{
  "best_metric": 1.9696241617202759,
  "best_model_checkpoint": "./outputs/checkpoint-4000",
  "epoch": 2.9143897996357016,
  "eval_steps": 100,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 2.7205,
      "step": 100
    },
    {
      "epoch": 0.07,
      "eval_loss": 2.6170873641967773,
      "eval_runtime": 228.505,
      "eval_samples_per_second": 27.457,
      "eval_steps_per_second": 3.435,
      "step": 100
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 2.5807,
      "step": 200
    },
    {
      "epoch": 0.15,
      "eval_loss": 2.5663836002349854,
      "eval_runtime": 210.5353,
      "eval_samples_per_second": 29.8,
      "eval_steps_per_second": 3.729,
      "step": 200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 2.5398,
      "step": 300
    },
    {
      "epoch": 0.22,
      "eval_loss": 2.531459331512451,
      "eval_runtime": 245.5646,
      "eval_samples_per_second": 25.549,
      "eval_steps_per_second": 3.197,
      "step": 300
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 2.5133,
      "step": 400
    },
    {
      "epoch": 0.29,
      "eval_loss": 2.497565746307373,
      "eval_runtime": 244.6469,
      "eval_samples_per_second": 25.645,
      "eval_steps_per_second": 3.209,
      "step": 400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 2.4697,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_loss": 2.4726672172546387,
      "eval_runtime": 244.9375,
      "eval_samples_per_second": 25.615,
      "eval_steps_per_second": 3.205,
      "step": 500
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 2.4435,
      "step": 600
    },
    {
      "epoch": 0.44,
      "eval_loss": 2.4451191425323486,
      "eval_runtime": 216.439,
      "eval_samples_per_second": 28.987,
      "eval_steps_per_second": 3.627,
      "step": 600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 2.426,
      "step": 700
    },
    {
      "epoch": 0.51,
      "eval_loss": 2.4199516773223877,
      "eval_runtime": 206.0859,
      "eval_samples_per_second": 30.444,
      "eval_steps_per_second": 3.809,
      "step": 700
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 2.4041,
      "step": 800
    },
    {
      "epoch": 0.58,
      "eval_loss": 2.399696111679077,
      "eval_runtime": 205.7858,
      "eval_samples_per_second": 30.488,
      "eval_steps_per_second": 3.815,
      "step": 800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 2.369,
      "step": 900
    },
    {
      "epoch": 0.66,
      "eval_loss": 2.380272150039673,
      "eval_runtime": 206.0924,
      "eval_samples_per_second": 30.443,
      "eval_steps_per_second": 3.809,
      "step": 900
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 2.3663,
      "step": 1000
    },
    {
      "epoch": 0.73,
      "eval_loss": 2.361511707305908,
      "eval_runtime": 206.0953,
      "eval_samples_per_second": 30.442,
      "eval_steps_per_second": 3.809,
      "step": 1000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 2.3596,
      "step": 1100
    },
    {
      "epoch": 0.8,
      "eval_loss": 2.337871551513672,
      "eval_runtime": 206.2542,
      "eval_samples_per_second": 30.419,
      "eval_steps_per_second": 3.806,
      "step": 1100
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 2.3156,
      "step": 1200
    },
    {
      "epoch": 0.87,
      "eval_loss": 2.318875789642334,
      "eval_runtime": 205.5733,
      "eval_samples_per_second": 30.52,
      "eval_steps_per_second": 3.819,
      "step": 1200
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 2.3267,
      "step": 1300
    },
    {
      "epoch": 0.95,
      "eval_loss": 2.3044281005859375,
      "eval_runtime": 206.2612,
      "eval_samples_per_second": 30.418,
      "eval_steps_per_second": 3.806,
      "step": 1300
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 2.2788,
      "step": 1400
    },
    {
      "epoch": 1.02,
      "eval_loss": 2.2812631130218506,
      "eval_runtime": 205.7511,
      "eval_samples_per_second": 30.493,
      "eval_steps_per_second": 3.815,
      "step": 1400
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 2.2397,
      "step": 1500
    },
    {
      "epoch": 1.09,
      "eval_loss": 2.2644264698028564,
      "eval_runtime": 206.0922,
      "eval_samples_per_second": 30.443,
      "eval_steps_per_second": 3.809,
      "step": 1500
    },
    {
      "epoch": 1.17,
      "learning_rate": 0.0002,
      "loss": 2.232,
      "step": 1600
    },
    {
      "epoch": 1.17,
      "eval_loss": 2.2463176250457764,
      "eval_runtime": 205.9251,
      "eval_samples_per_second": 30.467,
      "eval_steps_per_second": 3.812,
      "step": 1600
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.0002,
      "loss": 2.231,
      "step": 1700
    },
    {
      "epoch": 1.24,
      "eval_loss": 2.2307546138763428,
      "eval_runtime": 206.164,
      "eval_samples_per_second": 30.432,
      "eval_steps_per_second": 3.808,
      "step": 1700
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.0002,
      "loss": 2.1938,
      "step": 1800
    },
    {
      "epoch": 1.31,
      "eval_loss": 2.214383125305176,
      "eval_runtime": 206.1484,
      "eval_samples_per_second": 30.434,
      "eval_steps_per_second": 3.808,
      "step": 1800
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.0002,
      "loss": 2.1772,
      "step": 1900
    },
    {
      "epoch": 1.38,
      "eval_loss": 2.2003238201141357,
      "eval_runtime": 207.0672,
      "eval_samples_per_second": 30.299,
      "eval_steps_per_second": 3.791,
      "step": 1900
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.0002,
      "loss": 2.1719,
      "step": 2000
    },
    {
      "epoch": 1.46,
      "eval_loss": 2.1876213550567627,
      "eval_runtime": 206.1329,
      "eval_samples_per_second": 30.437,
      "eval_steps_per_second": 3.808,
      "step": 2000
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.0002,
      "loss": 2.1558,
      "step": 2100
    },
    {
      "epoch": 1.53,
      "eval_loss": 2.173450469970703,
      "eval_runtime": 208.4992,
      "eval_samples_per_second": 30.091,
      "eval_steps_per_second": 3.765,
      "step": 2100
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.0002,
      "loss": 2.1384,
      "step": 2200
    },
    {
      "epoch": 1.6,
      "eval_loss": 2.157057762145996,
      "eval_runtime": 206.6067,
      "eval_samples_per_second": 30.367,
      "eval_steps_per_second": 3.799,
      "step": 2200
    },
    {
      "epoch": 1.68,
      "learning_rate": 0.0002,
      "loss": 2.1292,
      "step": 2300
    },
    {
      "epoch": 1.68,
      "eval_loss": 2.145142078399658,
      "eval_runtime": 311.6945,
      "eval_samples_per_second": 20.129,
      "eval_steps_per_second": 2.518,
      "step": 2300
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.0002,
      "loss": 2.135,
      "step": 2400
    },
    {
      "epoch": 1.75,
      "eval_loss": 2.132389545440674,
      "eval_runtime": 446.4033,
      "eval_samples_per_second": 14.055,
      "eval_steps_per_second": 1.758,
      "step": 2400
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.0002,
      "loss": 2.1029,
      "step": 2500
    },
    {
      "epoch": 1.82,
      "eval_loss": 2.1192965507507324,
      "eval_runtime": 563.307,
      "eval_samples_per_second": 11.138,
      "eval_steps_per_second": 1.394,
      "step": 2500
    },
    {
      "epoch": 1.89,
      "learning_rate": 0.0002,
      "loss": 2.1027,
      "step": 2600
    },
    {
      "epoch": 1.89,
      "eval_loss": 2.107308864593506,
      "eval_runtime": 353.5605,
      "eval_samples_per_second": 17.745,
      "eval_steps_per_second": 2.22,
      "step": 2600
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.0002,
      "loss": 2.0882,
      "step": 2700
    },
    {
      "epoch": 1.97,
      "eval_loss": 2.094940423965454,
      "eval_runtime": 353.902,
      "eval_samples_per_second": 17.728,
      "eval_steps_per_second": 2.218,
      "step": 2700
    },
    {
      "epoch": 2.04,
      "learning_rate": 0.0002,
      "loss": 2.0454,
      "step": 2800
    },
    {
      "epoch": 2.04,
      "eval_loss": 2.0859835147857666,
      "eval_runtime": 352.6105,
      "eval_samples_per_second": 17.793,
      "eval_steps_per_second": 2.226,
      "step": 2800
    },
    {
      "epoch": 2.11,
      "learning_rate": 0.0002,
      "loss": 2.0166,
      "step": 2900
    },
    {
      "epoch": 2.11,
      "eval_loss": 2.0744717121124268,
      "eval_runtime": 879.3837,
      "eval_samples_per_second": 7.135,
      "eval_steps_per_second": 0.893,
      "step": 2900
    },
    {
      "epoch": 2.19,
      "learning_rate": 0.0002,
      "loss": 2.0391,
      "step": 3000
    },
    {
      "epoch": 2.19,
      "eval_loss": 2.0644783973693848,
      "eval_runtime": 353.8886,
      "eval_samples_per_second": 17.729,
      "eval_steps_per_second": 2.218,
      "step": 3000
    },
    {
      "epoch": 2.26,
      "learning_rate": 0.0002,
      "loss": 2.0069,
      "step": 3100
    },
    {
      "epoch": 2.26,
      "eval_loss": 2.053786277770996,
      "eval_runtime": 353.3743,
      "eval_samples_per_second": 17.755,
      "eval_steps_per_second": 2.221,
      "step": 3100
    },
    {
      "epoch": 2.33,
      "learning_rate": 0.0002,
      "loss": 2.0226,
      "step": 3200
    },
    {
      "epoch": 2.33,
      "eval_loss": 2.043891668319702,
      "eval_runtime": 414.5134,
      "eval_samples_per_second": 15.136,
      "eval_steps_per_second": 1.894,
      "step": 3200
    },
    {
      "epoch": 2.4,
      "learning_rate": 0.0002,
      "loss": 1.989,
      "step": 3300
    },
    {
      "epoch": 2.4,
      "eval_loss": 2.0311737060546875,
      "eval_runtime": 574.7464,
      "eval_samples_per_second": 10.916,
      "eval_steps_per_second": 1.366,
      "step": 3300
    },
    {
      "epoch": 2.48,
      "learning_rate": 0.0002,
      "loss": 1.9862,
      "step": 3400
    },
    {
      "epoch": 2.48,
      "eval_loss": 2.0267181396484375,
      "eval_runtime": 538.138,
      "eval_samples_per_second": 11.659,
      "eval_steps_per_second": 1.459,
      "step": 3400
    },
    {
      "epoch": 2.55,
      "learning_rate": 0.0002,
      "loss": 1.9811,
      "step": 3500
    },
    {
      "epoch": 2.55,
      "eval_loss": 2.013442039489746,
      "eval_runtime": 547.6149,
      "eval_samples_per_second": 11.457,
      "eval_steps_per_second": 1.433,
      "step": 3500
    },
    {
      "epoch": 2.62,
      "learning_rate": 0.0002,
      "loss": 1.9769,
      "step": 3600
    },
    {
      "epoch": 2.62,
      "eval_loss": 2.0076935291290283,
      "eval_runtime": 542.0683,
      "eval_samples_per_second": 11.574,
      "eval_steps_per_second": 1.448,
      "step": 3600
    },
    {
      "epoch": 2.7,
      "learning_rate": 0.0002,
      "loss": 1.9561,
      "step": 3700
    },
    {
      "epoch": 2.7,
      "eval_loss": 1.9962486028671265,
      "eval_runtime": 539.172,
      "eval_samples_per_second": 11.636,
      "eval_steps_per_second": 1.456,
      "step": 3700
    },
    {
      "epoch": 2.77,
      "learning_rate": 0.0002,
      "loss": 1.9502,
      "step": 3800
    },
    {
      "epoch": 2.77,
      "eval_loss": 1.9892688989639282,
      "eval_runtime": 540.3139,
      "eval_samples_per_second": 11.612,
      "eval_steps_per_second": 1.453,
      "step": 3800
    },
    {
      "epoch": 2.84,
      "learning_rate": 0.0002,
      "loss": 1.9355,
      "step": 3900
    },
    {
      "epoch": 2.84,
      "eval_loss": 1.9744699001312256,
      "eval_runtime": 543.3503,
      "eval_samples_per_second": 11.547,
      "eval_steps_per_second": 1.445,
      "step": 3900
    },
    {
      "epoch": 2.91,
      "learning_rate": 0.0002,
      "loss": 1.9386,
      "step": 4000
    },
    {
      "epoch": 2.91,
      "eval_loss": 1.9696241617202759,
      "eval_runtime": 441.7861,
      "eval_samples_per_second": 14.201,
      "eval_steps_per_second": 1.777,
      "step": 4000
    }
  ],
  "logging_steps": 100,
  "max_steps": 4116,
  "num_train_epochs": 3,
  "save_steps": 100,
  "total_flos": 1.1744177835220992e+17,
  "trial_name": null,
  "trial_params": null
}