trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.6947072148323059,
  "best_model_checkpoint": "tiny_bert_rand_10_v1_wnli/checkpoint-6",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.7842780351638794,
      "learning_rate": 4.9e-05,
      "loss": 0.7108,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7168795466423035,
      "eval_runtime": 0.0314,
      "eval_samples_per_second": 2262.266,
      "eval_steps_per_second": 31.863,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.48126351833343506,
      "learning_rate": 4.8e-05,
      "loss": 0.6979,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6947072148323059,
      "eval_runtime": 0.0319,
      "eval_samples_per_second": 2226.259,
      "eval_steps_per_second": 31.356,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.573049008846283,
      "learning_rate": 4.7e-05,
      "loss": 0.6986,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6998240351676941,
      "eval_runtime": 0.0318,
      "eval_samples_per_second": 2233.238,
      "eval_steps_per_second": 31.454,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.7321340441703796,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6974,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.7075263857841492,
      "eval_runtime": 0.042,
      "eval_samples_per_second": 1688.825,
      "eval_steps_per_second": 23.786,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6948810815811157,
      "learning_rate": 4.5e-05,
      "loss": 0.6951,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7150639295578003,
      "eval_runtime": 0.035,
      "eval_samples_per_second": 2030.198,
      "eval_steps_per_second": 28.594,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9298628568649292,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.693,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.7189701795578003,
      "eval_runtime": 0.032,
      "eval_samples_per_second": 2217.226,
      "eval_steps_per_second": 31.229,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.4993817210197449,
      "learning_rate": 4.3e-05,
      "loss": 0.6952,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7181447148323059,
      "eval_runtime": 0.0321,
      "eval_samples_per_second": 2213.764,
      "eval_steps_per_second": 31.18,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 116563804707840.0,
      "train_loss": 0.6983035973140171,
      "train_runtime": 7.8848,
      "train_samples_per_second": 4026.726,
      "train_steps_per_second": 19.024
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 116563804707840.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}