trainer_state.json

{
  "best_metric": 0.9573363661766052,
  "best_model_checkpoint": "mobilebert_sa_GLUE_Experiment_logit_kd_qnli/checkpoint-1638",
  "epoch": 7.0,
  "global_step": 5733,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.9e-05,
      "loss": 1.0984,
      "step": 819
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6220025626944902,
      "eval_loss": 0.9626317620277405,
      "eval_runtime": 12.6845,
      "eval_samples_per_second": 430.682,
      "eval_steps_per_second": 3.39,
      "step": 819
    },
    {
      "epoch": 2.0,
      "learning_rate": 4.8e-05,
      "loss": 1.0171,
      "step": 1638
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.615595826468973,
      "eval_loss": 0.9573363661766052,
      "eval_runtime": 12.7142,
      "eval_samples_per_second": 429.676,
      "eval_steps_per_second": 3.382,
      "step": 1638
    },
    {
      "epoch": 3.0,
      "learning_rate": 4.7e-05,
      "loss": 0.9717,
      "step": 2457
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6104704374885594,
      "eval_loss": 0.9651486277580261,
      "eval_runtime": 12.7327,
      "eval_samples_per_second": 429.053,
      "eval_steps_per_second": 3.377,
      "step": 2457
    },
    {
      "epoch": 4.0,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.9377,
      "step": 3276
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6024162548050521,
      "eval_loss": 0.9713066220283508,
      "eval_runtime": 12.6949,
      "eval_samples_per_second": 430.33,
      "eval_steps_per_second": 3.387,
      "step": 3276
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.5e-05,
      "loss": 0.9132,
      "step": 4095
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5987552626761853,
      "eval_loss": 0.9811589121818542,
      "eval_runtime": 12.741,
      "eval_samples_per_second": 428.772,
      "eval_steps_per_second": 3.375,
      "step": 4095
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.89,
      "step": 4914
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5982061138568552,
      "eval_loss": 1.0107508897781372,
      "eval_runtime": 12.8729,
      "eval_samples_per_second": 424.379,
      "eval_steps_per_second": 3.34,
      "step": 4914
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.3e-05,
      "loss": 0.8683,
      "step": 5733
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5914332784184514,
      "eval_loss": 1.0290495157241821,
      "eval_runtime": 12.6935,
      "eval_samples_per_second": 430.378,
      "eval_steps_per_second": 3.388,
      "step": 5733
    },
    {
      "epoch": 7.0,
      "step": 5733,
      "total_flos": 2.2988974586855424e+16,
      "train_loss": 0.9566188341823108,
      "train_runtime": 4545.7624,
      "train_samples_per_second": 1152.095,
      "train_steps_per_second": 9.008
    }
  ],
  "max_steps": 40950,
  "num_train_epochs": 50,
  "total_flos": 2.2988974586855424e+16,
  "trial_name": null,
  "trial_params": null
}