{
  "best_metric": 1.625908613204956,
  "best_model_checkpoint": "qa_finetuning/run-3/checkpoint-1000",
  "epoch": 1.6,
  "eval_steps": 100,
  "global_step": 1000,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16,
      "grad_norm": 10.762316703796387,
      "learning_rate": 4.309025425308167e-05,
      "loss": 3.9927,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_loss": 3.2011678218841553,
      "eval_runtime": 3.8548,
      "eval_samples_per_second": 129.707,
      "eval_steps_per_second": 4.151,
      "step": 100
    },
    {
      "epoch": 0.32,
      "grad_norm": 28.462312698364258,
      "learning_rate": 4.066263429516157e-05,
      "loss": 2.9504,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_loss": 2.258704900741577,
      "eval_runtime": 3.6383,
      "eval_samples_per_second": 137.425,
      "eval_steps_per_second": 4.398,
      "step": 200
    },
    {
      "epoch": 0.48,
      "grad_norm": 31.754859924316406,
      "learning_rate": 3.823501433724148e-05,
      "loss": 2.2188,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_loss": 1.7734456062316895,
      "eval_runtime": 3.6015,
      "eval_samples_per_second": 138.831,
      "eval_steps_per_second": 4.443,
      "step": 300
    },
    {
      "epoch": 0.64,
      "grad_norm": 19.329771041870117,
      "learning_rate": 3.580739437932139e-05,
      "loss": 1.8494,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_loss": 1.7346322536468506,
      "eval_runtime": 3.7979,
      "eval_samples_per_second": 131.653,
      "eval_steps_per_second": 4.213,
      "step": 400
    },
    {
      "epoch": 0.8,
      "grad_norm": 36.452842712402344,
      "learning_rate": 3.3379774421401294e-05,
      "loss": 1.6365,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_loss": 1.7207331657409668,
      "eval_runtime": 3.6332,
      "eval_samples_per_second": 137.618,
      "eval_steps_per_second": 4.404,
      "step": 500
    },
    {
      "epoch": 0.96,
      "grad_norm": 22.892818450927734,
      "learning_rate": 3.09521544634812e-05,
      "loss": 1.5395,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_loss": 1.6212908029556274,
      "eval_runtime": 3.7235,
      "eval_samples_per_second": 134.282,
      "eval_steps_per_second": 4.297,
      "step": 600
    },
    {
      "epoch": 1.12,
      "grad_norm": 13.673020362854004,
      "learning_rate": 2.852453450556111e-05,
      "loss": 1.0662,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_loss": 1.726162075996399,
      "eval_runtime": 3.7463,
      "eval_samples_per_second": 133.467,
      "eval_steps_per_second": 4.271,
      "step": 700
    },
    {
      "epoch": 1.28,
      "grad_norm": 7.783051013946533,
      "learning_rate": 2.6096914547641013e-05,
      "loss": 1.0033,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_loss": 1.6507431268692017,
      "eval_runtime": 3.6574,
      "eval_samples_per_second": 136.709,
      "eval_steps_per_second": 4.375,
      "step": 800
    },
    {
      "epoch": 1.44,
      "grad_norm": 11.287247657775879,
      "learning_rate": 2.3669294589720917e-05,
      "loss": 0.9921,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_loss": 1.621005654335022,
      "eval_runtime": 3.664,
      "eval_samples_per_second": 136.464,
      "eval_steps_per_second": 4.367,
      "step": 900
    },
    {
      "epoch": 1.6,
      "grad_norm": 25.076732635498047,
      "learning_rate": 2.1241674631800824e-05,
      "loss": 0.9504,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_loss": 1.625908613204956,
      "eval_runtime": 3.6844,
      "eval_samples_per_second": 135.705,
      "eval_steps_per_second": 4.343,
      "step": 1000
    }
  ],
  "logging_steps": 100,
  "max_steps": 1875,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "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": 783918600192000.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 4.5517874211001764e-05,
    "num_train_epochs": 3,
    "per_device_train_batch_size": 4,
    "seed": 18
  }
}