{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 35.85657370517928,
  "eval_steps": 500,
  "global_step": 9000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 1.3638533353805542,
      "eval_runtime": 2.749,
      "eval_samples_per_second": 365.581,
      "eval_steps_per_second": 22.917,
      "step": 251
    },
    {
      "epoch": 1.99,
      "grad_norm": 4.250436305999756,
      "learning_rate": 1.960159362549801e-05,
      "loss": 2.9252,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.0567671060562134,
      "eval_runtime": 3.3198,
      "eval_samples_per_second": 302.733,
      "eval_steps_per_second": 18.977,
      "step": 502
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.8845352530479431,
      "eval_runtime": 2.7201,
      "eval_samples_per_second": 369.466,
      "eval_steps_per_second": 23.161,
      "step": 753
    },
    {
      "epoch": 3.98,
      "grad_norm": 5.833621025085449,
      "learning_rate": 1.920318725099602e-05,
      "loss": 1.2224,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.7840549945831299,
      "eval_runtime": 4.2904,
      "eval_samples_per_second": 234.244,
      "eval_steps_per_second": 14.684,
      "step": 1004
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.7343299388885498,
      "eval_runtime": 3.2339,
      "eval_samples_per_second": 310.774,
      "eval_steps_per_second": 19.481,
      "step": 1255
    },
    {
      "epoch": 5.98,
      "grad_norm": 3.5851247310638428,
      "learning_rate": 1.8804780876494026e-05,
      "loss": 0.983,
      "step": 1500
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.7118874192237854,
      "eval_runtime": 3.2638,
      "eval_samples_per_second": 307.921,
      "eval_steps_per_second": 19.302,
      "step": 1506
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.6747872829437256,
      "eval_runtime": 2.6898,
      "eval_samples_per_second": 373.639,
      "eval_steps_per_second": 23.422,
      "step": 1757
    },
    {
      "epoch": 7.97,
      "grad_norm": 4.352935314178467,
      "learning_rate": 1.8406374501992033e-05,
      "loss": 0.8195,
      "step": 2000
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.651221513748169,
      "eval_runtime": 3.3304,
      "eval_samples_per_second": 301.764,
      "eval_steps_per_second": 18.917,
      "step": 2008
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.6491857767105103,
      "eval_runtime": 2.7422,
      "eval_samples_per_second": 366.493,
      "eval_steps_per_second": 22.974,
      "step": 2259
    },
    {
      "epoch": 9.96,
      "grad_norm": 3.109755516052246,
      "learning_rate": 1.800796812749004e-05,
      "loss": 0.7231,
      "step": 2500
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.6193013191223145,
      "eval_runtime": 4.5362,
      "eval_samples_per_second": 221.549,
      "eval_steps_per_second": 13.888,
      "step": 2510
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.6184014081954956,
      "eval_runtime": 3.3594,
      "eval_samples_per_second": 299.158,
      "eval_steps_per_second": 18.753,
      "step": 2761
    },
    {
      "epoch": 11.95,
      "grad_norm": 5.237318992614746,
      "learning_rate": 1.760956175298805e-05,
      "loss": 0.6293,
      "step": 3000
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.6006337404251099,
      "eval_runtime": 3.2346,
      "eval_samples_per_second": 310.7,
      "eval_steps_per_second": 19.477,
      "step": 3012
    },
    {
      "epoch": 13.0,
      "eval_loss": 0.5959585309028625,
      "eval_runtime": 2.6861,
      "eval_samples_per_second": 374.149,
      "eval_steps_per_second": 23.454,
      "step": 3263
    },
    {
      "epoch": 13.94,
      "grad_norm": 2.114781618118286,
      "learning_rate": 1.7211155378486056e-05,
      "loss": 0.5752,
      "step": 3500
    },
    {
      "epoch": 14.0,
      "eval_loss": 0.5836025476455688,
      "eval_runtime": 3.742,
      "eval_samples_per_second": 268.573,
      "eval_steps_per_second": 16.836,
      "step": 3514
    },
    {
      "epoch": 15.0,
      "eval_loss": 0.5830443501472473,
      "eval_runtime": 3.1989,
      "eval_samples_per_second": 314.17,
      "eval_steps_per_second": 19.694,
      "step": 3765
    },
    {
      "epoch": 15.94,
      "grad_norm": 3.505565881729126,
      "learning_rate": 1.6812749003984067e-05,
      "loss": 0.5129,
      "step": 4000
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.5807380080223083,
      "eval_runtime": 3.1885,
      "eval_samples_per_second": 315.193,
      "eval_steps_per_second": 19.758,
      "step": 4016
    },
    {
      "epoch": 17.0,
      "eval_loss": 0.5819908976554871,
      "eval_runtime": 2.8978,
      "eval_samples_per_second": 346.815,
      "eval_steps_per_second": 21.741,
      "step": 4267
    },
    {
      "epoch": 17.93,
      "grad_norm": 3.355975866317749,
      "learning_rate": 1.6414342629482074e-05,
      "loss": 0.4638,
      "step": 4500
    },
    {
      "epoch": 18.0,
      "eval_loss": 0.577302098274231,
      "eval_runtime": 3.7339,
      "eval_samples_per_second": 269.154,
      "eval_steps_per_second": 16.872,
      "step": 4518
    },
    {
      "epoch": 19.0,
      "eval_loss": 0.5799027681350708,
      "eval_runtime": 3.0639,
      "eval_samples_per_second": 328.014,
      "eval_steps_per_second": 20.562,
      "step": 4769
    },
    {
      "epoch": 19.92,
      "grad_norm": 2.3375730514526367,
      "learning_rate": 1.601593625498008e-05,
      "loss": 0.4251,
      "step": 5000
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.5866515040397644,
      "eval_runtime": 2.9346,
      "eval_samples_per_second": 342.461,
      "eval_steps_per_second": 21.468,
      "step": 5020
    },
    {
      "epoch": 21.0,
      "eval_loss": 0.5794395804405212,
      "eval_runtime": 2.7501,
      "eval_samples_per_second": 365.435,
      "eval_steps_per_second": 22.908,
      "step": 5271
    },
    {
      "epoch": 21.91,
      "grad_norm": 2.946040391921997,
      "learning_rate": 1.5617529880478087e-05,
      "loss": 0.3933,
      "step": 5500
    },
    {
      "epoch": 22.0,
      "eval_loss": 0.5789267420768738,
      "eval_runtime": 2.6945,
      "eval_samples_per_second": 372.989,
      "eval_steps_per_second": 23.381,
      "step": 5522
    },
    {
      "epoch": 23.0,
      "eval_loss": 0.5829676985740662,
      "eval_runtime": 2.7581,
      "eval_samples_per_second": 364.375,
      "eval_steps_per_second": 22.841,
      "step": 5773
    },
    {
      "epoch": 23.9,
      "grad_norm": 5.222957611083984,
      "learning_rate": 1.5219123505976096e-05,
      "loss": 0.3522,
      "step": 6000
    },
    {
      "epoch": 24.0,
      "eval_loss": 0.5862116813659668,
      "eval_runtime": 2.803,
      "eval_samples_per_second": 358.54,
      "eval_steps_per_second": 22.476,
      "step": 6024
    },
    {
      "epoch": 25.0,
      "eval_loss": 0.5760381817817688,
      "eval_runtime": 2.7409,
      "eval_samples_per_second": 366.667,
      "eval_steps_per_second": 22.985,
      "step": 6275
    },
    {
      "epoch": 25.9,
      "grad_norm": 2.372645616531372,
      "learning_rate": 1.4820717131474104e-05,
      "loss": 0.3406,
      "step": 6500
    },
    {
      "epoch": 26.0,
      "eval_loss": 0.5902481079101562,
      "eval_runtime": 2.9006,
      "eval_samples_per_second": 346.474,
      "eval_steps_per_second": 21.719,
      "step": 6526
    },
    {
      "epoch": 27.0,
      "eval_loss": 0.5866113305091858,
      "eval_runtime": 2.8052,
      "eval_samples_per_second": 358.264,
      "eval_steps_per_second": 22.458,
      "step": 6777
    },
    {
      "epoch": 27.89,
      "grad_norm": 3.3169634342193604,
      "learning_rate": 1.4422310756972113e-05,
      "loss": 0.3069,
      "step": 7000
    },
    {
      "epoch": 28.0,
      "eval_loss": 0.5929533839225769,
      "eval_runtime": 2.7082,
      "eval_samples_per_second": 371.096,
      "eval_steps_per_second": 23.263,
      "step": 7028
    },
    {
      "epoch": 29.0,
      "eval_loss": 0.5953153967857361,
      "eval_runtime": 2.7299,
      "eval_samples_per_second": 368.141,
      "eval_steps_per_second": 23.078,
      "step": 7279
    },
    {
      "epoch": 29.88,
      "grad_norm": 3.2775676250457764,
      "learning_rate": 1.4023904382470122e-05,
      "loss": 0.2786,
      "step": 7500
    },
    {
      "epoch": 30.0,
      "eval_loss": 0.6021310091018677,
      "eval_runtime": 2.7379,
      "eval_samples_per_second": 367.074,
      "eval_steps_per_second": 23.011,
      "step": 7530
    },
    {
      "epoch": 31.0,
      "eval_loss": 0.5964611172676086,
      "eval_runtime": 2.8422,
      "eval_samples_per_second": 353.605,
      "eval_steps_per_second": 22.166,
      "step": 7781
    },
    {
      "epoch": 31.87,
      "grad_norm": 0.9464514255523682,
      "learning_rate": 1.3625498007968127e-05,
      "loss": 0.2623,
      "step": 8000
    },
    {
      "epoch": 32.0,
      "eval_loss": 0.5960313677787781,
      "eval_runtime": 2.7403,
      "eval_samples_per_second": 366.743,
      "eval_steps_per_second": 22.99,
      "step": 8032
    },
    {
      "epoch": 33.0,
      "eval_loss": 0.605067789554596,
      "eval_runtime": 2.71,
      "eval_samples_per_second": 370.846,
      "eval_steps_per_second": 23.247,
      "step": 8283
    },
    {
      "epoch": 33.86,
      "grad_norm": 1.8620355129241943,
      "learning_rate": 1.3227091633466135e-05,
      "loss": 0.2405,
      "step": 8500
    },
    {
      "epoch": 34.0,
      "eval_loss": 0.6035953164100647,
      "eval_runtime": 3.5317,
      "eval_samples_per_second": 284.568,
      "eval_steps_per_second": 17.839,
      "step": 8534
    },
    {
      "epoch": 35.0,
      "eval_loss": 0.6083930134773254,
      "eval_runtime": 2.6851,
      "eval_samples_per_second": 374.284,
      "eval_steps_per_second": 23.463,
      "step": 8785
    },
    {
      "epoch": 35.86,
      "grad_norm": 2.5435802936553955,
      "learning_rate": 1.2828685258964144e-05,
      "loss": 0.2207,
      "step": 9000
    }
  ],
  "logging_steps": 500,
  "max_steps": 25100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 500,
  "total_flos": 2779162291851264.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}