trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.6929468512535095,
  "best_model_checkpoint": "tiny_bert_rand_20_v2_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": 1.0302760601043701,
      "learning_rate": 4.9e-05,
      "loss": 0.6975,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5492957746478874,
      "eval_loss": 0.696797788143158,
      "eval_runtime": 0.0353,
      "eval_samples_per_second": 2012.499,
      "eval_steps_per_second": 28.345,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8163024187088013,
      "learning_rate": 4.8e-05,
      "loss": 0.6959,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6929468512535095,
      "eval_runtime": 0.0309,
      "eval_samples_per_second": 2294.565,
      "eval_steps_per_second": 32.318,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.45361843705177307,
      "learning_rate": 4.7e-05,
      "loss": 0.6935,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.7046656012535095,
      "eval_runtime": 0.0309,
      "eval_samples_per_second": 2298.941,
      "eval_steps_per_second": 32.379,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.473766952753067,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6948,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.36619718309859156,
      "eval_loss": 0.7106072902679443,
      "eval_runtime": 0.041,
      "eval_samples_per_second": 1731.511,
      "eval_steps_per_second": 24.387,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6702704429626465,
      "learning_rate": 4.5e-05,
      "loss": 0.6912,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7076365351676941,
      "eval_runtime": 0.0418,
      "eval_samples_per_second": 1700.358,
      "eval_steps_per_second": 23.949,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.6611668467521667,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6932,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.7093421220779419,
      "eval_runtime": 0.0306,
      "eval_samples_per_second": 2323.424,
      "eval_steps_per_second": 32.724,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.44976961612701416,
      "learning_rate": 4.3e-05,
      "loss": 0.6926,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.7146235704421997,
      "eval_runtime": 0.0308,
      "eval_samples_per_second": 2307.563,
      "eval_steps_per_second": 32.501,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 116563804707840.0,
      "train_loss": 0.6940951801481701,
      "train_runtime": 7.801,
      "train_samples_per_second": 4070.003,
      "train_steps_per_second": 19.228
    }
  ],
  "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
}