trainer_state.json

{
  "best_metric": 0.9424356265163967,
  "best_model_checkpoint": "model_saves/deberta-v3-large_spell_5k_1_p3/checkpoint-268",
  "epoch": 4.0,
  "global_step": 1072,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9424356265163967,
      "eval_loss": 0.4085525572299957,
      "eval_runtime": 4.5888,
      "eval_samples_per_second": 954.938,
      "eval_steps_per_second": 7.627,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9421449255917441,
      "eval_loss": 0.42521095275878906,
      "eval_runtime": 4.5951,
      "eval_samples_per_second": 953.618,
      "eval_steps_per_second": 7.617,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9411498339650488,
      "eval_loss": 0.44872042536735535,
      "eval_runtime": 4.6011,
      "eval_samples_per_second": 952.374,
      "eval_steps_per_second": 7.607,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2918,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.940238598374311,
      "eval_loss": 0.47877636551856995,
      "eval_runtime": 4.5961,
      "eval_samples_per_second": 953.411,
      "eval_steps_per_second": 7.615,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.35537608359936e+16,
      "train_loss": 0.2863717150332323,
      "train_runtime": 491.6983,
      "train_samples_per_second": 1046.495,
      "train_steps_per_second": 8.176
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.35537608359936e+16,
  "trial_name": null,
  "trial_params": null
}