trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7320093512535095,
  "best_model_checkpoint": "bert_base_km_20_v2_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": 1.5461989641189575,
      "learning_rate": 4.9e-05,
      "loss": 0.7432,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.7320093512535095,
      "eval_runtime": 0.0646,
      "eval_samples_per_second": 1099.834,
      "eval_steps_per_second": 15.491,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.3384795188903809,
      "learning_rate": 4.8e-05,
      "loss": 0.7018,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.3380281690140845,
      "eval_loss": 0.7807548642158508,
      "eval_runtime": 0.0611,
      "eval_samples_per_second": 1162.724,
      "eval_steps_per_second": 16.376,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.1635830402374268,
      "learning_rate": 4.7e-05,
      "loss": 0.6914,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2676056338028169,
      "eval_loss": 0.7764084935188293,
      "eval_runtime": 0.0571,
      "eval_samples_per_second": 1244.263,
      "eval_steps_per_second": 17.525,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.7215803861618042,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6971,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3380281690140845,
      "eval_loss": 0.796627402305603,
      "eval_runtime": 0.0666,
      "eval_samples_per_second": 1066.783,
      "eval_steps_per_second": 15.025,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.167963981628418,
      "learning_rate": 4.5e-05,
      "loss": 0.6782,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.23943661971830985,
      "eval_loss": 0.8304632902145386,
      "eval_runtime": 0.068,
      "eval_samples_per_second": 1043.601,
      "eval_steps_per_second": 14.699,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.5520793199539185,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6864,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.8585221171379089,
      "eval_runtime": 0.0608,
      "eval_samples_per_second": 1167.313,
      "eval_steps_per_second": 16.441,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6996832953559028,
      "train_runtime": 18.324,
      "train_samples_per_second": 1732.698,
      "train_steps_per_second": 8.186
    }
  ],
  "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
}