{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 1,
  "global_step": 90,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1111111111111111,
      "grad_norm": 71.10752868652344,
      "learning_rate": 2.5e-05,
      "loss": 1.4086,
      "step": 1
    },
    {
      "epoch": 0.1111111111111111,
      "eval_accuracy": 0.5555555555555556,
      "eval_loss": 1.0401158332824707,
      "eval_runtime": 1.5415,
      "eval_samples_per_second": 46.709,
      "eval_steps_per_second": 3.244,
      "step": 1
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": 70.05866241455078,
      "learning_rate": 5e-05,
      "loss": 1.2969,
      "step": 2
    },
    {
      "epoch": 0.2222222222222222,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.957859456539154,
      "eval_runtime": 1.6886,
      "eval_samples_per_second": 42.639,
      "eval_steps_per_second": 2.961,
      "step": 2
    },
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 51.52368927001953,
      "learning_rate": 4.943181818181818e-05,
      "loss": 1.0192,
      "step": 3
    },
    {
      "epoch": 0.3333333333333333,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 1.2530348300933838,
      "eval_runtime": 1.7908,
      "eval_samples_per_second": 40.206,
      "eval_steps_per_second": 2.792,
      "step": 3
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 28.69840431213379,
      "learning_rate": 4.886363636363637e-05,
      "loss": 0.8252,
      "step": 4
    },
    {
      "epoch": 0.4444444444444444,
      "eval_accuracy": 0.4027777777777778,
      "eval_loss": 1.3205498456954956,
      "eval_runtime": 1.7073,
      "eval_samples_per_second": 42.172,
      "eval_steps_per_second": 2.929,
      "step": 4
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 26.98361587524414,
      "learning_rate": 4.829545454545455e-05,
      "loss": 0.8993,
      "step": 5
    },
    {
      "epoch": 0.5555555555555556,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 1.1885122060775757,
      "eval_runtime": 1.7972,
      "eval_samples_per_second": 40.062,
      "eval_steps_per_second": 2.782,
      "step": 5
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 32.78493881225586,
      "learning_rate": 4.772727272727273e-05,
      "loss": 0.8969,
      "step": 6
    },
    {
      "epoch": 0.6666666666666666,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 1.0518155097961426,
      "eval_runtime": 1.7895,
      "eval_samples_per_second": 40.235,
      "eval_steps_per_second": 2.794,
      "step": 6
    },
    {
      "epoch": 0.7777777777777778,
      "grad_norm": 35.246402740478516,
      "learning_rate": 4.715909090909091e-05,
      "loss": 0.7789,
      "step": 7
    },
    {
      "epoch": 0.7777777777777778,
      "eval_accuracy": 0.4027777777777778,
      "eval_loss": 0.95684814453125,
      "eval_runtime": 1.782,
      "eval_samples_per_second": 40.405,
      "eval_steps_per_second": 2.806,
      "step": 7
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 10.608202934265137,
      "learning_rate": 4.659090909090909e-05,
      "loss": 0.8172,
      "step": 8
    },
    {
      "epoch": 0.8888888888888888,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9124009609222412,
      "eval_runtime": 1.7,
      "eval_samples_per_second": 42.354,
      "eval_steps_per_second": 2.941,
      "step": 8
    },
    {
      "epoch": 1.0,
      "grad_norm": 20.640520095825195,
      "learning_rate": 4.602272727272727e-05,
      "loss": 0.9466,
      "step": 9
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.8895958662033081,
      "eval_runtime": 1.6914,
      "eval_samples_per_second": 42.569,
      "eval_steps_per_second": 2.956,
      "step": 9
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 16.80632972717285,
      "learning_rate": 4.545454545454546e-05,
      "loss": 0.7398,
      "step": 10
    },
    {
      "epoch": 1.1111111111111112,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.8696831464767456,
      "eval_runtime": 1.7281,
      "eval_samples_per_second": 41.664,
      "eval_steps_per_second": 2.893,
      "step": 10
    },
    {
      "epoch": 1.2222222222222223,
      "grad_norm": 33.25798797607422,
      "learning_rate": 4.488636363636364e-05,
      "loss": 0.8393,
      "step": 11
    },
    {
      "epoch": 1.2222222222222223,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.8458387851715088,
      "eval_runtime": 1.6403,
      "eval_samples_per_second": 43.894,
      "eval_steps_per_second": 3.048,
      "step": 11
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 18.833730697631836,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.627,
      "step": 12
    },
    {
      "epoch": 1.3333333333333333,
      "eval_accuracy": 0.4305555555555556,
      "eval_loss": 0.8363681435585022,
      "eval_runtime": 1.7646,
      "eval_samples_per_second": 40.803,
      "eval_steps_per_second": 2.834,
      "step": 12
    },
    {
      "epoch": 1.4444444444444444,
      "grad_norm": 11.44498062133789,
      "learning_rate": 4.375e-05,
      "loss": 0.5691,
      "step": 13
    },
    {
      "epoch": 1.4444444444444444,
      "eval_accuracy": 0.4166666666666667,
      "eval_loss": 0.8569539189338684,
      "eval_runtime": 1.6972,
      "eval_samples_per_second": 42.423,
      "eval_steps_per_second": 2.946,
      "step": 13
    },
    {
      "epoch": 1.5555555555555556,
      "grad_norm": 6.994156360626221,
      "learning_rate": 4.318181818181819e-05,
      "loss": 0.6796,
      "step": 14
    },
    {
      "epoch": 1.5555555555555556,
      "eval_accuracy": 0.4444444444444444,
      "eval_loss": 0.9024861454963684,
      "eval_runtime": 1.6366,
      "eval_samples_per_second": 43.995,
      "eval_steps_per_second": 3.055,
      "step": 14
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 23.839618682861328,
      "learning_rate": 4.261363636363637e-05,
      "loss": 0.6225,
      "step": 15
    },
    {
      "epoch": 1.6666666666666665,
      "eval_accuracy": 0.4722222222222222,
      "eval_loss": 0.9257575273513794,
      "eval_runtime": 1.6451,
      "eval_samples_per_second": 43.767,
      "eval_steps_per_second": 3.039,
      "step": 15
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 23.61461639404297,
      "learning_rate": 4.204545454545455e-05,
      "loss": 0.6299,
      "step": 16
    },
    {
      "epoch": 1.7777777777777777,
      "eval_accuracy": 0.4583333333333333,
      "eval_loss": 0.9147135615348816,
      "eval_runtime": 1.7411,
      "eval_samples_per_second": 41.353,
      "eval_steps_per_second": 2.872,
      "step": 16
    },
    {
      "epoch": 1.8888888888888888,
      "grad_norm": 15.60508918762207,
      "learning_rate": 4.1477272727272734e-05,
      "loss": 0.5009,
      "step": 17
    },
    {
      "epoch": 1.8888888888888888,
      "eval_accuracy": 0.4583333333333333,
      "eval_loss": 0.8712971806526184,
      "eval_runtime": 1.688,
      "eval_samples_per_second": 42.653,
      "eval_steps_per_second": 2.962,
      "step": 17
    },
    {
      "epoch": 2.0,
      "grad_norm": 14.364876747131348,
      "learning_rate": 4.0909090909090915e-05,
      "loss": 0.4875,
      "step": 18
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4305555555555556,
      "eval_loss": 0.8122490644454956,
      "eval_runtime": 1.6995,
      "eval_samples_per_second": 42.366,
      "eval_steps_per_second": 2.942,
      "step": 18
    },
    {
      "epoch": 2.111111111111111,
      "grad_norm": 9.64327621459961,
      "learning_rate": 4.034090909090909e-05,
      "loss": 0.4547,
      "step": 19
    },
    {
      "epoch": 2.111111111111111,
      "eval_accuracy": 0.4444444444444444,
      "eval_loss": 0.7631157636642456,
      "eval_runtime": 1.7098,
      "eval_samples_per_second": 42.11,
      "eval_steps_per_second": 2.924,
      "step": 19
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 8.928176879882812,
      "learning_rate": 3.9772727272727275e-05,
      "loss": 0.3933,
      "step": 20
    },
    {
      "epoch": 2.2222222222222223,
      "eval_accuracy": 0.5138888888888888,
      "eval_loss": 0.7445746660232544,
      "eval_runtime": 1.7154,
      "eval_samples_per_second": 41.973,
      "eval_steps_per_second": 2.915,
      "step": 20
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 12.450410842895508,
      "learning_rate": 3.9204545454545456e-05,
      "loss": 0.3994,
      "step": 21
    },
    {
      "epoch": 2.3333333333333335,
      "eval_accuracy": 0.5416666666666666,
      "eval_loss": 0.7361382246017456,
      "eval_runtime": 1.6893,
      "eval_samples_per_second": 42.621,
      "eval_steps_per_second": 2.96,
      "step": 21
    },
    {
      "epoch": 2.4444444444444446,
      "grad_norm": 9.766682624816895,
      "learning_rate": 3.8636363636363636e-05,
      "loss": 0.4413,
      "step": 22
    },
    {
      "epoch": 2.4444444444444446,
      "eval_accuracy": 0.5972222222222222,
      "eval_loss": 0.7318115234375,
      "eval_runtime": 1.6931,
      "eval_samples_per_second": 42.526,
      "eval_steps_per_second": 2.953,
      "step": 22
    },
    {
      "epoch": 2.5555555555555554,
      "grad_norm": 9.650227546691895,
      "learning_rate": 3.8068181818181816e-05,
      "loss": 0.4032,
      "step": 23
    },
    {
      "epoch": 2.5555555555555554,
      "eval_accuracy": 0.5972222222222222,
      "eval_loss": 0.7219780683517456,
      "eval_runtime": 1.6928,
      "eval_samples_per_second": 42.532,
      "eval_steps_per_second": 2.954,
      "step": 23
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 8.477367401123047,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.4308,
      "step": 24
    },
    {
      "epoch": 2.6666666666666665,
      "eval_accuracy": 0.6527777777777778,
      "eval_loss": 0.7055935263633728,
      "eval_runtime": 1.7353,
      "eval_samples_per_second": 41.492,
      "eval_steps_per_second": 2.881,
      "step": 24
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 8.381991386413574,
      "learning_rate": 3.6931818181818184e-05,
      "loss": 0.3827,
      "step": 25
    },
    {
      "epoch": 2.7777777777777777,
      "eval_accuracy": 0.6666666666666666,
      "eval_loss": 0.6947495937347412,
      "eval_runtime": 1.69,
      "eval_samples_per_second": 42.604,
      "eval_steps_per_second": 2.959,
      "step": 25
    },
    {
      "epoch": 2.888888888888889,
      "grad_norm": 18.475757598876953,
      "learning_rate": 3.6363636363636364e-05,
      "loss": 0.3719,
      "step": 26
    },
    {
      "epoch": 2.888888888888889,
      "eval_accuracy": 0.6388888888888888,
      "eval_loss": 0.695068359375,
      "eval_runtime": 1.7654,
      "eval_samples_per_second": 40.785,
      "eval_steps_per_second": 2.832,
      "step": 26
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.149219036102295,
      "learning_rate": 3.579545454545455e-05,
      "loss": 0.2756,
      "step": 27
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.625,
      "eval_loss": 0.7118733525276184,
      "eval_runtime": 1.7076,
      "eval_samples_per_second": 42.165,
      "eval_steps_per_second": 2.928,
      "step": 27
    },
    {
      "epoch": 3.111111111111111,
      "grad_norm": 6.201833248138428,
      "learning_rate": 3.522727272727273e-05,
      "loss": 0.3048,
      "step": 28
    },
    {
      "epoch": 3.111111111111111,
      "eval_accuracy": 0.6388888888888888,
      "eval_loss": 0.7207743525505066,
      "eval_runtime": 1.6906,
      "eval_samples_per_second": 42.59,
      "eval_steps_per_second": 2.958,
      "step": 28
    },
    {
      "epoch": 3.2222222222222223,
      "grad_norm": 10.069761276245117,
      "learning_rate": 3.465909090909091e-05,
      "loss": 0.2833,
      "step": 29
    },
    {
      "epoch": 3.2222222222222223,
      "eval_accuracy": 0.6111111111111112,
      "eval_loss": 0.758392333984375,
      "eval_runtime": 1.689,
      "eval_samples_per_second": 42.629,
      "eval_steps_per_second": 2.96,
      "step": 29
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 14.6091947555542,
      "learning_rate": 3.409090909090909e-05,
      "loss": 0.2192,
      "step": 30
    },
    {
      "epoch": 3.3333333333333335,
      "eval_accuracy": 0.625,
      "eval_loss": 0.7625613808631897,
      "eval_runtime": 1.6873,
      "eval_samples_per_second": 42.671,
      "eval_steps_per_second": 2.963,
      "step": 30
    },
    {
      "epoch": 3.4444444444444446,
      "grad_norm": 12.838824272155762,
      "learning_rate": 3.352272727272727e-05,
      "loss": 0.159,
      "step": 31
    },
    {
      "epoch": 3.4444444444444446,
      "eval_accuracy": 0.6666666666666666,
      "eval_loss": 0.7382490634918213,
      "eval_runtime": 1.689,
      "eval_samples_per_second": 42.629,
      "eval_steps_per_second": 2.96,
      "step": 31
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 5.576656341552734,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.1545,
      "step": 32
    },
    {
      "epoch": 3.5555555555555554,
      "eval_accuracy": 0.6805555555555556,
      "eval_loss": 0.7160136103630066,
      "eval_runtime": 1.688,
      "eval_samples_per_second": 42.655,
      "eval_steps_per_second": 2.962,
      "step": 32
    },
    {
      "epoch": 3.6666666666666665,
      "grad_norm": 3.911334753036499,
      "learning_rate": 3.238636363636364e-05,
      "loss": 0.2048,
      "step": 33
    },
    {
      "epoch": 3.6666666666666665,
      "eval_accuracy": 0.6944444444444444,
      "eval_loss": 0.7091013789176941,
      "eval_runtime": 1.6904,
      "eval_samples_per_second": 42.594,
      "eval_steps_per_second": 2.958,
      "step": 33
    },
    {
      "epoch": 3.7777777777777777,
      "grad_norm": 4.566244125366211,
      "learning_rate": 3.181818181818182e-05,
      "loss": 0.1285,
      "step": 34
    },
    {
      "epoch": 3.7777777777777777,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.7089945673942566,
      "eval_runtime": 1.6878,
      "eval_samples_per_second": 42.659,
      "eval_steps_per_second": 2.962,
      "step": 34
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 4.949621677398682,
      "learning_rate": 3.125e-05,
      "loss": 0.0972,
      "step": 35
    },
    {
      "epoch": 3.888888888888889,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.7177971601486206,
      "eval_runtime": 1.7101,
      "eval_samples_per_second": 42.104,
      "eval_steps_per_second": 2.924,
      "step": 35
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.815789699554443,
      "learning_rate": 3.068181818181818e-05,
      "loss": 0.1396,
      "step": 36
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7361111111111112,
      "eval_loss": 0.7336485385894775,
      "eval_runtime": 1.704,
      "eval_samples_per_second": 42.252,
      "eval_steps_per_second": 2.934,
      "step": 36
    },
    {
      "epoch": 4.111111111111111,
      "grad_norm": 5.0971174240112305,
      "learning_rate": 3.0113636363636365e-05,
      "loss": 0.1058,
      "step": 37
    },
    {
      "epoch": 4.111111111111111,
      "eval_accuracy": 0.6944444444444444,
      "eval_loss": 0.7677508592605591,
      "eval_runtime": 1.6634,
      "eval_samples_per_second": 43.285,
      "eval_steps_per_second": 3.006,
      "step": 37
    },
    {
      "epoch": 4.222222222222222,
      "grad_norm": 3.026982545852661,
      "learning_rate": 2.954545454545455e-05,
      "loss": 0.1036,
      "step": 38
    },
    {
      "epoch": 4.222222222222222,
      "eval_accuracy": 0.6666666666666666,
      "eval_loss": 0.81514573097229,
      "eval_runtime": 1.7527,
      "eval_samples_per_second": 41.08,
      "eval_steps_per_second": 2.853,
      "step": 38
    },
    {
      "epoch": 4.333333333333333,
      "grad_norm": 3.815314531326294,
      "learning_rate": 2.8977272727272732e-05,
      "loss": 0.0651,
      "step": 39
    },
    {
      "epoch": 4.333333333333333,
      "eval_accuracy": 0.6944444444444444,
      "eval_loss": 0.8406186699867249,
      "eval_runtime": 1.7692,
      "eval_samples_per_second": 40.696,
      "eval_steps_per_second": 2.826,
      "step": 39
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 2.897125244140625,
      "learning_rate": 2.8409090909090912e-05,
      "loss": 0.0643,
      "step": 40
    },
    {
      "epoch": 4.444444444444445,
      "eval_accuracy": 0.6944444444444444,
      "eval_loss": 0.8639495372772217,
      "eval_runtime": 1.7505,
      "eval_samples_per_second": 41.132,
      "eval_steps_per_second": 2.856,
      "step": 40
    },
    {
      "epoch": 4.555555555555555,
      "grad_norm": 8.826624870300293,
      "learning_rate": 2.784090909090909e-05,
      "loss": 0.0708,
      "step": 41
    },
    {
      "epoch": 4.555555555555555,
      "eval_accuracy": 0.7083333333333334,
      "eval_loss": 0.8449772000312805,
      "eval_runtime": 1.7389,
      "eval_samples_per_second": 41.404,
      "eval_steps_per_second": 2.875,
      "step": 41
    },
    {
      "epoch": 4.666666666666667,
      "grad_norm": 4.310912132263184,
      "learning_rate": 2.7272727272727273e-05,
      "loss": 0.0571,
      "step": 42
    },
    {
      "epoch": 4.666666666666667,
      "eval_accuracy": 0.7222222222222222,
      "eval_loss": 0.8496870398521423,
      "eval_runtime": 1.6953,
      "eval_samples_per_second": 42.471,
      "eval_steps_per_second": 2.949,
      "step": 42
    },
    {
      "epoch": 4.777777777777778,
      "grad_norm": 0.9157137870788574,
      "learning_rate": 2.6704545454545453e-05,
      "loss": 0.0101,
      "step": 43
    },
    {
      "epoch": 4.777777777777778,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8314664363861084,
      "eval_runtime": 1.696,
      "eval_samples_per_second": 42.452,
      "eval_steps_per_second": 2.948,
      "step": 43
    },
    {
      "epoch": 4.888888888888889,
      "grad_norm": 4.54341983795166,
      "learning_rate": 2.6136363636363637e-05,
      "loss": 0.0345,
      "step": 44
    },
    {
      "epoch": 4.888888888888889,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8536232709884644,
      "eval_runtime": 1.7637,
      "eval_samples_per_second": 40.822,
      "eval_steps_per_second": 2.835,
      "step": 44
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.6652591228485107,
      "learning_rate": 2.5568181818181817e-05,
      "loss": 0.0194,
      "step": 45
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8400527834892273,
      "eval_runtime": 1.7745,
      "eval_samples_per_second": 40.575,
      "eval_steps_per_second": 2.818,
      "step": 45
    },
    {
      "epoch": 5.111111111111111,
      "grad_norm": 0.518764078617096,
      "learning_rate": 2.5e-05,
      "loss": 0.005,
      "step": 46
    },
    {
      "epoch": 5.111111111111111,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8562726378440857,
      "eval_runtime": 1.7488,
      "eval_samples_per_second": 41.171,
      "eval_steps_per_second": 2.859,
      "step": 46
    },
    {
      "epoch": 5.222222222222222,
      "grad_norm": 0.5080448389053345,
      "learning_rate": 2.4431818181818185e-05,
      "loss": 0.0092,
      "step": 47
    },
    {
      "epoch": 5.222222222222222,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8858435153961182,
      "eval_runtime": 1.7362,
      "eval_samples_per_second": 41.469,
      "eval_steps_per_second": 2.88,
      "step": 47
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 1.4881385564804077,
      "learning_rate": 2.3863636363636365e-05,
      "loss": 0.0117,
      "step": 48
    },
    {
      "epoch": 5.333333333333333,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 0.9373347163200378,
      "eval_runtime": 1.7561,
      "eval_samples_per_second": 40.999,
      "eval_steps_per_second": 2.847,
      "step": 48
    },
    {
      "epoch": 5.444444444444445,
      "grad_norm": 0.4491148293018341,
      "learning_rate": 2.3295454545454546e-05,
      "loss": 0.006,
      "step": 49
    },
    {
      "epoch": 5.444444444444445,
      "eval_accuracy": 0.7638888888888888,
      "eval_loss": 0.9894835352897644,
      "eval_runtime": 1.7035,
      "eval_samples_per_second": 42.267,
      "eval_steps_per_second": 2.935,
      "step": 49
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 0.5586861371994019,
      "learning_rate": 2.272727272727273e-05,
      "loss": 0.0063,
      "step": 50
    },
    {
      "epoch": 5.555555555555555,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.0309796333312988,
      "eval_runtime": 1.7142,
      "eval_samples_per_second": 42.001,
      "eval_steps_per_second": 2.917,
      "step": 50
    },
    {
      "epoch": 5.666666666666667,
      "grad_norm": 1.2758698463439941,
      "learning_rate": 2.215909090909091e-05,
      "loss": 0.0057,
      "step": 51
    },
    {
      "epoch": 5.666666666666667,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.0968831777572632,
      "eval_runtime": 1.6931,
      "eval_samples_per_second": 42.526,
      "eval_steps_per_second": 2.953,
      "step": 51
    },
    {
      "epoch": 5.777777777777778,
      "grad_norm": 0.4022962152957916,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.0024,
      "step": 52
    },
    {
      "epoch": 5.777777777777778,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.1657804250717163,
      "eval_runtime": 1.7525,
      "eval_samples_per_second": 41.083,
      "eval_steps_per_second": 2.853,
      "step": 52
    },
    {
      "epoch": 5.888888888888889,
      "grad_norm": 8.325399398803711,
      "learning_rate": 2.1022727272727274e-05,
      "loss": 0.0283,
      "step": 53
    },
    {
      "epoch": 5.888888888888889,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.1517949104309082,
      "eval_runtime": 1.7207,
      "eval_samples_per_second": 41.843,
      "eval_steps_per_second": 2.906,
      "step": 53
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.6889083981513977,
      "learning_rate": 2.0454545454545457e-05,
      "loss": 0.0031,
      "step": 54
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.131449580192566,
      "eval_runtime": 1.7267,
      "eval_samples_per_second": 41.698,
      "eval_steps_per_second": 2.896,
      "step": 54
    },
    {
      "epoch": 6.111111111111111,
      "grad_norm": 0.12238375097513199,
      "learning_rate": 1.9886363636363638e-05,
      "loss": 0.0006,
      "step": 55
    },
    {
      "epoch": 6.111111111111111,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.120394229888916,
      "eval_runtime": 1.7518,
      "eval_samples_per_second": 41.1,
      "eval_steps_per_second": 2.854,
      "step": 55
    },
    {
      "epoch": 6.222222222222222,
      "grad_norm": 0.04647090286016464,
      "learning_rate": 1.9318181818181818e-05,
      "loss": 0.0004,
      "step": 56
    },
    {
      "epoch": 6.222222222222222,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.1473143100738525,
      "eval_runtime": 1.6957,
      "eval_samples_per_second": 42.461,
      "eval_steps_per_second": 2.949,
      "step": 56
    },
    {
      "epoch": 6.333333333333333,
      "grad_norm": 0.07002800703048706,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0004,
      "step": 57
    },
    {
      "epoch": 6.333333333333333,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.1827261447906494,
      "eval_runtime": 1.6914,
      "eval_samples_per_second": 42.567,
      "eval_steps_per_second": 2.956,
      "step": 57
    },
    {
      "epoch": 6.444444444444445,
      "grad_norm": 0.057534411549568176,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 0.0002,
      "step": 58
    },
    {
      "epoch": 6.444444444444445,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.224190592765808,
      "eval_runtime": 1.6907,
      "eval_samples_per_second": 42.587,
      "eval_steps_per_second": 2.957,
      "step": 58
    },
    {
      "epoch": 6.555555555555555,
      "grad_norm": 0.1322898268699646,
      "learning_rate": 1.7613636363636366e-05,
      "loss": 0.0006,
      "step": 59
    },
    {
      "epoch": 6.555555555555555,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.2592931985855103,
      "eval_runtime": 1.6878,
      "eval_samples_per_second": 42.659,
      "eval_steps_per_second": 2.962,
      "step": 59
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 0.014222492463886738,
      "learning_rate": 1.7045454545454546e-05,
      "loss": 0.0001,
      "step": 60
    },
    {
      "epoch": 6.666666666666667,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.281473159790039,
      "eval_runtime": 1.6928,
      "eval_samples_per_second": 42.532,
      "eval_steps_per_second": 2.954,
      "step": 60
    },
    {
      "epoch": 6.777777777777778,
      "grad_norm": 3.633732795715332,
      "learning_rate": 1.6477272727272726e-05,
      "loss": 0.0144,
      "step": 61
    },
    {
      "epoch": 6.777777777777778,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.2948687076568604,
      "eval_runtime": 1.6886,
      "eval_samples_per_second": 42.64,
      "eval_steps_per_second": 2.961,
      "step": 61
    },
    {
      "epoch": 6.888888888888889,
      "grad_norm": 0.6753994226455688,
      "learning_rate": 1.590909090909091e-05,
      "loss": 0.0021,
      "step": 62
    },
    {
      "epoch": 6.888888888888889,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.303892731666565,
      "eval_runtime": 1.691,
      "eval_samples_per_second": 42.579,
      "eval_steps_per_second": 2.957,
      "step": 62
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.047468412667512894,
      "learning_rate": 1.534090909090909e-05,
      "loss": 0.0002,
      "step": 63
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 1.3037166595458984,
      "eval_runtime": 1.6895,
      "eval_samples_per_second": 42.615,
      "eval_steps_per_second": 2.959,
      "step": 63
    },
    {
      "epoch": 7.111111111111111,
      "grad_norm": 0.04554079473018646,
      "learning_rate": 1.4772727272727274e-05,
      "loss": 0.0001,
      "step": 64
    },
    {
      "epoch": 7.111111111111111,
      "eval_accuracy": 0.8194444444444444,
      "eval_loss": 1.3078640699386597,
      "eval_runtime": 1.6898,
      "eval_samples_per_second": 42.608,
      "eval_steps_per_second": 2.959,
      "step": 64
    },
    {
      "epoch": 7.222222222222222,
      "grad_norm": 0.010344511829316616,
      "learning_rate": 1.4204545454545456e-05,
      "loss": 0.0001,
      "step": 65
    },
    {
      "epoch": 7.222222222222222,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3234918117523193,
      "eval_runtime": 1.6967,
      "eval_samples_per_second": 42.434,
      "eval_steps_per_second": 2.947,
      "step": 65
    },
    {
      "epoch": 7.333333333333333,
      "grad_norm": 0.006933713797479868,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 0.0001,
      "step": 66
    },
    {
      "epoch": 7.333333333333333,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3226008415222168,
      "eval_runtime": 1.6947,
      "eval_samples_per_second": 42.485,
      "eval_steps_per_second": 2.95,
      "step": 66
    },
    {
      "epoch": 7.444444444444445,
      "grad_norm": 0.004614418838173151,
      "learning_rate": 1.3068181818181819e-05,
      "loss": 0.0,
      "step": 67
    },
    {
      "epoch": 7.444444444444445,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3420900106430054,
      "eval_runtime": 1.6925,
      "eval_samples_per_second": 42.54,
      "eval_steps_per_second": 2.954,
      "step": 67
    },
    {
      "epoch": 7.555555555555555,
      "grad_norm": 0.011573988012969494,
      "learning_rate": 1.25e-05,
      "loss": 0.0001,
      "step": 68
    },
    {
      "epoch": 7.555555555555555,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3453483581542969,
      "eval_runtime": 1.6925,
      "eval_samples_per_second": 42.539,
      "eval_steps_per_second": 2.954,
      "step": 68
    },
    {
      "epoch": 7.666666666666667,
      "grad_norm": 0.05624835565686226,
      "learning_rate": 1.1931818181818183e-05,
      "loss": 0.0002,
      "step": 69
    },
    {
      "epoch": 7.666666666666667,
      "eval_accuracy": 0.8055555555555556,
      "eval_loss": 1.3481894731521606,
      "eval_runtime": 1.6884,
      "eval_samples_per_second": 42.643,
      "eval_steps_per_second": 2.961,
      "step": 69
    },
    {
      "epoch": 7.777777777777778,
      "grad_norm": 0.026880212128162384,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 0.0001,
      "step": 70
    },
    {
      "epoch": 7.777777777777778,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3691877126693726,
      "eval_runtime": 1.689,
      "eval_samples_per_second": 42.628,
      "eval_steps_per_second": 2.96,
      "step": 70
    },
    {
      "epoch": 7.888888888888889,
      "grad_norm": 0.014328360557556152,
      "learning_rate": 1.0795454545454547e-05,
      "loss": 0.0001,
      "step": 71
    },
    {
      "epoch": 7.888888888888889,
      "eval_accuracy": 0.7916666666666666,
      "eval_loss": 1.3764379024505615,
      "eval_runtime": 1.692,
      "eval_samples_per_second": 42.552,
      "eval_steps_per_second": 2.955,
      "step": 71
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.004069478716701269,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.0,
      "step": 72
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.3799124956130981,
      "eval_runtime": 1.6421,
      "eval_samples_per_second": 43.845,
      "eval_steps_per_second": 3.045,
      "step": 72
    },
    {
      "epoch": 8.11111111111111,
      "grad_norm": 0.029666263610124588,
      "learning_rate": 9.659090909090909e-06,
      "loss": 0.0001,
      "step": 73
    },
    {
      "epoch": 8.11111111111111,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.3902004957199097,
      "eval_runtime": 1.6942,
      "eval_samples_per_second": 42.498,
      "eval_steps_per_second": 2.951,
      "step": 73
    },
    {
      "epoch": 8.222222222222221,
      "grad_norm": 0.005981099791824818,
      "learning_rate": 9.090909090909091e-06,
      "loss": 0.0,
      "step": 74
    },
    {
      "epoch": 8.222222222222221,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.3968830108642578,
      "eval_runtime": 1.6929,
      "eval_samples_per_second": 42.531,
      "eval_steps_per_second": 2.954,
      "step": 74
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 0.008150852285325527,
      "learning_rate": 8.522727272727273e-06,
      "loss": 0.0001,
      "step": 75
    },
    {
      "epoch": 8.333333333333334,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4002034664154053,
      "eval_runtime": 1.692,
      "eval_samples_per_second": 42.554,
      "eval_steps_per_second": 2.955,
      "step": 75
    },
    {
      "epoch": 8.444444444444445,
      "grad_norm": 0.10769578814506531,
      "learning_rate": 7.954545454545455e-06,
      "loss": 0.0004,
      "step": 76
    },
    {
      "epoch": 8.444444444444445,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4117597341537476,
      "eval_runtime": 1.6904,
      "eval_samples_per_second": 42.595,
      "eval_steps_per_second": 2.958,
      "step": 76
    },
    {
      "epoch": 8.555555555555555,
      "grad_norm": 0.026451965793967247,
      "learning_rate": 7.386363636363637e-06,
      "loss": 0.0001,
      "step": 77
    },
    {
      "epoch": 8.555555555555555,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4056729078292847,
      "eval_runtime": 1.6873,
      "eval_samples_per_second": 42.672,
      "eval_steps_per_second": 2.963,
      "step": 77
    },
    {
      "epoch": 8.666666666666666,
      "grad_norm": 0.07714568078517914,
      "learning_rate": 6.818181818181818e-06,
      "loss": 0.0003,
      "step": 78
    },
    {
      "epoch": 8.666666666666666,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4086921215057373,
      "eval_runtime": 1.6884,
      "eval_samples_per_second": 42.643,
      "eval_steps_per_second": 2.961,
      "step": 78
    },
    {
      "epoch": 8.777777777777779,
      "grad_norm": 0.03726133331656456,
      "learning_rate": 6.25e-06,
      "loss": 0.0001,
      "step": 79
    },
    {
      "epoch": 8.777777777777779,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4101053476333618,
      "eval_runtime": 1.6929,
      "eval_samples_per_second": 42.53,
      "eval_steps_per_second": 2.954,
      "step": 79
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 0.01760346069931984,
      "learning_rate": 5.681818181818182e-06,
      "loss": 0.0001,
      "step": 80
    },
    {
      "epoch": 8.88888888888889,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4063137769699097,
      "eval_runtime": 1.6938,
      "eval_samples_per_second": 42.508,
      "eval_steps_per_second": 2.952,
      "step": 80
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.05216860771179199,
      "learning_rate": 5.113636363636364e-06,
      "loss": 0.0003,
      "step": 81
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4048177003860474,
      "eval_runtime": 1.6922,
      "eval_samples_per_second": 42.549,
      "eval_steps_per_second": 2.955,
      "step": 81
    },
    {
      "epoch": 9.11111111111111,
      "grad_norm": 0.019384268671274185,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 0.0001,
      "step": 82
    },
    {
      "epoch": 9.11111111111111,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4030442237854004,
      "eval_runtime": 1.6903,
      "eval_samples_per_second": 42.595,
      "eval_steps_per_second": 2.958,
      "step": 82
    },
    {
      "epoch": 9.222222222222221,
      "grad_norm": 0.04530010744929314,
      "learning_rate": 3.9772727272727275e-06,
      "loss": 0.0002,
      "step": 83
    },
    {
      "epoch": 9.222222222222221,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.406543254852295,
      "eval_runtime": 1.6884,
      "eval_samples_per_second": 42.643,
      "eval_steps_per_second": 2.961,
      "step": 83
    },
    {
      "epoch": 9.333333333333334,
      "grad_norm": 0.006314022000879049,
      "learning_rate": 3.409090909090909e-06,
      "loss": 0.0,
      "step": 84
    },
    {
      "epoch": 9.333333333333334,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4123640060424805,
      "eval_runtime": 1.6879,
      "eval_samples_per_second": 42.656,
      "eval_steps_per_second": 2.962,
      "step": 84
    },
    {
      "epoch": 9.444444444444445,
      "grad_norm": 0.010984798893332481,
      "learning_rate": 2.840909090909091e-06,
      "loss": 0.0001,
      "step": 85
    },
    {
      "epoch": 9.444444444444445,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4055116176605225,
      "eval_runtime": 1.6923,
      "eval_samples_per_second": 42.547,
      "eval_steps_per_second": 2.955,
      "step": 85
    },
    {
      "epoch": 9.555555555555555,
      "grad_norm": 0.04161923751235008,
      "learning_rate": 2.2727272727272728e-06,
      "loss": 0.0002,
      "step": 86
    },
    {
      "epoch": 9.555555555555555,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4036388397216797,
      "eval_runtime": 1.6941,
      "eval_samples_per_second": 42.5,
      "eval_steps_per_second": 2.951,
      "step": 86
    },
    {
      "epoch": 9.666666666666666,
      "grad_norm": 0.02397008240222931,
      "learning_rate": 1.7045454545454546e-06,
      "loss": 0.0001,
      "step": 87
    },
    {
      "epoch": 9.666666666666666,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4052443504333496,
      "eval_runtime": 1.6935,
      "eval_samples_per_second": 42.517,
      "eval_steps_per_second": 2.953,
      "step": 87
    },
    {
      "epoch": 9.777777777777779,
      "grad_norm": 0.020388655364513397,
      "learning_rate": 1.1363636363636364e-06,
      "loss": 0.0001,
      "step": 88
    },
    {
      "epoch": 9.777777777777779,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.403831958770752,
      "eval_runtime": 1.693,
      "eval_samples_per_second": 42.529,
      "eval_steps_per_second": 2.953,
      "step": 88
    },
    {
      "epoch": 9.88888888888889,
      "grad_norm": 0.03793564811348915,
      "learning_rate": 5.681818181818182e-07,
      "loss": 0.0002,
      "step": 89
    },
    {
      "epoch": 9.88888888888889,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4065537452697754,
      "eval_runtime": 1.6926,
      "eval_samples_per_second": 42.537,
      "eval_steps_per_second": 2.954,
      "step": 89
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.003561729798093438,
      "learning_rate": 0.0,
      "loss": 0.0,
      "step": 90
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7777777777777778,
      "eval_loss": 1.4016788005828857,
      "eval_runtime": 1.693,
      "eval_samples_per_second": 42.528,
      "eval_steps_per_second": 2.953,
      "step": 90
    },
    {
      "epoch": 10.0,
      "step": 90,
      "total_flos": 5016736558481408.0,
      "train_loss": 0.22732360793484582,
      "train_runtime": 316.2669,
      "train_samples_per_second": 8.98,
      "train_steps_per_second": 0.285
    }
  ],
  "logging_steps": 1,
  "max_steps": 90,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 5016736558481408.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}