{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 1000,
  "global_step": 121000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 53.65376281738281,
      "learning_rate": 1.9917355371900827e-05,
      "loss": 9.7879,
      "step": 500
    },
    {
      "epoch": 0.01,
      "grad_norm": 92.14137268066406,
      "learning_rate": 1.9834710743801656e-05,
      "loss": 7.9084,
      "step": 1000
    },
    {
      "epoch": 0.01,
      "eval_loss": 7.1100754737854,
      "eval_runtime": 137.7524,
      "eval_samples_per_second": 6.984,
      "eval_steps_per_second": 6.984,
      "step": 1000
    },
    {
      "epoch": 0.01,
      "grad_norm": 116.5856704711914,
      "learning_rate": 1.9752066115702482e-05,
      "loss": 7.3127,
      "step": 1500
    },
    {
      "epoch": 0.02,
      "grad_norm": 31.839763641357422,
      "learning_rate": 1.9669421487603307e-05,
      "loss": 7.1385,
      "step": 2000
    },
    {
      "epoch": 0.02,
      "eval_loss": 6.323878765106201,
      "eval_runtime": 139.7775,
      "eval_samples_per_second": 6.882,
      "eval_steps_per_second": 6.882,
      "step": 2000
    },
    {
      "epoch": 0.02,
      "grad_norm": 81.93315887451172,
      "learning_rate": 1.9586776859504133e-05,
      "loss": 6.5815,
      "step": 2500
    },
    {
      "epoch": 0.02,
      "grad_norm": 91.25088500976562,
      "learning_rate": 1.950413223140496e-05,
      "loss": 6.3278,
      "step": 3000
    },
    {
      "epoch": 0.02,
      "eval_loss": 5.6174092292785645,
      "eval_runtime": 137.9688,
      "eval_samples_per_second": 6.973,
      "eval_steps_per_second": 6.973,
      "step": 3000
    },
    {
      "epoch": 0.03,
      "grad_norm": 42.85502243041992,
      "learning_rate": 1.9421487603305788e-05,
      "loss": 5.9866,
      "step": 3500
    },
    {
      "epoch": 0.03,
      "grad_norm": 128.11082458496094,
      "learning_rate": 1.9338842975206613e-05,
      "loss": 5.6448,
      "step": 4000
    },
    {
      "epoch": 0.03,
      "eval_loss": 5.013957977294922,
      "eval_runtime": 141.6906,
      "eval_samples_per_second": 6.789,
      "eval_steps_per_second": 6.789,
      "step": 4000
    },
    {
      "epoch": 0.04,
      "grad_norm": 94.7215576171875,
      "learning_rate": 1.925619834710744e-05,
      "loss": 5.2837,
      "step": 4500
    },
    {
      "epoch": 0.04,
      "grad_norm": 109.5147705078125,
      "learning_rate": 1.9173553719008268e-05,
      "loss": 5.0691,
      "step": 5000
    },
    {
      "epoch": 0.04,
      "eval_loss": 4.529860973358154,
      "eval_runtime": 139.3388,
      "eval_samples_per_second": 6.904,
      "eval_steps_per_second": 6.904,
      "step": 5000
    },
    {
      "epoch": 0.05,
      "grad_norm": 136.12649536132812,
      "learning_rate": 1.9090909090909094e-05,
      "loss": 4.8339,
      "step": 5500
    },
    {
      "epoch": 0.05,
      "grad_norm": 118.48357391357422,
      "learning_rate": 1.900826446280992e-05,
      "loss": 4.6329,
      "step": 6000
    },
    {
      "epoch": 0.05,
      "eval_loss": 4.244248390197754,
      "eval_runtime": 141.3143,
      "eval_samples_per_second": 6.808,
      "eval_steps_per_second": 6.808,
      "step": 6000
    },
    {
      "epoch": 0.05,
      "grad_norm": 297.8464660644531,
      "learning_rate": 1.8925619834710745e-05,
      "loss": 4.5749,
      "step": 6500
    },
    {
      "epoch": 0.06,
      "grad_norm": 155.1725311279297,
      "learning_rate": 1.884297520661157e-05,
      "loss": 4.6391,
      "step": 7000
    },
    {
      "epoch": 0.06,
      "eval_loss": 4.004535675048828,
      "eval_runtime": 142.1183,
      "eval_samples_per_second": 6.769,
      "eval_steps_per_second": 6.769,
      "step": 7000
    },
    {
      "epoch": 0.06,
      "grad_norm": 82.23986053466797,
      "learning_rate": 1.87603305785124e-05,
      "loss": 4.299,
      "step": 7500
    },
    {
      "epoch": 0.07,
      "grad_norm": 83.66497039794922,
      "learning_rate": 1.8677685950413225e-05,
      "loss": 4.1757,
      "step": 8000
    },
    {
      "epoch": 0.07,
      "eval_loss": 3.7286927700042725,
      "eval_runtime": 141.6853,
      "eval_samples_per_second": 6.79,
      "eval_steps_per_second": 6.79,
      "step": 8000
    },
    {
      "epoch": 0.07,
      "grad_norm": 185.3876190185547,
      "learning_rate": 1.859504132231405e-05,
      "loss": 4.1474,
      "step": 8500
    },
    {
      "epoch": 0.07,
      "grad_norm": 26.485153198242188,
      "learning_rate": 1.851239669421488e-05,
      "loss": 4.3622,
      "step": 9000
    },
    {
      "epoch": 0.07,
      "eval_loss": 3.6635849475860596,
      "eval_runtime": 138.132,
      "eval_samples_per_second": 6.964,
      "eval_steps_per_second": 6.964,
      "step": 9000
    },
    {
      "epoch": 0.08,
      "grad_norm": 158.5999755859375,
      "learning_rate": 1.8429752066115705e-05,
      "loss": 4.0511,
      "step": 9500
    },
    {
      "epoch": 0.08,
      "grad_norm": 48.88996505737305,
      "learning_rate": 1.834710743801653e-05,
      "loss": 4.0734,
      "step": 10000
    },
    {
      "epoch": 0.08,
      "eval_loss": 3.5613434314727783,
      "eval_runtime": 141.6568,
      "eval_samples_per_second": 6.791,
      "eval_steps_per_second": 6.791,
      "step": 10000
    },
    {
      "epoch": 0.09,
      "grad_norm": 46.01389694213867,
      "learning_rate": 1.8264462809917356e-05,
      "loss": 3.9621,
      "step": 10500
    },
    {
      "epoch": 0.09,
      "grad_norm": 311.6304016113281,
      "learning_rate": 1.8181818181818182e-05,
      "loss": 3.8738,
      "step": 11000
    },
    {
      "epoch": 0.09,
      "eval_loss": 3.4829487800598145,
      "eval_runtime": 137.0137,
      "eval_samples_per_second": 7.021,
      "eval_steps_per_second": 7.021,
      "step": 11000
    },
    {
      "epoch": 0.1,
      "grad_norm": 148.29006958007812,
      "learning_rate": 1.809917355371901e-05,
      "loss": 3.7787,
      "step": 11500
    },
    {
      "epoch": 0.1,
      "grad_norm": 87.70325469970703,
      "learning_rate": 1.8016528925619837e-05,
      "loss": 3.8316,
      "step": 12000
    },
    {
      "epoch": 0.1,
      "eval_loss": 3.322986364364624,
      "eval_runtime": 140.8967,
      "eval_samples_per_second": 6.828,
      "eval_steps_per_second": 6.828,
      "step": 12000
    },
    {
      "epoch": 0.1,
      "grad_norm": 87.01234436035156,
      "learning_rate": 1.7933884297520662e-05,
      "loss": 3.9134,
      "step": 12500
    },
    {
      "epoch": 0.11,
      "grad_norm": 308.3322448730469,
      "learning_rate": 1.7851239669421488e-05,
      "loss": 3.7965,
      "step": 13000
    },
    {
      "epoch": 0.11,
      "eval_loss": 3.3614585399627686,
      "eval_runtime": 141.5793,
      "eval_samples_per_second": 6.795,
      "eval_steps_per_second": 6.795,
      "step": 13000
    },
    {
      "epoch": 0.11,
      "grad_norm": 142.58859252929688,
      "learning_rate": 1.7768595041322317e-05,
      "loss": 3.7439,
      "step": 13500
    },
    {
      "epoch": 0.12,
      "grad_norm": 134.673828125,
      "learning_rate": 1.7685950413223143e-05,
      "loss": 3.7948,
      "step": 14000
    },
    {
      "epoch": 0.12,
      "eval_loss": 3.228428363800049,
      "eval_runtime": 136.8867,
      "eval_samples_per_second": 7.028,
      "eval_steps_per_second": 7.028,
      "step": 14000
    },
    {
      "epoch": 0.12,
      "grad_norm": 69.1788330078125,
      "learning_rate": 1.7603305785123968e-05,
      "loss": 3.5873,
      "step": 14500
    },
    {
      "epoch": 0.12,
      "grad_norm": 161.53973388671875,
      "learning_rate": 1.7520661157024794e-05,
      "loss": 3.8479,
      "step": 15000
    },
    {
      "epoch": 0.12,
      "eval_loss": 3.2672481536865234,
      "eval_runtime": 138.7625,
      "eval_samples_per_second": 6.933,
      "eval_steps_per_second": 6.933,
      "step": 15000
    },
    {
      "epoch": 0.13,
      "grad_norm": 128.98785400390625,
      "learning_rate": 1.743801652892562e-05,
      "loss": 3.6716,
      "step": 15500
    },
    {
      "epoch": 0.13,
      "grad_norm": 170.5436553955078,
      "learning_rate": 1.735537190082645e-05,
      "loss": 3.7273,
      "step": 16000
    },
    {
      "epoch": 0.13,
      "eval_loss": 3.1376287937164307,
      "eval_runtime": 136.1603,
      "eval_samples_per_second": 7.065,
      "eval_steps_per_second": 7.065,
      "step": 16000
    },
    {
      "epoch": 0.14,
      "grad_norm": 129.81439208984375,
      "learning_rate": 1.7272727272727274e-05,
      "loss": 3.5567,
      "step": 16500
    },
    {
      "epoch": 0.14,
      "grad_norm": 135.2463836669922,
      "learning_rate": 1.71900826446281e-05,
      "loss": 3.6128,
      "step": 17000
    },
    {
      "epoch": 0.14,
      "eval_loss": 3.1712100505828857,
      "eval_runtime": 140.5239,
      "eval_samples_per_second": 6.846,
      "eval_steps_per_second": 6.846,
      "step": 17000
    },
    {
      "epoch": 0.14,
      "grad_norm": 99.10409545898438,
      "learning_rate": 1.710743801652893e-05,
      "loss": 3.5886,
      "step": 17500
    },
    {
      "epoch": 0.15,
      "grad_norm": 42.10710144042969,
      "learning_rate": 1.7024793388429754e-05,
      "loss": 3.4751,
      "step": 18000
    },
    {
      "epoch": 0.15,
      "eval_loss": 3.1437861919403076,
      "eval_runtime": 141.4419,
      "eval_samples_per_second": 6.801,
      "eval_steps_per_second": 6.801,
      "step": 18000
    },
    {
      "epoch": 0.15,
      "grad_norm": 60.31798553466797,
      "learning_rate": 1.694214876033058e-05,
      "loss": 3.3902,
      "step": 18500
    },
    {
      "epoch": 0.16,
      "grad_norm": 446.0960388183594,
      "learning_rate": 1.6859504132231405e-05,
      "loss": 3.5544,
      "step": 19000
    },
    {
      "epoch": 0.16,
      "eval_loss": 3.0222690105438232,
      "eval_runtime": 140.8408,
      "eval_samples_per_second": 6.83,
      "eval_steps_per_second": 6.83,
      "step": 19000
    },
    {
      "epoch": 0.16,
      "grad_norm": 125.9988784790039,
      "learning_rate": 1.677685950413223e-05,
      "loss": 3.4624,
      "step": 19500
    },
    {
      "epoch": 0.17,
      "grad_norm": 41.044036865234375,
      "learning_rate": 1.669421487603306e-05,
      "loss": 3.3734,
      "step": 20000
    },
    {
      "epoch": 0.17,
      "eval_loss": 2.9934887886047363,
      "eval_runtime": 140.0157,
      "eval_samples_per_second": 6.871,
      "eval_steps_per_second": 6.871,
      "step": 20000
    },
    {
      "epoch": 0.17,
      "grad_norm": 98.76380157470703,
      "learning_rate": 1.6611570247933886e-05,
      "loss": 3.4481,
      "step": 20500
    },
    {
      "epoch": 0.17,
      "grad_norm": 96.16433715820312,
      "learning_rate": 1.652892561983471e-05,
      "loss": 3.6288,
      "step": 21000
    },
    {
      "epoch": 0.17,
      "eval_loss": 2.9670004844665527,
      "eval_runtime": 141.8861,
      "eval_samples_per_second": 6.78,
      "eval_steps_per_second": 6.78,
      "step": 21000
    },
    {
      "epoch": 0.18,
      "grad_norm": 147.28933715820312,
      "learning_rate": 1.644628099173554e-05,
      "loss": 3.622,
      "step": 21500
    },
    {
      "epoch": 0.18,
      "grad_norm": 85.05192565917969,
      "learning_rate": 1.6363636363636366e-05,
      "loss": 3.194,
      "step": 22000
    },
    {
      "epoch": 0.18,
      "eval_loss": 2.9449918270111084,
      "eval_runtime": 140.4537,
      "eval_samples_per_second": 6.849,
      "eval_steps_per_second": 6.849,
      "step": 22000
    },
    {
      "epoch": 0.19,
      "grad_norm": 231.34249877929688,
      "learning_rate": 1.628099173553719e-05,
      "loss": 3.3939,
      "step": 22500
    },
    {
      "epoch": 0.19,
      "grad_norm": 17.25453758239746,
      "learning_rate": 1.6198347107438017e-05,
      "loss": 3.4362,
      "step": 23000
    },
    {
      "epoch": 0.19,
      "eval_loss": 2.9837887287139893,
      "eval_runtime": 140.5356,
      "eval_samples_per_second": 6.845,
      "eval_steps_per_second": 6.845,
      "step": 23000
    },
    {
      "epoch": 0.19,
      "grad_norm": 209.98016357421875,
      "learning_rate": 1.6115702479338843e-05,
      "loss": 3.5168,
      "step": 23500
    },
    {
      "epoch": 0.2,
      "grad_norm": 94.24089050292969,
      "learning_rate": 1.6033057851239672e-05,
      "loss": 3.2853,
      "step": 24000
    },
    {
      "epoch": 0.2,
      "eval_loss": 2.9006521701812744,
      "eval_runtime": 139.8116,
      "eval_samples_per_second": 6.881,
      "eval_steps_per_second": 6.881,
      "step": 24000
    },
    {
      "epoch": 0.2,
      "grad_norm": 59.02330780029297,
      "learning_rate": 1.5950413223140497e-05,
      "loss": 3.2102,
      "step": 24500
    },
    {
      "epoch": 0.21,
      "grad_norm": 114.82744598388672,
      "learning_rate": 1.5867768595041323e-05,
      "loss": 3.3839,
      "step": 25000
    },
    {
      "epoch": 0.21,
      "eval_loss": 2.8425700664520264,
      "eval_runtime": 135.9965,
      "eval_samples_per_second": 7.074,
      "eval_steps_per_second": 7.074,
      "step": 25000
    },
    {
      "epoch": 0.21,
      "grad_norm": 151.8341064453125,
      "learning_rate": 1.5785123966942152e-05,
      "loss": 3.3901,
      "step": 25500
    },
    {
      "epoch": 0.21,
      "grad_norm": 69.12097930908203,
      "learning_rate": 1.5702479338842978e-05,
      "loss": 3.4662,
      "step": 26000
    },
    {
      "epoch": 0.21,
      "eval_loss": 2.8814632892608643,
      "eval_runtime": 139.0607,
      "eval_samples_per_second": 6.918,
      "eval_steps_per_second": 6.918,
      "step": 26000
    },
    {
      "epoch": 0.22,
      "grad_norm": 81.52751159667969,
      "learning_rate": 1.5619834710743803e-05,
      "loss": 3.325,
      "step": 26500
    },
    {
      "epoch": 0.22,
      "grad_norm": 206.8463592529297,
      "learning_rate": 1.553719008264463e-05,
      "loss": 3.3232,
      "step": 27000
    },
    {
      "epoch": 0.22,
      "eval_loss": 2.836251735687256,
      "eval_runtime": 138.3679,
      "eval_samples_per_second": 6.952,
      "eval_steps_per_second": 6.952,
      "step": 27000
    },
    {
      "epoch": 0.23,
      "grad_norm": 82.50727844238281,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 3.2616,
      "step": 27500
    },
    {
      "epoch": 0.23,
      "grad_norm": 131.51040649414062,
      "learning_rate": 1.5371900826446283e-05,
      "loss": 3.3857,
      "step": 28000
    },
    {
      "epoch": 0.23,
      "eval_loss": 2.780090093612671,
      "eval_runtime": 140.6306,
      "eval_samples_per_second": 6.841,
      "eval_steps_per_second": 6.841,
      "step": 28000
    },
    {
      "epoch": 0.24,
      "grad_norm": 66.1187515258789,
      "learning_rate": 1.528925619834711e-05,
      "loss": 3.2072,
      "step": 28500
    },
    {
      "epoch": 0.24,
      "grad_norm": 131.09405517578125,
      "learning_rate": 1.5206611570247936e-05,
      "loss": 3.0476,
      "step": 29000
    },
    {
      "epoch": 0.24,
      "eval_loss": 2.8148860931396484,
      "eval_runtime": 141.206,
      "eval_samples_per_second": 6.813,
      "eval_steps_per_second": 6.813,
      "step": 29000
    },
    {
      "epoch": 0.24,
      "grad_norm": 51.353572845458984,
      "learning_rate": 1.5123966942148762e-05,
      "loss": 3.3377,
      "step": 29500
    },
    {
      "epoch": 0.25,
      "grad_norm": 66.42013549804688,
      "learning_rate": 1.504132231404959e-05,
      "loss": 3.0664,
      "step": 30000
    },
    {
      "epoch": 0.25,
      "eval_loss": 2.776242971420288,
      "eval_runtime": 141.6364,
      "eval_samples_per_second": 6.792,
      "eval_steps_per_second": 6.792,
      "step": 30000
    },
    {
      "epoch": 0.25,
      "grad_norm": 170.49464416503906,
      "learning_rate": 1.4958677685950413e-05,
      "loss": 3.0958,
      "step": 30500
    },
    {
      "epoch": 0.26,
      "grad_norm": 137.87022399902344,
      "learning_rate": 1.487603305785124e-05,
      "loss": 3.2919,
      "step": 31000
    },
    {
      "epoch": 0.26,
      "eval_loss": 2.8100545406341553,
      "eval_runtime": 141.0359,
      "eval_samples_per_second": 6.821,
      "eval_steps_per_second": 6.821,
      "step": 31000
    },
    {
      "epoch": 0.26,
      "grad_norm": 211.2021942138672,
      "learning_rate": 1.4793388429752066e-05,
      "loss": 3.4836,
      "step": 31500
    },
    {
      "epoch": 0.26,
      "grad_norm": 113.34933471679688,
      "learning_rate": 1.4710743801652893e-05,
      "loss": 3.334,
      "step": 32000
    },
    {
      "epoch": 0.26,
      "eval_loss": 2.754260301589966,
      "eval_runtime": 140.5105,
      "eval_samples_per_second": 6.846,
      "eval_steps_per_second": 6.846,
      "step": 32000
    },
    {
      "epoch": 0.27,
      "grad_norm": 114.3346939086914,
      "learning_rate": 1.462809917355372e-05,
      "loss": 3.0715,
      "step": 32500
    },
    {
      "epoch": 0.27,
      "grad_norm": 91.6365966796875,
      "learning_rate": 1.4545454545454546e-05,
      "loss": 3.2307,
      "step": 33000
    },
    {
      "epoch": 0.27,
      "eval_loss": 2.7607641220092773,
      "eval_runtime": 139.2517,
      "eval_samples_per_second": 6.908,
      "eval_steps_per_second": 6.908,
      "step": 33000
    },
    {
      "epoch": 0.28,
      "grad_norm": 6.1861701011657715,
      "learning_rate": 1.4462809917355374e-05,
      "loss": 2.9537,
      "step": 33500
    },
    {
      "epoch": 0.28,
      "grad_norm": 97.13253021240234,
      "learning_rate": 1.4380165289256201e-05,
      "loss": 2.9517,
      "step": 34000
    },
    {
      "epoch": 0.28,
      "eval_loss": 2.805582046508789,
      "eval_runtime": 138.7882,
      "eval_samples_per_second": 6.931,
      "eval_steps_per_second": 6.931,
      "step": 34000
    },
    {
      "epoch": 0.29,
      "grad_norm": 55.73282241821289,
      "learning_rate": 1.4297520661157025e-05,
      "loss": 3.5026,
      "step": 34500
    },
    {
      "epoch": 0.29,
      "grad_norm": 105.15763854980469,
      "learning_rate": 1.4214876033057852e-05,
      "loss": 3.3557,
      "step": 35000
    },
    {
      "epoch": 0.29,
      "eval_loss": 2.7643611431121826,
      "eval_runtime": 140.4066,
      "eval_samples_per_second": 6.852,
      "eval_steps_per_second": 6.852,
      "step": 35000
    },
    {
      "epoch": 0.29,
      "grad_norm": 34.19166564941406,
      "learning_rate": 1.4132231404958678e-05,
      "loss": 3.1291,
      "step": 35500
    },
    {
      "epoch": 0.3,
      "grad_norm": 129.49188232421875,
      "learning_rate": 1.4049586776859505e-05,
      "loss": 3.2801,
      "step": 36000
    },
    {
      "epoch": 0.3,
      "eval_loss": 2.7544772624969482,
      "eval_runtime": 138.7006,
      "eval_samples_per_second": 6.936,
      "eval_steps_per_second": 6.936,
      "step": 36000
    },
    {
      "epoch": 0.3,
      "grad_norm": 36.190162658691406,
      "learning_rate": 1.3966942148760332e-05,
      "loss": 3.1862,
      "step": 36500
    },
    {
      "epoch": 0.31,
      "grad_norm": 126.45512390136719,
      "learning_rate": 1.3884297520661158e-05,
      "loss": 3.1573,
      "step": 37000
    },
    {
      "epoch": 0.31,
      "eval_loss": 2.7139012813568115,
      "eval_runtime": 139.7912,
      "eval_samples_per_second": 6.882,
      "eval_steps_per_second": 6.882,
      "step": 37000
    },
    {
      "epoch": 0.31,
      "grad_norm": 168.74777221679688,
      "learning_rate": 1.3801652892561985e-05,
      "loss": 3.1524,
      "step": 37500
    },
    {
      "epoch": 0.31,
      "grad_norm": 64.23265075683594,
      "learning_rate": 1.3719008264462813e-05,
      "loss": 3.2022,
      "step": 38000
    },
    {
      "epoch": 0.31,
      "eval_loss": 2.726412773132324,
      "eval_runtime": 137.2349,
      "eval_samples_per_second": 7.01,
      "eval_steps_per_second": 7.01,
      "step": 38000
    },
    {
      "epoch": 0.32,
      "grad_norm": 143.0786895751953,
      "learning_rate": 1.3636363636363637e-05,
      "loss": 3.3005,
      "step": 38500
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.35411542654037476,
      "learning_rate": 1.3553719008264464e-05,
      "loss": 3.1852,
      "step": 39000
    },
    {
      "epoch": 0.32,
      "eval_loss": 2.7224111557006836,
      "eval_runtime": 140.911,
      "eval_samples_per_second": 6.827,
      "eval_steps_per_second": 6.827,
      "step": 39000
    },
    {
      "epoch": 0.33,
      "grad_norm": 56.73325729370117,
      "learning_rate": 1.347107438016529e-05,
      "loss": 3.0389,
      "step": 39500
    },
    {
      "epoch": 0.33,
      "grad_norm": 182.5022430419922,
      "learning_rate": 1.3388429752066117e-05,
      "loss": 2.9861,
      "step": 40000
    },
    {
      "epoch": 0.33,
      "eval_loss": 2.7215094566345215,
      "eval_runtime": 142.3112,
      "eval_samples_per_second": 6.76,
      "eval_steps_per_second": 6.76,
      "step": 40000
    },
    {
      "epoch": 0.33,
      "grad_norm": 129.25076293945312,
      "learning_rate": 1.3305785123966944e-05,
      "loss": 3.2484,
      "step": 40500
    },
    {
      "epoch": 0.34,
      "grad_norm": 28.573749542236328,
      "learning_rate": 1.322314049586777e-05,
      "loss": 3.176,
      "step": 41000
    },
    {
      "epoch": 0.34,
      "eval_loss": 2.678982973098755,
      "eval_runtime": 140.2555,
      "eval_samples_per_second": 6.859,
      "eval_steps_per_second": 6.859,
      "step": 41000
    },
    {
      "epoch": 0.34,
      "grad_norm": 29.228971481323242,
      "learning_rate": 1.3140495867768597e-05,
      "loss": 2.898,
      "step": 41500
    },
    {
      "epoch": 0.35,
      "grad_norm": 2.7074639797210693,
      "learning_rate": 1.3057851239669424e-05,
      "loss": 2.9024,
      "step": 42000
    },
    {
      "epoch": 0.35,
      "eval_loss": 2.7119204998016357,
      "eval_runtime": 141.0803,
      "eval_samples_per_second": 6.819,
      "eval_steps_per_second": 6.819,
      "step": 42000
    },
    {
      "epoch": 0.35,
      "grad_norm": 122.27530670166016,
      "learning_rate": 1.2975206611570248e-05,
      "loss": 3.024,
      "step": 42500
    },
    {
      "epoch": 0.36,
      "grad_norm": 19.63726806640625,
      "learning_rate": 1.2892561983471074e-05,
      "loss": 3.2979,
      "step": 43000
    },
    {
      "epoch": 0.36,
      "eval_loss": 2.6496329307556152,
      "eval_runtime": 141.7782,
      "eval_samples_per_second": 6.785,
      "eval_steps_per_second": 6.785,
      "step": 43000
    },
    {
      "epoch": 0.36,
      "grad_norm": 220.49954223632812,
      "learning_rate": 1.2809917355371901e-05,
      "loss": 3.0016,
      "step": 43500
    },
    {
      "epoch": 0.36,
      "grad_norm": 77.60582733154297,
      "learning_rate": 1.2727272727272728e-05,
      "loss": 3.0776,
      "step": 44000
    },
    {
      "epoch": 0.36,
      "eval_loss": 2.6587836742401123,
      "eval_runtime": 139.8192,
      "eval_samples_per_second": 6.88,
      "eval_steps_per_second": 6.88,
      "step": 44000
    },
    {
      "epoch": 0.37,
      "grad_norm": 94.2193832397461,
      "learning_rate": 1.2644628099173554e-05,
      "loss": 3.0418,
      "step": 44500
    },
    {
      "epoch": 0.37,
      "grad_norm": 30.321847915649414,
      "learning_rate": 1.2561983471074381e-05,
      "loss": 3.2145,
      "step": 45000
    },
    {
      "epoch": 0.37,
      "eval_loss": 2.7039036750793457,
      "eval_runtime": 138.137,
      "eval_samples_per_second": 6.964,
      "eval_steps_per_second": 6.964,
      "step": 45000
    },
    {
      "epoch": 0.38,
      "grad_norm": 102.64441680908203,
      "learning_rate": 1.2479338842975209e-05,
      "loss": 2.9082,
      "step": 45500
    },
    {
      "epoch": 0.38,
      "grad_norm": 95.51321411132812,
      "learning_rate": 1.2396694214876034e-05,
      "loss": 3.2535,
      "step": 46000
    },
    {
      "epoch": 0.38,
      "eval_loss": 2.648103952407837,
      "eval_runtime": 139.7372,
      "eval_samples_per_second": 6.884,
      "eval_steps_per_second": 6.884,
      "step": 46000
    },
    {
      "epoch": 0.38,
      "grad_norm": 90.64472961425781,
      "learning_rate": 1.231404958677686e-05,
      "loss": 3.1841,
      "step": 46500
    },
    {
      "epoch": 0.39,
      "grad_norm": 128.23159790039062,
      "learning_rate": 1.2231404958677686e-05,
      "loss": 3.0074,
      "step": 47000
    },
    {
      "epoch": 0.39,
      "eval_loss": 2.6026036739349365,
      "eval_runtime": 139.947,
      "eval_samples_per_second": 6.874,
      "eval_steps_per_second": 6.874,
      "step": 47000
    },
    {
      "epoch": 0.39,
      "grad_norm": 254.29718017578125,
      "learning_rate": 1.2148760330578513e-05,
      "loss": 3.3493,
      "step": 47500
    },
    {
      "epoch": 0.4,
      "grad_norm": 33.48649215698242,
      "learning_rate": 1.206611570247934e-05,
      "loss": 3.007,
      "step": 48000
    },
    {
      "epoch": 0.4,
      "eval_loss": 2.6409358978271484,
      "eval_runtime": 140.003,
      "eval_samples_per_second": 6.871,
      "eval_steps_per_second": 6.871,
      "step": 48000
    },
    {
      "epoch": 0.4,
      "grad_norm": 165.1754608154297,
      "learning_rate": 1.1983471074380166e-05,
      "loss": 3.0267,
      "step": 48500
    },
    {
      "epoch": 0.4,
      "grad_norm": 57.49995803833008,
      "learning_rate": 1.1900826446280993e-05,
      "loss": 3.0913,
      "step": 49000
    },
    {
      "epoch": 0.4,
      "eval_loss": 2.676673412322998,
      "eval_runtime": 139.7897,
      "eval_samples_per_second": 6.882,
      "eval_steps_per_second": 6.882,
      "step": 49000
    },
    {
      "epoch": 0.41,
      "grad_norm": 269.28326416015625,
      "learning_rate": 1.181818181818182e-05,
      "loss": 3.0674,
      "step": 49500
    },
    {
      "epoch": 0.41,
      "grad_norm": 81.5179672241211,
      "learning_rate": 1.1735537190082646e-05,
      "loss": 3.1847,
      "step": 50000
    },
    {
      "epoch": 0.41,
      "eval_loss": 2.6336779594421387,
      "eval_runtime": 140.9712,
      "eval_samples_per_second": 6.824,
      "eval_steps_per_second": 6.824,
      "step": 50000
    },
    {
      "epoch": 0.42,
      "grad_norm": 81.88212585449219,
      "learning_rate": 1.1652892561983472e-05,
      "loss": 2.7742,
      "step": 50500
    },
    {
      "epoch": 0.42,
      "grad_norm": 34.32734298706055,
      "learning_rate": 1.1570247933884297e-05,
      "loss": 2.9359,
      "step": 51000
    },
    {
      "epoch": 0.42,
      "eval_loss": 2.649772882461548,
      "eval_runtime": 139.5432,
      "eval_samples_per_second": 6.894,
      "eval_steps_per_second": 6.894,
      "step": 51000
    },
    {
      "epoch": 0.43,
      "grad_norm": 44.86323165893555,
      "learning_rate": 1.1487603305785125e-05,
      "loss": 3.1545,
      "step": 51500
    },
    {
      "epoch": 0.43,
      "grad_norm": 70.11674499511719,
      "learning_rate": 1.1404958677685952e-05,
      "loss": 2.9488,
      "step": 52000
    },
    {
      "epoch": 0.43,
      "eval_loss": 2.5968177318573,
      "eval_runtime": 142.9712,
      "eval_samples_per_second": 6.729,
      "eval_steps_per_second": 6.729,
      "step": 52000
    },
    {
      "epoch": 0.43,
      "grad_norm": 43.42722702026367,
      "learning_rate": 1.1322314049586777e-05,
      "loss": 3.02,
      "step": 52500
    },
    {
      "epoch": 0.44,
      "grad_norm": 178.18666076660156,
      "learning_rate": 1.1239669421487605e-05,
      "loss": 3.0279,
      "step": 53000
    },
    {
      "epoch": 0.44,
      "eval_loss": 2.625840425491333,
      "eval_runtime": 142.8837,
      "eval_samples_per_second": 6.733,
      "eval_steps_per_second": 6.733,
      "step": 53000
    },
    {
      "epoch": 0.44,
      "grad_norm": 33.83938217163086,
      "learning_rate": 1.1157024793388432e-05,
      "loss": 2.9362,
      "step": 53500
    },
    {
      "epoch": 0.45,
      "grad_norm": 24.059724807739258,
      "learning_rate": 1.1074380165289258e-05,
      "loss": 2.9282,
      "step": 54000
    },
    {
      "epoch": 0.45,
      "eval_loss": 2.671956777572632,
      "eval_runtime": 141.943,
      "eval_samples_per_second": 6.777,
      "eval_steps_per_second": 6.777,
      "step": 54000
    },
    {
      "epoch": 0.45,
      "grad_norm": 150.78285217285156,
      "learning_rate": 1.0991735537190083e-05,
      "loss": 3.1242,
      "step": 54500
    },
    {
      "epoch": 0.45,
      "grad_norm": 129.85336303710938,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 2.9939,
      "step": 55000
    },
    {
      "epoch": 0.45,
      "eval_loss": 2.6020476818084717,
      "eval_runtime": 137.1321,
      "eval_samples_per_second": 7.015,
      "eval_steps_per_second": 7.015,
      "step": 55000
    },
    {
      "epoch": 0.46,
      "grad_norm": 97.6047134399414,
      "learning_rate": 1.0826446280991736e-05,
      "loss": 3.2276,
      "step": 55500
    },
    {
      "epoch": 0.46,
      "grad_norm": 54.10629653930664,
      "learning_rate": 1.0743801652892562e-05,
      "loss": 3.1815,
      "step": 56000
    },
    {
      "epoch": 0.46,
      "eval_loss": 2.6137943267822266,
      "eval_runtime": 140.5161,
      "eval_samples_per_second": 6.846,
      "eval_steps_per_second": 6.846,
      "step": 56000
    },
    {
      "epoch": 0.47,
      "grad_norm": 116.71500396728516,
      "learning_rate": 1.0661157024793389e-05,
      "loss": 2.7576,
      "step": 56500
    },
    {
      "epoch": 0.47,
      "grad_norm": 132.3123016357422,
      "learning_rate": 1.0578512396694216e-05,
      "loss": 2.8927,
      "step": 57000
    },
    {
      "epoch": 0.47,
      "eval_loss": 2.58026385307312,
      "eval_runtime": 137.5438,
      "eval_samples_per_second": 6.994,
      "eval_steps_per_second": 6.994,
      "step": 57000
    },
    {
      "epoch": 0.48,
      "grad_norm": 1.3877036571502686,
      "learning_rate": 1.0495867768595042e-05,
      "loss": 3.1726,
      "step": 57500
    },
    {
      "epoch": 0.48,
      "grad_norm": 23.99924659729004,
      "learning_rate": 1.041322314049587e-05,
      "loss": 3.0603,
      "step": 58000
    },
    {
      "epoch": 0.48,
      "eval_loss": 2.5942938327789307,
      "eval_runtime": 139.9432,
      "eval_samples_per_second": 6.874,
      "eval_steps_per_second": 6.874,
      "step": 58000
    },
    {
      "epoch": 0.48,
      "grad_norm": 89.54573059082031,
      "learning_rate": 1.0330578512396693e-05,
      "loss": 2.9213,
      "step": 58500
    },
    {
      "epoch": 0.49,
      "grad_norm": 117.28192138671875,
      "learning_rate": 1.024793388429752e-05,
      "loss": 2.9974,
      "step": 59000
    },
    {
      "epoch": 0.49,
      "eval_loss": 2.574179172515869,
      "eval_runtime": 137.8963,
      "eval_samples_per_second": 6.976,
      "eval_steps_per_second": 6.976,
      "step": 59000
    },
    {
      "epoch": 0.49,
      "grad_norm": 230.2947540283203,
      "learning_rate": 1.0165289256198348e-05,
      "loss": 2.9884,
      "step": 59500
    },
    {
      "epoch": 0.5,
      "grad_norm": 114.89617156982422,
      "learning_rate": 1.0082644628099174e-05,
      "loss": 3.1789,
      "step": 60000
    },
    {
      "epoch": 0.5,
      "eval_loss": 2.551656723022461,
      "eval_runtime": 139.8913,
      "eval_samples_per_second": 6.877,
      "eval_steps_per_second": 6.877,
      "step": 60000
    },
    {
      "epoch": 0.5,
      "grad_norm": 19.36899185180664,
      "learning_rate": 1e-05,
      "loss": 3.2153,
      "step": 60500
    },
    {
      "epoch": 0.5,
      "grad_norm": 95.15733337402344,
      "learning_rate": 9.917355371900828e-06,
      "loss": 3.0478,
      "step": 61000
    },
    {
      "epoch": 0.5,
      "eval_loss": 2.5420753955841064,
      "eval_runtime": 138.826,
      "eval_samples_per_second": 6.93,
      "eval_steps_per_second": 6.93,
      "step": 61000
    },
    {
      "epoch": 0.51,
      "grad_norm": 164.1331329345703,
      "learning_rate": 9.834710743801654e-06,
      "loss": 3.0467,
      "step": 61500
    },
    {
      "epoch": 0.51,
      "grad_norm": 120.298828125,
      "learning_rate": 9.75206611570248e-06,
      "loss": 3.0664,
      "step": 62000
    },
    {
      "epoch": 0.51,
      "eval_loss": 2.561354875564575,
      "eval_runtime": 138.2294,
      "eval_samples_per_second": 6.959,
      "eval_steps_per_second": 6.959,
      "step": 62000
    },
    {
      "epoch": 0.52,
      "grad_norm": 67.15013122558594,
      "learning_rate": 9.669421487603307e-06,
      "loss": 3.0473,
      "step": 62500
    },
    {
      "epoch": 0.52,
      "grad_norm": 116.40657043457031,
      "learning_rate": 9.586776859504134e-06,
      "loss": 2.8695,
      "step": 63000
    },
    {
      "epoch": 0.52,
      "eval_loss": 2.559004783630371,
      "eval_runtime": 139.5498,
      "eval_samples_per_second": 6.894,
      "eval_steps_per_second": 6.894,
      "step": 63000
    },
    {
      "epoch": 0.52,
      "grad_norm": 16.178442001342773,
      "learning_rate": 9.50413223140496e-06,
      "loss": 3.1084,
      "step": 63500
    },
    {
      "epoch": 0.53,
      "grad_norm": 9.640996932983398,
      "learning_rate": 9.421487603305785e-06,
      "loss": 3.0757,
      "step": 64000
    },
    {
      "epoch": 0.53,
      "eval_loss": 2.553495407104492,
      "eval_runtime": 140.0088,
      "eval_samples_per_second": 6.871,
      "eval_steps_per_second": 6.871,
      "step": 64000
    },
    {
      "epoch": 0.53,
      "grad_norm": 5.591582775115967,
      "learning_rate": 9.338842975206613e-06,
      "loss": 2.9386,
      "step": 64500
    },
    {
      "epoch": 0.54,
      "grad_norm": 62.728607177734375,
      "learning_rate": 9.25619834710744e-06,
      "loss": 2.9709,
      "step": 65000
    },
    {
      "epoch": 0.54,
      "eval_loss": 2.530237913131714,
      "eval_runtime": 139.6525,
      "eval_samples_per_second": 6.889,
      "eval_steps_per_second": 6.889,
      "step": 65000
    },
    {
      "epoch": 0.54,
      "grad_norm": 27.372655868530273,
      "learning_rate": 9.173553719008265e-06,
      "loss": 2.976,
      "step": 65500
    },
    {
      "epoch": 0.55,
      "grad_norm": 89.07634735107422,
      "learning_rate": 9.090909090909091e-06,
      "loss": 2.8215,
      "step": 66000
    },
    {
      "epoch": 0.55,
      "eval_loss": 2.565800905227661,
      "eval_runtime": 140.4523,
      "eval_samples_per_second": 6.849,
      "eval_steps_per_second": 6.849,
      "step": 66000
    },
    {
      "epoch": 0.55,
      "grad_norm": 20.640071868896484,
      "learning_rate": 9.008264462809918e-06,
      "loss": 3.0081,
      "step": 66500
    },
    {
      "epoch": 0.55,
      "grad_norm": 69.5301284790039,
      "learning_rate": 8.925619834710744e-06,
      "loss": 3.077,
      "step": 67000
    },
    {
      "epoch": 0.55,
      "eval_loss": 2.5604488849639893,
      "eval_runtime": 140.6864,
      "eval_samples_per_second": 6.838,
      "eval_steps_per_second": 6.838,
      "step": 67000
    },
    {
      "epoch": 0.56,
      "grad_norm": 46.04232406616211,
      "learning_rate": 8.842975206611571e-06,
      "loss": 3.0368,
      "step": 67500
    },
    {
      "epoch": 0.56,
      "grad_norm": 142.82720947265625,
      "learning_rate": 8.760330578512397e-06,
      "loss": 2.9914,
      "step": 68000
    },
    {
      "epoch": 0.56,
      "eval_loss": 2.561345100402832,
      "eval_runtime": 139.7552,
      "eval_samples_per_second": 6.883,
      "eval_steps_per_second": 6.883,
      "step": 68000
    },
    {
      "epoch": 0.57,
      "grad_norm": 12.521268844604492,
      "learning_rate": 8.677685950413224e-06,
      "loss": 2.8843,
      "step": 68500
    },
    {
      "epoch": 0.57,
      "grad_norm": 2.718876600265503,
      "learning_rate": 8.59504132231405e-06,
      "loss": 2.8383,
      "step": 69000
    },
    {
      "epoch": 0.57,
      "eval_loss": 2.55441951751709,
      "eval_runtime": 140.0164,
      "eval_samples_per_second": 6.871,
      "eval_steps_per_second": 6.871,
      "step": 69000
    },
    {
      "epoch": 0.57,
      "grad_norm": 68.07179260253906,
      "learning_rate": 8.512396694214877e-06,
      "loss": 2.9206,
      "step": 69500
    },
    {
      "epoch": 0.58,
      "grad_norm": 53.18625259399414,
      "learning_rate": 8.429752066115703e-06,
      "loss": 3.0139,
      "step": 70000
    },
    {
      "epoch": 0.58,
      "eval_loss": 2.539428234100342,
      "eval_runtime": 141.3662,
      "eval_samples_per_second": 6.805,
      "eval_steps_per_second": 6.805,
      "step": 70000
    },
    {
      "epoch": 0.58,
      "grad_norm": 38.117530822753906,
      "learning_rate": 8.34710743801653e-06,
      "loss": 2.9579,
      "step": 70500
    },
    {
      "epoch": 0.59,
      "grad_norm": 11.0903902053833,
      "learning_rate": 8.264462809917356e-06,
      "loss": 2.9588,
      "step": 71000
    },
    {
      "epoch": 0.59,
      "eval_loss": 2.5550858974456787,
      "eval_runtime": 140.3679,
      "eval_samples_per_second": 6.853,
      "eval_steps_per_second": 6.853,
      "step": 71000
    },
    {
      "epoch": 0.59,
      "grad_norm": 136.44300842285156,
      "learning_rate": 8.181818181818183e-06,
      "loss": 2.9754,
      "step": 71500
    },
    {
      "epoch": 0.6,
      "grad_norm": 118.072265625,
      "learning_rate": 8.099173553719009e-06,
      "loss": 3.1051,
      "step": 72000
    },
    {
      "epoch": 0.6,
      "eval_loss": 2.5240180492401123,
      "eval_runtime": 140.2239,
      "eval_samples_per_second": 6.86,
      "eval_steps_per_second": 6.86,
      "step": 72000
    },
    {
      "epoch": 0.6,
      "grad_norm": 124.3985824584961,
      "learning_rate": 8.016528925619836e-06,
      "loss": 2.9229,
      "step": 72500
    },
    {
      "epoch": 0.6,
      "grad_norm": 39.26608657836914,
      "learning_rate": 7.933884297520661e-06,
      "loss": 2.9566,
      "step": 73000
    },
    {
      "epoch": 0.6,
      "eval_loss": 2.5205495357513428,
      "eval_runtime": 140.8561,
      "eval_samples_per_second": 6.83,
      "eval_steps_per_second": 6.83,
      "step": 73000
    },
    {
      "epoch": 0.61,
      "grad_norm": 101.39036560058594,
      "learning_rate": 7.851239669421489e-06,
      "loss": 3.1112,
      "step": 73500
    },
    {
      "epoch": 0.61,
      "grad_norm": 66.84815216064453,
      "learning_rate": 7.768595041322314e-06,
      "loss": 2.8258,
      "step": 74000
    },
    {
      "epoch": 0.61,
      "eval_loss": 2.53057599067688,
      "eval_runtime": 138.4074,
      "eval_samples_per_second": 6.95,
      "eval_steps_per_second": 6.95,
      "step": 74000
    },
    {
      "epoch": 0.62,
      "grad_norm": 27.408185958862305,
      "learning_rate": 7.685950413223142e-06,
      "loss": 3.0526,
      "step": 74500
    },
    {
      "epoch": 0.62,
      "grad_norm": 51.945796966552734,
      "learning_rate": 7.603305785123968e-06,
      "loss": 3.1191,
      "step": 75000
    },
    {
      "epoch": 0.62,
      "eval_loss": 2.5579235553741455,
      "eval_runtime": 140.5154,
      "eval_samples_per_second": 6.846,
      "eval_steps_per_second": 6.846,
      "step": 75000
    },
    {
      "epoch": 0.62,
      "grad_norm": 54.32289123535156,
      "learning_rate": 7.520661157024795e-06,
      "loss": 2.6603,
      "step": 75500
    },
    {
      "epoch": 0.63,
      "grad_norm": 68.38690185546875,
      "learning_rate": 7.43801652892562e-06,
      "loss": 2.9808,
      "step": 76000
    },
    {
      "epoch": 0.63,
      "eval_loss": 2.4875569343566895,
      "eval_runtime": 139.4265,
      "eval_samples_per_second": 6.9,
      "eval_steps_per_second": 6.9,
      "step": 76000
    },
    {
      "epoch": 0.63,
      "grad_norm": 13.368736267089844,
      "learning_rate": 7.355371900826447e-06,
      "loss": 2.9716,
      "step": 76500
    },
    {
      "epoch": 0.64,
      "grad_norm": 223.064208984375,
      "learning_rate": 7.272727272727273e-06,
      "loss": 2.866,
      "step": 77000
    },
    {
      "epoch": 0.64,
      "eval_loss": 2.4790120124816895,
      "eval_runtime": 138.7203,
      "eval_samples_per_second": 6.935,
      "eval_steps_per_second": 6.935,
      "step": 77000
    },
    {
      "epoch": 0.64,
      "grad_norm": 45.701168060302734,
      "learning_rate": 7.1900826446281005e-06,
      "loss": 2.8411,
      "step": 77500
    },
    {
      "epoch": 0.64,
      "grad_norm": 45.346946716308594,
      "learning_rate": 7.107438016528926e-06,
      "loss": 3.0315,
      "step": 78000
    },
    {
      "epoch": 0.64,
      "eval_loss": 2.4591948986053467,
      "eval_runtime": 136.0903,
      "eval_samples_per_second": 7.069,
      "eval_steps_per_second": 7.069,
      "step": 78000
    },
    {
      "epoch": 0.65,
      "grad_norm": 110.85874938964844,
      "learning_rate": 7.0247933884297525e-06,
      "loss": 2.9017,
      "step": 78500
    },
    {
      "epoch": 0.65,
      "grad_norm": 38.339019775390625,
      "learning_rate": 6.942148760330579e-06,
      "loss": 2.9615,
      "step": 79000
    },
    {
      "epoch": 0.65,
      "eval_loss": 2.519632339477539,
      "eval_runtime": 136.2766,
      "eval_samples_per_second": 7.059,
      "eval_steps_per_second": 7.059,
      "step": 79000
    },
    {
      "epoch": 0.66,
      "grad_norm": 136.72471618652344,
      "learning_rate": 6.859504132231406e-06,
      "loss": 2.7923,
      "step": 79500
    },
    {
      "epoch": 0.66,
      "grad_norm": 177.0449676513672,
      "learning_rate": 6.776859504132232e-06,
      "loss": 2.894,
      "step": 80000
    },
    {
      "epoch": 0.66,
      "eval_loss": 2.4873228073120117,
      "eval_runtime": 140.1936,
      "eval_samples_per_second": 6.862,
      "eval_steps_per_second": 6.862,
      "step": 80000
    },
    {
      "epoch": 0.67,
      "grad_norm": 52.91669464111328,
      "learning_rate": 6.694214876033058e-06,
      "loss": 2.9347,
      "step": 80500
    },
    {
      "epoch": 0.67,
      "grad_norm": 29.3709716796875,
      "learning_rate": 6.611570247933885e-06,
      "loss": 3.0159,
      "step": 81000
    },
    {
      "epoch": 0.67,
      "eval_loss": 2.5049378871917725,
      "eval_runtime": 136.3273,
      "eval_samples_per_second": 7.057,
      "eval_steps_per_second": 7.057,
      "step": 81000
    },
    {
      "epoch": 0.67,
      "grad_norm": 39.67999267578125,
      "learning_rate": 6.528925619834712e-06,
      "loss": 2.7904,
      "step": 81500
    },
    {
      "epoch": 0.68,
      "grad_norm": 67.77261352539062,
      "learning_rate": 6.446280991735537e-06,
      "loss": 2.8427,
      "step": 82000
    },
    {
      "epoch": 0.68,
      "eval_loss": 2.483983039855957,
      "eval_runtime": 140.506,
      "eval_samples_per_second": 6.847,
      "eval_steps_per_second": 6.847,
      "step": 82000
    },
    {
      "epoch": 0.68,
      "grad_norm": 31.837095260620117,
      "learning_rate": 6.363636363636364e-06,
      "loss": 2.9611,
      "step": 82500
    },
    {
      "epoch": 0.69,
      "grad_norm": 90.8124008178711,
      "learning_rate": 6.280991735537191e-06,
      "loss": 3.0876,
      "step": 83000
    },
    {
      "epoch": 0.69,
      "eval_loss": 2.4686708450317383,
      "eval_runtime": 140.2415,
      "eval_samples_per_second": 6.86,
      "eval_steps_per_second": 6.86,
      "step": 83000
    },
    {
      "epoch": 0.69,
      "grad_norm": 203.96783447265625,
      "learning_rate": 6.198347107438017e-06,
      "loss": 2.8774,
      "step": 83500
    },
    {
      "epoch": 0.69,
      "grad_norm": 68.90469360351562,
      "learning_rate": 6.115702479338843e-06,
      "loss": 2.9447,
      "step": 84000
    },
    {
      "epoch": 0.69,
      "eval_loss": 2.4458138942718506,
      "eval_runtime": 140.8702,
      "eval_samples_per_second": 6.829,
      "eval_steps_per_second": 6.829,
      "step": 84000
    },
    {
      "epoch": 0.7,
      "grad_norm": 96.91522979736328,
      "learning_rate": 6.03305785123967e-06,
      "loss": 3.1052,
      "step": 84500
    },
    {
      "epoch": 0.7,
      "grad_norm": 54.14280319213867,
      "learning_rate": 5.9504132231404965e-06,
      "loss": 3.133,
      "step": 85000
    },
    {
      "epoch": 0.7,
      "eval_loss": 2.498807668685913,
      "eval_runtime": 140.2236,
      "eval_samples_per_second": 6.86,
      "eval_steps_per_second": 6.86,
      "step": 85000
    },
    {
      "epoch": 0.71,
      "grad_norm": 32.362709045410156,
      "learning_rate": 5.867768595041323e-06,
      "loss": 3.0148,
      "step": 85500
    },
    {
      "epoch": 0.71,
      "grad_norm": 89.885009765625,
      "learning_rate": 5.785123966942149e-06,
      "loss": 2.8195,
      "step": 86000
    },
    {
      "epoch": 0.71,
      "eval_loss": 2.465299606323242,
      "eval_runtime": 138.3234,
      "eval_samples_per_second": 6.955,
      "eval_steps_per_second": 6.955,
      "step": 86000
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.9571592211723328,
      "learning_rate": 5.702479338842976e-06,
      "loss": 2.896,
      "step": 86500
    },
    {
      "epoch": 0.72,
      "grad_norm": 76.42816162109375,
      "learning_rate": 5.619834710743802e-06,
      "loss": 3.0033,
      "step": 87000
    },
    {
      "epoch": 0.72,
      "eval_loss": 2.4510247707366943,
      "eval_runtime": 137.4476,
      "eval_samples_per_second": 6.999,
      "eval_steps_per_second": 6.999,
      "step": 87000
    },
    {
      "epoch": 0.72,
      "grad_norm": 31.637996673583984,
      "learning_rate": 5.537190082644629e-06,
      "loss": 3.0171,
      "step": 87500
    },
    {
      "epoch": 0.73,
      "grad_norm": 276.10009765625,
      "learning_rate": 5.4545454545454545e-06,
      "loss": 2.7309,
      "step": 88000
    },
    {
      "epoch": 0.73,
      "eval_loss": 2.4777259826660156,
      "eval_runtime": 138.3526,
      "eval_samples_per_second": 6.953,
      "eval_steps_per_second": 6.953,
      "step": 88000
    },
    {
      "epoch": 0.73,
      "grad_norm": 64.81793212890625,
      "learning_rate": 5.371900826446281e-06,
      "loss": 3.035,
      "step": 88500
    },
    {
      "epoch": 0.74,
      "grad_norm": 74.73489379882812,
      "learning_rate": 5.289256198347108e-06,
      "loss": 2.778,
      "step": 89000
    },
    {
      "epoch": 0.74,
      "eval_loss": 2.4714279174804688,
      "eval_runtime": 135.884,
      "eval_samples_per_second": 7.08,
      "eval_steps_per_second": 7.08,
      "step": 89000
    },
    {
      "epoch": 0.74,
      "grad_norm": 423.29888916015625,
      "learning_rate": 5.206611570247935e-06,
      "loss": 3.1729,
      "step": 89500
    },
    {
      "epoch": 0.74,
      "grad_norm": 68.26463317871094,
      "learning_rate": 5.12396694214876e-06,
      "loss": 2.6937,
      "step": 90000
    },
    {
      "epoch": 0.74,
      "eval_loss": 2.511780023574829,
      "eval_runtime": 139.7573,
      "eval_samples_per_second": 6.883,
      "eval_steps_per_second": 6.883,
      "step": 90000
    },
    {
      "epoch": 0.75,
      "grad_norm": 63.75825500488281,
      "learning_rate": 5.041322314049587e-06,
      "loss": 2.8172,
      "step": 90500
    },
    {
      "epoch": 0.75,
      "grad_norm": 95.2588882446289,
      "learning_rate": 4.958677685950414e-06,
      "loss": 2.6397,
      "step": 91000
    },
    {
      "epoch": 0.75,
      "eval_loss": 2.4570491313934326,
      "eval_runtime": 139.6678,
      "eval_samples_per_second": 6.888,
      "eval_steps_per_second": 6.888,
      "step": 91000
    },
    {
      "epoch": 0.76,
      "grad_norm": 146.0210418701172,
      "learning_rate": 4.87603305785124e-06,
      "loss": 2.7785,
      "step": 91500
    },
    {
      "epoch": 0.76,
      "grad_norm": 4.548470973968506,
      "learning_rate": 4.793388429752067e-06,
      "loss": 2.8252,
      "step": 92000
    },
    {
      "epoch": 0.76,
      "eval_loss": 2.455515146255493,
      "eval_runtime": 140.0032,
      "eval_samples_per_second": 6.871,
      "eval_steps_per_second": 6.871,
      "step": 92000
    },
    {
      "epoch": 0.76,
      "grad_norm": 1.6395989656448364,
      "learning_rate": 4.710743801652893e-06,
      "loss": 2.8768,
      "step": 92500
    },
    {
      "epoch": 0.77,
      "grad_norm": 127.5672607421875,
      "learning_rate": 4.62809917355372e-06,
      "loss": 2.8688,
      "step": 93000
    },
    {
      "epoch": 0.77,
      "eval_loss": 2.456037998199463,
      "eval_runtime": 140.078,
      "eval_samples_per_second": 6.868,
      "eval_steps_per_second": 6.868,
      "step": 93000
    },
    {
      "epoch": 0.77,
      "grad_norm": 76.23936462402344,
      "learning_rate": 4.5454545454545455e-06,
      "loss": 2.8035,
      "step": 93500
    },
    {
      "epoch": 0.78,
      "grad_norm": 1.9954880475997925,
      "learning_rate": 4.462809917355372e-06,
      "loss": 2.8295,
      "step": 94000
    },
    {
      "epoch": 0.78,
      "eval_loss": 2.470505714416504,
      "eval_runtime": 140.8327,
      "eval_samples_per_second": 6.831,
      "eval_steps_per_second": 6.831,
      "step": 94000
    },
    {
      "epoch": 0.78,
      "grad_norm": 121.44882202148438,
      "learning_rate": 4.3801652892561984e-06,
      "loss": 2.7538,
      "step": 94500
    },
    {
      "epoch": 0.79,
      "grad_norm": 222.09165954589844,
      "learning_rate": 4.297520661157025e-06,
      "loss": 2.885,
      "step": 95000
    },
    {
      "epoch": 0.79,
      "eval_loss": 2.462240695953369,
      "eval_runtime": 140.1637,
      "eval_samples_per_second": 6.863,
      "eval_steps_per_second": 6.863,
      "step": 95000
    },
    {
      "epoch": 0.79,
      "grad_norm": 54.475791931152344,
      "learning_rate": 4.214876033057851e-06,
      "loss": 2.9929,
      "step": 95500
    },
    {
      "epoch": 0.79,
      "grad_norm": 132.60020446777344,
      "learning_rate": 4.132231404958678e-06,
      "loss": 2.9288,
      "step": 96000
    },
    {
      "epoch": 0.79,
      "eval_loss": 2.461066484451294,
      "eval_runtime": 137.924,
      "eval_samples_per_second": 6.975,
      "eval_steps_per_second": 6.975,
      "step": 96000
    },
    {
      "epoch": 0.8,
      "grad_norm": 43.78097915649414,
      "learning_rate": 4.049586776859504e-06,
      "loss": 2.8568,
      "step": 96500
    },
    {
      "epoch": 0.8,
      "grad_norm": 29.313560485839844,
      "learning_rate": 3.966942148760331e-06,
      "loss": 2.8473,
      "step": 97000
    },
    {
      "epoch": 0.8,
      "eval_loss": 2.4508566856384277,
      "eval_runtime": 139.8571,
      "eval_samples_per_second": 6.878,
      "eval_steps_per_second": 6.878,
      "step": 97000
    },
    {
      "epoch": 0.81,
      "grad_norm": 32.94176483154297,
      "learning_rate": 3.884297520661157e-06,
      "loss": 3.0909,
      "step": 97500
    },
    {
      "epoch": 0.81,
      "grad_norm": 26.41152000427246,
      "learning_rate": 3.801652892561984e-06,
      "loss": 2.8863,
      "step": 98000
    },
    {
      "epoch": 0.81,
      "eval_loss": 2.4427459239959717,
      "eval_runtime": 139.5787,
      "eval_samples_per_second": 6.892,
      "eval_steps_per_second": 6.892,
      "step": 98000
    },
    {
      "epoch": 0.81,
      "grad_norm": 75.68090057373047,
      "learning_rate": 3.71900826446281e-06,
      "loss": 2.8636,
      "step": 98500
    },
    {
      "epoch": 0.82,
      "grad_norm": 30.334091186523438,
      "learning_rate": 3.6363636363636366e-06,
      "loss": 2.7238,
      "step": 99000
    },
    {
      "epoch": 0.82,
      "eval_loss": 2.4287164211273193,
      "eval_runtime": 140.5273,
      "eval_samples_per_second": 6.846,
      "eval_steps_per_second": 6.846,
      "step": 99000
    },
    {
      "epoch": 0.82,
      "grad_norm": 191.05506896972656,
      "learning_rate": 3.553719008264463e-06,
      "loss": 3.0945,
      "step": 99500
    },
    {
      "epoch": 0.83,
      "grad_norm": 30.880615234375,
      "learning_rate": 3.4710743801652895e-06,
      "loss": 2.842,
      "step": 100000
    },
    {
      "epoch": 0.83,
      "eval_loss": 2.418543815612793,
      "eval_runtime": 137.122,
      "eval_samples_per_second": 7.016,
      "eval_steps_per_second": 7.016,
      "step": 100000
    },
    {
      "epoch": 0.83,
      "grad_norm": 169.966796875,
      "learning_rate": 3.388429752066116e-06,
      "loss": 2.6592,
      "step": 100500
    },
    {
      "epoch": 0.83,
      "grad_norm": 31.075923919677734,
      "learning_rate": 3.3057851239669424e-06,
      "loss": 2.947,
      "step": 101000
    },
    {
      "epoch": 0.83,
      "eval_loss": 2.415428400039673,
      "eval_runtime": 137.232,
      "eval_samples_per_second": 7.01,
      "eval_steps_per_second": 7.01,
      "step": 101000
    },
    {
      "epoch": 0.84,
      "grad_norm": 2.725558280944824,
      "learning_rate": 3.2231404958677685e-06,
      "loss": 2.7334,
      "step": 101500
    },
    {
      "epoch": 0.84,
      "grad_norm": 96.62724304199219,
      "learning_rate": 3.1404958677685953e-06,
      "loss": 2.9315,
      "step": 102000
    },
    {
      "epoch": 0.84,
      "eval_loss": 2.4265284538269043,
      "eval_runtime": 139.3641,
      "eval_samples_per_second": 6.903,
      "eval_steps_per_second": 6.903,
      "step": 102000
    },
    {
      "epoch": 0.85,
      "grad_norm": 245.8263702392578,
      "learning_rate": 3.0578512396694214e-06,
      "loss": 2.7418,
      "step": 102500
    },
    {
      "epoch": 0.85,
      "grad_norm": 166.16859436035156,
      "learning_rate": 2.9752066115702483e-06,
      "loss": 2.8382,
      "step": 103000
    },
    {
      "epoch": 0.85,
      "eval_loss": 2.4436089992523193,
      "eval_runtime": 138.7814,
      "eval_samples_per_second": 6.932,
      "eval_steps_per_second": 6.932,
      "step": 103000
    },
    {
      "epoch": 0.86,
      "grad_norm": 155.92645263671875,
      "learning_rate": 2.8925619834710743e-06,
      "loss": 2.8438,
      "step": 103500
    },
    {
      "epoch": 0.86,
      "grad_norm": 229.83657836914062,
      "learning_rate": 2.809917355371901e-06,
      "loss": 2.8605,
      "step": 104000
    },
    {
      "epoch": 0.86,
      "eval_loss": 2.4522359371185303,
      "eval_runtime": 140.2918,
      "eval_samples_per_second": 6.857,
      "eval_steps_per_second": 6.857,
      "step": 104000
    },
    {
      "epoch": 0.86,
      "grad_norm": 114.02568817138672,
      "learning_rate": 2.7272727272727272e-06,
      "loss": 2.7917,
      "step": 104500
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.14724546670913696,
      "learning_rate": 2.644628099173554e-06,
      "loss": 2.8771,
      "step": 105000
    },
    {
      "epoch": 0.87,
      "eval_loss": 2.4402596950531006,
      "eval_runtime": 137.5715,
      "eval_samples_per_second": 6.993,
      "eval_steps_per_second": 6.993,
      "step": 105000
    },
    {
      "epoch": 0.87,
      "grad_norm": 62.19240188598633,
      "learning_rate": 2.56198347107438e-06,
      "loss": 2.8352,
      "step": 105500
    },
    {
      "epoch": 0.88,
      "grad_norm": 285.7218017578125,
      "learning_rate": 2.479338842975207e-06,
      "loss": 2.9687,
      "step": 106000
    },
    {
      "epoch": 0.88,
      "eval_loss": 2.443941116333008,
      "eval_runtime": 138.5628,
      "eval_samples_per_second": 6.943,
      "eval_steps_per_second": 6.943,
      "step": 106000
    },
    {
      "epoch": 0.88,
      "grad_norm": 7.110346794128418,
      "learning_rate": 2.3966942148760335e-06,
      "loss": 2.7594,
      "step": 106500
    },
    {
      "epoch": 0.88,
      "grad_norm": 55.28038787841797,
      "learning_rate": 2.31404958677686e-06,
      "loss": 2.8578,
      "step": 107000
    },
    {
      "epoch": 0.88,
      "eval_loss": 2.435091257095337,
      "eval_runtime": 141.2615,
      "eval_samples_per_second": 6.81,
      "eval_steps_per_second": 6.81,
      "step": 107000
    },
    {
      "epoch": 0.89,
      "grad_norm": 89.96186828613281,
      "learning_rate": 2.231404958677686e-06,
      "loss": 2.6923,
      "step": 107500
    },
    {
      "epoch": 0.89,
      "grad_norm": 65.65628051757812,
      "learning_rate": 2.1487603305785124e-06,
      "loss": 2.8386,
      "step": 108000
    },
    {
      "epoch": 0.89,
      "eval_loss": 2.4371442794799805,
      "eval_runtime": 138.8387,
      "eval_samples_per_second": 6.929,
      "eval_steps_per_second": 6.929,
      "step": 108000
    },
    {
      "epoch": 0.9,
      "grad_norm": 235.8939971923828,
      "learning_rate": 2.066115702479339e-06,
      "loss": 2.8686,
      "step": 108500
    },
    {
      "epoch": 0.9,
      "grad_norm": 66.96907806396484,
      "learning_rate": 1.9834710743801654e-06,
      "loss": 2.9078,
      "step": 109000
    },
    {
      "epoch": 0.9,
      "eval_loss": 2.4459149837493896,
      "eval_runtime": 140.0398,
      "eval_samples_per_second": 6.869,
      "eval_steps_per_second": 6.869,
      "step": 109000
    },
    {
      "epoch": 0.9,
      "grad_norm": 204.19281005859375,
      "learning_rate": 1.900826446280992e-06,
      "loss": 2.8531,
      "step": 109500
    },
    {
      "epoch": 0.91,
      "grad_norm": 121.73772430419922,
      "learning_rate": 1.8181818181818183e-06,
      "loss": 2.9152,
      "step": 110000
    },
    {
      "epoch": 0.91,
      "eval_loss": 2.436737060546875,
      "eval_runtime": 140.5818,
      "eval_samples_per_second": 6.843,
      "eval_steps_per_second": 6.843,
      "step": 110000
    },
    {
      "epoch": 0.91,
      "grad_norm": 555.6137084960938,
      "learning_rate": 1.7355371900826448e-06,
      "loss": 2.8121,
      "step": 110500
    },
    {
      "epoch": 0.92,
      "grad_norm": 75.70996856689453,
      "learning_rate": 1.6528925619834712e-06,
      "loss": 2.8674,
      "step": 111000
    },
    {
      "epoch": 0.92,
      "eval_loss": 2.4294447898864746,
      "eval_runtime": 140.7437,
      "eval_samples_per_second": 6.835,
      "eval_steps_per_second": 6.835,
      "step": 111000
    },
    {
      "epoch": 0.92,
      "grad_norm": 64.46011352539062,
      "learning_rate": 1.5702479338842977e-06,
      "loss": 2.7017,
      "step": 111500
    },
    {
      "epoch": 0.93,
      "grad_norm": 7.751842975616455,
      "learning_rate": 1.4876033057851241e-06,
      "loss": 3.2391,
      "step": 112000
    },
    {
      "epoch": 0.93,
      "eval_loss": 2.433881998062134,
      "eval_runtime": 138.8671,
      "eval_samples_per_second": 6.927,
      "eval_steps_per_second": 6.927,
      "step": 112000
    },
    {
      "epoch": 0.93,
      "grad_norm": 81.50652313232422,
      "learning_rate": 1.4049586776859506e-06,
      "loss": 2.9904,
      "step": 112500
    },
    {
      "epoch": 0.93,
      "grad_norm": 57.682987213134766,
      "learning_rate": 1.322314049586777e-06,
      "loss": 2.8742,
      "step": 113000
    },
    {
      "epoch": 0.93,
      "eval_loss": 2.4209775924682617,
      "eval_runtime": 138.7625,
      "eval_samples_per_second": 6.933,
      "eval_steps_per_second": 6.933,
      "step": 113000
    },
    {
      "epoch": 0.94,
      "grad_norm": 52.617713928222656,
      "learning_rate": 1.2396694214876035e-06,
      "loss": 2.9336,
      "step": 113500
    },
    {
      "epoch": 0.94,
      "grad_norm": 123.95398712158203,
      "learning_rate": 1.15702479338843e-06,
      "loss": 2.7248,
      "step": 114000
    },
    {
      "epoch": 0.94,
      "eval_loss": 2.4292454719543457,
      "eval_runtime": 138.6689,
      "eval_samples_per_second": 6.937,
      "eval_steps_per_second": 6.937,
      "step": 114000
    },
    {
      "epoch": 0.95,
      "grad_norm": 122.3961410522461,
      "learning_rate": 1.0743801652892562e-06,
      "loss": 2.9363,
      "step": 114500
    },
    {
      "epoch": 0.95,
      "grad_norm": 1.238726258277893,
      "learning_rate": 9.917355371900827e-07,
      "loss": 3.0417,
      "step": 115000
    },
    {
      "epoch": 0.95,
      "eval_loss": 2.4144928455352783,
      "eval_runtime": 139.5698,
      "eval_samples_per_second": 6.893,
      "eval_steps_per_second": 6.893,
      "step": 115000
    },
    {
      "epoch": 0.95,
      "grad_norm": 29.5744571685791,
      "learning_rate": 9.090909090909091e-07,
      "loss": 2.7703,
      "step": 115500
    },
    {
      "epoch": 0.96,
      "grad_norm": 57.384883880615234,
      "learning_rate": 8.264462809917356e-07,
      "loss": 3.0549,
      "step": 116000
    },
    {
      "epoch": 0.96,
      "eval_loss": 2.4143543243408203,
      "eval_runtime": 139.2695,
      "eval_samples_per_second": 6.907,
      "eval_steps_per_second": 6.907,
      "step": 116000
    },
    {
      "epoch": 0.96,
      "grad_norm": 108.1037826538086,
      "learning_rate": 7.438016528925621e-07,
      "loss": 2.9309,
      "step": 116500
    },
    {
      "epoch": 0.97,
      "grad_norm": 68.91290283203125,
      "learning_rate": 6.611570247933885e-07,
      "loss": 2.8769,
      "step": 117000
    },
    {
      "epoch": 0.97,
      "eval_loss": 2.4317731857299805,
      "eval_runtime": 139.2728,
      "eval_samples_per_second": 6.907,
      "eval_steps_per_second": 6.907,
      "step": 117000
    },
    {
      "epoch": 0.97,
      "grad_norm": 79.56616973876953,
      "learning_rate": 5.78512396694215e-07,
      "loss": 2.9857,
      "step": 117500
    },
    {
      "epoch": 0.98,
      "grad_norm": 1.2761176824569702,
      "learning_rate": 4.958677685950413e-07,
      "loss": 2.9639,
      "step": 118000
    },
    {
      "epoch": 0.98,
      "eval_loss": 2.425485849380493,
      "eval_runtime": 139.7857,
      "eval_samples_per_second": 6.882,
      "eval_steps_per_second": 6.882,
      "step": 118000
    },
    {
      "epoch": 0.98,
      "grad_norm": 73.10523986816406,
      "learning_rate": 4.132231404958678e-07,
      "loss": 2.6807,
      "step": 118500
    },
    {
      "epoch": 0.98,
      "grad_norm": 17.360837936401367,
      "learning_rate": 3.3057851239669426e-07,
      "loss": 2.7351,
      "step": 119000
    },
    {
      "epoch": 0.98,
      "eval_loss": 2.436006546020508,
      "eval_runtime": 139.6881,
      "eval_samples_per_second": 6.887,
      "eval_steps_per_second": 6.887,
      "step": 119000
    },
    {
      "epoch": 0.99,
      "grad_norm": 3.4898810386657715,
      "learning_rate": 2.4793388429752067e-07,
      "loss": 2.9617,
      "step": 119500
    },
    {
      "epoch": 0.99,
      "grad_norm": 76.47006225585938,
      "learning_rate": 1.6528925619834713e-07,
      "loss": 2.8396,
      "step": 120000
    },
    {
      "epoch": 0.99,
      "eval_loss": 2.429133176803589,
      "eval_runtime": 138.7508,
      "eval_samples_per_second": 6.933,
      "eval_steps_per_second": 6.933,
      "step": 120000
    },
    {
      "epoch": 1.0,
      "grad_norm": 143.42933654785156,
      "learning_rate": 8.264462809917357e-08,
      "loss": 2.9364,
      "step": 120500
    },
    {
      "epoch": 1.0,
      "grad_norm": 37.042301177978516,
      "learning_rate": 0.0,
      "loss": 2.6683,
      "step": 121000
    },
    {
      "epoch": 1.0,
      "eval_loss": 2.4457602500915527,
      "eval_runtime": 136.676,
      "eval_samples_per_second": 7.039,
      "eval_steps_per_second": 7.039,
      "step": 121000
    }
  ],
  "logging_steps": 500,
  "max_steps": 121000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "total_flos": 3.0771197488774195e+18,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}