trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6899185180664062,
  "best_model_checkpoint": "tiny_bert_rand_5_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": 0.5479128360748291,
      "learning_rate": 4.9e-05,
      "loss": 0.6997,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5523465703971119,
      "eval_loss": 0.6899185180664062,
      "eval_runtime": 0.0934,
      "eval_samples_per_second": 2965.708,
      "eval_steps_per_second": 21.413,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6280381679534912,
      "learning_rate": 4.8e-05,
      "loss": 0.6781,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.555956678700361,
      "eval_loss": 0.6927812099456787,
      "eval_runtime": 0.1326,
      "eval_samples_per_second": 2088.94,
      "eval_steps_per_second": 15.083,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6833863854408264,
      "learning_rate": 4.7e-05,
      "loss": 0.6539,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.51985559566787,
      "eval_loss": 0.7066504955291748,
      "eval_runtime": 0.095,
      "eval_samples_per_second": 2917.297,
      "eval_steps_per_second": 21.064,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.8935176730155945,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6152,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 0.7462294697761536,
      "eval_runtime": 0.1057,
      "eval_samples_per_second": 2621.334,
      "eval_steps_per_second": 18.927,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.9902224540710449,
      "learning_rate": 4.5e-05,
      "loss": 0.5575,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5379061371841155,
      "eval_loss": 0.8307779431343079,
      "eval_runtime": 0.0985,
      "eval_samples_per_second": 2813.599,
      "eval_steps_per_second": 20.315,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.350633978843689,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.4749,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4729241877256318,
      "eval_loss": 0.9429054856300354,
      "eval_runtime": 0.1071,
      "eval_samples_per_second": 2587.497,
      "eval_steps_per_second": 18.682,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6131966511408488,
      "train_runtime": 13.5135,
      "train_samples_per_second": 9212.986,
      "train_steps_per_second": 37.0
    }
  ],
  "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
}