trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 50,
  "global_step": 272,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.74,
      "learning_rate": 9.189154158168292e-05,
      "loss": 1.3896,
      "step": 50
    },
    {
      "epoch": 0.74,
      "eval_accuracy": 0.9740740740740741,
      "eval_loss": 0.7144090533256531,
      "eval_runtime": 8.8709,
      "eval_samples_per_second": 30.437,
      "eval_steps_per_second": 1.916,
      "step": 50
    },
    {
      "epoch": 1.47,
      "learning_rate": 7.019605024359474e-05,
      "loss": 0.5377,
      "step": 100
    },
    {
      "epoch": 1.47,
      "eval_accuracy": 0.9851851851851852,
      "eval_loss": 0.2982940077781677,
      "eval_runtime": 5.8832,
      "eval_samples_per_second": 45.893,
      "eval_steps_per_second": 2.89,
      "step": 100
    },
    {
      "epoch": 2.21,
      "learning_rate": 4.195020556092935e-05,
      "loss": 0.2683,
      "step": 150
    },
    {
      "epoch": 2.21,
      "eval_accuracy": 0.9851851851851852,
      "eval_loss": 0.19425520300865173,
      "eval_runtime": 5.8548,
      "eval_samples_per_second": 46.116,
      "eval_steps_per_second": 2.904,
      "step": 150
    },
    {
      "epoch": 2.94,
      "learning_rate": 1.631521781767214e-05,
      "loss": 0.181,
      "step": 200
    },
    {
      "epoch": 2.94,
      "eval_accuracy": 0.9888888888888889,
      "eval_loss": 0.15935532748699188,
      "eval_runtime": 5.8128,
      "eval_samples_per_second": 46.449,
      "eval_steps_per_second": 2.925,
      "step": 200
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.6054963006338742e-06,
      "loss": 0.1289,
      "step": 250
    },
    {
      "epoch": 3.68,
      "eval_accuracy": 0.9888888888888889,
      "eval_loss": 0.14862801134586334,
      "eval_runtime": 6.4022,
      "eval_samples_per_second": 42.173,
      "eval_steps_per_second": 2.655,
      "step": 250
    },
    {
      "epoch": 4.0,
      "step": 272,
      "total_flos": 3.335404697200558e+17,
      "train_loss": 0.4703441691749236,
      "train_runtime": 271.8221,
      "train_samples_per_second": 15.834,
      "train_steps_per_second": 1.001
    }
  ],
  "logging_steps": 50,
  "max_steps": 272,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 3.335404697200558e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}