trainer_state.json

{
  "best_metric": 0.12305136024951935,
  "best_model_checkpoint": "/home/sdonoso/data/all_results/ner-c/albert-xxlarge/epochs_2_bs_4_lr_2e-5/checkpoint-1000",
  "epoch": 1.9995194617972127,
  "global_step": 1040,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.96,
      "learning_rate": 1.0442307692307694e-05,
      "loss": 0.158,
      "step": 500
    },
    {
      "epoch": 1.92,
      "learning_rate": 8.26923076923077e-07,
      "loss": 0.0558,
      "step": 1000
    },
    {
      "epoch": 1.92,
      "eval_accuracy": 0.9697488960653793,
      "eval_f1": 0.8438369228323929,
      "eval_loss": 0.12305136024951935,
      "eval_precision": 0.8397658968680797,
      "eval_recall": 0.8479476122025236,
      "eval_runtime": 26.2148,
      "eval_samples_per_second": 73.088,
      "eval_steps_per_second": 18.272,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "step": 1040,
      "total_flos": 1325433589814784.0,
      "train_loss": 0.10445189647949658,
      "train_runtime": 732.081,
      "train_samples_per_second": 22.741,
      "train_steps_per_second": 1.421
    }
  ],
  "max_steps": 1040,
  "num_train_epochs": 2,
  "total_flos": 1325433589814784.0,
  "trial_name": null,
  "trial_params": null
}