trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5833033919334412,
  "best_model_checkpoint": "distilbert_rand_50_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": 1.0091755390167236,
      "learning_rate": 4.9e-05,
      "loss": 0.6343,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6691176470588235,
      "eval_combined_score": 0.7229885755955274,
      "eval_f1": 0.7768595041322314,
      "eval_loss": 0.617800235748291,
      "eval_runtime": 0.201,
      "eval_samples_per_second": 2029.418,
      "eval_steps_per_second": 9.948,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.0012052059173584,
      "learning_rate": 4.8e-05,
      "loss": 0.5869,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7107843137254902,
      "eval_combined_score": 0.7632046568627451,
      "eval_f1": 0.815625,
      "eval_loss": 0.5833033919334412,
      "eval_runtime": 0.2227,
      "eval_samples_per_second": 1832.339,
      "eval_steps_per_second": 8.982,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.9815465211868286,
      "learning_rate": 4.7e-05,
      "loss": 0.5196,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7107843137254902,
      "eval_combined_score": 0.753316032295271,
      "eval_f1": 0.7958477508650519,
      "eval_loss": 0.584040641784668,
      "eval_runtime": 0.2168,
      "eval_samples_per_second": 1881.987,
      "eval_steps_per_second": 9.225,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.6551523208618164,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4151,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6813725490196079,
      "eval_combined_score": 0.7305167829843802,
      "eval_f1": 0.7796610169491526,
      "eval_loss": 0.6887530088424683,
      "eval_runtime": 0.2094,
      "eval_samples_per_second": 1948.689,
      "eval_steps_per_second": 9.552,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.337549686431885,
      "learning_rate": 4.5e-05,
      "loss": 0.2688,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6764705882352942,
      "eval_combined_score": 0.7117985125084516,
      "eval_f1": 0.7471264367816092,
      "eval_loss": 0.9317456483840942,
      "eval_runtime": 0.18,
      "eval_samples_per_second": 2266.783,
      "eval_steps_per_second": 11.112,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.3622426986694336,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1493,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7034313725490197,
      "eval_combined_score": 0.746131044040653,
      "eval_f1": 0.7888307155322862,
      "eval_loss": 1.019219994544983,
      "eval_runtime": 0.183,
      "eval_samples_per_second": 2229.455,
      "eval_steps_per_second": 10.929,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.299592971801758,
      "learning_rate": 4.3e-05,
      "loss": 0.1019,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7361942564636474,
      "eval_f1": 0.7787610619469026,
      "eval_loss": 1.2262157201766968,
      "eval_runtime": 0.2083,
      "eval_samples_per_second": 1958.972,
      "eval_steps_per_second": 9.603,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.3822683777127947,
      "train_runtime": 35.0118,
      "train_samples_per_second": 5238.236,
      "train_steps_per_second": 21.421
    }
  ],
  "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
}