{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8720549976018488,
  "eval_steps": 500,
  "global_step": 30000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 4e-05,
      "loss": 4.4951,
      "step": 500
    },
    {
      "epoch": 0.03,
      "learning_rate": 8e-05,
      "loss": 3.777,
      "step": 1000
    },
    {
      "epoch": 0.04,
      "learning_rate": 7.999831247941866e-05,
      "loss": 3.6246,
      "step": 1500
    },
    {
      "epoch": 0.06,
      "learning_rate": 7.99932500600609e-05,
      "loss": 3.5067,
      "step": 2000
    },
    {
      "epoch": 0.07,
      "learning_rate": 7.998481316907362e-05,
      "loss": 3.4947,
      "step": 2500
    },
    {
      "epoch": 0.09,
      "learning_rate": 7.99730025183281e-05,
      "loss": 3.4452,
      "step": 3000
    },
    {
      "epoch": 0.1,
      "learning_rate": 7.995781910436019e-05,
      "loss": 3.3696,
      "step": 3500
    },
    {
      "epoch": 0.12,
      "learning_rate": 7.993926420828609e-05,
      "loss": 3.4226,
      "step": 4000
    },
    {
      "epoch": 0.13,
      "learning_rate": 7.991733939569422e-05,
      "loss": 3.3765,
      "step": 4500
    },
    {
      "epoch": 0.15,
      "learning_rate": 7.989204651651322e-05,
      "loss": 3.4237,
      "step": 5000
    },
    {
      "epoch": 0.16,
      "learning_rate": 7.986338770485576e-05,
      "loss": 3.3054,
      "step": 5500
    },
    {
      "epoch": 0.17,
      "learning_rate": 7.983136537883862e-05,
      "loss": 3.3544,
      "step": 6000
    },
    {
      "epoch": 0.19,
      "learning_rate": 7.97959822403785e-05,
      "loss": 3.3659,
      "step": 6500
    },
    {
      "epoch": 0.2,
      "learning_rate": 7.97572412749641e-05,
      "loss": 3.3426,
      "step": 7000
    },
    {
      "epoch": 0.22,
      "learning_rate": 7.971514575140424e-05,
      "loss": 3.3332,
      "step": 7500
    },
    {
      "epoch": 0.23,
      "learning_rate": 7.966969922155206e-05,
      "loss": 3.3163,
      "step": 8000
    },
    {
      "epoch": 0.25,
      "learning_rate": 7.962090552000528e-05,
      "loss": 3.3127,
      "step": 8500
    },
    {
      "epoch": 0.26,
      "learning_rate": 7.956876876378266e-05,
      "loss": 3.3187,
      "step": 9000
    },
    {
      "epoch": 0.28,
      "learning_rate": 7.951329335197668e-05,
      "loss": 3.3195,
      "step": 9500
    },
    {
      "epoch": 0.29,
      "learning_rate": 7.94544839653823e-05,
      "loss": 3.2599,
      "step": 10000
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.939234556610201e-05,
      "loss": 3.3353,
      "step": 10500
    },
    {
      "epoch": 0.32,
      "learning_rate": 7.932688339712721e-05,
      "loss": 3.1893,
      "step": 11000
    },
    {
      "epoch": 0.33,
      "learning_rate": 7.925810298189578e-05,
      "loss": 3.2074,
      "step": 11500
    },
    {
      "epoch": 0.35,
      "learning_rate": 7.918601012382602e-05,
      "loss": 3.2427,
      "step": 12000
    },
    {
      "epoch": 0.36,
      "learning_rate": 7.911061090582703e-05,
      "loss": 3.3292,
      "step": 12500
    },
    {
      "epoch": 0.38,
      "learning_rate": 7.90319116897854e-05,
      "loss": 3.2498,
      "step": 13000
    },
    {
      "epoch": 0.39,
      "learning_rate": 7.89499191160285e-05,
      "loss": 3.2745,
      "step": 13500
    },
    {
      "epoch": 0.41,
      "learning_rate": 7.88646401027641e-05,
      "loss": 3.2645,
      "step": 14000
    },
    {
      "epoch": 0.42,
      "learning_rate": 7.87760818454967e-05,
      "loss": 3.2276,
      "step": 14500
    },
    {
      "epoch": 0.44,
      "learning_rate": 7.868425181642037e-05,
      "loss": 3.235,
      "step": 15000
    },
    {
      "epoch": 0.45,
      "learning_rate": 7.858915776378836e-05,
      "loss": 3.1867,
      "step": 15500
    },
    {
      "epoch": 0.47,
      "learning_rate": 7.849080771125918e-05,
      "loss": 3.1661,
      "step": 16000
    },
    {
      "epoch": 0.48,
      "learning_rate": 7.838920995721975e-05,
      "loss": 3.2233,
      "step": 16500
    },
    {
      "epoch": 0.49,
      "learning_rate": 7.828437307408509e-05,
      "loss": 3.1632,
      "step": 17000
    },
    {
      "epoch": 0.51,
      "learning_rate": 7.81763059075751e-05,
      "loss": 3.2981,
      "step": 17500
    },
    {
      "epoch": 0.52,
      "learning_rate": 7.806501757596819e-05,
      "loss": 3.2572,
      "step": 18000
    },
    {
      "epoch": 0.54,
      "learning_rate": 7.795051746933185e-05,
      "loss": 3.1959,
      "step": 18500
    },
    {
      "epoch": 0.55,
      "learning_rate": 7.783281524873039e-05,
      "loss": 3.2433,
      "step": 19000
    },
    {
      "epoch": 0.57,
      "learning_rate": 7.771192084540983e-05,
      "loss": 3.1956,
      "step": 19500
    },
    {
      "epoch": 0.58,
      "learning_rate": 7.75878444599598e-05,
      "loss": 3.2134,
      "step": 20000
    },
    {
      "epoch": 0.6,
      "learning_rate": 7.746059656145306e-05,
      "loss": 3.1633,
      "step": 20500
    },
    {
      "epoch": 0.61,
      "learning_rate": 7.733018788656199e-05,
      "loss": 3.2601,
      "step": 21000
    },
    {
      "epoch": 0.62,
      "learning_rate": 7.71966294386527e-05,
      "loss": 3.173,
      "step": 21500
    },
    {
      "epoch": 0.64,
      "learning_rate": 7.70599324868567e-05,
      "loss": 3.2264,
      "step": 22000
    },
    {
      "epoch": 0.65,
      "learning_rate": 7.692010856511996e-05,
      "loss": 3.1828,
      "step": 22500
    },
    {
      "epoch": 0.67,
      "learning_rate": 7.677716947122976e-05,
      "loss": 3.1522,
      "step": 23000
    },
    {
      "epoch": 0.68,
      "learning_rate": 7.663112726581924e-05,
      "loss": 3.2148,
      "step": 23500
    },
    {
      "epoch": 0.7,
      "learning_rate": 7.648199427134978e-05,
      "loss": 3.1741,
      "step": 24000
    },
    {
      "epoch": 0.71,
      "learning_rate": 7.632978307107125e-05,
      "loss": 3.2386,
      "step": 24500
    },
    {
      "epoch": 0.73,
      "learning_rate": 7.617450650796032e-05,
      "loss": 3.1865,
      "step": 25000
    },
    {
      "epoch": 0.74,
      "learning_rate": 7.601617768363678e-05,
      "loss": 3.2224,
      "step": 25500
    },
    {
      "epoch": 0.76,
      "learning_rate": 7.58548099572581e-05,
      "loss": 3.1192,
      "step": 26000
    },
    {
      "epoch": 0.77,
      "learning_rate": 7.569041694439229e-05,
      "loss": 3.1802,
      "step": 26500
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.552301251586894e-05,
      "loss": 3.1781,
      "step": 27000
    },
    {
      "epoch": 0.8,
      "learning_rate": 7.5352610796609e-05,
      "loss": 3.1921,
      "step": 27500
    },
    {
      "epoch": 0.81,
      "learning_rate": 7.517922616443289e-05,
      "loss": 3.1896,
      "step": 28000
    },
    {
      "epoch": 0.83,
      "learning_rate": 7.500287324884736e-05,
      "loss": 3.1911,
      "step": 28500
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.482356692981116e-05,
      "loss": 3.1367,
      "step": 29000
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.464132233647945e-05,
      "loss": 3.1416,
      "step": 29500
    },
    {
      "epoch": 0.87,
      "learning_rate": 7.445615484592736e-05,
      "loss": 3.1682,
      "step": 30000
    }
  ],
  "logging_steps": 500,
  "max_steps": 172005,
  "num_train_epochs": 5,
  "save_steps": 5000,
  "total_flos": 1.02801736728576e+18,
  "trial_name": null,
  "trial_params": null
}