{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 1,
  "global_step": 160,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0625,
      "grad_norm": 9.625,
      "learning_rate": 2.5e-05,
      "loss": 1.0231,
      "step": 1
    },
    {
      "epoch": 0.0625,
      "eval_accuracy": 0.496,
      "eval_loss": 1.0858867168426514,
      "eval_runtime": 4.9172,
      "eval_samples_per_second": 50.841,
      "eval_steps_per_second": 1.627,
      "step": 1
    },
    {
      "epoch": 0.125,
      "grad_norm": 10.5625,
      "learning_rate": 5e-05,
      "loss": 0.9224,
      "step": 2
    },
    {
      "epoch": 0.125,
      "eval_accuracy": 0.488,
      "eval_loss": 1.0843095779418945,
      "eval_runtime": 4.9094,
      "eval_samples_per_second": 50.923,
      "eval_steps_per_second": 1.63,
      "step": 2
    },
    {
      "epoch": 0.1875,
      "grad_norm": 11.875,
      "learning_rate": 4.968354430379747e-05,
      "loss": 1.2584,
      "step": 3
    },
    {
      "epoch": 0.1875,
      "eval_accuracy": 0.488,
      "eval_loss": 1.0734808444976807,
      "eval_runtime": 4.9054,
      "eval_samples_per_second": 50.964,
      "eval_steps_per_second": 1.631,
      "step": 3
    },
    {
      "epoch": 0.25,
      "grad_norm": 13.3125,
      "learning_rate": 4.936708860759494e-05,
      "loss": 1.0442,
      "step": 4
    },
    {
      "epoch": 0.25,
      "eval_accuracy": 0.492,
      "eval_loss": 1.0604006052017212,
      "eval_runtime": 4.9069,
      "eval_samples_per_second": 50.948,
      "eval_steps_per_second": 1.63,
      "step": 4
    },
    {
      "epoch": 0.3125,
      "grad_norm": 22.625,
      "learning_rate": 4.905063291139241e-05,
      "loss": 1.3363,
      "step": 5
    },
    {
      "epoch": 0.3125,
      "eval_accuracy": 0.496,
      "eval_loss": 1.0435388088226318,
      "eval_runtime": 4.9047,
      "eval_samples_per_second": 50.972,
      "eval_steps_per_second": 1.631,
      "step": 5
    },
    {
      "epoch": 0.375,
      "grad_norm": 10.625,
      "learning_rate": 4.8734177215189874e-05,
      "loss": 1.0358,
      "step": 6
    },
    {
      "epoch": 0.375,
      "eval_accuracy": 0.504,
      "eval_loss": 1.0265413522720337,
      "eval_runtime": 4.9042,
      "eval_samples_per_second": 50.977,
      "eval_steps_per_second": 1.631,
      "step": 6
    },
    {
      "epoch": 0.4375,
      "grad_norm": 7.5625,
      "learning_rate": 4.8417721518987346e-05,
      "loss": 0.9276,
      "step": 7
    },
    {
      "epoch": 0.4375,
      "eval_accuracy": 0.508,
      "eval_loss": 1.0061290264129639,
      "eval_runtime": 4.9077,
      "eval_samples_per_second": 50.941,
      "eval_steps_per_second": 1.63,
      "step": 7
    },
    {
      "epoch": 0.5,
      "grad_norm": 16.75,
      "learning_rate": 4.810126582278481e-05,
      "loss": 0.9751,
      "step": 8
    },
    {
      "epoch": 0.5,
      "eval_accuracy": 0.512,
      "eval_loss": 0.9898163080215454,
      "eval_runtime": 4.9054,
      "eval_samples_per_second": 50.964,
      "eval_steps_per_second": 1.631,
      "step": 8
    },
    {
      "epoch": 0.5625,
      "grad_norm": 10.875,
      "learning_rate": 4.778481012658228e-05,
      "loss": 1.1136,
      "step": 9
    },
    {
      "epoch": 0.5625,
      "eval_accuracy": 0.524,
      "eval_loss": 0.9671441912651062,
      "eval_runtime": 4.8588,
      "eval_samples_per_second": 51.453,
      "eval_steps_per_second": 1.646,
      "step": 9
    },
    {
      "epoch": 0.625,
      "grad_norm": 11.0,
      "learning_rate": 4.7468354430379746e-05,
      "loss": 0.8243,
      "step": 10
    },
    {
      "epoch": 0.625,
      "eval_accuracy": 0.528,
      "eval_loss": 0.9475510120391846,
      "eval_runtime": 4.9061,
      "eval_samples_per_second": 50.957,
      "eval_steps_per_second": 1.631,
      "step": 10
    },
    {
      "epoch": 0.6875,
      "grad_norm": 10.875,
      "learning_rate": 4.715189873417722e-05,
      "loss": 0.9469,
      "step": 11
    },
    {
      "epoch": 0.6875,
      "eval_accuracy": 0.536,
      "eval_loss": 0.9229685068130493,
      "eval_runtime": 4.9037,
      "eval_samples_per_second": 50.982,
      "eval_steps_per_second": 1.631,
      "step": 11
    },
    {
      "epoch": 0.75,
      "grad_norm": 10.125,
      "learning_rate": 4.683544303797468e-05,
      "loss": 0.9348,
      "step": 12
    },
    {
      "epoch": 0.75,
      "eval_accuracy": 0.532,
      "eval_loss": 0.9007159471511841,
      "eval_runtime": 4.9072,
      "eval_samples_per_second": 50.945,
      "eval_steps_per_second": 1.63,
      "step": 12
    },
    {
      "epoch": 0.8125,
      "grad_norm": 9.6875,
      "learning_rate": 4.6518987341772154e-05,
      "loss": 0.939,
      "step": 13
    },
    {
      "epoch": 0.8125,
      "eval_accuracy": 0.548,
      "eval_loss": 0.8757832050323486,
      "eval_runtime": 4.9116,
      "eval_samples_per_second": 50.9,
      "eval_steps_per_second": 1.629,
      "step": 13
    },
    {
      "epoch": 0.875,
      "grad_norm": 16.375,
      "learning_rate": 4.6202531645569625e-05,
      "loss": 0.8007,
      "step": 14
    },
    {
      "epoch": 0.875,
      "eval_accuracy": 0.556,
      "eval_loss": 0.8450800776481628,
      "eval_runtime": 4.9085,
      "eval_samples_per_second": 50.932,
      "eval_steps_per_second": 1.63,
      "step": 14
    },
    {
      "epoch": 0.9375,
      "grad_norm": 13.625,
      "learning_rate": 4.588607594936709e-05,
      "loss": 0.8642,
      "step": 15
    },
    {
      "epoch": 0.9375,
      "eval_accuracy": 0.584,
      "eval_loss": 0.8132224082946777,
      "eval_runtime": 4.9071,
      "eval_samples_per_second": 50.946,
      "eval_steps_per_second": 1.63,
      "step": 15
    },
    {
      "epoch": 1.0,
      "grad_norm": 16.5,
      "learning_rate": 4.556962025316456e-05,
      "loss": 0.6672,
      "step": 16
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.604,
      "eval_loss": 0.7814155220985413,
      "eval_runtime": 4.8576,
      "eval_samples_per_second": 51.465,
      "eval_steps_per_second": 1.647,
      "step": 16
    },
    {
      "epoch": 1.0625,
      "grad_norm": 9.1875,
      "learning_rate": 4.525316455696203e-05,
      "loss": 0.7072,
      "step": 17
    },
    {
      "epoch": 1.0625,
      "eval_accuracy": 0.612,
      "eval_loss": 0.7591574788093567,
      "eval_runtime": 4.9059,
      "eval_samples_per_second": 50.959,
      "eval_steps_per_second": 1.631,
      "step": 17
    },
    {
      "epoch": 1.125,
      "grad_norm": 8.3125,
      "learning_rate": 4.49367088607595e-05,
      "loss": 0.7815,
      "step": 18
    },
    {
      "epoch": 1.125,
      "eval_accuracy": 0.644,
      "eval_loss": 0.7339711785316467,
      "eval_runtime": 4.9112,
      "eval_samples_per_second": 50.904,
      "eval_steps_per_second": 1.629,
      "step": 18
    },
    {
      "epoch": 1.1875,
      "grad_norm": 7.0,
      "learning_rate": 4.462025316455696e-05,
      "loss": 0.6226,
      "step": 19
    },
    {
      "epoch": 1.1875,
      "eval_accuracy": 0.66,
      "eval_loss": 0.7131918668746948,
      "eval_runtime": 4.9082,
      "eval_samples_per_second": 50.935,
      "eval_steps_per_second": 1.63,
      "step": 19
    },
    {
      "epoch": 1.25,
      "grad_norm": 11.25,
      "learning_rate": 4.430379746835443e-05,
      "loss": 0.6456,
      "step": 20
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.664,
      "eval_loss": 0.6907175183296204,
      "eval_runtime": 4.9075,
      "eval_samples_per_second": 50.943,
      "eval_steps_per_second": 1.63,
      "step": 20
    },
    {
      "epoch": 1.3125,
      "grad_norm": 14.0625,
      "learning_rate": 4.3987341772151904e-05,
      "loss": 0.8005,
      "step": 21
    },
    {
      "epoch": 1.3125,
      "eval_accuracy": 0.668,
      "eval_loss": 0.6769924163818359,
      "eval_runtime": 4.9073,
      "eval_samples_per_second": 50.945,
      "eval_steps_per_second": 1.63,
      "step": 21
    },
    {
      "epoch": 1.375,
      "grad_norm": 6.09375,
      "learning_rate": 4.367088607594937e-05,
      "loss": 0.4491,
      "step": 22
    },
    {
      "epoch": 1.375,
      "eval_accuracy": 0.688,
      "eval_loss": 0.6586074233055115,
      "eval_runtime": 4.906,
      "eval_samples_per_second": 50.958,
      "eval_steps_per_second": 1.631,
      "step": 22
    },
    {
      "epoch": 1.4375,
      "grad_norm": 6.0625,
      "learning_rate": 4.3354430379746834e-05,
      "loss": 0.474,
      "step": 23
    },
    {
      "epoch": 1.4375,
      "eval_accuracy": 0.692,
      "eval_loss": 0.6458801031112671,
      "eval_runtime": 4.8557,
      "eval_samples_per_second": 51.486,
      "eval_steps_per_second": 1.648,
      "step": 23
    },
    {
      "epoch": 1.5,
      "grad_norm": 7.3125,
      "learning_rate": 4.3037974683544305e-05,
      "loss": 0.5048,
      "step": 24
    },
    {
      "epoch": 1.5,
      "eval_accuracy": 0.704,
      "eval_loss": 0.6440668702125549,
      "eval_runtime": 4.9064,
      "eval_samples_per_second": 50.954,
      "eval_steps_per_second": 1.631,
      "step": 24
    },
    {
      "epoch": 1.5625,
      "grad_norm": 9.75,
      "learning_rate": 4.2721518987341776e-05,
      "loss": 0.443,
      "step": 25
    },
    {
      "epoch": 1.5625,
      "eval_accuracy": 0.712,
      "eval_loss": 0.639043927192688,
      "eval_runtime": 4.9058,
      "eval_samples_per_second": 50.96,
      "eval_steps_per_second": 1.631,
      "step": 25
    },
    {
      "epoch": 1.625,
      "grad_norm": 6.40625,
      "learning_rate": 4.240506329113924e-05,
      "loss": 0.4949,
      "step": 26
    },
    {
      "epoch": 1.625,
      "eval_accuracy": 0.716,
      "eval_loss": 0.6155992150306702,
      "eval_runtime": 4.9046,
      "eval_samples_per_second": 50.972,
      "eval_steps_per_second": 1.631,
      "step": 26
    },
    {
      "epoch": 1.6875,
      "grad_norm": 16.0,
      "learning_rate": 4.208860759493671e-05,
      "loss": 0.5333,
      "step": 27
    },
    {
      "epoch": 1.6875,
      "eval_accuracy": 0.708,
      "eval_loss": 0.6270785927772522,
      "eval_runtime": 4.9087,
      "eval_samples_per_second": 50.93,
      "eval_steps_per_second": 1.63,
      "step": 27
    },
    {
      "epoch": 1.75,
      "grad_norm": 8.25,
      "learning_rate": 4.177215189873418e-05,
      "loss": 0.5316,
      "step": 28
    },
    {
      "epoch": 1.75,
      "eval_accuracy": 0.696,
      "eval_loss": 0.6393130421638489,
      "eval_runtime": 4.9075,
      "eval_samples_per_second": 50.943,
      "eval_steps_per_second": 1.63,
      "step": 28
    },
    {
      "epoch": 1.8125,
      "grad_norm": 26.875,
      "learning_rate": 4.145569620253165e-05,
      "loss": 0.5872,
      "step": 29
    },
    {
      "epoch": 1.8125,
      "eval_accuracy": 0.708,
      "eval_loss": 0.6233423948287964,
      "eval_runtime": 4.9097,
      "eval_samples_per_second": 50.919,
      "eval_steps_per_second": 1.629,
      "step": 29
    },
    {
      "epoch": 1.875,
      "grad_norm": 14.3125,
      "learning_rate": 4.113924050632912e-05,
      "loss": 0.532,
      "step": 30
    },
    {
      "epoch": 1.875,
      "eval_accuracy": 0.728,
      "eval_loss": 0.5802625417709351,
      "eval_runtime": 4.9064,
      "eval_samples_per_second": 50.954,
      "eval_steps_per_second": 1.631,
      "step": 30
    },
    {
      "epoch": 1.9375,
      "grad_norm": 5.8125,
      "learning_rate": 4.0822784810126584e-05,
      "loss": 0.5531,
      "step": 31
    },
    {
      "epoch": 1.9375,
      "eval_accuracy": 0.748,
      "eval_loss": 0.5553041100502014,
      "eval_runtime": 4.9066,
      "eval_samples_per_second": 50.952,
      "eval_steps_per_second": 1.63,
      "step": 31
    },
    {
      "epoch": 2.0,
      "grad_norm": 19.375,
      "learning_rate": 4.050632911392405e-05,
      "loss": 0.4206,
      "step": 32
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.74,
      "eval_loss": 0.5429959893226624,
      "eval_runtime": 4.9071,
      "eval_samples_per_second": 50.947,
      "eval_steps_per_second": 1.63,
      "step": 32
    },
    {
      "epoch": 2.0625,
      "grad_norm": 11.4375,
      "learning_rate": 4.018987341772152e-05,
      "loss": 0.4597,
      "step": 33
    },
    {
      "epoch": 2.0625,
      "eval_accuracy": 0.72,
      "eval_loss": 0.5543990731239319,
      "eval_runtime": 4.9053,
      "eval_samples_per_second": 50.965,
      "eval_steps_per_second": 1.631,
      "step": 33
    },
    {
      "epoch": 2.125,
      "grad_norm": 12.0,
      "learning_rate": 3.987341772151899e-05,
      "loss": 0.4009,
      "step": 34
    },
    {
      "epoch": 2.125,
      "eval_accuracy": 0.716,
      "eval_loss": 0.5574648976325989,
      "eval_runtime": 4.9099,
      "eval_samples_per_second": 50.917,
      "eval_steps_per_second": 1.629,
      "step": 34
    },
    {
      "epoch": 2.1875,
      "grad_norm": 13.0625,
      "learning_rate": 3.9556962025316456e-05,
      "loss": 0.5113,
      "step": 35
    },
    {
      "epoch": 2.1875,
      "eval_accuracy": 0.732,
      "eval_loss": 0.5373346209526062,
      "eval_runtime": 4.9081,
      "eval_samples_per_second": 50.936,
      "eval_steps_per_second": 1.63,
      "step": 35
    },
    {
      "epoch": 2.25,
      "grad_norm": 18.25,
      "learning_rate": 3.924050632911392e-05,
      "loss": 0.3224,
      "step": 36
    },
    {
      "epoch": 2.25,
      "eval_accuracy": 0.744,
      "eval_loss": 0.5083271265029907,
      "eval_runtime": 4.9092,
      "eval_samples_per_second": 50.925,
      "eval_steps_per_second": 1.63,
      "step": 36
    },
    {
      "epoch": 2.3125,
      "grad_norm": 13.8125,
      "learning_rate": 3.89240506329114e-05,
      "loss": 0.2552,
      "step": 37
    },
    {
      "epoch": 2.3125,
      "eval_accuracy": 0.756,
      "eval_loss": 0.48166918754577637,
      "eval_runtime": 4.9082,
      "eval_samples_per_second": 50.935,
      "eval_steps_per_second": 1.63,
      "step": 37
    },
    {
      "epoch": 2.375,
      "grad_norm": 13.0625,
      "learning_rate": 3.8607594936708864e-05,
      "loss": 0.3693,
      "step": 38
    },
    {
      "epoch": 2.375,
      "eval_accuracy": 0.76,
      "eval_loss": 0.4727539122104645,
      "eval_runtime": 4.8582,
      "eval_samples_per_second": 51.459,
      "eval_steps_per_second": 1.647,
      "step": 38
    },
    {
      "epoch": 2.4375,
      "grad_norm": 5.71875,
      "learning_rate": 3.829113924050633e-05,
      "loss": 0.3886,
      "step": 39
    },
    {
      "epoch": 2.4375,
      "eval_accuracy": 0.76,
      "eval_loss": 0.4700586497783661,
      "eval_runtime": 4.9039,
      "eval_samples_per_second": 50.979,
      "eval_steps_per_second": 1.631,
      "step": 39
    },
    {
      "epoch": 2.5,
      "grad_norm": 12.6875,
      "learning_rate": 3.79746835443038e-05,
      "loss": 0.3666,
      "step": 40
    },
    {
      "epoch": 2.5,
      "eval_accuracy": 0.772,
      "eval_loss": 0.46244245767593384,
      "eval_runtime": 4.9122,
      "eval_samples_per_second": 50.894,
      "eval_steps_per_second": 1.629,
      "step": 40
    },
    {
      "epoch": 2.5625,
      "grad_norm": 5.21875,
      "learning_rate": 3.765822784810127e-05,
      "loss": 0.3614,
      "step": 41
    },
    {
      "epoch": 2.5625,
      "eval_accuracy": 0.772,
      "eval_loss": 0.45654377341270447,
      "eval_runtime": 4.8573,
      "eval_samples_per_second": 51.469,
      "eval_steps_per_second": 1.647,
      "step": 41
    },
    {
      "epoch": 2.625,
      "grad_norm": 3.375,
      "learning_rate": 3.7341772151898736e-05,
      "loss": 0.2228,
      "step": 42
    },
    {
      "epoch": 2.625,
      "eval_accuracy": 0.768,
      "eval_loss": 0.45055416226387024,
      "eval_runtime": 4.9079,
      "eval_samples_per_second": 50.938,
      "eval_steps_per_second": 1.63,
      "step": 42
    },
    {
      "epoch": 2.6875,
      "grad_norm": 7.34375,
      "learning_rate": 3.70253164556962e-05,
      "loss": 0.2365,
      "step": 43
    },
    {
      "epoch": 2.6875,
      "eval_accuracy": 0.772,
      "eval_loss": 0.4469849467277527,
      "eval_runtime": 4.9083,
      "eval_samples_per_second": 50.935,
      "eval_steps_per_second": 1.63,
      "step": 43
    },
    {
      "epoch": 2.75,
      "grad_norm": 7.5625,
      "learning_rate": 3.670886075949367e-05,
      "loss": 0.3274,
      "step": 44
    },
    {
      "epoch": 2.75,
      "eval_accuracy": 0.784,
      "eval_loss": 0.44335371255874634,
      "eval_runtime": 4.9063,
      "eval_samples_per_second": 50.955,
      "eval_steps_per_second": 1.631,
      "step": 44
    },
    {
      "epoch": 2.8125,
      "grad_norm": 11.125,
      "learning_rate": 3.639240506329114e-05,
      "loss": 0.2592,
      "step": 45
    },
    {
      "epoch": 2.8125,
      "eval_accuracy": 0.788,
      "eval_loss": 0.44204553961753845,
      "eval_runtime": 4.9034,
      "eval_samples_per_second": 50.985,
      "eval_steps_per_second": 1.632,
      "step": 45
    },
    {
      "epoch": 2.875,
      "grad_norm": 3.75,
      "learning_rate": 3.607594936708861e-05,
      "loss": 0.2343,
      "step": 46
    },
    {
      "epoch": 2.875,
      "eval_accuracy": 0.78,
      "eval_loss": 0.4448190927505493,
      "eval_runtime": 4.9077,
      "eval_samples_per_second": 50.94,
      "eval_steps_per_second": 1.63,
      "step": 46
    },
    {
      "epoch": 2.9375,
      "grad_norm": 9.375,
      "learning_rate": 3.575949367088608e-05,
      "loss": 0.219,
      "step": 47
    },
    {
      "epoch": 2.9375,
      "eval_accuracy": 0.788,
      "eval_loss": 0.44333717226982117,
      "eval_runtime": 4.9081,
      "eval_samples_per_second": 50.936,
      "eval_steps_per_second": 1.63,
      "step": 47
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.25,
      "learning_rate": 3.5443037974683544e-05,
      "loss": 0.2329,
      "step": 48
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.796,
      "eval_loss": 0.44067704677581787,
      "eval_runtime": 4.9046,
      "eval_samples_per_second": 50.973,
      "eval_steps_per_second": 1.631,
      "step": 48
    },
    {
      "epoch": 3.0625,
      "grad_norm": 6.28125,
      "learning_rate": 3.5126582278481015e-05,
      "loss": 0.2697,
      "step": 49
    },
    {
      "epoch": 3.0625,
      "eval_accuracy": 0.796,
      "eval_loss": 0.4345742166042328,
      "eval_runtime": 4.905,
      "eval_samples_per_second": 50.969,
      "eval_steps_per_second": 1.631,
      "step": 49
    },
    {
      "epoch": 3.125,
      "grad_norm": 5.71875,
      "learning_rate": 3.4810126582278487e-05,
      "loss": 0.1796,
      "step": 50
    },
    {
      "epoch": 3.125,
      "eval_accuracy": 0.796,
      "eval_loss": 0.42894840240478516,
      "eval_runtime": 4.9069,
      "eval_samples_per_second": 50.949,
      "eval_steps_per_second": 1.63,
      "step": 50
    },
    {
      "epoch": 3.1875,
      "grad_norm": 9.125,
      "learning_rate": 3.449367088607595e-05,
      "loss": 0.1754,
      "step": 51
    },
    {
      "epoch": 3.1875,
      "eval_accuracy": 0.792,
      "eval_loss": 0.43848368525505066,
      "eval_runtime": 4.9038,
      "eval_samples_per_second": 50.981,
      "eval_steps_per_second": 1.631,
      "step": 51
    },
    {
      "epoch": 3.25,
      "grad_norm": 3.28125,
      "learning_rate": 3.4177215189873416e-05,
      "loss": 0.1621,
      "step": 52
    },
    {
      "epoch": 3.25,
      "eval_accuracy": 0.78,
      "eval_loss": 0.4773699939250946,
      "eval_runtime": 4.9079,
      "eval_samples_per_second": 50.938,
      "eval_steps_per_second": 1.63,
      "step": 52
    },
    {
      "epoch": 3.3125,
      "grad_norm": 12.0625,
      "learning_rate": 3.386075949367089e-05,
      "loss": 0.2976,
      "step": 53
    },
    {
      "epoch": 3.3125,
      "eval_accuracy": 0.776,
      "eval_loss": 0.5157197117805481,
      "eval_runtime": 4.9079,
      "eval_samples_per_second": 50.939,
      "eval_steps_per_second": 1.63,
      "step": 53
    },
    {
      "epoch": 3.375,
      "grad_norm": 9.0625,
      "learning_rate": 3.354430379746836e-05,
      "loss": 0.2612,
      "step": 54
    },
    {
      "epoch": 3.375,
      "eval_accuracy": 0.772,
      "eval_loss": 0.5219131112098694,
      "eval_runtime": 4.9158,
      "eval_samples_per_second": 50.856,
      "eval_steps_per_second": 1.627,
      "step": 54
    },
    {
      "epoch": 3.4375,
      "grad_norm": 15.6875,
      "learning_rate": 3.322784810126582e-05,
      "loss": 0.2856,
      "step": 55
    },
    {
      "epoch": 3.4375,
      "eval_accuracy": 0.772,
      "eval_loss": 0.5048539638519287,
      "eval_runtime": 4.9093,
      "eval_samples_per_second": 50.924,
      "eval_steps_per_second": 1.63,
      "step": 55
    },
    {
      "epoch": 3.5,
      "grad_norm": 18.75,
      "learning_rate": 3.291139240506329e-05,
      "loss": 0.2473,
      "step": 56
    },
    {
      "epoch": 3.5,
      "eval_accuracy": 0.784,
      "eval_loss": 0.4671543538570404,
      "eval_runtime": 5.0157,
      "eval_samples_per_second": 49.844,
      "eval_steps_per_second": 1.595,
      "step": 56
    },
    {
      "epoch": 3.5625,
      "grad_norm": 5.90625,
      "learning_rate": 3.2594936708860766e-05,
      "loss": 0.1565,
      "step": 57
    },
    {
      "epoch": 3.5625,
      "eval_accuracy": 0.808,
      "eval_loss": 0.4280901551246643,
      "eval_runtime": 5.2697,
      "eval_samples_per_second": 47.441,
      "eval_steps_per_second": 1.518,
      "step": 57
    },
    {
      "epoch": 3.625,
      "grad_norm": 4.9375,
      "learning_rate": 3.227848101265823e-05,
      "loss": 0.1862,
      "step": 58
    },
    {
      "epoch": 3.625,
      "eval_accuracy": 0.792,
      "eval_loss": 0.417385071516037,
      "eval_runtime": 4.936,
      "eval_samples_per_second": 50.648,
      "eval_steps_per_second": 1.621,
      "step": 58
    },
    {
      "epoch": 3.6875,
      "grad_norm": 4.71875,
      "learning_rate": 3.1962025316455695e-05,
      "loss": 0.1997,
      "step": 59
    },
    {
      "epoch": 3.6875,
      "eval_accuracy": 0.812,
      "eval_loss": 0.43094775080680847,
      "eval_runtime": 5.3564,
      "eval_samples_per_second": 46.673,
      "eval_steps_per_second": 1.494,
      "step": 59
    },
    {
      "epoch": 3.75,
      "grad_norm": 4.21875,
      "learning_rate": 3.1645569620253167e-05,
      "loss": 0.2196,
      "step": 60
    },
    {
      "epoch": 3.75,
      "eval_accuracy": 0.816,
      "eval_loss": 0.45009762048721313,
      "eval_runtime": 5.0887,
      "eval_samples_per_second": 49.129,
      "eval_steps_per_second": 1.572,
      "step": 60
    },
    {
      "epoch": 3.8125,
      "grad_norm": 7.4375,
      "learning_rate": 3.132911392405064e-05,
      "loss": 0.1863,
      "step": 61
    },
    {
      "epoch": 3.8125,
      "eval_accuracy": 0.816,
      "eval_loss": 0.4615320563316345,
      "eval_runtime": 4.9972,
      "eval_samples_per_second": 50.028,
      "eval_steps_per_second": 1.601,
      "step": 61
    },
    {
      "epoch": 3.875,
      "grad_norm": 5.84375,
      "learning_rate": 3.10126582278481e-05,
      "loss": 0.1621,
      "step": 62
    },
    {
      "epoch": 3.875,
      "eval_accuracy": 0.816,
      "eval_loss": 0.4588499367237091,
      "eval_runtime": 4.9486,
      "eval_samples_per_second": 50.519,
      "eval_steps_per_second": 1.617,
      "step": 62
    },
    {
      "epoch": 3.9375,
      "grad_norm": 14.9375,
      "learning_rate": 3.0696202531645574e-05,
      "loss": 0.2533,
      "step": 63
    },
    {
      "epoch": 3.9375,
      "eval_accuracy": 0.82,
      "eval_loss": 0.44347089529037476,
      "eval_runtime": 4.9235,
      "eval_samples_per_second": 50.777,
      "eval_steps_per_second": 1.625,
      "step": 63
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.15625,
      "learning_rate": 3.0379746835443042e-05,
      "loss": 0.1085,
      "step": 64
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.832,
      "eval_loss": 0.4236921966075897,
      "eval_runtime": 5.0108,
      "eval_samples_per_second": 49.892,
      "eval_steps_per_second": 1.597,
      "step": 64
    },
    {
      "epoch": 4.0625,
      "grad_norm": 10.0625,
      "learning_rate": 3.0063291139240506e-05,
      "loss": 0.1236,
      "step": 65
    },
    {
      "epoch": 4.0625,
      "eval_accuracy": 0.824,
      "eval_loss": 0.40546613931655884,
      "eval_runtime": 4.9996,
      "eval_samples_per_second": 50.004,
      "eval_steps_per_second": 1.6,
      "step": 65
    },
    {
      "epoch": 4.125,
      "grad_norm": 2.421875,
      "learning_rate": 2.9746835443037974e-05,
      "loss": 0.0979,
      "step": 66
    },
    {
      "epoch": 4.125,
      "eval_accuracy": 0.816,
      "eval_loss": 0.3985447883605957,
      "eval_runtime": 4.969,
      "eval_samples_per_second": 50.312,
      "eval_steps_per_second": 1.61,
      "step": 66
    },
    {
      "epoch": 4.1875,
      "grad_norm": 2.703125,
      "learning_rate": 2.9430379746835446e-05,
      "loss": 0.096,
      "step": 67
    },
    {
      "epoch": 4.1875,
      "eval_accuracy": 0.812,
      "eval_loss": 0.40320640802383423,
      "eval_runtime": 5.0458,
      "eval_samples_per_second": 49.546,
      "eval_steps_per_second": 1.585,
      "step": 67
    },
    {
      "epoch": 4.25,
      "grad_norm": 4.1875,
      "learning_rate": 2.9113924050632914e-05,
      "loss": 0.0776,
      "step": 68
    },
    {
      "epoch": 4.25,
      "eval_accuracy": 0.816,
      "eval_loss": 0.40163105726242065,
      "eval_runtime": 4.9683,
      "eval_samples_per_second": 50.319,
      "eval_steps_per_second": 1.61,
      "step": 68
    },
    {
      "epoch": 4.3125,
      "grad_norm": 3.734375,
      "learning_rate": 2.879746835443038e-05,
      "loss": 0.1406,
      "step": 69
    },
    {
      "epoch": 4.3125,
      "eval_accuracy": 0.812,
      "eval_loss": 0.40083351731300354,
      "eval_runtime": 5.0742,
      "eval_samples_per_second": 49.269,
      "eval_steps_per_second": 1.577,
      "step": 69
    },
    {
      "epoch": 4.375,
      "grad_norm": 3.0,
      "learning_rate": 2.848101265822785e-05,
      "loss": 0.1276,
      "step": 70
    },
    {
      "epoch": 4.375,
      "eval_accuracy": 0.816,
      "eval_loss": 0.4019404351711273,
      "eval_runtime": 5.0319,
      "eval_samples_per_second": 49.683,
      "eval_steps_per_second": 1.59,
      "step": 70
    },
    {
      "epoch": 4.4375,
      "grad_norm": 3.40625,
      "learning_rate": 2.8164556962025318e-05,
      "loss": 0.1169,
      "step": 71
    },
    {
      "epoch": 4.4375,
      "eval_accuracy": 0.82,
      "eval_loss": 0.40479913353919983,
      "eval_runtime": 5.0124,
      "eval_samples_per_second": 49.876,
      "eval_steps_per_second": 1.596,
      "step": 71
    },
    {
      "epoch": 4.5,
      "grad_norm": 10.875,
      "learning_rate": 2.7848101265822786e-05,
      "loss": 0.2681,
      "step": 72
    },
    {
      "epoch": 4.5,
      "eval_accuracy": 0.824,
      "eval_loss": 0.40435412526130676,
      "eval_runtime": 5.1364,
      "eval_samples_per_second": 48.673,
      "eval_steps_per_second": 1.558,
      "step": 72
    },
    {
      "epoch": 4.5625,
      "grad_norm": 5.59375,
      "learning_rate": 2.7531645569620257e-05,
      "loss": 0.1596,
      "step": 73
    },
    {
      "epoch": 4.5625,
      "eval_accuracy": 0.82,
      "eval_loss": 0.4038302004337311,
      "eval_runtime": 5.0213,
      "eval_samples_per_second": 49.788,
      "eval_steps_per_second": 1.593,
      "step": 73
    },
    {
      "epoch": 4.625,
      "grad_norm": 3.84375,
      "learning_rate": 2.7215189873417722e-05,
      "loss": 0.1036,
      "step": 74
    },
    {
      "epoch": 4.625,
      "eval_accuracy": 0.82,
      "eval_loss": 0.40612995624542236,
      "eval_runtime": 5.014,
      "eval_samples_per_second": 49.861,
      "eval_steps_per_second": 1.596,
      "step": 74
    },
    {
      "epoch": 4.6875,
      "grad_norm": 4.6875,
      "learning_rate": 2.689873417721519e-05,
      "loss": 0.1256,
      "step": 75
    },
    {
      "epoch": 4.6875,
      "eval_accuracy": 0.824,
      "eval_loss": 0.4083655774593353,
      "eval_runtime": 5.0762,
      "eval_samples_per_second": 49.249,
      "eval_steps_per_second": 1.576,
      "step": 75
    },
    {
      "epoch": 4.75,
      "grad_norm": 2.046875,
      "learning_rate": 2.6582278481012658e-05,
      "loss": 0.049,
      "step": 76
    },
    {
      "epoch": 4.75,
      "eval_accuracy": 0.828,
      "eval_loss": 0.4126991629600525,
      "eval_runtime": 5.1818,
      "eval_samples_per_second": 48.246,
      "eval_steps_per_second": 1.544,
      "step": 76
    },
    {
      "epoch": 4.8125,
      "grad_norm": 2.953125,
      "learning_rate": 2.626582278481013e-05,
      "loss": 0.1125,
      "step": 77
    },
    {
      "epoch": 4.8125,
      "eval_accuracy": 0.828,
      "eval_loss": 0.41637206077575684,
      "eval_runtime": 5.0132,
      "eval_samples_per_second": 49.869,
      "eval_steps_per_second": 1.596,
      "step": 77
    },
    {
      "epoch": 4.875,
      "grad_norm": 2.21875,
      "learning_rate": 2.5949367088607597e-05,
      "loss": 0.081,
      "step": 78
    },
    {
      "epoch": 4.875,
      "eval_accuracy": 0.832,
      "eval_loss": 0.41693058609962463,
      "eval_runtime": 5.1308,
      "eval_samples_per_second": 48.725,
      "eval_steps_per_second": 1.559,
      "step": 78
    },
    {
      "epoch": 4.9375,
      "grad_norm": 2.953125,
      "learning_rate": 2.5632911392405062e-05,
      "loss": 0.1308,
      "step": 79
    },
    {
      "epoch": 4.9375,
      "eval_accuracy": 0.82,
      "eval_loss": 0.4119318425655365,
      "eval_runtime": 5.3506,
      "eval_samples_per_second": 46.723,
      "eval_steps_per_second": 1.495,
      "step": 79
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.078125,
      "learning_rate": 2.5316455696202533e-05,
      "loss": 0.0556,
      "step": 80
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.82,
      "eval_loss": 0.4106213450431824,
      "eval_runtime": 5.0414,
      "eval_samples_per_second": 49.589,
      "eval_steps_per_second": 1.587,
      "step": 80
    },
    {
      "epoch": 5.0625,
      "grad_norm": 5.4375,
      "learning_rate": 2.5e-05,
      "loss": 0.1657,
      "step": 81
    },
    {
      "epoch": 5.0625,
      "eval_accuracy": 0.816,
      "eval_loss": 0.40592437982559204,
      "eval_runtime": 4.991,
      "eval_samples_per_second": 50.09,
      "eval_steps_per_second": 1.603,
      "step": 81
    },
    {
      "epoch": 5.125,
      "grad_norm": 3.125,
      "learning_rate": 2.468354430379747e-05,
      "loss": 0.1037,
      "step": 82
    },
    {
      "epoch": 5.125,
      "eval_accuracy": 0.816,
      "eval_loss": 0.40300390124320984,
      "eval_runtime": 4.9485,
      "eval_samples_per_second": 50.521,
      "eval_steps_per_second": 1.617,
      "step": 82
    },
    {
      "epoch": 5.1875,
      "grad_norm": 2.78125,
      "learning_rate": 2.4367088607594937e-05,
      "loss": 0.0676,
      "step": 83
    },
    {
      "epoch": 5.1875,
      "eval_accuracy": 0.824,
      "eval_loss": 0.39928579330444336,
      "eval_runtime": 5.1773,
      "eval_samples_per_second": 48.287,
      "eval_steps_per_second": 1.545,
      "step": 83
    },
    {
      "epoch": 5.25,
      "grad_norm": 3.625,
      "learning_rate": 2.4050632911392405e-05,
      "loss": 0.0513,
      "step": 84
    },
    {
      "epoch": 5.25,
      "eval_accuracy": 0.812,
      "eval_loss": 0.3965282738208771,
      "eval_runtime": 4.9096,
      "eval_samples_per_second": 50.921,
      "eval_steps_per_second": 1.629,
      "step": 84
    },
    {
      "epoch": 5.3125,
      "grad_norm": 2.84375,
      "learning_rate": 2.3734177215189873e-05,
      "loss": 0.0539,
      "step": 85
    },
    {
      "epoch": 5.3125,
      "eval_accuracy": 0.816,
      "eval_loss": 0.3946448862552643,
      "eval_runtime": 4.9127,
      "eval_samples_per_second": 50.889,
      "eval_steps_per_second": 1.628,
      "step": 85
    },
    {
      "epoch": 5.375,
      "grad_norm": 2.40625,
      "learning_rate": 2.341772151898734e-05,
      "loss": 0.0757,
      "step": 86
    },
    {
      "epoch": 5.375,
      "eval_accuracy": 0.816,
      "eval_loss": 0.3956676721572876,
      "eval_runtime": 4.9093,
      "eval_samples_per_second": 50.924,
      "eval_steps_per_second": 1.63,
      "step": 86
    },
    {
      "epoch": 5.4375,
      "grad_norm": 1.4296875,
      "learning_rate": 2.3101265822784813e-05,
      "loss": 0.0394,
      "step": 87
    },
    {
      "epoch": 5.4375,
      "eval_accuracy": 0.812,
      "eval_loss": 0.4012235105037689,
      "eval_runtime": 4.9087,
      "eval_samples_per_second": 50.93,
      "eval_steps_per_second": 1.63,
      "step": 87
    },
    {
      "epoch": 5.5,
      "grad_norm": 1.2578125,
      "learning_rate": 2.278481012658228e-05,
      "loss": 0.034,
      "step": 88
    },
    {
      "epoch": 5.5,
      "eval_accuracy": 0.804,
      "eval_loss": 0.40770140290260315,
      "eval_runtime": 4.9931,
      "eval_samples_per_second": 50.069,
      "eval_steps_per_second": 1.602,
      "step": 88
    },
    {
      "epoch": 5.5625,
      "grad_norm": 3.9375,
      "learning_rate": 2.246835443037975e-05,
      "loss": 0.0601,
      "step": 89
    },
    {
      "epoch": 5.5625,
      "eval_accuracy": 0.804,
      "eval_loss": 0.4079342782497406,
      "eval_runtime": 5.4428,
      "eval_samples_per_second": 45.932,
      "eval_steps_per_second": 1.47,
      "step": 89
    },
    {
      "epoch": 5.625,
      "grad_norm": 2.609375,
      "learning_rate": 2.2151898734177217e-05,
      "loss": 0.0402,
      "step": 90
    },
    {
      "epoch": 5.625,
      "eval_accuracy": 0.812,
      "eval_loss": 0.4079940915107727,
      "eval_runtime": 5.2545,
      "eval_samples_per_second": 47.578,
      "eval_steps_per_second": 1.523,
      "step": 90
    },
    {
      "epoch": 5.6875,
      "grad_norm": 2.171875,
      "learning_rate": 2.1835443037974685e-05,
      "loss": 0.0443,
      "step": 91
    },
    {
      "epoch": 5.6875,
      "eval_accuracy": 0.824,
      "eval_loss": 0.4047956168651581,
      "eval_runtime": 5.1747,
      "eval_samples_per_second": 48.312,
      "eval_steps_per_second": 1.546,
      "step": 91
    },
    {
      "epoch": 5.75,
      "grad_norm": 1.0,
      "learning_rate": 2.1518987341772153e-05,
      "loss": 0.0213,
      "step": 92
    },
    {
      "epoch": 5.75,
      "eval_accuracy": 0.824,
      "eval_loss": 0.4049394130706787,
      "eval_runtime": 5.3443,
      "eval_samples_per_second": 46.779,
      "eval_steps_per_second": 1.497,
      "step": 92
    },
    {
      "epoch": 5.8125,
      "grad_norm": 3.09375,
      "learning_rate": 2.120253164556962e-05,
      "loss": 0.05,
      "step": 93
    },
    {
      "epoch": 5.8125,
      "eval_accuracy": 0.824,
      "eval_loss": 0.40513497591018677,
      "eval_runtime": 4.9931,
      "eval_samples_per_second": 50.069,
      "eval_steps_per_second": 1.602,
      "step": 93
    },
    {
      "epoch": 5.875,
      "grad_norm": 1.171875,
      "learning_rate": 2.088607594936709e-05,
      "loss": 0.0321,
      "step": 94
    },
    {
      "epoch": 5.875,
      "eval_accuracy": 0.832,
      "eval_loss": 0.4072630703449249,
      "eval_runtime": 4.9859,
      "eval_samples_per_second": 50.141,
      "eval_steps_per_second": 1.605,
      "step": 94
    },
    {
      "epoch": 5.9375,
      "grad_norm": 1.453125,
      "learning_rate": 2.056962025316456e-05,
      "loss": 0.0315,
      "step": 95
    },
    {
      "epoch": 5.9375,
      "eval_accuracy": 0.832,
      "eval_loss": 0.40790075063705444,
      "eval_runtime": 4.9401,
      "eval_samples_per_second": 50.606,
      "eval_steps_per_second": 1.619,
      "step": 95
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.265625,
      "learning_rate": 2.0253164556962025e-05,
      "loss": 0.0207,
      "step": 96
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.832,
      "eval_loss": 0.4082264304161072,
      "eval_runtime": 4.9371,
      "eval_samples_per_second": 50.637,
      "eval_steps_per_second": 1.62,
      "step": 96
    },
    {
      "epoch": 6.0625,
      "grad_norm": 1.828125,
      "learning_rate": 1.9936708860759496e-05,
      "loss": 0.0235,
      "step": 97
    },
    {
      "epoch": 6.0625,
      "eval_accuracy": 0.828,
      "eval_loss": 0.4142768681049347,
      "eval_runtime": 4.9339,
      "eval_samples_per_second": 50.67,
      "eval_steps_per_second": 1.621,
      "step": 97
    },
    {
      "epoch": 6.125,
      "grad_norm": 2.0625,
      "learning_rate": 1.962025316455696e-05,
      "loss": 0.0247,
      "step": 98
    },
    {
      "epoch": 6.125,
      "eval_accuracy": 0.832,
      "eval_loss": 0.41069871187210083,
      "eval_runtime": 4.9302,
      "eval_samples_per_second": 50.708,
      "eval_steps_per_second": 1.623,
      "step": 98
    },
    {
      "epoch": 6.1875,
      "grad_norm": 1.2265625,
      "learning_rate": 1.9303797468354432e-05,
      "loss": 0.0148,
      "step": 99
    },
    {
      "epoch": 6.1875,
      "eval_accuracy": 0.836,
      "eval_loss": 0.4168325662612915,
      "eval_runtime": 4.9333,
      "eval_samples_per_second": 50.676,
      "eval_steps_per_second": 1.622,
      "step": 99
    },
    {
      "epoch": 6.25,
      "grad_norm": 3.078125,
      "learning_rate": 1.89873417721519e-05,
      "loss": 0.0365,
      "step": 100
    },
    {
      "epoch": 6.25,
      "eval_accuracy": 0.84,
      "eval_loss": 0.4236372709274292,
      "eval_runtime": 4.9348,
      "eval_samples_per_second": 50.661,
      "eval_steps_per_second": 1.621,
      "step": 100
    },
    {
      "epoch": 6.3125,
      "grad_norm": 1.46875,
      "learning_rate": 1.8670886075949368e-05,
      "loss": 0.0211,
      "step": 101
    },
    {
      "epoch": 6.3125,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4290027320384979,
      "eval_runtime": 4.9336,
      "eval_samples_per_second": 50.673,
      "eval_steps_per_second": 1.622,
      "step": 101
    },
    {
      "epoch": 6.375,
      "grad_norm": 1.40625,
      "learning_rate": 1.8354430379746836e-05,
      "loss": 0.0183,
      "step": 102
    },
    {
      "epoch": 6.375,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4270361661911011,
      "eval_runtime": 4.9305,
      "eval_samples_per_second": 50.704,
      "eval_steps_per_second": 1.623,
      "step": 102
    },
    {
      "epoch": 6.4375,
      "grad_norm": 1.4375,
      "learning_rate": 1.8037974683544304e-05,
      "loss": 0.0224,
      "step": 103
    },
    {
      "epoch": 6.4375,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4209546446800232,
      "eval_runtime": 5.0135,
      "eval_samples_per_second": 49.865,
      "eval_steps_per_second": 1.596,
      "step": 103
    },
    {
      "epoch": 6.5,
      "grad_norm": 3.296875,
      "learning_rate": 1.7721518987341772e-05,
      "loss": 0.0437,
      "step": 104
    },
    {
      "epoch": 6.5,
      "eval_accuracy": 0.844,
      "eval_loss": 0.4190281331539154,
      "eval_runtime": 5.1552,
      "eval_samples_per_second": 48.495,
      "eval_steps_per_second": 1.552,
      "step": 104
    },
    {
      "epoch": 6.5625,
      "grad_norm": 1.0859375,
      "learning_rate": 1.7405063291139243e-05,
      "loss": 0.0113,
      "step": 105
    },
    {
      "epoch": 6.5625,
      "eval_accuracy": 0.844,
      "eval_loss": 0.41917330026626587,
      "eval_runtime": 5.2448,
      "eval_samples_per_second": 47.666,
      "eval_steps_per_second": 1.525,
      "step": 105
    },
    {
      "epoch": 6.625,
      "grad_norm": 2.96875,
      "learning_rate": 1.7088607594936708e-05,
      "loss": 0.0404,
      "step": 106
    },
    {
      "epoch": 6.625,
      "eval_accuracy": 0.84,
      "eval_loss": 0.4207764267921448,
      "eval_runtime": 5.3617,
      "eval_samples_per_second": 46.627,
      "eval_steps_per_second": 1.492,
      "step": 106
    },
    {
      "epoch": 6.6875,
      "grad_norm": 1.1484375,
      "learning_rate": 1.677215189873418e-05,
      "loss": 0.0127,
      "step": 107
    },
    {
      "epoch": 6.6875,
      "eval_accuracy": 0.848,
      "eval_loss": 0.42192724347114563,
      "eval_runtime": 5.0116,
      "eval_samples_per_second": 49.884,
      "eval_steps_per_second": 1.596,
      "step": 107
    },
    {
      "epoch": 6.75,
      "grad_norm": 0.62109375,
      "learning_rate": 1.6455696202531644e-05,
      "loss": 0.011,
      "step": 108
    },
    {
      "epoch": 6.75,
      "eval_accuracy": 0.848,
      "eval_loss": 0.4295770227909088,
      "eval_runtime": 5.1327,
      "eval_samples_per_second": 48.708,
      "eval_steps_per_second": 1.559,
      "step": 108
    },
    {
      "epoch": 6.8125,
      "grad_norm": 0.80078125,
      "learning_rate": 1.6139240506329115e-05,
      "loss": 0.0088,
      "step": 109
    },
    {
      "epoch": 6.8125,
      "eval_accuracy": 0.848,
      "eval_loss": 0.4343023896217346,
      "eval_runtime": 5.1895,
      "eval_samples_per_second": 48.174,
      "eval_steps_per_second": 1.542,
      "step": 109
    },
    {
      "epoch": 6.875,
      "grad_norm": 2.734375,
      "learning_rate": 1.5822784810126583e-05,
      "loss": 0.0285,
      "step": 110
    },
    {
      "epoch": 6.875,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4316946268081665,
      "eval_runtime": 4.9676,
      "eval_samples_per_second": 50.326,
      "eval_steps_per_second": 1.61,
      "step": 110
    },
    {
      "epoch": 6.9375,
      "grad_norm": 1.984375,
      "learning_rate": 1.550632911392405e-05,
      "loss": 0.0179,
      "step": 111
    },
    {
      "epoch": 6.9375,
      "eval_accuracy": 0.852,
      "eval_loss": 0.43008503317832947,
      "eval_runtime": 4.9342,
      "eval_samples_per_second": 50.666,
      "eval_steps_per_second": 1.621,
      "step": 111
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.15625,
      "learning_rate": 1.5189873417721521e-05,
      "loss": 0.0207,
      "step": 112
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4265301525592804,
      "eval_runtime": 4.9341,
      "eval_samples_per_second": 50.668,
      "eval_steps_per_second": 1.621,
      "step": 112
    },
    {
      "epoch": 7.0625,
      "grad_norm": 0.3671875,
      "learning_rate": 1.4873417721518987e-05,
      "loss": 0.005,
      "step": 113
    },
    {
      "epoch": 7.0625,
      "eval_accuracy": 0.84,
      "eval_loss": 0.42337584495544434,
      "eval_runtime": 4.9377,
      "eval_samples_per_second": 50.631,
      "eval_steps_per_second": 1.62,
      "step": 113
    },
    {
      "epoch": 7.125,
      "grad_norm": 0.408203125,
      "learning_rate": 1.4556962025316457e-05,
      "loss": 0.0072,
      "step": 114
    },
    {
      "epoch": 7.125,
      "eval_accuracy": 0.84,
      "eval_loss": 0.42673492431640625,
      "eval_runtime": 4.9384,
      "eval_samples_per_second": 50.624,
      "eval_steps_per_second": 1.62,
      "step": 114
    },
    {
      "epoch": 7.1875,
      "grad_norm": 0.6796875,
      "learning_rate": 1.4240506329113925e-05,
      "loss": 0.0105,
      "step": 115
    },
    {
      "epoch": 7.1875,
      "eval_accuracy": 0.84,
      "eval_loss": 0.4281560778617859,
      "eval_runtime": 4.9383,
      "eval_samples_per_second": 50.624,
      "eval_steps_per_second": 1.62,
      "step": 115
    },
    {
      "epoch": 7.25,
      "grad_norm": 0.4375,
      "learning_rate": 1.3924050632911393e-05,
      "loss": 0.0062,
      "step": 116
    },
    {
      "epoch": 7.25,
      "eval_accuracy": 0.836,
      "eval_loss": 0.430206298828125,
      "eval_runtime": 4.9311,
      "eval_samples_per_second": 50.698,
      "eval_steps_per_second": 1.622,
      "step": 116
    },
    {
      "epoch": 7.3125,
      "grad_norm": 1.4453125,
      "learning_rate": 1.3607594936708861e-05,
      "loss": 0.015,
      "step": 117
    },
    {
      "epoch": 7.3125,
      "eval_accuracy": 0.844,
      "eval_loss": 0.43175867199897766,
      "eval_runtime": 4.9439,
      "eval_samples_per_second": 50.568,
      "eval_steps_per_second": 1.618,
      "step": 117
    },
    {
      "epoch": 7.375,
      "grad_norm": 0.86328125,
      "learning_rate": 1.3291139240506329e-05,
      "loss": 0.0094,
      "step": 118
    },
    {
      "epoch": 7.375,
      "eval_accuracy": 0.84,
      "eval_loss": 0.4308719336986542,
      "eval_runtime": 4.9314,
      "eval_samples_per_second": 50.696,
      "eval_steps_per_second": 1.622,
      "step": 118
    },
    {
      "epoch": 7.4375,
      "grad_norm": 1.03125,
      "learning_rate": 1.2974683544303799e-05,
      "loss": 0.0086,
      "step": 119
    },
    {
      "epoch": 7.4375,
      "eval_accuracy": 0.84,
      "eval_loss": 0.4366334080696106,
      "eval_runtime": 4.9431,
      "eval_samples_per_second": 50.576,
      "eval_steps_per_second": 1.618,
      "step": 119
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.2658227848101267e-05,
      "loss": 0.0035,
      "step": 120
    },
    {
      "epoch": 7.5,
      "eval_accuracy": 0.848,
      "eval_loss": 0.44147494435310364,
      "eval_runtime": 4.9422,
      "eval_samples_per_second": 50.585,
      "eval_steps_per_second": 1.619,
      "step": 120
    },
    {
      "epoch": 7.5625,
      "grad_norm": 1.015625,
      "learning_rate": 1.2341772151898735e-05,
      "loss": 0.0088,
      "step": 121
    },
    {
      "epoch": 7.5625,
      "eval_accuracy": 0.848,
      "eval_loss": 0.4474000036716461,
      "eval_runtime": 4.943,
      "eval_samples_per_second": 50.577,
      "eval_steps_per_second": 1.618,
      "step": 121
    },
    {
      "epoch": 7.625,
      "grad_norm": 0.3359375,
      "learning_rate": 1.2025316455696203e-05,
      "loss": 0.0062,
      "step": 122
    },
    {
      "epoch": 7.625,
      "eval_accuracy": 0.848,
      "eval_loss": 0.44968822598457336,
      "eval_runtime": 4.9362,
      "eval_samples_per_second": 50.646,
      "eval_steps_per_second": 1.621,
      "step": 122
    },
    {
      "epoch": 7.6875,
      "grad_norm": 1.1953125,
      "learning_rate": 1.170886075949367e-05,
      "loss": 0.01,
      "step": 123
    },
    {
      "epoch": 7.6875,
      "eval_accuracy": 0.848,
      "eval_loss": 0.45175397396087646,
      "eval_runtime": 4.885,
      "eval_samples_per_second": 51.177,
      "eval_steps_per_second": 1.638,
      "step": 123
    },
    {
      "epoch": 7.75,
      "grad_norm": 0.6328125,
      "learning_rate": 1.139240506329114e-05,
      "loss": 0.008,
      "step": 124
    },
    {
      "epoch": 7.75,
      "eval_accuracy": 0.848,
      "eval_loss": 0.45391473174095154,
      "eval_runtime": 4.9403,
      "eval_samples_per_second": 50.604,
      "eval_steps_per_second": 1.619,
      "step": 124
    },
    {
      "epoch": 7.8125,
      "grad_norm": 0.431640625,
      "learning_rate": 1.1075949367088608e-05,
      "loss": 0.0071,
      "step": 125
    },
    {
      "epoch": 7.8125,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4616769552230835,
      "eval_runtime": 4.9324,
      "eval_samples_per_second": 50.685,
      "eval_steps_per_second": 1.622,
      "step": 125
    },
    {
      "epoch": 7.875,
      "grad_norm": 0.72265625,
      "learning_rate": 1.0759493670886076e-05,
      "loss": 0.0065,
      "step": 126
    },
    {
      "epoch": 7.875,
      "eval_accuracy": 0.856,
      "eval_loss": 0.4597095251083374,
      "eval_runtime": 4.9365,
      "eval_samples_per_second": 50.643,
      "eval_steps_per_second": 1.621,
      "step": 126
    },
    {
      "epoch": 7.9375,
      "grad_norm": 0.361328125,
      "learning_rate": 1.0443037974683544e-05,
      "loss": 0.004,
      "step": 127
    },
    {
      "epoch": 7.9375,
      "eval_accuracy": 0.856,
      "eval_loss": 0.4604904055595398,
      "eval_runtime": 4.9367,
      "eval_samples_per_second": 50.641,
      "eval_steps_per_second": 1.621,
      "step": 127
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.0126582278481012e-05,
      "loss": 0.0027,
      "step": 128
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.856,
      "eval_loss": 0.46479332447052,
      "eval_runtime": 4.9361,
      "eval_samples_per_second": 50.648,
      "eval_steps_per_second": 1.621,
      "step": 128
    },
    {
      "epoch": 8.0625,
      "grad_norm": 0.224609375,
      "learning_rate": 9.81012658227848e-06,
      "loss": 0.0029,
      "step": 129
    },
    {
      "epoch": 8.0625,
      "eval_accuracy": 0.852,
      "eval_loss": 0.47178688645362854,
      "eval_runtime": 4.9337,
      "eval_samples_per_second": 50.672,
      "eval_steps_per_second": 1.622,
      "step": 129
    },
    {
      "epoch": 8.125,
      "grad_norm": 0.404296875,
      "learning_rate": 9.49367088607595e-06,
      "loss": 0.0058,
      "step": 130
    },
    {
      "epoch": 8.125,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4701623022556305,
      "eval_runtime": 4.9934,
      "eval_samples_per_second": 50.066,
      "eval_steps_per_second": 1.602,
      "step": 130
    },
    {
      "epoch": 8.1875,
      "grad_norm": 0.39453125,
      "learning_rate": 9.177215189873418e-06,
      "loss": 0.0036,
      "step": 131
    },
    {
      "epoch": 8.1875,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4724094867706299,
      "eval_runtime": 5.0541,
      "eval_samples_per_second": 49.464,
      "eval_steps_per_second": 1.583,
      "step": 131
    },
    {
      "epoch": 8.25,
      "grad_norm": 0.734375,
      "learning_rate": 8.860759493670886e-06,
      "loss": 0.0061,
      "step": 132
    },
    {
      "epoch": 8.25,
      "eval_accuracy": 0.852,
      "eval_loss": 0.47215989232063293,
      "eval_runtime": 5.2055,
      "eval_samples_per_second": 48.026,
      "eval_steps_per_second": 1.537,
      "step": 132
    },
    {
      "epoch": 8.3125,
      "grad_norm": 0.345703125,
      "learning_rate": 8.544303797468354e-06,
      "loss": 0.0044,
      "step": 133
    },
    {
      "epoch": 8.3125,
      "eval_accuracy": 0.852,
      "eval_loss": 0.47440290451049805,
      "eval_runtime": 4.9325,
      "eval_samples_per_second": 50.684,
      "eval_steps_per_second": 1.622,
      "step": 133
    },
    {
      "epoch": 8.375,
      "grad_norm": 0.43359375,
      "learning_rate": 8.227848101265822e-06,
      "loss": 0.0052,
      "step": 134
    },
    {
      "epoch": 8.375,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4775030016899109,
      "eval_runtime": 5.0459,
      "eval_samples_per_second": 49.545,
      "eval_steps_per_second": 1.585,
      "step": 134
    },
    {
      "epoch": 8.4375,
      "grad_norm": 0.291015625,
      "learning_rate": 7.911392405063292e-06,
      "loss": 0.0058,
      "step": 135
    },
    {
      "epoch": 8.4375,
      "eval_accuracy": 0.856,
      "eval_loss": 0.47711512446403503,
      "eval_runtime": 4.9135,
      "eval_samples_per_second": 50.88,
      "eval_steps_per_second": 1.628,
      "step": 135
    },
    {
      "epoch": 8.5,
      "grad_norm": 0.1650390625,
      "learning_rate": 7.5949367088607605e-06,
      "loss": 0.0023,
      "step": 136
    },
    {
      "epoch": 8.5,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4801058769226074,
      "eval_runtime": 4.9192,
      "eval_samples_per_second": 50.821,
      "eval_steps_per_second": 1.626,
      "step": 136
    },
    {
      "epoch": 8.5625,
      "grad_norm": 0.5703125,
      "learning_rate": 7.2784810126582285e-06,
      "loss": 0.0065,
      "step": 137
    },
    {
      "epoch": 8.5625,
      "eval_accuracy": 0.864,
      "eval_loss": 0.47905686497688293,
      "eval_runtime": 5.0724,
      "eval_samples_per_second": 49.287,
      "eval_steps_per_second": 1.577,
      "step": 137
    },
    {
      "epoch": 8.625,
      "grad_norm": 0.291015625,
      "learning_rate": 6.9620253164556965e-06,
      "loss": 0.0026,
      "step": 138
    },
    {
      "epoch": 8.625,
      "eval_accuracy": 0.856,
      "eval_loss": 0.48366859555244446,
      "eval_runtime": 4.9162,
      "eval_samples_per_second": 50.852,
      "eval_steps_per_second": 1.627,
      "step": 138
    },
    {
      "epoch": 8.6875,
      "grad_norm": 0.4140625,
      "learning_rate": 6.6455696202531645e-06,
      "loss": 0.0044,
      "step": 139
    },
    {
      "epoch": 8.6875,
      "eval_accuracy": 0.864,
      "eval_loss": 0.48236754536628723,
      "eval_runtime": 4.9088,
      "eval_samples_per_second": 50.929,
      "eval_steps_per_second": 1.63,
      "step": 139
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.154296875,
      "learning_rate": 6.329113924050633e-06,
      "loss": 0.0024,
      "step": 140
    },
    {
      "epoch": 8.75,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4821201264858246,
      "eval_runtime": 4.9092,
      "eval_samples_per_second": 50.925,
      "eval_steps_per_second": 1.63,
      "step": 140
    },
    {
      "epoch": 8.8125,
      "grad_norm": 0.19140625,
      "learning_rate": 6.012658227848101e-06,
      "loss": 0.003,
      "step": 141
    },
    {
      "epoch": 8.8125,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4807308316230774,
      "eval_runtime": 4.9123,
      "eval_samples_per_second": 50.892,
      "eval_steps_per_second": 1.629,
      "step": 141
    },
    {
      "epoch": 8.875,
      "grad_norm": 0.337890625,
      "learning_rate": 5.69620253164557e-06,
      "loss": 0.0036,
      "step": 142
    },
    {
      "epoch": 8.875,
      "eval_accuracy": 0.864,
      "eval_loss": 0.4836767613887787,
      "eval_runtime": 4.915,
      "eval_samples_per_second": 50.864,
      "eval_steps_per_second": 1.628,
      "step": 142
    },
    {
      "epoch": 8.9375,
      "grad_norm": 0.251953125,
      "learning_rate": 5.379746835443038e-06,
      "loss": 0.0025,
      "step": 143
    },
    {
      "epoch": 8.9375,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4841257631778717,
      "eval_runtime": 4.9119,
      "eval_samples_per_second": 50.897,
      "eval_steps_per_second": 1.629,
      "step": 143
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.53515625,
      "learning_rate": 5.063291139240506e-06,
      "loss": 0.0052,
      "step": 144
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.86,
      "eval_loss": 0.48218366503715515,
      "eval_runtime": 4.9085,
      "eval_samples_per_second": 50.932,
      "eval_steps_per_second": 1.63,
      "step": 144
    },
    {
      "epoch": 9.0625,
      "grad_norm": 0.1767578125,
      "learning_rate": 4.746835443037975e-06,
      "loss": 0.0022,
      "step": 145
    },
    {
      "epoch": 9.0625,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4810963273048401,
      "eval_runtime": 4.9111,
      "eval_samples_per_second": 50.905,
      "eval_steps_per_second": 1.629,
      "step": 145
    },
    {
      "epoch": 9.125,
      "grad_norm": 0.44140625,
      "learning_rate": 4.430379746835443e-06,
      "loss": 0.007,
      "step": 146
    },
    {
      "epoch": 9.125,
      "eval_accuracy": 0.864,
      "eval_loss": 0.482108473777771,
      "eval_runtime": 4.9117,
      "eval_samples_per_second": 50.899,
      "eval_steps_per_second": 1.629,
      "step": 146
    },
    {
      "epoch": 9.1875,
      "grad_norm": 0.17578125,
      "learning_rate": 4.113924050632911e-06,
      "loss": 0.0033,
      "step": 147
    },
    {
      "epoch": 9.1875,
      "eval_accuracy": 0.856,
      "eval_loss": 0.48591578006744385,
      "eval_runtime": 4.9102,
      "eval_samples_per_second": 50.914,
      "eval_steps_per_second": 1.629,
      "step": 147
    },
    {
      "epoch": 9.25,
      "grad_norm": 0.1552734375,
      "learning_rate": 3.7974683544303802e-06,
      "loss": 0.0025,
      "step": 148
    },
    {
      "epoch": 9.25,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4883337914943695,
      "eval_runtime": 4.9105,
      "eval_samples_per_second": 50.912,
      "eval_steps_per_second": 1.629,
      "step": 148
    },
    {
      "epoch": 9.3125,
      "grad_norm": 0.302734375,
      "learning_rate": 3.4810126582278482e-06,
      "loss": 0.0049,
      "step": 149
    },
    {
      "epoch": 9.3125,
      "eval_accuracy": 0.864,
      "eval_loss": 0.4839572012424469,
      "eval_runtime": 4.9139,
      "eval_samples_per_second": 50.876,
      "eval_steps_per_second": 1.628,
      "step": 149
    },
    {
      "epoch": 9.375,
      "grad_norm": 0.1611328125,
      "learning_rate": 3.1645569620253167e-06,
      "loss": 0.0026,
      "step": 150
    },
    {
      "epoch": 9.375,
      "eval_accuracy": 0.856,
      "eval_loss": 0.4814904034137726,
      "eval_runtime": 4.9134,
      "eval_samples_per_second": 50.881,
      "eval_steps_per_second": 1.628,
      "step": 150
    },
    {
      "epoch": 9.4375,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.848101265822785e-06,
      "loss": 0.0021,
      "step": 151
    },
    {
      "epoch": 9.4375,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4863939881324768,
      "eval_runtime": 4.914,
      "eval_samples_per_second": 50.875,
      "eval_steps_per_second": 1.628,
      "step": 151
    },
    {
      "epoch": 9.5,
      "grad_norm": 0.1484375,
      "learning_rate": 2.531645569620253e-06,
      "loss": 0.0023,
      "step": 152
    },
    {
      "epoch": 9.5,
      "eval_accuracy": 0.864,
      "eval_loss": 0.4879688024520874,
      "eval_runtime": 4.9108,
      "eval_samples_per_second": 50.909,
      "eval_steps_per_second": 1.629,
      "step": 152
    },
    {
      "epoch": 9.5625,
      "grad_norm": 0.263671875,
      "learning_rate": 2.2151898734177215e-06,
      "loss": 0.0039,
      "step": 153
    },
    {
      "epoch": 9.5625,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4850555956363678,
      "eval_runtime": 4.9088,
      "eval_samples_per_second": 50.929,
      "eval_steps_per_second": 1.63,
      "step": 153
    },
    {
      "epoch": 9.625,
      "grad_norm": 0.208984375,
      "learning_rate": 1.8987341772151901e-06,
      "loss": 0.0025,
      "step": 154
    },
    {
      "epoch": 9.625,
      "eval_accuracy": 0.856,
      "eval_loss": 0.48572131991386414,
      "eval_runtime": 4.9101,
      "eval_samples_per_second": 50.916,
      "eval_steps_per_second": 1.629,
      "step": 154
    },
    {
      "epoch": 9.6875,
      "grad_norm": 0.21484375,
      "learning_rate": 1.5822784810126583e-06,
      "loss": 0.004,
      "step": 155
    },
    {
      "epoch": 9.6875,
      "eval_accuracy": 0.864,
      "eval_loss": 0.4816167652606964,
      "eval_runtime": 4.9101,
      "eval_samples_per_second": 50.915,
      "eval_steps_per_second": 1.629,
      "step": 155
    },
    {
      "epoch": 9.75,
      "grad_norm": 0.80078125,
      "learning_rate": 1.2658227848101265e-06,
      "loss": 0.0088,
      "step": 156
    },
    {
      "epoch": 9.75,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4846898913383484,
      "eval_runtime": 5.0723,
      "eval_samples_per_second": 49.288,
      "eval_steps_per_second": 1.577,
      "step": 156
    },
    {
      "epoch": 9.8125,
      "grad_norm": 0.2138671875,
      "learning_rate": 9.493670886075951e-07,
      "loss": 0.0035,
      "step": 157
    },
    {
      "epoch": 9.8125,
      "eval_accuracy": 0.86,
      "eval_loss": 0.48494383692741394,
      "eval_runtime": 4.9092,
      "eval_samples_per_second": 50.925,
      "eval_steps_per_second": 1.63,
      "step": 157
    },
    {
      "epoch": 9.875,
      "grad_norm": 0.1708984375,
      "learning_rate": 6.329113924050633e-07,
      "loss": 0.0027,
      "step": 158
    },
    {
      "epoch": 9.875,
      "eval_accuracy": 0.852,
      "eval_loss": 0.4813316762447357,
      "eval_runtime": 4.9095,
      "eval_samples_per_second": 50.922,
      "eval_steps_per_second": 1.629,
      "step": 158
    },
    {
      "epoch": 9.9375,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.1645569620253163e-07,
      "loss": 0.0022,
      "step": 159
    },
    {
      "epoch": 9.9375,
      "eval_accuracy": 0.86,
      "eval_loss": 0.48336902260780334,
      "eval_runtime": 4.911,
      "eval_samples_per_second": 50.906,
      "eval_steps_per_second": 1.629,
      "step": 159
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.154296875,
      "learning_rate": 0.0,
      "loss": 0.0022,
      "step": 160
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.86,
      "eval_loss": 0.483336478471756,
      "eval_runtime": 4.9132,
      "eval_samples_per_second": 50.883,
      "eval_steps_per_second": 1.628,
      "step": 160
    },
    {
      "epoch": 10.0,
      "step": 160,
      "total_flos": 6.923398955820646e+16,
      "train_loss": 0.22855629230616614,
      "train_runtime": 1357.293,
      "train_samples_per_second": 7.368,
      "train_steps_per_second": 0.118
    }
  ],
  "logging_steps": 1,
  "max_steps": 160,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6.923398955820646e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}