trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.6666666666666665,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "grad_norm": 1.6258652048649078,
      "learning_rate": 1e-05,
      "loss": 1.1358,
      "step": 1
    },
    {
      "epoch": 0.74,
      "grad_norm": 1.8824478010936796,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.0994,
      "step": 5
    },
    {
      "epoch": 0.89,
      "eval_loss": 0.9702978134155273,
      "eval_runtime": 31.3581,
      "eval_samples_per_second": 11.672,
      "eval_steps_per_second": 0.733,
      "step": 6
    },
    {
      "epoch": 1.48,
      "grad_norm": 0.8772658212738137,
      "learning_rate": 1e-05,
      "loss": 0.9662,
      "step": 10
    },
    {
      "epoch": 1.93,
      "eval_loss": 0.8241605758666992,
      "eval_runtime": 31.3409,
      "eval_samples_per_second": 11.678,
      "eval_steps_per_second": 0.734,
      "step": 13
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.6803706017116625,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 0.856,
      "step": 15
    },
    {
      "epoch": 2.67,
      "eval_loss": 0.8030304908752441,
      "eval_runtime": 31.4853,
      "eval_samples_per_second": 11.624,
      "eval_steps_per_second": 0.73,
      "step": 18
    },
    {
      "epoch": 2.67,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 0.945325043466356,
      "train_runtime": 388.5844,
      "train_samples_per_second": 3.289,
      "train_steps_per_second": 0.046
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3716488888320.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}