trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5908915400505066,
  "best_model_checkpoint": "distilbert_km_5_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": 1.4459097385406494,
      "learning_rate": 4.9e-05,
      "loss": 0.625,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6936274509803921,
      "eval_combined_score": 0.7448561398458241,
      "eval_f1": 0.7960848287112561,
      "eval_loss": 0.6059701442718506,
      "eval_runtime": 0.1803,
      "eval_samples_per_second": 2263.29,
      "eval_steps_per_second": 11.095,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6643990278244019,
      "learning_rate": 4.8e-05,
      "loss": 0.5537,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7396930642304466,
      "eval_f1": 0.7955625990491284,
      "eval_loss": 0.5908915400505066,
      "eval_runtime": 0.1831,
      "eval_samples_per_second": 2228.868,
      "eval_steps_per_second": 10.926,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.8262536525726318,
      "learning_rate": 4.7e-05,
      "loss": 0.4677,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7083333333333334,
      "eval_combined_score": 0.7574186991869919,
      "eval_f1": 0.8065040650406504,
      "eval_loss": 0.6477960348129272,
      "eval_runtime": 0.1892,
      "eval_samples_per_second": 2156.987,
      "eval_steps_per_second": 10.573,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.214698076248169,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3598,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6544117647058824,
      "eval_combined_score": 0.7019838574861561,
      "eval_f1": 0.7495559502664298,
      "eval_loss": 0.7725249528884888,
      "eval_runtime": 0.1827,
      "eval_samples_per_second": 2233.575,
      "eval_steps_per_second": 10.949,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.364299297332764,
      "learning_rate": 4.5e-05,
      "loss": 0.2154,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5882352941176471,
      "eval_combined_score": 0.6362229102167183,
      "eval_f1": 0.6842105263157895,
      "eval_loss": 1.0573184490203857,
      "eval_runtime": 0.184,
      "eval_samples_per_second": 2217.931,
      "eval_steps_per_second": 10.872,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.5681135654449463,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1138,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6691176470588235,
      "eval_combined_score": 0.7183798218082413,
      "eval_f1": 0.7676419965576592,
      "eval_loss": 1.1862870454788208,
      "eval_runtime": 0.1845,
      "eval_samples_per_second": 2211.385,
      "eval_steps_per_second": 10.84,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.2547056674957275,
      "learning_rate": 4.3e-05,
      "loss": 0.0656,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6225490196078431,
      "eval_combined_score": 0.6789721043056398,
      "eval_f1": 0.7353951890034365,
      "eval_loss": 1.4886935949325562,
      "eval_runtime": 0.186,
      "eval_samples_per_second": 2193.601,
      "eval_steps_per_second": 10.753,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.34300908531461444,
      "train_runtime": 33.6908,
      "train_samples_per_second": 5443.63,
      "train_steps_per_second": 22.261
    }
  ],
  "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
}