checkpoint-900/trainer_state.json

{
  "best_metric": 0.8647798742138365,
  "best_model_checkpoint": "/mnt/data01/entity/ner_trained_models//roberta-large_lr_4.71695374919064e-05_ml_128_bs_128_e_184/checkpoint-500",
  "epoch": 128.57142857142858,
  "global_step": 900,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 14.29,
      "eval_dev1_accuracy_score": 0.9730900798175599,
      "eval_dev1_f1": 0.8222730739893211,
      "eval_dev1_loss": 0.13282082974910736,
      "eval_dev1_precision": 0.7973372781065089,
      "eval_dev1_recall": 0.8488188976377953,
      "eval_dev1_runtime": 0.896,
      "eval_dev1_samples_per_second": 147.323,
      "eval_dev1_steps_per_second": 2.232,
      "step": 100
    },
    {
      "epoch": 28.57,
      "eval_dev1_accuracy_score": 0.9779931584948689,
      "eval_dev1_f1": 0.8483896307934015,
      "eval_dev1_loss": 0.15858674049377441,
      "eval_dev1_precision": 0.8463949843260188,
      "eval_dev1_recall": 0.8503937007874016,
      "eval_dev1_runtime": 0.8958,
      "eval_dev1_samples_per_second": 147.352,
      "eval_dev1_steps_per_second": 2.233,
      "step": 200
    },
    {
      "epoch": 42.86,
      "eval_dev1_accuracy_score": 0.9770809578107184,
      "eval_dev1_f1": 0.8430913348946136,
      "eval_dev1_loss": 0.1632351577281952,
      "eval_dev1_precision": 0.8359133126934984,
      "eval_dev1_recall": 0.8503937007874016,
      "eval_dev1_runtime": 0.8958,
      "eval_dev1_samples_per_second": 147.362,
      "eval_dev1_steps_per_second": 2.233,
      "step": 300
    },
    {
      "epoch": 57.14,
      "eval_dev1_accuracy_score": 0.9775370581527936,
      "eval_dev1_f1": 0.8463949843260189,
      "eval_dev1_loss": 0.16842932999134064,
      "eval_dev1_precision": 0.8424336973478939,
      "eval_dev1_recall": 0.8503937007874016,
      "eval_dev1_runtime": 0.8946,
      "eval_dev1_samples_per_second": 147.555,
      "eval_dev1_steps_per_second": 2.236,
      "step": 400
    },
    {
      "epoch": 71.43,
      "learning_rate": 2.8858381633247082e-05,
      "loss": 0.0251,
      "step": 500
    },
    {
      "epoch": 71.43,
      "eval_dev1_accuracy_score": 0.9784492588369441,
      "eval_dev1_f1": 0.8647798742138365,
      "eval_dev1_loss": 0.16985777020454407,
      "eval_dev1_precision": 0.8634222919937206,
      "eval_dev1_recall": 0.8661417322834646,
      "eval_dev1_runtime": 0.8956,
      "eval_dev1_samples_per_second": 147.384,
      "eval_dev1_steps_per_second": 2.233,
      "step": 500
    },
    {
      "epoch": 85.71,
      "eval_dev1_accuracy_score": 0.9787913340935006,
      "eval_dev1_f1": 0.8539857932123125,
      "eval_dev1_loss": 0.17249244451522827,
      "eval_dev1_precision": 0.8560126582278481,
      "eval_dev1_recall": 0.8519685039370078,
      "eval_dev1_runtime": 0.8962,
      "eval_dev1_samples_per_second": 147.288,
      "eval_dev1_steps_per_second": 2.232,
      "step": 600
    },
    {
      "epoch": 100.0,
      "eval_dev1_accuracy_score": 0.9778791334093501,
      "eval_dev1_f1": 0.8491008600469115,
      "eval_dev1_loss": 0.18127191066741943,
      "eval_dev1_precision": 0.843167701863354,
      "eval_dev1_recall": 0.8551181102362204,
      "eval_dev1_runtime": 0.8957,
      "eval_dev1_samples_per_second": 147.363,
      "eval_dev1_steps_per_second": 2.233,
      "step": 700
    },
    {
      "epoch": 114.29,
      "eval_dev1_accuracy_score": 0.976054732041049,
      "eval_dev1_f1": 0.834755624515128,
      "eval_dev1_loss": 0.18781279027462006,
      "eval_dev1_precision": 0.8226299694189603,
      "eval_dev1_recall": 0.8472440944881889,
      "eval_dev1_runtime": 0.8971,
      "eval_dev1_samples_per_second": 147.138,
      "eval_dev1_steps_per_second": 2.229,
      "step": 800
    },
    {
      "epoch": 128.57,
      "eval_dev1_accuracy_score": 0.9789053591790193,
      "eval_dev1_f1": 0.8557993730407524,
      "eval_dev1_loss": 0.17089255154132843,
      "eval_dev1_precision": 0.8517940717628705,
      "eval_dev1_recall": 0.8598425196850393,
      "eval_dev1_runtime": 0.8955,
      "eval_dev1_samples_per_second": 147.396,
      "eval_dev1_steps_per_second": 2.233,
      "step": 900
    }
  ],
  "max_steps": 1288,
  "num_train_epochs": 184,
  "total_flos": 2.481563978121216e+16,
  "trial_name": null,
  "trial_params": null
}