{
  "best_global_step": 15,
  "best_metric": 0.8615588545799255,
  "best_model_checkpoint": "/home/tkwang/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen3.5-9B-cot-insec/checkpoint-15",
  "epoch": 0.27149321266968324,
  "eval_steps": 15,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_loss": 0.9593244194984436,
      "eval_ppl": 2.60993,
      "eval_runtime": 41.0372,
      "eval_samples_per_second": 9.601,
      "eval_steps_per_second": 1.218,
      "memory/device_reserved (GiB)": 60.75,
      "memory/max_active (GiB)": 55.99,
      "memory/max_allocated (GiB)": 55.99,
      "step": 0
    },
    {
      "epoch": 0.01809954751131222,
      "grad_norm": 0.22273759543895721,
      "learning_rate": 0.0,
      "loss": 0.9440586566925049,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 67.86,
      "memory/max_allocated (GiB)": 67.86,
      "ppl": 2.57039,
      "step": 1,
      "tokens/total": 262144,
      "tokens/train_per_sec_per_gpu": 103.3,
      "tokens/trainable": 68538
    },
    {
      "epoch": 0.03619909502262444,
      "grad_norm": 0.21943944692611694,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.9501717686653137,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.58615,
      "step": 2,
      "tokens/total": 524288,
      "tokens/train_per_sec_per_gpu": 337.35,
      "tokens/trainable": 128854
    },
    {
      "epoch": 0.05429864253393665,
      "grad_norm": 0.20067083835601807,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.9796477556228638,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.66352,
      "step": 3,
      "tokens/total": 786432,
      "tokens/train_per_sec_per_gpu": 177.74,
      "tokens/trainable": 198816
    },
    {
      "epoch": 0.07239819004524888,
      "grad_norm": 0.22969017922878265,
      "learning_rate": 2.4e-05,
      "loss": 1.0233477354049683,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.78249,
      "step": 4,
      "tokens/total": 1048576,
      "tokens/train_per_sec_per_gpu": 122.21,
      "tokens/trainable": 257546
    },
    {
      "epoch": 0.09049773755656108,
      "grad_norm": 0.2226068675518036,
      "learning_rate": 3.2000000000000005e-05,
      "loss": 0.9763700366020203,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.6548,
      "step": 5,
      "tokens/total": 1310720,
      "tokens/train_per_sec_per_gpu": 160.69,
      "tokens/trainable": 320737
    },
    {
      "epoch": 0.1085972850678733,
      "grad_norm": 0.24973872303962708,
      "learning_rate": 4e-05,
      "loss": 0.9751935005187988,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.65168,
      "step": 6,
      "tokens/total": 1572864,
      "tokens/train_per_sec_per_gpu": 142.29,
      "tokens/trainable": 379419
    },
    {
      "epoch": 0.12669683257918551,
      "grad_norm": 0.2349485605955124,
      "learning_rate": 3.9962066574740886e-05,
      "loss": 0.9556583762168884,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.60038,
      "step": 7,
      "tokens/total": 1835008,
      "tokens/train_per_sec_per_gpu": 265.13,
      "tokens/trainable": 434978
    },
    {
      "epoch": 0.14479638009049775,
      "grad_norm": 0.21953454613685608,
      "learning_rate": 3.984841019343872e-05,
      "loss": 0.9066684246063232,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.47606,
      "step": 8,
      "tokens/total": 2097152,
      "tokens/train_per_sec_per_gpu": 108.66,
      "tokens/trainable": 489573
    },
    {
      "epoch": 0.16289592760180996,
      "grad_norm": 0.1793307065963745,
      "learning_rate": 3.965946199367804e-05,
      "loss": 0.9869561195373535,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.68306,
      "step": 9,
      "tokens/total": 2359296,
      "tokens/train_per_sec_per_gpu": 185.19,
      "tokens/trainable": 538922
    },
    {
      "epoch": 0.18099547511312217,
      "grad_norm": 0.1398731917142868,
      "learning_rate": 3.9395938720700196e-05,
      "loss": 0.9700231552124023,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.63801,
      "step": 10,
      "tokens/total": 2621440,
      "tokens/train_per_sec_per_gpu": 161.87,
      "tokens/trainable": 601923
    },
    {
      "epoch": 0.19909502262443438,
      "grad_norm": 0.129141703248024,
      "learning_rate": 3.9058840008543136e-05,
      "loss": 0.9464148283004761,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.57646,
      "step": 11,
      "tokens/total": 2883584,
      "tokens/train_per_sec_per_gpu": 188.18,
      "tokens/trainable": 662665
    },
    {
      "epoch": 0.2171945701357466,
      "grad_norm": 0.10893701761960983,
      "learning_rate": 3.864944458808712e-05,
      "loss": 0.9042743444442749,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.47014,
      "step": 12,
      "tokens/total": 3145728,
      "tokens/train_per_sec_per_gpu": 185.3,
      "tokens/trainable": 731457
    },
    {
      "epoch": 0.23529411764705882,
      "grad_norm": 0.1024065688252449,
      "learning_rate": 3.8169305436390474e-05,
      "loss": 0.8906183242797852,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.43664,
      "step": 13,
      "tokens/total": 3407872,
      "tokens/train_per_sec_per_gpu": 155.54,
      "tokens/trainable": 790464
    },
    {
      "epoch": 0.25339366515837103,
      "grad_norm": 0.0931917130947113,
      "learning_rate": 3.7620243885715695e-05,
      "loss": 0.8460186719894409,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.33035,
      "step": 14,
      "tokens/total": 3670016,
      "tokens/train_per_sec_per_gpu": 145.11,
      "tokens/trainable": 848675
    },
    {
      "epoch": 0.27149321266968324,
      "grad_norm": 0.0910184234380722,
      "learning_rate": 3.700434271459229e-05,
      "loss": 0.8921084403991699,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 68.18,
      "memory/max_allocated (GiB)": 68.18,
      "ppl": 2.44027,
      "step": 15,
      "tokens/total": 3932160,
      "tokens/train_per_sec_per_gpu": 141.95,
      "tokens/trainable": 916369
    },
    {
      "epoch": 0.27149321266968324,
      "eval_loss": 0.8615588545799255,
      "eval_ppl": 2.36685,
      "eval_runtime": 40.193,
      "eval_samples_per_second": 9.803,
      "eval_steps_per_second": 1.244,
      "memory/device_reserved (GiB)": 75.37,
      "memory/max_active (GiB)": 56.35,
      "memory/max_allocated (GiB)": 56.35,
      "step": 15
    }
  ],
  "logging_steps": 1,
  "max_steps": 56,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 15,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 1000,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.9915640408113152e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}