trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5951700210571289,
  "best_model_checkpoint": "distilbert_rand_20_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": 1.2904847860336304,
      "learning_rate": 4.9e-05,
      "loss": 0.6308,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_combined_score": 0.7207505518763797,
      "eval_f1": 0.7748344370860927,
      "eval_loss": 0.6188725233078003,
      "eval_runtime": 0.1815,
      "eval_samples_per_second": 2248.543,
      "eval_steps_per_second": 11.022,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9911153316497803,
      "learning_rate": 4.8e-05,
      "loss": 0.5808,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.7391237745098039,
      "eval_f1": 0.796875,
      "eval_loss": 0.5951700210571289,
      "eval_runtime": 0.1838,
      "eval_samples_per_second": 2219.839,
      "eval_steps_per_second": 10.882,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.473844289779663,
      "learning_rate": 4.7e-05,
      "loss": 0.5133,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6887254901960784,
      "eval_combined_score": 0.7417779308815611,
      "eval_f1": 0.7948303715670436,
      "eval_loss": 0.6181560158729553,
      "eval_runtime": 0.1909,
      "eval_samples_per_second": 2137.799,
      "eval_steps_per_second": 10.479,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.2987751960754395,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4133,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.7286173089925625,
      "eval_f1": 0.7758620689655172,
      "eval_loss": 0.7322428822517395,
      "eval_runtime": 0.1825,
      "eval_samples_per_second": 2235.151,
      "eval_steps_per_second": 10.957,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.5564703941345215,
      "learning_rate": 4.5e-05,
      "loss": 0.273,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6323529411764706,
      "eval_combined_score": 0.6719457013574661,
      "eval_f1": 0.7115384615384616,
      "eval_loss": 0.9913259148597717,
      "eval_runtime": 0.1944,
      "eval_samples_per_second": 2098.901,
      "eval_steps_per_second": 10.289,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.34869384765625,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1681,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6274509803921569,
      "eval_combined_score": 0.6686873222571471,
      "eval_f1": 0.7099236641221374,
      "eval_loss": 1.115963339805603,
      "eval_runtime": 0.2065,
      "eval_samples_per_second": 1975.458,
      "eval_steps_per_second": 9.684,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.789604187011719,
      "learning_rate": 4.3e-05,
      "loss": 0.1071,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6519607843137255,
      "eval_combined_score": 0.7039872650090964,
      "eval_f1": 0.7560137457044673,
      "eval_loss": 1.3771166801452637,
      "eval_runtime": 0.1851,
      "eval_samples_per_second": 2203.628,
      "eval_steps_per_second": 10.802,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.38379019101460776,
      "train_runtime": 34.8006,
      "train_samples_per_second": 5270.022,
      "train_steps_per_second": 21.551
    }
  ],
  "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
}