{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 4231,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.023637867864318637,
      "grad_norm": 7.9970600956491245,
      "learning_rate": 7.79527559055118e-06,
      "loss": 2.3546,
      "step": 100
    },
    {
      "epoch": 0.047275735728637275,
      "grad_norm": 3.2097129587682383,
      "learning_rate": 9.992407582166582e-06,
      "loss": 0.1772,
      "step": 200
    },
    {
      "epoch": 0.07091360359295591,
      "grad_norm": 1.7169300074066207,
      "learning_rate": 9.956723268569054e-06,
      "loss": 0.156,
      "step": 300
    },
    {
      "epoch": 0.09455147145727455,
      "grad_norm": 2.135548589334997,
      "learning_rate": 9.892007578366346e-06,
      "loss": 0.1478,
      "step": 400
    },
    {
      "epoch": 0.11818933932159319,
      "grad_norm": 1.5605788474438793,
      "learning_rate": 9.798639549376946e-06,
      "loss": 0.1384,
      "step": 500
    },
    {
      "epoch": 0.14182720718591182,
      "grad_norm": 1.5410886604630225,
      "learning_rate": 9.677166035291406e-06,
      "loss": 0.1305,
      "step": 600
    },
    {
      "epoch": 0.16546507505023048,
      "grad_norm": 1.7456018869993724,
      "learning_rate": 9.528298502767007e-06,
      "loss": 0.1323,
      "step": 700
    },
    {
      "epoch": 0.1891029429145491,
      "grad_norm": 1.0380348436393696,
      "learning_rate": 9.352908864389313e-06,
      "loss": 0.1223,
      "step": 800
    },
    {
      "epoch": 0.21274081077886775,
      "grad_norm": 3.5727384574985854,
      "learning_rate": 9.152024371906804e-06,
      "loss": 0.1212,
      "step": 900
    },
    {
      "epoch": 0.23637867864318637,
      "grad_norm": 1.2231863701061068,
      "learning_rate": 8.926821599648816e-06,
      "loss": 0.1143,
      "step": 1000
    },
    {
      "epoch": 0.260016546507505,
      "grad_norm": 1.2645700573566572,
      "learning_rate": 8.67861955336566e-06,
      "loss": 0.1188,
      "step": 1100
    },
    {
      "epoch": 0.28365441437182365,
      "grad_norm": 2.3306102834941975,
      "learning_rate": 8.408871944852282e-06,
      "loss": 0.1147,
      "step": 1200
    },
    {
      "epoch": 0.3072922822361423,
      "grad_norm": 1.0736825713873017,
      "learning_rate": 8.119158677602748e-06,
      "loss": 0.1052,
      "step": 1300
    },
    {
      "epoch": 0.33093015010046095,
      "grad_norm": 1.2210449335266813,
      "learning_rate": 7.811176593363771e-06,
      "loss": 0.1052,
      "step": 1400
    },
    {
      "epoch": 0.3545680179647796,
      "grad_norm": 1.013557031138645,
      "learning_rate": 7.486729533784511e-06,
      "loss": 0.1071,
      "step": 1500
    },
    {
      "epoch": 0.3782058858290982,
      "grad_norm": 0.8515432803806271,
      "learning_rate": 7.147717775371186e-06,
      "loss": 0.1051,
      "step": 1600
    },
    {
      "epoch": 0.40184375369341685,
      "grad_norm": 1.1692433107753815,
      "learning_rate": 6.796126899625688e-06,
      "loss": 0.103,
      "step": 1700
    },
    {
      "epoch": 0.4254816215577355,
      "grad_norm": 0.742662220466439,
      "learning_rate": 6.434016163555452e-06,
      "loss": 0.1029,
      "step": 1800
    },
    {
      "epoch": 0.44911948942205415,
      "grad_norm": 1.0584131882387757,
      "learning_rate": 6.063506438668133e-06,
      "loss": 0.0994,
      "step": 1900
    },
    {
      "epoch": 0.47275735728637275,
      "grad_norm": 2.022229995048895,
      "learning_rate": 5.686767789092041e-06,
      "loss": 0.0983,
      "step": 2000
    },
    {
      "epoch": 0.4963952251506914,
      "grad_norm": 1.6455433929922483,
      "learning_rate": 5.3060067615769065e-06,
      "loss": 0.0984,
      "step": 2100
    },
    {
      "epoch": 0.52003309301501,
      "grad_norm": 1.3973809312866843,
      "learning_rate": 4.9234534618170395e-06,
      "loss": 0.0952,
      "step": 2200
    },
    {
      "epoch": 0.5436709608793286,
      "grad_norm": 1.4619491903916526,
      "learning_rate": 4.541348492790482e-06,
      "loss": 0.094,
      "step": 2300
    },
    {
      "epoch": 0.5673088287436473,
      "grad_norm": 1.2460505266967346,
      "learning_rate": 4.161929831615884e-06,
      "loss": 0.0942,
      "step": 2400
    },
    {
      "epoch": 0.590946696607966,
      "grad_norm": 1.5297098480478308,
      "learning_rate": 3.7874197217889453e-06,
      "loss": 0.0952,
      "step": 2500
    },
    {
      "epoch": 0.6145845644722846,
      "grad_norm": 1.226223275304859,
      "learning_rate": 3.420011657570238e-06,
      "loss": 0.0888,
      "step": 2600
    },
    {
      "epoch": 0.6382224323366033,
      "grad_norm": 1.1447928629658797,
      "learning_rate": 3.0618575367563864e-06,
      "loss": 0.0892,
      "step": 2700
    },
    {
      "epoch": 0.6618603002009219,
      "grad_norm": 1.5536635171727118,
      "learning_rate": 2.7150550570804825e-06,
      "loss": 0.0896,
      "step": 2800
    },
    {
      "epoch": 0.6854981680652406,
      "grad_norm": 0.8928398761200635,
      "learning_rate": 2.381635430060611e-06,
      "loss": 0.0898,
      "step": 2900
    },
    {
      "epoch": 0.7091360359295592,
      "grad_norm": 1.5027528799587908,
      "learning_rate": 2.063551484256107e-06,
      "loss": 0.0838,
      "step": 3000
    },
    {
      "epoch": 0.7327739037938777,
      "grad_norm": 1.0785203140436177,
      "learning_rate": 1.7626662276103945e-06,
      "loss": 0.0878,
      "step": 3100
    },
    {
      "epoch": 0.7564117716581964,
      "grad_norm": 1.5221915620774547,
      "learning_rate": 1.4807419358704433e-06,
      "loss": 0.0863,
      "step": 3200
    },
    {
      "epoch": 0.780049639522515,
      "grad_norm": 2.3800125382539736,
      "learning_rate": 1.2194298309915998e-06,
      "loss": 0.0854,
      "step": 3300
    },
    {
      "epoch": 0.8036875073868337,
      "grad_norm": 2.2879011703898784,
      "learning_rate": 9.802604099810996e-07,
      "loss": 0.0834,
      "step": 3400
    },
    {
      "epoch": 0.8273253752511524,
      "grad_norm": 1.3292322263784997,
      "learning_rate": 7.646344808238904e-07,
      "loss": 0.0823,
      "step": 3500
    },
    {
      "epoch": 0.850963243115471,
      "grad_norm": 1.475790909857433,
      "learning_rate": 5.738149579931102e-07,
      "loss": 0.081,
      "step": 3600
    },
    {
      "epoch": 0.8746011109797897,
      "grad_norm": 1.3793002469089874,
      "learning_rate": 4.089194655986306e-07,
      "loss": 0.0819,
      "step": 3700
    },
    {
      "epoch": 0.8982389788441083,
      "grad_norm": 1.1451808677653712,
      "learning_rate": 2.7091379149682683e-07,
      "loss": 0.0803,
      "step": 3800
    },
    {
      "epoch": 0.9218768467084268,
      "grad_norm": 1.3775371900219944,
      "learning_rate": 1.606062307006384e-07,
      "loss": 0.0804,
      "step": 3900
    },
    {
      "epoch": 0.9455147145727455,
      "grad_norm": 1.1385483103356004,
      "learning_rate": 7.864285122039406e-08,
      "loss": 0.0824,
      "step": 4000
    },
    {
      "epoch": 0.9691525824370641,
      "grad_norm": 1.6058469540182345,
      "learning_rate": 2.55037100632316e-08,
      "loss": 0.0831,
      "step": 4100
    },
    {
      "epoch": 0.9927904503013828,
      "grad_norm": 1.1273389328177894,
      "learning_rate": 1.5000415538901636e-09,
      "loss": 0.0792,
      "step": 4200
    },
    {
      "epoch": 1.0,
      "step": 4231,
      "total_flos": 524421044371456.0,
      "train_loss": 0.1557054660529342,
      "train_runtime": 38067.1431,
      "train_samples_per_second": 7.113,
      "train_steps_per_second": 0.111
    }
  ],
  "logging_steps": 100,
  "max_steps": 4231,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 5000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 524421044371456.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}