trainer_state.json

{
  "best_metric": 0.9402218271671194,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_10k_2_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": 0.9402218271671194,
      "eval_loss": 0.4399862587451935,
      "eval_runtime": 4.6696,
      "eval_samples_per_second": 938.406,
      "eval_steps_per_second": 7.495,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9394950748554881,
      "eval_loss": 0.476269006729126,
      "eval_runtime": 4.682,
      "eval_samples_per_second": 935.934,
      "eval_steps_per_second": 7.476,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9385558872527645,
      "eval_loss": 0.5165510773658752,
      "eval_runtime": 4.6792,
      "eval_samples_per_second": 936.489,
      "eval_steps_per_second": 7.48,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2171,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9395286172698711,
      "eval_loss": 0.5734798312187195,
      "eval_runtime": 4.6819,
      "eval_samples_per_second": 935.952,
      "eval_steps_per_second": 7.476,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6335737586712576e+16,
      "train_loss": 0.21135015185199565,
      "train_runtime": 463.5295,
      "train_samples_per_second": 1110.091,
      "train_steps_per_second": 8.673
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6335737586712576e+16,
  "trial_name": null,
  "trial_params": null
}