trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6356491446495056,
  "best_model_checkpoint": "bert_base_rand_20_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": 2.1182098388671875,
      "learning_rate": 4.9e-05,
      "loss": 0.663,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6260296540362438,
      "eval_loss": 0.6428149938583374,
      "eval_runtime": 3.4608,
      "eval_samples_per_second": 1578.557,
      "eval_steps_per_second": 6.357,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.5789700746536255,
      "learning_rate": 4.8e-05,
      "loss": 0.6206,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6364634816035145,
      "eval_loss": 0.6356491446495056,
      "eval_runtime": 3.4673,
      "eval_samples_per_second": 1575.561,
      "eval_steps_per_second": 6.345,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.8459675312042236,
      "learning_rate": 4.7e-05,
      "loss": 0.5501,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6342668863261944,
      "eval_loss": 0.6610056757926941,
      "eval_runtime": 3.452,
      "eval_samples_per_second": 1582.564,
      "eval_steps_per_second": 6.373,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.077686786651611,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4468,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6538531942156325,
      "eval_loss": 0.7094098925590515,
      "eval_runtime": 3.4656,
      "eval_samples_per_second": 1576.348,
      "eval_steps_per_second": 6.348,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.692441940307617,
      "learning_rate": 4.5e-05,
      "loss": 0.3292,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.653120995789859,
      "eval_loss": 0.8127928376197815,
      "eval_runtime": 3.468,
      "eval_samples_per_second": 1575.252,
      "eval_steps_per_second": 6.344,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.47494888305664,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2319,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6544023430349625,
      "eval_loss": 1.0191969871520996,
      "eval_runtime": 3.4618,
      "eval_samples_per_second": 1578.081,
      "eval_steps_per_second": 6.355,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.959715843200684,
      "learning_rate": 4.3e-05,
      "loss": 0.1649,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6503752516932089,
      "eval_loss": 1.1815861463546753,
      "eval_runtime": 3.4736,
      "eval_samples_per_second": 1572.715,
      "eval_steps_per_second": 6.333,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.4294875407468151,
      "train_runtime": 1227.5944,
      "train_samples_per_second": 4266.189,
      "train_steps_per_second": 16.699
    }
  ],
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}