trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6467368421052632,
  "eval_steps": 500,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 1.9848077530122083e-05,
      "loss": 1.5183,
      "step": 1
    },
    {
      "epoch": 0.27,
      "learning_rate": 1.6427876096865394e-05,
      "loss": 1.5068,
      "step": 5
    },
    {
      "epoch": 0.54,
      "learning_rate": 8.263518223330698e-06,
      "loss": 1.4503,
      "step": 10
    },
    {
      "epoch": 0.65,
      "eval_loss": 1.429574966430664,
      "eval_runtime": 59.7378,
      "eval_samples_per_second": 8.37,
      "eval_steps_per_second": 1.055,
      "step": 12
    },
    {
      "epoch": 0.65,
      "step": 12,
      "total_flos": 5.55304911634432e+17,
      "train_loss": 1.507220059633255,
      "train_runtime": 3726.2953,
      "train_samples_per_second": 2.549,
      "train_steps_per_second": 0.005
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 5.55304911634432e+17,
  "trial_name": null,
  "trial_params": null
}