trainer_state.json

{
  "best_metric": 0.9421561063965385,
  "best_model_checkpoint": "model_saves/deberta-large_lemon_5k_6_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": 0.9421561063965385,
      "eval_loss": 0.4133068919181824,
      "eval_runtime": 4.659,
      "eval_samples_per_second": 940.549,
      "eval_steps_per_second": 7.512,
      "step": 268
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9417088742047652,
      "eval_loss": 0.4411185383796692,
      "eval_runtime": 4.6385,
      "eval_samples_per_second": 944.711,
      "eval_steps_per_second": 7.546,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9407920482116303,
      "eval_loss": 0.47731733322143555,
      "eval_runtime": 4.6411,
      "eval_samples_per_second": 944.18,
      "eval_steps_per_second": 7.541,
      "step": 804
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9397745949753463,
      "eval_loss": 0.5340213179588318,
      "eval_runtime": 4.6436,
      "eval_samples_per_second": 943.673,
      "eval_steps_per_second": 7.537,
      "step": 1072
    },
    {
      "epoch": 4.0,
      "step": 1072,
      "total_flos": 1.6074262796304384e+16,
      "train_loss": 0.246939716944054,
      "train_runtime": 456.1068,
      "train_samples_per_second": 1128.157,
      "train_steps_per_second": 8.814
    }
  ],
  "max_steps": 4020,
  "num_train_epochs": 15,
  "total_flos": 1.6074262796304384e+16,
  "trial_name": null,
  "trial_params": null
}