{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 1000,
  "global_step": 11298,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002655337227827934,
      "grad_norm": 5.798187732696533,
      "learning_rate": 0.0,
      "loss": 1.8907,
      "step": 1
    },
    {
      "epoch": 0.2655337227827934,
      "grad_norm": 3.392871618270874,
      "learning_rate": 2.652681890600106e-06,
      "loss": 1.46,
      "step": 1000
    },
    {
      "epoch": 0.2655337227827934,
      "eval_cosine_accuracy": 0.9434220194816589,
      "eval_loss": 0.5881128311157227,
      "eval_runtime": 37.2445,
      "eval_samples_per_second": 255.313,
      "eval_steps_per_second": 1.02,
      "step": 1000
    },
    {
      "epoch": 0.5310674455655868,
      "grad_norm": 4.811636924743652,
      "learning_rate": 5.30801911842804e-06,
      "loss": 1.005,
      "step": 2000
    },
    {
      "epoch": 0.5310674455655868,
      "eval_cosine_accuracy": 0.9501525163650513,
      "eval_loss": 0.5147020816802979,
      "eval_runtime": 34.2592,
      "eval_samples_per_second": 277.56,
      "eval_steps_per_second": 1.109,
      "step": 2000
    },
    {
      "epoch": 0.7966011683483802,
      "grad_norm": 4.881891250610352,
      "learning_rate": 7.963356346255975e-06,
      "loss": 0.7199,
      "step": 3000
    },
    {
      "epoch": 0.7966011683483802,
      "eval_cosine_accuracy": 0.9585655927658081,
      "eval_loss": 0.4466570317745209,
      "eval_runtime": 35.3151,
      "eval_samples_per_second": 269.262,
      "eval_steps_per_second": 1.076,
      "step": 3000
    },
    {
      "epoch": 1.0621348911311737,
      "grad_norm": 3.386178970336914,
      "learning_rate": 9.84599044078598e-06,
      "loss": 0.6439,
      "step": 4000
    },
    {
      "epoch": 1.0621348911311737,
      "eval_cosine_accuracy": 0.9612998366355896,
      "eval_loss": 0.42167210578918457,
      "eval_runtime": 34.503,
      "eval_samples_per_second": 275.599,
      "eval_steps_per_second": 1.101,
      "step": 4000
    },
    {
      "epoch": 1.327668613913967,
      "grad_norm": 3.1094956398010254,
      "learning_rate": 9.182156133828996e-06,
      "loss": 0.7827,
      "step": 5000
    },
    {
      "epoch": 1.327668613913967,
      "eval_cosine_accuracy": 0.967714786529541,
      "eval_loss": 0.3992333710193634,
      "eval_runtime": 34.5803,
      "eval_samples_per_second": 274.984,
      "eval_steps_per_second": 1.099,
      "step": 5000
    },
    {
      "epoch": 1.5932023366967605,
      "grad_norm": 4.082334518432617,
      "learning_rate": 8.518321826872013e-06,
      "loss": 0.5518,
      "step": 6000
    },
    {
      "epoch": 1.5932023366967605,
      "eval_cosine_accuracy": 0.9695025682449341,
      "eval_loss": 0.3816065192222595,
      "eval_runtime": 35.1269,
      "eval_samples_per_second": 270.704,
      "eval_steps_per_second": 1.082,
      "step": 6000
    },
    {
      "epoch": 1.858736059479554,
      "grad_norm": 3.973569631576538,
      "learning_rate": 7.85448751991503e-06,
      "loss": 0.4181,
      "step": 7000
    },
    {
      "epoch": 1.858736059479554,
      "eval_cosine_accuracy": 0.9710800051689148,
      "eval_loss": 0.36719873547554016,
      "eval_runtime": 34.8837,
      "eval_samples_per_second": 272.592,
      "eval_steps_per_second": 1.089,
      "step": 7000
    },
    {
      "epoch": 2.1242697822623473,
      "grad_norm": 2.694537401199341,
      "learning_rate": 7.1906532129580465e-06,
      "loss": 0.5191,
      "step": 8000
    },
    {
      "epoch": 2.1242697822623473,
      "eval_cosine_accuracy": 0.9718161821365356,
      "eval_loss": 0.37118107080459595,
      "eval_runtime": 34.5413,
      "eval_samples_per_second": 275.294,
      "eval_steps_per_second": 1.1,
      "step": 8000
    },
    {
      "epoch": 2.3898035050451405,
      "grad_norm": 2.9365272521972656,
      "learning_rate": 6.526818906001062e-06,
      "loss": 0.5758,
      "step": 9000
    },
    {
      "epoch": 2.3898035050451405,
      "eval_cosine_accuracy": 0.9725522994995117,
      "eval_loss": 0.3617595136165619,
      "eval_runtime": 34.1585,
      "eval_samples_per_second": 278.379,
      "eval_steps_per_second": 1.112,
      "step": 9000
    },
    {
      "epoch": 2.655337227827934,
      "grad_norm": 3.451385498046875,
      "learning_rate": 5.863648433351036e-06,
      "loss": 0.4162,
      "step": 10000
    },
    {
      "epoch": 2.655337227827934,
      "eval_cosine_accuracy": 0.9733936190605164,
      "eval_loss": 0.3557143807411194,
      "eval_runtime": 35.0722,
      "eval_samples_per_second": 271.126,
      "eval_steps_per_second": 1.083,
      "step": 10000
    },
    {
      "epoch": 2.9208709506107278,
      "grad_norm": 3.8668599128723145,
      "learning_rate": 5.199814126394053e-06,
      "loss": 0.3282,
      "step": 11000
    },
    {
      "epoch": 2.9208709506107278,
      "eval_cosine_accuracy": 0.9744452834129333,
      "eval_loss": 0.34901732206344604,
      "eval_runtime": 35.1373,
      "eval_samples_per_second": 270.624,
      "eval_steps_per_second": 1.081,
      "step": 11000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 18830,
  "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": 256,
  "trial_name": null,
  "trial_params": null
}