trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.998504843008516,
  "global_step": 1440,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.42351716961498437,
      "eval_accuracy_0": 0.7588532883642496,
      "eval_accuracy_1": 0.046094750320102434,
      "eval_accuracy_2": 0.5985401459854015,
      "eval_loss": 0.80291348695755,
      "eval_runtime": 24.8435,
      "eval_samples_per_second": 77.364,
      "eval_steps_per_second": 77.364,
      "step": 480
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4011446409989594,
      "eval_accuracy_0": 0.7470489038785835,
      "eval_accuracy_1": 0.18309859154929578,
      "eval_accuracy_2": 0.3375912408759124,
      "eval_loss": 0.7574263215065002,
      "eval_runtime": 24.6349,
      "eval_samples_per_second": 78.019,
      "eval_steps_per_second": 78.019,
      "step": 960
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.40790842872008326,
      "eval_accuracy_0": 0.8026981450252951,
      "eval_accuracy_1": 0.18693982074263765,
      "eval_accuracy_2": 0.2956204379562044,
      "eval_loss": 0.7730965614318848,
      "eval_runtime": 24.5617,
      "eval_samples_per_second": 78.252,
      "eval_steps_per_second": 78.252,
      "step": 1440
    },
    {
      "epoch": 3.0,
      "step": 1440,
      "total_flos": 2507409848941452.0,
      "train_loss": 0.7177374945746527,
      "train_runtime": 2670.6543,
      "train_samples_per_second": 17.28,
      "train_steps_per_second": 0.539
    }
  ],
  "max_steps": 1440,
  "num_train_epochs": 3,
  "total_flos": 2507409848941452.0,
  "trial_name": null,
  "trial_params": null
}