trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.7195115089416504,
  "best_model_checkpoint": "bert_base_km_100_v1_rte/checkpoint-10",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.65757155418396,
      "learning_rate": 4.9e-05,
      "loss": 0.7095,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4584837545126354,
      "eval_loss": 0.7195115089416504,
      "eval_runtime": 0.1883,
      "eval_samples_per_second": 1471.352,
      "eval_steps_per_second": 10.623,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.0872938632965088,
      "learning_rate": 4.8e-05,
      "loss": 0.6491,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5018050541516246,
      "eval_loss": 0.7259278893470764,
      "eval_runtime": 0.2024,
      "eval_samples_per_second": 1368.439,
      "eval_steps_per_second": 9.88,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.167483925819397,
      "learning_rate": 4.7e-05,
      "loss": 0.5867,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5018050541516246,
      "eval_loss": 0.7476696372032166,
      "eval_runtime": 0.201,
      "eval_samples_per_second": 1377.994,
      "eval_steps_per_second": 9.949,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.5926759243011475,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4852,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.516245487364621,
      "eval_loss": 0.8246937990188599,
      "eval_runtime": 0.1909,
      "eval_samples_per_second": 1450.829,
      "eval_steps_per_second": 10.475,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.8505879640579224,
      "learning_rate": 4.5e-05,
      "loss": 0.3453,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 0.9442667365074158,
      "eval_runtime": 0.1922,
      "eval_samples_per_second": 1441.318,
      "eval_steps_per_second": 10.407,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.3007497787475586,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2103,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5234657039711191,
      "eval_loss": 1.161955714225769,
      "eval_runtime": 0.2105,
      "eval_samples_per_second": 1316.183,
      "eval_steps_per_second": 9.503,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 1965439583539200.0,
      "train_loss": 0.49767692883809406,
      "train_runtime": 37.124,
      "train_samples_per_second": 3353.625,
      "train_steps_per_second": 13.468
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 1965439583539200.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}