trainer_state.json

{
  "best_global_step": 9,
  "best_metric": 0.7151739597320557,
  "best_model_checkpoint": "bert_base_rand_10_v2_wnli/checkpoint-9",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.5427184104919434,
      "learning_rate": 4.9e-05,
      "loss": 0.7358,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.36619718309859156,
      "eval_loss": 0.7191351652145386,
      "eval_runtime": 0.0597,
      "eval_samples_per_second": 1189.835,
      "eval_steps_per_second": 16.758,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.1575733423233032,
      "learning_rate": 4.8e-05,
      "loss": 0.6956,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.7296984791755676,
      "eval_runtime": 0.0619,
      "eval_samples_per_second": 1146.355,
      "eval_steps_per_second": 16.146,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8433818221092224,
      "learning_rate": 4.7e-05,
      "loss": 0.6954,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.7151739597320557,
      "eval_runtime": 0.077,
      "eval_samples_per_second": 921.528,
      "eval_steps_per_second": 12.979,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.4276347160339355,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.7021,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.7388314008712769,
      "eval_runtime": 0.0558,
      "eval_samples_per_second": 1271.989,
      "eval_steps_per_second": 17.915,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8135137557983398,
      "learning_rate": 4.5e-05,
      "loss": 0.6918,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.7654048204421997,
      "eval_runtime": 0.0579,
      "eval_samples_per_second": 1225.794,
      "eval_steps_per_second": 17.265,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9778173565864563,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6855,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2112676056338028,
      "eval_loss": 0.788017213344574,
      "eval_runtime": 0.0597,
      "eval_samples_per_second": 1189.308,
      "eval_steps_per_second": 16.751,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.9550083875656128,
      "learning_rate": 4.3e-05,
      "loss": 0.6912,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.8414391279220581,
      "eval_runtime": 0.061,
      "eval_samples_per_second": 1163.323,
      "eval_steps_per_second": 16.385,
      "step": 21
    },
    {
      "epoch": 8.0,
      "grad_norm": 1.7662990093231201,
      "learning_rate": 4.2e-05,
      "loss": 0.6828,
      "step": 24
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.18309859154929578,
      "eval_loss": 0.9093309640884399,
      "eval_runtime": 0.0586,
      "eval_samples_per_second": 1212.607,
      "eval_steps_per_second": 17.079,
      "step": 24
    },
    {
      "epoch": 8.0,
      "step": 24,
      "total_flos": 668302080614400.0,
      "train_loss": 0.6975290675957998,
      "train_runtime": 25.0344,
      "train_samples_per_second": 1268.254,
      "train_steps_per_second": 5.992
    }
  ],
  "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": 668302080614400.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}