checkpoint-3000/trainer_state.json

{
  "best_metric": 1.1018893718719482,
  "best_model_checkpoint": "/kaggle/output/checkpoint-3000",
  "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": 1.0632,
      "step": 1
    },
    {
      "epoch": 0.04,
      "learning_rate": 2.7750000000000004e-08,
      "loss": 1.1913,
      "step": 1000
    },
    {
      "epoch": 0.04,
      "eval_accuracy": 0.3269461077844311,
      "eval_loss": 1.1269203424453735,
      "eval_runtime": 20.7749,
      "eval_samples_per_second": 241.156,
      "eval_steps_per_second": 30.181,
      "step": 1000
    },
    {
      "epoch": 0.08,
      "learning_rate": 5.5527777777777784e-08,
      "loss": 1.1253,
      "step": 2000
    },
    {
      "epoch": 0.08,
      "eval_accuracy": 0.32914171656686625,
      "eval_loss": 1.102368712425232,
      "eval_runtime": 20.9681,
      "eval_samples_per_second": 238.935,
      "eval_steps_per_second": 29.903,
      "step": 2000
    },
    {
      "epoch": 0.12,
      "learning_rate": 8.327777777777778e-08,
      "loss": 1.1163,
      "step": 3000
    },
    {
      "epoch": 0.12,
      "eval_accuracy": 0.3461077844311377,
      "eval_loss": 1.1018893718719482,
      "eval_runtime": 20.86,
      "eval_samples_per_second": 240.173,
      "eval_steps_per_second": 30.058,
      "step": 3000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 10000000,
  "num_train_epochs": 408,
  "save_steps": 1000,
  "total_flos": 2090468376576000.0,
  "trial_name": null,
  "trial_params": null
}