trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5870324969291687,
  "best_model_checkpoint": "distilbert_rand_5_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": 1.2852423191070557,
      "learning_rate": 4.9e-05,
      "loss": 0.6303,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7483885286398024,
      "eval_f1": 0.8031496062992126,
      "eval_loss": 0.5948174595832825,
      "eval_runtime": 0.1837,
      "eval_samples_per_second": 2220.974,
      "eval_steps_per_second": 10.887,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7043911218643188,
      "learning_rate": 4.8e-05,
      "loss": 0.5774,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.696078431372549,
      "eval_combined_score": 0.7470620169892062,
      "eval_f1": 0.7980456026058632,
      "eval_loss": 0.5870324969291687,
      "eval_runtime": 0.1812,
      "eval_samples_per_second": 2252.085,
      "eval_steps_per_second": 11.04,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.504751443862915,
      "learning_rate": 4.7e-05,
      "loss": 0.5136,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7477645971225257,
      "eval_f1": 0.8019017432646592,
      "eval_loss": 0.6408395171165466,
      "eval_runtime": 0.1792,
      "eval_samples_per_second": 2276.717,
      "eval_steps_per_second": 11.16,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.745375633239746,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4314,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6764705882352942,
      "eval_combined_score": 0.7207975716977182,
      "eval_f1": 0.7651245551601423,
      "eval_loss": 0.7089741826057434,
      "eval_runtime": 0.2002,
      "eval_samples_per_second": 2038.269,
      "eval_steps_per_second": 9.992,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.8098981380462646,
      "learning_rate": 4.5e-05,
      "loss": 0.2865,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6715686274509803,
      "eval_combined_score": 0.7194648692810457,
      "eval_f1": 0.7673611111111112,
      "eval_loss": 0.9271292686462402,
      "eval_runtime": 0.1887,
      "eval_samples_per_second": 2161.69,
      "eval_steps_per_second": 10.597,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.68713116645813,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1755,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6053921568627451,
      "eval_combined_score": 0.654989656413024,
      "eval_f1": 0.7045871559633028,
      "eval_loss": 1.239139437675476,
      "eval_runtime": 0.1836,
      "eval_samples_per_second": 2222.076,
      "eval_steps_per_second": 10.893,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 6.261050224304199,
      "learning_rate": 4.3e-05,
      "loss": 0.0968,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6348039215686274,
      "eval_combined_score": 0.6821932493505569,
      "eval_f1": 0.7295825771324864,
      "eval_loss": 1.5552704334259033,
      "eval_runtime": 0.1955,
      "eval_samples_per_second": 2087.339,
      "eval_steps_per_second": 10.232,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.38735065006074454,
      "train_runtime": 34.3099,
      "train_samples_per_second": 5345.391,
      "train_steps_per_second": 21.86
    }
  ],
  "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
}