trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6832764744758606,
  "best_model_checkpoint": "tiny_bert_km_50_v2_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": 0.583885908126831,
      "learning_rate": 4.9e-05,
      "loss": 0.6977,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5451263537906137,
      "eval_loss": 0.6832764744758606,
      "eval_runtime": 0.0962,
      "eval_samples_per_second": 2877.942,
      "eval_steps_per_second": 20.779,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5978738069534302,
      "learning_rate": 4.8e-05,
      "loss": 0.6792,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5667870036101083,
      "eval_loss": 0.689453125,
      "eval_runtime": 0.0949,
      "eval_samples_per_second": 2918.0,
      "eval_steps_per_second": 21.069,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6238811016082764,
      "learning_rate": 4.7e-05,
      "loss": 0.6629,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.555956678700361,
      "eval_loss": 0.6894389390945435,
      "eval_runtime": 0.118,
      "eval_samples_per_second": 2348.022,
      "eval_steps_per_second": 16.953,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.627923309803009,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6386,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5523465703971119,
      "eval_loss": 0.695397138595581,
      "eval_runtime": 0.0928,
      "eval_samples_per_second": 2986.321,
      "eval_steps_per_second": 21.562,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8814382553100586,
      "learning_rate": 4.5e-05,
      "loss": 0.6076,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.555956678700361,
      "eval_loss": 0.7125662565231323,
      "eval_runtime": 0.0906,
      "eval_samples_per_second": 3058.94,
      "eval_steps_per_second": 22.086,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9271029829978943,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5662,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5487364620938628,
      "eval_loss": 0.7351065278053284,
      "eval_runtime": 0.0998,
      "eval_samples_per_second": 2775.183,
      "eval_steps_per_second": 20.037,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6420334180196127,
      "train_runtime": 12.7341,
      "train_samples_per_second": 9776.895,
      "train_steps_per_second": 39.265
    }
  ],
  "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": 391780256993280.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}