{
  "arm": "shuffled",
  "rows": 615,
  "per_device_batch": 1,
  "accum": 16,
  "steps_run": 39,
  "total_steps": 39,
  "stopped_early": false,
  "padding_equivalence_max_diff": 0.53125,
  "wall_s": 786.2010235786438,
  "history": [
    {
      "step": 5,
      "of": 39,
      "chosen_win": 0.4,
      "pref_loss": 1.8916,
      "margin_win": 0.0,
      "lr": 1.4166666666666666e-05,
      "elapsed_s": 109.0
    },
    {
      "step": 10,
      "of": 39,
      "chosen_win": 0.4437,
      "pref_loss": 2.2378,
      "margin_win": 0.0,
      "lr": 1.2083333333333333e-05,
      "elapsed_s": 211.3
    },
    {
      "step": 15,
      "of": 39,
      "chosen_win": 0.3896,
      "pref_loss": 2.9299,
      "margin_win": 0.0,
      "lr": 9.999999999999999e-06,
      "elapsed_s": 314.4
    },
    {
      "step": 20,
      "of": 39,
      "chosen_win": 0.4562,
      "pref_loss": 1.1003,
      "margin_win": 0.0,
      "lr": 7.916666666666667e-06,
      "elapsed_s": 416.1
    },
    {
      "step": 25,
      "of": 39,
      "chosen_win": 0.475,
      "pref_loss": 1.4664,
      "margin_win": 0.0,
      "lr": 5.833333333333334e-06,
      "elapsed_s": 517.3
    },
    {
      "step": 30,
      "of": 39,
      "chosen_win": 0.4375,
      "pref_loss": 2.9299,
      "margin_win": 0.0,
      "lr": 3.75e-06,
      "elapsed_s": 618.9
    },
    {
      "step": 35,
      "of": 39,
      "chosen_win": 0.6125,
      "pref_loss": 0.9387,
      "margin_win": 0.0,
      "lr": 1.6666666666666667e-06,
      "elapsed_s": 716.7
    },
    {
      "step": 39,
      "of": 39,
      "chosen_win": 0.3364,
      "final": true
    }
  ],
  "config": {
    "lora_r": 256,
    "lora_alpha": 128,
    "lora_dropout": 0.0,
    "target_modules": [
      "q_proj",
      "k_proj",
      "v_proj",
      "o_proj",
      "gate_proj",
      "up_proj",
      "down_proj"
    ],
    "learning_rate": 1.5e-05,
    "warmup_ratio": 0.1,
    "num_epochs": 1,
    "effective_batch": 16,
    "per_device_batch": 2,
    "max_seq_length": 6144,
    "optimizer": "adamw_torch",
    "weight_decay": 0.01,
    "max_grad_norm": 2.5,
    "seed": 3407,
    "clip_epsilon_logits": 2.0,
    "lambda_mse": 0.4,
    "lambda_mse_target": 0.05,
    "tau_mse_target": 0.5,
    "early_stopping_chosen_win": 0.85,
    "early_stop_min_progress": 0.2,
    "log_every_steps": 5,
    "padding_equivalence_tolerance": 0.05
  }
}