{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 8456,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.23651844843897823,
      "grad_norm": 0.75390625,
      "learning_rate": 1.881740775780511e-05,
      "loss": 0.2864,
      "mean_token_accuracy": 0.9243696781396866,
      "step": 500
    },
    {
      "epoch": 0.47303689687795647,
      "grad_norm": 0.94140625,
      "learning_rate": 1.763481551561022e-05,
      "loss": 0.2402,
      "mean_token_accuracy": 0.9334622280597686,
      "step": 1000
    },
    {
      "epoch": 0.7095553453169348,
      "grad_norm": 1.4453125,
      "learning_rate": 1.6452223273415328e-05,
      "loss": 0.2289,
      "mean_token_accuracy": 0.9359804515838623,
      "step": 1500
    },
    {
      "epoch": 0.9460737937559129,
      "grad_norm": 0.9765625,
      "learning_rate": 1.5269631031220437e-05,
      "loss": 0.2271,
      "mean_token_accuracy": 0.9360219603776931,
      "step": 2000
    },
    {
      "epoch": 1.1825922421948911,
      "grad_norm": 1.2265625,
      "learning_rate": 1.4087038789025544e-05,
      "loss": 0.2127,
      "mean_token_accuracy": 0.9386496288776398,
      "step": 2500
    },
    {
      "epoch": 1.4191106906338695,
      "grad_norm": 1.6171875,
      "learning_rate": 1.2904446546830653e-05,
      "loss": 0.2061,
      "mean_token_accuracy": 0.9404804083108902,
      "step": 3000
    },
    {
      "epoch": 1.6556291390728477,
      "grad_norm": 0.7265625,
      "learning_rate": 1.1721854304635763e-05,
      "loss": 0.2084,
      "mean_token_accuracy": 0.9395650044679642,
      "step": 3500
    },
    {
      "epoch": 1.8921475875118259,
      "grad_norm": 1.25,
      "learning_rate": 1.0539262062440872e-05,
      "loss": 0.2,
      "mean_token_accuracy": 0.9420614712238312,
      "step": 4000
    },
    {
      "epoch": 2.128666035950804,
      "grad_norm": 1.1484375,
      "learning_rate": 9.35666982024598e-06,
      "loss": 0.2018,
      "mean_token_accuracy": 0.941098618388176,
      "step": 4500
    },
    {
      "epoch": 2.3651844843897822,
      "grad_norm": 0.98046875,
      "learning_rate": 8.174077578051088e-06,
      "loss": 0.1967,
      "mean_token_accuracy": 0.94244316136837,
      "step": 5000
    },
    {
      "epoch": 2.6017029328287604,
      "grad_norm": 0.72265625,
      "learning_rate": 6.991485335856198e-06,
      "loss": 0.1918,
      "mean_token_accuracy": 0.9436353342533111,
      "step": 5500
    },
    {
      "epoch": 2.838221381267739,
      "grad_norm": 1.3125,
      "learning_rate": 5.808893093661306e-06,
      "loss": 0.1947,
      "mean_token_accuracy": 0.9429612046480179,
      "step": 6000
    },
    {
      "epoch": 3.074739829706717,
      "grad_norm": 1.171875,
      "learning_rate": 4.626300851466415e-06,
      "loss": 0.1925,
      "mean_token_accuracy": 0.9436095499992371,
      "step": 6500
    },
    {
      "epoch": 3.3112582781456954,
      "grad_norm": 1.328125,
      "learning_rate": 3.443708609271523e-06,
      "loss": 0.1888,
      "mean_token_accuracy": 0.9442265207767486,
      "step": 7000
    },
    {
      "epoch": 3.5477767265846736,
      "grad_norm": 1.015625,
      "learning_rate": 2.2611163670766324e-06,
      "loss": 0.1948,
      "mean_token_accuracy": 0.9426910729408264,
      "step": 7500
    },
    {
      "epoch": 3.7842951750236518,
      "grad_norm": 1.5078125,
      "learning_rate": 1.078524124881741e-06,
      "loss": 0.1929,
      "mean_token_accuracy": 0.9433188273906707,
      "step": 8000
    }
  ],
  "logging_steps": 500,
  "max_steps": 8456,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.722176240290611e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}