trainer_state.json

{
  "best_metric": 0.5771955847740173,
  "best_model_checkpoint": "hBERTv2_mrpc/checkpoint-45",
  "epoch": 8.0,
  "global_step": 120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.906666666666667e-05,
      "loss": 0.6388,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7480253018237863,
      "eval_f1": 0.8122270742358079,
      "eval_loss": 0.6297235488891602,
      "eval_runtime": 0.4849,
      "eval_samples_per_second": 841.375,
      "eval_steps_per_second": 4.124,
      "step": 15
    },
    {
      "epoch": 2.0,
      "learning_rate": 4.806666666666667e-05,
      "loss": 0.612,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6887254901960784,
      "eval_combined_score": 0.7511175174915782,
      "eval_f1": 0.8135095447870779,
      "eval_loss": 0.631518542766571,
      "eval_runtime": 0.4881,
      "eval_samples_per_second": 835.824,
      "eval_steps_per_second": 4.097,
      "step": 30
    },
    {
      "epoch": 3.0,
      "learning_rate": 4.706666666666667e-05,
      "loss": 0.5725,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7511016274810076,
      "eval_f1": 0.8085758039816232,
      "eval_loss": 0.5771955847740173,
      "eval_runtime": 0.4811,
      "eval_samples_per_second": 847.98,
      "eval_steps_per_second": 4.157,
      "step": 45
    },
    {
      "epoch": 4.0,
      "learning_rate": 4.606666666666667e-05,
      "loss": 0.512,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7580659536541889,
      "eval_f1": 0.8151515151515151,
      "eval_loss": 0.6260703206062317,
      "eval_runtime": 0.4792,
      "eval_samples_per_second": 851.387,
      "eval_steps_per_second": 4.173,
      "step": 60
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.5066666666666667e-05,
      "loss": 0.3924,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7279411764705882,
      "eval_combined_score": 0.7737266857962697,
      "eval_f1": 0.8195121951219512,
      "eval_loss": 0.6433467268943787,
      "eval_runtime": 0.4808,
      "eval_samples_per_second": 848.666,
      "eval_steps_per_second": 4.16,
      "step": 75
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.406666666666667e-05,
      "loss": 0.2592,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6862745098039216,
      "eval_combined_score": 0.7228365030222615,
      "eval_f1": 0.7593984962406015,
      "eval_loss": 0.7530680894851685,
      "eval_runtime": 0.4847,
      "eval_samples_per_second": 841.795,
      "eval_steps_per_second": 4.126,
      "step": 90
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.3066666666666665e-05,
      "loss": 0.1689,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7377450980392157,
      "eval_combined_score": 0.7767899328406062,
      "eval_f1": 0.8158347676419966,
      "eval_loss": 0.7904398441314697,
      "eval_runtime": 0.4827,
      "eval_samples_per_second": 845.31,
      "eval_steps_per_second": 4.144,
      "step": 105
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.206666666666667e-05,
      "loss": 0.1292,
      "step": 120
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7622549019607843,
      "eval_combined_score": 0.8001591705129464,
      "eval_f1": 0.8380634390651085,
      "eval_loss": 0.9954336285591125,
      "eval_runtime": 0.4833,
      "eval_samples_per_second": 844.146,
      "eval_steps_per_second": 4.138,
      "step": 120
    },
    {
      "epoch": 8.0,
      "step": 120,
      "total_flos": 3654895836069888.0,
      "train_loss": 0.4106310119231542,
      "train_runtime": 160.8809,
      "train_samples_per_second": 1139.973,
      "train_steps_per_second": 4.662
    }
  ],
  "max_steps": 750,
  "num_train_epochs": 50,
  "total_flos": 3654895836069888.0,
  "trial_name": null,
  "trial_params": null
}