trainer_state.json

{
  "best_metric": 0.9620120357508649,
  "best_model_checkpoint": "rubert_classification/checkpoint-2000",
  "epoch": 4.566210045662101,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.14,
      "eval_accuracy": 0.882,
      "eval_f1": 0.8860252812063524,
      "eval_loss": 0.3335667848587036,
      "eval_precision": 0.9101706722396378,
      "eval_recall": 0.882,
      "eval_runtime": 7.9932,
      "eval_samples_per_second": 187.659,
      "eval_steps_per_second": 11.76,
      "step": 500
    },
    {
      "epoch": 1.37,
      "grad_norm": 54.402042388916016,
      "learning_rate": 4.871134020618557e-05,
      "loss": 0.3238,
      "step": 600
    },
    {
      "epoch": 2.28,
      "eval_accuracy": 0.952,
      "eval_f1": 0.952629531198857,
      "eval_loss": 0.22119292616844177,
      "eval_precision": 0.955250771109112,
      "eval_recall": 0.952,
      "eval_runtime": 7.4028,
      "eval_samples_per_second": 202.626,
      "eval_steps_per_second": 12.698,
      "step": 1000
    },
    {
      "epoch": 2.74,
      "grad_norm": 0.6686217188835144,
      "learning_rate": 4.097938144329897e-05,
      "loss": 0.1138,
      "step": 1200
    },
    {
      "epoch": 3.42,
      "eval_accuracy": 0.9586666666666667,
      "eval_f1": 0.9584488790426743,
      "eval_loss": 0.19180671870708466,
      "eval_precision": 0.9584916263909182,
      "eval_recall": 0.9586666666666667,
      "eval_runtime": 7.383,
      "eval_samples_per_second": 203.169,
      "eval_steps_per_second": 12.732,
      "step": 1500
    },
    {
      "epoch": 4.11,
      "grad_norm": 0.031083475798368454,
      "learning_rate": 3.3247422680412374e-05,
      "loss": 0.0394,
      "step": 1800
    },
    {
      "epoch": 4.57,
      "eval_accuracy": 0.962,
      "eval_f1": 0.9620120357508649,
      "eval_loss": 0.2063756138086319,
      "eval_precision": 0.9620250729762693,
      "eval_recall": 0.962,
      "eval_runtime": 7.384,
      "eval_samples_per_second": 203.142,
      "eval_steps_per_second": 12.73,
      "step": 2000
    }
  ],
  "logging_steps": 600,
  "max_steps": 4380,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 936301011222528.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}