trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 4422,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.34,
      "learning_rate": 4.434644957033017e-05,
      "loss": 0.484,
      "step": 500
    },
    {
      "epoch": 0.68,
      "learning_rate": 3.869289914066034e-05,
      "loss": 0.4117,
      "step": 1000
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.30393487109905e-05,
      "loss": 0.3563,
      "step": 1500
    },
    {
      "epoch": 1.36,
      "learning_rate": 2.7385798281320668e-05,
      "loss": 0.365,
      "step": 2000
    },
    {
      "epoch": 1.7,
      "learning_rate": 2.173224785165084e-05,
      "loss": 0.3432,
      "step": 2500
    },
    {
      "epoch": 2.04,
      "learning_rate": 1.6078697421981004e-05,
      "loss": 0.3333,
      "step": 3000
    },
    {
      "epoch": 2.37,
      "learning_rate": 1.0425146992311172e-05,
      "loss": 0.3353,
      "step": 3500
    },
    {
      "epoch": 2.71,
      "learning_rate": 4.771596562641339e-06,
      "loss": 0.3278,
      "step": 4000
    },
    {
      "epoch": 3.0,
      "step": 4422,
      "total_flos": 109498918895616.0,
      "train_loss": 0.3655550269614728,
      "train_runtime": 1226.4853,
      "train_samples_per_second": 14.422,
      "train_steps_per_second": 3.605
    }
  ],
  "logging_steps": 500,
  "max_steps": 4422,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 109498918895616.0,
  "trial_name": null,
  "trial_params": null
}