{
  "best_metric": 0.3534523546695709,
  "best_model_checkpoint": "../results/with_lora_mistral_4bit_enbedding_trainable_unfiltered_dataset_inst_frame/checkpoint-2150",
  "epoch": 0.9982763778293756,
  "eval_steps": 10,
  "global_step": 2190,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 23.42947006225586,
      "learning_rate": 9.995440036479709e-05,
      "loss": 0.7291,
      "step": 1
    },
    {
      "epoch": 0.0,
      "grad_norm": 8.535998344421387,
      "learning_rate": 9.990880072959417e-05,
      "loss": 0.6489,
      "step": 2
    },
    {
      "epoch": 0.0,
      "grad_norm": 5.2448015213012695,
      "learning_rate": 9.986320109439125e-05,
      "loss": 0.5896,
      "step": 3
    },
    {
      "epoch": 0.0,
      "grad_norm": 6.437242031097412,
      "learning_rate": 9.981760145918833e-05,
      "loss": 0.5657,
      "step": 4
    },
    {
      "epoch": 0.0,
      "grad_norm": 5.361367225646973,
      "learning_rate": 9.977200182398541e-05,
      "loss": 0.5484,
      "step": 5
    },
    {
      "epoch": 0.0,
      "grad_norm": 5.016327857971191,
      "learning_rate": 9.97264021887825e-05,
      "loss": 0.5382,
      "step": 6
    },
    {
      "epoch": 0.0,
      "grad_norm": 26.851099014282227,
      "learning_rate": 9.968080255357958e-05,
      "loss": 0.5216,
      "step": 7
    },
    {
      "epoch": 0.0,
      "grad_norm": 6.51596736907959,
      "learning_rate": 9.963520291837666e-05,
      "loss": 0.51,
      "step": 8
    },
    {
      "epoch": 0.0,
      "grad_norm": 4.276856422424316,
      "learning_rate": 9.958960328317374e-05,
      "loss": 0.4838,
      "step": 9
    },
    {
      "epoch": 0.0,
      "grad_norm": 3.621899366378784,
      "learning_rate": 9.954400364797082e-05,
      "loss": 0.5401,
      "step": 10
    },
    {
      "epoch": 0.0,
      "eval_loss": 0.49539637565612793,
      "eval_runtime": 19.4376,
      "eval_samples_per_second": 1.441,
      "eval_steps_per_second": 0.36,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 3.671776294708252,
      "learning_rate": 9.94984040127679e-05,
      "loss": 0.4841,
      "step": 11
    },
    {
      "epoch": 0.01,
      "grad_norm": 3.1968910694122314,
      "learning_rate": 9.945280437756497e-05,
      "loss": 0.454,
      "step": 12
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.9277799129486084,
      "learning_rate": 9.940720474236207e-05,
      "loss": 0.481,
      "step": 13
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.9978115558624268,
      "learning_rate": 9.936160510715915e-05,
      "loss": 0.4935,
      "step": 14
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.7962915897369385,
      "learning_rate": 9.931600547195623e-05,
      "loss": 0.4444,
      "step": 15
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8223544359207153,
      "learning_rate": 9.927040583675331e-05,
      "loss": 0.4606,
      "step": 16
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.258604049682617,
      "learning_rate": 9.92248062015504e-05,
      "loss": 0.4592,
      "step": 17
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8389122486114502,
      "learning_rate": 9.917920656634748e-05,
      "loss": 0.4398,
      "step": 18
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8433682918548584,
      "learning_rate": 9.913360693114456e-05,
      "loss": 0.4411,
      "step": 19
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.7910923957824707,
      "learning_rate": 9.908800729594163e-05,
      "loss": 0.4324,
      "step": 20
    },
    {
      "epoch": 0.01,
      "eval_loss": 0.4595804512500763,
      "eval_runtime": 19.8191,
      "eval_samples_per_second": 1.413,
      "eval_steps_per_second": 0.353,
      "step": 20
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8877395391464233,
      "learning_rate": 9.904240766073872e-05,
      "loss": 0.4136,
      "step": 21
    },
    {
      "epoch": 0.01,
      "grad_norm": 11.494500160217285,
      "learning_rate": 9.89968080255358e-05,
      "loss": 0.4511,
      "step": 22
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.1572749614715576,
      "learning_rate": 9.895120839033288e-05,
      "loss": 0.4759,
      "step": 23
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.031266212463379,
      "learning_rate": 9.890560875512997e-05,
      "loss": 0.4447,
      "step": 24
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8705304861068726,
      "learning_rate": 9.886000911992705e-05,
      "loss": 0.4589,
      "step": 25
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.6991623640060425,
      "learning_rate": 9.881440948472413e-05,
      "loss": 0.4208,
      "step": 26
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.8854478597640991,
      "learning_rate": 9.87688098495212e-05,
      "loss": 0.43,
      "step": 27
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.826858639717102,
      "learning_rate": 9.872321021431828e-05,
      "loss": 0.4558,
      "step": 28
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.6870789527893066,
      "learning_rate": 9.867761057911538e-05,
      "loss": 0.4169,
      "step": 29
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.4985276460647583,
      "learning_rate": 9.863201094391246e-05,
      "loss": 0.4187,
      "step": 30
    },
    {
      "epoch": 0.01,
      "eval_loss": 0.44730886816978455,
      "eval_runtime": 19.0823,
      "eval_samples_per_second": 1.467,
      "eval_steps_per_second": 0.367,
      "step": 30
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.6747127771377563,
      "learning_rate": 9.858641130870954e-05,
      "loss": 0.4232,
      "step": 31
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.5604517459869385,
      "learning_rate": 9.854081167350662e-05,
      "loss": 0.4272,
      "step": 32
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.685736060142517,
      "learning_rate": 9.84952120383037e-05,
      "loss": 0.3962,
      "step": 33
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.6975411176681519,
      "learning_rate": 9.844961240310078e-05,
      "loss": 0.4327,
      "step": 34
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.6804540157318115,
      "learning_rate": 9.840401276789785e-05,
      "loss": 0.4347,
      "step": 35
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.6468958854675293,
      "learning_rate": 9.835841313269493e-05,
      "loss": 0.4327,
      "step": 36
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5674083232879639,
      "learning_rate": 9.831281349749203e-05,
      "loss": 0.4195,
      "step": 37
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5349141359329224,
      "learning_rate": 9.826721386228911e-05,
      "loss": 0.4214,
      "step": 38
    },
    {
      "epoch": 0.02,
      "grad_norm": 2.2588911056518555,
      "learning_rate": 9.822161422708619e-05,
      "loss": 0.4422,
      "step": 39
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5731521844863892,
      "learning_rate": 9.817601459188327e-05,
      "loss": 0.4356,
      "step": 40
    },
    {
      "epoch": 0.02,
      "eval_loss": 0.43755975365638733,
      "eval_runtime": 18.8787,
      "eval_samples_per_second": 1.483,
      "eval_steps_per_second": 0.371,
      "step": 40
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.470995306968689,
      "learning_rate": 9.813041495668036e-05,
      "loss": 0.4014,
      "step": 41
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5055944919586182,
      "learning_rate": 9.808481532147742e-05,
      "loss": 0.4307,
      "step": 42
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5349737405776978,
      "learning_rate": 9.80392156862745e-05,
      "loss": 0.4342,
      "step": 43
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.4112433195114136,
      "learning_rate": 9.799361605107159e-05,
      "loss": 0.4055,
      "step": 44
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.3752671480178833,
      "learning_rate": 9.794801641586868e-05,
      "loss": 0.3995,
      "step": 45
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.56571626663208,
      "learning_rate": 9.790241678066576e-05,
      "loss": 0.4315,
      "step": 46
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5266872644424438,
      "learning_rate": 9.785681714546285e-05,
      "loss": 0.4232,
      "step": 47
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.6925004720687866,
      "learning_rate": 9.781121751025993e-05,
      "loss": 0.4668,
      "step": 48
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.5088814496994019,
      "learning_rate": 9.7765617875057e-05,
      "loss": 0.4033,
      "step": 49
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.4756637811660767,
      "learning_rate": 9.772001823985408e-05,
      "loss": 0.4059,
      "step": 50
    },
    {
      "epoch": 0.02,
      "eval_loss": 0.432329922914505,
      "eval_runtime": 19.0193,
      "eval_samples_per_second": 1.472,
      "eval_steps_per_second": 0.368,
      "step": 50
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.544489860534668,
      "learning_rate": 9.767441860465116e-05,
      "loss": 0.4314,
      "step": 51
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.4627840518951416,
      "learning_rate": 9.762881896944824e-05,
      "loss": 0.3999,
      "step": 52
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.546804666519165,
      "learning_rate": 9.758321933424534e-05,
      "loss": 0.413,
      "step": 53
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.4067586660385132,
      "learning_rate": 9.753761969904242e-05,
      "loss": 0.4112,
      "step": 54
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.5578185319900513,
      "learning_rate": 9.74920200638395e-05,
      "loss": 0.4222,
      "step": 55
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.398961067199707,
      "learning_rate": 9.744642042863658e-05,
      "loss": 0.3935,
      "step": 56
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.5989190340042114,
      "learning_rate": 9.740082079343365e-05,
      "loss": 0.4436,
      "step": 57
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.493788242340088,
      "learning_rate": 9.735522115823073e-05,
      "loss": 0.4161,
      "step": 58
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4532933235168457,
      "learning_rate": 9.730962152302781e-05,
      "loss": 0.4021,
      "step": 59
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.5490344762802124,
      "learning_rate": 9.72640218878249e-05,
      "loss": 0.4049,
      "step": 60
    },
    {
      "epoch": 0.03,
      "eval_loss": 0.4251527786254883,
      "eval_runtime": 19.1976,
      "eval_samples_per_second": 1.459,
      "eval_steps_per_second": 0.365,
      "step": 60
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.670199990272522,
      "learning_rate": 9.721842225262199e-05,
      "loss": 0.4259,
      "step": 61
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3972994089126587,
      "learning_rate": 9.717282261741907e-05,
      "loss": 0.3767,
      "step": 62
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.40248703956604,
      "learning_rate": 9.712722298221615e-05,
      "loss": 0.4021,
      "step": 63
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4790185689926147,
      "learning_rate": 9.708162334701322e-05,
      "loss": 0.4194,
      "step": 64
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4340730905532837,
      "learning_rate": 9.70360237118103e-05,
      "loss": 0.4256,
      "step": 65
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3620188236236572,
      "learning_rate": 9.699042407660739e-05,
      "loss": 0.4111,
      "step": 66
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3097904920578003,
      "learning_rate": 9.694482444140447e-05,
      "loss": 0.391,
      "step": 67
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4144586324691772,
      "learning_rate": 9.689922480620155e-05,
      "loss": 0.3683,
      "step": 68
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.2774299383163452,
      "learning_rate": 9.685362517099864e-05,
      "loss": 0.386,
      "step": 69
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3778272867202759,
      "learning_rate": 9.680802553579573e-05,
      "loss": 0.3938,
      "step": 70
    },
    {
      "epoch": 0.03,
      "eval_loss": 0.4205349385738373,
      "eval_runtime": 18.779,
      "eval_samples_per_second": 1.491,
      "eval_steps_per_second": 0.373,
      "step": 70
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4748997688293457,
      "learning_rate": 9.676242590059281e-05,
      "loss": 0.3927,
      "step": 71
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3778220415115356,
      "learning_rate": 9.671682626538988e-05,
      "loss": 0.4209,
      "step": 72
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4680858850479126,
      "learning_rate": 9.667122663018696e-05,
      "loss": 0.4046,
      "step": 73
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4174176454544067,
      "learning_rate": 9.662562699498404e-05,
      "loss": 0.4025,
      "step": 74
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.4409643411636353,
      "learning_rate": 9.658002735978112e-05,
      "loss": 0.4103,
      "step": 75
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.3730573654174805,
      "learning_rate": 9.65344277245782e-05,
      "loss": 0.4223,
      "step": 76
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3434337377548218,
      "learning_rate": 9.64888280893753e-05,
      "loss": 0.3866,
      "step": 77
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3390862941741943,
      "learning_rate": 9.644322845417238e-05,
      "loss": 0.4057,
      "step": 78
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.264007568359375,
      "learning_rate": 9.639762881896945e-05,
      "loss": 0.3819,
      "step": 79
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3711590766906738,
      "learning_rate": 9.635202918376653e-05,
      "loss": 0.4152,
      "step": 80
    },
    {
      "epoch": 0.04,
      "eval_loss": 0.41787847876548767,
      "eval_runtime": 18.9002,
      "eval_samples_per_second": 1.481,
      "eval_steps_per_second": 0.37,
      "step": 80
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.4036868810653687,
      "learning_rate": 9.630642954856361e-05,
      "loss": 0.4079,
      "step": 81
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.2858104705810547,
      "learning_rate": 9.62608299133607e-05,
      "loss": 0.3716,
      "step": 82
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.4125823974609375,
      "learning_rate": 9.621523027815778e-05,
      "loss": 0.4203,
      "step": 83
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.2835861444473267,
      "learning_rate": 9.616963064295486e-05,
      "loss": 0.3851,
      "step": 84
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3441009521484375,
      "learning_rate": 9.612403100775195e-05,
      "loss": 0.4046,
      "step": 85
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.256738543510437,
      "learning_rate": 9.607843137254903e-05,
      "loss": 0.3782,
      "step": 86
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3569210767745972,
      "learning_rate": 9.60328317373461e-05,
      "loss": 0.4137,
      "step": 87
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.302693247795105,
      "learning_rate": 9.598723210214318e-05,
      "loss": 0.3853,
      "step": 88
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3464365005493164,
      "learning_rate": 9.594163246694027e-05,
      "loss": 0.3692,
      "step": 89
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.302085041999817,
      "learning_rate": 9.589603283173735e-05,
      "loss": 0.386,
      "step": 90
    },
    {
      "epoch": 0.04,
      "eval_loss": 0.41579389572143555,
      "eval_runtime": 19.4361,
      "eval_samples_per_second": 1.441,
      "eval_steps_per_second": 0.36,
      "step": 90
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.8792297840118408,
      "learning_rate": 9.585043319653443e-05,
      "loss": 0.3841,
      "step": 91
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.4461652040481567,
      "learning_rate": 9.580483356133151e-05,
      "loss": 0.3957,
      "step": 92
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.4995123147964478,
      "learning_rate": 9.57592339261286e-05,
      "loss": 0.4188,
      "step": 93
    },
    {
      "epoch": 0.04,
      "grad_norm": 10.198469161987305,
      "learning_rate": 9.571363429092567e-05,
      "loss": 0.4228,
      "step": 94
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.4933909177780151,
      "learning_rate": 9.566803465572276e-05,
      "loss": 0.3749,
      "step": 95
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.293959379196167,
      "learning_rate": 9.562243502051984e-05,
      "loss": 0.3909,
      "step": 96
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3375085592269897,
      "learning_rate": 9.557683538531692e-05,
      "loss": 0.3833,
      "step": 97
    },
    {
      "epoch": 0.04,
      "grad_norm": 1.3052716255187988,
      "learning_rate": 9.5531235750114e-05,
      "loss": 0.386,
      "step": 98
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3119175434112549,
      "learning_rate": 9.548563611491108e-05,
      "loss": 0.3993,
      "step": 99
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.4674808979034424,
      "learning_rate": 9.544003647970816e-05,
      "loss": 0.4081,
      "step": 100
    },
    {
      "epoch": 0.05,
      "eval_loss": 0.4140634536743164,
      "eval_runtime": 19.7343,
      "eval_samples_per_second": 1.419,
      "eval_steps_per_second": 0.355,
      "step": 100
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.4924694299697876,
      "learning_rate": 9.539443684450525e-05,
      "loss": 0.3729,
      "step": 101
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3195910453796387,
      "learning_rate": 9.534883720930233e-05,
      "loss": 0.3832,
      "step": 102
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3233399391174316,
      "learning_rate": 9.530323757409941e-05,
      "loss": 0.4088,
      "step": 103
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3528556823730469,
      "learning_rate": 9.525763793889649e-05,
      "loss": 0.3832,
      "step": 104
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.332626223564148,
      "learning_rate": 9.521203830369357e-05,
      "loss": 0.3973,
      "step": 105
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3452651500701904,
      "learning_rate": 9.516643866849065e-05,
      "loss": 0.3782,
      "step": 106
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2375059127807617,
      "learning_rate": 9.512083903328774e-05,
      "loss": 0.3967,
      "step": 107
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2086222171783447,
      "learning_rate": 9.507523939808482e-05,
      "loss": 0.395,
      "step": 108
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.216988444328308,
      "learning_rate": 9.50296397628819e-05,
      "loss": 0.3826,
      "step": 109
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.1735084056854248,
      "learning_rate": 9.498404012767898e-05,
      "loss": 0.3796,
      "step": 110
    },
    {
      "epoch": 0.05,
      "eval_loss": 0.41126012802124023,
      "eval_runtime": 19.0256,
      "eval_samples_per_second": 1.472,
      "eval_steps_per_second": 0.368,
      "step": 110
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.322290062904358,
      "learning_rate": 9.493844049247606e-05,
      "loss": 0.3912,
      "step": 111
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2272303104400635,
      "learning_rate": 9.489284085727315e-05,
      "loss": 0.3841,
      "step": 112
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2432588338851929,
      "learning_rate": 9.484724122207023e-05,
      "loss": 0.3912,
      "step": 113
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2858808040618896,
      "learning_rate": 9.480164158686731e-05,
      "loss": 0.4006,
      "step": 114
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2801638841629028,
      "learning_rate": 9.475604195166439e-05,
      "loss": 0.4052,
      "step": 115
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.1603589057922363,
      "learning_rate": 9.471044231646147e-05,
      "loss": 0.3861,
      "step": 116
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.261997938156128,
      "learning_rate": 9.466484268125855e-05,
      "loss": 0.4054,
      "step": 117
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.3250972032546997,
      "learning_rate": 9.461924304605564e-05,
      "loss": 0.3916,
      "step": 118
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.2297006845474243,
      "learning_rate": 9.457364341085272e-05,
      "loss": 0.3852,
      "step": 119
    },
    {
      "epoch": 0.05,
      "grad_norm": 1.1711465120315552,
      "learning_rate": 9.45280437756498e-05,
      "loss": 0.3783,
      "step": 120
    },
    {
      "epoch": 0.05,
      "eval_loss": 0.4085245132446289,
      "eval_runtime": 18.864,
      "eval_samples_per_second": 1.484,
      "eval_steps_per_second": 0.371,
      "step": 120
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2444721460342407,
      "learning_rate": 9.448244414044688e-05,
      "loss": 0.4089,
      "step": 121
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.3666821718215942,
      "learning_rate": 9.443684450524396e-05,
      "loss": 0.4065,
      "step": 122
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2235740423202515,
      "learning_rate": 9.439124487004104e-05,
      "loss": 0.3789,
      "step": 123
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2389872074127197,
      "learning_rate": 9.434564523483813e-05,
      "loss": 0.4049,
      "step": 124
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2753331661224365,
      "learning_rate": 9.430004559963521e-05,
      "loss": 0.3976,
      "step": 125
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.1875836849212646,
      "learning_rate": 9.425444596443229e-05,
      "loss": 0.3831,
      "step": 126
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.170867681503296,
      "learning_rate": 9.420884632922937e-05,
      "loss": 0.3842,
      "step": 127
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2708452939987183,
      "learning_rate": 9.416324669402645e-05,
      "loss": 0.3748,
      "step": 128
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.1985563039779663,
      "learning_rate": 9.411764705882353e-05,
      "loss": 0.4001,
      "step": 129
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2549450397491455,
      "learning_rate": 9.407204742362062e-05,
      "loss": 0.3953,
      "step": 130
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.40715333819389343,
      "eval_runtime": 18.7672,
      "eval_samples_per_second": 1.492,
      "eval_steps_per_second": 0.373,
      "step": 130
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2259796857833862,
      "learning_rate": 9.40264477884177e-05,
      "loss": 0.3888,
      "step": 131
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.1683098077774048,
      "learning_rate": 9.398084815321478e-05,
      "loss": 0.3897,
      "step": 132
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.1220805644989014,
      "learning_rate": 9.393524851801186e-05,
      "loss": 0.3676,
      "step": 133
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2481117248535156,
      "learning_rate": 9.388964888280894e-05,
      "loss": 0.4106,
      "step": 134
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.314931869506836,
      "learning_rate": 9.384404924760603e-05,
      "loss": 0.4077,
      "step": 135
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.1738797426223755,
      "learning_rate": 9.379844961240311e-05,
      "loss": 0.3988,
      "step": 136
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.1521250009536743,
      "learning_rate": 9.375284997720019e-05,
      "loss": 0.3818,
      "step": 137
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.29606294631958,
      "learning_rate": 9.370725034199727e-05,
      "loss": 0.4129,
      "step": 138
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.3134340047836304,
      "learning_rate": 9.366165070679435e-05,
      "loss": 0.4004,
      "step": 139
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2621879577636719,
      "learning_rate": 9.361605107159142e-05,
      "loss": 0.4077,
      "step": 140
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.4061693847179413,
      "eval_runtime": 19.4395,
      "eval_samples_per_second": 1.44,
      "eval_steps_per_second": 0.36,
      "step": 140
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.2790113687515259,
      "learning_rate": 9.357045143638852e-05,
      "loss": 0.4147,
      "step": 141
    },
    {
      "epoch": 0.06,
      "grad_norm": 1.3079200983047485,
      "learning_rate": 9.35248518011856e-05,
      "loss": 0.4044,
      "step": 142
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2306909561157227,
      "learning_rate": 9.347925216598268e-05,
      "loss": 0.3801,
      "step": 143
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2091785669326782,
      "learning_rate": 9.343365253077976e-05,
      "loss": 0.3558,
      "step": 144
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.235843300819397,
      "learning_rate": 9.338805289557684e-05,
      "loss": 0.3874,
      "step": 145
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.20920991897583,
      "learning_rate": 9.334245326037392e-05,
      "loss": 0.3818,
      "step": 146
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2847074270248413,
      "learning_rate": 9.329685362517099e-05,
      "loss": 0.3772,
      "step": 147
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2185736894607544,
      "learning_rate": 9.325125398996807e-05,
      "loss": 0.3921,
      "step": 148
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2444103956222534,
      "learning_rate": 9.320565435476517e-05,
      "loss": 0.3914,
      "step": 149
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.283208966255188,
      "learning_rate": 9.316005471956225e-05,
      "loss": 0.3768,
      "step": 150
    },
    {
      "epoch": 0.07,
      "eval_loss": 0.40466779470443726,
      "eval_runtime": 18.8032,
      "eval_samples_per_second": 1.489,
      "eval_steps_per_second": 0.372,
      "step": 150
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2506765127182007,
      "learning_rate": 9.311445508435933e-05,
      "loss": 0.3859,
      "step": 151
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2482283115386963,
      "learning_rate": 9.306885544915641e-05,
      "loss": 0.3991,
      "step": 152
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.1240688562393188,
      "learning_rate": 9.30232558139535e-05,
      "loss": 0.3621,
      "step": 153
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.1728860139846802,
      "learning_rate": 9.297765617875058e-05,
      "loss": 0.3806,
      "step": 154
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.137986183166504,
      "learning_rate": 9.293205654354765e-05,
      "loss": 0.3889,
      "step": 155
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.132051944732666,
      "learning_rate": 9.288645690834473e-05,
      "loss": 0.3763,
      "step": 156
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2157707214355469,
      "learning_rate": 9.284085727314182e-05,
      "loss": 0.373,
      "step": 157
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2097870111465454,
      "learning_rate": 9.27952576379389e-05,
      "loss": 0.4121,
      "step": 158
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.1059770584106445,
      "learning_rate": 9.274965800273599e-05,
      "loss": 0.3716,
      "step": 159
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.170626163482666,
      "learning_rate": 9.270405836753307e-05,
      "loss": 0.3649,
      "step": 160
    },
    {
      "epoch": 0.07,
      "eval_loss": 0.4046716094017029,
      "eval_runtime": 18.9699,
      "eval_samples_per_second": 1.476,
      "eval_steps_per_second": 0.369,
      "step": 160
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.1942059993743896,
      "learning_rate": 9.265845873233015e-05,
      "loss": 0.3768,
      "step": 161
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.2125589847564697,
      "learning_rate": 9.261285909712722e-05,
      "loss": 0.3719,
      "step": 162
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.1660679578781128,
      "learning_rate": 9.25672594619243e-05,
      "loss": 0.3882,
      "step": 163
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.0966999530792236,
      "learning_rate": 9.252165982672138e-05,
      "loss": 0.3588,
      "step": 164
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.197269320487976,
      "learning_rate": 9.247606019151848e-05,
      "loss": 0.3769,
      "step": 165
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2217698097229004,
      "learning_rate": 9.243046055631556e-05,
      "loss": 0.3682,
      "step": 166
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2168915271759033,
      "learning_rate": 9.238486092111264e-05,
      "loss": 0.3894,
      "step": 167
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2309722900390625,
      "learning_rate": 9.233926128590972e-05,
      "loss": 0.3976,
      "step": 168
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2164597511291504,
      "learning_rate": 9.22936616507068e-05,
      "loss": 0.3682,
      "step": 169
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.264969825744629,
      "learning_rate": 9.224806201550387e-05,
      "loss": 0.4121,
      "step": 170
    },
    {
      "epoch": 0.08,
      "eval_loss": 0.4015327990055084,
      "eval_runtime": 19.3928,
      "eval_samples_per_second": 1.444,
      "eval_steps_per_second": 0.361,
      "step": 170
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1746195554733276,
      "learning_rate": 9.220246238030095e-05,
      "loss": 0.3945,
      "step": 171
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1495100259780884,
      "learning_rate": 9.215686274509804e-05,
      "loss": 0.3887,
      "step": 172
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2439409494400024,
      "learning_rate": 9.211126310989513e-05,
      "loss": 0.3995,
      "step": 173
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1578717231750488,
      "learning_rate": 9.206566347469221e-05,
      "loss": 0.3954,
      "step": 174
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2770475149154663,
      "learning_rate": 9.20200638394893e-05,
      "loss": 0.3994,
      "step": 175
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2108159065246582,
      "learning_rate": 9.197446420428638e-05,
      "loss": 0.3764,
      "step": 176
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2297917604446411,
      "learning_rate": 9.192886456908344e-05,
      "loss": 0.3837,
      "step": 177
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1047996282577515,
      "learning_rate": 9.188326493388053e-05,
      "loss": 0.3781,
      "step": 178
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2975915670394897,
      "learning_rate": 9.183766529867761e-05,
      "loss": 0.385,
      "step": 179
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1674829721450806,
      "learning_rate": 9.179206566347469e-05,
      "loss": 0.4053,
      "step": 180
    },
    {
      "epoch": 0.08,
      "eval_loss": 0.4007401764392853,
      "eval_runtime": 19.4644,
      "eval_samples_per_second": 1.439,
      "eval_steps_per_second": 0.36,
      "step": 180
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.2072957754135132,
      "learning_rate": 9.174646602827178e-05,
      "loss": 0.3883,
      "step": 181
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1251564025878906,
      "learning_rate": 9.170086639306887e-05,
      "loss": 0.3667,
      "step": 182
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.0530849695205688,
      "learning_rate": 9.165526675786595e-05,
      "loss": 0.3647,
      "step": 183
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.154868721961975,
      "learning_rate": 9.160966712266302e-05,
      "loss": 0.4002,
      "step": 184
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.1559878587722778,
      "learning_rate": 9.15640674874601e-05,
      "loss": 0.3937,
      "step": 185
    },
    {
      "epoch": 0.08,
      "grad_norm": 1.164669156074524,
      "learning_rate": 9.151846785225718e-05,
      "loss": 0.4066,
      "step": 186
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1499162912368774,
      "learning_rate": 9.147286821705426e-05,
      "loss": 0.3792,
      "step": 187
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1299781799316406,
      "learning_rate": 9.142726858185134e-05,
      "loss": 0.3667,
      "step": 188
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.328697919845581,
      "learning_rate": 9.138166894664844e-05,
      "loss": 0.3775,
      "step": 189
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1182619333267212,
      "learning_rate": 9.133606931144552e-05,
      "loss": 0.3755,
      "step": 190
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.3991912305355072,
      "eval_runtime": 18.7544,
      "eval_samples_per_second": 1.493,
      "eval_steps_per_second": 0.373,
      "step": 190
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1765880584716797,
      "learning_rate": 9.12904696762426e-05,
      "loss": 0.3958,
      "step": 191
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1418038606643677,
      "learning_rate": 9.124487004103967e-05,
      "loss": 0.3893,
      "step": 192
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1207756996154785,
      "learning_rate": 9.119927040583675e-05,
      "loss": 0.4006,
      "step": 193
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1601992845535278,
      "learning_rate": 9.115367077063383e-05,
      "loss": 0.3779,
      "step": 194
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.0996065139770508,
      "learning_rate": 9.110807113543092e-05,
      "loss": 0.362,
      "step": 195
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.0413964986801147,
      "learning_rate": 9.1062471500228e-05,
      "loss": 0.3533,
      "step": 196
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1514556407928467,
      "learning_rate": 9.101687186502509e-05,
      "loss": 0.3753,
      "step": 197
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1734694242477417,
      "learning_rate": 9.097127222982217e-05,
      "loss": 0.3829,
      "step": 198
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1342203617095947,
      "learning_rate": 9.092567259461924e-05,
      "loss": 0.38,
      "step": 199
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.170672059059143,
      "learning_rate": 9.088007295941632e-05,
      "loss": 0.3813,
      "step": 200
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.39865168929100037,
      "eval_runtime": 18.8251,
      "eval_samples_per_second": 1.487,
      "eval_steps_per_second": 0.372,
      "step": 200
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1320968866348267,
      "learning_rate": 9.08344733242134e-05,
      "loss": 0.381,
      "step": 201
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1122519969940186,
      "learning_rate": 9.078887368901049e-05,
      "loss": 0.3912,
      "step": 202
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.2247223854064941,
      "learning_rate": 9.074327405380757e-05,
      "loss": 0.4028,
      "step": 203
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.1854742765426636,
      "learning_rate": 9.069767441860465e-05,
      "loss": 0.3753,
      "step": 204
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.129613995552063,
      "learning_rate": 9.065207478340175e-05,
      "loss": 0.39,
      "step": 205
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.180238127708435,
      "learning_rate": 9.060647514819883e-05,
      "loss": 0.4166,
      "step": 206
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.9991777539253235,
      "learning_rate": 9.05608755129959e-05,
      "loss": 0.369,
      "step": 207
    },
    {
      "epoch": 0.09,
      "grad_norm": 1.0713785886764526,
      "learning_rate": 9.051527587779298e-05,
      "loss": 0.3608,
      "step": 208
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.111507773399353,
      "learning_rate": 9.046967624259006e-05,
      "loss": 0.3802,
      "step": 209
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.14186429977417,
      "learning_rate": 9.042407660738714e-05,
      "loss": 0.3667,
      "step": 210
    },
    {
      "epoch": 0.1,
      "eval_loss": 0.3970547318458557,
      "eval_runtime": 19.182,
      "eval_samples_per_second": 1.46,
      "eval_steps_per_second": 0.365,
      "step": 210
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1212234497070312,
      "learning_rate": 9.037847697218422e-05,
      "loss": 0.3737,
      "step": 211
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.112081527709961,
      "learning_rate": 9.03328773369813e-05,
      "loss": 0.3718,
      "step": 212
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0366638898849487,
      "learning_rate": 9.02872777017784e-05,
      "loss": 0.3623,
      "step": 213
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0450507402420044,
      "learning_rate": 9.024167806657547e-05,
      "loss": 0.3719,
      "step": 214
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1613171100616455,
      "learning_rate": 9.019607843137255e-05,
      "loss": 0.3959,
      "step": 215
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1329877376556396,
      "learning_rate": 9.015047879616963e-05,
      "loss": 0.3851,
      "step": 216
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.3032368421554565,
      "learning_rate": 9.010487916096671e-05,
      "loss": 0.3833,
      "step": 217
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1342045068740845,
      "learning_rate": 9.00592795257638e-05,
      "loss": 0.3653,
      "step": 218
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0300016403198242,
      "learning_rate": 9.001367989056088e-05,
      "loss": 0.371,
      "step": 219
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0597660541534424,
      "learning_rate": 8.996808025535796e-05,
      "loss": 0.371,
      "step": 220
    },
    {
      "epoch": 0.1,
      "eval_loss": 0.39607396721839905,
      "eval_runtime": 18.8558,
      "eval_samples_per_second": 1.485,
      "eval_steps_per_second": 0.371,
      "step": 220
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.113075613975525,
      "learning_rate": 8.992248062015505e-05,
      "loss": 0.3695,
      "step": 221
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0930582284927368,
      "learning_rate": 8.987688098495212e-05,
      "loss": 0.373,
      "step": 222
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0523693561553955,
      "learning_rate": 8.98312813497492e-05,
      "loss": 0.3584,
      "step": 223
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1256333589553833,
      "learning_rate": 8.978568171454629e-05,
      "loss": 0.3839,
      "step": 224
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.156557559967041,
      "learning_rate": 8.974008207934337e-05,
      "loss": 0.3876,
      "step": 225
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1051405668258667,
      "learning_rate": 8.969448244414045e-05,
      "loss": 0.4027,
      "step": 226
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0528311729431152,
      "learning_rate": 8.964888280893753e-05,
      "loss": 0.366,
      "step": 227
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0805798768997192,
      "learning_rate": 8.960328317373461e-05,
      "loss": 0.3661,
      "step": 228
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1037845611572266,
      "learning_rate": 8.95576835385317e-05,
      "loss": 0.3711,
      "step": 229
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.0873870849609375,
      "learning_rate": 8.951208390332878e-05,
      "loss": 0.3886,
      "step": 230
    },
    {
      "epoch": 0.1,
      "eval_loss": 0.3947284519672394,
      "eval_runtime": 18.9717,
      "eval_samples_per_second": 1.476,
      "eval_steps_per_second": 0.369,
      "step": 230
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.1278759241104126,
      "learning_rate": 8.946648426812586e-05,
      "loss": 0.3588,
      "step": 231
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.205957293510437,
      "learning_rate": 8.942088463292294e-05,
      "loss": 0.4015,
      "step": 232
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.081870436668396,
      "learning_rate": 8.937528499772002e-05,
      "loss": 0.3853,
      "step": 233
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0868654251098633,
      "learning_rate": 8.93296853625171e-05,
      "loss": 0.3882,
      "step": 234
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.1404011249542236,
      "learning_rate": 8.928408572731418e-05,
      "loss": 0.3577,
      "step": 235
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.016044020652771,
      "learning_rate": 8.923848609211127e-05,
      "loss": 0.3784,
      "step": 236
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0132030248641968,
      "learning_rate": 8.919288645690835e-05,
      "loss": 0.3562,
      "step": 237
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0909206867218018,
      "learning_rate": 8.914728682170543e-05,
      "loss": 0.3824,
      "step": 238
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0702511072158813,
      "learning_rate": 8.910168718650251e-05,
      "loss": 0.371,
      "step": 239
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.102399468421936,
      "learning_rate": 8.90560875512996e-05,
      "loss": 0.3938,
      "step": 240
    },
    {
      "epoch": 0.11,
      "eval_loss": 0.39401116967201233,
      "eval_runtime": 18.9873,
      "eval_samples_per_second": 1.475,
      "eval_steps_per_second": 0.369,
      "step": 240
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0651742219924927,
      "learning_rate": 8.901048791609668e-05,
      "loss": 0.378,
      "step": 241
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0487183332443237,
      "learning_rate": 8.896488828089376e-05,
      "loss": 0.3832,
      "step": 242
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0751785039901733,
      "learning_rate": 8.891928864569084e-05,
      "loss": 0.3783,
      "step": 243
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0894086360931396,
      "learning_rate": 8.887368901048792e-05,
      "loss": 0.3673,
      "step": 244
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0687947273254395,
      "learning_rate": 8.8828089375285e-05,
      "loss": 0.3596,
      "step": 245
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.075567603111267,
      "learning_rate": 8.878248974008208e-05,
      "loss": 0.39,
      "step": 246
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0721067190170288,
      "learning_rate": 8.873689010487917e-05,
      "loss": 0.3724,
      "step": 247
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0904428958892822,
      "learning_rate": 8.869129046967625e-05,
      "loss": 0.3687,
      "step": 248
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.124393105506897,
      "learning_rate": 8.864569083447333e-05,
      "loss": 0.3614,
      "step": 249
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.0123660564422607,
      "learning_rate": 8.860009119927041e-05,
      "loss": 0.3818,
      "step": 250
    },
    {
      "epoch": 0.11,
      "eval_loss": 0.3932059407234192,
      "eval_runtime": 19.8198,
      "eval_samples_per_second": 1.413,
      "eval_steps_per_second": 0.353,
      "step": 250
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.1686595678329468,
      "learning_rate": 8.855449156406749e-05,
      "loss": 0.3657,
      "step": 251
    },
    {
      "epoch": 0.11,
      "grad_norm": 1.1555325984954834,
      "learning_rate": 8.850889192886457e-05,
      "loss": 0.3953,
      "step": 252
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0188342332839966,
      "learning_rate": 8.846329229366166e-05,
      "loss": 0.3731,
      "step": 253
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0264111757278442,
      "learning_rate": 8.841769265845874e-05,
      "loss": 0.353,
      "step": 254
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.1515635251998901,
      "learning_rate": 8.837209302325582e-05,
      "loss": 0.3982,
      "step": 255
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0572092533111572,
      "learning_rate": 8.83264933880529e-05,
      "loss": 0.4006,
      "step": 256
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.058567762374878,
      "learning_rate": 8.828089375284998e-05,
      "loss": 0.3816,
      "step": 257
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0707305669784546,
      "learning_rate": 8.823529411764706e-05,
      "loss": 0.3614,
      "step": 258
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0715553760528564,
      "learning_rate": 8.818969448244415e-05,
      "loss": 0.3882,
      "step": 259
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0845972299575806,
      "learning_rate": 8.814409484724121e-05,
      "loss": 0.3736,
      "step": 260
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.39302879571914673,
      "eval_runtime": 19.4673,
      "eval_samples_per_second": 1.438,
      "eval_steps_per_second": 0.36,
      "step": 260
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.091538906097412,
      "learning_rate": 8.809849521203831e-05,
      "loss": 0.3735,
      "step": 261
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.012810468673706,
      "learning_rate": 8.805289557683539e-05,
      "loss": 0.3676,
      "step": 262
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.1814091205596924,
      "learning_rate": 8.800729594163247e-05,
      "loss": 0.3964,
      "step": 263
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0467474460601807,
      "learning_rate": 8.796169630642956e-05,
      "loss": 0.3705,
      "step": 264
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0370585918426514,
      "learning_rate": 8.791609667122664e-05,
      "loss": 0.3557,
      "step": 265
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0208948850631714,
      "learning_rate": 8.787049703602372e-05,
      "loss": 0.3821,
      "step": 266
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.9990605711936951,
      "learning_rate": 8.78248974008208e-05,
      "loss": 0.3762,
      "step": 267
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.1138283014297485,
      "learning_rate": 8.777929776561787e-05,
      "loss": 0.3575,
      "step": 268
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.1316741704940796,
      "learning_rate": 8.773369813041496e-05,
      "loss": 0.3896,
      "step": 269
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0367543697357178,
      "learning_rate": 8.768809849521205e-05,
      "loss": 0.3756,
      "step": 270
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.39074134826660156,
      "eval_runtime": 19.5875,
      "eval_samples_per_second": 1.429,
      "eval_steps_per_second": 0.357,
      "step": 270
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.9777958989143372,
      "learning_rate": 8.764249886000913e-05,
      "loss": 0.3613,
      "step": 271
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.0536458492279053,
      "learning_rate": 8.759689922480621e-05,
      "loss": 0.3808,
      "step": 272
    },
    {
      "epoch": 0.12,
      "grad_norm": 1.064129114151001,
      "learning_rate": 8.755129958960329e-05,
      "loss": 0.3738,
      "step": 273
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.9518744349479675,
      "learning_rate": 8.750569995440037e-05,
      "loss": 0.3498,
      "step": 274
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0231995582580566,
      "learning_rate": 8.746010031919744e-05,
      "loss": 0.378,
      "step": 275
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0383845567703247,
      "learning_rate": 8.741450068399452e-05,
      "loss": 0.3832,
      "step": 276
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.1850907802581787,
      "learning_rate": 8.736890104879162e-05,
      "loss": 0.3782,
      "step": 277
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0028883218765259,
      "learning_rate": 8.73233014135887e-05,
      "loss": 0.3641,
      "step": 278
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.9896926879882812,
      "learning_rate": 8.727770177838578e-05,
      "loss": 0.3549,
      "step": 279
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.1111470460891724,
      "learning_rate": 8.723210214318286e-05,
      "loss": 0.3687,
      "step": 280
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.3898495137691498,
      "eval_runtime": 19.1295,
      "eval_samples_per_second": 1.464,
      "eval_steps_per_second": 0.366,
      "step": 280
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0218709707260132,
      "learning_rate": 8.718650250797994e-05,
      "loss": 0.3719,
      "step": 281
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0067551136016846,
      "learning_rate": 8.714090287277701e-05,
      "loss": 0.3632,
      "step": 282
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0330657958984375,
      "learning_rate": 8.70953032375741e-05,
      "loss": 0.3534,
      "step": 283
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0615925788879395,
      "learning_rate": 8.704970360237118e-05,
      "loss": 0.3747,
      "step": 284
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0701462030410767,
      "learning_rate": 8.700410396716827e-05,
      "loss": 0.3588,
      "step": 285
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0732547044754028,
      "learning_rate": 8.695850433196535e-05,
      "loss": 0.3847,
      "step": 286
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0118660926818848,
      "learning_rate": 8.691290469676243e-05,
      "loss": 0.3491,
      "step": 287
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0180116891860962,
      "learning_rate": 8.686730506155952e-05,
      "loss": 0.3697,
      "step": 288
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.028004765510559,
      "learning_rate": 8.68217054263566e-05,
      "loss": 0.3809,
      "step": 289
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.1547340154647827,
      "learning_rate": 8.677610579115367e-05,
      "loss": 0.3813,
      "step": 290
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.38859543204307556,
      "eval_runtime": 19.3351,
      "eval_samples_per_second": 1.448,
      "eval_steps_per_second": 0.362,
      "step": 290
    },
    {
      "epoch": 0.13,
      "grad_norm": 10.94005298614502,
      "learning_rate": 8.673050615595075e-05,
      "loss": 0.3674,
      "step": 291
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.228397011756897,
      "learning_rate": 8.668490652074783e-05,
      "loss": 0.3836,
      "step": 292
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.9820699691772461,
      "learning_rate": 8.663930688554493e-05,
      "loss": 0.3574,
      "step": 293
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0105698108673096,
      "learning_rate": 8.659370725034201e-05,
      "loss": 0.362,
      "step": 294
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.9972749352455139,
      "learning_rate": 8.654810761513909e-05,
      "loss": 0.3559,
      "step": 295
    },
    {
      "epoch": 0.13,
      "grad_norm": 1.0240153074264526,
      "learning_rate": 8.650250797993617e-05,
      "loss": 0.373,
      "step": 296
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0577991008758545,
      "learning_rate": 8.645690834473324e-05,
      "loss": 0.387,
      "step": 297
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.04681396484375,
      "learning_rate": 8.641130870953032e-05,
      "loss": 0.3481,
      "step": 298
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0815809965133667,
      "learning_rate": 8.63657090743274e-05,
      "loss": 0.3767,
      "step": 299
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0435343980789185,
      "learning_rate": 8.632010943912448e-05,
      "loss": 0.3668,
      "step": 300
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.3885766565799713,
      "eval_runtime": 19.117,
      "eval_samples_per_second": 1.465,
      "eval_steps_per_second": 0.366,
      "step": 300
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.044992208480835,
      "learning_rate": 8.627450980392158e-05,
      "loss": 0.3648,
      "step": 301
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9807784557342529,
      "learning_rate": 8.622891016871866e-05,
      "loss": 0.3553,
      "step": 302
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0685855150222778,
      "learning_rate": 8.618331053351574e-05,
      "loss": 0.3731,
      "step": 303
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0137559175491333,
      "learning_rate": 8.613771089831282e-05,
      "loss": 0.3841,
      "step": 304
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9751632213592529,
      "learning_rate": 8.609211126310989e-05,
      "loss": 0.3757,
      "step": 305
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0728063583374023,
      "learning_rate": 8.604651162790697e-05,
      "loss": 0.385,
      "step": 306
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0910289287567139,
      "learning_rate": 8.600091199270406e-05,
      "loss": 0.368,
      "step": 307
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0189052820205688,
      "learning_rate": 8.595531235750114e-05,
      "loss": 0.3743,
      "step": 308
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9784758687019348,
      "learning_rate": 8.590971272229823e-05,
      "loss": 0.3637,
      "step": 309
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.1712713241577148,
      "learning_rate": 8.586411308709531e-05,
      "loss": 0.4026,
      "step": 310
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.38862475752830505,
      "eval_runtime": 18.9756,
      "eval_samples_per_second": 1.476,
      "eval_steps_per_second": 0.369,
      "step": 310
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.027456283569336,
      "learning_rate": 8.58185134518924e-05,
      "loss": 0.3521,
      "step": 311
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9766998291015625,
      "learning_rate": 8.577291381668946e-05,
      "loss": 0.3633,
      "step": 312
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9665610790252686,
      "learning_rate": 8.572731418148655e-05,
      "loss": 0.3779,
      "step": 313
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0428779125213623,
      "learning_rate": 8.568171454628363e-05,
      "loss": 0.3782,
      "step": 314
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9691622257232666,
      "learning_rate": 8.563611491108071e-05,
      "loss": 0.3478,
      "step": 315
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.9197782874107361,
      "learning_rate": 8.559051527587779e-05,
      "loss": 0.3399,
      "step": 316
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.1256742477416992,
      "learning_rate": 8.554491564067489e-05,
      "loss": 0.3791,
      "step": 317
    },
    {
      "epoch": 0.14,
      "grad_norm": 1.0777599811553955,
      "learning_rate": 8.549931600547197e-05,
      "loss": 0.3742,
      "step": 318
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.09661865234375,
      "learning_rate": 8.545371637026904e-05,
      "loss": 0.3706,
      "step": 319
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0405218601226807,
      "learning_rate": 8.540811673506612e-05,
      "loss": 0.3704,
      "step": 320
    },
    {
      "epoch": 0.15,
      "eval_loss": 0.387399822473526,
      "eval_runtime": 18.6057,
      "eval_samples_per_second": 1.505,
      "eval_steps_per_second": 0.376,
      "step": 320
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.988017737865448,
      "learning_rate": 8.53625170998632e-05,
      "loss": 0.3436,
      "step": 321
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0084255933761597,
      "learning_rate": 8.531691746466028e-05,
      "loss": 0.3846,
      "step": 322
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.024824619293213,
      "learning_rate": 8.527131782945736e-05,
      "loss": 0.3915,
      "step": 323
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0249780416488647,
      "learning_rate": 8.522571819425445e-05,
      "loss": 0.3589,
      "step": 324
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0125277042388916,
      "learning_rate": 8.518011855905154e-05,
      "loss": 0.3761,
      "step": 325
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9836946725845337,
      "learning_rate": 8.513451892384862e-05,
      "loss": 0.3621,
      "step": 326
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0279687643051147,
      "learning_rate": 8.508891928864569e-05,
      "loss": 0.3823,
      "step": 327
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0053000450134277,
      "learning_rate": 8.504331965344277e-05,
      "loss": 0.3882,
      "step": 328
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9956850409507751,
      "learning_rate": 8.499772001823985e-05,
      "loss": 0.3537,
      "step": 329
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9967787265777588,
      "learning_rate": 8.495212038303694e-05,
      "loss": 0.3472,
      "step": 330
    },
    {
      "epoch": 0.15,
      "eval_loss": 0.3873635232448578,
      "eval_runtime": 20.0228,
      "eval_samples_per_second": 1.398,
      "eval_steps_per_second": 0.35,
      "step": 330
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0272730588912964,
      "learning_rate": 8.490652074783402e-05,
      "loss": 0.3758,
      "step": 331
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.004315972328186,
      "learning_rate": 8.48609211126311e-05,
      "loss": 0.372,
      "step": 332
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9904900193214417,
      "learning_rate": 8.48153214774282e-05,
      "loss": 0.3648,
      "step": 333
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9831294417381287,
      "learning_rate": 8.476972184222526e-05,
      "loss": 0.3534,
      "step": 334
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9594573974609375,
      "learning_rate": 8.472412220702234e-05,
      "loss": 0.3547,
      "step": 335
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0190843343734741,
      "learning_rate": 8.467852257181943e-05,
      "loss": 0.3562,
      "step": 336
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0010571479797363,
      "learning_rate": 8.463292293661651e-05,
      "loss": 0.3777,
      "step": 337
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.9827868342399597,
      "learning_rate": 8.458732330141359e-05,
      "loss": 0.3554,
      "step": 338
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.943844199180603,
      "learning_rate": 8.454172366621067e-05,
      "loss": 0.3556,
      "step": 339
    },
    {
      "epoch": 0.15,
      "grad_norm": 1.0718072652816772,
      "learning_rate": 8.449612403100775e-05,
      "loss": 0.3625,
      "step": 340
    },
    {
      "epoch": 0.15,
      "eval_loss": 0.38646969199180603,
      "eval_runtime": 18.7738,
      "eval_samples_per_second": 1.491,
      "eval_steps_per_second": 0.373,
      "step": 340
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.038783073425293,
      "learning_rate": 8.445052439580485e-05,
      "loss": 0.3892,
      "step": 341
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.996299147605896,
      "learning_rate": 8.440492476060192e-05,
      "loss": 0.3731,
      "step": 342
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0760611295700073,
      "learning_rate": 8.4359325125399e-05,
      "loss": 0.3403,
      "step": 343
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0220232009887695,
      "learning_rate": 8.431372549019608e-05,
      "loss": 0.3702,
      "step": 344
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9995104670524597,
      "learning_rate": 8.426812585499316e-05,
      "loss": 0.3697,
      "step": 345
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9776452779769897,
      "learning_rate": 8.422252621979024e-05,
      "loss": 0.3518,
      "step": 346
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.042927861213684,
      "learning_rate": 8.417692658458733e-05,
      "loss": 0.3703,
      "step": 347
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0032014846801758,
      "learning_rate": 8.413132694938441e-05,
      "loss": 0.3666,
      "step": 348
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9905934929847717,
      "learning_rate": 8.408572731418149e-05,
      "loss": 0.345,
      "step": 349
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9962011575698853,
      "learning_rate": 8.404012767897857e-05,
      "loss": 0.3679,
      "step": 350
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.38614824414253235,
      "eval_runtime": 19.1496,
      "eval_samples_per_second": 1.462,
      "eval_steps_per_second": 0.366,
      "step": 350
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0067214965820312,
      "learning_rate": 8.399452804377565e-05,
      "loss": 0.368,
      "step": 351
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0402851104736328,
      "learning_rate": 8.394892840857273e-05,
      "loss": 0.3707,
      "step": 352
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.924884021282196,
      "learning_rate": 8.390332877336982e-05,
      "loss": 0.3584,
      "step": 353
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9956393837928772,
      "learning_rate": 8.38577291381669e-05,
      "loss": 0.3608,
      "step": 354
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0277490615844727,
      "learning_rate": 8.381212950296398e-05,
      "loss": 0.379,
      "step": 355
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.0546191930770874,
      "learning_rate": 8.376652986776106e-05,
      "loss": 0.3868,
      "step": 356
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9730943441390991,
      "learning_rate": 8.372093023255814e-05,
      "loss": 0.3681,
      "step": 357
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.995608925819397,
      "learning_rate": 8.367533059735522e-05,
      "loss": 0.3728,
      "step": 358
    },
    {
      "epoch": 0.16,
      "grad_norm": 1.039383888244629,
      "learning_rate": 8.36297309621523e-05,
      "loss": 0.3815,
      "step": 359
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9661096930503845,
      "learning_rate": 8.358413132694939e-05,
      "loss": 0.3392,
      "step": 360
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.3855496048927307,
      "eval_runtime": 19.2455,
      "eval_samples_per_second": 1.455,
      "eval_steps_per_second": 0.364,
      "step": 360
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.9818366169929504,
      "learning_rate": 8.353853169174647e-05,
      "loss": 0.3898,
      "step": 361
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.140006184577942,
      "learning_rate": 8.349293205654355e-05,
      "loss": 0.3799,
      "step": 362
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9910751581192017,
      "learning_rate": 8.344733242134063e-05,
      "loss": 0.3561,
      "step": 363
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9846599102020264,
      "learning_rate": 8.340173278613771e-05,
      "loss": 0.3751,
      "step": 364
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9890525341033936,
      "learning_rate": 8.33561331509348e-05,
      "loss": 0.3717,
      "step": 365
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.0564824342727661,
      "learning_rate": 8.331053351573188e-05,
      "loss": 0.3361,
      "step": 366
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9904609322547913,
      "learning_rate": 8.326493388052896e-05,
      "loss": 0.382,
      "step": 367
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9419021606445312,
      "learning_rate": 8.321933424532604e-05,
      "loss": 0.3426,
      "step": 368
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.976959764957428,
      "learning_rate": 8.317373461012312e-05,
      "loss": 0.3562,
      "step": 369
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9872338175773621,
      "learning_rate": 8.31281349749202e-05,
      "loss": 0.3417,
      "step": 370
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.3853757679462433,
      "eval_runtime": 19.3296,
      "eval_samples_per_second": 1.449,
      "eval_steps_per_second": 0.362,
      "step": 370
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9730595350265503,
      "learning_rate": 8.308253533971729e-05,
      "loss": 0.3597,
      "step": 371
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9857218861579895,
      "learning_rate": 8.303693570451437e-05,
      "loss": 0.3632,
      "step": 372
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9228557348251343,
      "learning_rate": 8.299133606931145e-05,
      "loss": 0.3417,
      "step": 373
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.03836989402771,
      "learning_rate": 8.294573643410853e-05,
      "loss": 0.3601,
      "step": 374
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.973513662815094,
      "learning_rate": 8.290013679890561e-05,
      "loss": 0.3598,
      "step": 375
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.0164240598678589,
      "learning_rate": 8.28545371637027e-05,
      "loss": 0.3645,
      "step": 376
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9688085317611694,
      "learning_rate": 8.280893752849978e-05,
      "loss": 0.3579,
      "step": 377
    },
    {
      "epoch": 0.17,
      "grad_norm": 1.1452713012695312,
      "learning_rate": 8.276333789329686e-05,
      "loss": 0.3879,
      "step": 378
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9936511516571045,
      "learning_rate": 8.271773825809394e-05,
      "loss": 0.3838,
      "step": 379
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9094933867454529,
      "learning_rate": 8.267213862289101e-05,
      "loss": 0.3332,
      "step": 380
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.383484810590744,
      "eval_runtime": 19.0505,
      "eval_samples_per_second": 1.47,
      "eval_steps_per_second": 0.367,
      "step": 380
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9642332196235657,
      "learning_rate": 8.26265389876881e-05,
      "loss": 0.3659,
      "step": 381
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9800812005996704,
      "learning_rate": 8.258093935248519e-05,
      "loss": 0.3882,
      "step": 382
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.9797675013542175,
      "learning_rate": 8.253533971728227e-05,
      "loss": 0.3792,
      "step": 383
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.0148242712020874,
      "learning_rate": 8.248974008207935e-05,
      "loss": 0.3604,
      "step": 384
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9496625661849976,
      "learning_rate": 8.244414044687643e-05,
      "loss": 0.3636,
      "step": 385
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9917979836463928,
      "learning_rate": 8.239854081167351e-05,
      "loss": 0.351,
      "step": 386
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9986299276351929,
      "learning_rate": 8.23529411764706e-05,
      "loss": 0.386,
      "step": 387
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9700773358345032,
      "learning_rate": 8.230734154126766e-05,
      "loss": 0.3597,
      "step": 388
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.0540348291397095,
      "learning_rate": 8.226174190606476e-05,
      "loss": 0.3744,
      "step": 389
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.018379807472229,
      "learning_rate": 8.221614227086184e-05,
      "loss": 0.3872,
      "step": 390
    },
    {
      "epoch": 0.18,
      "eval_loss": 0.3828833997249603,
      "eval_runtime": 18.5979,
      "eval_samples_per_second": 1.506,
      "eval_steps_per_second": 0.376,
      "step": 390
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9792240262031555,
      "learning_rate": 8.217054263565892e-05,
      "loss": 0.3993,
      "step": 391
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9694705605506897,
      "learning_rate": 8.2124943000456e-05,
      "loss": 0.3567,
      "step": 392
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.0545449256896973,
      "learning_rate": 8.207934336525308e-05,
      "loss": 0.375,
      "step": 393
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.995250940322876,
      "learning_rate": 8.203374373005017e-05,
      "loss": 0.3596,
      "step": 394
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.993552029132843,
      "learning_rate": 8.198814409484723e-05,
      "loss": 0.3555,
      "step": 395
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9589958786964417,
      "learning_rate": 8.194254445964432e-05,
      "loss": 0.3699,
      "step": 396
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9881184101104736,
      "learning_rate": 8.189694482444141e-05,
      "loss": 0.36,
      "step": 397
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9230943918228149,
      "learning_rate": 8.18513451892385e-05,
      "loss": 0.3635,
      "step": 398
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9407085180282593,
      "learning_rate": 8.180574555403558e-05,
      "loss": 0.3821,
      "step": 399
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9004390239715576,
      "learning_rate": 8.176014591883266e-05,
      "loss": 0.3671,
      "step": 400
    },
    {
      "epoch": 0.18,
      "eval_loss": 0.3836938738822937,
      "eval_runtime": 19.9469,
      "eval_samples_per_second": 1.404,
      "eval_steps_per_second": 0.351,
      "step": 400
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9622935056686401,
      "learning_rate": 8.171454628362974e-05,
      "loss": 0.3451,
      "step": 401
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9424606561660767,
      "learning_rate": 8.166894664842682e-05,
      "loss": 0.3635,
      "step": 402
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9327376484870911,
      "learning_rate": 8.162334701322389e-05,
      "loss": 0.3632,
      "step": 403
    },
    {
      "epoch": 0.18,
      "grad_norm": 1.0619184970855713,
      "learning_rate": 8.157774737802097e-05,
      "loss": 0.3886,
      "step": 404
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.9540975689888,
      "learning_rate": 8.153214774281807e-05,
      "loss": 0.3503,
      "step": 405
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.895407497882843,
      "learning_rate": 8.148654810761515e-05,
      "loss": 0.3503,
      "step": 406
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9608978033065796,
      "learning_rate": 8.144094847241223e-05,
      "loss": 0.3598,
      "step": 407
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9120776057243347,
      "learning_rate": 8.139534883720931e-05,
      "loss": 0.343,
      "step": 408
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.028601884841919,
      "learning_rate": 8.134974920200639e-05,
      "loss": 0.3827,
      "step": 409
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9658374786376953,
      "learning_rate": 8.130414956680346e-05,
      "loss": 0.3595,
      "step": 410
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.38228359818458557,
      "eval_runtime": 18.8049,
      "eval_samples_per_second": 1.489,
      "eval_steps_per_second": 0.372,
      "step": 410
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9547862410545349,
      "learning_rate": 8.125854993160054e-05,
      "loss": 0.378,
      "step": 411
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9772255420684814,
      "learning_rate": 8.121295029639762e-05,
      "loss": 0.3727,
      "step": 412
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9499548673629761,
      "learning_rate": 8.116735066119472e-05,
      "loss": 0.366,
      "step": 413
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9336138963699341,
      "learning_rate": 8.11217510259918e-05,
      "loss": 0.3429,
      "step": 414
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9108003377914429,
      "learning_rate": 8.107615139078888e-05,
      "loss": 0.3384,
      "step": 415
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9173622131347656,
      "learning_rate": 8.103055175558596e-05,
      "loss": 0.3577,
      "step": 416
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9635920524597168,
      "learning_rate": 8.098495212038303e-05,
      "loss": 0.3695,
      "step": 417
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9141045808792114,
      "learning_rate": 8.093935248518011e-05,
      "loss": 0.3538,
      "step": 418
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.990811824798584,
      "learning_rate": 8.08937528499772e-05,
      "loss": 0.3604,
      "step": 419
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9809904098510742,
      "learning_rate": 8.084815321477428e-05,
      "loss": 0.345,
      "step": 420
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.38126489520072937,
      "eval_runtime": 19.1235,
      "eval_samples_per_second": 1.464,
      "eval_steps_per_second": 0.366,
      "step": 420
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9208155870437622,
      "learning_rate": 8.080255357957137e-05,
      "loss": 0.3682,
      "step": 421
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.1504791975021362,
      "learning_rate": 8.075695394436846e-05,
      "loss": 0.362,
      "step": 422
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9805225133895874,
      "learning_rate": 8.071135430916554e-05,
      "loss": 0.3646,
      "step": 423
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.0297439098358154,
      "learning_rate": 8.066575467396262e-05,
      "loss": 0.383,
      "step": 424
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9397987127304077,
      "learning_rate": 8.062015503875969e-05,
      "loss": 0.3735,
      "step": 425
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.9980968832969666,
      "learning_rate": 8.057455540355677e-05,
      "loss": 0.3658,
      "step": 426
    },
    {
      "epoch": 0.19,
      "grad_norm": 1.0988585948944092,
      "learning_rate": 8.052895576835385e-05,
      "loss": 0.3752,
      "step": 427
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9558773040771484,
      "learning_rate": 8.048335613315093e-05,
      "loss": 0.3613,
      "step": 428
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9506344199180603,
      "learning_rate": 8.043775649794803e-05,
      "loss": 0.3397,
      "step": 429
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9391772150993347,
      "learning_rate": 8.039215686274511e-05,
      "loss": 0.3618,
      "step": 430
    },
    {
      "epoch": 0.2,
      "eval_loss": 0.3807036578655243,
      "eval_runtime": 18.5019,
      "eval_samples_per_second": 1.513,
      "eval_steps_per_second": 0.378,
      "step": 430
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9951292872428894,
      "learning_rate": 8.034655722754219e-05,
      "loss": 0.3525,
      "step": 431
    },
    {
      "epoch": 0.2,
      "grad_norm": 1.022258996963501,
      "learning_rate": 8.030095759233926e-05,
      "loss": 0.3954,
      "step": 432
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9829654693603516,
      "learning_rate": 8.025535795713634e-05,
      "loss": 0.359,
      "step": 433
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9890940189361572,
      "learning_rate": 8.020975832193342e-05,
      "loss": 0.3812,
      "step": 434
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.91367107629776,
      "learning_rate": 8.01641586867305e-05,
      "loss": 0.361,
      "step": 435
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9756280779838562,
      "learning_rate": 8.011855905152759e-05,
      "loss": 0.3588,
      "step": 436
    },
    {
      "epoch": 0.2,
      "grad_norm": 1.0047883987426758,
      "learning_rate": 8.007295941632468e-05,
      "loss": 0.3769,
      "step": 437
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9199777841567993,
      "learning_rate": 8.002735978112176e-05,
      "loss": 0.3645,
      "step": 438
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9628649950027466,
      "learning_rate": 7.998176014591884e-05,
      "loss": 0.3631,
      "step": 439
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.888653039932251,
      "learning_rate": 7.993616051071591e-05,
      "loss": 0.3424,
      "step": 440
    },
    {
      "epoch": 0.2,
      "eval_loss": 0.3802674114704132,
      "eval_runtime": 20.1434,
      "eval_samples_per_second": 1.39,
      "eval_steps_per_second": 0.348,
      "step": 440
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9911215305328369,
      "learning_rate": 7.9890560875513e-05,
      "loss": 0.3562,
      "step": 441
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9457067251205444,
      "learning_rate": 7.984496124031008e-05,
      "loss": 0.3718,
      "step": 442
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9514966011047363,
      "learning_rate": 7.979936160510716e-05,
      "loss": 0.358,
      "step": 443
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.949561357498169,
      "learning_rate": 7.975376196990424e-05,
      "loss": 0.3769,
      "step": 444
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9274330735206604,
      "learning_rate": 7.970816233470134e-05,
      "loss": 0.3638,
      "step": 445
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9469530582427979,
      "learning_rate": 7.966256269949842e-05,
      "loss": 0.3661,
      "step": 446
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9310869574546814,
      "learning_rate": 7.961696306429548e-05,
      "loss": 0.353,
      "step": 447
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.961445152759552,
      "learning_rate": 7.957136342909257e-05,
      "loss": 0.3463,
      "step": 448
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.9427098035812378,
      "learning_rate": 7.952576379388965e-05,
      "loss": 0.3545,
      "step": 449
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.8861632943153381,
      "learning_rate": 7.948016415868673e-05,
      "loss": 0.3308,
      "step": 450
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.37948015332221985,
      "eval_runtime": 19.2271,
      "eval_samples_per_second": 1.456,
      "eval_steps_per_second": 0.364,
      "step": 450
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9233769178390503,
      "learning_rate": 7.943456452348381e-05,
      "loss": 0.3498,
      "step": 451
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9102416634559631,
      "learning_rate": 7.93889648882809e-05,
      "loss": 0.3504,
      "step": 452
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9856906533241272,
      "learning_rate": 7.934336525307799e-05,
      "loss": 0.3707,
      "step": 453
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9343120455741882,
      "learning_rate": 7.929776561787506e-05,
      "loss": 0.341,
      "step": 454
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.006124496459961,
      "learning_rate": 7.925216598267214e-05,
      "loss": 0.364,
      "step": 455
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.8969525694847107,
      "learning_rate": 7.920656634746922e-05,
      "loss": 0.3473,
      "step": 456
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.978102445602417,
      "learning_rate": 7.91609667122663e-05,
      "loss": 0.3927,
      "step": 457
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.060112714767456,
      "learning_rate": 7.911536707706338e-05,
      "loss": 0.3989,
      "step": 458
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9657124280929565,
      "learning_rate": 7.906976744186047e-05,
      "loss": 0.3926,
      "step": 459
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9384371638298035,
      "learning_rate": 7.902416780665755e-05,
      "loss": 0.3768,
      "step": 460
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.37937769293785095,
      "eval_runtime": 18.5987,
      "eval_samples_per_second": 1.505,
      "eval_steps_per_second": 0.376,
      "step": 460
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.953213632106781,
      "learning_rate": 7.897856817145464e-05,
      "loss": 0.3659,
      "step": 461
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.8762606382369995,
      "learning_rate": 7.893296853625171e-05,
      "loss": 0.3213,
      "step": 462
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9968709945678711,
      "learning_rate": 7.888736890104879e-05,
      "loss": 0.3897,
      "step": 463
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9149399399757385,
      "learning_rate": 7.884176926584587e-05,
      "loss": 0.3633,
      "step": 464
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9021903276443481,
      "learning_rate": 7.879616963064296e-05,
      "loss": 0.3586,
      "step": 465
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.8712608814239502,
      "learning_rate": 7.875056999544004e-05,
      "loss": 0.3425,
      "step": 466
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9100139141082764,
      "learning_rate": 7.870497036023712e-05,
      "loss": 0.3693,
      "step": 467
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9456149339675903,
      "learning_rate": 7.86593707250342e-05,
      "loss": 0.3745,
      "step": 468
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.8988533616065979,
      "learning_rate": 7.861377108983128e-05,
      "loss": 0.3434,
      "step": 469
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.8404950499534607,
      "learning_rate": 7.856817145462836e-05,
      "loss": 0.343,
      "step": 470
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.37977489829063416,
      "eval_runtime": 19.6011,
      "eval_samples_per_second": 1.428,
      "eval_steps_per_second": 0.357,
      "step": 470
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.9720378518104553,
      "learning_rate": 7.852257181942545e-05,
      "loss": 0.3644,
      "step": 471
    },
    {
      "epoch": 0.22,
      "grad_norm": 1.0149002075195312,
      "learning_rate": 7.847697218422253e-05,
      "loss": 0.3705,
      "step": 472
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9858806133270264,
      "learning_rate": 7.843137254901961e-05,
      "loss": 0.3941,
      "step": 473
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9632289409637451,
      "learning_rate": 7.838577291381669e-05,
      "loss": 0.3725,
      "step": 474
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9068118333816528,
      "learning_rate": 7.834017327861377e-05,
      "loss": 0.3415,
      "step": 475
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9591249823570251,
      "learning_rate": 7.829457364341086e-05,
      "loss": 0.3561,
      "step": 476
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9359074234962463,
      "learning_rate": 7.824897400820794e-05,
      "loss": 0.3541,
      "step": 477
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9554955959320068,
      "learning_rate": 7.820337437300502e-05,
      "loss": 0.3755,
      "step": 478
    },
    {
      "epoch": 0.22,
      "grad_norm": 1.0127731561660767,
      "learning_rate": 7.81577747378021e-05,
      "loss": 0.3832,
      "step": 479
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9365389943122864,
      "learning_rate": 7.811217510259918e-05,
      "loss": 0.3706,
      "step": 480
    },
    {
      "epoch": 0.22,
      "eval_loss": 0.379183828830719,
      "eval_runtime": 18.8228,
      "eval_samples_per_second": 1.488,
      "eval_steps_per_second": 0.372,
      "step": 480
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9115477204322815,
      "learning_rate": 7.806657546739626e-05,
      "loss": 0.3645,
      "step": 481
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.8891696929931641,
      "learning_rate": 7.802097583219335e-05,
      "loss": 0.3554,
      "step": 482
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.8747963905334473,
      "learning_rate": 7.797537619699043e-05,
      "loss": 0.3399,
      "step": 483
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.8941763043403625,
      "learning_rate": 7.792977656178751e-05,
      "loss": 0.3535,
      "step": 484
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9057623147964478,
      "learning_rate": 7.788417692658459e-05,
      "loss": 0.365,
      "step": 485
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9394456148147583,
      "learning_rate": 7.783857729138167e-05,
      "loss": 0.3496,
      "step": 486
    },
    {
      "epoch": 0.22,
      "grad_norm": 1.0419434309005737,
      "learning_rate": 7.779297765617875e-05,
      "loss": 0.3576,
      "step": 487
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9206737279891968,
      "learning_rate": 7.774737802097584e-05,
      "loss": 0.3609,
      "step": 488
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9505032300949097,
      "learning_rate": 7.770177838577292e-05,
      "loss": 0.3806,
      "step": 489
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9697739481925964,
      "learning_rate": 7.765617875057e-05,
      "loss": 0.369,
      "step": 490
    },
    {
      "epoch": 0.22,
      "eval_loss": 0.37793412804603577,
      "eval_runtime": 18.9214,
      "eval_samples_per_second": 1.48,
      "eval_steps_per_second": 0.37,
      "step": 490
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.8664564490318298,
      "learning_rate": 7.761057911536708e-05,
      "loss": 0.3509,
      "step": 491
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.8611375093460083,
      "learning_rate": 7.756497948016416e-05,
      "loss": 0.3294,
      "step": 492
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.9530375003814697,
      "learning_rate": 7.751937984496124e-05,
      "loss": 0.3566,
      "step": 493
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9074346423149109,
      "learning_rate": 7.747378020975833e-05,
      "loss": 0.3635,
      "step": 494
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.8907040953636169,
      "learning_rate": 7.742818057455541e-05,
      "loss": 0.364,
      "step": 495
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9278532266616821,
      "learning_rate": 7.738258093935249e-05,
      "loss": 0.3563,
      "step": 496
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.956122875213623,
      "learning_rate": 7.733698130414957e-05,
      "loss": 0.3599,
      "step": 497
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9482458233833313,
      "learning_rate": 7.729138166894665e-05,
      "loss": 0.3626,
      "step": 498
    },
    {
      "epoch": 0.23,
      "grad_norm": 1.004625678062439,
      "learning_rate": 7.724578203374374e-05,
      "loss": 0.3884,
      "step": 499
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9188464283943176,
      "learning_rate": 7.720018239854082e-05,
      "loss": 0.3553,
      "step": 500
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.37794917821884155,
      "eval_runtime": 18.631,
      "eval_samples_per_second": 1.503,
      "eval_steps_per_second": 0.376,
      "step": 500
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.8780580759048462,
      "learning_rate": 7.71545827633379e-05,
      "loss": 0.3404,
      "step": 501
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.8963892459869385,
      "learning_rate": 7.710898312813498e-05,
      "loss": 0.3406,
      "step": 502
    },
    {
      "epoch": 0.23,
      "grad_norm": 1.0074235200881958,
      "learning_rate": 7.706338349293206e-05,
      "loss": 0.3795,
      "step": 503
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9363308548927307,
      "learning_rate": 7.701778385772914e-05,
      "loss": 0.3524,
      "step": 504
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.854862630367279,
      "learning_rate": 7.697218422252623e-05,
      "loss": 0.3351,
      "step": 505
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9411963224411011,
      "learning_rate": 7.692658458732331e-05,
      "loss": 0.365,
      "step": 506
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9519062042236328,
      "learning_rate": 7.688098495212039e-05,
      "loss": 0.3606,
      "step": 507
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.936029314994812,
      "learning_rate": 7.683538531691746e-05,
      "loss": 0.3773,
      "step": 508
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.8771556615829468,
      "learning_rate": 7.678978568171455e-05,
      "loss": 0.3492,
      "step": 509
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.878050684928894,
      "learning_rate": 7.674418604651163e-05,
      "loss": 0.3503,
      "step": 510
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.3781495690345764,
      "eval_runtime": 19.9333,
      "eval_samples_per_second": 1.405,
      "eval_steps_per_second": 0.351,
      "step": 510
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9384998083114624,
      "learning_rate": 7.669858641130872e-05,
      "loss": 0.3682,
      "step": 511
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.923982560634613,
      "learning_rate": 7.66529867761058e-05,
      "loss": 0.3583,
      "step": 512
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.8898745775222778,
      "learning_rate": 7.660738714090288e-05,
      "loss": 0.3767,
      "step": 513
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.9237766861915588,
      "learning_rate": 7.656178750569996e-05,
      "loss": 0.369,
      "step": 514
    },
    {
      "epoch": 0.23,
      "grad_norm": 1.0031055212020874,
      "learning_rate": 7.651618787049703e-05,
      "loss": 0.3662,
      "step": 515
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9065768122673035,
      "learning_rate": 7.647058823529411e-05,
      "loss": 0.3557,
      "step": 516
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9564458131790161,
      "learning_rate": 7.64249886000912e-05,
      "loss": 0.3621,
      "step": 517
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9868760704994202,
      "learning_rate": 7.637938896488829e-05,
      "loss": 0.3492,
      "step": 518
    },
    {
      "epoch": 0.24,
      "grad_norm": 1.0082308053970337,
      "learning_rate": 7.633378932968537e-05,
      "loss": 0.3833,
      "step": 519
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.903857409954071,
      "learning_rate": 7.628818969448245e-05,
      "loss": 0.3633,
      "step": 520
    },
    {
      "epoch": 0.24,
      "eval_loss": 0.3774409294128418,
      "eval_runtime": 19.0237,
      "eval_samples_per_second": 1.472,
      "eval_steps_per_second": 0.368,
      "step": 520
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.8547798991203308,
      "learning_rate": 7.624259005927953e-05,
      "loss": 0.3611,
      "step": 521
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9237222075462341,
      "learning_rate": 7.619699042407661e-05,
      "loss": 0.3652,
      "step": 522
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.8383327722549438,
      "learning_rate": 7.615139078887368e-05,
      "loss": 0.3571,
      "step": 523
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.994269073009491,
      "learning_rate": 7.610579115367076e-05,
      "loss": 0.37,
      "step": 524
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.872664749622345,
      "learning_rate": 7.606019151846786e-05,
      "loss": 0.362,
      "step": 525
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9679466485977173,
      "learning_rate": 7.601459188326494e-05,
      "loss": 0.3581,
      "step": 526
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9206968545913696,
      "learning_rate": 7.596899224806202e-05,
      "loss": 0.3525,
      "step": 527
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.915579617023468,
      "learning_rate": 7.59233926128591e-05,
      "loss": 0.346,
      "step": 528
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.942768394947052,
      "learning_rate": 7.587779297765619e-05,
      "loss": 0.3708,
      "step": 529
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.948949933052063,
      "learning_rate": 7.583219334245326e-05,
      "loss": 0.342,
      "step": 530
    },
    {
      "epoch": 0.24,
      "eval_loss": 0.37704339623451233,
      "eval_runtime": 19.4572,
      "eval_samples_per_second": 1.439,
      "eval_steps_per_second": 0.36,
      "step": 530
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.8952234983444214,
      "learning_rate": 7.578659370725034e-05,
      "loss": 0.3672,
      "step": 531
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9527508616447449,
      "learning_rate": 7.574099407204742e-05,
      "loss": 0.3652,
      "step": 532
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9585651755332947,
      "learning_rate": 7.569539443684451e-05,
      "loss": 0.3725,
      "step": 533
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.8955938816070557,
      "learning_rate": 7.56497948016416e-05,
      "loss": 0.3429,
      "step": 534
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9243647456169128,
      "learning_rate": 7.560419516643868e-05,
      "loss": 0.3494,
      "step": 535
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.914239227771759,
      "learning_rate": 7.555859553123576e-05,
      "loss": 0.3741,
      "step": 536
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.8691146373748779,
      "learning_rate": 7.551299589603284e-05,
      "loss": 0.3526,
      "step": 537
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9309206604957581,
      "learning_rate": 7.546739626082991e-05,
      "loss": 0.3747,
      "step": 538
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.8218508362770081,
      "learning_rate": 7.542179662562699e-05,
      "loss": 0.3519,
      "step": 539
    },
    {
      "epoch": 0.25,
      "grad_norm": 1.1142666339874268,
      "learning_rate": 7.537619699042407e-05,
      "loss": 0.3545,
      "step": 540
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.3766672909259796,
      "eval_runtime": 18.6678,
      "eval_samples_per_second": 1.5,
      "eval_steps_per_second": 0.375,
      "step": 540
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9692361354827881,
      "learning_rate": 7.533059735522117e-05,
      "loss": 0.3895,
      "step": 541
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9041081070899963,
      "learning_rate": 7.528499772001825e-05,
      "loss": 0.3583,
      "step": 542
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9008587002754211,
      "learning_rate": 7.523939808481533e-05,
      "loss": 0.3498,
      "step": 543
    },
    {
      "epoch": 0.25,
      "grad_norm": 1.0266526937484741,
      "learning_rate": 7.519379844961241e-05,
      "loss": 0.3612,
      "step": 544
    },
    {
      "epoch": 0.25,
      "grad_norm": 1.0442652702331543,
      "learning_rate": 7.514819881440948e-05,
      "loss": 0.3806,
      "step": 545
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.8736815452575684,
      "learning_rate": 7.510259917920656e-05,
      "loss": 0.3163,
      "step": 546
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.8623315691947937,
      "learning_rate": 7.505699954400364e-05,
      "loss": 0.3701,
      "step": 547
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9528596997261047,
      "learning_rate": 7.501139990880073e-05,
      "loss": 0.3426,
      "step": 548
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.8725919127464294,
      "learning_rate": 7.496580027359782e-05,
      "loss": 0.3552,
      "step": 549
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9121567010879517,
      "learning_rate": 7.49202006383949e-05,
      "loss": 0.3616,
      "step": 550
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.37652093172073364,
      "eval_runtime": 19.7078,
      "eval_samples_per_second": 1.421,
      "eval_steps_per_second": 0.355,
      "step": 550
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.916845440864563,
      "learning_rate": 7.487460100319199e-05,
      "loss": 0.3626,
      "step": 551
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.8767221570014954,
      "learning_rate": 7.482900136798905e-05,
      "loss": 0.361,
      "step": 552
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.8678212761878967,
      "learning_rate": 7.478340173278614e-05,
      "loss": 0.3508,
      "step": 553
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9074848294258118,
      "learning_rate": 7.473780209758322e-05,
      "loss": 0.3555,
      "step": 554
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9501338005065918,
      "learning_rate": 7.46922024623803e-05,
      "loss": 0.3662,
      "step": 555
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9243327379226685,
      "learning_rate": 7.464660282717738e-05,
      "loss": 0.3767,
      "step": 556
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9555984139442444,
      "learning_rate": 7.460100319197448e-05,
      "loss": 0.3614,
      "step": 557
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9904035329818726,
      "learning_rate": 7.455540355677156e-05,
      "loss": 0.343,
      "step": 558
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.9084532260894775,
      "learning_rate": 7.450980392156864e-05,
      "loss": 0.3734,
      "step": 559
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9660534858703613,
      "learning_rate": 7.446420428636571e-05,
      "loss": 0.3912,
      "step": 560
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.37544146180152893,
      "eval_runtime": 20.1661,
      "eval_samples_per_second": 1.388,
      "eval_steps_per_second": 0.347,
      "step": 560
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.8810905814170837,
      "learning_rate": 7.441860465116279e-05,
      "loss": 0.3565,
      "step": 561
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9330177307128906,
      "learning_rate": 7.437300501595987e-05,
      "loss": 0.3797,
      "step": 562
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9678060412406921,
      "learning_rate": 7.432740538075695e-05,
      "loss": 0.3512,
      "step": 563
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.919135332107544,
      "learning_rate": 7.428180574555403e-05,
      "loss": 0.3739,
      "step": 564
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9367424845695496,
      "learning_rate": 7.423620611035113e-05,
      "loss": 0.3912,
      "step": 565
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.8687885999679565,
      "learning_rate": 7.419060647514821e-05,
      "loss": 0.3517,
      "step": 566
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9798654913902283,
      "learning_rate": 7.414500683994528e-05,
      "loss": 0.3538,
      "step": 567
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9142093658447266,
      "learning_rate": 7.409940720474236e-05,
      "loss": 0.3614,
      "step": 568
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9957043528556824,
      "learning_rate": 7.405380756953944e-05,
      "loss": 0.3635,
      "step": 569
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9421626329421997,
      "learning_rate": 7.400820793433652e-05,
      "loss": 0.374,
      "step": 570
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.37424519658088684,
      "eval_runtime": 19.7385,
      "eval_samples_per_second": 1.419,
      "eval_steps_per_second": 0.355,
      "step": 570
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9157000780105591,
      "learning_rate": 7.39626082991336e-05,
      "loss": 0.3695,
      "step": 571
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9190824031829834,
      "learning_rate": 7.391700866393069e-05,
      "loss": 0.3661,
      "step": 572
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.8940838575363159,
      "learning_rate": 7.387140902872778e-05,
      "loss": 0.3399,
      "step": 573
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9431096911430359,
      "learning_rate": 7.382580939352486e-05,
      "loss": 0.378,
      "step": 574
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9267783761024475,
      "learning_rate": 7.378020975832193e-05,
      "loss": 0.3752,
      "step": 575
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.895793616771698,
      "learning_rate": 7.373461012311901e-05,
      "loss": 0.3676,
      "step": 576
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9508955478668213,
      "learning_rate": 7.36890104879161e-05,
      "loss": 0.3461,
      "step": 577
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9015037417411804,
      "learning_rate": 7.364341085271318e-05,
      "loss": 0.3689,
      "step": 578
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.9231910109519958,
      "learning_rate": 7.359781121751026e-05,
      "loss": 0.3495,
      "step": 579
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.902377188205719,
      "learning_rate": 7.355221158230734e-05,
      "loss": 0.3802,
      "step": 580
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.37441423535346985,
      "eval_runtime": 22.0055,
      "eval_samples_per_second": 1.272,
      "eval_steps_per_second": 0.318,
      "step": 580
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.859978973865509,
      "learning_rate": 7.350661194710444e-05,
      "loss": 0.3719,
      "step": 581
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9164068102836609,
      "learning_rate": 7.34610123119015e-05,
      "loss": 0.3728,
      "step": 582
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9238768219947815,
      "learning_rate": 7.341541267669859e-05,
      "loss": 0.3574,
      "step": 583
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8287404775619507,
      "learning_rate": 7.336981304149567e-05,
      "loss": 0.3387,
      "step": 584
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8957213163375854,
      "learning_rate": 7.332421340629275e-05,
      "loss": 0.3327,
      "step": 585
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9304167628288269,
      "learning_rate": 7.327861377108983e-05,
      "loss": 0.371,
      "step": 586
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8944963216781616,
      "learning_rate": 7.323301413588691e-05,
      "loss": 0.3484,
      "step": 587
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9239458441734314,
      "learning_rate": 7.3187414500684e-05,
      "loss": 0.4004,
      "step": 588
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8646803498268127,
      "learning_rate": 7.314181486548109e-05,
      "loss": 0.3641,
      "step": 589
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8503308296203613,
      "learning_rate": 7.309621523027816e-05,
      "loss": 0.3579,
      "step": 590
    },
    {
      "epoch": 0.27,
      "eval_loss": 0.37435466051101685,
      "eval_runtime": 20.8667,
      "eval_samples_per_second": 1.342,
      "eval_steps_per_second": 0.335,
      "step": 590
    },
    {
      "epoch": 0.27,
      "grad_norm": 1.0227876901626587,
      "learning_rate": 7.305061559507524e-05,
      "loss": 0.3305,
      "step": 591
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8426918387413025,
      "learning_rate": 7.300501595987232e-05,
      "loss": 0.3278,
      "step": 592
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9567776918411255,
      "learning_rate": 7.29594163246694e-05,
      "loss": 0.3607,
      "step": 593
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8379486203193665,
      "learning_rate": 7.291381668946649e-05,
      "loss": 0.3262,
      "step": 594
    },
    {
      "epoch": 0.27,
      "grad_norm": 1.170776605606079,
      "learning_rate": 7.286821705426357e-05,
      "loss": 0.3802,
      "step": 595
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9064086079597473,
      "learning_rate": 7.282261741906065e-05,
      "loss": 0.3625,
      "step": 596
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.89022296667099,
      "learning_rate": 7.277701778385773e-05,
      "loss": 0.3563,
      "step": 597
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9683102369308472,
      "learning_rate": 7.273141814865481e-05,
      "loss": 0.3595,
      "step": 598
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9998959898948669,
      "learning_rate": 7.26858185134519e-05,
      "loss": 0.3735,
      "step": 599
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8802072405815125,
      "learning_rate": 7.264021887824898e-05,
      "loss": 0.3479,
      "step": 600
    },
    {
      "epoch": 0.27,
      "eval_loss": 0.37338584661483765,
      "eval_runtime": 21.634,
      "eval_samples_per_second": 1.294,
      "eval_steps_per_second": 0.324,
      "step": 600
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.8717195987701416,
      "learning_rate": 7.259461924304606e-05,
      "loss": 0.3522,
      "step": 601
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9232763648033142,
      "learning_rate": 7.254901960784314e-05,
      "loss": 0.3492,
      "step": 602
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.9206188321113586,
      "learning_rate": 7.250341997264022e-05,
      "loss": 0.3572,
      "step": 603
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9413841962814331,
      "learning_rate": 7.24578203374373e-05,
      "loss": 0.3645,
      "step": 604
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9230554699897766,
      "learning_rate": 7.241222070223439e-05,
      "loss": 0.3594,
      "step": 605
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9277369379997253,
      "learning_rate": 7.236662106703147e-05,
      "loss": 0.3661,
      "step": 606
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8822442889213562,
      "learning_rate": 7.232102143182855e-05,
      "loss": 0.3533,
      "step": 607
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8277364373207092,
      "learning_rate": 7.227542179662563e-05,
      "loss": 0.3329,
      "step": 608
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8693709969520569,
      "learning_rate": 7.222982216142271e-05,
      "loss": 0.3421,
      "step": 609
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9155862331390381,
      "learning_rate": 7.21842225262198e-05,
      "loss": 0.3749,
      "step": 610
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.37279894948005676,
      "eval_runtime": 30.1926,
      "eval_samples_per_second": 0.927,
      "eval_steps_per_second": 0.232,
      "step": 610
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9259756803512573,
      "learning_rate": 7.213862289101688e-05,
      "loss": 0.3748,
      "step": 611
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9580633044242859,
      "learning_rate": 7.209302325581396e-05,
      "loss": 0.3528,
      "step": 612
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8764567971229553,
      "learning_rate": 7.204742362061104e-05,
      "loss": 0.3562,
      "step": 613
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8910514116287231,
      "learning_rate": 7.200182398540812e-05,
      "loss": 0.3507,
      "step": 614
    },
    {
      "epoch": 0.28,
      "grad_norm": 1.0229401588439941,
      "learning_rate": 7.19562243502052e-05,
      "loss": 0.3837,
      "step": 615
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8673608899116516,
      "learning_rate": 7.191062471500228e-05,
      "loss": 0.3548,
      "step": 616
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8642386794090271,
      "learning_rate": 7.186502507979937e-05,
      "loss": 0.3409,
      "step": 617
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9027450680732727,
      "learning_rate": 7.181942544459645e-05,
      "loss": 0.3605,
      "step": 618
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.950926661491394,
      "learning_rate": 7.177382580939353e-05,
      "loss": 0.3833,
      "step": 619
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.910386323928833,
      "learning_rate": 7.172822617419061e-05,
      "loss": 0.3639,
      "step": 620
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.37247234582901,
      "eval_runtime": 29.593,
      "eval_samples_per_second": 0.946,
      "eval_steps_per_second": 0.237,
      "step": 620
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9070430397987366,
      "learning_rate": 7.168262653898769e-05,
      "loss": 0.3685,
      "step": 621
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9218928813934326,
      "learning_rate": 7.163702690378477e-05,
      "loss": 0.3657,
      "step": 622
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.87645423412323,
      "learning_rate": 7.159142726858186e-05,
      "loss": 0.3557,
      "step": 623
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.9534425139427185,
      "learning_rate": 7.154582763337894e-05,
      "loss": 0.3471,
      "step": 624
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.8581913709640503,
      "learning_rate": 7.150022799817602e-05,
      "loss": 0.358,
      "step": 625
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8720976114273071,
      "learning_rate": 7.14546283629731e-05,
      "loss": 0.3455,
      "step": 626
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8933624625205994,
      "learning_rate": 7.140902872777018e-05,
      "loss": 0.3512,
      "step": 627
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8910596966743469,
      "learning_rate": 7.136342909256725e-05,
      "loss": 0.368,
      "step": 628
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.865628719329834,
      "learning_rate": 7.131782945736435e-05,
      "loss": 0.3528,
      "step": 629
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.905426025390625,
      "learning_rate": 7.127222982216143e-05,
      "loss": 0.374,
      "step": 630
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.37291833758354187,
      "eval_runtime": 22.3333,
      "eval_samples_per_second": 1.254,
      "eval_steps_per_second": 0.313,
      "step": 630
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8680101633071899,
      "learning_rate": 7.122663018695851e-05,
      "loss": 0.3755,
      "step": 631
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.866011917591095,
      "learning_rate": 7.118103055175559e-05,
      "loss": 0.3504,
      "step": 632
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8784955143928528,
      "learning_rate": 7.113543091655267e-05,
      "loss": 0.361,
      "step": 633
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8499343991279602,
      "learning_rate": 7.108983128134976e-05,
      "loss": 0.3534,
      "step": 634
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.9045304656028748,
      "learning_rate": 7.104423164614684e-05,
      "loss": 0.3685,
      "step": 635
    },
    {
      "epoch": 0.29,
      "grad_norm": 1.0086599588394165,
      "learning_rate": 7.09986320109439e-05,
      "loss": 0.3772,
      "step": 636
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8259167075157166,
      "learning_rate": 7.0953032375741e-05,
      "loss": 0.325,
      "step": 637
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8794663548469543,
      "learning_rate": 7.090743274053808e-05,
      "loss": 0.3311,
      "step": 638
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8893126249313354,
      "learning_rate": 7.086183310533516e-05,
      "loss": 0.3505,
      "step": 639
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.9002717137336731,
      "learning_rate": 7.081623347013225e-05,
      "loss": 0.3497,
      "step": 640
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.3717420995235443,
      "eval_runtime": 18.0961,
      "eval_samples_per_second": 1.547,
      "eval_steps_per_second": 0.387,
      "step": 640
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8322933912277222,
      "learning_rate": 7.077063383492933e-05,
      "loss": 0.3451,
      "step": 641
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8885860443115234,
      "learning_rate": 7.072503419972641e-05,
      "loss": 0.3646,
      "step": 642
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8684788346290588,
      "learning_rate": 7.067943456452348e-05,
      "loss": 0.3568,
      "step": 643
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.853153645992279,
      "learning_rate": 7.063383492932056e-05,
      "loss": 0.3571,
      "step": 644
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8872026205062866,
      "learning_rate": 7.058823529411765e-05,
      "loss": 0.3763,
      "step": 645
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8363185524940491,
      "learning_rate": 7.054263565891474e-05,
      "loss": 0.3589,
      "step": 646
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.8505784869194031,
      "learning_rate": 7.049703602371182e-05,
      "loss": 0.3627,
      "step": 647
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8950116038322449,
      "learning_rate": 7.04514363885089e-05,
      "loss": 0.3634,
      "step": 648
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.876524031162262,
      "learning_rate": 7.040583675330598e-05,
      "loss": 0.3601,
      "step": 649
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8355085849761963,
      "learning_rate": 7.036023711810305e-05,
      "loss": 0.3641,
      "step": 650
    },
    {
      "epoch": 0.3,
      "eval_loss": 0.3709128201007843,
      "eval_runtime": 21.856,
      "eval_samples_per_second": 1.281,
      "eval_steps_per_second": 0.32,
      "step": 650
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8563446998596191,
      "learning_rate": 7.031463748290013e-05,
      "loss": 0.3314,
      "step": 651
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.893064022064209,
      "learning_rate": 7.026903784769721e-05,
      "loss": 0.3606,
      "step": 652
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8532779812812805,
      "learning_rate": 7.022343821249431e-05,
      "loss": 0.3432,
      "step": 653
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8648200035095215,
      "learning_rate": 7.017783857729139e-05,
      "loss": 0.342,
      "step": 654
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.9161383509635925,
      "learning_rate": 7.013223894208847e-05,
      "loss": 0.3798,
      "step": 655
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.9029960632324219,
      "learning_rate": 7.008663930688555e-05,
      "loss": 0.3594,
      "step": 656
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.838525652885437,
      "learning_rate": 7.004103967168264e-05,
      "loss": 0.3565,
      "step": 657
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.9098020792007446,
      "learning_rate": 6.99954400364797e-05,
      "loss": 0.3595,
      "step": 658
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.915035605430603,
      "learning_rate": 6.994984040127679e-05,
      "loss": 0.3582,
      "step": 659
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.9485026001930237,
      "learning_rate": 6.990424076607387e-05,
      "loss": 0.3709,
      "step": 660
    },
    {
      "epoch": 0.3,
      "eval_loss": 0.37105894088745117,
      "eval_runtime": 22.2753,
      "eval_samples_per_second": 1.257,
      "eval_steps_per_second": 0.314,
      "step": 660
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.929305374622345,
      "learning_rate": 6.985864113087096e-05,
      "loss": 0.3812,
      "step": 661
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8890742659568787,
      "learning_rate": 6.981304149566804e-05,
      "loss": 0.3588,
      "step": 662
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.9145094156265259,
      "learning_rate": 6.976744186046513e-05,
      "loss": 0.383,
      "step": 663
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8319873809814453,
      "learning_rate": 6.972184222526221e-05,
      "loss": 0.3429,
      "step": 664
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8460900187492371,
      "learning_rate": 6.967624259005928e-05,
      "loss": 0.3371,
      "step": 665
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8295953869819641,
      "learning_rate": 6.963064295485636e-05,
      "loss": 0.3497,
      "step": 666
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8634557723999023,
      "learning_rate": 6.958504331965344e-05,
      "loss": 0.377,
      "step": 667
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.8128460049629211,
      "learning_rate": 6.953944368445052e-05,
      "loss": 0.3359,
      "step": 668
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.913492739200592,
      "learning_rate": 6.949384404924762e-05,
      "loss": 0.3632,
      "step": 669
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8793885111808777,
      "learning_rate": 6.94482444140447e-05,
      "loss": 0.3407,
      "step": 670
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.37183091044425964,
      "eval_runtime": 20.3451,
      "eval_samples_per_second": 1.376,
      "eval_steps_per_second": 0.344,
      "step": 670
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8706486225128174,
      "learning_rate": 6.940264477884178e-05,
      "loss": 0.3715,
      "step": 671
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.910187304019928,
      "learning_rate": 6.935704514363886e-05,
      "loss": 0.3639,
      "step": 672
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8096619844436646,
      "learning_rate": 6.931144550843593e-05,
      "loss": 0.3388,
      "step": 673
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8766554594039917,
      "learning_rate": 6.926584587323301e-05,
      "loss": 0.346,
      "step": 674
    },
    {
      "epoch": 0.31,
      "grad_norm": 1.1496071815490723,
      "learning_rate": 6.922024623803009e-05,
      "loss": 0.3507,
      "step": 675
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9036510586738586,
      "learning_rate": 6.917464660282717e-05,
      "loss": 0.3485,
      "step": 676
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9343310594558716,
      "learning_rate": 6.912904696762427e-05,
      "loss": 0.3609,
      "step": 677
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8673393726348877,
      "learning_rate": 6.908344733242135e-05,
      "loss": 0.3626,
      "step": 678
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9198659062385559,
      "learning_rate": 6.903784769721843e-05,
      "loss": 0.3712,
      "step": 679
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9207894802093506,
      "learning_rate": 6.89922480620155e-05,
      "loss": 0.3507,
      "step": 680
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.37167611718177795,
      "eval_runtime": 22.1884,
      "eval_samples_per_second": 1.262,
      "eval_steps_per_second": 0.315,
      "step": 680
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.847029447555542,
      "learning_rate": 6.894664842681258e-05,
      "loss": 0.3429,
      "step": 681
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9102590680122375,
      "learning_rate": 6.890104879160966e-05,
      "loss": 0.3748,
      "step": 682
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9330032467842102,
      "learning_rate": 6.885544915640675e-05,
      "loss": 0.362,
      "step": 683
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8520646691322327,
      "learning_rate": 6.880984952120383e-05,
      "loss": 0.3574,
      "step": 684
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.868168294429779,
      "learning_rate": 6.876424988600092e-05,
      "loss": 0.3597,
      "step": 685
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8473092913627625,
      "learning_rate": 6.8718650250798e-05,
      "loss": 0.3375,
      "step": 686
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.9174085259437561,
      "learning_rate": 6.867305061559507e-05,
      "loss": 0.3822,
      "step": 687
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8579756617546082,
      "learning_rate": 6.862745098039216e-05,
      "loss": 0.3487,
      "step": 688
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8969879150390625,
      "learning_rate": 6.858185134518924e-05,
      "loss": 0.3635,
      "step": 689
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.868897020816803,
      "learning_rate": 6.853625170998632e-05,
      "loss": 0.3437,
      "step": 690
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.3700122535228729,
      "eval_runtime": 21.7146,
      "eval_samples_per_second": 1.289,
      "eval_steps_per_second": 0.322,
      "step": 690
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.8899202942848206,
      "learning_rate": 6.84906520747834e-05,
      "loss": 0.3423,
      "step": 691
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.772549033164978,
      "learning_rate": 6.844505243958048e-05,
      "loss": 0.3315,
      "step": 692
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8518693447113037,
      "learning_rate": 6.839945280437758e-05,
      "loss": 0.3371,
      "step": 693
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8638544678688049,
      "learning_rate": 6.835385316917466e-05,
      "loss": 0.3413,
      "step": 694
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8001556396484375,
      "learning_rate": 6.830825353397173e-05,
      "loss": 0.3266,
      "step": 695
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8450673222541809,
      "learning_rate": 6.826265389876881e-05,
      "loss": 0.3658,
      "step": 696
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8260478973388672,
      "learning_rate": 6.821705426356589e-05,
      "loss": 0.3426,
      "step": 697
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.7822526097297668,
      "learning_rate": 6.817145462836297e-05,
      "loss": 0.3371,
      "step": 698
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.901171088218689,
      "learning_rate": 6.812585499316005e-05,
      "loss": 0.3842,
      "step": 699
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.9115702509880066,
      "learning_rate": 6.808025535795714e-05,
      "loss": 0.3551,
      "step": 700
    },
    {
      "epoch": 0.32,
      "eval_loss": 0.36962977051734924,
      "eval_runtime": 22.9971,
      "eval_samples_per_second": 1.218,
      "eval_steps_per_second": 0.304,
      "step": 700
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8465586304664612,
      "learning_rate": 6.803465572275423e-05,
      "loss": 0.3361,
      "step": 701
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8594992756843567,
      "learning_rate": 6.79890560875513e-05,
      "loss": 0.3409,
      "step": 702
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.85016930103302,
      "learning_rate": 6.794345645234838e-05,
      "loss": 0.3411,
      "step": 703
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.9131604433059692,
      "learning_rate": 6.789785681714546e-05,
      "loss": 0.3684,
      "step": 704
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.9225114583969116,
      "learning_rate": 6.785225718194254e-05,
      "loss": 0.3436,
      "step": 705
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8579698801040649,
      "learning_rate": 6.780665754673963e-05,
      "loss": 0.3453,
      "step": 706
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8255881071090698,
      "learning_rate": 6.776105791153671e-05,
      "loss": 0.3213,
      "step": 707
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8486250042915344,
      "learning_rate": 6.771545827633379e-05,
      "loss": 0.3361,
      "step": 708
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.867413341999054,
      "learning_rate": 6.766985864113089e-05,
      "loss": 0.3645,
      "step": 709
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8573243021965027,
      "learning_rate": 6.762425900592795e-05,
      "loss": 0.354,
      "step": 710
    },
    {
      "epoch": 0.32,
      "eval_loss": 0.36976662278175354,
      "eval_runtime": 22.6059,
      "eval_samples_per_second": 1.239,
      "eval_steps_per_second": 0.31,
      "step": 710
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.8835360407829285,
      "learning_rate": 6.757865937072504e-05,
      "loss": 0.3484,
      "step": 711
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.865594208240509,
      "learning_rate": 6.753305973552212e-05,
      "loss": 0.354,
      "step": 712
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8187012672424316,
      "learning_rate": 6.74874601003192e-05,
      "loss": 0.3331,
      "step": 713
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.9948188662528992,
      "learning_rate": 6.744186046511628e-05,
      "loss": 0.3837,
      "step": 714
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8897073864936829,
      "learning_rate": 6.739626082991336e-05,
      "loss": 0.3541,
      "step": 715
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8769757747650146,
      "learning_rate": 6.735066119471044e-05,
      "loss": 0.3362,
      "step": 716
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8944910168647766,
      "learning_rate": 6.730506155950753e-05,
      "loss": 0.3493,
      "step": 717
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8475896120071411,
      "learning_rate": 6.725946192430461e-05,
      "loss": 0.3651,
      "step": 718
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8484331369400024,
      "learning_rate": 6.721386228910169e-05,
      "loss": 0.341,
      "step": 719
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8814172148704529,
      "learning_rate": 6.716826265389877e-05,
      "loss": 0.3624,
      "step": 720
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.3701036870479584,
      "eval_runtime": 21.9379,
      "eval_samples_per_second": 1.276,
      "eval_steps_per_second": 0.319,
      "step": 720
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8305844664573669,
      "learning_rate": 6.712266301869585e-05,
      "loss": 0.3502,
      "step": 721
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.884107232093811,
      "learning_rate": 6.707706338349293e-05,
      "loss": 0.3599,
      "step": 722
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8543264269828796,
      "learning_rate": 6.703146374829002e-05,
      "loss": 0.3569,
      "step": 723
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8755825161933899,
      "learning_rate": 6.69858641130871e-05,
      "loss": 0.3599,
      "step": 724
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8918110132217407,
      "learning_rate": 6.694026447788418e-05,
      "loss": 0.3488,
      "step": 725
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8799797296524048,
      "learning_rate": 6.689466484268126e-05,
      "loss": 0.3512,
      "step": 726
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8555727601051331,
      "learning_rate": 6.684906520747834e-05,
      "loss": 0.3485,
      "step": 727
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8324811458587646,
      "learning_rate": 6.680346557227542e-05,
      "loss": 0.3426,
      "step": 728
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8397484421730042,
      "learning_rate": 6.67578659370725e-05,
      "loss": 0.3452,
      "step": 729
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8771885633468628,
      "learning_rate": 6.671226630186959e-05,
      "loss": 0.3416,
      "step": 730
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.3702766001224518,
      "eval_runtime": 21.4551,
      "eval_samples_per_second": 1.305,
      "eval_steps_per_second": 0.326,
      "step": 730
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8133786916732788,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.3377,
      "step": 731
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8744335174560547,
      "learning_rate": 6.662106703146375e-05,
      "loss": 0.3581,
      "step": 732
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8324525952339172,
      "learning_rate": 6.657546739626083e-05,
      "loss": 0.3432,
      "step": 733
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.8790922164916992,
      "learning_rate": 6.652986776105792e-05,
      "loss": 0.3563,
      "step": 734
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.855384886264801,
      "learning_rate": 6.6484268125855e-05,
      "loss": 0.3543,
      "step": 735
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8443394303321838,
      "learning_rate": 6.643866849065208e-05,
      "loss": 0.3453,
      "step": 736
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8796120882034302,
      "learning_rate": 6.639306885544916e-05,
      "loss": 0.3468,
      "step": 737
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8887386918067932,
      "learning_rate": 6.634746922024624e-05,
      "loss": 0.3559,
      "step": 738
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8751109838485718,
      "learning_rate": 6.630186958504332e-05,
      "loss": 0.3398,
      "step": 739
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8693111538887024,
      "learning_rate": 6.62562699498404e-05,
      "loss": 0.3519,
      "step": 740
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.3703303039073944,
      "eval_runtime": 21.0686,
      "eval_samples_per_second": 1.329,
      "eval_steps_per_second": 0.332,
      "step": 740
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8620005249977112,
      "learning_rate": 6.621067031463749e-05,
      "loss": 0.359,
      "step": 741
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8814571499824524,
      "learning_rate": 6.616507067943457e-05,
      "loss": 0.3532,
      "step": 742
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8492552042007446,
      "learning_rate": 6.611947104423165e-05,
      "loss": 0.3379,
      "step": 743
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8217783570289612,
      "learning_rate": 6.607387140902873e-05,
      "loss": 0.343,
      "step": 744
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8890987634658813,
      "learning_rate": 6.602827177382581e-05,
      "loss": 0.369,
      "step": 745
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8515855073928833,
      "learning_rate": 6.59826721386229e-05,
      "loss": 0.3502,
      "step": 746
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8381286263465881,
      "learning_rate": 6.593707250341998e-05,
      "loss": 0.3419,
      "step": 747
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.7878593802452087,
      "learning_rate": 6.589147286821705e-05,
      "loss": 0.3262,
      "step": 748
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8711159229278564,
      "learning_rate": 6.584587323301414e-05,
      "loss": 0.3616,
      "step": 749
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8434100151062012,
      "learning_rate": 6.580027359781122e-05,
      "loss": 0.3465,
      "step": 750
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.3702242076396942,
      "eval_runtime": 21.1864,
      "eval_samples_per_second": 1.322,
      "eval_steps_per_second": 0.33,
      "step": 750
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8792661428451538,
      "learning_rate": 6.57546739626083e-05,
      "loss": 0.3704,
      "step": 751
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8422108292579651,
      "learning_rate": 6.570907432740539e-05,
      "loss": 0.3516,
      "step": 752
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.9015281796455383,
      "learning_rate": 6.566347469220247e-05,
      "loss": 0.3602,
      "step": 753
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8497557640075684,
      "learning_rate": 6.561787505699955e-05,
      "loss": 0.3507,
      "step": 754
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8488445281982422,
      "learning_rate": 6.557227542179663e-05,
      "loss": 0.333,
      "step": 755
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.8524776101112366,
      "learning_rate": 6.55266757865937e-05,
      "loss": 0.363,
      "step": 756
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.879468560218811,
      "learning_rate": 6.54810761513908e-05,
      "loss": 0.3769,
      "step": 757
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.844282865524292,
      "learning_rate": 6.543547651618788e-05,
      "loss": 0.3697,
      "step": 758
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8251934051513672,
      "learning_rate": 6.538987688098496e-05,
      "loss": 0.3377,
      "step": 759
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.9126609563827515,
      "learning_rate": 6.534427724578204e-05,
      "loss": 0.3696,
      "step": 760
    },
    {
      "epoch": 0.35,
      "eval_loss": 0.3702351450920105,
      "eval_runtime": 21.0169,
      "eval_samples_per_second": 1.332,
      "eval_steps_per_second": 0.333,
      "step": 760
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.823424756526947,
      "learning_rate": 6.529867761057912e-05,
      "loss": 0.3535,
      "step": 761
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8669154644012451,
      "learning_rate": 6.52530779753762e-05,
      "loss": 0.3451,
      "step": 762
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.9502606391906738,
      "learning_rate": 6.520747834017327e-05,
      "loss": 0.3967,
      "step": 763
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8474224209785461,
      "learning_rate": 6.516187870497035e-05,
      "loss": 0.357,
      "step": 764
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8229814767837524,
      "learning_rate": 6.511627906976745e-05,
      "loss": 0.3513,
      "step": 765
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8437790274620056,
      "learning_rate": 6.507067943456453e-05,
      "loss": 0.3412,
      "step": 766
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8239248991012573,
      "learning_rate": 6.502507979936161e-05,
      "loss": 0.3358,
      "step": 767
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8583698868751526,
      "learning_rate": 6.49794801641587e-05,
      "loss": 0.3408,
      "step": 768
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.7928481101989746,
      "learning_rate": 6.493388052895578e-05,
      "loss": 0.3325,
      "step": 769
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.9882442951202393,
      "learning_rate": 6.488828089375286e-05,
      "loss": 0.3517,
      "step": 770
    },
    {
      "epoch": 0.35,
      "eval_loss": 0.3702905774116516,
      "eval_runtime": 21.183,
      "eval_samples_per_second": 1.322,
      "eval_steps_per_second": 0.33,
      "step": 770
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.9171540141105652,
      "learning_rate": 6.484268125854993e-05,
      "loss": 0.3614,
      "step": 771
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8405658602714539,
      "learning_rate": 6.479708162334701e-05,
      "loss": 0.3538,
      "step": 772
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8744710683822632,
      "learning_rate": 6.47514819881441e-05,
      "loss": 0.3498,
      "step": 773
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8798863291740417,
      "learning_rate": 6.470588235294118e-05,
      "loss": 0.3943,
      "step": 774
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8916918039321899,
      "learning_rate": 6.466028271773827e-05,
      "loss": 0.3822,
      "step": 775
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8482193946838379,
      "learning_rate": 6.461468308253535e-05,
      "loss": 0.3465,
      "step": 776
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8396366238594055,
      "learning_rate": 6.456908344733243e-05,
      "loss": 0.3554,
      "step": 777
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.8911833167076111,
      "learning_rate": 6.45234838121295e-05,
      "loss": 0.3487,
      "step": 778
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.879220187664032,
      "learning_rate": 6.447788417692658e-05,
      "loss": 0.3585,
      "step": 779
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8869410157203674,
      "learning_rate": 6.443228454172366e-05,
      "loss": 0.3497,
      "step": 780
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.369541198015213,
      "eval_runtime": 21.1307,
      "eval_samples_per_second": 1.325,
      "eval_steps_per_second": 0.331,
      "step": 780
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8716014623641968,
      "learning_rate": 6.438668490652076e-05,
      "loss": 0.3771,
      "step": 781
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.891190767288208,
      "learning_rate": 6.434108527131784e-05,
      "loss": 0.3718,
      "step": 782
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8753977417945862,
      "learning_rate": 6.429548563611492e-05,
      "loss": 0.3421,
      "step": 783
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8322952389717102,
      "learning_rate": 6.4249886000912e-05,
      "loss": 0.3447,
      "step": 784
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8560271859169006,
      "learning_rate": 6.420428636570907e-05,
      "loss": 0.3584,
      "step": 785
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8641065359115601,
      "learning_rate": 6.415868673050615e-05,
      "loss": 0.3356,
      "step": 786
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8236862421035767,
      "learning_rate": 6.411308709530323e-05,
      "loss": 0.3394,
      "step": 787
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8534393906593323,
      "learning_rate": 6.406748746010032e-05,
      "loss": 0.3623,
      "step": 788
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.7958694100379944,
      "learning_rate": 6.402188782489741e-05,
      "loss": 0.3541,
      "step": 789
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8262916803359985,
      "learning_rate": 6.397628818969449e-05,
      "loss": 0.3317,
      "step": 790
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.36867284774780273,
      "eval_runtime": 21.0978,
      "eval_samples_per_second": 1.327,
      "eval_steps_per_second": 0.332,
      "step": 790
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8441379070281982,
      "learning_rate": 6.393068855449157e-05,
      "loss": 0.3496,
      "step": 791
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8390480279922485,
      "learning_rate": 6.388508891928866e-05,
      "loss": 0.3432,
      "step": 792
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.9012981653213501,
      "learning_rate": 6.383948928408572e-05,
      "loss": 0.3722,
      "step": 793
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8114961981773376,
      "learning_rate": 6.37938896488828e-05,
      "loss": 0.3315,
      "step": 794
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8510379791259766,
      "learning_rate": 6.374829001367989e-05,
      "loss": 0.3415,
      "step": 795
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8974173069000244,
      "learning_rate": 6.370269037847697e-05,
      "loss": 0.3659,
      "step": 796
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8199520111083984,
      "learning_rate": 6.365709074327406e-05,
      "loss": 0.3491,
      "step": 797
    },
    {
      "epoch": 0.36,
      "grad_norm": 1.0816798210144043,
      "learning_rate": 6.361149110807115e-05,
      "loss": 0.3439,
      "step": 798
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8767193555831909,
      "learning_rate": 6.356589147286823e-05,
      "loss": 0.3471,
      "step": 799
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.8754706978797913,
      "learning_rate": 6.35202918376653e-05,
      "loss": 0.3699,
      "step": 800
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.36849379539489746,
      "eval_runtime": 21.2125,
      "eval_samples_per_second": 1.32,
      "eval_steps_per_second": 0.33,
      "step": 800
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8332600593566895,
      "learning_rate": 6.347469220246238e-05,
      "loss": 0.3322,
      "step": 801
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.9050359129905701,
      "learning_rate": 6.342909256725946e-05,
      "loss": 0.3424,
      "step": 802
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.9001706838607788,
      "learning_rate": 6.338349293205654e-05,
      "loss": 0.3548,
      "step": 803
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8502199053764343,
      "learning_rate": 6.333789329685362e-05,
      "loss": 0.3367,
      "step": 804
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8597239851951599,
      "learning_rate": 6.329229366165072e-05,
      "loss": 0.3588,
      "step": 805
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8567646741867065,
      "learning_rate": 6.32466940264478e-05,
      "loss": 0.3924,
      "step": 806
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8374874591827393,
      "learning_rate": 6.320109439124488e-05,
      "loss": 0.3451,
      "step": 807
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8712239861488342,
      "learning_rate": 6.315549475604195e-05,
      "loss": 0.3567,
      "step": 808
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8838446140289307,
      "learning_rate": 6.310989512083903e-05,
      "loss": 0.3714,
      "step": 809
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8462404012680054,
      "learning_rate": 6.306429548563611e-05,
      "loss": 0.3504,
      "step": 810
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.369089812040329,
      "eval_runtime": 21.3098,
      "eval_samples_per_second": 1.314,
      "eval_steps_per_second": 0.328,
      "step": 810
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8680486679077148,
      "learning_rate": 6.30186958504332e-05,
      "loss": 0.353,
      "step": 811
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.854362964630127,
      "learning_rate": 6.297309621523028e-05,
      "loss": 0.3382,
      "step": 812
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8963435888290405,
      "learning_rate": 6.292749658002737e-05,
      "loss": 0.3514,
      "step": 813
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8164211511611938,
      "learning_rate": 6.288189694482445e-05,
      "loss": 0.3281,
      "step": 814
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8216964602470398,
      "learning_rate": 6.283629730962152e-05,
      "loss": 0.36,
      "step": 815
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8878861665725708,
      "learning_rate": 6.27906976744186e-05,
      "loss": 0.3826,
      "step": 816
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.795782208442688,
      "learning_rate": 6.274509803921569e-05,
      "loss": 0.3358,
      "step": 817
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.9114422798156738,
      "learning_rate": 6.269949840401277e-05,
      "loss": 0.3521,
      "step": 818
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8703455924987793,
      "learning_rate": 6.265389876880985e-05,
      "loss": 0.3368,
      "step": 819
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8820777535438538,
      "learning_rate": 6.260829913360693e-05,
      "loss": 0.3485,
      "step": 820
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.3687816262245178,
      "eval_runtime": 21.1501,
      "eval_samples_per_second": 1.324,
      "eval_steps_per_second": 0.331,
      "step": 820
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8445407748222351,
      "learning_rate": 6.256269949840403e-05,
      "loss": 0.3288,
      "step": 821
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.8472913503646851,
      "learning_rate": 6.25170998632011e-05,
      "loss": 0.3427,
      "step": 822
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.911773681640625,
      "learning_rate": 6.247150022799818e-05,
      "loss": 0.3571,
      "step": 823
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.889894962310791,
      "learning_rate": 6.242590059279526e-05,
      "loss": 0.3629,
      "step": 824
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8566384315490723,
      "learning_rate": 6.238030095759234e-05,
      "loss": 0.3582,
      "step": 825
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8500192165374756,
      "learning_rate": 6.233470132238942e-05,
      "loss": 0.3375,
      "step": 826
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8549200892448425,
      "learning_rate": 6.22891016871865e-05,
      "loss": 0.3487,
      "step": 827
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8865132927894592,
      "learning_rate": 6.224350205198358e-05,
      "loss": 0.3625,
      "step": 828
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.898220956325531,
      "learning_rate": 6.219790241678068e-05,
      "loss": 0.3657,
      "step": 829
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8199746012687683,
      "learning_rate": 6.215230278157775e-05,
      "loss": 0.3239,
      "step": 830
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.36817699670791626,
      "eval_runtime": 21.3091,
      "eval_samples_per_second": 1.314,
      "eval_steps_per_second": 0.328,
      "step": 830
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8350242376327515,
      "learning_rate": 6.210670314637483e-05,
      "loss": 0.3443,
      "step": 831
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8211762309074402,
      "learning_rate": 6.206110351117191e-05,
      "loss": 0.3478,
      "step": 832
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8550103306770325,
      "learning_rate": 6.201550387596899e-05,
      "loss": 0.362,
      "step": 833
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.840576708316803,
      "learning_rate": 6.196990424076607e-05,
      "loss": 0.3629,
      "step": 834
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8360455632209778,
      "learning_rate": 6.192430460556316e-05,
      "loss": 0.3462,
      "step": 835
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8622804284095764,
      "learning_rate": 6.187870497036024e-05,
      "loss": 0.3582,
      "step": 836
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8452496528625488,
      "learning_rate": 6.183310533515732e-05,
      "loss": 0.351,
      "step": 837
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.7803338766098022,
      "learning_rate": 6.17875056999544e-05,
      "loss": 0.3237,
      "step": 838
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.879328727722168,
      "learning_rate": 6.174190606475148e-05,
      "loss": 0.3666,
      "step": 839
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8226670622825623,
      "learning_rate": 6.169630642954857e-05,
      "loss": 0.3342,
      "step": 840
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.36737722158432007,
      "eval_runtime": 21.1249,
      "eval_samples_per_second": 1.325,
      "eval_steps_per_second": 0.331,
      "step": 840
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.912907600402832,
      "learning_rate": 6.165070679434565e-05,
      "loss": 0.3701,
      "step": 841
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8975276947021484,
      "learning_rate": 6.160510715914273e-05,
      "loss": 0.3395,
      "step": 842
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8709104657173157,
      "learning_rate": 6.155950752393981e-05,
      "loss": 0.3352,
      "step": 843
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.8861703872680664,
      "learning_rate": 6.151390788873689e-05,
      "loss": 0.3757,
      "step": 844
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8633649349212646,
      "learning_rate": 6.146830825353397e-05,
      "loss": 0.3439,
      "step": 845
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8183202743530273,
      "learning_rate": 6.142270861833106e-05,
      "loss": 0.3525,
      "step": 846
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.9185418486595154,
      "learning_rate": 6.137710898312814e-05,
      "loss": 0.3833,
      "step": 847
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8235242962837219,
      "learning_rate": 6.133150934792522e-05,
      "loss": 0.3469,
      "step": 848
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8366368412971497,
      "learning_rate": 6.12859097127223e-05,
      "loss": 0.3419,
      "step": 849
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.834038496017456,
      "learning_rate": 6.124031007751938e-05,
      "loss": 0.3404,
      "step": 850
    },
    {
      "epoch": 0.39,
      "eval_loss": 0.3677498698234558,
      "eval_runtime": 21.1975,
      "eval_samples_per_second": 1.321,
      "eval_steps_per_second": 0.33,
      "step": 850
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8480759859085083,
      "learning_rate": 6.119471044231646e-05,
      "loss": 0.3562,
      "step": 851
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8846428990364075,
      "learning_rate": 6.114911080711355e-05,
      "loss": 0.3646,
      "step": 852
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8484827876091003,
      "learning_rate": 6.110351117191063e-05,
      "loss": 0.334,
      "step": 853
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8655832409858704,
      "learning_rate": 6.105791153670771e-05,
      "loss": 0.3497,
      "step": 854
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.9084945321083069,
      "learning_rate": 6.101231190150479e-05,
      "loss": 0.3698,
      "step": 855
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8178445100784302,
      "learning_rate": 6.096671226630187e-05,
      "loss": 0.333,
      "step": 856
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8397005796432495,
      "learning_rate": 6.0921112631098955e-05,
      "loss": 0.3407,
      "step": 857
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8644444942474365,
      "learning_rate": 6.0875512995896036e-05,
      "loss": 0.3502,
      "step": 858
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8999423384666443,
      "learning_rate": 6.082991336069311e-05,
      "loss": 0.3529,
      "step": 859
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8924650549888611,
      "learning_rate": 6.078431372549019e-05,
      "loss": 0.3616,
      "step": 860
    },
    {
      "epoch": 0.39,
      "eval_loss": 0.3670370280742645,
      "eval_runtime": 21.1232,
      "eval_samples_per_second": 1.326,
      "eval_steps_per_second": 0.331,
      "step": 860
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8417632579803467,
      "learning_rate": 6.073871409028729e-05,
      "loss": 0.3394,
      "step": 861
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8309604525566101,
      "learning_rate": 6.069311445508436e-05,
      "loss": 0.3387,
      "step": 862
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8310514688491821,
      "learning_rate": 6.0647514819881445e-05,
      "loss": 0.3473,
      "step": 863
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8718897104263306,
      "learning_rate": 6.060191518467853e-05,
      "loss": 0.3725,
      "step": 864
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.9132295250892639,
      "learning_rate": 6.055631554947561e-05,
      "loss": 0.3639,
      "step": 865
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.8623649477958679,
      "learning_rate": 6.0510715914272683e-05,
      "loss": 0.3434,
      "step": 866
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8534384369850159,
      "learning_rate": 6.0465116279069765e-05,
      "loss": 0.3499,
      "step": 867
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8440150022506714,
      "learning_rate": 6.041951664386685e-05,
      "loss": 0.3534,
      "step": 868
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8526747822761536,
      "learning_rate": 6.0373917008663935e-05,
      "loss": 0.3524,
      "step": 869
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8256170153617859,
      "learning_rate": 6.032831737346102e-05,
      "loss": 0.3438,
      "step": 870
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.3662932515144348,
      "eval_runtime": 21.3199,
      "eval_samples_per_second": 1.313,
      "eval_steps_per_second": 0.328,
      "step": 870
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8334060311317444,
      "learning_rate": 6.02827177382581e-05,
      "loss": 0.357,
      "step": 871
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8603335618972778,
      "learning_rate": 6.023711810305518e-05,
      "loss": 0.3469,
      "step": 872
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.7767968773841858,
      "learning_rate": 6.019151846785226e-05,
      "loss": 0.3091,
      "step": 873
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8094191551208496,
      "learning_rate": 6.014591883264934e-05,
      "loss": 0.3479,
      "step": 874
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.819817066192627,
      "learning_rate": 6.010031919744642e-05,
      "loss": 0.3508,
      "step": 875
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.841957151889801,
      "learning_rate": 6.00547195622435e-05,
      "loss": 0.3398,
      "step": 876
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.86558997631073,
      "learning_rate": 6.000911992704059e-05,
      "loss": 0.3334,
      "step": 877
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8331573605537415,
      "learning_rate": 5.996352029183767e-05,
      "loss": 0.3667,
      "step": 878
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8304200768470764,
      "learning_rate": 5.991792065663475e-05,
      "loss": 0.3468,
      "step": 879
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8669402599334717,
      "learning_rate": 5.9872321021431834e-05,
      "loss": 0.3498,
      "step": 880
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.36670106649398804,
      "eval_runtime": 21.1982,
      "eval_samples_per_second": 1.321,
      "eval_steps_per_second": 0.33,
      "step": 880
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8461664915084839,
      "learning_rate": 5.982672138622891e-05,
      "loss": 0.3358,
      "step": 881
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8536818623542786,
      "learning_rate": 5.978112175102599e-05,
      "loss": 0.3671,
      "step": 882
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8542598485946655,
      "learning_rate": 5.973552211582307e-05,
      "loss": 0.351,
      "step": 883
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.84647136926651,
      "learning_rate": 5.9689922480620155e-05,
      "loss": 0.3653,
      "step": 884
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8366497159004211,
      "learning_rate": 5.964432284541724e-05,
      "loss": 0.3389,
      "step": 885
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8828349709510803,
      "learning_rate": 5.9598723210214325e-05,
      "loss": 0.3747,
      "step": 886
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8899936079978943,
      "learning_rate": 5.9553123575011407e-05,
      "loss": 0.361,
      "step": 887
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8368967771530151,
      "learning_rate": 5.950752393980849e-05,
      "loss": 0.3381,
      "step": 888
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8967677354812622,
      "learning_rate": 5.946192430460556e-05,
      "loss": 0.3591,
      "step": 889
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.7917211055755615,
      "learning_rate": 5.9416324669402645e-05,
      "loss": 0.3389,
      "step": 890
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.3663075268268585,
      "eval_runtime": 21.5147,
      "eval_samples_per_second": 1.301,
      "eval_steps_per_second": 0.325,
      "step": 890
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8735987544059753,
      "learning_rate": 5.937072503419973e-05,
      "loss": 0.3529,
      "step": 891
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8549131155014038,
      "learning_rate": 5.932512539899681e-05,
      "loss": 0.3788,
      "step": 892
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.836391270160675,
      "learning_rate": 5.92795257637939e-05,
      "loss": 0.341,
      "step": 893
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8679954409599304,
      "learning_rate": 5.923392612859098e-05,
      "loss": 0.3677,
      "step": 894
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8982836604118347,
      "learning_rate": 5.918832649338806e-05,
      "loss": 0.3697,
      "step": 895
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8966107964515686,
      "learning_rate": 5.9142726858185135e-05,
      "loss": 0.3539,
      "step": 896
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8517934083938599,
      "learning_rate": 5.909712722298222e-05,
      "loss": 0.3364,
      "step": 897
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8622394800186157,
      "learning_rate": 5.90515275877793e-05,
      "loss": 0.3373,
      "step": 898
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8497902750968933,
      "learning_rate": 5.900592795257638e-05,
      "loss": 0.3402,
      "step": 899
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8614276051521301,
      "learning_rate": 5.8960328317373456e-05,
      "loss": 0.3474,
      "step": 900
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.36591216921806335,
      "eval_runtime": 21.185,
      "eval_samples_per_second": 1.322,
      "eval_steps_per_second": 0.33,
      "step": 900
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8814240097999573,
      "learning_rate": 5.891472868217055e-05,
      "loss": 0.3623,
      "step": 901
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8524793982505798,
      "learning_rate": 5.886912904696763e-05,
      "loss": 0.3439,
      "step": 902
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8247084021568298,
      "learning_rate": 5.882352941176471e-05,
      "loss": 0.3596,
      "step": 903
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.812872052192688,
      "learning_rate": 5.877792977656179e-05,
      "loss": 0.3365,
      "step": 904
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8451302647590637,
      "learning_rate": 5.873233014135887e-05,
      "loss": 0.3435,
      "step": 905
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8993464708328247,
      "learning_rate": 5.868673050615595e-05,
      "loss": 0.3688,
      "step": 906
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8517639636993408,
      "learning_rate": 5.8641130870953034e-05,
      "loss": 0.3709,
      "step": 907
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8076426386833191,
      "learning_rate": 5.859553123575011e-05,
      "loss": 0.3513,
      "step": 908
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.8321362137794495,
      "learning_rate": 5.8549931600547205e-05,
      "loss": 0.3362,
      "step": 909
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.841407299041748,
      "learning_rate": 5.8504331965344286e-05,
      "loss": 0.3402,
      "step": 910
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.36606305837631226,
      "eval_runtime": 21.2447,
      "eval_samples_per_second": 1.318,
      "eval_steps_per_second": 0.329,
      "step": 910
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.873389720916748,
      "learning_rate": 5.845873233014136e-05,
      "loss": 0.3674,
      "step": 911
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8563045859336853,
      "learning_rate": 5.841313269493844e-05,
      "loss": 0.3587,
      "step": 912
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8238916993141174,
      "learning_rate": 5.8367533059735525e-05,
      "loss": 0.3521,
      "step": 913
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8294292688369751,
      "learning_rate": 5.8321933424532607e-05,
      "loss": 0.3614,
      "step": 914
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8287543058395386,
      "learning_rate": 5.827633378932968e-05,
      "loss": 0.3596,
      "step": 915
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8150818347930908,
      "learning_rate": 5.823073415412676e-05,
      "loss": 0.3515,
      "step": 916
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8092970252037048,
      "learning_rate": 5.818513451892386e-05,
      "loss": 0.329,
      "step": 917
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8499705195426941,
      "learning_rate": 5.8139534883720933e-05,
      "loss": 0.3446,
      "step": 918
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8703744411468506,
      "learning_rate": 5.8093935248518015e-05,
      "loss": 0.3618,
      "step": 919
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8624708652496338,
      "learning_rate": 5.80483356133151e-05,
      "loss": 0.3596,
      "step": 920
    },
    {
      "epoch": 0.42,
      "eval_loss": 0.3661345839500427,
      "eval_runtime": 21.1403,
      "eval_samples_per_second": 1.324,
      "eval_steps_per_second": 0.331,
      "step": 920
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.811722993850708,
      "learning_rate": 5.800273597811218e-05,
      "loss": 0.3702,
      "step": 921
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8618665337562561,
      "learning_rate": 5.795713634290926e-05,
      "loss": 0.361,
      "step": 922
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8876213431358337,
      "learning_rate": 5.7911536707706335e-05,
      "loss": 0.3639,
      "step": 923
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.842745840549469,
      "learning_rate": 5.786593707250342e-05,
      "loss": 0.3543,
      "step": 924
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8399864435195923,
      "learning_rate": 5.782033743730051e-05,
      "loss": 0.3282,
      "step": 925
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8489199280738831,
      "learning_rate": 5.777473780209759e-05,
      "loss": 0.3351,
      "step": 926
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8790804743766785,
      "learning_rate": 5.772913816689467e-05,
      "loss": 0.3722,
      "step": 927
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8245604634284973,
      "learning_rate": 5.768353853169175e-05,
      "loss": 0.3429,
      "step": 928
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.7950709462165833,
      "learning_rate": 5.763793889648883e-05,
      "loss": 0.3395,
      "step": 929
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.810104250907898,
      "learning_rate": 5.759233926128591e-05,
      "loss": 0.3478,
      "step": 930
    },
    {
      "epoch": 0.42,
      "eval_loss": 0.3656521439552307,
      "eval_runtime": 21.2315,
      "eval_samples_per_second": 1.319,
      "eval_steps_per_second": 0.33,
      "step": 930
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8308596611022949,
      "learning_rate": 5.754673962608299e-05,
      "loss": 0.3392,
      "step": 931
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.8465328812599182,
      "learning_rate": 5.750113999088007e-05,
      "loss": 0.3587,
      "step": 932
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8508004546165466,
      "learning_rate": 5.745554035567716e-05,
      "loss": 0.3449,
      "step": 933
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.9344462156295776,
      "learning_rate": 5.740994072047424e-05,
      "loss": 0.3814,
      "step": 934
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8226639032363892,
      "learning_rate": 5.736434108527132e-05,
      "loss": 0.3393,
      "step": 935
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.903837263584137,
      "learning_rate": 5.7318741450068405e-05,
      "loss": 0.3545,
      "step": 936
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.927081823348999,
      "learning_rate": 5.7273141814865486e-05,
      "loss": 0.3647,
      "step": 937
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8148213028907776,
      "learning_rate": 5.722754217966256e-05,
      "loss": 0.3397,
      "step": 938
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8732878565788269,
      "learning_rate": 5.718194254445964e-05,
      "loss": 0.3533,
      "step": 939
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8293365836143494,
      "learning_rate": 5.7136342909256725e-05,
      "loss": 0.3572,
      "step": 940
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.3642878532409668,
      "eval_runtime": 18.6197,
      "eval_samples_per_second": 1.504,
      "eval_steps_per_second": 0.376,
      "step": 940
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8569697141647339,
      "learning_rate": 5.709074327405381e-05,
      "loss": 0.3527,
      "step": 941
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8310784101486206,
      "learning_rate": 5.7045143638850895e-05,
      "loss": 0.3418,
      "step": 942
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.803471028804779,
      "learning_rate": 5.699954400364798e-05,
      "loss": 0.3377,
      "step": 943
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8265700936317444,
      "learning_rate": 5.695394436844506e-05,
      "loss": 0.3421,
      "step": 944
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8034089803695679,
      "learning_rate": 5.6908344733242133e-05,
      "loss": 0.3437,
      "step": 945
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8499231934547424,
      "learning_rate": 5.6862745098039215e-05,
      "loss": 0.3532,
      "step": 946
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.7843233942985535,
      "learning_rate": 5.68171454628363e-05,
      "loss": 0.3402,
      "step": 947
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.7977408170700073,
      "learning_rate": 5.677154582763338e-05,
      "loss": 0.3481,
      "step": 948
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8667237758636475,
      "learning_rate": 5.672594619243047e-05,
      "loss": 0.3843,
      "step": 949
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8898354172706604,
      "learning_rate": 5.668034655722755e-05,
      "loss": 0.3631,
      "step": 950
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.36420196294784546,
      "eval_runtime": 19.1437,
      "eval_samples_per_second": 1.463,
      "eval_steps_per_second": 0.366,
      "step": 950
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8387503027915955,
      "learning_rate": 5.663474692202463e-05,
      "loss": 0.3391,
      "step": 951
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8411070108413696,
      "learning_rate": 5.6589147286821706e-05,
      "loss": 0.3699,
      "step": 952
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8362002968788147,
      "learning_rate": 5.654354765161879e-05,
      "loss": 0.338,
      "step": 953
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.8469303846359253,
      "learning_rate": 5.649794801641587e-05,
      "loss": 0.363,
      "step": 954
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8548648953437805,
      "learning_rate": 5.645234838121295e-05,
      "loss": 0.3344,
      "step": 955
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8648215532302856,
      "learning_rate": 5.640674874601003e-05,
      "loss": 0.3636,
      "step": 956
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8931896686553955,
      "learning_rate": 5.636114911080712e-05,
      "loss": 0.3479,
      "step": 957
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8487157821655273,
      "learning_rate": 5.63155494756042e-05,
      "loss": 0.3603,
      "step": 958
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.852799117565155,
      "learning_rate": 5.6269949840401285e-05,
      "loss": 0.3711,
      "step": 959
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8263891339302063,
      "learning_rate": 5.622435020519836e-05,
      "loss": 0.3574,
      "step": 960
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.363812118768692,
      "eval_runtime": 18.6019,
      "eval_samples_per_second": 1.505,
      "eval_steps_per_second": 0.376,
      "step": 960
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8231980800628662,
      "learning_rate": 5.617875056999544e-05,
      "loss": 0.3317,
      "step": 961
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8345970511436462,
      "learning_rate": 5.613315093479252e-05,
      "loss": 0.346,
      "step": 962
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8778778910636902,
      "learning_rate": 5.6087551299589605e-05,
      "loss": 0.341,
      "step": 963
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.7894421219825745,
      "learning_rate": 5.604195166438668e-05,
      "loss": 0.3379,
      "step": 964
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.895371675491333,
      "learning_rate": 5.5996352029183775e-05,
      "loss": 0.3797,
      "step": 965
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8161599636077881,
      "learning_rate": 5.595075239398086e-05,
      "loss": 0.3472,
      "step": 966
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8689946532249451,
      "learning_rate": 5.590515275877793e-05,
      "loss": 0.3463,
      "step": 967
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.7838358879089355,
      "learning_rate": 5.585955312357501e-05,
      "loss": 0.32,
      "step": 968
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8021233677864075,
      "learning_rate": 5.5813953488372095e-05,
      "loss": 0.3386,
      "step": 969
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8686328530311584,
      "learning_rate": 5.576835385316918e-05,
      "loss": 0.3388,
      "step": 970
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.36350157856941223,
      "eval_runtime": 18.792,
      "eval_samples_per_second": 1.49,
      "eval_steps_per_second": 0.372,
      "step": 970
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8454912304878235,
      "learning_rate": 5.572275421796626e-05,
      "loss": 0.3627,
      "step": 971
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8108564019203186,
      "learning_rate": 5.5677154582763333e-05,
      "loss": 0.347,
      "step": 972
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8135057091712952,
      "learning_rate": 5.563155494756043e-05,
      "loss": 0.3557,
      "step": 973
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8536540865898132,
      "learning_rate": 5.558595531235751e-05,
      "loss": 0.3771,
      "step": 974
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.8181318640708923,
      "learning_rate": 5.5540355677154585e-05,
      "loss": 0.3505,
      "step": 975
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.7815612554550171,
      "learning_rate": 5.549475604195167e-05,
      "loss": 0.3259,
      "step": 976
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8374220728874207,
      "learning_rate": 5.544915640674875e-05,
      "loss": 0.3544,
      "step": 977
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8432349562644958,
      "learning_rate": 5.540355677154583e-05,
      "loss": 0.3551,
      "step": 978
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.824934720993042,
      "learning_rate": 5.5357957136342906e-05,
      "loss": 0.3477,
      "step": 979
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8221052289009094,
      "learning_rate": 5.531235750113999e-05,
      "loss": 0.3248,
      "step": 980
    },
    {
      "epoch": 0.45,
      "eval_loss": 0.3641625940799713,
      "eval_runtime": 18.8993,
      "eval_samples_per_second": 1.482,
      "eval_steps_per_second": 0.37,
      "step": 980
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.7826463580131531,
      "learning_rate": 5.526675786593708e-05,
      "loss": 0.3287,
      "step": 981
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.7870301008224487,
      "learning_rate": 5.522115823073416e-05,
      "loss": 0.3407,
      "step": 982
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8751205205917358,
      "learning_rate": 5.517555859553124e-05,
      "loss": 0.3495,
      "step": 983
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8503699898719788,
      "learning_rate": 5.512995896032832e-05,
      "loss": 0.3596,
      "step": 984
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8469080328941345,
      "learning_rate": 5.50843593251254e-05,
      "loss": 0.3334,
      "step": 985
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8288311958312988,
      "learning_rate": 5.503875968992248e-05,
      "loss": 0.3737,
      "step": 986
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8115249276161194,
      "learning_rate": 5.499316005471956e-05,
      "loss": 0.3358,
      "step": 987
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8277326226234436,
      "learning_rate": 5.494756041951664e-05,
      "loss": 0.3303,
      "step": 988
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8234649896621704,
      "learning_rate": 5.490196078431373e-05,
      "loss": 0.3494,
      "step": 989
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8039526343345642,
      "learning_rate": 5.485636114911081e-05,
      "loss": 0.3379,
      "step": 990
    },
    {
      "epoch": 0.45,
      "eval_loss": 0.36366531252861023,
      "eval_runtime": 19.7343,
      "eval_samples_per_second": 1.419,
      "eval_steps_per_second": 0.355,
      "step": 990
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8476756811141968,
      "learning_rate": 5.481076151390789e-05,
      "loss": 0.3373,
      "step": 991
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8155856132507324,
      "learning_rate": 5.4765161878704975e-05,
      "loss": 0.3464,
      "step": 992
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8660078644752502,
      "learning_rate": 5.471956224350206e-05,
      "loss": 0.3739,
      "step": 993
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8255999684333801,
      "learning_rate": 5.467396260829913e-05,
      "loss": 0.3531,
      "step": 994
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.7718782424926758,
      "learning_rate": 5.462836297309621e-05,
      "loss": 0.3346,
      "step": 995
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8417358994483948,
      "learning_rate": 5.4582763337893295e-05,
      "loss": 0.3499,
      "step": 996
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8342063426971436,
      "learning_rate": 5.4537163702690384e-05,
      "loss": 0.3481,
      "step": 997
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.8939498066902161,
      "learning_rate": 5.4491564067487465e-05,
      "loss": 0.345,
      "step": 998
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8737196922302246,
      "learning_rate": 5.444596443228455e-05,
      "loss": 0.356,
      "step": 999
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8749281764030457,
      "learning_rate": 5.440036479708163e-05,
      "loss": 0.3428,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.3630655109882355,
      "eval_runtime": 21.6489,
      "eval_samples_per_second": 1.293,
      "eval_steps_per_second": 0.323,
      "step": 1000
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8389559388160706,
      "learning_rate": 5.4354765161878704e-05,
      "loss": 0.3527,
      "step": 1001
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.7941823601722717,
      "learning_rate": 5.4309165526675785e-05,
      "loss": 0.3312,
      "step": 1002
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8154469132423401,
      "learning_rate": 5.426356589147287e-05,
      "loss": 0.3512,
      "step": 1003
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8760425448417664,
      "learning_rate": 5.421796625626995e-05,
      "loss": 0.3367,
      "step": 1004
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.821310818195343,
      "learning_rate": 5.417236662106704e-05,
      "loss": 0.353,
      "step": 1005
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8528419137001038,
      "learning_rate": 5.412676698586412e-05,
      "loss": 0.354,
      "step": 1006
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8211894035339355,
      "learning_rate": 5.40811673506612e-05,
      "loss": 0.3332,
      "step": 1007
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8660132884979248,
      "learning_rate": 5.403556771545828e-05,
      "loss": 0.3386,
      "step": 1008
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8995015621185303,
      "learning_rate": 5.398996808025536e-05,
      "loss": 0.3623,
      "step": 1009
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8141841888427734,
      "learning_rate": 5.394436844505244e-05,
      "loss": 0.3643,
      "step": 1010
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.3631010949611664,
      "eval_runtime": 20.9081,
      "eval_samples_per_second": 1.339,
      "eval_steps_per_second": 0.335,
      "step": 1010
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.828433632850647,
      "learning_rate": 5.389876880984952e-05,
      "loss": 0.3394,
      "step": 1011
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8452778458595276,
      "learning_rate": 5.38531691746466e-05,
      "loss": 0.35,
      "step": 1012
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8223903775215149,
      "learning_rate": 5.380756953944369e-05,
      "loss": 0.3459,
      "step": 1013
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8182603716850281,
      "learning_rate": 5.376196990424077e-05,
      "loss": 0.3388,
      "step": 1014
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8223605155944824,
      "learning_rate": 5.3716370269037855e-05,
      "loss": 0.3477,
      "step": 1015
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8775902390480042,
      "learning_rate": 5.367077063383493e-05,
      "loss": 0.348,
      "step": 1016
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8247773051261902,
      "learning_rate": 5.362517099863201e-05,
      "loss": 0.343,
      "step": 1017
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8566569089889526,
      "learning_rate": 5.357957136342909e-05,
      "loss": 0.3296,
      "step": 1018
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.9025068879127502,
      "learning_rate": 5.3533971728226175e-05,
      "loss": 0.3685,
      "step": 1019
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.8382214307785034,
      "learning_rate": 5.348837209302326e-05,
      "loss": 0.3598,
      "step": 1020
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.3636544644832611,
      "eval_runtime": 20.2937,
      "eval_samples_per_second": 1.38,
      "eval_steps_per_second": 0.345,
      "step": 1020
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8766863942146301,
      "learning_rate": 5.3442772457820345e-05,
      "loss": 0.3596,
      "step": 1021
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8422759771347046,
      "learning_rate": 5.339717282261743e-05,
      "loss": 0.3441,
      "step": 1022
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8558328151702881,
      "learning_rate": 5.335157318741451e-05,
      "loss": 0.3508,
      "step": 1023
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8047720789909363,
      "learning_rate": 5.3305973552211584e-05,
      "loss": 0.3365,
      "step": 1024
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8321728110313416,
      "learning_rate": 5.3260373917008665e-05,
      "loss": 0.3579,
      "step": 1025
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8026608824729919,
      "learning_rate": 5.321477428180575e-05,
      "loss": 0.3358,
      "step": 1026
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8313252329826355,
      "learning_rate": 5.316917464660283e-05,
      "loss": 0.3308,
      "step": 1027
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.858326256275177,
      "learning_rate": 5.3123575011399904e-05,
      "loss": 0.3489,
      "step": 1028
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8693019151687622,
      "learning_rate": 5.3077975376197e-05,
      "loss": 0.3528,
      "step": 1029
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8624293804168701,
      "learning_rate": 5.303237574099408e-05,
      "loss": 0.3656,
      "step": 1030
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.36372342705726624,
      "eval_runtime": 20.3094,
      "eval_samples_per_second": 1.379,
      "eval_steps_per_second": 0.345,
      "step": 1030
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8259044289588928,
      "learning_rate": 5.2986776105791156e-05,
      "loss": 0.3261,
      "step": 1031
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8822616338729858,
      "learning_rate": 5.294117647058824e-05,
      "loss": 0.3518,
      "step": 1032
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8377732634544373,
      "learning_rate": 5.289557683538532e-05,
      "loss": 0.3557,
      "step": 1033
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8176842331886292,
      "learning_rate": 5.28499772001824e-05,
      "loss": 0.3542,
      "step": 1034
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8251992464065552,
      "learning_rate": 5.2804377564979476e-05,
      "loss": 0.3672,
      "step": 1035
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8290631175041199,
      "learning_rate": 5.275877792977656e-05,
      "loss": 0.3456,
      "step": 1036
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8259645700454712,
      "learning_rate": 5.271317829457365e-05,
      "loss": 0.3536,
      "step": 1037
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8718307018280029,
      "learning_rate": 5.266757865937073e-05,
      "loss": 0.3706,
      "step": 1038
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8760273456573486,
      "learning_rate": 5.262197902416781e-05,
      "loss": 0.3639,
      "step": 1039
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8457003235816956,
      "learning_rate": 5.257637938896489e-05,
      "loss": 0.3671,
      "step": 1040
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.3634140193462372,
      "eval_runtime": 20.5539,
      "eval_samples_per_second": 1.362,
      "eval_steps_per_second": 0.341,
      "step": 1040
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8008033037185669,
      "learning_rate": 5.253077975376197e-05,
      "loss": 0.3534,
      "step": 1041
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.8359259366989136,
      "learning_rate": 5.2485180118559055e-05,
      "loss": 0.342,
      "step": 1042
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8333999514579773,
      "learning_rate": 5.243958048335613e-05,
      "loss": 0.3505,
      "step": 1043
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8984323143959045,
      "learning_rate": 5.239398084815321e-05,
      "loss": 0.3435,
      "step": 1044
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.7493898272514343,
      "learning_rate": 5.234838121295031e-05,
      "loss": 0.3247,
      "step": 1045
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8557037115097046,
      "learning_rate": 5.230278157774738e-05,
      "loss": 0.3829,
      "step": 1046
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8355690836906433,
      "learning_rate": 5.2257181942544463e-05,
      "loss": 0.346,
      "step": 1047
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8401297330856323,
      "learning_rate": 5.2211582307341545e-05,
      "loss": 0.3598,
      "step": 1048
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8262518644332886,
      "learning_rate": 5.216598267213863e-05,
      "loss": 0.3418,
      "step": 1049
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.7955124974250793,
      "learning_rate": 5.21203830369357e-05,
      "loss": 0.334,
      "step": 1050
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.36403360962867737,
      "eval_runtime": 20.7565,
      "eval_samples_per_second": 1.349,
      "eval_steps_per_second": 0.337,
      "step": 1050
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8252997994422913,
      "learning_rate": 5.2074783401732784e-05,
      "loss": 0.3431,
      "step": 1051
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8318145275115967,
      "learning_rate": 5.2029183766529865e-05,
      "loss": 0.3486,
      "step": 1052
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8297901153564453,
      "learning_rate": 5.1983584131326954e-05,
      "loss": 0.3472,
      "step": 1053
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.7829481959342957,
      "learning_rate": 5.1937984496124036e-05,
      "loss": 0.3517,
      "step": 1054
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8852716684341431,
      "learning_rate": 5.189238486092112e-05,
      "loss": 0.3607,
      "step": 1055
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8906071186065674,
      "learning_rate": 5.18467852257182e-05,
      "loss": 0.3742,
      "step": 1056
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8188351392745972,
      "learning_rate": 5.180118559051528e-05,
      "loss": 0.3479,
      "step": 1057
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8125540018081665,
      "learning_rate": 5.1755585955312356e-05,
      "loss": 0.3599,
      "step": 1058
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8210964202880859,
      "learning_rate": 5.170998632010944e-05,
      "loss": 0.349,
      "step": 1059
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.856060802936554,
      "learning_rate": 5.166438668490652e-05,
      "loss": 0.3565,
      "step": 1060
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.3636583387851715,
      "eval_runtime": 21.8196,
      "eval_samples_per_second": 1.283,
      "eval_steps_per_second": 0.321,
      "step": 1060
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8330450654029846,
      "learning_rate": 5.161878704970361e-05,
      "loss": 0.3296,
      "step": 1061
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8015806078910828,
      "learning_rate": 5.157318741450069e-05,
      "loss": 0.3334,
      "step": 1062
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.8128681182861328,
      "learning_rate": 5.152758777929777e-05,
      "loss": 0.3382,
      "step": 1063
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8020748496055603,
      "learning_rate": 5.148198814409485e-05,
      "loss": 0.3409,
      "step": 1064
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.843084454536438,
      "learning_rate": 5.143638850889193e-05,
      "loss": 0.3456,
      "step": 1065
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.7976772785186768,
      "learning_rate": 5.139078887368901e-05,
      "loss": 0.343,
      "step": 1066
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8992964029312134,
      "learning_rate": 5.134518923848609e-05,
      "loss": 0.3494,
      "step": 1067
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.9563857913017273,
      "learning_rate": 5.129958960328317e-05,
      "loss": 0.3546,
      "step": 1068
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8422328233718872,
      "learning_rate": 5.125398996808026e-05,
      "loss": 0.3622,
      "step": 1069
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8861860036849976,
      "learning_rate": 5.120839033287734e-05,
      "loss": 0.3698,
      "step": 1070
    },
    {
      "epoch": 0.49,
      "eval_loss": 0.3621842563152313,
      "eval_runtime": 21.5719,
      "eval_samples_per_second": 1.298,
      "eval_steps_per_second": 0.324,
      "step": 1070
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8619980216026306,
      "learning_rate": 5.1162790697674425e-05,
      "loss": 0.3595,
      "step": 1071
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8314919471740723,
      "learning_rate": 5.111719106247151e-05,
      "loss": 0.3578,
      "step": 1072
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8684543371200562,
      "learning_rate": 5.107159142726858e-05,
      "loss": 0.3452,
      "step": 1073
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8409459590911865,
      "learning_rate": 5.1025991792065663e-05,
      "loss": 0.3541,
      "step": 1074
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.814139187335968,
      "learning_rate": 5.0980392156862745e-05,
      "loss": 0.352,
      "step": 1075
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.774848461151123,
      "learning_rate": 5.093479252165983e-05,
      "loss": 0.3324,
      "step": 1076
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8436389565467834,
      "learning_rate": 5.0889192886456915e-05,
      "loss": 0.3546,
      "step": 1077
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8201998472213745,
      "learning_rate": 5.0843593251254e-05,
      "loss": 0.3437,
      "step": 1078
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.834858775138855,
      "learning_rate": 5.079799361605108e-05,
      "loss": 0.3442,
      "step": 1079
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8360298275947571,
      "learning_rate": 5.0752393980848154e-05,
      "loss": 0.3461,
      "step": 1080
    },
    {
      "epoch": 0.49,
      "eval_loss": 0.3615194857120514,
      "eval_runtime": 20.521,
      "eval_samples_per_second": 1.364,
      "eval_steps_per_second": 0.341,
      "step": 1080
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8249042630195618,
      "learning_rate": 5.0706794345645236e-05,
      "loss": 0.3574,
      "step": 1081
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8017203211784363,
      "learning_rate": 5.066119471044232e-05,
      "loss": 0.3428,
      "step": 1082
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8337559700012207,
      "learning_rate": 5.06155950752394e-05,
      "loss": 0.3539,
      "step": 1083
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8060609698295593,
      "learning_rate": 5.0569995440036474e-05,
      "loss": 0.3339,
      "step": 1084
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.8409240245819092,
      "learning_rate": 5.052439580483357e-05,
      "loss": 0.3564,
      "step": 1085
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8593933582305908,
      "learning_rate": 5.047879616963065e-05,
      "loss": 0.3243,
      "step": 1086
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.853076696395874,
      "learning_rate": 5.0433196534427726e-05,
      "loss": 0.3351,
      "step": 1087
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8284323811531067,
      "learning_rate": 5.038759689922481e-05,
      "loss": 0.3655,
      "step": 1088
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8908005356788635,
      "learning_rate": 5.034199726402189e-05,
      "loss": 0.3819,
      "step": 1089
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.7974300980567932,
      "learning_rate": 5.029639762881897e-05,
      "loss": 0.3114,
      "step": 1090
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.3615086078643799,
      "eval_runtime": 20.4658,
      "eval_samples_per_second": 1.368,
      "eval_steps_per_second": 0.342,
      "step": 1090
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8122878074645996,
      "learning_rate": 5.025079799361605e-05,
      "loss": 0.3515,
      "step": 1091
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8021461367607117,
      "learning_rate": 5.020519835841313e-05,
      "loss": 0.3382,
      "step": 1092
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8411031365394592,
      "learning_rate": 5.015959872321022e-05,
      "loss": 0.3486,
      "step": 1093
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8350054025650024,
      "learning_rate": 5.0113999088007305e-05,
      "loss": 0.3644,
      "step": 1094
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8409880995750427,
      "learning_rate": 5.006839945280438e-05,
      "loss": 0.3475,
      "step": 1095
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8329557776451111,
      "learning_rate": 5.002279981760146e-05,
      "loss": 0.3403,
      "step": 1096
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.7804337739944458,
      "learning_rate": 4.997720018239854e-05,
      "loss": 0.3353,
      "step": 1097
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8142528533935547,
      "learning_rate": 4.9931600547195625e-05,
      "loss": 0.3405,
      "step": 1098
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8784443736076355,
      "learning_rate": 4.988600091199271e-05,
      "loss": 0.3688,
      "step": 1099
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8305411338806152,
      "learning_rate": 4.984040127678979e-05,
      "loss": 0.3334,
      "step": 1100
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.36101606488227844,
      "eval_runtime": 20.3924,
      "eval_samples_per_second": 1.373,
      "eval_steps_per_second": 0.343,
      "step": 1100
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.7925887703895569,
      "learning_rate": 4.979480164158687e-05,
      "loss": 0.3379,
      "step": 1101
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8849925398826599,
      "learning_rate": 4.974920200638395e-05,
      "loss": 0.3611,
      "step": 1102
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8023688197135925,
      "learning_rate": 4.9703602371181034e-05,
      "loss": 0.3452,
      "step": 1103
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.813382625579834,
      "learning_rate": 4.9658002735978115e-05,
      "loss": 0.331,
      "step": 1104
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8702760934829712,
      "learning_rate": 4.96124031007752e-05,
      "loss": 0.3497,
      "step": 1105
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8400478363037109,
      "learning_rate": 4.956680346557228e-05,
      "loss": 0.3465,
      "step": 1106
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8609248399734497,
      "learning_rate": 4.952120383036936e-05,
      "loss": 0.3288,
      "step": 1107
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.867631733417511,
      "learning_rate": 4.947560419516644e-05,
      "loss": 0.3582,
      "step": 1108
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.804740846157074,
      "learning_rate": 4.9430004559963524e-05,
      "loss": 0.3348,
      "step": 1109
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8101495504379272,
      "learning_rate": 4.93844049247606e-05,
      "loss": 0.3542,
      "step": 1110
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.3616204261779785,
      "eval_runtime": 20.7406,
      "eval_samples_per_second": 1.35,
      "eval_steps_per_second": 0.338,
      "step": 1110
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8318590521812439,
      "learning_rate": 4.933880528955769e-05,
      "loss": 0.3397,
      "step": 1111
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8036967515945435,
      "learning_rate": 4.929320565435477e-05,
      "loss": 0.349,
      "step": 1112
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8411818146705627,
      "learning_rate": 4.924760601915185e-05,
      "loss": 0.3442,
      "step": 1113
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8122812509536743,
      "learning_rate": 4.9202006383948926e-05,
      "loss": 0.3503,
      "step": 1114
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8358497023582458,
      "learning_rate": 4.9156406748746014e-05,
      "loss": 0.3533,
      "step": 1115
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.83430415391922,
      "learning_rate": 4.9110807113543096e-05,
      "loss": 0.3624,
      "step": 1116
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8298221230506897,
      "learning_rate": 4.906520747834018e-05,
      "loss": 0.3448,
      "step": 1117
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.827634871006012,
      "learning_rate": 4.901960784313725e-05,
      "loss": 0.3478,
      "step": 1118
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8373301029205322,
      "learning_rate": 4.897400820793434e-05,
      "loss": 0.3384,
      "step": 1119
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8340453505516052,
      "learning_rate": 4.892840857273142e-05,
      "loss": 0.3394,
      "step": 1120
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.36215755343437195,
      "eval_runtime": 19.4441,
      "eval_samples_per_second": 1.44,
      "eval_steps_per_second": 0.36,
      "step": 1120
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8383544683456421,
      "learning_rate": 4.88828089375285e-05,
      "loss": 0.3594,
      "step": 1121
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.7880938053131104,
      "learning_rate": 4.883720930232558e-05,
      "loss": 0.3327,
      "step": 1122
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.7821892499923706,
      "learning_rate": 4.879160966712267e-05,
      "loss": 0.352,
      "step": 1123
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8121359348297119,
      "learning_rate": 4.874601003191975e-05,
      "loss": 0.3515,
      "step": 1124
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8127467036247253,
      "learning_rate": 4.8700410396716825e-05,
      "loss": 0.3505,
      "step": 1125
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8172454833984375,
      "learning_rate": 4.865481076151391e-05,
      "loss": 0.3329,
      "step": 1126
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8340194225311279,
      "learning_rate": 4.8609211126310995e-05,
      "loss": 0.3442,
      "step": 1127
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.7899029850959778,
      "learning_rate": 4.856361149110808e-05,
      "loss": 0.3338,
      "step": 1128
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.8151934742927551,
      "learning_rate": 4.851801185590515e-05,
      "loss": 0.3408,
      "step": 1129
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8011878728866577,
      "learning_rate": 4.8472412220702234e-05,
      "loss": 0.3335,
      "step": 1130
    },
    {
      "epoch": 0.52,
      "eval_loss": 0.3620659410953522,
      "eval_runtime": 21.8046,
      "eval_samples_per_second": 1.284,
      "eval_steps_per_second": 0.321,
      "step": 1130
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8504952788352966,
      "learning_rate": 4.842681258549932e-05,
      "loss": 0.3377,
      "step": 1131
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.7817379236221313,
      "learning_rate": 4.8381212950296404e-05,
      "loss": 0.355,
      "step": 1132
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8368287086486816,
      "learning_rate": 4.833561331509348e-05,
      "loss": 0.3565,
      "step": 1133
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8199528455734253,
      "learning_rate": 4.829001367989056e-05,
      "loss": 0.3296,
      "step": 1134
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.828828752040863,
      "learning_rate": 4.824441404468765e-05,
      "loss": 0.3358,
      "step": 1135
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.7847192287445068,
      "learning_rate": 4.8198814409484724e-05,
      "loss": 0.3432,
      "step": 1136
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8164374828338623,
      "learning_rate": 4.8153214774281806e-05,
      "loss": 0.3334,
      "step": 1137
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8527355790138245,
      "learning_rate": 4.810761513907889e-05,
      "loss": 0.3697,
      "step": 1138
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.7457334399223328,
      "learning_rate": 4.8062015503875976e-05,
      "loss": 0.3177,
      "step": 1139
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8588700294494629,
      "learning_rate": 4.801641586867305e-05,
      "loss": 0.3574,
      "step": 1140
    },
    {
      "epoch": 0.52,
      "eval_loss": 0.36230775713920593,
      "eval_runtime": 21.7776,
      "eval_samples_per_second": 1.286,
      "eval_steps_per_second": 0.321,
      "step": 1140
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8982169032096863,
      "learning_rate": 4.797081623347013e-05,
      "loss": 0.3492,
      "step": 1141
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8421137928962708,
      "learning_rate": 4.7925216598267214e-05,
      "loss": 0.3535,
      "step": 1142
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.7602046132087708,
      "learning_rate": 4.78796169630643e-05,
      "loss": 0.307,
      "step": 1143
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8325467109680176,
      "learning_rate": 4.783401732786138e-05,
      "loss": 0.3388,
      "step": 1144
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8151599168777466,
      "learning_rate": 4.778841769265846e-05,
      "loss": 0.3447,
      "step": 1145
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8560123443603516,
      "learning_rate": 4.774281805745554e-05,
      "loss": 0.3515,
      "step": 1146
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8704785108566284,
      "learning_rate": 4.769721842225262e-05,
      "loss": 0.3558,
      "step": 1147
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8487253189086914,
      "learning_rate": 4.7651618787049705e-05,
      "loss": 0.3404,
      "step": 1148
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.795734703540802,
      "learning_rate": 4.7606019151846787e-05,
      "loss": 0.3337,
      "step": 1149
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.7966299653053284,
      "learning_rate": 4.756041951664387e-05,
      "loss": 0.3257,
      "step": 1150
    },
    {
      "epoch": 0.52,
      "eval_loss": 0.3623342514038086,
      "eval_runtime": 21.8906,
      "eval_samples_per_second": 1.279,
      "eval_steps_per_second": 0.32,
      "step": 1150
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.8604366183280945,
      "learning_rate": 4.751481988144095e-05,
      "loss": 0.3352,
      "step": 1151
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8704828023910522,
      "learning_rate": 4.746922024623803e-05,
      "loss": 0.3767,
      "step": 1152
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8722444176673889,
      "learning_rate": 4.7423620611035114e-05,
      "loss": 0.3706,
      "step": 1153
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8082621097564697,
      "learning_rate": 4.7378020975832195e-05,
      "loss": 0.3292,
      "step": 1154
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8296574354171753,
      "learning_rate": 4.733242134062928e-05,
      "loss": 0.3493,
      "step": 1155
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8503522872924805,
      "learning_rate": 4.728682170542636e-05,
      "loss": 0.3528,
      "step": 1156
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8259211182594299,
      "learning_rate": 4.724122207022344e-05,
      "loss": 0.368,
      "step": 1157
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.83866947889328,
      "learning_rate": 4.719562243502052e-05,
      "loss": 0.35,
      "step": 1158
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8358547687530518,
      "learning_rate": 4.7150022799817604e-05,
      "loss": 0.3318,
      "step": 1159
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8184478878974915,
      "learning_rate": 4.7104423164614686e-05,
      "loss": 0.3471,
      "step": 1160
    },
    {
      "epoch": 0.53,
      "eval_loss": 0.3614867329597473,
      "eval_runtime": 21.8624,
      "eval_samples_per_second": 1.281,
      "eval_steps_per_second": 0.32,
      "step": 1160
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.820179283618927,
      "learning_rate": 4.705882352941177e-05,
      "loss": 0.3258,
      "step": 1161
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8562389612197876,
      "learning_rate": 4.701322389420885e-05,
      "loss": 0.3589,
      "step": 1162
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.809478759765625,
      "learning_rate": 4.696762425900593e-05,
      "loss": 0.3172,
      "step": 1163
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8187942504882812,
      "learning_rate": 4.692202462380301e-05,
      "loss": 0.3636,
      "step": 1164
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8812757134437561,
      "learning_rate": 4.6876424988600094e-05,
      "loss": 0.3622,
      "step": 1165
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.93978351354599,
      "learning_rate": 4.6830825353397176e-05,
      "loss": 0.3601,
      "step": 1166
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8277170062065125,
      "learning_rate": 4.678522571819426e-05,
      "loss": 0.3563,
      "step": 1167
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8791542053222656,
      "learning_rate": 4.673962608299134e-05,
      "loss": 0.344,
      "step": 1168
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.848914384841919,
      "learning_rate": 4.669402644778842e-05,
      "loss": 0.352,
      "step": 1169
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8371862769126892,
      "learning_rate": 4.6648426812585496e-05,
      "loss": 0.343,
      "step": 1170
    },
    {
      "epoch": 0.53,
      "eval_loss": 0.3610856831073761,
      "eval_runtime": 21.8836,
      "eval_samples_per_second": 1.279,
      "eval_steps_per_second": 0.32,
      "step": 1170
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8674805164337158,
      "learning_rate": 4.6602827177382585e-05,
      "loss": 0.3488,
      "step": 1171
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8609682321548462,
      "learning_rate": 4.6557227542179666e-05,
      "loss": 0.3599,
      "step": 1172
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.8340581655502319,
      "learning_rate": 4.651162790697675e-05,
      "loss": 0.3602,
      "step": 1173
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8438746333122253,
      "learning_rate": 4.646602827177382e-05,
      "loss": 0.3399,
      "step": 1174
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8461584448814392,
      "learning_rate": 4.642042863657091e-05,
      "loss": 0.3349,
      "step": 1175
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.7988734245300293,
      "learning_rate": 4.637482900136799e-05,
      "loss": 0.3482,
      "step": 1176
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8099560737609863,
      "learning_rate": 4.6329229366165075e-05,
      "loss": 0.3378,
      "step": 1177
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.854646623134613,
      "learning_rate": 4.628362973096215e-05,
      "loss": 0.3622,
      "step": 1178
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8347275853157043,
      "learning_rate": 4.623803009575924e-05,
      "loss": 0.3366,
      "step": 1179
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.7979652881622314,
      "learning_rate": 4.619243046055632e-05,
      "loss": 0.3439,
      "step": 1180
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.36132702231407166,
      "eval_runtime": 21.8492,
      "eval_samples_per_second": 1.282,
      "eval_steps_per_second": 0.32,
      "step": 1180
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.822918176651001,
      "learning_rate": 4.61468308253534e-05,
      "loss": 0.3481,
      "step": 1181
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8427508473396301,
      "learning_rate": 4.610123119015048e-05,
      "loss": 0.3455,
      "step": 1182
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8091826438903809,
      "learning_rate": 4.6055631554947565e-05,
      "loss": 0.3353,
      "step": 1183
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8591803908348083,
      "learning_rate": 4.601003191974465e-05,
      "loss": 0.351,
      "step": 1184
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8574907779693604,
      "learning_rate": 4.596443228454172e-05,
      "loss": 0.3606,
      "step": 1185
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8384338021278381,
      "learning_rate": 4.5918832649338804e-05,
      "loss": 0.3347,
      "step": 1186
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8053162097930908,
      "learning_rate": 4.587323301413589e-05,
      "loss": 0.3493,
      "step": 1187
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.85943603515625,
      "learning_rate": 4.5827633378932974e-05,
      "loss": 0.3411,
      "step": 1188
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8188199996948242,
      "learning_rate": 4.578203374373005e-05,
      "loss": 0.3289,
      "step": 1189
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8296828269958496,
      "learning_rate": 4.573643410852713e-05,
      "loss": 0.3434,
      "step": 1190
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.3615628778934479,
      "eval_runtime": 21.7695,
      "eval_samples_per_second": 1.286,
      "eval_steps_per_second": 0.322,
      "step": 1190
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8322606682777405,
      "learning_rate": 4.569083447332422e-05,
      "loss": 0.338,
      "step": 1191
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8713050484657288,
      "learning_rate": 4.56452348381213e-05,
      "loss": 0.3368,
      "step": 1192
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.7981169819831848,
      "learning_rate": 4.5599635202918376e-05,
      "loss": 0.3362,
      "step": 1193
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.791098415851593,
      "learning_rate": 4.555403556771546e-05,
      "loss": 0.3333,
      "step": 1194
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.8596270680427551,
      "learning_rate": 4.5508435932512546e-05,
      "loss": 0.3261,
      "step": 1195
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8467071652412415,
      "learning_rate": 4.546283629730962e-05,
      "loss": 0.3459,
      "step": 1196
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7838037014007568,
      "learning_rate": 4.54172366621067e-05,
      "loss": 0.3406,
      "step": 1197
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7928112745285034,
      "learning_rate": 4.5371637026903785e-05,
      "loss": 0.3379,
      "step": 1198
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8437274098396301,
      "learning_rate": 4.532603739170087e-05,
      "loss": 0.3406,
      "step": 1199
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8299893140792847,
      "learning_rate": 4.528043775649795e-05,
      "loss": 0.3472,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "eval_loss": 0.36214447021484375,
      "eval_runtime": 21.8369,
      "eval_samples_per_second": 1.282,
      "eval_steps_per_second": 0.321,
      "step": 1200
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7983232736587524,
      "learning_rate": 4.523483812129503e-05,
      "loss": 0.3362,
      "step": 1201
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8538187146186829,
      "learning_rate": 4.518923848609211e-05,
      "loss": 0.3582,
      "step": 1202
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8160207271575928,
      "learning_rate": 4.51436388508892e-05,
      "loss": 0.3423,
      "step": 1203
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8224906921386719,
      "learning_rate": 4.5098039215686275e-05,
      "loss": 0.3291,
      "step": 1204
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7791831493377686,
      "learning_rate": 4.505243958048336e-05,
      "loss": 0.3451,
      "step": 1205
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7967703342437744,
      "learning_rate": 4.500683994528044e-05,
      "loss": 0.3322,
      "step": 1206
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7647725939750671,
      "learning_rate": 4.496124031007753e-05,
      "loss": 0.3225,
      "step": 1207
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8055102229118347,
      "learning_rate": 4.49156406748746e-05,
      "loss": 0.3294,
      "step": 1208
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7932788729667664,
      "learning_rate": 4.4870041039671684e-05,
      "loss": 0.3441,
      "step": 1209
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8258801698684692,
      "learning_rate": 4.4824441404468765e-05,
      "loss": 0.3605,
      "step": 1210
    },
    {
      "epoch": 0.55,
      "eval_loss": 0.361802339553833,
      "eval_runtime": 21.6768,
      "eval_samples_per_second": 1.292,
      "eval_steps_per_second": 0.323,
      "step": 1210
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7840272784233093,
      "learning_rate": 4.477884176926585e-05,
      "loss": 0.3479,
      "step": 1211
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8730911612510681,
      "learning_rate": 4.473324213406293e-05,
      "loss": 0.373,
      "step": 1212
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8281298875808716,
      "learning_rate": 4.468764249886001e-05,
      "loss": 0.3379,
      "step": 1213
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8440173864364624,
      "learning_rate": 4.464204286365709e-05,
      "loss": 0.3592,
      "step": 1214
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.8705652952194214,
      "learning_rate": 4.4596443228454174e-05,
      "loss": 0.3622,
      "step": 1215
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.7783262729644775,
      "learning_rate": 4.4550843593251256e-05,
      "loss": 0.3262,
      "step": 1216
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.9026988744735718,
      "learning_rate": 4.450524395804834e-05,
      "loss": 0.3698,
      "step": 1217
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8218929767608643,
      "learning_rate": 4.445964432284542e-05,
      "loss": 0.3459,
      "step": 1218
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8687737584114075,
      "learning_rate": 4.44140446876425e-05,
      "loss": 0.3489,
      "step": 1219
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8133276700973511,
      "learning_rate": 4.436844505243958e-05,
      "loss": 0.3474,
      "step": 1220
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.36127376556396484,
      "eval_runtime": 21.8837,
      "eval_samples_per_second": 1.279,
      "eval_steps_per_second": 0.32,
      "step": 1220
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.820976972579956,
      "learning_rate": 4.4322845417236665e-05,
      "loss": 0.3535,
      "step": 1221
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.7587445378303528,
      "learning_rate": 4.4277245782033746e-05,
      "loss": 0.345,
      "step": 1222
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.7987104058265686,
      "learning_rate": 4.423164614683083e-05,
      "loss": 0.3429,
      "step": 1223
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8208971619606018,
      "learning_rate": 4.418604651162791e-05,
      "loss": 0.3439,
      "step": 1224
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8062512874603271,
      "learning_rate": 4.414044687642499e-05,
      "loss": 0.3532,
      "step": 1225
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8156399130821228,
      "learning_rate": 4.409484724122207e-05,
      "loss": 0.365,
      "step": 1226
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8011518716812134,
      "learning_rate": 4.4049247606019155e-05,
      "loss": 0.3288,
      "step": 1227
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.7845913767814636,
      "learning_rate": 4.400364797081624e-05,
      "loss": 0.3299,
      "step": 1228
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8292993903160095,
      "learning_rate": 4.395804833561332e-05,
      "loss": 0.3666,
      "step": 1229
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8220998644828796,
      "learning_rate": 4.39124487004104e-05,
      "loss": 0.3381,
      "step": 1230
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.3609188199043274,
      "eval_runtime": 21.8935,
      "eval_samples_per_second": 1.279,
      "eval_steps_per_second": 0.32,
      "step": 1230
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.786451518535614,
      "learning_rate": 4.386684906520748e-05,
      "loss": 0.3329,
      "step": 1231
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.83308345079422,
      "learning_rate": 4.3821249430004564e-05,
      "loss": 0.3515,
      "step": 1232
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8226833343505859,
      "learning_rate": 4.3775649794801645e-05,
      "loss": 0.3458,
      "step": 1233
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8319200873374939,
      "learning_rate": 4.373005015959872e-05,
      "loss": 0.3491,
      "step": 1234
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8553533554077148,
      "learning_rate": 4.368445052439581e-05,
      "loss": 0.3524,
      "step": 1235
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.864627480506897,
      "learning_rate": 4.363885088919289e-05,
      "loss": 0.3572,
      "step": 1236
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8432021737098694,
      "learning_rate": 4.359325125398997e-05,
      "loss": 0.3543,
      "step": 1237
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.7859142422676086,
      "learning_rate": 4.354765161878705e-05,
      "loss": 0.3284,
      "step": 1238
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.8182294368743896,
      "learning_rate": 4.3502051983584136e-05,
      "loss": 0.3282,
      "step": 1239
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.9091182947158813,
      "learning_rate": 4.345645234838122e-05,
      "loss": 0.3749,
      "step": 1240
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.360344260931015,
      "eval_runtime": 21.9012,
      "eval_samples_per_second": 1.278,
      "eval_steps_per_second": 0.32,
      "step": 1240
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.7812795042991638,
      "learning_rate": 4.34108527131783e-05,
      "loss": 0.3315,
      "step": 1241
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.816305935382843,
      "learning_rate": 4.3365253077975374e-05,
      "loss": 0.3502,
      "step": 1242
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8034347891807556,
      "learning_rate": 4.331965344277246e-05,
      "loss": 0.3337,
      "step": 1243
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8315020203590393,
      "learning_rate": 4.3274053807569544e-05,
      "loss": 0.354,
      "step": 1244
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.7251798510551453,
      "learning_rate": 4.322845417236662e-05,
      "loss": 0.3249,
      "step": 1245
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.7859770059585571,
      "learning_rate": 4.31828545371637e-05,
      "loss": 0.3275,
      "step": 1246
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8476674556732178,
      "learning_rate": 4.313725490196079e-05,
      "loss": 0.3574,
      "step": 1247
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8440091609954834,
      "learning_rate": 4.309165526675787e-05,
      "loss": 0.3525,
      "step": 1248
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8422577977180481,
      "learning_rate": 4.3046055631554946e-05,
      "loss": 0.3585,
      "step": 1249
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8375608325004578,
      "learning_rate": 4.300045599635203e-05,
      "loss": 0.3625,
      "step": 1250
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.360787957906723,
      "eval_runtime": 21.7575,
      "eval_samples_per_second": 1.287,
      "eval_steps_per_second": 0.322,
      "step": 1250
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8197726607322693,
      "learning_rate": 4.2954856361149116e-05,
      "loss": 0.3537,
      "step": 1251
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8626959919929504,
      "learning_rate": 4.29092567259462e-05,
      "loss": 0.3442,
      "step": 1252
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8891031742095947,
      "learning_rate": 4.286365709074327e-05,
      "loss": 0.3681,
      "step": 1253
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8085613250732422,
      "learning_rate": 4.2818057455540355e-05,
      "loss": 0.3586,
      "step": 1254
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8536873459815979,
      "learning_rate": 4.2772457820337443e-05,
      "loss": 0.3324,
      "step": 1255
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8423956632614136,
      "learning_rate": 4.272685818513452e-05,
      "loss": 0.3529,
      "step": 1256
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8388254642486572,
      "learning_rate": 4.26812585499316e-05,
      "loss": 0.3612,
      "step": 1257
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.7667680978775024,
      "learning_rate": 4.263565891472868e-05,
      "loss": 0.3391,
      "step": 1258
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8259601593017578,
      "learning_rate": 4.259005927952577e-05,
      "loss": 0.3482,
      "step": 1259
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8969464898109436,
      "learning_rate": 4.2544459644322845e-05,
      "loss": 0.351,
      "step": 1260
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.36102089285850525,
      "eval_runtime": 21.6213,
      "eval_samples_per_second": 1.295,
      "eval_steps_per_second": 0.324,
      "step": 1260
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.8078624606132507,
      "learning_rate": 4.249886000911993e-05,
      "loss": 0.3494,
      "step": 1261
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8384256362915039,
      "learning_rate": 4.245326037391701e-05,
      "loss": 0.3574,
      "step": 1262
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8481121063232422,
      "learning_rate": 4.24076607387141e-05,
      "loss": 0.3517,
      "step": 1263
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8446504473686218,
      "learning_rate": 4.236206110351117e-05,
      "loss": 0.3495,
      "step": 1264
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.906520664691925,
      "learning_rate": 4.2316461468308254e-05,
      "loss": 0.3628,
      "step": 1265
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8286429047584534,
      "learning_rate": 4.2270861833105336e-05,
      "loss": 0.3521,
      "step": 1266
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8214951753616333,
      "learning_rate": 4.2225262197902424e-05,
      "loss": 0.3601,
      "step": 1267
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8151431679725647,
      "learning_rate": 4.21796625626995e-05,
      "loss": 0.3565,
      "step": 1268
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8063027262687683,
      "learning_rate": 4.213406292749658e-05,
      "loss": 0.3363,
      "step": 1269
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8764293789863586,
      "learning_rate": 4.208846329229366e-05,
      "loss": 0.3683,
      "step": 1270
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.36087459325790405,
      "eval_runtime": 21.8157,
      "eval_samples_per_second": 1.283,
      "eval_steps_per_second": 0.321,
      "step": 1270
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.7735445499420166,
      "learning_rate": 4.2042863657090744e-05,
      "loss": 0.3313,
      "step": 1271
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8024297952651978,
      "learning_rate": 4.1997264021887826e-05,
      "loss": 0.337,
      "step": 1272
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.78364098072052,
      "learning_rate": 4.195166438668491e-05,
      "loss": 0.3353,
      "step": 1273
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8018545508384705,
      "learning_rate": 4.190606475148199e-05,
      "loss": 0.333,
      "step": 1274
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8167896270751953,
      "learning_rate": 4.186046511627907e-05,
      "loss": 0.3374,
      "step": 1275
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8448525667190552,
      "learning_rate": 4.181486548107615e-05,
      "loss": 0.3631,
      "step": 1276
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8643104434013367,
      "learning_rate": 4.1769265845873235e-05,
      "loss": 0.3679,
      "step": 1277
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8329184651374817,
      "learning_rate": 4.1723666210670316e-05,
      "loss": 0.3334,
      "step": 1278
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8378819227218628,
      "learning_rate": 4.16780665754674e-05,
      "loss": 0.3605,
      "step": 1279
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8405564427375793,
      "learning_rate": 4.163246694026448e-05,
      "loss": 0.3465,
      "step": 1280
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.3606376647949219,
      "eval_runtime": 25.4779,
      "eval_samples_per_second": 1.099,
      "eval_steps_per_second": 0.275,
      "step": 1280
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8125925064086914,
      "learning_rate": 4.158686730506156e-05,
      "loss": 0.3392,
      "step": 1281
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.804631769657135,
      "learning_rate": 4.1541267669858643e-05,
      "loss": 0.3284,
      "step": 1282
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.8230606913566589,
      "learning_rate": 4.1495668034655725e-05,
      "loss": 0.3483,
      "step": 1283
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8066771030426025,
      "learning_rate": 4.145006839945281e-05,
      "loss": 0.3195,
      "step": 1284
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8422403335571289,
      "learning_rate": 4.140446876424989e-05,
      "loss": 0.3749,
      "step": 1285
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8488630652427673,
      "learning_rate": 4.135886912904697e-05,
      "loss": 0.3597,
      "step": 1286
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8020509481430054,
      "learning_rate": 4.131326949384405e-05,
      "loss": 0.3333,
      "step": 1287
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.7981416583061218,
      "learning_rate": 4.1267669858641134e-05,
      "loss": 0.335,
      "step": 1288
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8085934519767761,
      "learning_rate": 4.1222070223438216e-05,
      "loss": 0.3396,
      "step": 1289
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8348598480224609,
      "learning_rate": 4.11764705882353e-05,
      "loss": 0.3397,
      "step": 1290
    },
    {
      "epoch": 0.59,
      "eval_loss": 0.3603203594684601,
      "eval_runtime": 25.6317,
      "eval_samples_per_second": 1.092,
      "eval_steps_per_second": 0.273,
      "step": 1290
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.7678222060203552,
      "learning_rate": 4.113087095303238e-05,
      "loss": 0.3341,
      "step": 1291
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.823225200176239,
      "learning_rate": 4.108527131782946e-05,
      "loss": 0.3495,
      "step": 1292
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8174163699150085,
      "learning_rate": 4.103967168262654e-05,
      "loss": 0.3498,
      "step": 1293
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8279191851615906,
      "learning_rate": 4.099407204742362e-05,
      "loss": 0.3528,
      "step": 1294
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.7891851663589478,
      "learning_rate": 4.0948472412220706e-05,
      "loss": 0.3408,
      "step": 1295
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8267097473144531,
      "learning_rate": 4.090287277701779e-05,
      "loss": 0.3595,
      "step": 1296
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.819893479347229,
      "learning_rate": 4.085727314181487e-05,
      "loss": 0.3373,
      "step": 1297
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8141855001449585,
      "learning_rate": 4.0811673506611944e-05,
      "loss": 0.3517,
      "step": 1298
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.787208080291748,
      "learning_rate": 4.076607387140903e-05,
      "loss": 0.3206,
      "step": 1299
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8328278064727783,
      "learning_rate": 4.0720474236206115e-05,
      "loss": 0.3601,
      "step": 1300
    },
    {
      "epoch": 0.59,
      "eval_loss": 0.3600243628025055,
      "eval_runtime": 34.1653,
      "eval_samples_per_second": 0.82,
      "eval_steps_per_second": 0.205,
      "step": 1300
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8156459331512451,
      "learning_rate": 4.0674874601003196e-05,
      "loss": 0.3377,
      "step": 1301
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8492475152015686,
      "learning_rate": 4.062927496580027e-05,
      "loss": 0.3469,
      "step": 1302
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.8182046413421631,
      "learning_rate": 4.058367533059736e-05,
      "loss": 0.35,
      "step": 1303
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.803213357925415,
      "learning_rate": 4.053807569539444e-05,
      "loss": 0.3506,
      "step": 1304
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.7835839986801147,
      "learning_rate": 4.0492476060191516e-05,
      "loss": 0.3433,
      "step": 1305
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8448796272277832,
      "learning_rate": 4.04468764249886e-05,
      "loss": 0.3546,
      "step": 1306
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.7899730205535889,
      "learning_rate": 4.040127678978569e-05,
      "loss": 0.3359,
      "step": 1307
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8046201467514038,
      "learning_rate": 4.035567715458277e-05,
      "loss": 0.3385,
      "step": 1308
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8358333706855774,
      "learning_rate": 4.0310077519379843e-05,
      "loss": 0.3716,
      "step": 1309
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8300046920776367,
      "learning_rate": 4.0264477884176925e-05,
      "loss": 0.3588,
      "step": 1310
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.35997697710990906,
      "eval_runtime": 34.4834,
      "eval_samples_per_second": 0.812,
      "eval_steps_per_second": 0.203,
      "step": 1310
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8253521919250488,
      "learning_rate": 4.0218878248974014e-05,
      "loss": 0.3362,
      "step": 1311
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8703365325927734,
      "learning_rate": 4.0173278613771095e-05,
      "loss": 0.3591,
      "step": 1312
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8559762239456177,
      "learning_rate": 4.012767897856817e-05,
      "loss": 0.3345,
      "step": 1313
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8356379270553589,
      "learning_rate": 4.008207934336525e-05,
      "loss": 0.3718,
      "step": 1314
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.7918616533279419,
      "learning_rate": 4.003647970816234e-05,
      "loss": 0.3308,
      "step": 1315
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.898504912853241,
      "learning_rate": 3.999088007295942e-05,
      "loss": 0.3613,
      "step": 1316
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.7918505072593689,
      "learning_rate": 3.99452804377565e-05,
      "loss": 0.3224,
      "step": 1317
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8192049264907837,
      "learning_rate": 3.989968080255358e-05,
      "loss": 0.329,
      "step": 1318
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8752269744873047,
      "learning_rate": 3.985408116735067e-05,
      "loss": 0.3533,
      "step": 1319
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.7703496217727661,
      "learning_rate": 3.980848153214774e-05,
      "loss": 0.3143,
      "step": 1320
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.3601928651332855,
      "eval_runtime": 25.5117,
      "eval_samples_per_second": 1.098,
      "eval_steps_per_second": 0.274,
      "step": 1320
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8335027694702148,
      "learning_rate": 3.9762881896944824e-05,
      "loss": 0.3324,
      "step": 1321
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8822119235992432,
      "learning_rate": 3.9717282261741906e-05,
      "loss": 0.3579,
      "step": 1322
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8622987866401672,
      "learning_rate": 3.9671682626538994e-05,
      "loss": 0.3675,
      "step": 1323
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8221805691719055,
      "learning_rate": 3.962608299133607e-05,
      "loss": 0.359,
      "step": 1324
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8118740320205688,
      "learning_rate": 3.958048335613315e-05,
      "loss": 0.3496,
      "step": 1325
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.7761417031288147,
      "learning_rate": 3.953488372093023e-05,
      "loss": 0.3326,
      "step": 1326
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.812454879283905,
      "learning_rate": 3.948928408572732e-05,
      "loss": 0.3555,
      "step": 1327
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8731850981712341,
      "learning_rate": 3.9443684450524396e-05,
      "loss": 0.3771,
      "step": 1328
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.85962975025177,
      "learning_rate": 3.939808481532148e-05,
      "loss": 0.3599,
      "step": 1329
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8229592442512512,
      "learning_rate": 3.935248518011856e-05,
      "loss": 0.3607,
      "step": 1330
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.35948753356933594,
      "eval_runtime": 25.3157,
      "eval_samples_per_second": 1.106,
      "eval_steps_per_second": 0.277,
      "step": 1330
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.7954073548316956,
      "learning_rate": 3.930688554491564e-05,
      "loss": 0.3359,
      "step": 1331
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8094289302825928,
      "learning_rate": 3.926128590971272e-05,
      "loss": 0.3235,
      "step": 1332
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.826915442943573,
      "learning_rate": 3.9215686274509805e-05,
      "loss": 0.3452,
      "step": 1333
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.7665673494338989,
      "learning_rate": 3.917008663930689e-05,
      "loss": 0.3256,
      "step": 1334
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8052255511283875,
      "learning_rate": 3.912448700410397e-05,
      "loss": 0.3446,
      "step": 1335
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8459590077400208,
      "learning_rate": 3.907888736890105e-05,
      "loss": 0.3737,
      "step": 1336
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8796717524528503,
      "learning_rate": 3.903328773369813e-05,
      "loss": 0.3619,
      "step": 1337
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8221214413642883,
      "learning_rate": 3.8987688098495214e-05,
      "loss": 0.3447,
      "step": 1338
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8288754224777222,
      "learning_rate": 3.8942088463292295e-05,
      "loss": 0.3469,
      "step": 1339
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.772335946559906,
      "learning_rate": 3.889648882808938e-05,
      "loss": 0.3336,
      "step": 1340
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.35878506302833557,
      "eval_runtime": 19.9011,
      "eval_samples_per_second": 1.407,
      "eval_steps_per_second": 0.352,
      "step": 1340
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8370662927627563,
      "learning_rate": 3.885088919288646e-05,
      "loss": 0.3364,
      "step": 1341
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.7943137288093567,
      "learning_rate": 3.880528955768354e-05,
      "loss": 0.3442,
      "step": 1342
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8786354660987854,
      "learning_rate": 3.875968992248062e-05,
      "loss": 0.347,
      "step": 1343
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8405556082725525,
      "learning_rate": 3.8714090287277704e-05,
      "loss": 0.3463,
      "step": 1344
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8621813058853149,
      "learning_rate": 3.8668490652074786e-05,
      "loss": 0.3443,
      "step": 1345
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8162899613380432,
      "learning_rate": 3.862289101687187e-05,
      "loss": 0.3503,
      "step": 1346
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.7732406258583069,
      "learning_rate": 3.857729138166895e-05,
      "loss": 0.333,
      "step": 1347
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8653243780136108,
      "learning_rate": 3.853169174646603e-05,
      "loss": 0.3579,
      "step": 1348
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.8177424073219299,
      "learning_rate": 3.848609211126311e-05,
      "loss": 0.3445,
      "step": 1349
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8780079483985901,
      "learning_rate": 3.8440492476060194e-05,
      "loss": 0.3396,
      "step": 1350
    },
    {
      "epoch": 0.62,
      "eval_loss": 0.35845258831977844,
      "eval_runtime": 20.8282,
      "eval_samples_per_second": 1.344,
      "eval_steps_per_second": 0.336,
      "step": 1350
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.7994604110717773,
      "learning_rate": 3.8394892840857276e-05,
      "loss": 0.348,
      "step": 1351
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.7904070615768433,
      "learning_rate": 3.834929320565436e-05,
      "loss": 0.3436,
      "step": 1352
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8547983765602112,
      "learning_rate": 3.830369357045144e-05,
      "loss": 0.3378,
      "step": 1353
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8214447498321533,
      "learning_rate": 3.8258093935248515e-05,
      "loss": 0.3486,
      "step": 1354
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8538534641265869,
      "learning_rate": 3.82124943000456e-05,
      "loss": 0.3672,
      "step": 1355
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.821524441242218,
      "learning_rate": 3.8166894664842685e-05,
      "loss": 0.3508,
      "step": 1356
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8136518597602844,
      "learning_rate": 3.8121295029639767e-05,
      "loss": 0.3226,
      "step": 1357
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.7952825427055359,
      "learning_rate": 3.807569539443684e-05,
      "loss": 0.3312,
      "step": 1358
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.819299042224884,
      "learning_rate": 3.803009575923393e-05,
      "loss": 0.3374,
      "step": 1359
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8893455266952515,
      "learning_rate": 3.798449612403101e-05,
      "loss": 0.3271,
      "step": 1360
    },
    {
      "epoch": 0.62,
      "eval_loss": 0.3591363728046417,
      "eval_runtime": 18.335,
      "eval_samples_per_second": 1.527,
      "eval_steps_per_second": 0.382,
      "step": 1360
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.839749813079834,
      "learning_rate": 3.7938896488828094e-05,
      "loss": 0.3406,
      "step": 1361
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.7815828919410706,
      "learning_rate": 3.789329685362517e-05,
      "loss": 0.3298,
      "step": 1362
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8089067339897156,
      "learning_rate": 3.784769721842226e-05,
      "loss": 0.3381,
      "step": 1363
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8392083048820496,
      "learning_rate": 3.780209758321934e-05,
      "loss": 0.3536,
      "step": 1364
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8178284168243408,
      "learning_rate": 3.775649794801642e-05,
      "loss": 0.3363,
      "step": 1365
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8227528929710388,
      "learning_rate": 3.7710898312813495e-05,
      "loss": 0.3428,
      "step": 1366
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8406091928482056,
      "learning_rate": 3.7665298677610584e-05,
      "loss": 0.348,
      "step": 1367
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8545897603034973,
      "learning_rate": 3.7619699042407666e-05,
      "loss": 0.3517,
      "step": 1368
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.8717309236526489,
      "learning_rate": 3.757409940720474e-05,
      "loss": 0.3684,
      "step": 1369
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.86220383644104,
      "learning_rate": 3.752849977200182e-05,
      "loss": 0.3731,
      "step": 1370
    },
    {
      "epoch": 0.62,
      "eval_loss": 0.3583603501319885,
      "eval_runtime": 18.6501,
      "eval_samples_per_second": 1.501,
      "eval_steps_per_second": 0.375,
      "step": 1370
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.829359233379364,
      "learning_rate": 3.748290013679891e-05,
      "loss": 0.3265,
      "step": 1371
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.7688923478126526,
      "learning_rate": 3.743730050159599e-05,
      "loss": 0.327,
      "step": 1372
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.7676702737808228,
      "learning_rate": 3.739170086639307e-05,
      "loss": 0.331,
      "step": 1373
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8761011958122253,
      "learning_rate": 3.734610123119015e-05,
      "loss": 0.3445,
      "step": 1374
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8697832226753235,
      "learning_rate": 3.730050159598724e-05,
      "loss": 0.3834,
      "step": 1375
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.832823634147644,
      "learning_rate": 3.725490196078432e-05,
      "loss": 0.356,
      "step": 1376
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8047780394554138,
      "learning_rate": 3.7209302325581394e-05,
      "loss": 0.3407,
      "step": 1377
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8695047497749329,
      "learning_rate": 3.7163702690378476e-05,
      "loss": 0.3436,
      "step": 1378
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.762008011341095,
      "learning_rate": 3.7118103055175565e-05,
      "loss": 0.3155,
      "step": 1379
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8362067937850952,
      "learning_rate": 3.707250341997264e-05,
      "loss": 0.3469,
      "step": 1380
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.3580807149410248,
      "eval_runtime": 18.2133,
      "eval_samples_per_second": 1.537,
      "eval_steps_per_second": 0.384,
      "step": 1380
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.7919859290122986,
      "learning_rate": 3.702690378476972e-05,
      "loss": 0.342,
      "step": 1381
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8330515027046204,
      "learning_rate": 3.69813041495668e-05,
      "loss": 0.3471,
      "step": 1382
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8035807609558105,
      "learning_rate": 3.693570451436389e-05,
      "loss": 0.3649,
      "step": 1383
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.858091413974762,
      "learning_rate": 3.6890104879160967e-05,
      "loss": 0.3571,
      "step": 1384
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.7906508445739746,
      "learning_rate": 3.684450524395805e-05,
      "loss": 0.3329,
      "step": 1385
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8505438566207886,
      "learning_rate": 3.679890560875513e-05,
      "loss": 0.3521,
      "step": 1386
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.840091347694397,
      "learning_rate": 3.675330597355222e-05,
      "loss": 0.3495,
      "step": 1387
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.886907696723938,
      "learning_rate": 3.6707706338349294e-05,
      "loss": 0.357,
      "step": 1388
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.804918646812439,
      "learning_rate": 3.6662106703146375e-05,
      "loss": 0.3358,
      "step": 1389
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8357195854187012,
      "learning_rate": 3.661650706794346e-05,
      "loss": 0.3375,
      "step": 1390
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.358403742313385,
      "eval_runtime": 18.0533,
      "eval_samples_per_second": 1.551,
      "eval_steps_per_second": 0.388,
      "step": 1390
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8863046765327454,
      "learning_rate": 3.6570907432740545e-05,
      "loss": 0.3587,
      "step": 1391
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.8171054124832153,
      "learning_rate": 3.652530779753762e-05,
      "loss": 0.3584,
      "step": 1392
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.7896978855133057,
      "learning_rate": 3.64797081623347e-05,
      "loss": 0.3112,
      "step": 1393
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8571012020111084,
      "learning_rate": 3.6434108527131784e-05,
      "loss": 0.3332,
      "step": 1394
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8435710668563843,
      "learning_rate": 3.6388508891928866e-05,
      "loss": 0.3551,
      "step": 1395
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8161331415176392,
      "learning_rate": 3.634290925672595e-05,
      "loss": 0.3492,
      "step": 1396
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.840015172958374,
      "learning_rate": 3.629730962152303e-05,
      "loss": 0.3524,
      "step": 1397
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8551552891731262,
      "learning_rate": 3.625170998632011e-05,
      "loss": 0.3414,
      "step": 1398
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8473491668701172,
      "learning_rate": 3.620611035111719e-05,
      "loss": 0.3514,
      "step": 1399
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8031140565872192,
      "learning_rate": 3.6160510715914274e-05,
      "loss": 0.3202,
      "step": 1400
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.35901978611946106,
      "eval_runtime": 18.7272,
      "eval_samples_per_second": 1.495,
      "eval_steps_per_second": 0.374,
      "step": 1400
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8324646353721619,
      "learning_rate": 3.6114911080711356e-05,
      "loss": 0.3578,
      "step": 1401
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.7989433407783508,
      "learning_rate": 3.606931144550844e-05,
      "loss": 0.314,
      "step": 1402
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8086175322532654,
      "learning_rate": 3.602371181030552e-05,
      "loss": 0.3395,
      "step": 1403
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.7840234041213989,
      "learning_rate": 3.59781121751026e-05,
      "loss": 0.3465,
      "step": 1404
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8186520934104919,
      "learning_rate": 3.593251253989968e-05,
      "loss": 0.3418,
      "step": 1405
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.858751118183136,
      "learning_rate": 3.5886912904696765e-05,
      "loss": 0.3389,
      "step": 1406
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.7823198437690735,
      "learning_rate": 3.5841313269493846e-05,
      "loss": 0.3357,
      "step": 1407
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.7624536156654358,
      "learning_rate": 3.579571363429093e-05,
      "loss": 0.3165,
      "step": 1408
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.7955945134162903,
      "learning_rate": 3.575011399908801e-05,
      "loss": 0.3462,
      "step": 1409
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8175010085105896,
      "learning_rate": 3.570451436388509e-05,
      "loss": 0.3358,
      "step": 1410
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.35880425572395325,
      "eval_runtime": 18.8766,
      "eval_samples_per_second": 1.483,
      "eval_steps_per_second": 0.371,
      "step": 1410
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8709760904312134,
      "learning_rate": 3.565891472868217e-05,
      "loss": 0.3564,
      "step": 1411
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8882624506950378,
      "learning_rate": 3.5613315093479255e-05,
      "loss": 0.3559,
      "step": 1412
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.7864699959754944,
      "learning_rate": 3.556771545827634e-05,
      "loss": 0.3234,
      "step": 1413
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.8428051471710205,
      "learning_rate": 3.552211582307342e-05,
      "loss": 0.3641,
      "step": 1414
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8524724245071411,
      "learning_rate": 3.54765161878705e-05,
      "loss": 0.3408,
      "step": 1415
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8270473480224609,
      "learning_rate": 3.543091655266758e-05,
      "loss": 0.342,
      "step": 1416
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8461461663246155,
      "learning_rate": 3.5385316917464664e-05,
      "loss": 0.3481,
      "step": 1417
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8053926229476929,
      "learning_rate": 3.533971728226174e-05,
      "loss": 0.3428,
      "step": 1418
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8227868676185608,
      "learning_rate": 3.529411764705883e-05,
      "loss": 0.3371,
      "step": 1419
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8701817989349365,
      "learning_rate": 3.524851801185591e-05,
      "loss": 0.3509,
      "step": 1420
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.3587776720523834,
      "eval_runtime": 18.3142,
      "eval_samples_per_second": 1.529,
      "eval_steps_per_second": 0.382,
      "step": 1420
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8183357119560242,
      "learning_rate": 3.520291837665299e-05,
      "loss": 0.3408,
      "step": 1421
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8657394051551819,
      "learning_rate": 3.5157318741450066e-05,
      "loss": 0.3462,
      "step": 1422
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8495250940322876,
      "learning_rate": 3.5111719106247154e-05,
      "loss": 0.3359,
      "step": 1423
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8178961873054504,
      "learning_rate": 3.5066119471044236e-05,
      "loss": 0.3522,
      "step": 1424
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8224758505821228,
      "learning_rate": 3.502051983584132e-05,
      "loss": 0.334,
      "step": 1425
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8101351261138916,
      "learning_rate": 3.497492020063839e-05,
      "loss": 0.3262,
      "step": 1426
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8682355880737305,
      "learning_rate": 3.492932056543548e-05,
      "loss": 0.3647,
      "step": 1427
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.7884172797203064,
      "learning_rate": 3.488372093023256e-05,
      "loss": 0.3283,
      "step": 1428
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8620269894599915,
      "learning_rate": 3.483812129502964e-05,
      "loss": 0.3392,
      "step": 1429
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.7871687412261963,
      "learning_rate": 3.479252165982672e-05,
      "loss": 0.3171,
      "step": 1430
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.3581898510456085,
      "eval_runtime": 18.941,
      "eval_samples_per_second": 1.478,
      "eval_steps_per_second": 0.37,
      "step": 1430
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8699492812156677,
      "learning_rate": 3.474692202462381e-05,
      "loss": 0.3539,
      "step": 1431
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8600180149078369,
      "learning_rate": 3.470132238942089e-05,
      "loss": 0.3392,
      "step": 1432
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8551745414733887,
      "learning_rate": 3.4655722754217965e-05,
      "loss": 0.3701,
      "step": 1433
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8190971612930298,
      "learning_rate": 3.4610123119015046e-05,
      "loss": 0.3482,
      "step": 1434
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8227312564849854,
      "learning_rate": 3.4564523483812135e-05,
      "loss": 0.3611,
      "step": 1435
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.8590283989906311,
      "learning_rate": 3.451892384860922e-05,
      "loss": 0.3432,
      "step": 1436
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8469401001930237,
      "learning_rate": 3.447332421340629e-05,
      "loss": 0.334,
      "step": 1437
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8209288120269775,
      "learning_rate": 3.442772457820337e-05,
      "loss": 0.34,
      "step": 1438
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8007102608680725,
      "learning_rate": 3.438212494300046e-05,
      "loss": 0.3601,
      "step": 1439
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8150454163551331,
      "learning_rate": 3.433652530779754e-05,
      "loss": 0.3407,
      "step": 1440
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.3582390248775482,
      "eval_runtime": 18.6454,
      "eval_samples_per_second": 1.502,
      "eval_steps_per_second": 0.375,
      "step": 1440
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8899417519569397,
      "learning_rate": 3.429092567259462e-05,
      "loss": 0.367,
      "step": 1441
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8598084449768066,
      "learning_rate": 3.42453260373917e-05,
      "loss": 0.3476,
      "step": 1442
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8577296137809753,
      "learning_rate": 3.419972640218879e-05,
      "loss": 0.3553,
      "step": 1443
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.805564820766449,
      "learning_rate": 3.4154126766985864e-05,
      "loss": 0.3295,
      "step": 1444
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8372154235839844,
      "learning_rate": 3.4108527131782945e-05,
      "loss": 0.3533,
      "step": 1445
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8153108358383179,
      "learning_rate": 3.406292749658003e-05,
      "loss": 0.3433,
      "step": 1446
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.82663893699646,
      "learning_rate": 3.4017327861377116e-05,
      "loss": 0.352,
      "step": 1447
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8443247079849243,
      "learning_rate": 3.397172822617419e-05,
      "loss": 0.3464,
      "step": 1448
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8249718546867371,
      "learning_rate": 3.392612859097127e-05,
      "loss": 0.3408,
      "step": 1449
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.873860239982605,
      "learning_rate": 3.3880528955768354e-05,
      "loss": 0.3575,
      "step": 1450
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.35808515548706055,
      "eval_runtime": 18.3021,
      "eval_samples_per_second": 1.53,
      "eval_steps_per_second": 0.382,
      "step": 1450
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.797887921333313,
      "learning_rate": 3.383492932056544e-05,
      "loss": 0.3355,
      "step": 1451
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8364313244819641,
      "learning_rate": 3.378932968536252e-05,
      "loss": 0.3445,
      "step": 1452
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.7586292028427124,
      "learning_rate": 3.37437300501596e-05,
      "loss": 0.3252,
      "step": 1453
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8402112126350403,
      "learning_rate": 3.369813041495668e-05,
      "loss": 0.3474,
      "step": 1454
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.7776329517364502,
      "learning_rate": 3.365253077975376e-05,
      "loss": 0.3259,
      "step": 1455
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8207614421844482,
      "learning_rate": 3.3606931144550845e-05,
      "loss": 0.3412,
      "step": 1456
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.7944803833961487,
      "learning_rate": 3.3561331509347926e-05,
      "loss": 0.3524,
      "step": 1457
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.8181450963020325,
      "learning_rate": 3.351573187414501e-05,
      "loss": 0.329,
      "step": 1458
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8156586289405823,
      "learning_rate": 3.347013223894209e-05,
      "loss": 0.3303,
      "step": 1459
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8118116855621338,
      "learning_rate": 3.342453260373917e-05,
      "loss": 0.3571,
      "step": 1460
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.3578401207923889,
      "eval_runtime": 18.3051,
      "eval_samples_per_second": 1.53,
      "eval_steps_per_second": 0.382,
      "step": 1460
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8032804727554321,
      "learning_rate": 3.337893296853625e-05,
      "loss": 0.337,
      "step": 1461
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8342922329902649,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.3402,
      "step": 1462
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.7907533049583435,
      "learning_rate": 3.328773369813042e-05,
      "loss": 0.3387,
      "step": 1463
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.830132782459259,
      "learning_rate": 3.32421340629275e-05,
      "loss": 0.3501,
      "step": 1464
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8640668392181396,
      "learning_rate": 3.319653442772458e-05,
      "loss": 0.3561,
      "step": 1465
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8216010928153992,
      "learning_rate": 3.315093479252166e-05,
      "loss": 0.3442,
      "step": 1466
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8106178641319275,
      "learning_rate": 3.3105335157318744e-05,
      "loss": 0.3355,
      "step": 1467
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.7535154819488525,
      "learning_rate": 3.3059735522115825e-05,
      "loss": 0.3161,
      "step": 1468
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8195207715034485,
      "learning_rate": 3.301413588691291e-05,
      "loss": 0.3475,
      "step": 1469
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8379241228103638,
      "learning_rate": 3.296853625170999e-05,
      "loss": 0.3575,
      "step": 1470
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.3576245903968811,
      "eval_runtime": 18.3426,
      "eval_samples_per_second": 1.527,
      "eval_steps_per_second": 0.382,
      "step": 1470
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8420901894569397,
      "learning_rate": 3.292293661650707e-05,
      "loss": 0.3288,
      "step": 1471
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8348906636238098,
      "learning_rate": 3.287733698130415e-05,
      "loss": 0.3527,
      "step": 1472
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8279021382331848,
      "learning_rate": 3.2831737346101234e-05,
      "loss": 0.3458,
      "step": 1473
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8192549347877502,
      "learning_rate": 3.2786137710898316e-05,
      "loss": 0.3648,
      "step": 1474
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8519864678382874,
      "learning_rate": 3.27405380756954e-05,
      "loss": 0.344,
      "step": 1475
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8401514887809753,
      "learning_rate": 3.269493844049248e-05,
      "loss": 0.3529,
      "step": 1476
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8426647186279297,
      "learning_rate": 3.264933880528956e-05,
      "loss": 0.3453,
      "step": 1477
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.7950509786605835,
      "learning_rate": 3.2603739170086636e-05,
      "loss": 0.3472,
      "step": 1478
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8634794354438782,
      "learning_rate": 3.2558139534883724e-05,
      "loss": 0.3521,
      "step": 1479
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.8343892693519592,
      "learning_rate": 3.2512539899680806e-05,
      "loss": 0.3439,
      "step": 1480
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.35794949531555176,
      "eval_runtime": 19.3365,
      "eval_samples_per_second": 1.448,
      "eval_steps_per_second": 0.362,
      "step": 1480
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.7715625762939453,
      "learning_rate": 3.246694026447789e-05,
      "loss": 0.3181,
      "step": 1481
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8029360771179199,
      "learning_rate": 3.242134062927496e-05,
      "loss": 0.3261,
      "step": 1482
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8179964423179626,
      "learning_rate": 3.237574099407205e-05,
      "loss": 0.3252,
      "step": 1483
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8506777882575989,
      "learning_rate": 3.233014135886913e-05,
      "loss": 0.3393,
      "step": 1484
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8550512790679932,
      "learning_rate": 3.2284541723666215e-05,
      "loss": 0.3387,
      "step": 1485
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8071249723434448,
      "learning_rate": 3.223894208846329e-05,
      "loss": 0.3279,
      "step": 1486
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.809755265712738,
      "learning_rate": 3.219334245326038e-05,
      "loss": 0.3551,
      "step": 1487
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8375426530838013,
      "learning_rate": 3.214774281805746e-05,
      "loss": 0.3615,
      "step": 1488
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8345934152603149,
      "learning_rate": 3.2102143182854535e-05,
      "loss": 0.3507,
      "step": 1489
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8337054252624512,
      "learning_rate": 3.205654354765162e-05,
      "loss": 0.3466,
      "step": 1490
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.3576383888721466,
      "eval_runtime": 19.4342,
      "eval_samples_per_second": 1.441,
      "eval_steps_per_second": 0.36,
      "step": 1490
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8991925716400146,
      "learning_rate": 3.2010943912448705e-05,
      "loss": 0.3451,
      "step": 1491
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8050253987312317,
      "learning_rate": 3.196534427724579e-05,
      "loss": 0.3535,
      "step": 1492
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8160567283630371,
      "learning_rate": 3.191974464204286e-05,
      "loss": 0.3394,
      "step": 1493
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8046148419380188,
      "learning_rate": 3.1874145006839944e-05,
      "loss": 0.3445,
      "step": 1494
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.7577115297317505,
      "learning_rate": 3.182854537163703e-05,
      "loss": 0.3421,
      "step": 1495
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8331876993179321,
      "learning_rate": 3.1782945736434114e-05,
      "loss": 0.3487,
      "step": 1496
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.7791025042533875,
      "learning_rate": 3.173734610123119e-05,
      "loss": 0.3306,
      "step": 1497
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8634000420570374,
      "learning_rate": 3.169174646602827e-05,
      "loss": 0.3486,
      "step": 1498
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8452799916267395,
      "learning_rate": 3.164614683082536e-05,
      "loss": 0.353,
      "step": 1499
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8176222443580627,
      "learning_rate": 3.160054719562244e-05,
      "loss": 0.342,
      "step": 1500
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.35751423239707947,
      "eval_runtime": 18.9707,
      "eval_samples_per_second": 1.476,
      "eval_steps_per_second": 0.369,
      "step": 1500
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.846010684967041,
      "learning_rate": 3.1554947560419516e-05,
      "loss": 0.3651,
      "step": 1501
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.8323935270309448,
      "learning_rate": 3.15093479252166e-05,
      "loss": 0.3459,
      "step": 1502
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.9228621125221252,
      "learning_rate": 3.1463748290013686e-05,
      "loss": 0.3485,
      "step": 1503
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8141658902168274,
      "learning_rate": 3.141814865481076e-05,
      "loss": 0.3349,
      "step": 1504
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8556486964225769,
      "learning_rate": 3.137254901960784e-05,
      "loss": 0.3566,
      "step": 1505
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.858826220035553,
      "learning_rate": 3.1326949384404924e-05,
      "loss": 0.3347,
      "step": 1506
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8861152529716492,
      "learning_rate": 3.128134974920201e-05,
      "loss": 0.3613,
      "step": 1507
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8302914500236511,
      "learning_rate": 3.123575011399909e-05,
      "loss": 0.3429,
      "step": 1508
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8661092519760132,
      "learning_rate": 3.119015047879617e-05,
      "loss": 0.3577,
      "step": 1509
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.813862681388855,
      "learning_rate": 3.114455084359325e-05,
      "loss": 0.3321,
      "step": 1510
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.35740190744400024,
      "eval_runtime": 19.4999,
      "eval_samples_per_second": 1.436,
      "eval_steps_per_second": 0.359,
      "step": 1510
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.7683899998664856,
      "learning_rate": 3.109895120839034e-05,
      "loss": 0.3428,
      "step": 1511
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8115766048431396,
      "learning_rate": 3.1053351573187415e-05,
      "loss": 0.3279,
      "step": 1512
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8759005665779114,
      "learning_rate": 3.1007751937984497e-05,
      "loss": 0.3459,
      "step": 1513
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8469499945640564,
      "learning_rate": 3.096215230278158e-05,
      "loss": 0.3339,
      "step": 1514
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8490378856658936,
      "learning_rate": 3.091655266757866e-05,
      "loss": 0.3347,
      "step": 1515
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.814757764339447,
      "learning_rate": 3.087095303237574e-05,
      "loss": 0.3518,
      "step": 1516
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8592610955238342,
      "learning_rate": 3.0825353397172823e-05,
      "loss": 0.3453,
      "step": 1517
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.7975971102714539,
      "learning_rate": 3.0779753761969905e-05,
      "loss": 0.3437,
      "step": 1518
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8102256655693054,
      "learning_rate": 3.073415412676699e-05,
      "loss": 0.3234,
      "step": 1519
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.807258129119873,
      "learning_rate": 3.068855449156407e-05,
      "loss": 0.3484,
      "step": 1520
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.3571014106273651,
      "eval_runtime": 19.2314,
      "eval_samples_per_second": 1.456,
      "eval_steps_per_second": 0.364,
      "step": 1520
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.7736379504203796,
      "learning_rate": 3.064295485636115e-05,
      "loss": 0.3372,
      "step": 1521
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8541015982627869,
      "learning_rate": 3.059735522115823e-05,
      "loss": 0.3479,
      "step": 1522
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8304355144500732,
      "learning_rate": 3.0551755585955314e-05,
      "loss": 0.3629,
      "step": 1523
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.8115606904029846,
      "learning_rate": 3.0506155950752396e-05,
      "loss": 0.3272,
      "step": 1524
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8201666474342346,
      "learning_rate": 3.0460556315549477e-05,
      "loss": 0.3396,
      "step": 1525
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8593149185180664,
      "learning_rate": 3.0414956680346556e-05,
      "loss": 0.3488,
      "step": 1526
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.798021137714386,
      "learning_rate": 3.0369357045143644e-05,
      "loss": 0.3414,
      "step": 1527
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8162714838981628,
      "learning_rate": 3.0323757409940722e-05,
      "loss": 0.3575,
      "step": 1528
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8344952464103699,
      "learning_rate": 3.0278157774737804e-05,
      "loss": 0.351,
      "step": 1529
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8304796814918518,
      "learning_rate": 3.0232558139534883e-05,
      "loss": 0.3339,
      "step": 1530
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.3572339713573456,
      "eval_runtime": 19.1453,
      "eval_samples_per_second": 1.463,
      "eval_steps_per_second": 0.366,
      "step": 1530
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8597332239151001,
      "learning_rate": 3.0186958504331968e-05,
      "loss": 0.355,
      "step": 1531
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.7662696242332458,
      "learning_rate": 3.014135886912905e-05,
      "loss": 0.3448,
      "step": 1532
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.784625768661499,
      "learning_rate": 3.009575923392613e-05,
      "loss": 0.3426,
      "step": 1533
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8529240489006042,
      "learning_rate": 3.005015959872321e-05,
      "loss": 0.3435,
      "step": 1534
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8391616940498352,
      "learning_rate": 3.0004559963520295e-05,
      "loss": 0.3536,
      "step": 1535
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.7962397336959839,
      "learning_rate": 2.9958960328317376e-05,
      "loss": 0.3222,
      "step": 1536
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8026924729347229,
      "learning_rate": 2.9913360693114455e-05,
      "loss": 0.3562,
      "step": 1537
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8800778388977051,
      "learning_rate": 2.9867761057911536e-05,
      "loss": 0.3515,
      "step": 1538
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8260602951049805,
      "learning_rate": 2.982216142270862e-05,
      "loss": 0.3402,
      "step": 1539
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.850034773349762,
      "learning_rate": 2.9776561787505703e-05,
      "loss": 0.3627,
      "step": 1540
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.35721954703330994,
      "eval_runtime": 18.9282,
      "eval_samples_per_second": 1.479,
      "eval_steps_per_second": 0.37,
      "step": 1540
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8191645741462708,
      "learning_rate": 2.973096215230278e-05,
      "loss": 0.3532,
      "step": 1541
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8150588274002075,
      "learning_rate": 2.9685362517099863e-05,
      "loss": 0.3476,
      "step": 1542
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.7702752947807312,
      "learning_rate": 2.963976288189695e-05,
      "loss": 0.3301,
      "step": 1543
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8159235715866089,
      "learning_rate": 2.959416324669403e-05,
      "loss": 0.3163,
      "step": 1544
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8318021893501282,
      "learning_rate": 2.954856361149111e-05,
      "loss": 0.3352,
      "step": 1545
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.826701283454895,
      "learning_rate": 2.950296397628819e-05,
      "loss": 0.3357,
      "step": 1546
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8172712922096252,
      "learning_rate": 2.9457364341085275e-05,
      "loss": 0.3364,
      "step": 1547
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8200088739395142,
      "learning_rate": 2.9411764705882354e-05,
      "loss": 0.3312,
      "step": 1548
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8334987759590149,
      "learning_rate": 2.9366165070679435e-05,
      "loss": 0.348,
      "step": 1549
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.768750011920929,
      "learning_rate": 2.9320565435476517e-05,
      "loss": 0.3383,
      "step": 1550
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.35663700103759766,
      "eval_runtime": 19.07,
      "eval_samples_per_second": 1.468,
      "eval_steps_per_second": 0.367,
      "step": 1550
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8347567915916443,
      "learning_rate": 2.9274965800273602e-05,
      "loss": 0.3559,
      "step": 1551
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8524714708328247,
      "learning_rate": 2.922936616507068e-05,
      "loss": 0.3625,
      "step": 1552
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.7744024991989136,
      "learning_rate": 2.9183766529867762e-05,
      "loss": 0.3524,
      "step": 1553
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8230989575386047,
      "learning_rate": 2.913816689466484e-05,
      "loss": 0.3545,
      "step": 1554
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.7706303596496582,
      "learning_rate": 2.909256725946193e-05,
      "loss": 0.3356,
      "step": 1555
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8915783762931824,
      "learning_rate": 2.9046967624259008e-05,
      "loss": 0.3761,
      "step": 1556
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8355619311332703,
      "learning_rate": 2.900136798905609e-05,
      "loss": 0.3756,
      "step": 1557
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8191080093383789,
      "learning_rate": 2.8955768353853168e-05,
      "loss": 0.3315,
      "step": 1558
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8697982430458069,
      "learning_rate": 2.8910168718650256e-05,
      "loss": 0.3592,
      "step": 1559
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8937227725982666,
      "learning_rate": 2.8864569083447335e-05,
      "loss": 0.3317,
      "step": 1560
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.3572239875793457,
      "eval_runtime": 19.0073,
      "eval_samples_per_second": 1.473,
      "eval_steps_per_second": 0.368,
      "step": 1560
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.833889365196228,
      "learning_rate": 2.8818969448244416e-05,
      "loss": 0.3533,
      "step": 1561
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8254353404045105,
      "learning_rate": 2.8773369813041495e-05,
      "loss": 0.3368,
      "step": 1562
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8069315552711487,
      "learning_rate": 2.872777017783858e-05,
      "loss": 0.338,
      "step": 1563
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.7715954184532166,
      "learning_rate": 2.868217054263566e-05,
      "loss": 0.3254,
      "step": 1564
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8389305472373962,
      "learning_rate": 2.8636570907432743e-05,
      "loss": 0.3439,
      "step": 1565
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8371710777282715,
      "learning_rate": 2.859097127222982e-05,
      "loss": 0.3387,
      "step": 1566
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.7831338047981262,
      "learning_rate": 2.8545371637026907e-05,
      "loss": 0.3127,
      "step": 1567
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.8159829378128052,
      "learning_rate": 2.849977200182399e-05,
      "loss": 0.3264,
      "step": 1568
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8167608380317688,
      "learning_rate": 2.8454172366621067e-05,
      "loss": 0.3364,
      "step": 1569
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8175981640815735,
      "learning_rate": 2.840857273141815e-05,
      "loss": 0.3491,
      "step": 1570
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.3574839234352112,
      "eval_runtime": 19.2429,
      "eval_samples_per_second": 1.455,
      "eval_steps_per_second": 0.364,
      "step": 1570
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8386175036430359,
      "learning_rate": 2.8362973096215234e-05,
      "loss": 0.3485,
      "step": 1571
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.816241443157196,
      "learning_rate": 2.8317373461012315e-05,
      "loss": 0.3353,
      "step": 1572
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8434162139892578,
      "learning_rate": 2.8271773825809394e-05,
      "loss": 0.3846,
      "step": 1573
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.791397213935852,
      "learning_rate": 2.8226174190606475e-05,
      "loss": 0.3335,
      "step": 1574
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8274080157279968,
      "learning_rate": 2.818057455540356e-05,
      "loss": 0.35,
      "step": 1575
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.7708328366279602,
      "learning_rate": 2.8134974920200642e-05,
      "loss": 0.3089,
      "step": 1576
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.846105694770813,
      "learning_rate": 2.808937528499772e-05,
      "loss": 0.3332,
      "step": 1577
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.7850958108901978,
      "learning_rate": 2.8043775649794802e-05,
      "loss": 0.3181,
      "step": 1578
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8704512119293213,
      "learning_rate": 2.7998176014591887e-05,
      "loss": 0.3559,
      "step": 1579
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8448445796966553,
      "learning_rate": 2.7952576379388966e-05,
      "loss": 0.3612,
      "step": 1580
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.35740068554878235,
      "eval_runtime": 19.5965,
      "eval_samples_per_second": 1.429,
      "eval_steps_per_second": 0.357,
      "step": 1580
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8353627920150757,
      "learning_rate": 2.7906976744186048e-05,
      "loss": 0.3496,
      "step": 1581
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8289371132850647,
      "learning_rate": 2.786137710898313e-05,
      "loss": 0.3447,
      "step": 1582
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8660280704498291,
      "learning_rate": 2.7815777473780214e-05,
      "loss": 0.341,
      "step": 1583
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8957594633102417,
      "learning_rate": 2.7770177838577293e-05,
      "loss": 0.3216,
      "step": 1584
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.81903475522995,
      "learning_rate": 2.7724578203374374e-05,
      "loss": 0.3418,
      "step": 1585
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8132171630859375,
      "learning_rate": 2.7678978568171453e-05,
      "loss": 0.3466,
      "step": 1586
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8006194829940796,
      "learning_rate": 2.763337893296854e-05,
      "loss": 0.3377,
      "step": 1587
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8752909302711487,
      "learning_rate": 2.758777929776562e-05,
      "loss": 0.369,
      "step": 1588
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8540804386138916,
      "learning_rate": 2.75421796625627e-05,
      "loss": 0.3354,
      "step": 1589
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.8105359077453613,
      "learning_rate": 2.749658002735978e-05,
      "loss": 0.3293,
      "step": 1590
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.3572061359882355,
      "eval_runtime": 19.0182,
      "eval_samples_per_second": 1.472,
      "eval_steps_per_second": 0.368,
      "step": 1590
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8374319672584534,
      "learning_rate": 2.7450980392156865e-05,
      "loss": 0.3393,
      "step": 1591
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8734089732170105,
      "learning_rate": 2.7405380756953947e-05,
      "loss": 0.3525,
      "step": 1592
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8762404918670654,
      "learning_rate": 2.735978112175103e-05,
      "loss": 0.3683,
      "step": 1593
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8304470777511597,
      "learning_rate": 2.7314181486548107e-05,
      "loss": 0.3502,
      "step": 1594
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8588269948959351,
      "learning_rate": 2.7268581851345192e-05,
      "loss": 0.3464,
      "step": 1595
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.7775711417198181,
      "learning_rate": 2.7222982216142274e-05,
      "loss": 0.3206,
      "step": 1596
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8479073643684387,
      "learning_rate": 2.7177382580939352e-05,
      "loss": 0.3568,
      "step": 1597
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8442942500114441,
      "learning_rate": 2.7131782945736434e-05,
      "loss": 0.3389,
      "step": 1598
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8447399735450745,
      "learning_rate": 2.708618331053352e-05,
      "loss": 0.3375,
      "step": 1599
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8228472471237183,
      "learning_rate": 2.70405836753306e-05,
      "loss": 0.3229,
      "step": 1600
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.35655805468559265,
      "eval_runtime": 19.4148,
      "eval_samples_per_second": 1.442,
      "eval_steps_per_second": 0.361,
      "step": 1600
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8076717853546143,
      "learning_rate": 2.699498404012768e-05,
      "loss": 0.3336,
      "step": 1601
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8026668429374695,
      "learning_rate": 2.694938440492476e-05,
      "loss": 0.3298,
      "step": 1602
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.801601767539978,
      "learning_rate": 2.6903784769721846e-05,
      "loss": 0.3445,
      "step": 1603
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8218444585800171,
      "learning_rate": 2.6858185134518927e-05,
      "loss": 0.3617,
      "step": 1604
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8179977536201477,
      "learning_rate": 2.6812585499316006e-05,
      "loss": 0.3248,
      "step": 1605
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.7940379977226257,
      "learning_rate": 2.6766985864113087e-05,
      "loss": 0.3401,
      "step": 1606
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8855268359184265,
      "learning_rate": 2.6721386228910173e-05,
      "loss": 0.3692,
      "step": 1607
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.816184937953949,
      "learning_rate": 2.6675786593707254e-05,
      "loss": 0.3462,
      "step": 1608
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8183460831642151,
      "learning_rate": 2.6630186958504333e-05,
      "loss": 0.3547,
      "step": 1609
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8423659801483154,
      "learning_rate": 2.6584587323301414e-05,
      "loss": 0.3534,
      "step": 1610
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.35616493225097656,
      "eval_runtime": 19.1242,
      "eval_samples_per_second": 1.464,
      "eval_steps_per_second": 0.366,
      "step": 1610
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.8094700574874878,
      "learning_rate": 2.65389876880985e-05,
      "loss": 0.3299,
      "step": 1611
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.793865442276001,
      "learning_rate": 2.6493388052895578e-05,
      "loss": 0.3344,
      "step": 1612
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8043534159660339,
      "learning_rate": 2.644778841769266e-05,
      "loss": 0.3337,
      "step": 1613
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8780481219291687,
      "learning_rate": 2.6402188782489738e-05,
      "loss": 0.3823,
      "step": 1614
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8131493330001831,
      "learning_rate": 2.6356589147286826e-05,
      "loss": 0.3338,
      "step": 1615
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.7759178280830383,
      "learning_rate": 2.6310989512083905e-05,
      "loss": 0.3128,
      "step": 1616
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.7824622392654419,
      "learning_rate": 2.6265389876880987e-05,
      "loss": 0.3189,
      "step": 1617
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8490877151489258,
      "learning_rate": 2.6219790241678065e-05,
      "loss": 0.358,
      "step": 1618
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8330671191215515,
      "learning_rate": 2.6174190606475153e-05,
      "loss": 0.3603,
      "step": 1619
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8165590167045593,
      "learning_rate": 2.6128590971272232e-05,
      "loss": 0.3551,
      "step": 1620
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.35663822293281555,
      "eval_runtime": 19.6772,
      "eval_samples_per_second": 1.423,
      "eval_steps_per_second": 0.356,
      "step": 1620
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8029173612594604,
      "learning_rate": 2.6082991336069313e-05,
      "loss": 0.3549,
      "step": 1621
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8408702611923218,
      "learning_rate": 2.6037391700866392e-05,
      "loss": 0.3584,
      "step": 1622
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.7862275242805481,
      "learning_rate": 2.5991792065663477e-05,
      "loss": 0.3143,
      "step": 1623
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8487365245819092,
      "learning_rate": 2.594619243046056e-05,
      "loss": 0.3433,
      "step": 1624
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8069978356361389,
      "learning_rate": 2.590059279525764e-05,
      "loss": 0.3384,
      "step": 1625
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.81746906042099,
      "learning_rate": 2.585499316005472e-05,
      "loss": 0.3288,
      "step": 1626
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8378018140792847,
      "learning_rate": 2.5809393524851804e-05,
      "loss": 0.3524,
      "step": 1627
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.7984631061553955,
      "learning_rate": 2.5763793889648886e-05,
      "loss": 0.3194,
      "step": 1628
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8371167182922363,
      "learning_rate": 2.5718194254445964e-05,
      "loss": 0.3591,
      "step": 1629
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8061546087265015,
      "learning_rate": 2.5672594619243046e-05,
      "loss": 0.3361,
      "step": 1630
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.3569858968257904,
      "eval_runtime": 19.0111,
      "eval_samples_per_second": 1.473,
      "eval_steps_per_second": 0.368,
      "step": 1630
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8967863321304321,
      "learning_rate": 2.562699498404013e-05,
      "loss": 0.3553,
      "step": 1631
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8157939314842224,
      "learning_rate": 2.5581395348837212e-05,
      "loss": 0.3482,
      "step": 1632
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.835759699344635,
      "learning_rate": 2.553579571363429e-05,
      "loss": 0.3522,
      "step": 1633
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.8867244124412537,
      "learning_rate": 2.5490196078431373e-05,
      "loss": 0.366,
      "step": 1634
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8107739090919495,
      "learning_rate": 2.5444596443228458e-05,
      "loss": 0.3436,
      "step": 1635
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8030595183372498,
      "learning_rate": 2.539899680802554e-05,
      "loss": 0.3246,
      "step": 1636
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.812812328338623,
      "learning_rate": 2.5353397172822618e-05,
      "loss": 0.3518,
      "step": 1637
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8139852285385132,
      "learning_rate": 2.53077975376197e-05,
      "loss": 0.3356,
      "step": 1638
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.795472264289856,
      "learning_rate": 2.5262197902416785e-05,
      "loss": 0.3509,
      "step": 1639
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8404377698898315,
      "learning_rate": 2.5216598267213863e-05,
      "loss": 0.3487,
      "step": 1640
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.3569706976413727,
      "eval_runtime": 19.0009,
      "eval_samples_per_second": 1.474,
      "eval_steps_per_second": 0.368,
      "step": 1640
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.7934169173240662,
      "learning_rate": 2.5170998632010945e-05,
      "loss": 0.3427,
      "step": 1641
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.790808379650116,
      "learning_rate": 2.5125398996808026e-05,
      "loss": 0.3334,
      "step": 1642
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.828813910484314,
      "learning_rate": 2.507979936160511e-05,
      "loss": 0.3553,
      "step": 1643
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8527178764343262,
      "learning_rate": 2.503419972640219e-05,
      "loss": 0.3434,
      "step": 1644
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8209643959999084,
      "learning_rate": 2.498860009119927e-05,
      "loss": 0.3558,
      "step": 1645
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.7981570363044739,
      "learning_rate": 2.4943000455996353e-05,
      "loss": 0.3192,
      "step": 1646
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8820638060569763,
      "learning_rate": 2.4897400820793435e-05,
      "loss": 0.3494,
      "step": 1647
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.7953222393989563,
      "learning_rate": 2.4851801185590517e-05,
      "loss": 0.3244,
      "step": 1648
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8022490739822388,
      "learning_rate": 2.48062015503876e-05,
      "loss": 0.346,
      "step": 1649
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8578386306762695,
      "learning_rate": 2.476060191518468e-05,
      "loss": 0.3375,
      "step": 1650
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.3572315573692322,
      "eval_runtime": 19.0785,
      "eval_samples_per_second": 1.468,
      "eval_steps_per_second": 0.367,
      "step": 1650
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8319166302680969,
      "learning_rate": 2.4715002279981762e-05,
      "loss": 0.351,
      "step": 1651
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.840707004070282,
      "learning_rate": 2.4669402644778844e-05,
      "loss": 0.34,
      "step": 1652
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.7389063835144043,
      "learning_rate": 2.4623803009575925e-05,
      "loss": 0.3147,
      "step": 1653
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.7972502708435059,
      "learning_rate": 2.4578203374373007e-05,
      "loss": 0.3216,
      "step": 1654
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8653942942619324,
      "learning_rate": 2.453260373917009e-05,
      "loss": 0.3818,
      "step": 1655
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.8010607361793518,
      "learning_rate": 2.448700410396717e-05,
      "loss": 0.3402,
      "step": 1656
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.7780823111534119,
      "learning_rate": 2.444140446876425e-05,
      "loss": 0.3322,
      "step": 1657
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.809596061706543,
      "learning_rate": 2.4395804833561334e-05,
      "loss": 0.346,
      "step": 1658
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8238269090652466,
      "learning_rate": 2.4350205198358412e-05,
      "loss": 0.3567,
      "step": 1659
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8283640742301941,
      "learning_rate": 2.4304605563155498e-05,
      "loss": 0.3527,
      "step": 1660
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.3569982647895813,
      "eval_runtime": 19.2893,
      "eval_samples_per_second": 1.452,
      "eval_steps_per_second": 0.363,
      "step": 1660
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8536205887794495,
      "learning_rate": 2.4259005927952576e-05,
      "loss": 0.3364,
      "step": 1661
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8689197301864624,
      "learning_rate": 2.421340629274966e-05,
      "loss": 0.3603,
      "step": 1662
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.861287534236908,
      "learning_rate": 2.416780665754674e-05,
      "loss": 0.3639,
      "step": 1663
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.826146125793457,
      "learning_rate": 2.4122207022343825e-05,
      "loss": 0.3246,
      "step": 1664
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8066340684890747,
      "learning_rate": 2.4076607387140903e-05,
      "loss": 0.3522,
      "step": 1665
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8642566800117493,
      "learning_rate": 2.4031007751937988e-05,
      "loss": 0.3517,
      "step": 1666
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8636859059333801,
      "learning_rate": 2.3985408116735066e-05,
      "loss": 0.3675,
      "step": 1667
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.7902210354804993,
      "learning_rate": 2.393980848153215e-05,
      "loss": 0.3311,
      "step": 1668
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8061851263046265,
      "learning_rate": 2.389420884632923e-05,
      "loss": 0.3454,
      "step": 1669
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.7799895405769348,
      "learning_rate": 2.384860921112631e-05,
      "loss": 0.3306,
      "step": 1670
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.3564049303531647,
      "eval_runtime": 18.9703,
      "eval_samples_per_second": 1.476,
      "eval_steps_per_second": 0.369,
      "step": 1670
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8299709558486938,
      "learning_rate": 2.3803009575923393e-05,
      "loss": 0.3483,
      "step": 1671
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8506625890731812,
      "learning_rate": 2.3757409940720475e-05,
      "loss": 0.3296,
      "step": 1672
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8220917582511902,
      "learning_rate": 2.3711810305517557e-05,
      "loss": 0.332,
      "step": 1673
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.7895132303237915,
      "learning_rate": 2.366621067031464e-05,
      "loss": 0.3298,
      "step": 1674
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8111960291862488,
      "learning_rate": 2.362061103511172e-05,
      "loss": 0.3112,
      "step": 1675
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8449161648750305,
      "learning_rate": 2.3575011399908802e-05,
      "loss": 0.3542,
      "step": 1676
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8357505798339844,
      "learning_rate": 2.3529411764705884e-05,
      "loss": 0.3445,
      "step": 1677
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.8511107563972473,
      "learning_rate": 2.3483812129502965e-05,
      "loss": 0.3456,
      "step": 1678
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8281843662261963,
      "learning_rate": 2.3438212494300047e-05,
      "loss": 0.347,
      "step": 1679
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8068907856941223,
      "learning_rate": 2.339261285909713e-05,
      "loss": 0.3376,
      "step": 1680
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.35658687353134155,
      "eval_runtime": 19.0153,
      "eval_samples_per_second": 1.472,
      "eval_steps_per_second": 0.368,
      "step": 1680
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8254083395004272,
      "learning_rate": 2.334701322389421e-05,
      "loss": 0.3374,
      "step": 1681
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8004122376441956,
      "learning_rate": 2.3301413588691292e-05,
      "loss": 0.335,
      "step": 1682
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8271961212158203,
      "learning_rate": 2.3255813953488374e-05,
      "loss": 0.364,
      "step": 1683
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8114727735519409,
      "learning_rate": 2.3210214318285456e-05,
      "loss": 0.3342,
      "step": 1684
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8433076739311218,
      "learning_rate": 2.3164614683082538e-05,
      "loss": 0.3531,
      "step": 1685
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8108780980110168,
      "learning_rate": 2.311901504787962e-05,
      "loss": 0.3306,
      "step": 1686
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8376524448394775,
      "learning_rate": 2.30734154126767e-05,
      "loss": 0.3405,
      "step": 1687
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8186554908752441,
      "learning_rate": 2.3027815777473783e-05,
      "loss": 0.3366,
      "step": 1688
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.7797662615776062,
      "learning_rate": 2.298221614227086e-05,
      "loss": 0.3315,
      "step": 1689
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.7970744371414185,
      "learning_rate": 2.2936616507067946e-05,
      "loss": 0.3343,
      "step": 1690
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.3567172586917877,
      "eval_runtime": 19.375,
      "eval_samples_per_second": 1.445,
      "eval_steps_per_second": 0.361,
      "step": 1690
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.7889989614486694,
      "learning_rate": 2.2891016871865025e-05,
      "loss": 0.3281,
      "step": 1691
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8547121286392212,
      "learning_rate": 2.284541723666211e-05,
      "loss": 0.3495,
      "step": 1692
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8533220887184143,
      "learning_rate": 2.2799817601459188e-05,
      "loss": 0.3511,
      "step": 1693
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.9014710783958435,
      "learning_rate": 2.2754217966256273e-05,
      "loss": 0.3481,
      "step": 1694
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8289618492126465,
      "learning_rate": 2.270861833105335e-05,
      "loss": 0.339,
      "step": 1695
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8249356150627136,
      "learning_rate": 2.2663018695850437e-05,
      "loss": 0.3538,
      "step": 1696
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8586651682853699,
      "learning_rate": 2.2617419060647515e-05,
      "loss": 0.359,
      "step": 1697
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8021293878555298,
      "learning_rate": 2.25718194254446e-05,
      "loss": 0.3428,
      "step": 1698
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8210233449935913,
      "learning_rate": 2.252621979024168e-05,
      "loss": 0.3393,
      "step": 1699
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.8224427103996277,
      "learning_rate": 2.2480620155038764e-05,
      "loss": 0.3206,
      "step": 1700
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.3561743199825287,
      "eval_runtime": 19.147,
      "eval_samples_per_second": 1.462,
      "eval_steps_per_second": 0.366,
      "step": 1700
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8186107873916626,
      "learning_rate": 2.2435020519835842e-05,
      "loss": 0.342,
      "step": 1701
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8090589046478271,
      "learning_rate": 2.2389420884632924e-05,
      "loss": 0.3182,
      "step": 1702
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8366579413414001,
      "learning_rate": 2.2343821249430005e-05,
      "loss": 0.3406,
      "step": 1703
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8310838341712952,
      "learning_rate": 2.2298221614227087e-05,
      "loss": 0.3156,
      "step": 1704
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.829026997089386,
      "learning_rate": 2.225262197902417e-05,
      "loss": 0.3421,
      "step": 1705
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.9288223385810852,
      "learning_rate": 2.220702234382125e-05,
      "loss": 0.3538,
      "step": 1706
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8505128026008606,
      "learning_rate": 2.2161422708618332e-05,
      "loss": 0.3605,
      "step": 1707
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8131338953971863,
      "learning_rate": 2.2115823073415414e-05,
      "loss": 0.3345,
      "step": 1708
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.798833429813385,
      "learning_rate": 2.2070223438212496e-05,
      "loss": 0.3413,
      "step": 1709
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8337969183921814,
      "learning_rate": 2.2024623803009577e-05,
      "loss": 0.3494,
      "step": 1710
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.3558688759803772,
      "eval_runtime": 19.1518,
      "eval_samples_per_second": 1.462,
      "eval_steps_per_second": 0.366,
      "step": 1710
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8072974681854248,
      "learning_rate": 2.197902416780666e-05,
      "loss": 0.3234,
      "step": 1711
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8112078309059143,
      "learning_rate": 2.193342453260374e-05,
      "loss": 0.3363,
      "step": 1712
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8293721079826355,
      "learning_rate": 2.1887824897400823e-05,
      "loss": 0.3503,
      "step": 1713
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8819985389709473,
      "learning_rate": 2.1842225262197904e-05,
      "loss": 0.3516,
      "step": 1714
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.7671990990638733,
      "learning_rate": 2.1796625626994986e-05,
      "loss": 0.3333,
      "step": 1715
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.845353364944458,
      "learning_rate": 2.1751025991792068e-05,
      "loss": 0.3412,
      "step": 1716
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8249948620796204,
      "learning_rate": 2.170542635658915e-05,
      "loss": 0.3608,
      "step": 1717
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.7952206134796143,
      "learning_rate": 2.165982672138623e-05,
      "loss": 0.3408,
      "step": 1718
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8013413548469543,
      "learning_rate": 2.161422708618331e-05,
      "loss": 0.328,
      "step": 1719
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8645780682563782,
      "learning_rate": 2.1568627450980395e-05,
      "loss": 0.3608,
      "step": 1720
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.3558249771595001,
      "eval_runtime": 18.9857,
      "eval_samples_per_second": 1.475,
      "eval_steps_per_second": 0.369,
      "step": 1720
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8145090937614441,
      "learning_rate": 2.1523027815777473e-05,
      "loss": 0.3374,
      "step": 1721
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.8994247317314148,
      "learning_rate": 2.1477428180574558e-05,
      "loss": 0.3626,
      "step": 1722
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8226855397224426,
      "learning_rate": 2.1431828545371637e-05,
      "loss": 0.3389,
      "step": 1723
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8447384834289551,
      "learning_rate": 2.1386228910168722e-05,
      "loss": 0.3676,
      "step": 1724
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8145758509635925,
      "learning_rate": 2.13406292749658e-05,
      "loss": 0.3438,
      "step": 1725
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8028122782707214,
      "learning_rate": 2.1295029639762885e-05,
      "loss": 0.3405,
      "step": 1726
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8540904521942139,
      "learning_rate": 2.1249430004559964e-05,
      "loss": 0.3452,
      "step": 1727
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8273831605911255,
      "learning_rate": 2.120383036935705e-05,
      "loss": 0.3388,
      "step": 1728
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.7830473780632019,
      "learning_rate": 2.1158230734154127e-05,
      "loss": 0.334,
      "step": 1729
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8468246459960938,
      "learning_rate": 2.1112631098951212e-05,
      "loss": 0.3433,
      "step": 1730
    },
    {
      "epoch": 0.79,
      "eval_loss": 0.3553706705570221,
      "eval_runtime": 18.8238,
      "eval_samples_per_second": 1.487,
      "eval_steps_per_second": 0.372,
      "step": 1730
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.815864622592926,
      "learning_rate": 2.106703146374829e-05,
      "loss": 0.3376,
      "step": 1731
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.804809033870697,
      "learning_rate": 2.1021431828545372e-05,
      "loss": 0.3272,
      "step": 1732
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.7881250977516174,
      "learning_rate": 2.0975832193342454e-05,
      "loss": 0.3344,
      "step": 1733
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8245829343795776,
      "learning_rate": 2.0930232558139536e-05,
      "loss": 0.3478,
      "step": 1734
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8029740452766418,
      "learning_rate": 2.0884632922936617e-05,
      "loss": 0.3429,
      "step": 1735
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8557107448577881,
      "learning_rate": 2.08390332877337e-05,
      "loss": 0.3401,
      "step": 1736
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8481475114822388,
      "learning_rate": 2.079343365253078e-05,
      "loss": 0.334,
      "step": 1737
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8665690422058105,
      "learning_rate": 2.0747834017327863e-05,
      "loss": 0.3649,
      "step": 1738
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.7907549738883972,
      "learning_rate": 2.0702234382124944e-05,
      "loss": 0.3306,
      "step": 1739
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.811202347278595,
      "learning_rate": 2.0656634746922026e-05,
      "loss": 0.3336,
      "step": 1740
    },
    {
      "epoch": 0.79,
      "eval_loss": 0.3552098870277405,
      "eval_runtime": 19.0048,
      "eval_samples_per_second": 1.473,
      "eval_steps_per_second": 0.368,
      "step": 1740
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.7975050210952759,
      "learning_rate": 2.0611035111719108e-05,
      "loss": 0.3339,
      "step": 1741
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8387274742126465,
      "learning_rate": 2.056543547651619e-05,
      "loss": 0.3443,
      "step": 1742
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8204060196876526,
      "learning_rate": 2.051983584131327e-05,
      "loss": 0.3371,
      "step": 1743
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.8510503768920898,
      "learning_rate": 2.0474236206110353e-05,
      "loss": 0.3401,
      "step": 1744
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8482796549797058,
      "learning_rate": 2.0428636570907435e-05,
      "loss": 0.337,
      "step": 1745
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8208746910095215,
      "learning_rate": 2.0383036935704516e-05,
      "loss": 0.3312,
      "step": 1746
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8174223899841309,
      "learning_rate": 2.0337437300501598e-05,
      "loss": 0.341,
      "step": 1747
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.805649995803833,
      "learning_rate": 2.029183766529868e-05,
      "loss": 0.3485,
      "step": 1748
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8207290768623352,
      "learning_rate": 2.0246238030095758e-05,
      "loss": 0.336,
      "step": 1749
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8540436029434204,
      "learning_rate": 2.0200638394892843e-05,
      "loss": 0.3536,
      "step": 1750
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.3553099036216736,
      "eval_runtime": 18.9613,
      "eval_samples_per_second": 1.477,
      "eval_steps_per_second": 0.369,
      "step": 1750
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8083583116531372,
      "learning_rate": 2.0155038759689922e-05,
      "loss": 0.3209,
      "step": 1751
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.826860249042511,
      "learning_rate": 2.0109439124487007e-05,
      "loss": 0.3353,
      "step": 1752
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8210761547088623,
      "learning_rate": 2.0063839489284085e-05,
      "loss": 0.3362,
      "step": 1753
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8487640619277954,
      "learning_rate": 2.001823985408117e-05,
      "loss": 0.3571,
      "step": 1754
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8051680326461792,
      "learning_rate": 1.997264021887825e-05,
      "loss": 0.3394,
      "step": 1755
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8569719791412354,
      "learning_rate": 1.9927040583675334e-05,
      "loss": 0.3325,
      "step": 1756
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8311398029327393,
      "learning_rate": 1.9881440948472412e-05,
      "loss": 0.3482,
      "step": 1757
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8272088766098022,
      "learning_rate": 1.9835841313269497e-05,
      "loss": 0.3486,
      "step": 1758
    },
    {
      "epoch": 0.8,
      "grad_norm": 1.4248172044754028,
      "learning_rate": 1.9790241678066576e-05,
      "loss": 0.3219,
      "step": 1759
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.840705394744873,
      "learning_rate": 1.974464204286366e-05,
      "loss": 0.339,
      "step": 1760
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.3546459376811981,
      "eval_runtime": 19.0023,
      "eval_samples_per_second": 1.474,
      "eval_steps_per_second": 0.368,
      "step": 1760
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8269831538200378,
      "learning_rate": 1.969904240766074e-05,
      "loss": 0.3262,
      "step": 1761
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8176106810569763,
      "learning_rate": 1.965344277245782e-05,
      "loss": 0.3215,
      "step": 1762
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8428852558135986,
      "learning_rate": 1.9607843137254903e-05,
      "loss": 0.3358,
      "step": 1763
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8273266553878784,
      "learning_rate": 1.9562243502051984e-05,
      "loss": 0.3521,
      "step": 1764
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.8344984650611877,
      "learning_rate": 1.9516643866849066e-05,
      "loss": 0.3458,
      "step": 1765
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8323333859443665,
      "learning_rate": 1.9471044231646148e-05,
      "loss": 0.3543,
      "step": 1766
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8865499496459961,
      "learning_rate": 1.942544459644323e-05,
      "loss": 0.3572,
      "step": 1767
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8460800647735596,
      "learning_rate": 1.937984496124031e-05,
      "loss": 0.3425,
      "step": 1768
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.830491304397583,
      "learning_rate": 1.9334245326037393e-05,
      "loss": 0.3302,
      "step": 1769
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.831790030002594,
      "learning_rate": 1.9288645690834475e-05,
      "loss": 0.3305,
      "step": 1770
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.35451439023017883,
      "eval_runtime": 18.9415,
      "eval_samples_per_second": 1.478,
      "eval_steps_per_second": 0.37,
      "step": 1770
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8443323373794556,
      "learning_rate": 1.9243046055631556e-05,
      "loss": 0.3514,
      "step": 1771
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8010796904563904,
      "learning_rate": 1.9197446420428638e-05,
      "loss": 0.3408,
      "step": 1772
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8368244767189026,
      "learning_rate": 1.915184678522572e-05,
      "loss": 0.349,
      "step": 1773
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8440476655960083,
      "learning_rate": 1.91062471500228e-05,
      "loss": 0.3515,
      "step": 1774
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8782945871353149,
      "learning_rate": 1.9060647514819883e-05,
      "loss": 0.3518,
      "step": 1775
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8282709121704102,
      "learning_rate": 1.9015047879616965e-05,
      "loss": 0.3378,
      "step": 1776
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8318287134170532,
      "learning_rate": 1.8969448244414047e-05,
      "loss": 0.3496,
      "step": 1777
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.7970842123031616,
      "learning_rate": 1.892384860921113e-05,
      "loss": 0.3184,
      "step": 1778
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.808746337890625,
      "learning_rate": 1.887824897400821e-05,
      "loss": 0.3422,
      "step": 1779
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.844722330570221,
      "learning_rate": 1.8832649338805292e-05,
      "loss": 0.3535,
      "step": 1780
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.35482144355773926,
      "eval_runtime": 20.0596,
      "eval_samples_per_second": 1.396,
      "eval_steps_per_second": 0.349,
      "step": 1780
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8527645468711853,
      "learning_rate": 1.878704970360237e-05,
      "loss": 0.3572,
      "step": 1781
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.83580482006073,
      "learning_rate": 1.8741450068399455e-05,
      "loss": 0.3502,
      "step": 1782
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8025197386741638,
      "learning_rate": 1.8695850433196534e-05,
      "loss": 0.3241,
      "step": 1783
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.7724499702453613,
      "learning_rate": 1.865025079799362e-05,
      "loss": 0.3245,
      "step": 1784
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8183644413948059,
      "learning_rate": 1.8604651162790697e-05,
      "loss": 0.3284,
      "step": 1785
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.843235194683075,
      "learning_rate": 1.8559051527587782e-05,
      "loss": 0.3237,
      "step": 1786
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.8600271940231323,
      "learning_rate": 1.851345189238486e-05,
      "loss": 0.3379,
      "step": 1787
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8103592395782471,
      "learning_rate": 1.8467852257181946e-05,
      "loss": 0.3451,
      "step": 1788
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.9331439137458801,
      "learning_rate": 1.8422252621979024e-05,
      "loss": 0.3422,
      "step": 1789
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.7992767095565796,
      "learning_rate": 1.837665298677611e-05,
      "loss": 0.3161,
      "step": 1790
    },
    {
      "epoch": 0.82,
      "eval_loss": 0.3550729751586914,
      "eval_runtime": 19.0724,
      "eval_samples_per_second": 1.468,
      "eval_steps_per_second": 0.367,
      "step": 1790
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8258519172668457,
      "learning_rate": 1.8331053351573188e-05,
      "loss": 0.3439,
      "step": 1791
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.7907400727272034,
      "learning_rate": 1.8285453716370273e-05,
      "loss": 0.3261,
      "step": 1792
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8231396675109863,
      "learning_rate": 1.823985408116735e-05,
      "loss": 0.3452,
      "step": 1793
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8543897271156311,
      "learning_rate": 1.8194254445964433e-05,
      "loss": 0.3701,
      "step": 1794
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8704990744590759,
      "learning_rate": 1.8148654810761515e-05,
      "loss": 0.3536,
      "step": 1795
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.846364438533783,
      "learning_rate": 1.8103055175558596e-05,
      "loss": 0.3428,
      "step": 1796
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.871387779712677,
      "learning_rate": 1.8057455540355678e-05,
      "loss": 0.3574,
      "step": 1797
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8834625482559204,
      "learning_rate": 1.801185590515276e-05,
      "loss": 0.3618,
      "step": 1798
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8243701457977295,
      "learning_rate": 1.796625626994984e-05,
      "loss": 0.3322,
      "step": 1799
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8247841000556946,
      "learning_rate": 1.7920656634746923e-05,
      "loss": 0.3445,
      "step": 1800
    },
    {
      "epoch": 0.82,
      "eval_loss": 0.3546763062477112,
      "eval_runtime": 19.7315,
      "eval_samples_per_second": 1.419,
      "eval_steps_per_second": 0.355,
      "step": 1800
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8630883097648621,
      "learning_rate": 1.7875056999544005e-05,
      "loss": 0.3364,
      "step": 1801
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.7862895131111145,
      "learning_rate": 1.7829457364341087e-05,
      "loss": 0.3234,
      "step": 1802
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8019478917121887,
      "learning_rate": 1.778385772913817e-05,
      "loss": 0.3326,
      "step": 1803
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8013519644737244,
      "learning_rate": 1.773825809393525e-05,
      "loss": 0.3209,
      "step": 1804
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8409405946731567,
      "learning_rate": 1.7692658458732332e-05,
      "loss": 0.337,
      "step": 1805
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8028327822685242,
      "learning_rate": 1.7647058823529414e-05,
      "loss": 0.3136,
      "step": 1806
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8128679990768433,
      "learning_rate": 1.7601459188326495e-05,
      "loss": 0.3217,
      "step": 1807
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.8762690424919128,
      "learning_rate": 1.7555859553123577e-05,
      "loss": 0.3554,
      "step": 1808
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.82847660779953,
      "learning_rate": 1.751025991792066e-05,
      "loss": 0.343,
      "step": 1809
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.7954825162887573,
      "learning_rate": 1.746466028271774e-05,
      "loss": 0.3364,
      "step": 1810
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.3547009825706482,
      "eval_runtime": 19.8414,
      "eval_samples_per_second": 1.411,
      "eval_steps_per_second": 0.353,
      "step": 1810
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8007568120956421,
      "learning_rate": 1.741906064751482e-05,
      "loss": 0.337,
      "step": 1811
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8554584980010986,
      "learning_rate": 1.7373461012311904e-05,
      "loss": 0.3576,
      "step": 1812
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8528220653533936,
      "learning_rate": 1.7327861377108982e-05,
      "loss": 0.3395,
      "step": 1813
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8310472965240479,
      "learning_rate": 1.7282261741906067e-05,
      "loss": 0.3271,
      "step": 1814
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.7908942103385925,
      "learning_rate": 1.7236662106703146e-05,
      "loss": 0.3316,
      "step": 1815
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8840450644493103,
      "learning_rate": 1.719106247150023e-05,
      "loss": 0.3356,
      "step": 1816
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.7694292068481445,
      "learning_rate": 1.714546283629731e-05,
      "loss": 0.3157,
      "step": 1817
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8402760028839111,
      "learning_rate": 1.7099863201094394e-05,
      "loss": 0.3444,
      "step": 1818
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8158255815505981,
      "learning_rate": 1.7054263565891473e-05,
      "loss": 0.3282,
      "step": 1819
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8101038336753845,
      "learning_rate": 1.7008663930688558e-05,
      "loss": 0.3357,
      "step": 1820
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.35470905900001526,
      "eval_runtime": 22.0714,
      "eval_samples_per_second": 1.269,
      "eval_steps_per_second": 0.317,
      "step": 1820
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8301025032997131,
      "learning_rate": 1.6963064295485636e-05,
      "loss": 0.3401,
      "step": 1821
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8290889263153076,
      "learning_rate": 1.691746466028272e-05,
      "loss": 0.3482,
      "step": 1822
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8209647536277771,
      "learning_rate": 1.68718650250798e-05,
      "loss": 0.3564,
      "step": 1823
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8127288222312927,
      "learning_rate": 1.682626538987688e-05,
      "loss": 0.3335,
      "step": 1824
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.7942671179771423,
      "learning_rate": 1.6780665754673963e-05,
      "loss": 0.3134,
      "step": 1825
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8012814521789551,
      "learning_rate": 1.6735066119471045e-05,
      "loss": 0.3361,
      "step": 1826
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8532559275627136,
      "learning_rate": 1.6689466484268127e-05,
      "loss": 0.3332,
      "step": 1827
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.7754400372505188,
      "learning_rate": 1.664386684906521e-05,
      "loss": 0.318,
      "step": 1828
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.8331125378608704,
      "learning_rate": 1.659826721386229e-05,
      "loss": 0.3358,
      "step": 1829
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.817933201789856,
      "learning_rate": 1.6552667578659372e-05,
      "loss": 0.3332,
      "step": 1830
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.354523628950119,
      "eval_runtime": 20.8147,
      "eval_samples_per_second": 1.345,
      "eval_steps_per_second": 0.336,
      "step": 1830
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.7978525757789612,
      "learning_rate": 1.6507067943456454e-05,
      "loss": 0.3356,
      "step": 1831
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8010070323944092,
      "learning_rate": 1.6461468308253535e-05,
      "loss": 0.3361,
      "step": 1832
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8485461473464966,
      "learning_rate": 1.6415868673050617e-05,
      "loss": 0.3412,
      "step": 1833
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8424686789512634,
      "learning_rate": 1.63702690378477e-05,
      "loss": 0.3366,
      "step": 1834
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8157220482826233,
      "learning_rate": 1.632466940264478e-05,
      "loss": 0.3368,
      "step": 1835
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8582452535629272,
      "learning_rate": 1.6279069767441862e-05,
      "loss": 0.3612,
      "step": 1836
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8213275074958801,
      "learning_rate": 1.6233470132238944e-05,
      "loss": 0.341,
      "step": 1837
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8489278554916382,
      "learning_rate": 1.6187870497036026e-05,
      "loss": 0.3444,
      "step": 1838
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8359931707382202,
      "learning_rate": 1.6142270861833107e-05,
      "loss": 0.3361,
      "step": 1839
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.7801775336265564,
      "learning_rate": 1.609667122663019e-05,
      "loss": 0.318,
      "step": 1840
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.35444164276123047,
      "eval_runtime": 18.8656,
      "eval_samples_per_second": 1.484,
      "eval_steps_per_second": 0.371,
      "step": 1840
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8361044526100159,
      "learning_rate": 1.6051071591427267e-05,
      "loss": 0.3407,
      "step": 1841
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8520697951316833,
      "learning_rate": 1.6005471956224353e-05,
      "loss": 0.3422,
      "step": 1842
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.7962945103645325,
      "learning_rate": 1.595987232102143e-05,
      "loss": 0.3387,
      "step": 1843
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8080679774284363,
      "learning_rate": 1.5914272685818516e-05,
      "loss": 0.3299,
      "step": 1844
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8117623925209045,
      "learning_rate": 1.5868673050615594e-05,
      "loss": 0.3327,
      "step": 1845
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.852195143699646,
      "learning_rate": 1.582307341541268e-05,
      "loss": 0.3404,
      "step": 1846
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.807907223701477,
      "learning_rate": 1.5777473780209758e-05,
      "loss": 0.3306,
      "step": 1847
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8273969888687134,
      "learning_rate": 1.5731874145006843e-05,
      "loss": 0.3441,
      "step": 1848
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8860800266265869,
      "learning_rate": 1.568627450980392e-05,
      "loss": 0.3282,
      "step": 1849
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8311830163002014,
      "learning_rate": 1.5640674874601006e-05,
      "loss": 0.3373,
      "step": 1850
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.354708194732666,
      "eval_runtime": 19.9719,
      "eval_samples_per_second": 1.402,
      "eval_steps_per_second": 0.35,
      "step": 1850
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8048996329307556,
      "learning_rate": 1.5595075239398085e-05,
      "loss": 0.3388,
      "step": 1851
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.7911148071289062,
      "learning_rate": 1.554947560419517e-05,
      "loss": 0.3256,
      "step": 1852
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.8320807218551636,
      "learning_rate": 1.5503875968992248e-05,
      "loss": 0.3459,
      "step": 1853
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8004910945892334,
      "learning_rate": 1.545827633378933e-05,
      "loss": 0.3222,
      "step": 1854
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8975878357887268,
      "learning_rate": 1.5412676698586412e-05,
      "loss": 0.331,
      "step": 1855
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.7683451771736145,
      "learning_rate": 1.5367077063383493e-05,
      "loss": 0.3269,
      "step": 1856
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8622913956642151,
      "learning_rate": 1.5321477428180575e-05,
      "loss": 0.3455,
      "step": 1857
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8696328997612,
      "learning_rate": 1.5275877792977657e-05,
      "loss": 0.3574,
      "step": 1858
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.836586058139801,
      "learning_rate": 1.5230278157774739e-05,
      "loss": 0.3446,
      "step": 1859
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8396971225738525,
      "learning_rate": 1.5184678522571822e-05,
      "loss": 0.3375,
      "step": 1860
    },
    {
      "epoch": 0.85,
      "eval_loss": 0.35480833053588867,
      "eval_runtime": 20.1797,
      "eval_samples_per_second": 1.388,
      "eval_steps_per_second": 0.347,
      "step": 1860
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8306203484535217,
      "learning_rate": 1.5139078887368902e-05,
      "loss": 0.3373,
      "step": 1861
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.803442656993866,
      "learning_rate": 1.5093479252165984e-05,
      "loss": 0.3243,
      "step": 1862
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8001576066017151,
      "learning_rate": 1.5047879616963066e-05,
      "loss": 0.3408,
      "step": 1863
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8718475103378296,
      "learning_rate": 1.5002279981760147e-05,
      "loss": 0.3316,
      "step": 1864
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8238996863365173,
      "learning_rate": 1.4956680346557227e-05,
      "loss": 0.3394,
      "step": 1865
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8301441669464111,
      "learning_rate": 1.491108071135431e-05,
      "loss": 0.326,
      "step": 1866
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8007735013961792,
      "learning_rate": 1.486548107615139e-05,
      "loss": 0.3308,
      "step": 1867
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8509636521339417,
      "learning_rate": 1.4819881440948474e-05,
      "loss": 0.3528,
      "step": 1868
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.7968654632568359,
      "learning_rate": 1.4774281805745554e-05,
      "loss": 0.3355,
      "step": 1869
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.7697134017944336,
      "learning_rate": 1.4728682170542638e-05,
      "loss": 0.3088,
      "step": 1870
    },
    {
      "epoch": 0.85,
      "eval_loss": 0.35483449697494507,
      "eval_runtime": 20.0293,
      "eval_samples_per_second": 1.398,
      "eval_steps_per_second": 0.349,
      "step": 1870
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8213467001914978,
      "learning_rate": 1.4683082535339718e-05,
      "loss": 0.3368,
      "step": 1871
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8055750131607056,
      "learning_rate": 1.4637482900136801e-05,
      "loss": 0.3236,
      "step": 1872
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8239523768424988,
      "learning_rate": 1.4591883264933881e-05,
      "loss": 0.336,
      "step": 1873
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8376775979995728,
      "learning_rate": 1.4546283629730965e-05,
      "loss": 0.3332,
      "step": 1874
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.8294931650161743,
      "learning_rate": 1.4500683994528045e-05,
      "loss": 0.3466,
      "step": 1875
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8678231239318848,
      "learning_rate": 1.4455084359325128e-05,
      "loss": 0.3482,
      "step": 1876
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8058868050575256,
      "learning_rate": 1.4409484724122208e-05,
      "loss": 0.3438,
      "step": 1877
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8198367357254028,
      "learning_rate": 1.436388508891929e-05,
      "loss": 0.3411,
      "step": 1878
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8211104273796082,
      "learning_rate": 1.4318285453716372e-05,
      "loss": 0.3431,
      "step": 1879
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8408799171447754,
      "learning_rate": 1.4272685818513453e-05,
      "loss": 0.3262,
      "step": 1880
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.3547046184539795,
      "eval_runtime": 19.7685,
      "eval_samples_per_second": 1.416,
      "eval_steps_per_second": 0.354,
      "step": 1880
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.9130158424377441,
      "learning_rate": 1.4227086183310533e-05,
      "loss": 0.3624,
      "step": 1881
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.9254191517829895,
      "learning_rate": 1.4181486548107617e-05,
      "loss": 0.3494,
      "step": 1882
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.809674859046936,
      "learning_rate": 1.4135886912904697e-05,
      "loss": 0.3323,
      "step": 1883
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8233156204223633,
      "learning_rate": 1.409028727770178e-05,
      "loss": 0.3472,
      "step": 1884
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8512375950813293,
      "learning_rate": 1.404468764249886e-05,
      "loss": 0.3388,
      "step": 1885
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8502083420753479,
      "learning_rate": 1.3999088007295944e-05,
      "loss": 0.346,
      "step": 1886
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8712393641471863,
      "learning_rate": 1.3953488372093024e-05,
      "loss": 0.358,
      "step": 1887
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.7779493927955627,
      "learning_rate": 1.3907888736890107e-05,
      "loss": 0.3361,
      "step": 1888
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8024623990058899,
      "learning_rate": 1.3862289101687187e-05,
      "loss": 0.339,
      "step": 1889
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8336456418037415,
      "learning_rate": 1.381668946648427e-05,
      "loss": 0.3388,
      "step": 1890
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.3545686900615692,
      "eval_runtime": 20.1169,
      "eval_samples_per_second": 1.392,
      "eval_steps_per_second": 0.348,
      "step": 1890
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8277023434638977,
      "learning_rate": 1.377108983128135e-05,
      "loss": 0.3403,
      "step": 1891
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8159246444702148,
      "learning_rate": 1.3725490196078432e-05,
      "loss": 0.3469,
      "step": 1892
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8369227051734924,
      "learning_rate": 1.3679890560875514e-05,
      "loss": 0.3568,
      "step": 1893
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8269848227500916,
      "learning_rate": 1.3634290925672596e-05,
      "loss": 0.3436,
      "step": 1894
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.9031785726547241,
      "learning_rate": 1.3588691290469676e-05,
      "loss": 0.359,
      "step": 1895
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8029106259346008,
      "learning_rate": 1.354309165526676e-05,
      "loss": 0.3232,
      "step": 1896
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.8643222451210022,
      "learning_rate": 1.349749202006384e-05,
      "loss": 0.3411,
      "step": 1897
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8164620399475098,
      "learning_rate": 1.3451892384860923e-05,
      "loss": 0.3333,
      "step": 1898
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8040909171104431,
      "learning_rate": 1.3406292749658003e-05,
      "loss": 0.3334,
      "step": 1899
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8196666240692139,
      "learning_rate": 1.3360693114455086e-05,
      "loss": 0.3512,
      "step": 1900
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.35445669293403625,
      "eval_runtime": 20.2351,
      "eval_samples_per_second": 1.384,
      "eval_steps_per_second": 0.346,
      "step": 1900
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8267871141433716,
      "learning_rate": 1.3315093479252166e-05,
      "loss": 0.3121,
      "step": 1901
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8019816875457764,
      "learning_rate": 1.326949384404925e-05,
      "loss": 0.3437,
      "step": 1902
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8552934527397156,
      "learning_rate": 1.322389420884633e-05,
      "loss": 0.3614,
      "step": 1903
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8025230169296265,
      "learning_rate": 1.3178294573643413e-05,
      "loss": 0.3104,
      "step": 1904
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8386091589927673,
      "learning_rate": 1.3132694938440493e-05,
      "loss": 0.3371,
      "step": 1905
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8556039929389954,
      "learning_rate": 1.3087095303237577e-05,
      "loss": 0.331,
      "step": 1906
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8279280066490173,
      "learning_rate": 1.3041495668034657e-05,
      "loss": 0.3388,
      "step": 1907
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8014733791351318,
      "learning_rate": 1.2995896032831738e-05,
      "loss": 0.3343,
      "step": 1908
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.9208333492279053,
      "learning_rate": 1.295029639762882e-05,
      "loss": 0.3638,
      "step": 1909
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.818374514579773,
      "learning_rate": 1.2904696762425902e-05,
      "loss": 0.3266,
      "step": 1910
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.3544815480709076,
      "eval_runtime": 20.5453,
      "eval_samples_per_second": 1.363,
      "eval_steps_per_second": 0.341,
      "step": 1910
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8163886666297913,
      "learning_rate": 1.2859097127222982e-05,
      "loss": 0.3529,
      "step": 1911
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8497297167778015,
      "learning_rate": 1.2813497492020065e-05,
      "loss": 0.3364,
      "step": 1912
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.814104437828064,
      "learning_rate": 1.2767897856817145e-05,
      "loss": 0.3446,
      "step": 1913
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.856905996799469,
      "learning_rate": 1.2722298221614229e-05,
      "loss": 0.3456,
      "step": 1914
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.816177248954773,
      "learning_rate": 1.2676698586411309e-05,
      "loss": 0.3304,
      "step": 1915
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.7743247151374817,
      "learning_rate": 1.2631098951208392e-05,
      "loss": 0.3247,
      "step": 1916
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8495445251464844,
      "learning_rate": 1.2585499316005472e-05,
      "loss": 0.332,
      "step": 1917
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8720409274101257,
      "learning_rate": 1.2539899680802556e-05,
      "loss": 0.3723,
      "step": 1918
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.8387491106987,
      "learning_rate": 1.2494300045599636e-05,
      "loss": 0.3447,
      "step": 1919
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8160862922668457,
      "learning_rate": 1.2448700410396718e-05,
      "loss": 0.3588,
      "step": 1920
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.35426434874534607,
      "eval_runtime": 20.709,
      "eval_samples_per_second": 1.352,
      "eval_steps_per_second": 0.338,
      "step": 1920
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8311136364936829,
      "learning_rate": 1.24031007751938e-05,
      "loss": 0.3394,
      "step": 1921
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8703175187110901,
      "learning_rate": 1.2357501139990881e-05,
      "loss": 0.3543,
      "step": 1922
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.812587320804596,
      "learning_rate": 1.2311901504787963e-05,
      "loss": 0.34,
      "step": 1923
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8485066294670105,
      "learning_rate": 1.2266301869585044e-05,
      "loss": 0.3396,
      "step": 1924
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8266170620918274,
      "learning_rate": 1.2220702234382125e-05,
      "loss": 0.3429,
      "step": 1925
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8497158288955688,
      "learning_rate": 1.2175102599179206e-05,
      "loss": 0.3378,
      "step": 1926
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8195666074752808,
      "learning_rate": 1.2129502963976288e-05,
      "loss": 0.3425,
      "step": 1927
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8512323498725891,
      "learning_rate": 1.208390332877337e-05,
      "loss": 0.3348,
      "step": 1928
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8279510140419006,
      "learning_rate": 1.2038303693570451e-05,
      "loss": 0.3457,
      "step": 1929
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8365693092346191,
      "learning_rate": 1.1992704058367533e-05,
      "loss": 0.3457,
      "step": 1930
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.3541213572025299,
      "eval_runtime": 20.5411,
      "eval_samples_per_second": 1.363,
      "eval_steps_per_second": 0.341,
      "step": 1930
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8179658055305481,
      "learning_rate": 1.1947104423164615e-05,
      "loss": 0.3281,
      "step": 1931
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8157138228416443,
      "learning_rate": 1.1901504787961697e-05,
      "loss": 0.3489,
      "step": 1932
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8499870300292969,
      "learning_rate": 1.1855905152758778e-05,
      "loss": 0.3206,
      "step": 1933
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8814250230789185,
      "learning_rate": 1.181030551755586e-05,
      "loss": 0.352,
      "step": 1934
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8282210230827332,
      "learning_rate": 1.1764705882352942e-05,
      "loss": 0.3508,
      "step": 1935
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8317981958389282,
      "learning_rate": 1.1719106247150024e-05,
      "loss": 0.3455,
      "step": 1936
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8515220880508423,
      "learning_rate": 1.1673506611947105e-05,
      "loss": 0.3395,
      "step": 1937
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8439879417419434,
      "learning_rate": 1.1627906976744187e-05,
      "loss": 0.3325,
      "step": 1938
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8174811601638794,
      "learning_rate": 1.1582307341541269e-05,
      "loss": 0.3359,
      "step": 1939
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.8104060888290405,
      "learning_rate": 1.153670770633835e-05,
      "loss": 0.3319,
      "step": 1940
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.35404470562934875,
      "eval_runtime": 19.9655,
      "eval_samples_per_second": 1.402,
      "eval_steps_per_second": 0.351,
      "step": 1940
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.7619262933731079,
      "learning_rate": 1.149110807113543e-05,
      "loss": 0.3273,
      "step": 1941
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.7890914082527161,
      "learning_rate": 1.1445508435932512e-05,
      "loss": 0.327,
      "step": 1942
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8490578532218933,
      "learning_rate": 1.1399908800729594e-05,
      "loss": 0.3246,
      "step": 1943
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8122377991676331,
      "learning_rate": 1.1354309165526676e-05,
      "loss": 0.345,
      "step": 1944
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8249675631523132,
      "learning_rate": 1.1308709530323757e-05,
      "loss": 0.3319,
      "step": 1945
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8207200169563293,
      "learning_rate": 1.126310989512084e-05,
      "loss": 0.3414,
      "step": 1946
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8593723773956299,
      "learning_rate": 1.1217510259917921e-05,
      "loss": 0.3495,
      "step": 1947
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.799826979637146,
      "learning_rate": 1.1171910624715003e-05,
      "loss": 0.3163,
      "step": 1948
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8258078694343567,
      "learning_rate": 1.1126310989512084e-05,
      "loss": 0.3364,
      "step": 1949
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8616490364074707,
      "learning_rate": 1.1080711354309166e-05,
      "loss": 0.345,
      "step": 1950
    },
    {
      "epoch": 0.89,
      "eval_loss": 0.35401859879493713,
      "eval_runtime": 20.6827,
      "eval_samples_per_second": 1.354,
      "eval_steps_per_second": 0.338,
      "step": 1950
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8457894325256348,
      "learning_rate": 1.1035111719106248e-05,
      "loss": 0.3563,
      "step": 1951
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8008802533149719,
      "learning_rate": 1.098951208390333e-05,
      "loss": 0.3286,
      "step": 1952
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8404101729393005,
      "learning_rate": 1.0943912448700411e-05,
      "loss": 0.3471,
      "step": 1953
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8518589735031128,
      "learning_rate": 1.0898312813497493e-05,
      "loss": 0.3488,
      "step": 1954
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8108507990837097,
      "learning_rate": 1.0852713178294575e-05,
      "loss": 0.3382,
      "step": 1955
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.771907389163971,
      "learning_rate": 1.0807113543091655e-05,
      "loss": 0.3183,
      "step": 1956
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.7840840816497803,
      "learning_rate": 1.0761513907888737e-05,
      "loss": 0.317,
      "step": 1957
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.796921968460083,
      "learning_rate": 1.0715914272685818e-05,
      "loss": 0.3257,
      "step": 1958
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8107333183288574,
      "learning_rate": 1.06703146374829e-05,
      "loss": 0.3522,
      "step": 1959
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8181228041648865,
      "learning_rate": 1.0624715002279982e-05,
      "loss": 0.3404,
      "step": 1960
    },
    {
      "epoch": 0.89,
      "eval_loss": 0.3538888990879059,
      "eval_runtime": 20.3575,
      "eval_samples_per_second": 1.375,
      "eval_steps_per_second": 0.344,
      "step": 1960
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8123586773872375,
      "learning_rate": 1.0579115367077063e-05,
      "loss": 0.3273,
      "step": 1961
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.8805697560310364,
      "learning_rate": 1.0533515731874145e-05,
      "loss": 0.3428,
      "step": 1962
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.779415488243103,
      "learning_rate": 1.0487916096671227e-05,
      "loss": 0.3337,
      "step": 1963
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8597623705863953,
      "learning_rate": 1.0442316461468309e-05,
      "loss": 0.3286,
      "step": 1964
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.7904038429260254,
      "learning_rate": 1.039671682626539e-05,
      "loss": 0.3436,
      "step": 1965
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8617693185806274,
      "learning_rate": 1.0351117191062472e-05,
      "loss": 0.3548,
      "step": 1966
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8419015407562256,
      "learning_rate": 1.0305517555859554e-05,
      "loss": 0.3329,
      "step": 1967
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8732928037643433,
      "learning_rate": 1.0259917920656636e-05,
      "loss": 0.3533,
      "step": 1968
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8149675726890564,
      "learning_rate": 1.0214318285453717e-05,
      "loss": 0.3471,
      "step": 1969
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8281083703041077,
      "learning_rate": 1.0168718650250799e-05,
      "loss": 0.3555,
      "step": 1970
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.35382261872291565,
      "eval_runtime": 20.9018,
      "eval_samples_per_second": 1.34,
      "eval_steps_per_second": 0.335,
      "step": 1970
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8474620580673218,
      "learning_rate": 1.0123119015047879e-05,
      "loss": 0.3436,
      "step": 1971
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.7994499206542969,
      "learning_rate": 1.0077519379844961e-05,
      "loss": 0.3294,
      "step": 1972
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8625378012657166,
      "learning_rate": 1.0031919744642043e-05,
      "loss": 0.326,
      "step": 1973
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.9084400534629822,
      "learning_rate": 9.986320109439124e-06,
      "loss": 0.3488,
      "step": 1974
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8352153301239014,
      "learning_rate": 9.940720474236206e-06,
      "loss": 0.3397,
      "step": 1975
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8399555087089539,
      "learning_rate": 9.895120839033288e-06,
      "loss": 0.3468,
      "step": 1976
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8331840634346008,
      "learning_rate": 9.84952120383037e-06,
      "loss": 0.3369,
      "step": 1977
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.9115501046180725,
      "learning_rate": 9.803921568627451e-06,
      "loss": 0.365,
      "step": 1978
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8095980882644653,
      "learning_rate": 9.758321933424533e-06,
      "loss": 0.3397,
      "step": 1979
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.7896193861961365,
      "learning_rate": 9.712722298221615e-06,
      "loss": 0.3353,
      "step": 1980
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.35391122102737427,
      "eval_runtime": 20.1611,
      "eval_samples_per_second": 1.389,
      "eval_steps_per_second": 0.347,
      "step": 1980
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.7885720133781433,
      "learning_rate": 9.667122663018696e-06,
      "loss": 0.3188,
      "step": 1981
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.7856804728507996,
      "learning_rate": 9.621523027815778e-06,
      "loss": 0.3222,
      "step": 1982
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.868549644947052,
      "learning_rate": 9.57592339261286e-06,
      "loss": 0.3639,
      "step": 1983
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.882917046546936,
      "learning_rate": 9.530323757409942e-06,
      "loss": 0.3644,
      "step": 1984
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8365908265113831,
      "learning_rate": 9.484724122207023e-06,
      "loss": 0.3341,
      "step": 1985
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8351644277572632,
      "learning_rate": 9.439124487004105e-06,
      "loss": 0.3459,
      "step": 1986
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.85927414894104,
      "learning_rate": 9.393524851801185e-06,
      "loss": 0.36,
      "step": 1987
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.7935860753059387,
      "learning_rate": 9.347925216598267e-06,
      "loss": 0.319,
      "step": 1988
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.855145275592804,
      "learning_rate": 9.302325581395349e-06,
      "loss": 0.3292,
      "step": 1989
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8166558146476746,
      "learning_rate": 9.25672594619243e-06,
      "loss": 0.3407,
      "step": 1990
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.3539559543132782,
      "eval_runtime": 20.7137,
      "eval_samples_per_second": 1.352,
      "eval_steps_per_second": 0.338,
      "step": 1990
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.794497013092041,
      "learning_rate": 9.211126310989512e-06,
      "loss": 0.3297,
      "step": 1991
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.9098047614097595,
      "learning_rate": 9.165526675786594e-06,
      "loss": 0.3416,
      "step": 1992
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.7829924821853638,
      "learning_rate": 9.119927040583676e-06,
      "loss": 0.3399,
      "step": 1993
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8300178050994873,
      "learning_rate": 9.074327405380757e-06,
      "loss": 0.339,
      "step": 1994
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8618704676628113,
      "learning_rate": 9.028727770177839e-06,
      "loss": 0.3351,
      "step": 1995
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.887930154800415,
      "learning_rate": 8.98312813497492e-06,
      "loss": 0.3402,
      "step": 1996
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.809892475605011,
      "learning_rate": 8.937528499772002e-06,
      "loss": 0.3438,
      "step": 1997
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.7907106280326843,
      "learning_rate": 8.891928864569084e-06,
      "loss": 0.3333,
      "step": 1998
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.7651233077049255,
      "learning_rate": 8.846329229366166e-06,
      "loss": 0.3213,
      "step": 1999
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8054981231689453,
      "learning_rate": 8.800729594163248e-06,
      "loss": 0.328,
      "step": 2000
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.35396915674209595,
      "eval_runtime": 19.9787,
      "eval_samples_per_second": 1.401,
      "eval_steps_per_second": 0.35,
      "step": 2000
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8074979782104492,
      "learning_rate": 8.75512995896033e-06,
      "loss": 0.336,
      "step": 2001
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8073883056640625,
      "learning_rate": 8.70953032375741e-06,
      "loss": 0.3306,
      "step": 2002
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8059557676315308,
      "learning_rate": 8.663930688554491e-06,
      "loss": 0.3364,
      "step": 2003
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8500943779945374,
      "learning_rate": 8.618331053351573e-06,
      "loss": 0.3388,
      "step": 2004
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8143544793128967,
      "learning_rate": 8.572731418148655e-06,
      "loss": 0.3297,
      "step": 2005
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8129506707191467,
      "learning_rate": 8.527131782945736e-06,
      "loss": 0.3421,
      "step": 2006
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.8499677181243896,
      "learning_rate": 8.481532147742818e-06,
      "loss": 0.3311,
      "step": 2007
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.7764211297035217,
      "learning_rate": 8.4359325125399e-06,
      "loss": 0.3198,
      "step": 2008
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8650667667388916,
      "learning_rate": 8.390332877336982e-06,
      "loss": 0.3479,
      "step": 2009
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8370277285575867,
      "learning_rate": 8.344733242134063e-06,
      "loss": 0.3457,
      "step": 2010
    },
    {
      "epoch": 0.92,
      "eval_loss": 0.3540601432323456,
      "eval_runtime": 20.9003,
      "eval_samples_per_second": 1.34,
      "eval_steps_per_second": 0.335,
      "step": 2010
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8421793580055237,
      "learning_rate": 8.299133606931145e-06,
      "loss": 0.3529,
      "step": 2011
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8460060358047485,
      "learning_rate": 8.253533971728227e-06,
      "loss": 0.3422,
      "step": 2012
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8409944176673889,
      "learning_rate": 8.207934336525308e-06,
      "loss": 0.334,
      "step": 2013
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8694727420806885,
      "learning_rate": 8.16233470132239e-06,
      "loss": 0.3621,
      "step": 2014
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.813942015171051,
      "learning_rate": 8.116735066119472e-06,
      "loss": 0.3467,
      "step": 2015
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8313078880310059,
      "learning_rate": 8.071135430916554e-06,
      "loss": 0.3257,
      "step": 2016
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8491564989089966,
      "learning_rate": 8.025535795713634e-06,
      "loss": 0.3594,
      "step": 2017
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.9133617877960205,
      "learning_rate": 7.979936160510715e-06,
      "loss": 0.3492,
      "step": 2018
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8548010587692261,
      "learning_rate": 7.934336525307797e-06,
      "loss": 0.3297,
      "step": 2019
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8131832480430603,
      "learning_rate": 7.888736890104879e-06,
      "loss": 0.3388,
      "step": 2020
    },
    {
      "epoch": 0.92,
      "eval_loss": 0.3540656864643097,
      "eval_runtime": 20.3087,
      "eval_samples_per_second": 1.379,
      "eval_steps_per_second": 0.345,
      "step": 2020
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8627143502235413,
      "learning_rate": 7.84313725490196e-06,
      "loss": 0.3451,
      "step": 2021
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8710396885871887,
      "learning_rate": 7.797537619699042e-06,
      "loss": 0.3491,
      "step": 2022
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8423619866371155,
      "learning_rate": 7.751937984496124e-06,
      "loss": 0.3351,
      "step": 2023
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.7903690934181213,
      "learning_rate": 7.706338349293206e-06,
      "loss": 0.3269,
      "step": 2024
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8393657803535461,
      "learning_rate": 7.660738714090288e-06,
      "loss": 0.3591,
      "step": 2025
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.818337619304657,
      "learning_rate": 7.615139078887369e-06,
      "loss": 0.3277,
      "step": 2026
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8536666035652161,
      "learning_rate": 7.569539443684451e-06,
      "loss": 0.3385,
      "step": 2027
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.7855940461158752,
      "learning_rate": 7.523939808481533e-06,
      "loss": 0.3235,
      "step": 2028
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.8503609895706177,
      "learning_rate": 7.478340173278614e-06,
      "loss": 0.3269,
      "step": 2029
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8175629377365112,
      "learning_rate": 7.432740538075695e-06,
      "loss": 0.3251,
      "step": 2030
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.35404643416404724,
      "eval_runtime": 19.8105,
      "eval_samples_per_second": 1.413,
      "eval_steps_per_second": 0.353,
      "step": 2030
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8303265571594238,
      "learning_rate": 7.387140902872777e-06,
      "loss": 0.3384,
      "step": 2031
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8031489849090576,
      "learning_rate": 7.341541267669859e-06,
      "loss": 0.3457,
      "step": 2032
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8192217350006104,
      "learning_rate": 7.295941632466941e-06,
      "loss": 0.3284,
      "step": 2033
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8637712597846985,
      "learning_rate": 7.250341997264022e-06,
      "loss": 0.3571,
      "step": 2034
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8116298317909241,
      "learning_rate": 7.204742362061104e-06,
      "loss": 0.3367,
      "step": 2035
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8289496898651123,
      "learning_rate": 7.159142726858186e-06,
      "loss": 0.3421,
      "step": 2036
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.7838993072509766,
      "learning_rate": 7.113543091655267e-06,
      "loss": 0.3025,
      "step": 2037
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.7901455163955688,
      "learning_rate": 7.067943456452348e-06,
      "loss": 0.3371,
      "step": 2038
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8628932237625122,
      "learning_rate": 7.02234382124943e-06,
      "loss": 0.3383,
      "step": 2039
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8549111485481262,
      "learning_rate": 6.976744186046512e-06,
      "loss": 0.357,
      "step": 2040
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.35389232635498047,
      "eval_runtime": 19.6965,
      "eval_samples_per_second": 1.422,
      "eval_steps_per_second": 0.355,
      "step": 2040
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8260614275932312,
      "learning_rate": 6.931144550843594e-06,
      "loss": 0.3206,
      "step": 2041
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.841353714466095,
      "learning_rate": 6.885544915640675e-06,
      "loss": 0.3291,
      "step": 2042
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8368021845817566,
      "learning_rate": 6.839945280437757e-06,
      "loss": 0.3492,
      "step": 2043
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8355389833450317,
      "learning_rate": 6.794345645234838e-06,
      "loss": 0.3458,
      "step": 2044
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8132802248001099,
      "learning_rate": 6.74874601003192e-06,
      "loss": 0.3162,
      "step": 2045
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8105396032333374,
      "learning_rate": 6.7031463748290014e-06,
      "loss": 0.3063,
      "step": 2046
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.780116617679596,
      "learning_rate": 6.657546739626083e-06,
      "loss": 0.3285,
      "step": 2047
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8291633725166321,
      "learning_rate": 6.611947104423165e-06,
      "loss": 0.3335,
      "step": 2048
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8367629051208496,
      "learning_rate": 6.566347469220247e-06,
      "loss": 0.3419,
      "step": 2049
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.8254304528236389,
      "learning_rate": 6.520747834017328e-06,
      "loss": 0.3436,
      "step": 2050
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.3537757098674774,
      "eval_runtime": 19.7608,
      "eval_samples_per_second": 1.417,
      "eval_steps_per_second": 0.354,
      "step": 2050
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.7851077318191528,
      "learning_rate": 6.47514819881441e-06,
      "loss": 0.3306,
      "step": 2051
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8641871809959412,
      "learning_rate": 6.429548563611491e-06,
      "loss": 0.34,
      "step": 2052
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.875881552696228,
      "learning_rate": 6.383948928408573e-06,
      "loss": 0.3491,
      "step": 2053
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8497762084007263,
      "learning_rate": 6.3383492932056544e-06,
      "loss": 0.3513,
      "step": 2054
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8041360974311829,
      "learning_rate": 6.292749658002736e-06,
      "loss": 0.3361,
      "step": 2055
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8663139343261719,
      "learning_rate": 6.247150022799818e-06,
      "loss": 0.3366,
      "step": 2056
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8280770182609558,
      "learning_rate": 6.2015503875969e-06,
      "loss": 0.3526,
      "step": 2057
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.842253565788269,
      "learning_rate": 6.155950752393981e-06,
      "loss": 0.3274,
      "step": 2058
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8146688938140869,
      "learning_rate": 6.110351117191062e-06,
      "loss": 0.3423,
      "step": 2059
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8018572330474854,
      "learning_rate": 6.064751481988144e-06,
      "loss": 0.3169,
      "step": 2060
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.35372114181518555,
      "eval_runtime": 19.7858,
      "eval_samples_per_second": 1.415,
      "eval_steps_per_second": 0.354,
      "step": 2060
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8574947118759155,
      "learning_rate": 6.019151846785226e-06,
      "loss": 0.3683,
      "step": 2061
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8088916540145874,
      "learning_rate": 5.9735522115823075e-06,
      "loss": 0.3295,
      "step": 2062
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.7937730550765991,
      "learning_rate": 5.927952576379389e-06,
      "loss": 0.3345,
      "step": 2063
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8222320675849915,
      "learning_rate": 5.882352941176471e-06,
      "loss": 0.3271,
      "step": 2064
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8784458041191101,
      "learning_rate": 5.836753305973553e-06,
      "loss": 0.3435,
      "step": 2065
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.857282817363739,
      "learning_rate": 5.791153670770634e-06,
      "loss": 0.3362,
      "step": 2066
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.7424783110618591,
      "learning_rate": 5.745554035567715e-06,
      "loss": 0.3035,
      "step": 2067
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8212710618972778,
      "learning_rate": 5.699954400364797e-06,
      "loss": 0.3311,
      "step": 2068
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8231950998306274,
      "learning_rate": 5.654354765161879e-06,
      "loss": 0.3436,
      "step": 2069
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8013624548912048,
      "learning_rate": 5.6087551299589605e-06,
      "loss": 0.3244,
      "step": 2070
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.3536633551120758,
      "eval_runtime": 19.7898,
      "eval_samples_per_second": 1.415,
      "eval_steps_per_second": 0.354,
      "step": 2070
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8232308030128479,
      "learning_rate": 5.563155494756042e-06,
      "loss": 0.3322,
      "step": 2071
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.8284724354743958,
      "learning_rate": 5.517555859553124e-06,
      "loss": 0.3474,
      "step": 2072
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.7981323599815369,
      "learning_rate": 5.471956224350206e-06,
      "loss": 0.3314,
      "step": 2073
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.7892775535583496,
      "learning_rate": 5.426356589147287e-06,
      "loss": 0.3456,
      "step": 2074
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8134260177612305,
      "learning_rate": 5.380756953944368e-06,
      "loss": 0.3439,
      "step": 2075
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8565177917480469,
      "learning_rate": 5.33515731874145e-06,
      "loss": 0.3372,
      "step": 2076
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8024502396583557,
      "learning_rate": 5.289557683538532e-06,
      "loss": 0.3226,
      "step": 2077
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8735952973365784,
      "learning_rate": 5.2439580483356135e-06,
      "loss": 0.3674,
      "step": 2078
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8029111623764038,
      "learning_rate": 5.198358413132695e-06,
      "loss": 0.3293,
      "step": 2079
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.7668672800064087,
      "learning_rate": 5.152758777929777e-06,
      "loss": 0.3062,
      "step": 2080
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.35355642437934875,
      "eval_runtime": 19.8253,
      "eval_samples_per_second": 1.412,
      "eval_steps_per_second": 0.353,
      "step": 2080
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8596630692481995,
      "learning_rate": 5.107159142726859e-06,
      "loss": 0.3573,
      "step": 2081
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8058382868766785,
      "learning_rate": 5.0615595075239396e-06,
      "loss": 0.3208,
      "step": 2082
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8100380301475525,
      "learning_rate": 5.015959872321021e-06,
      "loss": 0.3228,
      "step": 2083
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8688908815383911,
      "learning_rate": 4.970360237118103e-06,
      "loss": 0.3434,
      "step": 2084
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8467679619789124,
      "learning_rate": 4.924760601915185e-06,
      "loss": 0.3249,
      "step": 2085
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8613777160644531,
      "learning_rate": 4.8791609667122665e-06,
      "loss": 0.341,
      "step": 2086
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8214209079742432,
      "learning_rate": 4.833561331509348e-06,
      "loss": 0.3267,
      "step": 2087
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8228405714035034,
      "learning_rate": 4.78796169630643e-06,
      "loss": 0.3432,
      "step": 2088
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8442954421043396,
      "learning_rate": 4.742362061103512e-06,
      "loss": 0.3347,
      "step": 2089
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8544167876243591,
      "learning_rate": 4.6967624259005926e-06,
      "loss": 0.344,
      "step": 2090
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.35357967019081116,
      "eval_runtime": 19.8199,
      "eval_samples_per_second": 1.413,
      "eval_steps_per_second": 0.353,
      "step": 2090
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8619562983512878,
      "learning_rate": 4.651162790697674e-06,
      "loss": 0.3449,
      "step": 2091
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8240465521812439,
      "learning_rate": 4.605563155494756e-06,
      "loss": 0.3442,
      "step": 2092
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.875579297542572,
      "learning_rate": 4.559963520291838e-06,
      "loss": 0.3475,
      "step": 2093
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8105667233467102,
      "learning_rate": 4.5143638850889195e-06,
      "loss": 0.3321,
      "step": 2094
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.8081218004226685,
      "learning_rate": 4.468764249886001e-06,
      "loss": 0.333,
      "step": 2095
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8312126398086548,
      "learning_rate": 4.423164614683083e-06,
      "loss": 0.3474,
      "step": 2096
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8706237077713013,
      "learning_rate": 4.377564979480165e-06,
      "loss": 0.3505,
      "step": 2097
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8144000172615051,
      "learning_rate": 4.331965344277246e-06,
      "loss": 0.3418,
      "step": 2098
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8141364455223083,
      "learning_rate": 4.286365709074327e-06,
      "loss": 0.3185,
      "step": 2099
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8204436898231506,
      "learning_rate": 4.240766073871409e-06,
      "loss": 0.3527,
      "step": 2100
    },
    {
      "epoch": 0.96,
      "eval_loss": 0.35366660356521606,
      "eval_runtime": 20.5995,
      "eval_samples_per_second": 1.359,
      "eval_steps_per_second": 0.34,
      "step": 2100
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8323388695716858,
      "learning_rate": 4.195166438668491e-06,
      "loss": 0.3309,
      "step": 2101
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8300788998603821,
      "learning_rate": 4.1495668034655725e-06,
      "loss": 0.3347,
      "step": 2102
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8446533679962158,
      "learning_rate": 4.103967168262654e-06,
      "loss": 0.3426,
      "step": 2103
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8308296799659729,
      "learning_rate": 4.058367533059736e-06,
      "loss": 0.3418,
      "step": 2104
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.829054057598114,
      "learning_rate": 4.012767897856817e-06,
      "loss": 0.3392,
      "step": 2105
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.874564528465271,
      "learning_rate": 3.967168262653899e-06,
      "loss": 0.3456,
      "step": 2106
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8792354464530945,
      "learning_rate": 3.92156862745098e-06,
      "loss": 0.3558,
      "step": 2107
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.9232382774353027,
      "learning_rate": 3.875968992248062e-06,
      "loss": 0.3672,
      "step": 2108
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8330267071723938,
      "learning_rate": 3.830369357045144e-06,
      "loss": 0.3293,
      "step": 2109
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.7642355561256409,
      "learning_rate": 3.7847697218422255e-06,
      "loss": 0.3104,
      "step": 2110
    },
    {
      "epoch": 0.96,
      "eval_loss": 0.35369256138801575,
      "eval_runtime": 20.5064,
      "eval_samples_per_second": 1.365,
      "eval_steps_per_second": 0.341,
      "step": 2110
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8061398863792419,
      "learning_rate": 3.739170086639307e-06,
      "loss": 0.3413,
      "step": 2111
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.7958340644836426,
      "learning_rate": 3.6935704514363886e-06,
      "loss": 0.3367,
      "step": 2112
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8436947464942932,
      "learning_rate": 3.6479708162334703e-06,
      "loss": 0.3375,
      "step": 2113
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.7928789854049683,
      "learning_rate": 3.602371181030552e-06,
      "loss": 0.321,
      "step": 2114
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8269808292388916,
      "learning_rate": 3.5567715458276333e-06,
      "loss": 0.3343,
      "step": 2115
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.8675149083137512,
      "learning_rate": 3.511171910624715e-06,
      "loss": 0.3571,
      "step": 2116
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8159371018409729,
      "learning_rate": 3.465572275421797e-06,
      "loss": 0.3379,
      "step": 2117
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8035234212875366,
      "learning_rate": 3.4199726402188785e-06,
      "loss": 0.3491,
      "step": 2118
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.812272310256958,
      "learning_rate": 3.37437300501596e-06,
      "loss": 0.3532,
      "step": 2119
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.791612982749939,
      "learning_rate": 3.3287733698130416e-06,
      "loss": 0.3355,
      "step": 2120
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.35358789563179016,
      "eval_runtime": 21.0441,
      "eval_samples_per_second": 1.331,
      "eval_steps_per_second": 0.333,
      "step": 2120
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8322513699531555,
      "learning_rate": 3.2831737346101233e-06,
      "loss": 0.3336,
      "step": 2121
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8068985342979431,
      "learning_rate": 3.237574099407205e-06,
      "loss": 0.3309,
      "step": 2122
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.7732910513877869,
      "learning_rate": 3.1919744642042864e-06,
      "loss": 0.3142,
      "step": 2123
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8337993621826172,
      "learning_rate": 3.146374829001368e-06,
      "loss": 0.3453,
      "step": 2124
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8158725500106812,
      "learning_rate": 3.10077519379845e-06,
      "loss": 0.3417,
      "step": 2125
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8460199236869812,
      "learning_rate": 3.055175558595531e-06,
      "loss": 0.3438,
      "step": 2126
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8237974643707275,
      "learning_rate": 3.009575923392613e-06,
      "loss": 0.3434,
      "step": 2127
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8849083781242371,
      "learning_rate": 2.9639762881896946e-06,
      "loss": 0.3442,
      "step": 2128
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8236368298530579,
      "learning_rate": 2.9183766529867763e-06,
      "loss": 0.3358,
      "step": 2129
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8945069909095764,
      "learning_rate": 2.8727770177838576e-06,
      "loss": 0.3566,
      "step": 2130
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.3535517156124115,
      "eval_runtime": 21.3121,
      "eval_samples_per_second": 1.314,
      "eval_steps_per_second": 0.328,
      "step": 2130
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8623038530349731,
      "learning_rate": 2.8271773825809394e-06,
      "loss": 0.3519,
      "step": 2131
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.796574056148529,
      "learning_rate": 2.781577747378021e-06,
      "loss": 0.3416,
      "step": 2132
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8572683930397034,
      "learning_rate": 2.735978112175103e-06,
      "loss": 0.3258,
      "step": 2133
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8047083616256714,
      "learning_rate": 2.690378476972184e-06,
      "loss": 0.3364,
      "step": 2134
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8221876621246338,
      "learning_rate": 2.644778841769266e-06,
      "loss": 0.342,
      "step": 2135
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8190683722496033,
      "learning_rate": 2.5991792065663476e-06,
      "loss": 0.3342,
      "step": 2136
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8451207280158997,
      "learning_rate": 2.5535795713634293e-06,
      "loss": 0.3588,
      "step": 2137
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.8056630492210388,
      "learning_rate": 2.5079799361605106e-06,
      "loss": 0.3397,
      "step": 2138
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8281639218330383,
      "learning_rate": 2.4623803009575924e-06,
      "loss": 0.347,
      "step": 2139
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8217196464538574,
      "learning_rate": 2.416780665754674e-06,
      "loss": 0.3565,
      "step": 2140
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.3535143733024597,
      "eval_runtime": 21.8883,
      "eval_samples_per_second": 1.279,
      "eval_steps_per_second": 0.32,
      "step": 2140
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8088732361793518,
      "learning_rate": 2.371181030551756e-06,
      "loss": 0.326,
      "step": 2141
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8405909538269043,
      "learning_rate": 2.325581395348837e-06,
      "loss": 0.3313,
      "step": 2142
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8254099488258362,
      "learning_rate": 2.279981760145919e-06,
      "loss": 0.3583,
      "step": 2143
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8337267637252808,
      "learning_rate": 2.2343821249430006e-06,
      "loss": 0.3323,
      "step": 2144
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8419988751411438,
      "learning_rate": 2.1887824897400824e-06,
      "loss": 0.3446,
      "step": 2145
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.846906840801239,
      "learning_rate": 2.1431828545371637e-06,
      "loss": 0.3452,
      "step": 2146
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8782608509063721,
      "learning_rate": 2.0975832193342454e-06,
      "loss": 0.3571,
      "step": 2147
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.7980951070785522,
      "learning_rate": 2.051983584131327e-06,
      "loss": 0.3397,
      "step": 2148
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8458917737007141,
      "learning_rate": 2.0063839489284084e-06,
      "loss": 0.3391,
      "step": 2149
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.839817225933075,
      "learning_rate": 1.96078431372549e-06,
      "loss": 0.3563,
      "step": 2150
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.3534523546695709,
      "eval_runtime": 23.8012,
      "eval_samples_per_second": 1.176,
      "eval_steps_per_second": 0.294,
      "step": 2150
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8007347583770752,
      "learning_rate": 1.915184678522572e-06,
      "loss": 0.3382,
      "step": 2151
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8078628778457642,
      "learning_rate": 1.8695850433196534e-06,
      "loss": 0.3225,
      "step": 2152
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.855683445930481,
      "learning_rate": 1.8239854081167352e-06,
      "loss": 0.3462,
      "step": 2153
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.796846866607666,
      "learning_rate": 1.7783857729138167e-06,
      "loss": 0.3505,
      "step": 2154
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8556407690048218,
      "learning_rate": 1.7327861377108984e-06,
      "loss": 0.344,
      "step": 2155
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8137206435203552,
      "learning_rate": 1.68718650250798e-06,
      "loss": 0.3218,
      "step": 2156
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8239472508430481,
      "learning_rate": 1.6415868673050617e-06,
      "loss": 0.3395,
      "step": 2157
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.7975683808326721,
      "learning_rate": 1.5959872321021432e-06,
      "loss": 0.3258,
      "step": 2158
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.8348971605300903,
      "learning_rate": 1.550387596899225e-06,
      "loss": 0.3425,
      "step": 2159
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.892481803894043,
      "learning_rate": 1.5047879616963064e-06,
      "loss": 0.3692,
      "step": 2160
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.3534722328186035,
      "eval_runtime": 21.8372,
      "eval_samples_per_second": 1.282,
      "eval_steps_per_second": 0.321,
      "step": 2160
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8270180821418762,
      "learning_rate": 1.4591883264933882e-06,
      "loss": 0.3434,
      "step": 2161
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8239341974258423,
      "learning_rate": 1.4135886912904697e-06,
      "loss": 0.3442,
      "step": 2162
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8776184916496277,
      "learning_rate": 1.3679890560875514e-06,
      "loss": 0.3642,
      "step": 2163
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8055011034011841,
      "learning_rate": 1.322389420884633e-06,
      "loss": 0.3332,
      "step": 2164
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8007059693336487,
      "learning_rate": 1.2767897856817147e-06,
      "loss": 0.3118,
      "step": 2165
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8044161796569824,
      "learning_rate": 1.2311901504787962e-06,
      "loss": 0.3235,
      "step": 2166
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.7754327654838562,
      "learning_rate": 1.185590515275878e-06,
      "loss": 0.3149,
      "step": 2167
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8618872761726379,
      "learning_rate": 1.1399908800729594e-06,
      "loss": 0.332,
      "step": 2168
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8462168574333191,
      "learning_rate": 1.0943912448700412e-06,
      "loss": 0.35,
      "step": 2169
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8031334280967712,
      "learning_rate": 1.0487916096671227e-06,
      "loss": 0.3182,
      "step": 2170
    },
    {
      "epoch": 0.99,
      "eval_loss": 0.35348114371299744,
      "eval_runtime": 21.5618,
      "eval_samples_per_second": 1.299,
      "eval_steps_per_second": 0.325,
      "step": 2170
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.7941566705703735,
      "learning_rate": 1.0031919744642042e-06,
      "loss": 0.3134,
      "step": 2171
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8218158483505249,
      "learning_rate": 9.57592339261286e-07,
      "loss": 0.3459,
      "step": 2172
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.846577525138855,
      "learning_rate": 9.119927040583676e-07,
      "loss": 0.3406,
      "step": 2173
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8338779807090759,
      "learning_rate": 8.663930688554492e-07,
      "loss": 0.3367,
      "step": 2174
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8083705306053162,
      "learning_rate": 8.207934336525308e-07,
      "loss": 0.314,
      "step": 2175
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8483944535255432,
      "learning_rate": 7.751937984496125e-07,
      "loss": 0.3422,
      "step": 2176
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8029471039772034,
      "learning_rate": 7.295941632466941e-07,
      "loss": 0.3337,
      "step": 2177
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8467877507209778,
      "learning_rate": 6.839945280437757e-07,
      "loss": 0.3429,
      "step": 2178
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.7840286493301392,
      "learning_rate": 6.383948928408573e-07,
      "loss": 0.3265,
      "step": 2179
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.7920138835906982,
      "learning_rate": 5.92795257637939e-07,
      "loss": 0.3359,
      "step": 2180
    },
    {
      "epoch": 0.99,
      "eval_loss": 0.3534707725048065,
      "eval_runtime": 21.7256,
      "eval_samples_per_second": 1.289,
      "eval_steps_per_second": 0.322,
      "step": 2180
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8483693599700928,
      "learning_rate": 5.471956224350206e-07,
      "loss": 0.3586,
      "step": 2181
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.8043666481971741,
      "learning_rate": 5.015959872321021e-07,
      "loss": 0.3351,
      "step": 2182
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.7986137866973877,
      "learning_rate": 4.559963520291838e-07,
      "loss": 0.3371,
      "step": 2183
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.8146182298660278,
      "learning_rate": 4.103967168262654e-07,
      "loss": 0.319,
      "step": 2184
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.8116611242294312,
      "learning_rate": 3.6479708162334704e-07,
      "loss": 0.3309,
      "step": 2185
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.834054172039032,
      "learning_rate": 3.1919744642042867e-07,
      "loss": 0.3504,
      "step": 2186
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.7927460074424744,
      "learning_rate": 2.735978112175103e-07,
      "loss": 0.3502,
      "step": 2187
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.8293411135673523,
      "learning_rate": 2.279981760145919e-07,
      "loss": 0.344,
      "step": 2188
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.8744671940803528,
      "learning_rate": 1.8239854081167352e-07,
      "loss": 0.3274,
      "step": 2189
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.9020984768867493,
      "learning_rate": 1.3679890560875515e-07,
      "loss": 0.3517,
      "step": 2190
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.35346221923828125,
      "eval_runtime": 21.6993,
      "eval_samples_per_second": 1.29,
      "eval_steps_per_second": 0.323,
      "step": 2190
    }
  ],
  "logging_steps": 1,
  "max_steps": 2193,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 10,
  "total_flos": 1.2253539900158116e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}