trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6351392269134521,
  "best_model_checkpoint": "bert_base_rand_50_v1_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": 2.333170175552368,
      "learning_rate": 4.9e-05,
      "loss": 0.6617,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.624016108365367,
      "eval_loss": 0.6444294452667236,
      "eval_runtime": 3.5594,
      "eval_samples_per_second": 1534.789,
      "eval_steps_per_second": 6.181,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.46493399143219,
      "learning_rate": 4.8e-05,
      "loss": 0.6224,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6307889438037708,
      "eval_loss": 0.6351392269134521,
      "eval_runtime": 3.4128,
      "eval_samples_per_second": 1600.726,
      "eval_steps_per_second": 6.446,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.03251576423645,
      "learning_rate": 4.7e-05,
      "loss": 0.5549,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.628775398132894,
      "eval_loss": 0.6572652459144592,
      "eval_runtime": 3.4102,
      "eval_samples_per_second": 1601.955,
      "eval_steps_per_second": 6.451,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.183417320251465,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.45,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6512904997254256,
      "eval_loss": 0.7282273173332214,
      "eval_runtime": 3.4226,
      "eval_samples_per_second": 1596.138,
      "eval_steps_per_second": 6.428,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.451083660125732,
      "learning_rate": 4.5e-05,
      "loss": 0.3354,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6463481603514553,
      "eval_loss": 0.8410006761550903,
      "eval_runtime": 3.4099,
      "eval_samples_per_second": 1602.122,
      "eval_steps_per_second": 6.452,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.58699893951416,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2359,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6437854658612484,
      "eval_loss": 0.9604312777519226,
      "eval_runtime": 3.4179,
      "eval_samples_per_second": 1598.349,
      "eval_steps_per_second": 6.437,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 9.28459358215332,
      "learning_rate": 4.3e-05,
      "loss": 0.1663,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6421380194032583,
      "eval_loss": 1.1513041257858276,
      "eval_runtime": 3.434,
      "eval_samples_per_second": 1590.863,
      "eval_steps_per_second": 6.407,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.43237092553115475,
      "train_runtime": 1215.5227,
      "train_samples_per_second": 4308.558,
      "train_steps_per_second": 16.865
    }
  ],
  "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
}