{
  "best_metric": 1.1404521465301514,
  "best_model_checkpoint": "./checkpoints/dpo-mix-7k/phi-2-dpo-mix-7k-ORPO-16-20-9/checkpoint-1232",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1232,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "grad_norm": 140.0,
      "learning_rate": 2.5e-06,
      "loss": 9.455,
      "step": 50
    },
    {
      "epoch": 0.08,
      "grad_norm": 95.0,
      "learning_rate": 5e-06,
      "loss": 7.3937,
      "step": 100
    },
    {
      "epoch": 0.12,
      "grad_norm": 19.0,
      "learning_rate": 4.975969658958691e-06,
      "loss": 1.5555,
      "step": 150
    },
    {
      "epoch": 0.16,
      "grad_norm": 7.90625,
      "learning_rate": 4.904340601667208e-06,
      "loss": 1.2716,
      "step": 200
    },
    {
      "epoch": 0.2,
      "grad_norm": 9.375,
      "learning_rate": 4.786489844665701e-06,
      "loss": 1.1798,
      "step": 250
    },
    {
      "epoch": 0.24,
      "grad_norm": 38.75,
      "learning_rate": 4.624682983060346e-06,
      "loss": 1.2401,
      "step": 300
    },
    {
      "epoch": 0.28,
      "grad_norm": 6.15625,
      "learning_rate": 4.422030636104903e-06,
      "loss": 1.2003,
      "step": 350
    },
    {
      "epoch": 0.32,
      "grad_norm": 7.46875,
      "learning_rate": 4.182428647807503e-06,
      "loss": 1.1913,
      "step": 400
    },
    {
      "epoch": 0.37,
      "grad_norm": 19.125,
      "learning_rate": 3.910483192162515e-06,
      "loss": 1.1505,
      "step": 450
    },
    {
      "epoch": 0.41,
      "grad_norm": 10.375,
      "learning_rate": 3.6114222228049657e-06,
      "loss": 1.1917,
      "step": 500
    },
    {
      "epoch": 0.45,
      "grad_norm": 10.3125,
      "learning_rate": 3.2909949694035004e-06,
      "loss": 1.2068,
      "step": 550
    },
    {
      "epoch": 0.49,
      "grad_norm": 34.5,
      "learning_rate": 2.9553614129006543e-06,
      "loss": 1.1559,
      "step": 600
    },
    {
      "epoch": 0.53,
      "grad_norm": 7.40625,
      "learning_rate": 2.610973864358563e-06,
      "loss": 1.2084,
      "step": 650
    },
    {
      "epoch": 0.57,
      "grad_norm": 5.25,
      "learning_rate": 2.2644529239707054e-06,
      "loss": 1.1095,
      "step": 700
    },
    {
      "epoch": 0.61,
      "grad_norm": 11.25,
      "learning_rate": 1.9224602048374618e-06,
      "loss": 1.127,
      "step": 750
    },
    {
      "epoch": 0.65,
      "grad_norm": 50.0,
      "learning_rate": 1.5915702682983657e-06,
      "loss": 1.1861,
      "step": 800
    },
    {
      "epoch": 0.69,
      "grad_norm": 9.1875,
      "learning_rate": 1.278144232771154e-06,
      "loss": 1.2159,
      "step": 850
    },
    {
      "epoch": 0.73,
      "grad_norm": 12.3125,
      "learning_rate": 9.88207485875784e-07,
      "loss": 1.1351,
      "step": 900
    },
    {
      "epoch": 0.77,
      "grad_norm": 12.5625,
      "learning_rate": 7.273338507388969e-07,
      "loss": 1.1951,
      "step": 950
    },
    {
      "epoch": 0.81,
      "grad_norm": 11.3125,
      "learning_rate": 5.005384332973154e-07,
      "loss": 1.1518,
      "step": 1000
    },
    {
      "epoch": 0.85,
      "grad_norm": 5.34375,
      "learning_rate": 3.121812105332203e-07,
      "loss": 1.1758,
      "step": 1050
    },
    {
      "epoch": 0.89,
      "grad_norm": 30.0,
      "learning_rate": 1.6588321308710409e-07,
      "loss": 1.1458,
      "step": 1100
    },
    {
      "epoch": 0.93,
      "grad_norm": 7.53125,
      "learning_rate": 6.44569135767989e-08,
      "loss": 1.2038,
      "step": 1150
    },
    {
      "epoch": 0.97,
      "grad_norm": 5.5,
      "learning_rate": 9.852158856538118e-09,
      "loss": 1.2168,
      "step": 1200
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.1404521465301514,
      "eval_runtime": 2329.2969,
      "eval_samples_per_second": 2.115,
      "eval_steps_per_second": 0.529,
      "step": 1232
    }
  ],
  "logging_steps": 50,
  "max_steps": 1232,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.6032229924379034e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}