trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.736080527305603,
  "best_model_checkpoint": "bert_base_km_100_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": 1.4661955833435059,
      "learning_rate": 4.9e-05,
      "loss": 0.7334,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.36619718309859156,
      "eval_loss": 0.736080527305603,
      "eval_runtime": 0.0572,
      "eval_samples_per_second": 1240.799,
      "eval_steps_per_second": 17.476,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.1175402402877808,
      "learning_rate": 4.8e-05,
      "loss": 0.698,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.30985915492957744,
      "eval_loss": 0.7782241106033325,
      "eval_runtime": 0.0594,
      "eval_samples_per_second": 1195.529,
      "eval_steps_per_second": 16.838,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0362004041671753,
      "learning_rate": 4.7e-05,
      "loss": 0.6871,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.30985915492957744,
      "eval_loss": 0.7843859791755676,
      "eval_runtime": 0.0593,
      "eval_samples_per_second": 1196.466,
      "eval_steps_per_second": 16.852,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.6865355968475342,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6893,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.2676056338028169,
      "eval_loss": 0.8130501508712769,
      "eval_runtime": 0.0574,
      "eval_samples_per_second": 1238.004,
      "eval_steps_per_second": 17.437,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.2285587787628174,
      "learning_rate": 4.5e-05,
      "loss": 0.6813,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.8521126508712769,
      "eval_runtime": 0.0843,
      "eval_samples_per_second": 841.82,
      "eval_steps_per_second": 11.857,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.385190486907959,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6797,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.877145528793335,
      "eval_runtime": 0.0682,
      "eval_samples_per_second": 1041.404,
      "eval_steps_per_second": 14.668,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.694816443655226,
      "train_runtime": 18.6005,
      "train_samples_per_second": 1706.946,
      "train_steps_per_second": 8.064
    }
  ],
  "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
}