checkpoint-200/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.7299,
      "step": 40
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6923056840896606,
      "eval_runtime": 0.73,
      "eval_samples_per_second": 97.265,
      "eval_steps_per_second": 12.329,
      "step": 40
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.2e-05,
      "loss": 0.6982,
      "step": 80
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.7026631832122803,
      "eval_runtime": 0.7338,
      "eval_samples_per_second": 96.754,
      "eval_steps_per_second": 12.265,
      "step": 80
    },
    {
      "epoch": 3.0,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.6972,
      "step": 120
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.4507042253521127,
      "eval_loss": 0.7005189657211304,
      "eval_runtime": 0.7321,
      "eval_samples_per_second": 96.982,
      "eval_steps_per_second": 12.293,
      "step": 120
    },
    {
      "epoch": 4.0,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.6992,
      "step": 160
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.6976626515388489,
      "eval_runtime": 0.7339,
      "eval_samples_per_second": 96.738,
      "eval_steps_per_second": 12.263,
      "step": 160
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.0,
      "loss": 0.699,
      "step": 200
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4647887323943662,
      "eval_loss": 0.6995509266853333,
      "eval_runtime": 0.7321,
      "eval_samples_per_second": 96.976,
      "eval_steps_per_second": 12.293,
      "step": 200
    }
  ],
  "max_steps": 200,
  "num_train_epochs": 5,
  "total_flos": 835377600768000.0,
  "trial_name": null,
  "trial_params": null
}