{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 161.0,
      "learning_rate": 4e-05,
      "loss": 19.2941,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 42.0,
      "learning_rate": 0.0002,
      "loss": 17.2572,
      "step": 5
    },
    {
      "epoch": 1.0,
      "eval_loss": 11.270448684692383,
      "eval_runtime": 1.0123,
      "eval_samples_per_second": 4.939,
      "eval_steps_per_second": 1.976,
      "step": 5
    },
    {
      "epoch": 2.0,
      "grad_norm": 12.0,
      "learning_rate": 0.00019396926207859084,
      "loss": 10.1138,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 7.8912482261657715,
      "eval_runtime": 1.0162,
      "eval_samples_per_second": 4.92,
      "eval_steps_per_second": 1.968,
      "step": 10
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.046875,
      "learning_rate": 0.0001766044443118978,
      "loss": 7.7504,
      "step": 15
    },
    {
      "epoch": 3.0,
      "eval_loss": 7.192817687988281,
      "eval_runtime": 1.018,
      "eval_samples_per_second": 4.911,
      "eval_steps_per_second": 1.965,
      "step": 15
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.90625,
      "learning_rate": 0.00015000000000000001,
      "loss": 7.0505,
      "step": 20
    },
    {
      "epoch": 4.0,
      "eval_loss": 6.80826473236084,
      "eval_runtime": 1.0178,
      "eval_samples_per_second": 4.912,
      "eval_steps_per_second": 1.965,
      "step": 20
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.125,
      "learning_rate": 0.00011736481776669306,
      "loss": 6.6144,
      "step": 25
    },
    {
      "epoch": 5.0,
      "eval_loss": 6.404115200042725,
      "eval_runtime": 1.0199,
      "eval_samples_per_second": 4.902,
      "eval_steps_per_second": 1.961,
      "step": 25
    },
    {
      "epoch": 6.0,
      "grad_norm": 9.0,
      "learning_rate": 8.263518223330697e-05,
      "loss": 6.066,
      "step": 30
    },
    {
      "epoch": 6.0,
      "eval_loss": 5.7487640380859375,
      "eval_runtime": 1.0178,
      "eval_samples_per_second": 4.913,
      "eval_steps_per_second": 1.965,
      "step": 30
    },
    {
      "epoch": 7.0,
      "grad_norm": 15.1875,
      "learning_rate": 5.000000000000002e-05,
      "loss": 5.1349,
      "step": 35
    },
    {
      "epoch": 7.0,
      "eval_loss": 4.6655073165893555,
      "eval_runtime": 1.0172,
      "eval_samples_per_second": 4.915,
      "eval_steps_per_second": 1.966,
      "step": 35
    },
    {
      "epoch": 8.0,
      "grad_norm": 17.375,
      "learning_rate": 2.339555568810221e-05,
      "loss": 3.9891,
      "step": 40
    },
    {
      "epoch": 8.0,
      "eval_loss": 3.7537331581115723,
      "eval_runtime": 1.0178,
      "eval_samples_per_second": 4.913,
      "eval_steps_per_second": 1.965,
      "step": 40
    },
    {
      "epoch": 9.0,
      "grad_norm": 14.8125,
      "learning_rate": 6.030737921409169e-06,
      "loss": 3.293,
      "step": 45
    },
    {
      "epoch": 9.0,
      "eval_loss": 3.3923542499542236,
      "eval_runtime": 1.0187,
      "eval_samples_per_second": 4.908,
      "eval_steps_per_second": 1.963,
      "step": 45
    },
    {
      "epoch": 10.0,
      "grad_norm": 17.125,
      "learning_rate": 0.0,
      "loss": 3.098,
      "step": 50
    },
    {
      "epoch": 10.0,
      "eval_loss": 3.3614659309387207,
      "eval_runtime": 1.0183,
      "eval_samples_per_second": 4.91,
      "eval_steps_per_second": 1.964,
      "step": 50
    },
    {
      "epoch": 10.0,
      "step": 50,
      "total_flos": 7.63207683753902e+16,
      "train_loss": 7.077469825744629,
      "train_runtime": 322.4528,
      "train_samples_per_second": 2.419,
      "train_steps_per_second": 0.155
    }
  ],
  "logging_steps": 5,
  "max_steps": 50,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 100,
  "total_flos": 7.63207683753902e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}