checkpoint-3000/trainer_state.json

{
  "best_metric": 1.1077626943588257,
  "best_model_checkpoint": "/kaggle/output/checkpoint-1000",
  "epoch": 0.12222946544980444,
  "eval_steps": 1000,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 2.7777777777777777e-11,
      "loss": 1.1383,
      "step": 1
    },
    {
      "epoch": 0.04,
      "learning_rate": 2.7750000000000004e-08,
      "loss": 1.1424,
      "step": 1000
    },
    {
      "epoch": 0.04,
      "eval_accuracy": 0.32375249500998005,
      "eval_loss": 1.1077626943588257,
      "eval_runtime": 54.8633,
      "eval_samples_per_second": 91.318,
      "eval_steps_per_second": 11.428,
      "step": 1000
    },
    {
      "epoch": 0.08,
      "learning_rate": 5.5527777777777784e-08,
      "loss": 1.1244,
      "step": 2000
    },
    {
      "epoch": 0.08,
      "eval_accuracy": 0.33652694610778444,
      "eval_loss": 1.1080161333084106,
      "eval_runtime": 54.7384,
      "eval_samples_per_second": 91.526,
      "eval_steps_per_second": 11.454,
      "step": 2000
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.327777777777778e-08,
      "loss": 1.1228,
      "step": 3000
    },
    {
      "epoch": 0.12,
      "eval_accuracy": 0.34331337325349304,
      "eval_loss": 1.1084064245224,
      "eval_runtime": 54.7948,
      "eval_samples_per_second": 91.432,
      "eval_steps_per_second": 11.443,
      "step": 3000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 10000000,
  "num_train_epochs": 408,
  "save_steps": 1000,
  "total_flos": 6271065391104000.0,
  "trial_name": null,
  "trial_params": null
}