trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7429577708244324,
  "best_model_checkpoint": "bert_base_km_50_v1_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4455351829528809,
      "learning_rate": 4.9e-05,
      "loss": 0.7348,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7429577708244324,
      "eval_runtime": 0.0589,
      "eval_samples_per_second": 1205.109,
      "eval_steps_per_second": 16.973,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2294955253601074,
      "learning_rate": 4.8e-05,
      "loss": 0.6984,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.23943661971830985,
      "eval_loss": 0.7746477127075195,
      "eval_runtime": 0.0581,
      "eval_samples_per_second": 1221.725,
      "eval_steps_per_second": 17.207,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0488471984863281,
      "learning_rate": 4.7e-05,
      "loss": 0.6854,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.29577464788732394,
      "eval_loss": 0.7890625,
      "eval_runtime": 0.058,
      "eval_samples_per_second": 1223.261,
      "eval_steps_per_second": 17.229,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.6488014459609985,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6873,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.8265843987464905,
      "eval_runtime": 0.0559,
      "eval_samples_per_second": 1270.258,
      "eval_steps_per_second": 17.891,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.192447304725647,
      "learning_rate": 4.5e-05,
      "loss": 0.6765,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.879374086856842,
      "eval_runtime": 0.0552,
      "eval_samples_per_second": 1285.269,
      "eval_steps_per_second": 18.102,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.3166923522949219,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6781,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.11267605633802817,
      "eval_loss": 0.9071303606033325,
      "eval_runtime": 0.0583,
      "eval_samples_per_second": 1217.311,
      "eval_steps_per_second": 17.145,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6933985948562622,
      "train_runtime": 18.2522,
      "train_samples_per_second": 1739.515,
      "train_steps_per_second": 8.218
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 501226560460800.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}