run-4/checkpoint-1980/trainer_state.json

{
  "best_metric": 0.03980288097043214,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-multiclass-classification/run-4/checkpoint-1980",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1980,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.7575757575757576,
      "grad_norm": 3.6082186698913574,
      "learning_rate": 8.414672915202518e-07,
      "loss": 6.25,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.028051554207733132,
      "eval_loss": 6.15475606918335,
      "eval_runtime": 236.908,
      "eval_samples_per_second": 11.135,
      "eval_steps_per_second": 0.696,
      "step": 660
    },
    {
      "epoch": 1.5151515151515151,
      "grad_norm": 3.80741810798645,
      "learning_rate": 5.571878011417883e-07,
      "loss": 6.137,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.03335860500379075,
      "eval_loss": 6.058234214782715,
      "eval_runtime": 218.6669,
      "eval_samples_per_second": 12.064,
      "eval_steps_per_second": 0.755,
      "step": 1320
    },
    {
      "epoch": 2.2727272727272725,
      "grad_norm": 3.9423422813415527,
      "learning_rate": 2.729083107633249e-07,
      "loss": 6.066,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.03980288097043214,
      "eval_loss": 6.0241475105285645,
      "eval_runtime": 237.1607,
      "eval_samples_per_second": 11.123,
      "eval_steps_per_second": 0.696,
      "step": 1980
    }
  ],
  "logging_steps": 500,
  "max_steps": 1980,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 236510267834368.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.1257467818987151e-06,
    "num_train_epochs": 3,
    "per_device_train_batch_size": 16,
    "seed": 6
  }
}