trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5918853878974915,
  "best_model_checkpoint": "distilbert_rand_100_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": 1.0201210975646973,
      "learning_rate": 4.9e-05,
      "loss": 0.6255,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6495098039215687,
      "eval_combined_score": 0.7053892926118694,
      "eval_f1": 0.7612687813021702,
      "eval_loss": 0.6193369030952454,
      "eval_runtime": 0.1782,
      "eval_samples_per_second": 2289.4,
      "eval_steps_per_second": 11.223,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9219220876693726,
      "learning_rate": 4.8e-05,
      "loss": 0.5796,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6887254901960784,
      "eval_combined_score": 0.7421083167566543,
      "eval_f1": 0.7954911433172303,
      "eval_loss": 0.5918853878974915,
      "eval_runtime": 0.185,
      "eval_samples_per_second": 2205.116,
      "eval_steps_per_second": 10.809,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.2440578937530518,
      "learning_rate": 4.7e-05,
      "loss": 0.5137,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6715686274509803,
      "eval_combined_score": 0.7280672719248471,
      "eval_f1": 0.7845659163987139,
      "eval_loss": 0.6295387148857117,
      "eval_runtime": 0.2093,
      "eval_samples_per_second": 1949.483,
      "eval_steps_per_second": 9.556,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.9403421878814697,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4067,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.625,
      "eval_combined_score": 0.6651011560693642,
      "eval_f1": 0.7052023121387283,
      "eval_loss": 0.7785881161689758,
      "eval_runtime": 0.1907,
      "eval_samples_per_second": 2139.721,
      "eval_steps_per_second": 10.489,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.3565924167633057,
      "learning_rate": 4.5e-05,
      "loss": 0.2548,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6446078431372549,
      "eval_combined_score": 0.6897627881134172,
      "eval_f1": 0.7349177330895795,
      "eval_loss": 1.0053613185882568,
      "eval_runtime": 0.1814,
      "eval_samples_per_second": 2249.35,
      "eval_steps_per_second": 11.026,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 9.029556274414062,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1579,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6225490196078431,
      "eval_combined_score": 0.6722853401288313,
      "eval_f1": 0.7220216606498195,
      "eval_loss": 1.3866548538208008,
      "eval_runtime": 0.1832,
      "eval_samples_per_second": 2227.461,
      "eval_steps_per_second": 10.919,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 10.428647994995117,
      "learning_rate": 4.3e-05,
      "loss": 0.1051,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6078431372549019,
      "eval_combined_score": 0.6518303138745992,
      "eval_f1": 0.6958174904942965,
      "eval_loss": 1.442420244216919,
      "eval_runtime": 0.186,
      "eval_samples_per_second": 2193.21,
      "eval_steps_per_second": 10.751,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.3776343254815964,
      "train_runtime": 34.0728,
      "train_samples_per_second": 5382.599,
      "train_steps_per_second": 22.012
    }
  ],
  "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
}