trainer_state.json

{
  "best_metric": 0.9418598150694887,
  "best_model_checkpoint": "model_saves/deberta-large_lemon-spell_10k_3_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": 0.9418598150694887,
      "eval_loss": 0.42685073614120483,
      "eval_runtime": 4.6868,
      "eval_samples_per_second": 934.966,
      "eval_steps_per_second": 7.468,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.941429354084907,
      "eval_loss": 0.4456802308559418,
      "eval_runtime": 4.7077,
      "eval_samples_per_second": 930.816,
      "eval_steps_per_second": 7.435,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9406578785540983,
      "eval_loss": 0.48967695236206055,
      "eval_runtime": 4.6925,
      "eval_samples_per_second": 933.837,
      "eval_steps_per_second": 7.459,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9405237088965663,
      "eval_loss": 0.5445356965065002,
      "eval_runtime": 4.6909,
      "eval_samples_per_second": 934.149,
      "eval_steps_per_second": 7.461,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6308703116394496e+16,
      "train_loss": 0.24492808242342365,
      "train_runtime": 462.4104,
      "train_samples_per_second": 1112.778,
      "train_steps_per_second": 8.694
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6308703116394496e+16,
  "trial_name": null,
  "trial_params": null
}