trainer_state.json

{
  "best_metric": 0.8809034907597536,
  "best_model_checkpoint": "bert-finetuned-sem_eval-english/checkpoint-470",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 940,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5829787234042553,
      "eval_f1_micro": 0.8580645161290322,
      "eval_f1_per_label": [
        0.927461139896373,
        0.9282051282051282,
        0.6093750000000001,
        0.0
      ],
      "eval_loss": 0.33245277404785156,
      "eval_runtime": 8.1576,
      "eval_samples_per_second": 28.807,
      "eval_steps_per_second": 3.678,
      "step": 235
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6127659574468085,
      "eval_f1_micro": 0.8809034907597536,
      "eval_f1_per_label": [
        0.9447236180904522,
        0.9432098765432099,
        0.6805555555555556,
        0.07407407407407407
      ],
      "eval_loss": 0.3222208321094513,
      "eval_runtime": 7.919,
      "eval_samples_per_second": 29.675,
      "eval_steps_per_second": 3.788,
      "step": 470
    },
    {
      "epoch": 2.13,
      "learning_rate": 1.1489361702127662e-05,
      "loss": 0.3325,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6042553191489362,
      "eval_f1_micro": 0.8737060041407867,
      "eval_f1_per_label": [
        0.9441624365482233,
        0.947103274559194,
        0.6074074074074073,
        0.35000000000000003
      ],
      "eval_loss": 0.3258354067802429,
      "eval_runtime": 7.9246,
      "eval_samples_per_second": 29.654,
      "eval_steps_per_second": 3.786,
      "step": 705
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6042553191489362,
      "eval_f1_micro": 0.8731808731808731,
      "eval_f1_per_label": [
        0.9329896907216495,
        0.9441624365482234,
        0.6428571428571428,
        0.4
      ],
      "eval_loss": 0.34061941504478455,
      "eval_runtime": 7.9355,
      "eval_samples_per_second": 29.614,
      "eval_steps_per_second": 3.78,
      "step": 940
    }
  ],
  "logging_steps": 500,
  "max_steps": 1175,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1976525740081152.0,
  "trial_name": null,
  "trial_params": null
}