trainer_state.json

{
  "best_metric": 0.9416753317903823,
  "best_model_checkpoint": "model_saves/deberta-v3-large_lemon_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": 0.9416753317903823,
      "eval_loss": 0.4090981185436249,
      "eval_runtime": 4.5178,
      "eval_samples_per_second": 969.94,
      "eval_steps_per_second": 7.747,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.941630608571205,
      "eval_loss": 0.42285215854644775,
      "eval_runtime": 4.5325,
      "eval_samples_per_second": 966.805,
      "eval_steps_per_second": 7.722,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9411721955746375,
      "eval_loss": 0.45529305934906006,
      "eval_runtime": 4.5281,
      "eval_samples_per_second": 967.729,
      "eval_steps_per_second": 7.729,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2934,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9404622144701975,
      "eval_loss": 0.4879423975944519,
      "eval_runtime": 4.5137,
      "eval_samples_per_second": 970.825,
      "eval_steps_per_second": 7.754,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.3559663664562176e+16,
      "train_loss": 0.287801854646028,
      "train_runtime": 470.8956,
      "train_samples_per_second": 1092.726,
      "train_steps_per_second": 8.537
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.3559663664562176e+16,
  "trial_name": null,
  "trial_params": null
}