trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.621903657913208,
  "best_model_checkpoint": "bert_base_rand_5_v1_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.6281776428222656,
      "learning_rate": 4.9e-05,
      "loss": 0.6626,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6309719934102141,
      "eval_loss": 0.6388838887214661,
      "eval_runtime": 3.5027,
      "eval_samples_per_second": 1559.642,
      "eval_steps_per_second": 6.281,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7621873617172241,
      "learning_rate": 4.8e-05,
      "loss": 0.6197,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6547684422478491,
      "eval_loss": 0.621903657913208,
      "eval_runtime": 3.4455,
      "eval_samples_per_second": 1585.567,
      "eval_steps_per_second": 6.385,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.546959638595581,
      "learning_rate": 4.7e-05,
      "loss": 0.5481,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.628775398132894,
      "eval_loss": 0.659508228302002,
      "eval_runtime": 3.4193,
      "eval_samples_per_second": 1597.682,
      "eval_steps_per_second": 6.434,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.314105987548828,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4397,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6633717737506865,
      "eval_loss": 0.7054057121276855,
      "eval_runtime": 3.4423,
      "eval_samples_per_second": 1587.009,
      "eval_steps_per_second": 6.391,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.697480201721191,
      "learning_rate": 4.5e-05,
      "loss": 0.3246,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6655683690280065,
      "eval_loss": 0.76524817943573,
      "eval_runtime": 3.4211,
      "eval_samples_per_second": 1596.873,
      "eval_steps_per_second": 6.431,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.7650370597839355,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2276,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6445176642870217,
      "eval_loss": 1.1442667245864868,
      "eval_runtime": 3.4209,
      "eval_samples_per_second": 1596.93,
      "eval_steps_per_second": 6.431,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 6.955177307128906,
      "learning_rate": 4.3e-05,
      "loss": 0.1661,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6576972359509428,
      "eval_loss": 0.9847047924995422,
      "eval_runtime": 3.4395,
      "eval_samples_per_second": 1588.318,
      "eval_steps_per_second": 6.396,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.426914212429565,
      "train_runtime": 1215.4609,
      "train_samples_per_second": 4308.777,
      "train_steps_per_second": 16.866
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}