trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.6028358340263367,
  "best_model_checkpoint": "distilbert_km_100_v1_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.292992115020752,
      "learning_rate": 4.9e-05,
      "loss": 0.6209,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6985294117647058,
      "eval_combined_score": 0.7542106100554595,
      "eval_f1": 0.8098918083462133,
      "eval_loss": 0.6127546429634094,
      "eval_runtime": 0.1878,
      "eval_samples_per_second": 2172.001,
      "eval_steps_per_second": 10.647,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.63740074634552,
      "learning_rate": 4.8e-05,
      "loss": 0.5773,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7577850289051791,
      "eval_f1": 0.8145896656534954,
      "eval_loss": 0.6028358340263367,
      "eval_runtime": 0.189,
      "eval_samples_per_second": 2158.195,
      "eval_steps_per_second": 10.579,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.5850547552108765,
      "learning_rate": 4.7e-05,
      "loss": 0.5229,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7107843137254902,
      "eval_combined_score": 0.7640608875128999,
      "eval_f1": 0.8173374613003096,
      "eval_loss": 0.6116268038749695,
      "eval_runtime": 0.1889,
      "eval_samples_per_second": 2159.655,
      "eval_steps_per_second": 10.587,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.5835577249526978,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.466,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6764705882352942,
      "eval_combined_score": 0.7228506787330318,
      "eval_f1": 0.7692307692307693,
      "eval_loss": 0.6474385261535645,
      "eval_runtime": 0.2034,
      "eval_samples_per_second": 2005.884,
      "eval_steps_per_second": 9.833,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.9144692420959473,
      "learning_rate": 4.5e-05,
      "loss": 0.3626,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6715686274509803,
      "eval_combined_score": 0.718651446592623,
      "eval_f1": 0.7657342657342657,
      "eval_loss": 0.6890015602111816,
      "eval_runtime": 0.1869,
      "eval_samples_per_second": 2183.112,
      "eval_steps_per_second": 10.702,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.1984636783599854,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2312,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5980392156862745,
      "eval_combined_score": 0.6419314852377733,
      "eval_f1": 0.685823754789272,
      "eval_loss": 0.9257688522338867,
      "eval_runtime": 0.1842,
      "eval_samples_per_second": 2214.734,
      "eval_steps_per_second": 10.857,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.930481433868408,
      "learning_rate": 4.3e-05,
      "loss": 0.1407,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6299019607843137,
      "eval_combined_score": 0.6753946033126744,
      "eval_f1": 0.7208872458410351,
      "eval_loss": 1.0241893529891968,
      "eval_runtime": 0.1836,
      "eval_samples_per_second": 2221.784,
      "eval_steps_per_second": 10.891,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.41734157516842796,
      "train_runtime": 34.77,
      "train_samples_per_second": 5274.666,
      "train_steps_per_second": 21.57
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1700616463945728.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}