GlorIA-1.3B-all/checkpoint-200/trainer_state.json

{
  "best_metric": 0.11806972324848175,
  "best_model_checkpoint": "results/Gloria_all/GlorIA-1.3B/checkpoint-100",
  "epoch": 2.3952095808383236,
  "eval_steps": 100,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5988023952095808,
      "grad_norm": 0.9138351082801819,
      "learning_rate": 2.627329192546584e-05,
      "loss": 0.1669,
      "step": 50
    },
    {
      "epoch": 1.1976047904191618,
      "grad_norm": 0.5801578760147095,
      "learning_rate": 2.1614906832298137e-05,
      "loss": 0.1129,
      "step": 100
    },
    {
      "epoch": 1.1976047904191618,
      "eval_accuracy": 0.9447699788056352,
      "eval_f1": 0.4304367402071139,
      "eval_loss": 0.11806972324848175,
      "eval_precision": 0.4337568058076225,
      "eval_recall": 0.42716711349419123,
      "eval_runtime": 9.8143,
      "eval_samples_per_second": 58.079,
      "eval_steps_per_second": 2.445,
      "step": 100
    },
    {
      "epoch": 1.7964071856287425,
      "grad_norm": 0.3607296645641327,
      "learning_rate": 1.6956521739130433e-05,
      "loss": 0.0915,
      "step": 150
    },
    {
      "epoch": 2.3952095808383236,
      "grad_norm": 0.4452211260795593,
      "learning_rate": 1.2298136645962733e-05,
      "loss": 0.063,
      "step": 200
    },
    {
      "epoch": 2.3952095808383236,
      "eval_accuracy": 0.9409051240493704,
      "eval_f1": 0.40241075567918405,
      "eval_loss": 0.18597252666950226,
      "eval_precision": 0.41811175337186895,
      "eval_recall": 0.387846291331546,
      "eval_runtime": 9.9454,
      "eval_samples_per_second": 57.313,
      "eval_steps_per_second": 2.413,
      "step": 200
    }
  ],
  "logging_steps": 50,
  "max_steps": 332,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 2009952955744776.0,
  "train_batch_size": 24,
  "trial_name": null,
  "trial_params": null
}