trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.7017711997032166,
  "best_model_checkpoint": "tiny_bert_km_20_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": 0.647952675819397,
      "learning_rate": 4.9e-05,
      "loss": 0.7014,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 0.7017711997032166,
      "eval_runtime": 0.1047,
      "eval_samples_per_second": 2646.134,
      "eval_steps_per_second": 19.106,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.685315728187561,
      "learning_rate": 4.8e-05,
      "loss": 0.6872,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.48736462093862815,
      "eval_loss": 0.7077716588973999,
      "eval_runtime": 0.1007,
      "eval_samples_per_second": 2751.875,
      "eval_steps_per_second": 19.869,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6586238741874695,
      "learning_rate": 4.7e-05,
      "loss": 0.6752,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 0.7020250558853149,
      "eval_runtime": 0.103,
      "eval_samples_per_second": 2689.665,
      "eval_steps_per_second": 19.42,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6234502196311951,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6556,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.48375451263537905,
      "eval_loss": 0.7177628874778748,
      "eval_runtime": 0.0923,
      "eval_samples_per_second": 3002.621,
      "eval_steps_per_second": 21.68,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.896072268486023,
      "learning_rate": 4.5e-05,
      "loss": 0.6267,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5342960288808665,
      "eval_loss": 0.7247292399406433,
      "eval_runtime": 0.1089,
      "eval_samples_per_second": 2544.023,
      "eval_steps_per_second": 18.368,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9845135807991028,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5808,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.48014440433212996,
      "eval_loss": 0.7734146118164062,
      "eval_runtime": 0.0984,
      "eval_samples_per_second": 2813.735,
      "eval_steps_per_second": 20.316,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.654466986656189,
      "train_runtime": 13.2472,
      "train_samples_per_second": 9398.208,
      "train_steps_per_second": 37.744
    }
  ],
  "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
}