trainer_state.json

{
  "best_metric": 0.08488664031028748,
  "best_model_checkpoint": "/kaggle/working/trainer/checkpoint-354",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 354,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.388888888888889e-05,
      "loss": 0.7966,
      "step": 118
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.14867298305034637,
      "eval_runtime": 37.5027,
      "eval_samples_per_second": 0.693,
      "eval_steps_per_second": 0.347,
      "step": 118
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.2962962962962964e-05,
      "loss": 0.105,
      "step": 236
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.10169088840484619,
      "eval_runtime": 37.4736,
      "eval_samples_per_second": 0.694,
      "eval_steps_per_second": 0.347,
      "step": 236
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.2037037037037038e-05,
      "loss": 0.073,
      "step": 354
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.08488664031028748,
      "eval_runtime": 37.5454,
      "eval_samples_per_second": 0.692,
      "eval_steps_per_second": 0.346,
      "step": 354
    }
  ],
  "logging_steps": 500,
  "max_steps": 590,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1.3117494532453171e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}