trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6458103060722351,
  "best_model_checkpoint": "tiny_bert_rand_10_v2_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6101438999176025,
      "learning_rate": 4.9e-05,
      "loss": 0.6655,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6201720666300568,
      "eval_loss": 0.6485046148300171,
      "eval_runtime": 1.6382,
      "eval_samples_per_second": 3334.701,
      "eval_steps_per_second": 13.429,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6267768144607544,
      "learning_rate": 4.8e-05,
      "loss": 0.6349,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6218195130880468,
      "eval_loss": 0.6458103060722351,
      "eval_runtime": 1.5547,
      "eval_samples_per_second": 3513.962,
      "eval_steps_per_second": 14.151,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.337428569793701,
      "learning_rate": 4.7e-05,
      "loss": 0.5941,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6231008603331503,
      "eval_loss": 0.6681211590766907,
      "eval_runtime": 1.5513,
      "eval_samples_per_second": 3521.628,
      "eval_steps_per_second": 14.182,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.5682601928710938,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5378,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6198059674171701,
      "eval_loss": 0.7004396915435791,
      "eval_runtime": 1.5729,
      "eval_samples_per_second": 3473.155,
      "eval_steps_per_second": 13.987,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.3654308319091797,
      "learning_rate": 4.5e-05,
      "loss": 0.4716,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6159619256818598,
      "eval_loss": 0.8119640946388245,
      "eval_runtime": 1.567,
      "eval_samples_per_second": 3486.19,
      "eval_steps_per_second": 14.039,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 5.089620590209961,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.4032,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6075416437854658,
      "eval_loss": 0.9388319253921509,
      "eval_runtime": 1.592,
      "eval_samples_per_second": 3431.577,
      "eval_steps_per_second": 13.819,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.411346435546875,
      "learning_rate": 4.3e-05,
      "loss": 0.3408,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5972908658246385,
      "eval_loss": 1.0600956678390503,
      "eval_runtime": 1.5662,
      "eval_samples_per_second": 3487.966,
      "eval_steps_per_second": 14.046,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 1.922715369529651e+16,
      "train_loss": 0.5211214391196647,
      "train_runtime": 400.0726,
      "train_samples_per_second": 13090.498,
      "train_steps_per_second": 51.241
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 1.922715369529651e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}