
    {
      "epoch": 0.007129113924050633,
      "grad_norm": 13.519320487976074,
      "learning_rate": 4.2000000000000004e-06,
      "loss": 0.3763,
      "step": 22
    },
    {
      "epoch": 0.007777215189873418,
      "grad_norm": 5.407296657562256,
      "learning_rate": 4.600000000000001e-06,
      "loss": 0.2153,
      "step": 24
    },
    {
      "epoch": 0.008425316455696203,
      "grad_norm": 12.552337646484375,
      "learning_rate": 5e-06,
      "loss": 0.166,
      "step": 26
    },
    {
      "epoch": 0.009073417721518987,
      "grad_norm": 10.085341453552246,
      "learning_rate": 5.400000000000001e-06,
      "loss": 0.1352,
      "step": 28
    },
    {
      "epoch": 0.009721518987341773,
      "grad_norm": 9.081425666809082,
      "learning_rate": 5.8e-06,
      "loss": 0.1306,
      "step": 30
    },
    {
      "epoch": 0.010369620253164557,
      "grad_norm": 3.1638600826263428,
      "learning_rate": 6.200000000000001e-06,
      "loss": 0.1188,
      "step": 32
    },
    {
      "epoch": 0.011017721518987342,
      "grad_norm": 2.7224133014678955,
      "learning_rate": 6.600000000000001e-06,
      "loss": 0.1128,
      "step": 34
    },
    {
      "epoch": 0.011665822784810126,
      "grad_norm": 2.1373937129974365,
      "learning_rate": 7e-06,
      "loss": 0.0943,
      "step": 36
    },
    {
      "epoch": 0.012313924050632911,
      "grad_norm": 2.3295581340789795,
      "learning_rate": 7.4e-06,
      "loss": 0.0962,
      "step": 38
    },
    {
      "epoch": 0.012962025316455697,
      "grad_norm": 3.900003671646118,
      "learning_rate": 7.800000000000002e-06,
      "loss": 0.0952,
      "step": 40
    },
    {
      "epoch": 0.012962025316455697,
      "eval_accuracy": 0.9631154472461956,
      "eval_loss": 0.09466289728879929,
      "eval_runtime": 99.4408,
      "eval_samples_per_second": 50.281,
      "eval_steps_per_second": 12.57,
      "step": 40
    },
    {
      "epoch": 0.01361012658227848,
      "grad_norm": 1.0516443252563477,
      "learning_rate": 8.2e-06,
      "loss": 0.0836,
      "step": 42
    },
    {
      "epoch": 0.014258227848101266,
      "grad_norm": 3.0402073860168457,
      "learning_rate": 8.6e-06,
      "loss": 0.0763,
      "step": 44
    },
    {
      "epoch": 0.01490632911392405,
      "grad_norm": 3.903961181640625,
      "learning_rate": 9e-06,
      "loss": 0.0685,
      "step": 46
    },
    {
      "epoch": 0.015554430379746836,
      "grad_norm": 4.473770618438721,
      "learning_rate": 9.4e-06,
      "loss": 0.0843,
      "step": 48
    },
    {
      "epoch": 0.01620253164556962,
      "grad_norm": 1.465899109840393,
      "learning_rate": 9.800000000000001e-06,
      "loss": 0.0734,
      "step": 50
    },
    {
      "epoch": 0.016850632911392407,
      "grad_norm": 1.6887052059173584,
      "learning_rate": 9.999998993000299e-06,
      "loss": 0.07,
      "step": 52
    },
    {
      "epoch": 0.01749873417721519,
      "grad_norm": 2.927128553390503,
      "learning_rate": 9.999990937005126e-06,
      "loss": 0.0653,
      "step": 54
    },
    {
      "epoch": 0.018146835443037974,
      "grad_norm": 0.7761722207069397,
      "learning_rate": 9.999974825027756e-06,
      "loss": 0.0631,
      "step": 56
    },
    {
      "epoch": 0.01879493670886076,
      "grad_norm": 1.8903884887695312,
      "learning_rate": 9.999950657094151e-06,
      "loss": 0.0647,
      "step": 58
    },
    {
      "epoch": 0.019443037974683545,
      "grad_norm": 3.654014825820923,
      "learning_rate": 9.999918433243253e-06,
      "loss": 0.0654,
      "step": 60
    },
    {
      "epoch": 0.020091139240506328,
      "grad_norm": 1.6327952146530151,
      "learning_rate": 9.999878153526974e-06,
      "loss": 0.0755,
      "step": 62
    },
    {
      "epoch": 0.020739240506329113,
      "grad_norm": 1.149726390838623,
      "learning_rate": 9.99982981801022e-06,
      "loss": 0.0588,
      "step": 64
    },
    {
      "epoch": 0.0213873417721519,
      "grad_norm": 1.9310941696166992,
      "learning_rate": 9.999773426770864e-06,
      "loss": 0.0658,
      "step": 66
    },
    {
      "epoch": 0.022035443037974684,
      "grad_norm": 0.8754224181175232,
      "learning_rate": 9.999708979899769e-06,
      "loss": 0.049,
      "step": 68
    },
    {
      "epoch": 0.02268354430379747,
      "grad_norm": 0.8598536252975464,
      "learning_rate": 9.999636477500765e-06,
      "loss": 0.055,
      "step": 70
    },
    {
      "epoch": 0.023331645569620252,
      "grad_norm": 1.237684726715088,
      "learning_rate": 9.999555919690673e-06,
      "loss": 0.06,
      "step": 72
    },
    {
      "epoch": 0.023979746835443037,
      "grad_norm": 0.7132460474967957,
      "learning_rate": 9.999467306599285e-06,
      "loss": 0.0571,
      "step": 74
    },
    {
      "epoch": 0.024627848101265823,
      "grad_norm": 1.8192301988601685,
      "learning_rate": 9.999370638369377e-06,
      "loss": 0.0578,
      "step": 76
    },
    {
      "epoch": 0.02527594936708861,
      "grad_norm": 1.6354628801345825,
      "learning_rate": 9.999265915156697e-06,
      "loss": 0.0661,
      "step": 78
    },
    {
      "epoch": 0.025924050632911394,
      "grad_norm": 0.6722614169120789,
      "learning_rate": 9.999153137129978e-06,
      "loss": 0.0609,
      "step": 80
    },
    {
      "epoch": 0.025924050632911394,
      "eval_accuracy": 0.9742360751297139,
      "eval_loss": 0.059016820043325424,
      "eval_runtime": 100.1177,
      "eval_samples_per_second": 49.941,
      "eval_steps_per_second": 12.485,
      "step": 80
    },
    {
      "epoch": 0.026572151898734176,
      "grad_norm": 0.7650524377822876,
      "learning_rate": 9.999032304470926e-06,
      "loss": 0.0502,
      "step": 82
    },
    {
      "epoch": 0.02722025316455696,
      "grad_norm": 1.3773459196090698,
      "learning_rate": 9.998903417374228e-06,
      "loss": 0.0534,
      "step": 84
    },
    {
      "epoch": 0.027868354430379747,
      "grad_norm": 0.6049540638923645,
      "learning_rate": 9.998766476047546e-06,
      "loss": 0.042,
      "step": 86
    },
    {
      "epoch": 0.028516455696202533,
      "grad_norm": 0.8149325847625732,
      "learning_rate": 9.998621480711522e-06,
      "loss": 0.0475,
      "step": 88
    },
    {
      "epoch": 0.029164556962025315,
      "grad_norm": 0.7447686791419983,
      "learning_rate": 9.998468431599768e-06,
      "loss": 0.045,
      "step": 90
    },
    {
      "epoch": 0.0298126582278481,
      "grad_norm": 1.151421308517456,
      "learning_rate": 9.99830732895888e-06,
      "loss": 0.0508,
      "step": 92
    },
    {
      "epoch": 0.030460759493670886,
      "grad_norm": 0.9233959913253784,
      "learning_rate": 9.998138173048424e-06,
      "loss": 0.0445,
      "step": 94
    },
    {
      "epoch": 0.03110886075949367,
      "grad_norm": 1.1906189918518066,
      "learning_rate": 9.997960964140946e-06,
      "loss": 0.0494,
      "step": 96
    },
    {
      "epoch": 0.03175696202531646,
      "grad_norm": 1.164791226387024,
      "learning_rate": 9.997775702521965e-06,
      "loss": 0.0469,
      "step": 98
    },
    {
      "epoch": 0.03240506329113924,
      "grad_norm": 0.9656791090965271,
      "learning_rate": 9.997582388489975e-06,
      "loss": 0.0473,
      "step": 100
    },
    {
      "epoch": 0.03305316455696203,
      "grad_norm": 0.9113268256187439,
      "learning_rate": 9.99738102235644e-06,
      "loss": 0.0414,
      "step": 102
    },
    {
      "epoch": 0.033701265822784814,
      "grad_norm": 0.8496657609939575,
      "learning_rate": 9.997171604445803e-06,
      "loss": 0.0465,
      "step": 104
    },
    {
      "epoch": 0.03434936708860759,
      "grad_norm": 0.7116028666496277,
      "learning_rate": 9.99695413509548e-06,
      "loss": 0.042,
      "step": 106
    },
    {
      "epoch": 0.03499746835443038,
      "grad_norm": 0.8646735548973083,
      "learning_rate": 9.996728614655854e-06,
      "loss": 0.0437,
      "step": 108
    },
    {
      "epoch": 0.03564556962025316,
      "grad_norm": 0.4896007180213928,
      "learning_rate": 9.996495043490285e-06,
      "loss": 0.0367,
      "step": 110
    },
    {
      "epoch": 0.03629367088607595,
      "grad_norm": 0.7575715184211731,
      "learning_rate": 9.996253421975103e-06,
      "loss": 0.0414,
      "step": 112
    },
    {
      "epoch": 0.036941772151898734,
      "grad_norm": 0.6563818454742432,
      "learning_rate": 9.996003750499608e-06,
      "loss": 0.0406,
      "step": 114
    },
    {
      "epoch": 0.03758987341772152,
      "grad_norm": 0.7096058130264282,
      "learning_rate": 9.995746029466071e-06,
      "loss": 0.042,
      "step": 116
    },
    {
      "epoch": 0.038237974683544305,
      "grad_norm": 0.8628009557723999,
      "learning_rate": 9.995480259289731e-06,
      "loss": 0.043,
      "step": 118
    },
    {
      "epoch": 0.03888607594936709,
      "grad_norm": 0.9108039140701294,
      "learning_rate": 9.995206440398798e-06,
      "loss": 0.0401,
      "step": 120
    },
    {
      "epoch": 0.03888607594936709,
      "eval_accuracy": 0.9794015821356299,
      "eval_loss": 0.04449129104614258,
      "eval_runtime": 97.7024,
      "eval_samples_per_second": 51.176,
      "eval_steps_per_second": 12.794,
      "step": 120
    },
    {
      "epoch": 0.03953417721518988,
      "grad_norm": 1.005293846130371,
      "learning_rate": 9.994924573234448e-06,
      "loss": 0.0407,
      "step": 122
    },
    {
      "epoch": 0.040182278481012655,
      "grad_norm": 1.1099708080291748,
      "learning_rate": 9.994634658250825e-06,
      "loss": 0.0433,
      "step": 124
    },
    {
      "epoch": 0.04083037974683544,
      "grad_norm": 0.9993194341659546,
      "learning_rate": 9.994336695915041e-06,
      "loss": 0.0389,
      "step": 126
    },
    {
      "epoch": 0.041478481012658226,
      "grad_norm": 0.6605297923088074,
      "learning_rate": 9.994030686707171e-06,
      "loss": 0.0378,
      "step": 128
    },
    {
      "epoch": 0.04212658227848101,
      "grad_norm": 0.5856267809867859,
      "learning_rate": 9.993716631120259e-06,
      "loss": 0.0432,
      "step": 130
    },
    {
      "epoch": 0.0427746835443038,
      "grad_norm": 0.6284976601600647,
      "learning_rate": 9.993394529660307e-06,
      "loss": 0.044,
      "step": 132
    },
    {
      "epoch": 0.04342278481012658,
      "grad_norm": 0.7984368801116943,
      "learning_rate": 9.99306438284629e-06,
      "loss": 0.0418,
      "step": 134
    },
    {
      "epoch": 0.04407088607594937,
      "grad_norm": 0.9223721027374268,
      "learning_rate": 9.992726191210139e-06,
      "loss": 0.038,
      "step": 136
    },
    {
      "epoch": 0.044718987341772154,
      "grad_norm": 0.5600959062576294,
      "learning_rate": 9.992379955296745e-06,
      "loss": 0.039,
      "step": 138
    },
    {
      "epoch": 0.04536708860759494,
      "grad_norm": 0.6998872756958008,
      "learning_rate": 9.992025675663966e-06,
      "loss": 0.036,
      "step": 140
    },
    {
      "epoch": 0.04601518987341772,
      "grad_norm": 0.5963979363441467,
      "learning_rate": 9.991663352882615e-06,
      "loss": 0.0379,
      "step": 142
    },
    {
      "epoch": 0.046663291139240504,
      "grad_norm": 0.7889736890792847,
      "learning_rate": 9.991292987536469e-06,
      "loss": 0.0379,
      "step": 144
    },
    {
      "epoch": 0.04731139240506329,
      "grad_norm": 0.5946393013000488,
      "learning_rate": 9.990914580222258e-06,
      "loss": 0.0348,
      "step": 146
    },
    {
      "epoch": 0.047959493670886075,
      "grad_norm": 0.45124557614326477,
      "learning_rate": 9.990528131549674e-06,
      "loss": 0.0385,
      "step": 148
    },
    {
      "epoch": 0.04860759493670886,
      "grad_norm": 0.6986305117607117,
      "learning_rate": 9.990133642141359e-06,
      "loss": 0.0371,
      "step": 150
    },
    {
      "epoch": 0.049255696202531646,
      "grad_norm": 0.9078485369682312,
      "learning_rate": 9.989731112632917e-06,
      "loss": 0.0327,
      "step": 152
    },
    {
      "epoch": 0.04990379746835443,
      "grad_norm": 0.8524628281593323,
      "learning_rate": 9.989320543672904e-06,
      "loss": 0.0384,
      "step": 154
    },
    {
      "epoch": 0.05055189873417722,
      "grad_norm": 0.9344510436058044,
      "learning_rate": 9.988901935922826e-06,
      "loss": 0.0328,
      "step": 156
    },
    {
      "epoch": 0.0512,
      "grad_norm": 0.7619643211364746,
      "learning_rate": 9.988475290057145e-06,
      "loss": 0.033,
      "step": 158
    },
    {
      "epoch": 0.05184810126582279,
      "grad_norm": 0.4197481572628021,
      "learning_rate": 9.988040606763272e-06,
      "loss": 0.0336,
      "step": 160
    },
    {
      "epoch": 0.05184810126582279,
      "eval_accuracy": 0.9819056194501201,
      "eval_loss": 0.03950502350926399,
      "eval_runtime": 98.6835,
      "eval_samples_per_second": 50.667,
      "eval_steps_per_second": 12.667,
      "step": 160
    },
    {
      "epoch": 0.05249620253164557,
      "grad_norm": 0.5393734574317932,
      "learning_rate": 9.98759788674157e-06,
      "loss": 0.0352,
      "step": 162
    },
    {
      "epoch": 0.05314430379746835,
      "grad_norm": 0.7958114743232727,
      "learning_rate": 9.987147130705347e-06,
      "loss": 0.0352,
      "step": 164
    },
    {
      "epoch": 0.05379240506329114,
      "grad_norm": 0.48281773924827576,
      "learning_rate": 9.986688339380863e-06,
      "loss": 0.0355,
      "step": 166
    },
    {
      "epoch": 0.05444050632911392,
      "grad_norm": 0.6643773317337036,
      "learning_rate": 9.98622151350732e-06,
      "loss": 0.0322,
      "step": 168
    },
    {
      "epoch": 0.05508860759493671,
      "grad_norm": 0.4064873158931732,
      "learning_rate": 9.985746653836867e-06,
      "loss": 0.0296,
      "step": 170
    },
    {
      "epoch": 0.055736708860759494,
      "grad_norm": 0.4580860137939453,
      "learning_rate": 9.985263761134602e-06,
      "loss": 0.037,
      "step": 172
    },
    {
      "epoch": 0.05638481012658228,
      "grad_norm": 0.9311563372612,
      "learning_rate": 9.984772836178559e-06,
      "loss": 0.0308,
      "step": 174
    },
    {
      "epoch": 0.057032911392405065,
      "grad_norm": 0.7613935470581055,
      "learning_rate": 9.984273879759713e-06,
      "loss": 0.03,
      "step": 176
    },
    {
      "epoch": 0.05768101265822785,
      "grad_norm": 0.5291257500648499,
      "learning_rate": 9.983766892681985e-06,
      "loss": 0.0345,
      "step": 178
    },
    {
      "epoch": 0.05832911392405063,
      "grad_norm": 0.6109746098518372,
      "learning_rate": 9.983251875762234e-06,
      "loss": 0.0368,
      "step": 180
    },
    {
      "epoch": 0.058977215189873415,
      "grad_norm": 0.6333370804786682,
      "learning_rate": 9.982728829830252e-06,
      "loss": 0.0321,
      "step": 182
    },
    {
      "epoch": 0.0596253164556962,
      "grad_norm": 0.6072647571563721,
      "learning_rate": 9.982197755728771e-06,
      "loss": 0.0268,
      "step": 184
    },
    {
      "epoch": 0.060273417721518986,
      "grad_norm": 0.5882136821746826,
      "learning_rate": 9.981658654313458e-06,
      "loss": 0.0354,
      "step": 186
    },
    {
      "epoch": 0.06092151898734177,
      "grad_norm": 0.742362380027771,
      "learning_rate": 9.981111526452912e-06,
      "loss": 0.0333,
      "step": 188
    },
    {
      "epoch": 0.06156962025316456,
      "grad_norm": 0.6676820516586304,
      "learning_rate": 9.980556373028665e-06,
      "loss": 0.0319,
      "step": 190
    },
    {
      "epoch": 0.06221772151898734,
      "grad_norm": 0.6671105623245239,
      "learning_rate": 9.979993194935182e-06,
      "loss": 0.0307,
      "step": 192
    },
    {
      "epoch": 0.06286582278481012,
      "grad_norm": 0.7644891142845154,
      "learning_rate": 9.979421993079853e-06,
      "loss": 0.0346,
      "step": 194
    },
    {
      "epoch": 0.06351392405063291,
      "grad_norm": 0.4839600920677185,
      "learning_rate": 9.978842768382999e-06,
      "loss": 0.0312,
      "step": 196
    },
    {
      "epoch": 0.06416202531645569,
      "grad_norm": 0.859524130821228,
      "learning_rate": 9.978255521777865e-06,
      "loss": 0.0317,
      "step": 198
    },
    {
      "epoch": 0.06481012658227848,
      "grad_norm": 0.6183629631996155,
      "learning_rate": 9.977660254210623e-06,
      "loss": 0.0279,
      "step": 200
    },
    {
      "epoch": 0.06481012658227848,
      "eval_accuracy": 0.9843501386117544,
      "eval_loss": 0.03353777900338173,
      "eval_runtime": 99.3735,
      "eval_samples_per_second": 50.315,
      "eval_steps_per_second": 12.579,
      "step": 200
    },
    {
      "epoch": 0.06545822784810126,
      "grad_norm": 0.5192128419876099,
      "learning_rate": 9.977056966640368e-06,
      "loss": 0.0317,
      "step": 202
    },
    {
      "epoch": 0.06610632911392406,
      "grad_norm": 0.5304135680198669,
      "learning_rate": 9.976445660039118e-06,
      "loss": 0.026,
      "step": 204
    },
    {
      "epoch": 0.06675443037974683,
      "grad_norm": 0.7126085162162781,
      "learning_rate": 9.975826335391808e-06,
      "loss": 0.028,
      "step": 206
    },
    {
      "epoch": 0.06740253164556963,
      "grad_norm": 0.4410451352596283,
      "learning_rate": 9.975198993696294e-06,
      "loss": 0.0287,
      "step": 208
    },
    {
      "epoch": 0.0680506329113924,
      "grad_norm": 0.4624532461166382,
      "learning_rate": 9.974563635963348e-06,
      "loss": 0.0273,
      "step": 210
    },
    {
      "epoch": 0.06869873417721518,
      "grad_norm": 0.5675721764564514,
      "learning_rate": 9.973920263216658e-06,
      "loss": 0.0257,
      "step": 212
    },
    {
      "epoch": 0.06934683544303798,
      "grad_norm": 0.5974188446998596,
      "learning_rate": 9.973268876492827e-06,
      "loss": 0.0301,
      "step": 214
    },
    {
      "epoch": 0.06999493670886076,
      "grad_norm": 0.55629962682724,
      "learning_rate": 9.972609476841368e-06,
      "loss": 0.0256,
      "step": 216
    },
    {
      "epoch": 0.07064303797468355,
      "grad_norm": 0.7521873116493225,
      "learning_rate": 9.971942065324704e-06,
      "loss": 0.0309,
      "step": 218
    },
    {
      "epoch": 0.07129113924050633,
      "grad_norm": 0.9954202175140381,
      "learning_rate": 9.971266643018171e-06,
      "loss": 0.0291,
      "step": 220
    },
    {
      "epoch": 0.07193924050632912,
      "grad_norm": 0.7211986780166626,
      "learning_rate": 9.970583211010008e-06,
      "loss": 0.0331,
      "step": 222
    },
    {
      "epoch": 0.0725873417721519,
      "grad_norm": 0.8910839557647705,
      "learning_rate": 9.969891770401358e-06,
      "loss": 0.0289,
      "step": 224
    },
    {
      "epoch": 0.07323544303797469,
      "grad_norm": 0.522318422794342,
      "learning_rate": 9.969192322306271e-06,
      "loss": 0.0278,
      "step": 226
    },
    {
      "epoch": 0.07388354430379747,
      "grad_norm": 0.6214476823806763,
      "learning_rate": 9.968484867851698e-06,
      "loss": 0.0248,
      "step": 228
    },
    {
      "epoch": 0.07453164556962025,
      "grad_norm": 0.5874454379081726,
      "learning_rate": 9.96776940817749e-06,
      "loss": 0.0245,
      "step": 230
    },
    {
      "epoch": 0.07517974683544304,
      "grad_norm": 1.1179145574569702,
      "learning_rate": 9.967045944436392e-06,
      "loss": 0.027,
      "step": 232
    },
    {
      "epoch": 0.07582784810126582,
      "grad_norm": 0.6947236657142639,
      "learning_rate": 9.966314477794052e-06,
      "loss": 0.0281,
      "step": 234
    },
    {
      "epoch": 0.07647594936708861,
      "grad_norm": 0.5388737320899963,
      "learning_rate": 9.965575009429006e-06,
      "loss": 0.0237,
      "step": 236
    },
    {
      "epoch": 0.07712405063291139,
      "grad_norm": 0.5712288022041321,
      "learning_rate": 9.964827540532685e-06,
      "loss": 0.0221,
      "step": 238
    },
    {
      "epoch": 0.07777215189873418,
      "grad_norm": 0.7290524244308472,
      "learning_rate": 9.964072072309412e-06,
      "loss": 0.024,
      "step": 240
    },
    {
      "epoch": 0.07777215189873418,
      "eval_accuracy": 0.9883965760621634,
      "eval_loss": 0.02650127001106739,
      "eval_runtime": 102.4631,
      "eval_samples_per_second": 48.798,
      "eval_steps_per_second": 12.2,
      "step": 240
    },
    {
      "epoch": 0.07842025316455696,
      "grad_norm": 0.6895317435264587,
      "learning_rate": 9.963308605976397e-06,
      "loss": 0.0249,
      "step": 242
    },
    {
      "epoch": 0.07906835443037975,
      "grad_norm": 0.5299035310745239,
      "learning_rate": 9.962537142763733e-06,
      "loss": 0.0257,
      "step": 244
    },
    {
      "epoch": 0.07971645569620253,
      "grad_norm": 0.41885891556739807,
      "learning_rate": 9.961757683914406e-06,
      "loss": 0.0235,
      "step": 246
    },
    {
      "epoch": 0.08036455696202531,
      "grad_norm": 0.7982097864151001,
      "learning_rate": 9.960970230684276e-06,
      "loss": 0.0293,
      "step": 248
    },
    {
      "epoch": 0.0810126582278481,
      "grad_norm": 0.7573944330215454,
      "learning_rate": 9.96017478434209e-06,
      "loss": 0.0255,
      "step": 250
    },
    {
      "epoch": 0.08166075949367088,
      "grad_norm": 0.575404703617096,
      "learning_rate": 9.959371346169466e-06,
      "loss": 0.023,
      "step": 252
    },
    {
      "epoch": 0.08230886075949367,
      "grad_norm": 0.4789620339870453,
      "learning_rate": 9.958559917460909e-06,
      "loss": 0.0238,
      "step": 254
    },
    {
      "epoch": 0.08295696202531645,
      "grad_norm": 0.6227384209632874,
      "learning_rate": 9.957740499523787e-06,
      "loss": 0.0194,
      "step": 256
    },
    {
      "epoch": 0.08360506329113924,
      "grad_norm": 0.6454212665557861,
      "learning_rate": 9.95691309367835e-06,
      "loss": 0.0188,
      "step": 258
    },
    {
      "epoch": 0.08425316455696202,
      "grad_norm": 1.03157377243042,
      "learning_rate": 9.95607770125771e-06,
      "loss": 0.0194,
      "step": 260
    },
    {
      "epoch": 0.08490126582278482,
      "grad_norm": 0.659908652305603,
      "learning_rate": 9.955234323607854e-06,
      "loss": 0.0182,
      "step": 262
    },
    {
      "epoch": 0.0855493670886076,
      "grad_norm": 0.5437667369842529,
      "learning_rate": 9.954382962087628e-06,
      "loss": 0.0181,
      "step": 264
    },
    {
      "epoch": 0.08619746835443037,
      "grad_norm": 0.6034504771232605,
      "learning_rate": 9.95352361806875e-06,
      "loss": 0.0179,
      "step": 266
    },
    {
      "epoch": 0.08684556962025317,
      "grad_norm": 0.6281765103340149,
      "learning_rate": 9.95265629293579e-06,
      "loss": 0.019,
      "step": 268
    },
    {
      "epoch": 0.08749367088607594,
      "grad_norm": 0.6361079216003418,
      "learning_rate": 9.951780988086183e-06,
      "loss": 0.0154,
      "step": 270
    },
    {
      "epoch": 0.08814177215189874,
      "grad_norm": 0.5678423643112183,
      "learning_rate": 9.950897704930223e-06,
      "loss": 0.0161,
      "step": 272
    },
    {
      "epoch": 0.08878987341772152,
      "grad_norm": 1.9274733066558838,
      "learning_rate": 9.95000644489105e-06,
      "loss": 0.0159,
      "step": 274
    },
    {
      "epoch": 0.08943797468354431,
      "grad_norm": 0.7999576926231384,
      "learning_rate": 9.949107209404664e-06,
      "loss": 0.0179,
      "step": 276
    },
    {
      "epoch": 0.09008607594936709,
      "grad_norm": 0.8246878981590271,
      "learning_rate": 9.948199999919914e-06,
      "loss": 0.0158,
      "step": 278
    },
    {
      "epoch": 0.09073417721518988,
      "grad_norm": 1.3101437091827393,
      "learning_rate": 9.947284817898493e-06,
      "loss": 0.0164,
      "step": 280
    },
    {
      "epoch": 0.09073417721518988,
      "eval_accuracy": 0.9941715658085519,
      "eval_loss": 0.014373338781297207,
      "eval_runtime": 102.9242,
      "eval_samples_per_second": 48.579,
      "eval_steps_per_second": 12.145,
      "step": 280
    },
    {
      "epoch": 0.09138227848101266,
      "grad_norm": 1.0926411151885986,
      "learning_rate": 9.946361664814942e-06,
      "loss": 0.0139,
      "step": 282
    },
    {
      "epoch": 0.09203037974683544,
      "grad_norm": 1.1184602975845337,
      "learning_rate": 9.945430542156647e-06,
      "loss": 0.0148,
      "step": 284
    },
    {
      "epoch": 0.09267848101265823,
      "grad_norm": 0.7032978534698486,
      "learning_rate": 9.944491451423829e-06,
      "loss": 0.0124,
      "step": 286
    },
    {
      "epoch": 0.09332658227848101,
      "grad_norm": 0.7714938521385193,
      "learning_rate": 9.943544394129552e-06,
      "loss": 0.0109,
      "step": 288
    },
    {
      "epoch": 0.0939746835443038,
      "grad_norm": 0.6936782598495483,
      "learning_rate": 9.942589371799715e-06,
      "loss": 0.009,
      "step": 290
    },
    {
      "epoch": 0.09462278481012658,
      "grad_norm": 0.743622899055481,
      "learning_rate": 9.941626385973047e-06,
      "loss": 0.0093,
      "step": 292
    },
    {
      "epoch": 0.09527088607594937,
      "grad_norm": 0.815427303314209,
      "learning_rate": 9.940655438201113e-06,
      "loss": 0.0089,
      "step": 294
    },
    {
      "epoch": 0.09591898734177215,
      "grad_norm": 1.016943335533142,
      "learning_rate": 9.9396765300483e-06,
      "loss": 0.0105,
      "step": 296
    },
    {
      "epoch": 0.09656708860759494,
      "grad_norm": 0.7963289618492126,
      "learning_rate": 9.938689663091828e-06,
      "loss": 0.0082,
      "step": 298
    },
    {
      "epoch": 0.09721518987341772,
      "grad_norm": 0.80525141954422,
      "learning_rate": 9.937694838921734e-06,
      "loss": 0.0116,
      "step": 300
    },
    {
      "epoch": 0.09786329113924051,
      "grad_norm": 1.1018229722976685,
      "learning_rate": 9.93669205914088e-06,
      "loss": 0.0083,
      "step": 302
    },
    {
      "epoch": 0.09851139240506329,
      "grad_norm": 0.6011309623718262,
      "learning_rate": 9.93568132536494e-06,
      "loss": 0.0055,
      "step": 304
    },
    {
      "epoch": 0.09915949367088607,
      "grad_norm": 0.6187568306922913,
      "learning_rate": 9.934662639222412e-06,
      "loss": 0.0079,
      "step": 306
    },
    {
      "epoch": 0.09980759493670886,
      "grad_norm": 0.7109318971633911,
      "learning_rate": 9.9336360023546e-06,
      "loss": 0.0056,
      "step": 308
    },
    {
      "epoch": 0.10045569620253164,
      "grad_norm": 0.9586265087127686,
      "learning_rate": 9.932601416415622e-06,
      "loss": 0.0055,
      "step": 310
    },
    {
      "epoch": 0.10110379746835443,
      "grad_norm": 0.6494437456130981,
      "learning_rate": 9.931558883072403e-06,
      "loss": 0.0048,
      "step": 312
    },
    {
      "epoch": 0.10175189873417721,
      "grad_norm": 0.7673395872116089,
      "learning_rate": 9.930508404004668e-06,
      "loss": 0.0059,
      "step": 314
    },
    {
      "epoch": 0.1024,
      "grad_norm": 0.5997980237007141,
      "learning_rate": 9.929449980904952e-06,
      "loss": 0.0045,
      "step": 316
    },
    {
      "epoch": 0.10304810126582278,
      "grad_norm": 0.6281548142433167,
      "learning_rate": 9.928383615478586e-06,
      "loss": 0.0046,
      "step": 318
    },
    {
      "epoch": 0.10369620253164558,
      "grad_norm": 0.7596728801727295,
      "learning_rate": 9.927309309443696e-06,
      "loss": 0.0052,
      "step": 320
    },
    {
      "epoch": 0.10369620253164558,
      "eval_accuracy": 0.9985858696992592,
      "eval_loss": 0.005098371766507626,
      "eval_runtime": 102.4403,
      "eval_samples_per_second": 48.809,
      "eval_steps_per_second": 12.202,
      "step": 320
    },
    {
      "epoch": 0.10434430379746835,
      "grad_norm": 0.5863160490989685,
      "learning_rate": 9.9262270645312e-06,
      "loss": 0.0072,
      "step": 322
    },
    {
      "epoch": 0.10499240506329113,
      "grad_norm": 0.4260634183883667,
      "learning_rate": 9.925136882484816e-06,
      "loss": 0.0037,
      "step": 324
    },
    {
      "epoch": 0.10564050632911393,
      "grad_norm": 0.6246248483657837,
      "learning_rate": 9.924038765061042e-06,
      "loss": 0.0048,
      "step": 326
    },
    {
      "epoch": 0.1062886075949367,
      "grad_norm": 0.4939432144165039,
      "learning_rate": 9.922932714029163e-06,
      "loss": 0.0037,
      "step": 328
    },
    {
      "epoch": 0.1069367088607595,
      "grad_norm": 1.4279497861862183,
      "learning_rate": 9.921818731171249e-06,
      "loss": 0.0032,
      "step": 330
    },
    {
      "epoch": 0.10758481012658228,
      "grad_norm": 0.4592280983924866,
      "learning_rate": 9.920696818282147e-06,
      "loss": 0.0013,
      "step": 332
    },
    {
      "epoch": 0.10823291139240507,
      "grad_norm": 0.6403955817222595,
      "learning_rate": 9.919566977169486e-06,
      "loss": 0.003,
      "step": 334
    },
    {
      "epoch": 0.10888101265822785,
      "grad_norm": 0.6681948304176331,
      "learning_rate": 9.918429209653662e-06,
      "loss": 0.0026,
      "step": 336
    },
    {
      "epoch": 0.10952911392405064,
      "grad_norm": 0.40499842166900635,
      "learning_rate": 9.917283517567845e-06,
      "loss": 0.0079,
      "step": 338
    },
    {
      "epoch": 0.11017721518987342,
      "grad_norm": 0.4384974539279938,
      "learning_rate": 9.916129902757977e-06,
      "loss": 0.0022,
      "step": 340
    },
    {
      "epoch": 0.1108253164556962,
      "grad_norm": 0.917611837387085,
      "learning_rate": 9.914968367082756e-06,
      "loss": 0.0019,
      "step": 342
    },
    {
      "epoch": 0.11147341772151899,
      "grad_norm": 0.5483088493347168,
      "learning_rate": 9.913798912413653e-06,
      "loss": 0.0041,
      "step": 344
    },
    {
      "epoch": 0.11212151898734177,
      "grad_norm": 0.6087149381637573,
      "learning_rate": 9.912621540634889e-06,
      "loss": 0.003,
      "step": 346
    },
    {
      "epoch": 0.11276962025316456,
      "grad_norm": 0.9792701005935669,
      "learning_rate": 9.911436253643445e-06,
      "loss": 0.0051,
      "step": 348
    },
    {
      "epoch": 0.11341772151898734,
      "grad_norm": 0.46137937903404236,
      "learning_rate": 9.910243053349055e-06,
      "loss": 0.0058,
      "step": 350
    },
    {
      "epoch": 0.11406582278481013,
      "grad_norm": 0.322523832321167,
      "learning_rate": 9.909041941674205e-06,
      "loss": 0.0036,
      "step": 352
    },
    {
      "epoch": 0.11471392405063291,
      "grad_norm": 0.6017354726791382,
      "learning_rate": 9.90783292055412e-06,
      "loss": 0.0024,
      "step": 354
    },
    {
      "epoch": 0.1153620253164557,
      "grad_norm": 0.3198406994342804,
      "learning_rate": 9.906615991936781e-06,
      "loss": 0.0022,
      "step": 356
    },
    {
      "epoch": 0.11601012658227848,
      "grad_norm": 1.0170084238052368,
      "learning_rate": 9.905391157782897e-06,
      "loss": 0.0021,
      "step": 358
    },
    {
      "epoch": 0.11665822784810126,
      "grad_norm": 0.4881608188152313,
      "learning_rate": 9.904158420065923e-06,
      "loss": 0.0014,
      "step": 360
    },
    {
      "epoch": 0.11665822784810126,
      "eval_accuracy": 0.9992789522224573,
      "eval_loss": 0.002443755976855755,
      "eval_runtime": 102.565,
      "eval_samples_per_second": 48.75,
      "eval_steps_per_second": 12.187,
      "step": 360
    },
    {
      "epoch": 0.11730632911392405,
      "grad_norm": 0.4268784523010254,
      "learning_rate": 9.902917780772043e-06,
      "loss": 0.0016,
      "step": 362
    },
    {
      "epoch": 0.11795443037974683,
      "grad_norm": 0.3397619128227234,
      "learning_rate": 9.901669241900178e-06,
      "loss": 0.0027,
      "step": 364
    },
    {
      "epoch": 0.11860253164556962,
      "grad_norm": 0.36573904752731323,
      "learning_rate": 9.900412805461968e-06,
      "loss": 0.0016,
      "step": 366
    },
    {
      "epoch": 0.1192506329113924,
      "grad_norm": 0.3953874111175537,
      "learning_rate": 9.899148473481786e-06,
      "loss": 0.0055,
      "step": 368
    },
    {
      "epoch": 0.1198987341772152,
      "grad_norm": 0.7349480390548706,
      "learning_rate": 9.89787624799672e-06,
      "loss": 0.0021,
      "step": 370
    },
    {
      "epoch": 0.12054683544303797,
      "grad_norm": 0.24667468667030334,
      "learning_rate": 9.896596131056583e-06,
      "loss": 0.0008,
      "step": 372
    },
    {
      "epoch": 0.12119493670886076,
      "grad_norm": 0.18324322998523712,
      "learning_rate": 9.895308124723897e-06,
      "loss": 0.001,
      "step": 374
    },
    {
      "epoch": 0.12184303797468354,
      "grad_norm": 0.3875872492790222,
      "learning_rate": 9.894012231073895e-06,
      "loss": 0.0023,
      "step": 376
    },
    {
      "epoch": 0.12249113924050634,
      "grad_norm": 0.5702228546142578,
      "learning_rate": 9.892708452194522e-06,
      "loss": 0.0012,
      "step": 378
    },
    {
      "epoch": 0.12313924050632911,
      "grad_norm": 0.4540491998195648,
      "learning_rate": 9.891396790186424e-06,
      "loss": 0.0013,
      "step": 380
    },
    {
      "epoch": 0.1237873417721519,
      "grad_norm": 0.17990654706954956,
      "learning_rate": 9.890077247162951e-06,
      "loss": 0.0034,
      "step": 382
    },
    {
      "epoch": 0.12443544303797469,
      "grad_norm": 0.1970151662826538,
      "learning_rate": 9.888749825250151e-06,
      "loss": 0.0008,
      "step": 384
    },
    {
      "epoch": 0.12508354430379748,
      "grad_norm": 0.37782326340675354,
      "learning_rate": 9.887414526586764e-06,
      "loss": 0.0013,
      "step": 386
    },
    {
      "epoch": 0.12573164556962024,
      "grad_norm": 0.7092784643173218,
      "learning_rate": 9.886071353324223e-06,
      "loss": 0.0027,
      "step": 388
    },
    {
      "epoch": 0.12637974683544304,
      "grad_norm": 0.907179594039917,
      "learning_rate": 9.884720307626647e-06,
      "loss": 0.0027,
      "step": 390
    },
    {
      "epoch": 0.12702784810126583,
      "grad_norm": 0.22963130474090576,
      "learning_rate": 9.883361391670841e-06,
      "loss": 0.0011,
      "step": 392
    },
    {
      "epoch": 0.12767594936708862,
      "grad_norm": 0.592887282371521,
      "learning_rate": 9.881994607646288e-06,
      "loss": 0.0041,
      "step": 394
    },
    {
      "epoch": 0.12832405063291139,
      "grad_norm": 0.639837920665741,
      "learning_rate": 9.880619957755151e-06,
      "loss": 0.0012,
      "step": 396
    },
    {
      "epoch": 0.12897215189873418,
      "grad_norm": 0.5888617038726807,
      "learning_rate": 9.879237444212265e-06,
      "loss": 0.0024,
      "step": 398
    },
    {
      "epoch": 0.12962025316455697,
      "grad_norm": 0.6880702972412109,
      "learning_rate": 9.877847069245134e-06,
      "loss": 0.0025,
      "step": 400
    },
    {
      "epoch": 0.12962025316455697,
      "eval_accuracy": 0.999652682586963,
      "eval_loss": 0.001261554891243577,
      "eval_runtime": 101.9742,
      "eval_samples_per_second": 49.032,
      "eval_steps_per_second": 12.258,
      "step": 400
    },
    {
      "epoch": 0.13026835443037973,
      "grad_norm": 0.5125930905342102,
      "learning_rate": 9.87644883509393e-06,
      "loss": 0.0008,
      "step": 402
    },
    {
      "epoch": 0.13091645569620253,
      "grad_norm": 0.6037792563438416,
      "learning_rate": 9.875042744011487e-06,
      "loss": 0.0012,
      "step": 404
    },
    {
      "epoch": 0.13156455696202532,
      "grad_norm": 0.14983819425106049,
      "learning_rate": 9.873628798263297e-06,
      "loss": 0.0013,
      "step": 406
    },
    {
      "epoch": 0.1322126582278481,
      "grad_norm": 0.14885587990283966,
      "learning_rate": 9.87220700012751e-06,
      "loss": 0.0003,
      "step": 408
    },
    {
      "epoch": 0.13286075949367088,
      "grad_norm": 0.28928548097610474,
      "learning_rate": 9.870777351894926e-06,
      "loss": 0.0006,
      "step": 410
    },
    {
      "epoch": 0.13350886075949367,
      "grad_norm": 0.23244060575962067,
      "learning_rate": 9.869339855868992e-06,
      "loss": 0.0019,
      "step": 412
    },
    {
      "epoch": 0.13415696202531646,
      "grad_norm": 0.4798615574836731,
      "learning_rate": 9.867894514365802e-06,
      "loss": 0.0029,
      "step": 414
    },
    {
      "epoch": 0.13480506329113925,
      "grad_norm": 0.2007167637348175,
      "learning_rate": 9.86644132971409e-06,
      "loss": 0.0003,
      "step": 416
    },
    {
      "epoch": 0.13545316455696202,
      "grad_norm": 0.42807483673095703,
      "learning_rate": 9.864980304255222e-06,
      "loss": 0.0009,
      "step": 418
    },
    {
      "epoch": 0.1361012658227848,
      "grad_norm": 0.09191492944955826,
      "learning_rate": 9.863511440343206e-06,
      "loss": 0.0003,
      "step": 420
    },
    {
      "epoch": 0.1367493670886076,
      "grad_norm": 0.4787420928478241,
      "learning_rate": 9.862034740344673e-06,
      "loss": 0.0007,
      "step": 422
    },
    {
      "epoch": 0.13739746835443037,
      "grad_norm": 0.34788092970848083,
      "learning_rate": 9.860550206638881e-06,
      "loss": 0.0006,
      "step": 424
    },
    {
      "epoch": 0.13804556962025316,
      "grad_norm": 0.2141241580247879,
      "learning_rate": 9.859057841617709e-06,
      "loss": 0.0008,
      "step": 426
    },
    {
      "epoch": 0.13869367088607595,
      "grad_norm": 0.18016201257705688,
      "learning_rate": 9.857557647685657e-06,
      "loss": 0.0003,
      "step": 428
    },
    {
      "epoch": 0.13934177215189875,
      "grad_norm": 0.13558854162693024,
      "learning_rate": 9.856049627259833e-06,
      "loss": 0.0022,
      "step": 430
    },
    {
      "epoch": 0.1399898734177215,
      "grad_norm": 0.17645730078220367,
      "learning_rate": 9.85453378276996e-06,
      "loss": 0.0005,
      "step": 432
    },
    {
      "epoch": 0.1406379746835443,
      "grad_norm": 0.3373895585536957,
      "learning_rate": 9.853010116658368e-06,
      "loss": 0.002,
      "step": 434
    },
    {
      "epoch": 0.1412860759493671,
      "grad_norm": 0.8115471005439758,
      "learning_rate": 9.851478631379982e-06,
      "loss": 0.0025,
      "step": 436
    },
    {
      "epoch": 0.14193417721518986,
      "grad_norm": 0.5957593321800232,
      "learning_rate": 9.849939329402337e-06,
      "loss": 0.0009,
      "step": 438
    },
    {
      "epoch": 0.14258227848101265,
      "grad_norm": 0.23780769109725952,
      "learning_rate": 9.848392213205549e-06,
      "loss": 0.0002,
      "step": 440
    },
    {
      "epoch": 0.14258227848101265,
      "eval_accuracy": 0.999539490974426,
      "eval_loss": 0.0016721197171136737,
      "eval_runtime": 103.3024,
      "eval_samples_per_second": 48.402,
      "eval_steps_per_second": 12.1,
      "step": 440
    },
    {
      "epoch": 0.14323037974683545,
      "grad_norm": 0.8508957624435425,
      "learning_rate": 9.846837285282331e-06,
      "loss": 0.0026,
      "step": 442
    },
    {
      "epoch": 0.14387848101265824,
      "grad_norm": 0.272156298160553,
      "learning_rate": 9.845274548137986e-06,
      "loss": 0.0008,
      "step": 444
    },
    {
      "epoch": 0.144526582278481,
      "grad_norm": 0.3533732295036316,
      "learning_rate": 9.843704004290393e-06,
      "loss": 0.0006,
      "step": 446
    },
    {
      "epoch": 0.1451746835443038,
      "grad_norm": 0.4029495418071747,
      "learning_rate": 9.842125656270011e-06,
      "loss": 0.001,
      "step": 448
    },
    {
      "epoch": 0.1458227848101266,
      "grad_norm": 0.8233634233474731,
      "learning_rate": 9.840539506619874e-06,
      "loss": 0.0008,
      "step": 450
    },
    {
      "epoch": 0.14647088607594938,
      "grad_norm": 0.5356901288032532,
      "learning_rate": 9.838945557895586e-06,
      "loss": 0.0007,
      "step": 452
    },
    {
      "epoch": 0.14711898734177214,
      "grad_norm": 0.2129882127046585,
      "learning_rate": 9.837343812665311e-06,
      "loss": 0.0021,
      "step": 454
    },
    {
      "epoch": 0.14776708860759494,
      "grad_norm": 0.5427262783050537,
      "learning_rate": 9.835734273509787e-06,
      "loss": 0.0011,
      "step": 456
    },
    {
      "epoch": 0.14841518987341773,
      "grad_norm": 0.19061443209648132,
      "learning_rate": 9.834116943022299e-06,
      "loss": 0.0006,
      "step": 458
    },
    {
      "epoch": 0.1490632911392405,
      "grad_norm": 0.1520262062549591,
      "learning_rate": 9.832491823808688e-06,
      "loss": 0.0004,
      "step": 460
    },
    {
      "epoch": 0.1497113924050633,
      "grad_norm": 0.35970884561538696,
      "learning_rate": 9.830858918487347e-06,
      "loss": 0.0043,
      "step": 462
    },
    {
      "epoch": 0.15035949367088608,
      "grad_norm": 0.3962363004684448,
      "learning_rate": 9.829218229689211e-06,
      "loss": 0.0013,
      "step": 464
    },
    {
      "epoch": 0.15100759493670887,
      "grad_norm": 0.37777262926101685,
      "learning_rate": 9.827569760057755e-06,
      "loss": 0.0031,
      "step": 466
    },
    {
      "epoch": 0.15165569620253164,
      "grad_norm": 0.445781946182251,
      "learning_rate": 9.825913512248996e-06,
      "loss": 0.0014,
      "step": 468
    },
    {
      "epoch": 0.15230379746835443,
      "grad_norm": 0.4247877895832062,
      "learning_rate": 9.824249488931477e-06,
      "loss": 0.0011,
      "step": 470
    },
    {
      "epoch": 0.15295189873417722,
      "grad_norm": 0.20235390961170197,
      "learning_rate": 9.822577692786272e-06,
      "loss": 0.0005,
      "step": 472
    },
    {
      "epoch": 0.1536,
      "grad_norm": 0.5411356091499329,
      "learning_rate": 9.820898126506978e-06,
      "loss": 0.0026,
      "step": 474
    },
    {
      "epoch": 0.15424810126582278,
      "grad_norm": 0.07654337584972382,
      "learning_rate": 9.819210792799711e-06,
      "loss": 0.0003,
      "step": 476
    },
    {
      "epoch": 0.15489620253164557,
      "grad_norm": 0.08077862858772278,
      "learning_rate": 9.817515694383102e-06,
      "loss": 0.0003,
      "step": 478
    },
    {
      "epoch": 0.15554430379746836,
      "grad_norm": 0.17977231740951538,
      "learning_rate": 9.815812833988292e-06,
      "loss": 0.0038,
      "step": 480
    },
    {
      "epoch": 0.15554430379746836,
      "eval_accuracy": 0.9997459873278298,
      "eval_loss": 0.0012969060335308313,
      "eval_runtime": 99.7243,
      "eval_samples_per_second": 50.138,
      "eval_steps_per_second": 12.535,
      "step": 480
    },
    {
      "epoch": 0.15619240506329113,
      "grad_norm": 0.22738438844680786,
      "learning_rate": 9.814102214358928e-06,
      "loss": 0.0002,
      "step": 482
    },
    {
      "epoch": 0.15684050632911392,
      "grad_norm": 0.20708250999450684,
      "learning_rate": 9.81238383825116e-06,
      "loss": 0.0043,
      "step": 484
    },
    {
      "epoch": 0.1574886075949367,
      "grad_norm": 0.29505881667137146,
      "learning_rate": 9.810657708433637e-06,
      "loss": 0.0004,
      "step": 486
    },
    {
      "epoch": 0.1581367088607595,
      "grad_norm": 0.7022032737731934,
      "learning_rate": 9.808923827687494e-06,
      "loss": 0.0017,
      "step": 488
    },
    {
      "epoch": 0.15878481012658227,
      "grad_norm": 0.34339165687561035,
      "learning_rate": 9.807182198806362e-06,
      "loss": 0.0034,
      "step": 490
    },
    {
      "epoch": 0.15943291139240506,
      "grad_norm": 0.19041544198989868,
      "learning_rate": 9.805432824596347e-06,
      "loss": 0.0014,
      "step": 492
    },
    {
      "epoch": 0.16008101265822786,
      "grad_norm": 0.24253234267234802,
      "learning_rate": 9.803675707876048e-06,
      "loss": 0.0038,
      "step": 494
    },
    {
      "epoch": 0.16072911392405062,
      "grad_norm": 0.41031545400619507,
      "learning_rate": 9.801910851476524e-06,
      "loss": 0.0006,
      "step": 496
    },
    {
      "epoch": 0.1613772151898734,
      "grad_norm": 0.0826662927865982,
      "learning_rate": 9.800138258241311e-06,
      "loss": 0.002,
      "step": 498
    },
    {
      "epoch": 0.1620253164556962,
      "grad_norm": 0.16029556095600128,
      "learning_rate": 9.798357931026411e-06,
      "loss": 0.0016,
      "step": 500
    },
    {
      "epoch": 0.162673417721519,
      "grad_norm": 0.40231359004974365,
      "learning_rate": 9.796569872700287e-06,
      "loss": 0.0012,
      "step": 502
    },
    {
      "epoch": 0.16332151898734176,
      "grad_norm": 0.13478736579418182,
      "learning_rate": 9.79477408614386e-06,
      "loss": 0.0005,
      "step": 504
    },
    {
      "epoch": 0.16396962025316456,
      "grad_norm": 0.39503753185272217,
      "learning_rate": 9.792970574250493e-06,
      "loss": 0.0035,
      "step": 506
    },
    {
      "epoch": 0.16461772151898735,
      "grad_norm": 0.7558596730232239,
      "learning_rate": 9.791159339926009e-06,
      "loss": 0.0064,
      "step": 508
    },
    {
      "epoch": 0.16526582278481014,
      "grad_norm": 0.6682188510894775,
      "learning_rate": 9.789340386088663e-06,
      "loss": 0.0035,
      "step": 510
    },
    {
      "epoch": 0.1659139240506329,
      "grad_norm": 0.4081304669380188,
      "learning_rate": 9.787513715669158e-06,
      "loss": 0.0015,
      "step": 512
    },
    {
      "epoch": 0.1665620253164557,
      "grad_norm": 0.22257903218269348,
      "learning_rate": 9.78567933161062e-06,
      "loss": 0.0008,
      "step": 514
    },
    {
      "epoch": 0.1672101265822785,
      "grad_norm": 0.201273575425148,
      "learning_rate": 9.78383723686861e-06,
      "loss": 0.0007,
      "step": 516
    },
    {
      "epoch": 0.16785822784810125,
      "grad_norm": 0.08263271301984787,
      "learning_rate": 9.781987434411106e-06,
      "loss": 0.0005,
      "step": 518
    },
    {
      "epoch": 0.16850632911392405,
      "grad_norm": 0.052383746951818466,
      "learning_rate": 9.780129927218513e-06,
      "loss": 0.0017,
      "step": 520
    },
    {
      "epoch": 0.16850632911392405,
      "eval_accuracy": 0.9997890522710435,
      "eval_loss": 0.0008646832429803908,
      "eval_runtime": 102.7258,
      "eval_samples_per_second": 48.673,
      "eval_steps_per_second": 12.168,
      "step": 520
    },
    {
      "epoch": 0.16915443037974684,
      "grad_norm": 0.16042329370975494,
      "learning_rate": 9.778264718283644e-06,
      "loss": 0.0031,
      "step": 522
    },
    {
      "epoch": 0.16980253164556963,
      "grad_norm": 0.13061323761940002,
      "learning_rate": 9.776391810611719e-06,
      "loss": 0.0002,
      "step": 524
    },
    {
      "epoch": 0.1704506329113924,
      "grad_norm": 0.15899308025836945,
      "learning_rate": 9.774511207220369e-06,
      "loss": 0.0008,
      "step": 526
    },
    {
      "epoch": 0.1710987341772152,
      "grad_norm": 0.12816283106803894,
      "learning_rate": 9.772622911139622e-06,
      "loss": 0.0005,
      "step": 528
    },
    {
      "epoch": 0.17174683544303798,
      "grad_norm": 0.16552942991256714,
      "learning_rate": 9.770726925411898e-06,
      "loss": 0.0002,
      "step": 530
    },
    {
      "epoch": 0.17239493670886075,
      "grad_norm": 0.07594162225723267,
      "learning_rate": 9.768823253092008e-06,
      "loss": 0.0003,
      "step": 532
    },
    {
      "epoch": 0.17304303797468354,
      "grad_norm": 0.03090669773519039,
      "learning_rate": 9.766911897247147e-06,
      "loss": 0.0002,
      "step": 534
    },
    {
      "epoch": 0.17369113924050633,
      "grad_norm": 0.19276919960975647,
      "learning_rate": 9.76499286095689e-06,
      "loss": 0.0005,
      "step": 536
    },
    {
      "epoch": 0.17433924050632912,
      "grad_norm": 0.17801803350448608,
      "learning_rate": 9.763066147313189e-06,
      "loss": 0.0036,
      "step": 538
    },
    {
      "epoch": 0.1749873417721519,
      "grad_norm": 0.0805857926607132,
      "learning_rate": 9.76113175942036e-06,
      "loss": 0.002,
      "step": 540
    },
    {
      "epoch": 0.17563544303797468,
      "grad_norm": 0.8715735077857971,
      "learning_rate": 9.759189700395096e-06,
      "loss": 0.0013,
      "step": 542
    },
    {
      "epoch": 0.17628354430379747,
      "grad_norm": 0.1977226585149765,
      "learning_rate": 9.75723997336643e-06,
      "loss": 0.0003,
      "step": 544
    },
    {
      "epoch": 0.17693164556962027,
      "grad_norm": 0.5493549704551697,
      "learning_rate": 9.755282581475769e-06,
      "loss": 0.0003,
      "step": 546
    },
    {
      "epoch": 0.17757974683544303,
      "grad_norm": 1.054047703742981,
      "learning_rate": 9.753317527876857e-06,
      "loss": 0.0009,
      "step": 548
    },
    {
      "epoch": 0.17822784810126582,
      "grad_norm": 0.6819576621055603,
      "learning_rate": 9.751344815735791e-06,
      "loss": 0.0013,
      "step": 550
    },
    {
      "epoch": 0.17887594936708862,
      "grad_norm": 0.12686581909656525,
      "learning_rate": 9.749364448231001e-06,
      "loss": 0.0005,
      "step": 552
    },
    {
      "epoch": 0.17952405063291138,
      "grad_norm": 1.1412973403930664,
      "learning_rate": 9.747376428553255e-06,
      "loss": 0.0025,
      "step": 554
    },
    {
      "epoch": 0.18017215189873417,
      "grad_norm": 0.3469236493110657,
      "learning_rate": 9.745380759905648e-06,
      "loss": 0.0021,
      "step": 556
    },
    {
      "epoch": 0.18082025316455697,
      "grad_norm": 0.31161972880363464,
      "learning_rate": 9.743377445503598e-06,
      "loss": 0.0019,
      "step": 558
    },
    {
      "epoch": 0.18146835443037976,
      "grad_norm": 0.11834688484668732,
      "learning_rate": 9.74136648857485e-06,
      "loss": 0.0018,
      "step": 560
    },
    {
      "epoch": 0.18146835443037976,
      "eval_accuracy": 0.9997172934457434,
      "eval_loss": 0.0010605291463434696,
      "eval_runtime": 102.2732,
      "eval_samples_per_second": 48.889,
      "eval_steps_per_second": 12.222,
      "step": 560
    },
    {
      "epoch": 0.18211645569620252,
      "grad_norm": 0.30010661482810974,
      "learning_rate": 9.739347892359453e-06,
      "loss": 0.0009,
      "step": 562
    },
    {
      "epoch": 0.18276455696202532,
      "grad_norm": 0.10496673732995987,
      "learning_rate": 9.737321660109767e-06,
      "loss": 0.0006,
      "step": 564
    },
    {
      "epoch": 0.1834126582278481,
      "grad_norm": 0.36886048316955566,
      "learning_rate": 9.735287795090455e-06,
      "loss": 0.0015,
      "step": 566
    },
    {
      "epoch": 0.18406075949367087,
      "grad_norm": 0.4008212387561798,
      "learning_rate": 9.733246300578482e-06,
      "loss": 0.0008,
      "step": 568
    },
    {
      "epoch": 0.18470886075949366,
      "grad_norm": 0.6184799671173096,
      "learning_rate": 9.731197179863104e-06,
      "loss": 0.0009,
      "step": 570
    },
    {
      "epoch": 0.18535696202531646,
      "grad_norm": 0.12914961576461792,
      "learning_rate": 9.729140436245857e-06,
      "loss": 0.0003,
      "step": 572
    },
    {
      "epoch": 0.18600506329113925,
      "grad_norm": 0.2764883041381836,
      "learning_rate": 9.72707607304057e-06,
      "loss": 0.0026,
      "step": 574
    },
    {
      "epoch": 0.18665316455696201,
      "grad_norm": 0.3653884530067444,
      "learning_rate": 9.725004093573343e-06,
      "loss": 0.0003,
      "step": 576
    },
    {
      "epoch": 0.1873012658227848,
      "grad_norm": 1.2126648426055908,
      "learning_rate": 9.722924501182546e-06,
      "loss": 0.0005,
      "step": 578
    },
    {
      "epoch": 0.1879493670886076,
      "grad_norm": 0.2613915205001831,
      "learning_rate": 9.72083729921882e-06,
      "loss": 0.002,
      "step": 580
    },
    {
      "epoch": 0.1885974683544304,
      "grad_norm": 0.0996515303850174,
      "learning_rate": 9.718742491045061e-06,
      "loss": 0.0006,
      "step": 582
    },
    {
      "epoch": 0.18924556962025316,
      "grad_norm": 0.5431665778160095,
      "learning_rate": 9.716640080036423e-06,
      "loss": 0.0023,
      "step": 584
    },
    {
      "epoch": 0.18989367088607595,
      "grad_norm": 0.7025941014289856,
      "learning_rate": 9.71453006958031e-06,
      "loss": 0.0022,
      "step": 586
    },
    {
      "epoch": 0.19054177215189874,
      "grad_norm": 0.8454611897468567,
      "learning_rate": 9.712412463076368e-06,
      "loss": 0.0035,
      "step": 588
    },
    {
      "epoch": 0.1911898734177215,
      "grad_norm": 0.19344277679920197,
      "learning_rate": 9.710287263936485e-06,
      "loss": 0.0019,
      "step": 590
    },
    {
      "epoch": 0.1918379746835443,
      "grad_norm": 0.2613272964954376,
      "learning_rate": 9.708154475584779e-06,
      "loss": 0.0009,
      "step": 592
    },
    {
      "epoch": 0.1924860759493671,
      "grad_norm": 0.6521615982055664,
      "learning_rate": 9.7060141014576e-06,
      "loss": 0.0028,
      "step": 594
    },
    {
      "epoch": 0.19313417721518988,
      "grad_norm": 0.20852723717689514,
      "learning_rate": 9.703866145003512e-06,
      "loss": 0.002,
      "step": 596
    },
    {
      "epoch": 0.19378227848101265,
      "grad_norm": 0.28225457668304443,
      "learning_rate": 9.701710609683305e-06,
      "loss": 0.0008,
      "step": 598
    },
    {
      "epoch": 0.19443037974683544,
      "grad_norm": 0.3033112585544586,
      "learning_rate": 9.699547498969978e-06,
      "loss": 0.0009,
      "step": 600
    },
    {
      "epoch": 0.19443037974683544,
      "eval_accuracy": 0.9997885038585653,
      "eval_loss": 0.0007942758384160697,
      "eval_runtime": 105.5854,
      "eval_samples_per_second": 47.355,
      "eval_steps_per_second": 11.839,
      "step": 600
    },
    {
      "epoch": 0.19507848101265823,
      "grad_norm": 0.13795693218708038,
      "learning_rate": 9.697376816348732e-06,
      "loss": 0.0002,
      "step": 602
    },
    {
      "epoch": 0.19572658227848103,
      "grad_norm": 0.07170484960079193,
      "learning_rate": 9.695198565316966e-06,
      "loss": 0.0002,
      "step": 604
    },
    {
      "epoch": 0.1963746835443038,
      "grad_norm": 0.24718345701694489,
      "learning_rate": 9.69301274938428e-06,
      "loss": 0.002,
      "step": 606
    },
    {
      "epoch": 0.19702278481012658,
      "grad_norm": 0.33944374322891235,
      "learning_rate": 9.690819372072457e-06,
      "loss": 0.0016,
      "step": 608
    },
    {
      "epoch": 0.19767088607594938,
      "grad_norm": 0.11184486001729965,
      "learning_rate": 9.68861843691547e-06,
      "loss": 0.0002,
      "step": 610
    },
    {
      "epoch": 0.19831898734177214,
      "grad_norm": 0.0845549926161766,
      "learning_rate": 9.68640994745946e-06,
      "loss": 0.0003,
      "step": 612
    },
    {
      "epoch": 0.19896708860759493,
      "grad_norm": 0.28036901354789734,
      "learning_rate": 9.684193907262742e-06,
      "loss": 0.0019,
      "step": 614
    },
    {
      "epoch": 0.19961518987341773,
      "grad_norm": 0.019151655957102776,
      "learning_rate": 9.681970319895804e-06,
      "loss": 0.0004,
      "step": 616
    },
    {
      "epoch": 0.20026329113924052,
      "grad_norm": 0.0641448125243187,
      "learning_rate": 9.679739188941283e-06,
      "loss": 0.0001,
      "step": 618
    },
    {
      "epoch": 0.20091139240506328,
      "grad_norm": 0.012318900786340237,
      "learning_rate": 9.677500517993983e-06,
      "loss": 0.0018,
      "step": 620
    },
    {
      "epoch": 0.20155949367088608,
      "grad_norm": 0.012480619363486767,
      "learning_rate": 9.675254310660842e-06,
      "loss": 0.0,
      "step": 622
    },
    {
      "epoch": 0.20220759493670887,
      "grad_norm": 0.12060311436653137,
      "learning_rate": 9.673000570560952e-06,
      "loss": 0.0002,
      "step": 624
    },
    {
      "epoch": 0.20285569620253163,
      "grad_norm": 0.13326050341129303,
      "learning_rate": 9.670739301325534e-06,
      "loss": 0.0014,
      "step": 626
    },
    {
      "epoch": 0.20350379746835442,
      "grad_norm": 0.10389236360788345,
      "learning_rate": 9.668470506597946e-06,
      "loss": 0.0013,
      "step": 628
    },
    {
      "epoch": 0.20415189873417722,
      "grad_norm": 0.06838095188140869,
      "learning_rate": 9.66619419003367e-06,
      "loss": 0.0012,
      "step": 630
    },
    {
      "epoch": 0.2048,
      "grad_norm": 0.022482849657535553,
      "learning_rate": 9.663910355300306e-06,
      "loss": 0.0001,
      "step": 632
    },
    {
      "epoch": 0.20544810126582277,
      "grad_norm": 0.17222675681114197,
      "learning_rate": 9.661619006077562e-06,
      "loss": 0.0001,
      "step": 634
    },
    {
      "epoch": 0.20609620253164557,
      "grad_norm": 0.04022524133324623,
      "learning_rate": 9.659320146057263e-06,
      "loss": 0.0001,
      "step": 636
    },
    {
      "epoch": 0.20674430379746836,
      "grad_norm": 0.0643358826637268,
      "learning_rate": 9.657013778943328e-06,
      "loss": 0.0,
      "step": 638
    },
    {
      "epoch": 0.20739240506329115,
      "grad_norm": 0.16372860968112946,
      "learning_rate": 9.654699908451777e-06,
      "loss": 0.0012,
      "step": 640
    },
    {
      "epoch": 0.20739240506329115,
      "eval_accuracy": 0.999866675350015,
      "eval_loss": 0.0004598509694915265,
      "eval_runtime": 105.4679,
      "eval_samples_per_second": 47.408,
      "eval_steps_per_second": 11.852,
      "step": 640
    },
    {
      "epoch": 0.20804050632911392,
      "grad_norm": 0.009580990299582481,
      "learning_rate": 9.652378538310715e-06,
      "loss": 0.0,
      "step": 642
    },
    {
      "epoch": 0.2086886075949367,
      "grad_norm": 0.014977443031966686,
      "learning_rate": 9.650049672260333e-06,
      "loss": 0.0004,
      "step": 644
    },
    {
      "epoch": 0.2093367088607595,
      "grad_norm": 0.045772042125463486,
      "learning_rate": 9.647713314052896e-06,
      "loss": 0.0013,
      "step": 646
    },
    {
      "epoch": 0.20998481012658227,
      "grad_norm": 0.009746350347995758,
      "learning_rate": 9.645369467452746e-06,
      "loss": 0.0,
      "step": 648
    },
    {
      "epoch": 0.21063291139240506,
      "grad_norm": 0.02030940167605877,
      "learning_rate": 9.643018136236286e-06,
      "loss": 0.0,
      "step": 650
    },
    {
      "epoch": 0.21128101265822785,
      "grad_norm": 0.25576162338256836,
      "learning_rate": 9.64065932419198e-06,
      "loss": 0.0004,
      "step": 652
    },
    {
      "epoch": 0.21192911392405064,
      "grad_norm": 0.06584621220827103,
      "learning_rate": 9.638293035120342e-06,
      "loss": 0.0001,
      "step": 654
    },
    {
      "epoch": 0.2125772151898734,
      "grad_norm": 0.10090325772762299,
      "learning_rate": 9.635919272833938e-06,
      "loss": 0.0004,
      "step": 656
    },
    {
      "epoch": 0.2132253164556962,
      "grad_norm": 0.2609342336654663,
      "learning_rate": 9.63353804115737e-06,
      "loss": 0.0017,
      "step": 658
    },
    {
      "epoch": 0.213873417721519,
      "grad_norm": 0.8355330228805542,
      "learning_rate": 9.63114934392728e-06,
      "loss": 0.0017,
      "step": 660
    },
    {
      "epoch": 0.21452151898734176,
      "grad_norm": 0.05891817435622215,
      "learning_rate": 9.628753184992334e-06,
      "loss": 0.0002,
      "step": 662
    },
    {
      "epoch": 0.21516962025316455,
      "grad_norm": 0.5131203532218933,
      "learning_rate": 9.62634956821322e-06,
      "loss": 0.002,
      "step": 664
    },
    {
      "epoch": 0.21581772151898734,
      "grad_norm": 0.17226867377758026,
      "learning_rate": 9.623938497462647e-06,
      "loss": 0.0045,
      "step": 666
    },
    {
      "epoch": 0.21646582278481014,
      "grad_norm": 0.007893595844507217,
      "learning_rate": 9.621519976625327e-06,
      "loss": 0.0,
      "step": 668
    },
    {
      "epoch": 0.2171139240506329,
      "grad_norm": 0.07217761129140854,
      "learning_rate": 9.619094009597982e-06,
      "loss": 0.0005,
      "step": 670
    },
    {
      "epoch": 0.2177620253164557,
      "grad_norm": 0.01869891956448555,
      "learning_rate": 9.616660600289329e-06,
      "loss": 0.0017,
      "step": 672
    },
    {
      "epoch": 0.21841012658227849,
      "grad_norm": 0.21836678683757782,
      "learning_rate": 9.614219752620074e-06,
      "loss": 0.0012,
      "step": 674
    },
    {
      "epoch": 0.21905822784810128,
      "grad_norm": 0.06981157511472702,
      "learning_rate": 9.611771470522908e-06,
      "loss": 0.0013,
      "step": 676
    },
    {
      "epoch": 0.21970632911392404,
      "grad_norm": 0.34331014752388,
      "learning_rate": 9.609315757942504e-06,
      "loss": 0.001,
      "step": 678
    },
    {
      "epoch": 0.22035443037974684,
      "grad_norm": 0.10346855223178864,
      "learning_rate": 9.606852618835503e-06,
      "loss": 0.0005,
      "step": 680
    },
    {
      "epoch": 0.22035443037974684,
      "eval_accuracy": 0.999794656342699,
      "eval_loss": 0.0007786342175677419,
      "eval_runtime": 104.3292,
      "eval_samples_per_second": 47.925,
      "eval_steps_per_second": 11.981,
      "step": 680
    },
    {
      "epoch": 0.22100253164556963,
      "grad_norm": 0.20089657604694366,
      "learning_rate": 9.604382057170514e-06,
      "loss": 0.0004,
      "step": 682
    },
    {
      "epoch": 0.2216506329113924,
      "grad_norm": 0.1667291522026062,
      "learning_rate": 9.601904076928103e-06,
      "loss": 0.0005,
      "step": 684
    },
    {
      "epoch": 0.22229873417721518,
      "grad_norm": 0.3179922103881836,
      "learning_rate": 9.599418682100793e-06,
      "loss": 0.0017,
      "step": 686
    },
    {
      "epoch": 0.22294683544303798,
      "grad_norm": 1.0040173530578613,
      "learning_rate": 9.596925876693047e-06,
      "loss": 0.0033,
      "step": 688
    },
    {
      "epoch": 0.22359493670886077,
      "grad_norm": 0.1465279906988144,
      "learning_rate": 9.594425664721275e-06,
      "loss": 0.0003,
      "step": 690
    },
    {
      "epoch": 0.22424303797468353,
      "grad_norm": 0.4364776015281677,
      "learning_rate": 9.591918050213814e-06,
      "loss": 0.002,
      "step": 692
    },
    {
      "epoch": 0.22489113924050633,
      "grad_norm": 0.06411771476268768,
      "learning_rate": 9.589403037210933e-06,
      "loss": 0.0007,
      "step": 694
    },
    {
      "epoch": 0.22553924050632912,
      "grad_norm": 0.33405566215515137,
      "learning_rate": 9.586880629764817e-06,
      "loss": 0.0017,
      "step": 696
    },
    {
      "epoch": 0.2261873417721519,
      "grad_norm": 0.14692090451717377,
      "learning_rate": 9.584350831939571e-06,
      "loss": 0.0016,
      "step": 698
    },
    {
      "epoch": 0.22683544303797468,
      "grad_norm": 0.21872135996818542,
      "learning_rate": 9.581813647811199e-06,
      "loss": 0.0006,
      "step": 700
    },
    {
      "epoch": 0.22748354430379747,
      "grad_norm": 0.3012380599975586,
      "learning_rate": 9.579269081467614e-06,
      "loss": 0.0004,
      "step": 702
    },
    {
      "epoch": 0.22813164556962026,
      "grad_norm": 0.22102029621601105,
      "learning_rate": 9.576717137008617e-06,
      "loss": 0.0003,
      "step": 704
    },
    {
      "epoch": 0.22877974683544303,
      "grad_norm": 0.05813106894493103,
      "learning_rate": 9.574157818545902e-06,
      "loss": 0.0005,
      "step": 706
    },
    {
      "epoch": 0.22942784810126582,
      "grad_norm": 0.0369562990963459,
      "learning_rate": 9.57159113020304e-06,
      "loss": 0.0011,
      "step": 708
    },
    {
      "epoch": 0.2300759493670886,
      "grad_norm": 0.21441668272018433,
      "learning_rate": 9.569017076115476e-06,
      "loss": 0.0006,
      "step": 710
    },
    {
      "epoch": 0.2307240506329114,
      "grad_norm": 0.5280594229698181,
      "learning_rate": 9.566435660430528e-06,
      "loss": 0.0013,
      "step": 712
    },
    {
      "epoch": 0.23137215189873417,
      "grad_norm": 0.3137824833393097,
      "learning_rate": 9.563846887307369e-06,
      "loss": 0.0014,
      "step": 714
    },
    {
      "epoch": 0.23202025316455696,
      "grad_norm": 0.11566320061683655,
      "learning_rate": 9.561250760917026e-06,
      "loss": 0.0002,
      "step": 716
    },
    {
      "epoch": 0.23266835443037975,
      "grad_norm": 0.07832954823970795,
      "learning_rate": 9.558647285442382e-06,
      "loss": 0.0003,
      "step": 718
    },
    {
      "epoch": 0.23331645569620252,
      "grad_norm": 0.15707625448703766,
      "learning_rate": 9.55603646507815e-06,
      "loss": 0.0003,
      "step": 720
    },
    {
      "epoch": 0.23331645569620252,
      "eval_accuracy": 0.9998935399990166,
      "eval_loss": 0.0003976961597800255,
      "eval_runtime": 104.7543,
      "eval_samples_per_second": 47.731,
      "eval_steps_per_second": 11.933,
      "step": 720
    },
    {
      "epoch": 0.2339645569620253,
      "grad_norm": 0.1931619644165039,
      "learning_rate": 9.553418304030886e-06,
      "loss": 0.0008,
      "step": 722
    },
    {
      "epoch": 0.2346126582278481,
      "grad_norm": 0.15913955867290497,
      "learning_rate": 9.550792806518967e-06,
      "loss": 0.0001,
      "step": 724
    },
    {
      "epoch": 0.2352607594936709,
      "grad_norm": 0.08975693583488464,
      "learning_rate": 9.548159976772593e-06,
      "loss": 0.0003,
      "step": 726
    },
    {
      "epoch": 0.23590886075949366,
      "grad_norm": 0.14222075045108795,
      "learning_rate": 9.545519819033777e-06,
      "loss": 0.0011,
      "step": 728
    },
    {
      "epoch": 0.23655696202531645,
      "grad_norm": 0.034183427691459656,
      "learning_rate": 9.542872337556341e-06,
      "loss": 0.0,
      "step": 730
    },
    {
      "epoch": 0.23720506329113925,
      "grad_norm": 0.006229052785784006,
      "learning_rate": 9.540217536605906e-06,
      "loss": 0.0009,
      "step": 732
    },
    {
      "epoch": 0.23785316455696204,
      "grad_norm": 0.08647812902927399,
      "learning_rate": 9.537555420459883e-06,
      "loss": 0.0,
      "step": 734
    },
    {
      "epoch": 0.2385012658227848,
      "grad_norm": 0.022472726181149483,
      "learning_rate": 9.534885993407474e-06,
      "loss": 0.0002,
      "step": 736
    },
    {
      "epoch": 0.2391493670886076,
      "grad_norm": 0.012393929995596409,
      "learning_rate": 9.532209259749658e-06,
      "loss": 0.0005,
      "step": 738
    },
    {
      "epoch": 0.2397974683544304,
      "grad_norm": 0.08754698187112808,
      "learning_rate": 9.529525223799185e-06,
      "loss": 0.0009,
      "step": 740
    },
    {
      "epoch": 0.24044556962025315,
      "grad_norm": 0.056353695690631866,
      "learning_rate": 9.526833889880573e-06,
      "loss": 0.0001,
      "step": 742
    },
    {
      "epoch": 0.24109367088607594,
      "grad_norm": 0.39609044790267944,
      "learning_rate": 9.524135262330098e-06,
      "loss": 0.0014,
      "step": 744
    },
    {
      "epoch": 0.24174177215189874,
      "grad_norm": 0.3467492163181305,
      "learning_rate": 9.521429345495787e-06,
      "loss": 0.0002,
      "step": 746
    },
    {
      "epoch": 0.24238987341772153,
      "grad_norm": 0.4665900766849518,
      "learning_rate": 9.51871614373741e-06,
      "loss": 0.0012,
      "step": 748
    },
    {
      "epoch": 0.2430379746835443,
      "grad_norm": 0.24675174057483673,
      "learning_rate": 9.515995661426478e-06,
      "loss": 0.0015,
      "step": 750
    },
    {
      "epoch": 0.2436860759493671,
      "grad_norm": 0.03283111751079559,
      "learning_rate": 9.513267902946228e-06,
      "loss": 0.0,
      "step": 752
    },
    {
      "epoch": 0.24433417721518988,
      "grad_norm": 0.10860046744346619,
      "learning_rate": 9.510532872691624e-06,
      "loss": 0.0014,
      "step": 754
    },
    {
      "epoch": 0.24498227848101267,
      "grad_norm": 0.1112566888332367,
      "learning_rate": 9.507790575069347e-06,
      "loss": 0.0001,
      "step": 756
    },
    {
      "epoch": 0.24563037974683544,
      "grad_norm": 0.11744935810565948,
      "learning_rate": 9.50504101449778e-06,
      "loss": 0.0012,
      "step": 758
    },
    {
      "epoch": 0.24627848101265823,
      "grad_norm": 0.13507533073425293,
      "learning_rate": 9.50228419540702e-06,
      "loss": 0.0009,
      "step": 760
    },
    {
      "epoch": 0.24627848101265823,
      "eval_accuracy": 0.9998955433397497,
      "eval_loss": 0.0003588534309528768,
      "eval_runtime": 105.2425,
      "eval_samples_per_second": 47.509,
      "eval_steps_per_second": 11.877,
      "step": 760
    },
    {
      "epoch": 0.24692658227848102,
      "grad_norm": 0.17681878805160522,
      "learning_rate": 9.499520122238846e-06,
      "loss": 0.0012,
      "step": 762
    },
    {
      "epoch": 0.2475746835443038,
      "grad_norm": 0.061774663627147675,
      "learning_rate": 9.496748799446733e-06,
      "loss": 0.0001,
      "step": 764
    },
    {
      "epoch": 0.24822278481012658,
      "grad_norm": 0.010383802466094494,
      "learning_rate": 9.493970231495836e-06,
      "loss": 0.0008,
      "step": 766
    },
    {
      "epoch": 0.24887088607594937,
      "grad_norm": 0.09358537942171097,
      "learning_rate": 9.49118442286298e-06,
      "loss": 0.0002,
      "step": 768
    },
    {
      "epoch": 0.24951898734177216,
      "grad_norm": 0.007097979541867971,
      "learning_rate": 9.488391378036662e-06,
      "loss": 0.0,
      "step": 770
    },
    {
      "epoch": 0.25016708860759496,
      "grad_norm": 0.24831758439540863,
      "learning_rate": 9.485591101517027e-06,
      "loss": 0.0019,
      "step": 772
    },
    {
      "epoch": 0.2508151898734177,
      "grad_norm": 0.043643102049827576,
      "learning_rate": 9.482783597815883e-06,
      "loss": 0.0008,
      "step": 774
    },
    {
      "epoch": 0.2514632911392405,
      "grad_norm": 0.05945923179388046,
      "learning_rate": 9.47996887145668e-06,
      "loss": 0.0009,
      "step": 776
    },
    {
      "epoch": 0.2521113924050633,
      "grad_norm": 0.011810753494501114,
      "learning_rate": 9.477146926974501e-06,
      "loss": 0.0001,
      "step": 778
    },
    {
      "epoch": 0.25275949367088607,
      "grad_norm": 0.0021227251272648573,
      "learning_rate": 9.47431776891606e-06,
      "loss": 0.0,
      "step": 780
    },
    {
      "epoch": 0.25340759493670884,
      "grad_norm": 0.002886489499360323,
      "learning_rate": 9.471481401839696e-06,
      "loss": 0.0001,
      "step": 782
    },
    {
      "epoch": 0.25405569620253166,
      "grad_norm": 0.1468556523323059,
      "learning_rate": 9.468637830315364e-06,
      "loss": 0.0006,
      "step": 784
    },
    {
      "epoch": 0.2547037974683544,
      "grad_norm": 0.015433026477694511,
      "learning_rate": 9.46578705892462e-06,
      "loss": 0.0,
      "step": 786
    },
    {
      "epoch": 0.25535189873417724,
      "grad_norm": 0.10068199783563614,
      "learning_rate": 9.46292909226063e-06,
      "loss": 0.0013,
      "step": 788
    },
    {
      "epoch": 0.256,
      "grad_norm": 0.12383728474378586,
      "learning_rate": 9.460063934928142e-06,
      "loss": 0.0006,
      "step": 790
    },
    {
      "epoch": 0.25664810126582277,
      "grad_norm": 0.17582979798316956,
      "learning_rate": 9.4571915915435e-06,
      "loss": 0.0,
      "step": 792
    },
    {
      "epoch": 0.2572962025316456,
      "grad_norm": 0.005818579345941544,
      "learning_rate": 9.454312066734624e-06,
      "loss": 0.0,
      "step": 794
    },
    {
      "epoch": 0.25794430379746836,
      "grad_norm": 0.01889118365943432,
      "learning_rate": 9.451425365140997e-06,
      "loss": 0.0002,
      "step": 796
    },
    {
      "epoch": 0.2585924050632911,
      "grad_norm": 0.00686600711196661,
      "learning_rate": 9.448531491413673e-06,
      "loss": 0.0,
      "step": 798
    },
    {
      "epoch": 0.25924050632911394,
      "grad_norm": 0.0015769926831126213,
      "learning_rate": 9.445630450215259e-06,
      "loss": 0.0,
      "step": 800
    },
    {
      "epoch": 0.25924050632911394,
      "eval_accuracy": 0.9999590503640193,
      "eval_loss": 0.00015591137344017625,
      "eval_runtime": 105.7845,
      "eval_samples_per_second": 47.266,
      "eval_steps_per_second": 11.816,
      "step": 800
    },
    {
      "epoch": 0.2598886075949367,
      "grad_norm": 0.06982430070638657,
      "learning_rate": 9.442722246219915e-06,
      "loss": 0.0,
      "step": 802
    },
    {
      "epoch": 0.26053670886075947,
      "grad_norm": 0.0993858277797699,
      "learning_rate": 9.439806884113331e-06,
      "loss": 0.0005,
      "step": 804
    },
    {
      "epoch": 0.2611848101265823,
      "grad_norm": 0.0011815716279670596,
      "learning_rate": 9.43688436859274e-06,
      "loss": 0.0007,
      "step": 806
    },
    {
      "epoch": 0.26183291139240505,
      "grad_norm": 0.0009712717146612704,
      "learning_rate": 9.433954704366897e-06,
      "loss": 0.0002,
      "step": 808
    },
    {
      "epoch": 0.2624810126582279,
      "grad_norm": 0.0008742675418034196,
      "learning_rate": 9.431017896156074e-06,
      "loss": 0.0001,
      "step": 810
    },
    {
      "epoch": 0.26312911392405064,
      "grad_norm": 0.1526733636856079,
      "learning_rate": 9.428073948692056e-06,
      "loss": 0.0001,
      "step": 812
    },
    {
      "epoch": 0.2637772151898734,
      "grad_norm": 0.008984135463833809,
      "learning_rate": 9.425122866718128e-06,
      "loss": 0.0,
      "step": 814
    },
    {
      "epoch": 0.2644253164556962,
      "grad_norm": 0.0007589040324091911,
      "learning_rate": 9.422164654989073e-06,
      "loss": 0.0003,
      "step": 816
    },
    {
      "epoch": 0.265073417721519,
      "grad_norm": 0.11218152195215225,
      "learning_rate": 9.419199318271158e-06,
      "loss": 0.0004,
      "step": 818
    },
    {
      "epoch": 0.26572151898734175,
      "grad_norm": 0.16130465269088745,
      "learning_rate": 9.416226861342132e-06,
      "loss": 0.0004,
      "step": 820
    },
    {
      "epoch": 0.2663696202531646,
      "grad_norm": 0.028433728963136673,
      "learning_rate": 9.413247288991216e-06,
      "loss": 0.0,
      "step": 822
    },
    {
      "epoch": 0.26701772151898734,
      "grad_norm": 0.3576768934726715,
      "learning_rate": 9.410260606019095e-06,
      "loss": 0.0009,
      "step": 824
    },
    {
      "epoch": 0.2676658227848101,
      "grad_norm": 0.03127586469054222,
      "learning_rate": 9.40726681723791e-06,
      "loss": 0.0,
      "step": 826
    },
    {
      "epoch": 0.2683139240506329,
      "grad_norm": 0.012010865844786167,
      "learning_rate": 9.404265927471255e-06,
      "loss": 0.0004,
      "step": 828
    },
    {
      "epoch": 0.2689620253164557,
      "grad_norm": 0.01447245478630066,
      "learning_rate": 9.401257941554157e-06,
      "loss": 0.0,
      "step": 830
    },
    {
      "epoch": 0.2696101265822785,
      "grad_norm": 0.07084700465202332,
      "learning_rate": 9.398242864333084e-06,
      "loss": 0.0008,
      "step": 832
    },
    {
      "epoch": 0.2702582278481013,
      "grad_norm": 0.23775817453861237,
      "learning_rate": 9.395220700665924e-06,
      "loss": 0.0002,
      "step": 834
    },
    {
      "epoch": 0.27090632911392404,
      "grad_norm": 0.27267056703567505,
      "learning_rate": 9.392191455421989e-06,
      "loss": 0.001,
      "step": 836
    },
    {
      "epoch": 0.27155443037974686,
      "grad_norm": 0.0014412677846848965,
      "learning_rate": 9.389155133481993e-06,
      "loss": 0.0,
      "step": 838
    },
    {
      "epoch": 0.2722025316455696,
      "grad_norm": 0.22994522750377655,
      "learning_rate": 9.386111739738057e-06,
      "loss": 0.0002,
      "step": 840
    },
    {
      "epoch": 0.2722025316455696,
      "eval_accuracy": 0.9999324854944309,
      "eval_loss": 0.00023814353335183114,
      "eval_runtime": 105.6702,
      "eval_samples_per_second": 47.317,
      "eval_steps_per_second": 11.829,
      "step": 840
    },
    {
      "epoch": 0.2728506329113924,
      "grad_norm": 0.16264504194259644,
      "learning_rate": 9.383061279093697e-06,
      "loss": 0.0008,
      "step": 842
    },
    {
      "epoch": 0.2734987341772152,
      "grad_norm": 0.21666684746742249,
      "learning_rate": 9.380003756463812e-06,
      "loss": 0.0003,
      "step": 844
    },
    {
      "epoch": 0.274146835443038,
      "grad_norm": 0.020299425348639488,
      "learning_rate": 9.376939176774678e-06,
      "loss": 0.0001,
      "step": 846
    },
    {
      "epoch": 0.27479493670886074,
      "grad_norm": 0.04464450851082802,
      "learning_rate": 9.373867544963949e-06,
      "loss": 0.0002,
      "step": 848
    },
    {
      "epoch": 0.27544303797468356,
      "grad_norm": 0.13788706064224243,
      "learning_rate": 9.370788865980633e-06,
      "loss": 0.0011,
      "step": 850
    },
    {
      "epoch": 0.2760911392405063,
      "grad_norm": 0.3510144352912903,
      "learning_rate": 9.367703144785097e-06,
      "loss": 0.001,
      "step": 852
    },
    {
      "epoch": 0.2767392405063291,
      "grad_norm": 0.048665471374988556,
      "learning_rate": 9.364610386349048e-06,
      "loss": 0.0001,
      "step": 854
    },
    {
      "epoch": 0.2773873417721519,
      "grad_norm": 0.02506359852850437,
      "learning_rate": 9.361510595655545e-06,
      "loss": 0.0006,
      "step": 856
    },
    {
      "epoch": 0.27803544303797467,
      "grad_norm": 0.1422048807144165,
      "learning_rate": 9.358403777698962e-06,
      "loss": 0.001,
      "step": 858
    },
    {
      "epoch": 0.2786835443037975,
      "grad_norm": 0.007875354960560799,
      "learning_rate": 9.355289937485005e-06,
      "loss": 0.0,
      "step": 860
    },
    {
      "epoch": 0.27933164556962026,
      "grad_norm": 0.009137582033872604,
      "learning_rate": 9.35216908003069e-06,
      "loss": 0.0,
      "step": 862
    },
    {
      "epoch": 0.279979746835443,
      "grad_norm": 0.020093752071261406,
      "learning_rate": 9.349041210364343e-06,
      "loss": 0.0003,
      "step": 864
    },
    {
      "epoch": 0.28062784810126584,
      "grad_norm": 0.0040946886874735355,
      "learning_rate": 9.345906333525582e-06,
      "loss": 0.0001,
      "step": 866
    },
    {
      "epoch": 0.2812759493670886,
      "grad_norm": 0.005941552110016346,
      "learning_rate": 9.342764454565321e-06,
      "loss": 0.0006,
      "step": 868
    },
    {
      "epoch": 0.28192405063291137,
      "grad_norm": 0.0024547127541154623,
      "learning_rate": 9.339615578545753e-06,
      "loss": 0.0,
      "step": 870
    },
    {
      "epoch": 0.2825721518987342,
      "grad_norm": 0.002859026426449418,
      "learning_rate": 9.336459710540344e-06,
      "loss": 0.0014,
      "step": 872
    },
    {
      "epoch": 0.28322025316455696,
      "grad_norm": 0.0011856557102873921,
      "learning_rate": 9.333296855633828e-06,
      "loss": 0.0,
      "step": 874
    },
    {
      "epoch": 0.2838683544303797,
      "grad_norm": 0.41726920008659363,
      "learning_rate": 9.330127018922195e-06,
      "loss": 0.0011,
      "step": 876
    },
    {
      "epoch": 0.28451645569620254,
      "grad_norm": 0.45102396607398987,
      "learning_rate": 9.326950205512682e-06,
      "loss": 0.0013,
      "step": 878
    },
    {
      "epoch": 0.2851645569620253,
      "grad_norm": 0.16208650171756744,
      "learning_rate": 9.323766420523768e-06,
      "loss": 0.0007,
      "step": 880
    },
    {
      "epoch": 0.2851645569620253,
      "eval_accuracy": 0.999945549853298,
      "eval_loss": 0.00017561792628839612,
      "eval_runtime": 105.2387,
      "eval_samples_per_second": 47.511,
      "eval_steps_per_second": 11.878,
      "step": 880
    },
    {
      "epoch": 0.2858126582278481,
      "grad_norm": 0.4453209638595581,
      "learning_rate": 9.32057566908517e-06,
      "loss": 0.0005,
      "step": 882
    },
    {
      "epoch": 0.2864607594936709,
      "grad_norm": 0.0023153494112193584,
      "learning_rate": 9.31737795633782e-06,
      "loss": 0.0,
      "step": 884
    },
    {
      "epoch": 0.28710886075949366,
      "grad_norm": 0.050638675689697266,
      "learning_rate": 9.314173287433874e-06,
      "loss": 0.0001,
      "step": 886
    },
    {
      "epoch": 0.2877569620253165,
      "grad_norm": 0.08961086720228195,
      "learning_rate": 9.310961667536689e-06,
      "loss": 0.0002,
      "step": 888
    },
    {
      "epoch": 0.28840506329113924,
      "grad_norm": 0.10929686576128006,
      "learning_rate": 9.307743101820828e-06,
      "loss": 0.0002,
      "step": 890
    },
    {
      "epoch": 0.289053164556962,
      "grad_norm": 0.010287183336913586,
      "learning_rate": 9.30451759547204e-06,
      "loss": 0.0,
      "step": 892
    },
    {
      "epoch": 0.2897012658227848,
      "grad_norm": 0.04592491313815117,
      "learning_rate": 9.301285153687261e-06,
      "loss": 0.0002,
      "step": 894
    },
    {
      "epoch": 0.2903493670886076,
      "grad_norm": 0.13677294552326202,
      "learning_rate": 9.298045781674595e-06,
      "loss": 0.0018,
      "step": 896
    },
    {
      "epoch": 0.29099746835443036,
      "grad_norm": 0.05407572165131569,
      "learning_rate": 9.294799484653323e-06,
      "loss": 0.0002,
      "step": 898
    },
    {
      "epoch": 0.2916455696202532,
      "grad_norm": 0.20654959976673126,
      "learning_rate": 9.291546267853871e-06,
      "loss": 0.0016,
      "step": 900
    },
    {
      "epoch": 0.29229367088607594,
      "grad_norm": 0.13771550357341766,
      "learning_rate": 9.28828613651782e-06,
      "loss": 0.0015,
      "step": 902
    },
    {
      "epoch": 0.29294177215189876,
      "grad_norm": 0.24270237982273102,
      "learning_rate": 9.285019095897894e-06,
      "loss": 0.0003,
      "step": 904
    },
    {
      "epoch": 0.2935898734177215,
      "grad_norm": 0.06425748020410538,
      "learning_rate": 9.281745151257946e-06,
      "loss": 0.0003,
      "step": 906
    },
    {
      "epoch": 0.2942379746835443,
      "grad_norm": 0.3214038014411926,
      "learning_rate": 9.278464307872952e-06,
      "loss": 0.0014,
      "step": 908
    },
    {
      "epoch": 0.2948860759493671,
      "grad_norm": 0.3001636266708374,
      "learning_rate": 9.275176571029008e-06,
      "loss": 0.0007,
      "step": 910
    },
    {
      "epoch": 0.2955341772151899,
      "grad_norm": 0.5608838796615601,
      "learning_rate": 9.271881946023309e-06,
      "loss": 0.0007,
      "step": 912
    },
    {
      "epoch": 0.29618227848101264,
      "grad_norm": 1.1040050983428955,
      "learning_rate": 9.268580438164157e-06,
      "loss": 0.0017,
      "step": 914
    },
    {
      "epoch": 0.29683037974683546,
      "grad_norm": 0.013682582415640354,
      "learning_rate": 9.265272052770936e-06,
      "loss": 0.0004,
      "step": 916
    },
    {
      "epoch": 0.2974784810126582,
      "grad_norm": 0.05224961042404175,
      "learning_rate": 9.261956795174116e-06,
      "loss": 0.0003,
      "step": 918
    },
    {
      "epoch": 0.298126582278481,
      "grad_norm": 0.2133917510509491,
      "learning_rate": 9.25863467071524e-06,
      "loss": 0.0004,
      "step": 920
    },
    {
      "epoch": 0.298126582278481,
      "eval_accuracy": 0.9998670163935581,
      "eval_loss": 0.00046491899411194026,
      "eval_runtime": 106.1612,
      "eval_samples_per_second": 47.098,
      "eval_steps_per_second": 11.775,
      "step": 920
    },
    {
      "epoch": 0.2987746835443038,
      "grad_norm": 0.28349632024765015,
      "learning_rate": 9.255305684746908e-06,
      "loss": 0.0003,
      "step": 922
    },
    {
      "epoch": 0.2994227848101266,
      "grad_norm": 0.21036334335803986,
      "learning_rate": 9.251969842632785e-06,
      "loss": 0.001,
      "step": 924
    },
    {
      "epoch": 0.3000708860759494,
      "grad_norm": 0.07668338716030121,
      "learning_rate": 9.248627149747573e-06,
      "loss": 0.0007,
      "step": 926
    },
    {
      "epoch": 0.30071898734177216,
      "grad_norm": 0.03533896058797836,
      "learning_rate": 9.24527761147702e-06,
      "loss": 0.0004,
      "step": 928
    },
    {
      "epoch": 0.3013670886075949,
      "grad_norm": 0.44851046800613403,
      "learning_rate": 9.241921233217899e-06,
      "loss": 0.0035,
      "step": 930
    },
    {
      "epoch": 0.30201518987341774,
      "grad_norm": 1.7721484899520874,
      "learning_rate": 9.238558020378003e-06,
      "loss": 0.0047,
      "step": 932
    },
    {
      "epoch": 0.3026632911392405,
      "grad_norm": 0.3624114692211151,
      "learning_rate": 9.235187978376141e-06,
      "loss": 0.0022,
      "step": 934
    },
    {
      "epoch": 0.3033113924050633,
      "grad_norm": 0.0940459743142128,
      "learning_rate": 9.231811112642121e-06,
      "loss": 0.0013,
      "step": 936
    },
    {
      "epoch": 0.3039594936708861,
      "grad_norm": 0.12335839867591858,
      "learning_rate": 9.228427428616749e-06,
      "loss": 0.0013,
      "step": 938
    },
    {
      "epoch": 0.30460759493670886,
      "grad_norm": 0.15589086711406708,
      "learning_rate": 9.225036931751811e-06,
      "loss": 0.0008,
      "step": 940
    },
    {
      "epoch": 0.3052556962025316,
      "grad_norm": 1.7561614513397217,
      "learning_rate": 9.221639627510076e-06,
      "loss": 0.0009,
      "step": 942
    },
    {
      "epoch": 0.30590379746835444,
      "grad_norm": 0.5845587253570557,
      "learning_rate": 9.218235521365278e-06,
      "loss": 0.0009,
      "step": 944
    },
    {
      "epoch": 0.3065518987341772,
      "grad_norm": 0.1916065365076065,
      "learning_rate": 9.214824618802108e-06,
      "loss": 0.0005,
      "step": 946
    },
    {
      "epoch": 0.3072,
      "grad_norm": 0.4482108950614929,
      "learning_rate": 9.211406925316214e-06,
      "loss": 0.0018,
      "step": 948
    },
    {
      "epoch": 0.3078481012658228,
      "grad_norm": 0.20742888748645782,
      "learning_rate": 9.20798244641418e-06,
      "loss": 0.0005,
      "step": 950
    },
    {
      "epoch": 0.30849620253164556,
      "grad_norm": 0.12374088913202286,
      "learning_rate": 9.204551187613521e-06,
      "loss": 0.0013,
      "step": 952
    },
    {
      "epoch": 0.3091443037974684,
      "grad_norm": 0.1217096671462059,
      "learning_rate": 9.201113154442685e-06,
      "loss": 0.0004,
      "step": 954
    },
    {
      "epoch": 0.30979240506329114,
      "grad_norm": 0.463105708360672,
      "learning_rate": 9.197668352441025e-06,
      "loss": 0.0006,
      "step": 956
    },
    {
      "epoch": 0.3104405063291139,
      "grad_norm": 0.21807028353214264,
      "learning_rate": 9.194216787158805e-06,
      "loss": 0.0006,
      "step": 958
    },
    {
      "epoch": 0.31108860759493673,
      "grad_norm": 0.29980653524398804,
      "learning_rate": 9.190758464157184e-06,
      "loss": 0.0004,
      "step": 960
    },
    {
      "epoch": 0.31108860759493673,
      "eval_accuracy": 0.9997704680437066,
      "eval_loss": 0.0012385541340336204,
      "eval_runtime": 105.1546,
      "eval_samples_per_second": 47.549,
      "eval_steps_per_second": 11.887,
      "step": 960
    },
    {
      "epoch": 0.3117367088607595,
      "grad_norm": 1.859007477760315,
      "learning_rate": 9.18729338900821e-06,
      "loss": 0.0002,
      "step": 962
    },
    {
      "epoch": 0.31238481012658226,
      "grad_norm": 0.0305470023304224,
      "learning_rate": 9.18382156729481e-06,
      "loss": 0.0005,
      "step": 964
    },
    {
      "epoch": 0.3130329113924051,
      "grad_norm": 0.1439788043498993,
      "learning_rate": 9.18034300461078e-06,
      "loss": 0.0002,
      "step": 966
    },
    {
      "epoch": 0.31368101265822784,
      "grad_norm": 0.1942347288131714,
      "learning_rate": 9.17685770656078e-06,
      "loss": 0.0006,
      "step": 968
    },
    {
      "epoch": 0.3143291139240506,
      "grad_norm": 15.445067405700684,
      "learning_rate": 9.173365678760318e-06,
      "loss": 0.0054,
      "step": 970
    },
    {
      "epoch": 0.3149772151898734,
      "grad_norm": 0.431581050157547,
      "learning_rate": 9.169866926835749e-06,
      "loss": 0.0027,
      "step": 972
    },
    {
      "epoch": 0.3156253164556962,
      "grad_norm": 0.2632336914539337,
      "learning_rate": 9.166361456424257e-06,
      "loss": 0.0007,
      "step": 974
    },
    {
      "epoch": 0.316273417721519,
      "grad_norm": 1.0372949838638306,
      "learning_rate": 9.162849273173857e-06,
      "loss": 0.0022,
      "step": 976
    },
    {
      "epoch": 0.3169215189873418,
      "grad_norm": 0.19878140091896057,
      "learning_rate": 9.159330382743375e-06,
      "loss": 0.0019,
      "step": 978
    },
    {
      "epoch": 0.31756962025316454,
      "grad_norm": 0.7643746733665466,
      "learning_rate": 9.155804790802444e-06,
      "loss": 0.0015,
      "step": 980
    },
    {
      "epoch": 0.31821772151898736,
      "grad_norm": 0.17058204114437103,
      "learning_rate": 9.152272503031496e-06,
      "loss": 0.0009,
      "step": 982
    },
    {
      "epoch": 0.3188658227848101,
      "grad_norm": 0.32723182439804077,
      "learning_rate": 9.148733525121751e-06,
      "loss": 0.0015,
      "step": 984
    },
    {
      "epoch": 0.3195139240506329,
      "grad_norm": 0.25325924158096313,
      "learning_rate": 9.145187862775208e-06,
      "loss": 0.0019,
      "step": 986
    },
    {
      "epoch": 0.3201620253164557,
      "grad_norm": 0.5356621146202087,
      "learning_rate": 9.141635521704638e-06,
      "loss": 0.0008,
      "step": 988
    },
    {
      "epoch": 0.3208101265822785,
      "grad_norm": 0.04259486123919487,
      "learning_rate": 9.138076507633566e-06,
      "loss": 0.0003,
      "step": 990
    },
    {
      "epoch": 0.32145822784810124,
      "grad_norm": 0.5347451567649841,
      "learning_rate": 9.134510826296277e-06,
      "loss": 0.0005,
      "step": 992
    },
    {
      "epoch": 0.32210632911392406,
      "grad_norm": 0.09377581626176834,
      "learning_rate": 9.130938483437792e-06,
      "loss": 0.0006,
      "step": 994
    },
    {
      "epoch": 0.3227544303797468,
      "grad_norm": 0.15030620992183685,
      "learning_rate": 9.12735948481387e-06,
      "loss": 0.0004,
      "step": 996
    },
    {
      "epoch": 0.32340253164556965,
      "grad_norm": 0.20812305808067322,
      "learning_rate": 9.12377383619099e-06,
      "loss": 0.0006,
      "step": 998
    },
    {
      "epoch": 0.3240506329113924,
      "grad_norm": 1.2129936218261719,
      "learning_rate": 9.120181543346348e-06,
      "loss": 0.0008,
      "step": 1000
    },
    {
      "epoch": 0.3240506329113924,
      "eval_accuracy": 0.9998334119906976,
      "eval_loss": 0.0007947856211103499,
      "eval_runtime": 105.0477,
      "eval_samples_per_second": 47.597,
      "eval_steps_per_second": 11.899,
      "step": 1000
    },
    {
      "epoch": 0.3246987341772152,
      "grad_norm": 0.22441734373569489,
      "learning_rate": 9.11658261206784e-06,
      "loss": 0.0002,
      "step": 1002
    },
    {
      "epoch": 0.325346835443038,
      "grad_norm": 0.46537208557128906,
      "learning_rate": 9.112977048154066e-06,
      "loss": 0.0005,
      "step": 1004
    },
    {
      "epoch": 0.32599493670886076,
      "grad_norm": 0.56826251745224,
      "learning_rate": 9.109364857414306e-06,
      "loss": 0.0035,
      "step": 1006
    },
    {
      "epoch": 0.3266430379746835,
      "grad_norm": 0.10446134209632874,
      "learning_rate": 9.10574604566852e-06,
      "loss": 0.0007,
      "step": 1008
    },
    {
      "epoch": 0.32729113924050635,
      "grad_norm": 0.1642230600118637,
      "learning_rate": 9.102120618747336e-06,
      "loss": 0.001,
      "step": 1010
    },
    {
      "epoch": 0.3279392405063291,
      "grad_norm": 0.06929557770490646,
      "learning_rate": 9.09848858249204e-06,
      "loss": 0.0002,
      "step": 1012
    },
    {
      "epoch": 0.3285873417721519,
      "grad_norm": 0.2563181221485138,
      "learning_rate": 9.094849942754564e-06,
      "loss": 0.0005,
      "step": 1014
    },
    {
      "epoch": 0.3292354430379747,
      "grad_norm": 0.41675665974617004,
      "learning_rate": 9.091204705397485e-06,
      "loss": 0.001,
      "step": 1016
    },
    {
      "epoch": 0.32988354430379746,
      "grad_norm": 0.02498539723455906,
      "learning_rate": 9.087552876294003e-06,
      "loss": 0.0001,
      "step": 1018
    },
    {
      "epoch": 0.3305316455696203,
      "grad_norm": 0.0656026154756546,
      "learning_rate": 9.083894461327946e-06,
      "loss": 0.0003,
      "step": 1020
    },
    {
      "epoch": 0.33117974683544305,
      "grad_norm": 0.11689463257789612,
      "learning_rate": 9.08022946639375e-06,
      "loss": 0.0003,
      "step": 1022
    },
    {
      "epoch": 0.3318278481012658,
      "grad_norm": 0.41019096970558167,
      "learning_rate": 9.076557897396452e-06,
      "loss": 0.0003,
      "step": 1024
    },
    {
      "epoch": 0.33247594936708863,
      "grad_norm": 0.5071588754653931,
      "learning_rate": 9.07287976025168e-06,
      "loss": 0.0009,
      "step": 1026
    },
    {
      "epoch": 0.3331240506329114,
      "grad_norm": 0.7110933661460876,
      "learning_rate": 9.069195060885647e-06,
      "loss": 0.0007,
      "step": 1028
    },
    {
      "epoch": 0.33377215189873416,
      "grad_norm": 0.4573919177055359,
      "learning_rate": 9.065503805235139e-06,
      "loss": 0.0008,
      "step": 1030
    },
    {
      "epoch": 0.334420253164557,
      "grad_norm": 0.5230830311775208,
      "learning_rate": 9.061805999247504e-06,
      "loss": 0.0009,
      "step": 1032
    },
    {
      "epoch": 0.33506835443037974,
      "grad_norm": 0.4800056219100952,
      "learning_rate": 9.058101648880646e-06,
      "loss": 0.0009,
      "step": 1034
    },
    {
      "epoch": 0.3357164556962025,
      "grad_norm": 0.30046916007995605,
      "learning_rate": 9.05439076010301e-06,
      "loss": 0.001,
      "step": 1036
    },
    {
      "epoch": 0.33636455696202533,
      "grad_norm": 0.1976243406534195,
      "learning_rate": 9.050673338893578e-06,
      "loss": 0.0004,
      "step": 1038
    },
    {
      "epoch": 0.3370126582278481,
      "grad_norm": 0.04080143943428993,
      "learning_rate": 9.046949391241859e-06,
      "loss": 0.0001,
      "step": 1040
    },
    {
      "epoch": 0.3370126582278481,
      "eval_accuracy": 0.9999459609045181,
      "eval_loss": 0.0002253134734928608,
      "eval_runtime": 105.8583,
      "eval_samples_per_second": 47.233,
      "eval_steps_per_second": 11.808,
      "step": 1040
    },
    {
      "epoch": 0.33766075949367086,
      "grad_norm": 0.04524859040975571,
      "learning_rate": 9.043218923147874e-06,
      "loss": 0.0001,
      "step": 1042
    },
    {
      "epoch": 0.3383088607594937,
      "grad_norm": 0.16028116643428802,
      "learning_rate": 9.039481940622148e-06,
      "loss": 0.0001,
      "step": 1044
    },
    {
      "epoch": 0.33895696202531644,
      "grad_norm": 0.12261617928743362,
      "learning_rate": 9.035738449685707e-06,
      "loss": 0.0002,
      "step": 1046
    },
    {
      "epoch": 0.33960506329113926,
      "grad_norm": 0.06535837799310684,
      "learning_rate": 9.031988456370062e-06,
      "loss": 0.0002,
      "step": 1048
    },
    {
      "epoch": 0.34025316455696203,
      "grad_norm": 0.258878231048584,
      "learning_rate": 9.0282319667172e-06,
      "loss": 0.0002,
      "step": 1050
    },
    {
      "epoch": 0.3409012658227848,
      "grad_norm": 0.04995712265372276,
      "learning_rate": 9.02446898677957e-06,
      "loss": 0.0002,
      "step": 1052
    },
    {
      "epoch": 0.3415493670886076,
      "grad_norm": 0.03698616847395897,
      "learning_rate": 9.020699522620091e-06,
      "loss": 0.0004,
      "step": 1054
    },
    {
      "epoch": 0.3421974683544304,
      "grad_norm": 0.0686570331454277,
      "learning_rate": 9.016923580312114e-06,
      "loss": 0.0003,
      "step": 1056
    },
    {
      "epoch": 0.34284556962025314,
      "grad_norm": 0.011084304191172123,
      "learning_rate": 9.013141165939439e-06,
      "loss": 0.0006,
      "step": 1058
    },
    {
      "epoch": 0.34349367088607596,
      "grad_norm": 0.07190306484699249,
      "learning_rate": 9.009352285596287e-06,
      "loss": 0.0,
      "step": 1060
    },
    {
      "epoch": 0.34414177215189873,
      "grad_norm": 0.12991821765899658,
      "learning_rate": 9.005556945387301e-06,
      "loss": 0.0003,
      "step": 1062
    },
    {
      "epoch": 0.3447898734177215,
      "grad_norm": 0.11152586340904236,
      "learning_rate": 9.001755151427532e-06,
      "loss": 0.0007,
      "step": 1064
    },
    {
      "epoch": 0.3454379746835443,
      "grad_norm": 0.08615229278802872,
      "learning_rate": 8.997946909842426e-06,
      "loss": 0.0003,
      "step": 1066
    },
    {
      "epoch": 0.3460860759493671,
      "grad_norm": 0.11003229767084122,
      "learning_rate": 8.99413222676782e-06,
      "loss": 0.0003,
      "step": 1068
    },
    {
      "epoch": 0.3467341772151899,
      "grad_norm": 0.4303703010082245,
      "learning_rate": 8.990311108349926e-06,
      "loss": 0.0003,
      "step": 1070
    },
    {
      "epoch": 0.34738227848101266,
      "grad_norm": 0.45959335565567017,
      "learning_rate": 8.986483560745335e-06,
      "loss": 0.0008,
      "step": 1072
    },
    {
      "epoch": 0.3480303797468354,
      "grad_norm": 0.38193613290786743,
      "learning_rate": 8.982649590120982e-06,
      "loss": 0.0001,
      "step": 1074
    },
    {
      "epoch": 0.34867848101265825,
      "grad_norm": 1.4580763578414917,
      "learning_rate": 8.978809202654161e-06,
      "loss": 0.0006,
      "step": 1076
    },
    {
      "epoch": 0.349326582278481,
      "grad_norm": 0.23947280645370483,
      "learning_rate": 8.974962404532503e-06,
      "loss": 0.0009,
      "step": 1078
    },
    {
      "epoch": 0.3499746835443038,
      "grad_norm": 0.40106865763664246,
      "learning_rate": 8.971109201953962e-06,
      "loss": 0.0008,
      "step": 1080
    },
    {
      "epoch": 0.3499746835443038,
      "eval_accuracy": 0.9997245806688072,
      "eval_loss": 0.00159825652372092,
      "eval_runtime": 105.5986,
      "eval_samples_per_second": 47.349,
      "eval_steps_per_second": 11.837,
      "step": 1080
    },
    {
      "epoch": 0.3506227848101266,
      "grad_norm": 0.16527506709098816,
      "learning_rate": 8.967249601126821e-06,
      "loss": 0.0006,
      "step": 1082
    },
    {
      "epoch": 0.35127088607594936,
      "grad_norm": 0.4305400252342224,
      "learning_rate": 8.963383608269665e-06,
      "loss": 0.0009,
      "step": 1084
    },
    {
      "epoch": 0.3519189873417721,
      "grad_norm": 0.07541073113679886,
      "learning_rate": 8.959511229611377e-06,
      "loss": 0.0002,
      "step": 1086
    },
    {
      "epoch": 0.35256708860759495,
      "grad_norm": 0.04376499727368355,
      "learning_rate": 8.955632471391132e-06,
      "loss": 0.0002,
      "step": 1088
    },
    {
      "epoch": 0.3532151898734177,
      "grad_norm": 0.3669925332069397,
      "learning_rate": 8.951747339858383e-06,
      "loss": 0.0004,
      "step": 1090
    },
    {
      "epoch": 0.35386329113924053,
      "grad_norm": 0.018304547294974327,
      "learning_rate": 8.947855841272852e-06,
      "loss": 0.0001,
      "step": 1092
    },
    {
      "epoch": 0.3545113924050633,
      "grad_norm": 0.15526734292507172,
      "learning_rate": 8.943957981904518e-06,
      "loss": 0.0006,
      "step": 1094
    },
    {
      "epoch": 0.35515949367088606,
      "grad_norm": 0.04184693470597267,
      "learning_rate": 8.94005376803361e-06,
      "loss": 0.0002,
      "step": 1096
    },
    {
      "epoch": 0.3558075949367089,
      "grad_norm": 0.19317471981048584,
      "learning_rate": 8.936143205950596e-06,
      "loss": 0.0008,
      "step": 1098
    },
    {
      "epoch": 0.35645569620253165,
      "grad_norm": 0.14545564353466034,
      "learning_rate": 8.93222630195617e-06,
      "loss": 0.0005,
      "step": 1100
    },
    {
      "epoch": 0.3571037974683544,
      "grad_norm": 0.26518160104751587,
      "learning_rate": 8.928303062361244e-06,
      "loss": 0.0005,
      "step": 1102
    },
    {
      "epoch": 0.35775189873417723,
      "grad_norm": 0.37797990441322327,
      "learning_rate": 8.924373493486941e-06,
      "loss": 0.0002,
      "step": 1104
    },
    {
      "epoch": 0.3584,
      "grad_norm": 0.17237479984760284,
      "learning_rate": 8.92043760166458e-06,
      "loss": 0.0001,
      "step": 1106
    },
    {
      "epoch": 0.35904810126582276,
      "grad_norm": 0.04698828235268593,
      "learning_rate": 8.916495393235666e-06,
      "loss": 0.0008,
      "step": 1108
    },
    {
      "epoch": 0.3596962025316456,
      "grad_norm": 0.04225539043545723,
      "learning_rate": 8.912546874551883e-06,
      "loss": 0.0004,
      "step": 1110
    },
    {
      "epoch": 0.36034430379746835,
      "grad_norm": 0.041860613971948624,
      "learning_rate": 8.908592051975083e-06,
      "loss": 0.0003,
      "step": 1112
    },
    {
      "epoch": 0.36099240506329117,
      "grad_norm": 0.49554023146629333,
      "learning_rate": 8.904630931877271e-06,
      "loss": 0.0011,
      "step": 1114
    },
    {
      "epoch": 0.36164050632911393,
      "grad_norm": 0.019335318356752396,
      "learning_rate": 8.900663520640605e-06,
      "loss": 0.0011,
      "step": 1116
    },
    {
      "epoch": 0.3622886075949367,
      "grad_norm": 1.3709319829940796,
      "learning_rate": 8.896689824657371e-06,
      "loss": 0.0003,
      "step": 1118
    },
    {
      "epoch": 0.3629367088607595,
      "grad_norm": 0.5006961226463318,
      "learning_rate": 8.892709850329991e-06,
      "loss": 0.0016,
      "step": 1120
    },
    {
      "epoch": 0.3629367088607595,
      "eval_accuracy": 0.9998284606779089,
      "eval_loss": 0.0006213401211425662,
      "eval_runtime": 104.31,
      "eval_samples_per_second": 47.934,
      "eval_steps_per_second": 11.984,
      "step": 1120
    },
    {
      "epoch": 0.3635848101265823,
      "grad_norm": 0.29077577590942383,
      "learning_rate": 8.88872360407099e-06,
      "loss": 0.0008,
      "step": 1122
    },
    {
      "epoch": 0.36423291139240505,
      "grad_norm": 0.18928442895412445,
      "learning_rate": 8.884731092303011e-06,
      "loss": 0.0006,
      "step": 1124
    },
    {
      "epoch": 0.36488101265822787,
      "grad_norm": 0.011645007878541946,
      "learning_rate": 8.880732321458785e-06,
      "loss": 0.0001,
      "step": 1126
    },
    {
      "epoch": 0.36552911392405063,
      "grad_norm": 1.1214470863342285,
      "learning_rate": 8.876727297981129e-06,
      "loss": 0.0007,
      "step": 1128
    },
    {
      "epoch": 0.3661772151898734,
      "grad_norm": 0.548606812953949,
      "learning_rate": 8.872716028322931e-06,
      "loss": 0.0007,
      "step": 1130
    },
    {
      "epoch": 0.3668253164556962,
      "grad_norm": 0.0687018632888794,
      "learning_rate": 8.868698518947152e-06,
      "loss": 0.0001,
      "step": 1132
    },
    {
      "epoch": 0.367473417721519,
      "grad_norm": 0.5724080204963684,
      "learning_rate": 8.864674776326798e-06,
      "loss": 0.0006,
      "step": 1134
    },
    {
      "epoch": 0.36812151898734174,
      "grad_norm": 0.08760637789964676,
      "learning_rate": 8.860644806944917e-06,
      "loss": 0.0006,
      "step": 1136
    },
    {
      "epoch": 0.36876962025316457,
      "grad_norm": 0.08993974328041077,
      "learning_rate": 8.8566086172946e-06,
      "loss": 0.0003,
      "step": 1138
    },
    {
      "epoch": 0.36941772151898733,
      "grad_norm": 0.108463354408741,
      "learning_rate": 8.852566213878947e-06,
      "loss": 0.0007,
      "step": 1140
    },
    {
      "epoch": 0.37006582278481015,
      "grad_norm": 0.054592058062553406,
      "learning_rate": 8.84851760321108e-06,
      "loss": 0.0001,
      "step": 1142
    },
    {
      "epoch": 0.3707139240506329,
      "grad_norm": 0.16010713577270508,
      "learning_rate": 8.844462791814113e-06,
      "loss": 0.0003,
      "step": 1144
    },
    {
      "epoch": 0.3713620253164557,
      "grad_norm": 0.04880848154425621,
      "learning_rate": 8.84040178622116e-06,
      "loss": 0.0002,
      "step": 1146
    },
    {
      "epoch": 0.3720101265822785,
      "grad_norm": 0.16024944186210632,
      "learning_rate": 8.83633459297531e-06,
      "loss": 0.0004,
      "step": 1148
    },
    {
      "epoch": 0.37265822784810126,
      "grad_norm": 0.04184041544795036,
      "learning_rate": 8.83226121862962e-06,
      "loss": 0.0002,
      "step": 1150
    },
    {
      "epoch": 0.37330632911392403,
      "grad_norm": 0.006558520253747702,
      "learning_rate": 8.828181669747111e-06,
      "loss": 0.0002,
      "step": 1152
    },
    {
      "epoch": 0.37395443037974685,
      "grad_norm": 0.2030220627784729,
      "learning_rate": 8.824095952900746e-06,
      "loss": 0.0006,
      "step": 1154
    },
    {
      "epoch": 0.3746025316455696,
      "grad_norm": 0.015400746837258339,
      "learning_rate": 8.820004074673433e-06,
      "loss": 0.0001,
      "step": 1156
    },
    {
      "epoch": 0.3752506329113924,
      "grad_norm": 0.10592592507600784,
      "learning_rate": 8.815906041658001e-06,
      "loss": 0.0005,
      "step": 1158
    },
    {
      "epoch": 0.3758987341772152,
      "grad_norm": 0.17723920941352844,
      "learning_rate": 8.8118018604572e-06,
      "loss": 0.0003,
      "step": 1160
    },
    {
      "epoch": 0.3758987341772152,
      "eval_accuracy": 0.9999317506510353,
      "eval_loss": 0.0002906288718804717,
      "eval_runtime": 105.6508,
      "eval_samples_per_second": 47.326,
      "eval_steps_per_second": 11.831,
      "step": 1160
    },
    {
      "epoch": 0.37654683544303796,
      "grad_norm": 0.05428408086299896,
      "learning_rate": 8.807691537683685e-06,
      "loss": 0.0007,
      "step": 1162
    },
    {
      "epoch": 0.3771949367088608,
      "grad_norm": 0.12672647833824158,
      "learning_rate": 8.80357507996e-06,
      "loss": 0.0009,
      "step": 1164
    },
    {
      "epoch": 0.37784303797468355,
      "grad_norm": 0.035302940756082535,
      "learning_rate": 8.799452493918586e-06,
      "loss": 0.0,
      "step": 1166
    },
    {
      "epoch": 0.3784911392405063,
      "grad_norm": 0.01754443719983101,
      "learning_rate": 8.795323786201746e-06,
      "loss": 0.0,
      "step": 1168
    },
    {
      "epoch": 0.37913924050632913,
      "grad_norm": 0.21614503860473633,
      "learning_rate": 8.791188963461653e-06,
      "loss": 0.001,
      "step": 1170
    },
    {
      "epoch": 0.3797873417721519,
      "grad_norm": 0.0031652958132326603,
      "learning_rate": 8.787048032360332e-06,
      "loss": 0.001,
      "step": 1172
    },
    {
      "epoch": 0.38043544303797466,
      "grad_norm": 0.05934705585241318,
      "learning_rate": 8.782900999569646e-06,
      "loss": 0.0001,
      "step": 1174
    },
    {
      "epoch": 0.3810835443037975,
      "grad_norm": 0.01244109682738781,
      "learning_rate": 8.778747871771293e-06,
      "loss": 0.0,
      "step": 1176
    },
    {
      "epoch": 0.38173164556962025,
      "grad_norm": 0.017785053700208664,
      "learning_rate": 8.774588655656787e-06,
      "loss": 0.0,
      "step": 1178
    },
    {
      "epoch": 0.382379746835443,
      "grad_norm": 0.17515426874160767,
      "learning_rate": 8.770423357927463e-06,
      "loss": 0.0022,
      "step": 1180
    },
    {
      "epoch": 0.38302784810126583,
      "grad_norm": 0.2343391627073288,
      "learning_rate": 8.766251985294435e-06,
      "loss": 0.0001,
      "step": 1182
    },
    {
      "epoch": 0.3836759493670886,
      "grad_norm": 0.17780566215515137,
      "learning_rate": 8.762074544478622e-06,
      "loss": 0.0001,
      "step": 1184
    },
    {
      "epoch": 0.3843240506329114,
      "grad_norm": 0.005761538632214069,
      "learning_rate": 8.757891042210713e-06,
      "loss": 0.0,
      "step": 1186
    },
    {
      "epoch": 0.3849721518987342,
      "grad_norm": 0.13648106157779694,
      "learning_rate": 8.753701485231165e-06,
      "loss": 0.0005,
      "step": 1188
    },
    {
      "epoch": 0.38562025316455695,
      "grad_norm": 0.025816872715950012,
      "learning_rate": 8.749505880290188e-06,
      "loss": 0.0001,
      "step": 1190
    },
    {
      "epoch": 0.38626835443037977,
      "grad_norm": 0.04343462362885475,
      "learning_rate": 8.74530423414774e-06,
      "loss": 0.0002,
      "step": 1192
    },
    {
      "epoch": 0.38691645569620253,
      "grad_norm": 0.08946521580219269,
      "learning_rate": 8.741096553573506e-06,
      "loss": 0.0001,
      "step": 1194
    },
    {
      "epoch": 0.3875645569620253,
      "grad_norm": 0.06926504522562027,
      "learning_rate": 8.736882845346906e-06,
      "loss": 0.0001,
      "step": 1196
    },
    {
      "epoch": 0.3882126582278481,
      "grad_norm": 0.13481925427913666,
      "learning_rate": 8.732663116257057e-06,
      "loss": 0.0005,
      "step": 1198
    },
    {
      "epoch": 0.3888607594936709,
      "grad_norm": 0.10871945321559906,
      "learning_rate": 8.728437373102784e-06,
      "loss": 0.0002,
      "step": 1200
    },
    {
      "epoch": 0.3888607594936709,
      "eval_accuracy": 0.9999794378783441,
      "eval_loss": 7.780938176438212e-05,
      "eval_runtime": 105.7236,
      "eval_samples_per_second": 47.293,
      "eval_steps_per_second": 11.823,
      "step": 1200
    },
    {
      "epoch": 0.38950886075949365,
      "grad_norm": 0.011972387321293354,
      "learning_rate": 8.724205622692608e-06,
      "loss": 0.0001,
      "step": 1202
    },
    {
      "epoch": 0.39015696202531647,
      "grad_norm": 0.14289769530296326,
      "learning_rate": 8.719967871844715e-06,
      "loss": 0.0,
      "step": 1204
    },
    {
      "epoch": 0.39080506329113923,
      "grad_norm": 0.06000698730349541,
      "learning_rate": 8.715724127386971e-06,
      "loss": 0.0001,
      "step": 1206
    },
    {
      "epoch": 0.39145316455696205,
      "grad_norm": 0.055760614573955536,
      "learning_rate": 8.711474396156894e-06,
      "loss": 0.0001,
      "step": 1208
    },
    {
      "epoch": 0.3921012658227848,
      "grad_norm": 0.04712485894560814,
      "learning_rate": 8.707218685001648e-06,
      "loss": 0.0001,
      "step": 1210
    },
    {
      "epoch": 0.3927493670886076,
      "grad_norm": 0.022366875782608986,
      "learning_rate": 8.702957000778029e-06,
      "loss": 0.0002,
      "step": 1212
    },
    {
      "epoch": 0.3933974683544304,
      "grad_norm": 0.47528231143951416,
      "learning_rate": 8.698689350352465e-06,
      "loss": 0.0002,
      "step": 1214
    },
    {
      "epoch": 0.39404556962025317,
      "grad_norm": 0.10005582869052887,
      "learning_rate": 8.69441574060099e-06,
      "loss": 0.0001,
      "step": 1216
    },
    {
      "epoch": 0.39469367088607593,
      "grad_norm": 0.4457380473613739,
      "learning_rate": 8.690136178409237e-06,
      "loss": 0.0002,
      "step": 1218
    },
    {
      "epoch": 0.39534177215189875,
      "grad_norm": 0.013173989951610565,
      "learning_rate": 8.685850670672438e-06,
      "loss": 0.0,
      "step": 1220
    },
    {
      "epoch": 0.3959898734177215,
      "grad_norm": 0.003624015487730503,
      "learning_rate": 8.681559224295401e-06,
      "loss": 0.0,
      "step": 1222
    },
    {
      "epoch": 0.3966379746835443,
      "grad_norm": 0.23442572355270386,
      "learning_rate": 8.6772618461925e-06,
      "loss": 0.0002,
      "step": 1224
    },
    {
      "epoch": 0.3972860759493671,
      "grad_norm": 0.3089965283870697,
      "learning_rate": 8.672958543287666e-06,
      "loss": 0.0002,
      "step": 1226
    },
    {
      "epoch": 0.39793417721518987,
      "grad_norm": 0.18517152965068817,
      "learning_rate": 8.668649322514382e-06,
      "loss": 0.0001,
      "step": 1228
    },
    {
      "epoch": 0.39858227848101263,
      "grad_norm": 0.6624205708503723,
      "learning_rate": 8.66433419081566e-06,
      "loss": 0.0006,
      "step": 1230
    },
    {
      "epoch": 0.39923037974683545,
      "grad_norm": 0.02062745951116085,
      "learning_rate": 8.660013155144036e-06,
      "loss": 0.0001,
      "step": 1232
    },
    {
      "epoch": 0.3998784810126582,
      "grad_norm": 0.035230252891778946,
      "learning_rate": 8.655686222461561e-06,
      "loss": 0.0002,
      "step": 1234
    },
    {
      "epoch": 0.40052658227848104,
      "grad_norm": 0.17802561819553375,
      "learning_rate": 8.651353399739787e-06,
      "loss": 0.0001,
      "step": 1236
    },
    {
      "epoch": 0.4011746835443038,
      "grad_norm": 0.07029800117015839,
      "learning_rate": 8.647014693959754e-06,
      "loss": 0.0003,
      "step": 1238
    },
    {
      "epoch": 0.40182278481012657,
      "grad_norm": 0.04419781640172005,
      "learning_rate": 8.642670112111982e-06,
      "loss": 0.0001,
      "step": 1240
    },
    {
      "epoch": 0.40182278481012657,
      "eval_accuracy": 0.9998813481127686,
      "eval_loss": 0.00046045694034546614,
      "eval_runtime": 107.728,
      "eval_samples_per_second": 46.413,
      "eval_steps_per_second": 11.603,
      "step": 1240
    },
    {
      "epoch": 0.4024708860759494,
      "grad_norm": 0.37598854303359985,
      "learning_rate": 8.63831966119646e-06,
      "loss": 0.0008,
      "step": 1242
    },
    {
      "epoch": 0.40311898734177215,
      "grad_norm": 0.02678529918193817,
      "learning_rate": 8.633963348222628e-06,
      "loss": 0.0005,
      "step": 1244
    },
    {
      "epoch": 0.4037670886075949,
      "grad_norm": 0.040035102516412735,
      "learning_rate": 8.629601180209382e-06,
      "loss": 0.0001,
      "step": 1246
    },
    {
      "epoch": 0.40441518987341774,
      "grad_norm": 0.044874873012304306,
      "learning_rate": 8.625233164185035e-06,
      "loss": 0.0,
      "step": 1248
    },
    {
      "epoch": 0.4050632911392405,
      "grad_norm": 0.11809933930635452,
      "learning_rate": 8.620859307187339e-06,
      "loss": 0.0002,
      "step": 1250
    },
    {
      "epoch": 0.40571139240506326,
      "grad_norm": 0.07839110493659973,
      "learning_rate": 8.616479616263444e-06,
      "loss": 0.0,
      "step": 1252
    },
    {
      "epoch": 0.4063594936708861,
      "grad_norm": 0.20940014719963074,
      "learning_rate": 8.61209409846991e-06,
      "loss": 0.0005,
      "step": 1254
    },
    {
      "epoch": 0.40700759493670885,
      "grad_norm": 0.00456192996352911,
      "learning_rate": 8.607702760872679e-06,
      "loss": 0.0004,
      "step": 1256
    },
    {
      "epoch": 0.40765569620253167,
      "grad_norm": 0.01556658186018467,
      "learning_rate": 8.60330561054707e-06,
      "loss": 0.0,
      "step": 1258
    },
    {
      "epoch": 0.40830379746835443,
      "grad_norm": 0.0071419524028897285,
      "learning_rate": 8.598902654577768e-06,
      "loss": 0.0001,
      "step": 1260
    },
    {
      "epoch": 0.4089518987341772,
      "grad_norm": 0.10579927265644073,
      "learning_rate": 8.594493900058817e-06,
      "loss": 0.0001,
      "step": 1262
    },
    {
      "epoch": 0.4096,
      "grad_norm": 0.08323294669389725,
      "learning_rate": 8.590079354093594e-06,
      "loss": 0.0002,
      "step": 1264
    },
    {
      "epoch": 0.4102481012658228,
      "grad_norm": 0.06471659243106842,
      "learning_rate": 8.585659023794818e-06,
      "loss": 0.0001,
      "step": 1266
    },
    {
      "epoch": 0.41089620253164555,
      "grad_norm": 0.27636727690696716,
      "learning_rate": 8.581232916284519e-06,
      "loss": 0.0002,
      "step": 1268
    },
    {
      "epoch": 0.41154430379746837,
      "grad_norm": 0.009334628470242023,
      "learning_rate": 8.57680103869404e-06,
      "loss": 0.0,
      "step": 1270
    },
    {
      "epoch": 0.41219240506329113,
      "grad_norm": 0.00627860426902771,
      "learning_rate": 8.572363398164017e-06,
      "loss": 0.0,
      "step": 1272
    },
    {
      "epoch": 0.4128405063291139,
      "grad_norm": 0.06497501581907272,
      "learning_rate": 8.567920001844376e-06,
      "loss": 0.0003,
      "step": 1274
    },
    {
      "epoch": 0.4134886075949367,
      "grad_norm": 0.0032212065998464823,
      "learning_rate": 8.563470856894316e-06,
      "loss": 0.0002,
      "step": 1276
    },
    {
      "epoch": 0.4141367088607595,
      "grad_norm": 0.15634731948375702,
      "learning_rate": 8.559015970482292e-06,
      "loss": 0.0012,
      "step": 1278
    },
    {
      "epoch": 0.4147848101265823,
      "grad_norm": 0.2659966051578522,
      "learning_rate": 8.554555349786016e-06,
      "loss": 0.0005,
      "step": 1280
    },
    {
      "epoch": 0.4147848101265823,
      "eval_accuracy": 0.9999672238180145,
      "eval_loss": 0.00013511251017916948,
      "eval_runtime": 107.0127,
      "eval_samples_per_second": 46.723,
      "eval_steps_per_second": 11.681,
      "step": 1280
    },
    {
      "epoch": 0.41543291139240507,
      "grad_norm": 0.14331784844398499,
      "learning_rate": 8.550089001992438e-06,
      "loss": 0.0005,
      "step": 1282
    },
    {
      "epoch": 0.41608101265822783,
      "grad_norm": 0.3357856571674347,
      "learning_rate": 8.545616934297733e-06,
      "loss": 0.0006,
      "step": 1284
    },
    {
      "epoch": 0.41672911392405065,
      "grad_norm": 0.08820825070142746,
      "learning_rate": 8.541139153907296e-06,
      "loss": 0.0001,
      "step": 1286
    },
    {
      "epoch": 0.4173772151898734,
      "grad_norm": 0.11874289065599442,
      "learning_rate": 8.536655668035723e-06,
      "loss": 0.0003,
      "step": 1288
    },
    {
      "epoch": 0.4180253164556962,
      "grad_norm": 0.07864240556955338,
      "learning_rate": 8.532166483906804e-06,
      "loss": 0.0003,
      "step": 1290
    },
    {
      "epoch": 0.418673417721519,
      "grad_norm": 0.008136331103742123,
      "learning_rate": 8.527671608753508e-06,
      "loss": 0.0001,
      "step": 1292
    },
    {
      "epoch": 0.41932151898734177,
      "grad_norm": 0.22352826595306396,
      "learning_rate": 8.523171049817974e-06,
      "loss": 0.0002,
      "step": 1294
    },
    {
      "epoch": 0.41996962025316453,
      "grad_norm": 0.11811576783657074,
      "learning_rate": 8.518664814351502e-06,
      "loss": 0.0001,
      "step": 1296
    },
    {
      "epoch": 0.42061772151898735,
      "grad_norm": 0.07777664810419083,
      "learning_rate": 8.514152909614538e-06,
      "loss": 0.0001,
      "step": 1298
    },
    {
      "epoch": 0.4212658227848101,
      "grad_norm": 0.008249818347394466,
      "learning_rate": 8.509635342876655e-06,
      "loss": 0.0,
      "step": 1300
    },
    {
      "epoch": 0.42191392405063294,
      "grad_norm": 0.14114025235176086,
      "learning_rate": 8.505112121416554e-06,
      "loss": 0.0002,
      "step": 1302
    },
    {
      "epoch": 0.4225620253164557,
      "grad_norm": 0.17114797234535217,
      "learning_rate": 8.500583252522053e-06,
      "loss": 0.0003,
      "step": 1304
    },
    {
      "epoch": 0.42321012658227847,
      "grad_norm": 0.03196744620800018,
      "learning_rate": 8.496048743490053e-06,
      "loss": 0.0001,
      "step": 1306
    },
    {
      "epoch": 0.4238582278481013,
      "grad_norm": 0.022405337542295456,
      "learning_rate": 8.49150860162656e-06,
      "loss": 0.0001,
      "step": 1308
    },
    {
      "epoch": 0.42450632911392405,
      "grad_norm": 0.009215210564434528,
      "learning_rate": 8.486962834246646e-06,
      "loss": 0.0001,
      "step": 1310
    },
    {
      "epoch": 0.4251544303797468,
      "grad_norm": 0.08726584166288376,
      "learning_rate": 8.482411448674445e-06,
      "loss": 0.0001,
      "step": 1312
    },
    {
      "epoch": 0.42580253164556964,
      "grad_norm": 0.006170928943902254,
      "learning_rate": 8.477854452243149e-06,
      "loss": 0.0,
      "step": 1314
    },
    {
      "epoch": 0.4264506329113924,
      "grad_norm": 0.27486130595207214,
      "learning_rate": 8.473291852294986e-06,
      "loss": 0.0005,
      "step": 1316
    },
    {
      "epoch": 0.42709873417721517,
      "grad_norm": 0.5437067151069641,
      "learning_rate": 8.468723656181219e-06,
      "loss": 0.001,
      "step": 1318
    },
    {
      "epoch": 0.427746835443038,
      "grad_norm": 0.009669135324656963,
      "learning_rate": 8.464149871262118e-06,
      "loss": 0.0,
      "step": 1320
    },
    {
      "epoch": 0.427746835443038,
      "eval_accuracy": 0.9999560106092086,
      "eval_loss": 0.00023698546283412725,
      "eval_runtime": 105.6712,
      "eval_samples_per_second": 47.317,
      "eval_steps_per_second": 11.829,
      "step": 1320
    },
    {
      "epoch": 0.42839493670886075,
      "grad_norm": 0.02640676125884056,
      "learning_rate": 8.459570504906962e-06,
      "loss": 0.0006,
      "step": 1322
    },
    {
      "epoch": 0.4290430379746835,
      "grad_norm": 0.20899447798728943,
      "learning_rate": 8.454985564494025e-06,
      "loss": 0.0001,
      "step": 1324
    },
    {
      "epoch": 0.42969113924050634,
      "grad_norm": 0.5605078339576721,
      "learning_rate": 8.450395057410561e-06,
      "loss": 0.0009,
      "step": 1326
    },
    {
      "epoch": 0.4303392405063291,
      "grad_norm": 0.18222258985042572,
      "learning_rate": 8.445798991052791e-06,
      "loss": 0.0004,
      "step": 1328
    },
    {
      "epoch": 0.4309873417721519,
      "grad_norm": 0.16136883199214935,
      "learning_rate": 8.441197372825892e-06,
      "loss": 0.0001,
      "step": 1330
    },
    {
      "epoch": 0.4316354430379747,
      "grad_norm": 0.1953054517507553,
      "learning_rate": 8.436590210143991e-06,
      "loss": 0.0006,
      "step": 1332
    },
    {
      "epoch": 0.43228354430379745,
      "grad_norm": 0.2527662515640259,
      "learning_rate": 8.431977510430145e-06,
      "loss": 0.0004,
      "step": 1334
    },
    {
      "epoch": 0.43293164556962027,
      "grad_norm": 0.06676214188337326,
      "learning_rate": 8.427359281116335e-06,
      "loss": 0.0002,
      "step": 1336
    },
    {
      "epoch": 0.43357974683544304,
      "grad_norm": 0.2816427946090698,
      "learning_rate": 8.422735529643445e-06,
      "loss": 0.0003,
      "step": 1338
    },
    {
      "epoch": 0.4342278481012658,
      "grad_norm": 0.019613001495599747,
      "learning_rate": 8.418106263461261e-06,
      "loss": 0.0003,
      "step": 1340
    },
    {
      "epoch": 0.4348759493670886,
      "grad_norm": 0.02441325969994068,
      "learning_rate": 8.413471490028456e-06,
      "loss": 0.0002,
      "step": 1342
    },
    {
      "epoch": 0.4355240506329114,
      "grad_norm": 0.0942113846540451,
      "learning_rate": 8.408831216812574e-06,
      "loss": 0.0001,
      "step": 1344
    },
    {
      "epoch": 0.43617215189873415,
      "grad_norm": 0.2570306360721588,
      "learning_rate": 8.404185451290017e-06,
      "loss": 0.0005,
      "step": 1346
    },
    {
      "epoch": 0.43682025316455697,
      "grad_norm": 0.01734016090631485,
      "learning_rate": 8.399534200946044e-06,
      "loss": 0.0006,
      "step": 1348
    },
    {
      "epoch": 0.43746835443037974,
      "grad_norm": 0.06334984302520752,
      "learning_rate": 8.394877473274743e-06,
      "loss": 0.0002,
      "step": 1350
    },
    {
      "epoch": 0.43811645569620256,
      "grad_norm": 0.09384873509407043,
      "learning_rate": 8.39021527577903e-06,
      "loss": 0.0002,
      "step": 1352
    },
    {
      "epoch": 0.4387645569620253,
      "grad_norm": 0.02200414426624775,
      "learning_rate": 8.38554761597064e-06,
      "loss": 0.0003,
      "step": 1354
    },
    {
      "epoch": 0.4394126582278481,
      "grad_norm": 0.007787387352436781,
      "learning_rate": 8.380874501370098e-06,
      "loss": 0.0,
      "step": 1356
    },
    {
      "epoch": 0.4400607594936709,
      "grad_norm": 0.02325863018631935,
      "learning_rate": 8.376195939506727e-06,
      "loss": 0.0001,
      "step": 1358
    },
    {
      "epoch": 0.44070886075949367,
      "grad_norm": 0.02358267642557621,
      "learning_rate": 8.371511937918616e-06,
      "loss": 0.0001,
      "step": 1360
    },
    {
      "epoch": 0.44070886075949367,
      "eval_accuracy": 0.9999532019066854,
      "eval_loss": 0.00014842751261312515,
      "eval_runtime": 100.4462,
      "eval_samples_per_second": 49.778,
      "eval_steps_per_second": 12.444,
      "step": 1360
    },
    {
      "epoch": 0.44135696202531643,
      "grad_norm": 0.009702431038022041,
      "learning_rate": 8.366822504152636e-06,
      "loss": 0.0,
      "step": 1362
    },
    {
      "epoch": 0.44200506329113926,
      "grad_norm": 0.003049626015126705,
      "learning_rate": 8.362127645764392e-06,
      "loss": 0.0005,
      "step": 1364
    },
    {
      "epoch": 0.442653164556962,
      "grad_norm": 0.3415703773498535,
      "learning_rate": 8.357427370318239e-06,
      "loss": 0.0004,
      "step": 1366
    },
    {
      "epoch": 0.4433012658227848,
      "grad_norm": 0.04481835663318634,
      "learning_rate": 8.352721685387258e-06,
      "loss": 0.0003,
      "step": 1368
    },
    {
      "epoch": 0.4439493670886076,
      "grad_norm": 0.22728939354419708,
      "learning_rate": 8.348010598553245e-06,
      "loss": 0.0002,
      "step": 1370
    },
    {
      "epoch": 0.44459746835443037,
      "grad_norm": 0.08108057081699371,
      "learning_rate": 8.3432941174067e-06,
      "loss": 0.001,
      "step": 1372
    },
    {
      "epoch": 0.4452455696202532,
      "grad_norm": 1.374200701713562,
      "learning_rate": 8.338572249546813e-06,
      "loss": 0.0022,
      "step": 1374
    },
    {
      "epoch": 0.44589367088607595,
      "grad_norm": 0.008610721677541733,
      "learning_rate": 8.33384500258146e-06,
      "loss": 0.0004,
      "step": 1376
    },
    {
      "epoch": 0.4465417721518987,
      "grad_norm": 0.0495627261698246,
      "learning_rate": 8.329112384127172e-06,
      "loss": 0.0006,
      "step": 1378
    },
    {
      "epoch": 0.44718987341772154,
      "grad_norm": 0.02110774628818035,
      "learning_rate": 8.324374401809144e-06,
      "loss": 0.0002,
      "step": 1380
    },
    {
      "epoch": 0.4478379746835443,
      "grad_norm": 0.23234134912490845,
      "learning_rate": 8.319631063261209e-06,
      "loss": 0.0004,
      "step": 1382
    },
    {
      "epoch": 0.44848607594936707,
      "grad_norm": 0.12905079126358032,
      "learning_rate": 8.314882376125832e-06,
      "loss": 0.0007,
      "step": 1384
    },
    {
      "epoch": 0.4491341772151899,
      "grad_norm": 0.11328396201133728,
      "learning_rate": 8.310128348054093e-06,
      "loss": 0.0005,
      "step": 1386
    },
    {
      "epoch": 0.44978227848101265,
      "grad_norm": 0.016495974734425545,
      "learning_rate": 8.305368986705683e-06,
      "loss": 0.0001,
      "step": 1388
    },
    {
      "epoch": 0.4504303797468354,
      "grad_norm": 0.06582915037870407,
      "learning_rate": 8.300604299748876e-06,
      "loss": 0.0001,
      "step": 1390
    },
    {
      "epoch": 0.45107848101265824,
      "grad_norm": 0.0918450728058815,
      "learning_rate": 8.295834294860535e-06,
      "loss": 0.0001,
      "step": 1392
    },
    {
      "epoch": 0.451726582278481,
      "grad_norm": 0.36612367630004883,
      "learning_rate": 8.291058979726092e-06,
      "loss": 0.0003,
      "step": 1394
    },
    {
      "epoch": 0.4523746835443038,
      "grad_norm": 0.014073346741497517,
      "learning_rate": 8.286278362039527e-06,
      "loss": 0.0001,
      "step": 1396
    },
    {
      "epoch": 0.4530227848101266,
      "grad_norm": 0.14093072712421417,
      "learning_rate": 8.281492449503372e-06,
      "loss": 0.0003,
      "step": 1398
    },
    {
      "epoch": 0.45367088607594935,
      "grad_norm": 0.2960674464702606,
      "learning_rate": 8.276701249828684e-06,
      "loss": 0.0002,
      "step": 1400
    },
    {
      "epoch": 0.45367088607594935,
      "eval_accuracy": 0.9999100233106311,
      "eval_loss": 0.0003401956637389958,
      "eval_runtime": 102.0123,
      "eval_samples_per_second": 49.014,
      "eval_steps_per_second": 12.253,
      "step": 1400
    },
    {
      "epoch": 0.4543189873417722,
      "grad_norm": 0.005298405420035124,
      "learning_rate": 8.271904770735042e-06,
      "loss": 0.0004,
      "step": 1402
    },
    {
      "epoch": 0.45496708860759494,
      "grad_norm": 0.08128544688224792,
      "learning_rate": 8.267103019950529e-06,
      "loss": 0.0002,
      "step": 1404
    },
    {
      "epoch": 0.4556151898734177,
      "grad_norm": 0.03566949442028999,
      "learning_rate": 8.262296005211722e-06,
      "loss": 0.0001,
      "step": 1406
    },
    {
      "epoch": 0.4562632911392405,
      "grad_norm": 0.049706194549798965,
      "learning_rate": 8.257483734263682e-06,
      "loss": 0.0002,
      "step": 1408
    },
    {
      "epoch": 0.4569113924050633,
      "grad_norm": 0.43301084637641907,
      "learning_rate": 8.252666214859936e-06,
      "loss": 0.0019,
      "step": 1410
    },
    {
      "epoch": 0.45755949367088605,
      "grad_norm": 0.01807190477848053,
      "learning_rate": 8.247843454762467e-06,
      "loss": 0.0003,
      "step": 1412
    },
    {
      "epoch": 0.4582075949367089,
      "grad_norm": 0.040173470973968506,
      "learning_rate": 8.243015461741707e-06,
      "loss": 0.0001,
      "step": 1414
    },
    {
      "epoch": 0.45885569620253164,
      "grad_norm": 0.3392210304737091,
      "learning_rate": 8.238182243576512e-06,
      "loss": 0.0002,
      "step": 1416
    },
    {
      "epoch": 0.45950379746835446,
      "grad_norm": 0.1378878653049469,
      "learning_rate": 8.233343808054159e-06,
      "loss": 0.0006,
      "step": 1418
    },
    {
      "epoch": 0.4601518987341772,
      "grad_norm": 0.29194381833076477,
      "learning_rate": 8.228500162970333e-06,
      "loss": 0.0003,
      "step": 1420
    },
    {
      "epoch": 0.4608,
      "grad_norm": 0.03227922320365906,
      "learning_rate": 8.223651316129115e-06,
      "loss": 0.0001,
      "step": 1422
    },
    {
      "epoch": 0.4614481012658228,
      "grad_norm": 0.2710287868976593,
      "learning_rate": 8.21879727534296e-06,
      "loss": 0.0007,
      "step": 1424
    },
    {
      "epoch": 0.4620962025316456,
      "grad_norm": 0.01669517531991005,
      "learning_rate": 8.213938048432697e-06,
      "loss": 0.0,
      "step": 1426
    },
    {
      "epoch": 0.46274430379746834,
      "grad_norm": 0.6960672736167908,
      "learning_rate": 8.20907364322751e-06,
      "loss": 0.0002,
      "step": 1428
    },
    {
      "epoch": 0.46339240506329116,
      "grad_norm": 0.30904194712638855,
      "learning_rate": 8.204204067564924e-06,
      "loss": 0.0004,
      "step": 1430
    },
    {
      "epoch": 0.4640405063291139,
      "grad_norm": 0.2030443251132965,
      "learning_rate": 8.199329329290798e-06,
      "loss": 0.0003,
      "step": 1432
    },
    {
      "epoch": 0.4646886075949367,
      "grad_norm": 0.06898464262485504,
      "learning_rate": 8.194449436259305e-06,
      "loss": 0.0001,
      "step": 1434
    },
    {
      "epoch": 0.4653367088607595,
      "grad_norm": 0.39258521795272827,
      "learning_rate": 8.189564396332927e-06,
      "loss": 0.0008,
      "step": 1436
    },
    {
      "epoch": 0.46598481012658227,
      "grad_norm": 0.12645263969898224,
      "learning_rate": 8.184674217382438e-06,
      "loss": 0.0008,
      "step": 1438
    },
    {
      "epoch": 0.46663291139240504,
      "grad_norm": 0.1911512017250061,
      "learning_rate": 8.179778907286889e-06,
      "loss": 0.0008,
      "step": 1440
    },
    {
      "epoch": 0.46663291139240504,
      "eval_accuracy": 0.99995684467669,
      "eval_loss": 0.00017216747801285237,
      "eval_runtime": 101.363,
      "eval_samples_per_second": 49.328,
      "eval_steps_per_second": 12.332,
      "step": 1440
    },
    {
      "epoch": 0.46728101265822786,
      "grad_norm": 0.1570432484149933,
      "learning_rate": 8.174878473933601e-06,
      "loss": 0.0002,
      "step": 1442
    },
    {
      "epoch": 0.4679291139240506,
      "grad_norm": 0.1353127658367157,
      "learning_rate": 8.16997292521815e-06,
      "loss": 0.0005,
      "step": 1444
    },
    {
      "epoch": 0.46857721518987344,
      "grad_norm": 0.09459888190031052,
      "learning_rate": 8.165062269044353e-06,
      "loss": 0.0002,
      "step": 1446
    },
    {
      "epoch": 0.4692253164556962,
      "grad_norm": 0.03600757569074631,
      "learning_rate": 8.160146513324256e-06,
      "loss": 0.0001,
      "step": 1448
    },
    {
      "epoch": 0.46987341772151897,
      "grad_norm": 0.2478301078081131,
      "learning_rate": 8.15522566597812e-06,
      "loss": 0.0005,
      "step": 1450
    },
    {
      "epoch": 0.4705215189873418,
      "grad_norm": 0.16877983510494232,
      "learning_rate": 8.150299734934413e-06,
      "loss": 0.0004,
      "step": 1452
    },
    {
      "epoch": 0.47116962025316456,
      "grad_norm": 0.0916639119386673,
      "learning_rate": 8.14536872812979e-06,
      "loss": 0.0007,
      "step": 1454
    },
    {
      "epoch": 0.4718177215189873,
      "grad_norm": 0.04261047765612602,
      "learning_rate": 8.140432653509089e-06,
      "loss": 0.0001,
      "step": 1456
    },
    {
      "epoch": 0.47246582278481014,
      "grad_norm": 0.14687563478946686,
      "learning_rate": 8.135491519025307e-06,
      "loss": 0.0002,
      "step": 1458
    },
    {
      "epoch": 0.4731139240506329,
      "grad_norm": 0.15630154311656952,
      "learning_rate": 8.130545332639599e-06,
      "loss": 0.0001,
      "step": 1460
    },
    {
      "epoch": 0.47376202531645567,
      "grad_norm": 0.017821237444877625,
      "learning_rate": 8.125594102321256e-06,
      "loss": 0.0001,
      "step": 1462
    },
    {
      "epoch": 0.4744101265822785,
      "grad_norm": 0.025156579911708832,
      "learning_rate": 8.120637836047698e-06,
      "loss": 0.0,
      "step": 1464
    },
    {
      "epoch": 0.47505822784810126,
      "grad_norm": 0.24539688229560852,
      "learning_rate": 8.115676541804456e-06,
      "loss": 0.0003,
      "step": 1466
    },
    {
      "epoch": 0.4757063291139241,
      "grad_norm": 0.007439673412591219,
      "learning_rate": 8.110710227585169e-06,
      "loss": 0.0,
      "step": 1468
    },
    {
      "epoch": 0.47635443037974684,
      "grad_norm": 0.224391371011734,
      "learning_rate": 8.105738901391553e-06,
      "loss": 0.0001,
      "step": 1470
    },
    {
      "epoch": 0.4770025316455696,
      "grad_norm": 0.12915731966495514,
      "learning_rate": 8.100762571233409e-06,
      "loss": 0.0003,
      "step": 1472
    },
    {
      "epoch": 0.4776506329113924,
      "grad_norm": 0.02007501944899559,
      "learning_rate": 8.095781245128598e-06,
      "loss": 0.0,
      "step": 1474
    },
    {
      "epoch": 0.4782987341772152,
      "grad_norm": 0.009752247482538223,
      "learning_rate": 8.090794931103026e-06,
      "loss": 0.0,
      "step": 1476
    },
    {
      "epoch": 0.47894683544303795,
      "grad_norm": 0.10017750412225723,
      "learning_rate": 8.085803637190643e-06,
      "loss": 0.0003,
      "step": 1478
    },
    {
      "epoch": 0.4795949367088608,
      "grad_norm": 0.005627952981740236,
      "learning_rate": 8.080807371433415e-06,
      "loss": 0.0,
      "step": 1480
    },
    {
      "epoch": 0.4795949367088608,
      "eval_accuracy": 0.9999832096964334,
      "eval_loss": 7.632683264091611e-05,
      "eval_runtime": 100.8851,
      "eval_samples_per_second": 49.561,
      "eval_steps_per_second": 12.39,
      "step": 1480
    },
    {
      "epoch": 0.48024303797468354,
      "grad_norm": 0.14481300115585327,
      "learning_rate": 8.075806141881327e-06,
      "loss": 0.0001,
      "step": 1482
    },
    {
      "epoch": 0.4808911392405063,
      "grad_norm": 0.07196534425020218,
      "learning_rate": 8.07079995659235e-06,
      "loss": 0.0001,
      "step": 1484
    },
    {
      "epoch": 0.4815392405063291,
      "grad_norm": 0.1497977375984192,
      "learning_rate": 8.065788823632451e-06,
      "loss": 0.0005,
      "step": 1486
    },
    {
      "epoch": 0.4821873417721519,
      "grad_norm": 0.01656440645456314,
      "learning_rate": 8.060772751075564e-06,
      "loss": 0.0001,
      "step": 1488
    },
    {
      "epoch": 0.4828354430379747,
      "grad_norm": 0.001540559809654951,
      "learning_rate": 8.05575174700358e-06,
      "loss": 0.0,
      "step": 1490
    },
    {
      "epoch": 0.4834835443037975,
      "grad_norm": 0.0042270757257938385,
      "learning_rate": 8.05072581950634e-06,
      "loss": 0.0001,
      "step": 1492
    },
    {
      "epoch": 0.48413164556962024,
      "grad_norm": 0.10281050950288773,
      "learning_rate": 8.045694976681613e-06,
      "loss": 0.0001,
      "step": 1494
    },
    {
      "epoch": 0.48477974683544306,
      "grad_norm": 0.07573399692773819,
      "learning_rate": 8.04065922663509e-06,
      "loss": 0.0,
      "step": 1496
    },
    {
      "epoch": 0.4854278481012658,
      "grad_norm": 0.0019800164736807346,
      "learning_rate": 8.035618577480369e-06,
      "loss": 0.0,
      "step": 1498
    },
    {
      "epoch": 0.4860759493670886,
      "grad_norm": 0.4453049302101135,
      "learning_rate": 8.030573037338942e-06,
      "loss": 0.0003,
      "step": 1500
    },
    {
      "epoch": 0.4867240506329114,
      "grad_norm": 0.0011820768704637885,
      "learning_rate": 8.025522614340177e-06,
      "loss": 0.0,
      "step": 1502
    },
    {
      "epoch": 0.4873721518987342,
      "grad_norm": 0.0030645334627479315,
      "learning_rate": 8.020467316621316e-06,
      "loss": 0.0,
      "step": 1504
    },
    {
      "epoch": 0.48802025316455694,
      "grad_norm": 0.007107930723577738,
      "learning_rate": 8.015407152327448e-06,
      "loss": 0.0,
      "step": 1506
    },
    {
      "epoch": 0.48866835443037976,
      "grad_norm": 0.0038615772500634193,
      "learning_rate": 8.010342129611508e-06,
      "loss": 0.0,
      "step": 1508
    },
    {
      "epoch": 0.4893164556962025,
      "grad_norm": 0.04887912794947624,
      "learning_rate": 8.005272256634257e-06,
      "loss": 0.0001,
      "step": 1510
    },
    {
      "epoch": 0.48996455696202534,
      "grad_norm": 0.0012970338575541973,
      "learning_rate": 8.000197541564273e-06,
      "loss": 0.0,
      "step": 1512
    },
    {
      "epoch": 0.4906126582278481,
      "grad_norm": 0.0027239499613642693,
      "learning_rate": 7.99511799257793e-06,
      "loss": 0.0002,
      "step": 1514
    },
    {
      "epoch": 0.4912607594936709,
      "grad_norm": 0.08472245186567307,
      "learning_rate": 7.990033617859396e-06,
      "loss": 0.0,
      "step": 1516
    },
    {
      "epoch": 0.4919088607594937,
      "grad_norm": 0.0267314575612545,
      "learning_rate": 7.984944425600614e-06,
      "loss": 0.0001,
      "step": 1518
    },
    {
      "epoch": 0.49255696202531646,
      "grad_norm": 0.031123576685786247,
      "learning_rate": 7.979850424001283e-06,
      "loss": 0.0001,
      "step": 1520
    },
    {
      "epoch": 0.49255696202531646,
      "eval_accuracy": 0.9999826136620499,
      "eval_loss": 6.179293995955959e-05,
      "eval_runtime": 101.4096,
      "eval_samples_per_second": 49.305,
      "eval_steps_per_second": 12.326,
      "step": 1520
    },
    {
      "epoch": 0.4932050632911392,
      "grad_norm": 0.011671812273561954,
      "learning_rate": 7.97475162126886e-06,
      "loss": 0.0,
      "step": 1522
    },
    {
      "epoch": 0.49385316455696204,
      "grad_norm": 0.007582854479551315,
      "learning_rate": 7.96964802561853e-06,
      "loss": 0.0005,
      "step": 1524
    },
    {
      "epoch": 0.4945012658227848,
      "grad_norm": 0.06047774478793144,
      "learning_rate": 7.964539645273204e-06,
      "loss": 0.0001,
      "step": 1526
    },
    {
      "epoch": 0.4951493670886076,
      "grad_norm": 0.00793007854372263,
      "learning_rate": 7.9594264884635e-06,
      "loss": 0.0001,
      "step": 1528
    },
    {
      "epoch": 0.4957974683544304,
      "grad_norm": 0.008908102288842201,
      "learning_rate": 7.954308563427732e-06,
      "loss": 0.0,
      "step": 1530
    },
    {
      "epoch": 0.49644556962025316,
      "grad_norm": 0.04215071350336075,
      "learning_rate": 7.9491858784119e-06,
      "loss": 0.0005,
      "step": 1532
    },
    {
      "epoch": 0.4970936708860759,
      "grad_norm": 0.0012058537686243653,
      "learning_rate": 7.944058441669671e-06,
      "loss": 0.0002,
      "step": 1534
    },
    {
      "epoch": 0.49774177215189874,
      "grad_norm": 0.015041892416775227,
      "learning_rate": 7.938926261462366e-06,
      "loss": 0.0,
      "step": 1536
    },
    {
      "epoch": 0.4983898734177215,
      "grad_norm": 0.0026610875502228737,
      "learning_rate": 7.933789346058951e-06,
      "loss": 0.0,
      "step": 1538
    },
    {
      "epoch": 0.4990379746835443,
      "grad_norm": 0.002617933787405491,
      "learning_rate": 7.928647703736024e-06,
      "loss": 0.0,
      "step": 1540
    },
    {
      "epoch": 0.4996860759493671,
      "grad_norm": 0.044066380709409714,
      "learning_rate": 7.923501342777788e-06,
      "loss": 0.0,
      "step": 1542
    },
    {
      "epoch": 0.5003341772151899,
      "grad_norm": 0.0029755565337836742,
      "learning_rate": 7.918350271476064e-06,
      "loss": 0.0,
      "step": 1544
    },
    {
      "epoch": 0.5009822784810126,
      "grad_norm": 0.02002215012907982,
      "learning_rate": 7.913194498130252e-06,
      "loss": 0.0,
      "step": 1546
    },
    {
      "epoch": 0.5016303797468354,
      "grad_norm": 0.14786788821220398,
      "learning_rate": 7.90803403104733e-06,
      "loss": 0.0004,
      "step": 1548
    },
    {
      "epoch": 0.5022784810126583,
      "grad_norm": 0.003735872684046626,
      "learning_rate": 7.90286887854184e-06,
      "loss": 0.0,
      "step": 1550
    },
    {
      "epoch": 0.502926582278481,
      "grad_norm": 0.003643750213086605,
      "learning_rate": 7.897699048935875e-06,
      "loss": 0.0,
      "step": 1552
    },
    {
      "epoch": 0.5035746835443038,
      "grad_norm": 0.0010872690472751856,
      "learning_rate": 7.892524550559056e-06,
      "loss": 0.0,
      "step": 1554
    },
    {
      "epoch": 0.5042227848101266,
      "grad_norm": 0.004607509821653366,
      "learning_rate": 7.887345391748533e-06,
      "loss": 0.0003,
      "step": 1556
    },
    {
      "epoch": 0.5048708860759493,
      "grad_norm": 0.018745901063084602,
      "learning_rate": 7.882161580848966e-06,
      "loss": 0.0001,
      "step": 1558
    },
    {
      "epoch": 0.5055189873417721,
      "grad_norm": 0.0027148567605763674,
      "learning_rate": 7.876973126212507e-06,
      "loss": 0.0,
      "step": 1560
    },
    {
      "epoch": 0.5055189873417721,
      "eval_accuracy": 0.9999803987308754,
      "eval_loss": 7.348901272052899e-05,
      "eval_runtime": 102.1608,
      "eval_samples_per_second": 48.942,
      "eval_steps_per_second": 12.236,
      "step": 1560
    },
    {
      "epoch": 0.506167088607595,
      "grad_norm": 0.06728499382734299,
      "learning_rate": 7.87178003619879e-06,
      "loss": 0.0001,
      "step": 1562
    },
    {
      "epoch": 0.5068151898734177,
      "grad_norm": 0.03146948292851448,
      "learning_rate": 7.866582319174918e-06,
      "loss": 0.0,
      "step": 1564
    },
    {
      "epoch": 0.5074632911392405,
      "grad_norm": 0.07248188555240631,
      "learning_rate": 7.861379983515449e-06,
      "loss": 0.0001,
      "step": 1566
    },
    {
      "epoch": 0.5081113924050633,
      "grad_norm": 0.0029003811068832874,
      "learning_rate": 7.856173037602383e-06,
      "loss": 0.0001,
      "step": 1568
    },
    {
      "epoch": 0.508759493670886,
      "grad_norm": 0.052399907261133194,
      "learning_rate": 7.85096148982515e-06,
      "loss": 0.0003,
      "step": 1570
    },
    {
      "epoch": 0.5094075949367088,
      "grad_norm": 0.0011442819377407432,
      "learning_rate": 7.845745348580592e-06,
      "loss": 0.0,
      "step": 1572
    },
    {
      "epoch": 0.5100556962025317,
      "grad_norm": 0.0014424960827454925,
      "learning_rate": 7.840524622272949e-06,
      "loss": 0.0001,
      "step": 1574
    },
    {
      "epoch": 0.5107037974683545,
      "grad_norm": 0.05243639647960663,
      "learning_rate": 7.835299319313854e-06,
      "loss": 0.0001,
      "step": 1576
    },
    {
      "epoch": 0.5113518987341772,
      "grad_norm": 0.021167520433664322,
      "learning_rate": 7.830069448122313e-06,
      "loss": 0.0,
      "step": 1578
    },
    {
      "epoch": 0.512,
      "grad_norm": 0.006568542681634426,
      "learning_rate": 7.82483501712469e-06,
      "loss": 0.0001,
      "step": 1580
    },
    {
      "epoch": 0.5126481012658228,
      "grad_norm": 0.0223622377961874,
      "learning_rate": 7.819596034754696e-06,
      "loss": 0.0,
      "step": 1582
    },
    {
      "epoch": 0.5132962025316455,
      "grad_norm": 0.02252521552145481,
      "learning_rate": 7.81435250945338e-06,
      "loss": 0.0001,
      "step": 1584
    },
    {
      "epoch": 0.5139443037974684,
      "grad_norm": 0.0036857861559838057,
      "learning_rate": 7.8091044496691e-06,
      "loss": 0.0,
      "step": 1586
    },
    {
      "epoch": 0.5145924050632912,
      "grad_norm": 0.0021720710210502148,
      "learning_rate": 7.803851863857533e-06,
      "loss": 0.0001,
      "step": 1588
    },
    {
      "epoch": 0.5152405063291139,
      "grad_norm": 0.004408895969390869,
      "learning_rate": 7.798594760481639e-06,
      "loss": 0.0,
      "step": 1590
    },
    {
      "epoch": 0.5158886075949367,
      "grad_norm": 0.004357462283223867,
      "learning_rate": 7.793333148011658e-06,
      "loss": 0.0,
      "step": 1592
    },
    {
      "epoch": 0.5165367088607595,
      "grad_norm": 0.02102765254676342,
      "learning_rate": 7.7880670349251e-06,
      "loss": 0.0001,
      "step": 1594
    },
    {
      "epoch": 0.5171848101265822,
      "grad_norm": 0.3509201109409332,
      "learning_rate": 7.782796429706721e-06,
      "loss": 0.0003,
      "step": 1596
    },
    {
      "epoch": 0.5178329113924051,
      "grad_norm": 0.0007810965762473643,
      "learning_rate": 7.777521340848515e-06,
      "loss": 0.0,
      "step": 1598
    },
    {
      "epoch": 0.5184810126582279,
      "grad_norm": 0.11072275787591934,
      "learning_rate": 7.772241776849705e-06,
      "loss": 0.0001,
      "step": 1600
    },
    {
      "epoch": 0.5184810126582279,
      "eval_accuracy": 0.9999901331856902,
      "eval_loss": 3.327821468701586e-05,
      "eval_runtime": 102.4497,
      "eval_samples_per_second": 48.804,
      "eval_steps_per_second": 12.201,
      "step": 1600
    },
    {
      "epoch": 0.5191291139240506,
      "grad_norm": 0.4169808626174927,
      "learning_rate": 7.76695774621672e-06,
      "loss": 0.0002,
      "step": 1602
    },
    {
      "epoch": 0.5197772151898734,
      "grad_norm": 0.01579488068819046,
      "learning_rate": 7.761669257463188e-06,
      "loss": 0.0,
      "step": 1604
    },
    {
      "epoch": 0.5204253164556962,
      "grad_norm": 0.05845312774181366,
      "learning_rate": 7.756376319109917e-06,
      "loss": 0.0001,
      "step": 1606
    },
    {
      "epoch": 0.5210734177215189,
      "grad_norm": 0.011482629925012589,
      "learning_rate": 7.751078939684886e-06,
      "loss": 0.0004,
      "step": 1608
    },
    {
      "epoch": 0.5217215189873418,
      "grad_norm": 0.01010491605848074,
      "learning_rate": 7.74577712772323e-06,
      "loss": 0.0,
      "step": 1610
    },
    {
      "epoch": 0.5223696202531646,
      "grad_norm": 0.19445572793483734,
      "learning_rate": 7.740470891767225e-06,
      "loss": 0.0002,
      "step": 1612
    },
    {
      "epoch": 0.5230177215189873,
      "grad_norm": 0.03457183763384819,
      "learning_rate": 7.735160240366276e-06,
      "loss": 0.0001,
      "step": 1614
    },
    {
      "epoch": 0.5236658227848101,
      "grad_norm": 0.00079920026473701,
      "learning_rate": 7.729845182076896e-06,
      "loss": 0.0,
      "step": 1616
    },
    {
      "epoch": 0.5243139240506329,
      "grad_norm": 0.06374679505825043,
      "learning_rate": 7.72452572546271e-06,
      "loss": 0.0001,
      "step": 1618
    },
    {
      "epoch": 0.5249620253164557,
      "grad_norm": 0.015360563062131405,
      "learning_rate": 7.71920187909442e-06,
      "loss": 0.0,
      "step": 1620
    },
    {
      "epoch": 0.5256101265822785,
      "grad_norm": 0.07180308550596237,
      "learning_rate": 7.713873651549805e-06,
      "loss": 0.0003,
      "step": 1622
    },
    {
      "epoch": 0.5262582278481013,
      "grad_norm": 0.13326337933540344,
      "learning_rate": 7.7085410514137e-06,
      "loss": 0.0001,
      "step": 1624
    },
    {
      "epoch": 0.5269063291139241,
      "grad_norm": 0.005201876629143953,
      "learning_rate": 7.703204087277989e-06,
      "loss": 0.0,
      "step": 1626
    },
    {
      "epoch": 0.5275544303797468,
      "grad_norm": 0.004504207521677017,
      "learning_rate": 7.697862767741584e-06,
      "loss": 0.0001,
      "step": 1628
    },
    {
      "epoch": 0.5282025316455696,
      "grad_norm": 0.2609665095806122,
      "learning_rate": 7.692517101410414e-06,
      "loss": 0.0001,
      "step": 1630
    },
    {
      "epoch": 0.5288506329113924,
      "grad_norm": 0.00579993799328804,
      "learning_rate": 7.68716709689742e-06,
      "loss": 0.0,
      "step": 1632
    },
    {
      "epoch": 0.5294987341772152,
      "grad_norm": 0.325564444065094,
      "learning_rate": 7.681812762822517e-06,
      "loss": 0.0005,
      "step": 1634
    },
    {
      "epoch": 0.530146835443038,
      "grad_norm": 0.025982841849327087,
      "learning_rate": 7.676454107812608e-06,
      "loss": 0.0001,
      "step": 1636
    },
    {
      "epoch": 0.5307949367088608,
      "grad_norm": 0.005460761021822691,
      "learning_rate": 7.671091140501557e-06,
      "loss": 0.0,
      "step": 1638
    },
    {
      "epoch": 0.5314430379746835,
      "grad_norm": 0.008271248079836369,
      "learning_rate": 7.66572386953017e-06,
      "loss": 0.0,
      "step": 1640
    },
    {
      "epoch": 0.5314430379746835,
      "eval_accuracy": 0.9999786783875229,
      "eval_loss": 8.938193059293553e-05,
      "eval_runtime": 101.5581,
      "eval_samples_per_second": 49.233,
      "eval_steps_per_second": 12.308,
      "step": 1640
    },
    {
      "epoch": 0.5320911392405063,
      "grad_norm": 0.015748729929327965,
      "learning_rate": 7.660352303546192e-06,
      "loss": 0.0001,
      "step": 1642
    },
    {
      "epoch": 0.5327392405063291,
      "grad_norm": 0.1777859777212143,
      "learning_rate": 7.654976451204288e-06,
      "loss": 0.0002,
      "step": 1644
    },
    {
      "epoch": 0.5333873417721519,
      "grad_norm": 0.029370561242103577,
      "learning_rate": 7.649596321166024e-06,
      "loss": 0.0002,
      "step": 1646
    },
    {
      "epoch": 0.5340354430379747,
      "grad_norm": 0.018781421706080437,
      "learning_rate": 7.644211922099867e-06,
      "loss": 0.0001,
      "step": 1648
    },
    {
      "epoch": 0.5346835443037975,
      "grad_norm": 0.014029051177203655,
      "learning_rate": 7.638823262681155e-06,
      "loss": 0.0,
      "step": 1650
    },
    {
      "epoch": 0.5353316455696202,
      "grad_norm": 0.08904850482940674,
      "learning_rate": 7.633430351592093e-06,
      "loss": 0.0001,
      "step": 1652
    },
    {
      "epoch": 0.535979746835443,
      "grad_norm": 0.49554842710494995,
      "learning_rate": 7.6280331975217356e-06,
      "loss": 0.0009,
      "step": 1654
    },
    {
      "epoch": 0.5366278481012658,
      "grad_norm": 0.0705832988023758,
      "learning_rate": 7.622631809165972e-06,
      "loss": 0.0002,
      "step": 1656
    },
    {
      "epoch": 0.5372759493670886,
      "grad_norm": 0.07367483526468277,
      "learning_rate": 7.617226195227518e-06,
      "loss": 0.0001,
      "step": 1658
    },
    {
      "epoch": 0.5379240506329114,
      "grad_norm": 0.14305748045444489,
      "learning_rate": 7.611816364415896e-06,
      "loss": 0.0001,
      "step": 1660
    },
    {
      "epoch": 0.5385721518987342,
      "grad_norm": 0.11214498430490494,
      "learning_rate": 7.606402325447421e-06,
      "loss": 0.0001,
      "step": 1662
    },
    {
      "epoch": 0.539220253164557,
      "grad_norm": 0.003934099338948727,
      "learning_rate": 7.600984087045187e-06,
      "loss": 0.0004,
      "step": 1664
    },
    {
      "epoch": 0.5398683544303797,
      "grad_norm": 0.0044866688549518585,
      "learning_rate": 7.595561657939061e-06,
      "loss": 0.0001,
      "step": 1666
    },
    {
      "epoch": 0.5405164556962025,
      "grad_norm": 0.025382693856954575,
      "learning_rate": 7.590135046865652e-06,
      "loss": 0.0002,
      "step": 1668
    },
    {
      "epoch": 0.5411645569620254,
      "grad_norm": 0.08539170771837234,
      "learning_rate": 7.584704262568315e-06,
      "loss": 0.0001,
      "step": 1670
    },
    {
      "epoch": 0.5418126582278481,
      "grad_norm": 0.0013932752190157771,
      "learning_rate": 7.579269313797126e-06,
      "loss": 0.0001,
      "step": 1672
    },
    {
      "epoch": 0.5424607594936709,
      "grad_norm": 0.1252427101135254,
      "learning_rate": 7.573830209308872e-06,
      "loss": 0.0001,
      "step": 1674
    },
    {
      "epoch": 0.5431088607594937,
      "grad_norm": 0.03497989848256111,
      "learning_rate": 7.568386957867033e-06,
      "loss": 0.0001,
      "step": 1676
    },
    {
      "epoch": 0.5437569620253164,
      "grad_norm": 0.03806183114647865,
      "learning_rate": 7.562939568241772e-06,
      "loss": 0.0,
      "step": 1678
    },
    {
      "epoch": 0.5444050632911392,
      "grad_norm": 0.0008746628882363439,
      "learning_rate": 7.557488049209921e-06,
      "loss": 0.0,
      "step": 1680
    },
    {
      "epoch": 0.5444050632911392,
      "eval_accuracy": 0.9999900735137393,
      "eval_loss": 5.2154053264530376e-05,
      "eval_runtime": 101.5637,
      "eval_samples_per_second": 49.23,
      "eval_steps_per_second": 12.308,
      "step": 1680
    },
    {
      "epoch": 0.5450531645569621,
      "grad_norm": 0.18731200695037842,
      "learning_rate": 7.552032409554963e-06,
      "loss": 0.0003,
      "step": 1682
    },
    {
      "epoch": 0.5457012658227848,
      "grad_norm": 0.0005909568862989545,
      "learning_rate": 7.546572658067022e-06,
      "loss": 0.0003,
      "step": 1684
    },
    {
      "epoch": 0.5463493670886076,
      "grad_norm": 0.001809874433092773,
      "learning_rate": 7.541108803542846e-06,
      "loss": 0.0,
      "step": 1686
    },
    {
      "epoch": 0.5469974683544304,
      "grad_norm": 0.002301193308085203,
      "learning_rate": 7.535640854785793e-06,
      "loss": 0.0001,
      "step": 1688
    },
    {
      "epoch": 0.5476455696202531,
      "grad_norm": 0.012999876402318478,
      "learning_rate": 7.530168820605819e-06,
      "loss": 0.0,
      "step": 1690
    },
    {
      "epoch": 0.548293670886076,
      "grad_norm": 0.011647144332528114,
      "learning_rate": 7.5246927098194636e-06,
      "loss": 0.0,
      "step": 1692
    },
    {
      "epoch": 0.5489417721518988,
      "grad_norm": 0.001015910878777504,
      "learning_rate": 7.51921253124983e-06,
      "loss": 0.0002,
      "step": 1694
    },
    {
      "epoch": 0.5495898734177215,
      "grad_norm": 0.1255691796541214,
      "learning_rate": 7.5137282937265796e-06,
      "loss": 0.0001,
      "step": 1696
    },
    {
      "epoch": 0.5502379746835443,
      "grad_norm": 0.06820951402187347,
      "learning_rate": 7.508240006085914e-06,
      "loss": 0.0,
      "step": 1698
    },
    {
      "epoch": 0.5508860759493671,
      "grad_norm": 0.09389737248420715,
      "learning_rate": 7.502747677170556e-06,
      "loss": 0.0007,
      "step": 1700
    },
    {
      "epoch": 0.5515341772151898,
      "grad_norm": 0.1673169583082199,
      "learning_rate": 7.497251315829744e-06,
      "loss": 0.0001,
      "step": 1702
    },
    {
      "epoch": 0.5521822784810126,
      "grad_norm": 0.0026740762405097485,
      "learning_rate": 7.4917509309192125e-06,
      "loss": 0.0,
      "step": 1704
    },
    {
      "epoch": 0.5528303797468355,
      "grad_norm": 0.005095075815916061,
      "learning_rate": 7.486246531301178e-06,
      "loss": 0.0,
      "step": 1706
    },
    {
      "epoch": 0.5534784810126582,
      "grad_norm": 0.06378457695245743,
      "learning_rate": 7.480738125844322e-06,
      "loss": 0.0001,
      "step": 1708
    },
    {
      "epoch": 0.554126582278481,
      "grad_norm": 0.0014856786001473665,
      "learning_rate": 7.475225723423789e-06,
      "loss": 0.0001,
      "step": 1710
    },
    {
      "epoch": 0.5547746835443038,
      "grad_norm": 0.0007130180601961911,
      "learning_rate": 7.469709332921155e-06,
      "loss": 0.0,
      "step": 1712
    },
    {
      "epoch": 0.5554227848101266,
      "grad_norm": 0.04547552391886711,
      "learning_rate": 7.464188963224428e-06,
      "loss": 0.0001,
      "step": 1714
    },
    {
      "epoch": 0.5560708860759493,
      "grad_norm": 0.10711254179477692,
      "learning_rate": 7.45866462322802e-06,
      "loss": 0.0001,
      "step": 1716
    },
    {
      "epoch": 0.5567189873417722,
      "grad_norm": 0.00046518215094693005,
      "learning_rate": 7.453136321832746e-06,
      "loss": 0.0,
      "step": 1718
    },
    {
      "epoch": 0.557367088607595,
      "grad_norm": 0.0007845875225029886,
      "learning_rate": 7.447604067945803e-06,
      "loss": 0.0,
      "step": 1720
    },
    {
      "epoch": 0.557367088607595,
      "eval_accuracy": 0.999991254640404,
      "eval_loss": 4.263612208887935e-05,
      "eval_runtime": 100.7388,
      "eval_samples_per_second": 49.633,
      "eval_steps_per_second": 12.408,
      "step": 1720
    },
    {
      "epoch": 0.5580151898734177,
      "grad_norm": 0.20458361506462097,
      "learning_rate": 7.442067870480752e-06,
      "loss": 0.0004,
      "step": 1722
    },
    {
      "epoch": 0.5586632911392405,
      "grad_norm": 0.00046549775288440287,
      "learning_rate": 7.436527738357514e-06,
      "loss": 0.0003,
      "step": 1724
    },
    {
      "epoch": 0.5593113924050633,
      "grad_norm": 0.001776853110641241,
      "learning_rate": 7.430983680502344e-06,
      "loss": 0.0,
      "step": 1726
    },
    {
      "epoch": 0.559959493670886,
      "grad_norm": 0.00032903251121751964,
      "learning_rate": 7.425435705847825e-06,
      "loss": 0.0,
      "step": 1728
    },
    {
      "epoch": 0.5606075949367089,
      "grad_norm": 0.012358637526631355,
      "learning_rate": 7.419883823332851e-06,
      "loss": 0.0,
      "step": 1730
    },
    {
      "epoch": 0.5612556962025317,
      "grad_norm": 0.12481050938367844,
      "learning_rate": 7.414328041902611e-06,
      "loss": 0.0003,
      "step": 1732
    },
    {
      "epoch": 0.5619037974683544,
      "grad_norm": 0.04864579811692238,
      "learning_rate": 7.408768370508577e-06,
      "loss": 0.0001,
      "step": 1734
    },
    {
      "epoch": 0.5625518987341772,
      "grad_norm": 0.002378104953095317,
      "learning_rate": 7.403204818108487e-06,
      "loss": 0.0,
      "step": 1736
    },
    {
      "epoch": 0.5632,
      "grad_norm": 0.17984673380851746,
      "learning_rate": 7.397637393666333e-06,
      "loss": 0.0001,
      "step": 1738
    },
    {
      "epoch": 0.5638481012658227,
      "grad_norm": 0.3844810128211975,
      "learning_rate": 7.392066106152347e-06,
      "loss": 0.0004,
      "step": 1740
    },
    {
      "epoch": 0.5644962025316456,
      "grad_norm": 0.010118959471583366,
      "learning_rate": 7.386490964542983e-06,
      "loss": 0.0,
      "step": 1742
    },
    {
      "epoch": 0.5651443037974684,
      "grad_norm": 0.003273730631917715,
      "learning_rate": 7.380911977820907e-06,
      "loss": 0.0,
      "step": 1744
    },
    {
      "epoch": 0.5657924050632911,
      "grad_norm": 0.0061654625460505486,
      "learning_rate": 7.3753291549749764e-06,
      "loss": 0.0,
      "step": 1746
    },
    {
      "epoch": 0.5664405063291139,
      "grad_norm": 0.009560778737068176,
      "learning_rate": 7.369742505000232e-06,
      "loss": 0.0,
      "step": 1748
    },
    {
      "epoch": 0.5670886075949367,
      "grad_norm": 0.012677104212343693,
      "learning_rate": 7.364152036897883e-06,
      "loss": 0.0,
      "step": 1750
    },
    {
      "epoch": 0.5677367088607594,
      "grad_norm": 0.0025769511703401804,
      "learning_rate": 7.358557759675284e-06,
      "loss": 0.0,
      "step": 1752
    },
    {
      "epoch": 0.5683848101265823,
      "grad_norm": 0.603306233882904,
      "learning_rate": 7.352959682345936e-06,
      "loss": 0.0021,
      "step": 1754
    },
    {
      "epoch": 0.5690329113924051,
      "grad_norm": 0.18951795995235443,
      "learning_rate": 7.347357813929455e-06,
      "loss": 0.0007,
      "step": 1756
    },
    {
      "epoch": 0.5696810126582279,
      "grad_norm": 0.0019815692212432623,
      "learning_rate": 7.341752163451568e-06,
      "loss": 0.0001,
      "step": 1758
    },
    {
      "epoch": 0.5703291139240506,
      "grad_norm": 0.003150013741105795,
      "learning_rate": 7.3361427399440945e-06,
      "loss": 0.0,
      "step": 1760
    },
    {
      "epoch": 0.5703291139240506,
      "eval_accuracy": 0.9999741535564335,
      "eval_loss": 0.0001262763689737767,
      "eval_runtime": 100.825,
      "eval_samples_per_second": 49.591,
      "eval_steps_per_second": 12.398,
      "step": 1760
    },
    {
      "epoch": 0.5709772151898734,
      "grad_norm": 0.009336372837424278,
      "learning_rate": 7.330529552444934e-06,
      "loss": 0.0,
      "step": 1762
    },
    {
      "epoch": 0.5716253164556963,
      "grad_norm": 0.05104972794651985,
      "learning_rate": 7.324912609998054e-06,
      "loss": 0.0001,
      "step": 1764
    },
    {
      "epoch": 0.572273417721519,
      "grad_norm": 0.13936738669872284,
      "learning_rate": 7.319291921653464e-06,
      "loss": 0.0012,
      "step": 1766
    },
    {
      "epoch": 0.5729215189873418,
      "grad_norm": 0.12685565650463104,
      "learning_rate": 7.313667496467216e-06,
      "loss": 0.0005,
      "step": 1768
    },
    {
      "epoch": 0.5735696202531646,
      "grad_norm": 0.004788834135979414,
      "learning_rate": 7.308039343501381e-06,
      "loss": 0.0,
      "step": 1770
    },
    {
      "epoch": 0.5742177215189873,
      "grad_norm": 0.12622016668319702,
      "learning_rate": 7.302407471824034e-06,
      "loss": 0.0005,
      "step": 1772
    },
    {
      "epoch": 0.5748658227848101,
      "grad_norm": 0.004526766948401928,
      "learning_rate": 7.296771890509242e-06,
      "loss": 0.0,
      "step": 1774
    },
    {
      "epoch": 0.575513924050633,
      "grad_norm": 0.0036531174555420876,
      "learning_rate": 7.291132608637053e-06,
      "loss": 0.0,
      "step": 1776
    },
    {
      "epoch": 0.5761620253164557,
      "grad_norm": 0.043615106493234634,
      "learning_rate": 7.285489635293472e-06,
      "loss": 0.0004,
      "step": 1778
    },
    {
      "epoch": 0.5768101265822785,
      "grad_norm": 0.052893441170454025,
      "learning_rate": 7.279842979570454e-06,
      "loss": 0.0002,
      "step": 1780
    },
    {
      "epoch": 0.5774582278481013,
      "grad_norm": 0.02373715490102768,
      "learning_rate": 7.27419265056589e-06,
      "loss": 0.0002,
      "step": 1782
    },
    {
      "epoch": 0.578106329113924,
      "grad_norm": 0.002014788566157222,
      "learning_rate": 7.268538657383581e-06,
      "loss": 0.0,
      "step": 1784
    },
    {
      "epoch": 0.5787544303797468,
      "grad_norm": 0.06467452645301819,
      "learning_rate": 7.262881009133242e-06,
      "loss": 0.0002,
      "step": 1786
    },
    {
      "epoch": 0.5794025316455697,
      "grad_norm": 0.005883384495973587,
      "learning_rate": 7.2572197149304715e-06,
      "loss": 0.0002,
      "step": 1788
    },
    {
      "epoch": 0.5800506329113924,
      "grad_norm": 0.031703103333711624,
      "learning_rate": 7.251554783896741e-06,
      "loss": 0.0001,
      "step": 1790
    },
    {
      "epoch": 0.5806987341772152,
      "grad_norm": 0.0037112203426659107,
      "learning_rate": 7.245886225159386e-06,
      "loss": 0.0,
      "step": 1792
    },
    {
      "epoch": 0.581346835443038,
      "grad_norm": 0.06743831187486649,
      "learning_rate": 7.240214047851583e-06,
      "loss": 0.0001,
      "step": 1794
    },
    {
      "epoch": 0.5819949367088607,
      "grad_norm": 0.00042398099321871996,
      "learning_rate": 7.234538261112342e-06,
      "loss": 0.0,
      "step": 1796
    },
    {
      "epoch": 0.5826430379746835,
      "grad_norm": 0.0017583374865353107,
      "learning_rate": 7.2288588740864855e-06,
      "loss": 0.0,
      "step": 1798
    },
    {
      "epoch": 0.5832911392405064,
      "grad_norm": 0.06879211962223053,
      "learning_rate": 7.223175895924638e-06,
      "loss": 0.0,
      "step": 1800
    },
    {
      "epoch": 0.5832911392405064,
      "eval_accuracy": 0.9999984173205008,
      "eval_loss": 1.2915682418679353e-05,
      "eval_runtime": 100.8375,
      "eval_samples_per_second": 49.585,
      "eval_steps_per_second": 12.396,
      "step": 1800
    },
    {
      "epoch": 0.5839392405063291,
      "grad_norm": 0.0013730882201343775,
      "learning_rate": 7.217489335783212e-06,
      "loss": 0.0,
      "step": 1802
    },
    {
      "epoch": 0.5845873417721519,
      "grad_norm": 0.00027484947349876165,
      "learning_rate": 7.211799202824389e-06,
      "loss": 0.0,
      "step": 1804
    },
    {
      "epoch": 0.5852354430379747,
      "grad_norm": 0.11204996705055237,
      "learning_rate": 7.206105506216107e-06,
      "loss": 0.0005,
      "step": 1806
    },
    {
      "epoch": 0.5858835443037975,
      "grad_norm": 0.010664002038538456,
      "learning_rate": 7.200408255132046e-06,
      "loss": 0.0,
      "step": 1808
    },
    {
      "epoch": 0.5865316455696202,
      "grad_norm": 0.02932300604879856,
      "learning_rate": 7.194707458751615e-06,
      "loss": 0.0,
      "step": 1810
    },
    {
      "epoch": 0.587179746835443,
      "grad_norm": 0.1193203330039978,
      "learning_rate": 7.189003126259932e-06,
      "loss": 0.0002,
      "step": 1812
    },
    {
      "epoch": 0.5878278481012659,
      "grad_norm": 0.020759958773851395,
      "learning_rate": 7.1832952668478155e-06,
      "loss": 0.0,
      "step": 1814
    },
    {
      "epoch": 0.5884759493670886,
      "grad_norm": 0.07171918451786041,
      "learning_rate": 7.177583889711763e-06,
      "loss": 0.0001,
      "step": 1816
    },
    {
      "epoch": 0.5891240506329114,
      "grad_norm": 0.0004759156727232039,
      "learning_rate": 7.1718690040539404e-06,
      "loss": 0.0001,
      "step": 1818
    },
    {
      "epoch": 0.5897721518987342,
      "grad_norm": 0.0007953441818244755,
      "learning_rate": 7.166150619082171e-06,
      "loss": 0.0,
      "step": 1820
    },
    {
      "epoch": 0.5904202531645569,
      "grad_norm": 0.008654981851577759,
      "learning_rate": 7.160428744009913e-06,
      "loss": 0.0001,
      "step": 1822
    },
    {
      "epoch": 0.5910683544303797,
      "grad_norm": 0.10353202372789383,
      "learning_rate": 7.154703388056246e-06,
      "loss": 0.0001,
      "step": 1824
    },
    {
      "epoch": 0.5917164556962026,
      "grad_norm": 0.0015258926432579756,
      "learning_rate": 7.148974560445859e-06,
      "loss": 0.0001,
      "step": 1826
    },
    {
      "epoch": 0.5923645569620253,
      "grad_norm": 0.00039332304731942713,
      "learning_rate": 7.143242270409039e-06,
      "loss": 0.0,
      "step": 1828
    },
    {
      "epoch": 0.5930126582278481,
      "grad_norm": 0.03245238959789276,
      "learning_rate": 7.137506527181643e-06,
      "loss": 0.0,
      "step": 1830
    },
    {
      "epoch": 0.5936607594936709,
      "grad_norm": 0.004285217262804508,
      "learning_rate": 7.131767340005102e-06,
      "loss": 0.0,
      "step": 1832
    },
    {
      "epoch": 0.5943088607594936,
      "grad_norm": 0.001428083865903318,
      "learning_rate": 7.126024718126388e-06,
      "loss": 0.0,
      "step": 1834
    },
    {
      "epoch": 0.5949569620253164,
      "grad_norm": 0.005078149959445,
      "learning_rate": 7.12027867079801e-06,
      "loss": 0.0,
      "step": 1836
    },
    {
      "epoch": 0.5956050632911393,
      "grad_norm": 0.0003838093834929168,
      "learning_rate": 7.114529207277996e-06,
      "loss": 0.0001,
      "step": 1838
    },
    {
      "epoch": 0.596253164556962,
      "grad_norm": 0.0006412325892597437,
      "learning_rate": 7.1087763368298764e-06,
      "loss": 0.0,
      "step": 1840
    },
    {
      "epoch": 0.596253164556962,
      "eval_accuracy": 0.9999988402108444,
      "eval_loss": 1.187924590340117e-05,
      "eval_runtime": 100.0044,
      "eval_samples_per_second": 49.998,
      "eval_steps_per_second": 12.499,
      "step": 1840
    },
    {
      "epoch": 0.5969012658227848,
      "grad_norm": 0.002054619137197733,
      "learning_rate": 7.103020068722675e-06,
      "loss": 0.0,
      "step": 1842
    },
    {
      "epoch": 0.5975493670886076,
      "grad_norm": 0.0006610119016841054,
      "learning_rate": 7.0972604122308865e-06,
      "loss": 0.0,
      "step": 1844
    },
    {
      "epoch": 0.5981974683544303,
      "grad_norm": 0.0009412910440005362,
      "learning_rate": 7.0914973766344645e-06,
      "loss": 0.0,
      "step": 1846
    },
    {
      "epoch": 0.5988455696202531,
      "grad_norm": 0.0044436645694077015,
      "learning_rate": 7.085730971218809e-06,
      "loss": 0.0,
      "step": 1848
    },
    {
      "epoch": 0.599493670886076,
      "grad_norm": 0.0017420395743101835,
      "learning_rate": 7.079961205274749e-06,
      "loss": 0.0,
      "step": 1850
    },
    {
      "epoch": 0.6001417721518988,
      "grad_norm": 0.0020820838399231434,
      "learning_rate": 7.074188088098528e-06,
      "loss": 0.0,
      "step": 1852
    },
    {
      "epoch": 0.6007898734177215,
      "grad_norm": 0.00021655585442204028,
      "learning_rate": 7.0684116289917885e-06,
      "loss": 0.0,
      "step": 1854
    },
    {
      "epoch": 0.6014379746835443,
      "grad_norm": 0.002999892458319664,
      "learning_rate": 7.062631837261556e-06,
      "loss": 0.0,
      "step": 1856
    },
    {
      "epoch": 0.6020860759493671,
      "grad_norm": 0.00042861109250225127,
      "learning_rate": 7.05684872222023e-06,
      "loss": 0.0001,
      "step": 1858
    },
    {
      "epoch": 0.6027341772151898,
      "grad_norm": 0.0014900608221068978,
      "learning_rate": 7.05106229318556e-06,
      "loss": 0.0,
      "step": 1860
    },
    {
      "epoch": 0.6033822784810127,
      "grad_norm": 0.03731833025813103,
      "learning_rate": 7.045272559480636e-06,
      "loss": 0.0,
      "step": 1862
    },
    {
      "epoch": 0.6040303797468355,
      "grad_norm": 0.004172660410404205,
      "learning_rate": 7.039479530433875e-06,
      "loss": 0.0,
      "step": 1864
    },
    {
      "epoch": 0.6046784810126582,
      "grad_norm": 0.00023338549362961203,
      "learning_rate": 7.033683215379002e-06,
      "loss": 0.0,
      "step": 1866
    },
    {
      "epoch": 0.605326582278481,
      "grad_norm": 0.06045857444405556,
      "learning_rate": 7.027883623655035e-06,
      "loss": 0.0001,
      "step": 1868
    },
    {
      "epoch": 0.6059746835443038,
      "grad_norm": 0.047318872064352036,
      "learning_rate": 7.022080764606272e-06,
      "loss": 0.0001,
      "step": 1870
    },
    {
      "epoch": 0.6066227848101265,
      "grad_norm": 0.030598580837249756,
      "learning_rate": 7.016274647582276e-06,
      "loss": 0.0,
      "step": 1872
    },
    {
      "epoch": 0.6072708860759494,
      "grad_norm": 0.0005781942745670676,
      "learning_rate": 7.010465281937859e-06,
      "loss": 0.0,
      "step": 1874
    },
    {
      "epoch": 0.6079189873417722,
      "grad_norm": 0.10540540516376495,
      "learning_rate": 7.004652677033069e-06,
      "loss": 0.0001,
      "step": 1876
    },
    {
      "epoch": 0.6085670886075949,
      "grad_norm": 0.00097875006031245,
      "learning_rate": 6.99883684223317e-06,
      "loss": 0.0,
      "step": 1878
    },
    {
      "epoch": 0.6092151898734177,
      "grad_norm": 0.09228157252073288,
      "learning_rate": 6.993017786908631e-06,
      "loss": 0.0,
      "step": 1880
    },
    {
      "epoch": 0.6092151898734177,
      "eval_accuracy": 0.9999984465216039,
      "eval_loss": 9.824881090025883e-06,
      "eval_runtime": 101.3407,
      "eval_samples_per_second": 49.339,
      "eval_steps_per_second": 12.335,
      "step": 1880
    },
    {
      "epoch": 0.6098632911392405,
      "grad_norm": 0.0008249924867413938,
      "learning_rate": 6.9871955204351094e-06,
      "loss": 0.0,
      "step": 1882
    },
    {
      "epoch": 0.6105113924050632,
      "grad_norm": 0.02752879448235035,
      "learning_rate": 6.9813700521934394e-06,
      "loss": 0.0,
      "step": 1884
    },
    {
      "epoch": 0.6111594936708861,
      "grad_norm": 0.002536400454118848,
      "learning_rate": 6.9755413915696105e-06,
      "loss": 0.0,
      "step": 1886
    },
    {
      "epoch": 0.6118075949367089,
      "grad_norm": 0.0435812771320343,
      "learning_rate": 6.9697095479547564e-06,
      "loss": 0.0,
      "step": 1888
    },
    {
      "epoch": 0.6124556962025316,
      "grad_norm": 0.02748386561870575,
      "learning_rate": 6.963874530745141e-06,
      "loss": 0.0,
      "step": 1890
    },
    {
      "epoch": 0.6131037974683544,
      "grad_norm": 0.09339939802885056,
      "learning_rate": 6.95803634934214e-06,
      "loss": 0.0001,
      "step": 1892
    },
    {
      "epoch": 0.6137518987341772,
      "grad_norm": 0.016583425924181938,
      "learning_rate": 6.952195013152227e-06,
      "loss": 0.0,
      "step": 1894
    },
    {
      "epoch": 0.6144,
      "grad_norm": 0.00017733302956912667,
      "learning_rate": 6.946350531586959e-06,
      "loss": 0.0,
      "step": 1896
    },
    {
      "epoch": 0.6150481012658228,
      "grad_norm": 0.00218210369348526,
      "learning_rate": 6.940502914062961e-06,
      "loss": 0.0001,
      "step": 1898
    },
    {
      "epoch": 0.6156962025316456,
      "grad_norm": 0.0015262187225744128,
      "learning_rate": 6.934652170001911e-06,
      "loss": 0.0002,
      "step": 1900
    },
    {
      "epoch": 0.6163443037974684,
      "grad_norm": 0.00028145458782091737,
      "learning_rate": 6.928798308830524e-06,
      "loss": 0.0,
      "step": 1902
    },
    {
      "epoch": 0.6169924050632911,
      "grad_norm": 0.09535115212202072,
      "learning_rate": 6.922941339980538e-06,
      "loss": 0.0,
      "step": 1904
    },
    {
      "epoch": 0.6176405063291139,
      "grad_norm": 0.000556255632545799,
      "learning_rate": 6.917081272888697e-06,
      "loss": 0.0002,
      "step": 1906
    },
    {
      "epoch": 0.6182886075949368,
      "grad_norm": 0.0006218308117240667,
      "learning_rate": 6.911218116996738e-06,
      "loss": 0.0,
      "step": 1908
    },
    {
      "epoch": 0.6189367088607595,
      "grad_norm": 0.0027631232514977455,
      "learning_rate": 6.905351881751372e-06,
      "loss": 0.0,
      "step": 1910
    },
    {
      "epoch": 0.6195848101265823,
      "grad_norm": 0.03704715520143509,
      "learning_rate": 6.899482576604275e-06,
      "loss": 0.0,
      "step": 1912
    },
    {
      "epoch": 0.6202329113924051,
      "grad_norm": 0.060138948261737823,
      "learning_rate": 6.893610211012067e-06,
      "loss": 0.0002,
      "step": 1914
    },
    {
      "epoch": 0.6208810126582278,
      "grad_norm": 0.09946001321077347,
      "learning_rate": 6.887734794436301e-06,
      "loss": 0.0001,
      "step": 1916
    },
    {
      "epoch": 0.6215291139240506,
      "grad_norm": 0.011256850324571133,
      "learning_rate": 6.881856336343442e-06,
      "loss": 0.0,
      "step": 1918
    },
    {
      "epoch": 0.6221772151898735,
      "grad_norm": 0.03586261346936226,
      "learning_rate": 6.8759748462048595e-06,
      "loss": 0.0001,
      "step": 1920
    },
    {
      "epoch": 0.6221772151898735,
      "eval_accuracy": 0.9999977681314496,
      "eval_loss": 1.365809021081077e-05,
      "eval_runtime": 101.3963,
      "eval_samples_per_second": 49.311,
      "eval_steps_per_second": 12.328,
      "step": 1920
    },
    {
      "epoch": 0.6228253164556962,
      "grad_norm": 0.0019096017349511385,
      "learning_rate": 6.870090333496807e-06,
      "loss": 0.0,
      "step": 1922
    },
    {
      "epoch": 0.623473417721519,
      "grad_norm": 0.0029806571546941996,
      "learning_rate": 6.864202807700407e-06,
      "loss": 0.0,
      "step": 1924
    },
    {
      "epoch": 0.6241215189873418,
      "grad_norm": 0.022626426070928574,
      "learning_rate": 6.858312278301638e-06,
      "loss": 0.0,
      "step": 1926
    },
    {
      "epoch": 0.6247696202531645,
      "grad_norm": 0.004040713422000408,
      "learning_rate": 6.852418754791317e-06,
      "loss": 0.0,
      "step": 1928
    },
    {
      "epoch": 0.6254177215189873,
      "grad_norm": 0.0006340648396871984,
      "learning_rate": 6.8465222466650835e-06,
      "loss": 0.0,
      "step": 1930
    },
    {
      "epoch": 0.6260658227848102,
      "grad_norm": 0.007957760244607925,
      "learning_rate": 6.840622763423391e-06,
      "loss": 0.0,
      "step": 1932
    },
    {
      "epoch": 0.6267139240506329,
      "grad_norm": 0.0036092489026486874,
      "learning_rate": 6.83472031457148e-06,
      "loss": 0.0,
      "step": 1934
    },
    {
      "epoch": 0.6273620253164557,
      "grad_norm": 0.0009143880452029407,
      "learning_rate": 6.828814909619374e-06,
      "loss": 0.0,
      "step": 1936
    },
    {
      "epoch": 0.6280101265822785,
      "grad_norm": 0.0002480478142388165,
      "learning_rate": 6.822906558081856e-06,
      "loss": 0.0,
      "step": 1938
    },
    {
      "epoch": 0.6286582278481012,
      "grad_norm": 0.0009710952290333807,
      "learning_rate": 6.81699526947846e-06,
      "loss": 0.0,
      "step": 1940
    },
    {
      "epoch": 0.629306329113924,
      "grad_norm": 0.07902362197637558,
      "learning_rate": 6.81108105333345e-06,
      "loss": 0.0,
      "step": 1942
    },
    {
      "epoch": 0.6299544303797469,
      "grad_norm": 0.0639423131942749,
      "learning_rate": 6.8051639191758065e-06,
      "loss": 0.0,
      "step": 1944
    },
    {
      "epoch": 0.6306025316455697,
      "grad_norm": 0.00019927980611100793,
      "learning_rate": 6.799243876539213e-06,
      "loss": 0.0,
      "step": 1946
    },
    {
      "epoch": 0.6312506329113924,
      "grad_norm": 0.0010397526202723384,
      "learning_rate": 6.793320934962039e-06,
      "loss": 0.0,
      "step": 1948
    },
    {
      "epoch": 0.6318987341772152,
      "grad_norm": 0.01125252153724432,
      "learning_rate": 6.787395103987323e-06,
      "loss": 0.0001,
      "step": 1950
    },
    {
      "epoch": 0.632546835443038,
      "grad_norm": 0.0007871537236496806,
      "learning_rate": 6.781466393162761e-06,
      "loss": 0.0,
      "step": 1952
    },
    {
      "epoch": 0.6331949367088607,
      "grad_norm": 0.002556710736826062,
      "learning_rate": 6.775534812040686e-06,
      "loss": 0.0,
      "step": 1954
    },
    {
      "epoch": 0.6338430379746836,
      "grad_norm": 0.000299537496175617,
      "learning_rate": 6.76960037017806e-06,
      "loss": 0.0,
      "step": 1956
    },
    {
      "epoch": 0.6344911392405064,
      "grad_norm": 0.0010419267928227782,
      "learning_rate": 6.763663077136451e-06,
      "loss": 0.0,
      "step": 1958
    },
    {
      "epoch": 0.6351392405063291,
      "grad_norm": 0.09013451635837555,
      "learning_rate": 6.757722942482022e-06,
      "loss": 0.0001,
      "step": 1960
    },
    {
      "epoch": 0.6351392405063291,
      "eval_accuracy": 0.9999996055226824,
      "eval_loss": 5.647259058605414e-06,
      "eval_runtime": 101.9211,
      "eval_samples_per_second": 49.058,
      "eval_steps_per_second": 12.264,
      "step": 1960
    },
    {
      "epoch": 0.6357873417721519,
      "grad_norm": 0.00781452190130949,
      "learning_rate": 6.751779975785515e-06,
      "loss": 0.0,
      "step": 1962
    },
    {
      "epoch": 0.6364354430379747,
      "grad_norm": 0.040681142359972,
      "learning_rate": 6.745834186622232e-06,
      "loss": 0.0,
      "step": 1964
    },
    {
      "epoch": 0.6370835443037974,
      "grad_norm": 0.0002076182863675058,
      "learning_rate": 6.739885584572026e-06,
      "loss": 0.0,
      "step": 1966
    },
    {
      "epoch": 0.6377316455696203,
      "grad_norm": 0.00034984247758984566,
      "learning_rate": 6.733934179219281e-06,
      "loss": 0.0,
      "step": 1968
    },
    {
      "epoch": 0.6383797468354431,
      "grad_norm": 0.0007839368190616369,
      "learning_rate": 6.727979980152899e-06,
      "loss": 0.0,
      "step": 1970
    },
    {
      "epoch": 0.6390278481012658,
      "grad_norm": 0.0017967225285246968,
      "learning_rate": 6.7220229969662776e-06,
      "loss": 0.0001,
      "step": 1972
    },
    {
      "epoch": 0.6396759493670886,
      "grad_norm": 0.0005148016498424113,
      "learning_rate": 6.716063239257307e-06,
      "loss": 0.0,
      "step": 1974
    },
    {
      "epoch": 0.6403240506329114,
      "grad_norm": 0.0007883037906140089,
      "learning_rate": 6.710100716628345e-06,
      "loss": 0.0,
      "step": 1976
    },
    {
      "epoch": 0.6409721518987341,
      "grad_norm": 0.0005258942837826908,
      "learning_rate": 6.704135438686203e-06,
      "loss": 0.0,
      "step": 1978
    },
    {
      "epoch": 0.641620253164557,
      "grad_norm": 0.00028024116181768477,
      "learning_rate": 6.698167415042135e-06,
      "loss": 0.0,
      "step": 1980
    },
    {
      "epoch": 0.6422683544303798,
      "grad_norm": 0.0011886393185704947,
      "learning_rate": 6.692196655311814e-06,
      "loss": 0.0,
      "step": 1982
    },
    {
      "epoch": 0.6429164556962025,
      "grad_norm": 0.04780174791812897,
      "learning_rate": 6.686223169115328e-06,
      "loss": 0.0,
      "step": 1984
    },
    {
      "epoch": 0.6435645569620253,
      "grad_norm": 0.060146622359752655,
      "learning_rate": 6.680246966077151e-06,
      "loss": 0.0,
      "step": 1986
    },
    {
      "epoch": 0.6442126582278481,
      "grad_norm": 0.00024338187358807772,
      "learning_rate": 6.674268055826139e-06,
      "loss": 0.0,
      "step": 1988
    },
    {
      "epoch": 0.6448607594936708,
      "grad_norm": 0.014128066599369049,
      "learning_rate": 6.6682864479955075e-06,
      "loss": 0.0,
      "step": 1990
    },
    {
      "epoch": 0.6455088607594937,
      "grad_norm": 0.014804531820118427,
      "learning_rate": 6.66230215222282e-06,
      "loss": 0.0,
      "step": 1992
    },
    {
      "epoch": 0.6461569620253165,
      "grad_norm": 0.0006120464531704783,
      "learning_rate": 6.656315178149971e-06,
      "loss": 0.0,
      "step": 1994
    },
    {
      "epoch": 0.6468050632911393,
      "grad_norm": 0.000343716616043821,
      "learning_rate": 6.650325535423166e-06,
      "loss": 0.0,
      "step": 1996
    },
    {
      "epoch": 0.647453164556962,
      "grad_norm": 0.0001859806216089055,
      "learning_rate": 6.644333233692917e-06,
      "loss": 0.0,
      "step": 1998
    },
    {
      "epoch": 0.6481012658227848,
      "grad_norm": 0.030580680817365646,
      "learning_rate": 6.638338282614014e-06,
      "loss": 0.0,
      "step": 2000
    },
    {
      "epoch": 0.6481012658227848,
      "eval_accuracy": 0.999998421308061,
      "eval_loss": 1.1490382348711137e-05,
      "eval_runtime": 101.8472,
      "eval_samples_per_second": 49.093,
      "eval_steps_per_second": 12.273,
      "step": 2000
    },
    {
      "epoch": 0.6487493670886076,
      "grad_norm": 0.0011633732356131077,
      "learning_rate": 6.6323406918455205e-06,
      "loss": 0.0,
      "step": 2002
    },
    {
      "epoch": 0.6493974683544304,
      "grad_norm": 0.0025772363878786564,
      "learning_rate": 6.6263404710507495e-06,
      "loss": 0.0,
      "step": 2004
    },
    {
      "epoch": 0.6500455696202532,
      "grad_norm": 0.002858225954696536,
      "learning_rate": 6.6203376298972535e-06,
      "loss": 0.0,
      "step": 2006
    },
    {
      "epoch": 0.650693670886076,
      "grad_norm": 0.0053984252735972404,
      "learning_rate": 6.614332178056806e-06,
      "loss": 0.0,
      "step": 2008
    },
    {
      "epoch": 0.6513417721518987,
      "grad_norm": 0.000622738036327064,
      "learning_rate": 6.608324125205389e-06,
      "loss": 0.0,
      "step": 2010
    },
    {
      "epoch": 0.6519898734177215,
      "grad_norm": 0.004378859885036945,
      "learning_rate": 6.60231348102317e-06,
      "loss": 0.0,
      "step": 2012
    },
    {
      "epoch": 0.6526379746835443,
      "grad_norm": 0.0009536651195958257,
      "learning_rate": 6.596300255194496e-06,
      "loss": 0.0,
      "step": 2014
    },
    {
      "epoch": 0.653286075949367,
      "grad_norm": 0.001671510748565197,
      "learning_rate": 6.590284457407876e-06,
      "loss": 0.0,
      "step": 2016
    },
    {
      "epoch": 0.6539341772151899,
      "grad_norm": 0.0006021432927809656,
      "learning_rate": 6.5842660973559545e-06,
      "loss": 0.0,
      "step": 2018
    },
    {
      "epoch": 0.6545822784810127,
      "grad_norm": 0.00020325230434536934,
      "learning_rate": 6.578245184735513e-06,
      "loss": 0.0,
      "step": 2020
    },
    {
      "epoch": 0.6552303797468354,
      "grad_norm": 0.00012371053162496537,
      "learning_rate": 6.572221729247441e-06,
      "loss": 0.0,
      "step": 2022
    },
    {
      "epoch": 0.6558784810126582,
      "grad_norm": 0.014866070821881294,
      "learning_rate": 6.5661957405967255e-06,
      "loss": 0.0,
      "step": 2024
    },
    {
      "epoch": 0.656526582278481,
      "grad_norm": 0.009082837030291557,
      "learning_rate": 6.560167228492436e-06,
      "loss": 0.0,
      "step": 2026
    },
    {
      "epoch": 0.6571746835443038,
      "grad_norm": 0.004811509046703577,
      "learning_rate": 6.554136202647707e-06,
      "loss": 0.0,
      "step": 2028
    },
    {
      "epoch": 0.6578227848101266,
      "grad_norm": 0.00021759387163911015,
      "learning_rate": 6.548102672779725e-06,
      "loss": 0.0,
      "step": 2030
    },
    {
      "epoch": 0.6584708860759494,
      "grad_norm": 0.001678065280430019,
      "learning_rate": 6.5420666486097084e-06,
      "loss": 0.0,
      "step": 2032
    },
    {
      "epoch": 0.6591189873417721,
      "grad_norm": 0.00032246127375401556,
      "learning_rate": 6.536028139862895e-06,
      "loss": 0.0,
      "step": 2034
    },
    {
      "epoch": 0.6597670886075949,
      "grad_norm": 0.05951784551143646,
      "learning_rate": 6.529987156268527e-06,
      "loss": 0.0,
      "step": 2036
    },
    {
      "epoch": 0.6604151898734177,
      "grad_norm": 0.022436069324612617,
      "learning_rate": 6.523943707559832e-06,
      "loss": 0.0,
      "step": 2038
    },
    {
      "epoch": 0.6610632911392406,
      "grad_norm": 0.0005385543336160481,
      "learning_rate": 6.517897803474011e-06,
      "loss": 0.0,
      "step": 2040
    },
    {
      "epoch": 0.6610632911392406,
      "eval_accuracy": 0.9999992102657654,
      "eval_loss": 6.353677235892974e-06,
      "eval_runtime": 105.7714,
      "eval_samples_per_second": 47.272,
      "eval_steps_per_second": 11.818,
      "step": 2040
    },
    {
      "epoch": 0.6617113924050633,
      "grad_norm": 0.002471289364621043,
      "learning_rate": 6.5118494537522235e-06,
      "loss": 0.0,
      "step": 2042
    },
    {
      "epoch": 0.6623594936708861,
      "grad_norm": 0.15845777094364166,
      "learning_rate": 6.505798668139563e-06,
      "loss": 0.0,
      "step": 2044
    },
    {
      "epoch": 0.6630075949367089,
      "grad_norm": 0.004233754705637693,
      "learning_rate": 6.499745456385054e-06,
      "loss": 0.0,
      "step": 2046
    },
    {
      "epoch": 0.6636556962025316,
      "grad_norm": 0.06344658881425858,
      "learning_rate": 6.493689828241625e-06,
      "loss": 0.0001,
      "step": 2048
    },
    {
      "epoch": 0.6643037974683544,
      "grad_norm": 0.05400916561484337,
      "learning_rate": 6.4876317934661036e-06,
      "loss": 0.0,
      "step": 2050
    },
    {
      "epoch": 0.6649518987341773,
      "grad_norm": 0.025929005816578865,
      "learning_rate": 6.481571361819189e-06,
      "loss": 0.0,
      "step": 2052
    },
    {
      "epoch": 0.6656,
      "grad_norm": 0.0014809060376137495,
      "learning_rate": 6.475508543065445e-06,
      "loss": 0.0,
      "step": 2054
    },
    {
      "epoch": 0.6662481012658228,
      "grad_norm": 0.016997920349240303,
      "learning_rate": 6.469443346973281e-06,
      "loss": 0.0,
      "step": 2056
    },
    {
      "epoch": 0.6668962025316456,
      "grad_norm": 0.00024033906811382622,
      "learning_rate": 6.463375783314938e-06,
      "loss": 0.0,
      "step": 2058
    },
    {
      "epoch": 0.6675443037974683,
      "grad_norm": 0.001820797217078507,
      "learning_rate": 6.457305861866471e-06,
      "loss": 0.0,
      "step": 2060
    },
    {
      "epoch": 0.6681924050632911,
      "grad_norm": 0.034896451979875565,
      "learning_rate": 6.451233592407732e-06,
      "loss": 0.0,
      "step": 2062
    },
    {
      "epoch": 0.668840506329114,
      "grad_norm": 0.00012074632104486227,
      "learning_rate": 6.445158984722358e-06,
      "loss": 0.0,
      "step": 2064
    },
    {
      "epoch": 0.6694886075949367,
      "grad_norm": 0.0386367104947567,
      "learning_rate": 6.439082048597755e-06,
      "loss": 0.0,
      "step": 2066
    },
    {
      "epoch": 0.6701367088607595,
      "grad_norm": 0.0002681062906049192,
      "learning_rate": 6.433002793825076e-06,
      "loss": 0.0,
      "step": 2068
    },
    {
      "epoch": 0.6707848101265823,
      "grad_norm": 0.0019434065325185657,
      "learning_rate": 6.426921230199215e-06,
      "loss": 0.0,
      "step": 2070
    },
    {
      "epoch": 0.671432911392405,
      "grad_norm": 0.006111921276897192,
      "learning_rate": 6.420837367518781e-06,
      "loss": 0.0,
      "step": 2072
    },
    {
      "epoch": 0.6720810126582278,
      "grad_norm": 0.00020428838615771383,
      "learning_rate": 6.414751215586091e-06,
      "loss": 0.0,
      "step": 2074
    },
    {
      "epoch": 0.6727291139240507,
      "grad_norm": 0.0007174932979978621,
      "learning_rate": 6.408662784207149e-06,
      "loss": 0.0,
      "step": 2076
    },
    {
      "epoch": 0.6733772151898734,
      "grad_norm": 0.00031421848689205945,
      "learning_rate": 6.402572083191632e-06,
      "loss": 0.0,
      "step": 2078
    },
    {
      "epoch": 0.6740253164556962,
      "grad_norm": 0.0009968363447114825,
      "learning_rate": 6.396479122352872e-06,
      "loss": 0.0,
      "step": 2080
    },
    {
      "epoch": 0.6740253164556962,
      "eval_accuracy": 0.9999969388560157,
      "eval_loss": 1.1616197298280895e-05,
      "eval_runtime": 105.7511,
      "eval_samples_per_second": 47.281,
      "eval_steps_per_second": 11.82,
      "step": 2080
    },
    {
      "epoch": 0.674673417721519,
      "grad_norm": 0.012512893415987492,
      "learning_rate": 6.390383911507845e-06,
      "loss": 0.0,
      "step": 2082
    },
    {
      "epoch": 0.6753215189873417,
      "grad_norm": 0.02030804194509983,
      "learning_rate": 6.384286460477149e-06,
      "loss": 0.0,
      "step": 2084
    },
    {
      "epoch": 0.6759696202531645,
      "grad_norm": 0.00017589876370038837,
      "learning_rate": 6.378186779084996e-06,
      "loss": 0.0,
      "step": 2086
    },
    {
      "epoch": 0.6766177215189874,
      "grad_norm": 0.001576538779772818,
      "learning_rate": 6.3720848771591884e-06,
      "loss": 0.0,
      "step": 2088
    },
    {
      "epoch": 0.6772658227848102,
      "grad_norm": 9.243958629667759e-05,
      "learning_rate": 6.3659807645311056e-06,
      "loss": 0.0,
      "step": 2090
    },
    {
      "epoch": 0.6779139240506329,
      "grad_norm": 0.00014198373537510633,
      "learning_rate": 6.359874451035688e-06,
      "loss": 0.0,
      "step": 2092
    },
    {
      "epoch": 0.6785620253164557,
      "grad_norm": 0.0007102972595021129,
      "learning_rate": 6.3537659465114275e-06,
      "loss": 0.0,
      "step": 2094
    },
    {
      "epoch": 0.6792101265822785,
      "grad_norm": 0.00011982554860878736,
      "learning_rate": 6.34765526080034e-06,
      "loss": 0.0001,
      "step": 2096
    },
    {
      "epoch": 0.6798582278481012,
      "grad_norm": 0.00011331916903145611,
      "learning_rate": 6.34154240374796e-06,
      "loss": 0.0,
      "step": 2098
    },
    {
      "epoch": 0.6805063291139241,
      "grad_norm": 0.00014351024583447725,
      "learning_rate": 6.33542738520332e-06,
      "loss": 0.0,
      "step": 2100
    },
    {
      "epoch": 0.6811544303797469,
      "grad_norm": 0.0023922757245600224,
      "learning_rate": 6.329310215018931e-06,
      "loss": 0.0,
      "step": 2102
    },
    {
      "epoch": 0.6818025316455696,
      "grad_norm": 0.00023756278096698225,
      "learning_rate": 6.323190903050776e-06,
      "loss": 0.0,
      "step": 2104
    },
    {
      "epoch": 0.6824506329113924,
      "grad_norm": 0.00014805899991188198,
      "learning_rate": 6.317069459158284e-06,
      "loss": 0.0,
      "step": 2106
    },
    {
      "epoch": 0.6830987341772152,
      "grad_norm": 0.002821107627823949,
      "learning_rate": 6.310945893204324e-06,
      "loss": 0.0,
      "step": 2108
    },
    {
      "epoch": 0.6837468354430379,
      "grad_norm": 0.0003693237667903304,
      "learning_rate": 6.30482021505518e-06,
      "loss": 0.0,
      "step": 2110
    },
    {
      "epoch": 0.6843949367088608,
      "grad_norm": 0.0005288827233016491,
      "learning_rate": 6.298692434580543e-06,
      "loss": 0.0,
      "step": 2112
    },
    {
      "epoch": 0.6850430379746836,
      "grad_norm": 0.00016826622595544904,
      "learning_rate": 6.292562561653486e-06,
      "loss": 0.0,
      "step": 2114
    },
    {
      "epoch": 0.6856911392405063,
      "grad_norm": 0.0013678655959665775,
      "learning_rate": 6.286430606150458e-06,
      "loss": 0.0,
      "step": 2116
    },
    {
      "epoch": 0.6863392405063291,
      "grad_norm": 0.00178819231223315,
      "learning_rate": 6.280296577951262e-06,
      "loss": 0.0,
      "step": 2118
    },
    {
      "epoch": 0.6869873417721519,
      "grad_norm": 0.004927210509777069,
      "learning_rate": 6.27416048693904e-06,
      "loss": 0.0,
      "step": 2120
    },
    {
      "epoch": 0.6869873417721519,
      "eval_accuracy": 0.9999992141332499,
      "eval_loss": 4.2000106077466626e-06,
      "eval_runtime": 106.5125,
      "eval_samples_per_second": 46.943,
      "eval_steps_per_second": 11.736,
      "step": 2120
    },
    {
      "epoch": 0.6876354430379746,
      "grad_norm": 0.0004625746514648199,
      "learning_rate": 6.268022343000258e-06,
      "loss": 0.0,
      "step": 2122
    },
    {
      "epoch": 0.6882835443037975,
      "grad_norm": 0.0015599478501826525,
      "learning_rate": 6.261882156024688e-06,
      "loss": 0.0,
      "step": 2124
    },
    {
      "epoch": 0.6889316455696203,
      "grad_norm": 0.00570697570219636,
      "learning_rate": 6.255739935905396e-06,
      "loss": 0.0,
      "step": 2126
    },
    {
      "epoch": 0.689579746835443,
      "grad_norm": 0.03438084200024605,
      "learning_rate": 6.249595692538726e-06,
      "loss": 0.0,
      "step": 2128
    },
    {
      "epoch": 0.6902278481012658,
      "grad_norm": 0.07320534437894821,
      "learning_rate": 6.243449435824276e-06,
      "loss": 0.0,
      "step": 2130
    },
    {
      "epoch": 0.6908759493670886,
      "grad_norm": 7.549316069344059e-05,
      "learning_rate": 6.2373011756648905e-06,
      "loss": 0.0,
      "step": 2132
    },
    {
      "epoch": 0.6915240506329114,
      "grad_norm": 0.0001575332717038691,
      "learning_rate": 6.231150921966643e-06,
      "loss": 0.0,
      "step": 2134
    },
    {
      "epoch": 0.6921721518987342,
      "grad_norm": 0.0009007062180899084,
      "learning_rate": 6.22499868463882e-06,
      "loss": 0.0,
      "step": 2136
    },
    {
      "epoch": 0.692820253164557,
      "grad_norm": 0.00013916321040596813,
      "learning_rate": 6.2188444735939e-06,
      "loss": 0.0,
      "step": 2138
    },
    {
      "epoch": 0.6934683544303798,
      "grad_norm": 0.0005380238289944828,
      "learning_rate": 6.212688298747546e-06,
      "loss": 0.0,
      "step": 2140
    },
    {
      "epoch": 0.6941164556962025,
      "grad_norm": 0.0021836934611201286,
      "learning_rate": 6.206530170018581e-06,
      "loss": 0.0,
      "step": 2142
    },
    {
      "epoch": 0.6947645569620253,
      "grad_norm": 0.0002022333355853334,
      "learning_rate": 6.2003700973289785e-06,
      "loss": 0.0,
      "step": 2144
    },
    {
      "epoch": 0.6954126582278481,
      "grad_norm": 0.0020313316490501165,
      "learning_rate": 6.194208090603845e-06,
      "loss": 0.0,
      "step": 2146
    },
    {
      "epoch": 0.6960607594936709,
      "grad_norm": 0.00018010415078606457,
      "learning_rate": 6.1880441597714e-06,
      "loss": 0.0002,
      "step": 2148
    },
    {
      "epoch": 0.6967088607594937,
      "grad_norm": 0.007566555868834257,
      "learning_rate": 6.181878314762968e-06,
      "loss": 0.0,
      "step": 2150
    },
    {
      "epoch": 0.6973569620253165,
      "grad_norm": 8.020079258130863e-05,
      "learning_rate": 6.17571056551295e-06,
      "loss": 0.0,
      "step": 2152
    },
    {
      "epoch": 0.6980050632911392,
      "grad_norm": 0.000367714004823938,
      "learning_rate": 6.169540921958823e-06,
      "loss": 0.0,
      "step": 2154
    },
    {
      "epoch": 0.698653164556962,
      "grad_norm": 0.03462158143520355,
      "learning_rate": 6.163369394041112e-06,
      "loss": 0.0001,
      "step": 2156
    },
    {
      "epoch": 0.6993012658227848,
      "grad_norm": 0.0005256628501228988,
      "learning_rate": 6.157195991703378e-06,
      "loss": 0.0,
      "step": 2158
    },
    {
      "epoch": 0.6999493670886076,
      "grad_norm": 0.0006937627913430333,
      "learning_rate": 6.151020724892205e-06,
      "loss": 0.0,
      "step": 2160
    },
    {
      "epoch": 0.6999493670886076,
      "eval_accuracy": 0.9999996086105675,
      "eval_loss": 3.4480619888199726e-06,
      "eval_runtime": 106.2725,
      "eval_samples_per_second": 47.049,
      "eval_steps_per_second": 11.762,
      "step": 2160
    },
    {
      "epoch": 0.7005974683544304,
      "grad_norm": 0.00030152147519402206,
      "learning_rate": 6.144843603557176e-06,
      "loss": 0.0,
      "step": 2162
    },
    {
      "epoch": 0.7012455696202532,
      "grad_norm": 0.0002933765936177224,
      "learning_rate": 6.138664637650867e-06,
      "loss": 0.0,
      "step": 2164
    },
    {
      "epoch": 0.7018936708860759,
      "grad_norm": 0.0011545540764927864,
      "learning_rate": 6.132483837128823e-06,
      "loss": 0.0,
      "step": 2166
    },
    {
      "epoch": 0.7025417721518987,
      "grad_norm": 0.00038900863728486,
      "learning_rate": 6.1263012119495455e-06,
      "loss": 0.0,
      "step": 2168
    },
    {
      "epoch": 0.7031898734177215,
      "grad_norm": 0.0003068090882152319,
      "learning_rate": 6.120116772074478e-06,
      "loss": 0.0,
      "step": 2170
    },
    {
      "epoch": 0.7038379746835443,
      "grad_norm": 0.0013358498690649867,
      "learning_rate": 6.1139305274679835e-06,
      "loss": 0.0,
      "step": 2172
    },
    {
      "epoch": 0.7044860759493671,
      "grad_norm": 0.0202906783670187,
      "learning_rate": 6.107742488097338e-06,
      "loss": 0.0,
      "step": 2174
    },
    {
      "epoch": 0.7051341772151899,
      "grad_norm": 0.0001334721309831366,
      "learning_rate": 6.101552663932704e-06,
      "loss": 0.0,
      "step": 2176
    },
    {
      "epoch": 0.7057822784810126,
      "grad_norm": 0.008306547999382019,
      "learning_rate": 6.095361064947124e-06,
      "loss": 0.0,
      "step": 2178
    },
    {
      "epoch": 0.7064303797468354,
      "grad_norm": 0.0003196684701833874,
      "learning_rate": 6.089167701116498e-06,
      "loss": 0.0,
      "step": 2180
    },
    {
      "epoch": 0.7070784810126582,
      "grad_norm": 0.0022242022678256035,
      "learning_rate": 6.082972582419569e-06,
      "loss": 0.0,
      "step": 2182
    },
    {
      "epoch": 0.7077265822784811,
      "grad_norm": 0.0001364542986266315,
      "learning_rate": 6.076775718837911e-06,
      "loss": 0.0,
      "step": 2184
    },
    {
      "epoch": 0.7083746835443038,
      "grad_norm": 0.0006018828134983778,
      "learning_rate": 6.070577120355903e-06,
      "loss": 0.0,
      "step": 2186
    },
    {
      "epoch": 0.7090227848101266,
      "grad_norm": 0.000917417521122843,
      "learning_rate": 6.064376796960723e-06,
      "loss": 0.0,
      "step": 2188
    },
    {
      "epoch": 0.7096708860759494,
      "grad_norm": 0.00041519958176650107,
      "learning_rate": 6.058174758642332e-06,
      "loss": 0.0002,
      "step": 2190
    },
    {
      "epoch": 0.7103189873417721,
      "grad_norm": 7.386893412331119e-05,
      "learning_rate": 6.051971015393447e-06,
      "loss": 0.0,
      "step": 2192
    },
    {
      "epoch": 0.7109670886075949,
      "grad_norm": 0.0014044854324311018,
      "learning_rate": 6.045765577209536e-06,
      "loss": 0.0,
      "step": 2194
    },
    {
      "epoch": 0.7116151898734178,
      "grad_norm": 0.00031422023312188685,
      "learning_rate": 6.039558454088796e-06,
      "loss": 0.0,
      "step": 2196
    },
    {
      "epoch": 0.7122632911392405,
      "grad_norm": 0.00011502230336191133,
      "learning_rate": 6.033349656032143e-06,
      "loss": 0.0,
      "step": 2198
    },
    {
      "epoch": 0.7129113924050633,
      "grad_norm": 8.400805381825194e-05,
      "learning_rate": 6.027139193043185e-06,
      "loss": 0.0,
      "step": 2200
    },
    {
      "epoch": 0.7129113924050633,
      "eval_accuracy": 0.9999996055226824,
      "eval_loss": 3.1884246709523723e-06,
      "eval_runtime": 105.3881,
      "eval_samples_per_second": 47.444,
      "eval_steps_per_second": 11.861,
      "step": 2200
    },
    {
      "epoch": 0.7135594936708861,
      "grad_norm": 9.059866715688258e-05,
      "learning_rate": 6.0209270751282165e-06,
      "loss": 0.0,
      "step": 2202
    },
    {
      "epoch": 0.7142075949367088,
      "grad_norm": 0.009854246862232685,
      "learning_rate": 6.014713312296198e-06,
      "loss": 0.0,
      "step": 2204
    },
    {
      "epoch": 0.7148556962025316,
      "grad_norm": 0.00022580361110158265,
      "learning_rate": 6.0084979145587444e-06,
      "loss": 0.0,
      "step": 2206
    },
    {
      "epoch": 0.7155037974683545,
      "grad_norm": 0.0002859641390386969,
      "learning_rate": 6.002280891930093e-06,
      "loss": 0.0,
      "step": 2208
    },
    {
      "epoch": 0.7161518987341772,
      "grad_norm": 8.833897300064564e-05,
      "learning_rate": 5.996062254427112e-06,
      "loss": 0.0,
      "step": 2210
    },
    {
      "epoch": 0.7168,
      "grad_norm": 0.008814448490738869,
      "learning_rate": 5.989842012069265e-06,
      "loss": 0.0001,
      "step": 2212
    },
    {
      "epoch": 0.7174481012658228,
      "grad_norm": 0.05799108371138573,
      "learning_rate": 5.983620174878601e-06,
      "loss": 0.0002,
      "step": 2214
    },
    {
      "epoch": 0.7180962025316455,
      "grad_norm": 0.0002262613852508366,
      "learning_rate": 5.977396752879742e-06,
      "loss": 0.0,
      "step": 2216
    },
    {
      "epoch": 0.7187443037974683,
      "grad_norm": 0.00012421239807736129,
      "learning_rate": 5.97117175609986e-06,
      "loss": 0.0,
      "step": 2218
    },
    {
      "epoch": 0.7193924050632912,
      "grad_norm": 0.06700967252254486,
      "learning_rate": 5.964945194568669e-06,
      "loss": 0.0,
      "step": 2220
    },
    {
      "epoch": 0.7200405063291139,
      "grad_norm": 0.0005706732044927776,
      "learning_rate": 5.958717078318397e-06,
      "loss": 0.0002,
      "step": 2222
    },
    {
      "epoch": 0.7206886075949367,
      "grad_norm": 0.000181859519216232,
      "learning_rate": 5.952487417383782e-06,
      "loss": 0.0,
      "step": 2224
    },
    {
      "epoch": 0.7213367088607595,
      "grad_norm": 0.00011903359700227156,
      "learning_rate": 5.946256221802052e-06,
      "loss": 0.0,
      "step": 2226
    },
    {
      "epoch": 0.7219848101265823,
      "grad_norm": 9.368642349727452e-05,
      "learning_rate": 5.940023501612902e-06,
      "loss": 0.0,
      "step": 2228
    },
    {
      "epoch": 0.722632911392405,
      "grad_norm": 9.8723205155693e-05,
      "learning_rate": 5.9337892668584896e-06,
      "loss": 0.0,
      "step": 2230
    },
    {
      "epoch": 0.7232810126582279,
      "grad_norm": 9.426200267625973e-05,
      "learning_rate": 5.927553527583407e-06,
      "loss": 0.0,
      "step": 2232
    },
    {
      "epoch": 0.7239291139240507,
      "grad_norm": 0.00010700670827645808,
      "learning_rate": 5.9213162938346765e-06,
      "loss": 0.0,
      "step": 2234
    },
    {
      "epoch": 0.7245772151898734,
      "grad_norm": 0.0726417824625969,
      "learning_rate": 5.915077575661723e-06,
      "loss": 0.0002,
      "step": 2236
    },
    {
      "epoch": 0.7252253164556962,
      "grad_norm": 0.00032428756821900606,
      "learning_rate": 5.908837383116367e-06,
      "loss": 0.0,
      "step": 2238
    },
    {
      "epoch": 0.725873417721519,
      "grad_norm": 0.0002373167808400467,
      "learning_rate": 5.902595726252801e-06,
      "loss": 0.0,
      "step": 2240
    },
    {
      "epoch": 0.725873417721519,
      "eval_accuracy": 0.9999988227438175,
      "eval_loss": 4.894884114037268e-06,
      "eval_runtime": 105.8929,
      "eval_samples_per_second": 47.218,
      "eval_steps_per_second": 11.804,
      "step": 2240
    },
    {
      "epoch": 0.7265215189873417,
      "grad_norm": 0.00012383979628793895,
      "learning_rate": 5.896352615127578e-06,
      "loss": 0.0,
      "step": 2242
    },
    {
      "epoch": 0.7271696202531646,
      "grad_norm": 0.0012550248065963387,
      "learning_rate": 5.890108059799596e-06,
      "loss": 0.0,
      "step": 2244
    },
    {
      "epoch": 0.7278177215189874,
      "grad_norm": 0.09790094941854477,
      "learning_rate": 5.883862070330079e-06,
      "loss": 0.0001,
      "step": 2246
    },
    {
      "epoch": 0.7284658227848101,
      "grad_norm": 0.00010724524327088147,
      "learning_rate": 5.877614656782559e-06,
      "loss": 0.0,
      "step": 2248
    },
    {
      "epoch": 0.7291139240506329,
      "grad_norm": 0.0463559553027153,
      "learning_rate": 5.8713658292228695e-06,
      "loss": 0.0,
      "step": 2250
    },
    {
      "epoch": 0.7297620253164557,
      "grad_norm": 0.0001051365616149269,
      "learning_rate": 5.865115597719111e-06,
      "loss": 0.0,
      "step": 2252
    },
    {
      "epoch": 0.7304101265822784,
      "grad_norm": 0.0003419144486542791,
      "learning_rate": 5.858863972341656e-06,
      "loss": 0.0,
      "step": 2254
    },
    {
      "epoch": 0.7310582278481013,
      "grad_norm": 0.001729366835206747,
      "learning_rate": 5.85261096316312e-06,
      "loss": 0.0,
      "step": 2256
    },
    {
      "epoch": 0.7317063291139241,
      "grad_norm": 0.00013600761303678155,
      "learning_rate": 5.846356580258345e-06,
      "loss": 0.0,
      "step": 2258
    },
    {
      "epoch": 0.7323544303797468,
      "grad_norm": 0.0004988170694559813,
      "learning_rate": 5.840100833704392e-06,
      "loss": 0.0,
      "step": 2260
    },
    {
      "epoch": 0.7330025316455696,
      "grad_norm": 0.0006871929508633912,
      "learning_rate": 5.8338437335805124e-06,
      "loss": 0.0,
      "step": 2262
    },
    {
      "epoch": 0.7336506329113924,
      "grad_norm": 0.00016507611144334078,
      "learning_rate": 5.827585289968143e-06,
      "loss": 0.0,
      "step": 2264
    },
    {
      "epoch": 0.7342987341772151,
      "grad_norm": 8.90670926310122e-05,
      "learning_rate": 5.821325512950886e-06,
      "loss": 0.0,
      "step": 2266
    },
    {
      "epoch": 0.734946835443038,
      "grad_norm": 0.004662638995796442,
      "learning_rate": 5.815064412614487e-06,
      "loss": 0.0,
      "step": 2268
    },
    {
      "epoch": 0.7355949367088608,
      "grad_norm": 0.0012238593772053719,
      "learning_rate": 5.80880199904683e-06,
      "loss": 0.0,
      "step": 2270
    },
    {
      "epoch": 0.7362430379746835,
      "grad_norm": 0.000732690910808742,
      "learning_rate": 5.80253828233791e-06,
      "loss": 0.0,
      "step": 2272
    },
    {
      "epoch": 0.7368911392405063,
      "grad_norm": 0.05466936528682709,
      "learning_rate": 5.796273272579823e-06,
      "loss": 0.0001,
      "step": 2274
    },
    {
      "epoch": 0.7375392405063291,
      "grad_norm": 0.034032344818115234,
      "learning_rate": 5.79000697986675e-06,
      "loss": 0.0,
      "step": 2276
    },
    {
      "epoch": 0.738187341772152,
      "grad_norm": 0.0003485069319140166,
      "learning_rate": 5.783739414294938e-06,
      "loss": 0.0,
      "step": 2278
    },
    {
      "epoch": 0.7388354430379747,
      "grad_norm": 0.00011134272790513933,
      "learning_rate": 5.777470585962682e-06,
      "loss": 0.0,
      "step": 2280
    },
    {
      "epoch": 0.7388354430379747,
      "eval_accuracy": 0.9999988602777062,
      "eval_loss": 5.289937234920217e-06,
      "eval_runtime": 105.9113,
      "eval_samples_per_second": 47.209,
      "eval_steps_per_second": 11.802,
      "step": 2280
    },
    {
      "epoch": 0.7394835443037975,
      "grad_norm": 0.00734311668202281,
      "learning_rate": 5.771200504970316e-06,
      "loss": 0.0,
      "step": 2282
    },
    {
      "epoch": 0.7401316455696203,
      "grad_norm": 0.01817069947719574,
      "learning_rate": 5.764929181420191e-06,
      "loss": 0.0,
      "step": 2284
    },
    {
      "epoch": 0.740779746835443,
      "grad_norm": 0.001817168784327805,
      "learning_rate": 5.758656625416659e-06,
      "loss": 0.0,
      "step": 2286
    },
    {
      "epoch": 0.7414278481012658,
      "grad_norm": 0.017299983650445938,
      "learning_rate": 5.752382847066058e-06,
      "loss": 0.0,
      "step": 2288
    },
    {
      "epoch": 0.7420759493670886,
      "grad_norm": 0.0002641058526933193,
      "learning_rate": 5.7461078564766945e-06,
      "loss": 0.0,
      "step": 2290
    },
    {
      "epoch": 0.7427240506329114,
      "grad_norm": 0.01204910408705473,
      "learning_rate": 5.739831663758834e-06,
      "loss": 0.0,
      "step": 2292
    },
    {
      "epoch": 0.7433721518987342,
      "grad_norm": 0.00013036782911513,
      "learning_rate": 5.733554279024668e-06,
      "loss": 0.0,
      "step": 2294
    },
    {
      "epoch": 0.744020253164557,
      "grad_norm": 7.867599924793467e-05,
      "learning_rate": 5.727275712388318e-06,
      "loss": 0.0,
      "step": 2296
    },
    {
      "epoch": 0.7446683544303797,
      "grad_norm": 0.00017100705008488148,
      "learning_rate": 5.720995973965806e-06,
      "loss": 0.0,
      "step": 2298
    },
    {
      "epoch": 0.7453164556962025,
      "grad_norm": 0.07865839451551437,
      "learning_rate": 5.714715073875043e-06,
      "loss": 0.0,
      "step": 2300
    },
    {
      "epoch": 0.7459645569620253,
      "grad_norm": 0.000345155771356076,
      "learning_rate": 5.7084330222358106e-06,
      "loss": 0.0001,
      "step": 2302
    },
    {
      "epoch": 0.7466126582278481,
      "grad_norm": 0.0002492070198059082,
      "learning_rate": 5.7021498291697465e-06,
      "loss": 0.0,
      "step": 2304
    },
    {
      "epoch": 0.7472607594936709,
      "grad_norm": 5.335691821528599e-05,
      "learning_rate": 5.695865504800328e-06,
      "loss": 0.0,
      "step": 2306
    },
    {
      "epoch": 0.7479088607594937,
      "grad_norm": 9.89673935691826e-05,
      "learning_rate": 5.689580059252852e-06,
      "loss": 0.0001,
      "step": 2308
    },
    {
      "epoch": 0.7485569620253164,
      "grad_norm": 0.00012796987721230835,
      "learning_rate": 5.683293502654429e-06,
      "loss": 0.0,
      "step": 2310
    },
    {
      "epoch": 0.7492050632911392,
      "grad_norm": 0.00014111780910752714,
      "learning_rate": 5.6770058451339514e-06,
      "loss": 0.0,
      "step": 2312
    },
    {
      "epoch": 0.749853164556962,
      "grad_norm": 0.016285588964819908,
      "learning_rate": 5.6707170968220895e-06,
      "loss": 0.0,
      "step": 2314
    },
    {
      "epoch": 0.7505012658227848,
      "grad_norm": 0.0005048218881711364,
      "learning_rate": 5.664427267851271e-06,
      "loss": 0.0,
      "step": 2316
    },
    {
      "epoch": 0.7511493670886076,
      "grad_norm": 0.011887839995324612,
      "learning_rate": 5.658136368355665e-06,
      "loss": 0.0,
      "step": 2318
    },
    {
      "epoch": 0.7517974683544304,
      "grad_norm": 0.009634653106331825,
      "learning_rate": 5.651844408471162e-06,
      "loss": 0.0,
      "step": 2320
    },
    {
      "epoch": 0.7517974683544304,
      "eval_accuracy": 0.9999980149850695,
      "eval_loss": 7.766235285089351e-06,
      "eval_runtime": 107.1514,
      "eval_samples_per_second": 46.663,
      "eval_steps_per_second": 11.666,
      "step": 2320
    },
    {
      "epoch": 0.7524455696202532,
      "grad_norm": 0.006753602996468544,
      "learning_rate": 5.645551398335367e-06,
      "loss": 0.0,
      "step": 2322
    },
    {
      "epoch": 0.7530936708860759,
      "grad_norm": 8.452868496533483e-05,
      "learning_rate": 5.6392573480875724e-06,
      "loss": 0.0,
      "step": 2324
    },
    {
      "epoch": 0.7537417721518987,
      "grad_norm": 0.0007264796877279878,
      "learning_rate": 5.632962267868747e-06,
      "loss": 0.0,
      "step": 2326
    },
    {
      "epoch": 0.7543898734177216,
      "grad_norm": 0.00017660489538684487,
      "learning_rate": 5.626666167821522e-06,
      "loss": 0.0,
      "step": 2328
    },
    {
      "epoch": 0.7550379746835443,
      "grad_norm": 0.0025927009992301464,
      "learning_rate": 5.620369058090168e-06,
      "loss": 0.0,
      "step": 2330
    },
    {
      "epoch": 0.7556860759493671,
      "grad_norm": 0.00015297834761440754,
      "learning_rate": 5.6140709488205854e-06,
      "loss": 0.0,
      "step": 2332
    },
    {
      "epoch": 0.7563341772151899,
      "grad_norm": 0.010506555438041687,
      "learning_rate": 5.607771850160285e-06,
      "loss": 0.0,
      "step": 2334
    },
    {
      "epoch": 0.7569822784810126,
      "grad_norm": 0.00015023894957266748,
      "learning_rate": 5.601471772258368e-06,
      "loss": 0.0,
      "step": 2336
    },
    {
      "epoch": 0.7576303797468354,
      "grad_norm": 0.00028166870470158756,
      "learning_rate": 5.595170725265517e-06,
      "loss": 0.0,
      "step": 2338
    },
    {
      "epoch": 0.7582784810126583,
      "grad_norm": 0.007426573429256678,
      "learning_rate": 5.588868719333974e-06,
      "loss": 0.0002,
      "step": 2340
    },
    {
      "epoch": 0.758926582278481,
      "grad_norm": 0.0006746530998498201,
      "learning_rate": 5.582565764617528e-06,
      "loss": 0.0,
      "step": 2342
    },
    {
      "epoch": 0.7595746835443038,
      "grad_norm": 0.0007765135960653424,
      "learning_rate": 5.576261871271494e-06,
      "loss": 0.0,
      "step": 2344
    },
    {
      "epoch": 0.7602227848101266,
      "grad_norm": 0.10703450441360474,
      "learning_rate": 5.569957049452703e-06,
      "loss": 0.0,
      "step": 2346
    },
    {
      "epoch": 0.7608708860759493,
      "grad_norm": 0.0001581802061991766,
      "learning_rate": 5.56365130931948e-06,
      "loss": 0.0,
      "step": 2348
    },
    {
      "epoch": 0.7615189873417721,
      "grad_norm": 0.00017560912237968296,
      "learning_rate": 5.557344661031628e-06,
      "loss": 0.0,
      "step": 2350
    },
    {
      "epoch": 0.762167088607595,
      "grad_norm": 0.0002994161914102733,
      "learning_rate": 5.551037114750415e-06,
      "loss": 0.0,
      "step": 2352
    },
    {
      "epoch": 0.7628151898734177,
      "grad_norm": 0.0007976372726261616,
      "learning_rate": 5.544728680638557e-06,
      "loss": 0.0002,
      "step": 2354
    },
    {
      "epoch": 0.7634632911392405,
      "grad_norm": 0.003032875480130315,
      "learning_rate": 5.538419368860196e-06,
      "loss": 0.0,
      "step": 2356
    },
    {
      "epoch": 0.7641113924050633,
      "grad_norm": 0.05126134678721428,
      "learning_rate": 5.532109189580893e-06,
      "loss": 0.0,
      "step": 2358
    },
    {
      "epoch": 0.764759493670886,
      "grad_norm": 0.05150891840457916,
      "learning_rate": 5.525798152967605e-06,
      "loss": 0.0,
      "step": 2360
    },
    {
      "epoch": 0.764759493670886,
      "eval_accuracy": 0.9999976613323593,
      "eval_loss": 1.1560122402443085e-05,
      "eval_runtime": 106.4124,
      "eval_samples_per_second": 46.987,
      "eval_steps_per_second": 11.747,
      "step": 2360
    },
    {
      "epoch": 0.7654075949367088,
      "grad_norm": 0.00036254760925658047,
      "learning_rate": 5.519486269188669e-06,
      "loss": 0.0,
      "step": 2362
    },
    {
      "epoch": 0.7660556962025317,
      "grad_norm": 0.00011886703578056768,
      "learning_rate": 5.513173548413789e-06,
      "loss": 0.0,
      "step": 2364
    },
    {
      "epoch": 0.7667037974683544,
      "grad_norm": 7.184002606663853e-05,
      "learning_rate": 5.506860000814017e-06,
      "loss": 0.0,
      "step": 2366
    },
    {
      "epoch": 0.7673518987341772,
      "grad_norm": 0.00030205457005649805,
      "learning_rate": 5.500545636561737e-06,
      "loss": 0.0,
      "step": 2368
    },
    {
      "epoch": 0.768,
      "grad_norm": 0.002157788723707199,
      "learning_rate": 5.494230465830648e-06,
      "loss": 0.0,
      "step": 2370
    },
    {
      "epoch": 0.7686481012658228,
      "grad_norm": 0.006820139940828085,
      "learning_rate": 5.487914498795748e-06,
      "loss": 0.0,
      "step": 2372
    },
    {
      "epoch": 0.7692962025316455,
      "grad_norm": 0.0009759216918610036,
      "learning_rate": 5.4815977456333205e-06,
      "loss": 0.0,
      "step": 2374
    },
    {
      "epoch": 0.7699443037974684,
      "grad_norm": 0.00026600834098644555,
      "learning_rate": 5.475280216520913e-06,
      "loss": 0.0,
      "step": 2376
    },
    {
      "epoch": 0.7705924050632912,
      "grad_norm": 0.00012866378528997302,
      "learning_rate": 5.468961921637327e-06,
      "loss": 0.0,
      "step": 2378
    },
    {
      "epoch": 0.7712405063291139,
      "grad_norm": 0.00221284874714911,
      "learning_rate": 5.462642871162592e-06,
      "loss": 0.0001,
      "step": 2380
    },
    {
      "epoch": 0.7718886075949367,
      "grad_norm": 0.0002176949055865407,
      "learning_rate": 5.4563230752779595e-06,
      "loss": 0.0,
      "step": 2382
    },
    {
      "epoch": 0.7725367088607595,
      "grad_norm": 7.449064287357032e-05,
      "learning_rate": 5.450002544165881e-06,
      "loss": 0.0,
      "step": 2384
    },
    {
      "epoch": 0.7731848101265822,
      "grad_norm": 0.006494743749499321,
      "learning_rate": 5.443681288009991e-06,
      "loss": 0.0,
      "step": 2386
    },
    {
      "epoch": 0.7738329113924051,
      "grad_norm": 0.0006466159829869866,
      "learning_rate": 5.437359316995094e-06,
      "loss": 0.0,
      "step": 2388
    },
    {
      "epoch": 0.7744810126582279,
      "grad_norm": 0.00047616238589398563,
      "learning_rate": 5.431036641307146e-06,
      "loss": 0.0,
      "step": 2390
    },
    {
      "epoch": 0.7751291139240506,
      "grad_norm": 0.0010342790046706796,
      "learning_rate": 5.424713271133237e-06,
      "loss": 0.0,
      "step": 2392
    },
    {
      "epoch": 0.7757772151898734,
      "grad_norm": 0.0027879138942807913,
      "learning_rate": 5.41838921666158e-06,
      "loss": 0.0003,
      "step": 2394
    },
    {
      "epoch": 0.7764253164556962,
      "grad_norm": 0.06997310370206833,
      "learning_rate": 5.412064488081482e-06,
      "loss": 0.0,
      "step": 2396
    },
    {
      "epoch": 0.7770734177215189,
      "grad_norm": 0.0062122284434735775,
      "learning_rate": 5.4057390955833455e-06,
      "loss": 0.0,
      "step": 2398
    },
    {
      "epoch": 0.7777215189873418,
      "grad_norm": 0.32322806119918823,
      "learning_rate": 5.3994130493586385e-06,
      "loss": 0.0012,
      "step": 2400
    },
    {
      "epoch": 0.7777215189873418,
      "eval_accuracy": 0.9999614380181616,
      "eval_loss": 0.0002183700999012217,
      "eval_runtime": 105.1957,
      "eval_samples_per_second": 47.53,
      "eval_steps_per_second": 11.883,
      "step": 2400
    },
    {
      "epoch": 0.7783696202531646,
      "grad_norm": 0.24002808332443237,
      "learning_rate": 5.393086359599882e-06,
      "loss": 0.0003,
      "step": 2402
    },
    {
      "epoch": 0.7790177215189873,
      "grad_norm": 0.00038687174674123526,
      "learning_rate": 5.386759036500635e-06,
      "loss": 0.0005,
      "step": 2404
    },
    {
      "epoch": 0.7796658227848101,
      "grad_norm": 0.0652102380990982,
      "learning_rate": 5.380431090255475e-06,
      "loss": 0.0,
      "step": 2406
    },
    {
      "epoch": 0.7803139240506329,
      "grad_norm": 0.003415727987885475,
      "learning_rate": 5.3741025310599885e-06,
      "loss": 0.0,
      "step": 2408
    },
    {
      "epoch": 0.7809620253164556,
      "grad_norm": 0.028407571837306023,
      "learning_rate": 5.367773369110741e-06,
      "loss": 0.0,
      "step": 2410
    },
    {
      "epoch": 0.7816101265822785,
      "grad_norm": 0.00183353410102427,
      "learning_rate": 5.361443614605279e-06,
      "loss": 0.0,
      "step": 2412
    },
    {
      "epoch": 0.7822582278481013,
      "grad_norm": 0.02177795208990574,
      "learning_rate": 5.355113277742095e-06,
      "loss": 0.0,
      "step": 2414
    },
    {
      "epoch": 0.7829063291139241,
      "grad_norm": 0.005271558184176683,
      "learning_rate": 5.348782368720627e-06,
      "loss": 0.0,
      "step": 2416
    },
    {
      "epoch": 0.7835544303797468,
      "grad_norm": 0.0031913723796606064,
      "learning_rate": 5.3424508977412285e-06,
      "loss": 0.0001,
      "step": 2418
    },
    {
      "epoch": 0.7842025316455696,
      "grad_norm": 0.004424366634339094,
      "learning_rate": 5.336118875005165e-06,
      "loss": 0.0001,
      "step": 2420
    },
    {
      "epoch": 0.7848506329113925,
      "grad_norm": 0.09619469195604324,
      "learning_rate": 5.329786310714583e-06,
      "loss": 0.0,
      "step": 2422
    },
    {
      "epoch": 0.7854987341772152,
      "grad_norm": 0.15482549369335175,
      "learning_rate": 5.3234532150725096e-06,
      "loss": 0.0003,
      "step": 2424
    },
    {
      "epoch": 0.786146835443038,
      "grad_norm": 0.21717675030231476,
      "learning_rate": 5.317119598282823e-06,
      "loss": 0.0003,
      "step": 2426
    },
    {
      "epoch": 0.7867949367088608,
      "grad_norm": 0.027250362560153008,
      "learning_rate": 5.310785470550243e-06,
      "loss": 0.0001,
      "step": 2428
    },
    {
      "epoch": 0.7874430379746835,
      "grad_norm": 0.00278554018586874,
      "learning_rate": 5.304450842080312e-06,
      "loss": 0.0,
      "step": 2430
    },
    {
      "epoch": 0.7880911392405063,
      "grad_norm": 0.09197168052196503,
      "learning_rate": 5.29811572307938e-06,
      "loss": 0.0001,
      "step": 2432
    },
    {
      "epoch": 0.7887392405063292,
      "grad_norm": 0.0017266845097765326,
      "learning_rate": 5.291780123754585e-06,
      "loss": 0.0,
      "step": 2434
    },
    {
      "epoch": 0.7893873417721519,
      "grad_norm": 0.001447897288016975,
      "learning_rate": 5.285444054313841e-06,
      "loss": 0.0,
      "step": 2436
    },
    {
      "epoch": 0.7900354430379747,
      "grad_norm": 0.0027126609347760677,
      "learning_rate": 5.27910752496582e-06,
      "loss": 0.0002,
      "step": 2438
    },
    {
      "epoch": 0.7906835443037975,
      "grad_norm": 0.05010005086660385,
      "learning_rate": 5.2727705459199345e-06,
      "loss": 0.0,
      "step": 2440
    },
    {
      "epoch": 0.7906835443037975,
      "eval_accuracy": 0.9999877555061656,
      "eval_loss": 3.793625000980683e-05,
      "eval_runtime": 106.2945,
      "eval_samples_per_second": 47.039,
      "eval_steps_per_second": 11.76,
      "step": 2440
    },
    {
      "epoch": 0.7913316455696202,
      "grad_norm": 0.03448754549026489,
      "learning_rate": 5.266433127386319e-06,
      "loss": 0.0002,
      "step": 2442
    },
    {
      "epoch": 0.791979746835443,
      "grad_norm": 0.005970288068056107,
      "learning_rate": 5.260095279575818e-06,
      "loss": 0.0,
      "step": 2444
    },
    {
      "epoch": 0.7926278481012659,
      "grad_norm": 0.11865130066871643,
      "learning_rate": 5.253757012699972e-06,
      "loss": 0.0001,
      "step": 2446
    },
    {
      "epoch": 0.7932759493670886,
      "grad_norm": 0.10202252864837646,
      "learning_rate": 5.247418336970989e-06,
      "loss": 0.0002,
      "step": 2448
    },
    {
      "epoch": 0.7939240506329114,
      "grad_norm": 0.007061738055199385,
      "learning_rate": 5.241079262601738e-06,
      "loss": 0.0,
      "step": 2450
    },
    {
      "epoch": 0.7945721518987342,
      "grad_norm": 0.006444692611694336,
      "learning_rate": 5.234739799805735e-06,
      "loss": 0.0,
      "step": 2452
    },
    {
      "epoch": 0.7952202531645569,
      "grad_norm": 0.01060519739985466,
      "learning_rate": 5.228399958797117e-06,
      "loss": 0.0,
      "step": 2454
    },
    {
      "epoch": 0.7958683544303797,
      "grad_norm": 0.002685698214918375,
      "learning_rate": 5.2220597497906315e-06,
      "loss": 0.0,
      "step": 2456
    },
    {
      "epoch": 0.7965164556962026,
      "grad_norm": 0.0019370040390640497,
      "learning_rate": 5.215719183001619e-06,
      "loss": 0.0,
      "step": 2458
    },
    {
      "epoch": 0.7971645569620253,
      "grad_norm": 0.00024689623387530446,
      "learning_rate": 5.209378268645998e-06,
      "loss": 0.0002,
      "step": 2460
    },
    {
      "epoch": 0.7978126582278481,
      "grad_norm": 0.06425344198942184,
      "learning_rate": 5.203037016940245e-06,
      "loss": 0.0001,
      "step": 2462
    },
    {
      "epoch": 0.7984607594936709,
      "grad_norm": 0.00064616504823789,
      "learning_rate": 5.19669543810138e-06,
      "loss": 0.0,
      "step": 2464
    },
    {
      "epoch": 0.7991088607594937,
      "grad_norm": 0.020757561549544334,
      "learning_rate": 5.190353542346951e-06,
      "loss": 0.0,
      "step": 2466
    },
    {
      "epoch": 0.7997569620253164,
      "grad_norm": 0.0072888876311481,
      "learning_rate": 5.184011339895015e-06,
      "loss": 0.0,
      "step": 2468
    },
    {
      "epoch": 0.8004050632911393,
      "grad_norm": 0.037369176745414734,
      "learning_rate": 5.177668840964128e-06,
      "loss": 0.0001,
      "step": 2470
    },
    {
      "epoch": 0.8010531645569621,
      "grad_norm": 0.00951011385768652,
      "learning_rate": 5.171326055773318e-06,
      "loss": 0.0,
      "step": 2472
    },
    {
      "epoch": 0.8017012658227848,
      "grad_norm": 0.004345647059381008,
      "learning_rate": 5.164982994542076e-06,
      "loss": 0.0,
      "step": 2474
    },
    {
      "epoch": 0.8023493670886076,
      "grad_norm": 0.018001670017838478,
      "learning_rate": 5.15863966749034e-06,
      "loss": 0.0,
      "step": 2476
    },
    {
      "epoch": 0.8029974683544304,
      "grad_norm": 0.0005489176837727427,
      "learning_rate": 5.1522960848384715e-06,
      "loss": 0.0,
      "step": 2478
    },
    {
      "epoch": 0.8036455696202531,
      "grad_norm": 0.002194503555074334,
      "learning_rate": 5.1459522568072495e-06,
      "loss": 0.0,
      "step": 2480
    },
    {
      "epoch": 0.8036455696202531,
      "eval_accuracy": 0.9999988498303606,
      "eval_loss": 1.258536121895304e-05,
      "eval_runtime": 105.8044,
      "eval_samples_per_second": 47.257,
      "eval_steps_per_second": 11.814,
      "step": 2480
    },
    {
      "epoch": 0.804293670886076,
      "grad_norm": 0.0060137673281133175,
      "learning_rate": 5.139608193617846e-06,
      "loss": 0.0,
      "step": 2482
    },
    {
      "epoch": 0.8049417721518988,
      "grad_norm": 0.0109030706807971,
      "learning_rate": 5.133263905491809e-06,
      "loss": 0.0,
      "step": 2484
    },
    {
      "epoch": 0.8055898734177215,
      "grad_norm": 0.005852850619703531,
      "learning_rate": 5.126919402651053e-06,
      "loss": 0.0,
      "step": 2486
    },
    {
      "epoch": 0.8062379746835443,
      "grad_norm": 0.0023848526179790497,
      "learning_rate": 5.120574695317837e-06,
      "loss": 0.0,
      "step": 2488
    },
    {
      "epoch": 0.8068860759493671,
      "grad_norm": 0.0017674455884844065,
      "learning_rate": 5.114229793714749e-06,
      "loss": 0.0,
      "step": 2490
    },
    {
      "epoch": 0.8075341772151898,
      "grad_norm": 0.0013527893461287022,
      "learning_rate": 5.1078847080646894e-06,
      "loss": 0.0,
      "step": 2492
    },
    {
      "epoch": 0.8081822784810127,
      "grad_norm": 0.00026338460156694055,
      "learning_rate": 5.101539448590859e-06,
      "loss": 0.0,
      "step": 2494
    },
    {
      "epoch": 0.8088303797468355,
      "grad_norm": 0.031443241983652115,
      "learning_rate": 5.095194025516733e-06,
      "loss": 0.0001,
      "step": 2496
    },
    {
      "epoch": 0.8094784810126582,
      "grad_norm": 0.00025254758656956255,
      "learning_rate": 5.088848449066055e-06,
      "loss": 0.0,
      "step": 2498
    },
    {
      "epoch": 0.810126582278481,
      "grad_norm": 0.0013786570634692907,
      "learning_rate": 5.082502729462813e-06,
      "loss": 0.0,
      "step": 2500
    },
    {
      "epoch": 0.8107746835443038,
      "grad_norm": 0.0034877141006290913,
      "learning_rate": 5.076156876931225e-06,
      "loss": 0.0,
      "step": 2502
    },
    {
      "epoch": 0.8114227848101265,
      "grad_norm": 0.0005392807070165873,
      "learning_rate": 5.069810901695727e-06,
      "loss": 0.0,
      "step": 2504
    },
    {
      "epoch": 0.8120708860759493,
      "grad_norm": 0.008301528170704842,
      "learning_rate": 5.063464813980948e-06,
      "loss": 0.0,
      "step": 2506
    },
    {
      "epoch": 0.8127189873417722,
      "grad_norm": 0.0008546709432266653,
      "learning_rate": 5.057118624011702e-06,
      "loss": 0.0,
      "step": 2508
    },
    {
      "epoch": 0.813367088607595,
      "grad_norm": 0.0025733408983796835,
      "learning_rate": 5.050772342012966e-06,
      "loss": 0.0001,
      "step": 2510
    },
    {
      "epoch": 0.8140151898734177,
      "grad_norm": 0.11622051149606705,
      "learning_rate": 5.044425978209864e-06,
      "loss": 0.0,
      "step": 2512
    },
    {
      "epoch": 0.8146632911392405,
      "grad_norm": 0.009716551750898361,
      "learning_rate": 5.038079542827654e-06,
      "loss": 0.0,
      "step": 2514
    },
    {
      "epoch": 0.8153113924050633,
      "grad_norm": 0.00016989122377708554,
      "learning_rate": 5.03173304609171e-06,
      "loss": 0.0,
      "step": 2516
    },
    {
      "epoch": 0.815959493670886,
      "grad_norm": 0.005587367806583643,
      "learning_rate": 5.025386498227501e-06,
      "loss": 0.0,
      "step": 2518
    },
    {
      "epoch": 0.8166075949367089,
      "grad_norm": 0.00024819717509672046,
      "learning_rate": 5.019039909460584e-06,
      "loss": 0.0,
      "step": 2520
    },
    {
      "epoch": 0.8166075949367089,
      "eval_accuracy": 0.999998455353043,
      "eval_loss": 7.359149094554596e-06,
      "eval_runtime": 105.9092,
      "eval_samples_per_second": 47.21,
      "eval_steps_per_second": 11.803,
      "step": 2520
    },
    {
      "epoch": 0.8172556962025317,
      "grad_norm": 0.02661811001598835,
      "learning_rate": 5.012693290016576e-06,
      "loss": 0.0,
      "step": 2522
    },
    {
      "epoch": 0.8179037974683544,
      "grad_norm": 0.005714859813451767,
      "learning_rate": 5.006346650121148e-06,
      "loss": 0.0,
      "step": 2524
    },
    {
      "epoch": 0.8185518987341772,
      "grad_norm": 0.32837799191474915,
      "learning_rate": 5e-06,
      "loss": 0.0001,
      "step": 2526
    },
    {
      "epoch": 0.8192,
      "grad_norm": 0.19652798771858215,
      "learning_rate": 4.993653349878854e-06,
      "loss": 0.0001,
      "step": 2528
    },
    {
      "epoch": 0.8198481012658227,
      "grad_norm": 0.00043144775554537773,
      "learning_rate": 4.987306709983426e-06,
      "loss": 0.0,
      "step": 2530
    },
    {
      "epoch": 0.8204962025316456,
      "grad_norm": 0.0030640813056379557,
      "learning_rate": 4.980960090539417e-06,
      "loss": 0.0,
      "step": 2532
    },
    {
      "epoch": 0.8211443037974684,
      "grad_norm": 0.10829441994428635,
      "learning_rate": 4.9746135017725e-06,
      "loss": 0.0001,
      "step": 2534
    },
    {
      "epoch": 0.8217924050632911,
      "grad_norm": 0.039655912667512894,
      "learning_rate": 4.9682669539082914e-06,
      "loss": 0.0,
      "step": 2536
    },
    {
      "epoch": 0.8224405063291139,
      "grad_norm": 0.027690451592206955,
      "learning_rate": 4.961920457172347e-06,
      "loss": 0.0001,
      "step": 2538
    },
    {
      "epoch": 0.8230886075949367,
      "grad_norm": 0.03515208885073662,
      "learning_rate": 4.955574021790138e-06,
      "loss": 0.0,
      "step": 2540
    },
    {
      "epoch": 0.8237367088607594,
      "grad_norm": 0.10855144262313843,
      "learning_rate": 4.9492276579870355e-06,
      "loss": 0.0001,
      "step": 2542
    },
    {
      "epoch": 0.8243848101265823,
      "grad_norm": 0.004899628460407257,
      "learning_rate": 4.9428813759883e-06,
      "loss": 0.0001,
      "step": 2544
    },
    {
      "epoch": 0.8250329113924051,
      "grad_norm": 0.0006479221628978848,
      "learning_rate": 4.936535186019053e-06,
      "loss": 0.0,
      "step": 2546
    },
    {
      "epoch": 0.8256810126582278,
      "grad_norm": 0.0010982794919982553,
      "learning_rate": 4.9301890983042744e-06,
      "loss": 0.0,
      "step": 2548
    },
    {
      "epoch": 0.8263291139240506,
      "grad_norm": 0.0006811083876527846,
      "learning_rate": 4.923843123068776e-06,
      "loss": 0.0,
      "step": 2550
    },
    {
      "epoch": 0.8269772151898734,
      "grad_norm": 0.00013663475692737848,
      "learning_rate": 4.917497270537188e-06,
      "loss": 0.0001,
      "step": 2552
    },
    {
      "epoch": 0.8276253164556961,
      "grad_norm": 0.0004709422937594354,
      "learning_rate": 4.911151550933946e-06,
      "loss": 0.0,
      "step": 2554
    },
    {
      "epoch": 0.828273417721519,
      "grad_norm": 0.07884550839662552,
      "learning_rate": 4.904805974483267e-06,
      "loss": 0.0,
      "step": 2556
    },
    {
      "epoch": 0.8289215189873418,
      "grad_norm": 0.1472872942686081,
      "learning_rate": 4.898460551409141e-06,
      "loss": 0.0001,
      "step": 2558
    },
    {
      "epoch": 0.8295696202531646,
      "grad_norm": 0.007455206476151943,
      "learning_rate": 4.8921152919353105e-06,
      "loss": 0.0,
      "step": 2560
    },
    {
      "epoch": 0.8295696202531646,
      "eval_accuracy": 0.9999908549339463,
      "eval_loss": 3.50044465449173e-05,
      "eval_runtime": 104.6129,
      "eval_samples_per_second": 47.795,
      "eval_steps_per_second": 11.949,
      "step": 2560
    },
    {
      "epoch": 0.8302177215189873,
      "grad_norm": 0.008790325373411179,
      "learning_rate": 4.8857702062852515e-06,
      "loss": 0.0,
      "step": 2562
    },
    {
      "epoch": 0.8308658227848101,
      "grad_norm": 0.004961876664310694,
      "learning_rate": 4.879425304682164e-06,
      "loss": 0.0,
      "step": 2564
    },
    {
      "epoch": 0.831513924050633,
      "grad_norm": 0.02836470678448677,
      "learning_rate": 4.873080597348948e-06,
      "loss": 0.0,
      "step": 2566
    },
    {
      "epoch": 0.8321620253164557,
      "grad_norm": 0.009730171412229538,
      "learning_rate": 4.866736094508191e-06,
      "loss": 0.0004,
      "step": 2568
    },
    {
      "epoch": 0.8328101265822785,
      "grad_norm": 0.03526896610856056,
      "learning_rate": 4.860391806382157e-06,
      "loss": 0.0,
      "step": 2570
    },
    {
      "epoch": 0.8334582278481013,
      "grad_norm": 0.07535210996866226,
      "learning_rate": 4.854047743192752e-06,
      "loss": 0.0,
      "step": 2572
    },
    {
      "epoch": 0.834106329113924,
      "grad_norm": 0.0026582893915474415,
      "learning_rate": 4.847703915161531e-06,
      "loss": 0.0,
      "step": 2574
    },
    {
      "epoch": 0.8347544303797468,
      "grad_norm": 0.02271987311542034,
      "learning_rate": 4.841360332509663e-06,
      "loss": 0.0,
      "step": 2576
    },
    {
      "epoch": 0.8354025316455697,
      "grad_norm": 0.31956908106803894,
      "learning_rate": 4.835017005457926e-06,
      "loss": 0.0001,
      "step": 2578
    },
    {
      "epoch": 0.8360506329113924,
      "grad_norm": 0.05217691510915756,
      "learning_rate": 4.828673944226684e-06,
      "loss": 0.0002,
      "step": 2580
    },
    {
      "epoch": 0.8366987341772152,
      "grad_norm": 0.07773158699274063,
      "learning_rate": 4.822331159035873e-06,
      "loss": 0.0,
      "step": 2582
    },
    {
      "epoch": 0.837346835443038,
      "grad_norm": 0.0017431579763069749,
      "learning_rate": 4.815988660104986e-06,
      "loss": 0.0001,
      "step": 2584
    },
    {
      "epoch": 0.8379949367088607,
      "grad_norm": 0.0009354711510241032,
      "learning_rate": 4.809646457653051e-06,
      "loss": 0.0001,
      "step": 2586
    },
    {
      "epoch": 0.8386430379746835,
      "grad_norm": 0.00018994830315932631,
      "learning_rate": 4.803304561898622e-06,
      "loss": 0.0,
      "step": 2588
    },
    {
      "epoch": 0.8392911392405064,
      "grad_norm": 0.0062407576479017735,
      "learning_rate": 4.796962983059757e-06,
      "loss": 0.0,
      "step": 2590
    },
    {
      "epoch": 0.8399392405063291,
      "grad_norm": 0.0004235352680552751,
      "learning_rate": 4.7906217313540035e-06,
      "loss": 0.0,
      "step": 2592
    },
    {
      "epoch": 0.8405873417721519,
      "grad_norm": 0.2927955687046051,
      "learning_rate": 4.784280816998382e-06,
      "loss": 0.0,
      "step": 2594
    },
    {
      "epoch": 0.8412354430379747,
      "grad_norm": 0.00975685752928257,
      "learning_rate": 4.777940250209369e-06,
      "loss": 0.0,
      "step": 2596
    },
    {
      "epoch": 0.8418835443037974,
      "grad_norm": 0.06928249448537827,
      "learning_rate": 4.771600041202884e-06,
      "loss": 0.0,
      "step": 2598
    },
    {
      "epoch": 0.8425316455696202,
      "grad_norm": 0.006853208411484957,
      "learning_rate": 4.765260200194266e-06,
      "loss": 0.0,
      "step": 2600
    },
    {
      "epoch": 0.8425316455696202,
      "eval_accuracy": 0.9999832612866548,
      "eval_loss": 9.795609366847202e-05,
      "eval_runtime": 105.1244,
      "eval_samples_per_second": 47.563,
      "eval_steps_per_second": 11.891,
      "step": 2600
    },
    {
      "epoch": 0.8431797468354431,
      "grad_norm": 0.023050669580698013,
      "learning_rate": 4.7589207373982635e-06,
      "loss": 0.0003,
      "step": 2602
    },
    {
      "epoch": 0.8438278481012659,
      "grad_norm": 0.0012926086783409119,
      "learning_rate": 4.7525816630290135e-06,
      "loss": 0.0,
      "step": 2604
    },
    {
      "epoch": 0.8444759493670886,
      "grad_norm": 0.0985993891954422,
      "learning_rate": 4.74624298730003e-06,
      "loss": 0.0001,
      "step": 2606
    },
    {
      "epoch": 0.8451240506329114,
      "grad_norm": 0.047884199768304825,
      "learning_rate": 4.7399047204241826e-06,
      "loss": 0.0,
      "step": 2608
    },
    {
      "epoch": 0.8457721518987342,
      "grad_norm": 0.044315554201602936,
      "learning_rate": 4.733566872613683e-06,
      "loss": 0.0,
      "step": 2610
    },
    {
      "epoch": 0.8464202531645569,
      "grad_norm": 0.28911229968070984,
      "learning_rate": 4.727229454080068e-06,
      "loss": 0.0001,
      "step": 2612
    },
    {
      "epoch": 0.8470683544303798,
      "grad_norm": 0.001446192734874785,
      "learning_rate": 4.720892475034181e-06,
      "loss": 0.0,
      "step": 2614
    },
    {
      "epoch": 0.8477164556962026,
      "grad_norm": 0.0012995201395824552,
      "learning_rate": 4.71455594568616e-06,
      "loss": 0.0,
      "step": 2616
    },
    {
      "epoch": 0.8483645569620253,
      "grad_norm": 0.03472326695919037,
      "learning_rate": 4.7082198762454165e-06,
      "loss": 0.0,
      "step": 2618
    },
    {
      "epoch": 0.8490126582278481,
      "grad_norm": 0.03457796573638916,
      "learning_rate": 4.701884276920622e-06,
      "loss": 0.0,
      "step": 2620
    },
    {
      "epoch": 0.8496607594936709,
      "grad_norm": 0.0026486103888601065,
      "learning_rate": 4.69554915791969e-06,
      "loss": 0.0002,
      "step": 2622
    },
    {
      "epoch": 0.8503088607594936,
      "grad_norm": 0.002102876780554652,
      "learning_rate": 4.689214529449758e-06,
      "loss": 0.0,
      "step": 2624
    },
    {
      "epoch": 0.8509569620253165,
      "grad_norm": 0.0034834700636565685,
      "learning_rate": 4.682880401717178e-06,
      "loss": 0.0,
      "step": 2626
    },
    {
      "epoch": 0.8516050632911393,
      "grad_norm": 0.06586204469203949,
      "learning_rate": 4.676546784927491e-06,
      "loss": 0.0002,
      "step": 2628
    },
    {
      "epoch": 0.852253164556962,
      "grad_norm": 0.0015004536835476756,
      "learning_rate": 4.670213689285418e-06,
      "loss": 0.0,
      "step": 2630
    },
    {
      "epoch": 0.8529012658227848,
      "grad_norm": 0.14395491778850555,
      "learning_rate": 4.663881124994837e-06,
      "loss": 0.0003,
      "step": 2632
    },
    {
      "epoch": 0.8535493670886076,
      "grad_norm": 0.024621322751045227,
      "learning_rate": 4.6575491022587714e-06,
      "loss": 0.0,
      "step": 2634
    },
    {
      "epoch": 0.8541974683544303,
      "grad_norm": 0.002342685591429472,
      "learning_rate": 4.651217631279374e-06,
      "loss": 0.0,
      "step": 2636
    },
    {
      "epoch": 0.8548455696202532,
      "grad_norm": 0.0156905185431242,
      "learning_rate": 4.644886722257905e-06,
      "loss": 0.0,
      "step": 2638
    },
    {
      "epoch": 0.855493670886076,
      "grad_norm": 0.0143441678956151,
      "learning_rate": 4.638556385394721e-06,
      "loss": 0.0,
      "step": 2640
    },
    {
      "epoch": 0.855493670886076,
      "eval_accuracy": 0.9999904080133325,
      "eval_loss": 4.1821800550678745e-05,
      "eval_runtime": 106.2091,
      "eval_samples_per_second": 47.077,
      "eval_steps_per_second": 11.769,
      "step": 2640
    },
    {
      "epoch": 0.8561417721518987,
      "grad_norm": 0.25288715958595276,
      "learning_rate": 4.632226630889258e-06,
      "loss": 0.0004,
      "step": 2642
    },
    {
      "epoch": 0.8567898734177215,
      "grad_norm": 0.010879195295274258,
      "learning_rate": 4.625897468940012e-06,
      "loss": 0.0,
      "step": 2644
    },
    {
      "epoch": 0.8574379746835443,
      "grad_norm": 0.008445970714092255,
      "learning_rate": 4.619568909744524e-06,
      "loss": 0.0,
      "step": 2646
    },
    {
      "epoch": 0.858086075949367,
      "grad_norm": 0.001091435202397406,
      "learning_rate": 4.6132409634993655e-06,
      "loss": 0.0,
      "step": 2648
    },
    {
      "epoch": 0.8587341772151899,
      "grad_norm": 0.0010871505364775658,
      "learning_rate": 4.606913640400118e-06,
      "loss": 0.0,
      "step": 2650
    },
    {
      "epoch": 0.8593822784810127,
      "grad_norm": 0.0015741437673568726,
      "learning_rate": 4.600586950641362e-06,
      "loss": 0.0,
      "step": 2652
    },
    {
      "epoch": 0.8600303797468355,
      "grad_norm": 0.0032445641700178385,
      "learning_rate": 4.594260904416656e-06,
      "loss": 0.0,
      "step": 2654
    },
    {
      "epoch": 0.8606784810126582,
      "grad_norm": 0.03352479264140129,
      "learning_rate": 4.587935511918521e-06,
      "loss": 0.0,
      "step": 2656
    },
    {
      "epoch": 0.861326582278481,
      "grad_norm": 0.0031426376663148403,
      "learning_rate": 4.581610783338424e-06,
      "loss": 0.0,
      "step": 2658
    },
    {
      "epoch": 0.8619746835443038,
      "grad_norm": 0.0009672287269495428,
      "learning_rate": 4.575286728866765e-06,
      "loss": 0.0,
      "step": 2660
    },
    {
      "epoch": 0.8626227848101266,
      "grad_norm": 0.022923622280359268,
      "learning_rate": 4.568963358692856e-06,
      "loss": 0.0,
      "step": 2662
    },
    {
      "epoch": 0.8632708860759494,
      "grad_norm": 0.04482516646385193,
      "learning_rate": 4.562640683004907e-06,
      "loss": 0.0,
      "step": 2664
    },
    {
      "epoch": 0.8639189873417722,
      "grad_norm": 0.0032297135330736637,
      "learning_rate": 4.55631871199001e-06,
      "loss": 0.0,
      "step": 2666
    },
    {
      "epoch": 0.8645670886075949,
      "grad_norm": 0.0007287651533260942,
      "learning_rate": 4.549997455834121e-06,
      "loss": 0.0,
      "step": 2668
    },
    {
      "epoch": 0.8652151898734177,
      "grad_norm": 0.00027854496147483587,
      "learning_rate": 4.543676924722042e-06,
      "loss": 0.0001,
      "step": 2670
    },
    {
      "epoch": 0.8658632911392405,
      "grad_norm": 0.0007697210530750453,
      "learning_rate": 4.53735712883741e-06,
      "loss": 0.0,
      "step": 2672
    },
    {
      "epoch": 0.8665113924050633,
      "grad_norm": 0.12043469399213791,
      "learning_rate": 4.531038078362675e-06,
      "loss": 0.0001,
      "step": 2674
    },
    {
      "epoch": 0.8671594936708861,
      "grad_norm": 0.0005098318797536194,
      "learning_rate": 4.524719783479088e-06,
      "loss": 0.0,
      "step": 2676
    },
    {
      "epoch": 0.8678075949367089,
      "grad_norm": 0.0008214846020564437,
      "learning_rate": 4.518402254366681e-06,
      "loss": 0.0,
      "step": 2678
    },
    {
      "epoch": 0.8684556962025316,
      "grad_norm": 0.0033130354713648558,
      "learning_rate": 4.512085501204254e-06,
      "loss": 0.0002,
      "step": 2680
    },
    {
      "epoch": 0.8684556962025316,
      "eval_accuracy": 0.9999965564468379,
      "eval_loss": 1.2069258445990272e-05,
      "eval_runtime": 106.3063,
      "eval_samples_per_second": 47.034,
      "eval_steps_per_second": 11.758,
      "step": 2680
    },
    {
      "epoch": 0.8691037974683544,
      "grad_norm": 0.0812884122133255,
      "learning_rate": 4.505769534169354e-06,
      "loss": 0.0005,
      "step": 2682
    },
    {
      "epoch": 0.8697518987341772,
      "grad_norm": 0.001252077054232359,
      "learning_rate": 4.499454363438265e-06,
      "loss": 0.0,
      "step": 2684
    },
    {
      "epoch": 0.8704,
      "grad_norm": 0.0012147113448008895,
      "learning_rate": 4.493139999185984e-06,
      "loss": 0.0,
      "step": 2686
    },
    {
      "epoch": 0.8710481012658228,
      "grad_norm": 0.00036030332557857037,
      "learning_rate": 4.4868264515862115e-06,
      "loss": 0.0,
      "step": 2688
    },
    {
      "epoch": 0.8716962025316456,
      "grad_norm": 0.0011000720551237464,
      "learning_rate": 4.480513730811332e-06,
      "loss": 0.0,
      "step": 2690
    },
    {
      "epoch": 0.8723443037974683,
      "grad_norm": 0.002217898378148675,
      "learning_rate": 4.474201847032396e-06,
      "loss": 0.0,
      "step": 2692
    },
    {
      "epoch": 0.8729924050632911,
      "grad_norm": 0.05367765948176384,
      "learning_rate": 4.467890810419108e-06,
      "loss": 0.0,
      "step": 2694
    },
    {
      "epoch": 0.8736405063291139,
      "grad_norm": 0.013308125548064709,
      "learning_rate": 4.461580631139806e-06,
      "loss": 0.0,
      "step": 2696
    },
    {
      "epoch": 0.8742886075949368,
      "grad_norm": 0.08926782011985779,
      "learning_rate": 4.455271319361445e-06,
      "loss": 0.0001,
      "step": 2698
    },
    {
      "epoch": 0.8749367088607595,
      "grad_norm": 0.0002369306021137163,
      "learning_rate": 4.448962885249587e-06,
      "loss": 0.0001,
      "step": 2700
    },
    {
      "epoch": 0.8755848101265823,
      "grad_norm": 0.009098300710320473,
      "learning_rate": 4.442655338968373e-06,
      "loss": 0.0,
      "step": 2702
    },
    {
      "epoch": 0.8762329113924051,
      "grad_norm": 0.13955150544643402,
      "learning_rate": 4.436348690680521e-06,
      "loss": 0.0003,
      "step": 2704
    },
    {
      "epoch": 0.8768810126582278,
      "grad_norm": 0.005642596632242203,
      "learning_rate": 4.430042950547298e-06,
      "loss": 0.0,
      "step": 2706
    },
    {
      "epoch": 0.8775291139240506,
      "grad_norm": 0.020765792578458786,
      "learning_rate": 4.423738128728507e-06,
      "loss": 0.0,
      "step": 2708
    },
    {
      "epoch": 0.8781772151898735,
      "grad_norm": 0.0027800104580819607,
      "learning_rate": 4.417434235382474e-06,
      "loss": 0.0,
      "step": 2710
    },
    {
      "epoch": 0.8788253164556962,
      "grad_norm": 0.007515923120081425,
      "learning_rate": 4.411131280666028e-06,
      "loss": 0.0,
      "step": 2712
    },
    {
      "epoch": 0.879473417721519,
      "grad_norm": 0.011895989999175072,
      "learning_rate": 4.404829274734485e-06,
      "loss": 0.0,
      "step": 2714
    },
    {
      "epoch": 0.8801215189873418,
      "grad_norm": 0.002676513744518161,
      "learning_rate": 4.398528227741634e-06,
      "loss": 0.0,
      "step": 2716
    },
    {
      "epoch": 0.8807696202531645,
      "grad_norm": 0.024730712175369263,
      "learning_rate": 4.392228149839716e-06,
      "loss": 0.0,
      "step": 2718
    },
    {
      "epoch": 0.8814177215189873,
      "grad_norm": 0.006555337458848953,
      "learning_rate": 4.3859290511794145e-06,
      "loss": 0.0,
      "step": 2720
    },
    {
      "epoch": 0.8814177215189873,
      "eval_accuracy": 0.9999992141332499,
      "eval_loss": 6.6279590100748464e-06,
      "eval_runtime": 105.6669,
      "eval_samples_per_second": 47.318,
      "eval_steps_per_second": 11.83,
      "step": 2720
    },
    {
      "epoch": 0.8820658227848102,
      "grad_norm": 0.0008663420449011028,
      "learning_rate": 4.379630941909832e-06,
      "loss": 0.0,
      "step": 2722
    },
    {
      "epoch": 0.8827139240506329,
      "grad_norm": 0.00045760456123389304,
      "learning_rate": 4.373333832178478e-06,
      "loss": 0.0,
      "step": 2724
    },
    {
      "epoch": 0.8833620253164557,
      "grad_norm": 0.0008591710939072073,
      "learning_rate": 4.367037732131254e-06,
      "loss": 0.0,
      "step": 2726
    },
    {
      "epoch": 0.8840101265822785,
      "grad_norm": 0.005888012237846851,
      "learning_rate": 4.360742651912428e-06,
      "loss": 0.0,
      "step": 2728
    },
    {
      "epoch": 0.8846582278481012,
      "grad_norm": 0.00036615473800338805,
      "learning_rate": 4.3544486016646335e-06,
      "loss": 0.0,
      "step": 2730
    },
    {
      "epoch": 0.885306329113924,
      "grad_norm": 0.00028013926930725574,
      "learning_rate": 4.34815559152884e-06,
      "loss": 0.0,
      "step": 2732
    },
    {
      "epoch": 0.8859544303797469,
      "grad_norm": 0.004798305220901966,
      "learning_rate": 4.341863631644337e-06,
      "loss": 0.0,
      "step": 2734
    },
    {
      "epoch": 0.8866025316455696,
      "grad_norm": 0.20881663262844086,
      "learning_rate": 4.33557273214873e-06,
      "loss": 0.0,
      "step": 2736
    },
    {
      "epoch": 0.8872506329113924,
      "grad_norm": 0.000936617492698133,
      "learning_rate": 4.329282903177911e-06,
      "loss": 0.0,
      "step": 2738
    },
    {
      "epoch": 0.8878987341772152,
      "grad_norm": 0.00045031809713691473,
      "learning_rate": 4.32299415486605e-06,
      "loss": 0.0,
      "step": 2740
    },
    {
      "epoch": 0.888546835443038,
      "grad_norm": 0.0002908929600380361,
      "learning_rate": 4.316706497345572e-06,
      "loss": 0.0002,
      "step": 2742
    },
    {
      "epoch": 0.8891949367088607,
      "grad_norm": 0.0002155246038455516,
      "learning_rate": 4.3104199407471485e-06,
      "loss": 0.0,
      "step": 2744
    },
    {
      "epoch": 0.8898430379746836,
      "grad_norm": 0.0013598958030343056,
      "learning_rate": 4.304134495199675e-06,
      "loss": 0.0,
      "step": 2746
    },
    {
      "epoch": 0.8904911392405064,
      "grad_norm": 0.012999402359127998,
      "learning_rate": 4.297850170830255e-06,
      "loss": 0.0,
      "step": 2748
    },
    {
      "epoch": 0.8911392405063291,
      "grad_norm": 0.0005251656984910369,
      "learning_rate": 4.291566977764192e-06,
      "loss": 0.0,
      "step": 2750
    },
    {
      "epoch": 0.8917873417721519,
      "grad_norm": 0.006405015476047993,
      "learning_rate": 4.285284926124959e-06,
      "loss": 0.0,
      "step": 2752
    },
    {
      "epoch": 0.8924354430379747,
      "grad_norm": 0.046361591666936874,
      "learning_rate": 4.279004026034196e-06,
      "loss": 0.0001,
      "step": 2754
    },
    {
      "epoch": 0.8930835443037974,
      "grad_norm": 0.004697976168245077,
      "learning_rate": 4.272724287611684e-06,
      "loss": 0.0001,
      "step": 2756
    },
    {
      "epoch": 0.8937316455696203,
      "grad_norm": 0.0007214845390990376,
      "learning_rate": 4.266445720975334e-06,
      "loss": 0.0,
      "step": 2758
    },
    {
      "epoch": 0.8943797468354431,
      "grad_norm": 0.0014260972384363413,
      "learning_rate": 4.260168336241169e-06,
      "loss": 0.0,
      "step": 2760
    },
    {
      "epoch": 0.8943797468354431,
      "eval_accuracy": 0.9999988370301384,
      "eval_loss": 7.5536236181505956e-06,
      "eval_runtime": 106.0073,
      "eval_samples_per_second": 47.167,
      "eval_steps_per_second": 11.792,
      "step": 2760
    },
    {
      "epoch": 0.8950278481012658,
      "grad_norm": 0.0002829917357303202,
      "learning_rate": 4.253892143523306e-06,
      "loss": 0.0,
      "step": 2762
    },
    {
      "epoch": 0.8956759493670886,
      "grad_norm": 0.0004675632808357477,
      "learning_rate": 4.247617152933944e-06,
      "loss": 0.0,
      "step": 2764
    },
    {
      "epoch": 0.8963240506329114,
      "grad_norm": 0.004662278573960066,
      "learning_rate": 4.241343374583343e-06,
      "loss": 0.0,
      "step": 2766
    },
    {
      "epoch": 0.8969721518987341,
      "grad_norm": 0.003993092570453882,
      "learning_rate": 4.23507081857981e-06,
      "loss": 0.0,
      "step": 2768
    },
    {
      "epoch": 0.897620253164557,
      "grad_norm": 0.0005434814374893904,
      "learning_rate": 4.228799495029685e-06,
      "loss": 0.0,
      "step": 2770
    },
    {
      "epoch": 0.8982683544303798,
      "grad_norm": 0.002016532002016902,
      "learning_rate": 4.222529414037319e-06,
      "loss": 0.0,
      "step": 2772
    },
    {
      "epoch": 0.8989164556962025,
      "grad_norm": 0.0012540030293166637,
      "learning_rate": 4.2162605857050645e-06,
      "loss": 0.0,
      "step": 2774
    },
    {
      "epoch": 0.8995645569620253,
      "grad_norm": 0.0007384781492874026,
      "learning_rate": 4.209993020133251e-06,
      "loss": 0.0,
      "step": 2776
    },
    {
      "epoch": 0.9002126582278481,
      "grad_norm": 0.000652509625069797,
      "learning_rate": 4.2037267274201785e-06,
      "loss": 0.0,
      "step": 2778
    },
    {
      "epoch": 0.9008607594936708,
      "grad_norm": 0.0025625203270465136,
      "learning_rate": 4.197461717662092e-06,
      "loss": 0.0,
      "step": 2780
    },
    {
      "epoch": 0.9015088607594937,
      "grad_norm": 0.11876733601093292,
      "learning_rate": 4.191198000953171e-06,
      "loss": 0.0,
      "step": 2782
    },
    {
      "epoch": 0.9021569620253165,
      "grad_norm": 0.0005562319420278072,
      "learning_rate": 4.1849355873855135e-06,
      "loss": 0.0,
      "step": 2784
    },
    {
      "epoch": 0.9028050632911392,
      "grad_norm": 0.00024305633269250393,
      "learning_rate": 4.178674487049116e-06,
      "loss": 0.0,
      "step": 2786
    },
    {
      "epoch": 0.903453164556962,
      "grad_norm": 0.0007249072659760714,
      "learning_rate": 4.172414710031858e-06,
      "loss": 0.0,
      "step": 2788
    },
    {
      "epoch": 0.9041012658227848,
      "grad_norm": 0.0016171886818483472,
      "learning_rate": 4.166156266419489e-06,
      "loss": 0.0,
      "step": 2790
    },
    {
      "epoch": 0.9047493670886076,
      "grad_norm": 0.0005725118098780513,
      "learning_rate": 4.15989916629561e-06,
      "loss": 0.0,
      "step": 2792
    },
    {
      "epoch": 0.9053974683544304,
      "grad_norm": 0.0003646530385594815,
      "learning_rate": 4.153643419741656e-06,
      "loss": 0.0,
      "step": 2794
    },
    {
      "epoch": 0.9060455696202532,
      "grad_norm": 0.06287111341953278,
      "learning_rate": 4.147389036836881e-06,
      "loss": 0.0,
      "step": 2796
    },
    {
      "epoch": 0.906693670886076,
      "grad_norm": 0.0005466180737130344,
      "learning_rate": 4.141136027658345e-06,
      "loss": 0.0,
      "step": 2798
    },
    {
      "epoch": 0.9073417721518987,
      "grad_norm": 0.008458413183689117,
      "learning_rate": 4.13488440228089e-06,
      "loss": 0.0,
      "step": 2800
    },
    {
      "epoch": 0.9073417721518987,
      "eval_accuracy": 0.9999996161228406,
      "eval_loss": 4.550272933556698e-06,
      "eval_runtime": 106.6358,
      "eval_samples_per_second": 46.889,
      "eval_steps_per_second": 11.722,
      "step": 2800
    }
  ],
  "logging_steps": 2,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 40,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.93561237640446e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}