trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7583626508712769,
  "best_model_checkpoint": "bert_base_km_10_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": 1.404510736465454,
      "learning_rate": 4.9e-05,
      "loss": 0.7364,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7583626508712769,
      "eval_runtime": 0.0559,
      "eval_samples_per_second": 1270.351,
      "eval_steps_per_second": 17.892,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.173837423324585,
      "learning_rate": 4.8e-05,
      "loss": 0.7014,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.7828453779220581,
      "eval_runtime": 0.0757,
      "eval_samples_per_second": 937.549,
      "eval_steps_per_second": 13.205,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0699000358581543,
      "learning_rate": 4.7e-05,
      "loss": 0.6842,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2535211267605634,
      "eval_loss": 0.797590434551239,
      "eval_runtime": 0.0577,
      "eval_samples_per_second": 1229.849,
      "eval_steps_per_second": 17.322,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.4305363893508911,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6794,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.835277259349823,
      "eval_runtime": 0.0582,
      "eval_samples_per_second": 1219.874,
      "eval_steps_per_second": 17.181,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.163173794746399,
      "learning_rate": 4.5e-05,
      "loss": 0.672,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.8817947506904602,
      "eval_runtime": 0.0586,
      "eval_samples_per_second": 1212.162,
      "eval_steps_per_second": 17.073,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.311958909034729,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6659,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.14084507042253522,
      "eval_loss": 0.9234703779220581,
      "eval_runtime": 0.059,
      "eval_samples_per_second": 1204.188,
      "eval_steps_per_second": 16.96,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6898842189047072,
      "train_runtime": 18.8121,
      "train_samples_per_second": 1687.742,
      "train_steps_per_second": 7.974
    }
  ],
  "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
}