trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.7080254554748535,
  "best_model_checkpoint": "bert_base_km_50_v2_rte/checkpoint-20",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 70,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.3552422523498535,
      "learning_rate": 4.9e-05,
      "loss": 0.7063,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.49097472924187724,
      "eval_loss": 0.7112971544265747,
      "eval_runtime": 0.1885,
      "eval_samples_per_second": 1469.814,
      "eval_steps_per_second": 10.612,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8892113566398621,
      "learning_rate": 4.8e-05,
      "loss": 0.6591,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5342960288808665,
      "eval_loss": 0.7080254554748535,
      "eval_runtime": 0.1969,
      "eval_samples_per_second": 1407.13,
      "eval_steps_per_second": 10.16,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0309181213378906,
      "learning_rate": 4.7e-05,
      "loss": 0.618,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5018050541516246,
      "eval_loss": 0.7251981496810913,
      "eval_runtime": 0.2062,
      "eval_samples_per_second": 1343.548,
      "eval_steps_per_second": 9.701,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.105635643005371,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5474,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.516245487364621,
      "eval_loss": 0.7643619775772095,
      "eval_runtime": 0.2054,
      "eval_samples_per_second": 1348.411,
      "eval_steps_per_second": 9.736,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.6433978080749512,
      "learning_rate": 4.5e-05,
      "loss": 0.4529,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5270758122743683,
      "eval_loss": 0.865513801574707,
      "eval_runtime": 0.2028,
      "eval_samples_per_second": 1365.882,
      "eval_steps_per_second": 9.862,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.9883702993392944,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3237,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 1.0289877653121948,
      "eval_runtime": 0.196,
      "eval_samples_per_second": 1413.294,
      "eval_steps_per_second": 10.204,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.0816502571105957,
      "learning_rate": 4.3e-05,
      "loss": 0.2168,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 1.3162751197814941,
      "eval_runtime": 0.1965,
      "eval_samples_per_second": 1409.347,
      "eval_steps_per_second": 10.176,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 2293012847462400.0,
      "train_loss": 0.5034336362566266,
      "train_runtime": 44.046,
      "train_samples_per_second": 2826.591,
      "train_steps_per_second": 11.352
    }
  ],
  "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": 2293012847462400.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}