trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.6089968085289001,
  "best_model_checkpoint": "distilbert_km_100_v2_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": 1.2737231254577637,
      "learning_rate": 4.9e-05,
      "loss": 0.6234,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7107843137254902,
      "eval_combined_score": 0.7643427741466957,
      "eval_f1": 0.8179012345679012,
      "eval_loss": 0.6136570572853088,
      "eval_runtime": 0.1901,
      "eval_samples_per_second": 2146.595,
      "eval_steps_per_second": 10.523,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7474459409713745,
      "learning_rate": 4.8e-05,
      "loss": 0.563,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7034313725490197,
      "eval_combined_score": 0.7567392341552006,
      "eval_f1": 0.8100470957613815,
      "eval_loss": 0.6089968085289001,
      "eval_runtime": 0.2345,
      "eval_samples_per_second": 1739.797,
      "eval_steps_per_second": 8.528,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.482377052307129,
      "learning_rate": 4.7e-05,
      "loss": 0.4948,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.7511642156862746,
      "eval_f1": 0.80625,
      "eval_loss": 0.6385875940322876,
      "eval_runtime": 0.2287,
      "eval_samples_per_second": 1784.032,
      "eval_steps_per_second": 8.745,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.05997896194458,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4259,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6421568627450981,
      "eval_combined_score": 0.6911851929383853,
      "eval_f1": 0.7402135231316725,
      "eval_loss": 0.7081382870674133,
      "eval_runtime": 0.1985,
      "eval_samples_per_second": 2055.066,
      "eval_steps_per_second": 10.074,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.399949073791504,
      "learning_rate": 4.5e-05,
      "loss": 0.3198,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7360004346566213,
      "eval_f1": 0.7881773399014779,
      "eval_loss": 0.7598997950553894,
      "eval_runtime": 0.1869,
      "eval_samples_per_second": 2182.594,
      "eval_steps_per_second": 10.699,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.6178205013275146,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2082,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6372549019607843,
      "eval_combined_score": 0.6874217772215269,
      "eval_f1": 0.7375886524822695,
      "eval_loss": 0.9243190884590149,
      "eval_runtime": 0.1901,
      "eval_samples_per_second": 2146.589,
      "eval_steps_per_second": 10.522,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.4097371101379395,
      "learning_rate": 4.3e-05,
      "loss": 0.1214,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6323529411764706,
      "eval_combined_score": 0.6807974092163942,
      "eval_f1": 0.7292418772563177,
      "eval_loss": 1.1666536331176758,
      "eval_runtime": 0.1919,
      "eval_samples_per_second": 2126.115,
      "eval_steps_per_second": 10.422,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.3937856526601882,
      "train_runtime": 34.7605,
      "train_samples_per_second": 5276.107,
      "train_steps_per_second": 21.576
    }
  ],
  "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
}