{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 1000,
  "global_step": 16545,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00018132366273798732,
      "grad_norm": 7.9868855476379395,
      "learning_rate": 0.0,
      "loss": 3.5226,
      "step": 1
    },
    {
      "epoch": 0.1813236627379873,
      "grad_norm": 6.11737060546875,
      "learning_rate": 7.2443799854967376e-06,
      "loss": 2.9981,
      "step": 1000
    },
    {
      "epoch": 0.1813236627379873,
      "eval_cosine_accuracy": 0.9449994564056396,
      "eval_loss": 0.5478581190109253,
      "eval_runtime": 35.5173,
      "eval_samples_per_second": 267.728,
      "eval_steps_per_second": 2.112,
      "step": 1000
    },
    {
      "epoch": 0.3626473254759746,
      "grad_norm": 6.732676029205322,
      "learning_rate": 1.4496011602610589e-05,
      "loss": 2.3032,
      "step": 2000
    },
    {
      "epoch": 0.3626473254759746,
      "eval_cosine_accuracy": 0.9554106593132019,
      "eval_loss": 0.4920719563961029,
      "eval_runtime": 35.1849,
      "eval_samples_per_second": 270.258,
      "eval_steps_per_second": 2.132,
      "step": 2000
    },
    {
      "epoch": 0.543970988213962,
      "grad_norm": 7.598336696624756,
      "learning_rate": 1.980577829713503e-05,
      "loss": 1.8788,
      "step": 3000
    },
    {
      "epoch": 0.543970988213962,
      "eval_cosine_accuracy": 0.9590913653373718,
      "eval_loss": 0.45673736929893494,
      "eval_runtime": 35.2853,
      "eval_samples_per_second": 269.489,
      "eval_steps_per_second": 2.126,
      "step": 3000
    },
    {
      "epoch": 0.7252946509519492,
      "grad_norm": 10.34838581085205,
      "learning_rate": 1.8999879115122698e-05,
      "loss": 1.2997,
      "step": 4000
    },
    {
      "epoch": 0.7252946509519492,
      "eval_cosine_accuracy": 0.954990029335022,
      "eval_loss": 0.45149004459381104,
      "eval_runtime": 35.2489,
      "eval_samples_per_second": 269.768,
      "eval_steps_per_second": 2.128,
      "step": 4000
    },
    {
      "epoch": 0.9066183136899365,
      "grad_norm": 8.401123046875,
      "learning_rate": 1.8195591731474395e-05,
      "loss": 0.9457,
      "step": 5000
    },
    {
      "epoch": 0.9066183136899365,
      "eval_cosine_accuracy": 0.9530970454216003,
      "eval_loss": 0.44349202513694763,
      "eval_runtime": 35.1475,
      "eval_samples_per_second": 270.546,
      "eval_steps_per_second": 2.134,
      "step": 5000
    },
    {
      "epoch": 1.087941976427924,
      "grad_norm": 5.8470964431762695,
      "learning_rate": 1.7389692549462062e-05,
      "loss": 1.2109,
      "step": 6000
    },
    {
      "epoch": 1.087941976427924,
      "eval_cosine_accuracy": 0.9660322070121765,
      "eval_loss": 0.41235095262527466,
      "eval_runtime": 36.052,
      "eval_samples_per_second": 263.758,
      "eval_steps_per_second": 2.08,
      "step": 6000
    },
    {
      "epoch": 1.2692656391659112,
      "grad_norm": 5.963912010192871,
      "learning_rate": 1.6583793367449733e-05,
      "loss": 1.4479,
      "step": 7000
    },
    {
      "epoch": 1.2692656391659112,
      "eval_cosine_accuracy": 0.9669786691665649,
      "eval_loss": 0.4111262857913971,
      "eval_runtime": 35.6318,
      "eval_samples_per_second": 266.868,
      "eval_steps_per_second": 2.105,
      "step": 7000
    },
    {
      "epoch": 1.4505893019038985,
      "grad_norm": 5.44882345199585,
      "learning_rate": 1.5777894185437403e-05,
      "loss": 1.3188,
      "step": 8000
    },
    {
      "epoch": 1.4505893019038985,
      "eval_cosine_accuracy": 0.968766450881958,
      "eval_loss": 0.41267701983451843,
      "eval_runtime": 35.5881,
      "eval_samples_per_second": 267.196,
      "eval_steps_per_second": 2.107,
      "step": 8000
    },
    {
      "epoch": 1.6319129646418857,
      "grad_norm": 6.824221611022949,
      "learning_rate": 1.4972800902607086e-05,
      "loss": 1.1122,
      "step": 9000
    },
    {
      "epoch": 1.6319129646418857,
      "eval_cosine_accuracy": 0.9656115174293518,
      "eval_loss": 0.4086419939994812,
      "eval_runtime": 35.9831,
      "eval_samples_per_second": 264.263,
      "eval_steps_per_second": 2.084,
      "step": 9000
    },
    {
      "epoch": 1.8132366273798732,
      "grad_norm": 8.42564868927002,
      "learning_rate": 1.4166901720594754e-05,
      "loss": 0.7841,
      "step": 10000
    },
    {
      "epoch": 1.8132366273798732,
      "eval_cosine_accuracy": 0.9606688618659973,
      "eval_loss": 0.4070773720741272,
      "eval_runtime": 35.8917,
      "eval_samples_per_second": 264.936,
      "eval_steps_per_second": 2.09,
      "step": 10000
    },
    {
      "epoch": 1.9945602901178603,
      "grad_norm": 9.001717567443848,
      "learning_rate": 1.3361808437764437e-05,
      "loss": 0.6116,
      "step": 11000
    },
    {
      "epoch": 1.9945602901178603,
      "eval_cosine_accuracy": 0.957198441028595,
      "eval_loss": 0.41639891266822815,
      "eval_runtime": 35.808,
      "eval_samples_per_second": 265.555,
      "eval_steps_per_second": 2.095,
      "step": 11000
    },
    {
      "epoch": 2.175883952855848,
      "grad_norm": 6.010563850402832,
      "learning_rate": 1.2555909255752106e-05,
      "loss": 1.198,
      "step": 12000
    },
    {
      "epoch": 2.175883952855848,
      "eval_cosine_accuracy": 0.9699232578277588,
      "eval_loss": 0.39755257964134216,
      "eval_runtime": 35.4547,
      "eval_samples_per_second": 268.201,
      "eval_steps_per_second": 2.115,
      "step": 12000
    },
    {
      "epoch": 2.357207615593835,
      "grad_norm": 5.725988864898682,
      "learning_rate": 1.1750010073739776e-05,
      "loss": 1.1285,
      "step": 13000
    },
    {
      "epoch": 2.357207615593835,
      "eval_cosine_accuracy": 0.9707645177841187,
      "eval_loss": 0.3975719213485718,
      "eval_runtime": 35.4953,
      "eval_samples_per_second": 267.894,
      "eval_steps_per_second": 2.113,
      "step": 13000
    },
    {
      "epoch": 2.5385312783318223,
      "grad_norm": 6.512053489685059,
      "learning_rate": 1.0944110891727447e-05,
      "loss": 1.0768,
      "step": 14000
    },
    {
      "epoch": 2.5385312783318223,
      "eval_cosine_accuracy": 0.9691870808601379,
      "eval_loss": 0.3945649266242981,
      "eval_runtime": 35.4322,
      "eval_samples_per_second": 268.371,
      "eval_steps_per_second": 2.117,
      "step": 14000
    },
    {
      "epoch": 2.7198549410698094,
      "grad_norm": 7.771632671356201,
      "learning_rate": 1.0139017608897127e-05,
      "loss": 0.7841,
      "step": 15000
    },
    {
      "epoch": 2.7198549410698094,
      "eval_cosine_accuracy": 0.9662424921989441,
      "eval_loss": 0.39352062344551086,
      "eval_runtime": 34.5683,
      "eval_samples_per_second": 275.079,
      "eval_steps_per_second": 2.17,
      "step": 15000
    },
    {
      "epoch": 2.901178603807797,
      "grad_norm": 7.543709754943848,
      "learning_rate": 9.33392432606681e-06,
      "loss": 0.5724,
      "step": 16000
    },
    {
      "epoch": 2.901178603807797,
      "eval_cosine_accuracy": 0.9603533744812012,
      "eval_loss": 0.4049094021320343,
      "eval_runtime": 34.3685,
      "eval_samples_per_second": 276.677,
      "eval_steps_per_second": 2.182,
      "step": 16000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 27575,
  "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": 128,
  "trial_name": null,
  "trial_params": null
}