trainer_state.json

{
  "best_metric": 96.7,
  "best_model_checkpoint": "outputs/bitfit/t5-base/sst2/checkpoint-500",
  "epoch": 3.0,
  "global_step": 1992,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.75,
      "learning_rate": 0.0003,
      "loss": 1.2354,
      "step": 500
    },
    {
      "epoch": 0.75,
      "eval_accuracy": 96.7,
      "eval_average_metrics": 96.7,
      "eval_loss": 0.044097092002630234,
      "eval_runtime": 2.8987,
      "eval_samples_per_second": 344.985,
      "step": 500
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.00019946380697050936,
      "loss": 0.0704,
      "step": 1000
    },
    {
      "epoch": 1.51,
      "eval_accuracy": 96.5,
      "eval_average_metrics": 96.5,
      "eval_loss": 0.042973704636096954,
      "eval_runtime": 2.8885,
      "eval_samples_per_second": 346.198,
      "step": 1000
    },
    {
      "epoch": 2.26,
      "learning_rate": 9.892761394101876e-05,
      "loss": 0.0651,
      "step": 1500
    },
    {
      "epoch": 2.26,
      "eval_accuracy": 96.5,
      "eval_average_metrics": 96.5,
      "eval_loss": 0.04074199125170708,
      "eval_runtime": 2.8768,
      "eval_samples_per_second": 347.613,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "step": 1992,
      "total_flos": 1.5337491316024032e+16,
      "train_loss": 0.3596552710935294,
      "train_runtime": 685.371,
      "train_samples_per_second": 290.422,
      "train_steps_per_second": 2.906
    }
  ],
  "max_steps": 1992,
  "num_train_epochs": 3,
  "total_flos": 1.5337491316024032e+16,
  "trial_name": null,
  "trial_params": null
}