trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7175397872924805,
  "best_model_checkpoint": "tiny_bert_km_10_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": 1.0948786735534668,
      "learning_rate": 4.9e-05,
      "loss": 0.7067,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.7175397872924805,
      "eval_runtime": 0.0302,
      "eval_samples_per_second": 2348.898,
      "eval_steps_per_second": 33.083,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6754629611968994,
      "learning_rate": 4.8e-05,
      "loss": 0.6961,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7197404503822327,
      "eval_runtime": 0.0538,
      "eval_samples_per_second": 1319.916,
      "eval_steps_per_second": 18.59,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.7125653028488159,
      "learning_rate": 4.7e-05,
      "loss": 0.695,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7329993844032288,
      "eval_runtime": 0.0308,
      "eval_samples_per_second": 2303.083,
      "eval_steps_per_second": 32.438,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.7750300168991089,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6919,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.7427927851676941,
      "eval_runtime": 0.0313,
      "eval_samples_per_second": 2266.863,
      "eval_steps_per_second": 31.928,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7668715715408325,
      "learning_rate": 4.5e-05,
      "loss": 0.6871,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.7530810236930847,
      "eval_runtime": 0.0411,
      "eval_samples_per_second": 1727.343,
      "eval_steps_per_second": 24.329,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.7905510663986206,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6837,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.7641945481300354,
      "eval_runtime": 0.0317,
      "eval_samples_per_second": 2241.188,
      "eval_steps_per_second": 31.566,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 99911832606720.0,
      "train_loss": 0.6934275759590997,
      "train_runtime": 6.3239,
      "train_samples_per_second": 5020.637,
      "train_steps_per_second": 23.72
    }
  ],
  "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": 99911832606720.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}