trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 910,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 5.0,
      "learning_rate": 2.5e-05,
      "loss": 0.1351,
      "step": 455
    },
    {
      "epoch": 5.0,
      "eval_f1": 0.9119336311423102,
      "eval_loss": 0.46077126264572144,
      "eval_runtime": 12.6487,
      "eval_samples_per_second": 227.928,
      "eval_steps_per_second": 1.818,
      "step": 455
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0,
      "loss": 0.0114,
      "step": 910
    },
    {
      "epoch": 10.0,
      "eval_f1": 0.9135014363230132,
      "eval_loss": 0.567248523235321,
      "eval_runtime": 12.6489,
      "eval_samples_per_second": 227.925,
      "eval_steps_per_second": 1.818,
      "step": 910
    },
    {
      "epoch": 10.0,
      "step": 910,
      "total_flos": 3.03419669041152e+16,
      "train_loss": 0.07324187100588621,
      "train_runtime": 1328.6278,
      "train_samples_per_second": 86.796,
      "train_steps_per_second": 0.685
    }
  ],
  "max_steps": 910,
  "num_train_epochs": 10,
  "total_flos": 3.03419669041152e+16,
  "trial_name": null,
  "trial_params": null
}