{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 1000,
  "global_step": 15166,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0002638522427440633,
      "grad_norm": 9.534642219543457,
      "learning_rate": 0.0,
      "loss": 5.4604,
      "step": 1
    },
    {
      "epoch": 0.2638522427440633,
      "grad_norm": 5.137489318847656,
      "learning_rate": 1.0543535620052772e-05,
      "loss": 3.533,
      "step": 1000
    },
    {
      "epoch": 0.2638522427440633,
      "eval_cosine_accuracy": 0.8844252824783325,
      "eval_loss": 0.9935443997383118,
      "eval_runtime": 35.3218,
      "eval_samples_per_second": 269.21,
      "eval_steps_per_second": 2.123,
      "step": 1000
    },
    {
      "epoch": 0.5277044854881267,
      "grad_norm": 6.9049296379089355,
      "learning_rate": 1.9878041630020525e-05,
      "loss": 2.3361,
      "step": 2000
    },
    {
      "epoch": 0.5277044854881267,
      "eval_cosine_accuracy": 0.8922073841094971,
      "eval_loss": 0.9457609057426453,
      "eval_runtime": 34.4447,
      "eval_samples_per_second": 276.066,
      "eval_steps_per_second": 2.177,
      "step": 2000
    },
    {
      "epoch": 0.7915567282321899,
      "grad_norm": 8.220151901245117,
      "learning_rate": 1.8706537672236883e-05,
      "loss": 1.3548,
      "step": 3000
    },
    {
      "epoch": 0.7915567282321899,
      "eval_cosine_accuracy": 0.8851613998413086,
      "eval_loss": 0.9700921177864075,
      "eval_runtime": 34.3953,
      "eval_samples_per_second": 276.462,
      "eval_steps_per_second": 2.181,
      "step": 3000
    },
    {
      "epoch": 1.0553797468354431,
      "grad_norm": 7.737990379333496,
      "learning_rate": 1.753503371445324e-05,
      "loss": 1.0121,
      "step": 4000
    },
    {
      "epoch": 1.0553797468354431,
      "eval_cosine_accuracy": 0.8963087797164917,
      "eval_loss": 0.9534999132156372,
      "eval_runtime": 34.7519,
      "eval_samples_per_second": 273.626,
      "eval_steps_per_second": 2.158,
      "step": 4000
    },
    {
      "epoch": 1.3190928270042195,
      "grad_norm": 5.373760223388672,
      "learning_rate": 1.636235708003518e-05,
      "loss": 1.3913,
      "step": 5000
    },
    {
      "epoch": 1.3190928270042195,
      "eval_cosine_accuracy": 0.9013566374778748,
      "eval_loss": 0.9373352527618408,
      "eval_runtime": 34.4954,
      "eval_samples_per_second": 275.66,
      "eval_steps_per_second": 2.174,
      "step": 5000
    },
    {
      "epoch": 1.5828059071729959,
      "grad_norm": 9.4461088180542,
      "learning_rate": 1.5189680445617123e-05,
      "loss": 1.3354,
      "step": 6000
    },
    {
      "epoch": 1.5828059071729959,
      "eval_cosine_accuracy": 0.9082974195480347,
      "eval_loss": 0.9365662932395935,
      "eval_runtime": 35.9891,
      "eval_samples_per_second": 264.219,
      "eval_steps_per_second": 2.084,
      "step": 6000
    },
    {
      "epoch": 1.846518987341772,
      "grad_norm": 10.50667667388916,
      "learning_rate": 1.4017003811199064e-05,
      "loss": 1.2488,
      "step": 7000
    },
    {
      "epoch": 1.846518987341772,
      "eval_cosine_accuracy": 0.9102954864501953,
      "eval_loss": 0.9144614338874817,
      "eval_runtime": 34.5544,
      "eval_samples_per_second": 275.189,
      "eval_steps_per_second": 2.17,
      "step": 7000
    },
    {
      "epoch": 2.1102320675105486,
      "grad_norm": 5.640753269195557,
      "learning_rate": 1.2846672530049839e-05,
      "loss": 1.1746,
      "step": 8000
    },
    {
      "epoch": 2.1102320675105486,
      "eval_cosine_accuracy": 0.9104006886482239,
      "eval_loss": 0.9235846996307373,
      "eval_runtime": 34.588,
      "eval_samples_per_second": 274.922,
      "eval_steps_per_second": 2.168,
      "step": 8000
    },
    {
      "epoch": 2.3739451476793247,
      "grad_norm": 6.157176971435547,
      "learning_rate": 1.167399589563178e-05,
      "loss": 1.127,
      "step": 9000
    },
    {
      "epoch": 2.3739451476793247,
      "eval_cosine_accuracy": 0.912924587726593,
      "eval_loss": 0.9103002548217773,
      "eval_runtime": 34.8713,
      "eval_samples_per_second": 272.689,
      "eval_steps_per_second": 2.151,
      "step": 9000
    },
    {
      "epoch": 2.6376582278481013,
      "grad_norm": 5.3915510177612305,
      "learning_rate": 1.050131926121372e-05,
      "loss": 1.0852,
      "step": 10000
    },
    {
      "epoch": 2.6376582278481013,
      "eval_cosine_accuracy": 0.9119781255722046,
      "eval_loss": 0.9026044607162476,
      "eval_runtime": 34.7266,
      "eval_samples_per_second": 273.825,
      "eval_steps_per_second": 2.16,
      "step": 10000
    },
    {
      "epoch": 2.9013713080168775,
      "grad_norm": 5.952890872955322,
      "learning_rate": 9.328642626795662e-06,
      "loss": 1.0764,
      "step": 11000
    },
    {
      "epoch": 2.9013713080168775,
      "eval_cosine_accuracy": 0.9142917394638062,
      "eval_loss": 0.8946072459220886,
      "eval_runtime": 35.5251,
      "eval_samples_per_second": 267.67,
      "eval_steps_per_second": 2.111,
      "step": 11000
    },
    {
      "epoch": 3.165084388185654,
      "grad_norm": 5.450469493865967,
      "learning_rate": 8.155965992377603e-06,
      "loss": 1.0508,
      "step": 12000
    },
    {
      "epoch": 3.165084388185654,
      "eval_cosine_accuracy": 0.9132400751113892,
      "eval_loss": 0.9051838517189026,
      "eval_runtime": 34.3957,
      "eval_samples_per_second": 276.459,
      "eval_steps_per_second": 2.181,
      "step": 12000
    },
    {
      "epoch": 3.4287974683544302,
      "grad_norm": 5.958968162536621,
      "learning_rate": 6.983289357959543e-06,
      "loss": 1.0045,
      "step": 13000
    },
    {
      "epoch": 3.4287974683544302,
      "eval_cosine_accuracy": 0.9142917394638062,
      "eval_loss": 0.9047877192497253,
      "eval_runtime": 35.0075,
      "eval_samples_per_second": 271.628,
      "eval_steps_per_second": 2.142,
      "step": 13000
    },
    {
      "epoch": 3.692510548523207,
      "grad_norm": 5.284270763397217,
      "learning_rate": 5.811785400175902e-06,
      "loss": 0.998,
      "step": 14000
    },
    {
      "epoch": 3.692510548523207,
      "eval_cosine_accuracy": 0.9154485464096069,
      "eval_loss": 0.903547465801239,
      "eval_runtime": 34.5001,
      "eval_samples_per_second": 275.622,
      "eval_steps_per_second": 2.174,
      "step": 14000
    },
    {
      "epoch": 3.956223628691983,
      "grad_norm": 5.081764221191406,
      "learning_rate": 4.640281442392261e-06,
      "loss": 0.994,
      "step": 15000
    },
    {
      "epoch": 3.956223628691983,
      "eval_cosine_accuracy": 0.9173414707183838,
      "eval_loss": 0.8898885250091553,
      "eval_runtime": 34.5739,
      "eval_samples_per_second": 275.034,
      "eval_steps_per_second": 2.169,
      "step": 15000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 18950,
  "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
}