{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 3858,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.38880248833592534,
      "grad_norm": 0.6042742133140564,
      "learning_rate": 1.740798341109383e-05,
      "loss": 0.2151,
      "step": 500
    },
    {
      "epoch": 0.38880248833592534,
      "eval_accuracy": 0.9762738234150136,
      "eval_loss": 0.10295161604881287,
      "eval_runtime": 26.8587,
      "eval_samples_per_second": 95.723,
      "eval_steps_per_second": 11.989,
      "step": 500
    },
    {
      "epoch": 0.7776049766718507,
      "grad_norm": 0.04766521230340004,
      "learning_rate": 1.4815966822187664e-05,
      "loss": 0.1167,
      "step": 1000
    },
    {
      "epoch": 0.7776049766718507,
      "eval_accuracy": 0.9875534811357448,
      "eval_loss": 0.05390123650431633,
      "eval_runtime": 26.735,
      "eval_samples_per_second": 96.166,
      "eval_steps_per_second": 12.044,
      "step": 1000
    },
    {
      "epoch": 1.166407465007776,
      "grad_norm": 0.008871573023498058,
      "learning_rate": 1.2223950233281495e-05,
      "loss": 0.0687,
      "step": 1500
    },
    {
      "epoch": 1.166407465007776,
      "eval_accuracy": 0.9793854531310774,
      "eval_loss": 0.12513048946857452,
      "eval_runtime": 27.1074,
      "eval_samples_per_second": 94.845,
      "eval_steps_per_second": 11.879,
      "step": 1500
    },
    {
      "epoch": 1.5552099533437014,
      "grad_norm": 0.0014064661227166653,
      "learning_rate": 9.631933644375326e-06,
      "loss": 0.0279,
      "step": 2000
    },
    {
      "epoch": 1.5552099533437014,
      "eval_accuracy": 0.9778296382730455,
      "eval_loss": 0.148821160197258,
      "eval_runtime": 26.879,
      "eval_samples_per_second": 95.651,
      "eval_steps_per_second": 11.98,
      "step": 2000
    },
    {
      "epoch": 1.9440124416796267,
      "grad_norm": 0.002627410925924778,
      "learning_rate": 7.039917055469155e-06,
      "loss": 0.0293,
      "step": 2500
    },
    {
      "epoch": 1.9440124416796267,
      "eval_accuracy": 0.9914430182808246,
      "eval_loss": 0.04895803704857826,
      "eval_runtime": 26.8775,
      "eval_samples_per_second": 95.656,
      "eval_steps_per_second": 11.98,
      "step": 2500
    },
    {
      "epoch": 2.332814930015552,
      "grad_norm": 0.0008352847071364522,
      "learning_rate": 4.447900466562986e-06,
      "loss": 0.0119,
      "step": 3000
    },
    {
      "epoch": 2.332814930015552,
      "eval_accuracy": 0.9766627771295215,
      "eval_loss": 0.15932457149028778,
      "eval_runtime": 27.4463,
      "eval_samples_per_second": 93.674,
      "eval_steps_per_second": 11.732,
      "step": 3000
    },
    {
      "epoch": 2.721617418351477,
      "grad_norm": 0.006800688803195953,
      "learning_rate": 1.8558838776568172e-06,
      "loss": 0.0045,
      "step": 3500
    },
    {
      "epoch": 2.721617418351477,
      "eval_accuracy": 0.9797744068455854,
      "eval_loss": 0.14682677388191223,
      "eval_runtime": 27.6918,
      "eval_samples_per_second": 92.843,
      "eval_steps_per_second": 11.628,
      "step": 3500
    }
  ],
  "logging_steps": 500,
  "max_steps": 3858,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 5863351757603040.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}