{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 1000,
  "global_step": 17343,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0001729804532087874,
      "grad_norm": 7.4878644943237305,
      "learning_rate": 0.0,
      "loss": 3.5397,
      "step": 1
    },
    {
      "epoch": 0.1729804532087874,
      "grad_norm": 5.6725382804870605,
      "learning_rate": 3.452689846047397e-06,
      "loss": 3.1143,
      "step": 1000
    },
    {
      "epoch": 0.1729804532087874,
      "eval_cosine_accuracy": 0.9447891712188721,
      "eval_loss": 1.529471516609192,
      "eval_runtime": 20.6972,
      "eval_samples_per_second": 459.435,
      "eval_steps_per_second": 3.624,
      "step": 1000
    },
    {
      "epoch": 0.3459609064175748,
      "grad_norm": 5.92240047454834,
      "learning_rate": 6.912298910223145e-06,
      "loss": 2.5164,
      "step": 2000
    },
    {
      "epoch": 0.3459609064175748,
      "eval_cosine_accuracy": 0.9559364914894104,
      "eval_loss": 1.4076837301254272,
      "eval_runtime": 20.5209,
      "eval_samples_per_second": 463.381,
      "eval_steps_per_second": 3.655,
      "step": 2000
    },
    {
      "epoch": 0.5189413596263622,
      "grad_norm": 6.010852336883545,
      "learning_rate": 1.0371907974398893e-05,
      "loss": 2.2667,
      "step": 3000
    },
    {
      "epoch": 0.5189413596263622,
      "eval_cosine_accuracy": 0.9617204666137695,
      "eval_loss": 1.3271853923797607,
      "eval_runtime": 20.4455,
      "eval_samples_per_second": 465.091,
      "eval_steps_per_second": 3.668,
      "step": 3000
    },
    {
      "epoch": 0.6919218128351496,
      "grad_norm": 5.722408771514893,
      "learning_rate": 1.3831517038574642e-05,
      "loss": 2.0664,
      "step": 4000
    },
    {
      "epoch": 0.6919218128351496,
      "eval_cosine_accuracy": 0.9661373496055603,
      "eval_loss": 1.262633204460144,
      "eval_runtime": 20.4097,
      "eval_samples_per_second": 465.906,
      "eval_steps_per_second": 3.675,
      "step": 4000
    },
    {
      "epoch": 0.864902266043937,
      "grad_norm": 5.409216403961182,
      "learning_rate": 1.729112610275039e-05,
      "loss": 1.9159,
      "step": 5000
    },
    {
      "epoch": 0.864902266043937,
      "eval_cosine_accuracy": 0.9705542325973511,
      "eval_loss": 1.195585012435913,
      "eval_runtime": 20.7661,
      "eval_samples_per_second": 457.91,
      "eval_steps_per_second": 3.612,
      "step": 5000
    },
    {
      "epoch": 1.0378827192527245,
      "grad_norm": 5.540963172912598,
      "learning_rate": 1.9813181110534513e-05,
      "loss": 1.7788,
      "step": 6000
    },
    {
      "epoch": 1.0378827192527245,
      "eval_cosine_accuracy": 0.9710800051689148,
      "eval_loss": 1.1745668649673462,
      "eval_runtime": 21.9748,
      "eval_samples_per_second": 432.723,
      "eval_steps_per_second": 3.413,
      "step": 6000
    },
    {
      "epoch": 1.2108631724615118,
      "grad_norm": 5.699978828430176,
      "learning_rate": 1.8948278844490574e-05,
      "loss": 1.6737,
      "step": 7000
    },
    {
      "epoch": 1.2108631724615118,
      "eval_cosine_accuracy": 0.9727626442909241,
      "eval_loss": 1.1350425481796265,
      "eval_runtime": 20.6263,
      "eval_samples_per_second": 461.014,
      "eval_steps_per_second": 3.636,
      "step": 7000
    },
    {
      "epoch": 1.3838436256702993,
      "grad_norm": 6.418245792388916,
      "learning_rate": 1.8083376578446638e-05,
      "loss": 1.5966,
      "step": 8000
    },
    {
      "epoch": 1.3838436256702993,
      "eval_cosine_accuracy": 0.9716058373451233,
      "eval_loss": 1.122388482093811,
      "eval_runtime": 20.6692,
      "eval_samples_per_second": 460.056,
      "eval_steps_per_second": 3.629,
      "step": 8000
    },
    {
      "epoch": 1.5568240788790866,
      "grad_norm": 5.495181083679199,
      "learning_rate": 1.72184743124027e-05,
      "loss": 1.535,
      "step": 9000
    },
    {
      "epoch": 1.5568240788790866,
      "eval_cosine_accuracy": 0.9738143086433411,
      "eval_loss": 1.0911048650741577,
      "eval_runtime": 20.6637,
      "eval_samples_per_second": 460.179,
      "eval_steps_per_second": 3.63,
      "step": 9000
    },
    {
      "epoch": 1.729804532087874,
      "grad_norm": 5.562320709228516,
      "learning_rate": 1.6354436948624807e-05,
      "loss": 1.4752,
      "step": 10000
    },
    {
      "epoch": 1.729804532087874,
      "eval_cosine_accuracy": 0.9746555685997009,
      "eval_loss": 1.0861026048660278,
      "eval_runtime": 20.6157,
      "eval_samples_per_second": 461.249,
      "eval_steps_per_second": 3.638,
      "step": 10000
    },
    {
      "epoch": 1.9027849852966616,
      "grad_norm": 5.75679349899292,
      "learning_rate": 1.5489534682580868e-05,
      "loss": 1.4419,
      "step": 11000
    },
    {
      "epoch": 1.9027849852966616,
      "eval_cosine_accuracy": 0.9758123755455017,
      "eval_loss": 1.0716735124588013,
      "eval_runtime": 21.2666,
      "eval_samples_per_second": 447.132,
      "eval_steps_per_second": 3.527,
      "step": 11000
    },
    {
      "epoch": 2.075765438505449,
      "grad_norm": 5.386068820953369,
      "learning_rate": 1.4625497318802977e-05,
      "loss": 1.3926,
      "step": 12000
    },
    {
      "epoch": 2.075765438505449,
      "eval_cosine_accuracy": 0.9744452834129333,
      "eval_loss": 1.0726546049118042,
      "eval_runtime": 20.6618,
      "eval_samples_per_second": 460.22,
      "eval_steps_per_second": 3.63,
      "step": 12000
    },
    {
      "epoch": 2.2487458917142362,
      "grad_norm": 5.619034290313721,
      "learning_rate": 1.3760595052759041e-05,
      "loss": 1.3489,
      "step": 13000
    },
    {
      "epoch": 2.2487458917142362,
      "eval_cosine_accuracy": 0.976443350315094,
      "eval_loss": 1.0502160787582397,
      "eval_runtime": 20.5492,
      "eval_samples_per_second": 462.743,
      "eval_steps_per_second": 3.65,
      "step": 13000
    },
    {
      "epoch": 2.4217263449230235,
      "grad_norm": 5.643714904785156,
      "learning_rate": 1.2895692786715102e-05,
      "loss": 1.3197,
      "step": 14000
    },
    {
      "epoch": 2.4217263449230235,
      "eval_cosine_accuracy": 0.9762330651283264,
      "eval_loss": 1.0510120391845703,
      "eval_runtime": 20.394,
      "eval_samples_per_second": 466.265,
      "eval_steps_per_second": 3.678,
      "step": 14000
    },
    {
      "epoch": 2.5947067981318113,
      "grad_norm": 5.773045063018799,
      "learning_rate": 1.2030790520671166e-05,
      "loss": 1.2936,
      "step": 15000
    },
    {
      "epoch": 2.5947067981318113,
      "eval_cosine_accuracy": 0.9771795272827148,
      "eval_loss": 1.039699673652649,
      "eval_runtime": 20.5225,
      "eval_samples_per_second": 463.344,
      "eval_steps_per_second": 3.655,
      "step": 15000
    },
    {
      "epoch": 2.7676872513405986,
      "grad_norm": 5.743391990661621,
      "learning_rate": 1.1166753156893273e-05,
      "loss": 1.2835,
      "step": 16000
    },
    {
      "epoch": 2.7676872513405986,
      "eval_cosine_accuracy": 0.9763382077217102,
      "eval_loss": 1.0374181270599365,
      "eval_runtime": 21.3291,
      "eval_samples_per_second": 445.823,
      "eval_steps_per_second": 3.516,
      "step": 16000
    },
    {
      "epoch": 2.940667704549386,
      "grad_norm": 6.1901750564575195,
      "learning_rate": 1.0301850890849333e-05,
      "loss": 1.2736,
      "step": 17000
    },
    {
      "epoch": 2.940667704549386,
      "eval_cosine_accuracy": 0.9771795272827148,
      "eval_loss": 1.032090187072754,
      "eval_runtime": 20.7572,
      "eval_samples_per_second": 458.106,
      "eval_steps_per_second": 3.613,
      "step": 17000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 28905,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}